From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754260AbYICHbb (ORCPT ); Wed, 3 Sep 2008 03:31:31 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1751466AbYICHbW (ORCPT ); Wed, 3 Sep 2008 03:31:22 -0400 Received: from E23SMTP04.au.ibm.com ([202.81.18.173]:41173 "EHLO e23smtp04.au.ibm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750913AbYICHbW (ORCPT ); Wed, 3 Sep 2008 03:31:22 -0400 Message-ID: <48BE3D43.7090903@linux.vnet.ibm.com> Date: Wed, 03 Sep 2008 13:01:15 +0530 From: Balbir Singh Reply-To: balbir@linux.vnet.ibm.com Organization: IBM User-Agent: Thunderbird 2.0.0.14 (X11/20080515) MIME-Version: 1.0 To: KAMEZAWA Hiroyuki CC: Nick Piggin , Andrew Morton , hugh@veritas.com, menage@google.com, xemul@openvz.org, linux-kernel@vger.kernel.org, linux-mm@kvack.org Subject: Re: [RFC][PATCH] Remove cgroup member from struct page References: <20080831174756.GA25790@balbir.in.ibm.com> <200809011656.45190.nickpiggin@yahoo.com.au> <20080901161927.a1fe5afc.kamezawa.hiroyu@jp.fujitsu.com> <200809011743.42658.nickpiggin@yahoo.com.au> <48BD0641.4040705@linux.vnet.ibm.com> <20080902190256.1375f593.kamezawa.hiroyu@jp.fujitsu.com> <48BD0E4A.5040502@linux.vnet.ibm.com> <20080902190723.841841f0.kamezawa.hiroyu@jp.fujitsu.com> <48BD119B.8020605@linux.vnet.ibm.com> <20080902195717.224b0822.kamezawa.hiroyu@jp.fujitsu.com> <48BD337E.40001@linux.vnet.ibm.com> <20080903123306.316beb9d.kamezawa.hiroyu@jp.fujitsu.com> In-Reply-To: <20080903123306.316beb9d.kamezawa.hiroyu@jp.fujitsu.com> Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org KAMEZAWA Hiroyuki wrote: > On Tue, 02 Sep 2008 18:07:18 +0530 > Balbir Singh wrote: >> I understand your concern and I am not trying to reduce memcg's performance - or >> add a fancy feature. I am trying to make memcg more friendly for distros. I see >> your point about the overhead. I just got back my results - I see a 4% overhead >> with the patches. Let me see if I can rework them for better performance. >> > Just an idea, by using atomic_ops page_cgroup patch, you can encode page_cgroup->lock > to page_cgroup->flags and use bit_spinlock(), I think. > (my new patch set use bit_spinlock on page_cgroup->flags for avoiding some race.) > > This will save extra 4 bytes. Exactly the next step I was thinking about (since we already use it, in the current form). Thanks for the suggestion! -- Balbir