From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1758125AbYJIHQQ (ORCPT ); Thu, 9 Oct 2008 03:16:16 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1755688AbYJIHP7 (ORCPT ); Thu, 9 Oct 2008 03:15:59 -0400 Received: from E23SMTP04.au.ibm.com ([202.81.18.173]:44823 "EHLO e23smtp04.au.ibm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754609AbYJIHP6 (ORCPT ); Thu, 9 Oct 2008 03:15:58 -0400 Message-ID: <48EDAFA0.1090808@linux.vnet.ibm.com> Date: Thu, 09 Oct 2008 12:45:44 +0530 From: Balbir Singh Reply-To: balbir@linux.vnet.ibm.com Organization: IBM User-Agent: Thunderbird 2.0.0.16 (X11/20080715) MIME-Version: 1.0 To: KAMEZAWA Hiroyuki CC: "linux-mm@kvack.org" , "linux-kernel@vger.kernel.org" , "nishimura@mxp.nes.nec.co.jp" , "akpm@linux-foundation.org" , Dave Jones Subject: Re: [PATCH] memcg: update patch set v7 References: <20081007190121.d96e58a6.kamezawa.hiroyu@jp.fujitsu.com> In-Reply-To: <20081007190121.d96e58a6.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: > Hi, Andrew. please allow me to test under -mm if ok. > > This series is against the newest -mmotm(stamp-2008-10-02-16-17) > and I think ready-to-go. > > All comments are reflected. > (and CONFIG_CGROUP_MEM_RES_CTLR=n case is fixed.) > > Including following patches. > > [1/6] ... account swap cache under lock > [2/6] ... set page->mapping to be NULL before uncharge > [3/6] ... avoid to account not-on-LRU pages. > [4/6] ... optimize per cpu statistics on memcg. > [5/6] ... make page_cgroup->flags atomic. > [6/6] ... allocate page_cgroup at boot. > > I did tests I can. But I think patch 6/6 needs wider testers. > It has some dependency to configs/archs. > > (*) the newest mmotm needs some patches to be driven. Kamezawa-San, Thanks for the patchset. I would like to see these tested in -mm as well. The complaint that I am hearing from Fedora is that for them to enable the memory controller, they would like to see the struct page overhead go (for 32 bit systems that have 32 byte cachelines). This series helps us address that issue and helps with performance. -- Balbir