From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754129AbYIBHgF (ORCPT ); Tue, 2 Sep 2008 03:36:05 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1752888AbYIBHfm (ORCPT ); Tue, 2 Sep 2008 03:35:42 -0400 Received: from e28smtp06.in.ibm.com ([59.145.155.6]:36314 "EHLO e28esmtp06.in.ibm.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1751021AbYIBHfl (ORCPT ); Tue, 2 Sep 2008 03:35:41 -0400 Message-ID: <48BCECB5.8070107@linux.vnet.ibm.com> Date: Tue, 02 Sep 2008 13:05:17 +0530 From: Balbir Singh Reply-To: balbir@linux.vnet.ibm.com Organization: IBM User-Agent: Thunderbird 2.0.0.16 (X11/20080724) MIME-Version: 1.0 To: Peter Zijlstra CC: Pavel Emelyanov , Andrew Morton , hugh@veritas.com, kamezawa.hiroyu@jp.fujitsu.com, menage@google.com, 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> <48BBAFDD.1000902@openvz.org> <48BBB326.3080505@linux.vnet.ibm.com> <1220275162.8426.61.camel@twins> In-Reply-To: <1220275162.8426.61.camel@twins> 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 Peter Zijlstra wrote: > On Mon, 2008-09-01 at 14:47 +0530, Balbir Singh wrote: >> Pavel Emelyanov wrote: >>> Balbir Singh wrote: >>>> This is a rewrite of a patch I had written long back to remove struct page >>>> (I shared the patches with Kamezawa, but never posted them anywhere else). >>>> I spent the weekend, cleaning them up for 2.6.27-rc5-mmotm (29 Aug 2008). >>>> >>>> I've tested the patches on an x86_64 box, I've run a simple test running >>>> under the memory control group and the same test running concurrently under >>>> two different groups (and creating pressure within their groups). I've also >>>> compiled the patch with CGROUP_MEM_RES_CTLR turned off. >>>> >>>> Advantages of the patch >>>> >>>> 1. It removes the extra pointer in struct page >>>> >>>> Disadvantages >>>> >>>> 1. It adds an additional lock structure to struct page_cgroup >>>> 2. Radix tree lookup is not an O(1) operation, once the page is known >>>> getting to the page_cgroup (pc) is a little more expensive now. >>> And besides, we also have a global lock, that protects even lookup >>> from this structure. Won't this affect us too much on bug-smp nodes? >> Sorry, not sure I understand. The lookup is done under RCU. Updates are done >> using the global lock. It should not be hard to make the radix tree per node >> later (as an iterative refinement). > > Or you could have a look at the concurrent radix tree, esp for dense > trees it can save a lot on lock bouncing. > > Latest code available at: > > http://programming.kicks-ass.net/kernel-patches/concurrent-pagecache/27-rc3/ I'll try them later. I see a lot of slab allocations/frees (expected). I am trying to experiment with alternatives to reduce them. -- Balbir