From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1758746AbYEFFfv (ORCPT ); Tue, 6 May 2008 01:35:51 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1753659AbYEFFfl (ORCPT ); Tue, 6 May 2008 01:35:41 -0400 Received: from e28smtp05.in.ibm.com ([59.145.155.5]:59890 "EHLO e28smtp05.in.ibm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753618AbYEFFfk (ORCPT ); Tue, 6 May 2008 01:35:40 -0400 Message-ID: <481FEDEF.9030901@linux.vnet.ibm.com> Date: Tue, 06 May 2008 11:04:39 +0530 From: Balbir Singh Reply-To: balbir@linux.vnet.ibm.com Organization: IBM User-Agent: Thunderbird 2.0.0.12 (X11/20080226) MIME-Version: 1.0 To: Andrew Morton CC: linux-mm@kvack.org, skumar@linux.vnet.ibm.com, yamamoto@valinux.co.jp, menage@google.com, lizf@cn.fujitsu.com, linux-kernel@vger.kernel.org, rientjes@google.com, xemul@openvz.org, kamezawa.hiroyu@jp.fujitsu.com Subject: Re: [-mm][PATCH 3/4] Add rlimit controller accounting and control References: <20080503213726.3140.68845.sendpatchset@localhost.localdomain> <20080503213814.3140.66080.sendpatchset@localhost.localdomain> <20080505152451.6dceec74.akpm@linux-foundation.org> In-Reply-To: <20080505152451.6dceec74.akpm@linux-foundation.org> 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 Andrew Morton wrote: > On Sun, 04 May 2008 03:08:14 +0530 > Balbir Singh wrote: > >> + if (res_counter_charge(&rcg->as_res, (mm->total_vm << PAGE_SHIFT))) > > I worry a bit about all the conversion between page-counts and byte-counts > in this code. > > For example, what happens if a process sits there increasing its rss with > sbrk(4095) or sbrk(4097) or all sorts of other scenarios? Do we get in a > situation in which the accounting is systematically wrong? > We already do all our accounting in pages for total_vm (field of mm_struct). task_vsize() for example multiplies PAGE_SIZE with total_vm before returning the result. > Worse, do we risk getting into that situation in the future, as unrelated > changes are made to the surrounding code? > I can't see that happening, but I'll look again and request reviewers to help me identify any such problems that can occur. > IOW, have we chosen the best, most maintainable representation for these > things? > That's a good question. From the sustenance point of view, resource counters have worked really well so far. Abstracting accounting and tracking from the controllers has been a good thing. One of the goals of the rlimit controller is to keep it open for extension, so that others can add their own control for other resources like mlock'ed pages. -- Warm Regards, Balbir Singh Linux Technology Center IBM, ISTL