Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753606Ab1BPR2d (ORCPT ); Wed, 16 Feb 2011 12:28:33 -0500 Received: from smtp-out.google.com ([74.125.121.67]:56545 "EHLO smtp-out.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753537Ab1BPR2b (ORCPT ); Wed, 16 Feb 2011 12:28:31 -0500 DomainKey-Signature: a=rsa-sha1; c=nofws; d=google.com; s=beta; h=mime-version:in-reply-to:references:from:date:message-id:subject:to :cc:content-type; b=b8jDAjTAmFMX7wyCwPHpAmB/GXA5LoAq1Zw65Pu12S+qRD63ytemtplPmEo9g3qxfe 1MNcsNtm7r1SJrAlEKCw== MIME-Version: 1.0 In-Reply-To: References: <4D51ED26.8050809@cn.fujitsu.com> <4D539821.1090703@cn.fujitsu.com> <20110210205722.GB2600@redhat.com> <4D55EEBB.8060304@cn.fujitsu.com> <20110214180428.GF13097@redhat.com> <4D59E728.3030209@cn.fujitsu.com> <20110215142752.GC27296@redhat.com> <4D5B2C07.9050909@cn.fujitsu.com> From: Divyesh Shah Date: Wed, 16 Feb 2011 09:28:07 -0800 Message-ID: Subject: Re: [PATCH 5/6 v4] cfq-iosched: CFQ group hierarchical scheduling and use_hierarchy interface To: Gui Jianfeng Cc: Vivek Goyal , Jens Axboe , Shaohua Li , lkml , Chad Talbott Content-Type: text/plain; charset=ISO-8859-1 X-System-Of-Record: true Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1333 Lines: 37 On Wed, Feb 16, 2011 at 9:22 AM, Divyesh Shah wrote: > On Tue, Feb 15, 2011 at 5:44 PM, Gui Jianfeng > wrote: >> Hmm, it seems memcg aggregating stats in parent group. >> But do we really need to do that in kernel? I think it's easier to do it in >> userland, and it makes kernel much simpler. >> > > I would prefer having stats aggregated up the hierarchy. One trick we > used at Google earlier was to do lazy updates for most stats. So we Note that this was on ancient version of the blkio controller :). As Vivek mentioned, it may be ok to add the hierarchical accounting later. > would accumulate stats for a given timeslice and at the end of that > timeslice propagate those counts all the way up to the parent. For > device info like this, the update can be instantaneous since its not a > very frequent event. It would also be very useful to distinguish > between stats for the cgroup itself vs total values accumulated from > the subtree rooted at that cgroup like memcg does. > > -- > Thanks, > Divyesh > -- Thanks, Divyesh -- 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/