From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1759420AbZEMLoD (ORCPT ); Wed, 13 May 2009 07:44:03 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1752806AbZEMLnp (ORCPT ); Wed, 13 May 2009 07:43:45 -0400 Received: from one.firstfloor.org ([213.235.205.2]:58848 "EHLO one.firstfloor.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750703AbZEMLno (ORCPT ); Wed, 13 May 2009 07:43:44 -0400 Date: Wed, 13 May 2009 13:49:07 +0200 From: Andi Kleen To: Minchan Kim Cc: Andi Kleen , KOSAKI Motohiro , LKML , linux-mm , Andrew Morton , Lee Schermerhorn , Minchan@firstfloor.org Subject: Re: [PATCH] Kconfig: CONFIG_UNEVICTABLE_LRU move into EMBEDDED submenu Message-ID: <20090513114907.GR19296@one.firstfloor.org> References: <20090513172904.7234.A69D9226@jp.fujitsu.com> <87r5ytl0nn.fsf@basil.nowhere.org> <2f11576a0905130418w1782f85j12cb938e92d256ff@mail.gmail.com> <20090513113817.GO19296@one.firstfloor.org> <28c262360905130441q8c904faq1d3e5152fada7a85@mail.gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <28c262360905130441q8c904faq1d3e5152fada7a85@mail.gmail.com> User-Agent: Mutt/1.4.2.1i Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Wed, May 13, 2009 at 08:41:21PM +0900, Minchan Kim wrote: > Hi, Andi. > > On Wed, May 13, 2009 at 8:38 PM, Andi Kleen wrote: > >> In past days, I proposed this. but Minchan found this config bloat kernel 7kb > >> and he claim embedded guys should have selectable chance. I agreed it. > > > > Well there's lots of code in the kernel and 7k doesn't seem worth bothering. > > If you just save two pages of memory somewhere you can save more. > > > >> Is this enough explanation? > > > > It's not a very good one. > > > > I would propose to just remove it or at least hide it completely > > and only make it dependent on CONFIG_MMU inside Kconfig. > > I thought this feature don't have a big impact on embedded. > At First, 7K is not important but as time goes by, it could be huge I don't follow. 7k is never huge, also not when time goes by. In general saving text size is not very fruitful compared to the savings you can get from optimizing dynamic memory allocation. Most of the memory waste is in dynamic allocation. That's easy to see because even on small systems there's much more (several magnitudes) dynamic memory than kernel text. -Andi -- ak@linux.intel.com -- Speaking for myself only.