From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754970AbYFKUac (ORCPT ); Wed, 11 Jun 2008 16:30:32 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1753303AbYFKU1j (ORCPT ); Wed, 11 Jun 2008 16:27:39 -0400 Received: from g1t0026.austin.hp.com ([15.216.28.33]:43841 "EHLO g1t0026.austin.hp.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756853AbYFKU1h (ORCPT ); Wed, 11 Jun 2008 16:27:37 -0400 Subject: Re: [PATCH -mm 11/24] pageflag helpers for configed-out flags From: Lee Schermerhorn To: Andrew Morton Cc: Rik van Riel , linux-kernel@vger.kernel.org, kosaki.motohiro@jp.fujitsu.com In-Reply-To: <20080611130145.2a202339.akpm@linux-foundation.org> References: <20080611184214.605110868@redhat.com> <20080611184339.493423475@redhat.com> <20080611130145.2a202339.akpm@linux-foundation.org> Content-Type: text/plain Organization: HP/OSLO Date: Wed, 11 Jun 2008 16:28:03 -0400 Message-Id: <1213216083.6436.66.camel@lts-notebook> Mime-Version: 1.0 X-Mailer: Evolution 2.22.1 Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Wed, 2008-06-11 at 13:01 -0700, Andrew Morton wrote: > On Wed, 11 Jun 2008 14:42:25 -0400 > Rik van Riel wrote: > > > Define proper false/noop inline functions for noreclaim page > > flags when !defined(CONFIG_NORECLAIM_LRU) > > I changed that to CONFIG_UNEVICTABLE_LRU. I noticed that the vmstat items [perhaps these will go away] still use "noreclaim". > > Did we agree that the presence of this config variable is undesirable? Not sure we did. It does add quite a bit of code that probably wouldn't be needed/wanted for, say, a laptop system or other small systems that still have an mmu. I initially made it configurable keeping in mind your speeches about not burdening smaller systems with larger system features, ... > If so, what's involved in making it go away? Just some busy work. I added a fair number of wrappers and such to minimize #ifdefs in the .c files. This resulted in quite a few such #ifdefs in the respective headers. Removing these should mostly involve removing the #ifdef and the #else case. Most of them that remain in .c files are for large blocks of code, mostly complete functions, and can just be removed. Let me know the consensus. I can send in a patch to remove them if it's agreed. Might want to wait until we're sure that this is the only barrier to merging, as the option might be useful during testing. Lee