From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1759900AbZAWQH2 (ORCPT ); Fri, 23 Jan 2009 11:07:28 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1756255AbZAWQHQ (ORCPT ); Fri, 23 Jan 2009 11:07:16 -0500 Received: from courier.cs.helsinki.fi ([128.214.9.1]:57800 "EHLO mail.cs.helsinki.fi" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756188AbZAWQHP (ORCPT ); Fri, 23 Jan 2009 11:07:15 -0500 Subject: Re: [PATCH] SLUB: revert direct page allocator pass through From: Pekka Enberg To: Nick Piggin Cc: Christoph Lameter , yanmin_zhang@linux.intel.com, Andi Kleen , Matthew Wilcox , linux-kernel@vger.kernel.org, akpm@linux-foundation.org In-Reply-To: <200901240254.01552.nickpiggin@yahoo.com.au> References: <200901240237.10522.nickpiggin@yahoo.com.au> <1232725468.6094.97.camel@penberg-laptop> <200901240254.01552.nickpiggin@yahoo.com.au> Date: Fri, 23 Jan 2009 18:07:13 +0200 Message-Id: <1232726833.6094.110.camel@penberg-laptop> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 7bit X-Mailer: Evolution 2.22.3.1 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Sat, 2009-01-24 at 02:37 +1100, Nick Piggin wrote: > > > However, I don't know if netperf udp 4K over loopback is totally > > > realistic. Maybe real network drivers have different allocation patterns. > > > If I were you I wouldn't be too hasty to make big changes based on that > > > alone, if it could introduce regression in somewhere more important. On Saturday 24 January 2009 02:44:28 Pekka Enberg wrote: > > Yup, I'm putting it in linux-next to see if anything pops up. It's of > > course possible that it will introduce regression somewhere but seeing > > the hit SLUB is taking from the page allocator, I'm bound to think that > > performance increase claims of commit > > aadb4bc4a1f9108c1d0fbd121827c936c2ed4217 ("SLUB: direct pass through of > > page size or higher kmalloc requests") don't hold anymore. On Sat, 2009-01-24 at 02:54 +1100, Nick Piggin wrote: > Well what workload were they relating to? That could easily be retested. No idea. I'm not sure I even reviewed that patch as it went in.