From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1761550AbXGKQqy (ORCPT ); Wed, 11 Jul 2007 12:46:54 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1751620AbXGKQqr (ORCPT ); Wed, 11 Jul 2007 12:46:47 -0400 Received: from smtp2.linux-foundation.org ([207.189.120.14]:54972 "EHLO smtp2.linux-foundation.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751706AbXGKQqr (ORCPT ); Wed, 11 Jul 2007 12:46:47 -0400 Date: Wed, 11 Jul 2007 09:46:00 -0700 From: Andrew Morton To: Andy Whitcroft Cc: linux-kernel@vger.kernel.org, Mel Gorman , Christoph Lameter , Peter Zijlstra Subject: Re: -mm merge plans -- lumpy reclaim Message-Id: <20070711094600.b2a525d5.akpm@linux-foundation.org> In-Reply-To: <4694A427.7070708@shadowen.org> References: <20070710013152.ef2cd200.akpm@linux-foundation.org> <46937DA2.4040302@shadowen.org> <4694A427.7070708@shadowen.org> X-Mailer: Sylpheed 2.4.1 (GTK+ 2.8.17; x86_64-unknown-linux-gnu) Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Wed, 11 Jul 2007 10:34:31 +0100 Andy Whitcroft wrote: > [Seems a PEBKAC occured on the subject line, resending lest it become a > victim of "oh thats spam".] > > Andy Whitcroft wrote: > > Andrew Morton wrote: > > > > [...] > >> lumpy-reclaim-v4.patch > >> have-kswapd-keep-a-minimum-order-free-other-than-order-0.patch > >> only-check-absolute-watermarks-for-alloc_high-and-alloc_harder-allocations.patch > >> > >> Lumpy reclaim. In a similar situation to Mel's patches. Stuck due to > >> general lack or interest and effort. > > > > The lumpy reclaim patches originally came out of work to support Mel's > > anti-fragmentation work. As such I think they have become somewhat > > attached to those patches. Whilst lumpy is most effective where > > placement controls are in place as offered by Mel's work, we see benefit > > from reduction in the "blunderbuss" effect when we reclaim at higher > > orders. While placement control is pretty much required for the very > > highest orders such as huge page size, lower order allocations are > > benefited in terms of lower collateral damage. > > > > There are now a few areas other than huge page allocations which can > > benefit. Stacks are still order 1. Jumbo frames want higher order > > contiguous pages for there incoming hardware buffers. SLUB is showing > > performance benefits from moving to a higher allocation order. All of > > these should benefit from more aggressive targeted reclaim, indeed I > > have been surprised just how often my test workloads trigger lumpy at > > order 1 to get new stacks. > > > > Truly representative work loads are hard to generate for some of these. > > Though we have heard some encouraging noises from those who can > > reproduce these problems. I'd expect that the main application for lumpy-reclaim is in keeping a pool of order-2 (say) pages in reserve for GFP_ATOMIC allocators. ie: jumbo frames. At present this relies upon the wakeup_kswapd(..., order) mechanism. How effective is this at solving the jumbo frame problem? (And do we still have a jumbo frame problem? Reports seems to have subsided)