Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1758750AbZCPQrs (ORCPT ); Mon, 16 Mar 2009 12:47:48 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1752408AbZCPQrj (ORCPT ); Mon, 16 Mar 2009 12:47:39 -0400 Received: from gir.skynet.ie ([193.1.99.77]:40367 "EHLO gir.skynet.ie" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752191AbZCPQri (ORCPT ); Mon, 16 Mar 2009 12:47:38 -0400 Date: Mon, 16 Mar 2009 16:47:36 +0000 From: Mel Gorman To: Christoph Lameter Cc: Linux Memory Management List , Pekka Enberg , Rik van Riel , KOSAKI Motohiro , Johannes Weiner , Nick Piggin , Linux Kernel Mailing List , Lin Ming , Zhang Yanmin , Peter Zijlstra Subject: Re: [PATCH 26/35] Use the per-cpu allocator for orders up to PAGE_ALLOC_COSTLY_ORDER Message-ID: <20090316164735.GM24293@csn.ul.ie> References: <1237196790-7268-1-git-send-email-mel@csn.ul.ie> <1237196790-7268-27-git-send-email-mel@csn.ul.ie> MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-15 Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.5.17+20080114 (2008-01-14) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 2059 Lines: 61 On Mon, Mar 16, 2009 at 12:26:07PM -0400, Christoph Lameter wrote: > On Mon, 16 Mar 2009, Mel Gorman wrote: > > > -static void free_hot_cold_page(struct page *page, int cold) > > +static void free_hot_cold_page(struct page *page, int order, int cold) > > { > > struct zone *zone = page_zone(page); > > struct per_cpu_pages *pcp; > > unsigned long flags; > > int clearMlocked = PageMlocked(page); > > > > + /* SLUB can return lowish-order compound pages that need handling */ > > + if (order > 0 && unlikely(PageCompound(page))) > > + if (unlikely(destroy_compound_page(page, order))) > > + return; > > + > > Isnt that also true for stacks and generic network objects ==- 8k? > I think they are vanilla high-order pages, not compound pages. > > again: > > cpu = get_cpu(); > > - if (likely(order == 0)) { > > + if (likely(order <= PAGE_ALLOC_COSTLY_ORDER)) { > > struct per_cpu_pages *pcp; > > + int batch; > > + int delta; > > > > pcp = &zone_pcp(zone, cpu)->pcp; > > + batch = max(1, pcp->batch >> order); > > local_irq_save(flags); > > if (!pcp->count) { > > - pcp->count = rmqueue_bulk(zone, 0, > > - pcp->batch, &pcp->list, migratetype); > > + delta = rmqueue_bulk(zone, order, batch, > > + &pcp->list, migratetype); > > + bulk_add_pcp_page(pcp, order, delta); > > if (unlikely(!pcp->count)) > > goto failed; > > The pcp adds a series of order N pages if an order N alloc occurs and the > queue is empty? > Nope, that would be bad. The calculation of batch is made above and is batch = max(1, pcp->batch >> order); so order is taken into account when deciding how many pages to allocate. -- Mel Gorman Part-time Phd Student Linux Technology Center University of Limerick IBM Dublin Software Lab -- 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/