Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752582AbZIURmU (ORCPT ); Mon, 21 Sep 2009 13:42:20 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1752180AbZIURmS (ORCPT ); Mon, 21 Sep 2009 13:42:18 -0400 Received: from gir.skynet.ie ([193.1.99.77]:49457 "EHLO gir.skynet.ie" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751421AbZIURmS (ORCPT ); Mon, 21 Sep 2009 13:42:18 -0400 Date: Mon, 21 Sep 2009 18:42:23 +0100 From: Mel Gorman To: Daniel Walker Cc: Nick Piggin , Pekka Enberg , Christoph Lameter , heiko.carstens@de.ibm.com, sachinp@in.ibm.com, linux-kernel@vger.kernel.org, linux-mm@kvack.org, Tejun Heo , Benjamin Herrenschmidt Subject: Re: [PATCH 1/3] powerpc: Allocate per-cpu areas for node IDs for SLQB to use as per-node areas Message-ID: <20090921174223.GR12726@csn.ul.ie> References: <1253549426-917-1-git-send-email-mel@csn.ul.ie> <1253549426-917-2-git-send-email-mel@csn.ul.ie> <1253553472.9654.236.camel@desktop> MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-15 Content-Disposition: inline In-Reply-To: <1253553472.9654.236.camel@desktop> 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: 2479 Lines: 56 On Mon, Sep 21, 2009 at 10:17:52AM -0700, Daniel Walker wrote: > On Mon, 2009-09-21 at 17:10 +0100, Mel Gorman wrote: > > SLQB uses DEFINE_PER_CPU to define per-node areas. An implicit > > assumption is made that all valid node IDs will have matching valid CPU > > ids. In memoryless configurations, it is possible to have a node ID with > > no CPU having the same ID. When this happens, a per-cpu are is not > > created and the value of paca[cpu].data_offset is some random value. > > This is later deferenced and the system crashes after accessing some > > invalid address. > > > > This patch hacks powerpc to allocate per-cpu areas for node IDs that > > have no corresponding CPU id. This gets around the immediate problem but > > it should be discussed if there is a requirement for a DEFINE_PER_NODE > > and how it should be implemented. > > > > Signed-off-by: Mel Gorman > > --- > > arch/powerpc/kernel/setup_64.c | 20 ++++++++++++++++++++ > > 1 files changed, 20 insertions(+), 0 deletions(-) > > > > diff --git a/arch/powerpc/kernel/setup_64.c b/arch/powerpc/kernel/setup_64.c > > index 1f68160..a5f52d4 100644 > > --- a/arch/powerpc/kernel/setup_64.c > > +++ b/arch/powerpc/kernel/setup_64.c > > @@ -588,6 +588,26 @@ void __init setup_per_cpu_areas(void) > > paca[i].data_offset = ptr - __per_cpu_start; > > memcpy(ptr, __per_cpu_start, __per_cpu_end - __per_cpu_start); > > } > > +#ifdef CONFIG_SLQB > > + /* > > + * SLQB abuses DEFINE_PER_CPU to setup a per-node area. This trick > > + * assumes that ever node ID will have a CPU of that ID to match. > > + * On systems with memoryless nodes, this may not hold true. Hence, > > + * we take a second pass initialising a "per-cpu" area for node-ids > > + * that SLQB can use > > + */ > > Very trivial, but there's a little trailing whitespace in the first line > of the comment (checkpatch warns on it.) D'oh, will clean it up in the next revision if the substance of the patch is agreed upon. > You also spelled initializing wrong. It's spelt correctly for British English. -- 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/