From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1757843AbYEIVDb (ORCPT ); Fri, 9 May 2008 17:03:31 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1753536AbYEIVDW (ORCPT ); Fri, 9 May 2008 17:03:22 -0400 Received: from netops-testserver-3-out.sgi.com ([192.48.171.28]:48705 "EHLO relay.sgi.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1753215AbYEIVDV (ORCPT ); Fri, 9 May 2008 17:03:21 -0400 Date: Fri, 9 May 2008 14:03:19 -0700 (PDT) From: Christoph Lameter X-X-Sender: clameter@schroedinger.engr.sgi.com To: Russ Anderson cc: linux-kernel@vger.kernel.org, linux-ia64@vger.kernel.org, Linus Torvalds , Andrew Morton , Tony Luck Subject: Re: [PATCH 3/3] ia64: Call migration code on correctable errors v3 In-Reply-To: <20080509151135.GD16523@sgi.com> Message-ID: References: <20080509151135.GD16523@sgi.com> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Fri, 9 May 2008, Russ Anderson wrote: > + local_irq_save(irq_flags); The page migration functions expect to be called in non atomic contexts since they use things like lock_page(). Can you just drop the irq disable? The spinlock should not be there either. Page migration serialize via the LRU. Pages that are to be migrated have to be taken off the LRU first. There is no danger of two threads trying to migrate the same page because the second one will not be able to take it off the LRU anymore.