From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S932410AbVJSIPF (ORCPT ); Wed, 19 Oct 2005 04:15:05 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S932411AbVJSIPF (ORCPT ); Wed, 19 Oct 2005 04:15:05 -0400 Received: from smtp.osdl.org ([65.172.181.4]:735 "EHLO smtp.osdl.org") by vger.kernel.org with ESMTP id S932410AbVJSIPD (ORCPT ); Wed, 19 Oct 2005 04:15:03 -0400 Date: Wed, 19 Oct 2005 01:14:20 -0700 From: Andrew Morton To: Andrea Arcangeli Cc: linux-kernel@vger.kernel.org, olof@austin.ibm.com Subject: Re: [PATCH] .text page fault SMP scalability optimization Message-Id: <20051019011420.032ccd6d.akpm@osdl.org> In-Reply-To: <20051019075255.GB30541@x30.random> References: <20051019075255.GB30541@x30.random> X-Mailer: Sylpheed version 1.0.4 (GTK+ 1.2.10; i386-redhat-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 Andrea Arcangeli wrote: > > We had a problem on ppc64 where with more than 4 threads a large system > wouldn't scale well while faulting in the .text (most of the time was > spent in the kernel despite it was an userland compute intensive app). > The reason is the useless overwrite of the same pte from all cpu. > > I fixed it this way (verified on an older kernel but the forward port is > almost identical). This will benefit all archs not just ppc64. How strange. Do you mena that all CPUs were entering the pagefault handler on behalf of the same pte all the time? That they're staying in lockstep? If so, there must be a bunch of page_table_lock contention too?