From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S965254AbXCTQdb (ORCPT ); Tue, 20 Mar 2007 12:33:31 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S965365AbXCTQdb (ORCPT ); Tue, 20 Mar 2007 12:33:31 -0400 Received: from ecfrec.frec.bull.fr ([129.183.4.8]:35507 "EHLO ecfrec.frec.bull.fr" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S965338AbXCTQd2 (ORCPT ); Tue, 20 Mar 2007 12:33:28 -0400 Message-ID: <46000C8B.1090600@bull.net> Date: Tue, 20 Mar 2007 17:32:11 +0100 From: Pierre Peiffer User-Agent: Thunderbird 1.5.0.10 (X11/20070302) MIME-Version: 1.0 To: Peter Zijlstra Cc: mingo@elte.hu, drepper@redhat.com, linux-kernel@vger.kernel.org, Jean-Pierre Dion Subject: Re: [PATCH 2.6.21-rc3-mm2 3/4] futex_requeue_pi optimization References: <20070313095203.154246923@bull.net> <20070313095548.003786249@bull.net> <1174040046.7124.27.camel@twins> <45FFFE72.6000300@bull.net> <1174406134.16478.37.camel@twins> In-Reply-To: <1174406134.16478.37.camel@twins> X-MIMETrack: Itemize by SMTP Server on ECN002/FR/BULL(Release 5.0.12 |February 13, 2003) at 20/03/2007 17:35:25, Serialize by Router on ECN002/FR/BULL(Release 5.0.12 |February 13, 2003) at 20/03/2007 17:35:27, Serialize complete at 20/03/2007 17:35:27 Content-Transfer-Encoding: 8bit Content-Type: text/plain; charset=UTF-8; format=flowed Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Peter Zijlstra a écrit : > > Unfortunately not, nonlinear vmas don't have a linear relation between > address and offset. What you would need to do is do a linear walk of the > page tables. But even that might not suffice if nonlinear vmas may form > a non-injective, surjective mapping. > > /me checks.. > > Hmm, yes that seems valid, so in general, this reverse mapping does not > uniquely exist for non-linear vmas. :-( > > What to do... disallow futexes in nonlinear mappings, > store the address in the key? << That seems to be the only solution... :-/ > > the vma_prio_tree would be able to give all vmas associated with a > mapping. > Thanks for your help. -- Pierre