From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754056AbXDVFA4 (ORCPT ); Sun, 22 Apr 2007 01:00:56 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1754082AbXDVFA4 (ORCPT ); Sun, 22 Apr 2007 01:00:56 -0400 Received: from ausmtp04.au.ibm.com ([202.81.18.152]:46328 "EHLO ausmtp04.au.ibm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754056AbXDVFAz (ORCPT ); Sun, 22 Apr 2007 01:00:55 -0400 Message-ID: <462AEBFE.80500@linux.vnet.ibm.com> Date: Sun, 22 Apr 2007 10:30:46 +0530 From: Balbir Singh Reply-To: balbir@linux.vnet.ibm.com Organization: IBM User-Agent: Thunderbird 1.5.0.10 (X11/20070306) MIME-Version: 1.0 To: Andrew Morton CC: Balbir Singh , nagar@watson.ibm.com, jlan@sgi.com, linux-kernel@vger.kernel.org Subject: Re: [PATCH -mm] Taskstats fix the structure members alignment issue References: <20070420164341.14718.5013.sendpatchset@balbir-laptop> <20070420121543.6a1a8426.akpm@linux-foundation.org> <462A0AA9.8000304@linux.vnet.ibm.com> <20070421123307.5dc2566f.akpm@linux-foundation.org> In-Reply-To: <20070421123307.5dc2566f.akpm@linux-foundation.org> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Andrew Morton wrote: > On Sat, 21 Apr 2007 18:29:21 +0530 Balbir Singh wrote: > >>>> The patch adds an __attribute__((aligned(8))) to the >>>> taskstats structure members so that 32 bit applications using taskstats >>>> can work with a 64 bit kernel. >>> But there might be 32-bit applications out there which are using the >>> present wrong structure? >>> >>> otoh, I assume that those applications would be using taskstats.h and would >>> hence encounter this bug and we would have heard about it, is that correct? >>> >> Yes, correct. >> >>> otoh^2, 32-bit applications running under 32-bit kernels will presently be >>> functioning correctly, and your change will require that those applications >>> be recompiled, I think? >>> >> Yes, correct. They would be broken with this fix. We could bump up the >> version TASKSTATS_VERSION to 4. Would you like a new patch the version >> bumped up? > > I can do that. Thanks > >>> This patch looks like 2.6.20 and 2.6.21 material, but very carefully... >> Yes, 2.6.20 and 2.6.21 sound correct. > > OK. I guess we have little choice but to slam it in asap, with a 2.6.20.x backport > before too many people start using the old interface. Thanks, again! -- Balbir Singh Linux Technology Center IBM, ISTL