From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from canuck.infradead.org ([209.217.80.40]:51389 "EHLO canuck.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752132AbXFOMBA (ORCPT ); Fri, 15 Jun 2007 08:01:00 -0400 Subject: Re: [PATCH] Introduce compat_u64 and compat_s64 types From: David Woodhouse In-Reply-To: <200706151355.57464.ak@suse.de> References: <200706150159.l5F1xNgM000459@hera.kernel.org> <200706151128.39566.arnd@arndb.de> <200706151131.38429.arnd@arndb.de> <200706151355.57464.ak@suse.de> Content-Type: text/plain Date: Fri, 15 Jun 2007 13:00:36 +0100 Message-Id: <1181908836.25228.452.camel@pmac.infradead.org> Mime-Version: 1.0 Content-Transfer-Encoding: 7bit Sender: linux-arch-owner@vger.kernel.org To: Andi Kleen Cc: Arnd Bergmann , Benjamin Herrenschmidt , Linux Kernel Mailing List , Dave Airlie , linux-arch@vger.kernel.org, Andrew Morton List-ID: On Fri, 2007-06-15 at 13:55 +0200, Andi Kleen wrote: > On Friday 15 June 2007 11:31:37 Arnd Bergmann wrote: > > One common problem with 32 bit system call and ioctl emulation > > is the different alignment rules between i386 and 64 bit machines. > > A number of drivers work around this by marking the compat > > structures as 'attribute((packed))', which is not the right > > solution because it breaks all the non-x86 architectures that > > want to use the same compat code. > > Why does it break them? It should just make them a little slower. > > The network code requires unaligned accesses to work > anyways so if your architecture doesn't support them it is already > remotely crashable. alignof(uint64_t) is 8 on just about every 32-bit architecture except i386. Using __attribute__((packed)) for the 32-on-64 compat code is thus wrong on every 64-bit architecture except x86_64 and ia64. It's the _location_ which is wrong; the handling of unaligned loads is irrelevant (and Linux actually supports a bunch of architecture on which fixups are impossible now, btw). -- dwmw2