From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from moutng.kundenserver.de ([212.227.126.187]:55907 "EHLO moutng.kundenserver.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755457AbXFPK1b (ORCPT ); Sat, 16 Jun 2007 06:27:31 -0400 From: Arnd Bergmann Subject: Re: [PATCH] Introduce compat_u64 and compat_s64 types Date: Sat, 16 Jun 2007 12:26:40 +0200 References: <200706150159.l5F1xNgM000459@hera.kernel.org> <200706151131.38429.arnd@arndb.de> <1181986686.25228.639.camel@pmac.infradead.org> In-Reply-To: <1181986686.25228.639.camel@pmac.infradead.org> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-15" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200706161226.40875.arnd@arndb.de> Sender: linux-arch-owner@vger.kernel.org To: David Woodhouse Cc: Benjamin Herrenschmidt , Linux Kernel Mailing List , Dave Airlie , linux-arch@vger.kernel.org, Andrew Morton List-ID: On Saturday 16 June 2007, David Woodhouse wrote: > Will GCC know that it needs to emit code to handle that (mis)alignment? I've tested this with gcc-4.0.3, and it does the right thing, which is to split a 4 byte aligned 64 bit load/store into two 32 bit accesses, if you pass -mstrict-align. > Preliminary tests show that it does load the value bytewise if we use > the 'packed' attribute structure on ppc64, but doesn't if we use > compat_u64. But then, I don't think it actually _needs_ to handle it n > ppc64 anyway, so maybe that's not such a good test. Right. Note that the behaviour of compat_u64 is the same as when you pass attribute((packed,aligned(4))) to the structure, where it also won't do byte accesses, and split the 64 bit access only if you pass -mstrict-align. Arnd <><