From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from moutng.kundenserver.de ([212.227.126.186]:57267 "EHLO moutng.kundenserver.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750863AbXI2AH7 convert rfc822-to-8bit (ORCPT ); Fri, 28 Sep 2007 20:07:59 -0400 From: Arnd Bergmann Subject: Re: [COMPAT] Add compat_merge64 helper Date: Sat, 29 Sep 2007 01:52:12 +0200 References: <1191018832-18069-1-git-send-email-kyle@mcmartin.ca> <200709290138.24562.arnd@arndb.de> <20070929000131.GD25346@fattire.cabal.ca> In-Reply-To: <20070929000131.GD25346@fattire.cabal.ca> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 8BIT Content-Disposition: inline Message-Id: <200709290152.12842.arnd@arndb.de> Sender: linux-arch-owner@vger.kernel.org To: Kyle McMartin Cc: linux-arch@vger.kernel.org, linux-kernel@vger.kernel.org List-ID: On Saturday 29 September 2007, Kyle McMartin wrote: > > On Sat, Sep 29, 2007 at 01:38:23AM +0200, Arnd Bergmann wrote: > > 1. Byte order matches the order in which 64 bit arguments are split > >    in system call conventions on all platforms. > > I checked powerpc, sparc, and mips, which are (besides parisc) the only > 64-bit with 32-bit userspace big endian architectures that I could think > of offhand. A quick grep shows sh64 too... Paul? > s390 is big-endian as well, and while it does have really weird C calling conventions for 64 bit arguments in 32 bit mode, these do not affect the system call ABI, so we should be fine here. ia64 and x86_64 are obviously little-endian, and I double-checked that they work with this logic. Arnd <><