From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754637AbYLLUNZ (ORCPT ); Fri, 12 Dec 2008 15:13:25 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1752642AbYLLUNQ (ORCPT ); Fri, 12 Dec 2008 15:13:16 -0500 Received: from caramon.arm.linux.org.uk ([78.32.30.218]:40783 "EHLO caramon.arm.linux.org.uk" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751464AbYLLUNQ (ORCPT ); Fri, 12 Dec 2008 15:13:16 -0500 Date: Fri, 12 Dec 2008 20:12:39 +0000 From: Russell King To: Gerd Hoffmann Cc: Scott Lurndal , Matthew Wilcox , linux-kernel@vger.kernel.org, linux-arch@vger.kernel.org, Ulrich Drepper Subject: Re: [PATCH v2] Add preadv and pwritev system calls. Message-ID: <20081212201239.GE5774@flint.arm.linux.org.uk> Mail-Followup-To: Gerd Hoffmann , Scott Lurndal , Matthew Wilcox , linux-kernel@vger.kernel.org, linux-arch@vger.kernel.org, Ulrich Drepper References: <1229090440-32120-1-git-send-email-kraxel@redhat.com> <20081212152929.GM26095@parisc-linux.org> <494287D4.2070909@redhat.com> <20081212155113.GO26095@parisc-linux.org> <49428AFD.5090009@redhat.com> <20081212182929.GA9631@pendragon.3leafnetworks.com> <20081212190728.GC5774@flint.arm.linux.org.uk> <4942C1D0.1080009@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <4942C1D0.1080009@redhat.com> User-Agent: Mutt/1.4.2.1i Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Fri, Dec 12, 2008 at 08:56:00PM +0100, Gerd Hoffmann wrote: > Russell King wrote: > >> should be similar to pread/pwrite, e.g: > >> > >> int preadv(fd, iovec, iovec_size, offset) > > > > Yes, and that's easy for glibc to achieve. > > This hints the ABI problem exists at syscall level only. Is that > correct? So we can have > > preadv(fd, vec, vlen, off) > > argument ordering at app <-> glibc level and > > preadv(fd, vec, off, vlen) > > ordering at glibc <-> kernel (aka syscall) level and it works fine for > ARM + MIPS + PARISC? Fine for ARM - and yes, the user visible API should be changed from the BSD standard. I don't think anyone in this thread was suggesting that the user visible argument ordering should be any different from the original. Having it in a different order from *BSD at the libc visible interface is just crazy from the OS portability point of view. -- Russell King Linux kernel 2.6 ARM Linux - http://www.arm.linux.org.uk/ maintainer of: