From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S267799AbUHRVg0 (ORCPT ); Wed, 18 Aug 2004 17:36:26 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S267746AbUHRVfu (ORCPT ); Wed, 18 Aug 2004 17:35:50 -0400 Received: from holomorphy.com ([207.189.100.168]:51897 "EHLO holomorphy.com") by vger.kernel.org with ESMTP id S267758AbUHRVfm (ORCPT ); Wed, 18 Aug 2004 17:35:42 -0400 Date: Wed, 18 Aug 2004 14:35:38 -0700 From: William Lee Irwin III To: Paul Jackson Cc: "David S. Miller" , linux-kernel@vger.kernel.org Subject: Re: Does io_remap_page_range() take 5 or 6 args? Message-ID: <20040818213538.GK11200@holomorphy.com> Mail-Followup-To: William Lee Irwin III , Paul Jackson , "David S. Miller" , linux-kernel@vger.kernel.org References: <20040818133348.7e319e0e.pj@sgi.com> <20040818135401.670f11bd.davem@redhat.com> <20040818141541.467e1e2d.pj@sgi.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20040818141541.467e1e2d.pj@sgi.com> User-Agent: Mutt/1.5.6+20040722i Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Dave Miller wrote: >> Each platform needs different args, unfortunately. On Wed, Aug 18, 2004 at 02:15:41PM -0700, Paul Jackson wrote: > Doesn't that make it kinda rough on the folks trying to write > arch-independent code, such as sound/core/pcm_native.c that I am > tripping over? > I can imagine a possible 'solution' something like (1) always passing > six args, and (2) providing arch-dependent macros to generate those last > two args, from some arch-generic value. > Just brainstorming ... This would not be useful. The extra argument is so that 64-bit values don't have to be passed. Keeping the minimal number of arguments and changing its meaning to be a pfn will suffice for the purposes here. -- wli