From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S267615AbUHRXdt (ORCPT ); Wed, 18 Aug 2004 19:33:49 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S267614AbUHRXdt (ORCPT ); Wed, 18 Aug 2004 19:33:49 -0400 Received: from holomorphy.com ([207.189.100.168]:60858 "EHLO holomorphy.com") by vger.kernel.org with ESMTP id S267615AbUHRXdd (ORCPT ); Wed, 18 Aug 2004 19:33:33 -0400 Date: Wed, 18 Aug 2004 16:33:24 -0700 From: William Lee Irwin III To: "David S. Miller" Cc: pj@sgi.com, linux-kernel@vger.kernel.org Subject: Re: Does io_remap_page_range() take 5 or 6 args? Message-ID: <20040818233324.GT11200@holomorphy.com> Mail-Followup-To: William Lee Irwin III , "David S. Miller" , pj@sgi.com, linux-kernel@vger.kernel.org References: <20040818133348.7e319e0e.pj@sgi.com> <20040818205338.GF11200@holomorphy.com> <20040818135638.4326ca02.davem@redhat.com> <20040818210503.GG11200@holomorphy.com> <20040818143029.23db8740.davem@redhat.com> <20040818214026.GL11200@holomorphy.com> <20040818220001.GN11200@holomorphy.com> <20040818225915.GQ11200@holomorphy.com> <20040818161658.49aa8de3.davem@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20040818161658.49aa8de3.davem@redhat.com> User-Agent: Mutt/1.5.6+20040722i Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Wed, 18 Aug 2004 15:59:15 -0700 >> Given this, will a pfn suffice? On Wed, Aug 18, 2004 at 04:16:58PM -0700, David S. Miller wrote: > There is an error in the calculations. 16TB "RAM", means "RAM". > On many systems, a large chunk of the physical address space is > taken up by I/O areas, not real memory. > Such areas do not take up mem_map[] array space. > Regardless, I think an "unsigned long" page frame number is sufficient > for now. Don't even make the new type. Oh, virtualspace footprint of IO areas is far worse, as the convention is to direct map them into a single address space if they're ever used. Of course this convention is much more loosely established than e.g. struct page is for RAM. Some analogue of kmap_atomic() for such machines to multiplex virtualspace in interrupt context would help, but is unrelated to physical address passing issues. -- wli