From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S267746AbUHRVko (ORCPT ); Wed, 18 Aug 2004 17:40:44 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S267796AbUHRVko (ORCPT ); Wed, 18 Aug 2004 17:40:44 -0400 Received: from holomorphy.com ([207.189.100.168]:55993 "EHLO holomorphy.com") by vger.kernel.org with ESMTP id S267746AbUHRVk3 (ORCPT ); Wed, 18 Aug 2004 17:40:29 -0400 Date: Wed, 18 Aug 2004 14:40:26 -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: <20040818214026.GL11200@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> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20040818143029.23db8740.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 14:05:03 -0700 William Lee Irwin III wrote: >> We should pass 64-bit values to remap_page_range() also, then. Or >> perhaps passing pfn's to both suffices, as it all has to be page >> aligned anyway. On Wed, Aug 18, 2004 at 02:30:29PM -0700, David S. Miller wrote: > Does not work on a system who has more physical address bits > than 32 + PAGE_SHIFT > Sparc32 does not fall into this category... but some other > might. All extant systems of this category I'm aware of are 32-bit kernels on 64-bit machines, which we don't really support. Also, the assumption that physical addresses are bounded by 1ULL << (BITS_PER_LONG + PAGE_SHIFT) is made broadly elsewhere, particularly in pfn_to_page() and the like. Making this assumption in remap_page_range() and io_remap_page_range() would save the overhead of passing additional arguments and/or passing 64-bit arguments on 32-bit machines. Using pgoff_t for pfn's may prove useful for such systems, but it's highly doubtful the kernel will ever be made clean for such or that they'll ever be brought to a usable state. -- wli