From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S272920AbTHKRai (ORCPT ); Mon, 11 Aug 2003 13:30:38 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S272898AbTHKR15 (ORCPT ); Mon, 11 Aug 2003 13:27:57 -0400 Received: from mx2.elte.hu ([157.181.151.9]:48618 "EHLO mx2.elte.hu") by vger.kernel.org with ESMTP id S272821AbTHKRYM (ORCPT ); Mon, 11 Aug 2003 13:24:12 -0400 Date: Mon, 11 Aug 2003 19:19:43 +0200 (CEST) From: Ingo Molnar Reply-To: Ingo Molnar To: Manfred Spraul Cc: linux-kernel@vger.kernel.org, Andrew Morton , Alexander Viro Subject: Re: 4GB+DEBUG_PAGEALLOC oopses with 2.6.0-test3-mm1 In-Reply-To: <3F37BED1.405@colorfullife.com> Message-ID: References: <3F361F5E.10106@colorfullife.com> <3F37B4B7.9010108@colorfullife.com> <3F37BED1.405@colorfullife.com> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Mon, 11 Aug 2003, Manfred Spraul wrote: > > this is not a bug technically, unless the mount options are in the > > last linearly mapped page. It is a bug to copy those unallocated > > bytes, but they do not get to relied upon. Note that the non-4G code > > copies them just as much. > > Or unless there are holes in the memory map, or unless pages were > unmapped from the kernel linear mapping for GART. > IMHO the currect code is unacceptable. IMHO the mount-options API is quite unclean - since when do we guarantee -EFAULT semantics? > It is possible to use direct_copy_ instead of memcpy for fs==KERNEL_DS > in mm/usercopy.c? yes. But i still think we should fix the API too. Ingo