From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752465Ab0HYXzR (ORCPT ); Wed, 25 Aug 2010 19:55:17 -0400 Received: from terminus.zytor.com ([198.137.202.10]:58989 "EHLO mail.zytor.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751997Ab0HYXzO (ORCPT ); Wed, 25 Aug 2010 19:55:14 -0400 Message-ID: <4C75AD28.8060101@zytor.com> Date: Wed, 25 Aug 2010 16:54:16 -0700 From: "H. Peter Anvin" User-Agent: Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.9.2.8) Gecko/20100806 Fedora/3.1.2-1.fc13 Thunderbird/3.1.2 MIME-Version: 1.0 To: LKML CC: Ingo Molnar , Thomas Gleixner , Arjan van de Ven , Jeremy Fitzhardinge , Andrew Morton Subject: RFC: remove or modify check_for_bios_corruption() Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org check_for_bios_corruption() was added to discover how many systems have BIOSes with problems corrupting low memory. The answer really is very simple... almost all. In fact, I have seen claims that Windows 7 doesn't use any memory below the 1 MiB point at all for this reason. As such, I have queued up a patch in -tip to reserve the low 64K for all BIOSes, with an option (suggested by Andrew) to adjust the reservation further. Since checking for corruption makes the memory unavailable anyway and since it is now established the memory is too unreliable to use, I am suggesting removing the check code completely. The other alternative is to have it optionally scan from the reserved threshold to some higher point, and only kick in if that gap is nonzero. That way it doesn't continually run for most users, but it might be useful to confirm corruption if we explicitly suspect it on some system. Thoughts? -hpa