From mboxrd@z Thu Jan 1 00:00:00 1970 From: ebiederm@xmission.com (Eric W. Biederman) Subject: Re: [PATCH 10/10] x86, ACPI: default to reboot via ACPI (again) Date: Tue, 11 Nov 2008 12:29:20 -0800 Message-ID: References: <007e7d616a5d1c2e16ad627d03f8b97799445e71.1226032943.git.len.brown@intel.com> <200811080930.21462.arvidjaar@mail.ru> <200811081050.25477.arvidjaar@mail.ru> <20081108115956.GE8354@elte.hu> <4916B38B.7050905@redhat.com> <4916B4DB.90602@zytor.com> <20081110083938.GD22392@elte.hu> <4917F6A9.3020408@redhat.com> <20081110090241.GH22392@elte.hu> <4919CE3F.4010008@zytor.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Received: from out01.mta.xmission.com ([166.70.13.231]:60417 "EHLO out01.mta.xmission.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751668AbYKKUfc (ORCPT ); Tue, 11 Nov 2008 15:35:32 -0500 In-Reply-To: <4919CE3F.4010008@zytor.com> (H. Peter Anvin's message of "Tue, 11 Nov 2008 10:26:07 -0800") Sender: linux-acpi-owner@vger.kernel.org List-Id: linux-acpi@vger.kernel.org To: "H. Peter Anvin" Cc: Ingo Molnar , Avi Kivity , Andrey Borzenkov , Len Brown , linux-acpi@vger.kernel.org, Len Brown , Thomas Gleixner , Eduardo Habkost , Andrew Morton "H. Peter Anvin" writes: > Ingo Molnar wrote: >> >> well then we could insert CF9 to before the triple fault, and solve >> some of the problems as well, without unnecessary risks. >> >> This is a separate patch from ACPI reboot itself, naturally. >> > > This is a pretty good point. There is extremely low risk of being > something at port CF9 that is something other than the reset register. > It may not be there, but if it isn't, the worst thing that happened is > we confused a device right before reset. > > Even better if we condition the port CF9 write on existence a PCI bus > (or even more specifically PCI Configuration Method #1 or #2). Port CF9 > is extremely unlikely to exist on a machine without a PCI bus, and > extremely unlikely to conflict with anything else on a machine with a > PCI bus. Yes. What we are guarding against in practice is a write that causes the system to hang. Anything else and we will make it to the next reset method. Eric