Re: 2.6.39.1 immediately reboots/resets on EFI system

From: Matthew Garrett <mjg59@srcf.ucam.org>
To: Yinghai Lu <yinghai@kernel.org>
Cc: Jim Bos <jim876@xs4all.nl>,
	"H. Peter Anvin" <hpa@linux.intel.com>,
	Ingo Molnar <mingo@elte.hu>, Thomas Gleixner <tglx@linutronix.de>,
	Maarten Lankhorst <m.b.lankhorst@gmail.com>,
	Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
	Greg KH <gregkh@suse.de>,
	Linus Torvalds <torvalds@linux-foundation.org>,
	Andrew Morton <akpm@linux-foundation.org>
Subject: Re: 2.6.39.1 immediately reboots/resets on EFI system
Date: Wed, 8 Jun 2011 20:29:50 +0100	[thread overview]
Message-ID: <20110608192950.GA29235@srcf.ucam.org> (raw)
In-Reply-To: <4DEFCD07.5020505@kernel.org>

On Wed, Jun 08, 2011 at 12:27:03PM -0700, Yinghai Lu wrote:
> On 06/08/2011 12:23 PM, Matthew Garrett wrote:
> > On Wed, Jun 08, 2011 at 12:17:52PM -0700, Yinghai Lu wrote:
> >> we need to revert patch from mjg.
> >>
> >> From 916f676f8dc016103f983c7ec54c18ecdbb6e349 Mon Sep 17 00:00:00 2001
> >> From: Matthew Garrett <mjg@redhat.com>
> >> Date: Wed, 25 May 2011 09:53:13 -0400
> >> Subject: [PATCH] x86, efi: Retain boot service code until after switching to
> >>  virtual mode
> > 
> > That's not a long-term option, since we have no way to distinguish 
> > between a machine that requires boot services code to be mapped and a 
> > machine that doesn't. Reverting just breaks the former set again. 
> 
> what former? you can not fix some corner case by breaking most other cases.

All Dell laptops, all new Apples, some Lenovos, various Intel server 
platforms. That I've found so far.

> > We need to ensure that the kernel isn't overlapping boot services code.
> 
> bootloader will put kernel from 16M, and boot services are on those area already.

And we need to be able to map the boot services code, so we can't put 
the kernel on top of it.
-- 
Matthew Garrett | mjg59@srcf.ucam.org