[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20081215185503.GA5620@alberich.amd.com>
Date: Mon, 15 Dec 2008 19:55:03 +0100
From: Andreas Herrmann <andreas.herrmann3@....com>
To: Andi Kleen <andi@...stfloor.org>
CC: Ingo Molnar <mingo@...e.hu>, Thomas Gleixner <tglx@...utronix.de>,
"H. Peter Anvin" <hpa@...or.com>, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 1/2] x86: don't enable MCE after suspend/resume when
system was booted with "nomce"
On Fri, Dec 12, 2008 at 08:08:28PM +0100, Andi Kleen wrote:
> Andreas Herrmann <andreas.herrmann3@....com> writes:
>
> > Impact: fix suspend/resume bug with MCE
> >
> > A suspend/resume cycle unconditionally enables MCE
> > for the boot CPU if MCE is compiled into the kernel.
> > I.e. even a system booted with "nomce" configures MCE for the boot CPU.
> >
> > This patch ensures that MCE is not turned on for systems booted with
> > "nomce".
>
> I have that fixed in a better way in
> http://git.kernel.org/?p=linux/kernel/git/mingo/linux-2.6-x86.git;a=commit;h=81c7af997dafbabad464ccefba89af2b247899da
Thanks for that pointer.
I didn't look at tip/x86/mce up to know. Is this .28 material?
My fix is intended for .28.
I didn't care about the lurking sysfs interface when booted with
"nomce". It's there and you could enable MCE with changing attributes
but then you are actively fiddling with MCE.
The fix is for the more severe problem that mce_init() is called on
resume and thus will enable MCE (setup cr4 and corresponding MSRs)
which shouldn't happen when booted with "nomce".
Regards,
Andreas
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists