Return-Path: <ak@linux.intel.com>
Received: from mga11.intel.com ([192.55.52.93])
	by Galois.linutronix.de with esmtps (TLS1.2:DHE_RSA_AES_256_CBC_SHA256:256)
	(Exim 4.80)
	(envelope-from <ak@linux.intel.com>)
	id 1fLyrG-0003Ul-34
	for speck@linutronix.de; Fri, 25 May 2018 00:37:18 +0200
Date: Thu, 24 May 2018 15:37:15 -0700
From: Andi Kleen <ak@linux.intel.com>
Subject: [MODERATED] Re: [PATCH v6 6/8] L1TFv6 3
Message-ID: <20180524223715.GT4486@tassilo.jf.intel.com>
References: <cover.1527195229.git.ak@linux.intel.com>
 <20180524210234.D1EA061155@crypto-ml.lab.linutronix.de>
 <ee48cd9e-3e08-7f57-aed9-6b2160e4963f@citrix.com>
MIME-Version: 1.0
In-Reply-To: <ee48cd9e-3e08-7f57-aed9-6b2160e4963f@citrix.com>
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: 8bit
To: speck@linutronix.de
List-ID: <speck.linutronix.de>

On Thu, May 24, 2018 at 11:30:35PM +0100, speck for Andrew Cooper wrote:
> On 24/05/2018 22:00, speck for Andi Kleen wrote:
> > From: Andi Kleen <ak@linux.intel.com>
> > Subject:  x86, l1tf: Report if too much memory for L1TF
> >  workaround
> >
> > If the system has more than MAX_PA/2 physical memory the
> > invert page workarounds don't protect the system against
> > the L1TF attack anymore, because an inverted physical address
> > will point to valid memory.
> 
> I realise I am late to this particular party (so apologies if it has
> been discussed before), but there is something which can be done here.
> 
> One option is to simply ignore any RAM above the PA/2 boundary.  If you

I had this in a earlier patch kit, but Linus didn't like it and pointed
out that the people who have that much memory will likely prefer
to use it. So now we just warn.

-Andi