From mboxrd@z Thu Jan  1 00:00:00 1970
From: Muli Ben-Yehuda <muli@il.ibm.com>
Subject: Re: [PATCH 4/4][VTD] vt-d specific files in KVM
Date: Tue, 10 Jun 2008 17:56:38 +0300
Message-ID: <20080610145638.GF6702@il.ibm.com>
References: <1FE6DD409037234FAB833C420AA843EC018831D5@orsmsx424.amr.corp.intel.com> <20080610102759.GG7307@il.ibm.com> <484E8EFC.10007@codemonkey.ws>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Cc: "Kay, Allen M" <allen.m.kay@intel.com>, kvm@vger.kernel.org,
	Amit Shah <amit.shah@qumranet.com>,
	Ben-Ami Yassour1 <BENAMI@il.ibm.com>,
	Avi Kivity <avik@qumranet.com>,
	Chris Wright <chrisw@redhat.com>,
	"Han, Weidong" <weidong.han@intel.com>
To: Anthony Liguori <anthony@codemonkey.ws>
Return-path: <kvm-owner@vger.kernel.org>
Received: from mtagate2.uk.ibm.com ([195.212.29.135]:16286 "EHLO
	mtagate2.uk.ibm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
	with ESMTP id S1752682AbYFJO4r (ORCPT <rfc822;kvm@vger.kernel.org>);
	Tue, 10 Jun 2008 10:56:47 -0400
Received: from d06nrmr1407.portsmouth.uk.ibm.com (d06nrmr1407.portsmouth.uk.ibm.com [9.149.38.185])
	by mtagate2.uk.ibm.com (8.13.8/8.13.8) with ESMTP id m5AEukOa134854
	for <kvm@vger.kernel.org>; Tue, 10 Jun 2008 14:56:46 GMT
Received: from d06av02.portsmouth.uk.ibm.com (d06av02.portsmouth.uk.ibm.com [9.149.37.228])
	by d06nrmr1407.portsmouth.uk.ibm.com (8.13.8/8.13.8/NCO v9.0) with ESMTP id m5AEujAh917652
	for <kvm@vger.kernel.org>; Tue, 10 Jun 2008 15:56:45 +0100
Received: from d06av02.portsmouth.uk.ibm.com (loopback [127.0.0.1])
	by d06av02.portsmouth.uk.ibm.com (8.12.11.20060308/8.13.3) with ESMTP id m5AEuimB017913
	for <kvm@vger.kernel.org>; Tue, 10 Jun 2008 15:56:45 +0100
Content-Disposition: inline
In-Reply-To: <484E8EFC.10007@codemonkey.ws>
Sender: kvm-owner@vger.kernel.org
List-ID: <kvm.vger.kernel.org>

On Tue, Jun 10, 2008 at 09:26:04AM -0500, Anthony Liguori wrote:

>> Checking against pfn_valid() isn't enough to differentiate between
>> RAM and MMIO areas. I think the consensus was that we also need to
>> check PageReserved(), i.e.,
>>
>> if (pfn_valid(pfn) && !PageReserved(pfn_to_page(pfn))) ...
>>   
>
> When checking the error return of gfn_to_pfn(), you should use
> is_error_pfn().  There's no need to differentiate mmio/ram pages in
> the code, the goal is just error checking.

I'd have to check the exact semantics of is_error_pfn() to see if it
fits, since strictly speaking what we are doing is not checking
pfn_to_page() for errors. We need to differentiate between gfns which
represent RAM (which needs to be mapped into the VT-d page tables) and
gfns which don't (e.g, slots which represent an MMIO region), which
should not be mapped in the VT-d page tables.

Cheers,
Muli