From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <linux-kernel-owner+willy=40w.ods.org-S1750766AbWFFRsQ@vger.kernel.org>
Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand
	id S1750766AbWFFRsQ (ORCPT <rfc822;willy@w.ods.org>);
	Tue, 6 Jun 2006 13:48:16 -0400
Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1750830AbWFFRsQ
	(ORCPT <rfc822;linux-kernel-outgoing>);
	Tue, 6 Jun 2006 13:48:16 -0400
Received: from mtagate2.uk.ibm.com ([195.212.29.135]:56155 "EHLO
	mtagate2.uk.ibm.com") by vger.kernel.org with ESMTP
	id S1750766AbWFFRsP (ORCPT <rfc822;linux-kernel@vger.kernel.org>);
	Tue, 6 Jun 2006 13:48:15 -0400
Subject: Re: 2.6.17-rc5-mm1
From: Martin Schwidefsky <schwidefsky@de.ibm.com>
Reply-To: schwidefsky@de.ibm.com
To: Christoph Lameter <clameter@sgi.com>
Cc: Hugh Dickins <hugh@veritas.com>, Andrew Morton <akpm@osdl.org>,
       mbligh@google.com, apw@shadowen.org, mbligh@mbligh.org,
       linux-kernel@vger.kernel.org, ak@suse.de
In-Reply-To: <Pine.LNX.4.64.0606061023080.27969@schroedinger.engr.sgi.com>
References: <447DEF49.9070401@google.com>
	 <20060531140652.054e2e45.akpm@osdl.org> <447E093B.7020107@mbligh.org>
	 <20060531144310.7aa0e0ff.akpm@osdl.org> <447E104B.6040007@mbligh.org>
	 <447F1702.3090405@shadowen.org> <44842C01.2050604@shadowen.org>
	 <Pine.LNX.4.64.0606051137400.17951@schroedinger.engr.sgi.com>
	 <44848DD2.7010506@shadowen.org>
	 <Pine.LNX.4.64.0606051304360.18543@schroedinger.engr.sgi.com>
	 <44848F45.1070205@shadowen.org> <44849075.5070802@google.com>
	 <Pine.LNX.4.64.0606051325351.18717@schroedinger.engr.sgi.com>
	 <Pine.LNX.4.64.0606051334010.18717@schroedinger.engr.sgi.com>
	 <20060605135812.30138205.akpm@osdl.org>
	 <Pine.LNX.4.64.0606060537460.6045@blonde.wat.veritas.com>
	 <Pine.LNX.4.64.0606060923250.27550@schroedinger.engr.sgi.com>
	 <Pine.LNX.4.64.0606061805380.28806@blonde.wat.veritas.com>
	 <Pine.LNX.4.64.0606061023080.27969@schroedinger.engr.sgi.com>
Content-Type: text/plain
Organization: IBM Corporation
Date: Tue, 06 Jun 2006 19:48:16 +0200
Message-Id: <1149616097.29059.65.camel@localhost>
Mime-Version: 1.0
X-Mailer: Evolution 2.6.1 
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
X-Mailing-List: linux-kernel@vger.kernel.org

On Tue, 2006-06-06 at 10:39 -0700, Christoph Lameter wrote:
> Ok. So we have a mystical expression here that needs to be explained:
> 
> swp_type(pte_to_swp_entry(swp_entry_to_pte(swp_entry(~0UL,0))))
> 
> swp_entry(~0L,0) ->
> swp_entry(0xffffffffffffffff, 0) ->
> 0xfffffffffffffff << SWP_TYPE_SHIFT ->
> 0x1f
> 
> Ok we got the highest swap type and this encodes MAX_SWAPFILES_SHIFT and 
> therefore limits the number of swap 
> 
> But now we convert this to a arch specific pte. The hope is that the arch
> will mask off any unsupported bits? This is done by s390?

Yes.

> Then we convert the result back into a swap entry and take its type which 
> will hopefully give us the maximum number of swap devices supported by the 
> arch?

Yes.

> Would it not be better for an arch to explicitly tell us how many swap 
> devices are supported? Then we could make the swap_info array shorter and 
> get rid of this cryptic test.

Do you want to calculate the correct number for all architectures? Have
fun trying to decrypt the different pte formats.

> This test also creates a problem for the migration entries: It is really 
> not clear until runtime how many swap devices are supported so we cannot 
> take the last 2 for page migration. In order for page migration to work 
> all NUMA arches that do not support 32 swap devices need to define 
> CONFIG_MIGRATION=n. Seems that this is only s390?

s390 is not a NUMA system, but it does support 32 swap devices.

-- 
blue skies,
  Martin.

Martin Schwidefsky
Linux for zSeries Development & Services
IBM Deutschland Entwicklung GmbH

"Reality continues to ruin my life." - Calvin.