Received: from nic.funet.fi (nic.funet.fi [128.214.248.6]) by odie.barnet.ac.uk (8.8.2/8.8.0) with ESMTP id DAA24401 for <willy@odie.fluff.org>; Fri, 16 May 1997 03:36:37 +0100
Received: from root@vger.rutgers.edu (port 50544 [128.6.190.2]) by nic.funet.fi with ESMTP id <751-10549>; Fri, 16 May 1997 05:35:59 +0300
Received: by vger.rutgers.edu id <971137-258>; Thu, 15 May 1997 22:33:18 -0400
Received: from sun3.nsfnet-relay.ac.uk ([128.86.8.50]) by vger.rutgers.edu with ESMTP id <971114-258>; Thu, 15 May 1997 22:32:46 -0400
Received: from bright.ecs.soton.ac.uk by sun3.nsfnet-relay.ac.uk with JANET SMTP (PP); Fri, 16 May 1997 03:34:26 +0100
Received: from by bright.ecs.soton.ac.uk; Thu, 15 May 97 22:43:34 BST
Received: from tanis.armlinux.org by caramon.armlinux.org; Thu, 15 May 1997 21:18:19 GMT
From: rmk92@ecs.soton.ac.uk
Message-Id: <133.199705152118@tanis.armlinux.org>
Subject: Re: ARM3/ARM6 compatibility
To: gilbertd@cs.man.ac.uk (David Alan Gilbert)
Date: Thu, 15 May 1997 22:18:08 +0100 (BST)
Cc: linux-arm@vger.rutgers.edu
In-Reply-To: <9705151511.AA13376@amu6.cs.man.ac.uk> from "David Alan Gilbert" at May 15, 97 04:11:36 pm
X-Phone: +44 (0)1737 360654
Reply-To: rmk92@ecs.soton.ac.uk
X-Mailer: ELM [version 2.4 PL24]
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Sender: owner-linux-arm@vger.rutgers.edu
Precedence: bulk
Status: RO
David Alan Gilbert writes:
> I'm wondering if the right thing to do is to move to fully 32 bit in
> the 2.1.x series when we move to ELF.
>
> The thing is the old machines (A400's etc) just aren't going to be
> able to cope with ELF; with only 128x32K pages and the extra page
> dirtying ELF is supposed to bring I don't think they stand a chance
> unless you are the lucky ones who have 8 or more MB in the old
> machines.
Under ELF, especially on the 386, the intermediate page between text and
data is mapped in twice - one after each other. This cannot be done under
the MEMC (since it uses a one to one physical -> logical mapping).
The only way that you could do it is to immediately create one extra dirty
page extra on top of the dirty pages caused by dynamic linking (if used).
I don't believe that it would be a good idea to do this to an already