Hi all,
I forward again a mail from Holger, it did not come through because of
strange configuration inside the GMD intranet.
Short WSE update: We talked with quite a lot developers about future
plans, especially about how to coordinate such a big project. I will
make a short summary soon and send that to all the people who
participated and some others as well who need to know that.
More will come after my diploma work (will be done at 13th of
December).
cu
Adrian
==================BEGIN FORWARDED MESSAGE==================
>Return-Path: <veit[ at ]borneo.gmd.de>
>Received: from mail.gmd.de (mail.gmd.de [129.26.8.90]) by warped.netlabs.org
> (Weasel v1.5) for <ktk[ at ]netlabs.org>; 06 Nov 2002 09:34:45
>Received: from borneo.gmd.de (borneo [129.26.12.20])
> by mail.gmd.de (8.9.3/8.9.3) with ESMTP id JAA10195
> for <ktk[ at ]netlabs.org>; Wed, 6 Nov 2002 09:34:45 +0100 (MET)
>Received: (from veit[ at ]localhost)
> by borneo.gmd.de (8.8.8/8.8.8) id JAA03862
> for ktk[ at ]netlabs.org; Wed, 6 Nov 2002 09:34:43 +0100 (MET)
>Date: Wed, 6 Nov 2002 09:34:43 +0100
>From: Holger Veit <holger.veit[ at ]ais.fhg.de>
>To: ktk[ at ]netlabs.org
>Subject: [holger.veit[ at ]ais.fhg.de: Re: [arch/i386os2] L4 Hurd]
>Message-ID: <[EMAIL PROTECTED]>
>Mime-Version: 1.0
>Content-Type: text/plain; charset=us-ascii
>Content-Disposition: inline
>User-Agent: Mutt/1.3.17i
>
Hi Adrian,
for some reason, the list manager at netlabs rejects my postings
to arch_i386os2, complaining that I am not subscribed. Could you please
have a look at that? The following is a rejected posting.
Holger
----- Forwarded message from Holger Veit <[EMAIL PROTECTED]> -----
Date: Wed, 6 Nov 2002 09:32:15 +0100
From: Holger Veit <holger.veit[ at ]ais.fhg.de>
To: arch_i386os2[ at ]netlabs.org
Cc: holger.veit[ at ]ais.fhg.de
Subject: Re: [arch/i386os2] L4 Hurd
User-Agent: Mutt/1.3.17i
In-Reply-To: <200211052042.000000EI[ at ]warped.netlabs.org>; from ktk[
at ]netlabs.org on Tue, Nov 05, 2002 at 08:42:40PM +0100
X-Spam-Status: No, hits=-4.5 required=5.0
tests=IN_REP_TO,USER_AGENT,REFERENCES
version=2.40
X-Spam-Level:
On Tue, Nov 05, 2002 at 08:42:40PM +0100, Adrian Gschwend wrote:
> Hi all,
>
> It's not really on topic but I found a short overview about the efforts
> of porting GNU/HURD to the L4 kernel architecture. I thought the
> GNU/HURD development is more or less frozen but if you find the right
> ressources you will notice that there are still some hardcore
> programmers working on it (do we know that? :-)
>
> I'm still fascinated by L4 and after my diploma work I try to compile a
> list of GNU/HURD ressources (basicly for me and a friend). The Kernel
> of "our" future OS/2 will be a topic (as you can see in Holgers mail)
> in the future.
>
> So far: http://www.8ung.at/shell/l4-port.html
I had long time ago looked at L4 (mainly because Jochen Liedtke was
once
a colleague of mine - now unfortunately dead) and proposed it as a
possible starting point. Infact, I had FIASCO, the clone of the Uni
Dresden, compiled with exclusively OS/2 tools (the old ELF tool suite
from XFree86/OS2) and booting a hello world example off a floppy.
Looking back, though, I found the kernel itself lean, as a microkernel
should be, but OTOH also rather intractable to base development on.
Frankly, I don't see much advantage to use it as a low level core where
to put an almost unmodified monolithic Linux kernel on top, as the
Dresden guys have done. The idea behind that is the "multiple
personalities" concept - take a microkernel and add certain
different OSses on it. Besides that I have never seen a microkernel
running more than a single OS at a time (due to lack of available
cores),
the implementations basically didn't need any real microkernel at all -
the Linux port on L4 basically does the following: it asks the
microkernel
for all available memory and then grab the CPU exclusively to maintain
all the resources by itself. The part that was saved was just a small
boot loader and a fraction of the paging memory management. This is not
how it was intended - you don't need a microkernel, you can do the same
on any OS with a decent VMWARE-like emulation.
Starting with a Linux kernel, as proposed on elfldr../doc/Targets.php,
as I understood it, will give us the advantage of a quick start, but
the whole monster is quite a large thing. I doubt it will be a good
idea to write ELF2OMF tools and ELF tools (we might try out my ELF gcc
stuff, BTW, as it already worked for fiasco) to pull the Linux
kernel over to OS/2. Linux is already a usable environment for
porting development. A first step might be, if you like, to make
the Linux HPFS support more bullet proof and setup a minimal Linux
system on a HPFS file system.
On the monster issue: I am sure that all the code in the Linux kernel
source tree is needed for some or other reason, but it is hardly
maintainable. Maybe we could extract some essential core, say a
minimum system with just Floppy, IDE, Console, HPFS support and
the kernel loader, to avoid getting lost in source files. The millions
of network and SCSI etc. drivers as well as odd legacy devices will
still be available through the kernel loader, if someone later
likes to add them. That core is what matters here to add the Dos*
API and an LX exe/dll loader.
Just some initial ideas.
Holger
--
Please update your tables to my new e-mail address:
holger.veit$ais.fhg.de (replace the '$' with '@' -- spam-protection)
----- End forwarded message -----
--
Please update your tables to my new e-mail address:
holger.veit$ais.fhg.de (replace the '$' with '@' -- spam-protection)
===================END FORWARDED MESSAGE===================
--
Adrian Gschwend
@ netlabs.org
ktk [a t] netlabs.org
-------
Free Software for OS/2 and eCS
http://www.netlabs.org
-----------
To unsubscribe yourself from this list, send the following message
to [EMAIL PROTECTED]
unsubscribe arch_i386os2
end
