[unios] An OS Foundation layer

[OJ Hickman]

Message-ID: <36E9DF8A.7BB8@peak.org>
Date: Fri, 12 Mar 1999 19:46:18 -0800
From: OJ Hickman <hickman1@peak.org>
X-Mailer: Mozilla 3.04 (Win95; I)
MIME-Version: 1.0
To: Pat Wendorf <beholder@ican.net>
Subject: Re: [unios] An OS Foundation layer
References: <36E962E6.7636@peak.org> <36E9781F.3E6BED70@ican.net>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Pat Wendorf wrote:

> This seems like a reasonable idea.  The only other idea along these
> lines that I had was a "boot partition" which contains a memory
> image of the run time system.  The image is loaded by a very simple
> boot loader, and it's started.  I think this method would support a
> kernel or nokernel based system, but would be difficult to modify
> (you'd need a program to create this image out of your current
> hardware set, not to mention the complexity of having a dedicated
> partition and such).  Other than the time to load x number of megs
> of hardware abstracts and system code into memory, I think it would
> be fairly quick.  Anyone got any other ideas?

I think this is done in some OSes (OS9?). They let you staticly link
components into a boot immage that can be put on a disk or burned into
a ROM.

The Foundation modual could be part of a far more complex, and versital,
boot seaquence by seting the Target to a modual loader that in turn
reads
a config text file and installs custom kernel moduals (scedular, IPC,
user mem manager . . .) then terminates from memory. Execution would
then
continue in a command interface or embeded app. Such a system could run
the spectrum from seting up a UNIX type monolithic kernel to an almost
non existant nanokernel for embeded systems.


_______________________________________________
"Imagination is more important than knowledge."
- Albert Einstein

Omer James Hickman - hickman1@peak.org - NullSpace@MailCity.com
http://members.tripod.com/~OJ_Hickman - updated 2/14/99
http://surf.to/NullSpace