[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Help-hurd] Mach Documentation from OSF

From: Farid Hajji
Subject: Re: [Help-hurd] Mach Documentation from OSF
Date: Thu, 28 Sep 2000 23:48:15 +0200

Hi Brent,

> Right -- and what's mk++?  Are these "improved" version of
> Mach?
I can't tell, since the sourcecode of mk++ was never released by OSF (AFAIK).
I'm trying to figure out the enhancements of mk++ from Mach, but the only
resource I'm having are the descriptions in the docs at opengroup's FTP site.

If we already talk about Mach, it may be interesting to compare osfmk
from the mklinux project and rtmach from keio/ntt to gnumach/oskit-mach.
gnumach is lacking e.g. norma/xmm (and of course the realtime semantics
of rtmach, though that would be overkill for now). OTOH, oskit-mach is
the right way to add Linux/FreeBSD driver code to Mach in a modular way.

It would be a very good idea to add norma/xmm to oskit-mach in the future.
NORMA-1 (as in cmu mach 3.0 mk83e) was utterly buggy/broken and NORMA-2
code was never released by OSF for misc. reasons. Building distributed
systems with NORMA may not be very efficient (as research in this area
showed), but it's at least an option. I wish we could have network wide
IPC and DSM in the Hurd, be it in gnumach and/or maybe as user-level
application libraries/servers.

Adding efficiency (in general as well as in the distributed case) may be
possible in the future, once the Hurd (a.k.a glibc sysdeps) is ported to L4.
Is anyone working on this project right now?


Farid Hajji -- Unix Systems and Network Admin | Phone: +49-2131-67-555
Broicherdorfstr. 83, D-41564 Kaarst, Germany  |
- - - - - - - - - - - - - - - - - - - - - - - + - - - - - - - - - - - -
Murphy's Law fails only when you try to demonstrate it, and thus succeeds.

reply via email to

[Prev in Thread] Current Thread [Next in Thread]