*BSD News Article 33051


Return to BSD News archive

Path: sserve!newshost.anu.edu.au!munnari.oz.au!constellation!news.uoknor.edu!ns1.nodak.edu!netnews.nwnet.net!news.clark.edu!spool.mu.edu!howland.reston.ans.net!europa.eng.gtefsd.com!MathWorks.Com!news.kei.com!bloom-beacon.mit.edu!ai-lab!life.ai.mit.edu!mycroft
From: mycroft@duality.gnu.ai.mit.edu (Charles M. Hannum)
Newsgroups: comp.os.386bsd.misc
Subject: Re: 4.4-lite?
Date: 20 Jul 1994 08:35:42 GMT
Organization: MIT Artificial Intelligence Lab
Lines: 47
Message-ID: <MYCROFT.94Jul20043542@duality.gnu.ai.mit.edu>
References: <2vgvc7$3tg@spruce.cic.net> <3097eh$m2h@pdq.coe.montana.edu>
	<Bs2yi5F.dysonj@delphi.com>
	<michaelv.774429899@ponderous.cc.iastate.edu>
	<30fh47$94i@pdq.coe.montana.edu>
NNTP-Posting-Host: duality.ai.mit.edu
In-reply-to: nate@bsd.coe.montana.edu's message of 19 Jul 1994 03:25:59 GMT


In article <30fh47$94i@pdq.coe.montana.edu> nate@bsd.coe.montana.edu
(Nate Williams) writes:

   The 4.4 'core' platform was the HP series, which is NOT a little
   endian machine.  Therefore, the plethora of endian-ness kernel bugs
   you allude to in your posting simply shouldn't be in 4.4 due to the
   x86 influence if 4.4 ran on HP's.  (Which is did.)

That statement doesn't make much sense.  The endian problem were with
little-endian machines.  Almost all late 4.4-Lite development was done
on hp300s and SPARCs; the two little-endian ports, pmax and i386,
didn't really work.  The i386 stuff didn't even come close to
compiling.

And I can assure you there were enough endian problems.  The inet code
thought my machine had a multicast address, and wouldn't respond to
any TCP open sequences initiated by another host.  Restores of dumps
from little-endian machines with old file systems didn't work.  LFS
was just totally hosed.  (It's still broken in NetBSD, but not as
badly.)  The list goes on.

All the ones I know of are fixed in NetBSD.  I'm running the code on
an i386 and a hp300 right now.

I've been there; I've done it.  I speak with authority on the subject.
Do you?

   However, since it was done speedily by the NetBSD folks those bugs
   can still be referenced by the FreeBSD folks or done completely
   from scratch.

Since I've been told that almost all of my patches (and most of them
*were* mine, since it was the file system and network code that had
problems, and I did that integration almost entirely myself) will be
in the second release of 4.4-Lite, you'd be a bit foolish to redo all
of it.

And yes, FreeBSD could simply pull them in, but I have to wonder why
people wouldn't rather use the system where the work is actually
getting done.

--
- Charles Hannum
  NetBSD group
  Working ports: i386, hp300, amiga, sparc, mac68k, pc532.
  In progress: pmax, sun3.