Path: utzoo!utgpu!news-server.csri.toronto.edu!mailrus!ames!eos!shelby!ctt.bellcore.com!lunt
From: lunt@ctt.bellcore.com (Steve Lunt)
Newsgroups: comp.protocols.kerberos
Subject: (none)
Message-ID: <9005312120.AA10939@dduck.ctt.bellcore.com>
Date: 31 May 90 21:20:18 GMT
Sender: daemon@shelby.Stanford.EDU
Organization: The Internet
Lines: 64

From: lunt (Steve Lunt)
To: kerberos
Date: 31 May 1990  17:08 EDT
Subject: Re: Kerberos on 386i

Kerberos Users and Developers,

	Has anyone else run into this problem in trying to get Kerberos
up on a new architecture?  Please send any help that you can.  Thanx.

-- Steve

       Steven J. Lunt         |  lunt@ctt.bellcore.com  |  RRC-1L213
Computer Security Technology  |-------------------------|  444 Hoes Lane
          Bellcore            |     (201) 699-4244      |  Piscataway, NJ 08854

Background:

Article 298 of comp.protocols.kerberos:

> From: lunt (Steve Lunt)

        I'm having trouble bringing Kerberos up on my Sun 386i.  Has anyone
gotten Kerberos up on this architecture?  How about you guys at Project Athena?
Any 386 machines running Kerberos?  I understand you run a lot of IBM 
hardware :-).  Below are the defines from conf-bsd386i.h which I created for my
386i.  Kerberos compiled with this but doesn't run.  BTW, I do have it running
on SPARCstations.

#define BITS32
#define BIG
#define LSBFIRST
#define BSDUNIX
---------------------------------------------------------------------------
> From: Jerome H Saltzer <Saltzer@mit.edu>

I'm not an expert on this area, but I discovered when I tried to get
the DES library to run again on the 8086 that the verify command that
came with the DES library helped me narrow down the range of
possibilities.  Does the verify command give wrong answers?

                                Jerry Saltzer
---------------------------------------------------------------------------
> From: lunt (Steve Lunt)
 
Jerry,
        The verify program runs successfully, but I get runtime errors
from other Kerberos programs.  It was a while ago that I tried it, but I used
a debugger to find out where the problem was (I think the error message was
something like "Generic kerberos error").  I narrowed it down to the part of the 
code that decodes the responses from Kerberos (in mk_req.c and get_ad_tkt.c).
There is code in there that looks machine dependent, but I didn't get into
the details to find out what might be going on.  Do you know if there were
problems with this code on other hardware platforms?  Are there plans on cleaning
this code up in the next version?  I think a lot of care was taken to make the des libraries portable, but I don't see the same care taken in the krb libraries.
-- Steve
---------------------------------------------------------------------------
> From: Jerome H Saltzer <Saltzer@mit.edu>

That almost certainly indicates that the four DEFINE parameters are
set correctly for your environment; the problem is somewhere else,
well beyond my knowledge.  Hope someone else can help.

                                        Jerry