*BSD News Article 41605


Return to BSD News archive

Newsgroups: comp.os.386bsd.questions
Path: sserve!newshost.anu.edu.au!harbinger.cc.monash.edu.au!msunews!uwm.edu!spool.mu.edu!howland.reston.ans.net!pipex!uunet!easix!news.uni-stuttgart.de!rz.uni-karlsruhe.de!stepsun.uni-kl.de!sun.rhrk.uni-kl.de!weber
From: weber@rhrk.uni-kl.de (Christoph Weber-Fahr [KIT])
Subject: Re: Installation problem(s)
Message-ID: <1995Jan26.144147.20673@rhrk.uni-kl.de>
Organization: University of Kaiserslautern, Germany
References: <1995Jan24.144419.24484@rhrk.uni-kl.de> <3g46fm$llc@agate.berkeley.edu>
Date: Thu, 26 Jan 1995 14:41:47 GMT
Lines: 93

Hello,

jkh@violet.berkeley.edu (Jordan K. Hubbard) writes:
>In article <1995Jan24.144419.24484@rhrk.uni-kl.de>,
>Christoph Weber-Fahr [KIT] <weber@rhrk.uni-kl.de> wrote:
>>- My main problem is that I try to run top. Both the precompiled binary
>>  and a new compiled top from the ports collection get killed by a signal 8
>>  floating point exception. Assuming the generic kernel lacks the 
>>  math-emulation I tried to compile a new one, but that didn't change 
>>  anything (BTW - are you supposed to get a few warnings when compiling
>>  a kernel ?) Or am I the only guy left without a coprocessor ?

>No idea, really.  Top works for me!  Are you running it as root or a normal
>user?  

I run it as root. 

> Perhaps your /dev/*mem permissions are overly restrictive.

I didn't change them. crw-r-----

I get a message on the console: 
    dido kernel: pid xxx: top: uid 0: exited on signal 8

and the program exits without a core dump, but with the line:
   floating exception

BTW - another problem is that still the reboot doesn't work (It also didn't so
with 1.1, but I seem to remember someone writing here that this problem was 
solved). The attempt to reboot hangs the system, after the string 'Rebooting.'.
Is there anything I can do about that ? It shouldn't be that hard to force a 
cold boot anyway... even OS/2 manages that without obvious difficulties.

>>- I can not write a new Bootstrap loader. disklabel complains about the 
>>  data created during installation ('partition c d e extend media size' or
>>  something like that) and refuses to write anything (I need a new bootstrap
>>  loader to boot from sd1a). Does that mean that the disklabel created
>>  during installation is somehow kaput, or does that mean the disklabel 
>>  utility has bugs ? And while we are at it - is it correct to set
>>  maj to 4 and unit to 1 in boot.c as default for sd1a ?

>You can still do the new bootstrap with -B.  Trying to rewrite the label
>will indeed have problems since the values that sysinstall chose were
>somewhat less than...  correct.. :-(  If you've really got courage, you
>can adjust them but you may lose a filesystem in the process!

didsklabel -B didn't work either. I did some investigation and I think 
I understood the problem:

The disk in question had been in use for FreeBSD before, but not as a 
boot drive. Therefore it was labeled with its native geometry values 
(2 heads etc).  Now Sysinstall correctly choose the Adaptec translations 
(64 heads etc) when modifying the label. But then it did not modify
partion c and d, respectively their maximum values (which are slightly 
larger with the original geometry - you loose about 500k with the Adaptec 
translation). 

Obviously disklabel has more built-in checks than sysinstall, discovered the
discrepancy and refused to write anything. I changed the maximum values
and then could write the disklabel. I didn't change anything with the file 
system, except a fsck - is that sufficient ?

>>- XFree86 3.1 needs about 40 seconds to start - is this normal (2.1 started 
>>  nearly immediately), and if not what can I do about that ?

>Well, it's a lot bigger! :-)  How much memory do you have?

8 Megs. And I doubt it is loading time. It loads a few seconds, changes 
the screen resolution, and then the whole machine, including the drives,
is dead for about 40 seconds. Then it shows the x screen background.

>>- another problem is that telnetd confuses my Lan Workplac DOS telnet 
>>  to death. Is there a workaround ? Another termcap entry for vt220 ?

>Have you tried just plain `ansi'?

Not an option here... would it work if I tried to take the termcap entry
from SunOS ?

[distfiles in ports/disfiles]
[pkg_* utils in /usr/sbin]

Found them, thanks !

Regards

Christoph Weber-Fahr

-- 
  Christoph Weber-Fahr                  |  E-Mail:  weber@rhrk.uni-kl.de 
  Universitaet Kaiserslautern,  KIT     |  S-Mail:  Postfach 3049
  Tel. 0631/205-3391                    |           D-67653 Kaiserslautern
--------------------------  My personal opinion only    ---------------------