From owner-freebsd-current  Fri May 12 13:15:12 1995
Return-Path: current-owner
Received: (from majordom@localhost)
          by freefall.cdrom.com (8.6.10/8.6.6) id NAA20769
          for current-outgoing; Fri, 12 May 1995 13:15:12 -0700
Received: from hda.com (hda.com [199.232.40.182])
          by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id NAA20762
          for <current@freebsd.org>; Fri, 12 May 1995 13:15:07 -0700
Received: (dufault@localhost) by hda.com (8.6.9/8.3) id QAA11912; Fri, 12 May 1995 16:16:03 -0400
From: Peter Dufault <dufault@hda.com>
Message-Id: <199505122016.QAA11912@hda.com>
Subject: Strange LKM problems
To: current@FreeBSD.org
Date: Fri, 12 May 1995 16:16:03 -0400 (EDT)
X-Mailer: ELM [version 2.4 PL24]
Content-Type: text
Content-Length: 1743      
Sender: current-owner@FreeBSD.org
Precedence: bulk

I have occasional failures loading up LKM modules, without any
obvious reasons.

I'm using this "./in" script to install the module:

> desmond# cat ./in
> modload -v -A /kernel -e sermux_init -o sermux.out sermux_mod.o
> ./setldisc /dev/ttyd4 7
> ./setldisc /dev/ttyd5 7
> desmond# 

And look at this failure:

> desmond# ./in  
> modload: not an a.out format file
> TIOCSETD: Operation not supported by device
> TIOCSETD: Operation not supported by device
> desmond# 
> desmond# sync
> desmond# rm sermux.out
> desmond# ./in
> Registered line discipline at 7.
> sermux loaded.
> Module loaded as ID 1

Here you see two load attempts in a row with nothing else happening
in between them, and note that it worked fine the second time
without my doing anything.  In another window I checked the output
file on the NFS server that has the object on it:

> hda# nm *.out
> nm: sermux.out: not object file or archive.

so after it failed the first time it wasn't an object file. (I
stupidly didn't ls it; I bet it was null). When the load worked it
was fine.

This is on an NFS mounted partition; the system that is doing the
loading is the client.

I also had a random crash during the load, and then checking
the .out file I had a complaint about the .stab entries.  It
is as if the file is short sometimes.

I've tried to reproduce the problem by continuously loading and
unloading the module in a command script but without any luck: If
it is going to fail at all, it is the first time after a
reboot and then I can't get it to happen again.

Any ideas?  It is really odd.

-- 
Peter Dufault               Real Time Machine Control and Simulation
HD Associates, Inc.         Voice: 508 433 6936
dufault@hda.com             Fax:   508 433 5267