From owner-freebsd-hackers  Tue Apr 29 20:32:43 1997
Return-Path: <owner-hackers>
Received: (from root@localhost)
          by hub.freebsd.org (8.8.5/8.8.5) id UAA29960
          for hackers-outgoing; Tue, 29 Apr 1997 20:32:43 -0700 (PDT)
Received: from genesis.atrad.adelaide.edu.au (genesis.atrad.adelaide.edu.au [129.127.96.120])
          by hub.freebsd.org (8.8.5/8.8.5) with ESMTP id UAA29955
          for <hackers@freebsd.org>; Tue, 29 Apr 1997 20:32:40 -0700 (PDT)
Received: (from msmith@localhost) by genesis.atrad.adelaide.edu.au (8.8.5/8.7.3) id NAA25320; Wed, 30 Apr 1997 13:02:34 +0930 (CST)
From: Michael Smith <msmith@atrad.adelaide.edu.au>
Message-Id: <199704300332.NAA25320@genesis.atrad.adelaide.edu.au>
Subject: Unloading LKMs (was Re: A Desparate Plea for Help...)
In-Reply-To: <199704291010.UAA30468@godzilla.zeta.org.au> from Bruce Evans at "Apr 29, 97 08:10:18 pm"
To: bde@zeta.org.au (Bruce Evans)
Date: Wed, 30 Apr 1997 13:02:33 +0930 (CST)
Cc: hackers@freebsd.org
X-Mailer: ELM [version 2.4ME+ PL28 (25)]
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Sender: owner-hackers@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk

Bruce Evans stands accused of saying:
> >Please keep me up to date on the results of your LKM-related tests.  If
> >it turns out that there's a problem with unloading LKMs leaving occupied
> >but invalid devsw entries around we'd better fix it 8)
> 
> Er, AFAIK unloading LKM drivers is broken in all cases.  It certainly
> doesn't work for either of the officially supported LKM cdevs (joy
> and qcam).  Unloading either of these and then attempting to opening
> the nonexistent device gives precisely the trap at _spec_open+0x6e that
> Simon reported (the devsw entry is not affected by unloading and points
> to garbage).  Unloading followed by reloading obviously can't work,
> because the driver only initializes the devsw once.

Um, the driver initialises the devsw on every load, does it not?  eg: 

static int 
qcam_load (struct lkm_table *lkmtp, int cmd)
{
        if (qcam_probe(&qcam_mod_dev)) {
                qcam_attach(&qcam_mod_dev);

                qcam_drvinit(NULL); /* XXX this shouldn't NEED to be here
                                     * the LKM code should be doing this
                                     * for us! */
...

Still, I take your point about the cdevsw not being updated.  I can't see
a clean way of doing this, either.

Doug R., how do you approach this in your New Module Structure?  (No, I
haven't had time to read it yet 8( 8( 8( )

> Bruce

-- 
]] Mike Smith, Software Engineer        msmith@gsoft.com.au             [[
]] Genesis Software                     genesis@gsoft.com.au            [[
]] High-speed data acquisition and      (GSM mobile)     0411-222-496   [[
]] realtime instrument control.         (ph)          +61-8-8267-3493   [[
]] Unix hardware collector.             "Where are your PEZ?" The Tick  [[