From owner-freebsd-hackers Mon Apr 28 21:43:12 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.5/8.8.5) id VAA23919 for hackers-outgoing; Mon, 28 Apr 1997 21:43:12 -0700 (PDT) Received: from sendero.i-connect.net (sendero-ppp.i-Connect.Net [206.190.143.100]) by hub.freebsd.org (8.8.5/8.8.5) with SMTP id VAA23879 for ; Mon, 28 Apr 1997 21:42:58 -0700 (PDT) Received: (qmail 4725 invoked by uid 1000); 29 Apr 1997 04:21:11 -0000 Message-ID: X-Mailer: XFMail 1.1-alpha [p0] on FreeBSD Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit MIME-Version: 1.0 In-Reply-To: <199704290247.MAA18906@genesis.atrad.adelaide.edu.au> Date: Mon, 28 Apr 1997 20:45:06 -0700 (PDT) Organization: iConnect Corp. From: Simon Shapiro To: Michael Smith Subject: Re: A Desparate Plea for Help... Cc: freebsd-hackers@freebsd.org Sender: owner-hackers@freebsd.org X-Loop: FreeBSD.org Precedence: bulk Hi Michael Smith; On 29-Apr-97 you wrote: ... > > > Are you loading/unloading LKM's on these systems? > > > > Yes. linux_mod, star_saver_mod, oss_mod (sound). > > Should I not? > > They're certainly prime suspects. Your crash looks/looked like it had > something to do with one of the device switch tables, and one possible > explanation would be that an LKM had been removed without zeroing > its switch entries. > > I have _definitely_ seen problems with the screensaver LKMs at times. > Make sure your LKMs are up to date with the kernel too. How do I do that (other than make world - which I am doing on a regular basis)? Yes, I have had severe problems with the screen savers, but they _seem_ fine now. They are not critical for this project. Not yet anyway. A prime suspect is the OSS driver, for which I have no sources. It is OFF. We may want to consider some mechanism to protect the system from LKM's blowing it up so easily. Like an unload function that will put a shorting plug in the interface, or something. I know nothing of FreeBSD LKM but in other O/S, I have seen the stubs being the default when the ``real thing'' is unplugged. I am sure we have something like that already. Another thing can be (again, if it is not there already) is a strict versioning option, which even linux has; It allows the kernel modules loader to verify that a given module is of the proper/matching version before loading. Simon