From owner-cvs-all Sun Feb 21 10:41:23 1999 Delivered-To: cvs-all@freebsd.org Received: from heathers.stdio.com (heathers.stdio.com [199.89.192.5]) by hub.freebsd.org (Postfix) with ESMTP id AA51610F37 for ; Sun, 21 Feb 1999 10:41:21 -0800 (PST) (envelope-from lile@stdio.com) Received: from heathers.stdio.com (lile@heathers.stdio.com [199.89.192.5]) by heathers.stdio.com (8.8.8/8.8.8) with ESMTP id NAA09627; Sun, 21 Feb 1999 13:22:42 -0500 (EST) (envelope-from lile@stdio.com) Date: Sun, 21 Feb 1999 13:22:41 -0500 (EST) From: Larry Lile To: Luigi Rizzo Cc: phk@critter.freebsd.dk, julian@whistle.com, jkh@zippy.cdrom.com, cvs-all@FreeBSD.ORG Subject: Re: Current status of the olicom fracas. In-Reply-To: <199902211423.PAA14194@labinfo.iet.unipi.it> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-cvs-all@FreeBSD.ORG Precedence: bulk On Sun, 21 Feb 1999, Luigi Rizzo wrote: > > > I personally advocate the position that any instruction executed > > > on the CPU must be available in source form, and therefore this > > i second this position. The idea being that you might not trust a > module and want to have a chance that it does not do bad things. Fine, warnings could be issued at compile time. Also, if deemed necesary, warning could also be issued during the device probe. > > Now how do we save the oltr driver and token-ring support for FreeBSD? > > Perhaps this would be a good case for a "kernel patch" type of > port, wouldn't it ? > I know by experience that kernel patches are problematic because > they are much harder to keep consistent with the system and/or other > patches, but at least, by going this way, one really has to know > what s/he does before using an external object module. You are exactly right about that. It is too difficult for networking code in my opinion. I just sent a reply to PHK with a possible solution. > The other possibility would be to modify the config structure so that > obkect-only modules can be clearly identified and either config or a > kernel build loudly remarks the use of such a module. I have no problem with that, as mentioned above. But could I add a compile option to knock them out? The user would have to specify it _explicity_ and by default it would complain. That should alleviate both security and aesthetic concerns. Larry Lile lile@stdio.com To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe cvs-all" in the body of the message