Received: from localhost by CS.UTK.EDU with SMTP (cf v2.9s-UTK) id NAA13376; Wed, 20 Mar 1996 13:32:43 -0500 Received: by CS.UTK.EDU (bulk_mailer v1.4); Wed, 20 Mar 1996 13:32:23 -0500 Received: from ng.netgate.net by CS.UTK.EDU with ESMTP (cf v2.9s-UTK) id NAA13338; Wed, 20 Mar 1996 13:32:17 -0500 Received: from [205.214.160.38] (d6.netgate.net [205.214.160.38]) by ng.netgate.net (8.6.12/8.6.9) with ESMTP id KAA13250 for ; Wed, 20 Mar 1996 10:41:54 -0800 X-Sender: dcrocker@ng.netgate.net Message-Id: In-Reply-To: <19960320173628.446.qmail@koobera.math.uic.edu> Mime-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Date: Wed, 20 Mar 1996 10:10:44 -0800 To: drums@cs.utk.edu From: Dave Crocker Subject: Re: msg-id Dan, At 9:36 AM 3/20/96, D. J. Bernstein wrote: >> > Your style of program design has been demonstrated to lead to security >> > holes. >> I be interested in specifics on this > >``Least privilege.'' Code should not execute with more privilege than it >needs. Three of the last seven sendmail security holes would have been >nothing more than core dumps if Eric Allman had followed this principle. I was asking about any specific security problems this issue will introduce. The principle you cite is fine as a principle -- and I wish more folks paid attention to it -- but doesn't give a concrete basis for warning one away from specific features. What, precisely, is the security concern for this specific suggestion? >doesn't matter how ``secure'' your messages are if an attacker takes >over your machine! how will the inclusion of the local host name permit this? >That's exactly the attitude that creates insecure programs. > >You don't always see bugs in advance, but you can give them several >hurdles to jump before they cause security holes. Adding unnecessary >code to setuid programs means flattening the hurdles. As for myself, I've always believed that the retail business would be a whole lot less stressful on sales people if we could just find a way to get rid of all those customers. Life is risks. operational security is a matter of balancing feature enhancement against potential risks. If you want perfect security, write an autistic program. No security problems at all. Not much utility either. We need a balance, not an absolute adherence to one extreme or the other. d/ -------------------- Dave Crocker +1 408 246 8253 Brandenburg Consulting fax: +1 408 249 6205 675 Spruce Dr. dcrocker@brandenburg.com Sunnyvale CA 94086 USA http://www.brandenburg.com Internet Mail Consortium http://www.imc.org, info@imc.org