From owner-freebsd-hackers@FreeBSD.ORG Fri Oct 17 03:22:31 2003 Return-Path: Delivered-To: freebsd-hackers@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 3CC2B16A4B3 for ; Fri, 17 Oct 2003 03:22:31 -0700 (PDT) Received: from essence.sdodson.com (67-50-89-87.br2.tbr.ga.frontiernet.net [67.50.89.87]) by mx1.FreeBSD.org (Postfix) with ESMTP id EA7F743FD7 for ; Fri, 17 Oct 2003 03:22:29 -0700 (PDT) (envelope-from samy@kerneled.com) Received: from beastie.freebsd.local (dial37-234.sbm.net.sa [212.46.37.234]) by essence.sdodson.com (8.12.10/8.12.10) with SMTP id h9HAM1in002375; Fri, 17 Oct 2003 10:22:16 GMT (envelope-from samy@kerneled.com) Date: Fri, 17 Oct 2003 13:22:55 +0300 From: Samy Al Bahra To: David Gilbert Message-Id: <20031017132255.197cd7b8.samy@kerneled.com> In-Reply-To: <16271.7039.150262.159805@canoe.dclg.ca> References: <1197083983.20031009074645@inbox.ru> <16271.7039.150262.159805@canoe.dclg.ca> Organization: Kerneled X-Mailer: Sylpheed version 0.9.5-gtk2-20030906 (GTK+ 2.2.1; i386-portbld-freebsd5.1) Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit cc: freebsd-hackers@freebsd.org cc: earthman@inbox.ru Subject: Re: On-line judgment kernel module X-BeenThere: freebsd-hackers@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Technical Discussions relating to FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 17 Oct 2003 10:22:31 -0000 On Thu, 16 Oct 2003 18:28:15 -0400 David Gilbert wrote: > As you conjecture, a syscall-less or syscall-restricted environment > *should* be safe ... if your syscall changes are bulletproof *_and_* > the rest of the runtime environment is bulletproof. Good system call policies are a WONDERFUL feature at a system administrator's hands. There is no such thing as a syscall-less environment but only a restricted (either at the same layer as the system calls or above in terms of code path). > Isn't a syscall required to finish off exit()? Yes, consult kern_exit.c How is this related to the discussion though? The fact is, most people would not even want to TOUCH sys_exit and friends since there are no real security advantages there. In otherwords, an exit system call remains completely the same. -- +-----------------------------------+ | Samy Al Bahra | samy@kerneled.com | |-----------------------------------| | B3A7 F5BE B2AE 67B1 AC4B | | 0983 956D 1F4A AA54 47CB | |-----------------------------------| | http://www.kerneled.com | +-----------------------------------+