Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site brl-tgr.ARPA
Path: utzoo!watmath!clyde!bonnie!akgua!sdcsvax!dcdwest!ittvax!decvax!genrad!panda!talcott!harvard!seismo!brl-tgr!tgr!root%bostonu.csnet@csnet-relay.arpa
From: BostonU SysMgr <root%bostonu.csnet@csnet-relay.arpa>
Newsgroups: net.unix-wizards
Subject: Re:  protect tape access
Message-ID: <8241@brl-tgr.ARPA>
Date: Mon, 11-Feb-85 16:28:53 EST
Article-I.D.: brl-tgr.8241
Posted: Mon Feb 11 16:28:53 1985
Date-Received: Fri, 15-Feb-85 03:46:19 EST
Sender: news@brl-tgr.ARPA
Organization: Ballistic Research Lab
Lines: 26


	I started to put one in my system, tell me if you
	find one but my scheme went like this:

	create a psuedo-user 'free' (or some such)
	free owns the tape drive (owner/group)
	make a tapealloc command called something which
	is setuid. It chowns the tape drive and forks a
	sub-shell. When the subshell exits it returns it
	to 'free'. The reason for the subshell is that
	any attempt to log out (eg. hanging up a phone)
	will free up the tape drive again (you may have
	to play with signals in tapealloc but it is straightforward.)

	The major problem is: The subshell method could make a
	few things awkward but this is not an unusual constraint
	and a user could still 'hog' a tape drive although at
	least now a 'ls -l /dev/?mt*' lists who is using what.
	Also, you will have to chown some set of devices
	(eg, /dev/mt0, /dev/rmt0, /dev/mt8 etc on 4.2bsd)
	but this is still reasonable.

		-Barry Shein

	[oh yeah, make sure you do your setuids correctly before
	forking that (setuid'd) subshell.]