Path: utzoo!attcan!uunet!lll-winken!lll-tis!oodis01!uplherc!nrc-ut!nrcvax!kvc
From: kvc@nrcvax.UUCP (Kevin Carosso)
Newsgroups: comp.os.vms
Subject: Re: (none)
Message-ID: <1486@nrcvax.UUCP>
Date: 25 May 88 01:47:12 GMT
References: <8805160319.AA08048@ucbvax.Berkeley.EDU>
Reply-To: kvc@minnie.UUCP (Kevin Carosso)
Organization: Network Research Corp.  Oxnard, CA
Lines: 51

In article <8805160319.AA08048@ucbvax.Berkeley.EDU> ORG5NMC@CMS1.UCS.LEEDS.AC.UK writes:
>
>Subject: Fix for PTY driver
>
>DEC's recent SECURITY UPDATE V2 patch has broken the CMU/Hughes
>PTY pseudo terminal driver. The patch to TTDRIVER.EXE has changed the default
>device protection for cloned devices (e.g VTAn virtual terminals). Thus
>the pseudo terminal device TPAn: receives the totally useless device protection
>of [S:WRLP O: G: W:]. It used to be [S:WRLP O:WRLP G:WRLP W:WRLP].
>
>This can be cured by adding two lines to TPDRIVER.MAR in the device UCB
>initialisation procedure TP$INITLINE, AFTER default UCB initialisation
>is done. Thus:

Right, that should do it.  Anyone wanna venture some ideas on how the device
protection might be a problem?  It seems to me that, in the general case,
it's not really desirable to have such an open protection on terminal
devices, cloned or otherwise and so I agree with DEC's patch.  On the
other hand, PHOTO requires access to the terminal device for its spawned
subprocess.

>I hope someone will put this fix into the PTY driver (if not already done so)

Since I'm sort of the keeper of the PTY drivers, I'll put this into my
copy unless someone comes up with a convincing argument as to why this is
a security hole and what to do about it for PHOTO programs.

Another thing.  The TP device name is no good, it has been usurped by the
PSI folks at DEC.  I am officially changing the driver to TWDRIVER.  PYDRIVER
will remain the same.  If you want to fix your current drivers, just go into
PYDRIVER.MAR and TPDRIVER.MAR and change every occurance of TP to TW.

If you are a site at which TW conflicts with some unnamed Digital product,
don't worry.  You might notice that your unnamed Digital product also has
a PY device and you might notice that PHOTO works with the Digital-supplied
devices...

>Does anyone know why the demo PTY.PAS program that comes with the driver kit
>does not work on WorkStation 2000's if your process TT device is a WTAn
>window device. It hangs and refuses to bring up the login sequence. I know
>that a SYS$QIOW does not complete ( the one in the infinite loop reading
>the PYAn device) but for what reason? PTY.PAS works on the workstations
>TTA2 line.

As the author of PTY.PAS, I'd LOVE to know why this happens.  I've seen it
before but never had enough time with a VAXstation to try and figure it
out.  Anyone got any clues?


        /Kevin Carosso                     kvc@nrc.com
         Network Research Co.              kvc@ymir.bitnet
                                           kvc@nrcvax.uucp