[comp.sys.sun] Sun-Spots Digest, v6n43

Sun-Spots-Request@RICE.EDU (William LeFebvre) (04/04/88)

SUN-SPOTS DIGEST           Sunday, 3 April 1988        Volume 6 : Issue 43

Today's Topics:
               Administrivia: regarding "for sale" notices
             Re: Problems configuring cua0 on a 3/160 zs0 (2)
                            Re: YP problem (2)
                          Re: SunOS 3.5 Problems
      Re: Problems with VME bus timeouts:  Ironics board on a Sun-3
           Re: Experience with Sun 3/60 memory from clearpoint
                      Re: MX mailer needed for Suns
                       Bug Sunpro make / SunOS 3.4
                        Bug in SUNlink X.25 (CUDF)
              Fortran program demonstrating f77 "getfd" bug
                      Renting a Sun3/160 in Phoenix?
                  Problem uudecoding suntools background
                            Network Security?
                           fstab swap entries?
                          Postscript to impress?

Send contributions to:  sun-spots@rice.edu
Send subscription add/delete requests to:  sun-spots-request@rice.edu
Bitnet readers can subscribe directly with the CMS command:
    TELL LISTSERV AT RICE SUBSCRIBE SUNSPOTS My Full Name
Recent backissues are stored on "titan.rice.edu".  For volume X, issue Y,
"get sun-spots/vXnY".  They are also accessible through the archive
server:  mail the word "help" to "archive-server@rice.edu".

----------------------------------------------------------------------

Date:    Sun,  3 Apr 88 21:38:52 CDT
From:    William LeFebvre <phil@Rice.edu>
Subject: Administrivia: regarding "for sale" notices

In the recent past I have received several messages advertising a
collection of used Sun equipment for sale.  The first one or two I
received I actually placed in a digest.  Since that time, I have been made
aware of the possibility that I may be violating certain ARPANet rules by
sending out such messages.  Until I am able to get a more definite idea of
the official policy on for-sale notices, I must refrain from including any
such messages in digests.  I will let the readership know when we finally
determine what is correct.  If you have solid (factual) information about
this topic, please send me mail about it.  In the meantime, if you are a
Usenet site and you are trying to sell some equipment, I would recommend
the newsgroup "misc.forsale".

William LeFebvre
Department of Computer Science
Rice University
<phil@Rice.edu>

------------------------------

Date:    Mon, 21 Mar 88 22:09:04 CST
From:    vixen!ronbo@cs.utexas.edu (Ron Hitchens)
Subject: Re: Problems configuring cua0 on a 3/160 zs0 (1)
Reference: v6n31

In v6n31 Michael van Biema <MICHAEL@cs.columbia.edu> asks why he can't get
/dev/cua0 working right so that he can dialin and dialout on the same
line.  He wraps up with:

> I assume that among the things I've supposedly configured correctly, there
> is something that I've overlooked or just plain screwed up, but maybe I've
> just missed something completely....

Yes, you've missed something completely.  The cua0 mechanism is NOT a
method of using one tty port for both dialin and dialout.  It is simply a
method of allowing a process to open a serial port, which has hardware
carrier detect configured, when there is no carrier present.  These are
two entirely different things.

All cua0 gives you is the ability to open a serial port without carrier so
that you can talk to the modem and tell it to dial.  If you open the
regular serial device, the open will hang (sleep in the kernel) until
carrier is detected on the port.  This is what normally happens when init
spawns a child process for a dialin port.  The child process initiates an
open on the serial port and the open sleeps until someone calls the modem
and it senses carrier at the other end.  At that point the open completes
and init execs getty which execs login, etc.

The cua0 only affects the carrier test at open time, it doesn't grant any
special priviledges or override someone elses exclusive access.  It does
nothing about getting init, getty or whoever else may be trying to use the
device out of the way.  The problems you're describing are simply tip and
init fighting for access to the serial port.

Unfortunately there is no simple method to use one modem for both dialin
and dialout like you're trying to do.  The setup as it exists doesn't
allow it.  Many people have implemented hacks to do it, using various
schemes, but none of them are really elegant.  The basic problem arises
from the need to have a process waiting to accept an incoming connection,
and the difficulty involved in gracefully moving that process out of the
way when you want to use the port for an outgoing call.  A good, seamless
solution would require some rather fundamental changes in several places,
so don't expect one very soon.

In the meantime, I'm afraid enabling and disabling the port by hand with
the "kill -HUP 1" nerve pinch is about all you can do.

Ron Hitchens		ronbo@vixen.uucp	hitchens@cs.utexas.edu

------------------------------

Date:    Tue, 22 Mar 88 13:52:38 PST
From:    roode@pisa.orc.olivetti.com (David Roode)
Subject: Re: Problems configuring cua0 on a 3/160 zs0 (2)
Reference: v6n31

It sounds like your RS232C device is delivering a continuous carrier
signal.  Since your cable passes that signal, which is normally desired,
the ttyd0 device will always be busy.  Can you configure your modem to not
provide carrier constantly?

------------------------------

Date:    23 Mar 88 16:33:36 GMT
From:    Andrew D. Bowen <psuvax1!adbst@cisunx.cs.psu.edu>
Subject: Re: YP problem (1)

We had the same problem, and the good word from above is that the docs are
wrong!  The YP never worked like that, and the security setup that they
suggest is simply not possible.

- Andy Bowen
  Dept. of Electrical Eng.
  U. of Pittsburgh

------------------------------

Date:    Wed, 23 Mar 88 08:57:36 PST
From:    jonab@cam.unisys.com (Jonathan P. Biggar)
Subject: Re: YP problem (2)

It turns out that yppasswdd requires that the password master source file
be in the same filesystem as /tmp on the master machine, because it does
link and unlink tricks to /tmp to store temporary files.

/etc/passwd is on the same filesystem as /tmp on most machines, however
/etc/yp is a link to /usr/etc/yp and is not on the same filesystem.

Jon Biggar
jonab@cam.unisys.com

------------------------------

Date:    22 Mar 1988 09:30-CST
From:    Rafael Bracho <ascway.UUCP!rxb@spar-20.spar.slb.com>
Subject: Re: SunOS 3.5 Problems
Reference: v6n34

We are running 3.4 but couldn't upgrade to 3.5 due to a bug in the nd
driver that stopped us from having two nd-swap partitions on a client.
Apparently there aren't that many sites having clients swap on nd1 and
nd2.

Rafael Bracho
RXB@SPAR-20.ARPA

------------------------------

Date:    Tue, 22 Mar 88 09:42:56 PST
From:    vsi1!tim@sally.utexas.edu (Tim Richardson (201))
Subject: Re: Problems with VME bus timeouts:  Ironics board on a Sun-3
Reference: v6n31

In reply to Bob Amstadt's question of bus timeouts using the Ironics
board...

This is most likely NOT, repeat not a problem with the speed at which the
Ironics board releases the bus.  In fact, the SUN 3 CPU boards, the
Carrera board, has numerous VMEbus interfacing problems.  The biggest one
is in the handling of interrupts.

Suggest first that you get the part number and rev level (look at the bar
code tag on the front of the pcb).  If you are using a Rev A board, inusre
that it is level 9 or higher.  If you are using a Rev B board, insure that
you are at level 5 or better.

Suggest secondly that you run your ironics board at interrupt priority
level 3 with rev a board and at priority level 2 with rev b boards.

There is a new ECO from SUN on the rev b boards which is a PAL change only
to rev b boards at either level 3 or 4 which will bring it to a level 5.
Your SUN customer service people can help get this for you.

Tim Richardson
VICOM Systems Inc.     sun!pyramid----\
2520 Junction Ave.     uunet!ubvax-----!vsi1!tim
San Jose, CA  95134    ucbvax!tolerant/
+1-408-432-8660

------------------------------

Date:    Tue, 22 Mar 88 09:50:15 PST
From:    vsi1!tim@sally.utexas.edu (Tim Richardson (201))
Subject: Re: Experience with Sun 3/60 memory from clearpoint
Reference: v6n32

Vicom has looked at both the Clearpoint memory expansion and the Parity
Systems.  We like the Parity board better...its cheaper and better made.
Also they have a version where you can insert your SUN SCSIbus Host
Adaptor (or other 6U VMEbus card) INTO the memory board, thus taking up
only ONE backplane slot.

Tim Richardson VP Engineering
VICOM Systems Inc.     sun!pyramid----\
2520 Junction Ave.     uunet!ubvax-----!vsi1!tim
San Jose, CA  95134    ucbvax!tolerant/
+1-408-432-8660

------------------------------

Date:    Wed, 23 Mar 88 16:33:08 EST
From:    lear@aramis.rutgers.edu (eliot lear)
Subject: Re: MX mailer needed for Suns
Reference: v6n34

Mel Pleasant has done a lot of work on sendmail here at Rutgers.  You can
pick up a copy through anonymous ftp from internet site RUTGERS.EDU.  The
file is src/RUsendmail.tar.Z.  It will handle MX records among other
things.

Regards,
--
Eliot Lear
[lear@rutgers.edu]

------------------------------

Date:    Tue, 22 Mar 88 09:36:10 +0100
From:    mcvax!presto.irisa.fr!hleroy@uunet.uu.net (Hugues Leroy -- iPSC team)
Subject: Bug Sunpro make / SunOS 3.4

I think there is a bug in Sunpro make. See the following example:

(I note BCK the back-slash , as i don't know if the mailers on the way
accept back-slash)

[[ I changed them back to backslashes, but I will note where they appear.
--wnl ]]

in my makefile :
----------------
.....

[[ The parens have backslashes before them ]]
LO=\( -name '*.l[om]' -o -name MODULES \)
listar:
        @find $(PREF)sources/comp_metal $(LO) -print
        @find $(PREF)sources/comp_ppml $(LO) -print
        @find $(PREF)sources/parser $(LO) -print
        @find $(PREF)sources/decompilo $(LO) -print
.....

make -n listar =>

.... [[ each close paren is preceded by a backslash ]]
find sources/comp_metal -name '*.l[om]' -o -name MODULES \) -print
find sources/comp_ppml -name '*.l[om]' -o -name MODULES \) -print
find sources/parser -name '*.l[om]' -o -name MODULES \) -print
find sources/decompilo -name '*.l[om]' -o -name MODULES \) -print

============================^^^^ NO backslash-( ======

BUT WITH /usr/sunpro/3.2/make -n listar =>
--------
..... [[ open and close parens are preceded by backslashes ]]
find sources/comp_metal \( -name '*.l[om]' -o -name MODULES \) -print
find sources/comp_ppml \( -name '*.l[om]' -o -name MODULES \) -print
find sources/parser \( -name '*.l[om]' -o -name MODULES \) -print
find sources/decompilo \( -name '*.l[om]' -o -name MODULES \) -print

That's work fine.

------------------------------

Date:    Wed, 23 Mar 88 09:36:02 CST
From:    Stephan Wasserroth <wasserroth@fokus.berlin.gmd.dbp.de>
Subject: Bug in SUNlink X.25 (CUDF)

There is a nasty bug in SUNlink X.25 (vers. 4.0 and 5.0): The
Call-User-Data-Field (contained in a Call-Request-Packet) is not set
correctly. For example, let the call contain the following CUDF:

    01  02  03                      (length 3 octets)

then the following is send out from SUNlink X.25:

    ??  01  02  03  ??  ??  ??      (length 7 octets)

The original user-data is doubled in length and shifted one octet. The
error seems to happen in the routine "connect".

DFN-EAN:       <wasserroth@fokus.berlin.gmd.dbp.de>
ARPA:          <wasserroth%fokus.berlin.gmd.dbp.de@relay.cs.net>

                        GMD-FOKUS
Stephan Wasserroth      Hardenbergplatz 2
System Manager for      D-1000 Berlin 12
VAXes and Sun-WS        Fed. Rep. of Germany

------------------------------

Date:    Wed, 23 Mar 88 16:57:21 GMT
From:    Ida <ida@eagle.warwick.ac.uk>
Subject: Fortran program demonstrating f77 "getfd" bug

C
C short Fortran program to demonstrate a bug with 'GETFD()'.
C compile with f77, create a temporary file called 'testfile' and then run.
C GETFD() does not return a sensible file descriptor.
C Am I missing something obvious?  Thanks for your help.    Russ.
C
C
      PROGRAM MAIN
C
      EXTERNAL GETFD,FLUSH
      INTEGER*4 ERROR
      LOGICAL*4 EX,OP
      CHARACTER*8 FNAME
C
      PRINT *,'inquire file on unit 20...'
      OPEN(20,FILE='testfile', STATUS='old')
      INQUIRE(NAME=FNAME, UNIT=20, IOSTAT=ERROR, EXIST=EX,
     &OPENED=OP)
      PRINT *,'error value=',ERROR
      PRINT *,'is it there?',EX
      PRINT *,'is it opened?',OP
      PRINT *,'name is ',FNAME
      PRINT *,'logical unit 20: the C fildes is:',GETFD(20)
      PRINT *,'finished'
      END

------------------------------

Date:    23 Mar 88 13:27:24 GMT
From:    bdrc!pdh@mcnc.org (Perry D. Haaland)
Subject: Renting a Sun3/160 in Phoenix?

I would like to rent a Sun3/160C workstation for a demonstration in
Phoenix on April 13, 1988.  I wish to demonstrate statistical graphics
using S-Plus so I also want to rent a projection system.  Names and
addresses of Sun rental companies in Phoenix would be greatly appreciated.
Advice, warnings, etc. would also be appreciated from anyone who could
share their experiences in running a Sun demo on rental equipment.  --
Perry Haaland - mcnc!bdrc!pdh

------------------------------

Date:    Wed, 23 Mar 88 08:52:41 PST
From:    Stuart Cracraft <cracraft@venera.isi.edu>
Subject: Problem uudecoding suntools background

The recently proferred suntools background isn't cooperating with
uudecode.  uudecode reports 'enterprise.Z: permission denied'. This is on
a SUN-3/140 running SUN release 3.4.

	Stuart

[[ Ahhhh, this has nothing to do with enterprise.Z!  This is because (a)
uudecode is installed setuid to uucp (this is true on every Unix system I
have encountered) and (b) the directory that is current when you tried to
run uudecode was not writable by uucp.  Three solutions:  (1) cd
appropriate-directory; chmod go+w . ; uudecode enterprise.Z ; chmod go-w .
(replace the permissions in that last chmod with whatever is appropriate).
(2) have the system administrator turn off the setuid bit for uudecode (I
think we determined the last time this came up that it doesn't really need
to be on).  (3) Place a copy of uudecode in a different bin directory
(such as your own or /usr/local) called, say, "decode", with the setuid
bit off.  --wnl ]]

------------------------------

Date:    Tue, 22 Mar 88 15:20:25 CST
From:    jsalmi@woods.uucp (John Salmi)
Subject: Network Security?

Here is the scenario.  I have a standalone node, connected to the NSFNet,
and I am considering connecting it to my production network.  We have a
few security concerns, such as IP forwarding, world access to our network,
etc...

I have heard rumors :-) of quite large networks, such as UCB, which have
literally hundreds of machines on their networks, and are also connected
to the world via arpa.

I need to sell this idea to my boss, who is somewhat paranoid (at least
someone is :-) about letting the world have access to our stuff.

Is there an rfc or other similar document which breaks down the essence of
this networking/security issue?  Or, if there is a system/network admin
out there who has faced this issue before and has some tips, I'd
appreciate hearing from you.

Thanks, as always, in advance!
                                   : ...!ames!eta!jsalmi
John Salmi 		Email      : jsalmi@mondo.unix.eta.com
Software Engineer	Voice      : (612) 642-3012
ETA Systems, Inc.	SnailMail  : 1450 energy Park Drive
				     St. Paul, Mn.  55108

------------------------------

Date:    Wed, 23 Mar 88 15:10:11 EST
From:    ted@braggvax.arpa
Subject: fstab swap entries?

What should a swap entry in /etc/fstab look like?

We have swap space on 0b and 1b, and would like for the 1b area to be
automatically used.  In /etc/rc sun does a swapon -a, which is supposed to
add any swap type partitions from /etc/fstab, but I have tried a couple of
permutations and haven't gotten it to work so far.  We can put a swapon
/dev/sd0b in /etc/rc, but I would prefer to do it the right way...

Any clues?  This is sun3.4 on a 2/120.

Ted Nolan
ted@braggvax.arpa

------------------------------

Date:    Wed, 23 Mar 88 14:59:52 EST
From:    ted@braggvax.arpa
Subject: Postscript to impress?

Well great!

The fig to screendump format should solve the immediate problem we had with
printing fig output (we had to convert to PIC, ship to braggvax, and itroff),
but the thought still remains:

Does anyone have a postscript to impress conversion filter?  This would help
us a lot with more and more Postscript sw coming out and our imagens being
deaf to it.

Thanks,

Ted Nolan
ted@braggvax.arpa

[[ I have never heard of such a beast, although I don't think it would be
too hard ot at least do a subset (you might want to direct your question
at the laser-lovers list).  However, Imagen printers are no longer deaf to
PostScript.  Imagen has added a language called "UltraScript" to its list
of available language for their printers.  As I understand it, it is a
PostScript superset.  Of course, it ain't cheap!  ---wnl ]]

------------------------------

End of SUN-Spots Digest
***********************