Sun-Spots-Request@RICE.EDU (Vicky Riffle) (08/21/87)
SUN-SPOTS DIGEST Thursday, 20 August 1987 Volume 5 : Issue 34
Today's Topics:
Problems with 3.4
Adding SCSI drives to 3/5x
extending monochrome monitor cables - how to do it
Tip Problems "Solved"
TCP/IP print spooling
HELP: not enough window displays!
Disk question
RCS Make?
Query: X on the high-resolution monitor of a 3/260?
32 bit disk controller for Sun 3?
Securing SUN workstations?
acquiring large amounts of data into Sun memory?
Memory Expansion for Sun 3/60?
Emacs icon?
Idle logout program for Suns
Wasting time on a Sun3
Self portrait
Yet another ICON
----------------------------------------------------------------------
Date: Sun, 16 Aug 87 00:08:39 PDT
From: edhall%rondo@rand-unix.arpa
Subject: Problems with 3.4
We have recently attempted to install 3.4, and have encountered two
problems serious enough to back out the installation and continue with
3.2:
1) Certain 8-bit PixWin operations on cgtwo color framebuffers cause
``Panic: Bus Error''. Several of our [not particularly elaborate]
color graphics applications are unusable as a result. Relinking
with the new libraries has no effect. Running the generic kernel
binary shipped with 3.4 doesn't help. Neither does building a new
kernel from the GENERIC configuration. And, yes, we supply the
proper vector in the config file (people using old config files
beware!).
2) TCP/IP connections with our ARPANET gateway (a VAX running vanilla
4.3) freeze up after the first 50,000 characters or so. Using
TCPDUMP shows the Sun3.4 host doing various bizarre things, such
as starting with an extra-large window (about 50,000 characters)
which dwindles to zero, after which it finally bounces back to
4096--but refuses to ACK any incoming data from that time on.
The VAX does everything completely by the book, even though the
Sun is doing other strange things like sending identical ACK's
in bursts of two, three, or even four (in time spans beneath
TCPDUMP's 20ms resolution). Other 4.3 VAXen don't seem to have
this problem; whether the IP address or gateway status is what
triggers this anomoly is unknown. But as I said, there is
absolutely nothing wrong with the VAX's handling of TCP--while the
Sun acts goofy from its very first ACK.
The second problem exists on both Sun-2's and Sun-3's. We don't have any
Sun-2's with cgtwo's, but the first problem occurs on all of our 3/160C's
without exception. Using another release tape didn't help.
There are more minor problems (e.g. the Browse feature in icontool seems
to be broken) as well. In all, I have to say that unless we're doing
something awfully stupid here, Sun has released a Real Turkey.
What are other people's experiences with 3.4? Are these problems peculiar
to us? Are there fixes? Should we wait for 3.5?
-Ed Hall
Information Sciences Department
The RAND Corporation
1700 Main Street
Santa Monica, CA 90406
(213) 393-0411 x7607
edhall@rand.org
------------------------------
Date: Thu 13 Aug 87 10:44:33 PDT
From: Dick <MEAD%hamal.usc.edu@oberon.usc.edu>
Subject: Adding SCSI drives to 3/5x
When you add your own SCSI drives to the 3/5x, using the ACB4000A controller,
you have the following Diag options: (not necessarily in order)
SKEW (Interleave) usually 1, could be 2, affects sectors/track.
Sectors/Track = 17 if skew=1, 18 if skew >1 (512 byte sectors)
The ACB4000 selects the SPT based on what you tell it the skew and sector
size is. Adaptec has a very nice manual for the ACB4000, which it appears
never gets sent unless you order it specifically.
------------------------------
Date: Fri 14 Aug 87 11:31:08
From: Niall Mansfield <MANSFIEL%EMBL.BITNET@wiscvm.wisc.edu>
Subject: extending monochrome monitor cables - how to do it
(I asked a while back about this but got no replies).
We asked the local Sun office, and they told us to use a co-ax cable for
EACH wire in the cable, so we have four co-ax cables bundled together,
with a tiny connector at each end! It looks funny, but it works.
(This is for the hi-res mono screen, on our 3/260).
Niall Mansfield
European Molecular Biology Laboratory
Meyerhofstrasse 1, 6900 Heidelberg, W. Germany
(06221) 387-0 [switchboard] (06221) 387-247 [direct]
------------------------------
Date: Sun, 16 Aug 87 22:25:05 EDT
From: Steve Simmons <umix!itivax!lokkur!scs@rutgers.edu>
Subject: Tip Problems "Solved"
We're the folks who had the wierd buffering problems with tip and our
VT-100 emulator on the Suns: after a screen clear the screen stayed blank
for a long time and then all the data flashed on at once. A number of
folks were very helpful in offering circumstances where similar things
did/didn't happen; many thanks to you all [you know who you are! ;-) ]
The ultimate "solution" came from someone who wished to remain anonymous,
proving it was a "bug" in the Sun shell window display. Seems it tries to
minimise screen repaints, and veiwed our situation as one where it can get
away with only one repaint. We are reporting this feature as a bug -- or
at least requesting a new feature that lets us make our own decisions
about how we want this to work.
:rutgers!husc6!necntc!gordon!josh (Josh Littlefield)
:ndd@duke.cs.duke.edu (Ned Danieley)
In addition, both these folks asked if our VT-100 emulator is PD or
licencable. Details are:
It's based on a PD VT-100 emulator someone wrote for a Sun-2, and required
heavy rework both for Sun-3 use and to clean up a lot of deficiencies. We
have asked the Powers That Be if they will permit the rewritten code to be
released to PD, and they're dithering. Caveat: this is *only* for a
Sun-3, using standard suntools. If you want to get a general one for
UNIX, you'd probably be better of writing your own. Aside from that, ours
is definately better than Suns -- it doesn't screw up your keyboard in
other windows, has pop-up keyboards mimicing a VT-100 board, etc, etc.
Main lacks are 132 columns and graphics character sets. One lever we're
using on the Powers is maybe some nice netlander will add these missing
items, or even upgrade it to VT220hood...we continue to nag them, and they
continue to dither. Sigh. If worst comes to worst we'll post the
original PD thing.
------------------------------
Date: Mon, 17 Aug 87 06:04 EDT
From: SIMON%M_SCRVX2%sdr.slb.com@relay.cs.net
Subject: TCP/IP print spooling
Thanks to those who replied to my query about driving printers from Suns
via a Bridge CS/100 terminal server running TCP/IP. Most of the necessary
information is contained in Application Note Number 12 Service Listeners
for Communications Servers with TCP/IP available from Bridge. However to
summarise what I did:
1. On the CS/100,
1.1 Associate an RS-232 port with an Internet address and host name.
1.2 Run the SYSGEN procedure to define port 2000 as a "Service Listener
Well Known Port".
2. On the Sun,
2.1 Include in /etc/hosts the host chosen in 1.1 above.
2.2 Write a program to open a TCP connnection to port 2000 on the chosen
host, and copy standard input across the connection. Sample programs
can be found in the Application Note or on page 21 of the Sun IPC Primer.
2.3 Make an /etc/printcap entry with :of=prog_name and :lp=/dev/null,
where prog_name points to the program written in 2.2 above.
Having done all these things I was able to transfer 2.5 Mbyte plot files
to our device with no apparent errors. Note that the program written in
2.2 has no knowledge of the telnet protocol, and makes no attempt to
handle flow control - these details are apparently taken care of by the
system software.
Simon Barnes
Schlumberger Cambridge Research
simon%m_scr%sdr.slb.com@relay.cs.net
------------------------------
Date: Fri, 14 Aug 87 11:00:00 EST
From: munnari!uacomsci.ua.oz!hawke@uunet.uu.net (Andrew Hawke)
Subject: HELP: not enough window displays!
I am a programmer at the Computer Science Dept of Adelaide working with
SUN 2/50's and a SUN 3/50 on a project that involves various processes,
each having their own window display, all communicating via sockets.
The problem is a limit on the number of window displays that I can create.
After I have created approximately 22 windows, using the window_create
call mechanism provided by Sunview, the operating system responds with the
message that it cannot open any more windows and then spits IOCTL numbers
at the screen. I have traced the problem back to the fact that every
window created has an open file associated with it, and that there is a
limit of approximately 25 open files for any single process. Evidently
there is a single process which manages the parent-window (i.e. screen)
and associated subwindows.
I suspect that there is no real solution to this problem apart from
writing one's own window manager from the pixrect level, a project I
shudder to think about. I guess a short term solution is to attempt to
reconfigure the kernel with a larger number of open files per process, but
I'm not at all sure of the consequences of such an action.
Is there anyone out there who could enlighten me?
Thanks in advance
Andrew Hawke.
- hawke@uacomsci.ua.oz -
------------------------------
Date: Wed, 12 Aug 87 10:15:17 -0700
From: Craig Rolandelli <craig@rome.uci.edu>
Subject: Disk question
We want to buy a Fujitsu 2344K disk and were wondering if anybody has
hooked one of these up to a Sun 3/2X0 machine. It is an 8 inch disk that
is 690 mbytes unformatted.
Any help is greatly appreciated.
Craig Rolandelli (craig@ics.uci.edu)
Computer Science Dept.
University of California, Irvine
Irvine Ca. 72717
(714) 856-4222
------------------------------
Date: Wed, 12 Aug 87 12:09:54 PDT
From: ultra!norm@ames.arpa (Norm Finn)
Subject: RCS Make?
We have Walter Tichey's RCS from Purdue. Does anyone have a make program
that includes both SCCS and RCS? How about a make that has all of Sun's
make features using RCS instead of SCCS?
------------------------------
Date: 13 Aug 87 10:51:25 GMT
From: mcvax!inria!shapiro@seismo.css.gov (Marc Shapiro)
Subject: Query: X on the high-resolution monitor of a 3/260?
We just received a new 3/260 with high-resolution monitor. Our version of
Xwindows (V10R4) doesn't work on this hardware, although it works on the
screen of 50's and 160's. Also, suntools works, so it's not a hardware or
a kernel problem.
The message I get when I run xinit is:
bw2_make size wrong 1 1600 1280
pixrect open failed for /dev/fb
I looked for this message in the sources and couldn't find it. adb was of
no help either. As I'm not an X expert (and don't intend to become one),
can someone tell me what the fix is ?
Marc Shapiro; INRIA Bat. 11; B.P. 105; 78153 Le Chesnay Cedex; France
tel. +33 (1) 39-63-53-25
uucp: seismo!mcvax!inria!shapiro; or: shapiro@inria.inria.fr
------------------------------
Date: Fri, 14 Aug 87 18:02:55 EDT
From: franklin@csv.rpi.edu (W. Randolph Franklin ( WRF ))
Subject: 32 bit disk controller for Sun 3?
A Xylogics salesman told me that his company has, and that Sun will later
this year release, a 32 bit disk controller for the VME bus. He says that
the current 16 bit controller is just a Multibus-VME converter.
It seems that the current 16 bit controller is a bottleneck given the 32
bit data bus and processor. Is this really so? If so, does anyone know
about the new 32 bit controller, whether it is worth waiting for, and
whether it can be retrofitted?
Wm. Randolph Franklin, RPI
Preferred net address: Franklin@csv.rpi.edu
Papermail: ECSE Dept, Rensselaer Polytechnic Institute,
Troy NY, 12180
Telephone: (518) 276-6077 ** Note the new exchange.
------------------------------
Date: Fri, 14 Aug 87 17:58:53 EDT
From: ihnp4!ho95b!gfm@seismo.css.gov
Subject: Securing SUN workstations?
We currently have several SUN 3/160C systems (with 141MB SCSI) and are
about to receive a couple of 3/280 file servers. Does anyone know of a
manufacturer that provides a physical key-locking mechanism that would
allow us to secure 3/160 and the 3/280 pedestal (i.e., prevent anyone from
using the system after hours) without resorting to a power down. Please
reply by email, I'll provide a summary if there's any interest.
George MacLachlan ihnp4!homxb!ho95b!gfm
------------------------------
Date: Fri, 14 Aug 87 09:13:52 EDT
From: Ned Danieley <ndd@sunbar.mc.duke.edu>
Subject: acquiring large amounts of data into Sun memory?
We have an data acquisition system that we would like to interface to a
Sun 3/160 using an IKON 10089 board (emulates a dr11w). We would like to
be able to acquire the data directly into memory. The 10089 has a 24 bit
word count, and we want to be able to get 24 bits (4 Meg) worth of data
without any paging. This obviously depends on having sufficient memory,
but it may also depend on a system constant. We bought a driver from Sun,
and it limits the transfer size to MAX_DR_BUF_SIZE. A comment says that
this can be no larger than DVMA size, less a little. DVMA is apparently 1
Meg. A request to transfer more than MAX_DR_BUF_SIZE results in multiple
calls to the strategy routine, which is probably going to be too slow (the
data rate is 264000 bytes per second). Is DVMA really a limit? If so, can
DVMA be increased? Or is there some other clever way to do this? We're
trying to avoid using a fifo or some other kind of temporary storage.
Ned Danieley (ndd@sunbar.mc.duke.edu)
Basic Arrhythmia Laboratory
Duke University Medical Center
Durham, NC 27710
(919) 684-6807 or 684-6942
------------------------------
Date: 14 Aug 87 17:13:11 GMT
From: km%emory@relay.cs.net (Ken Mandelberg)
Subject: Memory Expansion for Sun 3/60?
Maybe I'm a little premature here, but does anyone know about third party
sources for Sun 3/60 memory. It would really be nice if competition drove
those prices way down.
Ken Mandelberg | gatech!emory!km USENET
Emory University | km@emory CSNET,BITNET
Dept of Math and CS | km.emory@csnet-relay ARPANET
Atlanta, Ga 30322 | Phone: (404) 727-7963
------------------------------
Date: Fri, 14 Aug 87 09:59:56 GMT
From: lcp%computer-lab.cambridge.ac.uk@cs.ucl.ac.uk
Subject: Emacs icon?
I hear there is an icon for Emacs. Could someone please send a copy?
[[ I have not seen one pass this way all year. --wnl ]]
L C Paulson, U. of Cambridge (lcp@cl.cam.ac.uk)
------------------------------
Date: Fri, 14 Aug 87 14:27:55 EDT
From: cognac!elias@pt.cs.cmu.edu (Glenn S. Elias)
Subject: Idle logout program for Suns
Does anybody have an auto logout program for the suns? Since many
programs like Interleafs WPS leave /dev/console "idle", a program would
have to check all processes in a login. I assume that such a command would
work similarly to /usr/bin/screenblank that sun provides only logging out
the user instead of making the screen go blank.
Any help would be much appreciated.
Glenn S. Elias
Human Sciences
Westinghouse Research and Developement Center
cognac!elias@pt.cs.cmu.edu
------------------------------
Date: Wed, 12 Aug 87 09:53:47 EDT
From: rochester!srs!matt@seismo.css.gov (Matt Goheen)
Subject: Wasting time on a Sun3
Here is a little program I cooked up. This version is Sun3 only, if
I get enough responce, I'll adapt it to Sun2s (pretty trivial actually,
but see included README for details).
Matt Goheen
S.R. Systems
{seismo,rutgers,ames}!rochester!srs!matt
[[ This program was far too long to include here, but I am including part
of the README that describes what the thing does. The entire source is
available via anonymous FTP from "titan.rice.edu" under the name
"sun-source/brottool.shar". --wnl ]]
----------
Brottool is a little program that takes the following equation:
tmp = y - SIGN(x) * sqrt(fabs(b*x-c));
y = a - x;
x = tmp;
and goes nuts with it. Preferably, it should be started up like:
brottool -Ws 1100 700
since brottool scales its control panel according to the size of the frame,
this allows for greater accuracy when selecting the seed values.
Brottool clears the screen and restarts anytime one of the seeds changes,
the scale changes or the X or Y offsets change. It does NOT restart if you
scroll (the canvas is a "window" into the larger background, the X and Y
offsets control where in the background the program will plot, you can then
scroll your window around viewing the picture).
The f_sqrt() routine has restricted use because half of it was developed
on company time. Run "sccs what f_tab_sqrt.o" (once it is unpacked) on
the file to see the Copyright notice. The rest of the program is purely
public domain, although I would appreciate it if people sent me any neat
changes they make.
TAKE NOTE: brottool as distributed here will only run on a Sun3. That
is only because the f_sqrt() function uses some 68020
instructions. However, it wouldn't be much fun on a Sun2
anyway (too slow) so I didn't supply a Sun2 f_sqrt() routine.
Also, because the array used for plotting is so large (2 Meg),
most Sun2 systems will fall a bit short. This was developed
on Sun OS 3.2, no promises about other versions (no way for
2.x, probably works for any 3.x).
------------------------------
Date: Mon, 17 Aug 87 09:58:17 -0400
From: Chris Walters <walters%community-chest.mitre.org@gateway.mitre.org>
Subject: Self portrait
YAIC...
Usage (at least, how I use it) ---
suntools -pattern unixman
-- Chris Walters Dept:W94 Rm:H6225 MS:Z644
The Mitre Corporation
McLean, Va
[[ This really isn't an icon, but rather a background pattern. It is also
too large to include in the digest. It is in the anonymous FTP
collection, available as "sun-icons/cw-unixman.pat" (`pat' for pattern).
But I did get quite a kick out of it! --wnl ]]
------------------------------
Date: Wed, 12 Aug 87 09:32:08 EDT
From: blackje%sungod.tcpip@ge-crd.arpa (emmett black)
Subject: Yet another ICON
I got tired of seeing the same icon for my FTP and TELNET connections --
so I quickly threw together something I thought to be a bit more
appropriate (for FTP). .... and since we have begun to see an "ICON
Exchange" on this newsgroup, I thought I'd share it...
for what it's worth; --Emmett
BlackJE@GE-CRD.ARPA
...!steinmetz!crd!blackje
[[ In general, icon bits are small enough to include in the digest, and I
will probably continue to do so. But I'll try to put them at the very end
since the hex is rather boring to look at. I will continue to collect
them up in the sun-icons FTP directory. This one is available as
"sun-icons/jeb-ftp.icon". --wnl ]]
------------------*( cut here )*----------------------
/* An ICON for FTP: J.E.Black: Feb 1987 */
/* This version has a border at the edges */
/* Format_version=1, Width=64, Height=64, Depth=1, Valid_bits_per_item=16
*/
0x0000,0x0000,0x0000,0x0000,0x0000,0x0000,0x0000,0x0000,
0x7FFF,0xFFFF,0xFFFF,0xFFFF,0x4000,0x0000,0x0000,0x0001,
0x4000,0x0000,0x0000,0x0001,0x4000,0x0000,0x0000,0x0001,
0x4000,0x0000,0x0000,0x0001,0x4000,0x0000,0x0000,0x0001,
0x4000,0xA000,0x07CF,0x9E01,0x4015,0x5400,0x0402,0x1101,
0x402A,0xAA00,0x0402,0x1101,0x4055,0x5500,0x0402,0x1101,
0x40AA,0xAA80,0x0782,0x1E01,0x4155,0x5540,0x0402,0x1001,
0x42AA,0xAAA0,0x0402,0x1001,0x4155,0x5550,0x0402,0x1001,
0x42AA,0xAAA0,0x0402,0x1001,0x4555,0x5550,0x0000,0x0001,
0x42AA,0xAAA8,0x0000,0x0001,0x4555,0x5D50,0x0000,0x0001,
0x42AA,0xABA8,0x0000,0x0001,0x4555,0x55F0,0x0000,0x0001,
0x42AA,0xAAB8,0x0000,0x0001,0x4555,0x5557,0x0000,0x0001,
0x42AA,0xAAA0,0xC000,0x0001,0x4155,0x5550,0x3800,0x0001,
0x42AA,0xAAA0,0x0600,0x2A01,0x4155,0x5540,0x0181,0x5541,
0x40AA,0xAA80,0x0602,0xAAA1,0x4055,0x5500,0x1805,0x5551,
0x402A,0xAA00,0x300A,0xAAA9,0x4005,0x5000,0x0C15,0x5555,
0x4000,0x0000,0x038A,0xAAA9,0x4000,0x0000,0x0075,0xD555,
0x4000,0x0000,0x002F,0xAAAB,0x4000,0x0000,0x0017,0xD555,
0x4000,0x0000,0x002F,0xEAAB,0x4000,0x0000,0x0015,0xF555,
0x4000,0x0000,0x002A,0xAAAB,0x4000,0x0000,0x0015,0x5555,
0x4000,0x0000,0x000A,0xAAA9,0x4000,0x0000,0x0015,0x5555,
0x4000,0x0000,0x000A,0xAAA9,0x4000,0x0000,0x0005,0x5551,
0x4000,0x0000,0x0002,0xAAA1,0x4000,0x0000,0x0001,0x5541,
0x4000,0x0000,0x0000,0x2A01,0x4000,0x0000,0x0000,0x0001,
0x4000,0x0000,0x0000,0x0001,0x4000,0x0000,0x0000,0x0001,
0x4000,0x0000,0x0000,0x0001,0x4000,0x0000,0x0000,0x0001,
0x4000,0x0000,0x0000,0x0001,0x4000,0x0000,0x0000,0x0001,
0x4000,0x0000,0x0000,0x0001,0x4000,0x0000,0x0000,0x0001,
0x4000,0x0000,0x0000,0x0001,0x4000,0x0000,0x0000,0x0001,
0x4000,0x0000,0x0000,0x0001,0x4000,0x0000,0x0000,0x0001,
0x4000,0x0000,0x0000,0x0001,0x4000,0x0000,0x0000,0x0001,
0x4000,0x0000,0x0000,0x0001,0x7FFF,0xFFFF,0xFFFF,0xFFFF
------------------------------
End of SUN-Spots Digest
***********************