Sun-Spots-Request@RICE.EDU (William LeFebvre) (02/25/88)
SUN-SPOTS DIGEST Tuesday, 23 February 1988 Volume 6 : Issue 19
Today's Topics:
Re: New Fujitsu drive
Re: ADA compiler on the SUN
Re: Why doesn't SunOS Unix use hardware FPA when it's available
Re: Spurious level 3 interrupts!
Thanks for ND info
mods for calentool
block.c and 12" 1600BPI dumps
Two strange problems: PC-NFS and stuck in caps lock
Problems creating the mail spool areas
Sun 4 Multiuser SURVEY
fonts newsgroup?
Graphics Processor available?
HDB UUCP For Sun-3?
Multi-process debugging using Dbxtool?
24-bit video for Sun 3 or 4?
Expanding file serving capacity?
screendump from SUN 3/110 to laserprinter; best way?
Using Sunlink-X25?
Capacity of CDC Wren IV, diag/format problem?
SLIP and Smalltalk for Suns?
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: Wed, 10 Feb 88 7:15:17 EST
From: "Tim G. Smith" (Mechanical) <tsmith@cad.usna.mil>
Subject: Re: New Fujitsu drive
From what I have been told the controller that sun now supports is the
Xylogics 7053 which has a 2.4MB/s max transfer rate. The next model up is
the 7054 which supports 3.0MB/s. The 7053 is the "official sun" hardware
as the is the driver that sun sells. Xylogics is supposed to upgrade the
7053 to 3.0MB/s real soon like.
NOTE: All this info is hearsay from a sales rep at Software Associates. I
am only passing on what I have been told- I cannot vouch for the accuracy
of the above info. Hope this helps, Tim Smith
E-Mail :<tsmith@usna.mil>
US mail:Tim Smith
CADIG mailstop: 11G
US Naval Academy
Annapolis, MD 21402
Tel # :(301)267-4413
------------------------------
Date: Wed, 10 Feb 88 09:49:29 -0800
From: Craig Rolandelli <craig@ics.uci.edu>
Subject: Re: ADA compiler on the SUN
There is a company in Irvine, Ca. called ICC (Irvine Compiler Corp.).
They make an ada compiler the produces C code, 68020 code, or VAX code.
Their compiler is really fast, and produce code that is compact and fast.
They have an account here at U.C. Irvine that they use for mail. So you
can send mail to icc@rome.uci.edu and ask them for more info. A word of
warning about the Verdix compiler version 5.41. The compiler (Sun and
Sequent versions) has a bug that causes the compiler not to allocate
enough memory for variant records. You can work around this be putting in
dummy variable as long as the variant record is not part of a linked list.
I am sure this will be fixed in the next release, but if you decide to go
with Verdix, make sure and ask them about it.
Craig Rolandelli (craig@ics.uci.edu; {sun!sunkist,ucbvax}!ucivax!craig)
Computer Science Department
University of California, Irvine
Irvine Ca. 92717
(714) 856-4222
------------------------------
Date: Wed, 10 Feb 88 08:56:45 PST
From: Matt Wette <mwette%newton@hub.ucsb.edu>
Subject: Re: Why doesn't SunOS Unix use hardware FPA when it's available
The Sun OS kernel (and probably most of the utilities) are compiled
"-fsoft". There is at least one good reason for this as explained in the
"Sun Floating- Point Programmer's Guide" on page 7. The Sun FPA can only
support 32 simul- taneous processes. My guess is that "-fsoft" is a good
choice since kernel software is probably not too floating point intensive.
Question: Does "-fsoft" save time on context switches over "-ffpa"?
Matt
Matthew R. Wette
Scientific Computation Lab
Dept. Elec. and Comp. Eng.
University of California
Santa Barbara, CA 93106
ARPA: mwette%gauss@hub.ucsb.edu
UUCP: ucbvax!ucsbhub!gauss!mwette
------------------------------
Date: Thu, 11 Feb 88 09:58:28 PST
From: Jonathan Eisenhamer <jon@mira.astro.ucla.edu>
Subject: Re: Spurious level 3 interrupts!
To All Spots,
A couple of people have asked me for any info I have learned about level 3
interrupts. Since I seem not to be able to reach any of them, and I have
only received one reply, I felt that the whole net should have it. I have
not been able to pursue the situation, so I don't know whether the below
explanation works for my situation or not (I assume it does).
Jonathan Eisenhamer
UCLA Astronomy
jon@mira.astro.ucla.edu
jon@uclastro.bitnet
bonnie::jon (span 5828)
Forwarded message from: COM%"pms@Sun.COM"
Jonathan:
I believe that this message ("Spurious level n Interrupt") is printed by
the kernel when it gets an interrupt (from a board) that it doesn't know
what to do with. There are two reasons that spring to (my) mind for this:
1) a board is in the bus creating these interrupts, but hasn't been
properly (or at all) configured into the kernel; and 2) the interrupt is
being caught and dispatched to the right driver, but the driver isn't
ready for it. I suggest that you double check your configuration file
(/sys/conf/UPPERCASENAME) & see which boards are coinfigured with level 3
priority against which boards actually produce these interrupts (usually
set by jumpers on the board), and also check the drivers and hardware
manuals for the boards to see when they should and shouldn't be
interrupting. I hope this helps and isn't the n'th reply you've had!
patrick
------------------------------
Date: Wed, 10 Feb 88 18:31:24 EST
From: Mike Jipping <jipping@relay.cs.net>
Subject: Thanks for ND info
I want to say a public thanks to those who responded to my query about
adding a node to the ND partition. I got more replies than I had expected
-- about 29 or 30 -- and I now know lots more than I did about partitions
and ND allocation. This information will come quite handy for about four
or five more months, now that 4.0 won't use 'em.
Basically, there were three types of replies.
(1) Yes, I screwed up. Repartition your disk to make space and reset
the ND partitioning with SETUP -- it's the best bet for a rookie!
(2) Yes, I screwed up. Repartition and reset ND by hand! Here's how...
(3) Hey, you didn't screw up. Use a free, unused partition NEXT to
the ND partitions.
Because I'm a rookie to Sun Administration Land, I opted for (1). By
selecting NO optional software, I got setup to run fairly quickly, and I
was up a half a day later. I was a bit wiser this time...
Thanks to everyone again for the help. It's comforting to know there's
more out there than just Sun technical support.
-- Mike Jipping
Dept of Computer Science
Hope College
jipping@cs.hope.edu
------------------------------
Date: Wed, 10 Feb 88 14:24:30 PST
From: Bill Randle <billr@tekred.tek.com>
Subject: mods for calentool
A recent query to Sun-Spots asked if anyone had modified calentool (from
the 1987 SUG Tape) so that it understands leap years. I made this fix,
plus several others, which I am including here (see README2 for details).
The mods have been submitted to the SUG for futute tapes, as well as the
author.
-Bill Randle
Tektronix, Inc.
billr@tekred.TEK.COM
(The diffs are from the source as it comes off the tape.)
[[ The shar file also includes a program called "calenmail" which does as
the name implies. It is 47536 bytes in length and is stored in the
archives as "sun-source/calentool-diffs.shar". It can be retrieved via
anonymous FTP from the host "titan.rice.edu" or via the archive server
with the request "send sun-source calentool-diffs.shar". For more
information about the archive server, send a mail message containing the
word "help" to the address "archive-server@rice.edu". --wnl ]]
------------------------------
Date: Tue, 9 Feb 88 15:06:01 est
From: Sheila Hollenbaugh <shollen@thor.wright.edu>
Subject: block.c and 12" 1600BPI dumps
I recently obtained block.c from the archives, and have been attempting to
use it as directed, and with other block sizes. The results are as
follows:
1) The dump is declared to be done, possibly in record time.
2) The amount of tape space taken up by the dump is not
consistent with what the dump program says it should be.
3) The top level directory hierarchy seems to be as it should
be, but restore -i says that the top level file names "are not
directories".
We are running SUN OS 3.4 on a 3/180S, with a 1600BPI tape drive. We have
the same tape drive on two VAXen running ULTRIX 2.0, and we know for a
fact the the drive is capable of amazing speed. I was hopeful that
block.c would do the job on the Suns.
Has anyone gotten block.c to speed up dumps under similar conditions? If
so, you would earn my everlasting gratitude if you would enlighten me. I
am tired of spending 7+ hours a month doing epochal backups on this
system.
Sheila Hollenbaugh Wright State University Research Center
CSNET: shollen@SPOTS.Wright.Edu 3171 Research Blvd., Kettering, OH 45420
UUCP: ...!cbosgd!wright!shollen (513) 259-1382
------------------------------
Date: Tue, 09 Feb 88 14:20:48 CST
From: AARON KONSTAM <79343382@TRINITY.BITNET>
Subject: Two strange problems: PC-NFS and stuck in caps lock
I have two strange problems that I would like some help on. First, we are
running PC-NFS connected to a SUN 3/160 running SUNOS 3.2. We were having
no problems until we tried to run Turbo-C on the PC using the Sun to
support a C: disk where the Turbo-C ssytem resides. While running the
Turbo-C system can't seem to access the pseudo-C: disk on the Sun and the
following two messages appear on the Sun console over and over again.
xdr_bytes: bad size Failed svckudp_send: xdr_replymsg failed
What does this tell me? We have no trouble accessing the same disk using
other programs.
Second, there is some combination of keys that one can hit on the keyboard
that puts one irreversibly in upper case. When this happens and one is in
vi getting out of vi using ZZ puts you back in normal lower case. What
keys are being hit and how does one reverse this effect? You can answer
me directly if you think the answers are too trivia or too obvious to
interest every one else.
Aaron Konstam
Trinity University
San Antonio, Texas
79343382@trinity.bitnet
------------------------------
Date: 10 Feb 88 03:56:31 GMT
From: Andrew D. Bowen <psuvax1!adbst@cisunx.cs.psu.edu>
Subject: Problems creating the mail spool areas
We have had some bizarre problems with our mail system here at Pitt. We
have a server, (both disk and YP), called Jupiter and 4 clients, Io,
Europa, Ganymede and Callisto. Everyone's rc.local file begins with a
system call 'domainname jove'. So, we replaced 'sun' with 'jove' in the
sendmail.cf files on all the machines. Jupiter is listed as the mailhost
in the hosts file and ypmatch confirms it over the YP.
Now the problem.....
If a user sends mail from a client, and the recipient does not have a
space in the /usr.MC68020/spool/mail area, then the file is created, but
is owned by 'nobody'. We did some experiments and discovered that any
file created by root in that area on a client is owned by 'nobody' and
root cannot chown it to anyone else. Root on Jupiter does not encounter
these problems. /usr/spool/mail on the clients is a mounted file system
as per SUN's instructions.
Our Solution:
On pages 31-32 in the System Administrator's Guide, there is a patch that
can be made to nfs_server.o in the kernel for the NFS server machine. The
patch reassigns nobody to be some other value, presumablly root. This
'fix' solved our problems, but now we are wondering if that was the proper
fix, and/or is there a better fix.
Thanks in advance.
- Andy Bowen
University of Pittsburgh
Dept. of Electrical Engineering
------------------------------
Date: 12 Feb 88 22:24:20 GMT
From: amdcad!amdcad.AMD.COM!cdr@ames.arc.nasa.gov (Carl Rigney)
Subject: Sun 4 Multiuser SURVEY
Does anyone out there have any experience with Sun-4s running lots of
users (20+, or even 10+)? Information about a development environment
would be ideal (edit-compile cycles), but I'll take what I can get. I'm
also interested in hearing any stories about running news and rn on Suns,
particularly the Sun-4. Any problems or things to watch out for?
How's Sun's support been?
Please MAIL to me, DO NOT POST. If you're interested in a summary send me
mail and I'll summarize. If you don't want your remarks to appear in a
summary mention it in your reply.
Many thanks.
--Carl Rigney
USENET: {ames decwrl gatech ihnp4 pyramid sun ucbvax}!amdcad!cdr
cdr@amdcad.AMD.COM
USMAIL: MS 167; 901 Thompson Place; Sunnyvale, CA 94303
PHONE: 408-749-2453
------------------------------
Date: Tue Feb 9 18:19:57 1988
From: richard@gryphon.cts.com (Richard Sexton)
Subject: fonts newsgroup?
Are you sitting down? This massive crosspost actually has *technical
content*.
[[ It was posted to 12 different Usenet lists. --wnl ]]
There are many aspects to fonts that are device and computer independent.
It is my opinion that the exchange of fonts between different computers
should be more prevelant than it currently is.
And then there is the artistic side.
And the legal side.
If you are interested in fonts and would like to see a newsgroup dedicated
solely to that - drop me some email.
THIS IS NOT A REQUEST FOR VOTES. That will occur later if it seems like a
good idea.
"It's too dark to put my legs in my munitions"
richard@gryphon.CTS.COM
{ihnp4!scgvaxd!cadovax, rutgers!marque, codas!ddsw1} gryphon!richard
[[ Please mail responses to this to Richard and not to Sun-Spots. Is there
someone in ARPANet-land who is willing to moderate an ARPANet mailing list
about fonts? The two could "link up" in a similar way to Sun-Spots and
comp.sys.sun. --wnl ]]
------------------------------
Date: 10 Feb 1988 08:54-EST
From: Hans.Tallis@ml.ri.cmu.edu
Subject: Graphics Processor available?
I'm looking for a cheap software test to determine whether the current
machine has a graphics processor board installed. Currently I'm trying to
open the /dev/gpone0a file, which is rather kludgey. All pointers
appreciated. --Hans
------------------------------
Date: Tue, 9 Feb 88 9:32:26 EST
From: scs%lokkur.UUCP@umix.cc.umich.edu (Steve Simmons)
Subject: HDB UUCP For Sun-3?
Does anyone out there have a version of HDB uucp which they've ported to
the Sun-3 series? We're fully licenced for AT&T V.3 so legalisms will be
covered properly. We could do it ourselves, but if someone else has
already invented the wheel....
Steve Simmons Work: Schlumberger CAD/CAM
Voice: 313-995-6366 Reality: scs@lokkur.uucp
------------------------------
Date: Wed, 10 Feb 88 16:19:41 EST
From: ha@purdue.edu
Subject: Multi-process debugging using Dbxtool?
I read a paper[1] on Dbxtool which says that Dbxtool provides capabilities
to debug multiple processes, particularly processes that are not the
children of the debugger. It gives me the impression that while debugging
a program using Dbxtool, we can have a new Dbxtool window come up when the
current process forks a child process, so that from then on, both the
processes (the parent and the child) could be debugged in parallel in the
two Dbxtool windows. That is, at any time, there could be as many
instances of Dbxtool as there are active processes belonging to the
program.
However, I didn't find anything of that sort while debugging a program
that used fork and exec. Neither the man pages for Dbxtool or Dbx, nor the
document 'Debugging Tools for the Sun Workstation' talk about
multi-process debugging anywhere.
Is it that the above features have only been proposed/designed but not yet
implemented, or am I missing something? (Our Sun 3/50 runs Sun UNIX 4.2
Release 3.4.)
I would appreciate any help/references/pointers on this.
[1] Evan Adams and Steven S. Muchnick, "Dbxtool: A Window-Based
Symbolic Debugger for Sun Workstations", Software-Practice
and Experience, vol. 16, no. 7, pp. 653-669, July 1986.
------------------------------
Date: Wed, 10 Feb 88 17:34:26 EST
From: eric@tcgould.tn.cornell.edu (Eric Fielding)
Subject: 24-bit video for Sun 3 or 4?
I am interested in any rumors or plans people have heard about for 24 bit
video cards for the Sun 3 or 4. I am doing work combining satellite
imagery, which requires more than 8 bits of color, and vector drawing.
++Eric Fielding
eric@tcgould.tn.cornell.edu eric@crnlthry.bitnet !cornell!batcomputer!eric
------------------------------
Date: Wed, 10 Feb 88 23:27:05 PST
From: rob@presto.ig.com (Rob Liebschutz)
Subject: Expanding file serving capacity?
We are currently using a Sun 3/160 as a file server for 10 diskless
clients (4 3/50's, a 3/75, a 3/160, a 3/110's and 3 Sun 2's). We expect
to add several additional diskless machines in the next few months. Can
anyone comment on how effective it would be to upgrade the 3/160 to a
3/260 or a 4/260 for the purpose of increasing file serving capacity as
opposed to adding another file server? What sort of bottlenecks are
causing need for increased file serving resources? On a big file server
such as a 4/260, is it worthwhile to use multiple ethernet interfaces (on
seperate subnets)?
What kind of results are people having using machines like 3/280's and
4/280's as combination time sharing machines and file servers? Is it
reasonable to have 5-6 diskless clients on a 3/280 in addition to
interactive users (Note, these machines are different than those described
above; they would be 3/60's with at least 12 MB of memory)?
What about offloading large CPU intensive jobs to file servers with extra
CPU capacity? How might this effect performance of diskless machines?
Rob Liebschutz
IntelliGenetics
------------------------------
Date: Thu, 11 Feb 88 08:02:50 est
From: jsh@research.att.com (Juergen Schroeter)
Subject: screendump from SUN 3/110 to laserprinter; best way?
I need some ideas on the optimal way of getting screendumps from
a SUN 3/110G (greyscale, color) to some hardcopy device.
Are there laserprinters which can do greyscale?
Does anyone know about good quality color hardcopy devices?
A solution where the hardcopy device would be available to all
workstations on the local net would be best.
Thanks.
Juergen Schroeter, AT&T Bell Labs, Murray Hill, NJ 07974
(201) 582-7059; ....!alice!jsh
------------------------------
Date: Tue, 2 Feb 88 16:31:45 GMT
From: decvax!cvbnet!cvedg!kquinlan@decwrl.dec.com (Kevin Quinlan {SDQA-E Penn St. x269})
Subject: Using Sunlink-X25?
We are trying to set up a link using X25 between Boston and Amersham
England. We have Sunlink-x25 at both ends running on Sun 2's via the tty
ports. The problem is how can you use it sensibly? As far as I can tell
it is not possible to pipe stuff thru' the software pad and running an
internet router all the time is a non-starter due to the cost ($6.0 /
hour) of keeping the connection up, even with no data traffic.
The question: does anyone out there know of a way of setting this
software up to allow start-stop connections so that the connection is up
only when data is flowing?
Regards
Kevin Quinlan
Computervision {decvax,sun}!cvbnet!cvedg!kquinlan
Penn St, Amersham, + 44 494 714771
Bucks, HP7 0PX
England
------------------------------
Date: 10 Feb 88 11:57:21 GMT
From: Hans van Staveren <mcvax!cs.vu.nl!sater@uunet.uu.net>
Subject: Capacity of CDC Wren IV, diag/format problem?
We recently acquired a CDC Wren IV diskdrive and put it to work. That
worked quite well, as long as you run SunOs 3.4, but after formatting the
disk with diag we "only" had ~285Mb capacity, while the specs claim
~307Mb. There is however something funny about sector sizes in the manual
of the drive, and I wonder whether I formatted it wrong.
Does anyone know the nitty gritty details about this?
Hans van Staveren
Vrije Universiteit
Amsterdam, Holland
------------------------------
Date: Fri, 12 Feb 88 11:17:10 PST
From: trwrb!smpvax1!sbb@ucbvax.berkeley.edu
Subject: SLIP and Smalltalk for Suns?
I have a couple of questions that someone in the Sun-Spots readership
ought to know the answer to:
1) Is there a public domain Smalltalk available for Suns? If not, is there
someone at PARC that can be contacted regarding licensing?
2) (totally unrelated) Is there some specification for SLIP, the serial
line IP? Is there an RFC that talks about it?
Any information that you can provide will be greatly appreciated.
steve
Steve Byrne ...ucbvax!trwrb!smpvax1!sbb
Inference Corp. (213) 417-7997
5300 W. Century Blvd.
Los Angeles, CA 90045
------------------------------
End of SUN-Spots Digest
***********************