Sun-Spots-Request@RICE.EDU (William LeFebvre) (05/22/88)
SUN-SPOTS DIGEST Saturday, 21 May 1988 Volume 6 : Issue 94
Today's Topics:
Administrivia
Re: Darin Johnson's comments on DECnet
Re: 4 Fuji-2372's on a single xy451
MAXUSERS > 12 in SunOS 3.2 and 3.4
Anyone worked on inline libraries for SUN-3
Sun 3/50's: running 4.0, Clearpoint memory
Problems with subnetting and SunOS 3.5
Desperately seeking Sun network code patch!
NFS disk block sorting?
Is NFS "secure" in SunOS 4.0?
5 1/4'' drive on 386i?
Support for ISO 8859 on SUN?
rasterfile "shrinker"?
cmdtool windows?
Need help using SunCGI, mouse input, and multiple views
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 available via anonymous FTP from "titan.rice.edu".
For volume X, issue Y, "get sun-spots/vXnY". They are also accessible
through the archive server: mail the request "send sun-spots vXnY" to
"archive-server@rice.edu" or mail the word "help" to the same address
for more information.
----------------------------------------------------------------------
Date: Sat, 21 May 88 15:50:02 CDT
From: William LeFebvre <phil@Rice.edu>
Subject: Administrivia
We at Rice have been having some very serious networking problems of late.
For those of you who haven't noticed, bits coming in and out of the Rice
network have been doing slow very slowly, and in some cases not even at
all. It seems that the line between Rice and the IMP at UTexas is broken,
and Rice's network maintainers have been getting nothing but the runaround
when they try to get someone to fix it. Consequently, all of our bits
have to travel through NSFNet, and other problems exist there. I have no
idea when these problems will be fixed, but I do know that it has the
potential for getting worse. The now imminent midwest ARPANet shutdown
will render the UTexas IMP useless, and soon all of our traffic will be
forced to use NSFNet. I have been told that starting Monday, our traffic
on NSFNet will be gatewayed better (I'm trying very hard to avoid the
details here). I hope that by the middle of next week it will all be
better.
Unfortunately, I'm taking a Memorial weekend trip starting the middle of
next week and will be out for five days. If the network co-operates
between now and then, I should be almost completely caught up (have you
noticed that messages are now getting in to digests about 5 days after
they were sent?---it used to be a two week delay).
If you have been having problems FTP-ing to one of our machines, or
getting files from the archive server, you might want to try again after
Wednesday. I always get failed archive responses---usually from return
address that were incorrect. But lately I have gotten a few that were
returned because a host (in one case "sun.com"!) could not be reached for
7 days. These were due to the network problems. So please be patient,
and don't be surprised if no digests show up around the Memorial day
weekend.
Finally, there were a few files that I did not put in the archives as fast
as I announced them. My apologies. The summary of 4.0 is now available
in the directory "sun-spots" (note that this is the directory that
contains the backissues---NOT the source directory) as "4.0-notes.ms". It
is not a standard shar file: the leading "X" characters must be stripped
off by hand with "sed -e 's/^X//'" (this is the way it was sent to me).
The file "sun-source/braggtools.shar" has finally been updated: the
original shar file sent to me was incomplete. The shar files for
"calctool2" are now in the "sun-source" directory. Again I apologize for
the delay in making these available.
William LeFebvre
------------------------------
Date: Tue, 17 May 88 14:27:30 EDT
From: gfr%wolfgang@gateway.mitre.org (Glenn Roberts)
Subject: Re: Darin Johnson's comments on DECnet
Reference: v6n88
> >This product is a bit disappointing in terms of breadth of capabilities.
> >I have worked with DEC's DECnet DOS (for the PC), and feel that it offers
> >a much nicer set of capabilites, including sending E-mail, copying and
> >submitting batch files, and disk and printer peripheral sharing.
> ?? Gosh, you mean that you can do more with DECnet DOS than you can with
> the DECNET on VMS??? Without resorting to clusters, how can you submit
> batch files and share disks and printers between vaxen?
I didn't say you could do more with DECnet DOS than with the DECnet on
VMS. I *did* say you could do more with DEC's DECnet DOS on a PC than
with Sun's SunLink DNI on a Sun, including:
1. NFT COPY/SUBMIT to copy and submit a .COM file from a PC
2. E-mail (one way only - supplied with DEC's package)
3. Use NDU to make VAX printer or disk look like local PC peripheral
It is interesting that someone from DEC is *complaining* that I like DEC's
networking product better than Sun's!
Glenn Roberts, MITRE Corporation, McLean VA
703-883-6820
gfr%wolfgang@gateway.mitre.org
------------------------------
Date: Mon, 16 May 88 15:31:12 EDT
From: umix!bio-image.bio-image!mike@rutgers.edu (Mike Bernson)
Subject: Re: 4 Fuji-2372's on a single xy451
Reference: v6n76
There is a bug in pagging system that will cause any program that is not
in the first 512meg of the disk to break. This could cause the program you
are having with the 2372 disk drives.
Mike Bernson
------------------------------
Date: Wed, 18 May 88 09:15:36 EDT
From: ethan miller <emiller@berke.bbn.com>
Subject: MAXUSERS > 12 in SunOS 3.2 and 3.4
Reference: v6n89
->The failure only happens if you have enough physical memory. We had no
->problem with 8MB suns, but did for 16MB.
Actually, for 16Mb, the limit is at least 14 users. We crashed with 16,
but when we lowered it to 14, it worked fine. 15 may or may not work (I
don't remember what the Sun rep said).
ethan
ethan miller
BBN Laboratories
ARPAnet : emiller@bbn.com
PHONEnet: (617) 873-3091
------------------------------
Date: Wed, 18 May 88 16:59:22 +0300
From: leonid@TAURUS.BITNET
Subject: Anyone worked on inline libraries for SUN-3
Given the Inline facility with the SUN C compiler, it seems possible and
worthwile to rewrite some of the more popular C-library functions in
MC68000 assembly in the form of an inline library. I am currently
considering writing the string(3) functions this way. Has anyone already
done this, or something similar ?
If you have interest in such a thing, please drop me a line. If this
proves to be of interest to the net-world, I'll post it after it's tested.
------------------------------
Date: Tue, 17 May 88 13:35:12 PDT
From: rusty%math.Berkeley.EDU@cartan.berkeley.edu
Subject: Sun 3/50's: running 4.0, Clearpoint memory
Someone at Sun told me that 4.0 will run on 3/50's (i.e., only 4 meg) but
that it is "sludgy".
I just got off of the phone after talking to Michelle Sieber of ClearPoint
(408-727-2433). She said that they are getting ready to start selling
expansion memory for 3/50's. Good timing, eh? She didn't have any
prices, availability, etc. yet.
------------------------------
Date: Tue, 17 May 88 13:42:54 EDT
From: fedor@nisc.nyser.net (Mark S. Fedor)
Subject: Problems with subnetting and SunOS 3.5
Hi, I have noticed some strange behaviour in regards to subnetting on a
couple of Sun machines running SunOS 3.5.
It has to do with the following simple Scenario:
Clients Clients Clients
| | | | | | | | |
==================================== 128.230.3.0
|
Server
|
==================================== 128.230.7.0
|
Gateway
|
==================================== 128.230.1.0
Now all of the Sun machines shown above run SunOS 3.5 and have properly
configured subnet masks (255.255.255.0).
The clients have a default route to the server, that's it. (besides the
interface routes)
When I try to talk to a machine from a client off of subnet 128.230.3.0,
*the client ARPs for the machine* thinking that the machine from the other
subnet is on its cable. In other words, it seems like the clients think
that 128.230.0.0 is NOT subnetted even though the interface parameters are
set right.
Example:
- 128.230.3.15 (diskless client) tries to connect to 128.230.7.3
- 128.230.3.15 has a default route to the server (128.230.3.10)
- 128.230.3.15 ARPs for 128.230.7.3's hardware address.
- Of course, no one will answer 128.230.3.15's ARP request.
- The connection fails.
128.230.3.15 should not ARP for 128.230.7.3's hardware address, but should
ARP for the next hop gateways address if needed (128.230.3.10). The only
reason I can think of that would make 128.230.3.15 incorrectly ARP would
be that 128.230.3.15 thinks its interface to 128.230.0.0 is not subnetted.
Please note that this behaviour is only being exhibited by the diskless
clients and not the server. This makes it more interesting.
This is real strange. Is there a bug fix that I missed? It seems too
serious a bug to have gone unnoticed this long. Am I missing something in
my configuration?
I thought that my configuration was in some way messed up so I checked on
a totally different site running SunOS 3.5 and the same behaviour was
shown. That prompted me to write this letter.
Any info from anyone would be great.... And it would also cure some
headaches...
Thanks.
Mark Fedor
NYSERNet, Inc.
fedor@nisc.nyser.net
------------------------------
Date: 17 May 88 20:10:05 GMT
From: masscomp!garyo@uunet.uu.net (Gary Oberbrunner)
Subject: Desperately seeking Sun network code patch!
I heard through the grapevine that there is a patch to fix the ethernet
backoff problem in recent SunOS releases for the 3/60. Our symptom is
that under rlogin TO the Sun, screen update becomes extremely bursty, with
about 100 characters per burst and about 3-5 seconds between bursts.
Typing any character causes the next burst to come out, but obviously
that's not a real solution, since I'd have to type continuously to get
decent screen update.
I hear that this is a known bug in the net code, and that perhaps there's
a patch for it. If you have such a patch, or know where to get it via
ftp, please PLEASE mail it to me. We need it desperately here, since the
Suns are sort of useless with rlogin without it.
>> Please EMAIL to me, don't post, since I don't read this group. <<
My address is: ...{harvard,uunet,ulowell,allegra}!masscomp!garyo.
Thank you very much -- I really appreciate your help.
Sincerely,
Gary
x2445
Gary Oberbrunner
...!masscomp!garyo
....garyo@masscomp
------------------------------
Date: 17 May 88 22:20:31 EDT (Tue)
From: encore!cloud9!bob@talcott.harvard.edu (Bob Toxen)
Subject: NFS disk block sorting?
Reference: v6n65
When running NFS we have observed that the client requests blocks from the
server *after* sorting disk blocks for disk latency, i.e. block 0, 10,
20, 1, 11, 21, 2, 12, 22 ... How can one stop this so that blocks are
requested 0, 1, 2, 3, ... Our client is a Sun; our server ain't. (We
want this because we are handling the disk latency at a lower level.) I'm
sure the answer is easy. I just don't know it.
Please reply via mail. Thanks.
Bob Toxen {ucbvax!ihnp4,harvard,cloud9!es}!anvil!cavu!bob
Stratus Computer, Marlboro, MA
Pilot to Copilot: What's a mountain goat doing way up here in a cloud bank?
------------------------------
Date: Wed, 18 May 88 0:55:23 CDT
From: Naim Abdullah <naim@eecs.nwu.edu>
Subject: Is NFS "secure" in SunOS 4.0?
We haven't yet received our SunOS 4.0 upgrade but I have heard that it
incudes "secure RPC". And NFS in SunOS 4.0, is an application that uses
secure RPC.
I am unsure about what exactly this means. As far as I understand, secure
RPC uses encryption so it will no longer be possible to breach security by
using etherfind(1) or a network tap analyzer to watch data packets fly by
on the ethernet.
But Sun documents claim that secure RPC gives you the security level of a
conventional timesharing system. It seems that this claim is not true
regarding NFS because anybody with a workstation can assume any uid on
that workstation and if you export any file systems to that workstation,
that user can read/write any files (except those owned by uid 0). This is
certainly not the security level of a conventional time sharing system. Am
I missing something or is this matter of NFS being a user of secure RPC,
just gaining us security from ether watchers ?
Also, is the *server* allowed to specify "read-only" when it exports a
file system in SunOS 4.0 ? This is a feature that Pyramid's NFS has and I
wish Sun's implementation would have too.
Naim Abdullah
Dept. of EECS,
Northwestern University
Internet: naim@eecs.nwu.edu
Uucp: {ihnp4, chinet, gargoyle}!nucsrl!naim
------------------------------
Date: Tue, 17 May 88 16:26 EDT
From: Neil Bodick <Bodick@cis.upenn.edu>
Subject: 5 1/4'' drive on 386i?
Has anyone hooked up a 5 1/4" drive to a Road Runner on its AT bus? We're
interested, but worried that excessive interrupt latency or perhaps DMA
problems might cause an off the shelf drive not to work. Our loyalty to
Sun is so great that we don't have a single PC about to strip for parts.
Thanks.
Andre Marquis
bodick@cis.upenn.edu
------------------------------
Date: 17 May 88 12:25:16 GMT
From: mcvax!krafla!magnus@uunet.uu.net (Magnus Gislason)
Subject: Support for ISO 8859 on SUN?
Does anyone know whether/when Sun will support the ISO 8859 standards?
(In particular ISO 8859/1) ISO 8859 are international standards for coding
of non-English languages, ISO 8859/1 is for Western European languages.
------------------------------
Date: Wed, 18 May 88 1:02:06 CDT
From: Naim Abdullah <naim@eecs.nwu.edu>
Subject: rasterfile "shrinker"?
Recently there was a program called "pic2icon" that was posted to this
newsgroup. It translated a raster file to an icon. Unfortunately most of
the icons generated by this program are unsuitable because they are too
big.
What we need is a program that would "shrink" a rasterfile which can then
be handed to pic2icon to generate an icon. Obviously there would be loss
of information in the shrinkage but we'd be satisfied with a "best effort"
by the program.
I asked an image processing friend and he said something about "a digital
filter and then undersampling the image". Aparently the process is
nontrivial. So has somebody done it already and is willing to share it
with us ?
Naim Abdullah
Dept. of EECS,
Northwestern University
Internet: naim@eecs.nwu.edu
Uucp: {ihnp4, chinet, gargoyle}!nucsrl!naim
------------------------------
Date: Wed, 18 May 88 19:05:49 BST
From: Richard Tobin <richard%aiai.edinburgh.ac.uk@nss.cs.ucl.ac.uk>
Subject: cmdtool windows?
Does anyone know how to create a tool containing a cmdtool-type terminal
emulator? It should obviously be possible to use the same sort of code
the cmdtool does, but in 3.5 /usr/src/sun/suntools/cmdtool.c doesn't even
compile.
Actually, I know part of the answer to this. Declaring ttysw_cmdsw as an
external is enough to make it compile, but it then core dumps. Changing
the assignment to ttysw_cmdsw to assign a text subwindow to it (instead of
the integer 1 as in 3.2) makes it work, but I get a notifier error
"unknown client".
Anyone know what's really going on?
-- Richard
Richard Tobin, JANET: R.Tobin@uk.ac.ed
AI Applications Institute, ARPA: R.Tobin%uk.ac.ed@nss.cs.ucl.ac.uk
Edinburgh University. UUCP: ...!ukc!ed.ac.uk!R.Tobin
------------------------------
Date: 18 May 88 18:20:53 GMT
From: news@hubcap.clemson.edu (news)
Subject: Need help using SunCGI, mouse input, and multiple views
I am facing a problem of debugging a piece of code that involves using
SunCGI, mouse input, and multiple view surfaces.
1) When I wrote the program with just one view surface, setup_sigwinch(),
initialize_mouse..., etc. the program works fine. The problem is that I
am not able to debug it with gfxtool and dbx. when I do use dbx, pressing
the mouse button shows:
signal IO (possible input/output) in syscall at 0x709a8
syscall+0xc: bcc syscall+0x1a
(dbx)
and when I said "cont" , it goes into request_input(), and pressing a
mouse button still gives the same message. It is suppose to run the
function I requested on the panels.
2) With multiple view surfaces, the SunCGI(?), mouse(?), doesn't respond
to pressing the buttons at all. Using dbx with the current version of the
program shows the same message. and using print ivalue.xypt -> x gives
the same point I initialized. Arrrgh... :-(
3) Another problem when it comes to multiple view surfaces in SunCGI is
that the redraw() that I set up for SIGWINCH doesn't redraw the pictures
properly which was working before in the version with only one view
surface.
Thank you for any suggestions, comment, etc.
John K. Sun
Clemson University
Computer Science Dept.
Clemson, SC 29634
...!gatech!hubcap!scarle!citron
------------------------------
End of SUN-Spots Digest
***********************