Sun-Spots-Request@RICE.EDU (Vicky Riffle) (06/23/87)
SUN-SPOTS DIGEST Tuesday, 23 June 1987 Volume 5 : Issue 20
Today's Topics:
Re: Dynamic SunView object generation? (1)
Re: Dynamic SunView object generation? (2)
Re: DUMP tapes created on a VAX under 4.3BSD?
Re: Undocumented le0: Messages?
Re: query on dump to sun scsi tape?
3rd party disk drives
bug in SUN 3.x lint
CTS/RTS
IPC does not work on 3/280
4.3bsd tcp code for Sun available
xy450/xy451 failure modes
Why can't we have a Spectrum area?
questions about new 3/280?
SCSI driver for a OSI LaserDrive 1200 optical disk?
Help with Xylogics 472 tape controller and Multi-Vme board?
Sun OS 3.2 timed?
NFS strangeness on Suns?
Mail from Sun to Explorer?
Sun IPC board config. problems?
Undumping on a Sun III work Station?
SunCore Questions?
------------------------------------------------------------------------
Date: Wed, 17 Jun 87 23:26:20 PDT
From: weiser.pa@xerox.com (Mark Weiser)
Subject: Re: Dynamic SunView object generation? (1)
The game for the Sun I announced a few weeks ago on sun spots creates
quite a few dynamic windows--a silly number, perhaps. You might want to
look at the source, which is available for anonymous ftp on
mimsy.umd.edu in file sdi.shar.Z. (I am trying to get it to post to
net.sources.games, but the moderator of that group cannot get it to run
on his sun, and quite rightly is reluctant to post what to his mind may
be a rotten game. On the other hand, I have gotten picture postcards
from all over the world from people who HAVE successfully run it, so I
am not sure what is wrong.)
-mark
------------------------------
Date: Thu, 18 Jun 87 10:02:24 EDT
From: cb@mitre-bedford.arpa (Christopher Byrnes)
Subject: Re: Dynamic SunView object generation? (2)
With the exception of temporary response menus (created within windows
started with window_loop()), I've never had any success in dynamicly
creating SunView frames and windows from within the same UNIX process. I
don't know about your application, but mine required that the dynamic
frames/windows be very independent of each other (independently scrollable,
resizeable, destroyed, etc.). Trying to do this in the same process always
lead to trouble.
I finally had to make each dynamic frame/window into its own UNIX
process, started by a call to wmgr_forktool(). SunView seemed to handle
this situation without problems. The solution to your problem will depend
on how tightly coupled your frames and windows must be.
(the opinions of) Christopher Byrnes
(and not the opinions of) The MITRE Corporation
Burlington Road
M/S A156
Bedford, Mass. 01730
cb@Mitre-Bedford.ARPA
...!decvax!linus!mbunix!cb.UUCP
------------------------------
Date: Fri, 19 Jun 87 07:48:42 cdt
From: ables%pp.mcc.com@mcc.com (King Ables)
Subject: Re: DUMP tapes created on a VAX under 4.3BSD?
Suns and Vaxes have what is termed "the byte swap problem."
The hardware's idea of which bytes of a word are low and
high order is reversed between the two machines, hence when
things are written out in "image" form, the bytes are swapped.
You can use dd to read in from the tape and use the conv=swap
parameter with it to reverse them... you can either write that
back out to another tape or a disk file and feed it to dump
that way. You *may* even be able to pipe it into dump,
but I won't swear to that without looking it up first!
Hope that helps.
-king
ARPA: ables@mcc.com
UUCP: {gatech,ihnp4,nbires,seismo,ucb-vax}!ut-sally!im4u!milano!mcc-pp!ables
----
I tried relaxing ... but I feel more comfortable tense.
------------------------------
Date: Thu, 18 Jun 87 19:32:01 PDT
From: leres@lbl-rtsg.arpa (Craig Leres)
Subject: Re: Undocumented le0: Messages?
The Ethernet driver gives the LANCE chip a block of memory large enough
to hold 40 full sized packets. The errors:
le0: Received packet with STP bit in rmd cleared
le0: Received packet with ENP bit in rmd cleared
are indications that the LANCE received packets that were bigger than
the driver was expecting. The error:
le0: Receive buffer error - BUFF bit set in rmd
indicates that the LANCE chip ran out of memory to put incoming packets
in. The Intel interface will spew the message:
ie0: giant packet
when it receives packets that are too big.
Digital LanBridge-100's are one source of large packets. When the
bridge buffers incoming packets and then transmits them back-to-back,
it doesn't leave quite enough space between them. Some combinations of
transceivers and interfaces see these too-closely-spaced back-to-back
packets as a single large packet.
Another source large packets are old DEQNA Ethernet interfaces. When
they receive more packets than they can handle, they transmit all one's
for a short spell. This garbage looks like an impossibly large
broadcast packet.
Perhaps someone can give an example of non-DEC hardware causing large packets.
Craig
------------------------------
Date: Thu, 18 Jun 87 19:01:25 PDT
From: leres@lbl-rtsg.arpa (Craig Leres)
Subject: Re: query on dump to sun scsi tape?
The QIC-11 format makes 4 passes over the tape. Giving up 25 feet of
tape per pass for slop gives us:
(450 - 25) * 4 = 1700 feet
The scsi cartridge tape drive is 1000bpi. If you multiply this out, you
get 20.4 Mbytes per tape. (Note that these are the numbers you get with
the 'c' dump flag.)
The QIC-24 format makes 12 passes over the tape. This gives us:
(450 - 25) * 12 = 5100 feet
So this format holds 61.2 Mbytes per tape. When I do dumps, I specify
something similar to:
dump 0uncbsf 126 5100 /dev/rst8 /mnt
When using tapes that are not 450 feet long, use the above calculations
to figure out the tape length.
Craig
------------------------------
Date: Tue, 16 Jun 87 08:38:16 EDT
From: dms@hermes.ai.mit.edu (David M. Siegel)
Subject: 3rd party disk drives
Another possibility would be to get the 1.1 Gigabyte model of the
CDC-9715 drive. It's a half width (8 1/2 inches wide, so 2 fit side by
side in an equipment rack) drive, and costs less than 10K. It will run
on a Xylogics 451 controller, since it has a 2.4 megabyte per second
xfer rate.
We've used 850 Megabyte CDC-9772 drives on Interphase 4200
controllers, which is also a much more cost effective configuration
than that supplied by Sun.
-Dave
------------------------------
Date: 16 Jun 87 18:38:48 GMT
From: harvard!munsell!jwf@seismo.css.gov (Jim Franklin)
Subject: bug in SUN 3.x lint
There is a bug in SUN 3.0 and 3.2 lint related to use of static variables
declared in include files. For example, the simple C files shown below
will generate the following bogus lint messages:
$ lint foo.c
foo.c:
foo defined( ./foo.h(1) ), but never used
foo used( foo.c(6) ), but not defined
where foo.h is just
static int foo; /* private variable used by foo.c */
and where foo.c is
#include "foo.h"
main()
{
foo = 42;
(void) printf ("%d\n", foo);
}
Making the variable foo external (i.e., int foo) makes lint happy, but
this violates the principal of information hiding and shouldn't be
necessary.
This bug may also appear in 4.[23] BSD, but I can't test that here.
-----
{harvard!adelie,{decvax,allegra,talcott}!encore}!munsell!jwf
Jim Franklin, Eikonix Corp., 23 Crosby Drive, Bedford, MA 01730
Phone: (617) 663-2115 x415
------------------------------
Date: Thu, 18 Jun 87 10:57:22 EDT
From ted@braggvax.arpa (Ted Nolan)
Subject: CTS/RTS
We also recently needed CTS/RTS; Sun tech support gave me this, with the
warning that it is not something they support. It came off the net
originally, so I don't see any problem passing it on. I don't however know
if it really works - the device we were going to interface with went away
before we got the patch installed.
Ted Nolan
ted@braggvax.arpa
------- Forwarded Message
Received: from snail.sun.com by Sun.COM (4.0/SMI-3.2)
id AA07562; Tue, 26 May 87 15:34:08 PDT
Received: from mycreche.sun.com by snail.sun.com (3.2/SMI-3.2)
id AA21848; Tue, 26 May 87 15:33:35 PDT
Received: by mycreche.sun.com (3.2/SMI-3.2)
id AA02271; Tue, 26 May 87 15:36:56 PDT
Date: Tue, 26 May 87 15:36:56 PDT
From: jgath@Sun.COM (Jim Gath)
Message-Id: <8705262236.AA02271@mycreche.sun.com>
To: ted@braggvax.arpa
Subject: cts/rts flow control
.... here disassembled the /sys/OBJ/zs_async.o routine, and
re-enabled it. See included mail below. I'm impressed. BTW, it's the same
change for 68020 machines as well. I haven't had a chance to try this yet,
but I find it difficult to believe, that it would only work on a Sun1.
First some info that may be relevant:
Release: 3.2 Object only
CPU: 68010
*****************************************************
The zs (Zilog 8530 serial io chip) driver appears not to enable hardware (or
CTS) flow control.
Details (short version)
=======
The file /sys/OBJ/zs_async.o was patched as follows:
%cd /sys/OBJ
%adb -w zs_async.o << EOI
zsparam+0x2c?w 0x7c21
\$q
EOI
%
Now you can remake vmunix.
Details (long version)
=======
This forces the driver to set the 8530 AUTO_CD_CTS flag in write register 3
when the tty is opened or ioctled. To change it back to the original do the
above again but substitute 0x7c01 for 0x7c21.
This used to be done by the release 1.1 driver.
This is what to look for if you don't trust the method used above (comments
are indented - both input and output are shown):
%cd /sys/OBJ
%adb -w zs_asych.o
Now you should look for a group of instructions
which assign immediate values to registers
d7, d6, d5 and d4 - these values will be OR'ed
later on with bits depending on the tty flags set.
Zsparam is called whenever the device is opened
or ioctl'ed.
zsparam,15?ai
_zsparam:
_zsparam: _zsparam: linkw a6,#-0x2c
_zsparam+4: moveml d2/d3/d4/d5/d6/d7/a4/a5,sp@
_zsparam+8: movl a6@(8),a5
_zsparam+0xc: movl a5@(0x38),a4
_zsparam+0x10: tstb a5@(0x4b)
_zsparam+0x14: bnes _zsparam+0x2a
_zsparam+0x16: clrl sp@-
_zsparam+0x18: clrl sp@-
_zsparam+0x1a: pea a5@
_zsparam+0x1c: jsr _zsmctl:l
_zsparam+0x22: lea sp@(0xc),a7
_zsparam+0x26: bra _zsparam+0x330
_zsparam+0x2a: moveq #0x13,d7
_zsparam+0x2c: moveq #0x01,d6
_zsparam+0x2e: moveq #0x40,d5
_zsparam+0x30: moveq #0,d4
_zsparam+0x32: movb a4@(0x23),d4
_zsparam+0x36: andl #0x82,d4
_zsparam+0x3c: orb #8,d4
_zsparam+0x40: cmpb #4,a5@(0x4b)
_zsparam+0x46: bnes _zsparam+0x5a
These occur starting at zsparam+0x2a. The register of
interest is d6 which contains the value for ZSWR3 (see
/sys/sundev/zsreg.h) and the value 0x01 is ZSWR3_RX_ENABLE.
In release 1.1 this value was 0x21 or
(ZSWR3_AUTO_CD_CTS | ZSWR3_RX_ENABLE).
So to change this:
zsparam+0x2c?w 0x7c21
$q
%
------- End of Forwarded Message
------------------------------
Date: Thu, 18 Jun 87 11:50:09 PDT
From: ultra!shj@ames.arpa (Steve Jay)
Subject: IPC does not work on 3/280
We have just attempted to install an IPC card in our 3/280. It does not
work. It causes Bus Trap crashes on our 3/280. The crashes may
be related only to NFS usage from the IPC, and NFS is not required to use
the IPC (although without NFS there isn't much point in putting a PC in
the Sun). The IPC also looses up to half the characters that should appear
in the pctool window (the simulated console screen), making it essentially
unusable even if it doesn't crash.
Sun is well aware of the problems, and is working to resolve them, but
for now we cannot use the IPC on our 3/280. Unreliable rumor has says
that there are some IPC's on some 3/280's which do work, but for now,
I would strongly recommend that 3/280 owners stay far away from the IPC.
Our configuration is a 3/280 server with a bunch of 3/50 diskless
clients. We use the IPC only from the clients, so we have the
"multi-access" version of the software. I don't know if the problems
happen on the single user version (which requires that all IPC access
be done from a Sun console connected directly to the 3/280).
If you want additional information, contact me directly.
Steve Jay
Ultra Corporation
2140 Bering Drive
San Jose, CA 95131
(408) 922-0100
ultra!shj@ames.arpa
------------------------------
Date: Thu, 18 Jun 87 13:25:45 PDT
From: van@lbl-csam.arpa (Van Jacobson)
Subject: 4.3bsd tcp code for Sun available
A port of the 4.3bsd tcp code to Sun OS v3.3 is available for anonymous ftp
from host lbl-rtsg.arpa as compressed tar file 4.3tcp4sun.tar.Z (~ 50KB).
This code reflects what Berkeley is presently running (as of June 6th) and
includes all the bug fixes and performance enhancements made since the 4.3bsd
release.
In limitted testing we have found that the 4.3 code improves local net
performance by about a factor of two over the code that came with Sun v3.3.
E.g., throughput on an ftp between two Sun 3/50s went from 100KByte/sec to
190KByte/sec.
Most of the post-distribution changes in 4.3 have had to do with performance
over congested links (like the Arpanet). We expected, and observed, a large
performance improvement over the Sun (nee 4.2bsd) code. Unfortunately we
couldn't quantify the change: In six file transfer trials over the Arpanet
at midday, the 4.3 code delivered an average of 1.2KB/sec. The original Sun
code timed out without completing a transfer on all six trials. Thus, we are
postulating an infinite improvement.
The tarchive contains objects for a 68020. Since the tcp code is Berkeley's
and not in any way derived from Sun's code, all the source is included. Note
that this is only the 4.3 tcp and doesn't include the 4.3 fixes made to ip,
icmp or udp (this 4.3 tcp does not depend on any of these
fixes and will run in a Sun v3.3 system as-is).
- Van
------------------------------
Date: Fri, 19 Jun 87 15:56:12 PDT
From: guyton%condor@rand-unix.arpa (Jim Guyton)
Subject: xy450/xy451 failure modes
A word of warning for xy450/xy451 users.
We've been running a lot of Sun-2's and 3's with Xylogic 450/451's and
Eagles/Super-Eagles.
Over the years we've had at least three different controller boards fail on
us in such a way that the failure went UNDETECTED by the device driver.
And the data to/from the disk was sometimes corrupted.
No error messages on the console, no bad return codes from read(2) or
write(2). Nothing to indicate the data wasn't being written/read correctly
other than the fact that Unix slowly (or not so slowly) trashes your
filesystems and ND clients go down with VGETU panics. After a while the
server crashes due to freeing free inode or something similar. Running
fsck in such situations seems like playing russian roulette: will it
fix it or just make things worse?
We've nailed the problem by creating 10MB files, cp'ing them around and then
cmp'ing them to see if they're still the same file. Also by changing fsck
to re-read each block that it writes and comparing it to what it should
have been.
Each time replacing the controller fixed the problem. The failures have
been on three different file-servers, in two different machine rooms.
Both 450's and 451's.
Query: I've not heard of anyone else having this problem with their controller
boards. Are we alone?
We have had enough of these failures that we're thinking of writing some sort
of nightly "controller checkout" script.
-- Jim Guyton
guyton@rand-unix.arpa
------------------------------
Date: Thu, 11 Jun 87 01:04:20 GMT
From: mcvax!hatfield.ac.uk!guest@seismo.css.gov
Subject: Why can't we have a Spectrum area?
It's about time a Sinclair Spectrum area appeared on this system!!
What say you everyone?
------------------------------
Date: Thu, 18 Jun 87 11:24:38 EDT
From: ndd@cs.duke.edu (Ned D. Danieley)
Subject: questions about new 3/280?
We just got a new 3/280, and I have some questions:
1) We bought the high resolution monitor, and the screen fonts that I
was using (/usr/lib/fonts/fixedwidthfonts/screen.r.*) are now too
small for comfortable viewing. The 16 point courier font isn't too
bad, but what I'd really like is screen in 16 and 18. Has anyone
done this?
2) We'd like to move the monitor about 40 feet from the machine. We've
had pretty good luck hooking together two sun-supplied cables, but
that's only about 20 feet. Nothing else that we have seems to work
at the required distance (we've tried unshielded twisted pair cable,
but not shielded). Does anyone know how far the hi-res can be
extended, and what sort of cable would be best?
3) From the hardware documentation, I appear to have a Xylogics 472
tape controller and a Fujitsu GCR tape drive. From the documentation
in section 4 of the manual, it appears that I would access the
drive using xt8, say. However, there are no xt devices. That's
okay: /dev/MAKEDEV seems to make mt devices for either tm or xt
controllers, but the highest minor number is 12. How do I get to
the 6250 density? Can I just make mt16 and mt20, or does Sun use
some other method?
Ned Danieley
Basic Arrhythmia Laboratory
Duke University Medical Center
ndd@duke
------------------------------
Date: Thu, 18 Jun 87 11:37:42 EDT
From: nesheim@think.com (Bill Nesheim)
Subject: SCSI driver for a OSI LaserDrive 1200 optical disk?
Does anyone know of a sun 3.2 or better version of a driver for
the Optical Storage International LaserDrive 1200 optical disk?
Thanks,
-- Bill Nesheim; Thinking Machines Corporation, Cambridge, MA +1 617-876-1111
nesheim@think.com, {seismo,ihnp4}!think!nesheim, bill@crnlcs.bitnet
------------------------------
Date: Thu, 18 Jun 87 17:28:23 EDT
From: hull%buffalo.csnet@relay.cs.net (Jon Hull)
Subject: Help with Xylogics 472 tape controller and Multi-Vme board?
If you have this configuration and can spare the time to look or have them
handy, please send me the switch settings for these boards. We have
installed them here and set the switches according to the manuals.
The boards work just fine for small files but have occasional problems
writing large tars and dumps. Sometimes an interrupt is lost and the tar
or dump fails in the middle. Therefore, I would like the switch settings
to verify that our set-up is correct.
Thanks very much in advance,
Jon Hull
hull@cs.buffalo.edu
716-636-3195
P.S. This is the tape controller and adapter Sun sells
for their 1600/6250 bpi tape system.
I called the local Sun people but they
wouldn't even talk to me because I didn't purchase
the controller and drive from them.
------------------------------
Date: 18 Jun 1987 1915-EDT
From: ajw%beach.cis.ufl.edu@relay.cs.net (Andy Wilcox)
Subject: Sun OS 3.2 timed?
About a month ago I saw an advertisement of diffs for timed from 4.3
for 3.2. Alas, I lost the address. Could someone send me these?
Please? Pretty Please? Our entire campus thanks you,
Andy Wilcox
ajw@beach.cis.ufl.edu
------------------------------
Date: Fri, 19 Jun 87 21:01:36 PDT
From: tamir@cs.ucla.edu (Yuval Tamir)
Subject: NFS strangeness on Suns?
You have two Suns: sun1 and sun2. Your home directory is on sun1.
sun1 is exporting the file system containing your home directory on sun2.
Login to sun1 and run
% echo this is a test > foo
% chmod 600 foo
Logout of sun1.
Login to sun2. Don't touch file foo.
su to root.
% su
Run
# cat foo
You will get
cat: read error: Permission denied
since with an access from a remote root sun1 maps
the uid from 0 to -2 (as documented).
Now, exit from the root shell
# exit
Run (you are still on sun2)
% cat foo
You will get
cat: read error: Permission denied
So, you are not allowed to read your own file.
Apparently there is some caching of "credentials" going on.
Obviously it does not work very well . . .
Yuval Tamir
Internet: tamir@cs.ucla.edu
UUCP: ...!{ihnp4,ucbvax,sdcrdcf,trwspp,randvax,ism780}!ucla-cs!tamir
------------------------------
Date: Mon, 22 Jun 87 12:33:24
From: NCBITI%NUSVM.BITNET@wiscvm.wisc.edu (Vijay Shah)
Subject: Mail from Sun to Explorer?
Has anyone out there managed to get the sendmail config. fixed to allow a
Sun to send mail to a TI Explorer? The distributed version of sendmail.cf
caused the Sun to timeout after HELO. I put in the extra flags suggested
by Joe Wieclawek ("E=rn" to the Mether record) - this enabled the Sun to
talk to the Explorer but the mail exchange failed with the message
"some argument to - was of the wrong type". Any help you can offer will be
much appreciated - I am on my own here AND at the end of my rope. HELP!!!!
(NB. the Explorer has no problems getting mail to the Sun)
------------------------------
Date: 22 Jun 87 14:28:43 GMT
From: apctrc!lbrown@seismo.css.gov (Lawrence Brown)
Subject: Sun IPC board config. problems?
I am having trouble with our SUN IPC board. I am trying to get the IBM dos
side of it to communicate with the serial port. Is there anyone out there
who has experience with this system and would be willing to share some tips?
I am trying to configure the /dev/ttyb port to act as the COM2 port on the
MS-DOS side. I followed all instructions about how to configure the port but
the MODE (DOS) command keeps returning a "port not installed" message.
Please call or e-mail, and I'll summarize to the world the results.
Thanks in advance,
Lawrence Brown
USENET: ...!seismo!okstate!apctrc!cdf!zlhb0a or zlhb0a@cdf.apctrc.uucp (?)
Fidonet: 170/219 [ The Computercenter, 24 hrs, 918-245-3456 ]
Phone: (918-660-4389) 24 hrs, voice. USmail: 7325 E. 50th, Tulsa, OK 74145
Disclaimer: Opinions expressed are mine alone and not Amoco's. Thanks guys.
--
USENET: ...!seismo!okstate!apctrc!cdf!zlhb0a or zlhb0a@cdf.apctrc.uucp (?)
Fidonet: 170/219 [ The Computercenter, 24 hrs, 918-245-3456 ]
Phone: (918-660-4389) 24 hrs, voice. USmail: 7325 E. 50th, Tulsa, OK 74145
Disclaimer: Opinions expressed are mine alone and not Amoco's. Thanks guys.
------------------------------
Date: 17 Jun 87 13:37:11 GMT
From: dciem!nrcaer!grpthry!gsi2!alan%utzoo%math.waterloo.edu@relay.cs.net
Subject: Undumping on a Sun III work Station?
Does anyone have a WORKING undump for a SUN III work station. This program
allows a user to run a program to a point, kill it (making it dump core)
and then reload it so that the program starts up with its data pre-loaded.
This is used by GNU emacs and TeX.
Please help.... alan 8-)
------------------------------
Date: Mon, 22 Jun 87 17:08:24 CDT
From: jumper@invader.rice.edu (Greg Jumper)
Subject: SunCore Questions?
I would greatly appreciate anyone's help with the following two
questions/problems regarding SunCore:
1) Has anyone successfully used the valuator input function,
"await_any_button_get_valuator"? According to the documentation,
"Movement of the mouse left or right lowers or raises the value of the
valuator," and, for echo type one, "The current value of the valuator is
displayed on the screen starting at the echo reference point." However,
according to my experience, a value is displayed only when the timeout
period for the input function expires, and the user has no way of knowing
what that value will be if he has moved the mouse before the valuator
input function returns.
Is this behaviour a bug, am I misinterpreting the documentation, or am I
incorrectly using valuator input?
2) How can one (most straightforwardly) employ window routines while running
SunCore?
More specifically, since SunCore does not provide choice input (ala GKS),
I would like to use Suntool Menu Items in its place; furthermore, I would
(naturally) like the menus to appear over the window in which SunCore is
running. All of the examples I have seen in Sun's documentation using
either "menu_show" or "menu_show_using_fd" make use of a programmer-
created window; therefore, the application program "knows" the appropriate
window handle or file descriptor since the program has created the handle
or descriptor. However, my (much simpler) application needs to use a
pre-existing window. How can I obtain an appropriate handle to pass to
"menu_show" or file descriptor to pass to "menu_show_using_fd"?
Actually, the same question applies to any application running in a
Suntools window which would like to use the window from which the
application was invoked as the base frame for the application -- how does
one refer to the invoking window from inside the application? Perhaps
using the Suntools environment variables?
What if one would like to use window routines while running SunCore on a
"raw" device? Since SunCore is implemented on top of the window system,
and, in fact, creates a blanket window over the entire screen, it seems
like there should be a way to use the SunCore window itself. I don't
think the environment variables get set in this case.
I hope someone has some experience with these matters! Please send replies
directly to "jumper@rice.edu."
Thanks,
Greg Jumper
Department of Mathematics
Rice University
End of SUN-Spots Digest
***********************