Sun-Spots-Request@Rice.edu (William LeFebvre) (09/14/88)
SUN-SPOTS DIGEST Sunday, 11 September 1988 Volume 6 : Issue 226
Today's Topics:
Re: dialin/dialout on the same line
Re: TeX and floating point
yellow pages+sunos4.0+sunos3.5: forget it.
Writing to /dev/rst0 can hang s SUN 4/260S
cc dies after using dbxtool
GNU emacs hangs on ALM2
top 2.5 on 386i doesn't work
Wanted: Skeleton SCSI device driver
add a 2nd external ethernet controller?
Maxtor3380 and sun2's or sun3's 3.5?
Disabling the Left key
ID Prom?
Re: Notification on receiving new mail (shell scripts)
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: Fri, 9 Sep 88 15:22:16 BST
From: dmc%cam.sri.com@warbucks.ai.sri.com (David Carter)
Subject: Re: dialin/dialout on the same line
Reference: v6n53, v6n75
This may be ancient history for most of you, but...
Larry Blair wrote some time ago concerning dialin/dialout:
> What we have noticed, on our systems running 3.4, is that the flags for
> zs0 are being read wrong, at least for ttyb. We have tried setting the
> flags to 0x1 and it doesn't work. Just for the hell of it, we tried 0x2.
> That didn't work either. 0x0 works great.
and Brent Chapman said concerning the same problem
> I suspect something got broken in 3.5, but can't prove it... But
> if you're using zs0 ports and 3.5, it most definitely does _not_ work
> the way it is documented to... Have all the other people who've been
> having problems with this lately been running 3.5?
We have just upgraded from 3.2 to 3.5, and our dialin mechanism stopped
working - it answered the call but issued no login prompt. Sun UK could
not help. So as suggested above, I tried setting the flags to 0x0, which
solved the problem. It does therefore appear to be a bug introduced
between 3.2 and 3.5. Interestingly, our dialin port is ttya, not ttyb.
David Carter
SRI International Cambridge Computer Science Research Centre
U.K.
dmc@ai.sri.com
------------------------------
Date: Sat, 10 Sep 88 17:15:19 PDT
From: frame!troy!drf@sun.com (David Fuchs)
Subject: Re: TeX and floating point
Sorry to be disagreeable again, but while John Polstra is correct that TeX
does use fixed-point arithmetic for all calculations that might effect
line-breaking or page-breaking decisions, there are other, non-critical
computations that are carried out in floating-point. So, the speed at
which your CPU carries out floating-point operations will effect the speed
at which TeX runs.
-David Fuchs
[[ I hope this ends the discussion once and for all. --wnl ]]
------------------------------
Date: Fri, 09 Sep 88 14:14:46 -0400
From: Ronald G. Minnich <rminnich@super.org>
Subject: yellow pages+sunos4.0+sunos3.5: forget it.
I had written earlier with a question about a mixed sunos4.0 +sunos3.5
network i use. There were problems with yellow pages, among other things,
and i was curious as to whether (1) anybody else had them, and (2) anybody
else had fixed them.
The responses were interesting. Yes, people had problems, no, no one had
yet resolved them. After looking further and realizing there are some
*serious* design and performance problems with yellow pages (the
services.byname screwup is all too typical); after reading about people
patching binaries (libraries yet); after getting the allegedly working
named patch tape from sun that really does not work; we came to a simple
conclusion:
DITCH YELLOW PAGES
We have a pretty simple home-grown solution that works quite well. I
would have liked to use yp, there are aspects of it that seem well thought
out, but it looks like it was rushed to market and it just ain't there.
The home-grown solution works a whole lot better, and that is really the
final measure. If i could trash the gethostby* stuff in the libraries and
replace them with a pure named-based solution i could ditch yp completely.
As it is we are stuck with having to relay requests to named via ypserver.
ack.
Anybody know anything about the availability of Kerberos/Hesiod? I hate
to throw 150 copies of /etc/{passwd, etc.} everywhere if there is a better
solution. For now, however, homegrown's all right with me. ron
P.S. somebody did a 'df /dev/xy*'. Hangs the system on Sunos4.0. neat.
Also, 'ps ax | fgrep yp' will occasionally get a SEGV.
------------------------------
Date: Fri, 9 Sep 88 16:17:37 EDT
From: bugs@blitz.cs.psu.edu (Dan Ehrlich)
Subject: Writing to /dev/rst0 can hang s SUN 4/260S
Machine Type: Sun 4/260S
O/S Version: SunOS 4.0
Organization: Computer Science Department
The Pennsylvalia State University
333 Whitmore Laboratory
University Park, PA 16802
Description:
When trying to write blocks that are not sized modulo 512 to
/dev/rst0 vmunix continuosly generates error messages to the
console. Below is an extract from /usr/adm/messages. As the
system was not responding to the helm, we forced a crash, thus
the incomplete message and the panic.
vmunix: st0: stwrite: not modulo 512 block size
vmunix: 8b8 count= 63 (1600)
vmunix: cdb= a 1 0 0 3 0
vmunix: si0: siintr: dma in progress
vmunix: last phase= 0x87 (Cmd complete MSG)
vmunix: csr= 0x1107 bcr= 0 tcr= 0x4
vmunix: cbsr= 0x0 (BUS FREE) cdr= 0x0 mr= 0x0 bsr= 0x0
vmunix: target= 4, lun= 0 DMA addr= 0x48b8 count= 63 (1600)
vmunix: cdb= a 1 0 0 3 0
vmunix: si0: siintr: dma in progress
vmunix: last phase= 0x87 (Cmd complete MSG)
vmunix: csr= 0x1107 bcr= 0 tcr= 0x4
vmunix: cbsr= 0x0 (BUS FREE) cdr= 0x0 mr= 0x0 bsr= 0x0
vmunix: target= 4, lun= 0 DMA addr= 0panic: zero
vmunix: syncing file systems... [15] 1 [15] [11] [5] done
vmunix:
vmunix: dumping to vp ff005ca8, offset 0
Repeat-By:
% dd if=/vmunix ibs=80 obs=1600 cbs=1600 of=/dev/rst0
------------------------------
Date: Thu, 8 Sep 88 13:18:06 EDT
From: phri!bc-cis!fami!norm!serge@nyu.edu (Serge Sretschinsky)
Subject: cc dies after using dbxtool
I have been wondering for sometime now why, on any of the 3/50's we have
here, the first invocation of cc after using dbxtool usually dies. For
instance:
[norm] 97) dbxtool a.out
[norm] 98) make
cc -c -g Foo.c
cc: Fatal error in ccom: Hangup
*** Error code 1
Stop.
[norm] 99) make
cc -c -g Foo.c
ar rv Bar.a Foo.o
.
.
.
This is true for both 3.4 and 3.5, the two versions of SunOS I have used.
While this is mostly harmless, I am consumed with curiosity. Anyone out
there know what gives?
Serge Sretschinsky (...bc-cis!fami!serge) The Fantastic Animation Machine, Inc.
------------------------------
Date: Sat, 10 Sep 88 20:05:35 EDT
From: jdh@bu-it.bu.edu (Jason Heirtzler)
Subject: GNU emacs hangs on ALM2
GNU emacs (18.47, 18.50, 18.52 probably others) hangs when I suspend it
(^Z) or resume it (ala "fg"). I only have this when I use an ALM2. Other
programs (vi, etc..) are unaffected.
It seems to be triggered by:
#undef INTERRUPT_INPUT
at the end of config.h (which we need to do). It sounds like a SunOS bug,
but a work-around surely exists.
This is Sun-4/280 SunOS 4-3.2, but I've seen it on a Sun-3/180 with SunOS
3.5 as well.
Anyone got a fix ?
--Jason Heirtzler
Boston University
jdh@bu-it.bu.edu
------------------------------
Date: Fri, 9 Sep 88 13:34:44 MST
From: grandi@noao.edu (Steve Grandi)
Subject: top 2.5 on 386i doesn't work
top 2.5, which runs fine on a 3/160 running SunOS 4.0, does not run on a
Sun 386i running SunOS 4.0. The program bombs immediately with
"top: can't nlist 8 symbols." Pity.
[[ Sounds like NONE of the expected symbols were found. Are you sure that
the environment on the 386i is set up well? Is /vmunix protected somehow?
Is it maybe stripped? Is top installed correctly for 4.0? --wnl ]]
Steve Grandi, National Optical Astronomy Observatories, Tucson AZ, 602-325-9228
UUCP: {arizona,decvax,ncar}!noao!grandi or uunet!noao.edu!grandi
Internet: grandi@noao.edu SPAN/HEPNET: 5355::GRANDI or NOAO::GRANDI
------------------------------
Date: Fri, 9 Sep 88 23:10:06 PDT
From: harker@parns.nsc.com
Subject: Wanted: Skeleton SCSI device driver
I am looking for a skeleton SCSI device driver to use as a model for
writing a device driver for a custom SCSI device I am building. Having
listened to the net and read the sun documentation I am running into a
dead end on how to write a device driver for the Sun SCSI interface. If
you have written one or know of someone who has, I would like to see the
code your wrote or an outline of your code. I need help with how you
divide the device driver between the generic Sun SCSI controller half and
the custom half you write for the specific SCSI target. How do you pass
data between these two halfs? How do you initiate SCSI commands on the
SCSI bus?
Wouldn't it be nice if Sun provided a skeleton device driver and a model
device driver for SCSI devices like they do for MultiBus/VME devices?
I am working on designing a general propose SCSI microcontroller engine to
allow for easy attachment of I/O devices to the SCSI bus. I am doing this
as a hobby, trying to learn about hardware and UNIX device drivers. (if
your going to hack hardware, PCs are for wimps (:-) I am currently
designing it around a Motorola 68HC11 microcontroller with the New Micro
Inc. forth ROM burned in. To this I have attached a NCR 5380 SCSI
controller chip. I am writing a forth vocabulary of SCSI words to handle
reading SCSI commands, and reading and writing SCSI data. This
Microcontroller will have a simple 8 bit expansion bus for interfacing it
to other I/O peripherals. I am currently working on 24 bit digital I/O as
well as 12 bit A/D and D/A conversion.
I am finding forth to be an excellent hardware development environment.
It gives me a nice user interface for developing code and debugging
problems. It also allows for the simple inclusion of assembly code
routines for good performance. The New Micro Inc forth should also allow
me to generate ROMable code for embedded applications (like a SCSI forth
engine). I have tries other microprocessor development environments (CPM
& IBM PC) and found them hard to learn and use. I like the forth
microcontroller because it is well documented and easy to learn. It is
also cheap ($200)
As to leaning about SCSI I have found two very good sources of
information: The first is two articles by Steve Ciarcia in the April/May
1986 issues of Byte on adding a SCSI interface to the SB-180 computer.
The second is a series of articles by Rick Lehrbaum in "The Computer
Journal" (issues 22, 23, 24, and 25, 406-257-9119). These both are very
good introductory articles explaining how the SCSI bus works. For the
hardcore, both of the SCSI specs are very good, SCSI and the proposed SCSI
II spec, as well as the NCR 5380 manual.
Thanks in advance
Robert Harker, All around good guy.
harker@nsc.com, {sun,decwrl,hplabs}!nsc!harker
------------------------------
Date: Fri, 9 Sep 88 16:07:17 EDT
From: dms@wheaties.ai.mit.edu (David M Siegel)
Subject: add a 2nd external ethernet controller?
Hi,
I'd like to add a second ethernet controller board to a 3/180. That is,
I'd like 3 ethernet interfaces on the machine. Could someone suggest some
switch settings for the ethernet controller board for this? Sun doesn't
ship any documentation with the board, so it's very hard to figure out
what to do.
Thanks!
-Dave
------------------------------
Date: Fri, 9 Sep 88 18:37:58 PDT
From: decwrl!pyramid!amiga!dale@ucbvax.berkeley.edu (Dale Luck)
Subject: Maxtor3380 and sun2's or sun3's 3.5?
I have two maxtor3380 380mbyte embedded scsi drives that I am trying to
get working on a sun2/120 and a sun3/50. Both running os version 3.5.
After typing in the cyl/heads/sector information for either adaptec or
emulex, the drive is accessed, the light goes out and the sun hangs after
typing out scsi reset. Do I need a special driver just for the maxtor?
All the jumpers are set properly I believe. Please email to amiga!dale if
you have suggestions.
Dale Luck
amiga!dale
------------------------------
Date: Fri, 9 Sep 88 12:16:39 PDT
From: Francois Felix INGRAND <felix@ai.sri.com>
Subject: Disabling the Left key
On the sun 3, the LEFT key is recognize as the META key by GNUemacs...
Unfortunately, I am used to another keyboard (Lisp Machine) which has the
Control Key exactly at this place. You can imagine the mess under GNU
emacs when you send an esc instead of a control.
So I am looking for an utility to make this key be a CTRL modifier, or
even to disable it completly.
Please reply by E-mail as I do not always read this Newsgroup.
Thanks in advance,
Felix
------------------------------
Date: Sat, 10 Sep 88 14:17:11 CDT
From: drl@vuse.vanderbilt.edu (David R. Linn)
Subject: ID Prom?
Does anyone know whether the 4.0 document set contains info on how to
interpret the number returned by gethostid(2). I seem to recall something
from SUN that told how to interpret these numbers but I can't find that
reference.
If it is not in the document set, can someone in the know give a brief
rundown on the strategy for telling to what kind of machine a hostid
belongs.
David
David Linn - drl@vuse.vanderbilt.edu -or- ...!uunet!vuse!drl
------------------------------
Date: Fri, 9 Sep 88 23:25:52 MET DST
From: Maarten Litmaath <mcvax!cs.vu.nl!maart@uunet.uu.net>
Subject: Re: Notification on receiving new mail (shell scripts)
Reference: v6n221
Dear moderator, how about the following mailbox daemon shell hack 'maild'?
The 'ok' shell script is to let 'maild' stop nagging about unread mail.
The advantage over 'biff' and 'rcvalert' is that 'maild' keeps telling you
that there's mail, so if you were on a coffee break when the mail arrived,
and due to something the notification scrolled off the screen, there's
still no problem. Of course the best is a .maildelivery file with
something like
* - | R /foo/bar/rcvalarm
in it (man 5 maildelivery). 'rcvalarm' (compare with 'rcvalert') is
included too. BTW, there are a few ^G's in the notification messages; I
hope the mailer didn't object to them. [[ The bells have been removed
from the shar file. If you want to put them back in, they appeared in the
lines "Mail in..." and "You have mail." The shar file word-count checks
were adjusted accordingly. --wnl ]]
Maarten Litmaath @ Free U Amsterdam:
maart@cs.vu.nl, mcvax!botter!maart
: This is a shar archive. Extract with sh, not csh.
: This archive ends with exit, so do not worry about trailing junk.
: --------------------------- cut here --------------------------
PATH=/bin:/usr/bin:/usr/ucb
echo Extracting 'maild'
sed 's/^X//' > 'maild' << '+ END-OF-FILE ''maild'
X#! /bin/sh
X
Xcase $# in
X 0)
X set $HOME/mailbox
Xesac
X
Xfor i
Xdo
X if test -f $i
X then
X :
X else
X echo "Cannot find $i."
X exit 1
X fi
Xdone
X
Xmailboxes="$*"
Xuser=`whoami`
X
Xcase $timeout in
X '')
X timeout=180
Xesac
X
X(
Xwhile :
Xdo
X set sentry `who | sed -n "/^$user[ ]/{;p;q;}"`
X
X case $# in
X 1)
X exit 0
X esac
X
X tty=/dev/$3
X
X for i in $mailboxes
X do
X set sentry `ls -l $i`
X
X mod="$6$7$8"
X
X set sentry `ls -lu $i`
X
X if expr "$mod" \> "$6$7$8" > /dev/null
X then
X echo Mail in $i. > $tty
X fi
X done
X
X sleep $timeout
Xdone
X) &
X
Xexit 0
+ END-OF-FILE maild
chmod 'u=rwx,g=rx,o=rx' 'maild'
set `wc -c 'maild'`
count=$1
case $count in
532) :;;
*) echo 'Bad character count in ''maild' >&2
echo 'Count should be 532' >&2
esac
echo Extracting 'ok'
sed 's/^X//' > 'ok' << '+ END-OF-FILE ''ok'
X#! /bin/sh
X
Xcase $# in
X 0)
X set $HOME/mailbox
Xesac
X
Xexec touch $*
+ END-OF-FILE ok
chmod 'u=rwx,g=rx,o=rx' 'ok'
set `wc -c 'ok'`
count=$1
case $count in
67) :;;
*) echo 'Bad character count in ''ok' >&2
echo 'Count should be 67' >&2
esac
echo Extracting 'rcvalarm'
sed 's/^X//' > 'rcvalarm' << '+ END-OF-FILE ''rcvalarm'
X#! /bin/sh
X
Xuser=`whoami`
X
Xwhile :
Xdo
X set sentry `who | sed -n "/^$user[ ]/{;p;q;}"`
X
X case $# in
X 1)
X exit 0
X esac
X
X echo You have mail. Kill $$. > /dev/$3
X
X sleep 180
Xdone
+ END-OF-FILE rcvalarm
chmod 'u=rwx,g=rx,o=rx' 'rcvalarm'
set `wc -c 'rcvalarm'`
count=$1
case $count in
179) :;;
*) echo 'Bad character count in ''rcvalarm' >&2
echo 'Count should be 179' >&2
esac
exit 0
------------------------------
End of SUN-Spots Digest
***********************