Sun-Spots-Request@RICE.EDU (William LeFebvre) (06/22/88)
SUN-SPOTS DIGEST Tuesday, 21 June 1988 Volume 6 : Issue 119
Today's Topics:
Re: Doing the unexpected
Re: Sun 3/50 eyestrain?
Re: CDC Sabre series drives
NFS Protocol Rev BOF at USENIX
Sun 4 cc(1) bug
50's on thick and thin ethernets
anonymous ftp setup
Calctool problem
4.0 routed problems.
Xylogics 440 and Sun 2/120?
386 and a Sun?
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: Mon, 13 Jun 88 14:14:08 EDT
From: budd@bu-it.bu.edu (Phil Budne)
Subject: Re: Doing the unexpected
I posted this to comp.unix.wizards without realizing the original article
was cross posted from sun-spots!
All this is from experience managing a SOS 3.4 cluster, I have no
knowledge of how things may have changed in 4.0.
One example of how a user running /etc/rc* files can cause mayhem is a
user starting a private copy of inetd. The new inetd will try to bind
local sockets and fail for the ordinary tcp and udp based services, but
will re-register RPC based services with the portmapper.
This happened once here, and all rcp.mountd requests were being serviced
by and failing, returning EACCESS (presumably on an call to open with the
path of the mount (to get an fd to hand to getfh(2))
It took me QUITE a while to track this down after searching for and
finding no applicable references to EACCESS in the source for either
client or server nfs.
portmap and nfsd should be immune to this as they bind to reserved
sockets, but ypbind and ypserv may also have this weakness.
In short portmap does not provide rigorous defense of the ports it names.
In fact any user can call pmap_unset;
/* rpcrm.c -- p budne@bu/dsg */
main( c, v )
int c;
char *v[];
{
if( c != 3 ) {
printf("%s prog version\n", v[0] );
exit( 1 );
} /* wrong number of args */
if( pmap_unset( atoi( v[1] ), atoi( v[2] ) ) )
puts( "OK" );
else
puts( "failed" );
} /* main */
------------------------------
Date: 12 Jun 88 22:17:40 GMT
From: ur-valhalla!badri@valhalla.ee.rochester.edu (Badri Lokanathan)
Subject: Re: Sun 3/50 eyestrain?
Reference: v6n105
I work on a Sun 3-110 with a Hitachi color monitor in a lab with
fluoroscent lighting. I started with just black and white windows, but
soon discovered that the large quantity of light delivered by the
background was extremely tiring to the eye, so I switched to a grey
background. I have experimented with various backgrounds and font
colours, but none have been really soothing. The one that I like the most
is a yellow font on a dark grey background.
Another problem is a slight jitter in the screen, which is really annoying
after sometime.
I would like to hear from other users of colour Suns if they have any
solution to this problem.
badri@valhalla.ee.rochester.edu
{ames,cmcl2,columbia,cornell,
garp,harvard,ll-xn,rutgers}!
rochester!ur-valhalla!badri
------------------------------
Date: Mon, 13 Jun 88 14:42:01 EDT
From: dan@wind.bellcore.com (Daniel R. Strick)
Subject: Re: CDC Sabre series drives
We have been using CDC 9720 drives here, but only for a few months. All
of our drives have the SMD interface. We have the 368, 736, 850, and 1230
MB models. We have used these drives with Xylogics 451 and Interphase
4200 disk controllers.
The Sabre series drives have more than 1024 cylinders. They can be
configured to accept the high order address bit during either cylinder
select (aka TAG1, the address bit is passed on the signal line sometimes
called TAG4) or during head select (aka TAG2, the address bit is passed as
a very high order head address bit.) The Xylogics 451s seem to like high
order cylinder addressing via TAG1. The Interphase 4200 seems to like
high order cylinder addressing via TAG2. The drives can also be
configured for either the SMD-O (old, original, ...) or SMD-E (extended)
interface. The difference is basically that the SMD-E interface uses the
TAG4 signal line to trigger some extra drive status signals. As I recall,
the Xylogics 451 likes SMD-O and the Interphase 4200 doesn't care.
In addition to the usual hassle of working out sector switch settings, we
had a few problems with the newer disk control board (part of the drive)
that periodically sweeps the heads over the disks to "lower particle count
in the module". The problem is that the automatic sweep takes the drive
off cylinder and this tends to confuse SMD disk controllers. My Xylogics
451s didn't seem to care but my Interphase 4200s reported seek timeout
errors (apparently waiting for the drive to come back on cylinder). There
are several jumpers on the disk control board that disable or modify the
sweep activity. I installed the "return" jumper that causes the drive to
automatically return the heads to the original cylinder after an automatic
sweep and modified my disk driver to ignore the drive status change
resulting from the drive going off/on cylinder.
N.B: the 1230 MB model is a 24 MHz drive. It should be used only with
controllers rated for 24 MHz drives. The CIPRICO 3200, Emulex VM31, and
Interphase 4200 controllers are supposed to handle 24 MHz drives. The
Xylogics 451 and the current version of the 754 are not, though I
understand the 754 (and 753) will soon be "up to speed". I don't know
about the 7053.
------------------------------
Date: Mon, 13 Jun 88 17:46:36 PDT
From: sxn@sun.com (Stephen X. Nahm)
Subject: NFS Protocol Rev BOF at USENIX
If you are attending USENIX in San Francisco this month, you may want to
attend the NFS Protocol Revision BOF. The BOF will be held on Tuesday,
June 21 at 6 pm.
Rusty Sandberg, one of the initial NFS developers, will be presenting the
NFS protocol version 3. The protocol has many changes to incorporate
suggestions made by the ONC vendor community and to better support
non-Unix systems.
Hope to see you there,
Steve Nahm
Sun Portable ONC/NFS
------------------------------
Date: Mon, 13 Jun 88 14:41:13 EDT
From: clapper@nadc.arpa (Brian M. Clapper)
Subject: Sun 4 cc(1) bug
A colleague of mine discovered what appears to be a cc(1) bug on the Sun 4
(SunOS Release 'Sys4-3.2_REV2'). I've enclosed the program he used to
replicate the problem, along with a sample run. The problem occurs
primarily with functions that accept a floating point argument and return
a structure. We're reporting the bug to Sun, but I thought I'd publicize
it as well. I apologize if this one has already appeared in SunSpots; I
may have missed it.
Brian M. Clapper ARPA: clapper@nadc.ARPA
Code 7031 UUCP: ...!harvard!clapper@nadc.ARPA
Naval Air Development Center
Street and Jacksonville Roads Phone: (215) 441-2118
Warminster, PA 18974 AUTOVON: 441-2118
---------- start of program ----------
#include <stdio.h>
/*complex number template*/
struct complex
{
float real;
float imaginary;
};
struct complex makereal (f)
float f;
{
struct complex temp;
/*debugging print statement*/
fprintf(stderr,"Makereal got real value %f\n",f);
/*return complex number with real part f, imaginary part 0.*/
temp.real=f;
temp.imaginary=0.0;
return temp;
}
main(argc, argv)
int argc;
char *argv[];
{
struct complex value;
float f=10.0;
float atof();
if(argc>1)
f=atof(argv[1]);
value=makereal(f);
}
---------- end of program ----------
The program was compiled normally (i.e., 'cc float_bug.c'). Here's a
sample run:
% a.out
Makereal got real value 524288.000000
% a.out 0.0
Makereal got real value 0.000000
% a.out 1.0
Makereal got real value 0.007813
% a.out -1.0
Makereal got real value -0.007813
% a.out -2.0
Makereal got real value -2.000002
% a.out 2
Makereal got real value 2.000002
% a.out 3
Makereal got real value 32.000030
% a.out 4
Makereal got real value 512.000486
% a.out 5
Makereal got real value 2048.001945
% a.out 6
Makereal got real value 8192.007782
% a.out 7
Makereal got real value 32768.031128
% a.out 8
Makereal got real value 131072.124512
% a.out
Makereal got real value 524288.000000
% a.out 9
Makereal got real value 262144.249023
% a.out 12
Makereal got real value 2097153.992187
------------------------------
Date: Mon, 13 Jun 88 16:21:27 edt
From: mlijews@nswc-wo.arpa
Subject: 50's on thick and thin ethernets
I have 50's on both thick and thin Ethernet and can't tell any difference
in performance.
Mike
------------------------------
Date: 13 Jun 88 04:41:51 GMT
From: rutgers!uhnix1.uh.edu!cosc2mi@gatech.edu (Francis Kam)
Subject: anonymous ftp setup
I was trying to set up an anonymous ftp on a Sun 3/280 running release
3.5. When I tested it out after the ftp entry in /etc/passwd and mkdir
the appropriate directories, everything works fine except that 'ls'
returns nothing. I suspected I didn't use /usr/5bin/ls but putting it in
/bin as the ls does not help. 'ls' only returns something when I put '/'
(root) as the home directory for ftp in /etc/passwd. chroot() should work
ok for my previous default /usr/local/ftp since ftp didn't say 'error
setting guest privilege'.
Please give me some hints if possible.
-- Francis (UHCS)
Internet address: sun1.cs.uh.edu
CSNET Address : mkkam@houston.csnet
[[ Ahhh! A topic I really know something about (having personally set up
the anonymous FTP account on "titan". I'll tell you exactly what's
missing: /etc/passwd and /etc/group. Think about how "ls" maps uid and
gid to names and you will realize why you need them! Here is how our FTP
account is set up. The ftp line in the real passwd file is:
ftp:*:101:20:File Transfer:/mnt/ftp:/bin/true
The "*" means that no login password will work and "/bin/true" is the
useless login shell. "/mnt/ftp" must have at least the directories "bin"
and "etc". I set them up owned by root and not writable by anyone. "bin"
contains "ls" and "true", both mode 111 (I don't remember why I included
"true"). "etc" contains "passwd" and "group", both mode 444. DO NOT
place a copy of your real passwd file there. Instead make one up that
contains jsut a few entries. Ours just has lines for root, nobody,
daemon, bin, ftp, and all the password fields are "*". Similar comments
apply for "group". Remember, you cannot use symbolic links here, because
the contents of the link (the file's new name) gets interpreted in the
chroot-ed environment along with all other file names. Hard links will
work, but most setups won't have the anonymous area on the same partition
as /bin and /etc. As for the directory that will store the publically
accessible files, you might want to make it owned by "ftp" and set to mode
577. This lets any local user manipulate the directory, but does not
allow an anoymous FTP session to create, delete, or rename any of the
files there. We have a separate directory, called "incoming", where
anonymous people are allowed to "put" files. Hope this helps! --wnl ]]
------------------------------
Date: Mon, 13 Jun 88 13:04:13 PDT
From: ultra!ted@ames.arc.nasa.gov (Ted Schroeder)
Subject: Calctool problem
HELP! What'd I do wrong? I got calctool from the archives and (I think)
followed the instructions for installation properly. But instead of a
nicely layed out keyboard I get something that looks like it has one too
few rows (like the main window wasn't opened big enough or something).
It looks like:
+-------------------------------+
| Close | 0|
+-------+-----------------------+
| Inverse Scien( Erase) |
| Sto Rcl Exc ( ) EE |
| X log ln yx 7 8 |
| 9 - sin cos tan 4 |
| 5 6 + /x 1/x x! |
| 1 2 3 - pi Int |
| Fix . 0 +/- = |
+-------------------------------+
This can't be right, can it?
Ted Schroeder ultra!ted@Ames.arc.nasa.GOV
Ultra Network Technologies
2140 Bering drive with a domain server:
San Jose, CA 95131 ted@Ultra.COM
408-922-0100
[[ Did it find the right fonts, that is, the ones that came with it?
--wnl ]]
------------------------------
Date: Mon, 13 Jun 88 15:04:46 PDT
From: Tom Marazita <toad@hub.ucsb.edu>
Subject: 4.0 routed problems.
Here at UCSB we recently installed Sun's latest and greatest Unix version
4.0. on a few of our Sun 3/50s and 3/280s. So far I have only
encountered one problem:
When routed starts up, it receives all of the routing information from our
other gateways; however, after a few minutes all of the routes vanish,
never to be updated again. It seems the route broadcasts from other
systems are received initially, but then ignored by the 4.0 routed
thereafter. This has only happened on the few machines we have upgraded
to 4.0. All of our other machines (Sun, Vaxen, TeK systems, etc) all
function normally.
If anyone else has noticed similar problems with routed, or if anyone can
suggest any areas I might examine for possible configuration problems, I'd
be very interested in hearing from you.
Tom Marazita University of California
Center for Computational Sciences and Engineering.
3111 Engineering
Santa Barbara, CA 93106
INET: toad@hub.ucsb.edu
BITNET: toad@sbitp.bitnet
CSNET: toad%ucsb
UUCP: ...{ucbvax,ucsd,pyramid}!ucsbcsl!toad
VOICE: (805) 961-3221
------------------------------
Date: Mon, 13 Jun 88 14:15:08 ]Io
From: Douglas Humphrey <deh@eneevax.umd.edu>
Subject: Xylogics 440 and Sun 2/120?
This is a question about old hardware. I have a Sun 2/120, and also a
Xylogics 440 (the old 2 board version of the 450 I believe) SMD
controller, and an Ampex SMD drive (160mb). The drive and the controller
worked fine in a Masscomp system.
The question is, will the Sun OS support the Xylogics 440 ? I know that it
supports the 450/451 controllers. Will the 440 use the same software
drivers? If this controller turns out to be supported for early OS
releases, what is the situation with the current/new SunOS release?
Any answers or pointers would be much appriciated. Also, a second
question. Several times I have called Sun offices checking on the status
of the Sun 2 and software; is it frozen yet, is there current support,
will it run the new OS, etc. Each time they tell me that the Sun 2 is not
supported, and then after they go off and talk to people they tell me that
it is fully supported, and there are no plans to stop supporting it soon.
What is the real story folks?
Thanks,
Doug Humphrey
digex@ai.ai.mit.edu
------------------------------
Date: 13 Jun 88 16:09:00 GMT
From: cmcl2!phri!dasys1!jfreund@rutgers.edu (Jim Freund)
Subject: 386 and a Sun?
Has anyone had any experience (good or bad) upgrading a 286 to a 386 and
hooking it up into a Sun? I'd appreciate any help anyone can provide.
- Jim Freund -
Big Electric Cat Public UNIX ..!cmcl2!phri!dasys1!jfreund
Hour of the Wolf -- WBAI (99.5FM) NYC -- Saturdays 5 - 7 AM
------------------------------
End of SUN-Spots Digest
***********************