Sun-Spots Digest, v6n13
William LeFebvre
Sun-Spots-Request at RICE.EDU
Wed Feb 10 13:46:04 AEST 1988
SUN-SPOTS DIGEST Tuesday, 9 February 1988 Volume 6 : Issue 13
Today's Topics:
NFS mailing list formation notice
Re: MacPaint to Sun Rasterfile converter
Re: A weird NFS bug
Re: Problems with in.routed and the yp
Converting block numbers to file names
A solution to the pty problem
Floating-Point Programmer's Guide for the Sun Workstation
RPCSRC - I can't get them...
Software for Anonymous FTP
silo overflow error message?
Why doesn't SunOS Unix use hardware FPA when its available?
Cpu board repair?
How should line printer daemon output filters work?
New Fujitsu drive?
ADA compiler on the SUN?
Escher icon
Send contributions to: sun-spots at rice.edu
Send subscription add/delete requests to: sun-spots-request at rice.edu
Bitnet readers can subscribe directly with the CMS command:
TELL LISTSERV AT RICE SUBSCRIBE SUNSPOTS My Full Name
Recent backissues are stored on "titan.rice.edu". For volume X, issue Y,
"get sun-spots/vXnY". They are also accessible through the archive
server: mail the word "help" to "archive-server at rice.edu".
----------------------------------------------------------------------
Date: Fri, 29 Jan 88 02:58:51 CST
From: sob%bcm.tmc.edu at tmc.edu (Stan Barber)
Subject: NFS mailing list formation notice
As per a discussion at the TCP/IP conference in December, I have agreed to
host a discussion of NFS. While this discussion is mostly oriented toward
PC-NFS and MAC-NFS and so on, "large" computer NFS discussions are also
welcome.
If you would like to join the list, send mail to nfs-request at tmc.edu. If
you would like to send an article to the list, send mail to nfs at tmc.edu.
Stan Barber, NFS list coordinator
Baylor College of Medicine
------------------------------
Date: Wed, 27 Jan 88 11:33:37 CST
From: vixen!ronbo at im4u.utexas.edu (Ron Hitchens, Guru Emeritus)
Subject: Re: MacPaint to Sun Rasterfile converter
In Volume 6/Issue 4, strout at lll-lcc.llnl.gov (Robert E Strout) posted
a MacPaint to Sun raster filter, which included this comment:
> First, I must correct myself... I previously said that a fellow at Rice
> had developed this program. Getting back to the original, I find that
> actually someone (unknown) at Univ. of Texas did. Many thanks to
> him/her!!!
Well, I'm that unknown person, and you're quite welcome. When I saw that
note I took a peek at the file in the archives on titan.rice.edu and it is
indeed my old MacPaint filter. I know it's mine because my name is right
there in the source file.
That filter, paintraster, is rather ancient and has been kicking around
since the Mac was young, several years ago. It evovled (devolved?) from a
filter to convert MacPaint pictures to Impress which originated with a guy
named Dave Johnson at Brown University, who is better known as the author
of macput/macget. I sent paintraster to a guy at Sun about a year ago,
and it seems to have trickled it's way around the country since then.
I have other, similar hacks laying around, which sprang from several years
as Unix Staffer/Sun Guru as the UT CS department. Now that I'm semi-self
employed, I may try to scrub them up and post them. For instance, I have
a tool which displays a MacPaint on the Sun screen, and allows you to clip
an icon out of it. But I haven't gotten around to converting it to
Sunview style yet (it was written back in the 1.x days). These days I'm
into Amigas and am working on an Amiga picture display tool for my 3/60C.
I'll try one day soon to recount my adventures building my own shoebox
disk for my home Sun 3/60, but that's a whole nuther story.
Ron Hitchens ronbo at vixen.uucp (my sexy new Sun 3/60)
hitchens at sally.utexas.edu (a boring, but well-known, vax)
------------------------------
Date: Fri, 29 Jan 88 17:41:09 EST
From: Dan Trinkle <trinkle at purdue.edu>
Subject: Re: A weird NFS bug
There is definitely a "bug" in that permission is cached for a while.
After the buffer that has this information cached gets reused, everything
is back to Normal. This was documented with the system (4.3 BSD with NFS)
we got from Mt. Xinu. I don't know if it was every documented by Sun.
Here is the passage from the Mt. Xinu document:
Root access via NFS has always been restricted by setting root's uid
to "nobody" on the server. A bug has been introduced with version 3.2
that causes the result of permission checking to be remembered
regardless of the access. For example, if root gets denied permission
to a remote file because it is not world accessible, all further
access (even by the owner of the file) will get permission denied.
This is eventually corrected when the particular file leaves local
caches. We haven't had time to track this problem, and felt it wasn't
serious enough to hold up this release.
Please note that this is from Mt. Xinu and not Sun.
> By the way, what is the true logic of root privileges and NFS? I thought
> root is always dropped to -2 (nobody), but if chmod my .rhosts file to 0,
> root can still read it when I log in. Anybody care to enlighten? The
> NFS on that host is not locally hacked.
Is it possible that it is another permission caching problem? I know we
have had problems here with people that "protect' their .rhost files and
things suddenly stop working.
Daniel Trinkle trinkle at cs.purdue.edu ARPA
Computer Science Department trinkle%purdue.edu at relay.cs.net CSNET
Purdue University {ucbvax,decvax,ihnp4}!purdue!trinkle UUCP
West Lafayette, IN 47907 (317) 494-7844 PHONE
------------------------------
Date: Fri, 29 Jan 88 17:45:35 EST
From: Dan Trinkle <trinkle at purdue.edu>
Subject: Re: Problems with in.routed and the yp
Are you using subnets? The ypbind (and a few other binaries) distributed
with 3.4 were not the correct version. We later got (through a back door)
correct replacements. One of these was ypbind. If you subnet a machine,
start routed, then try to start ypbind, it will not work. To get around
this you can kill routed, flush the routes, start ypbind, then start
routed. The version of ypbind that we got with 3.4 does not understand
subnetting.
Dan
------------------------------
Date: 29 Jan 88 21:13:12 GMT
From: elroy!stevo%jane.Jpl.Nasa.Gov at ames.arc.nasa.gov (Steve Groom)
Subject: Converting block numbers to file names
> Is there a utility of the icheck/ncheck/fsck/... family which accepts a
> list of defective disk block numbers and spits out the names of the files
> of which these blocks are part of?
I just did this a few days ago. /usr/etc/icheck has a -b option to
convert block numbers to i-numbers (if the block is currently allocated).
Then, use /usr/etc/ncheck with the -i option to convert i-numbers to file
names.
You still have to use diag to fix the bad block. All that icheck/ncheck do
is convert disk block numbers to file names.
-steve
/* Steve Groom, MS 168-522, Jet Propulsion Laboratory, Pasadena, CA 91109
* Internet: stevo at elroy.jpl.nasa.gov UUCP: {ames,cit-vax}!elroy!stevo
* Disclaimer: (thick German accent) "I know noothingg! Noothingg!"
*/
------------------------------
Date: 27 Jan 88 16:58:53 GMT
From: Joe Pruett <joey%tessi at tektronix.tek.com>
Subject: A solution to the pty problem
Here is a program that will clear out slots in /etc/utmp caused by exiting
suntools programs semi-abnormally. It's dangerous in that people can make
themselves disappear, but it's useful...
Usage: cleartty ttypX ...
--------
#include <stdio.h>
#include <utmp.h>
main(argc, argv)
int argc;
char **argv;
{
struct utmp ut;
int fd = open("/etc/utmp", 2);
int i;
if (fd < 0) {
fprintf(stderr, "Couldn't open /etc/utmp\n");
exit(1);
}
while (read(fd, &ut, sizeof(ut)) > 0) {
for (i = 1; i < argc; i++) {
if (strcmp(ut.ut_line, argv[i]) == 0) {
for (i = 0; i < sizeof(ut.ut_name); i++) {
ut.ut_name[i] = '\0';
}
for (i = 0; i < sizeof(ut.ut_host); i++) {
ut.ut_host[i] = '\0';
}
ut.ut_time = 0;
if (lseek(fd, -sizeof(ut), 1) == -1) {
fprintf(stderr, "Couldn't reposition in /etc/utmp\n");
exit(1);
}
if (write(fd, &ut, sizeof(ut)) == -1) {
fprintf(stderr, "Couldn't write in /etc/utmp\n");
exit(1);
}
break;
}
}
}
exit(0);
}
[[ For this program to work, either /etc/utmp needs to be world writable
or the program needs to be setuid to root. --wnl ]]
------------------------------
Date: Wed, 27 Jan 88 17:40:54 PST
From: weiser.pa at xerox.com
Subject: Floating-Point Programmer's Guide for the Sun Workstation
Recently dgh at sun.com described the Floating-Point Programmer's Guide as
"everything you need to know about numerical scientific and engineering
computation on Suns that you can't find anywhere else."
But it is much more than that. It is also the ONLY documentation for some
of the nifty compile options which have nothing to do with numerical
computation. Did you know that the Sun compilers know how to do INLINE
code? And that Suns are shipped with inlinable versions of several
libraries? You didn't if you didn't read the Floating-Point Programmer's
Guide.
My suggestion for future revisions of this manual is: move this
non-numerical stuff where it belongs, to the CC documentation, or
equivalent.
-mark
------------------------------
Date: Fri, 8 Jan 88 20:11:41 +0100
From: Danny Backx <mcvax!kulcs!dannyb at uunet.uu.net>
Subject: RPCSRC - I can't get them...
I have ordered all files from the new RPC release in the sun-source directory.
All but two arrived intact.
The ones I didn't get are
rpcsrc.doc.11
rpcsrc.doc.13
which have lengths of 63406 and 49991 bytes, respectively.
Since rpcsrc.doc.13 certainly isn't too long to get to me, I retried. I
also retried the other one. (Which IS the longest one I ordered.)
The replies I got after re-ordering, where exactly the same as before.
I really see no way other than contacting you to get these files ordered.
Could you send them by hand, possibly fragmented ? Or could you check
whether the files are indeed correct in the database... ?
I truly appreciate your help. Thanks a lot !!
Danny
[[ The December version of each of the two files rpcsrc.doc.11 and
rcpsrc.doc.13 contained at least one line that had nothing but a dot on
it. This confuses many mailers (and also our archive server). On January
8, new versions of these shar files were installed. If you had problems
on or before that date and have not tried to retrieve new copies since
then, please try again. You should now get the files intact. --wnl ]]
------------------------------
Date: 29 Jan 1988 1432-EST
From: franceus at tycho.arpa (Paul Franceus)
Subject: Software for Anonymous FTP
Does anyone have a list of software (or know where a list can be gotten)
of software for the Sun available for anonymous FTP over the internet. I
am sure that there are some Gems out there. If there is no such list, can
people please let me know what they've got for FTP and I will compile the
list. I think it would be something very useful for the Internet's Sun
community.
Thanks,
Paul Franceus
Natl. Computer Security Center
[[ You can start by retrieving the file "sun-source/00index" from the host
"titan.rice.edu". This is the index for the Sun-Spots source archive,
which contains programs, diffs, shell scripts, etc. submitted to
Sun-Spots in the recent past. --wnl ]]
------------------------------
Date: Wed, 27 Jan 88 10:52:16 PST
From: merat at uw-ftcl.ee.washington.edu (Merat Bagha)
Subject: silo overflow error message?
Pardon my ignorance, but would a kind soul please explain the meaning of
the following "silo overflow" error message, and perhaps some advice on
how it can be avoided?
I'm running Suntools on a 3/50 with a SunView/SunCore application. The
problem occurs about one out of ten times upon exiting the program. The
offending program reported varies anywhere from "csh", "suntools" or
"rpc.rstatd" as below.
rpc.rstatd:
trap address 0x8, pid 965, pc = f0713a2, sr = 2104, stkfmt b, context 4
Bus Error Reg 80<INVALID>
data fault address 7c faultc 0 faultb 0 dfault 1 rw 1 size 0 fcode 5
KERNEL MODE
page map 2000000
D0-D7 2500 1 1 0 0 0 2100 427
A0-A7 0 f1eb878 ffff86dc 0 0 0 f074dcc f074dbc
Begin traceback...fp = f074dcc, sp = f074dbc
Called from f06ee0e, fp=f074dd8, args=f0941d8 f074e00 f06edda f0941d8
Called from f06edda, fp=f074e00, args=f0941d8 0 f0941ac 2100
Called from f0718ae, fp=f074e30, args=f0941d8 1 1 0
Called from f01b2ba, fp=f074e54, args=0 0 0 1
Called from f01b188, fp=f074e78, args=2000 0 2504 effffd8
Called from f004468, fp=ffff870a, args=f04b672 2000 0 e0000001
End traceback...
panic: Bus error
syncing disks... zs2: silo overflow
zs3: silo overflow
5 5 4 2 done
Many thanks.
Merat Bagha, Dept. of Electrical Engr., Univ. of Washington
merat at uw-ftcl.ee.washington.edu
------------------------------
Date: 27 Jan 88 16:25:00 EST
From: enger at bluto.scc.com
Subject: Why doesn't SunOS Unix use hardware FPA when its available?
I was building a new kernel recently and noticed that the C-compiler is
being run with the flag -fsoft. This flag tells the compiler
(collector/Linker?) to emulate floating point instructions in software on
the main cpu, rather than submitting them to the MC68881 hardware floating
point accelerator.
To a non-initiate such as myself, it seems like it would make sense to try
to utilize the resources available on the hardware platform to their
fullest. Since the operating system is one of the most important (and
time critical) programs that is run on the platform, it would seem to make
sense to try to optimize its speed by allowing it to utilize the hardware
FPA when it can. Unfortunately, I do not know enough about Unix to say
whether it would really make any substantial use of the FPA? Or even if
it could!
I consulted with Sun support and they told me that the .O files shipped in
my binary distribution are compiled -fsoft (because not all Sun-3's have
FPA). If I wanted to buy a source license, I could try to compile with
-f68881 flag, but they would not support it.
Can the SunOS Unix operating system benefit significantly from the use of
the MC68881 FPA? Are there really enough of the right type of instrucions
used in the Unix code that it would make sense to try to exploit the FPA
hardware?
I am NOT a recipient of the SUN-SPOTS mailings. I would appreciate it if
any replies were sent directly to me. Thanks in advance!
Bob Enger
Contel Federal Systems
enger at bluto.scc.com
------------------------------
Date: Wed 27 Jan 88 14:41:34-PDT
From: PAWKA at nosc-tecr.arpa
Subject: Cpu board repair?
Can anyone recommend a company in San Diego or L.A. who does good work on
repair of SUN CPU boards? Any help appreciated, please respond by mail as
I'm not yet on this list,
Mike Pawka
ARPA: PAWKA at NOSC-TECR.ARPA
------------------------------
Date: Wed, 27 Jan 88 19:22:30 EST
From: ingr!b11!jim at uunet.uu.net
Subject: How should line printer daemon output filters work?
I'm looking for some information on output filters for the printer
spooler, in particular what needs needs to be in the 'of' filter when
there are other filters specified in the same printcap entry.
Apparently the 'of' filter is treated differently by the system when there
are other filters defined in the printcap entry. I have determined that
the bannerpage data is fed to the 'of' filter as expected but it is
terminated by a ^A rather than EOF. If I let the filter exit when it sees
the ^A the system reports "unexpected exit of daemon, restarting", and it
promptly restarts the 'of' filter. If I let the filter wait for EOF the
process hangs and nothing is printed. There are two cases where the
filters work as documented. The first is when the printcap contains only
a definition for the 'of' filter. The second is when there are 'if' and
other filters specified, of course I loose the banner page then.
Has anyone else had any experience along these lines or know what I
need to put in the 'of' filter?
=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+
Jim Levie REMTECH Inc Huntsville, Al
The opinions expressed above are just that.
Ph. (205) 536-8581 email: uunet!ingr!b11!jim
[[ The most complete description of the lpr system that you'll be able to
find is in the "Unix System Manager's Manual that comes with 4.3. It's a
section entitled "4.3BSD Line Printer Spooler Manual". If you don't have
hardcopy, the troff source can be found online on an 4.3BSD system (in
other words *NOT* a Sun system) under "/usr/doc/smm/06.lpd". Sun might
have published the 4.2 equivalent to this manual somewhere in their stuff
as well. --wnl ]]
------------------------------
Date: Thu, 28 Jan 88 18:03:57 +0200
From: <leonid at TAURUS.BITNET>
Subject: New Fujitsu drive?
Fujitsu has announced a new 8" disk drive called M2388 with 1200 MB of
storage. Has anyone more information about this drive ? I know that it
does 3 MB/s transfer rate which seems to be too much for the Xylogics 451
controller. Xylogics book states that 2.4MB/s is the maximum. The M2388
drive comes with E-SMD interface and with IPI-2 interface. What is IPI-2?
This drive is about twice cheaper then the new disk Sun has announced.
Leonid
------------------------------
Date: 29 Jan 88 13:59:24 GMT
From: rti.uucp!jtm at mcnc.org (Jim Morrison)
Subject: ADA compiler on the SUN?
I would like to install an ADA compiler on a SUN 3/160C (70mb disk with
4mb ram) and would appreciate any advice on the available ADA compilers
for the SUN. I'm currently reading literature on the Alsys and Verdix
compilers. Are there others ? Any recommendations ?
------------------------------
Date: Thu, 28 Jan 88 18:36:19 +0100
From: mcvax!solaris!wyle at uunet.uu.net (Mitchell Wyle)
Subject: Escher icon
Here's one for you M. C. Escher fans: An icon which forms a symetric
background pattern
/* Format_version=1, Width=64, Height=64, Depth=1, Valid_bits_per_item=16
*/
0x8000,0x4000,0x0000,0x0000,0xC000,0x4000,0x0000,0x0001,
0xE000,0x4000,0x0000,0x0401,0xE000,0x2000,0x0000,0x0403,
0xF000,0x2000,0x0000,0x0707,0xF000,0x2000,0x0000,0x01FF,
0xF800,0x100F,0x0000,0x000F,0xF800,0x1030,0x8000,0x001F,
0xFC00,0x08C0,0x4000,0x003F,0xFC00,0x0100,0x2000,0x007F,
0xFE00,0x0200,0x2000,0x00FF,0xFE00,0x0C00,0x1000,0x03FF,
0xFF00,0x1C00,0x1000,0x07FF,0xFF00,0x3C00,0x0800,0x0FFF,
0xFF00,0x7E00,0x0800,0x3FFF,0xFF80,0xFE00,0x060C,0x707F,
0xFF80,0xFE00,0x01F1,0xCF9F,0xFF81,0xFF00,0x0003,0xFFEF,
0xFFC3,0xFF00,0x000F,0xFFEF,0xFFC7,0xFF00,0x001F,0xFFF7,
0xFFCF,0xFF80,0x003F,0xFFF7,0xFFBF,0xFF80,0x00FF,0xFFFB,
0xFF7F,0xFFC0,0x01F8,0x7FFB,0xFCEF,0xFFC0,0x03F7,0xBFFD,
0xF3F7,0xFFE0,0x07EF,0xDFFE,0x0FF7,0xFFE0,0x0FEC,0xDFFF,
0xFFFB,0xFFF0,0x1FEC,0xDFFF,0xFFFB,0xFFF0,0x1FEF,0xDFFF,
0xFFFB,0xFFF8,0x3FF7,0xBFFF,0xFFFD,0xFFF8,0x7FF8,0x7FFF,
0xFFFD,0xFFFC,0x7FFF,0xFFFF,0xFFFD,0xFFFE,0xFFFF,0xFFFF,
0xFFFD,0xFFFE,0xFFBF,0xFFFF,0xFFFD,0xFFFC,0x7FBF,0xFFFF,
0xFFFD,0xFFF8,0x3E3F,0xFFFF,0xFFFB,0xFFF8,0x00FF,0xFFFF,
0xFFFB,0xFFF0,0x1FFF,0xFFFF,0xFFFB,0xFFF0,0x1FFF,0xFFFF,
0x0FF7,0xFFE0,0x0FFF,0xFFFF,0xF3F7,0xFFE0,0x07FF,0xFFFE,
0xFCEF,0xFFC0,0x03FF,0xFFFD,0xFF7F,0xFFC0,0x01FF,0xFFFB,
0xFFBF,0xFF80,0x00FF,0xFFFB,0xFFCF,0xFF80,0x003F,0xFFF7,
0xFFC7,0xFF00,0x001F,0xFFF7,0xFFC3,0xFF00,0x000F,0xFFEF,
0xFF81,0xFF00,0x0003,0xFFEF,0xFF80,0xFE00,0x01F1,0xCF9F,
0xFF80,0xFE00,0x060C,0x707F,0xFF00,0x7E00,0x0800,0x3FFF,
0xFF00,0x3C00,0x0800,0x0FFF,0xFF00,0x1C00,0x1000,0x07FF,
0xFF00,0x0C00,0x1000,0x03FF,0xFE00,0x0200,0x2000,0x00FF,
0xFE00,0x0100,0x2001,0xE07F,0xFE00,0x08C0,0x4002,0x103F,
0xFC00,0x1030,0x8004,0x081F,0xF800,0x100F,0x0004,0xC80F,
0xF800,0x2000,0x0004,0xC807,0xF000,0x2000,0x0004,0x0807,
0xE000,0x2000,0x0002,0x1003,0xE000,0x4000,0x0001,0xE001,
0xC000,0x4000,0x0000,0x0001,0x8000,0x4000,0x0000,0x0000
-Mitchell F. Wyle wyle at ethz.uucp
Institut fuer Informatik wyle%ifi.ethz.ch at relay.cs.net
ETH Zentrum
8092 Zuerich, Switzerland +41 1 256-5237
[[ Also stored in the archives as "sun-icons/mw-escher.icon". --wnl ]]
------------------------------
End of SUN-Spots Digest
***********************
More information about the Comp.sys.sun
mailing list