Sun-Spots-Request@RICE.EDU (William LeFebvre) (02/17/88)
SUN-SPOTS DIGEST Tuesday, 16 February 1988 Volume 6 : Issue 16 Today's Topics: Sun-Symbolic-Math Mailing List Re: Setting up diskless clients without using Setup Re: Adding a node to a server-based system Re: adding more diskless clients Re: C++ on the Suns: which vendor to choose Re: "iebark" errors on Sun 3 Re: EPSON Printer Problems re: dumpregion by Richard Tobin, AIAI, Edinburgh A small bug in rarpd(8) Motorola MC68881/MC68882 manual published by Prentice Hall retrofitting 68881 chips Problems with Century C-2400 SMD drive Using 3rd and 4th serial port on Sun CPU board? Monochrome Pin-outs? 2 monitors off one Sun? lpd printing to a TCP terminal server attached printer? "Virtual" scrollbars? Need info on PostScript Laser Printers PostScript driver for SUN-GKS? 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 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@rice.edu". ---------------------------------------------------------------------- Date: Thu, 4 Feb 88 23:39:26 CST From: steve@ncsa.uiuc.edu (Steve Christensen) Subject: Sun-Symbolic-Math Mailing List The SUN User Group has asked that a number of Special Interest Group mailing lists be formed based on first meetings at the San Jose meeting. Following the "Suns-at-Home" mailing list setup by Dwight McKay in a previous Sun-Spots volume, I will be moderating a mailing list on symbolic manipulation systems on Suns called Sun-Symbolic-Math. I encourage any Sun Users interested in systems like Macsyma, SMP, Maple, Reduce, Sheep and others to join and contribute to this mailing list. I urge vendors to keep us informed of current and future developments of their products on SUN machines. If you wish to join the mailing list, send a message to Sun-Symbolic-Math-Request@spock.ncsa.uiuc.edu or steve@spock.ncsa.uiuc.edu. To submit an article or comment or other piece of information, send it to Sun-Symbolic-Math@spock.ncsa.uiuc.edu Steve Christensen, Moderator of Sun-Symbolic-Math steve@spock.ncsa.uiuc.edu National Center for Supercomputing Applications University of Illinois at Urbana-Chamgaign ------------------------------ Date: Wed, 3 Feb 88 10:48:40 PST From: Steve Schlaifer x4-3171 <steve@mahendo.jpl.nasa.gov> Subject: Re: Setting up diskless clients without using Setup Using setup is the easiest way to set up diskless clients only if it gets it right. When I went to set up a diskless client (my first and so far only one), I started off to use setup because that was what everyone said was best. But, when I fired it up it said I had xy0, xy1, sd0, and sd1 disks. Further, it said my xy0 and xy1 disks were Fujitsu eagles. This was completely wrong. My only disk is on xy0 and it is a Fujitsu super eagle. I could find no way to convince setup that it was wrong and was left with no recourse but to do the job by hand. Creating the disk partitions was the easy part. The hard part was figuring out what goes in the partitions and how to go about providing /tmp, /usr/spool/... etc. directories for the client. Larry Wall, who administers another server here, came up with the scheme we used for the first cut at the problem of what goes in the partitions. We just reproduced the contents of his server's client partitions on my machine. I poked around on another server for a while and saw how to use symbolic links to get private versions of the spool and tmp directories. I am up and running now but the whole process took much longer than it should have. It was annoying that setup didn't work and there was no way to make it work. It was more annoying that the documentation wasn't complete enough to tell me how to do the job by hand. --Steve Schlaifer (steve@mahendo.jpl.nasa.gov or jplgodo!steve) ------------------------------ Date: Wed, 3 Feb 88 17:17:49 EST From: Dick St.Peters <steinmetz!stpeters@uunet.uu.net> Subject: Re: Adding a node to a server-based system Reluctant to give up disk space now for clients you *might* get someday? Well, for once you can have your cake and eat it too. WARNING: Don't even think of trying this unless you really know what you're doing! You can allocate soft partitions for future clients and still use the space in the meantime, because you can NFS-export an ND partition. Twiddle nd.local to combine the future client's root and swap partitions into a single partition, rerun nd, make a filesystem on this ND partition, mount it, and export it. Then fiddle the server's rc.whatevers so that it doesn't try to mount the partition before running nd. (The NFS'd ND partition can't go in the server's fstab). When the planned-for client arrives, you just undo this, including making a root filesystem for the client and filling it with the needed files. As I recall, you can't just untwiddle nd.local and rerun nd: you have to actually reboot the server. Of course, you then face a squeeze: what to do with the files that were in the exported ND partition. Dick St.Peters GE Corporate R&D, Schenectady, NY stpeters@ge-crd.arpa uunet!steinmetz!stpeters ------------------------------ Date: Thu, 4 Feb 88 13:12:50 CST From: symcom!maeder@uxc.cso.uiuc.edu Subject: Re: adding more diskless clients In recent issues of sunspots there was some discussion about adding diskless clients. It seems to me there is a better solution than rerunning setup or repartitioning the whole disk and booting from tape. On a (homogenous) server (named orion in this case) the disk normally looks like this: partition fs a / b (swap) c nd root and swap d /usr/orion f /pub h /usr Partition c is the whole disk, but there is a 'hole' between b and d where the nd partitions are put in. d is normally the big partition with all the user files. So all you have to do is to extend the nd partitions at the expense of the d partition. The only change in the disk label is then to change the starting cylinder and size of the d partition. You should then remake a filesystem on it and edit nd.local. You can then reload the userfiles from the backup. All of this can be done without booting off the tape, you only need to run diag, which can be booted off the disk. (Changing the label does not destroy any partitions that were not changed, does it?) With the Rimfire controller which comes with a program like 'diag' which runs under Unix I would not even have to shut down the system! This is of course a bit like open heart surgery, so you should make a level 0 backup... Roman E. Maeder Dept. of Mathematics UUCP: ...!uiucuxc!symcom!maeder Univ. of Illinois at Urbana-Champaign Internet: maeder@symcom.math.uiuc.edu ------------------------------ Date: Wed, 3 Feb 88 09:32:51 CST From: strong%bell.cad.mcc.com@mcc.com (Michael Strong) Subject: Re: C++ on the Suns: which vendor to choose Michael Tiemann, here at MCC, modified the GNU C compiler and GNU gdb symbolic debugger to yield a C++ compiler and a version of gdb that can be used to debug C++ programs. It is public domain and has been given back to Richard Stallman's Free Software Foundation, whose goods are archived on AI.PREP.EDU at MIT. You might look into getting a copy of that from PREP via anonymous ftp. Michael Strong @ MCC VLSI CAD Program [512] 338-3642 P.O. Box 200195, Austin, TX 78720 3500 West Balcones Center Drive, Austin, Tx 78759 ARPA: strong@mcc.com UUCP: {ihnp4,seismo,harvard,gatech,pyramid}!ut-sally!strong%mcc.com ------------------------------ Date: Thu, 4 Feb 88 02:00:04 EST From: Chris Torek <chris@mimsy.umd.edu> Subject: Re: "iebark" errors on Sun 3 Iebark is one half of the Intel Ethernet watchdog timer (the other half is called iedog; the Intel Ethernet chip has so many bugs it takes two routines, and many pages of code, to avoid them all). The hang appears to be the result of a bug in the driver. We saw a spate of these some time ago; they subsequently vanished as mysteriously as they had appeared (or did someone replace a board? I forget). I stared at the driver for several hours while these were occurring, and twice thought I had found the problem, but was wrong both times. It may be another hardware bug. In-Real-Life: Chris Torek, Univ of MD Comp Sci Dept (+1 301 454 7163) Domain: chris@mimsy.umd.edu Path: uunet!mimsy!chris ------------------------------ Date: Wed, 3 Feb 88 07:54:43 EST From: Mike Jipping <jipping@frodo.cs.hope.edu> Subject: Re: EPSON Printer Problems > I added a "fc#0300" and "xc#40" to the printcap entry to shut off all > possible parity conflicts. > ... > [[ Ummm....Don't you mean "fs#0300", which would set "any parity"? And > "xs#040", which would set LITOUT mode? (The leading 0 tells printcap that > it's an octal value.) ... Out of curiosity, > what would turning off both EVEN and ODD parity mean? ... --wnl ]] Oops. You're right. It was "xs#040", which, upon looking back, invalidates the "fc" setting by ignoring parity. BTW -- Turning off both EVEN and ODD parity gets you space parity. [[ But, in my experience, so does turning ON both EVEN and ODD. How would one then get "mark" parity? --wnl ]] -- Mike Jipping Dept of Computer Science Hope College jipping@cs.hope.edu ------------------------------ From: Eric Ole Barber <mcvax!nw.stl.stc.co.uk!sizex@uunet.uu.net> Date: Thu, 4 Feb 88 09:41:44 GMT Subject: re: dumpregion by Richard Tobin, AIAI, Edinburgh Further to my recent message, Chris Uppal points out that dumpregion.c needs to be edited to enable option (iii). The edits are - [[ The "edits" are archived in the file "sun-source/dumpregion.diff". It can be retrieved via anonymous FTP from the host "titan.rice.edu" or via the archive server with the request "send sun-source dumpregion.diff". For more information about the archive server, send a mail message containing the word "help" to the address "archive-server@rice.edu". --wnl ]] ------------------------------ Date: Sat, 6 Feb 88 14:03:36 +0200 From: leonid@TAURUS.BITNET Subject: A small bug in rarpd(8) I have spent an hour trailing down this problem, and I wish to save you the pain if you are installing a new diskless client. Rarpd uses string comparison to macth an Ethernet address of a Client to the address in /etc/ethers. Thus leading zeroes in /etc/ethers will lead to a failure matching those addresses. It happened in 3.0-3.4, wheather you run YP or not. If you add an entry like this into /etc/ethers: 8:0:20:0:0b:11 venus the address will not be recognize. "0b" should be just "b" etc. Leonid ------------------------------ Date: Thu, 4 Feb 88 15:12:33 PST From: dhough@sun.com (David Hough) Subject: Motorola MC68881/MC68882 manual published by Prentice Hall I just received a copy of this manual, copyright by Motorola but published by Prentice Hall, so that, like the MC68020 manual, it may actually be obtainable in technical bookstores rather than by begging a Motorola sales office. At last there will be an answer for Sun's customers that I've told to read a 68881 manual without telling them how to obtain it. The MC68881 is the best hardware implementation of IEEE arithmetic from the standpoint of completeness and correctness, except for some minor complaints: transcendental functions are not monotonic in extended precision, log2(x) and 2**x aren't exact in some places they could be, there are no fmove-out instructions that round the source f register to the stored value not all the information that a user trap handler could exploit is available in user mode Transcendental functions are guided by the spirit rather than the specification of the IEEE standard, anyway, and most other hardware implementations don't come close. CISC enthusiasts should compare the 68881/2 instruction set to anything similar they may have been using to see the difference a clean orthogonal instruction set can make. RISC enthusiasts don't care about on-board transcendentals because, by their own admission, they are smart enough to code them faster using a smaller instruction set. They would be well advised, in general, not to be too pleased with their results until they are reasonably close to the 68881's in accuracy. I don't know whether or not to be surprised, but the page layout produced by Prentice Hall is not as aesthetically pleasing as that of the 68881 manual published by Motorola. David Hough ARPA: dhough@sun.com UUCP: {ucbvax,decvax,allegra,decwrl,cbosgd,ihnp4,seismo}!sun!dhough ------------------------------ Date: Wed, 3 Feb 88 22:05:37 PST From: Jim Frew <frew@hub.ucsb.edu> Subject: retrofitting 68881 chips Having stupidly ordered a 3/50 without the 68881 FPU, and also having a 3/160 with an old A79J 68881, I've been carefully following the discussion on adding the chips to your Sun-3's. Well, yesterday I ordered a couple of 68881's from Schweber Electronics (in Sherman Oaks CA, but they seem to be everywhere: I just called the Motorola phone number on the back of the 68020 spec sheet and asked for some distributors). Price $159 each. They arrived TODAY! Faster than some email ... The 3/50 was a plug-and-go operation. Pull the CPU card, plop the chip in, shove it back together. The 3/160 is almost as easy, except you have to move a CPU board jumper (at N-11, from 5-6 to 7-8) if you want the FPU clock to run at 16.67 instead of 12.5 MHz (see the Hardware Manual). Other than being careful to ground myself before touching the chips or the boards, I didn't follow any special procedures or use any weird tools. Unless I suffered from some incredible dumb luck, installing the chips yourself would seem to be a painless and VERY cost-effective procedure. -Jim Frew (frew@hub.ucsbcsl.edu) [[ Installing chips is usually pretty easy. It's getting them out again that almost always requires a special tool and patience. --wnl ]] ------------------------------ Date: Wed 3 Feb 88 07:08:20-PDT From: PAWKA@nosc-tecr.arpa Subject: Problems with Century C-2400 SMD drive I'm having a problem with a 3/160. I've connected a Century C-2400 SMD drive to a Xylogics 450 controller via an VME-Multibus adaptor. The system runs o.k. most of the time, however every once in a while it will reboot itself after printing the message: PANIC - REGS ACCESSED WHILE BUSY. Any help would be greatly appreciated. Mike Pawka Naval Ocean Systems Center PAWKA@NOSC-TECR.ARPA Comm: 619-553-4108 Av: 8-553-4108 ------------------------------ Date: Thu, 4 Feb 88 18:52:07 PST From: roode@pisa.orc.olivetti.com (David Roode) Subject: Using 3rd and 4th serial port on Sun CPU board? For those file servers without a Sun console, and hence no mouse and no keyboard, has anyone come up with a method of converting these 2 serial ports into additional modem connections, for UUCP, etc.? ------------------------------ Date: 5 Feb 88 10:50:00 EST From: "Dave Anderer" <dave@vax.oit.udel.edu> Subject: Monochrome Pin-outs? What at the pin assignments for the 9-pin connectors used for monochrome Sun-3s? Frequencies and levels where appropriate would be appreciated too. Thanks. ------------------------------ Date: 5 Feb 88 16:41:56 GMT From: asp@cos.com (Andrew S. Partan) Subject: 2 monitors off one Sun? We are looking for a cheap way of driving 2 (or more??) monitors off of one video signal (for instance - the signal that comes off the back of a 3/50). We thought that if we made a 'Y' cable, we could hook up 2 monitors to a Sun. Thanks for any info that anyone may have (even if it is telling us that we are off of our gourds). Please respond to Paul Serice (serice@cos.com). --asp (Andrew Partan @ Corporation for Open Systems) -- asp@cos.com or asp%cos.com@uunet.uu.net -- {uunet, sundc, decuac, hqda-ai, hadron}!cos!asp ASN.1 Object Identifier: "{joint-iso-ccitt mhs(6) group(6) 157}" ------------------------------ Date: 3 Feb 88 15:13:08 GMT From: yunexus!eriks@uunet.uu.net (Eriks Rugelis) Subject: lpd printing to a TCP terminal server attached printer? We have a Sun-3/280 running SunOS V3.4. We also have a Cisco TCP/IP terminal server. The server is capable of supporting 'reverse connections' so that a connection can be initiated from 'somewhere' on the network to one of the serial ports on the server. Reverse-connection serial ports are addressed in either of two ways: 1. you can connect to a particular TCP port at the IP address assigned to the server; a block of TCP port numbers are assigned to correspond to the serial ports on the server 2. you can 'alias' TCP ports on the server to distinct IP addresses The TCP ports can run the Telnet protocol or can be used as raw TCP data streams. The manual for the terminal server suggests that a raw TCP stream is usually the most desireable way of connecting to a port serving a printer device. Now come the questions: 1. How do I get lpd to speak to a raw TCP datastream? 2. Do I have to make source mods to lpd to make it understand how to 'dial' a TCP call? (or can I fake something into /dev and have lpd use that) 3. Does anyone else out there have some tools/sources/hints that they could throw our way to make this job a little easier? Presently I expect that the way to do this is to take lpd and hack it to open a socket to the desired TCP port on the server. From there on lpd would speak to the socket instead of the printer device that it would normally find in /dev. Any help/suggestions would be greatly appreciated. Voice: Eriks Rugelis Ma Bell: 416/736-5257 x.2688 NetNorth: eriks@yulibra Soon to be: eriks@libra.yorku.ca UUCP: seismo!mnetor!yunexus!eriks [[ Why not just write a back-end "filter" (either input or output, not sure which would be better), that does the connecting and the transferring? This is what Imagen does for their lpd support of an Ethernet-based printer. This way you don't have to modify lpd. --wnl ]] ------------------------------ Date: Thu, 4 Feb 88 11:36:01 EST From: rsalz@pineapple.bbn.com Subject: "Virtual" scrollbars? I want to create a text subwindow that scrolls through data that might not exist until you look at it. I'll know the total number of lines, but not their individual contents. Anyone got this, or hints on how to do it? (This is my first SunView program.) The ideal code would create the window and hook into two routines I specify, one to generate the line just before the one being shown, and one to generate the line just after. (I will maintain the line numbers internally.) Replies to me will get summarized for the list. ------------------------------ Date: Wed, 03 Feb 88 08:45:34 -0500 From: (Jeffrey A. Edelheit) <edelheit@community-chest.mitre.org> Subject: Need info on PostScript Laser Printers In the past, we have purchased Imagen printers. However, it has been suggested that we now start using PostScript printers. Thus, I am considering buying a PostScript laser printer to attach to a 3/280. A local supplier has suggested the QMS810, selling for $4,000. It uses the newer Canon engine (the same one as in the LaserJet II), and supposedly is much faster than the Apple LaserWriter Plus. Has anyone had experience with this printer? What other PostScript printers have people had experience with? Has anyone had any experience with Elan Computer Group and their Eroff package? Any suggestions for a source of device drivers and filters? [[ Specifically, those drivers and filters that run on SunOS. --wnl ]] Regards, Jeff Edelheit (edelheit@gateway.mitre.org) The MITRE Corporation 7525 Colshire Drive McLean, VA 22102 (703) 883-7586 ------------------------------ Date: 5 Feb 88 15:04 +0100 From: Igor Metz <metz@iam.unibe.ch> Subject: PostScript driver for SUN-GKS? A friend of mine asked me whether there exists a PostScript driver for GKS or a Metafile -> PostSCript translator. Has anybody of you seen such a beast? Thanks in advance Igor Metz X400: metz@iam.unibe.ch or metz@iam.unibe.chunet Institut fuer Informatik ARPA: metz%iam.unibe.ch@relay.cs.net und angewandte Mathematik UUCP: ..!uunet!mcvax!iam.unibe.ch!metz Universitaet Bern Switzerland Phone: (+31) 65 49 02 ------------------------------ End of SUN-Spots Digest ***********************