Sun-Spots-Request@RICE.EDU (William LeFebvre) (06/24/88)
SUN-SPOTS DIGEST Thursday, 23 June 1988 Volume 6 : Issue 122 Today's Topics: Re: text table full Re: slip on Suns Re: How do you use rex Re: Telebit Wants to Know: new features in TrailBlazers panic iechkcca setting subnet mask? Why can't I have multiple net IDs on a single ethernet board rf2spec in Tex? Why VT100 emulators? 8-bit shelltool? SCSI drivers for the Sun? Sun 2 Bug -- Need Assistance RPCGEN modifications for use with inetd(8) 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: Wed, 15 Jun 88 14:20:38 PDT From: Craig Leres <leres@helios.ee.lbl.gov> Subject: Re: text table full Like so many other things in SunOS, I've found that 4.3 BSD has a good solution to the problem. The first time I had trouble with too few text slots, I changed the calculation to match 4.3's param.c: RCS file: RCS/param.c,v retrieving revision 1.1 retrieving revision 1.2 diff -r1.1 -r1.2 52c52 < int ntext = 24 + MAXUSERS; --- > int ntext = 36 + MAXUSERS; I also config my kernels with a MAXUSERS of 8; I use 16 for "big memory" systems (16 Mbytes or more). Craig P.S. What kind of licensing arrangements have I been breaking by upping MAXUSERS? ------------------------------ Date: 16 Jun 88 01:41:26 GMT From: tekbspa!tss!joe@uunet.uu.net (Joe Angelo) Subject: Re: slip on Suns in article <1988.06.13.16.49.07.548.01790@rice.edu>, Mark G. Brown <markb@boulder.colorado.edu> says: | our situation sounds similar. we have a sub-net of suns off-campus served | by wasser, which is connected to sigi on campus via slip, and their | internet addresses are: | | 128.138.240.4 sigi # vax785, 4.3BSD | 128.138.240.2 wasser-gw # sun3/280, SunOS 3.5 | 128.138.239.1 wasser # sun3/280, SunOS 3.5 | I don't have the facilities (at the moment) to try this, but try typing: ifconfig sl0 down on the machine where sl0 and ie/le0 are interfaces for the same network addr. You'll see that sl0 will indeed come down, as it should; however, your gateway's view of it's own local net will be somewhat in error. That is, since sl0 was the interface for, say, 192.9.200, and the gateway was apart of 192.9.200 with ie0 also serving 192.9.200, bringing sl0 down will also bring 192.9.200 down on the gateway! Perhaps, in time, the routing tables will show that ie0 is the only remaining interface for 192.9.200 -- but that brings up a catch 22 scenerio ... think about it. I'd recommend making a seperate network number for the slip connection just as you would with an Internet Router connection, such as the below illustrates, ... 192.9.200.1 left_gateway 192.9.201.1 right_gateway 192.9.202.55 left_slip_gateway 192.9.202.66 right_slip_gateway ... and config the dst addr on left_gateway:sl0 as right_slip_gateway and the local addr of left_gateway:sl0 as left_slip_gateway (and visa versa). Routing tables look more logical this way and I personally feel it's a bad idea to have two interfaces to the same network addr (see netstat -r). Ofcourse, I don't have detail stats on the pros and cons and surely haven't fully tested the above. BUT -- if SLIP is truely an ASYNC/poor-man's version of the INR, it should work fine this way. Between 11 cats, 14 kittens, 5 dogs, 1 wife, 3 kids, 1 father-in-law, 37 acres of trees, 1 combat brother on leave, 1 VERY demanding boss, and a partridge... I have little time to read [study] standards and RFC -- anyone know what the standards for SLIP and INR net addrs are? Are multiple interfaces to the same network permitted? Joe Angelo -- Senior Systems Engineer/Systems Manager at Teknekron Software Systems, Palo Alto 415-325-1025 uunet!tekbspa!joe -OR- tekbspa!joe@uunet.uu.net ------------------------------ Date: Tue, 14 Jun 88 11:07:41 EDT From: rochester!srs!dan@rutgers.edu Subject: Re: How do you use rex Thanks for the tutorial on how to use rex, Steve. I've posted the source for rex_rexec(), an rexec() almost-replacement using rex, to Usenet's comp.sources.unix. ------------------------------ Date: 14 Jun 88 07:48:48 GMT From: munnari!ditmela.oz.au!A.Worsley@uunet.uu.net (Andrew Worsley) Subject: Re: Telebit Wants to Know: new features in TrailBlazers We are interested in using the modems to run a very cheap IP network over leased voice grade lines. They are much cheaper here than 9600 baud lines. The additional features we would like to see are: > NEW MODEM FEATURE IDEAS.... > > 3. TCP/SLIP Support in the modem Really IP/SLIP support although I don't know if there is much that can be done that is worth while. Prehaps improve the turnaround time so that rlogin works well across it? and sending the smaller IP packets ahead of larger ones that are going to different destinations (no sense unordering the data on a TCP connection) > 4. V.32 compatability (along with PEP) Always usefull, especially if you want other people's machines to call you up although the trailblzer appears to be almost a standard in itself. > 6. 4-wire Leased line support Very much, if we could have full duplex at the trailblazer speeds it would mean no response problems with running rlogin over the line. We can high 4 wire lines for about 20% more than a 2 wire line. We are actually so concerned about possible delays in response (because of the trailblazers half duplex nature) we were thinking of having two sets of modems, one pair for each direction. I guess this isn't such a big deal if you are just distributing news from a central site. > 7. Group III Fax support A rather nice idea as a lot of places in large organistations have fax machines but no e-mail. This provides at least reception of e-mail to them. Our OTC (Over seas Telecomunication Corporation) is providing on a test basis free of charge a X.400 to fax gateway which can use though. Andrew Worsley ------------------------------ Date: Wed, 15 Jun 88 10:44:53 EDT From: ileaf!io!leigh!ebr@eddie.mit.edu (Evan B. Ross) Subject: panic iechkcca I used to get the 'iechkcca' panic several times per day when doing printer development with a 3/160. As far as I could tell, there was some interaction between our Xylogics controller and a VPC parallel interface. One (or both) were not configured quite properly. We've played around with interrupt level settings on the cards as well as in the kernel. We found that the panic only occurred during times of high ethernet traffic combined with high utilization of the parallel interface. Sorry I can't be more specific, but we've changed our working environment so that we no longer have much of a problem. Now, if we could only stop the parallel interface/driver from zombie-ing, we'd be all set! :-) ...!{mit-eddie,bbn}!ileaf!ebr Evan B. Ross, Interleaf, Cambridge, Ma ...!sun!sunne!ileaf!ebr (617)577-9813 x5570 ------------------------------ Date: Wed, 15 Jun 88 14:14 EST From: "Yates, John H." <YATES@a.chem.upenn.edu> Subject: setting subnet mask? Soon after a server reboot the message "setting subnet mask to 0xffff0000" appears on our console server 3/280S. It serves 8 3/50 clients and did not do that under 3.2, and is now under 3.4. I can't be sure if it always has since the 3.4 upgrade. [We are on a campus Ethernet, and they supplied us with their sendmail configuration file that identifies other campus machines as domain servers- that may or may not be relevant]. Does anybody know what is going on? And where to look in the cardboard box of manuals? Nothing appears to be broken. John H. Yates , Ph.D. Director of the Chemistry Computer Facility Department of Chemistry University of Pennsylvania Philadelphia, PA 19104 yates%a.chem.upenn.edu@relay.upenn.edu (INTERNET) (215)898-4714 ------------------------------ Date: 15 Jun 88 23:29:07 GMT From: etnibsd!vsh@uunet.uu.net (Steve Harris) Subject: Why can't I have multiple net IDs on a single ethernet board We have a network consisting of a Sun 3/260 server, six 3/50 diskless clients, four 3/110 standalone systems, and a MicroVax. There are also some real-time controllers connected to the net. We currently use several bridges to isolate traffic on portions of the net. We are using class C net IDs. For reasons related to our application, we need to have clusters of systems each with a different net id. This means we need to gateway between different nets. What we would like to be able to do is this: from a machine on NET_A, send a packet to a machine on NET_B (i.e., construct and send a packet with a net ID different from that of MACHINE_A). However, it seems the OS will not allow such a packet to be transmitted. The error message is something like: network NET_B unreachable Sun support suggested we try: route add NET_B MACHINE_A 0 to fool the OS into acting as a gateway. The error message went away, but the packets were not sent. Sun support had no further ideas. I notice that etherfind goes into promiscuous mode to read packets not for NET_A; why can't the system write to the net promiscuously? As we hope to use 3/60s, we would really like to find a way around this problem. Any insights, solutions, or good ideas (or a good reason why things are done the way they are) gratefully accepted. Steve Harris uunet!etnibsd!vsh ------------------------------ Date: Wed, 15 Jun 88 12:18:22 EDT From: psuvax!idis!formtek!pen (Philip E. Nickerson, Jr.) Subject: rf2spec in Tex? I haven't had any problems including rasterfiles in LaTeX documents using rf2spec, but haven't yet sucessfully included one in a TeX document. Is there some trick I'm missing for Tex???? By the way, thank you to whoever submitted rf2spec to sunspots. (Some node along the way seems to be messing up return paths, so please use psuvax1!idis!formtek!pen@psuvax1.cs.psu.edu to reply, thanx!) -Phil Philip E. Nickerson,Jr. |UUCP {pitt,psuvax1}!idis!formtek!pen (412)937-4900|(800)FORMTEK| decvax!formtek!pen "Programming is simply |Snail Formative Technologies, Inc., Foster Plaza VII debugging a blank page!" | 661 Andersen Dr., Pittsburgh PA 15220 ------------------------------ Date: 15 Jun 88 02:16:22 GMT From: roy%phri@uunet.uu.net (Roy Smith) Subject: Why VT100 emulators? I'm sure this is a naive question, but what is it that vt100tool et al do that a normal shelltool window doesn't? Roy Smith, System Administrator Public Health Research Institute 455 First Avenue, New York, NY 10016 {allegra,philabs,cmcl2,rutgers}!phri!roy -or- phri!roy@uunet.uu.net [[ A real VT100 has many more features than are implemented in a shelltool. Such as double width and double height characters and different screen attributes. --wnl ]] ------------------------------ Date: Wed, 15 Jun 88 12:04:13 PDT From: ames!spar!navtech!mark@sally.utexas.edu (Mark Stevans) Subject: 8-bit shelltool? There has recently been mention of an 8-bit version of "shelltool" that would allow display of almost all of the 256 possible characters of a font. Where can I get a copy? Mark "Mr. DEC-10" Stevans ------------------------------ Date: Thu, 16 Jun 88 09:13:34 CDT From: bi@flora.wustl.edu (Hai Feng Bi) Subject: SCSI drivers for the Sun? Any information ( in addition to SUN's manual, " How to Write Device Driver ") on writing SCSI driver for SUN is APPRECIATED. Public domain drivers, routines, ideas etc are all helpful. Thanks in advance. My email address is: bi@flora.wustl.edu or bi%flora.wustl.edu@uunet.uu.net My regular mail address is Haifeng Bi Campus Box 1115 Washington University St. Louis, Missouri 63130 ------------------------------ Date: Thu, 16 Jun 88 13:41:48 -0200 From: Jacob Itzikowitz <yi@WISDOM.BITNET> Subject: Sun 2 Bug -- Need Assistance The following is a SUN/2 bug that we do not have an answer for. Would appreciate any help. Thanks. The environment is a SUN/2 file server, operating system level 3.4. The system is functioning ok except when one uses the vi editor with either a new or an old file and tries to use one of the q,w or wq commands in order to end the editing session, then the system crashs with the following message and then reboots itself. Parity Error! Bus Error Reg 83 <Valid,Uparerr Lparerr> Can't find parity error (transient ?) Panic: Parity Error Dumping to dev 301 offset 25408 dump succeded Rebooting Unix... Yaakov Itzikowitz Site & System Manager Applied Math & Computer Sc Dept Weizmann Institute of Sci Rehovot, Israel ------------------------------ Date: 16 Jun 88 10:56:07 GMT From: prlb2!kulcs!dannyb@uunet.uu.net (Danny Backx) Subject: RPCGEN modifications for use with inetd(8) The RPCGEN compiler provided with SUNRPC 3.9 does not allow writing servers suitable for use with inetd(8). The changes that have to be applied to rpcgen's output are minimal, though. (See the manual pages for inetd). The diffs provided here are minimal changes to the original rpcgen, in order to generate C programs for use with inetd directly. An option "-i" is added to rpcgen. If invoked, the generated programs are slightly different. No fundamentals in rpcgen's internals are changed. (Only a few printf's). The diffs are listed below. I used this for making a TCP-based server. If you do this, beware ! Your server should contain code to kill itself after it has provided you with the last answer. In my case, this means that the end of one of my service routines looks like this: %% #ifdef INETD %% alarm(5); %% signal(SIGALRM, exit); %% #endif INETD %% %% res = 1 ; %% return &res ; %% } <<< diff for rpc_main.c >>> 53a54 > int iflag; 64a66 > static int inetd = 0; 75c77 < "usage: %s infile\n", cmdname); --- > "usage: %s [-i] infile\n", cmdname); 77c79 < " %s [-c | -h | -l | -m] [-o outfile] [infile]\n", --- > " %s [-i] [-c | -h | -l | -m] [-o outfile] [infile]\n", 80c82 < " %s [-s udp|tcp]* [-o outfile] [infile]\n", --- > " %s [-i] [-s udp|tcp]* [-o outfile] [infile]\n", 83a86,87 > if (cmd.iflag) > inetd = 1; 285c289 < write_most(); --- > write_most(inetd); 337c341 < write_register(argv[i + 1]); --- > write_register(argv[i + 1], inetd); 369a374 > flag['i'] = 0; 383a389 > case 'i': 424a431 > cmd->iflag = flag['i']; <<< diff for rpc_svcout.c >>> 54c54,55 < write_most() --- > write_most(inetd) > int inetd; 74a76,79 > > if (inetd) > return; /* only 'pmap_unset' behind this */ > 91c96 < write_register(transp) --- > write_register(transp, inetd) 92a98 > int inetd; 99c105,108 < f_print(fout, "\t%s = svc%s_create(RPC_ANYSOCK", TRANSP, transp); --- > if (inetd) > f_print(fout, "\t%s = svc%s_create(0", TRANSP, transp); > else > f_print(fout, "\t%s = svc%s_create(RPC_ANYSOCK", TRANSP, transp); 119,120c128,132 < f_print(fout, ", IPPROTO_%s)) {\n", < streq(transp, "udp") ? "UDP" : "TCP"); --- > if (inetd) > f_print(fout, ", 0)) {\n"); > else > f_print(fout, ", IPPROTO_%s)) {\n", > streq(transp, "udp") ? "UDP" : "TCP"); Danny Backx | mail: Katholieke Universiteit Leuven Tel: +32 16 200656 x 3544 | Dept. Computer Science E-mail: dannyb@kulcs.UUCP | Celestijnenlaan 200 A ... mcvax!prlb2!kulcs!dannyb | B-3030 Leuven dannyb@blekul60.BITNET | Belgium ------------------------------ End of SUN-Spots Digest ***********************