#: 21304 S12/OS9/68000 (OSK) 14-Dec-95 06:21:48 Sb: #20977-#OS9 v3.0 ISP Fm: Jost Eberbach 73502,2041 To: M. Raabe 100327,1526 Hallo Martin, Niels ist leider nicht mehr bei unserer Firma, dafur bin ich jetzt hier. Wir haben leider immer noch gravierende Probleme mit dem MW-ISP. Kannst Du uns mal Informationen uber Eure TCP-Software zu kommen lassen? Vielleicht kann sie unsere Probleme ja wirklich beheben, das wurden wir gerne einmal ausprobieren! Unsere Addresse ist: ICT Integrated Circuit Testing GmbH z.Hd. Jost Eberbach Klausnerring 1a 85551 Kirchheim/Heimstetten Tel.: (089) 90 99 94 52 Im vorab schon mal vielen Dank, Jost There is 1 Reply. #: 21308 S12/OS9/68000 (OSK) 17-Dec-95 08:24:06 Sb: #21304-#OS9 v3.0 ISP Fm: Alptekin Aydogan 100527,324 To: Jost Eberbach 73502,2041 (X) >> gravierende Probleme mit dem MW-ISP.<< Hallo Jost, wir haben schon seit einiger Zeit den OS9TCP 2.0 von ASS im Einsatz. Die Hardware stammt ebenfalls von der Firma bei der Martin Raabe arbeitet. Probleme hatten wir, von einigen anfaenglichen OS9-Net Unvertraeglichkeiten mal abgesehen, bis jetzt nicht. Bisher hatten wir zwar keine gro_en Erfahrungen mit socket programming, aber seit ein paar Monaten bin ich dabei, ein Kommunikationsprogramm zwischen einem VME-System und einem PC auf dieser Basis aufzubauen. Die VME-Maschine soll MSR-Aufgaben uebernehmen, der PC fuer das GUI und Visualisierung verantwortlich sein. Das Problem Vermeidung von Polling habe ich geloest, in dem ich Signale und ss_sendsig() verwendet habe. Performance ist nicht so berauschend (Problem liegt wohl hauptsaechlich auf der PC-Seite, zwischen zwei VME-Systemen habe ich bis zu 250 MBaud erreicht) aber funktionieren tut's jetzt schon mal. Gruss Alptekin There is 1 Reply. #: 21309 S12/OS9/68000 (OSK) 18-Dec-95 06:22:15 Sb: #21308-#OS9 v3.0 ISP Fm: Jost Eberbach 73502,2041 To: Alptekin Aydogan 100527,324 (X) Hallo Alpetkin, danke fur Deine Info! Martin Raabe hat mir leider noch nicht geantwortet, er scheint sein email nicht so oft zu checken. Hast Du vielleicht eine Addresse oder Telefon-Nr. von der Firma? Gruss, Jost There is 1 Reply. #: 21315 S12/OS9/68000 (OSK) 20-Dec-95 02:40:48 Sb: #21309-OS9 v3.0 ISP Fm: Alptekin Aydogan 100527,324 To: Jost Eberbach 73502,2041 (X) Hallo Jost, bei dem extremen Messie-Aufkommen hier ist es sicherlich nicht verwunderlich, dass Antworten etwas laenger brauchen :-). Anscheinend sind alle privat hier, ich jedenfalls, keiner will sich so richtig outen mit Firma, Tel.-Nr. oder so. Na ja, ich glaube die Tel.-Nr. darf ich verraten, Du musst Dich allerdings verbinden lassen. Die Nummer des Vertriebs ist: 06131-918-100. Ja, es ist der VME-Hersteller in der Rheinland-Pfaelzischen Hauptstadt, dessen Name mit E anfaengt. Ich habe auch in der letzten Message was falsches gesagt, faellt mir ein. SS_SSIG funktioniert mit der ISP2.0 von MW aber -leider- nicht mit OS9TCP 2.0. Da muss man wohl oder uebel SS_SEVENT einsetzen. Man kann aber ein Event fuer alle Pfade verwenden, und dann nach Ueberpruefung des Puffers einlesen. Eine Select-Funktion ist ebenfalls in der Lib enthalten, das habe ich aber noch nicht ausprobiert. Gruss, Alptekin #: 21305 S12/OS9/68000 (OSK) 15-Dec-95 11:16:22 Sb: #nameserver Fm: Neal Wilsey 73757,2144 To: ALL My TCP/IP package on OS/9 requires me to have the hosts file in my /sys directory but has no provision for use of a domain nameserver. Is there any way I can take advantage of my institution's nameserver. It would cut down considerably the size of the hosts file. Neal Wilsey EMAIL: neal.wilsey@nrl.navy.mil There is 1 Reply. #: 21310 S12/OS9/68000 (OSK) 18-Dec-95 06:22:17 Sb: #21305-#nameserver Fm: Jost Eberbach 73502,2041 To: Neal Wilsey 73757,2144 (X) Neal, OS-9 uses the hosts (as well as the other network database files, i.e. networks, protocols and services) file only when it builds the inetdb data module. OS-9 actually doesn't use the files to get IP-Addresses, it only uses the data module. This allows embedded systems to use ISP aliases, even if they don't have any disk/file managers/devices installed. Usually the data module is created on startup using the idbgen utility from within a network startup script (startisp). You can tell the idbgen utility to use a different directory than the default directory by using the -d= option. could be the directory on your nameserver, assuming the drive is already mounted (requires NFS client installed). Of course this creates the the 'egg and hen' problem, what comes first? You would probably need a small local hosts file just for the nameserver, and then use the hosts file on the nameserver. By the way, the default directory for the hosts file should be /dd/etc, not /dd/sys. Regards, Jost There is 1 Reply. #: 21313 S12/OS9/68000 (OSK) 19-Dec-95 13:19:57 Sb: #21310-#nameserver Fm: Neal Wilsey 73757,2144 To: Jost Eberbach 73502,2041 (X) Jost, Thank you for your reply. Yes, my hosts file is in the ETC directory. From your reply, I gather that the OS-9 implementation of TCP/IP does not make use of the nameserver feature that has been in general use since 1991. Most unix systems have a file named resolve.cnf that contains the information needed to get the IP address from the nameserver. Some client systems such as NCSA Telnet and MacTCP place the domain and nameserver information in a configuration file so there is no need for a hosts file to exist on the local computer. I noted when I scanned the SOURCES file there was a UDP port associated with nameserver but found that there was no RESOLVE file. I was hoping to avoid downloading the nameserver hosts file during generation of the inetdb data module. Our institution's local hosts file is over 200K and the general hosts file is almost 1M! With the proliferation of networks, I really need to use the nameserver. Neal There is 1 Reply. #: 21317 S12/OS9/68000 (OSK) 20-Dec-95 09:26:36 Sb: #21313-nameserver Fm: Jost Eberbach 73502,2041 To: Neal Wilsey 73757,2144 (X) Hi Neal, wow, your hosts files are really big! You don't have to download them to your OS-9 system though. If you use NFS you can sinply mount the harddisk of Nameserver as a device for OS-9, then you can use the -d option of idbgen. But - do you really need all the entries in your nameserver's hosts file? You may get a problem with the inetdb getting too big. Unlike Unix, OS-9 has no virtual memory! Maybe you should create your own hosts file containing only the server aliases you really need. Regards, Jost #: 21340 S12/OS9/68000 (OSK) 24-Jan-96 05:48:18 Sb: #21315-OS9 v3.0 ISP Fm: Jost Eberbach 73502,2041 To: Alptekin Aydogan 100527,324 Hi Alptekin, mittlerweile habe ich von ein OS9TCP von ASS zur Probe bekommen, hatte aber noch wenig Erfolg damit. Ich weiss nicht genau, was ich falsch mache, aber ich krieg noch nicht einmal den TelnetDaemon zum Laufen, er steigt mit einer Fehlermeldung aus. Vielleicht liegt es an der OS-9 Version, welche OS-9 Version benutzt? Wir haben 3.0, lauft auf Eurocom 16 bzw. 17. Wurd mich freuen, wenn Du mir ein paar Tips zum OS9TCP geben konntest... Jost #: 21342 S12/OS9/68000 (OSK) 24-Jan-96 05:56:27 Sb: #20846-OS9 v3.0 ISP Fm: Jost Eberbach 73502,2041 To: Niels Peter Bogholm 100566,2262 Niels, >>WHY does the ISP crash when more than ca. 500 connections (499 TIME_WAIT and 1 established). Well I know why (namely 500 connections) but that's bad stuff...<< Yesterday I received a patch for the OS-9 kernel from Dr. Keil/MW, it fixes the >500 connections (>1024 vents) problem! Jost #: 21341 S12/OS9/68000 (OSK) 24-Jan-96 05:52:44 Sb: #21313-nameserver Fm: Jost Eberbach 73502,2041 To: Neal Wilsey 73757,2144 Hi Neal, I have some news for you and your nameserver problem. There is another ISP package for OS-9 besides the MicroWare-ISP available, called OS9TCP. It supports nameservers and is supposedly compatible with the MW-ISP, but has more features, like a select() function. It doesn't have NFS though, but can supposedly run with the NFS from Microware. Jost Press !> #: 21346 S12/OS9/68000 (OSK) 29-Jan-96 15:56:53 Sb: #21340-#OS9 v3.0 ISP Fm: Alptekin Aydogan 100527,324 To: Jost Eberbach 73502,2041 (X) Hallo Jost, wir hatten bis jetzt die Palette von E5-E17 unter OS9 2.4 mit OS9TCP laufen. Vor kurzem habe ich eine BAB-40 (Prozessor wie E17, aber andere Architektur) mit Version 3.0 (3.0.1 oder 3.0.2 weis ich im Moment nicht) in Betrieb genommen und hatte den gleichen Effekt, es ging nicht mit OS9TCP. Dafur funktioniert aber ISP von MW bis jetzt ohne Probleme (OS9-Net von MW vertragt sich nicht mit OS9-Net von ASS, aber wir setzen sowieso NFS ein). Info von Eltec war, OS9TCP sei fur die Version 3.0 noch nicht freigegeben. Da NFS, TELNET und FTP funktionierten, mein Programm (VME-PC-Kommunikation uber TCP-IP) auch mit kleinen Anderungen lief (Verkleinerung der Puffergrose beim Ubertragen), habe ich die Sache nicht mehr verfolgt. Ich nehme an, das irgendwann ein Update kommt oder so. Ich verwende noch SS_SEVENT aber werde es demnachst mit SS_SSIG versuchen, muste laut Dokumentation von MW funktionieren, aber man weis ja nie ;-). Grus, Alptekin There is 1 Reply. #: 21349 S12/OS9/68000 (OSK) 01-Feb-96 02:50:45 Sb: #21346-OS9 v3.0 ISP Fm: Jost Eberbach 73502,2041 To: Alptekin Aydogan 100527,324 (X) Hi Alptekin, Danke fur Deine Info. Hab ich mir schon gedacht, das OS9TCP nicht mit OS-9 3.0. Dasselbe ist mir schon mit verschiedenen seriellen und Grafiktreibern passiert. MW haben irgend etwas in der Speicherverwaltung fur Geratetreiber verandert, so dass viele alte Treiber nicht mehr funktionieren. Es ist aber erstaulich, das nach uber 2 Jahren, in denen 3.0 auf dem Markt ist, immer noch nicht alle anderen nachgezogen haben, und ihre Software angepasst haben! Wir haben von Dr.Keil ein patch fur den 3.0 Kernel bekommen, mit dem sich einige Probleme erledigt haben. Nur dass send() bleibt leider immer noch hangen. Naja. Jost Press !> #: 21359 S12/OS9/68000 (OSK) 06-Feb-96 23:19:42 Sb: #Checking on dead parents Fm: Bob van der Poel 76510,2203 To: All Is there any way a child process running in user mode can generate a signal if its parent dies. Here is my situation: parent processes spans child, parent/child communicate though shared data module created by parent, parent/child sync via events (event created by parent). If the parent dies a sudden death, the child not only hangs around forever, but even if you kill the child, the data module and the event still hang around. It would be nice if the child could be notified if the parent dies...then it could delete the data module (if it still exists) and delete the event. Any ideas on this appreciated! There is 1 Reply. #: 21361 S12/OS9/68000 (OSK) 07-Feb-96 04:06:12 Sb: #21359-Checking on dead parents Fm: Ian J Shearer 100410,2733 To: Bob van der Poel 76510,2203 (X) Bob, >> It would be nice if the child could be notified if the parent dies << Since you say 'if the parent dies a sudden death' I guess you can't signal the child from the parent, since the death is unexpected. You could have the child check that the parent is still there, however. It's a while since I did this, so you might need to improvise a little. Try having the parent pass the child the parent process id. At regular intervals, the child then fetches the parents process descriptor using _os_gprdsc(). If the parent is dead I THINK you should get an error from that call. I used to do a similar thing when I had problems with a process locking up on I/O. From the process descriptor (if I remember correctly) you can get the info normally shown by the 'procs' utility, so you can tell if all I/O activity on a process has halted, or if it's not taking any CPU time. Hope this helps. -Ian J Shearer, Onyx Systems Ltd. WWW - http://ourworld.compuserve.com/homepages/OnyxSystems #: 21360 S12/OS9/68000 (OSK) 06-Feb-96 23:20:07 Sb: #ppp Fm: Bob van der Poel 76510,2203 To: all Has anyone ported a ppp (point to point protocol) interface to osk? I have seen source code for a unix implementation of ppp and it doesn't seem to be all that complicated...but the unix code I saw created device drivers to do this. I'm not really knowledgeable about internet protocols, or unix device drivers to even start a port of this. It seems that my internet provider will no longer be supplying shell accounts and ppp will be the only access. I don't really need all the gfx stuff available under windows...so can this be done? There is 1 Reply. #: 21364 S12/OS9/68000 (OSK) 08-Feb-96 01:52:10 Sb: #21360-ppp Fm: David Breeding 72330,2051 To: Bob van der Poel 76510,2203 (X) > Has anyone ported a ppp (point to point protocol) interface to osk? > I have seen source code for a unix implementation of ppp and it doesn't > seem to be all that complicated...but the unix code I saw created device > drivers to do this. I'm not really knowledgeable about internet protocols, > or unix device drivers to even start a port of this. It looks like a pretty big undertaking. I believe I saw where someone said they might put a bug in your ear I have a zip file containing all the standards for all the protocols. The zip is about 833+ K Would you like for me to email it to you? > It seems that my internet provider will no longer be supplying shell > accounts and ppp will be the only access. I don't really need all the gfx > stuff available under windows...so can this be done? It certainly can be done, but whether anyone or group would have the time to do so is the big question. It is one thing I would have liked to have seen on our systems, but after taking a look at it, it looks like a tremendous amount of work ahead. As a matter of fact, just last night I installed Winsock and Eudora on my brother's PC just to see if I could and to become familiar with it. I set it up for my own account, (None of us have figured out how to come up with a complete logon script yet - if possible) and I don't think anyone could figure out the contortions we have to go through to get on Anyway, I went in with Eudora and got my mail, I think I had about 4 msgs. It was fantastic. In fact, I happened to have a MIME- encoded file attached and it came through automatically. Much simpler than using my system. It's not that bad with a shell account, but with the rates we are getting blistered with, you can't afford to spend so much time online reading and replying to mail. I've just about decided it's easier to switch than fight where Internet access is concerned. -- David Breeding -- CIS 72330,2051 | Delphi DBREEDING | dbree@duo-county.com Composed with InfoXpress/OSK Vr. 1.02 & VED Vr. 2.4.0 Press !> #: 21387 S12/OS9/68000 (OSK) 11-Mar-96 21:54:34 Sb: #21341-nameserver Fm: Neal Wilsey 73757,2144 To: Jost Eberbach 73502,2041 (X) Jost, Thank you for the information about OS9TCP. That sounds like the program I am looking for! Neal #: 21388 S12/OS9/68000 (OSK) 11-Mar-96 22:01:03 Sb: #21341-#nameserver Fm: Neal Wilsey 73757,2144 To: Jost Eberbach 73502,2041 (X) Jost, Thank you for the information about OS9TCP. Do you have any information on availability? Neal There is 1 Reply. #: 21389 S12/OS9/68000 (OSK) 12-Mar-96 12:01:42 Sb: #21388-nameserver Fm: Jost Eberbach 73502,2041 To: Neal Wilsey 73757,2144 >> Thank you for the information about OS9TCP. Do you have any information on availability? << Neal, OS9TCP has one big disadavantage, they only have a version for OS-9 2.4, not for 3.0. I've heard they're working on a new version. I don't know about the availability of OS9TCP in your country. In Germany OS9TCP is available from a company called ELTEC. Here is their address: ELTEC Elektronik Mainz GmbH Galileo-Galilei-Str. 11 Postfach 421363 55071 Mainz Germany Phone: ++49-6131 918-0 Fax: ++49-6131 918-199 Regards, Jost #: 21377 S12/OS9/68000 (OSK) 01-Mar-96 11:08:11 Sb: #MM-1 users? Fm: Jim Peasley [CA] 72726,1153 To: All Any MM-1 users still out there? What'cha using it for these days? ...Jim There is 1 Reply. #: 21379 S12/OS9/68000 (OSK) 01-Mar-96 20:55:12 Sb: #21377-#MM-1 users? Fm: John R. Wainwright 72517,676 To: Jim Peasley [CA] 72726,1153 (X) > Any MM-1 users still out there? > > What'cha using it for these days? > > ...Jim > Hey, its been a long time, Jim, Sure. There are still a fair number of MM/1 users around. Some have left CIS and hang out on the Delphi OS9 forum, but there are a few of us still here. My MM/1 has grown since I last talked to you. It now has a 68340 CPU (sort of a modified 68020), 9 Megs RAM, a CDROM and a tape drive along with 3 Hard drives of various sizes. It runs TeX/LaTeX2E, and Ghostscript, so I can produce and read Postscript documents as impressive as any PC or Mac. (With another program I wrote to produce Postscript drawings on the MM/1). It sends faxes -- I'm working on receiving them GRIN,-- but if I do that, the 486 next to it will lose another job. If I come up with an OSK scanner interface and a Graphical WEB browser, the 486 will get buried in dust bunnies. (Or at least dump messydos completely and run LINUX and OS-9000.) It keeps records for my business (SCULPTOR), and allows access from a surplus cheap terminal and modem at the office. Does all my CIS and DELPHI forum and email message work. Finally, and best of all, it allows me to snicker quietly at those who think Windoze-94.9996 (Pentium) is the ultimate system. John Wainwright There is 1 Reply. #: 21380 S12/OS9/68000 (OSK) 02-Mar-96 11:43:38 Sb: #21379-#MM-1 users? Fm: Jim Peasley [CA] 72726,1153 To: John R. Wainwright 72517,676 (X) >> My MM/1 has grown since I last talked to you. It now has a 68340 CPU (sort of a modified 68020), 9 Megs RAM, a CDROM and a tape drive along with 3 Hard drives of various sizes.<< John; All I can say is *WOW*! Sounds like you jacked up the MM/1 emblem and shoved a full-blown high-end Unix system under it! What'cha using for video? >>...snicker quietly at those who think Windoze-94.9996 (Pentium) is the ultimate system. << Heh, heh! Been there too! Although, as a WIN95 user on a 100mHz P5, I _DO_ enjoy going to the store and shopping for hardware/software that I can just come home and 'plug in'! For someone without a CS degree and a lot of time, this makes more sense than my MM/1. I was looking to see if there was enough activity that I could offer my MM/1 for sale, or if I should just quietly box it up and stuff it in the rafters. What'cha think? ...Jim There is 1 Reply. #: 21381 S12/OS9/68000 (OSK) 03-Mar-96 22:43:13 Sb: #21380-#MM-1 users? Fm: John R. Wainwright 72517,676 To: Jim Peasley [CA] 72726,1153 (X) > John; > All I can say is *WOW*! Sounds like > you jacked up the MM/1 emblem and > shoved a full-blown high-end Unix > system under it! What'cha using > for video? > The 68340 and 8-meg Ram were purchased at COCOfests, the software is available here and/or on the Delphi OS9 forum, plus at least one ftp site. The 68340 upgrade turns the machine into an "MM/1a". Still the old Tandy CM-8, with a Magnovox for a backup. That is one of the things that the SVGA machines can do better. Ghostscript on my MM/1 is pretty slow compared to the Linux version on my 486. I can see a whole page of Postscript on the MM/1 screen, but anything under 12 point type is a blur. There is a new machine out, being called an MM/1B by some, that will use a VGA card and monitor. Should be capable of better graphics with the right card and drivers. It also uses IDE hard drives. I think the basic motherboard goes for about $400. For the REAL details on the new machine, and what is available for the MM/1, EMAIL Bill Wittman -- WW2150@ACSPR1.acs.brockport.edu and David Graham -- NIMITZ@delphi.com > > I was looking to see if there was > enough activity that I could offer my > MM/1 for sale, or if I should just > quietly box it up and stuff it in the > rafters. What'cha think? > > ...Jim > Sure - post an offer here. If you like, I will tack it on the Delphi OS9 forum for you too. - Maybe on the COCOLIST also. I can't guarantee it will sell, but I can assure you that there are still a lot of COCO users around, and some of them are looking for a bigger OS9 machine (if the price is right GRIN). John There is 1 Reply. #: 21382 S12/OS9/68000 (OSK) 04-Mar-96 11:33:36 Sb: #21381-#MM-1 users? Fm: Jim Peasley [CA] 72726,1153 To: John R. Wainwright 72517,676 (X) >> Still the old Tandy CM-8, with a Magnovox for a backup. That is one of the things that the SVGA machines can do better.<< Yah, good grafix and resolution are a must if you spend any amount of time on a computer and have any vision difficulties! >>... an MM/1B by some, that will use a VGA card and monitor. ... It also uses IDE hard drives. << Cool! Should expand the usefulness of the machine greatly! >> For the REAL details on the new machine, and what is available for the MM/1, EMAIL Bill Wittman -- WW2150@ACSPR1.acs.brockport.edu and David Graham -- NIMITZ@delphi.com<< These the guys that are putting out the upgrades? Are Mark/Paul/Kevin still involved in the scene? >> If you like, I will tack it on the Delphi OS9 forum for you too. - Maybe on the COCOLIST also.<< That would be great! Let me put together a sale ad and get back to you. Shipping might be a problem if the CM-8 goes with it tho. ...Jim There is 1 Reply. #: 21383 S12/OS9/68000 (OSK) 04-Mar-96 20:59:13 Sb: #21382-#MM-1 users? Fm: John R. Wainwright 72517,676 To: Jim Peasley [CA] 72726,1153 (X) > for the MM/1, EMAIL Bill Wittman -- > WW2150@ACSPR1.acs.brockport.edu > and David Graham -- > NIMITZ@delphi.com<< > > These the guys that are putting out > the upgrades? Are Mark/Paul/Kevin > still involved in the scene? > Bill Wittman (Wittman Computer Products) and David Graham (Blackhawk) are MM/1 reps. Paul (Ward) left the scene a long time ago. Mark (Griffith) is still around, even checks in here now and then, but he is no longer active in the OS9 area. Kevin (Darling) also checks in here now and then, but he is really busy with his job -- programming OS9 boxes of course. Kevin (Pease) made the 68340 and 8-meg upgrades I mentioned earlier. The new WCP-306 aka MM/1b board is produced by Carl Krieder. John There is 1 Reply. #: 21384 S12/OS9/68000 (OSK) 05-Mar-96 14:01:09 Sb: #21383-MM-1 users? Fm: Mark Griffith 76070,41 To: John R. Wainwright 72517,676 (X) John, Yeah, I'm still lurking....more out of a decade of habit than anything else I guess (grin). See ya later. Mark Press !> #: 21395 S12/OS9/68000 (OSK) 26-Mar-96 09:37:25 Sb: #OS9_GETIME() Fm: Doug Bailey 73612,2146 To: Sysop (X) I am currently maintaining a piece of software running on V2.4 of OS9/68000. This software was compiled using Ultra C V1.1.1. In this software is a call to _OS9_GETIME in which the tick count is returned. The documentation indicates that the lower 16 bits contains the current tick. The original author assumed that the tick would count all the way to 65535 and then roll over. However, I am seeing the software roll over after the value 100. (There are 100 ticks in a second.) The documentation is woefully lacking in describing the values of the tick. Is this roll over at the number of ticks per second the way these functions always work? Thanks for your help! Doug Bailey is value. incredivb There is 1 Reply. #: 21396 S12/OS9/68000 (OSK) 30-Mar-96 18:06:58 Sb: #21395-#OS9_GETIME() Fm: Alptekin Aydogan 100527,324 To: Doug Bailey 73612,2146 (X) Doug, >> The documentation indicates that the lower 16 bits contains the current tick. << This is true, the higher two bytes indicate the tick-rate in ticks per second. If you have as usual 100 ticks per second, the u_int variable would count from $640000 up to $640063 and then start again with $640000 (this corresponds to 6553600 resp. 6553699 decimal). Regards, Alptekin There is 1 Reply. #: 21401 S12/OS9/68000 (OSK) 01-Apr-96 12:56:36 Sb: #21396-OS9_GETIME() Fm: Doug Bailey 73612,2146 To: Alptekin Aydogan 100527,324 (X) Thanks for the confirmation!! Press !> #: 21425 S12/OS9/68000 (OSK) 17-Jun-96 15:15:41 Sb: 38.4 KB SCF baud rate Fm: Doug Bailey 73612,2146 To: sysop (X) I am running OS-9 V2.4. I have an SCF driver driving a 68681 UART. Is there any way to set this bear up to drive a 38.4 KB serial port. I've seen some documentation where it tops out at 19.2KB. Thanks for your help Doug Bailey #: 21425 S12/OS9/68000 (OSK) 17-Jun-96 15:15:41 Sb: 38.4 KB SCF baud rate Fm: Doug Bailey 73612,2146 To: sysop (X) I am running OS-9 V2.4. I have an SCF driver driving a 68681 UART. Is there any way to set this bear up to drive a 38.4 KB serial port. I've seen some documentation where it tops out at 19.2KB. Thanks for your help Doug Bailey #: 21426 S12/OS9/68000 (OSK) 24-Jun-96 01:08:41 Sb: #21425-#38.4 KB SCF baud rate Fm: David Breeding 72330,2051 To: Doug Bailey 73612,2146 > I am running OS-9 V2.4. I have an SCF driver driving a 68681 UART. Is > there any way to set this bear up to drive a 38.4 KB serial port. I've > seen some documentation where it tops out at 19.2KB. The 68681 _is_ capable of 38.4 KB. However, most drivers are set up to only access 19.2. This is due to the fact, I think, that the 68681 is a "double-sided" thing, there are two sets of speed profiles, and 38.4 is one that is accessible to only one set. What system are you running? I have a Delmar System 5 and the driver for it was modified to access 38.4 I am not sure if driver would work with other systems or not. ** day++;dollar--; ** -- David Breeding -- CIS 72330,2051 | Delphi DBREEDING | dbree@duo-county.com Composed with InfoXpress/OSK Vr. 1.02 & VED Vr. 2.4.0 There is 1 Reply. #: 21427 S12/OS9/68000 (OSK) 25-Jun-96 23:53:39 Sb: #21426-#38.4 KB SCF baud rate Fm: Michael L. Smith 72762,3375 To: David Breeding 72330,2051 (X) The 68681 has 2 sets of baud rate tabels, and if you use the table that has 38.4 you loose the other table for both sides of the chip (which has some useful rates not in the 38.4 table). Therefore most drivers do not support 38.4kb. There is 1 Reply. #: 21428 S12/OS9/68000 (OSK) 27-Jun-96 23:42:35 Sb: #21427-#38.4 KB SCF baud rate Fm: David Breeding 72330,2051 To: Michael L. Smith 72762,3375 (X) > The 68681 has 2 sets of baud rate tabels, and if you use the table that > has 38.4 you loose the other table for both sides of the chip (which has > some useful rates not in the 38.4 table). Therefore most drivers do not > support 38.4kb. Yes, it's true that you lose some rates, but as I recall, some of the "best" ones are supported by both sides of the table. I believe 9600 is, and maybe 2400. Anyway, for my applications, it works pretty well. My system has two chips. My mouse is connected to /t0 one chip, and I connect my modem to the other (/t2). So these don't interfere with each other. The only thing you have to do, in case you _do_ need to use both sides of one chip is to not set an incompatible rate on the "other side". ** day++;dollar--; ** -- David Breeding -- CIS 72330,2051 | Delphi DBREEDING | dbree@duo-county.com Composed with InfoXpress/OSK Vr. 1.02 & VED Vr. 2.4.0 There is 1 Reply. #: 21429 S12/OS9/68000 (OSK) 28-Jun-96 20:01:19 Sb: #21428-#38.4 KB SCF baud rate Fm: Michael L. Smith 72762,3375 To: David Breeding 72330,2051 (X) That's correct, but if you select 38.4 you LOSE 19.2 which unacceptable to many users. Mike There is 1 Reply. #: 21431 S12/OS9/68000 (OSK) 29-Jun-96 02:31:54 Sb: #21429-#38.4 KB SCF baud rate Fm: David Breeding 72330,2051 To: Michael L. Smith 72762,3375 (X) > That's correct, but if you select 38.4 you LOSE 19.2 which unacceptable to > many users. I guess there might be cases when this might cause a problem. But in my case, I am using my system as a single-user system, and at this time, I am only using one side of two chips, one for my mouse and the other for the modem. My system does allow access to the full set. If you should need 19.2, you can still use it, just not on the other side of a chip using 38.4 I can only speak for myself, but for my application, I cannot see any time I would need 19.2 in favor of 38.4 .. In fact, to utilize a 28.8 modem, you _have_ to use the 38.4 setting to get 28.8 speed. ** day++;dollar--; ** -- David Breeding -- CIS 72330,2051 | Delphi DBREEDING | dbree@duo-county.com Composed with InfoXpress/OSK Vr. 1.02 & VED Vr. 2.4.0 There is 1 Reply. #: 21433 S12/OS9/68000 (OSK) 30-Jun-96 00:12:39 Sb: #21431-38.4 KB SCF baud rate Fm: Michael L. Smith 72762,3375 To: David Breeding 72330,2051 (X) I was _only_ referring to one chip. Of course you can have different baud rates on different chips. Also you can always find a work-around for most any problem. My only point is that the baud rate tables as they exist can cause problems. BTW Motorola has discontinued the chip. Philips still makes it. I don't know about about Thompson CSF. Mike #: 21430 S12/OS9/68000 (OSK) 28-Jun-96 23:15:20 Sb: #UniBasic Demos Fm: Michael L. Smith 72762,3375 To: SysOp (X) I have uploaded UniBasic Demos for OS9/68000, OS9/68020, and OS-9000. M.L> Smith There is 1 Reply. #: 21432 S12/OS9/68000 (OSK) 29-Jun-96 05:05:20 Sb: #21430-#UniBasic Demos Fm: Mike Ward 76703,2013 To: Michael L. Smith 72762,3375 (X) Thanks Michael! There is 1 Reply. #: 21434 S12/OS9/68000 (OSK) 30-Jun-96 00:28:05 Sb: #21432-#UniBasic Demos Fm: Michael L. Smith 72762,3375 To: Mike Ward 76703,2013 (X) Mike, The ub_tar.gz and Ub68ktar.gz files that I uploaded Friday are corrupted. Please remome them and replace with the new ones I just uploaded. The good files are ub_68000 and ub_68020. The OS-9000 file is ok as is. Sorry for the inconvenience. Mike There is 1 Reply. #: 21435 S12/OS9/68000 (OSK) 01-Jul-96 03:53:52 Sb: #21434-#UniBasic Demos Fm: Mike Ward 76703,2013 To: Michael L. Smith 72762,3375 (X) Okey doke. There is 1 Reply. #: 21436 S12/OS9/68000 (OSK) 01-Jul-96 13:04:52 Sb: #21435-UniBasic Demos Fm: Michael L. Smith 72762,3375 To: Mike Ward 76703,2013 (X) Thank you. Press !> #: 21437 S12/OS9/68000 (OSK) 09-Jul-96 10:54:36 Sb: FastTrack Fm: Gerhard Schumacher 100331,1572 To: SYSOP (X) Hi, does anyone have experience with the MW FastTrack cross compiler ? I have the problem that my output files have about double size as the ones compiled on a 68K system with cc. Regards, Thomas Press !> #: 21482 S12/OS9/68000 (OSK) 09-Oct-96 05:57:41 Sb: #Fast IRQ Fm: J. Eberbach 106147,521 To: all Hello all, I would like to use the _os_firq(..) function to install a C function as an interrupt service routine on the fast IRQ table. I understand the C function cannot called directly as an ISR, it needs some assembly language code around it. Does anybody have an example of the code needed to use a C function as a fast ISR? Regards, Jost There is 1 Reply. #: 21483 S12/OS9/68000 (OSK) 11-Oct-96 08:48:13 Sb: #21482-#Fast IRQ Fm: J. Eberbach 106147,521 To: J. Eberbach 106147,521 (X) since noone ever replied, and since I also didn'T get any help from Microware or Dr. Keil, I had to find out for myself. So here is an example: #include #include #include #include #include #include #include #include #include #include "/mwos/os9/src/defs/machine/reg.h" /*make this program system state */ _asm("_sysattr: equ 0xA001"); /**********************************************************************/ int irqsvc_f_t1(); void *get_global_base(void); int irq_timer1(char *system_globals) { /* do your interrupt service */ return(0); } /**********************************************************************/ #define CLEAN_UP \ /* cleanup paths */ \ prc->_path[0]=0; \ prc->_path[1]=0; \ prc->_path[2]=0; main () { error_code err; register procid *prc = sysglob(procid*, D_Proc); /* fix up standard i/o for a system state process */ stdin -> _fd = prc->_path[0]; stdout -> _fd = prc->_path[1]; stderr -> _fd = prc->_path[2]; _from_new(stdin); _from_new(stdout); _from_new(stderr); /* install interrupt service routines */ if (err = _os_firq(VMEVT1, 0, irqsvc_f_t1, get_global_base())) exit(_errmsg(error, "Can't install timer1 interrupt service routine.\n")); else{ printf ("Timer1 Interrupt service routine installed!\n"); } /* Setup hardware to produce interrupts */ /* enable interrupts */ /* do your stuff */ /* disable interrupts */ /* deinstall interrupt service routine */ _os_firq(VMEVT1, prior, NULL, get_global_base()); CLEAN_UP exit(0); } _asm("params equ 8"); _asm("irqsv set 0"); _asm("static set 4"); _asm("port set 8"); _asm("get_global_base:"); _asm(" move.l a6,d0"); _asm(" rts"); _asm("irqsvc_f_t1: move.l a6,-(a7)"); _asm(" movem.l d1/a0/a1/a3-a5,-(a7)"); _asm(" move.l a6,d0"); _asm(" movea.l a2,a6"); _asm(" bsr irq_timer1"); _asm(" move.l d0,d1"); _asm(" beq.s irqf_done"); _asm(" ori.b #01,ccr"); _asm("irqf_done movem.l (a7)+,d1/a0/a1/a3-a5"); _asm(" movea.l (a7)+,a6"); _asm(" rts"); There is 1 Reply. #: 21484 S12/OS9/68000 (OSK) 18-Oct-96 16:00:51 Sb: #21483-#Fast IRQ Fm: John Rodgers 73223,1204 To: J. Eberbach 106147,521 (X) There doesn't seem to be much activity here!! But thanks for the info -- it may come in handy. Thanks, John There is 1 Reply. #: 21485 S12/OS9/68000 (OSK) 26-Oct-96 09:46:15 Sb: #21484-Fast IRQ Fm: Jost Eberbach 73502,2041 To: John Rodgers 73223,1204 Hi John, no, there's hardly any activity here lately. I'm wondering if anyone else still uses OS-9 at all! Have they all gone to use VxWorks? Anyway, trying fast IRQs was a little disappointig for me. Using them may make a difference on a system with many different interrupt sources, but on my system it didn't increase the interrupt response time at all. With or without fast IRQs, OS-9 gave me an average 20 (+/- 2) microseconds interrupt response time on the MVME162. Could be a little faster, IMO, but for most applications it's probably good enough. Regards, Jost Press !> The OS-9 Forum Read Menu Read 1 [NEW] messages 2 Message NUMBER 3 WAITING messages for you (0) Search [new] messages 4 FROM (Sender) 5 SUBJECT 6 TO (Recipient) Enter choice !>1 #: 21482 S12/OS9/68000 (OSK) 09-Oct-96 05:57:41 Sb: #Fast IRQ Fm: J. Eberbach 106147,521 To: all Hello all, I would like to use the _os_firq(..) function to install a C function as an interrupt service routine on the fast IRQ table. I understand the C function cannot called directly as an ISR, it needs some assembly language code around it. Does anybody have an example of the code needed to use a C function as a fast ISR? Regards, Jost There is 1 Reply. #: 21483 S12/OS9/68000 (OSK) 11-Oct-96 08:48:13 Sb: #21482-#Fast IRQ Fm: J. Eberbach 106147,521 To: J. Eberbach 106147,521 (X) since noone ever replied, and since I also didn'T get any help from Microware or Dr. Keil, I had to find out for myself. So here is an example: #include #include #include #include #include #include #include #include #include #include "/mwos/os9/src/defs/machine/reg.h" /*make this program system state */ _asm("_sysattr: equ 0xA001"); /**********************************************************************/ int irqsvc_f_t1(); void *get_global_base(void); int irq_timer1(char *system_globals) { /* do your interrupt service */ return(0); } /**********************************************************************/ #define CLEAN_UP \ /* cleanup paths */ \ prc->_path[0]=0; \ prc->_path[1]=0; \ prc->_path[2]=0; main () { error_code err; register procid *prc = sysglob(procid*, D_Proc); /* fix up standard i/o for a system state process */ stdin -> _fd = prc->_path[0]; stdout -> _fd = prc->_path[1]; stderr -> _fd = prc->_path[2]; _from_new(stdin); _from_new(stdout); _from_new(stderr); /* install interrupt service routines */ if (err = _os_firq(VMEVT1, 0, irqsvc_f_t1, get_global_base())) exit(_errmsg(error, "Can't install timer1 interrupt service routine.\n")); else{ printf ("Timer1 Interrupt service routine installed!\n"); } /* Setup hardware to produce interrupts */ /* enable interrupts */ /* do your stuff */ /* disable inhrwhu atRsO- ema eae0(-2mreoditrprsoeeoteVE2 odbalt ae,M,b omsapctn tsrbb o ng. Rad, JtPs