From leonid@rst.co.il Mon Jul 1 06:55:38 1996 From: leonid@rst.co.il (Leonid Rosenboim) Date: Mon Jul 1 06:55:40 PDT 1996 Subject: Military References We have a potential customer from the Military segment who would like to gather some references to Military projects who use VxWorks for operative equipment (airborne, space, misile, tactical etc.) in order to be confident that VxWorks itself is reliable enough for these applications. I know many of you are out there on the net, but many of these issues are classified. So if there is any unclassified information of the type they are looking for, please send it to me. Unfortunately the customer themselves are too clasified to send this request to the net, so we are doing them the service. ----------------------------------------------------------------------- Leonid Rosenboim Phone: +972-3-559-8144 R S T Software Industries Ltd. Mobile: +972-50-307-142 P.O.Box 11502, AZUR 58017, Israel Fax: +972-3-559-8244 http://knight.trendline.co.il/~leonid/ E-Mail: leonid@rst.co.il From wbrown@csg.lbl.gov Mon Jul 1 07:19:59 1996 From: wbrown@csg.lbl.gov (Bill Brown) Date: Mon Jul 1 07:20:00 PDT 1996 Subject: remove Re: Chancellor Group (symbol = CHAG) You are sending these messages to an inappropriate mailing list. Please stop sending these messages at once! -bill wlbrown@lbl.gov wbrown@csg.lbl.gov Keeper of the VxWorks User Group mailing list - vxwexplo-request@lbl.gov > http://www.netamerica1.com > > I just got wind that Chancellor Group just announced a stock dividend for > July 31. SGA Goldstar has sent out a "buy" recommendation. I understand > other investment advisors are looking to recommend CHAG. The company has a > strong book value. The short sellers need to cover. This looks like a > good situation to me. What do you think? They are located at: > > http://www.netamerica1.com > > Bob Williams, 206-269-0846 > > To terminate from my Investment Opportunities, Reply to term@moneyworld.com > with "remove" in the subject field. From ddavies@worf.dvg.com Mon Jul 1 08:21:40 1996 From: ddavies@worf.dvg.com (Douglas Davies) Date: Mon Jul 1 08:21:42 PDT 1996 Subject: remove Junk email is truely bad form... From wbrown@csg.lbl.gov Mon Jul 1 09:20:02 1996 From: wbrown@csg.lbl.gov (Bill Brown) Date: Mon Jul 1 09:20:04 PDT 1996 Subject: Re: remove > Submitted-by: ddavies@worf.dvg.com (Douglas Davies) > > Junk email is truely bad form... > You got that right! It's starting to appear in many mailing lists. A couple that I subscribe to from my personal private ISP account are getting bombed by this kind of crap. I guess I'm gonna have to figure out how to become the list address in order to keep this BLEEP under control. But anyway, it's probably not productive to discuss it on the list. I only posted the reply I sent so that the subscribers would be aware that I'm trying to get it stopped. Disclaimer: Any opinions are my own and have | -bill nothing to do with the official policy or the | wlbrown@lbl.gov management of L.B.L, who probably couldn't | Berkeley, CA care less about employees who play with trains. | aka wlbrown@netcom.com From daemon@csg.lbl.gov Tue Jul 2 04:00:49 1996 From: daemon@csg.lbl.gov Date: Tue Jul 2 04:00:51 PDT 1996 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Tue Jul 2 04:00:46 PDT 1996 Subject: Problems with VxWorks5.2, MC68030 and LM629 Subject: Re: VxWorks Option INCLUDE_ADA Subject: Re: tFtpdTask Subject: Re: RPC UNDER VXWORKS : ANY EXPERIENCE Subject: Tornado makefiles(was: Re: Tornado Conversion) Subject: Using the MMU to catch stack over/underflows Subject: Re: Developing NIC drivers for VxWorks - how? Subject: Re: 5.3 BSP for HK80/V960D Subject: Re: Developing NIC drivers for VxWorks - how? Subject: Re: Q:static c++ objects under vxWorks ------------------------------------------------------- Newsgroups: comp.os.vxworks,comp.robotics.misc Subject: Problems with VxWorks5.2, MC68030 and LM629 Date: Thu, 20 Jun 1996 15:37:47 +0200 From: akaske@ensem.u-nancy.fr (Axel KASKE) Organization: CRAN-ENSEM, INPL Message-ID: Hello, we have a problem programming a motion controller LM629 by a MC68030 /25MHz running VxWorks5.2. We have an assembler programm running on a MC68000 (10MHz) (our VME-testboard) which works well. Now we use the same code on the MC68030. The programm exits without error, but the LM629 is badly programmed. We don't believe that it is a fault of VxWorks. Perhaps it is a timing problem and the MC68030 is to fast, but even with the debugger (vxgdb) we get not the same results as with the MC68000. Has anyone experience with a LM629. If you can give us a hint, please answer by email. Thanks in advance. Axel Kaske *************************************************************************************** Axel Kaske Tel.: (++33) 83.59.57.09 CRAN-ENSEM, INPL Fax.: (++33) 83.59.56.44 2, avenue de la Foret de Haye EMail.: akaske@ensem.u-nancy.fr 54516 Vandoeuvre les Nancy Cedex http://web.ensem.u-nancy.fr/Ensem/laboratoires/cran/equipes/vri/Perso/akaske/kaske.html *************************************************************************************** --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks Option INCLUDE_ADA Date: Fri, 28 Jun 1996 17:22:10 -0700 From: Magnus Ericson Organization: Ericsson Microwave Systems Message-ID: <31D47732.6268@kiemw.ericsson.se> References: <4qun90$n45@gde.GDEsystems.COM> Chet Kackman wrote: > > To anyone: > I was configuring VxWorks for NFS when I stumbled across the VxWorks > option INCLUDE_ADA in the configAll.h include file located in vw/config/all. > Most of our code is developed in ADA, and I was wondering if this option buys > us anything. We've never included it in previous kernels builds, and have had > no > adverse effects with our ada runtime. However, we would like to know what > purpose it serves. If any of you out there in VxWorks land have any ideas > cough em up. Thanks in Advance > Chet Kackman@GDESystems.com In the past there were a special "version" of VxWorks around that was built especially to support VADSWorks from Verdix. VADSWorks is no longer dependent on a special version of VxWorks. If the option has anything to do with this historical Ada-version it is of absolutly no use today and is hopefully doing nothing. - -- Magnus Ericson M.Sc, Ericsson Microwave Systems Torshamnsgatan 32C, S-164 84 STOCKHOLM Tel: +46 8 757 35 38 Fax: +46 8 750 74 42 Disclaimer: "The opinions expressed here are my own, and not those of my employer." --------------------------- Newsgroups: comp.os.vxworks Subject: Re: tFtpdTask Date: Fri, 28 Jun 1996 14:01:48 GMT From: mfisher@dyn.com (Mark Fisher) Organization: The World @ Software Tool & Die Message-ID: References: <4qujvv$1s5@galaxy.uci.agh.edu.pl> Reply-To: mfisher@dyn.com Sender: news@world.std.com (Mr Usenet Himself) piwek@student.uci.agh.edu.pl (Bartlomiej Piwek) wrote: >Hi, > I'm starting with VxWorks 5.1 system and I've got a problem with >built in ftp server. When I log in I can put any username and password >and system logs me in. Anyone can tell me how can I prevent this ? There is no way using the WRS FTP task because they only implement the minimum FTP commands, which I guess doesn't include security! Personally I think this is really a very very poor implementation decision by WRS. Anyone else agree? At present the only way to have security is to write your own ftp server task. Maybe if all those who consider this unacceptable would pester WRS about it they would provide a patch with security!! Mark Fisher --------------------------- Newsgroups: comp.os.vxworks Subject: Re: RPC UNDER VXWORKS : ANY EXPERIENCE Date: 20 Jun 1996 12:15:12 -0700 From: chaos@idiom.com (H.J. Bae) Organization: Peaceful Star, Oakland, CA Message-ID: <4qc800$7mv@idiom.com> References: <4qb8ck$isb@postman.jet.uk> One of the things to keep in mind is that original rpcgen from SUN generates non reentrant code. Most return types are going to be declared local static in clnt routines generated by rpcgen. This may not have produced your error, but it's something to keep in mind. - -- H.J. Bae http://www.peacefulstar.oakland.ca.us --------------------------- Newsgroups: comp.os.vxworks Subject: Tornado makefiles(was: Re: Tornado Conversion) Date: Wed, 26 Jun 96 11:02:08 GMT From: bergonz@labs.it (Michele Bergonzoni) Organization: Laboratori Fondazione G. Marconi Message-ID: <4qr5bg$2a4_001@gandalf.labs.it> References: In article , demars@terminus.com (Alan DeMars) wrote: >We've been trying to cutover to Tornado but have had lots of trouble with >makefile compatibility. Does the Windows version of Tornado's Make work >the same as the Unix version? It seems that we have to have a single flat >makefile rather than makefiles in each code subdirectory linked together >with a "master" makefile. Could this be true? I'm using this technique with a statement like: $(MAKE) -C c:\mypath\mydir -f c:\mypath\mydir\makefile mytarget It didn't work until I specified -C. I think that forward slash and back slash should be equivalent, but you should try to make sure. Hope this helps, Bergonz - --- Ing. Michele Bergonzoni - Laboratori Fondazione Guglielmo Marconi info and PGP key at http://www.dsnet.it/aziende_it/labs-main.html Phone:+39-51-846900 Fax:+39-51-846479 e-mail: bergonz@labs.it Telecommunications consultants, Microwave systems & circuits development. --------------------------- Newsgroups: comp.os.vxworks Subject: Using the MMU to catch stack over/underflows Date: 28 Jun 1996 15:35:50 GMT From: pww@cray.com (Paul Wells) Organization: Cray Research, Inc. Message-ID: <4r0u4m$hl@walter.cray.com> A while back, someone posted a message about using the MMU to catch stack over- and under-flows in VxWorks. This is a common type of error during the development phase of a project, and one that is often difficult to debug, since the resulting memory corruption may not be detected until long after it occurs. Also, VxWorks places the task's TCB at the high end of the stack, so errors such as over-indexing an automatic array can corrupt the TCB, while running off the other end of the stack will often corrupt some other task's TCB. To make our lives easier, we've developed a small replacement function for taskSpawn() that can help catch these problems. It takes the same arguments as the standard taskSpawn(), but the tasks it creates have the following differences: 1. The TCB is allocated from an array in the BSS segment, rather than from the system memory pool. In addition to being away from the stacks, this can make it easier to locate the TCBs when analyzing a post-mortem dump file. 2. The stacks remain in the system memory pool, but have one or more pages allocated at each end of the stack. These pages are marked read-only by the MMU, so an attempt to write off either end of the stack will result in an immediate data access exception. This adds some overhead at task creation and deletion time, but not when the task is running. In our application, task creation and deletion are not on the critical path, so this isn't a problem. The code is currently running with VxWorks 5.2 on a SPARC-based platform. It uses the vmBaseLib functions, so VxVMI is not required. If anyone is interested, send me some email at pww@cray.com and I'll send you the source (340 lines, including comments). - - Paul Wells Cray Research, Inc. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Developing NIC drivers for VxWorks - how? Date: 28 Jun 1996 08:42:15 GMT From: David Costain <100072.3624@CompuServe.COM> Organization: VG Organic Message-ID: <4r05t7$hn$1@mhade.production.compuserve.com> References: <4qec4b$ni3@wellspring.us.dg.com> Ask Wind River tech support/sales about the VxWorks BSP porting Kit. This consists of documentation and test/validation software. The documentation contains a section detailing the creation of a VxWorks Network Interface Driver. This section provides an example of a network interface driver, containing all the necessary hooks into VxWorks. All the developer has to do is to flesh the example code out with their NIC specific driver code. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: 5.3 BSP for HK80/V960D Date: Tue, 25 Jun 1996 10:35:43 -0500 From: James Gmeinder Organization: Heurikon Corporation Message-ID: <31D0074F.41C6@heurikon.com> References: <31CE700D.42D4@sagaut.de> Nicolas-Peter Pohland TM4 wrote: > > Hi all, > > I have a HK80/V960D board from Heurikon and a working 5.2 BSP. But I am > interested in > evaluating Tornado which requires 5.3. > > The problem is that the supplied 5.3 BSP assumes the the board is a > "Heurikon HK80/V960E". > > After reading "upgrading a VxWorks BSP for Tornado" I came to the > conclusion that I > need an appropriate sysSerial.c (not based on Z8530) to extend my 5.2 > BSP. > > Instead of trying to write this myself I was wondering if someone > already has a working > 5.3 BSP for a HK80/V960D or the matching sysSerial.c and so on ... > Hi, The V960D and V960E are indeed quite different and will not run each others BSPs. Wind River sells the V960E BSP, but not the V960D BSP, which is why you received only the V960E BSP. Heurikon has not yet released the V960D 5.3 BSP. We are currently working on the release of all our BSPs under 5.3, and the rollout should begin shortly. However, you should still be able to evaluate Tornado by simply using the Tornado "installOption" tool to re-install the 5.2 BSP for the V960D. The install procedure recognizes the BSP as version 5.2 and performs a conversion to a 5.3 BSP. Of course, it does not provide the serial driver required for system level debugging. Using this method, you should be able to perform a general evaluation of Tornado, with the exception of system level debugging. If you have any questions or concerns, please contact me directly. Regards, Jim Gmeinder - -- ============================================================================ James Gmeinder Heurikon Corporation (608) 831-5500 Application Engineer Madison, WI 53717 jmg@heurikon.com http://www.heurikon.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Developing NIC drivers for VxWorks - how? Date: Thu, 27 Jun 1996 20:53:32 GMT From: dap@network.com (Dave Peterson) Organization: Network Systems Corporation Message-ID: References: <4qec4b$ni3@wellspring.us.dg.com> Reply-To: dap@network.com Sender: news@network.com In article <4qec4b$ni3@wellspring.us.dg.com>, Yamudda Wezamibootz writes: > I've tried the manuals, the online help, and the tech support, > and all come up short. Wind River is kind enough to provide > sample code, but no documentation on the interface between the > driver and the system. And it's precisely that interface that > has me in knots. > > Can anyone point me to information on how to develop a network > driver for VxWorks? Profuse thanks in advance. > > > BTW, is there a FAQ for this group? > > Not sure what type of NIC you are developing but take a look at struct ifnet in h/net/if.h and possibly ether_attach(). =================================================================== Dave Peterson phone : 612-391-1008 Senior Engineer Network Systems Corp. email: dap@network.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Q:static c++ objects under vxWorks Date: 1 Jul 96 17:03:17 GMT From: bss@wrs.com (Benjamin Skrainka) Organization: Wind River Systems, Inc. Message-ID: References: <4r7vn2$46h@dub-news-svc-5.compuserve.com> 100567.1652@compuserve.com (Rolf Grzibek) writes: Greetings! >just a simple question. >Is it possible to have static objects under vxWorks? >Stroustrup states that the constructors of static object are called >before main(). As there is no main() under vxWorks it seems to be >impossible. But what if the constructor has no side effects except >initializing datas members (_no_ malloc() or other function calls)? >Does g++ just generate initialized data? vxWorks provides support for static objects via munching and some runtime support to call static constructors and destructors. The munching process builds two arrays, containing function pointers to the static constructors and destructors, respectively. If your vxWorks image has static Xtors, you should add a call to cplusCtorsLink() to call the the constructors at a convenient place when vxWorks boots. There is an analogous call cplusDtorsLink() which you should call when you shutdown your application. To deal with dynamic loading, you can configure vxWorks to either call the Xtors automatically (with cplusXtorSet()) or you can manually call your Xtors with cplusCtors() and cplusDtors(). For more information, see the manual page on cplusLib or look at the information in the Tornado User's Guide and the Programmer's Guide. Some of the information may be in the section on the Wind Foundation Classes (WFC), I think. At the risk of venturing into the abyss of religion, I would argue that you should write code that does not require static Xtors. Instead, use pointers and allocate the objects at a convenient place in your boot up process. That way you have control over when the Xtors are called. Your code will also be more portable. Good luck, Ben ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Benjamin S. Skrainka Wind River Systems Development Engineer 1010 Atlantic Avenue bss@wrs.com Alameda, CA 94501-1147 voice 510.814.2147 http://www.wrs.com/ facsimile 510.814.2164 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ --------------------------- End of New-News digest ********************** From ihw@rtp.co.uk Tue Jul 2 04:47:43 1996 From: Ian Willats Date: Tue Jul 2 04:47:44 PDT 1996 Subject: Miscellanous drivers required Hello all, Does anyone know if there are VxWorks drivers written for any of the following devices, and if so how one might get source code: * Intel 82595 Ethernet controller (not the 82596!) * Intel 28F008 flash memory (existing 28Fxxx driver goes up to 'F020, would this work??) * NCR 53C80 SCSI controller (not the 53C90; this one's prehistoric!) As always, all suggestions will be gratefully received. Ian ---------------------------------------------------------------- Ian Willats Real-Time Products Ltd. Unit 8, Holt Court, e-mail: ihw@rtp.co.uk Aston Science Park, Jennens Road, tel: +44 121 333 6955 Birmingham. B7 4EJ. England. fax: +44 121 333 5433 ---------------------------------------------------------------- From hiske@ict.nl Tue Jul 2 07:14:27 1996 From: Hiske Huiting Date: Tue Jul 2 07:14:28 PDT 1996 Subject: Supervisory ? Non-priviledge I have a 2 questions about VME communication. 1. What is the difference between Supervisory and Non-priviledge addressing modes? 2. What is the difference between Master and Slave access? Info: I have a Force-3CE SPARC board with VxWorks running connected via VME on a SGI-Power Challenge (OS: IRIX). An additional memoryboard is an option. Thanx, Hiske Huiting ICT Automatisering Deventer BV. e-mail: hiske@ict.nl From espin@serval.kla.com Tue Jul 2 08:43:26 1996 From: espin@serval.kla.com (Geoffrey Espin x2496) Date: Tue Jul 2 08:43:29 PDT 1996 Subject: rpc again > Subject: Re: RPC UNDER VXWORKS : ANY EXPERIENCE > From: chaos@idiom.com (H.J. Bae) > One of the things to keep in mind is that original rpcgen from SUN > generates non reentrant code. Most return types are going to be > declared local static in clnt routines generated by rpcgen. I wish you'd quit saying that, Hwa-Jin. :-) Undoubtedly the problem the original poster had was a missing rpcTaskInit() which is documented (not rpcInit() as someone else posted), but totally unexpected and unintuitive. The routine rpcTaskInit() just does a taskVarAdd() of the TCB's "rpcModList" which were statics in Sun's RPC (which was upgraded to RPC4.0 in 1989(!) by H.J. Bae.). I have been recently using Sun's rpcgen and writing reentrant code -- although I haven't had cause to "prove it" in the current development state. But the key points are: 1. Makefile: RPCGEN = rpcgen -N -M -C The "-N" is for nice :-) parameter passing as opposed to struct-city. Then use "-l" for client code, and "-m" for server. Use -Sc for example client template code using it -- not that useful. Use -Ss for example server template code using it -- very useful, except that you have to replace all the /* insert your code here */ comments, and then if you change your ".x" file you're screwed. Also, Sun inserts its own header files which I doubt are POSIX :-) so eleminate them in the Makefile, e.g. ... sed -e 's%%%' \ -e 's%^.*synch.h.*%%' \ -e 's%^.*thread.h.*%%' \ -e 's%^.*sysent.h.*%%' \ -e 's%^.*memory.h.*%%' \ -e 's%^.*stropts.h.*%%' \ -e 's%^.*resource.h.*%%' \ -e 's%^.*syslog.h.*%%' \ -e "s%netpath%tcp%" $@.tmp >> $@ [Excuse the brain-dead sed substitute commands, oh well, they work.] 2. Your server (xxxSvc.c) should contain something like : #define syslog(x, msg) logMsg("syslog: %s\n", (int)msg, 0,0,0,0,0) #include "logLib.h" #include "sysLib.h" #include "semLib.h" #define mutex_t SEM_ID #define mutex_lock(x) \ { \ int stat; \ if (*(x) == NULL) \ *(x) = semMCreate (SEM_DELETE_SAFE); \ stat = semTake (*(x), sysClkRateGet () * 1); \ if (stat == ERROR) \ logMsg ("mutex_lock: timeout!\n", 0, 0, 0, 0, 0, 0); \ } #define mutex_unlock(x) \ { \ if (*(x) == NULL) \ *(x) = semMCreate (SEM_DELETE_SAFE); \ semGive (*(x)); \ } #include "sdl_svc.c" /* "ver batim" rpcgen [...] -m ... */ #include "sdlSvcPrm.c" /* stubs filled in after rpcgen [...] -Ss ... */ There's some other massaging like eliminating "main"... but this will put you on track to doing things "the right way", that is, building from your ".x" file. Back to the rpcTaskInit() call, you might notice that if your task happens to make an NFS call before running your RPC code (say invoked from the VxWorks shell)... you won't observe the problem, because all NFS functions call -- guess what? :-) [This is a hint: all your client-side routines which might be called from any task must arrange an initial call to rpcTaskInit() for that task.] Geoff -- espin@idiom.com From ekins%sifvsj.dnet.@sifd01.solartron.com Tue Jul 2 10:02:30 1996 From: ekins%sifvsj.dnet.@sifd01.solartron.com (Phil Ekins - Solartron R&D - Farnborough UK) Date: Tue Jul 2 10:02:32 PDT 1996 Subject: VME multiplexer card wanted Hi, I am looking for some VME cards for a particular application, the one I have had no luck with yet is: - VME multiplexer card (giving 16:4 and 16:2 mux'ing, i.e. 16 channels in and giving optionally 4 lots of 4 out OR 8 lots of 2) The switching of the channels should be by BOSFET ideally (solid state relays, rather than mechanical relays, this is for reliability) This is to the switch the signal source to channels an a DSP analyser card, thus reducing the number of DSP analyser cards required in the system. The signal will be switched to the DSP analyser, which will then take its measurements and perform its calculations, then the mux will switch in the next channel (note, the muxer will switch about every 1/2 sec) I am also looking for a couple of other cards, which I have a couple of leads, but not satisfying the full requirement, so any recommendations will be appreciated: - Analogue i/p cards, with 16 bit ADC (preferably integrating), nominally 16 channels (but can be more): 2 readings per sec Programmable ranges: 0.2V to +/-12V Current measurement 4 - 20mA This is to measures signals such as temperature and pressure. - Digital i/p card, 8 channels total required: 0 to 48V 1ms timing required (can produce interrupt for CPU to do timing) As is always, we are price sensitive. Any info./pointers will be helpful, Thanks in advance, Phil. ******************************************************************* * Phil Ekins, Internet: pekins@solartron.com * * Solartron, Tel: (44) 1252 376666 Ext.2287 * * Victoria Road, Fax: (44) 1252 543854 * * Farnborough, GU14 7PW, * * England. * ******************************************************************* From BlairMJ@apmail.mpd.lord.com Tue Jul 2 13:55:45 1996 From: BlairMJ@apmail.mpd.lord.com Date: Tue Jul 2 13:55:46 PDT 1996 Subject: NDN:NDN:NDN:RE>What routine returns my target's IP address? PostalUnion-QM(tm) 'Non-Delivery Notice'! Your mail through the QuickMail 'INTERNET' Gateway could not be delivered! ***QM - No Valid Recipients! It has been deleted. Please re-send with correct addressing. From BlairMJ@apmail.mpd.lord.com Tue Jul 2 13:57:35 1996 From: BlairMJ@apmail.mpd.lord.com Date: Tue Jul 2 13:57:37 PDT 1996 Subject: NDN:NDN:NDN:Core File Checksum Mismatch PostalUnion-QM(tm) 'Non-Delivery Notice'! Your mail through the QuickMail 'INTERNET' Gateway could not be delivered! ***QM - No Valid Recipients! It has been deleted. Please re-send with correct addressing. From BlairMJ@apmail.mpd.lord.com Tue Jul 2 13:58:01 1996 From: BlairMJ@apmail.mpd.lord.com Date: Tue Jul 2 13:58:03 PDT 1996 Subject: NDN:NDN:NDN:RE>Core File Checksum Mismatch PostalUnion-QM(tm) 'Non-Delivery Notice'! Your mail through the QuickMail 'INTERNET' Gateway could not be delivered! ***QM - No Valid Recipients! It has been deleted. Please re-send with correct addressing. From BlairMJ@apmail.mpd.lord.com Tue Jul 2 13:58:13 1996 From: BlairMJ@apmail.mpd.lord.com Date: Tue Jul 2 13:58:15 PDT 1996 Subject: NDN:NDN:comp.os.vxworks newsdigest PostalUnion-QM(tm) 'Non-Delivery Notice'! Your mail through the QuickMail 'INTERNET' Gateway could not be delivered! ***QM - No Valid Recipients! It has been deleted. Please re-send with correct addressing. From beeman1@popcorn.llnl.gov Tue Jul 2 18:41:54 1996 From: Bart Beeman Date: Tue Jul 2 18:41:56 PDT 1996 Subject: Host/Target Config w/shared VME bus Hello, I'm new to this list as well as VxWorks and I have a basic system configuration problem. We have a host sparc5 and a target 162 living in the same VME crate. On power up the host takes longer to boot than the target. If the target is allowed to attempt to boot before the host is ready both machines hang. The host takes ~45sec to finish. Does anyone know how to sync these two guys or at least make the auto boot of the 162 something like 45sec insted of the standard 7sec? Any thoughts greatly appreciated! Thanks Bart From lepage@ftms.com Tue Jul 2 20:06:44 1996 From: Rob Lepage Date: Tue Jul 2 20:06:46 PDT 1996 Subject: mvme1603 com3&4 I'm using VxWorks 5.3 on an MVME1603. I need access to two serial ports on the front panel, and thought I could easily do this using ports 3 & 4, but found that VxWorks doesn't support use of these ports. The controller for these ports is the Zilog 85230. I don't need to do anything fancy like synchronous transfers or high data rates (simple 9600 baud will be sufficient). Is there a driver available to use these serial ports? Is the z85C30 close enough to the 85230 so that the driver in target/src/drv/serial could be used or adapted? Ideally, I would like to have port 3 act as com1/console, and port 4 as com2. Any help will be greatly appreciated. -Rob lepage@ftms.com From leonid@rst.co.il Wed Jul 3 00:02:27 1996 From: leonid@rst.co.il (Leonid Rosenboim) Date: Wed Jul 3 00:02:33 PDT 1996 Subject: Re: tFtpdTask > At present the only way to have security is to write your own ftp > server task. > Maybe if all those who consider this unacceptable would pester WRS > about it they would provide a patch with security!! Originally, the FTP server for VxWorks was written for a very particular purpose - enable slave VME cards to boot via backplane network, from a master board which has got a SCSI disk, which stores all bootable images for the entire create. For that purpose the implementation is quite sufficient, and performs well, and to add too much would be a waste of good DRAM. On the other hand, we have modified the WRS-supplied FTP server module to include some additional commands; Make Dir, Remove Dir, Delete File, Rename File, and login security using the WRS-supplied loginLib, which was designed for optional rlogin/telnet security. This security is rather minimal, but should be sufficient for most purposes. This modification was done for a customer who are using our FlashLib Flash File System on VxWorks, and use FTP to maintain the files on the Flash device from a remote management console. Unfortunately, for all those interested, we can not provide you with the patch, either source or object without specific permission from WRS. WRS SUpport team has got this patch from us, right after it was tested, so all of you who got a support contract with WRS should contact support, and ask for the modified FTP server. We dont know if and when these mods are going into the main release of VxWorks. I guess it does depend on customer demand. ----------------------------------------------------------------------- Leonid Rosenboim Phone: +972-3-559-8144 R S T Software Industries Ltd. Mobile: +972-50-307-142 P.O.Box 11502, AZUR 58017, Israel Fax: +972-3-559-8244 http://knight.trendline.co.il/~leonid/ E-Mail: leonid@rst.co.il From daemon@csg.lbl.gov Wed Jul 3 04:00:36 1996 From: daemon@csg.lbl.gov Date: Wed Jul 3 04:00:39 PDT 1996 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Wed Jul 3 04:00:32 PDT 1996 Subject: NTP on Nitro260 Subject: Re: excHookAdd() working??? Subject: Re: Telnetting to VxWorks. Subject: Re: Telnetting to VxWorks. Subject: MVME1604-011 booting Subject: Re: tFtpdTask Subject: Motorola PPC Bootroms Subject: Re: Q:static c++ objects under vxWorks Subject: gzip ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: NTP on Nitro260 Date: 24 Jun 1996 12:57:37 GMT From: Chris Marsey Organization: NASA/MSFC Message-ID: <31CE9114.4EC0@usscouts.org> Anyone using an NTP client for vxworks? On a Heurikon Nitro260? Any info/help would be appreciated. Chris Marsey cmarsey@usscouts.org --------------------------- Newsgroups: comp.os.vxworks Subject: Re: excHookAdd() working??? Date: Thu, 27 Jun 1996 12:42:10 -0500 From: Steve Strauss Organization: Lockheed Martin Tactical Aircraft Systems Message-ID: <31D2C7F2.7BEA@kuwait.lfwc.lockheed.com> References: <31CDDB24.3A07@ksl.co.il> Eli Eliezer wrote: > does any one of you, used the excHookAdd() function to capture an > exception? > I tryed it, but it not seem to work. > I am trying to catch bus-errors. I have been using it successfully for some time. I add a function that spawns a process that displays information about the exception. There's no real trick. I called it just like the documentation says to. The only reason my hook function spawns a process instead of doing the work itself is because I want to use the tt() function, but it doesn't seem to like being applied to the task that's calling it. Hope this helps, Steve --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Telnetting to VxWorks. Date: 2 Jul 1996 12:19:36 -0700 From: kla@leland.Stanford.EDU (Earl Mitchell) Organization: Stanford University, CA 94305, USA Message-ID: <4rbso8$do@elaine25.Stanford.EDU> References: <4qn02k$db3@samba.rahul.net> In article , Mark Fisher wrote: >Necromancer wrote: > >>I am trying to telnet to VxWorks from a remote system, and run my application >>main routine instead of the shell. I have set up vxWorks so that the shell is >>not built into the code, and my application is run. This works on the console >>but when I telnet into the box, the client connects to the VxWorks telnet serverbut then hangs. > >>Any ideas on how I can do this? > >If you wish to use the WRS telnet task then you must also be running >the WRS shell. I asked about this before and they indicated that the >shell and the telnet task were intertwined. The only way you can get >do what you want (run your own version of a shell and have telnet) is >to write your own telnet task and don't us ethe WRS version. This is correct. If you read the documentation (wink wink) it says that the telnet and rlogin daemons handle remote comms and then they redirect your keyboard input to the shell task. So the daemons are just feeding your remote input to the shell. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Telnetting to VxWorks. Date: Tue, 25 Jun 1996 14:25:32 GMT From: mfisher@dyn.com (Mark Fisher) Organization: The World @ Software Tool & Die Message-ID: References: <4qn02k$db3@samba.rahul.net> Reply-To: mfisher@dyn.com Sender: news@world.std.com (Mr Usenet Himself) Necromancer wrote: >I am trying to telnet to VxWorks from a remote system, and run my application >main routine instead of the shell. I have set up vxWorks so that the shell is >not built into the code, and my application is run. This works on the console >but when I telnet into the box, the client connects to the VxWorks telnet serverbut then hangs. >Any ideas on how I can do this? If you wish to use the WRS telnet task then you must also be running the WRS shell. I asked about this before and they indicated that the shell and the telnet task were intertwined. The only way you can get do what you want (run your own version of a shell and have telnet) is to write your own telnet task and don't us ethe WRS version. Mark --------------------------- Newsgroups: comp.os.vxworks Subject: MVME1604-011 booting Date: Tue, 02 Jul 1996 17:56:25 +0100 From: Alan Sparrow Organization: Roke Manor Research Ltd. Message-ID: <31D954B9.77DB@roke.co.uk> Dear All, I am using the Motorola MVME 1604-011 with Tornado for Windows NT. I would like to keep the Motorola PPCBug ROM's (instead of the VxWorks Boot ROM's) and use them to transfer the bootrom code via tftp. I can get the tftp bit to work, but when I execute the bootrom code it generates a illegal instruction error. Any suggestions gratefully received, Alan. - -- - ----------------------------------------------------------------- Alan Sparrow E-mail: alan.sparrow@roke.co.uk Roke Manor Research Ltd Romsey, Hants. SO51 0ZN. Tele: +44 (0)1794 833348 UK Fax: +44 (0)1794 833433 - ----------------------------------------------------------------- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: tFtpdTask Date: 2 Jul 1996 12:28:47 -0700 From: kla@leland.Stanford.EDU (Earl Mitchell) Organization: Stanford University, CA 94305, USA Message-ID: <4rbt9f$ic@elaine25.Stanford.EDU> References: <4qujvv$1s5@galaxy.uci.agh.edu.pl> In article <4qujvv$1s5@galaxy.uci.agh.edu.pl>, Bartlomiej Piwek wrote: >Hi, > > > I'm starting with VxWorks 5.1 system and I've got a problem with >built in ftp server. When I log in I can put any username and password >and system logs me in. Anyone can tell me how can I prevent this ? > >Bartek Piwek > >P.S. I'm working with SUN host, and load system from it via ethernet connecion. You need to define a login username and password. I believe they are defined in configAll.h. Look for the string "LOGIN". --------------------------- Newsgroups: comp.os.vxworks Subject: Motorola PPC Bootroms Date: Tue, 25 Jun 1996 21:51:56 +0100 From: John Gordon Organization: UnipalmPIPEX server (post doesn't reflect views of UnipalmPIPEX) Message-ID: <31D0516C.250D@dial.pipex.com> Hi, I am trying to program a flash rom to act as the VxWorks bootrom for a MVME1604-011 board. So far I have had no success though. I have an image of the bootrom which loads over the network successfully. To create the ROM version I changed the #define INCLUDE_MOT_BUG_ROMS to #undef INCLUDE_MOT_BUG_ROMS in config.h and rebuilt bootrom_uncmp & bootrom_uncmp.hex. The latter I then program onto an AM29F040 device using a PROM programmer (ie not using the pflash command in the Bug ROMs). The device verified correctly in the programmer so I feel I can trust the code that is there. I then put the flash rom in XU1 on the processor mezzanine (having first removed both the Motorola ROMs). [Just to make sure I have also tried the device in XU2 but it makes no difference to the result] When the board is powered up the Board Fail light remains lit, and there is no activity on the console serial link. If anybody has any ideas, or better still if any body has got this working, I'd appreciate a pointer! Thanks in advance, John... --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Q:static c++ objects under vxWorks Date: Mon, 01 Jul 1996 16:41:42 -0700 From: Oisin Hurley Organization: IONA Technologies Message-ID: <31D86236.15FB7483@iona.com> References: <4r7vn2$46h@dub-news-svc-5.compuserve.com> Rolf Grzibek wrote: > Is it possible to have static objects under vxWorks? > Stroustrup states that the constructors of static object are called > before main(). As there is no main() under vxWorks it seems to be > impossible. But what if the constructor has no side effects except > initializing datas members (_no_ malloc() or other function calls)? > Does g++ just generate initialized data? A simple answer: Yes. The centerline compiler for vxworks generates code that gets called when a module is ld'ed or unld'ed which will respectively construct or destruct static objects. There should be a similar procedure for the g++ compiler. You may have to call a flag-setting function from your startup script. When you have more than one static object ordering may become a problem if there is interdependence -- watch out for this. regards, Oisin - --- ohurley@iona.com --------------------------- Newsgroups: comp.os.vxworks Subject: gzip Date: 3 Jul 1996 02:08:27 GMT From: DPRUITT@stc.Net (Donald F. Pruitt) Organization: Maverick Labs, Inc. Keywords: compression Message-ID: <4rckmr$hqe@news1.sunbelt.net> Hi, Does anyone know where I could obtain a gzip implamentation for VxWorks? If not I would also like to know where I might obtain the source which I could port. Any help would be greatly appreciated. TIA Don Pruitt (dpruitt@stc.net) --------------------------- End of New-News digest ********************** From pardo@lassen.rti.com Wed Jul 3 10:14:02 1996 From: pardo@rti.com (Gerardo Pardo-Castellote) Date: Wed Jul 3 10:14:05 PDT 1996 Subject: Re: comp.os.vxworks newsdigest >> In article , Mark Fisher wrote: >> >Necromancer wrote: >> > >> >>I am trying to telnet to VxWorks from a remote system, and run my application >> >>main routine instead of the shell. I have set up vxWorks so that the shell is >> >>not built into the code, and my application is run. This works on the console >> >>but when I telnet into the box, the client connects to the VxWorks telnet serverbut then hangs. >> > >> >>Any ideas on how I can do this? >> > >> >If you wish to use the WRS telnet task then you must also be running >> >the WRS shell. I asked about this before and they indicated that the >> >shell and the telnet task were intertwined. The only way you can get >> >do what you want (run your own version of a shell and have telnet) is >> >to write your own telnet task and don't us ethe WRS version. >> >> This is correct. If you read the documentation (wink wink) it says that the >> telnet and rlogin daemons handle remote comms and then they redirect your >> keyboard input to the shell task. So the daemons are just feeding your remote >> input to the shell. Our RTILib utilities package includes a reentrant shell program called "rshell" that does not use telnet and will allow you to start any routine when you connect to the target. It will also allow you to have multiple of these simultaneosly. You can get details by sending email to info@rti.com -Gerardo =========================================================================== = = = = Gerardo Pardo-Castellote = email: pardo@rti.com = = Real-Time Innovations, Inc. = Phone: (408) 720-8312 = = 155A Moffett Park Drive, Suite 111 = Fax: (408) 734-5009 = = Sunnyvale, CA 94089 = http://www.rti.com = = = = =========================================================================== From sridhar@prc.allied.com Wed Jul 3 10:33:52 1996 From: sridhar@prc.allied.com (Sridhar Adumusilli) Date: Wed Jul 3 10:33:53 PDT 1996 Subject: Re: Q:static c++ objects under vxWorks Rolf Grzibek wrote: > Is it possible to have static objects under vxWorks? > Stroustrup states that the constructors of static object are called > before main(). As there is no main() under vxWorks it seems to be > impossible. But what if the constructor has no side effects except > initializing datas members (_no_ malloc() or other function calls)? > Does g++ just generate initialized data? After you compile the C++ code (using g++), run "nm", and "munch" command on the object files. Then, link the object modules, and load them to target using WindSh. Refer to Tornado User Guide. If your VxWorks build doesn't call the constructors automatically, you need to call "cplusCtors" from WindSh prompt. This calls all the constructors for the static objects in your code. Also you can use cplusDtors to call the destructors. After you have run the constructors, run the driver function (main/first function to be executed). There is a function call (I can't remember its name) made during the kernel initialization, that calls the constructors for global objects. I think you can use this to initialize your global objects, when you are ready to create a bootable application. Sridhar Sridhar Adusumilli AlliedSignal Aerospace From harvey@wrs.com Wed Jul 3 18:21:24 1996 From: Harvey Wong Date: Wed Jul 3 18:21:26 PDT 1996 Subject: drivers for sound chips? Hi, Just wondering if anyone has written drivers for any sound chips used in multimedia-type applications. If yes, which ones? Thanks, Harvey harvey@wrs.com From daemon@csg.lbl.gov Thu Jul 4 04:00:49 1996 From: daemon@csg.lbl.gov Date: Thu Jul 4 04:00:51 PDT 1996 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Thu Jul 4 04:00:44 PDT 1996 Subject: Re: Access Fault in VME-167 Subject: Re: Compiling on power pc 604 Subject: Re: printf() caveat Subject: Re: Q:static c++ objects under vxWorks Subject: Value returned by sizeof() Subject: HELP: FDDI driver problems, not cpu#0 Subject: Re: Q:static c++ objects under vxWorks Subject: Re: MIB-II System Group - NVRAM Subject: tFtpdTask Subject: How to setup a secure shell? Subject: Re: Value returned by sizeof() Subject: Re: MVME1604-011 booting ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Access Fault in VME-167 Date: Wed, 3 Jul 1996 09:11:10 GMT From: valentin@netcom.com (Valentin Pepelea) Organization: Netcom Communications, Santa Clara, CA Message-ID: References: <4r069p$pn3@f111.iassf.easams.com.au> Sender: valentin@netcom5.netcom.com In article <4r069p$pn3@f111.iassf.easams.com.au> dmr@f111.iassf.easams.com.au (Dewan RASHID) writes: > >I am trying to run a piece of my program >and consistently getting access fault. >... >Access Fault >Program Counter: 0x00eeac72 >Status Register: 0x3000 >Access Address : 0x05b80016 >Special Status : 0x0145 >Task: 0xe2563c "t8" > >I am not sure why the address 0x05b80016 is accessed by the program. > >The disassembled code at location 0x00eeac72 are given: > >-> l 0x00eeac72 >eeac72 3568 0006 0dc4 MOVE .W (0x6,A0),(0xdc4,A2) >eeac78 202a 0dbe MOVE .L (0xdbe,A2),D0 >eeac7c 2200 MOVE .L D0,D1 >eeac7e 2001 MOVE .L D1,D0 You are looking at the wrong address. The 68040 is a pipelined processor, and several instructions may have been loaded and started by the time the bad instruction causes an access fault. Hence, look two or three instructions before 0xeeac72. Your code is definitely accessing the wrong memory address, hence the access fault. Do not waste time inspecting configuration registers such as that of VmeChip2. The most likely problem is that you are using an uninitialized pointer. Another possible problem is that the pointer is a local variable (put on the stack) and that the stack is overflowing, thus getting the pointer corrupted. Another possibility is that some other task/ISR is overwriting the pointer. If your access fault consistenly occurs at the same address, you may put a breakpoint at the given location, and watch exactly why the address is being accessed. If the fault occurs intermittently, you may have to attach a logic analyzer, and break upon access to the access fault exception handler. Get one of the analyzers that can give you 68040 disassembly. Since the code you are running is most likely running out of the cache, you should consider disabling the cache, in order to get easy to follow traces from you analyzer. Personnally, I found these types of bugs easy enough to solve by simply looking at the C/assembly code with a source level debugger, and figuring out which variable contains the bad value, and determining slowly how it got it. Valentin - -- "An operating system without virtual memory Valentin Pepelea is an operating system without virtue." Embedded Sytems Consultant (415) 346-7105 - Ancient Inca Proverb valentin@netcom.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Compiling on power pc 604 Date: 3 Jul 1996 14:21:45 GMT From: Don Small Organization: Sandia National Laboratories Message-ID: <4rdvlp$s77@news.sandia.gov> References: <4qrp5j$2h3@hammer.msfc.nasa.gov> john.weir@msfc.nasa.gov (John Weir) wrote: >I have a Motorola 1604 with a power pc 604, running VxWorks 5.3 and >Tornado using the GNU 2.7 toolset. If I try to compile a program i get an >error cpp: Invalid option '-lang-c89'. I use a Makefile very similar >to one that works fine with VxWorks 5.2 running on a MC68060 CPU. The >makefile I use is shown below: > > John M. Weir > john.weir@msfc.nasa.gov > I hope you have found your solution already. If not I have a similar problem when I have not set the GCC_EXEC_PREFIX environment variable. What happens is ccppc is just a front end for the gnu compiler and it will default to a path which probably contains your hosts gcc compiler. The following line in my cshrc fixes this problem setenv GCC_EXEC_PREFIX ${WIND_BASE}/host/${WIND_HOST_TYPE}/lib/gcc-lib/ Hope this helps. Don --------------------------- Newsgroups: comp.os.vxworks Subject: Re: printf() caveat Date: Tue, 02 Jul 96 11:38:53 GMT From: bergonz@labs.it (Michele Bergonzoni) Organization: Laboratori Fondazione G. Marconi Message-ID: <4rb1od$190_001@gandalf.labs.it> References: <4r4b0s$m8n@newsfeed.cts.com> In article <4r4b0s$m8n@newsfeed.cts.com>, lees@cts.com (Lee Slaughter) wrote: >I'm wondering what possible evil resulted from using printf()? >give (void) >{ .. > logMsg("Semgive...\n",0,0,0,0,0,0); > /* do NOT use printf */ .. > wdStart (myWatchDogId, sysClkRateGet( ) * 3, give, NULL); > return(OK); >} give() will be called from a watchdog, i.e. in the interrupt context. You can't use printf() in interrupt context: you have to be in a task context (it is in the printf() manual page). Hope this helps, Bergonz - --- Ing. Michele Bergonzoni - Laboratori Fondazione Guglielmo Marconi info and PGP key at http://www.dsnet.it/aziende_it/labs-main.html Phone:+39-51-846900 Fax:+39-51-846479 e-mail: bergonz@labs.it Telecommunications consultants, Microwave systems & circuits development. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Q:static c++ objects under vxWorks Date: 1 Jul 1996 18:10:57 GMT From: varlese@tech.ascom.ch (Chris Varlese) Organization: Ascom Hasler AG, Berne, Switzerland Message-ID: <4r94bh$pj9@ascomax.hasler.ascom.ch> References: <4r7vn2$46h@dub-news-svc-5.compuserve.com> In article <4r7vn2$46h@dub-news-svc-5.compuserve.com>, 100567.1652@compuserve.com (Rolf Grzibek) writes: >Is it possible to have static objects under vxWorks? >Stroustrup states that the constructors of static object are called >before main(). As there is no main() under vxWorks it seems to be >impossible. But what if the constructor has no side effects except >initializing datas members (_no_ malloc() or other function calls)? Yes, VxWorks can support static objects. These are called at system startup time, because, as you say, there is no main(). Look in your VxWorks documentation for the subject "Munching". This is the process of finding and listing all the static objects that are to be constructed during the startup procedure. regards, Chris Varlese --------------------------- Newsgroups: comp.os.vxworks Subject: Value returned by sizeof() Date: Wed, 03 Jul 1996 11:04:53 -0400 From: "Bruce K. Gitlin" Message-ID: <31DA8C15.7900@wl.corning.com> TARGET: VxWorks 5.3-PowerPC-beta KERNEL: WIND version 2.4 CPU: Motorola MVME1600 - PowerPC 604 HOST: Sun Workstation/Solaris OS Has anyone else seen this problem? The value returned by the 'sizeof' function seems to be int oriented. It returns the value in bytes, but always a multiple of 4. The following code should return a 29 but instead returns 32. We use 'sizeof' on the MVME167 and have never experienced this problem. /*********************************************/ #include "vxWorks.h" typedef struct { float fil1[2]; /* 8 bytes */ int fil2[2]; /* 8 Bytes */ short fil3[5]; /* 10 bytes */ char fil4[2]; /* 2 bytes */ char fil5[1]; /* 1 byte */ } twenty_nine; /* 29 bytes */ int testSizeof() { return (sizeof (twenty_nine)); } /********************************************/ Any help would be appreciated, Thanks, Bruce gitlin_bk@wl.corning.com (910)397-8204 --------------------------- Newsgroups: comp.os.vxworks Subject: HELP: FDDI driver problems, not cpu#0 Date: Wed, 03 Jul 1996 12:14:26 -0700 From: Randy MacLeod Organization: Lawrence Berkeley National Lab Message-ID: <31DAC692.4DD1@lbl.gov> Hi, I'm trying to get an Interphase FDDI board (5211) to work. I have the driver source and it works fine if the host cpu is cpu # 0 but this is not an acceptable system configuration. After modifying the kernel to map my host (call it the sender) memory to 0x2000000 in vme address space, the problem appears to be that I am not getting an interrupt that the FDDI board supposed to generate when the driver calls get_mac_addr(). A clue may be that if I power on the sender as cpu #0, do a bootChange to cpu # 6, and do a software reboot, then the driver works fine. There is something about the vxworks kernel and/or the VME bus that I am missing. Help! I'm running VxWorks 5.2 on a Motorolla 166/7 and I'm talking to a Sparc 20 on the other end of the fibre. Thanks, Randy MacLeod P.S. The company that wrote the driver is not interested in making it work the way I need it to work! --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Q:static c++ objects under vxWorks Date: Wed, 03 Jul 1996 11:08:33 -0700 From: Oisin Hurley Organization: IONA Technologies Message-ID: <31DAB721.2C67412E@iona.com> References: <4r7vn2$46h@dub-news-svc-5.compuserve.com> <31DA0855.5AAB@possibility.com> Todd Hoff wrote: > The singleton idiom can be helpful here. For objects that > have one instance, which covers most static objects :-), > you can make the object create itself on access via > static method. Other calls to the static method will just > return the already allocated object. This way the system > just kind of creates itself on use. > > For example: > > class NotReally > { > public: > static NotReally* Singleton() > { > check if the object has been > created. If not create it otherwise return it. > } > private: > static NotReally* mpSingleton; > }; > > Get access to the object: > > NotReally* not= NotReally::Singleton(); This is a Good Right and True way to go. Note however that it is important to ensure that singleton initialisation be kept thread-safe. There's a short paper on an idiom called `Double-Checked locking' at http://siesta.cs.wustl.edu/~schmidt/patterns-cpd.html which deals with this. O[-< - -- oh@iona.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: MIB-II System Group - NVRAM Date: Sun, 30 Jun 1996 10:32:56 +0300 From: Ron Keidar Organization: Qualcomm Message-ID: <31D62DA8.76A5@qualcomm.com> References: <4qstd2$2d3@news.inforamp.net> Kha Sin Teow wrote: > > We have a need to support the MIB-II System Group on a > non-volatile memory (possibly FLASH-EPROM). > That is, we want to allow our customer to set some system variables > (e.g. sysLocation) once our product is installed on their site. > Reasonable enough. > > We are using vxWorks 5.2. > The SNMP agent is based on the SNMP Research code. > > The WindRiver Tech Support told me that I need to build a file system > on the FLASH. > We are not very happy with this suggestion. I believe you really work too hard. Please look at src/config/usrNetwork.c in function usrMib2Init() there is a call to m2SysInt() which initates the system group objects. Just edit this call the way you want and you are done. Also note the following bug: /* * BUG: * if k_mib2_initialize() called after m2SysInit() * the sysObjectID is nulled (i.e. {0, 0}). * If it get called before m2SysInit() , then * the sysDescr, sysContact, and sysLocation are * from configAll.h, no metter whether * the snmpd.cnf file exists or not. * * Ron Keidar May 16, 1996. */ Hope it helped. Ron Keidar > I tried the following experiment but it did not work: > > o undefined INCLUDE_MIB2_ALL > o defined INCLUDE_MIB2_IF > o definde INCLUDE_MIB_CUSTOM > o generated method routine stubs (k and v files) for the System Group > and compiled the agent as if the System Group is my "custom" MIB. > o rebooted vxWorks with the agent extension; the initialization routine > (that I wrote) was invoked when the agent came up. > > However, when I tested the agent with the "getone" on any System variable, > I got "no such variable". > > This is a long shot, but in case anyone had come across this similar > problem and had managed to solve it, I would appreciate a pointer. > > Since I don't read the News that often, please email your response > to khasin@inforamp.net. > I'll post the response if the group is interested. > > BTW, is there a FAQ archive somewhere on vxWorks? > > Thanks. > > khasin@inforamp.net. > > June 26, 1996. --------------------------- Newsgroups: comp.os.vxworks Subject: tFtpdTask Date: 27 Jun 1996 18:30:23 GMT From: piwek@student.uci.agh.edu.pl (Bartlomiej Piwek) Organization: University of Mining & Metallurgy Message-ID: <4qujvv$1s5@galaxy.uci.agh.edu.pl> Hi, I'm starting with VxWorks 5.1 system and I've got a problem with built in ftp server. When I log in I can put any username and password and system logs me in. Anyone can tell me how can I prevent this ? Bartek Piwek P.S. I'm working with SUN host, and load system from it via ethernet connecion. --------------------------- Newsgroups: comp.os.vxworks Subject: How to setup a secure shell? Date: 2 Jul 1996 12:43:28 -0700 From: kla@leland.Stanford.EDU (Earl Mitchell) Organization: Stanford University, CA 94305, USA Message-ID: <4rbu50$pa@elaine25.Stanford.EDU> VxWorks allows you to setup rlogin and telnet to require passwords etc. But access to the shell is unrestricted as long as you have physical access to the serial port. There is nothing to stop somebody from hooking up a terminal to the port. We could remove the port or not spawn the shell but we want to keep the shell accessible so that field service people can use it for diagnostics/debugging. Has anybody dealt with this problem before? A possible solution is ... Spawn a suspended shell and a task that will request a login and password. The new task grabs input the shell normally would. After getting a valid login it then resumes the shell and suspends itself. When the user is done they type logout which resumes the task and suspends the shell. Another variation of this would be to have the shell always running but the new task runs at a higher priority. When new task is running it gets shell input, when suspended shell gets input. Problem with these approaches is that telnet and rlogin will probably not work anymore, they will hang since shell is suspended. I would have to change source for telnet and rlogin to make this work. If anybody has dealt with this before I'd appreciate any suggestions. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Value returned by sizeof() Date: Wed, 03 Jul 1996 16:04:59 -0400 From: Kevin Rice Organization: NASA Goddard Space Flight Center -- Greenbelt, Maryland USA Message-ID: <31DAD26B.5C43@vlsi.gsfc.nasa.gov> References: <31DA8C15.7900@wl.corning.com> Bruce K. Gitlin wrote: > The value returned by the 'sizeof' function seems to be int oriented. > It returns the value in bytes, but always a multiple of 4. The > following code should return a 29 but instead returns 32. We use > 'sizeof' on the MVME167 and have never experienced this problem. > > /*********************************************/ > #include "vxWorks.h" > > typedef struct > { > float fil1[2]; /* 8 bytes */ > int fil2[2]; /* 8 Bytes */ > short fil3[5]; /* 10 bytes */ > char fil4[2]; /* 2 bytes */ > char fil5[1]; /* 1 byte */ > } twenty_nine; /* 29 bytes */ > Compilers often pad structures to align the fields on some architecture dependent boundary -- say a 32 bit size word. If you have to have something that is absolutely 29 bytes make the struct byte based (ie. all the types char .) You'll be left with job of assembling into the 'underlying' types you want at some point if there are any. ie. struct { char fil1[8]; /* float fil1[2] */ char fil2[8]; /* int fil2[2] */ char fil3[10]; /* short fil3[5] */ etc... }; If you have control, don't be so stingy with memory and align things yourself. Kevin Rice --------------------------- Newsgroups: comp.os.vxworks Subject: Re: MVME1604-011 booting Date: Wed, 03 Jul 1996 09:40:42 -0700 From: Geoffrey Espin x2496 Organization: KLA Message-ID: <31DAA28A.33E8@serval.kla.com> References: <31D954B9.77DB@roke.co.uk> Alan, > I am using the Motorola MVME 1604-011 with Tornado for Windows NT. > I would like to keep the Motorola PPCBug ROM's (instead of the VxWorks > Boot ROM's) and use them to transfer the bootrom code via tftp. > I can get the tftp bit to work, but when I execute the bootrom > code it generates a illegal instruction error. Ah, the pain of it all. WRS stopped "supporting" PPCBUG after EAR4 (pre-5.2 beta). Unfortunately, they unnecessarily deleted some #ifdef's from romInit.s. There are other little gotchas too. But to start with do this: ... sync /* Synchronize for ICE enable */ mtspr HID0,r4 /* Enable Instr Cache & Inval cache */ mtspr HID0,r3 /* using 2 consec instructions */ #ifndef INCLUDE_MOT_BUG_ROM /* KLA */ /* This following contains the entry code for the initialization code for the MPC 105, a PCI Bridge/Memory Controller used in Motorola's MV1600 and Ultra series PowerPC based boards. */ ... #endif /* INCLUDE_MOT_BUG_ROM / KLA */ li r0, 0x2000 /* approx decimal 8000 */ mtctr r0 wait8ref: ... There should be other stuff you can salvage in config.h, and of course in the Makefile: ROM_TEXT_ADRS=00780100 # ROM entry address forINCLUDE_MOT_BUG_ROM (8M) ... boot.bin : bootrom_uncmp $(VX_BSP_BASE)/bin/$(VX_HOST_TYPE)/elfToBin < bootrom_uncmp > $@ I had trouble with getting the image in "shape" (useable by PPCBUG) so try just downloading the file and dumping memory from PPCBUG first. At least the original mv1604/README is still there (by accident). Hope this helps. Geoff - -- Geoffrey Espin espin@idiom.com espin@serval.kla.com KLA (408)468-2496 KLA-FAX (408)434-4273 --------------------------- End of New-News digest ********************** From aitech@rugged.com Thu Jul 4 07:44:49 1996 From: aitech@rugged.com (Aitech ) Date: Thu Jul 4 07:44:51 PDT 1996 Subject: Re: VME multiplexer card wanted Phil, You should take a look at the Aitech I/O cards. It seems that the Aitech C407 board can meet your requirement for a multiplexer to switch the source signal to a DSP analyzer, using on-board solid state relays. This same board also has additional digital I/O capability that will meet your need for 8 channels of digital I/O (0 to 48V). Secondly, the Aitech C403 board is a combination of analog I/O and digital I/O that meets the requirements you specified. Aitech manufactures and sells these boards in rugged as well as commercial versions. You can get additional information in the Aitech web site: http://www.rugged.com I hope this helps, Dani From 100656.1507@compuserve.com Thu Jul 4 08:16:23 1996 From: John Shaw <100656.1507@compuserve.com> Date: Thu Jul 4 08:16:24 PDT 1996 Subject: VxWorks 5.2 slave bootup problems Greetings, We have recently upgraded from VxWorks 5.1.1 to 5.2 and have been successfully booting a crate full of mv16x cards with VxWorks 5.1 PROMS and with VxWorks 5.1.1 across the network. The host is a Sun running Solaris 2.3. Now, after having upgraded to VxWorks 5.2, we have found that backplane network booting is unreliable. We have a mv167 master, a mv167 slave and four, other mv162 slaves in the chassis. We have not, however, upgraded the VxWorks PROMS to 5.2. The release notes that come with VxWorks 5.2 have two known bugs reported, these are:- SPR 2623 - When using a mv167 BP-based master, the BP anchor and memory pool must be allocated in a cache-safe way. As shipped with 5.1 and 5.11, it does not. This can prevent a slave from booting. SPR 4658 - All sorts of problems with the mv177, particularly with cache. We've tried a mv177 as master with all the workarounds applied and the situation is even worse! Has anyone encountered similar problems booting using VxWorks 5.2, inspite of SPR 2623 which seems to suggest that VxWorks 5.1.1 should be the more unreliable? Is Tornado VxWorks 5.3 better? From John Shaw, Shade (Computer Services) Limited, 1/3 Patford Street, Calne, Wiltshire, England, SN11 0EF, email: 100656.1507@compuserve.com From hiske@ict.nl Fri Jul 5 02:36:50 1996 From: hiske@ict.nl (Hiske Huiting) Date: Fri Jul 5 02:36:52 PDT 1996 Subject: How to get a printer interrupt from VxWorks? I have a Force-3CE SPARC board with VxWorks running and I want to trigger an application on this board via an external pulse. I thought it would be easy to use the parallel port for this, so I connected a wire to the paper-error -pin in the parallel-port bus. The parallel port is enabled through a switch. Questions: 1. How to configure VxWorks for having a parallel port? 2. Can I use the standard LPT driver ? 3. How to connect this paper-error pin to a certain interupt? I could not find this information in the technical references of the board (chip 89C100) or in the VxWorks programmers guide or in the bsp-manual pages. Thanks, Hiske ____________________________________________________________________________ name: H.Huiting / ___ \ ICT Automatisering B.V. / Aerospace internet: hiske@ict.nl / / \ \ P.O.Box 701 fax: 31-570-621362 ( ( ICT ) 7400 AS Deventer tel: 31-570-633888 \ \___/ / The Netherlands ____________________________\_______/______________________________:-)______ From 100067.355@compuserve.com Fri Jul 5 05:45:07 1996 From: Philippe Weber <100067.355@compuserve.com> Date: Fri Jul 5 05:45:09 PDT 1996 Subject: sizeof > From: "Bruce K. Gitlin" > Message-ID: <31DA8C15.7900@wl.corning.com> > TARGET: VxWorks 5.3-PowerPC-beta > KERNEL: WIND version 2.4 > CPU: Motorola MVME1600 - PowerPC 604 > HOST: Sun Workstation/Solaris OS > Has anyone else seen this problem? > The value returned by the 'sizeof' function seems to be int oriented. > It returns the value in bytes, but always a multiple of 4. The > following code should return a 29 but instead returns 32. We use > 'sizeof' on the MVME167 and have never experienced this problem. > /*********************************************/ > #include "vxWorks.h" > typedef struct > { > float fil1[2]; /* 8 bytes */ > int fil2[2]; /* 8 Bytes */ > short fil3[5]; /* 10 bytes */ > char fil4[2]; /* 2 bytes */ > char fil5[1]; /* 1 byte */ > } twenty_nine; /* 29 bytes */ > int testSizeof() > { > return (sizeof (twenty_nine)); > } > /********************************************/ > Any help would be appreciated, > Thanks, > Bruce > gitlin_bk@wl.corning.com > (910)397-8204 Bruce, I think this behavior is normal. 68000 processors support unaligned structures, which means the next object in memory could effectively start at struct + 29. This is not the case with PowerPC, where structures must be 32-bit aligned. In other words, sizeof() returns 32 because 32 is really size occupied by your structure. _________________________ | Philippe Weber | | Themis Computer | /) +33 76 59 60 46 (\ / ) Email:phil@themis.com ( \ _( (|_________________________) ) /> (((\ \) /,) / ) / //))/ (\\\\ \_/ / \ \_/ ///// \ / \ / \ _/ \_ / ----/ /----------------------\ \---- / / \ \ From ozkose@venus.aselsan.com.tr Fri Jul 5 07:46:47 1996 From: Ozgu OZKOSE ERDOGAN Date: Fri Jul 5 07:46:56 PDT 1996 Subject: working of Solais and vxWorks on the same bus Hello, I am working on a project where two FORCE CPU-5V boards are used on a VMEbus. One board is configured as bus controller and uses Solaris as OS, the other one is configured as not being a controller and uses vxWorks as OS. We have a problem on the VME access from the second board (that use vxWorks). The 5V that runs Solaris can access to the VME as A32D32 and A16D16 without any problem (I have only this slaves, so I have tested only these modes). Yet, when the 5V that runs vxWorks want to access the bus it hangs up in a state its RUN led starts to blink as red and yellow. I have used the 5V with vxWorks on another VME without any problem, where the bus controller was a PC from Radisys. I have also tested the board with bus analyser, it worked without any problem. Have you ever seen such a problem before, I need any help/information on this problem. Any help will be apriciated. OzguERODGAN, MscEE Software Engineering Dept./MST ASELSAN Inc. ozkose@venus.aselsan.com.tr company profile: www.aselsan.com.tr From lindenstruth@csa5.lbl.gov Fri Jul 5 08:45:00 1996 From: lindenstruth@csa5.lbl.gov Date: Fri Jul 5 08:45:02 PDT 1996 Subject: unsubscribe unsubscribe From trgvxw@startrg.lbl.gov Fri Jul 5 08:46:59 1996 From: STAR Trigger VxWorks group account Date: Fri Jul 5 08:47:01 PDT 1996 Subject: subscribe subscribe From mea@mclean.sparta.com Fri Jul 5 11:36:12 1996 From: "Michael E. Anderson" Date: Fri Jul 5 11:36:54 PDT 1996 Subject: Re: FDDI Driver Problems On Thu, 4 Jul 1996, Randy MacLeod wrote: > Subject: HELP: FDDI driver problems, not cpu#0 > Date: Wed, 03 Jul 1996 12:14:26 -0700 > From: Randy MacLeod > Organization: Lawrence Berkeley National Lab > Message-ID: <31DAC692.4DD1@lbl.gov> > > Hi, > > I'm trying to get an Interphase FDDI board (5211) to > work. I have the driver source and it works fine if the > host cpu is cpu # 0 but this is not an acceptable system > configuration. After modifying the kernel to map my > host (call it the sender) memory to 0x2000000 in vme address > space, the problem appears to be that I am not getting > an interrupt that the FDDI board supposed to generate > when the driver calls get_mac_addr(). > A clue may be that if I power on the sender as cpu #0, > do a bootChange to cpu # 6, and do a software reboot, > then the driver works fine. There is something about the > vxworks kernel and/or the VME bus that I am missing. Help! > > I'm running VxWorks 5.2 on a Motorolla 166/7 and I'm > talking to a Sparc 20 on the other end of the fibre. > > Thanks, > Randy MacLeod > > P.S. The company that wrote the driver is not interested in > making it work the way I need it to work! > Randy, My guess is that the 5211 is having difficulty finding the requisite shared memory segment the it needs to communicate with the host processor. Since I haven't heard from you directly, I can only assume that you didn't purchase SPARTA's 5211 driver ;-). The driver we wrote will allow you to run with on or offboard memory regardless of the processor number. However, to get this driver to function, I believe that you'll have to make several mods to your BSP. Most notably, in sysLib.c you'll have to hack sysProcNumSet to map memory to the VMEbus even if you processor number's not 0, then you may need to also modify sysLocalToBusAddr if you're using one of the VME address modifiers that is not typically supported by VxWorks. (SPARTA's driver can run in both BLT (AMs 0x3b and 0xb) as well as VME64 MBLT (AMs 0x3c and 0xc) none of which are supported in most BSPs). The issue is that the 5211 is looking for the shared memory segment foretold in the short I/O setup commands and simply isn't finding it. Therefore, no interrupts. I'll be out of town for the next couple of weeks, but if you need more info, contact Keith Buchanan (tkb@mcelan.sparta.com) for info. Regards, =============================================================================== __ Real-Time System Development, Integration, Training and Services //\\ // \\ Mike Anderson // /\ \\ Chief Engineer Voice : (703) 448-0210 ext. 235 // / \ \\ SPARTA, Inc. FAX : (703) 734-3323 // \ \\ 7926 Jones Branch Drive EMAIL : mea@mclean.sparta.com \\ \ // Suite 900 Web : http://www.mclean.sparta.com \\ \ / // McLean, VA 22102 \\ \/ // "Software development is like making \\ // a baby... You can't make a baby in one \\// month by impregnating nine women. -- "Pride in Performance" Some things just take time." =============================================================================== From royg@semantic.com Fri Jul 5 14:32:34 1996 From: Roy Gordon Date: Fri Jul 5 14:32:36 PDT 1996 Subject: Re: sizeof Almost all modern compilers align and fill out structures (by default) according to the size of the strictest data type, recursively. Here there are two equal strictest data types: float and int. You can see if your compiler supports an option not to do so. If this results in misaligned data types its also necessary that the underlying system be able to handle misaligned accesses without bellying up. Being able to do so can slow things down. -- roy > The value returned by the 'sizeof' function seems to be int oriented. > It returns the value in bytes, but always a multiple of 4. The > following code should return a 29 but instead returns 32. We use > 'sizeof' on the MVME167 and have never experienced this problem. > /*********************************************/ > #include "vxWorks.h" > typedef struct > { > float fil1[2]; /* 8 bytes */ > int fil2[2]; /* 8 Bytes */ > short fil3[5]; /* 10 bytes */ > char fil4[2]; /* 2 bytes */ > char fil5[1]; /* 1 byte */ > } twenty_nine; /* 29 bytes */ > int testSizeof() > { > return (sizeof (twenty_nine)); > } > /********************************************/ > Any help would be appreciated, > Thanks, > Bruce > gitlin_bk@wl.corning.com > (910)397-8204 From 100656.1507@compuserve.com Sat Jul 6 05:06:37 1996 From: John Shaw <100656.1507@compuserve.com> Date: Sat Jul 6 05:07:23 PDT 1996 Subject: How to get a printer interrupt from VxWorks? Hiske, Does the Force-3CE SPARC board have the PCCchip2 like the Motorola MVME-167 card does? If so, I can submit a portion of a VxWorks Centronics Printer port driver that might be useful to you. From chwe@force.de Mon Jul 8 01:17:44 1996 From: chwe@force.de (Christof Wehner) Date: Mon Jul 8 01:17:49 PDT 1996 Subject: Re: working of Solaris and vxWorks on the same bus Dear Ms. Erdogan, I suspect your problem lies in a problem of the original MMU libraries. To be more exact, please provide to me the output of sysBoardShow to find out more about your board configuration. This will tell you the BSP revision, which (I guess) will be lower than 1.0/6. In later revisions, there is an additional patch which will fix multiple problems of VxWorks for SPARC, including changes to the MMU libraries. If you happen to have a lower BSP revision, please contact our sales office to get an upgrade to the latest revision (7). Best regards i.A. Christof Wehner -- Christof Wehner ____ ____ ____ ____ ____ Technical Support Engineer / / / / / / / / Customer Support Department /__ / / /___/ / /__ FORCE COMPUTERS GmbH / / / / \ / / Prof.-Messerschmitt-Str. 1 / /___/ / \ /___/ /___ D-85579 Neubiberg/Muenchen email: chwe@force.de C O M P U T E R S G m b H http://www.forcecomputers.com/ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - phone +49-89-60814-240 fax +49-89-609 77 93 From daemon@csg.lbl.gov Tue Jul 9 04:00:25 1996 From: daemon@csg.lbl.gov Date: Tue Jul 9 04:04:49 PDT 1996 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Tue Jul 9 04:00:18 PDT 1996 Subject: Re: mvme1603 com3&4 Subject: Device driver for Force ISIO/2 Subject: Using Tornado with one serial port Subject: Re: HELP: FDDI driver problems, not cpu#0 Subject: RPC Subject: Q:static c++ objects under vxWorks Subject: Re: Supervisory ? Non-priviledge Subject: printf() caveat Subject: Promiscuious Mode ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: mvme1603 com3&4 Date: Wed, 03 Jul 1996 09:13:58 -0700 From: Geoffrey Espin x2496 Organization: KLA Message-ID: <31DA9C46.3B04@serval.kla.com> References: <199607030306.WAA17420@ftms.ftms.com> Rob Lepage wrote: > > I'm using VxWorks 5.3 on an MVME1603. I need access to two serial ports > Is there a driver available to use these serial ports? Is the z85C30 > close enough to the 85230 so that the driver in target/src/drv/serial > could be used or adapted? Ideally, I would like to have port 3 act as > com1/console, and port 4 as com2. Yep, that's exactly what we've done. I started with the old serial/z8530Serial.c shipped with the mv147 and made just a few minor changes (unfortuately all the tyCoXX have to be changed to ty34XX). Important constants to know (or at least work for me): #define TY34_INTNUM 9 #define SERIAL_A (char*)(CPU_PCI_ISA_IO_ADRS | 0x00000842) #define SERIAL_B (char*)(CPU_PCI_ISA_IO_ADRS | 0x00000840) #define BAUD_CLK_FREQ 10000000 /* 10MHz "SBC Installation & Use" pg. C-2 */ I then added some code to bootConfig.c and usrConfig.c to look at the bootflags and decide which serial port to do the ioGlobalStdSet() on, then on the other port I print a message saying "use other port". [Oh, for a reentrant shell (not Tornado's).] Geoff - -- Geoffrey Espin espin@idiom.com espin@serval.kla.com KLA (408)468-2496 KLA-FAX (408)434-4273 --------------------------- Newsgroups: comp.os.vxworks Subject: Device driver for Force ISIO/2 Date: Thu, 04 Jul 1996 16:21:56 +0100 From: Andy Armitage Organization: European Space Agency Message-ID: <31DBE194.7FCB@wd.estec.esa.nl> A couple of weeks ago I posted a request for a VxWorks device driver for the Force ISIO/2. Well, we got some very good help from Force, and wrote the driver ourselves. The source is available (as-is, no-warranty, no-support) from ftp://ftp.estec.esa.nl/pub/ws/wsd/avecs/isio2.tar.Z There's a short PostScript doc in there as well Andy WSD/Estec European Space Agency --------------------------- Newsgroups: comp.os.vxworks Subject: Using Tornado with one serial port Date: 28 Jun 1996 16:47:07 GMT From: rbr@jhk.com Organization: jhk & Associates Message-ID: <4r12ab$1i9m@mule2.mindspring.com> Reply-To: rbr@jhk.com I want to set up a Tornado development environment on a Windows 95 box using only one serial port to both boot the target and run Tornado. (So my wife can work at home and avoid the hassle of traffic during the Olympics here in Atlanta.) The Tornado manual describes a scenario using a single serial connection, but has VxWorks booting out of ROM and using the serial connection as a raw connection for Tornado. But we'd really prefer to boot the target from the development host instead. My idea is to use SLIP to both boot the target and run Tornado over the same network connection. Obviously, Tornado can run over a network - it shouldn't care whether the physical connection is serial or Ethernet, right? My wife has tried this on her Sun at work, but couldn't get it working. We haven't actually tried it on the Win95 box yet. Has anyone else tried this? Can it work? Thanks, Robin - ---------------------------------------------------------------------------- Robin Ristow, Senior Engineer jhk & Associates (An SAIC Company) --------------------------- Newsgroups: comp.os.vxworks Subject: Re: HELP: FDDI driver problems, not cpu#0 Date: Wed, 03 Jul 1996 20:49:58 -0500 From: gwinn@res.ray.com (Joe Gwinn) Organization: Raytheon Electronic Systems Message-ID: References: <31DAC692.4DD1@lbl.gov> I would guess that in the setup for the driver and/or kernel on both CPUs, the interrupt masks were not set correctly, so the interrupt is going to CPU#0, not CPU#6. On CPU#0, the interrupt is ignored as an unexpected interrupt. In article <31DAC692.4DD1@lbl.gov>, Randy MacLeod wrote: > Hi, > > I'm trying to get an Interphase FDDI board (5211) to > work. I have the driver source and it works fine if the > host cpu is cpu # 0 but this is not an acceptable system > configuration. After modifying the kernel to map my > host (call it the sender) memory to 0x2000000 in vme address > space, the problem appears to be that I am not getting > an interrupt that the FDDI board supposed to generate > when the driver calls get_mac_addr(). > A clue may be that if I power on the sender as cpu #0, > do a bootChange to cpu # 6, and do a software reboot, > then the driver works fine. There is something about the > vxworks kernel and/or the VME bus that I am missing. Help! > > I'm running VxWorks 5.2 on a Motorolla 166/7 and I'm > talking to a Sparc 20 on the other end of the fibre. > > Thanks, > Randy MacLeod > > P.S. The company that wrote the driver is not interested in > making it work the way I need it to work! --------------------------- Newsgroups: comp.os.vxworks Subject: RPC Date: Thu, 04 Jul 1996 16:43:00 GMT From: q3931234@bonsai.fernuni-hagen.de Organization: FernUni Hagen Message-ID: <4rgsa2$g2q@ilex.FernUni-Hagen.de> I tried to use an RPC from a SUN to a VxWorks 5.2 VME-bus system (162). I used the same code and rpcgen which I used on the Sun. But the generated files have a lot of includes which are not available for VxWorks. What do I have to do to run the printmessage-demo from the RPC-package? (Server=VxWorks, Client=Sun). Where can I find sample code. I use NFS. Is it necessary to install the RPC 4.0 package? The RPC-package is in the unsupported directory of VxWorks. Will it be supported in future? --------------------------- Newsgroups: comp.os.vxworks Subject: Q:static c++ objects under vxWorks Date: Mon, 01 Jul 1996 07:45:52 GMT From: 100567.1652@compuserve.com (Rolf Grzibek) Organization: Philips BTS Message-ID: <4r7vn2$46h@dub-news-svc-5.compuserve.com> Hi to all, just a simple question. Is it possible to have static objects under vxWorks? Stroustrup states that the constructors of static object are called before main(). As there is no main() under vxWorks it seems to be impossible. But what if the constructor has no side effects except initializing datas members (_no_ malloc() or other function calls)? Does g++ just generate initialized data? TIA Rolf Grzibek 100567.1652@compuserve.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Supervisory ? Non-priviledge Date: Mon, 8 Jul 1996 16:14:49 GMT From: breti@bit3.com (Bret Indrelee) Organization: Bit 3 Computer Corp Message-ID: References: <31D92F1D.2FB9@ict.nl> In article <31D92F1D.2FB9@ict.nl>, Hiske Huiting wrote: >I have a 2 questions about VME communication. >1. What is the difference between Supervisory and Non-priviledge > addressing modes? The address modifier used. The VMEbus specification doesn't do anything except define them. Many processors make a differentiation between user and supervisory mode, where there are priviledged registers and instructions that can only be executed while the processor is in supervisory mode. Some interfaces will use this information when generating the address modifier. This is strickly an implementation issue, the VMEbus specification does not require this. Most VMEbus cards don't pay any attention to the two modes. If they do, they will usually have a jumper that allows them to only accept supervisory accesses. >2. What is the difference between Master and Slave access? The bus Master initiates the transaction. The bus Slave responds to the transaction. Memory cards are examples of bus slaves. Accesses by processors are always bus masters. SBC cards usually also allow slave access to the memory on the processor card as well. - -Bret - -- Bret Indrelee breti@bit3.com #include --------------------------- Newsgroups: comp.os.vxworks Subject: printf() caveat Date: 29 Jun 1996 22:34:04 GMT From: lees@cts.com (Lee Slaughter) Organization: CTS Network Services (CTSNET), San Diego, CA Message-ID: <4r4b0s$m8n@newsfeed.cts.com> Reply-To: lees@nosc.mil Hello helpful realtime gurus, I am new to this. At a recent WRS training I made a note here in a routine and cannot remember why except that I was having strange problems until instructor suggested using logMsg. I'm wondering what possible evil resulted from using printf()? thanks... give (void) { semGive(semId); logMsg("Semgive...\n",0,0,0,0,0,0); /* do NOT use printf */ /* Set timer to go off - after 1st time... */ wdStart (myWatchDogId, sysClkRateGet( ) * 3, give, NULL); return(OK); } --------------------------- Newsgroups: comp.os.vxworks Subject: Promiscuious Mode Date: Mon, 08 Jul 1996 15:35:35 -0500 From: andy.kollegger@gsfc.nasa.gov (Andreas Benno Kollegger) Organization: NASA Goddard Space Flight Center -- Greenbelt, Maryland USA Message-ID: Does anyone have experience with putting the 82596 ethernet driver into promiscuious mode (accepts all ethernet frames regardless of destination address)? I see in if_ei.h where the ac_config structure has a field for setting the chip into promiscuious mode, but am unclear about how to set that value. My reading of the Intel manual on the 82596 suggests that the config values get set during initialization and then the chip is off and running. That would mean that I'd have to get the source for the driver and change it. Is that truly necessary? Any help is appreciated, Andy --------------------------- End of New-News digest ********************** From SJohnson@systran.com Tue Jul 9 05:30:40 1996 From: Steve Johnson Date: Tue Jul 9 05:30:42 PDT 1996 Subject: Tornado Compiler for PPC Has anyone ever tried using the GNU compiler ( that comes with Tornado) on any of the PPC BSPs previous to the release of Tornado ? We have a need to change from the DIAB compiler but are not yet ready to change over to Tornado. From MikeMcQuade@prometheus.connectsoft.net Tue Jul 9 07:56:55 1996 From: Mike McQuade Date: Tue Jul 9 07:56:57 PDT 1996 Subject: Re: Tornado Compiler for PPC At 05:30 AM 7/9/96 PDT, you wrote: >Submitted-by SJohnson@systran.com Tue Jul 9 05:30:40 1996 >Submitted-by: Steve Johnson > > >Has anyone ever tried using the GNU compiler ( that comes with Tornado) on >any of the PPC BSPs previous to the release of Tornado ? We have a need to >change from the DIAB compiler but are not yet ready to change over to >Tornado. > I haven't done much with GNU on 5.2 BSPs, In fact since we got Tornado, I dont want to use 5.2 anymore, it was like steping out of a Impala and getting into a Corvette. But I have taken the "old Impala" for a ride using code compiled with GNU, and it did work. (VxWorks 5.2 image was copiled under Diab, objects ran were compiled with GNU) I am using both Diab and GNU (Diab 3.7a latest release and GNU 2.7.2) I like Diab alot more, It consistantly makes smaller (10-15%) code that runs faster (ie: on a big FIR filter, it was twice as fast). I have had a couple of instances where code compiled under GNU wouldn't work, but the same source compiled under Diab worked fine, these were quirky things, that were as result of my coding :) but still, Its been very nice to have both compilers available. From sri1@gvl.esys.com Tue Jul 9 11:18:15 1996 From: Shawn Ingram Date: Tue Jul 9 11:18:21 PDT 1996 Subject: NVRAM Usage Does anyone here have any experience with reading/writing (from under software control) to NVRAM on the Dec Alpha AXPvme 160? I don't mean the software error logging capabilities either (at least I don't believe that is what I want to do). What I want to do is store off various strings and constant values from under software control. For example I would like to record the total elapsed time my system has been active. When the unit came back in for servicing or repairs I could then maintain my own MTBF. Also I would like to record failures (using records designed by me) and other system status messages. I got the impression somewhere that part of the NVRAM was dedicated for user definition (I don't remember where I got this notion). I've checked the manpages for: AXPvme_sysNvRamGet AXPvme_sysNvRamSet AXPvme_sysNvrForceInit AXPvme_sysNvrInitStats AXPvme_sysNvrRead AXPvme_sysNvrValidate Of these only the first two are present in my present kernel's symbol table (used lkup from shell). Do I need to rebuild my kernel to gain access to the other functions? Also, where can I find a memory map that shows how the overall NVRAM area is organized? Any help offered would be greatly appreciated. Thanks. /********************************************* /* * /* S.R.INGRAM sri1@esygvl.com * /* Phone: (903) 457-5255 * /* * /*********************************************/ From stan@rti.com Tue Jul 9 11:46:19 1996 From: Stan Schneider Date: Tue Jul 9 11:46:21 PDT 1996 Subject: Re: printf() caveat >> From: lees@cts.com (Lee Slaughter) >> I'm wondering what possible evil resulted from using printf()? >> wdStart (myWatchDogId, sysClkRateGet( ) * 3, give, NULL); Watchdogs run at interrupt priority; stdio (printf) doesn't work in ISRs. This is instant death. -- Stan From tkb@mclean.sparta.com Tue Jul 9 15:13:52 1996 From: Thomas Keith Buchanan Date: Tue Jul 9 15:13:54 PDT 1996 Subject: Re: NVRAM Usage > Does anyone here have any experience with reading/writing (from under > software control) to NVRAM on the Dec Alpha AXPvme 160? I don't mean the > software error logging capabilities either (at least I don't believe that > is what I want to do). What I want to do is store off various strings > and constant values from under software control. For example I would VxWorks does not normally contain routines for accessing NVRAM other than those designed to store and retrieve the boot parameter string. You should have the source to these routines in your development tree somewhere, in mine the file is vw/src/drv/mem/nvRam.c. > access to the other functions? Also, where can I find a memory map that > shows how the overall NVRAM area is organized? Any help offered would be > greatly appreciated. Thanks. You need to find out two things, the virtual address of your NVRAM and how much of it is left over for your purposes. Look at the following lines in the file vw/config/all/configAll.h: ... #define BOOT_LINE_OFFSET 0x700 ... ... /* The bootroms put the boot line at the following address */ #define BOOT_LINE_ADRS ((char *) (LOCAL_MEM_LOCAL_ADRS+BOOT_LINE_OFFSET)) #define BOOT_LINE_SIZE 255 /* use 255 bytes for bootline */ /* The boot line is stored in non-volatile RAM at the following offset */ #define NV_BOOT_OFFSET 0 /* store the boot line at start of NVRAM */ ... Find the value of LOCAL_MEM_LOCAL_ADRS in your BSP directory and look for definitions that override the constants listed above. Look in your hardware manual and double check the address and size of NVRAM. If you have more non-volatile memory than VxWorks uses, you can write some simple functions to get at it. Just make sure you don't trash the boot parameters... adios ______________________________________________________ | |\ |Thomas Keith Buchanan Voice -> 703 448 1683 x233| | |tkb@mclean.sparta.com FAX -> 703 734 3323 | | |_____________________________________________________| | | | | | /\\ SPARTA, Inc. | | | ///\\ 7926 Jones Branch Drive | | | /////\\ Suite 900 | | | ///// \\ McLean, VA 22102 | | | / \\\\\ \\ 703-448-0210 | | | \ \\\\\ // | | | \ ////// | | | \////// SPARTA | | | \//// Pride In Performance | | | \// -------------------- | | |_____________________________________________________| | \______________________________________________________\| From uucp@relay1.fnet.fr Tue Jul 9 23:18:33 1996 From: Eric Mathie (stagiaire Univ. Nancy) Date: Tue Jul 9 23:18:36 PDT 1996 Subject: data compression for VxWorks 5.2 I am trying to uncompress an a.out executable file format for VxWorks 5.2 on my target. To do this, I have used the gzip tool that I have compiled. But I have some problems now to uncompress and load my executable file on the target. Has anybody done this and can help me. Thank you in advance, +-------------------------------------------------------------------------+ Mr Eric MATHIE EUROPEAN GAS TURBINES EGT Bat 93, 3 avenue des 3 chenes e-mail : emathie@egt.fr 90018 Belfort phone : 84-55-65-69 FRANCE fax : 84-55-35-21 +-------------------------------------------------------------------------+ From daemon@csg.lbl.gov Wed Jul 10 04:00:45 1996 From: daemon@csg.lbl.gov Date: Wed Jul 10 04:00:48 PDT 1996 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Wed Jul 10 04:00:40 PDT 1996 Subject: security in VxWorks Subject: Motorola MVME147 VME card available Subject: logTask Subject: Re: security in VxWorks Subject: Re: VxWorks 5.2 slave bootup problems Subject: Re: Interfacing Device Driver to TCP/IP prot. stack Subject: Re: fork(), PVM for vxWorks Subject: Re: Value returned by sizeof() Subject: Vxworks and Orbix (CORBA) with Windows-NT as host-plattform Subject: Re: FDDI Driver Problems SOLVED. Subject: Re: Value returned by sizeof() Subject: Re: logTask ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: security in VxWorks Date: 9 Jul 1996 12:09:31 GMT From: piwek@student.uci.agh.edu.pl (Bartlomiej Piwek) Organization: University of Mining & Metallurgy Message-ID: <4rti5r$l9a@galaxy.uci.agh.edu.pl> Here is a question about security: Can you explain me how goes authentication of user in VxWorks ? Here is the situation: I boot VxWorks via network from a UNIX host. I put my login name (which I use on the UNIX host) at user field, and leave blank the password field (because I'm going to use rsh).Of course I create proper .rhosts in my home directory on the UNIX host. After booting I work on target BUT I access the files on the UNIX host.So how the UNIX system recognizes me?Does it check my .rhosts file each time I try to access filesystem? - -- Bartek Piwek Dept. of Automatic Control, UMM, Krakow, Poland --------------------------- Newsgroups: comp.os.vxworks Subject: Motorola MVME147 VME card available Date: 9 Jul 1996 01:52:52 GMT From: innov8@ix.netcom.com(MaRtY) Organization: Netcom Message-ID: <4rse1k$p2a@sjx-ixn6.ix.netcom.com> I have some Motorola MVME147SA-1 25Mhz 8MB VME cards and a MVME147SB-1 25Mhz 16MB VME card available. E-mail if interested. Guaranteed. - -- E-mail: innov8@ix.netcom.com --------------------------- Newsgroups: comp.os.vxworks Subject: logTask Date: 9 Jul 1996 11:49:42 GMT From: piwek@student.uci.agh.edu.pl (Bartlomiej Piwek) Organization: University of Mining & Metallurgy Message-ID: <4rth0m$l9a@galaxy.uci.agh.edu.pl> Hi, I have a problem with logTask. What I want it to do is log all messages (generated by logMsg()) into file - for example :logfile. What I do is: 1. create file: > flieDesc=creat("logfile",2) 2. add a logging file desctiptor: > logFdAdd(fileDesc) 3. send a message to be logged: > logMsg("this is log test\n",0,0,0,0,0,0) 4. check what is in the log file: > copy "logfile" what I get is far from my message (what is it ?pointer to the message?) 5. close file :> close(fileDesc) 6. check what is in the log file:> copy "logfile" what I get is my message + info about tid and taskname. What is strange for me is :when I close the logfile (by close(fileDesc)) logTask is not able to add new logMsgs to it. Is it correct ? My questions :1. How can I access the logfile whent it's open ? (I want to run some tasks which send logMsg's to the log file, and during this I'd like to check what's going on by looking into the logfile) 2. Is it possible for logTask to log logMsg's when logfile is closed.(As a a matter of fact it knows logfile descriptor so it shouldn't has any problem with opening and closing the log file) Thanks in advance, - -- Bartek Piwek Dept. of Automatic Control, UMM, Krakow, Poland --------------------------- Newsgroups: comp.os.vxworks Subject: Re: security in VxWorks Date: 9 Jul 1996 11:25:11 -0700 From: kla@leland.Stanford.EDU (Earl Mitchell) Organization: Stanford University, CA 94305, USA Message-ID: <4ru867$33p@elaine25.Stanford.EDU> References: <4rti5r$l9a@galaxy.uci.agh.edu.pl> In article <4rti5r$l9a@galaxy.uci.agh.edu.pl>, Bartlomiej Piwek wrote: >Here is a question about security: >Can you explain me how goes authentication of user in VxWorks ? > >Here is the situation: > >I boot VxWorks via network from a UNIX host. >I put my login name (which I use on >the UNIX host) at user field, and leave blank the password field >(because I'm going to use rsh).Of course I create proper .rhosts >in my home directory on the UNIX host. >After booting I work on target BUT I access the files on the UNIX >host.So how the UNIX system recognizes me?Does it check my .rhosts >file each time I try to access filesystem? Checkout remote file access section in your Programmer's Guide. VxWorks uses 3 different protocols for remote file access 1. rsh 2. ftp 3. nfs Rsh will copy entire requested file. Rsh uses .rhosts and /etc/hosts.equiv. NFS mounts the directories from the UNIX hosts like a client workstation would. It uses the /etc/exports file. The programmer's guide has a brief description how these 3 methods work and how to set them up. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks 5.2 slave bootup problems Date: 8 Jul 1996 12:00:44 GMT From: Alex Goodyear Organization: JET Joint Undertaking Message-ID: <4rqt9c$bn6@postman.jet.uk> References: <960704151514_100656.1507_BHL62-1@CompuServe.COM> Hi, I have just experienced the same problem with MV162/7 cards but with VxWorks V5.1.1. but I may be able to shed some light onto your problems. You specifically mention BP as the backplane driver used in your old config, if you are now using SM then you are probably experiencing caching problems. BP specifically disables caching but SM assumes that it can use "cache safe" memory. Try disabling all data caching for the master board, not in the start- up script (its too late by then) but in the configuration include files. I can only assume that in our case, bus snooping isn't working or I have configured it incorrectly, anyway brute force data cache disable works. Alex Goodyear. =============================================================================== The above article is the personal view of the poster and should not be considered as an official comment from the JET Joint Undertaking =============================================================================== --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Interfacing Device Driver to TCP/IP prot. stack Date: 9 Jul 1996 13:55:47 GMT From: Chet Kackman Organization: MJ7 Message-ID: <4rtod3$3sp@gde.GDEsystems.COM> References: <31E1FCC9.3C1B@oslo.geco-prakla.slb.com> Arne For that many host you may want to look into using UDP connectionless prot. and just to the decom on the board. If you are working in a trusted envion forget about reliable UDP. Good Luck Chet Kackman@GDEsystems.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: fork(), PVM for vxWorks Date: 9 Jul 1996 14:05:31 GMT From: johncoby@ghgcorp.com (JOHN R. COBARRUVIAS) Organization: Personal Account Message-ID: References: <4rsvph$5js@news.rwth-aachen.de> In article <4rsvph$5js@news.rwth-aachen.de>, Stefan Benz wrote: > Hallo, > 1. I'd like to port PVM to vxWorks. Does anybody knows something about such a > port ? > 2. PVM uses fork(). Is there any way to emulate fork() in vxWorks ? fork() see taskSpawn(), but it's not a true unix fork. > > Please pot your answers here and e-mail also to: benz@aps.rwth-aachen.de > > Thanx Stefan - -- ,-_/ . .-,--. ,---. . ' | ,-. |-. ,-. `|__/ | -' ,-. |-. ,-. ,-. ,-. . . . , . ,-. ,-. | | | | | | | | \ ,. | -. | | | | ,-| | | | | | / | ,-| `-. | `-' ' ' ' ' `-' ` `' `---' `-' `-' `-^ ' ' `-' `' ' `-^ `-' / | John R. Cobarruvias johncoby@ghgcorp.com Houston, Tx `--' --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Value returned by sizeof() Date: 8 Jul 1996 12:48:19 GMT From: johncoby@ghgcorp.com (JOHN R. COBARRUVIAS) Organization: Personal Account Message-ID: References: <31DA8C15.7900@wl.corning.com> In article <31DA8C15.7900@wl.corning.com>, "Bruce K. Gitlin" wrote: > TARGET: VxWorks 5.3-PowerPC-beta > KERNEL: WIND version 2.4 > CPU: Motorola MVME1600 - PowerPC 604 > > HOST: Sun Workstation/Solaris OS > > Has anyone else seen this problem? > > The value returned by the 'sizeof' function seems to be int oriented. > It returns the value in bytes, but always a multiple of 4. The > following code should return a 29 but instead returns 32. We use > 'sizeof' on the MVME167 and have never experienced this problem. Alos, if the PPC chip allows it, check the switch on the GNU compiler, if that is what you are using. It could be set to force fullword alignment. > > /*********************************************/ > #include "vxWorks.h" > > typedef struct > { > float fil1[2]; /* 8 bytes */ > int fil2[2]; /* 8 Bytes */ > short fil3[5]; /* 10 bytes */ > char fil4[2]; /* 2 bytes */ > char fil5[1]; /* 1 byte */ > } twenty_nine; /* 29 bytes */ > > int testSizeof() > { > return (sizeof (twenty_nine)); > } > /********************************************/ > > Any help would be appreciated, > Thanks, > > Bruce > gitlin_bk@wl.corning.com > (910)397-8204 - -- ,-_/ . .-,--. ,---. . ' | ,-. |-. ,-. `|__/ | -' ,-. |-. ,-. ,-. ,-. . . . , . ,-. ,-. | | | | | | | | \ ,. | -. | | | | ,-| | | | | | / | ,-| `-. | `-' ' ' ' ' `-' ` `' `---' `-' `-' `-^ ' ' `-' `' ' `-^ `-' / | John R. Cobarruvias johncoby@ghgcorp.com Houston, Tx `--' --------------------------- Newsgroups: comp.os.vxworks Subject: Vxworks and Orbix (CORBA) with Windows-NT as host-plattform Date: 9 Jul 1996 23:12:57 GMT From: Alexander Sabbah Organization: iwb.mw.tu-muenchen.de Message-ID: <4rup1p$9b2@galadriel.iwb.mw.tu-muenchen.de> hi, I'm looking forward to build an application on an RT-OS. Because of different technical features VxWorks 5.3 seems like the right thing choose. As development plattform (host) I want to use Windows-NT. In order to develope a truly distributed objekt-orintated application I'm looking forward to use Orbix by Iona Technologies which is based on the CORBA (Common Objekt Request Broker Architecture). So far this all is no big deal. Unluckily I had to find out that: a) there is no NT-Hosted Centerline to develope Orbix based applications for VxWorks (not yet). Iona only supports SUN an HP-Hosted Centerlines. b) the Orbix libs developed by Iona are for VxWorks 5.1.1 Now as far as I know, when compiling an IDL-file with the local IDL-Compiler the output is plattform independent. That is why there should not by any problems when using the normal IDL-Compiler for Windows-NT to compile IDL-Files for VxWorks on the host system Windows-NT. If that is true, then the major problem lies in the compatibility of the Orbix-libs with the new VxWorks release 5.3. Is there any bode who has made similer experiance trying to using Orbix with VxWorks? Maybe you could give me a clue how to handel: a) the non exsistance of a NT-Hosted Centerline to compile my IDL-Files for VxWorks on an NT-system b) the problem with the different versions (5.1.1 und 5.3) so long ....... Alexander Sabbah email: sb@iwb.mw.tu-muenchen.de --------------------------- Newsgroups: comp.os.vxworks Subject: Re: FDDI Driver Problems SOLVED. Date: 10 Jul 1996 04:02:18 GMT From: rmacleod@csg.lbl.gov (randy macleod) Organization: Lawrence Berkeley Laboratory, California Message-ID: <4rva0a$elb@overload.lbl.gov> References: I had not enabled the proper address translation for VMECHIP2. See ~vxworks/h/drv/vme/vmechip2.h for info. I modified the functions sysLocalToBusAdrs and sysProcNumSet in sysLib.c as suggested by the very helpful folks at AP Labs (http://www.sd.aplabs.com). Thanks also to everyone else who made suggestions. I'm getting > 2,400 kb/s transfer using the zbuf kernel and FDDI compared to ~820 kb/s using no zbuf and ethernet. :-) The mvme166 is the limiting component. Thanks to all, Randy MacLeod > I'm trying to get an Interphase FDDI board (5211) to > work. I have the driver source and it works fine if the > host cpu is cpu # 0 but this is not an acceptable system > configuration. After modifying the kernel to map my > host (call it the sender) memory to 0x2000000 in vme address > space, the problem appears to be that I am not getting > an interrupt that the FDDI board supposed to generate > when the driver calls get_mac_addr(). > A clue may be that if I power on the sender as cpu #0, > do a bootChange to cpu # 6, and do a software reboot, > then the driver works fine. There is something about the > vxworks kernel and/or the VME bus that I am missing. Help! > > I'm running VxWorks 5.2 on a Motorolla 166/7 and I'm > talking to a Sparc 20 on the other end of the fibre. > > Thanks, > Randy MacLeod - -- God is real, unless declared integer. -FORTRAN God does not exist, unless declared; God is often a complex structure. -C God is in the creator and destructor functions. -C++ --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Value returned by sizeof() Date: 8 Jul 1996 13:44:27 -0700 From: mwette@mr-ed.jpl.nasa.gov (Matt Wette) Organization: Jet Propulsion Laboratory Message-ID: <4rrrvb$3fu@mr-ed.jpl.nasa.gov> References: <31DA8C15.7900@wl.corning.com> In article <31DA8C15.7900@wl.corning.com>, Bruce K. Gitlin wrote: >TARGET: VxWorks 5.3-PowerPC-beta > KERNEL: WIND version 2.4 > CPU: Motorola MVME1600 - PowerPC 604 > >HOST: Sun Workstation/Solaris OS > >Has anyone else seen this problem? > >The value returned by the 'sizeof' function seems to be int oriented. >It returns the value in bytes, but always a multiple of 4. The >following code should return a 29 but instead returns 32. We use >'sizeof' on the MVME167 and have never experienced this problem. > >/*********************************************/ > #include "vxWorks.h" > > typedef struct > { > float fil1[2]; /* 8 bytes */ > int fil2[2]; /* 8 Bytes */ > short fil3[5]; /* 10 bytes */ > char fil4[2]; /* 2 bytes */ > char fil5[1]; /* 1 byte */ > } twenty_nine; /* 29 bytes */ > > int testSizeof() > { > return (sizeof (twenty_nine)); > } >/********************************************/ The ANSI definition of sizeof requires that, if the argument is a structure, padding be added for alignment in an array. Since your target is a 68k, I would guess that sizeof(twenty_nine) would be 30, not 29, since the structure must be aligned on a 2-byte boundary. On a sparc I would guess sizeof to return 32, since this structure must be aligned on a 4-byte boundary. Matt - -- matthew.r.wette@jpl.nasa.gov --------------------------- Newsgroups: comp.os.vxworks Subject: Re: logTask Date: 9 Jul 1996 11:56:18 -0700 From: kla@leland.Stanford.EDU (Earl Mitchell) Organization: Stanford University, CA 94305, USA Message-ID: <4rua0i$3fi@elaine25.Stanford.EDU> References: <4rth0m$l9a@galaxy.uci.agh.edu.pl> In article <4rth0m$l9a@galaxy.uci.agh.edu.pl>, Bartlomiej Piwek wrote: >Hi, > > I have a problem with logTask. What I want it to do is log all >messages (generated by logMsg()) into file - for example :logfile. >What I do is: >1. create file: > flieDesc=creat("logfile",2) >2. add a logging file desctiptor: > logFdAdd(fileDesc) >3. send a message to be logged: > logMsg("this is log test\n",0,0,0,0,0,0) >4. check what is in the log file: > copy "logfile" > what I get is far from my message (what is it ?pointer to the message?) >5. close file :> close(fileDesc) >6. check what is in the log file:> copy "logfile" > what I get is my message + info about tid and taskname. > >What is strange for me is :when I close the logfile (by close(fileDesc)) >logTask is not able to add new logMsgs to it. Is it correct ? > >My questions :1. How can I access the logfile whent it's open ? > (I want to run some tasks which send logMsg's to > the log file, and during this I'd like to check > what's going on by looking into the logfile) > 2. Is it possible for logTask to log logMsg's when > logfile is closed.(As a a matter of fact it knows > logfile descriptor so it shouldn't has any problem > with opening and closing the log file) > You didn't say where the file is actually located? I am assuming the file is on a remote host and the file is written from target to hosts via NFS. I setup something like this a while back where a task on the target generated a file on a remote host and a monitor task running on the host would concurrently read the file as it was being generated. The two problems you have to watch out for are buffering and synchronization. The target task which generated the file had to flush the I/O all the way out to the remote host otherwise the different application layers would buffer it (e.g. stdio and nfs buffers). There is an iotcl call you can use to flush the I/O can't remember what it is. Remember UNIX like systems use a lot of caching and buffering. Closing a file does an implicit flush which is why your monitor task saw the written data after the close. When I wrote my code I set it up so that the writing task ALWAYS left the file open so it could continually write to it. The reading task would periodically open the file, read it, and close it. The second problem is that the monitor task on the host needs to know when something new has been written. It could do a read and find nothing new was written, something new was partially written, or something new was completely written. Depending on how the reader is using this data it may not care, but in most cases it will have problems if it reads partially written data records. To solve this the writer can can tell the reader when it has written something new, and how much. I had the writer send a socket message to the reader. Reader is given enough info to seek to location of newly written info and read to the extent of the new info. Example .. Writer 0. num_writes = 0; records_written = 0 ; 1. Create file. 2. Write n records (or lines for your case). 3. Flush the I/O out to host file system using ioctl call. 4. num_writes++ ; records_written += 3 ; 5. Send message to reader (via sockets, files, etc) (e.g. for first pass message would say write #1 wrote 3 records). 6. Goto 2 and repeat. Reader 0. Waits for message from writer. 1. Gets message (e.g. for first pass message would say write #1 wrote 3 records). 2. Opens file. 3. Reads 3 records. 4. Closes file. 5. Goto 0 and repeat. --------------------------- End of New-News digest ********************** From Mike.Milde.at.SRC-Home_Office@scires.com Wed Jul 10 04:25:39 1996 From: Mike Milde at SRC-Home_Office Date: Wed Jul 10 04:25:41 PDT 1996 Subject: HTTP server for VxWorks? A while back there were several messages about an HTTP server for VxWorks. What was the result of all that? I now have a need for one. The smaller and simpler the better. It only needs to be able to display a simple web page and pass some Java applets. Any help is greatly appreciated! Thanks, Mike Milde Scientific Research Corporation mmilde@scires.com From tkb@mclean.sparta.com Wed Jul 10 04:32:40 1996 From: Thomas Keith Buchanan Date: Wed Jul 10 04:32:42 PDT 1996 Subject: Re: comp.os.vxworks newsdigest > After booting I work on target BUT I access the files on the UNIX > host.So how the UNIX system recognizes me?Does it check my .rhosts > file each time I try to access filesystem? Every time you access a file, the target logs on to your Unix system using .rhosts in you home directory or /etc/hosts.equiv. adios ______________________________________________________ | |\ |Thomas Keith Buchanan Voice -> 703 448 1683 x233| | |tkb@mclean.sparta.com FAX -> 703 734 3323 | | |_____________________________________________________| | | | | | /\\ SPARTA, Inc. | | | ///\\ 7926 Jones Branch Drive | | | /////\\ Suite 900 | | | ///// \\ McLean, VA 22102 | | | / \\\\\ \\ 703-448-0210 | | | \ \\\\\ // | | | \ ////// | | | \////// SPARTA | | | \//// Pride In Performance | | | \// -------------------- | | |_____________________________________________________| | \______________________________________________________\| From tkb@mclean.sparta.com Wed Jul 10 04:39:39 1996 From: Thomas Keith Buchanan Date: Wed Jul 10 04:39:41 PDT 1996 Subject: Re: comp.os.vxworks newsdigest t > My questions :1. How can I access the logfile whent it's open ? > (I want to run some tasks which send logMsg's to > the log file, and during this I'd like to check > what's going on by looking into the logfile) > 2. Is it possible for logTask to log logMsg's when > logfile is closed.(As a a matter of fact it knows > logfile descriptor so it shouldn't has any problem > with opening and closing the log file) Your problems are probably not with 'logMsg'. Are you using netDrv to access files? They do not write any of the file's contents until you close the file. Note that your log file could grow large enough to consume the system memory pool before you got around to closing it. If you need this log file, setup and use an NFS file system. Your data will get written 8k at a time or when you flush it. adios -- ______________________________________________________ | |\ |Thomas Keith Buchanan Voice -> 703 448 1683 x233| | |tkb@mclean.sparta.com FAX -> 703 734 3323 | | |_____________________________________________________| | | | | | /\\ SPARTA, Inc. | | | ///\\ 7926 Jones Branch Drive | | | /////\\ Suite 900 | | | ///// \\ McLean, VA 22102 | | | / \\\\\ \\ 703-448-0210 | | | \ \\\\\ // | | | \ ////// | | | \////// SPARTA | | | \//// Pride In Performance | | | \// -------------------- | | |_____________________________________________________| | \______________________________________________________\| From namiq@wrs.com Wed Jul 10 19:07:16 1996 From: namiq@wrs.com Date: Wed Jul 10 19:07:18 PDT 1996 Subject: Dialogic Board Support Hello Folk: I am looking for a Dialogic E-1 board. Is anyone aware of a VxWorks driver for this Dialogic E-1 board. If a E-1 driver is not availabl,e a VxWorks driver for the Dialogic T-1 Board would also be helpful. Thanks in advance, Best Regards, ---------------------------------------------------------------------------- - Namiq Kunimoto Email: namiq@wrs.com Director of Sales Phone: 81-3-5467-5900 Asia Pacific Operations FAX: 81-3-5467-5877 Wind River Systems WWW: http://www.wrs.com Japan Office | POLA Ebisu Bldg. 11F +-------------------------------------------- 3-9-19, Higashi, Shibuya-ku, Tokyo, 150, Japan ---------------------------------------------------------------------------- - From namiq@wrs.com Wed Jul 10 19:10:47 1996 From: namiq@wrs.com Date: Wed Jul 10 19:10:49 PDT 1996 Subject: Mil Spec Hard Disk Drive Hi Folks: Can anyone recommend a 4 Giga Byte Ruggedised or Mil Spec SCSI Hard Disk Drive (other than Miltope). Thanks in advance, Best Regards, ---------------------------------------------------------------------------- - Namiq Kunimoto Email: namiq@wrs.com Director of Sales Phone: 81-3-5467-5900 Asia Pacific Operations FAX: 81-3-5467-5877 Wind River Systems WWW: http://www.wrs.com Japan Office | POLA Ebisu Bldg. 11F +-------------------------------------------- 3-9-19, Higashi, Shibuya-ku, Tokyo, 150, Japan ---------------------------------------------------------------------------- - From Zohar.Agrest@ecitele.com Thu Jul 11 01:33:56 1996 From: zohar agrest Date: Thu Jul 11 01:33:58 PDT 1996 Subject: format converter a.out to ieee695/tekhex Hi, We are currently working on Windriver Tornado for mc68k360. We also have an EST bdm for mc68k360. We would like to have the option of debugging the Tornado based software with the Est tools. In order to do that we a file converter A.OUT -> IEEE695 OR TEKHEX. If anybody have this software or know where we can get it,it will be very helpfull. Thanks, Zohar Agrest e-mail:zohar@ecitele.com phone:972-3-9266208 fax:972-3-9266411 From daemon@csg.lbl.gov Thu Jul 11 04:00:33 1996 From: daemon@csg.lbl.gov Date: Thu Jul 11 04:00:36 PDT 1996 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Thu Jul 11 04:00:29 PDT 1996 Subject: Re: How to setup a secure shell? Subject: Symbolic data extraction <--> ecoff Subject: UDP/IP Broadcast in VxWorks 5.2 Subject: Re: long long support in VxWorks 5.2 Subject: Re: telling when in an ISR? Subject: Debugger other than VxGDB? Subject: Re: HTTP server for {VxWorks,QNX}? ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to setup a secure shell? Date: Tue, 9 Jul 1996 04:46:03 GMT From: Ron Kellam Organization: Hewlett-Packard Co. Message-ID: <31E1E40B.53EA@aus.hp.com> References: <4rbu50$pa@elaine25.Stanford.EDU> Sender: news@hparc12.aus.hp.com (News Daemon ID) Earl Mitchell wrote: > > VxWorks allows you to setup rlogin and telnet to require passwords etc. > But access to the shell is unrestricted as long as you have physical > access to the serial port. There is nothing to stop somebody from > hooking up a terminal to the port. We could remove the port or not spawn > the shell but we want to keep the shell accessible so that field > service people can use it for diagnostics/debugging. > > Has anybody dealt with this problem before? > If anybody has dealt with this before I'd appreciate any suggestions. In usrRoot() in bootConfig.c & usrConfig.c, stdin, stdout, & stderr get attached to the consoleFd (i.e. the serial port) by a ioGlobalStdSet(). If you do not execute this segment of code, then the serial port is not usable for io, but rlogin & telnet will still work fine. You may find it a good idea to have a bit somewhere in NVRAM which controls whether std(in/out/err) is to be attached to the console or not. You could toggle this bit via an rlogin/telnet - this way you can still make use of the serial port during development, and easily disable it for shipment. Hope this solves the dilema... Cheers, Ron Kellam (ron_kellam@aus.hp.com) Australian Telecom Operation Hewlett Packard Australia --------------------------- Newsgroups: comp.os.vxworks Subject: Symbolic data extraction <--> ecoff Date: Thu, 11 Jul 1996 00:04:39 GMT From: wellsk@netcom.com (Kevin Wells) Organization: NETCOM On-line Communication Services (408 261-4700 guest) Message-ID: Sender: wellsk@netcom16.netcom.com VxWorks on MIPS processors uses ECOFF formatted object files. When creating the vxWorks.st image, a program called makeSymTbl is called to create a (c compilable) file called symTbl.c. This file contains symbol names & addresses extracted from the (debug?) information to be used in the built-in symbol table. Does anyone know where to get the source code for makeSymTbl.c (for ecoff) and if it is publically available? Kevin Wells wells@aisf.com --------------------------- Newsgroups: comp.os.vxworks Subject: UDP/IP Broadcast in VxWorks 5.2 Date: 9 Jul 1996 17:09:16 GMT From: tfm@sei.cmu.edu (Theodore Marz) Organization: The Software Engineering Institute Message-ID: <4ru3ns$4l9@news.sei.cmu.edu> All, Forgive me if I am not totally up to speed on this. I seem to have encountered a problem with VxWorks 5.2 on two different board environments (Both Motorola, one a -167, and a PPC card). In both cases, we were trying to do UDP/IP Broadcast transmittions from the on-board ethernet. What seems to happen is that the board runs fine for a time, but eventually, datagrams cease to be issued by the board. VxWorks sendto() does not report an error (reports that the correct number of bytes are issued), but no packet on the net. If we perform directed network I/O to the effected board (telnet or ping), the IP stack seems to wake up and datagrams are again issued. Conditions causing this problem seem in some way to be related to timed on an SGI. Both the SEI and McDonnell Douglas Training Systems have seen the problem and reported it to Wind River, with no help forthcoming. 1) Has anyone else seen this sort of problem? 2) If so, what did you do about it.... Thanks in advance Ted Marz tfm@sei.cmu.edu. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: long long support in VxWorks 5.2 Date: Tue, 09 Jul 1996 17:30:00 -0700 From: Duane C Austin Organization: A customer of Pacific Bell Internet Services Message-ID: <31E2F988.696E@pacbell.net> References: <4s1aqj$fnf@fountain.mindlink.net> Philip J. Wiebe wrote: > > GNU 2.2.3 specifies that it supports double-word integers (i.e. long > long and unsigned long long); however, bit shifting operations cause > the VxWorks 5.2 loader to come up with an undefined symbol. For > example, the code fragment > long long a, b; > ... > a = b << 2; > .. > results in an undefined symbol of ___lshldi3. > > Is there any additional libraries that must be linked in, or is there > no support for these types of operations on double words? > > Phil. > > ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~> Philip J. Wiebe "It has yet to be proven that intelligence > Software Engineer has any survival value." > Newnes Machine Ltd. - Arthur C. Clarke > P.O.Box 8, 3550-45th Street S.E. > Salmon Arm, BC Canada > email: Philip.Wiebe@newnes.com > ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~You need to link with some other libs... Search the various libs under gnuYOUR_TARGET for your undefined symbol. I think most of the long long functions all occur in software so... hope your not looking for speed. hope this helps If you cant find them let me know and I'll look them up for you dcaustin@pacbell.net --------------------------- Newsgroups: comp.os.vxworks Subject: Re: telling when in an ISR? Date: 9 Jul 1996 11:05:59 -0700 From: kla@leland.Stanford.EDU (Earl Mitchell) Organization: Stanford University, CA 94305, USA Message-ID: <4ru727$2sb@elaine25.Stanford.EDU> References: <31DEB306.151@possibility.com> In article <31DEB306.151@possibility.com>, Todd Hoff wrote: >It would be nice to know if code is executing in an ISR so >the same code may execute in an ISR context or in task >context. What we have decided to do is to increment >a global ISR counter when entering an ISR and decrement >the counter when leaving an ISR. Code can check the counter. >If the counter is positive then a function can choose not to >use a semaphore for mutual exclusion for example. Logging >functions can allocate from a special memory pool when >in an ISR, etc. etc. > >Is this a good strategy? Are there better ways? Use the routine intContext() instead. It returns TRUE if current execution state is in interrupt context and FALSE if in task context. --------------------------- Newsgroups: comp.os.vxworks Subject: Debugger other than VxGDB? Date: 8 Jul 1996 19:35:51 GMT From: jpollock@bnr.ca (Jason Pollock) Organization: Northern Telecom Canada Ltd. Message-ID: <4rrnun$4qp@bcrkh13.bnr.ca> I was wondering if it is possible to use a debugger other than VxGDB on software running under VxSIM on SunOS/Solaris. GDB is a good source level debugger, but I am hoping to get a higher level of access to C++ code, tracking object interactions and such. Any help would be appreciated, Jason Pollock jason.pollock.jpollock@nt.com --------------------------- Newsgroups: comp.os.vxworks,comp.os.qnx Subject: Re: HTTP server for {VxWorks,QNX}? Date: Thu, 11 Jul 1996 01:15:29 -0700 From: jbs@quiotix.com (Jeffrey B. Siegal) Organization: Quiotix Corporation Message-ID: References: In article <4rt3ng$8pe@ccnews.nctu.edu.tw>, edward@CCCA.NCTU.edu.tw (Feng-ming Guo) wrote: > Hi, > Does anyone know any WWW httpd under QNX, or how to complie > one http daemon under QNX?? > Thanks in advance !! > Sincerely, > Edward. In article , Mike Milde at SRC-Home_Office wrote: > A while back there were several messages about an HTTP server for > VxWorks. What was the result of all that? I now have a need for > one. The smaller and simpler the better. It only needs to be able to > display a simple web page and pass some Java applets. We have developed an embedded web (HTTP) server available for evaluation and commerical licensing, designed for environments such as VxWorks and QNX. For more information, send me email. Jeffrey Siegal Quiotix --------------------------- End of New-News digest ********************** From jorgee@bpo.hp.com Thu Jul 11 06:44:15 1996 From: Jorge Estevez Date: Thu Jul 11 06:44:17 PDT 1996 Subject: Which was running task before interrupt? Hi, I would like to know at ISR time which was the running task just before the interrupt was raised. Has anyone any suggestion? Thank you From uzgeit@zellweger.ch Thu Jul 11 07:39:10 1996 From: Geiter Paul Date: Thu Jul 11 07:39:12 PDT 1996 Subject: GNU Warning: trigraph Hi I am using the Tornado compiler on a NT Host and get the following warning message: mmi_menu.h:1 warning: 1 trigraph(s) encountered I did't find any strange characters in line 1 of the header file. Can anyone give me a hint, where to look for an error. Paul Geiter Zellweger Uster uzgeit@zellweger.ch From greg.brissey@nmr.varian.com Thu Jul 11 10:04:15 1996 From: greg.brissey@nmr.varian.com (Greg Brissey x6951) Date: Thu Jul 11 10:04:16 PDT 1996 Subject: Re: long long math library > > Philip J. Wiebe wrote: > > > > GNU 2.2.3 specifies that it supports double-word integers (i.e. long > > long and unsigned long long); however, bit shifting operations cause > > the VxWorks 5.2 loader to come up with an undefined symbol. For > > example, the code fragment > > long long a, b; > > ... > > a = b << 2; > > .. > > results in an undefined symbol of ___lshldi3. > > > > Is there any additional libraries that must be linked in, or is there > > no support for these types of operations on double words? > > > > Phil. > > > > ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~> Philip J. Wiebe "It has yet to be proven that intelligence > > Software Engineer has any survival value." > > Newnes Machine Ltd. - Arthur C. Clarke > > P.O.Box 8, 3550-45th Street S.E. > > Salmon Arm, BC Canada > > email: Philip.Wiebe@newnes.com > > ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~You need to link with some other libs... Search the various libs under > gnuYOUR_TARGET for your undefined symbol. I think most of the long long > functions all occur in software so... hope your not looking for speed. > > hope this helps > > If you cant find them let me know and I'll look them up for you > > dcaustin@pacbell.net Under 5.1.1 for mv162 you will find the library at /"where-you-loaded-wv"/gnu/sun4.68k/lib/gcc-lib/m68k-wrs-vxworks/cygnus-2.2.3.1 libgcc.a You path may vary some depending on the arch. We just extracted the routines we needed, so we didn't need to load the entire library. Greg Varian Palo Alto, CA greg.brissey@nmr.varin.com From bff@dspcon.com Thu Jul 11 10:19:12 1996 From: bff@dspcon.com (Robert F. Franz) Date: Thu Jul 11 10:19:13 PDT 1996 Subject: VxWorks Shared memory IP I would like to know if any body could tell me the format for the information for the vxWorks shared memory IP. I am trying to write my own IP for a differnt OS which will communcate to vxWorks. ------------------------------------------------------------------------------- DSPCon, Inc. -- leaders in digital signal processing solutions ------------------------------------------------------------------------------- Bob Franz 380 FootHill Rd. Bridgewater, NJ 08807 (908)722-5656 Fax:(908)722-3259 bff@dspcon.com ------------------------------------------------------------------------------- From stan@rti.com Thu Jul 11 14:10:16 1996 From: Stan Schneider Date: Thu Jul 11 14:10:17 PDT 1996 Subject: Re: Which was running task before interrupt? >> Submitted-by jorgee@bpo.hp.com Thu Jul 11 06:44:15 1996 >> I would like to know at ISR time which was the running >> task just before the interrupt was raised. Has anyone >> any suggestion? >> In order to properly profile multiple tasks simultaneously, ScopeProfile maintains a task-level profile along with the function breakdown. The code is a little architecture-dependent, but this works on a 68k: if (kernelIsIdle) { ss->tcb = (WIND_TCB *) PROFILE_IDLE_TID; } else if (kernelState || (taskIdCurrent == NULL)) { ss->tcb = (WIND_TCB *) PROFILE_KERNEL_TID; } else { ss->tcb = taskIdCurrent; } HTH, -- Stan ============================================================================= = = = = Stan Schneider = email: stan@rti.com = = Real-Time Innovations, Inc. = Phone: (408) 720-8312 x104 = = 155A Moffett Park Drive, Suite 111 = Fax: (408) 734-5009 = = Sunnyvale, CA 94089 = http://www.rti.com = = = = ============================================================================= From student@vs-ulm.dasa.de Fri Jul 12 01:50:27 1996 From: Sparky Date: Fri Jul 12 01:50:29 PDT 1996 Subject: Ftp for example code FTP site. Is there one as I have to usergroup access. Cheers _______________________________________________________________________________ |\/\/\/| | | | | | (o)(o) C _) | ,___| | / /____\ / \ Sparky Jonathan D Heath VS2E22 Daimler-Benz Aerospace AG Woerthstrasse 85 89077 Ulm Germany HTTP://www.uni-ulm.de/~s_sthiam/heath/index.html E-mail: student@vs-ulm.dasa.de _______________________________________________________________________________ From student@vs-ulm.dasa.de Fri Jul 12 02:10:09 1996 From: Sparky Date: Fri Jul 12 02:10:11 PDT 1996 Subject: Program to test serial ports I am looking for a program to test serial ports on an Eltec 17 68040 Target board. I only have e-mail. i.e. no access to usergroups via a newsreader. Also looking for a program to show two tasks running randomized number generators, one with higher priority than the other. _______________________________________________________________________________ |\/\/\/| | | | | | (o)(o) C _) | ,___| | / /____\ / \ Sparky Jonathan D Heath VS2E22 Daimler-Benz Aerospace AG Woerthstrasse 85 89077 Ulm Germany HTTP://www.uni-ulm.de/~s_sthiam/heath/index.html E-mail: student@vs-ulm.dasa.de _______________________________________________________________________________ From tiziana@vega.syswiz.it Fri Jul 12 02:20:44 1996 From: Tiziana Balegno Date: Fri Jul 12 02:20:45 PDT 1996 Subject: Multicast addresses I am using VxWorks with the following system configuration: HOST Win95 ARCH MC68020 BSP FRC30 Using the Am7990 Lance Ethernet driver, I need to receive packets from the following two different multicast addresses 0x09, 0x00, 0x2b, 0x00, 0x00, 0x04, 0xfe 0x09, 0x00, 0x2b, 0x00, 0x00, 0x05, 0xfe according to the International Standard ISO 9542 (End System to Intermediate System routeing protocol ...). How can I set the two multicast address? Thank you very much in advance Regards From daemon@csg.lbl.gov Fri Jul 12 04:00:25 1996 From: daemon@csg.lbl.gov Date: Fri Jul 12 04:00:31 PDT 1996 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Fri Jul 12 04:00:22 PDT 1996 Subject: Re: NFS version 3 Subject: fork(), PVM for vxWorks Subject: NFS version 3 Subject: Re: Debugger other than VxGDB? ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: NFS version 3 Date: 11 Jul 1996 13:21:40 GMT From: Chet Kackman Organization: MJ7 Message-ID: <4s2v54$lgn@gde.GDEsystems.COM> References: <4s0u4m$edv@nrchh52.rich.nt.com> I believe that NFS comes bundled with VxWorks 5.2. At least our package from WRS included NFS. And it also works to! Chet kackman@gdesystems.com --------------------------- Newsgroups: comp.os.vxworks Subject: fork(), PVM for vxWorks Date: 9 Jul 1996 06:55:45 GMT From: Stefan Benz Organization: RWTH -Aachen / Rechnerbetrieb Informatik Message-ID: <4rsvph$5js@news.rwth-aachen.de> Hallo, 1. I'd like to port PVM to vxWorks. Does anybody knows something about such a port ? 2. PVM uses fork(). Is there any way to emulate fork() in vxWorks ? Please pot your answers here and e-mail also to: benz@aps.rwth-aachen.de Thanx Stefan --------------------------- Newsgroups: comp.os.vxworks Subject: NFS version 3 Date: 10 Jul 1996 18:52:06 GMT From: ashrafis@nt.com (Siamak Razzaghe-Ashrafi) Organization: Northern Telecom Inc., Mission Park Message-ID: <4s0u4m$edv@nrchh52.rich.nt.com> Is NFS Version 3 "Free ware" and if so were can I get it. I need to port it to VxWorks 5.2. Thankx for all the help. ashrafis@nt.com - -- The thing is it so I believe it is the thing. -SubZero --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Debugger other than VxGDB? Date: 9 Jul 1996 11:11:00 -0700 From: kla@leland.Stanford.EDU (Earl Mitchell) Organization: Stanford University, CA 94305, USA Message-ID: <4ru7bk$2vh@elaine25.Stanford.EDU> References: <4rrnun$4qp@bcrkh13.bnr.ca> In article <4rrnun$4qp@bcrkh13.bnr.ca>, Jason Pollock wrote: > >I was wondering if it is possible to use a debugger other than >VxGDB on software running under VxSIM on SunOS/Solaris. GDB is >a good source level debugger, but I am hoping to get a higher >level of access to C++ code, tracking object interactions and such. I think ObjectCenter from Centerline software supports VxWorks but I don't know for sure. I've heard that they simply kick in gdb when connecting to VxWorks though. This may no longer be true. Don't know if they support VxSim either. --------------------------- End of New-News digest ********************** From dsoares@ccmail.crosscomm.com Fri Jul 12 05:47:38 1996 From: David Soares Date: Fri Jul 12 05:47:39 PDT 1996 Subject: Re:Which was running task before interrupt? jorgee@bpo.hp.com writes: >I would like to know at ISR time which was the running task just >before the interrupt was raised. An easy way to do this is to use taskHookLib to catch the current running task. When your system initializes, call the routine taskSwitchHookAdd() (make sure taskHookInit()is called first) with the routine you would like to have called every task context switch. This routine will be passed the previous task's TCB and the new task's TCB. All you need to do is save these in a global. When the ISR executes, just read the global. One thing I don't know how to do is to get the task's ID from the TCB, other than to use the name from the TCB to call taskNameToId(), which is very cumbersome. Regards, Dave Soares CrossComm Corporation dsoares@crosscomm.com From xsd@mee.hwm.com Fri Jul 12 06:13:12 1996 From: xsd@mee.hwm.com (Uwe Scheffold) Date: Fri Jul 12 06:13:14 PDT 1996 Subject: Re: GNU Warning: trigraph > > I am using the Tornado compiler on a NT Host and get the following warning > message: > > mmi_menu.h:1 warning: 1 trigraph(s) encountered > > I did't find any strange characters in line 1 of the header file. > > Can anyone give me a hint, where to look for an error. > > > Paul Geiter > Zellweger Uster > uzgeit@zellweger.ch We had the same problem, and we found the solution, it was caused by the following character sequences in C-comment: ??) ??( etc. these are so called ANSI-C-trigraph's. Best regards: Uwe Scheffold~. .~ From faill@eng.insol.com Fri Jul 12 06:38:12 1996 From: faill@eng.insol.com (Peter Faill) Date: Fri Jul 12 06:38:13 PDT 1996 Subject: Re: GNU Warning: trigraph > I did't find any strange characters in line 1 of the header file. > Can anyone give me a hint, where to look for an error. Most likely a ? character in a comment field. From mangane@tsf.laas.fr Fri Jul 12 07:39:53 1996 From: Laurent Mangane Date: Fri Jul 12 07:39:55 PDT 1996 Subject: New implementation of squeeze() Hello, I am looking for the re-implementation of the squeeze() of usrLib(1) but for DOS file system instead of RT-11. Any idea on where I could find it (either buy or reuse a PD SW) ? Thanks in advance for any hint. Laurent Mangane LAAS/LIS, Toulouse, France From wpitts@rational.com Fri Jul 12 08:50:41 1996 From: "Wayne Pitts" Date: Fri Jul 12 08:50:44 PDT 1996 Subject: Re: Mil Spec Hard Disk Drive > Hi Folks: > Can anyone recommend a 4 Giga Byte Ruggedised or Mil Spec SCSI Hard Disk > Drive (other than Miltope). > Namiq Kunimoto Email: namiq@wrs.com > Director of Sales Phone: 81-3-5467-5900 > Asia Pacific Operations FAX: 81-3-5467-5877 > Wind River Systems WWW: http://www.wrs.com > Japan Office | > POLA Ebisu Bldg. 11F +-------------------------------------------- > 3-9-19, Higashi, Shibuya-ku, Tokyo, 150, Japan Check out the July '96 issue of Military & Aerospace Electronics, URL: http://www.pennwell.com/aero.html for an article on Mil and Rugged Mass Storage. Two I have used are from Lockheed Martin Federal Systems (maliat@lfs.loral.com) and Raymond Engineering (jkupa-ray1@kaman.com) Wayne -- Wayne Pitts, Rational Software Corporation url; http://www.rational.com Desk, voice mail, & pager: 408-496-3946 Cell Phone: 408-309-4182 Alternate pager number: 800-759-7243 pin # 921-8060 From carl@themis.com Fri Jul 12 09:27:16 1996 From: Carl C. Chesbrough Date: Fri Jul 12 09:27:18 PDT 1996 Subject: Re: GNU Warning: trigraph I seem to remember this from my days at Wind River. Try looking for something with three question marks (???) in it. I added the following to a my sysLib.c file and got the trigraph warning along with some other errors: int test??? = 234; Hope that helps, -Carl. > I am using the Tornado compiler on a NT Host and get the following warning > message: > > mmi_menu.h:1 warning: 1 trigraph(s) encountered > > I did't find any strange characters in line 1 of the header file. > > Can anyone give me a hint, where to look for an error. > > > Paul Geiter > Zellweger Uster > uzgeit@zellweger.ch ******************************************************************** * * * Carl C. Chesbrough Telephone: (510) 252-0870 x 126 * * Themis Computer Facsimile: (510) 490-5529 * * 3185 Laurelview Court E-Mail : carl@themis.com * * Fremont, California 94538 * * * ******************************************************************** From daemon@csg.lbl.gov Sat Jul 13 04:00:50 1996 From: daemon@csg.lbl.gov Date: Sat Jul 13 04:00:52 PDT 1996 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sat Jul 13 04:00:47 PDT 1996 Subject: Interfacing Device Driver to TCP/IP prot. stack Subject: Re: telling when in an ISR? Subject: Can't get munching to work... Subject: VXSIM - Round Robin Scheduling Subject: long long support in VxWorks 5.2 ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Interfacing Device Driver to TCP/IP prot. stack Date: Tue, 09 Jul 1996 08:31:37 +0200 From: Arne Saeten Organization: Geco-Prakla Message-ID: <31E1FCC9.3C1B@oslo.geco-prakla.slb.com> Hei, I have to put some numbers on the approx. time it will take to interface the TCP/IP prot. stack to a propritery driver and network interface card. Also, is TCP/IP feasible for a acquisition system of 100 nodes, where each node contains a i960, VxWorks, 2MB SRAM, etc.?? Are there any particular problems I have to be aware of by using TCP/IP?? I would appriciate any comments/help. rgds, Arne - -- Arne Saeten email: saetenar@oslo.Geco-Prakla.slb.com Geco-Prakla - SAPD phone: +47 - 67 57 55 00 (switchboard) Kjorbokollen phone: +47 - 67 57 56 80 (direct) N-1300 Sandvika - NORWAY fax: +47 - 67 57 55 11 - ----------------------------------------------------------- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: telling when in an ISR? Date: Wed, 10 Jul 1996 22:32:50 -0500 From: vaccaro@mail.edgenet.net (Mike Vaccaro) Organization: Naval Undersea Warfare Center Message-ID: References: <31DEB306.151@possibility.com> <4rtrtn$3ms@hppadbk.waterloo.hp.com> In article <4rtrtn$3ms@hppadbk.waterloo.hp.com>, bowman@waterloo.hp.com (Don Bowman) wrote: > Todd Hoff (tmh@possibility.com) wrote: > : It would be nice to know if code is executing in an ISR so > : the same code may execute in an ISR context or in task > : context. What we have decided to do is to increment > : a global ISR counter when entering an ISR and decrement > : the counter when leaving an ISR. Code can check the counter. > : If the counter is positive then a function can choose not to > : use a semaphore for mutual exclusion for example. Logging > : functions can allocate from a special memory pool when > : in an ISR, etc. etc. > > : Is this a good strategy? Are there better ways? > > VxWorks provides intContext() for this. > > -- > bowman@waterloo.hp.com ** 519-883-3019 ** http://www.waterloo.hp.com/~bowman I have used the technique of writing a string to global memory. A low priority process reads the string, prints it and clears the memory. It's crude, but it works for unit testing ISRs for interrupts which do not happen very frequently. - -- - -Mike V --------------------------- Newsgroups: comp.os.vxworks Subject: Can't get munching to work... Date: 12 Jul 1996 13:56:40 GMT From: markham@bnr.ca (Andrew Markham) Organization: Bell Northern Research Message-ID: <4s5lio$5g6@nrtphba6.bnr.ca> Here's the code that I compile: foo.cc - ------ #include class bobo { public: // Constructor sets value of i bobo ( void ) { i = 9; } // Member function that prints value of i void printI ( void ) { cout << "i = " << i << endl; } private: int i; }; bobo myBOBO; void testfunc ( void ) { theBOBO.printI(); } - ------ Here are the compilation steps (I split up lines to make it under 80 chars): > c++hppa -c -DMTP_ENABLE_ANSI=1 -DVXWORKS=1 -DUSE_360=1 -fno-builtin -nostdinc -g -I/central/dev/vxsim5.3/target/h -I/central/dev/vxsim5.3/target/config/all -DSIMHPPACPU -DVXWORKSOSYS -DCPU=SIMHPPA -DDEBUG -o foo.o foo.cc > ld_09.74.08 -N -r -o foo_out.o.tmp foo.o > nm foo_out.o.tmp | munch > munch.c > c++hppa -c -DMTP_ENABLE_ANSI=1 -DVXWORKS=1 -DUSE_360=1 -fno-builtin -nostdinc -g -I/central/dev/vxsim5.3/target/h -I/central/dev/vxsim5.3/target/config/all -DSIMHPPACPU -DVXWORKSOSYS -DCPU=SIMHPPA -DDEBUG -traditional -fno-builtin -O -o munch.o munch.c > ld_09.74.08 -N -r -o foo_out.o munch.o foo_out.o.tmp Then I go to WindSh and do the following: - -> ld < foo_out.o - -> cplusCtors("foo_out.o") value = 0 = 0x0 - -> testfunc__Fv value = 1036320 = 0xfd020 = cout The value that shows up on the console is 0, when it should be 9. Can anyone tell me what I'm missing? - -- Andy Markham markham@nortel.ca Nortel, Inc. --------------------------- Newsgroups: comp.os.vxworks Subject: VXSIM - Round Robin Scheduling Date: Fri, 12 Jul 1996 17:01:04 GMT From: steve@exotech.com (Steve@Exotech.Com) Organization: MV Communications, Inc. Message-ID: Has anyone noticed problems with the round robin scheduling enabled by kernelTimeSlice()? It does not seem to run as one would expect under VXSIM. --------------------------- Newsgroups: comp.os.vxworks Subject: long long support in VxWorks 5.2 Date: Wed, 10 Jul 1996 22:27:37 GMT From: Philip.Wiebe@newnes.com (Philip J. Wiebe) Organization: MIND LINK! - British Columbia, Canada Message-ID: <4s1aqj$fnf@fountain.mindlink.net> GNU 2.2.3 specifies that it supports double-word integers (i.e. long long and unsigned long long); however, bit shifting operations cause the VxWorks 5.2 loader to come up with an undefined symbol. For example, the code fragment long long a, b; ... a = b << 2; .. results in an undefined symbol of ___lshldi3. Is there any additional libraries that must be linked in, or is there no support for these types of operations on double words? Phil. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Philip J. Wiebe "It has yet to be proven that intelligence Software Engineer has any survival value." Newnes Machine Ltd. - Arthur C. Clarke P.O.Box 8, 3550-45th Street S.E. Salmon Arm, BC Canada email: Philip.Wiebe@newnes.com ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ --------------------------- End of New-News digest ********************** From daemon@csg.lbl.gov Sun Jul 14 04:01:24 1996 From: daemon@csg.lbl.gov Date: Sun Jul 14 04:01:27 PDT 1996 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sun Jul 14 04:01:20 PDT 1996 Subject: Re: Supervisory ? Non-priviledge Subject: Re: When will Tornado be ported to Linux? Subject: Tape Devices Subject: Re: running remote commands Subject: VxGDB sources Subject: Re: VxGDB sources Subject: Tornado/Diab Data compiler Subject: CRT output ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Supervisory ? Non-priviledge Date: 9 Jul 1996 15:46:10 GMT From: kadionik@aramis (Patrice KADIONIK (IXL - CNRS URA 846 - FRANCE)) Organization: CRIBX1 , Universite de Bordeaux I , France Message-ID: <4rtus2$9dc@news.u-bordeaux.fr> References: <31D92F1D.2FB9@ict.nl> Bret Indrelee (breti@bit3.com) wrote: : In article <31D92F1D.2FB9@ict.nl>, Hiske Huiting wrote: : >I have a 2 questions about VME communication. : >1. What is the difference between Supervisory and Non-priviledge : > addressing modes? : The address modifier used. The VMEbus specification doesn't do anything : except define them. : Many processors make a differentiation between user and supervisory mode, : where there are priviledged registers and instructions that can only be : executed while the processor is in supervisory mode. Some interfaces will : use this information when generating the address modifier. This is strickly : an implementation issue, the VMEbus specification does not require this. : Most VMEbus cards don't pay any attention to the two modes. If they do, : they will usually have a jumper that allows them to only accept supervisory : accesses. : >2. What is the difference between Master and Slave access? : The bus Master initiates the transaction. The bus Slave responds to the : transaction. : Memory cards are examples of bus slaves. Accesses by processors are always : bus masters. SBC cards usually also allow slave access to the memory on : the processor card as well. : -Bret : -- : Bret Indrelee : breti@bit3.com #include Allthough VME norm is not based on an family of microprocessors, you find analogy between the 68k family and the VME (remember that VME was specified with Motorola). The AM codes have a same role than the FC lines of a 68k processor. That permits to make a memory partition between supervisor or user data and code. One can also distinguish A16, A24 and A32 address spaces too. So remember : a VME access is an address value plus a AM code value (precising the address space type and supervisor or user acces mode). Cheers; - -- Pat. P. Kadionik Laboratory of Microelectronics - IXL University of Bordeaux I Bordeaux - FRANCE \^ ^/ )@ @( +---------oOO--(_)-------------+ | Patrice Kadionik | | kadionik@ixl.u-bordeaux.fr | +-----------------Oooo--oOO----+ oooO ( ) ( ) ) / \ ( (_/ \_) --------------------------- Newsgroups: comp.os.vxworks Subject: Re: When will Tornado be ported to Linux? Date: 13 Jul 1996 07:57:56 -0400 From: dave@sanders.com (Dave Uhrenholdt) Organization: Sanders SSD Message-ID: References: <4s12v6$4p5@ringo.reno.nv.us> Sender: dave@cushing.nowhere.com In article <4s12v6$4p5@ringo.reno.nv.us> mike@ringo.reno.nv.us (Michael Morrison) writes: Linux runs on SPARC, Alpha, PPC and Intel and has all the bells and wistles when will WRS support Linux for a development platform? mike mike@ringo.reno.nv.us I asked this same question of Wind River in the open question and answer period at the Wind River User's conference in Burlington Mass. this past April 1. Their response was that some of their engineers had played with this internally, but they had no plans to market it. I think they feel that the place to go is Windows 95, Windows-NT. I have not used Tornado, I use VxWorks 5.2. I have been able to put together most of the important Cross development tools from the GNU distribution. I haven't yet tried them on Linux - I have built them for HP-UX and work ok. I have the linker, assembler, compiler and gdb working. I originally went with the GNU distribution because there was a bug in the linker, and they where unwilling to fix it. It turns out that the tools for 5.2 are four years out of date. When you go with GNU you loose a few things like TCL support within the debugger. The problem with Tornado is that alot more of their stuff on the host side is propietary. I hear that a good deal of it is written in TCL which is portable, but if it isn't all portable, how can you port it. Dave --------------------------- Newsgroups: comp.os.vxworks Subject: Tape Devices Date: 9 Jul 1996 15:40:48 GMT From: pjs@mc.com (Peter Satkwich) Organization: Mercury Computer System Message-ID: <4rtui0$dpb@newsy.mc.com> References: <9607051446.AA07806@lbl.gov> Reply-To: pjs@mc.com I would like to connect a scsi tape device to a mvme167 running vxWorks 5.2. Does anyone have any experience doing this. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: running remote commands Date: 12 Jul 1996 18:08:02 -0700 From: mwette@mr-ed.jpl.nasa.gov (Matt Wette) Organization: Jet Propulsion Laboratory Message-ID: <4s6sti$8qh@mr-ed.jpl.nasa.gov> References: <1996Jul12.140751.24774@hrbicf> In article <1996Jul12.140751.24774@hrbicf>, Michael E. Warren wrote: > >Is it possible to run VxWorks commands remotely? > >From our sun, I can run > > rsh nodename > >and I log in. But > > rsh nodename "ls" > >gives me a connection refused error. Is there something I'm missing? /* rshd.c - rshd for VxWorks * * Copyright (C) 1993 Matthew R. Wette -- all rights reserved. * * $Id: rshd.c,v 1.1 1995/09/27 04:03:24 mwette Exp $ * * You may use this program if this copyright notice is kept in tact and * you don't try to sell it. * * To use: * compile to rshd * ld < rshd * rshdInit */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include static char const rcsid[] = "$Id: rshd.c,v 1.1 1995/09/27 04:03:24 mwette Exp $"; #ifndef RSHD_PORT #define RSHD_PORT 514 #endif #ifndef RSHD_PRIO #define RSHD_PRIO 5 #endif #define SHELL_CMD 1 /* is a shell command */ #define RESET_CMD 2 /* is a reset command */ #define FEVAL_CMD 3 /* is a function eval command */ int rshdInit(void); int rshdTask(void); static int rshd(int sock, struct sockaddr_in*); static int getstr(int, char*, int, char*); static int rshdTid = 0; static int rshdStackSize = 0x8000; static char rcmd_errmsg[80]; static int cmdType(char *cmd); static int iCmdExec(char *cmd); static int qCmdPending = 0; static char qCmdBufr[128]; static int qCmd(char *cmd); static int qCmdExec(void); int rshdInit(void) { rshdTid = taskSpawn("aRshd", RSHD_PRIO, VX_STDIO, rshdStackSize, rshdTask, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); if (rshdTid == 0) return ERROR; return OK; } int rshdTask(void) { int status, serv_sock, clnt_sock, clen; struct sockaddr_in servaddr, clntaddr; memset(&servaddr, '\0', sizeof(servaddr)); memset(&clntaddr, '\0', sizeof(clntaddr)); /* Open the socket. */ serv_sock = socket(AF_INET, SOCK_STREAM, 0); if (serv_sock == ERROR) return ERROR; servaddr.sin_family = AF_INET; servaddr.sin_port = htons(RSHD_PORT); status = bind(serv_sock, (struct sockaddr *)&servaddr, sizeof(servaddr)); if (status == ERROR) { close(serv_sock); logMsg("rshd: bind failed\n", 0, 0, 0, 0, 0, 0); return ERROR; } status = listen(serv_sock, 2); if (status == ERROR) { logMsg("rshd: listen failed\n", 0, 0, 0, 0, 0, 0); close(serv_sock); return ERROR; } while (1) { clen = sizeof(clntaddr); clnt_sock = accept(serv_sock, (struct sockaddr*)&clntaddr, &clen); if (clnt_sock == ERROR) { logMsg("rshd: accept failed\n", 0, 0, 0, 0, 0, 0); continue; } status = rshd(clnt_sock, &clntaddr); close(clnt_sock); } return OK; } static int rshd(int sock, struct sockaddr_in *addr) { int n, status, was_stdout, was_stderr; int secdport, secdsock, loclport; char c, clntuser[16], servuser[16], bufr[128]; /* Get secondary port request. */ /*alarm(60);*/ secdport = 0; while ((n = read(sock, &c, 1)) == 1) { if (c == '\0') break; secdport = 10*secdport + c - '0'; } if (secdport != 0) { loclport = IPPORT_RESERVED - 1; secdsock = rresvport(&loclport); addr->sin_port = htons((unsigned short)secdport); status = connect(secdsock, (struct sockaddr*)addr, sizeof(*addr)); if (status == ERROR) { logMsg("rshd: connect() on second port %d failed\n", secdport, 0, 0, 0, 0, 0); close(sock); return status; } } /* get the command strings */ getstr(sock, clntuser, sizeof(clntuser), "clntuser"); getstr(sock, servuser, sizeof(servuser), "servuser"); getstr(sock, bufr, sizeof(bufr), "command"); # ifdef DEBUG logMsg("rshd: command=[%s]\n", bufr, 0, 0, 0, 0, 0); # endif switch (cmdType(bufr)) { case SHELL_CMD: /* -- lock shell -- */ status = shellLock(TRUE); if (status == FALSE) { logMsg("rshd: failed to lock shell\n", 0, 0, 0, 0, 0, 0); write(sock, "!", 1); remCurIdGet(servuser, 0); sprintf(rcmd_errmsg, "rshd: shell locked by user \"%s\"\n", servuser); if (strlen(rcmd_errmsg) > 80) rcmd_errmsg[79] = '\n'; write(sock, rcmd_errmsg, strlen(rcmd_errmsg)); goto exit2; } status = remCurIdSet(clntuser, 0); /* -- set I/O -- */ was_stdout = ioGlobalStdGet(1); ioGlobalStdSet(1, sock); if (secdport) { was_stderr = ioGlobalStdGet(2); ioGlobalStdSet(2, secdsock); } /* -- let the client know we're alive -- */ if (write(sock, "", 1) != 1) { logMsg("rshd: null write to client failed\n", 0, 0, 0, 0, 0, 0); goto exit3; } /* -- execute -- */ status = execute(bufr); if (status == ERROR) { logMsg("rshd: execute() failed\n", 0, 0, 0, 0, 0, 0); goto exit3; } break; case RESET_CMD: /* Check for internal command. */ status = iCmdExec(bufr); if (status < 0) { logMsg("rshd: internal command not known\n", 0, 0, 0, 0, 0, 0); goto exit3; } else if (0 < status) { goto exit1; } break; case FEVAL_CMD: default: logMsg("rshd: bad command\n", 0, 0, 0, 0, 0, 0); goto exit2; } exit3: /* Reestablish I/O */ ioGlobalStdSet(STD_OUT, was_stdout); if (secdport) ioGlobalStdSet(STD_ERR, was_stderr); exit2: /* unlock shell */ status = shellLock(FALSE); exit1: /* close sockets */ close(sock); if (secdport) close(secdsock); if (qCmdPending) qCmdExec(); return status; } static int getstr(int sock, char *bufr, int cnt, char* errmsg) { int i, n; char c; bufr[i=0] = '\0'; while ((n = read(sock, &c, 1)) == 1) { bufr[i++] = c; if (i == cnt) return ERROR; if (c == '\0') return OK; } if (n == -1) return ERROR; return OK; } /* cmdType() - classify the command type */ static int cmdType(char *cmd) { while (*cmd == ' ' || *cmd == '\t') cmd++; if (*cmd++ != '-') return SHELL_CMD; if (strcmp(cmd, "feval") == 0) return FEVAL_CMD; if (strcmp(cmd, "sysreset") == 0) return RESET_CMD; if (strcmp(cmd, "reboot") == 0) return RESET_CMD; return -1; } #if 0 static int evalCmd(char *cmd) { int argc; char *argv[10]; while (*cmd == ' ' || *cmd == '\t') cmd++; if (*cmd++ != '-') return -1; if (strcmp(cmd, "feval") != 0) return -1; cmd += strlen("feval"); for (argc = 0; argc < 10; argc++) { while (*cmd == ' ' || *cmd == '\t') cmd++; if (*cmd == '\0') break; argv[argc] = cmd; while (*cmd == ' ' || *cmd == '\t') cmd++; } argv[0] = cmd; argc = 1; return 0; } #endif static int iCmdExec(char *cmd) { qCmdPending = 0; while (*cmd == ' ' || *cmd == '\t') cmd++; if (*cmd++ != '-') return 0; if (strcmp(cmd, "vmereset") == 0) qCmd(cmd); else if (strcmp(cmd, "sysreset") == 0) qCmd(cmd); else if (strcmp(cmd, "reboot") == 0) qCmd(cmd); else if (strcmp(cmd, "coldboot") == 0) qCmd(cmd); else if (strcmp(cmd, "warmboot") == 0) qCmd(cmd); else if (strcmp(cmd, "quickboot") == 0) qCmd(cmd); else { return -1; } return 1; } static int qCmd(char *cmd) { strcpy(qCmdBufr, cmd); #ifdef DEBUG printf("qCmd: qCmdBufr=\"%s\"\n", cmd); #endif qCmdPending = 1; return 0; } static int qCmdExec(void) { taskDelay(sysClkRateGet()); qCmdPending = 0; # ifdef DEBUG printf("qCmdExec: qCmdBufr=\"%s\"\n", qCmdBufr); # endif if (strcmp(qCmdBufr, "vmereset") == 0) { *(unsigned short*)0xff0000e2 = (unsigned short)0x00f0; } else if (strcmp(qCmdBufr, "sysreset") == 0) { *(unsigned short*)0xff0000e2 = (unsigned short)0x00f0; } else if (strcmp(qCmdBufr, "reboot") == 0) { reboot(BOOT_NORMAL); } else if (strcmp(qCmdBufr, "coldboot") == 0) { reboot(BOOT_CLEAR); } if (strcmp(qCmdBufr, "warmboot") == 0) { reboot(BOOT_NO_AUTOBOOT); } if (strcmp(qCmdBufr, "quickboot") == 0) { reboot(BOOT_QUICK_AUTOBOOT); } return OK; } /* SYNOPSIS int rshdInit(void) - starts rshdTask int rshdTask(void) - server task int rshd() - service a request on Unix host: rsh DESCRIPTION This is an implementation of the remote shell daemon for VxWorks hosts. It allows users on Unix hosts to run shell commands via the rsh. ERRORS If an error occurs rsh will prints a single line error message. BUGS Surely some exist. SEE ALSO: rsh(1), rcmd(1), rshd(8), in.rshd(8) AUTHOR Matt Wette */ /* --- last line of rshd.c --- */ - -- matthew.r.wette@jpl.nasa.gov --------------------------- Newsgroups: comp.os.vxworks Subject: VxGDB sources Date: Thu, 11 Jul 1996 11:43:27 GMT From: lyorg@scorpio.com (scorpio) Organization: Scorpio Message-ID: Sender: news@actcom.co.il (News) Hello, We are running VxWorks 5.2 on top of SPARC cpu. We are looking into the possibility of porting the cross-development environment to Linux. When I looked at the GNU toolkit supplied along with the VxWorks distribution CD, there was no debugger source (!!!). I seem to remember having read in this newsgroup about someone who created his own version of the debugger based on "gdb" (which VxGDB itself admits in its warranty disclaimer to be based on...). a. Isn't WRS supposed to supply the source for VxGDB as article (3) of the FSF general license says ? b. VxGdb says it is based on "gdb4.6" - the best I could find on any GNU site is gdb4.16 - where is this gdb4.6 ? c. Has anyone ported successfully the VxGdb and can send hints as to how to do this ? I am mostly interested in the "attach" command which must be changed from its UNIX default (attaching processes) to its VxWorks analog (attaching tasks). Any other information would be welcome. Thank you, Lyor (lyorg@scorpio.com) --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxGDB sources Date: 11 Jul 1996 11:11:24 -0700 From: shebs@cygnus.com (Stan Shebs) Organization: Cygnus Support Message-ID: References: Sender: shebs@andros.cygnus.com In article lyorg@scorpio.com (scorpio) writes: From: lyorg@scorpio.com (scorpio) Date: Thu, 11 Jul 1996 11:43:27 GMT b. VxGdb says it is based on "gdb4.6" - the best I could find on any GNU site is gdb4.16 - where is this gdb4.6 ? Check the Museum of Software History. :-) GDB 4.6 is from a long time ago. I can probably turn up a copy around here, but it's not likely to be of much use to you, since WRS has certainly made modifications and improvements since then. You shouldn't have any problem getting a copy of the sources from them - I would suggest starting with your WRS field rep. Stan Shebs Cygnus Support shebs@cygnus.com --------------------------- Newsgroups: comp.os.vxworks Subject: Tornado/Diab Data compiler Date: Wed, 10 Jul 1996 07:42:04 -0500 From: Steve Woloschek Organization: GE Medical Systems Message-ID: <31E3A51C.ABD@ct.med.ge.com> I am currently evaluating Tornado.. I realize that the gnu tool chain is included.. But does anyone have any experience with the Diab Data tool chain with Tornado? - -- Steve Woloschek woloschekS@med.ge.com GE Medical Systems --------------------------- Newsgroups: comp.os.vxworks Subject: CRT output Date: Wed, 10 Jul 1996 18:14:54 -0700 From: Duane C Austin Organization: A customer of Pacific Bell Internet Services Message-ID: <31E4558E.E5A@pacbell.net> Hi, I am looking for an easy and low cpu usage solution of how to writing to a CRT display within my target. My target is a FORCE CPU10 card. I can't use the serial chip on the card (console device and ttyDriver) because each byte written to the display causes an interrupt and would be too cpu intensive. What are my other solutions? Can I use the ethernet interface? If yes, what is involved? Any other solutions? Thanks in advance, Duane Austin email: DCAustin@pacbell.net --------------------------- End of New-News digest ********************** From froeber@bbn.com Sun Jul 14 16:57:21 1996 From: Fred Roeber Date: Sun Jul 14 16:57:23 PDT 1996 Subject: Re:Which was running task before interrupt? David Soares wrote: > One thing I don't know how to do is to get the task's ID from the TCB, > other than to use the name from the TCB to call taskNameToId(), which > is very cumbersome. On all versions of VxWorks that I know of, the task ID and the TCB address are one and the same. In fact, most "id" values in VxWorks are actually just opaque pointers to the control blocks for the corresponding entities. This is true of tasks, semaphores, message queues... Since VxWorks comes with the include files that define a lot of the control structures you can easily poke around in the different control structures given their ID values. I wouldn't go changing values this way or even reading values in code you want to be portable but the technique can be useful at times. Fred | Fred J Roeber, BBN Systems & Technologies | | 50 Enterprise Place Middletown, RI 02842-5202 | | froeber@bbn.com 401-849-2543 (X48) | | See http://www.bbn.com/tv for TraceMaker info | From han2326@aminet.co.kr Mon Jul 15 22:56:32 1996 From: Han Jeonh Soo Date: Mon Jul 15 22:56:34 PDT 1996 Subject: how to use select() function Let me know how to use select() function, especially struct timeval. I want that the my application send command to a communication device, and then the device send the response data within 10 ms. If the device do not send the response data within 10 ms, Error routine will be proceed. Let me know how to check & how to set the struct timeval of select function. Best Regard, From daemon@csg.lbl.gov Tue Jul 16 04:00:36 1996 From: daemon@csg.lbl.gov Date: Tue Jul 16 04:00:39 PDT 1996 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Tue Jul 16 04:00:32 PDT 1996 Subject: Re: Debugger other than VxGDB? Subject: Booting i486 BSP from EPROM Subject: Ethernet hardware address. Subject: VxWorks freeware archive? Subject: running remote commands Subject: RPC trouble with rdb Subject: Re: Delays in download VxWorks symbol table using FTP Subject: Re: GNU Warning: trigraph ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Debugger other than VxGDB? Date: Fri, 12 Jul 1996 15:19:32 -0700 From: Bob Strickland Organization: Lockheed Martin Message-ID: <31E6CF73.1622@mm.ssd.lmsc.lockheed.com> References: <4rrnun$4qp@bcrkh13.bnr.ca> <4ru7bk$2vh@elaine25.Stanford.EDU> Reply-To: 1111, Lockheed, Way, MS, 32-40/156A Objectcenter works in two modes: + as described, it simply kicks in GDB from the objectcenter windows, + an emlation mode where your code is not actually compiled, linked and run. In this mode the code is interpreted by OC and provides very robust 'compiler' and run-time checks. The limitation is that your code can't run through vxWorks services such as sysBusToLocal, semGive, etc. We've been using both GDB and OC. Both have been helpful. - --Bob Lockheed Martin --------------------------- Newsgroups: comp.os.vxworks Subject: Booting i486 BSP from EPROM Date: 15 Jul 1996 15:53:10 GMT From: David Costain <100072.3624@CompuServe.COM> Organization: VG Organic Message-ID: <4sdph6$l7t$1@mhadg.production.compuserve.com> I am running the VxWorks 5.2, Intel i486 BSP on a Pentium. This BSP is currently booting from the floppy drive, then downloading the image and symbol table over the network. We are looking at using a minimal target system that has no floppy drive, but will have an in-house ISA pcb containing an EPROM socket(s), into which we intend to insert VxWorks bootrom(s), as we do on our 68k target systems. Has anyone been through this exercise ? Is it fairly straightforward, or are there problems ??. Any information gratefully accepted. From: Dave Costain 100072.3624 (dave.costain@micromass.co.uk) --------------------------- Newsgroups: comp.os.vxworks Subject: Ethernet hardware address. Date: 13 Jul 1996 13:01:36 -0400 From: duhrenho@sanders.com (David Uhrenholdt) Organization: Sanders SSD Message-ID: Sender: duhrenho@hickory I am interested in a subroutine call that provides the hardware address (48 bit address) of the ethernet address. I know you can get the address, because it shows up in ifShow. And ifunit provides most of the information output by ifShow, but it doesn't provide the hardware address. What I really am trying to do is RARP for a FDDI card that I am adding after download. Thanks for any help Dave - -- Dave Uhrenholdt duhrenho@vette.sanders.lockheed.com Phone (603) 885-7398 --------------------------- Newsgroups: comp.os.vxworks Subject: VxWorks freeware archive? Date: 15 Jul 96 14:39:13 -0500 From: hudgins_jc@corning.com (Jerry Hudgins) Organization: Corning Incorporated Message-ID: <1996Jul15.143914.1542@corning.com> Reply-To: hudgins_jc@corning.com I'd appreciate pointers to any net archives of VxWorks freeware. If such information is available via FAQ, a pointer to that would be likewise appreciated. - -- Jerry Hudgins | Corning Incorporated | hudgins_jc@corning.com --------------------------- Newsgroups: comp.os.vxworks Subject: running remote commands Date: 12 Jul 96 14:07:51 EST From: m1w@icf.hrb.com (Michael E. Warren) Organization: HRB Systems, Inc. Message-ID: <1996Jul12.140751.24774@hrbicf> Is it possible to run VxWorks commands remotely? From our sun, I can run rsh nodename and I log in. But rsh nodename "ls" gives me a connection refused error. Is there something I'm missing? Mike Warren HRB Systems State College, PA --------------------------- Newsgroups: comp.os.vxworks,comp.unix.aix Subject: RPC trouble with rdb Date: 16 Jul 1996 02:29:50 GMT From: rklar@photon.cs.tamu.edu (Robert A Klar) Organization: Texas A&M University, College Station, TX Keywords: aix,vxworks,rpc,rdb Message-ID: <4seuqu$rah@news.tamu.edu> Please respond to the newsgroup and to the email address listed at the bottom of this message - do not respond to the source email address. We are running the remote debugger (rdbTask) on VxWorks 5.1.1. We are using the MULTI debugger from Green Hills on an AIX 4.1.4 host. The VxWorks host is connected to the AIX host over a SLIP connection. When the rdb task on VxWorks starts up, it registers its RPC services with its local portmapper. It then apparently broadcasts its services to the AIX 4.1.4 host, which does not respond, or at least not fast enough because the rdb task issues a timeout errno (0xD0004 or something like that) immediately. We borrowed an AIX 3.2.5 host and, strangely, it does respond to the rdb's broadcast. So, we are currently using a kludgy bait-and-switch workaround by allowing rdb task to talk to the AIX 3.2.5 host first, then switching the SLIP connection to the AIX 4.1.4 host. Then everything works great. What is different between AIX 3.2.5 and AIX 4.1.4 which would cause this anomaly? Is it a security setting or configuration option? The port number is not in conflict with any port number in /etc/services. We have tried adding the port number to /etc/services. Nothing seems to make the AIX 4.1.4 machine respond. I appreciate any help you can provide. Ronnie Killough RKillough@swri.edu 210/522-3616 --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Delays in download VxWorks symbol table using FTP Date: Mon, 15 Jul 1996 22:08:42 +0100 From: John Gordon Organization: UnipalmPIPEX server (post doesn't reflect views of UnipalmPIPEX) Message-ID: <31EAB35A.5AB9@dial.pipex.com> References: <31EA9227.A6@po3.gi.com> I have six PPC boards sitting in a rack which all try to boot simultaneously from a SparcStation 10 over ethernet. I have noticed that my chances of all six booting correctly when the VME reset is asserted (via the slot one card reset) is slim. Assuming they get as far as running the bootrom (not always the case) they seem to fail at some stage either during kernel download or symbol table download. This is currently on my list of things to investigate! John... PS It is VxWorks 5.2 for PowerPC / Solaris 2.4 on the Sparc --------------------------- Newsgroups: comp.os.vxworks Subject: Re: GNU Warning: trigraph Date: 15 Jul 1996 15:23:51 -0700 From: kla@leland.Stanford.EDU (Earl Mitchell) Organization: Stanford University, CA 94305, USA Message-ID: <4segdn$d4m@elaine25.Stanford.EDU> References: <31E591FC@msmail.active.ch> In article <31E591FC@msmail.active.ch>, Geiter Paul wrote: > >Hi > >I am using the Tornado compiler on a NT Host and get the following warning >message: > > mmi_menu.h:1 warning: 1 trigraph(s) encountered > >Can anyone give me a hint, where to look for an error. I had this same problem a while back. It turned out that the compiler didn't like the following line of code: printf("Software Release ???-%s.\n", getSwVersion()) ; But when I replaced it with the following the trigraph warning went away... printf("Software Release UNKNOWN-%s.\n", getSwVersion()) ; The "?" chars somehow screwed up the parsing. The line above was part of a switch statement (see routine below): void versionShow(void) { switch ((int) HARDWAREMODEL) { case MOT6T1: printf(" Software Release MOT6T1-%s.\n", getSwVersion()) ; break ; case MEL4E1: printf(" Software Release MEL4E1-%s.\n", getSwVersion()) ; break ; case V514E1: printf(" Software Release V514E1-%s.\n", getSwVersion()) ; break ; default: printf(" Software Release UNKNOWN-%s.\n", getSwVersion()) ; } } You should run your program thu the preprocessor only and then see what is actually being fed to the compiler. If you can't find it this way then comment out most of code and iteratively uncomment stuff until you can zero in on which line of code introduces the trigraph warning. --------------------------- End of New-News digest ********************** From uchenick@tis4000.com Tue Jul 16 05:24:31 1996 From: Gordon Uchenick Date: Tue Jul 16 05:24:33 PDT 1996 Subject: RE: how to use select() function > Let me know how to check & how to set the struct timeval of select function. For this example, assume that you are reading the response on file descriptor fd. You may have to fiddle with the list of #include files here depending upon your specific release of vxWorks. #include #include #include /* Send the request to the device. If you are using TCP sockets, make sure you use setsockopt() to assert TCP_NODELAY. */ fd_set readMask; struct timeval timeOut; int err; timeOut.tv_sec = 0; timeOut.tv_usec = 10000; /* 10,000 usec = 10 msec */ FD_ZERO(&readMask); FD_SET(fd, &readMask); if ((err = select(fd + 1, &readMask, NULL, NULL, &timeOut)) == 0) { /* You timed out. */ } else if (err > 0) { /* At least one byte of the response came in. There are no guarantees about how many bytes have arrived or if any of those bytes were received in error. You'll have to issue a read() or one of the variants of recv() to find out. */ } else { /* Error in select call. */ } Hope this helps, Gordon Uchenick uchenick@tis4000.com From daemon@csg.lbl.gov Wed Jul 17 04:00:28 1996 From: daemon@csg.lbl.gov Date: Wed Jul 17 04:00:30 PDT 1996 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Wed Jul 17 04:00:25 PDT 1996 Subject: Re: Mil Spec Hard Disk Drive Subject: Network Functions Subject: Delays in download VxWorks symbol table using FTP Subject: Utils to run Checksums? Subject: Round-Robin Question Subject: Re: Application running "slower" in vxWorks.st mode. Help!! Subject: ARCNET driver Subject: Is there any pipeDevDelete() or similar function Subject: coff to ieee695 ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Mil Spec Hard Disk Drive Date: 16 Jul 1996 12:17:32 GMT From: msinger@griffon.oe.fau.edu (Matthew Singer) Organization: Florida Atlantic University Message-ID: <4sg18s$lds@cybernet.cse.fau.edu> References: <199607110210.AA23126@mail.wrs.com> Reply-To: msinger@griffon.oe.fau.edu (Matthew Singer) Sender: msinger@griffon (Matthew Singer) In article <199607110210.AA23126@mail.wrs.com>, namiq@wrs.com writes: |>Hi Folks: |> |>Can anyone recommend a 4 Giga Byte Ruggedised or Mil Spec SCSI Hard Disk |>Drive |> (other than Miltope). |> |>Thanks in advance, |>Best Regards, |> |> |> |>---------------------------------------------------------------------------- |>- |>Namiq Kunimoto Email: namiq@wrs.com |>Director of Sales Phone: 81-3-5467-5900 |>Asia Pacific Operations FAX: 81-3-5467-5877 |>Wind River Systems WWW: http://www.wrs.com |>Japan Office | |>POLA Ebisu Bldg. 11F +-------------------------------------------- |>3-9-19, Higashi, Shibuya-ku, Tokyo, 150, Japan |>---------------------------------------------------------------------------- |>- |> |> - -- Call Steve Lauria 607-751-2633 at Loral. They sell near space qaul. I haved used them in the past with good results. They are expensive however... - ------------------------------------------------------------------------------- Matthew R. Singer msinger@oe.fau.edu Systems Engineer Voice: 407-367-2947 Florida Atlantic University Fax : 407-367-3885 Dept of Ocean Engineering - ------------------------------------------------------------------------------- --------------------------- Newsgroups: comp.os.vxworks Subject: Network Functions Date: 16 Jul 1996 14:19:21 +0100 From: etlwepe@etlxdmx.ericsson.se (Wayne Price) Organization: Ericsson Message-ID: Sender: etlwepe@etlxd13m Greetings, I have some questions maybe someone out there can answer. In C there is a function call gethostent does vxworks have something like this ? I have looked at various host & inet functions. I need to get the aliases for all the machine and filter out the one's which I don't need. Then I want to convert all the alaises into the IP-Addresses. I've done this in C, however Vxworks does not have the functionality to get all the hostnames by there alaises and convert these into the IP-addesses. Thanks for your help/suggestions. Wayne. - -- "Men were real men. women were real .---. .----------- women, and small, furry creatures from / \ __ / ------ Alpha Centauri were REAL small, furry / / \( )/ ----- creatures from Alpha Centauri. ////// ' \/ ` --- Spirits were brave, men boldly split //// / // : : --- infinitives that no man had split // / / /` '-- before. Thus was the Empire forged." // //..\\ ====UU====UU==== -- The Hitchhiker's Guide to the Galaxy", '//||\\` Douglas Adams ''`` ________________________________________etlwepe@etlxdmx.ericsson.se_ Disclaimer: My views do not necessarily reflect those of my employer. --------------------------- Newsgroups: comp.os.vxworks Subject: Delays in download VxWorks symbol table using FTP Date: Mon, 15 Jul 1996 11:11:13 GMT From: Tom Fuda Organization: Northrop Grumman Norden Systems Message-ID: Sender: usenet@tron.bwi.wec.com (USEnet News Poster) I am using 486 targets (VME based single board computers from Logic Design Group with on board Ethernet and IDE disk drive) in a Sun host environment. I have noticed that if I try to download the VxWorks image (separate image and symbol table) to my target using FTP it takes approximately 1 minute to download the symbol table while the image itself is downloaded in a matter of seconds. If I use RSH to download the symbol table from the Sun host the transfer only takes a few seconds. This problem is observed whether my boot host is the Sun or another VxWorks target (using either Ethernet or shared memory network). I have also noticed that is two VxWorks targets simultaneously attempt to download images from the same host using FTP, one download usually fails while the other one succeeds after about a minute. Has anyone else experienced this problem? If so, do you know of a workaround or fix? Thanks in advance. Thomas Fuda Northrop Grumman Norden Systems (203) 852-7991 fuda@norden.com --------------------------- Newsgroups: comp.os.vxworks Subject: Utils to run Checksums? Date: 15 Jul 1996 13:06:26 GMT From: Skip Rye Organization: Raytheon Company Message-ID: <4sdfoi$tsl@swlab1.msd.ray.com> We have a vxworks computer that may need to run checksums on the software. Is there a utility to run checksums on programs? --------------------------- Newsgroups: comp.os.vxworks Subject: Round-Robin Question Date: 16 Jul 1996 18:35:29 GMT From: hac.es.com Organization: Hughes Aircraft Company Message-ID: Anyone, What is the easiest way to find out of my VxWorks image has Round-Robin scheduling enabled or not? Thanx MAPP --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Application running "slower" in vxWorks.st mode. Help!! Date: 16 Jul 1996 18:26:05 GMT From: johncoby@ghgcorp.com (JOHN R. COBARRUVIAS) Organization: Personal Account Message-ID: References: In article , loginid@server.jsc.nasa.gov (firstname lastname) wrote: Well firstname, lastname, call me. Your email bounces. X39357 > We are experiencing a very serious problem with the vxWorks.st running on > 486 target. This problem has been reported to Wind River for more than a > week. I still haven't got any clue or answer yet. The folowing is a simple > code I have "timexN"ed: > > void delay() > { > int i; > > for(i=0; i<163; i++); > } > > It took about 55 micro sec when "vxWorks" down loaded from host. > However, when in "vxWorks.st" mode, it took 233 micro sec. > > This is just an example. The whole application ran much slower. The > performance is not acceptable for the application. > I appreciate if any one can help. - -- ,-_/ . .-,--. ,---. . ' | ,-. |-. ,-. `|__/ | -' ,-. |-. ,-. ,-. ,-. . . . , . ,-. ,-. | | | | | | | | \ ,. | -. | | | | ,-| | | | | | / | ,-| `-. | `-' ' ' ' ' `-' ` `' `---' `-' `-' `-^ ' ' `-' `' ' `-^ `-' / | John R. Cobarruvias johncoby@ghgcorp.com Houston, Tx `--' --------------------------- Newsgroups: comp.os.vxworks Subject: ARCNET driver Date: Wed, 17 Jul 1996 05:09:29 GMT From: sgaeke@cris.com (Steve Gaeke) Organization: Concentric Internet Services Message-ID: <4shi2h$668@herald.concentric.net> I am looking at possibly using arcnet and was wondering if anyone has already written a VxWorks driver to allow it to work with TCP/IP? Thanks, steve sgaeke@cris.com --------------------------- Newsgroups: comp.os.vxworks Subject: Is there any pipeDevDelete() or similar function Date: 17 Jul 1996 13:13:17 +1000 From: dmr@hercules.iassf.easams.com.au (Dewan RASHID) Organization: EASAMS (Australia) Pty Ltd Message-ID: <4shlod$7d0@hercules.iassf.easams.com.au> Dear VxPerts: Is there a way to delete a pipe after it is created? Thanks, Dewan Sydney, Australia. --------------------------- Newsgroups: comp.os.vxworks Subject: coff to ieee695 Date: Mon, 15 Jul 1996 15:16:17 -0500 From: Steve Woloschek Organization: GE Medical Systems Message-ID: <31EAA711.167E@ct.med.ge.com> I thought I saw a post regarding a conversion utility for converting coff to ieee695? Sorry for the repost but.. Does anyone have knowledge about such a utility? - -- Steve Woloschek woloschekS@med.ge.com GE Medical Systems --------------------------- End of New-News digest ********************** From pean@mbari.org Wed Jul 17 08:12:35 1996 From: Andrew Pearce Date: Wed Jul 17 08:12:37 PDT 1996 Subject: Re: Round-Robin Question >Anyone, > >What is the easiest way to find out of my VxWorks image has Round-Robin >scheduling enabled or not? > >Thanx >MAPP Check the vxWorks global variable roundRobinOn. It is set to a value of 1 when round-robin scheduling is enabled and set to 0 when round-robin scheduling is disabled. You can list the function kernelTimeSlice for more information. -- Andrew Pearce Monterey Bay Aquarium Research Institute tel: (408) 775-1746 7700 Sandholt Rd fax: (408) 775-1646 Moss Landing, CA 95039 email: pean@mbari.org "Coping's easy. Not pureeing your loved ones, that's the hard part." Basil Fawlty From stan@rti.com Wed Jul 17 10:36:24 1996 From: Stan Schneider Date: Wed Jul 17 10:36:25 PDT 1996 Subject: Re: Application running "slower" in vxWorks.st mode. Help! Hi, > We are experiencing a very serious problem with the vxWorks.st running on > 486 target. This problem has been reported to Wind River for more than a > week. I still haven't got any clue or answer yet. The folowing is a simple > code I have "timexN"ed: > > void delay() > { > int i; > > for(i=0; i<163; i++); > } > > It took about 55 micro sec when "vxWorks" down loaded from host. > However, when in "vxWorks.st" mode, it took 233 micro sec. > > This is just an example. The whole application ran much slower. The > performance is not acceptable for the application. > I appreciate if any one can help. Are you linking your code with the kernel when building the vxWorks.st image? If so, don't use the "-fvolatile" flag (I think it's still set in the default makefile). It disables the optimizer. You can check by disassembling the code ("l delay"). Your example program should result in exactly zero code with the optimizer enabled. HTH. -- Stan stan@rti.com From pardo@rti.com Wed Jul 17 11:42:09 1996 From: Gerardo Pardo-Castellote Date: Wed Jul 17 11:42:11 PDT 1996 Subject: Re: Network Functions Check out the VxWorks man page for hostLib. Specifically the functions hostGetByName() and hostGetByAddr(). -Gerardo =========================================================================== = = = = Gerardo Pardo-Castellote = email: pardo@rti.com = = Real-Time Innovations, Inc. = Phone: (408) 720-8312 = = 155A Moffett Park Drive, Suite 111 = Fax: (408) 734-5009 = = Sunnyvale, CA 94089 = http://www.rti.com = = = = =========================================================================== >> I have some questions maybe someone out there can answer. >> In C there is a function call gethostent does vxworks have >> something like this ? >> I have looked at various host & inet functions. >> I need to get the aliases for all the machine and filter >> out the one's which I don't need. >> Then I want to convert all the alaises into the IP-Addresses. >> I've done this in C, however Vxworks does not have the >> functionality to get all the hostnames by there alaises >> and convert these into the IP-addesses. From mschnarc@madge.com Thu Jul 18 00:21:38 1996 From: "Miki Schnarch ESD-Tel" Date: Thu Jul 18 00:21:41 PDT 1996 Subject: vxworks global standard input/output Hello, I have tried to change the assignment of the global standard input/output to a a telnet file descriptor by using the ioGlobalStdSet routine. I did it in order to redirect globally all the input/output to the telnet port. But, somehow it is not working. When i did it for a specific task by using the ioTaskStdSet routine everything went OK and i got the information written from that specific task. I would be appreciated for any idea of how this can be fixed I would like also to know if the ioGlobalStdSet should effect also existing tasks or only new tasks that were created after the ioGlobalStdSet command. (I'm using vxWorks 5.2 for i960) thanks, miki. ======================================================= Miki Schnarch Email : mschnarc@madge.com Enterprise Systems Division, Voice : 972-3-6457669 Madge Networks (Israel) Ltd. Fax : 972-3-6487146 Atidim Technology Park, Bldg. 3, Tel-Aviv 61131, Israel. ======================================================= From student@vs-ulm.dasa.de Thu Jul 18 00:49:13 1996 From: Sparky Date: Thu Jul 18 00:49:18 PDT 1996 Subject: How to use virtual consoles (new user) I am using tornado 1.0, VxWorks 5.3 and an Eltec 17-5 board. My terminal output does not get any info so I wish to redirect the output to a virtual console. I have tried adding the -C flag to the tgtsvr command, nothing. I have used the I/O redirection on Shell Start up Tcl script pg 217. Still nothing. HELP _______________________________________________________________________________ |\/\/\/| | | | | | (o)(o) C _) | ,___| | / /____\ / \ Sparky Jonathan D Heath VS2E22 Daimler-Benz Aerospace AG Woerthstrasse 85 89077 Ulm Germany HTTP://www.uni-ulm.de/~s_sthiam/heath/index.html E-mail: student@vs-ulm.dasa.de _______________________________________________________________________________ From daemon@csg.lbl.gov Thu Jul 18 04:00:41 1996 From: daemon@csg.lbl.gov Date: Thu Jul 18 04:00:44 PDT 1996 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Thu Jul 18 04:00:29 PDT 1996 Subject: Re: Application running "slower" in vxWorks.st mode. Help!! Subject: Re: Delays in download VxWorks symbol table using FTP Subject: FREE CODE- You can't load symbol table via Network? Subject: Re: Round-Robin Question Subject: Re: Ethernet hardware address. Subject: Re: Utils to run Checksums? Subject: Re: Round-Robin Question Subject: Re: Delays in download VxWorks symbol table using FTP Subject: NFS verison 3 Subject: Application running "slower" in vxWorks.st mode. Help!! Subject: Re: Booting i486 BSP from EPROM ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Application running "slower" in vxWorks.st mode. Help!! Date: 17 Jul 1996 12:44:33 -0700 From: kla@leland.Stanford.EDU (Earl Mitchell) Organization: Stanford University, CA 94305, USA Message-ID: <4sjfr1$ar6@elaine25.Stanford.EDU> References: In article , firstname lastname wrote: >We are experiencing a very serious problem with the vxWorks.st running on >486 target. This problem has been reported to Wind River for more than a >week. I still haven't got any clue or answer yet. The folowing is a simple >code I have "timexN"ed: > >void delay() >{ > int i; > > for(i=0; i<163; i++); >} > >It took about 55 micro sec when "vxWorks" down loaded from host. >However, when in "vxWorks.st" mode, it took 233 micro sec. > >This is just an example. The whole application ran much slower. The >performance is not acceptable for the application. >I appreciate if any one can help. The download time for vxWorks.st increased because that image links in the symbol table. I'm not sure why this would affect run time performance though ... unless subroutine calls are being done via symbol table. Your best bet is to hook up a logic analyzer and see what's going on. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Delays in download VxWorks symbol table using FTP Date: 17 Jul 1996 16:16:16 GMT From: vfranco@cp10.es.xerox.com (Victor Franco) Organization: Xerox Message-ID: <4sj3kg$pf8@news.cp10.es.xerox.com> References: Reply-To: vfranco@cp10.es.xerox.com In article , Tom Fuda writes: >I am using 486 targets (VME based single board computers from Logic Design >Group with on board Ethernet and IDE disk drive) in a Sun host environment. >I have noticed that if I try to download the VxWorks image (separate image >and symbol table) to my target using FTP it takes approximately 1 minute to >download the symbol table while the image itself is downloaded in a matter >of seconds. If I use RSH to download the symbol table from the Sun host the >transfer only takes a few seconds. This problem is observed whether my boot >host is the Sun or another VxWorks target (using either Ethernet or shared >memory network). I have also noticed that is two VxWorks targets >simultaneously attempt to download images from the same host using FTP, one >download usually fails while the other one succeeds after about a minute. >Has anyone else experienced this problem? If so, do you know of a workaround >or fix? Thanks in advance. Tom, I saved the following article from a few months back. It seems to address the problem you are experiencing. I hope this helps. Credit, of course, goes to the author, Jeffrey Szczepanski. Good luck, Victor Franco Xerox Corp. - ---- Included article ---- From js@eroch.mc.xerox.com (Jeffrey R. Szczepanski) Newsgroups: comp.os.vxworks Subject: FREE CODE- You can't load symbol table via Network? Date: Thu Apr 18 09:30:02 PDT 1996 Organization: Lawrence Berkeley Laboratory, Berkeley CA Hello All! Fix your symbol table loading problems?! There has been talk here about the problem of loading the VxWorks symbol table immediately after a network boot of VxWorks. This problem occurs because the FTP load of the VxWorks image itself conflicts with the following network load of the symbol table. The usuall symptom is a ERROR loading the symbol table after a long timeout OR a successful load after a long timeout. Neither do I find acceptable. Some have explained this as a Sun problem, since this seems to be mostly associated with Sun, as far as I have heard. Although I have heard others complain about NT which has even a longer timeout....We have Suns here. Anyway, after doing some "homework" on TCP, I believe Sun implements correctly, what is known as the TIMED_WAIT phase or state of the TCP protocol. The problem arises because of the VxWorks implementation of TCP and how they assign the port numbers on the TARGET end. The boot code's use of FTP to load the VxWorks "application" image will use ports 1024 & 1025 for the FTP transfer (Control on 1024, data on 1025). This assumes that no other connects have been made by (or with) the target at the TCP level. When VxWorks restarts on the newly loaded image, the starting port numbers are reset back to the 1024 starting number. This means that the FTP transfer that will take place to load the symbol table will, again, occur quickly thereafter on 1024 & 1025 ports on the target. At this point, the HOST system is still in a TIMED_WAIT for a virtual connect between the FTP ports on the hosts and these specific port numbers on the target. The Sun is ignoring the requests in TIMED_WAIT, as it should. To eliminate this conflict you need to simply connect from the target to the FTP server on the host FROM a different set of ports. The simplest solution is to talk to something else first to increment the port numbers on your target. The code attached below, connects first, to TCP network standard TIME port on the host first. Any valid port connect will work, you can pick your favorite!!! (Besides, FTP, obviously) To use this patch, simply call ftpFix() immediately before attempting to load the symbol table. Good Luck, Jeff - ------------------------------------------------------------------------------ Jeffrey R. Szczepanski Phone: (716) 264-6613 Xerox Engineering Systems 300 Main Street East Rochester, NY 14445 Email: js@eroch.mc.xerox.com - ------------------------------------------------------------------------------ #include "vxWorks.h" #include #include #include #include #define TCP_TIME_PORT 37 /* ********************************************************************** * * NAME: ftpFix * * DESCRIPTION: Work around the issue with timeout * on FTP data sockets. Specifically, * the reuse of TCP port #1024 is avoided * when talking to the host FTP data port. * * PARAMETERS: void * * RETURN CODES: STATUS * * NOTES: This routine is used to work around * the issue of loading the symbol table * immediately after system boot. Generally, * the load will hang because the host * system has not timed out the socket "link" * to the same VxWorks target port number. ********************************************************************** */ static STATUS ftpFix(void) { struct sockaddr_in serverAddr; /* Server Port info */ int sock; /* Socket file descriptor */ if ((sock = socket(AF_INET, SOCK_STREAM, 0)) < 0) { ERROR(); return(FAILURE); } bzero((char *) (&serverAddr), sizeof(serverAddr)); serverAddr.sin_family = AF_INET; serverAddr.sin_port = htons(TCP_TIME_PORT); serverAddr.sin_addr.s_addr = hostGetByName(SysBootParams.hostName); if (connect(sock, &serverAddr, sizeof(serverAddr)) != OK) { ERROR(); return(FAILURE); } close(sock); return(SUCCESS); } --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Round-Robin Question Date: 17 Jul 1996 12:37:31 -0700 From: kla@leland.Stanford.EDU (Earl Mitchell) Organization: Stanford University, CA 94305, USA Message-ID: <4sjfdr$anc@elaine25.Stanford.EDU> References: In article , wrote: >Anyone, > >What is the easiest way to find out of my VxWorks image has Round-Robin >scheduling enabled or not? In 5.3 try sched_getscheduler(). Look it up in the reference manual. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Ethernet hardware address. Date: 17 Jul 1996 12:26:18 -0700 From: kla@leland.Stanford.EDU (Earl Mitchell) Organization: Stanford University, CA 94305, USA Message-ID: <4sjeoq$aj3@elaine25.Stanford.EDU> References: In article , David Uhrenholdt wrote: In VxWorks 5.2awe use m interested i. >hardware address (48 bit address) of the ethernet address. > >I know you can get the address, because it shows up in >ifShow. And ifunit provides most of the information >output by ifShow, but it doesn't provide the hardware >address. We use sysEnetAddrGet but we have a custom BSP version so this routine may have been defined by the guy here who wrote the ethernet driver. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Utils to run Checksums? Date: 17 Jul 1996 12:31:33 -0700 From: kla@leland.Stanford.EDU (Earl Mitchell) Organization: Stanford University, CA 94305, USA Message-ID: <4sjf2l$am1@elaine25.Stanford.EDU> References: <4sdfoi$tsl@swlab1.msd.ray.com> In article <4sdfoi$tsl@swlab1.msd.ray.com>, Skip Rye wrote: > >We have a vxworks computer that may need to run checksums on the software. >Is there a utility to run checksums on programs? I just wrote my own. Here it is if you want to use it. /* * return 16 bit ones complement of 16 bit ones complement sum * nwords = number of 16 bit words starting from buf */ U16 checkSum(U16 *buf, int nwords) { U32 sum ; for (sum=0; nwords > 0 ; nwords--) { sum += *buf++ ; } sum = (sum >> 16) + (sum & 0xffff) ; /* add in carry */ sum += (sum >> 16) ; /* maybe one more */ return(~sum) ; } --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Round-Robin Question Date: 17 Jul 1996 13:58:05 GMT From: johncoby@ghgcorp.com (JOHN R. COBARRUVIAS) Organization: Personal Account Message-ID: References: Round-Robin scheduling is disabled by default. You can turn it on using kernelTimeSlice(). See man pages for kernelLib. (ahhh, what the heck, here is the pages for round robin: ROUND-ROBIN SCHEDULING Round-robin scheduling allows the processor to be shared fairly by all tasks of the same priority. Without round- robin scheduling, when multiple tasks of equal priority must share the processor, a single non-blocking task can usurp the processor until preempted by a task of higher priority, thus never giving the other equal-priority tasks a chance to run. Round-robin scheduling is disabled by default. It can be enabled or disabled with the routine kernelTimeSlice(), which takes a parameter for the "time slice" (or interval) that each task will be allowed to run before relinquishing the processor to another equal-priority task. If the param- eter is zero, round-robin scheduling is turned off. If round-robin scheduling is enabled and preemption is enabled for the executing task, the routine tickAnnounce() will increment the task's time-slice count. When the specified time-slice interval is completed, the counter is cleared and the task is placed at the tail of the list of tasks at its priority. New tasks joining a given priority group are placed at the tail of the group with a run-time counter ini- tialized to zero. If a higher priority task preempts a task during its time- slice, the time-slice of the preempted task count is not changed for the duration of the preemption. If preemption is disabled during round-robin scheduling, the time-slice count of the executing task is not incremented. In article , hac.es.com wrote: > Anyone, > > What is the easiest way to find out of my VxWorks image has Round-Robin > scheduling enabled or not? > > Thanx > MAPP - -- ,-_/ . .-,--. ,---. . ' | ,-. |-. ,-. `|__/ | -' ,-. |-. ,-. ,-. ,-. . . . , . ,-. ,-. | | | | | | | | \ ,. | -. | | | | ,-| | | | | | / | ,-| `-. | `-' ' ' ' ' `-' ` `' `---' `-' `-' `-^ ' ' `-' `' ' `-^ `-' / | John R. Cobarruvias johncoby@ghgcorp.com Houston, Tx `--' --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Delays in download VxWorks symbol table using FTP Date: 17 Jul 1996 22:33:07 GMT From: watkins@spk.hp.com (Brian Watkins) Organization: Hewlett-Packard, Spokane Division Message-ID: <4sjpn3$bot@hpcvsnz.cv.hp.com> References: <4sj3kg$pf8@news.cp10.es.xerox.com> I have also observed the FTP delay correctly diagnosed by Jeffrey Szczepanski. I believe his solution does not entirely fix the problem; if the target system is re-booted before the host exits the TIME_WAIT state (usually 1-2 minutes) then the load of the VxWorks image or the symbol table may be delayed. I chose a different approach. It turns out that only one "side" of a TCP connection ends up in the TIME_WAIT state when a connection is released. The other side may re-use its socket immediately. The side that ends up in the TIME_WAIT state is the first one that attempts to release the connection. If the target were always the first to release the FTP connection, the host would never enter the TIME_WAIT state. In the TCP protocol, the CLOSE operation means "I have no more data to send" (see RFC 793 sec. 3.5). Therefore, if data will never be written to a socket, it is perfectly valid to CLOSE the connection before reading the data! I modified the bootrom of my mv162-523 board by adding a CLOSE of the FTP data socket after the transfer is initiated but before the data is read as follows in the netLoad() function in bootConfig.c: ... { if (bootFtp) { if (ftpXfer (hostName, usr, passwd, "", "RETR %s", "", fileName, &errFd, &fd) == ERROR) return (ERROR); /* BDW 7-14-95 * Disable further writes on the data socket. This will * hopefully initiate the graceful close of the socket * and prevent the FTP server from entering the TCP * TIME_WAIT state. */ shutdown (fd,1); } else { bootRsh = TRUE; sprintf (command, "cat %s", fileName); fd = rcmd (hostName, RSHD, usr, usr, command, &errFd); if (fd == ERROR) return (ERROR); } } if (bootLoadModule (fd, pEntry) != OK) goto readErr; ... This gets us to the same point as Mr. Szczepanski's solution; we must also modify the VxWorks image to ensure that the sockets used for the symbol table and the startup files get CLOSED (via shutdown()) before the data is read. Unfortunately these files are read via the network file driver (netDrv) which cannot be modified (by those of us without a source code license). Luckily netDrv uses the ftpDataConnGet() function to open the FTP data connection, and this function is simple enough to reverse-engineer and modify. I have written the following module to replace ftpDataConnGet(): /**** * * $Program name: ftpPatch.c $ * * $Description: Patch for VxWorks FTP client $ * * This file contains two functions that are used to alter the behavior * of the VxWorks FTP client code to eliminate (or at least reduce) the * possibility of the FTP server data port passing through the TCP TIME_WAIT * state. This state is entered when the FTP server is the first to close * the data port. We will attempt to create a pending close using the * shutdown() function for sockets that are only opened for reading. * * $Written by: Brian Watkins $ * * $Source$ * $Locker$ * $Modified: Mon Jul 17 10:32:32 1995 (watkins) watkins@hpspk7df $ ****/ static char * const rcsid = "$Header$"; #include "vxWorks.h" #include "ftpLib.h" #include "sockLib.h" #include "unistd.h" #include "stdio.h" u_short ftpReadOnly = 0; /**********************************************************************/ static int ftpPatchDataConnGet (int dataSock) /* This function duplicates the functionality of the VxWorks * ftpDataConnGet() function, with an additional call to * shutdown() if an external variable indicates that the FTP * data socket will only be used for reading. */ { int newSock; struct sockaddr addr; int addrlen; addrlen = sizeof (addr); newSock = accept (dataSock, &addr, &addrlen); close (dataSock); if ((newSock != ERROR) && (ftpReadOnly != 0)) { /* Initiate a graceful shutdown of the socket. */ shutdown (newSock,1); } return newSock; } /**********************************************************************/ void ftpPatchInit (void) { #define JMP_CODE 0x4ef9 u_short *short_ptr; u_long *long_ptr; /* Overwrite the beginning of the original ftpDataConnGet() * function with an absolute jump to the patched routine. * WARNING: This will not work if VxWorks is running out of * ROM. */ short_ptr = (u_short *) ftpDataConnGet; *short_ptr = JMP_CODE; if (*(short_ptr++) == JMP_CODE) { long_ptr = (u_long *) short_ptr; *long_ptr = (u_long *) ftpPatchDataConnGet; puts ("FTP client patch installed"); } else { puts ("Cannot patch FTP client"); } } /**********************************************************************/ Notice that the installation function is 680x0 specific, but it can be easily adapted to other CPUs. This function also assumes that VxWorks is running in RAM; the installation will fail if VxWorks is ROM based. The installation function should also invalidate the instruction cache in case the original ftpDataConnGet() function is in the cache. The FTP modifications are completed with the following changes in the usrRoot() function in usrConfig.c: ... /* initialize network */ #ifdef INCLUDE_NET_INIT usrBootLineInit (sysStartType); /* crack the bootline */ usrNetInit (BOOT_LINE_ADRS); /* initialize network support */ ftpPatchInit (); /* install FTP client patch */ ftpReadOnly = 1; /* Mark all FTP data sockets as */ /* read-only */ #endif /* INCLUDE_NET_INIT */ ... #ifdef INCLUDE_STARTUP_SCRIPT /* run a startup script */ if (sysBootParams.startupScript [0] != EOS) usrStartupScript (sysBootParams.startupScript); #endif /* INCLUDE_STARTUP_SCRIPT */ ftpReadOnly = 0; /* restore normal FTP */ /* client operation */ shellInit (SHELL_STACK_SIZE, TRUE); /* create the shell */ /* only include the simple demo if the shell is NOT included */ #else ftpReadOnly = 0; /* restore normal FTP */ /* client operation */ #if defined(INCLUDE_DEMO) /* create demo w/o shell */ taskSpawn ("demo", 20, 0, 2000, (FUNCPTR)usrDemo, 0,0,0,0,0,0,0,0,0,0); #endif /* mips cpp no elif */ #endif /* INCLUDE_SHELL */ ... Using the modified bootrom and VxWorks image, I rarely (if ever) observe FTP delays. Feel free to play around with this technique and/or e-mail me if you have any questions. - -------------------------- Brian Watkins watkins@spk.hp.com --------------------------- Newsgroups: comp.os.vxworks Subject: NFS verison 3 Date: 16 Jul 1996 21:12:30 GMT From: ashrafis@nt.com (Siamak Razzaghe-Ashrafi) Organization: Northern Telecom Inc., Mission Park Message-ID: <4sh0ju$qk2@nrchh52.rich.nt.com> Is NFS Version 3 "Free ware" and if so were can I get it. I need to port it to VxWorks 5.2. VxWorks 5.2 comes with NFS version 2(which sucks and that is why there is version 3) has anyone done/gotten the source of a NFS version 3 implementation on any unix that I can port to vxWorks 5.2. thankx for all the help ashrafis@nt.com - -- The thing is it so I believe it is the thing. -SubZero --------------------------- Newsgroups: comp.os.vxworks Subject: Application running "slower" in vxWorks.st mode. Help!! Date: Tue, 16 Jul 1996 12:57:54 -0600 From: loginid@server.jsc.nasa.gov (firstname lastname) Organization: NASA/JSC Message-ID: We are experiencing a very serious problem with the vxWorks.st running on 486 target. This problem has been reported to Wind River for more than a week. I still haven't got any clue or answer yet. The folowing is a simple code I have "timexN"ed: void delay() { int i; for(i=0; i<163; i++); } It took about 55 micro sec when "vxWorks" down loaded from host. However, when in "vxWorks.st" mode, it took 233 micro sec. This is just an example. The whole application ran much slower. The performance is not acceptable for the application. I appreciate if any one can help. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Booting i486 BSP from EPROM Date: Thu, 18 Jul 1996 09:11:26 +0100 From: Chris Wallett Organization: Cranfield Precision Message-ID: <31EDF1AE.6ABB@mail.compulink.co.uk> References: <4sdph6$l7t$1@mhadg.production.compuserve.com> We use a 1 MB FLASH disk on ISA. It looks like a IDE drive as far as the CPU is concerned. The FLASH disk contains minimnal DOS and loads the vxWorks bootrom code via vxload.com. Regards Chris Wallett Cranfield Precision --------------------------- End of New-News digest ********************** From currie@ghgcorp.com Thu Jul 18 05:51:34 1996 From: Alan Currie Date: Thu Jul 18 05:51:36 PDT 1996 Subject: Applications running slower than expected I've seen a similar problem than Bill has seen on his 486 computer with my DY4 68040 computer. Both computers are identical and have the same rev version of the board. Build is using the same makefile, except fp2 is a make vxWorks.st and fp1 is a make vxWorks.st_rom.hex fp1 is booting via ROM with network standalone flags set and fp2 is booting over the ethernet. Following is example code comparing speed. Does anyone have any suggestions? Thanks in advance. currie: telnet fp2 Trying 139.169.227.82 ... Connected to fp2. Escape character is '^]'. -> i NAME ENTRY TID PRI STATUS PC SP ERRNO DELAY ---------- ------------ -------- --- ---------- -------- -------- ------- ----- tExcTask _excTask f7ed64 0 PEND d3dac f7eccc d0003 0 tLogTask _logTask f7c464 0 PEND d3dac f7c3c8 3d0001 0 tShell _shell f3ff64 1 READY b15fc f3fc2c 30065 0 tRlogind _rlogind f53328 2 PEND 6cc54 f53000 0 0 tTelnetd _telnetd f51484 2 PEND 6cc54 f513b4 0 0 tTelnetOutT_telnetOutTa ecd198 2 READY 6cc54 eccf00 0 0 tTelnetInTa_telnetInTas ecbc80 2 READY 6cc2c ecb988 0 0 tRdbTask _rdbTask f4ea54 20 PEND 6cc54 f4e934 d0003 0 tNetTask _netTask f76168 50 READY afaf4 f76100 0 0 tFtpdTask _ftpdTask f4c59c 55 PEND 6cc54 f4c4e0 0 0 tPortmapd _portmapd f4ff6c 100 PEND 6cc54 f4fe38 16 0 value = 0 = 0x0 -> cd "/u4/currie/performance" value = 0 = 0x0 -> ld ls performance.c performance.o value = 0 = 0x0 -> ld perform value = 3 = 0x3 -> timexN perform timex: 1250 reps, time per rep = 966 +/- 16 (1%) microsecs value = 59 = 0x3b = ';' -> logout Connection closed by foreign host. currie: telnet fp1 Trying 139.169.227.73 ... Connected to fp1. Escape character is '^]'. -> i NAME ENTRY TID PRI STATUS PC SP ERRNO DELAY ---------- ------------ -------- --- ---------- -------- -------- ------- ----- tExcTask _excTask f7ed64 0 PEND d2928 f7eccc d0003 0 tLogTask _logTask f7c464 0 PEND d2928 f7c3c8 0 0 tShell _shell f41f88 1 READY b0178 f41c50 dd 0 tRlogind _rlogind f5534c 2 PEND 6b7d0 f55024 0 0 tTelnetd _telnetd f534a8 2 PEND 6b7d0 f533d8 0 0 tTelnetOutT_telnetOutTa f3eb20 2 READY 6b7d0 f3e888 0 0 tTelnetInTa_telnetInTas f3d608 2 READY 6b7a8 f3d310 0 0 tRdbTask _rdbTask f50a78 20 PEND 6b7d0 f50958 d0003 0 tNetTask _netTask f779a8 50 READY ae670 f77940 0 0 tFtpdTask _ftpdTask f4e5c0 55 PEND 6b7d0 f4e504 0 0 tPortmapd _portmapd f51f90 100 PEND 6b7d0 f51e5c 16 0 value = 0 = 0x0 -> cd "/u4/currie/performance" value = 0 = 0x0 -> ld timexN perform timex: 500 reps, time per rep = 1939 +/- 40 (2%) microsecs value = 59 = 0x3b = ';' -> From getker@sentientnet.com Thu Jul 18 06:45:37 1996 From: getker@sentientnet.com (Jim Getker) Date: Thu Jul 18 06:45:39 PDT 1996 Subject: Re: Application running "slower" in vxWorks.st mode. Help!! In article , firstname lastname wrote: >We are experiencing a very serious problem with the vxWorks.st running on >486 target. This problem has been reported to Wind River for more than a >week. I still haven't got any clue or answer yet. The folowing is a simple >code I have "timexN"ed: > >void delay() >{ > int i; > > for(i=0; i<163; i++); >} > >It took about 55 micro sec when "vxWorks" down loaded from host. >However, when in "vxWorks.st" mode, it took 233 micro sec. > >This is just an example. The whole application ran much slower. The >performance is not acceptable for the application. >I appreciate if any one can help. Is the standalone version of vxWorks or your code running out of ROM now? This might make a significant difference in the timing. Jim From rac@gti.ssr.upm.es Thu Jul 18 07:27:06 1996 From: rac@gti.ssr.upm.es (Roberto Artola Caballero) Date: Thu Jul 18 07:27:08 PDT 1996 Subject: Problem with floating-point operations Hello, We work with two Motorola MVME 1600-001 Power PC 604 boards in a single rack. We use VxWorks 5.2 beta for Power PC, the Kernel version is 2.4 and the BSP version is 1.0, the boards boot from a Sun 4m under SunOS 4.1.3. We are running a task which makes operations in floating-point. The problem is that the output of the program is different in subsequent executions. We get a little error of about a one per cent between different executions, i.e. the program execution is not deterministic. We think is not a problem of the code because we run also this program with exactly the same source code in a HP workstation and the output of the program is coherent and it's always the same in several executions (we executed it about 430 times). We could understand that the output of the program was different between the execution in HP and the execution in VxWorks; but we don't understand how the output of subsequent executions in VxWorks are different. Does anybody know about this type of problem? Could it be a problem with the FPU?, We use the option VX_FP_TASK for enabling the floating-point coprocessor support. Does anybody know if there is any problem like this with the management of the FPU by VxWorks. Thanks in advance, Roberto Artola ___________________________________________________________________________ Juan Roberto Artola Caballero _________________ / __________ ___ / / Grupo de Tratamiento de Imagenes / / ____ / / E.T.S. Ing. Telecomunicacisn (C-307) / / __ / / / / / Universidad Politecnica de Madrid / /____/ / / / / / 28040 Madrid - SPAIN /______ / / / / / tel: +34.1.336.7353 / / fax: +34.1.336.7350 e-mail: rac@gti.ssr.upm.es mr.chip@idg.encomix.com ___________________________________________________________________________ From eran@scorpio.com Thu Jul 18 09:22:38 1996 From: eran@scorpio.com (Eran Netanel) Date: Thu Jul 18 09:22:40 PDT 1996 Subject: looking for SNMP agent I'm looking for an extensible SNMP agent for VxWorks 5.2 running on SPARC, We are looking for the following features (preferably all): a. encoding/decoding of SNMP messages - V1.0 and/or V2.0 b. sources c. porting to VxWorks (would be nice 8-)) d. easy hook up of private SNMP MIB(s) Please e-mail or post article. thank you Eran From NEMMARTIN@aol.com Thu Jul 18 14:03:55 1996 From: NEMMARTIN@aol.com Date: Thu Jul 18 14:03:57 PDT 1996 Subject: vx works contractual assignments in CT I have an immediate need for 3 VXWORKS consultants for assignments that will continue at least 6 months in CT. If you are available for an assignment please contact me: EVELYN GUR SENIOR ACCOUNT MANAGER NEMETH/MARTIN CONSULTING 203-830-6300 From aisie.aisinc.com!aisinc.com!pew@msen.com Thu Jul 18 14:32:53 1996 From: pew@aisinc.com (Paul E. Wilt) Date: Thu Jul 18 14:32:55 PDT 1996 Subject: Re: Problem with floating-point operations Roberto: > We think is not a problem of the code because we run also this program > with exactly the same source code in a HP workstation and the output of the > program is coherent and it's always the same in several executions (we > executed it about 430 times). > > We could understand that the output of the program was different > between the execution in HP and the execution in VxWorks; but we don't > understand how the output of subsequent executions in VxWorks are different. > > Does anybody know about this type of problem? Perhaps this is a problem caused by global variables. Under Unix each time you execute an application, the initial values are always reset to a sane state. With vxWorks unless you download the application between different executions your global variables are not reset. Try downloading your code, between successive executions and compare the results. If the results are the same after the each first run then I would begin looking through your code for globals which need to be explicitly initialized before each run. Hope this helps! Paul -- ============================================================================= Paul Wilt, Sr Software Engineer (313) 332-7044 FAX : (313) 332-7077 Applied Intelligent Systems, Inc. pew@aisinc.com 110 Parkland Plaza Ann Arbor, MI 48103 ============================================================================= From student@vs-ulm.dasa.de Fri Jul 19 00:13:09 1996 From: Sparky Date: Fri Jul 19 00:13:11 PDT 1996 Subject: Re: Virtual IO in Tornado At 16:53 18.07.96 BST, you wrote: >Hi Jonathan, > >I think I can help with your problem. To redirect the output you need to >put the tcl you described in the directory .wind as file windsh.tcl. >When you start the target server from the launcher you should get an empty >virtual console window. To cause the IO to be redirected you then open a >shell from the launcher having selected your new target server. As the >shell starts up it looks for the .wind/windsh.tcl file and executes it. > >I enclose our copy of this file which has been working well for a few months >now, so much so that I couldn't remember how it worked. Incidentaly if you >don't have a virtual console (-V) option set then the script will direct >the IO to the shell window. > >Windsh.tcl >*********(Cut Here)******************** ># Set stdin, stdout, stderr to use virtual io >if { [shParse {tstz = open("/vio/0",2,0)}] != -1 } { > shParse {vf0 = tstz}; > shParse {ioGlobalStdSet (0,vf0)} ; > shParse {ioGlobalStdSet (1,vf0)} ; > shParse {ioGlobalStdSet (2,vf0)} ; > shParse {logFdAdd (vf0)} ; > shParse {printf("Std I/O set here!\n")} >} else { > shParse {printf("Std I/O unchanged.\n")} >} >*********(Cut Here)******************** > Thanks for the help, but that doesn't work either on my system. I think it is the Target Board. I don't get any virtual consoles starting up, or any thing in my shell window. My launch command is: tgtsvr.exe 149.207.86.6 -n vxworks -V -Bt 10 -C -Br 2 -Bd D:\TORNADO\log.txt You state that the (-V) is for virtual console, however I beleive that to be the (-C) switch. (-V) is the verbose switch. Never the less I have both. Anything wrong with my launch command? _______________________________________________________________________________ |\/\/\/| | | | | | (o)(o) C _) | ,___| | / /____\ / \ Sparky Jonathan D Heath VS2E22 Daimler-Benz Aerospace AG Woerthstrasse 85 89077 Ulm Germany HTTP://www.uni-ulm.de/~s_sthiam/heath/index.html E-mail: student@vs-ulm.dasa.de _______________________________________________________________________________ From daemon@csg.lbl.gov Fri Jul 19 04:00:35 1996 From: daemon@csg.lbl.gov Date: Fri Jul 19 04:00:37 PDT 1996 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Fri Jul 19 04:00:30 PDT 1996 Subject: Re: Delays in download VxWorks symbol table using FTP Subject: Re: [Q] tornado clock question Subject: Re: Delays in download VxWorks symbol table using FTP Subject: Re: Application running "slower" in vxWorks.st mode. Help!! Subject: Need tool to translate symbolic offset in assembly operand Subject: [Q]:Programming flask ---MVME162 vs. MVME162FX Subject: To get gw=... parameter from bootptab file Subject: Propietary shell(s) Subject: "Stealing" the console via telnet when no shell present Subject: VADSWorks 3.1 vs VADSWorks 3.0.0 ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Delays in download VxWorks symbol table using FTP Date: 18 Jul 1996 07:58:32 -0400 From: skb@frc.ri.cmu.edu (Scott K Boehmke) Organization: Field Robotics Center Message-ID: References: <4sj3kg$pf8@news.cp10.es.xerox.com> <4sjpn3$bot@hpcvsnz.cv.hp.com> Reply-To: skb@cmu.edu (Scott K Boehmke) Sender: skb@zinc.rec.ri.cmu.edu I too have experienced similar SymbolTable load delays, but I am using rsh to download... do your explanations hold water for such a case? I've found it to be somewhat random when trying to determine exactly when the board will enter the delay... it usually occurs exactly when I need the board the most. I scramble to figure out what's wrong, and eventually it works fine on its own. Watching the system with "snoop" shows that the VxWorks target continues to request things from the host, but that the host ignores (doesn't reply) until a timeout has occured. Scott - -- Scott Boehmke (412) 268-5679 Voice Research Engineer (412) 268-5895 Fax Field Robotics Center Carnegie Mellon University skb@cmu.edu 5000 Forbes Ave Pittsburgh, PA 15213 --------------------------- Newsgroups: comp.os.vxworks Subject: Re: [Q] tornado clock question Date: 18 Jul 1996 13:30:37 GMT From: tfm@sei.cmu.edu (Theodore Marz) Organization: The Software Engineering Institute Message-ID: <4sle9t$2fm@news.sei.cmu.edu> References: <31EE238C.3E8@ccii.co.za> In article <31EE238C.3E8@ccii.co.za>, Casey Crellin writes: |> Hi, |> |> Hopefully someone can throw some light on this for me. |> |> I am working with NTP on tornado/ vxWorks 5.3 |> |> Do the system calls |> 1) gettimeofday() |> 2) settimeofday() |> 3) adjtime() |> exist ? |> |> I have done the following: |> 1) grepped *.h and the libraries |> 2) looked in the subroutine and library section of the manual |> 3) looked for the calls under a different name |> 4) looked in the online help |> |> The port I have of NTP uses the calls, but this was a 5.2 port. |> |> Have these calls been POSIXED into oblivion or am I missing something |> (hopefully the latter). In fact if anyone knows of the status of these |> calls wrt POSIX 1003.1 please let me know. |> |> Any pointers appreciated. |> |> Casey |> |> -- |> Casey Crellin casey@ccii.co.za |> CCII Systems (Pty) Ltd |> ------------------------------------------------------------ I don't know the answer of how to do it in Tornado, but the POSIX operations are: int clock_gettime( clockid_t clock_id, struct timespec *tp) int clock_settime( clockid_t clock_id, const struct timespec *tp); int clock_getres ( clockid_t clock_id, struct timespec *res); one definition of clock_id is CLOCK_REALTIME a way to adjust the clock drift is: clock_setdrift(). The POSIX standard (B.14.2.1) "Interfaces were defined to support the synchronization of clocks with external time or between systems by using clock_settime() for gross adjustments in time, and by using clock_setdrift() to adjust the time precisely. When a clock needs to be set back in time, a negative drift can be used to slow down the clock so that it converges over time to the correct value without invalidating programs that expect time to be monotonically increasing. It also minimized the impact of changing the time base on armed per-process timers based on that clock." "There is not equivalent to the BSD adjtime() function in this standard. The adjtime() function could be implemented on the clock-drift functions described in this rationale. The virtual and profiling interval timing functions of BSD are not in this standard, but these could be implemented, as extensions, using new clock-id values." One problem is that I see no definition of the function clock_setdrift() in the copy of POSIX 1003.1b-1993 that I have. Hope this is some help.. Ted Marz tfm@sei.cmu.edu --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Delays in download VxWorks symbol table using FTP Date: 17 Jul 1996 12:59:41 GMT From: kowis@kowis@nsasrv.dseg.ti.com (Greg Kowis) Organization: Texas Instruments Message-ID: <4sio3t$qnc@superb.csc.ti.com> References: I've seen a similar delay using a DY4 SVME 162 booted from a Sun host. I don't have a work around yet. Greg Kowis Texas Instruments (214)952-4336 kowis@dlep1.itg.ti.com In article , fuda@norden.com says... > >I am using 486 targets (VME based single board computers from Logic Design >Group with on board Ethernet and IDE disk drive) in a Sun host environment. >I have noticed that if I try to download the VxWorks image (separate image >and symbol table) to my target using FTP it takes approximately 1 minute to >download the symbol table while the image itself is downloaded in a matter >of seconds. If I use RSH to download the symbol table from the Sun host the >transfer only takes a few seconds. This problem is observed whether my boot >host is the Sun or another VxWorks target (using either Ethernet or shared >memory network). I have also noticed that is two VxWorks targets >simultaneously attempt to download images from the same host using FTP, one >download usually fails while the other one succeeds after about a minute. >Has anyone else experienced this problem? If so, do you know of a workaround >or fix? Thanks in advance. > > > Thomas Fuda > Northrop Grumman Norden Systems > (203) 852-7991 > fuda@norden.com > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Application running "slower" in vxWorks.st mode. Help!! Date: 18 Jul 1996 18:27:04 GMT From: bowman@waterloo.hp.com (Don Bowman) Organization: H-P Panacom Div, Waterloo, ON Canada Message-ID: <4slvlo$mpq@hppadbk.waterloo.hp.com> References: firstname lastname (loginid@server.jsc.nasa.gov) wrote: : We are experiencing a very serious problem with the vxWorks.st running on : 486 target. This problem has been reported to Wind River for more than a : week. I still haven't got any clue or answer yet. The folowing is a simple : code I have "timexN"ed: : void delay() : { : int i; : : for(i=0; i<163; i++); : } : It took about 55 micro sec when "vxWorks" down loaded from host. : However, when in "vxWorks.st" mode, it took 233 micro sec. : This is just an example. The whole application ran much slower. The : performance is not acceptable for the application. : I appreciate if any one can help. How about this... The symbol table is quite different in the .st image. I bet a large portion of the time is due to the symbol lookup. Try using 'lkup' to find the symbol, and timex on the hex-address instead of the name. Or better, print the time in the loop instead of using an external influence. - -- bowman@waterloo.hp.com ** 519-883-3019 ** http://www.waterloo.hp.com/~bowman --------------------------- Newsgroups: comp.os.vxworks Subject: Need tool to translate symbolic offset in assembly operand Date: 18 Jul 1996 19:38:18 GMT From: egbert@torch.timeplex.com (Stephen L. Egbert) Organization: ascom Timplex, Inc. Keywords: assembly operand symbol Message-ID: <4sm3ra$pi7@bigguy.timeplex.com> Reply-To: egbert@torch.timeplex.com Does anyone know of any extended preprocessor (GNU cpp or otherwise) that can translate the C-styled structure offset for uses with assembly code using symbolic offset. At the moment, only #define can replace any matching symbolic offset in a typical assembly opcode/operand(s). The idea is to maintain a single include file for uses for both C and assembly using GNU tools. Steve - --- Stephen (Steve) L. Egbert Work: (214) 340-3300 x2123 ascom Timeplex Fax: (214) 341-3654 Dallas, Texas 75240 Net: egbert_steve@timeplex.com Pager: (214) 451-0437 --------------------------- Newsgroups: comp.os.vxworks Subject: [Q]:Programming flask ---MVME162 vs. MVME162FX Date: Wed, 17 Jul 1996 00:19:48 GMT From: Philip.Wiebe@newnes.com (Philip J. Wiebe) Organization: MIND LINK! - British Columbia, Canada Message-ID: <4shbl0$89m@fountain.mindlink.net> Is there any difference in programming the flash between the MVME162 and the MVME162FX series? I tried programming the FX in a similar fashion as the older 162s and all different scenerios didn't work. The commands used were: - ------- 162-Bug>lo 0 10000 # 'send text file bootrom.hex fom host' 162-Bug>pflash 10000 30fff ffa00000 - ---- One difference I noticed is that the FX series used more blocks. I.E. the response to the pflash command was: Source Starting/Ending Addresses =00010000/00030FFF Destination Starting/Ending =FFA00000/FFA20FFF Number of Effective Bytes =00021000 (&135168) Program FLASH Memory(Y/N)? y Erasing Block Number =00 ($FFA00000) Erasing Block Number =01 ($FFA10000) Erasing Block Number =02 ($FFA20000) Programming Block Number =00 ($FFA00000) Programming Block Number =01 ($FFA10000) Programming Block Number =02 ($FFA20000) FLASH Memory Programming Complete However, for the older series the response is: Source Starting/Ending Addresses =00010000/00030FFF Destination Starting/Ending =FFA00000/FFA20FFF Number of Effective Bytes =00021000 (&135168) Program FLASH Memory(Y/N)? y Erasing Block Number =00 ($FFA00000) Programming Block Number =00 ($FFA00000) FLASH Memory Programming Complete Phil. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Philip J. Wiebe "It has yet to be proven that intelligence Software Engineer has any survival value." Newnes Machine Ltd. - Arthur C. Clarke P.O.Box 8, 3550-45th Street S.E. Salmon Arm, BC Canada email: Philip.Wiebe@newnes.com ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ --------------------------- Newsgroups: comp.os.vxworks Subject: To get gw=... parameter from bootptab file Date: 18 Jul 1996 13:08:36 -0700 From: mitanu@synopsys.com (Mitanu Paul) Organization: Synopsys, Inc. Keywords: bootptab, gw parameter, bootpMsgSend, VxWorks Message-ID: <4sm5k4$of8@neemit.synopsys.com> Sender: mitanu@synopsys.com Hi, I would appreciate if someone can help me with the use of the bootpMsgSend() function. I wanted to obtain the gw=... parameter from the bootptab file in the set_inet_from_bootp() function (usrConfig.c file). I added the gw=... parameter in the bootptab file and when I call bootpParamsGet(), I see (using a network snooping software) that the gw parameter is sent in the BOOTP reply packet. Since bootpParamsGet() does not provide me with a way to get the gw value, I thought I might use the bootpMsgSend() function to obtain it. However, when I use it, the network snooping utility complains about "Unrecognized cookie: 0x434d5500" in the BOOTP reply packet. Here is the code snippet: { BOOTP_MSG bootpMsg; unsigned long bcast = 0; /* tried 0xFFFFFFFF, but same results */ bzero(&bootpMsg, sizeof(bootpMsg)); bootpMsgSend(netDev, &bcast, 67, &bootpMsg, 600); /* returns OK */ } The [cysg]iaddr fields in bootpMsg are filled correctly. The bp_vend field says "CMU". My question is: am I doing the right thing with bootpMsgSend()? I guess the ultimate question is whether there is any way to obtain the gw bootptab parameter using some VxWorks calls? I am using VxWorks 5.1.1 (Force 5V BSP). Thanks. - -- Mitanu Paul (mitanu@synopsys.com) --------------------------- Newsgroups: comp.os.vxworks Subject: Propietary shell(s) Date: Mon, 15 Jul 1996 11:35:05 GMT From: lyorg@Scorpio.Com (Lyor Goldstein) Organization: Scorpio Communications Ltd. Message-ID: Reply-To: lyorg@Scorpio.Com Sender: news@Scorpio.Com Has any one had experience with writing a propietary shell which would replace the standard VxWorks shell ? I would appreciate pointers and/or example source code on this, as well as known problems (possibly how to overcome them...8-)). Thank you - --- ************************************************************* * Lyor Goldstein. Email : lyorg@scorpio.com * * Scorpio Communications Phone : 972-3-5339654 ext 115 * * Software group Fax : 972-3-5339518 * * 1c Yoni Netanyahu st. * * Or Yehuda 60250 * * P.O.Box 564, Israel * ************************************************************* --------------------------- Newsgroups: comp.os.vxworks Subject: "Stealing" the console via telnet when no shell present Date: Mon, 15 Jul 1996 11:29:31 GMT From: lyorg@Scorpio.Com (Lyor Goldstein) Organization: Scorpio Communications Ltd. Message-ID: Reply-To: lyorg@Scorpio.Com Sender: news@Scorpio.Com I am looking for a way to "steal" the console of my VxWorks target by telnet-ing to it. One problem is that we are not running the shell because it competes with our console I/O task and intercepts the input intended for it. I am looking for either: a. a way to run a routine AUTOMATICALLY by the shell once it starts up or b. a way to re-direct all console I/O to a socket created on the telnet port NOTE: I have played a bit with this and am no novice, so I would appreciate example source code which does exactly this, and not general pointers... Thank you, - --- ************************************************************* * Lyor Goldstein. Email : lyorg@scorpio.com * * Scorpio Communications Phone : 972-3-5339654 ext 115 * * Software group Fax : 972-3-5339518 * * 1c Yoni Netanyahu st. * * Or Yehuda 60250 * * P.O.Box 564, Israel * ************************************************************* --------------------------- Newsgroups: comp.os.vxworks,comp.lang.ada Subject: VADSWorks 3.1 vs VADSWorks 3.0.0 Date: Thu, 18 Jul 1996 22:16:22 -0400 From: Paul McDonald Organization: AW Consulting Group Message-ID: <31EEEFF6.EE8@awcg.com> We have upgraded from using VADSWorks 3.0.0 to using VADSWorks 3.1. The script file we use to generate a ROMable image is now producing multiple defines and one undefined reference (script file stops). Rational says in their 3.1 documentation that either VxWorks 5.1.1 (what we have now) or VxWorks 5.2 will work. I'm trying to avoid upgrading to VxWorks 5.2 for schedule reasons, but I will if it fixes these problems. Any suggestions? Paul McDonald mcdonald@awcg.com Example follows: ld68k -o tmp.1 -X -r dataSegPad.o sysALib.o sysLib.o tyCoDrv.o if_ips.o if_cs220.o ncr5390Lib.o cs220_util.o /opt/rom31/hello.o tj_utl_c.obj tj_utl_d.obj \ usrConfig_st.o version.o /lor/VADS040/vxworks/lib/libMC68040gnudrv.a /lor/VADS040/vxworks/lib/libMC68040gnuvx.a windALib.o(.data+4): multiple definition of `STACK_LIMIT (.data)' v_i_mutex..NLPSa_strings..NLPS(.bss+128): first seen here windlib..NLPB(.text+a0c): multiple definition of `INT_IMAGE (.text)' more_memory..NLPBex_info..NLPB(.text+2c70): first seen here ... windlib..NLPB(.bss+c): multiple definition of `_STOP (.bss)' v_i_mutex..NLPSa_strings..NLPS(.bss+12c): first seen here cp -p tmp.1 tmp.o true tmp.o VX_CPU_FAMILY=68k VX_HOST_TYPE=sun4 /lor/VADS040/vxworks/bin/sun4/makeSymTbl tmp.o >symTbl.c sed -f editsymTbl symTbl.c > tmp.2 mv tmp.2 symTbl.c chmod g+w symTbl.c cc68k -c -O -fvolatile -fno-builtin -m68040 -m68040-only -Wall -I/h -I./src/drv/ - -I. -I/lor/VADS040/vxworks/config/all/ -I/lor/VADS040/vxworks/h -I/lor/VADS040/vxworks/h/ - -I/lor/VADS040/vxworks/config/romtest -I/lor/VADS040/vxworks/config/romtest/h - -I/lor/VADS040/vxworks/src/config/ -I/lor/VADS040/vxworks/src/drv/ - -I/lor/VADS040/vxworks/config/romtest/src/drv/ -ansi -pedantic -pipe -nostdinc - -DCPU=MC68040 -DVADSWORKS symTbl.c ld68k -o tmp.2 -X -N -M -e _sysInit \ -Ttext 00050000 tmp.1 symTbl.o ts_term..NLPBwindlib..NLPB(.text+c142): undefined reference to `_pTaskLastFpTcb' --------------------------- End of New-News digest ********************** From getker@sentientnet.com Fri Jul 19 06:21:48 1996 From: getker@sentientnet.com (Jim Getker) Date: Fri Jul 19 06:21:50 PDT 1996 Subject: i960 CF Cache Library Hi, I am doing a BSP port of vxWorks 5.2 for an i960CF processor. The 5.2 vxWorks only has a cache library for the Jx processor. Does anyone know if there is a set of CF cache control functions out there? Thanks, Jim ------ Jim Getker Sentient Networks Inc. 941-371-5454 ext. 30 getker@sentientnet.com From froeber@bbn.com Fri Jul 19 07:03:56 1996 From: Fred Roeber Date: Fri Jul 19 07:03:58 PDT 1996 Subject: Subject: Re: [Q] tornado clock question In article <31EE238C.3E8@ccii.co.za>, Casey Crellin writes: |> Hi, |> |> Hopefully someone can throw some light on this for me. |> |> I am working with NTP on tornado/ vxWorks 5.3 |> |> Do the system calls |> 1) gettimeofday() |> 2) settimeofday() |> 3) adjtime() |> exist ? Theodore Marz then responded: > I don't know the answer of how to do it in Tornado, but the POSIX > operations are: > int clock_gettime( clockid_t clock_id, struct timespec *tp) > int clock_settime( clockid_t clock_id, const struct timespec *tp); > int clock_getres ( clockid_t clock_id, struct timespec *res); > > one definition of clock_id is CLOCK_REALTIME > > a way to adjust the clock drift is: > clock_setdrift(). The POSIX timer support in VxWorks won't really do what Casey wants. It is really just a pretty thin layer over the standard VxWorks tickLib timer support. In fact, if you check the man page for clockLib, you see that there isn't even any clock_setres function. There is, however, an implementation of the BSD functions in the NTP port to VxWorks that Bob Herlien did quite a while ago. There was a usrTime library he did that supported them if my memory serves me. Check the NTP files in the VxWorks archive for the details. As I remember, the adjtime only provided very course adjustment; making it work better was something that still needed to be done. Fred -- | Fred J Roeber, BBN Systems & Technologies | | 50 Enterprise Place Middletown, RI 02842-5202 | | froeber@bbn.com 401-849-2543 (X48) | | See http://www.bbn.com/tv for TraceMaker info | From njsowko@sandia.gov Fri Jul 19 10:51:47 1996 From: "Nick J. Sowko" Date: Fri Jul 19 10:51:49 PDT 1996 Subject: Re[2]: Problem with floating-point operations > Submitted-by aisie.aisinc.com!aisinc.com!pew@msen.com Thu Jul 18 14:32:53 > 1996 > Submitted-by: pew@aisinc.com (Paul E. Wilt) > > Roberto: snip > > > > We could understand that the output of the program was different > > between the execution in HP and the execution in VxWorks; but we don't > > understand how the output of subsequent executions in VxWorks are > different. > > > > Does anybody know about this type of problem? > Perhaps this is a problem caused by global variables. Under Unix each time > you execute an application, the initial values are always reset to a sane > state. With vxWorks unless you download the application between different > executions your global variables are not reset. Try downloading your code, > between successive executions and compare the results. If the results are > the same after the each first run then I would begin looking through your > code for globals which need to be explicitly initialized before each run. > > Hope this helps! > Paul > -- I have found that the problem (different outputs of subsequent executions in VxWorks) also occurs when a variable is initialized in its declaration. Example: int i = 0; /* doomed to fail */ /* a better way */ int j; { j = 0; Hope this also helps, Nick Sowko Sandia National Laboratory 505-844-8019 From daemon@csg.lbl.gov Sat Jul 20 04:00:25 1996 From: daemon@csg.lbl.gov Date: Sat Jul 20 04:00:28 PDT 1996 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sat Jul 20 04:00:21 PDT 1996 Subject: TCL/TK & Tornado Subject: How is support from VxWorks? Subject: VxWorks Subject: Re: How to use virtual consoles (new user) Subject: Re: Delays in download VxWorks symbol table using FTP Subject: VxWorks 5.2 uncompress Subject: Help: Can't get Agent Online Subject: Re: Application running "slower" in vxWorks.st mode. Help!! ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: TCL/TK & Tornado Date: 18 Jul 1996 22:11:30 GMT From: llabar@qcktrn.com ( Leonard Labar ) Organization: Quickturn Design Systems, Inc., Mt. View, CA Message-ID: <4smcqi$g5f@newshost.quickturn.com> Reply-To: llabar@qcktrn.com Is anyone using the above combination or tried yet? If so, drop me a line. - --- - -Len LaBar Bicycled: CA, OR, MO, WA, IA, NE, Canada. http://www.geocities.com/Colosseum/2847 Len's Home Page http://www.geocities.com/Colosseum/2847/ICCC.html ICCC Web Page --------------------------- Newsgroups: comp.os.vxworks Subject: How is support from VxWorks? Date: Fri, 19 Jul 1996 09:49:15 +0200 From: Hartmut Pohl Organization: Hewlett-Packard Co. Message-ID: <31EF3DFB.303@bbn.hp.com> Hi there! We are planning to use VxWorks 5.3 on a i386 target and HP/UX 9.05 as development platform. What we would like to know is your statement about the following questions: - How is the support from Wind River Systems - Do you have a support contract? - Do you get support from US or a lokal service partner? - Any other statement about VxWorks. Any help would be greatly appreciated. Best regards, Hartmut - -- Hartmut Pohl HEWLETT-PACKARD GmbH email Hartmut_Pohl@bbn.hp.com BID phone +49-7031-14-2723 Herrenberger Strasse 130 fax +49-7031-14-7023 71034 Boeblingen telnet 778-2723 Germany http://www.uni-koblenz.de/~pohl --------------------------- Newsgroups: comp.os.vxworks Subject: VxWorks Date: 15 Jul 1996 20:12:11 GMT From: marsico@bnr.ca (David Marsico) Organization: Bell Northern Research Message-ID: <4se8mr$jkc@nrtphba6.bnr.ca> Is anybody using the GNU compiler for MPC 860? Please drop me a line... David Marsico marsico@nortel.ca Nortel --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to use virtual consoles (new user) Date: Fri, 19 Jul 1996 14:02:35 GMT From: sighg@sesig.abb.se (Henry Gessau) Organization: Adtranz Signal, Stockholm Message-ID: <31ef9368.23261458@news.seinf.abb.se> References: <1.5.4.32.19960718074527.002cf058@magic.vs-ulm.dasa.de> Reply-To: sighg@sesig.abb.se On Thu, 18 Jul 1996 09:45:27 +0200, Sparky wrote: > I am using tornado 1.0, VxWorks 5.3 and an Eltec 17-5 board. > > My terminal output does not get any info so I wish to redirect the output to > a virtual console. > > I have tried adding the -C flag to the tgtsvr command, nothing. > I have used the I/O redirection on Shell Start up Tcl script pg 217. Still > nothing. I don't know anything about the board you're using but it sounds strange to me that you don't get any output on your terminal. On my system (custom 68360 board), the terminal is connected to the first serial port. All BootRom output, VxWorks startup output, and all stdio goes there until redirected elsewhere. I also use the Tcl script on pg 217 (Tornado for Windows) and this redirects output to the target server's console when I start the target server. Tornado and VxWorks talk to each other via the LAN (ethernet). I don't know if my ramblings have helped you, but if you don't have a default terminal where all startup information is displayed, then I would suspect a problem with your BSP? - -- -------+ Henry Gessau System Designer/Software Engineer/Programmer/Hacker | | ABB Daimler-Benz Transportation Signal, Stockholm, Sweden | +------- All views and opinions expressed in this message are strictly my own --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Delays in download VxWorks symbol table using FTP Date: Mon, 15 Jul 1996 11:47:03 -0700 From: Greg Rasche Organization: General Instrument Message-ID: <31EA9227.A6@po3.gi.com> References: Tom, I've experienced similar problems with i960 targets on the VME bus loading from a SCSI disk which resides on one of the targets...but have not seen the problem when loading from a Sun Host's disk. We believe that our problem is related to the FTP Server running on one of the VxWorks targets. WRS is currently investigating solving this problem. Greg Rasche grasche@procy.gi.com Tom Fuda wrote: > > I am using 486 targets (VME based single board computers from Logic Design > Group with on board Ethernet and IDE disk drive) in a Sun host environment. > I have noticed that if I try to download the VxWorks image (separate image > and symbol table) to my target using FTP it takes approximately 1 minute to > download the symbol table while the image itself is downloaded in a matter > of seconds. If I use RSH to download the symbol table from the Sun host the > transfer only takes a few seconds. This problem is observed whether my boot > host is the Sun or another VxWorks target (using either Ethernet or shared > memory network). I have also noticed that is two VxWorks targets > simultaneously attempt to download images from the same host using FTP, one > download usually fails while the other one succeeds after about a minute. > Has anyone else experienced this problem? If so, do you know of a workaround > or fix? Thanks in advance. > > Thomas Fuda > Northrop Grumman Norden Systems > (203) 852-7991 > fuda@norden.com --------------------------- Newsgroups: comp.os.vxworks Subject: VxWorks 5.2 uncompress Date: Fri, 19 Jul 1996 21:35:22 GMT From: Thi D Nguyen Organization: StrataCom Corporation Message-ID: <31EFFF9A.41C67EA6@strata.com> Sender: news@strata.com I am in the process of upgrading an existing system using VxWorks 5.1 on (MIPS R3000) to 5.2. I am having trouble uncompressing the image from the resident boot. This used to work in our version using VxWorks 5.1. Looking at the logic analyzer execution trace, it looks like the uncompression was done; but the uncompress library call never returned. I'd apprieciate any help or pointer. Thanks in advance, Thi Nguyen --------------------------- Newsgroups: comp.os.vxworks Subject: Help: Can't get Agent Online Date: Fri, 19 Jul 1996 20:27:00 GMT From: np80eb@mail.telepac.pt (Pedro) Organization: TELEPAC, SA Message-ID: <31efef90.2338323@news.telepac.pt> We recently purchased a PEP VM62 Board and I would apreciate if you could help me with some problems. 1. History We purchased board with a VxWorks BSP and we got one with a OS-9 BSP. No problem, we sent it back and got a new one with VxWorks. BUT the manual weren't switched and we don't have the correct manuals (we will request them later). 2. Problem This is the procedure we followed to connect the board to a Sun workstation: a. Plug the VM62 to the VMEbus target. b. Connected a Terminal to it. c. Power on. d. In the terminal we press a key to abort the booting sequnce. e. typed the correct locations of the BSP files in the host system. f. pressed "@" g. the booting sequnce started and completed successfully. h. A big VXWORKS banner apeared on the terminal with version number (v1.3.1) and got an arrow prompt. What must I do to run the VxWorks Agent? I'm not able to run any comands on this "shell" (I don' have the manual!!!) In the VxWorks -Tornado I can't "Attatch" to the agent server. The "tgtsvr" function fails, and there seems to be some kind of VXN (or something) error. I can do a Telnet from the host to the Target and then I get the same arrow prompt. ********************** Pedro Alegria np80eb@mail.telepac.pt Porto ********************** --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Application running "slower" in vxWorks.st mode. Help!! Date: 18 Jul 1996 12:36:42 GMT From: bowman@waterloo.hp.com (Don Bowman) Organization: H-P Panacom Div, Waterloo, ON Canada Message-ID: <4slb4q$jiu@hppadbk.waterloo.hp.com> References: firstname lastname (loginid@server.jsc.nasa.gov) wrote: : We are experiencing a very serious problem with the vxWorks.st running on : 486 target. This problem has been reported to Wind River for more than a : week. I still haven't got any clue or answer yet. The folowing is a simple : code I have "timexN"ed: : void delay() : { : int i; : : for(i=0; i<163; i++); : } : It took about 55 micro sec when "vxWorks" down loaded from host. : However, when in "vxWorks.st" mode, it took 233 micro sec. : This is just an example. The whole application ran much slower. The : performance is not acceptable for the application. : I appreciate if any one can help. Well, is there any chance the optimisation level is different here? I would expect this loop to take 0us if any optimisation is on. Do an 'l' of the code in both shells to check this. Is it possible you got a different timex routine in the .st version, and that has a bug in it? Is it possible the .st version has spy enabled [or some other thing on the aux clock] that is causing more interrupts? Is it possible the .st version has a different #define for the clock frequency or something such that the real-time is off? Can you stop-watch it? ps... I can't imagine a 486 taking 55 or 233 us to run 163 increments... Something is up. - -- bowman@waterloo.hp.com ** 519-883-3019 ** http://www.waterloo.hp.com/~bowman --------------------------- End of New-News digest ********************** From daemon@csg.lbl.gov Sun Jul 21 04:00:22 1996 From: daemon@csg.lbl.gov Date: Sun Jul 21 04:00:26 PDT 1996 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sun Jul 21 04:00:17 PDT 1996 Subject: Re: Application running "slower" in vxWorks.st mode. Help!! Subject: Re: GNU Warning: trigraph Subject: [Q] tornado clock question Subject: Re: Multicast addresses Subject: upgrading to vxworks 5.2 Subject: Re: Application running "slower" in vxWorks.st mode. Help!! Subject: [Q] VxWorks target bootup problem sometimes ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Application running "slower" in vxWorks.st mode. Help!! Date: Wed, 17 Jul 1996 15:35:00 -0600 From: loginid@server.jsc.nasa.gov (Bill Ho) Organization: NASA/JSC Message-ID: References: <4sjfr1$ar6@elaine25.Stanford.EDU> In article <4sjfr1$ar6@elaine25.Stanford.EDU>, kla@leland.Stanford.EDU (Earl Mitchell) wrote: > In article , > firstname lastname wrote: > >We are experiencing a very serious problem with the vxWorks.st running on > >486 target. This problem has been reported to Wind River for more than a > >week. I still haven't got any clue or answer yet. The folowing is a simple > >code I have "timexN"ed: > > > >void delay() > >{ > > int i; > > > > for(i=0; i<163; i++); > >} > > > >It took about 55 micro sec when "vxWorks" down loaded from host. > >However, when in "vxWorks.st" mode, it took 233 micro sec. > > > >This is just an example. The whole application ran much slower. The > >performance is not acceptable for the application. > >I appreciate if any one can help. > > The download time for vxWorks.st increased because that image links > in the symbol table. I'm not sure why this would affect run time > performance though ... unless subroutine calls are being done via > symbol table. Your best bet is to hook up a logic analyzer and see > what's going on. Just clarify one thing here. " 55 microsec and 233 microsec" are not the down load time. The "55 micro sec" is the time to execute the subroutine delay() when running under "vxWorks". "233 microsec" is the time to take to run delay when running under "vxWorks.st". --------------------------- Newsgroups: comp.os.vxworks Subject: Re: GNU Warning: trigraph Date: Fri, 19 Jul 1996 16:11:36 +1000 From: Earl Chew Organization: Hewlett-Packard Co. Message-ID: <31EF2718.7C3E@hpato.aus.hp.com> References: <31E591FC@msmail.active.ch> <4segdn$d4m@elaine25.Stanford.EDU> Earl Mitchell wrote: > > In article <31E591FC@msmail.active.ch>, > Geiter Paul wrote: > > > >Hi > > > >I am using the Tornado compiler on a NT Host and get the following warning > >message: > > > > mmi_menu.h:1 warning: 1 trigraph(s) encountered > > > >Can anyone give me a hint, where to look for an error. This is a warning, and generated due to support for ANSI C trigraphs. ANSI C has defined character triples to substitute for characters that might not be available on all keyboards or character representations. The trigraphs begin with ?? with the next character indicating the actual character. This means that if you embed two question marks in a row in a string literal, an ANSI compiler will interpret that to mean that you are trying to start a trigraph sequence. This is most likely to occur in strings since you are unlikely to type two question marks in a row as part of a C statement sequence. (I might have some of the details wrong -- I don't have the standard handy at this time) Earl - -- +----------------------------------------------------------------------+ | Earl Chew Email: earl_chew@aus.hp.com | | Hewlett-Packard Australia Tel: +61 3 9210 5528 | | Australian Telecom Operation Fax: +61 3 9210 5550 | | 347 Burwood Highway GPS: 37 51'46"S 145 10'03"E | | Burwood East, Victoria 3151, Australia | +----------------------------------------------------------------------+ --------------------------- Newsgroups: comp.os.vxworks Subject: [Q] tornado clock question Date: Thu, 18 Jul 1996 13:44:12 +0200 From: Casey Crellin Organization: CCII Systems Message-ID: <31EE238C.3E8@ccii.co.za> Hi, Hopefully someone can throw some light on this for me. I am working with NTP on tornado/ vxWorks 5.3 Do the system calls 1) gettimeofday() 2) settimeofday() 3) adjtime() exist ? I have done the following: 1) grepped *.h and the libraries 2) looked in the subroutine and library section of the manual 3) looked for the calls under a different name 4) looked in the online help The port I have of NTP uses the calls, but this was a 5.2 port. Have these calls been POSIXED into oblivion or am I missing something (hopefully the latter). In fact if anyone knows of the status of these calls wrt POSIX 1003.1 please let me know. Any pointers appreciated. Casey - -- Casey Crellin casey@ccii.co.za CCII Systems (Pty) Ltd - ------------------------------------------------------------ --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Multicast addresses Date: 18 Jul 1996 03:21:06 GMT From: apaz@iis.com.br (armando paz) Organization: Inside Information Systems Message-ID: <4skaj2$jbp@ftp.iis.com.br> References: <9607120917.AA04224@vega.syswiz.it> In article <9607120917.AA04224@vega.syswiz.it>, Tiziana Balegno says: > >I am using VxWorks with the following system configuration: > >HOST Win95 >ARCH MC68020 >BSP FRC30 > >Using the Am7990 Lance Ethernet driver, I need to receive packets from the >following two different multicast addresses > >0x09, 0x00, 0x2b, 0x00, 0x00, 0x04, 0xfe >0x09, 0x00, 0x2b, 0x00, 0x00, 0x05, 0xfe > >according to the International Standard ISO 9542 (End System to Intermediate >System routeing protocol ...). >How can I set the two multicast address? >Thank you very much in advance >Regards > Configuring the ethernet adapter to handle multicasting is not possible through the functions available in the board's BSP. What I think you should do is refer to the data-sheet of the Am7990, where you'll find the information necessary to configure the registers of this chip and try to program it directly. Hope this is of some help ... Armando Paz apaz@iis.com.br --------------------------- Newsgroups: comp.os.vxworks Subject: upgrading to vxworks 5.2 Date: 19 Jul 1996 15:02:45 GMT From: don@mars.dgrc.doc.ca (Donald McLachlan) Organization: The Communications Research Centre Message-ID: <4so82l$ivi@crc-news.doc.ca> References: Reply-To: don@mars.dgrc.doc.ca I've been watching this group for a few days and have note seen a FAQ, is there one available? We have an application we are using on vxworks 5.0.1 which we intend to upgrade to use on vxworks 5.2. The 5.2 release notes say to see the VxWorks 5.1 Release Notes. Of course we don't have these. Is there an online version of these available? Thanks, Don - -- Donald McLachlan e-mail donald.mclachlan@crc.doc.ca Communications Research Centre / DRX office 613-998-2845 3701 Carling Ave., fax 613-998-9648 Ottawa, Ontario labs 613-998-2423 / 613-998-4118 K2H 8S2 home 613-599-6262 Canada --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Application running "slower" in vxWorks.st mode. Help!! Date: Thu, 18 Jul 1996 10:28:35 -0700 From: Andy Van Pelt Organization: QUALCOMM, Incorporated; San Diego, CA, USA Message-ID: References: <4sjfr1$ar6@elaine25.Stanford.EDU> Things that occur to me: Are the cache settings the same? Does the cache start out in the same condition? Is the difference in loading result in the code being loaded in a different area of memory with different behaviour? So the only difference is how the Vxworks image is loaded, the application is loaded the same in both cases? Makes me wonder if somehow it's taking VxWorks longer to "locate" the code one way than in the other. Maybe you could timexN a null routine to verify this. On 18 Jul 1996, JOHN R. COBARRUVIAS wrote: > In article , > loginid@server.jsc.nasa.gov (Bill Ho) wrote: > > > In article <4sjfr1$ar6@elaine25.Stanford.EDU>, kla@leland.Stanford.EDU > > (Earl Mitchell) wrote: > > > > > In article , > > > firstname lastname wrote: > > > >We are experiencing a very serious problem with the vxWorks.st running on > > > >486 target. This problem has been reported to Wind River for more than a > > > >week. I still haven't got any clue or answer yet. The folowing is a simple > > > >code I have "timexN"ed: > > > > > > > >void delay() > > > >{ > > > > int i; > > > > > > > > for(i=0; i<163; i++); > > > >} > > > > > > > >It took about 55 micro sec when "vxWorks" down loaded from host. > > > >However, when in "vxWorks.st" mode, it took 233 micro sec. > > > > > > > >This is just an example. The whole application ran much slower. The > > > >performance is not acceptable for the application. > > > >I appreciate if any one can help. > > > > > > The download time for vxWorks.st increased because that image links > > > in the symbol table. I'm not sure why this would affect run time > > > performance though ... unless subroutine calls are being done via > > > symbol table. Your best bet is to hook up a logic analyzer and see > > > what's going on. > > Just clarify one thing here. " 55 microsec and 233 microsec" are not the > > down load time. The "55 micro sec" is the time to execute the subroutine > > delay() when running under "vxWorks". "233 microsec" is the time to take > > to run delay when running under "vxWorks.st". > > RIGHT. I talked with Bill about this. I am stumped also. > Boot your target with vxWorks and load the symbol table and the > application, delay, runs in 55 microsec. (using timexN delay) > > BUT. Boot with vxWorks.st and the same application runs in 233 microsec. > > They have seen this on a 486 archeticture and a DY4 SVME162-163. > > I tried it on a PVM30 with run times of 1 microsec with or without the > built in symbol table. > > I am stumped. > > -- > > ,-_/ . .-,--. ,---. . > ' | ,-. |-. ,-. `|__/ | -' ,-. |-. ,-. ,-. ,-. . . . , . ,-. ,-. > | | | | | | | | \ ,. | -. | | | | ,-| | | | | | / | ,-| `-. > | `-' ' ' ' ' `-' ` `' `---' `-' `-' `-^ ' ' `-' `' ' `-^ `-' > / | John R. Cobarruvias johncoby@ghgcorp.com Houston, Tx > `--' > > - -- Andy Van Pelt Internet: avanpelt@qualcomm.com QUALCOMM Inc. Work : 619) 658-3214 --------------------------- Newsgroups: comp.unix.solaris,comp.os.vxworks Subject: [Q] VxWorks target bootup problem sometimes Date: Sat, 20 Jul 1996 17:28:41 GMT From: babru@netcom.com (Thatikunta ) Organization: NETCOM On-line Communication Services (408 261-4700 guest) Message-ID: Sender: babru@netcom20.netcom.com Hi, I have a Sun workstation with 2 ethernet interfaces. One is connected to the company network and the other to a VxWorks Target. This workstation acts as a host to the VxWorks target. Initially the target sends a arp packet, this is received by the host. Sometimes the host doesnot respond back to the target causing the target to hangup in the booting process. At this stage if I reboot the host everything becomes OK. I am running in.routed on this host as it is acting as a gateway to two networks. Secondly if I disconnect the host from the main network and isolate the network to the host and the target only, I never see any problem. In fact it improves the performance of the target. Could some body give me some hints on solving this problem. Thank You - -Babru (Rajababru Thatikunta) United Defence LP SJ babru_thatikunta@fmc.com babru@netcom.com Phone (408) 289 - 3359 --------------------------- End of New-News digest ********************** From hdang@hea.com Sun Jul 21 13:53:53 1996 From: hdang@hea.com (Huy Dang) Date: Sun Jul 21 13:53:55 PDT 1996 Subject: SPARC Emulator/Logic Analyzer Hi - We're looking for comments from people who are using or offering either a logic analyzer or an emulator on a SPARC target. We know HP has products in this area. We're interested to hear from other vendors and especially from the vxWorks community. Thanks - Huy Dang From jfd@rtp.co.uk Mon Jul 22 03:17:51 1996 From: Julian Day Date: Mon Jul 22 03:17:53 PDT 1996 Subject: Re: Application running "slower" in vxWorks.st mode. Help!! > >void delay() > >{ > > int i; > > > > for(i=0; i<163; i++); > >} > > > >It took about 55 micro sec when "vxWorks" down loaded from host. > >However, when in "vxWorks.st" mode, it took 233 micro sec. > > Have you tried compiling this using the -save-temps compiler flag or looking at the code using -> l delay to look at the code produced in the two cases. FYI -save-temps will cause the output from the pre-process and code generation phases of compilation to be written to disk, allowing you to see the assembly language generated. HTH, Julian --------------------------------------------------------------- Julian Day Real Time Products Ltd. Unit 8, Holt Court, e-mail: jfd@rtp.co.uk Aston Science Park, Jennens Road, tel: +44 121 333 6955 Birmingham. B7 4EJ. England. fax: +44 121 333 5433 --------------------------------------------------------------- From CJONES@digsys1.bwi.wec.com Mon Jul 22 07:22:49 1996 From: "CHRIS JONES" Date: Mon Jul 22 07:22:51 PDT 1996 Subject: Debugging a task at startup Does anyone have a method for attaching to a newly spawned tasked for debug purposes. By the time I get attached, the initialization code has already executed or crashed. -Thanks Chris Jones cjones@digsys1.bwi.wec.com (410) 765-7959 (410) 993-8822 FAX Northrop Grumman Electronic Sensors & Systems Division From uchenick@tis4000.com Mon Jul 22 08:41:23 1996 From: Gordon Uchenick Date: Mon Jul 22 08:41:25 PDT 1996 Subject: RE: Debugging a task at startup ---------- From: the vxWorks Users Group Exploder[SMTP:vxwexplo@lbl.gov] Sent: Monday, July 22, 1996 10:23 AM >Does anyone have a method for attaching to a newly spawned tasked for >debug purposes. By the time I get attached, the initialization code >has already executed or crashed. Use the "b" command on the console to put a breakpoint on the entry point of your task. When the breakpoint hits, attach to the task with vxgdb. After you've attached and before you do anything else, delete the console breakpoint. (I use bdall). Then you can step through the code or set breakpoints via vxgdb. HTH, Gordon Uchenick uchenick@tis4000.com From yfan@spar.ca Mon Jul 22 08:45:22 1996 From: Yibing Fan Date: Mon Jul 22 08:45:24 PDT 1996 Subject: Debugging a task at startup -Reply Hi Chris, You can suspend the newly spawned task by adding one line of code at the beginning of you task function: taskSuspend(0); After you attached to the task, resume the task from shell: taskResume(taskId) where taskId is the task number you get when you list tasks with command "i". If your other tasks have a short timeout waiting for the respond from the spawned task. You need to make them longer or add taskDelay in them or let them pending on the message from the spawned task. Hope this help. Yibing From bwenholz@pacesetter.com Mon Jul 22 09:17:44 1996 From: bwenholz@pacesetter.com (Bruce Wenholz) Date: Mon Jul 22 09:17:46 PDT 1996 Subject: Re: Debugging a task at startup )Submitted-by CJONES@digsys1.bwi.wec.com Mon Jul 22 07:22:49 1996 )Submitted-by: "CHRIS JONES" ) ) )Does anyone have a method for attaching to a newly spawned tasked for )debug purposes. By the time I get attached, the initialization code )has already executed or crashed. ) )-Thanks ) )Chris Jones )cjones@digsys1.bwi.wec.com ) )(410) 765-7959 )(410) 993-8822 FAX ) )Northrop Grumman )Electronic Sensors & Systems Division Sure, after you have loaded the file go to your shell and set a breakpoint on the entry point name for the task. Then when the task is spawned, it will break on the first instruction. You can then lookup the taskid and attach to that task in the debugger. ======================8<---work---------------------- e-mail: bwenholz@pacesetter.com (Bruce Wenholz) Pacesetter, Inc Sylmar, Ca 91342, USA ======================8<---personal------------------ e-mail: MMCo@DeltaNet.Com (Bruce Wenholz) WEB: http://www.deltanet.com/users/mmco/essay.html From burt@xylan.com Mon Jul 22 11:14:29 1996 From: burt@xylan.com (Burt Cyr) Date: Mon Jul 22 11:14:31 PDT 1996 Subject: Re: Debugging a task at startup Chris Jones writes: > Does anyone have a method for attaching to a newly spawned tasked for > debug purposes. By the time I get attached, the initialization code > has already executed or crashed. What I would do in this case is go into the tShell, set a breakpoint at the task entry before spawning it (using the shell's b command), and then spawn it without the VX_UNBREAKABLE flag set. It should get spwaned with state SUSPENDED. Then delete the breakpoint using the tshell's db or dball command. (dball is safest since you want to give vxGdb complete control over the target.) Your target will remain in state suspended. Then use the vxGdb tool to attach to it. Once attached, set a break at your init routine using vxGdb breakpoints. From the tshell, continue the task with a "c" command. The gdb tool should resume the debugging by breaking at your init routine. Hope this helps. Burt ================================================================== v _ Burt Cyr Xylan, Inc. ____(_)___ burt@xylan.com 26679 Agoura Rd. Suite 100 _-_-__-_-_-_ Calabasas, CA 91302 --_-_--_-_- Romans 5:1,2 (818) 878-4533 ================================================================== From KH@mps.ohio-state.edu Mon Jul 22 11:19:30 1996 From: "KLAUS HONSCHEID, OHIO STATE UNIVERSITY (614) 292-3287" Date: Mon Jul 22 11:19:32 PDT 1996 Subject: Serial Line Driver Hi, one of our students is designing a communications board based on Cirrus's CD1865 (software compatible to the CD1864) chip. Does anybody know of a serial line driver for this device and how and where we could get it. Thanks, Klaus Honscheid Dept. of Physics Ohio State University From jjs@mclean.sparta.com Mon Jul 22 19:56:08 1996 From: "John J. Smith" Date: Mon Jul 22 19:56:10 PDT 1996 Subject: jazz drives Hi all, Has anyone out there attempted to attach one of the Iomega Jazz drives to VxWorks? How about a Zip drive? We are currently considering one of these drives for a new project, and would like to know if it is a simple "bolt-on" exercise, or if it is a major software effort. We would also like an idea of the typical throughput rates for these drives. Any and all contributions would be greatly appreciated. Regards, _________________________________________________________ | |\ | John J Smith | Voice -> 703 448 1683 x253 | | | jjs@mclean.sparta.com | FAX -> 703 893 5494 | | | | | | Web -> http://www.mclean.sparta.com | | |_________________________________________________________| | | | | | /\\ SPARTA, Inc. | | | | ///\\ 7926 Jones Branch Drive | "Isn't Virtual | | | /////\\ Suite 900 | Reality an | | | ///// \\ McLean, VA 22102 | Oxymoron?" | | | / \\\\\ \\ 703-448-0120 | | | | \ \\\\\ // | | | | \ ////// | | | | \////// SPARTA | | | | \//// Pride In Performance | | | | \// -------------------- | | | |_________________________________________________________| | \__________________________________________________________\| From MikeMcQuade@prometheus.connectsoft.net Mon Jul 22 21:51:57 1996 From: Mike McQuade Date: Mon Jul 22 21:51:58 PDT 1996 Subject: Re: jazz drives At 07:56 PM 7/22/96 PDT, you wrote: >Submitted-by jjs@mclean.sparta.com Mon Jul 22 19:56:08 1996 >Submitted-by: "John J. Smith" > > >Hi all, > >Has anyone out there attempted to attach one of the Iomega Jazz drives to >VxWorks? How about a Zip drive? > >We are currently considering one of these drives for a new project, and would >like to know if it is a simple "bolt-on" exercise, or if it is a major >software effort. We would also like an idea of the typical throughput rates >for these drives. > Id be interested as well, we would like to use a Jazz drive on a project we have in the works Thanks Mike McQuade MikeMcQuade@connect.com From harvey@wrs.com Mon Jul 22 23:17:52 1996 From: Harvey Wong Date: Mon Jul 22 23:17:54 PDT 1996 Subject: Firewire support Dear VxWorks Users, Has anyone used VxWorks in an application using "Firewire" (the IEEE-394 fast serial digital audio-visual bus)? I'm interested in finding out what additional software is required to support this. Thanks in advance, Harvey Wong harvey@wrs.com From daemon@csg.lbl.gov Tue Jul 23 04:00:38 1996 From: daemon@csg.lbl.gov Date: Tue Jul 23 04:00:41 PDT 1996 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Tue Jul 23 04:00:35 PDT 1996 Subject: Re: How to write GNU C program to access VMEbus address under VxWork? Subject: Re: Debugging a task at startup ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to write GNU C program to access VMEbus address under VxWork? Date: 23 Jul 1996 03:21:18 GMT From: steved@pacifier.com@199.2.117.163 (Steve Doiel) Organization: Pacifier BBS, Vancouver, Wa. ((360) 693-0325) Message-ID: <4t1gfe$ece@news.pacifier.com> References: <4t0322$i59@news.ncu.edu.tw> Reply-To: steved@pacifier.com (Steve Doiel) In <4t0322$i59@news.ncu.edu.tw>, cjsjesus@ms1.hinet.net.tw (Johnson Chen) writes: >I am a beginner of VxWork operating system, and now I am starting a >new project in which we will use one 68040 CPU board to control >another speical I/O board through the writing command word into the >registers of I/O board. I can not find any function in reference >manual of VxWork5.1 about such "VMEbus address direct access" >function. Could anyone help me? >Thank! >Johnson Chen >san01@noc1.tyc.edu.tw > The first thing to check into is the function that keeps VxWorks from caching the part of memory you plan to access. Look in your board support manual for this. If you don't do this look forward to interesting behaviour. There may be many ways to access the memory, but what I have been doing follows (from memory since I don't have any source code handy). typedef short int16; typedef int int32; typedef unsigned int uInt32; typedef unsigned short uInt16 typedef volatile uInt16 uReg16; #define baseAddress ((uInt32)0x8000000) void SimpleExample( void ) { uReg16 *valPtr; valPtr = (uReg16 *)baseAddress; *valPtr = 0x0042; } I'm sure there are other ways, but this is what I have been doing. I hope this helps, Steve Doiel --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Debugging a task at startup Date: 23 Jul 1996 03:24:50 GMT From: steved@pacifier.com@199.2.117.163 (Steve Doiel) Organization: Pacifier BBS, Vancouver, Wa. ((360) 693-0325) Message-ID: <4t1gm2$ece@news.pacifier.com> References: Reply-To: steved@pacifier.com (Steve Doiel) In , "CHRIS JONES" writes: > >Does anyone have a method for attaching to a newly spawned tasked for >debug purposes. By the time I get attached, the initialization code >has already executed or crashed. > The simplest way is to put a taskSuspend(0) as the first statement in the task you are going to debug. The 0 argument causes the current task to suspend. If you want to suspend earlier than this I don't know. Steve Doiel --------------------------- End of New-News digest ********************** From daemon@csg.lbl.gov Wed Jul 24 04:00:25 1996 From: daemon@csg.lbl.gov Date: Wed Jul 24 04:00:27 PDT 1996 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Wed Jul 24 04:00:22 PDT 1996 Subject: usrTime package ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: usrTime package Date: 23 Jul 1996 18:17:54 GMT From: don@mars.dgrc.doc.ca (Donald McLachlan) Organization: The Communications Research Centre Message-ID: <4t350i$95d@crc-news.doc.ca> Reply-To: don@mars.dgrc.doc.ca We have been using the usrTime package (gettimeofday(), settimeofday(), date() etc) on vxWorks 5.0.2. We are porting our application to vxWorks 5.2 and are finding the macros for accessing the TCP (timer chip) have all changed. Has anyone out there ported the usrTime package (rtc.c in particular) to 5.2? Thanks, Don - -- Donald McLachlan e-mail donald.mclachlan@crc.doc.ca Communications Research Centre / DRX office 613-998-2845 3701 Carling Ave., fax 613-998-9648 Ottawa, Ontario labs 613-998-2423 / 613-998-4118 K2H 8S2 home 613-599-6262 Canada --------------------------- End of New-News digest ********************** From wcuan@net.com Wed Jul 24 21:25:22 1996 From: wcuan@net.com (William Cuan) Date: Wed Jul 24 21:25:24 PDT 1996 Subject: Debugging a spawned task Hello, I've been trying to follow these tips but haven't had much luck. I'm trying to attach to and debug a task that was spawned by another task that I'm debugging also. My develpment environment is VxWorks5.2 vxSim on Solaris. I'm trying this... 1. Start vxsim 2. start vxgdb, load, set breakpts in task#1 and run it -- task #2 is spawned 3. In vxSim shell, i() shows tasks as spawned and suspended 4. start another vxgdb and attempt an attach QUESTION: How do I bring in the code, set breakpoints for spawned tasks? Thanks in advance. Bill ------------- Excerpts of tips obtained.... Sure, after you have loaded the file go to your shell and set a breakpoint on the entry point name for the task. Then when the task is spawned, it will break on the first instruction. You can then lookup the taskid and attach to that task in the debugger. ======================8<---work---------------------- e-mail: bwenholz@pacesetter.com (Bruce Wenholz) Pacesetter, Inc Sylmar, Ca 91342, USA -------------------------------------------------------- Use the "b" command on the console to put a breakpoint on the entry point of your task. When the breakpoint hits, attach to the task with vxgdb. After you've attached and before you do anything else, delete the console breakpoint. (I use bdall). Then you can step through the code or set breakpoints via vxgdb. HTH, Gordon Uchenick uchenick@tis4000.com -------------------------------------------------------- You can suspend the newly spawned task by adding one line of code at the beginning of you task function: taskSuspend(0); After you attached to the task, resume the task from shell: taskResume(taskId) where taskId is the task number you get when you list tasks with command "i". If your other tasks have a short timeout waiting for the respond from the spawned task. You need to make them longer or add taskDelay in them or let them pending on the message from the spawned task. Yibing -------------------------------------------------------- -- +--------------------------+---------------------------------+ | William Cuan | Network Equipment Technologies | | email: wcuan@net.com | 800 Saginaw Drive, Bldg 23 | | phone: (415)780-5629 | Redwood City, CA 94063 | | fax : (415)780-4323 | M/S 23.1.1 | +--------------------------+---------------------------------+ From daemon@csg.lbl.gov Thu Jul 25 04:00:36 1996 From: daemon@csg.lbl.gov Date: Thu Jul 25 04:00:38 PDT 1996 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Thu Jul 25 04:00:31 PDT 1996 Subject: How to write GNU C program to access VMEbus address under VxWork? Subject: Re: How to write GNU C program to access VMEbus address under VxWork? Subject: Re: VxWorks 5.2 uncompress ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: How to write GNU C program to access VMEbus address under VxWork? Date: Mon, 22 Jul 1996 14:29:04 GMT From: cjsjesus@ms1.hinet.net.tw (Johnson Chen) Organization: NCUCC News Server Message-ID: <4t0322$i59@news.ncu.edu.tw> I am a beginner of VxWork operating system, and now I am starting a new project in which we will use one 68040 CPU board to control another speical I/O board through the wirting command word into the registers of I/O board. I can not find any function in reference manual of VxWork5.1 about such "VMEbus address direct access" function. Could anyone help me? Thank! Johnson Chen san01@noc1.tyc.edu.tw --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to write GNU C program to access VMEbus address under VxWork? Date: 24 Jul 1996 21:18:17 -0400 From: dave@sanders.com (Dave Uhrenholdt) Organization: Sanders SSD Message-ID: References: <4t0322$i59@news.ncu.edu.tw> <4t1gfe$ece@news.pacifier.com> Sender: dave@cushing.nowhere.com In article <4t1gfe$ece@news.pacifier.com> steved@pacifier.com@199.2.117.163 (Steve Doiel) writes: In <4t0322$i59@news.ncu.edu.tw>, cjsjesus@ms1.hinet.net.tw (Johnson Chen) writes: >I am a beginner of VxWork operating system, and now I am starting a >new project in which we will use one 68040 CPU board to control >another speical I/O board through the writing command word into the >registers of I/O board. I can not find any function in reference >manual of VxWork5.1 about such "VMEbus address direct access" >function. Could anyone help me? >Thank! >Johnson Chen >san01@noc1.tyc.edu.tw > The advice is all fine, but you should make sure you have the MMU setup correctly in the sysPhysMemDesc data structure. Depending on your board manufacture you may have certain addresses reserved for A16, and A24 space. Dave --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks 5.2 uncompress Date: 22 Jul 1996 20:54:24 GMT From: jfb@jma.com (John Bilger) Organization: appsig.com Message-ID: <4t0pq0$f4g@gate.appsig.com> References: <31EFFF9A.41C67EA6@strata.com> Reply-To: jfb@jma.com In article 41C67EA6@strata.com, Thi D Nguyen writes: >I am in the process of upgrading an existing system using VxWorks 5.1 >on (MIPS R3000) to 5.2. I am having trouble uncompressing the image >from the resident boot. This used to work in our version using VxWorks >5.1. Looking at the logic analyzer execution trace, it looks like the >uncompression was done; but the uncompress library call never returned. >I'd apprieciate any help or pointer. > >Thanks in advance, > >Thi Nguyen That sounds similar to a problem I had with uncompressing a boot rom image, although it was for an MV167 board. The boot rom size had to be set in $VX_BSP_BASE/config/mv167/romInit.s. The bottom line was that the default ROM size didn't match the ROM size I was using, so there was address discontinuity between the first pair or ROMs and the second pair of ROMs, which naturally caused trouble for the uncompress routine. Either way, it's one thing to think about. John Bilger --------------------------- End of New-News digest ********************** From rbakshi@hns.com Thu Jul 25 11:47:51 1996 From: "Ramandeep Bakshi" Date: Thu Jul 25 11:47:52 PDT 1996 Subject: Printing Activation Records Hi all, I am trying to debug some memory problems in one of the applications that is running on a Intel960. This application has been compiled for VxWorks using the gnu960 compiler. What I am required to do is to print out all the activation records in memory at a particular time. Anyone have any idea where I should start ? Any help will be much appreciated. Thanks a bunch, Raman Bakshi rbakshi@hns.com From MikeMcQuade@prometheus.connectsoft.net Thu Jul 25 15:14:46 1996 From: Mike McQuade Date: Thu Jul 25 15:14:48 PDT 1996 Subject: Unix / VxWorks Position SENIOR PROGRAMMER An international manufacturing firm is seeking a Senior Programmer to perform software design under a UNIX based operating system using HP 700 series computers. Duties will include prototyping new software products, supporting existing software products, interfacing w/ technical writer in creation of operating guides and interfacing with customers in troubleshooting system problems.Potential for providing field support in software design and maintenance. Desired experience includes 3 years programming experience, programming in C under a HPUX environment, user interface design implementing X-Windows and Motif, knowledge of HPUX system administration concepts, experience writing embedded software for the Power PC and experience using Wind Rivers VxWorks for embedded system design. Non-smoking environment. Competitive salary and Benefits, DOE. Equal Opportunity Employer. Reply to: pelwell@Zetec.com Pat Elwell, Human Resources Zetec, Inc. 1370 NW Mall Street PO Box 140 Issaquah WA 98027-0140 PH: (206) 392-5316 FAX: (206) 392-2086 From daemon@csg.lbl.gov Fri Jul 26 04:00:32 1996 From: daemon@csg.lbl.gov Date: Fri Jul 26 04:00:34 PDT 1996 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Fri Jul 26 04:00:28 PDT 1996 Subject: squeeze() for DOS file system Subject: Tracing Activation Records under VxWorks Subject: Call for Papers LCT-RTS '97 Subject: HTTP server for VxWorks Subject: --- PowerPC guru's --- Subject: vxWorks CAN Driver ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: squeeze() for DOS file system Date: Thu, 25 Jul 1996 16:47:14 +0200 From: mangane@laas.laas.fr (Laurent A. Mangane) Organization: LIS Message-ID: Hello VxWorks users, who knows where to find a version of squeeze() which could work with DOS file systems ? Thanks in advance for any hint. L.Mangane - -- To get random signatures put text files into a folder called 3Random Signatures2 into your Preferences folder. --------------------------- Newsgroups: comp.os.vxworks Subject: Tracing Activation Records under VxWorks Date: 22 Jul 1996 20:41:37 GMT From: rbakshi@wam.umd.edu (Raman Bakshi) Organization: University of Maryland, College Park, MD Message-ID: <4t0p21$kpf@dailyplanet.wam.umd.edu> Hi all, Herez the situation, we are writing for a system running vxworks on the Intel960 platform. The code is compiled on unix workstations using the gnu960 compiler and then transferred to the targer machine. What needs to be done is to somehow get a listing of all the activation records that are current. How can i do this ?? I have absolutely no idea where to start from. Anyone know ? Any Help much Appreciated, Raman Bakshi --------------------------- Newsgroups: comp.os.lynx,comp.os.qnx,comp.os.vxworks,comp.os.os9,comp.os.chorus,comp.os.mach Subject: Call for Papers LCT-RTS '97 Date: 25 Jul 1996 14:32:21 GMT From: whalley@sed.cs.fsu.edu (David Whalley) Organization: Florida State University Computer Science Department Message-ID: <4t80hl$6so@news.fsu.edu> Call for Papers ACM SIGPLAN Workshop on Languages, Compilers, and Tools for Real-Time Systems http://www.cs.pitt.edu/~gupta/lct-rts97.html Las Vegas, Nevada, 15 June, 1997 (In Conjunction with ACM SIGPLAN PLDI and PPoPP) ACM SIGPLAN LCT-RTS '97 is an interface between two dynamic fields of computer science and engineering: programming language implementation and real-time systems. Researchers in these areas are addressing many similar problems, but with different backgrounds and approaches. LCT-RTS is intended to expose researchers from either area to relevant work and interesting problems in the other area and provide a forum where they can interact. Until recently real-time systems development was performed by experienced specialists using a variety of custom kernels, non- standard languages, and vendor-specific device interfaces. System integration involved a complicated process of obtaining timing measurements, hand-tuning code, and re-measuring. These ad-hoc techniques do not scale well for modern systems. Also, the majority of real-time developers is no longer composed of embedded control experts. New software approaches are needed to support these new systems and provide appropriate tools for real- time programmers. Original submissions are invited in all areas relevant to this theme. Appropriate topics include (but are not restricted to) the following aspects of real-time systems. * Programming languages for real-time applications * Real-time design, specification, analysis * Exception & interrupt handling for real-time * Timing analysis: static & dynamic * Program optimization for real-time performance * Real-time profiling, measurement, & debugging * Real-time scheduling analysis * Real-time memory management & garbage collection * Language support for imprecise computation * Real-time on RISCs: caches, pipelines, windows * Real-time system integration & testing * Support for partitioning, mapping, & compression Papers should report new research and should not exceed 5000 words (approximately 10 typeset on 16-point spacing, or 15 typewritten double-spaced pages). Short papers that describe existing implementations or work-in-progress, or outline new problems or important issues are also welcome. Short papers should not exceed 3000 words (6 pages). All accepted papers will be presented at the workshop and published in the proceedings, which will be distributed at the workshop. PROGRAM COMMITTEE Rich Gerber (University of Maryland) Rajiv Gupta (University of Pittsburgh) Jane Liu (University of Illinois) Tom Marlowe (Seton Hall, NJIT RTCL) Sang-Lyul Min (University of Seoul) Frank Mueller (Humboldt University) Kelvin Nilsen (NewMonics Inc.) David Whalley (Florida State University) Andrew Wolfe (Princeton University) SUBMISSION Please e-mail a postscript copy of your submission to whalley@cs.fsu.edu by 5pm EST on January 10, 1997. If you are unable to send a postscript file via e-mail, then please send nine hard copies to one of the co-chairs. Authors will be notified by March 17. Final, camera-ready versions are due on May 15. Proceedings will be distributed at the workshop. Co-Chairs, ACM SIGPLAN LCT-RTS '97 Rajiv Gupta David Whalley Computer Science Department Computer Science Department University of Pittsburgh Florida State University Pittsburgh, PA 15260 Tallahassee, FL 32306-4019 e-mail: gupta@cs.pitt.edu e-mail: whalley@cs.fsu.edu phone: (412) 624-8421 phone: (904) 644-3506 fax: (412) 624-5249 fax: (904) 644-0058 --------------------------- Newsgroups: comp.os.vxworks Subject: HTTP server for VxWorks Date: 24 Jul 1996 17:02:08 -0700 From: jchang@physio-control.com (Jessie Chang) Organization: Physio-Control Message-ID: <4t6di0$10b@mistral.physio-control.com> Well, sorry for this repeated question. But I didn't catch the previous thread. Does anyone know a simple http server that can run under VxWorks? Thanks, Jessie - -- Jessie Chang, Physio-Control Co, Redmond, WA jchang@physio-control.com --------------------------- Newsgroups: comp.os.vxworks Subject: --- PowerPC guru's --- Date: 25 Jul 1996 21:08:09 GMT From: friedl@cher.heurikon.com (Ted Friedl) Organization: Heurikon Corporation Message-ID: <4t8nnp$a4t@badger.heurikon.com> Help! I am a PPC newbie and am looking at Wind River's romInit.s file for the mv1603. There are many instances of the following code example (initial condition: r2 = 0x80000cfc): fff001a0 7c80142c lwbrx r4, r0, r2 fff001a4 3c00403c addis r0, r0, 0x403c (16444) fff001a8 6000060c ori r0, r0, 0x60c fff001ac 7c80152c stwbrx r4, r0, r2 After reading the PPC processor manual, my interpretation of the code is this: fff001a0 r4 gets byte swapped MEM_WORD(0x80000cfc) fff001a4 r0 gets 0x403c0000 fff001a8 r0 gets 0x403c060c fff001ac MEM_WORD(0x80000cfc) gets byte swapped r4 In other words, MEM_WORD(0x80000cfc) is never changed. Am I correct in my understanding? Ted Friedl Heurikon Corporation --------------------------- Newsgroups: comp.os.vxworks,comp.realtime,comp.arch.embedded Subject: vxWorks CAN Driver Date: Thu, 25 Jul 1996 13:47:12 -0500 From: Steve Woloschek Organization: GE Medical Systems Message-ID: <31F7C130.FF6@ct.med.ge.com> Has anyone written, or know of a vxWorks Driver for the Intel 82527 CAN device? Thanks, Steve Woloschek woloschekS@med.ge.com GE Medical Systems --------------------------- End of New-News digest ********************** From darren@wrs.com Fri Jul 26 08:40:01 1996 From: Darren Cathey Date: Fri Jul 26 08:40:03 PDT 1996 Subject: Zilog 8440 Driver? Anyone out there in Exploder land know of a Zilog 8440 DUART driver for VxWorks? If so, I'd appreciate hearing about it... Thanks, Darren =========================================================== Darren Cathey Phone : 703-771-7934 Field Applications Engineer Fax : 703-771-9057 Wind River Systems Email : darren@wrs.com =========================================================== From sbadstuebner@ccgate.hac.com Fri Jul 26 08:54:13 1996 From: sbadstuebner@ccgate.hac.com Date: Fri Jul 26 08:54:14 PDT 1996 Subject: NTP availability? Does anyone know of an NTP port to VxWorks? thanks, Stefan Badstuebner sBadstuebner@ccgate.hac.com From daemon@csg.lbl.gov Sat Jul 27 04:00:59 1996 From: daemon@csg.lbl.gov Date: Sat Jul 27 04:01:00 PDT 1996 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sat Jul 27 04:00:55 PDT 1996 Subject: ? How to restrict VME Slave access Subject: SYS68K/PIO-1 (interrupts) Subject: Has anyone recieved this email? Subject: VxWorks Subject: Q: Signal useage (Can you use above _NSIGS) Subject: Re: ping and VxWorks 5.3 Subject: Timing Benchmarks Subject: Re: NTP availability? Subject: Re: Firewire support ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: ? How to restrict VME Slave access Date: Fri, 26 Jul 1996 07:47:05 GMT From: Hiske Huiting Organization: ICT Automatisering Deventer BV Message-ID: <31F877F9.327D@ict.nl> Sender: usenet@tweety.ict.nl (news) I have a Force 3CE sparc board with VxWorks 5.2 running. I use the WRS BSP and have installed VME Slave access. My whole memory is now accessible via VME, so anyone can overwrite my software and data. I would like to have a restricted area at a certain address allocated for VME Slave memory. Although I can change the size of the SLAVE memory area, I cannot find how to put this area to a certain address. I have tried this: - -create some memory space at 0xA000'0000 in structure sysPhysMemDesc - -map the SLAVE address in syslib.c via the ioMmuMicroSparcMap to this new address instead of the LOCAL_MEM_LOCAL_ADRS (0x0000'0000) It did not work: booting the system failed. I'm sure this is done before. Who helps? Thanks, Hiske ______________________________________________________________________ name: Hiske Huiting ) ICT Automatisering Deventer BV / Aerospace email: hiske@ict.nl ( P.O Box 701 tel: 31-570-633888 ) 7400 AS Deventer fax: 31-570-621362 ( The Netherlands ________________________)_____________________________:)______________ --------------------------- Newsgroups: comp.os.vxworks Subject: SYS68K/PIO-1 (interrupts) Date: Fri, 26 Jul 1996 11:26:20 +0200 From: Kurt Baumann Organization: Ingenieurschule ITR, Rapperswil, Switzerland Message-ID: <31F88F3C.41C67EA6@itr.ch> Have you ever programmed interrupt with VxWorks on the PIO-1 card. Our problem is: we have generated interrupts on the PIO-1 card wirh the handschake-pin H3. But the CPU-40 card (connected via VME bus) doesn't recognize the interrupt (IRQ6). The IRQ6 on the CPU-40 is connected to ISR with the VxWorks function intConnect(). After enabling the interrupts with sysIntEnable() the system breaks down. Have you ever programmed the PIO-1 card from FORCE Computers, then tell us the solution. Thanks and regards Kurt Baumann, ITR, Switzerland, kbaumann@itr.ch --------------------------- Newsgroups: comp.os.vxworks Subject: Has anyone recieved this email? Date: 25 Jul 1996 13:34:24 GMT From: johncoby@ghgcorp.com (JOHN R. COBARRUVIAS) Organization: Personal Account Message-ID: Hum................. X-POP3-Rcpt: johncoby@ghgcorp X-Sender: ami@vienna.ssds.com Mime-Version: 1.0 Date: Thu, 25 Jul 1996 09:10:24 -0600 To: johncoby@ghgcorp.com From: Ria Inkley Subject: VxWorks John, I came across your name as being someone who has experience with VxWorks. I'm with Scientific Research Corporation (SRC), an advanced technology engineering firm whose business focus spans a broad range of communication, radar, advanced processing, and environmental systems applications. I would like to have the chance to speak with you regarding the some of the work SRC is involved in, particularly in regards to real-time development using VxWorks. Please contact me at 770/989-9412, or e-mail me a phone number and a convenient time that I can reach you. My e-mail address is ami@ssds.com. You may also page me at 800/509-2146 during business hours, which are normally 8am - 6pm EST, M-F. Also, please visit our home page at http://www.scires.inter.net I look forward to speaking with you soon. Ria Inkley Scientific Research Corporation - -- ,-_/ . .-,--. ,---. . ' | ,-. |-. ,-. `|__/ | -' ,-. |-. ,-. ,-. ,-. . . . , . ,-. ,-. | | | | | | | | \ ,. | -. | | | | ,-| | | | | | / | ,-| `-. | `-' ' ' ' ' `-' ` `' `---' `-' `-' `-^ ' ' `-' `' ' `-^ `-' / | John R. Cobarruvias johncoby@ghgcorp.com Houston, Tx `--' --------------------------- Newsgroups: comp.os.vxworks Subject: Q: Signal useage (Can you use above _NSIGS) Date: Fri, 26 Jul 1996 10:38:42 -0700 From: Mike Harrelson Organization: Reticular Systems Inc. Message-ID: <31F902A2.41C67EA6@reticular.com> Hi, I'm considering using a signal as a way to control various context switches and task execution order. My dilema is whether or not I should use this instead of semaphores and what signals (besides SIGUSR1 and SIGUSR2) are available for use in vxWorks? I don't want to explain the entire scope of what I'm trying to do, but using semaphores complicates things. What I would like to do is raise signals to run tasks, and each different signal could start a different task. (Tasks would be run by the signal handler giving up their sempahores.) Thanks in advance... - -- - --------------------------------------------------------------------- Mike Harrelson harrelson@reticular.com A mind is like a parachute, it only works when it's open. - --------------------------------------------------------------------- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: ping and VxWorks 5.3 Date: 26 Jul 1996 16:01:31 GMT From: tfm@sei.cmu.edu (Theodore Marz) Organization: The Software Engineering Institute Message-ID: <4taq4r$387@news.sei.cmu.edu> References: <31F8DC91.72C5@fcfsrv1.lerc.nasa.gov> Hello, I just tried this in 5.2 on a MVME-167 card and it worked ok. I think the symbol INCLUDE_PING needs to be defined in .../config/all/configAll.h and your kernel remade Ping is not included in the default configuration of configAll.h as released. Ted Marz tfm@sei.cmu.edu About the keyword search not working... I know and I have sent mail to tech support letting them know. I also talked to an actual body there... He seemed to have no problem getting to the keyword search from the page. Seems like something is screwed up in someones nameserver. The URL for the keyword search string is a little screwy.. Looks like they used an internal name for the keyword search server, rather than a fully qualified name. In article <31F8DC91.72C5@fcfsrv1.lerc.nasa.gov>, Michael Charnoky writes: |> Hello, |> |> Our software lab recently installed VxWorks on a few remote 68040 targets. |> I've been writing code to test the ethernet connection between the boards. In |> order to troubleshoot some promlems, i tried to call the "ping" routine |> directly from the shell, but got an error: |> |> -> ping "ourtarget",1,0 |> undefined symbol: ping |> |> This seemed strange... so i wrote a short program, including "pingLib.h" to |> exercise "ping." Unfortunately, I got the same error. |> |> Has anyone else had this problem?? I'm stumped... the tech support web page |> lends no help - plus the keyword search doesn't even work. Any help is much |> appreciated!! |> |> mike |> ----------------------------------------------------------------------------- |> Michael Charnoky | phone: 216-433-3162 |> NASA Lewis Research Center | |> Cleveland, Ohio 44135 | email: mcharnok@fcfws3.lerc.nasa.gov |> ----------------------------------------------------------------------------- --------------------------- Newsgroups: comp.os.vxworks Subject: Timing Benchmarks Date: 26 Jul 1996 13:50:50 GMT From: milesm@tif531.ed.ray.com (Michael Miles) Organization: Raytheon Message-ID: <4taifq$6ks@news.res.ray.com> Hi-- A few months ago, I posted a request here for information on timing things in VxWorks. Someone from WRS was kind enough to fax me a copy of a WRS document on Methodology for doing timing benchmarks. Unfortunately, this fax mysteriously disappeared from my desk... :) I was hoping someone could point me to the document if it is online somewhere and also if whoever the WRS rep was could perhaps fax me a new copy. Thanks very much. Mike Miles --------------------------- Newsgroups: comp.os.vxworks Subject: Re: NTP availability? Date: 26 Jul 1996 17:00:19 GMT From: tfm@sei.cmu.edu (Theodore Marz) Organization: The Software Engineering Institute Message-ID: <4tatj4$e0r@news.sei.cmu.edu> References: <9606268383.AA838396366@CCGATE.HAC.COM> There is an NTP version 3 port code on the VxWorks archive. Refer to WRS's web page for how to get there. In article <9606268383.AA838396366@CCGATE.HAC.COM>, sbadstuebner@ccgate.hac.com writes: |> Does anyone know of an NTP port to VxWorks? |> |> thanks, |> |> Stefan Badstuebner |> sBadstuebner@ccgate.hac.com |> --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Firewire support Date: 23 Jul 1996 15:16:06 GMT From: Necromancer Organization: Huh? Message-ID: <4t2qbm$lqo@samba.rahul.net> References: <2.2.32.19960723061245.006fc8c0@mail.wrs.com> Reply-To: venu@rahul.net In <2.2.32.19960723061245.006fc8c0@mail.wrs.com> Harvey Wong writes: >Has anyone used VxWorks in an application using "Firewire" (the IEEE-394 >fast serial digital audio-visual bus)? >I'm interested in finding out what additional software is required to >support this. I haven't used it in a Firewire application, but I would imagine that you would need a P1394 driver. The driver will have to program the chipset you are using for the physical and link layers (TI, Apple). It will also need to have software layers for transaction support and bus management support. HTH. - -- Necromancer venu@rahul.net God's a Polytheist. venu@zeitnet.com --------------------------- End of New-News digest ********************** From daemon@csg.lbl.gov Sun Jul 28 04:00:34 1996 From: daemon@csg.lbl.gov Date: Sun Jul 28 04:00:36 PDT 1996 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sun Jul 28 04:00:31 PDT 1996 Subject: Re: Multicast addresses Subject: Why can't VxWorks handle multi-protocol sockets? ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Multicast addresses Date: Sat, 27 Jul 1996 21:09:16 -0500 From: blaine@texas.net (Blaine Kubesh) Organization: Texas Networking, Inc. Message-ID: References: <9607120917.AA04224@vega.syswiz.it> <4skaj2$jbp@ftp.iis.com.br> >In article <9607120917.AA04224@vega.syswiz.it>, Tiziana Balegno says: >> >>I am using VxWorks with the following system configuration: >> >>HOST Win95 >>ARCH MC68020 >>BSP FRC30 >> >>Using the Am7990 Lance Ethernet driver, I need to receive packets from the >>following two different multicast addresses >> >>0x09, 0x00, 0x2b, 0x00, 0x00, 0x04, 0xfe >>0x09, 0x00, 0x2b, 0x00, 0x00, 0x05, 0xfe >> >>according to the International Standard ISO 9542 (End System to Intermediate >>System routeing protocol ...). >>How can I set the two multicast address? > > Configuring the ethernet adapter to handle multicasting is >not possible through the functions available in the board's BSP. What >I think you should do is refer to the data-sheet of the Am7990, where >you'll find the information necessary to configure the registers of this >chip and try to program it directly. I have a different BSP but the functions should be similiar. You should be able to do an ioctl function call to add a multicast address to the NIC's hash table. Some BSP's may have this option and others may not. Check your man pages to be sure. The BSP I am using (68EN360) did not have this option, so I had to add it. Adding it requires that you have the source code for the ethernet driver. I used the Motorola data book to figure out how to add the multicast once I had the driver's ioctl source code. If you don't want to buy the source code for the driver, you could still write a function to do it. It would be more kludgy but it will work. WindRiver will sell you the source code if the driver is part of your BSP. Hope this helps, - -Blaine --------------------------- Newsgroups: comp.os.vxworks Subject: Why can't VxWorks handle multi-protocol sockets? Date: Sat, 27 Jul 1996 21:27:55 -0500 From: blaine@texas.net (Blaine Kubesh) Organization: Texas Networking, Inc. Message-ID: I really wish VxWorks could handle other protocols besides the PF_INET family. There are alot of other protocols out there, DecNet, SNA, LAT, IPX, etc. the list goes on. It seems really short sighted of Wind River to limit the capabilities of the sockLib to just TCP/IP protocols. I am really bummed because in order to add one of these other protocols to vxWorks I have to implement everything by hand. This means it will not be able support standard I/O (read, write, accept, select, etc.), which really stinks! Was it because of memory requirements? Too much overhead? I would really like to know. One answer I heard regarding this question is that I need to buy the streams interface. I could be wrong but I believe streams adds alot of overhead that could easily be done in sockets for this particular porpose. A discouraged developer, ~~~~~~~~~~~~~~~~ Blaine Kubesh BK Consulting blaine@texas.net "The bearing of a child takes nine months, no matter how many women are assigned." - The Mythical Man Month --------------------------- End of New-News digest ********************** From gdb@imsystems.com Sun Jul 28 20:58:55 1996 From: Greg Buzzard Date: Sun Jul 28 20:58:57 PDT 1996 Subject: Networking performance issue Can anyone shed any light on the following observation? Is this normal? Am I missing some configuration option? >From the time I do a "sendto" call at application level until the startOutput routine in my driver is called is about 9.8 msec (i.e., this time is apparently spent waiting somewhere in the OS). This drops to about 278 usec from the end of the first packet's DMA to the (intelligent) network interface card until the beginning of the 2nd packet getting into the driver for a back-to-back UDP send. Notes: 1. There is no signficant driver code on the path after the DMA is launched. 2. The times given are for a MVME 1603 (PowerPC board). 3. I am detecting no other PCI bus activity (i.e., no interrupts, etc.). 4. The driver is one that I am writing. -greg From P.H.A.Mutsaers@cycl.phys.tue.nl Mon Jul 29 08:44:45 1996 From: Peter Mutsaers Date: Mon Jul 29 08:44:47 PDT 1996 Subject: adjusting maximum size of socket-level send and receive buffers I have tried to adjust the maximum sizes of the socket-level send and receive buffers for both my client and server program using the setsockopt() routine (for example from the default value of 4096 bytes to 5000 bytes). And indeed, when I check on them with the getsockopt() routine, the values returned are the number of bytes I just assigned. However some strange things happen in this new situation: - When the client sends more then 5000 bytes, the number of bytes received by the server is only 2920. A reason for this could be that the number of bytes sends is larger then the maximum buffer size. However using the default socket level buffer sizes and sending more 4096 bytes the number of bytes received by the server is 4096, i.e. equal to the maximum buffer size. - When the client sends between 4097 and 5000 bytes, the number of bytes received by the server is again 2920, although in this situation the number of bytes sends is less then the maximum buffer size. - Finally, when the client sends less then 4097 bytes, the number of bytes received by the server is the number of bytes send. Is there anyone who can tell me why this happens, or what I do wrong? From rbakshi@hns.com Mon Jul 29 09:25:10 1996 From: "Ramandeep Bakshi" Date: Mon Jul 29 09:25:11 PDT 1996 Subject: Activation Records on VxWorks Hi all, I am desperately trying to get some info on how to trace Activation Records (the stack area of all the processes that are dynamically allocated for each function call) under VxWorks. Anyone have any idea where I can get this info ? Any help will be much appreciated. Thanks, Raman Bakshi From jhillman@wrs.com Mon Jul 29 10:44:15 1996 From: Jon Hillman Date: Mon Jul 29 10:44:17 PDT 1996 Subject: Re: Activation Records on VxWorks At 09:25 AM 7/29/96 PDT, Raman Bakshi wrote: > >Hi all, > >I am desperately trying to get some info on how to trace Activation Records >(the stack area of all the processes that are dynamically allocated for each >function call) under VxWorks. Anyone have any idea where I can get this info ? > Each vxWorks task has a stack which these "Activation Records" are stored (I usually refer to them as call stack to the function). To access them say from the debugger, set a break point before the first instruction in the function and then "inspect" the call stack (back trace) to let you know what parameters the function was called with. You can find the current stack pointer from the register inspect window, also. I hope this is what you were asking for. _________________________ | Jon Hillman, FAE | | Wind River Systems | /) (407) 273-4646 (\ / ) Email:jhillman@wrs.com ( \ _( (|_________________________) ) /> (((\ \) /,) / ) / //))/ (\\\\ \_/ / \ \_/ ///// \ / \ / \ _/ \_ / ----/ /----------------------\ \---- / / \ \ From wong@vms.masd.northgrum.com Mon Jul 29 15:51:47 1996 From: "J. Pierre Wong" Date: Mon Jul 29 15:51:49 PDT 1996 Subject: Standalone system setup Hello, I am trying to set up a standalone system using some old hardware we have store in the back of our cabinets. The main processor is a Motorola MVME136A that will drive some old Burr-Brown MPV 904's. The processing speed of this hardware is fine for our intended purposes (about 80Hz). The problem is that the software contains a data table of approximately 2 MegaBytes, and won't fit into the on-board EEPROMS. I found some old Force RR-2 memory boards that we could use to put the software on. My question is, how and where do I put the vxWorks software? Should a normal standalone vxWorks process be put on the 136A and have it spawn a task on the RR-2 board? If so, what steps do I need to take to compile and set up the code for the EEPROMS that reside on the RR-2 board. Does anybody have any other suggestions (other then getting newer and better hardware). Thanks in advance, Pierre Wong Vehicle Management Technology Systems Northrop Grumman Corporation Tel: 310.332.6840 Fax: 310.644.3652 Email: wong@vms.masd.northgrum.com From daemon@csg.lbl.gov Tue Jul 30 04:00:32 1996 From: daemon@csg.lbl.gov Date: Tue Jul 30 04:00:35 PDT 1996 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Tue Jul 30 04:00:29 PDT 1996 Subject: How to configure VxWorks interrupts ? Subject: Tornado/VxWorks 5.3 for PowerPC Subject: VMEbus accesses from MV1603 (VxWorks 5.3) Subject: VME Mailboxes (Location Monitors) and Shared Memory Driver Subject: Re: Why can't VxWorks handle multi-protocol sockets? ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: How to configure VxWorks interrupts ? Date: Mon, 29 Jul 1996 14:30:41 +0200 From: Hiske Huiting Organization: ICT Automatisering Deventer BV Message-ID: <31FCAEF1.AD@ict.nl> I have a Force 3CE SPARC board and want an external interrupt have triggered an application. I have mapped and configured the parallel port for this, but now I come to point where I want to tell the configuration that this interrupt should be connected to a certain interrupt vector. The function intConnect connects an ISR to a vector, but this vector much point to something (my parallel port paper error interrupt). Where is this configured and how to do? Thanks, Hiske - -- ______________________________________________________________________ name: Hiske Huiting ) ICT Automatisering Deventer BV / Aerospace email: hiske@ict.nl ( P.O Box 701 tel: 31-570-633888 ) 7400 AS Deventer fax: 31-570-621362 ( The Netherlands ________________________)_____________________________:)______________ --------------------------- Newsgroups: comp.os.vxworks Subject: Tornado/VxWorks 5.3 for PowerPC Date: Mon, 29 Jul 1996 15:38:25 GMT From: Graham Organization: University of Virginia Message-ID: <31FCDAF1.6598@cebaf.gov> Sender: usenet@murdoch.acc.Virginia.EDU Hello, does anyone have variable argument lists working on the PowerPC 604 using Tornado/VxWorks 5.3. We have some code which worked on the 68k and doesn't on the PowerPC. We wrote a little test to simulate the problem and have the same result. Please reply to timmer@cebaf.gov since he is working on the problem. I'm just his boss 8^) Graham - -<>-<>-<>-<>-<>-<>-<>-<>-<>-<>-<>-<>-<>-<>-<>-<>-<>-<>-<>-<>-<>-<>-<>-<>-<>-<>- Graham Heyes,CEBAF,12000 Jefferson Ave,Newport News,VA 23606,Tel:(804) 249-7030 http://codahp.cebaf.gov:3000/coda/coda_home.html & follow link under staff info --------------------------- Newsgroups: comp.os.vxworks Subject: VMEbus accesses from MV1603 (VxWorks 5.3) Date: Mon, 29 Jul 1996 12:32:33 GMT From: "Ross Microsystems, Inc." Organization: Ross Microsystems, Inc. Message-ID: <31FCAF61.CDF@qed.net> Hi all, I am having some difficulty accessing a VME board using the basic configuration of the mv1603 BSP supplied by WRS for Tornado 1.0 beta. If it makes any difference my host is a SunOS 4.1.3 box. All I would like to know is what address should I use for my VME board and then what address should I use in a "m" command from the VxWorks prompt. From this bit of knowledge, I can figure out what is going on. Any and all help is greatly appreciated. Jim McLean rmi@qed.net --------------------------- Newsgroups: comp.os.vxworks Subject: VME Mailboxes (Location Monitors) and Shared Memory Driver Date: Mon, 29 Jul 1996 14:12:46 -0700 From: Michael Levasseur Organization: GDE Systems Inc. Message-ID: <31FD28E0.41C67EA6@gdesystems.com> References: <01I7NJHVOID0B7B11F@cycl.phys.tue.nl> Hello, I'm working on a system that consists of multiple Baja R4700 VME64 boards with VxWorks 5.2B and I'm using the "shared memory network" driver. That is I'm using TCP/IP to communicate with the boards on the VME64 rack and with an Alpha. I'm curious if anyone has used VME Mailboxes on a system that's using the Shared Memory? Will the VME Mailboxes interfer with the Shared Memory socket communications. The VxWorks Programmer's Guide says that the Shared Memory network uses the VME Mailboxes. I'm curious what file contains the Mailbox Interrupt handler and what Mailbox value(s) the Shared Memory network is using? Is /tyCo/? the Shared Memory socket device driver or does the "sm= anchor Address" in the boot script do something else? I already have a work around to using Mailboxes and Shared Memory Driver, but it would be very nice to use both capabilities together. Thanx in advance...... --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Why can't VxWorks handle multi-protocol sockets? Date: Tue, 30 Jul 1996 09:07:19 GMT From: chaos@idiom.com (H.J. Bae) Organization: A poorly-installed InterNetNews site Message-ID: <4tkjcb$bc2@idiom.com> References: blaine@texas.net (Blaine Kubesh) wrote: >There are alot of other protocols out there, DecNet, SNA, LAT, IPX, etc. >the list goes on. It seems really short sighted of Wind River to limit the >capabilities of the sockLib to just TCP/IP protocols. >I am really bummed because in order to add one of these other protocols to >vxWorks I have to implement everything by hand. This means it will not be >able support standard I/O (read, write, accept, select, etc.), which >really stinks! VxWorks TCP/IP code is based on 4.3 BSD tahoe. Upgrading it to 4.4 BSD would give it OSI protocol suites. I have not seen DecNet or SNA in BSD based kernel implementation but I have seen IPX (similar to XNS) done in BSD based code. If you port 4.4 BSD into VxWorks and replace the native network code, you can retain standard socket I/O call compatibility. >Was it because of memory requirements? Too much overhead? I would really >like to know. As far as I know, it's because it's a lot of work to get it done. Not because of overhead. Not memory requirement (protocols can be optional). >One answer I heard regarding this question is that I need to buy the >streams interface. I could be wrong but I believe streams adds alot of >overhead that could easily be done in sockets for this particular >porpose. There are companies that have portable STREAMS based network stacks that have wider range of protocol support. WRS seems to have left the non TCP/IP support to those vendors for now. The STREAMS stuff I have seen, with exception of Mentat, are all pretty slow and bloated. --------------------------- End of New-News digest ********************** From tkb@mclean.sparta.com Tue Jul 30 13:42:44 1996 From: "Thomas K. Buchanan" Date: Tue Jul 30 13:42:46 PDT 1996 Subject: RS-6000 / 486 cross compiler Has anyone setup the GNU tools on a RS-6000 cross-compiling x86 code? ----------------------------------------- /\\ Keith Buchanan ///\\ /////\\ SPARTA, Inc. ///// \\ 7926 Jones Branch Drive / \\\\\ \\ Suite 900 \ \\\\\ // McLean, VA 22102 \ ////// \////// Voice: 703 448 0210 \//// FAX: 703 893 5494 \// WWW: www.mclean.sparta.com From cpires@br.homeshopping.com.br Tue Jul 30 15:02:27 1996 From: Carlos Pires Date: Tue Jul 30 15:02:29 PDT 1996 Subject: Word alignment Hi everyone ! Does anyone know how to align a VxWorks application to 4 (four) bytes. This is due to match a Solaris 2.4 driven CPU to a VxWorks one. Thanks to you all !! Bye for now. Carlos From hesham@baynetworks.com Tue Jul 30 18:49:15 1996 From: Hesham El Bakoury Date: Tue Jul 30 18:49:17 PDT 1996 Subject: Using BAT regsiters for PowerPc in Tornado I am trying to use the PowerPC MMU BAT registers for memory mapping in Tornado. The way PowerPC works is that it maps memory using BAT registers first if it finds an address match there. If not it uses segment registers and page tables. Since memory mapping using BAT registers is faster and does not consume memory to setup page tables, I prefer using them. Unfortunately VxWorks creates page tables whether or not BAT are used; and I do not know of any way to prevent these tables from being created if I decide to use BAT registers. Did any one configure MMU support in Tornado in such a way that BAT registers are used for memory mapping and page tables are never created. If yes, how did you define the sysBatDesc structure in sysLib.c and what macros in configAll.h you #define/#undef in config.h to make memory mapping (and caching) work using BAT registers without creating page tables and using segment registers. Regards, Hesham El Bakoury Bay Networks, CA. Hesham El Bakoury Bay Networks Tel: 408-486-3410 From tja@radstone.co.uk Wed Jul 31 03:20:08 1996 From: Tim Allen Date: Wed Jul 31 03:20:11 PDT 1996 Subject: Re: Using BAT regsiters for PowerPc in Tornado > I am trying to use the PowerPC MMU BAT registers for memory mapping > in Tornado. The way PowerPC works is that it maps memory using BAT > registers first if it finds an address match there. If not it uses > segment registers and page tables. > > Since memory mapping using BAT registers is faster and does not consume > memory to setup page tables, I prefer using them. > > Unfortunately VxWorks creates page tables whether or not BAT are used; > and I do not know of any way to prevent these tables from being created > if I decide to use BAT registers. As I understood it, VxWorks only creates the page tables based on the sysPhysMemDesc[] definition in sysLib.c, so if you keep this definition to a minimum and use sysBatDesc[] instead, the size of the page tables will be similarly reduced. > Did any one configure MMU support in Tornado in such a way that BAT registers > are used for memory mapping and page tables are never created. > > If yes, how did you define the sysBatDesc structure in sysLib.c and > what macros in configAll.h you #define/#undef in config.h to make > memory mapping (and caching) work using BAT registers without creating > page tables and using segment registers. The Radstone PPC603/4 Tornado BSP uses the sysBatDesc[] to map upto 256Mb of local memory, 256Mb of PCI memory space and 256Mb of PCI I/O space. The sysPhysMemDesc[] was then used to map the other, smaller areas eg bootrom, some VME space, and other resources. Using the BATs to map PCI space considerably reduces the initialisation time and page table size. Regards, Tim. +------------------------------------------------------------------+ | Tim Allen phone: +44 (0)1327 359444 | | Radstone Technology PLC, UK. fax: +44 (0)1327 358112 | +------------------------------------------------------------------+ From daemon@csg.lbl.gov Wed Jul 31 04:01:10 1996 From: daemon@csg.lbl.gov Date: Wed Jul 31 04:01:12 PDT 1996 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Wed Jul 31 04:01:06 PDT 1996 Subject: Re: NTP availability? Subject: Minimal pcnfsd For VxWorks? Subject: Time synchronization Subject: Re: NTP availability? Subject: printf() broken in Tornado's version of GNU C Subject: Re: ping and VxWorks 5.3 Subject: Re: NTP availability? Subject: Re: VME Mailboxes (Location Monitors) and Shared Memory Driver Subject: Re: HTTP server for VxWorks Subject: Re: Why can't VxWorks handle multi-protocol sockets? Subject: Re: Why can't VxWorks handle multi-protocol sockets? Subject: Re: Networking performance issue Subject: Re: Debugging a spawned task ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: NTP availability? Date: 30 Jul 1996 12:09:38 GMT From: don@mars.dgrc.doc.ca (Donald McLachlan) Organization: The Communications Research Centre Message-ID: <4tku22$rof@crc-news.doc.ca> References: <4tjug9$br7@mr-ed.jpl.nasa.gov> Reply-To: don@mars.dgrc.doc.ca >The clock on the HKV4F is not of very high precision if I remember right. >Something like 10ms comes to mind. You may want to just work with the >kernel tick by using clock_gettime and clock_settime. Anyway the routines >you want are in config/hkv4f/dp8750aRTC.c, more or less. I started >down this track a few years back and gave up. (I couldn't get xntpd >to converge). Below is what I've dug out from my efforts. In the >meantime I've been trying to work bottom-up from the NTP RFC's. > >Good luck and let us all know if you get it working! > >Matt We wrote our own cheapie NTP a while back. The trick to getting the time to converge is to adjust the tick rate so it turns out to be a factor of the clock driving the timer chip. We bumped the tick rate from 60 ticks (which never converged) to 64 and the time converged and was quite stable. In article <4tjug9$br7@mr-ed.jpl.nasa.gov>, mwette@mr-ed.jpl.nasa.gov (Matt Wette) writes: >In article <4tid19$mcm@crc-news.doc.ca>, >Donald McLachlan wrote: >> >>I work with Norm Young (who started this thread), we sort of have gettimeofday, >>settimeofday, adjtime etc. Currently we have this stuff working, but without >>being able to access the real time clock (even with the MMU support disabled) so >>time at boot up is Jan 1 1977, but xntpd should fix this :-). >>Wanna trade? >> >>Anyone out there able to access the real time clock on the hkv4f using the macros >>supplied by vxWorks 5.2? - -- Donald McLachlan e-mail donald.mclachlan@crc.doc.ca Communications Research Centre / DRX office 613-998-2845 3701 Carling Ave., fax 613-998-9648 Ottawa, Ontario labs 613-998-2423 / 613-998-4118 K2H 8S2 home 613-599-6262 Canada --------------------------- Newsgroups: comp.os.vxworks Subject: Minimal pcnfsd For VxWorks? Date: Tue, 30 Jul 1996 12:45:53 GMT From: mfisher@dyn.com (Mark Fisher) Organization: The World @ Software Tool & Die Message-ID: Reply-To: mfisher@dyn.com Sender: news@world.std.com (Mr Usenet Himself) Hi, I want to be able to mount a disk on vxWorks on a PC. I can get this to work by specifying a server that is running pcnfsd to do the user authentication. However, I need to be able to have the vxWorks system do the user authentication in the field. Is there any port (even a minimal one) of pcnfsd that is available that does this? Or does anyone have any suggestions as to how to get around this? Also is there any Windows 95 NFS client software that can be configured not to do the user authentication? Thanks, Mark Fisher Dynamic ReSolutions --------------------------- Newsgroups: comp.os.vxworks Subject: Time synchronization Date: Tue, 30 Jul 1996 10:53:00 +0200 From: Casey Crellin Organization: CCII Systems Message-ID: <31FDCD6C.2CA0@ccii.co.za> References: <4tid19$mcm@crc-news.doc.ca> <4tjug9$br7@mr-ed.jpl.nasa.gov> Hi, Over the past couple of days qute a bit has been written regarding ntp on Vxworks. I am in the process of compiling the archive version of ntp on Tornado. What we would actually like is a time synchronization method for a number of machines on an isolated network, with a very high degree of accuracy. (aiming at 200us offsets between clocks) Ntp is for internet time sync so I understand that it would not normally need to be that accurate. Is there any other protocol that is available that is more suited to this requirement? Casey - -- Casey Crellin casey@ccii.co.za CCII Systems (Pty) Ltd - ------------------------------------------------------------ --------------------------- Newsgroups: comp.os.vxworks Subject: Re: NTP availability? Date: 30 Jul 1996 18:29:27 GMT From: don@mars.dgrc.doc.ca (Donald McLachlan) Organization: The Communications Research Centre Message-ID: <4tlka7$6e8@crc-news.doc.ca> References: <4tku22$rof@crc-news.doc.ca> Reply-To: don@mars.dgrc.doc.ca All rightie then. I don't normally post a followup to my own message but there are some things one just has-ta-do :-) I thought I had vxWorks 5.2 rebuilt with the usrTime package fully installed, but no such luck. Following the usrTime installation instructions I have undefined INCLUDE_ANSI_TIME, I have added LIB_EXTRA in h/make/makeTarget2, and I have put the LIB_EXTRA macro ahead of all the other libraries in the LIBS macro. Unfortunatetely when I try to make the kernel it fails with ... 230 mars vxworks> make make -f Makefile.MC68040gnu CPU=MC68040 TOOL=gnu vxWorks `version.o' is up to date. ld68k -X -N -e _sysInit -Ttext 00001000 \ -o vxWorks dataSegPad.o sysALib.o sysLib.o tyCoDrv.o HostAdapt.obj ncr53c700Lib.obj scripts.obj usrConfig.o version.o ../../packages/usrTime/usrTime.a /usr/local/packages/vxworks5.2/vw/lib/libMC68040gnudrv.a /usr/local/packages/vxworks5.2/vw/lib/libMC68040gnuvx.a localtime.o(.text+0): multiple definition of `_localtime (.text)' /usr/local/packages/vxworks5.2/vw/packages/usrTime/usrTime.c:417: (localtime): first seen here gmtime.o(.text+0): multiple definition of `_gmtime (.text)' /usr/local/packages/vxworks5.2/vw/packages/usrTime/usrTime.c:402: (gmtime): first seen here *** Error code 1 make: Fatal error: Command failed for target `vxWorks' Current working directory /usr/local/packages/vxworks5.2/vw/config/hkv4f *** Error code 1 make: Fatal error: Command failed for target `vxWorks' Any idea how I can either remove localtime and gmtime from their libraries, or ignore this error so vxworks will build ( and hopefully sellect the usrTime versions of these functions) ? Don - -- Donald McLachlan e-mail donald.mclachlan@crc.doc.ca Communications Research Centre / DRX office 613-998-2845 3701 Carling Ave., fax 613-998-9648 Ottawa, Ontario labs 613-998-2423 / 613-998-4118 K2H 8S2 home 613-599-6262 Canada --------------------------- Newsgroups: comp.os.vxworks Subject: printf() broken in Tornado's version of GNU C Date: Sun, 28 Jul 1996 21:41:44 GMT From: brucew@phoenix.net (Bruce Wedding) Organization: BranPaul Systems Message-ID: <4tgn51$n34@uhura.phoenix.net> I'm having trouble getting either of the following lines to work properly in Tornado vxWorks 5.3. printf("%25s", buf); or printf("%*s", 25, buf); I'm passing a buffer that is not nul terminated, so if the length specification doesn't work, I print off into hyperspace. That is exactly what is happening now. Anyone else ever tried to use printf() this way? Is there a fix? Bruce Wedding --------------------------- Newsgroups: comp.os.vxworks Subject: Re: ping and VxWorks 5.3 Date: Tue, 30 Jul 1996 12:19:47 -0700 From: greg bruno Organization: ncr classic Message-ID: <31FE6053.122C@sandiegoca.ncr.com> References: <31F8DC91.72C5@fcfsrv1.lerc.nasa.gov> Michael Charnoky wrote: > > Hello, > > Our software lab recently installed VxWorks on a few remote 68040 targets. > I've been writing code to test the ethernet connection between the boards. In > order to troubleshoot some promlems, i tried to call the "ping" routine > directly from the shell, but got an error: > > -> ping "ourtarget",1,0 > undefined symbol: ping > > This seemed strange... so i wrote a short program, including "pingLib.h" to > exercise "ping." Unfortunately, I got the same error. i'm using 5.2 and if i define "INCLUDE_PING" in my config.h, ping works great. - greg bruno --------------------------- Newsgroups: comp.os.vxworks Subject: Re: NTP availability? Date: 30 Jul 1996 20:18:57 GMT From: tfm@sei.cmu.edu (Theodore Marz) Organization: The Software Engineering Institute Message-ID: <4tlqnh$90t@news.sei.cmu.edu> References: <4tku22$rof@crc-news.doc.ca> <4tlka7$6e8@crc-news.doc.ca> In article <4tlka7$6e8@crc-news.doc.ca>, don@mars.dgrc.doc.ca (Donald McLachlan) writes: |> All rightie then. I don't normally post a followup to my own message but there |> are some things one just has-ta-do :-) |> |> I thought I had vxWorks 5.2 rebuilt with the usrTime package fully installed, |> but no such luck. |> |> Following the usrTime installation instructions I have undefined INCLUDE_ANSI_TIME, |> I have added LIB_EXTRA in h/make/makeTarget2, and I have put the LIB_EXTRA macro |> ahead of all the other libraries in the LIBS macro. |> |> Unfortunatetely when I try to make the kernel it fails with ... |> |> 230 mars vxworks> make |> make -f Makefile.MC68040gnu CPU=MC68040 TOOL=gnu vxWorks |> `version.o' is up to date. |> ld68k -X -N -e _sysInit -Ttext 00001000 \ |> -o vxWorks dataSegPad.o sysALib.o sysLib.o tyCoDrv.o HostAdapt.obj |> ncr53c700Lib.obj scripts.obj usrConfig.o version.o |> ../../packages/usrTime/usrTime.a |> /usr/local/packages/vxworks5.2/vw/lib/libMC68040gnudrv.a |> /usr/local/packages/vxworks5.2/vw/lib/libMC68040gnuvx.a |> localtime.o(.text+0): multiple definition of `_localtime (.text)' |> /usr/local/packages/vxworks5.2/vw/packages/usrTime/usrTime.c:417: (localtime): first seen here |> gmtime.o(.text+0): multiple definition of `_gmtime (.text)' |> /usr/local/packages/vxworks5.2/vw/packages/usrTime/usrTime.c:402: (gmtime): first seen here |> *** Error code 1 |> make: Fatal error: Command failed for target `vxWorks' |> Current working directory /usr/local/packages/vxworks5.2/vw/config/hkv4f |> *** Error code 1 |> make: Fatal error: Command failed for target `vxWorks' |> |> Any idea how I can either remove localtime and gmtime from their libraries, or |> ignore this error so vxworks will build ( and hopefully sellect the usrTime |> versions of these functions) ? |> |> Don |> |> -- |> Donald McLachlan e-mail donald.mclachlan@crc.doc.ca |> Communications Research Centre / DRX office 613-998-2845 |> 3701 Carling Ave., fax 613-998-9648 |> Ottawa, Ontario labs 613-998-2423 / 613-998-4118 |> K2H 8S2 home 613-599-6262 |> Canada |> |> |> |> In the BSP for a MVME-167 card (68040 based) there is a _localtime and _gmtime in the library ..../libMC68040gnuvx.a Beats me if it is the same as the localtime and gmtime from usrTime/usrTime. But, this is (I believe) an ar archive, and so you could remove them with an ar d libMC68040gnuvx.a localtime.o ar d libMC68040gnuvx.a gmtime.o Good luck Ted Marz --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VME Mailboxes (Location Monitors) and Shared Memory Driver Date: 30 Jul 1996 19:57:49 GMT From: friedl@cher.heurikon.com (Ted Friedl) Organization: Heurikon Corporation Message-ID: <4tlpft$qv4@badger.heurikon.com> References: <01I7NJHVOID0B7B11F@cycl.phys.tue.nl> <31FD28E0.41C67EA6@gdesystems.com> <4tl6sp$eqa@xmission.xmission.com> In article <4tl6sp$eqa@xmission.xmission.com>, bgeer wrote: >This just in: a colleague assures me the BSP uses mailbox 0 for sm, & >supports mailboxes 1,2,3 for user use via intConnect() & a little >interrupt handling code. If there are really 8 mailboxes in the >hardware, I don't know what the deal is with 4,5,6,7. The VIC does contain 8 different mailboxes. Four of them (the ICMRs) enable any VME master to interrupt the VIC. The other four (the ICGRs) allow only a VIC to broadcast an interrupt to other VICs if and only if the VIC mailbox slave structures are overlapped. Overlapping the mailbox structure defeats the use of the ICMRs. The use of the ICMRs and ICGRs is therefore mutually exclusive. By far, the greatest consumer of mailbox interrupts is Wind River's shared memory backplane driver - which has no provision for broadcasted interrupts. We therefore chose to support ICMRs. Believe it, or not! :-> Ted Friedl Senior Software Engineer (and vicVme.c author!) Heurikon Corporation Madison, WI --------------------------- Newsgroups: comp.os.vxworks Subject: Re: HTTP server for VxWorks Date: 30 Jul 1996 15:28:56 -0700 From: nigel@idiom.com (Nigel Standing) Organization: A poorly-installed InterNetNews site Message-ID: References: <4t6di0$10b@mistral.physio-control.com> <4tja59$duv@idiom.com> chaos@idiom.com (H.J. Bae) writes: >Jessie Chang (jchang@physio-control.com) wrote: >: Well, sorry for this repeated question. But I didn't catch the >: previous thread. Does anyone know a simple http server that can run >: under VxWorks? > I have an extremely short VxWorks program (a hack hahaha) that will do text > only HTTP server function which works with clients like netscape > and mosaic. It'll be trivial to put in gif and cgi support into it. Actually porting the CERN httpd is pretty trivial; there are a couple of fork (or vforks) which need to be changed but if I recall correctly they can be made into function calls rather than taskSpawns. So just ftp the code from CERN and give it a try :-) Nigel - --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Why can't VxWorks handle multi-protocol sockets? Date: 30 Jul 1996 15:39:43 -0700 From: nigel@idiom.com (Nigel Standing) Organization: A poorly-installed InterNetNews site Message-ID: References: <4tkjcb$bc2@idiom.com> chaos@idiom.com (H.J. Bae) writes: > blaine@texas.net (Blaine Kubesh) wrote: >> There are alot of other protocols out there, DecNet, SNA, LAT, IPX, etc. >> the list goes on. > VxWorks TCP/IP code is based on 4.3 BSD tahoe. Upgrading it to 4.4 > BSD would give it OSI protocol suites. Rumor has it that WRS are in the process of upgrading VxWorks to BSD 4.4 networking (at long last). I'm not sure what this will actually mean (which protocols will be supported etc), perhaps you should call your local WRS Sales person for the low down. Nigel - --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Why can't VxWorks handle multi-protocol sockets? Date: Tue, 30 Jul 1996 18:37:48 -0500 From: blaine@texas.net (Blaine Kubesh) Organization: BK Consulting Message-ID: References: <4tkjcb$bc2@idiom.com> In article <4tkjcb$bc2@idiom.com>, chaos@idiom.com (H.J. Bae) wrote: >VxWorks TCP/IP code is based on 4.3 BSD tahoe. Upgrading it to 4.4 >BSD would give it OSI protocol suites. I have not seen DecNet or SNA >in BSD based kernel implementation but I have seen IPX (similar to >XNS) done in BSD based code. If you port 4.4 BSD into VxWorks and >replace the native network code, you can retain standard socket I/O >call compatibility. I don't know how to just replace the native network code without the source code to the VxWorks kernel. Do they distribute the kernel source? Thanks for your input! - -Blaine Kubesh --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Networking performance issue Date: Tue, 30 Jul 1996 09:10:41 GMT From: chaos@idiom.com (H.J. Bae) Organization: A poorly-installed InterNetNews site Message-ID: <4tkjil$bc2@idiom.com> References: <199607290405.VAA02457@puma.imsystems.com> Greg Buzzard wrote: >>From the time I do a "sendto" call at application level until the >startOutput routine in my driver is called is about 9.8 msec (i.e., >this time is apparently spent waiting somewhere in the OS). This >drops to about 278 usec from the end of the first packet's DMA to the >(intelligent) network interface card until the beginning of the 2nd >packet getting into the driver for a back-to-back UDP send. Sometimes people write network drivers to use queues for send side. This is commonly avoided in recent UNIX kernels. VxWorks may still have this way of handling output side, depending on your driver. It may also be due to a netTask. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Debugging a spawned task Date: Tue, 30 Jul 1996 20:24:03 -0700 From: SCSI DIAGS Organization: Winged Lizard Kingdoms Message-ID: <31FED1D3.3D86@FUZZY.BEAR> References: <199607250425.VAA01684@spyglass.net.com> Reply-To: johnx@wrs.com William Cuan wrote: > > Hello, > > I've been trying to follow these tips but haven't had much > luck. I'm trying to attach to and debug a task that > was spawned by another task that I'm debugging also. > My develpment environment is VxWorks5.2 vxSim on Solaris. > > I'm trying this... > > 1. Start vxsim > 2. start vxgdb, load, set breakpts in task#1 and run it > -- task #2 is spawned > 3. In vxSim shell, i() shows tasks as spawned and suspended > 4. start another vxgdb and attempt an attach > QUESTION: How do I bring in the code, set breakpoints > for spawned tasks? One method is to first use add-symbol-file command to load the objects symbols into the second vxgdb session. Then get the TID of the 2nd task (which is suspended) via the shell i or ti command and use "attach " in the 2nd gdb session. John > > Thanks in advance. > > Bill > > ------------- Excerpts of tips obtained.... > > Sure, after you have loaded the file go to your shell and set a breakpoint > on the entry point name for the task. Then when the task is spawned, it > will break on the first instruction. You can then lookup the taskid and > attach to that task in the debugger. > > ======================8<---work---------------------- > e-mail: bwenholz@pacesetter.com (Bruce Wenholz) > Pacesetter, Inc Sylmar, Ca 91342, USA > > -------------------------------------------------------- > Use the "b" command on the console to put a breakpoint on the entry point > of your task. When the breakpoint hits, attach to the task with vxgdb. After > you've attached and before you do anything else, delete the console breakpoint. > (I use bdall). Then you can step through the code or set breakpoints via vxgdb. > > HTH, > Gordon Uchenick > uchenick@tis4000.com > -------------------------------------------------------- > > You can suspend the newly spawned task by adding one line of code > at the beginning of you task function: > taskSuspend(0); > After you attached to the task, resume the task from shell: > taskResume(taskId) > where taskId is the task number you get when you list tasks with > command "i". > > If your other tasks have a short timeout waiting for the respond from > the spawned task. You need to make them longer or add taskDelay in > them or let them pending on the message from the spawned task. > > Yibing > > -------------------------------------------------------- > > -- > +--------------------------+---------------------------------+ > | William Cuan | Network Equipment Technologies | > | email: wcuan@net.com | 800 Saginaw Drive, Bldg 23 | > | phone: (415)780-5629 | Redwood City, CA 94063 | > | fax : (415)780-4323 | M/S 23.1.1 | > +--------------------------+---------------------------------+ --------------------------- End of New-News digest ********************** From William_Duvall@cscgt.gsfc.nasa.gov Wed Jul 31 05:49:27 1996 From: William_Duvall@cscgt.gsfc.nasa.gov (William Duvall) Date: Wed Jul 31 05:49:29 PDT 1996 Subject: unsubscribe UNSUBSCRIBE From js@eroch.mc.xerox.com Wed Jul 31 06:27:11 1996 From: js@eroch.mc.xerox.com (Jeffrey R. Szczepanski) Date: Wed Jul 31 06:27:13 PDT 1996 Subject: Re: Using BAT regsiters for PowerPc in VxWorks > Did any one configure MMU support in Tornado in such a way that BAT registers > are used for memory mapping and page tables are never created. > > If yes, how did you define the sysBatDesc structure in sysLib.c and > what macros in configAll.h you #define/#undef in config.h to make > memory mapping (and caching) work using BAT registers without creating > page tables and using segment registers. It is, in fact, possible to completely eliminate the initialization of the sysPhysMemDesc[] and therefore the existance of the page tables by not calling usrMmuInit() in the VxWorks startup. Of course, this leaves the init of the MMU to the developer. We have setup the MMU/BAT ourselves and eliminated the page tables as described above. To use the system caches, the MMU must indeed, be enabled in the proper manner with your own code. The Power PC documentation explains the proper way enable the MMU, etc....It turns out to be fairly easy to do. We setup the BATs and enable the MMU before braching into the "C" code at system startup. As a side note, Wind River has a cheesy global variable interface between their MMU library and the cache library. Therefore, to use their cache libraries (to enable caches, etc.) you must set the following globals after the MMU is enabled for both data and instructions: /* * For VxWorks Compatibility * * These variable need to be set to true when the MMU is * in use so that the cache can be enabled via cacheArchLib */ BOOL mmuPpcIEnabled = TRUE; BOOL mmuPpcDEnabled = TRUE; ------------------------------------------------------------------------------ Jeffrey R. Szczepanski Phone: (716) 264-6613 Xerox Engineering Systems 300 Main Street East Rochester, NY 14445 Email: js@eroch.mc.xerox.com ------------------------------------------------------------------------------ From wong@vms.masd.northgrum.com Wed Jul 31 08:47:37 1996 From: "J. Pierre Wong" Date: Wed Jul 31 08:47:38 PDT 1996 Subject: Serial Communication Controller Greeting, I am trying to set up a Zilog Z85C30 Serial Communications Controller in an asynchronous polling mode. Does anyone know the correct initialization sequence to put the SCC in this mode? Pierre Wong Vehicle Management Systems Technologies Northrop Grumman Corp Email: wong@vms.masd.northgrum.com Tel: 310.332.6840 Fax: 310.644.3652