From zmau@ksl.co.il Wed Nov 1 00:32:26 1995 From: Maurice Zamir Date: Wed Nov 1 00:32:29 PST 1995 Subject: Re: talking to vxWorks from Unix shell script? > carl@ccnet.com (Carl Farrington) wrote: > >We have a test setup with vxWorks running on 2 different boards. > >Starting the test requires typing a few commands to board1, then some to > >board2, then back to board 1 for a while, etc. > > > >I'd like to automate this with a shell script (or something). Is there > >any easy way to send a single line command from a Unix program over the > >ethernet to a vxWorks shell, and get the output back? Ideally, I'd like > >the equivalent of rshd running under vxWorks. > > > >Any suggestions? > > > > I'd suggest using TCL (with expect). That's what we've been using here, and > it's what WRS has developed the BSP test scripts in. > > TCL with the "expect" extension is undoubtedly the simplest and most flexible > way of remotely automating tests on a vxWorks machine (or any other machine for > that matter..) > Can someone give more information about TCL (How to get ?) Thanks Zamir Maurice Zamir Maurice | zmau@ksl.co.il Kulicke & Soffa Ltd. | Phone: (04) - 545379 A.T.C. Haifa, Israel | fax: (04) - 550827 From daemon@csg.lbl.gov Wed Nov 1 04:00:26 1995 From: daemon@csg.lbl.gov Date: Wed Nov 1 04:00:29 PST 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Wed Nov 1 04:00:20 PST 1995 Subject: PCMCIA ATA / IDE Drivers for VxWorks Subject: sysClkRateGet Subject: Re: help !!! Subject: Re: Passing FILE types as a function parameter Subject: Re: NFS and rpcTaskInit() Subject: Re: How to profile VxWorks tasks? ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: PCMCIA ATA / IDE Drivers for VxWorks Date: 31 Oct 1995 16:09:30 GMT From: farshid@ix.netcom.com (farshid sabet) Organization: Netcom Message-ID: <475hnq$khq@ixnews6.ix.netcom.com> References: <46p0e0$4s1@bug.rahul.net> I am looking for drivers for PCMCIA ATA and Standard IDE harddisks ? Does anybody know where can I find these drivers ? Thanks, Farshid farshid@ix.netcom.com --------------------------- Newsgroups: comp.os.vxworks Subject: sysClkRateGet Date: 31 Oct 1995 17:09:50 GMT From: stevecor@nwlink.com (Steve Cornett) Organization: Bullseye Software Message-ID: <475l8u$gkh@texas.nwlink.com> Reply-To: steve@bullseye.com How can I obtain the system clock rate (ticks/second) in a way which is portable to many or all VxWorks implementations and versions? The VxWorks reference manual describes a function named sysClkRateGet but implies that for different BSPs the function may not be implemented or might be named differently or might behave differently. What is the danger with using this function? - -- Steve Cornett Bullseye Software Voice 800-278-4268, Fax 206-524-3575 Web http://www.bullseye.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: help !!! Date: 31 Oct 1995 21:48:14 GMT From: Rendell Fong Organization: svl.trw.com Message-ID: <4765iu$mi4@gatekeeper.svl.trw.com> References: <46tak1$jmt@fnnews.fnal.gov> Don't know if this is the problem but the floating point context of the counter task may be getting destroyed. This condition may occur when a context switch occurs swapping in some other higher priority task which also performs floating point operations. When the counter task becomes active again the floating point coprocessor registers are not necessarily restored. It can be prevented by specifying the VX_FP_TASK option during taskSpawn() call. In addition, other tasks which also perform floating point operations should do this. taskSpawn ("t_Monitor", 80, VX_FP_TASK, 20000, display); Good Luck, Rendell Fong --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Passing FILE types as a function parameter Date: 31 Oct 1995 23:38:20 -0800 From: pwatson@eworld.com (PWatson) Organization: eWorld Message-ID: <47785c$2hn@hp5.online.apple.com> References: Reply-To: pwatson@eworld.com (PWatson) Sender: root@news-relay.eworld.com Pass func(fp); Since it's already a pointer, no need for '&'. They just don't want you to pass func (*fp); Philip Watson Matsushita Avionics Systems Corp., Irvine, CA --------------------------- Newsgroups: comp.os.vxworks Subject: Re: NFS and rpcTaskInit() Date: 31 Oct 1995 23:40:01 -0800 From: pwatson@eworld.com (PWatson) Organization: eWorld Message-ID: <47788h$2ho@hp5.online.apple.com> References: Reply-To: pwatson@eworld.com (PWatson) Sender: root@news-relay.eworld.com No. Philip Watson Matsushita Avionics Systems Corp., Irvine, CA --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to profile VxWorks tasks? Date: 31 Oct 1995 23:37:03 -0800 From: pwatson@eworld.com (PWatson) Organization: eWorld Message-ID: <47782v$2hk@hp5.online.apple.com> References: <46rbsn$8id@kannews.ca.newbridge.com> Reply-To: pwatson@eworld.com (PWatson) Sender: root@news-relay.eworld.com Check out spyLib. Philip Watson Matsushita Avionics Systems Corp., Irvine, CA --------------------------- End of New-News digest ********************** From lynn@taurustech.com Wed Nov 1 07:02:33 1995 From: Lynn Date: Wed Nov 1 07:02:35 PST 1995 Subject: GreenSprings IPMP (Z16C30) Hi All, Im looking for companies with drivers for sale or freebies that will control the GreenSprings IP-Multi-Protocol Serial board which utilizes the Z16C30. I will be putting four of these IPs on a Motorola MV162 runing VxWorks 5.1. My intended configuration is SDLC protocal with FM1 encoding. On person gave me some info but due to a system glitch I I lost his mail. Thanks in advance Beers Lynn From mpc@globesat.lanl.gov Wed Nov 1 07:46:13 1995 From: mpc@globesat.lanl.gov (Michael Caffrey) Date: Wed Nov 1 07:46:15 PST 1995 Subject: profiling vxworks Hi, I saw a message a while back asking about profiling vxworks tasks, but no reply. I would like to ask that question again, what is the easiest way to profile vxworks? I have briefly looked at windview and it looks like it may be possible to use, but it isn't clear how to interpret the time figures it produces, are they clock ticks or ms ? Also, it isn't clear whether the act of profiling actually effects the speed of execution; windview has a server task and the code must be compiled with a -g, both of which must slow a task. Is this the uncertainty principle in action? does it give a worst case only? Do any of you grizzled old vxperts have an opinion? Thanks a bunch.... ---------------------------------------------------------------- Michael Caffrey PHONE:(505) 667-2422 MS: D440 FAX:(505)665-4197 Los Alamos National Laboratory EMAIL: mpc@lanl.gov Los Alamos, NM 87545 GROUP: NIS-3 From jhillman@wrs.com Wed Nov 1 08:33:52 1995 From: jhillman@wrs.com (Jon Hillman) Date: Wed Nov 1 08:33:58 PST 1995 Subject: Re: profiling vxworks Michael Caffrey writes: > > Hi, > I saw a message a while back asking about profiling vxworks > tasks, but no reply. I would like to ask that question again, what is > the easiest way to profile vxworks? I have briefly looked at windview > and it looks like it may be possible to use, but it isn't clear how to > interpret the time figures it produces, are they clock ticks or ms ? > Also, it isn't clear whether the act of profiling actually effects the > speed of execution; windview has a server task and the code must be > compiled with a -g, both of which must slow a task. Is this the > uncertainty principle in action? does it give a worst case only? > Do any of you grizzled old vxperts have an opinion? > The easiest way to profile vxWorks tasks is with the spy profiler. There are other profilers, namely ScopeProfile from RTI which is part of Stethoscope. Timex and timexN are routines which can provide actual times of execution for routines based on when they begin and end. Provided with Windview is typically functions for doing timestamp so you can install your own timestamp points if you wish. I.e., see sysTimestamp, sysTimestampEnable, sysTimestampPeriod, etc. These would be included in the timer's driver source file. Wind View can be used to provide profile information but currently you have to gather the times for a period of time for all the task you want to profile and add them up, etc. WindView provides (when a timestamp timer is installed) up to nanosecond or microsecond resulotion for the times provided. Typically one would click the left mouse button at a couple of points (horizontally) on the graph and this would bound the area and print the time difference. The act of profiling typically would not affect the act of execution if the spy profiler task was a lessor priority than you application; otherwise it may. compiling with -g does not necessarily slow down the speed of execution, however it does not call the compiler's optimization either. I hope this gives you a clue on how you might do what you need. Jon Hillman Field Applications Engineer Wind River Systems Orlando, FL (407) 273-4646 jhillman@wrs.com From lynn@taurustech.com Wed Nov 1 08:42:30 1995 From: Lynn Date: Wed Nov 1 08:42:32 PST 1995 Subject: MVME162 IP interface Hi All, Does anbody know what the bus rate is for the IP interfaces on a Motorola MVME162, ie 8 or 32MHz? Thanks Beers Lynn From guthrie@power.amasd.anatcp.rockwell.com Wed Nov 1 08:58:36 1995 From: guthrie@power.amasd.anatcp.rockwell.com (Bob Guthrie) Date: Wed Nov 1 08:58:39 PST 1995 Subject: C++ with VxWorks To All: I have vxWorks 5.1.1 for the 680x0 and up till now have been using "C". I would like to start using C++ and have the following questions: 1) Does the basic GNU compiler support C++? 2) Does the basic vxGDB support C++? 3) Does VxWorks handle global object constructors and destructors? 4) What does the INCLUDE_CPLUS #define do? Robert Guthrie From stan@rti.com Wed Nov 1 09:13:19 1995 From: Stan Schneider Date: Wed Nov 1 09:13:21 PST 1995 Subject: Re: profiling vxworks >> Submitted-by mpc@globesat.lanl.gov Wed Nov 1 07:46:13 1995 >> Submitted-by: mpc@globesat.lanl.gov (Michael Caffrey) >> >> I saw a message a while back asking about profiling vxworks >> tasks, but no reply. I would like to ask that question again, what is >> the easiest way to profile vxworks? I have briefly looked at windview >> and it looks like it may be possible to use, but it isn't clear how to >> interpret the time figures it produces, are they clock ticks or ms ? >> Also, it isn't clear whether the act of profiling actually effects the >> speed of execution; windview has a server task and the code must be >> compiled with a -g, both of which must slow a task. Is this the >> uncertainty principle in action? does it give a worst case only? >> Do any of you grizzled old vxperts have an opinion? Hi, I've appended a short (marketing) description of our ScopeProfile product. This is included with StethoScope. -- Stan ============================================================================= = = = = Stan Schneider = email: stan@rti.com = = Real-Time Innovations, Inc. = Phone: (408) 720-8312 = = 954 Aster, Sunnyvale, CA 94086 = Fax: (408) 720-8419 = = = = ============================================================================= =============================================================================== Product name: ScopeProfile Overview: Real-time dynamic execution profiler. Category: Software, Development tools Highlights: Detailed procedure-by-procedure analysis of CPU usage. Tree or flat structure model. Quickly spot performance bottlenecks. Minimally intrusive. Run your code at full execution speed. No special compilation. Analyze already running code. ScopeProfile is a dynamic execution profiler for VxWorks. It shows you exactly where you're spending your CPU cycles. Profiling task 0x3b20e0: net4 Current sum % direct % 10.43 .12 _TcpWrite 10.31 0.01 . _write 10.27 .18 . . _iosWrite 10.06 .16 . . . _send 9.35 .64 . . . . _sosend 7.92 .27 . . . . . _tcp_usrreq 7.25 .88 . . . . . . _tcp_output 4.97 .49 . . . . . . . _ip_output 3.87 0.06 . . . . . . . . _eiOutput 3.80 .51 . . . . . . . . . _ether_output 2.32 .31 . . . . . . . . . . _eiTxStartup 1.12 1.12 . . . . . . . . . . . _bcopy .69 .69 . . . . . . . _cksum .63 .63 . . . . . _bcopy The example shows a network server's TCP write activity. The profile clearly shows the overhead incurred at each layer: the I/O system, the socket layer, TCP, IP, and the ethernet driver. Both cumulative (including all subroutines) and direct CPU usage of each routine are captured. You can easily see how much time was required to copy data (bcopy) and calculate checksums (cksum). The output can be plotted in real-time with StethoScope. ScopeProfile also keeps a task-level execution record, and can function as a graphical "spy". ScopeProfile is both powerful and easy to use. ScopeProfile will help you tune your time-critical system for maximum performance. From mhamilto@bbn.com Wed Nov 1 09:23:55 1995 From: Mark Hamilton Date: Wed Nov 1 09:23:57 PST 1995 Subject: unsubscribe... I'm currently trying to unsubscribe. I no longer have subscription info. Can somebody help... mark... ------------------------------------------------------------------------ Mark Hamilton Bolt Beranek & Newman Cambridge, MA 02138 voice: (617) 873-3927 email: mhamilton@bbn.com "DON'T PANIC!", Hitchhikers Guide to the Galaxy (Douglas Adams) From kwoods@aspect.win.net Wed Nov 1 10:07:19 1995 From: kwoods@aspect.win.net (Aspect Engineering) Date: Wed Nov 1 10:07:22 PST 1995 Subject: Connecting into RealTime Clock on PC motherboard with VXWorks I am porting to vxworks on a PC motherboard and WRS doesn't make use of the RealTime Clock in there BSP! Does anybody find it strange that a BSP for the PC doesn't support fundamental issues? They support IDE hard drives and a DOS file system, but if you modify a file on that IDE drive the date is 1/1/1980... The response from WRS is that most users don't care about "time" in the embedded market. What rock have they been under? The embedded market has been evolving and changing and along with this has been the need for supporting such devices as PCMCIA cards, Hard drives, floppy drives, and TIME. Has anybody addressed this issue and if so, can you give me some help. I am new to vxworks and therefore don't know all the tricks and secrets. Secondly, I have been experimenting with the floppy disk support, and find that if you modify a disk, remove it, and insert another disk, that the second disk is corrupted (FAT tables are altered with parts of the FAT from the previously modified disk). I can understand the first disk being corrupt if the FAT et. al. were not flushed to the disk prior to removal, but why the second disk? Doesn't there (WRS) floppy driver recognize disk changes? Has anyone seen this? Have a work around? And how do you mount the floppy file system if no disk is installed? Any help with these issues would be greatly appreciated... Kevin Woods kwoods@aspect.win.net Aspect Electronics 12740 Earhart Ave. Auburn CA 95602 (916) 887-1008 From ddavies@xmission.com Wed Nov 1 10:10:38 1995 From: Doug Davies Date: Wed Nov 1 10:10:40 PST 1995 Subject: Re: GreenSprings IPMP (Z16C30) > Submitted-by: Lynn > > Hi All, Im looking for companies with drivers for sale or freebies > that will control the GreenSprings IP-Multi-Protocol Serial board which > utilizes the Z16C30. I will be putting four of these IPs on a Motorola > MV162 runing VxWorks 5.1. My intended configuration is SDLC protocal with > FM1 encoding. On person gave me some info but due to a system glitch I I > lost his mail. Thanks in advance We contracted "Systems Integration Plus" to write a driver for the afore mentioned UART. You may want to contact them. We have used the driver and it works well. Gunnar Buzzard System Intregration Plus (602) 443-0454 gunnar_buzzard@imsystems.com www.sip.vme.com/sip Good luck, -Doug ddavies@xmission.com Douglas Davies From froeber@bbn.com Wed Nov 1 10:22:15 1995 From: Fred Roeber Date: Wed Nov 1 10:22:18 PST 1995 Subject: Re: MVME162 IP interface Lynn asks: > Hi All, Does anbody know what the bus rate is for the IP interfaces >on a Motorola MVME162, ie 8 or 32MHz? The MVME162 and MVME162LX boards have the 8MHz IP interfaces. I think the newer MVME162FX board has the 32MHz interfaces. Fred | Fred J Roeber, Bolt Beranek and Newman Inc | | 50 Enterprise Place Middletown, RI 02842-5202 | | froeber@bbn.com 401-849-2543 (X48) | From jhillman@wrs.com Wed Nov 1 10:27:03 1995 From: jhillman@wrs.com (Jon Hillman) Date: Wed Nov 1 10:27:06 PST 1995 Subject: Re: C++ with VxWorks Robert Guthrie writes: > > I have vxWorks 5.1.1 for the 680x0 and up till now have > been using "C". I would like to start using C++ and > have the following questions: > > 1) Does the basic GNU compiler support C++? There is a current version of G++ (GNU C++) which is rather good. GNU 2.6. You can obtain this from CYGNUS. > 2) Does the basic vxGDB support C++? No, the basic vxGDB does not. There is an add-on to vxworks 5.1.1 called WindC++ which can provide gdb support for C++ but it uses a cfront pre-processor, not G++. > 3) Does VxWorks handle global object constructors > and destructors? Yes, with the add-on windC++ product there is a library and mechanism to support static constructors and destructors. > 4) What does the INCLUDE_CPLUS #define do? This is for the add-on windC++ product. It is unbundled. To add, the Tornado product (with vxworks 5.3) provides C++ via the GNU C++ standard. Also available with Tornado are the Wind Foundation Classes which provide wrapper classes to the vxWorks OS; and also are iostreams, tools.h++ and booch classes. To find out more call your local sales rep or 1-800-545-9463 or e-mail to inquiry@wrs.com. Jon Hillman Field Applications Engineer Wind River Systems Orlando, FL (407) 273-4646 jhillman@wrs.com From jcn@kirok.atcwest.com Wed Nov 1 12:22:41 1995 From: jcn@kirok.atcwest.com (Chris Nickles) Date: Wed Nov 1 12:22:44 PST 1995 Subject: Re: MVME162 IP interface Lynn, On the MVME162FX it is jumper selectable. See p 2-13 of the 162FX installation manual. On the earlier boards it is 8 MHz. BTW, the master clock on the FX is 25 or 32 (not 33) MHz, for compatibility with the IP2 industry pack controller chip. This makes the system ticks and the bus timeouts run a little slow on the 32 MHz version. It can be fixed in sysLib.c; look for the comment "for 33MHz board" in two places. Chris Nickles Alliant Techsystems Advanced Technology Applications From froeber@bbn.com Wed Nov 1 12:38:06 1995 From: Fred Roeber Date: Wed Nov 1 12:38:09 PST 1995 Subject: Re: C++ with VxWorks Robert Guthrie asks: >I have vxWorks 5.1.1 for the 680x0 and up till now have >been using "C". I would like to start using C++ and >have the following questions: > >1) Does the basic GNU compiler support C++? >2) Does the basic vxGDB support C++? >3) Does VxWorks handle global object constructors >and destructors? >4) What does the INCLUDE_CPLUS #define do? Seems the easiest way to go to C++ with VxWorks is to go with the new Tornado product form WRS. Does cost a bit to upgrade :-( but gives you fully integrated C++ support. Other alternatives either also cost money (getting CenterLine C++ product WRS sells) or involve some pain (building your own GNU toolchain for C++). Fred | Fred J Roeber, Bolt Beranek and Newman Inc | | 50 Enterprise Place Middletown, RI 02842-5202 | | froeber@bbn.com 401-849-2543 (X48) | From bqv@se19.wg2.waii.com Wed Nov 1 13:25:10 1995 From: Bang Vu Date: Wed Nov 1 13:25:13 PST 1995 Subject: Re: profiling vxworks the vxWorks Users Group Exploder penned: >Submitted-by mpc@globesat.lanl.gov Wed Nov 1 07:46:13 1995 >Submitted-by: mpc@globesat.lanl.gov (Michael Caffrey) > > >Hi, > I saw a message a while back asking about profiling vxworks >tasks, but no reply. I would like to ask that question again, what is >the easiest way to profile vxworks? I have briefly looked at windview >and it looks like it may be possible to use, but it isn't clear how to >interpret the time figures it produces, are they clock ticks or ms ? >Also, it isn't clear whether the act of profiling actually effects the >speed of execution; windview has a server task and the code must be >compiled with a -g, both of which must slow a task. Is this the >uncertainty principle in action? does it give a worst case only? >Do any of you grizzled old vxperts have an opinion? > >Thanks a bunch.... > >---------------------------------------------------------------- >Michael Caffrey PHONE:(505) 667-2422 >MS: D440 FAX:(505)665-4197 >Los Alamos National Laboratory EMAIL: mpc@lanl.gov >Los Alamos, NM 87545 GROUP: NIS-3 > I use ScopeProfile which is part of StethoScope from RTI for profiling purposes and like the product. You may want to give it a try. You can ask for a demo license, contact RTI or WRS. -- Bang Vu | QC Tools, Inc. (a subsidiary of I/O, Inc.) vu@wg2.waii.com | Software, Systems, and Consultancy for the Voice 713-964-6192 | Exploration Industry. Fax 713-964-6218 | 3600 Briarpark, Houston, TX 77042 From froeber@bbn.com Wed Nov 1 15:10:36 1995 From: Fred Roeber Date: Wed Nov 1 15:10:41 PST 1995 Subject: Re: profiling vxworks Michael Caffrey asked about performance measurement tools and Jon Hillman responded with some good overview info. I am going to try to fill in some holes since I feel that the tools that VxWorks supports in this area are real useful and important in getting projects done in a timely fashion. Jon's comments are delimited by '>'. >The easiest way to profile vxWorks tasks is with the spy profiler. I agree, using the spyLib routines that come with VxWorks are the easiest mechanism to use to measure performance. However, they only give you information about CPU utilization at the task level. If you want to see what percentage of the CPU is being used by each task then they can be ok. One thing to note, though, is that the spy routines use the "auxiliary" clock on your CPU. This is not a problem unless your CPU doesn't have a second clock or you are already using it. Also, spy uses a "sampling" approach to measuring performance; it sees what task is running each time a periodic interrupt occurs. In some applications, this approach to performance measurement can yield inaccurate results. >There are other profilers, namely ScopeProfile from RTI which is >part of Stethoscope. The ScopeProfile routines in Stethoscope are quite powerful for doing profiling. The Stethoscope User Manual has a good description of the issues involved with profiling; where it works well and what the limitations are. Like any "profile" based performance measurement, ScopeProfile works by keeping track of what routines are running at periodic intervals. It gives information on what percentage of execution time is being used by the various global functions in VxWorks and your application. Results can be skewed if you have a lot of local functions or if your application has the same periodicity as the profiling clock (which, as with spyLib, is the "auxiliary" clock by default). Using the profiler is the best way to get an idea of percentage CPU utilization at the function level. >Timex and timexN are routines which can provide actual times >of execution for routines based on when they begin and end. These routines run a function and measure how many clock ticks elapse. You need to run enough iterations such that the elapsed time is a reasonable number of clock ticks. The resulting time never accounts for the overhead of the interrupts that occur while the timing is being done. These routines can be useful for testing simple functions or sequential pieces of an application. They are not typically much use for measuring performance of a complete application. >Wind View can be used to provide profile information but currently >you have to gather the times for a period of time for all the task >you want to profile and add them up, etc. WindView provides >(when a timestamp timer is installed) up to nanosecond or microsecond >resulotion for the times provided. Typically one would click the >left mouse button at a couple of points (horizontally) on the graph >and this would bound the area and print the time difference. Using WindView provides very different information than using the profiling tools. WindView shows a time line of critical events within a processor. Wind River has instrumented the VxWorks kernel to automatically provide information on kernel events. You can easily see the overhead due to interrupts and the effect of task switching. You can add event logging calls to your code to measure the duration of specific pieces of code that you are interested in. If your processor has a high resolution clock that the timestamp routines can use, then you can get very accurate measurements of how long things take. I have been using these specific VxWorks tools for a few years but have used event logging and profiling tools in general for many years. My experience is that profiling tools are good for optimizing application code. Tools like WindView are of much more general use for detecting why an application isn't working. I typically find event logging tools to be the primary debug tool during the stage of a project where you are integrating all the pieces together and want to see multitasking effects or find out why things stop working. >The act of profiling typically would not affect the act of execution >if the spy profiler task was a lessor priority than you application; >otherwise it may. Using the spy or profiling routines adds very little execution overhead to your application since they only collect information periodically. Using WindView can add a bit more overhead since events are typically logged at a fairly high rate and there is overhead for offloading the collected event data to a host for display. For a "normal" application something in the ballpark of 5% CPU utilization for event logging is about what you could expect. If you log lots of events, this figure can go up. It turns out that my company is getting ready to release a product that makes WindView operate even better in a number of ways. To avoid too much commercial content, I will simply say our product (called TraceMaker) works in VME based systems and: - takes care of offloading event data so overhead of logging is lower - makes WindView "post-mortem" mode easy to use - provides a microsecond clock that is used for all event timestamping; this allows accurate event timing within and between processors Anyway, we will be demoing the product at the VxWorks User Group meeting later this month; stop by and take a look if you want. Now back to the program. >compiling with -g does not necessarily slow down the speed of execution, >however it does not call the compiler's optimization either. Using the -g option with the GNU tool chain (unlike with many compilers) has no effect on code optimization. The -g option for debugging causes symbol table information to be included in your program. This makes the object code bigger and makes it take a little longer to load but doesn't make it run any differently. You need -g so that tools like scopeProfile can tell what part of your code is running when they sample execution. You don't need -g for spy or WindView use. >I hope this gives you a clue on how you might do what you need. I provided more detail here on the VxWorks performance assessment tools because I view them as an important resource in getting projects done well. Fred | Fred J Roeber, Bolt Beranek and Newman Inc | | 50 Enterprise Place Middletown, RI 02842-5202 | | froeber@bbn.com 401-849-2543 (X48) | From stan@rti.com Wed Nov 1 15:53:46 1995 From: Stan Schneider Date: Wed Nov 1 15:53:48 PST 1995 Subject: Re: profiling vxworks >> Submitted-by froeber@bbn.com Wed Nov 1 15:10:36 1995 >> Submitted-by: Fred Roeber >> Great summary! Thanks, Fred. A couple very minor corrections: * ScopeProfile does not require the "-g" flag, but it does require that you load the local symbols (ld "syms" flag = 1) * local functions shouldn't affect the profile * ScopeProfile can use any clock source (in fact, we ship (free to all!) drivers for the other two clocks available on VIC chips, e.g. the mv167 board. (send requests to "info@rti.com") We've had many requests to explain the diff between StethoScope, ScopeProfile, and WindView. Here's our canned answer; it seems appropriate here: >> What is the difference between StethoScope and WindView? >> There is a superficial similarity---both graphically represent information gathered from an executing VxWorks target---but there's very little real functional overlap. There are three "types" of visualization/analysis tools for VxWorks: a source-level debugger (vxgdb), a real-time data monitor (StethoScope), and a operating system visualization tool (WindView). A debugger can be thought of as similar to a voltmeter. It measures static values. It's nearly useless for finding peak values, occasional "glitches", or time-dependent quantities like durations of events or noise characteristics. StethoScope is like a digital oscilloscope; it displays all these with ease. WindView is like a logic analyzer. It doesn't show analog values (program variables), but it excels at displaying events, such as task switches, semaphore activity, interrupts, etc. In particular, StethoScope is a window into your *application*. You can view any variable or memory location in your system. For example, you can see how much overshoot your controller has, measure how long it took a door to close, analyze noise in a sensor, or see how long a data queue is getting. StethoScope will graphically display your data, collect it for analysis, and save it to disk. StethoScope helps you understand what your program is doing. WindView, on the other hand, allows you to visualize the relationships and timing between operating system events. For example, you can measure how long it took for an interrupt to execute, how much later that interrupt was serviced, etc. It's also useful for watching and understanding task switching activity: which tasks are running, why tasks are preempted, how long the task switch took, etc. Issues like semaphore priority inversion, lock-outs, and interrupt latency are easily revealed. So, StethoScope shows you what your VxWorks application is doing, and WindView shows you how VxWorks is running your application. There's a big difference. There's almost nothing WindView can show that could be viewed with StethoScope and vice-versa. On another note, StethoScope also includes a dynamic execution profiler known as ScopeProfile. ScopeProfile displays which routines in your system are using the CPU, as a percentage of total CPU time. It's very useful for tuning your application code for maximum performance. WindView is also useful for this purpose, but it shows entirely different information. The tools are quite complementary. For example, if your problem is that your application is thrashing (spending most of its time context switching), ScopeProfile will simply report that some large percentage of the CPU is being spent in the kernel. It can show you which kernel routines are active, but that may not be of much help. WindView will show you which tasks are switching, and display the events (e.g. interrupts) that are causing the switches. On the other hand, if the problem is that you have one task with inefficient code, WindView will only tell you that one task is running most of the time. ScopeProfile will break down the individual routines within the task that are burning the CPU in great detail. It will give you a direct map of what that task is doing, what routines are being called, what routines they call, and point out exactly where the inefficiencies are. ScopeProfile provides a detailed function-by-function analysis. So, both WindView and ScopeProfile will help tune the performance of your code. However, they provide different perspectives: WindView offers precise measurements of event timing, ScopeProfile provides a detailed analysis of where your CPU time is being used. -- Stan ============================================================================= = = = = Stan Schneider = email: stan@rti.com = = Real-Time Innovations, Inc. = Phone: (408) 720-8312 = = 954 Aster, Sunnyvale, CA 94086 = Fax: (408) 720-8419 = = = = ============================================================================= From daemon@csg.lbl.gov Thu Nov 2 04:00:28 1995 From: daemon@csg.lbl.gov Date: Thu Nov 2 04:00:34 PST 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Thu Nov 2 04:00:24 PST 1995 Subject: Green Hills Compilers/Tools - Good/Bad ? Subject: Re: VxWorks External Interrupt on MIPS processor Subject: *Synchronous* mode driver for Z85230(8530) needed Subject: Re: crosscompiling with gcc ? Subject: Re: Help on Command Line Argument passing Subject: Re: re: Bugs due to typecasts Subject: VXWORKS C UNIX Subject: Re: Multi-Cast Addressing Subject: Re: profiling vxworks Subject: Re: crosscompiling with gcc ? Subject: Need HELP with PCMCIA drivers ! ------------------------------------------------------- Newsgroups: comp.realtime,comp.os.vxworks,comp.unix.aix Subject: Green Hills Compilers/Tools - Good/Bad ? Date: 31 Oct 95 00:11:29 GMT From: gregb@den.mmc.com (gregb) Organization: Lockheed Martin -- Denver InterNetNews site Message-ID: Reply-To: gregb@gemini.den.mmc.com I have seen a demo of the Green Hills C compiler and their MULTI software development environment for a VxWorks target (from an AIX host). It looked to be quite good. Does Green Hills build a good reliable compiler? Is their support good? Any comments, negative or positive are welcome. gregb@gemini.den.mmc.com - -- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks External Interrupt on MIPS processor Date: 1 Nov 1995 14:26:17 GMT From: Don Small Organization: Sandia National Laboratories Message-ID: <478029$ocf@news.sandia.gov> References: Bob wrote: > Can someone point me towards some sample code that shows how to set up a > Heurikon R3xxx processor operating under VxWorks 5.x to accept/enable/etc > external interrupts over the VME backplane? > OUR SETUP: > R3500/VxWorks 5.1 in slot_0 on a VME backplane with an additional board > that generates VME interrupts. To test the software, we set up a watchdog > timerer that uses intConnect to kick-off an ISR when the wdTimer times > out. Now I'd like to use the same ISR to be kicked off when an external > interrupt is received, but we're lost in the overlapping options of > intEnable, intConnect, intVectEtc, etc, etc... > If there is a template that shows the steps necessary to set-up (enable, > identify, connect, etc) an interrupt, I'd love to see it. The first step is to configure vxWorks properly. It is shipped to not use vectored interrupts. You will probably want vectored interrupts which can be enabled by editing 'configAll.h'. The use of vectored interrupts allows vxWorks to automatically acknowledge the interrupt and inquire for the interrupt vector. Look for the following section: #if CPU_FAMILY==MIPS #define INT_LOCK_LEVEL 0x1 /* R3K interrupt disable mask */ #define ROOT_STACK_SIZE (20000) /* size of root's stack, in bytes */ #define SHELL_STACK_SIZE (100000) /* size of shell's stack, in bytes */ #define ISR_STACK_SIZE (5000) /* size of ISR stack, in bytes */ #define VEC_BASE_ADRS ((char *) 0x0) /* meaningless in R3k land */ #define VME_VECTORED TRUE /* use vectored VME interrupts */ #define TRAP_DEBUG 0 /* trap 0 - breakpoint trap */ #endif /* CPU_FAMILY==MIPS */ making sure that VME_VECTORED is set to TRUE. Rebuild vxWorks. The next step is to connect to the proper interrupt vector using intConnect. intConnect uses intVec??? to set up the vectors so these calls are not necessary. An example, we have an interrupt handler called 'mbIntHandler' (wow, what an original name) that should be executed when a board acknowledges using interrupt vector 0x85. The intConnect should look like: intConnect ( (NUM_TO_IVEC(0x85), mbIntHandler, NULL); Next, program your board to acknowledge with the proper vector number (in this example an 0x85) and interrupt priority level. The level is a priority between 1 & 7 with priority 7 being the highest. The loading of these values is board specific but must be done to properly generate interrupts on the CPU. A note for the Huerikon processor. There are two types of interrupts defined the VME specification. RORA (Release on register access) and ROAK (release on acknowledgement). The Heurikon processor only works correctly with ROAK interrupts. If the interrupting board uses RORA type interrupts then a second interrupt will be generated while the CPU is processing the first and vxWorks will panic and halt with an error (I can not remember the actual wording of the error but it is something like 'Exception during exception: halting'). The final step necessary is to enable interrupts on the CPU board. Interrupts come into the Mips CPU in a heirarchical manor. The CPU only has 8 interrupts of which 6 are hardware and 2 are software. intEnable()/intDisable() will enable/disable interrupts at this level with all VME interrupts physically tied to hardware interrupt 0 (the BSP [sysLib.c] allows you to prioritize CPU level interrupts from 0->7 or from 7->0). The BSP already enables VME interrupts at the CPU level so the use of intEnable/intDisable would be useful in your code only to stop all VME interrupts. The VME prioritized interrupts are all mapped to the single hardware interrupt 0 but have their own enable/disable. These are the levels that I mentioned in the previous paragraph. To enable/disable a particular interrupt level use sysIntEnable(level)/sysIntDisble(level). We usually set our boards to level 2 hence a sysIntEnable(2) is the proper value. A couple of suggestions. If a board interrupts a processor at a vector in which a interrupt handler does not yet exist, then vxWorks has a default handler that tells you and will inform you of the vector. I would suggest updating vxWorks (VME_VECTORED = TRUE), configuring your hardware, and using the sysIntEnable() as your first steps. Leave out the intConnect() so you don't have to debug you service routine yet. Once you see the vxWorks message that an interrupt is occuring you know you are on the right track. Then insert the intConnect and debug you service routine. Hope this helps. Don Small Sandia National Labs --------------------------- Newsgroups: comp.arch.embedded,comp.arch.bus.vmebus,comp.realtime,comp.os.vxworks Subject: *Synchronous* mode driver for Z85230(8530) needed Date: 1 Nov 1995 13:24:32 GMT From: michaelv@qualcomm.com (Michael Vakulenko) Organization: Qualcomm Israel Message-ID: Hi all, Does anybody know where I can find source code for Z85230(8530) driver for communication in SYNCHRONOUS mode? ~~~~~~~~~~~ I'm going to use Industry Pack of Actis Computers with two Z85230's under VxWorks. Thanx in advance, __________________________________________________________ Michael Vakulenko, Voice: +972-4-577999 Software Engineer, Fax: +972-4-577998 Qualcomm Israel, Ltd Email: michaelv@qualcomm.com __________________________________________________________ --------------------------- Newsgroups: comp.os.vxworks Subject: Re: crosscompiling with gcc ? Date: Wed, 01 Nov 1995 08:14:43 -0800 From: Strickland_Bob@eagle.lmsc.lockheed.com (Bob Strickland) Organization: Lockheed Martin Missiles & Space Co. Message-ID: References: <4737lr$d2m@news1.svc> In article <4737lr$d2m@news1.svc>, Arnaud.Girsch@marben.com wrote: > Hi, > > Has anyone ever configured (and made it work ! :-) ) a gcc on a Unix > system to x-compile some code for vxworks ? > I'm interested in experiences. Thanks. > > Arnaud. > > -- > Arnaud Girsch -+- agirsch@marben.com -+- Marben Products, Inc. - San Jose, CA > \- OSI Networking Products. We just set up our VxWorks environment for 5.1. There are two directroy trees, one vw/... the other gnu/... We set up a script to make the definitions and alias' (VX_VSP, _HOST, etc) used by the WRS makefiles. It's a simple matter to compile MIS/WxWorks objects on the Sun SPARC host and 'ld References: <473kp4$5jgp@unix1.sncc.lsu.edu> In article <473kp4$5jgp@unix1.sncc.lsu.edu>, eeliu@unix1.sncc.lsu.edu (Yu Liu) wrote: > I know I am going to ask a stupid question. But I do need the answer. > I wrote a C program which passed the command line argument to the program. > After I compiled and linked the program, I don't know how to run it. It > seems "run prog.exe argumant" does not work. Do I use "MCR"? What is it? > Is there any help online?(I could not find it). > > Any help is appreciated. > YU VxWorks 'programs' are different than 'normal' c programs, for starters, you don't need a 'main' statement. If you can get the WRS calss notes, there are some examples on how to compile/load user code. The basic process is: + boot the target + compile the file (ie foo.c) BUT use the WRS compiler and swtiches to prevent building an a.out file. + from the VxWorks shell do a 'ld < foo.o' to load the obj file. + from the shell type 'foo parameter' to run function foo and pass in 'parameter'. Good Luck Bob --------------------------- Newsgroups: comp.os.vxworks Subject: Re: re: Bugs due to typecasts Date: 1 Nov 1995 22:36:10 GMT From: Gene Bruce Organization: Texas Instruments Message-ID: <478soq$2li@mksrv1.dseg.ti.com> References: <9510311537.AA04547@bach.jhuapl.edu> Be careful with altering WRS-supplied files. When you purchase an upgrade, you could have a BIG mess on your hands attempting to incorporate these alterations that you are now dependent on into the upgrade files! Yuk! - -- Gene Bruce Texas Instruments History teaches that wars Systems Group begin when governments Electronic Systems Division believe the price of _/_/_/_/_/ _/_/_/ Ocean Surveillance Radar aggression is cheap. _/ _/ 2501 W. University, MS 8049 - Ronald Reagan _/ _/ McKinney, Texas 75070 _/ _/ Phone/Voice Mail: (214)952-4937 _/ _/_/_/ FAX: (214)952-2104 --------------------------- Newsgroups: comp.os.vxworks Subject: VXWORKS C UNIX Date: 1 Nov 1995 23:03:54 GMT From: jeff@scali.midcom.com (Jeff Schadoff) Organization: MIDCOM Corporation Message-ID: <478ucq$17b@liberty.liberty.com> Contract to perm opportunity for a vxworks unix c p/a in southern california Please fax resumes to 800-888-7899 or email to jeff@midcom.com Immediate hire for right individual - -- MIDCOM Corporation | MIDCOM specializes in contract programming, 4175 E. La Palma Ave., Ste. 200 | engineering and technical services. (Agency) Anaheim, CA 92807 | DISCLAIMER: Any opinion expressed is my own (714)579-3000 (714)961-1767 Fax | and is not necessarily that of MIDCOM Corp. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Multi-Cast Addressing Date: 1 Nov 1995 07:21:59 GMT From: squeeby@tuna.hooked.net (squeedy squeeby) Organization: Peaceful Star Message-ID: <47776n$575@its.hooked.net> References: <309555A2@gicpo.gic.gi.com> Reply-To: squeeby@hooked.net On Mon, 30 Oct 1995 13:44:00 -0800 (PST), HT-MS wrote: : : Question 1: Does this hook intercept the message AFTER the ARP resolution : has occurred? : in general yes. the problem is that not all network drivers in vxworks support etherOutputHook facility. it is not done consistently. : Question 2: Will I have to calculate the MAC layer checksum, or does the : driver handle that? : in general, you need not worry about MAC layer checksum. all of this hook-related software add-ons sound rather excruciating for working around lack of proper multicast support in vxworks. vxworks needs an update in their network software. bsd 4.4 network code includes all the multicast support. vxworks' bsd tahoe release based network code is at least 5 years old, and behind the times. http://www.hooked.net:80/users/squeeby/ --------------------------- Newsgroups: comp.os.vxworks Subject: Re: profiling vxworks Date: 1 Nov 1995 19:26:38 GMT From: Don Small Organization: Sandia National Laboratories Message-ID: <478hle$23d@news.sandia.gov> References: <9511011542.AA13229@globesat> mpc@globesat.lanl.gov (Michael Caffrey) wrote: > >Hi, > I saw a message a while back asking about profiling vxworks >tasks, but no reply. I would like to ask that question again, what is >the easiest way to profile vxworks? We use a Vmetro VME bus analyzer. By simply writing to a memory location a value of say 0x0 at the beginning of the function and a value of 0x1 at the end then the Vmetro displays the precise time required for the function. The only overhead is the actual VME writes which could be determined also with the Vmetro and removed from the overall time of the function. Of course, this is a real pain if you have either a lot of code or a need for very fine granularity over the entire program. This method has proven very useful for our needs. Don --------------------------- Newsgroups: comp.os.vxworks Subject: Re: crosscompiling with gcc ? Date: Wed, 01 Nov 1995 16:29:32 -0700 From: Barnacle Wes Organization: inteleNET Internet Services Message-ID: <309802DC.167EB0E7@intele.net> References: <4737lr$d2m@news1.svc> Arnaud Girsch wrote: > Has anyone ever configured (and made it work ! :-) ) a gcc on a Unix > system to x-compile some code for vxworks ? > I'm interested in experiences. Thanks. We do this at BTS. We run VxWorks on MC68302 boards, and use GCC 2.6 as a cross-compiler on HP 9000/712 workstations. Actually, GCC is built to generate code for SunOS 4.1 on a Sun-3 (MC68020) system, and then the VxWorks linker is used to link the object modules with the VxWorks system. I've only been working here for three days, so I'm not all that familiar with the system yet. Send me e-mail at wes@btsslc.com and I'll forward your questions to the staff tool builders. - -- Wes Peters | Yes I am a pirate, two hundred years too late Softweyr | The cannons don't thunder, there's nothing to plunder Consulting | I'm an over forty victim of fate... wes@intele.net | Jimmy Buffet --------------------------- Newsgroups: comp.os.vxworks Subject: Need HELP with PCMCIA drivers ! Date: 2 Nov 1995 06:18:10 GMT From: farshid@ix.netcom.com (farshid sabet) Organization: Netcom Message-ID: <479nr2$pj1@ixnews3.ix.netcom.com> References: <478ucq$17b@liberty.liberty.com> Does anybody know of PCMCIA ATA drivers for VxWorks? Farshid --------------------------- End of New-News digest ********************** From chh@everest.esec.ch Thu Nov 2 08:03:15 1995 From: Christopher Hegarty Date: Thu Nov 2 08:03:18 PST 1995 Subject: Job Offer at ESEC SA, Cham Switzerland ESEC SA is the leading supplier of precision equipment and automation systems for semiconductor assembly. We are based in Cham, Switzerland and have sales and services offices in Europe, Asia and the USA. We are looking for an expert in real-time machine control systems to lead a group in the development of a new generation of die bonders and wire bonders. The duties of this position include the following: + Based on the known requirements for a new generation of equipment, develop a concept for a new software and hardware architecture for a control system. + Evaluate the chosen software and hardware system and develop a "proof of concept" prototype. + Lead a team to develop the first generation of equipment based on the new control system. Successful candidates should have the following qualifications: + A least four years of experience in developing real-time software for high-speed control systems. + Familiarity with modern software and hardware architectures for machine control. + Experience with modern software engineering methods. A major advantage would be experience with object-oriented design and analysis methods (e.g., Booch, Rumbaugh or ROOM). + Good communication skills and ability to work well in a team: the project involves intensive cooperation with the developers of our existing control systems and marketing. + A degree in software engineering with detailed knowledge of computer hardware or electrical engineering. Spoken or written knowledge of German would be an advantage, but is not essential, so long as a willingness to learn German is present. Experience in managing a group of engineers would be valuable, but is not essential. ESEC offers competitive compensation and a bonus program to reward outstanding contributions. Support for relocation will be provided. If you are interested in this position, please send a resume and cover letter to: Mrs. B. Golay ESEC SA Hinterbergstr. 32 CH-6330 CHAM Switzerland Phone: +41 42 44 22 77 Fax: +41 42 41 64 84 Or reply to aljobs@al.esec.ch. Please do not reply to this account. From tsb@mclean.sparta.com Thu Nov 2 12:06:14 1995 From: "Todd S. Brackett" Date: Thu Nov 2 12:06:21 PST 1995 Subject: SM available for SOLARIS 2.4 PRESS RELEASE 2 November 1995 SPARTA Custom Computer Products Directorate 7926 Jones Branch Drive, Suite 900 McLean, VA 22102 (703) 448-0210 SPARTA ANNOUNCES INTRODUCTION OF VXWORKS(R) SM(TM) DRIVER SUPPORT FOR SOLARIS 2.4. VxWorks(R) Shared Memory backplane network support allows TCP/IP networked communication between a SOLARIS(R) 2.4 VMEbus host and a VxWorks VMEbus Target. McLEAN, VIRGINIA , 2 NOVEMBER 1995 -- SPARTA, INC. announces the introduction of a VxWorks(R) SM(TM) (Wind River Systems, Alameda, CA) device driver for SOLARIS(R) 2.4. The initial target support is available on the THEMIS 5/64(TM) microSPARC based VMEbus SBC (Themis Computer, Fremont, CA). The driver fully supports full TCP/IP communications between multiple VxWorks and SOLARIS 2.4 hosts over the VMEbus backplane. Peer-To-Peer SOLARIS support is also offered without the requirement for a VxWorks target in the system. This development was done as a cooperative effort between SPARTA, Wind River Systems and Themis Computer. Follow on versions will be available for the THEMIS 10MP, 20MP, LXE+ and other SPARC boards. The product will sell for a $4500.00 per project fee with $295.00 right to use license fee per additional SOLARIS target. Source code license is also available. Call 1-800-228-8914 or e-mail info@mclean.sparta.com for more details. http://www.mclean.sparta.com. ------------------------------------------------------------ |Todd S. Brackett | Voice -> 703 448 1683 x249 | |Director, Computer Products| FAX -> 703 734 3323 | |EMAIL: | On the WEB! | |tsb@mclean.sparta.com | http://www.mclean.sparta.com | |___________________________|______________________________| | | | | /\\ SPARTA, Inc. | "Never try and | | ///\\ 7926 Jones Branch Drive | teach a pig to | | /////\\ Suite 900 | sing. It wastes | | ///// \\ McLean, VA 22102 | your time and | | / \\\\\ \\ 703-448-0120 | annoys the pig." | | \ \\\\\ // | | | \ ////// | | | \////// SPARTA | -M. Twain | | \//// Pride In Performance | | | \// -------------------- | | ------------------------------------------------------------ From ddavies@xmission.com Thu Nov 2 12:13:50 1995 From: Doug Davies Date: Thu Nov 2 12:13:53 PST 1995 Subject: Re: C++ with VxWorks > > 2) Does the basic vxGDB support C++? > > No, the basic vxGDB does not. There is an add-on to vxworks 5.1.1 > called WindC++ which can provide gdb support for C++ but it > uses a cfront pre-processor, not G++. Realize that the current WindC++ product (based on cfront) does not support 'volatile'. It also has a problem handling templates. We have run into a couple of other problems with WindC++ that we have had to work around. > To add, the Tornado product (with vxworks 5.3) provides > C++ via the GNU C++ standard. Also available with Tornado > are the Wind Foundation Classes which provide wrapper classes > to the vxWorks OS; and also are iostreams, tools.h++ and booch > classes. Hopefully, this will be an improvement... -Doug ddavies@xmission.com Douglas Davies From daemon@csg.lbl.gov Fri Nov 3 04:00:30 1995 From: daemon@csg.lbl.gov Date: Fri Nov 3 04:00:49 PST 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Fri Nov 3 04:00:25 PST 1995 Subject: Re: Loading VADS over VxWorks Subject: Re: reloading a program in vxworks Subject: Re: crosscompiling with gcc ? Subject: Access Fault under VxWorks Subject: Re: C++ with VxWorks Subject: Re: crosscompiling with gcc ? Subject: Problem with lseek() call Subject: Alpha AXPvme VME bus Access Subject: Shells for VxWorks ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Loading VADS over VxWorks Date: 2 Nov 1995 08:36:52 -0500 From: jcreem@rapnet.sanders.lockheed.com (Jeffrey M. Creem x5700) Organization: Lockheed Sanders Message-ID: <47ahhk$dvc@rapnet.sanders.lockheed.com> References: In article David Rourke {78523} writes: > > >I would like to know if it is possible to load a Rational VADS micro >kernel and TDM over VxWorks. > >My environment is: > >VADSCross, Version 39 E Sun (Sun OS) to 68360 >Matrix CPU36E target board >VxWorks BSP Version 1.0 >VxWorks Version 5.2 >Wind Kernel Version 2.4 > - -- STUFF DELETED >This version of VADS doesn't support VxWorks and our ultimate target >hardware won't have VxWorks. This is the only board I have for initial >development. > Doesn't sound quite right. The TDM expects that it has full control of the machine and I know the kernel is going to muck with low level registers. Sound like what you need is 1) Remove the vxWorks Proms and stick in some debugger prom that allows for serial transfer of S-recors: or 2) burn the tdm into roms and stick those in the board or 3) Get the VADSworks compiler (that supports vxWorks) use it for development and debug (And make sure people only use packages from target/standard and target/vads_cross and it should port fairly well to a bare VADScross machine. Jeff Creem --------------------------- Newsgroups: comp.os.vxworks Subject: Re: reloading a program in vxworks Date: 2 Nov 1995 16:04:11 GMT From: john.weir@msfc.nasa.gov (John Weir) Organization: NASA MSFC Message-ID: <47aq5r$99o@hammer.msfc.nasa.gov> References: There is also the reload command - reld which does an unld followed by a ld. This should free up the RAM occupied by the code loaded the first time to allow enough space to load it again. This command is available undel VxWorks 5.2 I know In article , matusewicz@norden.com says... > >After I load and run my program once onto our power pc 604 boards, I load >the program again and I get the following message: >0x1fc32d0 (TShell): memPartAlloc: block too big - 7539968 in partition >0x1b5928. >could not allocate text and data segments. > >The only way I have been able to correct this problem is to reboot the >board. Is there a command that can clear the memory for me? > >Thank you for your help. >Mindy Matusewicz > --------------------------- Newsgroups: comp.os.vxworks,gnu.gcc.help Subject: Re: crosscompiling with gcc ? Date: Thu, 02 Nov 1995 08:11:48 -0800 From: steve@tsunami.jpl.nasa.gov (Stephen Mitchell) Organization: JPL Message-ID: References: <4737lr$d2m@news1.svc> <309802DC.167EB0E7@intele.net> In article <309802DC.167EB0E7@intele.net>, Barnacle Wes wrote: > Arnaud Girsch wrote: > > Has anyone ever configured (and made it work ! :-) ) a gcc on a Unix > > system to x-compile some code for vxworks ? > > I'm interested in experiences. Thanks. > > We do this at BTS. We run VxWorks on MC68302 boards, and use GCC 2.6 > as a cross-compiler on HP 9000/712 workstations. Actually, GCC is > built to generate code for SunOS 4.1 on a Sun-3 (MC68020) system, > and then the VxWorks linker is used to link the object modules with > the VxWorks system. We're doing this here at JPL as well. We have gcc 2.6 on SunOS4 and Solaris2 working as a cross-compiler for 68k and r3000 SBC's. We have been trying to get gcc 2.6 or 2.7 to work as a cross-compiler for the PowerPC and/or a RAD/6000 (RS/6000 based) SBC. We've not had any luck in this area, and currently have to use an AIX 3.2 machine to do development for our RAD board. If anybody out there has got the gcc compiler to cross-compile to the RS/6000, I sure would like to hear from you.. - --steve - -- Steve Mitchell KD6BET TIP#168 steve@tsunami.jpl.nasa.gov "Everything in this post may be wrong" / "Make love, not Perl code." --------------------------- Newsgroups: comp.os.vxworks Subject: Access Fault under VxWorks Date: 2 Nov 1995 16:56:03 GMT From: john.weir@msfc.nasa.gov (John Weir) Organization: NASA MSFC Message-ID: <47at73$aqt@hammer.msfc.nasa.gov> I am attempting to execute a program on a Heurikon Nitro 60 under VxWorks 5.2 which uses a lot of double precision floating point. I keep getting an Access Fault error with an Access Address the same as the PC when the fault occurs. I can dump the code at this address using l and the code looks just fine. Also, I ran under VxGdb and found that the error occurs at different lines in the program if I run more than once and does not occur the first time the offending line is reached but after a few iterations. Does anybody out there have any idea what this Access Fault means and how I make it go away? BTW we do not have the optional virtual memory products. Thanks! John Weir --------------------------- Newsgroups: comp.os.vxworks Subject: Re: C++ with VxWorks Date: 2 Nov 1995 19:38:18 GMT From: Graham Waters Organization: The University of British Columbia Message-ID: <47b6na$3ep@nntp.ucs.ubc.ca> References: <9511011633.AA17067@power.amasd.anatcp.rockwell.com> guthrie@power.amasd.anatcp.rockwell.com (Bob Guthrie) wrote: > > To All: > > I have vxWorks 5.1.1 for the 680x0 and up till now have > been using "C". I would like to start using C++ and > have the following questions: > > 1) Does the basic GNU compiler support C++? > 2) Does the basic vxGDB support C++? > 3) Does VxWorks handle global object constructors > and destructors? > 4) What does the INCLUDE_CPLUS #define do? > > > Robert Guthrie Answers 1) gnu compiler 2.5.8 and above supports C/C++ Binaries for sun SPARCstations are available from the vxWorks archives (ftp://ftp.atd.ucar.edu/pub/vxworks/) 2) vxGDB supports gnu C++, I use it all the time. 3) maybe someone else has a good answer 4) I've never come across this one Check out FAQ item 9. Graham Waters E-mail waters@triumf.ca Control System Eng TRIUMF Meson Research Facility Vancouver, CANADA --------------------------- Newsgroups: comp.os.vxworks Subject: Re: crosscompiling with gcc ? Date: Thu, 2 Nov 1995 22:00:25 GMT From: marklz@topaz.technion.ac.il (Mark Zusman) Organization: Technion, Israel Institute of Technology Message-ID: References: <4737lr$d2m@news1.svc> <309802DC.167EB0E7@intele.net> Sender: news@wang.com Barnacle Wes (wes@intele.net) wrote: : Arnaud Girsch wrote: : > Has anyone ever configured (and made it work ! :-) ) a gcc on a Unix : > system to x-compile some code for vxworks ? : > I'm interested in experiences. Thanks. I did it for gcc/g++ 2.6.3 from HP-UX 9/Linux 1.2/SunOS4 to VxWorks 5.1.1 - it works perfectly. I haven't tried 2.7.0 buty it should be even easier - wrs* now is supported in configure . Mark Zusman. --------------------------- Newsgroups: comp.os.vxworks Subject: Problem with lseek() call Date: 2 Nov 1995 22:44:32 GMT From: ggross@inetgw.fsc.ibm.com (Glenn Gross) Organization: Loral Federal Systems Message-ID: <47bhkg$dgp@lfsserv1.fsc.ibm.com> Reply-To: ggross@lfs.loral.com I am having trouble with some code I wrote to open a file, position within that file and then read data from that file. This code has worked under HP-UX and AIX but I can't get it to work under VxWorks. Here is the section of code that I am having problems with: int msgs_fd, rc, pm_msg_num = 1; msgs_fd = open ("uccpm_msgs.data", O_RDONLY, 0); rc = lseek(msgs_fd, (sizeof(PM_MSG_INFO_TYPE) * pm_msg_num), 0); printf("return code from lseek = %d; errno = %d\n", rc, errno); PM_MSG_UNFO_TYPE is a valid typedef and it is 92 bytes. The code compiles fine and runs, as I said under HP-UX and AIX. However, when I run it under VxWorks the return code I get from the lseek call is -1 and the errno = 786434. Any help as to what the problem is would be appreciated. Glenn --------------------------- Newsgroups: comp.os.vxworks Subject: Alpha AXPvme VME bus Access Date: 3 Nov 1995 05:17:58 GMT From: "Phil S. Wilshire" Organization: Internet On-Ramp, Inc. Message-ID: <47c8m6$odi@express.ior.com> Help ! I am trying to rush a program into VxWorks. I have spent a few weeks perfecting an application under Digital Unix only to find that it currently can't do fast stuff and handle file access. So off to VxWorks we go. The problem is How do I access the vme bus outbound ?? I've RTFM managed to get the system booting and got some user test programs running fine. I've looked at adpCreateMapPath but I think I need a working example of how to make this work. I understand more than I want to about sparse and dense addressing space and about A32 A24 A16 bytes ,ints, quads and words. I just can't get a simple (??) call to adpCreateMapPath working yet. Any clues out there ??? Thanks I am in a bit of a rush to get this done. Hope to return the favor. Phil Wilshire I do computer rolling mill control stuff in Spokane WA. --------------------------- Newsgroups: comp.os.vxworks Subject: Shells for VxWorks Date: Thu, 2 Nov 1995 21:00:28 GMT From: Perry Wintner Organization: Ariel Corporation Message-ID: Sender: news@ariel.com Does anyone have an improved shell for VxWorks 5.1.1 or 5.2. I am looking for a shell with Emacs-like or tcsh-like (or both) commands including command history and text editing keystrokes. Any help would be appreciated. Perry - -- _____________________________________________________________________________ _ Perry Wintner Email: perry.wintner@ariel.com Applications Engineer Phone: (908) 249-2900 ARIEL CORPORATION Fax : (908) 249-2123 ariel@ariel.com BBS : (908) 249-2124 _____________________________________________________________________________ _ --------------------------- End of New-News digest ********************** From aisie.aisinc.com!aisinc.com!pew@msen.com Fri Nov 3 07:33:59 1995 From: pew@aisinc.com (Paul E. Wilt) Date: Fri Nov 3 07:34:02 PST 1995 Subject: Re: Problem with lseek() call ggross@inetgw.fsc.ibm.com (Glenn Gross) wrote: >I am having trouble with some code I wrote to open a file, position within that >file and then read data from that file. This code has worked under HP-UX and >AIX but I can't get it to work under VxWorks. Here is the section of code that >I am having problems with: > > int msgs_fd, rc, pm_msg_num = 1; > > msgs_fd = open ("uccpm_msgs.data", O_RDONLY, 0); > rc = lseek(msgs_fd, (sizeof(PM_MSG_INFO_TYPE) * pm_msg_num), 0); > printf("return code from lseek = %d; errno = %d\n", rc, errno); > >PM_MSG_UNFO_TYPE is a valid typedef and it is 92 bytes. > >The code compiles fine and runs, as I said under HP-UX and AIX. However, when >I run it under VxWorks the return code I get from the lseek call is -1 and the >errno = 786434. > >Any help as to what the problem is would be appreciated. Glenn: I believe the errno is telling you that the file does not exist. If you print the error number out in hex you get 0xc0002. Go to the h/ directory under the top level vxWorks directory tree and you will find errno.h and vwModNum.h. The header vwModNum.h contains the following: [From vwModNum.h]: #define M_ioLib (12 << 16) which is 0xc0000. This tells you that the error is in an ioLib function. The 0x0002 indicates an error number of 2: [From errno.h]: #define ENOENT 2 /* No such file or directory */ I would imagine that the open actually failed and you did not catch it. Are you using an NFS-mounted file system or are you using some other file system? In either case you need to change to the same directory where the file exists. Hope this helps, Paul -- ============================================================================= Paul Wilt, Sr Software Engineer (313) 995-2035 Applied Intelligent Systems, Inc. pew@aisinc.com 110 Parkland Plaza John 3:16 Ann Arbor, MI 48103 For God so loved the world ... ============================================================================= From randy@opid330.dsd.litton.com Fri Nov 3 09:42:14 1995 From: randy@opid330.dsd.litton.com (Randy Ryan) Date: Fri Nov 3 09:42:17 PST 1995 Subject: Re: Problem with lseek() call > int msgs_fd, rc, pm_msg_num = 1; > > msgs_fd = open ("uccpm_msgs.data", O_RDONLY, 0); > rc = lseek(msgs_fd, (sizeof(PM_MSG_INFO_TYPE) * pm_msg_num), 0); > printf("return code from lseek = %d; errno = %d\n", rc, errno); Just in case you didn't already, double check and make sure you didn't get a file error when attempting to open the file. Periodically, there can be file permissions problem quirks that pop up due to any number of things with a VxWorks target. Good luck, Randy From stan@rti.com Fri Nov 3 09:47:09 1995 From: Stan Schneider Date: Fri Nov 3 09:47:11 PST 1995 Subject: Re: Shells for VxWorks >> Date: Thu, 2 Nov 1995 21:00:28 GMT >> From: Perry Wintner >> Does anyone have an improved shell for VxWorks 5.1.1 or 5.2. I am looking >> for a shell with Emacs-like or tcsh-like (or both) commands including command >> history and text editing keystrokes. Any help would be appreciated. RTILib includes a utility called "cle" that lets any tty program use emacs/tcsh command-line editing. That includes our "rshell" reentrant shell and the VxWorks native shell. -- Stan ============================================================================= = = = = Stan Schneider = email: stan@rti.com = = Real-Time Innovations, Inc. = Phone: (408) 720-8312 = = 954 Aster, Sunnyvale, CA 94086 = Fax: (408) 720-8419 = = = = ============================================================================= From perryd@boris.lbl.gov Fri Nov 3 10:27:59 1995 From: "David B. Perry" Date: Fri Nov 3 10:28:02 PST 1995 Subject: Re: Problem with lseek() call On Nov 3, 7:34am, the vxWorks Users Group Exploder wrote: > Subject: Re: Problem with lseek() call > Submitted-by: pew@aisinc.com (Paul E. Wilt) > ggross@inetgw.fsc.ibm.com (Glenn Gross) wrote: > >I am having trouble with some code I wrote to open a file, position within that > >file and then read data from that file. This code has worked under HP-UX and > >AIX but I can't get it to work under VxWorks. Here is the section of code that > >I am having problems with: > > > > int msgs_fd, rc, pm_msg_num = 1; > > > > msgs_fd = open ("uccpm_msgs.data", O_RDONLY, 0); > > rc = lseek(msgs_fd, (sizeof(PM_MSG_INFO_TYPE) * pm_msg_num), 0); > > printf("return code from lseek = %d; errno = %d\n", rc, errno); > > > > Glenn: > > I believe the errno is telling you that the file does not exist. If you > print the error number out in hex you get 0xc0002. > > ... > > I would imagine that the open actually failed and you did not catch it. > Are you using an NFS-mounted file system or are you using some other file > system? In either case you need to change to the same directory where the > file exists. > > Hope this helps, > Paul > > -- > ============================================================================= > Paul Wilt, Sr Software Engineer (313) 995-2035 > Applied Intelligent Systems, Inc. pew@aisinc.com > 110 Parkland Plaza John 3:16 > Ann Arbor, MI 48103 For God so loved the world ... > ============================================================================= > This looks like another facet of the problem I posted some time back. It seems that if you are NOT using NFS, vxWorks will NOT give you *ANY* kind of warning or error return if you open or fopen a file that doesn't exist on another machine. This despite the verbage to the contrary in the vxWorks reference manual. Another 'feature'... thanks WRS. -- ===================================================================== David B. Perry TRW e-mail: perryd@boris.msfc.nasa.gov 213 Wynn Dr. phone: (205) 961-1033 Huntsville, AL 35805 fax: (205) 544-2238 ===================================================================== From daemon@csg.lbl.gov Sat Nov 4 04:00:37 1995 From: daemon@csg.lbl.gov Date: Sat Nov 4 04:00:41 PST 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sat Nov 4 04:00:28 PST 1995 Subject: Re: help !!! Subject: long long but only add and subtract Subject: Re: Connecting into RealTime Clock on PC motherboard with VXWorks Subject: re: MVME162 IP interface Subject: mouse driver details Subject: re: Problem with lseek() call Subject: Re: Shells for VxWorks Subject: Re: SLIP vs. NFS issues and problems Subject: Re: Serial file transfer s/w? Subject: Re: mouse driver details ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: help !!! Date: 3 Nov 1995 12:05:55 GMT From: msinger@griffon.oe.fau.edu (Matthew Singer) Organization: Florida Atlantic University Message-ID: <47d0j3$l14@cybernet.cse.fau.edu> References: <46tak1$jmt@fnnews.fnal.gov> Reply-To: msinger@griffon.oe.fau.edu (Matthew Singer) Sender: msinger@griffon (Matthew Singer) - -- In article <46tak1$jmt@fnnews.fnal.gov>, kccheng@hycpsg01.fnal.gov (Kuang-chun Cheng) writes: |>Hi, |> |> I am using vxWorks 5.1 in MV167 cpu. |> Could somebody tell me what's wrong of the following code ? |> |> double counter=0.; |> display() { |> printf("\n"); |> for(;;) { |> printf("%E\r", counter); |> taskDelay(100); |> } |> } |> add_number() { |> taskSpawn("t_Monitor", 80, 0, 20000, display); |> for(;;) { |> counter += 1.; |> } |> |> As I run this code in my MV167, I got the number just stay between |> x.xxE+05 and x.xxE+06 where 0. < x.xx < 10.. It didn't add one into |> counter !!!! |> |> |> Thanks for any help. |> |> Kuang-chun Cheng (kccheng@phys.sinica.edu.tw |> In the taskSpawn, you are not telling the kernel that the task uses floating point, so when a context switch occurs at the taskDelay, the registers are trashed... #include #include double counter=0.; display() { printf("\n"); for(;;) { printf("%f %e\n", counter, counter); taskDelay(100); } } add_number() { taskSpawn(0, 80, VX_FP_TASK, 20000, display,0,0,0,0,0,0,0,0,0,0); for(;;) { counter += 1.; } } - -------------------------------------------------------------------------- Matthew R. Singer msinger@oe.fau.edu Systems Engineer Voice: 407-367-2947 Florida Atlantic University Fax : 407-367-3885 Dept of Ocean Engineering - ------------------------------------------------------------------------------- --------------------------- Newsgroups: comp.os.vxworks Subject: long long but only add and subtract Date: Fri, 03 Nov 1995 02:29:57 GMT From: duane6@ix.netcom.com Organization: Netcom Message-ID: <47buu3$lde@ixnews7.ix.netcom.com> The following code results in some undefined symbols at "ld" time /* compile line is: ccsparc -c -DCPU=SPARC dca1.c */ double foo(void* SECtime) { return (double)(((*(unsigned long long*)SECtime)>>9)*5)/10000; } /* ld results are: */ /* - -> ld < dca1.o undefined sybol: ___lshrdi3 undefined symbol: ___ashldi3 undefined symbol: ___floatdidf undefined symbol: ___cmpdi2 ld error: error reading file (errno = 0x1c0001). value = 0 = 0x0 */ WRS says only add and subtract are supportted with "long long" data types. Please tell me it ain't so!!!! Anyone doing 64 bit divides, multiplies???? Note: my target is a FORCE10... Thanks to all that respond... duane6@ix.netcom.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Connecting into RealTime Clock on PC motherboard with VXWorks Date: Fri, 03 Nov 1995 02:30:59 GMT From: duane6@ix.netcom.com Organization: Netcom Message-ID: <47bv03$lde@ixnews7.ix.netcom.com> References: <90@aspect.win.net> kwoods@aspect.win.net (Aspect Engineering) wrote: >I am porting to vxworks on a PC motherboard and WRS doesn't make >use of the RealTime Clock in there BSP! Does anybody find it >strange that a BSP for the PC doesn't support fundamental issues? >They support IDE hard drives and a DOS file system, but if you >modify a file on that IDE drive the date is 1/1/1980... The >response from WRS is that most users don't care about "time" in >the embedded market. What rock have they been under? The embedded >market has been evolving and changing and along with this has been >the need for supporting such devices as PCMCIA cards, Hard drives, >floppy drives, and TIME. Has anybody addressed this issue and if >so, can you give me some help. I am new to vxworks and therefore >don't know all the tricks and secrets. I learned like you that the BSP for the FORCE10 cpu does not support the realtime clock. But using the documentation that came with the board I wrote some code to set and read the realtime clock (RTC). If your card is like the FORCE10 it's probably in BCD and with a little work you could set the clock and then come up with a scheme for updating the time words used by the DOS file system. fsDosSetTime (I think?) and fsDosSetDate (I think?). There's also a routine that you can hook a routine to that returns time to the DOS file system so that when the file system needs to update its time it calls your routine (returns the RTC). At least for you they support the serial interface .... they didn't on the FORCE10... Again for me a write your own problem... Does seem like a rip off... Good Luck Duane >Secondly, I have been experimenting with the floppy disk support, >and find that if you modify a disk, remove it, and insert another >disk, that the second disk is corrupted (FAT tables are altered >with parts of the FAT from the previously modified disk). I can >understand the first disk being corrupt if the FAT et. al. were >not flushed to the disk prior to removal, but why the second disk? >Doesn't there (WRS) floppy driver recognize disk changes? Has >anyone seen this? Have a work around? And how do you mount the >floppy file system if no disk is installed? >Any help with these issues would be greatly appreciated... >Kevin Woods >kwoods@aspect.win.net >Aspect Electronics >12740 Earhart Ave. >Auburn CA 95602 >(916) 887-1008 --------------------------- Newsgroups: comp.os.vxworks Subject: re: MVME162 IP interface Date: Fri, 3 Nov 95 14:10:58 CET From: kai@ade.no (Kai Atle Myrvang) Message-ID: <3a0aolu5f@ade.no> References: <9511011822.AA10511@lbl.gov> Reply-To: kai@ade.no >Lynn asks: > >> Hi All, Does anbody know what the bus rate is for the IP interfaces >>on a Motorola MVME162, ie 8 or 32MHz? > >The MVME162 and MVME162LX boards have the 8MHz IP interfaces. I think the >newer MVME162FX board has the 32MHz interfaces. Fred > > | Fred J Roeber, Bolt Beranek and Newman Inc | > | 50 Enterprise Place Middletown, RI 02842-5202 | > | froeber@bbn.com 401-849-2543 (X48) | I'm not sure, but don't you have to have same speed for all IPs on one module? There are non-Mota alternatives that let you select separate clocks for each IP. Nice if you have one IP that must work with 8MHz. - -- * Kai Atle Myrvang (kai@ade.no) * * AD Elektronikk AS, P.O.Box 104, N-1405 Langhus, Norway * * Tel: int+47 64 86 99 70 | Fax: int+47 64 86 99 20 * ********************************************************************* --------------------------- Newsgroups: comp.os.msdos.programmer,comp.sys.ibm.pc.misc,comp.sys.ibm.pc.hardware.misc,sci.electronics,comp.os.vxworks Subject: mouse driver details Date: 3 Nov 1995 16:13:15 GMT From: peter@cs.strath.ac.uk (Peter Barrie) Organization: Computer Science Dept., Strathclyde University., Glasgow, Scotland. Keywords: mouse driver Message-ID: <47df2r$ut@dunlop.cs.strath.ac.uk> I need to write a mouse driver for a PC motherboard running VxWorks real-time kernel. The serial ports are available but I need to get the spec for ANY serial mouse (baud rate , no-of-bits, parity, serial protocol , commands and data-packet spec) . Can anybody help with this or point me in the right direction ? I need to set this up fairly urgently. Thanks Pete Barrie --------------------------- Newsgroups: comp.os.vxworks Subject: re: Problem with lseek() call Date: 3 Nov 1995 18:22:41 GMT From: ggross@inetgw.fsc.ibm.com (Glenn Gross) Organization: Loral Federal Systems Message-ID: <47dmlh$8f1@lfsserv1.fsc.ibm.com> References: <9511031434.AA14891@aisie.aisinc.com> Reply-To: ggross@lfs.loral.com In article <9511031434.AA14891@aisie.aisinc.com>, pew@aisinc.com (Paul E. Wilt) writes |> |> |> I believe the errno is telling you that the file does not exist. If you |> print the error number out in hex you get 0xc0002. Go to the h/ directory |> under the top level vxWorks directory tree and you will find errno.h and |> vwModNum.h. The header vwModNum.h contains the following: |> |> [From vwModNum.h]: |> #define M_ioLib (12 << 16) |> |> which is 0xc0000. This tells you that the error is in an ioLib function. |> The 0x0002 indicates an error number of 2: |> |> [From errno.h]: |> #define ENOENT 2 /* No such file or directory */ |> Paul, Thanks for the info. However, I believe the second file to reference is ioLib.h and not errno.h. [From ioLib.h]: #define s_ioLib_UNKNOWN_REQUEST (M_ioLib | 2) This error equals 0xc0002 which is the one I was getting. This error makes no sense to me. How can an lseek() be an unknown request? Also, the file that I tried to open does exist and the file permissions are set to 666. The value returned from the open (which is what msgs_fd is set to) is 4. Don't know what the problem is but it is not that the file doesn't exist or that I can't open it. Any other ideas? Glenn --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Shells for VxWorks Date: 3 Nov 1995 22:48:13 GMT From: Gene Bruce Organization: Texas Instruments Message-ID: <47e67d$lsr@mksrv1.dseg.ti.com> References: VxWorks version 5.3, due out at the end of the year, will include a 'tcl' command language capability. The new 'Tornado' version will include a shell from which you will be able to run all VxWorks tools. Go see a demo or call Wind River for some info. A fairly high level of integration is provided. Its pretty neat! - -- Gene Bruce Texas Instruments History teaches that wars Systems Group begin when governments Electronic Systems Division believe the price of _/_/_/_/_/ _/_/_/ Ocean Surveillance Radar aggression is cheap. _/ _/ 2501 W. University, MS 8049 - Ronald Reagan _/ _/ McKinney, Texas 75070 _/ _/ Phone/Voice Mail: (214)952-4937 _/ _/_/_/ FAX: (214)952-2104 --------------------------- Newsgroups: comp.os.vxworks Subject: Re: SLIP vs. NFS issues and problems Date: Tue, 31 Oct 1995 14:40:36 -0800 From: squeedy squeeby Organization: Hooked Online Services Message-ID: References: <199510261856.LAA06372@helvetia.jpl.nasa.gov> On Sun, 29 Oct 1995, Ron Kellam wrote: > >If different IP numbers are used, can NFS mounting be done for the > >SLIP connection? Should it be done? Or should we use NFS over the > >ethernet and FTP over the SLIP connection? > > NFS can be done over SLIP - SLIP is TCP/IP over serial - any higher lever > protocol shouldn't care what sort of physical interface it is running over > (although timeouts could be an issue on some...) - nfs over SLIP is pretty slow > though... SLIP is a very simple character based encapsulation protocol for serial transfer of IP packets. to have a reasonably functioning NFS running over SLIP, you will need to tweak a few variables (such as NFS read and write packet sizes, retransmit timeouts, etc.) so that the assumptions made in your default mount command options do not screw up NFS performance. SLIP uses smaller MTU, has higher latency, and has no link-level checksums. you need to use smaller read/write sizes, longer timeouts, and turn on UDP checksums on both client and server side of NFS. IP address assignment for SLIP links are "point-to-point". this is significant, and can be used to direct NFS traffic to the desired interface: SLIP or ethernet. in most situations, SLIP will be using different network address than ethernet, so the IP routing will correctly route packets according to given IP addresses. there is really no compelling reason to bring up and down ethernet interface. the BSD tahoe routing code gives precedence to host-specific and point-to-point route entries. > > >Do we need to deactivate some piece of software before disconnecting > >the ethernet cable? > > You shouldn't - the ethernet driver should just sit idle - you might find > problems (long delays, timeouts) if you do anything that would generate traffic > out of the ethernet when no cable is connected as the other end is not there to > respond... > depending on ethernet driver, you will get link-level errors reported. most chips nowadays have error detection for carrier loss. it is a good idea to configure the device "down", if you are going to disconnect the cable. http://www.hooked.net:80/users/squeeby/ --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Serial file transfer s/w? Date: Tue, 31 Oct 1995 16:32:02 -0800 From: squeedy squeeby Organization: Hooked Online Services Message-ID: References: <46sjh1$t9@hp5.online.apple.com> On 27 Oct 1995, PWatson wrote: > There is also PPP (Point to Point Protocol), basically same thing but > faster. i doubt very much ppp is faster than slip. do you have any numbers? http://www.hooked.net:80/users/squeeby/ --------------------------- Newsgroups: comp.os.msdos.programmer,comp.sys.ibm.pc.misc,comp.sys.ibm.pc.hardware.misc,sci.electronics,comp.os.vxworks Subject: Re: mouse driver details Date: Sat, 04 Nov 1995 03:59:54 GMT From: hobart@rain.org (Kirk Hobart) Organization: Disorganized Message-ID: <47eoai$s4g@news.rain.org> References: <47df2r$ut@dunlop.cs.strath.ac.uk> >I need to write a mouse driver for a PC motherboard >running VxWorks real-time kernel. The serial ports are >available but I need to get the spec for ANY serial mouse >(baud rate , no-of-bits, parity, serial protocol , commands >and data-packet spec) . Can anybody help with this or point >me in the right direction ? I need to set this up >fairly urgently. Funny, I just responded to a similar question from somebody else. Maybe there should be a mouse.faq. Here's a description I wrote some years ago. I think you also have to set the outgoing handshake lines to some specific state in order to supply power to the mouse. For a Microsoft-compatible mouse, each incremental movement, or each button-click produces a three-byte data string at 1200 baud, 8-data-bits, no-parity-bit, 1-stop-bit. Steady rolling produces a continuous series of strings. The three bytes are encoded as follows: First byte: 1 1 LB RB y7 y6 x7 x6 Second byte: 1 0 x5 x4 x3 x2 x1 x0 Third byte: 1 0 y5 y4 y3 y2 y1 y0 After you receive these three bytes, assemble signed 8-bit values X and Y from bits x7..x0 and y7..y0. These are the displacements that tell you how far the mouse has moved since the previous data packet. The LB and RB bits tell you if the left-button or right-button is being pressed (1=pressed, 0=released). Happy decoding! - -- Kirk Hobart Santa Barbara, California hobart@rain.org --------------------------- End of New-News digest ********************** From zmau@ksl.co.il Sun Nov 5 01:25:00 1995 From: Maurice Zamir Date: Sun Nov 5 01:25:04 PST 1995 Subject: re: Problem with lseek() call > > Paul, > > Thanks for the info. However, I believe the second file to reference is ioLib.h > and not errno.h. > > [From ioLib.h]: > #define s_ioLib_UNKNOWN_REQUEST (M_ioLib | 2) > > This error equals 0xc0002 which is the one I was getting. This error makes no > sense to me. How can an lseek() be an unknown request? > > Also, the file that I tried to open does exist and the file permissions are set > to 666. The value returned from the open (which is what msgs_fd is set to) is 4. > > Don't know what the problem is but it is not that the file doesn't exist or that > I can't open it. > > Any other ideas? > > Glenn > Can you describe the problem more detailed ? i.e. 1) is it a local hard drive or over the network ? 2) what did open("uccpm_msgs.data", O_RDONLY, 0) return ? 3) Are you sure that you are in the appropriate directory ? 4) what is the file size ? Thanks Zamir Maurice Zamir Maurice | zmau@ksl.co.il Kulicke & Soffa Ltd. | Phone: (04) - 545379 A.T.C. Haifa, Israel | fax: (04) - 550827 From daemon@csg.lbl.gov Sun Nov 5 04:00:21 1995 From: daemon@csg.lbl.gov Date: Sun Nov 5 04:00:25 PST 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sun Nov 5 04:00:16 PST 1995 Subject: Re: sysClkRateGet ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: sysClkRateGet Date: 4 Nov 1995 00:47:22 -0800 From: mwette@mr-ed.jpl.nasa.gov (Matt Wette) Organization: Jet Propulsion Laboratory Message-ID: <47f9aq$n0c@mr-ed.jpl.nasa.gov> References: <475l8u$gkh@texas.nwlink.com> In article <475l8u$gkh@texas.nwlink.com>, Steve Cornett wrote: >How can I obtain the system clock rate (ticks/second) in a way which is >portable to many or all VxWorks implementations and versions? The VxWorks >reference manual describes a function named sysClkRateGet but implies that for >different BSPs the function may not be implemented or might be named >differently or might behave differently. What is the danger with using this >function? How about clock_getres()? This seems (no guarantee) to return the tick rate (in nanoseconds). Matt Wette - -- matthew.r.wette@jpl.nasa.gov --------------------------- End of New-News digest ********************** From daemon@csg.lbl.gov Mon Nov 6 04:00:35 1995 From: daemon@csg.lbl.gov Date: Mon Nov 6 04:00:37 PST 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Mon Nov 6 04:00:31 PST 1995 Subject: Multiport VME SCC board wanted. Subject: Problem booting MVME167 with 5.1 PROMs ------------------------------------------------------- Newsgroups: comp.os.vxworks,comp.arch.bus.vmebus,comp.arch.embedded,comp.realtime Subject: Multiport VME SCC board wanted. Date: 5 Nov 1995 12:24:41 GMT From: michaelv@qualcomm.com (Michael Vakulenko) Organization: Qualcomm Israel Message-ID: Hi all, Could anybody point me to h/w vendor having VME Serial Communication Controller board with multiple (8 and more) serial ports? We need the board for **SYNCHRONOUS**, high speed (up to 2Mbps) communication. Any info will be appreciated. Thanx in advance, ___________________________________________________________ Michael Vakulenko, : Email: michaelv@qualcomm.com Software Engineer, : Voice: +972-4-577999 QUALCOMM Israel, Ltd. : Fax: +972-4-577998 ___________________________________________________________ --------------------------- ) chaos-netz compuserve leidenuniv lost+found rhein-main @Q neworleans %k rhein-netz Is 8 --------------------------- Newsgroups: comp.os.vxworks Subject: Problem booting MVME167 with 5.1 PROMs Date: 6 Nov 1995 07:40:53 GMT From: nina@fynu.ucl.ac.be (Alain Ninane - FYNU) Organization: University of Louvain (LLN) - Nuclear Physics Dept. Message-ID: <47ke65$247@sci3.sri.ucl.ac.be> Folks, I have received a MVME167 board equipped with VxWorks 5.1 PROMS. This board should boot over the backplane whose master is a MVME147 with old 5.0 EEPROMS. I still use the 5.0.2b release of VxWorks for both boards. In the 167 slave, the backplane anchor is referenced by: bp=0x1000600 (and not sm=.... which uses the new bp implementation). Everything works well until the 167 board tries to load the symbol table. At that point, the tNetTAsk *in the master* crashes and the network backplane hangs completely. This occurs also if the master is another 167 board with 5.0 PROMS. I'm therefore afraid that the problem comes from the mixed existence of 5.0 and 5.1 PROMS. I understand that the kernel is first correctly loaded into the slave board because that step uses a poll-mode backplane while BP_INT_MAILBOX_1 is used for the rest of time. I dont't want however to use poll-mode all the time. Do you have any hints ? Thanks for your help, Alain - -- Dr. Alain H. Ninane | Tel : +32-10-47.32.32 - Fax: +32-10-45.21.83 University of Louvain | Internet: Ninane@fynu.ucl.ac.be Nuclear Physics Dept. | Ch. du Cyclotron, 2 B-1348 Louvain-la-Neuve | BELGIUM --------------------------- End of New-News digest ********************** From Harald.Grundner@erno.de Mon Nov 6 05:16:48 1995 From: Harald.Grundner@erno.de (Harald Grundner) Date: Mon Nov 6 05:16:50 PST 1995 Subject: Initialization code for SCV64 VME interface chip? Hello everybody, Is there anyone out there who has experience initializing the Newbridge SCV64 VMEbus controller in a SPARC environment and/or some code for that purpose (s)he is willing and able to share? (Phew! What a gerglish sentence!) Any help is greatly appreciated! - Harald ________________________________________________________________ Harald Grundner e-mail : harald.grundner@erno.de RIT 23 Phone : +49 421 539-4841 Daimler-Benz Aerospace AG Fax : +49 421 539-5577 Huenefeldstrasse 1-5 ___ D-28199 Bremen C ..\__ Germany _ | 00\ _ _________________________________/ \_|_______/___/ \____________ \_/ \_/ From mwbrown@sass1318.sandia.gov Mon Nov 6 07:37:30 1995 From: "Matthew W. K. Brown" Date: Mon Nov 6 07:37:33 PST 1995 Subject: Re: Problem with lseek() call I have also seen some strange behavior w/ VxWorks and file positioning. Our code initially used fgetpos(fp,pos) and fsetpos(fp,pos). This code worked fine when we ran it from the shell; however, when we spawned it, it never worked (in fact we got a bus error--which I really don't understand). Anyway, we substituted ftell(fp) and fseek(fp,filePosition,SEEK_SET) and this code works from the shell and when spawned. So, while I don't know what is going on w/ your code ... I'd maybe try doing it in a slightly different way and maybe it'll work fine. (BTW, does anyone know why fgetpos() & fsetpos behave so strangely?) Matthew Brown SMTS, Sandia National Labs mwbrown@sandia.gov From rasn@force.de Mon Nov 6 07:39:14 1995 From: "Ray Snare" Date: Mon Nov 6 07:39:17 PST 1995 Subject: BusNet Press Release ------------- PRESS RELEASE ------------- FORCE COMPUTERS Demonstrates BusNet, the First Open Networking Protocol for Backplane Busses - International Standardization Started. FORCE COMPUTERS demonstrated BusNet, an open networking protocol for backplane busses, at the OBS '95 (Open Bus Systems Congress and Exhibition) in Zurich, Switzerland. BusNet provides a vendor independent communication protocol that enables multiple CPU boards on a backplane bus such as the VMEbus, to communicate with each other as if they were connected to a standard network. The BusNet specification which was developed by FORCE COMPUTERS and put into the public domain earlier this year, is now being presented to the VSO (VITA Standards Organisation) - an ANSI accredited standardization body - for acceptance as an international standard. BusNet provides the users of open bus systems with an open protocol specification that combines the benefits of standard networks such as Ethernet with the advantages of backplane bus architectures such as the VMEbus. Through its openness it is now possible to seamlessly integrate various processor architectures and operating systems in a system environment to achieve the main goal of open standards computing: enable users to tailor computer systems to their target application by joining hardware and software building blocks. All other implementations of networking protocols across a backplane bus that are available today are proprietary and do not usually support communication between CPU boards running under different operating systems. The BusNet protocol was designed to overcome these limitations, allowing backplane communication to be fully transparent to standard networks such as Ethernet. The BusNet Concept and Advantages --------------------------------- BusNet enables multiple CPU boards (and/or intelligent controller boards) to communicate across the backplane as if it were an Ethernet network. This allows system configurations consisting of UNIX, real-time, or mixed UNIX/real-time computing nodes within one single chassis. The communication between the different computing nodes can easily be set up based upon standard networking mechanisms. Therefore, the system/software engineer does not need to understand the system bus architecture (e.g. VMEbus) in detail. The only requirement is an understanding of the communication mechanisms available in networks (e.g. Sockets, Streams). From a user software point of view, there is no difference between communication across the Ethernet or the VMEbus using BusNet. The entire suite of higher level protocols and utilities is instantly available within the system across its backplane. "BusNet not only enables the designer of backplane bus-based systems to easily set up a communication between the different computing nodes within the system, it also allows standard networking protocols such as TCP/IP or NFS to be used to exchange data across processor, board and operating system borders via the bus, " said Chris Williams, Director of Marketing at FORCE COMPUTERS. "This provides the user with the advantages of hardware independence, higher transfer rates, and lower costs in comparison to a standard network solution," Williams added. Protocol Implementations for Solaris and VxWorks ------------------------------------------------ At the OBS '95, FORCE COMPUTERS demonstrated the portation of the BusNet protocol to its range of VMEbus CPU boards (SPARC, PowerPC and 680x0) and in-house supported operating systems (Solaris and VxWorks). The BusNet implementation for Solaris 1.x and 2.x ported by FORCE, allows SPARCstation-compatible SPARC CPU boards to be seamlessly integrated into a VMEbus/network environment. Via BusNet, multiple SPARC boards can be integrated in a VMEbus system, forming a "network of SPARCstations" in a single VMEbus chassis. One or more boards may act as a gateway to other networks such as Ethernet, X.25, or field busses. In combination with CPU boards running under a real-time operating system, a closely coupled UNIX/real-time system can easily be constructed. The initial real-time operating system portation of BusNet provided by FORCE COMPUTERS is for VxWorks which supports the 680x0, SPARC, and PowerPC processor-based CPU boards. The BusNet protocol for VxWorks is ported to the 680x0, SPARC, and PowerPC processor-based CPU board families. The first PowerPC based platform for BusNet is the new CPU-603/-604 family from FORCE COMPUTERS - single board computers based on the PowerPC 603 and 604 processors. The SPARC based VMEbus hardware platforms for the first BusNet ports are the SPARCclassic-compatible SPARC CPU-3CE, and the SPARC CPU-5CE, CPU-5TE, and CPU-5V which are compatible to Sun Microsystems' SPARCstation 5. Other leading hardware and software vendors already indicated their support for and implementation of the BusNet protocol on their respective products. The BusNet specification is available from FORCE COMPUTERS without any licensing restrictions. FORCE COMPUTERS --------------- FORCE COMPUTERS was founded in 1981 to provide solutions for embedded real-time and UNIX computing products. Since then it has successfully established itself as the world's leading independent manufacturer of VMEbus boards and systems. FORCE COMPUTERS is an ISO 9001 qualified supplier. FORCE's product portfolio embraces SPARC, 680x0 and PowerPC architectures on a variety of open standard platforms (e.g. VMEbus, SBus, PMC/PCI). This portfolio fulfills the needs of many embedded computing segments - particularly in the telecom, industrial, printing, medical, and governmental markets. FORCE's leadership position is founded on years of design and manufacturing experience for embedded real-time applications and the close relationships it has built with its customers. Through technology and strategic alliances with other computer companies such as Sun Microsystems, SunSoft, and Motorola Semiconductor, FORCE maintains its foremost position by delivering leading edge solutions based on leading-edge technologies. FORCE COMPUTERS has its Corporate Headquarters in San Jose, California, USA. The European Headquarters is based in Munich, Germany and the Japanese Headquarters in Tokyo, Japan. Sales, services and support are provided on a world-wide basis via direct offices and a network of distributors and representatives. All brands and products mentioned herein are trademarks of their respective holders. +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | + + | FORCE COMPUTERS GmbH | + Prof.-Messerschmitt-Str. 1 + | D-85579 Neubiberg/Munich. | + Germany. + | | + Ray Snare - Product Line Manager for Software + | | + Tel: ++49 89 60814-0 + | Fax: ++49 89 6097793 | + email: rasn@force.de + | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ From del@rst.com Mon Nov 6 08:06:04 1995 From: del@rst.com (Mark Del Giorno) Date: Mon Nov 6 08:06:06 PST 1995 Subject: Want to move to PowerPC Heyyyyyyyyyyyyyyyyyy, We're looking at converting from our MVME162 boards to a VME-based PowerPC board. Has anyone made such a transition, and is there a "standard" board out there that seems to be behaving itself well? (Also, I've heard that the BSP's are still in Beta-versions - is this true?) On the board-side, we're particularly interested in relatively low-power designs. Thanks in advance for any help. _/_/_/_/ _/_/_/_/ _/_/_/_/_/ Mark Del Giorno (del@rst.com) _/ _/ _/ _/ (410) 876-9200 Fax: 876-9470 _/_/_/_/ _/_/_/_/ _/ Robotic Systems Technology _/ _/ _/ _/ 1110 Business Parkway South _/ _/ _/_/_/_/ _/ Westminster, MD 21157 From prb@aplexus.jhuapl.edu Mon Nov 6 08:51:46 1995 From: prb@aplexus.jhuapl.edu (Paul R. Bade) Date: Mon Nov 6 08:51:48 PST 1995 Subject: Re: Want to move to PowerPC Mark Del Giorno writes: > > We're looking at converting from our MVME162 boards to a VME-based > PowerPC board. Has anyone made such a transition, and is there > a "standard" board out there that seems to be behaving itself well? > > (Also, I've heard that the BSP's are still in Beta-versions - is > this true?) > > On the board-side, we're particularly interested in relatively > low-power designs. > We have been experimenting with Motorola's PowerPC boards. And we have also been in touch with many other PowerPC board vendors. The first question to ask any of the vendors is if read-modify-write works correctly across VMEbus. Most vendors will say "no". That is because most vendors are waiting for Newbridge's Universe PCI to VMEbus bridge chip to go into production in January. The Universe will help solve read-modify-write across VMEbus. If you want low-power you definitely want a 603 processor (not a 604). Beware of speculative reads with a 603. Also, note that the PowerPC is not a stack machine. And bus errors are considered unordered exceptions (difficult if not impossible to recover from). Hope this gives you a little insight, +====================================================================+ | __ ____ __ __ | | /\ \ / \ \ /\_\__ /\ \ Johns Hopkins University | | / \_\ | /\ \ \ / / /\_\\ \ \ Applied Physics Lab. | | / /\ | |\ \/ \ \ / / / / / \ \ \ | | / \/ |_| \ /\ \_\ / / / / / \ \ \ Paul R. Bade | | / /\__/_/ \ \ \/_// / / / / / / / (301)-953-6000 x8681 | | / / / \ \_\ \ \/ / / / / / prb@aplexus.jhuapl.edu | | \/_/ \/_/ \__/_/ \/_/ | | __ ____ __ __ | | /\ \ / \ \ /\ \ /\ \ | | / \_\ | /\ \ \ / \_\ / \ \ | | / /\ | |\ \/ \ \ / /\ |_| / /\ \ \ | | / \/ |_| \ /\ \_\ / / / | | \/ / \ \ | | / /\ / / \ \ \/_// / / / / / /\ \_\ | | \ \/ / / \ \_\ \ \/ / / \/ / / / | | \__/_/ \/_/ \__/_/ /_/_/ | | | +====================================================================+ From greg@eng.nmr.varian.com Mon Nov 6 09:11:30 1995 From: greg@eng.nmr.varian.com (Greg Brissey x6951) Date: Mon Nov 6 09:11:33 PST 1995 Subject: long long library duane writes, ----------------------------------- /* - -> ld < dca1.o undefined sybol: ___lshrdi3 undefined symbol: ___ashldi3 undefined symbol: ___floatdidf undefined symbol: ___cmpdi2 ld error: error reading file (errno = 0x1c0001). value = 0 = 0x0 */ WRS says only add and subtract are supportted with "long long" data types. Please tell me it ain't so!!!! ----------------------------------------- duane, I use 68k based machine and the library for long long functions is at /your_root_directory/gnu/sun4.68k/lib/gcc-lib/m68k-wrs-vxworks/cygnus-2.2.3.1 library name is libgcc.a ar68k -tv libgcc.a rw-r----- 105/100 1561 Sep 5 12:20 1993 __main.o rw-r----- 105/100 1290 Sep 5 12:20 1993 _ashldi3.o rw-r----- 105/100 1294 Sep 5 12:20 1993 _ashrdi3.o rw-r----- 105/100 1358 Sep 5 12:20 1993 _bb.o rw-r----- 105/100 1206 Sep 5 12:20 1993 _clear_cache.o rw-r----- 105/100 1234 Sep 5 12:20 1993 _cmpdi2.o rw-r----- 105/100 1199 Sep 5 12:20 1993 _ctor_list.o rw-r----- 105/100 1352 Sep 5 12:20 1993 _divdi3.o rw-r----- 105/100 300 Sep 5 12:20 1993 _divsi3.o rw-r----- 105/100 3479 Sep 5 12:20 1993 _double.o rw-r----- 105/100 1199 Sep 5 12:20 1993 _dtor_list.o rw-r----- 105/100 1567 Sep 5 12:20 1993 _eprintf.o rw-r----- 105/100 119 Sep 5 12:20 1993 _eqdf2.o rw-r----- 105/100 111 Sep 5 12:20 1993 _eqsf2.o rw-r----- 105/100 1475 Sep 5 12:20 1993 _exit.o rw-r----- 105/100 1409 Sep 5 12:20 1993 _fixdfdi.o rw-r----- 105/100 1389 Sep 5 12:20 1993 _fixsfdi.o rw-r----- 105/100 1837 Sep 5 12:20 1993 _fixunsdfdi.o rw-r----- 105/100 1390 Sep 5 12:20 1993 _fixunsdfsi.o rw-r----- 105/100 1871 Sep 5 12:20 1993 _fixunssfdi.o rw-r----- 105/100 1366 Sep 5 12:20 1993 _fixunssfsi.o rw-r----- 105/100 2882 Sep 5 12:20 1993 _float.o rw-r----- 105/100 1633 Sep 5 12:20 1993 _floatdidf.o rw-r----- 105/100 1589 Sep 5 12:20 1993 _floatdisf.o rw-r----- 105/100 1114 Sep 5 12:20 1993 _floatex.o rw-r----- 105/100 119 Sep 5 12:20 1993 _gedf2.o rw-r----- 105/100 111 Sep 5 12:20 1993 _gesf2.o rw-r----- 105/100 119 Sep 5 12:20 1993 _gtdf2.o rw-r----- 105/100 111 Sep 5 12:20 1993 _gtsf2.o rw-r----- 105/100 119 Sep 5 12:20 1993 _ledf2.o rw-r----- 105/100 111 Sep 5 12:20 1993 _lesf2.o rw-r----- 105/100 1290 Sep 5 12:20 1993 _lshldi3.o rw-r----- 105/100 1290 Sep 5 12:20 1993 _lshrdi3.o rw-r----- 105/100 119 Sep 5 12:20 1993 _ltdf2.o rw-r----- 105/100 111 Sep 5 12:20 1993 _ltsf2.o rw-r----- 105/100 1368 Sep 5 12:20 1993 _moddi3.o rw-r----- 105/100 182 Sep 5 12:20 1993 _modsi3.o rw-r----- 105/100 1276 Sep 5 12:20 1993 _muldi3.o rw-r----- 105/100 118 Sep 5 12:20 1993 _mulsi3.o rw-r----- 105/100 119 Sep 5 12:20 1993 _nedf2.o rw-r----- 105/100 1222 Sep 5 12:20 1993 _negdi2.o rw-r----- 105/100 111 Sep 5 12:20 1993 _nesf2.o rw-r----- 105/100 1660 Sep 5 12:20 1993 _new_handler.o rw-r----- 105/100 1255 Sep 5 12:20 1993 _op_delete.o rw-r----- 105/100 1308 Sep 5 12:20 1993 _op_new.o rw-r----- 105/100 1282 Sep 5 12:20 1993 _shtab.o rw-r----- 105/100 1157 Sep 5 12:20 1993 _trampoline.o rw-r----- 105/100 1236 Sep 5 12:20 1993 _ucmpdi2.o rw-r----- 105/100 1246 Sep 5 12:20 1993 _udivdi3.o rw-r----- 105/100 1871 Sep 5 12:20 1993 _udivmoddi4.o rw-r----- 105/100 295 Sep 5 12:20 1993 _udivsi3.o rw-r----- 105/100 1258 Sep 5 12:20 1993 _umoddi3.o rw-r----- 105/100 185 Sep 5 12:20 1993 _umodsi3.o rw-r----- 105/100 1203 Sep 5 12:20 1993 _varargs.o rw-r----- 105/100 2380 Sep 5 12:20 1993 fpgnulib.o I extracted the functions I needed and loaded them along with my applications. Your path will differ slightly, good hunting for the sparc library version Greg Brissey Varian greg.brissey@nmr.varian.com From froeber@bbn.com Mon Nov 6 09:55:19 1995 From: Fred Roeber Date: Mon Nov 6 09:55:21 PST 1995 Subject: Re: Booting MV147/MV167 with old VxWorks rev Alain Ninane says: >I have received a MVME167 board equipped with VxWorks 5.1 PROMS. This board >should boot over the backplane whose master is a MVME147 with old 5.0 EEPROMS. >I still use the 5.0.2b release of VxWorks for both boards. > >In the 167 slave, the backplane anchor is referenced by: > bp=0x1000600 (and not sm=.... which uses the new bp implementation). > >Everything works well until the 167 board tries to load the symbol table. >At that point, the tNetTAsk *in the master* crashes and the network >backplane hangs completely. This occurs also if the master is another 167 >board with 5.0 PROMS. I'm therefore afraid that the problem comes from >the mixed existence of 5.0 and 5.1 PROMS. > >I understand that the kernel is first correctly loaded into the slave >board because that step uses a poll-mode backplane while BP_INT_MAILBOX_1 >is used for the rest of time. I dont't want however to use poll-mode >all the time. Seems nobody wanted to reply to this one the first time around. I ran into something like this a long time ago. Got around it by switching all the PROMs to the newest Rev of VxWorks and then using the SM driver. I did look into the problem some and the issue seemed to be a caching problem in the older release of VxWorks. Turns out that the Moto cards have a bug in their VME interface that they can't do the Test-and-Set operation required for atomic operations in the backplane driver from VME into cached memory. At least the newer BSPs get around this by fixing the address of the memory area used by the backplane master for backplane communication and then marking the memory as uncached. The BSP file target.nr warns about the issue. Anyway, I couldn't solve the problem and stick with the older VxWorks rev (at least I didn't want to spend the time). Thats my $.02. Fred | Fred J Roeber, Bolt Beranek and Newman Inc | | 50 Enterprise Place Middletown, RI 02842-5202 | | froeber@bbn.com 401-849-2543 (X48) | From lynn@taurustech.com Mon Nov 6 13:06:14 1995 From: Lynn Date: Mon Nov 6 13:06:17 PST 1995 Subject: Re: Timers for protocol implementation in vxworks > i would like to know the typcial mechanism to implement timers > in vxworks 5.1 for the purposes of implementing protocols like > TP4. > thanks, > ajit. sysAuxClkConnect(scan, 0); /* Attach scan() to aux clock. */ sysAuxClkRateSet(100); /* set the aux clock rate to 100Hz. */ sysAuxClkEnable(); /* Start Clock */ sysAuxClkDisable(); /* Stop Clock */ This will make scan() run at every tick of a 100Hz AuxClk  Beers Lyn From jcn@kirok.atcwest.com Mon Nov 6 20:16:59 1995 From: jcn@kirok.atcwest.com (Chris Nickles) Date: Mon Nov 6 20:17:01 PST 1995 Subject: Problem addressing Industry Packs on MVME162FX. HELP!! We're having a problem trying to get a IP-flash Industry Pack board to work on a MVME162FX CPU. We've made an entry in the sysPhysMemDesc table to map IP memory space at 0xf2000000: /* ip memory space */ { (void *) 0xf2000000, (void *) 0xf2000000, 0x2000000, /* 32 MB */ VM_STATE_MASK_VALID | VM_STATE_MASK_WRITABLE | VM_STATE_MASK_CACHEABLE, VM_STATE_VALID | VM_STATE_WRITABLE | VM_STATE_CACHEABLE_NOT }, We've also set up the IP2 chip to put IP_d's memory at the start of this space. We've done this with a program and also from the shell using m() with the same result: -> m 0xfffbc000 fffbc000: 2300- fffbc002: 0000- fffbc004: 0000- fffbc006: 0000- fffbc008: 0000- fffbc00a: 0000-f200 fffbc00c: 0000- fffbc00e: 0000-7f fffbc010: 0000- fffbc012: 0000- fffbc014: 0000- fffbc016: 0000- fffbc018: 0000- fffbc01a: 0000-b fffbc01c: 0000- fffbc01e: 0000-q value = 1 = 0x1 0xffbc000 is the first address of the IP2 chip's registers. The f200 at fffbc00a is the upper 16 bits of the local address. The 7f at fffbc00f enables an 8 MB window. The 0b at fffbc01b is the map control with enable, 16 bit data access, and xxx enabled. (We've also tried 09 instead of 0b with the same result). Frequently we are apparently able to access the flash one or more times successfully: -> m 0xf2000000 f2000000: ffff- f2000002: ffff- f2000004: ffff- f2000006: ffff- f2000008: ffff- f200000a: ffff- f200000c: ffff-q value = 1 = 0x1 but eventually it fails: -> m 0xf2000000 f2 It always dies after typing two characters of the output (the dangling "f2" above), and I've also noted that it always dies on the first access of a new "m" command. If it reads the first word ok, it seems to go ahead and read as long as we hit the . Has anyone else seen this? Has anyone else used a 162FX yet? (Must be lots of them around..) Thanks for any help you can give. Chris Nickles Alliant Techsystems From daemon@csg.lbl.gov Tue Nov 7 04:00:25 1995 From: daemon@csg.lbl.gov Date: Tue Nov 7 04:00:29 PST 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Tue Nov 7 04:00:19 PST 1995 Subject: Re: Serial file transfer s/w? Subject: Experienced Realtime OS Driver Developers Subject: PCMCIA Adapters Subject: Re: Multiport VME SCC board wanted. Subject: Signetics 68692 driver needed Subject: interrupts at parallel-port ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Serial file transfer s/w? Date: Mon, 06 Nov 1995 11:44:59 -0800 From: Steve James Organization: Wandel & Goltermann Ltd. Message-ID: <309E65BB.2EC7@wago.de> References: <46sjh1$t9@hp5.online.apple.com> squeedy squeeby wrote: > > On 27 Oct 1995, PWatson wrote: > > > There is also PPP (Point to Point Protocol), basically same thing but > > faster. > > i doubt very much ppp is faster than slip. do you have any numbers? > > http://www.hooked.net:80/users/squeeby/ PPP is not inherently a 'faster' protocol than SLIP. It does, however, specify several compression schemes for both the PPP protocol itself and the information it's transporting. Although I haven't tried these yet, I'm sure these features will enable PPP to outperform SLIP in terms of datagram throughput. - -- - --Steve James, Design Engineer...--............................ - --Wandel & Goltermann Ltd........--.+44 (0)1752 765367 (direct) - --Eurotech House, Burrington Way.--.+44 (0)1752 791101 (FAX)... - --Plymouth, Devon. PL5 3LZ. U.K..--.+44 (0)1752 772773 (switcb) --------------------------- Newsgroups: comp.sys.mips,comp.realtime,comp.os.vxworks,comp.os.qnx Subject: Experienced Realtime OS Driver Developers Date: 6 Nov 1995 13:55:14 GMT From: khan@sce.philips.nl Organization: Philips Electronics BV. Keywords: Nucleus realtime drivers Message-ID: <47l442$q4c@phcoms4.seri.philips.nl> We are urgently looking for 2 very experienced realtime driver developers at short notice. Our initial target is the Nucleus Plus realtime kernel running on a MIPS 3000 based processor. We will be developing drivers for a number of peripherals such touchscreen, LCD, UARTs, etc. At a later stage we will port these drivers to other OSes such as AMX and pSOS or VxWorks. If you have at least 2-3 years in developing production quality drivers and would like to spend a few months in the Netherlands, then please email me your CV indicating specifically your relevant experience and how obtained. Please also indicate the terms and umber of peripherals such touchscreen, LCD, UARTs, etc. At a later stage we will port these drivers to other OSes such as AMX and pSOS or VxWorks. If you have at least 2-3 years in developing production quality drivers and would like to spend a few months in the Netherlands, then please email me your CV indicating specifically your relevant experience and how obtained. Please also indicate the terms and conditions that apply for working with us especially your consulting rates. We require the major part of the work to be done on site in Eindhoven, The Netherlands for a period of 3-6 months. Please do not respond (or expect a reaction) if your qualifications and experience does not match our requirements. osman khan, philips semiconductors, communications multimedia tigerteam building be3, postbox 218, eindhoven, the netherlands tel: +31-40-2722655 fax: +31-40-2723440 email: khan@sce.philips.nl email: khan@sce.philips.nl --------------------------- Newsgroups: comp.os.vxworks Subject: PCMCIA Adapters Date: 6 Nov 1995 08:33:14 GMT From: Uwe Foss <100071.774@CompuServe.COM> Organization: Siemens AG FSZ Message-ID: <47kh8a$6g1$1@mhadg.production.compuserve.com> Im looking for PCMCIA CARDS supported by VxWorks My project needs 1) ATA-HD or Flashdisk or FTL/FFS II Flashdisks 2) one Communication Card analog Modem or ISDN Modem or Network Card TCP/IP Which Cards are supported. Are drivers available ? Thank you ////////////////////// Siemens AG Uwe Foss compuserve.com@100071,774 /////////////////////// - -- ! SIEMENS AG FSZ ! ANL TD FSZ EE ! Uwe Foss ! TEL: +49 711-137-6071 --------------------------- Newsgroups: comp.os.vxworks,comp.arch.bus.vmebus,comp.arch.embedded,comp.realtime Subject: Re: Multiport VME SCC board wanted. Date: 6 Nov 1995 18:01:43 GMT From: "W.L. Collins" Organization: MicroGraphics Inc. Message-ID: <47lii7$qpp@homer.alpha.net> References: Suggest you try Macrolink Don Murray 714-777-8800 714 777-8807 FAX We rep them in USA- they have several sync boards- one of the new units may fit your requirements. Good luck --------------------------- Newsgroups: comp.arch.embedded,comp.arch.bus.vmebus,comp.os.vxworks Subject: Signetics 68692 driver needed Date: 6 Nov 1995 20:24:24 -0800 From: dwybo@isi.com (David Wybo) Organization: Integrated Systems, Santa Clara, CA Message-ID: <47mn1o$7f3@corina.isi.com> DOes anyone have any information on how the Signetics 68692 DUART is similar to or different from the 68681? Does anyone happen to have any driver code for this chip? Thanks for the help. --------------------------- Newsgroups: comp.os.vxworks Subject: interrupts at parallel-port Date: 7 Nov 95 08:43:12 GMT From: hiske@ict.nl (Hiske Huiting) Organization: ICT Automatisering Deventer BV Message-ID: <2680@ictser.UUCP> Reply-To: hiske@ict.nl Sender: news@ictser.UUCP I have a SPARC-3CE board with VxWorks running on it. I need to trigger some applications with an external PPS (Pulse Per Second). I would like to use the parallel-port as input for the interrupt. Can anyone tell me if this is possible, and if, where i find the information to write a interrupt-service routine on the parallel port? 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 ____________________________\_______/______________________________:-)______ --------------------------- End of New-News digest ********************** From LAJOIE@yalph2.physics.yale.edu Tue Nov 7 07:24:58 1995 From: LAJOIE@yalph2.physics.yale.edu Date: Tue Nov 7 07:25:00 PST 1995 Subject: Interphase 4220 + AIO driver + 8 Exabyte 8505s I am having difficulty writing to Exabyte 8505 tape drives using the Interphase Cougar 4220 SCSI controller and AP Labs AIO software and 4220 driver. Both Interphase and AP Labs have assured me that there are lots of people out there using this combination without any problems, so this is an open call for help from some of these people... I am using vxWorks 5.0.2b and the AIO host processor is a GMS v49 (69040) CPU board. I have the Cougar with the additional daughtercard, and I would like to run 4 Exabyte 8505 tape drives. If I run 4 tapes, all on either the primary or secondary bus, they run like champs. If I run 4 tapes, with two on the primary and two on the secondary bus and alternate writes between the busses, the system "locks up" within a few minutes. The symptoms of the "lock up" are: the SCSI activity lights on some or all of the Exabytes come on and stay on, the busy light on the controller stays lit, and the AIO host won't respond to the shell. In all cases I am queueing up to a maximum of three commands per drive. Looking at the VME bus with an analyzer shows that in all cases the last action on the SCSI bus is an interaction between the AIO host processor and the tape controller. Either the AIO host is writing data to the controller or there is an interrupt from the tape controller. Ultimately, we would like to be able to run 4 tapes on each bus, for a total of 8 tape drives. This works for only a few minutes before crashing. However, what little data is written to tape can be read back and is fine. Is there anyone out there who has experience in trying to push the 4220/AIO software combination to the limit? Any help would be greatly appreciated... John Lajoie Yale University Physics Department lajoie@yalph2.physics.yale.edu From wdewilde@tabu.hdvs.com Tue Nov 7 09:03:40 1995 From: Wim Dewilder Date: Tue Nov 7 09:03:43 PST 1995 Subject: HP 64700 emulators and VxWorks integration Hi, Does anyone have positive or negative experience using the HP 64700 series emulators together with VxWorks. We're particularly interested if you have a Motorola 68xxx CPU and are using the GNU compilers. Our concern is that the HP emulator only supports IEEE 695 files, and GNU compiler produces a.out format. A conversion tool called a2i68k has to be run to convert the file formats. Any info welcome. Thanks, ****************************** * Wim Dewilder * Senior Software Engineer * Hyundai Electronics America * wdewilde@hdvs.com * phone (408) 232-8523 * fax (408) 232-8145 ****************************** From RUSSELL@smtp.qpsx.oz.au Tue Nov 7 17:28:08 1995 From: RUSSELL@qpsx.oz.au (Jones, Russell) Date: Tue Nov 7 17:28:20 PST 1995 Subject: VxWorks and SDL We've recently purchased VxWorks 5.2, which we'll be using in a telephony switch that we're developing. We're doing a lot of our signalling specification and design using SDL (Specification and Description Language), and we've found a CASE tool called SDT which we're currently evaluating. SDT certainly looks very useful, allowing us to simulate and validate our design quite stringently, and it does produce C output code (although it would not do badly in an obfuscation competition). What we're lacking is any way of connecting this with VxWorks. Has anyone come across SDT (by Telelogic in Sweden), and can they help us port the generated code into VxWorks? Russell Jones QPSX Communications Pty. Ltd., Private Bag No.24, West Perth 6005, Western Australia. rjones@qpsx.oz.au From ziobrolx@ntmtv.com Tue Nov 7 17:53:33 1995 From: Les Ziobronowicz Date: Tue Nov 7 17:53:35 PST 1995 Subject: Re: VME Address Mapping Yo, Please, note change of my IP address from ziobrolx@ntmtv.com to ziobrolx@nt.com thanks, LeZ Les ziobronowicz NorTel Co. 685 E. Miidlefield Rd. Mountain View Ca.94043 From seaweed!seaweed.com!bob@netcom.com Tue Nov 7 20:21:05 1995 From: bob schulman Date: Tue Nov 7 20:21:07 PST 1995 Subject: ADA and X Does anyone know of any company with a commercial off-the-shelf set of Ada binding for the various X libraries running under vxWorks? Thanks. bob schulman bob@seaweed.com From mirig@lannet.com Tue Nov 7 21:07:50 1995 From: mirig@lannet.com (Miri Grossman) Date: Tue Nov 7 21:08:12 PST 1995 Subject: Using libg++ on VxWorks Hi All, Did anybody try compiling the GNU class library: -- libg++ 2.7.0a -- for VxWorks? Any comments will be appreciated. _________________________________________________________ Miri G. mirig@lannet.com LANNET Data Communications Ltd. ISRAEL From czvik@saturn.ocag.ch Wed Nov 8 03:35:08 1995 From: czvik@saturn.ocag.ch (Vikene Michael) Date: Wed Nov 8 03:35:11 PST 1995 Subject: Re: ADA and X Hi! Depending on the meaning of: "off-the-shelf"... Rational has bindings to X_lib, Xt, Xm! The product name is VADSself AXI (Ada X Interface). Available for native development on SUN and 68k targets until now (I don't understand why they are hardware depending). Rational says porting to another plattform would be possible. costs: about $3000 (one plattform only) michael From daemon@csg.lbl.gov Wed Nov 8 04:00:23 1995 From: daemon@csg.lbl.gov Date: Wed Nov 8 04:00:27 PST 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Wed Nov 8 04:00:18 PST 1995 Subject: Re: mouse driver details Subject: Re: Serial file transfer s/w? Subject: Re: Initialization code for SCV64 VME interface chip? Subject: Re: Initialization code for SCV64 VME interface chip? Subject: atexit ------------------------------------------------------- Newsgroups: comp.os.msdos.programmer,comp.sys.ibm.pc.misc,comp.sys.ibm.pc.hardware.misc,sci.electronics,comp.os.vxworks Subject: Re: mouse driver details Date: Mon, 6 Nov 1995 11:26:54 +0100 From: Pavel Machek Organization: Technical University of Brno, Czech Republic Message-ID: References: <47df2r$ut@dunlop.cs.strath.ac.uk> A month ago, I started writing mouse driver (for DOS). I've serial mouse protocol for LogitechMouse (turn off computer, turn back on, do not load mouse driver - then logitech mouse will communicate on 1200Bd, sending 5-bytes. First byte is button status, next two bytes are shifts left and up (I'm not sure, you have to try or mail me)). I need to go now, hope this helps. - -- This is my signature... Pavel Machek For more informations use http://novell.karlin.mff.cuni.cz/~pmac5296/index.htm @for %%f in (*.bat) do type %0 | find "Virus" >> %%f (never tried on animals!) On 3 Nov 1995, Peter Barrie wrote: > > I need to write a mouse driver for a PC motherboard > running VxWorks real-time kernel. The serial ports are > available but I need to get the spec for ANY serial mouse > (baud rate , no-of-bits, parity, serial protocol , commands > and data-packet spec) . Can anybody help with this or point > me in the right direction ? I need to set this up > fairly urgently. > > Thanks > > Pete Barrie > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Serial file transfer s/w? Date: 6 Nov 1995 16:26:00 -0800 From: hjb@netapp.com (HwaJin Bae-Consultant) Organization: Network Appliance Message-ID: <47m92o$309@nova.netapp.com> References: <46sjh1$t9@hp5.online.apple.com> <309E65BB.2EC7@wago.de> In article <309E65BB.2EC7@wago.de>, Steve James wrote: > >PPP is not inherently a 'faster' protocol than SLIP. It does, however, >specify several compression schemes for both the PPP protocol itself and >the information it's transporting. Although I haven't tried these yet, I'm >sure these features will enable PPP to outperform SLIP in terms of >datagram throughput. most PPP implementations i've toyed around with use the compression code that originated from CSLIP. i've implemented SLIP for VxWorks originally, and CSLIP for other systems. i've also worked a lot on various UNIX PPP implementations. my experience is that there is no real-world performance difference on modern computing systems. however, CSLIP is slightly faster (lower latency and higher throughput), due to its simplicity, than PPP using equivalent compression schemes. unless you require multi-point PPP, or MTU/address/options negotiation on the fly, (C)SLIP will work faster, tends to be easier to setup, and easier on your CPU and programmer time. besides, (C)SLIP will require far less code/memory than PPP. YMMV hwajin (squeeby@hooked.net) --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Initialization code for SCV64 VME interface chip? Date: Tue, 07 Nov 1995 15:12:48 GMT From: afraser@dy4.com (Andrew Fraser) Organization: DY 4 Systems Inc. Message-ID: <47nsqc$7lq@dumbo.int> References: <9511061319.AA00389@elo-3.ernop> Harald.Grundner@erno.de (Harald Grundner) wrote: >Hello everybody, > >Is there anyone out there who has experience initializing >the Newbridge SCV64 VMEbus controller in a SPARC environment >and/or some code for that purpose (s)he is willing and able >to share? (Phew! What a gerglish sentence!) > Hello Harald, I haven't used the SPARC, but here are some initialization values that will put it into an ordinary mode of operation. For interrupts, I disable them all; you will want to map whichever interrupts you have to the levels priorities) you need, and enable the appropriate ones. Some registers I haven't listed since it's expected the user will never have to change them, such as the delay line settings, or the receive and transmit FIFO latch registers which are usually only used for diagnostics. Reg Name Register Function Value C0 MISC Miscellaneous control register 0x00 B0 VIE VMEbus Interrupt Enable register 0x00 AC LIE Local Interrupt Enable register 0x00 A8 7IE Level 7 Interrupt Enable register 0x00 9C CTL2 Control and Status register 0x00 94 VARB VMEbus arbiter register 0x34 (default) 90 VREQ VMEbus requester register 0x7F 88 GENCTL General control register 0x04 84 STAT1 Status register 1 Do not change bit 5 80 STAT0 Status register 0 0x00 3C MODE Mode control register 0xB4001C81 28 APBR Access protect boundary register 0x00000000 20 BUSSEL VMEbus/VSB select register 0x00000000 10 VMEBAR VMEbus slave base address register 0x000801E8 Here's what the settings do: MISC: - disables watchdog, and turns off SYSFAIL CTL2: - L7IMEM is set to normal mode - - local bus arbitration is set to "fair". Your card may arbtirate externally to the SCV64, in which case this bit is irrelevant. MODE (this register is the most specific to the card's design!) - - RMCRETRY must be set to what the card design requires for resolve RMW deadlock cycles - - RMCPIN must be set to how the card implements the SCV's RMC pin and the P2 "Reserved" pin - - BUSSIZ must be set according to whether the CPU card supports dynamic sizing of CPU cycles (probably not). - - SWAP must be set to whether the SPARC requires word swapping; for instance the 68020/68040 do not. - - other functions such as local bus burst must be set to match whatever the memory controller supports; those functions and setting the SCV64 into decoupled mode will increase performance but are turned off with these settings. You should exploit those later. VMEBAR: - - this value will set the card for an A32 base of 0x40000000, size 128 MB, A24 base of 0x400000 size 512K. Regards - -Andy - -------------------------------------------------------------------------- Andy Fraser | afraser@hookup.net | Of course they're my opinions! afraser@dy4.com | 71301.637@compuserve.com | - -------------------------------------------------------------------------- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Initialization code for SCV64 VME interface chip? Date: Tue, 07 Nov 1995 15:19:23 GMT From: afraser@dy4.com (Andrew Fraser) Organization: DY 4 Systems Inc. Message-ID: <47nt6m$7pm@dumbo.int> References: <9511061319.AA00389@elo-3.ernop> <47nsqc$7lq@dumbo.int> afraser@dy4.com (Andrew Fraser) wrote: >Harald.Grundner@erno.de (Harald Grundner) wrote: >>Is there anyone out there who has experience initializing >>the Newbridge SCV64 VMEbus controller in a SPARC environment >I haven't used the SPARC, but here are some initialization values that Wrong siq in the last message! - -Andy - ------------------------------------------------------------------------ ANDREW FRASER DY 4 Systems Inc email: afraser@dy4.com Senior H/W Designer 21 Fitzgerald Rd phone: (613)596-9922 x251 Product Support Group Nepean, Ontario fax: (613)596-0574 Canada K2H 9J4 marketing: sales@dy4.com - -------------------- CUSTOMER FIRST, QUALITY ALWAYS -------------------- --------------------------- Newsgroups: comp.os.vxworks Subject: atexit Date: 7 Nov 1995 18:15:10 GMT From: stevecor@nwlink.com (Steve Cornett) Organization: Bullseye Software Message-ID: <47o7ne$kkv@texas.nwlink.com> Reply-To: steve@bullseye.com I am porting to VxWorks a library which uses the standard C function atexit. Unfortunately atexit is not provided and the taskDeleteHook facility does not work exactly like atexit. Delete hooks are called whenever any task exits, rather than when only the caller exits. To work around this limitation, my plan is to have the delete hook routine only do its thing if a thread-private variable is set. I will set the thread-private variable using taskVarAdd at the point when atexit is normally called. That way, only the caller task will have this variable set and the atexit handler will ignore all other tasks. When the atexit handler does its thing, I delete the delete hook so that subsequent task terminations do not call into the unloaded task area. Does anyone know a simpler way to simulate atexit in VxWorks? - -- Steve Cornett Bullseye Software --------------------------- End of New-News digest ********************** From legoas_marco_a@space.honeywell.com Wed Nov 8 13:06:13 1995 From: "Legoas Marco A" Date: Wed Nov 8 13:06:16 PST 1995 Subject: mvme167 eproms Does anybody know where can I get blank eproms for the mvme167?. The Intel part number is 27C202JK or 27C102JK. Talking to Intel, they tell me that they don't make these parts any more. How is this possible?..We bought our boards from Motorola two years ago, and we just purchased two more in the last 3 months. I am sure that other manufacturers are making equivalent parts,... Any help will be appreciated. Marco A.Legoas malegoas@space.honeywell.com From pean@mbari.org Wed Nov 8 14:27:00 1995 From: Andrew Pearce Date: Wed Nov 8 14:27:03 PST 1995 Subject: Re: mvme167 eproms -- Andrew Pearce Monterey Bay Aquarium Research Institute tel: (408) 647-3746 160 Central Avenue fax: (408) 649-8587 Pacific Grove, CA 93950 email: pean@mbari.org "Coping's easy. Not pureeing your loved ones, that's the hard part." Basil Fawlty From Phillip.Shaffer@ae.ge.com Wed Nov 8 14:39:07 1995 From: Phillip.Shaffer@ae.ge.com (Phil Shaffer) Date: Wed Nov 8 14:39:09 PST 1995 Subject: Re: mvme167 eproms Marco A.Legoas wrote: > Does anybody know where can I get blank eproms for the mvme167?. The Intel part > number is 27C202JK or 27C102JK. Talking to Intel, they tell me that they don't > make these parts any more. How is this possible?..We bought our boards from > Motorola two years ago, and we just purchased two more in the last 3 months. I > am sure that other manufacturers are making equivalent parts,... Any help will > be appreciated. I am using Mitsubishi M5M27C202JK-15 EPROMs on our MVME167s. I bought these about 2 years ago through a local distributor. I don't know if they are still available. Phillip L. Shaffer phillip.shaffer@ae.ge.com GE Aircraft Engines, MS G57 1 Neumann Way Cincinnati, OH 45215 From sidney@peter.boeing.com Wed Nov 8 16:19:00 1995 From: sidney@peter.boeing.com Date: Wed Nov 8 16:19:02 PST 1995 Subject: Re: mvme167 eproms Mitsubishi part# M5M27C202JK should do the job for you. From daemon@csg.lbl.gov Thu Nov 9 04:00:24 1995 From: daemon@csg.lbl.gov Date: Thu Nov 9 04:00:28 PST 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Thu Nov 9 04:00:19 PST 1995 Subject: (no subject) Subject: Multiprocessor Jobs Subject: CAPI for ISDN Modems Subject: Re: Using libg++ on VxWorks Subject: Multiprocessor Jobs Subject: Re: Signetics 68692 driver needed Subject: DBASE using lstLib Subject: Gnat for VxWorks?! ------------------------------------------------------- Newsgroups: comp.realtime,comp.dsp,comp.os.os9,comp.arch.parallel,comp.arch.bus.vmebus,comp.os.vxworks,comp.arch.embedded Subject: (no subject) Date: 8 Nov 1995 14:35:51 GMT From: Ted Thompson Organization: Axiom Technology, Inc. Message-ID: <47qf87$jc6@tsunami.traveller.com> Regional Sales Representatives Axiom Technology is seeking individuals experienced in the multiprocessor market to represent our products and services. Candidates should have previous technical experience in multiprocessing hardware or software plus 1-2 years sales experience. Independent sales representatives currently representing other multiprocessor products are also desired. For more details about Axiom or our products see our WWW page: http://www.axiomtech.com. Axiom Technology develops software products for high performance multiprocessor systems. We offer standard software products, customized software products, consultation and systems integration services for: Multiprocessor Software Development Real Time Applications and Embedded Systems Digital Signal Processing (DSP) Applications VME based Software Tools for Force, Themis, Mercury, SKY, and CSPI Systems Axioms DataFlo MP is a complete graphical programming environment for multiprocessor and signal processing systems. Users choose functions from an extensive set of libraries containing hundreds of signal and data processing functions, or they can write their own library functions. Functions are represented as icons which are connected together to form a block diagram called a graph. Graphs can be executed on workstation hosts or on multiprocessor systems. DataFlo MP has an integrated set of data visualization and editing tools that can be used to analyze and debug applications graphically on the multiprocessor target. Built-in profiling tools allow users to optimize run-time performance. DataFlo MP is now available on the following platforms: Sun, Silicon Graphics, Hewlett-Packard, Mercury, CSPI, SKY, Force, and Themis. If you wish to apply please email your resume to Human.Resources@axiomtech.com, FAX a copy of your resume to 205-721-0805 attn. Human Resources, or send to: Axiom Technology, Inc. Attn: Human Resources 115 Murry Drive Madison, AL 35758 Ted.Thompson@axiomtech.com 205-722-0075 (V) 205-721-0805 (F) Axiom Technology, Inc. 115 Murry Drive Madison, AL 35758 Graphical Tools For Multiprocessor Software Development http://www.axiomtech.com --------------------------- Newsgroups: comp.realtime,comp.dsp,comp.os.os9,comp.os.vxworks,comp.arch.bus.vmebus,comp.arch.embedded,comp.arch.parallel Subject: Multiprocessor Jobs Date: 8 Nov 1995 14:38:36 GMT From: Ted Thompson Organization: Axiom Technology, Inc. Message-ID: <47qfdc$jc6@tsunami.traveller.com> Regional Sales Representatives Axiom Technology is seeking individuals experienced in the multiprocessor market to represent our products and services. Candidates should have previous technical experience in multiprocessing hardware or software plus 1-2 years sales experience. Independent sales representatives currently representing other multiprocessor products are also desired. For more details about Axiom or our products see our WWW page: http://www.axiomtech.com. Axiom Technology develops software products for high performance multiprocessor systems. We offer standard software products, customized software products, consultation and systems integration services for: Multiprocessor Software Development Real Time Applications and Embedded Systems Digital Signal Processing (DSP) Applications VME based Software Tools for Force, Themis, Mercury, SKY, and CSPI Systems Axioms DataFlo MP is a complete graphical programming environment for multiprocessor and signal processing systems. Users choose functions from an extensive set of libraries containing hundreds of signal and data processing functions, or they can write their own library functions. Functions are represented as icons which are connected together to form a block diagram called a graph. Graphs can be executed on workstation hosts or on multiprocessor systems. DataFlo MP has an integrated set of data visualization and editing tools that can be used to analyze and debug applications graphically on the multiprocessor target. Built-in profiling tools allow users to optimize run-time performance. DataFlo MP is now available on the following platforms: Sun, Silicon Graphics, Hewlett-Packard, Mercury, CSPI, SKY, Force, and Themis. If you wish to apply please email your resume to Human.Resources@axiomtech.com, FAX a copy of your resume to 205-721-0805 attn. Human Resources, or send to: Axiom Technology, Inc. Attn: Human Resources 115 Murry Drive Madison, AL 35758 - -- Ted.Thompson@axiomtech.com 205-722-0075 (V) 205-721-0805 (F) Axiom Technology, Inc. 115 Murry Drive Madison, AL 35758 Graphical Tools For Multiprocessor Software Development http://www.axiomtech.com --------------------------- Newsgroups: comp.os.qnx,comp.os.vxworks Subject: CAPI for ISDN Modems Date: 7 Nov 1995 08:28:49 GMT From: Uwe Foss <100071.774@CompuServe.COM> Organization: Siemens AG FSZ Message-ID: <47n5c1$cqn$1@mhafm.production.compuserve.com> Hi all, is a CAPI interface for this realtime os available ? I am interested in ISDN products and CAPI drivers. (ISA, PCMCIA) /////////// Siemens AG Uwe Foss compuserve.com@100071,774 ////////// - -- ! SIEMENS AG FSZ ! ANL TD FSZ EE ! Uwe Foss ! TEL: +49 711-137-6071 --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Using libg++ on VxWorks Date: 8 Nov 1995 15:37:48 GMT From: thor@ymir.atd.ucar.edu (Rich Neitzel) Organization: National Center for Atmosphereic Research Message-ID: <47qisc$pjs@ncar.ucar.edu> References: <9511080507.AA00374@tut.lannet.com> In article <9511080507.AA00374@tut.lannet.com>, mirig@lannet.com (Miri Grossman) writes: |> |> Did anybody try compiling the GNU class library: -- libg++ 2.7.0a -- |> for VxWorks? |> Yes and it works great. Look at the file 'vx_cplusplus' in the user group archive. - -- Richard Neitzel thor@atd.ucar.edu Torren med sitt skjegg National Center For Atmospheric Research lokkar borni under sole-vegg Box 3000 Boulder, CO 80307-3000 Gjx'i med sitt shinn 303-497-2057 jagar borni inn. --------------------------- Newsgroups: comp.realtime,comp.dsp,comp.arch.bus.vmebus,comp.os.os9,comp.parallel,comp.arch.embedded,comp.os.vxworks Subject: Multiprocessor Jobs Date: Tue, 7 Nov 1995 15:32:17 GMT From: Ted Thompson Organization: Axiom Technology Message-ID: <47nu61$bv8@tsunami.traveller.com> Sender: news@relay.acadiau.ca Regional Sales Representatives Axiom Technology is seeking individuals experienced in the multiprocessor market to represent our products and services. Candidates should have previous technical experience in multiprocessing hardware or software plus 1-2 years sales experience. Independent sales representatives currently representing other multiprocessor products are also desired. For more details about Axiom or our products see our WWW page: http://www.axiomtech.com. Axiom Technology develops software products for high performance multiprocessor systems. We offer standard software products, customized software products, consultation and systems integration services for: Multiprocessor Software Development Real Time Applications and Embedded Systems Digital Signal Processing (DSP) Applications VME based Software Tools for Force, Themis, Mercury, SKY, and CSPI Systems Axiom=92s DataFlo MP is a complete graphical programming environment for multiprocessor and signal processing systems. Users choose functions from an extensive set of libraries containing hundreds of signal and data processing functions, or they can write their own library functions. Functions are represented as icons which are connected together to form a block diagram called a graph. Graphs can be executed on workstation hosts or on multiprocessor systems. DataFlo MP has an integrated set of data visualization and editing tools that can be used to analyze and debug applications graphically on the multiprocessor target. Built-in profiling tools allow users to optimize run-time performance. DataFlo MP is now available on the following platforms: Sun, Silicon Graphics, Hewlett-Packard, Mercury, CSPI, SKY, Force, and Themis. If you wish to apply please email your resume to Human.Resources@axiomtech.com, FAX a copy of your resume to 205-721-0805 attn. Human Resources, or send to: Axiom Technology, Inc. Attn: Human Resources 115 Murry Drive Madison, AL 35758 Ted.Thompson@axiomtech.com 205-722-0075 (V) 205-721-0805 (F) Axiom Technology, Inc. 115 Murry Drive Madison, AL 35758 Graphical Tools For Multiprocessor Software Development http://www.axiomtech.com --------------------------- Newsgroups: comp.arch.embedded,comp.arch.bus.vmebus,comp.os.vxworks Subject: Re: Signetics 68692 driver needed Date: 8 Nov 1995 12:27:37 +0100 From: jj@ekf.werries.de ("J. Jansen") Organization: EKF Elektronik GmbH, Hamm FRG Message-ID: <47q479$a2k@ekf.werries.de> References: <47mn1o$7f3@corina.isi.com> dwybo@isi.com (David Wybo) writes: >DOes anyone have any information on how the Signetics 68692 DUART is similar to or different from the 68681? >Does anyone happen to have any driver code for this chip? >Thanks for the help. I think the 68692 has slight enhancements compared to the 68C681 (power down mode, receiver timeout). Pinning is identically, 681 drivers should run. If there is no other source, EKF has an OS-9 3.0 source code package for the 68C681 available (sorry, not for free). - --------------------- EKF Elektronik Messtechnik GmbH --------------------- Joachim Jansen e-mail: jj@ekf.werries.de Philipp Reis Str. 4 phone: ++49 (0)2381-6890-0 D-59065 Hamm (Germany) fax: ++49 (0)2381-6890-90 - ------------------- Advanced VMEbus Hardware + Software ------------------- - -- - --------------------------------------------------------------------------- jj@ekf.werries.de - --------------------------------------------------------------------------- --------------------------- Newsgroups: comp.os.vxworks Subject: DBASE using lstLib Date: Thu, 9 Nov 1995 02:51:59 GMT From: sayeed@netcom.com (Syed Sayeed Ahmed) Organization: NETCOM On-line Communication Services (408 261-4700 guest) Message-ID: Sender: sayeed@netcom9.netcom.com Hello, wonder if anyone knows of any third party database applications/managers that are out there, and have been built using the lstLib set of functions and the vxWorks mutual exclusion mechanism??? thanks. - --sayeed. - -- ************************************************************ Sayeed Ahmed e-mail:sayeed@netcom.com 330 N. Mathilda Ave, # 901 phone: 408-739-2142. Sunnyvale, California, 94086. ************************************************************ --------------------------- Newsgroups: comp.os.vxworks Subject: Gnat for VxWorks?! Date: Thu, 09 Nov 1995 07:58:29 -0800 From: emwmei Organization: Ericsson Message-ID: <30A22525.5EDC@kiemw.ericsson.se> I have heard some rumors about work in progres to port the freely available Ada 95 compiler (gnat) to VxWorks. Anybody that can confirm this and maybe give some more info about the status of the project? Regards Magnus --------------------------- End of New-News digest ********************** From rhin!wrsec.fr!joergb@lbl.gov Thu Nov 9 06:43:00 1995 From: rhin!wrsec.fr!joergb@lbl.gov ( =?iso-8859-1?Q?J=F6rg?= Bertholdt) Date: Thu Nov 9 06:43:02 PST 1995 Subject: Ethernet for 68332 Greetings, I`m looking for a VxWorks driver for an ethernet controller to integrate on a MC68332 custom board. Due to the architecture of the 332 a controller that does not use DMA is preferred. Alternatively info about a BSP for an 68332 Board with Ethernet is also appreciated. Thanks in advance Joerg ************************************************************************* * Joerg Bertholdt Wind River Systems GmbH * * Field Application Engineer Freisinger Strasse 34 * * Email: joergb@wrsec.fr D-85737 Ismaning * * Tel: +49-89-962445-42 Fax: +49-89-962445-55 * * http://www.wrs.com * ************************************************************************* From perryd@boris.lbl.gov Thu Nov 9 08:24:22 1995 From: "David B. Perry" Date: Thu Nov 9 08:24:24 PST 1995 Subject: Code Formatter Hello out there, Anyone know of any C source code formatters? Free would be good, tailorable would be great. We have several styles going in use and would like to 'painlessly' standardize. David -- ===================================================================== David B. Perry TRW e-mail: perryd@boris.msfc.nasa.gov 213 Wynn Dr. phone: (205) 961-1033 Huntsville, AL 35805 fax: (205) 544-2238 ===================================================================== From greg@eng.nmr.varian.com Thu Nov 9 09:56:08 1995 From: greg@eng.nmr.varian.com (Greg Brissey x6951) Date: Thu Nov 9 09:56:10 PST 1995 Subject: Re: Code Formatter David writes: Anyone know of any C source code formatters? . . snip ------- There is the standard UNIX utility indent, for reformating C programs INDENT(1) USER COMMANDS INDENT(1) NAME indent - indent and format a C program source file SYNOPSIS indent input-file [ output-file ] [ [ -bap | -nbap ] [ -bacc | -nbacc ] [ -bad | -nbad ] [ -bbb | -nbbb ] [ -bc | -nbc ] [ -bl ] [ -br ] [ -bs | -nbs ] [ -cn ] [ -cdn ] [ -cdb | -ncdb ] [ -ce | -nce ] [ -cin ] [ -clin ] [ -dn ] [ -din ] [ -eei | -neei ] [ -fc1 | -nfc1 ] [ -in ] [ -ip | -nip ] [ -ln ] [ -lcn ] [ -lp | -nlp ] [ -pcs | -npcs ] [ -npro ] [ -psl | -npsl ] [ -sc | -nsc ] [ -sob | -nsob ] [ -st ] [ -troff ] [ -v | -nv ] DESCRIPTION indent is a C program formatter. It reformats the C program in the input-file according to the switches. The switches which can be specified are described below. They may appear before or after the file names. For more do a man on indent. Greg Brissey Varian greg.brissey@nmr.varian.com From daemon@csg.lbl.gov Fri Nov 10 04:00:22 1995 From: daemon@csg.lbl.gov Date: Fri Nov 10 04:00:27 PST 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Fri Nov 10 04:00:18 PST 1995 Subject: SLIP or PPP on Solaris2.x Subject: Floating Point Math Library Support Needed Subject: Orbix CORBA Implementation Subject: CORBA on VxWorks ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: SLIP or PPP on Solaris2.x Date: 9 Nov 1995 22:21:17 GMT From: Gary Lin Organization: Xerox Corp. Message-ID: <47tust$6v5@news.cp10.es.xerox.com> Hi, I have a question regarding SLIP/PPP under solaris2.x. Currently, we are using sunos4.1.3 and SLIP in communicating to our embedded board thru SPARC 5 station 's serial port. Everything works fine. Now we are going to upgrade from SunOS4.1.3 to Solaris2.x. We are told that SLIP no longer exist on solaris2.x and we have to use PPP . We tried use PPP to talk to the board thru serial port and it does not work ( no response to ping). Most of the PPP document are geared toward modem connection . there maybe some configuration parameter we are missing. Anyway do you know where we can find a SLIP software running on Solaris2.x. Do we know VxWorks supports PPP protocol ? Thanks a lot !! Gary Corporate Research and Technology Xerox corporation --------------------------- Newsgroups: gnu.g++.help,comp.realtime,comp.os.vxworks Subject: Floating Point Math Library Support Needed Date: 9 Nov 1995 14:34:11 -0800 From: kla@leland.Stanford.EDU (Earl Mitchell) Organization: Stanford University, CA 94305, USA Message-ID: <47tvl3$ddj@elaine25.Stanford.EDU> A friend of mine is working on an embedded system which does not have support for floating point math. He's writing his code in C and what's to know if there exist any public domain floating point math libraries he could use. I believe GNU C has built in support for this but I'm not sure. If anybody knows for sure could you please post more info on this option. Also interested to here any other possible alternatives. --------------------------- Newsgroups: comp.os.vxworks Subject: Orbix CORBA Implementation Date: 10 Nov 1995 00:08:50 GMT From: Earl Chew Organization: Hewlett-Packard Co. Message-ID: <47u56i$pod@hpscit.sc.hp.com> I'd like to get in touch with anyone who is using this (or any other) implementation of CORBA under VxWorks. - -- +-------------------------------------------------------------------------+ | Earl Chew Email: earl_chew@aus.hp.com | | Australian Telecom Operation Tel: +61 3 9272 8380 | | Hewlett-Packard Australia Ltd Fax: +61 3 9898 9257 | | 31 Joseph St, Blackburn 3130, Australia GPS: 37 48'44"S 145 8'51"E | +-------------------------------------------------------------------------+ --------------------------- Newsgroups: comp.os.vxworks Subject: CORBA on VxWorks Date: 9 Nov 1995 22:30:16 GMT From: jblack@theo.apl.washington.edu (John Black) Organization: University of Washington Message-ID: <47tvdo$fkm@nntp5.u.washington.edu> Has anyone had any experience with VxWorks implementations of the Common Object Request Broker Architecture (CORBA)? We're interested in using CORBA in a distributed application that runs on Unix machines (Solaris 2.x) and embedded computers (VxWorks). We've found two venders that claim to have VxWorks implementations of CORBA, but we don't necessarily want to be their first customer. Has anyone actually used CORBA and VxWorks to develop a large application? Any help would be appreciated. John Black, jblack@apl.washington.edu --------------------------- End of New-News digest ********************** From tkb@mclean.sparta.com Fri Nov 10 04:38:43 1995 From: Thomas Keith Buchanan Date: Fri Nov 10 04:38:45 PST 1995 Subject: Re: SLIP or PPP on Solaris2.x > > Do we know VxWorks supports PPP protocol ? > I just tested a beta release of 68k PPP from Wind River and it worked well. ______________________________________________________ | |\ |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 | | | / \\\\\ \\ 703-448-0120 | | | \ \\\\\ // | | | \ ////// | | | \////// SPARTA | | | \//// Pride In Performance | | | \// -------------------- | | |_____________________________________________________| | \______________________________________________________\| From froeber@bbn.com Fri Nov 10 05:29:47 1995 From: Fred Roeber Date: Fri Nov 10 05:29:50 PST 1995 Subject: Re: 68681 serial driver >dwybo@isi.com (David Wybo) writes: > >>DOes anyone have any information on how the Signetics 68692 DUART is similar to or different from the 68681? > >>Does anyone happen to have any driver code for this chip? Joachim Jansen responds: >I think the 68692 has slight enhancements compared to the 68C681 >(power down mode, receiver timeout). Pinning is identically, 681 >drivers should run. If there is no other source, EKF has an OS-9 3.0 >source code package for the 68C681 available (sorry, not for free). Wind River has a standard driver for the 68681 chip since that is used on a number of boards. I am not sure how you get access to it. Probably should ask WRS support. Then again, I notice that in the standard VxWorks distribution in the unsupported/config tree there appear to be 68681 supplied for a number of the boards. You might be able to get what you need. Fred | Fred J Roeber, Bolt Beranek and Newman Inc | | 50 Enterprise Place Middletown, RI 02842-5202 | | froeber@bbn.com 401-849-2543 (X48) | From froeber@bbn.com Fri Nov 10 05:29:47 1995 From: Fred Roeber Date: Fri Nov 10 05:29:52 PST 1995 Subject: Re: 68681 serial driver >dwybo@isi.com (David Wybo) writes: > >>DOes anyone have any information on how the Signetics 68692 DUART is similar to or different from the 68681? > >>Does anyone happen to have any driver code for this chip? Joachim Jansen responds: >I think the 68692 has slight enhancements compared to the 68C681 >(power down mode, receiver timeout). Pinning is identically, 681 >drivers should run. If there is no other source, EKF has an OS-9 3.0 >source code package for the 68C681 available (sorry, not for free). Wind River has a standard driver for the 68681 chip since that is used on a number of boards. I am not sure how you get access to it. Probably should ask WRS support. Then again, I notice that in the standard VxWorks distribution in the unsupported/config tree there appear to be 68681 supplied for a number of the boards. You might be able to get what you need. Fred | Fred J Roeber, Bolt Beranek and Newman Inc | | 50 Enterprise Place Middletown, RI 02842-5202 | | froeber@bbn.com 401-849-2543 (X48) | From kapoorsx@nt.com Fri Nov 10 11:58:37 1995 From: Sanjay Kapoor Date: Fri Nov 10 11:58:40 PST 1995 Subject: Corba on VxWorks I noticed a couple of posting regarding Corba on VxWorks. We are currenly using Iona Technologies ORBIX implementation of Corba on VxWorks. So far, it has been a very interesting experience. Some things I've noticed: 1. As CORBA is a moderately complex technology, the problems associated with using can be quite subtle. Having devlopers strongly rooted in OO methodology is definitely a plus. 2. There is a performance penalty with using ORBIX(or any other CORBA implementation for that matter). The question of how big a penalty is still yet to be quantified. I can say, however, that we do have severe real-time constraints, but are managing quite well with ORBIX. All in all, I feel that although going with ORBIX has definitely had a ramp-up time, I think it has allowed us to build a distributed platform that will re-usable by many other groups. Sanjay Kapoor Northern Telecom kapoorsx@nt.com From brett@rti.com Fri Nov 10 15:18:36 1995 From: Brett Murphy Date: Fri Nov 10 15:18:38 PST 1995 Subject: Re: Corba on VxWorks >> Submitted-by kapoorsx@nt.com Fri Nov 10 11:58:37 1995 >> Submitted-by: Sanjay Kapoor >> >> I noticed a couple of posting regarding Corba on VxWorks. We are >> currenly using Iona Technologies ORBIX implementation of Corba >> on VxWorks. So far, it has been a very interesting experience. >> >> 1. As CORBA is a moderately complex technology, the problems associated >> with using can be quite subtle. Having devlopers strongly rooted in OO >> methodology is definitely a plus. >> >> 2. There is a performance penalty with using ORBIX(or any >> other CORBA implementation for that matter). The question of how... ORBIX is a reasonable port of CORBA to VxWorks, but CORBA was never designed for real time. It was designed by committee for a huge range of things from databases to file systems. If you're interfacing VxWorks to CORBA-based apps on Unix boxes, then CORBA makes sense. If you're transmitting data between many real-time nodes, it's going to be a forced fit. #ifdef CrassCommercialOK If you're designing a distributed real-time system, you should consider our Network Data Delivery Service (NDDS) product. It's designed from the ground up for VxWorks and real-time applications. It implements a "dissemination" communications structure; it's much more robust and efficient than any client server design, including CORBA. It also runs on most all flavors of Unix (& soon NT/95). It's also vastly easier to use, has true real-time guaranteed update deadlines and timestamps, offers full control of delivery semantics, handles many-to-many communications, provides instantaneous hot-swap backups, and does asynchronous production/reception [The ability to "produce" data in one task, but transmit it with another. This is critical for real-time applications; you don't want your control loop doing any network transactions.]. Stan Schneider and Gerardo Pardo wrote an interesting short technical paper on information flow for distributed real-time systems. If you're facing these issues, you should check out this paper by two real VxWorks experts. Contact "info@rti.com" for details. #endif -- Brett =============================================================================== = = = = Brett Murphy = email: brett@rti.com = = Real-Time Innovations, Inc. = Phone: (408) 720-8312 = = 155A Moffet Park Drive, Suite 111 = Fax: (408) 720-8419 = = Sunnyvale, CA 94089 = = = = = =============================================================================== From daemon@csg.lbl.gov Sat Nov 11 04:00:27 1995 From: daemon@csg.lbl.gov Date: Sat Nov 11 04:00:32 PST 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sat Nov 11 04:00:22 PST 1995 Subject: Re: Gnat for VxWorks?! Subject: Re: CORBA on VxWorks Subject: Re: SLIP or PPP on Solaris2.x Subject: Automated testing tools Subject: "rsh" capability for vxWorks Subject: i82596 ethernet driver problem ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Gnat for VxWorks?! Date: 10 Nov 1995 03:27:46 GMT From: steved@pop3.pacifier.com (Steve Doiel) Organization: Pacifier BBS, Vancouver, Wa. ((360) 693-0325) Message-ID: <47ugri$40a@news.pacifier.com> References: <30A22525.5EDC@kiemw.ericsson.se> Reply-To: steved@pacifier.com (Steve Doiel) In <30A22525.5EDC@kiemw.ericsson.se>, emwmei writes: >I have heard some rumors about work in progres to port the >freely available Ada 95 compiler (gnat) to VxWorks. >Anybody that can confirm this and maybe give some more info >about the status of the project? >Regards >Magnus While I can't say whether or not anyone is doing a port, I can say that the GNAT authors are not in the process, but are available for hire to do so (they formed a support company Ada Core Technologies). You can check out their web site at http://www.gnat.com (I'd be interested in GNAT for VxWorks too!) Steve Doiel --------------------------- Newsgroups: comp.os.vxworks Subject: Re: CORBA on VxWorks Date: Fri, 10 Nov 1995 18:25:54 GMT From: skv@pcdoh.pgh.wec.com (Suresh K. Vishnubhatla) Organization: Westinghouse Electric Co. Message-ID: References: <47tvdo$fkm@nntp5.u.washington.edu> Reply-To: skv@pcdoh.pgh.wec.com Sender: news@elmo.pgh.wec.com (News Group Owner) In article <47tvdo$fkm@nntp5.u.washington.edu>, jblack@theo.apl.washington.edu (John Black) writes: >Has anyone had any experience with VxWorks implementations >of the Common Object Request Broker Architecture (CORBA)? > >We're interested in using CORBA in a distributed application >that runs on Unix machines (Solaris 2.x) and embedded >computers (VxWorks). > >We've found two venders that claim to have VxWorks implementations >of CORBA, but we don't necessarily want to be their first customer. > >Has anyone actually used CORBA and VxWorks to develop a large >application? > >Any help would be appreciated. > >John Black, jblack@apl.washington.edu > > > I think Post Modern Computing has something like that. Their phone (415) 967-6169 Suresh --------------------------- Newsgroups: comp.os.vxworks Subject: Re: SLIP or PPP on Solaris2.x Date: 10 Nov 1995 18:09:37 GMT From: gmelton@cs.stanford.edu (Gary L. Melton) Organization: Stanford University Computer Science Department Message-ID: References: <47tust$6v5@news.cp10.es.xerox.com> Reply-To: gmelton@xenon.stanford.edu In article <47tust$6v5@news.cp10.es.xerox.com> Gary Lin writes: I have a question regarding SLIP/PPP under solaris2.x. Currently, we are using sunos4.1.3 and SLIP in communicating to our embedded board thru SPARC 5 station 's serial port. Everything works fine. Now we are going to upgrade from SunOS4.1.3 to Solaris2.x. We are told that SLIP no longer exist on solaris2.x and we have to use PPP . We tried use PPP to talk to the board thru serial port and it does not work ( no response to ping). Most of the PPP document are geared toward modem connection . there maybe some configuration parameter we are missing. Anyway do you know where we can find a SLIP software running on Solaris2.x. Do we know VxWorks supports PPP protocol ? I have a very similar problem. We have recently updated from SunOS to Solaris 2.4 and still need to use SLIP for accessing our host. In comp.unix.solaris, Tim Rylance posted a message that recommended these packages: There's a Solaris 2 SLIP written by Bruce Orchard at ftp://ftp.ece.wisc.edu/pub/slip/solaris-slip-5.tar.Z It's easy to install and I'm told it works OK. The latest version of the PPP package "dp" does SLIP as well as PPP on both Solaris 1 and Solaris 2. It's not trivial to install, but it seems very reliable. Get it from ftp://www.acn.purdue.edu/dp/dp-4.0.tar.Z I got these packages (actually, now, there's a solaris-slip-6), and decided to go with solaris-slip-6 just because it seemed easier to install and because we don't need the features for a dial-up on demand connection. Now, however, I'm befuddled on how to set up the serial ports on the Sun workstation. Can anyone provide pointers? Thanks for any help. Gary Melton gmelton@xenon.stanford.edu --------------------------- Newsgroups: comp.os.vxworks Subject: Automated testing tools Date: 10 Nov 1995 14:38:36 -0500 From: clare@bisco.kodak.com (Bill Clare) Organization: Eastman Kodak Message-ID: <4809ns$frb@bissun.bisco.kodak.com> I am interested in automated testing tools that may be used with a VxWorks(tm) development environment. Specifically, I am using a Motorola MVME162LX target, a Sun host, VxWorks 5.1.1 with the associated board support package. We are using the GNU toolkit to compile and debug. Thanks for any help. billc - -- Bill Clare, Eastman Kodak Company Internet: clare@bisco.kodak.com mail code 35424 -- 901 Elmgrove Road Phone: (716) 726-9419 Fax: (716) 726-0129 Rochester, New York 14653-5424 Founding member of the Bill Hajt fan club. Any opinions expressed herein belong to me and not to my employer. --------------------------- Newsgroups: comp.os.vxworks Subject: "rsh" capability for vxWorks Date: Sat, 11 Nov 1995 01:14:46 GMT From: adl@eskimo.com (al lykken) Organization: lyktron Message-ID: Sender: news@eskimo.com (News User Id) We have written our own rsh demon for our vxWorks targets but it is kludgy at best. Does anyone know of or had success with a version of 'rsh' for vxWorks???? We have frozen on version 5.1 and are using FORCE 68040 CPUs. --------------------------- Newsgroups: comp.os.vxworks Subject: i82596 ethernet driver problem Date: Fri, 10 Nov 1995 23:52:27 GMT From: wellsk@netcom.com (Kevin Wells) Organization: NETCOM On-line Communication Services (408 261-4700 guest) Message-ID: Sender: wellsk@netcom7.netcom.com We have a custom R3000 board with an i82596 ethernet interface. The interface contains 16K of shared memory. The 596 (which is an SX) interfaces to the memory in 16-bits. The CPU also interfaces in 16-bits, with the upper 16-bits of the 32-bit bus not used. Will this cause problems for the 596 generic driver? (Is it expecting contiguous 32-bit memory?) We also don't use the port function for the 596. (The PORT pin is physically pulled high in the system). Will the generic driver have a problem with a sys596Port() function that doesn't do anything? Kevin Wells, Lear Astronics Corporation, Software Engineering (310) 915-8358 --------------------------- End of New-News digest ********************** From leonid@rst.co.il Sat Nov 11 23:53:46 1995 From: leonid@rst.co.il (Leonid Rosenboim) Date: Sat Nov 11 23:53:49 PST 1995 Subject: Re: i82596 ethernet driver problem > Date: Fri, 10 Nov 1995 23:52:27 GMT > From: wellsk@netcom.com (Kevin Wells) > We have a custom R3000 board with an i82596 ethernet interface. The > interface contains 16K of shared memory. The 596 (which is an SX) > interfaces to the memory in 16-bits. The CPU also interfaces in > 16-bits, with the upper 16-bits of the 32-bit bus not used. Sounds like typical sloppy hardware design. Seen such cases all around. When for are these hardware people are going to learn that it is not worth to save a couple of ICs and later pay much more in software development time ? > Will this cause problems for the 596 generic driver? (Is it > expecting contiguous 32-bit memory?) Certainly. The generic 596 driver WILL NOT WORK in this hardware configuration. The driver does have macros to handle tge 596 pointers, but only changing these macros wont do the trick. You will have to re-work the driver quote seriously, or have a consultant do that for you. > We also don't use the port function for the 596. (The PORT pin > is physically pulled high in the system). Will the generic driver > have a problem with a sys596Port() function that doesn't do > anything? You will not be able to operate the 596 without the PORT function, because this is the only way to convey commands to the 596 controller. Now if you can't command it, it will not be able to do anything with it. We have done some work on the 596 driver, our last project was to eliminate the memory-to-memory copy during transmission, plus some other performance-related tuneups. ----------------------------------------------------------------------- Leonid Rosenboim Phone: +972-3-559-8144 R S T Software Industries Ltd. Mobile: +972-50-307-142 3 Hashikma st. Industrial Zone, Fax: +972-3-559-8244 P.O.Box 11502, AZUR 58017, Israel E-Mail: leonid@rst.co.il From daemon@csg.lbl.gov Mon Nov 13 04:00:22 1995 From: daemon@csg.lbl.gov Date: Mon Nov 13 04:00:26 PST 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Mon Nov 13 04:00:18 PST 1995 Subject: Re: atexit Subject: Re: VxWorks and SDL Subject: Re: CAPI for ISDN Modems Subject: Re: CORBA on VxWorks Subject: Re: Green Hills Compilers/Tools - Good/Bad ? Subject: Deleting a named pipe ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: atexit Date: Sun, 12 Nov 1995 20:50:15 GMT From: mfisher@dyn.com (Mark Fisher) Organization: The World @ Software Tool & Die Message-ID: References: <47o7ne$kkv@texas.nwlink.com> Sender: news@world.std.com (Mr Usenet Himself) stevecor@nwlink.com (Steve Cornett) wrote: >I am porting to VxWorks a library which uses the standard C function >atexit. Unfortunately atexit is not provided and the taskDeleteHook >facility does not work exactly like atexit. Delete hooks are called >whenever any task exits, rather than when only the caller exits. >To work around this limitation, my plan is to have the delete hook >routine only do its thing if a thread-private variable is set. I will >set the thread-private variable using taskVarAdd at the point when >atexit is normally called. That way, only the caller task will have >this variable set and the atexit handler will ignore all other tasks. >When the atexit handler does its thing, I delete the delete hook so >that subsequent task terminations do not call into the unloaded task > area. >Does anyone know a simpler way to simulate atexit in VxWorks? >-- >Steve Cornett >Bullseye Software Hi, I have also been thinking about this lately. While I haven't tried it yet; I was going to add a delete hook like you say, but instead of using taskVarAdd I would use the user area of the WIND_TCB. In one of those (int sized) elements I would store a pointer to a structure I defined. In that structure I would have a list of routines to call at exit. If the pointer was NULL delete hook would do nothing; making the overhead as small as possible for tasks not using the feature. I don't know if this is easier but is is probably more extensible. There isn't too much said abouit the user area of the WIND_TCB structure in the VxWorks manuals so I have emailed WRS tech support for some example usages etc. Mark Fisher Dynamic ReSolutions --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks and SDL Date: Sun, 12 Nov 1995 10:33:35 GMT From: Henry.Gessau@abc.se (Henry Gessau) Organization: None Message-ID: <484ii3$cgt@oden.abc.se> References: <199511080127.JAA25922@webster.qpsx.oz.au> On Wed, 08 Nov 1995 09:11 WAT, RUSSELL@qpsx.oz.au (Jones, Russell) wrote: > We've recently purchased VxWorks 5.2, which we'll be using in a telephony > switch that we're developing. We're doing a lot of our signalling > specification and design using SDL (Specification and Description > Language), and we've found a CASE tool called SDT which we're currently > evaluating. SDT certainly looks very useful, allowing us to simulate and > validate our design quite stringently, and it does produce C output code > (although it would not do badly in an obfuscation competition). What > we're lacking is any way of connecting this with VxWorks. Has anyone come > across SDT (by Telelogic in Sweden), and can they help us port the > generated code into VxWorks? We are in the process of doing the same thing with SDT 3.02 and VxWorks 5.3. Currently we have a project using SDT 2.3 with code generation and a real-time OS called OSE. This works well, so we are hoping that we won't have too much trouble with the new version of SDT, although we have not used VxWorks before. Telelogic provides adaptation modules for many real-time OS's like OSE, VxWorks, pSOS, etc, or it can run OS-less with its own real-time executive. Do you know that there is a competing SDL tool called Object Geode by Verilog? You should check it out too and compare it to SDT if you're in the evaluation stage. Verilog is a French company (you guys in Oz still boycotting froggy products?). Keep in touch. - -- Henry Gessau ABB Signal, Stockholm, Sweden --------------------------- Newsgroups: comp.os.qnx,comp.os.vxworks Subject: Re: CAPI for ISDN Modems Date: Thu, 09 Nov 1995 14:54:44 +0000 From: 0735128830-0001@t-online.de (T.Snitko) Message-ID: <47t884$s32@news00.btx.dtag.de> References: <47n5c1$cqn$1@mhafm.production.compuserve.com> Re: CAPI for ISDN Modems Hi, Use a V24 ISDN-Modem. i.e. ELINK-310 from EEH-Datalink. You needn't a CAPI ! It works fine with qtalk. Thaddaeus Snitko --------------------------- Newsgroups: comp.os.vxworks Subject: Re: CORBA on VxWorks Date: 12 Nov 1995 22:21:12 GMT From: Earl Chew Organization: Hewlett-Packard Co. Message-ID: <485s0o$m2k@hpscit.sc.hp.com> References: <47tvdo$fkm@nntp5.u.washington.edu> We are also interested in similar things. Would you be able to forward or summarise any replies that you get. I know of two vendors: o Post Modern Computing has an implementation. o Orbix has an implementation. I am interested in any experiences people have had with these products on VxWorks. Earl - -- +-------------------------------------------------------------------------+ | Earl Chew Email: earl_chew@aus.hp.com | | Australian Telecom Operation Tel: +61 3 9272 8380 | | Hewlett-Packard Australia Ltd Fax: +61 3 9898 9257 | | 31 Joseph St, Blackburn 3130, Australia GPS: 37 48'44"S 145 8'51"E | +-------------------------------------------------------------------------+ --------------------------- Newsgroups: comp.realtime,comp.os.vxworks,comp.unix.aix Subject: Re: Green Hills Compilers/Tools - Good/Bad ? Date: 13 Nov 1995 00:22:37 GMT From: yotakiza Organization: Internet Initiative Japan Inc., Tokyo, JAPAN Message-ID: <48634d$n5m@news2.iij.ad.jp> References: > I have seen a demo of the Green Hills C compiler and > their MULTI software development environment for a > VxWorks target (from an AIX host). It looked to be quite good. > > Does Green Hills build a good reliable compiler? > Is their support good? > > Any comments, negative or positive are welcome. > > gregb@gemini.den.mmc.com > -- > Hello, I will try "Green Hills C compiler : MULTI" for AXP on PC/Windows3.1. Soon, I think that send the report about "MULTI". Bye. --------------------------- Newsgroups: comp.os.vxworks Subject: Deleting a named pipe Date: 13 Nov 1995 06:17:10 GMT From: John Hughes Organization: The University of Arizona Message-ID: <486nt6$d3g@news.ccit.arizona.edu> Does anyone happen to know how to remove a named pipe in the event that an error occurs and it is not feasible to open it? Example: pipe is created with a call to pipeDevCreate(), but some error condition prevents it from being opened for subsequent use. What happens if the pipeDevCreate() call is done again using the same name? Since VxWorks pipes are a pseudo-device built on the message queue facility, it doesn't look like the iosDevDelete() will work to rid it from the device descriptor tables. In my application I may need to be able to reuse a pipe name, since it is being determined a function which would have no knowledge of the existance of a pipe prior to its request. Of course, I could assume that I'll never have an error condition, I suppose... ;) Any suggestions/hints/tips cheerfully accepted. - ------------------------------------------------------------------------------ John M. Hughes VOX: 520-621-7804 Senior Systems Programmer FAX: 520-621-9843 Steward Observatory, University of Arizona jmh@canopus.as.arizona.edu --------------------------- End of New-News digest ********************** From del@rst.com Mon Nov 13 06:41:43 1995 From: del@rst.com (Mark Del Giorno) Date: Mon Nov 13 06:41:45 PST 1995 Subject: Test-and-Set Instruction(s)? Heyyyyyyyyyyyyyyyyyyyy, I'm using the vxTas() instruction to set a semaphore that exists on a VME memory board. It seems to work fine, but when I dissassemble the code itself, it's not quite what I expected: -> l vxTas _vxTas: 00a384 7000 MOVEQ #0,D0 00a386 206f 0004 MOVEA .L (0x4,A7),A0 00a38a 4e71 NOP 00a38c 4ad0 TAS (A0) 00a38e 4e71 NOP 00a390 6b08 BMI 0x0000a39a 00a392 4a80 TST .L D0 00a394 6bf4 BMI 0x0000a38a 00a396 7001 MOVEQ #0x1,D0 00a398 4e75 RTS 00a39a 4a80 TST .L D0 00a39c 6bec BMI 0x0000a38a 00a39e 4e75 RTS [My target is a MVME162] More out of curiousity than anything else, does anyone understand the reason for all the "fluff" around the TAS instruction? - why the NOP's? - why do the 2nd/3rd BMI's (which seem to serve no purpose) branch back to do another TAS (which would do another VME bus cycle)? _/_/_/_/ _/_/_/_/ _/_/_/_/_/ Mark Del Giorno (del@rst.com) _/ _/ _/ _/ (410) 876-9200 Fax: 876-9470 _/_/_/_/ _/_/_/_/ _/ Robotic Systems Technology _/ _/ _/ _/ 1110 Business Parkway South _/ _/ _/_/_/_/ _/ Westminster, MD 21157 From froeber@bbn.com Mon Nov 13 09:33:29 1995 From: Fred Roeber Date: Mon Nov 13 09:33:31 PST 1995 Subject: Re: Test-and-Set Instruction(s)? Mark Del Giorno asked: >I'm using the vxTas() instruction to set a semaphore that exists >on a VME memory board. It seems to work fine, but when I dissassemble >the code itself, it's not quite what I expected: > >-> l vxTas > _vxTas: >00a384 7000 MOVEQ #0,D0 >00a386 206f 0004 MOVEA .L (0x4,A7),A0 >00a38a 4e71 NOP >00a38c 4ad0 TAS (A0) >00a38e 4e71 NOP >00a390 6b08 BMI 0x0000a39a >00a392 4a80 TST .L D0 >00a394 6bf4 BMI 0x0000a38a >00a396 7001 MOVEQ #0x1,D0 >00a398 4e75 RTS >00a39a 4a80 TST .L D0 >00a39c 6bec BMI 0x0000a38a >00a39e 4e75 RTS > >[My target is a MVME162] > >More out of curiousity than anything else, does anyone understand >the reason for all the "fluff" around the TAS instruction? > - why the NOP's? > - why do the 2nd/3rd BMI's (which seem to serve no purpose) branch > back to do another TAS (which would do another VME bus cycle)? There are some wierd issues with 68040s and atomic VMEbus operations. The TAS instruction does a RMW (read-modify-write) cycle. This will be done across the VMEbus in some cases. The possibility of the instruction being aborted due to a deadlock condition can happen. The CPU receives a bus error in such cases. In some 680X0 processors there is support for automatic retries of the instruction in such cases. The 68040 doesn't have such a retry capability so some explicit way of retrying the instruction is needed. If I remember right from when I looked at this a few years back, that is where the D0 check comes in; the bus error handler uses D0 to signal the need for a retry. As far as the NOPs being there, they take care of flushing the instruction pipeline in the processor so that if a bus error does happen, the error PC address reflects the correct instruction location and memory doesn't get corrupted on the return. Having standard code to handle all the wierd errors and idiosynchrasies built into the different processor chips out there is one of the big things that I find VxWorks buys you. It is easy to overlook that benefit too. Fred | Fred J Roeber, Bolt Beranek and Newman Inc | | 50 Enterprise Place Middletown, RI 02842-5202 | | froeber@bbn.com 401-849-2543 (X48) | From leec@manta.nosc.mil Mon Nov 13 14:25:40 1995 From: leec@manta.nosc.mil (Lam Lee) Date: Mon Nov 13 14:25:43 PST 1995 Subject: TCL, TK, and VxWorks Hi there! I am trying to make TCL and TK to work with VxWorks from WRS. Is anyone successfully doing that yet? If so, please let me know how did you accomplish that. Thank you. Lam Lee leec@manta.nosc.mil (619) 553-6454 From daemon@csg.lbl.gov Tue Nov 14 04:00:21 1995 From: daemon@csg.lbl.gov Date: Tue Nov 14 04:00:26 PST 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Tue Nov 14 04:00:15 PST 1995 Subject: GPIB esp488 driver for AXPvme Subject: test Subject: VxVMI experiences (?) Subject: Appl. can run on VxSim but crash when run on actual target ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: GPIB esp488 driver for AXPvme Date: 13 Nov 1995 16:00:44 GMT From: "Phil S. Wilshire" Organization: Internet On-Ramp, Inc. Message-ID: <487q3c$l56@express.ior.com> Hi, I am just starting to port the NI esp488 driver for their GPIB 1014 vme board to the Alpha AXP vme platform. Has anyone already done this ??? If not is any one interested in the final product ?? The port will take about a week to complete and about two more to test. I'll have to talk to NI to see if they would let me publish the completed port. Thanks Phil Wilshire --------------------------- Newsgroups: comp.os.vxworks Subject: test Date: 13 Nov 1995 23:58:47 GMT From: Gary Lin Organization: Xerox Corp. Message-ID: <488m3n$18h@news.cp10.es.xerox.com> 3D""3D""3D""= 3D""3D""
Newsgroup: comp.os.vxworks

Newsgroup: comp.os.vxworks

3D""3D""3D""= 3D""3D"" --------------------------- Newsgroups: comp.os.vxworks Subject: VxVMI experiences (?) Date: 13 Nov 95 21:00:49 GMT From: gregb@den.mmc.com (gregb) Organization: Lockheed Martin -- Denver InterNetNews site Message-ID: Reply-To: gregb@gemini.den.mmc.com What type of memory and process protection features are added with the VxVMI module? Is the performance impact (size/speed)? another 10% (?) gregb@gemini.den.mmc.com - -- --------------------------- Newsgroups: comp.os.vxworks Subject: Appl. can run on VxSim but crash when run on actual target Date: Tue, 14 Nov 1995 14:58:44 -0800 From: "Chonticha P." Organization: National Electronics and Computer Technology Center, Bangkok Message-ID: <30A91F24.2337@ttl.nectec.or.th> I 've recently got the SNMPV2 from the vxworks archive (ftp.atd.ucar.edu /pub/vxworks/vx/snmp2.01-snmp2.40).It compiled and run quite well on VxSim 5.2 (which running on sparc 5 machine using SunOS 4.1.3_U1).But when I port this to our target which is a 80486 machine with 8MB. RAM installed (with minor modification in a header file to set the correct byte order according to information inside the code) it is total not works. The code compiled smoothly quite the same as in VxSim, but when I try to run application it prints out several lines of the following message: 0x7d8e14 (xxv) memPartFree : invalid block 0x7d7630 in partition 0x17b488. . . . and crash the sytem in the middle of application (or sometime when the application terminate) and print out the following message: General Protection Fault Program Counter: 0x0013c4d0 Status Register: 0x00010282 Error Code: 0x007d000 Task : 0x7f9d44 "tExcTask" What I do not understand is why the same application can run on VxSim but not on the target.I have checked the compile option that I used on the target several time and found no problem (per my knowledge and experience) Does anyone know or have the similar experience that can point me out of this ? Thank you. -- Chonticha P. TTL-LAB, NECTEC, Thailand. --------------------------- End of New-News digest ********************** From nora@wrs.com Tue Nov 14 16:24:25 1995 From: nora@wrs.com (Nora Mitchell) Date: Tue Nov 14 16:24:28 PST 1995 Subject: Wind River Users Conference & Exhibition FINAL NOTICE! There is still time to register for the Wind River Users Conference and Exhibition this Thursday, November 16. To register simply email: wrug@wrs.com, or register throught the Web at:http://www/wrs/com/wruscon.html, or call 800-545-WIND. While we prefer that you register in advance, we will accept on-site registrants. Registration starts at 7:30am, conference starts at 8:00am. For more information please call 510-814-2177. Location: Renaissance Meeting Center at Techmart 5201 Great America Parkway Santa Clara, CA 408-562-6111 We look forward to seeing you there. -Nora ============================ Nora Mitchell Marketing Promotions Manager Wind River Systems 1010 Atlantic Ave Alameda, CA 94501 510-814-2052 tel 510-814-2010 fax nora@wrs.com http://www.wrs.com From daemon@csg.lbl.gov Wed Nov 15 04:00:20 1995 From: daemon@csg.lbl.gov Date: Wed Nov 15 04:00:24 PST 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Wed Nov 15 04:00:15 PST 1995 Subject: Booting from Flash Subject: PPP VS. SLIP on Solaris2.x Subject: Available compilers? Subject: VxWorks / Alpha AXP ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Booting from Flash Date: 14 Nov 1995 18:46:03 GMT From: getz@getz.manassas.ibm.com Organization: Loral Federal Systems, Manassas Message-ID: <48ao5b$965@lfsserv1.fsc.ibm.com> Reply-To: getz@getz.manassas.ibm.com Is it possible to boot VxWorks on a Force 5CE processor from FLASH memory using the VxWorks boot prom? If so, what is the proper way to do this? I'm told it can be done using the OpenBoot prom, but we don't have the OpenBoot prom on our 5CE's. Thanks, Fred Getz --------------------------- Newsgroups: comp.os.vxworks Subject: PPP VS. SLIP on Solaris2.x Date: 14 Nov 1995 22:13:35 GMT From: Gary Lin Organization: Xerox Corp. Message-ID: <48b4af$4qa@news.cp10.es.xerox.com> Hi, Does anyone in this group configure a PPP connection from Sun host to embedded board thru serial port successfully ? We have a Sun host running SunOS4.1.3 , and have been using the SLIP to talk to our embedded board thru the serial port , no problem . Now we want to move to solaris2.x and my questions are: 1. Is there a SLIP module for solaris2.x we can use ? 2. If not, has anyone tried PPP connection to their embedded board thru serial port and succeeded. We tried PPP on solaris2.x and it did not work. Most of the manual on PPP are modem related stuff. Any help would be appreciated Gary --------------------------- Newsgroups: comp.os.vxworks Subject: Available compilers? Date: Wed, 15 Nov 1995 02:36:08 GMT From: sgaeke@erinet.com (Steve Gaeke) Organization: EriNet Online 513 436-9915 Message-ID: <48b8md$2o3@eri1.erinet.com> We have been looking at VxWorks in an embedded 80486 environment. My one complaint is that it uses the GNU compiler. Are there any other compilers that can build a VxWorks application under an 80486. We have had excellent results with the Watcom compiler in the past and would love to continue to use it. Anyone have any suggestions on how it might play with VxWorks and the debugger? steve. --------------------------- Newsgroups: comp.os.vxworks Subject: VxWorks / Alpha AXP Date: 15 Nov 1995 10:19:42 GMT From: David Costain <100072.3624@CompuServe.COM> Organization: VG Organic Message-ID: <48ceru$p2e$1@mhade.production.compuserve.com> We are currently using VxWorks RTV3.2 for OSF/1 Alpha running on a DEC Alpha AXP PCI 33 embedded system. We are alarmed that even before loading our application, the basic configuration consumes 8MBytes of RAM. We already ship a MC68020 version of the application, and this has very modest memory requirements ( App + VxWorks < 1MByte). The Alpha CONSOLE uses seems to use somewhere between 2 and 6 MByte (DEC in the UK seem unsure of this!). We are desperate to get the memory size down - can anyone help! Thanks in anticipation! --------------------------- End of New-News digest ********************** From leonid@rst.co.il Wed Nov 15 04:57:42 1995 From: leonid@rst.co.il (Leonid Rosenboim) Date: Wed Nov 15 04:57:45 PST 1995 Subject: MC68302 serial driver bug The UART serial driver supplied by WindRiver for the MC68302 SCC with VxWorks has a bug in the interrupt service routine. THis bug would casue the transmitted to freeze under certain situations. This problem is more probable when the 302 is supplied with a low frequency clock ( e.g. 16Mhz and below), and when the UART is heavily used. Attached is the diff to the driver source which is supplied by WindRiver with the relevant BSP. We hope this patch will be useful to some of you. *** m68302Serial.c.orig Sun Feb 28 11:20:50 1993 --- m68302Serial.c Wed Nov 15 13:47:18 1995 *************** *** 6,11 **** --- 6,12 ---- /* modification history -------------------- + 01z,15dec95,rst fixed race condition in tyCoInt(). 01l,01dec92,jdi NOMANUAL for tyCoInt() - SPR 1808. 01k,28oct92,caf tweaked tyCoDevCreate() documentation and parm checking. 01j,22oct92,jcf increased transmit buffer size. *************** *** 369,375 **** /* acknowledge interrupt */ ! pDv->pSccReg->scce = UART_SCCE_RX | UART_SCCE_TX; /* clear events */ *IMP_ISR = pDv->intAck; /* ack. interrupt */ if (!pDv->created) --- 370,376 ---- /* acknowledge interrupt */ ! pDv->pSccReg->scce = event; /* clear events */ *IMP_ISR = pDv->intAck; /* ack. interrupt */ if (!pDv->created) ----------------------------------------------------------------------- Leonid Rosenboim Phone: +972-3-559-8144 R S T Software Industries Ltd. Mobile: +972-50-307-142 3 Hashikma st. Industrial Zone, Fax: +972-3-559-8244 P.O.Box 11502, AZUR 58017, Israel E-Mail: leonid@rst.co.il From ZINGLEBJ@ocd231.bwi.wec.com Wed Nov 15 05:30:51 1995 From: "Zingler, Bill J." Date: Wed Nov 15 05:30:53 PST 1995 Subject: vxWorks Training Class in D.C. area We have 3 to 4 engineers that need to participate in the VxWorks Training Workshop and the company cannot afford to send them to Orlando. If there are any other individuals or groups organizing a VxWorks Training Workshop in the D.C. area in the next few months and are willing to have others join you, please let me know. We would also be willing to organize a workshop at our site if we get enough interest. Thanks for your time, Bill _________________________________________________________________ Bill Zingler Senior Software Engineer Westinghouse Electric Corporation Oceanic Division/D.W. Complex P.O. Box1488/MS-9RDW Annapolis, MD 21404 Phone: 410-260-5507 Fax: 410-260-5916 Email: zinglebj@ocd231.bwi.wec.com _________________________________________________________________ From mstull@fci.com Wed Nov 15 07:49:22 1995 From: mstull@fci.com (Mark C. Stull) Date: Wed Nov 15 07:49:25 PST 1995 Subject: Re:Booting from Flash Fred, The BSP from FORCE COMPUTERS, includes routines to program and display the user flash onboard the CPU-5CE. These routines can be found in config/cpu5ce/bspsrc/flashExt.c. If the WRS 5.2BSP you are using is missing this file, you may also use the CPU-5TE 5.2BSP which you should receive today. If you have any questions, please feel free to contact me directly. Best Regards, Mark > >Newsgroups: comp.os.vxworks >Subject: Booting from Flash >Date: 14 Nov 1995 18:46:03 GMT >From: getz@getz.manassas.ibm.com >Organization: Loral Federal Systems, Manassas >Message-ID: <48ao5b$965@lfsserv1.fsc.ibm.com> >Reply-To: getz@getz.manassas.ibm.com > >Is it possible to boot VxWorks on a Force 5CE processor from FLASH memory using >the VxWorks boot prom? If so, what is the proper way to do this? >I'm told it can be done using the OpenBoot prom, but we don't have the OpenBoot >prom on our 5CE's. > >Thanks, >Fred Getz > > ****************************************************************** Mark C. Stull Voice: (410) 674-0003 Field Applications Engineer Fax: (410) 674-0008 FORCE COMPUTERS, Inc. Internet: mstull@fci.com 1130 Annapolis Rd - Suite 203 Odenton, Maryland 21113 ****************************************************************** From lahtinen@tsl.uu.se Wed Nov 15 08:18:01 1995 From: lahtinen@tsl.uu.se Date: Wed Nov 15 08:18:04 PST 1995 Subject: Backplane boot problem Hi All, I have tried to setup backplane network for two MVME167 boards to boot VxWorks 5.2 from a SUN host (SPARCstation LX running Solaris 2.4). If both of those boards are booted directly via Ethernet there are no problems. If, however, the second 167 is trying to boot via the shared memory network the booting will end up with an error message: Backplane anchor at 0xf0400600... Attaching network interface sm0... done. registering proxy client: xxx.xxx.67.72.....client registered failed 4c0006 Can't load boot file ! At the time when the slave prints out 'registering proxy client ...' the master hangs up completely and must be reset. The slave comes back to the 'VxWorks boot:' prompt. According to the vwModNum.h and proxyArpLib.h files the above mentioned error is a timeout coming from the M_proxyArpLib module. The shared memory network resides in the same subnet as the boot host and the proxy server features are enabled in the configAll.h file. For some reason it seems that the slave is not able to get through the gateway and fetch its boot files. I would really appreciate if someone could give me some ideas of what to test next. I've run out of my own. Thanks in advance, Mikko Lahtinen lahtinen@tsl.uu.se The Svedberg Laboratory S-75121 Uppsala SWEDEN PS. Some more detailed information of our setup follows. The boot parameters for the master and slave CPU are: Parameter SVEDAQEC1 (master) SVEDAQED1 (slave) ---------- ------------------ ----------------- boot device: ei sm=0xf0400600 processor number: 0 4 host name: sundaq1 sundaq1 file name: /sd/boot/ec1/vxWorks /sd/boot/ed1/vxWorks inet on ethernet (e): xxx.xxx.67.70 inet on backplane (b): xxx.xxx.67.71 xxx.xxx.67.72 host inet (h): xxx.xxx.67.136 xxx.xxx.67.136 gateway inet (g): xxx.xxx.67.71 user (u): vxworks vxworks ftp password (pw): flags (f): 0x0 0x100 target name (tn): svedaqec1 svedaqed1 startup script (s): other (0): Definitions done on the configuration files: - INCLUDE_RPC defined - INCLUDE_NFS defined - INCLUDE_SM_NET defined - INCLUDE_PROXY_SERVER defined - LOCAL_MEM_BUS_A24 defined to be 0x400000 for the master and 0xc00000 for the slave - LOCAL_MEM_BUS_ADRS defined to be 0x1000000 for the master and 0x2000000 for the slave - sysPhysMemDecs() has been modified to map the above mentioned address spaces - on-board SM pool is allocated from the default location on the master - the names of the master, slave and boot host are in the /etc/host and /etc/host.equiv files - routing table on the boot host has been changed to contain an entry for the proxy slave - the permissions of the VxWorks image and symbol table files (and their parent directories) should be correct These things seem to work: - smNetShow on the master is showing that the SM heartbeat is ticking - the slave can access the SM anchor at the address 0xf0400600 - the slave sees the heartbeat ticking - separately both CPUs can be booted directly from the boot host without problems From greg@eng.nmr.varian.com Wed Nov 15 08:51:37 1995 From: greg@eng.nmr.varian.com (Greg Brissey x6951) Date: Wed Nov 15 08:51:39 PST 1995 Subject: Re: Slip on Solaris Gary writes: We have a Sun host running SunOS4.1.3 , and have been using the SLIP to talk to our embedded board thru the serial port , no problem . Now we want to move to solaris2.x and my questions are: 1. Is there a SLIP module for solaris2.x we can use ? 2. If not, has anyone tried PPP connection to their embedded board thru serial port and succeeded. We tried PPP on solaris2.x and it did not work. Most of the manual on PPP are modem related stuff. snip ------------------- We used slip on SunOS4.1.3, then when moving to Solaris we obtain a source of the slip driver either from Wind River or SUN itself, however this driver was not usable on Solaris 2.X. We did get it to work but if you powered down the target or any other glitch in slip communication resulted in the SUN completely hanging, requiring to reboot it. We gave up and used a SUNOS 4.1.3 machine as a slip gateway. Then we move to a backplane net. This was a year ago so maybe the driver has been improved. I'd check with Wind River or SUN for a slip driver. Greg Varian Palo Alto, CA greg.brissey@nmr.varian.com From lori_hanning@postman.trilogic.com Wed Nov 15 13:17:59 1995 From: lori_hanning@postman.trilogic.com Date: Wed Nov 15 13:18:02 PST 1995 Subject: MATRIX 540 BSP Available The following is being submitted under the Exploder's 'occasional commercial notice' rules. TRILOGIC is New England's leading sales and support organization for VMEbus products. Providing you with specific VMEbus components or with complete multi-vendor VMEbus systems, TRILOGIC can help you deliver your products to market quickly. TRILOGIC offers a complete selection of CPU technologies including 68000, SPARC, ALPHA, and 486/Pentium. We augment our hardware offerings with real-time operating systems, debug and development tools, and standard and custom software drivers for many off-the-shelf Third-party VMEbus boards. ########## What's New in VxWorks... ############### The MATRIX CPU540 VxWorks Solution ---------------------------------- The CPU540 from MATRIX Corporation features: o Rugged 6U VMEbus processor board o Targeted at the most demanding control and communications intensive applications. o Supports concurrent processing on its MC68040 module and I/O on its two MC68360 modules. VxWorks Support from Target Technologies features: o VxWorks kernel on all three processors o Standard BSP functionality/compatibility o Ethernet or UART support for 68360 nodes o Inter-processor communications - TCP/IP or high speed virtual channel - Point to point between each 360 & the 040 - Interrupt/event driven - Guaranteed packet sequencing o Flash support for customizing the bootrom or for application usage For more information contact: TRILOGIC 187 Ballardvale Street, Wilmington, MA 01887 Phone (800)486-3575 Fax(508)657-5927 info@trilogic.com From leonid@rst.co.il Thu Nov 16 05:52:34 1995 From: leonid@rst.co.il (Leonid Rosenboim) Date: Thu Nov 16 05:52:37 PST 1995 Subject: VME 100baseT We have a customer looking for a VME Fast Ethernet card that supports 100BaseT. Any information would be greatly appreciated. Leonid From prb@aplexus.jhuapl.edu Thu Nov 16 06:29:24 1995 From: prb@aplexus.jhuapl.edu (Paul R. Bade) Date: Thu Nov 16 06:29:26 PST 1995 Subject: Re: VME 100baseT Leonid Rosenboim writes: > > We have a customer looking for a VME Fast Ethernet card that supports 100BaseT. > Any information would be greatly appreciated. > VI computer is building a PowerPC 604 with 100 baseT4 ethernet based on our SPECs. We chose 100 baseT4 because we needed all I/O via P2 and we were concerned about the coupling of 125MHZ of 100 BaseTX to P2. You can contact hwhite@vicomp.com to get more information and the current schedule. +====================================================================+ | __ ____ __ __ | | /\ \ / \ \ /\_\__ /\ \ Johns Hopkins University | | / \_\ | /\ \ \ / / /\_\\ \ \ Applied Physics Lab. | | / /\ | |\ \/ \ \ / / / / / \ \ \ | | / \/ |_| \ /\ \_\ / / / / / \ \ \ Paul R. Bade | | / /\__/_/ \ \ \/_// / / / / / / / (301)-953-6000 x8681 | | / / / \ \_\ \ \/ / / / / / prb@aplexus.jhuapl.edu | | \/_/ \/_/ \__/_/ \/_/ | | __ ____ __ __ | | /\ \ / \ \ /\ \ /\ \ | | / \_\ | /\ \ \ / \_\ / \ \ | | / /\ | |\ \/ \ \ / /\ |_| / /\ \ \ | | / \/ |_| \ /\ \_\ / / / | | \/ / \ \ | | / /\ / / \ \ \/_// / / / / / /\ \_\ | | \ \/ / / \ \_\ \ \/ / / \/ / / / | | \__/_/ \/_/ \__/_/ /_/_/ | | | +====================================================================+ Acct # 255 Site ID: 159 We now have a 3 year site license # 7513 PC# 707115 Previous Support IDs: 1726, 1944, 2031 Host: Sun Sparc Targets: MV167, MV197, PowerPC From Hicks_Brian/macom_pbl_ctc@mailback.macom.com Thu Nov 16 10:35:04 1995 From: Hicks_Brian/macom_pbl_ctc@mailback.macom.com Date: Thu Nov 16 10:35:07 PST 1995 Subject: Force-10 NVRAM read/writes Item Subject: Message text We are running a Force-10 board, and are trying to solve a board lockup problem by using NVRAM area. I know there is available space in the NVRAM area, and need to know if anyone knows how to access it (read/write) directly or via VxWorks. Thanks, Brian Hicks M/A-COM Lowell, MA. From burns.pentek.com!tkl@pentek.com Thu Nov 16 13:28:42 1995 From: tkl@burns.pentek.com (Terry K. Lim) Date: Thu Nov 16 13:28:49 PST 1995 Subject: 162LX D64 DMA Hello, Does anyone have sample code on setting up the VMEchip2 on the MVME162LX for D64 DMA transfers? Any info at all would help. Thanks, -- Terry Lim | Voice: (201) 767-7100 Director of Software Engineering | FAX: (201) 767-7179 Pentek, Inc. | BBS: (201) 767-8733 From uszvnzw6@ibmmail.com Thu Nov 16 13:34:36 1995 From: uszvnzw6@ibmmail.com Date: Thu Nov 16 13:34:39 PST 1995 Subject: MULTI is bad n >> I have seen a demo of the Green Hills C compiler and >> their MULTI software development environment for a >> VxWorks target (from an AIX host). It looked to be quite good. >> >> Does Green Hills build a good reliable compiler? >> Is their support good? >> >> Any comments, negative or positive are welcome. >> >> gregb@gemini.den.mmc.com > >-- >> >Hello, I will try "Green Hills C compiler : MULTI" for AXP on PC/Windows3.1. >Soon, I think that send the report about "MULTI". >Bye. > yotakiza We have been using MULTI on PC/Windows 3.1 host for a target 486 PC for the past 6 months in our project. Overall, we all hate MULTI and look forward to the upcoming Tornado upgrade within the next couple of months. I have reported a number of bugs with MULTI but still have not yet received any bug fix from Green Hills via WRS technical support. For example, one annoying bug is that multiple cut and paste operations among several MULTI editor windows sometimes cause the host PC to lock up. To anyone who is stuck with using MULTI on a PC host, I can only recommend that you use MULTI strictly to perform the builds and use a separate Windows-based editor. Since it is very likely that MULTI was originally developed for the UNIX environment, it is possible that it may be more reliable on a UNIX host if you don't mind the hostile user interface. So far, I find that WRS support for PC host and x86 target tends to lag behind that for Unix host and 68K target. From rml@barnacle.ucsd.edu Thu Nov 16 15:45:13 1995 From: rml@barnacle.ucsd.edu (Bob Lawhead) Date: Thu Nov 16 15:45:15 PST 1995 Subject: 3U video capture H/W Does Anyone know of a manufacture of 3U VME video capture cards? Card should accept "NTSC" video input, 512x512. 8-bit GS or color OK. My application requires only an occasional "frame grab". Any suggestions appreciated. Robert M. Lawhead, Senior Programmer/Analyst Marine Physical Laboratory Scripps Institution of Oceanography University of California, San Diego San Diego, CA 92093-0205 voice: (619) 534-6205 FAX: (619) 534-6849 email: rlawhead@ucsd.edu From jlf@se04.wg2.waii.com Thu Nov 16 23:16:29 1995 From: jlf@se04.wg2.waii.com (Jeff Franks) Date: Thu Nov 16 23:16:32 PST 1995 Subject: Re: TCL, TK, and VxWorks the vxWorks Users Group Exploder writes: > Submitted-by leec@manta.nosc.mil Mon Nov 13 14:25:40 1995 > Submitted-by: leec@manta.nosc.mil (Lam Lee) > > Hi there! > > I am trying to make TCL and TK to work with VxWorks from WRS. Is anyone > successfully doing that yet? If so, please let me know how did you > accomplish that. > > Thank you. > > Lam Lee > leec@manta.nosc.mil > (619) 553-6454 > snip... We've used TCL with VxWorks 5.1.1 by following the instructions and readme files, and qualifying with the test suite. As I recall, there was very little problem in making TCL work. We had found a bug when using TCL in more than one task and using the async event handler. We fixed this by changing some global's related to async event lists and incorating them into the TCL object itself such that each TCL interpreter has it's own async event list to handle. We've not used TK at all with VxWorks. regards, +--- |Jeff Franks | QC Tools, Inc. | No need to worry |franks@wg2.waii.com | of Input/Output, Inc. #####---- | about a schedule, |franksjl@neosoft.com | 3600 Briarpark \ I / O / | I have no time left |(713) 964-6199 | Houston, TX 77042 ---#### | for that. From kelvin.martin@def.bae.co.uk Fri Nov 17 01:53:25 1995 From: kelvin.martin@def.bae.co.uk (MARTIN) Date: Fri Nov 17 01:53:27 PST 1995 Subject: Multi >> I have seen a demo of the Green Hills C compiler and >> their MULTI software development environment for a >> VxWorks target (from an AIX host). It looked to be quite good. >> >> Does Green Hills build a good reliable compiler? >> Is their support good? >> >> Any comments, negative or positive are welcome. >> >> gregb@gemini.den.mmc.com > >-- >> >Hello, I will try "Green Hills C compiler : MULTI" for AXP on PC/Windows3.1. >Soon, I think that send the report about "MULTI". >Bye. We have been using MULTI on a 16MB 486 PC using Windows 3.11 for over a year. Its horrible. Its an X-Windows program ported to the PC Windows OS. It has many 'quirks' and it takes a huge amount of memory for what it actually does (It needs 16MB to run at anything other than a snail's pace). Its editor is just about useable but it does not use any of the PC Windows conventions. We use MULTI mainly for building. We do not use its Revision Control System. I would summarise by saying MULTI is adequate, .... just! ================================================================== Kelvin Martin, Senior Software Engineer, Computing Systems & Technology Department, British Aerospace Defence Limited (Dynamics Division), PB 230, Six Hills Way, Stevenage, Herts. SG1 2DA Phone: 01438 754306 Fax: 01438 756293 Email: kelvin.martin@def.bae.co.uk ================================================================== From daemon@csg.lbl.gov Fri Nov 17 04:00:23 1995 From: daemon@csg.lbl.gov Date: Fri Nov 17 04:00:26 PST 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Fri Nov 17 04:00:19 PST 1995 Subject: Re: Green Hills Compilers/Tools - Good/Bad ? Subject: copy_to_mbufs routine Subject: Updated FAQ Subject: Re: PPP VS. SLIP on Solaris2.x Subject: US-CA-Los Angeles - Robotics R&D @ JPL Subject: VxWindows ------------------------------------------------------- Newsgroups: comp.realtime,comp.os.vxworks,comp.unix.aix Subject: Re: Green Hills Compilers/Tools - Good/Bad ? Date: Wed, 15 Nov 1995 08:55:46 GMT From: 100072.1646@compuserve.com (Andy Bridge) Organization: APV Engineering Message-ID: <48c9ud$n6p@dub-news-svc-6.compuserve.com> References: <48634d$n5m@news2.iij.ad.jp> gregb@gemini.den.mmc.com wrote: >> I have seen a demo of the Green Hills C compiler and >> their MULTI software development environment for a >> VxWorks target (from an AIX host). It looked to be quite good. >> >> Does Green Hills build a good reliable compiler? >> Is their support good? >> >> Any comments, negative or positive are welcome. >> >> gregb@gemini.den.mmc.com >> -- >> We have had a copy of Multi for 3 months running under AIX and have never been able to get the debugger to debug at C source code level. We have complained to our supplier and as yet have no solution. Further more our licence timed out after 3 months for some inexplicable reason so we can't use it at all now. I had assumed that maybe the RS6000 was a new port and that was why we were having problems but since you say you have seen a demo running properly maybe the problems we have are fixed. On the whole MULTI looks a good package but we've never been able to use it in anger. I'd be interested to know if anybody else is running MULTI on AIX and if so what version of MULTI? Regards, Andy Bridge APV Engineering PO Box 4 Gatwick Road Crawley West Sussex RH10 2QB Tel: +44 1293 527777 Fax: +44 1293 552640 Email: 100072.1646@compuserve.com --------------------------- Newsgroups: comp.os.vxworks Subject: copy_to_mbufs routine Date: 15 Nov 1995 16:29:52 GMT From: akr@acsu.buffalo.edu (Ajit K Ranganathan) Organization: UB Message-ID: <48d4i0$hae@azure.acsu.buffalo.edu> the copy_to_mbufs routine in vxworks 5.1 takes 4 arguments. copy_to_mbufs(buf, totlen, off, ifp) i would like to know the use of the last argument ifp. i guess it is the interface pointer. supposing i make this a NULL pointer, will this routine work just to copy data from the buffer buf to a new mbuf chain. will it work? thanks, --------------------------- Newsgroups: comp.os.vxworks Subject: Updated FAQ Date: Wed, 15 Nov 95 16:33:30 PDT From: sridhar@fcseng.ip.portal.com Organization: Portal Communications (service) Message-ID: Can somebody let me know where I can get the updated FAQ for this group? I have the FAQ dated 1993 and need updated info on some FTP sites from where I can download additional information. For example, the FAQ I have states that I can obtain a list of BSPs supported from ftp.netcom.com under the /pub/hjb/ directory but I cannot find a directory with that name when I use anon FTP to that site. Thanks in advance. Sridhar. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: PPP VS. SLIP on Solaris2.x Date: Wed, 15 Nov 1995 20:13:41 GMT From: dab@netcom.com (Don Brooks) Organization: NETCOM On-line Communication Services (408 261-4700 guest) Message-ID: References: <48b4af$4qa@news.cp10.es.xerox.com> Sender: dab@netcom14.netcom.com Gary Lin (glin) wrote: : Hi, : Does anyone in this group configure a PPP connection from Sun host : to embedded board thru serial port successfully ? : We have a Sun host running SunOS4.1.3 , and have been using the SLIP : to talk to our embedded board thru the serial port , no problem . : : Now we want to move to solaris2.x and my questions are: : : 1. Is there a SLIP module for solaris2.x we can use ? : 2. If not, has anyone tried PPP connection to their embedded board : thru serial port and succeeded. We tried PPP on solaris2.x and it : did not work. Most of the manual on PPP are modem related stuff. : Any help would be appreciated : Gary I've been able to establish a PPP connection between vxWorks 5.1.1 and a Sun workstation running 4.1.3_U1 by using ppp-2.1.2 on the Sun and a public domain port of ppp-2.1.2 for vxWorks. The port is available via anonymous ftp from ftp.netcom.com as /pub/da/dab/ppp-2.1.2.tar.gz. Don Brooks dab@netcom.com --------------------------- Newsgroups: comp.os.vxworks Subject: US-CA-Los Angeles - Robotics R&D @ JPL Date: 16 Nov 1995 22:19:12 GMT From: Bruce Bon Organization: JPL, Automation and Control Section Message-ID: <48gdd0$jp3@netline-fddi.jpl.nasa.gov> ROBOTICS JOB OPPORTUNITY 11/16/95 Required: - advanced degree in engineering or CS with robotics emphasis - strong C/Unix programming background - hands-on experience in robotics - 3-D graphics/real-time experience - SEE BELOW Robotics programs at the Jet Propulsion Laboratory, California Institute of Technology, are immediately seeking self-motivated individuals, capable of first-rate research and development in several areas of robotics. Opportunities exist to participate in funded multiyear research projects in space robotics: autonomous rovers for planetary surface exploration; free-flying telerobots for servicing missions; space station robotics; and terrestrial applications of space robotics. Examples of relevant technologies for these projects are: machine vision and perception for autonomous navigation and automated inspection in space; dexterous manipulator control for free-flying robots; modeling of complex space robots in space; and flying robots in planetary atmospheres. Opportunities exist for initiating new areas of R & D in applied robotics, with emphasis on space applications. Candidates should be strongly motivated to contribute innovative concepts to robotics problems in space. They should also be very articulate with strong skills in written and oral communications. Expertise required in the fundamental disciplines of robotics, including kinematics, dynamics, sensing, and control, supported by demonstrated research record within these areas. Strong background required in C and Unix programming; hands-on experience in robotics laboratory experimentation and hardware; familiarity with 3-D graphics programming and real-time systems. Graduate studies are necessary, preferrably leading to the Ph. D. degree. For further information on JPL robotics research, check out: http://robotics.jpl.nasa.gov/ Please send your resume to: Homayoun.Seraji@jpl.nasa.gov Phone: (818) 354-4839. --------------------------- Newsgroups: comp.os.vxworks Subject: VxWindows Date: 17 Nov 1995 09:31:10 GMT From: rolf@davis.no (Rolf Berge) Organization: DAVIS A/S Message-ID: <48hkou$ftq@ulke.hiMolde.no> Have anyone experience with VxWindows? Can I use it as an event driven GUI? What performance can I expect on a i386 processor? Rolf Berge Email: rolf@davis.no DAVIS AS NORWAY --------------------------- End of New-News digest ********************** From pardoe@ncp.gpt.co.uk Fri Nov 17 04:32:55 1995 From: "Matthew Pardoe" Date: Fri Nov 17 04:32:57 PST 1995 Subject: Problems booting over SLIP Hi, We are running using 2 motorola 6302 boards running vxWorks 5.1 In the following configuration. The 1st board has two processors that communicate over shared memory using vxMP The first processor boots up from SCSI, the second boots ffrom shared memory. The second board is connected toprocessor 2 of the first board using SLIP. On power on the 1st board boots up, the second board starts up it's SLIP network and then waits for a period until the first board has completed it's boot and set up it's end of the SLIP link. The second board then attempts to boot over the SLIP link. This only suceeds about 50% of the time. The rest of the time the network appears to hang and the load stops. It appears as if the SLIP link is dying for some reason. I am not seeing any erros reported and so are finding it difficult to find out why it fails Has anyone had any experience of running SLIP between two vxWorks platforms, If so what problems have you had, what areas are likely to be causing the problems. ? Thanks Matt Pardoe =============================================================================== Matthew Pardoe _/_/ _/_/ _/ _/_/_/_/ email : pardoe@ncp.gpt.co.uk GPT Ltd. _/ _/ _/ _/ _/ _/ tel : +44 1202 782653 Sopers Lane _/ _/ _/ _/ _/ _/_/_/_/ fax : +44 1202 782163 Poole BH17 7EQ _/ _/ _/_/_/_/ _/ sneakernet : AFF (the dark side) England towards south end =============================================================================== From cmiller@robosun.tyndall.af.mil Fri Nov 17 07:36:53 1995 From: cmiller@robosun.tyndall.af.mil Date: Fri Nov 17 07:36:55 PST 1995 Subject: help From stan@rti.com Fri Nov 17 14:29:37 1995 From: Stan Schneider Date: Fri Nov 17 14:29:39 PST 1995 Subject: i960 BSP Hi all, I got a call from someone interested in hiring a contractor to write a BSP for a custom i960 board. If you're interested, contact john@armadasoft.com. (Posted as a favor; please reply directly) -- Stan stan@rti.com From punamia@newkla.kla.com Fri Nov 17 16:22:47 1995 From: punamia@newkla.kla.com (Manoj Punamia) Date: Fri Nov 17 16:22:49 PST 1995 Subject: Re: 3U video capture H/W Try Datacube. They have VME based board for video. I do not have their phone # . Manoj Punamia From daemon@csg.lbl.gov Sat Nov 18 04:00:35 1995 From: daemon@csg.lbl.gov Date: Sat Nov 18 04:00:54 PST 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sat Nov 18 04:00:21 PST 1995 Subject: ELF <---> ECOFF Subject: Re: Test-and-Set Instruction(s)? Subject: NFS weirdness Subject: RAD6000 experience ? Subject: Re: Installing Interrupt Handler with a pending interrupt Subject: JOB OPENING: US-CA-Los Angeles - real-time robotics programming @ JPL Subject: Re: ELF <---> ECOFF Subject: Re: Test-and-Set Instruction(s)? Subject: Deleting an unwanted directory Subject: SCSI / FDDI problems on 16Meg Force CPU-5CE and VxWorks 5.2 Subject: Installing Interrupt Handler with a pending interrupt Subject: Re: Deleting a named pipe Subject: Problem booting AXPvme 160 with VxWorks Subject: RAD6000 experience ? ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: ELF <---> ECOFF Date: Wed, 15 Nov 1995 23:24:54 GMT From: wellsk@netcom.com (Kevin Wells) Organization: NETCOM On-line Communication Services (408 261-4700 guest) Message-ID: Sender: wellsk@netcom5.netcom.com Does anyone know of a program that converts ELF-BIG object code to ECOFF format? (ld-sde doesn't seem to support it!) We are trying to load object code into a R3000 emulator that only supports ECOFF formatted files? Thanks. Kevin Wells, Lear Astronics Software Engineering (310)915-8358 --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Test-and-Set Instruction(s)? Date: Fri, 17 Nov 1995 04:15:54 GMT From: afraser@hookup.net (Andrew Fraser) Organization: HookUp Communication Corporation, Oakville, Ontario, CANADA Message-ID: <48gump$3vp@noc.tor.hookup.net> References: <9511131436.AA01564@rst.com> del@rst.com (Mark Del Giorno) wrote: >Heyyyyyyyyyyyyyyyyyyyy, > >I'm using the vxTas() instruction to set a semaphore that exists >on a VME memory board. It seems to work fine, but when I dissassemble >the code itself, it's not quite what I expected: > >-> l vxTas > _vxTas: >00a384 7000 MOVEQ #0,D0 >00a386 206f 0004 MOVEA .L (0x4,A7),A0 >00a38a 4e71 NOP >00a38c 4ad0 TAS (A0) >00a38e 4e71 NOP >00a390 6b08 BMI 0x0000a39a >00a392 4a80 TST .L D0 >00a394 6bf4 BMI 0x0000a38a >00a396 7001 MOVEQ #0x1,D0 >00a398 4e75 RTS >00a39a 4a80 TST .L D0 >00a39c 6bec BMI 0x0000a38a >00a39e 4e75 RTS > >[My target is a MVME162] > >More out of curiousity than anything else, does anyone understand >the reason for all the "fluff" around the TAS instruction? > - why the NOP's? > - why do the 2nd/3rd BMI's (which seem to serve no purpose) branch > back to do another TAS (which would do another VME bus cycle)? > The '040 has the really interesting ability to perform reads and writes in a different order than they were in the source - specifically, if the processor core is stalled waiting for data, then the read for that data may be performed ahead of a write cycle previously completed but still waiting in the "output" queue of the 040. If the addresses of that pending write and the promoted read match, then the write will be allowed to proceed first. Otherwise, you can have out of order reads and writes. More info in the 68040 data book, section 8.6, "Access Serialization and Bus Synchronization". The book says it's possible to force in-order memory accesses, by inserting a NOP which apparently freezes instruction execution until all pending cycles are done. That's probably why the compiler inserted the NOP before *and* after the TAS. I don't know what the BMI instructions do though. - -Andy - --------------------------------------------------------------------- ANDREW FRASER DY 4 Systems Inc email afraser@dy4.com Senior Hardware Designer 21 Fitzgerald Rd voice (613)596-9922x251 Product Support Group Nepean, Ontario fax (613)596-0574 (veni, vidi, fixi) Canada support support@dy4.com K2H 9J4 sales sales@dy4.com - ------------------- CUSTOMER FIRST, QUALITY ALWAYS ------------------ --------------------------- Newsgroups: comp.os.vxworks Subject: NFS weirdness Date: Fri, 17 Nov 1995 09:11:16 -0600 From: Shash Chatterjee Organization: Lockheed Martin M & S, Sunnyvale, CA Message-ID: <30ACA614.1254@lfwc.lockheed.com> Hi! Has anybody seen the following NFS problem? I'm using VxWorks/68k 5.1.1 on a Force CPU-40. The NFS server is a Sun running SunOS 4.1.1. I have /home/users/rai/data -> /vol1/users/rai/raidata (symbolic link). The directories are NFS mounted as: nfSMount("rai1","/vol1","/vol1") nfsMount("rai1","/vol1/users","/home/users") raidata and data have 0777 permission. If I 'cd "/home/users/rai/raidata"' and open a file for write, it works. But if I 'cd "/home/users/rai/data"' I cannot open a file for write, although reading a file and directory listings work fine. If I logon to the Sun using the same userid/group as defined in the BSP (NFS_USER_ID/NFS_GROUP_ID in config.h) I can write to files OK. The "user" parameter in EEPROM is also same as NFS_USER_ID. Any ideas? Thanks, Shash PS: The above NFS mounts are done because "rai:/home/users" is NFS mounted everywhere on our net, but we want different "/home/users/rai/ data" based on which host we mount "/vol1" from. - -- +-----------------------------------------------------------------+ + ChatterjeeS@lfwc.lockheed.com + + Sasvata Chatterjee | Electronic Systems Design & Integration + + "SHASH-WATA" | Lockheed Martin Tactical Aircraft Systems+ + (817) 763-1495 | Lockheed Boulevard, MZ2118 + + (817) 777-2115 (FAX) | Fort Worth, TX 76108 + +-----------------------------------------------------------------+ --------------------------- Newsgroups: comp.os.vxworks Subject: RAD6000 experience ? Date: Thu, 16 Nov 1995 23:41:01 GMT From: mmccall@netcom.com (Mark McCall) Organization: NETCOM On-line Communication Services (408 261-4700 guest) Message-ID: Sender: mmccall@netcom.netcom.com I'm working in the space technology group of Lawrence Livermore Labs in Livermore CA. We're evaluating processors to be used for a control system on a deep space probe. We may be required to use a radiation hardened processor. We're interested in Loral's RAD6000 and would like to hear from anyone who has had some experience with this machine. We would be developing code in C/C++, and would be using vxworks. Our application is very compute intensive. Could you tell me: How long you have been using the RAD6000, and have you experienced any hardware problems ? What operating system and version are you using ? Any problems with the board support package from Wind River ? How has the OS support been ? What is your development system and are you satisfied with it ? We would like to avoid using an AIX host. Any word on when Wind River will support a SPARC development environment for the RAD6000 ? Green Hills claims to have a SPARC development environment for the RAD6000 in their MULTI product, including a simulator and debugger. Any experience with this ? I'd really appreciate your input. Thanks in advance, Mark McCall (510) 423-7566 mccall1@llnl.gov mmccall@netcom.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Installing Interrupt Handler with a pending interrupt Date: Fri, 17 Nov 1995 19:16:50 -000 From: anj@ast.cam.ac.uk (Andrew Johnson) Organization: Royal Greenwich Observatory Message-ID: <48in5m$164@lyra.csx.cam.ac.uk> References: In article luna@cod.nosc.mil wrote: > Is there a way to install an interrupt handler for a particular interrupt > level while there is an interrupt pending? I notice that a call to > intConnect() just hangs. This situation comes up because we have a board > that receives external data (which is continuously being generated) and > informs the processor of this via an interrupt. Well, on power-up or > after a system reset the board will generate an interrupt before the > processor board has completed its boot and initialization sequence. Is > there a way around this problem? > > Motorola MVME167 board running VxWorks 5.2 Have you tried disabling the relevant VME interrupt level before installing the handler - ie do something like (assuming your external device uses level 5 IRQs): sysIntDisable(5); intConnect(...); sysIntEnable(5); This should prevent any problems with interrupts occurring while inside intConnect(). Note that the manual pages on intLock() imply you can't use that to do the same thing. - - Andrew - -- O o . Andrew Johnson, Royal Greenwich Observatory __\\_ Madingley Road, Cambridge, CB3 0EZ _|ooooo|_ Tel +44 (0)1223 374823 Fax 374700 ~~\_______/~~ WWW http://www.ast.cam.ac.uk/~anj/ --------------------------- Newsgroups: comp.realtime,comp.os.vxworks Subject: JOB OPENING: US-CA-Los Angeles - real-time robotics programming @ JPL Date: 17 Nov 1995 17:01:13 GMT From: Bruce Bon Organization: JPL, Automation and Control Section Message-ID: <48if4p$i86@netline-fddi.jpl.nasa.gov> REAL-TIME ROBOTICS JOB OPPORTUNITY 11/17/95 Required experience/knowledge: VxWorks, device driver code development, real-time C programming Desired experience/knowledge: RTI Control Shell development environment, C++, X Programming, Control Theory, Robotics The Rover Technology Research Program at the Jet Propulsion Laboratory, California Institute of Technology, is looking for a real-time systems programmer. This is a research program to build prototype Mars rovers and develop technology and capabilities for upcoming missions to Mars. For instance, our current efforts are in building Rocky 7, a 10kg fully autonomous, rough terrain mobile robot. Rocky 7 has a 3U VME chasis with 68040, A/D, DIO, framegrabber, peripheral LM629 motor control, 13 actuators, 6 cameras, etc. See http://robotics.jpl.nasa.gov/ for information about robotics R&D at JPL. See http://robotics.jpl.nasa.gov/tasks/rovertech/ for more information about the Rover Technology Research Program. Please send resumes to: Dr. Homayoun Seraji Mail Stop 198-219 Jet Propulsion Laboratory 4800 Oak Grove Drive Pasadena CA 91109 seraji@jpl.nasa.gov --------------------------- Newsgroups: comp.os.vxworks Subject: Re: ELF <---> ECOFF Date: 16 Nov 1995 20:10:58 GMT From: dyer@alx.sticomet.com (Doug Dyer) Organization: Naval Research Laboratory Message-ID: <48g5si$scu@ra.nrl.navy.mil> References: wellsk@netcom.com (Kevin Wells) writes: >Does anyone know of a program that converts ELF-BIG object >code to ECOFF format? (ld-sde doesn't seem to support it!) >We are trying to load object code into a R3000 emulator >that only supports ECOFF formatted files? >Thanks. >Kevin Wells, Lear Astronics Software Engineering >(310)915-8358 There is an easier way if you have the source. Use the "-coff" compiler option (atleast for SGIs) and you should get a coff rather than elf. Have fun :) Doug - -- Doug Dyer - dyer@alx.sticomet.com | *play Mail Signature Challenge* Software Technology, Inc. (STI) | click mouse in areas for points: DC office: (703) 329-9707 | [10 points][20 points][5 points] --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Test-and-Set Instruction(s)? Date: 14 Nov 1995 10:02:25 GMT From: kadionik@ranko (KADIONIK Patrice) Organization: ENSERB, Bordeaux, France Message-ID: <489pfh$ah8@serveur.cribx1.u-bordeaux.fr> References: <9511131436.AA01564@rst.com> Mark Del Giorno (del@rst.com) wrote: : Heyyyyyyyyyyyyyyyyyyyy, : I'm using the vxTas() instruction to set a semaphore that exists : on a VME memory board. It seems to work fine, but when I dissassemble : the code itself, it's not quite what I expected: : -> l vxTas : _vxTas: : 00a384 7000 MOVEQ #0,D0 : 00a386 206f 0004 MOVEA .L (0x4,A7),A0 : 00a38a 4e71 NOP : 00a38c 4ad0 TAS (A0) : 00a38e 4e71 NOP : 00a390 6b08 BMI 0x0000a39a : 00a392 4a80 TST .L D0 : 00a394 6bf4 BMI 0x0000a38a : 00a396 7001 MOVEQ #0x1,D0 : 00a398 4e75 RTS : 00a39a 4a80 TST .L D0 : 00a39c 6bec BMI 0x0000a38a : 00a39e 4e75 RTS : [My target is a MVME162] : More out of curiousity than anything else, does anyone understand : the reason for all the "fluff" around the TAS instruction? : - why the NOP's? : - why do the 2nd/3rd BMI's (which seem to serve no purpose) branch : back to do another TAS (which would do another VME bus cycle)? Hi; The NOP operation must not be understood as a null operation. It intends to get access to the local bus . You must have the bus acces before you do your RMW cycle with the TAS assembly instruction. By memory, there 's just one or two lines in the m68k (68020) user's manual on this subject. There 's a BMI to test gained bus access (after the NOP opcode (its seems to affect the CCR register !)) and BMI to test D0 register after the TAS opcode. There were also a posts on NOP role in this newsgroup or in the comp.sys.m68k in the past weeks... Regards; Pat. Patrice Kadionik Laboratory of Microelectronics University of Bordeaux - FRANCE --------------------------- Newsgroups: comp.os.vxworks Subject: Deleting an unwanted directory Date: 18 Nov 1995 00:45:17 GMT From: ernest@pongo.sd.gat.com Organization: General Atomics - ASI Message-ID: <48jaat$e7t@rosebud.sdsc.edu> Reply-To: ernest@pongo.sd.gat.com Running VxWorks 5.1.1, I'm trying to delete a directory that contains invalid files (i.e. "##################", "ld < utils.o", ...) that seemed to have originated from an ascii file during an ftp session. Since VxWorks doesn't use wildcards, and I can't delete these illegal files individually, and I really don't want to resort to reformatting the disk, I'm stuck. Also, rmdir("string") doesn't delete a directory that isn't empty. If you have any thoughts send me mail at ernest@pongo.sd.gat.com --------------------------- Newsgroups: comp.os.vxworks Subject: SCSI / FDDI problems on 16Meg Force CPU-5CE and VxWorks 5.2 Date: 17 Nov 1995 21:25:59 GMT From: getz@getz.manassas.ibm.com Organization: Loral Federal Systems, Manassas Message-ID: <48iul7$ae2@lfsserv1.fsc.ibm.com> Reply-To: getz@getz.manassas.ibm.com We obtained VxWorks 5.2 and the Force CPU-5CE shortly after VxWorks 5.2 was released. From the start, we had SCSI disk problems. Attempting to create a dos filesystem on the disk would cause a kernel WorkQ panic. In order to avoid this problem, we had to run off SCSI_DMA. This made our SCSI disk slower than a PC floppy disk drive. WRS tech support worked on the SCSI disk problem for us but never fully resolved it. We've been having mysterious problems with a new version of the Interphase 5211 FDDI card also. Older firmware versions worked fine on our system, but the A07 firmware boards would not boot properly. Sparta worked on this problem for us and they determined that it was caused by a problem in the BSP. Force recently provided us with the beta version of their CPU-5CE Board Support Package. There was one small glitch with the BSP that I was able to fix myself. Talk about killing two birds with one stone.... this solved both problems. Fred Getz L.F.S. Manassas --------------------------- Newsgroups: comp.os.vxworks Subject: Installing Interrupt Handler with a pending interrupt Date: Thu, 16 Nov 1995 19:19:38 GMT From: luna@cod.nosc.mil (Robert E. Luna) Organization: NCCOSC RDTE DIVISION Keywords: interrupts Message-ID: Reply-To: luna@cod.nosc.mil Sender: news@nosc.mil Is there a way to install an interrupt handler for a particular interrupt level while there is an interrupt pending? I notice that a call to intConnect() just hangs. This situation comes up because we have a board that receives external data (which is continuously being generated) and informs the processor of this via an interrupt. Well, on power-up or after a system reset the board will generate an interrupt before the processor board has completed its boot and initialization sequence. Is there a way around this problem? Motorola MVME167 board running VxWorks 5.2 Thanks, Robert E. Luna luna@cod.nosc.mil --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Deleting a named pipe Date: 17 Nov 1995 06:25:23 GMT From: mcs@goblin.caltech.edu (Martin Shepherd) Organization: California Institute of Technology. Message-ID: References: <486nt6$d3g@news.ccit.arizona.edu> In article <486nt6$d3g@news.ccit.arizona.edu> John Hughes wrote: >Does anyone happen to know how to remove a named pipe in the event that an >error occurs and it is not feasible to open it? Example: pipe is created with >a call to pipeDevCreate(), but some error condition prevents it from being >opened for subsequent use. What happens if the pipeDevCreate() call is done >again using the same name? Since VxWorks pipes are a pseudo-device built on >the message queue facility, it doesn't look like the iosDevDelete() will work >to rid it from the device descriptor tables. In my application I may need to >be able to reuse a pipe name, since it is being determined a function which >would have no knowledge of the existance of a pipe prior to its request. Of >course, I could assume that I'll never have an error condition, I suppose... >;) >Any suggestions/hints/tips cheerfully accepted. I would also like some clarification on this. When I ran into the same quandary a few months ago, I wrote a simple utility function as follows: #include void util_delete_pipe(char *name) { char *dummy; DEV_HDR *pipe_dev = iosDevFind(name, &dummy); if(pipe_dev) iosDevDelete(pipe_dev); return; } This successfully removes the device entry of the pipe from the device table (as verified by use of iosDevShow), and one can thereafter open a new pipe with the same name, but unfortunately it leaves 212 bytes of memory unreclaimed (under VxWorks 5.1). My suspicion is that the device that the above function deletes is the underlying message queue of the pipe and that deletion of this doesn't reclaim the resources that were used to abstract the message queue as a pipe. So what is the correct way to delete a pipe? Martin Shepherd (mcs@astro.caltech.edu) --------------------------- Newsgroups: comp.os.vxworks Subject: Problem booting AXPvme 160 with VxWorks Date: Fri, 17 Nov 1995 10:31:22 GMT From: FRED ZWARTS Organization: Rekencentrum der Rijksuniversiteit Groningen Message-ID: Sender: root@rug.nl (Operator) A few months ago a decision was made that we would base our real-time enviroment on AXPvme boards running VxWorks. Because of that we installed Digital-Unix (OSF/1) and VXworks on a DEC 3000 Model 300 system. Recently we also bought an AXPvme 160 SBC. To make a start, I'm trying the example in chapter 5 "Introduction to VXworks", but this gives some problems. I've configured de Unix server for bootp and tftp. I made a shadow directory tree in /users/zwarts/demo. In the file /users/zwarts/demo/config/all/configAll.h, I've brought the #define INCLUDE_RDB outside the #if FALSE block (this does not make any difference by the way). I've ran "make" in the /users/zwarts/demo/config/AXPvme directory, which created a new "system" file. With this I tried to boot the AXPvme. In first instance all seems to work as expected. Bootp shows the correct IP adresses for server and target. Tftp gets the VXworks network loader. The VXworks network loader seems to read /etc/vxbootlines because it starts to load /users/zwarts/demo/config/AXPvme/system. At last I receive 2 lines saying: netload: calling vxentry Setting up vme base and size from config.h Then things start to go wrong in 2 different ways. One time the system continously prints Unexpected interrupt, adp vector = 0xc00000 Only way to escape this is to halt or reset the system. Another time the AXPvme seems to be locked up. Even a halt doen't work. A reset is necessary to get the AXPvme back responding. When I boot with the option -fl v, kdebug get's send down by tftp. The system quiets after the message netload: calling kdbentry In this case I'm able to get out with a halt or reset. I noticed that in /var/adm/syslog.dated.dd-mmm-hh:mm/daemon.log all files send over are mentioned twice. One time with "sending file" the other time with "finished sending". Only /users/zwarts/demo/config/AXPvme/system is not mentioned with "finished sending". What could be wrong ? We are running Digital-Unix V3.2c, VXworks V3.1A. The AXPvme has SRM console V14.0-0, VMS palcode X5.48-107, OSF palcode X1.35-76, 32MB memory and VIC64 enabled. Another strange thing is that the AXPvme (in console mode) knows the command vme_setup. This command is not documented and the system continuosly prints machine checks when using this command. Should I worry about this? Fred Zwarts KVI Internet: F.Zwarts@KVI.nl Phone: (+31)50-3633619 Telefax: (+31)50-3634003 X400: C=nl;ADMD=400net;PRMD=surf;O=KVI;S=Zwarts;G=Fred Zernikelaan 25, 9747 AA Groningen, The Netherlands --------------------------- Newsgroups: comp.os.vxworks Subject: RAD6000 experience ? Date: Thu, 16 Nov 1995 23:46:06 GMT From: mmccall@netcom.com (Mark McCall) Organization: NETCOM On-line Communication Services (408 261-4700 guest) Message-ID: Sender: mmccall@netcom.netcom.com I'm working in the space technology group of Lawrence Livermore Labs in Livermore CA. We're evaluating processors to be used for a control system on a deep space probe. We may be required to use a radiation hardened processor. We're interested in Loral's RAD6000 and would like to hear from anyone who has had some experience with this machine. We would be developing code in C/C++, and would be using vxworks. Our application is very compute intensive. Could you tell me: How long you have been using the RAD6000, and have you experienced any hardware problems ? What operating system and version are you using ? Any problems with the board support package from Wind River ? How has the OS support been ? What is your development system and are you satisfied with it ? We would like to avoid using an AIX host. Any word on when Wind River will support a SPARC development environment for the RAD6000 ? Green Hills claims to have a SPARC development environment for the RAD6000 in their MULTI product, including a simulator and debugger. Any experience with this ? I'd really appreciate your input. Thanks in advance, Mark McCall (510) 423-7566 mccall1@llnl.gov mmccall@netcom.com --------------------------- End of New-News digest ********************** From daemon@csg.lbl.gov Sun Nov 19 04:00:20 1995 From: daemon@csg.lbl.gov Date: Sun Nov 19 04:00:25 PST 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sun Nov 19 04:00:17 PST 1995 Subject: Re: JOB OPENING: US-CA-Los Angeles - real-time robotics programming @ JPL Subject: x does not reboot Subject: Re: Multiport VME SCC board wanted. Subject: Re: ELF <---> ECOFF ------------------------------------------------------- Newsgroups: comp.realtime,comp.os.vxworks Subject: Re: JOB OPENING: US-CA-Los Angeles - real-time robotics programming @ JPL Date: Sat, 18 Nov 1995 08:26:42 UNDEFINED From: efa@teleport.com (Edwin Armstrong) Organization: Salem House Message-ID: References: <48if4p$i86@netline-fddi.jpl.nasa.gov> >See http://robotics.jpl.nasa.gov/ for information about robotics R&D >at JPL. See http://robotics.jpl.nasa.gov/tasks/rovertech/ for more >information about the Rover Technology Research Program. >Please send resumes to: > Dr. Homayoun Seraji > Mail Stop 198-219 > Jet Propulsion Laboratory > 4800 Oak Grove Drive > Pasadena CA 91109 > seraji@jpl.nasa.gov I worked at the Jet Propulsion Laboratory for 17 years. It is the best company that I have ever worked at. The people are great and the atmosphere for getting things done is wonderful. If you are thinking about this job, do yourself a favor and don't pass this one up - really :-) Cheers, Edwin Armstrong ... --------------------------- Newsgroups: comp.os.vxworks Subject: x does not reboot Date: 18 Nov 1995 02:38:18 GMT From: duane6@ix.netcom.com (Duane C Austin ) Organization: Netcom Message-ID: <48jguq$g9t@ixnews6.ix.netcom.com> I can't get x to cause a reboot with my target FORCE10 with 64MB of memory. However x works fine with another FORCE10 card with 32MB of memory. Anyone seen this problem or have any ideas? Thanks to all that respond. Duane Austin email - duane6@ix.netcom.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Multiport VME SCC board wanted. Date: 17 Nov 1995 13:32:52 -0500 From: accesstec@aol.com (AccessTec) Organization: America Online, Inc. (1-800-827-6364) Message-ID: <48ikgk$caa@newsbf02.news.aol.com> References: <47lii7$qpp@homer.alpha.net> Sender: root@newsbf02.news.aol.com You should also look at Performance Technologies for the PT-VME-330A. There are options for RAM between 1MB and 4MB on-board. Even better for sync serial is the PT-VME-340, which is a high speed sync serial WAN controller. It uses these little Line Adapter Boardlets which allows you to mix two separate two-channel interfaces on a single host board. There is a version for SBus and desk-top PCI and probably a PMC version as well. These guys are very competitively priced and the tech support is outstanding, with software support thru a sister company called UconX. You may see these boards sold under a third party label who specializes in communication products - a good testimonial to the products success. Good Luck! Russ Robertson AccessTec --------------------------- Newsgroups: comp.os.vxworks Subject: Re: ELF <---> ECOFF Date: 19 Nov 1995 02:44:06 GMT From: chris@news.algor.co.uk (Chris Dearman) Organization: Algorithmics Ltd. Message-ID: <48m5lm$n39@moorgate.algor.co.uk> References: Kevin Wells (wellsk@netcom.com) wrote: : Does anyone know of a program that converts ELF-BIG object : code to ECOFF format? (ld-sde doesn't seem to support it!) I wrote a program that generates an ECOFF file from an ELF executable, but you can convince the linker to generate an ECOFF file directly. When performing the final link, specify the option -oformat ecoff-bigmips Let me know how you get on. Chris - -- Chris Dearman, Algorithmics Ltd, 3 Drayton Park, London, N5 1NU, England Phone: +44 171 700 3301 FAX: +44 171 700 3400 Email: chris@algor.co.uk --------------------------- End of New-News digest ********************** From sergi@bpo.hp.com Sun Nov 19 22:52:50 1995 From: Sergi Casas Date: Sun Nov 19 22:52:52 PST 1995 Subject: Re: NFS weirdness Sasvata Chatterjee wrote: > Has anybody seen the following NFS problem? > > I'm using VxWorks/68k 5.1.1 on a Force CPU-40. The NFS server > is a Sun running SunOS 4.1.1. > > I have /home/users/rai/data -> /vol1/users/rai/raidata (symbolic link). > > The directories are NFS mounted as: > nfSMount("rai1","/vol1","/vol1") > nfsMount("rai1","/vol1/users","/home/users") > > raidata and data have 0777 permission. > > If I 'cd "/home/users/rai/raidata"' and open a file for write, it works. > But if I 'cd "/home/users/rai/data"' I cannot open a file for write, > although reading a file and directory listings work fine. > > If I logon to the Sun using the same userid/group as defined > in the BSP (NFS_USER_ID/NFS_GROUP_ID in config.h) I can write to > files OK. The "user" parameter in EEPROM is also same as NFS_USER_ID. > > Any ideas? > I guess you are having trouble with the interesting limitation of NFS of not being able to jump more than one machine to access an NFS mounted file. This means that any NFS-mounted directory in vxWorks must reside in the machine you give as NFS server, and should NOT be a link or contain links to other NFS-mounted files in other NFS servers, as vxWorks would have to jump twice to access those files. Well, of course, the directory may contain links if you don't intend to acces them from vxWorks. Hope this helps. -- |============================================================================| | Sergi Casas ##### / ##### Firmware Engineer (R&D) | | Hewlett-Packard Company ### /_ _ ### | | Barcelona Division (BCD) ## / / / / ## e-mail: sergi@hp-bpo.bpo.hp.com | | Avda. Graells, 501 ## / / /_/ ## Phone : +34 3 582 1475 | | 08190 S. Cugat (Barcelona) ### / ### Fax : +34 3 582 1410 | | SPAIN ##### / ##### HP-Telnet: 712 1475 | |============================================================================| From daemon@csg.lbl.gov Mon Nov 20 04:00:25 1995 From: daemon@csg.lbl.gov Date: Mon Nov 20 04:00:30 PST 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Mon Nov 20 04:00:21 PST 1995 Subject: Re: NFS weirdness Subject: copying from a structure to an mbuf Subject: re: 3U video capture H/W ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: NFS weirdness Date: 19 Nov 1995 11:29:42 -0800 From: mwette@mr-ed.jpl.nasa.gov (Matt Wette) Organization: Jet Propulsion Laboratory Message-ID: <48o0j6$b87@mr-ed.jpl.nasa.gov> References: <30ACA614.1254@lfwc.lockheed.com> In article <30ACA614.1254@lfwc.lockheed.com>, Shash Chatterjee wrote: >Hi! > >Has anybody seen the following NFS problem? >I have /home/users/rai/data -> /vol1/users/rai/raidata (symbolic link). > >The directories are NFS mounted as: > nfSMount("rai1","/vol1","/vol1") > nfsMount("rai1","/vol1/users","/home/users") > >raidata and data have 0777 permission. > >If I 'cd "/home/users/rai/raidata"' and open a file for write, it works. >But if I 'cd "/home/users/rai/data"' I cannot open a file for write, >although reading a file and directory listings work fine. We've experienced similar problems. We mount a directory of all user home directories a symbolic links to other mounted filesystems. They don't work. This forced users to all mount their home directories by hand. Matt - -- matthew.r.wette@jpl.nasa.gov --------------------------- Newsgroups: comp.os.vxworks Subject: copying from a structure to an mbuf Date: 18 Nov 1995 16:27:00 GMT From: akr@acsu.buffalo.edu (Ajit K Ranganathan) Organization: UB Message-ID: <48l1gk$a4j@azure.acsu.buffalo.edu> i want to copy a structure into an mbuf. this is what i am doing. copy_struct_mbuf( struct mbuf **mp, struct sockaddr_iso *buf, int buflen, int type) { m = m_get(M_WAIT, type); if ( m == NULL) return (ENOBUFS); m->m_len = buflen; m->m_off=0; m->m_ctype = MC_NO_CLUSTER; m->m_act = NULL; bcopy(buf, mtod(m, struct sockaddr_iso *),(u_int) buflen); } something seems to go wrong in bcopy. the mbuf data area does not contain what it has to.. is it bcos of some memory alignment problem.. the structure i want to copy is struct sockaddr_iso { u_char siso_len; u_char siso_family; u_int siso_tlen; struct iso_addr siso_addr; }; struct iso_addr { u_char isoa_len; char isoa_genaddr[20]; }; any help is appreciated, thanks --------------------------- Newsgroups: comp.os.vxworks Subject: re: 3U video capture H/W Date: 20 Nov 1995 12:59:38 +1030 From: wes@spock.atlantek.com.au (Wesley Hosking) Organization: Atlantek Microsystem Pty. Ltd, Adelaide, Australia. Message-ID: References: <9511180020.AA20964@newkla.kla.com> Sender: wes@spock.atlantek.com.au >>>>> "Manoj" == Manoj Punamia writes: In article <9511180020.AA20964@newkla.kla.com> punamia@newkla.kla.com (Manoj Punamia) writes: Manoj> Try Datacube. They have VME based board for video. I do not have Manoj> their phone # . Manoj> Manoj Punamia Their phone number is : DATACUBE, INC 0011 1 508 777 4200 300 Rosewood Drive, Danvers, MA, 01923, USA. . I think that have email at sales@datacube.com However - I think they only make 6U VME equipment (from the last sales blurb that I have seen). wes - -- - ------------------------------------------------------------------------------ Wesley Hosking Atlantek Microsystems Endeavour House, Technology Park, South Australia wes@atlantek.com.au +61-8-260-8990 --------------------------- End of New-News digest ********************** From bqv@se19.wg2.waii.com Mon Nov 20 07:32:47 1995 From: Bang Vu Date: Mon Nov 20 07:32:49 PST 1995 Subject: Re: comp.os.vxworks newsdigest duane6@ix.netcom.com (Duane C Austin ) penned: > >I can't get x to cause a reboot with my target FORCE10 with 64MB >of memory. However x works fine with another FORCE10 card with >32MB of memory. > >Anyone seen this problem or have any ideas? > >Thanks to all that respond. > >Duane Austin > >email - duane6@ix.netcom.com > The trap-to-monitor character defaults to ^X. However, it can be changed by the command tyMonitorTrapSet(). Somebody has used this command to force your target to use another character for reboot? Look for tyMonitorTrapSet in some of your init script or you can use it to revert back to ^X. -- Bang Vu | QC Tools, Inc. (a subsidiary of I/O, Inc.) vu@wg2.waii.com | Software, Systems, and Consultancy for the Voice 713-964-6192 | Exploration Industry. Fax 713-964-6218 | 3600 Briarpark, Houston, TX 77042 From jjs@mclean.sparta.com Mon Nov 20 09:00:00 1995 From: jjs@mclean.sparta.com (John J. Smith) Date: Mon Nov 20 09:00:03 PST 1995 Subject: promiscuous anyone? Hi all, Has anyone out there been able to get the ei driver (Intel 82596) to go into promiscuous mode? I've been trying to do this on a Heurikon V3D running vxWorks 5.1.1 without much success. I've tried the obvious things like using ifFlagsSet(), but still, no luck. I would be grateful for any assistance. Thanks in advance, _________________________________________________________ | |\ |John J Smith | Voice -> 703 448 1683 x253 | | |jjs@mclean.sparta.com | FAX -> 703 734 3323 | | |_________________________________________________________| | | | | | /\\ 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 daemon@csg.lbl.gov Tue Nov 21 04:00:20 1995 From: daemon@csg.lbl.gov Date: Tue Nov 21 04:00:24 PST 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Tue Nov 21 04:00:15 PST 1995 Subject: Help! bootup prb. Subject: Unable to open an file Subject: I have more MVME167 boards available ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Help! bootup prb. Date: 20 Nov 1995 19:41:10 GMT From: leticia.montanez@jpl.nasa.gov (Leticia Montanez) Organization: JPL Message-ID: I was wondering if anyone has experienced the following problem: upon booting up via serial (SunOS 4.1.1, Vxworks 5.1) when loading a file I get the following error undefined symbol: _tioDa ld error:error reading file (errno = 0x1c0001). value = -1 = 0xffffffff Now the undefined symbol is different at times when this error is issued when ever it appears. The only thing that is wierd is that the error message prints the symbol with a capitial letter in it when it really doesn't have one. I thought that it was a bug in the way I load but it also happens to Vxworks predefined symbols. Could this be the serial bug? Any hints. I have found that if the Heurikon is rebooted, it loads and boots up correctly. --------------------------- Newsgroups: comp.os.vxworks Subject: Unable to open an file Date: 20 Nov 1995 21:42:28 GMT From: walid@minerva.cis.yale.edu (Walid Majid) Organization: Yale University Message-ID: <48qso4$sbh@news.ycc.yale.edu> i am unable to open any files on the host (alpha osf/1) from our target board (running vxworks 5.1.1). i can login on the host, run pwd, ls and cd commands. all the files appear to be visible to the target, however, i am not able to download or run any file (scripts) from the host. the message i get is: Can't open input 'fn' S_netDrv_UNIX_FILE_ERROR at first i thought the protections were probably not set correctly, but that wasn't the case, since i allowed all kind of access to some files, but still was not able to download them. any ideas, anyone? thanks in advance. walid --------------------------- Newsgroups: comp.os.vxworks Subject: I have more MVME167 boards available Date: 21 Nov 1995 05:25:19 GMT From: innov8@ix.netcom.com (M.W. ) Organization: Netcom Message-ID: <48rnrv$f5u@ixnews5.ix.netcom.com> I have some more MVME167 boards available. E-mail if interested. --------------------------- End of New-News digest ********************** From jhillman@wrs.com Tue Nov 21 06:39:46 1995 From: jhillman@wrs.com (Jon Hillman) Date: Tue Nov 21 06:39:48 PST 1995 Subject: Re: promiscuous anyone? John J Smith writes: > > > Hi all, > > Has anyone out there been able to get the ei driver (Intel 82596) to go > into promiscuous mode? I've been trying to do this on a Heurikon V3D running > vxWorks 5.1.1 without much success. I've tried the obvious things like using > ifFlagsSet(), but still, no luck. I would be grateful for any assistance. > > Thanks in advance, > John, The i82596 ethernet driver does promiscuous mode merely by setting a single bit (I forget which one). I can put a binary copy on the ftp site for you. I'll put it in pub/ethernet/if_eipm.o The entry point is eiattach_pm, but it cannot be included along with eiattach (the normal if_ei). I hope this helps. Jon Hillman From stan@rti.com Tue Nov 21 17:49:11 1995 From: Stan Schneider Date: Tue Nov 21 17:49:14 PST 1995 Subject: PC 104 cards with VxWorks? Hi all, Does anyone have experience with PC-104 format cards running VxWorks? Which ones? What's available? -- Stan ============================================================================= = = = = Stan Schneider = email: stan@rti.com = = Real-Time Innovations, Inc. = Phone: (408) 720-8312 = = 954 Aster, Sunnyvale, CA 94086 = Fax: (408) 720-8419 = = = = ============================================================================= From punamia@newkla.kla.com Tue Nov 21 18:18:12 1995 From: punamia@newkla.kla.com (Manoj Punamia) Date: Tue Nov 21 18:18:14 PST 1995 Subject: unsubscribe unsubscribe From leonid@rst.co.il Wed Nov 22 00:30:41 1995 From: leonid@rst.co.il (Leonid Rosenboim) Date: Wed Nov 22 00:30:44 PST 1995 Subject: Re: promiscuous anyone? > From: jjs@mclean.sparta.com (John J. Smith) > Date: Mon Nov 20 09:00:03 PST 1995 > Has anyone out there been able to get the ei driver (Intel 82596) to go > into promiscuous mode? I've been trying to do this on a Heurikon V3D running > vxWorks 5.1.1 without much success. I've tried the obvious things like using > ifFlagsSet(), but still, no luck. I would be grateful for any assistance. All promiscous mode support has been removed from VxWorks drivers with release 5.1, and new drivers do not support this mode. We have added promiscous mode to some drivers by means of modifying the source code, which turned to be quite easy to do, once you got the source. Hope this helps. ----------------------------------------------------------------------- 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 ftp://knight.trendline.co.il/pub/rst E-Mail: leonid@rst.co.il From daemon@csg.lbl.gov Wed Nov 22 04:00:25 1995 From: daemon@csg.lbl.gov Date: Wed Nov 22 04:00:28 PST 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Wed Nov 22 04:00:21 PST 1995 Subject: Re: promiscuous anyone? Subject: vw5.2 for MIPS recall? Subject: Re: VxWorks / Alpha AXP Subject: Address of "SIMPAK"? ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: promiscuous anyone? Date: 21 Nov 1995 11:42:48 GMT From: "Jeff Daly." Organization: Lockheed Sanders, Inc., Nashua, NH Message-ID: <48sdvo$kub@rapnet.sanders.lockheed.com> References: <9511201659.AA13275@mclean.sparta.com> jjs@mclean.sparta.com (John J. Smith) wrote: > >Hi all, > > Has anyone out there been able to get the ei driver (Intel 82596) to go >into promiscuous mode? I've been trying to do this on a Heurikon V3D running >vxWorks 5.1.1 without much success. I've tried the obvious things like using >ifFlagsSet(), but still, no luck. I would be grateful for any assistance. > you need to set a bit in the configuration bytes. if you have the source, if_ei.c the function is eiConfig() the line will look like: pDrvCtrl->pCfd->cfdConfig.ccByte16 = 0x00; /* promiscuous off */ you need to change it to: pDrvCtrl->pCfd->cfdConfig.ccByte16 = 0x01; /* promiscuous on */ this function only gets called once on initialization, so you need to do it this way. - -jeff --------------------------- Newsgroups: comp.os.vxworks Subject: vw5.2 for MIPS recall? Date: 21 Nov 1995 15:23:18 GMT From: "Jeff Daly." Organization: Lockheed Sanders, Inc., Nashua, NH Message-ID: <48sqt6$kub@rapnet.sanders.lockheed.com> i heard yesterday that wrs has brought vxWorks 5.2 for MIPS targets back into engineering because of some major bugs. is there anyone out there that can substantiate this, and if so what the 'bugs' are? i'm sure this would be interesting news for all MIPS vxWorks users. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks / Alpha AXP Date: 21 Nov 1995 15:36:57 GMT From: msinger@oe.fau.edu (Matthew Singer) Organization: Cybernet, Boca Raton, Florida. Message-ID: <48srmp$ijl@cybernet.cse.fau.edu> References: <48ceru$p2e$1@mhade.production.compuserve.com> In article <48ceru$p2e$1@mhade.production.compuserve.com>, David Costain <100072.3624@CompuServe.COM> writes: >We are currently using VxWorks RTV3.2 for OSF/1 Alpha running on >a DEC Alpha AXP PCI 33 embedded system. We are alarmed that even >before loading our application, the basic configuration consumes >8MBytes of RAM. > >We already ship a MC68020 version of the application, and this >has very modest memory requirements ( App + VxWorks < 1MByte). > >The Alpha CONSOLE uses seems to use somewhere between 2 and 6 >MByte (DEC in the UK seem unsure of this!). > >We are desperate to get the memory size down - can anyone help! > >Thanks in anticipation! True, the console uses a lot of ram, both at the bottom of memory and some at the top. I dont recall exactly how much. In addition, if I recall correctly, loading VxWorks into Flash makes things worse as the flash area for the boot roms gets mapped over ram. To get the most amount of ram, you want boot via bootp from the AXP console (not vxworks rom image). You also want to tune the addresses in MakeSkel. They are probably set to allow the ram that the bootrom image uses. - -- - ------------------------------------------------------------------------------- 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,comp.arch.bus.vmebus Subject: Address of "SIMPAK"? Date: 22 Nov 1995 08:11:28 GMT From: michaelv@qualcomm.com (Michael Vakulenko) Organization: Qualcomm Israel Message-ID: Hi, I heard, that there is a VME board vendor company called "SIMPAK". Does anybody know their phone, e-mail or address? Thanx in advance, ___________________________________________________________ Michael Vakulenko, : Email: michaelv@qualcomm.com Software Engineer, : Voice: +972-4-577999 QUALCOMM Israel, Ltd. : Fax: +972-4-577998 ___________________________________________________________ --------------------------- End of New-News digest ********************** From lrichard@redwood.dn.hac.com Wed Nov 22 09:55:38 1995 From: lrichard@redwood.dn.hac.com (Larry Richards) Date: Wed Nov 22 09:55:40 PST 1995 Subject: Deleting a Bogus Directory A directory on my DOS formatted disk somehow got corrupted, and now I cannot even remove it. It seems as maybe the name of the directory got changed and the new name has some unreadable characters at the end, but because vxWorks has no wild cards I do not know of a way to remove it. Any help? -------------------------------------------------------------------- Larry Richards Staff Engineer Internet: lrichard@redwood.dn.hac.com Hughes Aircraft Phone: (303)344-6630 BLDG: S75 MS: A1715 Fax: (303)344-2903 16800 E CentreTech Pkwy Aurora, CO 80011 -------------------------------------------------------------------- From ozkose@venus.aselsan.com.tr Thu Nov 23 00:25:24 1995 From: Ozgu OZKOSE ERDOGAN Date: Thu Nov 23 00:25:27 PST 1995 Subject: unsubscribe unsubscribe From daemon@csg.lbl.gov Thu Nov 23 04:00:25 1995 From: daemon@csg.lbl.gov Date: Thu Nov 23 04:00:28 PST 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Thu Nov 23 04:00:20 PST 1995 Subject: Any experience with Tornado? Subject: VxWorks FAQ Wanted Subject: Re: atexit Subject: GNU c++ Stream IO Subject: VME Addressing configuration with MVME162 & VxWorks Subject: Re: atexit ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Any experience with Tornado? Date: Wed, 22 Nov 95 12:54:56 CET From: kai@ade.no (Kai Atle Myrvang) Message-ID: <3836mmn5f@ade.no> Reply-To: kai@ade.no After looking at the glossy papers from Wind River regarding Tornado, I just wonder if there is anyone outh ther who has had the opportunity to use this environment. I'm particularly interested in Win/NT host and/or i960 target, but are not sure they are released yet. - -- * Kai Atle Myrvang (kai@ade.no) * * AD Elektronikk AS, P.O.Box 104, N-1405 Langhus, Norway * * Tel: int+47 64 86 99 70 | Fax: int+47 64 86 99 20 * ********************************************************************* --------------------------- Newsgroups: comp.os.vxworks Subject: VxWorks FAQ Wanted Date: 22 Nov 1995 15:55:30 GMT From: plward@ll.mit.edu (paula l ward) Organization: MIT Lincoln Laboratory Message-ID: <48vh5i$clt@llnews.ll.mit.edu> Sender: plw@bart (paula l ward) I am in the process of designing a system that will use vxWorks. Is there a FAQ available? advTHANKSance --------------------------- Newsgroups: comp.os.vxworks Subject: Re: atexit Date: Thu, 23 Nov 1995 03:11:02 GMT From: mfisher@dyn.com (Mark Fisher) Organization: The World @ Software Tool & Die Message-ID: References: <47o7ne$kkv@texas.nwlink.com> Sender: news@world.std.com (Mr Usenet Himself) stevecor@nwlink.com (Steve Cornett) wrote: >I am porting to VxWorks a library which uses the standard C function >atexit. Unfortunately atexit is not provided and the taskDeleteHook >facility does not work exactly like atexit. Delete hooks are called >whenever any task exits, rather than when only the caller exits. >To work around this limitation, my plan is to have the delete hook >routine only do its thing if a thread-private variable is set. I will >set the thread-private variable using taskVarAdd at the point when >atexit is normally called. That way, only the caller task will have >this variable set and the atexit handler will ignore all other tasks. >When the atexit handler does its thing, I delete the delete hook so >that subsequent task terminations do not call into the unloaded task > area. >Does anyone know a simpler way to simulate atexit in VxWorks? >-- >Steve Cornett >Bullseye Software Hi, I have been thinking about this myself lately. While I haven't yet tried this I was thinking of adding a delete hook like you sya, but instead of --------------------------- Newsgroups: comp.os.vxworks Subject: GNU c++ Stream IO Date: 23 Nov 1995 01:03:06 GMT From: Graham Waters Organization: The University of British Columbia Message-ID: <490h8a$c4k@nntp.ucs.ubc.ca> I have successfully build c++ applications using gcc2.7.0. For consol io etc I have been happy to use printf. However I would like to be able to use cout etc. So far I have been able to compile and link by including iostream.h. No errors occur but neither dous any output. I expect I have to open a stream somehow. Has anybody done this. My target is an mvme162. Thanks Graham Waters E-mail waters@triumf.ca - ---------------------------------------------------------------- Graham Waters Control Systems Engineer TRIUMF Meson Research Facility Vancouver, CANADA. - ---------------------------------------------------------------- --------------------------- Newsgroups: comp.os.vxworks,comp.arch.bus.vmebus Subject: VME Addressing configuration with MVME162 & VxWorks Date: Wed, 22 Nov 1995 21:10:06 GMT From: mdscott@novice.uwaterloo.ca (Mark David Scott) Organization: University of Waterloo Message-ID: Sender: news@novice.uwaterloo.ca (Mr. News) Hi all, I'm currently experiencing some difficulty in a VMEbus based system. The VMEbus system has traditionally been running a FORCE board baced on a Sparc processor, and is the VMEBbus master. Now, we're trying to add a MV162 as a slave, but when the 162 boots (under VxWorks), the Force can't (under UNIX) due to a memory conflict of some sort. I've tried editing the A24 and A32 address space mappings for the 162 board in config.h, but I've had to suces in getting the Force board to boot. We're reasonably sure (but not positive!) that this is a memory address conflict. If the MV162 is plugged in, but VxWorks boots, there is no conflict. (Since VxWorks confiures the 162 at boot time, this is what I'd expect). Any suggestions would be greatly appreciated! - -- - ------------------------------------------------------------------------------ Mark David Scott mdscott@novice.uwaterloo.ca 2B Computer Engineering 268 Phillip Street, Apt. Ca-31 University of Waterloo Waterloo, Ontario N2L 3G1 --------------------------- Newsgroups: comp.os.vxworks Subject: Re: atexit Date: 23 Nov 1995 03:56:34 GMT From: steved@pop3.pacifier.com (Steve Doiel) Organization: Pacifier BBS, Vancouver, Wa. ((360) 693-0325) Message-ID: <490rdi$3ma@news.pacifier.com> References: <47o7ne$kkv@texas.nwlink.com> Reply-To: steved@pacifier.com (Steve Doiel) >stevecor@nwlink.com (Steve Cornett) wrote: > >>I am porting to VxWorks a library which uses the standard C function >>atexit. Unfortunately atexit is not provided and the taskDeleteHook >>facility does not work exactly like atexit. Delete hooks are called >>whenever any task exits, rather than when only the caller exits. > >>To work around this limitation, my plan is to have the delete hook >>routine only do its thing if a thread-private variable is set. I will >>set the thread-private variable using taskVarAdd at the point when >>atexit is normally called. That way, only the caller task will have >>this variable set and the atexit handler will ignore all other tasks. >>When the atexit handler does its thing, I delete the delete hook so >>that subsequent task terminations do not call into the unloaded task >> area. > >>Does anyone know a simpler way to simulate atexit in VxWorks? >>-- >>Steve Cornett >>Bullseye Software I implemented C exception handling using setjmp/longjmp and a couple of macro's to create something similar to a C++ try/catch block. To do this I needed a task variable containing a jump buffer. I first attempted using taskVarAdd in the createTaskHook routine, but after a little frustration, decided to take a simpler approach. First I created a small function that takes a pointer to a function as it's first parameter. Then I created a CreateTask function takes the same arguments as taskSpawn, but instead of using the function passed to CreateTask, I use my small function and pass the address of the function passed into CreateTask as the first argument in the taskSpawn call. I copy the rest of the arguments to CreateTask to the call to taskSpawn, and propagate these to the original function. This permits me to execute my own code before and after the function of the task is executed. It adds a level of indirection, but is dirt simple to implement and works. Hope this helps Steve Doiel --------------------------- End of New-News digest ********************** From daemon@csg.lbl.gov Fri Nov 24 04:00:22 1995 From: daemon@csg.lbl.gov Date: Fri Nov 24 04:00:26 PST 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Fri Nov 24 04:00:17 PST 1995 Subject: Re: Any experience with Tornado? Subject: FAQ? Subject: VxWorks device driver for parallel port (MV167)? Subject: Re: vw5.2 for MIPS recall? Subject: kbhit-like routine for VxWorks? ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Any experience with Tornado? Date: Thu, 23 Nov 1995 15:03:08 GMT From: avlasblo@ms.philips.nl (Arjon Vlasblom) Organization: High Tech Automation BV. Message-ID: <1995Nov23.150308.27450@ms.philips.nl> References: <3836mmn5f@ade.no> Sender: news@ms.philips.nl Kai Atle Myrvang (kai@ade.no) wrote: :>After looking at the glossy papers from Wind River regarding :>Tornado, I just wonder if there is anyone outh ther who has :>had the opportunity to use this environment. I'm particularly :>interested in Win/NT host and/or i960 target, but are not sure :>they are released yet. I asked sales support in Europe this question and I got the following answer from them. - ----- Included Message ----- - From a technical point of view, VxWorks 5.2 for i960 is available now. I believe it would have been in beta at the time the order was received, which is why version 5.1.1 was shipped. - VxWorks 5.3 is the version of the RTOS which is included as part of our new Tornado development environment. Tornado is in beta at the moment for 68k, cpu32 and Intel 80x86 targets (386, 486, Pentium). Support for i960 is included as part of the 'second tier' release which is currently following about 3 months behind the first tier. (The main reason for releasing Tornado in tiers is to manage the workload on the QA team.) We could send you an information pack on Tornado if this would be useful. - ---------------------------- Regards, Arjon Vlasblom. //// (o o) ____________________________________oOO_( )_OOo______________________________ / \ | Arjon.Vlasblom@HTA.nl : Eindhoven : Bilthoven | | High Tech Automation BV : Voice +31 40 2447465 : Voice +31 30 2251700 | | The Netherlands : Fax +31 40 2441260 : Fax +31 30 2251717 | | | |-----------------------------------------------------------------------------| | Disclaimer: This message does not necessarily reflect the company's opinion,| | just my own humble one. | \_____________________________________________________________________________/ --------------------------- Newsgroups: comp.os.vxworks Subject: FAQ? Date: Thu, 23 Nov 1995 08:45:42 GMT From: 100072.1646@compuserve.com (Andy Bridge) Organization: APV Engineering Message-ID: <491cba$fgr@dub-news-svc-4.compuserve.com> Is there a FAQ for this group, I've been watching patiently but haven't seen any mention of it? Andy Bridge APV Engineering PO Box 4 Gatwick Road Crawley West Sussex RH10 2QB Tel: +44 1293 527777 Fax: +44 1293 552640 Email: 100072.1646@compuserve.com --------------------------- Newsgroups: comp.os.vxworks Subject: VxWorks device driver for parallel port (MV167)? Date: 23 Nov 1995 23:42:28 GMT From: glenn@tauon.ph.unimelb.edu.au (Glenn Moloney) Organization: University of Melbourne Message-ID: Reply-To: glenn@physics.unimelb.edu.au Hi all, I'm looking for a character device driver for the parallel port (on the transition module) for the MV167 board (also interested in MV147 - but less urgent). If anyone can point me to a location for such a beast please let me know. If there is none out there I'll write one (fairly simple I would expect) so any pointers on useful docs for the PCCchip parallel port registers would be greatly appreciated. All I've got is the list of register addresses and their mnemonic names. I might be able to guess the right functionality (he says hopefully ;-). cheers and thanks, glenn. - -- Glenn Moloney glenn@physics.unimelb.edu.au School of Physics, Phone: +61 3 9344 5081 University of Melbourne, Fax: +61 3 9347 4783 Parkville, Australia 3052. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: vw5.2 for MIPS recall? Date: 21 Nov 1995 17:57:09 GMT From: dyer@alx.sticomet.com (Doug Dyer) Organization: Naval Research Laboratory Message-ID: <48t3tl$j97@ra.nrl.navy.mil> References: <48sqt6$kub@rapnet.sanders.lockheed.com> "Jeff Daly." writes: >i heard yesterday that wrs has brought vxWorks 5.2 for MIPS targets back into >engineering because of some major bugs. is there anyone out there that can >substantiate this, and if so what the 'bugs' are? i'm sure this would be >interesting news for all MIPS vxWorks users. ^^^^^^^^^^^^^^^^^^^^^ Yeah, all three of us! (:-) - -- Doug Dyer - dyer@alx.sticomet.com | *play Mail Signature Challenge* Software Technology, Inc. (STI) | click mouse in areas for points: DC office: (703) 329-9707 | [10 points][20 points][5 points] --------------------------- Newsgroups: comp.os.vxworks Subject: kbhit-like routine for VxWorks? Date: Wed, 15 Nov 1995 20:58:55 GMT From: wlynde@nmsu.edu (William Lynde) Organization: New Mexico State University Message-ID: <48ddeb$7sb@bubba.NMSU.Edu> I am just getting started with VxWorks and VME in general, so please excuse any ignorance on my part. Does VxWorks have a routine like kbhit() in Borland? I would like to be able to look at keystrokes in the keyboard buffer, and I don't see a way to do this easily. Any help would be greatly appreciated. My target is a FORCE CPU-44C and my host is a Sun. Thanks. Bill Lynde --------------------------- End of New-News digest ********************** From pak@rtp.co.uk Fri Nov 24 09:25:28 1995 From: pak@rtp.co.uk.\(\).\(\)(\) (Paul Keir \(e-mail only\)) Date: Fri Nov 24 09:25:31 PST 1995 Subject: X25 for VxWorks? Hi, We have a potential application which needs to run on an MVME 162 or 167, but which requires connection to an X25 WAN. Does anybody know of any off-the-shelf solutions to this - in terms of hardware, software or both (and what physical connections are provided?) Any help/suggestions would be gratefully received. Paul. ------ | | ****** |******| ****** Paul Keir Real Time Products Ltd ** ** | ** | ** ** Unit 8, Holt Court ****** | ** | ****** e-mail: pak@rtp.co.uk Aston Science Park ** ** | ** | ** Birmingham ** ** | ** | ** tel: +44 121 333 6955 B7 4EJ ** ** | ** | ** fax: +44 121 333 5433 United Kingdom | | | -- | ------ From daemon@csg.lbl.gov Sat Nov 25 04:00:32 1995 From: daemon@csg.lbl.gov Date: Sat Nov 25 04:00:35 PST 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sat Nov 25 04:00:26 PST 1995 Subject: Re: Address of "SIMPAK"? ------------------------------------------------------- Newsgroups: comp.os.vxworks,comp.arch.bus.vmebus Subject: Re: Address of "SIMPAK"? Date: 25 Nov 1995 02:02:01 -0500 From: micrology@aol.com (Micrology) Organization: America Online, Inc. (1-800-827-6364) Message-ID: <496f19$lt3@newsbf02.news.aol.com> References: Reply-To: micrology@aol.com (Micrology) Sender: root@newsbf02.news.aol.com >I heard, that there is a VME board vendor company called "SIMPAK". >Does anybody know their phone, e-mail or address? I believe you are referring to Simpact Associates. You can phone them at 619-595-1865 or fax them at 619-292-8015. Sounds like you need a (free) subscription to VMEbus Systems Magazine. We publish an annual Buyer's Guide issue that lists the phone number and fax number for all of the VMEbus suppliers. (See the subscription form posted elsewhere on this newsgroup to apply for a free subscription.) John Black Editor-in-Chief VMEbus Systems magazine micrology@aol.com --------------------------- End of New-News digest ********************** From daemon@csg.lbl.gov Sun Nov 26 04:00:27 1995 From: daemon@csg.lbl.gov Date: Sun Nov 26 04:00:30 PST 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sun Nov 26 04:00:23 PST 1995 Subject: GPIB drivers ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: GPIB drivers Date: 24 Nov 1995 19:01:57 GMT From: isr@ix.netcom.com (Robert C. Smith ) Organization: Netcom Message-ID: <4954r5$ja6@ixnews6.ix.netcom.com> Does anyone know of shareware drivers for National Instruments or TI gpib interface chips other than the ones on the Nat. Instr. ftp site? Also, anyone have any experience implementing talker or listener drivers into the VxWorks IO system? Thanks, Bob Smith isr@ix.netcom.com --------------------------- End of New-News digest ********************** From anatolyf@telecomm.tadiran.co.il Mon Nov 27 01:05:37 1995 From: "Anatoly Feigin (1725)" Date: Mon Nov 27 01:05:39 PST 1995 Subject: Help on TELENETWORKS required Does anybody use TELENETWORK HDLC Communications packet under VxWorks ? We run TELENETWORKS v.4.02 under VxWorks 5.1.1. on i960 platform and have some problems with it. Any suggestions will be very appreciated. ============================================================ Anatoly Feigin - Software Design Engineer Igal Chernobylsky - Software Design Engineer Access Systems Division TADIRAN TELECOMMUNICATIONS Ltd. Tel.: 972-3-9261025 18 Ha-Sivim Str., p.o.b.500 Fax.: 972-3-9262567 Petah-Tikva 49104 ISRAEL e-mail: anatolyf@telecomm.tadiran.co.il igal@telecomm.tadiran.co.il ============================================================ From msk@wrsec.fr Mon Nov 27 07:53:31 1995 From: msk@wrsec.fr (Michel Smolak) Date: Mon Nov 27 07:53:33 PST 1995 Subject: ATM/VxWorks Hi, Is there somewhere available a DRIVER for VxWorks for : - The SBUS board from SUN, Ref: x10060A SUN ATM 155/MFiber SBUS Adapter 2.0. This product will be used on a THEMIS LXE board based on the MicroSPARC 2 microprocessor. Thank you for your ongoing help. Regards ***************** * Michel SMOLAK * ************************************************************* * Wind River Systems * Phone : (33) 1 69 07 78 78 * * 27, Avenue de la Baltique * Fax : (33) 1 69 07 08 26 * * 91962 Les Ulis Cedex, FRANCE * email : msk@wrsec.f * ************************************************************* From elmer!key@diasonx.com Mon Nov 27 10:56:43 1995 From: elmer!key@diasonx.com (Charles Key) Date: Mon Nov 27 10:56:46 PST 1995 Subject: Re: Deleting a Bogus Directory > >From lrichard@redwood.dn.hac.com Wed Nov 22 09:55:38 1995 >From: lrichard@redwood.dn.hac.com (Larry Richards) >Date: Wed Nov 22 09:55:40 PST 1995 >Subject: Deleting a Bogus Directory >A directory on my DOS formatted disk somehow got corrupted, and now I cannot >even remove it. It seems as maybe the name of the directory got changed >and the new name has some unreadable characters at the end, but >because vxWorks has no wild cards I do not know of a way to remove it. >Any help? > How about something like the following: #define PSEUDOCODE void findBogusDir(char * dirName) { pDir = opendir(dirname); while (pEnt = readdir(pDir)) { // decide if this is the file to delete by: // - printf'ing pDir-d_name and scanf'ing a response (y or n) // - filtering on unique substring in filename using 'strstr' // - it is because you moved all good files somewhere else if (bogusName) rm(pEnt->d_name); } closedir(pdir); } ___________________________________________________________________________ Charles Key Diasonics Ultrasound key@diasonx.com From prb@aplexus.jhuapl.edu Mon Nov 27 11:48:30 1995 From: prb@aplexus.jhuapl.edu (Paul R. Bade) Date: Mon Nov 27 11:48:34 PST 1995 Subject: ATM PMC Hi, Does anybody have a driver available for the Interphase 4515 ATM PMC (PCI Mezzanine Card)? Or does anyone know of any other ATM PMCs for which there is a vxWorks driver? I am particularly interested in a driver that has support for ATM LAN Emulation. Is there any vxWorks ATM driver that supports LAN Emulation and could be adapted to the Interphase ATM PMC? Thanks, +====================================================================+ | __ ____ __ __ | | /\ \ / \ \ /\_\__ /\ \ Johns Hopkins University | | / \_\ | /\ \ \ / / /\_\\ \ \ Applied Physics Lab. | | / /\ | |\ \/ \ \ / / / / / \ \ \ | | / \/ |_| \ /\ \_\ / / / / / \ \ \ Paul R. Bade | | / /\__/_/ \ \ \/_// / / / / / / / (301)-953-6000 x8681 | | / / / \ \_\ \ \/ / / / / / prb@aplexus.jhuapl.edu | | \/_/ \/_/ \__/_/ \/_/ | | __ ____ __ __ | | /\ \ / \ \ /\ \ /\ \ | | / \_\ | /\ \ \ / \_\ / \ \ | | / /\ | |\ \/ \ \ / /\ |_| / /\ \ \ | | / \/ |_| \ /\ \_\ / / / | | \/ / \ \ | | / /\ / / \ \ \/_// / / / / / /\ \_\ | | \ \/ / / \ \_\ \ \/ / / \/ / / / | | \__/_/ \/_/ \__/_/ /_/_/ | | | +====================================================================+ From pak@rtp.co.uk Mon Nov 27 12:45:11 1995 From: pak@rtp.co.uk.\(\).\(\)(\) (Paul Keir \(e-mail only\)) Date: Mon Nov 27 12:45:14 PST 1995 Subject: PCMCIA Hi, Does anyone have Token Ring support for VxWorks available ? We have a customer wanting support who is using a PC104 card. Any help or info would be greatly appreciated, Regards, Paul. ------ | | ****** |******| ****** Paul Keir Real Time Products Ltd ** ** | ** | ** ** Unit 8, Holt Court ****** | ** | ****** e-mail: pak@rtp.co.uk Aston Science Park ** ** | ** | ** Birmingham ** ** | ** | ** tel: +44 121 333 6955 B7 4EJ ** ** | ** | ** fax: +44 121 333 5433 United Kingdom | | | -- | ------ From harvey@wrs.com Mon Nov 27 13:37:39 1995 From: Harvey Wong Date: Mon Nov 27 13:37:41 PST 1995 Subject: CD-ROM file system Dear VxWorks Users, Has anyone implemented drivers and file system for CD-ROMs for VxWorks? Any info or leads would be appreciated. Regards, Harvey Wong Wind River Systems harvey@wrs.com From senthil@mcs.com Mon Nov 27 13:39:19 1995 From: "Senthil Natesan" Date: Mon Nov 27 13:39:21 PST 1995 Subject: reg exp support Hi all I would like to know about regular expression availablity for vxworks version 5.2. Have you used or do you know any third party who supports regex for solaris host platform (target is 386)? Thank you. Senthil Natesan (email : senthil@mcs.com ) From lrichard@redwood.dn.hac.com Mon Nov 27 13:54:54 1995 From: lrichard@redwood.dn.hac.com (Larry Richards) Date: Mon Nov 27 13:54:56 PST 1995 Subject: Re: Deleting a Bogus Directory > >From lrichard@redwood.dn.hac.com Wed Nov 22 09:55:38 1995 > >From: lrichard@redwood.dn.hac.com (Larry Richards) > >Date: Wed Nov 22 09:55:40 PST 1995 > >Subject: Deleting a Bogus Directory > >A directory on my DOS formatted disk somehow got corrupted, and now I cannot > >even remove it. It seems as maybe the name of the directory got changed > >and the new name has some unreadable characters at the end, but > >because vxWorks has no wild cards I do not know of a way to remove it. > >Any help? > > > > How about something like the following: > > #define PSEUDOCODE > void findBogusDir(char * dirName) > { > pDir = opendir(dirname); > while (pEnt = readdir(pDir)) { > // decide if this is the file to delete by: > // - printf'ing pDir-d_name and scanf'ing a response (y or n) > // - filtering on unique substring in filename using 'strstr' > // - it is because you moved all good files somewhere else > if (bogusName) > rm(pEnt->d_name); > } > closedir(pdir); > } > > ___________________________________________________________________________ > Charles Key > Diasonics Ultrasound key@diasonx.com > Good idea, but I tried a similar program and it did not work. I have learned it is more that a bogus name. As some of you have pointed out, it is a bug/deficiency with vxWorks. -------------------------------------------------------------------- Larry Richards Staff Engineer Internet: lrichard@redwood.dn.hac.com Hughes Aircraft Phone: (303)344-6630 BLDG: S75 MS: A1715 Fax: (303)344-2903 16800 E CentreTech Pkwy Aurora, CO 80011 -------------------------------------------------------------------- From banta@cebaf.gov Mon Nov 27 14:38:55 1995 From: Paul Banta Date: Mon Nov 27 14:38:57 PST 1995 Subject: Re: ATM PMC >>>>> " " == the vxWorks Users Group Exploder writes: > Submitted-by prb@aplexus.jhuapl.edu Mon Nov 27 11:48:30 1995 > Submitted-by: prb@aplexus.jhuapl.edu (Paul R. Bade) > Does anybody have a driver available for the Interphase 4515 > ATM PMC (PCI Mezzanine Card)? Or does anyone know of any other > ATM PMCs for which there is a vxWorks driver? > I am particularly interested in a driver that has support for > ATM LAN Emulation. Is there any vxWorks ATM driver that > supports LAN Emulation and could be adapted to the Interphase > ATM PMC? > Thanks, We, at CEBAF, are in the process of porting a driver for the 4515. Ultimately we want an RFC 1577 compliant driver, but first we'll have one that supports only PVCs. We bought a 4515 Driver Developer's Kit from Interphase that included source code for their Solaris 2.4 x86 version of the driver. I've just been working on it for a couple of weeks, now, and I've had a lot of learning to do about the PCI system architecture...so not much is done. (However, I'm near the top of the learning curve and I expect things to pick up soon.) I, too, would be interested if anyone has *any* ATM driver that they would be willing to share that supports LAN Emulation. One issue that I'm sure is holding up some of the development is that ATM hardware address registration (among other things) is done via SNMP and the Interim Local Management Interface (ILMI) MIB. This means that an SNMP agent that supports this stuff has to be running. There is quite a bit of work to be done before I'm finished with what we want. Even then, I'm not sure how much I could share because we had to sign a non-disclosure agreement with Interphase for their code. (Much of which is going in the garbage because it's all Solaris overhead.) -- Paul Banta voice: 804-249-5829 CEBAF, MS 12H fax: 804-249-5800 12000 Jefferson Avenue email: banta@cebaf.gov Newport News, VA 23606 From daemon@csg.lbl.gov Tue Nov 28 04:00:54 1995 From: daemon@csg.lbl.gov Date: Tue Nov 28 04:00:57 PST 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Tue Nov 28 04:00:51 PST 1995 Subject: Re: Address of "SIMPAK"? Subject: DSET GDMO Agent for vxworks Subject: HW/SW Openings @ Ariel Subject: Ethernet looping back Subject: 20meg Flash Card Subject: Epilogue Envoy SNMP Port Subject: US-CA Embedded VXWORKS Software Engineers, 2 openings (Sr & Jr) STELLCOM TECH. Subject: Problems booting off network? (Hangs after start of loading) Subject: Re: VME Addressing configuration with MVME162 & VxWorks ------------------------------------------------------- Newsgroups: comp.os.vxworks,comp.arch.bus.vmebus Subject: Re: Address of "SIMPAK"? Date: 27 Nov 1995 14:03:02 GMT From: m_thompson@ids.net (Michael Thompson) Organization: Schroff, Inc. Message-ID: <49cgem$qgk@paperboy.ids.net> References: In article , michaelv@qualcomm.com says... > >Hi, > >I heard, that there is a VME board vendor company called "SIMPAK". >Does anybody know their phone, e-mail or address? Simpact 9210 Sky Park Court San Diego,CA 92123 - -- Michael Thompson E-Mail m_thompson@ids.net Schroff, Inc. Phone 401-732-3770 Warwick, RI 02886 FAX 401-739-7599 --------------------------- Newsgroups: comp.os.vxworks Subject: DSET GDMO Agent for vxworks Date: Sun, 26 Nov 95 12:31:58 PDT From: Kevin Glavin Organization: Best Internet Communications Message-ID: Has anybody any experience of the GDMO Agent toolkit by DSET which I have seen in an announcement in the fall edition of the WindWord Magazine? I would also appreciate any contact information for DSET especially www pointers if they exist. - --Kevin. --------------------------- Newsgroups: nj.jobs,nyc.jobs,misc.jobs.offered,comp.realtime,comp.dsp,comp.arch.vmebus,comp.os.vxworks Subject: HW/SW Openings @ Ariel Date: Mon, 27 Nov 1995 14:13:20 GMT From: tje@master.ariel.com (Jeff Enderwick) Organization: Ariel Corporation, Highland Park, NJ Message-ID: Sender: news@ariel.com The following positions are open at Ariel: 1) Senior Software Engineer - experience with real-time, device-drivers and DSP. Should have outstanding software architecture, development and debug skills. Familiarity with VxWorks, pSOS, UNIX, and/or Windows NT. Must understand hardware at a low-level. 2) Senior Hardware Engineer - should have an outstanding track record of designing high-speed RISC or DSP-based multiprocessor systems. Software, ASIC and/or leadership skills a plus. VME or PCI experience also a plus. These are not entry-level positions. Both openings are in central New Jersey. Currently we're located in Highland Park, however we'll be moving to Cranbury in January. Ariel designs and manufactures high-performance multiprocessor DSP systems. Hours are flexible, and the atmosphere is fast-paced but fun. Please send resumes to: Jeff Enderwick Ariel Corp. 433 River Road Highland Park, NJ 08904 FAX: 908-249-2123 --------------------------- Newsgroups: comp.os.vxworks Subject: Ethernet looping back Date: 27 Nov 1995 17:17:09 GMT From: akr@acsu.buffalo.edu (Ajit K Ranganathan) Organization: UB Message-ID: <49crql$5ri@azure.acsu.buffalo.edu> hi, i am working on OSI TP4 over null CLNP . for testing, i need to loop back the packet to test my receiver. i can view the packets that i am transmitting using a network analyzer(sniffer). the packet has the source and destination ethernet address fields the same, but i dont seem to receive the packet. how do i go about looping the packet back to itself, since i dont have IP loop back address to use. i am using a fujitsu nice ethernet driver and need to know whether the hardware checks if the packet is for itself before transmitting, which does not seem to be the case as far as i have seen the driver output routine.. but even with this, once the packet has been transmitted, the ethernet address should be matching with the sender and should get back or not???? thanks in advance, ajit. akr@acsu.buffalo.edu --------------------------- Newsgroups: comp.os.vxworks Subject: 20meg Flash Card Date: Mon, 27 Nov 1995 17:14:22 GMT From: sycamore@resumix.portal.com Organization: Motorola, Scottsdale, AZ Message-ID: <1995Nov27.171422.10624@schbbs.mot.com> References: <490rdi$3ma@news.pacifier.com> Reply-To: sycamore@resumix.portal.com Sender: news@schbbs.mot.com (SCHBBS News Account) Our project is using VADSWorks 3.1 with VxWorks 5.2 operating system on many 68040 boards. We would like to store our executables on a 20 megabyte FLASH card and down load them when the power is supplied to the system. Has anyone written any drivers or has any information for accessing a FLASH card using VxWorks? Thanks, sycamore@andy.geg.mot.com --------------------------- Newsgroups: comp.os.vxworks Subject: Epilogue Envoy SNMP Port Date: 28 Nov 1995 00:52:29 GMT From: vineet@skorpio3.usask.ca (Vineet Chikarmane) Organization: University of Saskatchewan Message-ID: <49dmgd$gjh@tribune.usask.ca> I am interested in implementing the Epilogue Envoy SNMPv2 MIB II agent in VxWorks 5.2 Has anybody done this before? I would be gratefull for glue code examples. (Traps en Table lookups) Regards, Vineet +-----------------------------------------------------------------------+ | URL : http://www.cs.usask.ca/grads/vineet/index.html | | Comp Sc Dept, Univ of Saskatchewan & Telecommunications Research Labs | | Vox: (306)966-8654 Fax: (306)966-4884 E : vineet@cs.usask.ca | +-----------------------------------------------------------------------+ --------------------------- Newsgroups: sdnet.jobs,comp.os.vxworks Subject: US-CA Embedded VXWORKS Software Engineers, 2 openings (Sr & Jr) STELLCOM TECH. Date: 27 Nov 1995 22:11:41 GMT From: resumes Organization: Stellcom Technologies, Inc. Message-ID: <49dd2t$f5i@news.cerf.net> This is a multi-part message in MIME format. - ---------------------------------115992313814673 Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=us-ascii Immediate position open for 2 Software Engineers with experience using VxWorks Development tools. C and/or C++. - ---------------------------------115992313814673 Content-Transfer-Encoding: 7bit Content-Type: text/plain =========================================================================== Stellcom Technologies The leading engineering consulting and project management firm with over 150 full time Software/Hardware engineering professionals in San Diego. Benefits include a 401(K) plan, 125 flexible spending plan, health/dental insurance plans, Microsoft Certification, Technical Training, et.al. Brian Leggee, Sr. Technical Manager (619)298-3860 480 Camino Del Rio South, Suite 226 (619)298-6940 fax San Diego, CA 92108 careers@stellcom.com =========================================================================== - ---------------------------------115992313814673-- --------------------------- Newsgroups: comp.os.vxworks Subject: Problems booting off network? (Hangs after start of loading) Date: Tue, 28 Nov 1995 09:21:31 GMT From: mmigdol@crl.mew.co.jp (Michael Migdol) Organization: Matsushita Electric Works, Ltd. Message-ID: Sender: news@isc.mew.co.jp (USENET News System) Hi all, I'm trying to boot our target (a 486) off the network through an NE2000. The host is a Sparc5 running Solaris 2.4. If I boot the target off of floppy, I can rlogin into the host with no problem, so I know that the physical connection is intact. But when I try to boot off of the network, I see: Attaching network interface ene0...done. Subnet Mask: 0xffffff00 Attaching network interface lo0... done. Loading... 420096 And then the system hangs; I can get back to the boot prompt by hitting ^X a few times. Netstat on the host reveals the following: host.shell target.1022 4096 1459 8760 0 ESTABLISHED host.1023 target.1020 4096 0 8760 0 ESTABLISHED Any suggestions on where to look? The host is newly installed, so there are still a few kinks in it, and I wouldn't be surprised at all if it were something on that end. Thanks in advance Michael Migdol Matsushita Electric Works, Ltd. - -- mmigdol@crl.mew.co.jp (Michael Migdol) Matushita Electric Works Central Research Lab Signal Processing and Software Research Group マイケル　ミグドル 松下電工(株)　中央研究所　信号処理・ソフトウェア研究室 TEL (06)906-8075 FAX (06)904-7104 MIC 7-711-2667 --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VME Addressing configuration with MVME162 & VxWorks Date: 23 Nov 1995 08:56:27 GMT From: kadionik@ixl.u-bordeaux.fr (KADIONIK Patrice) Organization: IXL Lab - MICROELECTRONIC Dept of UNIVERSITY OF BORDEAUX (FRANCE) Message-ID: <491cvr$eq@serveur.cribx1.u-bordeaux.fr> References: Reply-To: kadionik@ixl.u-bordeaux.fr Hi Mark; First point : I suppose that you have uninstalled the SYSCON jumper of your MVME162 board. Second point : you have uninstalled the five jumpers of the daisy chain of the VME slot occupied by your MVME162 board. Two ideas : 1. have you redefined the VME base address of your MVME162 board. It must not be $00000000 (reserved for your Force Board). If not, you have a conflict access. 2. But I think you must verify this point. Have you redefined the VME base address of the GCR registers (Global Control Registers) of your MVME162 board. Motorola have defined for each of their own boards short I/O (A16) global registers (256 bytes) in order to implement for example mailbox interrupts (It's a way to introduce multiprocessing communications). By default, this address is set to FFFFC000. So verify if there's not a overlap with short I/O registers of your Force board and those of your MVME162 board. (I don't know the Force board). If yes, change the GCR base address of your MVME162 board (saved in the NVRAM) and set it to avoid conflict accesses (to FFFFCx00) between your two boards. May these reflexions help you. Regards. Pat. Patrice Kadionik Laboratory of Microelectronics University of Bordeaux - FRANCE --------------------------- End of New-News digest ********************** From twang@gic.gi.com Tue Nov 28 07:30:30 1995 From: "Wang, Tom (HT-MS)" Date: Tue Nov 28 07:30:33 PST 1995 Subject: Block IO driver for the flash memory (28F016SA). Hi, I am looking for the block-IO device driver for the flash memory (28F016SA, the intel FLASHFILE memory 2M x 8), so that I can use the DOS file system and use standard routine to access the files for my vxWorks application. Does anyone have information on this ? Thanks in advance. -Tom Wang =================================================================== Tom Wang E-Mail: twang@gic.gi.com GIC Tel: 215-773-8526 2200 Byberry Rd Fax:215-773-8536 PBC - Dept#1322 Hatboro, PA 19040 =================================================================== From zmau@ksl.co.il Tue Nov 28 12:29:10 1995 From: Maurice Zamir Date: Tue Nov 28 12:29:12 PST 1995 Subject: General questions Hi! 1) Has any one had problems porting from vxWorks 5.1.1 to 5.2 ? 2) Has any one had problems porting the development host from sunOS 4.x to Solaris 2.x ? 3) Does anyone has experience with internationalization & wxWorks ? I am especially interested in Japaneez language (Kanji, Hiragana, Katakana). Thanks Zamir Maurice Zamir Maurice | zmau@ksl.co.il Kulicke & Soffa Ltd. | Phone: (04) - 8545379 A.T.C. Haifa, Israel | fax: (04) - 8550827 From zot!Schilling.com!stevec@netcom.com Tue Nov 28 16:36:55 1995 From: stevec@schilling.com (Steve Cohan) Date: Tue Nov 28 16:36:57 PST 1995 Subject: Access Fault I'm experiencing occassional VME Access Faults in my 3 processor system. The problem occurs during access by two of the processors to a buffer on the 3rd processor via the bus. Since two processors are accessing the same buffer, I'm using vxTas to provide mutual exclusion. The memory is allocated with cacheDmaMalloc (on the 3rd processor), using the RTILib Shm utility. The 3rd processor's memory is accessed A32/D32. The fault always occurs on the same processor, which resides in slot 1 and is system controller: Access Fault Program Counter: 0x000a6888 Status Register: 0x3011 Access Address : 0x821cba60 Special Status : 0x00ed Task: 0x308278 "tSlow" -> tt tSlow a8018 _vxTaskEntry +10 : _HeapArchInstall (0, 0, 0, 0, 0, 0, 0, 0, 0, 0) 3768e0 _HeapArchInstall+42 : _DQI_WriteSlow ([0, a801a, 0, 0, 0]) 34a40e _DQI_WriteSlow +12 : _SMI_Write ([0, 308248, 3768f0, 0, a801a]) 34cc62 _SMI_Write +4a : _bcopy (30a450, 821cba60, 2e) Address 0x821cba60 (the VME address of the shared buffer on the 3rd processor) is valid. At times, the crash occurs a bit deeper in the buffer: Access Fault Program Counter: 0x000a68c6 Status Register: 0x3010 Access Address : 0x821cba70 <--- crash occured 16 bytes further along Special Status : 0x00ed Task: 0x30b55c "tSlow" What does Special Status 0x00ed mean? Do I have a bus contention problem (seems like that would cause a bus error). Thanks, ...................................................................... . . . Steve Cohan Telerobotics for nuclear, . . Chief Engineer, Software/Controls subsea, and other extreme . . Schilling Robotic Systems, Inc. environments . . . . . . 1632 Da Vinci Court tel : 916 753 6718 . . Davis, CA 95616 fax : 916 753 8092 . . email : stevec@schilling.com . . . ...................................................................... From francis@wrs.com Tue Nov 28 16:56:38 1995 From: "Francis St. Amant" Date: Tue Nov 28 16:56:41 PST 1995 Subject: VxWorks 5.2 for MIPS Hello, VxWorks 5.2 for MIPS is as sound as any other VxWorks 5.2 and is definitely not being brought back to WRS. The only news here is that there is no news at all. ------------------------------------------------------------------------------ Francis St. Amant | 510-748-4100 voice Director of Customer Services | 510-814-2010 FAX Wind River Systems | 1010 Atlantic Avenue francis@wrs.com | Alameda, California, 94501 USA ------------------------------------------------------------------------------ > Newsgroups: comp.os.vxworks > Subject: vw5.2 for MIPS recall? > Date: 21 Nov 1995 15:23:18 GMT > From: "Jeff Daly." > Organization: Lockheed Sanders, Inc., Nashua, NH > Message-ID: <48sqt6$kub@rapnet.sanders.lockheed.com> > > i heard yesterday that wrs has brought vxWorks 5.2 for MIPS targets back into > engineering because of some major bugs. is there anyone out there that can > substantiate this, and if so what the 'bugs' are? i'm sure this would be > interesting news for all MIPS vxWorks users. From stan@rti.com Tue Nov 28 17:08:08 1995 From: Stan Schneider Date: Tue Nov 28 17:08:10 PST 1995 Subject: Re: Access Fault Hi Steve, >> >> I'm experiencing occassional VME Access Faults in my 3 processor system. Are all your processors on different bus grant lines? They should be. The daisy chain on BG3 (the default on most cards) won't work reliably for intensive bus activity (like spinning on a TAS). >> The fault always occurs on the same processor, which resides in slot 1 and is >> system controller: >> The fact that it's CPU 0 argues against the bus grant problem. However, you still could be locked out by the other CPU's spinning on TAS. If spreading the BG lines out doesn't help, try increasing the VME timeout on the system controller... HTH, -- Stan stan@rti.com From isys!sherrill@uunet.uu.net Tue Nov 28 18:43:18 1995 From: isys!sherrill@uunet.uu.net (Sherrill Lavagnino) Date: Tue Nov 28 18:43:21 PST 1995 Subject: Alameda, CA Electronics Design Engineer High tech imaging company seeks individual with minimum 4 years experience designing sophisticated digital electronics systems. Experience with CCD cameras, VME, Xilinx, FPGAs, High Speed Memory, VRAMS, DRAMS and PALS a plus. Prefer BSEE/MSEE degree. Isys Controls specializes in high-performance automated inspection systems. Isys inspection systems provide 100% real-time defect information to manufacturers of metal, paper, textiles, non-wovens, plastic film and other web or strip products. We have been in business since 1989, and have systems installed around the U.S. as well as in Europe and the Far East. Please send resume to: Isys Controls, Inc. HW Engineering Dept. 2417 Mariner Square Loop Alameda, CA 94501 or FAX to: (510) 865-9927 or email (plain text only, please) to: jobs@isys.com From daemon@csg.lbl.gov Wed Nov 29 04:00:53 1995 From: daemon@csg.lbl.gov Date: Wed Nov 29 04:00:56 PST 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Wed Nov 29 04:00:48 PST 1995 Subject: loading vox from scsi or network? Subject: GUI display software for host wanted Subject: Re: Ethernet looping back Subject: Re: VxWorks FAQ Wanted Subject: Re: Problems booting off network? (Hangs after start of loading) Subject: Re: GPIB drivers Subject: VXGdb unable to upload object file ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: loading vox from scsi or network? Date: 28 Nov 1995 15:03:34 GMT From: jimenez@shadow (Javier Jimenez) Organization: Lockheed Martin M & S, Sunnyvale, CA Message-ID: <49f8c6$9cj@butch.lmsc.lockheed.com> We have the following situation: We are booting our Motorola 167 boards using vxworks 5.1.1 from a SCSI drive connected to board number 0. The other boards boot from that board. When we "cd" to a network directory the boards can no longer access the SCSI drive. Does anyone know of a way to do this with having to put the boot board in the cdpath? - -> pwd yrt-0: value = 7 = 0x7 - -> ld < /sd0/builds/test/disptime.vox value = 31808588 = 0x1e55c4c ####### Currently I have to do this.... - -> cd "yrt-0:/users/jimenez/temp" value = 0 = 0x0 - -> pwd yrt-0:/users/jimenez/temp value = 26 = 0x1a - -> ld < /sd0/builds/test/disptime.vox value = 32308740 = 0x1ecfe04 ######## I want to be able to do this... - -> cd "/users/jimenez/temp" value = 0 = 0x0 - -> pwd /users/jimenez/temp value = 20 = 0x14 - -> ld < /sd0/builds/test/disptime.vox can't open input '/sd0/builds/test/disptime.vox' errno = 0xd (S_errno_EACCES) Thanks in advance, Javier - -- _ ____ _ _ _ ____ ____ _ _ _ _ ____ __ _ ____ ___ ___| |--| \/ | |=== |--< ___| | |\/| |=== | \| |=== /__ Lockheed Martin Aeronautical Systems 86 South Cobb Dr. jimenez@gelac.lasc.lockheed.com Dept. 73-6M Zone 0670 V:770-494-3060 F:770-494-6989 Marietta, GA 30063-0670 --------------------------- Newsgroups: comp.os.vxworks Subject: GUI display software for host wanted Date: 28 Nov 1995 17:53:20 GMT From: plward@ll.mit.edu (paula l ward) Organization: MIT Lincoln Laboratory Message-ID: <49fiag$el4@llnews.ll.mit.edu> Sender: plw@bart (paula l ward) I have an application for a display package to run on an SGI Indy host computer. I am creating a data stream under VxWorks and sending it TCPIP to the host. I am looking for a somewhat 'canned' display software package that is quickly set up and changed. Something along the lines of Labview but for Unix. I am looking at AP Labs, Veda, XRT, Tcl and Dataviews. Any suggestions or recommendations? Thanks. Paula Ward --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Ethernet looping back Date: Tue, 28 Nov 1995 18:28:42 GMT From: jfinley@netcom.com (John Finley) Organization: NETCOM On-line Communication Services (408 261-4700 guest) Message-ID: References: <49crql$5ri@azure.acsu.buffalo.edu> Sender: jfinley@netcom10.netcom.com Ajit K Ranganathan (akr@acsu.buffalo.edu) wrote: : hi, : i am working on OSI TP4 over null CLNP . : for testing, i need to loop back the packet : to test my receiver. : i can view the packets that i am transmitting : using a network analyzer(sniffer). : the packet has the source and destination ethernet : address fields the same, but i dont seem : to receive the packet. : how do i go about looping the packet back to itself, : since i dont have IP loop back address to use. : i am using a fujitsu nice ethernet driver : and need to know whether the hardware checks : if the packet is for itself before transmitting, which : does not seem to be the case as far as i have seen the : driver output routine.. but even with this, once the : packet has been transmitted, the ethernet address should : be matching with the sender and should get back or not???? : thanks in advance, : ajit. : akr@acsu.buffalo.edu My guess is that you do not have full duplex Ethernet, so you can't transmit and receive at the same time. A packet cannot come in while you are transmitting; that is a collision. John --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks FAQ Wanted Date: Tue, 28 Nov 1995 14:08:03 -0800 From: Strickland_Bob@mm.rdd.lmsc.lockheed.com (Bob Strickland) Organization: Lockheed Martin Missiles & Space Co. Message-ID: References: <48vh5i$clt@llnews.ll.mit.edu> In article <48vh5i$clt@llnews.ll.mit.edu>, plward@ll.mit.edu (paula l ward) wrote: > I am in the process of designing a system that will use vxWorks. > Is there a FAQ available? > > advTHANKSance > Start with the URL http://www.wrs.com (the Wind Rivers Homepage). This leads to various itmes including a users group. - -- Neither I nor Lockheed Martin speak for the other. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Problems booting off network? (Hangs after start of loading) Date: Tue, 28 Nov 1995 14:32:13 -0800 From: Strickland_Bob@mm.rdd.lmsc.lockheed.com (Bob Strickland) Organization: Lockheed Martin Missiles & Space Co. Message-ID: References: In article , mmigdol@crl.mew.co.jp (Michael Migdol) wrote: > Hi all, > > I'm trying to boot our target (a 486) off the network through an NE2000. [SNIP...] > > Attaching network interface ene0...done. > Subnet Mask: 0xffffff00 > Attaching network interface lo0... done. > Loading... 420096 > > And then the system hangs; I can get back to the boot prompt by hitting > ^X a few times. > [SNIP...] Take a look in the 'VxWorks Programmer's Guide (para 2.6.4 for 5.2 or 2.5.4 for 5.1.1). The 'Loading ... nnnn' indicates that the boot process found the file, but note #5 (in the Prog Guide) indicates there may be a problem with the symbol table. Good Luck - -- Neither I nor Lockheed Martin speak for the other. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: GPIB drivers Date: 28 Nov 95 22:23:27 EST From: r4s@icf.hrb.com (Robert M. Sanders) Organization: HRB Systems, Inc. Message-ID: <1995Nov28.222327.24008@hrbicf> References: <4954r5$ja6@ixnews6.ix.netcom.com> In article <4954r5$ja6@ixnews6.ix.netcom.com>, isr@ix.netcom.com (Robert C. Smith ) writes: > Does anyone know of shareware drivers for National Instruments or TI > gpib interface chips other than the ones on the Nat. Instr. ftp site? > Also, anyone have any experience implementing talker or listener > drivers into the VxWorks IO system? I second this request! I'm looking for VxWorks device driver code for inter- facing the MVME167 to the MVME300. I've talked with John Fabiani from Wind River, and he pointed out a number of resources that I'm currently checking in to, but if any one has a definate answer I'd be interested. Rob - -- Robert M. Sanders Internet: R4S@ICF.HRB.COM HRB Systems, Inc. Bitnet : R4S%HRB@PSUECL.BITNET State College, PA USA UUCP : ...!psuvax1!hrbicf!r4s (814)-238-4311 --------------------------- Newsgroups: comp.os.vxworks Subject: VXGdb unable to upload object file Date: Wed, 29 Nov 1995 09:45:50 GMT From: mmigdol@crl.mew.co.jp (Michael Migdol) Organization: Matsushita Electric Works, Ltd. Message-ID: Sender: news@isc.mew.co.jp (USENET News System) Hi all, What mechanism does vxgdb use to upload the object file to the target? When I try to do the load, I get: "Downloading (filename) to target (target)...target system unable to load file." What exactly does this mean? Thanks in advance, Michael - -- mmigdol@crl.mew.co.jp (Michael Migdol) Matushita Electric Works Central Research Lab Signal Processing and Software Research Group マイケル　ミグドル 松下電工(株)　中央研究所　信号処理・ソフトウェア研究室 TEL (06)906-8075 FAX (06)904-7104 MIC 7-711-2667 --------------------------- End of New-News digest ********************** From prb@aplexus.jhuapl.edu Wed Nov 29 05:50:31 1995 From: prb@aplexus.jhuapl.edu (Paul R. Bade) Date: Wed Nov 29 05:50:34 PST 1995 Subject: Re: Access Fault (Steve Cohan) wrote: > > > I'm experiencing occassional VME Access Faults in my 3 processor system. > > The problem occurs during access by two of the processors to a buffer on > the 3rd processor via the bus. Since two processors are accessing the same > buffer, I'm using vxTas to provide mutual exclusion. The memory is allocated > with cacheDmaMalloc (on the 3rd processor), using the RTILib Shm utility. The > 3rd processor's memory is accessed A32/D32. > > The fault always occurs on the same processor, which resides in slot 1 and is > system controller: > > Access Fault > Program Counter: 0x000a6888 > Status Register: 0x3011 > Access Address : 0x821cba60 > Special Status : 0x00ed > Task: 0x308278 "tSlow" > > -> tt tSlow > a8018 _vxTaskEntry +10 : _HeapArchInstall (0, 0, 0, 0, 0, 0, 0, 0, 0, 0) > 3768e0 _HeapArchInstall+42 : _DQI_WriteSlow ([0, a801a, 0, 0, 0]) > 34a40e _DQI_WriteSlow +12 : _SMI_Write ([0, 308248, 3768f0, 0, a801a]) > 34cc62 _SMI_Write +4a : _bcopy (30a450, 821cba60, 2e) > > Address 0x821cba60 (the VME address of the shared buffer on the 3rd processor) > is valid. At times, the crash occurs a bit deeper in the buffer: > > Access Fault > Program Counter: 0x000a68c6 > Status Register: 0x3010 > Access Address : 0x821cba70 <--- crash occured 16 bytes further along > Special Status : 0x00ed > Task: 0x30b55c "tSlow" > > What does Special Status 0x00ed mean? Do I have a bus contention problem > (seems like that would cause a bus error). > It sounds like a timeout trying to get access to the VMEbus. If you are using a card that supports fairness (such as the MVME-167), then you will want the following setup: - make sure that all of your 167 processor cards are requesting on the same VMEbus level. In our case, we have chosen level 2. And we put peripherals at level 1. - set up the bus requester to be fair and to release on request. Fairness basicly says not to request the bus again until nobody else is requesting the bus. Thus, each board takes its turn and then waits. *VMECHIP2_DMACR1 = (DMACR1_SINK_DATA | DMACR1_LVFAIR | DMACR1_LVROR | DMACR1_LVREQ_L2); (note our cache is copyback so we setup snooping to sink data) - Also if you are using a 167, note the timeouts for accessing the bus: /*********************************************************/ /*********************************************************/ /* @prb 4/26/93 WARNING */ /* */ /* The global VMEbus timeout VGTO must be set to 256Us */ /* because the ethernet chip (in un-throtled mode) can */ /* hog the local bus for ~80Us. If another board tries */ /* to access this board while the ethernet chip is */ /* hogging the bus and the VGTO is less than 80Us */ /* the board would timeout and get a bus error */ /* Similarly, the VGTO must be greater than the LBTO */ /* or a timeout could occur if the local CPU is probing */ /* to see if the VSBchip is present (i.e. am I an MV166) */ /* So I recommend: VGTO_256US, VATO_32MS, LBTO_8US */ /*********************************************************/ /*********************************************************/ *VMECHIP2_TIMEOUTCR = (TIMEOUTCR_VGTO_256US | TIMEOUTCR_VATO_32MS | TIMEOUTCR_LBTO_8US | TIMEOUTCR_WDTO_32MS | ((256 - cpuSpeed) & 0xff)); Hope that this helps, +====================================================================+ | __ ____ __ __ | | /\ \ / \ \ /\_\__ /\ \ Johns Hopkins University | | / \_\ | /\ \ \ / / /\_\\ \ \ Applied Physics Lab. | | / /\ | |\ \/ \ \ / / / / / \ \ \ | | / \/ |_| \ /\ \_\ / / / / / \ \ \ Paul R. Bade | | / /\__/_/ \ \ \/_// / / / / / / / (301)-953-6000 x8681 | | / / / \ \_\ \ \/ / / / / / prb@aplexus.jhuapl.edu | | \/_/ \/_/ \__/_/ \/_/ | | __ ____ __ __ | | /\ \ / \ \ /\ \ /\ \ | | / \_\ | /\ \ \ / \_\ / \ \ | | / /\ | |\ \/ \ \ / /\ |_| / /\ \ \ | | / \/ |_| \ /\ \_\ / / / | | \/ / \ \ | | / /\ / / \ \ \/_// / / / / / /\ \_\ | | \ \/ / / \ \_\ \ \/ / / \/ / / / | | \__/_/ \/_/ \__/_/ /_/_/ | | | +====================================================================+ From goodrich@noao.edu Wed Nov 29 09:53:27 1995 From: goodrich@noao.edu (Bret Goodrich) Date: Wed Nov 29 09:53:29 PST 1995 Subject: VME transfer rates Hello All... I need some real world advice about VME transfer rates. We are proposing to put 3 68060 CPUs in a VME chassis, along with 3 in-house boards which provide data at a 2 MB/sec rate (2 KB at 1 KHz). Realistically, is a cumulative 6 MB/sec continuous rate achievable on the VME backplane? How? simple bcopies or DMA transfers? What other ways can you feed a CPU at these rates? Any comments, experience or theoretical, would be greatly appreciated. bret -------------------- Bret Goodrich, Gemini 8 Meter Telescopes Project, National Optical Astronomy Observatories, Tucson, Arizona, USA Internet: goodrich@gemini.edu SPAN/HEPNET: noao::goodrich Phonenet: 1.520.318.8522 Snailnet: PO Box 26732, Tucson, AZ 85726 From isys!bt@uunet.uu.net Wed Nov 29 10:14:39 1995 From: isys!bt@uunet.uu.net (Bret Tilson) Date: Wed Nov 29 10:14:41 PST 1995 Subject: Ariel COMMIO-IP Hi Folks, I am interested if anyone has experience (good or bad) with the Ariel COMMIO-IP DPS board or the Ariel Hydra-II board in a vxWorks system. All responses will be appreciated. Thanks, ------------------------------------------------------------------------- Bret Tilson email: tilson@isys.com Isys Controls, Inc. voice: +510-865-9882 2417 Mariner Square Loop fax: +510-865-9927 Alameda, CA 94501 USA From froeber@bbn.com Wed Nov 29 13:07:23 1995 From: Fred Roeber Date: Wed Nov 29 13:07:25 PST 1995 Subject: Re: VME transfer rates Bret Goodrich asked: >I need some real world advice about VME transfer rates. We are proposing >to put 3 68060 CPUs in a VME chassis, along with 3 in-house boards which >provide data at a 2 MB/sec rate (2 KB at 1 KHz). Realistically, is a >cumulative 6 MB/sec continuous rate achievable on the VME backplane? >How? simple bcopies or DMA transfers? What other ways can you feed >a CPU at these rates? I have worked on a fair number of VME systems where backplane performance was a critical issue. Measuring performance in a number of systems has shown that the maximum throughput is quite variable and depends on the exact boards you are using. In particular, the VME interface chips you are using makes a difference. At any rate, I have found that a loading of 3-4 M transfers a second is usually easy to achieve on a VMEbus. You don't say but if your device supports D32 accesses, then you should have no problem achieving 12-16 MB/sec throughput. There are a number of ways to achieve (much) higher throughputs. Some things you can do (sort of in order of less to more effective in boosting rates): - If application is set up such that processors read data in non overlapping fashion, switch system controller to use release-on-request (ROR) arbitration vs the release-when-done (RWD) arbitration. - If you are writing data to a card that supports it, turn on write posting on the card so that it buffers the data and acknowledges the transfer immediately. - Use a DMA engine to transfer data rather using than program transfer (i.e. bcopy). Should be able to get at least 20MB/sec in many cases. - Do burst transfers (i.e. BLT transfers). Used with DMA, you might get 30MB/sec. - Use D64 transfers if they are supported (this is usually combined with DMA and burst transfers to get rates that can approach 60-70MB/sec) Well, thats some rough types of numbers that I use in sizing systems. As I said, your mileage may vary. Fred | Fred J Roeber, BBN Systems & Technologies | | 50 Enterprise Place Middletown, RI 02842-5202 | | froeber@bbn.com 401-849-2543 (X48) | From phil@naic.edu Wed Nov 29 13:44:52 1995 From: phil@naic.edu (Phil Perillat) Date: Wed Nov 29 13:44:55 PST 1995 Subject: Re: VME transfer rates > Submitted-by goodrich@noao.edu Wed Nov 29 09:53:27 1995 > Submitted-by: goodrich@noao.edu (Bret Goodrich) > > Hello All... > > I need some real world advice about VME transfer rates. We are proposing > to put 3 68060 CPUs in a VME chassis, along with 3 in-house boards which > provide data at a 2 MB/sec rate (2 KB at 1 KHz). Realistically, is a > cumulative 6 MB/sec continuous rate achievable on the VME backplane? > How? simple bcopies or DMA transfers? What other ways can you feed > a CPU at these rates? > > Any comments, experience or theoretical, would be greatly appreciated. I don't have any experience with 68060 boards, but for an mv167 board (68040) running at 33 mhz we get: 4-5mb/sec single word i/o across vmebus 15 mb/sec using 32bit dma xfers to a custom a/d board we built 28 mb/sec using D64 xfers between a mv162 at 25mhz and a mv167 at 33mhz. If each of your boads only had to do 2mb/sec you could get away with single word i/o but then your processor wouldn't have as much time to do any computations that were needed. On the other hand if you want to do dma, you'll have to make your home built device support it. I'd also thing about buffering the data in your home built boards so you wouldn't have to read 2kb every millisecond. At that rate, the dma startup and interrupt overhead might eat into your i/o rate. phil perillat arecibo observatory phil@naic.edu From bqv@se19.wg2.waii.com Wed Nov 29 14:56:51 1995 From: Bang Vu Date: Wed Nov 29 14:56:53 PST 1995 Subject: Re: loading vox from scsi or network? Javier Jimenez penned: > >We have the following situation: We are booting our Motorola 167 boards >using vxworks 5.1.1 from a SCSI drive connected to board number 0. The >other boards boot from that board. When we "cd" to a network directory >the boards can no longer access the SCSI drive. Does anyone know of a way >to do this with having to put the boot board in the cdpath? > > >-> pwd >yrt-0: >value = 7 = 0x7 >-> ld < /sd0/builds/test/disptime.vox >value = 31808588 = 0x1e55c4c > >####### Currently I have to do this.... >-> cd "yrt-0:/users/jimenez/temp" >value = 0 = 0x0 >-> pwd >yrt-0:/users/jimenez/temp >value = 26 = 0x1a >-> ld < /sd0/builds/test/disptime.vox >value = 32308740 = 0x1ecfe04 > >######## I want to be able to do this... >-> cd "/users/jimenez/temp" >value = 0 = 0x0 >-> pwd >/users/jimenez/temp >value = 20 = 0x14 >-> ld < /sd0/builds/test/disptime.vox >can't open input '/sd0/builds/test/disptime.vox' > errno = 0xd (S_errno_EACCES) > > >Thanks in advance, > >Javier >-- > _ ____ _ _ _ ____ ____ _ _ _ _ ____ __ _ ____ ___ >___| |--| \/ | |=== |--< ___| | |\/| |=== | \| |=== /__ >Lockheed Martin Aeronautical Systems 86 South Cobb Dr. > jimenez@gelac.lasc.lockheed.com Dept. 73-6M Zone 0670 > V:770-494-3060 F:770-494-6989 Marietta, GA 30063-0670 When you try to do a , you have to be able to NFS mount the /users. vxWorks 5.1 doesn't have the ability to act as a NFS server (vxWorks 5.2 does). Therefore, i don't think you can do what you want. Somebody may want to correct me on this. BTW, the works for you because file access is done through rsh with the remote host being yrt-0. -- Bang Vu | QC Tools, Inc. (a subsidiary of I/O, Inc.) vu@wg2.waii.com | Software, Systems, and Consultancy for the Voice 713-964-6192 | Exploration Industry. Fax 713-964-6218 | 3600 Briarpark, Houston, TX 77042 From leec@ashbeech.nosc.mil Wed Nov 29 18:56:03 1995 From: leec@ashbeech.nosc.mil (Lam Lee--NP Code 824) Date: Wed Nov 29 18:56:05 PST 1995 Subject: Boot from SCSI Hi there! I just set up a backplane using several PT-VME 151 boards. The master board, processor #0, is booting from a SCSI drive. How do I get the rest of my boards to boot from the same SCSI drive? Any ideas would be appreciated. Lam From matsuo@ashbeech.nosc.mil Wed Nov 29 19:18:07 1995 From: matsuo@ashbeech.nosc.mil (Eric Matsuo--NP code 824 10/10/95-1/10/96) Date: Wed Nov 29 19:18:09 PST 1995 Subject: register NAME: Eric Matsuo Company: NRaD Phone Number: (619) 553-6454 E-mail: matsuo@cod.nosc.mil Hello, I just set up a backplane using several PT-VME 151 boards. My master board is booting from a SCSI drive. I want get my other PT 151 boards to boot from the same SCSI drive. Does anyone know how to do this? Any ideas? From tkb@mclean.sparta.com Wed Nov 29 19:22:09 1995 From: Thomas Keith Buchanan Date: Wed Nov 29 19:22:12 PST 1995 Subject: Re: Boot from SCSI > I just set up a backplane using several PT-VME 151 boards. The master > board, processor #0, is booting from a SCSI drive. How do I get the > rest of my boards to boot from the same SCSI drive? Any ideas would > be appreciated. Set up the others with boot device "sm" or "sm=?" for a SM anchor address that doesn't match the defaults in config.h. Put a bogus password in the ftp password field. The boot host will be the backplane host name and IP address of the CPU with the SCSI disks attached. The other processors should boot via ftp from processor #0. Sometimes the timing is a little tricky, the SCSI device needs to be available before SM is fully initialized. good luck ______________________________________________________ | |\ |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 | | | / \\\\\ \\ 703-448-0120 | | | \ \\\\\ // | | | \ ////// | | | \////// SPARTA | | | \//// Pride In Performance | | | \// -------------------- | | |_____________________________________________________| | \______________________________________________________\| From jupiter!bolz%mars.egt.fr@lbl.gov Thu Nov 30 00:51:23 1995 From: Bernard Bolz Date: Thu Nov 30 00:51:26 PST 1995 Subject: unsubscribe request Please unsubscribe me. B. BOLZ. -- +------------------------------------+------------------------------------+ | Bernard BOLZ | European Gas Turbines | | EGT PEDL | 3, avenue des Trois Chenes | | tel: (33) 84.55.49.57 | 90018 BELFORT Cedex | | fax: (33) 84.55.35.21 | FRANCE | | E-Mail: bbolz@egt.fr | | +------------------------------------+------------------------------------+ From daemon@csg.lbl.gov Thu Nov 30 04:01:12 1995 From: daemon@csg.lbl.gov Date: Thu Nov 30 04:01:19 PST 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Thu Nov 30 04:00:58 PST 1995 Subject: Re: X25 for VxWorks? Subject: Only add and subtract with long long Subject: No X with 64MB SPARC10 Subject: bcopy_to_mbufs routine bug Subject: VxMP multiprocessor accessory Subject: (no subject) Subject: (no subject) ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: X25 for VxWorks? Date: Mon, 27 Nov 95 09:31:37 CET From: kai@ade.no (Kai Atle Myrvang) Message-ID: <251fjms5f@ade.no> References: <9511241549.AA11024@rtp.co.uk> Reply-To: kai@ade.no > >Hi, > >We have a potential application which needs to run on an MVME 162 or 167, >but which requires connection to an X25 WAN. > >Does anybody know of any off-the-shelf solutions to this - in terms of >hardware, software or both (and what physical connections are provided?) > VIX25 from PEP Modular Computers provides you with a X.25 connection. It's an intelligent slave board using dp memory as master interface. Physical connections are 15 pin Dsub X.21bis. There is lots more info in the datasheet. Since your sig. says UK, PEP's UK office phone/fax: 0273 44 11 88 (phone) 0273 44 11 99 (fax) or e-mail sales@pep-kaufbeuren.de. - -- * Kai Atle Myrvang (kai@ade.no) * * AD Elektronikk AS, P.O.Box 104, N-1405 Langhus, Norway * * Tel: int+47 64 86 99 70 | Fax: int+47 64 86 99 20 * ********************************************************************* --------------------------- Newsgroups: comp.os.vxworks Subject: Only add and subtract with long long Date: Mon, 27 Nov 1995 02:05:00 GMT From: duane6@ix.netcom.com Organization: Netcom Message-ID: <49b6lc$9fv@ixnews2.ix.netcom.com> WRS says they only support addition and subtraction with "long long" data types. I have a FORCE10 target and BSP. Anyone doing multiplies, divides with "long long" data types??? --------------------------- Newsgroups: comp.os.vxworks Subject: No X with 64MB SPARC10 Date: Mon, 27 Nov 1995 02:05:12 GMT From: duane6@ix.netcom.com Organization: Netcom Message-ID: <49b6ln$9fv@ixnews2.ix.netcom.com> x does not cause a reboot with a 64MB Force SPARC (BSP frc10) target. All works well with a 32MB board. Anyone solved this one? Thanks to all that respond!! Duane --------------------------- Newsgroups: comp.os.vxworks Subject: bcopy_to_mbufs routine bug Date: 29 Nov 1995 20:57:38 GMT From: akr@acsu.buffalo.edu (Ajit K Ranganathan) Organization: UB Message-ID: <49ihg2$8rp@azure.acsu.buffalo.edu> i was trying to use this routine at the ethernet driver to copy the ethernet frame data to mbufs. the arguments to this routine are 1. pointer to the ethernet frame buffer. 2. length to be copied 3. offset 4. pointer to interface structure. when the routine returns, the length field in the mbuf , namely m->m_len does not contain what i gave as the argument. it contains a number greater than what i gave. this causes problems later.. dont know why?/ any idea.. ajit. --------------------------- Newsgroups: comp.os.vxworks Subject: VxMP multiprocessor accessory Date: Wed, 29 Nov 1995 21:07:39 GMT From: cjones@digsys1.bwi.wec.com (Chris Jones) Organization: Westinghouse Electric Corp. Message-ID: Sender: usenet@tron.bwi.wec.com (USEnet News Poster) Does anyone have any experience with the VxMP multiprocessor accessory? We have an application which will require several PowerPC's to share the systems tasks. Throughput is very important and I would like to find out how well this accessory works. Also, does anyone have experience with other RTOS's and their multiprocessing capabilities, specifically VRTX and pSOS? Response to cjones@digsys1.bwi.wec.com if you can. Thanks ! Chris Jones Electrical Engineer Westinghouse Electronic Systems --------------------------- Newsgroups: comp.os.vxworks Subject: (no subject) Date: 28 Nov 1995 20:24:12 GMT From: Maurice Zamir Organization: NetVision LTD. Message-ID: <49fr5c$nv9@news.netvision.net.il> Hi! 1) Has any one had problems porting from vxWorks 5.1.1 to 5.2 ? 2) Has any one had problems porting the development host from sunOS 4.x to Solaris 2.x ? 3) Does anyone has experience with internationalization & wxWorks ? I am especially interested in Japaneez language (Kanji, Hiragana, Katakana). Thanks Zamir Maurice Zamir Maurice | zmau@ksl.co.il Kulicke & Soffa Ltd. | Phone: (04) - 8545379 A.T.C. Haifa, Israel | fax: (04) - 8550827 --------------------------- Newsgroups: comp.os.vxworks Subject: (no subject) Date: 28 Nov 1995 20:24:08 GMT From: Maurice Zamir Organization: NetVision LTD. Message-ID: <49fr58$nv9@news.netvision.net.il> Hi! 1) Has any one had problems porting from vxWorks 5.1.1 to 5.2 ? 2) Has any one had problems porting the development host from sunOS 4.x to Solaris 2.x ? 3) Does anyone has experience with internationalization & wxWorks ? I am especially interested in Japaneez language (Kanji, Hiragana, Katakana). Thanks Zamir Maurice Zamir Maurice | zmau@ksl.co.il Kulicke & Soffa Ltd. | Phone: (04) - 8545379 A.T.C. Haifa, Israel | fax: (04) - 8550827 --------------------------- End of New-News digest ********************** From uchenick@tate.com Thu Nov 30 05:26:42 1995 From: uchenick@tate.com (Gordon Uchenick) Date: Thu Nov 30 05:26:44 PST 1995 Subject: Hardware/Software for V.35 comm Can anyone recommend a VME hardware/vxWorks software combination for supporting one V.35 comm line @ 64K bps? Also, how about the same recommendation for RS-232 (synchronous) @ 64K bps? TIA, Gordon Uchenick uchenick@tate.com Phone: 410-581-0422 Fax: 410-581-5738 From kennyk@lannet.com Thu Nov 30 05:55:15 1995 From: kennyk@lannet.com (Kenny Krupnik) Date: Thu Nov 30 05:55:18 PST 1995 Subject: RIP on VxWorks Hi ! I would like to know if anyone has tried to port routed (RIP) on VxWorks. If so, is it possible to obtain the sources ? Thanks, Kenny Krupnik (E-mail: kennyk@lannet.com) From bqv@se19.wg2.waii.com Thu Nov 30 07:30:35 1995 From: Bang Vu Date: Thu Nov 30 07:30:37 PST 1995 Subject: Re: comp.os.vxworks newsdigest duane6@ix.netcom.com penned: > >x does not cause a reboot with a 64MB Force SPARC (BSP frc10) >target. All works well with a 32MB board. Anyone solved this one? >Thanks to all that respond!! > >Duane > I thought i had responded to this one. You may have missed it. Anyway, ^X can be changed to something else by the tyMonitorTrapSet command (one of the tyLib function calls). You may want to use this command yourself to set the reboot character to ^D, for example. If ^D reboots your target then this means that someone has used tyMonitorTrapSet (you may want to look for it in your init script) to change the reboot character. If ^D doesn't reboot your target then i have no idea why the reboot (aka trap-to-monitor) doesn't work for you. -- Bang Vu | QC Tools, Inc. (a subsidiary of I/O, Inc.) vu@wg2.waii.com | Software, Systems, and Consultancy for the Voice 713-964-6192 | Exploration Industry. Fax 713-964-6218 | 3600 Briarpark, Houston, TX 77042 From guthrie@power.amasd.anatcp.rockwell.com Thu Nov 30 08:29:49 1995 From: guthrie@power.amasd.anatcp.rockwell.com (Bob Guthrie) Date: Thu Nov 30 08:29:51 PST 1995 Subject: Re: VME Transfer Times I was able to get about 78 MB/sec between a MVME166 and a Micro Memory 128 MB board using VME64 BLT with the VMEChip2. Robert Guthrie From phil@naic.edu Thu Nov 30 10:27:19 1995 From: phil@naic.edu (Phil Perillat) Date: Thu Nov 30 10:27:22 PST 1995 Subject: Re: VME Transfer Times > Submitted-by guthrie@power.amasd.anatcp.rockwell.com Thu Nov 30 08:29:49 1995 > Submitted-by: guthrie@power.amasd.anatcp.rockwell.com (Bob Guthrie) > > I was able to get about 78 MB/sec between a MVME166 and a Micro Memory > 128 MB board using VME64 BLT with the VMEChip2. > > Robert Guthrie When i measure speeds like this (i have in the past) ,the first thing i do is check my timing routine, then I get out the logic analyzer to see that the data is actually being moved. Although i haven't worked with a mv166 board ,i'm pretty sure they are just an mv167 with a vsb bus added. For the 167/162,162fx i use the following as a sanity check: the best the local bus can do is a burst xfer of 4 1 1 1 (at least this is the best i've seen in the quoted specs for normal memory configurations). That's 16 bytes in 7 clock cycles (and some of the above boards can't even do that because of their memory configuration) for a 25mhz board this gives 16/(7*40ns)= 57mb/sec. for a 33mhz board this gives 16/(7*33ns)=76mb/sec. This is 100% utilization of the local bus by the dmac xfer. That leaves nothing for any processing by the cpu, and no local bus bandwidth for generating or bringing the data in. So you may be able to get the above stated i/o rates as a burst but i'd be really surprised if you could sustain it (unless you are willing to send the same data over and over to the memory board). so much for my $.02... Now for my questions.... We are using 4 mv162fx boards running at 32mhz and a mv167 board in a vme crate. On each mv162fx, the ip bus is running at 32mhz. We read from the ip bus, process the data a little, and then use the vmechip2 to dma the data out to the mv167 board. These 3 operations are all overlapped on each mv162fx board. The problem is kind of a catch 22.. We have interrupts on the local bus that need to be serviced very often. At the same time we want to do long dma xfers out over the vmebus to get the best vmebus i/o. As far as i can tell , the local bus priorities are fixed on the mv162fx.. ipchip highest local bus priority scsi chip vmechip2 cpu lowest bus priority.... So i see the ipDevice generate an interrupt, but the interrupt acknowledge cycle takes a long time to occur since the vmechip2 is doing dma over the local bus. Although the vmechip2 has a bunch of options to change it's priority access to the vmebus , i haven't found anything that allows you to modify the local bus access priorities. I know the correct answer is to slow down the local bus interrupts by more buffering on the local ip device... but the board's already made... so my questions are: 1. does anyone have any ideas on how to boost the cpu's local bus priority so it can execute the interrupt service routine ahead of the vmechip'2 dma xfer. 2. does anyone know how to hook up an hp1660 logic analyzer to the local bus of an mv162fx computer? The 68040 is socketed, and there is a mezzanine memory card (although i don't know if it is looking at all the local bus signals). I would maybe unplug the 68040, plug in an adaptor that would then take the 68040 and give me some pins to hook my logic analyzer to... thanks, phil perillat arecibo observatory phil@naic.edu