From akbiocca@lbl.gov Fri Jan 2 19:16:58 1998 From: Alan K Biocca Date: Fri Jan 2 19:17:02 PST 1998 Subject: Re: Advertising Hi, Russ. I'd like a good, simple solution (since I have to write and maintain it). Registration of users is tough - we have many hundreds, and they often post from other than their normal address, and many more users on the newsgroup. Your post caused me to consider another methodology. At the current time we have about 30 filters installed that reject quite a bit of junk. Obviously these are insufficient. However it might be simpler and still effective to only accept messages that had certain words in them. A list of enabling words would probably not be hard to come up with that would not occur in these offensive advertising postings. I am open to other suggestions. Regarding "Help Wanted" queries - do folk feel that vxworks related jobs should be posted here? Thanks, Alan K Biocca Berkeley Lab Advanced Light Source Controls Section Leader VxWorks Exploder software author akbiocca@lbl.gov >I don't know about the rest of you out there, but I am getting pretty >fed up with the unsolicited advertising as of late. This is SUPPOSSED to >be a forum for users of vxWorks. I can accept "advertising" from WRS or >other vendors of products related to our work, however advertising for >life insurance, on-line gambling and just about any product you can >think of, is out of line. What's worse, is when these items these items >get incorporated into the daily digest. Personally, I believe "help >wanted" ads are also out of place here, but TMHO. > >Isn't there something that can be done, maybe require registration of >users, to eliminate some of this clutter ? > >Russ >-- From bmg@pacbell.net Sat Jan 3 09:04:27 1998 From: mac Date: Sat Jan 3 09:04:31 PST 1998 Subject: Re: Advertising > >Regarding "Help Wanted" queries - do folk feel that vxworks related jobs >should be posted here? > Yes. Emphatically yes. Brian From froeber@BBN.COM Sat Jan 3 10:03:41 1998 From: Fred Roeber Date: Sat Jan 3 10:03:44 PST 1998 Subject: Re: Advertising On Fri, 2 Jan 1998, Alan K Biocca wrote about filtering out the reams of junk email that comes through the exploder: > Hi, Russ. I'd like a good, simple solution (since I have to write and > maintain it). Registration of users is tough - we have many hundreds, and > they often post from other than their normal address, and many more users > on the newsgroup. > > Your post caused me to consider another methodology. At the current time we > have about 30 filters installed that reject quite a bit of junk. Obviously > these are insufficient. However it might be simpler and still effective to > only accept messages that had certain words in them. A list of enabling > words would probably not be hard to come up with that would not occur in > these offensive advertising postings. This sounds like a very good method. Trying to filter out junk email in general is a real bear since there are so many sources of it and the different messages follow no simple formulas (hey, if it was easy to filter it wouldn't be as effective). With the exploder, however, we have the strong advantage of having common subject matter. We could make something as simple as requiring any post to include the words "vxworks" or "wrs" somewhere in them. If someone wasn't going to reference one of these words directly they could simply tack one of them on to the end of the message. The idea works for me. This would also cut down on some of the irrelavent cross postings that sometimes show up (e.g. the thread on year 2000 and different facts about dates has been lingering around too long fueled by cross postings from other groups). > Regarding "Help Wanted" queries - do folk feel that vxworks related jobs > should be posted here? I don't find these posts helpful or of interest but I can see where others may. Seems it would be better if there was some central web site for this information where interested people could look. Just my $.02. Fred | Fred J Roeber, BBN Systems & Technologies | | 4 John Clarke Road Middletown, RI 02842-5202 | | froeber@bbn.com 401-848-3548 | From norberto@dem.uc.pt Mon Jan 5 04:21:23 1998 From: "J. Norberto Pires" Date: Mon Jan 5 04:21:28 PST 1998 Subject: Re: Advertising At 11:39 AM 1/5/98, you wrote: >>Regarding "Help Wanted" queries - do folk feel that vxworks related jobs >>should be posted here? Yep! Sometimes this is a nice and faster way of having some help or hint. Isn't that one of the reasons to have a mail-list? Happy new year! norberto From mwbrown@sandia.gov Mon Jan 5 07:40:18 1998 From: "Matthew W. K. Brown" Date: Mon Jan 5 07:40:21 PST 1998 Subject: Re: Advertising the vxWorks Users Group Exploder wrote: > > Submitted-by akbiocca@lbl.gov Fri Jan 2 19:16:58 1998 > Submitted-by: Alan K Biocca > ... > > Regarding "Help Wanted" queries - do folk feel that vxworks related jobs > should be posted here? > > Yes!!! From del@rst.com Mon Jan 5 16:15:17 1998 From: del@rst.com (Mark Del Giorno) Date: Mon Jan 5 16:15:21 PST 1998 Subject: fopen() fails on big files Heyyyyyyyyyyyyyyyyyyyyyyyy, vxWorks 5.2 Target: MVME162 I'm trying to fopen() a file, and for big files, it fails with: >0x3ae498 (tShell): memPartAlloc: block too big - 524 in partition 0x99332. >TIGER:cat: write error: Broken pipe >S_netDrv_UNIX_FILE_ERROR The file is 2.3 MB. I can successfully open 1.6 MB. What are the vxWorks memory requirements? Must I have 2.3 MB of memory available to fopen() the file? [I hope the answer to that is "no".] I have #INCLUDE_NFS in configAll.h, which I thought was supposed to solve the problem. Thanks in advance for any info. - Mark _/_/_/_/ _/_/_/_/ _/_/_/_/_/ Mark Del Giorno (del@rst.com) _/ _/ _/ _/ Robotic Systems Technology _/_/_/_/ _/_/_/_/ _/ 1234 Tech Court _/ _/ _/ _/ Westminster, MD 21157 _/ _/ _/_/_/_/ _/ (410) 876-9200 Fax: 876-9470 http://www.rst.com From froeber@BBN.COM Mon Jan 5 19:45:41 1998 From: Fred Roeber Date: Mon Jan 5 19:45:49 PST 1998 Subject: Re: fopen() fails on big files On Mon, 5 Jan 1998, Mark Del Giorno wrote: > I'm trying to fopen() a file, and for big files, it fails with: > > >0x3ae498 (tShell): memPartAlloc: block too big - 524 in partition 0x99332. > >TIGER:cat: write error: Broken pipe > >S_netDrv_UNIX_FILE_ERROR > > > The file is 2.3 MB. I can successfully open 1.6 MB. > What are the vxWorks memory requirements? > Must I have 2.3 MB of memory available to fopen() the file? > [I hope the answer to that is "no".] > > I have #INCLUDE_NFS in configAll.h, which I thought was supposed to solve > the problem. Including NFS IS a key first step in getting to a solution to this problem. The problem is that the file is being opened using the default VxWorks file system driver; netDrv. Note the error code S_netDrv_UNIX_FILE_ERROR which references the system service reporting the error and is always a good clue to the problem source. To get around the use of netDrv, it isn't enough to include the NFS driver in your VxWorks image. You also have to mount a file system using the NFS driver and then make sure to access (i.e. open) a file on the mounted file system. To mount a file system use the call: nfsMount(host, fileSystem, localName) The VxWorks programmer guide goes into detail on mounting file systems using NFS. Once the system is mounted (check the return status from nfsMount), opening a file that starts with the localName cited above will result in a file that is accessed using NFS. NFS writes the files in blocks (nominally 8KB). Thus, you don't need a huge memory to access a large file. You don't say what operations you do on the file but there are some minor differences between the operations supported by the two drivers (e.g. flushing and appending to files). Good luck. Fred | Fred J Roeber, BBN Systems & Technologies | | 4 John Clarke Road Middletown, RI 02842-5202 | | froeber@bbn.com 401-848-3548 | From jagan@wipinfo.soft.net Mon Jan 5 23:15:47 1998 From: Jagan Mohan Date: Mon Jan 5 23:15:58 PST 1998 Subject: VxGdb Hi ! Can someone give me the deatils of using VxGdb 2.0.. I have a remote MIPS target and am doing source debugging using Slip. (Remote CPU : R3000 MIPS, OS : Vxworks 5.1, compiler : ccx) Though I have done the configuration as given in the manual, my target CPU crashes after debugging for a few minutes. More the breakpoints I have in the source code, faster is the crash. Also, without this source level debugging, code seems fine and target CPU never hangs. Someone had similar problem using VxGDB ..? Thanx in advance Jagan From c.j.slominski@larc.larc.nasa.gov Tue Jan 6 11:35:39 1998 From: "Christopher J. Slominski" Date: Tue Jan 6 11:35:42 PST 1998 Subject: Serial I/O Group, I am going to use the standard VxWorks serial driver for COM2 on my PowerPC 604. Documentation for the "read" function implies that it returns immediately, providing the number of bytes read up to the maximum specified in the command. I also have read a warning in WRS documentation that the "read" and "write" functions may block. Does the "read" ever wait for the device, or will it always return the number of bytes available (0 to maxBytes) if no ERROR occurs? Is this configurable with "ioctl"? Do the device driver's functions (and other VxWorks system calls) become part of the task thread calling them, and can be interrupted by higher priority tasks? Does "read" or other system calls have critical sections where they become non-interruptable? If so, for any significant time? There must be some semaphore work done in "read" which is treated as a critical section. I am trying to assess how a "read" in a low priority background task, fetching continuous slow rate data, will effect my system determinism. *----------------------------------------* | Christopher J. Slominski | | Computer Sciences Corporation (CSC) | | 3217 N. Armistead Ave. | | Hampton, VA 23666 | | (757) 766-8258 | | c.j.slominski@larc.nasa.gov | | NASA Mail Stop 931 | *----------------------------------------* From koch@tcville.ES.HAC.COM Tue Jan 6 16:05:21 1998 From: koch@tcville.ES.HAC.COM (Kim Chacon) Date: Tue Jan 6 16:05:24 PST 1998 Subject: question on how to do non-blocking keyboard reads Hello fellow VxWorks users: I am using VxWorks 5.3 on a Vista PowerPC processor. I need to be able to read the keyboard, one character at a time, in a non-blocking (polling) mode. I have successfully used the ioctl() and read() calls to read a single character at a time, but the read() call is blocking until a key gets entered on the keyboard. Does anyone know how to configure the file descriptor, so that the read call is non-blocking? I've checked the VxWorks documentation, but have not found a solution yet. Any help would be appreciated. Thank you and regards, Kim M. Chacon Raytheon Systems El Segundo, California phone: (310) 616-6471 email: chacon@tcville.es.hac.com (at internet) From AKBiocca@lbl.gov Tue Jan 6 17:16:00 1998 From: Alan K Biocca Date: Tue Jan 6 17:16:04 PST 1998 Subject: ALL READ - New Exploder Filter Based on the response I received regarding advertising annoyance I have tenatively installed a filter that REQUIRES the word 'vxworks' in the message body or subject (case not important) in order that it be exploded. This does not apply to the newsgroup (we have no control over that at the current time), or to the digests from the newsgroup (this will take additional programming and could be done in the future). If an article submitted does not have the word 'vxworks' somewhere within it - it will not be posted, and will silently sink into the abyss of rejected postings. We may consider allowing rejects to be viewed with a web page for the five days they hang around in the system. Give us feedback on this 'new' feature - do we want it or not? Thanks, and back to our regular service -- Alan K Biocca exploder software author From AKBiocca@lbl.gov Tue Jan 6 17:42:58 1998 From: Alan K Biocca Date: Tue Jan 6 17:43:02 PST 1998 Subject: ALL READ - New Exploder Filter Based on the response I received regarding advertising annoyance I have tenatively installed a filter that REQUIRES the word 'vxworks' in the message body or subject (case not important) in order that it be exploded. This does not apply to the newsgroup (we have no control over that at the current time), or to the digests from the newsgroup (this will take additional programming and could be done in the future). If an article submitted does not have the word 'vxworks' somewhere within it - it will not be posted, and will silently sink into the abyss of rejected postings. We may consider allowing rejects to be viewed with a web page for the five days they hang around in the system. Give us feedback on this 'new' feature - do we want it or not? Thanks, and back to our regular service -- Alan K Biocca exploder software author From daemon@csg.lbl.gov Wed Jan 7 04:00:33 1998 From: daemon@csg.lbl.gov Date: Wed Jan 7 04:00:36 PST 1998 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Wed Jan 7 04:00:29 PST 1998 Subject: Re: Advertising ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Advertising Date: Tue, 6 Jan 1998 13:07:17 -0800 From: "Jerry Asher" Organization: DNAI ( Direct Network Access ) Message-ID: <68v4f5$2rv$1@castor.dnai.com> References: <3.0.1.32.19980103090334.00b7e6a0@pacbell.net> >>Regarding "Help Wanted" queries - do folk feel that vxworks related jobs >>should be posted here? >> > >Yes. Emphatically yes. Seconded with a resounding YES! Jerry --------------------------- End of New-News digest ********************** From uchenick@tis4000.com Wed Jan 7 04:05:45 1998 From: Gordon Uchenick Date: Wed Jan 7 04:05:48 PST 1998 Subject: RE: question on how to do non-blocking keyboard reads Vxworks vxworks vxWorks VXWORKS Does anyone know how to configure the file descriptor, so that the read call is non-blocking? Three methods come to mind. First, you could use the FIONREAD ioctl to determine if there were any bytes available in the input buffer. If the return is positive, then there are bytes in the buffer and a read call will not block. If zero, then a read would block, so DON'T do it. I have used this technique successfully. Second way, try setting FIONBIO in your ioctl. This sets the fd to non-blocking. Doing a read call on an fd with no data ready returns ERROR with ERRNO set to EWOULDBLOCK. I have not used this technique on serial lines, but I have used it on sockets with good results. Third, it seems to me that using the select call to determine if the fd is ready for reading is the most straightforward and portable way to solve your problem. Select transfers to other operating systems easily. Contact me directly if you want to discuss this further. Gordon uchenick@tis4000.com From currie@ghg.net Wed Jan 7 05:49:36 1998 From: currie@ghg.net (Alan Currie) Date: Wed Jan 7 05:49:41 PST 1998 Subject: vxworks Re: ALL READ - New Exploder Filter I like it. From a.g.kirke@ieee.org Wed Jan 7 07:35:15 1998 From: Al Kirke Date: Wed Jan 7 07:35:19 PST 1998 Subject: Re: vxworks Re: ALL READ - New Exploder Filter At 05:49 AM 1/7/98 PST, you wrote: >Submitted-by currie@ghg.net Wed Jan 7 05:49:36 1998 >Submitted-by: currie@ghg.net (Alan Currie) > >I like it. > >********** vxworks Should this message have filtered? vxworks From AKBiocca@lbl.gov Wed Jan 7 08:53:28 1998 From: Alan K Biocca Date: Wed Jan 7 08:53:31 PST 1998 Subject: Re: ALL READ - New Exploder Filter Apparently the filter is working. So far it caught a couple of respondents who liked the filter and were responding positively to its existence, but didn't actually include the word vxworks in their messages. Perhaps they were testing it. I did actually send the notification twice during my testing, so that is why there were two messages broadcast. The exploder takes quite awhile to finish mailing each message. Thanks, Alan From CBarkley@NLvl.com Wed Jan 7 10:02:10 1998 From: "Barkley, Charles (HT-EX)" Date: Wed Jan 7 10:02:14 PST 1998 Subject: RE: ALL READ - New Exploder Filter >Based on the response I received regarding advertising annoyance I have >tenatively installed a filter that REQUIRES the word 'vxworks' in the >message body or subject (case not important) in order that it be exploded. This seems like a reasonable compromise. It doesn't necessarily block help wanted ads, only those not specifically stating 'vxworks' (after all, this IS a vxworks group), while keeping unrelated advertising and other spam out. >===================================================== >Chuck Barkley >NextLevel Systems, Inc. Tel: (215) 773-8509 >2200 Byberry Road Fax: (215) 773-8536 >Hatboro, PA 19040 Email: cbarkley@nlvl.com >===================================================== > From cpinkard@falcon.kla-tencor.com Wed Jan 7 10:21:49 1998 From: Christopher Pinkard Date: Wed Jan 7 10:21:52 PST 1998 Subject: Re: vxworks Re: ALL READ - New Exploder Filter the vxWorks Users Group Exploder wrote: > > Submitted-by a.g.kirke@ieee.org Wed Jan 7 07:35:15 1998 > Submitted-by: Al Kirke > > At 05:49 AM 1/7/98 PST, you wrote: > >Submitted-by currie@ghg.net Wed Jan 7 05:49:36 1998 > >Submitted-by: currie@ghg.net (Alan Currie) > > > >I like it. > > > >********** > vxworks Should this message have filtered? vxworks > > ********** > Glancing through the news group directly, it looks like one out of 3 or 4 valid messages might get filtered. What if we had a list of trigger words, like tornado, 5.3, 5.2, wrs, ... -- ---------------------------------------------------- Chris Pinkard Sr. Staff SW Engineer KLA-Tencor RAPID Engineering 160 Rio Robles C2-3400 San Jose, CA 95161-9055 cpinkard@falcon.kla.com 408-875-4346 From AKBiocca@lbl.gov Wed Jan 7 11:12:36 1998 From: Alan K Biocca Date: Wed Jan 7 11:12:39 PST 1998 Subject: The Vxworks Exploder itself VxWorks Exploder History, Credit, Notes I would like to take this opportunity to review the history and give some credit for the effort in the VxWorks User's Group Exploder. As I recall, this all started when Sypko Andreae of Lawrence Berkeley Laboratory was president of the users group (don't have the date). As was requested at the users group meeting, he set up an exploder here at LBNL on the Computing Services mail system, a VMS VAX. After a time the popularity of this list and the difficulty of managing it as well as problems with large lists on VMS prompted us to convince him to move it to a small dedicated Unix box, a Sun 3/80. I was asked to specify the hardware, and around that time I got more involved and wrote some code to make the list work better, eventually evolving that into the code running the list today. Sypko retired and maintenance of the system fell to Mike Bordua, Gary Jung (our unix administrators), and I. Mike finally retired the 3/80 and moved the code over to the Sparc. He and Gary continue to manage this machine, and I the programs. The larger part of the work is in actually adding and deleting names from the list. This is alternated between Bill Brown and Loren Shalz. Thanks to Bill, Loren, Gary, and Mike for their support! Also thanks to R.P. Singh, the Software Engineering Department Head who funds the machine it runs on, and to Advanced Light Source management (Ben Feinberg, Henry Lancaster et al) who allow us to spend a few minutes on it. Alan K Biocca Berkeley Labs (LBL/LBNL) Advanced Light Source Controls Section Leader From aram.nahidipour@spdg.COM Wed Jan 7 11:26:39 1998 From: aram.nahidipour@spdg.COM (Aram Nahidipour) Date: Wed Jan 7 11:26:44 PST 1998 Subject: Re: vxworks Re: ALL READ - New Exploder Filter No because the subject had vxworks From daemon@csg.lbl.gov Thu Jan 8 04:00:32 1998 From: daemon@csg.lbl.gov Date: Thu Jan 8 04:00:36 PST 1998 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Thu Jan 8 04:00:17 PST 1998 Subject: NFS for Windows NT 4.0 ??? ------------------------------------------------------- Newsgroups: comp.os.vxworks,comp.realtime,comp.os.ms-windows.nt.admin.networking,comp.os.ms-windows.nt.software.services Subject: NFS for Windows NT 4.0 ??? Date: Wed, 07 Jan 1998 13:16:37 +0100 From: Maciek Czech Organization: PHU ControlSoft Message-ID: <34B37225.1E656C2D@wawel.cyfronet.krakow.pl> Has anybody such a software ? Where can I buy it ? Generally I need to mount NT4.0 hard drive on some Unix-like system (VxWorks). Thanks for help. Maciek --------------------------- End of New-News digest ********************** From mea@mclean.sparta.com Thu Jan 8 07:14:41 1998 From: "Mike Anderson" Date: Thu Jan 8 07:14:45 PST 1998 Subject: Re: NFS for Windows NT 4.0 ??? >From: Maciek Czech >Organization: PHU ControlSoft >Message-ID: <34B37225.1E656C2D@wawel.cyfronet.krakow.pl> > >Has anybody such a software ? Where can I buy it ? > >Generally I need to mount NT4.0 hard drive on some Unix-like system >(VxWorks). > Yes. FTP Software, Inc. makes such a creature for NT 4.0 Server. It allows you to mount the NT drives with any standard NFS implementation. I haven't tried VxWorks specifically, but it *should* work. HTH, --------------------------------------------------------------------- |Michael E. Anderson | Voice -> 703 448 1683 x235 | |Chief Engineer | FAX -> 703 893 5494 | |mea@mclean.sparta.com | http://www.mclean.sparta.com | |___________________________|_______________________________________| | SPARTA, Inc. | | | /\\ C4I Operation |"Software Development is | | ///\\ 7926 Jones Branch Drive | like making a baby... | | /////\\ Suite 900 | You can't make a baby in | | ///// \\ McLean, VA 22102 | one month by impregnating | | / \\\\\ \\ 703-448-0210 | nine women. | | \ \\\\\ // | | | \ ////// | Some things just take | | \////// SPARTA | time." | | \//// Pride In Performance | | | \// -------------------- | | --------------------------------------------------------------------- From wong@vms.masd.northgrum.com Thu Jan 8 08:41:56 1998 From: wong@vms.masd.northgrum.com (J. Pierre Wong) Date: Thu Jan 8 08:42:00 PST 1998 Subject: Uniform Driver Interface (UDI) Greetings, Has anyone ever heard of Uniform Driver Interface (UDI)? Is it an addition to POSIX? Or how is it different? Does anyone know if WRS plans on implementing the Uniform Driver Interface (UDI) into vxworks? Thanks, Pierre Wong wong@vms.masd.northgrum.com From Eli_Wittner@cti-pet.com Thu Jan 8 11:29:59 1998 From: "Eli Wittner" Date: Thu Jan 8 11:30:03 PST 1998 Subject: Re: NFS for Windows NT 4.0 ??? >From: Maciek Czech >Organization: PHU ControlSoft >Message-ID: <34B37225.1E656C2D@wawel.cyfronet.krakow.pl> > >Has anybody such a software ? Where can I buy it ? > >Generally I need to mount NT4.0 hard drive on some Unix-like system >(VxWorks). > I'm not sure if this is what you are asking for, but we use a shareware product called TropicNFS to mount an NT 4.0 hard drive partition on vxWorks. We tried several others, but this one works the best and is <$20.00 per copy. Check out http://www.shareware.com Hope this helps. Eli Wittner CTI Cyclotron Systems, Inc. From hawthorne@aptec.com Thu Jan 8 11:39:06 1998 From: "James B. Hawthorne" Date: Thu Jan 8 11:39:10 PST 1998 Subject: Re: NFS for Windows NT 4.0 ??? the vxWorks Users Group Exploder wrote: >From: Maciek Czech >Organization: PHU ControlSoft >Message-ID: <34B37225.1E656C2D@wawel.cyfronet.krakow.pl> > >Has anybody such a software ? Where can I buy it ? > >Generally I need to mount NT4.0 hard drive on some Unix-like system >(VxWorks). > Hummingbird Communications, Ltd (http://www.hcl.com) has a product called Maestro NFS Server (http://www.hummingbird.com/products/maestro/) that should do just fine. I have used Hummingbird products in the past and found them to do tremendously well. The Maestro NFS Client package worked real well at my previous place of employment. Sun also has has a product called PCNFS Pro. However, the last time I used this product it was not available for the NT platform and did not run well on Windows 95 along side of the Microsoft network client. Best Regards, Jim ================================================================= James B. Hawthorne | voice: (505)767-1228 Realtime Systems Engineer | fax: (505)768-1379 Applied Technology Associates | 1900 Randolph Road SE | email: hawthorne@aptec.com Albuquerque, NM 87106 | http://www.aptec.com ================================================================= From mmcquade@slave2.aa.net Thu Jan 8 14:06:41 1998 From: Mike McQuade Date: Thu Jan 8 14:06:45 PST 1998 Subject: Re: NFS for Windows NT 4.0 ??? At 11:39 AM 1/8/98 PST, you wrote: >>From: Maciek Czech >>Organization: PHU ControlSoft >>Message-ID: <34B37225.1E656C2D@wawel.cyfronet.krakow.pl> >> >>Has anybody such a software ? Where can I buy it ? vxworks We use Diskshare from Intergraph, we use it on a nearly daily basis to write data from our VxWorks targets (mv2600, VxWorks 5.3.1) to PC running NT 4.0 sp 3. The software is easy to install, runs great, and you can get a fully functional demo from: http://www.intergraph.com/iss/products/prodtool/nfs/ Mike From daemon@csg.lbl.gov Fri Jan 9 04:00:36 1998 From: daemon@csg.lbl.gov Date: Fri Jan 9 04:00:41 PST 1998 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Fri Jan 9 04:00:30 PST 1998 Subject: What's VxWorks Subject: Re: VxWorks in Flight Software? ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: What's VxWorks Date: 8 Jan 1998 13:22:12 GMT From: pdumon@eduserv1.rug.ac.be (Pieter Dumon) Organization: University of Ghent, Belgium Message-ID: <692ju4$ff0$1@inf6serv.rug.ac.be> I never heard of VxWorks before visiting this newsgroup. I know it's a shame, but if anyone could give me a nice introduction ... - -- ######################################## # # # Pieter Dumon # # # # Pieter.Dumon@rug.ac.be # # # # http://studwww.rug.ac.be/~pdumon # ######################################## --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks in Flight Software? Date: Thu, 08 Jan 1998 13:22:53 -0800 From: Matt Cosby Organization: DERA Message-ID: <34B543AD.140D@scs.dra.hmg.gb> References: <903015884273D111A8FA00805FCCEB4319FB67@fl51m01.space.honeywell.com> <34AA941E.6FCF@csi.jpl.nasa.gov> <34B3C699.167E@nist.gov> Reply-To: m_cosby@scs.dra.hmg.gb - -- We at DERA, in the UK, are using VxWorks for the control software for an experiment on-board our STRV 1d satellite. This is scheduled to be launched in 1999. Regards Matt. - ------------------------------------------------ Matt Cosby SpaceCraft Environment & Protection, Space Department, Room G081, Arthur C Clarke Building, DERA Farnborough, Hants, UK. GU14 0LX. 44-(0)1252-396313 - ------------------------------------------------ --------------------------- End of New-News digest ********************** From daemon@csg.lbl.gov Sat Jan 10 04:02:12 1998 From: daemon@csg.lbl.gov Date: Sat Jan 10 04:02:15 PST 1998 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sat Jan 10 04:02:10 PST 1998 Subject: VxWorks/Tornado for DSP Subject: VxWorks/Tornado for DSP Subject: Re: File I/O problems Subject: TSR 16443 - MV167 VME TAS patch. Subject: [Help] VxWorks network !! ------------------------------------------------------- Newsgroups: comp.os.vxworks,comp.realtime,comp.dsp,sci.engr.control Subject: VxWorks/Tornado for DSP Date: Fri, 09 Jan 1998 18:37:13 +0100 From: Friedrich Scheurer / ZFS Organization: Zentrum Fertigungstechnik Stuttgart Message-ID: <34B66049.4D72@zfssun6.tz.rus.uni-stuttgart.de> References: <34B2BB9A.81C3F459@integralaccess.com> Hi, Is there anyone who already uses VxWorks/TORNADO for DSPs? What state of development has WRS at this point? Are there any experiences? What platforms are supported? Who offers complete CPU-I/O-Systems? F. Scheurer +--------------------------------------------+ | | | Dipl.-Ing. F. Scheurer | | | | c/o ZFS | | Zentrum Fertigungstechnik Stuttgart | | Center for Production Technology Stuttgart | | Nobelstrasse 15 | | D-70569 Stuttgart | | | | Telefon +49 711 13162-41(-0) | | Fax +49 711 13162-11 | | E-Mail scheurer@zfs.rus.uni-stuttgart.de | | | +--------------------------------------------+ | | | privat: | | Telefon +49 7161 686362 | | Fax +49 7161 686301 | | E-Mail Friedrich.Scheurer@t-online.de | | | +--------------------------------------------+ --------------------------- Newsgroups: comp.os.vxworks Subject: VxWorks/Tornado for DSP Date: Fri, 09 Jan 1998 19:28:37 +0100 From: Friedrich Scheurer / ZFS Organization: Zentrum Fertigungstechnik Stuttgart Message-ID: <34B66C55.298A@zfssun6.tz.rus.uni-stuttgart.de> References: <34B2BB9A.81C3F459@integralaccess.com> <34B66049.4D72@zfssun6.tz.rus.uni-stuttgart.de> Hi, Is there anyone who already uses VxWorks/TORNADO for DSPs? What state of development has WRS at this point? Are there any experiences? What platforms are supported? Who offers complete CPU-I/O-Systems? F. Scheurer +--------------------------------------------+ | | | Dipl.-Ing. F. Scheurer | | | | c/o ZFS | | Zentrum Fertigungstechnik Stuttgart | | Center for Production Technology Stuttgart | | Nobelstrasse 15 | | D-70569 Stuttgart | | | | Telefon +49 711 13162-41(-0) | | Fax +49 711 13162-11 | | E-Mail scheurer@zfs.rus.uni-stuttgart.de | | | +--------------------------------------------+ | | | privat: | | Telefon +49 7161 686362 | | Fax +49 7161 686301 | | E-Mail Friedrich.Scheurer@t-online.de | | | +--------------------------------------------+ --------------------------- Newsgroups: comp.os.vxworks Subject: Re: File I/O problems Date: 9 Jan 98 18:20:15 GMT From: georg@sgl.crestech.ca (Georg Feil) Organization: Centre for Research in Earth and Space Technology Message-ID: References: <691co9$sn4@clarknet.clark.net> <34B53331.6725@most.fw.hac.com> "Troy A. Moon" writes: >Jeff wrote: >> >> I have a nagging File I/O problem that I can't seem to pinpoint. The >> problem occurs during file close() and/or fclose() function calls that cause >> memPartFree (invalid block) errors, and subsequently suspends the offending >> task. Using the stack trace, I can see that the file descriptors/pointers >> are valid, and they're listed when I use iosFdShow from the shell (after the >> error). >> >> I have about a dozen tasks in my app; about 8 perform file I/O. Every file >> I/O task is equally capable of generating the error, and there's no >> commonality among the files being closed. There's no pattern specific to >> any event or task that causes this error; at least none that I can find. >> >> What's particularly annoying is that the error is very infrequent, on the >> order of 1 "crash" per month. >> >> More details; >> >> VxWorks 5.1.1 >> MVME167 SBC >> 4Gig SCSI hard drive >> dosFs filesystem >> >> I would appreciate any suggestions on debugging this problem. Thanks. >> >> Jeff Clements >> Embedded Systems Design, Inc. >> jeff.clements@embedded-sys.com While this could very well be caused by memory overwrites, we had a very similar problem back in '94 that was caused by a VxWorks bug. More exactly, it's a CPU hardware bug that the VxWorks BSP didn't handle right. Back then we had mv167 BSP version 1.0 (last change to sysLib.c was 02d,27feb93,ccc), so if your BSP is newer this bug may already be fixed. Have a look, if so this should be obvious. Implementing the following patch solved our problem. This applies only to VxWorks 5.1.1! Good luck, Georg Date: Tue, 5 Apr 1994 14:32:11 -0400 From: Jerald Pendleton Subject: TSR 16443 - MV167 VME TAS patch. Status: RO ] Date: 03-02-1994 ] Submitted by: Vinai Kolli ] ] ] Sub: Configuration changes ] which provide workaround for the ] back plane problem for the motorolla mvme162. ] ----------------------------------------------------------------------------- ] ] Due to snoop problems with the Test And Set (TAS) command from the ] VMEbus the following modifications are necessary until the TAS ] command can be snooped from the VMEbus. The shared memory pool ] will be defined in low memory, and will be marked as non-cacheable ] by the local processor. The VMEchip will have the snoop control ] bits turned off. This will not effect the cache and snooping for ] the on-board devices such as ethernet and SCSI, as snooping will ] still be turned on for those devices. ] The size of the memory required for the shared memory should be a multiple ] of 8k which is the default page size for the mmu. ] Any interprocessor communication which is done through a shared memory ] should make sure that the cache is disabled for that shared memory location. ] These changes would be in effect until motorolla fixes the snooping ] problems. The following changes apply to mvme162 and mvme167 ] ] File Changes ] ------------------------------------------------------------------------------ ] config.h define RAM_LOW_ADRS as 0x20000 ] define SM_MEM_SIZE as 0xe000 (64k -8k) ] define SM_OBJ_MEM_SIZE as 0x10000 (64k) ] define SM_MEM_ADRS as ] RAM_LOW_ADRS - (SM_MEM_SIZE + SM_OBJ_MEM_SIZE) ] define SM_OBJ_MEM_ADRS as ] RAM_LOW_ADRS - SM_OBJ_MEM_SIZE ] Specifically to mvme167 the macro ] USER_D_CACHE_MODE should be defined as ] (CACHE_COPYBACK | CACHE_SNOOP_ENABLE) ] ] MakeSkel change RAM_LOW_ADRS to 0x20000 ] ] sysLib.c In the function sysHwInit2() add a function code ] shown below to the end of the function which disables ] the cache for the pages dedicated to the shared ] memory. ] ] void sysHwInit2 (void) ] { ] ] .... ] .... ] .... ] ] #ifdef INCLUDE_MMU_BASIC ] (void) vmBaseStateSet (NULL, (void *)SM_MEM_ADRS, ] SM_MEM_SIZE + SM_OBJ_MEM_SIZE, ] VM_STATE_MASK_CACHEABLE, ] VM_STATE_CACHEABLE_NOT); ] #endif /* INCLUDE_MMU_BASIC */ ] ] #ifdef INCLUDE_MMU_FULL /* unbundled mmu product */ ] (void) vmStateSet (NULL, (void *)SM_MEM_ADRS, ] SM_MEM_SIZE + SM_OBJ_MEM_SIZE, ] VM_STATE_MASK_CACHEABLE, ] VM_STATE_CACHEABLE_NOT); ] #endif /* INCLUDE_MMU_FULL */ ] } ] ] sysLib.c In the function sysProcNumSet() inhibit the snooping ] by programming the snoop bits in the vmechip2 as ] shown below ] change the following Macros ] from to ] ----------------------------------------------- ] (write sink data) (snoop inhibit) ] VSAMSR2_SNP_WSD VSAMSR2_SNP_INHIBIT ] VSAMSR1_SNP_WSD VSAMSR1_SNP_INHIBIT ] ] ] - -- Georg Feil | http://www.sgl.crestech.ca/ Space Geodynamics Laboratory | Email: georg@sgl.crestech.ca CRESTech | Phone: (416) 665-5458 4850 Keele St./North York/Ont/Canada/M3J 3K1 | Fax: (416) 665-1815 --------------------------- Newsgroups: comp.os.vxworks Subject: [Help] VxWorks network !! Date: Fri, 9 Jan 1998 14:00:39 -0500 From: "Huy M.T.T Nguyen" Organization: Raytheon Electronic Systems Message-ID: References: <01bd194c$8cd57450$fa35afce@marcia> Hi All !! I am just beginning to learn how to use Vxworks functions for my project. Therefore, I want to ask all of you to help or give me some directions of using VxWorks Here is my task... There is a server machine that is running under VxWorks operation system. And my job is write an interface GUI to get all the data from users, then send that data into the server by either TCP/IP or UDP. I know the IP address of the server. So I can make a connection from my workstation to that server and send all the data. BUT I have a problem that How can I know that the server get all the data correctly and How can I get that data from the server.. Thus, I can compare my input with the output of the server. What kind of functions of VxWorks should I use ? and How can I get the data from the server side ?? Thanks !! /-_-\ / / \ / / \ @}-,'-- Huy M.T.T Nguyen --,'-{@ \ \ / \__\__/ Raytheon Electronic System, Inc. \\ --Seasparrow Missile Team -- -\\ ____ \\ / / " Life is too short to waste, better have fun ____ \\/___/ while you still can. " \ \ -// \___\//- -// \\ --------------------------- End of New-News digest ********************** From mam@pluto.dspt.com Sat Jan 10 10:24:58 1998 From: "Mark Menge" Date: Sat Jan 10 10:25:01 PST 1998 Subject: Searching for DOSFS workarounds, bandages, tourniquets We have been seeing more flakiness with vxWorks' DOSFS. Sometimes a disk will get into a reproduceable state where several k of memory is lost after creating and closing a new file. Reformatting is the only cure. I understand it normally loses about 20 bytes for each new file on the disk. Since our system can write lots of data we have cases where our memory monitor shows the memory getting quickly sucked up 1. Has anyone seen this? 2. Does anyone have source code to check if files are cross linked? I heard if the system is turned off with a open file, you could get cross linked files. 3. Anyone else know of other nonboard specific gotchas and fixes? From mabrouk@nortel.ca Sun Jan 11 10:55:05 1998 From: "Mabrouk Aboughanaima" Date: Sun Jan 11 10:55:08 PST 1998 Subject: MPC860 Interrupts Hi All, I'm developing a C++ application using vxWorks 5.3.1 for a PowerPC860 platform. I am trying to connect an ISR to handle an external interrupt which is a level interrupt but I have no luck yet. When I invoke the function that connects the ISR fron windSh or CrossWind, I neve get control back, it seems the system is too busy doing something even though the exeternal device's interrupt is not enabled yet. When I enable the exeternal device interrupt and then invoke the function that connect the ISR I get the message lost connection to target from windSh. Here is the code. //function that clears/sets the external interrupt. void setInterrupt(int state) { unsigned char data = 0; int ret = -1; if ( state == 1) { if (( !read(0x59, data))) //read control register { data |= (1 << 1); ret = write(0x19, data); } } else { if (!read(0x59, data)) //read control register { data &= (~(1 << 1)); ret = write(0x19, data); } } logMsg("\nin setInterrupt(return) %d\n", ret, 0,0,0,0,0); } void isr(void) { unsigned long pimmr = vxImmrGet(); setInterrupt(0); //clear interrupt *SIPEND(pimmr) |= SIEL_ED4; // reset the SIPEND bit corressponding to IRQ4 logMsg("\nin isr\n", 0, 0,0,0,0,0); } int initISR() { unsigned long pimmrbase = vxImmrGet(); if ( intConnect(IV_IRQ4, (VOIDFUNCPTR) isr, 0) == OK ) { *SIEL(pimmrbase) &= ~SIPEND_IRQ4; *SIMASK(pimmrbase) |= SIMASK_IRM4; //This is where it seems to be stuck at return 0; } else { printf("\nintConnect failed\n"); return -1; } } Any help will be greatly apprecaited. Please e_mail or call me. Thanks a million. ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ + Mabrouk Aboughanaima + Tel: (613) 763-6000 + + CDMA Radio Software + ESN: 393-6000 + + Nortel Wireless Networks + FAX: (613) 763-2686 + + P.O. Box 3511, Station 'C' + Internet: mabrouk@nortel.ca + + Ottawa K1Y 4H7 Canada + + ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ From daemon@csg.lbl.gov Tue Jan 13 04:00:51 1998 From: daemon@csg.lbl.gov Date: Tue Jan 13 04:00:56 PST 1998 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Tue Jan 13 04:00:49 PST 1998 Subject: POSIX Threads Subject: RogueWave and VxWorks Subject: Re: RPC Between Windows NT 4.0 and VxWorks 5.3.1 ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: POSIX Threads Date: Mon, 12 Jan 1998 06:40:18 -0500 From: "John A. Koback" Organization: Erol's Internet Services Message-ID: <69cvcb$6na$1@winter.news.erols.com> Can anyone tell me if VxWorks supports a Posix Threads C++ implementation. Are there any other news groups or sites to visit that will have this info. Anything information that you could provide would be appreciated. Thanks John (jakoback@erols.com) --------------------------- Newsgroups: comp.os.vxworks Subject: RogueWave and VxWorks Date: 12 Jan 1998 07:25:40 -0500 From: waldbuss@tif252.ed.ray.com (Lorraine Waldbusser) Organization: Raytheon Company. Message-ID: Reply-To: waldbuss@ed.ray.com Sender: waldbuss@tif252.ed.ray.com Hi, I'm interested in finding out if anyone has ever tried to use RogueWave on VxWorks. If so, what C++ compiler were you using and what issues/problems you had with the attempt. Thanks. Lorraine Cash - -- ************************************************************* * Lorraine L. Cash Raytheon Electronic Systems * * Software Engineering Lab * * RAPID Laboratory * * Tel: (508)440-3236 528 Boston Post Road * * (508)490-2574 Sudbury, MA 01776 * * Email: waldbuss@ed.ray.com * ************************************************************* --------------------------- Newsgroups: comp.os.vxworks Subject: Re: RPC Between Windows NT 4.0 and VxWorks 5.3.1 Date: Mon, 12 Jan 1998 15:50:28 -0800 From: "John Finley" Organization: Netcom Message-ID: <69eavu$st2@sjx-ixn8.ix.netcom.com> References: <696dhm$7q3$1@ha2.rdc1.az.home.com> Brian Tucker wrote in message <696dhm$7q3$1@ha2.rdc1.az.home.com>... >I'm fairly new to VxWorks OS so I have no idea what RPC services it >provides. I have looked over the manuals a little, but I have to say they >are lacking in information about RPC. I have done several projects using >RPC in the NT environment, but now I need a way to connect a NT box with >several VxWorks boards. I was hoping that someone out there has done the >same and could provide a little sample code that I might be able to look >over. I think you need ONC RPC, not the Microsoft flavor. There is a free ONC RPC for NT at http://set.gmd.de/~mfg/oncrpc.htm which also has some links to commercial products. John - -------------------------------------------------- John Finley Engineering Consultant john@finley.com Real Time Systems (619) 689-0032 Networking Protocols - -------------------------------------------------- --------------------------- End of New-News digest ********************** From bobperry@bellatlantic.net Tue Jan 13 05:09:40 1998 From: "Robert L. Perry" Date: Tue Jan 13 05:09:51 PST 1998 Subject: 860 Interrupts One possible problem could be that the cause of the interrupt is not being reset -writing to the SIPEND register doesn't do anything (see p.12-5 section 12.2.1.3.1) Perhaps your 860 is interrupt'n like crazy. We had alot of trouble w/ CPM interrupt support on the 860, here was one find: many bits in the 860 registers are reset when a '1' is written to the bit position of interrest. One example on how *not* to access the CISR register can be found in ppc860CpmIntrDemux() in ppc860Intr.c /* clear the bit in the CISR */ #if 1 * CISR(regBase) = (1 << intNum); /* This is correct */ #else * CISR(regBase) |= (1 << intNum); /* This is not correct */ #endif The OR operation above appears to be the straigth forward approach to writing a '1' to the proper bit. The problem is that the compiler will read the CISR value, OR the proper bit, then perform a write. If more than one bit is set in the CISR, they will all be reset after the |= operation, which in this case is very sad. We visited ppc860Intr.c while getting all of the scc's operational and there is a TSR open w/ WindRiver -to date, our changes have not been blessed. If anyone is interrested, please contact me directly as I don't know what the rules are regarding code segments postings from copyright material. Bob Perry bobperry@bellatlantic.net At 10:55 AM 1/11/98 PST, you wrote: >Submitted-by mabrouk@nortel.ca Sun Jan 11 10:55:05 1998 >Submitted-by: "Mabrouk Aboughanaima" > >Hi All, > >I'm developing a C++ application using vxWorks 5.3.1 for a PowerPC860 >platform. I am trying to connect an ISR to handle an external interrupt which is >a level interrupt but I have no luck yet. When I invoke the function that connects >the ISR fron windSh or CrossWind, I neve get control back, it seems the system is too busy doing something >even though the exeternal device's interrupt is not enabled yet. > >When I enable the exeternal device interrupt and then invoke the function that connect the ISR >I get the message lost connection to target from windSh. > >Here is the code. > >//function that clears/sets the external interrupt. > >void setInterrupt(int state) >{ > unsigned char data = 0; > int ret = -1; > > if ( state == 1) > { > if (( !read(0x59, data))) //read control register > { > > data |= (1 << 1); > > ret = write(0x19, data); > } > > } > else > { > if (!read(0x59, data)) //read control register > { > > data &= (~(1 << 1)); > > ret = write(0x19, data); > } > > } > logMsg("\nin setInterrupt(return) %d\n", ret, 0,0,0,0,0); >} > >void isr(void) >{ > > unsigned long pimmr = vxImmrGet(); > > setInterrupt(0); //clear interrupt > > *SIPEND(pimmr) |= SIEL_ED4; // reset the SIPEND bit corressponding to IRQ4 > > > logMsg("\nin isr\n", 0, 0,0,0,0,0); > >} > > >int initISR() >{ > unsigned long pimmrbase = vxImmrGet(); > > > if ( intConnect(IV_IRQ4, (VOIDFUNCPTR) isr, 0) == OK ) > { > *SIEL(pimmrbase) &= ~SIPEND_IRQ4; > > *SIMASK(pimmrbase) |= SIMASK_IRM4; //This is where it seems to be stuck at > return 0; > } > else > { > printf("\nintConnect failed\n"); > return -1; > } > >} > > > > >Any help will be greatly apprecaited. Please e_mail or call me. > >Thanks a million. > >++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ >+ Mabrouk Aboughanaima + Tel: (613) 763-6000 + >+ CDMA Radio Software + ESN: 393-6000 + >+ Nortel Wireless Networks + FAX: (613) 763-2686 + >+ P.O. Box 3511, Station 'C' + Internet: mabrouk@nortel.ca + >+ Ottawa K1Y 4H7 Canada + + >++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ > > From bf@kayser-threde.de Tue Jan 13 06:01:50 1998 From: Florian Bittner Date: Tue Jan 13 06:01:58 PST 1998 Subject: Tornado PPC2604 and FDDI Hello, has anybody out there used any type of FDDI board together with the Motorola MVME 2604 yet. Preferably I would like to hear about PCI boards and the performance of tcp/ip transfers. Thank you, Florian Bittner Software Engineer __ __ Kayser-Threde GmbH Tel +49 341 9125230 | |/ / Dessauerstr.9 Fax +49 341 9122160 | ====< 04129 Leipzig Germany |__|\__\ From ptp@mclean.sparta.com Tue Jan 13 06:46:42 1998 From: "Patrick T. Pinkowski" Date: Tue Jan 13 06:46:45 PST 1998 Subject: Re: Tornado PPC2604 and FDDI > has anybody out there used any type of FDDI board together with the > Motorola MVME 2604 yet. Preferably I would like to hear about PCI > boards > and the performance of tcp/ip transfers. > Florian, We (SPARTA, Inc.) have a VxWorks Device Driver for the MVME-2604 on the Interphase 5211 VME Board. We also have a driver working for the Motorola MVME 2604 on the Interphase 4511 PMC/PCI Card, but it would take some work to get that fully optimized. The performance on the VME Board is 7 MBytes per Second using tcp/ip transfers. Regards Pat -------------------------------------------------- /\ Patrick T. Pinkowski, Senior Engineer /##\ ptp@mclean.sparta.com /####\ 703.448.1683 x228 /####/-- /####/---- SPARTA, Inc. -\####\__--- 7926 Jones Branch Drive, Suite 900 ---\##\ /---- McLean, Virginia 22102 ----\##\/\---- (703)448-0210 (Main) ----\####\-- (703)893-5494 (Facsimile) ----\####/ http://www.mclean.sparta.com ---/###/ -/###/ SPARTA \##/ ~~~~~~ \/ Pride In Performance From mea@mclean.sparta.com Tue Jan 13 07:44:53 1998 From: Mike Anderson Date: Tue Jan 13 07:44:57 PST 1998 Subject: RE: Tornado PPC2604 and FDDI On Tuesday, January 13, 1998 9:02 AM, Florian Bittner wrote: > Submitted-by: Florian Bittner > > Hello, > > has anybody out there used any type of FDDI board together with the > Motorola MVME 2604 yet. Preferably I would like to hear about PCI boards > and the performance of tcp/ip transfers. > Yes, we have a device driver for the Interphase 5211 FDDI card working with the 2604. Sustained data rates are in the 7+ MBytes/sec for TCP transfers. We did have quite a few problems with the original port of the driver to the 2604. Most of these problems were related to bugs both on the board (hardware -- the Universe chip has some major flaws), in the BSP (the PCI Raven chip support) and in WRS's PPC port. After a couple of months of working with both Motorola and WRS, we were finally able to get everything working. But, the 2604 exhibits some really bizzare interrupt latency problems related to the PCI <-> VME interface. Most notably related to Ethernet activity (via the DEC Ethernet chip) causing VME interrupts to be deferred for 50-70 us. So, caveat emptor on the 2604. BTW, our driver for the 5211 also supports SMT 7.3 and we provide you the software to burn the flash on the 5211 so you can establish a baseline firmware release for any systems you build :-). We also have a driver for the Interphase 4511 PMC FDDI card. We have tested it on the 2604 with good success. However, the firmware setup for the 4511 is such that it works well for receiving packets but not so well for sending packets. This is supposed to be fixed with a forthcoming firmware release, but we haven't seen it yet ;-). HTH, --------------------------------------------------------------------- |Michael E. Anderson | Voice -> 703 448 1683 x235 | |Chief Engineer | FAX -> 703 893 5494 | |mea@mclean.sparta.com | http://www.mclean.sparta.com | |___________________________|_______________________________________| | SPARTA, Inc. | | | /\\ C4I Operation |"Software Development is | | ///\\ 7926 Jones Branch Drive | like making a baby... | | /////\\ Suite 900 | You can't make a baby in | | ///// \\ McLean, VA 22102 | one month by impregnating | | / \\\\\ \\ 703-448-0210 | nine women. | | \ \\\\\ // | | | \ ////// | Some things just take | | \////// SPARTA | time." | | \//// Pride In Performance | | | \// -------------------- | | --------------------------------------------------------------------- From cruz_nojunk@xyplex.com Tue Jan 13 09:10:59 1998 From: Roger Cruz Date: Tue Jan 13 09:11:03 PST 1998 Subject: Re:Searching for DOSFS workarounds, bandages, tourniquets > I've had a lot of problems with the DOS FS. The latest is the cross-linked files. I've just installed a new patch from WRS. SPR #8632 which they claim fixes. I can't say that it does for sure because I haven't run it long enough to test it. Please keep me posted on ANY news you find about the DOS FS lib. Remove "_nojunk" from my email address. > >From mam@pluto.dspt.com Sat Jan 10 10:24:58 1998 > From: "Mark Menge" > Date: Sat Jan 10 10:25:01 PST 1998 > Subject: Searching for DOSFS workarounds, bandages, tourniquets > > We have been seeing more flakiness with vxWorks' DOSFS. > > Sometimes a disk will get into a reproduceable state where several k of memory is > lost after creating and closing a new file. Reformatting is the only > cure. I understand it normally loses about 20 bytes for each new file on the disk. > Since our system can write lots of data we have cases where our memory monitor shows the > memory getting quickly sucked up > > 1. Has anyone seen this? > > 2. Does anyone have source code to check if files are cross linked? > I heard if the system is turned off with a open file, you could get > cross linked files. > > 3. Anyone else know of other nonboard specific gotchas and fixes? > -- Roger Cruz cruz_nojunk@xyplex.com Xyplex Networks w: 508-952-4783 295 Foster Street f: 508-952-4887 Littleton, MA 01460 From rtp.co.uk!jge@rtp.co.uk Tue Jan 13 09:11:25 1998 From: John Efstathiades Date: Tue Jan 13 09:11:29 PST 1998 Subject: RE: Tornado PPC2604 and FDDI >Submitted-by bf@kayser-threde.de Tue Jan 13 06:01:50 1998 >Submitted-by: Florian Bittner > >Hello, > >has anybody out there used any type of FDDI board together with the >Motorola MVME 2604 yet. Preferably I would like to hear about PCI boards >and the performance of tcp/ip transfers. Florian, Real Time Products has developed a VxWorks network interface driver for the Interphase 4511/5511 (PMC/PCI) FDDI cards. The driver integrates into VxWorks in the same way as an ethernet driver and allows socket applications to run over an FDDI ring. If you would like some more details please e-mail me at the address shown below. Regards, John Efstathiades ---------------------------------------------------------------- John Efstathiades Real Time Products Ltd e-mail: jge@rtp.co.uk Chancery House, tel: +44 (0)121 234 6600 8 Edward Street, fax: +44 (0)121 234 6611 Birmingham. B1 2RX. England. web: http://www.rtp.co.uk ---------------------------------------------------------------- From JOSIPDI@euromsx.gemse.fr Tue Jan 13 09:19:00 1998 From: "Josipovici, Diane (MED)" Date: Tue Jan 13 09:19:04 PST 1998 Subject: EST Cpu32 backend for Tornado 1.0 Hi vxworkers ! I am trying to use an EST emulator to debug a custom board, MC68360 based. I have the EST Cpu32 backend for Tornado 1.0, and installed it for HPUX V9 host. But now, Tornado doesn't communicate with the estcpu32 backend. error message is : Attaching backend... Error: Can't open tornado/host/parisc-hpux9/lib/backend/estcpu32.sl file failed. Error: No backend initialization routine. I verified that it was not a problem of acces to the file estcpu32.sl (it is corectly installed). And the answer of EST support is that they don't support any more estcpu32 backend for Tornado 1.0, because windRiver itself consider it as obsolete !!! So I can't find this old release from EST to verify if I have the correct file. Has anyone an idea to solve this problem ? Thanks Diane > g GE Medical Systems > _____________________________________________ > Diane JOSIPOVICI - mailto:Diane.Josipovici@med.ge.com > Software Engineer - Global X-Ray Engineering > 283, rue de la Miniere, BP 34 > 78533 Buc Cedex, France > Tel: 33 (0)1 30 70 98 29 > Fax: 33 (0)1 30 70 46 70 > > From currie@ghg.net Tue Jan 13 12:18:57 1998 From: currie@ghg.net (Alan Currie) Date: Tue Jan 13 12:19:01 PST 1998 Subject: Vista FDDI device driver for vxWorks Does anyone have or know of a vxWorks driver for the Vista VC-FDDI-INP2 board? Thanks. From kalalau!jmk@uu9.psi.com Tue Jan 13 21:10:07 1998 From: jmk@acut.com (Joel Katz) Date: Tue Jan 13 21:10:11 PST 1998 Subject: RPC is not RPC >Brian Tucker wrote in message <696dhm$7q3$1@ha2.rdc1.az.home.com>... >>I'm fairly new to VxWorks OS so I have no idea what RPC services it >>provides. I have looked over the manuals a little, but I have to say they >>are lacking in information about RPC. I have done several projects using >>RPC in the NT environment, but now I need a way to connect a NT box with >>several VxWorks boards. I was hoping that someone out there has done the >>same and could provide a little sample code that I might be able to look >>over. > >I think you need ONC RPC, not the Microsoft flavor. >There is a free ONC RPC for NT at >http://set.gmd.de/~mfg/oncrpc.htm >which also has some links to commercial products. > >John Just to add to the Noise, there are at least two RPC standards in use today.= The first is ONC RPC which was developed by Sun Microsystems and sits under= NFS. This is the RPC that comes with vxWorks. The other OSF/DCE RPC, was= developed by Apollo Computers(I believe) adopted by OSF as a standard, in= the spirit of not cooperating with SUN(my own comments) and later by= Microsoft. OSF/DCE RPC is the RPC package available in Windows 95 and= Windows NT 4.0. Of course Microsoft did change many of the API's from the= OSF standard ... Anyway, the easiest way to do NFS between Windows NT and vxWorks is to use= ONC RPC. We've been working with the commercial product offering from= Noblenet and have been pleased with it. Check out www.Noblenet.com for more= information on their product. Joel Katz, Principal Engineer --- (603)-577-5895 Acuity Imaging, 9 Townsend West, Nashua, NH 03063 --- (603)-577-5965 (FAX) From daemon@csg.lbl.gov Wed Jan 14 04:00:37 1998 From: daemon@csg.lbl.gov Date: Wed Jan 14 04:00:41 PST 1998 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Wed Jan 14 04:00:35 PST 1998 Subject: Re: Network problem ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Network problem Date: Tue, 13 Jan 1998 08:19:44 -0500 From: Kevin Bradley Organization: Doctoral student, Electrical and Computer Engineering, Carnegie Mellon, Pittsburgh, PA Message-ID: <4oiqbkC00iWP01TOE0@andrew.cmu.edu> References: <34BAACD7.EA64AE79@wacos.com> Excerpts from netnews.comp.os.vxworks: 13-Jan-98 Network problem by thomas Li@wacos.com > My pc486 target can be booted by using floppy disk, no any error if > using floppy disk boot it. > The config,h is set as follow: > #define DEFAULT_BOOT_LINE \ > fd=0,0(0,0)Wacosserver1:/fd0/vxWorks e=192.168.0.7 h=192.168.0.102 > u=Administrator pw=zhoudd tn=Thomas.Li s=/target/pc486/boot f=0x0" You probably also need to set a gateway parameter (g=192.168.0.xxx), which would be the same as your host's gateway, judging from your network parameters as listed. You also probably need to add a netmask so that packets are filtered properly. My PC-based system boots from the network with the following line (I have 3com 3c509 network cards, hence the ELT): #define DEFAULT_BOOT_LINE \ "elt(0,0)kalamari:pc386/vxWorks_z h=128.2.102.215 e=128.2.102.70:ffffff00 g=128.2.102.254 tn=zaxxon u=vxworks pw=vxworks o=elt" The "ffffff00" after the e command is the netmask. On my system it's 255.255.255.0, which translates to the above. If you're not sure of what the proper mask is, look it up on the NT host; it's in the same tab as the IP address. > And target server cant launch successful, the error occurs as > same as using floppy disk. I guess that is a network problem, how > can I test the network in this condition? I had similar problems when I was starting out. First, make sure the following things are enabled in your config.h / configdb.h: #define INCLUDE_PING #define INCLUDE_SHOW_ROUTINES #define INCLUDE_NET_SHOW #define INCLUDE_SHELL This will add a "ping" function to your target, so you can attempt to ping other systems once it boots up. It also adds the target- based shell, which you can use with the PC keyboard (since you say you're booting a 486 PC system, you can use this to your advantage). Finally, it adds NET_SHOW, which will display network information. Keep on booting from floppy (your first boot line) until the system is working properly. Once your system boots and you get a shell prompt, try the following: 1) "ifShow": see what network activity is occurring, and if your network card is configured properly. 2) "ping": ping your host, and see if you get any responses. 3) Manual FTP: try manually FTP'ing to the host, and see if you can communicate to the server. The directions are in one of the manuals (I think the programmer's guide). My problem was getting the netmask right. After that, all was copacetic. Hope this helps... -- Kevin --------------------------- End of New-News digest ********************** From pat.shaw@canada.cdev.com Wed Jan 14 05:59:18 1998 From: "Shaw, Pat" Date: Wed Jan 14 05:59:24 PST 1998 Subject: Macrolink Hi: Has anyone had any experience with the Macrolink MPC VME 16 card and Vxworks?. I was wondering if there are any hidden problems with using it? Cheers Patrick Shaw COMPUTING DEVICES CANADA ? ,,, (o o) -------------------------------oOO--(_)--OOo---------------------------- -------- From dlarson@etcconnect.com Wed Jan 14 08:02:55 1998 From: David Larson Date: Wed Jan 14 08:03:02 PST 1998 Subject: RE: network problem... > Excerpts from netnews.comp.os.vxworks: 13-Jan-98 Network problem by > thomas Li@wacos.com > > My pc486 target can be booted by using floppy disk, no any error if > > using floppy disk boot it. > > The config,h is set as follow: > > #define DEFAULT_BOOT_LINE \ > > fd=0,0(0,0)Wacosserver1:/fd0/vxWorks e=192.168.0.7 > h=192.168.0.102 > > u=Administrator pw=zhoudd tn=Thomas.Li s=/target/pc486/boot f=0x0" > Make sure #INCLUDE_NETWORK_INIT is set. And that usrNetInit() is being called from usrInit() is usrConfig.c OR that if #INCLUDE_NET_INIT is not defined, that you are calling usrNetInit() explicitly after you have booted from floppy. Make sure that the network device is included in the NETIF structure in /wrs/target/src/config/usrNetwork.c. You will have to give usrNetInit () a pointer to a string with the same format as your DEFAULT_BOOT_LINE that contains the name of your network device in place of "fd". If you use your current DEFAULT_BOOT_LINE, with usrNetInit, nothing will happen, because there is no network device specified. If you are using a network device that comes bundled with vxWorks, things should start working in pretty short order. If you are writing your own driver, you need to make sure that there is an entry for it in the NETIF structure, so that usrNetInit will call your attach() function to init the driver/hardware. > Keep on booting from floppy (your first boot line) until the > system is working properly. > > Once your system boots and you get a shell prompt, try the > following: > verify that the ISR for the driver is installed properly, and that the interrupt is not masked, and is getting enabled. You need to make sure that the driver is pulling packets off the wire. > 1) "ifShow": see what network activity is occurring, and if > your network card is configured properly. > > 2) "ping": ping your host, and see if you get any responses. I would suggest spawing ping as a task: -> sp ping, "192,168.0.102", 20, 0 This will keep ping () from hanging the shell if it dies. From t_wmoles@qualcomm.com Wed Jan 14 08:27:40 1998 From: Hugh Molesworth Date: Wed Jan 14 08:27:43 PST 1998 Subject: Re: 860 Interrupts Level vs. Edge When a level-triggered interrupt is received on most machines the interrupt service routine (ISR) must clear the source of the interrupt or disable that interrupt until it is known that the source of the interrupt is clear. By "source of the interrupt" I mean the hardware signal connected to the processor interrupt line (internally or externally). Normally this signal would be designed to go inactive when the ISR performs it's task otherwise when the ISR completes the interrupt will be immediately activated again and bingo the processor latches up in a ISR-ReturnFromInterrupt-ISR state until the source goes inactive (if it ever does). Most of us have probably done this at some time or other. (By the way, I'm not familiar with the 860). For this reason interrupts are often specified as edge-triggered. Most processors now correctly latch such events provided they meet minimum pulse-width requirements. The danger is that such an interrupt is used as a handshake and one gets missed, which of course can't happen :-), and this causes some process or other to hang up awaiting the next signal (the one we just missed). To avoid this most such systems would incorporate some kind of timeout checking or method to instigate a check that an interrupt had been issued but not processed, perhaps by the occasional read of a status register. You can save a lot of grief by checking for that (impossible) missed interrupt. Finally level-triggered interrupts can be missed whereas edge-triggered interrupts "can't" be missed. If a level-triggered interrupt signal goes active whilst a higher priority interrupt ISR is in progress but goes inactive before the higher-priority ISR completes then the interrupt will be missed on lots of machines because it is not latched; in this case (assuming you wanted to catch it) you should be using an edge-triggered interrupt or make sure it doesn't go away until the ISR executes. Most level-triggered interrupt systems are designed to stay active until the ISR services the interrupt. Power-fail or other fatal non-maskable interrupts are often level-triggered simply because they often never go away (and you definitely don't want to miss them ...). Hugh hugh@qualcomm.com At 05:09 AM 1/13/98 PST, you wrote: >Submitted-by bobperry@bellatlantic.net Tue Jan 13 05:09:40 1998 >Submitted-by: "Robert L. Perry" > >One possible problem could be that the cause of the interrupt is not being >reset -writing to the SIPEND register doesn't do anything (see p.12-5 >section 12.2.1.3.1) >Perhaps your 860 is interrupt'n like crazy. > >We had alot of trouble w/ CPM interrupt support on the 860, here was one >find: many bits in the 860 registers are reset when a '1' is written to the >bit position of interrest. One example on how *not* to access the CISR >register can be found in ppc860CpmIntrDemux() in ppc860Intr.c > > /* clear the bit in the CISR */ >#if 1 > * CISR(regBase) = (1 << intNum); /* This is correct */ >#else > * CISR(regBase) |= (1 << intNum); /* This is not correct */ >#endif > >The OR operation above appears to be the straigth forward approach to >writing a '1' to the proper bit. The problem is that the compiler will >read the CISR value, OR the proper bit, then perform a write. If more than >one bit is set in the CISR, they will all be reset after the |= operation, >which in this case is very sad. > >We visited ppc860Intr.c while getting all of the scc's operational and >there is a TSR open w/ WindRiver -to date, our changes have not been >blessed. If anyone is interrested, please contact me directly as I don't >know what the rules are regarding code segments postings from copyright >material. > >Bob Perry >bobperry@bellatlantic.net > >At 10:55 AM 1/11/98 PST, you wrote: >>Submitted-by mabrouk@nortel.ca Sun Jan 11 10:55:05 1998 >>Submitted-by: "Mabrouk Aboughanaima" >> >>Hi All, >> >>I'm developing a C++ application using vxWorks 5.3.1 for a PowerPC860 >>platform. I am trying to connect an ISR to handle an external interrupt >which is >>a level interrupt but I have no luck yet. When I invoke the function that >connects >>the ISR fron windSh or CrossWind, I neve get control back, it seems the >system is too busy doing something >>even though the exeternal device's interrupt is not enabled yet. >> >>When I enable the exeternal device interrupt and then invoke the function >that connect the ISR >>I get the message lost connection to target from windSh. >> >>Here is the code. >> >>//function that clears/sets the external interrupt. >> >>void setInterrupt(int state) >>{ >> unsigned char data = 0; >> int ret = -1; >> >> if ( state == 1) >> { >> if (( !read(0x59, data))) //read control register >> { >> >> data |= (1 << 1); >> >> ret = write(0x19, data); >> } >> >> } >> else >> { >> if (!read(0x59, data)) //read control register >> { >> >> data &= (~(1 << 1)); >> >> ret = write(0x19, data); >> } >> >> } >> logMsg("\nin setInterrupt(return) %d\n", ret, 0,0,0,0,0); >>} >> >>void isr(void) >>{ >> >> unsigned long pimmr = vxImmrGet(); >> >> setInterrupt(0); //clear interrupt >> >> *SIPEND(pimmr) |= SIEL_ED4; // reset the SIPEND bit corressponding to IRQ4 >> >> >> logMsg("\nin isr\n", 0, 0,0,0,0,0); >> >>} >> >> >>int initISR() >>{ >> unsigned long pimmrbase = vxImmrGet(); >> >> >> if ( intConnect(IV_IRQ4, (VOIDFUNCPTR) isr, 0) == OK ) >> { >> *SIEL(pimmrbase) &= ~SIPEND_IRQ4; >> >> *SIMASK(pimmrbase) |= SIMASK_IRM4; //This is where it seems to be >stuck at >> return 0; >> } >> else >> { >> printf("\nintConnect failed\n"); >> return -1; >> } >> >>} >> >> >> >> >>Any help will be greatly apprecaited. Please e_mail or call me. >> >>Thanks a million. >> >>++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ >>+ Mabrouk Aboughanaima + Tel: (613) 763-6000 + >>+ CDMA Radio Software + ESN: 393-6000 + >>+ Nortel Wireless Networks + FAX: (613) 763-2686 + >>+ P.O. Box 3511, Station 'C' + Internet: mabrouk@nortel.ca + >>+ Ottawa K1Y 4H7 Canada + + >>++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ > VxWorks Tornado From daemon@csg.lbl.gov Thu Jan 15 04:00:27 1998 From: daemon@csg.lbl.gov Date: Thu Jan 15 04:00:31 PST 1998 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Thu Jan 15 04:00:25 PST 1998 Subject: PPP between VxWorks and Windows NT ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: PPP between VxWorks and Windows NT Date: Tue, 13 Jan 1998 13:44:16 -0600 From: Wu Yapeng Organization: University of Regina Message-ID: I'm trying to set up a PPP connection across a serial line with a null modem between my workstation running Windows NT 4.0 and a 5.3.1 VxWorks MPC860 target. Has anyone done this and can tell me what PPP parameters were used on the VxWorks side and what was done on the Windows side to establish the link? Thanks for any help, --------------------------- End of New-News digest ********************** From moneil@ccmgate.eng.brooktrout.com Thu Jan 15 04:30:53 1998 From: moneil@brooktrout.com Date: Thu Jan 15 04:31:00 PST 1998 Subject: WindView questions with Tornado 1.0.1 Hey I am using the Tornado 1.0.1 development software and a 401gf BSP. I am attempting to get the WindView tool up and running. I do not have an ethernet port on the target. I understand that the default configuration is for ethernet, but does it have to have one or can a seriel port be used and if so has anyone accomplished this? Currently teh WindView does not work and I would like to get it running. Any help would be a great help. Thanks O'Neill From clare@bis.kodak.COM Thu Jan 15 06:17:00 1998 From: clare@kodak.com (Bill Clare) Date: Thu Jan 15 06:17:03 PST 1998 Subject: Re: WindView questions with Tornado 1.0.1 > > Hey > > I am using the Tornado 1.0.1 development software and a 401gf BSP. I > am attempting to get the WindView tool up and running. I do not have > an ethernet port on the target. > I understand that the default configuration is for ethernet, but does > it have to have one or can a seriel port be used and if so has anyone > accomplished this? > Currently teh WindView does not work and I would like to get it > running. Any help would be a great help. Unless I am mistaken you need the ethernet connection for WindView. I was not able to run until I got the network connection working. Unlike the other Tornado Tools, WindView doesn't work over the serial back end. On page 4 of the WindView 1.0.2 Release Notes, it mentions that "For continuous monitoring of event data, your target must be equipped with an Ethernet interface." It then goes on to mention that it can be over VME through another board (I was able to use WindView through a PC104-based network card attached to my x86 based CPU board). The rest of the paragraph seems to indicated that if you are willing to use the post-mortem mode then you don't need an ethernet interface. billc Bill Clare, Software Engineer, Eastman Kodak Company Email: clare@kodak.com Mail code: 2/5/EP 35425 Phone: (726)726-9419 US Mail: 901 Elmgrove Road Rochester NY 14653-5425 Fax: (716)726-0129 Any opinions expressed herein are mine and not those of Eastman Kodak Company. From espin@idiom.com Thu Jan 15 06:34:40 1998 From: Geoffrey Espin Date: Thu Jan 15 06:34:43 PST 1998 Subject: Re: comp.os.vxworks newsdigest Wu, > Subject: PPP between VxWorks and Windows NT > Date: Tue, 13 Jan 1998 13:44:16 -0600 > From: Wu Yapeng > I'm trying to set up a PPP connection across a serial line > with a null modem between my workstation running Windows > NT 4.0 and a 5.3.1 VxWorks MPC860 target. > Has anyone done this and can tell me what PPP parameters > were used on the VxWorks side and what was done on the > Windows side to establish the link? I could not get this to work, WRS Tech Support confirmed it doesn't. Microsoft's PPP (server) is incompatible with standard PPP. At best, you should be able to insert a "terminal" server, like a PortMaster, and use it to do the negotiation/conversion. Alternatively, you can use WDBSERIAL (what I'm currently using) but then you have no network (!) and are very limited to what windsh can do for you, e.g. copy a data file back and forth. Geoff -- Geoffrey Espin espin@idiom.com From kb+@andrew.cmu.edu Thu Jan 15 06:36:56 1998 From: Kevin Bradley Date: Thu Jan 15 06:36:59 PST 1998 Subject: Re: WindView questions with Tornado 1.0.1 Excerpts from mail: 15-Jan-98 WindView questions with Tor.. by the U. G. Exploder@lbl.g > I am using the Tornado 1.0.1 development software and a 401gf BSP. I > am attempting to get the WindView tool up and running. I do not have > an ethernet port on the target. > I understand that the default configuration is for ethernet, but does > it have to have one or can a seriel port be used and if so has anyone > accomplished this? I would think that as long as you can open a socket using the serial port you'd be ok. WindView as far as I can tell uses sockets for communication. As the serial port has limited bandwidth compared to a network card, I'd suggest running WindView in post-mortem mode, and being patient uploading the buffer. On my system, WindView can saturate the host uploading events in continuous mode. In post-mortem mode, I get a more accurate picture of system activity as well, as the system isn't interrupted by network activity every couple hundred milliseconds. See if you can write a small program on the host to wait on a port (e.g. 6164, which is WindView's default), and then open a socket to it from the target using the serial port. If that works, you should be in business. I don't know of a restriction in VxWorks that prevents the serial port from being uses in socket communications, so it should be ok. -- Kevin From uchenick@tis4000.com Thu Jan 15 07:37:03 1998 From: Gordon Uchenick Date: Thu Jan 15 07:37:06 PST 1998 Subject: RE: comp.os.vxworks newsdigest There was a posting from someone who got this to work. The posting was about 1-2 months ago. If I can find it at home tonight I'll send it to you. > Subject: PPP between VxWorks and Windows NT > Date: Tue, 13 Jan 1998 13:44:16 -0600 > From: Wu Yapeng > I'm trying to set up a PPP connection across a serial line > with a null modem between my workstation running Windows > NT 4.0 and a 5.3.1 VxWorks MPC860 target. > Has anyone done this and can tell me what PPP parameters > were used on the VxWorks side and what was done on the > Windows side to establish the link? From chuckvb@ix.netcom.com Thu Jan 15 10:06:38 1998 From: Chuck Van Blaricom Date: Thu Jan 15 10:06:41 PST 1998 Subject: Re: comp.os.vxworks newsdigest > Subject: PPP between VxWorks and Windows NT > Wu Yapeng Wrote: > > I'm trying to set up a PPP connection across a serial line > with a null modem between my workstation running Windows > NT 4.0 and a 5.3.1 VxWorks MPC860 target. > > Has anyone done this and can tell me what PPP parameters > were used on the VxWorks side and what was done on the > Windows side to establish the link? > > Thanks for any help, > Submitted-by espin@idiom.com Wed Nov 26 10:31:00 1997 >Submitted-by: Geoffrey Espin >I've been unsuccessful in using NT4.0 as a pppd/server for >an i960 vxworks target. > >I installed DUN/RAS, setup up a modem (pc 2 pc cable and/or 9600 modem) >on COM1 and/or COM2 ports, tested h/w connectivity (i.e. was able to see >control chars come into NT4.0 hypertrm session from the vxworks target), >so what gives!? > Microsoft developed two new authentication protocols to use with PPP. MSCHAP or CHAP80 is simillar, but incompatible to CHAP (as far as I can remember, one of them uses DES and the other uses MD5). Shiva PAP is the Microsoft equivalent of PAP. Shiva PAP uses encripted passwords while PAP uses cleartext. In order to allow vxWorks to connect to NT, some registry editing is necessary. Do a search through the registry for SPAP / Shiva PAP and delete them. Then enable PAP at the vxWorks end and try to connect. The reason that this works is that during the negotiation for authentication method, an each implementation which knows of more than one method will prioritise methods. Where a peer attempts to use a method which is known, but not preferred, an implementation may reject it in favour of a more secure authentication. By deleting the registry settings which tell NT which DLL holds Shiva PAP, the negotiation is reduced to two methods known by each peer. Once MSCHAP is rejected, NT will now agree to use PAP. Further information can be found in Linux documentation on connecting to an NT server using PPP. Regards, Pete -- Chuck Van Blaricom Communication Network Systems, 2525 Walsh Ave, Santa Clara, CA 95051 (408) 235-7039 FAX: (408) 235-7085 EMail: chuckvb@ix.netcom.com From davidg@telegate.co.il Fri Jan 16 03:32:12 1998 From: "David Gasul" Date: Fri Jan 16 03:32:17 PST 1998 Subject: Any experience with TRI's TargetDB on VxWorks? This is a multi-part message in MIME format. ------=_NextPart_000_0132_01BD2283.ADBC4960 Content-Type: text/plain; charset="iso-8859-8" Content-Transfer-Encoding: quoted-printable Hi, Does anybody have any experience with Technology Rendezvous Inc. (TRI) TargetDB database on VxWorks? Sincerely, Dave ------=_NextPart_000_0132_01BD2283.ADBC4960 Content-Type: text/html; charset="iso-8859-8" Content-Transfer-Encoding: quoted-printable
Hi,
 
Does anybody have any experience with Technology = Rendezvous=20 Inc.
(TRI) TargetDB database on = VxWorks? 
 
Sincerely,
Dave 
 
------=_NextPart_000_0132_01BD2283.ADBC4960-- From daemon@csg.lbl.gov Fri Jan 16 04:00:20 1998 From: daemon@csg.lbl.gov Date: Fri Jan 16 04:00:30 PST 1998 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Fri Jan 16 04:00:18 PST 1998 Subject: Porting GNAT to vxWorks 5.3.1 ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Porting GNAT to vxWorks 5.3.1 Date: Thu, 15 Jan 1998 11:36:42 -0800 From: Javier Leon Organization: trw Message-ID: <34BE654A.841D0250@elrond.sp.trw.com> I am trying to run an Ada benchmark that will be given to us from a potential customer on our Motorola PowerPC 603e board that is running vxWorks. I am also trying to do this on $0 budget on a very short time frame. Has anybody ported GNAT to run on a PowerPC running GNAT ???? Can you help me ?? - -- - - Thank You - >>>>>>>>>>>>>>>>>>>>>>>>><<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<< Javier Leon TRW Software Staff Engineer Avionics Systems Engineering R4/2020 Phone: (310) 814-6826 One Space Park fax : (310) (310)814-4941] Redondo Beach, CA 90728 jleon@elrond.sp.trw.com Sun (preferred) Javier.Leon@trw.com Quickmail >>>>>>>>>>>>>>>>>>>>>>>>><<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<< --------------------------- End of New-News digest ********************** From BOOJ@kidd.Co.Za Fri Jan 16 05:08:39 1998 From: Johan Boot Date: Fri Jan 16 05:08:45 PST 1998 Subject: Communication interface Dear Reader We are developing target hardware around a PowerPC processor, using Tornado. We have a serial line available for the target connection to the host, but are also looking for something with more bandwidth. The harware design does not provide for DMA transfer or PCI, which rules out (I think), the use of ehternet. We have to use a device that can be accessed using only the address and data busses, and the read, write and interrupt control lines. One such device that we found is a network controller from Standard Microsystems Corporation (SMC), implementing an ARCNET protocol engine. (the COM20020-5 ULANC). Do you know if there are drivers (for target server and debug agent) available for this device (or similar devices). Is there a different option for implementing the target connection to the host, that we could use, given the restrictions of using only the address, data and control bus? Any help or comments are appreciated and can be sent to: booj@kidd.co.za Regards, Johan From Susan.J.Henry@lmco.com Fri Jan 16 07:04:50 1998 From: Susan J Henry Date: Fri Jan 16 07:04:53 PST 1998 Subject: Diagnostics on a MVME1603 board running under VxWorks Hi, We're working on an embedded stand-alone system (no networking) that resides on a ship. If a failure is reported or the system isn't functioning properly, the sailor will run diagnostics on the PPC MVME1603 board. Since there is no built-in-test on this board, we need to develop diagnostic tests to check out the integrity of the board. Motorola does deliver PPCBUG with the board, however we need to run diagnostics while our application is running under VxWorks. In addition, the sailor is not allowed to remove and replace components on the board. Has anyone developed a set a diagnostics for the 1603 that runs under VxWorks? Has anyone successfully incorporated PPCBUGs source code into an application that runs under vxWorks? TIA, Sue Henry Lockheed Martin Launching Systems 410.682.0776 susan.j.henry@lmco.com From POST@IMS.fhg.de Fri Jan 16 08:52:07 1998 From: Rainer Stelzer Date: Fri Jan 16 08:52:11 PST 1998 Subject: Re: Communication interface In <9801161308.AA24224@csg.lbl.gov>, the vxWorks Users Group Exploder wrote: >Submitted-by BOOJ@kidd.Co.Za Fri Jan 16 05:08:39 1998 >Submitted-by: Johan Boot > >Dear Reader > >We are developing target hardware around a PowerPC >processor, using Tornado. We have a serial line available for >the target connection to the host, but are also looking for >something with more bandwidth. The harware design does >not provide for DMA transfer or PCI, which rules out (I think), >the use of ehternet. We have to use a device that can be >accessed using only the address and data busses, and the >read, write and interrupt control lines. > >One such device that we found is a network controller from >Standard Microsystems Corporation (SMC), implementing an >ARCNET protocol engine. (the COM20020-5 ULANC). >Do you know if there are drivers (for target server and debug >agent) available for this device (or similar devices). > >Is there a different option for implementing the target >connection to the host, that we could use, given the >restrictions of using only the address, data and control bus? > Hi Johan, we had the same problem around a MC68332 and so I think I can give you some helpfull hints: I don't know if you are experienced in hardware design. If so excuse the following lesson: First have to distinguish between passive and active DMA. In the first case the CPU carries the DMA controller. If your CPU does not provide a DMAC this is of course no solution for you. In the second case the network controller has a DMAC. To use this option you have to meet the following two requirements. 1. The CPU on the desgin must provide a bus arbitration Interface. I think most 32Bit CPU's fit this option, even controller cores. 2. The Network Controller's memory interface must be able to handle your system memory. (most of them use intel and/or motorola or a generic timing). Especially in controller applications you can't avoid the usage of glue logic to adapt the interfaces. This is because most of them have a "programmable glue logic" on chip, and this logic isn't accessable by a external busmaster (DMAC) We successfully used a veteran under the ethernet controllers in one of our systems: the AMD 7990. But there are a lot of more modern ones. However, you do not have to use DMA at all. DMA is very popular because the nc's can use the system memory for paket backup which makes the system cheap and fast (because less CPU copy operations take place and no further memory is neccessary) There are nc's that manage the paket buffer on theire own. This kind of nc's have the interface that you are looking for. In a second system we fitted such kind of NC (NE2000 PCMCIA (ISA Interface)) to a MC68332 system. look at: SMC, Fujitsu, Intel, AMD, National Semiconductor and Western Digital. Before you choose a nc, you should first contact your vxworks-dealer. They have a list of supported ethernet controllers on chip level. I think this is very important because the hardware design takes less time than the software design. Believe me ;-) hope this will help. regards, rainer -- ********************************************************************** fraunhofer institute of microelectronic circuits and systems department sat name rainer stelzer address finkenstrasse 61 47057 duisburg germany phone +0049/203/3783-238 fax +0049/203/3783-266 email rstelzer@ims.fhg.de ********************************************************************** From mam@pluto.dspt.com Fri Jan 16 12:34:59 1998 From: "Mark Menge" Date: Fri Jan 16 12:35:04 PST 1998 Subject: Real Time NT We are considering using NT with real-time extensions. Our system contains serial, gpib, and ethernet real time interfaces to over 30 different data acquisition devices. Since our application requires an extensive GUI support and NT now can do Real Time, switching to a NT target and host will be drastically reduce per unit shipment costs (not to mention development costs). Wind River has no plans to support real time NT. They do have some product that does windows, but that doesn't allow any Microsoft windows programs to run. I can't understand in the long run how they can hope to compete with a real time NT O.S. when GUIs are required. Is it just a matter of time before vxWorks will only be in small embedded systems like printers? Has anyone our there done an evaluation on Real Time NT O.S.s? From AKBiocca@lbl.gov Fri Jan 16 13:38:22 1998 From: Alan K Biocca Date: Fri Jan 16 13:38:26 PST 1998 Subject: recent list post rejections Here are the posting rejections currently in the queue. It keeps them around for five days or so. Before each message is the error message indicating the reason for rejection. Descriptions are below. incoming trailer reject - indicates the exploder found the trailer on the incoming message. This happens frequently when poor mailers bounce bad delivery problems back to the list. The presence of the trailer causes the message to reject. Please EDIT THE TRAILERS OFF WHEN REPLYING. no vxworks reject - indicates no use of the word vxworks or tornado in the posting. This indicates possible SPAM and gets rejected. ALWAYS INCLUDE VXWORKS OR TORNADO IN POSTINGS. mailer daemon reject - There is one user (hao@ct) that mail coming from has the header line "from mailer-daemon" rather than the actual user. The mailer daemons of some systems send us lots of bounce junk and are filtered out. They caused some serious runaway loops at times in the past, hence the filter. I've been unsuccessful at getting email directly to hao@ct, so perhaps this will help him diagnose the problem. The short summary is that the filter got one very large spam posting (I left the header below but wiped out the hundreds of content lines), and blocked perhaps 3-4 valid postings. In byte count this was a win, but it does interfere with some good comments that were blocked. Two subscription messages incorrectly sent to the list were also filtered. Please edit off the trailers and include the word vxworks in your postings! -- Alan K Biocca akbiocca@lbl.gov >:::::::::::::: >11748.incoming_trailer_reject >:::::::::::::: >>From brett.smith@wg.com Thu Jan 15 11:52:00 1998 >Received: from lbl.gov by csg.lbl.gov (4.1/0.0) > id AA11744; Thu, 15 Jan 98 11:52:00 PST >Received: from chaos.wg.com by lbl.gov (SMI-8.6/SMI-SVR4) > id LAA01220; Thu, 15 Jan 1998 11:52:14 -0800 >Received: from smtp.wg.com ([198.85.45.14] (may be forged)) by chaos.wg.com (8.8.6/8.7.3) with SMTP id OAA11220 for ; Thu, 15 Jan 1998 14:51:55 -0500 (EST) >Received: from ccMail by smtp.wg.com (ccMail Link to SMTP R6.01.01 BETA) > id AA884894040; Thu, 15 Jan 98 14:54:01 -0500 >Message-Id: <9801158848.AA884894040@smtp.wg.com> >X-Mailer: ccMail Link to SMTP R6.01.01 BETA >Date: Thu, 15 Jan 98 14:48:15 -0500 >From: "Brett Smith" >To: >Subject: Re[2]: WindView questions with Tornado 1.0.1 >Mime-Version: 1.0 >Content-Type: text/plain; charset=US-ASCII >Content-Transfer-Encoding: 7bit > > > Ethernet is not required for WindView, at least not with an x86 > target. I had WindView running over a serial target server for a > p-133. > > The only thing I had to do was build with WDB configured for serial > and included the INSTRUMENTATION directive: > > #define INCLUDE_INSTRUMENTATION > > in config.h > > BWS > >______________________________ Reply Separator _________________________________ >Subject: Re: WindView questions with Tornado 1.0.1 >Author: at wg-ral-smtp >Date: 1/15/98 6:17 AM > > >Submitted-by clare@bis.kodak.COM Thu Jan 15 06:17:00 1998 >Submitted-by: clare@kodak.com (Bill Clare) > >> >> Hey >> >> I am using the Tornado 1.0.1 development software and a 401gf BSP. I >> am attempting to get the WindView tool up and running. I do not have >> an ethernet port on the target. >> I understand that the default configuration is for ethernet, but does >> it have to have one or can a seriel port be used and if so has anyone >> accomplished this? >> Currently teh WindView does not work and I would like to get it >> running. Any help would be a great help. > >Unless I am mistaken you need the ethernet connection for WindView. I was not >able to run until I got the network connection working. Unlike the other >Tornado Tools, WindView doesn't work over the serial back end. > >On page 4 of the WindView 1.0.2 Release Notes, it mentions that "For continuous >monitoring of event data, your target must be equipped with an Ethernet >interface." It then goes on to mention that it can be over VME through another >board (I was able to use WindView through a PC104-based network card attached >to my x86 based CPU board). The rest of the paragraph seems to indicated that >if you are willing to use the post-mortem mode then you don't need an ethernet >interface. > >billc > >Bill Clare, Software Engineer, Eastman Kodak Company >Email: clare@kodak.com Mail code: 2/5/EP 35425 Phone: (726)726-9419 >US Mail: 901 Elmgrove Road Rochester NY 14653-5425 Fax: (716)726-0129 >Any opinions expressed herein are mine and not those of Eastman Kodak Company. > >********** > > This is a ... > Mail articles ... > Include the ... > edit off the ... > Send subscription ... > > > >:::::::::::::: >12517.no_vxworks_reject >:::::::::::::: >>From billy@bingo.com Thu Jan 15 12:40:31 1998 >Received: from lbl.gov by csg.lbl.gov (4.1/0.0) > id AA12510; Thu, 15 Jan 98 12:40:31 PST >Received: from mail.mymail.net by lbl.gov (SMI-8.6/SMI-SVR4) > id MAA03630; Thu, 15 Jan 1998 12:40:45 -0800 >Received: from netrepreneur.mymail.net (1Cust216.tnt13.lax3.da.uu.net [153.37.88.216]) by mail.mymail.net (8.8.5/8.7.3) with SMTP id PAA29946; Thu, 15 Jan 1998 15:05:22 -0600 (CST) >Date: Thu, 15 Jan 1998 15:05:22 -0600 (CST) >To: billy@bingo.com >From: billy@bingo.com (Me) >Comments: Authenticated sender is >Reply-To: cyberink@spectra.net >Subject: I never thought I'd be the one telling you this >Message-Id: <199801151326XAA4365@post.mymail.net> > ... A REAL LONG SPAM POSTING WAS DELETED HERE ... >:::::::::::::: >12922.no_vxworks_reject >:::::::::::::: >>From cjgarcia@unex.es Mon Jan 12 09:28:05 1998 >Received: from lbl.gov by csg.lbl.gov (4.1/0.0) > id AA12918; Mon, 12 Jan 98 09:28:05 PST >Received: from guadiana.unex.es by lbl.gov (SMI-8.6/SMI-SVR4) > id JAA08697; Mon, 12 Jan 1998 09:28:08 -0800 >Received: from seminario1.unex.es ([158.49.24.130]) > by guadiana.unex.es (8.8.5/8.8.5) with SMTP id SAA14091 > for ; Mon, 12 Jan 1998 18:27:41 +0100 >Received: by seminario1.unex.es with Microsoft Mail > id <01BD1F87.DAB6C8E0@seminario1.unex.es>; Mon, 12 Jan 1998 18:28:18 +0100 >Message-Id: <01BD1F87.DAB6C8E0@seminario1.unex.es> >From: Carlos Javier Garcia Orellana >To: "'Lista vxworks'" >Subject: About 3Com net cards drivers >Date: Mon, 12 Jan 1998 18:23:16 +0100 >Mime-Version: 1.0 >Content-Type: text/plain; charset="iso-8859-1" >Content-Transfer-Encoding: quoted-printable > >Hello, > >Has anybody a driver for de Fast Ethernet 3Com network cards? > >Thanks in advance. > >Carlos Javier Garc=EDa Orellana >Dpto. de Electronica >Universidad de Extremadura >SPAIN > >e-mail: carlos@nernet.unex.es >:::::::::::::: >18614.mailer-daemon_reject >:::::::::::::: >>From Mailer-Daemon@lbl.gov Wed Jan 14 09:14:03 1998 >Received: from lbl.gov by csg.lbl.gov (4.1/0.0) > id AA18610; Wed, 14 Jan 98 09:14:03 PST >Received: from gemsgw.med.ge.com by lbl.gov (SMI-8.6/SMI-SVR4) > id JAA07964; Wed, 14 Jan 1998 09:14:18 -0800 >Received: from gemed.med.ge.com (gemed.med.ge.com [3.7.12.4]) by gemsgw.med.ge.com (8.6.12/8.6.12) with ESMTP id LAA12206 for ; Wed, 14 Jan 1998 11:10:49 -0600 >Received: from ct.med.ge.com (ct [3.70.56.18]) by gemed.med.ge.com (8.6.12/8.6.12) with SMTP id LAA04502; Wed, 14 Jan 1998 11:13:38 -0600 >Received: from egret by ct.med.ge.com (4.1/SMI-4.0) > id AA24853; Wed, 14 Jan 98 11:13:33 CST >Received: from egret by egret (950413.SGI.8.6.12) id LAA05116; Wed, 14 Jan 1998 11:15:23 -0600 >Sender: hao@ct >Message-Id: <34BCF2AA.167E@ct.med.ge.com> >Date: Wed, 14 Jan 1998 11:15:23 -0600 >From: "L. Hao" >Organization: General Electric Company >X-Mailer: Mozilla 3.01Gold (X11; I; IRIX 6.2 IP22) >Mime-Version: 1.0 >To: vxwexplo@lbl.gov >Subject: fcntl questions >Content-Type: text/plain; charset=us-ascii >Content-Transfer-Encoding: 7bit > >Hi, all VxWorks users, > Does anyone know how to get around VxWorks's missing fcntl() to >successfully setup non-blocking I/O. ioctl(sFd, FIONBIO, 1) only works >on sockets. I have fds for pipes and am trying to find out how to make >it nonblocking. > Any help is greatly appreciated. Thanks > > >-- >Li Hao >G.E. Medical Systems >email:Li.Hao@ct.med.ge.com >Phone:5-4267 >:::::::::::::: >19103.no_vxworks_reject >:::::::::::::: >>From awai@info.kobelco.co.jp Mon Jan 12 18:27:31 1998 >Received: from lbl.gov by csg.lbl.gov (4.1/0.0) > id AA19099; Mon, 12 Jan 98 18:27:31 PST >Received: from aquarius.kobelco.co.jp by lbl.gov (SMI-8.6/SMI-SVR4) > id SAA07842; Mon, 12 Jan 1998 18:27:43 -0800 >Received: by aquarius.kobelco.co.jp; id LAA17970; Tue, 13 Jan 1998 11:29:15 +0900 (JST) >Received: from cancer.kobelco.co.jp(133.109.30.2) by aquarius.kobelco.co.jp via smap (3.2) > id xma017963; Tue, 13 Jan 98 11:29:10 +0900 >Received: from info0.snmy.info.kobelco.co.jp ([10.16.99.1]) by cancer.kobelco.co.jp (8.8.5+2.7Wbeta5/3.4W-98010613) with SMTP id LAA02318 for ; Tue, 13 Jan 1998 11:26:24 +0900 (JST) >Received: from [10.18.96.40] by info0.snmy.info.kobelco.co.jp (8.6.12+2.4W/6.4J.1-Ver3.0) > id LAA01133; Tue, 13 Jan 1998 11:24:10 +0900 >Message-Id: <199801130224.LAA01133@info0.snmy.info.kobelco.co.jp> >X-Sender: awai@info0 >X-Mailer: Macintosh Eudora Pro Version 3.1.1-J >Mime-Version: 1.0 >Content-Type: text/plain; charset="us-ascii" >Date: Tue, 13 Jan 1998 11:28:51 +0900 >To: vxwexplo@lbl.gov >From: Syoichi Awai >Subject: help > >help > > >:::::::::::::: >20303.no_vxworks_reject >:::::::::::::: >>From tkb@mclean.sparta.com Wed Jan 14 11:09:21 1998 >Received: from lbl.gov by csg.lbl.gov (4.1/0.0) > id AA20299; Wed, 14 Jan 98 11:09:21 PST >Received: from Sparta.Com by lbl.gov (SMI-8.6/SMI-SVR4) > id LAA14789; Wed, 14 Jan 1998 11:09:36 -0800 >Received: from Mclean.Sparta.Com (heisenberg.mclean.sparta.com [157.185.42.1]) > by Sparta.Com (8.8.5/8.8.5) with ESMTP id OAA09210 > for ; Wed, 14 Jan 1998 14:09:17 -0500 (EST) >Received: from hometkb ([157.185.44.51]) > by Mclean.Sparta.Com (8.8.5/8.8.5) with SMTP id OAA03358 > for ; Wed, 14 Jan 1998 14:08:56 -0500 (EST) >Message-Id: <3.0.32.19980114141011.006aeb28@mailhost.mclean.sparta.com> >X-Sender: tkb@mailhost.mclean.sparta.com >X-Mailer: Windows Eudora Pro Version 3.0 (32) >Date: Wed, 14 Jan 1998 14:10:14 -0500 >To: vxwexplo@lbl.gov (the vxWorks Users Group Exploder) >From: "Thomas K. Buchanan" >Subject: Re: 860 Interrupts >Mime-Version: 1.0 >Content-Type: text/plain; charset="us-ascii" > >>When a level-triggered interrupt is received on most machines the interrupt >>service routine (ISR) must clear the source of the interrupt or disable >>that interrupt until it is known that the source of the interrupt is clear. >>By "source of the interrupt" I mean the hardware signal connected to the >>processor interrupt line (internally or externally). Normally this signal >>would be designed to go inactive when the ISR performs it's task otherwise >>when the ISR completes the interrupt will be immediately activated again >>and bingo the processor latches up in a ISR-ReturnFromInterrupt-ISR state >>until the source goes inactive (if it ever does). Most of us have probably >>done this at some time or other. (By the way, I'm not familiar with the 860). >> >>For this reason interrupts are often specified as edge-triggered. Most > >Wrong! Edge triggered interrupts exist becauses it makes handling the >interrupt signals easier for the hardware on the CPU. It is a tradeoff that >Intel decided to make and Motorola did not. > >>Finally level-triggered interrupts can be missed whereas edge-triggered >>interrupts "can't" be missed. If a level-triggered interrupt signal goes > >In my experience, this is the opposite of the truth. Level edge >interruptors never deassert the signal unless they get serviced. In >addition, edge triggered interrupts are subject to noise and other spurious >problems. >I understand the hardware decisions behind using edge triggered interrupts >though I don't agree with them. I don't understand why should it ever be OK >to miss the source of an interrupt. What possible advantage does that give >us? > >adios >:::::::::::::: >20781.no_vxworks_reject >:::::::::::::: >>From thomas.arand@oen.siemens.de Fri Jan 16 01:30:54 1998 >Received: from lbl.gov by csg.lbl.gov (4.1/0.0) > id AA20777; Fri, 16 Jan 98 01:30:54 PST >Received: from gorilla.mchh.siemens.de by lbl.gov (SMI-8.6/SMI-SVR4) > id BAA03418; Fri, 16 Jan 1998 01:31:09 -0800 >Received: from moody.mchh.siemens.de (moody-i.mchh.siemens.de [194.138.158.26]) > by gorilla.mchh.siemens.de (8.8.7/8.8.7) with ESMTP id KAA06300 > for ; Fri, 16 Jan 1998 10:30:49 +0100 (MET) >Received: from descn121.oen.siemens.de (descn121 [132.88.71.55]) > by moody.mchh.siemens.de (8.8.7/8.8.7) with SMTP id KAA27861 > for ; Fri, 16 Jan 1998 10:32:03 +0100 (MET) >Received: from Microsoft Mail (PU Serial #1895) > by descn121.oen.siemens.de (PostalUnion/SMTP(tm) v2.1.9h for Windows NT(tm)) > id AA-1998Jan16.101430.1895.210274; Fri, 16 Jan 1998 10:30:15 +0100 >From: thomas.arand@oen.siemens.de (Arand, Thomas) >Return-Receipt-To: thomas.arand@oen.siemens.de (Arand, Thomas) >To: vxwexplo@lbl.gov ('vxwexplo@lbl.gov') >Message-Id: <1998Jan16.101430.1895.210274@descn121.oen.siemens.de> >X-Mailer: Microsoft Mail via PostalUnion/SMTP for Windows NT >Mime-Version: 1.0 >Content-Type: text/plain; charset="ISO-8859-1" >Content-Transfer-Encoding: quoted-printable >Organization: Siemens AG >Date: Fri, 16 Jan 1998 10:30:15 +0100 >Subject: Symbol s_userGet in VxSim target > > >Hi VxSim world, > >does anyone know about the symbol s_userGet in the VxSim target? I = >sometimes get the error message =22s_userGet: getenv failed=22 when VxSim = >= >starts up. Can anyone explain this? > >Thanks in advance, > >Thomas Arand > ---------------------------------------------------------------------- >Thomas Arand Siemens AG >Tel. +49 - 89 - 722 42296 >Fax. +49 - 89 - 722 26572 >thomas.arand=40oen.siemens.de > > >:::::::::::::: >29316.incoming_trailer_reject >:::::::::::::: >>From t_wmoles@qualcomm.com Tue Jan 13 09:15:18 1998 >Received: from lbl.gov by csg.lbl.gov (4.1/0.0) > id AA29312; Tue, 13 Jan 98 09:15:18 PST >Received: from adept.qualcomm.com by lbl.gov (SMI-8.6/SMI-SVR4) > id JAA24254; Tue, 13 Jan 1998 09:15:33 -0800 >Received: from t_wmoles (t-wmoles.qualcomm.com [129.46.159.54]) by adept.qualcomm.com (8.8.5/1.4/8.7.2/1.14) with SMTP id JAA23446 for ; Tue, 13 Jan 1998 09:15:12 -0800 (PST) >Message-Id: <3.0.1.32.19980113091513.009dc310@emrys.qualcomm.com> >X-Sender: t_wmoles@emrys.qualcomm.com >X-Mailer: Windows Eudora Pro Version 3.0.1 (32) >Date: Tue, 13 Jan 1998 09:15:13 -0800 >To: vxwexplo@lbl.gov (the vxWorks Users Group Exploder) >From: Hugh Molesworth >Subject: Re: 860 Interrupts >In-Reply-To: <9801131309.AA25524@csg.lbl.gov> >Mime-Version: 1.0 >Content-Type: text/plain; charset="us-ascii" > >Level vs. Edge > >When a level-triggered interrupt is received on most machines the interrupt >service routine (ISR) must clear the source of the interrupt or disable >that interrupt until it is known that the source of the interrupt is clear. >By "source of the interrupt" I mean the hardware signal connected to the >processor interrupt line (internally or externally). Normally this signal >would be designed to go inactive when the ISR performs it's task otherwise >when the ISR completes the interrupt will be immediately activated again >and bingo the processor latches up in a ISR-ReturnFromInterrupt-ISR state >until the source goes inactive (if it ever does). Most of us have probably >done this at some time or other. (By the way, I'm not familiar with the 860). > >For this reason interrupts are often specified as edge-triggered. Most >processors now correctly latch such events provided they meet minimum >pulse-width requirements. The danger is that such an interrupt is used as a >handshake and one gets missed, which of course can't happen :-), and this >causes some process or other to hang up awaiting the next signal (the one >we just missed). To avoid this most such systems would incorporate some >kind of timeout checking or method to instigate a check that an interrupt >had been issued but not processed, perhaps by the occasional read of a >status register. You can save a lot of grief by checking for that >(impossible) missed interrupt. > >Finally level-triggered interrupts can be missed whereas edge-triggered >interrupts "can't" be missed. If a level-triggered interrupt signal goes >active whilst a higher priority interrupt ISR is in progress but goes >inactive before the higher-priority ISR completes then the interrupt will >be missed on lots of machines because it is not latched; in this case >(assuming you wanted to catch it) you should be using an edge-triggered >interrupt or make sure it doesn't go away until the ISR executes. Most >level-triggered interrupt systems are designed to stay active until the ISR >services the interrupt. > >Power-fail or other fatal non-maskable interrupts are often level-triggered >simply because they often never go away (and you definitely don't want to >miss them ...). > >Hugh >hugh@qualcomm.com > >At 05:09 AM 1/13/98 PST, you wrote: >>Submitted-by bobperry@bellatlantic.net Tue Jan 13 05:09:40 1998 >>Submitted-by: "Robert L. Perry" >> >>One possible problem could be that the cause of the interrupt is not being >>reset -writing to the SIPEND register doesn't do anything (see p.12-5 >>section 12.2.1.3.1) >>Perhaps your 860 is interrupt'n like crazy. >> >>We had alot of trouble w/ CPM interrupt support on the 860, here was one >>find: many bits in the 860 registers are reset when a '1' is written to the >>bit position of interrest. One example on how *not* to access the CISR >>register can be found in ppc860CpmIntrDemux() in ppc860Intr.c >> >> /* clear the bit in the CISR */ >>#if 1 >> * CISR(regBase) = (1 << intNum); /* This is correct */ >>#else >> * CISR(regBase) |= (1 << intNum); /* This is not correct */ >>#endif >> >>The OR operation above appears to be the straigth forward approach to >>writing a '1' to the proper bit. The problem is that the compiler will >>read the CISR value, OR the proper bit, then perform a write. If more than >>one bit is set in the CISR, they will all be reset after the |= operation, >>which in this case is very sad. >> >>We visited ppc860Intr.c while getting all of the scc's operational and >>there is a TSR open w/ WindRiver -to date, our changes have not been >>blessed. If anyone is interrested, please contact me directly as I don't >>know what the rules are regarding code segments postings from copyright >>material. >> >>Bob Perry >>bobperry@bellatlantic.net >> >>At 10:55 AM 1/11/98 PST, you wrote: >>>Submitted-by mabrouk@nortel.ca Sun Jan 11 10:55:05 1998 >>>Submitted-by: "Mabrouk Aboughanaima" >>> >>>Hi All, >>> >>>I'm developing a C++ application using vxWorks 5.3.1 for a PowerPC860 >>>platform. I am trying to connect an ISR to handle an external interrupt >>which is >>>a level interrupt but I have no luck yet. When I invoke the function that >>connects >>>the ISR fron windSh or CrossWind, I neve get control back, it seems the >>system is too busy doing something >>>even though the exeternal device's interrupt is not enabled yet. >>> >>>When I enable the exeternal device interrupt and then invoke the function >>that connect the ISR >>>I get the message lost connection to target from windSh. >>> >>>Here is the code. >>> >>>//function that clears/sets the external interrupt. >>> >>>void setInterrupt(int state) >>>{ >>> unsigned char data = 0; >>> int ret = -1; >>> >>> if ( state == 1) >>> { >>> if (( !read(0x59, data))) //read control register >>> { >>> >>> data |= (1 << 1); >>> >>> ret = write(0x19, data); >>> } >>> >>> } >>> else >>> { >>> if (!read(0x59, data)) //read control register >>> { >>> >>> data &= (~(1 << 1)); >>> >>> ret = write(0x19, data); >>> } >>> >>> } >>> logMsg("\nin setInterrupt(return) %d\n", ret, 0,0,0,0,0); >>>} >>> >>>void isr(void) >>>{ >>> >>> unsigned long pimmr = vxImmrGet(); >>> >>> setInterrupt(0); //clear interrupt >>> >>> *SIPEND(pimmr) |= SIEL_ED4; // reset the SIPEND bit corressponding to >IRQ4 >>> >>> >>> logMsg("\nin isr\n", 0, 0,0,0,0,0); >>> >>>} >>> >>> >>>int initISR() >>>{ >>> unsigned long pimmrbase = vxImmrGet(); >>> >>> >>> if ( intConnect(IV_IRQ4, (VOIDFUNCPTR) isr, 0) == OK ) >>> { >>> *SIEL(pimmrbase) &= ~SIPEND_IRQ4; >>> >>> *SIMASK(pimmrbase) |= SIMASK_IRM4; //This is where it seems to be >>stuck at >>> return 0; >>> } >>> else >>> { >>> printf("\nintConnect failed\n"); >>> return -1; >>> } >>> >>>} >>> >>> >>> >>> >>>Any help will be greatly apprecaited. Please e_mail or call me. >>> >>>Thanks a million. >>> >>>++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ >>>+ Mabrouk Aboughanaima + Tel: (613) 763-6000 + >>>+ CDMA Radio Software + ESN: 393-6000 + >>>+ Nortel Wireless Networks + FAX: (613) 763-2686 + >>>+ P.O. Box 3511, Station 'C' + Internet: mabrouk@nortel.ca + >>>+ Ottawa K1Y 4H7 Canada + + >>>++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ >> >>> >>> >> >> >>********** >> >> This is a ...trailer removed... >> >> >:::::::::::::: >7013.no_vxworks_reject >:::::::::::::: >>From gprimiano@hayes.com Thu Jan 15 07:40:38 1998 >Received: from lbl.gov by csg.lbl.gov (4.1/0.0) > id AA07006; Thu, 15 Jan 98 07:40:38 PST >Received: from atl_xch_srvr1.hayes.com by lbl.gov (SMI-8.6/SMI-SVR4) > id HAA17886; Thu, 15 Jan 1998 07:40:51 -0800 >Received: by atl_xch_srvr1.hayes.com with Internet Mail Service (5.0.1457.3) > id ; Thu, 15 Jan 1998 10:40:08 -0500 >Message-Id: <6F8A1DCCA47ED011ADB600609733E58D010A20C1@atl_xch_srvr1.hayes.com> >From: "Primiano, Guy" >To: "'VX'" >Subject: subscribe >Date: Thu, 15 Jan 1998 10:40:07 -0500 >X-Priority: 3 >X-Ms-Tnef-Correlator: <6F8A1DCCA47ED011ADB600609733E58D010A20C1@atl_xch_srvr1.hayes.com> >Mime-Version: 1.0 >X-Mailer: Internet Mail Service (5.0.1457.3) >Content-Type: multipart/mixed; > boundary="---- =_NextPart_000_01BD21A1.F30D5370" > >This message is in MIME format. Since your mail reader does not understand >this format, some or all of this message may not be legible. > >------ =_NextPart_000_01BD21A1.F30D5370 >Content-Type: text/plain > >I would like to subscribe. > >------ =_NextPart_000_01BD21A1.F30D5370 >Content-Type: application/ms-tnef >Content-Transfer-Encoding: base64 > >eJ8+IgkPAQaQCAAEAAAAAAABAAEAAQeQBgAIAAAA5AQAAAAAAADoAAEIgAcAGAAAAElQTS5NaWNy >b3NvZnQgTWFpb ... base64 erased >------ =_NextPart_000_01BD21A1.F30D5370-- > > From tkb@mclean.sparta.com Fri Jan 16 14:44:04 1998 From: "Thomas K. Buchanan" Date: Fri Jan 16 14:44:07 PST 1998 Subject: Re: Real Time NT >Wind River has no plans to support real time NT. They do have some >product that does windows, but that doesn't allow any Microsoft >windows programs to run. I can't understand in the long run >how they can hope to compete with a real time NT O.S. when GUIs >are required. Is it just a matter of time before vxWorks will only be in >small embedded systems like printers? They don't have a product that "does" windows. They have a Windows-based development environment. The cost of development is mostly labor, not the tools. Choosing Win32 for real-time applications to save development dollars is a questionable decision. It's a great platform for GUIs and client/server apps. I usually use it for the GUI and another target running VxWorks for hard real-time duties. I have heard several horror stories about the various "real-time" extensions to NT, especially since Microsoft moved all the graphics code into the kernel in version 4.0 and above. The hardware requirements for NT are huge compared to VxWorks. You can enable every option in VxWorks and still embed it in 2MB of flash. On some x86 platforms, the real competition might be with WinCE. Why would Wind River want to compete with "real-time" NT? I think printers, scanners, engines, telecommunications, consumer applicances, military gear etc. are a big enough market. The ratio of desktop/VME platforms to smaller embedded targets is very small and shrinking every day. I have 4 Win32 boxes between work and home, there are 12-15 embedded computers in my cars alone and dozens more around the house. adios -------------------------------------------- /\ Thomas Keith Buchanan /##\ Principal Engineer /####\ /####/-- /####/---- SPARTA, Inc. -\####\__--- Computer Products Division ---\##\ /---- 7926 Jones Branch Drive ----\##\/\---- Suite 900 ----\####\-- McLean, VA 22102 ----\####/ ---/###/ Voice: 703 448 0210 -/###/ FAX: 703 893 5494 \##/ WWW: www.mclean.sparta.com \/ From mmcquade@pop.aa.net Fri Jan 16 15:08:17 1998 From: Mike McQuade Date: Fri Jan 16 15:08:20 PST 1998 Subject: Re: Real Time NT vxworks At 12:35 PM 1/16/98 PST, you wrote: >Submitted-by mam@pluto.dspt.com Fri Jan 16 12:34:59 1998 >Submitted-by: "Mark Menge" >Wind River has no plans to support real time NT. They do have some >product that does windows, but that doesn't allow any Microsoft >windows programs to run. I can't understand in the long run >how they can hope to compete with a real time NT O.S. when GUIs >are required. Is it just a matter of time before vxWorks will only be in >small embedded systems like printers? > Mark, We have a GUI on our VxWorks system, its not a windows system, but it works very well, and our customers like it :) We use the graphics package from: http://www.rastergr.com see our instrument screen at: http://www.aa.net/~mmcquade/graphic.htm I have a couple friends that are device driver support techs at Microsoft, they consider the real-time extensions to NT a plague. I wouldn't go there unless the customer demanded a Windows interface. Maybe a better option would be the real-time Linux extensions, you can read more at: http://luz.cs.nmt.edu/~rtlinux/ I would think the transition from VxWorks to Linux would be less painful than jumping into Win32. Mike. From daemon@csg.lbl.gov Sat Jan 17 04:03:14 1998 From: daemon@csg.lbl.gov Date: Sat Jan 17 04:03:18 PST 1998 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sat Jan 17 04:03:13 PST 1998 Subject: Driver for 3Com Fast Ethernet cards Subject: Re: unbuffered ANSI streams Subject: VxSim-to-UNIX Interface Subject: Re: Bootroms for the MVME167? ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Driver for 3Com Fast Ethernet cards Date: Fri, 16 Jan 1998 09:41:03 +0100 From: "Carlos J. Garcia Orellana" Organization: Universidad de Extremadura Message-ID: <34BF1D1F.3055@nernet.unex.es> Hello, Anybody knowns if there are a driver for 3Com Fast Ethernet cards?. Thanks. Carlos. Universidad de Extremadura (SPAIN) carlos@nernet.unex.es --------------------------- Newsgroups: comp.os.vxworks Subject: Re: unbuffered ANSI streams Date: 16 Jan 1998 12:00:14 GMT From: gj@mail.ast.cam.ac.uk (Gareth Jones) Organization: Royal Greenwich Observatory, Cambridge (UK) Message-ID: <69ni4e$igr$1@lyra.csx.cam.ac.uk> References: FISH wote: > I have NFS mounted a unix disk. I opened a file > on this using fopen() and set no-buffering option using > setvbuf(). I cannot get the system to write the data > to the file unless I fclose() it !! fflush() does not > help either. I have encountered this too, but in the system I worked on it was not an issue. However, I remember reading/being-told some time ago that the reason for this is that the directory entry on the host system (presumably SUN?) is not updated until the file is closed? Is this actually the case? could someone shed a little light on this one? Thanks, Gareth Jones RGO Technology Division ... all my own, etc ... --------------------------- Newsgroups: comp.os.vxworks Subject: VxSim-to-UNIX Interface Date: 16 Jan 1998 21:15:37 GMT From: "Brenda Spaur" Organization: BrenCorp Message-ID: <01bd22c4$bb616aa0$224fa8ce@aries3.spaur.com.206.168.79.31> I am running VxSim (Tornado) on a Sparc 20/Solaris workstation. I need to be able to create and manipulate UNIX semaphores from within VxSim but have not found a way to do so (long story which I won't go into). The simulated VxWorks semLib routines in this environment do not create UNIX semaphores and UNIX system calls such as semget, semctl, semop, etc. show up as undefined symbols. It also looks like the UNIX "system" call is not implemented. Seems like there should be but ... is there a way to call standard UNIX system calls from applications running under VxSim? I think I will need access to other UNIX system calls that don't appear to be implemented in VxSim so a generic solution would be good. Thanks in advance, Brenda Spaur The Johns Hopkins University Applied Physics Laboratory bms@aplexus.jhuapl.edu --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Bootroms for the MVME167? Date: Fri, 16 Jan 1998 13:05:44 +0100 From: Markus Muehlbauer Message-ID: <34BF4D17.A7E7918D@ubaclu.unibas.ch> References: <34bd618d.0@news.randori.com> micron wrote: > > I'm using vxWorks 5.0.2 on a SPARC 10 to create new bootroms to include > loading vxWorks from a local scsi drive, and was wondering if there are any > special procedures for getting proms to boot the Motorola MVME167-34B (33 Mhz > w/32MB ram) board properly? I have edited the bootConfig.c file and used the > 'make bootrom.hex' (this is a compressed hex file, I have also used the > uncompressed version) command successfully. (I'm presently using previously > created bootroms in the xu1 and xu2 PLCC sockets and they work fine, the > proms are M5M27C202 128K x 16 bits). I then take the Motorola S record > bootrom.hex file and mcopy the file to a DOS formatted disk. Using an Allpro 88 > prom programmer I read in the bootrom.hex file from the DOS floppy using the > Motorola S format. It should also be noted that the bootrom.hex file is in a > combination of Motorola S1 and S2 format. After reading in the file I write it > back to the hard drive converted in the Motorola S3 format, this is done to > facilitate a 32 bit word format. The S3 format bootrom.hex file is broken down > to two different files by the Allpro 88, one for even addresses and one for odd > addresses. The two proms are programmed and their checksums are verified. > Unfortunately when the proms are put into MVME167 sockets and the board is > powered up the SYSFAIL and STAT leds remain on and I see no indication of life > on the terminal screen. So my question is, is there some sort of checksum that > needs to be calculated and put in a proper address location to give control > over to the proms like in the 167bug proms? If not are there any other > suggestions. > > -Frustrated Craig Hi Craig - recently I made rootroms for my VCC2118 - also a MC68xxx guy using a 27C1024 (16Bit) rom. I've made 'bootrom.hex' on my DEC Unix system, put it on a DOS disk, loaded it into our DOS based EPROM burner in Motorola format and burned an EPROM. And with the new one our CPU doesn't boot. I took the old (working) EPROM put it into the Burner and looked at its contents - and i found the byteorder of the new one was different from the old one! Therfore I changed the byteorder and reburned the new one and boot! So put your old ones into the burner and compare them with the new ones (in the first bytes you will find the entry point and within the first few hundred bytes the copyright message - both helpfull in finding the byte order!) and swap the bytes accordingly. Hope this will help Markus byte took this file on a - -------------------------------------------------------------------- Markus Muehlbauer Tel.: xx41/(0)61-267 3746 Institut fuer Physik Fax.: xx41/(0)61-267 3784 Universitaet Basel Kingelbergstr. 82 EMail: CH-4056 Basel muehlbauer@ubaclu.unibas.ch - -------------------------------------------------------------------- --------------------------- End of New-News digest ********************** From bberry@nortel.ca Sat Jan 17 09:13:21 1998 From: "Bo Berry" Date: Sat Jan 17 09:13:24 PST 1998 Subject: dosFS workarounds, bandages, tourniquets Mark: We've experienced a similar list of problems with dosFs. Here's what we found so far: o Recieved a patch from Wind River to correct a problem with the rename function of the DOS filesystem. This problem was corrupting our file structure, a real bleeder.... o check your scsi bus termination. We had a problem on some drives where the scsi bus was not properly terminated, causing I/O errors. o verify that the remap option for the scsi drive is enabled. This is the option that allows the drive to automatically remap bad blocks. Note: drives have a finite number of blocks available for automatic remapping. o we also found a few instances of 'improper' coding when code was compiled for the unix file system, tornado sim, and VxWorks run time. May wish to scrub through code validating parmeters and usage if it is cross compiled on several targets. o if the system is powered down or a crashes during disk I/O, then most likely file structure problems will occur. We also experience this. A chkdisk like utility does the trick - recovering lost chains and the like. Caution though, a critical file may not exist when the system reboots and losts chain is freed. o SCSI specs can be found at: ftp://ftp.std.com/obi/Standards/ANSI/SCSI-2/ Hope this helps, ---------------------------------------------------------------- Bo Berry Manager, Board Level Systems Nortel - Signaling Solutions Group voice: 919-905-3461 ESN: 263 507 Airport Blvd, Suite 115 fax: 919-905-3918 Morrisville, NC 27560 email: bberry@nortel.com Mailstop: 08Z001 ---------------------------------------------------------------- In message "Vx Exploder Digest", you write: > Subject: Searching for DOSFS workarounds, bandages, tourniquets > > ********** > > From mam@pluto.dspt.com Sat Jan 10 10:24:58 1998 > From: "Mark Menge" > Date: Sat Jan 10 10:25:01 PST 1998 > Subject: Searching for DOSFS workarounds, bandages, tourniquets > > We have been seeing more flakiness with vxWorks' DOSFS. > > Sometimes a disk will get into a reproduceable state where several k of memory is > lost after creating and closing a new file. Reformatting is the only > cure. I understand it normally loses about 20 bytes for each new file on the disk. > Since our system can write lots of data we have cases where our memory monitor shows the > memory getting quickly sucked up > > 1. Has anyone seen this? > > 2. Does anyone have source code to check if files are cross linked? > I heard if the system is turned off with a open file, you could get > cross linked files. > > 3. Anyone else know of other nonboard specific gotchas and fixes? From espin@idiom.com Sat Jan 17 13:42:09 1998 From: Geoffrey Espin Date: Sat Jan 17 13:42:12 PST 1998 Subject: Re: unbuffered ANSI streams > Subject: Re: unbuffered ANSI streams > Date: 16 Jan 1998 12:00:14 GMT > From: gj@mail.ast.cam.ac.uk (Gareth Jones) > Organization: Royal Greenwich Observatory, Cambridge (UK) > FISH wote: > > I have NFS mounted a unix disk. I opened a file > > on this using fopen() and set no-buffering option using > > setvbuf(). I cannot get the system to write the data > > to the file unless I fclose() it !! fflush() does not > > help either. > > I have encountered this too, but in the system I worked on it > was not an issue. However, I remember reading/being-told > some time ago that the reason for this is that the directory > entry on the host system (presumably SUN?) is not updated until > the file is closed? Is this actually the case? could someone > shed a little light on this one? > > Thanks, > Gareth Jones > RGO Technology Division The VxWorks NFS "driver" has an internal 8K cache. Try this, usrConfig.c:usrRoot(), extern int nfsMaxMsgLen; nfsMaxMsgLen = 0; /* default is 8K */ I don't recall if I've ever tried a zero size cache. Or one can (from the shell set "nfsCachSize = " as long as a file is already open. Geoff --- Geoffrey Espin espin@idiom.com From espin@idiom.com Sat Jan 17 13:49:26 1998 From: Geoffrey Espin Date: Sat Jan 17 13:49:30 PST 1998 Subject: Symbol s_userGet in VxSim target Thomas, (from the vxWorks newsgroup reject pile) > >From: thomas.arand@oen.siemens.de (Arand, Thomas) > >To: vxwexplo@lbl.gov ('vxwexplo@lbl.gov') > >Content-Type: text/plain; charset="ISO-8859-1" > >Content-Transfer-Encoding: quoted-printable > >Organization: Siemens AG > >Date: Fri, 16 Jan 1998 10:30:15 +0100 > >Subject: Symbol s_userGet in VxSim target > >Hi VxSim world, > > > >does anyone know about the symbol s_userGet in the VxSim target? I = > >sometimes get the error message =22s_userGet: getenv failed=22 when VxSim = > >= > >starts up. Can anyone explain this? > > > >Thanks in advance, I think this is an age-old bug, I can't recall if I ever knew what triggered it. s_userGet() is an internal "glue" routine, which calls u_getenv(), u_gethostname(), u_xxx, to get from your UNIX enviroment info useful to VxSim (like calculating your bootline automagically). Anyways, getenv("USER") can sometimes return null/error. You can probably continue to ignore it. Geoff -- Geoffrey Espin espin@idiom.com From froeber@BBN.COM Sun Jan 18 19:05:47 1998 From: Fred Roeber Date: Sun Jan 18 19:05:58 PST 1998 Subject: Re: NFS data not showing up on remote disk FISH wrote about vxworks: > > I have NFS mounted a unix disk. I opened a file > on this using fopen() and set no-buffering option using > setvbuf(). I cannot get the system to write the data > to the file unless I fclose() it !! fflush() does not > help either. Gareth Jones responded: > I have encountered this too, but in the system I worked on it > was not an issue. However, I remember reading/being-told > some time ago that the reason for this is that the directory > entry on the host system (presumably SUN?) is not updated until > the file is closed? Is this actually the case? could someone > shed a little light on this one? First off, you have to make sure that you are actually opening the file you want on the NFS mounted disk; not the default netDrv mounted disk. If you don't specify the correct path name on your open, you can end up with the file actually being opened on the wrong device. Use the iosFdShow command to indicate what drive the file is being opened with. The iosDrvShow command shows entry points for each of the drives. Using lkAddr on the driver entry points can tell you which drive is which. Second, there is the issue of buffering. By default, NFS buffers up data in 8K chunks and doesn't write it out to disk until that much data has accumulated. Probably one of these reasons is why the data doesn't seem to be getting written out. If you write more than 8KB to a file that is actually being handled by NFS, it has always showed up on the other system in my experience (even if you don't flush the data). Fred | Fred J Roeber, BBN Systems & Technologies | | 4 John Clarke Road Middletown, RI 02842-5202 | | froeber@bbn.com 401-848-3548 | From tony.mayes@[194.184.137.2] Mon Jan 19 00:12:39 1998 From: tony.mayes@[194.184.137.2] Date: Mon Jan 19 00:12:43 PST 1998 Subject: Profibus and VxWorks Can anyone help me with a source of, or supplier of a Profibus driver for Vxworks? We want one for a system running on a PC/104 bus with Intel 486 / Pentium CPU's. We would prefer a supplier of both hardware and software. I am interested in both slave and master drivers. Thanks in advance. Tony Mayes Tony Mayes Email address tony.mayes@epid.eurotherm.co.uk Phone 031 975243 Fax 031 977512 Eurotherm SpA, Via XXIV Maggio, 22070 Guanzate (CO), Italy (NOTE: The opinions expressed within are my own and are not necessarily those of my employer) From brett.smith@wg.com Mon Jan 19 05:09:47 1998 From: "Brett Smith" Date: Mon Jan 19 05:09:50 PST 1998 Subject: Re[2]: WindView questions with Tornado 1.0.1 Ethernet is not required for WindView, at least not with an x86 target. I had WindView running over a serial target server for a p-133. The only thing I had to do was build with WDB configured for serial and included the INSTRUMENTATION directive: #define INCLUDE_INSTRUMENTATION in config.h BWS ______________________________ Reply Separator _________________________________ Subject: Re: WindView questions with Tornado 1.0.1 Author: at wg-ral-smtp Date: 1/15/98 6:17 AM Submitted-by clare@bis.kodak.COM Thu Jan 15 06:17:00 1998 Submitted-by: clare@kodak.com (Bill Clare) > > Hey > > I am using the Tornado 1.0.1 development software and a 401gf BSP. I > am attempting to get the WindView tool up and running. I do not have > an ethernet port on the target. > I understand that the default configuration is for ethernet, but does > it have to have one or can a seriel port be used and if so has anyone > accomplished this? > Currently teh WindView does not work and I would like to get it > running. Any help would be a great help. Unless I am mistaken you need the ethernet connection for WindView. I was not able to run until I got the network connection working. Unlike the other Tornado Tools, WindView doesn't work over the serial back end. On page 4 of the WindView 1.0.2 Release Notes, it mentions that "For continuous monitoring of event data, your target must be equipped with an Ethernet interface." It then goes on to mention that it can be over VME through another board (I was able to use WindView through a PC104-based network card attached to my x86 based CPU board). The rest of the paragraph seems to indicated that if you are willing to use the post-mortem mode then you don't need an ethernet interface. billc Bill Clare, Software Engineer, Eastman Kodak Company Email: clare@kodak.com Mail code: 2/5/EP 35425 Phone: (726)726-9419 US Mail: 901 Elmgrove Road Rochester NY 14653-5425 Fax: (716)726-0129 Any opinions expressed herein are mine and not those of Eastman Kodak Company. From wae@interspeed.com Mon Jan 19 07:18:47 1998 From: William Edmondson Date: Mon Jan 19 07:18:50 PST 1998 Subject: RE: WindView questions with Tornado 1.0.1 [wae] this is probably correct. i was thinking of crosswind. bill interSPEED > > Hey > > I am using the Tornado 1.0.1 development software and a 401gf BSP. I > am attempting to get the WindView tool up and running. I do not have > an ethernet port on the target. > I understand that the default configuration is for ethernet, but does > it have to have one or can a seriel port be used and if so has anyone > accomplished this? > Currently teh WindView does not work and I would like to get it > running. Any help would be a great help. Unless I am mistaken you need the ethernet connection for WindView. I was not able to run until I got the network connection working. Unlike the other Tornado Tools, WindView doesn't work over the serial back end. On page 4 of the WindView 1.0.2 Release Notes, it mentions that "For continuous monitoring of event data, your target must be equipped with an Ethernet interface." It then goes on to mention that it can be over VME through another board (I was able to use WindView through a PC104-based network card attached to my x86 based CPU board). The rest of the paragraph seems to indicated that if you are willing to use the post-mortem mode then you don't need an ethernet interface. From Huang.Wu@nrc.ca Mon Jan 19 08:27:07 1998 From: "Wu, Huang" Date: Mon Jan 19 08:27:10 PST 1998 Subject: SCSI devices Hello Everyone, Just very recently I upgrade vxworks from 5.3 to 5.3.1 in order to get tape drive support. All my scsi drives (hard disk and floppy disk) work well before upgrading. For version 5.3, I can detect tape drive when scsi_auto_config is defined. But after upgrading, none of these devices can be detected using scsi_auto_config.I also tried to configure them manualy but I got nothing: scsiShow does not display any scsi devices detected. The basic error message is Thread execution failed, ScsiCommand Error and Request sense failed. I am wondering whether I forgot to do something with this new version. In version 5.3.1, include_scsi2 is defined by default: In ../mv167/config.h, I include #if TRUE /* change FALSE to TRUE for SCSI interface */ #define INCLUDE_SCSI_DMA /* Enable DMA for SCSI */ #define INCLUDE_SCSI /* include ncr710 driver */ #define INCLUDE_SCSI_BOOT /* include ability to boot from SCSI */ #define INCLUDE_DOSFS /* file system to be used */ #define INCLUDE_TAPEFS /* undef tape file system if necessary */ /*#define SYS_SCSI_CONFIG*/ /* BSP provides sysScsiConfig, in sysScsi.c */ #define SCSI_AUTO_CONFIG #endif /* FALSE/TRUE */ Can anyone show me some light? Many thanks Huang Flight Research Lab huang.wu@nrc.ca From kalalau!jmk@uu9.psi.com Mon Jan 19 09:14:05 1998 From: jmk@acut.com (Joel Katz) Date: Mon Jan 19 09:14:09 PST 1998 Subject: Re: NFS data not showing up on remote disk >Submitted-by froeber@BBN.COM Sun Jan 18 19:05:47 1998 >Submitted-by: Fred Roeber > >FISH wrote about vxworks: >> >> I have NFS mounted a unix disk. I opened a file >> on this using fopen() and set no-buffering option using >> setvbuf(). I cannot get the system to write the data >> to the file unless I fclose() it !! fflush() does not >> help either. > >Gareth Jones responded: > >> I have encountered this too, but in the system I worked on it >> was not an issue. However, I remember reading/being-told >> some time ago that the reason for this is that the directory >> entry on the host system (presumably SUN?) is not updated until >> the file is closed? Is this actually the case? could someone >> shed a little light on this one? > >First off, you have to make sure that you are actually opening the >file you want on the NFS mounted disk; not the default netDrv mounted >disk. If you don't specify the correct path name on your open, you >can end up with the file actually being opened on the wrong device. >Use the iosFdShow command to indicate what drive the file is being >opened with. The iosDrvShow command shows entry points for each of >the drives. Using lkAddr on the driver entry points can tell you >which drive is which. > >Second, there is the issue of buffering. By default, NFS buffers up >data in 8K chunks and doesn't write it out to disk until that much >data has accumulated. Probably one of these reasons is why the >data doesn't seem to be getting written out. If you write more than >8KB to a file that is actually being handled by NFS, it has always >showed up on the other system in my experience (even if you don't >flush the data). Fred > > | Fred J Roeber, BBN Systems & Technologies | > | 4 John Clarke Road Middletown, RI 02842-5202 | > | froeber@bbn.com 401-848-3548 | > >********** I've experienced the same phenomenon with disk logging. If you want to flush the output to the NFS server without closing the file, you can either use fsync(open file descriptor) or ioctl (open file descriptor, FIOSYNC, 0) I've used this for logging status information on my target to a Unix host, so I really didn't want to keep opening and closing the file, but I also wasn't transmitting enough information to automagically flush the cache. Joel Katz, Principal Engineer --- (603)-577-5895 Acuity Imaging, 9 Townsend West, Nashua, NH 03063 --- (603)-577-5965 (FAX) From frederick.g.linnemann@lmco.com Tue Jan 20 07:03:39 1998 From: frederick.g.linnemann@lmco.com Date: Tue Jan 20 07:03:43 PST 1998 Subject: Re: vxworks-users Diagnostics on a MVME1603 board running un You might try Algorithmics - a British company that has BSPs for Wind River Systems. They have several low cost $1K -$5K packages including debug sw routines ... we have used their MIPS ISA package with success. They may have one for PPC ... their on the WWW - suggest you do a find as I don't recall their Website address. Fred Linnemann Sanders 603-885-4139 ______________________________ Reply Separator _________________________________ Subject: vxworks-users Diagnostics on a MVME1603 board running under Author: Susan J Henry at INTERNET Date: 1/16/98 8:04 AM Hi, We're working on an embedded stand-alone system (no networking) that resides on a ship. If a failure is reported or the system isn't functioning properly, the sailor will run diagnostics on the PPC MVME1603 board. Since there is no built-in-test on this board, we need to develop diagnostic tests to check out the integrity of the board. Motorola does deliver PPCBUG with the board, however we need to run diagnostics while our application is running under VxWorks. In addition, the sailor is not allowed to remove and replace components on the board. Has anyone developed a set a diagnostics for the 1603 that runs under VxWorks? Has anyone successfully incorporated PPCBUGs source code into an application that runs under vxWorks? TIA, Sue Henry Lockheed Martin Launching Systems 410.682.0776 susan.j.henry@lmco.com ======================================================================== Help regarding Majordomo: http://www.epic.lmco.com/Majordomo/Docs/majordomo.basic.html Archives for this list: http://www.epic.lmco.com/Majordomo/Archives/vxworks-users Reply to the whole list: mailto:vxworks-users@epic.lmco.com ======================================================================== From cgrames@mdc.com Tue Jan 20 14:26:38 1998 From: Charlie Grames Date: Tue Jan 20 14:26:42 PST 1998 Subject: Bad VME Interrupt 0 VxWorks 5.3.1 MVME2604 1.1/3 All, I am getting a sporadic system hang on one of my systems which contains an MVME2604 running AIX and another running VxWorks. It also contains a VMIC interrupter module which interrupts on VME level 5. The hang is preceded with the following messages at the console: interrupt: received unknown interrupt 0 interrupt: bad vme interrupt 0 Has anyone seen this before, and do you know the cause and/or solution? Any help would be much appreciated. Charlie Grames The Boeing Company (314) 233-1956 Charles.R.Grames@boeing.com From bm417929@muenchen.org Tue Jan 20 15:37:33 1998 From: bm417929@muenchen.org (Max Seeger) Date: Tue Jan 20 15:37:37 PST 1998 Subject: Support for PPC750 and Pentium?? Hi, does anybody know, when WRS plans to release the VxWorks for a PowerPC 750 and the Pentium? Thanks for help, Max From enrico.ferro@elsag.it Wed Jan 21 02:44:22 1998 From: Enrico Ferro Date: Wed Jan 21 02:44:27 PST 1998 Subject: BSP from 1.0 to 1.0.1 Hi, we have developed a simple BSP for a board having a Motorola 68360 CPU under Tornado 1.0 / VxWorks 5.3. Last week I received the upgrade package Tornado 1.0.1 and VXWorks 5.3.1. I installed it and immediately tried to rebuild my BSP. The problem is, I got plenty of compilation errors and I could not rebuild my BSP. I did not expect to encounter any problems just because I upgraded. Is there someone out there who has experienced similar problems? Thanks in advance to anyone who may reply Enrico Ferro Elsag Bailey Hartmann & Braun S.p.A. Via Hermada 6, 16154 Genova - Italy Tel. 0039-10-6586222 --- Fax. 0039-10-6586210 Home 0039-10-6517797 Work E-mail Enrico.Ferro@elsag.it Backup work e-mail address Enrico.Ferro@usa.net Home e-mail cluinfo@clu.assicomitalia.it From daemon@csg.lbl.gov Wed Jan 21 04:00:26 1998 From: daemon@csg.lbl.gov Date: Wed Jan 21 04:00:30 PST 1998 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Wed Jan 21 04:00:24 PST 1998 Subject: Re: Real Time NT ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Real Time NT Date: Tue, 20 Jan 1998 10:51:50 GMT From: "Marc de Bruin" Organization: Oce-Technologies B.V. - Research & Development Message-ID: References: <199801162035.MAA26420@lbl.gov> Sender: news@oce.nl (The Daily News @ nntp01.oce.nl) >Wind River has no plans to support real time NT. They do have some >product that does windows, but that doesn't allow any Microsoft >windows programs to run. I can't understand in the long run >how they can hope to compete with a real time NT O.S. when GUIs >are required. Is it just a matter of time before vxWorks will only be in >small embedded systems like printers? > >Has anyone our there done an evaluation on Real Time NT O.S.s? Take a look at http://www.lp-elektronik.com/ They provide vxWorks and WinNT, running together on one processor. I am using it, and I like it a lot! Marc. --------------------------- End of New-News digest ********************** From Richard.W.Mattox@lmco.com Wed Jan 21 05:07:43 1998 From: Richard W Mattox Date: Wed Jan 21 05:07:47 PST 1998 Subject: Ada vs C/C++ Hello Fellow VxWorks Users, I requested input about which language or languages developers were using and got several responses. First of all, thanks to all of those who did respond. Secondly, a few people requested that I post any statistics that I collected. Well, here are the results: 43% C++ 43% C 14% Ada83 *0% Ada95 14 Total Responses received. *Note : one person did say they were looking at moving from Ada83 to Ada95 I printed out all of the responses I received and saved them. If anyone would like a copy of the responses faxed to them I would be glad to do so. Just drop me an e-mail message with your fax number. Again, thanks to all that participated. Rick Mattox 410.682.2315 richard.w.mattox@lmco.com From cgrames@mdc.com Wed Jan 21 06:19:03 1998 From: Charlie Grames Date: Wed Jan 21 06:19:06 PST 1998 Subject: Support for PPC750 and Pentium?? -Reply Wind River has BSPs for x86 processors, but I don't know any details. As far as the PPC750 is concerned, what platform are you considering? Motorola's MVME2700, which incorporates the 750, is in production now. Motorola currently supports the BSP, which is a derivative of the MVME2604 BSP, and hopefully Wind River will pick up support for it in the not-too-distant future. HTH Charlie Grames The Boeing Company (314) 233-1956 Charles.R.Grames@boeing.com >>> the vxWorks Users Group Exploder 01/20/98 05:37pm >>> Submitted-by bm417929@muenchen.org Tue Jan 20 15:37:33 1998 Submitted-by: bm417929@muenchen.org (Max Seeger) Hi, does anybody know, when WRS plans to release the VxWorks for a PowerPC 750 and the Pentium? Thanks for help, Max From Richard.W.Mattox@lmco.com Wed Jan 21 06:29:04 1998 From: Richard W Mattox Date: Wed Jan 21 06:29:07 PST 1998 Subject: Results of the language poll Hello Fellow VxWorks Users, I will merge the results of the poll into a text file so that I can send electronic copies instead of faxes. So if anyone is interested (2 people so far) I will send you the results electronically if you send me e-mail message directly to the address below. I again reiterate that this is not an exhaustive poll and it does not state that no one is using Ada95. It states only that from my limited sample there was no one using Ada95 and that one person was planning on moving to Ada95 from Ada83. Rick Mattox 410.682.2315 richard.w.mattox@lmco.com From hao@ct.med.ge.com Wed Jan 21 08:18:47 1998 From: "L. Hao 5-4267 (Pager)558-6419" Date: Wed Jan 21 08:18:52 PST 1998 Subject: fcntl questions Hi, all VxWorks users, Does anyone know how to get around VxWorks's missing fcntl() to successfully setup non-blocking I/O. ioctl(sFd, FIONBIO, 1) only works for sockets. I have fds for pipes and am trying to find out how to make it nonblocking. Any help is greatly appreciated. Thanks -- Li Hao G.E. Medical Systems email:Li.Hao@med.ge.com Phone:785-4267 Pager:558-6419 From alai@squid.jpl.nasa.gov Wed Jan 21 09:57:06 1998 From: Anthony Lai Date: Wed Jan 21 09:57:10 PST 1998 Subject: Poll for ControlShell experience ... Hi, I have seen some very nice demo of a state-machine type CASE tool called ControlShell made by RTI (http://www.rti.com/products/cs.html). I have a thought to use this CASE tool to control my VxWorks application and perform repeatable testing. I want to know if anyone can send me a few comments about their experience with this tool. Thanks a lot. Anthony ******************************************************************** - Anthony S. Lai --------------------------------------------------- - Jet Propulsion Laboratory --------------------------- = 4800 Oak Grove Drive, Mail Stop: 79-24 * ======= ** ======= = Pasadena, CA 91109 ** == = = ** ====== = Voice: (818) 354-0473 ** ======= ** ===== = Fax: (818) 393-4878 ****** == ****** === *************************************** ***** == ******* ** ******************************************************************** From cgrames@mdc.com Wed Jan 21 12:00:24 1998 From: Charlie Grames Date: Wed Jan 21 12:00:28 PST 1998 Subject: Bad VME Interrupt 0 VxWorks 5.3.1 MVME2604 1.1/3 Thanks to all who responded to my posting about "bad VME interrupt 0." I located the errata regarding VME interrupts on the Universe chip and have duly noted the information. It turns out, in our case, that the VMIC interrupter module we were using went bad. We moved it to another system, and the problem moved with it. We have another module that seems to be working OK. Charlie Grames The Boeing Company (314) 233-1956 Charles.R.Grames@boeing.com From Stephen.Carruthers.scarruth@nt.com Wed Jan 21 22:35:24 1998 From: "Stephen Carruthers" Date: Wed Jan 21 22:35:27 PST 1998 Subject: FW: Queued signal problem in Tornado >Calling all VxWorks hackers, > >We are having a problem queuing signals in Tornado, as follows. Does anyone >else know about this? > >Thanks muchly, > - Steve > >Stephen Carruthers >Nortel/Northern Telecom >Wireless Networks Product Development >Calgary, Alberta, Canada > >------------------------ >The code below can be compiled to one task. When it is running, it blocks the >signal MY_SIG until it calls sigwaitinfo(). According to POSIX, if this task >receives MY_SIG before calling sigwaitinfo(), the signal should be queued and >will be handled after sigwaitinfo(). >However, in VxWorks Tornado 1.0.1, such signal is not queued. It is lost if >it is sent before sigwaitinfo(). > >/*================================================== >sig_rcvr.cxx >==================================================*/ >#include >#include >#include >#include >#include > >#define MY_SIG SIGRTMIN + 1 > >void sig_handler(char *info) >{ > cout << "Signal received!" << endl; >} > >void sig_receiver() >{ > // Setup signal mask > sigset_t mysigset; > > sigemptyset(&mysigset); > sigaddset(&mysigset, MY_SIG); > sigprocmask(SIG_BLOCK, &mysigset, NULL); > > // Do something > for (int counter = 0; counter < 30; counter++) { > for (long j = 0; j < 2000000; j++); > cout << counter << '\r' << flush; > } > > // Handle signal > siginfo_t mysiginfo; > > if (sigwaitinfo(&mysigset, &mysiginfo) < 0) { > perror("sigwaitinfo"); > exit(1); > } > > void *mystr = mysiginfo.si_value.sival_ptr; > sig_handler((char *) mystr); >} > > > From sergi@bpo.hp.com Thu Jan 22 01:21:03 1998 From: Sergi Casas Date: Thu Jan 22 01:21:07 PST 1998 Subject: Re: FW: Queued signal problem in Tornado This is a multi-part message in MIME format. --------------2CE515E83765A78C89C84315 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit the vxWorks Users Group Exploder wrote: > Submitted-by Stephen.Carruthers.scarruth@nt.com Wed Jan 21 22:35:24 1998 > Submitted-by: "Stephen Carruthers" > > >Calling all VxWorks hackers, > > > >We are having a problem queuing signals in Tornado, as follows. Does anyone > >else know about this? > > > ... > >/*================================================== > >sig_rcvr.cxx > >==================================================*/ > >#include > >#include > >#include > >#include > >#include > > > >#define MY_SIG SIGRTMIN + 1 I have also experienced that many vxWorks signals do not block correctly using sigsetmask() or similars. If not properly blocked the symptom would be what you are seeing, so I'd try with some different signal number (signals below 20 seem to work better; I use SIGTERM and it blocks well). Hope it helps. -- Sergi Casas Software Engineer (R&D) Hewlett-Packard Co. Barcelona Division - Spain --------------2CE515E83765A78C89C84315 Content-Type: text/x-vcard; charset=us-ascii; name="vcard.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Sergi Casas Content-Disposition: attachment; filename="vcard.vcf" begin: vcard fn: Sergi Casas n: Casas;Sergi org: Hewlett Packard - Barcelona Division email;internet: sergi@bpo.hp.com title: Software Engineer (R&D) x-mozilla-cpt: ;0 x-mozilla-html: TRUE version: 2.1 end: vcard --------------2CE515E83765A78C89C84315-- From jshaw_shade@compuserve.com Thu Jan 22 02:13:37 1998 From: John Shaw Date: Thu Jan 22 02:13:41 PST 1998 Subject: SLIP between VxWorks and PC DOS Greetings, Has anyone succeeded in establishing a SLIP-based connection between VxWorks (5.2) and a PC running MS-DOS 5 or above? I have had a hunt around for public domain software to provide a TCP/IP stack and SLIP support without much success. I want the connection between a processor running VxWorks and the PC to be direct, port to port, without modems. I also want to avoid any Windows packages... the application I want to run on the PC must be entirely DOS-based. I would like to use SLIP to establish TCP connections across a serial line between user applications written for either end. Thanks. =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D John Shaw, Shade (Computer Services) Limited =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D Email: jshaw_shade@compuserve.com= From tei@adv.sbc.sony.co.jp Thu Jan 22 02:46:20 1998 From: Tom Isaacson Date: Thu Jan 22 02:46:25 PST 1998 Subject: Re: BSP from 1.0 to 1.0.1 At 02:44 21/01/98 PST, you wrote: >Hi, >we have developed a simple BSP for a board having a Motorola 68360 CPU under >Tornado 1.0 / VxWorks 5.3. > >Last week I received the upgrade package Tornado 1.0.1 and VXWorks 5.3.1. I >installed it and immediately tried to rebuild my BSP. The problem is, I got >plenty of compilation errors and I could not rebuild my BSP. > >I did not expect to encounter any problems just because I upgraded. Is there >someone out there who has experienced similar problems? > >Thanks in advance to anyone who may reply >Enrico Ferro >Elsag Bailey Hartmann & Braun S.p.A. >Via Hermada 6, 16154 Genova - Italy >Tel. 0039-10-6586222 --- Fax. 0039-10-6586210 >Home 0039-10-6517797 >Work E-mail Enrico.Ferro@elsag.it >Backup work e-mail address Enrico.Ferro@usa.net >Home e-mail cluinfo@clu.assicomitalia.it I am also having problems getting Tornado 1.0.1 working (on a Motorola 63840), but I didn't have any problems compiling the BSP. Have you made the necessary changes to your setup that are specified in the installation manual? What is your host platform? I am using Unix (sun4-solaris2). I have managed to compile my BSP succeessfully, but I can't get my code to download to the board when I am running the BSP. It rejects some of the object files. Has anyone else experienced this problem? ____________________________________________________________________ Tom Isaacson . o o O O O Systems Product Development (SPD), . o o O O Sony Broadcast & Professional Europe, . o o O O Jays Close, Viables, . o o O O O Basingstoke. Tel: +44 (0) 1256 355011 . o o O O O O RG22 4SB DDI: +44 (0) 1256 483460 . o o O O O O United Kingdom Fax: +44 (0) 1256 483591 . o o O O O mailto:tei@adv.sbc.sony.co.jp ____________________________________________________________________ From daemon@csg.lbl.gov Thu Jan 22 04:01:59 1998 From: daemon@csg.lbl.gov Date: Thu Jan 22 04:02:03 PST 1998 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Thu Jan 22 04:01:57 PST 1998 Subject: Re: Soundblaster in VxWorks? Subject: Re: NFS for Windows NT 4.0 ??? Subject: VME, Motorola and VxWorks ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Soundblaster in VxWorks? Date: Wed, 21 Jan 1998 08:10:57 -0500 From: Kevin Bradley Organization: Doctoral student, Electrical and Computer Engineering, Carnegie Mellon, Pittsburgh, PA Message-ID: <0olTDVG00iWP01NLs0@andrew.cmu.edu> References: <34C5AAC5.79DF8D0F@cts.com> Excerpts from netnews.comp.os.vxworks: 20-Jan-98 Soundblaster in VxWorks? by Ingo Pakleppa@cts.com > I just got the task to throw together some software for using a > SoundBlaster under VxWorks (486 hardware). I don't know much about > VxWorks yet and was wondering whether somebody already wrote a driver > and has that for sale, or alternately has some partially working (or > even non-working) code that would give me a head start. Any help at all > would be appreciated! I've ported something called SMIX from Ethan Brodsky for VxWorks for the SoundBlaster 16 only; it could probably be migrated to other SoundBlaster platforms. SMIX comes from: http://www.xraylith.wisc.edu/~ebrodsky/ There's also a commercial driver from 4Front Technologies: http://www.4front-tech.com/vxworks.html I've not tried their driver, but I know they support Linux and other Unixes quite well, and provide full-featured drivers. My driver only does output and MIDI read, 'cause that's what I needed for my project; I tried simultaneous input / output, but got hung up and just skipped it. I'd already gotten it working when I heard about 4Front's, but there Driver features: 1) up to 20 simultaneous output channels at different sampling rates supported. All outputs are mixed together at 44.1 kHz stereo. 2) Blocking write. 3) Driver does sampling rate and multi-channel conversions. Unfortunately, performance is rather spotty -- zero-order hold sampling rate conversion runs at about 20% CPU on a Pentium / 75 (11.25 kHz to 44.1), polyphase at about 80%. I also have code for a bandlimited interpolator, but it's got some problems still. Anyway, the driver is usable in a variety of applications. I've used it in MPEG audio players and video games on VxWorks, and in a short-lived software MIDI synthesizer (read MIDI event, spawn note). If you're interested, drop me a line, and I'll tar up what I have. -- Kevin --------------------------- Newsgroups: comp.os.vxworks,comp.realtime Subject: Re: NFS for Windows NT 4.0 ??? Date: Wed, 14 Jan 1998 21:25:38 +0000 From: Philippe Le Foll Organization: Wanadoo - (Client of French Internet Provider) Message-ID: <34BD2D51.46AA913A@fridu.com> References: <34B37225.1E656C2D@wawel.cyfronet.krakow.pl> <698p71$3dg$1@winter.news.erols.com> <34BA1CF6.2F6754F0@ornl.gov> Dennis Depp wrote: > Dave, > > While Samba works very well to map Unix drives on an NT server, I don't think > it will work in the other direction, ie. you can't map a disk on an NT server > to a Unix workstation. > > Dennis Depp, d5p@ornl.gov > This mainly depend on your UNIX, Linux can have some routines of samba include directely in 2.0.x kernel, this allows you to mount a NT or WIN95 disk from your Unix box whit almost no configuration (and no security ...). If you want to access a disk from VxWorks, you could also port some routine of samba on VxWorks, it should not be a huge job. I personnally ported SUN ONC/RPC on NT, and SUN C code almost compile with no modification under NT, some distributions are freely avaliable, and one include a small and pretty huggly NFS server that could be enough for VxWorks to write some data on your PC disk. The best alternative to me is surely to port Samba on VxWorks I even don't understand why Wind River don't make it. Philippe --------------------------- Newsgroups: comp.os.vxworks,comp.dsp,comp.realtime,sci.engr.control,comp.arch.embedded,comp.arch.bus.vmebus Subject: VME, Motorola and VxWorks Date: Thu, 15 Jan 1998 19:36:43 +0100 From: Friedrich Scheurer / ZFS Organization: Zentrum Fertigungstechnik Stuttgart Message-ID: <34BE573B.A3@zfssun6.tz.rus.uni-stuttgart.de> Hi, searching for a Control-Prototyping-Hardware that runs under VxWorks several questions occured. Since i am not so familiar with electronics, bus systems and computer-hardware i am posting my questions: What throughput time can i get on the VME-Bus using single read actions? Does it mostly depend on the used master-CPU? What improvement can be expected transfering connected datablocks of 10-15 words? What throughput time has the TUNDRA-Universe PCI-to-VME-Bridge, recently used on single board computers, for those above described accesses? I heard about some problems of the MVME 26XX PowerPC-Boards in accessing VME-Bus through the TUNDRA-Bridge? - - interrupt latency of some 10 microseconds - - considerable context switching time under VxWorks F. Scheurer +-------------------------------------------------+ | | | Dipl.-Ing. F. Scheurer | | | | c/o Zentrum Fertigungstechnik Stuttgart | | Center for Production Technology Stuttgart | | Nobelstrasse 15 | | D-70569 Stuttgart | | | | Telefon +49 711 13162-41(-0) | | Fax +49 711 13162-11 | | E-Mail scheurer@zfs.rus.uni-stuttgart.de | | zfssrr@zfssun6.tz.rus.uni-stuttgart.de | | | +-------------------------------------------------+ | | | privat: | | Telefon +49 7161 686362 | | Fax +49 7161 686301 | | E-Mail Friedrich.Scheurer@t-online.de | | | +-------------------------------------------------+ --------------------------- End of New-News digest ********************** From msaunders@taz.dra.hmg.gb Thu Jan 22 04:37:24 1998 From: "Land Systems 4" Date: Thu Jan 22 04:37:28 PST 1998 Subject: (Fwd) DCRsi interface to VME-VxWorks We need to interface a high speed digitiser with a DCRsi output into a PowerPC based VME system running VxWorks, has anybody got any ideas? ls4@taz.dra.hmg.gb From cjgarcia@unex.es Thu Jan 22 06:29:06 1998 From: "Carlos J. Garcia Orellana" Date: Thu Jan 22 06:29:18 PST 1998 Subject: About 3Com net cards drivers for VxWorks Hello, Has anybody a driver for de Fast Ethernet 3Com network cards for = VxWorks? Thanks in advance. Carlos Javier Garc=EDa Orellana Dpto. de Electronica Universidad de Extremadura SPAIN e-mail: carlos@nernet.unex.es From prager@tcville.es.hac.com Thu Jan 22 07:48:58 1998 From: Kenneth Prager Date: Thu Jan 22 07:49:02 PST 1998 Subject: Re: (Fwd) DCRsi interface to VME-VxWorks >Submitted-by msaunders@taz.dra.hmg.gb Thu Jan 22 04:37:24 1998 >Submitted-by: "Land Systems 4" > >We need to interface a high speed digitiser with a DCRsi output into >a PowerPC based VME system running VxWorks, has anybody got any >ideas? >ls4@taz.dra.hmg.gb There is a company named Tri-Plex Systems which makes a VME-based DCRSi interface. Their telephone number was 301-290-7711 the last time I had contact with them (3 or 4 years ago). Ken Prager From cjgarcia@unex.es Thu Jan 22 08:00:28 1998 From: "Carlos J. Garcia Orellana" Date: Thu Jan 22 08:00:32 PST 1998 Subject: Bootting VxWorks from PCI net card ROMs Hello, I works with VxWorks for Intel x86, and I=B4m interested in bootting = VxWorks with a EPROM, so, I have thought it could be using the PCI net = card boot ROM.=20 But, how can I do it? Thanks. Carlos Javier Garcia Orellana Universidad de Extremadura SPAIN e-mail: carlos@nernet.unex.es From eugenel@flash.condorsys.com Thu Jan 22 10:01:58 1998 From: eugenel@flash.condorsys.com (Eugene Leung x714) Date: Thu Jan 22 10:02:02 PST 1998 Subject: compress and decompress application Hi all, The application that we are working with got too large for the available flash. We are using tornado 1.0.1 with the MVME2603. We want to compress the application to flash and during boot, uncompress and load the application. My questions are: What are the procedures to compress (using deflate or compress?)? Do I simply run one of these routines? how to get the size be determined by the bootcode? When calling inflate to decompress, should I malloc the required space to inflate the application to prior to doing a ld? Any gotcha I should be aware of? Please reply directly to eugenel@condorsys.com. TIA. Eugene Leung From frank@info-age.net Fri Jan 23 01:08:32 1998 From: Frank Fischer Date: Fri Jan 23 01:08:38 PST 1998 Subject: Problem with USER_RESERVED_MEM This is a multi-part message in MIME format. --------------23EFC651F6BC84696AC7619F Content-Type: text/plain; charset=us-ascii Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Content-Transfer-Encoding: 7bit Hi... I am using Tornado 1.0.1 on a PowerPC ( 860/821 ). We have to split the memory in two parts: one with cache enabled and another without cache. All the memory is on the same module and is handeled by the same memory controller. The memory is then used by external devices. To keep the memory located at the very end of the memory map clear of vxWorks I set the define USER_RESERVED_MEM on a valid value. The effects are quite curious. Everything seems to run normal. All functions sys* are running as expected. But functions like printf return the first argument and nothing happens at all. As soon as I set USER_RESERVED_MEM to zero everything is allright. If anyone could explain this effect and fix my world...!! Thanx... CU 0xff Frank Fischer Embedded Systems Design DHF - Gesellschaft fuer Datenverarbeitung --------------23EFC651F6BC84696AC7619F Content-Type: text/x-vcard; charset=us-ascii; name="vcard.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Frank Fischer Content-Disposition: attachment; filename="vcard.vcf" begin: vcard fn: Frank Fischer n: Fischer;Frank org: DHF - Gesellschaft für Datenverarbeitung email;internet: frank@info-age.net x-mozilla-cpt: ;0 x-mozilla-html: FALSE end: vcard --------------23EFC651F6BC84696AC7619F-- From info@triada.si Fri Jan 23 01:50:35 1998 From: Jure Ferbezar Date: Fri Jan 23 01:50:39 PST 1998 Subject: Re: SLIP between VxWorks and PC DOS >Has anyone succeeded in establishing a SLIP-based connection between >VxWorks (5.2) and a PC running MS-DOS 5 or above? yes we were using telnet client on the PC conected over slip to Vxworks 5.3 The following drivers and SW worked: slip driver by Peter Tattam trumpet TCP packet driver by Peter Tattam telnet client MUDCALLER by Peter Haight NCSA telnet client Try with this. ----------------------------------------------------------- Jure Ferbezar Triada d.o.o. Rozna dolina 9/3 1000 Ljubljana Slovenia Europe fax: +386-61-1234-166 tel: +386-61-1234-891 ----------------------------------------------------------- From daemon@csg.lbl.gov Fri Jan 23 04:01:22 1998 From: daemon@csg.lbl.gov Date: Fri Jan 23 04:01:25 PST 1998 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Fri Jan 23 04:01:20 PST 1998 Subject: Re: ppp implementation for vxworks ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: ppp implementation for vxworks Date: Sat, 17 Jan 1998 05:35:34 GMT From: jdeffner@west.net (john deffner) Organization: West.Net Communications Message-ID: <69pfjm$7kj@daffy.sb.west.net> References: <68vamr$f46$1@news00.btx.dtag.de> <34BDADC6.B48DF1E8@xmission.com> On Wed, 14 Jan 1998 23:33:43 -0700, Wes Peters wrote: >Martin Sauter wrote: >> does anyone know of companies that offer ppp solutions for VxWorks? On a reasonably related subject, does the slip implementation in tornado 1.0 have any known problems? I've been trying to use one of the serial ports on my x86 target as a slip port with little luck. I'm including the proper libraries and defines I think, but when I call slipInit from the shell it just comes back with -1. Not very informative. There's a slip client on the other end of the wire, and the hardware is known good. I'd hate to spend much more time on it and then find out that everyone but me knows it's always been broken. Any thoughts? --------------------------- End of New-News digest ********************** From bernard@iotek.ns.ca Fri Jan 23 06:54:53 1998 From: Bernard Whalen Date: Fri Jan 23 06:57:57 PST 1998 Subject: Booting from flash on MVME2604 Hi all, Hi have been attempting to boot an vxworks application from flash on a MVME2604 rather than from the network. I have been successful in putting this application in flash but I am having a problem with the network when this application runs. The application opens a tcp socket with a workstation. This works correctly when booting from the network but not when booting from flash. This board has the Open Firmware which trashes the boot line so I run bootParamsPrompt and sysNvRamSet to setup the network interface. After doing this the I see the familiar network initialization ("Attaching network interface dc0 ... done", etc). If I run tcpstatShow I see that the MVME2604 is receiving packets as the workstation attempts to connect but the MVME2604 is not sending any packets. I've gotten no response from WRS support on this one and I've run out of ideas to try. Has anyone successfully booted an application from flash on the MVME2604 that uses the network after booting? I am using Tornado 1.0.1 with all the lastest PPC patches. Thanks for any help. Bernie -- Bernie Whalen | MDA, 1000 Windmill Rd, Suite 60 | Ph: (902)481-3528 bernard@mda.ca | Dartmouth, N.S., Canada, B3B 1L7 | Fax:(902)468-2278 From mmcquade@aa.net Fri Jan 23 10:30:59 1998 From: Mike McQuade Date: Fri Jan 23 10:31:02 PST 1998 Subject: Re: Booting from flash on MVME2604 vxworks At 06:58 AM 1/23/98 PST, you wrote: >Submitted-by bernard@iotek.ns.ca Fri Jan 23 06:54:53 1998 >Submitted-by: Bernard Whalen > > >I've gotten no response from WRS support on this one and I've run out of >ideas to try. Has anyone successfully booted an application from flash on >the MVME2604 that uses the network after booting? > >I am using Tornado 1.0.1 with all the lastest PPC patches. > Bernie, Our instrument runs out of flash, and uses the network (FTP, NFS, HTTP), we have a problem with the LATEST version of the BSP (rev /4), with rev /4, we our application wont even startup. My point is that rev /4 of the BSP changed something, whatever it was, it hosed our application. We had to drop back to rev /3 of the BSP with ALL patches. You might want to update to PPC BUG, this will keep your NVRAM from getting trashed. The update is painless, and available on Motorola's FTP site. Mike. From espin@idiom.com Fri Jan 23 16:46:09 1998 From: Geoffrey Espin Date: Fri Jan 23 16:46:13 PST 1998 Subject: i960RPLV vxWorks 5.3.1 problems Folks, Problems with (Cyclone) i960RP BSP on VxWorks 5.3.1. Version # (from sysLib.c): "01d,19mar97,myz changed sysPhysMemTop for i960RD support." I've been trying to figure out bus errors for this board and have found and enabled BMER/MEAR registers, but it just a one-shot and I guess an NMI int-handler needs to reset this. Has anyone done this? Ultimately, we want to catch signals (SIGBUS/SEGV). However, I then wrote some test code to make sure signal handling in general was working... it doesn't look like it!?? E.g. I trap all signals 1-31 and then generate a divide-by-zero... only the built-in "arithmetic exception" is displayed. If I use "kill(tid, )" that works fine. Yep, INCLUDE_SIGNALS is defined, also INCLUDE_EXC_TASK/HANDLING. It appears that sysIntConnect(RP_VIRT_CORE_NMI, rtn, arg) probably needs to be hooked up... but how does "excExcHandle" and/or the underlying _func_sigExcKill, etc. get run? The supplied sysIntDefaultHandler() seems quite inadequate. Other problems: We have a 4M board, it appears the romInit.s calculation always comes up with 8M... anyone seen this? I've searched the Windsurf Reports to no avail. TIA, Geoff -- Geoffrey Espin espin@idiom.com From daemon@csg.lbl.gov Sat Jan 24 04:00:34 1998 From: daemon@csg.lbl.gov Date: Sat Jan 24 04:00:37 PST 1998 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sat Jan 24 04:00:31 PST 1998 Subject: Initialising Video Mode 640x400/800x600 on PC ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Initialising Video Mode 640x400/800x600 on PC Date: Sat, 17 Jan 1998 20:28:59 +0100 From: Torsten Rupp Organization: University of Karlsruhe, Germany Message-ID: <34C1067A.794BDF32@fzi.de> Hello, I'm looking for a source code fragment to initialise the video mode 640x400 or 800x600 (256 colors) on a standard PC without using a BIOS function (because the PC is running with vxWorks :-)). All examples I can find use PC BIOS interrupt 10h to enable the graphics mode. Thanx. Bye, Torsten - -- Computer Science Research Center (FZI) phone : +49-721-9654-302 Mobility Management and Robotics fax : +49-721-9654-309 Haid-und-Neu-Str. 10-14 e-mail: rupp@fzi.de D-76131 Karlsruhe (Germany) www : www.fzi.de/robot --------------------------- End of New-News digest ********************** From yzur@qualcomm.com Sun Jan 25 03:24:47 1998 From: Joseph Zur Date: Sun Jan 25 03:24:50 PST 1998 Subject: VxWorks on a PC card Hi, I am going to run VxWorks on a standard off the shelf PC card (386 or 486). I need some information on how the VxWorks runs on the CPU, does it come up after the BIOS or does it replaces the BIOS , how to eliminate the DOS, has anyone good pointers or references that apply ? Thanks in advance --- '0|' From daemon@csg.lbl.gov Sun Jan 25 04:01:24 1998 From: daemon@csg.lbl.gov Date: Sun Jan 25 04:01:28 PST 1998 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sun Jan 25 04:01:22 PST 1998 Subject: VxWorks drivers Subject: Re: Real Time NT Subject: Fibre Channel + vxworks Subject: CAN bus drivers? ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: VxWorks drivers Date: Fri, 16 Jan 1998 15:23:00 -0600 From: jeff.ogden@das.honeywell.com Organization: Deja News Posting Service Message-ID: <884985560.1173553161@dejanews.com> Reply-To: jeff.ogden@das.honeywell.com Is there a current FAQ for this group? The only FAQ I've been able to find was dated 1994. The main reason I ask is that I sure would like to know of lists or sites that might contain various VxWorks device drivers. Specifically, I'm looking for a driver for the AMD87C970 ethernet chip and a driver for the MPC106 chip (this is similair to the Raven chip on Motorola MVME2603 boards). Thanks! Jeff Ogden Honeywell jeff.ogden@das.honeywell.com - -------------------==== Posted via Deja News ====----------------------- http://www.dejanews.com/ Search, Read, Post to Usenet --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Real Time NT Date: Fri, 16 Jan 1998 16:45:11 -0500 From: Kevin Bradley Organization: Doctoral student, Electrical and Computer Engineering, Carnegie Mellon, Pittsburgh, PA Message-ID: References: <199801162035.MAA26420@lbl.gov> Excerpts from netnews.comp.os.vxworks: 16-Jan-98 Real Time NT by "Mark Menge"@pluto.dspt. > Wind River has no plans to support real time NT. They do have some > product that does windows, but that doesn't allow any Microsoft > windows programs to run. I can't understand in the long run > how they can hope to compete with a real time NT O.S. when GUIs > are required. Is it just a matter of time before vxWorks will only be in > small embedded systems like printers? Just as an aside, I think that targets requiring Microsoft products and hard real-time are an extreme subset. Unless, of course, the Martians want to run Word. Also, the real-time NT OS's are _still_ resource hungry. Having a 32 MB + 1 GB hard drive requirement on all target-platforms may be a bit expensive compared to a card with 8 MB + 1MB boot ROM. I'm pulling requirements for NT out of thin air; I think if you really work at it, you can run NT in 8 MB of RAM + 10 MB hard drive, but then you probably won't be able to run Word or PowerPoint, which is not what you're talking about. There's also an issue of "WindRiver supporing real-time NT". What does that mean, in general? Should VxWorks run Word, or support Win32, or what? There's a product that supports Win32 called RT-Win, which allows concurrent VxWorks and Win95. There's also a variety of user interfaces for VxWorks platforms. > Has anyone our there done an evaluation on Real Time NT O.S.s? There was a flurry of activity about this on comp.realtime a few months ago, so you'd probably be able to find posts there. No-one has really done a hard look at the real-time extensions yet. The consensus seems to be that they help out, but are still not perfect. We were contemplating doing an analysis, but the extensions we got from VenturCom were not instrumented at all, so we're punting to a later date (if at all). When I say instrumented, I mean something akin to WindView that would really easily allow their behavior to be observed and analyzed. Having to instrument the OS without source-code is a pain. There was one interesting evaluation made by some guys at GM, but it didn't seem to draw any conclusions. http://www.arcweb.com/omac/Documents/doc_list.htm Also, the real-time encyclopedia has some information. http://www.Realtime-info.Be/encyc/market/rtos/ntspecial.Htm -- Kevin --------------------------- Newsgroups: comp.os.vxworks Subject: Fibre Channel + vxworks Date: Sat, 24 Jan 1998 18:55:17 +0000 From: Michael Morrison Organization: My House Message-ID: <34CA3915.C68598CC@ringo.reno.nv.us> Hi there, Does anyone know of a manufactorer of PCI Fibre Channel cards with VxWorks drivers? Thanks mike@ringo.reno.nv.us mikem@sdlabs.com --------------------------- Newsgroups: comp.os.vxworks Subject: CAN bus drivers? Date: Mon, 19 Jan 1998 15:09:58 -0800 From: Mike Horan Message-ID: <34C3DD46.101F@dynatem.com> Does anyone know if VxWorks drivers for the CAN bus are available? - -- Mike Horan PH: 714-855-3235 Dynatem Inc. FAX: 714-770-3481 23263 Madero, Suite C Email: steve@dynatem.com Mission Viejo, CA 92691 Web: http://www.dynatem.com --------------------------- End of New-News digest ********************** From jj_boom@signaal.nl Mon Jan 26 00:35:33 1998 From: jj_boom@signaal.nl (Boom J.J.) Date: Mon Jan 26 00:35:38 PST 1998 Subject: VME-DCRsi interface msaunders@taz.dra.hmg.gb wrote >We need to interface a high speed digitiser with a DCRsi output into >a PowerPC based VME system running VxWorks, has anybody got any >ideas? There is a company named MYRIAD Logic which makes a VME-based DCRsi interface. The address is: MYRIAD Logic 1109 Spring Street Silver Spring, MD20910 FAX (301) 588-0605 (301) 588-1900 Email: roc@myriadlogic.com Jan Boom ---------------[This message contains Unclassified information]---------------- | | | Jan Boom (db646) | | System Integration & Testing Department | | Hollandse Signaalapparaten B.V., | | P.O. Box 42 ____ ____ ____ ____ | | 7550 GD Hengelo, The Netherlands / / / / / /| / / / / / / | | Phone : +31-(0)74-2483118 /___ / /___ / | / /___/ /___/ / | | Fax : +31-(0)74-2484042 / / / / / | / / / / / / | | E-Mail : jj_boom@signaal.nl /___/ / /___/ / |/ / / / / /___ | | | ------------------------------------------------------------------------------- From thomas.arand@oen.siemens.de Mon Jan 26 00:42:05 1998 From: thomas.arand@oen.siemens.de (Arand, Thomas) Date: Mon Jan 26 00:42:08 PST 1998 Subject: IDTS 465 evaluation board Hi VxWorks World, does anyone in the world use the IDTS 465 evaluation board with the MIPS = = 4700 processor on it? When trying to download modules to the target I often get a = LOADER_RELOCATION_ERROR. Did anyone encounter this also? Thanks for answers, Thomas ---------------------------------------------------------------------- Thomas Arand Siemens AG Tel. +49 - 89 - 722 42296 Fax. +49 - 89 - 722 26572 thomas.arand=40oen.siemens.de From igal@tadirantele.com Mon Jan 26 07:53:16 1998 From: Igal Chernobelsky Date: Mon Jan 26 07:53:21 PST 1998 Subject: Remove from immediate VxWorks mail list This is remove from immediate VxWorks mail list request Thanks -- Igal Chernobelsky Wireless Systems Tadiran Telecommunications, Israel Tel: 972-3-9263600 Fax: 972-3-9263678 From bpringle@teklogix.com Mon Jan 26 16:00:51 1998 From: Bill Pringlemeir Date: Mon Jan 26 16:00:55 PST 1998 Subject: VxWorks, PPC 860/821 Data Cache and Decrementer. Hello, Has anyone investigated an alternative to using the PPC decrementer for the VxWorks timer? Would it be possible to use the CPM's PIT instead. According to the Errata for the data cache, the decrementer seems to be the only SPR that would be used during regular operation. We have A3 silicon (or better?). The idea is to use the PIT as a supplement to ppcDecTimer.c with a cpmPIT.c or something, then setup the MMU and data cache. Thanks, Bill Pringlemeir Software Engineer Teklogix Inc. Toronto, On. From daemon@csg.lbl.gov Tue Jan 27 04:01:58 1998 From: daemon@csg.lbl.gov Date: Tue Jan 27 04:02:01 PST 1998 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Tue Jan 27 04:01:56 PST 1998 Subject: Re: What do you think of WxWorks? ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: What do you think of WxWorks? Date: 26 Jan 1998 13:23:49 -0800 From: bny@crl9.crl.com (Bradley Yearwood) Organization: CRL Dialup Internet Access (415) 705-6060 [Login: guest] Message-ID: <6aiut5$cog@crl9.crl.com> References: <34C8C29A.F629296D@post4.tele.dk> <34CB4551.6E9CBD1A@ecitele.com> <34CC4D98.448B8F60@post4.tele.dk> In article <34CC4D98.448B8F60@post4.tele.dk>, Peter Holm wrote: > > >Chanoch Gotlieb wrote: > >> >> >> What exactly did you hear? What are the problems related to? It's fair > >I have a friend who just started working with VxWorks. She's experiencedin >Unix device drivers. > >The purpose of my posting was to get a more balanced impression. > >Based on the responses I conclude that: > >1) If you expect a Unix environment, where you can easily port public domain > >software, your out of luck. VxWorks is not Unix. No, but it offers more Unix-like services than one typically sees from a realtime operating system. This can substantially reduce developers' learning time, and makes the embedded target readily accessible to, for example, hardware engineers who want to write various sorts of casual test code. >2) Other than that (based on my limited research :-)) people seems more or >less >satisfied with the product. > >On the plus side : >- availability of ported 3rd party software >- support for lots of different CPUs >- documentation is OK The documentation is actually remarkably good. (I have found that it's even better on the rare occasions that I bother to read it.) >On the other side : >-disappointing technical support >-not robust and versatile All technical support for all software products is disappointing, unless you do the right thing and insist upon source code. Same applies to versatility. You, the customer, are in complete control of the demand side of the market. If you keep buying the wrong product, the market is encouraged to keep selling the wrong product. VxWorks seems robust (with the exception of what seem to be some annoying problems in management of FTP socket close in the boot loader, which is not relevant once you quit network booting, and load or run the OS from ROM), and is more versatile than any other realtime OS product that I have had experience with. This comes at a price, both in memory space (but VxWorks is quite modular, so this can be tailored to fit actual needs), and in cash. The payback in time-to-results can be very good, but one needs to judge this according to one's own situation. [I don't own any Wind River stock, unless it's buried deep in a mutual fund somewhere. Until and unless something gnice happens to the realtime world, I think VxWorks is well worth considering.] Brad Yearwood Cotati, California --------------------------- End of New-News digest ********************** From Justin.Beaver-1@kmail.ksc.nasa.gov Tue Jan 27 10:19:33 1998 From: "Beaver-1, Justin" Date: Tue Jan 27 10:19:44 PST 1998 Subject: VxWorks FTP Daemon does not support all FTP Commands VxWorks Community, Wind River's VxWorks FTP Daemon currently does not support the full complement of FTP commands (e.g. mkdir). Does anyone know of an alternate VxWorks FTP Daemon available that supports all of the FTP commands? Thanks, Justin Beaver __________________________________________________________________ Justin Beaver Mail Code: DP-3 Kennedy Space Center, FL 32899 (407) 861-7683 Justin.Beaver-1@ksc.nasa.gov From zhengyu.guo@kla-tencor.com Tue Jan 27 11:03:56 1998 From: Zhengyu Guo Date: Tue Jan 27 11:04:04 PST 1998 Subject: Re: comp.os.vxworks newsdigest Hi VxWorks Gurus, I have an application that we need to send live video over ethernet (10 BaseT or 100BaseT). Fortunately the images that need to be transmitted are semi-static. We are thinking of using MPEG. Given that I don't have much knowledge on MPEG, I have the following questions: 1). Which way to go is better: a). using hardware MPEG encoder board to send the compressed image, and then using software decoder on the receiving side. b). Completely use software solution (software encoder and decoder). Or are there any other alternatives? We'd like to get at least 20 frames/sec. 2). If a) is prefered, what board do you recommand for PCI or VME bus? We will have a SBC running VxWorks to control this board. Is it very difficult to write interface code/driver for this board? 3). The image transmitted is for review (either by human inspection or go through image processing algorithm), we'd like to control the loss during compression and de-compression. If the image difference is very small, we may decide not to compress it at all. Ideally we'd like to have a lossless transmission. Which method above can give us more control on loss? 4). If b) is preferred, are there any encoder or decoder source code to share/buy? I'd really appreciate your input. Forgive me, some of the questions may sound stupid. Thanks in advance! -Guo -- ----------------------------------------- Zhengyu Guo KLA-Tencor Corporation (408) 875-4643 Zhengyu.Guo@kla-tencor.com From mrana@ebmail.gdeb.com Tue Jan 27 11:17:38 1998 From: mrana@ebmail.gdeb.com Date: Tue Jan 27 11:17:46 PST 1998 Subject: Bad VME Interrupt I am getting a sporadic system hang on my systems which contain a MVME2604 running VxWorks 5.3.1. I am using this configuration in an ATM network. When the system hangs I get the following message on my screen: interrupt: received unknown interrupt 0 interrupt: bad vme interrupt 0 If anyone has seen this situation before, and knows the cause and/or solution please let me know? Mehul Rana General Dynamics mrana@ebmail.gdeb.com (860)433-4999 From mea@mclean.sparta.com Tue Jan 27 11:46:58 1998 From: Mike Anderson Date: Tue Jan 27 11:47:03 PST 1998 Greetings! Anyone out there in netland have a pointer or source for a VxWorks driver for the Greensprings dual PI/T IP module (Motorola 8230)? PD is preferred, but commercial is acceptable if it's in source. TIA, --------------------------------------------------------------------- |Michael E. Anderson | Voice -> 703 448 1683 x235 | |Chief Engineer | FAX -> 703 893 5494 | |mea@mclean.sparta.com | http://www.mclean.sparta.com | |___________________________|_______________________________________| | SPARTA, Inc. | | | /\\ C4I Operation |"Software Development is | | ///\\ 7926 Jones Branch Drive | like making a baby... | | /////\\ Suite 900 | You can't make a baby in | | ///// \\ McLean, VA 22102 | one month by impregnating | | / \\\\\ \\ 703-448-0210 | nine women. | | \ \\\\\ // | | | \ ////// | Some things just take | | \////// SPARTA | time." | | \//// Pride In Performance | | | \// -------------------- | | --------------------------------------------------------------------- From tkb@mclean.sparta.com Tue Jan 27 13:01:59 1998 From: Keith Buchanan Date: Tue Jan 27 13:02:06 PST 1998 Subject: Re: At 11:47 AM 1/27/98 PST, you wrote: >Submitted-by mea@mclean.sparta.com Tue Jan 27 11:46:58 1998 >Submitted-by: Mike Anderson > > >Greetings! > > Anyone out there in netland have a pointer or source for a >VxWorks driver for the Greensprings dual PI/T IP module (Motorola >8230)? PD is preferred, but commercial is acceptable if it's in >source. Todd had one at one point. adios From mfischer@shearwater.qualcomm.com Tue Jan 27 20:48:02 1998 From: Mark Fischer Date: Tue Jan 27 20:48:13 PST 1998 Subject: VxWorks- RAM disks? All, Does anyone know if VxWorks supports RAM disks? If so, how? Mark Fischer Qualcomm, Inc. From bob_anderson@void.mclean.mcd.mot.com Tue Jan 27 22:24:39 1998 From: Bob Anderson Date: Tue Jan 27 22:24:51 PST 1998 Subject: Serial Port 3 and 4 on 2600 Does anyone know where or have driver source code, that could be shared, for serial ports 3 and 4 on Motorola's MVME2600 running VxWorks? It is desired that it support the synchronous operation if possible. Thanks, Bob From jshaw_shade@compuserve.com Wed Jan 28 02:36:38 1998 From: John Shaw Date: Wed Jan 28 02:36:56 PST 1998 Subject: Re: VxWorks- RAM disks? Regarding the following:- >All, >Does anyone know if VxWorks supports RAM disks? If so, how? >Mark Fischer >Qualcomm, Inc. Yes is the answer. VxWorks 5.2 does and I would be surprised if Tornado did not. The steps I will outline below are for VxWorks 5.2.... 1) Add INCLUDE_RAMDRV to config.h in the directory containing the VxWorks image you wish to load. Remake the image. 2) After the remade VxWorks has loaded, run the following from the VxWorks shell:- descriptor=3DramDevCreate(0,512,400,400,0) dosFsMkfs("/ram0", descriptor) this assumes that support for DOS filesystems has been added to the VxWorks image. Now a ram disk, called /ram0, can be used and has been allocated from VxWorks memory. A DOS file- system has been placed on the disk. The disk occupies 200Kb. Local SRAM could be used as a ram disk by replacing the first ramDevCreate argument with the address of SRAM. If this is battery backed, the disk could become permanent. In this case use dosFsMkfs when creating the disk for the first time and use dosFsDevInit("/ram0", descriptor, 0) to re-establish any existing RAM disk contents after a reboot of the target. But local SRAM might not occupy 200Kb! Reduce the 400,400 args. I have successfully used RAM disks myself on many occasions. John Shaw Shade (Computer Services) Limited, Calne, England, UK. Email: jshaw_shade@compuserve.com= From nrd1jrb@nrd.ups.com Wed Jan 28 05:08:40 1998 From: Jim Bucciferro Date: Wed Jan 28 05:08:55 PST 1998 Subject: Re: VxWorks- RAM disks? the vxWorks Users Group Exploder wrote: > Submitted-by mfischer@shearwater.qualcomm.com Tue Jan 27 20:48:02 1998 > Submitted-by: Mark Fischer > > All, > > Does anyone know if VxWorks supports RAM disks? If so, how? > > Mark Fischer > Qualcomm, Inc. Mark, Yes, VxWorks supports RAM disks. To use do the following: 1. Add INCLUDE_RAMDISK in the bsp configuration. 2. To create the RAM disk: a. #include "ramdrv.h" #define RAMDISK_ID "E:" b. This is a function I use to create the RAM disk. STATUS setramdrive( int bytesblk, /* bytes per block */ int blksTrack, /* blocks per track */ int nBlocks, /* number of blocks */ char* devname /* name of device */ ) { BLK_DEV *pBlkDev; DOS_VOL_DESC *pVolDesc; if( bytesblk == 0 ) bytesblk = 512; if( blksTrack == 0 ) blksTrack = 400; if( nBlocks == 0 ) nBlocks = 400; pBlkDev = ramDevCreate(0,bytesblk,blksTrack,nBlocks,0); if( pBlkDev == NULL ) { printf("Creation of RAM Drive failed\n"); return(ERROR); } pVolDesc = dosFsMkfs( devname, pBlkDev ); if( pVolDesc == NULL ) { printf("Creation of RAM Drive File System failed\n"); return(ERROR); } printf("Creation of RAM disk complete!\n"); return (OK); } This is how I call it: setramdrive( 2048, 512, 2048, RAMDISK_ID ); It uses defaults if the values are 0. Let me know if you need any help JB From agiovino@stisierra.com Wed Jan 28 08:01:48 1998 From: "Anthony M. Giovino" Date: Wed Jan 28 08:02:09 PST 1998 Subject: Sparta MS-CPU36E VxWorks BSP references I am looking for contacts who have used the Sparta MS-CPU36E BSP. Here's a little background information on our situation. We have a working application that runs using pSOS+ (pRISM IDE under Solaris) and we're looking to port it to VxWorks. We have concerns with using third party BSP vendors and would like to ask a few questions to current users. Thanks in advance, Tony G. -- Anthony M. Giovino Sierra Research Division of STI Software Engineer 485 Cayuga Rd MS345 Tel: (716) 631-7389 Buffalo, N.Y. 14225 Fax: (716) 631-7849 email: agiovino@stisierra.com From art@iotek.ns.ca Wed Jan 28 11:23:15 1998 From: Art Cole Date: Wed Jan 28 11:26:19 PST 1998 Subject: FibreXpress Network cards Hi all: Does anyone out there have any experience in using a Systran FibreXpress PMC with a 2604 board running VxWorks 3.1?? We're not interested in specifics, but we would like to know things such as how long it took to get the thing working properly and/or if it ever worked properly (or at least as advertised). Thanks Art Cole ****************************************************************** * Art Cole MacDonald Dettwiler & Associates * * Systems Engineer 1000 Windmill Road, Suite 60 * * phone: (902)-481-3523 Dartmouth, Nova Scotia * * fax: (902)-468-2278 Canada * * email: art@iotek.ns.ca B3B 1L7 * ****************************************************************** From bdf@amherst.com Wed Jan 28 12:28:43 1998 From: Bruce Friedman Date: Wed Jan 28 12:28:48 PST 1998 Subject: VxWorks And The Socket Receive Function VxWorks Gurus - I hope someone can help as this problem has been driving me crazy for almost a week now -> PROBLEM: I am trying to create a function that will continually examine and read data from a non-blocking UDP Socket (the non-blocking part is important here). I can create the socket okay, but I cannot get it to read in a non-blocking mode. If I use the ioctl() function to set FIONBIO to be non blocking - it seems to never read anything; just passing over the recvfrom function like it wasn't there. However - if I put a statement (printf();) just before or just after the recvfrom statement, it works fine. I have also been able to successfully get it to block read - waiting until something is there - but this is no good for me. If I leave out the ioctl() call, it defaults to blocking. I have also put in an external delay timer as well as making all variables global. Why oh why does inserting a printf() work???? Since I can not have a printf() in the final version of my code, unfortunately this is not an option. I have placed the code that I am using at the end of this E-Mail in hopes that someone can find something wrong with it SYSTEM: VxWorks version V5.2C KERNEL: WIND version 2.4 Copyright Wind River Systems, Inc., 1984-1997 CPU: Digital Alpha VME 4/288. Processor #0 Memory Size: 0xd18400. BSP version V5.2C DEC Tulip Ethernet Drivers SAMPLE CODE: typedef unsigned int in_addr_t; #include "errno.h" #include "sys/types.h" #include "time.h" #include "vxWorks.h" #include "fioLib.h" /* printf */ #include "ioLib.h" #include "stdioLib.h" #include "taskLib.h" #include "memLib.h" /* free malloc */ /* #include "varargs.h" */ #include "sys/socket.h" /* bind recvfrom sendto setsockopt shutdown socket */ #include "sys/ioctl.h" /* bind recvfrom sendto setsockopt shutdown socket */ #include "netinet/in.h" /* inet_addr inet_ntoa_b */ #include "socketvar.h" /* bind recvfrom sendto setsockopt shutdown socket */ #include "arpa/inet.h" #include "inetLib.h" /* inet_addr inet_ntoa_b */ #include "hostLib.h" /* gethostname */ #include "etherLib.h" #include "ifLib.h" #include "if_tu.h" #include "netdb.h" typedef UINT_32 SOCKET; #define INVALID_SOCKET (SOCKET)(~0) #define SOCKET_ERROR (-1) void main(void) { char Recv_Data[4096]; SOCKET The_Socket; UINT_16 The_Port = 5001; UINT_16 The_Type = SOCK_DGRAM; UINT_16 The_Peers_IP = INADDR_ANY; int Sockaddr_Size = sizeof(struct sockaddr_in); int Bytes_Rxed = 0; unsigned long Sock_Opt_ul = 0; struct sockaddr_in P2P_Rx_Sockaddr_Info; struct sockaddr_in Rx_Sockaddr_Info; double Run_Duration_Sec = 100.0; int Done = 0; int fionbio = FIONBIO; int Counter = 0; /*------------------------------------*/ /* Create the peer to peer Rx socket. */ bzero((char *)&P2P_Rx_Sockaddr_Info, Sockaddr_Size); P2P_Rx_Sockaddr_Info.sin_family = AF_INET; P2P_Rx_Sockaddr_Info.sin_addr.s_addr = htonl(The_Peers_IP); P2P_Rx_Sockaddr_Info.sin_port = htons(The_Port); The_Socket = socket(AF_INET, SOCK_DGRAM, 0); ioctl(The_Socket, fionbio, &Sock_Opt_ul); bind(The_Socket, (struct sockaddr *) &P2P_Rx_Sockaddr_Info, Sockaddr_Size); printf("\nI'm looking for friends on port %u.\n", The_Port); /* Messaging loop. */ while (!Done) { /* Receive data on this port (non-blocking) and process it. */ Bytes_Rxed = recvfrom(The_Socket, Recv_Data, 4096, 0, (struct sockaddr *) &Rx_Sockaddr_Info, &Sockaddr_Size); if (Bytes_Rxed > 0) printf("I received %d Bytes of Data \n", Bytes_Rxed); Counter++; Done = (Counter == 1000); } /* Goodbye. */ shutdown(The_Socket, 2); close (The_Socket); } -------------------------------------------------------------------------- _/ _/_/_/_/_/_/ Bruce Friedman | Amherst Systems, Inc FROM _/_/_/_/ bdf@amherst.com | 30 Wilson Rd. _/_/_/_/_/_/ Voice 716-631-0610 x213 | Buffalo, NY 14221 _/ Fax 716-631-0629 | From uchenick@tis4000.com Wed Jan 28 13:49:08 1998 From: Gordon Uchenick Date: Wed Jan 28 13:49:12 PST 1998 Subject: RE: VxWorks And The Socket Receive Function Here's what's wrong with your sample code. I have no idea why the = additional printf would make things work or not work. It seems to me = that the select() function is much more straightforward. Why don't you = consider using that? PROBLEM: I am trying to create a function that will continually examine and read data from a non-blocking UDP Socket (the non-blocking part is important here). I can create the socket okay, but I cannot get it to read in a non-blocking mode.=20 void main(void) { char Recv_Data[4096]; SOCKET The_Socket; UINT_16 The_Port =3D 5001; UINT_16 The_Type =3D SOCK_DGRAM; UINT_16 The_Peers_IP =3D INADDR_ANY; int Sockaddr_Size =3D sizeof(struct = sockaddr_in); int Bytes_Rxed =3D 0; HERE'S THE PROBLEM! Sock_Opt_ul should be 1. You are turning OFF = non-blocking I/O. unsigned long Sock_Opt_ul =3D 0; struct sockaddr_in P2P_Rx_Sockaddr_Info; struct sockaddr_in Rx_Sockaddr_Info; double Run_Duration_Sec =3D 100.0; int Done =3D 0; int fionbio =3D FIONBIO; int Counter =3D 0; /*------------------------------------*/ /* Create the peer to peer Rx socket. */ bzero((char *)&P2P_Rx_Sockaddr_Info, Sockaddr_Size); P2P_Rx_Sockaddr_Info.sin_family =3D AF_INET; P2P_Rx_Sockaddr_Info.sin_addr.s_addr =3D htonl(The_Peers_IP); P2P_Rx_Sockaddr_Info.sin_port =3D htons(The_Port); The_Socket =3D socket(AF_INET, SOCK_DGRAM, 0); ioctl(The_Socket, fionbio, &Sock_Opt_ul); bind(The_Socket, (struct sockaddr *) &P2P_Rx_Sockaddr_Info, = Sockaddr_Size); printf("\nI'm looking for friends on port %u.\n", The_Port); /* Messaging loop. */ while (!Done) { /* Receive data on this port (non-blocking) and process it. */ YOU SHOULD PUT THIS INTO EACH LOOP ITERATION: SockAddr_Size =3D sizeof(Rx_Sockaddr_Info); Bytes_Rxed =3D recvfrom(The_Socket, Recv_Data, 4096, 0, (struct sockaddr *) &Rx_Sockaddr_Info, &Sockaddr_Size); if (Bytes_Rxed > 0) printf("I received %d Bytes of Data \n", Bytes_Rxed); Counter++; Done =3D (Counter =3D=3D 1000); } From msa@redifon.com Thu Jan 29 07:22:44 1998 From: Mark Armstrong Date: Thu Jan 29 07:22:47 PST 1998 Subject: Re: Driving a Z16C30 serial chip in SDLC mode.. Hello VxWorks type people. I have a problem that one of you out there may be able to help with. Our system is running several tens of RS422 serial receivers on Green Spring MP-Serial IP modules in SDLC mode on Motorola MVME162s. Because of the high data rates the serial controller chips are polled sequentially. Several of these channels have appalling reception rates with almost 100% rejection of received data while others are perfect. The manual for the Zilog Z16C30 chip in question is quite sketchy as to why it sets certain exception bits that are the root of the problem. The data on the bad channels is lost because of "first byte exceptions". If you remove the first byte exception code the data received is full of erroneous information. A summarised description of First Byte Error is "This bit is set when oldest byte in RxFIFO had RxBound, Parity Error, Over Run set when Receiver Data Register was last read." RxBound is the most likely candidate as this is set at the end of an SDLC frame. It indicates when the controller should hunt for a new SDLC sync word. But I know the external hardware I'm syncing with resyncs the SDLC every 4 bytes, this leads me to think I should start hunting every 4 bytes. Does this sound feasible? The following code segment runs for each poll, the FIFO is emptied every time it gets half full (16 bytes). If anyone can see a problem, or tell me how to drive the chip with more finesse I'll be very greatful. Hope someone can help, Mark Armstrong Redifon MEL CODE SEGMENT: ------------ BOOL pollChannel(POLL_OBJECT *this, int channel) { TY_CO_DEV * pTyCoDv = &(this->ipmpsTyCoDv[channel]); RX_OBJECT * rx = &(this->receiver[channel]); uint16_t fifoLevel; volatile short rcsr; uint32_t byte_count=0; char * rx_msg_ptr = rx->currentPos; rcsr = *Z1630_REG (pTyCoDv->chanBase, IPMPS_REG_OFFSET, Z1630_RCSR); /* * if receiver has overrun reset the chip */ if (rcsr & Z1630_RCSR_OVERRUN) { *Z1630_REG (pTyCoDv->chanBase, IPMPS_REG_OFFSET, Z1630_CCAR) = 0xc800; rx->overRunCount++; return (FALSE); } else if (rcsr & (Z1630_RCSR_IDLE | Z1630_RCSR_BREAK_ABORT)) { *Z1630_REG (pTyCoDv->chanBase, IPMPS_REG_OFFSET, Z1630_RCSR) = rcsr; *Z1630_REG (pTyCoDv->chanBase, IPMPS_REG_OFFSET, Z1630_CCAR) = 0xc800; rx->idleErrorCount++; return (FALSE); } /* * Monitor FIFO level, wait until 16 bytes in fifo */ fifoLevel = *pTyCoDv->ricReg; fifoLevel = (fifoLevel & Z1630_RICR_FIFO_CTRL_STATUS_MASK) >> Z1630_RICR_FIFO_CTRL_STATUS_SHIFT; if (fifoLevel >= REACT_FIFO_DEPTH) { for(byte_count=0; byte_countrxChar; /* * read rcsr to flush register on last byte of message */ rcsr = *Z1630_REG (pTyCoDv->chanBase, IPMPS_REG_OFFSET, Z1630_RCSR); if ( !(rcsr & Z1630_RCSR_FIRST_BYTE_ERROR)) { /* The problems here somewhere.*/ rx->firstByteExCount++; /* Reset the serial channel forcing SDLC resync. */ *Z1630_REG (pTyCoDv->chanBase, IPMPS_REG_OFFSET, Z1630_CCAR) = 0xc800; return (FALSE); } } /* * swing to the next buffer */ rx->whichBuffer = (rx->whichBuffer++) % INT_SWING_BUFFERS; rx->currentPos = rx->swingBuffer[rx->whichBuffer]; return (TRUE); } else return (FALSE); } From bdf@amherst.com Thu Jan 29 07:51:41 1998 From: Bruce Friedman Date: Thu Jan 29 07:51:45 PST 1998 Subject: Can't Load vxWorks Bootfile Problem Okay, Here is another one for you - I have a very large program that I am trying to run under vxWorks. At its minimal configuration it loaded and ran fine. However, I have since added a number of "features" (yes features!) that have caused the code to significantly increase in size since the last time that I tried to run it. I have seen this situation described in the manuals, but I am not sure how to correct it (the manuals are either very vague on the matter or I am just not getting it - probably the latter if I know me). When I try to load the vxWorks.st file the following lines of text are displayed: (note that I am fairly sure that I have to modify an address (or multiple addresses) in one (or more) of the makefiles, but I am not sure which ones or what to modify them to.) ---------------- Load file text/data/bss addresses will collide with ours. Rebuild load file so its .bss ends below our start (0xfffffc000057dc00) Error loading file: errno = 0x0 Can't load boot file!! My vxWorks.st compile line looks like this: ------------------------------------------ cc -migrate -o sysLib_st.o -c -O3 -I/h -I. -I/home6/bdf/vxworks_52c/config/all -I/home6/bdf/vxworks_52c/h -I/home6/bdf/vxworks_52c/h/ -I/home6/bdf/vxworks_52c/src/config -I/home6/bdf/vxworks_52c/src/config/ - I/home6/bdf/vxworks_52c/src/drv/ -DCPU=21064 -DRAM_LOW_ADRS=0xfffffc0000200000 -std1 -Wf,-no_ansi_args -D_ANSI_C_SOURCE -DSTANDALONE sysLib.c My vxWorks.st link line looks like this --------------------------------------- ld -N -EL -e sysInit -Ttext 0xfffffc0000200000 -o vxWorks.st tmp.2 symTbl.o true vxWorks.st My bootrom_alpha link line looks like this --------------------------------------- ld -N -EL -e sysInit -Ttext fffffc0000580000 -o tmp sysALib2.o sysLib.o tyCo Drv.o sysAdpLib.o nvram.o version.o bootConfig.o /home6/bdf/vxworks_52c/lib/lib21064osfdrv.a /home6/bdf/vxworks_52c/lib/lib21064osfvx.a /home6/bdf/vxworks_52c/lib/lib21064osfmath.a /home6/bdf/vxworks_52c/lib/lib21064osfots.a /home6/bdf/vxworks_52c/lib/lib21064osfvx.a The end of compilation looks like this -------------------------------------- (reading 22 KB from /home6/bdf/vxworks_52c/bin/osf/secload) examine coff secondary loader will occupy 20000000 to 20007eb0 (reading 773 KB from tmp) vx system (text, data, bss) fffffc0000580000 to fffffc0000651950 compressing... ... 773 KB compressed to 448 KB (42 % reduction) comparing sizes Flash rom size = 3584 KB Code size of (secondary loader + vx system) = 481 KB (Must be < ROM size) Flash rom unused = 3102 KB (flash checksum = 0x11606ba9) (writing 512 KB to bootrom_alpha) TIA Bruce P.S. Thanks to those who responded to my previous post re-non-blocking sockets. I have not implemented the select() function yet, but I now understand why the problem was occuring. -------------------------------------------------------------------------- _/ _/_/_/_/_/_/ Bruce Friedman | Amherst Systems, Inc FROM _/_/_/_/ bdf@amherst.com | 30 Wilson Rd. _/_/_/_/_/_/ Voice 716-631-0610 x213 | Buffalo, NY 14221 _/ Fax 716-631-0629 | From johnny@as.wm.edu Fri Jan 30 05:57:51 1998 From: Johnny Tang Date: Fri Jan 30 05:57:55 PST 1998 Subject: vxworks subscrib VxWorks Tornado From atkin.robert@orbital.com Fri Jan 30 08:00:24 1998 From: "Atkin.Robert" Date: Fri Jan 30 08:00:27 PST 1998 Subject: Re[2]: Driving a Z16C30 serial chip in SDLC mode.. Mark- I don't know if this has anything to do with the general problem you described, but I think it is a problem non-the-less (and perhaps related as a bonus!). In the test where you check to see that the fifo level is greater than or equal to REACT_FIFO_DEPTH, you then only read REACT_FIFO_DEPTH characters, possibly leaving some characters in the fifo ... this could potentially result in an overrun condition, which i think might cause your error. I would either use a while loop and re-check the fifo level each time, or simply read fifoLevel characters instead of REACT_FIFO_DEPTH characters (change the for loop termination to " < fifoLevel" ) I hope this helps. -Rob Atkin Tiger Innovations Real-Time Software Engineering Consultants ratkin@ix.netcom.com vxworks Submitted-by msa@redifon.com Thu Jan 29 07:22:44 1998 Submitted-by: Mark Armstrong [ snip ] CODE SEGMENT: ------------ BOOL pollChannel(POLL_OBJECT *this, int channel) { [snip] >>> >>> The fifo level can be > REACT_FIFO_DEPTH but only REACT... chars >>> are removed from the fifo >>> /* * Monitor FIFO level, wait until 16 bytes in fifo */ fifoLevel = *pTyCoDv->ricReg; fifoLevel = (fifoLevel & Z1630_RICR_FIFO_CTRL_STATUS_MASK) >> Z1630_RICR_FIFO_CTRL_STATUS_SHIFT; if (fifoLevel >= REACT_FIFO_DEPTH) { for(byte_count=0; byte_countrxChar; /* * read rcsr to flush register on last byte of message */ rcsr = *Z1630_REG (pTyCoDv->chanBase, IPMPS_REG_OFFSET, Z1630_RCSR); if ( !(rcsr & Z1630_RCSR_FIRST_BYTE_ERROR)) { /* The problems here somewhere.*/ rx->firstByteExCount++; /* Reset the serial channel forcing SDLC resync. */ *Z1630_REG (pTyCoDv->chanBase, IPMPS_REG_OFFSET, Z1630_CCAR) = 0xc800; return (FALSE); } } /* * swing to the next buffer */ rx->whichBuffer = (rx->whichBuffer++) % INT_SWING_BUFFERS; rx->currentPos = rx->swingBuffer[rx->whichBuffer]; return (TRUE); } else return (FALSE); } From pouivet@ocegr.fr Fri Jan 30 08:30:48 1998 From: Mickael Pouivet Date: Fri Jan 30 08:30:52 PST 1998 Subject: Tornado for Java I'm trying to use the Java Virtual Machine (1.1 beta) for Tornado on a ultra604 Power PC board running VxWorks5.3.1. When getting standard classes.zip from a network location, everything is OK. But when changing JAVA_CLASS_PATH and/or JAVA_HOME so that to use a classes.zip located on our target local IDE disk the folllowing error is displayed : "Invoking Java VM with command line: java myClass Unable to initialize threads: cannot find class java/lang/Thread" Note that the problem only occurs with classes.zip but neither with any of ours plain nor with the "COM.wrs.demo.HelloWorld" file that we can get from ultra604 local IDE disk. Has anybody already got that problem? Thanks for any suggestion. Regards. Mickael Pouivet. -- --------------------------------------------------------------------- Mickael POUIVET OCE Industries S.A. FRANCE Tel : 01 48 98 81 30 E-mail: pouivet@ocegr.fr --------------------------------------------------------------------- From mea@mclean.sparta.com Fri Jan 30 08:58:05 1998 From: Mike Anderson Date: Fri Jan 30 08:59:09 PST 1998 VxWorks Greetings! Can someone out there tell me the difference between a task "switch" hook and a task "swap" hook? It looks like when the switchHook is called, the scheduler has already run and selected the next task to run. What I'd like to do is to get in *before* the scheduler is run to tinker with the task priorities and, in effect, implement a new scheduling policy. TIA, --------------------------------------------------------------------- |Michael E. Anderson | Voice -> 703 448 1683 x235 | |Chief Engineer | FAX -> 703 893 5494 | |mea@mclean.sparta.com | http://www.mclean.sparta.com | |___________________________|_______________________________________| | SPARTA, Inc. | | | /\\ C4I Operation |"Software Development is | | ///\\ 7926 Jones Branch Drive | like making a baby... | | /////\\ Suite 900 | You can't make a baby in | | ///// \\ McLean, VA 22102 | one month by impregnating | | / \\\\\ \\ 703-448-0210 | nine women. | | \ \\\\\ // | | | \ ////// | Some things just take | | \////// SPARTA | time." | | \//// Pride In Performance | | | \// -------------------- | | --------------------------------------------------------------------- From YF.Siu@eng.efi.com Fri Jan 30 12:32:53 1998 From: "Yuet Fung Siu" Date: Fri Jan 30 12:32:57 PST 1998 Subject: task switch/swap hooks As far as I know, both task swap hooks (undocumented ?) and switch hooks will be called right before the dispatcher, and after the task ready queue had been finalized. One major difference between the two is that switch hooks will be called on every task switch, whereas for swap hooks, you can specify when the hook will be called (swap in or out) and which tasks. Swap hooks will be called only after they have been registered via taskSwapHookAdd and taskSwapHookAttach. The floating-point context save/restore is a typical implementation using swap hooks. So I think may be you need to replace the VxWorks schedular: reschedule. Hope this helps and good luck, YF + + Can someone out there tell me the difference between a task "switch" + hook and a task "swap" hook? It looks like when the switchHook is + called, the scheduler has already run and selected the next task to + run. What I'd like to do is to get in *before* the scheduler is run + to tinker with the task priorities and, in effect, implement a new + scheduling policy. + + TIA, + From Bob.Ferrara@digital.com Fri Jan 30 12:49:25 1998 From: Bob Ferrara Date: Fri Jan 30 12:49:28 PST 1998 Subject: VxWorks - system diagnostics Hi all - I am required to produce some level of diagnostics for the various components on my system, such as, Ethernet device, disk drive, floppy driver and graphics card. I believe all selected components are currently supported under VxWorks, thus, drivers exist. Is it possible to create shell scripts to "talk" to these devices and exercise them ? For example, are there shell commands that could put the Ethernet into a loopback mode and then send packets to verify its functionality? Otherwise, would I be required to purchase a VxWorks Developer's License to permit writing/compiling/downloading diagnostics written in C? Thanks for any suggestions, Bob Ferrara Bob.Ferrara@Digital.com From anthony.le@lamrc.com Fri Jan 30 13:16:55 1998 From: "Anthony T. Le" Date: Fri Jan 30 13:16:59 PST 1998 Subject: MV162 Bus Error when accessing VME bus Hello: Does anyone know how to set up for vxWorks for MVME162 board (or configure the MMU) to make it talk to the VME bus A24/A32/D16/D32 address space? Anthony From cruz_nojunk@xyplex.com Fri Jan 30 13:49:09 1998 From: Roger Cruz Date: Fri Jan 30 13:49:12 PST 1998 Subject: Re[2]: WindView questions with Tornado 1.0.1 Anyone knows if this also works with tornado 1.0? >From brett.smith@wg.com Mon Jan 19 05:09:47 1998 From: "Brett Smith" Date: Mon Jan 19 05:09:50 PST 1998 Subject: Re[2]: WindView questions with Tornado 1.0.1 Ethernet is not required for WindView, at least not with an x86 target. I had WindView running over a serial target server for a p-133. The only thing I had to do was build with WDB configured for serial and included the INSTRUMENTATION directive: #define INCLUDE_INSTRUMENTATION in config.h BWS -- Roger Cruz cruz@xyplex.com Xyplex Networks w: 508-952-4783 295 Foster Street f: 508-952-4887 Littleton, MA 01460 From cbriles@lanl.gov Fri Jan 30 14:20:48 1998 From: Carolyn Briles Date: Fri Jan 30 14:20:51 PST 1998 Subject: acromag ip502 module Hello All, Does anyone have experience using the Acromag IP502 RS485 module with vxWorks? I am using the modules on a MVME162LX. The drivers provided are OS-9, but are easy to read and port. I am getting some unexpected Line Status errors and was wondering if anyone had been down this road. Thanks! Carolyn /~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.\ | Carolyn Briles | | | Los Alamos National Lab | (505) 665-1198 (office) | | P-21, MS D454 | | | P.O. Box 1663 | (505) 667-7684 (FAX) | | Los Alamos, NM 87545 | | \~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~./ From greg.brissey@nmr.varian.com Fri Jan 30 15:44:23 1998 From: greg.brissey@nmr.varian.com (Greg Brissey x6951) Date: Fri Jan 30 15:44:27 PST 1998 Subject: accessing VME bus vxworks Anthony This may not be what you are looking for but we use the TRANSPARET Transation Register of the 68040 to allow VME access from $10000000 -> $5fffffff address space we do this instead of programing the sysPhysMemDesc[] struct for the MMU. Doing it this way speeds up up the bootup. * * bits: * 31-24: logical Address base * 23-16: logical address mask * 15: Enable Translation * 14-13: Address mode (00-user,01-Super, 1X-any) * 6-5: cache mode: 00-cache wr/through, 01-cache copyback, * 10-noncache,serialized, 11-noncache * 2: Write - 0 R/W, 1 R_Only * * compile for mc68040, otherwise compiler errors will result */ #define DTT0_VALUE 0x10EFC040 /* from $10000000 -> $5fffffff */ #define DTT1_VALUE 0x20DFC040 /* from $20000000 -> $2fffffff */ asm ("movec %0,dtt0" : : "r" (DTT0_VALUE)); asm ("movec %0,dtt1" : : "r" (DTT1_VALUE)); Actually I obtain this trick from previous discussion on the exploder about VME access. Greg Brissey greg.brissey@nmr.varian.com Varian NMR R&D Palo Alto CA