From owner-vxwexplo-process Thu Apr 1 01:49:58 1999 From: Chatel Guy Date: Thu Apr 1 01:50:01 PST 1999 Subject: Re: fft algorithms for PowerPC the vxWorks Users Group Exploder a écrit: > Does anyone know of a "DSP like" optimized math library (FFT algorithms, > etc.) for the PowerPC family of CPU's? > A colleague of mine has found some interesting things on the MIT web site (http://theory.lcs.mit.edu/~fftw). You may have a look at the John Green's algorithm. Hope this help Bye -- -----------------oooOooo------------------ Guy Châtel Thomson CSF-DETEXIS 10 av. de la 1ere DFL 29283 BREST CEDEX - FRANCE fax: +33 298 312 443 voice: +33 298 312 601 email: guy.chatel@detexis.thomson-csf.com -----------------oooOooo------------------ From owner-vxwexplo-process Thu Apr 1 04:00:19 1999 From: daemon@csg.lbl.gov Date: Thu Apr 1 04:00:23 PST 1999 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Thu Apr 1 04:00:16 PST 1999 Subject: Re: Ampro Pentium CPU board & VxWorks Subject: How to drive ? Subject: Re: StrongARM SA1100 (BRUTUS) ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Ampro Pentium CPU board & VxWorks Date: Wed, 31 Mar 1999 12:24:39 -0500 From: Lawrence Pfeffer Organization: NIST Message-ID: <37025A57.CF86253A@nist.gov> References: Tim Allen wrote: > >2. Anyone which has experience with some other Ampro CPU boards and > >VxWorks? Tim, Sorry I haven't read the group for a bit, or I'd have replied earlier. No exp. w/ Ampro P5x, but we've got P5i boards. We're running running Vx 5.3.1, using WRS' x86 BSP, *without* SENS. Let me know if you need details, but it sounds like you have the P5i config already. We found an Ampro guy who clued us in on setup for the P5i. You might try contacting Bill Wobber (bwobber@ampro.com.) He may be able to answer questions on VxWorks on the P5x Unfortunately, I can't help you with the PCI issues on the P5x. Good luck. I encourage you to post a summary if/when you get a new target working. Side issue: We are booting the P5i from a DOC2000 solidState disk (into VxWorks, although this is probably indep of O/S.) Setting the SSD to be the boot device apparently has a wrinkle that was not obvious to us. Depending on how you set up a P5i you can have either (HardDisk=C: + DOC=D:) or (DOC=C: only.) Several of the steps to making the DOC2000 the boot device apparently work *only* in the latter case. We do not understand why the same commands applied to the DOC (when it's accessable as drive D:) do *not* have the same effect as when it is C: See the P5i quickstart card for details. HTH, - -- Lawrence E. Pfeffer, Ph.D. (Larry), voice: (301)975-6067, FAX: (301)869-6275 *Preferred* email: lawrence.pfeffer@nist.gov U.S. mail: NIST, Bldg. 226/Rm. B150, Gaithersburg, MD 20899 --------------------------- Newsgroups: comp.os.vxworks Subject: How to drive ? Date: Wed, 31 Mar 1999 17:09:50 -0500 From: Lei Sun Organization: University of Central Florida Message-ID: <37029D2E.9CFD2235@pegasus.cc.ucf.edu> Hi: I am trying to develop a data acquisition and control system running under Vx-Work. the data acquistion board is VMIVME4514A analogue I/O. According to the manual, the address of output channel 1 is 0xffff8840, In windsh, I did the following: m 0xFFFF8840 the value in this address is 0xFF80, I changed it to 0x03FF. the output voltage is calculated by V=-10+20*Value(DEC)/4096, So the output voltage should be -10v. I measured the voltage it's true that I got this value, But after I connected the output line with a physical device which I want to control. the output voltage becomes 0,nomatter how much I write in the output register. Can anybody give me any hint about this. I have got lots of help from this group. thank you ! lei sun --------------------------- Newsgroups: comp.os.vxworks Subject: Re: StrongARM SA1100 (BRUTUS) Date: Wed, 31 Mar 1999 22:22:51 -0600 From: "Gopal Miglani" Organization: AT&T WorldNet Services Message-ID: <7dus1k$n6f$1@bgtnsc01.worldnet.att.net> References: <199903301703.SAA03162@dsl-2.tadpole.co.uk> Just FYI... Microware (www.microware.com ) is just now releasing SideARM/Sidekick (StrongARM SA-1100 Multimedia Development board with companion SA-1101 development board support package in their latest StrongARM release. It has support for the SA-1100 and 1101 peripherals. It's rumored that the VGA is extremely tricky (bus bandwidth) and Microware's VGA driver controls the graphics very nicely without bus contention with PCMCIA Ethernet going. It's tons less than 8000 lines of source for the device drivers. The OS-9 drivers are also already compiled and run out of the box. Regards, Gopal Miglani SoftProse, Inc. 1776 22nd Street, Suite 100, West Des Moines, IA 50266 Tel: 515-221-2220, Fax: 515-225-0660, Mobile: 515-988-4157 email: gopalm@att.net, URL: http://www.softproseinc.com Contract Engineering & Products for: Digital TV, Communication Protocols, OS Ports, Device Drivers, Embedded Systems, BSPs =============================== David Laight wrote in message <199903301703.SAA03162@dsl-2.tadpole.co.uk>... > >The SA1100 evaluation board is indeed 'BRUTUS'. > >Intel also have a 'StrongARM SA-1100 Multimedia Development board with Companion >SA-1101 development board' (MESA in some places). > >The brutus BSP runs on the former, the latter is different enough to require a >BSP port. In particular the PCMCIA and display devices are different - so its >hard working out what is going on. > >We now have prototypes of our SA-1100/SA-1101 board working under vxWorks (about >8000 lines of source for device drivers!) > > David > >---------------------------------------------------------------- >David Laight email: dsl@tadpole.co.uk >Tadpole Technology plc phone: +44 1223 278 256 >Cambridge, UK fax: +44 1223 278 201 > --------------------------- End of New-News digest ********************** From owner-vxwexplo-process Thu Apr 1 11:35:22 1999 From: Kevin Sapp Date: Thu Apr 1 11:35:26 PST 1999 Subject: Breakpoints Hello, The documentation implies that "global" breakpoints are the default and thread breakpoints the thread has to be specified. What I have found is that if I attach to a task and set the break point and then continue and the function is called in the "attached" threads context the break is caught. If the break occurs in another context the break is missed ? When a file is loaded via tornado and a break is set, then execution starts if the break is in a different task (a new one created by sp, ...) the break point is not detected. What am I doing incorrectly ? I just want to sent an unconditional break point that breaks NO MATTER WHO CALLS IT FROM WHAT THREAD. Thanks for the help Kevin From owner-vxwexplo-process Thu Apr 1 19:44:21 1999 From: mmenge@dspt.com Date: Thu Apr 1 19:44:24 PST 1999 Subject: Re: Breakpoints Mark Menge@DSPT 04/01/99 10:43 PM If you are using tornado 1.0 the WDB (host source debugger) only does task specific breaks. (I know what the ____ are they thinking?). Next version is supposed to fix this. tornado vxworks From owner-vxwexplo-process Fri Apr 2 04:00:18 1999 From: daemon@csg.lbl.gov Date: Fri Apr 2 04:00:22 PST 1999 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Fri Apr 2 04:00:15 PST 1999 Subject: RTX X server not started Subject: make VxWorks from command prompt Subject: gdb for vxworks on linux ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: RTX X server not started Date: Thu, 01 Apr 1999 22:54:31 +0800 From: LIM KIMPANG Organization: Subscriber of Pacific Internet Message-ID: <370388A7.EF72AAD7@pacific.net.sg> Reply-To: tetotaller@pacific.net.sg Hi, Anyone familiar with RTX?? I am having problems in setting up RTX for vxworks. I am using CDFB server. Below is the failure message: RTX Err: Xserver: Failed to establish listening sockets NAME ENTRY TID SIZE CUR HIGH MARGIN - ------------ ------------ -------- ----- ----- ----------- tRootTask 0x000000a9b8 ffbe78 15956 140 2920 13036 tExcTask 0x00001d1b3c ff7094 7988 188 232 7756 tLogTask 0x00001d1f7c ff478c 4988 188 268 4720 tWdbTask 0x00001cc294 ff3120 8180 244 288 7892 tXserver 0x00000185c4 ff0984 49988 220 1372 48616 rtx 0x000000c2b0 ff0f88 1020 152 828 192 INTERRUPT 1792 0 192 1600 - SVC - 256 0 24 232 - IRQ - NAME ENTRY TID PRI STATUS PC SP ERRNO DELAY - ---------- ------------ -------- --- ----------------- -------- ------- - ----- tRootTask a9b8 ffbe78 0 READY 1d9104 ffb0a4 d0003 0 tExcTask 1d1b3c ff7094 0 PEND 1db0f4 ff6fd8 0 0 tLogTask 1d1f7c ff478c 0 PEND 1db0f4 ff46d0 0 0 tWdbTask 1cc294 ff3120 3 PEND 1b09d8 ff302c 0 0 tXserver 185c4 ff0984 75 SUSPEND 1af5ac ff08a8 0 0 rtx c2b0 ff0f88 100 PEND 1b09d8 ff0ef0 0 0 Error status symbol table not included (errno =0xd003). vxlibinit: X11 libraries could not be initialized --------------------------- Newsgroups: comp.os.vxworks Subject: make VxWorks from command prompt Date: Thu, 01 Apr 1999 15:45:42 GMT From: weatp@syntron.com (Peter Weatherall) Message-ID: <370393b0.229543500@news.phoenix.net> Hello I can build VxWorks within Tornado, fby selecting the menu option under "Project"->"make XXX" -> etc. How do I run same Makefile from the shell prompt ? Thanks Peter --------------------------- Newsgroups: comp.os.vxworks Subject: gdb for vxworks on linux Date: Thu, 01 Apr 1999 19:35:49 -0800 From: Brian Craft Organization: Cisco Systems, Inc. Message-ID: <37043B15.F339EA09@cisco.com> Hi -- I've been trying to build gdb for vxworks on a rh5.2 system. I've tried versions 4.15 & 4.16. It always chokes at run-time, claiming memory corruption, after calling getcwd(). I wonder if it's a GNU libc problem. Anyone get this to work? b.c. --------------------------- End of New-News digest ********************** From owner-vxwexplo-process Fri Apr 2 14:56:29 1999 From: Charlie Grames Date: Fri Apr 2 14:56:32 PST 1999 Subject: Re: make VxWorks from command prompt Peter, I'm assuming you are using the Windows version of Tornado. If you ensure that C:\Tornado\host\x86-win32\bin is in your path, all you need to do is change to the BSP directory and type make to build vxWorks and vxWorks.sym or make clean to clean up. Other make targets can be found by looking in target\h\make\rules.bsp. You do need to make sure that the environment variables WIND_BASE and WIND_HOST_TYPE are defined. WIND_BASE is typically C:\Tornado, and WIND_HOST_TYPE for Windows hosts is x86-win32. Charlie Grames The Boeing Company (314) 233-1956 Charles.R.Grames@boeing.com Newsgroups: comp.os.vxworks Subject: make VxWorks from command prompt Date: Thu, 01 Apr 1999 15:45:42 GMT From: weatp@syntron.com (Peter Weatherall) Message-ID: <370393b0.229543500@news.phoenix.net> Hello I can build VxWorks within Tornado, fby selecting the menu option under "Project"->"make XXX" -> etc. How do I run same Makefile from the shell prompt ? Thanks Peter From owner-vxwexplo-process Sat Apr 3 04:00:21 1999 From: daemon@csg.lbl.gov Date: Sat Apr 3 04:00:25 PST 1999 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sat Apr 3 04:00:18 PST 1999 Subject: Re: How to draw dynamic graphic using Wind Web Server Subject: Re: setting MK48*** time-of-day clock Subject: Re: vxWorks.st_rom - network does not init Subject: Re: gdb for vxworks on linux ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to draw dynamic graphic using Wind Web Server Date: Tue, 30 Mar 1999 09:14:19 -0500 From: bernard@ultranet.com (Bernard Farrell) Organization: Farrell Family Message-ID: References: <36FB5373.F355A580@ms.chttl.com.tw> Sender: bernard@hfa2c.s84f5.baynetworks.com In article <36FB5373.F355A580@ms.chttl.com.tw>, ycliu@ms.chttl.com.tw says... > From: "¼B¤_æÐ" > > I have to design web-based management program. > It's purpose is to monitor the status of a device and > needs to be real-time updated by the web server > according to the status of the device. > Can any body give me some advices? I did this in a job I was working on by building a Java client that would open pages on the web server from time to time. The web pages would only return raw figures that the Java client could break into tokens and then use to draw graphs. The client would request the pages every minute or so, though this is something you could easily change. It's not real-time, but it was enough for our needs. Java code was only a few hundred lines and it did not require any changes to the web server. Bernard --------------------------- Newsgroups: comp.os.vxworks Subject: Re: setting MK48*** time-of-day clock Date: Fri, 02 Apr 1999 15:58:51 -0800 From: Scott Kamara Organization: AP Labs Message-ID: <370559BB.B96352CC@pop.sttl.uswest.net> References: <3704C6E3.3D5B37C9@psi.ch> Reply-To: spk@uswest.net You must set the write bit to 1 first (bit D7 of ctlr reg(1ff8h)) to halt timekeeper updates. Then do your time-of-day load. Then reset (clr to 0) the write bit. Should do the same thing when reading time-of-day. Set D6 of ctlr reg 1ff8h, read time-of-day, reset D6 to 0. The M48T59/M48T59Y datasheet (programming manual) is on the web somewhere in the 'ST' (S??? Thomson) web site. Martin Grossmann wrote: > > vxWorks 5.3.1, MV2604 target > > I use sysNvRead to get the time-of-day set by the onboard clock chip > from the NVRAM. I thought I could put together a time structure and use > sysNvWrite to set a time-of-day. Unfortunately I find that the values I > put in the NVRAM get overwritten within a second or so with the current > time from the clock chip. > > What is the right way to set a time for the chip? > > Thanks alot, > > martin > > martin.grossmann@psi.ch - -- - ----------------------- Scott Kamara AP Data Systems Seattle Engr. Office (206) 525-8592 scott@aplabs.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: vxWorks.st_rom - network does not init Date: Sat, 03 Apr 1999 01:33:38 GMT From: ckrause@duckharbor.com (Chuck Krause) Message-ID: <37056fbf.3947439@newsdesk.tidalwave.net> References: Reply-To: ckrause@NOSPAM.duckharbor.com Did you include INCLUDE_STANDALONE_NET ? On Thu, 01 Apr 1999 21:52:51 -0600, petekz@airmail.net (Pete Kockritz) wrote: >Using Tornado 1.0.1 for PPC860. > >I am building the vxWorks.st_rom for programming into flash on our board. >I built the image and programmed it into flash and it boots, but the >network is not initialized. I can init the network with the following >commands entered in the target shell: > > usrBootLineInit(2) > usrNetworkInit(0x4200) > >This starts the network so that I can ping and so forth, but does not >start the WDB agent. Previously, I was building the vxWorks target and >downloading it from our Sun workstations and everything worked fine. I >do have INCLUDE_NET_INIT and INCLUDE_WDB in my configdb.h. I'm using the >same configdb.h for the rom build as I am for the network download build. > >I read a posting a few weeks ago that was similar. A reponse was to >add 'cpm' to the bootline in the 'other' field. This doesn't help me, since >the bootline is not being read from our NVRAM. I verified this by looking >in memory at BOOT_LINE_ADRS (0x4200 on the PPC) and there was nothing there. >I checked it again after doing the 'usrBootLineInit(2)' and the bootline >was there. The bootline reading works fine in the net boot version. > >This is probably something really simple, but so far it has escaped me. > >Regards, >Pete Zone 6b/7a North Virginia Watch out for the SPAM trap in email address. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: gdb for vxworks on linux Date: Sat, 03 Apr 1999 10:42:41 +0400 From: "Victor V. Vengerov" Organization: TEPKOM Message-ID: <3705B861.81EABDD6@tepkom.ru> References: <37043B15.F339EA09@cisco.com> - --------------87240844B85BE74E98BE6F64 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hi! Brian Craft wrote: > Hi -- I've been trying to build gdb for vxworks on a rh5.2 system. I've > tried > versions 4.15 & 4.16. It always chokes at run-time, claiming memory > corruption, > after calling getcwd(). I wonder if it's a GNU libc problem. > > Anyone get this to work? > It is needed to comment and uncomment some ilnes in Makefile to use system malloc, not malloc from library distributed with gdb. Look comments about it in Makefile.in, or get patch from Linux gdb source rpm. > > b.c. - -- Victor V. Vengerov TEPKOM State Enterprise Bibliotechnaya sq., 2, office 3384, St.-Petersburg 198904, Russia 7-812-4284659 - --------------87240844B85BE74E98BE6F64 Content-Type: text/html; charset=us-ascii Content-Transfer-Encoding: 7bit     Hi!
Brian Craft wrote:
Hi -- I've been trying to build gdb for vxworks on a rh5.2 system. I've
tried
versions 4.15 & 4.16. It always chokes at run-time, claiming memory
corruption,
after calling getcwd(). I wonder if it's a GNU libc problem.

Anyone get this to work?
 

It is needed to comment and uncomment some ilnes in Makefile to use system
malloc, not malloc from library distributed with gdb. Look comments about it
in Makefile.in, or get patch from Linux gdb source rpm.
 
b.c.
-- 
Victor V. Vengerov
TEPKOM State Enterprise
Bibliotechnaya sq., 2, office 3384, St.-Petersburg 198904, Russia
7-812-4284659
  - --------------87240844B85BE74E98BE6F64-- --------------------------- End of New-News digest ********************** From owner-vxwexplo-process Sat Apr 3 08:23:06 1999 From: "Kirnum, Robert" Date: Sat Apr 3 08:23:09 PST 1999 Subject: Symbols - COFF to ASCII Anyone know if the VxWorks / Tornado tools allow for extracting symbols to an ASCII file from COFF? Thanks, Bob From owner-vxwexplo-process Sun Apr 4 04:00:12 1999 From: daemon@csg.lbl.gov Date: Sun Apr 4 04:00:15 PDT 1999 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sun Apr 4 04:00:08 PDT 1999 Subject: Re: Sys Admin. Ref. for Wind River Vxworks Subject: Re: How do you verify data written to disk? Subject: Re: How do you verify data written to disk? Subject: Re: How do you verify data written to disk? ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Sys Admin. Ref. for Wind River Vxworks Date: Sat, 03 Apr 1999 09:28:20 -0600 From: me@nowhere.jp (me) Organization: Self Message-ID: References: <37038AA6.3FD3A2DB@crosslink.net> In article <37038AA6.3FD3A2DB@crosslink.net>, stephen mcdermott wrote: As tart as this sounds, it is about the best advice. Read the books. Really. The Programmers guide is a pretty easy to follow document. It should get you a long way. There really isnt a sys adm book johncoby@orbitworld.net NASA/JSC > Could someone refer me to a where I may find information about System > Administration for Wind River Vxworks. A 'turn-key' system just became > my responsibility. > > tia. - -- See "Closing Costs" by Brad Tyer, The Houston Press for a feature story on Ryland Homes and my webpage at: http://www.houstonpress.com/1998/102998/feature2-1.html --------------------------- Newsgroups: comp.os.vxworks,comp.arch.embedded,comp.realtime,comp.arch.storage Subject: Re: How do you verify data written to disk? Date: Sat, 03 Apr 1999 17:34:57 GMT From: aerosoft@agt.net Organization: Aerospace Software Ltd. Message-ID: <37065ED8.4A5C@agt.net> References: <7e4agr$sl1$1@nnrp1.dejanews.com> <3705E1CA.30AF4F25@trailing-edge.com> Reply-To: aerosoft@agt.net Depending on how paranoid you are, you may have to write the file and save it with a cryptographic checksum - also known as a cyclic redundancy check or cryptographic signature (different Profs, different names). Then you can periodically scan the disk and verify the integrity. I would use the CRC32 algorithm on blocks of data. That should be good enough. Cheers, Herman http://www.agt.net/public/aerosoft Tim Shoppa wrote: > > paulvroberts@my-dejanews.com wrote: > > > > Hi, > > > > In our application, we log data to disk, and the integrity of the data > > is highly important. > > > > I am seeking out thoughts and ideas on the best way of verifying that data > > that has just been written to disk was written correctly. Note that we are > > using an RTOS (VxWorks), Motorola MVME2700, SCSI disks, and are writing the > > software using C/C++. > > > > Presumably when you do a default write to hard disk (eg with C write() or > > fwrite()) the data is not actually verified. If we were to attempt > > verification ourselves in the application code (eg with a C read() function), > > then would this actually work or would we just get data from one of the > > numerous caches (C buffers, smartdrive type cache, cpu caches, disk > > controller cache, etc) along the way? > > > > There is no way I can find for specifying "verification on" under VxWorks - I > > wonder what is done in other OSes (eg UNIX) for mission-critical data > > recording systems? > > It's not just an "OS" issue, it's also a hardware issue. The > SCSI or IDE host adapter may decide to cache written data for > future reads, and the drive can do this as well. With SCSI > drives there are bits that can be toggled in the mode page > to select/deselect caching in the drive itself, though not all > drives will obey these bits. You need to get the full SCSI interface > specs from the drive manufacturer for your specific drive's firmware > to be really sure that you've turned of caching. > > Several widely used operating systems - notably RSX-11M > and VMS - allow files and entire volumes to be opened with > "data checking". The OS makes sure that the device driver and > the device itself have caching disabled in this case. > This can be specified as a characteristic of the file or the > volume; for example, the VMS INIT command can take the /DATA_CHECK > qualifier: > > $HELP INIT/DATA_CHECK > > INITIALIZE > > /DATA_CHECK > > /DATA_CHECK[=(option[,...])] > > Checks all read and write operations on the disk. By default, > no data checks are made. Specify one or both of the following > options: > > READ Checks all read operations. > > WRITE Checks all write operations; default if only the /DATA_ > CHECK qualifier is specified. > > To override the checking you specify at initialization for disks, > enter a MOUNT command to mount the volume. > > -- > Tim Shoppa Email: shoppa@trailing-edge.com > Trailing Edge Technology WWW: http://www.trailing-edge.com/ > 7328 Bradley Blvd Voice: 301-767-5917 > Bethesda, MD, USA 20817 Fax: 301-767-5927 --------------------------- Newsgroups: comp.os.vxworks,comp.arch.embedded,comp.realtime,comp.arch.storage Subject: Re: How do you verify data written to disk? Date: Sat, 03 Apr 1999 13:14:17 -0400 From: Tim Shoppa Organization: Trailing Edge Technology Message-ID: <37061428.76274908@trailing-edge.com> References: <7e4agr$sl1$1@nnrp1.dejanews.com> <3705E1CA.30AF4F25@trailing-edge.com> <37065ED8.4A5C@agt.net> aerosoft@agt.net wrote: > > Depending on how paranoid you are, you may have to write the file and > save it with a cryptographic checksum - also known as a cyclic > redundancy check or cryptographic signature (different Profs, different > names). Then you can periodically scan the disk and verify the > integrity. > > I would use the CRC32 algorithm on blocks of data. That should be good > enough. You might think so, until you run across a disk drive which fails by pretending that it is writing blocks, but is not actually doing so - the old information (with the old CRC) is still there. You might tell me that this will never happen in practice, but I've seen it happen twice in my (short) career! Tim. --------------------------- Newsgroups: comp.os.vxworks,comp.arch.embedded,comp.realtime,comp.arch.storage Subject: Re: How do you verify data written to disk? Date: Sat, 03 Apr 1999 16:19:15 -0800 From: "Christian G. Taranti" Organization: pop.ece.nps.navy.mil Message-ID: <3706B003.C520E532@ece.nps.navy.mil> References: <7e4agr$sl1$1@nnrp1.dejanews.com> For what you are describing, the development cost is more important than the final product cost. The main problem is that you are going to have caches all along: C/C++, operating system, controller and hard drive. You can take control over a few of them, but not over all of them. One way to approach this question is: wht is the error probability that you can afford? Zero is not a good answer, because NO scheme will give you zero probrability of undetected error... In this case, one solution is to use hardware redundancy: two different boards, two diferent hard disks. Write everything twice. Since the probability of failure is very small, the new probability will be much smaller. Some sanity check of the data written will make your probability even smaller. Maybe this solution is more conveniente than digging into manuals and trying to disable the board cache, the hard drive cache, the processor cache and the software buffers. It's a different way to look at the question. paulvroberts@my-dejanews.com wrote: > Hi, > > In our application, we log data to disk, and the integrity of the data > is highly important. > > I am seeking out thoughts and ideas on the best way of verifying that data > that has just been written to disk was written correctly. Note that we are > using an RTOS (VxWorks), Motorola MVME2700, SCSI disks, and are writing the > software using C/C++. > > Presumably when you do a default write to hard disk (eg with C write() or > fwrite()) the data is not actually verified. If we were to attempt > verification ourselves in the application code (eg with a C read() function), > then would this actually work or would we just get data from one of the > numerous caches (C buffers, smartdrive type cache, cpu caches, disk > controller cache, etc) along the way? > > There is no way I can find for specifying "verification on" under VxWorks - I > wonder what is done in other OSes (eg UNIX) for mission-critical data > recording systems? > > Thanks for any help, > > Paul Roberts. > > -----------== Posted via Deja News, The Discussion Network ==---------- > http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own --------------------------- End of New-News digest ********************** From owner-vxwexplo-process Mon Apr 5 05:01:04 1999 From: "Blake, Don" Date: Mon Apr 5 05:01:08 PDT 1999 Subject: RE: comp.os.vxworks newsdigest My guess is that you analog I/O board doesn't have enough drive. How much current does your physical device require? What is the maximum drive current of you analog I/O board? Don R. Blake Sr. Programmer Lockheed Martin Federal Systems Owego, NY -----Original Message----- From: Lei Sun I am trying to develop a data acquisition and control system running under Vx-Work. the data acquistion board is VMIVME4514A analogue I/O. According to the manual, the address of output channel 1 is 0xffff8840, In windsh, I did the following: m 0xFFFF8840 the value in this address is 0xFF80, I changed it to 0x03FF. the output voltage is calculated by V=-10+20*Value(DEC)/4096, So the output voltage should be -10v. I measured the voltage it's true that I got this value, But after I connected the output line with a physical device which I want to control. the output voltage becomes 0,nomatter how much I write in the output register. Can anybody give me any hint about this. I have got lots of help from this group. From owner-vxwexplo-process Mon Apr 5 05:15:10 1999 From: Venkatesh Prakash PS Date: Mon Apr 5 05:15:13 PDT 1999 Subject: ECP Support by VxWorks PPP. Hi all, I want to know 1. Does Vxworks version 5.3.1 supports ECP(Encryption Control Protocol)in Point-to-point Protocol(PPP). 2. What are the encryption methods does it supports? Thanks in advance. Regards, vprakash -- -------------------------------------------------- Venkatesh Prakash PS Lucent Technologies Golf View Campus Wind Tunnel Road Bangalore 560 017 Voice +91 80 527 1771 +91 80 527 1772 Extn 2183 E-mail vps@lucent.com vps@bell-labs.com -------------------------------------------------- From owner-vxwexplo-process Mon Apr 5 14:10:40 1999 From: John Ford Date: Mon Apr 5 14:10:44 PDT 1999 Subject: mv167 and A24 vxworks gurus: I need to set up a multi-processor mv167 chassis so that an A24-only compatible DMA controller can access the on-board memory of one of my CPU cards. Until now, all my CPU cards have been mapped onto the bus in A32 space only, and so I'm proceeding slowly with this, and would like some advice, if anyone has sone a similar thing. It appears that I can edit sysLib.c, enable an alternate translation register, configure the address modifier to respond to A24, and be done. Is it that simple? All my cards have either 8,16, or 32 MB of memory. Thanks! -- John Ford National Radio Astronomy Observatory Green Bank, WV 24944-0002 jford@nrao.edu From owner-vxwexplo-process Mon Apr 5 15:57:48 1999 From: "Hao, Li (MED)" Date: Mon Apr 5 15:57:52 PDT 1999 Subject: question about VxWorks object file size I tried to load an object file to my VxWorks target from a mounted NFS filesystem. When I did "ls -l" on my host machine, it tells me the file size is 4708217. But on the target, the memory difference before and after the load according to memShow is 3266648. If it was bigger, it would have made more sense because loader would try to update symbol table. Does anybody know why? Thanks. Li From owner-vxwexplo-process Mon Apr 5 19:12:59 1999 From: Sid Madhavan Date: Mon Apr 5 19:13:03 PDT 1999 Subject: Socket connection problem Hi Can any one of the vxworks gurus throw me some light on the socket connect between HPUX10.20 and vx works5.2 My old cofiguration is HPUX9.05 talking to vxworks 5.2 subsystem via tcp/ip. In this setup I could connect to all the ports and worked fine. Now we migrated to HPUX10.20 and vxworks 5.2. The problem now I have is that some of the ports that we used to connect are not connecting. The vxworks side is the server and HPUX side is a client. While trying to connect to some well known ports the client side hangs in the connect call. ( This problem does not come in HPUX9.05 compiled executables). Any help will be greatly appriciated. Thanks. -- Sridhar(Sid)Madhavan Email: Sridhar.Madhavan@amermsx.med.ge.com FAX: (414) 548 5197 Voice: (414) 548-5117 3000 N. Grandview Waukesha, WI 53188 From owner-vxwexplo-process Mon Apr 5 23:13:14 1999 From: Venkatesh Prakash PS Date: Mon Apr 5 23:13:17 PDT 1999 Subject: Memory estimate and PPP issue Hi All, 1. Can I know the memory estimates for Vxworks Network Stack as supplied as layer2 package by Vx works and PPPlayer. 2. I have a small clarification, in the Vxworks PPP feature sheet it has been told that upto 16 interfaces can be supported Is this limitation because of i.The Vxworks system/target ii.Vxworks PPP iii.Or is it just a compilation constant ( MAX_USERS) which defaults to 16 and can be changed according to the requirements. Thanks in advance, regards vprakash -- -------------------------------------------------- Venkatesh Prakash PS Lucent Technologies Golf View Campus Wind Tunnel Road Bangalore 560 017 Voice +91 80 527 1771 +91 80 527 1772 Extn 2183 E-mail vps@lucent.com vps@bell-labs.com -------------------------------------------------- From owner-vxwexplo-process Tue Apr 6 04:00:14 1999 From: daemon@csg.lbl.gov Date: Tue Apr 6 04:00:18 PDT 1999 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Tue Apr 6 04:00:11 PDT 1999 Subject: Re: make VxWorks from command prompt Subject: Re: make VxWorks from command prompt Subject: Re: VxWorks + Strong ARM SA110 Subject: C++ Overload Question ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: make VxWorks from command prompt Date: Mon, 05 Apr 1999 09:20:12 -0400 From: Douglas Fraser Organization: Lucent Technologies Message-ID: <3708B88C.1C7FC501@lucent.com> References: There is a file, torVars.bat, in the Tornado host bin directory that will set up your environment based on your installation. (that is, for a Windows host....) So, execute torVars.bat from your DOS shell, then just run make as Charlie mentions below. Doug Charlie Grames wrote: > > Peter, > > I'm assuming you are using the Windows version of Tornado. If you ensure > that > > C:\Tornado\host\x86-win32\bin > > is in your path, all you need to do is change to the BSP directory and type > > make > > to build vxWorks and vxWorks.sym or > > make clean > > to clean up. Other make targets can be found by looking in > target\h\make\rules.bsp. > > You do need to make sure that the environment variables WIND_BASE and > WIND_HOST_TYPE are defined. WIND_BASE is typically C:\Tornado, and > WIND_HOST_TYPE for Windows hosts is x86-win32. > > Charlie Grames > The Boeing Company > (314) 233-1956 > Charles.R.Grames@boeing.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: make VxWorks from command prompt Date: Mon, 05 Apr 1999 15:50:45 GMT From: weatp@syntron.com (Peter Weatherall) Message-ID: <3708db19.140184156@news.phoenix.net> References: <3708B88C.1C7FC501@lucent.com> Thanks. For the NT host, are VxWorks developers in general using the win32 command shell or the cygwin32 (UNIX shell emulation) from Cygnus ? Peter On Mon, 05 Apr 1999 09:20:12 -0400, Douglas Fraser wrote: >There is a file, torVars.bat, in the Tornado host bin directory >that will set up your environment based on your installation. >(that is, for a Windows host....) So, execute torVars.bat from >your DOS shell, then just run make as Charlie mentions below. > >Doug > >Charlie Grames wrote: >> >> Peter, >> >> I'm assuming you are using the Windows version of Tornado. If you ensure >> that >> >> C:\Tornado\host\x86-win32\bin >> >> is in your path, all you need to do is change to the BSP directory and type >> >> make >> >> to build vxWorks and vxWorks.sym or >> >> make clean >> >> to clean up. Other make targets can be found by looking in >> target\h\make\rules.bsp. >> >> You do need to make sure that the environment variables WIND_BASE and >> WIND_HOST_TYPE are defined. WIND_BASE is typically C:\Tornado, and >> WIND_HOST_TYPE for Windows hosts is x86-win32. >> >> Charlie Grames >> The Boeing Company >> (314) 233-1956 >> Charles.R.Grames@boeing.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks + Strong ARM SA110 Date: 5 Apr 1999 16:48:27 GMT From: Hwa-Jin Bae Organization: PSO Systems Inc (http://www.pso.com) Message-ID: <7eapgr$73j$2@news.idiom.com> References: <3708B77C.9154A23B@lucent.com> > Is that cache incoherency problem documented anywhere? > BTW, I just booted an EBSA285 here on the passive backplane > with a DEC21140, and it runs fine. We also have a 21143 running > with no problems. SA110 and 21285 footbridge do not implement cache coherency via snooping the bus. If you run VxWorks with basic MMU support and have cache turned off on device accessible DMA memory you won't see any problems. If you run with full MMU support, or if you run Linux or BSD Unix on the same HW, you will have to modify the drivers to flush the cache lines on bus mastering devices. We just went through modifying bunch of drivers for Linux on custom HW similar to EBSA285. - -- Hwa-Jin Bae PSO Systems, Inc. http://www.pso.com mailto:hjb@pso.com --------------------------- Newsgroups: comp.os.vxworks Subject: C++ Overload Question Date: Mon, 05 Apr 1999 17:45:36 -0400 From: Tim Sohacki Organization: Nortel (Northern Telecom) Message-ID: <37092F00.418C2901@nortelnetworks.com> Under vxWorks 5.3.1 on the PPC (the default 2.7+ gnu toolset), a fellow designer is attempting the following: processMsg (baseClassInstancePtr); processMsg (aSubClassInstancePtr); where aSubClassInstancePtr is a instance of a subclass of the class of baseClassInstancePtr. Whenever processMsg() is called with aSubClassInstancePtr as a parameter, processMsg (baseClassInstancePtr) executes. I was told that this behaves as expected under Microsoft Visual C++; processMsg (aSubClassInstancePtr) executes. Thoughts? Thanks, Tim - -- Tim Sohacki, Nortel Networks email: sohacki@nortelnetworks.com --------------------------- End of New-News digest ********************** From owner-vxwexplo-process Tue Apr 6 07:43:09 1999 From: Bill Brown Date: Tue Apr 6 07:43:12 PDT 1999 Subject: Re: mv167 and A24 - vxWorks the vxWorks Users Group Exploder wrote: > > Submitted-by owner-vxwexplo-process Mon Apr 5 14:10:40 1999 > Submitted-by: John Ford > > vxworks gurus: > > I need to set up a multi-processor mv167 chassis so that an A24-only > compatible DMA controller can access the on-board memory of one of my > CPU cards. Until now, all my CPU cards have been mapped onto the bus in > A32 space only, and so I'm proceeding slowly with this, and would like > some advice, if anyone has sone a similar thing. It's been a few years since I poked around in thaat area, but I believe I may have done something similar. > > It appears that I can edit sysLib.c, enable an alternate translation > register, configure the address modifier to respond to A24, and be > done. Is it that simple? IMHO, _anyting_ to do with the memory-mapping registers can hardly be called simple; otherwise the answer is "yes." There are, I think, cache issues which you need to watch out for. Also, keep in mind that all cpus in a bin _except_ cpu #0 normally have bus-access to on-board memory disabled. If you're going to change this, be sure to watch out for address space collisions between cpu boards. At least one of the Motorola cpu boards is set up such that mapping of "mail-box" memory is controlled by the over-all memory mapping. I don't know if it applies to the '167, but it is something to keep in mind. > > All my cards have either 8,16, or 32 MB of memory. > > Thanks! > > -- > John Ford > National Radio Astronomy Observatory > Green Bank, WV 24944-0002 > jford@nrao.edu Disclaimer: Any opinions are my own and have | -bill nothing to do with the official policy or the | wlbrown@lbl.gov management of L.B.N.L, who probably couldn't | Berkeley, CA care less about employees who play with trains. | aka wlbrown@netcom.com From owner-vxwexplo-process Tue Apr 6 07:50:47 1999 From: francois pertin Date: Tue Apr 6 07:50:50 PDT 1999 Subject: C++ static constructors ????? Does any VXWORKS user already succeed to call module static constructors with the function cplusCtors of the vx librarie cplusLib ????? Any help is realy apreciated.... PS: I'm using tornado 1.0 (vx 5.3) Thanks Francois From owner-vxwexplo-process Wed Apr 7 01:45:04 1999 From: David Laight Date: Wed Apr 7 01:45:08 PDT 1999 Subject: Re: question about VxWorks object file size > Subject: question about VxWorks object file size > > I tried to load an object file to my VxWorks target from a mounted NFS > filesystem. When I did "ls -l" on my host machine, it tells me the file > size is 4708217. But on the target, the memory difference before and > after the load according to memShow is 3266648. If it was bigger, it > would have made more sense because loader would try to update symbol > table. > > Does anybody know why? Thanks. There are all sorts of extra info in the COFF file (ELF has even more). Use 'size' to find the sizes of the sections in the COFF file - but remember it doesn't give the sizes of 'common' data (uninitialised externs in a relocatable C file). David ---------------------------------------------------------------- David Laight email: dsl@tadpole.co.uk Tadpole Technology plc phone: +44 1223 278 256 Cambridge, UK fax: +44 1223 278 201 From owner-vxwexplo-process Wed Apr 7 04:01:26 1999 From: daemon@csg.lbl.gov Date: Wed Apr 7 04:01:30 PDT 1999 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Wed Apr 7 04:01:23 PDT 1999 Subject: Recommend a VME Data Storage Solution? Subject: Re: mv167 and A24 Subject: AZ:Tempe - VxWorks OS/Driver Contract - Agency Subject: Re: Select and Sockets Subject: Re: VxWorks + Strong ARM SA110 Subject: Any luck with GDB? Subject: x86/disk/formatting Subject: NTP Subject: tNetTask: Receive Structure Invalid Subject: Re: Solaris 7 and VxWorks ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Recommend a VME Data Storage Solution? Date: Tue, 06 Apr 1999 15:50:35 +0100 From: Matt Cowley Organization: Simoco Europe Limited Message-ID: <370A1F3B.9D354B85@cambridge.simoco.com> - --------------1067734508BB42162EE2A176 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Can anyone recommend a manufacturer of a single slot 6U/3U VME/VME64 disk controller/data storage solution. Moreover, who would you not recommend and why? I'm looking for an inexpensive (<$1000), low bandwidth (10Mps) and small storage capacity (<4G) solution. It doesn't have to be a Hard Disk! However it must have VxWorks Driver Support as I don't have time to write my own, and be compatible with off-the-shelf processor cards. i.e. I don't want to use/buy into a vendors proprietary HW. Thanks Matt - -- - -------------------------------------------------------------- Matthew Cowley Simoco Europe Ltd. Email Matthew.Cowley@cambridge.simoco.com P.O. Box 24 Web Site: http://www.simoco.com St Andrews Road Tel. +44 1223 877544 CAMBRIDGE CB4 1DP Fax. +44 1223 314812 ENGLAND (This is not an official communication) - --------------1067734508BB42162EE2A176 Content-Type: text/html; charset=us-ascii Content-Transfer-Encoding: 7bit Can anyone recommend a manufacturer of a single slot 6U/3U VME/VME64 disk controller/data storage solution.  Moreover, who would you not recommend and why?

I'm looking for an inexpensive (<$1000), low bandwidth (10Mps) and  small storage capacity (<4G) solution.  It doesn't have to be a Hard Disk! However it must have VxWorks Driver Support as I don't have time to write my own, and be compatible with off-the-shelf processor cards.   i.e. I don't want to use/buy into a vendors proprietary HW.

Thanks

Matt 

-- 
- --------------------------------------------------------------
                        Matthew Cowley

Simoco Europe Ltd.              Email Matthew.Cowley@cambridge.simoco.com
P.O. Box 24                     Web Site: http://www.simoco.com
St Andrews Road                 Tel. +44 1223 877544
CAMBRIDGE CB4 1DP               Fax. +44 1223 314812
ENGLAND           (This is not an official communication)
  - --------------1067734508BB42162EE2A176-- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: mv167 and A24 Date: Tue, 06 Apr 1999 08:47:52 -0700 From: Scott Kamara Organization: AP Labs Message-ID: <370A2CA7.D7DF415C@pop.sttl.uswest.net> References: <199904052110.RAA24546@bara.gb.nrao.edu> Reply-To: spk@uswest.net To change how a mv167's on-board dram may be accessed as a vme slave, you should only have to mess with the function sysProcNumSet() in sysLib.c. If you need to, you can read about the key registers in the mv167 programmer's guide which is online in the motorola web site (www.mcg.mot.com). By default, only cpu#0's memory is mapped to the vme bus: if cpu#0 has 4 MB, then that 4 MB is mapped to A24. if cpu#0 has >= 4MB, then the memory is mapped to A32. You will have to tailor sysProcNumSet() to do what you want for a given cpu#N and given on-board memory size. Remember! A24 space is 16MB wide. So, if you map all of cpu#1's 16MB, for instance, to A24 then all of A24 is now used up. So, I would choose the 8MB cpu card as the card to allow mapping of all on-board ram to A24. This way, half of A24 space gets used (perhaps you have other a24 slaves). Using the 32 MB mv167 as your a24 slave would be a mess. John Ford wrote: > > vxworks gurus: > > I need to set up a multi-processor mv167 chassis so that an A24-only > compatible DMA controller can access the on-board memory of one of my > CPU cards. Until now, all my CPU cards have been mapped onto the bus in > A32 space only, and so I'm proceeding slowly with this, and would like > some advice, if anyone has sone a similar thing. > > It appears that I can edit sysLib.c, enable an alternate translation > register, configure the address modifier to respond to A24, and be > done. Is it that simple? > > All my cards have either 8,16, or 32 MB of memory. > > Thanks! > > -- > John Ford > National Radio Astronomy Observatory > Green Bank, WV 24944-0002 > jford@nrao.edu - -- - ----------------------- Scott Kamara AP Data Systems Seattle Engr. Office (206) 525-8592 scott@aplabs.com --------------------------- Newsgroups: comp.os.vxworks,comp.realtime,misc.jobs.contract Subject: AZ:Tempe - VxWorks OS/Driver Contract - Agency Date: Tue, 06 Apr 1999 15:46:21 GMT From: Tate Consulting Organization: Tate Consulting, Inc. Message-ID: <923413581.745.11@news.remarQ.com> Fortune 100 Company needs contractor to work on VxWorks Drivers and OS tuning. Refer to Job Code: JC-A10078-01. Please reference the JOB CODE listed above in any correspondence to us. Send your resume in ASCII or MSWord format to: Email: hotjobs@tateinc.com Fax: (561) 852-2385 Snail: Tate Consulting, Inc. 10647 Avenida Santa Ana Boca Raton, FL 33498 Please visit our Hi-Tech Job Database at http://www.tateinc.com Tate Consulting is a Computer Consulting and Placement Company in business since 1991. Andrew Tate, President/Owner Tate Consulting, Inc. atate@tateinc.com (561) 852-TATE (8283) We are an EOE. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Select and Sockets Date: Tue, 06 Apr 1999 17:08:01 +0100 From: Phil Ashby Organization: BT Wireplay Message-ID: <370A3161.540A@wireplay.bt.co.uk> References: <7dkktn$s7g3@overload.lbl.gov> <36FFE475.7294FD51@spamthanks.com> Joe User (really? - I thought you were Brian :) wrote: > How would you handle timeouts in the solution(s?) you > proposed? Are you suggesting that each task that owns > a socket still use select? It seemed to me that you were > infering that select should be avoided. > > Is there a way of interrupting read() if a timeout occurs, > or were you suggesting another API other than read() > to block on? We generally close() the errant socket descriptor from a timeout mechanism (in this case a timer task), which has so far proved very successful in causing read() to return an error, whereupon the blocked task can clean up / terminate gracefully. - -- !--------------------------- Phil "Phlash" Ashby ---------------------! ! BT Wireplay - Lead Programmer and general technical trivia supplier ! ! Phone: 01473-644348 WWW: http://www.wireplay.com/ ! ! Snail: B81/G41, BT Labs, Martlesham Heath, Ipswich IP5 3RE, England.! --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks + Strong ARM SA110 Date: Tue, 06 Apr 1999 17:41:47 GMT From: Peter Smith Message-ID: <370A476E.CE75A1F6@salerio.freeuk.co.uk> References: <8J6O2.30634$FZ5.11609@news.rdc1.sfba.home.com> So to summarise, at present Tornado 1 for ARM supports SENS but not at level 1.1. What does this mean practically, e.g. can I still use the MUX to filter incoming packets ? Does the driver use a format different to END as described in the SENS manual ? Robert Armstrong wrote: > T2 for ARM and all else soon, at that point ARM will have the same networking > source as all other architectures. ARM SENS is actually more like 1.05, it's a > little bit more than 1.0. > > Bob > > In article <8J6O2.30634$FZ5.11609@news.rdc1.sfba.home.com>, "Sudhakar Ravi" > wrote: > >I've been using the EBSA285 for a while now, and I ran into a few surprises: > > > >- If you read a misaligned short/long, you won't get what you'd expect. For > >example, a long read to offset 2 won't give you bytes 2,3,4,5. Instead > >you'll get a rotated combination of 2,3,0,1. > >- To catch the above types of problems, you can enable alignment faults in > >the MMU. BUT, some WRS code is doing misaligned transfers, so you'll get > >faults. For example, the TCP checksum routine has this problem. > >- Because EBSA285 only has SENS 1.0 and not 1.1, there are some things to > >watch for. The default number of sockets, mBlks, clBlks is really low, > >you'll want to increase these. > >- EtherOutputHook doesn't work, you won't receive anything beyond the IP > >header. I had an SPR filed on this. > > > >What's meant by "do not use the full featured memory manager". Is there > >something that needs to be turned off in the MMU or cache? > > > >Patrick Drolet wrote in message ... > >>I do. Now... what's the question ? > >> > >>If it's regarding using the BSP as is, you'll need a NIC with a DEC21143 > >>(i.e.: the new DE500, not the old one). Remember that the EBSA needs to be > >>in SYSTEM slot. I use the NIC in slot 4. Do not use the full featured > >>memory manager... there is a cache incoherency problem with bus master > >>devices. > >> > >>PatD. > >> > >>Peter Smith wrote in message ... > >>>I understand that there is a Wind River supported BSP for the Intel > >>EBSA-285 > >>>Evaluation Board (which contains a StrongARM SA110 + companion core > >logic). > >>>Does anyone have any experience using this board/BSP with the Intel > >passive > >>>PCI backplane, particulary with respect to supported PCI Ethernet cards > >and > >>>END drivers. > >> > >> > >> > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Any luck with GDB? Date: Tue, 06 Apr 1999 18:20:55 GMT From: Aaron Williams Organization: FlowWise Networks, Inc. Message-ID: <370A5081.388C0196@delete.this.FlowWise.com> Hi, I'm trying to get GDB for Tornado 1.0.1 to work for the Mips environment, but every time I try and start it I get a Segmentation fault. I am running Solaris 7 although this also seems to occur on other versions. The GDB binary, gdbmips, was supplied with VxWorks. Has anyone else experienced this? If so, has anyone found a solution? - -Aaron Williams - -- FlowWise Networks, Inc. (408) 546-7216 Note that my return email address has been munged to block spam. --------------------------- Newsgroups: comp.os.vxworks Subject: x86/disk/formatting Date: Tue, 06 Apr 1999 15:35:07 -0400 From: Steve Sherin Organization: Syncro Technology Corp. Message-ID: <370A61EB.F5BF20D9@syncro-tech.com> Hello, Has anyone successfully made a bootable hard disk (ata) for the pc486 BSP using long file names? The boot sector seems to get clobbered when I do a dosFsMkfs() to re-format. The disk is readable and correctly configured otherwise (after a couple days digging around). Technical support is looking into it.. Thanks, Steve Sherin ssherin@syncro-tech.com --------------------------- Newsgroups: comp.os.vxworks Subject: NTP Date: 6 Apr 1999 20:25:30 GMT From: chris@raf.atd.ucar.edu (Chris Webster) Organization: NCAR Research Aviation Facility, Boulder, CO Message-ID: <7edqjq$b84$1@ncar.ucar.edu> Reply-To: chris@raf.atd.ucar.edu Is there a NTP deamon/task for vxWorks out there? - --Chris --------------------------- Newsgroups: comp.os.vxworks Subject: tNetTask: Receive Structure Invalid Date: Sat, 3 Apr 1999 14:44:04 -0500 From: "Mike Reed" Organization: MindSpring Enterprises Message-ID: <7e5reo$sdd$1@camel21.mindspring.com> Reply-To: "Mike Reed" Hello, Has anyone ever heard of the error 'tNetTask: Receive Structure Invalid'? I am using a DY4 PowerPPC 603e BSP with a DY4 783 Graphics board. Anybody heard of this and how to correct it? Any help provided is greatly appreciated. Mike Reed reedm@battelle.org --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Solaris 7 and VxWorks Date: Wed, 07 Apr 1999 05:35:33 GMT From: Aaron Williams Organization: FlowWise Networks, Inc. Message-ID: <370AEEA0.6F8616F0@delete.this.FlowWise.com> References: Martin Shepherd wrote: > > I am thinking about upgrading our Sun Ultra 1-170 Tornado/VxWorks host > to Solaris 7. Unfortunately I haven't been able to find any mention of > Solaris 7 on wrs web site, so I don't know whether it is officially > supported. > > Is anybody aware of any problems with Tornado/VxWorks in this > environment? > > Martin Shepherd (mcs@astro.caltech.edu) It's working fine for me. I'm running Solaris 7 on a Sun Ultra 5. The only problem I have is that gdbmips coredumps every time I try and use it. It has the same problem on everyone else's system as well, and they're running earlier versions of Solaris. - -Aaron --------------------------- End of New-News digest ********************** From owner-vxwexplo-process Wed Apr 7 04:43:39 1999 From: "Kirnum, Robert" Date: Wed Apr 7 04:43:43 PDT 1999 Subject: RE: question about VxWorks object file size You can also use objdump (objdump, objdump960, objdumparm, etc,). It has many options for displaying object file headers. It even disassembles an object file for an assembly listing. Bob << > Subject: question about VxWorks object file size > > I tried to load an object file to my VxWorks target from a mounted NFS > filesystem. When I did "ls -l" on my host machine, it tells me the file > size is 4708217. But on the target, the memory difference before and > after the load according to memShow is 3266648. If it was bigger, it > would have made more sense because loader would try to update symbol > table. > > Does anybody know why? Thanks. There are all sorts of extra info in the COFF file (ELF has even more). Use 'size' to find the sizes of the sections in the COFF file - but remember it doesn't give the sizes of 'common' data (uninitialised externs in a relocatable C file). David ---------------------------------------------------------------- David Laight email: dsl@tadpole.co.uk Tadpole Technology plc phone: +44 1223 278 256 Cambridge, UK fax: +44 1223 278 201 >> From owner-vxwexplo-process Wed Apr 7 08:20:15 1999 From: David Laight Date: Wed Apr 7 08:20:19 PDT 1999 Subject: Re: Symbols - COFF to ASCII > Subject: Symbols - COFF to ASCII > > Anyone know if the VxWorks / Tornado tools allow for extracting symbols to > an ASCII file from COFF? > > Thanks, > Bob > nm ---------------------------------------------------------------- David Laight email: dsl@tadpole.co.uk Tadpole Technology plc phone: +44 1223 278 256 Cambridge, UK fax: +44 1223 278 201 From owner-vxwexplo-process Wed Apr 7 08:21:03 1999 From: Cagatay ALTAN Date: Wed Apr 7 08:21:06 PDT 1999 Subject: VxWorks interrupts. Hello all, I am struggling with a bug and hence using WindView to locate it. My question is not about the bug, but about something I noticed while using WindView: I'm developing a sortware on a mvme162. Mvme162 is communicating with a custom card on VME via INT 3; that is when data is ready the custom card issues an INT 3, with vector 0x80. When I was using WindView I noticed that there are other INT3's that are not issued by the custom card, but that seem to be issued by the vxWorks itself. There happens either a wdCancel or (semGive + wdCancel) during or following that interrupt and tNetTask seem to activate just after. That gives the clue that INT3 is used by network mutual exclusion mechanism. It is also documented that INT4 is used by tty (shell), but I could not reached a documentation that lists which interrupts are used by which calls. I would be grateful if someone could supply me with a pointer. Thanks for your patience. Tevfik Cagatay ALTAN, BScEE mail: altan@mst.aselsan.com.tr ***************************************************************** * "No questions I'm not alone, somehow I'll find my way home." * ***************************************************************** Microwave and Systems Technologies / Tel: 90-312-3851900 / 2926(ext.) Software Engineering Dept. Fax: 90-312-3545205 ASELSAN web: www.aselsan.com.tr From owner-vxwexplo-process Wed Apr 7 11:48:22 1999 From: Roger Cruz Date: Wed Apr 7 11:48:26 PDT 1999 Subject: Missing interrupts on x86 board.. Dear VxWorks users We've been having some random problems with one of targets where we seem to lose interrupts after some time of operation. The time appears to be random and I can't correlate it to any one event. We've instrumented all the ISRs and we only receive the TIMER interrupt after we lose all the other interrupts (serial port and PCI). Our chipset is the intel 430HX which has 2 i8259 PIC controllers cascaded together via IRQ2. IRQ0 (the highest priority) is the timer interrupt which continues to run. IRQ3 & 4 are the serial ports and are the lowest priority than IRQ2 (the slave PIC) to which IRQ8-IRQ15 are connected. These are never seen again after the "condition" occurs. An interrupt in our WRS provided BSP is dismissed via the sysIntEOI code below: 0x10acc0 : pushl %ebp 0x10acc1 : movl %esp,%ebp 0x10acc3 : subl $0x4,%esp 0x10acc6 : call 0x5cbb90 0x10accb : movl %eax,%eax 0x10accd : movl %eax,0xfffffffc(%ebp) 0x10acd0 : pushl $0x20 0x10acd2 : pushl $0x20 0x10acd4 : call 0x1081a8 0x10acd9 : addl $0x8,%esp 0x10acdc : pushl $0x20 0x10acde : pushl $0xa0 0x10ace3 : call 0x1081a8 0x10ace8 : addl $0x8,%esp 0x10aceb : movl 0xfffffffc(%ebp),%eax 0x10acee : pushl %eax 0x10acef : call 0x5cbba0 0x10acf4 : addl $0x4,%esp 0x10acf7 : movl %ebp,%esp 0x10acf9 : popl %ebp 0x10acfa : ret The 2 sysOutBytes actually perform the EOI command, first to the master and then to the slave. This code is executed for ALL interrupts, no matter if they are came from the master or the slave. We've looked at other code from the net and we find that interrupts from the master are issue only one EOI, rather than 2 as the code above shows. Another difference is the order in which the EOIs take place when the interrupt is from the slave. The slave is always done first and the master second. Can anyone tell me if these differences really matter and if so, can they be causing the hangs? If you can provide me with a scenario that proves your theory, it would be extremely appreciated it. Thank you very much. Roger PS: Remove "_nojunk" from email address. vxworks, VxWorks. -- Roger Cruz cruz_nojunk@xyplex.com Xyplex Networks w: 978-952-4783 295 Foster Street f: 978-952-4887 Littleton, MA 01460 From owner-vxwexplo-process Wed Apr 7 11:56:32 1999 From: "Blake, Don" Date: Wed Apr 7 11:56:35 PDT 1999 Subject: RE: Symbols - COFF to ASCII Take a look at objdump - for the PPC it's called objdumpppc. There's also makeSymTbl but I don't know if it does COFF. Don Blake Sr. Programmer Lockheed Martin Federal Systems Owego, NY -----Original Message----- From: vxwexplo@lbl.gov [SMTP:vxwexplo@lbl.gov] Sent: Wednesday, April 07, 1999 11:20 AM To: vxworks_users@csg.lbl.gov Subject: Re: Symbols - COFF to ASCII Submitted-by owner-vxwexplo-process Wed Apr 7 08:20:15 1999 Submitted-by: David Laight > > Subject: Symbols - COFF to ASCII > > Anyone know if the VxWorks / Tornado tools allow for extracting symbols to > an ASCII file from COFF? > > Thanks, > Bob > nm ---------------------------------------------------------------- David Laight email: dsl@tadpole.co.uk Tadpole Technology plc phone: +44 1223 278 256 Cambridge, UK fax: +44 1223 278 201 ********** This is a user group mailing list for vxWorks related topics Mail articles to vxwexplo@lbl.gov for 'explosion' Include the word VxWorks or Tornado to penetrate SPAM filters edit off the trailer to avoid bounce filters Send subscription requests & comments to vxwexplo-request@lbl.gov From owner-vxwexplo-process Wed Apr 7 12:45:27 1999 From: Fred Roeber Date: Wed Apr 7 12:45:31 PDT 1999 Subject: Re: VxWorks interrupts. Cagatay ALTAN asked: > I am struggling with a bug and hence using WindView to locate it. My > question is not about the bug, but about something I noticed while using > WindView: > > I'm developing a sortware on a mvme162. Mvme162 is communicating with a > custom card on VME via INT 3; that is when data is ready the custom card > issues an INT 3, with vector 0x80. > > When I was using WindView I noticed that there are other INT3's that are > not issued by the custom card, but that seem to be issued by the vxWorks > itself. There happens either a wdCancel or (semGive + wdCancel) during or > following that interrupt and tNetTask seem to activate just after. That > gives the clue that INT3 is used by network mutual exclusion > mechanism. It is also documented that INT4 is used by tty (shell), but I > could not reached a documentation that lists which interrupts are used by > which calls. I would be grateful if someone could supply me with a > pointer. From the config/mv162/mv162.h file, the interrupts used for this target are: #define LANC_IRQ_LEVEL 3 /* LANC IRQ level */ #define SCC_IRQ_LEVEL 4 /* serial comm IRQ level */ #define SYS_CLK_LEVEL 6 /* interrupt level for sysClk */ #define AUX_CLK_LEVEL 5 /* interrupt level for auxClk */ #define SCSI_IRQ_LEVEL 2 /* SCSI interrupt level */ #define ABORT_IRQ_LEVEL 7 /* ABORT interrupt level */ Thus, the Lance ethernet chip is generating the level 3 interrupts. Hope that helps. Fred -- | Fred J Roeber, BBN Systems & Technologies | | 4 John Clarke Road Middletown, RI 02842-5202 | | froeber@bbn.com 401-848-3548 | | TraceMaker product manager -> www.tracemaker.bbn.com | From owner-vxwexplo-process Thu Apr 8 04:00:47 1999 From: daemon@csg.lbl.gov Date: Thu Apr 8 04:00:51 PDT 1999 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Thu Apr 8 04:00:43 PDT 1999 Subject: Re: Reusing socket numbers Subject: Re: VxWorks + Strong ARM SA110 Subject: Re: NTP Subject: Re: Force CPU-30 Subject: Re: A networking/vxworks/NT longshot ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Reusing socket numbers Date: Wed, 07 Apr 1999 13:58:30 +0100 From: Phil Ashby Organization: BT Wireplay Message-ID: <370B5676.6705@wireplay.bt.co.uk> References: <370AC739.75D30545@mediaone.net> Ephraim & Margo Ezekiel wrote: > We are trying to decide if we have a timing problem between > the time a connection is closed and a new one is created. If > an application deletes a socket but there is data for that > socket in its queue, then the application has to know that > the data or information is stale. Err. I'm confused by this statement. Are you talking about TCP sockets or UDP sockets? If a TCP socket is close()ed then any data that was in the receive buffers will be discarded before control is returned to the calling code. The TCP port number will remain in a CLOSE_WAIT state for a significant amount of time (I think it's 120 seconds), to prevent delayed packets destined for this port from confusing any new/existing TCP sessions. This behaviour may be overidden by enabling the SO_REUSE_ADDRESS flag, in which case the port number can be explicitly rebound (using bind()) immediately after the original socket has been closed (and any unread data flushed). In this case the new socket may be confused by delayed packets arriving (which will probably have an invalid sequence number). Caveat implementor! UDP sockets are connectionless, and no guarantees can be made about packet timing/sequencing. Therefore UDP ports do not have a CLOSE_WAIT state and can be rebound immediately in all cases. UDP sockets do have a receive buffer however, which is flushed when a socket is close()ed. > ... But there may be a timing > problem, before the application gets scheduled to read from > its queue, it opens a new socket. If it happens to get the > same socket number as the one just deleted, there is no way > the application can know if the data in its queue is for the > old socket or the new one. As stated previously: all existing data in the receive queue will be flushed when the original socket is closed, so if 'old' data is appearing on a 'new' socket, it must be due to network delays and the SO_REUSE_ADDRESS flag being set, or you are using UDP and some data has been delayed in the network. HTH, - -- !--------------------------- Phil "Phlash" Ashby ---------------------! ! BT Wireplay - Lead Programmer and general technical trivia supplier ! ! Phone: 01473-644348 WWW: http://www.wireplay.com/ ! ! Snail: B81/G41, BT Labs, Martlesham Heath, Ipswich IP5 3RE, England.! --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks + Strong ARM SA110 Date: Wed, 07 Apr 1999 16:44:38 +0100 From: j.news.9904@csd.co.uk Message-ID: <370B7D66.15C367EE@csd.co.uk> References: <3708B77C.9154A23B@lucent.com> <7eapgr$73j$2@news.idiom.com> Patrick Drolet wrote: > Do not use the full featured memory manager... there is a cache > incoherency problem with bus master devices. Could you explain further, please? What is the problem and to which version of VxWorks are you referring? Hwa-Jin Bae wrote: > SA110 and 21285 footbridge do not implement cache coherency via snooping > the bus. If you run VxWorks with basic MMU support and have cache > turned off on device accessible DMA memory you won't see any problems. > If you run with full MMU support, or if you run Linux or BSD Unix > on the same HW, you will have to modify the drivers to flush the > cache lines on bus mastering devices. We just went through modifying > bunch of drivers for Linux on custom HW similar to EBSA285. Whether you use basic MMU support or full MMU support (VxVMI), the Ethernet driver can allocate any device-accessible memory using cacheDmaMalloc(), so that the memory is marked as not cacheable by the SA-110. In that case, flushing the SA-110 cache is not required. This is the approach taken by the WRS dec21x4xEnd driver, so this should not need modification for use with VxVMI. Of course, if the drivers you are using do not use cacheDmaMalloc (or otherwise assume the existence of bus-snooping) then problems will ensue. - -- Jem Davies --------------------------- Newsgroups: comp.os.vxworks Subject: Re: NTP Date: 7 Apr 1999 16:18:32 GMT From: chris@raf.atd.ucar.edu (Chris Webster) Organization: NCAR Research Aviation Facility, Boulder, CO Message-ID: <7eg0go$3fb$1@ncar.ucar.edu> References: <370B5DFE.A208CF4D@lmco-NOSPAM.com> Reply-To: chris@raf.atd.ucar.edu > In fact, since at least v3.4, VxWorks specific mods have been in the > source code. What platform are you targeting? I can only speak for the > v3.4 on a Moto MVME 2604 board. I can send you my stuff if applicable. We are using the Motorola MV162. - --Chris --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Force CPU-30 Date: 8 Apr 1999 01:24:20 GMT From: Hwa-Jin Bae Organization: PSO Systems Inc (http://www.pso.com) Message-ID: <7eh0g4$31u$1@news.idiom.com> References: <7eba4a$923$1@idiom.com> <7ee84f$cc1$1@news.idiom.com> <7eg4ep$38c$1@idiom.com> <7egj56$7f1$1@news.idiom.com> <7egqn6$95i$1@idiom.com> Oh yeah, Force CPU-30... how retro of you :-) At least some of the CPU-30's had problems with Lance. Remember I had to actually take one of the CPU-30's down to Campbell and show the Force engineers that I can make the ethernet generate all kinds of frame errors at will. They finally tracked it down in Germany (where these Force boards were originally designed) that it was due to missing caps between the MAC and PHY. So I hope you're using the one that has that problem fixed. Another thing I remember about this board is that the ethernet address was not hardwired. You actually had to software program that address on the fly. (Force used to do this a lot on their boards -- of course this is technically illegal since you can program anything you want into ethernet MAC address). Depending on the rev of HW and SW you use, the source of this info comes from various places (NVRAM or some magic cookie elsewhere). I would go about getting a LAN analyser (tcpdump on Linux works fine) and also a good debugger (shell debugger should work OK) and start tracing it down. Good luck! - -- Hwa-Jin Bae PSO Systems, Inc. http://www.pso.com mailto:hjb@pso.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: A networking/vxworks/NT longshot Date: Sun, 4 Apr 1999 11:57:18 -0700 From: "Dave Craig" Organization: Verio Northern California's Usenet News Service Message-ID: <7e8cn8$n3h$1@news.ncal.verio.com> References: <7e3hud$e5v$1@news.ncal.verio.com> <7e5iti$ik5$1@q.seanet.com> No but a good try! The NT workstation installation can be fresh onto a newly formatted disk and the problem is still there. The SBC behaves in the same way with any machine running NT workstation (and all of the SBCs we have behave in this way)... Dave. seanet wrote in message <7e5iti$ik5$1@q.seanet.com>... >Long shot response - you aren't by any chance running some sort of VPN >software on the NT side, are you? This is similar to the behavior I saw >when I tried to run an FTP server on a machine running Aventail's VPN >client - the FTP open command worked, the GET command looked like it went >through, but no data was actually transferred. I solved it by just removing >Aventail from that machine. > >James Moore >james@banshee.com > >Dave Craig wrote in message >news:7e3hud$e5v$1@news.ncal.verio.com... >> Here's the problem, we have a Pentium based SBC with 100BaseT full duplex >> ethernet that will ftp on bootup from a host that is running the Wftpd32 >FTP >> server on Windows NT Server or Windows95. However, if the host is running >> Windows NT workstation the FTP fails to get started - a connection is >> usually made but that's about all. The host PC has been varied, as have >the >> switch and cables but to no avail. Has anyone else seen anything like this >> at all? At 10BaseT there is no problem... >> >> Dave Craig >> Palo Alto, CA. >> >> > > --------------------------- End of New-News digest ********************** From owner-vxwexplo-process Thu Apr 8 05:10:10 1999 From: "Tony Mayes" Date: Thu Apr 8 05:10:14 PDT 1999 Subject: Flash File Problems Hi, We are using the FlashLib, flash file system for VxWorks from RST with Intel 28F016 / 32 flash chips. I have noticed that on occassion accesses to the file system are taking 100's of milli secs. A delete or write I can understand to a degree, since an erase of a block could take 300 - 400 mSecs. However on a sequence of Stat, open, close I cannot understand why this should be. Any ideas gratefully considered. Tony Tony Mayes (The opinions expressed within are my own) Email address tony.mayes@epid.eurotherm.co.uk Phone (0044)(0)1903 205277 Fax (0044)(0)1903 524016 Eurotherm Process Instrumentation Div, Southdownview Way, Worthing, Sussex, UK, BN14 8NN From owner-vxwexplo-process Thu Apr 8 05:40:30 1999 From: Venkatesh V Date: Thu Apr 8 05:40:33 PDT 1999 Subject: Doubt in WDOG timers!!! Hi VxWorks GURUS, I have a basic doubt regarding watch dog timers implemetation. I require a watchdog timer which will expire after say 125 ms. so i call wdStart as wdStart (id, time, routine, NULL); In the function "routine", do i implement the functionality i require to be executed on watchdog expiry OR do i spawn a task in "routine" which will implement the functionality??? ie. for clarity sake, void routine () { taskSpawn (..., functionality, ..); } OR can i pass the address of functionality () directly to wdStart ???? The reason behind my question is Since the routines posted to watchdog timers execute in ISR context, it should not be allowed to do processing which consume lot of time ..... Hope i am clear... Eagerly awaiting a reply ... Rgds, Venkat... From owner-vxwexplo-process Thu Apr 8 07:33:53 1999 From: "Richard B Asen" Date: Thu Apr 8 07:33:57 PDT 1999 Subject: 82592 Device Driver for VxWorks Does anyone out there have a device drive for the 82592 Ethernet controller? Richard B. Asen Raytheon Systems Company richard_b_asen@raytheon.com 410-583-4477 From owner-vxwexplo-process Thu Apr 8 07:34:41 1999 From: Bill Atkins Date: Thu Apr 8 07:34:45 PDT 1999 Subject: Re: Doubt in WDOG timers!!! > Hi VxWorks GURUS, > I have a basic doubt regarding watch dog timers implemetation. > > I require a watchdog timer which will expire after say 125 ms. so i call wdStart as > > wdStart (id, time, routine, NULL); > > In the function "routine", do i implement the functionality i require to be executed on watchdog expiry OR do i spawn a task in "routine" which will implement the functionality??? > > ie. for clarity sake, > void routine () > { > taskSpawn (..., functionality, ..); > } > > OR > > can i pass the address of functionality () directly to wdStart ???? > The reason behind my question is > Since the routines posted to watchdog timers execute in ISR context, it should not be allowed to do processing which consume lot of time ..... > You are correct. If the functionality needed is simple, you can perform it directly in 'routine', provided, as you say, that the instructions are suitable for ISR context. Else you may want to spawn a task to handle the processing. Sometimes you specifically want to execute things in ISR context. For example, a timer might be started at the same time an I/O operation is started. If the timer expires, it means the I/O did not complete, so you want to use the ISR context to cancel the failed I/O. This is commonly used to prevent hardware problems from hanging up the software. The last argument passed to wdStart is typically used to pass a context pointer so the timer routine is reentrant. If you're using c++, it might be the address of an object which the timer is associated with. Then if you have multiple instances of that object, the same timer routine works for all of them. Check all the restrictions of ISR context - the one I have trouble remembering is NO FLOATING POINT ARITHMETIC. From owner-vxwexplo-process Thu Apr 8 11:35:59 1999 From: David Laight Date: Thu Apr 8 11:36:03 PDT 1999 Subject: Re: Doubt in WDOG timers!!! > Subject: Doubt in WDOG timers!!! > > Submitted-by: Venkatesh V > > Hi VxWorks GURUS, > I have a basic doubt regarding watch dog timers implemetation. > > I require a watchdog timer which will expire after say 125 ms. so i call > wdStart as > > wdStart (id, time, routine, NULL); > > In the function "routine", do i implement the functionality i require to be > executed on watchdog expiry OR do i spawn a task in "routine" which will > implement the functionality??? It really depends on whether the actions you want to perform are valid in an interrupt context, and whether your system can stand interrupts being disabled for the time it takes to do the actions. > > ie. for clarity sake, > void routine () > { > taskSpawn (..., functionality, ..); > } I very much doubt that taskSpawn is valid from an ISR, even if it were the execution time is probably excessive. > > OR > > can i pass the address of functionality () directly to wdStart ???? > The reason behind my question is > Since the routines posted to watchdog timers execute in ISR context, it should > not be allowed to do processing which consume lot of time ..... > > Hope i am clear... > Eagerly awaiting a reply ... > > Rgds, > Venkat... If you need a task to execute your timeout code (and other stuff) netJobAdd will cause tNetwork to execute the work. Just handshake it enough to stop multiple requests blowing tNetwork's job queue. (You do have to be careful you don't have a deadly embrace - eg if your actions require network activity.) David ---------------------------------------------------------------- David Laight email: dsl@tadpole.co.uk Tadpole Technology plc phone: +44 1223 278 256 Cambridge, UK fax: +44 1223 278 201 From owner-vxwexplo-process Thu Apr 8 12:00:59 1999 From: johill@lanl.gov (Jeff Hill) Date: Thu Apr 8 12:01:02 PDT 1999 Subject: RE: Missing interrupts on x86 board.. On Wednesday, April 07, 1999 12:48 PM, the vxWorks Users Group Exploder [SMTP:vxwexplo@lbl.gov] wrote: > Submitted-by owner-vxwexplo-process Wed Apr 7 11:48:22 1999 > Submitted-by: Roger Cruz > > We've been having some random problems with one of targets where we > seem to lose interrupts after some time of operation. The time appears > to be random and I can't correlate it to any one event. We also see this problem here, and ended up adding a watchdog to the BSP supplied COM port serial driver as a workaround for a hang in continuous serial communication with remote Allen Bradley PLC devices that occurred from one to three times a day. This problem was reported to WRS tech support somewhat more than a year ago. Jeff From owner-vxwexplo-process Thu Apr 8 18:14:53 1999 From: helbakou@nortelnetworks.com (Hesham Elbakoury) Date: Thu Apr 8 18:14:56 PDT 1999 Subject: Nucleus Event Flags in VxWorks Did someone implement the Nucleus event flags/groups in VxWorks. Thanks From owner-vxwexplo-process Thu Apr 8 18:59:20 1999 From: "Robert L. Perry" Date: Thu Apr 8 18:59:24 PDT 1999 Subject: Re: SCSI2 Lib Information vxWorks / Tornado Users, I am writing a scsi controller driver and would like to use the scsi 2 library. To understand the program flow I am currently looking at source code for an ncr device along w/ the messages generated when the scsiDebug and scsiIntsDebug flags are enabled. This is a painful process. If anyone has any words of wisdom, a state diagram of scsi thread execution, or information on the scsi manager event notification, please share. Thanks, Bob Perry From owner-vxwexplo-process Wed Mar 31 17:15:15 1999 From: Debi Dawson-Lujan Date: Fri Apr 9 01:54:18 PDT 1999 Subject: New Products from Intoto Inc. Hello, Previously you had contacted Technology Rendezvous Inc. for embedded system software solutions. Therefore, we would like to take this opportunity to inform you of our new spin-off called INTOTO INC. and the products we offer. Intoto Inc. provides embedded software solutions for Networking and Interconnectivity market. - IEEE-1394, IKE, IPSEC, NAT, USB, PCMCIA and Flash Software solutions. Below is the information regarding the new company Intoto Inc. and the solutions we offer. Should you like to receive some additional information please do not hesitate to contact me or visit our web site. http://www.intototinc.com Debi Dawson-Lujan Marketing & Sales Coordinator 3160 De La Cruz Blvd. Suite # 100 Santa Clara CA - 95054 USA T: 408 844 0480 x 303 F: 408 844 0488 ==================================================== Corporate Overview ---------------------------- Intoto Inc. develops, markets and manufactures reliable, off-the-shelf, and easy-to-use customized embedded system solutions that accelerate the customer's product development in the Networking and Interconnectivity market. Interconectivity Solutions ------------------------------------ - eFireStack - IEEE-1394 software and Protocols for Embedded Systems - LINCOLN - IEEE-1394 Link Layer and other related Cores - ePCMCIA - PCMCIA software with Card and Socket Services - eFMM - Flash software supporting on-board and PCMCIA Flash Devices - eUSBStack - Universal Serial Bus (USB) software with Class Drivers (Q3) Networking Solutions ------------------------------ - eNAT - Network Address Translation (NAT) Software - eIKE & eIPSEC - IPSEC software for Embedded Systems (Q2/Q3) - HTTP Server, DHCP and SNMP V1 Supported CPUs : x86, StrongARM, ARM, MIPS, SH, 68k, DSP, etc. Supported RTOS : VxWorks, pSOS, Nucleus, ThreadX, OSE, OS/9 Windows CE, QNX, etc. Markets and Associations -------------------------------------- Intoto's solutions address the needs of the Embedded market for digital multimedia, consumer electronics such as HDTV, Settop boxes and DVD players; for computer peripherals such as Printers, Scanners, Disk Drivers, MFPs; and for Networking, Virtual Private Networks such as VPN router, Switch-Router, Load Balancer, etc. Intoto is an active member of 1394TA, HAVi, PWG-C, USB, OpenCable, VESA. MS SerifINTOTO, INC. THE Solution Provider for Embedded Systems: IEEE-1394 PCMCIA Flash IPSEC USB sIKE+IPSEC Debi Dawson-Lujan Marketing & Sales Coordinator 3160 De La Cruz Blvd. Suite # 100 Santa Clara CA 95054 T: 408 844 0480 x 303 F: 408 844 0488 From owner-vxwexplo-process Fri Apr 9 01:54:15 1999 Received: from postal1.lbl.gov (root@postal1.lbl.gov [128.3.7.82]) by csg.lbl.gov (8.8.7/8.8.7) with ESMTP id BAA15449 for ; Fri, 9 Apr 1999 01:54:14 -0700 (PDT) Received: from SpamWall.lbl.gov (root@localhost [127.0.0.1]) by postal1.lbl.gov (8.9.0/8.9.0) with ESMTP id BAA17059 for ; Fri, 9 Apr 1999 01:54:14 -0700 (PDT) Received: from mustard.roke.co.uk (exchange.roke.co.uk [193.118.192.1]) by SpamWall.lbl.gov (8.9.0/8.9.0) with SMTP id BAA17043 for ; Fri, 9 Apr 1999 01:53:58 -0700 (PDT) Received: from rsys002a.roke.co.uk by mustard.roke.co.uk with SMTP (PP) with ESMTP; Fri, 9 Apr 1999 09:49:09 +0100 Received: by RSYS002A with Internet Mail Service (5.5.2448.0) id <2DH3CQ9F>; Fri, 9 Apr 1999 09:53:12 +0100 Message-ID: From: "Moy, Ian" To: "'vxwexplo@lbl.gov'" Subject: if_fei and sys557Init. Date: Fri, 9 Apr 1999 09:53:01 +0100 X-Mailer: Internet Mail Service (5.5.2448.0) MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Hi, I am running VxWorks on a PowerPC, MBX860 and would like to add a Fast Ethernet module. The ethernet module is based on the Intel 82558 ethernet controller, is for the pc/104-Plus standard, made by E.E.P.D and is called Profive ethernet-100. If anyone has experience with this module, is it any good? Does anybody know of any more pc/104-Plus ethernet modules? Will if_fei driver this device? What has to be in the sys557Init function for the above setup? Is the function specific to the main board, the ethernet module or both? Any help or advice is welcome, Ian. > -------------------------------------------------- > Ian Moy, Roke Manor Research Ltd > (email) ian.moy@roke.co.uk > -------------------------------------------------- > > From owner-vxwexplo-process Fri Apr 9 04:00:12 1999 From: daemon@csg.lbl.gov Date: Fri Apr 9 04:00:15 PDT 1999 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Fri Apr 9 04:00:09 PDT 1999 Subject: Software quality assurance information ------------------------------------------------------- Newsgroups: comp.arch.embedded,comp.os.vxworks,comp.realtime,comp.software.config-mgmt,comp.software.testing,comp.software-eng Subject: Software quality assurance information Date: Thu, 8 Apr 1999 18:46:35 -0500 From: "Peter Stephens" Message-ID: <%dbP2.521$E4.2411@news1.iquest.net> Reply-To: "Peter Stephens" To everyone out there in embedded land. I am writing a thesis on software quality assurance and was wondering if anyone out there knows or has statistics about SQA in the embedded world? I am interested in things like: % of embedded companies that have independent SQA groups % of embedded companies that have SQA programs % of embedded companies that are SEI level x % of embedded companies that have ISO9001 - software certification etc... Any ideas on how to get information like this? - -- Pete Stephens pastephe@holli.com peter.a.stephens@delphiauto.com --------------------------- End of New-News digest ********************** From owner-vxwexplo-process Fri Apr 9 05:20:40 1999 From: RobertK54@aol.com Date: Fri Apr 9 05:20:45 PDT 1999 Subject: Tornado ARM Compiler Bugs CRITICAL PROBLEM Can anyone tell me if they have encountered problems with the ARM compiler (ccarm) from Tornado 1.0.1. For that matter, can anyone tell me if they have encountered bugs with any of the Tornado compilers (cc960, ccppc)? I found a case where it is generating bad assembly code. Take for example the following code which is similar to the code that causes the problem. (I did not try to compile this example so I can't say if this exact code duplicates my problem.) testfile.c: typedef struct { int f; int g; } aMsg; static struct { int x; int y; emum {A, B, C} z; } a[5]; void func(aMsg *m) { switch (m->f) { case SOMETHING: { int i; i = m->g; printf ("0x%x, 0x%x, 0x%x, 0x%x\n", a[1].x, a[i].x, *((int *) &a[i].x)), a[m->g].x); } } If the array a[] is initialized as follows: a[0] = 0, 0, A a[1] = 0x1234, 0, A a[2] = 0, 0, A a[3] = 0, 0, A a[4] = 0, 0, A The printf would generate the following: "0x1234, 0x0, 0x1234, 0x1234" Clearly the second item is wrong. Looking at the disassembled object file as well as a logic analyzer trace, I verified that it is the code that is wrong. This is NOT a runtime issue of some pointer accessing the wrong locations. The assembly code indicates that the second item is read as follows: 1) load &a into r5 (r2 = ???; r5 = &a; r6 = ???) 2) load r5 indexed by 0 into r6 (r6 = a[0].x) (r2 = ???; r5 = &a; r6 = 0x00000000) 3) load r5 indexed by r6 into r2 (r2 = a[a[0].x].x) (r2 = 0x00000000; r5 = &a; r6 = 0x00000000) In the above r2 is the register containing the parameter to the printf call. It is clear (to me!) that the compiler is not consistent. It is able to properly use 'i' to index into the a[] array in one case, but not another. How widespread is this problem?? Thanks, Bob From owner-vxwexplo-process Fri Apr 9 06:43:31 1999 From: Rainer Stelzer Date: Fri Apr 9 06:43:35 PDT 1999 Subject: MBX860 BSP questions: SPI,WindView Dies ist eine mehrteilige Nachricht im MIME-Format. --------------D61C5319F371B7105EF19762 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit hi, i've two questions to WRS MBX860 board support pakage : 1. Including the Windview option by setting #define INCLUDE_INSTRUMENTATION in config.h makes the vxworks image unbootable. Any hint ? 2. We tried to utilize the SPI Interface in master mode, but until now we didn't notice any SPI-action at the port pin at all.(see code attached below) What we find out until now: - setting the ports using PBDAT,PBDDIR works fine, (-> interface is electrical ok) - the clock brgclk is supplied (seen on baudrate output) - the CPM comand causes the copy of the rbase + tbase pointers to rbptr+ tbptr - this is the last action we are able to notice, not even the internal state of the SPI-machine seems to change after SPI start. The BSP includes the I2C/SPI microcode patch. Although the microcode patch is called I2C/SPI patch we can not seen any changes in the SPI interface handling caused by the patch , so we use the SPI DP RAM adresses, as discribed in the MPC manual. Does anybody out there use the SPI Interface on the MBX860 board successfully ? any hint is welcome best regards rainer vxworks tornado ********************************************************************** fraunhofer institute of microelectronic circuits and systems department SAT (systems and applications) name rainer stelzer address finkenstrasse 61 47057 duisburg germany phone +0049/203/3783-238 fax +0049/203/3783-266 email rstelzer@ims.fhg.de web www.fhg.de/depts/ims-de-e.html ********************************************************************** --------------D61C5319F371B7105EF19762 Content-Type: application/x-unknown-content-type-c_auto_file; name="spi_test1.c" Content-Transfer-Encoding: base64 Content-Disposition: inline; filename="spi_test1.c" I2luY2x1ZGUgInZ4d29ya3MuaCINCiNpbmNsdWRlIDx0YXNrbGliLmg+DQojaW5jbHVkZSA8 ZHJ2L2ludHJDdGwvcHBjODYwSW50ci5oPg0KI2luY2x1ZGUgIml2LmgiDQojaW5jbHVkZSA8 aW50TGliLmg+DQojaW5jbHVkZSA8c3BpX2RlZnMuaD4NCg0KDQoNCnZvaWQgdGVzdHNwaSh2 b2lkKQ0Kew0KCVNQSV9QQVJBTSAJKnNwaXB0cjsNCglTUElfRFJWX0JEICAqc3BpZGVzOw0K CVVJTlQ4CQlzbmRidWZmZXJbU1BJX1RYQkxfTUFYXTsNCglVSU5UOAkJcmN2YnVmZmVyW1NQ SV9SWEJMX01BWF07DQoJRkFTVCBVSU5UMzIgaW1tclZhbCA9IElOVEVSTkFMX01FTV9NQVBf QUREUjsgLyogQmFzZSBhZGRyZXNzIG9mIEludGVybmFsIE1lbW9yeSAqLw0KCVVJTlQzMiBs b29wY291bnQgPSAwOw0KCVVJTlQxNiBwaXBjID0gMDsNCglVSU5UOCBpbmhhbHQgPSAwOw0K CWludCBpOw0KCWludCBtOw0KDQojaWZkZWYgUEExX1RPR0dMRQ0KDQovKmFsbGdlbWVpbmUg QlJHQ0xLICBkdXJjaHNjaGFsdGVuIGF1ZiBCUkdPNChQQTEpKi8NCipCUkdDNChpbW1yVmFs KSA9IDB4MDAwMTAwMDA7DQoNCgkvKiBkZWZpbmUgUEFEQVQgdmFsdWUqLw0KCSpQQURBVChp bW1yVmFsKSAmPSB+UE9SVF9BX1BJTlNfQTE7IC8qIGxvZXNjaGUgbGF0Y2ggZnVlciBQQVsx XSovDQoJSU9fU1lOQzsNCg0KCS8qIGRlZmluZSBwb3J0IEEgUGluIDEqLw0KICAgICpQQVBB UihpbW1yVmFsKSB8PSBQT1JUX0FfUElOU19BMTsgIC8qIHBoZXJpcGhlcmFsIGZ1bmN0aW9u ICovDQogICAgSU9fU1lOQzsNCg0KICAgICpQQURJUihpbW1yVmFsKSB8PSBQT1JUX0FfUElO U19BMTsgIC8qIG91dHB1dCEqLw0KICAgIElPX1NZTkM7DQoNCiAgICAqUEFPRFIoaW1tclZh bCkgJj0gflBPUlRfQV9QSU5TX0ExOyAvKiBlaWdlbmRsaWNoIGVnYWwsIHp1IHRlc3R6d2Vj a2VuIGFscyBBS1RJVkVSIFRyZWliZXIgZ2VzZXR6dCEqLw0KICAgIElPX1NZTkM7DQoNCglm b3IgKGxvb3Bjb3VudCA9IDA7IGxvb3Bjb3VudCA8IDEwMDsgbG9vcGNvdW50KyspDQoJCXsN CgkNCgkJdGFza0RlbGF5KDEwKTsgIC8qIG51bWJlciBvZiB0aWNrcyB0byBkZWxheSB0YXNr ICovDQoNCgkgICAgKlBBREFUKGltbXJWYWwpIHw9IFBPUlRfQV9QSU5TX0ExOyAJLyogc2V0 emUgcGluID0gaGlnaCAqLw0KCSAgICBJT19TWU5DOw0KCQ0KCQl0YXNrRGVsYXkoMTApOyAg LyogbnVtYmVyIG9mIHRpY2tzIHRvIGRlbGF5IHRhc2sgKi8NCg0KCSAgICAqUEFEQVQoaW1t clZhbCkgJj0gflBPUlRfQV9QSU5TX0ExOyAJLyogc2V0emUgcGluID0gbG93ICovDQoJICAg IElPX1NZTkM7DQoNCgkJfQ0KI2VuZGlmDQoNCg0KDQoNCg0KDQoNCg0KLypzcGVpY2hlciAm IGJ1ZmZlciovDQoNCgkvKiBpbml0aWFsaXNpZXJlIFNQSS1QYXJhbWV0ZXIgUkFNKi8NCglz cGlwdHIgPSAoU1BJX1BBUkFNICopICAweGZhMjAzZDgwOyAvKiBJTU1SK0RQUl9CQVNFKzB4 MWQ4MCovDQoNCgkvKiBsZWdlIEJEIGluIGZyZWllbSBEUFJBTS1CZXJlaWNoIGFuKi8NCglz cGlkZXMgPSAoU1BJX0RSVl9CRCAqKSAoSU5URVJOQUxfTUVNX01BUF9BRERSICsgU1BJX0JE X09GRlNFVCk7IC8qIDB4ZmEyMDAwMDAgKyAweDAwMDAyQTAwICovDQoNCg0KI2lmbmRlZiBO SUUNCglmb3IoaT0wO2k8c2l6ZW9mKFNQSV9QQVJBTSk7aSsrKQ0KCXsNCgkJKChVSU5UOCop c3BpcHRyKVtpXT0wOw0KCX0NCglmb3IoaT0wO2k8c2l6ZW9mKFNQSV9EUlZfQkQpO2krKykN Cgl7DQoJCSgoVUlOVDgqKXNwaWRlcylbaV09MDsNCgl9DQojZW5kaWYNCg0KDQoJaW50Q29u bmVjdChJVl9TUEksIChWT0lERlVOQ1BUUilzcGlpbnQsaW1tclZhbCk7DQoNCi8qIHJlZ2lz dGVyICovDQoNCiAgICAvKiBkaXNhYmxlIHNwaSBvcGVyYXRpb24gKi8NCi8qICAgICpTUE1P REUoaW1tclZhbCkgJj0gflNQTU9ERV9FTjsqLw0KICAgICpTUE1PREUoaW1tclZhbCkgPSAg KFNQTU9ERV9ESVYxNiB8IFNQTU9ERV9NUyB8IFNQTU9ERV9MRU5fOF9CSVRTIHwgU1BNT0RF X1BNX0RJVjQpOyAgDQogICAgSU9fU1lOQzsNCg0KDQogICAgDQogCXBpcGMgPSAoVUlOVDE2 KSAqUElQQyhpbW1yVmFsKSA7IA0KCS8qZGlzYWJsZSBQYXJhbGxlbCBJbnRlcmZhY2UgUG9y dCAoUElQKSovDQogICAgKlBJUEMoaW1tclZhbCkgJj0gfihQSVBDX01PRExfR0VOSU8gfCBQ SVBDX01PREhfR0VOSU8pOyANCiAgICBJT19TWU5DOw0KDQoNCgkvKiBkZWZpbmUgUEJEQVQg dmFsdWUqLw0KCSpQQkRBVChpbW1yVmFsKSAmPSB+KFBPUlRfQl9QSU5TX1NQSVNFTG91dCB8 IFBPUlRfQl9QSU5TX1NQSSB8IFBPUlRfQl9QSU5TX1NQSVNFTGluICk7IC8qIGxvZXNjaGUg bGF0Y2ggZnVlciBQQlsyNSwyOC4uMzFdKi8NCglJT19TWU5DOw0KDQoJLyogZW5hYmxlIHRo ZSBTUElNSVNPLFNQSU1PU0kgYW5kIFNQSUNMSyBwaW5zICovDQogICAgKlBCUEFSKGltbXJW YWwpIHw9IFBPUlRfQl9QSU5TX1NQSTsgIC8qIHNwaS1tb2R1cyAqLw0KICAgIElPX1NZTkM7 DQoNCiAgICAqUEJESVIoaW1tclZhbCkgfD0gUE9SVF9CX1BJTlNfU1BJOyAgLyogcGhlcmlw aGVyYWwgZnVuY3Rpb24sIHdlbm4gUEJQQVIgaW0gc3BpLW1vZHVzKi8NCiAgICBJT19TWU5D Ow0KDQogICAgKlBCT0RSKGltbXJWYWwpICY9IH5QT1JUX0JfUElOU19TUEk7IC8qIEFMTEUg enUgdGVzdHp3ZWNrZW4gYWxzIEFLVElWRVIgVHJlaWJlciBnZXNldHp0ISovDQogICAgSU9f U1lOQzsNCg0KCQ0KCS8qIGNvbmZpZ3VyZSB+U1BJU0VMaW4gcGluICghIGJ6dy4gYWxsZy4g cG9ydC1vdXRwdXQgd2cuIG11bHRpbWFzdGVyLWVycm9yKSovDQogICAgKlBCUEFSKGltbXJW YWwpICY9IH5QT1JUX0JfUElOU19TUElTRUxpbjsgLyogenUgdGVzdHp3ZWNrZW4gYWxzIHBv cnQtc2lnbmFsIGdlc2V0enQhKi8NCiAgICBJT19TWU5DOw0KDQogICAgKlBCRElSKGltbXJW YWwpIHw9IFBPUlRfQl9QSU5TX1NQSVNFTGluOwkvKiBwaGVyaXBoZXJhbCBmdW5jdGlvbiwg d2VubiBQQlBBUiBpbSBzcGktbW9kdXMqLw0KICAgIElPX1NZTkM7DQoNCiAgICAqUEJPRFIo aW1tclZhbCkgJj0gflBPUlRfQl9QSU5TX1NQSVNFTGluOyAvKiB6dSB0ZXN0endlY2tlbiBh bHMgQUtUSVZFUiBUcmVpYmVyIGdlc2V0enQhKi8NCiAgICBJT19TWU5DOw0KDQoNCiNpZmRl ZiBUT0dHTEVfUE9SVFBJTg0KCWZvciAobG9vcGNvdW50ID0gMDsgbG9vcGNvdW50IDwgMTAw OyBsb29wY291bnQrKykNCgkJew0KCQ0KCQl0YXNrRGVsYXkoMTApOyAgLyogbnVtYmVyIG9m IHRpY2tzIHRvIGRlbGF5IHRhc2sgKi8NCg0KCSAgICAqUEJEQVQoaW1tclZhbCkgfD0gUE9S VF9CX1BJTlNfU1BJU0VMaW47IAkvKiBzZXR6ZSBwaW4gPSBoaWdoICovDQoJICAgIElPX1NZ TkM7DQoJDQoJCXRhc2tEZWxheSgxMCk7ICAvKiBudW1iZXIgb2YgdGlja3MgdG8gZGVsYXkg dGFzayAqLw0KDQoJICAgICpQQkRBVChpbW1yVmFsKSAmPSB+UE9SVF9CX1BJTlNfU1BJU0VM aW47IAkvKiBzZXR6ZSBwaW4gPSBsb3cgKi8NCgkgICAgSU9fU1lOQzsNCg0KCQl9DQojZWxz ZQ0KCSAgICAqUEJEQVQoaW1tclZhbCkgJj0gflBPUlRfQl9QSU5TX1NQSVNFTGluOyAJLyog c2V0emUgcGluID0gbG93ICovDQogCSAgICBJT19TWU5DOw0KI2VuZGlmDQoNCgkvKiBTUEkt UGFyYW1ldGVyIHNldHplbiBmdWVyIEJEIGJhc2UgYWRyZXNzZSovIA0KCQkJCS8qIG9mZnNl dCBkZXMgU1BJX0JEIHp1ciBEUFJBTV9CQVNFIGVybWl0dGVsbiAqLw0KCQkJCS8qIDB4RkEy MDJBMDAgLSAoMHhGQTIwMDAwMCArIDB4MDAwMDIwMDApID0gMHgwMDAwMEEwMCAqLw0KCQkJ CS8qIDB4RkEyMDJBMDggLSAoMHhGQTIwMDAwMCArIDB4MDAwMDIwMDApID0gMHgwMDAwMEEw OCAqLw0KCQ0KCXNwaXB0ci0+cmJhc2UgPSAoVUlOVDE2KSgoVUlOVDMyKSgmc3BpZGVzLT5y eGJkWzBdKSAtIChVSU5UMzIpKCBJTlRFUk5BTF9NRU1fTUFQX0FERFIgK0RQUl9CQVNFKSk7 DQoJc3BpcHRyLT50YmFzZSA9IChVSU5UMTYpKChVSU5UMzIpKCZzcGlkZXMtPnR4YmRbMF0p IC0gKFVJTlQzMikoIElOVEVSTkFMX01FTV9NQVBfQUREUiArRFBSX0JBU0UpKTsNCi8qCXNw aXB0ci0+cmJhc2UgPSAweDBBMDA7DQoJc3BpcHRyLT50YmFzZSA9IDB4MEEwODsNCiovDQoN CgkvKiBpbml0aWFsaXplIFNETUEqLw0KI2lmZGVmIE5JRQ0KICAgICpTRENSKGltbXJWYWwp ID0gMHgwMDAxOwkvKiBTRENSX1JBSURfQlI1IDB4MDAwMDAwMDEgKi8gLyogVS1CVVMgYXJi aXRyYXRpb24gcHJpb3JpdHkgNSAoQlI1KSAqLyANCiAgICBJT19TWU5DOw0KI2VuZGlmDQoN CgkvKiB3ZWl0ZXJlIFNQSS1QYXJhbWV0ZXIgc2V0emVuICovDQoNCiAgICBzcGlwdHItPnJm Y3IgID0gMHgwODsNCiAgICBzcGlwdHItPnRmY3IgID0gMHgwODsNCg0KCS8qIGlzc3VlICdp bml0aWFsaXplIFJ4IGFuZCBUeCBwYXJhbWV0ZXJzJyBjb21tYW5kICovDQogICAgbWJ4U3Bp ZENjciAoaW1tclZhbCwgQ1BNX0NSX09QQ09ERV9JTklUX1JUKTsNCg0KDQoJLyogd2VpdGVy ZSBTUEktUGFyYW1ldGVyIHNldHplbiAqLw0KICAgIHNwaXB0ci0+bXJibHIgPSBTUElfUlhC TF9NQVg7DQogIC8qICBzcGlwdHItPnJpcyAgID0gMHgwMDAwMDAwMDsNCiAgICBzcGlwdHIt PnJicHRyID0gc3BpcHRyLT5yYmFzZTsNCiAgICBzcGlwdHItPnRpcyAgID0gMHgwMDAwMDAw MDsNCiAgICBzcGlwdHItPnRicHRyID0gc3BpcHRyLT50YmFzZTsqLw0KDQoJLyogaW5pdGlh bGlzaWVyZSBSWC1CRCovDQoJc3BpZGVzLT5yeGJkWzBdLnN0YXR1c01vZGUgID0gKCBCRF9S WF9FTVBUWV9CSVQgfCBCRF9SWF9XUkFQX0JJVCApOw0KCXNwaWRlcy0+cnhiZFswXS5zdGF0 dXNNb2RlICAmPSB+KCBCRF9SWF9DT05fTU9ERV9CSVQgfCBCRF9SWF9JTlRFUlJVUFRfQklU KSA7DQoNCglzcGlkZXMtPnJ4YmRbMF0uZGF0YUxlbmd0aCAgPSAweDAwMDA7CQkvKiBpbml0 aWFsbHkgY29udGFpbmluZyB6ZXJvIGNoYXJzICovDQoJc3BpZGVzLT5yeGJkWzBdLmRhdGFQ b2ludGVyID0gJnJjdmJ1ZmZlclswXTsNCg0KCS8qIGluaXRpYWxpc2llcmUgVFgtQkQqLw0K CXNwaWRlcy0+dHhiZFswXS5zdGF0dXNNb2RlICB8PSAoQkRfVFhfUkVBRFlfQklUfEJEX1RY X0lOVEVSUlVQVF9CSVR8QkRfVFhfV1JBUF9CSVR8QkRfVFhfTEFTVF9CSVQpOw0KCXNwaWRl cy0+dHhiZFswXS5zdGF0dXNNb2RlICAmPSB+QkRfVFhfQ09OX01PREVfQklUIDsNCg0KCXNw aWRlcy0+dHhiZFswXS5kYXRhTGVuZ3RoICA9IDB4MDAwNTsJCS8qIGNvbnRhaW5zIGZpdmUg Y2hhcmFjdGVycyAqLw0KCXNwaWRlcy0+dHhiZFswXS5kYXRhUG9pbnRlciA9ICZzbmRidWZm ZXJbMF07DQoNCgkvKiBkdW1teSB3ZXJ0IHp1bSBzZW5kZW4gKi8NCi8qCWZvcihtPTA7bTxz aXplb2Yoc25kYnVmZmVyW1NQSV9UWEJMX01BWF0pO2krKykNCgl7DQoJCXNuZGJ1ZmZlcltt XT0weDVhOw0KCX0NCiovDQoNCglzbmRidWZmZXJbMF0gPSAweDVhOw0KCXNuZGJ1ZmZlclsx XSA9IDB4YTU7DQoJc25kYnVmZmVyWzJdID0gMHg1YTsNCglzbmRidWZmZXJbM10gPSAweGE1 Ow0KCXNuZGJ1ZmZlcls0XSA9IDB4NWE7DQoNCiNpZmRlZiBCVUZGRVJJTkhBTFQNCglpbmhh bHQgPSAoVUlOVDgpICooc3BpZGVzLT50eGJkWzBdLmRhdGFQb2ludGVyKTsNCiNlbmRpZg0K DQogICAgLyoNCiAgICAgKiBjbGVhciBhbnkgcHJldmlvdXMgaW50ZXJydXB0IHN0YXR1cw0K ICAgICAqIGVuYWJsZSBhbGwgc3BpIGludGVycnVwdHMgDQogICAgICovDQogICAgKlNQSUUo aW1tclZhbCkgPSBTUElfSU5UX0NMUjsNCiAgICBJT19TWU5DOw0KDQogICAgKlNQSU0oaW1t clZhbCkgPSBTUElfSU5UX0VOQkw7DQogICAgSU9fU1lOQzsNCg0KICAgIC8qIGVuYWJsZSBT UEktZ2VuZXJhdGVkIHN5c3RlbSBpbnRlcnJ1cHRzICovDQogICAgKkNJTVIoaW1tclZhbCkg fD0gQ0lNUl9TUEk7DQogICAgSU9fU1lOQzsNCg0KI2lmZGVmIElOVA0KICAgIC8qIEludGVy cnVwdCBlbmFibGUuIE1hc3RlciBlbmFibGUgZm9yIENQTSBpbnRlcnJ1cHRzLiAqLw0KICAg ICpDSUNSKGltbXJWYWwpID0gKENJQ1JfSUVOIHwgIA0KDQogICAgIC8qIEludGVycnVwdCBy ZXF1ZXN0IGxldmVsLiBDb250YWlucyB0aGUgcHJpb3JpdHkgcmVxdWVzdCBsZXZlbCBvZiB0 aGUgDQogICAgICogaW50ZXJydXB0IGZyb20gdGhlIENQTSB0aGF0IGlzIHNlbnQgdG8gdGhl IFNJVS4gTGV2ZWwgMCBpbmRpY2F0ZXMNCiAgICAgKiBoaWdoZXN0IHByaW9yaXR5LiBJUkwg aXMgaW5pdGlhbGl6ZWQgdG8gemVybyBkdXJpbmcgcmVzZXQuIEluIG1vc3Qgc3lzdGVtcywN CiAgICAgKiB2YWx1ZSAwYjEwMCBpcyBhIGdvb2QgdmFsdWUgdG8gY2hvb3NlIGZvciBJUkwu Ki8NCgkJCShDSUNSX0lSTF9NU0sgJiBDSUNSX0lSTF9MVkw1KSk7DQogICAgIElPX1NZTkM7 CSANCiNlbmRpZg0KDQogICAgLyoNCiAgICAgKiBjb25maWd1cmUgdG8gbWFzdGVyIG1vZGUN CiAgICAgKiBlbmFibGUgU1BJIG9wZXJhdGlvbiwgY2hhckxlbmdodCA9IDggYml0LCBzcGlj bGs9QlJHQ0xLLygxNng0KQ0KICAgICAqLw0KDQogICAgKlNQTU9ERShpbW1yVmFsKSB8PSAg KFNQTU9ERV9FTiApOyAgDQoJSU9fU1lOQzsNCi8qCSpTUE1PREUoaW1tclZhbCkgJj0gfigg U1BNT0RFX0xPT1AgfCBTUE1PREVfQ0kgfCBTUE1PREVfQ1ApOw0KICAgIElPX1NZTkM7Ki8N Cg0KDQojaWZkZWYgTU1FDQogICAgLyogYXNzZXJ0IFNQSSBzZWxlY3Qgb3V0cHV0IHNpZ25h bCovDQogICAgKlBCREFUKGltbXJWYWwpIHw9IFBPUlRfQl9QSU5TX1NQSVNFTGluOyAvKiBv bmUgZm9yIHRoZSB0b2dnbGUgZWZmZWN0Li4uKi8NCiAgICBJT19TWU5DOw0KDQogICAgKlBC REFUKGltbXJWYWwpICY9IH5QT1JUX0JfUElOU19TUElTRUxpbjsgDQogICAgSU9fU1lOQzsN CiNlbmRpZg0KDQoJaT0gCXNwaXB0ci0+dGlzOw0KDQogICAgLyogc3RhcnQgc3BpIHRyYW5z ZmVyKi8NCiAgICAqU1BDT00oaW1tclZhbCkgfD0gU1BDT01fU1RSOw0KICAgIElPX1NZTkM7 ICAgIA0Kd2hpbGUoIGkgPT0gc3BpcHRyLT50aXMpOw0KDQoNCg0KDQp9ICAvKiBlbmQgc3Bp X3Rlc3QxKCkqLw0KDQoNCi8qKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioq KioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioNCioNCiogbWJ4U3Bp ZENjciAtIGV4ZWN1dGUgY2hhbm5lbCBjb21tYW5kDQoqIA0KKiBUaGlzIGNvbXBvbmVudCdz IHB1cnBvc2UgaXMgdG8gZXhlY3V0ZSB0aGUgY29tbWFuZA0KKiBwYXNzZWQuICBUaGUgY2hh bm5lbCBjb21tYW5kIHJlZ2lzdGVyIGlzIHJlYWQgcHJpb3INCiogdG8gZXhlY3V0aW9uIHRv IHZlcmlmeSB0aGF0IHRoZXJlIGFyZSBubyBvdXRzdGFuZGluZw0KKiBjb21tYW5kcywgdGhp cyBpcyBkb25lIGJ5IHdhaXRpbmcgZm9yIGEgemVybyBzdGF0dXMuDQoqIFRoZSBjaGFubmVs IGNvbW1hbmQgcmVnaXN0ZXIgaXMgYWxzbyByZWFkIGFmdGVyIHRoZSBjb21tYW5kDQoqIGV4 ZWN1dGlvbiB0byBpbnN1cmUgY29tbWFuZCBjb21wbGV0aW9uIHByaW9yIHRvIHJldHVybmlu Zy4NCioNCiogUkVUVVJOUzogTi9BDQoqLw0KDQp2b2lkIG1ieFNwaWRDY3INCiAgICAoDQog ICAgRkFTVCBVSU5UIGltbXJWYWwsICAgICAgICAgIC8qIEJhc2UgYWRkcmVzcyBvZiBJbnRl cm5hbCBNZW1vcnkgUmVncyAqLyANCiAgICBGQVNUIFVJTlQgY21kICAgICAgICAgICAgICAg LyogY29tbWFuZCB0byBleGVjdXRlICovDQogICAgKQ0KICAgIHsNCglGQVNUIFVJTlQzMgls b29wY291bnQyID0gMDsNCiAgICAvKiB3YWl0IGZvciBhbnkgcGVuZGluZyBjb21tYW5kcyB0 byBjb21wbGV0ZSAqLw0KDQogICAgd2hpbGUgKCpDUENSKGltbXJWYWwpICYgQ1BNX0NSX0ZM RykNCgk7DQoNCiAgICAvKiBpc3N1ZSBjb21tYW5kICovDQoNCiAgICAqQ1BDUihpbW1yVmFs KSA9IChjbWQgfCBDUE1fQ1JfQ0hBTk5FTF9TUEkgfCBDUE1fQ1JfRkxHKTsgDQogICAgSU9f U1lOQzsNCg0KICAgIC8qIHdhaXQgZm9yIGlzc3VlZCBjb21tYW5kIHRvIGNvbXBsZXRlICov DQoJbG9vcGNvdW50MisrOw0KDQogICAgd2hpbGUgKCpDUENSKGltbXJWYWwpICYgQ1BNX0NS X0ZMRykNCgl7DQoJbG9vcGNvdW50MisrOw0KCX0NCgk7DQogICAgfQ0KDQoNCnZvaWQgc3Bp aW50KFVJTlQzMiBpbW1yVmFsKQ0Kew0KCVVJTlQxNiBpbnRtYXNrOw0KDQoJaW50bWFzayA9 ICpTUElFKGltbXJWYWwpOw0KDQoJbG9nTXNnKCJTUEkgSW50ZXJydXB0ICUwMnggJTA4eFxu IixpbnRtYXNrLGltbXJWYWwsMCwwLDAsMCk7DQoNCiAgIC8qDQogICAgICogY2xlYXIgYW55 IHByZXZpb3VzIGludGVycnVwdCBzdGF0dXMNCiAgICAgKi8NCiAgICAqU1BJRShpbW1yVmFs KSA9IFNQSV9JTlRfQ0xSOw0KICAgIElPX1NZTkM7DQoNCiNpZmRlZglDTEVBUklOVA0KCSpD SVNSKGltbXJWYWwpIHw9IENJU1JfU1BJOyAvKiBjbGVhciB0aGUgaW50IGJ5IHdyaXRpbmcg YSBPTkUgdG8gQ0lTUltTUEldICovDQoJSU9fU1lOQzsNCg0KCV9fYXNtX18oIiAgICAgIHJm aSIpOw0KI2VuZGlmDQoNCn0NCg== --------------D61C5319F371B7105EF19762 Content-Type: text/plain; charset=us-ascii; name="vxworks_MBX860_DPRAM_MAP.txt" Content-Transfer-Encoding: 7bit Content-Disposition: inline; filename="vxworks_MBX860_DPRAM_MAP.txt" Data Parameter RAM (DPRAM) MAP The diagram below shows the Data Parameter Ram layout for the VxWorks configuration: ----------------------------- DPRAM base (address = IMMR + 0x2000) | 0x200 (512) bytes | | I2C/SCC1 conflict | | microcode patch part 1 | |---------------------------| end microcode patch (0x2200) | 8 bytes per descriptor | UART Tx Buffer Descriptor (0x2200) |===========================| | 16 descriptors @ | UART Rx Buffer Descriptors (0x2210) | 8 bytes per descriptor | |===========================| end UART Rx BDs (0x2290) | 80 hex bytes allowed | UART Tx Buffer (0x2300) |===========================| | one receive char/buffer | UART Rx Buffer (0x2380) |===========================| PIP Tx BD Buffer (0x2400) | 5 bytes CSR,Len,Addr | |---------------------------| end PIP Tx Buffer (0x2405) : : : : |---------------------------| | 16 bytes allowed | I2C Buffer Descriptors (0x2500) |---------------------------| | 32 bytes allowed | I2C Buffer (0x2510) |---------------------------| end @ 0x2530 : : : : |---------------------------| | 256 bytes allowed | SROM temp. buffer (0x2600) |---------------------------| | 64 bytes allowed | SROM temp. system values (0x2700) |---------------------------| end @ 0x2740 : : : : |---------------------------| | 32 descriptors @ | Enet Tx Buffer Descriptors (0x2800) | 8 bytes per descriptor | |---------------------------| | 32 descriptors @ | Enet Rx Buffer Descriptors (0x2900) | 8 bytes per descriptor | |---------------------------| end @ 0x2A00 : : : : inserted! |---------------------------| | 1 descriptor @ | SPI RX Buffer Descriptors (0x2A00) | 8 bytes per descriptor | |---------------------------| | 1 descriptor @ | SPI TX Buffer Descriptors (0x2A08) | 8 bytes per descriptor | |---------------------------| end @ 0x2A16 : : : : |---------------------------| romInit temp. stack frame (0x2EE0) | 32 bytes | |---------------------------| end romInit temp. stack (0x2F00) | 0x200 (512) bytes | microcode patch (0x2F00) | I2C/SCC1 conflict | | microcode patch part 2 | |---------------------------| end @ 0x3100 : : : : |---------------------------| | A4 hex bytes of parameter | SCC1 Parameter Area (0x3c00) Enet | info incl. Rx and Tx BD | | pointers, func codes etc. | |---------------------------| end @ 0x3CA4 : : : : inserted! |---------------------------| | 48 bytes availiable, | SPI Parameter Area (0x3D80) | 40 bytes used for | | parameter info | end @ 0x3DAF |---------------------------| : : : : |---------------------------| | 3A hex bytes of parameter | SMC1 Parameter Area (0x3e80) UART | info including Rx and Tx | | BD pointers, func codes | | etc... | |---------------------------| end @ 0x3EBA : : --------------D61C5319F371B7105EF19762 Content-Type: text/plain; charset=us-ascii; name="spi_defs.h" Content-Transfer-Encoding: 7bit Content-Disposition: inline; filename="spi_defs.h" #undef TOGGLE_PORTPIN #undef BUFFERINHALT #undef PA1_TOGGLE #define IO_SYNC __asm__(" sync") /* Macro for all i/o operations to use */ #define FAST register #define SDCR(base) (UINT32 *) (base + 0x0030) /* SDMA Config. Reg */ #define CPCR(base) (UINT16 *) (base + 0x09C0) /* Com Processor Cmd*/ #define SPI_DRV_NBDS 1 /* Number of BDs*/ #define SPI_RXBL_MAX 0x0010 /*16*/ /* maximum RX buffer length */ #define SPI_TXBL_MAX 0x0010 /*16*/ /* maximum TX buffer length */ #define SPI_INT_CLR 0xff /* clear interrupt status */ #define SPI_INT_DSBL 0x0 /* disable all interrupts */ #define SPI_INT_ENBL 0x37 /* enable all possible SPI interrupts */ /* SPI register set */ #define SPMODE(base) (UINT16 *) (base + 0x0AA0) /* SPI Mode Reg */ #define SPIE(base) (UINT8 *) (base + 0x0AA6) /* SPI Event Reg */ #define SPIM(base) (UINT8 *) (base + 0x0AAA) /* SPI Mask Reg */ #define SPCOM(base) (UINT8 *) (base + 0x0AAD) /* SPI Command Reg */ /* SPI Mode Register bit definition (SPMODE - 0x0AA0) */ #define SPMODE_LOOP 0x4000 /* Loop Mode */ #define SPMODE_CI 0x2000 /* Clock Invert */ #define SPMODE_CI_LOW 0x0000 /* Inactive state is low */ #define SPMODE_CI_HIGH 0x2000 /* Inactive state is high */ #define SPMODE_CP 0x1000 /* Clock Phase */ #define SPMODE_CP_MIDDLE 0x0000 /* Clock Phase: middle of the data */ #define SPMODE_CP_BEGIN 0x1000 /* Clock Phase: Beginning of the data */ #define SPMODE_DIV16 0x0800 /* BRGCLK Divide by 16 */ #define SPMODE_REV 0x0400 /* Reverse Data */ #define SPMODE_MS 0x0200 /* SPI Master */ #define SPMODE_EN 0x0100 /* Enable SPI */ #define SPMODE_LEN_MSK 0x00f0 /* Character Length */ #define SPMODE_LEN_1_BIT 0x0000 /* Character Length: 1 bit */ #define SPMODE_LEN_2_BITS 0x0010 /* Character Length: 2 bits */ #define SPMODE_LEN_3_BITS 0x0020 /* Character Length: 3 bits */ #define SPMODE_LEN_4_BITS 0x0030 /* Character Length: 4 bits */ #define SPMODE_LEN_5_BITS 0x0040 /* Character Length: 5 bits */ #define SPMODE_LEN_6_BITS 0x0050 /* Character Length: 6 bits */ #define SPMODE_LEN_7_BITS 0x0060 /* Character Length: 7 bits */ #define SPMODE_LEN_8_BITS 0x0070 /* Character Length: 8 bits */ #define SPMODE_LEN_9_BITS 0x0080 /* Character Length: 9 bits */ #define SPMODE_LEN_10_BITS 0x0090 /* Character Length: 10 bits */ #define SPMODE_LEN_11_BITS 0x00a0 /* Character Length: 11 bits */ #define SPMODE_LEN_12_BITS 0x00b0 /* Character Length: 12 bits */ #define SPMODE_LEN_13_BITS 0x00c0 /* Character Length: 13 bits */ #define SPMODE_LEN_14_BITS 0x00d0 /* Character Length: 14 bits */ #define SPMODE_LEN_15_BITS 0x00e0 /* Character Length: 15 bits */ #define SPMODE_LEN_16_BITS 0x00f0 /* Character Length: 16 bits */ #define SPMODE_PM_MSK 0x000f /* BRGCLK Prescale Modulus select mask*/ #define SPMODE_PM_DIV4 0x0000 /* BRGCLK clock divided by 4 */ #define SPMODE_PM_DIV8 0x0001 /* BRGCLK clock divided by 8 */ #define SPMODE_PM_DIV12 0x0002 /* BRGCLK clock divided by 12 */ #define SPMODE_PM_DIV16 0x0003 /* BRGCLK clock divided by 16 */ #define SPMODE_PM_DIV20 0x0004 /* BRGCLK clock divided by 20 */ #define SPMODE_PM_DIV24 0x0005 /* BRGCLK clock divided by 24 */ #define SPMODE_PM_DIV28 0x0006 /* BRGCLK clock divided by 28 */ #define SPMODE_PM_DIV32 0x0007 /* BRGCLK clock divided by 32 */ #define SPMODE_PM_DIV36 0x0008 /* BRGCLK clock divided by 36 */ #define SPMODE_PM_DIV40 0x0009 /* BRGCLK clock divided by 40 */ #define SPMODE_PM_DIV44 0x000a /* BRGCLK clock divided by 44 */ #define SPMODE_PM_DIV48 0x000b /* BRGCLK clock divided by 48 */ #define SPMODE_PM_DIV52 0x000c /* BRGCLK clock divided by 52 */ #define SPMODE_PM_DIV56 0x000d /* BRGCLK clock divided by 56 */ #define SPMODE_PM_DIV60 0x000e /* BRGCLK clock divided by 60 */ #define SPMODE_PM_DIV64 0x000f /* BRGCLK clock divided by 64 */ /* SPI Command Register (SPCOM - 0x0AAD) */ #define SPCOM_STR 0x80 /* Start Transmit */ /* Command Register definitions (CPCR - 0x9C0) */ #define CPM_CR_FLG 0x0001 /* flag - command executing */ #define CPM_CR_RESET 0x8000 /* software reset command */ #define CPM_CR_OPCODE 0x0f00 /* command opcode */ #define CPM_CR_CHAN(x) (((x) & 0x000f) << 4) #define CPM_CR_OP(x) (((x) & 0x000f) << 8) /* CPCR - Channel Numbers */ #define CPM_CR_CHANNEL_SCC1 CPM_CR_CHAN(0x0) /* SCC1 channel */ #define CPM_CR_CHANNEL_I2C CPM_CR_CHAN(0x1) /* I2C channel */ #define CPM_CR_CHANNEL_SCC2 CPM_CR_CHAN(0x4) /* SCC2 channel */ #define CPM_CR_CHANNEL_SPI CPM_CR_CHAN(0x5) /* SPI channel */ #define CPM_CR_CHANNEL_RTMR CPM_CR_CHAN(0x5) /* RISC timer channel */ #define CPM_CR_CHANNEL_SCC3 CPM_CR_CHAN(0x8) /* SCC3 channel */ #define CPM_CR_CHANNEL_SMC1 CPM_CR_CHAN(0x9) /* SMC1 channel */ #define CPM_CR_CHANNEL_IDMA1 CPM_CR_CHAN(0x9) /* IDMA1 channel */ #define CPM_CR_CHANNEL_SCC4 CPM_CR_CHAN(0xc) /* SCC4 channel */ #define CPM_CR_CHANNEL_SMC2 CPM_CR_CHAN(0xd) /* SMC2 channel */ #define CPM_CR_CHANNEL_PIP CPM_CR_CHAN(0xd) /* PIP channel */ #define CPM_CR_CHANNEL_IDMA2 CPM_CR_CHAN(0xd) /* IDMA2 channel */ /* CPCR - opcodes */ #define CPM_CR_OPCODE_INIT_RT CPM_CR_OP(0x0) /* Init rx and tx */ #define CPM_CR_OPCODE_INIT_R CPM_CR_OP(0x1) /* init rx only */ #define CPM_CR_OPCODE_INIT_T CPM_CR_OP(0x2) /* init tx only */ #define CPM_CR_OPCODE_HUNT CPM_CR_OP(0x3) /* rx frame hunt mode */ #define CPM_CR_OPCODE_STOP CPM_CR_OP(0x4) /* stop tx */ #define CPM_CR_OPCODE_GRSTOP CPM_CR_OP(0x5) /* gracefully stop tx */ #define CPM_CR_OPCODE_IDMA_INIT CPM_CR_OP(0x5) /* init idma */ #define CPM_CR_OPCODE_RESTART CPM_CR_OP(0x6) /* restart tx */ #define CPM_CR_OPCODE_CLOSE CPM_CR_OP(0x7) /* close rx buffer */ #define CPM_CR_OPCODE_SET_GROUP CPM_CR_OP(0x8) /* set group address */ #define CPM_CR_OPCODE_SET_ITMR CPM_CR_OP(0x8) /* set itimer */ #define CPM_CR_OPCODE_GCI_TMO CPM_CR_OP(0x9) /* gci timeout */ #define CPM_CR_OPCODE_RESET_BCS CPM_CR_OP(0xa) /* blk chk seq reset */ #define CPM_CR_OPCODE_GCI_ABRT CPM_CR_OP(0xa) /* gci abort request */ #define CPM_CR_OPCODE_IDMA_STOP CPM_CR_OP(0xb) /* stop idma */ #define CPM_CR_OPCODE_DSP_START CPM_CR_OP(0xc) /* start DSP */ #define CPM_CR_OPMODE_DSP_INIT CPM_CR_OP(0xd) /* init DSP */ /* CPM Interrupt Control registers */ #define CIVR(base) (UINT16 *) (base + 0x0930) /* CP Int vector */ #define CICR(base) (UINT32 *) (base + 0x0940) /* CP Int config */ #define CIPR(base) (UINT32 *) (base + 0x0944) /* CP Int pending */ #define CIMR(base) (UINT32 *) (base + 0x0948) /* CP Int mask reg */ #define CISR(base) (UINT32 *) (base + 0x094C) /* CP Int in-service*/ /* CPM Interrupt Mask Register (CIMR - 0x948) */ #define CIMR_PC15 0x80000000 /* Parallel I/O Port C[15] */ #define CIMR_SCC1 0x40000000 /* SCC 1 */ #define CIMR_SCC2 0x20000000 /* SCC 2 */ #define CIMR_SCC3 0x10000000 /* SCC 3 */ #define CIMR_SCC4 0x08000000 /* SCC 4 */ #define CIMR_PC14 0x04000000 /* Parallel I/O Port C[14] */ #define CIMR_TIMER1 0x02000000 /* Timer 1 */ #define CIMR_PC13 0x01000000 /* Parallel I/O Port C[13] */ #define CIMR_PC12 0x00800000 /* Parallel I/O Port C[12] */ #define CIMR_SDMA 0x00400000 /* SDMA bus error */ #define CIMR_IDMA1 0x00200000 /* IDMA 1 */ #define CIMR_IDMA2 0x00100000 /* IDMA 2 */ #define CIMR_TIMER2 0x00040000 /* Timer 2 */ #define CIMR_R_TT 0x00020000 /* Risc Timer Table */ #define CIMR_I2C 0x00010000 /* I2C */ #define CIMR_PC11 0x00008000 /* Parallel I/O Port C[11] */ #define CIMR_PC10 0x00004000 /* Parallel I/O Port C[10] */ #define CIMR_TIMER3 0x00001000 /* Timer 3 */ #define CIMR_PC9 0x00000800 /* Parallel I/O Port C[9] */ #define CIMR_PC8 0x00000400 /* Parallel I/O Port C[8] */ #define CIMR_PC7 0x00000200 /* Parallel I/O Port C[7] */ #define CIMR_TIMER4 0x00000080 /* Timer 4 */ #define CIMR_PC6 0x00000040 /* Parallel I/O Port C[6] */ #define CIMR_SPI 0x00000020 /* SPI */ #define CIMR_SMC1 0x00000010 /* SMC 1 */ #define CIMR_SMC2_PIP 0x00000008 /* SMC 1 */ #define CIMR_PC5 0x00000004 /* Parallel I/O Port C[5] */ #define CIMR_PC4 0x00000002 /* Parallel I/O Port C[4] */ #define CIMR_ALL 0xfff7defe /* all interrupt masks */ /* CPM Interrupt configuration Register (CICR - 0x940) */ #define CICR_SCCDP_MSK 0xc00000 /* SCCd Priority Order Mask */ #define CICR_SCCDP_SCC1 0x000000 /* SCC1 asserts the SCCd */ #define CICR_SCCDP_SCC2 0x400000 /* SCC2 asserts the SCCd */ #define CICR_SCCDP_SCC3 0x800000 /* SCC3 asserts the SCCd */ #define CICR_SCCDP_SCC4 0xc00000 /* SCC4 asserts the SCCd */ #define CICR_SCCCP_MSK 0x300000 /* SCCc Priority Order Mask */ #define CICR_SCCCP_SCC1 0x000000 /* SCC1 asserts the SCCc */ #define CICR_SCCCP_SCC2 0x100000 /* SCC2 asserts the SCCc */ #define CICR_SCCCP_SCC3 0x200000 /* SCC3 asserts the SCCc */ #define CICR_SCCCP_SCC4 0x300000 /* SCC4 asserts the SCCc */ #define CICR_SCCBP_MSK 0x0c0000 /* SCCb Priority Order Mask */ #define CICR_SCCBP_SCC1 0x000000 /* SCC1 asserts the SCCb */ #define CICR_SCCBP_SCC2 0x040000 /* SCC2 asserts the SCCb */ #define CICR_SCCBP_SCC3 0x080000 /* SCC3 asserts the SCCb */ #define CICR_SCCBP_SCC4 0x0c0000 /* SCC4 asserts the SCCb */ #define CICR_SCCAP_MSK 0x030000 /* SCCa Priority Order Mask */ #define CICR_SCCAP_SCC1 0x000000 /* SCC1 asserts the SCCa */ #define CICR_SCCAP_SCC2 0x010000 /* SCC2 asserts the SCCa */ #define CICR_SCCAP_SCC3 0x020000 /* SCC3 asserts the SCCa */ #define CICR_SCCAP_SCC4 0x030000 /* SCC4 asserts the SCCa */ #define CICR_IRL_MSK 0x00e000 /* Interrupt Request Level */ #define CICR_IRL_LVL0 0x000000 /* Interrupt Request Level 0 */ #define CICR_IRL_LVL1 0x002000 /* Interrupt Request Level 1 */ #define CICR_IRL_LVL2 0x004000 /* Interrupt Request Level 2 */ #define CICR_IRL_LVL3 0x006000 /* Interrupt Request Level 3 */ #define CICR_IRL_LVL4 0x008000 /* Interrupt Request Level 4 */ #define CICR_IRL_LVL5 0x00a000 /* Interrupt Request Level 5 */ #define CICR_IRL_LVL6 0x00c000 /* Interrupt Request Level 6 */ #define CICR_IRL_LVL7 0x00e000 /* Interrupt Request Level 7 */ #define CICR_HP_MSK 0x001f00 /* Highest Priority */ #define CICR_IEN 0x000080 /* Interrupt Enable */ #define CICR_SPS 0x000001 /* Spread Priority Scheme */ /* CPM Interrupt in Service Register (CISR - 0x94c) */ #define CISR_PC15 0x80000000 /* Parallel I/O Port C[15] */ #define CISR_SCC1 0x40000000 /* SCC 1 */ #define CISR_SCC2 0x20000000 /* SCC 2 */ #define CISR_SCC3 0x10000000 /* SCC 3 */ #define CISR_SCC4 0x08000000 /* SCC 4 */ #define CISR_PC14 0x04000000 /* Parallel I/O Port C[14] */ #define CISR_TIMER1 0x02000000 /* Timer 1 */ #define CISR_PC13 0x01000000 /* Parallel I/O Port C[13] */ #define CISR_PC12 0x00800000 /* Parallel I/O Port C[12] */ #define CISR_SDMA 0x00400000 /* SDMA bus error */ #define CISR_IDMA1 0x00200000 /* IDMA 1 */ #define CISR_IDMA2 0x00100000 /* IDMA 2 */ #define CISR_TIMER2 0x00040000 /* Timer 2 */ #define CISR_R_TT 0x00020000 /* Risc Timer Table */ #define CISR_I2C 0x00010000 /* I2C */ #define CISR_PC11 0x00008000 /* Parallel I/O Port C[11] */ #define CISR_PC10 0x00004000 /* Parallel I/O Port C[10] */ #define CISR_TIMER3 0x00001000 /* Timer 3 */ #define CISR_PC9 0x00000800 /* Parallel I/O Port C[9] */ #define CISR_PC8 0x00000400 /* Parallel I/O Port C[8] */ #define CISR_PC7 0x00000200 /* Parallel I/O Port C[7] */ #define CISR_TIMER4 0x00000080 /* Timer 4 */ #define CISR_PC6 0x00000040 /* Parallel I/O Port C[6] */ #define CISR_SPI 0x00000020 /* SPI */ #define CISR_SMC1 0x00000010 /* SMC 1 */ #define CISR_SMC2_PIP 0x00000008 /* SMC 1 */ #define CISR_PC5 0x00000004 /* Parallel I/O Port C[5] */ #define CISR_PC4 0x00000002 /* Parallel I/O Port C[4] */ #define SDMR(base) (UINT8 *) (base + 0x090C) /* SDMA Mask Reg */ /* Buffer Descriptor Pointer Definitions */ /* Receive BD status bits 16-bit value */ #define BD_RX_EMPTY_BIT 0x8000 /* buffer is empty */ #define BD_RX_WRAP_BIT 0x2000 /* last BD in chain */ #define BD_RX_INTERRUPT_BIT 0x1000 /* set interrupt when filled */ #define BD_RX_CON_MODE_BIT 0x0200 /* Continuous Mode bit */ /* Transmit BD status bits 16-bit value */ #define BD_TX_READY_BIT 0x8000 /* Transmit ready/busy bit */ #define BD_TX_WRAP_BIT 0x2000 /* last BD in chain */ #define BD_TX_INTERRUPT_BIT 0x1000 /* set interrupt when emptied */ #define BD_TX_LAST_BIT 0x0800 /* BD contains last char of message */ #define BD_TX_CON_MODE_BIT 0x0200 /* Continuous Mode bit */ /* Baud Rate Generation Register set */ #define BRGC1(base) (UINT32 *) (base + 0x09F0) /* BRG1 Config Reg */ #define BRGC2(base) (UINT32 *) (base + 0x09F4) /* BRG2 Config Reg */ #define BRGC3(base) (UINT32 *) (base + 0x09F8) /* BRG3 Config Reg */ #define BRGC4(base) (UINT32 *) (base + 0x09FC) /* BRG4 Config Reg */ /* settings and pins for port b ...*/ #define PIPC(base) (UINT16 *) (base + 0x0AB2) /* PIP Config Reg */ #define PIPC_MODL_GENIO ((1<<5)|(1<<4)) /* Parallel Interface Port (PIP) PB[24-31]*/ #define PIPC_MODH_GENIO ((1<<3)|(1<<2)) /* Parallel Interface Port (PIP) PB[14-23]*/ #define PBDIR(base) (UINT32 *) (base + 0x0AB8) /* Port B data direction */ #define PBPAR(base) (UINT32 *) (base + 0x0ABC) /* Port B pin assign */ #define PBODR(base) (UINT16 *) (base + 0x0AC2) /* Port B open drain */ #define PBDAT(base) (UINT32 *) (base + 0x0AC4) /* Port B Data Reg */ #define PORT_B_PINS_SPI ((1<<3)|(1<<2)|(1<<1)) /* SPI port B pins (MISO,MOSI,CLK)(28,29,30) */ #define PORT_B_PINS_SPIMISO (1<<3) /* SPIMISO port B pins (28) */ #define PORT_B_PINS_SPIMOSI (1<<2) /* SPIMOSI port B pins (29) */ #define PORT_B_PINS_SPICLK (1<<1) /* SPICLK port B pins (30) */ #define PORT_B_PINS_SPISELin (1<<0) /* SPISELin port B pin (31) */ #define PORT_B_PINS_SPISELout (1<<6) /* SPISELout port B pin (25) */ /* Input/Output Port A register */ #define PADIR(base) (UINT16 *) (base + 0x0950) /* Port A data dir */ #define PAPAR(base) (UINT16 *) (base + 0x0952) /* Port A pin assign*/ #define PAODR(base) (UINT16 *) (base + 0x0954) /* Port A open drain*/ #define PADAT(base) (UINT16 *) (base + 0x0956) /* Port A data reg */ #define PORT_A_PINS_A1 (1<<14) /* mask for signal PA1 @ port a*/ /* MBX DPRAM addresses */ #define INTERNAL_MEM_MAP_ADDR 0xFA200000 #define SPI_BD_OFFSET 0x00002A00 /* DPRAM Bereich 0x2A00 - 0x2EDF auf MBX860 verfuegbar*/ #define DPR_BASE 0x00002000 /* DPRAM-offset zum IMMR*/ /* typedefs */ /* Buffer */ typedef struct /* SPI_BUF */ { UINT16 statusMode; /* status and control */ UINT16 dataLength; /* length of data buffer in bytes */ /* points to data buffer; datapointer uses 32 bits in structure, *datapointer points to an UINT8 field*/ UINT8 * dataPointer; } SPI_BUF; typedef struct spi_drv_bd /* spi_DRV_BD */ { SPI_BUF rxbd[SPI_DRV_NBDS]; /* receive */ SPI_BUF txbd[SPI_DRV_NBDS]; /* transmit */ } SPI_DRV_BD; typedef struct /* SPI_PARAM */ { /* offset description */ UINT16 rbase; /* 00 Rx buffer descriptors base address */ UINT16 tbase; /* 02 Tx buffer descriptors base address */ UINT8 rfcr; /* 04 Rx function code */ UINT8 tfcr; /* 05 Tx function code */ UINT16 mrblr; /* 06 maximum receive buffer length */ UINT32 ris; /* 08 Rx internal state */ UINT32 ridp; /* 0c Rx internal data pointer */ UINT16 rbptr; /* 10 Rx buffer descriptor pointer */ UINT16 ribc; /* 12 Rx internal byte count */ UINT32 rt; /* 14 Rx temp */ UINT32 tis; /* 18 Tx internal state */ UINT32 tidp; /* 1c Tx internal data pointer */ UINT16 tbptr; /* 20 Tx buffer descriptor pointer */ UINT16 tibc; /* 22 Tx internal byte count */ UINT32 tt; /* 24 Tx temp */ } SPI_PARAM;/* 40 Bytes von 48 verfuegbaren Bytes belegt*/ /* locals */ void mbxSpidCcr(); /* execute Spi command */ void spiint(UINT32); /* handle Spi-INT */ --------------D61C5319F371B7105EF19762-- From owner-vxwexplo-process Fri Apr 9 08:43:51 1999 From: Roger Cruz Date: Fri Apr 9 08:43:55 PDT 1999 Subject: RE: Missing interrupts on x86 board.. John, Our problem is not only with the serial ports but with all other interrupts except IRQ0 (timer). What you described, we already encountered about a year ago and solved. Ask to see TSR 101529 (the serial mutex was not made inversion safe). About our current problem, I think we have found the solution. We currently believe that it has to do with a spurious interrupt from the slave PIC not being EOI which prevents any lower priority interrupts from ever being serviced. We're working with WRS to solve it. This is another time where the cost of the emulator has paid for itself. Roger -------- From: johill@lanl.gov (Jeff Hill) We also see this problem here, and ended up adding a watchdog to the BSP supplied COM port serial driver as a workaround for a hang in continuous serial communication with remote Allen Bradley PLC devices that occurred from one to three times a day. Vxworks, vxworks PS: Remove "_nojunk" from email address -- Roger Cruz cruz_nojunk@xyplex.com Xyplex Networks w: 978-952-4783 295 Foster Street f: 978-952-4887 Littleton, MA 01460 From owner-vxwexplo-process Sat Apr 10 04:00:10 1999 From: daemon@csg.lbl.gov Date: Sat Apr 10 04:00:14 PDT 1999 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sat Apr 10 04:00:07 PDT 1999 Subject: Re: question about VxWorks object file size Subject: Re: Porting Web Server to VxWorks Subject: core dump using (nmppc),(ldppc) and (stripppc) on large .a files ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: question about VxWorks object file size Date: 6 Apr 1999 12:34:05 GMT From: chc@nasa2.ksc.nasa.gov (Charles H. Chapman) Organization: NASA, Kennedy Space Center Message-ID: References: <7ebg5m$6jk1@overload.lbl.gov> On Mon, 5 Apr 1999 15:14:37 -0500, Hao, Li (MED) wrote: >I tried to load an object file to my VxWorks target from a mounted NFS >filesystem. When I did "ls -l" on my host machine, it tells me the file >size is 4708217. But on the target, the memory difference before and >after the load according to memShow is 3266648. If it was bigger, it >would have made more sense because loader would try to update symbol >table. > >Does anybody know why? Thanks. I don't know any of the details but there's a lot of "junk" in object files (debugging info, symbol info, etc.) besides just code and data. That stuff doesn't get loaded into memory. Chuck --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Porting Web Server to VxWorks Date: Fri, 9 Apr 1999 14:58:02 -0700 From: "Dave Morse" Organization: Posted via RemarQ Communities, Inc. Message-ID: <923694850.854.24@news.remarQ.com> References: <7d5qal$m203@overload.lbl.gov> You can use Polyhedra as a web server on VxWorks. see www.ployhedra.com Dave M. Bhavin Parikh wrote in message news:7d5qal$m203@overload.lbl.gov... >Hi Everybody, > >Currently I am working hard with windweb server and VxWorks. Actually >WindWeb server doesn't support everything whatever I need like java >servlet and some more. >So now I am looking for other web server, who are available on VxWorks. > >I need your help in following issues. It can save my lots of time. >1. Does anybody has some recommedations and solutions so WindWeb server >can be extended for achieving servlet support? >2. Is there a freeware HTTP server with servlet support already ported >to VxWorks available? >3. I found Apache is really good for servlet support. What is porting >effect required for Apache Web Server porting to VxWorks. > >I would appreciate your reply and help. > >Thanks >Bhavin Parikh >Microsoft Certified Solution Developer, > --------------------------- Newsgroups: comp.os.vxworks Subject: core dump using (nmppc),(ldppc) and (stripppc) on large .a files Date: Fri, 09 Apr 1999 16:10:41 -0700 From: Jim Rucker Organization: Jet Propulsion Laboratory - Pasadena CA Message-ID: <370E88F1.6BE30187@jpl.nasa.gov> I'm compiling a set of libraries that produce a file 'liborbsvcs.a', which is 12 megs. When I perform nmppc, ldppc or strippc these programs core dump. By reducing the side of this file down to 9.5 megs they work fine. If I bring it up to 9.7 megs, these utilities exit with an 'Out of Memory' error. Using the solaris 'nm' tool produces the right output, but I can't use ld since it's not compatable. Are there any fixes for this? Thanks in advance - -Jim Rucker jim.rucker@jpl.nasa.gov --------------------------- End of New-News digest ********************** From owner-vxwexplo-process Mon Apr 12 04:00:15 1999 From: daemon@csg.lbl.gov Date: Mon Apr 12 04:00:19 PDT 1999 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Mon Apr 12 04:00:12 PDT 1999 Subject: rpc server example? Subject: Re: tcl/tk for vxworks shell ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: rpc server example? Date: Thu, 08 Apr 1999 09:27:47 +0200 From: Martin Grossmann Organization: PSI news server Message-ID: <370C5A73.DB12D0B9@psi.ch> anybody has an example for an rpc server on vxWorks (and a client on some unix)? i don't quite manage to use the output from rpcgen on solaris to get an rpc server running on vxWorks. any help greatly appreciated martin - -- ============================================================================== Martin Grossmann Radiation Medicine Paul Scherrer Institute 5232 Villigen PSI tel. (+41) 56 310 36 33 Switzerland fax (+41) 56 310 35 15 Internet: Martin.Grossmann@psi.ch X.400: WWW: http://www1.psi.ch/~grossmann ============================================================================== --------------------------- Newsgroups: comp.lang.tcl,comp.os.vxworks Subject: Re: tcl/tk for vxworks shell Date: Wed, 07 Apr 1999 22:11:19 -0400 From: Zach Frey Organization: BrightNet Ohio Message-ID: <370C1046.30F7A482@bright.net> References: <3703988F.CD81D051@huey.jpl.nasa.gov> Reply-To: zfrey@bright.net Leonard J. Reder wrote: > Does any one know of a TCL/TK implementation which > can be loaded onto VxWorks? If you're using a more recent version of VxWorks (one that includes the Tornado(tm) development environment), Tcl is included as a target shell. For older versions of VxWorks (5.2 and earlier), several versions of Tcl have been ported, and should be available at the VxWorks FTP archive (whose URL I've misplaced, sorry). I don't know if there has been a successful port of Tcl 8.0 or later to VxWorks; all of the ports I've seen were of Tcl 7.[3-6]. I am not aware of a port of Tk to VxWorks. Good luck, Zach --------------------------- End of New-News digest ********************** From owner-vxwexplo-process Mon Apr 12 11:29:36 1999 From: David Laight Date: Mon Apr 12 11:29:39 PDT 1999 Subject: PPP over dialup links I've just succeeded in getting PPP to connect over a dialup link. However I don't know what I should do when the 'phone disconnects (ie carrier detect goes inactive). pppDelete() will try to send LCP stuff to the remote system - which isn't there. I guess it will time out someday, but surely there must be a way to tell PPP the link is dead (nailed to its perch...). Any thoughts? Any thoughts as how to get vxworks to get a PPP link established in response to a TCP socket connect() (or even sendto()) call? Or do we have to put an explicit request in the app all the time? David ---------------------------------------------------------------- David Laight email: dsl@tadpole.co.uk Tadpole Technology plc phone: +44 1223 278 256 Cambridge, UK fax: +44 1223 278 201 From owner-vxwexplo-process Tue Apr 13 05:33:36 1999 From: stevens Date: Tue Apr 13 05:33:40 PDT 1999 Subject: FTP-Server for vxworks Hi VxWorkers ! I want to get access to a VxWorks-Target-System via FTP. A FTP-Server is installed, but I don´t know the (standard?) username and password. Torsten -- ************************************************************************** fraunhofer institute of microelectronic circuits and systems department sat name torsten stevens mailto:stevens@ims.fhg.de address finkenstrasse 61 phone +0049/203/3783-238 47057 duisburg fax +0049/203/3783-266 germany http://www.fhg.de/depts/ims-du-e.html ************************************************************************** From owner-vxwexplo-process Tue Apr 13 07:23:52 1999 From: "Platz, Allen" Date: Tue Apr 13 07:23:56 PDT 1999 Subject: vxWorks and NFS File rename Has anybody found a work around for the vxWorks dos file "rename" function not working on a NFS volume ? I checked the WindRiver support web pages and the nfsDrv and nfsLib functions with no luck (ioctl rename function not supported in NFS). Is there a third party software package that might support this ? Thanks Allen Platz Senior Engineering Specialist LORD Corp 800-458-0581 ext 6741 allen_platz@lord.com From owner-vxwexplo-process Tue Apr 13 07:32:39 1999 From: Reyhan ERGUN Date: Tue Apr 13 07:32:43 PDT 1999 Subject: using EPPCBug This is a multi-part message in MIME format. --------------9046091473E1C2C916E8592F Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit I use MBX860-002 board and Tornado. I have some questions about EPPCBug (The PowerPC debugger) which is contained in both the socketed "Boot ROM" device and the four soldered 32-pin PLCC Flash devices on the MBX860 board. Using EPPCBug, you can flash bootrom.bin into the soldered Flash device. I expect EPPCBug is placed at locations ($FC000000) to ($FC000100), since EPPCBug firmware takes up 512 KB (128K words) of space and I flash bootrom.bin starting from ($FC000100). Q1.) "EPPC-Bug>plash 4000:83fff fc000100" command erases block# 0000 ($FC000000). Does it leave ($FC000000) to ($FC000100) unerased? Q2.) At that moment, I use EPPCBug when the selected device is socketed "Boot ROM" by installing corresponding jumper (J4) across pins 1 and 2. How can I use EPPCBug, when the current selected device is soldered Flash? Q3.) In the manuals for MBX860, it is written that various EPPCBug routines are available to user programs through the system call handler, how? I appreciate your interest, thanks Warmest regards, Reyhan --------------9046091473E1C2C916E8592F Content-Type: text/x-vcard; charset=us-ascii; name="vcard.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Reyhan ERGUN Content-Disposition: attachment; filename="vcard.vcf" begin: vcard fn: Reyhan ERGUN n: ;Reyhan ERGUN org: ASELSAN INC. adr: Microwave and System Technologies Division ;;Software Engineering Dept.;ANKARA;;;TURKEY email;internet: findik@mst.aselsan.com.tr tel;work: (90-312) 385 19 00 (Ext. 2668) tel;fax: (90-312) 354 52 05 note: http://www.aselsan.com.tr x-mozilla-cpt: ;0 x-mozilla-html: FALSE version: 2.1 end: vcard --------------9046091473E1C2C916E8592F-- From owner-vxwexplo-process Tue Apr 13 09:10:00 1999 From: Bill Brown Date: Tue Apr 13 09:10:03 PDT 1999 Subject: Re: vxWorks and NFS File rename the vxWorks Users Group Exploder wrote: > > Submitted-by owner-vxwexplo-process Tue Apr 13 07:23:52 1999 > Submitted-by: "Platz, Allen" > > Has anybody found a work around for the vxWorks dos file "rename" function > not working on a NFS volume ? I checked the WindRiver support web pages and > the nfsDrv and nfsLib functions with no luck (ioctl rename function not > supported in NFS). Is there a third party software package that might > support this ? > Thanks > Allen Platz > Senior Engineering Specialist > LORD Corp > 800-458-0581 ext 6741 > allen_platz@lord.com Years ago when faced with this problem, in principle I just did copy followed by rm Of course, I checked the error status before doing the latter operation..... Don't remember the details, but I did find a workable solution "down this path." Disclaimer: Any opinions are my own and have | -bill nothing to do with the official policy or the | wlbrown@lbl.gov management of L.B.N.L, who probably couldn't | Berkeley, CA care less about employees who play with trains. | aka wlbrown@netcom.com From owner-vxwexplo-process Wed Apr 14 01:35:16 1999 From: "Moy, Ian" Date: Wed Apr 14 01:35:21 PDT 1999 Subject: if_fei and sys557Init. > Hi, > > I am running VxWorks on a PowerPC, MBX860 and would like to add a Fast > Ethernet module. > > The ethernet module is based on the Intel 82558 ethernet controller, is > for the pc/104-Plus standard, made by E.E.P.D and is called Profive > ethernet-100. > > If anyone has experience with this module, is it any good? Does anybody > know of any more pc/104-Plus ethernet modules? > > Will if_fei driver this device? > > What has to be in the sys557Init function for the above setup? Is the > function specific to the main board, the ethernet module or both? > > Any help or advice is welcome, > > Ian. > > -------------------------------------------------- > Ian Moy, Roke Manor Research Ltd > (email) ian.moy@roke.co.uk > -------------------------------------------------- > > From owner-vxwexplo-process Wed Apr 14 04:00:56 1999 From: daemon@csg.lbl.gov Date: Wed Apr 14 04:01:00 PDT 1999 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Wed Apr 14 04:00:52 PDT 1999 Subject: INCLUDE_NET_INIT seems undefined even I defined it!!! Subject: Re: SRAM checking Subject: C++ exceptions / STL Subject: Re: SRAM checking Subject: Re: FTP-Server for vxworks Subject: Page Fault Message Subject: Re: SRAM checking Subject: mvme2604 and serial i/o Subject: Re: SRAM checking Subject: Re: Shell scripts - Windows NT problem? Subject: Re: Breakpoints ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: INCLUDE_NET_INIT seems undefined even I defined it!!! Date: Tue, 13 Apr 1999 15:35:52 +0300 From: Bahri Okuroglu Organization: Netas - Northern Electric Telecom AS Message-ID: <37133A27.132DF355@rnd.netas.com.tr> Hi all, My target is pc486 and i boot from a networked image. Even this means vxworks can initialize and access network, I see that network is not initialized when I check just after the boot. INCLUDE_NETWORK and INCLUDE_NET_INIT are also defined (using Tornado BSP Configuration Tool). But INCLUDE_NET_INIT seems undefined in usrConfig.c (where the network initialization routine is called). Waiting comments from VxWorks gurus... Cheers, - -- ====================================================== Bahri OKUROGLU Software Design Engineer Netas R&D RT6 mailto:bahrio@rnd.netas.com.tr http://www.netas.com.tr mailto:bahrio@yahoo.com http://193.140.184.5/r748/member/bahrio (internal) Netas Alemdag Cad. Umraniye 81244 ISTANBUL TURKEY ====================================================== --------------------------- Newsgroups: comp.arch.embedded,comp.realtime,comp.os.vxworks,comp.os.psos,comp.os.qnx,comp.sys.powerpc,comp.sys.m68k,comp.os.chorus,comp.os.linux,comp.os.os9 Subject: Re: SRAM checking Date: Tue, 13 Apr 1999 13:13:02 GMT From: aerosoft@agt.net Organization: Aerospace Software Ltd. Message-ID: <3713506A.6774@agt.net> References: <371046BC.5DC63B7C@chromatis.com> <3712DDE0.2935@nowhere.tennyson.com.au> <7eusib$d2f$1@thelma.netpower.no> <371301cc.0@seat> Reply-To: aerosoft@agt.net Hi there, A good way to trap addressing problems, is to fill the memory with a known pattern and then calculate a checksum on it, which should then match the known checksum. Cheers, Herman http://www.agt.net/public/aerosoft Liam Friel wrote: > > David Brown wrote in message <7eusib$d2f$1@thelma.netpower.no>... > >I am afraid "quick-and-dirty" is a good description of Tom's algorithm - it > >is quick, but a memory test algorithm must test for independance of the > data > >pins. > > Depending on your situation, another very useful thing is to use a pattern > that repeats every 3 locations, not every single location. > So write 0x55, 0xAA, 0xFF, 0x55, 0xAA, 0xFF and so on (or whatever complex > pattern you want). > > That way, if you have address bus errors (ie you think you have 128K of > memory but because of a layout or wiring error you can only address 64K) > your test will catch it. Otherwise, you'll think everything is OK when in > fact it ain't. > > As an earlier poster said, errors in actual memory chips are rare (unless > you zapped the chip yourself). > But if you're making your own board, layout errors are much more common ... > > Liam > > ------------------------------------------------------------ > Liam Friel || email: liamf@s3two.ie > Software Division || > Silicon & Software Systems || tel: +353-1-218 5532 > Newstead, Clonskeagh || fax: +353-1-218 5590 > Dublin 14, Ireland || www: http://www.s3group.com > ------------------------------------------------------------ --------------------------- Newsgroups: comp.os.vxworks Subject: C++ exceptions / STL Date: Tue, 13 Apr 1999 14:58:21 +0100 From: "Paul Crowley" Message-ID: <37134d3c.0@nnrp1.news.uk.psi.net> I've done a few tests to determine whether VxWorks supports C++ exceptions. It appears that it doesn't but the GCC does (the test code compiles but produces link errors related to unwinding the stack when loaded on the target). Is my logic correct? If exceptions are not supported is it still possible to use STLPort (ie. can it be configured not to generate exceptions)? Thanks. --------------------------- Newsgroups: comp.arch.embedded,comp.realtime,comp.os.vxworks,comp.os.psos,comp.os.qnx,comp.sys.powerpc,comp.sys.m68k,comp.os.chorus,comp.os.linux,comp.os.os9 Subject: Re: SRAM checking Date: Tue, 13 Apr 1999 13:10:37 -0400 From: Matthew Donadio Organization: Image & Signal Processing, Inc. Message-ID: <37137A8D.BD1A1DA@isptechinc.com> References: <371046BC.5DC63B7C@chromatis.com> <3712DDE0.2935@nowhere.tennyson.com.au> <7eusib$d2f$1@thelma.netpower.no> David Brown wrote: > It is worth noting that errors in RAM chips are extremly rare - for most > systems, the danger is production errors (short-circuits or poorly soldered > pins). This is true once the manufacturing process has been nailed down. Early on, hot spots can damage chips. Static blows can also damage chips, and unfortunately, these can be hard to find. > ... Check for data pin independance using a 0x01, 0x02, .. 0x80 > sequence. Check for address pin independance by writing to address 0x0000, > ... A little explanation is worthwhile here. When you are routing the boards, you have the option of saying that certain signals can be interchanged. For example, you may end up with D7 of a bus hooked up to D1 on a RAM. The same can happen with address lines. It makes dubugging tricky, but it may be the routes shorter. Along with the walking-ones, you should also test use walking zeros. I have also seen an instance where some bad solder joints were only found by filling the RAM with pseudo random data and then verifying it. Depending on the situation, the ramtest should also be designed to give an address/expected/read report (unless you just need pass or fail). It's easy to touch up the solder on an entire part, but you need the full report to track down shorts, especially whacky ones like data lines shorted to address lines and data/address shorted to control signals or clocks. - -- Matt Donadio (donadio@isptechinc.com) | 43 Leopard Rd, Suite 102 Sr. Software Engineer | Paoli, PA 19301-1552 Image & Signal Processing, Inc. | Phone: +1 610 407 4391 http://www.isptechinc.com | FAX: +1 610 407 4405 --------------------------- Newsgroups: comp.os.vxworks Subject: Re: FTP-Server for vxworks Date: Tue, 13 Apr 1999 10:05:03 -0700 From: "k.york" Organization: Cisco Systems, Inc. Message-ID: <3713793F.19692D47@cisco.com> References: <37134740.661D361F@ims.fhg.de> Greetings, stevens wrote: > > Hi VxWorkers ! > I want to get access to a VxWorks-Target-System via FTP. A FTP-Server is installed, but I don´t know > the (standard?) username and password. > > Torsten > > -- Look up ``loginLib'' in the manuals. You need to add the user/password yourself, there is no default. The doc's could do a better job explaining this, like maybe a link in the ftp section to the loginLib - --kyle --------------------------- Newsgroups: comp.os.vxworks Subject: Page Fault Message Date: Tue, 13 Apr 1999 11:58:17 -0700 From: "Glenn Dicus" Organization: Nomadix LLC Message-ID: <7f03mn$2ov$1@la-mail4.digilink.net> Does anybody know the exact meaning of the page fault message on an x86 platform. For instance, the message includes the Status Register, but I am unsure as to what register this maps to on the K6 processor. Likewise, I do not know the Error Code protocol as well. Any pointers would be appreciated. thanks glenn dicus Nomadix Inc --------------------------- Newsgroups: comp.arch.embedded,comp.realtime,comp.os.vxworks,comp.os.psos Subject: Re: SRAM checking Date: Tue, 13 Apr 1999 21:41:09 GMT From: jimf@impaq.com.au (Jim (from Oz)) Organization: To a limited extent Message-ID: <3713c7bc.1894322@news0.optus.net.au> References: <371046BC.5DC63B7C@chromatis.com> <3712DDE0.2935@nowhere.tennyson.com.au> <7eusib$d2f$1@thelma.netpower.no> <371301cc.0@seat> <3713506A.6774@agt.net> <7evlo2$1t8$1@nnrp1.dejanews.com> Reply-To: jimf@impaq.com.au On Tue, 13 Apr 1999 14:52:54 GMT, Robert Gush wrote: >In article <3713506A.6774@agt.net>, > aerosoft@agt.net wrote: >> Hi there, >> >> A good way to trap addressing problems, is to fill the memory with a >> known pattern and then calculate a checksum on it, which should then >> match the known checksum. >The most solid method I have used so far is: > 1. Fill memory with 55h, verifying as you go and go to step 2 on error > or if limit is reached > 2. Reset pointer to start address > 3. Check that contents still OK, if not end of mem reached > 4. Write 0AAh to location, if verify bad then end of mem > 5. Write 00h to location and increment the pointer and loop to step 3. > 6. End Of Memory - pointer is 1 count past valid memory > >It's a bit slow, but it is pretty thorough. Maybe other patterns could be >used to take into account the layout of the data lines, but in many >applications the data bus is also used for program execution and it often >doesn't run when there's a databus short. This was developed by my good mate Pete, we've both used it for years and it really seems to catch most faults, are you there Pete? Flow chart for Pete's walking bit memory test Start Init. pattern to 01h ---------- Fill memory with pattern | | Point to start of memory /|\ | ------ Test (pointer) for valid pattern | | | | Complement (pointer) | /|\ | | Increment pointer | | No | ------ End of memory ? /|\ | Shift pattern left 1 bit | No ---------- Carry set ? End Jim (from Oz) --------------------------- Newsgroups: comp.os.vxworks Subject: mvme2604 and serial i/o Date: Tue, 13 Apr 1999 18:00:49 -0500 From: "Frank Martinez" Organization: AT&T WorldNet Services Message-ID: <7f0iia$l09$1@bgtnsc02.worldnet.att.net> Hi, I'm new to vxWorks and I have a problem. I am trying to create a user prompt for an application running on my 2604 board using the serial i/o. If I try to use scanf() or getchar() my program locks up and I have to escape or reboot to regain control. Any ideas on what I may be doing wrong? thanks in advance. --------------------------- Newsgroups: comp.arch.embedded,comp.realtime,comp.os.vxworks,comp.os.psos Subject: Re: SRAM checking Date: Tue, 13 Apr 1999 20:28:22 -0700 From: Ken Message-ID: <37140B56.F72BDBE9@well.com> References: <371046BC.5DC63B7C@chromatis.com> <3712DDE0.2935@nowhere.tennyson.com.au> <7eusib$d2f$1@thelma.netpower.no> <371301cc.0@seat> <3713506A.6774@agt.net> <7evlo2$1t8$1@nnrp1.dejanews.com> <3713c7bc.1894322@news0.optus.net.au> Note that filling memory with a single value and then testing it only finds stuck data bits. It doesn't find stuck address bits. To find stuck address bits, you need a pattern that varies with address. The simplest test is to fill each location with its address, then go back and make sure that pattern stuck. Follow by testing with the complement of the address. - -- Ken mailto:shiva@well.com http://www.well.com/user/shiva/ --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Shell scripts - Windows NT problem? Date: Sat, 10 Apr 1999 17:43:30 -0700 From: Jerry Hudgins Organization: Western Digital Design, Marin County, California Message-ID: <370FF032.B6937028@e-farm.com> References: <370DD078.393C@gecm.com> David Rigg wrote: > > If I connect a host shell to my card and type '< loadall.scp' > all the files get loaded happily > > However, If I try it with the target shell, I get: > > .ld < fooa.o > .Error: File C:\TORNADO\TARGET\CONFIG\MV2304\fooa.o does not exist > .'an't open input 'fooa.o > . errno = 0x226 > > and it falls over at the first file. > BUT > > I can immediately type ld < fooa.o at the shell prompt I just tried > the above from and it loads happily!!! I've run into this problem in startup scripts loaded from an NT boot host. It appears that the line terminator is being included in the file name, leading to both the load failure and the garbled error message. The equivalent command: ld 1,0,"fooa.o" works fine for me under these circumstances - apparently the quoted file name prevents picking up the line terminator. Give that a try. I've never seen the problem on Solaris, BTW, just on NT. - -- Jerry Hudgins (jerry@e-farm.com) --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Breakpoints Date: Wed, 14 Apr 1999 10:27:26 GMT From: brunella@txt.it (Mauro Brunella) Organization: I.NET S.p.A Message-ID: <7f1qlb$1354$1@urano.inet.it> References: <3703C994.163668FA@tekelec.com> Kevin, there two kinds of breakpoints in Tornado: - - Global breakpoints, managed via shell; - - Per-task breaks, managed by GDB (crosswind). Global breakpoints is what you are asking for, some that stops every tasks that executes a code. If you read the shell documentation on Tornado user's manual you will find the commands to be used. For example: - - b
or b , create a breakpoint in the specified address. Breakpoint is global, as you want. I can issue a breakpoint for the taskEntry of your application-spawned task to debug it from the beginning. - - bd
, destroy a existing break. - - bdall , destroy every breaks. - - b , without parameters shows the breakpoints present in the system. Note that the GDB-breakpoints are listed, too. They are active only for the task attached in GDB at the moment. I think GDB uses the same mechanism we can use in the shell. - - c , continues ( like in GDB) the last breaked task. Don't use tr (taskResume) because the task will break on the same breakpoint. I'm using Tornado on Windows NT now, but I used it on unix (HP-UX and sun-os). I'm sure that "tr" does not restart a breaked task on NT, but It seems to me that on unix it did. When a task breaks on a breakpoint its state changes to SUSPEND and you can attach it with GDB (if you have debug information for that piece of code GDB will open the source immediatly). "Shell" breakpoints can be created refering to a functionName. This function must be visible by shell (e.g. not "static" in your source code). You can issue a breakpoint on a function if you can "see" it with lkup "functionName". Hope this could help, good work. Mauro P.S.= Sorry for my bad English. In article <3703C994.163668FA@tekelec.com>, Kevin Sapp wrote: >Hello, > >The documentation implies that "global" breakpoints are the >default and thread breakpoints the thread has to be specified. > >What I have found is that if I attach to a task and set the >break point and then continue and the function is called in the >"attached" threads context the break is caught. If the break >occurs in another context the break is missed ? > >When a file is loaded via tornado and a break is set, then execution >starts if the break is in a different task (a new one created by >sp, ...) the break point is not detected. > >What am I doing incorrectly ? > >I just want to sent an unconditional break point that breaks NO MATTER >WHO CALLS IT FROM WHAT THREAD. > >Thanks for the help >Kevin --------------------------- End of New-News digest ********************** From owner-vxwexplo-process Wed Apr 14 06:33:58 1999 From: Hakan Karakas Date: Wed Apr 14 06:34:02 PDT 1999 Subject: Re: FTP-server for vxworks Hi Torsten, FTP user name and password is set on boot time via boot parameters. You can see your current boot parameters by typing the following on the shell. > bootChange Hakan Karakas h.karakas@ieee.org --------------------------------------------------------------------------------------------- Submitted-by owner-vxwexplo-process Tue Apr 13 05:33:36 1999 Submitted-by: stevens Hi VxWorkers ! I want to get access to a VxWorks-Target-System via FTP. A FTP-Server is installed, but I don´t know the (standard?) username and password. Torsten From owner-vxwexplo-process Wed Apr 14 06:45:08 1999 From: "Cui Jian" Date: Wed Apr 14 06:45:13 PDT 1999 Subject: wtx Error? Dear vxworks guru: I encounted a serious problem about Tornado's wdb agent. I don't know why this happened, if you have any comments please tell me. When I started vxworks 5.0.3, I could use "i" to show a summary of tasks information in windsh. But after I download my own programs and spawn the tasks, I couldn't use "i" to show the information, and in crosswind I couldn't attach the tasks. The error showed "WTX_AGENT_GOPHER_FAULT", but I could see the tasks in browser. My programs is rather large-about 2MB in all. I had worked on the programs for a year, and I have debugged it successfully, but something unknown made tornado stop working several days ago. Who can tell me why? From owner-vxwexplo-process Wed Apr 14 08:07:30 1999 From: "Blake, Don" Date: Wed Apr 14 08:07:35 PDT 1999 Subject: RE: VxWorks SRAM checking An addressing test that I've used quite often is to fill the area of memory with a pseudo-random pattern then read it back and verify each location. I don't know where I got this algorithm - maybe I invented it! I don't like to use a checksum for the following reasons: A checksum provides little, if any, diagnostic information A simple checksum (e.g., 32-bit 2's complement addition) doesn't necessarily provide good fault detection Here's some sample code I've used for generating a pseudo-random patterns. The NEXT_DATA_PATTERN and PREVIOUS_DATA_PATTERN macros allow for tricks like writing memory from low to high address then reading it back from high to low address. /* macros for pseudo-random pattern generation */ #define POLYNOMIAL 0xEDB88320 /* 32-bit polynomial for pseudo random data pattern generation */ #define NEXT_DATA_PATTERN(d) \ if ( d & 0x00000001 ) { \ d = ( d >> 1 ) ^ POLYNOMIAL; \ } else { \ d >>= 1; \ } /* end if */ #define PREVIOUS_DATA_PATTERN(d) \ if ( d & 0x80000000 ) { \ d = ( ( d ^ POLYNOMIAL ) << 1 ) | 0x00000001; \ } else { \ d <<= 1; \ } /* end if */ /* sample loop to write pseudo-random data pattern in PMC SRAM - replace SRAM_WRITE macro call with whatever is appropriate in your application */ expSramData = RANDOM_PATTERN_3; for ( address = 0; address < FCPMC_SRAM_SIZE; address += 4 ) { SRAM_WRITE( address, expSramData ); NEXT_DATA_PATTERN( expSramData ) } /* end for */ Don Blake Sr. Programmer Lockheed Martin Federal Systems Owego, NY -----Original Message----- Hi there, A good way to trap addressing problems, is to fill the memory with a known pattern and then calculate a checksum on it, which should then match the known checksum. Cheers, Herman http://www.agt.net/public/aerosoft Liam Friel wrote: > > David Brown wrote in message <7eusib$d2f$1@thelma.netpower.no>... > >I am afraid "quick-and-dirty" is a good description of Tom's algorithm - it > >is quick, but a memory test algorithm must test for independance of the > data > >pins. > > Depending on your situation, another very useful thing is to use a pattern > that repeats every 3 locations, not every single location. > So write 0x55, 0xAA, 0xFF, 0x55, 0xAA, 0xFF and so on (or whatever complex > pattern you want). > > That way, if you have address bus errors (ie you think you have 128K of > memory but because of a layout or wiring error you can only address 64K) > your test will catch it. Otherwise, you'll think everything is OK when in > fact it ain't. > > As an earlier poster said, errors in actual memory chips are rare (unless > you zapped the chip yourself). > But if you're making your own board, layout errors are much more common ... > > Liam > > ------------------------------------------------------------ > Liam Friel || email: liamf@s3two.ie > Software Division || > Silicon & Software Systems || tel: +353-1-218 5532 > Newstead, Clonskeagh || fax: +353-1-218 5590 > Dublin 14, Ireland || www: http://www.s3group.com > ------------------------------------------------------------ From owner-vxwexplo-process Wed Apr 14 10:30:35 1999 From: Larry Sinn Date: Wed Apr 14 10:30:39 PDT 1999 Subject: Re: wtx Error? the vxWorks Users Group Exploder wrote: > Submitted-by owner-vxwexplo-process Wed Apr 14 06:45:08 1999 > Submitted-by: "Cui Jian" > > Dear vxworks guru: > I encounted a serious problem about Tornado's wdb agent. I don't know > why this happened, if you have any comments please tell me. > When I started vxworks 5.0.3, I could use "i" to show a summary of tasks > information in windsh. But after I download my own programs and spawn the > tasks, I couldn't use "i" to show the information, and in crosswind I > couldn't attach the tasks. The error showed "WTX_AGENT_GOPHER_FAULT", but I > could see the tasks in browser. My programs is rather large-about 2MB in > all. I had worked on the programs for a year, and I have debugged it > successfully, but something unknown made tornado stop working several days > ago. Who can tell me why? > This happens to me all the time. The fault is not in vxWorks but in my program clobbering something it shouldn't have. Can you run your program of several day ago? After you get something running successfully, add in the new stuff a piece at a time until things break. Larry. Larry Sinn Spectral Dynamics (408)474-1746 voice 1983 Concourse Drive (408)474-1780 fax San Jose, Ca. 95131-1708 sinnl@sd-star.com From owner-vxwexplo-process Wed Apr 14 11:11:09 1999 From: Michael Morrison Date: Wed Apr 14 11:11:12 PDT 1999 Subject: Re: wtx Error? the vxWorks Users Group Exploder wrote: > > Submitted-by owner-vxwexplo-process Wed Apr 14 06:45:08 1999 > Submitted-by: "Cui Jian" > > Dear vxworks guru: > I encounted a serious problem about Tornado's wdb agent. I don't know > why this happened, if you have any comments please tell me. > When I started vxworks 5.0.3, I could use "i" to show a summary of tasks > information in windsh. But after I download my own programs and spawn the > tasks, I couldn't use "i" to show the information, and in crosswind I > couldn't attach the tasks. The error showed "WTX_AGENT_GOPHER_FAULT", but I > could see the tasks in browser. My programs is rather large-about 2MB in > all. I had worked on the programs for a year, and I have debugged it > successfully, but something unknown made tornado stop working several days > ago. Who can tell me why? Check your stacks for overflow, or for a buffer overflow somewhere in your program. Sounds like your program is out stomping where it shouldn't -- Michael Morrison, Sierra Design Labs http://www.sdlabs.com PGP Key: http://pgp5.ai.mit.edu:11371/pks/lookup?op=get&search=0x7FB0919F From owner-vxwexplo-process Wed Apr 14 15:04:01 1999 From: Kim Middleton Date: Wed Apr 14 15:04:05 PDT 1999 Subject: VxWorks: cksum: out of data error Hello! We recently upgraded our BSP for the mv2303 board to work with the SENS drivers and have noticed that we're seeing a recurring error: cksum: out of data This seems to be benign (can't see anything wrong), but I'm wondering if anyone else has seen this and/or knows what the cause is. TIA! Kim -- Kim Middleton Hewlett-Packard Company kimberly@spk.hp.com 24001 E. Mission Ave Voice: (509) 921-3817 Liberty Lake, WA 99019 FAX: (509) 921-3991 From owner-vxwexplo-process Wed Apr 14 15:33:33 1999 From: Charlie Grames Date: Wed Apr 14 15:33:37 PDT 1999 Subject: Re: Shell scripts - Windows NT problem? The reason this problem occurs on NT and not on UNIX is that VxWorks' notion of end-of-line in an ASCII file is the same as that of UNIX--a single line feed character--whereas NT uses a CRLF combination. Most of the time, the extra carriage return does not cause any problem. However, in cases where strings are being parsed to end-of-line (such as when using redirection), the extra carriage return will be included in the string. This causes problems not only with redirection but with any algorithm that assumes that everything preceding the line feed is valid. To compensate in the general case, either the algorithm must be modified, or the offending carriage return characters must be stripped out of the file. Charlie Grames The Boeing Company (314) 233-1956 Charles.R.Grames@boeing.com Newsgroups: comp.os.vxworks Subject: Re: Shell scripts - Windows NT problem? Date: Sat, 10 Apr 1999 17:43:30 -0700 From: Jerry Hudgins Organization: Western Digital Design, Marin County, California Message-ID: <370FF032.B6937028@e-farm.com> References: <370DD078.393C@gecm.com> David Rigg wrote: > > If I connect a host shell to my card and type '< loadall.scp' > all the files get loaded happily > > However, If I try it with the target shell, I get: > > .ld < fooa.o > .Error: File C:\TORNADO\TARGET\CONFIG\MV2304\fooa.o does not exist > .'an't open input 'fooa.o > . errno = 0x226 > > and it falls over at the first file. > BUT > > I can immediately type ld < fooa.o at the shell prompt I just tried > the above from and it loads happily!!! I've run into this problem in startup scripts loaded from an NT boot host. It appears that the line terminator is being included in the file name, leading to both the load failure and the garbled error message. The equivalent command: ld 1,0,"fooa.o" works fine for me under these circumstances - apparently the quoted file name prevents picking up the line terminator. Give that a try. I've never seen the problem on Solaris, BTW, just on NT. - -- Jerry Hudgins (jerry@e-farm.com) From owner-vxwexplo-process Thu Apr 15 04:00:29 1999 From: daemon@csg.lbl.gov Date: Thu Apr 15 04:00:33 PDT 1999 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Thu Apr 15 04:00:26 PDT 1999 Subject: Re: Simple Database Software For VxWorks Subject: Re: SRAM checking Subject: renaming tasks Subject: Re: VxWorks: cksum: out of data error Subject: Position Available, Santa Clara CA S/W Engineer, RTOS ------------------------------------------------------- Newsgroups: comp.os.vxworks,comp.arch.embedded Subject: Re: Simple Database Software For VxWorks Date: Fri, 9 Apr 1999 15:03:53 -0700 From: "Dave Morse" Organization: Posted via RemarQ Communities, Inc. Message-ID: <923695202.596.86@news.remarQ.com> References: <36F8F9B5.287DEB53@kronos.com> You can use Polyhedra. The unbundeled database engine (without all the features FT, Historian, etc.) is available for $1495 and supports flash disk, fixed/var records, microsecond access/update/insert, etc. performance Many customers use just the basic SQL/relational model/features. See www.polyhedra.com Dave M. Dave Leney wrote in message news:36F8F9B5.287DEB53@kronos.com... >We are looking for a simple, flash-friendly, inexpensive database or >other type of indexed storage software which is available for VxWorks. >Our project is using an MPC860T with an M-Systems disk-on-a-chip flash >device for persistant storage. The VxWorks dosFS will be thr primary >interface to the flash device. > >We need to store both fixed and variable length data records in flash, >allowing for fast retrieval and for updating which includes insertion, >replacement and deletion. However, we don't want soething which will >cause un-necessary wear on the flash device. > >Thanks in advance for any responses > --------------------------- Newsgroups: comp.arch,embedded,comp.realtime,comp.os.vxworks,comp.os.psos,comp.os.qnx,comp.sys.powerpc,comp.sys.m68k,comp.os.chorus,comp.os.linux,comp.os.os9 Subject: Re: SRAM checking Date: Tue, 13 Apr 1999 16:02:08 +1000 From: Tom Evans Organization: OzEmail Ltd. Message-ID: <3712DDE0.2935@nowhere.tennyson.com.au> References: <371046BC.5DC63B7C@chromatis.com> Rami Gideoni wrote: > Looking for patterns and algorithms to check DRAM (0x55, 0xAA ...) My quick-and-dirty favourite: 1. For all memory, fill with zeros, as fast as you can make it run, 2. For all memory: 2a. Read a byte/word/long/ultra-long/whatever 2b. Decrement the value you just read, 2c. Write it back, 2d. If you didn't get a "borrow" from the adder you've got an error, so print address/read/expected. 3. For all memory: 3a. Read a byte/word/long/ultra-long/whatever 3b. Increment the value you just read, 3c. Write it back, 3d. If you didn't get a "carry" from the adder you've got an error, so print address/read/expected. 4. For all memory: 4a. Read and verify all-zeros. 5. Now disable the (usually) slow memory-clear loop in your initialisation code because the memory test has already done it for you. The above has been implemented in simple, dumb hardware. It is pretty quick (and usually good enough) in software too. If you're on a modern processor you should pipeline it. It checks all locations and all address lines, but not shorted data lines - those you check separately (but you don't need to do this for the whole of memory). Tom Evans InitialSurname@somewhere --------------------------- Newsgroups: comp.os.vxworks Subject: renaming tasks Date: Wed, 14 Apr 1999 10:43:06 +0200 From: "Me" Organization: Philips Medical Systems BV XRD-Best Message-ID: <924079086.540693@z289> Hi All, Is it possible to rename a task ascii name (on the fly)? Regards, Joost Kramer Philips Medical Systems --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks: cksum: out of data error Date: Wed, 14 Apr 1999 16:21:19 -0700 From: Stephen Macmanus Organization: Wind River Systems Message-ID: <371522EF.4ED1@wrs.com> References: <7f3416$fmb11@overload.lbl.gov> Kim Middleton wrote: > > Hello! > > We recently upgraded our BSP for the mv2303 board to > work with the SENS drivers and have noticed that we're > seeing a recurring error: > > cksum: out of data > > This seems to be benign (can't see anything wrong), but > I'm wondering if anyone else has seen this and/or knows > what the cause is. > > TIA! > The checksum calculation routine takes two parameters: an MBlk chain and a length. The message indicates that the given length value exceeds the total amount of data reported in the individual length values for each piece of the Mblk chain. In addition to the mBlkPktHdr.len field, which is set for the first block in the chain to the total data size, make sure that the mBlkHdr.mLen field is set to the size of each fragment for the first block and any subsequent blocks. Of course, these values will be equal for the case of a single-element chain. Stephen > Kim Middleton > Hewlett-Packard Company kimberly@spk.hp.com > 24001 E. Mission Ave Voice: (509) 921-3817 > Liberty Lake, WA 99019 FAX: (509) 921-3991 - -- - ------------------ Stephen Macmanus #include Software Engineer Wind River Systems --------------------------- Newsgroups: comp.os.vxworks Subject: Position Available, Santa Clara CA S/W Engineer, RTOS Date: 12 Apr 1999 03:41:34 GMT From: twhit57023@aol.com (TWhit57023) Organization: AOL http://www.aol.com Message-ID: <19990411234134.06277.00001810@ng01.aol.com> OVERVIEW: Work with a leader of embedded communications software; pre-ipo start-up slated to go public in 8-12 months; strong management team from major companies. Seeking talented engineers with good skills or ability to interface with key major accounts. Compensation includes Salary, Stock Options, and Full Benefits. RESPONSIBILITIES: 1. Provide dedicated major account support. 2. Post-sales support; provide answers to implementation-related questions or documentation. 3. Prepare, present product and application training material to customers. 4. Support customer's development, integration, test and acceptance project phases. 5. Develop diagnostic methodology and problem work-arounds. 6. Participate in engineering reviews; create and suggest diagnostic enhancements. 7. Participate in system integration and test activity. 8. Inter-work issues and customer-requested enhancements with sales and marketing. 9. Lite travel (10% U.S. & International) may be required. REQUIRED QUALIFICATIONS: 1. Minimum 3-8 years of real-time communications systems software development or integration experience with 1-3 in embedded software. 2. Direct development experience using the following: assembler, C, C++, pSOS or some other RTOS, TCP/IP, FR, ATM, digital transmission techniques DESIRABLE QUALIFICATIONS: 1. Hands-on experience using programming scopes, analyzers and related equipment in a software-intensive development environment. 2. Switching communications techniques and systems knowledge. 3. B.S. in E.E., Computer Science or equivalent. Tom Whiting Enterprise Resource Group (voice) (415) 440-1290 2711 Pacific Avenue (fax) (415) 775-2908 San Francisco, CA 94115 (email) tom@eresourcegroup.com --------------------------- End of New-News digest ********************** From owner-vxwexplo-process Thu Apr 15 04:44:00 1999 From: akarpov@my-dejanews.com Date: Thu Apr 15 04:44:03 PDT 1999 Subject: Re: Page Fault Message Interrupt 14 on x86 CPU (PC386/486 BSP of VxWorks )is generated when Page-Fault Exception is detected. Page-Fault Exception indicates that, with paging enabled the processor detected one of the following conditions while using the page-translation mechanism to translate a linear address to a physical address: (o) Present bit indicates that a page table or the page containing the operand is not present in physical memory. (o) The procedure does not have sufficient privilege to access the indicated page (that is, a procedure running in user mode attempts to access a supervisor-mode page). (o) Code running in user mode attempts to write to a read-only page. In the Intel486™ and later processors, the page fault will also be triggered by code running in supervisor mode that tries to write to a read-only user-mode page. Also Page fault may be caused by inconsistent stack pointer (SS:ESP). ----------------------- Alex Karpov Software Engineer Kollmorgen-Servotronix (www.kollmorgen.com) ----------------------- "Glenn Dicus" wrote: > Does anybody know the exact meaning of the page fault message on an x86 > platform. For instance, the message includes the Status Register, but I am > unsure as to what register this maps to on the K6 processor. Likewise, I do > not know the Error Code protocol as well. > > Any pointers would be appreciated. > > thanks > > glenn dicus > Nomadix Inc > From owner-vxwexplo-process Thu Apr 15 07:47:06 1999 From: Roger Cruz Date: Thu Apr 15 07:47:10 PDT 1999 Subject: Re: wtx Error? I encountered this problem a long time ago. It does not only affect the "i" command, but any command that sends a lot of data over the WDB protocol. I tracked it down to a 1500 byte limit (see TSR #47072). WRS said that they would fixed in Tornado 1.0.1. I never received a fix for Tornado 1.0 (vx5.3). I hope this helps you some how. Roger From: "Cui Jian" Date: Wed Apr 14 06:45:13 PDT 1999 Subject: wtx Error? Dear vxworks guru: I encounted a serious problem about Tornado's wdb agent. I don't know why this happened, if you have any comments please tell me. When I started vxworks 5.0.3, I could use "i" to show a summary of tasks information in windsh. But after I download my own programs and spawn the tasks, I couldn't use "i" to show the information, and in crosswind I couldn't attach the tasks. The error showed "WTX_AGENT_GOPHER_FAULT", but I could see the tasks in browser. My programs is rather large-about 2MB in all. I had worked on the programs for a year, and I have debugged it successfully, but something unknown made tornado stop working several days ago. Who can tell me why? ------- Ps Remove "_nojunk" from email address -- Roger Cruz cruz_nojunk@xyplex.com Xyplex Networks w: 978-952-4783 295 Foster Street f: 978-952-4887 Littleton, MA 01460 From owner-vxwexplo-process Fri Apr 16 04:00:09 1999 From: daemon@csg.lbl.gov Date: Fri Apr 16 04:00:13 PDT 1999 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Fri Apr 16 04:00:05 PDT 1999 Subject: Do I have to call sysBusIntAck before my interrupt handler returns?. Subject: Re: Adding my END driver to Vxworks Subject: Looking for information on TASK VARIABLES Subject: Re: RTxWindows problem ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Do I have to call sysBusIntAck before my interrupt handler returns?. Date: Mon, 12 Apr 1999 10:05:45 -0700 From: "rvsaripa" Organization: Intel Corporation Message-ID: <7etdme$nkv@news.or.intel.com> I have a driver for a PCI device for which I have an interrupt handler. The handler itself is getting called, but currently I am calling sysBusIntAck() just before my handler returns back. Do I have to do this?. My driver seems to work fine. Just curious. Please let me know. rk All opinions are mine & mine only. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Adding my END driver to Vxworks Date: Thu, 15 Apr 1999 21:09:51 GMT From: rbwilliams2@lucent.com Organization: Deja News - The Leader in Internet Discussion Message-ID: <7f5kio$7pq$1@nnrp1.dejanews.com> References: <7nOO2.1035$MR1.5043136@news.magma.ca> Yie Wang, Looks like you've done almost everything right. There's one statement missing from your configNet.h file: #define IP_MAX_UNITS - set this to one more than you have right now. This define sets the maximum number of END drivers the system will load and connect to the IP stack. Search through all your code for IP_MAX_UNITS. You'll probably find it set to "1" in usrNetwork.c. If you define it in your configNet.h to "2", then your driver should load. Rich Williams Software Engineer Lucent Technologies Murray Hill, NJ In article <7nOO2.1035$MR1.5043136@news.magma.ca>, "Yie Wang" wrote: > Hi everyone, > You help would be greatly appreciated. > I wrote a very simple dummy END driver. This dirver does nothing, > myEndSend just have one printf(). > > My problem is that when I try to add this dummy END driver to Vxworks > board (mpc 860), it makes the board does not work. > > I followed the procedures in the SENS for Tornado manual and does the > following: > 1) I compled the code of this dummy END driver > 2) Include it into the MACH_EXTRA = mydummy_driver.o into the Makefile of > BSP > 3) In the config.h, I added > #define INCLUDE_END > #define END_OVERRIDE > 4) I also made a configNet.h as follows: > > #ifndef INCnetConfigh > #define INCnetConfigh > > #include "vxWorks.h" > #include "end.h" > #include "c:/tornado/host/src/ywang/drv1/myend.h" > #define END_LOAD_FUNC myEndLoad > > /* unit number, this dummy only contains the unit number, no interrupt */ > > #define END_LOAD_STRING "1" /* asume the unit number is 1*/ > > END_TBL_ENTRY endDevTbl [] = > { > { 1, END_LOAD_FUNC, END_LOAD_STRING, 0, 0}, > { 0, END_TBL_END, 0, NULL}, > }; > > #endif /* INCnetConfigh */ > recreate the vxworks image and it compiles without error. > > 5) Create the vxworks image and it compiles without an error. > I did not rebuild my bootROM. But, I defined END_OVERRIDE in the BSP's > config.h. > > 6) Then, I reboot the board and failed. > > Can you give me some ideas? > > I also tried this: > 1) I run muxDevLoad in the shell (using the previous image without including > END) > It runs without error. > 2) I run muxDevStart and no error ocurrs. > But, the END structure does not get inserted in the OS's END Link List. > endFindByName returns witha NULL and the muxShow shows nothing there. > Could you give me some hints? > > Thank you very much for your time. > Regards, > Yie Wang > > - -----------== Posted via Deja News, The Discussion Network ==---------- http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own --------------------------- Newsgroups: comp.os.vxworks Subject: Looking for information on TASK VARIABLES Date: Thu, 15 Apr 1999 17:50:00 -0400 From: "Keith Smith" Organization: Nortel Message-ID: <7f5n0u$54r$1@bcarh8ab.ca.nortel.com> Hi: I'm finding the information in the VxWorks manuals about TaskVariables a little thin... Can anyone out there point me to some additional resources, or provide some insight? My questions surround the scoping of the variables... probably because I don't really understand what's going on exactly. e.g. code snipet... int * p_int; void foo (opcode) { 1 ... taskVarAdd(tid, &p_int); 2 ... } void bar(opcode) { 3 } What is actually happening here? Is it the contents of the pointer p_int that are being restored/saved at task context switch time, or is it &p_int ? What would we see at point #1, and point #2? Assume multiple tasks are all calling these functions... at point #2, does p_int act functionally the same as a local variable in terms of scope and accessability by other tasks? What happens if instead of declaring the global int *p_int, we declare it at point #1 as a static int *p_int. Note that foo() is being used as a callback which is registered and called repeatedly with an opcode the specify the necessary action(e.g. database query, database write etc.). There are multiple work tasks running concurrently performing these actions . Thanks in advance, Keith --------------------------- Newsgroups: comp.os.vxworks Subject: Re: RTxWindows problem Date: Thu, 15 Apr 1999 15:37:24 -0700 From: "Russell Savio" Organization: EarthLink Network, Inc. Message-ID: <7f5pam$jk2$1@ash.prod.itd.earthlink.net> References: <3713514A.9B9B0A5E@sainsel.es> When you started the rtx server, rtxClientLibInit was called and doesn't need to be called again in your client application. Your client application "prueba.c" should instead call: rtxLibInit( X11LIB | XEXTLIB | XMULIB | XTLIB | XAWLIB ); Ildefonso Junquero Martin-Arroyo wrote in message news:3713514A.9B9B0A5E@sainsel.es... > > > Hi! > > i've tryed to create an rtxwindows very simple application, but i can't > get it works. Can anyone help me? > > thanks! > > Ilde. > > -- > ______________________________________________________________________ > Ildefonso Junquero e@mail : junquero@sainsel.es > ___ __o Software Engineer > ____ _'\<,_ Sainsel Sistemas Navales S.A. > ______ (*)/ (*) C/ Manuel Velasco Pando, N. 7 > 41007 Sevilla (Spain) > Go faster! Phone : +34-95-493 64 65 > Use the byke! Fax : +34-95-493 64 33 > ______________________________________________________________________ - ---------------------------------------------------------------------------- - ---- > /* prueba .c */ > /* ---------------------------------------------- */ > > #include "rtxcnfg.h" > > #include > #include > #include > #include > > /* forward declarations */ > extern int rtxClientLibInit(void); > > int prueba(void) > { > XtAppContext xtcontext; > > Display *display; > > printf("Iniciando cliente\n"); > > /*#ifdef INCLUDE_RTX_CLIENTS*/ > rtxClientLibInit(); > /*#endif */ > > /* AT this point the X server has been initialized an the client > * libraries are initialized. Start any applications. > */ > > printf("Abriendo display\n"); > > display = XOpenDisplay("150.150.8.5:0"); > > if (!display) > printf("NO DISPLAY\n"); > > printf("DISPLAY!!!\n"); > return (0); > } > > /* -------------------------------------------------- */ > /* End of prueba.c */ > > /* Compilation command > > cc68k -B/usr/wind/host/sun4-solaris2/lib/gcc-lib/ -c -m68040 -nostdinc -O -f volatile -fno-builtin -DCPU=MC68040 -DVXWORKS -DVX5x -I. -I/usr/wind/targe t/h/rtx -I/usr/wind/target/h -I/usr/wind/target/src -I/usr/wind/target/confi g/rtx -c prueba.c -o prueba.o > > */ > > > /* > > Output : > > boot device : lm > processor number : 0 > host name : orko > file name : /usr/wind/target/config/cm40/vxWorks > inet on ethernet (e) : 150.150.8.89:ffff0000 > host inet (h) : 150.150.8.5 > user (u) : sirius > flags (f) : 0x0 > target name (tn) : dlt > startup script (s) : /home/users/ijm/tmp/rtx/script > > Attaching network interface lm0... done. > Attaching network interface lo0... done. > Loading... 384532 + 43104 + 25464 > Starting at 0x1000... > > Attaching network interface lm0... done. > Attaching network interface lo0... done. > Loading symbol table from orko:/usr/wind/target/config/cm40/vxWorks.sym ...done > > > ]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]] > ]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]] > ]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]] > ]]]]]]]]]]] ]]]] ]]]]]]]]]] ]] ]]]] (R) > ] ]]]]]]]]] ]]]]]] ]]]]]]]] ]] ]]]] > ]] ]]]]]]] ]]]]]]]] ]]]]]] ] ]] ]]]] > ]]] ]]]]] ] ]]] ] ]]]] ]]] ]]]]]]]]] ]]]] ]] ]]]] ]] ]] ]]] > ]]]] ]]] ]] ] ]]] ]] ]]]]] ]]]]]] ]] ]]]]]]] ]]]] ]] ]]] ] > ]]]]] ] ]]]] ]]]]] ]]]]]]]] ]]]] ]] ]]]] ]]]]]]] ]] ]] > ]]]]]] ]]]]] ]]]]]] ] ]]]]] ]]]] ]] ]]]] ]]]]]]]] ] ]]] > ]]]]]]] ]]]]] ] ]]]]]] ] ]]] ]]]] ]] ]]]] ]]]] ]]]] ]]]] > ]]]]]]]] ]]]]] ]]] ]]]]]]] ] ]]]]]]] ]]]] ]]]] ]]]] ]] ]]] > ]]]]]]]]]]]]]]]]]]]]]]]]]]]]]] > ]]]]]]]]]]]]]]]]]]]]]]]]]]]]] Development System > ]]]]]]]]]]]]]]]]]]]]]]]]]]]] > ]]]]]]]]]]]]]]]]]]]]]]]]]]] VxWorks version 5.3.1 > ]]]]]]]]]]]]]]]]]]]]]]]]]] KERNEL: WIND version 2.5 > ]]]]]]]]]]]]]]]]]]]]]]]]] Copyright Wind River Systems, Inc., 1984-1997 > > CPU: CM-Computer CM-CPU-40. Processor #0. > Memory Size: 0x7ee000. BSP version 1.1/2. > > Executing startup script /home/users/ijm/tmp/rtx/script ... > cd "/usr/wind/target/config/rtx" > value = 0 = 0x0 > ld < rtx.o > value = 7811204 = 0x773084 > rtx > X11 libs loaded > Xext libs loaded > Xmu libs loaded > Xt libs loaded > Xaw libs loaded > Xm libs not loaded > value = 0 = 0x0 > cd "/usr/wind/target/lib/rtx/MC68040gnu" > value = 0 = 0x0 > ld < crtx.o > value = 7883276 = 0x784a0c = crtx.o_bss + 0x8 > > cd "/home/users/ijm/tmp/rtx" > value = 0 = 0x0 > ld < prueba.o > value = 7892632 = 0x786e98 = prueba.o_bss + 0xe18 > > Done executing startup script /home/users/ijm/tmp/rtx/script > -> prueba > Iniciando cliente > X11 libs loaded > Xext libs loaded > Xmu libs loaded > Xt libs loaded > Xaw libs loaded > Xm libs not loaded > Abriendo display > > Access Fault > Program Counter: 0x00044422 > Status Register: 0x3000 > Access Address : 0x83120000 > Special Status : 0x0525 > > 6a8310 _XOpenDisplay +9e : 442f0 ([1, 53c, &_yyval, 3, 3]) > 44308 __x_fpsp_fline +24 : __x_mem_read ([41000000, &_excIntStub, 0, &__x_fpsp_ > fline, 785f5d]) > shell restarted. > > -> > > */ --------------------------- End of New-News digest ********************** From owner-vxwexplo-process Fri Apr 16 07:02:38 1999 From: Charlie Grames Date: Fri Apr 16 07:02:41 PDT 1999 Subject: Re: Do I have to call sysBusIntAck before my interrupt handler RK, I may be wrong, but I think sysBusIntAck() only applies to the VMEbus. I do know that many VMEbus interfaces automatically handle interrupt acknowledge, so calling sysBusIntAck() really does nothing useful. However, for portability, sysBusIntAck() should be used. Again, I do not think sysBusIntAck() applies to the PCI bus. Anyone who knows better, please jump in. Charlie Grames The Boeing Company (314) 233-1956 Charles.R.Grames@boeing.com Newsgroups: comp.os.vxworks Subject: Do I have to call sysBusIntAck before my interrupt handler returns?. Date: Mon, 12 Apr 1999 10:05:45 -0700 From: "rvsaripa" Organization: Intel Corporation Message-ID: <7etdme$nkv@news.or.intel.com> I have a driver for a PCI device for which I have an interrupt handler. The handler itself is getting called, but currently I am calling sysBusIntAck() just before my handler returns back. Do I have to do this?. My driver seems to work fine. Just curious. Please let me know. rk From owner-vxwexplo-process Fri Apr 16 10:58:05 1999 From: "Alan T. Pickering" Date: Fri Apr 16 10:58:09 PDT 1999 Subject: program load time Fellow Vxworkers, Does anyone have an approximate idea of how many seconds a program load from Flash ROM to RAM would take ? Please assume two scenarios: 1) the program image and Vxworks have been compressed in order to fit into an 8 Meg Flash memory and actually take up about 6 out of the 8 Meg 2) the program image and Vxworks have not been compressed and take up the full 8 Meg. My platform is a 300 Mz PowerPC (750). Thanks. Allan Pickering Raytheon Portsmouth, RI Email: atp@ssd.ray.com From owner-vxwexplo-process Fri Apr 16 18:41:53 1999 From: Janusz Kolodziejczyk Date: Fri Apr 16 18:41:57 PDT 1999 I have been experiencing problems with WindLicense Server. The license server is responding very slowly and incorrectly. The time of response sometimes is up to a few minutes. In addition every time I close the target console, I have to unregister it manually, otherwise it is still registered as a target in use. In order to eliminate potential network problems, I installed Tornado and WindLicense Server on my local machine, however it did not change the behavior. So far the WindRiver Technical Support has not been able to help me to resolve this problem. If anyone experienced this type of problems or has some ideas how to solve it, please let me know. ------------------------------------------------------------- Janusz Kolodziejczyk voice: (650) 327-1200 x.138 SBS Technologies, Modular I/O fax: (650) 327-3808 181 Constitution Drive e-mail: januszk@sbs-mio.com Menlo Park CA 94025 URL: www.sbs.com ------------------------------------------------------------- From owner-vxwexplo-process Sun Apr 18 04:00:16 1999 From: daemon@csg.lbl.gov Date: Sun Apr 18 04:00:20 PDT 1999 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sun Apr 18 04:00:13 PDT 1999 Subject: tNetTask: Receive Structure Invalid Subject: Re: tNetTask: Receive Structure Invalid Subject: Re: Corba and VxWorks ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: tNetTask: Receive Structure Invalid Date: Fri, 16 Apr 1999 05:58:12 -0700 From: "Reed, Michael" Message-ID: References: <37091AE2.67E6@wrs.com> Sender: "Reed, Michael" 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_001_01BE8808.C8BF0892 Content-Type: text/plain; charset="iso-8859-1" How do I turn off the data cache? WindConfig? - -----Original Message----- From: Stephen Macmanus [mailto:stephenm@wrs.com] Posted At: Monday, April 05, 1999 4:20 PM Posted To: vxworks Conversation: tNetTask: Receive Structure Invalid Subject: Re: tNetTask: Receive Structure Invalid What network driver are you using? The BSD Sonic driver is the only portion of our code which generates a similar error message. In that case, it indicates that the receive descriptor is invalid. Possible causes of that error include a reported frame lengths outside the available range. It can also result from problems handling the data cache, if once is available for the BSP. If you are not using the Sonic driver, then this message is probably generated by custom modifications to the stack. Good luck. Stephen Mike Reed wrote: > > Hello, > > Has anyone ever heard of the error 'tNetTask: Receive Structure > Invalid'? I am using a DY4 PowerPPC 603e BSP with a DY4 783 Graphics board. > Anybody heard of this and how to correct it? Any help provided is greatly > appreciated. > > Mike Reed > reedm@battelle.org - -- - ------------------ Stephen Macmanus #include Software Engineer Wind River Systems - ------_=_NextPart_001_01BE8808.C8BF0892 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable

How do I turn off the data cache?  = WindConfig?

-----Original Message-----
From: Stephen Macmanus [mailto:stephenm@wrs.com]
Posted At: Monday, April 05, 1999 4:20 PM
Posted To: vxworks
Conversation: tNetTask: Receive Structure = Invalid
Subject: Re: tNetTask: Receive Structure = Invalid


What network driver are you using? The BSD Sonic = driver is the
only portion of our code which generates a similar = error message.
In that case, it indicates that the receive = descriptor is invalid.
Possible causes of that error include a reported = frame lengths outside
the available range. It can also result from = problems handling the
data cache, if once is available for the BSP.

    If you are not using the Sonic = driver, then this message is
probably generated by custom modifications to the = stack. Good luck.

          &nb= sp;           &nb= sp;           &nb= sp;           &nb= sp;        Stephen

Mike Reed wrote:
>
> Hello,
>
>     Has anyone ever heard = of the error 'tNetTask: Receive Structure
> Invalid'?  I am using a DY4 PowerPPC 603e = BSP with a DY4 783 Graphics board.
> Anybody heard of this and how to correct = it?  Any help provided is greatly
> appreciated.
>
> Mike Reed
> reedm@battelle.org

--
------------------
Stephen = Macmanus          &nbs= p;           &nbs= p;  #include <std_disclaimer.h>
Software Engineer
Wind River Systems

- ------_=_NextPart_001_01BE8808.C8BF0892-- --------------------------- Newsgroups: comp.object.corba,comp.os.vxworks,comp.soft-sys.ace Subject: Re: Corba and VxWorks Date: 16 Apr 1999 10:59:50 -0500 From: schmidt@tango.cs.wustl.edu (Douglas C. Schmidt) Organization: Computer Science Department, Washington University. Message-ID: <7f7mpm$mvg@tango.cs.wustl.edu> References: <37146374.6ED554D5@post.rwth-aachen.de> <7f2q6i$phe$1@nnrp1.dejanews.com> <7f7ee9$isb@nsfhh5.alcatel.fr> Hi YP, >> Is it practicable? Yes, indeed. Check out http://www.cs.wustl.edu/~schmidt/TAO-boeing.html for a recent example of where ACE+TAO were used for avionics mission computing over VxWorks/VME on a Harrier jet. >> I have downloaded ACE and TAO and I have not been able to run a >> single example because I had not enough memory with 8 MBytes. You may be running into bugs with the GNU C++ compiler distributed with VxWorks. It's got some problems with templates, which is why we use GreenHills for our projects. Please see http://www.cs.wustl.edu/~schmidt/ACE-INSTALL.html for more information. >> As far as I can remember, the minimal footprint for Iona Orb is >> about 50 Kbytes that, I think, is quite correct. I find that rather hard to believe ;-) Take care, Doug - -- Dr. Douglas C. Schmidt, Associate Professor Department of Computer Science, Washington University St. Louis, MO 63130. Work #: (314) 935-4215; FAX #: (314) 935-7302 schmidt@cs.wustl.edu, www.cs.wustl.edu/~schmidt/ --------------------------- End of New-News digest ********************** From owner-vxwexplo-process Mon Apr 19 01:57:46 1999 From: David Laight Date: Mon Apr 19 01:57:50 PDT 1999 Subject: Running two END drivers "Yie Wang" wrote: > I also tried this: > 1) I run muxDevLoad in the shell (using the previous image without including > END) > It runs without error. > 2) I run muxDevStart and no error ocurrs. > But, the END structure does not get inserted in the OS's END Link List. > endFindByName returns witha NULL and the muxShow shows nothing there. > Could you give me some hints? > > Thank you very much for your time. > Regards, > Yie Wang > I found that endFindByName doesn't work! Apparantly it was NEVER tested with more than one END driver. I haven't the source, but I don't even think the required loop is present! This problem was assigned SPR21055 (I recall in June '98), the fix was 'rolled into' patch 21644 which is obsoleted by SENS 1.1. This is fine unless, like ourselves, you are using the ARM - where SENS 1.1 has not been generated. As usual WRS support were not exceptionally helpful (they are good at managing bug reports, not very good at fixing bugs...) The following implementation fixed the problem: #include #include #include END_OBJ * tad_endFindByName( char *name, int unit ) { END_TBL_ROW *row; END_OBJ *end; extern int *muxDevExists; for (row = (void *)lstFirst( endList ); row; row = (void *)lstNext( &row->node )) { if (strcmp( name, row->name )) continue; for (end = (void *)lstFirst( &row->units ); end; end = (void *)lstNext( &end->node )) if (end->devObject.unit == unit) return end; } return 0; } void patch_endFindByName(void) { extern int endFindByName; /* plant 'b tad_endFindByName' over first instruction of endFinfByName. */ endFindByName = 0xea000000 | (((int *)&tad_endFindByName - &endFindByName - 2) & 0xffffff); } The patch must be applies before the code is made read only. The branch generated is an ARM one, the PPC equivalent will be different! You will need to find 'endList' - it's not a visible symbol (WRS are good at that too) - You should be able to write some code that finds the value that the existing endFindByName uses (The address may be at a fixed offset from endFindByName). ---------------------------------------------------------------- David Laight email: dsl@tadpole.co.uk Tadpole Technology plc phone: +44 1223 278 256 Cambridge, UK fax: +44 1223 278 201 From owner-vxwexplo-process Mon Apr 19 04:00:09 1999 From: daemon@csg.lbl.gov Date: Mon Apr 19 04:00:13 PDT 1999 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Mon Apr 19 04:00:06 PDT 1999 Subject: Re: 2 byte vs 4 byte addressing on VME Subject: Re: Zinc discontinues desktop versions ------------------------------------------------------- Newsgroups: comp.os.vxworks,jurgan.almlie@lmco.com Subject: Re: 2 byte vs 4 byte addressing on VME Date: Wed, 14 Apr 1999 21:46:51 -0400 From: "Tim Stanka" Organization: EarthLink Network, Inc. Message-ID: <7f3gbv$f76$1@ash.prod.itd.earthlink.net> References: <37151806.28971AE0@lmco.com> Looks like you have an A32D32 slave which does not support A32D16 and may not support A32D08(even,odd). If you have a VME bus analyzer I'm sure you would see the bus timeout monitor (part of system controller bd.) assert BUSERR. If I remember correctly the VME spec says if you support A32D32 you should support D16 and D08 but many people ignore this. Probably your slave device is memory in a D32 arrangement. One reason people choose not to support D16 and D08 is because the data may appear on different byte lanes for a given address depending on data width. You would need to design in a means to direct data to the correct byte lanes depending on address value and datawidth. Tim --------------------------- Newsgroups: comp.os.vxworks,alt.comp.zinc Subject: Re: Zinc discontinues desktop versions Date: Mon, 19 Apr 1999 09:19:09 +0200 From: Ole Asbjorn Fadum Organization: Telenor Online Public Access Message-ID: <371AD8ED.742ADDF9@scanmar.no> References: <371436E5.B47DABDB@scanmar.no> <37170910.10562057@news.fr.bosch.de> Hi, Yes I received an official letter from Zinc sw, signed Larry Macfarlane President Zinc Software. The letter contain three topics: 1. Release of Zinc V 5.3 2. Zinc Desktop Versions Discontinued 3. UK Office Closed. Sorry but it is on paper so I could not send you a copy. Jan Koeppen wrote: > On Wed, 14 Apr 1999 08:34:13 +0200, Ole Asbjorn Fadum > wrote: > > >Hi, > > > >We deeply regret the mail we received today, regarding the discontinued > >development of the Zinc graphical toolkit for the > >Windows/Motif/DOS/Machintosh and OS/2. > > Since we are currently looking into ZINC as a graphics framework I am > interested in the essence of this mail (i.e. is it an official > statement from ZINC / WRS) > > --------------------------- End of New-News digest ********************** From owner-vxwexplo-process Mon Apr 19 13:38:32 1999 From: "Kozel, Mark Mr ITT/FSC" Date: Mon Apr 19 13:38:36 PDT 1999 Subject: VxWorks and Peritek Cards Hi, I'm having trouble getting my PPC 300 mhz (2604) to talk with my Peritek card (VCL-V). I'm using vxworks 5.3.1 with patch. When I spawn pkstub I get: Bus Fault @ LAR 0xFC00 pkSizeMemory found 0 Megabytes of memory pkSizeMemory Failed! Illegal RPC Op -1 Illegal RPC Op -1 Illegal RPC Op -1 <- this message scroll Thanks, Mark ------------------------------------------ Mark Kozel * kozelm@fsc1.vafb.af.mil | "I climb the walls of my mind ITT Federal Services - bldg. 6525 ms 5320 | just like I'm climbing on a PO Box 5728 VAFB, CA 93437 (805) 606-7928 | jungle gym. | I am more that content with | the state of mind I am in..." -BNL "Crazy" From owner-vxwexplo-process Mon Apr 19 20:59:22 1999 From: "John McQueen" Date: Mon Apr 19 20:59:26 PDT 1999 Subject: END vs. NID Hey all, With the release of Tornado 2.0 coming soon... I am going to have to convert a several network drivers to SENS. The porting from the 4.3 to 4.4 BSD stack looks trivial; however, I am looking into porting several of these drivers to the mux model. Can someone enumerate the advantages / disadvantages of moving to the mux model? Is there overhead in the mux model? thanks, John McQueen From owner-vxwexplo-process Tue Apr 20 06:48:43 1999 From: "Wright, Robert B." Date: Tue Apr 20 06:48:46 PDT 1999 Subject: UML tool is there a preferred UML tool to use with Tornado/VxWorks? Rob begin 600 winmail.dat M>)\^(B8-`0:0"``$```````!``$``0>0!@`(````Y`0```````#H``$(@`<` M&````$E032Y-:6-R;W-O9G0@36%I;"Y.;W1E`#$(`06``P`.````SP<$`!0` M"``P`!X``@!&`0$@@`,`#@```,\'!``4``@`,``C``(`2P$!"8`!`"$```!% M-30U.30X-T0W1C9$,C$Q.48V1#`P03!#.4$U,#4P-@`5!P$$@`$`"0```%5- M3"!T;V]L`,P"`0V`!``"`````@`"``$#D`8`&`8``#$````"`7$``0```!8` M```!OHLT>/_P;B67]O`1TH51`,"5[D60```#`-X_KV\```,``H`((`8````` M`,````````!&`````%*%``#C%0``'@`#@`@@!@``````P````````$8````` M5(4```$````$````."XU``L`=X`((`8``````,````````!&``````:%```` M`````P`$@`@@!@``````P````````$8``````84````````+``"`""`&```` M``#`````````1@`````#A0````````L`!8`((`8``````,````````!&```` M``Z%`````````P`!@`@@!@``````P````````$8`````$(4````````#``:` M""`&``````#`````````1@`````1A0````````,`!X`((`8``````,`````` M``!&`````!B%````````'@`(@`@@!@``````P````````$8`````-H4```$` M```!`````````!X`"8`((`8``````,````````!&`````#>%```!`````0`` M```````>``J`""`&``````#`````````1@`````XA0```0````$````````` M"P`8@`L@!@``````P````````$8``````(@````````+`!F`"R`&``````#` M````````1@`````%B`````````(!"1`!````MP```+,```#T````3%I&=4J$ M>P4#``H``'```0````D` M``!534P@=&]O;``````+``(``0````,`_3_D!```0``Y`)`6DGDTB[X!`P#Q M/PD$```>`#%``0````D```!W`/@_`0```!(```!7`#E``0````D```!W32+O@%```@P\O(H?#2+O@$>`#T``0````$`````````'@`=#@$` M```)````54U,('1O;VP`````'@`U$`$````_````/$8T.# Date: Tue Apr 20 08:48:42 PDT 1999 Subject: GNU Compiler for Windows and/or MSDOS VxWorkers- I am looking for a version of the GNU Compiler (or any other free compiler for that matter) that will run on a laptop running Windows 98 and compile 68K code for Motorola MVME162s and MVME167s running VxWorks. Does anybody know of such a package? Jim Hawthorne Applied Technology Associates Albuquerque, New Mexico From owner-vxwexplo-process Wed Apr 21 04:01:05 1999 From: daemon@csg.lbl.gov Date: Wed Apr 21 04:01:08 PDT 1999 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Wed Apr 21 04:01:00 PDT 1999 Subject: Re: timing between 2 points in a program Subject: Re: help no gethostbyname Subject: Re: Object Oriented operating system resources Subject: pre-SENS IP stack routing bug Subject: scsi problem in x86 platform. Subject: Driver test ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: timing between 2 points in a program Date: Tue, 20 Apr 1999 18:06:27 GMT From: Bill Pringlemeir Message-ID: References: <371C49CC.295D@gecm.com> <371CA5C0.86A166BB@adc.com> Sender: bpringle@DeadDuck Actually, a better scheme is to not use interrupts at all. The timestamp drivers support this with some hardware. Since you don't want to 'schedule' something, but just want to know what the current time is, reading a register is fine. Using interrupts (especially at 1000Hz) maybe too much for some boards. It would also use a lot of power for a battery operated device. The decrementer timer library for the MPC860 is _pathetic_ compared to the RISC and 32 bit timers in the CPM. The fact that WRS didn't support them in the MBX is pure sloth. The drivers must have been in a 68360 BSP. Then again, the fact that you can write your own is a plus for VxWorks. ...so don't use interrupts, read a register directly. The sysTimestamp() function probably does this in whatever BSP you are using. You have to (INCLUDE_TIMESTAMP) as John wisely noted. hth, Bill >>>>> David Rigg wrote: >> Does anyone know of a function like time() but which gives a >> finer granularity i.e. gives hundredths of a second? >>>>> "John" == John Blackwood writes: John> If you want hundredths of a second, set the tick rate to 100 John> Hz (60 Hz default) by 'sysClkRateSet (100)' and then calling John> 'tickGet' to read the time. Since you can be anywhere within John> the tick period when 'tickGet' is called, the average John> resolution will be 15 msec. John> For better resolution, mod the 'sysClkInt' routine to John> interrupt at a 1000Hz and keep a separate counter. This John> would require calling 'tickAnnounce' on some multiple of John> these interrupts. Reading this count would then give you a 1 John> msec time. John> Even better, is using time stamp (INCLUDE_TIMESTAMP). This John> works by using a decrement counter for the system clock John> interrupt. By reading this decrement count and 'tickGet' John> your get good resolution with little overhead. I'm using John> this currently on a MPC860 (see John> src/drv/timer/ppcDecTimer.c). This driver was modified to John> return ticks along with the decrement count. I did the same John> thing several years ago on a i486 board. - -- Brave New Waves, "http://www.radio.cbc.ca/programs/bnw/" --------------------------- Newsgroups: comp.os.vxworks Subject: Re: help no gethostbyname Date: Sun, 18 Apr 1999 07:34:35 -0700 From: "Mike Campbell" Message-ID: <7ff2aj$8d$1@autumn.news.rcn.net> References: <37146951.5B668184@webrider.net.au> <37158071.3C5F7B3F@siemens.at> might also have to kick the arp in the butt to get IP resolved, using etherAddrResolve (sp?) Andreas Lugbauer wrote in message <37158071.3C5F7B3F@siemens.at>... >Hi ! > >Maybe you can use the vxWorks function hostGetByName(). It returns the ip >address >corresponding to a host name. >The host must be in the vxWorks host table, which can be done by the >hostAdd() function. > >I hope this helps > >regards > >Andreas > >Stuart and Kerry Clark schrieb: > >> i am new to vxworks. >> >> i am writing some network code. on unix i usually read in a "logical" >> name for a network, eg "stuart", from file. in my code i then call >> gethostbyname("stuart"). gethostbyname returns the ip address of stuart. >> on unix this is ok because there is an /etc/hosts file. i looked in the >> vx programmers guide and found gethostbyname isnt there. how do you vx >> gurus usually do such a thing. >> >> 1)someone a work came up with the idea of sending a message to our unix >> machine which would then call gethostbyname and then send a message back >> to the sbc with the ip address. >> >> in unix when you do a network write the os determines which interface >> (ehternet card say) to send the packet out using the /etc/hosts file and >> by the interface being preconfigured to a certain ip address. >> >> question - how does vx determine which interface to write a packet to. i >> understand you configure it by associating a driver with an interface. >> how are a number of ip address of networks get assigned. is there an >> equivalent mechanism of /etc/hosts. i know one interface is assigned >> for the down load of the actual vx from a server but what about other >> networks ? >> >> sorry for the dumb question > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Object Oriented operating system resources Date: Fri, 16 Apr 1999 09:27:41 +0200 From: "Markus Mitterer" Organization: Customer of EUnet Austria Message-ID: <7f6osi$sg4$1@fleetstreet.Austria.EU.net> References: <371595ab.0@nnrp1.news.uk.psi.net> Hallo Paul, I am enthusiastic about that you have an idea to improve C++ support under VxWorks. In my company we were thinking about the same but we are going to use an simplier method than you are. Do you use the normal TaskLib-Functions for your class? If it is possible, please let me have a look at your source. best regards Markus Mitterer, STK --------------------------- Newsgroups: comp.os.vxworks Subject: pre-SENS IP stack routing bug Date: Tue, 20 Apr 1999 16:21:58 -0500 From: Andy Mauro Organization: Nortel Networks Message-ID: <371CEFF6.99D40C29@broadbandnetworks.com> Howdy all. My problem is caused by one bug and one unsupported feature of the VxWorks 4.3 BSD stack (i.e. pre-SENS stack). The unsupported feature is that VxWorks does not support the binding of a subnet mask to a routing table entry. Thus, all routing comparisons are done based upon the application of the locally attached interface's subnet mask. In our case the customer would like to be able to add a routing table entry of the following form : Destination Gateway 10.0.0.0 10.40.26.39 However, the subnet mask of the locally attached interface (10.40.26.49) is /26 bit. Thus the application of a 26 bit subnet mask to any outgoing packets destined for the 10 network (e.g. say we ping 10.1.1.1 from our box) does not render a match with a 10.0.0.0 routing table entry. What we would like to be able to do is enter a routing table entry of the following form: Destination Gateway Subnetmask 10.0.0.0 10.40.26.39 255.0.0.0 The SENS stack support the addition of subnet masks to routes, but we are not yet ready to move to SENS...problem. The second problem is a bug that causes us some more grief (i.e. harder workarounds for our situation). The bug is that if a particular box has multiple interfaces, and two or more of those interfaces are on the same network, within the same class (e.g. two interfaces on the 10.x.x.x network - they will of course be on different subnetworks, but their class A is the same), then ONLY the subnet mask of the first attached interface is applied for routing comparisons. This means that instead of the routing algorithm using each of the interface's subnet masks for routing comparisons, only the first interface's subnet mask is used. As you might suspect, unless all these interfaces have the same subnet mask, or the first interface has the least restrictive subnet mask, some routing problems will occur. *** I am not looking for workarounds. I am looking for someone who knows where we might pick up (buy) modified VxWorks 4.3 BSD source code that correctly implements IP routing. *** Thanks, - -Andy Andy Mauro Software Designer Nortel Networks Broadband Wireless Access Unit mailto:mauroa@nortelnetworks.com --------------------------- Newsgroups: comp.os.vxworks Subject: scsi problem in x86 platform. Date: Wed, 21 Apr 1999 10:22:35 +0900 From: Kim Wonkook Organization: System Engineering Research Institute (SERI) Message-ID: <371D285B.172CD105@cs.kyungpook.ac.kr> hello. to use scsi disk, i define "INCLUDE_SCSI" in config.h. and compile. then some error occur. Error message is "sysScsiInit" Not found.!! I could not find scsilib.o in object files. only "scsilib.h" exist. i use pc486 bsp. how can i do it? x86 does not support scsi??? thanks. have a nice day. - - Wonkook Kim I --------------------------- Newsgroups: comp.os.vxworks Subject: Driver test Date: Wed, 21 Apr 1999 09:53:04 +0300 From: Bahri Okuroglu Organization: Netas - Northern Electric Telecom AS Message-ID: <371D75CE.12825AE5@rnd.netas.com.tr> Hi all, I'm just trying to test the code giving in the VxWorks 5.3.1 Programmer's Guide Example 3-9 Hypothetical Driver. Compiled the below code and successfully installed the driver in Vx. However I could not manage to create a device. Can anyone tell me the problem with my code? Cheers, ================ THE CODE ================= #include "vxWorks.h" #include "iosLib.h" int dummyDrvNum; DEV_HDR *dummyDev; STATUS dummyDrv(); STATUS dummyDevCreate(char *); int dummyOpen(char *, char *, int); int dummyRead(char *, char *, int); int dummyWrite(char *, char *, int); STATUS dummyDrv() { dummyDrvNum = iosDrvInstall(dummyDevCreate, 0, dummyOpen, 0, dummyRead, dummyWrite, 0); if (dummyDrvNum==ERROR) { printf("There is no room for the driver\n"); return ERROR; } else { printf("\n\nIn initialization routine\nReturn value is %d\n\n", dummyDrvNum); return OK; } } STATUS dummyDevCreate(name) char *name; { STATUS status; status = iosDevAdd(dummyDev, name, dummyDrvNum); if (status==ERROR) { printf("Could not add the device %s\n", name); return ERROR; } else { printf("in DevCreate\n%s\t%d\t%d\n", dummyDev->name, dummyDev->drvNum, status); return status; } } int dummyOpen(dummyDev, remainder, mode) char *dummyDev; char *remainder; int mode; { printf("in dummyopen\n"); return (int) dummyDev; } int dummyRead(dummyDev, buffer, nBytes) char *dummyDev; char *buffer; int nBytes; { printf("in dummyRead"); return nBytes; } int dummyWrite(dummyDev, buffer, nBytes) char *dummyDev; char *buffer; int nBytes; { printf("in dummyWrite"); return nBytes; } ============= END OF THE CODE ================ ============= COMMAND LINE OUTPUT ================= - -> ld < dummyDrv.o value = 1073894456 = 0x40025438 - -> dummyDrv value = 0 = 0x0 In initialization routine Return value is 11 - -> dummyDevCreate("/test"); value = 0 = 0x0 in DevCreate /test 11 0 - -> devs drv name 0 /null 1 /tyCo/0 1 /tyCo/1 2 /pcConsole/0 2 /pcConsole/1 7 bahrio: 8 /pty/rlogin.S 9 /pty/rlogin.M 8 /pty/telnet.S 9 /pty/telnet.M 10 /vio value = 0 = 0x0 - -> iosDrvShow drv create delete open close read write ioctl 1 146c60 0 146c60 0 111c70 111b90 146c70 2 108688 0 108688 0 111c70 111b90 108698 3 0 0 14c81c 0 14c85c 14c8ac 14c96c 4 11db48 11df38 1215e8 11cb18 121aa8 123188 120d38 5 14e0a8 0 14e0a8 0 14e0c8 14e0e8 14e388 6 0 0 0 131efc 132b0c 1326cc 15e738 7 134b30 134c90 134ea0 134f70 135c70 135df0 135f60 8 1626e8 0 1626e8 1626f8 162748 1627d8 162838 9 1626d8 0 1626d8 162728 162768 1627f8 162858 10 1423d8 0 1423d8 142498 111c70 111b90 142508 11 58c64c 0 58c6cc 0 58c6fc 58c72c 0 value = 0 = 0x0 - -> ============ END OF THE COMMAND LINE OUTPUT =========== - -- ====================================================== Bahri OKUROGLU Software Design Engineer Netas R&D RT6 mailto:bahrio@rnd.netas.com.tr http://www.netas.com.tr mailto:bahrio@yahoo.com http://193.140.184.5/r748/member/bahrio (internal) Netas Alemdag Cad. Umraniye 81244 ISTANBUL TURKEY ====================================================== --------------------------- End of New-News digest ********************** From owner-vxwexplo-process Wed Apr 21 05:00:10 1999 From: Florian Bittner Date: Wed Apr 21 05:00:14 PDT 1999 Subject: Re: GNU Compiler for Windows and/or MSDOS At 08:48 20.04.99 -0700, you wrote: >Submitted-by owner-vxwexplo-process Tue Apr 20 08:48:38 1999 >Submitted-by: "Jim Hawthorne" > >VxWorkers- > >I am looking for a version of the GNU Compiler (or any other free compiler >for that matter) that will run on a laptop running Windows 98 and compile >68K code for Motorola MVME162s and MVME167s running VxWorks. > >Does anybody know of such a package? Hi, you may extract the compiler portion from tornado ( cc68k and friends ) or you may use the cygnus win32 stuff and compiler a cross compiler yourself. It´s all documented and vxworks is supported as a build target, you need to add the vxworks libraries. By the way, the c compiler is on the tornado cd´s in source form. Greetings Florian Bittner Kayser-Threde GmbH Tel +49 421 2235133 Hauptmann-Boese-Weg 24 Fax +49 421 2235134 28213 Bremen Germany From owner-vxwexplo-process Wed Apr 21 07:54:35 1999 From: "James O. Prewitt" Date: Wed Apr 21 07:54:40 PDT 1999 Subject: Looking for VxWorks device driver for Intel 21143 chip based ethernet communications... I have a customer who is interested in using a PC from Trenton who is well known in industrial PC circles. The board has Celeron II processor and the board name is CBX 5690. It has on-board ethernet support which uses the Intel 21143 chip. In the past I have been able to use the plain Jane pc486 BSP to handle most of these kinds of boards. Does anyone have any experience with this board running on VxWorks 5.3.1, or know of a ethernet driver that might work on this board? Thanks in advance. Jim Prewitt Adventa Control Technologies, Inc. j-prewitt@adventact.com 972-543-1687 From owner-vxwexplo-process Wed Apr 21 10:04:16 1999 From: lee.wheat@wg.com Date: Wed Apr 21 10:04:20 PDT 1999 Subject: Re: help no gethostbyname you can provide your own version of the /etc/hosts file by using hostAdd() to add entries to vxWorks' host table. this way, as long as the desired host is in the table, the call to hostGetByName() will be able to perform the resolution. if you want to handle any case there are two options (that i know of): 1. implement a name resolver. you don't have to do a full blown one. a stub resolver will do if your DNS server is not limited in some way (some freeware ones don't implement everything). the source to nslookup is available on the net and can be trimmed down to the bare essentials to get you what you want. 2. switch to SENS. one of the new features in SENS is a name resolver, so you don't have to implement it yourself. as to how vxWorks picks an interface to use when sending a packet, that is determined by the route the packet is going to take. vxWorks build up a routing table from the addresses associated with the interface cards it knows about and additional entries provided by routeAdd(). based on the destination address of the packet that is being sent, vxWorks picks the card that has the best path. for an example, do a routeShow(). the output shows the current contents of the routing table. along with each entry, vxWorks will display the card it will use for that route. HTH, lee wheat wheat@wg.com > > i am new to vxworks. > > i am writing some network code. on unix i usually read in a "logical" > name for a network, eg "stuart", from file. in my code i then call > gethostbyname("stuart"). gethostbyname returns the ip address of stuart. > on unix this is ok because there is an /etc/hosts file. i looked in the > vx programmers guide and found gethostbyname isnt there. how do you vx > gurus usually do such a thing. > > 1)someone a work came up with the idea of sending a message to our unix > machine which would then call gethostbyname and then send a message back > to the sbc with the ip address. > > in unix when you do a network write the os determines which interface > (ehternet card say) to send the packet out using the /etc/hosts file and > by the interface being preconfigured to a certain ip address. > > question - how does vx determine which interface to write a packet to. i > understand you configure it by associating a driver with an interface. > how are a number of ip address of networks get assigned. is there an > equivalent mechanism of /etc/hosts. i know one interface is assigned > for the down load of the actual vx from a server but what about other > networks ? > > sorry for the dumb question From owner-vxwexplo-process Wed Apr 21 10:27:38 1999 From: lee.wheat@wg.com Date: Wed Apr 21 10:27:42 PDT 1999 we also needed to handle routing in a manner not supported in vxWorks and we weren't willing to move to SENS to do it. SENS also did not provide exactly the type of functionality we were looking for either. what we ended up doing was explaining, in detail, exactly what we were trying to do and why vxWorks' method of routing was not what we needed to the wind river sales engineer. he eventually gave us the source to ip_output.c. this is the chunk of the kernel that performs IP routing in vxWorks. we changed it to use our routing methodology and removed the original ip_output.o from the library. everything worked fine. the wind river lawyers would probably descend upon me if i sent you the file. just bug your sales engineer. if your support contract is paid up, it should be no problem. HTH, lee wheat wheat@wg.com >Howdy all. > >My problem is caused by one bug and one unsupported feature of the >VxWorks 4.3 BSD stack (i.e. pre-SENS stack). The unsupported feature is >that VxWorks does not support the binding of a subnet mask to a routing >table entry. Thus, all routing comparisons are done based upon the >application of the locally attached interface's subnet mask. In our >case the customer would like to be able to add a routing table entry of >the following form : > >Destination Gateway >10.0.0.0 10.40.26.39 > >However, the subnet mask of the locally attached interface (10.40.26.49) >is /26 bit. Thus the application of a 26 bit subnet mask to any >outgoing packets destined for the 10 network (e.g. say we ping 10.1.1.1 >from our box) does not render a match with a 10.0.0.0 routing table >entry. What we would like to be able to do is enter a routing table >entry of the following form: > >Destination Gateway Subnetmask >10.0.0.0 10.40.26.39 255.0.0.0 > >The SENS stack support the addition of subnet masks to routes, but we >are not yet ready to move to SENS...problem. > >The second problem is a bug that causes us some more grief (i.e. harder >workarounds for our situation). The bug is that if a particular box has >multiple interfaces, and two or more of those interfaces are on the same >network, within the same class (e.g. two interfaces on the 10.x.x.x >network - they will of course be on different subnetworks, but their >class A is the same), then ONLY the subnet mask of the first attached >interface is applied for routing comparisons. > >This means that instead of the routing algorithm using each of the >interface's subnet masks for routing comparisons, only the first >interface's subnet mask is used. As you might suspect, unless all these >interfaces have the same subnet mask, or the first interface has the >least restrictive subnet mask, some routing problems will occur. > >*** >I am not looking for workarounds. I am looking for someone who knows >where we might pick up (buy) modified VxWorks 4.3 BSD source code that >correctly implements IP routing. >*** > >Thanks, >- -Andy > > >Andy Mauro >Software Designer >Nortel Networks >Broadband Wireless Access Unit >mailto:mauroa@nortelnetworks.com From owner-vxwexplo-process Wed Apr 21 12:16:57 1999 From: Jochen Linkohr Date: Wed Apr 21 12:17:00 PDT 1999 Subject: Re: UML tool Hi, vxwexplo@lbl.gov wrote: > is there a preferred UML tool to use with Tornado/VxWorks? The only thing I know is that ILogix' Rhapsody supports VxWorks as target OS. I do not know how useful the generated code is. I anybody does so, please let me know... (by Email..) bye Jochen From owner-vxwexplo-process Wed Apr 21 13:02:52 1999 From: Kim Middleton Date: Wed Apr 21 13:02:55 PDT 1999 Subject: VxWorks: running off an IDE drive? Hi! Curiousity has me wondering if anyone has ever made an executable from their application code, put it on a hard-drive in your VxWorks system, and then tried to run the code off the hard drive without loading it into memory in its entirety? If so, could you give me some pointers? Thanks! Kim -- Kim Middleton Hewlett-Packard Company kimberly@spk.hp.com 24001 E. Mission Ave Voice: (509) 921-3817 Liberty Lake, WA 99019 FAX: (509) 921-3991 From owner-vxwexplo-process Wed Apr 21 14:34:47 1999 From: Vasudev Dasari Date: Wed Apr 21 14:34:51 PDT 1999 Subject: Memory leaks Hello, Can someone tell me how to dettect memory leaks. Do we have any tools from Tornado to check them. Thanks Vasu From owner-vxwexplo-process Wed Apr 21 18:40:18 1999 From: Turley Scott-P27856 Date: Wed Apr 21 18:40:22 PDT 1999 Subject: Symbolic links in a vxworks supported file system Does vxworks support symbolic links? We can't find any evidence in the documentation so our opinion is no, but we'd like to get a confirmation. Thanks in advance. Scott Turley From owner-vxwexplo-process Wed Apr 21 23:35:14 1999 From: Arand Thomas Date: Wed Apr 21 23:35:18 PDT 1999 Subject: AW: UML tool Hi, we are using together from Object International (http://www.oisoft.com), and it suites our needs. However, it has no special Tornado support. Ciao, Thomas ---------------------------------------------------------------------- Thomas Arand Siemens AG Tel. +49 - 89 - 722 42296 Fax. +49 - 89 - 722 26572 thomas.arand@icn.siemens.de ----Ursprüngliche Nachricht----- Von: vxwexplo@lbl.gov [SMTP:vxwexplo@lbl.gov] Gesendet am: Dienstag, 20. April 1999 15:49 An: vxworks_users@csg.lbl.gov Betreff: UML tool Submitted-by owner-vxwexplo-process Tue Apr 20 06:48:43 1999 Submitted-by: "Wright, Robert B." is there a preferred UML tool to use with Tornado/VxWorks? Rob From owner-vxwexplo-process Thu Apr 22 02:16:13 1999 From: stevens Date: Thu Apr 22 02:16:17 PDT 1999 Subject: IP-Routing for vxworks ? Hi VxWorkers ! We use Tornado 1.0.1 and must route packets from ISDN/PPP to Ethernet. Does VxWorks support this feature like other unix based systems ? If not, who wrote a software and can share this code (maybe Linux code) ? We are interested in NAT-Software for vxworks, too. Thanks, Torsten -- ************************************************************************** fraunhofer institute of microelectronic circuits and systems department sat name torsten stevens mailto:stevens@ims.fhg.de address finkenstrasse 61 phone +0049/203/3783-238 47057 duisburg fax +0049/203/3783-266 germany http://www.fhg.de/depts/ims-du-e.html ************************************************************************** From owner-vxwexplo-process Thu Apr 22 04:00:11 1999 From: daemon@csg.lbl.gov Date: Thu Apr 22 04:00:15 PDT 1999 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Thu Apr 22 04:00:08 PDT 1999 Subject: NY-White Plains - VxWorks, C++, Windows 95, GUI Developer needed Subject: WRS information policy (WAS: Zinc discontinues desktop versions) Subject: timing between 2 points in a program Subject: Re: UML tool ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: NY-White Plains - VxWorks, C++, Windows 95, GUI Developer needed Date: Wed, 21 Apr 1999 11:29:53 GMT From: vailj@ibm.net (John Vail) Organization: Advantis Message-ID: <372fb6af.1148953@news-s01.ny.us.ibm.net> Reply-To: vailj@ibm.net Does anyone know of a developer who would be interested in a Contract R&D effort utlilizing VxWorks, C++, and MS Windows GUI? If so, please have them contact me or email a resume. The specs for the project is as follows: - ---------------------------------------------------------------------------------------------------------------------------- Project: C++, VxWorks RealTime programming Job Location: White Plains , New York Area code: (914) Contact: John Vail email: vailj@ibm.net Skills desired: C++, VxWorks, RealTime programming, Unix, RISC processor knowledge, embedded OS, C . Windows, Visual C++, MFC, GUI We are looking to hire a Software Engineer with a strong background in C++, VxWorks development. Candidate would become part of a R&D team developing applications in the areas of Computer Integrated Telephony, IVR and Speech Recognition and Graphics. Knowledge of programming for handheld computers (PalmOS, Geos, Windows CE, etc) and knowledge of VxWorks Internals, kernel, and device driver development is a plus. If you wish to send a resume please email Mr. John Vail at vailj@ibm.net. --------------------------- Newsgroups: comp.os.vxworks,alt.comp.zinc Subject: WRS information policy (WAS: Zinc discontinues desktop versions) Date: Tue, 20 Apr 1999 08:15:20 GMT From: jan.koeppen@bosch.com (Jan Koeppen) Organization: The Bosch Group (Robert Bosch GmbH, Germany) Message-ID: <371c35e2.11934671@news.fr.bosch.de> References: <371436E5.B47DABDB@scanmar.no> <37170910.10562057@news.fr.bosch.de> <371AD8CD.A1345484@scanmar.no> On Mon, 19 Apr 1999 09:18:37 +0200, Ole Asbjorn Fadum wrote: >Hi, > >Yes I received an official letter from Zinc sw, signed Larry Macfarlane >President Zinc Software. The letter contain three topics: >1. Release of Zinc V 5.3 >2. Zinc Desktop Versions Discontinued >3. UK Office Closed. > >Sorry but it is on paper so I could not send you a copy. > Thanks, this information is very good. I am just a little surprised by WRS's information policy as they did not tell us anything about this when they visited us last week. We discussed the problem of customer presentation on a windows machine using the same code in the target and they said nothing. :-( It seems to me that they are more interested in a quick singular commercial success than in a long term cooperation. Jan Koeppen jan.koeppen@bosch.com - ------------------------------------------- contents of this positng are my personal oppinions and does not necessarily conform to those of my employer --------------------------- Newsgroups: comp.os.vxworks Subject: timing between 2 points in a program Date: Tue, 20 Apr 1999 10:33:00 +0100 From: David Rigg Organization: GEC Marconi Avionics Ltd Message-ID: <371C49CC.295D@gecm.com> Reply-To: david.rigg@gecm.com Hi VxWorkers Does anyone know of a function like time() but which gives a finer granularity i.e. gives hundredths of a second? Thanks very much David - -- David Rigg \ _ / Mission Avionics Division, x__________\_(0)_/__________x GEC Marconi Avionics, Scotland @ @ (](_o_)[) @ @ Tel : (+44) (0)131 314 8223 --------------------------- Newsgroups: comp.os.vxworks Subject: Re: UML tool Date: Thu, 22 Apr 1999 08:56:08 +0200 From: "Martin Weiss" Organization: Ericsson Austria AG Message-ID: <7fmg7j$isu@newstoo.ericsson.se> References: <7flbka$p0g2@overload.lbl.gov> 'RationalRose Realtime' from Rational Software supports Tornado http://www.rational.com BR /Martin Jochen Linkohr wrote in message <7flbka$p0g2@overload.lbl.gov>... >Hi, > >vxwexplo@lbl.gov wrote: > >> is there a preferred UML tool to use with Tornado/VxWorks? > >The only thing I know is that ILogix' Rhapsody supports VxWorks as target >OS. I do not know how useful the generated code is. I anybody does so, >please let me know... (by Email..) > >bye > Jochen > --------------------------- End of New-News digest ********************** From owner-vxwexplo-process Thu Apr 22 05:42:11 1999 From: jmenon@lucent.com Date: Thu Apr 22 05:42:15 PDT 1999 Subject: Re: IP-Routing for vxworks ? ISDN/PPP uses IP encapsulated packets. So you only need a MAC layer address translation. You don't really need routing. Just change the MAC layer headers and forward to ethernet driver. As for the software, it is really dependent on how your hardware is setup and if you have any CAMS. From owner-vxwexplo-process Thu Apr 22 08:48:07 1999 From: "Barkley, Charles (HT-EX)" Date: Thu Apr 22 08:48:11 PDT 1999 Subject: VxWorks vs. Linux We are heavy users of VxWorks (several products based on it) and are considering it for use in a new product. However, the application is not embedded nor real-time. I am interested in any comparison studies that have been done comparing VxWorks to Linux for use in non-embedded, non-real-time applications. Actually, ANY study that compares the two would be helpful -- I can weed out the embedded and real-time issues as long as there are other considerations. Thanks, Chuck Barkley General Instrument, Inc. cbarkley@gi.com From owner-vxwexplo-process Thu Apr 22 11:47:09 1999 From: "=?iso-8859-1?B?TWFydO0gUGk=?=" Date: Thu Apr 22 11:47:13 PDT 1999 Subject: RE: UML tool Hi Rob, We are using Rational Rose for the UML part. We are also using Rational Unified Process for the software development process. Best regards, Marti ------------------------------------------------------------ Martí Pi i Puig email: martipi@iac.es Gran Telescopio Canarias (GTC) http://www.gtc.iac.es Control Group Instituto de Astrofísica de Canarias http://www.iac.es c/ Vía Láctea s/n Phone: 34-922-395137 38200 La Laguna (Tenerife) Fax: 34-922-315032 -----Original Message----- From: the vxWorks Users Group Exploder [mailto:vxwexplo@lbl.gov] Sent: martes 20 de abril de 1999 14:49 To: vxworks_users@csg.lbl.gov Subject: UML tool Submitted-by owner-vxwexplo-process Tue Apr 20 06:48:43 1999 Submitted-by: "Wright, Robert B." is there a preferred UML tool to use with Tornado/VxWorks? Rob From owner-vxwexplo-process Thu Apr 22 18:13:55 1999 From: "Mike Anderson" Date: Thu Apr 22 18:13:59 PDT 1999 Subject: Mike Anderson is Leaving SPARTA VxWorks Greetings! For those of you who know me or that I have helped over the last 12 years, this is just a note to let you know that I'm finally leaving SPARTA. I'm leaving to become Chief Scientist for a startup RTOS company known as Integrated Chipware (www.chipware.com). Watch for us at the embedded shows. My new e-mail is: mike@chipware.com For those of you who don't have the slightest idea who I am or don't care ;-), my apologies for the wasted bandwidth. Best regards and good luck in the future... Mike Anderson Chief Scientist Integrated Chipware, Inc. mike@chipware.com (703) 736-3500 From owner-vxwexplo-process Fri Apr 23 04:00:24 1999 From: daemon@csg.lbl.gov Date: Fri Apr 23 04:00:28 PDT 1999 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Fri Apr 23 04:00:20 PDT 1999 Subject: Re: Symbolic links supported on vxworks file system Subject: Re: Seeking 100baseT performance data on VxWorks targets, VME bus Subject: Re: How to "catch" FP exceptions Subject: Re: SRAM checking Subject: Urgent... Lose Weight now - Maigrissez maintenant Subject: Converting GMT time to suitable time in VxWorks. ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Symbolic links supported on vxworks file system Date: 22 Apr 1999 15:53:24 GMT From: Hwa-Jin Bae Organization: PSO Systems Inc (http://www.pso.com) Message-ID: <7fnglk$d13$1@news.idiom.com> References: <7fluoe$qgc$1@schbbs.mot.com> The answer is correctly no ; vxWorks filesystems (rt11, netdrv, dosfs) do not support symlinks. Scott wrote: > Are symbolic links supported? We couldn't find anything in the > documentation, so our feeling is no. However, we'd like to confirm this. > Thanks in advance, > Scott Turley > p27856@email.mot.com - -- Hwa-Jin Bae PSO Systems, Inc. http://www.pso.com mailto:hjb@pso.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Seeking 100baseT performance data on VxWorks targets, VME bus Date: Thu, 22 Apr 1999 12:23:39 -0400 From: Bill West Organization: BWI Message-ID: <371F4D0B.777399BE@bwi.com> References: <7flgso$j78@lindy.stanford.edu> Robert A. Lerche wrote: > I have a client that needs to make a hardware decision. Network > performance on a 100baseT link is a critical requirement. > > We have been very disappointed with the performance of the Heurikon > Baja68K (68060, 66Mhz) -- a TCP throughput test soaks up 100% of the > CPU achieving only about 2Mbytes/second. (Dec 21140 PCI/PMC adapter) > > Can anyone recommend VME bus boards that support higher throughput on > 100baseT? Thank you. We provide an OS9 BSP for the BAJA 68K. I don't have measured throughput numbers, but it should be much better than you are seeing. Let me know if you would like an eval copy for testing. Bill West --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to "catch" FP exceptions Date: Tue, 20 Apr 1999 13:51:43 +0100 From: Kees van der Bent Organization: News Service (http://www.news-service.com/) Message-ID: <371C7858.D694322C@support.nl> References: <371B7BA0.BBF3FF71@bnl.gov> Reply-To: vdbent@support.nl About the code Lawrence wrote: #include #include #include static int restore(int sig){ return ~0; } double protectedCalculate(int i) { jmp_buf env; if(setjmp(env)) return 0.0; signal(SIGFPE, restore); return calculate(i); } A longjmp() call is missing, so even if the signal was caught and restore() invoked, calculate() would simply continue at the place where the exception took place. Apparently VxWorks masks SIGFPE when it occurs (refer to VxWorks manual for confirmation). So when restore() returns, the signal is generated again, but now SIGFPE is masked and your task will be suspended. The signal handler must be returning void: void (*handler)(int). What you must do is unmask SIGFPE, you can do this by calling: signal(SIGFPE, restore); inside restore(), or you could use sigsetmask(). So the following code should work: #include #include #include static jmp_buf jmpBuf; static void handleSigfpe( int sig) { signal(SIGFPE, handleSigfpe); longjmp(jmpBuf); } double protectedCalculate( int i) { if(setjmp(env) != 0) return 0.0; signal(SIGFPE, handleSigfpe); return calculate(i); } But if you'd rather write: #include ... ... ... #include "Except.h" double protectedCalculate int i) { try { calculate(i) } catch (EX_SIGFPE, e) { printf("%s\n", e->getMessage()); return 0.0; } finally; } let me know, and I will send you my free C exception handling package (tell me if you would like gzipped tar, jar or winzip file). All the best, Kees van der Bent vdbent@support.nl --------------------------- Newsgroups: comp.arch.embedded,comp.realtime,comp.os.vxworks,comp.os.psos,comp.os.qnx,comp.sys.powerpc,comp.sys.m68k,comp.os.chorus,comp.os.linux,comp.os.os9 Subject: Re: SRAM checking Date: Thu, 22 Apr 99 18:06:39 GMT From: dom@i-cubed.co.uk (Dom) Organization: i-cubed Message-ID: <7fnoef$b7s$1@shiny.i-cubed.co.uk> References: <371046BC.5DC63B7C@chromatis.com> <3712DDE0.2935@nowhere.tennyson.com.au> <7eusib$d2f$1@thelma.netpower.no> <371301cc.0@seat> <371E8A7D.4D2F@nowhere.tennyson.com.au> <7fmm4f$7e2$1@thelma.netpower.no> In article <7fmm4f$7e2$1@thelma.netpower.no>, "David Brown" wrote: >Tom Evans wrote in message <371E8A7D.4D2F@nowhere.tennyson.com.au>... >>If your data pins are shorted in an embedded product with >>the code memory on the same external bus as the RAM then >>no memory test is going to be able to find AND REPORT >>shorted data bus lines, is it? >> >Good point. On many smaller microcontrollers, however, the program (or at >least a boot program which, among other things, checks the RAM) is on >internal FRAM or EPROM. If the program and/or minimal required RAM is on >the external databus, then you are in trouble if this is shorted. If there >are open connections (pins not soldered properly, for example), then >diagnostic printouts can speed testing and correcting enormously. At a low level, using I/O lines connected to LEDs allows diagnostic information to be put out without using the address or data bus. If you boot from an internal EPROM, you're away. Even if not, there's still a lot you can do: for instance if you are booting from an 8 bit wide EPROM located in high memory before switching to 32 bits, only part of the system needs to be working for a diagnostic test to be run. --------------------------- Newsgroups: comp.os.vxworks Subject: Urgent... Lose Weight now - Maigrissez maintenant Date: Tuesday, 20 Apr 1999 13:46:06 -0600 From: no.email.address.entered@none444.yet Organization: Message-ID: <20049913.4606@none444.yet> Reply-To: gestion_sante@hotmail.com Attention recherche 250 personnes désirent perdre 25 lbs et + possibilité très payantes de votre ordinateur. pour plus d'informations communiquer : gestion_sante@hotmail.com ou par téléphone : 514-990-7393 ext: 7393 - --------------------------------------------------------------------------------------------------------------------------------------------------------------------- Attention seeking 250 individuals whom wish to lose 25 lbs or more - could earn revenus with your computer, inquire at: gestion_sante@hotmail.com or by phone: (514) 990-7393 ext: 7393 --------------------------- Newsgroups: comp.os.vxworks Subject: Converting GMT time to suitable time in VxWorks. Date: Tue, 20 Apr 1999 16:54:39 -0500 From: Norman Dack Organization: Nortel BWA Message-ID: <371CF79E.145ADE45@nortelnetworks.com> This is a multi-part message in MIME format. - --------------CE33B9C4B0C99CC05700F5C2 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit I start with a signed long variable that contains the number of seconds since midnight GMT on January 1, 1900. I wish to get the month, day, year, hours, minutes, and seconds from this number. Do any of you have suggestions as to how I can do this? Thanks. - -- Norman Dack Software Designer Nortel Networks BWA ndack@nortelnetworks.com - --------------CE33B9C4B0C99CC05700F5C2 Content-Type: text/x-vcard; charset=us-ascii; name="vcard.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Norman Dack Content-Disposition: attachment; filename="vcard.vcf" begin: vcard fn: Norman Dack n: Dack;Norman org: Nortel - Broadband Wireless Access adr;dom: 14 Fultz Boulevard;;;Winnipeg;Manitoba;R3Y 1V3; email;internet: ndack@nortelnetworks.com title: Software Designer tel;work: (204) 631-2456 tel;fax: (204) 631-2475 x-mozilla-cpt: ;0 x-mozilla-html: TRUE version: 2.1 end: vcard - --------------CE33B9C4B0C99CC05700F5C2-- --------------------------- End of New-News digest ********************** From owner-vxwexplo-process Fri Apr 23 09:07:38 1999 From: Brett Murphy Date: Fri Apr 23 09:07:41 PDT 1999 Subject: Re: Memory leaks In response to your request for memory tools: Wind River and RTI recently announced the latest Wind Power product for Tornado -- RTI PerformancePak. It includes ScopeProfile - the dynamic execution profiler, and MemScope - the memory analysis tool. MemScope will help you pinpoint memory leaks. I've included the press release below. You can check with your Wind River rep. Best Regards, -- Brett the vxWorks Users Group Exploder wrote: > > Submitted-by owner-vxwexplo-process Wed Apr 21 14:34:47 1999 > Submitted-by: Vasudev Dasari > > Hello, > > Can someone tell me how to dettect memory leaks. Do we have any tools > from Tornado to check them. > > Thanks > Vasu > ============================================================================== = = = = Brett Murphy = Director of Marketing = = Real-Time Innovations, Inc. = email: brett@rti.com = = 155A Moffett Park Drive, Suite 111 = Phone: (408) 720-8312 = = Sunnyvale, CA 94089 = Fax: (408) 734-5009 = = = = ============================================================================== FOR IMMEDIATE RELEASE Press contact: Jenny Horton: (408) 720-8312 x205 jenny@rti.com Tabitha Steager, Wind River Systems, 510 749 2470 tabitha.steager@wrs.com Reader contact: Tel: (408) 720-8312 FAX: (408) 734-5009 http://www.rti.com Real-Time Innovations and Wind River Systems Introduce RTI PerformancePak, the Ultimate Tornado Performance Toolkit PerformancePak for Tornado combines two key performance tools: ScopeProfile(tm) and MemScope(tm). SUNNYVALE, CA, March 8, 1999 - Real-Time Innovations, Inc. and Wind River Systems (NASDAQ: WIND) today announced the availability of RTI PerformancePak(tm). Real-time embedded systems must be fast and reliable. ScopeProfile is a graphical execution profiler that helps Tornado(tm) developers get the highest performance out of their embedded processor. MemScope is a graphical memory analyzer that helps control memory usage and increase reliability of an embedded application. Together, the PerformancePak tools form a powerful toolkit with overwhelming value. Stan Schneider, President of Real-Time Innovations, said, "Visualization is an amazingly powerful capability and visualization tools let the developer see what an application is doing, understand it, and make it work. The PerformancePak tools are critical for any programmer concerned about execution speed or long-term reliability." Avi Schlank, Director of Product Marketing for Wind River Systems added, "StethoScope(tm) is one of our most popular WindPower(tm) tools. With the release of the complementary PerformancePak, RTI has delivered two additional tools that will benefit all our customers. Now, Tornado developers can tune their applications for top performance to ensure that they will run flawlessly in the field. PerformancePak helps make Tornado the highest-performance, most-reliable real-time development environment on the market." About RTI PerformancePak PerformancePak includes the ScopeProfile execution profiler and the MemScope memory analysis tool. Both feature intuitive graphical interfaces, and are intimately integrated into the Tornado development environment. ScopeProfile ScopeProfile helps Tornado developers get the top performance that is critical for real-time systems. ScopeProfile's dynamic display shows exactly where a program is spending time, whether in the code or the operating system. It gives a full, accurate tree profile with no special compilation or hardware. ScopeProfile enables developers to see where the CPU is spending its cycles, understand performance bottlenecks, and make it run fast. Combined with RTI's StethoScope graphical data monitor, ScopeProfile can also display dynamic execution loads. This lets developers quickly isolate bottleneck functions during transient conditions. Dynamic profiling is the key to top performance. MemScope MemScope helps developers ship reliable products. Since real-time programs execute with limited resources for long periods, they must use memory carefully. Memory "leaks" - the gradual loss of small memory blocks - will eventually cause failure. MemScope graphically analyzes memory usage in a running Tornado program and provides a live summary of each allocated block in the system, including a full call-stack history of each memory allocation. MemScope lets Tornado developers see memory leaks, understand memory usage, and use memory efficiently and safely. About Tornado Available for UNIX, Windows 95, and Windows NT host platforms and more than 20 target processor families, Tornado is the industry's most comprehensive development environment for embedded systems. Tornado seamlessly integrates the industry's most powerful development tools and VxWorks(r), a feature-rich, high performance and scalable real-time operating system. Completely open and customizable, Tornado also facilitates the use of third-party and custom-developed tools to enhance its capabilities and meet unique customer requirements. Tornado has more than 280 software and hardware partners and is used by thousands of developers worldwide. Availability RTI PerformancePak for Tornado is currently available for most Tornado configurations. Interested customers can get more information by calling Wind River Systems at 800-545-WIND or 510-748-4100, calling Real-Time Innovations at (408) 720-8312, or visiting http://www.wrs.com or http://www.rti.com. Wind River Systems, Inc. Wind River Systems, Inc., an ISO 9001 registered company, is the leading provider of integrated software development tools for real-time embedded applications in the Internet, telecommunications, data communications, office automation, networking, computer peripherals, medical, automotive, industrial, aerospace/defense and multimedia markets. Through the industry's most comprehensive product line and customer support, Wind River enables customers to develop and reuse application software across a variety of products and platforms, shortening development cycles and improving time-to-market. Incorporated in 1983, Wind River is headquartered in Alameda, California. Wind River reported revenues of $129.4 million in fiscal year 1999, a 40% increase from fiscal year 1998. Real-Time Innovations, Inc. Real-Time Innovations, Inc. (RTI) is a leading developer of new tools and architectures for the growing real-time software market. The company's products help real-time developers analyze and understand embedded systems, speed development of distributed real-time systems, and coordinate the work of teams of programmers developing large projects. The company's products are used in thousands of companies worldwide. ### PerformancePak, ScopeProfile, MemScope and StethoScope are trademarks for Real-Time Innovations. Tornado is a trademark and VxWorks, Wind River Systems, and the Wind River Systems logo are registered trademarks of Wind River Systems, Inc. All other names mentioned are trademarks, registered trademarks or service marks of their respective companies From owner-vxwexplo-process Fri Apr 23 10:29:22 1999 From: Larry Sinn Date: Fri Apr 23 10:29:25 PDT 1999 Subject: Re: Symbolic links supported on vxworks file system --------------470FABE44C9F3518F48AC8E9 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit > The answer is correctly no ; vxWorks filesystems (rt11, netdrv, dosfs) > do not support symlinks. > > Scott wrote: > > Are symbolic links supported? We couldn't find anything in the > > documentation, so our feeling is no. However, we'd like to confirm this. > > > Thanks in advance, > > > Scott Turley > > p27856@email.mot.com > > > > - -- > Hwa-Jin Bae > PSO Systems, Inc. > http://www.pso.com > mailto:hjb@pso.com > The real answer is sort of. If the symbolic and source file are in the same directory, symbolic links work. Larry. Larry Sinn License #: 11561 Spectral Dynamics (408) 474-1746 (voice) 1983 Concourse Drive (408) 474-1780 (fax) San Jose, Ca. 95131-1708 sinnl@sd-star.com (email) --------------470FABE44C9F3518F48AC8E9 Content-Type: text/html; charset=us-ascii Content-Transfer-Encoding: 7bit 
The answer is correctly no ; vxWorks filesystems (rt11, netdrv, dosfs)
do not support symlinks.

Scott  wrote:
> Are symbolic links supported?   We couldn't find anything in the
> documentation, so our feeling is no.  However, we'd like to confirm this.

> Thanks in advance,

> Scott Turley
> p27856@email.mot.com



- -- 
Hwa-Jin Bae         
PSO Systems, Inc. 
http://www.pso.com
mailto:hjb@pso.com


The real answer is sort of.
If the symbolic and source file are in the same directory, symbolic links work.

Larry.

Larry Sinn                      License #: 11561
Spectral Dynamics               (408) 474-1746 (voice)
1983 Concourse Drive            (408) 474-1780 (fax)
San Jose, Ca. 95131-1708        sinnl@sd-star.com (email) --------------470FABE44C9F3518F48AC8E9-- From owner-vxwexplo-process Fri Apr 23 10:34:38 1999 From: Larry Sinn Date: Fri Apr 23 10:34:42 PDT 1999 Subject: Re: Seeking 100baseT performance data on VxWorks targets, VME bus --------------63ADE957017DF5658535D09B Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit > Robert A. Lerche wrote: > > > I have a client that needs to make a hardware decision. Network > > performance on a 100baseT link is a critical requirement. > > > > We have been very disappointed with the performance of the Heurikon > > Baja68K (68060, 66Mhz) -- a TCP throughput test soaks up 100% of the > > CPU achieving only about 2Mbytes/second. (Dec 21140 PCI/PMC adapter) > > > > Can anyone recommend VME bus boards that support higher throughput on > > 100baseT? Thank you. > > We provide an OS9 BSP for the BAJA 68K. I don't have measured > throughput > numbers, but it should be much better than you are seeing. Let me know > if you > would like an eval copy for testing. > > Bill West > We're using Motorla mvme230x boards (603 or 604 PPC and on board 100baseT) and get good performance (11Mbytes/sec between the 230x and a Sun on a dedicated network or a crossover cable). The 230x boards are simple (1 console port, 1 ethernet port and 2 PMC sites) and inexpensive. Larry. Larry Sinn License #: 11561 Spectral Dynamics (408) 474-1746 (voice) 1983 Concourse Drive (408) 474-1780 (fax) San Jose, Ca. 95131-1708 sinnl@sd-star.com (email) --------------63ADE957017DF5658535D09B Content-Type: text/html; charset=us-ascii Content-Transfer-Encoding: 7bit 

Robert A. Lerche wrote:

> I have a client that needs to make a hardware decision.  Network
> performance on a 100baseT link is a critical requirement.
>
> We have been very disappointed with the performance of the Heurikon
> Baja68K (68060, 66Mhz) -- a TCP throughput test soaks up 100% of the
> CPU achieving only about 2Mbytes/second.  (Dec 21140 PCI/PMC adapter)
>
> Can anyone recommend VME bus boards that support higher throughput on
> 100baseT?  Thank you.

  We provide an OS9 BSP for the BAJA 68K.  I don't have measured
throughput
numbers, but it should be much better than you are seeing.  Let me know
if you
would like an eval copy for testing.

Bill West


We're using Motorla mvme230x boards (603 or 604 PPC and on board 100baseT) and get good performance (11Mbytes/sec between the 230x and a Sun on a dedicated network or a crossover cable). The 230x boards are simple (1 console port, 1 ethernet port and 2 PMC sites) and inexpensive.

Larry.

Larry Sinn                      License #: 11561
Spectral Dynamics               (408) 474-1746 (voice)
1983 Concourse Drive            (408) 474-1780 (fax)
San Jose, Ca. 95131-1708        sinnl@sd-star.com (email)
 
  --------------63ADE957017DF5658535D09B-- From owner-vxwexplo-process Fri Apr 23 12:50:27 1999 From: "Thomas Keith Buchanan" Date: Fri Apr 23 12:50:31 PDT 1999 Subject: Re: Seeking 100baseT performance data on VxWorks targets, VME bus > We have been very disappointed with the performance of the Heurikon > Baja68K (68060, 66Mhz) -- a TCP throughput test soaks up 100% of the > CPU achieving only about 2Mbytes/second. (Dec 21140 PCI/PMC adapter) That does not jive with some tests I have done. I've seen the Baja68K run at over 4MBytes/s over the backplane network consuming arounf 30% of the CPU bandwidth. Since the 21140 uses DMA, the numbers you are seeing appear low. Have you varies the size of the socket buffers? Are you doing some sort of proprietary processing on either end? Are you using the sockets in blocking mode? > Can anyone recommend VME bus boards that support higher throughput on > 100baseT? Thank you. Given that the NIC hardware drivers are in good shape, TCP throughput is increases linearly with the intger performance of the CPU. As usual, the answer is get more horspower o;). The new SENS stack is rumored to be a little faster. I'm quite familiar with VxWorks and OS-9, there is no reason that OS-9 would be faster. In fact, the I/O subsystem in VxWorks is a little simpler and thus faster. HTH ------------------------------------ /\ Thomas Keith Buchanan /**\ Principal Engineer /****\ SPARTA, Inc. /****/.. 205 East IH 30 /****/.... Rockwall, TX 75087 .\****\__... ...\**\ /.... Voice: 972 772 4487 ....\**\/\.... FAX: 972 722 3201 ....\****\.. PCS: 972 672 9657 ....\****/ WWW: www.sparta.com .../***/ ./***/ SPARTA \**/ ====== \/ Pride In Performance From owner-vxwexplo-process Fri Apr 23 16:04:18 1999 From: Charlie Grames Date: Fri Apr 23 16:04:21 PDT 1999 Subject: Re: Converting GMT time to suitable time in VxWorks. Norman, Unless you are only interested in dates up to 1968, you can't. A signed long cannot store a large enough value to represent more than about 68 years. That is why UNIX and VxWorks do not have a year 2000 problem but DO have a year 2038 problem (because the baseline time in the implementation is January 1, 1970, 0000). I don't know how the UNIX community plans to deal with this problem. Perhaps they are hoping that (at least) 64-bit implementations will be enough commonplace by then that a signed 64-bit value can be used in place of the current signed 32-bit. Anyway, what UNIX does in 2038 doesn't help you now. One possible solution would be to use a structure with two long int fields: one that represents days elapsed since 1900, and the other that represents seconds elapsed within the current day. This would carry you until mankind is probably no longer in existence, but you're still limited when translating into the VxWorks scheme. I hope this helps. Charlie Grames The Boeing Company (314) 233-1956 Charles.R.Grames@boeing.com ----- Newsgroups: comp.os.vxworks Subject: Converting GMT time to suitable time in VxWorks. Date: Tue, 20 Apr 1999 16:54:39 -0500 From: Norman Dack Organization: Nortel BWA Message-ID: <371CF79E.145ADE45@nortelnetworks.com> This is a multi-part message in MIME format. - --------------CE33B9C4B0C99CC05700F5C2 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit I start with a signed long variable that contains the number of seconds since midnight GMT on January 1, 1900. I wish to get the month, day, year, hours, minutes, and seconds from this number. Do any of you have suggestions as to how I can do this? Thanks. - -- Norman Dack Software Designer Nortel Networks BWA ndack@nortelnetworks.com From owner-vxwexplo-process Sat Apr 24 04:00:28 1999 From: daemon@csg.lbl.gov Date: Sat Apr 24 04:00:32 PDT 1999 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sat Apr 24 04:00:25 PDT 1999 Subject: Re: Converting GMT time to suitable time in VxWorks. Subject: Re: cacheDmaMalloc on mv2306 Subject: Re: cacheDmaMalloc on mv2306 ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Converting GMT time to suitable time in VxWorks. Date: Wed, 21 Apr 1999 08:33:57 -0400 From: "Bob Beers" Organization: CyberGate, Inc. Message-ID: <7fkgl8$2lc8$1@news.gate.net> References: <371CF79E.145ADE45@nortelnetworks.com> You might try ansiTime function gmtime(). It converts calendar time in seconds to vxworks "broken-down time" expressed as UTC. The structure for broken-down time is also listed in the ansiTime section (pg. 1-24 of my VxWorks Reference Manual 5.3.1) Bob Norman Dack wrote in message <371CF79E.145ADE45@nortelnetworks.com>... >I start with a signed long variable that contains the number of seconds >since midnight GMT on January 1, 1900. I wish to get the month, day, >year, hours, minutes, and seconds from this number. Do any of you have >suggestions as to how I can do this? > >Thanks. > >-- >Norman Dack >Software Designer >Nortel Networks BWA >ndack@nortelnetworks.com > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: cacheDmaMalloc on mv2306 Date: 23 Apr 1999 17:07:30 -0700 From: Christopher Ogrady Organization: Stanford Linear Accelerator Center Message-ID: References: Thanks Scott for your reply: >I 'believe' that the Moto MV2xxx PowerPC architectures >come under the category of 'perfect harvard architecture' >as described in the cacheLib manual section. Hmmm. I don't know alot about these things, but I thought harvard architecture meant just split instruction and data caches. Is there something else in this architecture that guarantees cache-coherency? I re-read the cacheLib section of the manual, but didn't see anything about this. >I havent tried turning off snooping to see if the macros return >false. I should have been more specific in my original post. My impression is that the *default* behaviour for the 2306 is to NOT have hardware snooping. I say this because I look at the page-table entries (for the dec21140 ethernet driver buffers, for example) and the "M" bit is turned OFF. "I" (cache inhibiting) is also set to 0, so the cache is active as well. (it's my understanding that the "M" bit being set to 1, plus the GBL* line on the MPC bus are what guarantee the coherency, by enabling the ppc604 bus snooping hardware). So the powerPC hardware mechanism for cache coherency (bus snooping) seems to be disabled, so far as I can tell. I believe that the cache is indeed coherent (otherwise the dec21140 driver wouldn't work). I just don't understand what enforces the coherency. If anyone has some insight into this, I'd be very interested. Thanks! chris ====================================================================== Christopher O'Grady Stanford Linear Accelerator Center cpo@slac.stanford.edu 650-926-4282 --------------------------- Newsgroups: comp.os.vxworks Subject: Re: cacheDmaMalloc on mv2306 Date: 23 Apr 1999 19:11:31 -0700 From: Christopher Ogrady Organization: Stanford Linear Accelerator Center Message-ID: References: I think that by reading the powerPC manuals some more I've answered my own question. I'm posting it here in case anybody else ever spends time wondering about this: The key bit of information I was missing was that the "M" bit in the powerPC memory attributes is the signal for the MPC bus master to set the GBL* line. I had (mistakenly) thought that it was needed to enable the 604 bus snooper. Since the ravenMpic has GBL* set for *every* transaction, the snooper picks it up automatically. It would only need to be set if, for example, there were two CPU's on the MPC bus with two caches that needed to be kept coherent. So the final conclusion for this thread is (if anyone cares): on the mv2306 cacheDmaMalloc only has to do a normal malloc because the hardware bus snooping (responding to GBL*) takes care of the cache coherency. chris ====================================================================== Christopher O'Grady Stanford Linear Accelerator Center cpo@slac.stanford.edu 650-926-4282 --------------------------- End of New-News digest ********************** From owner-vxwexplo-process Sun Apr 25 04:00:11 1999 From: daemon@csg.lbl.gov Date: Sun Apr 25 04:00:15 PDT 1999 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sun Apr 25 04:00:08 PDT 1999 Subject: Re: GNU Compiler for Windows and/or MSDOS Subject: Re: Mil-std 1553 driver ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: GNU Compiler for Windows and/or MSDOS Date: Sat, 24 Apr 1999 20:15:32 GMT From: David Hamel Organization: @Home Network Message-ID: <3722268D.E1C6A245@ntanet.net> References: <7ficmt$p0g1@overload.lbl.gov> djgpp will have a hard time producing 68k binaries. Robert Armstrong wrote: > Try www.delorie.com, there you will find djgpp which I use. If someone gets > through the Cygnus site and finds the URL where I can actually get theirs I'd > be very pleased to see it posted here (the URL, not the code). > > One of my friends asked about building from the Gnu source provided by Wind > River and was told that building for Solaris or HPUX is trivial but compiling > for win32 is an art form. Wind River don't short change us on source, you get > what they use, it's just very involved... I decided not to bother :) > > Bob > > In article <7ficmt$p0g1@overload.lbl.gov>, "Jim Hawthorne" > wrote: > >VxWorkers- > > > >I am looking for a version of the GNU Compiler (or any other free compiler > >for that matter) that will run on a laptop running Windows 98 and compile > >68K code for Motorola MVME162s and MVME167s running VxWorks. > > > >Does anybody know of such a package? > > > >Jim Hawthorne > >Applied Technology Associates > >Albuquerque, New Mexico > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Mil-std 1553 driver Date: Sun, 25 Apr 1999 11:10:45 +0100 From: Mark Durham Message-ID: References: <7eng69$fcv$1@nnrp1.dejanews.com> <371BF262.C7DBF0DD@earthlink.net> The new hybrid chips from ILC-DDC are quite complicated beasts. Check out DDCs website as suggested but also try and get hold of the Mil-Std- 1553 Designers Guide - also from DDC - I've got edition 5. It explains all about 1553 including busses, bus controllers, remote terminals, monitors, and the 1553 protocol. Don't go too deep into the protocol stuff - just get a feel of what 16-bit words go out onto the 1553 bus for RT-BC, BC-RT & RT-RT transfers - - there are a couple of good figures in the book that show the transfers and the order of command words, status words and data. You don't have to sort out the protocol itself as the chip does all that for you! The problem with starting out on 1553 is that it's not cheap (compared to PC ethernet) and because it's a data bus, you will need something to talk to whilst you develop your code ( an RT if you are going to be to BC or a BC if you going to be the RT ) and something to monitor the data bus. There are several PC cards out the that do BC, RT or BM and some that are "full function" i.e. they can do all three at once. We use the AIM PBA cards under DOS. Also make sure that your data bus is set up correctly - the easy way round this is to get hold of a bus coupler that has several stubs. That way you only need to connect terminators to either end of the coupler. In article <371BF262.C7DBF0DD@earthlink.net>, Anthony Wilson writes >imohammed@mailexcite.com wrote: > >> I am working on a project that uses 1553 protocol. I am trying to write a >> driver for 1553 under Vxworks. I am new to Vxworks and 1553. Is there any >> documentation or books about 1553 driver development? >> >> Thanks >> Irfan Mohammed >> >> -----------== Posted via Deja News, The Discussion Network ==---------- >> http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own > >try here http://www.ilcddc.com/ > > - -- - --------------------------------------------------------------------------- | Mark Durham | | Home: mark@rosedene.demon.co.uk (Weekends) | | Work: mjdurham@dera.gov.uk | - --------------------------------------------------------------------------- --------------------------- End of New-News digest ********************** From owner-vxwexplo-process Mon Apr 26 04:00:21 1999 From: daemon@csg.lbl.gov Date: Mon Apr 26 04:00:24 PDT 1999 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Mon Apr 26 04:00:17 PDT 1999 Subject: Looking for ATM card supported for x86 platform Subject: Can DRAM and SRAM corroborate Subject: VME CPU and Data Aquisition CHEAP! PPC and 68K Many Boards in Many Configurations ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Looking for ATM card supported for x86 platform Date: Thu, 22 Apr 1999 11:00:59 -0400 From: "Tom Truman" Organization: Lucent Technologies, Columbus, Ohio Message-ID: <7fnee1$3ht@nntpb.cb.lucent.com> I'm looking for an ATM card that is supported on the x86 platform running Tornado. Any help is appreciated. Thanks, Tom --------------------------- Newsgroups: comp.os.vxworks Subject: Can DRAM and SRAM corroborate Date: Mon, 26 Apr 1999 11:06:44 +0800 From: "²éöÁº£" Organization: Alcatel/Bell Message-ID: <3723d8f5.0@dnews.sbell.com.cn> Hi, everybody: There are both 8MB DRAM and 1MB SRAM on my 860 board. Now the problem is when the DRAM_BASE_ADRS and LOCAL_MEM_BASE_ADRS is set to 0x0000_0000 and SRAM_BASE_ADRS is set to 0x1000_0000, everything is OK and vxworks can start correctly, but for the addrss is not continuous, I set the LOCAL_MEM_SIZE to 0x0800_0000, then memory usage in browser shows 0x0800_0000. When I change SRAM_BASE_ADRS to 0x0800_0000 and LOCAL_MEM_SIZE to 0x0900_0000, I can start the serial and load the vxworks from host, but that is all, the vxworks can't start. In both cases mentioned above the RAM_HIGH_ADRS is set to 0x0020_0000 and RAM_LOW_ADRS is set to 0x0001_0000. All your answer will be appreciated, thank you very much. Best Regards Zha Junhai --------------------------- Newsgroups: misc.forsale.computers.net-hardware,comp.arch.bus.vmebus,comp.sys.m68k,comp.arch.embedded,comp.os.vxworks,comp.realtime,comp.sys.powerpc.misc Subject: VME CPU and Data Aquisition CHEAP! PPC and 68K Many Boards in Many Configurations Date: Sun, 25 Apr 1999 23:07:40 -0500 From: "Sun Harbor Sales" Message-ID: <3723e6c3@tejani.ghg.net> We're selling most of the boards below at 30% list or less. Everything is guaranteed 30 days to be 100% functional. Click on the link below for a listing of Motorola VME PPC and 68K... http://www.ghg.net/wmcalpin/shsvmefs.htm Click on the link below for a listing of VMIC data aquisition boards, D/A, A/D, D/D, driver, etc... http://www.ghg.net/wmcalpin/vmivmefs.htm --------------------------- End of New-News digest ********************** From owner-vxwexplo-process Mon Apr 26 06:30:01 1999 From: Charlie Grames Date: Mon Apr 26 06:30:06 PDT 1999 Subject: Re: Converting GMT time to suitable time in VxWorks. -Reply Scott, I stand corrected. VxWorks indeed defines time_t as "unsigned long", so the "2038 problem" that will occur in some UNIX implementations is postponed until 2106 for VxWorks. Thanks for pointing out my mistake. Charlie Grames The Boeing Company (314) 233-1956 Charles.R.Grames@boeing.com >>> Scott Kamara 04/23/99 07:34pm >>> Charlie, With regards to vxWorks having a year 2038 problem: Under vxWorks 5.3.X, time_t is typed as 'unsigned long'. My time-of-day support (clock_settime(),time(),gmtime()) is working through the year 2105. (time() is returning 'secs from 1/1/70') Charlie Grames wrote: > > Norman, > > Unless you are only interested in dates up to 1968, you can't. A signed > long cannot store a large enough value to represent more than about 68 > years. That is why UNIX and VxWorks do not have a year 2000 problem but DO > have a year 2038 problem (because the baseline time in the implementation is > January 1, 1970, 0000). > > I don't know how the UNIX community plans to deal with this problem. > Perhaps they are hoping that (at least) 64-bit implementations will be > enough commonplace by then that a signed 64-bit value can be used in place > of the current signed 32-bit. > > Anyway, what UNIX does in 2038 doesn't help you now. One possible solution > would be to use a structure with two long int fields: one that represents > days elapsed since 1900, and the other that represents seconds elapsed > within the current day. This would carry you until mankind is probably no > longer in existence, but you're still limited when translating into the > VxWorks scheme. > > I hope this helps. > > Charlie Grames > The Boeing Company > (314) 233-1956 > Charles.R.Grames@boeing.com > > ----- > > Newsgroups: comp.os.vxworks > Subject: Converting GMT time to suitable time in VxWorks. > Date: Tue, 20 Apr 1999 16:54:39 -0500 > From: Norman Dack > Organization: Nortel BWA > Message-ID: <371CF79E.145ADE45@nortelnetworks.com> > > This is a multi-part message in MIME format. > - --------------CE33B9C4B0C99CC05700F5C2 > Content-Type: text/plain; charset=us-ascii > Content-Transfer-Encoding: 7bit > > I start with a signed long variable that contains the number of seconds > since midnight GMT on January 1, 1900. I wish to get the month, day, > year, hours, minutes, and seconds from this number. Do any of you have > suggestions as to how I can do this? > > Thanks. > > - -- > Norman Dack > Software Designer > Nortel Networks BWA > ndack@nortelnetworks.com -- ----------------------- Scott Kamara AP Data Systems Seattle Engr. Office (206) 525-8592 scott@aplabs.com From owner-vxwexplo-process Mon Apr 26 06:37:23 1999 From: "Richard B Asen" Date: Mon Apr 26 06:37:27 PDT 1999 Subject: VxWorks - TFFS and multiple partitions Has anyone tried to set up multiple "DOS Partitions" in a TFFS? According to Wind River its not possible. (I know the DOS File System does not support FDISK) I need to setup two distinct areas (partitions or blocks) in the flash system. The first, call it "C:" , contains code that is never changed or erased, the second, call it "D:", contains code that must be erasable. I don't need to know where the files are in D: but I must be able to erase all of D: without touching C: Is this possible? Richard B. Asen Raytheon Systems Company richard_b_asen@raytheon.com 410-583-4477 From owner-vxwexplo-process Mon Apr 26 07:14:17 1999 From: mmenge@dspt.com Date: Mon Apr 26 07:14:20 PDT 1999 Subject: Help! we are under attack by the (tNetTask): dc0: no carrier Mark Menge@DSPT 04/26/99 10:13 AM Help! we have several (3) customers that run powerpc mvme2604 cards that are complain about 0xbcc250 (tNetTask): dc0: no carrier messages. This message is accompanied by a temporary loss of ethernet communications. 1. Has anyone seen this message on any other architecture? what was done to fix it? 2. Does everyone who runs a 2604 with heavy ethernet usage see this message? 3. What does this message really mean? where should I look to get more technical information? Any information would be highly appreciated. vxWorks tornado spam -> (tNetTask): dc0: no carrier (tNetTask): dc0: no carrier (tNetTask): dc0: no carrier From owner-vxwexplo-process Mon Apr 26 11:10:18 1999 From: Charlie Grames Date: Mon Apr 26 11:10:21 PDT 1999 Subject: Re: Help! we are under attack by the (tNetTask): dc0: no Mark, I have seen this message when attempting to use the dc Ethernet device with a switch that auto-negotiates full-duplex operation. The hardware will happily auto-negotiate full duplex, but, by default, the software is configured for half duplex only. We have solved this problem by setting the switch to explicit half-duplex operation. If you want to use full duplex, you need to set DC_MODE in config.h to 0x18. HTH Charlie Grames The Boeing Company (314) 233-1956 Charles.R.Grames@boeing.com >>> the vxWorks Users Group Exploder 04/26/99 09:14am >>> Submitted-by owner-vxwexplo-process Mon Apr 26 07:14:17 1999 Submitted-by: mmenge@dspt.com Mark Menge@DSPT 04/26/99 10:13 AM Help! we have several (3) customers that run powerpc mvme2604 cards that are complain about 0xbcc250 (tNetTask): dc0: no carrier messages. This message is accompanied by a temporary loss of ethernet communications. 1. Has anyone seen this message on any other architecture? what was done to fix it? 2. Does everyone who runs a 2604 with heavy ethernet usage see this message? 3. What does this message really mean? where should I look to get more technical information? Any information would be highly appreciated. vxWorks tornado spam -> (tNetTask): dc0: no carrier (tNetTask): dc0: no carrier (tNetTask): dc0: no carrier From owner-vxwexplo-process Mon Apr 26 13:49:18 1999 From: Roger Cruz Date: Mon Apr 26 13:49:22 PDT 1999 Subject: Socket error number discrepancy? I've ported some code that uses the select() on a socket and checks the result and errno. In one case, it is looking for EBADF (bad file number) according to errno.h. When I run the code, however, all checks fail and errno is set to S_iosLib_INVALID_FILE_DESCRIPTOR. The descriptor was a good socket used prior to calling select, but very likely, it was closed when the connection was broken. My question is whether I should consider S_iosLib_INVALID_FILE_DESCRIPTOR equivalent to EBADF? Why doesn't the select() routine return with the appropriate BSD errno value? Thanks. R. PS: Remove "_nojunk" from email address vxworks, VxWorks -- Roger Cruz cruz_nojunk@xyplex.com Xyplex Networks w: 978-952-4783 295 Foster Street f: 978-952-4887 Littleton, MA 01460 From owner-vxwexplo-process Tue Apr 27 03:46:33 1999 From: "Tony Mayes" Date: Tue Apr 27 03:46:37 PDT 1999 Subject: Pointer Verification Hi all, Has anyone come across a function or mechanism to easily check that a pointer value is valid, ie will not generate a PAGE FAULT. I guess vxworks must have something to be able to generate the error but I can't find a function to do this. Thanks Tony Mayes (The opinions expressed within are my own) Email address tony.mayes@epid.eurotherm.co.uk Phone (0044)(0)1903 205277 Fax (0044)(0)1903 524016 Eurotherm Process Instrumentation Div, Southdownview Way, Worthing, Sussex, UK, BN14 8NN From owner-vxwexplo-process Tue Apr 27 04:00:16 1999 From: daemon@csg.lbl.gov Date: Tue Apr 27 04:00:19 PDT 1999 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Tue Apr 27 04:00:13 PDT 1999 Subject: Re: DiskOnChip2000 ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: DiskOnChip2000 Date: Mon, 26 Apr 1999 07:49:33 -0700 From: "Dave Craig" Organization: Verio Northern California's Usenet News Service Message-ID: <7g1ueu$met$1@news.ncal.verio.com> References: <7fqo8t$qc0$1@news.gate.net> <37210432.1938153@news.texas.net> >You have to purchase the VxWorks True Flash File system. It will give you >usrTffsConfig() No you don't - DO NOT buy True Flash File system for Disk On Chip 2000 it has a built in flash filing system. Attaching to the drive using ATA config is the right way to go. Dave Craig Palo Alto. --------------------------- End of New-News digest ********************** From owner-vxwexplo-process Tue Apr 27 08:04:36 1999 From: Vasudev Dasari Date: Tue Apr 27 08:04:39 PDT 1999 Subject: GCC Compiler ... Hi, We are using GCC compiler for our application in vxWorks on MIPS platform. I found some strange results. "socket" system call is taking around 300 bytes and that memory is not released. "inet_ntoa()" is also taking around 22 bytes and that memory is not released. I have instantiated every object on stack. Program is like this : void ms() { struct sockaddr_in serverAddr; int sockAddrSize; int sFd; memShow(0); // Here around 300 bytes is used by socket system call, which is not released even after the scope of this sFD goes away. if ((sFd = socket (AF_INET, SOCK_STREAM, 0)) == ERROR) { printf ("socket"); return (ERROR); } memShow(0); sockAddrSize = sizeof (struct sockaddr_in); bzero ((char *) &serverAddr, sockAddrSize); serverAddr.sin_family = AF_INET; serverAddr.sin_port = ntohs (2000); memShow(0); serverAddr.sin_addr.s_addr = inet_addr ("127.1.0.0"); memShow(0); char *p = inet_ntoa(serverAddr.sin_addr); memShow(0); if(p) delete p; // I have to delete this, it is not an error, if I don't do it is amemory leak. I think it is not the same on on SPARC compiler memShow(0); We have an application that is working on Sparc with no memory leaks. I changed that app to run on vxWorks on MIPS platoform. Only sstem calls will be different between these two, which I have taken care of. I am finding some memory leaks in my app. Is it some thing to do with the memory is allocated and released by GCC compiler or where could be the problem. Please help me ... Thanks Vasu From owner-vxwexplo-process Tue Apr 27 15:17:08 1999 From: Charlie Grames Date: Tue Apr 27 15:17:12 PDT 1999 Subject: Re: GCC Compiler ... Vasu, The shortcomings of inet_ntoa() are discussed in the VxWorks Reference Manual. You should use inet_ntoa_b() with VxWorks, instead. As far as the socket is concerned, I don't see where you close it. Unlike Unix, VxWorks does not automatically close open file descriptors when your code stops executing. You are responsible for it. (That's good programming practice, anyway). HTH Charlie Grames The Boeing Company (314) 233-1956 Charles.R.Grames@boeing.com >>> the vxWorks Users Group Exploder 04/27/99 10:04am >>> Submitted-by owner-vxwexplo-process Tue Apr 27 08:04:36 1999 Submitted-by: Vasudev Dasari Hi, We are using GCC compiler for our application in vxWorks on MIPS platform. I found some strange results. "socket" system call is taking around 300 bytes and that memory is not released. "inet_ntoa()" is also taking around 22 bytes and that memory is not released. I have instantiated every object on stack. Program is like this : void ms() { struct sockaddr_in serverAddr; int sockAddrSize; int sFd; memShow(0); // Here around 300 bytes is used by socket system call, which is not released even after the scope of this sFD goes away. if ((sFd = socket (AF_INET, SOCK_STREAM, 0)) == ERROR) { printf ("socket"); return (ERROR); } memShow(0); sockAddrSize = sizeof (struct sockaddr_in); bzero ((char *) &serverAddr, sockAddrSize); serverAddr.sin_family = AF_INET; serverAddr.sin_port = ntohs (2000); memShow(0); serverAddr.sin_addr.s_addr = inet_addr ("127.1.0.0"); memShow(0); char *p = inet_ntoa(serverAddr.sin_addr); memShow(0); if(p) delete p; // I have to delete this, it is not an error, if I don't do it is amemory leak. I think it is not the same on on SPARC compiler memShow(0); We have an application that is working on Sparc with no memory leaks. I changed that app to run on vxWorks on MIPS platoform. Only sstem calls will be different between these two, which I have taken care of. I am finding some memory leaks in my app. Is it some thing to do with the memory is allocated and released by GCC compiler or where could be the problem. Please help me ... Thanks Vasu From owner-vxwexplo-process Wed Apr 28 04:03:25 1999 From: daemon@csg.lbl.gov Date: Wed Apr 28 04:03:29 PDT 1999 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Wed Apr 28 04:03:22 PDT 1999 Subject: Re: Converting GMT time to suitable time in VxWorks. Subject: Re: vprintf on the PowerPC based system ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Converting GMT time to suitable time in VxWorks. Date: Fri, 23 Apr 1999 17:34:01 -0700 From: Scott Kamara Organization: AP Labs Message-ID: <37211179.6AFC550D@pop.sttl.uswest.net> References: Reply-To: spk@uswest.net Charlie, With regards to vxWorks having a year 2038 problem: Under vxWorks 5.3.X, time_t is typed as 'unsigned long'. My time-of-day support (clock_settime(),time(),gmtime()) is working through the year 2105. (time() is returning 'secs from 1/1/70') Charlie Grames wrote: > > Norman, > > Unless you are only interested in dates up to 1968, you can't. A signed > long cannot store a large enough value to represent more than about 68 > years. That is why UNIX and VxWorks do not have a year 2000 problem but DO > have a year 2038 problem (because the baseline time in the implementation is > January 1, 1970, 0000). > > I don't know how the UNIX community plans to deal with this problem. > Perhaps they are hoping that (at least) 64-bit implementations will be > enough commonplace by then that a signed 64-bit value can be used in place > of the current signed 32-bit. > > Anyway, what UNIX does in 2038 doesn't help you now. One possible solution > would be to use a structure with two long int fields: one that represents > days elapsed since 1900, and the other that represents seconds elapsed > within the current day. This would carry you until mankind is probably no > longer in existence, but you're still limited when translating into the > VxWorks scheme. > > I hope this helps. > > Charlie Grames > The Boeing Company > (314) 233-1956 > Charles.R.Grames@boeing.com > > ----- > > Newsgroups: comp.os.vxworks > Subject: Converting GMT time to suitable time in VxWorks. > Date: Tue, 20 Apr 1999 16:54:39 -0500 > From: Norman Dack > Organization: Nortel BWA > Message-ID: <371CF79E.145ADE45@nortelnetworks.com> > > This is a multi-part message in MIME format. > - --------------CE33B9C4B0C99CC05700F5C2 > Content-Type: text/plain; charset=us-ascii > Content-Transfer-Encoding: 7bit > > I start with a signed long variable that contains the number of seconds > since midnight GMT on January 1, 1900. I wish to get the month, day, > year, hours, minutes, and seconds from this number. Do any of you have > suggestions as to how I can do this? > > Thanks. > > - -- > Norman Dack > Software Designer > Nortel Networks BWA > ndack@nortelnetworks.com - -- - ----------------------- Scott Kamara AP Data Systems Seattle Engr. Office (206) 525-8592 scott@aplabs.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: vprintf on the PowerPC based system Date: Tue, 27 Apr 1999 21:47:51 -0400 From: "Jeff Szczepanski" Organization: Road Runner - NY Regional Operations Center Message-ID: References: <37250E78.EC6424BA@terayon.com> Are you using the GNU compiler with the "software floating point option"? I forget the exact option..."-msoft-math", maybe. The VxWorks documentation tells you that it is supported, but it has problems in certain circumstances if the FPU is disabled...I have been meaning to write up a problem report on this one, but haven't got around to it. If so, this blows up calls into the VxWorks libraries for variable argument list function calls. (Anything with func(x,y,...)). This happens because of what I believe to be non-EABI compliant behavior of the gcc compiler. Drop me a note if you want a more detailed description. Regards, Jeff PS: Another possibility for you problem is stack overflow. All the printf like routines use a fair amount of stack space....try checkStack() to verify stack sizes and to check for overflow. =========================== Jeffrey R. Szczepanski InSciTek Microsystems, Inc. 635 CrossKeys Office Park Fairport, NY 14450 Phone: 716-421-3850 sasha wrote in message <37250E78.EC6424BA@terayon.com>... Hi, My PowerPC system crashes on the calls to vprintf or vsprintf function. Does somebody know what the problem can be ? I guess it is problem of the BSP configuration. Thanks - -- ***************************************************************** * Alexander Fertelmeister email: sasha@terayon.com * * Terayon Corporation home: * * 2952 Bunker Hill Lane voice: 408 486-5275 * * Santa Clara CA, 95054 fax: 408 727-6204 * ***************************************************************** --------------------------- End of New-News digest ********************** From owner-vxwexplo-process Wed Apr 28 07:57:05 1999 From: Kevin Kernin Date: Wed Apr 28 07:57:09 PDT 1999 Subject: vxWorks in ROM Hi VxWorks Users, We are using Tornado 1.0.1 and vxworks 5.3.1 with a MVME2306 board. I am looking into putting our vxWorks kernel in ROM and running it from RAM (later I'll see if we can fit our application in ROM). I ran make vxWorks.st_rom and it died with the following error: VX_CPU_FAMILY=ppc /pkgs/wind_1.1.4plus/host/sun4-solaris2/bin/makeSymTbl tmp.o >symTbl.c ccppc -B/pkgs/wind_1.1.4plus/host/sun4-solaris2/lib/gcc-lib/ -c -traditional -mstrict-align -ansi -nostdinc -O2 -fvolatile -fno-builtin -fno-for-scope -Wall -I/h -I. -I/pkgs/wind_1.1.4plus/target/config/all -I/pkgs/wind_1.1.4plus/target/h -I/pkgs/wind_1.1.4plus/target/src/config -I/pkgs/wind_1.1.4plus/target/src/drv -DCPU=PPC604 -DMV2300 -DTARGET_DIR="\"mv2304\"" symTbl.c symTbl.c:88: nondigits in number and not hexadecimal symTbl.c:88: nondigits in number and not hexadecimal symTbl.c:88: numeric constant contains digits beyond the radix symTbl.c:88: syntax error before `0000f6f8' make: *** [vxWorks.st.Z.s] Error 1 Below is an excerpt of symTbl.c starting with line 88. 0000f6f8 R WTX_EVT_CALL_GOT_EXC 0000f6ec R WTX_EVT_CALL_RETURN 0000f68c R WTX_EVT_CTX_EXIT 0000f680 R WTX_EVT_CTX_START 0000f6e0 R WTX_EVT_DATA_ACCESS 0000f698 R WTX_EVT_EXCEPTION 0000f664 R WTX_EVT_OBJ_LOADED 0000f670 R WTX_EVT_OBJ_UNLOADED 0000f64c R WTX_EVT_SYM_ADDED 0000f658 R WTX_EVT_SYM_REMOVED 0000f6d4 R WTX_EVT_TEXT_ACCESS 0000f640 R WTX_EVT_TGT_RESET 0000f6b0 R WTX_EVT_TOOL_ATTACH 0000f6bc R WTX_EVT_TOOL_DETACH 0000f6c8 R WTX_EVT_TOOL_MSG 0000f710 R WTX_EVT_UNKNOWN 0000f708 R WTX_EVT_USER 0000f6a4 R WTX_EVT_VIO_WRITE IMPORT Write__6RWFilePCUwUi (); Any ideas what I'm doing wrong? Thanks, Kevin ================================ Kevin Kernin Moore Research Center, Inc. (716)773-0658 FAX : (716)773-0210 email: kernin@research.moore.com ================================= From owner-vxwexplo-process Thu Apr 29 01:52:41 1999 From: David Laight Date: Thu Apr 29 01:52:45 PDT 1999 Subject: Re: vxWorks in ROM > Subject: vxWorks in ROM > > Submitted-by: Kevin Kernin > > Hi VxWorks Users, > > We are using Tornado 1.0.1 and vxworks 5.3.1 with a MVME2306 board. > I am looking into putting our vxWorks kernel in ROM and running it from > RAM (later I'll see if we can fit our application in ROM). > I ran make vxWorks.st_rom and it died with the following error: > > symTbl.c:88: nondigits in number and not hexadecimal > > ... > > Any ideas what I'm doing wrong? I guess that the code that generates symtbl.c is stuffed! However I wonder if you need all the symbols? On my system the symbol table (loaded over the network) uses about 1Mb of memory. I constructed a static symbol table containing all the show routines and a few other commands (If I'm missing a command I add it for next build). I thus have a symbol table for the shell - and nothing in it is inappropriate to call from the shell. For some debugging I call netLoadSymTbl (after sysFlags |= 2) to get ALL the symbols loaded. None of the standard vxWorks images seemed right. I link a normal vxWorks, compress the contents, and add it to the boot rom uncompresser (linked to high memory). In my case all the object files are linked into vxWorks (4.5Mb uncompressed, 1.8Mb compressed) - although the actual application is in java and loaded from filestore. David > > Thanks, > Kevin > ---------------------------------------------------------------- David Laight email: dsl@tadpole.co.uk Tadpole Technology plc phone: +44 1223 278 256 Cambridge, UK fax: +44 1223 278 201 From owner-vxwexplo-process Thu Apr 29 04:00:16 1999 From: daemon@csg.lbl.gov Date: Thu Apr 29 04:00:19 PDT 1999 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Thu Apr 29 04:00:14 PDT 1999 Subject: Re: VMetro DPIO (PCI) & VxWorks Subject: dec21x40End and promiscuous ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: VMetro DPIO (PCI) & VxWorks Date: Wed, 28 Apr 1999 13:02:41 -0400 From: Frederick Byrne Organization: Lockheed Martin - GES Message-ID: <37273F31.F1D08930@lmco.com> References: Kai Virkki wrote: > > Hello! > > Has someone any experience on using VMetro DPIO-modules (PCI-mezzanine) > with Unix-version of Tornado, and possibly a Motorola CPU-board? I am > having some problems with using the driver of the module and would be > happy to discuss with any expert! > > I got a DPIO-driver to be used with VMetro Midas-board, which has > i900-processor but I have a Motorola-board. The first problems were > Endian-issues... I have been porting that driver now for some time and it > seems to install ok, but some functions still give strange output... I am > trying to do some DMA-transfers between Input- & Output-modules but I'm > not quite sure how to work this out. > > If you think you could give me some hints, please drop me some mail! > > Regards, > > Kai Virkki > 4th year physics student > (trying to learn VxWorks & PCI) We've used a VMETRO vxWorks driver for the Motorola MVME2600 series. It does (or did) exist. - -- ============================================================== The opinions and ideas expressed here are my own (unless they make money and then they belong to my employer :-) Frederick "Rick" Byrne frederick.j.byrne@lmco.com Lockheed Martin Government Electronic Systems --------------------------- Newsgroups: comp.os.vxworks Subject: dec21x40End and promiscuous Date: Wed, 28 Apr 1999 15:09:29 -0500 From: Claude Jibouleau Organization: Rockwell Collins Message-ID: <37276AF9.507CE90C@collins.rockwell.com> - --------------7036489F543B4B953209F513 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hi all ! I am using a PPC 604 VME board with a ZNYX346 NIC card. I need to set the card in promiscuous mode with the SENS dec21x40End driver. No success... First I tried muxIoctl(pEnd, EIOCSFLAGS, (caddr_t) (IFF_PROMISC | IFF_ALLMULTI)) no error but it doesnt work. so then I tried ifFlagChange( name, IFF_PROMISC, TRUE ) ifFlagChange( name, IFF_ALLMULTI, TRUE ) still no error and still doesnt work. I only get packets if the destination MAC address correspond to the NIC card's MAC address. I'm using a UDP socket to read the packets. Any ideas of what I am doing wrong? - --------------7036489F543B4B953209F513 Content-Type: text/html; charset=us-ascii Content-Transfer-Encoding: 7bit  
Hi all !

I am using a PPC 604 VME board with a ZNYX346 NIC card.  I need to set the card in promiscuous mode with the SENS dec21x40End driver.  No success...

First I tried

muxIoctl(pEnd, EIOCSFLAGS, (caddr_t) (IFF_PROMISC | IFF_ALLMULTI))

no error but it doesnt work. so then I tried

ifFlagChange( name, IFF_PROMISC, TRUE )
ifFlagChange( name, IFF_ALLMULTI, TRUE )

still no error and still doesnt work.  I only get packets if the destination MAC address correspond to the NIC card's MAC address.  I'm using a UDP socket to read the packets.  Any ideas of what I am doing wrong?
  - --------------7036489F543B4B953209F513-- --------------------------- End of New-News digest ********************** From owner-vxwexplo-process Thu Apr 29 07:04:27 1999 From: "Tony Mayes" Date: Thu Apr 29 07:04:30 PDT 1999 Subject: VxWorks & COM Hi all, Does anyone know if there is a COM interface for Vxworks? Is anyone out there working on this? Tony Mayes (The opinions expressed within are my own) Email address tony.mayes@epid.eurotherm.co.uk Phone (0044)(0)1903 205277 Fax (0044)(0)1903 524016 Eurotherm Process Instrumentation Div, Southdownview Way, Worthing, Sussex, UK, BN14 8NN From owner-vxwexplo-process Thu Apr 29 14:02:38 1999 From: Kevin Sapp Date: Thu Apr 29 14:02:41 PDT 1999 Subject: Re: proxy arp Hello, I am running VxWorks on a Pentium/cPCI bus system with hot swapable modules. There is a defined protocol for hot swapping the boards, but there are not physical locks that prevent a person from just ripping the board out of the system. I wanted to trap the PageFault/Seg Violation and, see if we were in the Shared Memory driver and report an error without suspending the task, and then continue on my way. Something like... catchPageFault(); ... do the shared memory stuff ... uncatchPageFault(); Signals do not work for this, simply returning from the interrupt does not work (page fault will try to reexecute the last instruction) This would be like a catch/throw or longjmp, but these cannot handle the interrupt context. It is not possible to vxMemProbe each access because of size of the data and vxMemProbe access time is too slow. Any ideas... Thanks Kevin From owner-vxwexplo-process Thu Apr 29 23:02:12 1999 From: "percheron cecile" Date: Thu Apr 29 23:02:15 PDT 1999 Subject: linking objects under vxWorks I try to link files **.o I compiled with a special GNU outside Tornado with other **.o compiled under Tornado. Under Tornado we get specific *.o who can't be linked. Does any body have seen this problem? /************************************** Cecile PERCHERON percheron@enertec.avicore.com tel: 01-41-28-88-35 ENERTEC. 92 Clamart -FRANCE /************************************** From owner-vxwexplo-process Fri Apr 30 07:47:26 1999 From: "F. Scheurer" Date: Fri Apr 30 07:47:30 PDT 1999 Subject: VxWorks on MV2700 using IPPC2636 Hello VxWorkers, i'm trying to connect an IPPC 2636 carrier board to a MVME 2700. I'm using the BSP from May 1998. I got a description for the procedure using the MVME 2604. There are multiple differences between my actual BSP and the MVME2604 (for example pciIntXXX() have been moved to pciIntLib.c (comment in pciConfigLib.c)). Is there someone out there having experience using this configuration? Do i really have to do all these changes to mv2600.h, pciXXXLib.c, sysLib.c and so on? Or is it now just a part of the BSP? But i even didn't find pciIntLib.X. TIA -- +----------------------------------------------------------------+ | Friedrich Scheurer Zentrum Fertigungstechnik Stuttgart | | Dipl.-Ing. Arbeitsbereich Steuerungstechnik | | Tel +49 711 13162-41(-0) Nobelstrasse 15 | | Fax +49 711 13162-11 D-70569 Stuttgart | | E-Mail scheurer@zfs.rus.uni-stuttgart.de | +----------------------------------------------------------------+ | http://zfs.rus.uni-stuttgart.de | +----------------------------------------------------------------+ | privat: | | Telefon +49 7161 686362 | | Fax +49 7161 686301 | | E-Mail Friedrich.Scheurer@t-online.de | +----------------------------------------------------------------+ From owner-vxwexplo-process Fri Apr 30 08:20:10 1999 From: Jean-Marie Campiglia Date: Fri Apr 30 08:20:13 PDT 1999 Subject: Tornado 1.0.1 and STL I am using Tornado 1.0.1 on Motorolla mtx60x boards. I need to use the from STL. I have tried the STLport based on the SGI STL but I can't compile it with GNU GCC 2.8.1 for Power PC because some header files related to threads are missing. Has anybody already used a STL (and map) with vxWorks? Any help and pointers (STL location) would be greatly appreciated. JM From owner-vxwexplo-process Fri Apr 30 13:50:46 1999 From: mmenge@dspt.com Date: Fri Apr 30 13:50:50 PDT 1999 Subject: T2 Mark Menge@DSPT 04/30/99 04:50 PM When is tornado version 2 coming out? tornado vxWorks From owner-vxwexplo-process Fri Apr 30 15:16:24 1999 From: David Fraser Date: Fri Apr 30 15:16:28 PDT 1999 Subject: Re: T2 >Mark Menge@DSPT >04/30/99 04:50 PM >When is tornado version 2 coming out? Tornado II is in the final stages of release validation and will be shipping very soon. Rest assured, we're all looking forward to it here :-) Dave Fraser VP Engineering, Wind River Systems From owner-vxwexplo-process Fri Apr 30 17:33:22 1999 From: "Patrick T. Pinkowski" Date: Fri Apr 30 17:33:26 PDT 1999 Subject: When is the new DosFs Coming Out When is the new VxWorks DosFs coming out? (FAT32, long file names, FTP GUI access, etc.) 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 owner-vxwexplo-process Fri Apr 30 19:37:11 1999 From: ktsubota@keck.hawaii.edu (Kevin Tsubota) Date: Fri Apr 30 19:37:15 PDT 1999 Subject: how do I see what ISR is attached to an interrupt vector Hi, I have a MVME230x PowerPC board running vxWorks 5.3.1. Can anyone tell me how to find out what ISR function is attached to a given interrupt vector? I'm experiencing a problem where my first interrupt vector is generated and processed but the next interrupt vector although generated and attached to a different ISR function seems to be calling the same ISR funtion as the first one. This same code works fine for a frc40 running vxWorks 5.3.1. Thank you. Kevin Tsubota California Assoc. for Research in Astronomy (808)885-7887 From owner-vxwexplo-process Fri Apr 30 21:02:20 1999 From: ktsubota@keck.hawaii.edu (Kevin Tsubota) Date: Fri Apr 30 21:02:24 PDT 1999 Subject: how do I see what ISR is attached to an interrupt vector Hi, I have a MVME230x PowerPC board running VxWorks 5.3.1. Can anyone tell me how to find out what ISR function is attached to a given interrupt vector? I'm experiencing a problem where my first interrupt vector is generated and processed but the next interrupt vector although generated and attached to a different ISR function seems to be calling the same ISR funtion as the first one. This same code works fine for a frc40 running VxWorks 5.3.1. Thank you. Kevin Tsubota California Assoc. for Research in Astronomy (808)885-7887