From vxwexplo-errs@csg.lbl.gov Fri Feb 1 04:03:22 2002 From: Vxworks Exploder Date: Fri Feb 1 04:03:25 PST 2002 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Fri Feb 1 04:03:17 PST 2002 Subject: Re: How do I access both serial ports? Subject: Re: How do I access both serial ports? Subject: Re: rapidcontrol for CLI Subject: Re: semTake and ISR Subject: Re: Connect more than one ISR to single IV?? Subject: Determining the target IP address Subject: Re: Connect more than one ISR to single IV?? Subject: Re: 16550 UART driver Subject: Re: Tornado2 and PVCS Subject: Re: tffsDevFormat fails half way thru.. Subject: Re: Determining the target IP address Subject: Re: How do I access both serial ports? Subject: Re: Does nucleus plus support MMU? Subject: Re: changing the target name Subject: Re: Tornado2 and PVCS Subject: Re: Tornado2 and PVCS Subject: tffsDevFormat() fails after progress-reporting 100% completed. Subject: Re: MPC8260 Ethernet Driver Subject: How to set the host as current working directory? Subject: Re: "__fixdfdi" reference Subject: DOS key like features in vxworks Subject: Re: How to set the host as current working directory? Subject: Registers list in Crosswind Subject: Re: Tornado2 and PVCS Subject: rcmd for VxWorks Subject: Re: Tornado2 and PVCS Subject: Re: DOS key like features in vxworks Subject: Re: Tornado2 and PVCS Subject: Re: Determining the target IP address Subject: open changes the priority of my task ! Subject: Re: open changes the priority of my task ! Subject: Re: Default scheduler in VxSim ? Subject: Re: Source for WRS build utilities? Subject: Vxworks PPP: Login problem Subject: vxWorks and vxWorks.st Subject: Re: 16550 UART driver Subject: Problem binding socket to IP address Subject: Re: rcmd for VxWorks Subject: Re: what are these Assembler messages... Subject: How to auto format code in Tornado Subject: Re: How to auto format code in Tornado Subject: Re: vxWorks and vxWorks.st Subject: Re: open changes the priority of my task ! Subject: Re: Configuration of a serial device to run full duplex Subject: Design of Flash file system ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: How do I access both serial ports? Date: Thu, 31 Jan 2002 13:43:35 +0100 From: "Werner Schiendl" Message-ID: <3c593bf7$1@brateggebdc5.br-automation.co.at> References: hi, check NUM_TTY in your project (hardware | peripherals | serial | SIO) if it's (N_UART_CHANNELS) like in my BSP, check this macro in your BSP's config.h hth Werner "Ole-Johan Ellingsen" wrote in message news:a3b4l2$8pm$1@tyfon.itea.ntnu.no... > Hi. > > I have a Pentium target with two serial ports > (http://www.eurotech.it/main/ProductNFE.asp?SubM=Catalog&Sub2M=PC/104&Sub3M= > Cpu_module&Sub4M=CPU-1231&Lang=E). I use one of the serial ports, but am > unable to enable the second one for use. When I type 'devs' in the console, > I only get one of them listed. The target has onboard VGA and Ethernet > adapter, so I don't use any serial line for communication with Tornado or > anything. I have tried to remove SLIP from the project, but it still doesn't > enable the last serial port. Does anyone have any idea what I can do? > > Regards, > Ole-Johan Ellingsen > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How do I access both serial ports? Date: Thu, 31 Jan 2002 14:59:18 +0100 From: "Ole-Johan Ellingsen" Organization: Norwegian university of science and technology Message-ID: References: <3c593bf7$1@brateggebdc5.br-automation.co.at> I did, and it is set to 2. The funny thing is that when I type 'devs', it lists 'tyCo/1', which means that 'tyCo/0' should be in there somewhere... :) I know that different settings can occupy a serial port (like SLIP, WDB and console), but I have excluded SLIP, use END for WDB and have included PC Console. What more can I do? :) Ole-Johan "Werner Schiendl" wrote in message news:3c593bf7$1@brateggebdc5.br-automation.co.at... > hi, > > check NUM_TTY in your project > (hardware | peripherals | serial | SIO) > > if it's (N_UART_CHANNELS) like in my BSP, check this macro in your BSP's > config.h > > hth > Werner > > "Ole-Johan Ellingsen" wrote in message > news:a3b4l2$8pm$1@tyfon.itea.ntnu.no... > > Hi. > > > > I have a Pentium target with two serial ports > > > (http://www.eurotech.it/main/ProductNFE.asp?SubM=Catalog&Sub2M=PC/104&Sub3M= > > Cpu_module&Sub4M=CPU-1231&Lang=E). I use one of the serial ports, but am > > unable to enable the second one for use. When I type 'devs' in the > console, > > I only get one of them listed. The target has onboard VGA and Ethernet > > adapter, so I don't use any serial line for communication with Tornado or > > anything. I have tried to remove SLIP from the project, but it still > doesn't > > enable the last serial port. Does anyone have any idea what I can do? > > > > Regards, > > Ole-Johan Ellingsen > > > > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: rapidcontrol for CLI Date: 31 Jan 2002 06:08:15 -0800 From: tut_pharoh@yahoo.com (tutankhamun) Organization: http://groups.google.com/ Message-ID: <88c8dc81.0201310608.a6134a9@posting.google.com> References: Assuming you're using RapidControl CLI v3.12 (or later) the following might help. A global no form can be achieve by #define __APPLY_NO_TO_ALL__ in Future Extension. This will apply no to all child subnodes. A node w/ just one handler is defaulted to take no or non-no command. You can call: Boolean RCC_DB_IsNoCommand(paramList *pParamList) within handler to determine whether or not its a no command. To enforce no handler must have 'no', #define __NO_MEANS_NO__ You can also play around with __EXEC_LEAF_NODE_ONLY__ in rcc_db.c tutankhamun ilyasg@yahoo.com (Ilyas G) wrote in message news:... > hi i am using rapid control for CLI > > We know that : > The handlers are executed one by one through the command tree > ie : For a command "create ip route", the handlers of the "create" > ,"ip" , "route" will be executed one by one. > > the problem is with the no form. > if i want to have "no" form for this command, i have to declare the > handler within the last node "route". and the command will be > # create ip no route > > wich is not good, i prefer to have > # no create ip route > > Of course i can test on the rest of the command in the first handler, > but if many nodes commands exists it will be as hard as developing a > new "no" branch. > i think. > > any hints ???? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: semTake and ISR Date: Thu, 31 Jan 2002 09:17:02 -0500 From: Doug Dotson Organization: Verio Message-ID: <3C5951DE.3CB80120@digidata.com> References: <9b6f8883.0201301813.e5ac79c@posting.google.com> semTake doesn't basically do an IntLock. Having code that seems the need to take a semaphore in an ISR probably needs to be rethought. If the code works as written, then the semTake never blocks which would seem to indicate that is is not needed. Suggest you try and understand what the original author was trying to accomplish. doug Mark Chang wrote: > I know that > The semTake( ) routine is not callable from interrupt service routines. > > So what happens when I have a semTake in an IRS? > Does it just return? > > Any damage? Since semTake basically does an IntLock. > > And to answer your question as to why I would > do this. I don't, but some code I was debugging > did and I'm wondering about the effects, if any. > > TIA > Mark --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Connect more than one ISR to single IV?? Date: 31 Jan 2002 06:19:21 -0800 From: mistry_amit@yahoo.com (Amit) Organization: http://groups.google.com/ Message-ID: <3ca555c5.0201310619.13d062aa@posting.google.com> References: <3ca555c5.0201302211.4f9744ef@posting.google.com> <3c591f4b$1@brateggebdc5.br-automation.co.at> Hi Werner, Thanks for your reply. Let me explain my problem in detail. Actually, I dont want to patch vxworks's system files. The reason behind not modifying original ISR is whenever I will give source code of module then I need to provide patch file of original ISR. And that is the limitation. My second ISR should be called whenever there is a keyboard's interrupt. My module's ISR will read keycode and take seperate action as per my application. Is it possible? Any guidence in this direction. One of my friend told me that: Using "signal", we can generate software interrupt for specific keys but not all keys. Thanks, Amit "Werner Schiendl" wrote in message news:<3c591f4b$1@brateggebdc5.br-automation.co.at>... > Hi, > > you can get the current interrupt service routine with intVecGet() > Depending on your target you *may* be able to call this vector in your own > handler and thus daisy chain the two. > > If you have the source for the first ISR (and the code that attaches it, in > particular) you can probably just use pciIntConnect instead of intConnect. > The former manages a list as it is used for shared interrupts on PCI bus. > > If you just want to do additional processing to the same interrupt (i. e. > from the same device), the best way would be modifying the original ISR. > Appart from getting the source code, this will cause the least headaches. > > hth > Werner > > "Amit" wrote in message > news:3ca555c5.0201302211.4f9744ef@posting.google.com... > > Hi friends, > > > > Do I connect more than one Interrupt Service routine to same Interrupt > > Vector means same Interrupt source?? > > > > I want to connect another ISR to Keyboard interrupt source. > > > > Is there any other way to achieve this?? > > > > Expecting replies from VxWorks experts!! > > > > Thanks, > > Amit --------------------------- Newsgroups: comp.os.vxworks Subject: Determining the target IP address Date: Thu, 31 Jan 2002 10:16:58 -0500 From: Jeremiah Frick Organization: Improv Systems, Inc. Message-ID: <3C595FEA.4AA9D8FA@improvsys.com> Hi all! I'm trying to programatically determine the IP address of my VxWorks Malta box. Unfortunately, I'm getting an error when I try to run this code on the box: { char addrArray[10]; ifAddrGet( "lnPci0", addrArray ); printf( "IP Address: %s", addrArray ); } The error is: (KLSI CartRootTaskd found. ): Can't attach unknown device lnpci (unit 0). I'm fairly certain lnPci is the correct device. Any thoughts? Jeremy --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Connect more than one ISR to single IV?? Date: Thu, 31 Jan 2002 17:04:29 +0100 From: "Werner Schiendl" Message-ID: <3c596b0e@brateggebdc5.br-automation.co.at> References: <3ca555c5.0201302211.4f9744ef@posting.google.com> <3c591f4b$1@brateggebdc5.br-automation.co.at> <3ca555c5.0201310619.13d062aa@posting.google.com> Hi, "Amit" wrote in message news:3ca555c5.0201310619.13d062aa@posting.google.com... > Hi Werner, > > Thanks for your reply. > Let me explain my problem in detail. What platform and what type of keyboard are you using? PS/2 keyboard on PC-like platform? > Actually, I dont want to patch vxworks's system files. > The reason behind not modifying original ISR is whenever I will give > source code of module then I need to provide patch file of original > ISR. And that is the limitation. > Well, although this is a good idea in general, it is sometimes simple not practicable in VxWorks. If you have the driver in source, the best would probably be to derive your own version of it. Include that instead of the original one, then you don't need to modify the original sources. However, you need to supply your modified driver as part of your stuff. > My second ISR should be called whenever there is a keyboard's > interrupt. My module's ISR will read keycode and take seperate action > as per my application. What is the reason, why you do not just _replace_ the whole driver?? What do you expect the existing driver to do?? And what purpose is your hacking?? Do you think, the original ISR can work, if you already read the keycode(s) from the hardware (depends on the type of keyboard). > > Is it possible? Any guidence in this direction. Did you try the intVecGet API? It delivers the pointer to the original interrupt service routine. Try invoking it from (the end) of your own ISR an see if it works. If it does not work that way, you can probably hack it in assembler - if you really like. > > One of my friend told me that: > Using "signal", we can generate software interrupt for specific keys > but not all keys. > Hm, never heard that this should work.... Where should the keys go, normal target shell? UGL? WindML? hth Werner --------------------------- Newsgroups: comp.os.vxworks Subject: Re: 16550 UART driver Date: 31 Jan 2002 11:23:24 -0500 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: References: <3C580FC8.8DAA7C55@sepro-robotique.com> <3C5901E5.B7FC9082@sepro-robotique.com> Sender: bpringlemeir@DeadDuck >>>>> "EH" == Emmanuel Herbreteau writes: EH> Hi, I'm looking for a driver for the NS16550 UART with FIFO EH> enabled (usefull for speed over 19200 bauds). Does anyone know EH> where to find such driver ?? >>> Bill Pringlemeir wrote: >> Yes, you can find one in your imagination! The source code is in >> $(WIND_BASE)/target/src/drv/sio/ns16550Sio.c. I am guessing that >> you didn't know this? [patches to enable FIFO snipped by Emmanuel] EH> Hi, I'm very surprised to find this driver in EH> /src/drv/serial and (also) in /src/drv/sio. And it's not EH> the same driver ! EH> But, unfortunately, despite the FIFO_ENABLE bit, i'm quite EH> sure it's still an old driver for 16450 (without FIFO) EH> because the interrupts routines looks like : [snip] EH> Does anyone know how to enable the FIFO mode on 16550 ? Well, first off, you have broken a cardinal rule of comp.os.vxworks. What is the version of vxWorks, Tornado, and the BSP that you are using. With Tornado 2.0, 2.0.1, and 2.1.1 for the ARM the $(WIND_BASE)/target/src/drv/sio/ns16550Sio.c code is sufficient to enable the FIFO with the patches I provided in my original reply. Apparently you like to ignore free advice. Perhaps you can give me the address of your company so that I can come and work for you for free? regards, Bill Pringlemeir. - -- I mean, furthermore, you fascist, Nicaraguan people are pathetic. I don't want to hear about your operation. vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado2 and PVCS Date: 31 Jan 2002 16:20:35 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: <3c587646$1@brateggebdc5.br-automation.co.at> <9ec53d11.0201310148.3c625ef5@posting.google.com> "Ofer Goren" wrote: > Hi. > Let me rephrase my question: > Fact 1: We use Tornado 2.0.2 > Fact 2: WE USE PVCS. > > I was wonder what is the best way to work with this version control > and Tornado: > like, do you use archives, or flat trees? > Do you put your project files under PVCS? > Do you put your makefiles, prjParams, prgConfig etc under PVCS? > If you use archives, do you put them also under PVCS? > what do you do with you .out files? your .o files? your .exe files? > you vxWorks files in case of bootable application? > Do you compile your archives with your own makefiles, or use the > Tornado GUI? > > All those question effect the way one work with PVCS, and the number > of files under the control of the source control utility. These are more general CM questions, and on some of them I can give my opinion. One thing is that you should only put original files under CM or VM. So files that can be regenerated should not be stored in CM/VM. Groeten, Johan - -- o o o o o o o . . . _____________________________ o _____ || Johan Borkhuis | .][__n_n_|DD[ ====_____ | borkhuis@agere.com | >(________|__|_[_________]_|__________________________| _/oo OOOOO oo` ooo ooo 'o!o!o o!o!o` === VxWorks FAQ: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html === --------------------------- Newsgroups: comp.os.vxworks Subject: Re: tffsDevFormat fails half way thru.. Date: 31 Jan 2002 08:34:31 -0800 From: roswel_ajf@hotmail.com (roswell) Organization: http://groups.google.com/ Message-ID: <6dcae13d.0201310834.61d42eb5@posting.google.com> References: <6dcae13d.0201301352.383bde3d@posting.google.com> well, the noOfMTDs is 1 also, and now when i run tffsDevFormat 0, 0 it picks up my write/erase routines (god knows why it didn't pick up before), anyway, but afer a long time (erase/write), tffsDevFormat() fails. Here is display of what i see in my first 2 erase units. - -> tffsDevFormat 0, 0 (takes the default FormatParams) ;;;;;;;lots of my debug prints;;;;;; value = -1 = 0xffffffff = udbinfo + 0xffed79af (this is the failing msg, no clue what it is) - -> statprint (for each routine, i increment a counter myStat.routinename++) cfiscsIdentify = 2 cfiscsInterleavedWordWrite = 2511 cfiscsInterleavedWordErase = 256 cfiscsWordIdentify = 2 flash2byteswrite = 17567 flMap = 4282 sysTffsInit = 1 rfaRegister = 1 rfaCardDetected = 2 rfaVccOn = 1 rfaVccOff = 1 rfaVppOn = 2 rfaVppOff = 3 rfaSocketInit = 1 rfaSetWindow = 3 rfaSetMappingContext = 2 rfaGetAndClearCardChangeIndicator = 4 rfaWriteProtected = 2767 flFitInSocketWindow = 0 flDelayLoop = 2 ..Errors.. cfiscsInterleavedWordWrite = 0 cfiscsInterleavedWordErase = 0 cfiscsWordIdentify = 0 flash2byteswrite = 0 According to this, i never had an error happenning while write/erase. But why my tffsDevFormat faile. .....FLASH DISPLAY..... d 0x78000000 (EU 0) 78000000: 1303 4349 5346 ff00 4654 4c31 3030 0001 *..CISF..FTL100..* 78000010: 0100 0000 0000 0913 0000 0001 0032 c207 *.............2..* 78000020: 0000 0100 d207 00ff 0000 0000 0000 0000 *................* 78000030: 4400 0000 ffff ffff ffff ffff ffff ffff *D...............* 78000040: ffff ffff 3000 0000 3000 0000 3000 0000 *....0...0...0...* 78000050: 3000 0000 3000 0000 3000 0000 3000 0000 *0...0...0...0...* 78000060: 3000 0000 3000 0000 405c f0ff 405e f0ff *0...0...@\..@^..* 78000070: 4060 f0ff 4062 f0ff 4064 f0ff 4066 f0ff *@`..@b..@d..@f..* 78000080: 4068 f0ff 406a f0ff 406c f0ff 406e f0ff *@h..@j..@l..@n..* 78000090: 4070 f0ff 4072 f0ff 4074 f0ff 4076 f0ff *@p..@r..@t..@v..* 780000a0: 4078 f0ff 407a f0ff 407c f0ff 407e f0ff *@x..@z..@|..@~..* 780000b0: 4080 f0ff 4082 f0ff 4084 f0ff 4086 f0ff *@...@...@...@...* 780000c0: 4088 f0ff 408a f0ff 408c f0ff 408e f0ff *@...@...@...@...* 780000d0: 4090 f0ff 4092 f0ff 4094 f0ff 4096 f0ff *@...@...@...@...* 780000e0: 4098 f0ff 409a f0ff 409c f0ff 409e f0ff *@...@...@...@...* 780000f0: 40a0 f0ff 40a2 f0ff 40a4 f0ff 40a6 f0ff *@...@...@...@...* d 0x78080000 (EU 0) 78080000: 1303 4349 5346 ff00 4654 4c31 3030 0001 *..CISF..FTL100..* 78080010: 0200 0000 0100 0913 0000 0001 0032 c207 *.............2..* 78080020: 0000 0100 d207 00ff 0000 0000 0000 0000 *................* 78080030: 4400 0000 ffff ffff ffff ffff ffff ffff *D...............* 78080040: ffff ffff 3000 0000 3000 0000 3000 0000 *....0...0...0...* 78080050: 3000 0000 3000 0000 3000 0000 3000 0000 *0...0...0...0...* 78080060: 3000 0000 3000 0000 ffff ffff ffff ffff *0...0...........* 78080070: ffff ffff ffff ffff ffff ffff ffff ffff *................* 78080080: ffff ffff ffff ffff ffff ffff ffff ffff *................* 78080090: ffff ffff ffff ffff ffff ffff ffff ffff *................* 780800a0: ffff ffff ffff ffff ffff ffff ffff ffff *................* 780800b0: ffff ffff ffff ffff ffff ffff ffff ffff *................* 780800c0: ffff ffff ffff ffff ffff ffff ffff ffff *................* 780800d0: ffff ffff ffff ffff ffff ffff ffff ffff *................* 780800e0: ffff ffff ffff ffff ffff ffff ffff ffff *................* 780800f0: ffff ffff ffff ffff ffff ffff ffff ffff *................* d 0x78100000 (EU 1) 78100000: 1303 4349 5346 ff00 4654 4c31 3030 0001 *..CISF..FTL100..* 78100010: 0300 0000 0200 0913 0000 0001 0032 c207 *.............2..* 78100020: 0000 0100 d207 00ff 0000 0000 0000 0000 *................* 78100030: 4400 0000 ffff ffff ffff ffff ffff ffff *D...............* 78100040: ffff ffff 3000 0000 3000 0000 3000 0000 *....0...0...0...* 78100050: 3000 0000 3000 0000 3000 0000 3000 0000 *0...0...0...0...* 78100060: 3000 0000 3000 0000 ffff ffff ffff ffff *0...0...........* 78100070: ffff ffff ffff ffff ffff ffff ffff ffff *................* 78100080: ffff ffff ffff ffff ffff ffff ffff ffff *................* 78100090: ffff ffff ffff ffff ffff ffff ffff ffff *................* 781000a0: ffff ffff ffff ffff ffff ffff ffff ffff *................* 781000b0: ffff ffff ffff ffff ffff ffff ffff ffff *................* 781000c0: ffff ffff ffff ffff ffff ffff ffff ffff *................* 781000d0: ffff ffff ffff ffff ffff ffff ffff ffff *................* 781000e0: ffff ffff ffff ffff ffff ffff ffff ffff *................* 781000f0: ffff ffff ffff ffff ffff ffff ffff ffff *................* roswel_ajf@hotmail.com (roswell) wrote in message news:<6dcae13d.0201301352.383bde3d@posting.google.com>... > Hi vxfolks.. > > well, tffsDevFormat 0, 0 fails. (value = -1 = 0xffffffff = udbinfo + > 0xffed84f3). I guess, it cannot find my write/erase routines. yes, i > never got the printout in my cfiscsIdentify() routine. But, i saw my > xxxRegiset() is getting called. i can see noOfDrives is 1. BTW, i > erased all erseunit before doing this tffsDevformat thing. now, my > entire flash filled with 0xfffff...till theend. > > i think, the connection between FlFlash and FlSocket is not > established, may. I thought the cfiscsIdentify() makes the connection. > then, tffsDevFormat needs to write/erase etc.. > > please please help. > > thanks in advance > > (THe most funny thing is one time tffsDevFormat was actually invoking > my MTD's write routine, but half was failed with the same error as > above. then, via other project i got eraseunit() to erase the entire > flash, then trying tffsDevFormat() fails. Does the tffsDevFormat() > expect any manufacture data to be foudn in flash? ). my core is > ppc750/vx5.2 --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Determining the target IP address Date: Thu, 31 Jan 2002 17:29:02 +0100 From: "Werner Schiendl" Message-ID: <3c5970ce@brateggebdc5.br-automation.co.at> References: <3C595FEA.4AA9D8FA@improvsys.com> Hi, "Jeremiah Frick" wrote in message news:3C595FEA.4AA9D8FA@improvsys.com... > Hi all! > > I'm trying to programatically determine the IP address of my VxWorks > Malta box. Unfortunately, I'm getting an error when I try to run this > code on the box: > > { > char addrArray[10]; > ifAddrGet( "lnPci0", addrArray ); > > printf( "IP Address: %s", addrArray ); > } > > The error is: > > (KLSI CartRootTaskd found. > ): Can't attach unknown device lnpci (unit 0). > Is this a typo, or does the error message _indead_ display the device name in all lowercase? A compare of "lnPci" with "lnpci" will fail, if str(n)cmp() function is used! > I'm fairly certain lnPci is the correct device. Any thoughts? > Try a different name for the network device, or a network board with a name in all-lowercase letters. hth Werner --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How do I access both serial ports? Date: Thu, 31 Jan 2002 17:26:07 +0100 From: "Werner Schiendl" Message-ID: <3c59701e$1@brateggebdc5.br-automation.co.at> References: <3c593bf7$1@brateggebdc5.br-automation.co.at> Hi, Indees, '/tyCo/0' should be there, if '/tyCo/1' is :-) See the file $(WIND_BASE)/target/config/comps/src/usrSerial.c This is where the serial ports are initialized, at least if you are building from Tornado II project facility. Check, that all prerequesites are met, e. g. that INCLUDE_WDB_COMM_SERIAL is not set by accident. If that does not help, try to initialize the interface manually to see if and which error occurs. hth Werner "Ole-Johan Ellingsen" wrote in message news:a3bigv$gtq$1@tyfon.itea.ntnu.no... > I did, and it is set to 2. > > The funny thing is that when I type 'devs', it lists 'tyCo/1', which means > that 'tyCo/0' should be in there somewhere... :) I know that different > settings can occupy a serial port (like SLIP, WDB and console), but I have > excluded SLIP, use END for WDB and have included PC Console. What more can I > do? :) > > Ole-Johan > > > > > "Werner Schiendl" wrote in message > news:3c593bf7$1@brateggebdc5.br-automation.co.at... > > hi, > > > > check NUM_TTY in your project > > (hardware | peripherals | serial | SIO) > > > > if it's (N_UART_CHANNELS) like in my BSP, check this macro in your BSP's > > config.h > > > > hth > > Werner > > > > "Ole-Johan Ellingsen" wrote in message > > news:a3b4l2$8pm$1@tyfon.itea.ntnu.no... > > > Hi. > > > > > > I have a Pentium target with two serial ports > > > > > > (http://www.eurotech.it/main/ProductNFE.asp?SubM=Catalog&Sub2M=PC/104&Sub3M= > > > Cpu_module&Sub4M=CPU-1231&Lang=E). I use one of the serial ports, but am > > > unable to enable the second one for use. When I type 'devs' in the > > console, > > > I only get one of them listed. The target has onboard VGA and Ethernet > > > adapter, so I don't use any serial line for communication with Tornado > or > > > anything. I have tried to remove SLIP from the project, but it still > > doesn't > > > enable the last serial port. Does anyone have any idea what I can do? > > > > > > Regards, > > > Ole-Johan Ellingsen > > > > > > > > > > > > > > > --------------------------- Newsgroups: comp.arch.embedded,comp.os.vxworks Subject: Re: Does nucleus plus support MMU? Date: 31 Jan 2002 14:59:54 GMT From: "Mark A. Odell" Organization: Embedded Firmware Design, Inc. Message-ID: References: "WJ" wrote in news:u5h5mltim0dee2@corp.supernews.com: > Hi, > > I'd like to know if Nucleus Plus supports MMU for MIPS4K? It does, at least in a flat one-to-one mapping (phys->log) on the PowerPC. I don't see why they wouldn't answer this question for you if you asked them. If you mean use the MMU for virtual address space protection for tasks, then the answer is no. - -- - - Mark A. Odell - - Embedded Firmware Design, Inc. - - http://www.embeddedfw.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: changing the target name Date: Thu, 31 Jan 2002 09:51:45 +0100 From: "Michael Lawnick" Organization: Buergernetz Dillingen Message-ID: References: <3C5726DA.A30B2B06@dsto.defence.gov.au> Reply-To: "Michael Lawnick" Sender: mlawnick@pd9e6515a.dip.t-dialin.net Hi Trish, in addition to Tim: think twice, whether you get what you want: there is a difference between standard .out and .o ! .out is munched, i.e. added with c++ creators and destructors (which might be empty). You can't link two .out files because of doubled symbols therefore. - -- Mit freundlichen Grüßen, Michael Lawnick ============================================== SOFTEC GmbH Tel +49-731-96600-0 Promenade 17 Fax +49-731-96600-23 D-89073 Ulm Michael Lawnick Germany lawnick@softec.de ============================================== "Tim Shaw" schrieb im Newsbeitrag news:3C5726DA.A30B2B06@dsto.defence.gov.au... > You can either create a new rule in the project properties menu, or add mv > xx.out xx.o to the post build rule in the properties/macros tab. > > Trish Bennett wrote: > > > We simply (?) want to change the name of the target from xx.out to > > xx.o. How can this be done? I tried replacing all occurences of > > xx.out to xx.o in the .wpj file (outside of the Tornado IDE), but that > > doesn't work. > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado2 and PVCS Date: Thu, 31 Jan 2002 17:48:22 +0200 From: "Ofer Goren" Organization: Verio Message-ID: References: <3c587646$1@brateggebdc5.br-automation.co.at> <9ec53d11.0201310148.3c625ef5@posting.google.com> Hi. Let me rephrase my question: Fact 1: We use Tornado 2.0.2 Fact 2: WE USE PVCS. I was wonder what is the best way to work with this version control and Tornado: like, do you use archives, or flat trees? Do you put your project files under PVCS? Do you put your makefiles, prjParams, prgConfig etc under PVCS? If you use archives, do you put them also under PVCS? what do you do with you .out files? your .o files? your .exe files? you vxWorks files in case of bootable application? Do you compile your archives with your own makefiles, or use the Tornado GUI? All those question effect the way one work with PVCS, and the number of files under the control of the source control utility. Thanks, - -- Ofer Goren Iamba Technologies "Ignacio González" wrote in message news:9ec53d11.0201310148.3c625ef5@posting.google.com... > We use RCS (free) for DOS with our Tornado2 projects, plus a > MS-Windows front-end developed by myself, and a set of equivalent > batch procedures for the anti-windows people. > > In fact, we have been using PVCS (for DOS) for many years, and we > changed to RCS when we were forced to use long names, like some of the > sources distributed by Tornado (not supported by our robust, > efficient, 2ky-compliant but very old PVCS 1990 version.) > > RCS is almost identical to bare PVCS, although has a little less > options (which I had to supplement with some awk scripts called from > the batch files and MS-Windows front-end.) > > As pointed by Werner, we also had to treat the project and workspace > files as binary files (-b option in RCS) and the rest as text files > (-kv option.) We learned it the hard way. > > We don't integrate RCS as a tool in Tornado. In fact, we use MultiEdit > as the main IDE. From here we launch Tornado and RCS-Windows as > separte applications. We also configure Tornado to use MultiEdit as > the external editor. MultiEdit has the nice feature of updating > automatically the contents and attributes of the files opened in its > IDE and modified by other applications (like RCS or Tornado.) > > "Werner Schiendl" wrote in message news:<3c587646$1@brateggebdc5.br-automation.co.at>... > > Hi, > > > > I don't know PVCS, we use MS SourceSafe for some VxWorks projects. > > > > As far as I was told, PVCS and SourceSafe are pretty much the same stuff (e. > > g. exclusive check-out paradigm). > > > > There is no need for any third party tools (especially if they are not for > > free - or you own them anyway) > > Just use the Tool that comes with your sourcecode management system. > > (for SourceSafe, its named "Visual SourceSafe Explorer" - and allows > > browsing the repository, checking-in and out, getting the latest > > version....) > > > > + you can configure Tornado to allow check-in and check-out directly using > > command-line (we do not use this) > > > > One thing to be careful are the project and workspace file. > > They contain Unix line endings by default, and if they are changed to DOS > > Line-Endings, Tornado will cease to work. > > In SourceSafe, we checked them in as binary files - this solves the problem > > for us. > > > > Alternatively, you can patch the TCL files of your Tornado install. > > Search the group, there was a description on how to do it some time ago > > (maybe its in the FAQ too). > > > > > > hth > > Werner > > > > > > "Martin Raabe" wrote in message > > news:a394lp$dbu$03$1@news.t-online.com... > > > Hello Ofer, > > > I recommend to use SNiFF+ PRO as a WindPowerTool on top of Tornado. > > > For more derails look at: > > > www.windriver.com for a free eval download. > > > > > > Ciao > > > > > > Martin Raabe > > > > > > "Ofer Goren" schrieb im Newsbeitrag > > > news:newscache$ku9rqg$lck$1@lnews.actcom.co.il... > > > > Hi. > > > > Does anyone work with this combination, and if so, any recommendations > > about > > > > the best way to work? > > > > > > > > Tornado2.02 for win2kSp2 > > > > PVCS for windows, version 6.7.11 > > > > > > > > Best reagrds, > > > > > > > > -- > > > > Ofer Goren > > > > Iamba Technologies > > > > > > > > > > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado2 and PVCS Date: Thu, 31 Jan 2002 08:11:01 -0800 From: "Michael R. Kesti" Organization: MK Associates Message-ID: <3C596C95.6B3F4ECF@gv.net> References: <3c587646$1@brateggebdc5.br-automation.co.at> <9ec53d11.0201310148.3c625ef5@posting.google.com> Reply-To: mkesti@gv.net Ofer Goren wrote: >Hi. >Let me rephrase my question: >Fact 1: We use Tornado 2.0.2 >Fact 2: WE USE PVCS. My experience is with ClearCase, but I think the concepts are sufficiently similar. >I was wonder what is the best way to work with this version control and >Tornado: >like, do you use archives, or flat trees? Flat trees. >Do you put your project files under PVCS? YES! The project files are a very important part of the source! Sometimes, though, it is a hassle to have them under control. For example, if a project generates several builds and you wish to compile one of the builds not currently selected by the project file, you will have to check the project file out to make it writable even though you don't intend to make any real changes to it. >Do you put your makefiles, prjParams, prgConfig etc under PVCS? No. These are derived files and the project files have the information needed to generated them. >If you use archives, do you put them also under PVCS? No comment. >what do you do with you .out files? your .o files? your .exe files? you >vxWorks files in case of bootable application? Again, these are derived files and need not be controlled. >Do you compile your archives with your own makefiles, or use the Tornado >GUI? I use the GUI. - -- ======================================================================== Michael Kesti | "And like, one and one don't make | two, one and one make one." mkesti@gv.net | - The Who, Bargain --------------------------- Newsgroups: comp.os.vxworks Subject: tffsDevFormat() fails after progress-reporting 100% completed. Date: 31 Jan 2002 09:39:58 -0800 From: roswel_ajf@hotmail.com (roswell) Organization: http://groups.google.com/ Message-ID: <6dcae13d.0201310939.5cc537fe@posting.google.com> hi, I added a FormatProgress() callback routine to std_format_parameters (see systffsformat). So, formatting is reported incrementally upto 100%, which is good. then, tffsDevFormat() start writing 4 bytes at a time continously. eventually, it dies. some printouts shown below. And my write routine doesn't show any error conditions. thanks..any tips are welcome. mtd: writing 4 bytes at addr 0x78000ff8 mtd: writing 4 bytes at addr 0x78000ff8 mtd: writing 4 bytes at addr 0x78000ffc mtd: writing 4 bytes at addr 0x78000ffc mtd: writing 4 bytes at addr 0x78001000 mtd: writing 4 bytes at addr 0x78001000 mtd: writing 4 bytes at addr 0x78001004 mtd: writing 4 bytes at addr 0x78001004 value = -1 = 0xffffffff = udbinfo + 0xffed7787 my flash base is 0x78000000. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: MPC8260 Ethernet Driver Date: 31 Jan 2002 09:43:11 -0800 From: ryeung@earthlink.net (Raymond Yeung) Organization: http://groups.google.com/ Message-ID: References: I've been using WRS's MPC8260 Ethernet driver with zero copy, and don't have any problems. You may want to take a closer look at whether the driver flushes/invalidate caches. Presumably in the copy-tx mode, you don't need these operation. But you would need them in zero-copy mode. IP and above probably use cacheable memory. "Grant" wrote in message news:... > Hi > > I am having a problem with the ethernet driver WRS provides for MPC8260. > > The default ethernet driver coming with Vxworks 5.4 and MPC 8260 BSP only > supports copy on transmission. Although inside motFccEnd.c there is a > function called motFccPktTransmit() which is supposed to transmit packets > with zero copy, the reality is if I use this API instead of > motFccPktCopyTransmit(), the vxworks cannot boot up. Actually it hangs when > SYS_FCC_ENET_ENABLE is invoked. I am wondering if anybody tried the > zero-copy transmission or this is a hardware problem that MPC8260 only > supports copy and transmit? > > Thanks a lot for any comments --------------------------- Newsgroups: comp.os.vxworks Subject: How to set the host as current working directory? Date: 31 Jan 2002 17:32:24 GMT From: Charly Message-ID: <2002131-183224-364691@foorum.com> References: <2002130-143411-880483@foorum.com> <3c581520$1@brateggebdc5.br-automation.co.at> <2002130-191651-67831@foorum.com> <3c587481$1@brateggebdc5.br-automation.co.at> But How to set the host as current working directory? - -- Use our news server 'news.foorum.com' from anywhere. More details at: http://nnrpinfo.go.foorum.com/ --------------------------- Newsgroups: comp.os.vxworks Subject: Re: "__fixdfdi" reference Date: 31 Jan 2002 11:11:47 -0800 From: Jeffrey.Steinsnyder@L-3Com.com (Jeff Steinsnyder) Organization: http://groups.google.com/ Message-ID: References: sj@shell1.aracnet.com (Scott Johnson) wrote in message news:... > In article , > Jeff Steinsnyder wrote: > >We are using VxWorks v5.4. When the 64-bit floating point number > >converted to 64-bit integer the compiler generates a reference to > >"__fixdfdi" which can not be resolved at link time. What do I need to > >do to resolve this reference? > > > I'm asuming you are using gcc. If you are using Diab or another compiler; > I do not know the answer. > > Try adding -lgcc to your final link (either of your kernel/app, or of your > loaded module.) You might also have to kill the -nostdlib option that is > getting passed to the compiler/linker. > > The necessary function is in libgcc.a (the compiler's support library) > which Wind River likes to pretend doesn't exist. Look in > $WIND_BASE/host/hosttype/lib/gcc-lib/powerpc-wrs-vxworks/cygnus-whatever/libgcc.a > > > e_s Thank You very much for your suggestions. --------------------------- Newsgroups: comp.os.vxworks Subject: DOS key like features in vxworks Date: 31 Jan 2002 11:17:07 -0800 From: mark_ck_chang@yahoo.com (Mark Chang) Organization: http://groups.google.com/ Message-ID: <9b6f8883.0201311117.2e1b5206@posting.google.com> Does anyone know of code or have code that will do DOS Key like features? You know, up arrow show history, line editing, insert and replace. I think it will have to replace tylib.c - TIA Mark --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to set the host as current working directory? Date: Thu, 31 Jan 2002 20:29:38 +0100 From: "Werner Schiendl" Message-ID: <3c599b20$1@brateggebdc5.br-automation.co.at> References: <2002130-143411-880483@foorum.com> <3c581520$1@brateggebdc5.br-automation.co.at> <2002130-191651-67831@foorum.com> <3c587481$1@brateggebdc5.br-automation.co.at> <2002131-183224-364691@foorum.com> ioDefPathSet "Charly" wrote in message news:2002131-183224-364691@foorum.com... > > But > How to set the host as current working directory? > > -- > Use our news server 'news.foorum.com' from anywhere. > More details at: http://nnrpinfo.go.foorum.com/ --------------------------- Newsgroups: comp.os.vxworks Subject: Registers list in Crosswind Date: Thu, 31 Jan 2002 15:36:00 -0500 From: "Jim S" Organization: Lucent Technologies, Columbus, Ohio Message-ID: Anyone know how to edit the register list in the "Registers" window of the Crosswind debugger? Is it possible to add any/all processor registers? Platform is PPC405L, BSP we got seems to be lacking. Only registers shown are the GP's 0-31 and about half dozen others (normal?). I'd like to be able to see others...started to sift through some tcl files, but thought someone may have a quicker answer. Any help is appreciated. Thanks, Jim S. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado2 and PVCS Date: Thu, 31 Jan 2002 16:07:48 -0500 From: Doug Dotson Organization: Verio Message-ID: <3C59B223.C23D941@digidata.com> References: <3c587646$1@brateggebdc5.br-automation.co.at> <9ec53d11.0201310148.3c625ef5@posting.google.com> Isn't that just like a newsgroup. We use exactly what you do, altough we don't try to get fancy. For example, we have never made a serious attempt to allow Tornado automatically check files in and out. Each developer has his own local workspace directory and gets all files off of the baseline to build the project. Anyone wanting to make a change checks out the necessary files, makes the change, tests, and then checks the files back in. Comments inserted below. doug Ofer Goren wrote: > Hi. > Let me rephrase my question: > Fact 1: We use Tornado 2.0.2 > Fact 2: WE USE PVCS. > > I was wonder what is the best way to work with this version control and > Tornado: > like, do you use archives, or flat trees? Not sure what you mean here. > > Do you put your project files under PVCS? No, just the source code. > > Do you put your makefiles, prjParams, prgConfig etc under PVCS? No. Everybody has their own on their local machine. > > If you use archives, do you put them also under PVCS? By archives, do you mean libraries like fred.a? If so, no we have not created any application level archives. > > what do you do with you .out files? your .o files? your .exe files? you > vxWorks files in case of bootable application? They all end up in the diectory created by Tornado when you create a project. Nothing in those directories is used with PVCS. > > Do you compile your archives with your own makefiles, or use the Tornado > GUI? We use the Tornado GUI exclusively. At least so far. But we don't build any archives. > > > All those question effect the way one work with PVCS, and the number of > files under the control of the source control utility. > > Thanks, > > -- > Ofer Goren > Iamba Technologies > "Ignacio González" wrote in message > news:9ec53d11.0201310148.3c625ef5@posting.google.com... > > We use RCS (free) for DOS with our Tornado2 projects, plus a > > MS-Windows front-end developed by myself, and a set of equivalent > > batch procedures for the anti-windows people. > > > > In fact, we have been using PVCS (for DOS) for many years, and we > > changed to RCS when we were forced to use long names, like some of the > > sources distributed by Tornado (not supported by our robust, > > efficient, 2ky-compliant but very old PVCS 1990 version.) > > > > RCS is almost identical to bare PVCS, although has a little less > > options (which I had to supplement with some awk scripts called from > > the batch files and MS-Windows front-end.) > > > > As pointed by Werner, we also had to treat the project and workspace > > files as binary files (-b option in RCS) and the rest as text files > > (-kv option.) We learned it the hard way. > > > > We don't integrate RCS as a tool in Tornado. In fact, we use MultiEdit > > as the main IDE. From here we launch Tornado and RCS-Windows as > > separte applications. We also configure Tornado to use MultiEdit as > > the external editor. MultiEdit has the nice feature of updating > > automatically the contents and attributes of the files opened in its > > IDE and modified by other applications (like RCS or Tornado.) > > > > "Werner Schiendl" wrote in message > news:<3c587646$1@brateggebdc5.br-automation.co.at>... > > > Hi, > > > > > > I don't know PVCS, we use MS SourceSafe for some VxWorks projects. > > > > > > As far as I was told, PVCS and SourceSafe are pretty much the same stuff > (e. > > > g. exclusive check-out paradigm). > > > > > > There is no need for any third party tools (especially if they are not > for > > > free - or you own them anyway) > > > Just use the Tool that comes with your sourcecode management system. > > > (for SourceSafe, its named "Visual SourceSafe Explorer" - and allows > > > browsing the repository, checking-in and out, getting the latest > > > version....) > > > > > > + you can configure Tornado to allow check-in and check-out directly > using > > > command-line (we do not use this) > > > > > > One thing to be careful are the project and workspace file. > > > They contain Unix line endings by default, and if they are changed to > DOS > > > Line-Endings, Tornado will cease to work. > > > In SourceSafe, we checked them in as binary files - this solves the > problem > > > for us. > > > > > > Alternatively, you can patch the TCL files of your Tornado install. > > > Search the group, there was a description on how to do it some time ago > > > (maybe its in the FAQ too). > > > > > > > > > hth > > > Werner > > > > > > > > > "Martin Raabe" wrote in message > > > news:a394lp$dbu$03$1@news.t-online.com... > > > > Hello Ofer, > > > > I recommend to use SNiFF+ PRO as a WindPowerTool on top of Tornado. > > > > For more derails look at: > > > > www.windriver.com for a free eval download. > > > > > > > > Ciao > > > > > > > > Martin Raabe > > > > > > > > "Ofer Goren" schrieb im Newsbeitrag > > > > news:newscache$ku9rqg$lck$1@lnews.actcom.co.il... > > > > > Hi. > > > > > Does anyone work with this combination, and if so, any > recommendations > > > about > > > > > the best way to work? > > > > > > > > > > Tornado2.02 for win2kSp2 > > > > > PVCS for windows, version 6.7.11 > > > > > > > > > > Best reagrds, > > > > > > > > > > -- > > > > > Ofer Goren > > > > > Iamba Technologies > > > > > > > > > > > > > > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: rcmd for VxWorks Date: Thu, 31 Jan 2002 14:44:07 -0800 From: "Raj Sistla" Organization: Intel Corporation Message-ID: Hi, Does VxWorks have a Linux like "rcmd" command? I am trying to invoke a Linux executable from VxWorks. Thanks a bunch Raj --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado2 and PVCS Date: Thu, 31 Jan 2002 23:55:26 +0100 From: "Werner Schiendl" Message-ID: <3c59cb5b$1@brateggebdc5.br-automation.co.at> References: <3c587646$1@brateggebdc5.br-automation.co.at> <9ec53d11.0201310148.3c625ef5@posting.google.com> Hi, "Ofer Goren" wrote in message news:newscache$om9tqg$jhn$1@lnews.actcom.co.il... > Hi. > Let me rephrase my question: > Fact 1: We use Tornado 2.0.2 > Fact 2: WE USE PVCS. > No need for shouting, basically most source control software works similar. Using PVCS is no reason to not accept tips from people using ClearCase or SourceSafe or .... :-) > I was wonder what is the best way to work with this version control and > Tornado: > like, do you use archives, or flat trees? What exactly do you mean with this? We check in the plain source files, of course. > Do you put your project files under PVCS? Yes, because they contain many important information about your project. E. g. the VxWorks configuration for your bootable VxWorks applications. Build settings, file lists, etc. As already mentioned, be careful that PVCS does not change the line-endings of these files. > Do you put your makefiles, prjParams, prgConfig etc under PVCS? NO, never put any derived files into source control. You will have troubles whenever you try a build. > If you use archives, do you put them also under PVCS? What do you mean with archives? Third party libraries? Yes - it is just plain convenient if you can do a checkout / get latest version and have everything in place exactly as it was when you need to track down some problem in an old version of the software. > what do you do with you .out files? your .o files? your .exe files? you Nothing, usually. Make sure that your source control system is complete from time to time. It's probably worth to go through the trouble of trying a build on a blank machine to verify everything is under source control. If you want to archive your released software for easy access (and testing) later on, do so. But I wouldn't check in those binary files on a regular basis - just released versions. And do not check them in from your normal build directory (at least if PVCS marks checked in files as read-only, as VSS does). This would cause problems when building. > vxWorks files in case of bootable application? The same as above. If you check in the resulting objects, do so only for the end result (vxWorks, app.out). The intermediate files are of no use and eat up a lot of space in source control. > Do you compile your archives with your own makefiles, or use the Tornado > GUI? > Both, Tornado is used for development, everyone can build everything at any time from within Tornado. The official build is done in batch mode, w/o user intervention - based on the same configuration files. > All those question effect the way one work with PVCS, and the number of > files under the control of the source control utility. > hth Werner --------------------------- Newsgroups: comp.os.vxworks Subject: Re: DOS key like features in vxworks Date: Fri, 01 Feb 2002 09:54:51 +1030 From: Tim Shaw Organization: Defence Science and Technology Organisation Message-ID: <3C59D243.3192A23@dsto.defence.gov.au> References: <9b6f8883.0201311117.2e1b5206@posting.google.com> They're not the same as DOSKEY, but you can use vi commands for similar effect. Mark Chang wrote: > Does anyone know of code or have > code that will do DOS Key like features? > You know, up arrow show history, line editing, insert and replace. > I think it will have to replace tylib.c - > > TIA > Mark --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado2 and PVCS Date: 31 Jan 2002 19:29:50 -0500 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: References: <3c587646$1@brateggebdc5.br-automation.co.at> <9ec53d11.0201310148.3c625ef5@posting.google.com> <3c59cb5b$1@brateggebdc5.br-automation.co.at> Sender: bpringlemeir@DeadDuck >>>>> "Werner" == Werner Schiendl writes: [snip] >> Do you put your makefiles, prjParams, prgConfig etc under PVCS? Werner> NO, never put any derived files into source control. You Werner> will have troubles whenever you try a build. It is arguable whether the Makefile is `generated' or not. Certainly you must put the `.wpj'. I have never tried to regenerate the Makefile from the `.wpj'. Unless of course you have assumed that no one would use the Tornado projects? >> what do you do with you .out files? your .o files? your .exe >> files? you Werner> Nothing, usually. The best question (IMHO) would be what do you do with the Tornado header files. Developer X has T2.0 installed with WindML (no patches). Developer Y has T2.0.1 with WindML CP 1 and Developer Z has T2.1.1 with WindML CP2. Do you want whats behind door number 1, 2, or 3? I have been toying with the idea of putting the Tornado headers under source control so that every developer doesn't have to install every patch that WRS has ever produced. fwiw, Bill Pringlemeir. - -- I shall be well-worn. vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Determining the target IP address Date: Fri, 01 Feb 2002 04:04:55 GMT From: pkockritz@mac.com (Pete Kockritz) Organization: AT&T Broadband Message-ID: References: <3C595FEA.4AA9D8FA@improvsys.com> In article <3C595FEA.4AA9D8FA@improvsys.com>, Jeremiah Frick wrote: > Hi all! > > I'm trying to programatically determine the IP address of my VxWorks > Malta box. Unfortunately, I'm getting an error when I try to run this > code on the box: > > { > char addrArray[10]; > ifAddrGet( "lnPci0", addrArray ); > > printf( "IP Address: %s", addrArray ); > } You do realize that the array for the return string must contain at least INET_ADDR_LEN bytes (which is 18). You're probably clobbering your stack. Try this: #include #include { char addrArray[INET_ADDR_LEN]; if (ifAddrGet( "lnPci0", addrArray ) == OK) { printf( "IP Address: %s", addrArray ); } else { printf ("ifAddrGet failed\n"); } } - -- Regards, Pete Kockritz --------------------------- Newsgroups: comp.os.vxworks Subject: open changes the priority of my task ! Date: Thu, 31 Jan 2002 23:43:36 -0500 From: Yves Organization: Bell Sympatico Message-ID: <3C5A1CF8.72BA15AA@sympatico.ca> Hi all, we have C++ code that seems fine until it gets to an open statement, where the task priority changes from its value of 55 to 50 (tNetTask priority). Now stackTrace show plenty of room on the stack (before and after), the file gets opened and can be read correctly, but how can my task priority get changed (as seen using "i" command before and after the call). There is one other user task running, it polls until a flag gets set (so it should not be doing anything really). Stepped through with the debugger, the file name is a bit long (~64 characters), but is well formed. Any suggestions ? BTW: The file system on which the file resides is accessed via ftp. the file gets opened "read only". Thanks. Yves --------------------------- Newsgroups: comp.os.vxworks Subject: Re: open changes the priority of my task ! Date: Fri, 01 Feb 2002 05:23:01 GMT From: dale Message-ID: References: <3C5A1CF8.72BA15AA@sympatico.ca> In article <3C5A1CF8.72BA15AA@sympatico.ca>, yves.boudreault2 @sympatico.ca says... > Hi all, > > we have C++ code that seems fine until it gets to an open statement, > where the task priority changes from its value of 55 to 50 (tNetTask > priority). Now stackTrace show plenty of room on the stack (before and > after), the file gets opened and can be read correctly, but how can my > task priority get changed (as seen using "i" command before and after > the call). There is one other user task running, it polls until a flag > gets set (so it should not be doing anything really). Stepped through > with the debugger, the file name is a bit long (~64 characters), but is > well formed. Any suggestions ? > > > BTW: The file system on which the file resides is accessed via ftp. the > file gets opened "read only". > > Thanks. > > Yves > Your task is encountering a mutex semaphore "get" in a system call, where that mutex is in use by the nettask. Priority inversion applied to the mutex-get will change the owning and waiting task priorities to the higher of the two tasks. - ---- Dale --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Default scheduler in VxSim ? Date: 31 Jan 2002 22:04:11 -0800 From: deepaks@sasken.com (Deepak Shetty) Organization: http://groups.google.com/ Message-ID: <421ece65.0201312204.46ebe764@posting.google.com> References: <421ece65.0201300525.73bf4466@posting.google.com> <3c581398$1@brateggebdc5.br-automation.co.at> <421ece65.0201302142.39ce2dba@posting.google.com> <3c5919cb@brateggebdc5.br-automation.co.at> > I guess you are using the timerLib functions - which are (or at least try to > be) POSIX compatible. > POSIX by itself is not particularly designed for real-time apps. Can u elaborate on this more ? any links to info where its explained as to why POSIX is not suitable for realtime apps ? > > Anyway the timerLib builds on signals, which I did not use much 'til now. > If you want some kind of alarm or trigger after a certain amount of time, I > think the easiest approach in VxWorks is using wdLib (the VxWorks watchdog > facility). > The routine you've connected in to the watchdog will be called from > interrupt context (basically the system clock interrupt). > > Depending on what you are using the timer for, you can e. g. give a binary > semaphore to trigger a task which should perform the real action. > If you are using a watchdog as a guard to make sure something does not take > longer than acceptable, you may do whatever brings your device in a safe > fatal-error-state (shut down, reboot, whatever is appropriate). > > A note to the timers: You need not keep the task running in order to receive > the signal. The signal handler interrupts the current code anyway - so you > should of course receive the signal, if your task e. g. pends on a > semaphore - which you may then give in the signal handler. U need not keep the task running, in case of watchdog timers, as they execute the handler routine irrespective of the state of the task, even if its deleted, as told in teh doc, but in case of timerLib timers u need to keep the task active,as the routine in this case executes in this task's context, again as said in teh doc. > > > > > To make sure everything works as expected, you usually have watchdogs > > > guarding the software. > > > There are software watchdogs (like wdLib in VxWorks) and hardware > watchdogs. If wdLib is for software watchdog, what is the equivalent of hardware watchdog ? > > > > How is this guarding achieved ? > > For the software watchdog see above, and see the VxWorks documenation on > watchdogs (wdLib). > The action you take is completely dependent on your device. For some device, > shut down may be the safest - for other devices this would not be > acceptable. > > A hardware watchdog is a timer which must be reset from the software before > it expires. > If it is not, then (usually) the board is reset by hardware. > > This makes sure, that the system is running - if it locks up, e. g. in an > crazy ISR function, the hardware watchdog will reset the system to a safe > state. Does this mean that i shud use software watchdogs ( wdLib ) instead of POSIX timers as i was doign before, and this is more suited for real time apps under vxworks ? also if wdLib is software watchdog, where is hardware watch dog documented ? how do i access it, in case i need to ? In the doc. of wdLib, there is an example in which he uses this software watchdog to guard a routine, but from ur point i deduce that one needs to use hardware watchdog for the same ? I am bit confused in this hardware Vs software watchdog, are they the same, and their names ( h/w or s/w ) varies depending on the context in which it is used, like if used for guarding some routine, its called h/w wd else if used as a timer its called s/w wd ? Thanx, deepak --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Source for WRS build utilities? Date: Fri, 01 Feb 2002 03:41:11 GMT From: pkockritz@mac.com (Pete Kockritz) Organization: AT&T Broadband Message-ID: References: In article , pausmith@nortelnetworks.com wrote: > 2) deflate. I read somewhere that this tool used normal zlib > compression, but I haven't been able to get any zlib-based tool I > have to uncompress a file created by deflate, nor to compress a > file to be the same as the one deflate creates. I wrote a routine to uncompress an image compressed by deflate. The trick is that you have to skip the first byte. I think it's a 0x08; it's a tag that WRS adds to the start of the compressed block. Other than that it was just a matter of using the zlib routines to decompress it. I haven't tried to compress an image with anything other than deflate, but it shouldn't be too hard. I bet if you asked your FAE nicely, he'd get you the source to deflate and inflateLib. - -- Regards, Pete Kockritz --------------------------- Newsgroups: comp.os.vxworks Subject: Vxworks PPP: Login problem Date: Fri, 01 Feb 2002 06:39:43 GMT From: "v.papavinasam" Organization: AT&T Broadband Message-ID: Hello, I have a login problem over PPP link. I'm using Tornado1 Vxworks on a PPC system. After the PPP connection is established, starting a telnet session to the PPP IP-Address brings up the default windSh prompt ("->") but it is not accepting any keys at all, so I couldn't use the connection. In WRS Vxworks Network guide, there is a discussion about PPP connect and disconnect hooks, but I'm not sure whether I need to define those hook routines to fix this problem or this is a serial driver issue. If this is a hook routines problem, how they should be implemented? Please help. Thanks in advance. / V.P --------------------------- Newsgroups: comp.os.vxworks Subject: vxWorks and vxWorks.st Date: 31 Jan 2002 22:51:34 -0800 From: ramit.bhalla@wipro.com (Ramit Bhalla) Organization: http://groups.google.com/ Message-ID: <78ccdd99.0201312251.54553c53@posting.google.com> Hi, I'm a newbie to vxworks and I'm facing a few teething problems, hope some of you guy could help me out. I'm using Tornado 1.0.1 for the i86 architecture. I need to build the bootrom_uncmp image bootloader using a BSD style driver since it does not support an END style driver. There after when it downloads the vxWorks image from the host, it contains an END style driver. The problem is that is a build a vxWorks image, after it downloads the image it gives me a bind failure. If I use a vxWorks.st image, it work fine but the console comes up on the PC Console and the target server cannot attach. Why can't I use a vxWorks image ? What's the difference ? Also, when the bootloader loads, it initializes the Network card. When the vxWorks image loads does it restart eveything for uses the current network connection ??? Thanks, Ramit. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: 16550 UART driver Date: Fri, 01 Feb 2002 08:43:02 +0100 From: Emmanuel Herbreteau Organization: SEPRO Robotique (BEP) Message-ID: <3C5A4706.4C766158@sepro-robotique.com> References: <3C580FC8.8DAA7C55@sepro-robotique.com> <3C5901E5.B7FC9082@sepro-robotique.com> Hi Bill, Thanks for your "free advice". No, I don't need your help to get an UART driver working. I was just asking if a different driver can be find (I don't want to do something that could already be a part of Tornado). The latest news are : - - "ns16550sio.c" doesn't use FIFO and is the driver I have to change - - "st16552sio.c" use FIFO and it is the driver I need (june 1999 revision) This was the answer I was looking for. Regards Emmanuel. Bill Pringlemeir wrote: > Apparently you like to ignore free advice. Perhaps you can give me > the address of your company so that I can come and work for you for > free? --------------------------- Newsgroups: comp.os.vxworks Subject: Problem binding socket to IP address Date: Fri, 01 Feb 2002 07:51:54 GMT From: "Michael Rice" Message-ID: Hello, I am trying to bind() a socket to the IP address of one of my interfaces, but it fails and sets errno to 49 (EADDRNOTAVAIL). When I bind to INADDR_ANY, everything works fine. I am running VxWorks 5.4 on a PPC604. I have included some sample code (ServerTest shown below). I've also included the output from 'ifShow', 'routeShow', 'inetstatShow', and 'hostShow'. I would very much appreciate any suggestions as to what might be causing this problem. Thank you. Michael - -> ServerTest("192.168.1.200", 8000) Trying: bind() to 192.168.1.200:8000 Error binding socket: errno = 49! - -> ServerTest("", 8000) Trying: bind() to :8000 Trying: listen() to :8000 Everything is cool! - -> ifShow esyf (unit number 0): Flags: (0x263) UP BROADCAST RUNNING ARP Type: ETHERNET_CSMACD Internet address: 192.168.1.200 Broadcast address: 192.168.1.255 Netmask 0xffffff00 Subnetmask 0xffffff00 Ethernet address is 00:80:f6:01:40:39 Metric is 0 Maximum Transfer Unit size is 1500 1756 packets received; 0 packets sent 556 multicast packets received 12 multicast packets sent 0 input errors; 0 output errors 0 collisions; 0 dropped lo (unit number 0): Flags: (0x8069) UP LOOPBACK RUNNING ARP MULTICAST Type: SOFTWARE_LOOPBACK Internet address: 127.0.0.1 Netmask 0xff000000 Subnetmask 0xff000000 Metric is 0 Maximum Transfer Unit size is 32768 108 packets received; 108 packets sent 0 multicast packets received 0 multicast packets sent 0 input errors; 0 output errors 0 collisions; 0 dropped - -> routeShow ROUTE NET TABLE destination gateway flags Refcnt Use Interface - ---------------------------------------------------------------------------- 192.168.1.0 192.168.1.200 101 0 0 esyf0 - ---------------------------------------------------------------------------- ROUTE HOST TABLE destination gateway flags Refcnt Use Interface - ---------------------------------------------------------------------------- 127.0.0.1 127.0.0.1 5 1 105 lo0 - ---------------------------------------------------------------------------- - -> inetstatShow Active Internet connections (including servers) PCB Proto Recv-Q Send-Q Local Address Foreign Address (state) - -------- ----- ------ ------ ------------------ ------------------ ------- 17ae268 TCP 0 0 0.0.0.0.111 0.0.0.0.0 LISTEN 17ae0dc TCP 0 0 0.0.0.0.23 0.0.0.0.0 LISTEN 17ae058 TCP 0 0 0.0.0.0.513 0.0.0.0.0 LISTEN 17ade48 TCP 0 0 0.0.0.0.21 0.0.0.0.0 LISTEN 17ae1e4 UDP 0 0 0.0.0.0.111 0.0.0.0.0 17ae160 UDP 0 0 0.0.0.0.69 0.0.0.0.0 17adfd4 UDP 0 0 127.0.0.1.1024 127.0.0.1.17185 17adf50 UDP 0 0 0.0.0.0.17185 0.0.0.0.0 - -> hostShow hostname inet address aliases - -------- ------------ ------- uci 192.168.1.200 localhost 127.0.0.1 int ServerTest(char* szIPAddress, unsigned short nPort) { STATUS nStatus; int nSockServ; struct sockaddr_in sinServ; nSockServ = socket(AF_INET, SOCK_STREAM, 0); if(nSockServ != ERROR) { sinServ.sin_family = AF_INET; sinServ.sin_addr.s_addr = inet_addr(szIPAddress); sinServ.sin_port = htons(nPort); sinServ.sin_len = sizeof(struct sockaddr_in); printf("Trying: bind() to %s:%hu\n", szIPAddress, nPort); nStatus = bind(nSockServ, (struct sockaddr*)&sinServ, sizeof(sinServ)); if(nStatus == OK) { printf("Trying: listen() to %s:%hu\n", szIPAddress, nPort); nStatus = listen(nSockServ, 5); if(nStatus == OK) { printf("Everything is cool!\n"); } else { printf("Error listening on socket: errno = %d!\n", errno); } } else { printf("Error binding socket: errno = %d!\n", errno); } close(nSockServ); } else { printf("Error creating socket: errno = %d!\n", errno); } return 0; } --------------------------- Newsgroups: comp.os.vxworks Subject: Re: rcmd for VxWorks Date: Fri, 01 Feb 2002 07:58:25 GMT From: "Michael Rice" Message-ID: References: "Raj Sistla" wrote in message news:a3cheu$20a@news.or.intel.com... > Hi, > > Does VxWorks have a Linux like "rcmd" command? I am trying to invoke a Linux > executable from VxWorks. > > Thanks a bunch > Raj > > Yes, it's in remLib. You need to include RPC into VxWorks. The command looks like: rcmd() - execute a shell command on a remote machine int rcmd ( char * host, /* host name or inet address */ int remotePort, /* remote port to connect to (rshd) */ char * localUser, /* local user name */ char * remoteUser, /* remote user name */ char * cmd, /* command */ int * fd2p /* if this pointer is non-zero, stderr socket is */ /* and socket descriptor is filled in */ ) --------------------------- Newsgroups: comp.os.vxworks Subject: Re: what are these Assembler messages... Date: Thu, 31 Jan 2002 23:59:29 -0800 From: "GaryM" Message-ID: References: What is your host platform? "0xd" is the the ASCII Carriage Return character, which Unix hosts are loathe to recognize. If you are trying to port source files from a Microsoft host to a Unix host, you should run it through a filter (dos2unix) to strip out the CR's. "Ali" wrote in message news:b305b5bd.0201302344.6131ff25@posting.google.com... > Hi All > I am currently working on BSP for our custom board having MPC8260 > processor. > When i am compiling romInit.s file it shows me the assembler messages > like shown below.. > as a result my file does not get build can any body tell me why this > is so and what actually are these messages .. > How can i get rid of these messages.I am using tornado2.0 tool and my > operating system is vxworks. > The gnu tool im using is PPCEC603 . > What have to do so that i can build my file without this all going > on. > thanx in advance > ali > here is snapshot of these messages > > cd /home/alimohd/tornado/target/proj/eps8260Asm/PPCEC603gnu > make -f ../Makefile BUILD_SPEC=PPCEC603gnu rominit.o > ccppc -g -mstrict-align -ansi -nostdinc -fvolatile -fno-builtin > -fno-for-scope -P -x assembler-with-cpp -I. > -I/home/alimohd/tornado/target/h -DCPU=PPCEC603 -c > /home/alimohd/tornado/target/config/eps8260/rominit.s -o rominit.o > /var/tmp/ccpuaWfd.s: Assembler messages: > /var/tmp/ccpuaWfd.s:1: Error: Rest of line ignored. First ignored > character valued 0xd. > /var/tmp/ccpuaWfd.s:2: Error: Rest of line ignored. First ignored > character valued 0xd. > /var/tmp/ccpuaWfd.s:3: Error: Rest of line ignored. First ignored > character valued 0xd. > /var/tmp/ccpuaWfd.s:4: Error: Bad expression > /var/tmp/ccpuaWfd.s:5: Error: Rest of line ignored. First ignored > character valued 0xd. > /var/tmp/ccpuaWfd.s:6: Error: Rest of line ignored. First ignored > character valued 0xd. > /var/tmp/ccpuaWfd.s:7: Error: Rest of line ignored. First ignored > character valued 0xd. > /var/tmp/ccpuaWfd.s:14: Error: Rest of line ignored. First ignored > character valued 0xd. > /var/tmp/ccpuaWfd.s:15: Error: Rest of line ignored. First ignored > character valued 0xd. > /var/tmp/ccpuaWfd.s:27: Error: Rest of line ignored. First ignored > character valued 0xd. > /var/tmp/ccpuaWfd.s:28: Error: Rest of line ignored. First ignored > character valued 0xd. > /var/tmp/ccpuaWfd.s:4135: Error: Rest of line ignored. First ignored > character valued 0xd. > /var/tmp/ccpuaWfd.s:4138: Error: Rest of line ignored. First ignored > character valued 0xd. > /var/tmp/ccpuaWfd.s:4139: Error: Rest of line ignored. First ignored > character valued 0xd. > /var/tmp/ccpuaWfd.s:4140: Error: Rest of line ignored. First ignored > character valued 0xd. > /var/tmp/ccpuaWfd.s:4141: Error: Rest of line ignored. First ignored > character valued 0xd. > /var/tmp/ccpuaWfd.s:4142: Error: Rest of line ignored. First ignored > character valued 0xd. > /var/tmp/ccpuaWfd.s:4143: Error: Rest of line ignored. First ignored > character valued 0xd. > /var/tmp/ccpuaWfd.s:4144: Error: Rest of line ignored. First ignored > character valued 0xd. > /var/tmp/ccpuaWfd.s:4145: Error: Rest of line ignored. First ignored > character valued 0xd. > /var/tmp/ccpuaWfd.s:4147: Error: Rest of line ignored. First ignored > character valued 0xd. > /var/tmp/ccpuaWfd.s:4196: Error: Rest of line ignored. First ignored > character valued 0xd. > /var/tmp/ccpuaWfd.s:4197: Error: Rest of line ignored. First ignored > character valued 0xd. > /var/tmp/ccpuaWfd.s:4199: Error: Rest of line ignored. First ignored > character valued 0xd. > /var/tmp/ccpuaWfd.s:4200: Error: Bad expression > /var/tmp/ccpuaWfd.s:4201: Error: Rest of line ignored. First ignored > character valued 0xd. > /var/tmp/ccpuaWfd.s:4202: Error: Rest of line ignored. First ignored > character valued 0xd. > /var/tmp/ccpuaWfd.s:4203: Error: Rest of line ignored. First ignored > character valued 0xd. > /var/tmp/ccpuaWfd.s:4204: Error: Rest of line ignored. First ignored > character valued 0xd. > /var/tmp/ccpuaWfd.s:4205: Error: Rest of line ignored. First ignored > character valued 0xd. > /var/tmp/ccpuaWfd.s:4206: Error: Rest of line ignored. First ignored > character valued 0xd. > /var/tmp/ccpuaWfd.s:4852: Error: Rest of line ignored. First ignored > character valued 0xd. > ' but expected `,'s:4853: Error: syntax error; found ` > /var/tmp/ccpuaWfd.s:4855: Error: Rest of line ignored. First ignored > character valued 0xd. > ' but expected `,'s:4856: Error: syntax error; found ` > ' but expected `,'s:4857: Error: syntax error; found ` > ' but expected `,'s:4858: Error: syntax error; found ` > ' but expected `,'s:4859: Error: syntax error; found ` > /var/tmp/ccpuaWfd.s:4861: Error: Rest of line ignored. First ignored > character valued 0xd. > /var/tmp/ccpuaWfd.s:4862: Error: Rest of line ignored. First ignored > character valued 0xd. > /var/tmp/ccpuaWfd.s:4863: Error: Rest of line ignored. First ignored > character valued 0xd. > ' but expected `,'s:4864: Error: syntax error; found ` > ' but expected `,'s:4866: Error: syntax error; found ` > ' but expected `,'s:4868: Error: syntax error; found ` > ' but expected `,'s:4870: Error: syntax error; found ` > ' but expected `,'s:4872: Error: syntax error; found ` > ' but expected `,'s:4874: Error: syntax error; found ` > ' but expected `,'s:4876: Error: syntax error; found ` > ' but expected `,'s:4878: Error: syntax error; found ` > ' but expected `,'s:4880: Error: syntax error; found ` > ' but expected `,'s:4882: Error: syntax error; found ` > ' but expected `,'s:4884: Error: syntax error; found ` > ' but expected `,'s:4886: Error: syntax error; found ` > ' but expected `,'s:4888: Error: syntax error; found ` > ' but expected `,'s:4890: Error: syntax error; found ` > ' but expected `,'s:4892: Error: syntax error; found ` > ' but expected `,'s:4894: Error: syntax error; found ` > /var/tmp/ccpuaWfd.s:4896: Error: Rest of line ignored. First ignored > character valued 0xd. > /var/tmp/ccpuaWfd.s:4897: Error: Rest of line ignored. First ignored > character valued 0xd. > /var/tmp/ccpuaWfd.s:4898: Error: Rest of line ignored. First ignored > character valued 0xd. > ' but expected `,'s:4899: Error: syntax error; found ` > ' but expected `,'s:4901: Error: syntax error; found ` > ' but expected `,'s:4903: Error: syntax error; found ` > ' but expected `,'s:4905: Error: syntax error; found ` > /var/tmp/ccpuaWfd.s:4907: Error: Rest of line ignored. First ignored > character valued 0xd. > /var/tmp/ccpuaWfd.s:4908: Error: Rest of line ignored. First ignored > character valued 0xd. > /var/tmp/ccpuaWfd.s:4909: Error: Rest of line ignored. First ignored > character valued 0xd. > ' but expected `,'s:4910: Error: syntax error; found ` > ' but expected `,'s:4912: Error: syntax error; found ` > ' but expected `,'s:4914: Error: syntax error; found ` > ' but expected `,'s:4916: Error: syntax error; found ` > /var/tmp/ccpuaWfd.s:4918: Error: Rest of line ignored. First ignored > character valued 0xd. > /var/tmp/ccpuaWfd.s:4919: Error: Rest of line ignored. First ignored > character valued 0xd. > /var/tmp/ccpuaWfd.s:4920: Error: Rest of line ignored. First ignored > character valued 0xd. > ' but expected `,'s:4921: Error: syntax error; found ` > ' but expected `,'s:4922: Error: syntax error; found ` > ' but expected `,'s:4923: Error: syntax error; found ` > /var/tmp/ccpuaWfd.s:4924: Error: Rest of line ignored. First ignored > character valued 0xd. > /var/tmp/ccpuaWfd.s:4925: Error: Rest of line ignored. First ignored > character valued 0xd. > /var/tmp/ccpuaWfd.s:4926: Error: Rest of line ignored. First ignored > character valued 0xd. > /var/tmp/ccpuaWfd.s:4996: Error: Rest of line ignored. First ignored > character valued 0xd. > ' but expected `,'s:4997: Error: syntax error; found ` > ' but expected `,'s:4998: Error: syntax error; found ` > /var/tmp/ccpuaWfd.s:4999: Error: Rest of line ignored. First ignored > character valued 0xd. > ' but expected `,'s:5000: Error: syntax error; found ` > /var/tmp/ccpuaWfd.s:5001: Error: Rest of line ignored. First ignored > character valued 0xd. > make: *** [rominit.o] Error 1 > hit ENTER to exit --------------------------- Newsgroups: comp.os.vxworks Subject: How to auto format code in Tornado Date: 1 Feb 2002 00:10:28 -0800 From: deepaks@sasken.com (Deepak Shetty) Organization: http://groups.google.com/ Message-ID: <421ece65.0202010010.2ca001c1@posting.google.com> Hi, I am interested in auto-formatting of the code as i write it under tornado, how do i acheive that ? ie when i write an if / while statement, & open a '{' then the next code shud auto start from 4/8 spaces after the column where i typed '{' in the last line deepak --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to auto format code in Tornado Date: 1 Feb 2002 08:17:31 GMT From: Denis Perelyubskiy Message-ID: References: <421ece65.0202010010.2ca001c1@posting.google.com> * Deepak Shetty [02-Fri-02 00:10 -0800]: > > Hi, > I am interested in auto-formatting of the code as i write it under > tornado, how do i acheive that ? dont write it in tornado :)... there are plenty of editors available for writing code. many can be automated to do compilation, etc... but that depends on your setup. i.e, vim, emacs (free), textpad (pretty cheap), visual slickedit, sniff+... more expensive, and a boatload of other things i haven't mentioned yet... its especially easy in t2, where you can define external editors... anyway, the point of this post: dont write it in tornado's default editor. its not very good.. denis - -- if you want to reply by mail, please tweak my address in a reasonable way. sorry for the inconvenience. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: vxWorks and vxWorks.st Date: 1 Feb 2002 10:49:18 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: <78ccdd99.0201312251.54553c53@posting.google.com> ramit.bhalla@wipro.com (Ramit Bhalla) wrote: > The problem is that is a build a vxWorks image, after it downloads the > image it gives me a bind failure. If I use a vxWorks.st image, it work > fine but the console comes up on the PC Console and the target server > cannot attach. > > Why can't I use a vxWorks image ? What's the difference ? The standalone version (vxWorks.st) does not initialise the network interfaces. But you can try to initialise the network card from the target shell and see where it fails. (see the file target/config/comps/src/net/usrNetConfigIf.c, function usrNetIfConfig) > Also, when the bootloader loads, it initializes the Network card. When > the vxWorks image loads does it restart eveything for uses the current > network connection ??? No, the VxWorks image restarts the network interface. Groeten, Johan - -- o o o o o o o . . . _____________________________ o _____ || Johan Borkhuis | .][__n_n_|DD[ ====_____ | borkhuis@agere.com | >(________|__|_[_________]_|__________________________| _/oo OOOOO oo` ooo ooo 'o!o!o o!o!o` === VxWorks FAQ: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html === --------------------------- Newsgroups: comp.os.vxworks Subject: Re: open changes the priority of my task ! Date: Fri, 1 Feb 2002 11:15:22 +0000 (UTC) From: David Laight Organization: BT Openworld Message-ID: <3C5A791F.E812FD36@l8s.co.uk> References: <3C5A1CF8.72BA15AA@sympatico.ca> > > > > we have C++ code that seems fine until it gets to an open statement, > > where the task priority changes from its value of 55 to 50 (tNetTask > > priority). > > Your task is encountering a mutex semaphore "get" in a system call, where > that mutex is in use by the nettask. Priority inversion applied to the > mutex-get will change the owning and waiting task priorities to the > higher of the two tasks. Your task must also hold another mutex that has the 'priority inversion' bit set. vxWorks is too stupid to drop your priority when releasing a mutex if you hold another one. Clearly the full set of checks is hard, however some simple tests - 'is any process waiting on a mutex that has priority inversion enabled' would help! David --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Configuration of a serial device to run full duplex Date: Fri, 1 Feb 2002 10:17:40 +0100 From: "Friedrich Ensslin" Message-ID: References: <3c57b060$0$89074$edfadb0f@dspool01.news.tele.dk> Hi Jakob, it seems to me, that the problem might have to do with character buffering inside the driver. Your observations might be caused by the fact, that the receiver side of the driver does no sufficient data buffering. When your code arrives at the read instruction, the first three bytes are already lost (what's the data rate on /tyCo/1 ?). To clarify this, it might be helpful to look at the driver's configuration sequence and sourcecode if available. What's the size of the buffers (especially RX) ? Try to start a parallel task doing the while/read loop before writing to the loopback. Friedrich Ensslin ________________________________________ AM3 AutoMotive MultiMedia AG friedrich.ensslin@am3.com ________________________________________ "Jakob Holm-Petersen" schrieb im Newsbeitrag news:3c57b060$0$89074$edfadb0f@dspool01.news.tele.dk... > Hi, > > I'm right now working on a project where we need to make a physical loop > back on a serial device. The loop back function is used to verify that all > transmitted chars on the device are properly transmitted. My problem is that > it seems like that the serial device is running half duplex. > > My source code looks like this: > > fd = open ("/tyCo/1", O_RDWR, 1); > write (fd, "abcde", 5); > i=0; > while (i < 5) > { > i += read (fd, &rxBuffer[i], 1); > } > > The result is that the while loop is newer completed because the read > function only read the last 2 chars "de" in the string "abcde". > > Do anyone have a solution of an explanation of this problem? > > Best regards > > Jakob Holm-Petersen > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Design of Flash file system Date: 1 Feb 2002 03:58:10 -0800 From: kamala.kodi@wipro.com (Kamala) Organization: http://groups.google.com/ Message-ID: <27203e86.0202010358.70ffeac6@posting.google.com> Hi, We need to develop Flash file system for Intel E28F128J3A flash on vxworks. pls. let me know where can i get Links / documents / source code / etc.... for the same. Thanks & regards, kamala. --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Sat Feb 2 04:03:22 2002 From: Vxworks Exploder Date: Sat Feb 2 04:03:25 PST 2002 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sat Feb 2 04:03:15 PST 2002 Subject: Re: Default scheduler in VxSim ? Subject: transition from bootloader to vxworks Subject: Re: Tornado2 and PVCS Subject: Re: Determining the target IP address Subject: Re: Tornado2 and PVCS Subject: Re: transition from bootloader to vxworks Subject: Re: semTake and ISR Subject: Re: routing problem ??? Subject: Still "dosFsDevInit failed". Subject: Re: Still "dosFsDevInit failed". Subject: Re: Tornado2 and PVCS Subject: can't compile a simple mib Subject: Re: Tornado2 and PVCS Subject: Re: Tornado2 and PVCS Subject: Re: Design of Flash file system Subject: Problem with multicast using vxsim Subject: Problem with multicast using vxsim Subject: target server (serial) connect fails if cache is disabled... Subject: Re: Tornado2 and PVCS Subject: Re: can't compile a simple mib Subject: Re: login issue Subject: Re: login issue Subject: Re: Tornado2 and PVCS Subject: Re: login issue Subject: Re: Tornado2 and PVCS Subject: Re: Tornado2 and PVCS Subject: Re: Design of Flash file system Subject: Re: 16550 UART driver Subject: Re: Tornado2 and PVCS Subject: Re: Tornado2 and PVCS Subject: Re: login issue Subject: RTOS Evaluation Comments Request Subject: +++ C-Designer (sehr flexibel) für Realtime/C/Systeme sucht neues Projekt +++ Subject: vxEieio Subject: Re: RTOS Evaluation Comments Request Subject: Re: can't compile a simple mib Subject: What's usage of ccppc? Subject: Re: What's usage of ccppc? Subject: Tornado-2.1/mips and cp7000 BSP (Ocelot MIPS board) Subject: Freebie VxWorks Tornado? Subject: This is Cool 90548 Subject: Returning memory Subject: MPC8260, vector table @ 0xfff00000 Subject: Re: Returning memory ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Default scheduler in VxSim ? Date: Fri, 1 Feb 2002 13:42:30 +0100 From: "Werner Schiendl" Message-ID: <3c5a8d33$1@brateggebdc5.br-automation.co.at> References: <421ece65.0201300525.73bf4466@posting.google.com> <3c581398$1@brateggebdc5.br-automation.co.at> <421ece65.0201302142.39ce2dba@posting.google.com> <3c5919cb@brateggebdc5.br-automation.co.at> <421ece65.0201312204.46ebe764@posting.google.com> Hi, "Deepak Shetty" wrote in message news:421ece65.0201312204.46ebe764@posting.google.com... > > I guess you are using the timerLib functions - which are (or at least try to > > be) POSIX compatible. > > > POSIX by itself is not particularly designed for real-time apps. > > Can u elaborate on this more ? any links to info where its explained > as to why POSIX is not suitable for realtime apps ? > I did NOT say it is not suitable, just not particularly designed for this purpose. Therefore, some idioms are not ideally adapted to VxWorks. AFAIK most posix APIs are designed on top of the VxWorks native APIs. So if size does matter for your target image, you may be able to safe some kB using VxWorks' native APIs. If you need to have the software as portable as possible, POSIX usually is a good choice. > > > > Anyway the timerLib builds on signals, which I did not use much 'til now. > > If you want some kind of alarm or trigger after a certain amount of time, I > > think the easiest approach in VxWorks is using wdLib (the VxWorks watchdog > > facility). > > The routine you've connected in to the watchdog will be called from > > interrupt context (basically the system clock interrupt). > > > > Depending on what you are using the timer for, you can e. g. give a binary > > semaphore to trigger a task which should perform the real action. > > If you are using a watchdog as a guard to make sure something does not take > > longer than acceptable, you may do whatever brings your device in a safe > > fatal-error-state (shut down, reboot, whatever is appropriate). > > > > A note to the timers: You need not keep the task running in order to receive > > the signal. The signal handler interrupts the current code anyway - so you > > should of course receive the signal, if your task e. g. pends on a > > semaphore - which you may then give in the signal handler. > > U need not keep the task running, in case of watchdog timers, as they > execute the handler routine irrespective of the state of the task, > even if its deleted, as told in teh doc, but in case of timerLib > timers u need to keep the task active,as the routine in this case > executes in this task's context, again as said in teh doc. > Yes, this is how I read the docs. BUT: Keeping active means only that you must not delete the task. Pending tasks receive signals (at least they should), because the signal routine is not dependent (or - it MUST NOT be dependent) on the state of the tasks program flow. Signals can occur at any time while the task lives. If you need to understand how the routine is invoked in detail (using timerLib) - see the documentation on signals. timerLib is using signals to invoke the routine (as stated in the docs as well). > > > > > > > > To make sure everything works as expected, you usually have watchdogs > > > > guarding the software. > > > > There are software watchdogs (like wdLib in VxWorks) and hardware > > watchdogs. > > If wdLib is for software watchdog, what is the equivalent of hardware > watchdog ? > This is completely hardware dependent. Most hardware watchdogs are simply triggered by writing to a particular memory or IO address. Maybe you need to alter the value written to this location between to subsequent triggers. However, hardware watchdogs and how to use them must be documented with your hardware. > > > > > > How is this guarding achieved ? > > > > For the software watchdog see above, and see the VxWorks documenation on > > watchdogs (wdLib). > > The action you take is completely dependent on your device. For some device, > > shut down may be the safest - for other devices this would not be > > acceptable. > > > > A hardware watchdog is a timer which must be reset from the software before > > it expires. > > If it is not, then (usually) the board is reset by hardware. > > > > This makes sure, that the system is running - if it locks up, e. g. in an > > crazy ISR function, the hardware watchdog will reset the system to a safe > > state. > > Does this mean that i shud use software watchdogs ( wdLib ) instead of > POSIX timers as i was doign before, and this is more suited for real > time apps under vxworks ? As stated above, this depends on what you want to achieve. wdLib is somewhat easier to use - in my opinion - an has less overhead (no signals). Therefore I prefer them, if they are sufficient. Of course it is perfectly ok to use timerLib if you need or want to. > also if wdLib is software watchdog, where is > hardware watch dog documented ? how do i access it, in case i need to > ? In the doc. of wdLib, there is an example in which he uses this > software watchdog to guard a routine, but from ur point i deduce that > one needs to use hardware watchdog for the same ? I am bit confused in > this hardware Vs software watchdog, are they the same, and their names > ( h/w or s/w ) varies depending on the context in which it is used, > like if used for guarding some routine, its called h/w wd else if used > as a timer its called s/w wd ? > Hardware watchdogs are very reliable, compared to software watchdogs. Do you know a bit of x86 ASM? how about CLI ; lock all interrupts JMP $ ; spin forever Some code like this one can lock your system completely, a software watchdog cannot become effective - since interrupts are locked. A hardware watchdog, however, is independent of the CPU to go off and reset the CPU (and hardware) into a safe state. The difference between software and hardware watchdogs is, that hardware watchdogs are solded onto the CPU board (and therefore specific to the target in use, the VxWorks simulator has _NO_ hardware watchdogs), while software watchdogs are a software concept. While you can have only a very limited number of hardware watchdogs (because of board size, cost, etc.), software watchdogs are basically limited by system memory and performance (eventually) of the system. Does this clarify the difference between the two concepts? hth Werner --------------------------- Newsgroups: comp.os.vxworks Subject: transition from bootloader to vxworks Date: 1 Feb 2002 04:54:15 -0800 From: ramit.bhalla@wipro.com (Ramit Bhalla) Organization: http://groups.google.com/ Message-ID: <78ccdd99.0202010454.6cabbf7d@posting.google.com> Hi, Hoping some of the experts could give me some tips. When using Tornado 1.0.1 ot Tornado 2.0.2 what is the transition scheme like. When the bootloader loads, it initialized the ethernet interface to download the image. When the image is loaded and starts execution does it reuse the same network interface or does it reinitialize the whole network again ??? Thanks, Ramit. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado2 and PVCS Date: Thu, 31 Jan 2002 01:19:44 +0000 From: molochai Message-ID: References: Martin Raabe wrote: > Hello Ofer, > I recommend to use SNiFF+ PRO as a WindPowerTool on top of Tornado. > For more derails look at: > www.windriver.com for a free eval download. > Didn't I read somewhere (probably on here in fact) that WRS employees are strictly prohibitted by WRS from posting to comp.os.vxworks? I'm sure that I'm not the only one to be getting totally pissed off with your repeated attempts to make sales in responce to questions. M > Ciao > > Martin Raabe > > "Ofer Goren" schrieb im Newsbeitrag > news:newscache$ku9rqg$lck$1@lnews.actcom.co.il... >> Hi. >> Does anyone work with this combination, and if so, any recommendations > about >> the best way to work? >> >> Tornado2.02 for win2kSp2 >> PVCS for windows, version 6.7.11 >> >> Best reagrds, >> >> -- >> Ofer Goren >> Iamba Technologies >> >> --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Determining the target IP address Date: Fri, 01 Feb 2002 08:06:49 -0500 From: Jeremiah Frick Organization: Improv Systems, Inc. Message-ID: <3C5A92E9.6BCF22D6@improvsys.com> References: <3C595FEA.4AA9D8FA@improvsys.com> <3c5970ce@brateggebdc5.br-automation.co.at> Werner, That was the problem. I actually noticed it about a half an hour after I posted this here. Thanks all! Jeremy Werner Schiendl wrote: > Hi, > > "Jeremiah Frick" wrote in message > news:3C595FEA.4AA9D8FA@improvsys.com... > > Hi all! > > > > I'm trying to programatically determine the IP address of my VxWorks > > Malta box. Unfortunately, I'm getting an error when I try to run this > > code on the box: > > > > { > > char addrArray[10]; > > ifAddrGet( "lnPci0", addrArray ); > > > > printf( "IP Address: %s", addrArray ); > > } > > > > The error is: > > > > (KLSI CartRootTaskd found. > > ): Can't attach unknown device lnpci (unit 0). > > > > Is this a typo, or does the error message _indead_ display the device name > in all lowercase? > A compare of "lnPci" with "lnpci" will fail, if str(n)cmp() function is > used! > > > I'm fairly certain lnPci is the correct device. Any thoughts? > > > > Try a different name for the network device, or a network board with a name > in all-lowercase letters. > > hth > Werner --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado2 and PVCS Date: Fri, 1 Feb 2002 14:01:13 +0100 From: "Werner Schiendl" Message-ID: <3c5a9197@brateggebdc5.br-automation.co.at> References: <3c587646$1@brateggebdc5.br-automation.co.at> <9ec53d11.0201310148.3c625ef5@posting.google.com> <3c59cb5b$1@brateggebdc5.br-automation.co.at> Hi, The makefile is regenerated automatically if it is missing. It is also _overwritten_ whenever you change the project file. The only thing that causes some troubles is generating these files _without_ the Tornado II IDE. There are some suggestions for that in Johan's FAQ pages on how to achieve this. I've hacked some changes into prjConfig.c for testing myself, since it safes you changing the configlettes (which are global for the installation). However, this is not a goog idea for production use, since it is overwritten whenever Tornado thinks the project file has changed. Checking in the Tornado target tree as a whole is a good idea in my opinion. The biggest advantage is, that service packs and patches are propagated automatically to other team members and _even_more_important_ to the build machine. An other advantage is, that if you need to debug something in an old version, you have the correct header files and libraries. WARNING: If you check in the library files of VxWorks (which we did, for the reasons mentioned) - do not forget to check them out _before_ installing an add on package like WindWeb Server. Otherwise the installer will SILENTLY fail to update your libraries :-) best regards Werner "Bill Pringlemeir" wrote in message news:ug04mhwn5.fsf@yahoo.com... > >>>>> "Werner" == Werner Schiendl writes: > > [snip] > > >> Do you put your makefiles, prjParams, prgConfig etc under PVCS? > > Werner> NO, never put any derived files into source control. You > Werner> will have troubles whenever you try a build. > > It is arguable whether the Makefile is `generated' or not. Certainly > you must put the `.wpj'. I have never tried to regenerate the > Makefile from the `.wpj'. Unless of course you have assumed that no > one would use the Tornado projects? > > >> what do you do with you .out files? your .o files? your .exe > >> files? you > > Werner> Nothing, usually. > > The best question (IMHO) would be what do you do with the Tornado > header files. Developer X has T2.0 installed with WindML (no > patches). Developer Y has T2.0.1 with WindML CP 1 and Developer Z has > T2.1.1 with WindML CP2. Do you want whats behind door number 1, 2, or > 3? > > I have been toying with the idea of putting the Tornado headers under > source control so that every developer doesn't have to install every > patch that WRS has ever produced. > > fwiw, > Bill Pringlemeir. > > -- > I shall be well-worn. > vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" --------------------------- Newsgroups: comp.os.vxworks Subject: Re: transition from bootloader to vxworks Date: Fri, 1 Feb 2002 14:27:32 +0100 From: "Werner Schiendl" Message-ID: <3c5a97c1$1@brateggebdc5.br-automation.co.at> References: <78ccdd99.0202010454.6cabbf7d@posting.google.com> Hi, VxWorks uses only the bootline from the bootrom image, and initializes everything from scratch. At least for pc486 and pcPentium BSP. This is true for both, Tornado 1.0.1 and Tornado 2.0.x. Unless the address, where the bootline is stored, has changes since 1.0.1 - there is no immediate need to compile a new bootrom, AFAIK. The biggest transition work is to change your config.h modifications into a working VxWorks bootable project. You can search for the macro names in the VxWorks configuration tool, if you can get all your changes to config.h and friends from source control (or memory ;-) ) Your application should run without changes, unless you hacked something in VxWorks 5.3's sources. hth Werner "Ramit Bhalla" wrote in message news:78ccdd99.0202010454.6cabbf7d@posting.google.com... > Hi, > > Hoping some of the experts could give me some tips. > > When using Tornado 1.0.1 ot Tornado 2.0.2 what is the transition > scheme like. > > When the bootloader loads, it initialized the ethernet interface to > download the image. > > When the image is loaded and starts execution does it reuse the same > network interface or does it reinitialize the whole network again ??? > > Thanks, > Ramit. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: semTake and ISR Date: Fri, 1 Feb 2002 15:40:00 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <9b6f8883.0201301813.e5ac79c@posting.google.com> semTake does not do anything in ISR, because all semLib functions have two distinct implementations, one for task context and one for Interrupt context, and there is no implementation of semTake for ISR context, so it will return ERROR, that's all. By the way, there are different implementations of semTake which also depend on the type of semaphore used, B, M or C. "Mark Chang" wrote in message news:9b6f8883.0201301813.e5ac79c@posting.google.com... > I know that > The semTake( ) routine is not callable from interrupt service routines. > > So what happens when I have a semTake in an IRS? > Does it just return? > > Any damage? Since semTake basically does an IntLock. > > And to answer your question as to why I would > do this. I don't, but some code I was debugging > did and I'm wondering about the effects, if any. > > TIA > Mark --------------------------- Newsgroups: comp.os.vxworks Subject: Re: routing problem ??? Date: Fri, 1 Feb 2002 15:45:33 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: Hey Xpert, I dont want to sound rude, but your knowledge of routing is similar to my knowldge of airoplanes. Would you imagine I developed an airpolane ? ... neither should you attempt deveoping a router. "it2xpert" wrote in message news:a2jojc$6eu$1@hiline.shinbiro.com... > Hi , vxworks seniors !!! > I desinged the mpc850sr board for router... > I implemented the scc2,scc3 for ethernet mode... > > I use the T2, vxworks 5.4... > I want to implement the Layer 3 router..... > > 192.168.1.0 -----------------192.168.1.1 (scc3) > > (scc2) > 210.118.156.240 -------------------210.118.156.0 > > I used the RIP v.1 > Routing table is following.... > > -> routeShow > > ROUTE NET TABLE > destination gateway flags Refcnt Use Interface > -------------------------------------------------------------------------- - -- > 0.0.0.0 210.118.156.254 3 0 21 cpm0 > 192.168.1.0 192.168.1.1 101 0 0 cpm1 > 210.118.156.0 210.118.156.240 101 0 0 cpm0 > -------------------------------------------------------------------------- - -- > > ROUTE HOST TABLE > destination gateway flags Refcnt Use Interface > -------------------------------------------------------------------------- - -- > 127.0.0.1 127.0.0.1 5 1 65 lo0 > -------------------------------------------------------------------------- - -- > value = 77 = 0x4d = 'M' > -> > > I trid the ping test !!! > My pc is connected the 192.168.1.0 network (192.168.1.3) > ping 192.168.1.1 ----- OK, > ping 210.118.156.240 ---- OK > > but i did't reach the 210.118.156.xxx point.... > > I tried ping to the target server console.... > -> ping "210.118.156.xxx" > this is OK.... > > what's wrong ???? > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Still "dosFsDevInit failed". Date: 01 Feb 2002 12:46:37 GMT From: Charly Message-ID: <200221-134637-250830@foorum.com> References: <2002130-143411-880483@foorum.com> Thanks for your help, The problem is still there "dosFsDevInit failed". When I start a big application on my PCPentium target, the shell display "dosFsDevInit failed". Then, the target seems dead, no Shell and no target server connection. This application works fine on an other BSP (Lp-elektronik). Do you have an idea where I could look? VxWorks configuration? All my files are on a FTP server. I use also a RAMDISK. Here are the result of some shell command. - -> devs drv name 0 /null 1 /tyCo/0 1 /tyCo/1 2 /pcConsole/0 2 /pcConsole/1 6 host: 7 /vio value = 0 = 0x0 - -> ls Volume in drive D has no label Volume Serial Number is 12E3-344A Directory of D:\kern\Awd_num\T2mak\rtl [.] [..] RTL.WPJ USRAPP~1.C PRJCOMPS.H PRJPAR~1.H LINKSYMS.C PRJCON~1.C MAKEFILE [DEFAULT] PRJOBJS.LST ADD_VXW.C USRAPP~2.C [DEBUG] 10 file(s) 98,644 bytes 4 dir(s) 1,170,317,312 bytes free value = 0 = 0x0 - -> version VxWorks (for PC PENTIUM) version 5.4.2. Kernel: WIND version 2.5 Made on Feb 1 2002, 13:09:39. Boot line: ene(0,0)host:vxWorks e=192.0.1.2:ffffff00 h=192.0.1.1 u=target pw=vxworks f=0x8 tn=target o=ene. value = 0 = 0x0 - -- Use our news server 'news.foorum.com' from anywhere. More details at: http://nnrpinfo.go.foorum.com/ --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Still "dosFsDevInit failed". Date: Fri, 1 Feb 2002 15:21:55 +0100 From: "Werner Schiendl" Message-ID: <3c5aa480$1@brateggebdc5.br-automation.co.at> References: <2002130-143411-880483@foorum.com> <200221-134637-250830@foorum.com> Hi Charly, As I see from your shell commands' output, you are trying those in WindShell (Tornado II host shell). This can cause confusion, because devs lists the devices on the target, while ls lists the directory of your Windows Box. Type @ls to execute the command on your target. Try @pwd as well, if the above does not help. If you do not have those included in your image, you can try ioDefPathGet(), alternatively. Something like: bufx=malloc(1000) ioDefPathGet(bufx) printf(bufx) in the WindShell. You need to find out, what your target tries to access, when you get that error message. It's probably using some 'relative' path. Did you try to access a file on your FTP server with it's absolute path? like (in WindShell) fdx = open("host:/test.txt", 0) read(fdx, bufx, 1000) d bufx Does this work? (note the _forward_ slash) Report your findings. hth Werner "Charly" wrote in message news:200221-134637-250830@foorum.com... > > Thanks for your help, > The problem is still there "dosFsDevInit failed". > > When I start a big application on my PCPentium target, the shell display > "dosFsDevInit failed". Then, the target seems dead, no Shell and no target > server connection. This application works fine on an other BSP (Lp-elektronik). > Do you have an idea where I could look? VxWorks configuration? > All my files are on a FTP server. > I use also a RAMDISK. > > Here are the result of some shell command. > -> devs > drv name > 0 /null > 1 /tyCo/0 > 1 /tyCo/1 > 2 /pcConsole/0 > 2 /pcConsole/1 > 6 host: > 7 /vio > value = 0 = 0x0 > > -> ls > > Volume in drive D has no label > Volume Serial Number is 12E3-344A > Directory of D:\kern\Awd_num\T2mak\rtl > > [.] [..] RTL.WPJ USRAPP~1.C PRJCOMPS.H > PRJPAR~1.H LINKSYMS.C PRJCON~1.C MAKEFILE [DEFAULT] > PRJOBJS.LST ADD_VXW.C USRAPP~2.C [DEBUG] > 10 file(s) 98,644 bytes > 4 dir(s) 1,170,317,312 bytes free > value = 0 = 0x0 > > -> version > VxWorks (for PC PENTIUM) version 5.4.2. > Kernel: WIND version 2.5 > Made on Feb 1 2002, 13:09:39. > Boot line: > ene(0,0)host:vxWorks e=192.0.1.2:ffffff00 h=192.0.1.1 u=target pw=vxworks f=0x8 > tn=target > o=ene. > value = 0 = 0x0 > > > -- > Use our news server 'news.foorum.com' from anywhere. > More details at: http://nnrpinfo.go.foorum.com/ --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado2 and PVCS Date: Fri, 1 Feb 2002 15:25:46 +0100 From: "Werner Schiendl" Message-ID: <3c5aa567$1@brateggebdc5.br-automation.co.at> References: Hi, "molochai" wrote in message news:g36a3a.na3.ln@vapour-trail.demon.co.uk... > Martin Raabe wrote: > > For more derails look at: > > www.windriver.com for a free eval download. And I thought derails was just a typo :-] > > Didn't I read somewhere (probably on here in fact) that WRS employees are > strictly prohibitted by WRS from posting to comp.os.vxworks? > > I'm sure that I'm not the only one to be getting totally pissed off with > your repeated attempts to make sales in responce to questions. > > M > --------------------------- Newsgroups: comp.os.vxworks,comp.protocols.snmp Subject: can't compile a simple mib Date: 1 Feb 2002 06:38:00 -0800 From: ilyasg@yahoo.com (Ilyas G) Organization: http://groups.google.com/ Message-ID: HJi i have the fellowing ASN1 declaration - ----------------------------------- MYPERSONAL-MIB DEFINITIONS ::= BEGIN IMPORTS OBJECT-TYPE FROM RFC-1212; internet OBJECT IDENTIFIER ::= { 1 3 6 1 } experimental OBJECT IDENTIFIER ::= { internet 3 } myDir OBJECT IDENTIFIER ::= { experimental 17 } myObject OBJECT-TYPE SYNTAX OCTET STRING (SIZE (0..255)) ACCESS read-write STATUS mandatory DESCRIPTION "Un object string cree par moi meme a des fin de test." ::= { myDir 1 } END - ----------------------------------- i compile it using the VxWorks mibcompile, and it generate a c file mymib.c the c file generated contain static object - --------------- STATIC MIBARC_T a_zeroDotZero[] = { END_OF_ARC_LIST }; - ---------------------- the problem is that END_OF_ARC_LIST is not declared anywhere I am really novice on this stuff ! thanks for help --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado2 and PVCS Date: 1 Feb 2002 06:40:14 -0800 From: igtorque@eliop.es (=?ISO-8859-1?Q?Ignacio_Gonz=E1lez?=) Organization: http://groups.google.com/ Message-ID: <9ec53d11.0202010640.5085a2bc@posting.google.com> References: <3c587646$1@brateggebdc5.br-automation.co.at> <9ec53d11.0201310148.3c625ef5@posting.google.com> <3c59cb5b$1@brateggebdc5.br-automation.co.at> Bill Pringlemeir wrote in message news:... [snip] > > I have been toying with the idea of putting the Tornado headers under > source control so that every developer doesn't have to install every > patch that WRS has ever produced. > We put under version control every Tornado file (.c or .h) that has to be changed to support one of our products. For instance, we have put files like: C:\Tornado\target\config\comps\src\romStart.c C:\Tornado\target\config\comps\src\sysClkInit.c C:\Tornado\target\config\comps\src\usrSerial.c C:\Tornado\target\config\comps\src\version.c C:\Tornado\target\config\comps\src\net\usrNetConfigIf.c because our different products must use different revisions of these files in order to be correctly built. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado2 and PVCS Date: Fri, 01 Feb 2002 10:02:41 -0500 From: Doug Dotson Organization: Verio Message-ID: <3C5AAE11.70A9752@digidata.com> References: <3c587646$1@brateggebdc5.br-automation.co.at> <9ec53d11.0201310148.3c625ef5@posting.google.com> <3c59cb5b$1@brateggebdc5.br-automation.co.at> Technically, everything that is used to build a project should be under version control. That includes system libraries, system header files, the compilers, linkers, etc. I've never worked in a place that actually does it but a guy I worked with from HP said it is SOP there. doug Bill Pringlemeir wrote: > >>>>> "Werner" == Werner Schiendl writes: > > [snip] > > >> Do you put your makefiles, prjParams, prgConfig etc under PVCS? > > Werner> NO, never put any derived files into source control. You > Werner> will have troubles whenever you try a build. > > It is arguable whether the Makefile is `generated' or not. Certainly > you must put the `.wpj'. I have never tried to regenerate the > Makefile from the `.wpj'. Unless of course you have assumed that no > one would use the Tornado projects? > > >> what do you do with you .out files? your .o files? your .exe > >> files? you > > Werner> Nothing, usually. > > The best question (IMHO) would be what do you do with the Tornado > header files. Developer X has T2.0 installed with WindML (no > patches). Developer Y has T2.0.1 with WindML CP 1 and Developer Z has > T2.1.1 with WindML CP2. Do you want whats behind door number 1, 2, or > 3? > > I have been toying with the idea of putting the Tornado headers under > source control so that every developer doesn't have to install every > patch that WRS has ever produced. > > fwiw, > Bill Pringlemeir. > > -- > I shall be well-worn. > vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Design of Flash file system Date: Fri, 01 Feb 2002 10:08:35 -0500 From: Doug Dotson Organization: Verio Message-ID: <3C5AAF73.7AFA025E@digidata.com> References: <27203e86.0202010358.70ffeac6@posting.google.com> Try searching for Tiny File System. It's a free, small file system for flash. We used it successfully but decided that TFFS was better for our purposes. Try this link: www.embedded.com/internet/0012/0012ia2.htm doug Kamala wrote: > Hi, > > We need to develop Flash file system for Intel E28F128J3A flash on > vxworks. pls. let me know where can i get Links / documents / source > code / etc.... for the same. > > Thanks & regards, > kamala. --------------------------- Newsgroups: comp.os.vxworks Subject: Problem with multicast using vxsim Date: Fri, 1 Feb 2002 09:56:54 -0500 From: "Bhaskar Rao" Message-ID: <3c5ab126$0$17869$4c41069e@reader1.ash.ops.us.uu.net> Hi, I am a new comer to vxworks. I am facing some basic problem. I am using full version of VxWorks simulator. I builded the vxworks.exe image for full networking support(By including basic network components, changing the flag to FALSE in config.h, installed ULIP driver, giving MCAST ROUTING options etc... as per the instructions in SPR 30084, as I am using Win2K). I am assigning IP address for the simulated target as 90.0.0.1 and from another host if I try to communicate to this simulated target, using unicast sockets I succeed and everything works fine. But when my application running on simulated target try to receive multicast packet(after joining the multicast group with IP_ADD_MEMBERSHIP options.. using setsockopt), it won't receive any multicast packets(Multicast cast address is in the proper range224.0.0.0 to 239.255.255.255) where as setsockopt returns success and more over on the shell if I do ifShow instead of showing IFF_MULTICAST flag as explained in the manual, it just shows MULTICAST. Is there anything wrong I am doing for this or missing out anything for multicast. (I am posting message to the news group for the first time, any suggestions for the way of posting messages are welcomed.) thanks in advance regards Bhaskar UTStarcom 33 Wood Avenue South Iselin, NJ08830 Tel :(7325167536) --------------------------- Newsgroups: comp.os.vxworks Subject: Problem with multicast using vxsim Date: Fri, 1 Feb 2002 10:16:06 -0500 From: "Bhaskar Rao" Message-ID: <3c5ab127$0$17869$4c41069e@reader1.ash.ops.us.uu.net> Hi, I am a new comer to vxworks. I am facing some basic problem. I am using full version of VxWorks simulator. I builded the vxworks.exe image for full networking support(By including basic network components, changing the flag to FALSE in config.h, installed ULIP driver, giving MCAST ROUTING options etc... as per the instructions in SPR 30084, as I am using Win2K). I am assigning IP address for the simulated target as 90.0.0.1 and from another host if I try to communicate to this simulated target, using unicast sockets I succeed and everything works fine. But when my application running on simulated target try to receive multicast packet(after joining the multicast group with IP_ADD_MEMBERSHIP options.. using setsockopt), it won't receive any multicast packets(Multicast cast address is in the proper range224.0.0.0 to 239.255.255.255) where as setsockopt returns success and more over on the shell if I do ifShow instead of showing IFF_MULTICAST flag as explained in the manual, it just shows MULTICAST. Is there anything wrong I am doing for this or missing out anything for multicast. (I am posting message to the news group for the first time, any suggestions for the way of posting messages are welcomed.) thanks in advance regards Bhaskar UTStarcom 33 Wood Avenue South Iselin, NJ08830 Tel :(7325167536) --------------------------- Newsgroups: comp.os.vxworks Subject: target server (serial) connect fails if cache is disabled... Date: 1 Feb 2002 07:41:52 -0800 From: fillmore2@arcor.de (Achim) Organization: http://groups.google.com/ Message-ID: <9d082700.0202010741.14bed109@posting.google.com> Hi all, I need to disable the cache to debug some driver adaption for a x86 BSP (Pentium): When I'm trying to disable caching (either 'cache enable' setting in VxWorks or cacheDisable() call during initialization), the target server (serial connection) fails to connect. I've already tried other settings for the serial connection and checked for code corruption during initialization... Have I been missing some additional cache initialization - or any other ideas? Thanks. Achim --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado2 and PVCS Date: Fri, 1 Feb 2002 15:50:05 +0000 (UTC) From: David Laight Organization: BT Openworld Message-ID: <3C5AB981.44CF4FE9@l8s.co.uk> References: <3c587646$1@brateggebdc5.br-automation.co.at> <9ec53d11.0201310148.3c625ef5@posting.google.com> <3c59cb5b$1@brateggebdc5.br-automation.co.at> <3C5AAE11.70A9752@digidata.com> Doug Dotson wrote: > > Technically, everything that is used to build a project should be under > version control. That includes system libraries, system header files, > the compilers, linkers, etc. I've never worked in a place that actually > does it but a guy I worked with from HP said it is SOP there. A good check - take a 'new' disk, install the OS, install any 'obvious' additional packages, load in the archive of your source control, extract them all (or maybe extract the 'magic' file that knows how to extract everyting else), type 'make'. If you are missing something, go put it under source control. (now work out how to put those config options you selected on a 'gui' interface under source control.) David --------------------------- Newsgroups: comp.os.vxworks,comp.protocols.snmp Subject: Re: can't compile a simple mib Date: 1 Feb 2002 15:55:39 GMT From: Juergen Schoenwaelder Message-ID: <1012578939.423733@salvator.ibr.cs.tu-bs.de> References: In comp.protocols.snmp Ilyas G wrote: > i have the fellowing ASN1 declaration I am not sure why this specific compiler fails, but I strongly suggest to use SMIv2 rather than SMIv1. As a quick start, here is your module in SMIv2 format: MYPERSONAL-MIB DEFINITIONS ::= BEGIN IMPORTS OBJECT-TYPE, MODULE-IDENTITY, experimental FROM SNMPv2-SMI; myDir MODULE-IDENTITY LAST-UPDATED "200202010000Z" ORGANIZATION "..." CONTACT-INFO "Ilyas G " DESCRIPTION "..." ::= { experimental 17 } myObject OBJECT-TYPE SYNTAX OCTET STRING (SIZE (0..255)) MAX-ACCESS read-write STATUS current DESCRIPTION "Un object string cree par moi meme a des fin de test." ::= { myDir 1 } END You can easily turn that into an SMIv1 module by using smidump or other MIB compilers - if you really need to do that. Note that SMIv2 is out since 1994 and a full Standard since 1999. Note also that using { experimental 17 } is only allowed if you never use it anywhere outside of your lab. /js - -- Juergen Schoenwaelder University of Osnabrueck Dept. of Mathematics and Computer Science Phone: +49 541 969 2483 Albrechtstr. 28, 49069 Osnabrueck, Germany Fax: +49 541 969 2770 --------------------------- Newsgroups: comp.os.vxworks Subject: Re: login issue Date: Fri, 01 Feb 2002 10:55:27 -0500 From: david lindauer Message-ID: <3C5ABA6F.E5D4982B@notifier-is.net> References: <26ca8a8f.0201300030.47a0fa90@posting.google.com> <3C58628B.50702@notifier-is.net> <26ca8a8f.0201302203.893596c@posting.google.com> hi, we found out that the encrypted password 'LOGIN_PASSWORD' does NOT match the password given by vxencrypt. We cannot add new users/passwords, but we can successfully delete and readd the default password, but only if we use the LOGIN_PASSWORD version rather than the encrypted given by vxencrypt. Trying to delete and readd the default using the vxencrypt version of the password does not work. This used to work because we have a project that successfully adds and deletes passwords based on the vxencrypt value; we are hypothesizing that the transition to windweb 2.0 broke vxencrypt but don't know for sure. It is somewhat puzzling becase we can do loginUserDelete and it will detect the vxEncrypt password properly (you can't delete a user unless you know his password), but if we do loginUserVerify it will not. We are still unsure what to do... I'm thinking about installing the version of tornado that came with windweb 2.0 to see if I get a version of vxEncrypt that works properly. David senthil wrote: > Initally, i got the encrypted password using the exe vxencrpt in bin > folder. Even after adding a new user with the encrypted password i was > not able to login. Inorder to override the encrypt call back we > installed our routine at the start of the application. As below > > /**** insall user defined encrption routine ***/ > loginEncryptInstall((FUNCPTR)encryptRoutine,0); > > /* our encryption test routine just have the same passoword added */ > STATUS > encryptRoutine(char *password, char *encryptedPassword) > { > strcpy(encryptedPassword, password); > return OK; > } > > Using this encryption routine the default > user : target pswrd: password got added as we expected. Then we added > another user . We were able to login using target user but not with > any other user. We even had some debug mechanisms to prove that the > user defined encryptRouting( ) is called when the loginUserVerify( ) > function is called. The loginUserVerify( ) entered into our encrypt > routine but it always returned -1 for new users added and was > sucessful for target user. > > Please let me know how to overcome this issue. > > regards > senthil --------------------------- Newsgroups: comp.os.vxworks Subject: Re: login issue Date: Fri, 01 Feb 2002 11:18:33 -0500 From: david lindauer Message-ID: <3C5ABFD9.EDC813EF@notifier-is.net> References: <26ca8a8f.0201300030.47a0fa90@posting.google.com> <3C58628B.50702@notifier-is.net> <26ca8a8f.0201302203.893596c@posting.google.com> <3C5ABA6F.E5D4982B@notifier-is.net> hm, vxencrypt is giving the wrong encrypted values. We ran our password through loginDefaultEncrypt() at run-time and were able to log in ok. David david lindauer wrote: > hi, > > we found out that the encrypted password 'LOGIN_PASSWORD' does NOT match > the password given by vxencrypt. We cannot add new users/passwords, but > we can successfully delete and readd the default password, but only if we > use the LOGIN_PASSWORD version rather than the encrypted given by > vxencrypt. Trying to delete and readd the default using the vxencrypt > version of the password does not work. > > This used to work because we have a project that successfully adds and > deletes passwords based on the vxencrypt value; we are hypothesizing that > the transition to windweb 2.0 broke vxencrypt but don't know for sure. It > is somewhat puzzling becase we can do loginUserDelete and it will detect > the vxEncrypt password properly (you can't delete a user unless you know > his password), but if we do loginUserVerify it will not. > > We are still unsure what to do... I'm thinking about installing the > version of tornado that came with windweb 2.0 to see if I get a version of > vxEncrypt that works properly. > > David > senthil wrote: > > > Initally, i got the encrypted password using the exe vxencrpt in bin > > folder. Even after adding a new user with the encrypted password i was > > not able to login. Inorder to override the encrypt call back we > > installed our routine at the start of the application. As below > > > > /**** insall user defined encrption routine ***/ > > loginEncryptInstall((FUNCPTR)encryptRoutine,0); > > > > /* our encryption test routine just have the same passoword added */ > > STATUS > > encryptRoutine(char *password, char *encryptedPassword) > > { > > strcpy(encryptedPassword, password); > > return OK; > > } > > > > Using this encryption routine the default > > user : target pswrd: password got added as we expected. Then we added > > another user . We were able to login using target user but not with > > any other user. We even had some debug mechanisms to prove that the > > user defined encryptRouting( ) is called when the loginUserVerify( ) > > function is called. The loginUserVerify( ) entered into our encrypt > > routine but it always returned -1 for new users added and was > > sucessful for target user. > > > > Please let me know how to overcome this issue. > > > > regards > > senthil --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado2 and PVCS Date: Fri, 1 Feb 2002 17:45:39 +0100 From: "Werner Schiendl" Message-ID: <3c5ac630$1@brateggebdc5.br-automation.co.at> References: <3c587646$1@brateggebdc5.br-automation.co.at> <9ec53d11.0201310148.3c625ef5@posting.google.com> <3c59cb5b$1@brateggebdc5.br-automation.co.at> <3C5AAE11.70A9752@digidata.com> <3C5AB981.44CF4FE9@l8s.co.uk> Hi, "David Laight" wrote in message news:3C5AB981.44CF4FE9@l8s.co.uk... ...snip... > > If you are missing something, go put it under source control. > > (now work out how to put those config options you selected on a 'gui' > interface under source control.) > > David this seems to be one of Tornado's strengths, everything seems to be in the .wpj file. And the source files of course :-) regards Werner --------------------------- Newsgroups: comp.os.vxworks Subject: Re: login issue Date: Fri, 1 Feb 2002 17:43:00 +0100 From: "Werner Schiendl" Message-ID: <3c5ac592@brateggebdc5.br-automation.co.at> References: <26ca8a8f.0201300030.47a0fa90@posting.google.com> <3C58628B.50702@notifier-is.net> <26ca8a8f.0201302203.893596c@posting.google.com> <3C5ABA6F.E5D4982B@notifier-is.net> <3C5ABFD9.EDC813EF@notifier-is.net> Hi, I remember to have read something about a changed password hashing in VxWorks some time ago. Don't know if this has to do anything with WindWeb, some Cumulative Patch, or so. I think there should have been an updated version of vxencrypt.exe as well. Maybe you give it a try and search for it on WindSurf... hth Werner "david lindauer" wrote in message news:3C5ABFD9.EDC813EF@notifier-is.net... > hm, vxencrypt is giving the wrong encrypted values. We ran our password > through loginDefaultEncrypt() at run-time and were able to log in ok. > > David > > david lindauer wrote: > > > hi, > > > > we found out that the encrypted password 'LOGIN_PASSWORD' does NOT match > > the password given by vxencrypt. We cannot add new users/passwords, but > > we can successfully delete and readd the default password, but only if we > > use the LOGIN_PASSWORD version rather than the encrypted given by > > vxencrypt. Trying to delete and readd the default using the vxencrypt > > version of the password does not work. > > > > This used to work because we have a project that successfully adds and > > deletes passwords based on the vxencrypt value; we are hypothesizing that > > the transition to windweb 2.0 broke vxencrypt but don't know for sure. It > > is somewhat puzzling becase we can do loginUserDelete and it will detect > > the vxEncrypt password properly (you can't delete a user unless you know > > his password), but if we do loginUserVerify it will not. > > > > We are still unsure what to do... I'm thinking about installing the > > version of tornado that came with windweb 2.0 to see if I get a version of > > vxEncrypt that works properly. > > > > David > > senthil wrote: > > > > > Initally, i got the encrypted password using the exe vxencrpt in bin > > > folder. Even after adding a new user with the encrypted password i was > > > not able to login. Inorder to override the encrypt call back we > > > installed our routine at the start of the application. As below > > > > > > /**** insall user defined encrption routine ***/ > > > loginEncryptInstall((FUNCPTR)encryptRoutine,0); > > > > > > /* our encryption test routine just have the same passoword added */ > > > STATUS > > > encryptRoutine(char *password, char *encryptedPassword) > > > { > > > strcpy(encryptedPassword, password); > > > return OK; > > > } > > > > > > Using this encryption routine the default > > > user : target pswrd: password got added as we expected. Then we added > > > another user . We were able to login using target user but not with > > > any other user. We even had some debug mechanisms to prove that the > > > user defined encryptRouting( ) is called when the loginUserVerify( ) > > > function is called. The loginUserVerify( ) entered into our encrypt > > > routine but it always returned -1 for new users added and was > > > sucessful for target user. > > > > > > Please let me know how to overcome this issue. > > > > > > regards > > > senthil > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado2 and PVCS Date: 01 Feb 2002 12:26:04 -0500 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: References: <3c587646$1@brateggebdc5.br-automation.co.at> <9ec53d11.0201310148.3c625ef5@posting.google.com> <3c59cb5b$1@brateggebdc5.br-automation.co.at> <9ec53d11.0202010640.5085a2bc@posting.google.com> Sender: bpringlemeir@DeadDuck >>>> Bill Pringlemeir writes: >> I have been toying with the idea of putting the Tornado headers >> under source control so that every developer doesn't have to >> install every patch that WRS has ever produced. >>>>> "Ignacio" == Ignacio González writes: Ignacio> We put under version control every Tornado file (.c or .h) Ignacio> that has to be changed to support one of our products. For Ignacio> instance, we have put files like: Ignacio> C:\Tornado\target\config\comps\src\romStart.c [snip] Well, there is one problem. The "C:/Tornado/target". I want to make the build drive independent (note, this is a deficiency in Windoze) and the WPJ files and other include absolute information and drive letters. In another post, David Laight gave the step of installing the OS. This in itself _COULD_ cause a difference in generated code. I suppose that that is the key point. You should have everything under version control that is needed to replicate the output [vxWorks images]. I feel much more confident in recommending that the Tornado files are put under revision control. I may change the location of the files to be something other than $(WIND_BASE). People often run out of space on the "C:" drive. Especially if this is where the "Windows" directory lies... `mount' is a beautiful thing! Regards, Bill Pringlemeir. - -- I'm meditating on the FORMALDEHYDE and the ASBESTOS leaking into my PERSONAL SPACE!! vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado2 and PVCS Date: Fri, 1 Feb 2002 17:31:20 +0000 (UTC) From: David Laight Organization: BT Openworld Message-ID: <3C5AD13D.814EF53B@l8s.co.uk> References: <3c587646$1@brateggebdc5.br-automation.co.at> <9ec53d11.0201310148.3c625ef5@posting.google.com> <3c59cb5b$1@brateggebdc5.br-automation.co.at> <3C5AAE11.70A9752@digidata.com> <3C5AB981.44CF4FE9@l8s.co.uk> <3c5ac630$1@brateggebdc5.br-automation.co.at> > > this seems to be one of Tornado's strengths, everything seems to be in the > .wpj file. > And the source files of course :-) Dunno - I did try generating a 'project' once, got the error 'List does not have an even number of elements', never got a resolution :-) Mind you the 'project' stuff seemed to assume your BSP was a fixed item and would never change.... David --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Design of Flash file system Date: Fri, 1 Feb 2002 19:27:30 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <27203e86.0202010358.70ffeac6@posting.google.com> Once upon a time there was a FlashLib, a true Flash File System, in contrast to FTL (Flash Translation Layer) like Tffs. "Kamala" wrote in message news:27203e86.0202010358.70ffeac6@posting.google.com... > Hi, > > We need to develop Flash file system for Intel E28F128J3A flash on > vxworks. pls. let me know where can i get Links / documents / source > code / etc.... for the same. > > Thanks & regards, > kamala. begin 666 flashLib.new.datasheet.pdf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`G]/^3YHY++K3W0O'0UE;F1S=')E86T-96YD M;V)J#3"!;(# @," U.34@.#0R M(%T@#2]#'0@72 - M+T9O;G0@/#P@+T8S(#0R(# @4B O1C0@,S<@,"!2("]&-2 T," P(%(@+T8V M(#,V(# @4B O1C<@-3$@,"!2("]&." T.2 P(%(@#2]&.2 T-R P(%(@/CX@ M#2]%>'1'4W1A=&4@/#P@+T=3,2 V," P(%(@+T=3,B V,2 P(%(@/CX@#2]# M;VQONZ-3N<>ZKLF3$F@RW7$9D6+C>[O88/&-O!6.\5'7A M\ZP8KHR'@RWZ[>T/[.A./'XKQBGV__*)>TTFA009UPG$7(*@I7SHC M"V^_"1>GY 3PJG=Z@W91K<@]Q6ZD? M\&&+SAUG.#H;BAI OA91`8+@^3QJXJ;0D?8']( H@%;.VT\Z!:D_R1^$P3!! M[')+-^.ZCZGL%J6;+^T,W3)ZMZD)I)F_+D!=UQ296!+*RE"*OB;1JI]YA@`\ MOR6DI ?W#8.COL!H[?RTH?U,A<']6--U,TA4DV&R4=MV!>2K,'D-*SHDM<2> M@MN7TT%JAQ5]0$>>KK0KUS%&AJ5?!:^41F6]^EM#%H"J68*L.2PE,1#A60I/ M,'J K]%:^,T6DU^91+DP&I=WM@S+H&`ZHG@=]K^Q? 92ML;P@Z ME'94NTJ5/RZRY7=)Y/"L'#D:=WZPIC,011&?NPM0O; G=5X+0(\R=J=^O,B6 M7[Y;#<'!0$/==8X\+QNL_.JO/>*GF%.ISV#4@>:%8A7=;7:1KB +Y%,WY^F- M!IB6_#K%)=E0`KT67YXJOTA3U\F:63=QJY^&\(UF]X..;X M2ZR::I;$V7H1*B!1,+AD;\>=&8H-6&P+8>_0O6HGT$%MO\!S3X/IKIG,MS(( M\&",$L^;N$D!]-2D8A<4.2<]31&P@D?0)*6$]"+!X SST4-$!,ALKEEEK24] M)'_ +<[;')@$^]=]GJ:/$%KS]S^$8^TD.2OA3^/3[U]B*+Y J%:-C47.@W9_ M%6V4OH':3JJ7:N<"C5OE8M9M]J;?MV;@1G[W<-RV\[ #$+\JLOE9CO3R?M[3 M:2QE6A&!$$;:4"-=FD>OQO%&6$VZN#'N+(%.Y#5BUA,7[Q4:SC[TE M#FU$630;D6%L6NF%771U W01<$=G[T%E2'6 M_"%1SP7X]LF!$7!UBZ7RCD0%^'>4L_>V# $+V7<5`;H3`S,/`IBPOCE*_=YG MZ3GMUC4TH%T722.F2QO+%>+X06^=K-[V;PD*1)_QJR<1)+";F3@$E3WF`&"9 MB,4[M3-P-H_]X6%>_A2[812ZT7[PF!OZ@=65I\^Y"-O_?)R:20GOW0+G*"5D MWV9H5]N_TM=_RG'F^OZ/HTM$1>YFD\=/W=#,KZLR)DS3"^O"9. B5IM1. "P MD291(\>"HG#X3X!Q<'T8.+%5D^+\<.M6>I1E0O M'D64?RPO%:;PS*:8=BX^ MITI#7N5L].4T? \M=91'8=.>E/M#V8>GPM4$7^,,7"FY?3Z,(:";*+@%B*#A M&] E*,TH1^7"+NO1Q%]TOZ^Y=Q/G^:U'JVR@5*]@7[T_K0BW54Y,%4' M&;JW`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`]N>R3I!-]D1K_@VA-P>8*UOOKN'E3OP%*A^B3: '0QH*>22 M%%\U1\XD3SAB/:H;XW"@JA*<%+$_-M[P//57)%MQ0TX+64%T&R&'P: M*(58W;K['>VTHY20D]Q[Y$01WO\@0"]NHR$FDV$P7K^$].+@.TGG?UQ=J^-_ MN1.!`@@&GK((#^&+<4Y&";H$*8C!^I01-X-XASS]J=SV2E0)\Q\B]>NJ)+0C M\6ZO"QG4+5T`5'][\_LK(@$830[:!:=!K,*K&W>V8T_$FODLB]9-QSACK3IY M\8II5(/H-)OCF0!09.1H&940IO^S-6(:[%7IV>"T38:9!ZGW:7I_'X`,++F*,#9RBK*'>:^A(=TCP7DS54 M4O!1ZVE8T +!:_\'AE[_Z_TRFBT+4H*D(YYAE9)\%3NRNLT6$TI[O-H.=QFE M[R^B-XUKE 2-R3$[-+"G^=(N3@/'H ,QQU67DR'+@>:FGT/C6Y[4`A!#`IQ@ M>_\-"R+H\N[B`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`QGR;>6R.:B0K.DQ_[XFHYF;6JDQVZ/7 MEI4ZTX$Z=IG%B-D]U>/7*T=2TTTN":G9L'N'`A4V2/N(!@;3-$@--O"/O'V# M>S&#@Y71P3\\/W3QLU0J>2EN_8YW@-10KQ@*E5K-3NHT8[QJ`9<^DP=M(4?@ MO7CM40K_ M)CM'/;BOR6HBWML,_GC1KS?6B7?&M(&,(JP9[#6L#,GX+A@&2@ZVU=6"'X<% M$W=OR=,*9 E^`6'S&)][]-!*L UE;F1S=')E86T-96YD;V)J#3,Q(# @;V)J M#3$T-C<@#65N9&]B:@TS,B P(&]B:@T\/" O1FEL=&5R("]&;&%T941E8V]D M92 O3&5N9W1H(#,Q(# @4B ^/B -T/Q045#U2S;H+$8<> M]N/*7+8`"\C#WH7+\VNVUX. M>81 `J1I`F!BGM0^."@P]8:U#NG#3"RP??J0:E[XF#%8%%;7"HYW-XV(*L^2Y"'J)#P+5J7ZW=)VIB8IZ2S7:B"!:AKD"8#@G M0=>FYX$*#Q)SULC&?(Q\A_$D>I9\LO8T0MU*:B(-"E_Y+Q&88):XB^T9S DQ MUQY1(7/\9^-O+*5"[L96YR,J-^XX[]PER14HZ)"TG(8VXCMP5=RJ3X*1F9DH M4;4)ISRO/!9K[XE2TCFBZTU&D:B?7Y,DW(\MP&7RXDB9+IW[J M@$=\6P3]T">W:"S9!L=S;9N[5C(`MO?04T6@%![OJ;JHD[TSW)86ARJG0++K MS$ZHAJ/L+G%IYJK 1QAFD+LI%7F(R( M0-=;EO-YK,5288*LF,Q^-\1(]L/19?VIK[_YHB>?\GYJX SW`E89[U XB).J MKEC0;M\VOBMG6G;]K4JB=A03I>N_IZU:.NX(7#:]'Y60*,A"[9;X MCR1T0%@:4'4*A-SC?PT[]O,`(;T7U3%F]R>!U)Q2L(CA D\=OOVGX:JQ; YJ7^](OQ](OM.OZ@3Z=\TFC+V<(OH M4#.K>K#G;>0!6XB!P])VK. )5&BL\";?Q6HL)5(J^#\0(FM2<0X!<=, M>8MD3[#/27^6AB[H;YWY(O95PMB[XX%GV;UH]/J6GE-G^4_Y/\0,O_&%R*]9 M:^ Z]9?E&"8QO@3ZP72VXI\C]I6V,B"NY"9?'UA?0D) XWD-7O^._=?L+]:+ MI/@=(%A+F0V4<]W>U0Q,Z F>NPX7)T_PWSGM+1)WG)=;D87,@ZM:2T1:OKOK MMXZT3V"P(X(K?LGK`.V2,%QW0CS8HZQSV,>%C'3[MN/_;W(!TJK_/P197-^P M$)B@B'QHVWV=B MV $RS)NGT96A<4@-96YD5D6>10<(7 &%D@8,U/3C)S<:;@=PC##<]'?BA!\- MPPKXB0LHQ1(ACUCQ!NV]'"K4)VSDN]%TYOB ?9A0&U:.PBS<5J2:30C%/3S) M;.#Z_#Q*5P!_7O:EF>=:\-"F.T\0$?IHD0US'9<7ALDF,Y!NGY0MB!]VASW& MRJDF#C66CYH4!=D7Y0(.\3V%Y8]+`GM!,_'O<\Y:TCDA`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`KF3P@_C$GQ>15F#>JBVRM2RFJHM M>'&D=?C(-&>69-K4>0RGDL)3P%-SNC1C?8UUA31J784C/*5"\ZK$P-2D1R(`-TR+ M_#"TI4C1I/CO#1F]IF&NW3\DY1<`:DV/-@8;9B &H;0J_2[G5:B;^O[22$<- MT=*T+3/"CKAF/5%P`6&M>&0&L=+.D?\;K!C^BU0VMXT3M*91=Q'OPF?9R3*: MJ!?C"S7J:.\^93=I-CADX+D8]WOQC)QU]10D!X>EK;D))7\9?6!.RCC ,1M$ M[@D M;Q:0]J?G\7'(J(X_N;CM%J,'\0OM*R?U$O]=>N"4L9T-]@ZCW>K%XK(4B<]Q MTNE"%?K'(P'^!-HA)5J5UW!QS]0"%SY2=&T&R.F\,ODV];B>!91`GH.1=YQ,,$DB.!G MM+_YL!&_W5MWDE$C"0AU8V[RT#P0.P`:YN3HFFY374 DR> M70*T/.52XF=9HG:(`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`4_.6U38'NEP-756]R<2/J_FF=VKNX='2$"O<6/(I M#2]&;VYT1FEL93,@-C,@,"!2( T^/B -96YD;V)J#3,Y(# @;V)J#3P\( TO M5'EP92 O1F]N=$1EM1=8'9)]<)(3%G0QM]^.):]<#8:47S"<%M"P M`>P46'W_KG"P2%%;`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`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`@H+ M^*P9A9WB?*OPU=RX,[=)`R)(E#4I*[6HWF_Y<^_G(>]!=FM^?DK#AYB]O,$# M_5E^!Q+P,S@^"XU,S]VF*1WU,38B`[Z"!6O$%]WFP1DX+"<[UZ=>+P<<>?:" M-6#I6*22*ET@%<"<+O4WO=59#,9Q1MV="+$=0:-7#*C2K2HSMXANO[2U_QFR M>[[S"^UM"9T&#\$W5>&S(LTM(?44UB(=E$I5,(-2K3 Y[]5S)L496CKY:5%:\^5;65!G^M *"9T>S&@ M=VN!F%E8"'1LLU$(RC-EH.NA(MK^^NU=,+^BWS>OT0A/9+0`U#O@]3,9!*#Q M^FM.?1DVS3.3=%%1%+6AF0^W+@(IU138QT%M CPCTA]1!CP MQ5:/E/V09>6]%#QLO$NQ[7C=;+[:_9> M`W1U'?44H?0?:814L>T8045=-$) -38`FO-$J^ G%ZH.A++IT3Q*+-7F1N1M M%?='^GXK*J\;S9QA@&/Q4"CXI2Y>L$@0LQ3'P]N'Z)19K]S=_Y(\T;&EG9/L M+]&93=-X$*ASB$.W^0X5/]]6[,J\.L\AOBG5I*:G2V#L17'8`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`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`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`:O7VT(N,@KNJ7YM1/WF=XD_(#-Y,J^XOS M!TZ":/MM,FCS$X2NQ97?I*(O4E/!82]!TFA82'Q(Q3?/)VZ7IQ>?41)N65JJ M)+>X6S(\^X-JMV\_NV^/<3DXY&K3WMA2,6(G/B7YBT69@UY2;PL]3K!)HPK' M+,)SHS(1%#R12U](&L19X%=&*(>G=R(=4#%GM BK&5 D(.=VFBX$&E(`(?\Y M[HU7_C0X6"OK)6DD11:_@(9S?QF]F!8\!89F*S=BN,M6U?-ZTMP#`.($'[PU MBL1-X4U6Y\B,OJ/F7:BRJ1 NO=K&KU2K)EG=\0-5SU67:UP&DL*2'(FM>WS^ M2/C!!5J2]P1Q7F&Y:(Y$@8P!\Y]]+1YCWZF;1'YZFSR4EO,S^,M-VDY*!BV) MWPH>$.SYYX1SW$# 4,6[(8X&`Y'<_:^SE0C0>BVB1HXJ>JG4^@Q_DXNM,?I_ MZ@'W\BL)L*SR?C[BMIOJ[W=Z@KQ+8*/38E5!&L@[P(+61/X;Y@+P.A6M9I7O M$'[RF?)1]!,2*M@.J 'G_Z4MGS-]3'\B!(]9@W' ZI)F`-/G^%S"+J2>E3>V MKIV38[VVX1<(1 VZKN&@O$ZZ7:LUN/^5/X$J5"LRMQF@_A4F&>L YK-XF\9L"C(F?C:S;YQCOYLOY2 MI_L%V05!L%R62*&:- M!1,D1),KHT6T(BSJ8H"V8I0QV4U#.:WB(@/L+Z/%Y]P7-%T-96YD@&T]A_ST<2Y]9ED#;S).T5@0TQ:F%F='*("8>?N>6UJ(+: MDZITX%=(H@>.'6VYS%B%<1$.X3?.[TF<^FC*\&.&XB1F4)&?;_"QN20L8B.H>^WMK1>=UJER1/]/0@%2T)#E8:LQF MPN6#;L.$S\HY/H%]&S5.!V0'7_U^K[.H;>QT184`BP8?=FL&J: 5E%EAQHER MRPKFEKKX//29PH.L7C?'Q8-[&* 8;"9$1MS4XT$S3E&?7--%TO:?57R_A@X6=1643#' MM]TJ$C/T;"==#P:B[LJ";XY8@BT$$'Q2U`#?P2V+DE\65;=USU?*FGRP'NIKSH_ MPMY@6&DO.[(U\HS/XH\?VR8@#6,:;=XYD^W'GC4KE:R"Z\DAI5= 6F44?3%( MKSJP[09"![W!'8PW*^T2#>A[L4N-%M]Q"QW_;I9CJ6#_;XP8-7$0Y3-AY*\` MK732'QX,4M?3M#2E#^O20$@V_6G@,"D^7LM^LOU'3V@PY*K,>9G)K5T+:3CH*=*MDO;7E5,FC+DHTM*L/;R?\@Q. MQP'N$FGX-2\H`T9FKI\\.4Q!NH1[-*9@2)\U'DO[T)@BI5&WY[;\;RQF6)F6 M+QM&[\VYB7N'=L%F=(JE1SZ](^,J?EI5DFD#GL2D!/ `0H("`+ [-[!G>H= MBU'!C+WQT:DEHC/^F,;>@SL`-YT]WQ.U&[%R_B%*UN\\NDQM)4&,#1.7$B(E MY'C3-*_S?,:L?#WN[-$TK"^F?!<96*7CP/ )$M"K: E1X6_ S(("^!_]@)N3 MTTE G_*=1DAU/GL87MP\4(?]H1 <%QW8%*,MGM_B+9?A>?AD, @#%,K;:BGN M@A&\;HT-,FS>U!59*[WJ29P^X0.M#WT=;_KQ["2>11"(M"-R5/=&^,'D_659 MN*D3#-S.EPO&E@[4G=Y.!D6BT$X<]GC:/'*RJL#= M-\P!U^EE.%AOXU754:7%+OO6EMC4(8PA>T(7A@:?';'!8-;'4P$Q2>@;+5J_ M%HJM2DB'3"1:)Y';ZV_<\E">D^(6R! MH.1#;]KA_PR G2Q:H9,?8&J#;0XIB\ 27G[VEC_6VW )%=>>H=E+R PH5\;Z M.MLMK#TG41BN<>V@L\?5`>7#&V(,`F1B]S;+*-);P A0>>^,FO"6*T(J7M+B M?7)T4@C-3H"@+R,#II!]];EZ,2HIHV*I7:.>CR@GH@ZS/Q3U2<*,?J"VVAA= M@;@ROP[&XH/<34BQUR,[#F6#S2NNOOD&>?7R2CT?ARK "7WRUW3_Q5]#A+>D ME\JLMO16&G8?$8B;4/X&6R$=3Q?3/,EK;5LNQH@`]SR(I*>6:3&\$5;"WEU3 MXOCWUYX2KO"TDE*-4>#?M_ASS!V9)R?R3<'&2?F=#>^5%5(695Y7+@1_OIYN M1]K.= Z1;DS*MQW1IB7<.FP$Z(PA12OHY1(;[D3$4?;Y3JH?LU28AVA3LDS]1I;M91 M@*K)BDRFC:_7#,]3@@AH,GVE6D&)<0#;[<-&3]!BYP=Z[0UZ#65N9'-T2J$%E!7WN`?Z&2/Z[VT;!HV45DU++*WUW&9<-)2*_T69VI2 ` M86ZI>*O?[6>A-%Y<%6LDY5OY*%3=*8L9^4M:509B9K8X$-B^PAO\&-@S"IUZ M\'?#.NY1NI>/$0QP5+)U3T5D[6<7@C=E6L[_,;>KM3'X)^]I(B(O!9L;`<#W MBA/)O]#Z&"5M6L(4>'E*^<,@:] 8-G4-1?"2B'(US'+JM8CUB%W3Z]A,EC4V M/P$A4[+"?+_G2-2-_XCNQY,+H6OK%O0+;-&HTZG8DD9KU02%/ MB7:V]0US4(1\E N&Q#8`5G/VB[.PL8KYRB!EZ#]D_.P91*D)TH24*DC=9F:;MAX[,;-SE4@\:*5- MXQYH3ID?7U1?8N!8A:?A"L/R),:5#U]97B:3 MRSOYT"*W/!+.88;PB]%T+[D<>3,*TQ4^OY-G#[HHA:9F4/L$F]R[0(7)DB)>$'"&C?C(!AZGM,9C('A3,/2.%T,EG[6^ MQUMMN+N45!N,Y(=\)?VVR3US@7<@-! M2^)=`F)]+13:\^<8#%4MD>U[)KB.-]B@RXRE)6VPB*I>]9:?(TTXNU,(5_;K M0H22Z@ZUG5'?NTV:#T[]=*=$`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`*V!PX&CE-R;'?4-%G' MDH)[;$Z0YV2A75//"ZEY\V%\EPQ& M6ZER:'*!7^5=]N+K8;HO/ 8UMWEQ#:JV#9YTCM-:%DEJYSVU8$3H7-CM(3CT M:Z^HLF2-=+>E<[R]LK8$L2'I$ T:J5B?J=IEC WJM& JRT.XT!K2=^(2'.NI M\IM"+OV7FQ>J0^$Y7Q'[@0)4)CRC*B=D6R!Q];:9R52B5]B*HQ:/-$@=WG!1 M<19P5JFJ`@8`?,'C@..;W'CD,!-/35!TUMXG6PT.\),IJ+P1*8(TORX&99,L M+6G1+?@_+62J*47D8+*<,%6&'"3K"V3 `)I4K^?0U74#.D'BT63TT4K$G]J$ MWLZF!XA7FUM\^^T>F+$90PZ?ZROI>MU!J[A9@??,N-=?!AM$PN$Y0HFBO>@I M^A@_@X_6TI6C)U7QIC77_^V$T,,AE/J E./7,FB^5N?(YO"66_L"\])H.2I0 MO]:CF3TQ3.WL<;VEG,<(E A?M*)679@JB$:^TZZ?/"$3`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`#+Y7I_2 M]Q!Z1CQ 1>&HWZURL=%!O3VH&M^F1#1UM;QD23OMKN(6:->$[U/G;("(QCB5 M&=OV&83@CA@4;^"9?Q&'ERXUNV4!D$"!FU6=;T/0`P[4XL@>P;C?!1%*W+N@ M5N5Q!O[.5R-C:4_VYM=>IG**6">B('/)AX#AH31YX 3>2Q_TL&G;GM+@BSO1 M!/'5[6Q$65EA0-].B]-F.1Z@/_OBV.OQA#]")&TJMJE.ZY:A:KP=DXD M"PJQYM,QUY#5Z^5?7?5[#"K$#09Y!F:--6+FY63*.WZ3RV_/YM^A5.S>XR9< M.<:W!!,U="[()<&A,4O&Y#6)I;9MOHQM+DL6;#A?T1QY!.')`"6-E6L]'_F$ M6?[,CJO,*FTQM5NJZ'P7P4)'W<)?45-+KC&9C*/,("K:DW>I!S3O2#H5B09J5? MV/-'U>O-OM/\\6WM,\'34%SP@YDSP C.T0N@QIM"=B7Q"L6(DWWKR!*WO,64 M^K]J$7.6TMM-!;^"^9%\=_1704N)KOF<7!VWY-6*?AD"H85O.[\-K6K7(4=M M2=ETAJ_U@GFZ#GVVB8/!Q'2D;V*,B\@"H/ -/X%G.L\@-7-".INZ`][5B+2J MISP))P7KR\N%WDV09!L%)52 8F%#<*0ZYVM^#0[D V+>M.04PUE;F1S M=')E86T-96YD;V)J#34W(# @;V)J#3P\("]&:6QT97(@+T9L871E1&5C;V1E M("],96YG=&@@-S R.2 O4W5B='EP92 O5'EP93%#(#X^( US=')E86T-"JVN MV_C'69>C-TQM9%*GA/W*"\!+W3^_.;_RK%5-5"^VA,2+_*H=;3]4=97V)S.( M[U'Y`S;TZ9%$YS%B"HV:`PC-KDWF@+ %OX$5:TV:)1B/XQ'F,-0,)':JIT.C M70. >,ZU&ZBH8S=RP?HU1#T;8J7'M?BS`AL0F?4ZF]ZQA6>>U)S%!1=,F7'M MT1U?:_3]&R^]_A\=E-",$N]@88'0R52RKL=$UC72$O.DBX=.2'L&?5"B[+UG M![6E]@=1C%(Q:IP.4Q*4ONRX9UZ[W.=(B/^N">\*Y1_&HCZB5WNH#^\AR]Y8 M"()?UKYMAPTIJE:2.V&C=O0D$;H:CD.NN4)/ZV!7CGZZQ>$;<)O3(H"V.IRM MX)_HYY#H1"A2]@<*:IMYD9_Q7'$_X=D&OU-H-B1D)IM#$"]:5I>C^)*YD ?) MZ!0CZ0%EC#W/JUV1O9N53B*&-&)2)DAVV-F@0G78#J37.R;7HMRQWX]JX231*7-\ MI Y*\E0?<*,M9NL-P> Y@R9@#I"%1'ZSL"S:>!ZW(,6(:H,;ZT4\V"+-XGGL MYLB2U;UBCC>.^YQ9FJA:U(P1>M#]WI M-D[THI?%GWSC+&;^6(1!R#CG!_S:6 LGO.L1P *16W\Z[OT!#<-8I]Z%=O=C M3% G*VY+"ZS4,";&6T_'G0#WZ&68ZT)"*,*W42[-S3W,3/`BH%H)T^]"8SOWKQL'5$SL+A2M3Z+V1W#8K$AYL,; ;$0:HB M."L@ZDR#F(44\ID MZA;-B;#\$R\+F7Q84>3&/SAQL;_?:7@_:XGS+1&'W MH-_01I4(6^I \$<,-%2XYF5N;NI,+@]I10GJ^-#R"6[9,W>Z1W+<#K9UM%6S M<8\&A"3A2=,"/! J?Z*#\J/G9':/N7XA^BT1Z;_8":N[A@.5<+ZO_>6RVU(W MZ1(H6<<)_^$;;L<(Y2S/`(IMH*WP?P:IPN-,;Y52U:NW&XYN>V=.39@1WY71U.!H0Y(D]MX1HL]:E02N7>Y$DS'VGHVK6DY#6FAW M M3H52Z1?38.W$]E+MA(D/-&!O#5A3Z14U=5V)E?B /)7IET"R3'B-\(2=BH;> M=0@'N8]I8>M/&.*0FLE;77:3OD:.XL^E@HH)T3)O/.$IJ-F5E+,OOBSUY_02 MJWK"/Y)AWNW3LEIESM$MV_QP6+[AC'6=V2*C$E!U_UHA.Y!W$=HB:5P^&H!5 M%K#:3WXZ(P,J.QZQH6Q7S*0=D?29V^Z@56W>:]>O\]_/:DK(__49)N9<;KC8 M%R]2;'13O)92B\OPO"-;@@GD#*,7_%T5LK.4&TG\6$<3?1OES7$ M36U?Z#B7M=6PLS;9WZ=:_[H=.97K>N_#4)V:D]3X%2W_RJ1VU-)76JZ7?Y;L M0"SB'P$H;'%@J>[5> 0N+(J8PGAN.#*[P>%0D_1V`B?57^?-0'L\*"MAN[#4 M.S<3;W3VM8+)E=&%`U.;R-NR(,#-C6,%LVNB# "V_[/G_U_-UUEBMN5[PSFI MNHM0PL-[G-0-EG^PE@:3@2>$!V+AX"4S8VH*_;,!K1J1TNDCYV435;SEPLP" MQ343`?PQ^H7-CY,GP/%.S4L1N]1^>8B_\(A6\[:2L'.$NMB008O>ST#2K2K4 M_D8-COGF_Z&C^;P])[I$Q:]:<&L!5XD<)*1V*>O Z^9G,9NMZYV!LCLT7MA' M7(J'+6_7/UWTR@[.9:(I6[5NG; MKKXRD]<==+-V!6U..M"98=C#T0)S73TBUEMR(-]EK317IZ]Z9.P;%5Z0*$94DN^7J9^SFY&4%"T^?7>!35O`4ZS1F(] MMEF.NDP2Z0Q;OBKCK$,?O>Y>95->$!O]-&C4AQBM)JCV?);NCEY,_M$(Z9N: MC_]JRUG]$HI<>&CM)5G9^!^CNBTDG?([V[FLWHV*=JXEQRR[Y_Z@W9/7V(EO M: XW#51?!ZQ&33)K^D%"1H*NDI5Q;NQ[FS"D&%]X:D!P>2]RWB*D8J"8Y:LM*+"4K'='2$%9@WNZ.+X 4*?:G^FMK2BW= M%0[)"2(Q8(03%S&[8P^E>06O>SPFTG*4S+\$WPE2OB57Z]4S)>CZ)WP'KSP= MQ1XP19VB^(\P=PU2"4#RN==%F>H%I-9FG5$>A-=1S);21C8:U\F:72*Y+V[< MX)-[SEZ8S3?9NZKY2!E?6OF7,?]LJ4#.I_MPJ:0PIL"^*/MKU<:20_ I$'S7 MG@)CPWOAU:O\")W 4LT,L-L2N#^ZXLN!@0`2V,'(&]&9_F(G+9QQ!AVD)-V? MWB+DPRY=1(S^`VX-0_J&LLP\6L_^[J($*.PF]8,+^DD'^43(9*;,KY9H9UD" MJEGJS%RGA\>.=(S3CNP3YCZJ-E_\>)>XZ(3H;R \0^$O5582TY3%.L^;!#G3 MW!(CDQJ[_#$8BYUT:XT;8_OW\)03[DN&Y]LB(Z D8"L]9#"[GB"%/,T/J?$W MWBP$&HIY*:QJ\BE2\Q_60IXM?&Z*9HA8G*OQ_X9%8:S;\^.=M+E[+4YXCAB% MZ%FRM$+UPF&ZV[F2C/K\]1L+<&W0KJZ(](V_!T*T?Q3QHIIJCCR1-WGFDG@L MHB=Q^=5G(1(%#V:M^/'U'C]6$^V/'G#QT6^)C:?3&>'-(;<7*B3:' $<"D&D M^ !#?[:BGM.6SR3:.JY$0%HB[B<#C=T]9F_+LJ\X>1L2KY:@9R&@`)H2R1$/ M?R;@10/H/NH``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`!7KGS(J5*" EWCC#C\W M*+IRPHF04HJZ&#P(J:,0*T=&2 Q<2F%MIN@3!.V_3\!!F%-IG/7'?8SF%0G"V'$UZ4'LI.B['"HKQ$_,@U$[>HJXG8W%:H,Y@ZP!E$9R9'XY9:/5PND&J:?YL_[P MF,ER3,\]46QW\*V')>OP.(4"9!TS8R22:*6_Y"_EDP]<>TY"IUVQKUVF-HBL MGL.X7*P/:R!=+U*LQ9BI1)G3SK_*G=57RC4!'5/DM@Q(LQRPK/?\`T%@L.3= ML8C6VNX/K;'9>)D;YC@9=SX*&364P;1B@A-\E94AVSEQ&RIOF$">S:(2_;)E MTT_8Y>";<6NC,\7'N@O)VG^VOT*Z@)^DZZ]2%Q4@16)8__^70IT7&M*,CQ*G MV=_L442X=[N:"VC0>\AJ$YX_E19]'07$<&*<7A!5#4\R6@2JAW1V34=EW2P& M%S->TV&S\7NI%=",ZV9G6=?,6]3.R0!?P^['M ,`4T;10KQ=I[* M]8U?H[>K=2%JZ4L^&2Z>A :(+E5%4U&JLU]85=(6IT=+<*-@3BFELU>8G M+P!!R?13B+$S"586U(O1]T<%!%G>YD%F10CU! 2B'1C7-EZ_R#*/2-BXV_A$ ME8NH955/KAAO-MSA#''EB61F!ALD2X9/XT;SO"Y-U062$5[F`LZ O87TS^]Z M,/7N5?=TEJ)E7! Q3#ENE-]%T :\-.-6+9Y1&^:J4K!0QAT6KP&202$PR>LK M5IYC,-[P/(MDG#G(5GB]3.E)#3TBSATT`O@ONK '*J8F)%0Z5W/?U@G?FT!O M!$RL^1 9G_AS_-IBLJ41J_7'/P\[?/+UK&V4]O"%E2",1/T$\&H:\/'KJ. O M\1.,EC@7WO()!7#8/&*X'@M6"ZG9S!]K,SMST90DI>@IV@#::B9*-18%$G # M0](9;X%+=3.4!M8%&@:+%067C;P$W(=2R/.:-[B MD^ `"=DFF]?5>$C^5BT=TS>[)`4N/X&%$8]+8&236=?4^9LY6$(EMQOA6 M\?J(\*(M/J9E*VRDL?!/?G-CA5>B"ZZ/?E]-"+BB= )S8'N\41D-N&/.A@') M)"KBT?N_WMX-DD[>/M=/#$,4QJI\ M%)/SFI<.* #WNPRMY/,L>M $V]V"F='W.;Z*'"IDONQ!)&E/R7E=0?D(+#O@ M-)K9\>J@\-&,&:BED(KW-\-\-?>SHL`KI8$1C)"1\.NM+= M`13O4A&Y5%4%IO2E>P.\OVUCCH>^]'Y9%-@L)&DV;_)7\:ESJ I\Z5EV3=][ M)=CQS->Q9T UBH7_\G,.ECNXHY2PB]=Q$FH:0DJU*S:[K/'%,/(V6X>^2P>E M0V\AO\#-:IJWK+B8> MV$OV$H7G:5GX^8EJX_@;I7[W-L;WDE1Y)\MS4H6,RA_ROHC;H+.KBS[D^%,` M\81Y=,KT`$\1RD/8:;",ICIB2UIO->T-]HD*+<;(<+I2_!:\,?CFVR_CF9]V MZJ$6M)6VNFE ]D$D+QH-S>8LW/RKR4D5!E"=5<9O4O OZJ>4Z70:A:QS+0QK M-\7U'V6*S_ ^=7=R<"N&,J$P]@,EK9*:]N8)/]*D)T[J%,;2SH^\1YF%33." M"7*D5$'78(>E:0PV0P!#;^V?!297 #EOWD7 MUJ7[J0NR:YLXL;>7>;\AMV:1:M>B2MC<^.+Q% ^$*P9U<_QOFCO05NC,A !: M"FX1G ]^_$/&??%#F\^TW.0WG3N7$%"V'4LM`(*BNGGP/E4Z`*C-MP3#* 4" M8G[)H7E*S KLB3%8^T40AYI!2ZL!L(\5-X<<(;6HLR??V_80,4]/;) M%X"MT/Z"0)FOBU98XGP%9=X#E*HT$C50HP3?XWCL"IYKP[]K7">AJS:J0X%A MTYQ;!*D9UVC73_OO.:'S.:.?4-;FF :79DVUP@(Z=,@P;3LYZ5//K.YZWGHT M))"*TKTXSC'(8(6(=IFX1>Q$E;3T1=*W.3#>;HZ9"M^_NM'OG'9)B/2_00\T M6_FM[D;'_4.XW;MC+9T]8OB4YX=>JJ(90M;9B(Z05XE;>,WPB;CS42LC2_6& M"B+G+-T$`2Q1D&Q\0N#4DB9B.]$<(+=8."#3!E/M!U8$(723B+)#-_8$FU2" MX!./=\_:PZT!.5NQ.B_+-;E)X;RP_:*%$O;A>/65R&>_#5>1V>^Z8_2G:7F M/7!ISIC=+*3 7>%81%^U9V_DK'?DJ+,SX@7QK6N[5;<%MW"@(&Y?#D=6Q9!K M8H\$;I8V%\:*!H;Z[$GLPGL\1"8W2+>EGLW^4@(8/0G-`P%,>1Q8#F@A# MU M*-!X_%PNR3=R=I_QYBI^918U8E8K& =&6V_D.^WG`P'A/[Y?+>:'W=H".#]@H_+YBZ*VIU3O+X;]0#S%,F1^3\\ MRJ?6-+ 49/.F53J\IH:6RYG4^EHZ52M];-85D6;VG.MB[H(,\ ;%5G?4Y\D/ MZEV:!Q,+\@E26Z_#;GE\!R .2>3KX>,B_4DV#4FV&W2@8BB67\ZLZT]\2"7, M6-BE03.?9^14:5PE$_1[!'(-JTZ,2HKY`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`MP58P2/3FK168?KZ_GWU:K[T1_JTR1 Z+A!U$EV:,U\F.C ALN= ME0*9]5'LE_P!Y99;?B( M1.VS=7TPZV)3XCH3NS(.A$)19PPW!OE58:X-`U!YFL_=TOQ.)?#;7-9-3EZ* MK:CX3F3/]6KX=2R]=-8Q$64UT ?\=E0<;IXM@1507Z;P*@$%UD/"=2\T7@S( M&.+%6 'G'+X(+DK=V5PFR(;/3NS^6, 5"=I;_28ZS@[7;!9>/Q12JUN=RIK> M5OS#T8=1ILOBQ.6Q%>=BCNE8 JH-T?@VN8AYVCR/3GDBP]>!JO.*[&Q>Y8.(,3J?F159]@9^ M-\OMZ-=\]PLN8",,K7&92C*ERYX3^N+^^2?9K>RP2QM_0<,>/\Q))(-R7OH[ M*E=A%1JVY9W'NTZTBU,"HDL3N_9@]8HES%?A^#;W>!B(R]!Q6&71 MX(#>MIG^L3I='3(C(>XZW0X"J5XOZ E8SSEA?)W()MM7)RS2LA(><+Q)\>9N M,GS.Q$JURE,$G%&N! T,LSA^6;E%1GFW]=A1Q#]>?&LFH%M^,>F3OB6/(5/; MWB.^ZP`CV:!42%)<1! 2ZNCGR=U)%H<7W=M2?3R^?)&G(W:D'-!5O*!T]:-)[@)4C;G9W24N0Q*E)'FK1UY/=><109K8$S3!_!1 M[C69>K==UESBPG?UD[^ M=YV>6L!9\1U$-@4+D#O1%%[[*",;0-N: M&>41A.P-EV:>PY+?\1Z.0NJ.76:7JK95H3P&KV^^I&=4YZV=S6RGQYNV?Q+# M5 PV*F2#V@UE;F1S=')E86T-96YD;V)J#34X(# @;V)J#3P\("]&:6QT97(@ M+T9L871E1&5C;V1E("],96YG=&@@.#,Q("]3=6)T>7!E("]4>7!E,4,@/CX@ M#7-T546-@TGYCU:3F#V!S M[ ]Y``WX&>F6F\_.GPR%U,(I0+*2\_'HXI.@3*%!\*SS^4A.!09["I+S&=\"26M-IW=' MUDLC)E&]=@&_*_H?[(P\(;F<(?4,G:YTI!0`URT9)S4H"=^>M'[)HH&(\JG MTT5*]*/R2CWWG(H9W/W!HPL31]S0S 0HBR-A(7(23JB+<8*)\V1X'OZ5`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`]P!A72+<;':E$2&1EVPBUW0@=S&D M+FD>KX[9SLN...)NVR])?J(?G>:9#)-G>;B5A(]7J,R#`A/P:&1TMFG2+X&U M16A=>D@$HW%+Y76Q9EW)+I6&;QUXV(# +C^&J $]L[=V$%[6H,- A\# W(B1,9[FZ8!C@:T? MJ!/CZ7N@)/V/)B[4'A-92(C*\H.7A#BMXD;0,4/CE6QM!H^U)I\A6;>I0S3A6"F],"6(0&/ M3'_?[GE=#26)6-/:^A*:'!-V6>@L\4W\ILV;1XOT;Z#X(G@DEV2A36L?$<@+ M9/&XB:T,K4GA?&S&$%/]0J-L\'!+'HX&W_RA9M8A\J:+2U\Y*YK7!1G"R2R* M,PH1J,MU1.8(FG$Y4 LL5''!\B*SV4 6E,=1=B[6*.=$%L\L.TR# MJ@_6BXI"\/9B906/@4Y(^"I0B'YZN$QP#FU=Q7L&0Y($EDEJ')/ZOVY-:[F! M`_[V79Q%CXCP.0!CER[#KA=?VR$@LR;?QYQ^TXBYK*E]*KQD5GY#Q#'2SP,F M.JM.0W_24BG3);U7L?+D_QTP 0WT(]IBD�MJOB?1M0GZ(^M%# M/+%?4/;;\"4E+9T\K?.K"]>F;VTSH3 M4P.T76M_K8Z&R.A\(JOSY(9!%[C4Q.Z"C\UKHT"F1?CSKI[3@?F]R#-@$3AO MA64.!*.$+WY[0/*P@\ML?J)PO_;?G-EJ?5**)RMEU6HV`-E"?$/7YYD.CGUI M=<:NC[27;'[[=>WAY"(_2S!PV5B2)",+@?74'/.@ZH0$L8/:E. D\>=I]7R! M5H;#(-X;5TD`-G\-SHYRL+CRIB%CFUGN2LA]+RS/!!P+$%G'8EL53HPKXD"V M/K8`:7Z[SQTM;4LYC=G6V+@!5S7!G;.WX0"@E4B6+3^32][N5^)__ZN_9DS+ ME:Y*-I&)Z+2=_4Q@IERA_"9,`QZE,35%RJA"^&+$)ME.LIVPFO6MX :K-8HQ M-\LR_E!-6RIA^<6'K7 XX-!+*':<07W'MFQOSPVR59->: M$&24QV"P1"71Z!YJV*]%_GHT?=>X%2*,-U6JVT&\(AM URPH28:AM-CMO\,2 MK/"[E]UK+]CI,Z,1?;>H;EOX"G.)'ED`Y@FT`*%L7,@KEJ!LLV6#182QI+C> MG8J9)0],`.XEI5\,3MG$V1Y^'HH4QXUFV0J[)5@.8TI_RHY@7Q3$^V/:3K62 MT7!\ATMJ" ]J2GLY,2 30H?L2-.?0A.DW[T=-0UE;F1S=')E86T-96YD;V)J M#38P(# @;V)J#3P\( TO5'EP92 O17AT1U-T871E( TO4T$@9F%L+Y^#FOT:.7[#6A^A/A5KEH]NJ:$T?>X"^#9&?``'J ME%<]``2K4[3UT_9LWK#R0@C8_?\--O! $FOPAL6YJ6N,):#_YZK<242O/5CE MU5B_YDY?3PBKSPD#20>M=.W=5;]=`:#\G_HG2=D0'R6CJT&C&.<33?LC1& 9 MD:@B@#0'^*KJ"DY1H#&2Q2F06>=HOC?7HHGG\K=%_DI/PI3 MP7E(C& YTZ1WE" ,:\PR42\?[2=W.9I,( 5(&=?K*/.6UW;RT#(=:2B=Q=.) M7TR$_<3JX(7LX6AQOGUIAN4OR+T&. %B!^0XPK8N;0AP%. "W?1 G-=%6[P6 ME^EP\JC/3MF\4N3]"NJ19R Y:GD54S!J$54!UJO=!PL[Y!S/F&%P\VYE.%2& M(OLJ+6(`)4?I!"RV% 6@=YZ:K$5+]A! V\8'JSB2@-Q0\TW\C.8EH^K\,B/6 M=SXQ2>W M%_!9?T!?3.+FMT('2JY3$9LLR3,@<.X+7*;2!ZR@D\%I^U_Z.>/C_WCY6ZL@ MTB@TD(PZ%10Z=8CLQN]!+;$3U>S-;%B5K,J$QNMPPJR.R\H![$AL0Y,P9+[ M]IS9[S2]9>)^QISWX3M!YXN$U5L4.A0. M^=_YY=#>+!E2D4.O:*VSYQUA$;/TYZS)::Y:^G)LC]@GHE-P(14# M%:+YW(DQ3WJ1MM11G*B5=.%=C08-KVG$HT7<_-K>*2USA1<9:RGHH8,WJ'3 MU?"[.I=#&^5I3$8/+I!GZKD\'ENHNHKZ7?/[YK)L9N)7RCZ:( )=A@`T%D/K MQ@#-Y^C&071+AL%%@J]AH*1D)=V/U#"]W.IN;CH8AMEN<,.;#BVF]:^]H5Y3 M,V 94"H_Y!8+^]^Z\'KF]4TVP "(#[>^)%WM>NS9.!,N932ICY.I@@C]\9GZ M_$<'_P\ZH]/KS0'[5*:G5EV$3+#2T+D1,4$Z=1Q;U(VX=W\/.$.OZ!OR2@MP MC(M!.J2=(!FUS;-:WM+N#7XY7/9V(5WZG MD5;;(IDH?' QNJ770UP&6MY*G83;DI_X[:$)-<3EBW6).N^QZ8T/OONZ2$$% M!XZ,FI HX.,2= MD1;)+^5C:_FFH9@M)G- ]%4L`T/9#OZ*0#T\?0P(ZVD5ZO]C_I3+M0PBR>.G M4YV_4_#8=\LW^OHZL!I+\:+";W &T, <%(N2(4JL+(.[-XN:Q4OC-4***>2W MAE?=B"1WV[]%K,*P"J$-31HE=^+_7S&G6-+&.,1&I$=5OLLIX8]51"/'MWS%.=/3E5.K2\%R_@RY.ZQR(_2A9(H&1F=E<3A M+L[GBTYN_<;Z$D)"@(-\7^N!2+8V2F `#'ZETLU2*Q>\4+>.D@__X..]G89D M/W<`G5]KVZ_C0V.$^X9SU#W:9UIAWF+?\Q?YB&,><>/88:QOA"P!QE1!KV=?=EC2IU>X!;$(])JP4\ZC<8T:Q&UN&T-0+V:(VC\>H]4 MH:Q'X<2.'D@7:E:7RBF:?PLE'O6OY?X3QJ(5_HI6 M%#F0NX$%NW7O[L$$^Q%0DN2WRC8DBXXJK4V>:D!LF8\(3]90L@=]QQ4]2CG? MIHDKS,@Q!<79-^]$[T<>(O5-3S_05F5@FTI"SXI=@BU//UGJ>E69I[RF%+6P M#K*#M/1+2<2N_ MA;66M%AT'&4&ZFN"/O+N6OHA8,('>KQ6WUGC3 M/IR8=.VIXJ>PRT?E+!"=?/&E;)L[U<,47F*#5YY] (&%61_QQO7 M&O%;/TVC<1Z2H2;R(YY8>0PW347N,'(M/NSWIN/-D+-:A9)H.L(*,_,/9764 MH!5]Z(VU:1<]'8JRQ2>-!9PZK'_=%LI?H3;6>JID`^.P6H=5K!P['6B^Z;(9 M]FD,KKB:%P*_M70XJ*@>P4!>6#W\-W4('WDN#QC? M(@"^*Z4.GD?]I*]Y-"=1%Y@J=$CA:/4\=BN]*'!8.=.9;I8Q7JXQ7.]`;6SO#=F&^VREG6:'K?J42O >UI8R]0G M&_#F)=Z1+W(QG>];_ UE;F1S=')E86T-96YD;V)J#38S(# @;V)J#3P\("]& M:6QT97(@+T9L871E1&5C;V1E("],96YG=&@@,C0T-" O4W5B='EP92 O5'EP M93%#(#X^( US=')E86T-"A;]8 @`X%YG<)H%9$]U++NSC6.[GE=9L,D*&1;F MY"9MZ\4+M.^VO(Y?WRV137YX%G>>&8LIC.X@R#N%;YI+66XJ1W1X/8?V>/A? MO,86#(OW8,F7,5?"'OE.*GL<3##P=R$+%O40!N)%N!OG8D9B#)SWB(N=>/KO MGG1#_VGOGW,\CT\?\O^DW&+)Z?6>-W1B&UT41I=*NU!2/R%7@+,]]<)N"6/* M6>,"'VB/"'XAE!BKXAA^R*;DJGB>%^L(IU/M34[CV*WJGKTS#_L!>#=+@Z0[ MK>H(!MA*XHYAK6(M4E#>?;!9V-9QJY6_X,-OR)M^8/&/O*U0@1FS*5UZFQ=N M^--K_RH^\D/A75K/"+&CS6#U]>+-G84STC)Y]!R7_GFU4X'Y:)BB\7D(7G47 M+>V74PFW3%=>AU>K[-WJ$L^"RSU5QDL($0>1V*M01N^T@K)H`TKD4 M`4L4DH2/=,MJ7_TY1DF)=;&;X/6E@F@B&F^R7\C?K+;S4O[8_BT*"-W I?<" M=S1*?-4.[AY)3@R6?GA3Q"_C72%B)BXN\R/6%0?B$.,5C[]&>4X(Q2(H=%G\ M,QWO''^X//'D/I>L-$(N:_52\(SGUC6H[]XW00*?ET.\+N92VHZ9AGV\`Y+7 M#)D[%RW+OIE!&@GS<2<;@$?XEGSK3[Y@%HKS\BEW8JQU0B@L#]\[R!G.60"5 M1P=R,N_M'O)4I/;R'N/%( V V=^.#Q4I=K%F"17P#DL!]=6U>/#3RJ M%%DB;(CW0/#-EZ1%/P'CASZ#A[@C/6:7<]%N';%Y`1##7"1<:C3_$WX#ND&4 M14C(RS3C5X,NS16W`:63.:>+3Z02[_,8-6U8_=!UW&4/T;O0UJ6,/][:<7 Q M3/H,?:D2D[/T2<,N2 F!_M-QA">*GG!@>X4+U1N?](37^["8DS=F1V'02;10O'8>E%2R5_;"B3UI*M4J"Z*Q/L=^+BR[4FJ=Z8ZRV%FD-,U< MF<%D.4RIT!N]K*+T/<$_2=L[`ONR0,IVDT)E+'%IEKGJ0 0XU]IK3;'IV+$$ M_58<+)/[2@*9![TVMWPQA4;@\WNJ!>'(AR)M^"M.GN6L0NZ_XGR MY!OC$!6"VW]B]S%0L&E#/+?Z=B#Q$VOL$P#[(U_7Q$L$S%XE5$5Y^F6?LE5J M^=P].*HI%"/59S?2\V2O-RZ,=?VY;_-+W0]`UW[@?C5^.I MSY?F3A(W9YLYRZ]<]=,<$]X$**JK?4V7.;V5K01@K).*G>OY586?XZ(.;WO[ MJ"I6@)8H]@1M!F]ZEW- ^+B2@#,T#V'Y:5*/_BDLIM_@.=];K_W!_!+):O#3 MHOA 3[-!.<367"98R>)8$J&949N3`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`\KGC'YL63XL M(CM[)NX)GXW#K)B5`234.0MUDA.A'\V*L7[R)12:*%#SJF,:6SO -R7ZP* ` M_$5#FEAV[H6A\5-WK#+M0A@8H?MG>@SIB +%+@K(STIBD*(-[*GVP"D^+@;/MR!(#1B3]L5MB(X2N4)PD:171,HW']SL>.KHWVTOG1:I;/P/`6T54GDF/1C?*[!8A$5. M2HTPL",4HI[+Y&',=V'RGA3<=)B1*2,;@VQKGLO,$E1H< KF#@HMD@JG%CP3 MGV;@]:VVF?2_@\#*P*$0;BG.!7!\-"07!X3TT8)*8?\J63XH*J"O@4FXF"[C M?(U%9UXD^J$@+)%/-L[G[B+GHQMHTRP<*\V*-OSBM%TE3D5J8GRMIW$JP28; M,_&'X-F0-]FH!=8G9AX$5HG30.TF]=4MRY]:%XD(#65N9'-T7!E("]4>7!E,4,@/CX@#7-TWSP:(#B8<+CPN77 MO29&(2J-DKY15W;VUK9>GY*;YFDZT_L&55*YMJ 9QG-/M&E%""WI;ZG6%/2S MFK[$<=@`S%*\4LSWHT%*[E#\=V;%%DD/]WT6F779Q^N<+02S;% ^G$MT7R@_ M7'\O+*GLGC]!V^CNG\$>+WWZDN[]=19%0 _F)F+BH/W'J7,QUXHGKARH/L2S M^-H)LZ4G39//OJM#G,7%J>JH8?MA#,4`UIV3#_A#"3Y!WOF)VXK6BFN0U&U^ M4+),=M%$+/;D)+'90MO-IV(&RKTWVQ#-^_A9F#QGR D/AL9T7G54\]4,C\AE M_O)-!/^@M01F_UIJMKP/NDG-*A??:@EC,8I(^4;Z`>-@AG=$^NX%)WI>[E/* ME_8DI[S0YP3SH$YEI9?)-OF/E(XH0_6\(05]+B[+GZL:)DI9?"YZS?WS?7%P M\ALIRBW%ZFCO1'A@N1X60W,PNT8+3H/MT"[EA,^MYDBBQ#,6HPZE0A($[BF M]'P%WO;]-LE6>6(@9NSBK:6\TV@_U;9^NINSS/)?B&VS947P]ZZ"^:5L7KT, M2;.(D,(R)(!]`YE7`G2 4*4G/X^/VTR_W%T&-X.18G\>Q-YL>;)X]*VB969! MR'2_)O24RD47, M\SL:B$:97E:",,1[U&E@#6[Z!\""VY[/& 6@:Q-*_)!\IMM@S9,X7Q5()N9* MS/._6/QNT,0;I8"/*_;-21_95V,`?,SA#4];* >-@#FD2V[JMIA77*0[&"6S MV>8H+3@B%KJFC+LWK '.;L($%[+&WX.QDTG3_77YY^!,_RR&62)Z M*9D\K!P\I26];F%K>$4@3[_Y:A;B[6.<]0P'P3-4=:'-G9HK*\_;Q0I6]2?E MY^: ?AAZPG:T_PT=(K!W2"7 /C%2K #=$/G&N$W>B6%TMCYFJ;&=\X[MIM-B MK!40_NY:2FUK@4)5HZ'_S2?VE^B6KH0#/)QQC$+Q7.\AKC;EB]%\,:M'M-<6 MQW(#2F[!C7PJW&N:G^6E?RGH`JA13N[3$!G$OQ.!S6RQ=27A@5.X+Y@J?+3A M'U#M5F4>2=IYK\0`K;4N57J8NSKVT$>Z^=6RPJ2S3BU\PL)@V*G^).9BZ*H= ML.?UDXI]K$H4V!BE&!8,BO"*I";1P1.VN(:JRY8CS.JY@EG'W"X4C.^F_G[P M-^;9$L%R(#R/]$:%(>4A&_0!+0S4;>]]*>(T3Q9?QFO;`T>P43PX.[.^K8#[ M=+N__Q2!Y^.O0KR?NZB_\NX5S&N('1./D],S]NJ-QY*N`8'8\0,U+]9B=HK\ M[G@5JZ+Z/#Q@=T&DTU/4OZCP#^W\<+3?_W@3F F9.^$49PL*3@1/B ME5XAF 3B;2OUW.;0\>24M: P%0*6'?&[ID<+^!@%D!O/2M+P(UUDZGJ%_65] M6K^-0-E0:_3RHH4*:Y[WZ;TU%D$ :CC@2KVSM0VEX+J3N=T*TH8O,MS*O M=:;/2Z#SLZ^%GG.D9]??I,P;&A[;E0'3M>!UZSL8=MEL9G?.FDV*8[6A"9[? M8?KHAMXG#H(9!X.(VIA,^U-K\YO%/*?J9>:PS2=L:0PPU)IU*9^$6Z6T)/.+ M]04)<0%-_J;"M9_6$!7/W-!$9 D)<^#'&TM@DLD?9"[3=R#V<;M9I4^WMT>( M+-?Z+WO!UH->\LU(D.1TLEI9^7JXUYUK@%>,"X>B:AF)GPQ3*;N.J-7J!DF:(7:&F@5H8R629'U/UF MQL 9$W66!GFF[!4`YD627F7?<+^VVS,.?/A"0?;^K>&)"&ZG#^/XV2<=5^J; M-.?\H)UF%"?DWM?VT4T'A\[ MVF]0ZM-SD,V>Z #\#KT44!$'?M;"@M<,PFB"%6\%^F8YM)U?''4C,C[V:ER/ M-,+Z_N/M#1"NFV/*`!?W]&U?<'BZC9- \KUQ9@W0T[DXZ]_WK2VR,1O10M4W M)8!5QV^."V9+,NWX="EW527 ;0VBV0NY<3C_`ZR&+$)V7N[$`!O HR#W?$+0Z1ND[6]FDTG M13+8$-:UX531U$NWO^;_+)4"?"YTOPV!$HN DB/Z9^VW$]-4<=&(/ZFA@V&L M4*O&;N'C+4L#[1L]N 1>)A8(>1ZO81A%< 653_FMY&0CG8 827+*D_?K+>\1 MV8XD`$H]TZIR<@EY1.)?G#Y>:IP\MCSBR=(DP*!WM$=M,#M8-H0>`N$ZGX'( M3,AG\3J9_*V?@->1'6MC5N"+J%OCBS?EAZ'/I0,X=".J*L$SW(A-"B(>B/]])FG)'NF.(V@:'J854_$8O\N MA##*"RUR).-0=G,*F4MNW7>8(?S>_:$:`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`VU53RB\+$/0B5C^6GOU'OB\`I*G_B:N.=X+XK,YTGS93,=BG^^,KF1/=G"S)P?MR .L(QR118O2^FI+=\: MF=M=:FA562_\2]/%;2EXWMA*Z6,Y!_FEZX HPW)^4A):W4J"[N-YN"T):KA% M9BE&1;1-P//>(&39HLJO([;*MN[S&FFJK5U'\3A'Y?X.#C:M@-*;4Q@HV#YUF$6T`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`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`/NHA\KZGW:!*S:/+N)(AG^7ZOD.X78+%/.1 M_W#<]7]!)GA#A[^N#Y^[Q(B[`:LLHVEQ9S RCQV^3N$&UFXZC)C)5PUM80)N%1T).Y ',:@19X,;,Q% MK,;B%3\I5%&88/0DFV!-NH EYA=7>@=5>6=FV T8X^0'EET[H%2#:H_;FL/\ M!$*CK]WG[3C\K-QR"V;*(_Y"[^51.T<(AOYG.D_"^>,OBDS58>/L6:RL'K_K M';Q)@A^/V^V8N2,%L"L=W^8H8K<^VAX0_W?R9CG5@DMI&$#O179ZXT/IW&J! MV;WZ].4JJ-EB.0O.::>*NO* >86)R%%B3HB,R&V!`UHB-/$@O3B+..YVLLS* MYY#9$XW2EF[)PPZ;R\;PXHUQ2X=X,"+5):<,V+&07F'=/D4)^Z.?A6&SN7$6 M"C+[TS5)!CM!K>JR^N>NE:*^J$D--[%!M*!/.*US/N$1\O.SRRG=5%K1Q+IV MJ9$%:3-4T2:./#O6%8 A!KZ*=";.*I%$7>,[+%SO`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`PDORX"<3[8`-Q-7VNGB M%B<2V92^=G=MX>B='8^Q3_Z/+GW8I;A* MLJ=M4Y12MY-@-[#TK+X!V>Z#&469X8J-X;G?@O$I76(D2!0A\6-05\:9JDUC MT9"K+;,(@FSHV=XASK].-_>25:WF)(27H9V!N[E_@9U0QK?!L>3NX/-567&$ M-0?VJ;\QBUJ<@VL.ESM[4BP>O MX#!2*[HB98QS# I=Z15LZ62\+@U8Y#<%['\+GF4,P@,="=;@7K3%8*MK`32^ M$_C.E&>#[S,YW@43K)(4)X"3)7_,(WNIH_FJ'QGBO[YWPY +Q-4$TVI2D M7;["U8@`"W/R*'VV>CC!A-F%&9.+GV[E*,>SU@`.SO='6\W+A8+C#(-A--SW ML>DR+^K<\T?^_O>_(ZD(]7>7'GF[(1(7NSH11#+0#*H>M;Y")7J3=UD5`M"[ M\7/V9 ^6%G(`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

Hi Friends,

We are assigned the task of developing a Flash driver & integrate the same with a File system. The system is an ARM7TDMI based board. We have to use VxWorks for this purpose. Now we are standing in the design phase.  The chip we are planning to use is MBM29-Fujitsu flash. The whole system needs to boot from the FLASH. So the Flash would hold the BOOT image as well as the data area( need to store HTML pages). The problem is that, we can't use the TrueFFS since it is an optional component. The data area of the flash has to store the native system configurations while power down. So the flash would operate in both read/write mode. However in Tornado help manual it is specified that the , Flash that stores the VxWorks image should only be a ROM & shouldn't be Written. If the flash(with image) would undergo write cycles, then the whole image needs to be transferred to RAM at the strt-up. Is my above understanding correct.

Next regarding the file system: Can I implement the dos file system on the Falsh which stores the image & the data memory. The dos FS would have a FAT table. Will the FAT be a part of Data area. My primary worry is  , Can I have a dos FS or RawFS over the flash?. I am not bothered about the Falsh wear cycles at this stage. In case , if the above said file systems can't be used with flash, what would be proper alternative. Above the file system layer there would be the FTP & HTTP servers running. However there would be frequent write cycles into the flash.

Next problem is, The user should be able to update the VxWorks boot image. Perhaps if the user dumps a corrupted image , then how to bring back the system to live condition. Should I need to go for a seperate boot loader which can point to both old & new image.  

My last query is , Can a NAND flash device be made bootable, where the flash contains both Image & data.

Thanx & Regards                                                                                                                                            Prabavathy Ganesh

 

 

 

Get Your Private, Free E-mail from Indiatimes at http://email.indiatimes.com
Buy Music, Video, CD-ROM, Audio-Books and Music Accessories from http://www.planetm.co.in --=_MAILER_ATTACH_BOUNDARY1_20022411744381931706506-- From vxwexplo-errs@csg.lbl.gov Mon Feb 4 06:16:52 2002 From: "Lakshmi Reddy P" Date: Mon Feb 4 06:16:54 PST 2002 Subject: Image downloading problem Hai friends, i am trying to download the vxworks image to target board.it is giving error in downloading the file. the image size is 1.4mb.i tryed to download the image by using both tftp and ftp . but it is giving the fillowing error. Using TFTP : boot device : fei unit number : 0 processor number : 0 host name : poip6 file name : c:\target\image\vxWorks inet on ethernet (e) : 10.4.5.245:ff000000 inet on backplane (b): 0.0.0.0 host inet (h) : 10.4.5.242 gateway inet (g) : 10.4.5.245 flags (f) : 0x80 target name (tn) : ocelot Attached TCP/IP interface to fei0. Attaching network interface lo0... done. Loading... Transfer Timed Out. tftpGet: Error occurred while transferring the file. Erroneous header read Error loading file: errno = 0x0. Using FTP : boot device : fei unit number : 0 processor number : 0 host name : poip6 file name : vxWorks inet on ethernet (e) : 10.4.5.245:ffffff00 host inet (h) : 10.4.5.242 gateway inet (g) : 10.4.5.245 user (u) : prodtest ftp password (pw) : prodtest flags (f) : 0x0 target name (tn) : ocelot Attached TCP/IP interface to fei0. Attaching network interface lo0... done. Loading... Error loading file: errno = 0xd0003. Thanks in advance for any help Lakshmi Reddy .P From vxwexplo-errs@csg.lbl.gov Mon Feb 4 10:19:52 2002 From: =?iso-8859-1?q?ram=20mohan?= Date: Mon Feb 4 10:19:55 PST 2002 Subject: Re: Image downloading problem hi, first and foremost it seems that u have not started the ftp server on ur host unless otherwise u start u r ftp server ur host can't send the boot image to target this should solve ur prob if u still face prob feel free to ask atb regards Mohan --- the vxWorks Users Group Exploder wrote: > Submitted-by vxwexplo-errs@csg.lbl.gov Mon Feb 4 > 06:16:52 2002 > Submitted-by: "Lakshmi Reddy P" > > > Hai friends, > i am trying to download the vxworks image to > target board.it is giving > error in downloading the file. > the image size is 1.4mb.i tryed to download the > image by using both tftp and > ftp . but it is giving the fillowing error. > > Using TFTP : > > > boot device : fei > unit number : 0 > processor number : 0 > host name : poip6 > file name : > c:\target\image\vxWorks > inet on ethernet (e) : > 10.4.5.245:ff000000 > inet on backplane (b): 0.0.0.0 > host inet (h) : 10.4.5.242 > gateway inet (g) : 10.4.5.245 > flags (f) : 0x80 > target name (tn) : ocelot > > Attached TCP/IP interface to fei0. > Attaching network interface lo0... done. > Loading... Transfer Timed Out. > tftpGet: Error occurred while transferring > the file. > Erroneous header read > > Error loading file: errno = 0x0. > > Using FTP : > > boot device : fei > unit number : 0 > processor number : 0 > host name : poip6 > file name : vxWorks > inet on ethernet (e) : 10.4.5.245:ffffff00 > host inet (h) : 10.4.5.242 > gateway inet (g) : 10.4.5.245 > user (u) : prodtest > ftp password (pw) : prodtest > flags (f) : 0x0 > target name (tn) : ocelot > > Attached TCP/IP interface to fei0. > Attaching network interface lo0... done. > Loading... > Error loading file: errno = 0xd0003. > > > > Thanks in advance for any help > > Lakshmi Reddy .P > > > > > ********** > > This is a user group mailing list for vxWorks > related topics > see http://www-vxw.lbl.gov/vxworks for posting > and subscription info From vxwexplo-errs@csg.lbl.gov Mon Feb 4 12:12:24 2002 From: "Rajaram Soundararajan" Date: Mon Feb 4 12:12:26 PST 2002 Subject: Re: File system - reg Hi, You can partition Flash device into various functional areas. 1. A portion of it can be defined for dosFs or rawFs which is read/writable as any other filesystem. 2. Boot image need not be in the filesystem. May be you make two iamges and flash your boot image into both images. Make sure the booting code switches from one to another in case a crash/fatal exception happened with the earlier image. 3. With TtrueFFS, flash burning may be weathered with long use. Without that I think a portion being written very too often might get burnt very too soon. I dont think the entire image needs to get written back to RAM everytime there is a write cycle. But, of course, when you boot, the image should flow to RAM and execute from there to fire up. Rajaram S. the vxWorks Users Group Exploder wrote: > Submitted-by vxwexplo-errs@csg.lbl.gov Mon Feb 4 04:17:32 2002 > Submitted-by: "praba_ganesh" > > --=_MAILER_ATTACH_BOUNDARY1_20022411744381931706506 > Content-Type: text/plain; charset=us-ascii > > Hi Friends, > > We are assigned the task of developing a Flash driver & integrate the same with a File system. The system is an ARM7TDMI based board. We have to use VxWorks for this purpose. Now we are standing in the design phase. The chip we are planning to use is MBM29-Fujitsu flash. The whole system needs to boot from the FLASH. So the Flash would hold the BOOT image as well as the data area( need to store HTML pages). The problem is that, we can't use the TrueFFS since it is an optional component. The data area of the flash has to store the native system configurations while power down. So the flash would operate in both read/write mode. However in Tornado help manual it is specified that the , Flash that stores the VxWorks image should only be a ROM & shouldn't be Written. If the flash(with image) would undergo write cycles, then the whole image needs to be transferred to RAM at the strt-up. Is my above understanding correct. > > Next regarding the file system: Can I implement the dos file system on the Falsh which stores the image & the data memory. The dos FS would have a FAT table. Will the FAT be a part of Data area. My primary worry is , Can I have a dos FS or RawFS over the flash?. I am not bothered about the Falsh wear cycles at this stage. In case , if the above said file systems can't be used with flash, what would be proper alternative. Above the file system layer there would be the FTP & HTTP servers running. However there would be frequent write cycles into the flash. > > Next problem is, The user should be able to update the VxWorks boot image. Perhaps if the user dumps a corrupted image , then how to bring back the system to live condition. Should I need to go for a seperate boot loader which can point to both old & new image. > > My last query is , Can a NAND flash device be made bootable, where the flash contains both Image & data. > > Thanx & Regards Prabavathy Ganesh From vxwexplo-errs@csg.lbl.gov Mon Feb 4 20:19:47 2002 From: "srujana R" Date: Mon Feb 4 20:19:49 PST 2002 Subject: Re: Image downloading problem On Mon, 04 Feb 2002 the vxWorks Users Group Exploder wrote : > Submitted-by vxwexplo-errs@csg.lbl.gov Mon Feb 4 > 06:16:52 2002 > Submitted-by: "Lakshmi Reddy P" > > > Hai friends, > i am trying to download the vxworks image to > target board.it is giving > error in downloading the file. > the image size is 1.4mb.i tryed to download the image > by using both tftp and > ftp . but it is giving the fillowing error. > > Using TFTP : > > > boot device : fei > unit number : 0 > processor number : 0 > host name : poip6 > file name : > c:\target\image\vxWorks > inet on ethernet (e) : 10.4.5.245:ff000000 > inet on backplane (b): 0.0.0.0 > host inet (h) : 10.4.5.242 > gateway inet (g) : 10.4.5.245 > flags (f) : 0x80 > target name (tn) : ocelot > > Attached TCP/IP interface to fei0. > Attaching network interface lo0... done. > Loading... Transfer Timed Out. > tftpGet: Error occurred while transferring > the file. > Erroneous header read > > Error loading file: errno = 0x0. > > Using FTP : > > boot device : fei > unit number : 0 > processor number : 0 > host name : poip6 > file name : vxWorks > inet on ethernet (e) : 10.4.5.245:ffffff00 > host inet (h) : 10.4.5.242 > gateway inet (g) : 10.4.5.245 > user (u) : prodtest > ftp password (pw) : prodtest > flags (f) : 0x0 > target name (tn) : ocelot > > Attached TCP/IP interface to fei0. > Attaching network interface lo0... done. > Loading... > Error loading file: errno = 0xd0003. > > > > Thanks in advance for any help > > Lakshmi Reddy .P > > > > > ********** > > This is a user group mailing list for vxW lbl.gov/vxworks for posting and > subscription info From vxwexplo-errs@csg.lbl.gov Mon Feb 4 23:14:42 2002 From: "Sairam" Date: Mon Feb 4 23:14:45 PST 2002 Subject: ISR using C++ Hi Friends, I am writing an ISR using C++. This ISR routine is a member function of a class A. I am writing the intConnect() within a member function of class B. Now in class B I am initiating a pointer to class A and using this pointer I am passing the ISR routine of class A to intConnect(). But the ISR is not getting accessed. Here I am setting flag to 1 in the ISR but the value is remaining 0 when accessed in root. #include "vxWorks.h" #include "intLib.h" #include "pc.h" #include "sysLib.h" #include "config.h" #include "iostream.h" extern "C" { void root(void); } class A{ public: int flag; void isr() { logMsg("in isr****",0,0,0,0,0,0); flag = 1; } }; class B { public: A *ptr; void connect(); }; void B:: connect() { printf("\n connect status :%d \n",intConnect(INUM_TO_IVEC(COM1_INT_VEC),(VOIDFUNCPTR)ptr->isr,0)); printf("\n ERROR is %d OK id %d",ERROR,OK); } void root(void) { class B b; class A a; a.flag =0; b.ptr = &a; b.connect(); sysOutByte(0x3f8 + 1,0x02); sysOutByte(0x21,(sysInByte(0x21)&0xef)); while(1) printf("%d \n",c.ptr->flag); /* 1 should be displayed */ } From vxwexplo-errs@csg.lbl.gov Tue Feb 5 04:03:13 2002 From: Vxworks Exploder Date: Tue Feb 5 04:03:16 PST 2002 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Tue Feb 5 04:03:10 PST 2002 Subject: 8260 - Tx Error occurred in 10Mbps Subject: ways to get inbound/outbound IP packets? Subject: Re: vxEieio Subject: Re: DOS key like features in vxworks Subject: Re: Returning memory Subject: Expert at HTML, JavaScript...? Subject: pciFindDevice()- index 0 based or 1 based ? Subject: Re: Tornado2 and PVCS Subject: Re: CLI Parser for VxWorks Subject: Re: frequent rebooting for an ethernet interface looses connectivity?? Subject: Re: udbinfo error status msg, what is it? Subject: Initializing a semaphore Subject: Load microcode into ixp1200 processor using vxworks application. Subject: Re: Initializing a semaphore Subject: Prototyper in Windows 2000 ! Subject: Re: What's usage of ccppc? Subject: Re: frequent rebooting for an ethernet interface looses connectivity?? Subject: How to control a modem? Subject: adding compilation flag for good - how? Subject: Re: pciFindDevice()- index 0 based or 1 based ? Subject: [HelpMe] unld() error ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: 8260 - Tx Error occurred in 10Mbps Date: 4 Feb 2002 04:05:52 -0800 From: scsuh@secui.com (S.Chull Suh) Organization: http://groups.google.com/ Message-ID: Hi, I'm using FCC1 and 2 as a fast ethernet controller to make network component. Here are my problems. 1. when I use FCC1 or 2 in 100Mbps(full-duplex or half-duplex), everything is fine. But when I use them in 10Mbps channels reports many tx errors. 2. and when Tx error in one controller occurres both controllers are dead. My environment is below 1. vxWorks 5.4 2. MPC8260 - XPC8260ZU166A 166/133/66MHZ, 1K22A CRQAX0029 HKG 3. LX970A PHY chip Now, I set following things. 1. FPSMR is 0x00000080 -> this means (i think) half-duplex. 2. CR SR ANA ANLPA IER ISR CONF CSR ------------------------------------------------------------- 1000 7829 01e1 0021 0000 4000 0000 2089 Link Status: motfcc1: Auto Nego - 10Mbps Half Duplex I think LX970 is set correctly. I wonder that 1. why TX ERROR occurs 2. something more is set when I use 10Mbps Thanks... I hope that anybody tells me why TX error occurs. --------------------------- Newsgroups: comp.os.vxworks Subject: ways to get inbound/outbound IP packets? Date: 4 Feb 2002 04:18:58 -0800 From: sitalakshmi.gautham@wipro.com (Sitalakshmi Gautham) Organization: http://groups.google.com/ Message-ID: <2fe901df.0202040418.3389a7b1@posting.google.com> Are there alternatives to receiving Incoming IP packets (other than muxBind(...MUX_PROTO_SNARF) or IpFilterHooks)? Since WDB is already bound to MUX_PROTO_SNARF, this cannot be done - also Hooks support may not be available in Future VxWorks versions. Similarly are there any alternatives to receiving outbound IP packets (other than muxBind(...MUX_PROTO_OUTPUT..)? basically i am looking at this for writing a packet filtering code, which works on outbound/incoming IP packets. Also when IpDetach() is done over an interface, is there a way for my module to know that IP stack no long is attached to a particular interface? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: vxEieio Date: Mon, 4 Feb 2002 12:14:44 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: "Vinh" wrote in message news:c129947c.0202011419.73b4ab76@posting.google.com... > Hi, > I found a usefull routine 'vxEieio' in 'LibPPC604gnuvx.a' but there > are no prototypes for it in any of vxWorks header file. I think the > routine is a macro with one assembly instruction 'eieio', but i can't > verified this with the 'objdumpppc'. > > I would like to know if any one has try using this routine, and if > so how can i ensured that is it just a macro 'eieio'. H051> l vxEieio vxEieio: 0x112e7c 7c0006ac eieio 0x112e80 4e800020 blr Go for it. DaveK - -- moderator of alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! Proud Member of the Exclusive "I have been plonked by Davee because he thinks I'm interesting" List Member # Master of Many Meowing Minions Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage above and beyond the call of hilarity. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: DOS key like features in vxworks Date: Mon, 4 Feb 2002 12:33:15 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: <9b6f8883.0201311117.2e1b5206@posting.google.com> "Mark Chang" wrote in message news:9b6f8883.0201311117.2e1b5206@posting.google.com... > Does anyone know of code or have > code that will do DOS Key like features? > You know, up arrow show history, line editing, insert and replace. > I think it will have to replace tylib.c - It's built in: look up ledlib.h for vxworks line editing features. DaveK - -- moderator of alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! Proud Member of the Exclusive "I have been plonked by Davee because he thinks I'm interesting" List Member # Master of Many Meowing Minions Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage above and beyond the call of hilarity. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Returning memory Date: Sat, 2 Feb 2002 09:56:29 -0800 From: "GaryM" Message-ID: References: <28809481.0202012009.10cf7545@posting.google.com> No. Since VxWorks cannot determine what the memory block is used for and with what other components it is being shared, memory blocks are never deleted automatically when the allocating task is deleted. "rasayani" wrote in message news:28809481.0202012009.10cf7545@posting.google.com... > If a task allocates memory using malloc/calloc and is deleted, is the > memory released to the free memory pool? > > -vs --------------------------- Newsgroups: comp.os.vxworks Subject: Expert at HTML, JavaScript...? Date: Mon, 04 Feb 2002 14:43:39 GMT From: "MX" Organization: Bentium Ltd. (CN99) Message-ID: Are you guys expert at HTML, JavaScript, Internet Protocol, Window Sockets, and 101 many other things? We have a very special browser for you. * Customize web pages in ways you like. * Prevent malicious web sites from invading your system. * Restrict any applications' connections to the outside world. * Chat or exchange files with people without a server in between. * Trace the Windows Socket activities that underlie network applications .......... Visit http://alanet.51.net/ for more! --------------------------- Newsgroups: comp.os.vxworks Subject: pciFindDevice()- index 0 based or 1 based ? Date: 4 Feb 2002 06:46:36 -0800 From: arun4vxworks@indiatimes.com (Arun Prasad) Organization: http://groups.google.com/ Message-ID: Hi, I have a doubt on the index parameter of pciFindDevice funcion. I have only one PCI device in my system and I'm getting success if I call the function with index value passed either 0 or 1. Can anyone tell me whether this index value starts with 0 or 1? I'm using IBM PPC 405GP. Thanks in advance. Regards Arun Arun Prasad Software Engineer Adaptec India Pvt. Limited (A wholly owned subsidiary of Adaptec Inc., USA) 6-3-1086, IV Floor, Vista Grand Towers, Rajbhavan Road, Somajiguda, Hydereabad - 500082 Phone: (O):91-40-6661555/56/57/58 (Ext):234 Email: arun_prasad@adaptecin.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado2 and PVCS Date: Mon, 4 Feb 2002 15:23:27 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: <3c587646$1@brateggebdc5.br-automation.co.at> "Martin Raabe" wrote in message news:a3asq2$j3r$07$1@news.t-online.com... > Also MS Source Safe doe not support branches, which might be changes being > - customer speciffic > - bug fixing > - new feature testing > - etc. In case anyone has been fooled by this statement, it is of course garbage. MS SourceSafe fully supports branching and merging. Martin, does your management know you are using a WRS email address to post potentially libellous mis-statements in public about a competitor's product ? Do they approve ? How can you even imagine for one minute that SourceSafe doesn't have branching ? That would be an incredibly vital feature to omit from *any* source control software. > I mentioned SNiFF+ here, because I don't do any development without it, > is it doen with Tornado or pRISM+ or Visual Studio. Not to forget the fact that you *always* plug SNiFF+ in here. > Just my two new Euro Cents. Is that how much commission you get on each sale ? DaveK - -- moderator of alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! Proud Member of the Exclusive "I have been plonked by Davee because he thinks I'm interesting" List Member # Master of Many Meowing Minions Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage above and beyond the call of hilarity. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: CLI Parser for VxWorks Date: Mon, 4 Feb 2002 15:51:31 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: <182ee52a.0201251347.2a1d1421@posting.google.com> <9b6f8883.0201251727.781eea1a@posting.google.com> "Mark Chang" wrote in message news:9b6f8883.0201251727.781eea1a@posting.google.com... > dpraveen@hotmail.com (Praveen Dulam) wrote in message news:<182ee52a.0201251347.2a1d1421@posting.google.com>... > > Hi > > > > I am looking for CLI commands parser for VxWorks. > > > > TIA > > Praveen > > If you need it free - use TCL (we use 7.3 in our embedded box) > or use one of WindRiver's RapidControl for CLI. > Mark > [assuming CLI refers to a command-line here] Other good options are the free source utilities lexx/yacc and bison, all of which provide a simple C-like syntax for constructing parsers of pretty much any complexity you like. DaveK - -- moderator of alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! Proud Member of the Exclusive "I have been plonked by Davee because he thinks I'm interesting" List Member # Master of Many Meowing Minions Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage above and beyond the call of hilarity. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: frequent rebooting for an ethernet interface looses connectivity?? Date: Mon, 04 Feb 2002 14:42:56 -0500 From: Babu Yama Organization: Lucent Technologies Message-ID: <3C5EE440.F4C4B94@lucent.com> References: <3C5DF5A5.1F6CFFCC@lucent.com> <9ec53d11.0202040118.514e0357@posting.google.com> Sorry, I forgot the following information. My target details are: = target: MPC8260 (FCC2 ethernet port) - when ping fails FCC1 is fine and t= alking. vxWorks: 5.4 BSP: wrSbc8260 BSP 33 (wind river's single board compter FCC driver) Any ideas as to why the FCC2 ethernet port does not talk after some reboo= ts.... Thanks, Babu gnacio Gonz=E1lez wrote: > = > Babu Yama wrote in message news:<3C5DF5A5.1F6CFFC= C@lucent.com>... > > Hi, > > My ethernet device driver runs on a target that reboots "often". When= the target > > reboots, sometimes, immediately following the reboot, the target cann= 't be > > pinged from the network. Then, the subsequent reboot restores the net= work > > connectivity to the target. This happens intermittently. > > > > Any ideas as to what might be causing this. Is ARP playing a role her= e? Some > > thing I need to do when I bring up the interface each time? > > Apprecaite your help, > > > > Thanks in advance, > > Babu > = > The description of the problem is too vague for me. Which version of > VxWorks do you have ? BSP ? Kind of target ? > = > I had a problem similar to yours: after a reboot, my target sometimes > refused to send messages through Ethernet (as a response to ping or > when taking the initiative of sending messages). It was caused by a > bug in the Ethernet driver that I managed to fix (later, the driver's > supplier learned of it and also fixed the driver.) --------------------------- Newsgroups: comp.os.vxworks Subject: Re: udbinfo error status msg, what is it? Date: Mon, 04 Feb 2002 15:05:53 -0600 From: Andrew Johnson Organization: APS, Argonne National Laboratory Message-ID: <3C5EF7B1.284FF522@aps.anl.gov> References: <6dcae13d.0202041150.3eb2d165@posting.google.com> roswell wrote: > > i am tyring to understand what does this: > value = -1 = 0xffffffff = udbinfo + 0xffed79af > > means? is the symbtbl srewed up?.. No, it just means that whatever command you just executed returned the value -1, and the highest symbol in your symbol table is probably udbinfo (which in any case has value the 0xffffffff - 0xffed79af = 0x00128650). It is normal for the shell to try and match return values to the nearest preceeding symbol in case you want to know the offset from that symbol, although IIRC it doesn't always do so; if you don't want to know that offset, just ignore that bit. - - Andrew - -- Perfection is reached, not when there is no longer anything to add, but when there is no longer anything to take away. - - Antoine de Saint-Exupery --------------------------- Newsgroups: comp.os.vxworks Subject: Initializing a semaphore Date: Mon, 4 Feb 2002 17:51:17 -0500 From: "nitin kumar" Organization: Marconi Message-ID: Hi, I have a structure of the following type .... static struct foo { foobar_t *ptr; SEMID sem; uint32_t refcnt; } I need to initialize this struct only once during the execution of my code. How can I do something like ... static struct foo { foobar_t *ptr; SEMID sem; uint32_t refcnt; } = { NULL, semBCreate(), 0 }; I know I cannot do a semBCreate() in the initializing value. How do I initialize the semaphore there ?...I remember pthreads having a PTHREAD_MUTEX_INITIALIZER ...thingie which I could put over there... The only way I can think of now is to have aanother pointer to struct foo, .. static struct foo { }..*foo_ptr = NULL; ANd then do somrething like.. if ( !foo_ptr ) { foo_ptr = malloc( sizeof( struct foo )); memset( foo_ptr, 0, sizeof( *foo_ptr )); foo_ptr->sem = semBCreate(); } Thanks all nitin --------------------------- Newsgroups: comp.os.vxworks Subject: Load microcode into ixp1200 processor using vxworks application. Date: 4 Feb 2002 17:31:29 -0800 From: hansc1@chollian.net (Seungchul Han) Organization: http://groups.google.com/ Message-ID: I want to load my microcode(*.uof) into ixp1200 processor using vxworks application. Because I have debuged my program with workbench till now, I have no problem. Now I will release products, but I don't know how to operate system stand-alone. I would be grateful for your advice on the problems I am experiencing. Thanks in advance. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Initializing a semaphore Date: Mon, 04 Feb 2002 20:30:50 -0500 From: joe durusau Organization: Bellsouth.Net Message-ID: <3C5F35CA.19B69434@bellsouth.net> References: nitin kumar wrote: > Hi, > I have a structure of the following type .... > > static struct foo { > foobar_t *ptr; > SEMID sem; > uint32_t refcnt; > } > > I need to initialize this struct only once during the execution of my code. > How > can I do something like ... > > static struct foo { > foobar_t *ptr; > SEMID sem; > uint32_t refcnt; > } = { NULL, semBCreate(), 0 }; > > I know I cannot do a semBCreate() in the initializing value. How do I > initialize the > semaphore there ?...I remember pthreads having a PTHREAD_MUTEX_INITIALIZER > ...thingie > which I could put over there... > > The only way I can think of now is to have aanother pointer to struct foo, > .. > > static struct foo { > }..*foo_ptr = NULL; > > ANd then do somrething like.. > > if ( !foo_ptr ) { > foo_ptr = malloc( sizeof( struct foo )); > memset( foo_ptr, 0, sizeof( *foo_ptr )); > foo_ptr->sem = semBCreate(); > } > > Thanks all > > nitin Well, one way is: static int first =1; ... if(first==1) { first =0; semBCreate() } The stuff between the first two braces will execute one and only one time, barring reloading of the executable or rebooting. Speaking only for myself, Joe Durusau --------------------------- Newsgroups: comp.os.vxworks Subject: Prototyper in Windows 2000 ! Date: Tue, 05 Feb 2002 04:48:52 GMT From: "Jamilur Rahman" Message-ID: I've proper license and environment set for the 'WRSLicense.lic' file to run prototyper. But my target server doesn't start. Has anybody worked on Prototyper in W2000 ? Any help will be appreciated. /Rokeya. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: What's usage of ccppc? Date: 5 Feb 2002 05:52:54 GMT From: Denis Perelyubskiy Message-ID: References: * Liang Yang [02-Fri-02 17:05 -0800]: > > Hi, > Which company developed ccppc used in Tornado? Where can I get detailed > manul about the usage of ccppc? try ccppc --version to see if maybe you get something useful back. i'd venture a guess that you would. wrs is waaaay behind times in compilers (at least for x86 they are), so i would not hold my breath as far as getting a newer/working one goes :) but in any case, that's something you'd probably want to ask from your salesperson or fai, or whoever it is you're supposed to be talking to ... they will need to know your current version anyway though denis - -- if you want to reply by mail, please tweak my address in a reasonable way. sorry for the inconvenience. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: frequent rebooting for an ethernet interface looses connectivity?? Date: Tue, 5 Feb 2002 08:24:58 +0100 From: Mark Koennecke Message-ID: References: <3C5DF5A5.1F6CFFCC@lucent.com> <9ec53d11.0202040118.514e0357@posting.google.com> <3C5EE440.F4C4B94@lucent.com> Dear Babu, On Mon, 4 Feb 2002, Babu Yama wrote: > Sorry, I forgot the following information. > > My target details are: > > target: MPC8260 (FCC2 ethernet port) - when ping fails FCC1 is fine and talking. > vxWorks: 5.4 > BSP: wrSbc8260 BSP 33 (wind river's single board compter FCC driver) > > Any ideas as to why the FCC2 ethernet port does not talk after some reboots.... > May be this is of some help: I had massive networking problems with vxWorks on a mvme1603, it hung up and often ARP error messages appeared at the console. The reason was a broadcast storm on the network. As the size of networking buffers and response to trouble on the net is OS dependent you might consider checking this. Regards, Mark Koennecke --------------------------- Newsgroups: comp.os.vxworks Subject: How to control a modem? Date: 4 Feb 2002 23:39:39 -0800 From: chwin@china.com (Adom) Organization: http://groups.google.com/ Message-ID: <92dbccc1.0202042339.5e273cd7@posting.google.com> Hi,All! i have decide to research a gateway for home . it's hardware platform is ppc860 and there is one ethernet adapter. i prepare for use it's serial to control a modem to dial up. after i use at* command , the modem has establisted a link between modem and ISP, now ,what should i do to init this ppp interface and communicate with peer? In vxworks, pppInit(...) will create two task , first is pppD, do LCP and IPCP config, then pppWrite task, attach and init ppp interface. But now, the LCP and IPCP have been set by modem and ISP, how can i init ppp interface ! Any suggestions would be appreciated Thank you very much! Adom Chan 2002-2-5 leadit@hotmail.com --------------------------- Newsgroups: comp.os.vxworks Subject: adding compilation flag for good - how? Date: Tue, 5 Feb 2002 12:17:30 +0200 From: "Ofer Goren" Organization: Verio Message-ID: - -- - ------------------------ Win2k Sp2 Tornado 2.02 for PPC Diab 4.4b Envoy 9.2 - ------------------------ Hi. I'm trying to add a compilation flag to my Tornado, so it will always be used, instead of adding it manually in each new project I create. I'm using diab, so I tried changing ../config/h/make/defs.diab file. However, it didn't changed me compilation. Even after deleteing the project, shuthing down Torando, re-open it, and creating a new project. The flags I changed are CC_COMPILER and C++COMPILER (added -Xlint=1). So, I went to all possible locations, even run "wtxtcl gnuInfoGen.tcl [vendor]", as suggested in gnuInfoGet.tcl file. Nothing helped. Any idea? - -- Ofer Goren Iamba Technologies --------------------------- Newsgroups: comp.os.vxworks Subject: Re: pciFindDevice()- index 0 based or 1 based ? Date: 5 Feb 2002 10:41:06 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: arun4vxworks@indiatimes.com (Arun Prasad) wrote: > Hi, > > I have a doubt on the index parameter of pciFindDevice funcion. > > I have only one PCI device in my system and I'm getting success if I > call the function with index value passed either 0 or 1. > > Can anyone tell me whether this index value starts with 0 or 1? As far as I know the PCI bridge has ID 0, and your device has ID 1. Check the device ID's that you are getting back from pciDeviceShow and pciHeaderShow. Groeten, Johan - -- o o o o o o o . . . _____________________________ o _____ || Johan Borkhuis | .][__n_n_|DD[ ====_____ | borkhuis@agere.com | >(________|__|_[_________]_|__________________________| _/oo OOOOO oo` ooo ooo 'o!o!o o!o!o` === VxWorks FAQ: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html === --------------------------- Newsgroups: comp.os.vxworks Subject: [HelpMe] unld() error Date: Tue, 05 Feb 2002 03:36:59 GMT From: "Kim, Jeong-Hwan" Organization: Dacom Message-ID: Hi, everyone Could you tell me in which case unld() function fails? Thanks in advance Kim. --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Tue Feb 5 04:28:55 2002 From: "Hans-Peter Kabout" Date: Tue Feb 5 04:28:57 PST 2002 Subject: RE: ISR using C++ This is a multi-part message in MIME format. ------=_NextPart_000_000D_01C1AE48.C3F31290 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Hi C++ writers, A nonstatic member function uses the this keyword, which is a const pointer to the object for which the function was called. flag = 1; is implicitely translated to this->flag = 1; because both, isr and flag have internal linkage. You cannot call functions from ASM or C code that have internal linkage because *this is unknown in then ASM and C code. To get external linkage you need class A { public: static int flag; static void isr(); }; Static member functions are considered to have class scope. In contrast to nonstatic member functions, these functions have no implicit this argument; therefore, they can use only static data members, enumerators, or nested types directly. Static member functions can be accessed without using an object of the corresponding class type. Regards, H.-P. Kabout, IT-Consulting e-mail: Hans-Peter@Kabout.de Phone: +49 (0)9085 - 1017 Mobile: +49 (0)160 - 7556913 www.kabout.com -----Original Message----- From: the vxWorks Users Group Exploder [mailto:vxwexplo@lbl.gov] Sent: Tuesday, February 05, 2002 08:32 To: vxworks_users@csg.lbl.gov Subject: ISR using C++ Submitted-by vxwexplo-errs@csg.lbl.gov Mon Feb 4 23:14:42 2002 Submitted-by: "Sairam" Hi Friends, I am writing an ISR using C++. This ISR routine is a member function of a class A. I am writing the intConnect() within a member function of class B. Now in class B I am initiating a pointer to class A and using this pointer I am passing the ISR routine of class A to intConnect(). But the ISR is not getting accessed. Here I am setting flag to 1 in the ISR but the value is remaining 0 when accessed in root. #include "vxWorks.h" #include "intLib.h" #include "pc.h" #include "sysLib.h" #include "config.h" #include "iostream.h" extern "C" { void root(void); } class A{ public: int flag; void isr() { logMsg("in isr****",0,0,0,0,0,0); flag = 1; } }; class B { public: A *ptr; void connect(); }; void B:: connect() { printf("\n connect status :%d \n",intConnect(INUM_TO_IVEC(COM1_INT_VEC),(VOIDFUNCPTR)ptr->isr,0)); printf("\n ERROR is %d OK id %d",ERROR,OK); } void root(void) { class B b; class A a; a.flag =0; b.ptr = &a; b.connect(); sysOutByte(0x3f8 + 1,0x02); sysOutByte(0x21,(sysInByte(0x21)&0xef)); while(1) printf("%d \n",c.ptr->flag); /* 1 should be displayed */ } ------=_NextPart_000_000D_01C1AE48.C3F31290 Content-Type: text/html; charset="us-ascii" Content-Transfer-Encoding: quoted-printable

Hi C++ = writers,

 

A nonstatic member function uses the this keyword, which = is a const pointer to the object for which the function was called. flag =3D 1; is implicitely translated to

this->flag =3D = 1;

because both, isr and flag have internal linkage.

 

You cannot call functions from = ASM or C code that have internal linkage because *this is = unknown in then ASM and C code.

 

To get external linkage you = need

class A

{

    public:

        = static int  = flag;

        = static void isr();

};

Static member functions are = considered to have class scope. In contrast to nonstatic member functions, these functions have no implicit this argument; therefore, = they can use only static data members, enumerators, or nested types directly. = Static member functions can be accessed without using an object of the = corresponding class type.

 

Regards,

H.-P. Kabout, IT-Consulting =

e-mail: <A HREF=3D"mailto:Hans-Peter@Kabout.de">Hans-Peter@Kabout.de</A><= /p>

 Phone: +49 (0)9085 - 1017 =

Mobile= : +49 (0)160 - 7556913 =

<A = HREF=3D"http://www.kabout.com/index.html">www.kabout.com<= /A>

 

-----Original = Message-----

From: = the vxWorks Users Group = Exploder [mailto:vxwexplo@lbl.gov]

Sent: Tuesday, February 05, 2002 08:32

To: = vxworks_users@csg.lbl.gov

Subject: ISR using = C++

 

Submitted-by vxwexplo-errs@csg.lbl.gov  Mon Feb  4 23:14:42 = 2002

Submitted-by: "Sairam" <sairam@cmcltd.com>

 

Hi = Friends,

 

I am writing an ISR using C++. = This ISR routine is a member function of a class A. I am writing the intConnect() = within a member function of class B. Now in class B I am initiating a pointer to class A = and using this pointer I am passing the ISR routine of class A to intConnect(). But the ISR is not getting accessed. Here I am setting flag to 1 in the = ISR but the value is remaining 0 when accessed in = root.

 

#include "vxWorks.h"

#include "intLib.h"

#include "pc.h"

#include "sysLib.h"

#include "config.h"

#include "iostream.h"

 

extern = "C"

{

 void = root(void);

}

 

class A{

 

public:

int flag;

 

void isr()

{

logMsg("in isr****",0,0,0,0,0,0);

flag =3D = 1;

}

};

 

 

class B

{

public:

A *ptr;

void = connect();

};

 

void B:: = connect()

{

printf("\n connect status :%d \n",intConnect(INUM_TO_IVEC(COM1_INT_VEC),(VOID= FUNCPTR)ptr->isr,0));

printf("\n ERROR is %d OK id %d",ERROR,OK);

}

 

 

void = root(void)

{

class B b;

class A a;

a.flag = =3D0;

b.ptr =3D = &a;

b.connect();

 

sysOutByte(0x3f8 + 1,0x02); sysOutByte(0x21,(sysInByte(0x21)&0xef));

while(1)

printf("%d \n",c.ptr->flag);   /* 1 should be displayed = */

}

 

------=_NextPart_000_000D_01C1AE48.C3F31290-- From vxwexplo-errs@csg.lbl.gov Tue Feb 5 07:34:02 2002 From: "Shah, Ranjana" Date: Tue Feb 5 07:34:04 PST 2002 Subject: Re: ISR using C++ VxWorks Sairam, When a class member function is called, it needs the pointer to the corresponding object as the first parameter. So, try passing the pointer to the object a (instance of class A) when you call intConnect() as the first parameter instead of 0. i.e., the first parameter before any explicit parameters. This is because when a member function is called, the C++ compiler generates code such that it looks for the address of the object as the first parameter implicitly. If the function takes no explicit parameters (which is true in your example), even then pass the pointer to the object a casted with int. Possibly, the following call might work: intConnect(INUM_TO_IVEC(COM1_INT_VEC),(VOIDFUNCPTR)ptr->isr, (int)ptr) as part of the printf. Hope this helps. Ranjana Shah Staff Engineer Telephonics Corp. Voice: (631) 549-6104 FAX: (631) 755-7642 From vxwexplo-errs@csg.lbl.gov Wed Feb 6 01:04:13 2002 From: Atkinson Andrew Date: Wed Feb 6 01:04:15 PST 2002 Subject: Cache / CIO Query - vxworks Hi A quick question of cache. I need to interface to a Zilog Z8536 CIO. My BSP already has code to do some of what I want, but the code contains the following which I am unclear on: /* CIO initialize: insure register sync */ tc = 60; /* default time constant */ cioCtl = ZCIO_CNTRL_ADRS; temp = *cioCtl; /* see CIO manual sec 2.3 */ *cioCtl = ZERO; temp = *cioCtl; CACHE_PIPE_FLUSH(); /* reset CIO chip */ *cioCtl = ZCIO_MIC; /* Master Interrupt Control */ *cioCtl = ZCIO_MIC_RESET; /* device RESET */ temp = CIO_RESET_DELAY; /* Reset delay */ do ; /* wait for reset */ while (temp --); I do not want to go OTT with unnecessary details on the device. Rather I am hoping that someone with some experience of this device (or similar) can explain the section which has been indented. In particular, what is the relationship between this code and the devices state machine and also what rules are dictating the use of the CACHE_PIPE_FLUSH macro. Any help appreciated. Andy Atkinson Software Engineer Bldg 31a QinetiQ, Bincleaves Technology Park Weymouth Dorset DT4 8UR Tel : 01305 764660 Fax : 01305 764661 e-mail : AATKINSON1@QinetiQ.com -- The Information contained in this E-Mail and any subsequent correspondence is private and is intended solely for the intended recipient(s). For those other than the recipient any disclosure, copying, distribution, or any action taken or omitted to be taken in reliance on such information is prohibited and may be unlawful. From vxwexplo-errs@csg.lbl.gov Wed Feb 6 01:41:53 2002 From: "Lakshmi Reddy P" Date: Wed Feb 6 01:41:56 PST 2002 Subject: Downloading an application on to the target Hi Friends, We are using ocelot board (CP7000 processor)and VxWorks tornado version 2.0 We have downloaded BSP for ths processor on to the target. I am not able to ping to the interface and i am not able to download an application on to the taget. If anybody get the idea to download an application on to the target, please reply yo me. Thanx& Regards Laxmi Reddy From vxwexplo-errs@csg.lbl.gov Wed Feb 6 03:03:09 2002 From: "Pradeepa Venkatswamy" Date: Wed Feb 6 03:03:11 PST 2002 Subject: how to get stack pointer for x86 in vxworks This is a multi-part message in MIME format. ------=_NextPartTM-000-c41f3e56-1ae9-11d6-a217-0000e22173f5 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Hi, We would like to know how we can get the present value of stack pointer for x86 target in vxworks.Please help us. Regards, V.Pradeepa. ------=_NextPartTM-000-c41f3e56-1ae9-11d6-a217-0000e22173f5 Content-Type: text/plain; name="InterScan_Disclaimer.txt" Content-Transfer-Encoding: 7bit Content-Disposition: attachment; filename="InterScan_Disclaimer.txt" **************************Disclaimer************************************ Information contained in this E-MAIL being proprietary to Wipro Limited is 'privileged' and 'confidential' and intended for use only by the individual or entity to which it is addressed. You are notified that any use, copying or dissemination of the information contained in the E-MAIL in any manner whatsoever is strictly prohibited. ******************************************************************** ------=_NextPartTM-000-c41f3e56-1ae9-11d6-a217-0000e22173f5-- From vxwexplo-errs@csg.lbl.gov Wed Feb 6 03:34:40 2002 From: S Ramesh Date: Wed Feb 6 03:34:42 PST 2002 Subject: Deleting pipes Hi, In VxWorks pipes are created using pipeDevCreate call. I think we can delete the pipe using iosDevDelete call. But the whole memory allocated for the pipe is not being deleted and it is created memory leaks. Is this a feature or bug or how to overcome this problem. Thanks Ramesh From vxwexplo-errs@csg.lbl.gov Wed Feb 6 03:36:54 2002 From: Jim Bryden Date: Wed Feb 6 03:36:56 PST 2002 Subject: BSP for GMS Atlantis Hello list, Has anyone any experience with the C191-Atlantis board from General Micro Systems? Its a PPC74xx based cPCI board. We were thinking using it with VxWorks, so any information on the maturity of the bsp and support tools would be of interest. Thank you in advance. Jim Bryden This e-mail is intended solely for the above mentioned recipient and it may contain confidential or privileged information. If you have received it in error, please notify us immediately by telephone and delete the email. You must not copy, distribute, disclose or take any action in reliance on it. Whilst every reasonable precaution has been taken, Thales Optronics can not accept liability for any damage which you sustain as a result of software viruses that may be contained in, or attached to this email. -------------------------------------------------- From vxwexplo-errs@csg.lbl.gov Wed Feb 6 04:03:18 2002 From: Vxworks Exploder Date: Wed Feb 6 04:03:20 PST 2002 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Wed Feb 6 04:03:14 PST 2002 Subject: Re: why is it so Subject: does my MIB leaf exists ? Subject: Re: Configuration of a serial device to run full duplex Subject: Use of DMA controller on PC486 board Subject: Wind Web Server question Subject: Re: pciFindDevice()- index 0 based or 1 based ? Subject: motFccEnd.c Subject: Re: Initializing a semaphore Subject: why is it so Subject: UDF 1.5 support for VxWorks Subject: handling multiple interrupts Subject: Re: handling multiple interrupts Subject: Re: Use of DMA controller on PC486 board Subject: getproc_and_instance MIB Subject: WindView.win32.tcl error in T2 Subject: Re: CLI Parser for VxWorks Subject: Re: curses Subject: Re: motFccEnd.c Subject: Re: UDF 1.5 support for VxWorks Subject: Re: DHCP over PPP Subject: Re: DHCP over PPP Subject: Symbol table synchronization issue (Tornado 2/PPC/Solaris) Subject: Re: Symbol table synchronization issue (Tornado 2/PPC/Solaris) Subject: Re: Tornado2 and PVCS Subject: Serial Driver Lockup? Subject: tornado shell, tasks, printf not working Subject: Re: pciFindDevice()- index 0 based or 1 based ? Subject: Time Base register update rate Subject: Re: Time Base register update rate Subject: Re: motFccEnd.c Subject: about lex Subject: Re: why is it so Subject: Intermittent reboots Subject: Re: WindView.win32.tcl error in T2 ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: why is it so Date: Tue, 5 Feb 2002 13:08:08 +0100 From: "Michael Lawnick" Organization: Buergernetz Dillingen Message-ID: References: Reply-To: "Michael Lawnick" Sender: mlawnick@pd95245d7.dip.t-dialin.net Hi Ali, just as your output says: multiple definition of `endDevTbl' first in sysMotFccEnd.c, included by sysLib.c, scnd in usrBootLine.c, included by prjConfig.c AFAIK it is an array of structures. You should have only one defined, so making them local won't solve your problem. - -- Mit freundlichen Grüßen, Michael Lawnick ============================================== SOFTEC GmbH Tel +49-731-96600-0 Promenade 17 Fax +49-731-96600-23 D-89073 Ulm Michael Lawnick Germany lawnick@softec.de ============================================== "Ali" schrieb im Newsbeitrag news:b305b5bd.0202050028.4252dd51@posting.google.com... > hi all > When i build my bsp it gives me two peculiar errors at last can > anybody tell me what it actually looks for i tried much to solve this > but it comes again and again.Though the lines it is specifying has no > entry like this.... > here is snapshot of this build > > ccppc -B/home/alimohd/tornado/host/sun4-solaris2/lib/gcc-lib/ > -nostdlib -r -Wl,-X \ > -Wl,@/home/alimohd/tornado/target/proj/ep8260boot/prjObjs.lst > version.o /home/alimohd/tornado/target/lib/libPPCEC603gnuvx.a -o > partialImage.o > prjConfig.o: In function `usrBootLineInit': > /home/alimohd/tornado/target/config/comps/src/net/usrBootLine.c:36: > multiple definition of `endDevTbl' > sysLib.o:/home/alimohd/tornado/target/config/ep8260/sysMotFccEnd.c:180: > first defined here --------------------------- Newsgroups: comp.os.vxworks Subject: does my MIB leaf exists ? Date: 5 Feb 2002 05:48:30 -0800 From: ilyasg@yahoo.com (Ilyas G) Organization: http://groups.google.com/ Message-ID: Hi i have created a personal MIB i have followed the example shown in the guide, I download it then add the tree to the MIB. I use Vxsim !!! no probleme, everything Ok ! i have installed a snmp manager and it's able to see the target, but if I try a get or set , a timeout error occurs. is there any way to check if a MIB variable exist, I have created under experimental.17.1 . does someone have a little routine to check/set the value from it's OID ??? thanks --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Configuration of a serial device to run full duplex Date: Tue, 05 Feb 2002 14:25:36 GMT From: soren_abildgaard@hotmail.com (Sřren Abildgaard) Organization: TDC Internet Message-ID: <3c5fe909.19193765@news.tele.dk> References: <3c57b060$0$89074$edfadb0f@dspool01.news.tele.dk> Hi Friedrich Jakob has tried starting 2 tasks for read and write, and the problem still exists. The baudrate is 19.2 kbps. I don´t know the buffer size, but it should be sufficient. We know the VxWorks deliveres driver dosn´t support the 16550 FIFO´s, but are they neccesary at this baud rate? Any other suggestions on where the problem could be ? Regards, Sřren Abildgaard Co-Worker On Fri, 1 Feb 2002 10:17:40 +0100, "Friedrich Ensslin" wrote: >Hi Jakob, > >it seems to me, that the problem might have to do with character buffering >inside the driver. Your observations might be caused by the fact, that the >receiver side of the driver does no sufficient data buffering. When your >code arrives at the read instruction, the first three bytes are already lost >(what's the data rate on /tyCo/1 ?). >To clarify this, it might be helpful to look at the driver's configuration >sequence and sourcecode if available. What's the size of the buffers >(especially RX) ? >Try to start a parallel task doing the while/read loop before writing to the >loopback. > >Friedrich Ensslin > >________________________________________ > >AM3 AutoMotive MultiMedia AG >friedrich.ensslin@am3.com >________________________________________ > > >"Jakob Holm-Petersen" schrieb im >Newsbeitrag news:3c57b060$0$89074$edfadb0f@dspool01.news.tele.dk... >> Hi, >> >> I'm right now working on a project where we need to make a physical loop >> back on a serial device. The loop back function is used to verify that all >> transmitted chars on the device are properly transmitted. My problem is >that >> it seems like that the serial device is running half duplex. >> >> My source code looks like this: >> >> fd = open ("/tyCo/1", O_RDWR, 1); >> write (fd, "abcde", 5); >> i=0; >> while (i < 5) >> { >> i += read (fd, &rxBuffer[i], 1); >> } >> >> The result is that the while loop is newer completed because the read >> function only read the last 2 chars "de" in the string "abcde". >> >> Do anyone have a solution of an explanation of this problem? >> >> Best regards >> >> Jakob Holm-Petersen >> >> >> > > --------------------------- Newsgroups: comp.os.vxworks Subject: Use of DMA controller on PC486 board Date: Tue, 05 Feb 2002 14:34:50 GMT From: soren_abildgaard@hotmail.com (Sřren Abildgaard) Organization: TDC Internet Message-ID: <3c5fec9e.20110390@news.tele.dk> Basically how do I use the DMA comtroller to do a number of writes to memory ? Does anyone have som example code ? Maybe some public doc on the subject ? Regards, Sřren Abildgaard --------------------------- Newsgroups: comp.os.vxworks Subject: Wind Web Server question Date: Tue, 5 Feb 2002 16:54:26 +0200 From: "Yaron Cohen" Message-ID: <3c5ff2e7@news.012.net.il> Hi, I am learning Wind Web Server and I have two questions: 1. Where can I find the online "Wind Web Server Reference" ? 2. Is it possible to associate two user RPMs for the same page ? I would like the page to be treated both by the SSI RPM and by the function multiplexer RPM. Thanks, Yaron --------------------------- Newsgroups: comp.os.vxworks Subject: Re: pciFindDevice()- index 0 based or 1 based ? Date: Tue, 5 Feb 2002 09:11:55 +0100 From: "Michael Lawnick" Organization: Buergernetz Dillingen Message-ID: References: Reply-To: "Michael Lawnick" Sender: mlawnick@pd95245d7.dip.t-dialin.net Hi Arun, is the output the same for 0 and 1 ? On Walnut you may expect two devices: CPUtoPCI-Bridge and VGA-adapter ! HTH - -- Mit freundlichen Grüßen, Michael Lawnick ============================================== SOFTEC GmbH Tel +49-731-96600-0 Promenade 17 Fax +49-731-96600-23 D-89073 Ulm Michael Lawnick Germany lawnick@softec.de ============================================== "Arun Prasad" schrieb im Newsbeitrag news:e38d912c.0202040646.24fec1a6@posting.google.com... > Hi, > > I have a doubt on the index parameter of pciFindDevice funcion. > > I have only one PCI device in my system and I'm getting success if I > call the function with index value passed either 0 or 1. > > Can anyone tell me whether this index value starts with 0 or 1? > > I'm using IBM PPC 405GP. > > Thanks in advance. > Regards > Arun > > Arun Prasad > Software Engineer > Adaptec India Pvt. Limited > (A wholly owned subsidiary of Adaptec Inc., USA) > 6-3-1086, IV Floor, Vista Grand Towers, > Rajbhavan Road, Somajiguda, > Hydereabad - 500082 > Phone: (O):91-40-6661555/56/57/58 > (Ext):234 > Email: arun_prasad@adaptecin.com --------------------------- Newsgroups: comp.os.vxworks Subject: motFccEnd.c Date: 5 Feb 2002 00:22:45 -0800 From: alimohd@logiceastern.com (Ali) Organization: http://groups.google.com/ Message-ID: hi all We are using FCC2 and 3 for Ethernet devices with the END driver. I suppose number og people have done this. Can anybody give me a source code(motFccEnd.c) for this as i am not having it with me.What chabges i have to make in sysMotFccend.c and motFccEnd.h..... i will be greatly obliged to him. thanx in advance ali --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Initializing a semaphore Date: Tue, 5 Feb 2002 09:20:11 +0100 From: "Michael Lawnick" Organization: Buergernetz Dillingen Message-ID: References: <3C5F35CA.19B69434@bellsouth.net> Reply-To: "Michael Lawnick" Sender: mlawnick@pd95245d7.dip.t-dialin.net Hi joe, nitin, joe wrote: > Well, one way is: > > static int first =1; > ... > if(first==1) > { > first =0; > semBCreate() > > } > > and this is the first security hole in your application ! It should be: static int first =1; UINT32 state; ... if(first==1) { state=taskLock(); if(first==1) { first =0; semBCreate(); } taskUnlock(state); } as you don't use an atomic test-and-set for entry into the braces ! - -- Mit freundlichen Grüßen, Michael Lawnick ============================================== SOFTEC GmbH Tel +49-731-96600-0 Promenade 17 Fax +49-731-96600-23 D-89073 Ulm Michael Lawnick Germany lawnick@softec.de ============================================== "joe durusau" schrieb im Newsbeitrag news:3C5F35CA.19B69434@bellsouth.net... > > > nitin kumar wrote: > > > Hi, > > I have a structure of the following type .... > > > > static struct foo { > > foobar_t *ptr; > > SEMID sem; > > uint32_t refcnt; > > } > > > > I need to initialize this struct only once during the execution of my code. > > How > > can I do something like ... > > > > static struct foo { > > foobar_t *ptr; > > SEMID sem; > > uint32_t refcnt; > > } = { NULL, semBCreate(), 0 }; > > > > I know I cannot do a semBCreate() in the initializing value. How do I > > initialize the > > semaphore there ?...I remember pthreads having a PTHREAD_MUTEX_INITIALIZER > > ...thingie > > which I could put over there... > > > > The only way I can think of now is to have aanother pointer to struct foo, > > .. > > > > static struct foo { > > }..*foo_ptr = NULL; > > > > ANd then do somrething like.. > > > > if ( !foo_ptr ) { > > foo_ptr = malloc( sizeof( struct foo )); > > memset( foo_ptr, 0, sizeof( *foo_ptr )); > > foo_ptr->sem = semBCreate(); > > } > > > > Thanks all > > > > nitin > > Well, one way is: > > static int first =1; > ... > if(first==1) > { > first =0; > semBCreate() > > } > > > > The stuff between the first two braces will execute one and only one time, > barring reloading of the executable or rebooting. > > Speaking only for myself, > > Joe Durusau > > --------------------------- Newsgroups: comp.os.vxworks Subject: why is it so Date: 5 Feb 2002 00:28:13 -0800 From: alimohd@logiceastern.com (Ali) Organization: http://groups.google.com/ Message-ID: hi all When i build my bsp it gives me two peculiar errors at last can anybody tell me what it actually looks for i tried much to solve this but it comes again and again.Though the lines it is specifying has no entry like this.... here is snapshot of this build ccppc -B/home/alimohd/tornado/host/sun4-solaris2/lib/gcc-lib/ - -nostdlib -r -Wl,-X \ -Wl,@/home/alimohd/tornado/target/proj/ep8260boot/prjObjs.lst version.o /home/alimohd/tornado/target/lib/libPPCEC603gnuvx.a -o partialImage.o prjConfig.o: In function `usrBootLineInit': /home/alimohd/tornado/target/config/comps/src/net/usrBootLine.c:36: multiple definition of `endDevTbl' sysLib.o:/home/alimohd/tornado/target/config/ep8260/sysMotFccEnd.c:180: first defined here --------------------------- Newsgroups: comp.os.vxworks Subject: UDF 1.5 support for VxWorks Date: 5 Feb 2002 07:17:43 -0800 From: stampfl@ict.tuwien.ac.at (Norbert) Organization: http://groups.google.com/ Message-ID: Hello, Does anyone know if VxWorks supports the UDF1.5 or the ISO9660 file system. I would need it to read and write CD-RWs. Or does hav anyone experience in SW to burns CD-R/RWs in UDF1.5 format? Thanks in advance Norbert --------------------------- Newsgroups: comp.os.vxworks Subject: handling multiple interrupts Date: 5 Feb 2002 01:35:59 -0800 From: senthilb@comneti.com (senthil) Organization: http://groups.google.com/ Message-ID: <26ca8a8f.0202050135.72964a8b@posting.google.com> Hi, I have a doubt regarding multiple interrupt handling by VxWorks for MPC860. Consider a case a CPM interrupt at (level4 in siu) occurs. At that instant another interrupt is also raised from another source in the same CPM but with hight priority to the previous one. In such cases nested interrupts of multiple interrupts arises. While the first interrupt calls the ISR, it masks of the low priority interrupts but the higher priority interrupts will not be disabled. In such cases the mpc manual explain procedures like saving msr,PC in srr0,1 registers and recovering them back in case of nested interrupts. Does VxWorks handle nested interrupts in this manner or should we handle multiple interrupts by our own. This doubt is because our target board gets reset rarely by multiple interrupts. Please calrify me . Regards senthil --------------------------- Newsgroups: comp.os.vxworks Subject: Re: handling multiple interrupts Date: Tue, 5 Feb 2002 09:39:57 +0000 (UTC) From: David Laight Organization: BT Openworld Message-ID: <3C5FA8ED.49B8AEC2@l8s.co.uk> References: <26ca8a8f.0202050135.72964a8b@posting.google.com> senthil wrote: > > Hi, > I have a doubt regarding multiple interrupt handling by VxWorks for > MPC860. > Consider a case a CPM interrupt at (level4 in siu) occurs. At that > instant another interrupt is also raised from another source in the > same CPM but with hight priority to the previous one. In such cases > nested interrupts of multiple interrupts arises. While the first > interrupt calls the ISR, it masks of the low priority interrupts but > the higher priority interrupts will not be disabled. In such cases the > mpc manual explain procedures like saving msr,PC in srr0,1 registers > and recovering them back in case of nested interrupts. Does VxWorks > handle nested interrupts in this manner or should we handle multiple > interrupts by our own. This doubt is because our target board gets > reset rarely by multiple interrupts. I suspect that vxWorks gets nested interrupts right. However there may not be enough stack space for both interrupts. David --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Use of DMA controller on PC486 board Date: Tue, 5 Feb 2002 15:57:06 +0000 (UTC) From: David Laight Organization: BT Openworld Message-ID: <3C60014F.EE59A65C@l8s.co.uk> References: <3c5fec9e.20110390@news.tele.dk> "Sřren Abildgaard" wrote: > > Basically how do I use the DMA comtroller to > do a number of writes to memory ? The PC hardware has 2 intel 8257 (I think) DMA controllers [1] cascaded off each other. One does 16bit transfers, the other 8. In general they can only by used for transfers to/from IO devices that support DMA. [1] These days it is all embedded in the south? bridge, with a few non-standard registers to get wider addresses, but this can be ignored. David --------------------------- Newsgroups: comp.os.vxworks Subject: getproc_and_instance MIB Date: 5 Feb 2002 08:58:57 -0800 From: ilyasg@yahoo.com (Ilyas G) Organization: http://groups.google.com/ Message-ID: Hi after compiling the MIB, i had in the skeleton skel.c file a call to a function snmpdGroupByGetprocAndInstance , wich is defined as group_by_getprov_and_instance however, in the bootable project I have found the procedure snmpdGetByGroupprocAndInstance. this cause a probleme when i try to download the project , an error occurs saying that _group_by_getproc_and_instance doesn't exists !!!!!! does someone had had this kind of problem, and have the clue ? --------------------------- Newsgroups: comp.os.vxworks Subject: WindView.win32.tcl error in T2 Date: Tue, 5 Feb 2002 09:50:59 -0800 From: "kdb" Organization: cmc Message-ID: When I start up Tornado v2.0 I get the following error message: couldn't read file "TRY=localhost\host\resource\tcl\WindView.win32.tcl": no such file or directory When you click OK, Tornado continues its startup and can then be used. Using Tornado v2.0 on a Win2000 PC host, targeted to PPC 8245 (603). Had v2.0.2 and did not get this error; however, must use v2.0 for this project. WindView option does not appear in Tornado menu or toolbar (did in v2.0.2). The named TCL file is present, at C:\Tornado\host\resource\tcl\WindView.win32.tcl. Tried copying this to all permutations of the suggested pathname but still get the error. Pathname seems to be used in C:\Tornado\host\x86-win32\bin\WindView.dll, but is given there with a leading backslash (ie, \host\resource\...). Anybody know how to get around this error? Not only is it annoying, it would be nice to be able to use WindView. Just started reading this newsgroup; I've searched the FAQ and didn't find any mention of this. thx - K --------------------------- Newsgroups: comp.os.vxworks Subject: Re: CLI Parser for VxWorks Date: Tue, 5 Feb 2002 15:43:10 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: <182ee52a.0201251347.2a1d1421@posting.google.com> <9b6f8883.0201251727.781eea1a@posting.google.com> "Dave Korn" wrote in message news:oZz78.1617$5o.406163@newsr2.u-net.net... > "Mark Chang" wrote in message > news:9b6f8883.0201251727.781eea1a@posting.google.com... > > dpraveen@hotmail.com (Praveen Dulam) wrote in message > news:<182ee52a.0201251347.2a1d1421@posting.google.com>... > > > Hi > > > > > > I am looking for CLI commands parser for VxWorks. > > > > > > TIA > > > Praveen > > > > If you need it free - use TCL (we use 7.3 in our embedded box) > > or use one of WindRiver's RapidControl for CLI. > > Mark > > > > [assuming CLI refers to a command-line here] > > Other good options are the free source utilities lexx/yacc and bison, all And antlr. Forgot that one. DaveK - -- moderator of alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! Proud Member of the Exclusive "I have been plonked by Davee because he thinks I'm interesting" List Member # Master of Many Meowing Minions Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage above and beyond the call of hilarity. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: curses Date: Tue, 5 Feb 2002 15:52:27 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: "Dave Korn" wrote in message news:aJB58.39$5o.67861@newsr2.u-net.net... > "Weddick" wrote in message > news:KgqX7.12617$kK1.91125@rwcrnsc52... > > Has anyone done a port or know of a port for using curses on a X86 Vxworks > > System? > > > > Thanks, > > > Did you get an answer to this one yet? If not, do a google search for > "vxworks ncurses ftp.ucar.edu" and you should find it. > > > DaveK Or check Johan's post news: which gives the site as ftp://ftp.atd.ucar.edu/pub/archive/vxworks/vx/ DaveK - -- moderator of alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! Proud Member of the Exclusive "I have been plonked by Davee because he thinks I'm interesting" List Member # Master of Many Meowing Minions Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage above and beyond the call of hilarity. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: motFccEnd.c Date: Tue, 5 Feb 2002 11:09:03 -0800 From: "Dan Gold" Organization: Mailgate.ORG Server - http://www.Mailgate.ORG Message-ID: References: Sender: gold@ensemble.com "Ali" wrote in message news:b305b5bd.0202050022.3fc4cb0d@posting.google.com... > hi all > We are using FCC2 and 3 for Ethernet devices with the END driver. > I suppose number og people have done this. > Can anybody give me a source code(motFccEnd.c) for this as i am not > having it with me.What chabges i have to make in sysMotFccend.c and > motFccEnd.h..... > i will be greatly obliged to him. [Shades of one of George's more recent posts here... :-)] Ali, If you're using vxWorks (and I believe you are), the WRS 8260/FCC driver I believe already supports this. It may be a bit buggy, in case they haven't released the patch yet, but you should be able to obtain a correct version. If it's not fixed (or released yet), or you still want the source to the FCC driver, ask your WRS FAE. They give it to you, as they wrote it, and I think usually only release it in binary form. hth, Dan Gold gold@ensemble.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: UDF 1.5 support for VxWorks Date: Tue, 5 Feb 2002 21:24:17 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: AN ISO 9660 file system - cdromFsLib, was developed by myself and my team back in 1995, and now is integral part of VxWorks (since Tornado 2.0). This library however only deals with read-only disks, because seriously, ISO9660 is not a writable file system in the complete sense. We had a plan for UDF before I left Wind River, but I recon they never got around to do it since. I do remember seing a pretty lean ISO9660 mastering application in the open source some years back, which could be ported to VxWorks. There is a big difference between CD-ROM mastering and a truly writable file system, I hope you dig that. "Norbert" wrote in message news:f4669478.0202050717.3b424db7@posting.google.com... > Hello, > > Does anyone know if VxWorks supports the UDF1.5 or the ISO9660 file > system. I would need it to read and write CD-RWs. > > Or does hav anyone experience in SW to burns CD-R/RWs in UDF1.5 > format? > > Thanks in advance > Norbert --------------------------- Newsgroups: comp.os.vxworks Subject: Re: DHCP over PPP Date: 5 Feb 2002 12:12:59 -0800 From: jzhang@drs.ca (Jason) Organization: http://groups.google.com/ Message-ID: References: Hello Yusuf, I create the ppp link to the dhcp server and call the following codes to initialize the dhcp service: pIf = ifunit ("ppp0"); if (pIf == NULL) { printf ("Invalid device \"%s\"\n", netDev); return (ERROR); } pCookie = dhcpcInit (pIf, TRUE); if (pCookie == NULL) { printErrno(errnoGet()); printf ("Error initializing DHCP boot lease.\n"); return (ERROR); } /* Set any lease options here. */ result = dhcpcOptionSet (pCookie,_DHCP_SUBNET_MASK_TAG); if (result != OK) { printErrno(errnoGet()); printf( "dhcpcOptionSet Failed \n"); return (ERROR); } if (dhcpcBind (pCookie, FALSE) != OK) { printErrno(errnoGet()); printf ("dhcpcBind Failed.\n"); pCookie = NULL; return (ERROR); } return (OK); And then the ppp interface went down. I set the interface flag of ppp to be the broadcast enable (which is mentioned in network programmer's guide), but got the same result. It looks like that the dhcpcBind() checks the attributes of the interface and only support a Ethernet-like port. Do you know how to make ppp looks like Ethernet to make dhcp client support it ? Any comments will be greatly appreciated. Regards. "Yusuf Motiwala" wrote in message news:... > You mean client right. Why do you need to implement, you can use WRS > supplied DHCP client. However, by default it uses boot device to get DHCP > lease which you need to change. > > Yusuf > > > "Jason" wrote in message > news:e555a185.0201141146.5099300b@posting.google.com... > > I am going to implement the DHCP over PPP. By default, the vxWorks > > only supports DHCP over Ethernet. It would be greatly appreciated if > > anyone can give me some ideas or sample codes about DHCP over PPP. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: DHCP over PPP Date: 05 Feb 2002 15:21:48 -0500 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: References: Sender: bpringlemeir@DeadDuck >>>>> "Jason" == Jason writes: Jason> Hello Yusuf, I create the ppp link to the dhcp server and call Jason> the following codes to initialize the dhcp service: Jason> pIf = ifunit ("ppp0"); [etc...] In an attempt to save you some time, or perhaps inject some orthadox conformity, you can not do Bootp and DHCP over PPP. You do not have a MAC address. The code you posted will attempt to setup the PPP interface. The interface is probably already setup with the PPP parameters, PPP_OPT_DEFAULT_ROUTE, OPT_DEFAULT_ROUTE, default_route PPP_OPT_IPCP_ACCEPT_LOCAL, OPT_IPCP_ACCEPT_LOCAL, ipcp_accept_local PPP_OPT_IPCP_ACCEPT_REMOTE, OPT_IPCP_ACCEPT_REMOTE,ipcp_accept_remote IPCP is the usual method of setting up PPP. The IP addresses are not really that important with PPP as you have two point network on that interface. If you need something else, it is probably addressed by another standard. I have no answer to your questions. I just wonder "Why you are trying to do it?" hth, Bill Pringlemeir. - -- class struggle munitions cryptographic $400 million in gold bullion Legion of Doom genetic Honduras Ft. Bragg terrorist quiche jihad strategic Cocaine fissionable Ft. Meade vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" --------------------------- Newsgroups: comp.os.vxworks Subject: Symbol table synchronization issue (Tornado 2/PPC/Solaris) Date: 5 Feb 2002 20:55:22 GMT From: sj@shell1.aracnet.com (Scott Johnson) Organization: aracnet.com -- Portland's loudest electrons (info@aracnet.com) Message-ID: Have an interesting symbol table synchronization issue. Target is built in two parts; a "kernel" which has vxWorks itself and several libraries; and an application which has the useful code. :) When kernel boots, it uses loadModule() to load application into memory. All is well except when trying to debug; Crosswind is able to see kernel symbols, but NOT application symbols. Kernel is vxWorks.st type; and has its own symbol table. Kernel IS built with INCLUDE_SYM_TBL_SYNC (or whatever); this appears to be working. When I launch a target server, I see the following: Connecting to target agent... succeeded. Attaching C++ interface... succeeded. Attaching elf OMF reader for PPC CPU family... succeeded. Added target_modules to target-server.....done No error messages present anywhere. If I use ld (on the target; or @ld from windsh) to load an object file AFTER starting the target server, it works fine. However, the object loaded automatically on boot (but loaded bEFORE I start the target server) does not get its symbols sync'd properly. As an example, there is a symbol in the application file called "instSetMode". Typing from windsh: - -> lkup "instSetMode" value = 0 = 0x0 - -> @lkup "instSetMode" instSetModeCursorType 0x006bb2c4 text (application) instSetMode 0x006a28cc text (application) instSetModeCursorActive 0x006bb704 text (application) value = 0 = 0x0 Any ideas? The host is Solaris, and the problem is not incorrect registry configuration (a common problem with Windows hosts, it seems.) I only have problems with the module loaded onto the target BEFORE connecting the target server. Target server is invoked thusly: tgtsvr amdam10 -A -V -s -n amdam10 -m 4000000 -c A[horrendous pathname to vxWorks.st file ] -RW engineer_scotty - -- - ------------------------------------------------------------------------ engineer_scotty (no, not that one) -- consumer of fine ales and lagers some days you're the fire hydrant, some days you're the dog | go blazers no small furry creatures were harmed in the creation of this .signature --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Symbol table synchronization issue (Tornado 2/PPC/Solaris) Date: 05 Feb 2002 16:42:37 -0500 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: References: Sender: bpringlemeir@DeadDuck >>>>> "Scott" == Scott Johnson writes: Scott> Have an interesting symbol table synchronization issue. ^^^^^^^^^^^^^^^ Use the source.... Scott> Target is built in two parts; a "kernel" which has vxWorks Scott> itself and several libraries; and an application which has the Scott> useful code. :) When kernel boots, it uses loadModule() to Scott> load application into memory. All is well except when trying Scott> to debug; Crosswind is able to see kernel symbols, but NOT Scott> application symbols. Kernel is vxWorks.st type; and has its Scott> own symbol table. Are you sure "has its own symbol table" is correct? Is a symTbl.c linked with the image? I think that you might need the standalone symbol table. I have done what you are talking about, although I nixed the synchronization to the host. Sometimes even if the symbol table is linked, it isn't used! I had to goof around with usrConfig.c, do lots of tricks with config.h, and create a custom Makefile. A standalone symbol table with networking support doesn't seem to be a vxWorks option. However, you need this to load an application without the network (and tools) and you need the network to debug the same image. I believe that WRS has postulated that you would have either (NOT OR, Ie. eXclusive OR). Anyways if you have a shell on a console, you can verify that something is rotten in Alameda. The shell will show what is happening when a network is not connected. If you don't have a shell, you'll have to take my word on it. hth, Bill Pringlemeir. - -- Only a ignorant beast like you would say that most Polish people have computers . Let me tell you something, you rat, all MicroVAX II-loving dudes should get system calls . vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado2 and PVCS Date: Tue, 5 Feb 2002 22:44:40 +0100 From: "Martin Raabe" Organization: T-Online Message-ID: References: <3c587646$1@brateggebdc5.br-automation.co.at> Dear David, Visual Source Safe does NOT support branching a file in the meaning other leading version control tools do this. Visual Source Safe allows "branching" only, if you share a file in another project (and therefor in another directory) and then branch the file. This is simple not acceptable. A branch in the meaning of Version Control Tools (RCS, CVS, PVCS, Continuus, ClearCase ....) is a logical state, which is stored inside the corresponding version control database / repository file. It has absolutly nothing to do, that the file have to be relocated into another directory. A branch is a logical administration structure of the version control tool, not a simple copy of the file. Therefor, the information inside the Visual Source Safe integration documentation is probably a little bit unprecise, but at the end correct. As ever, a branch is for Microsoft something different as for everyone else. VSS simple do not support branches in the general meaning of the other version control tools and therefor SNiFF+ do not support the VSS way of branching. Ciao Martin Raabe "Dave Korn" schrieb im Newsbeitrag news:nZz78.1616$5o.406163@newsr2.u-net.net... > "Martin Raabe" wrote in message > news:a3asq2$j3r$07$1@news.t-online.com... > > > Also MS Source Safe doe not support branches, which might be changes being > > - customer speciffic > > - bug fixing > > - new feature testing > > - etc. > > In case anyone has been fooled by this statement, it is of course garbage. > MS SourceSafe fully supports branching and merging. > > Martin, does your management know you are using a WRS email address to > post potentially libellous mis-statements in public about a competitor's > product ? Do they approve ? How can you even imagine for one minute that > SourceSafe doesn't have branching ? That would be an incredibly vital > feature to omit from *any* source control software. > > > I mentioned SNiFF+ here, because I don't do any development without it, > > is it doen with Tornado or pRISM+ or Visual Studio. > > Not to forget the fact that you *always* plug SNiFF+ in here. > > > Just my two new Euro Cents. > > Is that how much commission you get on each sale ? > > DaveK > -- > moderator of > alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow > Burn your ID card! http://www.optional-identity.org.uk/ > Help support the campaign, copy this into your .sig! > Proud Member of the Exclusive "I have been plonked by Davee because he > thinks I'm interesting" List Member # > Master of Many Meowing Minions > Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage above > and beyond the call of hilarity. > > --------------------------- Newsgroups: comp.os.vxworks Subject: Serial Driver Lockup? Date: Tue, 05 Feb 2002 22:40:49 GMT From: "Chad" Organization: Excite@Home - The Leader in Broadband http://home.com/faster Message-ID: I have 2 tasks running at equal priority and actively listening to serial 485 ports. Eventually one of the listeners gets stuck on a semaphore: 2b2d8c write__2IOPCvi +d4 : write ([]) 166e34 write +10 : iosWrite ([]) 168398 iosWrite +cc : tyWrite ([]) 19ae0c tyWrite +5c : semTake ([]) 18c288 semTake +134: semBTake ([]) The serial driver is still not finalized and has a few potential issues: 1 - The driver does some internal timing to manage half duplex states 2 - The uart is in character mode so the interrupt traffic is currently quite heavy I'm not overly familiar with drivers, so can anybody suggest where I might start? Why would there be contention for a write semaphore between two separate devices? This must be whats happening as each port functions fine individually. Any advice welcome, Chad --------------------------- Newsgroups: comp.os.vxworks Subject: tornado shell, tasks, printf not working Date: 5 Feb 2002 14:59:52 -0800 From: bdutta@hotmail.com (Banibrata Dutta) Organization: http://groups.google.com/ Message-ID: hi, i am new to vxwork programming. i've a problem with printf()s. my main task creates 2 more tasks that call printf(), scanf(). the problem is once the task is created, neither of the tasks seem to output anything in their printf()s. i am running this on VxSIM on NT machine, thru the Tornado Prototyper. my hunch is that the other task's I/O is somehow not redirected to the WindSh shell. can someone tell me what exactly is going on. thanks, bdutta. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: pciFindDevice()- index 0 based or 1 based ? Date: 5 Feb 2002 15:42:01 -0800 From: arun4vxworks@indiatimes.com (Arun Prasad) Organization: http://groups.google.com/ Message-ID: References: Hi, When I use pciDeviceShow() it gives 3 entires as output like this, Scanning function 0 of each PCI device on bus 0 Using configuration mechanism 0 bus device function vendorID deviceID class 00000000 00000000 00000000 00001014 00000156 00060000 00000000 00000002 00000000 0000143a 00004e53 00028000 00000000 00000004 00000000 00005333 00008a01 00030000 The middle entry is for our card. And I am able to find our card using pciFindDevice() by passing our device ID and vendor ID. However I'm failing to find the remaining two devices by giving the deive and Vendor IDs shown by the pciDeviceShow(). But I left it as its not necessary for me and I concern only about our Card. My doubt is, I'm calling pciFindDevice twice with our card's device & Vendor ID and the index as 0 and 1. For both the calls I'm getting success and while scanning, Device number varies, IRQ no. remains the same, Device ID and Vendor ID same. But the remaining values of our card specific are incorrect for index argument 1. Where as pciHeaderShow() for both these are same. Since the I'm getting success for both 0 and 1 my loop to find our devices getting executed twice and the device array also filled with two entries and the routine to find no. of device returns 2. Regards Arun - ----- Arun Prasad Software Engineer Adaptec India Pvt. Limited (A wholly owned subsidiary of Adaptec Inc., USA) 6-3-1086, IV Floor, Vista Grand Towers, Rajbhavan Road, Somajiguda, Hydereabad - 500082 Phone: (O):91-40-6661555/56/57/58 (Ext):234 Email: arun_prasad@adaptecin.com Johan Borkhuis wrote in message news:... > arun4vxworks@indiatimes.com (Arun Prasad) wrote: > > > Hi, > > > > I have a doubt on the index parameter of pciFindDevice funcion. > > > > I have only one PCI device in my system and I'm getting success if I > > call the function with index value passed either 0 or 1. > > > > Can anyone tell me whether this index value starts with 0 or 1? > > As far as I know the PCI bridge has ID 0, and your device has ID 1. Check > the device ID's that you are getting back from pciDeviceShow and > pciHeaderShow. > > Groeten, > Johan --------------------------- Newsgroups: comp.os.vxworks Subject: Time Base register update rate Date: 5 Feb 2002 18:12:54 -0800 From: thongMb359@hotmail.com (Thong Huynh) Organization: http://groups.google.com/ Message-ID: Hi, I have a PPC750 card. I am looking for a way to find out the update rate of the Time Base register (TBU and TBL). Any pointer are greatly appreciated. Thanks, Thong. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Time Base register update rate Date: Wed, 06 Feb 2002 05:14:29 GMT From: pkockritz@mac.com (Pete Kockritz) Organization: AT&T Broadband Message-ID: References: In article , thongMb359@hotmail.com (Thong Huynh) wrote: > Hi, > > I have a PPC750 card. I am looking for a way to find out the update rate > of the Time Base register (TBU and TBL). Any pointer are greatly appreciated. > > Thanks, > > Thong. IIRC, on most PPCs the TB register is clocked at 1/4 the rate of external bus. A simple way to get a ballpark value is to write a simple function to read the timebase register, taskDelay for a period of time (e.g., 1 or 10 seconds) and read the TB again. It won't be exact, since the actual taskDelay time will vary: void findTbRate (int seconds) { UINT32 startTbl, startTbu, endTbl, endTbu; if (seconds <= 0) seconds = 1; taskDelay( 1 ); vxTimeBaseGet( &startTbu, &startTbl ); taskDelay( sysClkRateGet() * seconds ); vxTimeBaseGet( &endTbu, &endTbl ); printf ( "delta is TBU: %08x TBL: %08x\n", endTbu - startTbu, endTbl - startTbl ); } - -- Regards, Pete Kockritz --------------------------- Newsgroups: comp.os.vxworks Subject: Re: motFccEnd.c Date: 5 Feb 2002 21:28:54 -0800 From: alimohd@logiceastern.com (Ali) Organization: http://groups.google.com/ Message-ID: References: "Dan Gold" wrote in message news:... > "Ali" wrote in message > news:b305b5bd.0202050022.3fc4cb0d@posting.google.com... > > hi all > > We are using FCC2 and 3 for Ethernet devices with the END driver. > > I suppose number og people have done this. > > Can anybody give me a source code(motFccEnd.c) for this as i am not > > having it with me.What chabges i have to make in sysMotFccend.c and > > motFccEnd.h..... > > i will be greatly obliged to him. > > [Shades of one of George's more recent posts here... :-)] > > Ali, > > If you're using vxWorks (and I believe you are), the WRS 8260/FCC driver I > believe already supports this. It may be a bit buggy, in case they haven't > released the patch yet, but you should be able to obtain a correct version. > If it's not fixed (or released yet), or you still want the source to the FCC > driver, ask your WRS FAE. They give it to you, as they wrote it, and I > think usually only release it in binary form. > > hth, > Dan Gold gold@ensemble.com Hi Dan Thanx for ur reply yes i have the source files for sysmotfccend.c and configNet.h but the situation there is only for one instance of end driver can u pls tell me what extra changes i have to make to these two files so that i can support two fcc's for ethernet and that two both of them in full duplex mode. waiting for the reply thanx ali --------------------------- Newsgroups: comp.os.vxworks Subject: about lex Date: 5 Feb 2002 22:34:53 -0800 From: cwnam@korealink.ne.kr (changwoo) Organization: http://groups.google.com/ Message-ID: <3d64ee1a.0202052234.26dc4574@posting.google.com> I am tired to use lex in vxWorks but there is some problem to compile the files. First I check the source file and successed in Unix environment. with the same source, In vxWorks, first I compiled it with lex.(lex sampel.l) after then outfile was created "lex.yy.c" and I include the file to workspace and compiled, but there was error "inttypes.h : no such file" showed. how can I fix this. And what I have to do other thing to use "lex" in vxworks ? Help this situation... thanks --------------------------- Newsgroups: comp.os.vxworks Subject: Re: why is it so Date: Wed, 06 Feb 2002 07:20:12 GMT From: "drdiags" Organization: AT&T Broadband Message-ID: References: Ali, The endDevTbl[] array exists in the configNet.h file, here is an example: END_TBL_ENTRY endDevTbl [] = { { 0, WIN_LOAD_FUNC, WIN_LOAD_STRING, WIN_BUFF_LOAN, NULL, FALSE}, { 0, END_TBL_END, NULL, 0, NULL, FALSE}, }; Each BSP defines this differently. All other references to it should be like so: pcClient\sysDec21x4xEnd.c(161): IMPORT END_TBL_ENTRY endDevTbl[]; pcClient\sysDec21x4xEnd.c(214): endDevTbl[0].endLoadString = endLoadStr; pcClient\sysNetif.c(1224): IMPORT END_TBL_ENTRY endDevTbl[]; /* end device table */ So as Michael states, your BSP defines this and since you are building a vxWorks kernel (prjConfig.c created in Tornado IDE), look for the usrBootLine.c file in your %WIND_BASE%/target/config/comps/src/net directory and see if this array has been declared again. Don't know your BSP, but that shouldn't matter because AFAIK (always wanted to do that), it is the one that creates the instance of the array. My copy of the file usrBootLine.c doesn't contain this reference. Did someone modify this? "Ali" wrote in message news:b305b5bd.0202052152.8a2cb7c@posting.google.com... > "Michael Lawnick" wrote in message news:... > > Hi Ali, > > > > just as your output says: > > multiple definition of `endDevTbl' > > first in sysMotFccEnd.c, included by sysLib.c, > > scnd in usrBootLine.c, included by prjConfig.c > > AFAIK it is an array of structures. You should have only one defined, so > > making them local won't solve your problem. > > > > -- > > Mit freundlichen Grüßen, > > Michael Lawnick > > ============================================== > > SOFTEC GmbH Tel +49-731-96600-0 > > Promenade 17 Fax +49-731-96600-23 > > D-89073 Ulm Michael Lawnick > > Germany lawnick@softec.de > > ============================================== > > > hi Michel > thanx for suggestion but im wondering as there is no such definition > in usrBootLine.c or in sysMotFccEnd.c and the line numbers its > giving does not match at all > can u tell me exactly what is this. > this array is not defined in any one of these files but ofcourse it is > defined in configNet.h only. > thanx in advance > ali > > > "Ali" schrieb im Newsbeitrag > > news:b305b5bd.0202050028.4252dd51@posting.google.com... > > > hi all > > > When i build my bsp it gives me two peculiar errors at last can > > > anybody tell me what it actually looks for i tried much to solve this > > > but it comes again and again.Though the lines it is specifying has no > > > entry like this.... > > > here is snapshot of this build > > > > > > ccppc -B/home/alimohd/tornado/host/sun4-solaris2/lib/gcc-lib/ > > > -nostdlib -r -Wl,-X \ > > > -Wl,@/home/alimohd/tornado/target/proj/ep8260boot/prjObjs.lst > > > version.o /home/alimohd/tornado/target/lib/libPPCEC603gnuvx.a -o > > > partialImage.o > > > prjConfig.o: In function `usrBootLineInit': > > > /home/alimohd/tornado/target/config/comps/src/net/usrBootLine.c:36: > > > multiple definition of `endDevTbl' > > > sysLib.o:/home/alimohd/tornado/target/config/ep8260/sysMotFccEnd.c:180: > > > first defined here --------------------------- Newsgroups: comp.os.vxworks Subject: Intermittent reboots Date: 6 Feb 2002 07:24:51 GMT From: Denis Perelyubskiy Message-ID: hello, we have an x86-based target. once in a while, we experience a reboot, but no exception information is printed, our exception handlers are not called, we're sure its not us who call the reboot function, etc. while i realize that this is a broad question, can someone suggest the possible areas to look for the cause of reboot? one is a watchdog timer, since this is just hardware, which would not print any info upon reboot. i do believe we have an MMU, and protect page zero, where an interrupt table is located, so we'd get an exception if we were to write there. somehow, it seems that even if we have stack overruns (which we checked for), we'd eventually dereference something we should not have, and get an exception...but could anyone think of what could cause a quiet reboot other then watchdog? any ideas, no matter how far-fetched you think they are, would be greatly appreciated. thanks, denis - -- if you want to reply by mail, please tweak my address in a reasonable way. sorry for the inconvenience. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: WindView.win32.tcl error in T2 Date: Wed, 06 Feb 2002 07:26:09 GMT From: "drdiags" Organization: AT&T Broadband Message-ID: References: KDB, I think there is an SPR Patch on WindSurf for patching Tornado 2.0 to handle this problem, which is related to the Win2K Host. The patch became part of the T2.0.2 release, that is why you didn't see this problem then. HTH. "kdb" wrote in message news:jYU78.32$84.3822@newsfeed.avtel.net... > When I start up Tornado v2.0 I get the following error message: > > couldn't read file "TRY=localhost\host\resource\tcl\WindView.win32.tcl": > no such file or directory > > When you click OK, Tornado continues its startup and can then be used. > Using Tornado v2.0 on a Win2000 PC host, targeted to PPC 8245 (603). Had > v2.0.2 and did not get this error; however, must use v2.0 for this project. > WindView option does not appear in Tornado menu or toolbar (did in v2.0.2). > The named TCL file is present, at > C:\Tornado\host\resource\tcl\WindView.win32.tcl. Tried copying this to all > permutations of the suggested pathname but still get the error. Pathname > seems to be used in C:\Tornado\host\x86-win32\bin\WindView.dll, but is given > there with a leading backslash (ie, \host\resource\...). > > Anybody know how to get around this error? Not only is it annoying, it > would be nice to be able to use WindView. Just started reading this > newsgroup; I've searched the FAQ and didn't find any mention of this. > > thx - K > > > --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Wed Feb 6 18:42:06 2002 From: Ernie Rael Date: Wed Feb 6 18:42:08 PST 2002 Subject: MPC860 floating point emulation We use both MPC860 and PPC750. I'd like to compile only for PPC750, and when running on MPC860 handle the software emulation exception to implement floating point. There isn't much floating point; performance isn't much of an issue. Does anyone know of a package that handles this. I can do any glue if it doesn't hook directly into a vxworks exception. I haven't looked at the gnu source for the soft-float stuff, but I'm guessing that it would take a bit of work to adapt it. FWIW, I'm using tornado 2, vxWorks 5.4. -ernie From vxwexplo-errs@csg.lbl.gov Thu Feb 7 01:36:11 2002 From: ulf.fiederling@dornier.eads.net Date: Thu Feb 7 01:36:13 PST 2002 Subject: How to get special Preprocessor output Hi all We are using Tornado 2 on an NT Host with a Pentium Target. Is there any possability to get the output of the preprocessor for our sources or for the whole project substituting only the macros used, and not !!! the #includes? Thanks in advance Ulf Fiederling From vxwexplo-errs@csg.lbl.gov Thu Feb 7 04:03:16 2002 From: Vxworks Exploder Date: Thu Feb 7 04:03:18 PST 2002 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Thu Feb 7 04:03:11 PST 2002 Subject: Re: Symbol table synchronization issue (Tornado 2/PPC/Solaris) Subject: RPC on VxWorks Subject: 5.3.1 --> 5.4 Subject: Re: Symbol table synchronization issue (Tornado 2/PPC/Solaris) Subject: Re: Intermittent reboots Subject: Re: Intermittent reboots Subject: Re: Freebie VxWorks Tornado? Subject: Re: motFccEnd.c Subject: Re: Very important, I need the Login/password of my Hi-Focus ADSL modem (ECI) Subject: Re: Tornado-2.1/mips and cp7000 BSP (Ocelot MIPS board) Subject: Re: DOS key like features in vxworks Subject: Re: PCI Device Base Address Subject: problem compiling curses with T2/vxworks5.4 Subject: Re: How to change shell's default directory? Subject: Re: Returning memory Subject: DOS key with timeouts, idea? Subject: i have downloaded the tornado 2.0 protoyper, but the key as not been emailed to me Subject: Re: frequent rebooting for an ethernet interface looses connectivity?? Subject: Re: problem compiling curses with T2/vxworks5.4 Subject: Re: problem with flash that contains dosFs and bootload Subject: Re: problem with flash that contains dosFs and bootload Subject: Re: DOS key with timeouts, idea? Subject: Re: Intermittent reboots Subject: No echo when using gets() with a WindShell Subject: Re: DOS key with timeouts, idea? followup using Select() Subject: Where is the Setup program?! Subject: Re: problem with flash that contains dosFs and bootload Subject: Re: VxWorks & RS422 Subject: Re: VxWorks & RS422 Subject: how i build bootrom ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Symbol table synchronization issue (Tornado 2/PPC/Solaris) Date: Wed, 6 Feb 2002 16:06:57 +0100 From: "Michael Lawnick" Organization: Buergernetz Dillingen Message-ID: References: Reply-To: "Michael Lawnick" Sender: mlawnick@p5088f3f8.dip.t-dialin.net Hi Scott, as far as I remeber, there was a SPR about that. It was that the first object loaded wasn't included in sync, but next should be. If there isn't a patch, you could load a little dummy .o before your app ... HTH - -- Mit freundlichen Grüßen, Michael Lawnick ============================================== SOFTEC GmbH Tel +49-731-96600-0 Promenade 17 Fax +49-731-96600-23 D-89073 Ulm Michael Lawnick Germany lawnick@softec.de ============================================== "Scott Johnson" schrieb im Newsbeitrag news:a3pgrq016a9@enews2.newsguy.com... > > Have an interesting symbol table synchronization issue. > > Target is built in two parts; a "kernel" which has vxWorks itself and > several libraries; and an application which has the useful code. :) When > kernel boots, it uses loadModule() to load application into memory. All > is well except when trying to debug; Crosswind is able to see kernel > symbols, but NOT application symbols. Kernel is vxWorks.st type; and has > its own symbol table. > > Kernel IS built with INCLUDE_SYM_TBL_SYNC (or whatever); this appears to > be working. When I launch a target server, I see the following: > > Connecting to target agent... succeeded. > Attaching C++ interface... succeeded. > Attaching elf OMF reader for PPC CPU family... succeeded. > Added target_modules to target-server.....done > > > No error messages present anywhere. If I use ld (on the target; or @ld > from windsh) to load an object file AFTER starting the target server, it > works fine. However, the object loaded automatically on boot (but loaded > bEFORE I start the target server) does not get its symbols sync'd > properly. > > As an example, there is a symbol in the application file called > "instSetMode". Typing from windsh: > > -> lkup "instSetMode" > value = 0 = 0x0 > -> @lkup "instSetMode" > instSetModeCursorType 0x006bb2c4 text (application) > instSetMode 0x006a28cc text (application) > instSetModeCursorActive 0x006bb704 text (application) > value = 0 = 0x0 > > Any ideas? The host is Solaris, and the problem is not incorrect registry > configuration (a common problem with Windows hosts, it seems.) I only > have problems with the module loaded onto the target BEFORE connecting the > target server. > > Target server is invoked thusly: > > tgtsvr amdam10 -A -V -s -n amdam10 -m 4000000 -c > A[horrendous pathname to vxWorks.st file ] -RW > > engineer_scotty > > -- > ------------------------------------------------------------------------ > engineer_scotty (no, not that one) -- consumer of fine ales and lagers > some days you're the fire hydrant, some days you're the dog | go blazers > no small furry creatures were harmed in the creation of this .signature --------------------------- Newsgroups: comp.os.vxworks Subject: RPC on VxWorks Date: 6 Feb 2002 07:31:14 -0800 From: hding@fwion.com (Roger Ding) Organization: http://groups.google.com/ Message-ID: Hi, I have problem to run RPC on VxWorks. First I tried run RPC Server on VxWorks and RPC Client on Linux, it is OK. Then I tried to run RPC Client also on VxWorks. But it crashs at clnt_create(). See below: ==================================================== program Exception current instruction address: 0x00001d80 Machine Status Register: 0x00089032 Condition Register: 0x44400022 11e070 vxTaskEntry +5c : shell () 1a56b8 shell +180: 1a56e4 () 1a58f4 shell +3bc: execute () 1a5a74 execute +d4 : yyparse () 1e95c4 yyparse +790: 1e7580 () 1e76f0 yystart +878: fw_rpcClient (61, 5) ca6f14 fw_rpcClient +3c : clnt_create () 1a78c0 clnt_create +114: clnttcp_create () 1a80ac clnttcp_create +98 : pmap_getport () 1a9748 pmap_getport +f4 : 1d80 () shell restarted. ==================================================== Following is the Source code: I call fw_rpcClient from shell to invoke RPC call. Thanks, Roger /****************************************************************************** File: ioss_rpc_clnt_main.c ******************************************************************************/ #include #include "ioss_rpc_api_vxworks.h" int fw_rpcClient(char *server, int choice) { CLIENT *cl; long *result_1; long *result_2; XCONN_INFO *result_3; XCONN_LIST *result_4; CHAN_LIST *result_5; XCONN_INFO xConnInfo; long chanId; cl = clnt_create(server, IOSS_RPC_API_PROG, IOSS_RPC_API_VERS, "tcp"); if (cl == (CLIENT *)NULL) { clnt_pcreateerror(server); exit(16); } chanId = 0; switch (choice) { case 1: { printf("Input the XCONN_INFO info :\n"); printf("InChanIdA = "); scanf("%u", &(xConnInfo.InChanIdA)); printf("EgChanIdA = "); scanf("%u", &(xConnInfo.EgChanIdA)); printf("InChanIdB = "); scanf("%u", &(xConnInfo.InChanIdB)); printf("EgChanIdB = "); scanf("%u", &(xConnInfo.EgChanIdB)); printf("WavelengthA = "); scanf("%u", &(xConnInfo.WavelengthA)); printf("WavelengthB = "); scanf("%u", &(xConnInfo.WavelengthB)); printf("ConnectionType = "); scanf("%u", &(xConnInfo.ConnectionType)); printf("OperStatus = "); scanf("%u", &(xConnInfo.OperStatus)); /* PrintOutXConnInfo(&xConnInfo); */ result_1 = ioss_setup_cross_connect_1(&xConnInfo, cl); if (result_1 == (long *)NULL) { clnt_perror(cl, server); exit(16); } if (*result_1 == 0) { fprintf(stderr, "ioss_setup_cross_connect_1 fail\n"); exit(16); } printf("ioss_setup_cross_connect_1 success\n"); break; } case 2: { printf("Input the XCONN_INFO info :\n"); printf("InChanIdA = "); scanf("%u", &(xConnInfo.InChanIdA)); printf("EgChanIdA = "); scanf("%u", &(xConnInfo.EgChanIdA)); printf("InChanIdB = "); scanf("%u", &(xConnInfo.InChanIdB)); printf("EgChanIdB = "); scanf("%u", &(xConnInfo.EgChanIdB)); printf("WavelengthA = "); scanf("%u", &(xConnInfo.WavelengthA)); printf("WavelengthB = "); scanf("%u", &(xConnInfo.WavelengthB)); printf("ConnectionType = "); scanf("%u", &(xConnInfo.ConnectionType)); printf("OperStatus = "); scanf("%u", &(xConnInfo.OperStatus)); /* PrintOutXConnInfo(&xConnInfo); */ result_2 = ioss_remove_cross_connect_1(&xConnInfo, cl); if (result_2 == (long *)NULL) { clnt_perror(cl, server); exit(16); } if (*result_2 == 0) { fprintf(stderr, "ioss_remove_cross_connect_1 fail\n"); exit(16); } printf("ioss_remove_cross_connect_1 success\n"); break; } case 3: { printf("Input the XCONN_INFO info :\n"); printf("InChanIdA = "); scanf("%u", &(xConnInfo.InChanIdA)); printf("EgChanIdA = "); scanf("%u", &(xConnInfo.EgChanIdA)); printf("InChanIdB = "); scanf("%u", &(xConnInfo.InChanIdB)); printf("EgChanIdB = "); scanf("%u", &(xConnInfo.EgChanIdB)); printf("WavelengthA = "); scanf("%u", &(xConnInfo.WavelengthA)); printf("WavelengthB = "); scanf("%u", &(xConnInfo.WavelengthB)); printf("ConnectionType = "); scanf("%u", &(xConnInfo.ConnectionType)); printf("OperStatus = "); scanf("%u", &(xConnInfo.OperStatus)); /* PrintOutXConnInfo(&xConnInfo); */ result_3 = ioss_query_cross_connect_1(&xConnInfo, cl); if (result_3 == (XCONN_INFO *)NULL) { clnt_perror(cl, server); exit(16); } printf("ioss_query_cross_connect_1 return with following result:\n"); PrintOutXConnInfo(result_3); break; } case 4: { printf("Input the XCONN_INFO info :\n"); printf("InChanIdA = "); scanf("%u", &(xConnInfo.InChanIdA)); printf("EgChanIdA = "); scanf("%u", &(xConnInfo.EgChanIdA)); printf("InChanIdB = "); scanf("%u", &(xConnInfo.InChanIdB)); printf("EgChanIdB = "); scanf("%u", &(xConnInfo.EgChanIdB)); printf("WavelengthA = "); scanf("%u", &(xConnInfo.WavelengthA)); printf("WavelengthB = "); scanf("%u", &(xConnInfo.WavelengthB)); printf("ConnectionType = "); scanf("%u", &(xConnInfo.ConnectionType)); printf("OperStatus = "); scanf("%u", &(xConnInfo.OperStatus)); /* PrintOutXConnInfo(&xConnInfo); */ result_4 = ioss_retrieve_cross_connect_list_1(&xConnInfo, cl); if (result_4 == (XCONN_LIST *)NULL) { clnt_perror(cl, server); exit(16); } if (result_4->noOfConnections == 0) { fprintf(stderr, "ioss_retrieve_cross_connect_list_1 fail\n"); exit(16); } printf("ioss_retrieve_cross_connect_list_1 success, return following info :\n"); PrintOutXConnList(result_4); break; } case 5: { int i; printf("Input the channel id info :\n"); printf("chanId = "); scanf("%u", &chanId); result_5 = ioss_retrieve_channel_list_1(&chanId, cl); if (result_5 == (CHAN_LIST *)NULL) { clnt_perror(cl, server); exit(16); } if (result_5->NoOfChannels == 0) { fprintf(stderr, "ioss_retrieve_channel_list_1 fail\n"); exit(16); } printf("ioss_retrieve_channel_list_1 success\n"); printf("NoOfChannels = %u\n", result_5->NoOfChannels); for (i = 0; i < result_5->NoOfChannels; i++) { printf("Channel %u : ChanId = %u, OperStatus = %u\n", i, result_5->ChanId[i], result_5->OperStatus[i]); } break; } case 0: { exit(16); break; } default: { break; } } return 0; } --------------------------- Newsgroups: comp.os.vxworks Subject: 5.3.1 --> 5.4 Date: Wed, 6 Feb 2002 10:41:21 -0500 From: "nitin kumar" Organization: Marconi Message-ID: Hi, We have a BSP using VxWorks 5.3.1. Has been working fine for ages and we are OK with it. I need to investigate the advantages we could get by upgrading to VxWorks 5.4.1. Is there some place I could get the new feature set introduced by 5.4.1 and see if we need them ? thanks a lot in advance. nitin --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Symbol table synchronization issue (Tornado 2/PPC/Solaris) Date: 06 Feb 2002 11:19:34 -0500 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: References: Sender: bpringlemeir@DeadDuck >>>>> "Michael" == Michael Lawnick writes: Michael> Hi Scott, as far as I remeber, there was a SPR about Michael> that. It was that the first object loaded wasn't included in Michael> sync, but next should be. If there isn't a patch, you could Michael> load a little dummy .o before your app ... If Michael's suggestion doesn't work, does the shell say anything when the `.o' (or `.out') is loading? Like unresolved, or cant use, or ??? regards, Bill Pringlemeir. - -- LOOK!! Sullen American teens wearing MADRAS shorts and ``Flock of Seagulls'' HAIRCUTS! vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Intermittent reboots Date: 06 Feb 2002 11:30:16 -0500 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: References: Sender: bpringlemeir@DeadDuck >>>>> "Denis" == Denis Perelyubskiy writes: >> we have an x86-based target. once in a while, we experience a >> reboot, but no exception information is printed, our exception >> handlers are not called, we're sure its not us who call the reboot >> function, etc. [snip info on vector table and stack protection] Denis> hate to followup to my own emails, sorry. but, i think i Denis> forgot to mention that i could think of another reason for no Denis> exception info being printed, and that's if we do not protect Denis> 'text', and we overwrite an exception handler, i think... I don't know if there is such a thing as a `double bus fault' in the x86. However, this often causes a reset on other processors. Ie, an exception in an exception. If bad data is sent to your interrupt routines, then this might happen. Another common problem is that the reset circuits find a voltage drop and the CPU will reset. This could be due to a driver causing some operation to draw a lot of power from the system. If you have some sort of manual/PDF/datasheet on the processor you are using, then look at the reset section (hopefully they have one). Usually they list every reason why a processor can reset. Go through each one and think about ways that that could happen... at least that is what I have done! Best of luck, Bill Pringlemeir. - -- I don't want to hear about your religion. No nukes! vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Intermittent reboots Date: 06 Feb 2002 17:24:22 +0000 (GMT) From: Martin Read Organization: Linux Unlimited Message-ID: References: In article , Bill Pringlemeir wrote: >I don't know if there is such a thing as a `double bus fault' in the >x86. However, this often causes a reset on other processors. Ie, an >exception in an exception. Intel architectures from 286 onwards define a "double fault" exception in protected mode, which occurs if certain fault-type exceptions occur during the attempt to start the handler for another exception (e.g. page fault upon trying to access the page fault handler). If a further exception occurs during the attempt to start the DF handler, a triple fault results and a CPU reset occurs. m. - -- \_\/_/| Martin Read - my opinions are my own. share them if you wish. \ / | i tried to tell her 'bout marx and engels, god and angels \/ | i don't really know what for, but she looked good in ribbons - ------+ --- sisters of mercy, "ribbons" --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Freebie VxWorks Tornado? Date: Sat, 02 Feb 2002 14:59:22 -0500 From: joe durusau Organization: BELLSOUTH.net Message-ID: <3C5C451A.100832E1@bellsouth.net> References: <57d4545a.0202011927.77254ba7@posting.google.com> Embed77 wrote: > I am interesting in moving from BSD development towards VxWorks. > Is there a non commercial development enviroment available for download? > Not that I've heard of, except for the simulator. > > Is there a x86 version of VxWorks OS that is downloadable? > Yes. Go to the website, but you only get the simulator that way. > > If not , what is the cheapest way to get into VxWorks development. > For a real target (not the simulator), think about $30K. You can learn about it for free, but doing real things takes real money. > > Thanks > > Sam. Speaking only for myself, Joe Durusau --------------------------- Newsgroups: comp.os.vxworks Subject: Re: motFccEnd.c Date: Wed, 6 Feb 2002 09:46:20 -0800 From: "Dan Gold" Organization: Mailgate.ORG Server - http://www.Mailgate.ORG Message-ID: References: Sender: gold@ensemble.com "Ali" wrote in message > > Hi Dan > Thanx for ur reply yes i have the source files for sysmotfccend.c and > configNet.h but the situation there is only for one instance of end > driver can u pls tell me what extra changes i have to make to these > two files so that i can support two fcc's for ethernet and that two > both of them in full duplex mode. > waiting for the reply > thanx I'm not using sysMotFccEnd.c, but my configNet.h file for multiple END devices (1 SCC and 2 FCC) looks like: END_TBL_ENTRY endDevTbl [] = { { 0, PQ2_SCC_LOAD_FUNC, PQ2_SCC1_LOAD_STRING,PQ2_SCC_BUFF_LOAN,NULL, FALSE}, { 3, PQ2_FCC_LOAD_FUNC, PQ2_FCC3_LOAD_STRING,PQ2_FCC_BUFF_LOAN,NULL, FALSE}, { 1, PQ2_FCC_LOAD_FUNC, PQ2_FCC1_LOAD_STRING,PQ2_FCC_BUFF_LOAN,NULL, FALSE}, { 0, END_TBL_END, NULL, 0, NULL, FALSE}, }; That should cause SENS to run the xxxEndLoad/xxxEndStart functions for each entry. To actually bring up the interface (since vxWorks only brings up the boot device in usrNetInit()), see the vxWorks FAQ. They have a little code snippet (that I wrote) to bring up more interfaces. As far as setting it for full duplex, I don't know where you specify that - I just autodetect, and then read my PHY to see if it's full or half duplex. hth, Dan Gold gold@ensemble.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Very important, I need the Login/password of my Hi-Focus ADSL modem (ECI) Date: 6 Feb 2002 09:54:08 -0800 From: support.metier5@wanadoo.fr (badmarsh) Organization: http://groups.google.com/ Message-ID: References: <3c4348e3$0$203$626a54ce@news.free.fr> salut, MDR, mais t'as raison faut tout essayer. Sauf que j'ai pas bien compris la réponse. "Mirak" wrote in message news:<3c4348e3$0$203$626a54ce@news.free.fr>... > Who knows how to get it ? > I can't configurate the modem without this log/pass. > Bye. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado-2.1/mips and cp7000 BSP (Ocelot MIPS board) Date: Wed, 06 Feb 2002 18:42:36 GMT From: Andray Kaganovsky Organization: Primus Canada Message-ID: References: Johan Borkhuis wrote in news:Xns91AB66D15BFBDborkhuisagerecom@135.7.153.102: > Andray Kaganovsky wrote: > >> Greetings, >> >> I am using Tornado-2.1/mips and cp7000 BSP (Ocelot board by >> Momentum Computers). The original version of this BSP that >> came with the board was for Tornado-2.0x; I got T-2.1 compatible >> version of this BSP directly from WindRiver. >> >> Here is my problems: >> >> -> printf "%d %d %d %d %d %d %d %d\n",1,2,3,4,5,6,7,8 >> 1 2 3 4 5 6 7 8 >> >> -> sp printf,"%d %d %d %d %d %d %d %d\n",1,2,3,4,5,6,7,8 >> 1 2 3 0 0 536912869 -33554961 60401228 > > I tried it on a IDT board using MIPS32sfgnu library and running the > commands using the target shell, and I got the right output. Did you try > it using the target shell or the host shell? > > Groeten, > Johan > [AndrayK 2]. On the host shell. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: DOS key like features in vxworks Date: 6 Feb 2002 11:11:55 -0800 From: mark_ck_chang@yahoo.com (Mark Chang) Organization: http://groups.google.com/ Message-ID: <9b6f8883.0202061111.43325042@posting.google.com> References: <9b6f8883.0201311117.2e1b5206@posting.google.com> "Dave Korn" wrote in message news:... > "Mark Chang" wrote in message > news:9b6f8883.0201311117.2e1b5206@posting.google.com... > > Does anyone know of code or have > > code that will do DOS Key like features? > > You know, up arrow show history, line editing, insert and replace. > > I think it will have to replace tylib.c - > > It's built in: look up ledlib.h for vxworks line editing features. > > DaveK > -- > moderator of > alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow > Burn your ID card! http://www.optional-identity.org.uk/ > Help support the campaign, copy this into your .sig! > Proud Member of the Exclusive "I have been plonked by Davee because he > thinks I'm interesting" List Member # > Master of Many Meowing Minions > Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage above > and beyond the call of hilarity. Thanks - I took a look at ledlib.c, it's not what I want and it's too much. So I ended up writing my own DOS key supporting the basic keys (DEL, Arrows, Home, End, BackSpace). Mark --------------------------- Newsgroups: comp.os.vxworks Subject: Re: PCI Device Base Address Date: Wed, 06 Feb 2002 13:11:54 -0600 From: Andrew Johnson Organization: APS, Argonne National Laboratory Message-ID: <3C617FFA.814299A3@aps.anl.gov> References: Arun Prasad wrote: > > I'm using IBM PPC405GP walnut board. > > I'm facing some problem in getting the Base address of my PCI based > NIC Card. > > I'm not getting the correct base address of our PCI based NIC card. > I'm getting a value which doesn't fall within the PCI memory or I/O > range. I feel like, my device hasn't been properly initiated. I'm able > to find my card using pciFindDevice() call. I am able to find its IRQ > no. I'm able to dump the PCI header for our card using pciConfigInXXX > functions. But base address are wrong. I don't know whats wrong with > that. Don't forget that the base address is an address on the PCIbus, and will need translation to convert it into a CPU address. I've never used the walnut so I don't know how its BSP maps PCI addresses to CPU addresses, but it's likely that you just need to add the base address of the PCI memory (or I/O) range. > ********How to choose Base Address for my device on my wish? > How to assign base address for our card statically? > ********How to program the PCI configuration space to mention > I have chosen my own base address for my card? That typically depends whether your sysLib.c file includes "drv/pci/pciConfigLib.h" or "drv/pci/pciAutoConfigLib.h". If only the former then you will have to choose the address yourself, but if the latter then it's probably being set for you and you should not attempt to change the base address register yourself, just read it and convert to a CPU address. - - Andrew - -- Perfection is reached, not when there is no longer anything to add, but when there is no longer anything to take away. - - Antoine de Saint-Exupery --------------------------- Newsgroups: comp.os.vxworks Subject: problem compiling curses with T2/vxworks5.4 Date: Wed, 06 Feb 2002 14:15:27 -0500 From: Gilad Katz Organization: Mailgate.ORG Server - http://www.Mailgate.ORG Message-ID: Sender: bpatel@fwion.com Hi, I am using Tornado2.0/vxWorks5.4 on powerpc target with host windows2000. I downloaded the curses package from ftp://ftp.atd.ucar.edu. After building the projects for termlib and curses subdirectory, I went ahead to build the libraries libtermlib.a and libcurses.a. Building libtermlib.a went without any hitches. However compiling the source files under curses subdirectory is giving me problems related to mvscanw() functions. I am not able to figure out why I am getting compilation error.Attached is the curses.h, mwscanw.c and the compile log showing the error. I will appreciate if any porting pitfalls are informed to me. thank you for your help. - ---------------------------------------------- compile log - ----------------------------------------------- vxrm *.o *.rpo ctdt.c symTbl.c vxApp* vxrm C:\vw_curses\proj_curses\prjComps.h C:\vw_curses\proj_curses\prjParams.h C:\vw_curses \proj_curses\prjConfig.c C:\vw_curses\proj_curses\linkSyms.c vxrm C:\vw_curses\proj_curses\libs.nm C:\vw_curses\proj_curses\libs.size ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT - -fvolatile -fno-bui ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. - -IC:\Tornado\target\h -DCPU=PPCEC60 3 -DVxWorks -c C:\vw_curses\curses\unctrl.c ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT - -fvolatile -fno-bui ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. - -IC:\Tornado\target\h -DCPU=PPCEC60 3 -DVxWorks -c C:\vw_curses\curses\addch.c ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT - -fvolatile -fno-bui ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. - -IC:\Tornado\target\h -DCPU=PPCEC60 3 -DVxWorks -c C:\vw_curses\curses\addstr.c ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT - -fvolatile -fno-bui ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. - -IC:\Tornado\target\h -DCPU=PPCEC60 3 -DVxWorks -c C:\vw_curses\curses\box.c ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT - -fvolatile -fno-bui ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. - -IC:\Tornado\target\h -DCPU=PPCEC60 3 -DVxWorks -c C:\vw_curses\curses\clear.c ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT - -fvolatile -fno-bui ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. - -IC:\Tornado\target\h -DCPU=PPCEC60 3 -DVxWorks -c C:\vw_curses\curses\clrtobot.c ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT - -fvolatile -fno-bui ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. - -IC:\Tornado\target\h -DCPU=PPCEC60 3 -DVxWorks -c C:\vw_curses\curses\clrtoeol.c ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT - -fvolatile -fno-bui ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. - -IC:\Tornado\target\h -DCPU=PPCEC60 3 -DVxWorks -c C:\vw_curses\curses\cr_put.c ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT - -fvolatile -fno-bui ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. - -IC:\Tornado\target\h -DCPU=PPCEC60 3 -DVxWorks -c C:\vw_curses\curses\cr_tty.c C:\vw_curses\curses\cr_tty.c:46: warning: `tgoto' declared `static' but never defined ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT - -fvolatile -fno-bui ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. - -IC:\Tornado\target\h -DCPU=PPCEC60 3 -DVxWorks -c C:\vw_curses\curses\curses.c ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT - -fvolatile -fno-bui ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. - -IC:\Tornado\target\h -DCPU=PPCEC60 3 -DVxWorks -c C:\vw_curses\curses\delch.c ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT - -fvolatile -fno-bui ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. - -IC:\Tornado\target\h -DCPU=PPCEC60 3 -DVxWorks -c C:\vw_curses\curses\deleteln.c ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT - -fvolatile -fno-bui ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. - -IC:\Tornado\target\h -DCPU=PPCEC60 3 -DVxWorks -c C:\vw_curses\curses\delwin.c ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT - -fvolatile -fno-bui ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. - -IC:\Tornado\target\h -DCPU=PPCEC60 3 -DVxWorks -c C:\vw_curses\curses\endwin.c ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT - -fvolatile -fno-bui ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. - -IC:\Tornado\target\h -DCPU=PPCEC60 3 -DVxWorks -c C:\vw_curses\curses\erase.c ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT - -fvolatile -fno-bui ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. - -IC:\Tornado\target\h -DCPU=PPCEC60 3 -DVxWorks -c C:\vw_curses\curses\fullname.c ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT - -fvolatile -fno-bui ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. - -IC:\Tornado\target\h -DCPU=PPCEC60 3 -DVxWorks -c C:\vw_curses\curses\getch.c ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT - -fvolatile -fno-bui ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. - -IC:\Tornado\target\h -DCPU=PPCEC60 3 -DVxWorks -c C:\vw_curses\curses\getstr.c C:\vw_curses\curses\getstr.c: In function `wgetstr': C:\vw_curses\curses\getstr.c:33: warning: comparison is always 1 due to limited range of d ata type C:\vw_curses\curses\getstr.c:35: warning: comparison is always 0 due to limited range of d ata type ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT - -fvolatile -fno-bui ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. - -IC:\Tornado\target\h -DCPU=PPCEC60 3 -DVxWorks -c C:\vw_curses\curses\id_subwins.c ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT - -fvolatile -fno-bui ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. - -IC:\Tornado\target\h -DCPU=PPCEC60 3 -DVxWorks -c C:\vw_curses\curses\idlok.c ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT - -fvolatile -fno-bui ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. - -IC:\Tornado\target\h -DCPU=PPCEC60 3 -DVxWorks -c C:\vw_curses\curses\initscr.c C:\vw_curses\curses\initscr.c: In function `initscr': C:\vw_curses\curses\initscr.c:81: warning: comparison between pointer and integer ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT - -fvolatile -fno-bui ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. - -IC:\Tornado\target\h -DCPU=PPCEC60 3 -DVxWorks -c C:\vw_curses\curses\insch.c ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT - -fvolatile -fno-bui ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. - -IC:\Tornado\target\h -DCPU=PPCEC60 3 -DVxWorks -c C:\vw_curses\curses\insertln.c ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT - -fvolatile -fno-bui ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. - -IC:\Tornado\target\h -DCPU=PPCEC60 3 -DVxWorks -c C:\vw_curses\curses\longname.c ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT - -fvolatile -fno-bui ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. - -IC:\Tornado\target\h -DCPU=PPCEC60 3 -DVxWorks -c C:\vw_curses\curses\move.c ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT - -fvolatile -fno-bui ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. - -IC:\Tornado\target\h -DCPU=PPCEC60 3 -DVxWorks -c C:\vw_curses\curses\mvprintw.c ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT - -fvolatile -fno-bui ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. - -IC:\Tornado\target\h -DCPU=PPCEC60 3 -DVxWorks -c C:\vw_curses\curses\mvscanw.c C:\vw_curses\curses\mvscanw.c: In function `mvwscanw': C:\vw_curses\curses\mvscanw.c:51: argument `win' doesn't match prototype C:\vw_curses\curses\curses.h:166: prototype declaration C:\vw_curses\curses\mvscanw.c:51: argument `x' doesn't match prototype C:\vw_curses\curses\curses.h:166: prototype declaration C:\vw_curses\curses\mvscanw.c:51: argument `fmt' doesn't match prototype C:\vw_curses\curses\curses.h:166: prototype declaration C:\vw_curses\curses\mvscanw.c:51: number of arguments doesn't match prototype C:\vw_curses\curses\curses.h:166: prototype declaration make: *** [mvscanw.o] Error 0x1 Done. - --------------------------------------------------------------------------------- curses.h file - --------------------------------------------------------------------------------- * Copyright (c) 1981 Regents of the University of California. * All rights reserved. * * Redistribution and use in source and binary forms are permitted * provided that the above copyright notice and this paragraph are * duplicated in all such forms and that any documentation, * advertising materials, and other materials related to such * distribution and use acknowledge that the software was developed * by the University of California, Berkeley. The name of the * University may not be used to endorse or promote products derived * from this software without specific prior written permission. * THIS SOFTWARE IS PROVIDED ``AS IS'' AND WITHOUT ANY EXPRESS OR * IMPLIED WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED * WARRANTIES OF MERCHANTIBILITY AND FITNESS FOR A PARTICULAR PURPOSE. * * @(#)curses.h 5.4 (Berkeley) 6/30/88 */ # ifndef WINDOW # include #ifndef VxWorks # include #else # include # include #endif # define bool char # define reg register #ifdef TRUE # undef TRUE # undef FALSE #endif # define TRUE (1) # define FALSE (0) # define ERR (-1) # define CURSES_ERR ERR # define CURSES_OK OK # define _ENDLINE 001 # define _FULLWIN 002 # define _SCROLLWIN 004 # define _FLUSH 010 # define _FULLLINE 020 # define _IDLINE 040 # define _STANDOUT 0200 # define _NOCHANGE -1 # define _puts(s) tputs(s, 0, _putchar) #ifndef VxWorks typedef struct sgttyb SGTTY; #endif /* * Capabilities from termcap */ extern bool AM, BS, CA, DA, DB, EO, HC, HZ, IN, MI, MS, NC, NS, OS, UL, XB, XN, XT, XS, XX; extern char *AL, *BC, *BT, *CD, *CE, *CL, *CM, *CR, *CS, *DC, *DL, *DM, *DO, *ED, *EI, *K0, *K1, *K2, *K3, *K4, *K5, *K6, *K7, *K8, *K9, *HO, *IC, *IM, *IP, *KD, *KE, *KH, *KL, *KR, *KS, *KU, *LL, *MA, *ND, *NL, *RC, *SC, *SE, *SF, *SO, *SR, *TA, *TE, *TI, *UC, *UE, *UP, *US, *VB, *VS, *VE, *AL_PARM, *DL_PARM, *UP_PARM, *DOWN_PARM, *LEFT_PARM, *RIGHT_PARM; extern char PC; /* * From the tty modes... */ extern bool GT, NONL, UPPERCASE, normtty, _pfast; struct _win_st { short _cury, _curx; short _maxy, _maxx; short _begy, _begx; short _flags; short _ch_off; bool _clear; bool _leave; bool _scroll; char **_y; short *_firstch; short *_lastch; struct _win_st *_nextp, *_orig; }; # define WINDOW struct _win_st extern bool My_term, _echoit, _rawmode, _endwin; extern char *Def_term, ttytype[]; extern int LINES, COLS, _tty_ch, _res_flg; #ifndef VxWorks extern SGTTY _tty; #endif extern WINDOW *stdscr, *curscr; /* * Define VOID to stop lint from generating "null effect" * comments. */ # ifdef lint int __void__; # define VOID(x) (__void__ = (int) (x)) # else # define VOID(x) (x) # endif /* * standard curses functions. * */ #ifdef __STDC__ extern WINDOW * newwin(int lines, int cols, int sy, int sx); extern WINDOW * renewwin(int lines, int cols, int sy, int sx, WINDOW* win); extern WINDOW * subwin(WINDOW *w, int lines, int cols, int sy, int sx); extern WINDOW * initscr(); extern int box (WINDOW*, char, char); extern int delwin(WINDOW*); extern int redelwin(WINDOW*); extern int mvcur(int, int, int, int); extern int overlay(WINDOW*, WINDOW*); extern int overwrite(WINDOW*, WINDOW*); extern int scroll(WINDOW*); extern int touchline(WINDOW*, int, int, int); extern int hideline(WINDOW*, int, int, int); extern int touchoverlap(WINDOW*, WINDOW*); extern int hideoverlap(WINDOW*, WINDOW*); extern int touchwin(WINDOW*); extern int hidewin(WINDOW*); extern int waddbytes(WINDOW* win, const char* bytes, int count); extern int waddch(WINDOW*, char); extern int waddstr(WINDOW*, const char*); extern int wclear(WINDOW*); extern int wclrtobot(WINDOW*); extern int wclrtoeol(WINDOW*); extern int wdelch(WINDOW*); extern int wdeleteln(WINDOW*); extern int werase(WINDOW*); extern int wgetch(WINDOW*); extern int wgetstr(WINDOW*, char*); extern int winch(WINDOW*); extern int winsch(WINDOW*, char); extern int winsertln(WINDOW*); extern int wmove(WINDOW*, int, int); extern int wrefresh(WINDOW*); extern int wstandend(WINDOW*); extern int wstandout(WINDOW*); extern int printw(const char * fmt, ...); extern int wprintw(WINDOW*, const char * fmt, ...); extern int mvprintw(int y, int x, const char * fmt, ...); extern int mvwprintw(WINDOW*, int y, int x, const char * fmt, ...); extern int wscanw(WINDOW*, const char *, ...); extern int mvwscanw(int, int, WINDOW*, const char*, ...); extern int endwin(); extern char *longname(char *, char *); extern char *getcap(char *); #else WINDOW *initscr(), *newwin(), *renewwin(), *subwin(); char *longname(), *getcap(); #endif __STDC__ /* * pseudo functions for standard screen */ # define addch(ch) VOID(waddch(stdscr, ch)) # define getch() VOID(wgetch(stdscr)) # define addbytes(da,co) VOID(waddbytes(stdscr, da,co)) # define addstr(str) VOID(waddbytes(stdscr, str, strlen(str))) # define getstr(str) VOID(wgetstr(stdscr, str)) # define move(y, x) VOID(wmove(stdscr, y, x)) # define clear() VOID(wclear(stdscr)) # define erase() VOID(werase(stdscr)) # define clrtobot() VOID(wclrtobot(stdscr)) # define clrtoeol() VOID(wclrtoeol(stdscr)) # define insertln() VOID(winsertln(stdscr)) # define deleteln() VOID(wdeleteln(stdscr)) # define refresh() VOID(wrefresh(stdscr)) # define inch() VOID(winch(stdscr)) # define insch(c) VOID(winsch(stdscr,c)) # define delch() VOID(wdelch(stdscr)) # define standout() VOID(wstandout(stdscr)) # define standend() VOID(wstandend(stdscr)) /* * mv functions * We should really lock between the wmove and the action. (***) */ #define mvwaddch(win,y,x,ch) VOID(wmove(win,y,x)==ERR?ERR:waddch(win,ch)) #define mvwgetch(win,y,x) VOID(wmove(win,y,x)==ERR?ERR:wgetch(win)) #define mvwaddbytes(win,y,x,da,co) \ VOID(wmove(win,y,x)==ERR?ERR:waddbytes(win,da,co)) #define mvwaddstr(win,y,x,str) \ VOID(wmove(win,y,x)==ERR?ERR:waddbytes(win,str,strlen(str))) #define mvwgetstr(win,y,x,str) VOID(wmove(win,y,x)==ERR?ERR:wgetstr(win,str)) #define mvwinch(win,y,x) VOID(wmove(win,y,x) == ERR ? ERR : winch(win)) #define mvwdelch(win,y,x) VOID(wmove(win,y,x) == ERR ? ERR : wdelch(win)) #define mvwinsch(win,y,x,c) VOID(wmove(win,y,x) == ERR ? ERR:winsch(win,c)) #define mvaddch(y,x,ch) mvwaddch(stdscr,y,x,ch) #define mvgetch(y,x) mvwgetch(stdscr,y,x) #define mvaddbytes(y,x,da,co) mvwaddbytes(stdscr,y,x,da,co) #define mvaddstr(y,x,str) mvwaddstr(stdscr,y,x,str) #define mvgetstr(y,x,str) mvwgetstr(stdscr,y,x,str) #define mvinch(y,x) mvwinch(stdscr,y,x) #define mvdelch(y,x) mvwdelch(stdscr,y,x) #define mvinsch(y,x,c) mvwinsch(stdscr,y,x,c) /* * psuedo functions */ #define clearok(win,bf) (win==NULL ? ERR : (win->_clear = bf)) #define leaveok(win,bf) (win==NULL ? ERR : (win->_leave = bf)) #define scrollok(win,bf) (win==NULL ? ERR : (win->_scroll = bf)) #define flushok(win,bf) (win==NULL ? ERR : (bf ? (win->_flags |= _FLUSH):(win->_flags &= ~_FLUSH))) #define getyx(win,y,x) (win==NULL ? ERR : (y = win->_cury, x = win->_curx)) #define winch(win) (win==NULL ? ERR : (win->_y[win->_cury][win->_curx] & 0177)) #ifndef VxWorks #define raw() (_tty.sg_flags|=RAW, _pfast=_rawmode=TRUE, stty(_tty_ch,&_tty)) #define noraw() (_tty.sg_flags&=~RAW,_rawmode=FALSE,_pfast=!(_tty.sg_flags&CRMOD),stty(_tty_ch,&_tty)) #define cbreak() (_tty.sg_flags |= CBREAK, _rawmode = TRUE, stty(_tty_ch,&_tty)) #define nocbreak() (_tty.sg_flags &= ~CBREAK,_rawmode=FALSE,stty(_tty_ch,&_tty)) #define crmode() cbreak() /* backwards compatability */ #define nocrmode() nocbreak() /* backwards compatability */ #define echo() (_tty.sg_flags |= ECHO, _echoit = TRUE, stty(_tty_ch, &_tty)) #define noecho() (_tty.sg_flags &= ~ECHO, _echoit = FALSE, stty(_tty_ch, &_tty)) #define nl() (_tty.sg_flags |= CRMOD,_pfast = _rawmode,stty(_tty_ch, &_tty)) #define nonl() (_tty.sg_flags &= ~CRMOD, _pfast = TRUE, stty(_tty_ch, &_tty)) #define savetty() ((void) gtty(_tty_ch, &_tty), _res_flg = _tty.sg_flags) #define resetty() (_tty.sg_flags = _res_flg, (void) stty(_tty_ch, &_tty)) #define erasechar() (_tty.sg_erase) #define killchar() (_tty.sg_kill) #define baudrate() (_tty.sg_ospeed) #else #define raw() (cbreak()) #define noraw() (nocbreak()) #define cbreak() (_rawmode = TRUE,\ ioctl(0,FIOSETOPTIONS,ioctl(0,FIOGETOPTIONS,0) & ~OPT_LINE) ) #define nocbreak() (_rawmode = FALSE,\ ioctl(0,FIOSETOPTIONS,ioctl(0,FIOGETOPTIONS,0) | OPT_LINE) ) #define crmode() cbreak() /* backwards compatability */ #define nocrmode() nocbreak() /* backwards compatability */ #define echo() (_echoit = TRUE,\ ioctl(0,FIOSETOPTIONS,ioctl(0,FIOGETOPTIONS,0) | OPT_ECHO) ) #define noecho() (_echoit = FALSE,\ ioctl(0,FIOSETOPTIONS,ioctl(0,FIOGETOPTIONS,0) & ~OPT_ECHO) ) #define nl() (OK) #define nonl() ( \ fprintf(stderr,"curses.h: nonl() not supported\n"); \ exit(1); \ ) #define savetty() (_res_flg = ioctl(0,FIOGETOPTIONS,0)) #define resetty() (ioctl(0,FIOSETOPTIONS,_res_flg)) #define erasechar() ('') #define killchar() ( \ fprintf(stderr,"curses.h: killchar() not supported\n"); \ exit(1); \ ) #define baudrate() (13) /* this is the termio/gtty code for 9600 baud, used by tputs() in the termlib library */ #endif /* * Used to be in unctrl.h. */ #define unctrl(c) _unctrl[(c) & 0177] extern char *_unctrl[]; # endif - ------------------------------------------------------------------------------------ mvwscanw.c file - ------------------------------------------------------------------------------------ * * Copyright (c) 1981 Regents of the University of California. * All rights reserved. * * Redistribution and use in source and binary forms are permitted * provided that the above copyright notice and this paragraph are * duplicated in all such forms and that any documentation, * advertising materials, and other materials related to such * distribution and use acknowledge that the software was developed * by the University of California, Berkeley. The name of the * University may not be used to endorse or promote products derived * from this software without specific prior written permission. * THIS SOFTWARE IS PROVIDED ``AS IS'' AND WITHOUT ANY EXPRESS OR * IMPLIED WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED * WARRANTIES OF MERCHANTIBILITY AND FITNESS FOR A PARTICULAR PURPOSE. */ #ifndef lint static char sccsid[] = "@(#)mvscanw.c 5.3 (Berkeley) 6/30/88"; #endif /* not lint */ # include "curses.ext" /* * implement the mvscanw commands. Due to the variable number of * arguments, they cannot be macros. Another sigh.... * */ mvscanw(y, x, fmt, args) reg int y, x; char *fmt; int args; { int retval; curses_P(); if (move(y, x) != OK) { curses_V(); return ERR; } retval= _sscans(stdscr, fmt, &args); curses_V(); return retval; } mvwscanw(win, y, x, fmt, args) reg WINDOW *win; reg int y, x; char *fmt; int args; { int retval; curses_P(); if (win == NULL) { curses_V(); return ERR; } if (wmove(win, y, x) != OK) { curses_V(); return ERR; } retval= _sscans(win, fmt, &args); curses_V(); return retval; } - -------------------------------------------------------------------------------------- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to change shell's default directory? Date: 2 Feb 2002 21:36:56 GMT From: Denis Perelyubskiy Message-ID: References: <9ec53d11.0202020631.21afbef6@posting.google.com> * Ignacio González [02-Sat-02 06:31 -0800]: > > Tornado 2.0.2 for Windows NT. > When I run a local shell from the Tornado IDE, I usually perform some > setting activities (like ?shConfig ... etc.) I've learned to perform > these things writing text scripts and typing shell's prompt. [...] > How could I specify this directory (as a default directory) > without having to write the entire path every time I call > "script.txt"? sorry, i dont do this, so i dont remember, but if you check shell's manual i think i saw something in there about changing things in .wind directory. i think there is a script you can place there, which would get executed every time you start a shell. perhaps this will be of use. in the worst case, if not running all the scripts from that location, perhaps you could do the cd "myfolder" and change to myfolder. anyway, i dont know if this is THE way to do this, but i think it should work. denis - -- if you want to reply by mail, please tweak my address in a reasonable way. sorry for the inconvenience. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Returning memory Date: 2 Feb 2002 21:39:18 GMT From: Denis Perelyubskiy Message-ID: References: <28809481.0202012009.10cf7545@posting.google.com> * rasayani [02-Fri-02 20:09 -0800]: > > If a task allocates memory using malloc/calloc and is deleted, is the > memory released to the free memory pool? as others have stated "no", but i think there is a taskDeleteHookAdd() or something of that nature. you could play with it, to make that function deallocate your memory if you can't do it before deleting the task. i am not sure if that fn is called every time the task goes away for whatever reason, or only on taskDelete() called by a user, but it is something to look at denis - -- if you want to reply by mail, please tweak my address in a reasonable way. sorry for the inconvenience. --------------------------- Newsgroups: comp.os.vxworks Subject: DOS key with timeouts, idea? Date: 6 Feb 2002 11:48:38 -0800 From: mark_ck_chang@yahoo.com (Mark Chang) Organization: http://groups.google.com/ Message-ID: <9b6f8883.0202061148.5789ec09@posting.google.com> I finished my DOS key function in Vxworks. Basically a simple set to support history and a few keys (bkspace, home, end, arrows, del) Now I want to add a timeout. So if the keyboard is idle for a set amount of time. It will logout the CLI. My basic function for getting keys is as such... do { ch = getchar(); if (ch == D_ESC) { ch = getchar(); if (ch == SPEC_KEY) { ch = getchar()+256; } } SNIP } while (ch != D_LF); I have 2 ideas Lets say the user sets a timeout of 30 seconds. 1) Spawn a task before this and set the timeout to 10 seconds, set a sleep count to 3 (3x10=30 seconds) Every 10 seconds, Check a flag (set in the above routine) to see if the keyboard was touched. A) If TRUE, clear the flag, clear a sleep count. B) If the touch flag is FALSE, increment the sleep count by 1. C) If the sleep count == 3, logout Repeat A-C My second idea is to use select on the fd passed back by ioTaskStdGet() for input. Will the select return if a key is hit and will my getchar still have the data? Thanks, Mark --------------------------- Newsgroups: comp.os.vxworks Subject: i have downloaded the tornado 2.0 protoyper, but the key as not been emailed to me Date: Sat, 2 Feb 2002 22:53:02 +0100 From: "Mirak" Organization: Wanadoo, l'internet avec France Telecom Message-ID: Do you have a key ? Or how can i obtain it ? It should be emailed to me but it isn't. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: frequent rebooting for an ethernet interface looses connectivity?? Date: 6 Feb 2002 12:17:30 -0800 From: igtorque@eliop.es (=?ISO-8859-1?Q?Ignacio_Gonz=E1lez?=) Organization: http://groups.google.com/ Message-ID: <9ec53d11.0202061217.22d8e165@posting.google.com> References: <3C5DF5A5.1F6CFFCC@lucent.com> <9ec53d11.0202040118.514e0357@posting.google.com> <3C5EE440.F4C4B94@lucent.com> Babu Yama wrote in message news:<3C5EE440.F4C4B94@lucent.com>... > Sorry, I forgot the following information. > > My target details are: > > > target: MPC8260 (FCC2 ethernet port) - when ping fails FCC1 is fine and t > alking. > vxWorks: 5.4 > BSP: wrSbc8260 BSP 33 (wind river's single board compter FCC driver) > > Any ideas as to why the FCC2 ethernet port does not talk after some reboo > ts.... > > Thanks, > Babu > Pity, my problematic target was a custom i386EX-based board, VxWorks 5.4, with a custom cs8900A-based Ethernet board. We used the driver supplied by Cirrus for free (END version 3.04). After fixing the bug and reporting it, Cirrus kindly gave us the last this-bug-free version (END version 3.06) The problem was a race condition between the net task and the driver's ISR, which could surface after a transmission underrun, or when CPU was slower than expected by the driver. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: problem compiling curses with T2/vxworks5.4 Date: Wed, 6 Feb 2002 22:54:21 +0100 From: "Werner Schiendl" Message-ID: <3c61a606@brateggebdc5.br-automation.co.at> References: Hi, > mvwscanw(win, y, x, fmt, args) > reg WINDOW *win; > reg int y, x; > char *fmt; > int args; { > two things... well... three: 1) The prototype states, that the first two parameters should be int - so I'd guess y and x - then is a WINDOW *. Whatever is wrong, needs be changed (lookup occurences of this function-call in the source and determine whats right). 2) The prototype states fmt as const char *, the function omits the const. Either add it here (or if that cannot be done because the function does not obey const-ness - remove it in the prototype. 3) I have no idea how you could pass the variable args (the ... in the prototype on to _sscans() function). I've never seen this syntax, using an int - maybe someone else here has and knows how to work around. hth Werner "Gilad Katz" wrote in message news:a3rvl4$htp$1@newsreader.mailgate.org... > Hi, > > I am using Tornado2.0/vxWorks5.4 on powerpc target with host > windows2000. > I downloaded the curses package from ftp://ftp.atd.ucar.edu. After > building the > projects for termlib and curses subdirectory, I went ahead to build the > libraries > libtermlib.a and libcurses.a. Building libtermlib.a went without any > hitches. However compiling the source files under curses subdirectory is > giving me problems related to > mvscanw() functions. I am not able to figure out why I am getting > compilation error.Attached is the curses.h, mwscanw.c and the compile > log showing the error. I will appreciate if any porting pitfalls are > informed to me. > thank you for your help. > > ---------------------------------------------- > compile log > ----------------------------------------------- > vxrm *.o *.rpo ctdt.c symTbl.c vxApp* > vxrm C:\vw_curses\proj_curses\prjComps.h > C:\vw_curses\proj_curses\prjParams.h C:\vw_curses > \proj_curses\prjConfig.c C:\vw_curses\proj_curses\linkSyms.c > vxrm C:\vw_curses\proj_curses\libs.nm C:\vw_curses\proj_curses\libs.size > > ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT > -fvolatile -fno-bui > ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. > -IC:\Tornado\target\h -DCPU=PPCEC60 > 3 -DVxWorks -c C:\vw_curses\curses\unctrl.c > ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT > -fvolatile -fno-bui > ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. > -IC:\Tornado\target\h -DCPU=PPCEC60 > 3 -DVxWorks -c C:\vw_curses\curses\addch.c > ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT > -fvolatile -fno-bui > ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. > -IC:\Tornado\target\h -DCPU=PPCEC60 > 3 -DVxWorks -c C:\vw_curses\curses\addstr.c > ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT > -fvolatile -fno-bui > ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. > -IC:\Tornado\target\h -DCPU=PPCEC60 > 3 -DVxWorks -c C:\vw_curses\curses\box.c > ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT > -fvolatile -fno-bui > ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. > -IC:\Tornado\target\h -DCPU=PPCEC60 > 3 -DVxWorks -c C:\vw_curses\curses\clear.c > ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT > -fvolatile -fno-bui > ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. > -IC:\Tornado\target\h -DCPU=PPCEC60 > 3 -DVxWorks -c C:\vw_curses\curses\clrtobot.c > ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT > -fvolatile -fno-bui > ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. > -IC:\Tornado\target\h -DCPU=PPCEC60 > 3 -DVxWorks -c C:\vw_curses\curses\clrtoeol.c > ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT > -fvolatile -fno-bui > ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. > -IC:\Tornado\target\h -DCPU=PPCEC60 > 3 -DVxWorks -c C:\vw_curses\curses\cr_put.c > ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT > -fvolatile -fno-bui > ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. > -IC:\Tornado\target\h -DCPU=PPCEC60 > 3 -DVxWorks -c C:\vw_curses\curses\cr_tty.c > C:\vw_curses\curses\cr_tty.c:46: warning: `tgoto' declared `static' but > never defined > ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT > -fvolatile -fno-bui > ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. > -IC:\Tornado\target\h -DCPU=PPCEC60 > 3 -DVxWorks -c C:\vw_curses\curses\curses.c > ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT > -fvolatile -fno-bui > ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. > -IC:\Tornado\target\h -DCPU=PPCEC60 > 3 -DVxWorks -c C:\vw_curses\curses\delch.c > ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT > -fvolatile -fno-bui > ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. > -IC:\Tornado\target\h -DCPU=PPCEC60 > 3 -DVxWorks -c C:\vw_curses\curses\deleteln.c > ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT > -fvolatile -fno-bui > ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. > -IC:\Tornado\target\h -DCPU=PPCEC60 > 3 -DVxWorks -c C:\vw_curses\curses\delwin.c > ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT > -fvolatile -fno-bui > ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. > -IC:\Tornado\target\h -DCPU=PPCEC60 > 3 -DVxWorks -c C:\vw_curses\curses\endwin.c > ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT > -fvolatile -fno-bui > ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. > -IC:\Tornado\target\h -DCPU=PPCEC60 > 3 -DVxWorks -c C:\vw_curses\curses\erase.c > ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT > -fvolatile -fno-bui > ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. > -IC:\Tornado\target\h -DCPU=PPCEC60 > 3 -DVxWorks -c C:\vw_curses\curses\fullname.c > ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT > -fvolatile -fno-bui > ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. > -IC:\Tornado\target\h -DCPU=PPCEC60 > 3 -DVxWorks -c C:\vw_curses\curses\getch.c > ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT > -fvolatile -fno-bui > ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. > -IC:\Tornado\target\h -DCPU=PPCEC60 > 3 -DVxWorks -c C:\vw_curses\curses\getstr.c > C:\vw_curses\curses\getstr.c: In function `wgetstr': > C:\vw_curses\curses\getstr.c:33: warning: comparison is always 1 due to > limited range of d > ata type > C:\vw_curses\curses\getstr.c:35: warning: comparison is always 0 due to > limited range of d > ata type > ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT > -fvolatile -fno-bui > ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. > -IC:\Tornado\target\h -DCPU=PPCEC60 > 3 -DVxWorks -c C:\vw_curses\curses\id_subwins.c > ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT > -fvolatile -fno-bui > ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. > -IC:\Tornado\target\h -DCPU=PPCEC60 > 3 -DVxWorks -c C:\vw_curses\curses\idlok.c > ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT > -fvolatile -fno-bui > ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. > -IC:\Tornado\target\h -DCPU=PPCEC60 > 3 -DVxWorks -c C:\vw_curses\curses\initscr.c > C:\vw_curses\curses\initscr.c: In function `initscr': > C:\vw_curses\curses\initscr.c:81: warning: comparison between pointer > and integer > ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT > -fvolatile -fno-bui > ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. > -IC:\Tornado\target\h -DCPU=PPCEC60 > 3 -DVxWorks -c C:\vw_curses\curses\insch.c > ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT > -fvolatile -fno-bui > ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. > -IC:\Tornado\target\h -DCPU=PPCEC60 > 3 -DVxWorks -c C:\vw_curses\curses\insertln.c > ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT > -fvolatile -fno-bui > ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. > -IC:\Tornado\target\h -DCPU=PPCEC60 > 3 -DVxWorks -c C:\vw_curses\curses\longname.c > ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT > -fvolatile -fno-bui > ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. > -IC:\Tornado\target\h -DCPU=PPCEC60 > 3 -DVxWorks -c C:\vw_curses\curses\move.c > ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT > -fvolatile -fno-bui > ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. > -IC:\Tornado\target\h -DCPU=PPCEC60 > 3 -DVxWorks -c C:\vw_curses\curses\mvprintw.c > ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT > -fvolatile -fno-bui > ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. > -IC:\Tornado\target\h -DCPU=PPCEC60 > 3 -DVxWorks -c C:\vw_curses\curses\mvscanw.c > C:\vw_curses\curses\mvscanw.c: In function `mvwscanw': > C:\vw_curses\curses\mvscanw.c:51: argument `win' doesn't match prototype > > C:\vw_curses\curses\curses.h:166: prototype declaration > C:\vw_curses\curses\mvscanw.c:51: argument `x' doesn't match prototype > C:\vw_curses\curses\curses.h:166: prototype declaration > C:\vw_curses\curses\mvscanw.c:51: argument `fmt' doesn't match prototype > > C:\vw_curses\curses\curses.h:166: prototype declaration > C:\vw_curses\curses\mvscanw.c:51: number of arguments doesn't match > prototype > C:\vw_curses\curses\curses.h:166: prototype declaration > make: *** [mvscanw.o] Error 0x1 > > Done. > -------------------------------------------------------------------------- - ------- > > curses.h file > -------------------------------------------------------------------------- - ------- > > * Copyright (c) 1981 Regents of the University of California. > * All rights reserved. > * > * Redistribution and use in source and binary forms are permitted > * provided that the above copyright notice and this paragraph are > * duplicated in all such forms and that any documentation, > * advertising materials, and other materials related to such > * distribution and use acknowledge that the software was developed > * by the University of California, Berkeley. The name of the > * University may not be used to endorse or promote products derived > * from this software without specific prior written permission. > * THIS SOFTWARE IS PROVIDED ``AS IS'' AND WITHOUT ANY EXPRESS OR > * IMPLIED WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED > * WARRANTIES OF MERCHANTIBILITY AND FITNESS FOR A PARTICULAR PURPOSE. > * > * @(#)curses.h 5.4 (Berkeley) 6/30/88 > */ > > # ifndef WINDOW > > # include > > #ifndef VxWorks > # include > #else > # include > # include > #endif > > # define bool char > # define reg register > > #ifdef TRUE > # undef TRUE > # undef FALSE > #endif > # define TRUE (1) > # define FALSE (0) > # define ERR (-1) > # define CURSES_ERR ERR > # define CURSES_OK OK > > # define _ENDLINE 001 > # define _FULLWIN 002 > # define _SCROLLWIN 004 > # define _FLUSH 010 > # define _FULLLINE 020 > # define _IDLINE 040 > # define _STANDOUT 0200 > # define _NOCHANGE -1 > > # define _puts(s) tputs(s, 0, _putchar) > > #ifndef VxWorks > typedef struct sgttyb SGTTY; > #endif > > /* > * Capabilities from termcap > */ > > extern bool AM, BS, CA, DA, DB, EO, HC, HZ, IN, MI, MS, NC, NS, OS, > UL, > XB, XN, XT, XS, XX; > extern char *AL, *BC, *BT, *CD, *CE, *CL, *CM, *CR, *CS, *DC, *DL, > *DM, *DO, *ED, *EI, *K0, *K1, *K2, *K3, *K4, *K5, *K6, > *K7, *K8, *K9, *HO, *IC, *IM, *IP, *KD, *KE, *KH, *KL, > *KR, *KS, *KU, *LL, *MA, *ND, *NL, *RC, *SC, *SE, *SF, > *SO, *SR, *TA, *TE, *TI, *UC, *UE, *UP, *US, *VB, *VS, > *VE, *AL_PARM, *DL_PARM, *UP_PARM, *DOWN_PARM, > *LEFT_PARM, *RIGHT_PARM; > extern char PC; > > /* > * From the tty modes... > */ > > extern bool GT, NONL, UPPERCASE, normtty, _pfast; > > struct _win_st { > short _cury, _curx; > short _maxy, _maxx; > short _begy, _begx; > short _flags; > short _ch_off; > bool _clear; > bool _leave; > bool _scroll; > char **_y; > short *_firstch; > short *_lastch; > struct _win_st *_nextp, *_orig; > }; > > # define WINDOW struct _win_st > > extern bool My_term, _echoit, _rawmode, _endwin; > > extern char *Def_term, ttytype[]; > > extern int LINES, COLS, _tty_ch, _res_flg; > > #ifndef VxWorks > extern SGTTY _tty; > #endif > > extern WINDOW *stdscr, *curscr; > > /* > * Define VOID to stop lint from generating "null effect" > * comments. > */ > # ifdef lint > int __void__; > # define VOID(x) (__void__ = (int) (x)) > # else > # define VOID(x) (x) > # endif > > /* > * standard curses functions. > * > */ > > #ifdef __STDC__ > extern WINDOW * newwin(int lines, int cols, int sy, int sx); > extern WINDOW * renewwin(int lines, int cols, int sy, int sx, WINDOW* > win); > extern WINDOW * subwin(WINDOW *w, int lines, int cols, int sy, int sx); > extern WINDOW * initscr(); > extern int box (WINDOW*, char, char); > extern int delwin(WINDOW*); > extern int redelwin(WINDOW*); > extern int mvcur(int, int, int, int); > extern int overlay(WINDOW*, WINDOW*); > extern int overwrite(WINDOW*, WINDOW*); > extern int scroll(WINDOW*); > extern int touchline(WINDOW*, int, int, int); > extern int hideline(WINDOW*, int, int, int); > extern int touchoverlap(WINDOW*, WINDOW*); > extern int hideoverlap(WINDOW*, WINDOW*); > extern int touchwin(WINDOW*); > extern int hidewin(WINDOW*); > extern int waddbytes(WINDOW* win, const char* bytes, int count); > extern int waddch(WINDOW*, char); > extern int waddstr(WINDOW*, const char*); > extern int wclear(WINDOW*); > extern int wclrtobot(WINDOW*); > extern int wclrtoeol(WINDOW*); > extern int wdelch(WINDOW*); > extern int wdeleteln(WINDOW*); > extern int werase(WINDOW*); > extern int wgetch(WINDOW*); > extern int wgetstr(WINDOW*, char*); > extern int winch(WINDOW*); > extern int winsch(WINDOW*, char); > extern int winsertln(WINDOW*); > extern int wmove(WINDOW*, int, int); > extern int wrefresh(WINDOW*); > extern int wstandend(WINDOW*); > extern int wstandout(WINDOW*); > extern int printw(const char * fmt, ...); > extern int wprintw(WINDOW*, const char * fmt, ...); > extern int mvprintw(int y, int x, const char * fmt, ...); > extern int mvwprintw(WINDOW*, int y, int x, const char * fmt, ...); > > extern int wscanw(WINDOW*, const char *, ...); > extern int mvwscanw(int, int, WINDOW*, const char*, ...); > extern int endwin(); > > extern char *longname(char *, char *); > extern char *getcap(char *); > > #else > WINDOW *initscr(), *newwin(), *renewwin(), *subwin(); > char *longname(), *getcap(); > > #endif __STDC__ > > /* > * pseudo functions for standard screen > */ > # define addch(ch) VOID(waddch(stdscr, ch)) > # define getch() VOID(wgetch(stdscr)) > # define addbytes(da,co) VOID(waddbytes(stdscr, da,co)) > # define addstr(str) VOID(waddbytes(stdscr, str, strlen(str))) > # define getstr(str) VOID(wgetstr(stdscr, str)) > # define move(y, x) VOID(wmove(stdscr, y, x)) > # define clear() VOID(wclear(stdscr)) > # define erase() VOID(werase(stdscr)) > # define clrtobot() VOID(wclrtobot(stdscr)) > # define clrtoeol() VOID(wclrtoeol(stdscr)) > # define insertln() VOID(winsertln(stdscr)) > # define deleteln() VOID(wdeleteln(stdscr)) > # define refresh() VOID(wrefresh(stdscr)) > # define inch() VOID(winch(stdscr)) > # define insch(c) VOID(winsch(stdscr,c)) > # define delch() VOID(wdelch(stdscr)) > # define standout() VOID(wstandout(stdscr)) > # define standend() VOID(wstandend(stdscr)) > > /* > * mv functions > * We should really lock between the wmove and the action. (***) > */ > #define mvwaddch(win,y,x,ch) > VOID(wmove(win,y,x)==ERR?ERR:waddch(win,ch)) > #define mvwgetch(win,y,x) VOID(wmove(win,y,x)==ERR?ERR:wgetch(win)) > #define mvwaddbytes(win,y,x,da,co) \ > VOID(wmove(win,y,x)==ERR?ERR:waddbytes(win,da,co)) > #define mvwaddstr(win,y,x,str) \ > VOID(wmove(win,y,x)==ERR?ERR:waddbytes(win,str,strlen(str))) > #define mvwgetstr(win,y,x,str) > VOID(wmove(win,y,x)==ERR?ERR:wgetstr(win,str)) > #define mvwinch(win,y,x) VOID(wmove(win,y,x) == ERR ? ERR : winch(win)) > #define mvwdelch(win,y,x) VOID(wmove(win,y,x) == ERR ? ERR : > wdelch(win)) > #define mvwinsch(win,y,x,c) VOID(wmove(win,y,x) == ERR ? > ERR:winsch(win,c)) > #define mvaddch(y,x,ch) mvwaddch(stdscr,y,x,ch) > #define mvgetch(y,x) mvwgetch(stdscr,y,x) > #define mvaddbytes(y,x,da,co) mvwaddbytes(stdscr,y,x,da,co) > #define mvaddstr(y,x,str) mvwaddstr(stdscr,y,x,str) > #define mvgetstr(y,x,str) mvwgetstr(stdscr,y,x,str) > #define mvinch(y,x) mvwinch(stdscr,y,x) > #define mvdelch(y,x) mvwdelch(stdscr,y,x) > #define mvinsch(y,x,c) mvwinsch(stdscr,y,x,c) > > /* > * psuedo functions > */ > > #define clearok(win,bf) (win==NULL ? ERR : (win->_clear = bf)) > #define leaveok(win,bf) (win==NULL ? ERR : (win->_leave = bf)) > #define scrollok(win,bf) (win==NULL ? ERR : (win->_scroll = bf)) > #define flushok(win,bf) (win==NULL ? ERR : (bf ? (win->_flags |= > _FLUSH):(win->_flags &= ~_FLUSH))) > #define getyx(win,y,x) (win==NULL ? ERR : (y = win->_cury, x = > win->_curx)) > #define winch(win) (win==NULL ? ERR : (win->_y[win->_cury][win->_curx] > & 0177)) > > #ifndef VxWorks > #define raw() (_tty.sg_flags|=RAW, _pfast=_rawmode=TRUE, > stty(_tty_ch,&_tty)) > #define noraw() > (_tty.sg_flags&=~RAW,_rawmode=FALSE,_pfast=!(_tty.sg_flags&CRMOD),stty(_tty_ ch,&_tty)) > > #define cbreak() (_tty.sg_flags |= CBREAK, _rawmode = TRUE, > stty(_tty_ch,&_tty)) > #define nocbreak() (_tty.sg_flags &= > ~CBREAK,_rawmode=FALSE,stty(_tty_ch,&_tty)) > #define crmode() cbreak() /* backwards compatability */ > #define nocrmode() nocbreak() /* backwards compatability */ > #define echo() (_tty.sg_flags |= ECHO, _echoit = TRUE, stty(_tty_ch, > &_tty)) > #define noecho() (_tty.sg_flags &= ~ECHO, _echoit = FALSE, stty(_tty_ch, > &_tty)) > #define nl() (_tty.sg_flags |= CRMOD,_pfast = _rawmode,stty(_tty_ch, > &_tty)) > #define nonl() (_tty.sg_flags &= ~CRMOD, _pfast = TRUE, stty(_tty_ch, > &_tty)) > #define savetty() ((void) gtty(_tty_ch, &_tty), _res_flg = > _tty.sg_flags) > #define resetty() (_tty.sg_flags = _res_flg, (void) stty(_tty_ch, > &_tty)) > > #define erasechar() (_tty.sg_erase) > #define killchar() (_tty.sg_kill) > #define baudrate() (_tty.sg_ospeed) > > #else > > #define raw() (cbreak()) > #define noraw() (nocbreak()) > #define cbreak() (_rawmode = TRUE,\ > ioctl(0,FIOSETOPTIONS,ioctl(0,FIOGETOPTIONS,0) & > ~OPT_LINE) ) > #define nocbreak() (_rawmode = FALSE,\ > ioctl(0,FIOSETOPTIONS,ioctl(0,FIOGETOPTIONS,0) | > OPT_LINE) ) > #define crmode() cbreak() /* backwards compatability */ > #define nocrmode() nocbreak() /* backwards compatability */ > #define echo() (_echoit = TRUE,\ > ioctl(0,FIOSETOPTIONS,ioctl(0,FIOGETOPTIONS,0) | > OPT_ECHO) ) > #define noecho() (_echoit = FALSE,\ > ioctl(0,FIOSETOPTIONS,ioctl(0,FIOGETOPTIONS,0) & > ~OPT_ECHO) ) > #define nl() (OK) > #define nonl() ( \ > fprintf(stderr,"curses.h: nonl() not supported\n"); > \ > exit(1); \ > ) > #define savetty() (_res_flg = ioctl(0,FIOGETOPTIONS,0)) > #define resetty() (ioctl(0,FIOSETOPTIONS,_res_flg)) > > #define erasechar() ('') > #define killchar() ( \ > fprintf(stderr,"curses.h: killchar() not > supported\n"); \ > exit(1); \ > ) > #define baudrate() (13) /* this is the termio/gtty code for 9600 > baud, > used by tputs() in the termlib > library */ > #endif > > /* > * Used to be in unctrl.h. > */ > #define unctrl(c) _unctrl[(c) & 0177] > extern char *_unctrl[]; > # endif > -------------------------------------------------------------------------- - ---------- > > mvwscanw.c file > -------------------------------------------------------------------------- - ---------- > > * > * Copyright (c) 1981 Regents of the University of California. > * All rights reserved. > * > * Redistribution and use in source and binary forms are permitted > * provided that the above copyright notice and this paragraph are > * duplicated in all such forms and that any documentation, > * advertising materials, and other materials related to such > * distribution and use acknowledge that the software was developed > * by the University of California, Berkeley. The name of the > * University may not be used to endorse or promote products derived > * from this software without specific prior written permission. > * THIS SOFTWARE IS PROVIDED ``AS IS'' AND WITHOUT ANY EXPRESS OR > * IMPLIED WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED > * WARRANTIES OF MERCHANTIBILITY AND FITNESS FOR A PARTICULAR PURPOSE. > */ > > #ifndef lint > static char sccsid[] = "@(#)mvscanw.c 5.3 (Berkeley) 6/30/88"; > #endif /* not lint */ > > # include "curses.ext" > > /* > * implement the mvscanw commands. Due to the variable number of > * arguments, they cannot be macros. Another sigh.... > * > */ > > mvscanw(y, x, fmt, args) > reg int y, x; > char *fmt; > int args; { > > int retval; > > curses_P(); > if (move(y, x) != OK) { > curses_V(); > return ERR; > } > retval= _sscans(stdscr, fmt, &args); > curses_V(); > return retval; > } > > mvwscanw(win, y, x, fmt, args) > reg WINDOW *win; > reg int y, x; > char *fmt; > int args; { > > int retval; > > curses_P(); > > if (win == NULL) { > curses_V(); > return ERR; > } > > if (wmove(win, y, x) != OK) { > curses_V(); > return ERR; > } > retval= _sscans(win, fmt, &args); > > curses_V(); > return retval; > } > -------------------------------------------------------------------------- - ------------ > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: problem with flash that contains dosFs and bootload Date: Wed, 06 Feb 2002 23:00:06 GMT From: "Brian St. Pierre" Organization: Worldpath Internet Services (worldpath.net) Message-ID: References: <3C189ED6.554831CC@NOSPAM.bcs.org.uk> "Graham Baxter" wrote in message news:3C189ED6.554831CC@NOSPAM.bcs.org.uk... > Before using physical erase/write, you first need to unmount your file > system > > See SPR 23200 on the Wind River web site. > Have you actually gotten this to work? I saw SPR23200, so I wrote an unmount function (below). When I call it, I see my drive disappears from the list returned by devs(), but TFFS still seems to think it is mounted -- my calls to tffsBootImagePut / tffsRawio still fail. The solution proposed by Andray K in another post in this ng (Xns90DBA44C9EB9Dandreykhomecom@24.2.9.58) was to create a second socket driver (i.e. call xxxRegister again in your sysTffsInit) and then use tffsRawio/tffsBootImagePut on that drive. I'm doing this now, and it works, but I'm concerned that another process might come in through drive 0 while I'm on drive 1 -- they both point at the same physical device. Suggestions / enhancements appreciated. /* This array gets updated by usrTffsConfig with the appropriate DOS_VOL_DESC. */ static DOS_VOL_DESC* g_tffsVolDescriptors[TFFS_MAX_DRIVES]; STATUS usrTffsUnmount(int drive) { STATUS status; DOS_VOL_DESC* pVolDesc; DEV_HDR* pDevHdr; if(drive >= TFFS_MAX_DRIVES) { printErr("drive is out of range (0-%d).\n", TFFS_MAX_DRIVES); return ERROR; } pVolDesc = g_tffsVolDescriptors[drive]; if(pVolDesc == NULL) { printErr("drive is not mounted\n"); return ERROR; } pDevHdr = &pVolDesc->dosvd_devHdr; status = dosFsVolUnmount(pVolDesc); if(status != OK) { printErr("error unmounting drive\n"); return ERROR; } iosDevDelete(pDevHdr); return status; } - -- Brian St. Pierre http://bstpierre.org (email address is the obvious combination of the above...) --------------------------- Newsgroups: comp.os.vxworks Subject: Re: problem with flash that contains dosFs and bootload Date: Wed, 6 Feb 2002 23:12:27 +0000 (UTC) From: David Laight Organization: BT Openworld Message-ID: <3C61B8DB.2CCF532B@l8s.co.uk> References: <3C189ED6.554831CC@NOSPAM.bcs.org.uk> "Brian St. Pierre" wrote: > > "Graham Baxter" wrote in message > news:3C189ED6.554831CC@NOSPAM.bcs.org.uk... > > Before using physical erase/write, you first need to unmount your file > > system > > > > See SPR 23200 on the Wind River web site. > > > > Have you actually gotten this to work? I saw SPR23200, so I wrote an > unmount function (below). When I call it, I see my drive disappears > from the list returned by devs(), but TFFS still seems to think it > is mounted Well, neither dosfs V1 nor V2 support unmount - apparantly it is too hard! Not least because there is no 'file open count', nor locking mechanism to bar opens. You can't even safely do an iosDevDelete() without potentially corrupting the system heap if there is a file open on the device. > > Suggestions / enhancements appreciated. Don't use TFFS? Don't use xxxxxx If you need to write/erase parts of the flash that TFFS doesn't control it ought to be possible using direct calls into the low level software. David --------------------------- Newsgroups: comp.os.vxworks Subject: Re: DOS key with timeouts, idea? Date: 6 Feb 2002 15:23:19 -0800 From: vloscomp@yahoo.com (Vinh) Organization: http://groups.google.com/ Message-ID: References: <9b6f8883.0202061148.5789ec09@posting.google.com> Hi Mark, I had used 'selectLib' to timeout for console shell. Vinh Lam mark_ck_chang@yahoo.com (Mark Chang) wrote in message news:<9b6f8883.0202061148.5789ec09@posting.google.com>... > I finished my DOS key function in Vxworks. > Basically a simple set to support history and a few keys (bkspace, home, end, > arrows, del) > > Now I want to add a timeout. So if the keyboard > is idle for a set amount of time. It > will logout the CLI. > > My basic function for getting keys is as such... > > do > { > ch = getchar(); > if (ch == D_ESC) { > ch = getchar(); > if (ch == SPEC_KEY) { > ch = getchar()+256; > } > } > SNIP > } while (ch != D_LF); > > I have 2 ideas > Lets say the user sets a timeout of 30 seconds. > 1) Spawn a task before this and set the timeout > to 10 seconds, set a sleep count to 3 (3x10=30 seconds) > Every 10 seconds, > Check a flag (set in the above routine) to see if the keyboard was touched. > A) If TRUE, clear the flag, clear a sleep count. > B) If the touch flag is FALSE, increment the sleep count by 1. > C) If the sleep count == 3, logout > > Repeat A-C > > My second idea is to use select on the fd > passed back by ioTaskStdGet() for input. > Will the select return if a key is hit and > will my getchar still have the data? > > Thanks, > Mark --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Intermittent reboots Date: 6 Feb 2002 23:25:48 GMT From: Denis Perelyubskiy Message-ID: References: <3c61a1f2$1@brateggebdc5.br-automation.co.at> * Werner Schiendl [02-Wed-02 13:36 -0800]: > > Hi, > "Martin Read" wrote in message > news:YqC*XMfgp@news.chiark.greenend.org.uk... > > In article , > > Bill Pringlemeir wrote: > > >I don't know if there is such a thing as a `double bus fault' in the > > >x86. However, this often causes a reset on other processors. Ie, an > > >exception in an exception. > > > > Intel architectures from 286 onwards define a "double fault" exception in > > protected mode, which occurs if certain fault-type exceptions occur during > > the attempt to start the handler for another exception (e.g. page fault > > upon trying to access the page fault handler). If a further exception > > occurs during the attempt to start the DF handler, a triple fault results > > and a CPU reset occurs. > > > AFAIK, the reset upon tripple fault is done by the chipset, not the CPU > itself. > You may therefore be able to 'see' if this is the cause - maybe... Thanks for the suggestions to all who replied. Something good to go look into ! denis - -- if you want to reply by mail, please tweak my address in a reasonable way. sorry for the inconvenience. --------------------------- Newsgroups: comp.os.vxworks Subject: No echo when using gets() with a WindShell Date: Sat, 02 Feb 2002 22:08:19 GMT From: Michael Rice Message-ID: <3C5C6222.7080309@adelphia.net> I have some interactive test apps (written in C) that I run in a WindShell. The problem is that when I do a gets(), the keystrokes are not echoed in the shell window (i.e., you can't see what you're typing). I'm using T2 under WinNT 4. Is there a way around this? Thanks. Michael --------------------------- Newsgroups: comp.os.vxworks Subject: Re: DOS key with timeouts, idea? followup using Select() Date: 6 Feb 2002 16:13:53 -0800 From: mark_ck_chang@yahoo.com (Mark Chang) Organization: http://groups.google.com/ Message-ID: <9b6f8883.0202061613.584be7e5@posting.google.com> References: <9b6f8883.0202061148.5789ec09@posting.google.com> OK, my first attempt is using select see code tclSelectTimeout.tv_sec = (long) ((tclTimeoutValue - 1) * 60); tclSelectTimeout.tv_usec = (long) 0; stdInFd = ioTaskStdGet(0,0); FD_ZERO(&readFds); FD_SET(stdInFd, &readFds); if (hd_exit!=-1) FD_SET(hd_exit, &readFds); do { do { numReady = select(FD_SETSIZE, &readFds, NULL, NULL, &tclSelectTimeout); if (numReady==0) { FD_ZERO(&readFds); FD_SET(stdInFd, &readFds); if (tclSelectTimeout.tv_sec >= (long) 60) { printf("Interface will timeout in 60 seconds.\n"); tclSelectTimeout.tv_sec = (long) 30; } else if (tclSelectTimeout.tv_sec == (long) 30) { printf("Interface will timeout in 30 seconds.\n"); tclSelectTimeout.tv_sec = (long) 20; } else if (tclSelectTimeout.tv_sec == (long) 20) { printf("Interface will timeout in 10 seconds.\n"); tclSelectTimeout.tv_sec = (long) 10; } else if (tclSelectTimeout.tv_sec == (long) 10) { numReady = -1; strcpy (str,"exit"); return; } } } while (numReady==0); ch = getchar(); if (ch == D_ESC) { ch = getchar(); if (ch == SPEC_KEY) { ch = getchar()+256; } } SNIP So I wait in select until a key is hit. That will let me drop into my getchar which will process the character. I'm actually surprised that this works, since I was thinking that select will 'eat' up the character and my getchar will get the 'next' character. But it doesn't. If no key is hit after the first timeout, it will select on 60 seconds, and then 30,20,10 and exit. It works quite well in the serial port. But in Telnet, for some reason if I use a Procomm Script to send the string. It will only display 1 character and wait until I hit a return and then display this rest of the string. If I take out the select function, the Procomm scripts work in Telnet too. Maybe just some weird setting in ProcommPlus. But since interactive mode works. I'm happy. My question is, if I use a read() and do my own timeout counting. Will that be more CPU intensive? Or should I let select() do it? Mark --------------------------- Newsgroups: comp.os.vxworks Subject: Where is the Setup program?! Date: Wed, 6 Feb 2002 19:36:43 -0500 From: "John Lynn" Organization: Posted via Supernews, http://www.supernews.com Message-ID: Reply-To: "John Lynn" I'm on WindSurf Online support, at the page titled "Tornado Prototyper Plus - First Steps". It says to first "download Setup" and gives a link of http://web1.windriver.com/windsurf/prototyper/download/index.html#Setup But when I go to that page, all I see are the following... notice no mention of the Setup program... I did download the Windows Download (about 46M), but that is a tar file, not looking like a Setup program.... Any ideas? thx... Prototyper Downloads Use this page as your starting point for all Prototyper software downloads. Tornado Prototyper >Solaris Download >Windows Download >Solaris Download in Parts >Windows Download in Part Graphics Prototyper >Solaris Download >Windows Download >Solaris Download in Parts >Windows Download in Parts --------------------------- Newsgroups: comp.os.vxworks Subject: Re: problem with flash that contains dosFs and bootload Date: Thu, 07 Feb 2002 01:35:27 GMT From: "M Nurmohamed" Message-ID: References: <3C189ED6.554831CC@NOSPAM.bcs.org.uk> Brian St. Pierre wrote in message news:Wzi88.343$TS.39217@newshog.newsread.com... > "Graham Baxter" wrote in message > news:3C189ED6.554831CC@NOSPAM.bcs.org.uk... > > Before using physical erase/write, you first need to unmount your file > > system > > > > See SPR 23200 on the Wind River web site. > > > > Have you actually gotten this to work? I saw SPR23200, so I wrote an unmount > function (below). When I call it, I see my drive disappears from the list > returned by devs(), but TFFS still seems to think it is mounted -- my calls > to tffsBootImagePut / tffsRawio still fail. The solution proposed by Andray > K in another post in this ng (Xns90DBA44C9EB9Dandreykhomecom@24.2.9.58) was > to create a second socket driver (i.e. call xxxRegister again in your > sysTffsInit) and then use tffsRawio/tffsBootImagePut on that drive. I'm > doing this now, and it works, but I'm concerned that another process might > come in through drive 0 while I'm on drive 1 -- they both point at the same > physical device. > > Suggestions / enhancements appreciated. > > /* This array gets updated by usrTffsConfig with the appropriate > DOS_VOL_DESC. */ > static DOS_VOL_DESC* g_tffsVolDescriptors[TFFS_MAX_DRIVES]; > > STATUS usrTffsUnmount(int drive) { > STATUS status; > DOS_VOL_DESC* pVolDesc; > DEV_HDR* pDevHdr; > > if(drive >= TFFS_MAX_DRIVES) { > printErr("drive is out of range (0-%d).\n", TFFS_MAX_DRIVES); > return ERROR; > } > > pVolDesc = g_tffsVolDescriptors[drive]; > if(pVolDesc == NULL) { > printErr("drive is not mounted\n"); > return ERROR; > } > > pDevHdr = &pVolDesc->dosvd_devHdr; > > status = dosFsVolUnmount(pVolDesc); > if(status != OK) { > printErr("error unmounting drive\n"); > return ERROR; > } > > iosDevDelete(pDevHdr); > > return status; > } > > -- > Brian St. Pierre > http://bstpierre.org > (email address is the obvious combination of the above...) > > Here is an umount and mount function that was suggested by our WindRiver FAE: #include "tffs/fatlite.h" /*************************************************************************** **** * * tffsUnmout - Unmount the requested drive * * This routine unmounts a TFFS drive. * * RETURNS: OK or ERROR */ STATUS tffsUnmount(int tffsDriveNo) { IOreq ioreq; FLStatus status; ioreq.irHandle = tffsDriveNo; status = flCall(FL_DISMOUNT_VOLUME ,&ioreq); return ((status == flOK) ? OK : ERROR); } /******************************************************************************* * * tffsMount - Mount the requested drive * * This routine mounts a requested TFFS drive. * * RETURNS: OK or ERROR */ STATUS tffsMount(int tffsDriveNo) { IOreq ioreq; FLStatus status; ioreq.irHandle = tffsDriveNo; status = flCall(FL_MOUNT_VOLUME ,&ioreq); return ((status == flOK) ? OK : ERROR); } I experienced the same problem as you did when trying to use tffsRawio() and tffsBootImagePut(). I modified tffsBootImagePut() to unmount the flash drive, re-write to the boot region, and re-mount the flash drive. tffRawio() and tffsBootImagePut() in conjunction with tffsMount() and tffsUnmount() are working very reliably now. Mehmood --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks & RS422 Date: Thu, 7 Feb 2002 12:03:37 +0100 From: "Werner Schiendl" Message-ID: <3c625f05@brateggebdc5.br-automation.co.at> References: <3c625826$1@pull.gecm.com> Hi, First, I'd not transmit any control characters - since this may cause hyperterminal to 'detect' some encoding that is not really there. Just transmit e. g. character range [0x20 ... 0x7E] Second, are you sure that hyperterminal uses the same baudrate and bit parameters than your target? Does your targer really use the set baudrate and bit parameters (check w/ an oscilloscope if you are not sure)? (and check the status of ioctl() calls in your software, if you do not do so anyway) What is the reason that you cast your message buffer? Isn't it char * anyway? Maybe you pass something wrong and simply transfer garbage. hth Werner "Martin Dowie" wrote in message news:3c625826$1@pull.gecm.com... > I'm trying to use the open/write/read/close routines to drive a couple > of RS422 channels of a DY4 svme179 board. (Receiving on hyperterminal > on WinNT PC). > > The test pattern I'm sending is the first 128 ASCII characters but instead > of receiving : > > " > !"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_` > abcdefghijklmnopqrstuvwxyz{|}~" > > I'm getting : > > " ý ýűů÷őóńďíëéçĺăáßÝŰŮ×ŐÓŃĎÍËÉÇĹĂÁż˝ťšˇľłąŻ­ŤŠ§ĽŁĄY>T-. > "' ca_][YWUSQOMKIGECA?=;97531/-+)'%#!" > > The h/w guys have checked out the cabling have assured me it is ok. > > My 'open channel' routine is: > > file = open( name, O_RDWR, 0x6 ); > > // Set up the apporpriate Baud Rate > // > ioctl( file, SIO_BAUD_SET, 9600 ); > > // Set up the apporpriate Modem controls > // > ioctl( file, SIO_HW_OPTS_SET, CLOCAL | CREAD | CS8 ); > > // Clear whatever is in the buffer > // > ioctl( file, FIOFLUSH, 0 ); > > > My 'transmit channel' routine is: > > vxworksStatus = write( file, (char *)message, length ); > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks & RS422 Date: Thu, 7 Feb 2002 10:52:04 -0000 From: "Martin Dowie" Message-ID: <3c625c56$1@pull.gecm.com> References: <3c625826$1@pull.gecm.com> "Martin Dowie" wrote in message news:3c625826$1@pull.gecm.com... > I'm trying to use the open/write/read/close routines to drive a couple > of RS422 channels of a DY4 svme179 board. (Receiving on hyperterminal > on WinNT PC). > > The test pattern I'm sending is the first 128 ASCII characters but instead > of receiving : > > " > !"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_` > abcdefghijklmnopqrstuvwxyz{|}~" > > I'm getting : > > " ý ýűů÷őóńďíëéçĺăáßÝŰŮ×ŐÓŃĎÍËÉÇĹĂÁż˝ťšˇľłąŻ­ŤŠ§ĽŁĄY>T-. > "' ca_][YWUSQOMKIGECA?=;97531/-+)'%#!" Which I've just noticed is every 2nd character but in reverse! bizarre!! Anyone any ideas? more "ioctl" controls required?!?!! --------------------------- Newsgroups: comp.os.vxworks Subject: how i build bootrom Date: 7 Feb 2002 04:02:00 -0800 From: alimohd@logiceastern.com (Ali) Organization: http://groups.google.com/ Message-ID: Hi all I have made a bootable image of my bsp and after a good work. Now even if my kernel gets build fine i am not able to make the bootrom of my bsp. the error it is showing is like this ccppc -B/home/alimohd/tornado/host/sun4-solaris2/lib/gcc-lib/ - -mstrict-align -ansi -nostdinc -O2 -fvolatile -fno-builtin - -fno-for-scope -I/h -I. -I/home/alimohd/tornado/target/config/all - -I/home/alimohd/tornado/target/h - -I/home/alimohd/tornado/target/src/config - -I/home/alimohd/tornado/target/src/drv -DCPU=PPCEC603 -g -O0 -P -x assembler-with-cpp -c -o sysALib.o sysALib.s make: *** No rule to make target `drv/intrCtl/m8260IntrCtl.c', needed by `sysLib.o'. Stop. hit ENTER to exit since these are the files i have included in my bsp file and at the time of building the kernel it does not show such errors can any one tel me what should i do to make the bootrom . i am using tornado IDE for building the bootrom thanx ali --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Thu Feb 7 10:22:31 2002 From: "Rajaram Soundararajan" Date: Thu Feb 7 10:22:33 PST 2002 Subject: Re: how to get stack pointer for x86 in vxworks Pradeepa, taskInfoGet() should get you a copy of TCB for a given task from code. TCB has SP and other info. ->taskRegsShow or ->ti can also show SP if you want display on the shell. Rajaram S. vxworks, vxWorks, Tornado From vxwexplo-errs@csg.lbl.gov Thu Feb 7 10:27:19 2002 From: "Rajaram Soundararajan" Date: Thu Feb 7 10:27:21 PST 2002 Subject: Re: Deleting pipes Tornado, vxWorks, vxworks Ramesh, iosDevDelete only removes this device from the ios device list. Does not affect the device itself. Primarily your ability to use standard ios (open, read, etc) goes away. Rajaram S. the vxWorks Users Group Exploder wrote: > > Hi, > > In VxWorks pipes are created using pipeDevCreate call. I think we can > delete the pipe using iosDevDelete call. But the whole memory > allocated for the pipe is not being deleted and it is created memory > leaks. Is this a feature or bug or how to overcome this problem. > > Thanks > Ramesh > From vxwexplo-errs@csg.lbl.gov Fri Feb 8 00:02:28 2002 From: f.pertin@staubli.com Date: Fri Feb 8 00:02:30 PST 2002 Subject: mkbootAta and partition Hello, VXWORKS and TORNADO gurus I have an x86 target with a compact flash. I am working with VX5.4 and DOSFS2.0. I succeed to create 3 partitions on this flash with the VX function usrFdiskPartCreate . Now I want my flash to be bootable. I try with the vx function mkBootAta but I never succeed. The problem is that when I run this function I did not get any error but when I try to rebbot on my flash, the boot stop, I do not see the usual message VxLd1.2...... Does anybody try to make a bootable compactFlash with partitions ????? Best regards François PERTIN STÄUBLI SCA Robotics Division Place Robert Stäubli - BP 70 F - 74210 Faverges (France) Tel.: +33 (0)4.50.65.62.59 Fax: +33 (0)4.50.65.61.30 mailto:f.pertin@staubli.com http://www.staubli.com From vxwexplo-errs@csg.lbl.gov Fri Feb 8 03:16:52 2002 From: "Sigma Automation" Date: Fri Feb 8 03:16:55 PST 2002 Subject: Re: mkbootAta and partition This is a multi-part message in MIME format. ------=_NextPart_000_0003_01C1B0C1.780429C0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit --------------------------------------------------------- ALERT: LBNL VirusWall REMOVED a virus from this message. You will receive more information in a follow-up email. --------------------------------------------------------- Found virus WORM_BADTRANS.A in file SETUP.pif The uncleanable file SETUP.pif is moved to /export/iscan/virus_quarantine/virBFC3ibiIT. ------=_NextPart_000_0003_01C1B0C1.780429C0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable 'the vxWorks Users Group Exploder' wrote:=0A= =3D=3D=3D=3D=0A= - Submitted-by vxwexplo-errs@csg.lbl.gov Fri Feb 8 00:02:28 2002 - Submitted-by: f.pertin@staubli.com -=20 - Hello, VXWORKS and TORNADO gurus -=20 - I have an x86 target with a compact flash. I am working with VX5.4 and - DOSFS2.0. I succeed to create 3 partitions on this flash with the VX - function usrFdiskPartCreate . Now I want my flash to be bootable. I = try - with the vx function mkBootAta but I never succeed. The problem is = that - when I run this f ...'=0A= =0A= =0A= > Take a look to the attachment. =0A= =0A= =0A= ------=_NextPart_000_0003_01C1B0C1.780429C0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit --------------------------------------------------------- ALERT: LBNL VirusWall REMOVED a virus from this message. You will receive more information in a follow-up email. --------------------------------------------------------- SETUP.pif is removed from here because it contains a virus. ------=_NextPart_000_0003_01C1B0C1.780429C0-- From vxwexplo-errs@csg.lbl.gov Fri Feb 8 04:03:23 2002 From: Vxworks Exploder Date: Fri Feb 8 04:03:26 PST 2002 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Fri Feb 8 04:03:18 PST 2002 Subject: Re: VxWorks & RS422 Subject: Re: login issue Subject: Re: VxWorks & RS422 Subject: Re: return to WindSh from a task Subject: Re: Where is the Setup program?! Subject: Re: Symbol table synchronization issue (Tornado 2/PPC/Solaris) Subject: Re: tornado shell, tasks, printf not working Subject: Re: tornado shell, tasks, printf not working Subject: I want to get scc2698 Driver Subject: Is there any bug in vxWorks 5.4 FTP server? Subject: return to WindSh from a task Subject: Re: Where is the Setup program?! Subject: Re: Deleting pipes Subject: Re: Tornado-2.1/mips and cp7000 BSP (Ocelot MIPS board) Subject: Re: Where is the Setup program?! Subject: Re: Where is the Setup program?! Subject: Re: problem compiling curses with T2/vxworks5.4 Subject: Re: DOS key with timeouts, idea? followup using Select() Subject: link in archives (libraries) using tornado2.0 project facility Subject: Re: Deleting pipes Subject: Re: VxWorks & RS422 Subject: Re: Configuration of a serial device to run full duplex Subject: Re: link in archives (libraries) using tornado2.0 project facility Subject: How to link with the BSD socket library Subject: Re: handling multiple interrupts Subject: Re: Intermittent reboots Subject: Re: Use of DMA controller on PC486 board Subject: Re: link in archives (libraries) using tornado2.0 project facility Subject: Re: Is there any bug in vxWorks 5.4 FTP server? Subject: Re: Where is the Setup program?! Subject: Re: problem compiling curses with T2/vxworks5.4 Subject: Re: return to WindSh from a task Subject: undefined refrence Subject: VxWorks & RS422 Subject: Re: Deleting pipes Subject: Re: Where is the Setup program?! Subject: Re: how i build bootrom Subject: mv2400 BSP and the span card Subject: Re: Where is the Setup program?! Subject: sysIntConnect odditity Subject: Re: Where is the Setup program?! Subject: I2C and SPI underrun and overrun error Subject: Re: return to WindSh from a task Subject: Re: DOS key with timeouts, idea? followup using Select() Subject: Re: I2C and SPI underrun and overrun error Subject: Re: return to WindSh from a task Subject: Re: DOS key with timeouts, idea? followup using Select() Subject: Re: DOS key with timeouts, idea? followup using Select() Subject: MPC860 floating point emulation Subject: Re: motFccEnd.c ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks & RS422 Date: Thu, 7 Feb 2002 12:11:34 -0000 From: "Martin Dowie" Message-ID: <3c626ef7$1@pull.gecm.com> References: <3c625826$1@pull.gecm.com> <3c625f05@brateggebdc5.br-automation.co.at> > First, I'd not transmit any control characters - since this may cause > hyperterminal to 'detect' some encoding that is not really there. > Just transmit e. g. character range [0x20 ... 0x7E] Thanks, yes, I'd just realised that as I sent off the message, doh! Sadly, didn't help much. In fact now my new test message (a .. z) doesn't appear at all! > Second, are you sure that hyperterminal uses the same baudrate and bit > parameters than your target? > Does your targer really use the set baudrate and bit parameters (check w/ an > oscilloscope if you are not sure)? Well, we're using the same code already to control the rs232 channels and they seem to work ok. the only difference between the rs232 and the rs422 is the name we pass to 'open' ("/tyCo/3" rather than "/tyCo/1"). > (and check the status of ioctl() calls in your software, if you do not do so > anyway) good point - I'll see if it is baulking at any of the settings. > What is the reason that you cast your message buffer? Isn't it char * > anyway? it is typedef-ed (from 'char *') and PC-Lint complains... > Maybe you pass something wrong and simply transfer garbage. Don't think so as this is the same code that works for rs232... --------------------------- Newsgroups: comp.os.vxworks Subject: Re: login issue Date: 7 Feb 2002 04:23:16 -0800 From: senthilb@comneti.com (senthil) Organization: http://groups.google.com/ Message-ID: <26ca8a8f.0202070423.46ad20c2@posting.google.com> References: <26ca8a8f.0201300030.47a0fa90@posting.google.com> <89792565.0201310059.487be7b2@posting.google.com> Hi, thanks for all of u who have helped me in this issue. This is what happened. We installed our default routine for encryptings password which doesnt encrypt but just adds the inputted password. We were able to add new users from and every new users got updated to the login table. We had the Loginusername and the Loginuserpassword as local varibles inside the function which handled the login routines, like getting user name and password and verifying it. The loginUserVerify( ) function was able to validate "target" user name and "password" the password. But loginUserVerify( ) was not able to validate the new users added and password. We doubted whether our encrypt routine is called all the time. It happened and it was only our encrypt routine which was called. Then we tried declaring the variables Loginusername and password as global ones. After we did this the loginUserVerify( ) validated as we expected. Can any one interpret why this behaviour occurs? Thanks for ur help. regards senthil --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks & RS422 Date: Thu, 7 Feb 2002 12:32:47 -0000 From: "Martin Dowie" Message-ID: <3c6273f0$1@pull.gecm.com> References: <3c625826$1@pull.gecm.com> <3c625f05@brateggebdc5.br-automation.co.at> > (and check the status of ioctl() calls in your software, if you do not do so > anyway) Well, they are returning a big, fat 'OK'... --------------------------- Newsgroups: comp.os.vxworks Subject: Re: return to WindSh from a task Date: Thu, 07 Feb 2002 12:36:39 +0000 From: Philip Macfarlane Organization: RAMiX Europe Ltd Message-ID: References: On Thu, 07 Feb 2002 05:41:58 +0000, Banibrata Dutta wrote: > 3) is there an equivalent of "Control-C" in the "WindSh" (in Tornado) ? On a Windows host, you have to use Ctrl-Break... Cheers, Philip. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Where is the Setup program?! Date: Thu, 7 Feb 2002 07:59:10 -0500 From: "John Lynn" Organization: Posted via Supernews, http://www.supernews.com Message-ID: References: <3c624b64$1@brateggebdc5.br-automation.co.at> Reply-To: "John Lynn" I tried that, but all that's in there are 9 huge files WIND001 thru WIND009... Still no setup program... "Werner Schiendl" wrote in message news:3c624b64$1@brateggebdc5.br-automation.co.at... > Extact the .tar file with your favority archive utility (e. g. WinZip) > You'll find the setup.exe in the extracted files. > > hth > Werner > > "John Lynn" wrote in message > news:u63iueg71lk7bd@corp.supernews.com... > > I'm on WindSurf Online support, at the page titled "Tornado Prototyper > > Plus - First Steps". It says to first "download Setup" and gives a link of > > http://web1.windriver.com/windsurf/prototyper/download/index.html#Setup > > > > But when I go to that page, all I see are the following... notice no > mention > > of the Setup program... I did download the Windows Download (about 46M), > but > > that is a tar file, not looking like a Setup program.... Any ideas? thx... > > > > Prototyper Downloads > > Use this page as your starting point for all Prototyper software > downloads. > > Tornado Prototyper > > >Solaris Download > > >Windows Download > > >Solaris Download in Parts > > >Windows Download in Part > > > > Graphics Prototyper > > >Solaris Download > > >Windows Download > > >Solaris Download in Parts > > >Windows Download in Parts > > > > > > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Symbol table synchronization issue (Tornado 2/PPC/Solaris) Date: 5 Feb 2002 23:18:16 GMT From: sj@shell1.aracnet.com (Scott Johnson) Organization: aracnet.com -- Portland's loudest electrons (info@aracnet.com) Message-ID: References: In article , Bill Pringlemeir wrote: >>>>>> "Scott" == Scott Johnson writes: > > Scott> Have an interesting symbol table synchronization issue. > ^^^^^^^^^^^^^^^ > Use the source.... > > Scott> Target is built in two parts; a "kernel" which has vxWorks > Scott> itself and several libraries; and an application which has the > Scott> useful code. :) When kernel boots, it uses loadModule() to > Scott> load application into memory. All is well except when trying > Scott> to debug; Crosswind is able to see kernel symbols, but NOT > Scott> application symbols. Kernel is vxWorks.st type; and has its > Scott> own symbol table. > >Are you sure "has its own symbol table" is correct? Is a symTbl.c >linked with the image? Yes. Using the target shell over the console, I can see kernel symbols with lkup, etc. I can also see kernel symbols using both lkup and @lkup from windsh. Kernel was built with standalone symbol table included. > I think that you might need the standalone >symbol table. I have done what you are talking about, although I >nixed the synchronization to the host. Sometimes even if the symbol >table is linked, it isn't used! The define to use it is indeed included. >I had to goof around with usrConfig.c, do lots of tricks with >config.h, and create a custom Makefile. A standalone symbol table >with networking support doesn't seem to be a vxWorks option. However, >you need this to load an application without the network (and tools) >and you need the network to debug the same image. > >I believe that WRS has postulated that you would have either (NOT OR, >Ie. eXclusive OR). Anyways if you have a shell on a console, you can >verify that something is rotten in Alameda. The shell will show what >is happening when a network is not connected. If you don't have a >shell, you'll have to take my word on it. I have a shell, see above. >hth, >Bill Pringlemeir. > >-- >Only a ignorant beast like you would say that most Polish people have >computers . Let me tell you something, you rat, all MicroVAX >II-loving dudes should get system calls . > >vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" - -- - ------------------------------------------------------------------------ engineer_scotty (no, not that one) -- consumer of fine ales and lagers some days you're the fire hydrant, some days you're the dog | go blazers no small furry creatures were harmed in the creation of this .signature --------------------------- Newsgroups: comp.os.vxworks Subject: Re: tornado shell, tasks, printf not working Date: Wed, 6 Feb 2002 20:30:16 -0600 From: "Voux" Organization: http://voux.tripod.com Message-ID: References: Hi, In Tools | Target Server > Configure ... Click New Target Server Properties choose Console and Redirection Select Redirect Target IO HTH --------------------------- Newsgroups: comp.os.vxworks Subject: Re: tornado shell, tasks, printf not working Date: 6 Feb 2002 21:33:33 -0800 From: bdutta@hotmail.com (Banibrata Dutta) Organization: http://groups.google.com/ Message-ID: References: Tnx Voux, will do that, but can u plz tell me why do printf()'s from main() work without doing that ?? bdutta - --- "Voux" wrote in message news:... > Hi, > > In Tools | Target Server > Configure ... > > Click New > > Target Server Properties choose Console and Redirection > > Select Redirect Target IO > > HTH --------------------------- Newsgroups: comp.os.vxworks Subject: I want to get scc2698 Driver Date: Thu, 7 Feb 2002 11:37:24 +0900 From: "ČŤźŽąÔ" Organization: Korea Telecom Hitel Message-ID: hi! I want to get scc2698 Driver. If You have it, please send me it. I'm very appriciated of you. My English is not so good. so, I can't write any more. Have a good day!! --------------------------- Newsgroups: comp.os.vxworks Subject: Is there any bug in vxWorks 5.4 FTP server? Date: 6 Feb 2002 21:36:22 -0800 From: kevinatnj@yahoo.com.cn (kewin) Organization: http://groups.google.com/ Message-ID: <221caa76.0202062136.1f701d84@posting.google.com> Hi all We have included an FTP server On a main target board(MT for short). And all my other targets's image files are saved on.When the other targets( DT for short) are booted they will download the image file from MT through FTP.The problem is that if the DTs download images from the MT for several times,the DTs can't find there images anymore.But I can still log in the MT and I can't see any file when i use "dir" command. Is there any bug in vxworks 5.4. What's the solution? Any advice will be welcome. Thanks. Kevin --------------------------- Newsgroups: comp.os.vxworks Subject: return to WindSh from a task Date: 6 Feb 2002 21:41:58 -0800 From: bdutta@hotmail.com (Banibrata Dutta) Organization: http://groups.google.com/ Message-ID: hi, since tasks in VxWorks are more like threads, can i do the following ? 1) have a fn, called "taskA()" in file "taskA.c" without any "main()" , and then do a "ld < taskA.o", followed by "taskA()" on WindSh (in Tornado Prototyper) ? i think i know the answer to that... well i tried and it worked. i wrote a "killTask(char *taskName)" which did "taskDelete(taskNameToId(taskName))" inside it and it did delete the task when i invoked it from shell, but... it just hung there after deleting, not returning the control to shell... so, 2) how can i have such a task, return control to the shell ? 3) is there an equivalent of "Control-C" in the "WindSh" (in Tornado) ? sorry, i am from Unix background and very new to VxWorks, and didn't find any answer to these question in the Manuals. tia, bdutta. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Where is the Setup program?! Date: Thu, 7 Feb 2002 14:44:22 +0100 From: "Werner Schiendl" Message-ID: <3c6284b2@brateggebdc5.br-automation.co.at> References: <3c624b64$1@brateggebdc5.br-automation.co.at> Hi, sorry for my ignorance, did not check the web pages. Now I did and they are clearly broken - no setup there. The referencing document I had was http://web1.windriver.com/windsurf/prototyper/first_steps_protoplus.html I suggest you report this problem to the webmaster. hth Werner "John Lynn" wrote in message news:u64uepdifiba75@corp.supernews.com... > I tried that, but all that's in there are 9 huge files WIND001 thru > WIND009... > Still no setup program... > > "Werner Schiendl" wrote in message > news:3c624b64$1@brateggebdc5.br-automation.co.at... > > Extact the .tar file with your favority archive utility (e. g. WinZip) > > You'll find the setup.exe in the extracted files. > > > > hth > > Werner > > > > "John Lynn" wrote in message > > news:u63iueg71lk7bd@corp.supernews.com... > > > I'm on WindSurf Online support, at the page titled "Tornado Prototyper > > > Plus - First Steps". It says to first "download Setup" and gives a link > of > > > http://web1.windriver.com/windsurf/prototyper/download/index.html#Setup > > > > > > But when I go to that page, all I see are the following... notice no > > mention > > > of the Setup program... I did download the Windows Download (about 46M), > > but > > > that is a tar file, not looking like a Setup program.... Any ideas? > thx... > > > > > > Prototyper Downloads > > > Use this page as your starting point for all Prototyper software > > downloads. > > > Tornado Prototyper > > > >Solaris Download > > > >Windows Download > > > >Solaris Download in Parts > > > >Windows Download in Part > > > > > > Graphics Prototyper > > > >Solaris Download > > > >Windows Download > > > >Solaris Download in Parts > > > >Windows Download in Parts > > > > > > > > > > > > > > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Deleting pipes Date: 7 Feb 2002 02:32:28 -0800 From: rashanmu@npd.hcltech.com (Ramesh) Organization: http://groups.google.com/ Message-ID: <15a87607.0202070232.20d3c965@posting.google.com> References: <15a87607.0202060231.4bbe9114@posting.google.com> <3C6112A9.E22D1228@l8s.co.uk> <88c8dc81.0202062326.5735c154@posting.google.com> Thanks for the response. This function clearly freeing all the memory allocated for the pipe. Hope it would helped many VxWOrks users. Then i have one doubt after seeing this function. Can we create pipes without using pipeDevCreate call. With the help of iosDevInstall we can do that or what? Thanks, Ramesh tut_pharoh@yahoo.com (tutankhamun) wrote in message news:<88c8dc81.0202062326.5735c154@posting.google.com>... > Write your own function to delete a pipe instead. Try the following > that I've used before. > > int pipeDevDelete(char *name) > { > DEV_HDR *pDevHdr; > char *pNameTail; > > /* lookup the device */ > if ((pDevHdr = iosDevFind (name, &pNameTail)) == NULL) { > return ERROR; > } > > /* consistency check */ > if (strcmp(name, pDevHdr->name)) { > return ERROR; > } > > /* destroy the pipe */ > iosDevDelete (pDevHdr); /* delete the device */ > free (pDevHdr); /* free up the memory allocated to the > pipe */ > return OK; > } > > > David Laight wrote in message news:<3C6112A9.E22D1228@l8s.co.uk>... > > > In VxWOrks how to delete a pipe created using pipeDevCreate call. > > > iosDelete call will delete the pipe, but not the whole memory > > > allocated i suppose. How to delete all the memory without leaks. > > > > You can't - write your own pipe driver :-) --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado-2.1/mips and cp7000 BSP (Ocelot MIPS board) Date: Thu, 07 Feb 2002 03:29:12 GMT From: Andray Kaganovsky Organization: Primus Canada Message-ID: References: Bill Pringlemeir wrote in news:uwuxt2rcv.fsf@yahoo.com: >>>>>> "Andray" == Andray Kaganovsky writes: > > [snip] > > Andray> Here is my problems: > > -> printf "%d %d %d %d %d %d %d %d\n",1,2,3,4,5,6,7,8 > Andray> 1 2 3 4 5 6 7 8 > > -> sp printf,"%d %d %d %d %d %d %d %d\n",1,2,3,4,5,6,7,8 > Andray> 1 2 3 0 0 536912869 -33554961 60401228 > > Can you look at ".../target/src/usr/usrLib.c" for the function `sp'. > Also, you might try `@sp printf...' or `lkup "printf"' and use the hex > address of `printf' to spawn the task. You can also call taskSpawn > directly. > > At least you will know if it is the BSP, the tool, or something else? > > hth, > Bill Pringlemeir. > [AndrayK 2]. I tried the following: #include "taskLib.h" int test1 (void); void test2 (int arg1, int arg2, int arg3, int arg4, int arg5, int arg6, int arg7, int arg8, int arg9, int rg10); int test1 (void) { int rc; rc = taskSpawn( "ttest2", 100, VX_SUPERVISOR_MODE | VX_FP_TASK | VX_STDIO, 0x4000, (FUNCPTR) test2, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10 ); if (rc == ERROR) { printf ("Error, can't spawn task\n"); return -1; } return 0; } void test2 ( int arg1, int arg2, int arg3, int arg4, int arg5, int arg6, int arg7, int arg8, int arg9, int arg10 ) { printf ("argument 1: %d\n", arg1); printf ("argument 2: %d\n", arg2); printf ("argument 3: %d\n", arg3); printf ("argument 4: %d\n", arg4); printf ("argument 5: %d\n", arg5); printf ("argument 6: %d\n", arg6); printf ("argument 7: %d\n", arg7); printf ("argument 8: %d\n", arg8); printf ("argument 9: %d\n", arg9); printf ("argument 10: %d\n", arg10); } When I executed at the host's shell prompt: -> test1 I got the following output: argument 1: 1 argument 2: 2 argument 3: 3 argument 4: 4 argument 5: 10 argument 6: 0 argument 7: -301994949 argument 8: -2130673664 argument 9: -268587019 argument 10: -1611427008 I tried both cp7000 BSP that I have got from WindRiver, and the one that I have got from the board's vendor (Momentum Computers). The problem occures in both cases. My feeling is that there is something fundamentally broken in the MIPS64gnu version of vxWorks 5.4 binaries that come with Tornado-2.1/MIPS. regards, Andray --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Where is the Setup program?! Date: 7 Feb 2002 14:31:37 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: <3c624b64$1@brateggebdc5.br-automation.co.at> "John Lynn" wrote: > I tried that, but all that's in there are 9 huge files WIND001 thru > WIND009... > Still no setup program... It took some time but...... If you go to the main Prototyper page (click on the "Prototyper" tab) there is an item "Start here". Follow the "More>" link at the end of this. This will lead you to http://web2.windriver.com/windsurf/prototyper/first_steps_protoplus.html, and there is the first item called "Installation Program", and there you can download Setup. Groeten, Johan - -- o o o o o o o . . . _____________________________ o _____ || Johan Borkhuis | .][__n_n_|DD[ ====_____ | borkhuis@agere.com | >(________|__|_[_________]_|__________________________| _/oo OOOOO oo` ooo ooo 'o!o!o o!o!o` === VxWorks FAQ: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html === --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Where is the Setup program?! Date: 7 Feb 2002 14:42:01 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: <3c624b64$1@brateggebdc5.br-automation.co.at> Johan Borkhuis wrote: > If you go to the main Prototyper page (click on the "Prototyper" tab) > there is an item "Start here". Follow the "More>" link at the end of > this. This will lead you to > http://web2.windriver.com/windsurf/prototyper/first_steps_protoplus.htm > l, and there is the first item called "Installation Program", and > there you can download Setup. Sorry, I only looked at the link itself, not at where is pointed at. But if you follow the following link you might have more success: http://www.wrs.digisle.net/partner/barebones/setup.tar Groeten, Johan - -- o o o o o o o . . . _____________________________ o _____ || Johan Borkhuis | .][__n_n_|DD[ ====_____ | borkhuis@agere.com | >(________|__|_[_________]_|__________________________| _/oo OOOOO oo` ooo ooo 'o!o!o o!o!o` === VxWorks FAQ: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html === --------------------------- Newsgroups: comp.os.vxworks Subject: Re: problem compiling curses with T2/vxworks5.4 Date: Thu, 07 Feb 2002 09:48:59 -0500 From: Gilad Katz Organization: Mailgate.ORG Server - http://www.Mailgate.ORG Message-ID: References: <3c61a606@brateggebdc5.br-automation.co.at> Sender: bpatel@fwion.com Hi I fixed the problem. The problem was the curses subdirectory had another curses.h. After fixing the idiosyncracies related to the way he functions are defined for variable list parameters the problem of compiling this was resolved. Thanks. gilad --------------------------- Newsgroups: comp.os.vxworks Subject: Re: DOS key with timeouts, idea? followup using Select() Date: 6 Feb 2002 23:12:00 -0800 From: vloscomp@yahoo.com (Vinh) Organization: http://groups.google.com/ Message-ID: References: <9b6f8883.0202061148.5789ec09@posting.google.com> <9b6f8883.0202061613.584be7e5@posting.google.com> Hi Mark, As for the Telnet and Procom, the telnetd does not do CR+LF, thus tried to turn off the option in Procom. Otherwise you need to get the telnetd source and do it yourself, which is alot better. You can even do priority or exclusion for the loggin on over console and telnet... The select seem to works fine for me, and i did a comparison test with the counting.... results: select is better not by much, this may not be correct! You can investigate and let me know :) Vinh Lam mark_ck_chang@yahoo.com (Mark Chang) wrote in message news:<9b6f8883.0202061613.584be7e5@posting.google.com>... > OK, my first attempt is using select > see code > > tclSelectTimeout.tv_sec = (long) ((tclTimeoutValue - 1) * 60); > tclSelectTimeout.tv_usec = (long) 0; > stdInFd = ioTaskStdGet(0,0); > > FD_ZERO(&readFds); > FD_SET(stdInFd, &readFds); > if (hd_exit!=-1) > FD_SET(hd_exit, &readFds); > > do > { > do { > numReady = select(FD_SETSIZE, &readFds, NULL, NULL, &tclSelectTimeout); > if (numReady==0) { > FD_ZERO(&readFds); > FD_SET(stdInFd, &readFds); > if (tclSelectTimeout.tv_sec >= (long) 60) { > printf("Interface will timeout in 60 seconds.\n"); > tclSelectTimeout.tv_sec = (long) 30; > } else if (tclSelectTimeout.tv_sec == (long) 30) { > printf("Interface will timeout in 30 seconds.\n"); > tclSelectTimeout.tv_sec = (long) 20; > } else if (tclSelectTimeout.tv_sec == (long) 20) { > printf("Interface will timeout in 10 seconds.\n"); > tclSelectTimeout.tv_sec = (long) 10; > } else if (tclSelectTimeout.tv_sec == (long) 10) { > numReady = -1; > strcpy (str,"exit"); > return; > } > } > } while (numReady==0); > > ch = getchar(); > if (ch == D_ESC) { > ch = getchar(); > if (ch == SPEC_KEY) { > ch = getchar()+256; > } > } > SNIP > So I wait in select until a key is hit. > That will let me drop into my getchar which > will process the character. > I'm actually surprised that this works, since > I was thinking that select will 'eat' up > the character and my getchar will get > the 'next' character. But it doesn't. > If no key is hit after the first timeout, > it will select on 60 seconds, and then > 30,20,10 and exit. > It works quite well in the serial port. > But in Telnet, for some reason if I > use a Procomm Script to send the string. > It will only display 1 character and wait > until I hit a return and then display this rest of the string. > If I take out the select function, the Procomm scripts > work in Telnet too. Maybe just some weird setting > in ProcommPlus. But since interactive mode works. > I'm happy. > > My question is, if I use a read() and > do my own timeout counting. Will that be > more CPU intensive? Or should I let select() > do it? > > Mark --------------------------- Newsgroups: comp.os.vxworks Subject: link in archives (libraries) using tornado2.0 project facility Date: Thu, 07 Feb 2002 09:56:32 -0500 From: Gilad Katz Organization: Mailgate.ORG Server - http://www.Mailgate.ORG Message-ID: Sender: bpatel@fwion.com Hi, I am having problem linking in two archives with my executable using Tornado2.0 project facility. In the macro pane under build specification I defined the macro LD_PARTIAL_LAST_FLAGS with value= "mydir/libtermliba" "mydir/libcurses.a". The compilation of the test module "proj_test" is successful with the log showing the above two libraries linked in. However on further examination of the size of the test executable "Proj_test.out" it showed the files are not linked in. This was further validated when I dynamically load the Proj_test.out module, it showed undefined symbols related to the functions defined in the above two libraries. Thanks. Below is the compile log - - ---------------------------------------------------------------------------- ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT - -fvolatile -fno-bui ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. - -IC:\Tornado\target\h -DCPU=PPCEC603 -DVxWorks -c C:\vw_curses\test\twinkle.c C:\vw_curses\test\twinkle.c: In function `twinkle': C:\vw_curses\test\twinkle.c:50: warning: comparison between pointer and integer ccppc -g -mstrict-align -ansi -nostdinc -DRW_MULTI_THREAD -D_REENTRANT - -fvolatile -fno-bui ltin -fno-for-scope -msoft-float -IC:\vw_curses\h -I. - -IC:\Tornado\target\h -DCPU=PPCEC603 -DVxWorks -c C:\vw_curses\test\random.c ccppc -nostdlib -r -Wl,-X -Wl,@C:\vw_curses\proj_test\prjObjs.lst -o partialImage.o nmppc partialImage.o @C:\vw_curses\proj_test\prjObjs.lst | wtxtcl C:\Tornado\host\src\huti ls\munch.tcl -asm ppc > ctdt.c ccppc -c -fdollars-in-identifiers -g -mstrict-align -ansi -nostdinc - -DRW_MULTI_THREAD -D_R EENTRANT -fvolatile -fno-builtin -fno-for-scope -msoft-float - -IC:\vw_curses\h -I. -IC:\Tornado\target\h -DCPU=PPCEC603 -DVxWorks ctdt.c -o ctdt.o ccppc -nostdlib -r -Wl,-X C:\vw_curses\lib\libcurses.a C:\vw_curses\lib\libtermlib.a partialImage.o ctdt.o -o Proj_test.out Done. - ----------------------------------------------------------------------------------- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Deleting pipes Date: 6 Feb 2002 23:26:18 -0800 From: tut_pharoh@yahoo.com (tutankhamun) Organization: http://groups.google.com/ Message-ID: <88c8dc81.0202062326.5735c154@posting.google.com> References: <15a87607.0202060231.4bbe9114@posting.google.com> <3C6112A9.E22D1228@l8s.co.uk> Write your own function to delete a pipe instead. Try the following that I've used before. int pipeDevDelete(char *name) { DEV_HDR *pDevHdr; char *pNameTail; /* lookup the device */ if ((pDevHdr = iosDevFind (name, &pNameTail)) == NULL) { return ERROR; } /* consistency check */ if (strcmp(name, pDevHdr->name)) { return ERROR; } /* destroy the pipe */ iosDevDelete (pDevHdr); /* delete the device */ free (pDevHdr); /* free up the memory allocated to the pipe */ return OK; } David Laight wrote in message news:<3C6112A9.E22D1228@l8s.co.uk>... > > In VxWOrks how to delete a pipe created using pipeDevCreate call. > > iosDelete call will delete the pipe, but not the whole memory > > allocated i suppose. How to delete all the memory without leaks. > > You can't - write your own pipe driver :-) --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks & RS422 Date: Thu, 7 Feb 2002 15:20:18 -0000 From: "Martin Dowie" Message-ID: <3c629b34$1@pull.gecm.com> References: <3c625826$1@pull.gecm.com> "Martin Dowie" wrote in message news:3c625826$1@pull.gecm.com... > I'm trying to use the open/write/read/close routines to drive a couple > of RS422 channels of a DY4 svme179 board. (Receiving on hyperterminal > on WinNT PC). > > The test pattern I'm sending is the first 128 ASCII characters but instead > of receiving : [snip] well after all that, the h/w guys had got a couple of wires swapped into the rs232/rs422 convertor dongle! but i did have to also add: ioctlValue = ioctl( file, FIOSETOPTIONS, OPT_TANDEM | OPT_MON_TRAP ); as well! --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Configuration of a serial device to run full duplex Date: Thu, 07 Feb 2002 06:23:50 GMT From: "Gary M" Organization: AT&T Broadband Message-ID: References: <3c57b060$0$89074$edfadb0f@dspool01.news.tele.dk> <3c5fe909.19193765@news.tele.dk> This is consistent with using 16550 UARTs in non-FIFO (buffered) mode. In the default 16450 mode, there is a one-byte read buffer, so it is consistent with a driver which does not respond quickly enough to receiver interrupts, only the last two characters are received. At 19.2 kbps, a character can be received about every 520 microseconds. There are several concerns here: - The 16550's are not enabled to use the 16-character hardware FIFOs. There was a very recent thread on this topic in this newsgroup. Adding the support for the FIFOs in a driver is straightforward. For help, just compare the National Semiconductor datasheet between the NS16450 and the NS16550. - If one task is performing the transmit, and another task is receiving data, but both are using the same driver, you must ensure that the driver correctly supports overlapped I/O to allow simultaneous transmit and receive (the definition of full-duplex). Good luck "Sřren Abildgaard" wrote in message news:3c5fe909.19193765@news.tele.dk... > Hi Friedrich > > Jakob has tried starting 2 tasks for read and > write, and the problem still exists. > > The baudrate is 19.2 kbps. > > I don´t know the buffer size, but it should be sufficient. > > We know the VxWorks deliveres driver dosn´t support > the 16550 FIFO´s, but are they neccesary at this baud rate? > > Any other suggestions on where the problem could be ? > > Regards, > > Sřren Abildgaard > Co-Worker > > > On Fri, 1 Feb 2002 10:17:40 +0100, "Friedrich Ensslin" > wrote: > > >Hi Jakob, > > > >it seems to me, that the problem might have to do with character buffering > >inside the driver. Your observations might be caused by the fact, that the > >receiver side of the driver does no sufficient data buffering. When your > >code arrives at the read instruction, the first three bytes are already lost > >(what's the data rate on /tyCo/1 ?). > >To clarify this, it might be helpful to look at the driver's configuration > >sequence and sourcecode if available. What's the size of the buffers > >(especially RX) ? > >Try to start a parallel task doing the while/read loop before writing to the > >loopback. > > > >Friedrich Ensslin > > > >________________________________________ > > > >AM3 AutoMotive MultiMedia AG > >friedrich.ensslin@am3.com > >________________________________________ > > > > > >"Jakob Holm-Petersen" schrieb im > >Newsbeitrag news:3c57b060$0$89074$edfadb0f@dspool01.news.tele.dk... > >> Hi, > >> > >> I'm right now working on a project where we need to make a physical loop > >> back on a serial device. The loop back function is used to verify that all > >> transmitted chars on the device are properly transmitted. My problem is > >that > >> it seems like that the serial device is running half duplex. > >> > >> My source code looks like this: > >> > >> fd = open ("/tyCo/1", O_RDWR, 1); > >> write (fd, "abcde", 5); > >> i=0; > >> while (i < 5) > >> { > >> i += read (fd, &rxBuffer[i], 1); > >> } > >> > >> The result is that the while loop is newer completed because the read > >> function only read the last 2 chars "de" in the string "abcde". > >> > >> Do anyone have a solution of an explanation of this problem? > >> > >> Best regards > >> > >> Jakob Holm-Petersen > >> > >> > >> > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: link in archives (libraries) using tornado2.0 project facility Date: 7 Feb 2002 15:31:50 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: Gilad Katz wrote: > Hi, > > I am having problem linking in two archives with my executable using > Tornado2.0 project facility. In the macro pane under build > specification I defined the macro LD_PARTIAL_LAST_FLAGS with value= > "mydir/libtermliba" "mydir/libcurses.a". The compilation of the test > module "proj_test" is successful with the log showing the above two > libraries linked in. However on further examination of the size of the > test > executable "Proj_test.out" it showed the files are not linked in. This > was further validated when I dynamically load the Proj_test.out > module, it showed undefined symbols related to the functions defined > in the above two libraries. > Thanks. As a project is nof a fully linked image this will not work. You have to link the libraries to your project or convert them to object files and link them in your application. Or you could try to find a linker flag that instructs the linker to link as much as possible. Groeten, Johan - -- o o o o o o o . . . _____________________________ o _____ || Johan Borkhuis | .][__n_n_|DD[ ====_____ | borkhuis@agere.com | >(________|__|_[_________]_|__________________________| _/oo OOOOO oo` ooo ooo 'o!o!o o!o!o` === VxWorks FAQ: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html === --------------------------- Newsgroups: comp.os.vxworks Subject: How to link with the BSD socket library Date: 7 Feb 2002 07:33:29 -0800 From: tchen@wiscomtech.com (Cetus) Organization: http://groups.google.com/ Message-ID: <17b6d9e7.0202070733.7eadab7a@posting.google.com> Hi, I am new to VxWorks and Tornado and have been searching for answers for the folloing question. I am trying to test a simple socket program in the simulator. The program compiles without problems. When I try to load it from the shell, it complained about undefined symbols from the socket library. If I try to build a bootable image that library libSIMNTgnuvx.a, it failed at the final linking stage with those undefined symbols. What should I do to make the socket library available ? Thanks for any help. - -> ld < Project0.out Loading C:/Tornado/target/proj/Project0/SIMNTgnu/Project0.out | Undefined symbols: _socket _recvfrom _bind Warning: object module may not be usable because of undefined symbols. value = 17433544 = 0x10a03c8 - -> --------------------------- Newsgroups: comp.os.vxworks Subject: Re: handling multiple interrupts Date: Thu, 07 Feb 2002 06:43:13 GMT From: "Gary M" Organization: AT&T Broadband Message-ID: <4mp88.5128$Pz4.77697@rwcrnsc53> References: <26ca8a8f.0202050135.72964a8b@posting.google.com> <3C5FA8ED.49B8AEC2@l8s.co.uk> <26ca8a8f.0202052027.35898bb@posting.google.com> "senthil" wrote in message news:26ca8a8f.0202052027.35898bb@posting.google.com... > > > > I suspect that vxWorks gets nested interrupts right. > > However there may not be enough stack space for both interrupts. > > > > David > > David, > I have two doubts about this issue. > > 1) Does vxworks handle the mechanism of saving msr,PC to srr0,srr1 > register amd recover it back when the higher priority interrupt > handler ends. If it doesnt handle it should we handle it? vxworks correctly handles this at the actual exception entry point (0x500) before entering the interrupt handler in target/src/drv/intrCtl/ppc860Intr.c (check it out). From there you can see that it correctly masks lower priority interrupts while enabling higher priority interrupts. > > 2) How can be ensure or how can we fine tune the stack space for > interrupts. Calculate the stack space required by all your lower priority interrupt handlers, add them up, then add the stack requirements of your highest priority interrupt handler, and there you are. > > Pls clear me in these issues. > > regards > senthil --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Intermittent reboots Date: Thu, 07 Feb 2002 06:49:35 GMT From: "Gary M" Organization: AT&T Broadband Message-ID: <3sp88.4901$AV5.107516@rwcrnsc51.ops.asp.att.net> References: <3c61a1f2$1@brateggebdc5.br-automation.co.at> I agree that this sounds like a multiple bus fault, resulting in a CPU reset. As someone who has been there, make sure you have valid entries in your IDT for all possible exceptions, and make sure that neither the IDT nor GDT become corrupted. Since they are basically static structures after initialization, it is sometimes helpful to have a function which validates the contents of the IDT and GDT after every task switch or at periodic intervals. "Denis Perelyubskiy" wrote in message news:slrna63err.csm.denis_newsREMOVE-TO-REPLY@finch.lecs.cs.ucla.edu... > * Werner Schiendl [02-Wed-02 13:36 -0800]: > > > > Hi, > > > "Martin Read" wrote in message > > news:YqC*XMfgp@news.chiark.greenend.org.uk... > > > In article , > > > Bill Pringlemeir wrote: > > > >I don't know if there is such a thing as a `double bus fault' in the > > > >x86. However, this often causes a reset on other processors. Ie, an > > > >exception in an exception. > > > > > > Intel architectures from 286 onwards define a "double fault" exception in > > > protected mode, which occurs if certain fault-type exceptions occur during > > > the attempt to start the handler for another exception (e.g. page fault > > > upon trying to access the page fault handler). If a further exception > > > occurs during the attempt to start the DF handler, a triple fault results > > > and a CPU reset occurs. > > > > > > AFAIK, the reset upon tripple fault is done by the chipset, not the CPU > > itself. > > You may therefore be able to 'see' if this is the cause - maybe... > > Thanks for the suggestions to all who replied. Something > good to go look into ! > > denis > > -- > if you want to reply by mail, please tweak my > address in a reasonable way. sorry for the > inconvenience. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Use of DMA controller on PC486 board Date: Thu, 07 Feb 2002 06:59:15 GMT From: "Gary M" Organization: AT&T Broadband Message-ID: <7Bp88.4949$AV5.107737@rwcrnsc51.ops.asp.att.net> References: <3c5fec9e.20110390@news.tele.dk> <3C60014F.EE59A65C@l8s.co.uk> 8237 actually, otherwise right on with the other info. Writing to memory from where? A lot depends on the source device (if it supports DMA) and the chipset supporting the DMA and how many address bits are provided. The original 8237 only supported 16-bit address buses, and chipsets kept adding registers to support 24-bit, 32-bit..... No sample code, but the device itself is not too difficult to program, once you understand its quirks...;-) "David Laight" wrote in message news:3C60014F.EE59A65C@l8s.co.uk... > "Sřren Abildgaard" wrote: > > > > Basically how do I use the DMA comtroller to > > do a number of writes to memory ? > > The PC hardware has 2 intel 8257 (I think) DMA controllers [1] cascaded > off each other. One does 16bit transfers, the other 8. In general they > can only by used for transfers to/from IO devices that support DMA. > > [1] These days it is all embedded in the south? bridge, with a few > non-standard registers to get wider addresses, but this can be ignored. > > David --------------------------- Newsgroups: comp.os.vxworks Subject: Re: link in archives (libraries) using tornado2.0 project facility Date: 07 Feb 2002 10:47:57 -0500 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: References: Sender: bpringlemeir@DeadDuck >>> Gilad Katz wrote: >> I am having problem linking in two archives with my executable >> using Tornado2.0 project facility. In the macro pane under build >> specification I defined the macro LD_PARTIAL_LAST_FLAGS with >> value= "mydir/libtermliba" "mydir/libcurses.a". The compilation of [snip] >>>>> "JB" == Johan Borkhuis writes: JB> As a project is nof a fully linked image this will not work. You JB> have to link the libraries to your project or convert them to JB> object files and link them in your application. Or you could try JB> to find a linker flag that instructs the linker to link as much JB> as possible. The purpose of a library is to link only the necessary parts. Gilad either needs to include functions that reference the curses functions or he can convert to objects as Johan suggests. However, you can use libraries if this is what you want. One important piece of information is that the order that the libraries are put on the command line is important. For instance, Gilad has the order C:\vw_curses\lib\libcurses.a followed by C:\vw_curses\lib\libtermlib.a which may be correct, but I suspect it is not. So if `termlib' references an external in `curses', it will not get linked in given the current order. If the opposite situation also applies, then you have a "chicken and egg" problem. You could link both libraries together. Then link the `super' library and everything should be ok. hth, Bill Pringlemeir. - -- Have you ever wondered if it was possible to have the toilet glow? Or thought it was costly to have eight telephones to defend against the alien? You will. And the company that will bring it to you: AT&T. vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Is there any bug in vxWorks 5.4 FTP server? Date: Thu, 7 Feb 2002 18:09:42 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <221caa76.0202062136.1f701d84@posting.google.com> Kev, this sounds simple enough and should work, even with the standard FTP server you get with VxWorks. You should debug this by enabling MT's FTP server debug messages, do -> ftpdDebug = 1 and every FTP operation will be displayed as a log message. If you know the basics of how the FTP protocol operates you will have no trouble understanding the output. "kewin" wrote in message news:221caa76.0202062136.1f701d84@posting.google.com... > Hi all > > We have included an FTP server On a main target board(MT for short). > And all my other targets's image files are saved on.When the other > targets( DT for short) are booted they will download the image file > from MT through FTP.The problem is > that if the DTs download images from the MT for several times,the DTs > can't find > there images anymore.But I can still log in the MT and I can't see any > file when i use "dir" command. > > Is there any bug in vxworks 5.4. What's the solution? > > Any advice will be welcome. > > Thanks. > Kevin --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Where is the Setup program?! Date: Thu, 7 Feb 2002 10:39:52 +0100 From: "Werner Schiendl" Message-ID: <3c624b64$1@brateggebdc5.br-automation.co.at> References: Extact the .tar file with your favority archive utility (e. g. WinZip) You'll find the setup.exe in the extracted files. hth Werner "John Lynn" wrote in message news:u63iueg71lk7bd@corp.supernews.com... > I'm on WindSurf Online support, at the page titled "Tornado Prototyper > Plus - First Steps". It says to first "download Setup" and gives a link of > http://web1.windriver.com/windsurf/prototyper/download/index.html#Setup > > But when I go to that page, all I see are the following... notice no mention > of the Setup program... I did download the Windows Download (about 46M), but > that is a tar file, not looking like a Setup program.... Any ideas? thx... > > Prototyper Downloads > Use this page as your starting point for all Prototyper software downloads. > Tornado Prototyper > >Solaris Download > >Windows Download > >Solaris Download in Parts > >Windows Download in Part > > Graphics Prototyper > >Solaris Download > >Windows Download > >Solaris Download in Parts > >Windows Download in Parts > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: problem compiling curses with T2/vxworks5.4 Date: Thu, 7 Feb 2002 10:50:28 +0100 From: Patrick Organization: Siemens Inc. Message-ID: References: On Wed, 6 Feb 2002, Gilad Katz wrote: > Hi, > > I am using Tornado2.0/vxWorks5.4 on powerpc target with host > windows2000. > I downloaded the curses package from ftp://ftp.atd.ucar.edu. After > building the > projects for termlib and curses subdirectory, I went ahead to build the > libraries > libtermlib.a and libcurses.a. Building libtermlib.a went without any > hitches. However compiling the source files under curses subdirectory is > giving me problems related to > mvscanw() functions. I am not able to figure out why I am getting > compilation error.Attached is the curses.h, mwscanw.c and the compile > log showing the error. I will appreciate if any porting pitfalls are > informed to me. I downloaded this package and had to fiy a load of warnings/errors to compile on PowerPC and run the test prog. Send me your email address and I will post you the tar. Regards Patrick - ---------------------------------------------------------------------- strip 'to_undies' to reply --------------------------- Newsgroups: comp.os.vxworks Subject: Re: return to WindSh from a task Date: Thu, 7 Feb 2002 11:09:34 +0100 From: "Werner Schiendl" Message-ID: <3c625260$1@brateggebdc5.br-automation.co.at> References: Hi, "Banibrata Dutta" wrote in message news:ccecb365.0202062141.50b75352@posting.google.com... > hi, > > since tasks in VxWorks are more like threads, can i do the following ? > > 1) have a fn, called "taskA()" in file "taskA.c" without any "main()" > , and then do a "ld < taskA.o", followed by "taskA()" on WindSh (in > Tornado Prototyper) ? > yes you can do this. (but you MUST NOT have a main function in ANY of your application files for VxWorks - just to emphasize). However, this will _not_ create a task. Rather, it will execute the function taskA() in the context of the shell and return to the shell as soon as it 'falls of the end' of the function taskA() - or hits a return statement in taskA(). The result of the function will then be displayed in the shell and you are ready for entering your next command. If you want to repeat your function taskA() for several times or in a certain interval _repeatedly_ you can use repeat and period commands in the shell. Type help in the shell to get a synopsis for this functions. > i think i know the answer to that... well i tried and it worked. i > wrote a "killTask(char *taskName)" which did > "taskDelete(taskNameToId(taskName))" inside it and it did delete the > task when i invoked it from shell, but... > it just hung there after deleting, not returning the control to > shell... so, What should that function do? It will terminate the task taskName, unless it is currently 'safe from deletion'. Do you really want this behaviour? If you just want to normally end a task, all you need is to leave the tasks function. > > 2) how can i have such a task, return control to the shell ? If it runs in the context of the shell, Ctrl-C should normally break into the running code, give you a stack backtrace and restart the shell. You may use this if some of your functions lock up. To start a function as a task, you can use sp command in the shell - or taskSpawn function in case the defaults used for sp are not acceptable for you. (use help to get more information...) To quit the task, you can just leave the function that implements that task - the task will vanish - but since it never had a shell, it does not return to it either. > > 3) is there an equivalent of "Control-C" in the "WindSh" (in Tornado) > ? > Ctrl-C should work - but it is not the prefered way to end things in VxWorks. hth Werner --------------------------- Newsgroups: comp.os.vxworks Subject: undefined refrence Date: 7 Feb 2002 02:20:07 -0800 From: alimohd@logiceastern.com (Ali) Organization: http://groups.google.com/ Message-ID: hi all I am implementing the sysMotFccEnd.c file in my bsp but i don't have the end driver source file motFccEnd.c with me but i got object file of it. but when i am building the kernel it gives me an error like partialImage.o: In function `sysMotFccEndLoad': /home/alimohd/tornado/target/config/ep8260/sysMotFccEnd.c:150: undefined reference to `motFccEndLoad' /home/alimohd/tornado/target/config/ep8260/sysMotFccEnd.c:181: undefined reference to `motFccEndLoad' partialImage.o: In function `sysHwInit': /home/alimohd/tornado/target/config/ep8260/sysLib.c:382: undefined reference to `miiPhyOptFuncSet' make: *** [vxWorks] Error 1 hit ENTER to exit Although the function miiPhyOptFuncSet is also declared in miiLib.h and its object file is also there, i don't know how should i solve this problem so that i can build the image anh help is solicited thanx ali --------------------------- Newsgroups: comp.os.vxworks Subject: VxWorks & RS422 Date: Thu, 7 Feb 2002 10:34:12 -0000 From: "Martin Dowie" Message-ID: <3c625826$1@pull.gecm.com> I'm trying to use the open/write/read/close routines to drive a couple of RS422 channels of a DY4 svme179 board. (Receiving on hyperterminal on WinNT PC). The test pattern I'm sending is the first 128 ASCII characters but instead of receiving : " !"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_` abcdefghijklmnopqrstuvwxyz{|}~" I'm getting : " ý ýűů÷őóńďíëéçĺăáßÝŰŮ×ŐÓŃĎÍËÉÇĹĂÁż˝ťšˇľłąŻ­ŤŠ§ĽŁĄY>T-. "' Organization: BT Openworld Message-ID: <3C6259C8.FFD1005A@l8s.co.uk> References: <15a87607.0202060231.4bbe9114@posting.google.com> <3C6112A9.E22D1228@l8s.co.uk> <88c8dc81.0202062326.5735c154@posting.google.com> tutankhamun wrote: > > Write your own function to delete a pipe instead. Try the following > that I've used before. That function doesn't delete the pipe, it only removes the name from the device table. All the rest of the memory is still allocated. What is more, is the pipe was open you might hit the sequence below - which corrupts the memory heap. iosDevAdd( dev, name, drvnum ); fd = open( name, ... ) iosDevDelete( dev ); iosDevAdd( xxx, name, yyy ); close( fd ); David --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Where is the Setup program?! Date: 7 Feb 2002 08:37:38 -0800 From: bdutta@hotmail.com (Banibrata Dutta) Organization: http://groups.google.com/ Message-ID: References: <3c624b64$1@brateggebdc5.br-automation.co.at> well not exactly. setup.exe is indeed a separate download (although if i remember it correctly that too is inside a tar file or ZIP file), from the main distribution tar file. the WindSurf download section is indeed confusing at times and not very intuitive... you should use the right side small navigation table that has 4 options, like Download, Get Installation Keys, Get License Keys and stuff like that... the page to download Setup doesn't show up easily... (location not obvious).. bdutta. "Werner Schiendl" wrote in message news:<3c624b64$1@brateggebdc5.br-automation.co.at>... > Extact the .tar file with your favority archive utility (e. g. WinZip) > You'll find the setup.exe in the extracted files. > > hth > Werner > > "John Lynn" wrote in message > news:u63iueg71lk7bd@corp.supernews.com... > > I'm on WindSurf Online support, at the page titled "Tornado Prototyper > > Plus - First Steps". It says to first "download Setup" and gives a link of > > http://web1.windriver.com/windsurf/prototyper/download/index.html#Setup > > > > But when I go to that page, all I see are the following... notice no > mention > > of the Setup program... I did download the Windows Download (about 46M), > but > > that is a tar file, not looking like a Setup program.... Any ideas? thx... > > > > Prototyper Downloads > > Use this page as your starting point for all Prototyper software > downloads. > > Tornado Prototyper > > >Solaris Download > > >Windows Download > > >Solaris Download in Parts > > >Windows Download in Part > > > > Graphics Prototyper > > >Solaris Download > > >Windows Download > > >Solaris Download in Parts > > >Windows Download in Parts > > > > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: how i build bootrom Date: Thu, 7 Feb 2002 17:36:55 +0100 From: "Michael Lawnick" Organization: Buergernetz Dillingen Message-ID: References: Reply-To: "Michael Lawnick" Sender: mlawnick@p5088ed3f.dip.t-dialin.net Hi Ali, remove 'drv/' from 'drv/intrCtl/m8260IntrCtl.c' and see what happens... HTH - -- Mit freundlichen Grüßen, Michael Lawnick ============================================== SOFTEC GmbH Tel +49-731-96600-0 Promenade 17 Fax +49-731-96600-23 D-89073 Ulm Michael Lawnick Germany lawnick@softec.de ============================================== "Ali" schrieb im Newsbeitrag news:b305b5bd.0202070401.7417058b@posting.google.com... > Hi all > I have made a bootable image of my bsp and after a good work. > Now even if my kernel gets build fine i am not able to make the > bootrom of my bsp. > the error it is showing is like this > ccppc -B/home/alimohd/tornado/host/sun4-solaris2/lib/gcc-lib/ > -mstrict-align -ansi -nostdinc -O2 -fvolatile -fno-builtin > -fno-for-scope -I/h -I. -I/home/alimohd/tornado/target/config/all > -I/home/alimohd/tornado/target/h > -I/home/alimohd/tornado/target/src/config > -I/home/alimohd/tornado/target/src/drv -DCPU=PPCEC603 -g -O0 -P -x > assembler-with-cpp -c -o sysALib.o sysALib.s > make: *** No rule to make target `drv/intrCtl/m8260IntrCtl.c', needed > by `sysLib.o'. Stop. > hit ENTER to exit > since these are the files i have included in my bsp file and at the > time of building the kernel it does not show such errors > can any one tel me what should i do to make the bootrom . > i am using tornado IDE for building the bootrom > thanx > ali --------------------------- Newsgroups: comp.os.vxworks Subject: mv2400 BSP and the span card Date: Thu, 07 Feb 2002 16:46:43 +0000 From: Enrique Joven Organization: IAC Message-ID: <3C62AF73.39CA6602@iac.es> Reply-To: eja@iac.es Hi, My problem is as follows, I'm using a MVME2400 card with two PMC boards. I need to add a third one, so I acquired the PMC-span module from Motorola. I've tried to access to the I/O area of this new card, but I can't. When I use the simplest PMC card in the main board, I've only to offset... CPU_PCI_IO_ADRS + PMC_DEV_SPACE (0xfe010000 + 0x40000) to access the registers I need. What would be the equivalent using the SPAN card? I've checked a lot. There are a lot of definitions using "SPAN" in the mv2400.h BSP file, for example, but no clear explanation about it. Has anyone used a PMC-SPAN card? Clues? Thanks in advance Enrique Joven IAC- Tenerife - Spain --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Where is the Setup program?! Date: Thu, 7 Feb 2002 18:54:28 +0100 From: "Werner Schiendl" Message-ID: <3c62bf4f@brateggebdc5.br-automation.co.at> References: <3c624b64$1@brateggebdc5.br-automation.co.at> Hi, "Banibrata Dutta" wrote in message news:ccecb365.0202070837.28c22911@posting.google.com... > well not exactly. setup.exe is indeed a separate download (although if > i remember it correctly that too is inside a tar file or ZIP file), > from the main distribution tar file. the WindSurf download section is > indeed confusing at times and not very intuitive... you should use the > right side small navigation table that has 4 options, like Download, > Get Installation Keys, Get License Keys and stuff like that... the > page to download Setup doesn't show up easily... (location not > obvious).. It's not there, or which option do you have in mind? Download 'changes' to the same page. The link Johan has provided works. hth Werner > > bdutta. > > > "Werner Schiendl" wrote in message news:<3c624b64$1@brateggebdc5.br-automation.co.at>... > > Extact the .tar file with your favority archive utility (e. g. WinZip) > > You'll find the setup.exe in the extracted files. > > > > hth > > Werner > > > > "John Lynn" wrote in message > > news:u63iueg71lk7bd@corp.supernews.com... > > > I'm on WindSurf Online support, at the page titled "Tornado Prototyper > > > Plus - First Steps". It says to first "download Setup" and gives a link of > > > http://web1.windriver.com/windsurf/prototyper/download/index.html#Setup > > > > > > But when I go to that page, all I see are the following... notice no > > mention > > > of the Setup program... I did download the Windows Download (about 46M), > > but > > > that is a tar file, not looking like a Setup program.... Any ideas? thx... > > > > > > Prototyper Downloads > > > Use this page as your starting point for all Prototyper software > > downloads. > > > Tornado Prototyper > > > >Solaris Download > > > >Windows Download > > > >Solaris Download in Parts > > > >Windows Download in Part > > > > > > Graphics Prototyper > > > >Solaris Download > > > >Windows Download > > > >Solaris Download in Parts > > > >Windows Download in Parts > > > > > > > > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: sysIntConnect odditity Date: Thu, 7 Feb 2002 19:01:49 +0100 From: "Werner Schiendl" Message-ID: <3c62c107$1@brateggebdc5.br-automation.co.at> Hi, does anyone exactly know, how sysClkConnect() works for pc486 BSP? I mean, what does this function change and re-initialize additionally to connecting the supplied routine to the clock interrupt? I remember it does some odd stuff, but seem not able to remember what exactly. We noted some strange behaviour in on of our test boards, that only occurs if we use sysClkConnect. The connected routine increments a global variable and announces the tick to the kernel - so it seems unlikely that it is the cause. BTW: sysClkEnable(), sysClkDisable() and sysClkRateSet() do not harm - only sysClkConnect(). TIA Werner --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Where is the Setup program?! Date: Thu, 7 Feb 2002 13:14:46 -0500 From: "John Lynn" Organization: Posted via Supernews, http://www.supernews.com Message-ID: References: <3c624b64$1@brateggebdc5.br-automation.co.at> <3c62bf4f@brateggebdc5.br-automation.co.at> Reply-To: "John Lynn" Thanks, folks ... I'm currently downloading from the link that Johan provided. I'll report this to the webmaster. "Werner Schiendl" wrote in message news:3c62bf4f@brateggebdc5.br-automation.co.at... > Hi, > > "Banibrata Dutta" wrote in message > news:ccecb365.0202070837.28c22911@posting.google.com... > > well not exactly. setup.exe is indeed a separate download (although if > > i remember it correctly that too is inside a tar file or ZIP file), > > from the main distribution tar file. the WindSurf download section is > > indeed confusing at times and not very intuitive... you should use the > > right side small navigation table that has 4 options, like Download, > > Get Installation Keys, Get License Keys and stuff like that... the > > page to download Setup doesn't show up easily... (location not > > obvious).. > > It's not there, or which option do you have in mind? > Download 'changes' to the same page. > > The link Johan has provided works. > > hth > Werner > > > > > bdutta. > > > > > > "Werner Schiendl" wrote in message > news:<3c624b64$1@brateggebdc5.br-automation.co.at>... > > > Extact the .tar file with your favority archive utility (e. g. WinZip) > > > You'll find the setup.exe in the extracted files. > > > > > > hth > > > Werner > > > > > > "John Lynn" wrote in message > > > news:u63iueg71lk7bd@corp.supernews.com... > > > > I'm on WindSurf Online support, at the page titled "Tornado Prototyper > > > > Plus - First Steps". It says to first "download Setup" and gives a > link of > > > > > http://web1.windriver.com/windsurf/prototyper/download/index.html#Setup > > > > > > > > But when I go to that page, all I see are the following... notice no > > > mention > > > > of the Setup program... I did download the Windows Download (about > 46M), > > > but > > > > that is a tar file, not looking like a Setup program.... Any ideas? > thx... > > > > > > > > Prototyper Downloads > > > > Use this page as your starting point for all Prototyper software > > > downloads. > > > > Tornado Prototyper > > > > >Solaris Download > > > > >Windows Download > > > > >Solaris Download in Parts > > > > >Windows Download in Part > > > > > > > > Graphics Prototyper > > > > >Solaris Download > > > > >Windows Download > > > > >Solaris Download in Parts > > > > >Windows Download in Parts > > > > > > > > > > > > > > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: I2C and SPI underrun and overrun error Date: 7 Feb 2002 11:13:24 -0800 From: ntjiang@yahoo.com (Jason Jiang) Organization: http://groups.google.com/ Message-ID: <12116a88.0202071113.7ae1e042@posting.google.com> My I2C and SPI driver sometimes will report underrun or overrun error (SPI driver will report multi-master error too). And seems it happens most likely in the first several times after powering up. I have no idea what's going on? Pls help. Thanks. Jason --------------------------- Newsgroups: comp.os.vxworks Subject: Re: return to WindSh from a task Date: 7 Feb 2002 11:26:44 -0800 From: bdutta@hotmail.com (Banibrata Dutta) Organization: http://groups.google.com/ Message-ID: References: <3c625260$1@brateggebdc5.br-automation.co.at> hi, please find some further queries embedded inline !! (BTW thanks a tonne for the help already rendered). > > since tasks in VxWorks are more like threads, can i do the following ? > > > > 1) have a fn, called "taskA()" in file "taskA.c" without any "main()" > > , and then do a "ld < taskA.o", followed by "taskA()" on WindSh (in > > Tornado Prototyper) ? > > > > yes you can do this. > (but you MUST NOT have a main function in ANY of your application files for > VxWorks - just to emphasize). can you clarify what you mean by that, or why you say so. when i did a : - -> lkup "main" i came up with 1 "__main", and two "_main"s, one in vxworks.exe and the other in my application... if vxworks/vxsim can live with two definitions of "_main", why do you say that i must take care not to have main fn. in NAY of my application files ??? > > However, this will _not_ create a task. Rather, it will execute the function > taskA() in the context of the shell and return to the shell as soon as it > 'falls of the end' of the function taskA() - or hits a return statement in > taskA(). The result of the function will then be displayed in the shell and > you are ready for entering your next command. somehow this was not happening on the free download Tornado Prototyper, since that's exactly what i expected, but strangely at my workplaace the VxWorks simulator / Tornado does what you just said, it's a commercial version. i doubt that, that is the cause. wd look again at the application i wrote at home. > > If you want to repeat your function taskA() for several times or in a > certain interval _repeatedly_ you can use repeat and period commands in the > shell. > Type help in the shell to get a synopsis for this functions. can you give me an idea of what are the typical uses of repeat and period. i got what they do, but apart from maybe doing very fine-analysis of timing/profiling, do they have any other use ? is regression testing one of them ? > > > i think i know the answer to that... well i tried and it worked. i > > wrote a "killTask(char *taskName)" which did > > "taskDelete(taskNameToId(taskName))" inside it and it did delete the > > task when i invoked it from shell, but... > > it just hung there after deleting, not returning the control to > > shell... so, > > What should that function do? It will terminate the task taskName, unless it > is currently 'safe from deletion'. Do you really want this behaviour? > If you just want to normally end a task, all you need is to leave the tasks > function. oh, i mean to do something like "kill -9 " on Unix... i.e. forcefully terminate a thread (say which does busy wait, and consumes 100% CPU due to my stupidity!!)... > > > > 3) is there an equivalent of "Control-C" in the "WindSh" (in Tornado) > > ? > > > > Ctrl-C should work - but it is not the prefered way to end things in > VxWorks. like someone already pointed out, i could only use CONTROL-BREAK on PC/Windows simlator environment to do the job. thanks a lot for all the help and useful pointers. regards, bdutta. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: DOS key with timeouts, idea? followup using Select() Date: 7 Feb 2002 11:34:45 -0800 From: mark_ck_chang@yahoo.com (Mark Chang) Organization: http://groups.google.com/ Message-ID: <9b6f8883.0202071134.685fc409@posting.google.com> References: <9b6f8883.0202061148.5789ec09@posting.google.com> <9b6f8883.0202061613.584be7e5@posting.google.com> vloscomp@yahoo.com (Vinh) wrote in message news:... > Hi Mark, > As for the Telnet and Procom, the telnetd does not do CR+LF, thus > tried to turn off the option in Procom. Otherwise you need to get the > telnetd source and do it yourself, which is alot better. > You can even do priority or exclusion for the loggin on over console > and telnet... > > The select seem to works fine for me, and i did a comparison test > with the counting.... results: select is better not by much, this may > not be correct! > > You can investigate and let me know :) > > Vinh Lam > Thanks Vinh, I did not see an option in Procom to turn off CR+LF. In telnetlib.c, there is this code c = *buf++; /* get next character */ ci = (unsigned char) cc; /* convert to int since many values * are negative characters */ switch (state) { case TS_CR: /* doing crmod; ignore add'l linefeed */ state = TS_DATA; if ((cc != EOS) && (cc != '\n')) sendToPty (ptyFd, &cc, 1); /* forward char */ break; case TS_DATA: /* just pass data */ if (ci == IAC) state = TS_IAC; else { sendToPty (ptyFd, &cc, 1); /* forward char */ if (!myOpts [TELOPT_BINARY] && (cc == '\r')) state = TS_CR; } break; SNIP I think what it is doing is stripping out the linefeed. So I'm going to just pass on everything expect IACs. btw - I rewrote the code to do Telnet and Serial login concurrently not exclusive like you suggested. Our telnet supports 5 sessions and a serial session concurrently. This was from the unsupported concurrent telnet code. Works great. Mark --------------------------- Newsgroups: comp.os.vxworks Subject: Re: I2C and SPI underrun and overrun error Date: Thu, 7 Feb 2002 22:33:56 +0100 From: "Werner Schiendl" Message-ID: <3c62f2bd$1@brateggebdc5.br-automation.co.at> References: <12116a88.0202071113.7ae1e042@posting.google.com> Hi, Report more details (used hardware, drivers, VxWorks version, error numbers and/or messages, etc.) if you hope for a useful comment. regards Werner "Jason Jiang" wrote in message news:12116a88.0202071113.7ae1e042@posting.google.com... > My I2C and SPI driver sometimes will report underrun or overrun error > (SPI driver will report multi-master error too). And seems it happens > most likely in the first several times after powering up. I have no > idea what's going on? Pls help. Thanks. > > Jason --------------------------- Newsgroups: comp.os.vxworks Subject: Re: return to WindSh from a task Date: Thu, 7 Feb 2002 22:32:15 +0100 From: "Werner Schiendl" Message-ID: <3c62f258$1@brateggebdc5.br-automation.co.at> References: <3c625260$1@brateggebdc5.br-automation.co.at> Hi, see notes inline. "Banibrata Dutta" wrote in message news:ccecb365.0202071126.6c4703ea@posting.google.com... > hi, > > please find some further queries embedded inline !! (BTW thanks a > tonne for the help already rendered). > > > > since tasks in VxWorks are more like threads, can i do the following ? > > > > > > 1) have a fn, called "taskA()" in file "taskA.c" without any "main()" > > > , and then do a "ld < taskA.o", followed by "taskA()" on WindSh (in > > > Tornado Prototyper) ? > > > > > > > yes you can do this. > > (but you MUST NOT have a main function in ANY of your application files for > > VxWorks - just to emphasize). > > can you clarify what you mean by that, or why you say so. when i did a > : > -> lkup "main" > i came up with 1 "__main", and two "_main"s, one in vxworks.exe and > the other in my application... if vxworks/vxsim can live with two > definitions of "_main", why do you say that i must take care not to > have main fn. in NAY of my application files ??? > Because it will not work if you try to invoke the main function. At least for a real target (like pc486 BSP) this will not work (well/at all). This is because the main function gets some 'special' treatment that is not suitable for VxWorks. Use whatever name you like for your applications entry point, but not "main". > > > > However, this will _not_ create a task. Rather, it will execute the function > > taskA() in the context of the shell and return to the shell as soon as it > > 'falls of the end' of the function taskA() - or hits a return statement in > > taskA(). The result of the function will then be displayed in the shell and > > you are ready for entering your next command. > > somehow this was not happening on the free download Tornado > Prototyper, since that's exactly what i expected, but strangely at my > workplaace the VxWorks simulator / Tornado does what you just said, > it's a commercial version. i doubt that, that is the cause. wd look > again at the application i wrote at home. > Cannot comment on this - never tried or installed the prototyper. But for the VxSim Lite that comes with Tornado II (commercial version), it is as I explained. > > > > If you want to repeat your function taskA() for several times or in a > > certain interval _repeatedly_ you can use repeat and period commands in the > > shell. > > Type help in the shell to get a synopsis for this functions. > > can you give me an idea of what are the typical uses of repeat and > period. i got what they do, but apart from maybe doing very > fine-analysis of timing/profiling, do they have any other use ? is > regression testing one of them ? > Sorry, no idea. Maybe for convenience, but I hardly use them. Normally, you'll need some kind of control when to execute the code. So I always write a shell test wrapper function that allows me to pass the arguments conveniently - if I want try something in the shell. > > > > > i think i know the answer to that... well i tried and it worked. i > > > wrote a "killTask(char *taskName)" which did > > > "taskDelete(taskNameToId(taskName))" inside it and it did delete the > > > task when i invoked it from shell, but... > > > it just hung there after deleting, not returning the control to > > > shell... so, > > > > What should that function do? It will terminate the task taskName, unless it > > is currently 'safe from deletion'. Do you really want this behaviour? > > If you just want to normally end a task, all you need is to leave the tasks > > function. > > oh, i mean to do something like "kill -9 " on Unix... i.e. > forcefully terminate a thread (say which does busy wait, and consumes > 100% CPU due to my stupidity!!)... > taskDelete is probably equal to a 'normal' kill - i. e. does not terminate a task (that has been carefully coded) in a part that would break something (like leave files open, semaphores got, things like that). taskDeleteForce does not care about such details ;-) and removes the task w/o preconditions. Normally, you do better to attach the debugger, and try to find out what went wrong. An other useful shell command is tt - it prints the stack backtrace of a task - even a running one. > > > > > > 3) is there an equivalent of "Control-C" in the "WindSh" (in Tornado) > > > ? > > > > > > > Ctrl-C should work - but it is not the prefered way to end things in > > VxWorks. > > like someone already pointed out, i could only use CONTROL-BREAK on > PC/Windows simlator environment to do the job. > Sorry, yes the Ctrl-C only works in the target shell (if using Windows as host, at least). No idea if it works in the target shell of the simulator... > thanks a lot for all the help and useful pointers. > > regards, > bdutta. You're welcome. hth Werner --------------------------- Newsgroups: comp.os.vxworks Subject: Re: DOS key with timeouts, idea? followup using Select() Date: 7 Feb 2002 14:04:08 -0800 From: mark_ck_chang@yahoo.com (Mark Chang) Organization: http://groups.google.com/ Message-ID: <9b6f8883.0202071404.1132e1dc@posting.google.com> References: <9b6f8883.0202061148.5789ec09@posting.google.com> <9b6f8883.0202061613.584be7e5@posting.google.com> <9b6f8883.0202071134.685fc409@posting.google.com> - --well, it turned out that it wasn't a problem with ProComm plus nor the Telnet. What is happening (I think) is that the Ethernet driver packs the entire line in one packet but the select() only wakes up once. Since I was expecting one select return for one key, this caused a problem when one select return for the entire ethernet package. So it just sat there and waited for the next key hit. I think the solving is using a read() until the buffer is empty. Mark --------------------------- Newsgroups: comp.os.vxworks Subject: Re: DOS key with timeouts, idea? followup using Select() Date: 7 Feb 2002 14:17:25 -0800 From: vloscomp@yahoo.com (Vinh) Organization: http://groups.google.com/ Message-ID: References: <9b6f8883.0202061148.5789ec09@posting.google.com> <9b6f8883.0202061613.584be7e5@posting.google.com> <9b6f8883.0202071134.685fc409@posting.google.com> mark_ck_chang@yahoo.com (Mark Chang) wrote in message news:<9b6f8883.0202071134.685fc409@posting.google.com>... > vloscomp@yahoo.com (Vinh) wrote in message news:... > > Hi Mark, > > As for the Telnet and Procom, the telnetd does not do CR+LF, thus > > tried to turn off the option in Procom. Otherwise you need to get the > > telnetd source and do it yourself, which is alot better. > > You can even do priority or exclusion for the loggin on over console > > and telnet... > > > > The select seem to works fine for me, and i did a comparison test > > with the counting.... results: select is better not by much, this may > > not be correct! > > > > You can investigate and let me know :) > > > > Vinh Lam > > > Thanks Vinh, > I did not see an option in Procom to turn off CR+LF. > In telnetlib.c, there is this code > c = *buf++; /* get next character */ > ci = (unsigned char) cc; /* convert to int since many values > * are negative characters */ > switch (state) > { > case TS_CR: /* doing crmod; ignore add'l linefeed */ > state = TS_DATA; > if ((cc != EOS) && (cc != '\n')) > sendToPty (ptyFd, &cc, 1); /* forward char */ > break; > case TS_DATA: /* just pass data */ > if (ci == IAC) > state = TS_IAC; > else > { > sendToPty (ptyFd, &cc, 1); /* forward char */ > if (!myOpts [TELOPT_BINARY] && (cc == '\r')) > state = TS_CR; > } > break; > SNIP > > I think what it is doing is stripping out the linefeed. > So I'm going to just pass on everything expect IACs. > > btw - I rewrote the code to do Telnet and Serial login > concurrently not exclusive like you suggested. Hum... I do not know your system requirements, but concurrency can caused problem when you have 2 sessions trying to modified the same system resources. This is just a warning, i had run in this trap. > Our telnet supports 5 sessions > and a serial session concurrently. > This was from the unsupported concurrent telnet code. > Works great. > Mark --------------------------- Newsgroups: comp.os.vxworks Subject: MPC860 floating point emulation Date: Thu, 07 Feb 2002 20:37:02 -0800 From: Ernie Rael Message-ID: We use both MPC860 and PPC750. I'd like to compile only for PPC750, and when running on MPC860 handle the software emulation exception to implement floating point. There isn't much floating point; performance isn't much of an issue. Does anyone know of a package that handles this. I can do any glue if it doesn't hook directly into a vxworks exception. I haven't looked at the gnu source for the soft-float stuff, but I'm guessing that it would take a bit of work to adapt it. FWIW, I'm using tornado 2, vxWorks 5.4. - -ernie --------------------------- Newsgroups: comp.os.vxworks Subject: Re: motFccEnd.c Date: 7 Feb 2002 17:31:19 -0500 From: vanbaren@falcon.si.com (vanbaren_gerald) Organization: Smiths Industries Message-ID: <3c630037$1@news.si.com> References: "Dan Gold" writes: >"Ali" wrote in message > >> Hi Dan >> Thanx for ur reply yes i have the source files for sysmotfccend.c and >> configNet.h but the situation there is only for one instance of end >> driver can u pls tell me what extra changes i have to make to these >> two files so that i can support two fcc's for ethernet and that two >> both of them in full duplex mode. >> waiting for the reply >> thanx >I'm not using sysMotFccEnd.c, but my configNet.h file for multiple END >devices (1 SCC and 2 FCC) looks like: >END_TBL_ENTRY endDevTbl [] = >{ > { 0, PQ2_SCC_LOAD_FUNC, PQ2_SCC1_LOAD_STRING,PQ2_SCC_BUFF_LOAN,NULL, >FALSE}, > { 3, PQ2_FCC_LOAD_FUNC, PQ2_FCC3_LOAD_STRING,PQ2_FCC_BUFF_LOAN,NULL, >FALSE}, > { 1, PQ2_FCC_LOAD_FUNC, PQ2_FCC1_LOAD_STRING,PQ2_FCC_BUFF_LOAN,NULL, >FALSE}, > { 0, END_TBL_END, NULL, 0, NULL, FALSE}, >}; >That should cause SENS to run the xxxEndLoad/xxxEndStart functions for each >entry. To actually bring up the interface (since vxWorks only brings up the >boot device in usrNetInit()), see the vxWorks FAQ. They have a little code >snippet (that I wrote) to bring up more interfaces. As far as setting it >for full duplex, I don't know where you specify that - I just autodetect, >and then read my PHY to see if it's full or half duplex. > hth, > Dan Gold gold@ensemble.com I have not done it myself, but I recall seeing warnings that you need to make sure the two FCCs are using different areas of DPRAM for their buffers (RIPTR, TIPTR). My recollection is that the WR driver can support any FCC port, but it has a hardcoded offset that it uses for RIPTR and TIPTR. gvb - -- +-----------------------------------------------------------------------+ | Jerry Van Baren / vanbaren_gerald@si.com / Grand Rapids, MI | | My employer is a company. Companies are artifacts of a legal system. | |_____________Artifacts are incapable of having opinions._______________| --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Fri Feb 8 09:25:50 2002 From: Wu@mosaid.com Date: Fri Feb 8 09:25:53 PST 2002 Subject: 64-bit data memory transaction with pentium VxWorks/Tornado world, This is not exactly a vxWorks question but more on pentium architecture, but I guess there might be someone who knows the answer. I'm using a pentium III with MMX assembler support, but I have never seen a 64-bit memory transaction when using an logic analyzer. I used Microsoft visual C++ 6.0 to compile and link the code. When I used Microsoft visual C++ debugger disassemble the MMX assembler, I can see the code is disassembled different from MMX code. Any guess where I could be wrong? Visual C++ compiler, or DIMM, or logic analyzer, or pentium memroy controller configured for 64-bit (is there such a bit?), CS for different memory banks? Any suggestions are appreciated. Thanks Huang Wu In the following, the line number followed by " : " is my original MMX code, the other are those disassembled by visual C++ debugger. 27: _asm 28: { 29: mov eax, pData0 00401065 mov eax,dword ptr [ebp-20h] 30: mov ebx, pData1 00401068 mov ebx,dword ptr [ebp-24h] 31: mov ecx, pData2 0040106B mov ecx,dword ptr [ebp-28h] 32: mov edx, pData3 0040106E mov edx,dword ptr [ebp-2Ch] 33: 34: movq mm1, [eax] 00401071 movq mm1,mmword ptr [eax] 35: movq mm2, [ebx] 00401074 movq mm2,mmword ptr [ebx] 36: movq mm3, [ecx] 00401077 movq mm3,mmword ptr [ecx] 37: movq mm4, [edx] 0040107A movq mm4,mmword ptr [edx] 38: 39: punpcklwd mm2, mm1 0040107D punpcklwd mm2,mm1 40: punpcklwd mm4, mm3 00401080 punpcklwd mm4,mm3 41: 42: punpckldq mm4, mm2 00401083 punpckldq mm4,mm2 43: 44: mov edx, pDword 00401086 mov edx,dword ptr [ebp-14h] 45: movq [edx], mm4 00401089 movq mmword ptr [edx],mm4 46: emms 0040108C emms 47: } From vxwexplo-errs@csg.lbl.gov Fri Feb 8 10:57:03 2002 From: =?iso-8859-1?q?ram=20mohan?= Date: Fri Feb 8 10:57:06 PST 2002 Subject: Re: mkbootAta and partition Hi i havnt worked on Flash systems ,i can suggest u to mail to my friend who can help u, sumi_reddy77@hotmail.com, he is working on the flash file system. bye Mohan --- the vxWorks Users Group Exploder wrote: > Submitted-by vxwexplo-errs@csg.lbl.gov Fri Feb 8 > 00:02:28 2002 > Submitted-by: f.pertin@staubli.com > > Hello, VXWORKS and TORNADO gurus > > I have an x86 target with a compact flash. I am > working with VX5.4 and > DOSFS2.0. I succeed to create 3 partitions on this > flash with the VX > function usrFdiskPartCreate . Now I want my flash to > be bootable. I try > with the vx function mkBootAta but I never succeed. > The problem is that > when I run this function I did not get any error but > when I try to rebbot > on my flash, the boot stop, I do not see the usual > message VxLd1.2...... > > Does anybody try to make a bootable compactFlash > with partitions ????? > > Best regards > > > > François PERTIN > > > STÄUBLI SCA > Robotics Division > Place Robert Stäubli - BP 70 > F - 74210 Faverges (France) > Tel.: +33 (0)4.50.65.62.59 > Fax: +33 (0)4.50.65.61.30 > mailto:f.pertin@staubli.com > http://www.staubli.com > > > ********** > > This is a user group mailing list for vxWorks > related topics > see http://www-vxw.lbl.gov/vxworks for posting > and subscription info From vxwexplo-errs@csg.lbl.gov Fri Feb 8 15:22:32 2002 From: "Nadar, Jaybal J" Date: Fri Feb 8 15:22:34 PST 2002 Subject: VxWorks 5.4/MPC8260 We are working on a driver for the MPC8260. During the testing, we have determined that there is a cache coherency problem on the MPC8260 on certain hardware configuration. We are currently using the Motorola MPC8260 ADS board for our testing. In order to guarantee cache coherency, we have included the cache flushing and cache invalidate functions as described in the example illustrated in the vxWorks reference manual (1-42). We are using the standard vxWorks call cacheFlush for the flushing of the cache before sending data and cacheInvalidate for the invalidation on reception of data. The problem that we are experiencing is that everything works fine without these caching functions on the Motorola MPC8260 ADS board. We have determined that after adding the caching functions to the driver, the cacheInvalidate call caused memory corruptions. We have come to this conclusion after an indepth investigation involving the following: 1. Running the driver in TX mode only and all was well. 2. Running the driver tests in loopback mode with the cache Invalidation call removed - everything runs fine 3. Running the driver tests in loopback mode with the the cache Invalidation call once again insert - we get memory corruption. As part of the investigar\tion into resolving this, we have ensured the following: 1. Our buffers are always 32 bytes aligned. (TRUE) 2. The length is always a multiple of 32 bytes. (TRUE) The questions we place before you are: Is this a known problem regarding the cache flush and invalidate for which there is a fix or a workaround? Are there any changes / set up parameters that need to be made to the driver / WxWorks when using the cache invalidate call? Regards, Jaybal From vxwexplo-errs@csg.lbl.gov Sat Feb 9 04:03:07 2002 From: Vxworks Exploder Date: Sat Feb 9 04:03:10 PST 2002 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sat Feb 9 04:03:04 PST 2002 Subject: How to build a image without host? ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: How to build a image without host? Date: 8 Feb 2002 19:20:08 -0800 From: forbes_bofeng@sohu.com (forbes) Organization: http://groups.google.com/ Message-ID: <7d3efabf.0202081920.516e2717@posting.google.com> When I debug the board for PowerPC 860, I made a BSP, Connection betwween host and target is network(CPM). After I made a bootrom, board could connect to host, download file from host and download program to run. I want to ask that how to build a image(include VxWorks and user application) that run without host???? thanks.... --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Sun Feb 10 04:03:10 2002 From: Vxworks Exploder Date: Sun Feb 10 04:03:12 PST 2002 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sun Feb 10 04:03:07 PST 2002 Subject: Re: how i build bootrom Subject: Again, I need SCC2698( Philips) and SC28L198 VxWorks Driver Subject: Re: tNetTask takes too long? Subject: ln7990End Working on Force 68K w/ Lance ? Subject: Questions about vxmalloc.c Subject: Re: How to build a image without host? Subject: Re: Still "dosFsDevInit failed". Subject: Re: dosFs size limit 32MB? Subject: Re: DOS key with timeouts, idea? followup using Select() Subject: I/O redirection problems Subject: ~!~ C-Designer (mit RoseRT Ausbildung) für Realtime/C/Systeme sucht neues Projekt ~!~ Subject: Re: about lex ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: how i build bootrom Date: 9 Feb 2002 05:06:54 -0800 From: alimohd@logiceastern.com (Ali) Organization: http://groups.google.com/ Message-ID: References: <4f09bb68.0202081309.2e2c0c0d@posting.google.com> hi this file is there buti need this file in building my image so when i remove this from tree i can't build the image and if i keep this i am unable to build bootrom . ali akarpov@my-deja.com (Alex) wrote in message news:<4f09bb68.0202081309.2e2c0c0d@posting.google.com>... > Go and double check if there are such a file in the main Tornado tree > Tornado/target/src/drv/intrCtl/m8260IntrCtl.c > or in your BSP : > $(BSP_DIR)/src/drv/intrCtl/m8260IntrCtl.c > > > alimohd@logiceastern.com (Ali) wrote in message news:... > > > > > make: *** No rule to make target `drv/intrCtl/m8260IntrCtl.c', needed > > > > by `sysLib.o'. Stop. > > vxworks --------------------------- Newsgroups: comp.os.vxworks Subject: Again, I need SCC2698( Philips) and SC28L198 VxWorks Driver Date: Fri, 8 Feb 2002 18:06:54 +0900 From: "william hong" Organization: Korea Telecom Hitel Message-ID: Hi. I need serial chip bsp. If you have it, send me please. Thank you. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: tNetTask takes too long? Date: Sat, 09 Feb 2002 01:08:44 GMT From: "Kelly Hornsby" Organization: Cox Communications Message-ID: References: I also had this problem. I have a system that adds and removes a PMC networking card based on lab or field use. We found that our system would freeze for 1 second every 5 seconds... really weird. It turns out it's the tNetTask doing it. I had to add in a pciFindDevice to see if my networking card was there, and if not, to delete the tNetTask task. I think it's like the posts imply, the tNetTask pends all tasks until it determines the state of the networking devices. A hack, but it solved the problem for both configuraitons. --------------------------- Newsgroups: comp.os.vxworks Subject: ln7990End Working on Force 68K w/ Lance ? Date: 8 Feb 2002 23:47:03 -0800 From: greene@kmr.ll.mit.edu (Dave) Organization: http://groups.google.com/ Message-ID: The ln7990End "example" driver for the Lance on-board controller seems almost tailor made for a Force 68K board; any idea why it doesn't work for me ? I haven't made any changes to it, just set up configNet.h(LN_LOAD_STRING). Is this driver actually workable or does it need reparations just to get it going under T2.0.2./Vx5.4 ? I build and boot VxWorks and it "hangs" right after load and Go. I don't get a single logmsg indicating "ln" startup. I'd like to debug the startup but don't really know how ? Can anyone get me started ? Thx, Dave. --------------------------- Newsgroups: comp.os.vxworks Subject: Questions about vxmalloc.c Date: 8 Feb 2002 18:42:01 -0800 From: manbert@rahul.net (Rush Manbert) Organization: http://groups.google.com/ Message-ID: I have gotten a copy of vxmalloc.c, and I was trying to figure out the changes relative to Doug Lea's original, so I could apply the same changes to the latest version. I came across a number of macros that were redefined in vxmalloc.c, and I'm trying to understand why the changes were made. These include MALLOC_ALIGNMENT, chunk2mem, and mem2chunk. So my question is, has anyone used this code and analyzed why these changes exist, and/or does anyone know a current email address for Hwa-Jin Bae, the author? Thanks, Rush --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to build a image without host? Date: Sat, 09 Feb 2002 10:26:36 -0800 From: "Michael R. Kesti" Organization: MK Associates Message-ID: <3C6569DC.D629CE75@gv.net> References: <7d3efabf.0202081920.516e2717@posting.google.com> Reply-To: mkesti@gv.net forbes wrote: >When I debug the board for PowerPC 860, I made a BSP, And burned a boot ROM that making the BSP builds. I like to call this the "development boot ROM". >Connection betwween host and target is network(CPM). >After I made a bootrom, board could connect to host, >download file from host and download program to run. There are usually two downloads. The first is an operating system image downloaded by the boot ROM. I like to call this the "development OS". It is typically built in Tornado as a "bootable vxWorks image" using the "vxWorks" rule with INCLUDE_USER_APPL undefined. The second download is the application program image that you download via the debugger. It too is typically built in Tornado, but as a "downloadable vxWorks application" using the ".out" build rule. >I want to ask that how to build a image(include VxWorks >and user application) that run without host???? >thanks.... There are several ways to go about this. The vxWorks Programmer's Guide describes a couple, but I never could figure out exactly how they meant for them to work. What I have done and will describe has worked well for me. Start by creating another Tornado project of the "bootable vxWorks image" type that builds using the "vxWorks_rom.hex" rule. I like to call this the "deployment OS" and its build will be used to burn another boot ROM that I call the "deployment boot ROM". Include your "downloadable vxWorks image" that is your application in this image by setting this new project's EXTRA_MODULES macro with the path to your application's .out file. Also define the INCLUDE_USER_APPL macro in this project and add code to its usrApplInit.c module to call your application's entry point. Finally, burn the result of this build to a ROM, replay your development boot ROM with this new ROM, and, presto-chango, you have a system that runs your application without the need for host support! You might, at first, think it inappropriate to have two OS projects associated with your target. This actually works out quite well, though, as it allows you to include OS components to support debugging in your development OS, but exclude them from you deployment OS. A good example of this is the target shell. You will certainly want to include it in your development OS, but, for security reasons and to prevent your customers from messing things up for themseleves, you may wish to exclude it from your deployment OS. Good luck! - -- ======================================================================== Michael Kesti | "And like, one and one don't make | two, one and one make one." mkesti@gv.net | - The Who, Bargain --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Still "dosFsDevInit failed". Date: Sat, 9 Feb 2002 19:44:01 -0000 From: "cecil" Message-ID: <1013284083.9121.0.nnrp-01.d4e4b8d1@news.demon.co.uk> References: <2002130-143411-880483@foorum.com> <200221-134637-250830@foorum.com> We had this problem a couple of years a go. By default, our shell was assigned to a serial port. If a ctrl-s was received by the port that the shell was using while the machine as booting, the boot process was suspended, until you sent a ctl-q. You couldn't ping it, nothing. We would also occasionally see the message you have seen. That the power of xon-xoff!!!!! We were getting ctl-s randomly. This was because when vxworks prints out all that vxworks booting stuff, and our serial port was disconnected the output to the shell was been reflected back as random characters in to the shell. Now that may be specific to our application...but the moral to the sort is that issuing xon/xoff to the shell while it is booting kills that machine in the way you described. Hope that helps Nigel Charly wrote in message <200221-134637-250830@foorum.com>... > >Thanks for your help, >The problem is still there "dosFsDevInit failed". > >When I start a big application on my PCPentium target, the shell display >"dosFsDevInit failed". Then, the target seems dead, no Shell and no target >server connection. This application works fine on an other BSP (Lp-elektronik). >Do you have an idea where I could look? VxWorks configuration? >All my files are on a FTP server. >I use also a RAMDISK. > >Here are the result of some shell command. >-> devs >drv name > 0 /null > 1 /tyCo/0 > 1 /tyCo/1 > 2 /pcConsole/0 > 2 /pcConsole/1 > 6 host: > 7 /vio >value = 0 = 0x0 > >-> ls > > Volume in drive D has no label > Volume Serial Number is 12E3-344A > Directory of D:\kern\Awd_num\T2mak\rtl > >[.] [..] RTL.WPJ USRAPP~1.C PRJCOMPS.H >PRJPAR~1.H LINKSYMS.C PRJCON~1.C MAKEFILE [DEFAULT] >PRJOBJS.LST ADD_VXW.C USRAPP~2.C [DEBUG] > 10 file(s) 98,644 bytes > 4 dir(s) 1,170,317,312 bytes free >value = 0 = 0x0 > >-> version >VxWorks (for PC PENTIUM) version 5.4.2. >Kernel: WIND version 2.5 >Made on Feb 1 2002, 13:09:39. >Boot line: >ene(0,0)host:vxWorks e=192.0.1.2:ffffff00 h=192.0.1.1 u=target pw=vxworks f=0x8 >tn=target >o=ene. >value = 0 = 0x0 > > >-- >Use our news server 'news.foorum.com' from anywhere. >More details at: http://nnrpinfo.go.foorum.com/ --------------------------- Newsgroups: comp.os.vxworks Subject: Re: dosFs size limit 32MB? Date: Fri, 08 Feb 2002 19:32:02 GMT From: Andray Kaganovsky Organization: Primus Canada Message-ID: References: <6dcae13d.0202071948.73976ad8@posting.google.com> "Michael Lawnick" wrote in news:a40227$g58$1 @snoopy.bndlg.de: > Hi roswell, > > AFAIK, your limit should not be 32MB, but 40MB. You'll find in > target/h/tffs/flCustom.h: > #define MAX_VOLUME_MBYTES 40 > I had to increase SECTOR_SIZE_BITS (512 byte->1024 byte) to get over 64MB. > > Change this defines and recompile TFFS-sources if available or ask your FAE. > > -- > Mit freundlichen Grüßen, > Michael Lawnick >============================================== > SOFTEC GmbH Tel +49-731-96600-0 > Promenade 17 Fax +49-731-96600-23 > D-89073 Ulm Michael Lawnick > Germany lawnick@softec.de > ============================================== > > "roswell" schrieb im Newsbeitrag > news:6dcae13d.0202071948.73976ad8@posting.google.com... >> Hi vxworkers, >> >> I want to dosFs to use my entire flash (128MB), but it is using only >> 32MB, according to what i see on dosFsConfigShow("/tffs0/"). Is there >> an option to increase this? >> >> many thanks.. > > [AndrayK]. It is true that TFFS limits size of the single TFFS volume to 40MBytes. If you happened to have TFFS source code, then Michael's suggestion above will allow you to work around this limit. If you don't have TFFS source code, then you can "partition" your physical flash array into independent TFFS volumes. You will have to do minor change in your TFFS's socket code in /sysTffs.c to do address wrap around 32MBytes offsets. This will allow you to install four socket components, with base addresses as follows: socket # address range 0 ... + 32MB 1 + 32MB ... + 64MB 2 + 64MB ... + 96MB 3 + 96MB ... + 128MB Each of the memory ranges above can be formatted and mounted as independent TFFS volume. Depending on the size of the physical flash parts that comprise your flash array, and flash array's interleaving, you can do I/O in parallel to all four TFFS volumes, which can greatly improve 'write' performance. Andray --------------------------- Newsgroups: comp.os.vxworks Subject: Re: DOS key with timeouts, idea? followup using Select() Date: 8 Feb 2002 11:54:07 -0800 From: mark_ck_chang@yahoo.com (Mark Chang) Organization: http://groups.google.com/ Message-ID: <9b6f8883.0202081154.3b243fab@posting.google.com> References: <9b6f8883.0202061148.5789ec09@posting.google.com> <9b6f8883.0202061613.584be7e5@posting.google.com> <9b6f8883.0202071134.685fc409@posting.google.com> vloscomp@yahoo.com (Vinh) wrote in message news:... > mark_ck_chang@yahoo.com (Mark Chang) wrote in message news:<9b6f8883.0202071134.685fc409@posting.google.com>... > > vloscomp@yahoo.com (Vinh) wrote in message news:... > > > Hi Mark, > > > As for the Telnet and Procom, the telnetd does not do CR+LF, thus > > > tried to turn off the option in Procom. Otherwise you need to get the > > > telnetd source and do it yourself, which is alot better. > > > You can even do priority or exclusion for the loggin on over console > > > and telnet... > > > > > > The select seem to works fine for me, and i did a comparison test > > > with the counting.... results: select is better not by much, this may > > > not be correct! > > > > > > You can investigate and let me know :) > > > > > > Vinh Lam > > > > > Thanks Vinh, > > I did not see an option in Procom to turn off CR+LF. > > In telnetlib.c, there is this code > > c = *buf++; /* get next character */ > > ci = (unsigned char) cc; /* convert to int since many values > > * are negative characters */ > > switch (state) > > { > > case TS_CR: /* doing crmod; ignore add'l linefeed */ > > state = TS_DATA; > > if ((cc != EOS) && (cc != '\n')) > > sendToPty (ptyFd, &cc, 1); /* forward char */ > > break; > > case TS_DATA: /* just pass data */ > > if (ci == IAC) > > state = TS_IAC; > > else > > { > > sendToPty (ptyFd, &cc, 1); /* forward char */ > > if (!myOpts [TELOPT_BINARY] && (cc == '\r')) > > state = TS_CR; > > } > > break; > > SNIP > > > > I think what it is doing is stripping out the linefeed. > > So I'm going to just pass on everything expect IACs. > > > > btw - I rewrote the code to do Telnet and Serial login > > concurrently not exclusive like you suggested. > > Hum... I do not know your system requirements, but concurrency can > caused problem when you have 2 sessions trying to modified the same > system resources. > This is just a warning, i had run in this trap. > > > Our telnet supports 5 sessions > > and a serial session concurrently. > > This was from the unsupported concurrent telnet code. > > Works great. > > Mark Each session is just another task and is treated as such. We were using TCL7.3 and rewrote it to be reentrant. btw - the problem turned out to be the way ethernet sends the data. In a ProComm script, it was one ethernet package that contain the entire line and this trigger the select only once. So my routine only saw the first character. That's why using the keyboard, hitting one key at a time works. The ethernet package only contained one byte and my routine expected that. The way I fixed it was to read the entire buffer and parse each character into my keyboard routine. It's works really well now - thanks. Mark --------------------------- Newsgroups: comp.os.vxworks Subject: I/O redirection problems Date: Fri, 8 Feb 2002 11:15:19 -0800 From: "Ricardo J. Hassan II" Organization: Jet Propulsion Laboratory - Pasadena CA Message-ID: Hi, I am using Tornado II, and I am trying to get all IO directed to the wind shell. Right now my printf's are being directed to the windshell, but logMsg is being directed to the target serial port. I have used ioGlobalStdGet to direct all stdfd's to the virtual device /vio/o. I called logInit for this fd, and then called logFdSet to the same fd. Does anyone have any suggestions? Thanks. Ricardo --------------------------- Newsgroups: comp.os.vxworks Subject: ~!~ C-Designer (mit RoseRT Ausbildung) für Realtime/C/Systeme sucht neues Projekt ~!~ Date: Fri, 8 Feb 2002 21:05:30 +0100 From: "M.Thoemel" Organization: T-Online Message-ID: Guten Tag, ich bin freier Sofwareentwickler mit einem sehr breiten Spektrum. News: Aktuell habe ich einen Kurs über UML und Rational Rose RT absolviert und werde diese SOftware jetzt zwei Monate für ein VxWorks Target einsetzen. Ab dem April 2002 stehe ich wieder zur Verfügung. Meine Schwerpunkte sind: - - C (aber auch andere Sprachen!) - - Realtimesysteme - - systemnahe Programmierung - - spezielle Hardware-Schnittstellen - - Netzwerke - - Treiber - - Streamsdriver - - Grafische Datenauswertung/Darstellung - - Software-Messungen - - Gruppenleitung - - Projektierung - - Projektleitung etc... Sie können sich im Detail informieren unter: http://matthias.thoemel.bei.t-online.de Ich würde mich über einen Kontakt mit Ihnen freuen. Ihr Matthias Thömel (matthias.thoemel@t-online.de) --------------------------- Newsgroups: comp.os.vxworks Subject: Re: about lex Date: 8 Feb 2002 20:24:05 GMT From: Troy Cauble Organization: Bell Labs, Holmdel, NJ, USA Message-ID: References: <3d64ee1a.0202052234.26dc4574@posting.google.com> changwoo wrote: : I am tired to use lex in vxWorks : but there is some problem to compile the files. : First I check the source file and successed in Unix environment. : with the same source, : In vxWorks, first I compiled it with lex.(lex sampel.l) : after then outfile was created "lex.yy.c" and I include the file to : workspace and compiled, but there was error "inttypes.h : no such : file" showed. : how can I fix this. : And what I have to do other thing to use "lex" in vxworks ? : Help this situation... thanks Lex is Unix centric. VxWorks is not Unix. inttypes.h is a header file that exists in your Unix environment. It doesn't exist in your VxWorks compile environment. Try removing the inttypes.h line in lex.yy.c. You may have to remove other #includes. You may have to add VxWorks #includes to replace them. You may have to replace other things... Did you know that lex generated code calls exit() on failure? Probably not what you want in VxWorks. Did you know it used global data? Will you run more than one lexer at a time? Did you know lex code allocates memory on the fly? This may be important to your app. Do you need/want to support FILE* io in your app? You might consider a script (awk or perl) to post process lex generated code and make it more portable. This will be non trivial. Or you might consider not using lex. - -troy --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Sun Feb 10 07:57:06 2002 From: Ronen Levy Date: Sun Feb 10 07:57:08 PST 2002 Subject: Bus Error on PPC860 with vxWorks Hi, I have a BUS ERROR in my PPC860 target board, and I am trying to catch it and figure out what is the address in the program that causes this Bus Error. Below, is a quote that states that I should disconnect the vxWorks built-in Bus Error Handler, and provide my own. I have seen the posting for this function: STATUS intDisconnect(int vector) {} But I could NOT find/understand in MPC860 User Manual what is the vector number for BUS ERROR. I would appreciate if you email me an example that DELETES Bus Error, and hook up a user interrupt Bus Error handler. Appreciate your resonse, Ronen Levy. "I did this because I found it was not possible to use the VxWorks intConnect() call to set up a bus error exception handler. Upon initiation of a bus error from the task level, the system would hang in the intExit() wrapper code set up by the intConnect() call." From vxwexplo-errs@csg.lbl.gov Sun Feb 10 14:57:13 2002 From: hnguyen@pdx.edu Date: Sun Feb 10 14:57:15 PST 2002 Subject: VxWork CD-RW Driver for TEAC Does anyone has the source code driver in C/C++ for CD-RW that work in Linux or VxWorks? I'm working in a senior project and need your help. Thanks Howard Nguyen hnguyen@pdx.edu 503-515-5920 From vxwexplo-errs@csg.lbl.gov Mon Feb 11 04:03:08 2002 From: Vxworks Exploder Date: Mon Feb 11 04:03:11 PST 2002 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Mon Feb 11 04:03:05 PST 2002 Subject: Re: Equivalent of setitimer Subject: multicast on 'ene' ne2000 ethernet driver Subject: Re: Problem with IP Forwarding Subject: Beginner Client/Server problems! Subject: Re: Again, I need SCC2698( Philips) and SC28L198 VxWorks Driver Subject: Re: How do I get started learning VxWorks? Subject: Re: Measuring performance of Embedded Software. Subject: Re: How do I use intConnect Subject: Re: 64-bit operations ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Equivalent of setitimer Date: 11 Feb 2002 00:30:02 -0800 From: bryanramesh@yahoo.com (Bryan) Organization: http://groups.google.com/ Message-ID: References: <3C66A964.A7BDD3D0@gv.net> Hi Mic, Yeah i already had a look at watchdog timers but they will run in ISR context right. I dont want ISR context timers. Atleast signal handler level or task context level timers i need. I tried with timer_settime. It worked. THanks for the response Bryan "Michael R. Kesti" wrote in message news:<3C66A964.A7BDD3D0@gv.net>... > Bryan wrote: > > >In Linux setitimer is there to set a timer and when the timer expires > >SIGALRM will be sent. In VxWorks what is the equivalent of that? In > >timer_create call where shall we mention the timer value. > > See the vxWorks docs for "wdLib", where "wd" stands for "watchdog" as in > "watchdog timers". --------------------------- Newsgroups: comp.os.vxworks Subject: multicast on 'ene' ne2000 ethernet driver Date: 11 Feb 2002 00:33:11 -0800 From: rsramkee@mailcity.com (Ramki) Organization: http://groups.google.com/ Message-ID: Hi All, I found the following while searching in the comp.os.vxworks archive. -8<------- : > I'm using VxWorks with a PC386 target, NE2000 NIC. : > Does anyone know if the 'ene' driver supports multicast? : > If so, how do we activate it? How do we set the multicast : > address(es)? : The Vxworks TCP/IP stack doesn't support multicast. A number of : organizations have hacked in support for multicast at various times, : but you have to have source to the TCP/IP stack to use this, AND : must know how to adapt your ethernet driver to support it. I'm trying to use the 'ene' driver that's built into VxWorks. Are you saying that driver does not support multicast? I don't care about the TCP/IP stack, I have my own stack. -------8<- I have the same problem with the same situation. I have my own IP stack. I use the 'ene' driver and it doesnt seem to support multicast. I use VxWorks 5.4 on x86. However, multicast works for another ethernet Card ('elPci'). I've pasted the output I got from ifShow('ene') at the end of this post. I would greatly appreciate any ideas/pointers. Thanks for your time Ramki ene (unit number 1): Flags: (0x8163) UP BROADCAST MULTICAST PROMISCUOUS ARP RUNNING Type: ETHERNET_CSMACD Netmask 0x19ee00 Subnetmask 0x19ee00 Ethernet address is 00:40:33:33:d4:8c Metric is 0 Maximum Transfer Unit size is 1500 0 packets received; 0 packets sent 0 multicast packets received 0 multicast packets sent 0 input errors; 0 output errors 0 collisions; 0 dropped elPci (unit number 2): Flags: (0x8963) UP BROADCAST MULTICAST PROMISCUOUS ARP RUNNING Type: ETHERNET_CSMACD Netmask 0x19ee00 Subnetmask 0x19ee00 Ethernet address is 00:10:5a:85:bb:da Metric is 0 Maximum Transfer Unit size is 1500 0 packets received; 0 packets sent 0 multicast packets received 0 multicast packets sent 0 input errors; 0 output errors 0 collisions; 0 dropped value = 29 = 0x1d --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Problem with IP Forwarding Date: Mon, 11 Feb 2002 10:56:16 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <1a938d02.0202102123.71ad4c26@posting.google.com> Hemant, you obviously need to study TCP/IP. Your ISP does not route your internap IP network which is configured on the ethernet side, hence you dont get any replies. What you probably need is to add NAT on the VxWorks target so that every packet from the ethernet will come out on PPP as if it was from the 8260 itself, which gets a dynamic IP address from the ISP, thus is globally routed. I recomment you read some white papers about NAT and NAPT on the Cisco web site. - - Leonid "Hemant Passi" wrote in message news:1a938d02.0202102123.71ad4c26@posting.google.com... > Hi, > > I'm developing an application that connects several host machines to > the internet via the 8260 board (acting as the router). The > application essentially involves dialing an ISP through the modem, > connected to the com port of the target, and the establishment of the > PPP over the serial channel. > > The problem description is like this: > > After the successful establishment of the connection with the ISP, > when an IP address is pinged from the target shell, the ping returns > successfully with valid bytes echoed. > > The same IP when is pinged from the host machine (the host machine is > connected to the target on the 100BT using a cross ethernet cable), > the ping times out. > > On using the commands like ipstatShow & pppstatShow, the following > things are observed: > 1. The PPP is running fine. > 2. The IP forwards the packets requested from the host to the PPP. > 3. The packets are not received by the serial driver; (this I checked > by putting a break point at tyWrite). > > I have tried with various PPP options but all in vain. > > Request you to kindly revert in this regard. > > Thanks. > > Hemant Passi --------------------------- Newsgroups: comp.os.vxworks Subject: Beginner Client/Server problems! Date: 8 Feb 2002 08:32:22 -0800 From: joeycook@mail.com (Joseph Cook) Organization: http://groups.google.com/ Message-ID: I am trying to send a message through a socket from a Client to a Server and then have the Server return that message. I am trying to do this in a loop as many times as I want. The message changes each time as well to show which iteration it is.. It hangs though on the send or receive calls, and I can't figure out what is wrong... Client: char messBuff[8]; char replyBuff[8]; for (i=0; i<5; i++) { sprintf(messBuff, "Message%d", i); if(write(Fd, messBuff, sizeof(messBuff))==ERROR) //ERROR if(read(Fd, replyBuff, sizeof(replyBuff)) < 0) //ERROR } Server: char messBuff[8]; char replyBuff[8]; for(i=0; i<5; i++) { if((nRead = read(Fd, messBuff, sizeof(messBuff))) == ERROR { //ERROR } if(write (Fd, replyBuff, sizeof(replyBuff) == ERROR)) { //ERROR } Seems simple enough...what could be wrong? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Again, I need SCC2698( Philips) and SC28L198 VxWorks Driver Date: 11 Feb 2002 09:31:48 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: "william hong" wrote: > Hi. > > I need serial chip bsp. Try contacting Compware (see their Vx/IP-OctalSerial driver) for the SCC2698. Groeten, Johan - -- o o o o o o o . . . _____________________________ o _____ || Johan Borkhuis | .][__n_n_|DD[ ====_____ | borkhuis@agere.com | >(________|__|_[_________]_|__________________________| _/oo OOOOO oo` ooo ooo 'o!o!o o!o!o` === VxWorks FAQ: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html === --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How do I get started learning VxWorks? Date: 11 Feb 2002 10:09:49 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: I.need.work@com.net (Unemployed Windows Developer) wrote: > I am an unemployed Windows (VC++) developer and would like to learn > the VxWorks OS. Does anyone have any tips on what I should do - > books, development kits, etc.? Thanx Start by getting some information on embedded and realtime SW development (try a seach on amazon or fatbrain using embedded and realtime). Also download Tornado Prototyper, this is a VxWorks simulator that runs on a PC. You can get s feeling of how to develop for VxWorks using this. Groeten, Johan - -- o o o o o o o . . . _____________________________ o _____ || Johan Borkhuis | .][__n_n_|DD[ ====_____ | borkhuis@agere.com | >(________|__|_[_________]_|__________________________| _/oo OOOOO oo` ooo ooo 'o!o!o o!o!o` === VxWorks FAQ: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html === --------------------------- Newsgroups: comp.os.psos,comp.realtime,comp.arch.embedded,comp.os.vxworks Subject: Re: Measuring performance of Embedded Software. Date: Mon, 11 Feb 2002 08:25:23 +0100 From: "Martin Raabe" Organization: Wind River GmbH Message-ID: References: Hello ahosoto, please use ESp (Embedded System Profiler) for the first approach. Also you may use the profiler of pROBE+. Hope it helps Ciao Martin Raabe "akosoto" schrieb im Newsbeitrag news:d64bec.0202102221.34c20ff@posting.google.com... > Hi all, > We were trying to measure the performance of our real time software which is a > PSOS based system. We basically want to measure the performance of our device > drivers, DLC software and the kernel. Can you please give us guidelines on > how to go about it. Any methods,samples,software,links etc... > with regards, > Ashokc. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How do I use intConnect Date: 11 Feb 2002 03:11:31 -0800 From: danielmartinfuchs@yahoo.co.uk (Daniel Fuchs) Organization: http://groups.google.com/ Message-ID: <78cc994f.0202110311.3a3493b7@posting.google.com> References: <78cc994f.0201300758.3d442448@posting.google.com> <3c587774$1@brateggebdc5.br-automation.co.at> <7f5i5ukjchuibfa52uobmhf0clcthlcqsd@4ax.com> Bertram Metz wrote in message news:<7f5i5ukjchuibfa52uobmhf0clcthlcqsd@4ax.com>... > Sorry, > there was a wrong email address in my last posting. > > Bertram Just a quick note to thank you all for your hlep.. most useful. Thanks. Your quick responce(s) were greatly appreciated. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: 64-bit operations Date: Mon, 11 Feb 2002 06:41:44 -0500 From: joe durusau Organization: BELLSOUTH.net Message-ID: <3C67ADF8.55BA8A64@bellsouth.net> References: This is just a guess, but have you tried splitting the variable into bytes (a union or some such) an dprinting the bytes? It is entirely possible that the bsp does not handle long long and therefore prints them as zero. On most BSPs, you fill find that trying to print a float with logMsg will do similar things, because logMsg just doesn't support floats. Speaking only for myself, Joe Durusau Nejhdeh Ghevondian wrote: > Dear anyone, > I am trying to implement 64-bit operations in my VxWorks operations (using > the GNU compilier, on PowerPC platform). > I am trying to multiply and divide unsigned long long int variables. > However, I always get 0 as value. > I acnt even do a printf("%llu\n", 64bit_var) > > Can anyone help me??? > > Regards Nejhdeh Ghevondian --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Mon Feb 11 04:13:28 2002 From: Ronen Levy Date: Mon Feb 11 04:13:30 PST 2002 Subject: recover Bus Error Exception - sysIntTbl Hi, I am trying to recover from BUS ERROR exception that occurs on my board. In the process of deatch vxWorks Exception handler, and trying to put my own handler. For that, I am looking for the 'sysIntTbl', where vxWorks holds the Exception Handler in a table. The problem that I am facing: 'sysIntTbl' is Not found. Q: Is that the Correct symbol table? Any more comments are welcome. Can some From vxwexplo-errs@csg.lbl.gov Mon Feb 11 06:31:36 2002 From: "Fei, Hong-Bing" Date: Mon Feb 11 06:31:38 PST 2002 Subject: access to comp.os.vxworks Hi there, Could you tell me how to access the news group form web browser in a Win2000 machine? Thanks! Hong-Bing Fei .............................................................................. Hong-Bing Fei Embedded software engineer, Ciena Corp Tel ( 410) 981 7236 ............................................................................. From vxwexplo-errs@csg.lbl.gov Mon Feb 11 06:43:48 2002 From: "Mike Anderson" Date: Mon Feb 11 06:43:50 PST 2002 Subject: RE: access to comp.os.vxworks VxWorsk Greetings! > > Hi there, > > Could you tell me how to access the news group form web > browser in a Win2000 machine? > Go to http://www.deja.com and type your queries there. However, if you really want to just scan all of the traffic, you'll need access to a news reader such as Outlook Express and a news feed. HTH, Mike -- ======================================================= T h e P T R G r o u p, I n c. ======================================================= Embedded, Real-Time Solutions and Training Mike Anderson mailto:mike@ThePTRGroup.com Chief Scientist http://www.ThePTRGroup.com C: (703) 585-9384 V: (703) 430-3748 ======================================================= "Software development is like making a baby... You can't make a baby in one month by impregnating nine women. Some things just take time." From vxwexplo-errs@csg.lbl.gov Mon Feb 11 14:13:17 2002 From: Rainer Stelzer Date: Mon Feb 11 14:13:19 PST 2002 Subject: rebooting the master in a shared memory network Hi, we run a shared memory network on a VMEBus system. (master CPU (mvme 2304) and twelve slaves) While rebooting single slaves is not a problem, rebooting the master causes the complete system to be rebooted, because the slaves didn't reconnect to the master. Is this a common feature of the SM network implementation or a special one caused by our configuration ? I can't see that there is a principle reason for that. Neither for hardware limitations (the sm master is not the VMEBus controller, VMEBus slave access is locked while rebooting and the anchor adress doesn't change) nor for software reasons. (That's IP's job to deal with this kind of exceptions) in other words: is the slave side of shared memory network driver able to reconnect to a master that has rebooted ? Thanx in advance. regards R.Stelzer vxworks Tornado ===================================================== Gebrüder Stoye GmbH department: traffic control systems name: Rainer Stelzer adr: Longericher Strasse 177 city: Cologne, Germany zip: 50739 mobil: 0173/7266090 phone: ++49/221/1707019 fax: ++49/221/1707010 mailto: r.stelzer@stoye.de www: http://www.stoye.de ===================================================== From vxwexplo-errs@csg.lbl.gov Tue Feb 12 04:03:20 2002 From: Vxworks Exploder Date: Tue Feb 12 04:03:23 PST 2002 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Tue Feb 12 04:03:15 PST 2002 Subject: entry symbol lost / changed (ld386) ? Subject: Tornado Training Subject: Re: motFccEnd.c Subject: Problem with IP Forwarding Subject: Re: motFccEnd.c Subject: Re: DNS question. Subject: Re: DOS key with timeouts, idea? followup using Select() Subject: Re: return to WindSh from a task Subject: Re: Tornado-2.1/mips and cp7000 BSP (Ocelot MIPS board) Subject: Re: How to make Ctrl-X handler wait for a routine to finish Subject: Re: Wind Web Server question Subject: Re: Intermittent reboots Subject: Measuring performance of Embedded Software. Subject: 64-bit operations Subject: Re: How to make Ctrl-X handler wait for a routine to finish Subject: system debug mode Subject: Re: multiple instances of a variable Subject: Re: how i build bootrom Subject: Re: How to link with the BSD socket library Subject: Wind Web Server - problem in httpSymTabAdd Subject: Re: ~!~ C-Designer (mit RoseRT Ausbildung) für Realtime/C/Systeme sucht neues Projekt ~!~ Subject: Re: motFccEnd.c Subject: How to make Ctrl-X handler wait for a routine to finish Subject: Re: Questions about vxmalloc.c Subject: Re: PCI Device Base Address Subject: Re: Tornado Training Subject: Re: pciFindDevice()- index 0 based or 1 based ? Subject: Re: XON-XOFF on the Console port Subject: Re: Questions about vxmalloc.c Subject: Re: Equivalent of setitimer Subject: Re: Returning data from S-func in RTW Subject: Re: +++ C-Designer (sehr flexibel) für Realtime/C/Systeme sucht neues Projekt +++ Subject: newgroup comp.os.vxworks Subject: Re: about lex Subject: Re: MPC860 floating point emulation Subject: Re: New WindSerf, TSR's and SPR's. Subject: Re: Beginner Client/Server problems! Subject: tNetTask takes too long? Subject: stack USB from WinRiver on All BSP? Subject: Sockets Subject: generating a makefile (on command line) for bootable image Subject: Re: +++ C-Designer (sehr flexibel) für Realtime/C/Systeme sucht neues Projekt +++ ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: entry symbol lost / changed (ld386) ? Date: 11 Feb 2002 04:17:51 -0800 From: fillmore2@arcor.de (Achim) Organization: http://groups.google.com/ Message-ID: <9d082700.0202110417.1bfcda75@posting.google.com> Hi, somehow the entry symbol for the ld386 was changed from _sysInit to sysInit, giving me the following warning: "ld386: warning: cannot find entry symbol sysInit" Could anyone tell me where this entry point is set? I haven't been able to locate it in any of the project settings... Thanks. Achim --------------------------- Newsgroups: comp.os.vxworks Subject: Tornado Training Date: Sun, 10 Feb 2002 23:56:56 -0500 From: "news.sympatico.ca" Organization: Bell Sympatico Message-ID: Hello all, Tornado Training is being held in Toronto Ontario, has anyone gone to it and if so was it worth the $$? Patrick Egan Eganconsulting.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: motFccEnd.c Date: Fri, 8 Feb 2002 09:41:50 -0800 From: "Dan Gold" Organization: Mailgate.ORG Server - http://www.Mailgate.ORG Message-ID: References: <3c630037$1@news.si.com> Sender: gold@ensemble.com "Ali" wrote in message news:b305b5bd.0202072202.76bc37a4@posting.google.com... > hi Dan > Thanx for ur reply and i think it worked here since im in > development phase and infact the board we are going to make ourselves > has two fcc full duplex ports and the PHY device is LXT973(INTEL) now > tell me pls if u can how i bring up the interfaces and how to i assign > the private ip addresses to each port (ipAttach()) and where should i > add this much of code in BSP. How to bring up the additional interfaces: The first one (boot device) is brought up in usrNetInit(). To bring up others, look at the code in usrNetInit(), put it into a separate function, and call it from wherever you want in the BSP. Alternatively, you can go (as I mentioned before) to Johan's (useful, informative, well-maintained, great) vxWorks FAQ. There you'll find someone has already created a function to do what you're asking. Where do you call this function ? Wherever you want. You can call it from usrNetInit(), call it from usrConfig.c after usrNetInit(), at the end of usrRoot(), from your application, etc... > to get the ethernet mac address of each port and to enable the > interrupts do i have to call the functions > sysFccEnetAddrGet() > sysFccEnetEnable() > sysFccMiiBitWr() > sysFccMiiBitRd() > two times and do my work after.Also does these two fccs have > harcoded offsets in DPRAM and if it is so how can i know those > addresses.Does it seems to be right approach . > waiting for the reply Yes you need to call the above functions, once for each interface. Notice they (should) take a parameter, corresponding to the unit number of the device. So sysFccEnetAddrGet() should check the unit number and return the MAC address for the correct unit number. And sysFccEnetEnable() should setup the correct 8260 port pins for the particular unit number, and enable it. The bit Rd/Wr functions are with the WRS BSP and I believe they're used for talking to the PHY. Those will also need to look at the unit number so they can address the correct PHY. Since I don't use those, I can't help you much there. Which hardcoded offsets in DPRAM are you talking about ? The RIPTR/TIPTR ? As I said before, *you* must talk to your WRS FAE and make sure you have their latest driver that includes this fix, or get source and fix it yourself. You can find out if you have the fix by dumping the FCC parameter RAM and looking at the values for riptr/tiptr, and seeing if they're the same. hth, Dan Gold gold@ensemble.com --------------------------- Newsgroups: comp.os.vxworks Subject: Problem with IP Forwarding Date: 10 Feb 2002 21:23:06 -0800 From: hemant_passi@satyam.com (Hemant Passi) Organization: http://groups.google.com/ Message-ID: <1a938d02.0202102123.71ad4c26@posting.google.com> Hi, I'm developing an application that connects several host machines to the internet via the 8260 board (acting as the router). The application essentially involves dialing an ISP through the modem, connected to the com port of the target, and the establishment of the PPP over the serial channel. The problem description is like this: After the successful establishment of the connection with the ISP, when an IP address is pinged from the target shell, the ping returns successfully with valid bytes echoed. The same IP when is pinged from the host machine (the host machine is connected to the target on the 100BT using a cross ethernet cable), the ping times out. On using the commands like ipstatShow & pppstatShow, the following things are observed: 1. The PPP is running fine. 2. The IP forwards the packets requested from the host to the PPP. 3. The packets are not received by the serial driver; (this I checked by putting a break point at tyWrite). I have tried with various PPP options but all in vain. Request you to kindly revert in this regard. Thanks. Hemant Passi --------------------------- Newsgroups: comp.os.vxworks Subject: Re: motFccEnd.c Date: Fri, 8 Feb 2002 09:47:26 -0800 From: "Dan Gold" Organization: Mailgate.ORG Server - http://www.Mailgate.ORG Message-ID: References: <3c630037$1@news.si.com> Sender: gold@ensemble.com "I/me/my alter-ego" wrote in message news:a412r1$nm$3@newsreader.mailgate.org... > "Ali" wrote in message > news:b305b5bd.0202072202.76bc37a4@posting.google.com... > > to get the ethernet mac address of each port and to enable the > > interrupts do i have to call the functions > > sysFccEnetAddrGet() > > sysFccEnetEnable() > > sysFccMiiBitWr() > > sysFccMiiBitRd() > > two times and do my work after.> > Yes you need to call the above functions, once for each interface. Notice I should have also noted that you don't call these functions explicitly. They should be called by the FCC driver when the xxxEndLoad() and xxxEndStart() functions in the driver are called. hth, Dan Gold gold@ensemble.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: DNS question. Date: Fri, 8 Feb 2002 15:03:13 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: "Howchin" wrote in message news:a40478$t4i$1@news.seed.net.tw... > struct hostent * spstdtime; > char szBuff[256]; > if ( resolvInit("168.95.1.1","glorytel.com.tw",NULL)==OK) > { > spstdtime=resolvGetHostByName("clock.stdtime.gov.tw",szBuff,256); > if(spstdtime!=NULL) > { > printf("h_name=%s\n",(spstdtime->h_name)); > printf("h_aliases=%s\n",*(spstdtime->h_aliases)); > printf("h_addrtype=%d\n",spstdtime->h_addrtype); > printf("h_length=%d\n",spstdtime->h_length); > printf("h_addr_list=%s\n",*(spstdtime->h_addr_list)); > } > } > This is the result. > > h_name=clock.stdtime.gov.tw > h_aliases=(null) > h_addrtype=2 > h_length=4 > h_addr_list= > > > Who can tell me what happen ? I use netxray to capture the packets. You did get the address; you just misunderstood what format you got it in! h_addr_list is not supposed to be a string. h_addrtype will be AF_INET, h_length will be 4, and h_addr_list[0] will point to a 4 byte IP addr in network order. To fix your program, try this instead: - -----------replace this----------- printf("h_addr_list=%s\n",*(spstdtime->h_addr_list)); - -----------with this----------- unsigned int addr; addr = ntohl (*(unsigned int *)(spstdtime->h_addr_list[0])); printf ("h_addr_list = %d.%d.%d.%d\n", (addr >> 24) & 0xff, (addr >> 16) & 0xff, (addr >> 8) & 0xff, addr & 0xff); - ----------------------------- DaveK - -- moderator of alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! Proud Member of the Exclusive "I have been plonked by Davee because he thinks I'm interesting" List Member # Master of Many Meowing Minions Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage above and beyond the call of hilarity. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: DOS key with timeouts, idea? followup using Select() Date: Fri, 8 Feb 2002 15:33:46 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: <9b6f8883.0202061148.5789ec09@posting.google.com> <9b6f8883.0202061613.584be7e5@posting.google.com> <9b6f8883.0202071134.685fc409@posting.google.com> <9b6f8883.0202071404.1132e1dc@posting.google.com> "Mark Chang" wrote in message news:9b6f8883.0202071404.1132e1dc@posting.google.com... > --well, it turned out that it wasn't a problem with > ProComm plus nor the Telnet. > What is happening (I think) is that the Ethernet > driver packs the entire line in one packet > but the select() only wakes up once. > Since I was expecting one select return for > one key, this caused a problem when one select return > for the entire ethernet package. > So it just sat there and waited for the next key hit. > > I think the solving is using a read() until the > buffer is empty. > > Mark After a select, you should use an ioctl FIONREAD to find out how much data has actually arrived, then read it all. You should be careful how you use read and pay attention to whether the socket is blocking or non-blocking (ioctl FIONBIO) since that is how you decide whether you want it to return 0 bytes read when there is no data or for it to suspend your task until some data arrives. Look at the documentation for sockLib and ioctl. DaveK - -- moderator of alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! Proud Member of the Exclusive "I have been plonked by Davee because he thinks I'm interesting" List Member # Master of Many Meowing Minions Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage above and beyond the call of hilarity. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: return to WindSh from a task Date: Fri, 8 Feb 2002 15:41:18 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: <3c625260$1@brateggebdc5.br-automation.co.at> "Banibrata Dutta" wrote in message news:ccecb365.0202071126.6c4703ea@posting.google.com... > > However, this will _not_ create a task. Rather, it will execute the function > > taskA() in the context of the shell and return to the shell as soon as it > > 'falls of the end' of the function taskA() - or hits a return statement in > > taskA(). The result of the function will then be displayed in the shell and > > you are ready for entering your next command. > > somehow this was not happening on the free download Tornado > Prototyper, since that's exactly what i expected, but strangely at my > workplaace the VxWorks simulator / Tornado does what you just said, > it's a commercial version. i doubt that, that is the cause. wd look > again at the application i wrote at home. You just have to follow the implications of what your code is doing: tShell task is running. You say "taskA()" to shell and press return tShell task does not run taskA() as a separate task, because you did not use "sp" command to spawn it. tShell task in fact looks up address of taskA() and calls it as a subroutine. Function taskA(), running in context of tShell, looks up current task - which is tShell - and deletes it. tShell is correctly terminated by this call in taskA. tShell is no longer running. Just as your program wished! The correct way to exit from taskA and return to the shell is to place a 'return' statement in taskA, or let the code flow reach the end of taskA and fall through. DaveK - -- moderator of alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! Proud Member of the Exclusive "I have been plonked by Davee because he thinks I'm interesting" List Member # Master of Many Meowing Minions Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage above and beyond the call of hilarity. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado-2.1/mips and cp7000 BSP (Ocelot MIPS board) Date: Fri, 8 Feb 2002 17:41:12 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: <_nU88.2234$5o.598776@newsr2.u-net.net> References: "Andray Kaganovsky" wrote in message news:Xns91ADE4F242793andrayk@216.254.136.26... > > My feeling is that there is something fundamentally broken in > the MIPS64gnu version of vxWorks 5.4 binaries that come with > Tornado-2.1/MIPS. > > regards, Andray I don't know much about MIPS so I'll just offer up one quick guess: is there some calling-convention restriction on the alignment of the stack pointer that isn't being complied with by the task startup code, perhaps? It looks very similar to problems that can arise with the 8-byte-aligned ppc eabi sp vs. the 16-byte aligned altivec ppc case; the compiler may make assumptions about the low bits of the sp value when performing sp-relative arithmetic/addressing and optimize an add or subtract to an OR or AND that goes wrong if those low bits aren't as expected. DaveK - -- moderator of alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! Proud Member of the Exclusive "I have been plonked by Davee because he thinks I'm interesting" List Member # Master of Many Meowing Minions Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage above and beyond the call of hilarity. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to make Ctrl-X handler wait for a routine to finish Date: Fri, 8 Feb 2002 10:04:22 -0800 From: "Mark Galecki" Message-ID: <3c641302$0$5916$4c41069e@reader1.ash.ops.us.uu.net> References: <3c63076c$0$29348$4c41069e@reader1.ash.ops.us.uu.net> Yes of course I have :). However, rebootHookAdd() does not do what I want - it lets you plug your own routines into the Ctrl-X handler. My problem is different - I want to _synchronize_ the routine that was _already_ running with the Ctrl-X handler. I cannot pass information from that routine to the routine that I plug into the Ctrl-X handler (if I could, I would make the handler simply finish the work). If I try to use the filesystem to deposit the information, I will still have a synchronisation problem - the handler might be called between the time I take an action and the time it is recorded. So I think the only way is to make the handler wait until I finish - but how? "timallen" wrote in message news:cCN88.7459$as2.1282635@news6-win.server.ntlworld.com... > Have you looked at rebootHookAdd in rebootLib? I'm not sure at what point in > the reboot sequence the hooks are called though. > > Tim. > > "Mark Galecki" wrote in message > news:3c63076c$0$29348$4c41069e@reader1.ash.ops.us.uu.net... > > Hello, > > > > I want to make the Ctrl-X handler wait for a routine to finish. That is, > > after that routine starts execution and before it finishes, I do not want > > the soft system reboot to happen. Can I do that and how? > > > > Thank you, > > Mark Galecki > > SBS Technologies > > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Wind Web Server question Date: Fri, 08 Feb 2002 13:07:32 -0500 From: david lindauer Message-ID: <3C6413E4.E83772C7@notifier-is.net> References: <3c5ff2e7@news.012.net.il> Yaron Cohen wrote: > Hi, > > I am learning Wind Web Server and I have two questions: > > 1. Where can I find the online "Wind Web Server Reference" ? > don't know. There is a reference (HTML pages) that gets put on your hard drive from install, if using wind web 2.0 you can (in the tornado ide) select 'help' 'manuals' and then find the wind web server manual. Otherwise you have to knwo a file name to give to IE... but the manuals are kind of lacking, I usually end up looking in the header files for function definitions then just guessing at what functions will perform the activities I want... > > 2. Is it possible to associate two user RPMs for the same page ? > It is possible, but only one of the RPMs can actually spit page content. Additionally you may have to write the RPMs yourself since the system RPMs return the code meaning 'quit processing the URL'. > > I would like the page to be treated both by the SSI RPM and by the function > multiplexer RPM. > it might be more natural to use the evaluate RPM in conjunction with the SSI RPM for this. Just glancing at the book it doesn't look like the function multiplexer will buy you anything over you can't get other ways... you might want to look at their examples of password configuration (e.g. find the pwd_conf.c file in step 7 of the tutorials and look at the httpRpmDynPwdConf routine; try to match it to webfiles\accessctrl\user\user.htm). That will show you how to intercept a submit button after running the evaluate RPM manually; then you process the submit button and finally do a manual call to the SSI RPM to get it to serve up another version of the page (assuming you are doing get-oriented submits). You can either have multiple submit buttons, or use the symbol table to associate a combo box with a variable and use the variable to determine what function the submit will cause to happen. David > Thanks, > > Yaron --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Intermittent reboots Date: 11 Feb 2002 05:59:02 GMT From: Denis Perelyubskiy Message-ID: References: <3c61a1f2$1@brateggebdc5.br-automation.co.at> <3sp88.4901$AV5.107516@rwcrnsc51.ops.asp.att.net> * Gary M [02-Wed-02 22:49 -0800]: > > I agree that this sounds like a multiple bus fault, resulting in a CPU > reset. > As someone who has been there, make sure you have valid entries in your IDT > for all possible exceptions, and make sure that neither the IDT nor GDT > become corrupted. Since they are basically static structures after > initialization, it is sometimes helpful to have a function which validates > the contents of the IDT and GDT after every task switch or at periodic > intervals. just wanted to thank everyone, and just for the benefit of whoever is going to google upon this thread, state that our particular problem seems to have been the watchdog resetting the box the suggestions were very educating nevertheless. as far as this specific issues related to IDT/GDT, *i think* that the pages where they are stored are write-protected after initialization... (will make sure) thanks. > "Denis Perelyubskiy" wrote in > message > news:slrna63err.csm.denis_newsREMOVE-TO-REPLY@finch.lecs.cs.ucla.edu... > > * Werner Schiendl [02-Wed-02 13:36 -0800]: > > > > > > Hi, > > > > > "Martin Read" wrote in message > > > news:YqC*XMfgp@news.chiark.greenend.org.uk... > > > > In article , > > > > Bill Pringlemeir wrote: > > > > >I don't know if there is such a thing as a `double bus fault' in the > > > > >x86. However, this often causes a reset on other processors. Ie, > an > > > > >exception in an exception. > > > > > > > > Intel architectures from 286 onwards define a "double fault" > exception in > > > > protected mode, which occurs if certain fault-type exceptions occur > during > > > > the attempt to start the handler for another exception (e.g. page > fault > > > > upon trying to access the page fault handler). If a further > exception > > > > occurs during the attempt to start the DF handler, a triple fault > results > > > > and a CPU reset occurs. > > > > > > > > > AFAIK, the reset upon tripple fault is done by the chipset, not the CPU > > > itself. > > > You may therefore be able to 'see' if this is the cause - maybe... > > > > Thanks for the suggestions to all who replied. Something > > good to go look into ! > > > > denis > > > > -- > > if you want to reply by mail, please tweak my > > address in a reasonable way. sorry for the > > inconvenience. denis - -- if you want to reply by mail, please tweak my address in a reasonable way. sorry for the inconvenience. --------------------------- Newsgroups: comp.os.psos,comp.realtime,comp.arch.embedded,comp.os.vxworks Subject: Measuring performance of Embedded Software. Date: 10 Feb 2002 22:21:05 -0800 From: ashokchandran@chequemail.com (akosoto) Organization: http://groups.google.com/ Message-ID: Hi all, We were trying to measure the performance of our real time software which is a PSOS based system. We basically want to measure the performance of our device drivers, DLC software and the kernel. Can you please give us guidelines on how to go about it. Any methods,samples,software,links etc... with regards, Ashokc. --------------------------- Newsgroups: comp.os.vxworks Subject: 64-bit operations Date: Mon, 11 Feb 2002 06:25:53 GMT From: "Nejhdeh Ghevondian" Organization: Integrated Spectronics Message-ID: Reply-To: "Nejhdeh Ghevondian" Dear anyone, I am trying to implement 64-bit operations in my VxWorks operations (using the GNU compilier, on PowerPC platform). I am trying to multiply and divide unsigned long long int variables. However, I always get 0 as value. I acnt even do a printf("%llu\n", 64bit_var) Can anyone help me??? Regards Nejhdeh Ghevondian --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to make Ctrl-X handler wait for a routine to finish Date: Fri, 8 Feb 2002 10:51:21 -0800 From: "Mark Galecki" Message-ID: <3c641e06$0$5915$4c41069e@reader1.ash.ops.us.uu.net> References: <3c63076c$0$29348$4c41069e@reader1.ash.ops.us.uu.net> <3c641302$0$5916$4c41069e@reader1.ash.ops.us.uu.net> OK so I see one way to do it: 1. use intLock() to lock out Ctrl-X 2. save the state in a global variable 3. intUnlock() 4. add a reboot handler hook that will look at the state and finish the work from there 5. then the system can do its own normal soft reset I will try it, not now, later, so does anyone see any problem with the above scenario? Thank you. "Mark Galecki" wrote in message news:3c641302$0$5916$4c41069e@reader1.ash.ops.us.uu.net... > Yes of course I have :). However, rebootHookAdd() does not do what I want - > it lets you plug your own routines into the Ctrl-X handler. My problem is > different - I want to _synchronize_ the routine that was _already_ running > with the Ctrl-X handler. > > I cannot pass information from that routine to the routine that I plug into > the Ctrl-X handler (if I could, I would make the handler simply finish the > work). If I try to use the filesystem to deposit the information, I will > still have a synchronisation problem - the handler might be called between > the time I take an action and the time it is recorded. So I think the only > way is to make the handler wait until I finish - but how? > "timallen" wrote in message > news:cCN88.7459$as2.1282635@news6-win.server.ntlworld.com... > > Have you looked at rebootHookAdd in rebootLib? I'm not sure at what point > in > > the reboot sequence the hooks are called though. > > > > Tim. > > > > "Mark Galecki" wrote in message > > news:3c63076c$0$29348$4c41069e@reader1.ash.ops.us.uu.net... > > > Hello, > > > > > > I want to make the Ctrl-X handler wait for a routine to finish. That > is, > > > after that routine starts execution and before it finishes, I do not > want > > > the soft system reboot to happen. Can I do that and how? > > > > > > Thank you, > > > Mark Galecki > > > SBS Technologies > > > > > > > > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: system debug mode Date: 11 Feb 2002 05:19:13 -0800 From: ruchama@septier.com (Ruchama Goldshmidt) Organization: http://groups.google.com/ Message-ID: <5df340e0.0202110519.46251276@posting.google.com> Hello vxWorkers!! I am working with tornado 2 on a PPC8260 board. Now,in my project I try to include the system debug mode utility but , when I open a target shell I get a timeout error I checkout and see that all the congiguration details are correct !! (WDBSERIAL :38400,COM2) is anyone has an idea what else should I do?? thanks in advance!!! Ruchama Goldshmidt --------------------------- Newsgroups: comp.os.vxworks Subject: Re: multiple instances of a variable Date: 8 Feb 2002 12:58:18 -0800 From: akarpov@my-deja.com (Alex) Organization: http://groups.google.com/ Message-ID: <4f09bb68.0202081258.4f0bf8ca@posting.google.com> References: <3C63EA94.E47EFE94@res.raytheon.com> Nick, Your tasks share the same global variables unless you spawn your tasks in different domains under VxWorks AE. Alex willisn wrote in message news:<3C63EA94.E47EFE94@res.raytheon.com>... > Hi all, > > My question concerns multiple instances of a single global variable. > > I have 3 tasks running, all have an instance of a global variable (say > an integer). > > If I want to know the current value of this integer for any one of the > tasks, I need to do a lkup on the variable and then dereference the > address associated with the task I am interested in. > > My question is, if you simply type the name of the variable in the > console, how does vxworks decide which instance of the variable to give > you? > > Thanks in advance for any help! > > Nick --------------------------- Newsgroups: comp.os.vxworks Subject: Re: how i build bootrom Date: 8 Feb 2002 13:09:51 -0800 From: akarpov@my-deja.com (Alex) Organization: http://groups.google.com/ Message-ID: <4f09bb68.0202081309.2e2c0c0d@posting.google.com> References: Go and double check if there are such a file in the main Tornado tree Tornado/target/src/drv/intrCtl/m8260IntrCtl.c or in your BSP : $(BSP_DIR)/src/drv/intrCtl/m8260IntrCtl.c alimohd@logiceastern.com (Ali) wrote in message news:... > > > make: *** No rule to make target `drv/intrCtl/m8260IntrCtl.c', needed > > > by `sysLib.o'. Stop. vxworks --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to link with the BSD socket library Date: Thu, 7 Feb 2002 14:23:27 -0500 From: "Bhaskar Rao" Message-ID: <3c67d0ea$0$25356$4c41069e@reader1.ash.ops.us.uu.net> References: <17b6d9e7.0202070733.7eadab7a@posting.google.com> 1. Do this only if you have full simulator - Go to $WIND_BASE\target\config\simpc\config.h file(Line Number 52, make TRUE as FALSE and save the file. IMPORTANT, you have to modify this, before creating the workspace and project 2. Now open up Tornado and create a new project for bootable vxworks image(Custom configured) and select BSP(simpc) in the option for basing your project. Select VxWorks tab in the workspace window and include the following components (if the components are already included, then ignore) a. network components->basic network init add "basic network support" b. network components->bootline processing->network init add "init network at boot" c. network components->networking protocols-> core TCP/IP add ICMPv4 add IGMPv4 d. development tools->WDB agent-> select WDB connection remove "WDB simulator pipe" add "WDB END driver" e. Include Ping Client under network components > network protocols > network applications > PING client f. Locate network components-> network protocols-> network debugging and select UDP show routines TCP show routines ICMP show routines IGMP show routines Network show routines g. Under development tool components, locate show routines and include memory show routines stdio show routines symbol table show routines task show routines h. Also under development tool components, include the downloaded symbol table: symbol table components symbol table initialization components select symbol table initialization downloaded symbol table i. Include error status table in symbol table components -> error status table j. Locate target shell component and include it development tool components ->target shell components -> target shell (target debugging will automatically be included.) k. Right-click in the VxWorks tab of the Project Facility and select "Rebuild all (vxWorks.exe). 3. Launch a VxSim, from the dialog box that appears, select "Custom-built simulator". Click on "Browse" and navigate to the directory where the vxWorks.exe image is located, i.e. $WIND_BASE\target\proj\simPROJ\default\vxWorks.exe. When successfully launched a message appears "A Target Server named "vxsim@xxx" is required and will be started." IMPORTANT: click on CANCEL Clicking on OK launches a target server that fails: (The automatically launched target server fails because the name of the target is vxTarget, as hostShow will indicate later. The name of the target is hardcoded in configAll.h to vxTarget.) INSTEAD, from Tornado menu: select Tools > Target Server Configure the target server to use wdbrpc backend, to attach to a target with IP address 90.0.0.1. The target server command line should look like this: tgtsvr.exe 90.0.0.1 -n vxTarget1 -V -B wdbrpc Click on "Launch" Confirm that the Target Server connected to the target by double-clicking on the small red and white, bulls-eye icon that appears in the lower right hand corner of the desktop. A window called vxTarget1@xxx Log Console appears. Now your simulated target with IP address 90.0.0.1 is available. (Ensure you have installed ULIP driver properly before you begin) - - Bhaskar Rao "Cetus" wrote in message news:17b6d9e7.0202070733.7eadab7a@posting.google.com... > Hi, > > I am new to VxWorks and Tornado and have been > searching for answers for the folloing question. > I am trying to test a simple socket program > in the simulator. The program compiles without > problems. When I try to load it from the shell, > it complained about undefined symbols from the socket > library. > If I try to build a bootable image that > library libSIMNTgnuvx.a, it failed > at the final linking stage with those undefined symbols. > What should I do to make the socket library available ? > > Thanks for any help. > > -> ld < Project0.out > Loading C:/Tornado/target/proj/Project0/SIMNTgnu/Project0.out | > Undefined symbols: > _socket > _recvfrom > _bind > Warning: object module may not be usable because of undefined symbols. > value = 17433544 = 0x10a03c8 > -> --------------------------- Newsgroups: comp.os.vxworks Subject: Wind Web Server - problem in httpSymTabAdd Date: Mon, 11 Feb 2002 16:54:20 +0200 From: "Yaron Cohen" Message-ID: <3c67dbde@news.012.net.il> Hi, I am getting "symbol xxx not found! SSI failed " (8 times for each error) after trying to add symbols to httpSymTabAdd() in the follwing way: - ----------------------------------------------------------- char buf[50]; for(i = 1; i < 9; i++) { sprintf(buf,"AutomaticTAOS%d", i); symId = httpSymTabAdd (httpGlobalSymTabId, buf, "", HTTP_SYM_TYPE_BOOL_ID, 1, &T1_interface[i-1].BitsClockEnable); } - ---------------------------------------------------------------- When I am trying to do it without using 'buf' it works fine (without the loop): - ----------------------------------------------------------- symId = httpSymTabAdd (httpGlobalSymTabId, AutomaticTAOS1, "", HTTP_SYM_TYPE_BOOL_ID, 1, &T1_interface[0].BitsClockEnable); symId = httpSymTabAdd (httpGlobalSymTabId, AutomaticTAOS2, "", HTTP_SYM_TYPE_BOOL_ID, 1, &T1_interface[1].BitsClockEnable); - ------- . . . | symId = httpSymTabAdd (httpGlobalSymTabId, AutomaticTAOS8, "", HTTP_SYM_TYPE_BOOL_ID, 1, &T1_interface[7].BitsClockEnable); - --------------------------------------------------- Any idea ? Thanks, Yaron Cohen --------------------------- Newsgroups: comp.os.vxworks Subject: Re: ~!~ C-Designer (mit RoseRT Ausbildung) für Realtime/C/Systeme sucht neues Projekt ~!~ Date: Mon, 11 Feb 2002 17:09:54 +0100 From: "Werner Schiendl" Message-ID: <3c67ecd2$1@brateggebdc5.br-automation.co.at> References: You do not seem very well adaptable. At least you are not able to adjust your spamming to the proper language used for this list. You've already been advised to do so. regards Werner "M.Thoemel" wrote in message news:a41b2f$7pd$04$1@news.t-online.com... > Guten Tag, > > ich bin freier Sofwareentwickler mit einem sehr breiten Spektrum. > > News: > > Aktuell habe ich einen Kurs über UML und Rational Rose RT absolviert und > werde diese SOftware jetzt zwei Monate für ein VxWorks Target einsetzen. > > Ab dem April 2002 stehe ich wieder zur Verfügung. > > Meine Schwerpunkte sind: > > - C (aber auch andere Sprachen!) > - Realtimesysteme > - systemnahe Programmierung > - spezielle Hardware-Schnittstellen > - Netzwerke > - Treiber > - Streamsdriver > - Grafische Datenauswertung/Darstellung > - Software-Messungen > - Gruppenleitung > - Projektierung > - Projektleitung > etc... > > Sie können sich im Detail informieren unter: > > http://matthias.thoemel.bei.t-online.de > > Ich würde mich über einen Kontakt mit Ihnen freuen. > > Ihr Matthias Thömel > (matthias.thoemel@t-online.de) > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: motFccEnd.c Date: Thu, 7 Feb 2002 14:54:13 -0800 From: "Dan Gold" Organization: Mailgate.ORG Server - http://www.Mailgate.ORG Message-ID: References: <3c630037$1@news.si.com> Sender: gold@ensemble.com "vanbaren_gerald" wrote in message news:3c630037$1@news.si.com... > I have not done it myself, but I recall seeing warnings that you need > to make sure the two FCCs are using different areas of DPRAM for their > buffers (RIPTR, TIPTR). My recollection is that the WR driver can > support any FCC port, but it has a hardcoded offset that it uses for > RIPTR and TIPTR. Correct. I pointed that out in an earlier post (that was deleted for brevity) when I said the WRS driver was buggy. The OP needs to talk to their local WRS FAE and make sure they either have a fixed version of the driver, or source so he can fix it himself. Dan Gold gold@ensemble.com --------------------------- Newsgroups: comp.os.vxworks Subject: How to make Ctrl-X handler wait for a routine to finish Date: Thu, 7 Feb 2002 15:02:42 -0800 From: "Mark Galecki" Message-ID: <3c63076c$0$29348$4c41069e@reader1.ash.ops.us.uu.net> Hello, I want to make the Ctrl-X handler wait for a routine to finish. That is, after that routine starts execution and before it finishes, I do not want the soft system reboot to happen. Can I do that and how? Thank you, Mark Galecki SBS Technologies --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Questions about vxmalloc.c Date: 11 Feb 2002 12:25:21 -0500 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: References: Sender: bpringlemeir@DeadDuck >>>>> "Rush" == Rush Manbert writes: Rush> I have gotten a copy of vxmalloc.c, and I was trying to figure Rush> out the changes relative to Doug Lea's original, so I could Rush> apply the same changes to the latest version. I came across a Rush> number of macros that were redefined in vxmalloc.c, and I'm Rush> trying to understand why the changes were made. These include Rush> MALLOC_ALIGNMENT, chunk2mem, and mem2chunk. Rush> So my question is, has anyone used this code and analyzed why Rush> these changes exist, and/or does anyone know a current email Rush> address for Hwa-Jin Bae, the author? Don't use the vxMalloc.c on the FAQ sight. * VERSION 2.7.0 Sun Mar 11 14:14:06 2001 Doug Lea (dl at gee) ftp://gee.cs.oswego.edu/pub/misc/malloc.c This is a more recent version with nice little fixes for people wanting to use Doug's allocator on vxWorks. It is also not clouded with the BSD code that `Hwa-Jin' included. There is a provision to use a custom header file. I ended up modifying the source. I have include pertinent info at the end of this message. I have removed the stupid inter-twining of the vxWorks memory libraries. Ie, memPartLib contains only partition functions. valloc, calloc, etc are in memLib and they have been replaced by `vxMalloc.c'. You might have to code another vxSbrk if you don't do this. The info about MALLOC_ALIGNMENT, chunk2mem, mem2chunk is provided in Doug Lea's web page "http://g.oswego.edu/dl/html/malloc.html". I doubt that I could provide a more insightful comment. Times, they are a changin' Bill Pringlemeir. [start changes] #define MORECORE vxSbrk #define MORECORE_CONTIGUOUS 0 #define MORECORE_CANNOT_TRIM 1 #define malloc_getpagesize (64*1024) struct malloc_state { /* The maximum chunk size to be eligible for fastbin */ [snip...] /* vxWorks semaphore. */ SEMAPHORE mem_semaphore; }; #define MALLOC_PREACTION semTake(&(get_malloc_state()->mem_semaphore), \ WAIT_FOREVER) #define MALLOC_POSTACTION semGive(&((get_malloc_state()->mem_semaphore))) void vxMemInit(void) { semBInit(&(get_malloc_state()->mem_semaphore), SEM_Q_PRIORITY, SEM_FULL); } void *vxSbrk(int size) { void *tmp; static void *previous_top; if(size) { if(!(tmp = memPartAlloc(memSysPartId, size+MALLOC_ALIGNMENT))) return (void*)-1; /* Must be aligned. */ tmp = (void*)(((int)tmp + MALLOC_ALIGN_MASK) & ~MALLOC_ALIGN_MASK); previous_top = (void*)((int)tmp + size + MALLOC_ALIGNMENT); return tmp; } return previous_top; } [end changes] - -- quiche counter-intelligence domestic disruption $400 million in gold bullion arrangements Rule Psix Marxist South Africa bomb class struggle cracking [Hello to all my fans in domestic surveillance] ammunition plutonium Soviet vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" --------------------------- Newsgroups: comp.os.vxworks Subject: Re: PCI Device Base Address Date: 11 Feb 2002 09:35:24 -0800 From: arun4vxworks@indiatimes.com (Arun Prasad) Organization: http://groups.google.com/ Message-ID: References: <3C617FFA.814299A3@aps.anl.gov> Hi, Our sysLib.c includes "drv/pci/pciAutoConfigLib.h". But still we faced the problem of not getting the proper Base Address. Our Card requires two 64bit Base Address Registers. Somehow they are not getting initialized properly. We tried to use pciDevConfig() function. But that function is meant only for 32 bit base address registers and that too for not more than one base address register. This is what pciConfigLibc. says........ ========= The function pciDevConfig() is used to configure PCI devices that require no more than one Memory Space and one I/O Space. According to the PCI standard, a device may have up to six 32-bit Base Address Registers (BARs) each of which can have either a Memory Space or I/O Space base address. In 64-bit PCI devices, the registers double up to give a maximum of three 64-bit BARs. The 64-bit BARs are not supported by this function nor are more than one 32-bit BAR of each type, Memory or I/O. ............ ............ /* * Due to system constraints, this is a partial implementation * of enabling the Base Address Registers (BARs). * It is hoped that all PCI devices only require at most one * I/O space and/or one memory space. * If not, this code will re-allocate the device's memory to * each BAR implemented. Sounds like future trouble! */ ========= Atlast we wrote our own pciDevConfig() to handle multiple 64bit base address registers. This has been done by using the existing pciDevConfig() routine with the required changes taken from Linux implementation of the equivalent function. Now we are getting the expected values in the Base Address Registers fields. Still we are in the process of testing Register Acccess. I still wonder/doubt whether VxWorks supports PCI cards with multiple 64 bit Base addresses or not. It's sure that thru the existing pciDevConfig(), we can't handle this scenario. We went for configuring our device on our own, just because we haven't got the proper/correct Base Adresses. Regards Arun Andrew Johnson wrote in message news:<3C617FFA.814299A3@aps.anl.gov>... > Arun Prasad wrote: > > > > I'm using IBM PPC405GP walnut board. > > > > I'm facing some problem in getting the Base address of my PCI based > > NIC Card. > > > > I'm not getting the correct base address of our PCI based NIC card. > > I'm getting a value which doesn't fall within the PCI memory or I/O > > range. I feel like, my device hasn't been properly initiated. I'm able > > to find my card using pciFindDevice() call. I am able to find its IRQ > > no. I'm able to dump the PCI header for our card using pciConfigInXXX > > functions. But base address are wrong. I don't know whats wrong with > > that. > > Don't forget that the base address is an address on the PCIbus, and will > need translation to convert it into a CPU address. I've never used the > walnut so I don't know how its BSP maps PCI addresses to CPU addresses, > but it's likely that you just need to add the base address of the PCI > memory (or I/O) range. > > > ********How to choose Base Address for my device on my wish? > > How to assign base address for our card statically? > > ********How to program the PCI configuration space to mention > > I have chosen my own base address for my card? > > That typically depends whether your sysLib.c file includes > "drv/pci/pciConfigLib.h" or "drv/pci/pciAutoConfigLib.h". If only the > former then you will have to choose the address yourself, but if the > latter then it's probably being set for you and you should not attempt to > change the base address register yourself, just read it and convert to a > CPU address. > > - Andrew --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado Training Date: Mon, 11 Feb 2002 09:43:01 -0800 From: "Michael R. Kesti" Organization: MK Associates Message-ID: <3C6802A5.4F0ECF64@gv.net> References: Reply-To: mkesti@gv.net "news.sympatico.ca" wrote: >Tornado Training is being held in Toronto Ontario, has anyone gone to it and >if so was it worth the $$? I've taken the Tornado class in Alameda. If I had been paying for it, rather than my employer, I might have thought that it was a lightweight for the price. OTOH, it did give me a jump start that easily saved me weeks of self-teaching/learning and therefore probably did pay off. If the class is local enough that you stay home, I'd say it's easily worth the money. If you add the price of away-from-home accomodations, the value probably depends on the cost of those accomodations. - -- ======================================================================== Michael Kesti | "And like, one and one don't make | two, one and one make one." mkesti@gv.net | - The Who, Bargain --------------------------- Newsgroups: comp.os.vxworks Subject: Re: pciFindDevice()- index 0 based or 1 based ? Date: 11 Feb 2002 09:47:39 -0800 From: arun4vxworks@indiatimes.com (Arun Prasad) Organization: http://groups.google.com/ Message-ID: References: <3c612508$1@netnews.web.de> Hi, somehow, something messed up in initializing our PCI based NIC card. It gave wrong Base addresses. We tried to use pciDevConfig() after that our pciFindDevice() problem solved automaticaly. But later we found that the current implementation of pciDevConfig() doesn't support our card as our card requires multiple (two) 64-bit base address registers. Atlast we made some changes in pciDevConfig() and now it seems everything fine as far as pciFindDevice() and Base address registers are considered. We are working on accessing the registers of our card. After that we'll be in a position to conclude everything works fine. Regards Arun ====== Arun Prasad Software Engineer Adaptec India Pvt. Limited (A wholly owned subsidiary of Adaptec Inc., USA) 6-3-1086, IV Floor, Vista Grand Towers, Rajbhavan Road, Somajiguda, Hydereabad - 500082 Phone: (O):91-40-6661555/56/57/58 (Ext):234 Email: arun_prasad@adaptecin.com Reinhold Wirth wrote in message news:<3c612508$1@netnews.web.de>... > Hi Arun, > > > to your frist question: > * Can anyone tell me whether this index value starts with > 0 or 1? (of the function pciFindDevice(..,..,index,....)) > The source code should be a part of the installation > (in ..../target/src/drv/pci/pciConfigLib.c) > The index=0 finds the 1. "pci-card" with the defined deviceId and vendorId. ... > > > I think, you have to get OK from the function pciFindDevice with the vendorId, > deviceId and index=0. > You wrote, you "got a success" with the index=1. Ar you sure? > > > Your funktion calls should look like: > pciFindDevice (0x1014, 0x0156, 0, &busNo, &deviceNo, &fktNo) > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: XON-XOFF on the Console port Date: Mon, 11 Feb 2002 15:02:57 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: <2002130-143411-880483@foorum.com> <200221-134637-250830@foorum.com> <1013284083.9121.0.nnrp-01.d4e4b8d1@news.demon.co.uk> "Leonid Rosenboim" wrote in message news:newscache$604brg$i12$1@lnews.actcom.co.il... > Cecil, I beleive you have the case of badly designed RS232 hardware. > If the designers follow the standard closely, and use whatever application > notes they can get from whoever makes the driver chip they use, this would > not have happened. EIA calls for a low impedance of the Rx signal, > which would absorb any EMI generated signals, and would supress any > xross talk (what you refer to as reflection) from the Tx signal. > Your hardware designers created the Rx with a high impedance, > which will cause trouble (e.g. FEXT) even if the cable was connected. > Cheap and rubbishy PCs often spit out some garbage down the serial port when the power comes on. Followed by some more garbage, *deliberately*, during booting to try and detect a serial mouse. It can be a right nuisnance if you are used to leaving your PC and your target board permanently connected by a serial cable and switching them on simultaneously at a single mains socket... DaveK - -- moderator of alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! Proud Member of the Exclusive "I have been plonked by Davee because he thinks I'm interesting" List Member # Master of Many Meowing Minions Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage above and beyond the call of hilarity. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Questions about vxmalloc.c Date: Mon, 11 Feb 2002 14:30:42 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: "Rush Manbert" wrote in message news:a294ea3e.0202081842.42362789@posting.google.com... > I have gotten a copy of vxmalloc.c, and I was trying to figure out the > changes relative to Doug Lea's original, so I could apply the same > changes to the latest version. I came across a number of macros that > were redefined in vxmalloc.c, and I'm trying to understand why the > changes were made. These include MALLOC_ALIGNMENT, chunk2mem, and > mem2chunk. > > So my question is, has anyone used this code and analyzed why these > changes exist, and/or does anyone know a current email address for > Hwa-Jin Bae, the author? > > Thanks, > Rush Sounds like he wanted to change the memory alignment quantum to me. Wanna post the macros in before-and-after versions ? DaveK - -- moderator of alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! Proud Member of the Exclusive "I have been plonked by Davee because he thinks I'm interesting" List Member # Master of Many Meowing Minions Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage above and beyond the call of hilarity. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Equivalent of setitimer Date: Mon, 11 Feb 2002 14:33:29 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: <3C66A964.A7BDD3D0@gv.net> "Bryan" wrote in message news:cc857445.0202110030.d1946fb@posting.google.com... > "Michael R. Kesti" wrote in message news:<3C66A964.A7BDD3D0@gv.net>... > > Bryan wrote: > > > > >In Linux setitimer is there to set a timer and when the timer expires > > >SIGALRM will be sent. In VxWorks what is the equivalent of that? In > > >timer_create call where shall we mention the timer value. > > > > See the vxWorks docs for "wdLib", where "wd" stands for "watchdog" as in > > "watchdog timers". > Hi Mic, > > Yeah i already had a look at watchdog timers but they will run in ISR > context right. I dont want ISR context timers. Atleast signal handler > level or task context level timers i need. I tried with timer_settime. > It worked. THanks for the response > > Bryan > just want to point out that an ISR can call kill(...) to send a signal. but then timer_settime is probably doing pretty much that internally anyway! DaveK - -- moderator of alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! Proud Member of the Exclusive "I have been plonked by Davee because he thinks I'm interesting" List Member # Master of Many Meowing Minions Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage above and beyond the call of hilarity. --------------------------- Newsgroups: comp.soft-sys.matlab,comp.os.vxworks Subject: Re: Returning data from S-func in RTW Date: Mon, 11 Feb 2002 15:13:58 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: "Ole-Johan Ellingsen" wrote in message news:a3lhm9$srg$1@tyfon.itea.ntnu.no... > Hi again. > > Sorry, but the problem seems to have been obvious. :) > Using parrOutput[2] instead of *(parrOutput+2) did the trick. > > Ole-Johan Those two expressions *ought* to be exactly identical in every way. What was your definition of real_T ? DaveK - -- moderator of alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! Proud Member of the Exclusive "I have been plonked by Davee because he thinks I'm interesting" List Member # Master of Many Meowing Minions Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage above and beyond the call of hilarity. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: +++ C-Designer (sehr flexibel) für Realtime/C/Systeme sucht neues Projekt +++ Date: Mon, 11 Feb 2002 15:09:36 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: <3c5e5151@brateggebdc5.br-automation.co.at> > "M.Thoemel" wrote in message > news:a3f403$bi0$04$1@news.t-online.com... > > Guten Tag, "Werner Schiendl" wrote in message news:3c5e5151@brateggebdc5.br-automation.co.at... > Hi, > > If you use this list for your advertising, _AT_LEAST_ use the correct > language: > This is an english language list! > > - Werner It is? If you're referring to the newgroup, I can't see anything that specifies what language should be used: - --------------begin quote-------------- Control: newgroup comp.os.vxworks Newsgroups: comp.os.vxworks.ctl Path: rpi!tale From: tale@rpi.edu (David C Lawrence) Subject: newgroup comp.os.vxworks Message-ID: Sender: tale@cs.rpi.edu Nntp-Posting-Host: cs.rpi.edu Date: 13 Jan 92 16:43:31 GMT Approved: tale@rpi.edu Lines: 16 comp.os.vxworks is an unmoderated newsgroup which passed its vote for creation by 295:17 as reported in news.announce.newgroups on 6 Jan 1992. For your newsgroups file: comp.os.vxworks The VxWorks real-time operating system. The charter, culled from the call for votes: comp.os.vxworks will be a forum for discussion of topics related to the VxWorks real-time operating system and development environment. Current and prospective users of VxWorks will be able to discuss technical problems, report experiences with various hardware components, share code, report bugs and fixes, and exchange any information of interest to the VxWorks user community. Relevant new product announcements may be made by commercial participants, subject to a limit - e.g. 1/month - on the frequency of such posts. - ---------------end quote--------------- Whereas if you're referring to the exploder, well, it isn't Matthias' fault if the exploder went and copied a post from the newsgroup that it shouldn't have. DaveK - -- moderator of alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! Proud Member of the Exclusive "I have been plonked by Davee because he thinks I'm interesting" List Member # Master of Many Meowing Minions Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage above and beyond the call of hilarity. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: about lex Date: Mon, 11 Feb 2002 15:22:10 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: <3d64ee1a.0202052234.26dc4574@posting.google.com> "Troy Cauble" wrote in message news:a41c55$15v@nntpb.cb.lucent.com... > changwoo wrote: > : I am tired to use lex in vxWorks > : but there is some problem to compile the files. > Did you know it used global data? Will you run more than one > lexer at a time? Yep, with great certainty, since the VxWorks shell commandline parser was built using lex, so there will already be many many global symbols beginning "yy" - -> lkup "yy" yyerrflag 0x001a93d0 bss yychk 0x0019e990 data yyparse 0x0017991c text yyr1 0x0019e798 data yypgo 0x0019e778 data yystart 0x0017788c text yyexca 0x0019d8e8 data yyval 0x001b04a0 bss yychar 0x001a95fc bss yypact 0x0019e5a8 data yyr2 0x0019e894 data yylval 0x001b0b50 bss yyact 0x0019d900 data yydebug 0x001a9548 bss yydef 0x0019eb60 data yynerrs 0x001a9464 bss value = 0 = 0x0 - -> DaveK - -- moderator of alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! Proud Member of the Exclusive "I have been plonked by Davee because he thinks I'm interesting" List Member # Master of Many Meowing Minions Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage above and beyond the call of hilarity. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: MPC860 floating point emulation Date: Mon, 11 Feb 2002 14:57:42 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: "Ernie Rael" wrote in message news:acl66uojgmoch330fm6aiqbiotrddnnqal@4ax.com... > We use both MPC860 and PPC750. I'd like to compile only for > PPC750, and when running on MPC860 handle the software > emulation exception to implement floating point. There isn't > much floating point; performance isn't much of an issue. > > Does anyone know of a package that handles this. I can do > any glue if it doesn't hook directly into a vxworks exception. > I haven't looked at the gnu source for the soft-float stuff, but > I'm guessing that it would take a bit of work to adapt it. Presumably the gnu soft-float code is just in the form of subroutines that take arguments, so if you're as happy as you say to write the exception handling glue, why isn't it the answer you're looking for? Bear in mind however that if you use gnu *source* in your work, you are obliged to open your source to any customers who you distribute your binaries containing the work-derivative-from-gnu-code to. That's probably far more of an issue than adapting the code. I don't have a recommendation about what code you can use, but I thought I'd better drop in the oops-that's-gnu-SOURCE-you're-referring-to-there problem. DaveK - -- moderator of alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! Proud Member of the Exclusive "I have been plonked by Davee because he thinks I'm interesting" List Member # Master of Many Meowing Minions Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage above and beyond the call of hilarity. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: New WindSerf, TSR's and SPR's. Date: Mon, 11 Feb 2002 14:47:46 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: <488e459a.0201221531.41030518@posting.google.com> "Johan Borkhuis" wrote in message news:Xns91AB692444D63borkhuisagerecom@135.7.153.102... > Johan Borkhuis wrote: > > You could try the new message boards, there is one for WindSurf, but I > > don't know if that will help much. > > One thing that is different on these message boards is that WindRiver > employees are allowed to post there, but they are not allowed to post in > the newsgroups. > > Groeten, > Johan That doesn't stop Martin Raabe from using his WRS email address to advertise the fuck out of sniff and similar packages! DaveK - -- moderator of alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! Proud Member of the Exclusive "I have been plonked by Davee because he thinks I'm interesting" List Member # Master of Many Meowing Minions Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage above and beyond the call of hilarity. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Beginner Client/Server problems! Date: Mon, 11 Feb 2002 14:56:55 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: "Joseph Cook" wrote in message news:b6f5d6a9.0202080832.22c36d98@posting.google.com... > I am trying to send a message through a socket from a Client to a > Server and then have the Server return that message. I am trying to > do this in a loop as many times as I want. The message changes each > time as well to show which iteration it is.. > > It hangs though on the send or receive calls, and I can't figure out > what is wrong... > > Client: > char messBuff[8];\ > char replyBuff[8]; > for (i=0; i<5; i++) > { > sprintf(messBuff, "Message%d", i); buffer overflow here. 'Message' = seven chars. %d will use one since i is 0-5. That gives 8..... Plus one more for the NUL terminator = 9. Bingo, that's one byte past the end of the messBuff[8]. > if(write(Fd, messBuff, sizeof(messBuff))==ERROR) > file://ERROR > if(read(Fd, replyBuff, sizeof(replyBuff)) < 0) > file://ERROR > } > Server: > char messBuff[8]; > char replyBuff[8]; > for(i=0; i<5; i++) > { > if((nRead = read(Fd, messBuff, sizeof(messBuff))) == ERROR > { > file://ERROR > } > if(write (Fd, replyBuff, sizeof(replyBuff) == ERROR)) > { > file://ERROR > } > > Seems simple enough...what could be wrong? If there's anything else wrong, it'll be in the bits of code you didn't show here. DaveK - -- moderator of alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! Proud Member of the Exclusive "I have been plonked by Davee because he thinks I'm interesting" List Member # Master of Many Meowing Minions Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage above and beyond the call of hilarity. --------------------------- Newsgroups: comp.os.vxworks Subject: tNetTask takes too long? Date: Thu, 7 Feb 2002 17:08:22 -0600 From: Dinker Charak Organization: Fermi national Accelerator Lab Message-ID: Hello, Tornado 2.0.2/vxWorks 5.4.2 We have a task running at priority 130 that just eats CPU cyclles (no net activity). We observed (on MV2301 and MV2400 - both 5.4.2) that every ~5 s tNetTask switches in, stays for ~4.5ms and then switches out. There is no network activity, sniffer shows nothing on wire. Exactly same setup with 5.4 shows no such abnormaly long activity. Why is tNetTask taking so long? What is it doing? Anybody noticed anything like this? Any idea or remarks? Thanks, Dinker --------------------------- Newsgroups: comp.os.vxworks Subject: stack USB from WinRiver on All BSP? Date: 11 Feb 2002 17:05:54 GMT From: Charly Message-ID: <2002211-18554-313711@foorum.com> Hello, Are there somebody that are using USB stack from WindRiver. Is-it easy to configure and install? Does it use BSP capabilities or direct h/w registers to access USB's h/w ? My BSP builder doesn't know anything on USB, Do you thing I could access directly to USB component even if nothing is written for that in the BSP. Thanks you charly - -- Use our news server 'news.foorum.com' from anywhere. More details at: http://nnrpinfo.go.foorum.com/ --------------------------- Newsgroups: comp.os.vxworks Subject: Sockets Date: Thu, 7 Feb 2002 17:14:12 -0600 From: Dinker Charak Organization: Fermi national Accelerator Lab Message-ID: Hi, On linux we can send and receive using the same socket. I have observed that in VxWorks in the same program if I have to send and receive, I can not use the same socket. We need to have two sockets for this. Is my observation consistent. Can any one explain why so? I have also observed the max number of sockets that are opened is usually half (well some sockts are used by standard vxWorks activity) the number in NUM_FILES. This seems consistent with above observation. Any comments? Any idea what is happening? Thanks, Dinker --------------------------- Newsgroups: comp.os.vxworks Subject: generating a makefile (on command line) for bootable image Date: 11 Feb 2002 11:42:36 -0800 From: trish.bennett@caenewnes.com (Trish Bennett) Organization: http://groups.google.com/ Message-ID: I have successfully created makefiles for downloadable images on the command line using the following command: wtxtcl makeGen.tcl project.wpj Now I want to do the same for bootable images. I use the command wtxtcl configGen.tcl project.wpj and it creates 3 files: config.h, sysConfig.c and linksyms.c but no makefile. Can anybody help? Thanks. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: +++ C-Designer (sehr flexibel) für Realtime/C/Systeme sucht neues Projekt +++ Date: Mon, 11 Feb 2002 22:08:07 +0100 From: "Werner Schiendl" Message-ID: <3c6832b7@brateggebdc5.br-automation.co.at> References: <3c5e5151@brateggebdc5.br-automation.co.at> Dave, "Dave Korn" wrote in message news:PGT98.2307$5o.893553@newsr2.u-net.net... > > "M.Thoemel" wrote in message > > news:a3f403$bi0$04$1@news.t-online.com... > > > Guten Tag, > > "Werner Schiendl" wrote in message > news:3c5e5151@brateggebdc5.br-automation.co.at... > > Hi, > > > > If you use this list for your advertising, _AT_LEAST_ use the correct > > language: > > This is an english language list! > > > > - Werner > > It is? If you're referring to the newgroup, I can't see anything that > specifies what language should be used: > - -- snip -- that might be true, indeed. Anyway: How many posts in this newsgroup - apart from spam - are in an other language?? None (or almost, since sometimes a newbie may post in some other language). Specified or not: effectively this -is- an english language list / newsgroup. best regards Werner P. S. I hate spam, including Martin R's and this one. _Particularly_ if it is repeated every week. > --------------begin quote-------------- > Control: newgroup comp.os.vxworks > Newsgroups: comp.os.vxworks.ctl > Path: rpi!tale > From: tale@rpi.edu (David C Lawrence) > Subject: newgroup comp.os.vxworks > Message-ID: > Sender: tale@cs.rpi.edu > Nntp-Posting-Host: cs.rpi.edu > Date: 13 Jan 92 16:43:31 GMT > Approved: tale@rpi.edu > Lines: 16 > > comp.os.vxworks is an unmoderated newsgroup which passed its vote for > creation by 295:17 as reported in news.announce.newgroups on 6 Jan 1992. > > For your newsgroups file: > comp.os.vxworks The VxWorks real-time operating system. > > The charter, culled from the call for votes: > > comp.os.vxworks will be a forum for discussion of topics related to > the VxWorks real-time operating system and development environment. > Current and prospective users of VxWorks will be able to discuss > technical problems, report experiences with various hardware components, > share code, report bugs and fixes, and exchange any information of interest > to > the VxWorks user community. Relevant new product announcements may be > made by commercial participants, subject to a limit - e.g. 1/month - > on the frequency of such posts. > ---------------end quote--------------- > > Whereas if you're referring to the exploder, well, it isn't Matthias' > fault if the exploder went and copied a post from the newsgroup that it > shouldn't have. > > DaveK > -- > moderator of > alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow > Burn your ID card! http://www.optional-identity.org.uk/ > Help support the campaign, copy this into your .sig! > Proud Member of the Exclusive "I have been plonked by Davee because he > thinks I'm interesting" List Member # > Master of Many Meowing Minions > Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage above > and beyond the call of hilarity. > > > --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Tue Feb 12 08:57:58 2002 From: Ronen Levy Date: Tue Feb 12 08:58:00 PST 2002 Subject: reboot from Target Shell fails Hi, I am trying to reboot the target (evaluation board: estMdpBasic821/860 from Wind River). The BSP loads the vxworks image from the Network, and then I get the Target Shell prompt. At this point I am executing: reboot 0 The problem: Evaluation Board does Not respond. I have VisionClick debugger thru the BDM port, that lets me know of Machine Check occuring. Q: I would like to know if what I am trying to achieve (reboot the target) is doable. What could possibly be the problem? Ronen. Press any key to stop auto-boot... 0 auto-booting... boot device : cpm unit number : 0 processor number : 0 host name : diags file name : d:\tornado\target\proj\vxworks\default\vxworks inet on ethernet (e) : 172.17.23.13 host inet (h) : 172.17.12.52 gateway inet (g) : 172.16.0.1 user (u) : anonymous ftp password (pw) : user flags (f) : 0x8 target name (tn) : estMdpBasic8xx Attached TCP/IP interface to cpm0. Attaching network interface lo0... done. Loading... 836776 Starting at 0x100000... Attached TCP/IP interface to cpm unit 0 Attaching interface lo0...done Adding 3119 symbols for standalone. VxWorks Copyright 1984-1998 Wind River Systems, Inc. CPU: EST Corp. estMdpBasic821/860 -- Mini PowerPC VxWorks: 5.4 BSP version: 1.2/3 Creation date: Feb 12 2002 WDB: Ready. -> reboot 0 From vxwexplo-errs@csg.lbl.gov Tue Feb 12 20:28:50 2002 From: "Sajjad Ahmed" Date: Tue Feb 12 20:28:52 PST 2002 Subject: Problem in Flushing the Routing entries.(STRANGE READ THIS) Hello VxWorks Gurus, I met a typical problem problem this morning, taking your less time let me explain to you the short scenario. 1) i create an ethernet interface "eth1" using ether_attach. ->routeShow destination gateway flags Refcnt Use Interface ---------------------------------------------------------------------------- 20.0.0.1 20.0.0.1 101 0 0 eth1 the physical connection as follows 20.0.0.1 |--------------| |-----------|20.0.0.4 | |HUB| | ROUTER | | | | HOST 2)telnet to a host on 20.0.0.0 n/w with ip address 20.0.0.4 i see telnet TCP session established src 20.0.0.1 <--- ESTAB -----> 20.0.0.4 dst 3)keeping the connection in ESTABLISHED state i removed the 20.0.0.1 ("eth1") interface using if_dettach() and ifRouteDelete() I find still the TCP session in established state src 20.0.0.1 <--- ESTAB -----> 20.0.0.4 dst but no data transfer occur when i analyse the packets on the net. 4)Once again created "eth1" But with different interface ip address "30.0.0.1" ->routeShow destination gateway flags Refcnt Use Interface ---------------------------------------------------------------------------- 30.0.0.1 30.0.0.1 101 0 0 eth1 I see the Now the TCP traffic moving out through the net BUT WITH 20.0.0.1 as the SOURCE IP ADDRESS -- This looks strange I doubted the interface not being flushed properly. Further diagonising found that if_dettach has BUG in flushing the entries and so added the following code void in_ifaddr_remove(ifp) struct ifnet *ifp; /* ifp = ifunit("sl0") before slipDelete() called */ { struct in_ifaddr *ia, *prev_ia; if (!ifp) return; prev_ia = 0; for (ia = in_ifaddr; ia; ia = ia->ia_next) { if (ia->ia_ifp == ifp) { if (prev_ia) prev_ia->ia_next = ia->ia_next; else in_ifaddr = ia->ia_next; return; } prev_ia = ia; } } But the problem as in step 4 still exist. Any gurus to further point me out where should i further proceed. THANKS, "VXWORKS GURUS" IN ADVANCE _Sajjad. From vxwexplo-errs@csg.lbl.gov Tue Feb 12 20:39:01 2002 From: "Sajjad Ahmed" Date: Tue Feb 12 20:39:04 PST 2002 Subject: Problem in Flushing the Routing entries.(STRANGE READ THIS) Hello VxWorks Gurus, I met a typical problem problem this morning, taking your less time let me explain to you the short scenario. 1) i create an ethernet interface "eth1" using ether_attach. ->routeShow destination gateway flags Refcnt Use Interface ---------------------------------------------------------------------------- 20.0.0.1 20.0.0.1 101 0 0 eth1 the physical connection as follows 20.0.0.1 |--------------| |-----------|20.0.0.4 | |HUB| | ROUTER | | | | HOST 2)telnet to a host on 20.0.0.0 n/w with ip address 20.0.0.4 i see telnet TCP session established src 20.0.0.1 <--- ESTAB -----> 20.0.0.4 dst 3)keeping the connection in ESTABLISHED state i removed the 20.0.0.1 ("eth1") interface using if_dettach() and ifRouteDelete() I find still the TCP session in established state src 20.0.0.1 <--- ESTAB -----> 20.0.0.4 dst but no data transfer occur when i analyse the packets on the net. 4)Once again created "eth1" But with different interface ip address "30.0.0.1" ->routeShow destination gateway flags Refcnt Use Interface ---------------------------------------------------------------------------- 30.0.0.1 30.0.0.1 101 0 0 eth1 I see the Now the TCP traffic moving out through the net BUT WITH 20.0.0.1 as the SOURCE IP ADDRESS -- This looks strange I doubted the interface not being flushed properly. Further diagonising found that if_dettach has BUG in flushing the entries and so added the following code void in_ifaddr_remove(ifp) struct ifnet *ifp; /* ifp = ifunit("sl0") before slipDelete() called */ { struct in_ifaddr *ia, *prev_ia; if (!ifp) return; prev_ia = 0; for (ia = in_ifaddr; ia; ia = ia->ia_next) { if (ia->ia_ifp == ifp) { if (prev_ia) prev_ia->ia_next = ia->ia_next; else in_ifaddr = ia->ia_next; return; } prev_ia = ia; } } But the problem as in step 4 still exist. Any gurus to further point me out where should i further proceed. THANKS, "VXWORKS GURUS" IN ADVANCE _Sajjad. From vxwexplo-errs@csg.lbl.gov Wed Feb 13 08:48:56 2002 From: karunakaran nair Date: Wed Feb 13 08:48:58 PST 2002 Subject: Re: comp.os.vxworks newsdigest here is a question my friend. suppose two devices are talking to each other. let us call them X and Y. X has multiple threads and multiple channels to talk to Y. When X sends a message along a channel, it expects Y to come back with an acknowledgement. Given the scenario of multiple channels what is the safest way to do it with vxworks ? my thoughts are going to use of counting semaphores. so, assuming 10 channels, we would have a count of 10 on a counting semaphore. you are the vxworks guy. have anything to say on this. also, i want to know about backtracing on the shell. when a program crashes, how to back trace ont the shell ? another question, how does crosswind get to know which file to attach when debugging ? if assembly code comes up, what are the checks to make ? what are static loads ? what is the difference between .st and .st.bin files ? how to debug when static loads are running ? thanks, KN __________________________________________________ Do You Yahoo!? Send FREE Valentine eCards with Yahoo! Greetings! http://greetings.yahoo.com From vxwexplo-errs@csg.lbl.gov Thu Feb 14 03:29:08 2002 From: Anand Vardhan Date: Thu Feb 14 03:29:11 PST 2002 Subject: downloading Hi, I am a new user of Tornado. I am using an evaluation copy of it. I have used hostLib.h and netLib.h libraries in a program. I have compiled it and when I started downladiong it onto the target, it pops up an error message and gives the list of methods I have used in these particular libraries. I believe its an error of linker. I would like to know anykind of explanation in this regards thanks in advance Anand __________________________________________________ Do You Yahoo!? Send FREE Valentine eCards with Yahoo! Greetings! http://greetings.yahoo.com From vxwexplo-errs@csg.lbl.gov Thu Feb 14 04:03:25 2002 From: Vxworks Exploder Date: Thu Feb 14 04:03:27 PST 2002 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Thu Feb 14 04:03:20 PST 2002 Subject: Re: How do I get started learning VxWorks? Subject: Re: Questions about vxmalloc.c Subject: dosFs size limit 32MB? Subject: Re: how i build bootrom Subject: Re: motFccEnd.c Subject: Re: Deleting pipes Subject: vxWorks BSP developments with Galileo GT64260 Subject: Re: line feeds, carriage returns, oh my -- BSP problem??? Subject: Re: does my MIB leaf exists ? Subject: How to add a new mib to vxworks snmp agent? Subject: Re: tNetTask takes too long? Subject: Re: dosFs size limit 32MB? Subject: Re: dosFs size limit 32MB? Subject: DNS question. Subject: Re: stack USB from WinRiver on All BSP? Subject: Using lex in vxWorks . Subject: Re: How to make Ctrl-X handler wait for a routine to finish Subject: Reading from serial ports - Blocking or interrupts? Subject: IDMA to SCC performance via vxWorks Subject: Re: Deleting pipes Subject: Re: ULIP driver ... Subject: line feeds, carriage returns, oh my -- BSP problem??? Subject: How do I get started learning VxWorks? Subject: Re: dosFs size limit 32MB? Subject: ProfileScope data logging Subject: multiple instances of a variable Subject: ProfileScope data logging Subject: ProfileScope data logging Subject: Re: tNetTask takes too long? Subject: Subject: Re: USB on the PCPentium BSP? Subject: Re: disassember in vxWorks 5.4? Subject: Re: 64-bit operations Subject: Re: does my MIB leaf exists ? Subject: Re: Checking size of string with arbitrary characters Subject: Re: Checking size of string with arbitrary characters Subject: Re: [AE] Problem building a bootApp with RAM_HIGH_ADRS higher than 0x03F00000 Subject: Re: XON-XOFF on the Console port Subject: Re: Beginner Client/Server problems! Subject: Re: Help on SDRAM performance consideration Subject: Re: multiple instances of a variable Subject: Equivalent of setitimer Subject: FS: VME and CPCI, Current SBCs and Chassis to 700MHz Subject: Re: how i build bootrom Subject: Re: Equivalent of setitimer Subject: how should i start Subject: Re: anticipating machine check error... Subject: Re: adding new function to be called from prjConfig.c Subject: Project-relative files? Subject: Handling select() function failures Subject: Re: ULIP driver ... Subject: Re: Equivalent of setitimer Subject: Unix : getting symbols for PowerPc Subject: Re: Help on SDRAM performance consideration Subject: Re: +++ C-Designer (sehr flexibel) für Realtime/C/Systeme sucht neues Projekt +++ Subject: Re: I2C and SPI underrun and overrun error Subject: Re: Tornado II Prototyper installation failure Subject: Tool for checking memory leak Subject: Re: Reading from serial ports - Blocking or interrupts? Subject: Re: Unix : getting symbols for PowerPc Subject: Re: ULIP driver ... Subject: Re: MPC860 interrupt latency variation Subject: Memory Management: Questions about Partitions ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: How do I get started learning VxWorks? Date: Wed, 13 Feb 2002 11:49:28 -0600 From: "Tim" Organization: Nicolet Technologies Message-ID: References: Reply-To: "Tim" Sender: moeller@niti.com ARE YOU SURE you really want to do this? :-) Seriously, why? Having come from the Windows software development world (many years, SDK, MSVC, MFC, etc.) to VxWorks myself (out of necessity, because the previous embedded guy left the company before the project was finished), I would offer some caveats: First of all, you will find WindRiver's tools (and the Gnu compiler, which is the compiler engine around which the WindRiver tools are wrapped) somewhat frustrating (e.g., forget about source browsing by reference, like you used in MSVC with the F11 key.) Also, I don't think there's much money or opportunity in VxWorks programming, at least not more than Windows programming; so why don't you sharpen your Windows skills instead? Unless you're a glutton for punishment. :-) These views are my own. -- Tim "Unemployed Windows Developer" wrote in message news:a40o58$9oq$1@slb0.atl.mindspring.net... > I am an unemployed Windows (VC++) developer and would like to learn the VxWorks > OS. Does anyone have any tips on what I should do - books, development kits, > etc.? Thanx > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Questions about vxmalloc.c Date: Wed, 13 Feb 2002 12:43:31 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: "Bill Pringlemeir" wrote in message news:uvgd4j5ha.fsf@yahoo.com... > -- > quiche counter-intelligence domestic disruption $400 million in gold > bullion arrangements Rule Psix Marxist South Africa bomb class > struggle cracking [Hello to all my fans in domestic surveillance] > ammunition plutonium Soviet comrade pringlemeir, in accordance with secret world conspiracy instruction #23783, you should have changed the first word in that list to 'pretzel' by now... DaveK - -- moderator of alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! Proud Member of the Exclusive "I have been plonked by Davee because he thinks I'm interesting" List Member # Master of Many Meowing Minions Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage above and beyond the call of hilarity. --------------------------- Newsgroups: comp.os.vxworks Subject: dosFs size limit 32MB? Date: 7 Feb 2002 19:48:48 -0800 From: roswel_ajf@hotmail.com (roswell) Organization: http://groups.google.com/ Message-ID: <6dcae13d.0202071948.73976ad8@posting.google.com> Hi vxworkers, I want to dosFs to use my entire flash (128MB), but it is using only 32MB, according to what i see on dosFsConfigShow("/tffs0/"). Is there an option to increase this? many thanks.. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: how i build bootrom Date: 7 Feb 2002 21:45:03 -0800 From: alimohd@logiceastern.com (Ali) Organization: http://groups.google.com/ Message-ID: References: hi michael thanx for ur reply but it did not worked when i do this the image does not get build any other alternative pls. ali "Michael Lawnick" wrote in message news:... > Hi Ali, > > remove 'drv/' from 'drv/intrCtl/m8260IntrCtl.c' and see what happens... > HTH > -- > Mit freundlichen Grüßen, > Michael Lawnick > ============================================== > SOFTEC GmbH Tel +49-731-96600-0 > Promenade 17 Fax +49-731-96600-23 > D-89073 Ulm Michael Lawnick > Germany lawnick@softec.de > ============================================== > > "Ali" schrieb im Newsbeitrag > news:b305b5bd.0202070401.7417058b@posting.google.com... > > Hi all > > I have made a bootable image of my bsp and after a good work. > > Now even if my kernel gets build fine i am not able to make the > > bootrom of my bsp. > > the error it is showing is like this > > ccppc -B/home/alimohd/tornado/host/sun4-solaris2/lib/gcc-lib/ > > -mstrict-align -ansi -nostdinc -O2 -fvolatile -fno-builtin > > -fno-for-scope -I/h -I. -I/home/alimohd/tornado/target/config/all > > -I/home/alimohd/tornado/target/h > > -I/home/alimohd/tornado/target/src/config > > -I/home/alimohd/tornado/target/src/drv -DCPU=PPCEC603 -g -O0 -P -x > > assembler-with-cpp -c -o sysALib.o sysALib.s > > make: *** No rule to make target `drv/intrCtl/m8260IntrCtl.c', needed > > by `sysLib.o'. Stop. > > hit ENTER to exit > > since these are the files i have included in my bsp file and at the > > time of building the kernel it does not show such errors > > can any one tel me what should i do to make the bootrom . > > i am using tornado IDE for building the bootrom > > thanx > > ali --------------------------- Newsgroups: comp.os.vxworks Subject: Re: motFccEnd.c Date: 7 Feb 2002 22:02:44 -0800 From: alimohd@logiceastern.com (Ali) Organization: http://groups.google.com/ Message-ID: References: <3c630037$1@news.si.com> hi Dan Thanx for ur reply and i think it worked here since im in development phase and infact the board we are going to make ourselves has two fcc full duplex ports and the PHY device is LXT973(INTEL) now tell me pls if u can how i bring up the interfaces and how to i assign the private ip addresses to each port (ipAttach()) and where should i add this much of code in BSP. to get the ethernet mac address of each port and to enable the interrupts do i have to call the functions sysFccEnetAddrGet() sysFccEnetEnable() sysFccMiiBitWr() sysFccMiiBitRd() two times and do my work after.Also does these two fccs have harcoded offsets in DPRAM and if it is so how can i know those addresses.Does it seems to be right approach . waiting for the reply yhanx ali vanbaren@falcon.si.com (vanbaren_gerald) wrote in message news:<3c630037$1@news.si.com>... > "Dan Gold" writes: > > >"Ali" wrote in message > > >> Hi Dan > >> Thanx for ur reply yes i have the source files for sysmotfccend.c and > >> configNet.h but the situation there is only for one instance of end > >> driver can u pls tell me what extra changes i have to make to these > >> two files so that i can support two fcc's for ethernet and that two > >> both of them in full duplex mode. > >> waiting for the reply > >> thanx > > >I'm not using sysMotFccEnd.c, but my configNet.h file for multiple END > >devices (1 SCC and 2 FCC) looks like: > > >END_TBL_ENTRY endDevTbl [] = > >{ > > { 0, PQ2_SCC_LOAD_FUNC, PQ2_SCC1_LOAD_STRING,PQ2_SCC_BUFF_LOAN,NULL, > >FALSE}, > > { 3, PQ2_FCC_LOAD_FUNC, PQ2_FCC3_LOAD_STRING,PQ2_FCC_BUFF_LOAN,NULL, > >FALSE}, > > { 1, PQ2_FCC_LOAD_FUNC, PQ2_FCC1_LOAD_STRING,PQ2_FCC_BUFF_LOAN,NULL, > >FALSE}, > > { 0, END_TBL_END, NULL, 0, NULL, FALSE}, > >}; > > >That should cause SENS to run the xxxEndLoad/xxxEndStart functions for each > >entry. To actually bring up the interface (since vxWorks only brings up the > >boot device in usrNetInit()), see the vxWorks FAQ. They have a little code > >snippet (that I wrote) to bring up more interfaces. As far as setting it > >for full duplex, I don't know where you specify that - I just autodetect, > >and then read my PHY to see if it's full or half duplex. > > > hth, > > Dan Gold gold@ensemble.com > > I have not done it myself, but I recall seeing warnings that you need > to make sure the two FCCs are using different areas of DPRAM for their > buffers (RIPTR, TIPTR). My recollection is that the WR driver can > support any FCC port, but it has a hardcoded offset that it uses for > RIPTR and TIPTR. > > gvb --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Deleting pipes Date: 7 Feb 2002 22:04:57 -0800 From: rashanmu@npd.hcltech.com (Ramesh) Organization: http://groups.google.com/ Message-ID: <15a87607.0202072204.3e00f7d7@posting.google.com> References: <15a87607.0202060231.4bbe9114@posting.google.com> <3C6112A9.E22D1228@l8s.co.uk> <88c8dc81.0202062326.5735c154@posting.google.com> <3C6259C8.FFD1005A@l8s.co.uk> David, That function deletes the memory allocated for the pipe also. If iosDevDelete alone is called means, it will delete the pipe alone. But the next call of free(pipeDevHdr) actually frees all the allocated memory. Just check out with memShow after creating a pipe and after calling the call. You will recover all the memory allocated without leaks. Also when a pipe is created and opened it is returned with a 'fd'. Even when pipeDevDelete is called the 'fd' assinged will not get closed unless you close the 'fd' with close call. Thanks, Ramesh David Laight wrote in message news:<3C6259C8.FFD1005A@l8s.co.uk>... > tutankhamun wrote: > > > > Write your own function to delete a pipe instead. Try the following > > that I've used before. > > That function doesn't delete the pipe, it only removes the name from the > device table. All the rest of the memory is still allocated. What is > more, is the pipe was open you might hit the sequence below - which > corrupts the memory heap. > > iosDevAdd( dev, name, drvnum ); > fd = open( name, ... ) > iosDevDelete( dev ); > iosDevAdd( xxx, name, yyy ); > close( fd ); > > David --------------------------- Newsgroups: comp.os.vxworks Subject: vxWorks BSP developments with Galileo GT64260 Date: 13 Feb 2002 18:23:25 -0800 From: jchen@tollbridgetech.com (john c) Organization: http://groups.google.com/ Message-ID: <1f1b48b5.0202131823.2c3e9702@posting.google.com> Anyone working with the Galileo GT64260 Dual PCI bridge Controller? Just curious. We're using IBM 750 CXe (600MHz PowerPC) interfacing to Galileo GT 64260 dual PCI bridge controller. Our vendor did a really crappy job porting vxWorks BSP for our needs, and we ended up developing most of the drivers and debugging ourselves. I'm just wondering if anyone else out there is trying to get their GT64260 to work properly that I can share ideas with. We measured SDRAM performance yesterday with the 750 and the Galileo. Basically, the 60x bus of the 750 doesn't seem to handle SDRAM bursts very well. the result is very limited bandwidth for the 750 to read and write to the SDRAM that is connected through the Galileo chip. YIKES. If you are building your system, try to use a 74xx series controller that takes advantage of the MPX bus.... :-) anyhow. just wondering if I'm alone in fighting with the GT64260. smiles john --------------------------- Newsgroups: comp.os.vxworks Subject: Re: line feeds, carriage returns, oh my -- BSP problem??? Date: Wed, 13 Feb 2002 16:08:43 GMT From: "Jim" Organization: Magma Communications Ltd. Message-ID: References: is your terminal set to vt100 ;-) i am windows guy but if i remember right ... setenv TERM vt100 ... but i am also full of crap in my spare time so i may be mistaken ;-) Jim "Philip" wrote in message news:dd1860e3.0202130728.15995091@posting.google.com... > Hi, I'm in the process of writing a BSP. We have almost everything > done, except we have this one nagging serial port problem. > > During the execution of the bootrom AND while we are using > "bootChange" in the vxWorks kernel we do not get a line feed, but we > do get a carriage return. > So when I hit enter, I do not see a new line, but have to type over > the previous characters of the old line. > > We are using Solaris to tip into our serial port. Both windsh, and the > regular shell seem to work OK. This is something specific about our > BSP, since tipping into an mcp750 next slot over works fine. > > I know that NT and Unix are different wrt what they need at the ends > of lines, but what the hell is this all about? Is it some funky ioctl > setting that we have to do in bootConfig.c or usrConfig.c when we > create the serial port? > > We are using a brain dead "FPGA" serial port, so the serial driver I > wrote does nothing with the ioctls when they come in. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: does my MIB leaf exists ? Date: 13 Feb 2002 08:38:02 -0800 From: ilyasg@yahoo.com (Ilyas G) Organization: http://groups.google.com/ Message-ID: References: Yes, the problem is caused by the community . when i have debugged the snmpd i noticed that the community wasn't set for the agent. Do you know haw i can set it? --------------------------- Newsgroups: comp.os.vxworks Subject: How to add a new mib to vxworks snmp agent? Date: 7 Feb 2002 23:55:37 -0800 From: yong_bvcom@sina.com (bigtornado) Organization: http://groups.google.com/ Message-ID: <5624c1bf.0202072355.76388f67@posting.google.com> Any answers and suggestions are appreciated! Jeffy --------------------------- Newsgroups: comp.os.vxworks Subject: Re: tNetTask takes too long? Date: Fri, 8 Feb 2002 09:52:14 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: tNetTask is sort of "catch-all" task for network related processing, so among other things, it will take care of all TCP timer processing, so if you have many TCP connections, tNetTask will process each one of them periodically to check if any of their timers has expired. ANother typical use for the task is for "watch dog" processing on behalf of the network driver, as the ISR needs to be minimal, some drivers have to process interrupts other than packet reception or transmission at tNetTask. This cliuld be "anti lockup" checking, PHY link renegotiation control etc. The 5 seconds period seems odd though, as does 4.5ms processing time, except maybe the driver's PHY code, which can be rather CPU intensive due to MDIO - an I2C-like interface. You can use profiler ur simple assembly debugging to get to the bottom of this. "Dinker Charak" wrote in message news:a3v19m$17o$1@info4.fnal.gov... > Hello, > > Tornado 2.0.2/vxWorks 5.4.2 > > We have a task running at priority 130 that just eats CPU cyclles (no net > activity). We observed (on MV2301 and MV2400 - both 5.4.2) that every ~5 s > tNetTask switches in, stays for ~4.5ms and then switches out. There is no > network activity, sniffer shows nothing on wire. Exactly same setup with > 5.4 shows no such abnormaly long activity. > > Why is tNetTask taking so long? What is it doing? Anybody noticed anything > like this? Any idea or remarks? > > Thanks, > Dinker > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: dosFs size limit 32MB? Date: Fri, 8 Feb 2002 09:25:41 +0100 From: "Michael Lawnick" Organization: Buergernetz Dillingen Message-ID: References: <6dcae13d.0202071948.73976ad8@posting.google.com> Reply-To: "Michael Lawnick" Sender: mlawnick@pd9e65702.dip.t-dialin.net Hi roswell, AFAIK, your limit should not be 32MB, but 40MB. You'll find in target/h/tffs/flCustom.h: #define MAX_VOLUME_MBYTES 40 I had to increase SECTOR_SIZE_BITS (512 byte->1024 byte) to get over 64MB. Change this defines and recompile TFFS-sources if available or ask your FAE. - -- Mit freundlichen Grüßen, Michael Lawnick ============================================== SOFTEC GmbH Tel +49-731-96600-0 Promenade 17 Fax +49-731-96600-23 D-89073 Ulm Michael Lawnick Germany lawnick@softec.de ============================================== "roswell" schrieb im Newsbeitrag news:6dcae13d.0202071948.73976ad8@posting.google.com... > Hi vxworkers, > > I want to dosFs to use my entire flash (128MB), but it is using only > 32MB, according to what i see on dosFsConfigShow("/tffs0/"). Is there > an option to increase this? > > many thanks.. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: dosFs size limit 32MB? Date: Fri, 8 Feb 2002 09:33:52 +0100 From: "Werner Schiendl" Message-ID: <3c638d69$1@brateggebdc5.br-automation.co.at> References: <6dcae13d.0202071948.73976ad8@posting.google.com> Hi, dosFs (and dosFs2 for that matter) have no such limit. How did you format / create the drive? With VxWorks or some external tool (DOS,... which...). If VxWorks, how exactly did you do setup the disk? The only limitation I can remember for 32MB is for FAT-12 format (very old style). AFAIK dosFs chooses the proper format (FAT-12 or FAT-16) automatically. If not, be sure to specify FAT-16. hth Werner "roswell" wrote in message news:6dcae13d.0202071948.73976ad8@posting.google.com... > Hi vxworkers, > > I want to dosFs to use my entire flash (128MB), but it is using only > 32MB, according to what i see on dosFsConfigShow("/tffs0/"). Is there > an option to increase this? > > many thanks.. --------------------------- Newsgroups: comp.os.vxworks Subject: DNS question. Date: Fri, 8 Feb 2002 17:03:39 +0800 From: "Howchin" Organization: SEEDNet News Service Message-ID: This is my function. struct hostent * spstdtime; char szBuff[256]; if ( resolvInit("168.95.1.1","glorytel.com.tw",NULL)==OK) { spstdtime=resolvGetHostByName("clock.stdtime.gov.tw",szBuff,256); if(spstdtime!=NULL) { printf("h_name=%s\n",(spstdtime->h_name)); printf("h_aliases=%s\n",*(spstdtime->h_aliases)); printf("h_addrtype=%d\n",spstdtime->h_addrtype); printf("h_length=%d\n",spstdtime->h_length); printf("h_addr_list=%s\n",*(spstdtime->h_addr_list)); } #if 0 struct hostent { char * h_name; /* official name of host */ char ** h_aliases; /* alias list */ int h_addrtype; /* address type */ int h_length; /* length of address */ char ** h_addr_list; /* list of addresses from name server */ unsigned int h_ttl; /* Time to Live in Seconds for this entry */ } #endif } This is the result. h_name=clock.stdtime.gov.tw h_aliases=(null) h_addrtype=2 h_length=4 h_addr_list= Who can tell me what happen ? I use netxray to capture the packets. 168.95.1.1(dns.hinet.net) have sent the resolved message to me. and resolvGetHostByName return value isn't NULL. Why I cant get the IP?? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: stack USB from WinRiver on All BSP? Date: 12 Feb 2002 11:59:10 GMT From: Charly Message-ID: <2002212-125910-501455@foorum.com> References: <2002211-18554-313711@foorum.com> Thanks Johan, My target is a standard PC. I could use the USB included on the mother board or an additive PCI USB card. Charly - -- Use our news server 'news.foorum.com' from anywhere. More details at: http://nnrpinfo.go.foorum.com/ --------------------------- Newsgroups: comp.os.vxworks Subject: Using lex in vxWorks . Date: 8 Feb 2002 01:46:47 -0800 From: cwnam@korealink.ne.kr (changwoo) Organization: http://groups.google.com/ Message-ID: <3d64ee1a.0202080146.3f9b5d0b@posting.google.com> Hi? all I tried to use lex compiler in vxWorks but I couldn't In Unix and linux environment, I compiled and excuted it. Nothing was wrong to use it in Unix. With the same source in vxWorks, I couldn't even compile. The procedure is like this. first, I compiled it with lex(unix\> lex sampel.l) and the result of compile was "lex.yy.c". After then I included the file to workspace in vxWorks and compiled, but there was a error "inttypes.h : no such file" showed. What I have to do to use "lex" in vxworks? Do I have get lex compiler for vxWorks instead of that for Unix? Please take a few time and help this ... --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to make Ctrl-X handler wait for a routine to finish Date: Fri, 8 Feb 2002 10:15:34 -0000 From: "timallen" Organization: ntlworld News Service Message-ID: References: <3c63076c$0$29348$4c41069e@reader1.ash.ops.us.uu.net> Reply-To: "timallen" Have you looked at rebootHookAdd in rebootLib? I'm not sure at what point in the reboot sequence the hooks are called though. Tim. "Mark Galecki" wrote in message news:3c63076c$0$29348$4c41069e@reader1.ash.ops.us.uu.net... > Hello, > > I want to make the Ctrl-X handler wait for a routine to finish. That is, > after that routine starts execution and before it finishes, I do not want > the soft system reboot to happen. Can I do that and how? > > Thank you, > Mark Galecki > SBS Technologies > > --------------------------- Newsgroups: comp.os.vxworks Subject: Reading from serial ports - Blocking or interrupts? Date: 13 Feb 2002 06:24:22 -0800 From: vtail@hotmail.com Organization: http://groups.google.com/ Message-ID: <37449c9d.0202130624.43ffc949@posting.google.com> I am working on an application using VxWorks 5.4 to read from a number of serial ports simultaneously. Each serial port is read by its own separate VxWorks task. The serial ports receive pre-defined, but variable-length (over some fairly close limits) messages. Each serial port comes with a 128-byte buffer, and the receive interrupt can be set to fire on any number of characters in the buffer from 1 to 128. It can also generate an interrupt if 4 character times pass with no input, with the FIFO only partially filled (i.e. not filled enough to reach its interrupt level). Timing for processing incoming messages is critical, i.e. each task must process its received message as soon as practical. My question is, which of the following scenarios is likely to be the most efficient in terms of use of processor resources, while still meeting the objective of processing incoming messages as quickly as possible: 1. Set the interrupt to interrupt on the length of the shortest received message. The interrupt sends a signal to the input task, which then reads and stores the data received so far, then goes back to sleep. The tail end of the message is handled by the 4-character timeout-with-no-data interrupt. 2. Generate an interrupt on each character, which sends a signal to the input task which processes incoming characters on a character-by-character basis. 3. Use a blocking read function to read from the serial port. No interrupts are needed, and incoming data is handled on a character-by-character basis. It seems to me that #3 is likely to be the most efficient, assuming the serial port board's BSP supports it (we don't have the BSP documentation yet so I can't be sure but I'd be very surprised if it didn't). But, that raises another question ... In general, what is the overhead for unblocking read functions, compared with the overhead of processing an interrupt? I'm concerned that there might be some goofy internal hidden stuff that goes on in scenario #3 such that #2 is actually faster and more efficnent. Also, when a task is blocked, i.e. on a read of a serial port, does it ALWAYS relinquish the CPU for use by lower-priority tasks, and then take over the CPU again when data appears on the serial port? Thanks... Eric --------------------------- Newsgroups: comp.os.vxworks Subject: IDMA to SCC performance via vxWorks Date: Fri, 08 Feb 2002 12:25:04 GMT From: "Jose" Organization: Magma Communications Ltd. Message-ID: Hello any one seen better peformance on an 860 than this: IP->Protocol->MPC860_IDMA->Recieve DMAed Buffers->Strip Unecessary Header->Send to IP stack->out ethernet When dmaing 128 bytes buffers containing header + 64 bytes ip packet + padding we see transfer rates of only 1.5 Mbits or so when sending 2.0 Mbits of possible thruput. Data in this situation is being discarded by the device which is being dma'd from. If the hand off to the IP stack is removed and packets are discarded before the "Send to IP stack" then no data is lost by device being dma'd from. DMA takes advantage of "free" bus cycles ... so it is understandable some performance hit can occur ... but this does not explain clearly why the IP stack causes the bus to be so active ... I was very careful to dma directly into mblk,cblk,cluster pools so that there would be ZERO copy between DMA and ethernet SCC. Does anyone know if the vxWorks IP stack is doing any unecessary "copy" or what would be so expensive in the IP stack ? Even if you think this is impossible please send me your comments ... Jose ps for larger packets we get very high rates however because the vxWorks IP stack is used less often. pps It says in motorola documentation that DMA transfer rate is 2.0 Mbytes per second ... which is cut in half when we enable other CPM functionality ...I see around 0.8 Mbytes for large packets ... but much poorer results for small ones ... --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Deleting pipes Date: 8 Feb 2002 05:02:03 -0800 From: tut_pharoh@yahoo.com (tutankhamun) Organization: http://groups.google.com/ Message-ID: <88c8dc81.0202080502.23776917@posting.google.com> References: <15a87607.0202060231.4bbe9114@posting.google.com> <3C6112A9.E22D1228@l8s.co.uk> <88c8dc81.0202062326.5735c154@posting.google.com> <3C6259C8.FFD1005A@l8s.co.uk> <15a87607.0202072204.3e00f7d7@posting.google.com> rashanmu@npd.hcltech.com (Ramesh) wrote in message news:<15a87607.0202072204.3e00f7d7@posting.google.com>... > David, > > That function deletes the memory allocated for the pipe also. If > iosDevDelete alone is called means, it will delete the pipe alone. But > the next call of free(pipeDevHdr) actually frees all the allocated > memory. Just check out with memShow after creating a pipe and after > calling the call. You will recover all the memory allocated without > leaks. Also when a pipe is created and opened it is returned with a > 'fd'. Even when pipeDevDelete is called the 'fd' assinged will not get > closed unless you close the 'fd' with close call. > > Thanks, > Ramesh > IMHO if you strictly adhere to the following sequence of events, there's no room for any memory corruption. pipeDevDelete() frees up all the associated memory. /* create a pipe device */ pipeDevCreate ("xxx", 10, 100); /* open the device */ fd = open ("xxx", O_RDWR); : : : /* you have a valid file descriptor here - do what you want with it */ : : /* close the device */ close (fd); /* delete the pipe */ pipeDevDelete ("xxx"); regards, tutankhamun --------------------------- Newsgroups: comp.os.vxworks Subject: Re: ULIP driver ... Date: Thu, 14 Feb 2002 03:09:22 GMT From: "Kim Letkeman" Message-ID: References: Dave, have you spent any time with XP? I've been running it for over a year as a beta tester (did the same for Win2k, so I have extensive experience with both.) There is no comparison. Win2k is old news in almost every way. Just my opinion of course. By the way, if anyone is interested, I found the magic mechanism for adding drivers to XP (they make it harder, for good reason) and found that both drivers will install and both drivers will blue screen on boot. Windows XP's recovery mechanism is so strong that you merely have to fall back to a dated check point and the bad stuff just disappears. A few minutes and you are running perfectly again. Very cool, but very unfortunate that there is no way to easily run the current generation of ulip drivers. The failure, for those who may know a trick to get this stuff to work, is actually in the NDIS driver, so it is likely some change introduced in XP that's frying their bacon, not anything specifically wrong with the ulip driver. But busted is busted ..... I'd bet someone at WindRiver could get this sucker to work in a day or less .... "Dave Korn" wrote in message news:T7ta8.2382$5o.1004928@newsr2.u-net.net... > Why dual boot when you could just revert back to 2k for good ? I > guess it's not what you want to hear, but frankly, unless it's *really* > important to you that M$ fill your machine full of backdoors to let in > advertisers and the mpaa/riaa/bsa, w2k is so much more safe stable reliable > useable and secure than XP that I'd consider reverting to be an upgrade. > IMO. > > DaveK > -- > moderator of > alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow > Burn your ID card! http://www.optional-identity.org.uk/ > Help support the campaign, copy this into your .sig! > Proud Member of the Exclusive "I have been plonked by Davee because he > thinks I'm interesting" List Member # > Master of Many Meowing Minions > Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage above > and beyond the call of hilarity. > > --------------------------- Newsgroups: comp.os.vxworks Subject: line feeds, carriage returns, oh my -- BSP problem??? Date: 13 Feb 2002 07:28:07 -0800 From: p-m@yahoo.com (Philip) Organization: http://groups.google.com/ Message-ID: Hi, I'm in the process of writing a BSP. We have almost everything done, except we have this one nagging serial port problem. During the execution of the bootrom AND while we are using "bootChange" in the vxWorks kernel we do not get a line feed, but we do get a carriage return. So when I hit enter, I do not see a new line, but have to type over the previous characters of the old line. We are using Solaris to tip into our serial port. Both windsh, and the regular shell seem to work OK. This is something specific about our BSP, since tipping into an mcp750 next slot over works fine. I know that NT and Unix are different wrt what they need at the ends of lines, but what the hell is this all about? Is it some funky ioctl setting that we have to do in bootConfig.c or usrConfig.c when we create the serial port? We are using a brain dead "FPGA" serial port, so the serial driver I wrote does nothing with the ioctls when they come in. --------------------------- Newsgroups: comp.os.vxworks Subject: How do I get started learning VxWorks? Date: 8 Feb 2002 14:42:48 GMT From: I.need.work@com.net (Unemployed Windows Developer) Organization: MindSpring Enterprises Message-ID: I am an unemployed Windows (VC++) developer and would like to learn the VxWorks OS. Does anyone have any tips on what I should do - books, development kits, etc.? Thanx --------------------------- Newsgroups: comp.os.vxworks Subject: Re: dosFs size limit 32MB? Date: Fri, 8 Feb 2002 16:41:31 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <6dcae13d.0202071948.73976ad8@posting.google.com> DosFs will use whatever it gets from its subordinate block driver, as per the block device structure field specifying the sectors per volume. Your block driver is apparently TFFS, which lets you allocate only some partition of your overall Flash to DosFs, and use the rest to store the BootROM. I feel that if you go back to your TFFS manual you can pretty much figure it out yourself. - - Leonid "roswell" wrote in message news:6dcae13d.0202071948.73976ad8@posting.google.com... > Hi vxworkers, > > I want to dosFs to use my entire flash (128MB), but it is using only > 32MB, according to what i see on dosFsConfigShow("/tffs0/"). Is there > an option to increase this? > > many thanks.. --------------------------- Newsgroups: comp.os.vxworks Subject: ProfileScope data logging Date: Thu, 14 Feb 2002 03:26:48 GMT From: "Jamilur Rahman" Message-ID: Hi Community, I am currently working with ProfileScope (an RTI tool from W.R.S.). Can I save those profile data in a file (as a log file) of Excell or other data based format for future reference ? I also need some recommendation about a freeware graph drawing software that I could use for plotting two dimensional graph plotting. /jamil, Ottawa, Ontario. --------------------------- Newsgroups: comp.os.vxworks Subject: multiple instances of a variable Date: Fri, 08 Feb 2002 10:11:16 -0500 From: willisn Organization: Raytheon Company Message-ID: <3C63EA94.E47EFE94@res.raytheon.com> Hi all, My question concerns multiple instances of a single global variable. I have 3 tasks running, all have an instance of a global variable (say an integer). If I want to know the current value of this integer for any one of the tasks, I need to do a lkup on the variable and then dereference the address associated with the task I am interested in. My question is, if you simply type the name of the variable in the console, how does vxworks decide which instance of the variable to give you? Thanks in advance for any help! Nick --------------------------- Newsgroups: comp.os.vxworks Subject: ProfileScope data logging Date: Thu, 14 Feb 2002 03:34:11 GMT From: "Jamilur Rahman" Message-ID: I am currently working with ProfileScope (an RTI tool from WindRiver). Is it possible to save the profile data in Excell or other data base format for future reference ? I also need recommnedation about a graph drawing (freeware) software for plotting 2-dimensional graph. Any help would be very much appreciated. /jamil, Ottawa, Ontario. --------------------------- Newsgroups: comp.os.vxworks Subject: ProfileScope data logging Date: Thu, 14 Feb 2002 03:43:49 GMT From: "Jamilur Rahman" Message-ID: How can I save the ProfilrScope data in Excel or other data base format for future reference ? I also need recommendation of graph drawing software (preferably freeware) for plotting 2-dimensional plot. Thanks in advance for any help. /jamil, Ottawa, Ontario. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: tNetTask takes too long? Date: Fri, 8 Feb 2002 09:33:47 -0600 From: Dinker Charak Organization: Fermi national Accelerator Lab Message-ID: References: Subject: Re: tNetTask takes too long? Date: Fri, 08 Feb 2002 10:36:16 +0100 From: Peter Salewsky To: "'dinker@fnal.gov'" Hi Dinker, I seem to be unable to post to the NG so I answer via mail. > We have a task running at priority 130 that just eats CPU cyclles (no net > activity). We observed (on MV2301 and MV2400 - both 5.4.2) that every ~5 s > tNetTask switches in, stays for ~4.5ms and then switches out. There is no > network activity, sniffer shows nothing on wire. Exactly same setup with > 5.4 shows no such abnormaly long activity. > > Why is tNetTask taking so long? What is it doing? Anybody noticed anything > like this? Any idea or remarks? We have a similar (even worse) problem on an MVME2401 (VxWorks 5.4). Our task runs at priority 2 and sends messages to a client viw a TCP socket every 5 ms. A WindView analysis shows the following: Every ~2 s tNetTask is activated via a software watchdog and runs for ~30 ms. When our task calls send() during this period it is blocked on a mutex semaphore that had been taken by tNetTask - so we are unable to send messages for 30 ms! Disabling the sysClk solves the problem (the watchdog timers are managed via sysClk) - - this might help in your case also if you do not need the system clock otherwise. We are currently further investigating the problem; it might be related to the Ethernet card driver (we do not use the on-board interface but an additional PMC card). Please let me know if you have any new ideas. Bye, Peter --------------------------- Newsgroups: comp.os.vxworks Subject: Re: USB on the PCPentium BSP? Date: Tue, 12 Feb 2002 13:12:38 -0000 From: "Richard Horton" Organization: Newport Networks Ltd. Message-ID: <1013519562.44534@azores.network-i.net> References: <2002212-131239-482717@foorum.com> "Charly" wrote in message news:2002212-131239-482717@foorum.com... > > Hello vxworks users, > > I wonder if it is possible to use the WindRiver USB Stack on PCPentium BSP. > Does anyone use it? > What is the development cost to : > install and configure the USB stack. > To write and tesy an easy request on it. > > Thanks for your experiences. > charly Charly, You only need ask your question once... it was only yesterday when you first asked about USB stack support... - -- All opinions, unless otherwise stated are my own and not those of my employer. Richard Horton Newport Networks Ltd. "Users are like a virus, each causing a thousand tiny crises until the host finally dies" --------------------------- Newsgroups: comp.os.vxworks Subject: Re: disassember in vxWorks 5.4? Date: Tue, 12 Feb 2002 19:23:35 -0600 From: Andrew Johnson Organization: APS, Argonne National Laboratory Message-ID: <3C69C017.8D790622@aps.anl.gov> References: <3C692108.41057703@ll.mit.edu> Michael Reed wrote: > > We're making a shopping list for an upcoming project... does anyone know > if vxWOrks 5.4 has a disassember bundled w/ it or the GNU tools? Two: the target shell provides the 'l' command (which may also work in the WindShell, I never use that so I'm not sure), and the GNU toolset's objdump program also disassembles object code for the relevent architecture. - - Andrew - -- Perfection is reached, not when there is no longer anything to add, but when there is no longer anything to take away. - - Antoine de Saint-Exupery --------------------------- Newsgroups: comp.os.vxworks Subject: Re: 64-bit operations Date: 12 Feb 2002 08:34:28 GMT From: Denis Perelyubskiy Message-ID: References: <3C67ADF8.55BA8A64@bellsouth.net> hello, * joe durusau [02-Mon-02 03:41 -0800]: > > This is just a guess, but have you tried splitting the variable into bytes > (a union or some such) an dprinting the bytes? It is entirely possible that > the bsp does not handle long long and therefore prints them as zero. > On most BSPs, you fill find that trying to print a float with logMsg > will do similar things, because logMsg just doesn't support floats. what does the BSP have to do with 64 bit operations on the printf()? i'd thought it was all about the compiler? thanks, > Nejhdeh Ghevondian wrote: > > Dear anyone, > > I am trying to implement 64-bit operations in my VxWorks operations (using > > the GNU compilier, on PowerPC platform). > > I am trying to multiply and divide unsigned long long int variables. > > However, I always get 0 as value. > > I acnt even do a printf("%llu\n", 64bit_var) > > > > Can anyone help me??? > > > > Regards Nejhdeh Ghevondian denis - -- if you want to reply by mail, please tweak my address in a reasonable way. sorry for the inconvenience. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: does my MIB leaf exists ? Date: 13 Feb 2002 22:22:59 -0800 From: tut_pharoh@yahoo.com (tutankhamun) Organization: http://groups.google.com/ Message-ID: <88c8dc81.0202132222.5cbb9287@posting.google.com> References: ilyasg@yahoo.com (Ilyas G) wrote in message news:... > Yes, the problem is caused by the community . when i have debugged the > snmpd i noticed that the community wasn't set for the agent. > Do you know haw i can set it? If you're using WindNet SNMP (or Epilogue) on VxWorks, then you may want to lookup the function snmpIoCommunityValidate() defined in your C:\Tornado\target\src\snmpv1\agent\snmpIoLib.c... This is where the SNMP community is validated according to the view. This function gets called every time the agent has to process the SNMP PDU. regards, tut --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Checking size of string with arbitrary characters Date: Tue, 12 Feb 2002 17:17:08 +0200 From: Alex Vinokur Message-ID: <3C6931F4.6230394A@bigfoot.com> References: <3C6930F9.A6B5EB16@bigfoot.com> Followup-To: comp.protocols.snmp,comp.os.psos,comp.os.vxworks ============================ pRISM+ for pSOSystem/PowerPC pRISM+ 1.2.3 . pSOSystem 2.2.2 Diab Data D-CC Optimizing C compiler (Version 4.1a) SNMP Release 7.4 (Protocol SNMPv1) Epilogue Technology Emissary SNMP MIB Compiler, version 6.3 Host : Windows98 Target : PowerPC (MPC860) ============================ My mib file contains the following item : - -- --------------------------------------- item1 OBJECT-TYPE SYNTAX OCTET STRING (SIZE ( 5 .. 5 )) MAX-ACCESS read-write STATUS current DESCRIPTION "" ::= { top 1 } - -- --------------------------------------- This string may hold any character, including 0. So, we can't use strlen() in the item1-test-async() function in order to check size of item1. How does 'item1-test-async' have to check size of item1 in SetRequest? Thanks in advance, =========================== Alex Vinokur mailto:alexvn@bigfoot.com mailto:alexvnw@yahoo.com http://up.to/alexvn http://go.to/alexv_math =========================== --------------------------- Newsgroups: comp.protocols.snmp,comp.os.psos,comp.os.vxworks Subject: Re: Checking size of string with arbitrary characters Date: Tue, 12 Feb 2002 17:23:57 +0200 From: Alex Vinokur Message-ID: <3C69338D.1F68A098@bigfoot.com> References: <3C6930F9.A6B5EB16@bigfoot.com> Alex Vinokur wrote: > ============================ > pRISM+ for pSOSystem/PowerPC > pRISM+ 1.2.3 . pSOSystem 2.2.2 > Diab Data D-CC Optimizing C compiler (Version 4.1a) > SNMP Release 7.4 (Protocol SNMPv1) > Epilogue Technology Emissary SNMP MIB Compiler, version 6.3 > Host : Windows98 > Target : PowerPC (MPC860) > ============================ > > My mib file contains the following item : > > -- --------------------------------------- > item1 OBJECT-TYPE > SYNTAX OCTET STRING (SIZE ( 5 .. 5 )) > MAX-ACCESS read-write > STATUS current > DESCRIPTION "" > ::= { top 1 } > -- --------------------------------------- > > This string may hold any character, including 0. > > So, > we can't use strlen() > in the item1-test-async() function > in order to check size of item1. > > How does 'item1-test-async' have to check size of item1 in SetRequest? > > Thanks in advance, > > Is this a solution ? // --------- from stub.c ------------ case LEAF_ird_control_packet: { ALENGTH_T length = EBufferUsed(VB_GET_STRING(group_vbp)); /* !!! These are the lengths from the MIB. Modify if your * implementation differs. */ if (length != SIZE_ird_control_packet) { testproc_error(pktp, group_vbp, WRONG_VALUE); break; } } testproc_good(pktp, group_vbp); break; // ---------------------------------- - -- =========================== Alex Vinokur mailto:alexvn@bigfoot.com mailto:alexvnw@yahoo.com http://up.to/alexvn http://go.to/alexv_math =========================== --------------------------- Newsgroups: comp.os.vxworks Subject: Re: [AE] Problem building a bootApp with RAM_HIGH_ADRS higher than 0x03F00000 Date: Thu, 14 Feb 2002 07:11:57 GMT From: "Gary M" Organization: AT&T Broadband Message-ID: <1rJa8.50967$fK1.4145568@rwcrnsc54> References: <14fc63a8.0202131233.2e9aec64@posting.google.com> My hunch is that the assembly language module excConnectALib.s, which sounds like it provides support for connecting to the exception vectors in low memory, uses instructions with 24-bit effective address operands. Relocating the code where you did makes this exceed the 24 bits available to generate the effective address. If you have the source code, you should be able to see and possibly correct the problem; otherwise, you are restricted to addresses below 0x00ffffff. Out of curiosity, it sounds like the hardware engineer decided to route the SDRAM chip selects to an extremely high part of the address space. If so, can this be changed? "David Masse" wrote in message news:14fc63a8.0202131233.2e9aec64@posting.google.com... > Hi, > > I'm trying to build a bootApp in Tornado AE of a custom BSP (A > modified version of ads860 BSP). My board needs to have ROM_BASE_ADRS > = 0x00000000 and LOCAL_MEM_LOCAL_ADRS = 0x50000000. I set > RAM_HIGH_ADRS = 0x50300000, RAM_LOW_ADRS = 0x50010000, ROM_TEXT_ADRS = > 0x00000100 and ROM_SIZE = 0x00280000 in both 00bsp.cdf and makefile. > When I try to build the bootApp, I always received the following > error: > > ldppc -X -e usrInit -Ttext 50300000 $\(PRJ_OBJS\) $\(bootKernel_OBJS\) > -o bootApp_reloc > C:\Tornado_AE/target/lib/objPPC860gnuvx/excConnectALib.o: In function > `excConnectCode': > C:\Tornado_AE/target/lib/objPPC860gnuvx/excConnectALib.o(.data+0x28): > relocation truncated to fit: R_PPC_ADDR24 excEnt > C:\Tornado_AE/target/lib/objPPC860gnuvx/excConnectALib.o(.data+0x34): > relocation truncated to fit: R_PPC_ADDR24 excExcHandle > C:\Tornado_AE/target/lib/objPPC860gnuvx/excConnectALib.o(.data+0x3c): > relocation truncated to fit: R_PPC_ADDR24 excExit > C:\Tornado_AE/target/lib/objPPC860gnuvx/excConnectALib.o(.data+0x48): > relocation truncated to fit: R_PPC_ADDR24 excEnt > C:\Tornado_AE/target/lib/objPPC860gnuvx/excConnectALib.o(.data+0x54): > relocation truncated to fit: R_PPC_ADDR24 excExcHandle > C:\Tornado_AE/target/lib/objPPC860gnuvx/excConnectALib.o(.data+0x5c): > relocation truncated to fit: R_PPC_ADDR24 excExit > C:\Tornado_AE/target/lib/objPPC860gnuvx/excConnectALib.o(.data+0x68): > relocation truncated to fit: R_PPC_ADDR24 excEnt > C:\Tornado_AE/target/lib/objPPC860gnuvx/excConnectALib.o(.data+0x74): > relocation truncated to fit: R_PPC_ADDR24 excExcHandle > C:\Tornado_AE/target/lib/objPPC860gnuvx/excConnectALib.o(.data+0x7c): > relocation truncated to fit: R_PPC_ADDR24 excExit > > make[1]: *** [bootApp_reloc] Error 0x1 > make[1]: Leaving directory `R:\bootApp\test_ads\PPC860gnu_romCopy' > make: Leaving directory `R:\bootApp\test_ads' > make: *** [bootApp_romCopy] Error 0x2 > > It seems that I cannot modified RAM_LOW_ADRS and RAM_HIGH_ADRS to be > higher than 0x03F00000 approximately. > > Is their anyone that have an idea about the nature of my problem? > > Thanks. > > David --------------------------- Newsgroups: comp.os.vxworks Subject: Re: XON-XOFF on the Console port Date: Sun, 10 Feb 2002 08:49:38 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <2002130-143411-880483@foorum.com> <200221-134637-250830@foorum.com> <1013284083.9121.0.nnrp-01.d4e4b8d1@news.demon.co.uk> Cecil, I beleive you have the case of badly designed RS232 hardware. If the designers follow the standard closely, and use whatever application notes they can get from whoever makes the driver chip they use, this would not have happened. EIA calls for a low impedance of the Rx signal, which would absorb any EMI generated signals, and would supress any xross talk (what you refer to as reflection) from the Tx signal. Your hardware designers created the Rx with a high impedance, which will cause trouble (e.g. FEXT) even if the cable was connected. As to VxWorks, it seems quite silly to have XON/XOFF enabled on the console port, WindRiver have put it in, but you could change that by modifying the console intialization parameters in usrConfig.c. Its really easy to do except most embedded developers are unaware of this. "cecil" wrote in message news:1013284083.9121.0.nnrp-01.d4e4b8d1@news.demon.co.uk... > We had this problem a couple of years a go. By default, our shell was > assigned to a serial port. If a ctrl-s was received by the port that the > shell was using while the machine as booting, the boot process was > suspended, until you sent a ctl-q. You couldn't ping it, nothing. We would > also occasionally see the message you have seen. That the power of > xon-xoff!!!!! We were getting ctl-s randomly. This was because when > vxworks prints out all that vxworks booting stuff, and our serial port was > disconnected the output to the shell was been reflected back as random > characters in to the shell. Now that may be specific to our > application...but the moral to the sort is that issuing xon/xoff to the > shell while it is booting kills that machine in the way you described. > Hope that helps > > Nigel > Charly wrote in message <200221-134637-250830@foorum.com>... > > > >Thanks for your help, > >The problem is still there "dosFsDevInit failed". > > > >When I start a big application on my PCPentium target, the shell display > >"dosFsDevInit failed". Then, the target seems dead, no Shell and no target > >server connection. This application works fine on an other BSP > (Lp-elektronik). > >Do you have an idea where I could look? VxWorks configuration? > >All my files are on a FTP server. > >I use also a RAMDISK. > > > >Here are the result of some shell command. > >-> devs > >drv name > > 0 /null > > 1 /tyCo/0 > > 1 /tyCo/1 > > 2 /pcConsole/0 > > 2 /pcConsole/1 > > 6 host: > > 7 /vio > >value = 0 = 0x0 > > > >-> ls > > > > Volume in drive D has no label > > Volume Serial Number is 12E3-344A > > Directory of D:\kern\Awd_num\T2mak\rtl > > > >[.] [..] RTL.WPJ USRAPP~1.C PRJCOMPS.H > >PRJPAR~1.H LINKSYMS.C PRJCON~1.C MAKEFILE [DEFAULT] > >PRJOBJS.LST ADD_VXW.C USRAPP~2.C [DEBUG] > > 10 file(s) 98,644 bytes > > 4 dir(s) 1,170,317,312 bytes free > >value = 0 = 0x0 > > > >-> version > >VxWorks (for PC PENTIUM) version 5.4.2. > >Kernel: WIND version 2.5 > >Made on Feb 1 2002, 13:09:39. > >Boot line: > >ene(0,0)host:vxWorks e=192.0.1.2:ffffff00 h=192.0.1.1 u=target pw=vxworks > f=0x8 > >tn=target > >o=ene. > >value = 0 = 0x0 > > > > > >-- > >Use our news server 'news.foorum.com' from anywhere. > >More details at: http://nnrpinfo.go.foorum.com/ > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Beginner Client/Server problems! Date: Thu, 14 Feb 2002 07:20:14 GMT From: "Gary M" Organization: AT&T Broadband Message-ID: References: Well, you have told us nothing about your hardware architecture or network interfaces, so I don't know how much help you can expect. Have you looked at your network statistics? "Joseph Cook" wrote in message news:b6f5d6a9.0202131121.79ad0d13@posting.google.com... > Ok. The problem was in the bits of code unshown. So, now I am timing > this sequence for a 256 byte buffer, 512, 1K, 2K, 4K, etc. Everything > looks fine (even if terrible) until I hit 4K buffer size, and then the > performance comes to a screaching halt, outputting 40Kb/s (when the 2K > buffer size provides 6 Mb/s). I'm the only process on this system, > and I'm my own independent network...any ideas? > > J.C. > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Help on SDRAM performance consideration Date: Thu, 14 Feb 2002 07:29:19 GMT From: "Gary M" Organization: AT&T Broadband Message-ID: References: Since I don't know which BSP you are using, I'd suggest looking at romInit.s for starters, since the memory controller and the SDRAM have to be initialized before bootInit starts using it. You'll need to get cozy with the data sheet on the memory controller and the SDRAM to convince yourself that the SDRAM system is optimized. On the MPC8xx family, SDRAM performance is limited by the inability to overlap operations in the UPM; however, with cache and Memory Management enabled, 16-byte bursts are typically performed. "Raymond Yeung" wrote in message news:d83bb8e3.0202131030.43232839@posting.google.com... > I inherited the BSP from WindRiver. How do I check whether SDRAM > bursting has been enabled (i.e. which register(s))? > > > "Gary M" wrote in message news:... > > "Joe Chung" wrote in message > > news:rpsu1smif2c.fsf@nospam.invalid... > > > > "Higher clock speed translates to faster SDRAM access time". > > > > Based on some reading I've done so far, this statement is probably not > > true. > > > > There're first access latency, pipelining, bursting. > > > > > > The S in SDRAM means that data (after an initial latency) is able to burst > > > out at the rate of bus speed, synchronized on the bus clock edge(s). > > > > > > > And unless the system actually uses the burst mode (and the larger the burst > > size, the better), the benefits of SDRAM are basically lost. SDRAMs also > > support a great deal of overlap between command phases and data phases, and > > the memory controller and the operating system must be capable of utilizing > > this feature. > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: multiple instances of a variable Date: 10 Feb 2002 04:14:33 -0800 From: rashanmu@npd.hcltech.com (Ramesh) Organization: http://groups.google.com/ Message-ID: <15a87607.0202100408.57072dbb@posting.google.com> References: <3C63EA94.E47EFE94@res.raytheon.com> <4f09bb68.0202081258.4f0bf8ca@posting.google.com> Nick, In VxWorks, there is only a single flat address space. Hence when u name a global variable from a task, there will be memory created for that global variable. So from any other task if you access the global variable, it will give whatever value stored since the last operation done on that variable. So lkup will give the value of a variable which has only one instance. You can have a global variable, which can be local to your task by having it as task variable. Ramesh ================================================================ akarpov@my-deja.com (Alex) wrote in message news:<4f09bb68.0202081258.4f0bf8ca@posting.google.com>... > Nick, > > Your tasks share the same global variables unless you spawn your tasks > in different domains under VxWorks AE. > > Alex > > willisn wrote in message news:<3C63EA94.E47EFE94@res.raytheon.com>... > > Hi all, > > > > My question concerns multiple instances of a single global variable. > > > > I have 3 tasks running, all have an instance of a global variable (say > > an integer). > > > > If I want to know the current value of this integer for any one of the > > tasks, I need to do a lkup on the variable and then dereference the > > address associated with the task I am interested in. > > > > My question is, if you simply type the name of the variable in the > > console, how does vxworks decide which instance of the variable to give > > you? > > > > Thanks in advance for any help! > > > > Nick --------------------------- Newsgroups: comp.os.vxworks Subject: Equivalent of setitimer Date: 10 Feb 2002 06:26:24 -0800 From: bryanramesh@yahoo.com (Bryan) Organization: http://groups.google.com/ Message-ID: Hi, In Linux setitimer is there to set a timer and when the timer expires SIGALRM will be sent. In VxWorks what is the equivalent of that? In timer_create call where shall we mention the timer value. THanks, Bryan --------------------------- Newsgroups: comp.os.vxworks,comp.os.os9 Subject: FS: VME and CPCI, Current SBCs and Chassis to 700MHz Date: 10 Feb 2002 07:48:33 -0800 From: wmcalpin@ghgcorp.com (Bill) Organization: http://groups.google.com/ Message-ID: Look here: http://www.sunharborsales.com/shsvmefs.htm --------------------------- Newsgroups: comp.os.vxworks Subject: Re: how i build bootrom Date: 12 Feb 2002 21:11:59 -0800 From: alimohd@logiceastern.com (Ali) Organization: http://groups.google.com/ Message-ID: References: <4f09bb68.0202081309.2e2c0c0d@posting.google.com> hi I was successful in creating a bootrom_uncmp of my bsp. also the bootrom_uncmp_hex is also made but when i try to make bootrom or botrom_res it gives errors like sysClkRateGet() not refrenced in miiLib.o and errors like undefined parameters like ROM_DST_ADRS.. what changes i have to make further . thanx alimohd@logiceastern.com (Ali) wrote in message news:... > hi > this file is there buti need this file in building my image so when i > remove this from tree i can't build the image and if i keep this i am > unable to build bootrom . > ali > akarpov@my-deja.com (Alex) wrote in message news:<4f09bb68.0202081309.2e2c0c0d@posting.google.com>... > > Go and double check if there are such a file in the main Tornado tree > > Tornado/target/src/drv/intrCtl/m8260IntrCtl.c > > or in your BSP : > > $(BSP_DIR)/src/drv/intrCtl/m8260IntrCtl.c > > > > > > alimohd@logiceastern.com (Ali) wrote in message news:... > > > > > > > make: *** No rule to make target `drv/intrCtl/m8260IntrCtl.c', needed > > > > > by `sysLib.o'. Stop. > > > > vxworks --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Equivalent of setitimer Date: Sun, 10 Feb 2002 09:09:56 -0800 From: "Michael R. Kesti" Organization: MK Associates Message-ID: <3C66A964.A7BDD3D0@gv.net> References: Reply-To: mkesti@gv.net Bryan wrote: >In Linux setitimer is there to set a timer and when the timer expires >SIGALRM will be sent. In VxWorks what is the equivalent of that? In >timer_create call where shall we mention the timer value. See the vxWorks docs for "wdLib", where "wd" stands for "watchdog" as in "watchdog timers". - -- ======================================================================== Michael Kesti | "And like, one and one don't make | two, one and one make one." mkesti@gv.net | - The Who, Bargain --------------------------- Newsgroups: comp.os.vxworks Subject: how should i start Date: 12 Feb 2002 21:18:59 -0800 From: alimohd@logiceastern.com (Ali) Organization: http://groups.google.com/ Message-ID: Hi All I want to make support a tffs in my project. The flash we are using is 28F128J3A/28F640J3A these are intel strata flashes. The MTDs for these flashes are not given in WRS supplied bsp's .. so can anybody tell me how should i start with. Some have told that it is better to start with cfiscs.c .But i suppose it is for 8 bit interleaving but i need 16 bit interleaving. My flash is onboard 4M flash and is a single chip, can anybody tell me how should i start with and where to make necessary changes to support all this. Should i consider it a SIMM or DOC. I am sure number of guys have worked on this file system so i am sure of getting a good help. thanx in advance ali --------------------------- Newsgroups: comp.os.vxworks Subject: Re: anticipating machine check error... Date: Thu, 14 Feb 2002 09:44:16 +0100 From: "Michael Lawnick" Organization: Buergernetz Dillingen Message-ID: References: <8fe86b76.0202131506.3c00702d@posting.google.com> Reply-To: "Michael Lawnick" Sender: mlawnick@pd95247e2.dip.t-dialin.net Hi David, if your problem is only 'first access', your function is vxMemProbe(). If you fear card fail while operating, use excHookAdd() and setjmp/longjmp. If you don't like the logMsg of the vxWorks excHandler, you will need excVecSet(). Use setjmp/longjmp here too. Incrementing PC would be for 'transparent operation fail', i.e. if failing task shouldn't see its fail. IMO this is a solution for players, not a solid one. - -- Mit freundlichen Grüßen, Michael Lawnick ============================================== SOFTEC GmbH Tel +49-731-96600-0 Promenade 17 Fax +49-731-96600-23 D-89073 Ulm Michael Lawnick Germany lawnick@softec.de ============================================== "David" schrieb im Newsbeitrag news:8fe86b76.0202131506.3c00702d@posting.google.com... > I was hoping somebody could verify this for me: > > I am porting an application from PSOS to VxWorks 5.4 on a Power PC > platform (mv5100). > > The existing application has a set of exception handling routines that > are installed (in the PSOS assembly code) to handle machine check > errors. The errors in question occur when card signal processing is > attempted on a card that is missing from the chassis. We don't wan't > the system to halt if a card is missing or not working correctly. > > From reading previous posts, it looks like I will have to use the > excVecSet() function, but those posts also mention manually > incrementing the Program Counter which sounds a bit hairy. My current > handlers get passed: > > vector offset for the exception, > addr - next program location when the exception occurred > ExceptionFramePtr - exception frame pointer > > but the only thing I do with this information is display it in the > error log. > > Any ideas on how to proceed would be greatly appreciated, > > David --------------------------- Newsgroups: comp.os.vxworks Subject: Re: adding new function to be called from prjConfig.c Date: Thu, 14 Feb 2002 09:46:35 +0100 From: "Michael Lawnick" Organization: Buergernetz Dillingen Message-ID: References: Reply-To: "Michael Lawnick" Sender: mlawnick@pd95247e2.dip.t-dialin.net Hi Ofer, edit the apropriate *.cdf file in target/config/comps/vxWorks or add a new one (with incremented prefix) in your BSP folder. HTH - -- Mit freundlichen Grüßen, Michael Lawnick ============================================== SOFTEC GmbH Tel +49-731-96600-0 Promenade 17 Fax +49-731-96600-23 D-89073 Ulm Michael Lawnick Germany lawnick@softec.de ============================================== "Ofer Goren" schrieb im Newsbeitrag news:newscache$nkjfrg$yk8$1@lnews.actcom.co.il... > > -- > ------------------------ > Win2k Sp2 > Tornado 2.02 for PPC > Diab 4.4b > Envoy 9.2 > ------------------------ > Hi. > > How do I add a function to be called from orjConfig.c? This is a > auto-generated file, so whenever I rebuild my project, my changes are > gone.... > > Thanks, > > -- > Ofer Goren > Iamba Technologies > > --------------------------- Newsgroups: comp.os.vxworks Subject: Project-relative files? Date: Thu, 14 Feb 2002 01:05:52 -0800 From: Bob Bradley Organization: Uh huh huh...It says Organ Message-ID: Is it possible to convince Tornado to reference files by relative path that are above the project file in the hierarchy when added to the project? It seems to use project-relative paths when the files are below it in the hierarchy, but uses full paths for files above it. For example, I might have a hierarchy as follows: Sources Make MyProject.wpj Misc MyFile.c When I add "MyFile.c" to "MyProject.wpj", I'd like it to be saved in the project file (and the resulting makefile) as "$(PRJ_DIR)../Misc/MyFile.c" instead of using the full path to MyFile.c. Our project is built on a variety of systems so using a full path doesn't work. Is this possible? --------------------------- Newsgroups: comp.os.vxworks Subject: Handling select() function failures Date: Sun, 10 Feb 2002 14:15:14 -0500 From: Chandwani Organization: Lucent Technologies Message-ID: <3C66C6C2.C3B3CA0E@ubbsrv3.ho.lucent.com> Hi ! I am facing a problem with select() function. Sometimes this function returns with an ERROR (all fds are fine). The vxWorks manual doesn't say exactly why thsi happens and how to handle this situation. My suspiscion is that if I put select() in a FOREVER loop and it returns an ERROR, the task will go in an infinite loop. Why does select() return an ERROR and how should I handle it in a FOREVER loop then ? One more question : UDP broadcasts to 135.0.0.0 network don't work. They work well to 192.0.0.0 network. Has anyone else seen this behavior in their environment ? Thanks, Rakesh --------------------------- Newsgroups: comp.os.vxworks Subject: Re: ULIP driver ... Date: Wed, 13 Feb 2002 05:41:08 GMT From: "Kim Letkeman" Message-ID: References: Oops ... I just noticed that the 2k version is in the x86 bin directory. All I really need now is a setup mechanism that works for XP. Anyone hacked something up yet? "Kim Letkeman" wrote in message news:RCla8.2135$oG.1544@news2.bloor.is... > A few posts over the years have alluded to people running the ULIP driver > for vxsim under Windows 2000. Were those all just for PPC architecture? Was > there ever a ULIP driver for Windows 2000 x86? I ask because I am unable to > run ULIP at home (XP) and I'd like to be able to test there. I _hate_ the > thought of dual booting just for this purpose. > > Is there anything available? > > Why in the world would WindRiver not release the ULIP source code into the > public domain if they refuse to support it on OSs newer than 6 years old? > > Kim > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Equivalent of setitimer Date: 12 Feb 2002 21:59:11 -0800 From: bryanramesh@yahoo.com (Bryan) Organization: http://groups.google.com/ Message-ID: References: <3C66A964.A7BDD3D0@gv.net> Dave, Yeah without the help of interrupts we cannot implement timerhandlers, since we are interacting with the system clock. Anyway what i thought is instead of providing the user the functionality of executing the timer expiry code, i thought timer_settime will be bit safer comparitively. But the implementation of timer_settime will be hidden from the user and user is not handling the ISR and the only thing he can d is execute a piece of code and a handler to a signal. which is not that much dangerous like the ISR eventhough signal handler also will be executed immediately when received, even the task is blocked at that time. Hope this makes my arguement clear. Thanks, Bryan. =============================================== "Dave Korn" wrote in message news:... > > > > just want to point out that an ISR can call kill(...) to send a signal. but > then timer_settime is probably doing pretty much that internally anyway! > > DaveK --------------------------- Newsgroups: comp.os.vxworks Subject: Unix : getting symbols for PowerPc Date: Thu, 14 Feb 2002 10:37:57 +0100 From: "Pradel, Jean-Marc [CTF:4907:EXCH]" Organization: Nortel Networks Message-ID: <3C6B8575.C64E03ED@europem01.nt.com> Hello, I'm working on an UNIX system with core PowerPc. I'm writing a trap handler in order to get registers, stack and symbols corresponding to the different calls. I managed decoding stack and finding closest symbol (like any debugger is doing that) but I still got some problems to find symbols in shared libraries. I think I'll have to analyse TOC table to get the indirection but I would like to know if something is not already existing. Thanks a lot for your help. Jean-Marc PRADEL ( 6-574-7313 ) --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Help on SDRAM performance consideration Date: Wed, 13 Feb 2002 06:19:05 GMT From: "Gary M" Organization: AT&T Broadband Message-ID: References: "Joe Chung" wrote in message news:rpsu1smif2c.fsf@nospam.invalid... > > "Higher clock speed translates to faster SDRAM access time". > > Based on some reading I've done so far, this statement is probably not true. > > There're first access latency, pipelining, bursting. > > The S in SDRAM means that data (after an initial latency) is able to burst > out at the rate of bus speed, synchronized on the bus clock edge(s). > And unless the system actually uses the burst mode (and the larger the burst size, the better), the benefits of SDRAM are basically lost. SDRAMs also support a great deal of overlap between command phases and data phases, and the memory controller and the operating system must be capable of utilizing this feature. > -jc > -- > (apply 'concat (reverse (list "com" > (char-to-string 46) "yahoo" > (char-to-string 64) "joechung"))) --------------------------- Newsgroups: comp.os.vxworks Subject: Re: +++ C-Designer (sehr flexibel) für Realtime/C/Systeme sucht neues Projekt +++ Date: Thu, 14 Feb 2002 10:46:43 +0100 From: "Werner Schiendl" Message-ID: <3c6b877e$1@brateggebdc5.br-automation.co.at> References: <3c5e5151@brateggebdc5.br-automation.co.at> <3c6832b7@brateggebdc5.br-automation.co.at> Hi Dave, - -- snip -- > > Oh I agree with you; it is de facto an english language list/group; but > there's no need to tell him it's incorrect to post in german, it would be > more accurate to say it's *pointless* to post in german. And speaking only > for myself, I like seeing the odd bit of foreign language show up. It gives > me an incentive to try and remember some of the foreign language lessons I > had in school and I always like a challenge! > agreed > > > > P. S. I hate spam, including Martin R's and this one. _Particularly_ if it > > is repeated every week. > > > > mmm. I don't like those either, but there ya go. If the BI stays < 20 in > 45 days, it's not considered spam by the current usenet > standards/guidelines, except in groups which explicitly have more stringent > restrictions. Even though we're in big8 not alt here, it's still an > unmoderated group. > Hm, but if anyone currently unemployed or looking for some consulting engagement posts to the list/newsgroup, we'll soon have tons of such messages in the group. I'm not here to find out that Mr. X living in India needs a job, but to improve VxWorks skills, learn about the system, volunteer some help and sometimes to ask something myself. I think this is what most readers/posters expect from the list. Hopefully, this kind of advertisment does not yield a business or employment - then it might regulate itself. best regards Werner --------------------------- Newsgroups: comp.os.vxworks Subject: Re: I2C and SPI underrun and overrun error Date: 10 Feb 2002 14:14:49 -0800 From: ntjiang@yahoo.com (Jason Jiang) Organization: http://groups.google.com/ Message-ID: <12116a88.0202101414.19b19f91@posting.google.com> References: <12116a88.0202071113.7ae1e042@posting.google.com> <3c62f2bd$1@brateggebdc5.br-automation.co.at> Thanks, Werner. I'm using PPC855t, vxWorks version 5.4.2. And the clock settings in the drivers are: I2C: SCL = I2CCLK = BRGCLK / I2MOD[PDIV] / (2 * (I2BRG[DIV]+3) ) = (VCOOUT / SCCR[DFBRG]) / I2MOD[PDIV] / (2 * (I2BRG[DIV]+3) ) = (OSCCLK * (PLPRCR[MF]+1) / SCCR[DFBRG]) / I2MOD[PDIV] / (2*(I2BRG[DIV]+3)) = 4M * 12 / 1 / 8 / (2 * (5+3) ) = 375 kHz SPI: SPICLK = BRGCLK / SPMODE[DIV16] / (4*(SPMODE[PM]+1)) = (VCOOUT / SCCR[DFBRG]) / SPMODE[DIV16] / (4*(SPMODE[PM]+1)) = (OSCCLK * (PLPRCR[MF]+1) / SCCR[DFBRG]) / SPMODE[DIV16] / (4*(SPMODE[PM]+1)) = 4M * 12 / 1 / 1 / (4 * (15+1) ) = 750 kHz Let me know if you need more information. Thanks a lot. Jason "Werner Schiendl" wrote in message news:<3c62f2bd$1@brateggebdc5.br-automation.co.at>... > Hi, > > Report more details (used hardware, drivers, VxWorks version, error numbers > and/or messages, etc.) if you hope for a useful comment. > > regards > Werner > > "Jason Jiang" wrote in message > news:12116a88.0202071113.7ae1e042@posting.google.com... > > My I2C and SPI driver sometimes will report underrun or overrun error > > (SPI driver will report multi-master error too). And seems it happens > > most likely in the first several times after powering up. I have no > > idea what's going on? Pls help. Thanks. > > > > Jason --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado II Prototyper installation failure Date: Thu, 14 Feb 2002 11:06:20 +0100 From: "Werner Schiendl" Message-ID: <3c6b8c17$1@brateggebdc5.br-automation.co.at> References: <576881c0.0202121417.e1df9b0@posting.google.com> Hi, it seems there is something broken with your downloaded file. I have downloaded the Prototyper .tar file a couple days ago and it contains the following files: Disk_id 80 Bytes Wind.000 2,535,146 Bytes Wind.001 5,727,500 ... Wind.002 4,945,701 Wind.003 5,905,079 Wind.004 4,670,421 Wind.005 5,555,023 Wind.006 5,246,791 Wind.007 4,014,090 Wind.008 4,332,725 Wind.009 4,915,579 The archive file downloaded is 47,858,688 bytes in size. Maybe something went wrong while downloading the file. If your archive does not contain the above files, try re-downloading the archive. I used WinZip 8.0 to view/extract the archive. hth Werner "Nick Forrette" wrote in message news:576881c0.0202121417.e1df9b0@posting.google.com... > First off, I'm very new to VxWorks and have been assigned to learn a > bit about it. As part of my exploration, I've tried to install the > Tornado II Prototyper under Windows 2000. I am having two problems > with the installer. > > 1. The Tornado II Prototyper tar file they ship includes two files > named "inatr:intf". Of course colons are illegal characters for > filenames in Windows. I worked around this by extracting everything > but those two offending files. > > 2. When running the installer, everything goes smoothly till it > actually starts to copy file files. The progress bar jumps immediately > to 100% done and hangs there. After being patient (waiting 30 minutes+ > just to be sure), I canceled the install. > > Maybe problem #2 is a result of the files I couldn't extract because > of #1. Any thoughts? > > TIA, > Nick --------------------------- Newsgroups: comp.os.vxworks Subject: Tool for checking memory leak Date: Thu, 14 Feb 2002 10:40:09 GMT From: Glenn Ivar Klausen Message-ID: <3C6B932E.C080423B@oslo.online.no> Is there a tool simular to purifi or boundchecker in vxWorks? - -- Glenn --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Reading from serial ports - Blocking or interrupts? Date: Thu, 14 Feb 2002 11:36:42 +0100 From: "Werner Schiendl" Message-ID: <3c6b9334$1@brateggebdc5.br-automation.co.at> References: <37449c9d.0202130624.43ffc949@posting.google.com> Hi, see comments inside. wrote in message news:37449c9d.0202130624.43ffc949@posting.google.com... > I am working on an application using VxWorks 5.4 to read from a number > of serial ports simultaneously. > > Each serial port is read by its own separate VxWorks task. > > The serial ports receive pre-defined, but variable-length (over some > fairly close limits) messages. > > Each serial port comes with a 128-byte buffer, and the receive > interrupt can be set to fire on any number of characters in the buffer > from 1 to 128. It can also generate an interrupt if 4 character times > pass with no input, with the FIFO only partially filled (i.e. not > filled enough to reach its interrupt level). > > Timing for processing incoming messages is critical, i.e. each task > must process its received message as soon as practical. > > My question is, which of the following scenarios is likely to be the > most efficient in terms of use of processor resources, while still > meeting the objective of processing incoming messages as quickly as > possible: > > 1. Set the interrupt to interrupt on the length of the shortest > received message. The interrupt sends a signal to the input task, > which then reads and stores the data received so far, then goes back > to sleep. The tail end of the message is handled by the 4-character > timeout-with-no-data interrupt. > In my opinion, this is probably the most efficient model. If you can live with 4-character times delay, of course. > 2. Generate an interrupt on each character, which sends a signal to > the input task which processes incoming characters on a > character-by-character basis. > This will produce a _very_ high load on your CPU. Keep in mind, that the 16550 FIFOs in the PC's serial ports have been introduced to allow for fast transfer modes (57600, 115200) on 486 (maybe 386) machines. You have a 'number' of serial ports, so you may end up being slower that option #1 just because processing each character seperately wastes a lot of time for context switches. The costs of a context switch depends on your hardware, but even Pentium processors spend some usec to enter and leave an interrupt. > 3. Use a blocking read function to read from the serial port. No > interrupts are needed, and incoming data is handled on a > character-by-character basis. > How should a blocking read function work? It will use an interrupt handler internally, to unblock the task on the arrival of a new character. So, if I understand what you wanna try correct, this is basically the same then option #2. 4) An other option you have is to use a polling scheme. This means, your UARTs do net generate an interrupt when new data arrives. Instead you poll them in a resolution suitable for your system to find out if new data has arrived. If handling these UARTs is the 'function' of your device, this may be an appropriate design - otherwise it may waste to much processor time for the polling. 5) Write a special driver that allows polling in addition to the normal, interrupt driver transfer. If you are able to determine the size of an unfinished packet, you'll know when there is not a full 128 buffer left to read and read the rest in polled mode. Or - if the UART allows changing the interrupt level 'on the fly' - change it to single byte interrupts until the packet is complete, then back to the big buffer. > It seems to me that #3 is likely to be the most efficient, assuming > the serial port board's BSP supports it (we don't have the BSP > documentation yet so I can't be sure but I'd be very surprised if it > didn't). For option #4 you need a driver that can do 'polled' mode. WRS drivers support this (usually) for system mode debugging with CrossWinds. See the documentation on how to use polled mode serial drivers. Option #5 will most probably need you coding your own driver. > > But, that raises another question ... > > In general, what is the overhead for unblocking read functions, > compared with the overhead of processing an interrupt? I'm concerned > that there might be some goofy internal hidden stuff that goes on in > scenario #3 such that #2 is actually faster and more efficnent. > It is the same, since unblocking always requires an interrupt (theoretically, it could be unblocked using a polling task, but this is probably not a common design for a serial driver - this would require a speperate task the checks the hardware on a cyclical basis). > Also, when a task is blocked, i.e. on a read of a serial port, does it > ALWAYS relinquish the CPU for use by lower-priority tasks, and then > take over the CPU again when data appears on the serial port? > Yes, if - of course - the driver has been properly coded. hth Werner --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Unix : getting symbols for PowerPc Date: Thu, 14 Feb 2002 11:56:26 +0100 From: "Michael Lawnick" Organization: Buergernetz Dillingen Message-ID: References: <3C6B8575.C64E03ED@europem01.nt.com> Reply-To: "Michael Lawnick" Sender: mlawnick@pd95247e2.dip.t-dialin.net Hi Jean-Marc, what's about lkAddr() ? - -- Mit freundlichen Grüßen, Michael Lawnick ============================================== SOFTEC GmbH Tel +49-731-96600-0 Promenade 17 Fax +49-731-96600-23 D-89073 Ulm Michael Lawnick Germany lawnick@softec.de ============================================== "Pradel, Jean-Marc [CTF:4907:EXCH]" schrieb im Newsbeitrag news:3C6B8575.C64E03ED@europem01.nt.com... > Hello, > > I'm working on an UNIX system with core PowerPc. > I'm writing a trap handler in order to get registers, stack and symbols > corresponding to the different calls. > I managed decoding stack and finding closest symbol (like any debugger > is doing that) but I still got some problems to find symbols in shared > libraries. > I think I'll have to analyse TOC table to get the indirection but I > would like to know if something is not already existing. > > Thanks a lot for your help. > > Jean-Marc PRADEL ( 6-574-7313 ) --------------------------- Newsgroups: comp.os.vxworks Subject: Re: ULIP driver ... Date: Wed, 13 Feb 2002 07:50:15 GMT From: "Kim Letkeman" Message-ID: References: Well, I found the magic mechanism to try to install the two drivers (NT & Win2k). Bad news ... NDIS.SYS is very unhappy with ulip ... so it's blue screen city for both. Bummer, really ... it would be so nice. "Kim Letkeman" wrote in message news:U%ma8.8983$LL.8472@news1.bloor.is... > Oops ... I just noticed that the 2k version is in the x86 bin directory. All > I really need now is a setup mechanism that works for XP. Anyone hacked > something up yet? > > > "Kim Letkeman" wrote in message > news:RCla8.2135$oG.1544@news2.bloor.is... > > A few posts over the years have alluded to people running the ULIP driver > > for vxsim under Windows 2000. Were those all just for PPC architecture? > Was > > there ever a ULIP driver for Windows 2000 x86? I ask because I am unable > to > > run ULIP at home (XP) and I'd like to be able to test there. I _hate_ the > > thought of dual booting just for this purpose. > > > > Is there anything available? > > > > Why in the world would WindRiver not release the ULIP source code into the > > public domain if they refuse to support it on OSs newer than 6 years old? > > > > Kim > > > > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: MPC860 interrupt latency variation Date: Wed, 13 Feb 2002 08:13:59 GMT From: "Gary M" Organization: AT&T Broadband Message-ID: References: <3C694008.DF0FB22D@st2e.com> What model of MPC860 and at what speed? By "interrupt latency" do you mean the time from the interrupt assertion until your ISR begins execution? We've seen latencies (as defined above) on an MPC855T at 50 MHz on the order of 7-8 microseconds. This is longer if any higher priority ISR is already executing or if interrupts are masked. Locking critical code in cache would certainly improve these numbers. "Christophe Couturier" wrote in message news:3C694008.DF0FB22D@st2e.com... > > Hi all, > > My application is synchronised by an external signal causing an > interrupt. > Althought I don't care if the interrupt latency is long, I need the > variation of this latency to be very short (~+/-5ľs). > Does anyone knows how to configure VxWorks and the PowerPc to reach this > goal. > > I use VxWorks 5.4 and a MPC860. > > Thanks. > > Chrsitophe > --------------------------- Newsgroups: comp.os.vxworks Subject: Memory Management: Questions about Partitions Date: Wed, 13 Feb 2002 10:06:51 +0100 From: "Juergen Ferber" Organization: - Message-ID: Hello all, I am evaluating the Tornado Prototyper in combination with VxSim, after it took me only some days to get around the download and license trouble ;-) To have some usefull code at the end, I am trying to write a sort of Memory Management System. This should decrease memory fragmentation. With the Prototyper I got a lot of documentation. Thousands of pages, but the detailed information is missing and I could not find it on the web. Here my design: I have some partitions containing buffers. One partition with a buffersize of 16 bytes, one with 64,... and one partition for all memory requests that do not fit to the others. Those partitions are within a big "Master" partition and are created during initialisation. When memory is allocated, my Manager should allocate a buffer that fits best and return it to the pool when it is not needed any more. I am using the memPartLib and mainly the routines memPartCreate( ) and memPartAlloc( ). 1. Should I keep the "Master" partition or should I remove it(advantages/disadvantages)? 2. How can I calculate the real size of a partition? I think a partition containing 100 bytes needs some bytes for a header, or am I wrong? 3. Is the documentation more detailed, when buying Tornado? Thanks in advance for any help! Jürgen --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Thu Feb 14 04:54:08 2002 From: "Tyler, Paul A (DBY)" Date: Thu Feb 14 04:54:11 PST 2002 Subject: High stack usage during object creation. I have a come across a potential problem in porting programs for VxWorks 5.3.1 to 5.4. As yet I do not understand the cause. I launch the program from the start script using taskSpawn The program is as follows: // PressureBricksClient - shell routine for class int PressureBricksClient( char* reg_path ) { CPressureBricksClient *client; int s; client = new CPressureBricksClient; client->set_reg_path( reg_path ); s = client->main(); delete client; return s; } If I use the default stack size of 20,000, I don't see a problem using checkStack, but the tasks launched earlier in the script behave unpredictably, sometimes reporting General Protection Fault, sometimes rebooting. If I increase the stack size to 200,000, I see a high water mark about the 160,000. Current stack usage is always low, about the 400 mark. Putting checkStack into the code shows that the high water mark is reached during the "new". If I try to debug this the debug session fails when I step into the "new". I have not seen this problem with other client programs which use this technique. Similarly I have not seen this problem with VxWorks 5.3.1 (I went back to check). Questions: 1. Why would stack be used (albeit temporarily), when the "new" command suggests allocation of memory from the pool? 2. How can I increase the debugger stack size to try to get to the bottom of this? I am using Tornado 2 on a PC host, with a VMIVME-7698 (Pentium based) target. Thanks Paul Tyler Measurement Systems Specialist Rig Test Systems Team, Facility Engineering Dept. Rolls-Royce plc. e-mail Paul.A.Tyler@Rolls-Royce.com Tel. +44 (1332) 247808 Fax. +44 (1332) 247928 The data contained in, or attached to, this e-mail, may contain confidential information. If you have received it in error you should notify the sender immediately by reply e-mail, delete the message from your system and contact +44(0)1332 248119 (the Rolls-Royce IT Security Director) if you need assistance. Please do not copy it for any purpose, or disclose its contents to any other person. An e-mail response to this address may be subject to interception or monitoring for operational reasons or for lawful business practices. (c) 2002 Rolls-Royce plc From vxwexplo-errs@csg.lbl.gov Thu Feb 14 12:58:48 2002 From: "Bob Kindel" Date: Thu Feb 14 12:58:50 PST 2002 Subject: RE: [Fwd: ProfileScope data logging] Hi Jamil- You can open ProfileScope data in Excel by doing the following: 1. In ProfileScope press the "Save Text..." button on the top bar, or use the pulldown menu File->Save Text. You can do this on either a snapshot or the live data window. 2. Specify the name of the file that you want to save it to. 3. Open Excel 4. Go to Data->Get External Data->Import Text File 5. The wizard will guide you through importing the data. You probably only want to import the data after line 4. Regards, Bob Kindel Real-Time Innovations, Inc. > > -------- Original Message -------- > Subject: ProfileScope data logging > Date: Thu, 14 Feb 2002 03:43:49 GMT > From: "Jamilur Rahman" > Newsgroups: comp.os.vxworks > > How can I save the ProfilrScope data in Excel or other data base > format for > future reference ? > I also need recommendation of graph drawing software (preferably freeware) > for plotting 2-dimensional plot. > Thanks in advance for any help. > > /jamil, > Ottawa, Ontario. > From vxwexplo-errs@csg.lbl.gov Thu Feb 14 13:53:28 2002 From: karunakaran nair Date: Thu Feb 14 13:53:30 PST 2002 Subject: comp.os.vxworks newsdigest some times it is easy to use binary flags which are executed atomically. when an event occurs we can set this flag, a task can then check on this flag and take decisions. is there anything like this in vxWorks ? thanks, KN __________________________________________________ Do You Yahoo!? Send FREE Valentine eCards with Yahoo! Greetings! http://greetings.yahoo.com From vxwexplo-errs@csg.lbl.gov Thu Feb 14 13:59:59 2002 From: karunakaran nair Date: Thu Feb 14 14:00:01 PST 2002 Subject: comp.os.vxworks newsdigest suppose I have to change a BSP working with respect to a specific bus X to another bus Y. what files need to be changed ? somebody already has this experience and could be willing to share this for me. can someone tell me what an END driver is ? thanks, KN __________________________________________________ Do You Yahoo!? Send FREE Valentine eCards with Yahoo! Greetings! http://greetings.yahoo.com From vxwexplo-errs@csg.lbl.gov Thu Feb 14 14:51:01 2002 From: "Mike Anderson" Date: Thu Feb 14 14:51:04 PST 2002 Subject: RE: What are END drivers? VxWorks Greetings! > Submitted-by: karunakaran nair > > suppose I have to change a BSP working with respect to > a specific bus X to another bus Y. what files need to > be changed ? somebody already has this experience and > could be willing to share this for me. It all depends on which bus X and which bus Y. There are lots of examples for VME, (c)PCI and ISA bus. There are a few for Sun's SBus. > > can someone tell me what an END driver is ? > Hmm, let's see... A nightmare, a pain in the *ss, a hellish morass in which your packets check in but may never check out, all of these descriptions come to mind ;-)... However, an END (Extensible Network Device) driver is the code that goes from your physical interface, such as Ethernet, and marries that with the multiplexor (MUX) layer to move packets in and out of the protocol stack. These types of drivers are being supplanted by the Network Protocol Toolkit (NPT) drivers that offer a superset of END functionality. The MUX layer and NPT/END drivers can be quite a bit confusing owing to the use of cluster and MBLK buffers and the descriptions of their use in the WRS docs. But, once you get the hang of it, they're not *too* bad. Look at WRS's WindSurf website for the docs that describe NPT drivers and prepare to put on your thinking cap... HTH, Mike -- ======================================================= T h e P T R G r o u p, I n c. ======================================================= Embedded, Real-Time Solutions and Training Mike Anderson mailto:mike@ThePTRGroup.com Chief Scientist http://www.ThePTRGroup.com C: (703) 585-9384 V: (703) 430-3748 ======================================================= "Software development is like making a baby... You can't make a baby in one month by impregnating nine women. Some things just take time." From vxwexplo-errs@csg.lbl.gov Thu Feb 14 14:51:03 2002 From: "Mike Anderson" Date: Thu Feb 14 14:51:07 PST 2002 Subject: RE: Binary flags VxWorks Greetings! > Submitted-by: karunakaran nair > > some times it is easy to use binary flags which are > executed atomically. when an event occurs we can set > this flag, a task can then check on this flag and take > decisions. is there anything like this in vxWorks ? > thanks, Yes, this type of functionality can be implemented using semaphore operations. Alternatively, you can simply use a bit of data with the TAS (test-and-set) instruction, but you'd have to architect things quite carefully with the TAS approach to avoid CPU starvation problems. HTH, Mike -- ======================================================= T h e P T R G r o u p, I n c. ======================================================= Embedded, Real-Time Solutions and Training Mike Anderson mailto:mike@ThePTRGroup.com Chief Scientist http://www.ThePTRGroup.com C: (703) 585-9384 V: (703) 430-3748 ======================================================= "Software development is like making a baby... You can't make a baby in one month by impregnating nine women. Some things just take time." From vxwexplo-errs@csg.lbl.gov Fri Feb 15 04:03:31 2002 From: Vxworks Exploder Date: Fri Feb 15 04:03:34 PST 2002 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Fri Feb 15 04:03:25 PST 2002 Subject: PROXY ARP on Multiple Ethernet Interfaces Subject: Automatic test case generators Subject: Munch for Linux? Subject: serial port settings Subject: Re: vxWorks BSP developments with Galileo GT64260 Subject: Re: MPC860 floating point emulation Subject: Help on SDRAM performance consideration Subject: Re: stack USB from WinRiver on All BSP? Subject: Re: serial port settings Subject: Re: +++ C-Designer (sehr flexibel) für Realtime/C/Systeme sucht neues Projekt +++ Subject: MPC860 interrupt latency variation Subject: USB on the PCPentium BSP? Subject: Question:Capturing stack trace of a taskwhen it crashes Subject: disassember in vxWorks 5.4? Subject: Guru Question: IP stack and Mblks Subject: how to define view mask for mib tree Subject: Re: Reading from serial ports - Blocking or interrupts? Subject: Re: login issue Subject: CodeTEST/Memory or PerformancePak (MemScope) ? Subject: adding new function to be called from prjConfig.c Subject: winmodem support Subject: What's vxImmrGet() function for? Subject: Re: [OT] ULIP driver ... Subject: Re: Reading from serial ports - Blocking or interrupts? Subject: LSI53C1000 driver? Subject: Guru Question # 2: netBufLib and concurency Subject: Re: USB on the PCPentium BSP? Subject: Failover TCP/IP stack for vxworks Subject: prjObjs.lst - error 0x80 Subject: Guru Question # 2: netBufLib and concurency Subject: Flash Drivers ->TFFS & DosFs2.0 Subject: Re: [Help Me] Socket binding error Subject: Re: problem with -mlongcall ( with one more interesting point) Subject: Re: [HELP] network connection with Solaris Subject: Re: MPC860 floating point emulation Subject: Re: 64-bit operations Subject: Re: [HELP] network connection with Solaris Subject: Re: [HELP] network connection with Solaris Subject: Re: Help on SDRAM performance consideration Subject: June 2001 v/s T2CP4 Subject: VirtuallAlloc failed error on compilation Subject: Re: What's vxImmrGet() function for? Subject: Re: Anyone tried to modify 8260 interrupt priority? Subject: Re: prjObjs.lst - error 0x80 Subject: Tornado II Prototyper installation failure Subject: Re: disassember in vxWorks 5.4? Subject: Re: What's vxImmrGet() function for? Subject: Re: 64-bit operations Subject: Re: vxWorks BSP developments with Galileo GT64260 Subject: Re: What's vxImmrGet() function for? Subject: Re: Munch for Linux? Subject: Re: Questions about vxmalloc.c Subject: Re: Questions about vxmalloc.c Subject: Re: Tool for checking memory leak Subject: S 626 PCI card Subject: S 626 PCI card Subject: ULIP driver ... Subject: Re: CodeTEST/Memory or PerformancePak (MemScope) ? Subject: pap and secret files Subject: Re: CodeTEST/Memory or PerformancePak (MemScope) ? Subject: Re: Wind Web Server - problem in httpSymTabAdd Subject: Re: Munch for Linux? Subject: Re: CodeTEST/Memory or PerformancePak (MemScope) ? Subject: Re: Project-relative files? Subject: Re: Tornado II Prototyper installation failure Subject: User defined events not triggering windview? Subject: What happens beneath the 860SAR while setting PDPAR.UT bit ? Subject: Re: VirtuallAlloc failed error on compilation Subject: Re: What's vxImmrGet() function for? Subject: Re: Tool for checking memory leak Subject: Re: Tornado II Prototyper installation failure Subject: Re: Wind Web Server - problem in httpSymTabAdd Subject: Re: CodeTEST/Memory or PerformancePak (MemScope) ? Subject: Re: Project-relative files? Subject: Re: Handling select() function failures ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: PROXY ARP on Multiple Ethernet Interfaces Date: 13 Feb 2002 02:53:03 -0800 From: ramit.bhalla@wipro.com (Ramit Bhalla) Organization: http://groups.google.com/ Message-ID: <78ccdd99.0202130253.1b0fb9d2@posting.google.com> Hi, I was wondering if anyone could help me figure out if it's possible to enable a proxy ARP server on multiple ethernet interfaces. The manual just talks about a proxy arp between shared memory and in ethernet interface. I was trying to enable proxy arp between multiple ethernet interfaces. Can this be done ?? Thanks, Regards, Ramit. --------------------------- Newsgroups: comp.os.vxworks Subject: Automatic test case generators Date: Tue, 12 Feb 2002 04:01:46 GMT From: "Madhav Ancha" Organization: AT&T Broadband Message-ID: Hi, What are some of the famous automatic test case generators for vxWorks? Any idea? --------------------------- Newsgroups: comp.os.vxworks Subject: Munch for Linux? Date: Thu, 14 Feb 2002 15:16:25 +0100 From: "news" Organization: XS4ALL Internet BV Message-ID: I need the VxWorks Munch application (or similar functions) to run under Linux on a PC - any ideas on how or where to get such a thing? Any help or pointers appreciated. BR;s Sten --------------------------- Newsgroups: comp.os.vxworks Subject: serial port settings Date: Thu, 14 Feb 2002 09:30:21 -0500 From: "Suresh Kannan" Organization: Georgia Institute of Technology Message-ID: Hi, I am trying to read data from and industrial helicopter computer whose serial settings ar 9600 Baud, 8E1. How in vxworks do I set the 8 bit, Even parit and 1 stop bit parameters. I tried looking at the ioctl FIOxxx calls but could not find a corresponding FIOxxx flag. I do not seem to be able to do the standard posix thing i.e., get the termios structure and manipulate that. It seems vxworks does not support that at all. regards, suresh. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: vxWorks BSP developments with Galileo GT64260 Date: Thu, 14 Feb 2002 08:23:36 -0600 From: JH Organization: Rockwell International Message-ID: <3C6BC868.F80FE901@collins.rockwell.com> References: <1f1b48b5.0202131823.2c3e9702@posting.google.com> This is an attention grabber. Our hardware folks are designing a new in house board that is using the same chips, IBM 750cxe and GT64260. The group I'm in is going to be responsible for porting the in house OS to this new board when it's finally here. I'm curious about the performance results your seeing. Can you give a little more detail about it?? I must admit we are just starting to comb through the user manuals etc for the porting effort. Maybe there is a pitfall that we could avoid with this combination. John john c wrote: > > Anyone working with the Galileo GT64260 Dual PCI bridge Controller? > > Just curious. We're using IBM 750 CXe (600MHz PowerPC) interfacing to > Galileo GT 64260 dual PCI bridge controller. Our vendor did a really > crappy job porting vxWorks BSP for our needs, and we ended up > developing most of the drivers and debugging ourselves. I'm just > wondering if anyone else out there is trying to get their GT64260 to > work properly that I can share ideas with. > > We measured SDRAM performance yesterday with the 750 and the Galileo. > Basically, the 60x bus of the 750 doesn't seem to handle SDRAM bursts > very well. the result is very limited bandwidth for the 750 to read > and write to the SDRAM that is connected through the Galileo chip. > YIKES. If you are building your system, try to use a 74xx series > controller that takes advantage of the MPX bus.... :-) > > anyhow. just wondering if I'm alone in fighting with the GT64260. > > smiles > john --------------------------- Newsgroups: comp.os.vxworks Subject: Re: MPC860 floating point emulation Date: Mon, 11 Feb 2002 23:48:43 -0800 From: Ernie Rael Message-ID: <71ih6u092rjmtk11gccujljce0vhd98v4m@4ax.com> References: I'll have to review the license, I thought that the library license would not require opening our source. Any changes to the library would have to be open. But that's not really an issue here anyway. I would only be referring to the source for the API; I can write wrappers, in asm or C, to hook to the subroutines. It's not really the exception handler glue, but rather decoding all the fp instructions and hooking them to the floating lib that I was trying to avoid; you're right, its not that big of a deal. I'll work hard to be lazy and I like to re-use when possible. If I'm going from scratch, I guess I'll set up a taskVar to point to the floating point registers. We use longlong so they have to be there for every task. Hmmm, I'll have to check... I don't remember if I can call malloc from the taskCreateHook... Thanks for the feedback, - -ernie "Dave Korn" wrote: >"Ernie Rael" wrote in message >news:acl66uojgmoch330fm6aiqbiotrddnnqal@4ax.com... >> We use both MPC860 and PPC750. I'd like to compile only for >> PPC750, and when running on MPC860 handle the software >> emulation exception to implement floating point. There isn't >> much floating point; performance isn't much of an issue. >> >> Does anyone know of a package that handles this. I can do >> any glue if it doesn't hook directly into a vxworks exception. >> I haven't looked at the gnu source for the soft-float stuff, but >> I'm guessing that it would take a bit of work to adapt it. > > Presumably the gnu soft-float code is just in the form of subroutines that >take arguments, so if you're as happy as you say to write the exception >handling glue, why isn't it the answer you're looking for? > > Bear in mind however that if you use gnu *source* in your work, you are >obliged to open your source to any customers who you distribute your >binaries containing the work-derivative-from-gnu-code to. That's probably >far more of an issue than adapting the code. I don't have a recommendation >about what code you can use, but I thought I'd better drop in the >oops-that's-gnu-SOURCE-you're-referring-to-there problem. > > DaveK --------------------------- Newsgroups: comp.os.vxworks Subject: Help on SDRAM performance consideration Date: 11 Feb 2002 23:54:45 -0800 From: ryeung@earthlink.net (Raymond Yeung) Organization: http://groups.google.com/ Message-ID: "Higher clock speed translates to faster SDRAM access time". Based on some reading I've done so far, this statement is probably not true. There're first access latency, pipelining, bursting. But if I want to determine whether a particular SDRAM module is fast enough to match an upgrade to a faster processor (and bus speed), what are the key parameters I should look for in the SDRAM spec. sheet? How would you decide that you need a even faster SDRAM module than what you're using now? A pointer to a website that sheds some light on this topic would also be welcome. THanks! --------------------------- Newsgroups: comp.os.vxworks Subject: Re: stack USB from WinRiver on All BSP? Date: 12 Feb 2002 08:31:16 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: <2002211-18554-313711@foorum.com> Charly wrote: > Are there somebody that are using USB stack from WindRiver. > Is-it easy to configure and install? > Does it use BSP capabilities or direct h/w registers to access USB's > h/w ? > > My BSP builder doesn't know anything on USB, Do you thing I could > access directly to USB component even if nothing is written for that > in the BSP. Some general info on this issue. You always have to adapt the generic driver to your specific HW. Depending on the controller you are using a lot may already have been done, but there will always be some modification, for example other memory locations, connecting interrupts, etc. So it depends on how standard your HW is, and how good the chipset is supported. Groeten, Johan - -- o o o o o o o . . . _____________________________ o _____ || Johan Borkhuis | .][__n_n_|DD[ ====_____ | borkhuis@agere.com | >(________|__|_[_________]_|__________________________| _/oo OOOOO oo` ooo ooo 'o!o!o o!o!o` === VxWorks FAQ: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html === --------------------------- Newsgroups: comp.os.vxworks Subject: Re: serial port settings Date: Thu, 14 Feb 2002 14:53:38 -0000 From: "Martin Dowie" Message-ID: <3c6bcf66$1@pull.gecm.com> References: > I am trying to read data from and industrial helicopter computer whose > serial settings ar 9600 Baud, 8E1. How in vxworks do I set the 8 bit, Even > parit and 1 stop bit parameters. I tried looking at the ioctl FIOxxx calls > but could not find a corresponding FIOxxx flag. > > I do not seem to be able to do the standard posix thing i.e., get the > termios structure and manipulate that. It seems vxworks does not support > that at all. Try: #include "vxWorks.h" #include "ioLib.h" ... // name = "tyCo/0" | "tyCo/1" | "tyCo/2" | "tyCo/3" // flags = O_RDONLY | O_WRONLY | O_RDWR // mode = 2 | 4 | 6 (actually not used in our device) // file = open( name, mode, mode ); int ioctlValue; // baudRate = 9600 | 19200 | 38400 // ioctlValue = ioctl( file, SIO_BAUD_SET, baudRate ); // modemControl = (CLOCAL | CREAD) /* read */ | // CLOCAL /* write */ // (CLOCAL | CREAD) /* read & write */ // // parity = PARODD /* odd */ | PARENB /* even */ | 0 /* no parity */ // // characterSize = CS7 | CS8 // ioctlValue = ioctl( file, SIO_HW_OPTS_SET, modemControl | parity | characterSize ); ioctlValue = ioctl( file, FIOSETOPTIONS, OPT_TANDEM | OPT_MON_TRAP ); ioctlValue = ioctl( file, FIOFLUSH, 0 ); Hope this helps! Martin --------------------------- Newsgroups: comp.os.vxworks Subject: Re: +++ C-Designer (sehr flexibel) für Realtime/C/Systeme sucht neues Projekt +++ Date: Tue, 12 Feb 2002 10:20:44 +0100 From: "Michael Lawnick" Organization: Buergernetz Dillingen Message-ID: References: <3c5e5151@brateggebdc5.br-automation.co.at> <3c6832b7@brateggebdc5.br-automation.co.at> Reply-To: "Michael Lawnick" Sender: mlawnick@p5088f16b.dip.t-dialin.net Hallo Werner, es ist sicher nicht sehr zweckmäßig hier nicht in Englisch zu schreiben, aber einen Flame deswegen loszubrechen bringt's auch nicht ! Hi Werner, I doesn't seem to be much of purpose to post here in English, but to put on a flame because of that doesn't seem to meet ether. So besser ? - -- Mit freundlichen Grüßen, Michael Lawnick ============================================== SOFTEC GmbH Tel +49-731-96600-0 Promenade 17 Fax +49-731-96600-23 D-89073 Ulm Michael Lawnick Germany lawnick@softec.de ============================================== "Werner Schiendl" schrieb im Newsbeitrag news:3c6832b7@brateggebdc5.br-automation.co.at... > Dave, > > "Dave Korn" wrote in message > news:PGT98.2307$5o.893553@newsr2.u-net.net... > > > "M.Thoemel" wrote in message > > > news:a3f403$bi0$04$1@news.t-online.com... > > > > Guten Tag, > > > > "Werner Schiendl" wrote in message > > news:3c5e5151@brateggebdc5.br-automation.co.at... > > > Hi, > > > > > > If you use this list for your advertising, _AT_LEAST_ use the correct > > > language: > > > This is an english language list! > > > > > > - Werner > > > > It is? If you're referring to the newgroup, I can't see anything that > > specifies what language should be used: > > > -- snip -- > > that might be true, indeed. > > Anyway: How many posts in this newsgroup - apart from spam - are in an other > language?? > None (or almost, since sometimes a newbie may post in some other language). > > Specified or not: effectively this -is- an english language list / > newsgroup. > > best regards > Werner > > P. S. I hate spam, including Martin R's and this one. _Particularly_ if it > is repeated every week. > > > --------------begin quote-------------- > > Control: newgroup comp.os.vxworks > > Newsgroups: comp.os.vxworks.ctl > > Path: rpi!tale > > From: tale@rpi.edu (David C Lawrence) > > Subject: newgroup comp.os.vxworks > > Message-ID: > > Sender: tale@cs.rpi.edu > > Nntp-Posting-Host: cs.rpi.edu > > Date: 13 Jan 92 16:43:31 GMT > > Approved: tale@rpi.edu > > Lines: 16 > > > > comp.os.vxworks is an unmoderated newsgroup which passed its vote for > > creation by 295:17 as reported in news.announce.newgroups on 6 Jan 1992. > > > > For your newsgroups file: > > comp.os.vxworks The VxWorks real-time operating system. > > > > The charter, culled from the call for votes: > > > > comp.os.vxworks will be a forum for discussion of topics related to > > the VxWorks real-time operating system and development environment. > > Current and prospective users of VxWorks will be able to discuss > > technical problems, report experiences with various hardware components, > > share code, report bugs and fixes, and exchange any information of > interest > > to > > the VxWorks user community. Relevant new product announcements may be > > made by commercial participants, subject to a limit - e.g. 1/month - > > on the frequency of such posts. > > ---------------end quote--------------- > > > > Whereas if you're referring to the exploder, well, it isn't Matthias' > > fault if the exploder went and copied a post from the newsgroup that it > > shouldn't have. > > > > DaveK > > -- > > moderator of > > alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow > > Burn your ID card! http://www.optional-identity.org.uk/ > > Help support the campaign, copy this into your .sig! > > Proud Member of the Exclusive "I have been plonked by Davee because he > > thinks I'm interesting" List Member # > > Master of Many Meowing Minions > > Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage > above > > and beyond the call of hilarity. > > > > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: MPC860 interrupt latency variation Date: Tue, 12 Feb 2002 17:17:12 +0100 From: Christophe Couturier Organization: Guest of France Telecom Oleane's newsreading service Message-ID: <3C694008.DF0FB22D@st2e.com> Hi all, My application is synchronised by an external signal causing an interrupt. Althought I don't care if the interrupt latency is long, I need the variation of this latency to be very short (~+/-5ľs). Does anyone knows how to configure VxWorks and the PowerPc to reach this goal. I use VxWorks 5.4 and a MPC860. Thanks. Chrsitophe --------------------------- Newsgroups: comp.os.vxworks Subject: USB on the PCPentium BSP? Date: 12 Feb 2002 12:12:39 GMT From: Charly Message-ID: <2002212-131239-482717@foorum.com> Hello vxworks users, I wonder if it is possible to use the WindRiver USB Stack on PCPentium BSP. Does anyone use it? What is the development cost to : install and configure the USB stack. To write and tesy an easy request on it. Thanks for your experiences. charly - -- Use our news server 'news.foorum.com' from anywhere. More details at: http://nnrpinfo.go.foorum.com/ --------------------------- Newsgroups: comp.os.vxworks Subject: Question:Capturing stack trace of a taskwhen it crashes Date: 14 Feb 2002 07:19:52 -0800 From: sunil@dharwad.com (Sunil K) Organization: http://groups.google.com/ Message-ID: <396b9cc7.0202140719.7109204c@posting.google.com> Hi VxWorkers, I need to capture stack trace in a file when any task crashes. (Even if the crash leads to a reboot of vxWorks) Any ideas how to do it ? Should I replcae the default interupt handlers that suspend the task when it does any illegal operation ? How should I go about it ? Your help is greatly appreciated. Thanks! Sunil --------------------------- Newsgroups: comp.os.vxworks Subject: disassember in vxWorks 5.4? Date: Tue, 12 Feb 2002 09:04:56 -0500 From: Michael Reed Message-ID: <3C692108.41057703@ll.mit.edu> Hello, We're making a shopping list for an upcoming project... does anyone know if vxWOrks 5.4 has a disassember bundled w/ it or the GNU tools? THanks, Mike --------------------------- Newsgroups: comp.os.vxworks Subject: Guru Question: IP stack and Mblks Date: Tue, 12 Feb 2002 14:05:40 GMT From: "Jim" Organization: Magma Communications Ltd. Message-ID: Hello, I a driver for a highly customized asic that DMAs data into a set of buffers/clusters ... The data is of the following format. MBLK--->CBLK--->buffer( [HEADER][IP PACKET][CRC] ) I want to but this in the IP stack with zero copy. So I will do the following: mPacketLen -= length(HEADER) mLen -= (length(HEADER) + length(CRC)) Then I will insert it into the IP Queue. Do I need room in the Mblk chain for the ethernet header and crc? Or does vxworks add that? Should I use a seperate pool from the vxWorks IP stack buffer and mBlk pool? Help please! Thnx and cheers, Jim --------------------------- Newsgroups: comp.os.vxworks,comp.protocols.snmp Subject: how to define view mask for mib tree Date: 12 Feb 2002 06:55:58 -0800 From: ilyasg@yahoo.com (Ilyas G) Organization: http://groups.google.com/ Message-ID: Hi all, i have an snmp agent, that i have created, it contains MIB2. the target is vx simulator, when i try to access some variable an error occurs. when i had debugged snmpd, i have noticed that when i call snmp_view_family_check , the access view is rejected with the error VIEW_EXCLUDED. which don't surprise me a lot, because i don't know how to convigure it. even if it works with older version of vxsim . question : How do we configure the view access on the agents ? thanks --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Reading from serial ports - Blocking or interrupts? Date: Thu, 14 Feb 2002 17:47:45 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <37449c9d.0202130624.43ffc949@posting.google.com> <3c6b9334$1@brateggebdc5.br-automation.co.at> I have done a very high performance serial driver for 68302 a few years back, and the scheme I used was - 8 byte buffer per channel, and 2-character idle time would trigger an interrupt, or buffer full interrupt. This worked great for an application which used packets on top of async serial channels, and turned out that inside a packet there was no idle time whatsoever, and a 2-character delay was not a problem either. "Werner Schiendl" wrote in message news:3c6b9334$1@brateggebdc5.br-automation.co.at... > Hi, > > see comments inside. > > wrote in message > news:37449c9d.0202130624.43ffc949@posting.google.com... > > I am working on an application using VxWorks 5.4 to read from a number > > of serial ports simultaneously. > > > > Each serial port is read by its own separate VxWorks task. > > > > The serial ports receive pre-defined, but variable-length (over some > > fairly close limits) messages. > > > > Each serial port comes with a 128-byte buffer, and the receive > > interrupt can be set to fire on any number of characters in the buffer > > from 1 to 128. It can also generate an interrupt if 4 character times > > pass with no input, with the FIFO only partially filled (i.e. not > > filled enough to reach its interrupt level). > > > > Timing for processing incoming messages is critical, i.e. each task > > must process its received message as soon as practical. > > > > My question is, which of the following scenarios is likely to be the > > most efficient in terms of use of processor resources, while still > > meeting the objective of processing incoming messages as quickly as > > possible: > > > > 1. Set the interrupt to interrupt on the length of the shortest > > received message. The interrupt sends a signal to the input task, > > which then reads and stores the data received so far, then goes back > > to sleep. The tail end of the message is handled by the 4-character > > timeout-with-no-data interrupt. > > > > In my opinion, this is probably the most efficient model. > If you can live with 4-character times delay, of course. > > > 2. Generate an interrupt on each character, which sends a signal to > > the input task which processes incoming characters on a > > character-by-character basis. > > > > This will produce a _very_ high load on your CPU. > Keep in mind, that the 16550 FIFOs in the PC's serial ports have been > introduced to allow for fast transfer modes (57600, 115200) on 486 (maybe > 386) machines. You have a 'number' of serial ports, so you may end up being > slower that option #1 just because processing each character seperately > wastes a lot of time for context switches. The costs of a context switch > depends on your hardware, but even Pentium processors spend some usec to > enter and leave an interrupt. > > > 3. Use a blocking read function to read from the serial port. No > > interrupts are needed, and incoming data is handled on a > > character-by-character basis. > > > > How should a blocking read function work? It will use an interrupt handler > internally, to unblock the task on the arrival of a new character. > So, if I understand what you wanna try correct, this is basically the same > then option #2. > > 4) An other option you have is to use a polling scheme. This means, your > UARTs do net generate an interrupt when new data arrives. Instead you poll > them in a resolution suitable for your system to find out if new data has > arrived. If handling these UARTs is the 'function' of your device, this may > be an appropriate design - otherwise it may waste to much processor time for > the polling. > > 5) Write a special driver that allows polling in addition to the normal, > interrupt driver transfer. If you are able to determine the size of an > unfinished packet, you'll know when there is not a full 128 buffer left to > read and read the rest in polled mode. Or - if the UART allows changing the > interrupt level 'on the fly' - change it to single byte interrupts until the > packet is complete, then back to the big buffer. > > > > It seems to me that #3 is likely to be the most efficient, assuming > > the serial port board's BSP supports it (we don't have the BSP > > documentation yet so I can't be sure but I'd be very surprised if it > > didn't). > > For option #4 you need a driver that can do 'polled' mode. > WRS drivers support this (usually) for system mode debugging with > CrossWinds. > See the documentation on how to use polled mode serial drivers. > > Option #5 will most probably need you coding your own driver. > > > > > But, that raises another question ... > > > > In general, what is the overhead for unblocking read functions, > > compared with the overhead of processing an interrupt? I'm concerned > > that there might be some goofy internal hidden stuff that goes on in > > scenario #3 such that #2 is actually faster and more efficnent. > > > > It is the same, since unblocking always requires an interrupt > (theoretically, it could be unblocked using a polling task, but this is > probably not a common design for a serial driver - this would require a > speperate task the checks the hardware on a cyclical basis). > > > Also, when a task is blocked, i.e. on a read of a serial port, does it > > ALWAYS relinquish the CPU for use by lower-priority tasks, and then > > take over the CPU again when data appears on the serial port? > > > > Yes, if - of course - the driver has been properly coded. > > hth > Werner > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: login issue Date: 12 Feb 2002 07:08:08 -0800 From: macpete@gmx.de (Marc Peter) Organization: http://groups.google.com/ Message-ID: <24315948.0202120708.4b49f61a@posting.google.com> References: <26ca8a8f.0201300030.47a0fa90@posting.google.com> <89792565.0201310059.487be7b2@posting.google.com> <26ca8a8f.0202070423.46ad20c2@posting.google.com> senthilb@comneti.com (senthil) wrote in message news:<26ca8a8f.0202070423.46ad20c2@posting.google.com>... > [...] Then we tried declaring the variables Loginusername > and password as global ones. After we did this the loginUserVerify( ) > validated as we expected. > Can any one interpret why this behaviour occurs? loginLib only stores a pointer to the encrypted password you provide to loginUserAdd. That means it has to stay there for as long as the user is known to loginLib. This explains why your local variables couldn't work. They were simply overwritten with bogus stuff by subsequent function calls. There is one nice thing about this behaviour, though. You can change the password in place without having to allocate new space for it. I hope this clarifies things up a bit Marc --------------------------- Newsgroups: comp.os.vxworks Subject: CodeTEST/Memory or PerformancePak (MemScope) ? Date: Tue, 12 Feb 2002 17:02:59 +0200 From: "Pavius" Organization: Internet Gold, ISRAEL Message-ID: Hi. Sorry if this is a bit off topic, but could any seasoned vets give recommendations as to which of the above is better (CodeTEST/Memory / Memscope) ? --------------------------- Newsgroups: comp.os.vxworks Subject: adding new function to be called from prjConfig.c Date: Tue, 12 Feb 2002 18:28:03 +0200 From: "Ofer Goren" Organization: Verio Message-ID: - -- - ------------------------ Win2k Sp2 Tornado 2.02 for PPC Diab 4.4b Envoy 9.2 - ------------------------ Hi. How do I add a function to be called from orjConfig.c? This is a auto-generated file, so whenever I rebuild my project, my changes are gone.... Thanks, - -- Ofer Goren Iamba Technologies --------------------------- Newsgroups: comp.os.vxworks Subject: winmodem support Date: Tue, 12 Feb 2002 17:01:48 GMT From: "Peter Marqui" Organization: Verio Message-ID: <0Uba8.932$A%3.15288@ord-read.news.verio.net> I know this is a stupid question (I'm reluctant to even ask) Does anyone know if vxWorks supports any winmodem's ? I realize that winmodem's are nothing more than a (DAA) data access arrangement, and a codec, to perform ADC/DC or does any one know of a cheap internal controller based modem ?. 3com (formly U.S. Robotics), still makes controller based PCI, and ISA based modems priced at $80 (56K* V.92 Performance Pro Modem )and $170 (Courier V.everything)respectively thanks --------------------------- Newsgroups: comp.os.vxworks Subject: What's vxImmrGet() function for? Date: Thu, 14 Feb 2002 09:36:33 -0700 From: "Liang Yang" Organization: RTS-LAB Message-ID: Reply-To: "Liang Yang" Hi, When I need to set the ISR for my vxWorks program, I always have to invoke vxImmrGet() first, what's the vxImmrGet() for? Thank you. Liang --------------------------- Newsgroups: comp.os.vxworks Subject: Re: [OT] ULIP driver ... Date: Thu, 14 Feb 2002 16:12:17 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: "Kim Letkeman" wrote in message news:CTFa8.947$Vba.893@news2.bloor.is... > Dave, have you spent any time with XP? I've been running it for over a year > as a beta tester (did the same for Win2k, so I have extensive experience > with both.) There is no comparison. Win2k is old news in almost every way. > Just my opinion of course. Nope, I haven't used XP, but my PC already has an OS that boots and works, so why would I? ISTM that you have conflated "newer" with "better", or at any rate your comment that "2k is old news" suggests that you didn't really have any actual positive *reason* for going to XP. So apart from "not running the software that you vitally need", what other advantages does XP give *you* personally over 2k in terms of the way you use your pc from day-to-day ? Pardon my manner of speaking; I am genuinely curious to know why you feel something that was fine yesterday is suddenly no use at all today. I just can't make any sense of your suggestion that there is no comparison: they're almost exactly the same kernel, almost exactly the same architecture; most of what has changed is window dressing and high-level application support apis, but of course none of the tools I use from day-to-day are going to suddenly start making any use of any new features. Your suggestion that 2k is not an OS "..newer than 6 years old.." must be because you regard it as a variant of NT; well then, XP isn't any newer by that standard. For me, my PC is the tool of my trade, so to me "better" == "does the job I require of it either more efficiently or more correctly", not "newer", not "nicer graphics in the gui", not "has flashy instant messaging service builtin", and particularly not "won't actually be any use for months until the driver I want gets rewritten to work with NDIS5.0". I thought about it, and couldn't see the slightest purpose in breaking everything for no advantages that I actually had any use for. I guess YMMV :) depending on what else you use your computer for. > driver. But busted is busted ..... I'd bet someone at WindRiver could get > this sucker to work in a day or less .... I suspect it'd take a little more than that, assuming that the ULIP works as a filter over the serial device, then rewriting an NDIS3 or NDIS4 filter driver to be NDIS5 is at least a modest amount of work. DaveK - -- moderator of alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! Proud Member of the Exclusive "I have been plonked by Davee because he thinks I'm interesting" List Member # Master of Many Meowing Minions Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage above and beyond the call of hilarity. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Reading from serial ports - Blocking or interrupts? Date: Thu, 14 Feb 2002 16:29:57 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: <1IRa8.2442$5o.1075855@newsr2.u-net.net> References: <37449c9d.0202130624.43ffc949@posting.google.com> wrote in message news:37449c9d.0202130624.43ffc949@posting.google.com... [SNIP!] > Each serial port is read by its own separate VxWorks task. > > The serial ports receive pre-defined, but variable-length (over some > fairly close limits) messages. > > Each serial port comes with a 128-byte buffer, and the receive > interrupt can be set to fire on any number of characters in the buffer > from 1 to 128. It can also generate an interrupt if 4 character times > pass with no input, with the FIFO only partially filled (i.e. not > filled enough to reach its interrupt level). > > Timing for processing incoming messages is critical, i.e. each task > must process its received message as soon as practical. > > My question is, which of the following scenarios is likely to be the > most efficient in terms of use of processor resources, while still > meeting the objective of processing incoming messages as quickly as > possible: [SNIP!] The answer to this question somewhat depends on the answer to this question: Can you determine how long the packet is going to be by inspecting the first one or two chars ? DaveK - -- moderator of alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! Proud Member of the Exclusive "I have been plonked by Davee because he thinks I'm interesting" List Member # Master of Many Meowing Minions Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage above and beyond the call of hilarity. --------------------------- Newsgroups: comp.os.vxworks Subject: LSI53C1000 driver? Date: Tue, 12 Feb 2002 17:16:06 -0000 From: "Employee" Organization: Newport Networks Ltd. Message-ID: <1013534167.152994@azores.network-i.net> Hi all, is anyone writing a VxWorks driver for LSI53C1000? I am trying to upgrade the sym895 driver to work with it, but I am having problems handling phase mismatches. Regards. Dimitris --------------------------- Newsgroups: comp.os.vxworks Subject: Guru Question # 2: netBufLib and concurency Date: Tue, 12 Feb 2002 17:34:44 GMT From: "Jim" Organization: Magma Communications Ltd. Message-ID: References: <1013534167.152994@azores.network-i.net> Hello, If two threads are using a pool allocated by netBufLib in vxWorks ... is concurent access provided or do I have to do that ... I really hope it is! Help please ;-) Cheers, Jim "Employee" wrote in message news:1013534167.152994@azores.network-i.net... > Hi all, > > is anyone writing a VxWorks driver for LSI53C1000? > > I am trying to upgrade the sym895 driver to work with it, but I am having > problems handling phase mismatches. > > Regards. > > Dimitris > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: USB on the PCPentium BSP? Date: 12 Feb 2002 14:56:12 GMT From: Charly Message-ID: <2002212-155612-337225@foorum.com> References: <2002212-131239-482717@foorum.com> <1013519562.44534@azores.network-i.net> I am sorry for that. I didn't want to post it 2 times. But this question was different than yesterday. Charly - -- Use our news server 'news.foorum.com' from anywhere. More details at: http://nnrpinfo.go.foorum.com/ --------------------------- Newsgroups: comp.os.vxworks Subject: Failover TCP/IP stack for vxworks Date: 12 Feb 2002 10:21:29 -0800 From: anmol@eternal-systems.com (Anmol Mishra) Organization: http://groups.google.com/ Message-ID: Hi, there are various FT technique for TCP stacks, based on Virtual IP, or taking over the IP address of a failed node. However I am looking for a stack that, not just takes over the IP address of a failed node, but also keeps existing socket connection alive and valid. Thus the client should see a completely transparent failover. Compaq has a product for their Himalaya servers that does this, Parallel TCP/IP. However I am looking at something similar in the embedded world. Thus, if client c connects to server s1 on a socket. s1 fails and s2 takes over from s1, then the socket connection should still be alive.. Existing methods of Virtual IP, IP address takeover, NAT based failover DO NOT keep existing sockets alive. You need to reset the existing connection from the client side and connect again.... Any pointer, etc to similar research projects will be appreciated.. --------------------------- Newsgroups: comp.os.vxworks,gnu.gcc.help Subject: prjObjs.lst - error 0x80 Date: Thu, 14 Feb 2002 19:07:39 +0200 From: "Ofer Goren" Organization: Verio Message-ID: - ------------------------ Win2k Sp2 Tornado 2.02 for PPC GNU compiler Envoy 9.2 - ------------------------ Hi. I have a strange problem, when compiling a project that worked well a day ago. When I compile, I get an make: *** /prjObjs.lst Error 0x80 I tried creating new project, under a new worksapce, with no files, the works. Nothing helped. No makefile under ../target/h/make/ was modified during the last 6 months... Anyway, any help, before re-installing Tornado (and all other utilities after) ... :(( Thanks, Ofer - -- Ofer Goren Iamba Technologies --------------------------- Newsgroups: comp.os.vxworks Subject: Guru Question # 2: netBufLib and concurency Date: Tue, 12 Feb 2002 18:48:58 GMT From: "Jim" Organization: Magma Communications Ltd. Message-ID: Hello, If two threads are using a pool allocated by netBufLib in vxWorks ... is concurent access provided or do I have to do that ... I really hope it is! Help please ;-) Cheers, Jim "Employee" wrote in message news:1013534167.152994@azores.network-i.net... > Hi all, > > is anyone writing a VxWorks driver for LSI53C1000? > > I am trying to upgrade the sym895 driver to work with it, but I am having > problems handling phase mismatches. > > Regards. > > Dimitris > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Flash Drivers ->TFFS & DosFs2.0 Date: 12 Feb 2002 11:09:41 -0800 From: vloscomp@yahoo.com (Vinh) Organization: http://groups.google.com/ Message-ID: Hi, I am trying to write a flash driver for the TFFS and DosFs2.0. Does any have a clear documentation for the MTD, Socket interfaces to TFFS? (Not the VxWorks delivered craps... I need a bit detail description of interfaces to TFFS) Vinh Lam. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: [Help Me] Socket binding error Date: Tue, 12 Feb 2002 19:31:59 GMT From: talberts Organization: The Boeing Company Message-ID: <3C696DAF.C23B4C0@please.nospam> References: <3c581122$1@brateggebdc5.br-automation.co.at> Sender: nntp@news.boeing.com (Boeing NNTP News Access) Also, via TCP/IP protocol rules, there is a timeout value for TCP sockets. I think the default in VxWorks is 60 seconds, so you should not be able to reuse the exact same socket for 60 seconds after closing it by default. I think you can get around this rule but I don't remember off the top of my head. Werner Schiendl wrote: > > Hi, > > Do you close the socket before you unload your code? > You cannot have to sockets bind to the same port > > hth > Werner > > "Kim, Jeong-Hwan" wrote in message > news:EDJ58.9680$Of.192302@news.bora.net... > > Hi, everyone > > > > When socket server task is loaded again after it is unloaded ( not board > > reset ), > > socket binding routine occurs error indicating "address in use". > > So I set socket option value "SO_REUSEADDR" but it also has same error. > > Please let me know how I can solve this problem ?? > > > > > > My socket server code is as follows : > > > > > > > > ................ > > > > m_hSocket = socket(AF_INET, SOCK_STREAM, 0); > > > > if ( m_hSocket == INVALID_SOCKET ) > > { > > error_print("(%s, %d) DM Server socket initialization failure", > > file_name, __LINE__); > > perror("[DM SERVER : socket]"); > > return(1); > > } > > else > > printf("[DM SERVER] SOCKET created\n"); > > > > optval = 1; > > setsockopt(m_hSocket, SOL_SOCKET, SO_REUSEADDR, (char *)&optval, > > sizeof(int)); > > > > server_address.sin_family = AF_INET; > > server_address.sin_addr.s_addr = htonl(INADDR_ANY); /* > > inet_addr("127.0.0.1"); */ > > server_address.sin_port = htons(tcp_port_number); > > > > > > err = bind(m_hSocket, (struct sockaddr *)&server_address, > > sizeof(server_address)); > > > > if ( err ) > > { > > error_print("(%s, %d) DM Server socket bind failure", file_name, > > __LINE__); > > perror("[DM SERVER : bind]"); > > close(m_hSocket); > > return(1); > > } > > > > > > > > > > Thanks in advance > > > > > > Kim > > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: problem with -mlongcall ( with one more interesting point) Date: Thu, 14 Feb 2002 16:39:19 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: <307af50b.0201290317.689ad29e@posting.google.com> <9JB58.38$5o.67861@newsr2.u-net.net> <307af50b.0201300435.637b546@posting.google.com> <307af50b.0201310153.36844789@posting.google.com> "pavan" wrote in message news:307af50b.0201310153.36844789@posting.google.com... > hi, Hi again Pavan; sorry for taking a while to reply. > i am posting again for the same problem with more inputs now. > One more interesting ( / strange ?? ) problem is that, if you comment > out the destructor for class A then Machine check exception doesn't come. > > The machine check occurs only in the second iteration of the while loop with > destructor for class A. Pavan, when you download your application to the shell, does it claim that __builtin_delete is missing ? Also, from your other post: about the flags: > + /home/tornado/host/sun4-solaris2/bin/ccppc -mlongcall -DCPU=PPC604 > -Wall > -I.. -I/home/tornado/target/h/arch/ppc > -I/home/tornado/target/h > -I/home/tornado/host/sun4-solaris2/lib/gcc-lib/powerpc-wrs-vxworks/cygnus- 2.7.2-960126/include > -c ../test.cpp > could you please let me know if you had any more flags in your > compilation. Yes, quite a few, some of which may be fairly important, many of which don't matter: ccppc -BF:\Tornado/host/x86-win32/lib/gcc-lib/ -mlongcall -memb -fsigned-ch ar -Wconversion -Wmissing-prototypes -mstrict-align -ansi -nostdinc -DRW_MUL TI_THREAD -D_REENTRANT -O2 -fno-builtin -fno-for-scope -I/h -I. -IF:\To rnado\target\config\all -IF:\Tornado\target/h -IF:\Tornado\target/src/config -IF:\Tornado\target/src/drv -DCPU=PPC604 -DD3 -DNEW_CPU -DRUN_ON_750 - Wnested-externs -Wmissing-declarations -Wreturn-type -Wnested-externs -Wimpl icit -Wstrict-prototypes -Wparentheses -fenum-int-equiv -Wchar-subscripts - DH050REVISION_B -mcpu=604 -c testcase.cpp If I were you I would *definitely* start using all the flags between "-memb" and "-fno-builtin" in that list, although for the -O2 you may want to change the optimization level as appropriate. In particular, if you haven't used -mstrict-align, things could easily go badly wrong, I think... DaveK - -- moderator of alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! Proud Member of the Exclusive "I have been plonked by Davee because he thinks I'm interesting" List Member # Master of Many Meowing Minions Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage above and beyond the call of hilarity. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: [HELP] network connection with Solaris Date: Thu, 14 Feb 2002 16:52:22 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: <3C696BE5.D057C4BF@please.nospam> "talberts" wrote in message news:3C696BE5.D057C4BF@please.nospam... > chihwankim wrote: > > > > I have a target with MPC8260 CPU and vxWorks 5.4. > > > > There is little bit funny problem between target and Solaris workstation. > > I assume that Solaris workstation's IP address is 'A', and target 'B'. > > > > The 'Ping' from target to Solaris is ok. > > >ping B > > > > But 'Ping' from Solaris to target is not ok. Through 'icmpstatShow', I > > verified that target 'B' received all the icmp packets and made replies for > > them. But Solaris displayed 'no answer from B'. And 'telnet' to target 'B' > > won't work. > > > > Of course there is no problem between target 'B' and window 2000 based > > PC/Linux/another target. I mean 'ping' and 'telnet' work well. > > > > What's the wrong? Has anybody have this kind of problem? > > > > chkim. > > If your target and host are on the same subnet, make sure your netmasks > are identical on both machines. Also check that the ARP caches on both machines have the correct contents. There may be a problem going on there. DaveK - -- moderator of alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! Proud Member of the Exclusive "I have been plonked by Davee because he thinks I'm interesting" List Member # Master of Many Meowing Minions Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage above and beyond the call of hilarity. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: MPC860 floating point emulation Date: Thu, 14 Feb 2002 17:25:25 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: <71ih6u092rjmtk11gccujljce0vhd98v4m@4ax.com> "Ernie Rael" wrote in message news:71ih6u092rjmtk11gccujljce0vhd98v4m@4ax.com... > I'll have to review the license, I thought that the library > license would not require opening our source. Any changes > to the library would have to be open. But that's not > really an issue here anyway. I would only be referring to > the source for the API; I can write wrappers, in asm or C, > to hook to the subroutines. D'oh, of course: you'll just link them with the library rather than incorporate or modify the source. Ignore my previous comment, since doing *that* is fine under either GPL *or* LGPL. DaveK - -- moderator of alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! Proud Member of the Exclusive "I have been plonked by Davee because he thinks I'm interesting" List Member # Master of Many Meowing Minions Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage above and beyond the call of hilarity. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: 64-bit operations Date: 12 Feb 2002 12:55:36 -0800 From: Joe Chung Message-ID: References: > Dear anyone, > I am trying to implement 64-bit operations in my VxWorks operations (using > the GNU compilier, on PowerPC platform). > I am trying to multiply and divide unsigned long long int variables. > However, I always get 0 as value. > I acnt even do a printf("%llu\n", 64bit_var) > > Can anyone help me??? > > > Regards Nejhdeh Ghevondian My setup: T2.0.2, gcc 2.7.2, MCP750. long long works fine (as long as certain functions from the host-side libgcc.a is linked into my app), but printf %llu does not work. Too bad, because it's rather trivial to implement. Anyway, I wrote my own longlong_to_string conversion function in order to print 64 bit ints, but again - all "long long" operations worked flawlessly. - -jc - -- (apply 'concat (reverse (list "com" (char-to-string 46) "yahoo" (char-to-string 64) "joechung"))) --------------------------- Newsgroups: comp.os.vxworks Subject: Re: [HELP] network connection with Solaris Date: Tue, 12 Feb 2002 19:22:52 GMT From: talberts Organization: The Boeing Company Message-ID: <3C696B8C.72E1DC01@please.nospam> References: Sender: nntp@news.boeing.com (Boeing NNTP News Access) If both target and host are on the same network, make sure your netmasks are identical. chihwankim wrote: > > I have a target with MPC8260 CPU and vxWorks 5.4. > > There is little bit funny problem between target and Solaris workstation. > I assume that Solaris workstation's IP address is 'A', and target 'B'. > > The 'Ping' from target to Solaris is ok. > >ping B > > But 'Ping' from Solaris to target is not ok. Through 'icmpstatShow', I > verified that target 'B' received all the icmp packets and made replies for > them. But Solaris displayed 'no answer from B'. And 'telnet' to target 'B' > won't work. > > Of course there is no problem between target 'B' and window 2000 based > PC/Linux/another target. I mean 'ping' and 'telnet' work well. > > What's the wrong? Has anybody have this kind of problem? > > chkim. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: [HELP] network connection with Solaris Date: Tue, 12 Feb 2002 19:24:21 GMT From: talberts Organization: The Boeing Company Message-ID: <3C696BE5.D057C4BF@please.nospam> References: Sender: nntp@news.boeing.com (Boeing NNTP News Access) If your target and host are on the same subnet, make sure your netmasks are identical on both machines. chihwankim wrote: > > I have a target with MPC8260 CPU and vxWorks 5.4. > > There is little bit funny problem between target and Solaris workstation. > I assume that Solaris workstation's IP address is 'A', and target 'B'. > > The 'Ping' from target to Solaris is ok. > >ping B > > But 'Ping' from Solaris to target is not ok. Through 'icmpstatShow', I > verified that target 'B' received all the icmp packets and made replies for > them. But Solaris displayed 'no answer from B'. And 'telnet' to target 'B' > won't work. > > Of course there is no problem between target 'B' and window 2000 based > PC/Linux/another target. I mean 'ping' and 'telnet' work well. > > What's the wrong? Has anybody have this kind of problem? > > chkim. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Help on SDRAM performance consideration Date: 12 Feb 2002 13:08:11 -0800 From: Joe Chung Message-ID: References: > "Higher clock speed translates to faster SDRAM access time". > Based on some reading I've done so far, this statement is probably not true. > There're first access latency, pipelining, bursting. The S in SDRAM means that data (after an initial latency) is able to burst out at the rate of bus speed, synchronized on the bus clock edge(s). > But if I want to determine whether a particular SDRAM module is fast > enough to match an upgrade to a faster processor (and bus speed), what > are the key parameters I should look for in the SDRAM spec. sheet? How > would you decide that you need a even faster SDRAM module than what you're > using now? It's not the CPU speed, but the BUS SPEED that you have to worry about. If your bus speed is, say 133MHz, then you'll need an SDRAM rated at 133MHz or faster. Using faster SDRAM than your bus speed will gain you nothing, and using slower SDRAM than your bus speed will just not work, most of the time. > A pointer to a website that sheds some light on this topic would also > be welcome. THanks! Try http://www.tomshardware.com/mainboard/98q4/981024/index.html - -jc - -- (apply 'concat (reverse (list "com" (char-to-string 46) "yahoo" (char-to-string 64) "joechung"))) --------------------------- Newsgroups: comp.os.vxworks Subject: June 2001 v/s T2CP4 Date: Tue, 12 Feb 2002 15:28:42 -0600 From: Dinker Charak Organization: Fermi national Accelerator Lab Message-ID: Hello, Does someone know which came first? T2CP4 says it was released on June 2001. So is it that all driver updates and bug-fixes in Jume 2001 driver pack release are part of T2CP4? Thanks, Dinker --------------------------- Newsgroups: comp.os.vxworks Subject: VirtuallAlloc failed error on compilation Date: 14 Feb 2002 10:11:16 -0800 From: isro@yahoo.com (isro) Organization: http://groups.google.com/ Message-ID: <89441eba.0202141011.1e4c5bd@posting.google.com> Hi, I am using Tornado compiler 2.1.1 for XSCALE 2.9-010413 Any idea, why i am getting the following error on compilation commit_and_inc: VirtualAlloc failed The file i am copiling is cpp file with size 126K. My computer has about 256K RAM. Thanks in advance isro --------------------------- Newsgroups: comp.os.vxworks Subject: Re: What's vxImmrGet() function for? Date: Thu, 14 Feb 2002 18:22:41 +0000 From: Philip Macfarlane Organization: RAMiX Europe Ltd Message-ID: References: On Thu, 14 Feb 2002 16:36:33 +0000, Liang Yang wrote: > When I need to set the ISR for my vxWorks program, I always have to > invoke vxImmrGet() first, what's the vxImmrGet() for? If I remember correctly - the IMMR is the Internal Memory Map Register. The PowerPC has internal memory for controlling peripherals, which can be mapped into several different places, under control of pullup/down pins. The software needs to know where this is mapped so it can control peripherals etc... vxImmrGet reads the location that was set in hardware. Or something like that. It's been a while. Cheers, Philip. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Anyone tried to modify 8260 interrupt priority? Date: 14 Feb 2002 10:30:45 -0800 From: ryeung@earthlink.net (Raymond Yeung) Organization: http://groups.google.com/ Message-ID: References: ryeung@earthlink.net (Raymond Yeung) wrote in message news:... > I need to modify Timer2's interrupt priority such that > auxiliary clock used by SPY has the highest priority on > the processor. This would allow me to measure ISR CPU > utilization as well. > > I first thought all I need to do is to program SICR with > the interrupt number of Timer2. However this doesn't work. > It seems VxWorks interrupt controller masks out Timer2 > interrupts in other high speed device (e.g. FCC) ISRs. > > I'd need to modify the hardcoded interrupt masks, such > that Timer2 interrups are no longer masked by interrupts > I want to measure. Then regenerate the library archive > and rebuild my vxWorks image. > > Does this sound right? This procedure has been confirmed. In particular, I need to modify the masks for FCC1 and FCC2 (the 2 high runner we're using) to NOT mask out TIMER2 interrupt, and to have TIMER2 mask out other interrupts (it's the highest priority interrupt). All these masks are implemented in an 8260 specific interrupt controller file under ..\src\drv directory. I don't need to regenerate library. Looks like sysLib.c includes the interrupt controller .c file. So I just need to regenerate the VxWorks image. --------------------------- Newsgroups: comp.os.vxworks,gnu.gcc.help Subject: Re: prjObjs.lst - error 0x80 Date: Thu, 14 Feb 2002 18:22:41 -0000 From: "Martin Dowie" Message-ID: <3c6c0065@pull.gecm.com> References: > I have a strange problem, when compiling a project that worked well a day > ago. > When I compile, I get an make: *** /prjObjs.lst Error 0x80 > > I tried creating new project, under a new worksapce, with no files, the > works. Nothing helped. > > No makefile under ../target/h/make/ was modified during the last 6 months... > > Anyway, any help, before re-installing Tornado (and all other utilities > after) ... :(( I had the same problem and reported it to Wind River (ref 268815). They had never seen this before, couldn't reproduce it. Every suggestion they came up with didn't work. In the end they said to re-install everything. As we out-source our IT, this hasn't happened yet! :-( If you do re-install tornado and it "cures" this problem, could you let me know? Cheers, Martin --------------------------- Newsgroups: comp.os.vxworks Subject: Tornado II Prototyper installation failure Date: 12 Feb 2002 14:17:39 -0800 From: google@drforehead.net (Nick Forrette) Organization: http://groups.google.com/ Message-ID: <576881c0.0202121417.e1df9b0@posting.google.com> First off, I'm very new to VxWorks and have been assigned to learn a bit about it. As part of my exploration, I've tried to install the Tornado II Prototyper under Windows 2000. I am having two problems with the installer. 1. The Tornado II Prototyper tar file they ship includes two files named "inatr:intf". Of course colons are illegal characters for filenames in Windows. I worked around this by extracting everything but those two offending files. 2. When running the installer, everything goes smoothly till it actually starts to copy file files. The progress bar jumps immediately to 100% done and hangs there. After being patient (waiting 30 minutes+ just to be sure), I canceled the install. Maybe problem #2 is a result of the files I couldn't extract because of #1. Any thoughts? TIA, Nick --------------------------- Newsgroups: comp.os.vxworks Subject: Re: disassember in vxWorks 5.4? Date: 12 Feb 2002 14:39:02 -0800 From: pkockritz@mac.com (Pete Kockritz) Organization: http://groups.google.com/ Message-ID: <9423c55b.0202121439.cb81d2@posting.google.com> References: <3C692108.41057703@ll.mit.edu> Michael Reed wrote in message news:<3C692108.41057703@ll.mit.edu>... > Hello, > We're making a shopping list for an upcoming project... does anyone know > if vxWOrks 5.4 has a disassember bundled w/ it or the GNU tools? Yes: 1) the target shell and windSh have a command (l - lowercase L) that can disassemble memory. Just give it an address and it will show you the assy beginning at that address (but no source interleaved). 2) the CrossWind debugger can show source and assy interleaved. 3) objdump (GNU binutils) can disassemble an object file and optionally interleave the source with it. On Windows hosts, there's a problem with munging of the source filename as stored in the object file, so objdump can't find it. But there are ways around that. Pete Kockritz - -- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: What's vxImmrGet() function for? Date: Thu, 14 Feb 2002 11:46:31 -0700 From: "Liang Yang" Organization: RTS-LAB Message-ID: References: Reply-To: "Liang Yang" Thank you for your explanation. But why should I always need to call this routine before I set my ISR. What's relationship between the Internal Memory Map Register and my own ISR? "Philip Macfarlane" wrote in message news:u6o03hef65ht25@corp.supernews.com... > On Thu, 14 Feb 2002 16:36:33 +0000, Liang Yang wrote: > > > When I need to set the ISR for my vxWorks program, I always have to > > invoke vxImmrGet() first, what's the vxImmrGet() for? > > If I remember correctly - the IMMR is the Internal Memory Map Register. > The PowerPC has internal memory for controlling peripherals, which can be > mapped into several different places, under control of pullup/down pins. > > The software needs to know where this is mapped so it can control > peripherals etc... vxImmrGet reads the location that was set in hardware. > > Or something like that. It's been a while. > > Cheers, > Philip. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: 64-bit operations Date: 12 Feb 2002 15:03:25 -0800 From: manbert@rahul.net (Rush Manbert) Organization: http://groups.google.com/ Message-ID: References: "Nejhdeh Ghevondian" wrote in message news:... > Dear anyone, > I am trying to implement 64-bit operations in my VxWorks operations (using > the GNU compilier, on PowerPC platform). > I am trying to multiply and divide unsigned long long int variables. > However, I always get 0 as value. > I acnt even do a printf("%llu\n", 64bit_var) > > Can anyone help me??? > > > Regards Nejhdeh Ghevondian I was recently doing the same on the same platform. I'm not sure about multiply/divide, except that a divide of two long long unsigned ints generates a reference to udivdi3. Someone here told me that we had to get a replacement for libgcc.a from Wind River before this would work. As for printing, forget it. There is no conversion for a 64 bit integer that I have been able to find. You need to display it as two 32 bit numbers, 0x%08x%08x works for hex, but you must isolate the two halves of the word. I was using a right shift of 32 bits for the high half, but the compiler seems to generate bad shift code for that (the assembler complains about the shift count being out of the range 0-31) so I switched to dividing by the unsigned 64-bit integer constant 0x100000000ULL, which worked. (But I can divide two 64 bit integers without any problems.) Good luck. - - Rush --------------------------- Newsgroups: comp.os.vxworks Subject: Re: vxWorks BSP developments with Galileo GT64260 Date: 14 Feb 2002 10:48:56 -0800 From: jchen@tollbridgetech.com (john c) Organization: http://groups.google.com/ Message-ID: <1f1b48b5.0202141048.5a83c3c0@posting.google.com> References: <1f1b48b5.0202131823.2c3e9702@posting.google.com> <3C6BC868.F80FE901@collins.rockwell.com> Well, I'm not too sure how much of this I can disclose before I get in trouble with Galileo, but I'll tell you what we did and what our results were. We're using Micron SDRAM Chips, currently running 100MHz on the CPU bus with CAS, SRAS, and SRAS to SCAS Delay set to 0x2 (3 cycles). Our test script is written in Assembly and it basically copies 4MB of data into SDRAM. We have timers (in 100MHz ticks) that give us the starting tick and ending tick count, and we measured the total performance. In our test, each "write" to SDRAM from CPU seems to take about 80ns, for a 64-bit double word. That gives us 100MB/s throughput to the SDRAM. We potted up one of our boards and looked at the signals, and it seems like there are gaps between each transaction. Also, we weren't able to get the CPU to write to the SDRAM through Galileo in 64-bit chunks. It does 2 transactions of 32-bits to each of our banks. The other thing is that we can't seem to get the CPU to "burst" to the SDRAM. We're stuck doing 64-bit reads and writes. We can't seem to give SDRAM the starting address and just burst data into SDRAM. Nothing else is running on our system, but this seems to be a limitation of the 60x bus. Do you know much about the 60x bus vs. MPX bus (the bus between Galileo and the CPU)? I'm not too famaliar with the concepts of pipe-lining, either (still learning).... Right now, we're trying to change the CAS latency to 2 on the memory banks. We tried it with the 750 emulator and got another 23% improvement. But I'm having trouble with the script (the system freezes). We'll see how things go. In theory, we should be able to get 800MB/s throughput to the SDRAM, right? Assuming 8 bytes * 100MHz = 800MB/s?!?!? Any thoughts? john JH wrote in message news:<3C6BC868.F80FE901@collins.rockwell.com>... > This is an attention grabber. Our hardware folks are designing a new in > house board that is using the same chips, IBM 750cxe and GT64260. The > group I'm in is going to be responsible for porting the in house OS to > this new board when it's finally here. > > I'm curious about the performance results your seeing. Can you give a > little more detail about it?? I must admit we are just starting to comb > through the user manuals etc for the porting effort. Maybe there is a > pitfall that we could avoid with this combination. > > John > > john c wrote: > > > > Anyone working with the Galileo GT64260 Dual PCI bridge Controller? > > > > Just curious. We're using IBM 750 CXe (600MHz PowerPC) interfacing to > > Galileo GT 64260 dual PCI bridge controller. Our vendor did a really > > crappy job porting vxWorks BSP for our needs, and we ended up > > developing most of the drivers and debugging ourselves. I'm just > > wondering if anyone else out there is trying to get their GT64260 to > > work properly that I can share ideas with. > > > > We measured SDRAM performance yesterday with the 750 and the Galileo. > > Basically, the 60x bus of the 750 doesn't seem to handle SDRAM bursts > > very well. the result is very limited bandwidth for the 750 to read > > and write to the SDRAM that is connected through the Galileo chip. > > YIKES. If you are building your system, try to use a 74xx series > > controller that takes advantage of the MPX bus.... :-) > > > > anyhow. just wondering if I'm alone in fighting with the GT64260. > > > > smiles > > john --------------------------- Newsgroups: comp.os.vxworks Subject: Re: What's vxImmrGet() function for? Date: Thu, 14 Feb 2002 18:53:30 +0000 From: Philip Macfarlane Organization: RAMiX Europe Ltd Message-ID: References: On Thu, 14 Feb 2002 18:46:31 +0000, Liang Yang wrote: > But why should I always need to call this routine before I set my ISR. > What's relationship between the Internal Memory Map Register and my own > ISR? Calling it on it's own (and disregarding the value returned) doesn't do anything, as far as I know. The function returns the value of the IMMR, so you can subsequently use it to access some peripheral. e.g. volatile unsigned long *immr = vxGetImmr(); immr[foo] = bar; /* Yuck - Put the foo peripheral into bar mode. */ (Not a brilliant example, but hopefully useful.) Cheers, Philip. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Munch for Linux? Date: Thu, 14 Feb 2002 18:51:41 +0000 (UTC) From: David Laight Organization: BT Openworld Message-ID: <3C6C078F.4010807@l8s.co.uk> References: news wrote: > I need the VxWorks Munch application (or similar functions) to run under > Linux on a PC - any ideas on how or where to get such a thing? Any help or > pointers appreciated. Well, the munch in T2.0.1 (solaris for arm) is a Tcl script..... David --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Questions about vxmalloc.c Date: 12 Feb 2002 15:14:20 -0800 From: manbert@rahul.net (Rush Manbert) Organization: http://groups.google.com/ Message-ID: References: Bill Pringlemeir wrote in message news:... > >>>>> "Rush" == Rush Manbert writes: > > Rush> I have gotten a copy of vxmalloc.c, and I was trying to figure > Rush> out the changes relative to Doug Lea's original, so I could > Rush> apply the same changes to the latest version. I came across a > Rush> number of macros that were redefined in vxmalloc.c, and I'm > Rush> trying to understand why the changes were made. These include > Rush> MALLOC_ALIGNMENT, chunk2mem, and mem2chunk. > > Rush> So my question is, has anyone used this code and analyzed why > Rush> these changes exist, and/or does anyone know a current email > Rush> address for Hwa-Jin Bae, the author? > > Don't use the vxMalloc.c on the FAQ sight. > > * VERSION 2.7.0 Sun Mar 11 14:14:06 2001 Doug Lea (dl at gee) > ftp://gee.cs.oswego.edu/pub/misc/malloc.c > > This is a more recent version with nice little fixes for people > wanting to use Doug's allocator on vxWorks. It is also not clouded > with the BSD code that `Hwa-Jin' included. There is a provision to > use a custom header file. I ended up modifying the source. I have > include pertinent info at the end of this message. I have removed the > stupid inter-twining of the vxWorks memory libraries. Ie, memPartLib > contains only partition functions. valloc, calloc, etc are in memLib > and they have been replaced by `vxMalloc.c'. You might have to code > another vxSbrk if you don't do this. > > The info about MALLOC_ALIGNMENT, chunk2mem, mem2chunk is provided in > Doug Lea's web page "http://g.oswego.edu/dl/html/malloc.html". I > doubt that I could provide a more insightful comment. > > Times, they are a changin' > Bill Pringlemeir. > > [start changes] > > #define MORECORE vxSbrk > #define MORECORE_CONTIGUOUS 0 > #define MORECORE_CANNOT_TRIM 1 > #define malloc_getpagesize (64*1024) > > struct malloc_state > { > /* The maximum chunk size to be eligible for fastbin */ > [snip...] > /* vxWorks semaphore. */ > SEMAPHORE mem_semaphore; > }; > > #define MALLOC_PREACTION semTake(&(get_malloc_state()->mem_semaphore), \ > WAIT_FOREVER) > #define MALLOC_POSTACTION semGive(&((get_malloc_state()->mem_semaphore))) > > void vxMemInit(void) > { > semBInit(&(get_malloc_state()->mem_semaphore), SEM_Q_PRIORITY, SEM_FULL); > } > > void *vxSbrk(int size) > { > void *tmp; > static void *previous_top; > > if(size) > { > > if(!(tmp = memPartAlloc(memSysPartId, size+MALLOC_ALIGNMENT))) > return (void*)-1; > > /* Must be aligned. */ > tmp = (void*)(((int)tmp + MALLOC_ALIGN_MASK) & ~MALLOC_ALIGN_MASK); > previous_top = (void*)((int)tmp + size + MALLOC_ALIGNMENT); > return tmp; > } > > return previous_top; > } > > [end changes] Thanks. This confirms what I suspected - that the alignment stuff did not need to be changed in order to work for VxWorks. There was a lot of "cleanup" done in vxmalloc that makes it hard to port the changes forward into newer versions. And thanks for the code snippet. I already had a copy of 2.7.0, and I see how what you've done fits in. - - Rush --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Questions about vxmalloc.c Date: 12 Feb 2002 15:15:44 -0800 From: manbert@rahul.net (Rush Manbert) Organization: http://groups.google.com/ Message-ID: References: "Dave Korn" wrote in message news:... > "Rush Manbert" wrote in message > news:a294ea3e.0202081842.42362789@posting.google.com... > > I have gotten a copy of vxmalloc.c, and I was trying to figure out the > > changes relative to Doug Lea's original, so I could apply the same > > changes to the latest version. I came across a number of macros that > > were redefined in vxmalloc.c, and I'm trying to understand why the > > changes were made. These include MALLOC_ALIGNMENT, chunk2mem, and > > mem2chunk. > > > > So my question is, has anyone used this code and analyzed why these > > changes exist, and/or does anyone know a current email address for > > Hwa-Jin Bae, the author? > > > > Thanks, > > Rush > > Sounds like he wanted to change the memory alignment quantum to me. Wanna > post the macros in before-and-after versions ? > I think the answer is that they didn't need to be changed. (See previous response). Thanks. - - Rush --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tool for checking memory leak Date: 14 Feb 2002 11:20:23 -0800 From: mark_ck_chang@yahoo.com (Mark Chang) Organization: http://groups.google.com/ Message-ID: <9b6f8883.0202141120.42e1ae04@posting.google.com> References: <3C6B932E.C080423B@oslo.online.no> Glenn Ivar Klausen wrote in message news:<3C6B932E.C080423B@oslo.online.no>... > Is there a tool simular to purifi or boundchecker in vxWorks? I have been using RTI's Memscope available from WindRiver. It will display a 'real-time' listing of functions that call malloc, free and the size of memory. --------------------------- Newsgroups: comp.os.vxworks Subject: S 626 PCI card Date: Wed, 13 Feb 2002 00:22:07 GMT From: "Mika" Organization: [Infostrada] Message-ID: Hi everybody! I'm having many problems with a PCI card. It is a Sensoray 626. Does anybody know it? Unluckly I'm a "beginner" in VxWorks facilities so I need some help to understood how I can install this card in my VxWorks target. I used INCLUDE_PCI and INCLUDE_PCI_PARAMS (from Tornado) to scan the PCI bus but I don't know how to use these information. How can I translate the BAR0 (the only one base address)into a CPU address? Isn't PCI configuration space a part of MMU? How can I use offsets specified on user's manual? (no word are written about the address they start from...it is reasonably BAR0, right?). I mean, once the card is correcly installed, can I read/write card registers using pointers for example (or functions like pciConfigOutWord/Long ecc)? I know, probably my questions are too much generic! I few words I need a quick guide from the plug to a correct memory-mapped card. Thank you for your attention...you could make a student happy! Michael. --------------------------- Newsgroups: comp.os.vxworks Subject: S 626 PCI card Date: Wed, 13 Feb 2002 01:01:50 GMT From: "Mika" Organization: [Infostrada] Message-ID: <2Wia8.9576$Jo.284693@twister2.libero.it> I've forgotten just one thingh in my previous message: I work with a x86 target. Bye! --------------------------- Newsgroups: comp.os.vxworks Subject: ULIP driver ... Date: Wed, 13 Feb 2002 04:06:09 GMT From: "Kim Letkeman" Message-ID: A few posts over the years have alluded to people running the ULIP driver for vxsim under Windows 2000. Were those all just for PPC architecture? Was there ever a ULIP driver for Windows 2000 x86? I ask because I am unable to run ULIP at home (XP) and I'd like to be able to test there. I _hate_ the thought of dual booting just for this purpose. Is there anything available? Why in the world would WindRiver not release the ULIP source code into the public domain if they refuse to support it on OSs newer than 6 years old? Kim --------------------------- Newsgroups: comp.os.vxworks Subject: Re: CodeTEST/Memory or PerformancePak (MemScope) ? Date: Tue, 12 Feb 2002 15:21:42 -0600 From: "J. Min Kim" Organization: Nortel Message-ID: References: I have only used the MemScope. I find it very easy to use. It searches for 'new', 'delete', 'malloc', 'free' operations. "Pavius" wrote in message news:a4bar4$fug$1@news.inter.net.il... > Hi. > > Sorry if this is a bit off topic, but could any seasoned vets give > recommendations as to which of the above is better (CodeTEST/Memory / > Memscope) ? > > --------------------------- Newsgroups: comp.os.vxworks Subject: pap and secret files Date: Thu, 14 Feb 2002 20:25:27 GMT From: "alain tamdem" Organization: Rauland Borg Corporation Message-ID: Reply-To: "alain tamdem" Hi Iam trying to use the secret file to specify in me ppp options files how to ge the secret files, but its still not working... can somebody tell me how to use pppSecretAdd() , I mean the prototype... --------------------------- Newsgroups: comp.os.vxworks Subject: Re: CodeTEST/Memory or PerformancePak (MemScope) ? Date: Thu, 14 Feb 2002 12:42:08 -0800 From: Lori Fraleigh Organization: Real-Time Innovations Message-ID: <3C6C2120.D559AF1A@rti.com> References: I may be a little biased in my opinion of which tool is better, :-), but here are some facts: CodeTEST Memory requires instrumentation which can slow down your code. It only lets you track memory leaks in code you can recompile. RTI's MemScope requires no instrumentation. It patches the memory allocation routines so that it can intercept all memory allocation calls regardless of where they originate: your own code, third-party libraries, or the operating system. If you have other questions regarding MemScope, I would be happy to answer them! Regards, Lori Pavius wrote: > > Hi. > > Sorry if this is a bit off topic, but could any seasoned vets give > recommendations as to which of the above is better (CodeTEST/Memory / > Memscope) ? - -- Lori Fraleigh Director of Engineering Real-Time Innovations, Inc. http://www.rti.com Phone: 408-734-4200 x121 Fax: 408-734-5009 --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Wind Web Server - problem in httpSymTabAdd Date: Thu, 14 Feb 2002 15:46:45 -0500 From: "Jing Ma" Message-ID: <3c6c2236.0@alognews.analogic.com> References: <3c67dbde@news.012.net.il> what if that function has been changed to > for(i = 1; i < 9; i++) > { > sprintf(buf,"AutomaticTAOS%d", i); > > symId = httpSymTabAdd (httpGlobalSymTabId, > variablename[i], > "", > HTTP_SYM_TYPE_BOOL_ID, > 1, > &T1_interface[i-1].BitsClockEnable); > } "Yaron Cohen" wrote in message news:3c67dbde@news.012.net.il... > Hi, > I am getting "symbol xxx not found! SSI failed " (8 times for each error) > after > > trying to add symbols to httpSymTabAdd() in the follwing way: > > ----------------------------------------------------------- > char buf[50]; > > for(i = 1; i < 9; i++) > { > sprintf(buf,"AutomaticTAOS%d", i); > > symId = httpSymTabAdd (httpGlobalSymTabId, > buf, > "", > HTTP_SYM_TYPE_BOOL_ID, > 1, > &T1_interface[i-1].BitsClockEnable); > } > > ---------------------------------------------------------------- > > When I am trying to do it without using 'buf' it works fine (without the > loop): > > ----------------------------------------------------------- > symId = httpSymTabAdd (httpGlobalSymTabId, > AutomaticTAOS1, > "", > HTTP_SYM_TYPE_BOOL_ID, > 1, > &T1_interface[0].BitsClockEnable); > > symId = httpSymTabAdd (httpGlobalSymTabId, > AutomaticTAOS2, > "", > HTTP_SYM_TYPE_BOOL_ID, > 1, > &T1_interface[1].BitsClockEnable); > > ------- > . > . > . > | > > symId = httpSymTabAdd (httpGlobalSymTabId, > AutomaticTAOS8, > "", > HTTP_SYM_TYPE_BOOL_ID, > 1, > &T1_interface[7].BitsClockEnable); > --------------------------------------------------- > > Any idea ? > > Thanks, > > Yaron Cohen > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Munch for Linux? Date: Thu, 14 Feb 2002 22:08:45 +0100 From: "xs4all" Organization: XS4ALL Internet BV Message-ID: References: <3C6C078F.4010807@l8s.co.uk> Thanks for the hint David - only problem is that I don't have Tornado 2.0.1 so if anyone happens to have the script and accidentally hit the send mail button I would be very grateful. sten@xs4all.nl "David Laight" wrote in message news:3C6C078F.4010807@l8s.co.uk... > news wrote: > > > I need the VxWorks Munch application (or similar functions) to run under > > Linux on a PC - any ideas on how or where to get such a thing? Any help or > > pointers appreciated. > > > Well, the munch in T2.0.1 (solaris for arm) is a Tcl script..... > > David > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: CodeTEST/Memory or PerformancePak (MemScope) ? Date: 14 Feb 2002 16:11:38 -0500 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: References: <3C6C2120.D559AF1A@rti.com> Sender: bpringlemeir@DeadDuck >>>>> "Lori" == Lori Fraleigh writes: Lori> I may be a little biased in my opinion of which tool is better, Lori> :-), but here are some facts: Lori> CodeTEST Memory requires instrumentation which can slow down Lori> your code. It only lets you track memory leaks in code you can Lori> recompile. Lori> RTI's MemScope requires no instrumentation. It patches the Lori> memory allocation routines so that it can intercept all memory Lori> allocation calls regardless of where they originate: your own Lori> code, third-party libraries, or the operating system. [snip] My apologies, I thought MemScope was a WindRiver product. Maybe that is a compliment... well sort of! Having used instrumented code, I can attest that it is a pain. However, as I noted in another post, the MemScope product requires that the image be in RAM pages that are writable. This is a smaller problem that requiring source IMHO. I will give the OP this link, "http://www.mozilla.org/performance/tools.html" These are good for people on a budget. I have thought that CodeTEST used veneers or something to track all allocation... perhaps that was the code coverage tool. Anyways, I hope that people don't get too uppity and chastise you for say "Hey, look what I make." Regards, Bill Pringlemeir. ps, I don't think Martin is really a spammer either. Titles like "Cute Asian girls", "Make money Fa$t" and "Hack your Gameboy in five easy steps" are spam. - -- Kibo could tamper with the Pope. vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Project-relative files? Date: Fri, 15 Feb 2002 09:12:48 +1030 From: Tim Shaw Organization: Defence Science and Technology Organisation Message-ID: <3C6C3D68.5E14B2EB@dsto.defence.gov.au> References: We tried to do that about 12 months ago, using T2 on Solaris. But our local support person said it wasn't possible. Bob Bradley wrote: > Is it possible to convince Tornado to reference files by relative path > that are above the project file in the hierarchy when added to the > project? It seems to use project-relative paths when the files are below > it in the hierarchy, but uses full paths for files above it. For > example, I might have a hierarchy as follows: > > Sources > Make > MyProject.wpj > Misc > MyFile.c > > When I add "MyFile.c" to "MyProject.wpj", I'd like it to be saved in the > project file (and the resulting makefile) as > "$(PRJ_DIR)../Misc/MyFile.c" instead of using the full path to MyFile.c. > Our project is built on a variety of systems so using a full path > doesn't work. > > Is this possible? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado II Prototyper installation failure Date: 14 Feb 2002 15:59:44 -0800 From: google@drforehead.net (Nick Forrette) Organization: http://groups.google.com/ Message-ID: <576881c0.0202141559.6a9ac550@posting.google.com> References: <576881c0.0202121417.e1df9b0@posting.google.com> <3c6b8c17$1@brateggebdc5.br-automation.co.at> "Werner Schiendl" wrote... > it seems there is something broken with your downloaded file. That wouldn't surprise me. > I have downloaded the Prototyper .tar file a couple days ago and it contains > the following files: > > Disk_id 80 Bytes > Wind.000 2,535,146 Bytes > Wind.001 5,727,500 ... > Wind.002 4,945,701 > Wind.003 5,905,079 > Wind.004 4,670,421 > Wind.005 5,555,023 > Wind.006 5,246,791 > Wind.007 4,014,090 > Wind.008 4,332,725 > Wind.009 4,915,579 > > The archive file downloaded is 47,858,688 bytes in size. I don't have the archive handy, but IIRC it was on the order of 60-70MB, included all those files plus a whole lot more (e.g. setup.exe). > Maybe something went wrong while downloading the file. > If your archive does not contain the above files, try re-downloading the > archive. I'll try it again. Maybe they put a botched version of the tar file on the web. > I used WinZip 8.0 to view/extract the archive. Ditto. Thanks, Nick --------------------------- Newsgroups: comp.os.vxworks Subject: User defined events not triggering windview? Date: Fri, 15 Feb 2002 01:09:36 GMT From: "Chad" Organization: Excite@Home - The Leader in Broadband http://home.com/faster Message-ID: I can't seem to get windview to trigger from a user defined event (trgEvent(MYEVENTID)); It will trigger based upon variable conditions etc however. Is there something that must be done aside from defining and downloading triggers then calling trgEvent()? I tried the helloString example in the Triggering section of the WindView User's guide but it did not work. --------------------------- Newsgroups: comp.os.vxworks Subject: What happens beneath the 860SAR while setting PDPAR.UT bit ? Date: 14 Feb 2002 17:14:07 -0800 From: 10month@hanmail.net (Changwon Lee) Organization: http://groups.google.com/ Message-ID: Hi... I'm trying to activating UTOPIA interface in 860SAR without any PHY chip. I want to just check signals(SOC, TxCav, TxEnb, UTPD[0-8]) using scope. But when I set up PDPAR.UT bit, target fails. When I set up PDPAR and PDDIR as follows, myIMMR->pio_pdpar |= 0xdf7f; myIMMR->pio_pddir &= 0x2080; What's happening inside the 860SAR. For example, Does 860SAR reads port D after PDPAR.UT bit is set? Thanks for any help... ^^ --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VirtuallAlloc failed error on compilation Date: Fri, 15 Feb 2002 05:40:03 GMT From: "drdiags" Organization: AT&T Broadband Message-ID: References: <89441eba.0202141011.1e4c5bd@posting.google.com> Hello Isro, I would contact Wind River support. if(HOST == WINDOWS_HOST) { I hear there are a couple of ways to get around this problem, which is based on the fact that the compiler needs to increase it's virtual memory allocation to a larger scratchpad size. You might mention the keyword "pepatch.exe" which is a way to increase the size of the virtual memory chunk that was compiled in the cplus1.exe (?) and another file, but I don't have the information in front of me. Check Wind Surf to see if a compiler patch for this problem has been released. There is also a built-in limit of how big you can make virtual size increase, so other options are to use the -fno-inline or change the optimization. And there was another -f flag, but if you find out all the correct details, could you post so the next fellow will have the real story instead of my ramblings. NOTE: You do mean 256MB of ram on your Host, right? } else { printf ("NEVER MIND\n"); } "isro" wrote in message news:89441eba.0202141011.1e4c5bd@posting.google.com... > Hi, > I am using Tornado compiler 2.1.1 for XSCALE > 2.9-010413 > > Any idea, why i am getting the following error on compilation > > commit_and_inc: VirtualAlloc failed > > The file i am copiling is cpp file with size 126K. My computer has about 256K RAM. > Thanks in advance > isro --------------------------- Newsgroups: comp.os.vxworks Subject: Re: What's vxImmrGet() function for? Date: Fri, 15 Feb 2002 07:43:48 GMT From: "Gary M" Organization: AT&T Broadband Message-ID: References: If you delete the variable definition and its assignment to vxImmrGet(), and the compiler doesn't complain, then you don't need it. If it does complain, then its use is probably buried in a macro inside the ISR. "Philip Macfarlane" wrote in message news:u6o1tb4q7bc33e@corp.supernews.com... > On Thu, 14 Feb 2002 18:46:31 +0000, Liang Yang wrote: > > > But why should I always need to call this routine before I set my ISR. > > What's relationship between the Internal Memory Map Register and my own > > ISR? > > Calling it on it's own (and disregarding the value returned) doesn't do > anything, as far as I know. The function returns the value of the IMMR, so > you can subsequently use it to access some peripheral. e.g. > > volatile unsigned long *immr = vxGetImmr(); > immr[foo] = bar; /* Yuck - Put the foo peripheral into bar mode. */ > > (Not a brilliant example, but hopefully useful.) > > Cheers, > Philip. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tool for checking memory leak Date: Fri, 15 Feb 2002 08:40:50 +0000 From: Danny Chapman Organization: Lucent Technologies, Columbus, Ohio Message-ID: References: <3C6B932E.C080423B@oslo.online.no> <9b6f8883.0202141120.42e1ae04@posting.google.com> On 14 Feb 2002, Mark Chang wrote: > Glenn Ivar Klausen wrote in message news:<3C6B932E.C080423B@oslo.online.no>... > > Is there a tool simular to purifi or boundchecker in vxWorks? > > I have been using RTI's Memscope available > from WindRiver. It will display a 'real-time' listing of > functions that call malloc, free and the size of memory. Unfortunately memscope won't tell you anything about corruption. The only way I know of is to write your code so that it also builds and runs under (for example) Solaris, where you can run purify. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado II Prototyper installation failure Date: Fri, 15 Feb 2002 09:53:11 +0100 From: "Werner Schiendl" Message-ID: <3c6ccc7d$1@brateggebdc5.br-automation.co.at> References: <576881c0.0202121417.e1df9b0@posting.google.com> <3c6b8c17$1@brateggebdc5.br-automation.co.at> <576881c0.0202141559.6a9ac550@posting.google.com> Hi Nick, I downloaded the file on February 7, 2002. At that time, the prototyper download did not include the setup.exe. Instead, the setup tool was a separate download. Did they merge it in the meantime into one download? Maybe, an error was introduced in that process? Johan Borkhuis posted a link to download the setup tool, if you do not have it anyway - search the NG - I don't have it handy now. hth Werner "Nick Forrette" wrote in message news:576881c0.0202141559.6a9ac550@posting.google.com... > "Werner Schiendl" wrote... > > it seems there is something broken with your downloaded file. > > That wouldn't surprise me. > > > I have downloaded the Prototyper .tar file a couple days ago and it contains > > the following files: > > > > Disk_id 80 Bytes > > Wind.000 2,535,146 Bytes > > Wind.001 5,727,500 ... > > Wind.002 4,945,701 > > Wind.003 5,905,079 > > Wind.004 4,670,421 > > Wind.005 5,555,023 > > Wind.006 5,246,791 > > Wind.007 4,014,090 > > Wind.008 4,332,725 > > Wind.009 4,915,579 > > > > The archive file downloaded is 47,858,688 bytes in size. > > I don't have the archive handy, but IIRC it was on the order of > 60-70MB, included all those files plus a whole lot more (e.g. > setup.exe). > > > Maybe something went wrong while downloading the file. > > If your archive does not contain the above files, try re-downloading the > > archive. > > I'll try it again. Maybe they put a botched version of the tar file on > the web. > > > I used WinZip 8.0 to view/extract the archive. > > Ditto. > > Thanks, > Nick --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Wind Web Server - problem in httpSymTabAdd Date: Fri, 15 Feb 2002 10:00:19 +0100 From: "Werner Schiendl" Message-ID: <3c6cce1f@brateggebdc5.br-automation.co.at> References: <3c67dbde@news.012.net.il> Hi, just a shot in the dark, because I don't have Wind Web Server at hand. But are you sure, that httpSymTabAdd() does copy the given variable name?? There are a lot of occasions, where VxWorks only stores a POINTER to what you hand it. Obviously, the buffer would then be overwritten in each interation of your loop. hth Werner "Yaron Cohen" wrote in message news:3c67dbde@news.012.net.il... > Hi, > I am getting "symbol xxx not found! SSI failed " (8 times for each error) > after > > trying to add symbols to httpSymTabAdd() in the follwing way: > > ----------------------------------------------------------- > char buf[50]; > > for(i = 1; i < 9; i++) > { > sprintf(buf,"AutomaticTAOS%d", i); > > symId = httpSymTabAdd (httpGlobalSymTabId, > buf, > "", > HTTP_SYM_TYPE_BOOL_ID, > 1, > &T1_interface[i-1].BitsClockEnable); > } > > ---------------------------------------------------------------- > > When I am trying to do it without using 'buf' it works fine (without the > loop): > > ----------------------------------------------------------- > symId = httpSymTabAdd (httpGlobalSymTabId, > AutomaticTAOS1, > "", > HTTP_SYM_TYPE_BOOL_ID, > 1, > &T1_interface[0].BitsClockEnable); > > symId = httpSymTabAdd (httpGlobalSymTabId, > AutomaticTAOS2, > "", > HTTP_SYM_TYPE_BOOL_ID, > 1, > &T1_interface[1].BitsClockEnable); > > ------- > . > . > . > | > > symId = httpSymTabAdd (httpGlobalSymTabId, > AutomaticTAOS8, > "", > HTTP_SYM_TYPE_BOOL_ID, > 1, > &T1_interface[7].BitsClockEnable); > --------------------------------------------------- > > Any idea ? > > Thanks, > > Yaron Cohen > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: CodeTEST/Memory or PerformancePak (MemScope) ? Date: Fri, 15 Feb 2002 10:13:18 +0100 From: "Werner Schiendl" Message-ID: <3c6cd127@brateggebdc5.br-automation.co.at> References: <3C6C2120.D559AF1A@rti.com> Hi, "Bill Pringlemeir" wrote in message news:ulmdvpy45.fsf@yahoo.com... - -- snip -- > allocation... perhaps that was the code coverage tool. Anyways, I > hope that people don't get too uppity and chastise you for say "Hey, > look what I make." - -- snip -- I cannot see a reason for chasing someone for answering _the_ original posters question. Stressing the advantages of ones product - if this was the question - is no fault in my opinion. This is very different from trying to sell a tool that is not in the least necessary to solve the problem in question. best regards Werner --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Project-relative files? Date: Fri, 15 Feb 2002 10:25:32 +0100 From: "Werner Schiendl" Message-ID: <3c6cd405@brateggebdc5.br-automation.co.at> References: <3C6C3D68.5E14B2EB@dsto.defence.gov.au> Hi, you could try to manually edit the .wpj file manually. Use an editor that conserves Unix line endings if you work on Windows (LF only), otherwise your .wpj will not work afterwards. The format of the file is pretty self-explanatory. I never tried if Tornado can properly work with $(PRJ_DIR)/../someFile.c - you need give it a try. hth Werner "Tim Shaw" wrote in message news:3C6C3D68.5E14B2EB@dsto.defence.gov.au... > We tried to do that about 12 months ago, using T2 on Solaris. But our local > support person said it wasn't possible. > > Bob Bradley wrote: > > > Is it possible to convince Tornado to reference files by relative path > > that are above the project file in the hierarchy when added to the > > project? It seems to use project-relative paths when the files are below > > it in the hierarchy, but uses full paths for files above it. For > > example, I might have a hierarchy as follows: > > > > Sources > > Make > > MyProject.wpj > > Misc > > MyFile.c > > > > When I add "MyFile.c" to "MyProject.wpj", I'd like it to be saved in the > > project file (and the resulting makefile) as > > "$(PRJ_DIR)../Misc/MyFile.c" instead of using the full path to MyFile.c. > > Our project is built on a variety of systems so using a full path > > doesn't work. > > > > Is this possible? > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Handling select() function failures Date: Sun, 10 Feb 2002 14:33:23 -0500 From: joe durusau Organization: Bellsouth.Net Message-ID: <3C66CB03.7F03BF22@bellsouth.net> References: <3C66C6C2.C3B3CA0E@ubbsrv3.ho.lucent.com> Chandwani wrote: > Hi ! > > I am facing a problem with select() function. Sometimes this function returns > with an ERROR (all fds are fine). The vxWorks manual doesn't say exactly why > thsi happens and how to handle this situation. My suspiscion is that if I put > select() in a FOREVER loop and it returns an ERROR, the task will go in an > infinite loop. Why does select() return an ERROR and how should I handle it in a > FOREVER loop then ? > Check the global variable errno. It's value should tell you what went wrong. > > One more question : > > UDP broadcasts to 135.0.0.0 network don't work. They work well to 192.0.0.0 > network. Has anyone else seen this behavior in their environment ? > Never tried it, maybe someone else has. > > Thanks, > Rakesh Speaking only for myself, Joe Durusau --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Fri Feb 15 04:41:18 2002 From: Ronen Levy Date: Fri Feb 15 04:41:21 PST 2002 Subject: SNMP product survey Hi, Survey for SNMP agent for vxWorks plus some NMS on window. I have looked at WindRiver documentation for WindNet and Envoy SNMP agent. As I understand, these products are for the SNMP agent ALONE. Do they come with any mib browser or GUI tester? What I am looking for is an SNMP agent, cost effective, preliminary for testing purposes. I also considering an HTTP support, which give me access thru Internet Explorer or Netscape. There are so many out there, and any recommendations are appreciated. Ronen. From vxwexplo-errs@csg.lbl.gov Fri Feb 15 09:14:11 2002 From: Dale Luck Date: Fri Feb 15 09:14:13 PST 2002 Subject: Re: Tool for checking memory leak 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_01C1B643.3CBE8EF0 Content-Type: text/plain; charset="iso-8859-1" We use Seaweed softwares drop in memLib library replacement for vxworks. Contact them for price. We received good technical support. http://www.seaweed.com/page3.html#The%20memLib It has a memleak checker, also can check for memory corruption. Dale Luck Shoreline Communications ------_=_NextPart_001_01C1B643.3CBE8EF0 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Re: Tool for checking memory leak

We use Seaweed softwares drop in memLib library = replacement for vxworks.
Contact them for price. We received good technical = support.
http://www.seaweed.com/page3.html#The%20memLib

It has a memleak checker, also can check for memory = corruption.

Dale Luck
Shoreline Communications

------_=_NextPart_001_01C1B643.3CBE8EF0-- From vxwexplo-errs@csg.lbl.gov Fri Feb 15 16:05:21 2002 From: "Bob Kindel" Date: Fri Feb 15 16:05:23 PST 2002 Subject: RE: [Fwd: Re: Tool for checking memory leak] MemScope can tell you about heap corruption through the HeapUtilities. These utilities ship with MemScope and are available both from the command line (using the C inteface) and from the MemScope GUI. In the GUI select View->Heap Utilities to access the graphical interface. HeapCheck() walks the heap, checking for inconsistencies. This can allow you to determine that you have a problem with the heap and where the problem occurs. From there you can either try to back out where in your code you might access the memory (for instance if it is at the end of an array, you can look for code that might overwrite the end of the array). Another common use is to "protect" some functions by checking the heap both before and after the function is called. For details on the heap utilities take a look at the manual (section 3.3 in the ScopeTools 2.0b manual for the graphical interface and page 3-3->3-24 in the RTITools 4.0 reference manual for the C reference). The manuals are normally shipped in electronic form in $WIND_BASE/rti/scopetools.x.x/docs in case you don't have a paper copy. Best Regards, Bob Kindel Real-Time Innovations, Inc. > > -------- Original Message -------- > Subject: Re: Tool for checking memory leak > Date: Fri, 15 Feb 2002 08:40:50 +0000 > From: Danny Chapman > Organization: Lucent Technologies, Columbus, Ohio > Newsgroups: comp.os.vxworks > References: <3C6B932E.C080423B@oslo.online.no> > <9b6f8883.0202141120.42e1ae04@posting.google.com> > > > > On 14 Feb 2002, Mark Chang wrote: > > > Glenn Ivar Klausen wrote in > message news:<3C6B932E.C080423B@oslo.online.no>... > > > Is there a tool simular to purifi or boundchecker in vxWorks? > > > > I have been using RTI's Memscope available > > from WindRiver. It will display a 'real-time' listing of > > functions that call malloc, free and the size of memory. > > Unfortunately memscope won't tell you anything about corruption. The only > way I know of is to write your code so that it also builds and runs under > (for example) Solaris, where you can run purify. > From vxwexplo-errs@csg.lbl.gov Fri Feb 15 16:13:07 2002 From: "Bob Kindel" Date: Fri Feb 15 16:13:09 PST 2002 Subject: RE: [Fwd: CPU usage thru wind shell] Hi Rakesh- You can effectively run ProfileScope in target only mode by running ProfilePrompt (see Chapter 4 of the ProfileScope manual). This is effectively ProfileScope running in command-line mode. You will need to load the necessary libraries for ProfileScope on your machine and then run: > ProfilePrompt The following is from the manual on using ProfilePrompt: Using ProfilePrompt() The default initialization starts the profiler in tree mode (granularity 2); the routines are sorted by their position in the execution call stack. >ProfilePrompt Do you need help? (y/[n]) 1-Aux; 2=System; 3=none. Clock =? [1] StethoScope index (-1 => don't use StethoScope) =? [0] ProfilePrompt Command =? [none] After you have answered the initialization questions, a summary of the active-call tree can be printed via the show command: ProfilePrompt command =? [none] show routines If the system is not heavily loaded, this display is likely to be uninteresting; most of the activity will be idle time, which is not shown by default. On the other hand, if your system is complex, you may wish to limit the display to rou-tines called by a particular function or to routines using some minimum percentage of the CPU. To set these filters, type show parameters. You can enter flat mode (granularity 0) by typing mode and answering all questions "no." In this mode, the call stack is not important; each routine appears in the profile only once. Allow a short period for the profiler to collect data, and then display the system activity via show routines. Best Regards, Bob Kindel Real-Time Innovations, Inc. > > -------- Original Message -------- > Subject: CPU usage thru wind shell > Date: Fri, 15 Feb 2002 15:31:16 -0500 > From: Chandwani > Organization: Lucent Technologies > Newsgroups: comp.os.vxworks > > Hi ! > > Is there a way to see the CPU usage ( on per task basis ) by a wind shell > command ? I know about the RTI tools like "Profilescope" etc. but > I am having > tough time running such data intensive GUI on my target boards ( > ppcec603 arch. > ). > > Any commands from windshell giving even the cursory data will > help as my target > board is becoming sluggish after some time and is also tasks are > leaking memory. > > Thanks, > -- > Rakesh Chandwani > Lambda Router AOS SW Dev.| Lucent Technologies, > Room 3N-320 | 101 Crawfords Corner Road, > Ph:732-949-0576 | Holmdel, NJ - 07748 > From vxwexplo-errs@csg.lbl.gov Sat Feb 16 04:03:15 2002 From: Vxworks Exploder Date: Sat Feb 16 04:03:18 PST 2002 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sat Feb 16 04:03:10 PST 2002 Subject: Re: Reading from serial ports - Blocking or interrupts? Subject: Re: Reading from serial ports - Blocking or interrupts? Subject: ipFilterHookAdd() Problem Subject: memShow crash Subject: for PAP configuration Subject: Re: Tool for checking memory leak Subject: Re: Tool for checking memory leak Subject: Re: What's vxImmrGet() function for? Subject: Test Subject: Re: Tornado II Prototyper installation failure Subject: Re: Beginner Client/Server problems! Subject: Re: VirtuallAlloc failed error on compilation Subject: Re: Measuring performance of Embedded Software. Subject: Re: What's vxImmrGet() function for? Subject: Re: memShow crash Subject: Re: VirtuallAlloc failed error on compilation Subject: Re: VirtuallAlloc failed error on compilation Subject: Re: 64-bit operations Subject: Re: What's vxImmrGet() function for? Subject: CPU usage thru wind shell Subject: Problem booting four systems at once over the network. Subject: Re: What's vxImmrGet() function for? Subject: test Subject: Under Windows ME, Tornado II Prototyper support network? Subject: Re: memShow crash Subject: Re: Problem booting four systems at once over the network. ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Reading from serial ports - Blocking or interrupts? Date: Fri, 15 Feb 2002 12:05:03 GMT From: vtail@hotmail.com Organization: MV Communications, Inc. Message-ID: <3c6cf920.89418883@news.netis.com> References: <37449c9d.0202130624.43ffc949@posting.google.com> <3c6b9334$1@brateggebdc5.br-automation.co.at> Good morning, Werner. Thank you very much for your comprehensive and informative reply. It tells me exactly what I needed and also confirms what I suspected about the unblocking mechanism. Eric --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Reading from serial ports - Blocking or interrupts? Date: Fri, 15 Feb 2002 12:10:03 GMT From: vtail@hotmail.com Organization: MV Communications, Inc. Message-ID: <3c6cf987.89521798@news.netis.com> References: <37449c9d.0202130624.43ffc949@posting.google.com> <1IRa8.2442$5o.1075855@newsr2.u-net.net> On Thu, 14 Feb 2002 16:29:57 -0000, "Dave Korn" wrote: >Can you determine how long the packet is going to be by inspecting the first >one or two chars ? Good morning, Dave. Actually, the packets on some of the serial ports are fixed length, so on those it's not a problem. Others are NMEA sentences that have variable-length value fields (e.g. a water depth might be 4.7 meters or it might be 258.3 meters) thus varying the length of the sentence. There is no way to detect this until the trailing CR/LF is received. So, the plan (finalized with the help of you all here in the NG) is to set the interrupt level for the minimum sentence length and then handle the rest of the sentence using the 4-character-time no-data interrupt. Thanks... Eric --------------------------- Newsgroups: comp.os.vxworks Subject: ipFilterHookAdd() Problem Date: Fri, 15 Feb 2002 08:48:29 -0500 From: John Sestak Organization: Raytheon Company Message-ID: <3C6D11AD.385877BC@fallschurch.esys.com> Hi all, Well basically what I'm trying to do is allow my host to receive tunneled IP packets. The original IP header contains a protocol ID of 36 (which is for XTP) but we have to tunnel this original packet inside another packet than has a IP header with a protocol ID of 4 (IP Multicast). So, the packet essential arrives at my host looking like: [Ethernet header] [New IP Header (prot 4)] [Original IP header (prot 36)] [ XTP header] [Data] The way I tried to fix this problem is by using the 'ipFilterHookAdd()' function to replace the new IP headers' protocol ID of 4 with a protocol ID of 36 and stamp the first 20 bytes where the original IP header (protocol 36) used to be with a special sequence of bytes indicating that this packet was originally a tunneled packet and so I should skip the first 20 bytes where the original IP header was and start reading the XTP header from there. This has caused two problems. 1) It seems as though if I try to run anymore than a few instructions in the ipFilterHook function that gets called when a packet arrives, the program will crash. Originally, I was going to check all the fields of the original IP header to make sure it was valid, and do a few other things, etc. but it always crashed that way. When I scaled it down to just replacing the protocol ID of 4 with a 36 and stamping the first 20 bytes of the original IP header with a special sequence of bytes then it seems to run ok but, this it only runs ok when I'm running a very SMALL test application. Once I try it with a larger application (multiple tasks running at once) that sends and receives on many different sockets it dies. Essentially, what I'm saying is that the hook function doesn't seem to work well if it stays in this function for too long or if the processor is under a heavy load. Any ideas as to why??? 2) First of all, I can't seem to find a *.h file anywhere in the Tornado directory that has a function prototype for ipFilterHookAdd() so when I compile obviously the compiler says 'ipFilterHookAdd() implicitly declared'. I tried running the program even with the warning that it gave and it seemed to work with a small test program I wrote. So, I then tried adding my own function prototype to get rid of the compiler warning but when I ran it, the program crashes. Where in the world is the function prototype for this 'ipFilterHookAdd()' function??? I found a *.h file that had a function prototype for 'ipFilterHookInit()' but not ipFilterHookAdd(). Any ideas as to how to fix this hook function or maybe another way I could receive tunneled packets (i.e. grab the tunneled, protocol 4 packets and change them to protocol 36 packets so that the raw socket listening for protocol 36 can receive these packets)??? Any help would be much appreciated. Thank you. John --------------------------- Newsgroups: comp.os.vxworks Subject: memShow crash Date: Fri, 15 Feb 2002 09:25:22 -0500 From: "Vlad Falcon" Organization: Bell Sympatico Message-ID: here is a snipet of shell commands that i use to monitor the memory usage on my running system. occasionaly this crashes in different ways, "machine check" error being the most often crash cause. is there any restriction for this kind of usage from the tornado shell? - ---------------------------------------------------------------------------- - --- - -> period 10,memShow A host function cannot be used as an argument. - -> - -> lkup "memShow" memShowInit 0x000f5cfc text (Bsrc-VxBSP1PB2.10.40) memShow 0x000f5d38 text (Bsrc-VxBSP1PB2.10.40) value = 0 = 0x0 - -> - -> - -> period 10,0x000f5d38 task spawned: id = 5d5140, name = s1u3 value = 6115648 = 0x5d5140 - -> status bytes blocks avg block max block ------ --------- -------- ---------- ---------- current free 3367232 25 134689 3357288 alloc 10671608 1707 6251 - cumulative alloc 182428120 1651347 110 - en for StationId 0. PDU is destroyed status bytes blocks avg block max block ------ --------- -------- ---------- ---------- current free 3367232 25 134689 3357288 alloc 10671608 1707 6251 - cumulative alloc 182428120 1651347 110 - status bytes blocks avg block max block ------ --------- -------- ---------- ---------- current free 3367232 25 134689 3357288 alloc 10671608 1707 6251 - cumulative alloc 182428120 1651347 110 - status bytes blocks avg block max block ------ --------- -------- ---------- ---------- current free 3367232 25 134689 3357288 alloc 10671608 1707 6251 - cumulative alloc 182428120 1651347 110 - en for StationId 0. PDU is destroyed status bytes blocks avg block max block ------ --------- -------- ---------- ---------- current free 3367232 25 134689 3357288 alloc 10671608 1707 6251 - cumulative alloc 182428120 1651347 110 - - ---------------------------------------------------------------------------- - --- Our Configuration: Tornado 2.0, PPC860, host Win-NT 4.0 (SP6, build 1381) Thanks, Vlad. --------------------------- Newsgroups: comp.os.vxworks Subject: for PAP configuration Date: Fri, 15 Feb 2002 15:03:58 GMT From: "alain tamdem" Organization: Rauland Borg Corporation Message-ID: Reply-To: "alain tamdem" To allow authentification using PAP do i need only to turn the option PPP_OPT_NO_PAP to 1 or also PPP_OPT_REQUIRE_PAP to 1 ?? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tool for checking memory leak Date: Fri, 15 Feb 2002 15:33:15 -0000 From: Organization: Posted via Supernews, http://www.supernews.com Message-ID: References: <3C6B932E.C080423B@oslo.online.no> <9b6f8883.0202141120.42e1ae04@posting.google.com> Danny Chapman wrote: > Unfortunately memscope won't tell you anything about corruption. The only > way I know of is to write your code so that it also builds and runs under > (for example) Solaris, where you can run purify. *Commercial plug warning* Seaweed Systems has a written-from-scratch memLib replacement. It has all the functionality which Wind River's memLib has...and more, One of its additions is a corruption checker. It is not anything as slick as purify, but it's alot better then no corruption checker. Please call or email for information: bob@seaweed.com +1 425.895-1721 bob --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tool for checking memory leak Date: Fri, 15 Feb 2002 15:54:10 +0000 (UTC) From: David Laight Organization: BT Openworld Message-ID: <3C6D2F77.8010009@l8s.co.uk> References: <3C6B932E.C080423B@oslo.online.no> Glenn Ivar Klausen wrote: > Is there a tool simular to purifi or boundchecker in vxWorks? You could always? Intercept all calls to memPartAlignedAlloc (eg by overwriting the first intruction with a branch to the interceptor - which must finish by jumping to memPartAlignedAlloc+4 after issueing the first inst.) Add (probably) 8 bytes to the length, allocate the memory, put a 'tag' and the address of caller (follow stack frames for expected callers). Set tag=1, run the system, run your test once. Set tag=2, run your test again, Set tag=3, run your test a third time. Scan the heap area (trivial format - once you have found the start!) looking for allocated items with tag=2 - they are probably orphaned. See who allocated them, wotk out why they weren't freed. Share and enjoy.... :-) David --------------------------- Newsgroups: comp.os.vxworks Subject: Re: What's vxImmrGet() function for? Date: Fri, 15 Feb 2002 10:30:39 -0700 From: "Liang Yang" Organization: RTS-LAB Message-ID: References: Reply-To: "Liang Yang" But I can't find any description about the function vxGetImmr() in VxWorks library (it seems it doesn't belong to any library), nor can I find its macro definion in any .h files in tornado. It is very strange. Where does vxgetImmr() come from? "Gary M" wrote in message news:U_2b8.65175$AV5.354273@rwcrnsc51.ops.asp.att.net... > If you delete the variable definition and its assignment to vxImmrGet(), and > the compiler doesn't complain, then you don't need it. If it does complain, > then its use is probably buried in a macro inside the ISR. > > "Philip Macfarlane" wrote in message > news:u6o1tb4q7bc33e@corp.supernews.com... > > On Thu, 14 Feb 2002 18:46:31 +0000, Liang Yang wrote: > > > > > But why should I always need to call this routine before I set my ISR. > > > What's relationship between the Internal Memory Map Register and my own > > > ISR? > > > > Calling it on it's own (and disregarding the value returned) doesn't do > > anything, as far as I know. The function returns the value of the IMMR, so > > you can subsequently use it to access some peripheral. e.g. > > > > volatile unsigned long *immr = vxGetImmr(); > > immr[foo] = bar; /* Yuck - Put the foo peripheral into bar mode. */ > > > > (Not a brilliant example, but hopefully useful.) > > > > Cheers, > > Philip. > > --------------------------- Newsgroups: comp.os.vxworks Subject: Test Date: Fri, 15 Feb 2002 17:36:31 GMT From: jackblack@yahoo.com (Bluesman) Message-ID: <3c6d470f.67109868@news.sunwave.net> test --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado II Prototyper installation failure Date: Fri, 15 Feb 2002 15:07:46 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: <576881c0.0202121417.e1df9b0@posting.google.com> <576881c0.0202131100.2c440d89@posting.google.com> "Nick Forrette" wrote in message news:576881c0.0202131100.2c440d89@posting.google.com... > "Dave Korn" wrote in message news:... > > "Nick Forrette" wrote in message > > news:576881c0.0202121417.e1df9b0@posting.google.com... > > > First off, I'm very new to VxWorks and have been assigned to learn a > > > bit about it. As part of my exploration, I've tried to install the > > > Tornado II Prototyper under Windows 2000. I am having two problems > > > with the installer. > > > > > > 1. The Tornado II Prototyper tar file they ship includes two files > > > named "inatr:intf". Of course colons are illegal characters for > > > filenames in Windows. I worked around this by extracting everything > > > but those two offending files. > > > > That's bizarre. Those sort of names *are* allowed on NTFS under either NT > > or 2K: they refer to an alternate data stream called 'intf' in a file > > called 'inatr'. But it seems a bit unlikely that WRS would really mean > > that, so I reckon that something is pretty wrong here. > > Odd. I can't seem to create files with those names. Wordpad reports > "You can not save in the URL location you specified. Please choose > another location." when I tried to save a file called "asdf:fdsa.txt" > or "asdf.txt:fdsa". I had simmilar issues with the old "edit" command > under "cmd.exe". The colon may be a legal character as far as the > filesystem is concerned is concerned, but the Windows sure does its > darnest to prevent you from actually creating files with them. Heh, I was going to post an example to prove it works, and that it's perhaps the windows file requester filtering the colon, but i couldn't make it work. Then I remembered all my partitions on my PC at work here are FAT, not NTFS, so there's no wonder. But try this at a dos prompt on an NTFS partition: echo. >file.txt Hello world dir >>file.txt:stream2 type file.txt type file.txt:stream2 Note that it could conceivably go wrong because the dos prompt might thing you were referring to a file called 'stream2' in the current directory of drive "file.txt:", by analogy to names like C:file.txt, so you might need to specify the full drive letter and path in that example. Also, go to http://www.heysoft.de and download LADS. Drop some graphics files in your "My pictures" folder, enable the thumbnail view, then run LADS on it. > I did a little checking. I'm pretty sure (but not positive) that isn't > the problem. Windows NT itself has a few files / directories with > those names in them (e.g the directory C:\WINNT\System32\Com\ in my > Win2k install). Fair enough, but it still applies to some names: here's the results of a little test I did C:\WINNT\SYSTEM32>dir c*/ad Volume in drive C has no label. Volume Serial Number is D845-2F2F Directory of C:\WINNT\SYSTEM32 24/03/98 16:20 config 20/12/99 11:27 CertSrv 2 File(s) 0 bytes 262,799,360 bytes free C:\WINNT\SYSTEM32>cd com The system cannot find the path specified. C:\WINNT\SYSTEM32>mkdir com C:\WINNT\SYSTEM32>cd com C:\WINNT\SYSTEM32\com>dir Volume in drive C has no label. Volume Serial Number is D845-2F2F Directory of C:\WINNT\SYSTEM32\com 15/02/02 15:03 . 15/02/02 15:03 .. 2 File(s) 0 bytes 262,766,592 bytes free C:\WINNT\SYSTEM32\com>dir >com C:\WINNT\SYSTEM32\com>dir Volume in drive C has no label. Volume Serial Number is D845-2F2F Directory of C:\WINNT\SYSTEM32\com 15/02/02 15:03 . 15/02/02 15:03 .. 15/02/02 15:04 339 com 3 File(s) 339 bytes 262,733,824 bytes free C:\WINNT\SYSTEM32\com>type com Volume in drive C has no label. Volume Serial Number is D845-2F2F Directory of C:\WINNT\SYSTEM32\com 15/02/02 15:03 . 15/02/02 15:03 .. 15/02/02 15:04 0 com 3 File(s) 0 bytes 262,733,824 bytes free C:\WINNT\SYSTEM32\com>dir >aux Access is denied. C:\WINNT\SYSTEM32\com>dir >prn C:\WINNT\SYSTEM32\com> After that last one, a listing of the directory came out of my printer; despite there being no : after the name prn. Note also that the name 'aux' was forbidden. DaveK - -- moderator of alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! Proud Member of the Exclusive "I have been plonked by Davee because he thinks I'm interesting" List Member # Master of Many Meowing Minions Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage above and beyond the call of hilarity. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Beginner Client/Server problems! Date: Fri, 15 Feb 2002 14:33:59 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: "Joseph Cook" wrote in message news:b6f5d6a9.0202131121.79ad0d13@posting.google.com... > Ok. The problem was in the bits of code unshown. So, now I am timing > this sequence for a 256 byte buffer, 512, 1K, 2K, 4K, etc. Everything > looks fine (even if terrible) until I hit 4K buffer size, and then the > performance comes to a screaching halt, outputting 40Kb/s (when the 2K > buffer size provides 6 Mb/s). I'm the only process on this system, > and I'm my own independent network...any ideas? That's almost certainly related to the number of mbufs of various sizes that the wrs stack initializes: the fact that it happens at 4k, rather than anything related to say ethernet maximum frame size suggests so, anyway. There's a document on windsurf somewhere about how to tune the stack's mbuf allocations according to your applications needs. See if you can find it and it should help some. 'fraid I don't know the link myself. DaveK - -- moderator of alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! Proud Member of the Exclusive "I have been plonked by Davee because he thinks I'm interesting" List Member # Master of Many Meowing Minions Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage above and beyond the call of hilarity. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VirtuallAlloc failed error on compilation Date: Fri, 15 Feb 2002 14:58:00 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: <89441eba.0202141011.1e4c5bd@posting.google.com> "isro" wrote in message news:89441eba.0202141011.1e4c5bd@posting.google.com... > Hi, > I am using Tornado compiler 2.1.1 for XSCALE > 2.9-010413 > > Any idea, why i am getting the following error on compilation > > commit_and_inc: VirtualAlloc failed > > The file i am copiling is cpp file with size 126K. My computer has about 256K RAM. > Thanks in advance > isro If you've run out of memory, maybe either your windows swapfile or your *nix pagefile is probably full, perhaps even the HD partition containing it is full. Check these things and take any steps you can to grow the size of it. 256K of ram isn't very much these days. It could simply be the case there isn't enough memory for the essential non-pageable OS structures/pagetables relating to a larger swapfile. Or perhaps you meant to say you had 256M of ram ? DaveK - -- moderator of alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! Proud Member of the Exclusive "I have been plonked by Davee because he thinks I'm interesting" List Member # Master of Many Meowing Minions Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage above and beyond the call of hilarity. --------------------------- Newsgroups: comp.os.psos,comp.realtime,comp.arch.embedded,comp.os.vxworks Subject: Re: Measuring performance of Embedded Software. Date: Tue, 12 Feb 2002 07:20:09 GMT From: u913848876@spawnkill.ip-mobilphone.net Message-ID: References: Sender: u913848876@spawnkill.ip-mobilphone.net > > Hi all, > We were trying to measure the performance of our real time software which is a > PSOS based system. We basically want to measure the performance of our device > drivers, DLC software and the kernel. Can you please give us guidelines on > how to go about it. Any methods,samples,software,links etc... > with regards, > Ashokc. There are many ways to define and measure performance. Here is a simple technique for dynamically measuring the total CPU load time in the system (including all tasks and interrupt handlers). a) Define your own idle task. This should basically be an infinite loop incrementing a static counter (32-bit or even 64-bit long; just make sure that it is done atomically). This task should be given a priority that is lower than any other task on your system, yet higher than that of the system's idle task. The idle loop should be coded identically to the code that is used for auto- calibration (see d below). b) Periodically (say every 100 msec. or every 500 msec.) have the above counter read and cleared in the context of another (periodic) task in your application. Also need to do this atomically. c) Sample the system's time upon each such reading. The value that has accumulated in the counter divided by the length of the accumulation period is directly indicative of the (average) idle time in your system for that measurement period. d) In order to be able to convert the result of the above calculation to a meaningful value (percent CPU load), you need to know the rate at which the idle loop increments the counter on your particular CPU. To accomplish that, perform an auto-calibration measurement where you run the loop for a known number of times and measure the elapsed time. This measurement should be done once at system initialization time. Preferably, this measurement should be done using a hardware timer, in which case you can run it with the interrupts locked out. If this is not feasible, use the system clock (that would make it less precise). Obviously, you can give up auto-calibration and instead measure the idle loop once and define a constant in your code. However this is hardly a good choice (tomorrow or next year you or somebody else may need to port your system to a higher-performance CPU). HTH, Beni Falk - -- Spam protected message from: Sent by falk_beni from rada in area com Posted via http://www.usenet-replayer.com/cgi/content/new --------------------------- Newsgroups: comp.os.vxworks Subject: Re: What's vxImmrGet() function for? Date: Fri, 15 Feb 2002 18:04:19 +0000 From: Philip Macfarlane Organization: RAMiX Europe Ltd Message-ID: References: On Fri, 15 Feb 2002 17:30:39 +0000, Liang Yang wrote: > It is very strange. Where does vxgetImmr() come from? On my installation, it appears to be in libPPC860gnuvx.a Your mileage may vary. :-) Cheers, Philip. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: memShow crash Date: 15 Feb 2002 10:15:51 -0800 From: pkockritz@mac.com (Pete Kockritz) Organization: http://groups.google.com/ Message-ID: <9423c55b.0202151015.1bd3f655@posting.google.com> References: "Vlad Falcon" wrote in message news:... > here is a snipet of shell commands that i use to monitor the memory usage on > my running system. occasionaly this crashes in different ways, "machine > check" error being the most often crash cause. is there any restriction for > this kind of usage from the tornado shell? No restrictions that I'm aware of. This could be an indication of memory corruption, however. You might try turning on some of the debug options in memLib using memOptionsSet and see if it catches anything. > -> period 10,memShow > A host function cannot be used as an argument. Try this: period 10,@memShow Also, you could use memPartInfoGet to get the memory usage in your program, have it periodically check the amount free and panic if it gets too low. #include MEM_PART_STATS memStats; memPartInfoGet (memSysPartId, &memStats); Hope this helps, Pete Kockritz - -- pkockritz at mac dot com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VirtuallAlloc failed error on compilation Date: 15 Feb 2002 13:15:49 -0500 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: References: <89441eba.0202141011.1e4c5bd@posting.google.com> Sender: bpringlemeir@DeadDuck >>>>> "isro" == isro writes: isro> Hi, I am using Tornado compiler 2.1.1 for XSCALE 2.9-010413 isro> Any idea, why i am getting the following error on compilation isro> commit_and_inc: VirtualAlloc failed isro> The file i am copiling is cpp file with size 126K. My computer isro> has about 256K RAM. Thanks in advance isro FAE or WRS support has code for pepatch. Alternatively you can download the mingw toolset and use objcopy to change the stack size to something sane. The stack size has been set wrong for a great many WRS tools that are release with 2.0 and 2.1.1. You can definitely avoid the error by compiling without optimization. This is not a good solution, I just present it as a bandaid in case this is a critical path item. I have attached my conversation with WRS support on this matter. {start comedy} [Message 1] Hello Bill, This is XXXXXXX from Wind River Systems. I have been assigned to work on your TSR# 270458: "memory exhaustion problem". Before I can start my research I need the following from you. 1) Setup.log (located in the root of tornado directory) 2) Are you building through project facility or command line? 3) If command line, can you send me the command line options that you are using. 4) Complete build output. (once again) [Message 2] Bill, Here is the zip file. [Message 3] Bill, Thank you for sending the information. Can you please try increasing the stack size to 0x8000000. This seems to be a magic number that has worked in the past. Let me know if this helps, [Message 4] Bill, Thank you for writing. In most cases the problem with virtual alloc fail is resolved with increasing the stack size using pepatch. However, there is an SPR # 70780 filed against pepatch not resolving the problem. As per our conversation , I am sending you a new cygwin32.dll. To install, go to %WIND_BASE\host\x86-win32\bin Rename cygwin32.dll to oldcygwin32.dll copy this file in as cygwin32.dll. Let me know if this works, [Message 5] Bill, I am glad that changing the .dll file helped resolve access violation problem. To answer your other question, we have no memory requirement for the compiler to build. Since you are trying to build 10 megabyte file the speed will be a factor. I am not sure if you are aware of the difference between T 2.0.1 and T 2.1.1. T 2.0.1 using GNU 2.72 and T 2.1.1 is using GNU 2.96+ patches this may cause slow performance. Specially using templates. As a side note, -O3 optimization not supported please use lower than -O3 for optimization. Having faster processor may help decreasing build time. Since this is a performance issue; I am not sure if there is anything I can do besides filing an SPR. Let me know if you like me to file an SPR against GNU 2.96. We're you able to build your project after increasing stack size and changing the .dll file? [Message 6 from me...] The file did build after more than one hour (with -O2). I was not using -O3 on this file, just -O2. I have seen an SPR for -O3 taking a long time. Can you verify that a faster processor is needed or is it more memory. I am in charge of deciding whether to upgrade to T2.1.1 or stay with T2.0.1 At the very least I must tell management that developers need faster machines and/or memory and the capitol cost must be justified against my labour to port our software to another OS such as Linux or building a Mingw cross compiler. I am aware that WRS has increased the gcc version to 2.95 (are you sure it is 2.96?). Compiling the source with a Linux native compiler or Mingw GCC does not take this long. This lengthy time is only associated with the WRS toolset for whatever reason. I can give specific performance results if you like. I think it would be especially instructive to give a comparison of the T2.0.1 vs the T2.1.1 compiler. I have told management that templates are not good and that we should convert to simple C++ container classes. However, no matter what we do to our source, the Zinc C++ GUI (from WRS) that we use includes template code. Regards, Bill Pringlemeir. [message 7, six days later] Bill, I hope I can clear up some questions, our GNU 2.96 uses 2.9-based toolchain. We have named our version to gnu 2.96. Gnu 2.96 has better support for C++ templates than GNU 2.72, so it is benificial for you to use gnu 2.96. Another side note that you may want to consider is all of our new releases will be built with gnu 2.96. As far as I performance issue, please contact your sales person who can better questions. Since the you are no longer having problems building your image, the original issue for this TSR has been resolved. We like to keep one issue per TSR for our database purposes. I think sales person can break down the difference between T 2.0.1 and T2.1.1. Let me know if I can close this TSR since you are no longer having problems building. {end comedy} Admittedly I am a jerk to deal with. However, this support person gave me the typical `run-around' that makes me slam their support every time. I have had good tech support people, but I remember dialogs like this... It seems that they always solicit my opinion of the tech support when a good person answers. I am not sure of my legal requirements regarding the pepatch code or the cygwin dll that has been distributed to me. However, the compiler is a little bug ridden when it comes to NT, C++ and the Windows host. I suppose WRS management need to create a reason for their `Tech support' so they can generate more revenue. Too bad, it give some of the good guys at WRS a bad name. I guess a file that did compile in 10 minutes and now takes over 60 is status quo. hth, Bill Pringlemeir - -- Have you ever joined a church where everyone but you can sing with your rhinocerous? Or been forced by your telephone to eat bananas? You will. And the company that will bring it to you: AT&T. vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VirtuallAlloc failed error on compilation Date: 15 Feb 2002 13:21:34 -0500 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: References: <89441eba.0202141011.1e4c5bd@posting.google.com> Sender: bpringlemeir@DeadDuck >>>>> "Bill" == Bill Pringlemeir writes: Bill> [message 7, six days later] [...] Bill> {end comedy} Bill> Admittedly I am a jerk to deal with. However, this support [snip] Sorry, six days was from the start of the conversation. It seemed misleading when I re-read it. fwiw, Bill Pringlemeir. - -- SANTA CLAUS comes down a FIRE ESCAPE wearing bright blue LEG WARMERS.. He scrubs the POPE with a mild soap or detergent for 15 minutes, starring JANE FONDA!! vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" --------------------------- Newsgroups: comp.os.vxworks Subject: Re: 64-bit operations Date: Tue, 12 Feb 2002 06:25:27 -0500 From: joe durusau Organization: Bellsouth.Net Message-ID: <3C68FBA7.43FCD6FF@bellsouth.net> References: <3C67ADF8.55BA8A64@bellsouth.net> Denis Perelyubskiy wrote: > hello, > > * joe durusau [02-Mon-02 03:41 -0800]: > > > > This is just a guess, but have you tried splitting the variable into bytes > > (a union or some such) an dprinting the bytes? It is entirely possible that > > the bsp does not handle long long and therefore prints them as zero. > > On most BSPs, you fill find that trying to print a float with logMsg > > will do similar things, because logMsg just doesn't support floats. > > what does the BSP have to do with 64 bit operations on the > printf()? i'd thought it was all about the compiler? > > thanks, > > > Nejhdeh Ghevondian wrote: > > > > Dear anyone, > > > I am trying to implement 64-bit operations in my VxWorks operations (using > > > the GNU compilier, on PowerPC platform). > > > I am trying to multiply and divide unsigned long long int variables. > > > However, I always get 0 as value. > > > I acnt even do a printf("%llu\n", 64bit_var) > > > > > > Can anyone help me??? > > > > > > Regards Nejhdeh Ghevondian > > denis > > -- > if you want to reply by mail, please tweak my > address in a reasonable way. sorry for the > inconvenience. The compiler just generates a call to routines in the BSP which do the actual work. In a more 'normal' situation, there is a C (or C++) library on the system that does the work, but in vxWorks, all that stuff is inside the BSP. Use -s and look at the assembler output. You won't see the compiler generating any big conversion routines. They are all in some 'library' code that gets called at run time. Again, I'm just guessing as to the root cause of you problem, but you should be able to investigate it easily enough and see whether I'm all wet. Speaking only for myself, Joe Durusau --------------------------- Newsgroups: comp.os.vxworks Subject: Re: What's vxImmrGet() function for? Date: Fri, 15 Feb 2002 13:57:21 -0600 From: Andrew Johnson Organization: APS, Argonne National Laboratory Message-ID: <3C6D6821.AD479A16@aps.anl.gov> References: Philip Macfarlane wrote: > > On Fri, 15 Feb 2002 17:30:39 +0000, Liang Yang wrote: > > > It is very strange. Where does vxgetImmr() come from? > > On my installation, it appears to be in libPPC860gnuvx.a The header file that refers to it is /target/h/arch/ppc/vxPpcLib.h - - Andrew - -- Perfection is reached, not when there is no longer anything to add, but when there is no longer anything to take away. - - Antoine de Saint-Exupery --------------------------- Newsgroups: comp.os.vxworks Subject: CPU usage thru wind shell Date: Fri, 15 Feb 2002 15:31:16 -0500 From: Chandwani Organization: Lucent Technologies Message-ID: <3C6D7014.A91B6C5E@lucent.com> Hi ! Is there a way to see the CPU usage ( on per task basis ) by a wind shell command ? I know about the RTI tools like "Profilescope" etc. but I am having tough time running such data intensive GUI on my target boards ( ppcec603 arch. ). Any commands from windshell giving even the cursory data will help as my target board is becoming sluggish after some time and is also tasks are leaking memory. Thanks, - -- Rakesh Chandwani Lambda Router AOS SW Dev.| Lucent Technologies, Room 3N-320 | 101 Crawfords Corner Road, Ph:732-949-0576 | Holmdel, NJ - 07748 --------------------------- Newsgroups: comp.os.vxworks Subject: Problem booting four systems at once over the network. Date: 15 Feb 2002 13:34:57 -0800 From: tmehrkam@i-o.com (Thomas O Mehrkam) Organization: http://groups.google.com/ Message-ID: We have A system made up of four VME chassis and one Unix system, The VME chassis each contain one mvme167 Computer running Tornado 2.02 patched to combined patch level 2. The Unix system is a Sun Ultra I system running Solaris 2.7. The network connection is 10 base tx ethernet connected to an HP hub. The interface on the Sun system is a qe type interface. If the real time system goes out of sync then we have a method of resetting all four of the VME chassis. The chassis reboot and reconnect to the GUI on the unix system. This system has worked this way for about 5 years with 20-30 systems fielded. In the latest release of the software we upgraded from Tornado 1.0 to Tornado 2.02. We also shortened the auto boot delay from 8 seconds to 2 seconds. Now when the system reboots one or two of the chassis will hang at the "load... " message. The serial shell will still echo carrage returns so the processor is not completly down. We use rsh to download the vxworks image from the sun workstation and it hangs at the rsh call in the boot loader. The two changes were: 1) Changing from tornado 1.0 to 2.02 2) shorting the autoboot time. If I put in the old boot roms then everything works fine. I suspect that the systems are closer in sync because the number of times through the reboot countdown is less and the actual boot time is closer in sync. Using snoop to monitor the ethernet port I observed that the first thing the targets do is transmit an ARP broadcast with their ip address. The targets that boot sucessfully transmit a second ARP message with the unix hosts IP address. They get a reply then start doing RSH commands to download VxWorks. The chassis that failes goes silent after the first ARP. Observation: 1) The broadcasts are very close together and there are collisions. 2) The unix system is still trying to communicate with the target that failes. Maybe that has something to do with it. There was on open socket at the time the chassis was rebooted and it takes a while for the unix system to give up. Has anyone else had this problem booting several targets at once? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: What's vxImmrGet() function for? Date: Sat, 16 Feb 2002 04:24:58 GMT From: "Gary M" Organization: AT&T Broadband Message-ID: References: It's vxImmrGet(), not vxGetImmr(). All it does is return the base of the SIU memory space, typically 0xFF000000. Again, refer to the User Manual for your particular processor regarding the details of what resides there. Others have posted where the prototype and objects reside. "Liang Yang" wrote in message news:a4jgfc$4km$1@news.asu.edu... > But I can't find any description about the function vxGetImmr() in VxWorks > library > (it seems it doesn't belong to any library), nor can I find its macro > definion > in any .h files in tornado. > > It is very strange. Where does vxgetImmr() come from? > --------------------------- Newsgroups: comp.os.vxworks Subject: test Date: Sat, 16 Feb 2002 05:07:22 GMT From: "Jamilur Rahman" Message-ID: /**************** test test test test test test test test end ****************/ Ottawa. --------------------------- Newsgroups: comp.os.vxworks Subject: Under Windows ME, Tornado II Prototyper support network? Date: Sat, 16 Feb 2002 05:40:23 GMT From: "david" Message-ID: Hi, Thanks in advance. please tell me. I use Win98, but does not support network function, I would like to update to win2000, but not sure which version Windows ME or Windows 2000 Pro version, my computer is PII 450. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: memShow crash Date: Sat, 16 Feb 2002 12:07:52 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: Intermittent machine check exceptions used to be the norm on the 860 processor, but I was under the impression that its causes where fixed by now. The way you use memShow is quite acceptable, and there could be only two possible reasons for it to crash: 1. Perhaps your memory pool gets corrupted, which in turn causes memShow to access an illegal address and generage an exception. You can verify if this is the case by carefully inspecting the "offendinbg address" which caused the exception. 2. You could be experiencing the cache coherency hardware problem that where present in early versions of the 860 chip. If this is the cuase, you can work around it by disabling data cache. HTH, - - Leonid "Vlad Falcon" wrote in message news:UT8b8.31075$Hv5.3476629@news20.bellglobal.com... > > here is a snipet of shell commands that i use to monitor the memory usage on > my running system. occasionaly this crashes in different ways, "machine > check" error being the most often crash cause. is there any restriction for > this kind of usage from the tornado shell? > > -------------------------------------------------------------------------- - -- > --- > -> period 10,memShow > A host function cannot be used as an argument. > -> > -> lkup "memShow" > memShowInit 0x000f5cfc text (Bsrc-VxBSP1PB2.10.40) > memShow 0x000f5d38 text (Bsrc-VxBSP1PB2.10.40) > value = 0 = 0x0 > -> > -> > -> period 10,0x000f5d38 > task spawned: id = 5d5140, name = s1u3 > value = 6115648 = 0x5d5140 > -> status bytes blocks avg block max block > ------ --------- -------- ---------- ---------- > current > free 3367232 25 134689 3357288 > alloc 10671608 1707 6251 - > cumulative > alloc 182428120 1651347 110 - > en for StationId 0. PDU is destroyed > status bytes blocks avg block max block > ------ --------- -------- ---------- ---------- > current > free 3367232 25 134689 3357288 > alloc 10671608 1707 6251 - > cumulative > alloc 182428120 1651347 110 - > status bytes blocks avg block max block > ------ --------- -------- ---------- ---------- > current > free 3367232 25 134689 3357288 > alloc 10671608 1707 6251 - > cumulative > alloc 182428120 1651347 110 - > status bytes blocks avg block max block > ------ --------- -------- ---------- ---------- > current > free 3367232 25 134689 3357288 > alloc 10671608 1707 6251 - > cumulative > alloc 182428120 1651347 110 - > en for StationId 0. PDU is destroyed > status bytes blocks avg block max block > ------ --------- -------- ---------- ---------- > current > free 3367232 25 134689 3357288 > alloc 10671608 1707 6251 - > cumulative > alloc 182428120 1651347 110 - > -------------------------------------------------------------------------- - -- > --- > > > Our Configuration: > Tornado 2.0, PPC860, host Win-NT 4.0 (SP6, build 1381) > > Thanks, > > Vlad. > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Problem booting four systems at once over the network. Date: Sat, 16 Feb 2002 12:19:50 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: I can see several reasons for this problem to happen: 1. The Unix RSH daemon has a sanity limit on number of request per second. Look ath the in.rshd manual page. If this is the cause, you should see a message "rsh looping" on the Unix console or something similar. 2. Excessive colissions could be causing some of the targets to timeout, I feel this is unlikely, but your description of the packet trace hits in this direction. 3. In the past I seen systems that had a MAC address inconsistency between the Boot ROM and the booted VxWorks image (BSP nafu), this could cause the hand when combined with the reduced boot delay. 4. The Unix TCP sockets have a TIME_WAIT interval set to a high value by default, but you can easily reduce it using "ndd" utility. HTH, - - Leonid "Thomas O Mehrkam" wrote in message news:d19fa6df.0202151334.4fa7cdee@posting.google.com... > We have A system made up of four VME chassis and one Unix system, > The VME chassis each contain one mvme167 Computer running Tornado 2.02 > patched to combined patch level 2. The Unix system is a Sun Ultra I > system running Solaris 2.7. > > The network connection is 10 base tx ethernet connected to an > HP hub. The interface on the Sun system is a qe type interface. > > If the real time system goes out of sync then we have a method of > resetting all four of the VME chassis. The chassis reboot and reconnect > to the GUI on the unix system. This system has worked this way for > about 5 years with 20-30 systems fielded. In the latest release > of the software we upgraded from Tornado 1.0 to Tornado 2.02. We also > shortened the auto boot delay from 8 seconds to 2 seconds. Now > when the system reboots one or two of the chassis will hang at the > "load... " message. The serial shell will still echo carrage returns so > the processor is not completly down. We use rsh to download the vxworks > image from the sun workstation and it hangs at the rsh call in the boot > loader. > > The two changes were: > > 1) Changing from tornado 1.0 to 2.02 > 2) shorting the autoboot time. > > If I put in the old boot roms then everything works fine. I suspect > that the systems are closer in sync because the number of times through > the reboot countdown is less and the actual boot time is closer in sync. > > Using snoop to monitor the ethernet port I observed that the first > thing the targets do is transmit an ARP broadcast with their ip address. > The targets that boot sucessfully transmit a second ARP message with the > unix hosts IP address. They get a reply then start doing RSH commands > to download VxWorks. The chassis that failes goes silent after the > first ARP. > > Observation: > > 1) The broadcasts are very close together and there are collisions. > 2) The unix system is still trying to communicate with the target that > failes. Maybe that has something to do with it. There was on open > socket at the time the chassis was rebooted and it takes a while for > the unix system to give up. > > Has anyone else had this problem booting several targets at once? --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Sun Feb 17 04:03:11 2002 From: Vxworks Exploder Date: Sun Feb 17 04:03:14 PST 2002 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sun Feb 17 04:03:08 PST 2002 Subject: PCI card Subject: Re: What's vxImmrGet() function for? Subject: Monitoring Interrupt Stack? Subject: clock_gettime fails with write() but not fwrite Subject: Target server with serial backend freezing, trash wakes it up. Subject: Re: ipFilterHookAdd() Problem Subject: Re: multiple instances of a variable Subject: Re: Project-relative files? Subject: Re: CPU usage thru wind shell Subject: Re: Problem booting four systems at once over the network. Subject: Re: Monitoring Interrupt Stack? Subject: Re: Project-relative files? Subject: Re: Monitoring Interrupt Stack? Subject: Re: prjObjs.lst - error 0x80 ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: PCI card Date: Sat, 16 Feb 2002 12:52:34 GMT From: "Michael B." Organization: [Infostrada] Message-ID: Hi! I'm working on a I/O PCI card and using pciHeaderShow I obtained a base address 0 (BAR0) equal to 0xe4000000 and some other items (are they important?). How can I use BAR0 to access to card's register? What about I/O mapped and/or memory mapped devices? My BSP is a default PcPentium. Thanks. Mic. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: What's vxImmrGet() function for? Date: Sat, 16 Feb 2002 09:43:03 -0800 From: "Michael R. Kesti" Organization: MK Associates Message-ID: <3C6E9A27.A6860E37@gv.net> References: Reply-To: mkesti@gv.net Gary M wrote: >It's vxImmrGet(), not vxGetImmr(). All it does is return the base of the SIU >memory space, Specifically, it returns the IMMR (Internal Memory Map Register) with the PARTNUM and MASKNUM fields mask to zero. > typically 0xFF000000. Yes, typically, but it can be located in other places and that is why it is best to use the function rather than asumming its location. - -- ======================================================================== Michael Kesti | "And like, one and one don't make | two, one and one make one." mkesti@gv.net | - The Who, Bargain --------------------------- Newsgroups: comp.os.vxworks Subject: Monitoring Interrupt Stack? Date: Sat, 16 Feb 2002 21:58:17 GMT From: "Chad" Organization: Excite@Home - The Leader in Broadband http://home.com/faster Message-ID: Can anybody point me to a facility for checking/monitoring the interrupt stack? --------------------------- Newsgroups: comp.os.vxworks Subject: clock_gettime fails with write() but not fwrite Date: Sat, 16 Feb 2002 23:00:18 GMT From: "Rory O'Connor" Organization: AT&T Broadband Message-ID: <6wBb8.3387$zl4.2197904@typhoon3.we.ipsvc.net> Hi, I am using Tornado 2.1 and I can't use the function "clock_gettime" with "write()" to a SCSI harddrive. For some reason, the write tampers with the clock counter and messes up clock_gettime. I don't see this behavior with "fwrite()". Does anyone know a work around so that I can use "write" and also measure it performance with "clock_gettime"? --------------------------- Newsgroups: comp.os.vxworks Subject: Target server with serial backend freezing, trash wakes it up. Date: 16 Feb 2002 16:19:49 -0800 From: eric.e.lesch@lmco.com (Eric Lesch) Organization: http://groups.google.com/ Message-ID: Hello, Hosts: Several Win2000 machines with T2.0.2 Target: PPC603, custom hardware. No ethernet. Backend: Serial 57600bps The target server comes up fine and connects to target. A target shell can be opened and commands can be issued with an occasional delay of several seconds. The debugger can be used briefly with some occasional delays and then I get a WTX timeout in tornado. The target server window displays no errors. No tools can communicate with the target server. If the target server is shutdown then I still have to kill its process in the task manager and then unregister it. When I restart the target server it reconnects even though the target has not been rebooted. Target side seems ok, but host side target server seems to have gotten hung. I have been able to extend the period of debugging because I discovered that, after the target server gets hung, I can use the console to force some trash down the debug channel to wakeup the target server and everything comes back for a little bit longer before hanging again. I have been able to force this to happen on a MVME2603 by repeating the "i" command very quickly at the prompt. I am convinced that the target server is where the problem exists since the target can be reconnected to again after killing, unregistering and restarting the target server. Also, why would sending trash cause it to continue working? Even with sending trash eventually is freezes for good. Our target does not support hardware flow control, is it needed? Anyways, I appreciate any help that anybody can provide. Thanks, Eric Lesch --------------------------- Newsgroups: comp.os.vxworks Subject: Re: ipFilterHookAdd() Problem Date: Sun, 17 Feb 2002 00:52:32 GMT From: "drdiags" Organization: AT&T Broadband Message-ID: References: <3C6D11AD.385877BC@fallschurch.esys.com> Hello John, I found this for ipFilterHookAdd(): STATUS ipFilterHookAdd ( FUNCPTR ipFilterHook /* routine to receive raw ip packets */ ) from what I can tell the ipFilterHook routine should expect four arguments STATUS ipFilterHook (struct ifnet, struct mbuf **, struct ip **, int len ); I am guessing on the first parameter. But here is what is being passed: m->m_pkthdr.rcvif I thought I saw on Wind Surf an example of using a ipFilterHook. Is this feature being deprecated? Maybe......... "John Sestak" wrote in message news:3C6D11AD.385877BC@fallschurch.esys.com... > Hi all, > Well basically what I'm trying to do is allow my host to receive > tunneled IP packets. The original IP header contains a protocol ID of > 36 (which is for XTP) but we have to tunnel this original packet inside > another packet than has a IP header with a protocol ID of 4 (IP > Multicast). So, the packet essential arrives at my host looking like: > > [Ethernet header] [New IP Header (prot 4)] [Original IP header (prot > 36)] [ XTP header] [Data] > > The way I tried to fix this problem is by using the 'ipFilterHookAdd()' > function to replace the new IP headers' protocol ID of 4 with a protocol > ID of 36 and stamp the first 20 bytes where the original IP header > (protocol 36) used to be with a special sequence of bytes indicating > that this packet was originally a tunneled packet and so I should skip > the first 20 bytes where the original IP header was and start reading > the XTP header from there. This has caused two problems. > > 1) It seems as though if I try to run anymore than a few instructions in > the ipFilterHook function that gets called when a packet arrives, the > program will crash. Originally, I was going to check all the fields of > the original IP header to make sure it was valid, and do a few other > things, etc. but it always crashed that way. When I scaled it down to > just replacing the protocol ID of 4 with a 36 and stamping the first 20 > bytes of the original IP header with a special sequence of bytes then it > seems to run ok but, this it only runs ok when I'm running a very SMALL > test application. Once I try it with a larger application (multiple > tasks running at once) that sends and receives on many different sockets > it dies. Essentially, what I'm saying is that the hook function doesn't > seem to work well if it stays in this function for too long or if the > processor is under a heavy load. Any ideas as to why??? > > 2) First of all, I can't seem to find a *.h file anywhere in the Tornado > directory that has a function prototype for ipFilterHookAdd() so when I > compile obviously the compiler says 'ipFilterHookAdd() implicitly > declared'. I tried running the program even with the warning that it > gave and it seemed to work with a small test program I wrote. So, I > then tried adding my own function prototype to get rid of the compiler > warning but when I ran it, the program crashes. Where in the world is > the function prototype for this 'ipFilterHookAdd()' function??? I found > a *.h file that had a function prototype for 'ipFilterHookInit()' but > not ipFilterHookAdd(). > > Any ideas as to how to fix this hook function or maybe another way I > could receive tunneled packets (i.e. grab the tunneled, protocol 4 > packets and change them to protocol 36 packets so that the raw socket > listening for protocol 36 can receive these packets)??? > > Any help would be much appreciated. Thank you. > > John > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: multiple instances of a variable Date: Sun, 17 Feb 2002 03:36:21 GMT From: Pierre Frigon Organization: AT&T Broadband Message-ID: <3C6F245A.F78D2A82@yahoo.com> References: <3C63EA94.E47EFE94@res.raytheon.com> There is only one instance of a global variable. The exception is if you have the same global variable defined in more than one object file and all these are dynamically linked (downloaded). In this case you will have more than one instance of the same global. A clean way of doing this is to use task variables under VxWorks 5.3.x. I beleive they are documented in taskVarLib. willisn wrote: > > Hi all, > > My question concerns multiple instances of a single global variable. > > I have 3 tasks running, all have an instance of a global variable (say > an integer). > > If I want to know the current value of this integer for any one of the > tasks, I need to do a lkup on the variable and then dereference the > address associated with the task I am interested in. > > My question is, if you simply type the name of the variable in the > console, how does vxworks decide which instance of the variable to give > you? > > Thanks in advance for any help! > > Nick --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Project-relative files? Date: Sun, 17 Feb 2002 03:46:21 GMT From: Pierre Frigon Organization: AT&T Broadband Message-ID: <3C6F26B2.60F64B0E@yahoo.com> References: Can't be done unless you edit the wpj files and makefiles by hand. Note that you will have to do this every time you change the project. So we traded the project facility GUI for a text editor. It worked nicely ever since. Bob Bradley wrote: > > Is it possible to convince Tornado to reference files by relative path > that are above the project file in the hierarchy when added to the > project? It seems to use project-relative paths when the files are below > it in the hierarchy, but uses full paths for files above it. For > example, I might have a hierarchy as follows: > > Sources > Make > MyProject.wpj > Misc > MyFile.c > > When I add "MyFile.c" to "MyProject.wpj", I'd like it to be saved in the > project file (and the resulting makefile) as > "$(PRJ_DIR)../Misc/MyFile.c" instead of using the full path to MyFile.c. > Our project is built on a variety of systems so using a full path > doesn't work. > > Is this possible? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: CPU usage thru wind shell Date: Sun, 17 Feb 2002 03:51:33 GMT From: Pierre Frigon Organization: AT&T Broadband Message-ID: <3C6F27EA.67012792@yahoo.com> References: <3C6D7014.A91B6C5E@lucent.com> I beleive you can do this using the shell. Take a look at the documentation for spyLib. Chandwani wrote: > > Hi ! > > Is there a way to see the CPU usage ( on per task basis ) by a wind shell > command ? I know about the RTI tools like "Profilescope" etc. but I am having > tough time running such data intensive GUI on my target boards ( ppcec603 arch. > ). > > Any commands from windshell giving even the cursory data will help as my target > board is becoming sluggish after some time and is also tasks are leaking memory. > > Thanks, > -- > Rakesh Chandwani > Lambda Router AOS SW Dev.| Lucent Technologies, > Room 3N-320 | 101 Crawfords Corner Road, > Ph:732-949-0576 | Holmdel, NJ - 07748 --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Problem booting four systems at once over the network. Date: Sun, 17 Feb 2002 03:54:37 GMT From: Pierre Frigon Organization: AT&T Broadband Message-ID: <3C6F28A2.5CD55C74@yahoo.com> References: Have you tried using ftp instead of rsh? Thomas O Mehrkam wrote: > > We have A system made up of four VME chassis and one Unix system, > The VME chassis each contain one mvme167 Computer running Tornado 2.02 > patched to combined patch level 2. The Unix system is a Sun Ultra I > system running Solaris 2.7. > > The network connection is 10 base tx ethernet connected to an > HP hub. The interface on the Sun system is a qe type interface. > > If the real time system goes out of sync then we have a method of > resetting all four of the VME chassis. The chassis reboot and reconnect > to the GUI on the unix system. This system has worked this way for > about 5 years with 20-30 systems fielded. In the latest release > of the software we upgraded from Tornado 1.0 to Tornado 2.02. We also > shortened the auto boot delay from 8 seconds to 2 seconds. Now > when the system reboots one or two of the chassis will hang at the > "load... " message. The serial shell will still echo carrage returns so > the processor is not completly down. We use rsh to download the vxworks > image from the sun workstation and it hangs at the rsh call in the boot > loader. > > The two changes were: > > 1) Changing from tornado 1.0 to 2.02 > 2) shorting the autoboot time. > > If I put in the old boot roms then everything works fine. I suspect > that the systems are closer in sync because the number of times through > the reboot countdown is less and the actual boot time is closer in sync. > > Using snoop to monitor the ethernet port I observed that the first > thing the targets do is transmit an ARP broadcast with their ip address. > The targets that boot sucessfully transmit a second ARP message with the > unix hosts IP address. They get a reply then start doing RSH commands > to download VxWorks. The chassis that failes goes silent after the > first ARP. > > Observation: > > 1) The broadcasts are very close together and there are collisions. > 2) The unix system is still trying to communicate with the target that > failes. Maybe that has something to do with it. There was on open > socket at the time the chassis was rebooted and it takes a while for > the unix system to give up. > > Has anyone else had this problem booting several targets at once? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Monitoring Interrupt Stack? Date: Sun, 17 Feb 2002 03:56:01 GMT From: Pierre Frigon Organization: AT&T Broadband Message-ID: <3C6F28F6.44FFC43A@yahoo.com> References: The routine "checkStack" allows you to monitor the stack usage. You can call this routine from the shell with no arguments. Chad wrote: > > Can anybody point me to a facility for checking/monitoring the interrupt > stack? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Project-relative files? Date: Sun, 17 Feb 2002 09:29:32 +0100 From: "Martin Raabe" Organization: Wind River GmbH Message-ID: References: Hello Bob, use SNiFF+ 4.0.2 for project management. Ciao Martin Raabe "Bob Bradley" schrieb im Newsbeitrag news:bob-D507E6.01055214022002@news.apple.com... > Is it possible to convince Tornado to reference files by relative path > that are above the project file in the hierarchy when added to the > project? It seems to use project-relative paths when the files are below > it in the hierarchy, but uses full paths for files above it. For > example, I might have a hierarchy as follows: > > Sources > Make > MyProject.wpj > Misc > MyFile.c > > When I add "MyFile.c" to "MyProject.wpj", I'd like it to be saved in the > project file (and the resulting makefile) as > "$(PRJ_DIR)../Misc/MyFile.c" instead of using the full path to MyFile.c. > Our project is built on a variety of systems so using a full path > doesn't work. > > Is this possible? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Monitoring Interrupt Stack? Date: Sun, 17 Feb 2002 11:43:18 -0000 From: "cecil" Message-ID: <1013946407.21382.0.nnrp-10.d4e4b8d1@news.demon.co.uk> References: <3C6F28F6.44FFC43A@yahoo.com> Be aware that if your running on a Intel platform then vxworks does not use the interrupt stack, it uses the stack of the interrupted task. This been the case, you will see the margin grow less when the machine has been running for a few hours. It is therefore advisable to give your tasks more stack than seems necessary... Nige Pierre Frigon wrote in message <3C6F28F6.44FFC43A@yahoo.com>... >The routine "checkStack" allows you to monitor the stack usage. You can >call this routine from the shell with no arguments. > >Chad wrote: >> >> Can anybody point me to a facility for checking/monitoring the interrupt >> stack? --------------------------- Newsgroups: comp.os.vxworks,gnu.gcc.help Subject: Re: prjObjs.lst - error 0x80 Date: Sun, 17 Feb 2002 13:50:40 +0200 From: "Ofer Goren" Organization: Verio Message-ID: References: <3c6c0065@pull.gecm.com> do you have cygwin installed? if so, check for existance of cygwin.dll. We are starting to believe that there is some connection. If you can't find the file, try to restore it from another PC, and place it where it should be. Let me know if it worked... - -- Ofer Goren Iamba Technologies "Martin Dowie" wrote in message news:3c6c0065@pull.gecm.com... > > I have a strange problem, when compiling a project that worked well a day > > ago. > > When I compile, I get an make: *** /prjObjs.lst Error 0x80 > > > > I tried creating new project, under a new worksapce, with no files, the > > works. Nothing helped. > > > > No makefile under ../target/h/make/ was modified during the last 6 > months... > > > > Anyway, any help, before re-installing Tornado (and all other utilities > > after) ... :(( > > I had the same problem and reported it to Wind River (ref 268815). They > had never seen this before, couldn't reproduce it. Every suggestion > they came up with didn't work. In the end they said to re-install > everything. As we out-source our IT, this hasn't happened yet! :-( > > If you do re-install tornado and it "cures" this problem, could > you let me know? > > Cheers, > Martin > > > --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Mon Feb 18 04:03:08 2002 From: Vxworks Exploder Date: Mon Feb 18 04:03:11 PST 2002 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Mon Feb 18 04:03:05 PST 2002 Subject: Re: Monitoring Interrupt Stack? Subject: Re: Monitoring Interrupt Stack? Subject: Major dosFs2 problem with Hard Drive Subject: Re: Target server with serial backend freezing, trash wakes it up. Subject: Re: Serial Driver Lockup? (HTML) Subject: Re: Question:Capturing stack trace of a taskwhen it crashes Subject: Launching Shell from Command line Subject: Re: Major dosFs2 problem with Hard Drive Subject: Re: pap and secret files Subject: Re: Launching Shell from Command line ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Monitoring Interrupt Stack? Date: 17 Feb 2002 07:56:18 -0800 From: pkockritz@mac.com (Pete Kockritz) Organization: http://groups.google.com/ Message-ID: <9423c55b.0202170756.21cf556e@posting.google.com> References: <3C6F28F6.44FFC43A@yahoo.com> <1013946407.21382.0.nnrp-10.d4e4b8d1@news.demon.co.uk> "cecil" wrote in message news:<1013946407.21382.0.nnrp-10.d4e4b8d1@news.demon.co.uk>... > Be aware that if your running on a Intel platform then vxworks does not use > the interrupt stack, it uses the stack of the interrupted task. This been > the case, you will see the margin grow less when the machine has been > running for a few hours. It is therefore advisable to give your tasks more > stack than seems necessary... Also, on other targets (PPC for example), depending on what you do in the interrupt some work may get deffered and run in the context of _any_ task. So it is safest to keep the stack requirements of your ISRs small, and to make sure you have enough margin on all your task's stacks for the ISR to execute. You may also need to take into account interrupt nesting. The windSh version of checkStack doesn't show the interrupt stack usage. You can use the target shell or force the windsh to use the target resident version of checkStack with '@checkStack'. Pete Kockritz - -- pkockritz at mac dot com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Monitoring Interrupt Stack? Date: Sun, 17 Feb 2002 19:52:07 GMT From: "Chad" Organization: Excite@Home - The Leader in Broadband http://home.com/faster Message-ID: References: <3C6F28F6.44FFC43A@yahoo.com> <1013946407.21382.0.nnrp-10.d4e4b8d1@news.demon.co.uk> <9423c55b.0202170756.21cf556e@posting.google.com> Thanks all! I'm using a PPC, but was unaware of the @checkStack trick (or that the two versions differed at all). Some great info here. "Pete Kockritz" wrote in message news:9423c55b.0202170756.21cf556e@posting.google.com... > "cecil" wrote in message news:<1013946407.21382.0.nnrp-10.d4e4b8d1@news.demon.co.uk>... > > Be aware that if your running on a Intel platform then vxworks does not use > > the interrupt stack, it uses the stack of the interrupted task. This been > > the case, you will see the margin grow less when the machine has been > > running for a few hours. It is therefore advisable to give your tasks more > > stack than seems necessary... > > Also, on other targets (PPC for example), depending on what you do in > the interrupt some work may get deffered and run in the context of > _any_ task. So it is safest to keep the stack requirements of your > ISRs small, and to make sure you have enough margin on all your task's > stacks for the ISR to execute. You may also need to take into account > interrupt nesting. > > The windSh version of checkStack doesn't show the interrupt stack > usage. You can use the target shell or force the windsh to use the > target resident version of checkStack with '@checkStack'. > > Pete Kockritz > -- > pkockritz at mac dot com --------------------------- Newsgroups: comp.os.vxworks Subject: Major dosFs2 problem with Hard Drive Date: 17 Feb 2002 12:18:04 -0800 From: ilevi@rugged.com (Ilan) Organization: http://groups.google.com/ Message-ID: <2d7af320.0202171218.65f9ceaf@posting.google.com> Hi, In am working on 80Gb Hard drive ("big drive") and In my application I have task that open files for writing ,write 30 Mb of data and close the files and so on in while loop for more files. after I run the chkdsk function I see that I have a lot error messages as below for example: "/AT3/mov493.dat" file larger then its chain of clusters Entry will be deleted. "/AT3 - lost chains " has illegal cluster in chain "/AT3 - lost chains " cross-linked with another chain I initialize my dosFs system in that way with the option DOS_OPT_CHANGENOWARN: if (pVolDesc=dosFsDevInit(bootDir, pBlkDev, NULL)) == (DOS_VOL_DESC *) NULL) { printErr ("Error during dosFsDevInit: %x\n", errno); return (ERROR); } if(dosFsVolOptionsSetpVolDesc,DOS_OPT_CHANGENOWARN)==ERROR) { perror("dosFsVolOptionsSet"); } I as well call the ioctl with the options FIOSYNC or FIOFLUSH after every write to take care of the synchronization (FAT32) but it doesn't help. The HD is ATA/100 (Lba). the printing of the function dosFsConfigShow() is: current volume configuration: - volume label: NO LABEL ; (in boot sector: ) - volume Id: 0x54414438 - total number of sectors: 156,301,488 - bytes per sector: 512 - # of sectors per cluster: 128 - # of reserved sectors: 32 - FAT entry size: FAT32 - # of sectors per FAT copy: 9,540 - # of FAT table copies: 2 - # of hidden sectors: 0 - first cluster is in sector # 19,112 - directory structure: VFAT - root dir start cluster: 2 FAT handler information: - ------------------------ - allocation group size: 123 clusters - free space on volume: 76,867,371,046 bytes I will be glad if someone have an idea how to solve this problem. this is very important. Thank you very much, Ilan --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Target server with serial backend freezing, trash wakes it up. Date: Sun, 17 Feb 2002 20:59:22 GMT From: "Iosif Harutyunov" Organization: AT&T Broadband Message-ID: References: I am experiencing similar problems with my W2K Target server connected to x86 target using serial channel. I tried to play with serial channel speed and WDB timeouts with no success. BTW, problem not persistent. Sometime debugger session lasts about 10-15 minutes before it hung, sometime I could not get any response (except WTX timeout) from very first 'i' command issued and shell prompt. Does anybody experience similar problems with different OS (NT, 98)? Thank you. Iosif,_ "Eric Lesch" wrote in message news:e3d1aa11.0202161619.27f7832f@posting.google.com... > Hello, > > Hosts: Several Win2000 machines with T2.0.2 > Target: PPC603, custom hardware. No ethernet. > Backend: Serial 57600bps > > The target server comes up fine and connects to target. A target > shell can be opened and commands can be issued with an occasional > delay of several seconds. The debugger can be used briefly with some > occasional delays and then I get a WTX timeout in tornado. The target > server window displays no errors. No tools can communicate with the > target server. If the target server is shutdown then I still have to > kill its process in the task manager and then unregister it. When I > restart the target server it reconnects even though the target has not > been rebooted. Target side seems ok, but host side target server > seems to have gotten hung. I have been able to extend the period of > debugging because I discovered that, after the target server gets > hung, I can use the console to force some trash down the debug channel > to wakeup the target server and everything comes back for a little bit > longer before hanging again. > > I have been able to force this to happen on a MVME2603 by repeating > the "i" command very quickly at the prompt. > > I am convinced that the target server is where the problem exists > since the target can be reconnected to again after killing, > unregistering and restarting the target server. Also, why would > sending trash cause it to continue working? Even with sending trash > eventually is freezes for good. > > Our target does not support hardware flow control, is it needed? > > Anyways, I appreciate any help that anybody can provide. > Thanks, > Eric Lesch --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Serial Driver Lockup? (HTML) Date: Sun, 17 Feb 2002 21:57:27 GMT From: "Chad" Organization: Excite@Home - The Leader in Broadband http://home.com/faster Message-ID: References: This problem is still persisting. It appears to be related to the semaphores taken in tyWrite and/or tyRead for synchronisation. This semaphores can be taken during a read but not released in the case that a buffer is full in the case of a write, but is then released when the driver empties the buffer past a threshold. The only explanation I can find for my problem is that the semaphore isn't being release properly. Here's some details I logged when the system was last locked: version VxWorks (for MPC823) version 5.4. Kernel: WIND version 2.5 Made on Feb 4 2002, 14:54:58. GC4000. value = 0 = 0x0 - -> i NAME ENTRY TID PRI STATUS PC SP ERRNO DELAY - ---------- ------------ -------- --- ---------- ---- --- ---------- ----- MyTask Task::do_mai ffddb0 99 PEND 18b74c ffd860 3006b 0 value = 0 = 0x0 - -> tt 0xffddb0 2b2a4c write__2IOPCvi +d4 : write ([]) 166e34 write +10 : iosWrite ([]) 168398 iosWrite +cc : tyWrite ([]) 19ae0c tyWrite +5c : semTake ([]) 18c288 semTake +134: semBTake ([]) value = 0 = 0x0 - -> d 0xffd860 00ffd860: 00ff d878 0018 b74c 0000 0000 0000 00b4 *...x...L........* 00ffd870: ffff ffff 00fe af04 00ff d890 0018 c288 *................* ** note: presumed 0xfeaf04 to be the SEM_ID because the semTake() call takes both the ID and WAIT_FOREVER (-1 or 0xfffffff) as arguments - -> semShow 0xfeaf04,1 Semaphore Id : 0xfeaf04 Semaphore Type : BINARY Task Queueing : PRIORITY Pended Tasks : 1 State : EMPTY ** note: semaphore empty. This basically has to be one or another of the two semaphores dealt with in tyWrite(). How do I tell which one?? semTake (&pTyDev->wrtSyncSem, WAIT_FOREVER); OR semTake (&pTyDev->mutexSem, WAIT_FOREVER); ** however, it's probably the wrtSyncSem since there is a condition where a write does not give it back after its call is complete, the mutex is ALWAYS given back. if (rngFreeBytes (pTyDev->wrtBuf) > 0) semGive (&pTyDev->wrtSyncSem); semGive (&pTyDev->mutexSem); NAME TID PRI TIMEOUT - ---------- -------- --- ------- MyTask ffddb0 99 0 value = 0 = 0x0 - -> d 0xfeaf04 (SEMAPHORE struct) 00feaf00: 001e 1b80 0001 0000 00ff ddb0 * ............* 00feaf10: 00ff ddb0 0000 0000 001d 5188 00ff ddb0 *..........Q.....* 00feaf20: 0001 0000 0000 0000 0004 0000 0017 0ae0 *................* typedef struct semaphore /* SEMAPHORE */ { OBJ_CORE objCore; /* 0x00: object management */ UINT8 semType; UINT8 options; UINT16 recurse; Q_HEAD qHead; union { UINT count; struct windTcb *owner; } state; } SEMAPHORE; I also noticed that doing a semGive() on the command line did not seem to kick things back into action. Another fellow I have been corresponding with (who is experiencing the same condition) said that the tyShow() function [which I don't seem to have?] shows the ty device ring buffer and it is, in fact, full. I'm not sure how I can verify this on my end. It would explain why the synch semaphore was not given back, so the question that remains is why this call: /* when we pass the write threshhold, give write synchonization * and release tasks pended in select */ if (rngFreeBytes (ringId) == tyWrtThreshold) { semGive (&pTyDev->wrtSyncSem); if (_func_selWakeupAll != NULL) (* _func_selWakeupAll) (&pTyDev->selWakeupList, SELWRITE); } } At the end of tyITx() did not seem to give it back either, or perhaps it didn't have a chance as the interrupts stopped being raised to clear out the queue?? Perhaps its because of the low interrupt priority of the SMC ports on the 823? We do have timers in the system, I am looking into whether those all clear their interrupt pending bits etc properly so as to allow lower priority interrupts to be processed. Thats about where I'm at now. Either my tx interrupts are dying from time to time (or being starved) and the driver doesn't recover, or somebody is failing a call to get a char from the ty buffer. The latter might indicate a problem with this macro call in tyITx(): /* check for more characters to output */ else if (RNG_ELEM_GET (ringId, pChar, nn) == 0) /* no more chars */ pTyDev->wrtState.busy = FALSE; else The macro is defined as (in rngLib.h): /*************************************************************************** * * RNG_ELEM_GET - get one character from a ring buffer * * This macro gets a single character from the specified ring buffer. * Must supply temporary variable (register int) 'fromP'. * * RETURNS: 1 if there was a char in the buffer to return, 0 otherwise * * NOMANUAL */ #define RNG_ELEM_GET(ringId, pCh, fromP) \ ( \ fromP = (ringId)->pFromBuf, \ ((ringId)->pToBuf == fromP) ? \ 0 \ : \ ( \ *pCh = (ringId)->buf[fromP], \ (ringId)->pFromBuf = ((++fromP == (ringId)->bufSize) ? 0 : fromP), \ 1 \ ) \ ) If anybody could shed any further insite into this problem, or perhaps point out some facilities I could use to get a better look at the relevant state of the system, I would appreciate it and shower you in praise. - -Chad --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Question:Capturing stack trace of a taskwhen it crashes Date: 17 Feb 2002 21:56:38 -0800 From: mistry_amit@yahoo.com (Amit) Organization: http://groups.google.com/ Message-ID: <3ca555c5.0202172156.10963c11@posting.google.com> References: <396b9cc7.0202140719.7109204c@posting.google.com> sunil@dharwad.com (Sunil K) wrote in message news:<396b9cc7.0202140719.7109204c@posting.google.com>... > Hi VxWorkers, > > I need to capture stack trace in a file when any task crashes. > (Even if the crash leads to a reboot of vxWorks) > > Any ideas how to do it ? Should I replcae the default interupt > handlers that suspend the task when it does any illegal operation ? > How should I go about it ? > > Your help is greatly appreciated. > > Thanks! > Sunil Of course, you can do with the help of WindView configuring in Post-Mortem mode. post-mortem mode is helpful in collecting data on an application/task failure resulting in a target reboot. You will get more information about post-mortem mode in windview help. Amit --------------------------- Newsgroups: comp.os.vxworks Subject: Launching Shell from Command line Date: 17 Feb 2002 22:11:56 -0800 From: kumar_amit@delhi.tcs.co.in (Amit) Organization: http://groups.google.com/ Message-ID: <7421c9fd.0202172211.534ae489@posting.google.com> Hi, I want to launch windsh from command prompt in win32 environment. Can I do the same. If this is possible, then can I compile and download my project from the same shell itself. Any help will be greatly appreciated. Regards, Amit. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Major dosFs2 problem with Hard Drive Date: Mon, 18 Feb 2002 09:29:43 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <2d7af320.0202171218.65f9ceaf@posting.google.com> Ilan, I would recommend that you first make sure that you are using a version of DosFs2.0 which has all the lastest patches. Also, you should search the archives of this newsgroup, e.g. on Google Groups, and you will find a couple of known problems and their solution. My best guess is that you are using PPC604, and the problem you experience is due to the compier improper handling of 64-bit integer arithmetics, which is used extensively throughout dosFs2.0. For that problem, there is a binary patch from WIndSurf or a source work around (if you got dosFs2.0 source code). Send my regards to both Davids and Daniel. - - Leonid "Ilan" wrote in message news:2d7af320.0202171218.65f9ceaf@posting.google.com... > Hi, > > In am working on 80Gb Hard drive ("big drive") and In my application I > have task that open files for writing ,write 30 Mb of data and close > the files and so on in while loop for more files. > after I run the chkdsk function I see that I have a lot error messages > as below for example: > > "/AT3/mov493.dat" file larger then its chain of clusters Entry will be > deleted. > "/AT3 - lost chains " has illegal cluster in chain > "/AT3 - lost chains " cross-linked with another chain > > I initialize my dosFs system in that way with the option > DOS_OPT_CHANGENOWARN: > > if (pVolDesc=dosFsDevInit(bootDir, pBlkDev, NULL)) == > (DOS_VOL_DESC *) NULL) > { > printErr ("Error during dosFsDevInit: %x\n", errno); > return (ERROR); > } > > if(dosFsVolOptionsSetpVolDesc,DOS_OPT_CHANGENOWARN)==ERROR) > { > perror("dosFsVolOptionsSet"); > } > > I as well call the ioctl with the options FIOSYNC or FIOFLUSH after > every write > to take care of the synchronization (FAT32) but it doesn't help. > The HD is ATA/100 (Lba). > > the printing of the function dosFsConfigShow() is: > > current volume configuration: > - volume label: NO LABEL ; (in boot sector: ) > - volume Id: 0x54414438 > - total number of sectors: 156,301,488 > - bytes per sector: 512 > - # of sectors per cluster: 128 > - # of reserved sectors: 32 > - FAT entry size: FAT32 > - # of sectors per FAT copy: 9,540 > - # of FAT table copies: 2 > - # of hidden sectors: 0 > - first cluster is in sector # 19,112 > - directory structure: VFAT > - root dir start cluster: 2 > > FAT handler information: > ------------------------ > - allocation group size: 123 clusters > - free space on volume: 76,867,371,046 bytes > > I will be glad if someone have an idea how to solve this problem. > this is very important. > > Thank you very much, > Ilan --------------------------- Newsgroups: comp.os.vxworks Subject: Re: pap and secret files Date: 17 Feb 2002 23:49:03 -0800 From: Scott@svis.com (Scott A. Vigil) Organization: http://groups.google.com/ Message-ID: References: "alain tamdem" wrote in message news:... > Hi Iam trying to use the secret file to specify in me ppp options files how > to ge the secret files, > but its still not working... > can somebody tell me how to use pppSecretAdd() , I mean the prototype... These were the steps I used to configure the password and the PPP secret for an experiment. I place them in the context some of the other major steps since they have to be in agreement. 1. Create the encrypted version of YourPassword F> vxencrypt please enter password: YourPassword encrypted password is cRQyS9dcQQ 2. Place the encrypted password (cRQyS9dcQQ in this case) into RLOGIN/Telnet password configuration box in the VxWorks tab of the Tornado Workspace window. This would be found under "network componenets/networking protocols/network applications". Then, re-build image and download. 3. Add a password to the VxWorks password database. pppSecretAdd("*", "*", "YourPassword", ""); Scott Vigil Embedded Software Consultant Scott@svis.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Launching Shell from Command line Date: 18 Feb 2002 08:21:36 GMT From: Denis Perelyubskiy Message-ID: References: <7421c9fd.0202172211.534ae489@posting.google.com> hello, * Amit [02-Sun-02 22:11 -0800]: > > I want to launch windsh from command prompt in win32 > environment. Can I do the same. If this is possible, > then can I compile and download my project from the > same shell itself. Well, you do need to have the target server running. When you set it up (at least in T1, and I think in t2 as well) in that same dialog box on the bottom, there is a command-line string, which you can copy and paste into a shortcut. Then, starting this shortcut will start your shell. As far as compiling/downloading from the shell, I don't see how you could do that, because the shell actually runs on the target, not on the host. There is no compiler there, and whatnot. denis - -- if you want to reply by mail, please tweak my address in a reasonable way. sorry for the inconvenience. --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Tue Feb 19 02:52:01 2002 From: jw@ocean.ioa.ac.cn (Jiang wei) Date: Tue Feb 19 02:52:03 PST 2002 Subject: Hard driver problem in VxWorks This is a multi-part message in MIME format. ------=_NextPart_000_001B_01C1B976.E9ABE280 Content-Type: text/plain; charset="gb2312" Content-Transfer-Encoding: base64 SGVsbG8gYWxsLA0KDQpJJ20gdXNpbmcgVG9ybmFkbyAyLjAgd2l0aCBhIHg4NiB0YXJnZXQuIFRo ZXJlIGFyZSB0d28gaGFyZCBkaXNrcyBvbiBib2FyZCBhcyBmb2xsb3c6DQoxLiAgICBQcmltYXJ5 IE1hc3RlcjogICAgICAgIEZ1aml0c3UgICAgICAgICAgICAgICAgICAgIDExRyAgICANCjIuICAg IFNlY29uZGFyeSBNYXN0ZXI6ICAgIElERSBGbGFzaERyaXZlICAgICAxNTNNDQpXaGVuIEkgdHJ5 IHRvIHVzZSBhdGFEZXZDcmVhdGUoICksIGl0IGFsd2F5cyBmYWlscy4gDQoNCkkgaGF2ZSBpbmNs dWRlZCBhdGEgaGFyZCBkcml2ZXIgaW4gaW1hZ2UgY29uZmlndXJhdGlvbi4NCg0KQlRXLCB3aGF0 J3MgdGhlIGRpZmZlcmVuY2UgYmV0d2VlbiBBVEEgYW5kIElERSBoYXJkIGRpc2sgc3VwcG9ydD8N Cg0KVGhhbmtzIQ0KDQpUb255DQo= ------=_NextPart_000_001B_01C1B976.E9ABE280 Content-Type: text/html; charset="gb2312" Content-Transfer-Encoding: base64 PCFET0NUWVBFIEhUTUwgUFVCTElDICItLy9XM0MvL0RURCBIVE1MIDQuMCBUcmFuc2l0aW9uYWwv L0VOIj4NCjxIVE1MPjxIRUFEPg0KPE1FVEEgaHR0cC1lcXVpdj1Db250ZW50LVR5cGUgY29udGVu dD0idGV4dC9odG1sOyBjaGFyc2V0PWdiMjMxMiI+DQo8TUVUQSBjb250ZW50PSJNU0hUTUwgNS41 MC40NTIyLjE4MDAiIG5hbWU9R0VORVJBVE9SPg0KPFNUWUxFPjwvU1RZTEU+DQo8L0hFQUQ+DQo8 Qk9EWSBiZ0NvbG9yPSNmZmZmZmY+DQo8RElWPjxGT05UIGZhY2U9IlRpbWVzIE5ldyBSb21hbiI+ SGVsbG8gYWxsLDwvRk9OVD48L0RJVj4NCjxESVY+PEZPTlQgZmFjZT0iVGltZXMgTmV3IFJvbWFu Ij48L0ZPTlQ+Jm5ic3A7PC9ESVY+DQo8RElWPjxGT05UIGZhY2U9IlRpbWVzIE5ldyBSb21hbiI+ SSdtIHVzaW5nIFRvcm5hZG8gMi4wIHdpdGggYSB4ODYgdGFyZ2V0LiBUaGVyZSANCmFyZSB0d28g aGFyZCBkaXNrcyBvbiBib2FyZCBhcyBmb2xsb3c6PC9GT05UPjwvRElWPg0KPERJVj48Rk9OVCBm YWNlPSJUaW1lcyBOZXcgUm9tYW4iPjEuJm5ic3A7Jm5ic3A7Jm5ic3A7IFByaW1hcnkgDQpNYXN0 ZXI6Jm5ic3A7Jm5ic3A7Jm5ic3A7ICZuYnNwOyZuYnNwOyZuYnNwOyBGdWppdHN1Jm5ic3A7Jm5i c3A7Jm5ic3A7IA0KJm5ic3A7Jm5ic3A7Jm5ic3A7ICZuYnNwOyZuYnNwOyZuYnNwOyAmbmJzcDsm bmJzcDsmbmJzcDsgJm5ic3A7Jm5ic3A7Jm5ic3A7IA0KMTFHPC9GT05UPjxGT05UIHNpemU9Mj4m bmJzcDsmbmJzcDsmbmJzcDsgPC9GT05UPjwvRElWPg0KPERJVj48Rk9OVCBmYWNlPSJUaW1lcyBO ZXcgUm9tYW4iPjIuJm5ic3A7Jm5ic3A7Jm5ic3A7IFNlY29uZGFyeSANCk1hc3RlcjombmJzcDsm bmJzcDsmbmJzcDsgSURFIEZsYXNoRHJpdmUmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsgDQoxNTNN PC9GT05UPjwvRElWPg0KPERJVj48Rk9OVCBmYWNlPSJUaW1lcyBOZXcgUm9tYW4iPldoZW4gSSB0 cnkgdG8gdXNlIGF0YURldkNyZWF0ZSggKSwgaXQgYWx3YXlzIA0KZmFpbHMuIDwvRk9OVD48L0RJ Vj4NCjxESVY+PEZPTlQgZmFjZT0iVGltZXMgTmV3IFJvbWFuIj48L0ZPTlQ+Jm5ic3A7PC9ESVY+ DQo8RElWPjxGT05UIGZhY2U9IlRpbWVzIE5ldyBSb21hbiI+SSBoYXZlIGluY2x1ZGVkIGF0YSBo YXJkIGRyaXZlciBpbiZuYnNwO2ltYWdlIA0KY29uZmlndXJhdGlvbi48L0ZPTlQ+PC9ESVY+DQo8 RElWPjxGT05UIGZhY2U9IlRpbWVzIE5ldyBSb21hbiI+PC9GT05UPiZuYnNwOzwvRElWPg0KPERJ Vj48Rk9OVCBmYWNlPSJUaW1lcyBOZXcgUm9tYW4iPkJUVywgd2hhdCdzIHRoZSBkaWZmZXJlbmNl IGJldHdlZW4gQVRBIGFuZCBJREUgDQpoYXJkIGRpc2sgc3VwcG9ydD88L0ZPTlQ+PC9ESVY+DQo8 RElWPjxGT05UIGZhY2U9IlRpbWVzIE5ldyBSb21hbiI+PC9GT05UPiZuYnNwOzwvRElWPg0KPERJ Vj48Rk9OVCBmYWNlPSJUaW1lcyBOZXcgUm9tYW4iPlRoYW5rcyE8L0ZPTlQ+PC9ESVY+DQo8RElW PjxGT05UIGZhY2U9IlRpbWVzIE5ldyBSb21hbiI+PC9GT05UPiZuYnNwOzwvRElWPg0KPERJVj48 Rk9OVCBmYWNlPSJUaW1lcyBOZXcgUm9tYW4iPlRvbnk8L0ZPTlQ+PC9ESVY+PC9CT0RZPjwvSFRN TD4NCg== ------=_NextPart_000_001B_01C1B976.E9ABE280-- From vxwexplo-errs@csg.lbl.gov Tue Feb 19 04:00:45 2002 From: jw@ocean.ioa.ac.cn (Jiang wei) Date: Tue Feb 19 04:00:47 PST 2002 Subject: Hard driver problem in VxWorks Hello all, I'm using Tornado 2.0 with a x86 target. There are two hard disks on board as follow: 1. Primary Master: Fujitsu 11G 2. Secondary Master: IDE FlashDrive 153M When I try to use ataDevCreate( ), it always fails. I have included ata hard driver in image configuration. BTW, what's the difference between ATA and IDE hard disk support? Thanks! Tony From vxwexplo-errs@csg.lbl.gov Tue Feb 19 04:03:16 2002 From: Vxworks Exploder Date: Tue Feb 19 04:03:19 PST 2002 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Tue Feb 19 04:03:11 PST 2002 Subject: share flash chip between TFFS and registry? Subject: Re: share flash chip between TFFS and registry? Subject: Re: Target server with serial backend freezing, trash wakes it up. Subject: ANN: Embedded Systems Club / Ada UK conference & exhibition Subject: Re: share flash chip between TFFS and registry? Subject: Re: share flash chip between TFFS and registry? Subject: Re: Problem booting four systems at once over the network. Subject: Re: Serial Driver Lockup? (HTML) Subject: Re: system debug mode Subject: Re: Wind Web Server - problem in httpSymTabAdd Subject: Re: for PAP configuration Subject: Re: PCI card Subject: Re: clock_gettime fails with write() but not fwrite Subject: Re: Monitoring Interrupt Stack? Subject: Re: prjObjs.lst - error 0x80 Subject: Re: Reading from serial ports - Blocking or interrupts? Subject: Re: clock_gettime fails with write() but not fwrite Subject: Writing a BSP from scratch Subject: Re: anticipating machine check error... Subject: Re: Writing a BSP from scratch Subject: how to release a mutex when the process is killed Subject: problem during debbuging!! Subject: Re: how to release a mutex when the process is killed Subject: Re: how to release a mutex when the process is killed Subject: Is there a free database can be used on vxworks Subject: What means this sentens...? Subject: Re: What means this sentens...? Subject: Re: anticipating machine check error... Subject: MUX Load failed Subject: Re: anticipating machine check error... Subject: Re: Writing a BSP from scratch Subject: Re: anticipating machine check error... Subject: Re: anticipating machine check error... Subject: Re: prjObjs.lst - error 0x80 - solution Subject: Re: Question:Capturing stack trace of a taskwhen it crashes ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: share flash chip between TFFS and registry? Date: Mon, 18 Feb 2002 12:25:02 +0000 From: Owain Phillips Organization: Siemens Inc. Message-ID: <3C70F29E.AA0548FF@siemenscomms.co.uk> This is a multi-part message in MIME format. - --------------000553DB3BF173811A77027D Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Dear vxWorkers, I have the requirement that I want to have a flash filesystem and a simple registry in the same flash chip. Is TFFS geared up to this sharing of the device? The higher level filesystem drivers must ensure mutual exclusion to the drive by accessing a common low level driver. Where is this common driver? All the best, Owain - --------------000553DB3BF173811A77027D Content-Type: text/x-vcard; charset=us-ascii; name="owain.phillips.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Owain Phillips Content-Disposition: attachment; filename="owain.phillips.vcf" begin:vcard n:Phillips;Owain tel;fax:+44 115 943 4969 tel;work:+44 115 943 2167 x-mozilla-html:TRUE url:http://www.siemenscomms.co.uk org:Siemens Comunications Ltd version:2.1 email;internet:owain.phillips@siemenscomms.co.uk title:Software Engineer adr;quoted-printable:;;Technology Drive=0D=0ABeeston;Nottingham;;NG9 1LA;United Kingdom fn:Owain Phillips end:vcard - --------------000553DB3BF173811A77027D-- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: share flash chip between TFFS and registry? Date: Mon, 18 Feb 2002 15:01:22 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <3C70F29E.AA0548FF@siemenscomms.co.uk> TFFS has a feature to partition the Flash into a BootROM and TFFS areas. So it can make two partitions, and I doubt it can do more. So if you have a Boot area on the Flash, you'd better locate your registry (whatever that means) as a file or a group of files on top of the file system. Registry is typically an application-level database, while the Boot sector and Flash partitioning is the Hardware-Interface (aka BSP) section of expertise. Registry as file keeps the work properly assigned. "Owain Phillips" wrote in message news:3C70F29E.AA0548FF@siemenscomms.co.uk... > Dear vxWorkers, > > I have the requirement that I want to have a flash filesystem and a > simple registry in the same flash chip. Is TFFS geared up to this > sharing of the device? > > The higher level filesystem drivers must ensure mutual exclusion to the > drive by accessing a common low level driver. Where is this common > driver? > > All the best, > Owain > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Target server with serial backend freezing, trash wakes it up. Date: 18 Feb 2002 05:27:05 -0800 From: niall.leonard@ncr.com (Niall Leonard) Organization: http://groups.google.com/ Message-ID: <14e76c83.0202180527.4206bd62@posting.google.com> References: "Iosif Harutyunov" wrote in message news:... > I am experiencing similar problems with my W2K Target server connected to > x86 target using serial channel. I tried to play with serial channel > speed and WDB timeouts with no success. > BTW, problem not persistent. Sometime debugger session lasts about 10-15 > minutes before it hung, sometime I could not get any response (except WTX timeout) > from very first 'i' command issued and shell prompt. > Does anybody experience similar problems with different OS (NT, 98)? > > Thank you. > Iosif,_ > > "Eric Lesch" wrote in message news:e3d1aa11.0202161619.27f7832f@posting.google.com... > > Hello, > > > > Hosts: Several Win2000 machines with T2.0.2 > > Target: PPC603, custom hardware. No ethernet. > > Backend: Serial 57600bps > > > > The target server comes up fine and connects to target. A target > > shell can be opened and commands can be issued with an occasional > > delay of several seconds. The debugger can be used briefly with some > > occasional delays and then I get a WTX timeout in tornado. The target > > server window displays no errors. No tools can communicate with the > > target server. If the target server is shutdown then I still have to > > kill its process in the task manager and then unregister it. When I > > restart the target server it reconnects even though the target has not > > been rebooted. Target side seems ok, but host side target server > > seems to have gotten hung. I have been able to extend the period of > > debugging because I discovered that, after the target server gets > > hung, I can use the console to force some trash down the debug channel > > to wakeup the target server and everything comes back for a little bit > > longer before hanging again. > > > > I have been able to force this to happen on a MVME2603 by repeating > > the "i" command very quickly at the prompt. > > > > I am convinced that the target server is where the problem exists > > since the target can be reconnected to again after killing, > > unregistering and restarting the target server. Also, why would > > sending trash cause it to continue working? Even with sending trash > > eventually is freezes for good. > > > > Our target does not support hardware flow control, is it needed? > > > > Anyways, I appreciate any help that anybody can provide. > > Thanks, > > Eric Lesch I'm pretty sure this is due to a known problem with the serial debugger and win2K. There is a patch available to fix this - you may need to ask your WRS support contact for it. There is also a workaround which involves raising the priority of the target server executable to realtime. It works, but there is a snag - if your target reboots, the target server will start up again at normal priority and you will be back to square one ! Regards, Niall Leonard Consultant - Real Time Software, Platform Technology Development, NCR Financial Solutions Group Ltd, Discovery Centre, 3 Fulton Road, Dundee, DD2 4SW Tel: +44 (0)1382 592866 (Direct) e-mail: niall.leonard@ncr.com --------------------------- Newsgroups: comp.os.vxworks Subject: ANN: Embedded Systems Club / Ada UK conference & exhibition Date: Mon, 18 Feb 2002 14:08:05 +0000 From: John Robinson Organization: John Robinson & Associates Message-ID: We are pleased to announce a joint Embedded Systems Club/Ada UK conference to be held on March 14th 2002 in Swindon, UK. Attendance for members will cost 50 GBP (82 Euro). A full programme and registration forms can be found on the following web sites: http://www.AdaUK.org.uk http://www.EmbeddedSystemsClub.com The keynote presentation, given by Kevlin Henney of Curbralan, will pose the question: * "Are Agile Methods Applicable to Embedded Systems?". Papers include: * Developing High Integrity Systems with VxWorks and ACT GNAT Ada * Modern RTOS Architecture: Meeting Today's Demands * Real Time Operating Systems for use in Safety Critical Systems * Dynamic Storage Allocation with C++ in Embedded Systems * Putting Design Patterns to Use in a Real-Time, Embedded Design * UML 2.0 - The Next Step in the Embedded Development Mega-Trend? * Managing Safety Requirements: A Use Case Driven Approach * Conformity! A Practical Integration of Standards: A case for using UML with Ada 83 & Ada 95 * Proposed HOOD to UML Migration Process * HOOD and UML: You Can Use Both! * Does Software Engineering Have a Future? * Raven in Space * Have the Cake & Eat it Too: A State-of-the-art GNU IDE for Ada 95 * Designing for Testability: Testing an Ada "Stack" Class * SPARK - What Does the 'R' Stand For? * Helping You to Deliver Ada Applications to Your Customer's Requirements and Quality Standards * Right on Target: Target Testing Past, Present & Future For full details of the 18 papers and two panel sessions, either visit the websites or contact the Club & Event Manager Hazel Lawton at mailto:Hazel@Adaxia.com. Best regards John Robinson Conference Director mailto:John@JohnRobinsonAndAssociates.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: share flash chip between TFFS and registry? Date: Mon, 18 Feb 2002 14:24:05 +0000 (UTC) From: David Laight Organization: BT Openworld Message-ID: <3C710EDE.5080700@spamm.me.l8s.co.uk> References: <3C70F29E.AA0548FF@siemenscomms.co.uk> Leonid Rosenboim wrote: > TFFS has a feature to partition the Flash into a BootROM and TFFS > areas. So it can make two partitions, and I doubt it can do more. Hasn't it been said before that you cannot write to this reserved area while the TFFS FS is mounted? Generally causes problems trying to update the boot area... >>I have the requirement that I want to have a flash filesystem and a >>simple registry in the same flash chip. Is TFFS geared up to this >>sharing of the device? I did a 'simple' environment storage is a flash block using 3 bytes 'type' (usually printable) 1 byte 'length' (in 32 bit words) and variable length data in a single flash sector. The sector was erased to ff, if an entry had to be updated the 'type' was set to 0 (leaving the length intact). Dead space was deleted when it was getting too big. The advantage is that you can parse it very easily - eg in romInit.s >> >>The higher level filesystem drivers must ensure mutual exclusion to the >>drive by accessing a common low level driver. Where is this common >>driver? God knows - I didn't use TFFS, just a small write cache. David --------------------------- Newsgroups: comp.os.vxworks Subject: Re: share flash chip between TFFS and registry? Date: Mon, 18 Feb 2002 14:39:51 +0000 From: Owain Phillips Organization: Siemens Inc. Message-ID: <3C711237.507E7A97@siemenscomms.co.uk> References: <3C70F29E.AA0548FF@siemenscomms.co.uk> This is a multi-part message in MIME format. - --------------9C0F7EA79E61031DA5F793B6 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit We already have a record based storgage scheme that works pretty reliably; and have tried to merge in the TFFS. We cocked this up we added semaphores in the low level MTD; but this didn't take into acount that TFFS could read and sneak in behind the semaphore.... I just looked in the TFFS handbook and it looks as if you can use tffsRawio to go over the same TFFS drive as used for the TFFS filesystem. I just wondered if anyone had used these tffsRawio routines in this way. The docs say tffsBootImagePut uses this tffsRawio. But then again I heard of problems trying to use tffsBootImagePut at the same time as TFFS drive is mounted. Does not bode well!! All the best, Owain Leonid Rosenboim wrote: > TFFS has a feature to partition the Flash into a BootROM and TFFS > areas. So it can make two partitions, and I doubt it can do more. > So if you have a Boot area on the Flash, you'd better locate > your registry (whatever that means) as a file or a group of files > on top of the file system. > > Registry is typically an application-level database, while the Boot > sector and Flash partitioning is the Hardware-Interface (aka BSP) > section of expertise. Registry as file keeps the work properly assigned. > > "Owain Phillips" wrote in message > news:3C70F29E.AA0548FF@siemenscomms.co.uk... > > Dear vxWorkers, > > > > I have the requirement that I want to have a flash filesystem and a > > simple registry in the same flash chip. Is TFFS geared up to this > > sharing of the device? > > > > The higher level filesystem drivers must ensure mutual exclusion to the > > drive by accessing a common low level driver. Where is this common > > driver? > > > > All the best, > > Owain > > - --------------9C0F7EA79E61031DA5F793B6 Content-Type: text/x-vcard; charset=us-ascii; name="owain.phillips.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Owain Phillips Content-Disposition: attachment; filename="owain.phillips.vcf" begin:vcard n:Phillips;Owain tel;fax:+44 115 943 4969 tel;work:+44 115 943 2167 x-mozilla-html:TRUE url:http://www.siemenscomms.co.uk org:Siemens Comunications Ltd version:2.1 email;internet:owain.phillips@siemenscomms.co.uk title:Software Engineer adr;quoted-printable:;;Technology Drive=0D=0ABeeston;Nottingham;;NG9 1LA;United Kingdom fn:Owain Phillips end:vcard - --------------9C0F7EA79E61031DA5F793B6-- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Problem booting four systems at once over the network. Date: Mon, 18 Feb 2002 09:13:29 -0600 From: Thomas O Mehrkam Organization: Input/Output, Inc. Message-ID: <3C711A19.698F7A9E@wt.net> References: <3C6F28A2.5CD55C74@yahoo.com> Reply-To: mehrkam@wt.net No. The boot parameters are hard coded. The boot rom gets a chassis ID from the backpland and modified the boot paramaters based on the ID. We did not use FTP because we did not want to provide a password and rsh worked just fine. That is a possibility if I cannot get RSH to work. Thankyou T.O.M. Pierre Frigon wrote: > Have you tried using ftp instead of rsh? > > Thomas O Mehrkam wrote: > > > > We have A system made up of four VME chassis and one Unix system, > > The VME chassis each contain one mvme167 Computer running Tornado 2.02 > > patched to combined patch level 2. The Unix system is a Sun Ultra I > > system running Solaris 2.7. > > > > The network connection is 10 base tx ethernet connected to an > > HP hub. The interface on the Sun system is a qe type interface. > > > > If the real time system goes out of sync then we have a method of > > resetting all four of the VME chassis. The chassis reboot and reconnect > > to the GUI on the unix system. This system has worked this way for > > about 5 years with 20-30 systems fielded. In the latest release > > of the software we upgraded from Tornado 1.0 to Tornado 2.02. We also > > shortened the auto boot delay from 8 seconds to 2 seconds. Now > > when the system reboots one or two of the chassis will hang at the > > "load... " message. The serial shell will still echo carrage returns so > > the processor is not completly down. We use rsh to download the vxworks > > image from the sun workstation and it hangs at the rsh call in the boot > > loader. > > > > The two changes were: > > > > 1) Changing from tornado 1.0 to 2.02 > > 2) shorting the autoboot time. > > > > If I put in the old boot roms then everything works fine. I suspect > > that the systems are closer in sync because the number of times through > > the reboot countdown is less and the actual boot time is closer in sync. > > > > Using snoop to monitor the ethernet port I observed that the first > > thing the targets do is transmit an ARP broadcast with their ip address. > > The targets that boot sucessfully transmit a second ARP message with the > > unix hosts IP address. They get a reply then start doing RSH commands > > to download VxWorks. The chassis that failes goes silent after the > > first ARP. > > > > Observation: > > > > 1) The broadcasts are very close together and there are collisions. > > 2) The unix system is still trying to communicate with the target that > > failes. Maybe that has something to do with it. There was on open > > socket at the time the chassis was rebooted and it takes a while for > > the unix system to give up. > > > > Has anyone else had this problem booting several targets at once? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Serial Driver Lockup? (HTML) Date: Mon, 18 Feb 2002 11:36:22 -0500 From: david lindauer Message-ID: <3C712D86.46705085@notifier-is.net> References: it might be that your UART is locking up. Older uarts on IBM-PC compatibles had exactly this problem, that sometimes transmit interrupts would get lost. I'm having a problem now where newer UARTs will sometimes lock up and not receive data. Before I dug too much further into the driver I would be using sysInByte and sysOutByte to probe the state of the UART and make sure that it is responding properly. You might also check things like the CTS/RTS lines; if you put a UART in the wrong mode it may block while waiting for those things to go into the appropriate state. David Chad wrote: > This problem is still persisting. It appears to be related to the > semaphores taken in tyWrite and/or tyRead for synchronisation. This > semaphores can be taken during a read but not released in the case that a > buffer is full in the case of a write, but is then released when the driver > empties the buffer past a threshold. The only explanation I can find for my > problem is that the semaphore isn't being release properly. > > Here's some details I logged when the system was last locked: > > version > VxWorks (for MPC823) version 5.4. > Kernel: WIND version 2.5 > Made on Feb 4 2002, 14:54:58. > GC4000. > value = 0 = 0x0 > > -> i > NAME ENTRY TID PRI STATUS PC SP > ERRNO DELAY > ---------- ------------ -------- --- ---------- ---- --- > ---------- ----- > MyTask Task::do_mai ffddb0 99 PEND 18b74c ffd860 > 3006b 0 > value = 0 = 0x0 > > -> tt 0xffddb0 > 2b2a4c write__2IOPCvi +d4 : write ([]) > 166e34 write +10 : iosWrite ([]) > 168398 iosWrite +cc : tyWrite ([]) > 19ae0c tyWrite +5c : semTake ([]) > 18c288 semTake +134: semBTake ([]) > value = 0 = 0x0 > > -> d 0xffd860 > 00ffd860: 00ff d878 0018 b74c 0000 0000 0000 00b4 *...x...L........* > 00ffd870: ffff ffff 00fe af04 00ff d890 0018 c288 *................* > > ** note: presumed 0xfeaf04 to be the SEM_ID because the semTake() call takes > both the ID and WAIT_FOREVER (-1 or 0xfffffff) as arguments > > -> semShow 0xfeaf04,1 > Semaphore Id : 0xfeaf04 > Semaphore Type : BINARY > Task Queueing : PRIORITY > Pended Tasks : 1 > State : EMPTY > > ** note: semaphore empty. This basically has to be one or another of the > two semaphores dealt with in tyWrite(). How do I tell which one?? > > semTake (&pTyDev->wrtSyncSem, WAIT_FOREVER); > OR > semTake (&pTyDev->mutexSem, WAIT_FOREVER); > > ** however, it's probably the wrtSyncSem since there is a condition where a > write does not give it back after its call is complete, the mutex is ALWAYS > given back. > > if (rngFreeBytes (pTyDev->wrtBuf) > 0) > semGive (&pTyDev->wrtSyncSem); > semGive (&pTyDev->mutexSem); > > NAME TID PRI TIMEOUT > ---------- -------- --- ------- > MyTask ffddb0 99 0 > value = 0 = 0x0 > > -> d 0xfeaf04 (SEMAPHORE struct) > 00feaf00: 001e 1b80 0001 0000 00ff ddb0 * ............* > 00feaf10: 00ff ddb0 0000 0000 001d 5188 00ff ddb0 *..........Q.....* > 00feaf20: 0001 0000 0000 0000 0004 0000 0017 0ae0 *................* > > typedef struct semaphore /* SEMAPHORE */ > { > OBJ_CORE objCore; /* 0x00: object management */ > UINT8 semType; > UINT8 options; > UINT16 recurse; > Q_HEAD qHead; > union > { > UINT count; > struct windTcb *owner; > } state; > } SEMAPHORE; > > I also noticed that doing a semGive() on the command line did not seem to > kick things back into action. Another fellow I have been corresponding with > (who is experiencing the same condition) said that > the tyShow() function [which I don't seem to have?] shows the ty device ring > buffer and it is, in fact, full. I'm not sure how I can verify this on my > end. It would explain why the synch semaphore was not given back, so the > question that remains is why this call: > > /* when we pass the write threshhold, give write synchonization > * and release tasks pended in select */ > > if (rngFreeBytes (ringId) == tyWrtThreshold) > { > semGive (&pTyDev->wrtSyncSem); > if (_func_selWakeupAll != NULL) > (* _func_selWakeupAll) (&pTyDev->selWakeupList, SELWRITE); > } > } > > At the end of tyITx() did not seem to give it back either, or perhaps it > didn't have a chance as the interrupts stopped being raised to clear out the > queue?? Perhaps its because of the low interrupt priority of the SMC ports > on the 823? We do have timers > in the system, I am looking into whether those all clear their interrupt > pending bits etc properly so as to allow lower priority interrupts to be > processed. > > Thats about where I'm at now. Either my tx interrupts are dying from time > to time (or being starved) and the driver doesn't recover, or somebody is > failing a call to get > a char from the ty buffer. The latter might indicate a problem with this > macro call in tyITx(): > > /* check for more characters to output */ > > else if (RNG_ELEM_GET (ringId, pChar, nn) == 0) /* no more chars */ > pTyDev->wrtState.busy = FALSE; > else > > The macro is defined as (in rngLib.h): > > /*************************************************************************** > * > * RNG_ELEM_GET - get one character from a ring buffer > * > * This macro gets a single character from the specified ring buffer. > * Must supply temporary variable (register int) 'fromP'. > * > * RETURNS: 1 if there was a char in the buffer to return, 0 otherwise > * > * NOMANUAL > */ > > #define RNG_ELEM_GET(ringId, pCh, fromP) \ > ( \ > fromP = (ringId)->pFromBuf, \ > ((ringId)->pToBuf == fromP) ? \ > 0 \ > : \ > ( \ > *pCh = (ringId)->buf[fromP], \ > (ringId)->pFromBuf = ((++fromP == (ringId)->bufSize) ? 0 : fromP), \ > 1 \ > ) \ > ) > > If anybody could shed any further insite into this problem, or perhaps point > out some facilities I could use to get a better look at the relevant state > of the system, I would appreciate it and shower you in praise. > > -Chad --------------------------- Newsgroups: comp.os.vxworks Subject: Re: system debug mode Date: Mon, 18 Feb 2002 11:39:18 -0500 From: david lindauer Message-ID: <3C712E36.765ABDC1@notifier-is.net> References: <5df340e0.0202110519.46251276@posting.google.com> I think if you want to do system level debugging you have to be connected via an ethernet port. David Ruchama Goldshmidt wrote: > Hello vxWorkers!! > > I am working with tornado 2 on a PPC8260 board. > Now,in my project I try to include the system debug mode utility but , > when I open a target shell I get a timeout error I checkout and see > that all the congiguration details are correct !! (WDBSERIAL > :38400,COM2) > is anyone has an idea what else should I do?? > > thanks in advance!!! > Ruchama Goldshmidt --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Wind Web Server - problem in httpSymTabAdd Date: Mon, 18 Feb 2002 11:42:11 -0500 From: david lindauer Message-ID: <3C712EE3.12E14CE5@notifier-is.net> References: <3c67dbde@news.012.net.il> <3c6cce1f@brateggebdc5.br-automation.co.at> yes it only stores a pointer in this case. I used httpStrDup to get around this one, but it would work just as well if you made a global variable for each symbol. David Werner Schiendl wrote: > Hi, > > just a shot in the dark, because I don't have Wind Web Server at hand. > But are you sure, that httpSymTabAdd() does copy the given variable name?? > > There are a lot of occasions, where VxWorks only stores a POINTER to what > you hand it. > Obviously, the buffer would then be overwritten in each interation of your > loop. > > hth > Werner > > "Yaron Cohen" wrote in message > news:3c67dbde@news.012.net.il... > > Hi, > > I am getting "symbol xxx not found! SSI failed " (8 times for each error) > > after > > > > trying to add symbols to httpSymTabAdd() in the follwing way: > > > > ----------------------------------------------------------- > > char buf[50]; > > > > for(i = 1; i < 9; i++) > > { > > sprintf(buf,"AutomaticTAOS%d", i); > > > > symId = httpSymTabAdd (httpGlobalSymTabId, > > buf, > > "", > > HTTP_SYM_TYPE_BOOL_ID, > > 1, > > &T1_interface[i-1].BitsClockEnable); > > } > > > > ---------------------------------------------------------------- > > > > When I am trying to do it without using 'buf' it works fine (without the > > loop): > > > > ----------------------------------------------------------- > > symId = httpSymTabAdd (httpGlobalSymTabId, > > AutomaticTAOS1, > > "", > > HTTP_SYM_TYPE_BOOL_ID, > > 1, > > &T1_interface[0].BitsClockEnable); > > > > symId = httpSymTabAdd (httpGlobalSymTabId, > > AutomaticTAOS2, > > "", > > HTTP_SYM_TYPE_BOOL_ID, > > 1, > > &T1_interface[1].BitsClockEnable); > > > > ------- > > . > > . > > . > > | > > > > symId = httpSymTabAdd (httpGlobalSymTabId, > > AutomaticTAOS8, > > "", > > HTTP_SYM_TYPE_BOOL_ID, > > 1, > > &T1_interface[7].BitsClockEnable); > > --------------------------------------------------- > > > > Any idea ? > > > > Thanks, > > > > Yaron Cohen > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: for PAP configuration Date: Mon, 18 Feb 2002 11:48:41 -0500 From: david lindauer Message-ID: <3C713069.2999E109@notifier-is.net> References: alain tamdem wrote: > To allow authentification using PAP > do i need only to turn the option > PPP_OPT_NO_PAP to 1 or also PPP_OPT_REQUIRE_PAP to 1 ?? set OPT_NO_PAP to 0 and OPT_REQUIRE_PAP to 1 David --------------------------- Newsgroups: comp.os.vxworks Subject: Re: PCI card Date: Mon, 18 Feb 2002 11:53:14 -0500 From: david lindauer Message-ID: <3C71317A.A31D4A45@notifier-is.net> References: if the card is designed to have both memory and I/O ports associated with it, generally that will be setable in the PCI configuration space and then you just access the card directly and ignore PCI. This configuration will be card dependent. For a simple ATA PCI controller I set BAR0 and BAR1 to the two ATA addresses to be used, and PCI_CFG_DEV_INT_LINE to the interrupt to use, and it worked ok. If you have DOS/BIOS you can sometimes get a driver to configure your device, then read the PCI configuration registers to determine what to do next. David "Michael B." wrote: > Hi! > I'm working on a I/O PCI card and using pciHeaderShow I obtained a base > address 0 (BAR0) equal to 0xe4000000 and some other items (are they > important?). > How can I use BAR0 to access to card's register? > What about I/O mapped and/or memory mapped devices? > My BSP is a default PcPentium. > > Thanks. > Mic. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: clock_gettime fails with write() but not fwrite Date: Mon, 18 Feb 2002 16:31:10 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: <6wBb8.3387$zl4.2197904@typhoon3.we.ipsvc.net> "Rory O'Connor" wrote in message news:6wBb8.3387$zl4.2197904@typhoon3.we.ipsvc.net... > Hi, > > I am using Tornado 2.1 and I can't use the function "clock_gettime" with > "write()" to a SCSI harddrive. For some reason, the write tampers with the > clock counter and messes up clock_gettime. I don't see this behavior with > "fwrite()". Does anyone know a work around so that I can use "write" and > also measure it performance with "clock_gettime"? > > > This is a known bug that's been discussed here a couple of times. The solution is to get onto WRS, tell them you've hit TSR# 144712, and you want them to send you the source code to clockLib.c; then replace "vxTicks" with "vxAbsTicks" throughout. DaveK - -- moderator of alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! Proud Member of the Exclusive "I have been plonked by Davee because he thinks I'm interesting" List Member # Master of Many Meowing Minions Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage above and beyond the call of hilarity. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Monitoring Interrupt Stack? Date: Mon, 18 Feb 2002 16:39:20 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: <3C6F28F6.44FFC43A@yahoo.com> <1013946407.21382.0.nnrp-10.d4e4b8d1@news.demon.co.uk> <9423c55b.0202170756.21cf556e@posting.google.com> "Chad" wrote in message news:HRTb8.169329$qr2.23242575@news2.rdc1.bc.home.com... > Thanks all! I'm using a PPC, but was unaware of the @checkStack trick (or > that the two versions differed at all). Some great info here. Just BTW and FYI, the @ prefix forces the host shell to forward the command down to the target shell for execution, and applies to all shell commands. Most commands should run the same on either shell, but if you have non-synchronized symbol tables it's important to know which one you're inspecting; likewise, when using "cd" and similar filing-system related commands, you need to be aware of whether you're changing the host's or target's current directory, since they both generally have different filing systems. DaveK - -- moderator of alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! Proud Member of the Exclusive "I have been plonked by Davee because he thinks I'm interesting" List Member # Master of Many Meowing Minions Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage above and beyond the call of hilarity. --------------------------- Newsgroups: comp.os.vxworks,gnu.gcc.help Subject: Re: prjObjs.lst - error 0x80 Date: Mon, 18 Feb 2002 17:15:13 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: <3c6c0065@pull.gecm.com> "Ofer Goren" wrote in message news:newscache$v3gorg$5q$1@lnews.actcom.co.il... > do you have cygwin installed? if so, check for existance of cygwin.dll. We > are starting to believe that there is some connection. If you can't find the > file, try to restore it from another PC, and place it where it should be. > Let me know if it worked... If you suspect cygwin problems, run "cygcheck -s -v -r -h", and look in particular for multiple copies of the cygwin dll being found in the last couple of sections of the output. That would always be a source of trouble. DaveK - -- moderator of alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! Proud Member of the Exclusive "I have been plonked by Davee because he thinks I'm interesting" List Member # Master of Many Meowing Minions Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage above and beyond the call of hilarity. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Reading from serial ports - Blocking or interrupts? Date: Mon, 18 Feb 2002 16:52:38 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: <37449c9d.0202130624.43ffc949@posting.google.com> <1IRa8.2442$5o.1075855@newsr2.u-net.net> <3c6cf987.89521798@news.netis.com> wrote in message news:3c6cf987.89521798@news.netis.com... > On Thu, 14 Feb 2002 16:29:57 -0000, "Dave Korn" > wrote: > > >Can you determine how long the packet is going to be by inspecting the first > >one or two chars ? > > Good morning, Dave. > > Actually, the packets on some of the serial ports are fixed length, so > on those it's not a problem. > > Others are NMEA sentences that have variable-length value fields (e.g. > a water depth might be 4.7 meters or it might be 258.3 meters) thus > varying the length of the sentence. There is no way to detect this > until the trailing CR/LF is received. So, the plan (finalized with > the help of you all here in the NG) is to set the interrupt level for > the minimum sentence length and then handle the rest of the sentence > using the 4-character-time no-data interrupt. That's one of those things where a little hardware support makes possible something that simply could not be done in software. You'd never believe some of the things that hw designers have expected me to be able to do in software that would trash the performance of an entire system just to save a few pennies on the hw side.... DaveK - -- moderator of alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! Proud Member of the Exclusive "I have been plonked by Davee because he thinks I'm interesting" List Member # Master of Many Meowing Minions Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage above and beyond the call of hilarity. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: clock_gettime fails with write() but not fwrite Date: Mon, 18 Feb 2002 18:12:10 +0000 (UTC) From: David Laight Organization: BT Openworld Message-ID: <3C714455.4050207@spamm.me.l8s.co.uk> References: <6wBb8.3387$zl4.2197904@typhoon3.we.ipsvc.net> Dave Korn wrote: > > This is a known bug that's been discussed here a couple of times. The > solution is to get onto WRS, tell them you've hit TSR# 144712, and you want > them to send you the source code to clockLib.c; then replace "vxTicks" with > "vxAbsTicks" throughout. Has anyone tracked down why/where the scsi code uses a -ve timeout? That is another bug lurking in there somewhere, changing from vxTicks to vxAbsTicks just hides one symptom of it. Mind you the date code needs integrating better anyway, far too much arithmetic goes on... David --------------------------- Newsgroups: comp.os.vxworks Subject: Writing a BSP from scratch Date: Mon, 18 Feb 2002 15:06:13 -0500 From: "Harjinder Chayra" Organization: Storm Internet Services Message-ID: Sender: terayon@gw-terayon.storm.ca Hello all, I am planning to write a BSP for the ARM CPU from scratch with the help of the BSP porting kit from WRS. If you have been involved in this level of development I would be interested in learning, how helpful (or not) the BSP porting kit was how long it took to get the BSP working any recommendations e.g. base the new BSP on an exisiting working one. Looking forward to your comments! Harjinder --------------------------- Newsgroups: comp.os.vxworks Subject: Re: anticipating machine check error... Date: 18 Feb 2002 12:59:06 -0800 From: dwh8@yahoo.com (David) Organization: http://groups.google.com/ Message-ID: <8fe86b76.0202181259.e622c15@posting.google.com> References: <8fe86b76.0202131506.3c00702d@posting.google.com> Michael, Thanks for the info. I was able to write a little test program which works, but there is a problem. The code section that will be generating the exceptions also relies on a pending call (msgQReceive). In my test code I reduplicated that effect. When the pending call is reached, the system completely fails and must be rebooted. Any Ideas on how to resolve this? Below is the test code I wrote: - ------------------------------------------------------------------- jmp_buf exceptionPoint; /* handle exception */ void expHandler ( int task, int vecNum, ESFPPC *pEsf) { printf("\nTask %i caused vector # %i error", task, vecNum); longjmp(exceptionPoint, 1); } int gen(int memLocation) { ULONG * buf = (ULONG *) memLocation; // 0xffffffff causes exception char cbuf[10]; MSG_Q_ID msgQ = msgQCreate(10,10,0); excHookAdd((FUNCPTR)expHandler); for(;;){ msgQReceive(msgQ,cbuf,10,20); if (setjmp(exceptionPoint) == 0){ printf("\ntask %i ready to print\n",taskIdSelf()); printf("\n%x",*buf); // generate exception }else{ printf("\ntask %i resumed\n",taskIdSelf()); } } // end of for(;;) } // end of gen() - ------------------------------------------------------------- "Michael Lawnick" wrote in message news:... > Hi David, > > if your problem is only 'first access', your function is vxMemProbe(). > If you fear card fail while operating, use excHookAdd() and setjmp/longjmp. > If you don't like the logMsg of the vxWorks excHandler, you will need > excVecSet(). Use setjmp/longjmp here too. > > Incrementing PC would be for 'transparent operation fail', i.e. if failing > task shouldn't see its fail. IMO this is a solution for players, not a solid > one. > > -- > Mit freundlichen Grüßen, > Michael Lawnick > ============================================== > SOFTEC GmbH Tel +49-731-96600-0 > Promenade 17 Fax +49-731-96600-23 > D-89073 Ulm Michael Lawnick > Germany lawnick@softec.de > ============================================== > > "David" schrieb im Newsbeitrag > news:8fe86b76.0202131506.3c00702d@posting.google.com... > > I was hoping somebody could verify this for me: > > > > I am porting an application from PSOS to VxWorks 5.4 on a Power PC > > platform (mv5100). > > > > The existing application has a set of exception handling routines that > > are installed (in the PSOS assembly code) to handle machine check > > errors. The errors in question occur when card signal processing is > > attempted on a card that is missing from the chassis. We don't wan't > > the system to halt if a card is missing or not working correctly. > > > > From reading previous posts, it looks like I will have to use the > > excVecSet() function, but those posts also mention manually > > incrementing the Program Counter which sounds a bit hairy. My current > > handlers get passed: > > > > vector offset for the exception, > > addr - next program location when the exception occurred > > ExceptionFramePtr - exception frame pointer > > > > but the only thing I do with this information is display it in the > > error log. > > > > Any ideas on how to proceed would be greatly appreciated, > > > > David --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Writing a BSP from scratch Date: 18 Feb 2002 17:55:52 -0500 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: References: Sender: bpringlemeir@DeadDuck >>>>> "HC" == Harjinder Chayra writes: HC> Hello all, I am planning to write a BSP for the ARM CPU from HC> scratch with the help of the BSP porting kit from WRS. If you HC> have been involved in this level of development I would be HC> interested in learning, HC> how helpful (or not) the BSP porting kit was how long it took to HC> get the BSP working any recommendations e.g. base the new BSP on HC> an exisiting working one. I would suggest getting an ARM BSP of some form that matches your CPU. Ie, StrongARM, Cirrus, etc. At the very least, a BSP that has the same ARM CPU version as yours. If you can not do this for some sort of IP reasons, then the BSP porting kit is helpful. However, you should be able to create your own BSP without the guide, given that you have another similar BSP. hth, Bill Pringlemeir. - -- I firmly believe that all young dudes should get source licenses . I couldn't care less about your venereal disease. vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" --------------------------- Newsgroups: comp.os.vxworks Subject: how to release a mutex when the process is killed Date: Mon, 18 Feb 2002 15:44:08 -0800 From: "Mark Galecki" Message-ID: <3c719169$0$5154$724ebb72@reader3.ash.ops.us.uu.net> Hello, I have what must be a common issue when writing a communication device driver and surprisingly I don't find an easy solution in VxWorks manuals. I am experienced with writing drivers for other OSs, but not VxWorks, and probably I don't see an easy solution, so please help... - ---- Description of problem: To prevent interleaving under multiple threads, the driver read() handler first grabs a read mutex and it releases it when it returns. In this way only one read handler is executing at any given time. This is probably the way all drivers like that are written. Now, inside the handler I have to wait for data. This is a blocking read. So I wait on a semaphore, and when the data arrives, the read ISR signals me and I can read the data. This also is probably very common. The problem is this: the user starts a read app and it hangs, waiting for data. Then the user notices something is wrong and no data will be coming, so they kill the user app. I do it with the Tornado debugger by interrupting it and/or stopping it, but there are other ways, such as sending a kill from command line. Then the read mutex which was taken, is not released, so upon fixing the problem, if one tries to start the read app again, the read handler won't even enter. I was expecting that, upon a kill, first the semaphore wait would exit with an ERROR so I can clean up in the handler but that does not happen. I was also hoping that maybe the close handler would get called so I could attempt a cleanup based on a task ID, but no, the close handler does not get called either. So I have to reboot the machine if no data is coming - hardly a satisfactory solution. Thank you, Mark Galecki --------------------------- Newsgroups: comp.os.vxworks Subject: problem during debbuging!! Date: 18 Feb 2002 21:31:32 -0800 From: zvikag@yahoo.com (zvi) Organization: http://groups.google.com/ Message-ID: Hello dear vxWorkers! we develop a new board base on a ppc8260 processor. We use Tornado2 with vxWorks 5.4. We have in our board 2 serial channels, SMC and SCC and now we want to use the SMC channel for debugging purpose. In our project we defined the WDB serial connection(INCLUDE_WDB_COMM_SERIAL)and the WDB system debugging (INCLUDE_WDB_SYS) We set the boad rate to 38400 . In order to verify the connection, we check the printout in the terminal when it boots and we see the prompt: WDB READY which as explained in the manuals it demonstrates that all the connections are correct. We also succeeded to attach the debugger (usually, but sometimes it complains about rpc timeout error), but here the problem starts, when trying to attach the "system" mode in the 'attach' option in the debug menu ,there is an error: attach system attaching to system failed could not suspend system context: API_REQUEST_TIMED_OUT If we try to attach any of the task the debugger work as expected. We really want to use system debug mode, and we will be gratefull for a quick advice of how enable correctlly this debug mode, Thanks in advance, Best Regards, Zvi Gold. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: how to release a mutex when the process is killed Date: 19 Feb 2002 06:02:15 GMT From: Denis Perelyubskiy Message-ID: References: <3c719169$0$5154$724ebb72@reader3.ash.ops.us.uu.net> Mark, * Mark Galecki [02-Mon-02 15:44 -0800]: [...] > Now, inside the handler I have to wait for data. This is > a blocking read. So I wait on a semaphore, and when the > data arrives, the read ISR signals me and I can read the > data. This also is probably very common. > The problem is this: the user starts a read app and it > hangs, waiting for data. Then the user notices something > is wrong and no data will be coming, so they kill the > user app. I do it with the Tornado debugger by > interrupting it and/or stopping it, but there are other > ways, such as sending a kill from command line. > Then the read mutex which was taken, is not released, so > upon fixing the problem, if one tries to start the read > app again, the read handler won't even enter. > I was expecting that, upon a kill, first the semaphore > wait would exit with an ERROR so I can clean up in the > handler but that does not happen. I was also hoping that > maybe the close handler would get called so I could > attempt a cleanup based on a task ID, but no, the close > handler does not get called either. I do not know about the original source of your problem, and whether what you are seeing is normal, but... If what you are seeing is normal, and you're going about it the right way, perhaps the taskDeleteHookAdd() is what you are looking for. Do a search for this taskDelete hook on groups.google.com, and within a couple of weeks you'll find another discussion that had to do with freeing memory. At least one interesting thing, is that you can not call free(), since this hook is not guaranteed to be executing in the context of any task. semGive() is ok though, I imagine. This is, provided that the 'close' handler you are talking about is not the same hook I am talking about. Another option (well, I dont know if this is an option for you), is perhaps in your init routine (I assume you have something of that sort) to do a semDelete()? I mean, you must initialize a semaphore somewhere, right? So, why not delete it before initializing? ... just a thought... denis - -- if you want to reply by mail, please tweak my address in a reasonable way. sorry for the inconvenience. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: how to release a mutex when the process is killed Date: 18 Feb 2002 22:13:09 -0800 From: chwin@china.com (Adom) Organization: http://groups.google.com/ Message-ID: <92dbccc1.0202182213.4889ad40@posting.google.com> References: <3c719169$0$5154$724ebb72@reader3.ash.ops.us.uu.net> i think it would be ok when you create the task ,add a taskdelete hook function, in this function, release the mutex Adom Chan > > Description of problem: > To prevent interleaving under multiple threads, the driver read() handler > first grabs a read mutex and it releases it when it returns. In this way > only one read handler is executing at any given time. This is probably the > way all drivers like that are written. > > Now, inside the handler I have to wait for data. This is a blocking read. > So I wait on a semaphore, and when the data arrives, the read ISR signals me > and I can read the data. This also is probably very common. > > The problem is this: the user starts a read app and it hangs, waiting for > data. Then the user notices something is wrong and no data will be coming, > so they kill the user app. I do it with the Tornado debugger by > interrupting it and/or stopping it, but there are other ways, such as > sending a kill from command line. > > Then the read mutex which was taken, is not released, so upon fixing the > problem, if one tries to start the read app again, the read handler won't > even enter. > > I was expecting that, upon a kill, first the semaphore wait would exit with > an ERROR so I can clean up in the handler but that does not happen. I was > also hoping that maybe the close handler would get called so I could > attempt a cleanup based on a task ID, but no, the close handler does not get > called either. > --------------------------- Newsgroups: comp.os.vxworks Subject: Is there a free database can be used on vxworks Date: 18 Feb 2002 22:18:18 -0800 From: chwin@china.com (Adom) Organization: http://groups.google.com/ Message-ID: <92dbccc1.0202182218.72990392@posting.google.com> Hello, now , i'm designing a model for ip traffic monitor on vxworks. i want to use database to store data,but i don't know wether there is a free database can be used on vxworks? Thanks for any suggestion! Adom Chan 2002-2-19 --------------------------- Newsgroups: comp.os.vxworks Subject: What means this sentens...? Date: Tue, 19 Feb 2002 16:17:05 +0900 From: "kimp" Organization: Korea Research Environment Open Network Message-ID: Hi, I look into the BSP for MPC860. But I couldn't undrestand the meaning of following sentence in romInit.s of BSP. Please let me know what these means. - ---------------------------------------- /* Setup SYPCR for normal and proper background mode operation */ lis r5,0xffff /* ffffff88*/ ori r5,r5,0xff03 stw r5,SYPCR(0)(r4) ==> I couldn't understand this line. - --------------------------------------- Thanks to your attention :-) --------------------------- Newsgroups: comp.os.vxworks Subject: Re: What means this sentens...? Date: Tue, 19 Feb 2002 07:48:43 GMT From: "Gary M" Organization: AT&T Broadband Message-ID: References: SYPCR is the System Protection Control Register, residing in the System Interface Unit (SIU). The instruction you specifically refer to stores the contents of R5 (0xFFFFFF03) into this register. R4 points to the base of the internal memory map. Note that this register is write-once, meaning that it can be written to only one time after reset. The value of 0xFFFFFF03 will disable the Bus Monitor and the Software Watchdog. Refer to section 11.4.3, page 11-9, in the MPC860 User Manual. "kimp" wrote in message news:a4su05$mh$1@news.kreonet.re.kr... > Hi, > > I look into the BSP for MPC860. > But I couldn't undrestand the meaning of following sentence in romInit.s of > BSP. > Please let me know what these means. > > ---------------------------------------- > > /* Setup SYPCR for normal and proper background mode operation */ > > lis r5,0xffff /* ffffff88*/ > ori r5,r5,0xff03 > stw r5,SYPCR(0)(r4) ==> I couldn't understand this line. > > --------------------------------------- > > Thanks to your attention :-) > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: anticipating machine check error... Date: 18 Feb 2002 23:51:32 -0800 From: kannan@comneti.com (kannan) Organization: http://groups.google.com/ Message-ID: <2ede7703.0202182351.7abcc9cf@posting.google.com> References: <8fe86b76.0202131506.3c00702d@posting.google.com> <8fe86b76.0202181259.e622c15@posting.google.com> dwh8@yahoo.com (David) wrote in message news:<8fe86b76.0202181259.e622c15@posting.google.com>... > Michael, > > Thanks for the info. I was able to write a little test program which > works, but there is a problem. > > The code section that will be generating the exceptions also relies on > a pending call (msgQReceive). In my test code I reduplicated that > effect. When the pending call is reached, the system completely fails > and must be rebooted. > > Any Ideas on how to resolve this? > > Below is the test code I wrote: > ------------------------------------------------------------------- > jmp_buf exceptionPoint; > > /* handle exception */ > void expHandler ( int task, int vecNum, ESFPPC *pEsf) > { > printf("\nTask %i caused vector # %i error", task, vecNum); > longjmp(exceptionPoint, 1); > } > > int gen(int memLocation) > { > ULONG * buf = (ULONG *) memLocation; // 0xffffffff causes > exception > char cbuf[10]; > > MSG_Q_ID msgQ = msgQCreate(10,10,0); > > excHookAdd((FUNCPTR)expHandler); > > for(;;){ > > msgQReceive(msgQ,cbuf,10,20); > > if (setjmp(exceptionPoint) == 0){ > printf("\ntask %i ready to print\n",taskIdSelf()); > printf("\n%x",*buf); // generate exception > }else{ > printf("\ntask %i resumed\n",taskIdSelf()); > } > > } // end of for(;;) > } // end of gen() > > ------------------------------------------------------------- > > "Michael Lawnick" wrote in message news:... > > Hi David, > > > > if your problem is only 'first access', your function is vxMemProbe(). > > If you fear card fail while operating, use excHookAdd() and setjmp/longjmp. > > If you don't like the logMsg of the vxWorks excHandler, you will need > > excVecSet(). Use setjmp/longjmp here too. > > > > Incrementing PC would be for 'transparent operation fail', i.e. if failing > > task shouldn't see its fail. IMO this is a solution for players, not a solid > > one. > > > > -- > > Mit freundlichen Grüßen, > > Michael Lawnick > > ============================================== > > SOFTEC GmbH Tel +49-731-96600-0 > > Promenade 17 Fax +49-731-96600-23 > > D-89073 Ulm Michael Lawnick > > Germany lawnick@softec.de > > ============================================== > > > > "David" schrieb im Newsbeitrag > > news:8fe86b76.0202131506.3c00702d@posting.google.com... > > > I was hoping somebody could verify this for me: > > > > > > I am porting an application from PSOS to VxWorks 5.4 on a Power PC > > > platform (mv5100). > > > > > > The existing application has a set of exception handling routines that > > > are installed (in the PSOS assembly code) to handle machine check > > > errors. The errors in question occur when card signal processing is > > > attempted on a card that is missing from the chassis. We don't wan't > > > the system to halt if a card is missing or not working correctly. > > > > > > From reading previous posts, it looks like I will have to use the > > > excVecSet() function, but those posts also mention manually > > > incrementing the Program Counter which sounds a bit hairy. My current > > > handlers get passed: > > > > > > vector offset for the exception, > > > addr - next program location when the exception occurred > > > ExceptionFramePtr - exception frame pointer > > > > > > but the only thing I do with this information is display it in the > > > error log. > > > > > > Any ideas on how to proceed would be greatly appreciated, > > > > > > David Hi, The longjmp( ) function executes correctly in contexts of signal handlers. But it doesn't work in case of exception or interrupt handlers can you try signal handlers for exceptions. Thanks and Regards, Kannan --------------------------- Newsgroups: comp.os.vxworks Subject: MUX Load failed Date: Tue, 19 Feb 2002 09:25:34 +0100 From: Franz Ettengruber Organization: Siemens AG Message-ID: <3C720BFE.8F260574@icn.siemens.de> Hello! We are doing our first steps with VxWorks and try to Boot it on a PC-Pentium Target. We are using a pre-built Boot Rom that should support pci network cards with Intel chipset, but we tried several different (all?) of them and always get this error message: "MUX Load failed". In our opinion it must have something to do with networking. So our question is where can we find information about what types of network cards VxWorks does support and what parameter to specify for "boot device", when booting VxWorks. Any help would be very appreciated. Regards, Franz --------------------------- Newsgroups: comp.os.vxworks Subject: Re: anticipating machine check error... Date: Tue, 19 Feb 2002 09:57:12 +0100 From: "Michael Lawnick" Organization: Buergernetz Dillingen Message-ID: References: <8fe86b76.0202131506.3c00702d@posting.google.com> <8fe86b76.0202181259.e622c15@posting.google.com> Reply-To: "Michael Lawnick" Sender: mlawnick@pd9524cdc.dip.t-dialin.net Hi David, Kannan seems to be correct, I couldn't get your code working too. A secondary exception occurs and system hangs. It has nothing to do with your msgQ. Manual states "Do not use longjmp( ) or setjmp( ) from an ISR.", but exception handling is done in task context ... #@! So you still may use vxMemProbe for all critical accesses. This should avoid any exception at all. - -- Mit freundlichen Grüßen, Michael Lawnick ============================================== SOFTEC GmbH Tel +49-731-96600-0 Promenade 17 Fax +49-731-96600-23 D-89073 Ulm Michael Lawnick Germany lawnick@softec.de ============================================== "David" schrieb im Newsbeitrag news:8fe86b76.0202181259.e622c15@posting.google.com... > Michael, > > Thanks for the info. I was able to write a little test program which > works, but there is a problem. > > The code section that will be generating the exceptions also relies on > a pending call (msgQReceive). In my test code I reduplicated that > effect. When the pending call is reached, the system completely fails > and must be rebooted. > > Any Ideas on how to resolve this? > > Below is the test code I wrote: > ------------------------------------------------------------------- > jmp_buf exceptionPoint; > > /* handle exception */ > void expHandler ( int task, int vecNum, ESFPPC *pEsf) > { > printf("\nTask %i caused vector # %i error", task, vecNum); > longjmp(exceptionPoint, 1); > } > > int gen(int memLocation) > { > ULONG * buf = (ULONG *) memLocation; // 0xffffffff causes > exception > char cbuf[10]; > > MSG_Q_ID msgQ = msgQCreate(10,10,0); > > excHookAdd((FUNCPTR)expHandler); > > for(;;){ > > msgQReceive(msgQ,cbuf,10,20); > > if (setjmp(exceptionPoint) == 0){ > printf("\ntask %i ready to print\n",taskIdSelf()); > printf("\n%x",*buf); // generate exception > }else{ > printf("\ntask %i resumed\n",taskIdSelf()); > } > > } // end of for(;;) > } // end of gen() > > ------------------------------------------------------------- > > "Michael Lawnick" wrote in message news:... > > Hi David, > > > > if your problem is only 'first access', your function is vxMemProbe(). > > If you fear card fail while operating, use excHookAdd() and setjmp/longjmp. > > If you don't like the logMsg of the vxWorks excHandler, you will need > > excVecSet(). Use setjmp/longjmp here too. > > > > Incrementing PC would be for 'transparent operation fail', i.e. if failing > > task shouldn't see its fail. IMO this is a solution for players, not a solid > > one. > > > > -- > > Mit freundlichen Grüßen, > > Michael Lawnick > > ============================================== > > SOFTEC GmbH Tel +49-731-96600-0 > > Promenade 17 Fax +49-731-96600-23 > > D-89073 Ulm Michael Lawnick > > Germany lawnick@softec.de > > ============================================== > > > > "David" schrieb im Newsbeitrag > > news:8fe86b76.0202131506.3c00702d@posting.google.com... > > > I was hoping somebody could verify this for me: > > > > > > I am porting an application from PSOS to VxWorks 5.4 on a Power PC > > > platform (mv5100). > > > > > > The existing application has a set of exception handling routines that > > > are installed (in the PSOS assembly code) to handle machine check > > > errors. The errors in question occur when card signal processing is > > > attempted on a card that is missing from the chassis. We don't wan't > > > the system to halt if a card is missing or not working correctly. > > > > > > From reading previous posts, it looks like I will have to use the > > > excVecSet() function, but those posts also mention manually > > > incrementing the Program Counter which sounds a bit hairy. My current > > > handlers get passed: > > > > > > vector offset for the exception, > > > addr - next program location when the exception occurred > > > ExceptionFramePtr - exception frame pointer > > > > > > but the only thing I do with this information is display it in the > > > error log. > > > > > > Any ideas on how to proceed would be greatly appreciated, > > > > > > David --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Writing a BSP from scratch Date: Tue, 19 Feb 2002 09:26:58 +0000 (UTC) From: David Laight Organization: BT Openworld Message-ID: <3C721ABC.4030605@spamm.me.l8s.co.uk> References: Harjinder Chayra wrote: > Hello all, > I am planning to write a BSP for the ARM CPU from scratch with > the help of the BSP porting kit from WRS > If you have been involved in this level of development I would > be interested in learning, > how helpful (or not) the BSP porting kit was > how long it took to get the BSP working > any recommendations e.g. base the new BSP on an exisiting working one. I would start from one of the WRS BSP for something similar, then add in the extra functionality as you need it. If you can write device drivers and system setup code I'm not sure the BSP porting kit contains anything useful! Once you have romInit.s done vxWorks will probably boot. Then you need to fix WRS broken ARM interrupt and mmu code, stop the RTC drifting and write device drivers for your hardware. Get the serial port working early on, and write debug strings to it - in romInit.s. Since this is all 'ongoing' work I don't know how you use the 'project' facility as it seems to snapshot the BSP! (It barfed on one of my BSP header files, so never generated a project!) David --------------------------- Newsgroups: comp.os.vxworks Subject: Re: anticipating machine check error... Date: 19 Feb 2002 01:58:21 -0800 From: kannan@comneti.com (kannan) Organization: http://groups.google.com/ Message-ID: <2ede7703.0202190158.1e19c22a@posting.google.com> References: <8fe86b76.0202131506.3c00702d@posting.google.com> <8fe86b76.0202181259.e622c15@posting.google.com> dwh8@yahoo.com (David) wrote in message news:<8fe86b76.0202181259.e622c15@posting.google.com>... > Michael, > > Thanks for the info. I was able to write a little test program which > works, but there is a problem. > > The code section that will be generating the exceptions also relies on > a pending call (msgQReceive). In my test code I reduplicated that > effect. When the pending call is reached, the system completely fails > and must be rebooted. > > Any Ideas on how to resolve this? > > Below is the test code I wrote: > ------------------------------------------------------------------- > jmp_buf exceptionPoint; > > /* handle exception */ > void expHandler ( int task, int vecNum, ESFPPC *pEsf) > { > printf("\nTask %i caused vector # %i error", task, vecNum); > longjmp(exceptionPoint, 1); > } > > int gen(int memLocation) > { > ULONG * buf = (ULONG *) memLocation; // 0xffffffff causes > exception > char cbuf[10]; > > MSG_Q_ID msgQ = msgQCreate(10,10,0); > > excHookAdd((FUNCPTR)expHandler); > > for(;;){ > > msgQReceive(msgQ,cbuf,10,20); > > if (setjmp(exceptionPoint) == 0){ > printf("\ntask %i ready to print\n",taskIdSelf()); > printf("\n%x",*buf); // generate exception > }else{ > printf("\ntask %i resumed\n",taskIdSelf()); > } > > } // end of for(;;) > } // end of gen() > > ------------------------------------------------------------- > > "Michael Lawnick" wrote in message news:... > > Hi David, > > > > if your problem is only 'first access', your function is vxMemProbe(). > > If you fear card fail while operating, use excHookAdd() and setjmp/longjmp. > > If you don't like the logMsg of the vxWorks excHandler, you will need > > excVecSet(). Use setjmp/longjmp here too. > > > > Incrementing PC would be for 'transparent operation fail', i.e. if failing > > task shouldn't see its fail. IMO this is a solution for players, not a solid > > one. > > > > -- > > Mit freundlichen Grüßen, > > Michael Lawnick > > ============================================== > > SOFTEC GmbH Tel +49-731-96600-0 > > Promenade 17 Fax +49-731-96600-23 > > D-89073 Ulm Michael Lawnick > > Germany lawnick@softec.de > > ============================================== > > > > "David" schrieb im Newsbeitrag > > news:8fe86b76.0202131506.3c00702d@posting.google.com... > > > I was hoping somebody could verify this for me: > > > > > > I am porting an application from PSOS to VxWorks 5.4 on a Power PC > > > platform (mv5100). > > > > > > The existing application has a set of exception handling routines that > > > are installed (in the PSOS assembly code) to handle machine check > > > errors. The errors in question occur when card signal processing is > > > attempted on a card that is missing from the chassis. We don't wan't > > > the system to halt if a card is missing or not working correctly. > > > > > > From reading previous posts, it looks like I will have to use the > > > excVecSet() function, but those posts also mention manually > > > incrementing the Program Counter which sounds a bit hairy. My current > > > handlers get passed: > > > > > > vector offset for the exception, > > > addr - next program location when the exception occurred > > > ExceptionFramePtr - exception frame pointer > > > > > > but the only thing I do with this information is display it in the > > > error log. > > > > > > Any ideas on how to proceed would be greatly appreciated, > > > > > > David Hi Gurus, The above code i tried working with excVecSet it works fine. The code is changed as follows: - ------------------------------------------------------------------- > jmp_buf exceptionPoint; > #if 0 /*commented this will not work*/ > /* handle exception */ > void expHandler ( int task, int vecNum, ESFPPC *pEsf) > { > printf("\nTask %i caused vector # %i error", task, vecNum); > longjmp(exceptionPoint, 1); > } #endif /* handle exception */ static int machineCheckIsr (ESFPPC * pEsf ) { logMsg("Machine Check\n",0,0,0,0,0,0); longjmp(exceptionPoints, 1); } > > int gen(int memLocation) > { > ULONG * buf = (ULONG *) memLocation; // 0xffffffff causes > exception > char cbuf[10]; > > MSG_Q_ID msgQ = msgQCreate(10,10,0); > >/* excHookAdd((FUNCPTR)expHandler);*//*commented this will not work*/ excVecSet ((FUNCPTR *)_EXC_OFF_MACH, FUNCREF (machineCheckIsr));/*Do this*/ > > for(;;){ > > msgQReceive(msgQ,cbuf,10,20); > > if (setjmp(exceptionPoint) == 0){ > printf("\ntask %i ready to print\n",taskIdSelf()); > printf("\n%x",*buf); // generate exception > }else{ > printf("\ntask %i resumed\n",taskIdSelf()); > } > > } // end of for(;;) > } // end of gen() > > ------------------------------------------------------------- what i have posted previously is correct the longjmp( ) function executes correctly in the contexts of signal handlers ,But it will not work in the case of Interrupt Handlers and some Exception handlers.(e.g.excHookAdd function) Thanks and Regards kannan --------------------------- Newsgroups: comp.os.vxworks Subject: Re: anticipating machine check error... Date: Tue, 19 Feb 2002 11:07:59 +0100 From: "Werner Schiendl" Message-ID: <3c7223ee$1@brateggebdc5.br-automation.co.at> References: <8fe86b76.0202131506.3c00702d@posting.google.com> <8fe86b76.0202181259.e622c15@posting.google.com> Hi, "Michael Lawnick" wrote in message news:a4t419$p56$1@snoopy.bndlg.de... > Hi David, > > Kannan seems to be correct, I couldn't get your code working too. A > secondary exception occurs and system hangs. It has nothing to do with your > msgQ. > Manual states "Do not use longjmp( ) or setjmp( ) from an ISR.", but > exception handling is done in task context ... #@! > -- snip -- Are you sure that the exception hook is run in task context? I could not find details on this in the manuals right now. Anyway, if it is task context, then it is excTask where it is called. IIRC, setjmp and longjmp are task bound functions (i. e. every task has its own longjmp() destination set up). Your code longjmps excTask(), which will naturally fail. An other problem you'll face is, that your task is (AFAIK unconditionally) suspended after the excHook you've added. I'd recommend use vxMemProbe, if you can do so. Alternatively, you may try to find out how vxMemProbe does the trick. hth Werner --------------------------- Newsgroups: comp.os.vxworks,gnu.gcc.help Subject: Re: prjObjs.lst - error 0x80 - solution Date: Tue, 19 Feb 2002 13:16:30 +0200 From: "Ofer Goren" Organization: Verio Message-ID: References: Well, the problem is in the combination w/ cygwin, god knows why. If cygwin was installed before Torando, removing cygwin after the torando installation will hard Tornado. The solution is to re-install cygwin, and update the window's PATH variable to the location of the cygwin's bin directory. - -- Ofer Goren Iamba Technologies "Ofer Goren" wrote in message news:newscache$brajrg$zle$1@lnews.actcom.co.il... > ------------------------ > Win2k Sp2 > Tornado 2.02 for PPC > GNU compiler > Envoy 9.2 > ------------------------ > Hi. > > I have a strange problem, when compiling a project that worked well a day > ago. > When I compile, I get an make: *** /prjObjs.lst Error 0x80 > > I tried creating new project, under a new worksapce, with no files, the > works. Nothing helped. > > No makefile under ../target/h/make/ was modified during the last 6 months... > > Anyway, any help, before re-installing Tornado (and all other utilities > after) ... :(( > > Thanks, > > Ofer > > -- > Ofer Goren > Iamba Technologies > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Question:Capturing stack trace of a taskwhen it crashes Date: 19 Feb 2002 03:49:18 -0800 From: sunil@dharwad.com (Sunil K) Organization: http://groups.google.com/ Message-ID: <396b9cc7.0202190349.3ae5f1c5@posting.google.com> References: <396b9cc7.0202140719.7109204c@posting.google.com> <3ca555c5.0202172156.10963c11@posting.google.com> mistry_amit@yahoo.com (Amit) wrote in message news:<3ca555c5.0202172156.10963c11@posting.google.com>... > sunil@dharwad.com (Sunil K) wrote in message news:<396b9cc7.0202140719.7109204c@posting.google.com>... > > Hi VxWorkers, > > > > I need to capture stack trace in a file when any task crashes. > > (Even if the crash leads to a reboot of vxWorks) > > > > Any ideas how to do it ? Should I replcae the default interupt > > handlers that suspend the task when it does any illegal operation ? > > How should I go about it ? > > > > Your help is greatly appreciated. > > > > Thanks! > > Sunil > > Of course, you can do with the help of WindView configuring in > Post-Mortem mode. > post-mortem mode is helpful in collecting data on an application/task > failure resulting in a target reboot. > You will get more information about post-mortem mode in windview help. > > Amit Thanks Amit. But I am looking for collecting the dump in the run time environment, not in debug settings. If some task crashes at a customer site, I want to be able to save the task related data in a file. Customer can send this file to me for analyzing what went wrong. Any pointers ?? Thanks again, Sunil --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Wed Feb 20 04:03:21 2002 From: Vxworks Exploder Date: Wed Feb 20 04:03:24 PST 2002 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Wed Feb 20 04:03:16 PST 2002 Subject: Re: VirtuallAlloc failed error on compilation Subject: Re: how to release a mutex when the process is killed Subject: Re: ipFilterHookAdd() Problem Subject: Re: anticipating machine check error... Subject: Re: Writing a BSP from scratch Subject: Re: Writing a BSP from scratch Subject: Re: problem during debbuging!! Subject: Re: Is there a free database can be used on vxworks Subject: Re: Writing a BSP from scratch Subject: Re: Writing a BSP from scratch Subject: Re: MUX Load failed Subject: Why do I need to do eieio behavior after sysInByte() and sysOutByte()? Subject: IDE (ATA) interface on an MPC8260-based target Subject: Re: anticipating machine check error... Subject: Re: Writing a BSP from scratch Subject: Re: IDE (ATA) interface on an MPC8260-based target Subject: Re: IDE (ATA) interface on an MPC8260-based target Subject: Re: Why do I need to do eieio behavior after sysInByte() and sysOutByte()? Subject: Re: prjObjs.lst - error 0x80 - solution Subject: Re: IDE (ATA) interface on an MPC8260-based target Subject: Re: how to release a mutex when the process is killed Subject: Re: VirtuallAlloc failed error on compilation Subject: RTOS Feedback? Subject: Re: IDE (ATA) interface on an MPC8260-based target Subject: Re: how to release a mutex when the process is killed Subject: Re: RTOS Feedback? Subject: Re: PCI card Subject: PowerPC860 power-up reset Subject: Re: Writing a BSP from scratch Subject: fatal txStall error with motFccEnd drv wrSbc8260 bsp ver 33?? Subject: Re: Writing a BSP from scratch Subject: Re: What means this sentens...? Subject: Just a test(NULL!) Subject: Re: anticipating machine check error... Subject: Re: RTOS Feedback? Subject: Re: VirtuallAlloc failed error on compilation Subject: Help from the Networking Guru's of VxWorks. Subject: Re: Is there a free database can be used on vxworks Subject: structure sigwait not found in any header file Subject: MPC860 with Vision ICE Subject: Re: PCI card Subject: Wind Web Server memory problem ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: VirtuallAlloc failed error on compilation Date: Tue, 19 Feb 2002 12:22:54 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: <89441eba.0202141011.1e4c5bd@posting.google.com> "Bill Pringlemeir" wrote in message news:ud6z6obl6.fsf@yahoo.com... > I guess a file that did compile in 10 minutes and now takes over 60 is > status quo. As the fsf gcc developers went through post 2.95.3 development leading into the 3.0.x series, there were indeed *major* issues with compilation time: lots of the new C++ ABI stuff caused major increases. Towards the 3.0.0 release, they started to put some effort into bringing the increases back down, but 3.0.x is still far slower than 2.95.x which in turn is slower than 2.7.2. It's not like there's no reason for it: the compiler is doing more work, and generating better code and optimization as a result. So it's possible that WRS chose a bad point at which to branch off their 2.96 version from the FSF sources, in between the slowdown and the effort to speed it back up again. As you've realised, WRS' code does not correspond to any official fsf release version: 2.96 is a pretty good description of it though, since it brings to mind the brouhaha surrounding Red Hat's release of a buggy intermediate gcc version with RH 7.0 (or was it 7.1?) I forget exactly how it's done, but it is possible to get Gcc to profile itself and spit out some timing info at the end of a compile. Without looking it up, I'm not sure whether you'd need to recompile gcc itself with different configure options or whether you just need to pass in a particular flag to cc1plus. This information might give some useful clue - maybe you could use one of the -f flags to turn off a particular optimisation that was eating up most of the time. And at least some increase in compile time is unavoidable, particularly with templates, since there are combinatorially-growing search issues (IIUIC) involved with resolving them through several layers. This didn't used to be a problem only because Gcc never did it correctly before. DaveK - -- moderator of alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! Proud Member of the Exclusive "I have been plonked by Davee because he thinks I'm interesting" List Member # Master of Many Meowing Minions Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage above and beyond the call of hilarity. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: how to release a mutex when the process is killed Date: Tue, 19 Feb 2002 12:35:17 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: <3c719169$0$5154$724ebb72@reader3.ash.ops.us.uu.net> "Mark Galecki" wrote in message news:3c719169$0$5154$724ebb72@reader3.ash.ops.us.uu.net... VxWorks is not *nix. In particular, there is no automatic resource freeing when you kill a task. (Except under AE, which is another matter altogether). File handles are not closed, semaphores are not released, memory is not freed. You have to manage it all yourself at the application level. This is specifically warned about in the VxWorks programmer's guide, section 2.3.4 Tasking Control, subheading "Task Deletion and Deletion Safety". You should read the entire chapter. > I was expecting that, upon a kill, first the semaphore wait would exit with > an ERROR so I can clean up in the handler but that does not happen. Nope, upon killing a task, it just stops. You could use a signal handler or a taskdeletehook, as others have suggested, but one way or another, you must manage a graceful cleaup-and-exit process yourself. DaveK - -- moderator of alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! Proud Member of the Exclusive "I have been plonked by Davee because he thinks I'm interesting" List Member # Master of Many Meowing Minions Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage above and beyond the call of hilarity. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: ipFilterHookAdd() Problem Date: Tue, 19 Feb 2002 12:25:24 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: <3C6D11AD.385877BC@fallschurch.esys.com> "drdiags" wrote in message news:k9Db8.73821$AV5.372645@rwcrnsc51.ops.asp.att.net... > Hello John, > > I found this for ipFilterHookAdd(): It's all fully documented in the WRS manuals: look up ipFilterLib. I was assuming John's problem was that the routine doesn't behave as documented; if he hasn't read the manpage, though, he should, and just copy+paste the prototype supplied there... DaveK - -- moderator of alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! Proud Member of the Exclusive "I have been plonked by Davee because he thinks I'm interesting" List Member # Master of Many Meowing Minions Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage above and beyond the call of hilarity. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: anticipating machine check error... Date: Tue, 19 Feb 2002 14:44:06 +0100 From: "Michael Lawnick" Organization: Buergernetz Dillingen Message-ID: References: <8fe86b76.0202131506.3c00702d@posting.google.com> <8fe86b76.0202181259.e622c15@posting.google.com> <2ede7703.0202190158.1e19c22a@posting.google.com> Reply-To: "Michael Lawnick" Sender: mlawnick@pd9e65572.dip.t-dialin.net Hi kannan, if this works -congratulations-, but you have done what is forbidden in manual. Connecting via excVecSet will result in an call within interrupt context. Your prior index to signal handler seems the correct 2nd way (beside vxMemProbe) to me. Here is the platform setjmp/longjmp are implemented for (AFAIK). - -- Mit freundlichen Grüßen, Michael Lawnick ============================================== SOFTEC GmbH Tel +49-731-96600-0 Promenade 17 Fax +49-731-96600-23 D-89073 Ulm Michael Lawnick Germany lawnick@softec.de ============================================== "kannan" schrieb im Newsbeitrag news:2ede7703.0202190158.1e19c22a@posting.google.com... > dwh8@yahoo.com (David) wrote in message news:<8fe86b76.0202181259.e622c15@posting.google.com>... > > Michael, > > > > Thanks for the info. I was able to write a little test program which > > works, but there is a problem. > > > > The code section that will be generating the exceptions also relies on > > a pending call (msgQReceive). In my test code I reduplicated that > > effect. When the pending call is reached, the system completely fails > > and must be rebooted. > > > > Any Ideas on how to resolve this? > > > > Below is the test code I wrote: > > ------------------------------------------------------------------- > > jmp_buf exceptionPoint; > > > > /* handle exception */ > > void expHandler ( int task, int vecNum, ESFPPC *pEsf) > > { > > printf("\nTask %i caused vector # %i error", task, vecNum); > > longjmp(exceptionPoint, 1); > > } > > > > int gen(int memLocation) > > { > > ULONG * buf = (ULONG *) memLocation; // 0xffffffff causes > > exception > > char cbuf[10]; > > > > MSG_Q_ID msgQ = msgQCreate(10,10,0); > > > > excHookAdd((FUNCPTR)expHandler); > > > > for(;;){ > > > > msgQReceive(msgQ,cbuf,10,20); > > > > if (setjmp(exceptionPoint) == 0){ > > printf("\ntask %i ready to print\n",taskIdSelf()); > > printf("\n%x",*buf); // generate exception > > }else{ > > printf("\ntask %i resumed\n",taskIdSelf()); > > } > > > > } // end of for(;;) > > } // end of gen() > > > > ------------------------------------------------------------- > > > > "Michael Lawnick" wrote in message news:... > > > Hi David, > > > > > > if your problem is only 'first access', your function is vxMemProbe(). > > > If you fear card fail while operating, use excHookAdd() and setjmp/longjmp. > > > If you don't like the logMsg of the vxWorks excHandler, you will need > > > excVecSet(). Use setjmp/longjmp here too. > > > > > > Incrementing PC would be for 'transparent operation fail', i.e. if failing > > > task shouldn't see its fail. IMO this is a solution for players, not a solid > > > one. > > > > > > -- > > > Mit freundlichen Grüßen, > > > Michael Lawnick > > > ============================================== > > > SOFTEC GmbH Tel +49-731-96600-0 > > > Promenade 17 Fax +49-731-96600-23 > > > D-89073 Ulm Michael Lawnick > > > Germany lawnick@softec.de > > > ============================================== > > > > > > "David" schrieb im Newsbeitrag > > > news:8fe86b76.0202131506.3c00702d@posting.google.com... > > > > I was hoping somebody could verify this for me: > > > > > > > > I am porting an application from PSOS to VxWorks 5.4 on a Power PC > > > > platform (mv5100). > > > > > > > > The existing application has a set of exception handling routines that > > > > are installed (in the PSOS assembly code) to handle machine check > > > > errors. The errors in question occur when card signal processing is > > > > attempted on a card that is missing from the chassis. We don't wan't > > > > the system to halt if a card is missing or not working correctly. > > > > > > > > From reading previous posts, it looks like I will have to use the > > > > excVecSet() function, but those posts also mention manually > > > > incrementing the Program Counter which sounds a bit hairy. My current > > > > handlers get passed: > > > > > > > > vector offset for the exception, > > > > addr - next program location when the exception occurred > > > > ExceptionFramePtr - exception frame pointer > > > > > > > > but the only thing I do with this information is display it in the > > > > error log. > > > > > > > > Any ideas on how to proceed would be greatly appreciated, > > > > > > > > David > > > Hi Gurus, > > The above code i tried working with excVecSet it works fine. > The code is changed as follows: > ------------------------------------------------------------------- > > jmp_buf exceptionPoint; > > > #if 0 /*commented this will not work*/ > > /* handle exception */ > > void expHandler ( int task, int vecNum, ESFPPC *pEsf) > > { > > printf("\nTask %i caused vector # %i error", task, vecNum); > > longjmp(exceptionPoint, 1); > > } > #endif > > /* handle exception */ > static int machineCheckIsr (ESFPPC * pEsf ) > { > logMsg("Machine Check\n",0,0,0,0,0,0); > longjmp(exceptionPoints, 1); > } > > > > > int gen(int memLocation) > > { > > ULONG * buf = (ULONG *) memLocation; // 0xffffffff causes > > exception > > char cbuf[10]; > > > > MSG_Q_ID msgQ = msgQCreate(10,10,0); > > > >/* excHookAdd((FUNCPTR)expHandler);*//*commented this will not > work*/ > excVecSet ((FUNCPTR *)_EXC_OFF_MACH, FUNCREF > (machineCheckIsr));/*Do this*/ > > > > > for(;;){ > > > > msgQReceive(msgQ,cbuf,10,20); > > > > if (setjmp(exceptionPoint) == 0){ > > printf("\ntask %i ready to print\n",taskIdSelf()); > > printf("\n%x",*buf); // generate exception > > }else{ > > printf("\ntask %i resumed\n",taskIdSelf()); > > } > > > > } // end of for(;;) > > } // end of gen() > > > > ------------------------------------------------------------- > what i have posted previously is correct the longjmp( ) function > executes correctly in the contexts of signal handlers ,But it will not > work in > the case of Interrupt Handlers and some Exception > handlers.(e.g.excHookAdd > function) > > Thanks and Regards > kannan --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Writing a BSP from scratch Date: Tue, 19 Feb 2002 16:07:12 +0100 From: Markus Pietrek Organization: FS Forth Systeme GmbH Message-ID: References: <3C721ABC.4030605@spamm.me.l8s.co.uk> David Laight wrote: > I would start from one of the WRS BSP for something similar, then > add in the extra functionality as you need it. > If you can write device drivers and system setup code I'm not > sure the BSP porting kit contains anything useful! Just did this on the NetARM based on pid7t. Took me some time to get all running but buying the expensive BSP porting kit was not necessary. You should have some knowledge of vxWorks before starting :-) > > Once you have romInit.s done vxWorks will probably boot. > Then you need to fix WRS broken ARM interrupt and mmu code, > stop the RTC drifting and write device drivers for your > hardware. > > Get the serial port working early on, and write debug strings > to it - in romInit.s. > > Since this is all 'ongoing' work I don't know how you use the > 'project' facility as it seems to snapshot the BSP! Yes. But if the workspace is closed and opened again, the new or modified entries are present in the project facility. Of course there is still the possibility to use "make bootrom" and "make vxWorks" on the command line in the BSP directory > (It barfed on one of my BSP header files, so never generated > a project!) I don't had such problems. Bye, Markus Pietrek --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Writing a BSP from scratch Date: Tue, 19 Feb 2002 16:23:40 +0100 From: Markus Pietrek Organization: FS Forth Systeme GmbH Message-ID: References: <3C721ABC.4030605@spamm.me.l8s.co.uk> BTW: Does anyone know how to change the compiler options for projects created by the project facility. It always creates a Project.wpj and Makefile which uses the compiler option "-mcpu=arm8" instead of the correct "-march=armv4". Bye, Markus Pietrek --------------------------- Newsgroups: comp.os.vxworks Subject: Re: problem during debbuging!! Date: Tue, 19 Feb 2002 10:39:02 -0500 From: david lindauer Message-ID: <3C727195.E7BFD963@notifier-is.net> References: I'm pretty sure you cannot do this when debugging over a serial port. Check the manuals. David zvi wrote: > Hello dear vxWorkers! > > we develop a new board base on a ppc8260 processor. > We use Tornado2 with vxWorks 5.4. > We have in our board 2 serial channels, SMC and SCC > and now we want to use the SMC channel for debugging purpose. > In our project we defined the WDB serial > connection(INCLUDE_WDB_COMM_SERIAL)and the WDB system debugging > (INCLUDE_WDB_SYS) > > We set the boad rate to 38400 . > In order to verify the connection, we check the printout in the > terminal when it boots and we see the prompt: > > WDB READY > > which as explained in the manuals it demonstrates that all the > connections are correct. > > We also succeeded to attach the debugger (usually, but sometimes it > complains about rpc timeout error), > > but here the problem starts, when trying to attach the "system" mode > in the 'attach' option in the debug menu ,there is an error: > > attach system > > attaching to system failed > > could not suspend system context: > > API_REQUEST_TIMED_OUT > > If we try to attach any of the task the debugger work as expected. > > We really want to use system debug mode, and we will be gratefull for > a quick advice of how enable correctlly this debug mode, > > Thanks in advance, > > Best Regards, > > Zvi Gold. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Is there a free database can be used on vxworks Date: Tue, 19 Feb 2002 10:43:42 -0500 From: david lindauer Message-ID: <3C7272AD.8C931D1A@notifier-is.net> References: <92dbccc1.0202182218.72990392@posting.google.com> http://www.sleepycat.com/ Adom wrote: > Hello, > now , i'm designing a model for ip traffic monitor on vxworks. > i want to use database to store data,but i don't know wether > there is a free database can be used on vxworks? > > Thanks for any suggestion! > > Adom Chan > 2002-2-19 --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Writing a BSP from scratch Date: 19 Feb 2002 10:41:30 -0500 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: References: <3C721ABC.4030605@spamm.me.l8s.co.uk> Sender: bpringlemeir@DeadDuck >>> Harjinder Chayra wrote: >> Hello all, I am planning to write a BSP for the ARM CPU from >> scratch with the help of the BSP porting kit from WRS >> If you have been involved in this level of development I would be >> interested in learning, how helpful (or not) the BSP porting kit >> was how long it took to get the BSP working any recommendations >> e.g. base the new BSP on an exisiting working one. >>>>> "David" == David Laight writes: David> I would start from one of the WRS BSP for something similar, David> then add in the extra functionality as you need it. If you [snip more good advice] This is the way to go unless you are a silicon manufacturer. I was speculating that Harjinder might be in this position. "http://www.imedia.com/index2.html" At least if his `Terayon' is the same as this one. If you are a CPU manufacturer who is distributing BSPs to customers, then you would need the BSP porting guide and probably wish to setup the GUI/project interface. Otherwise it is not that valuable, as David noted... I wouldn't say it is worthless. It does give a good overview of the pieces that are needed. Unfortunately many of the vxWorks BSP hooks appear to be undocumented. A great deal of kernel functions will call a BSP hook (in varying ways). This is always a sneaky thing when writing a BSP and that is why existing source is a good place to start. hth, Bill Pringlemeir. - -- You can teach others to disappear. vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Writing a BSP from scratch Date: Tue, 19 Feb 2002 16:09:13 +0000 (UTC) From: David Laight Organization: BT Openworld Message-ID: <3C727903.6010407@spamm.me.l8s.co.uk> References: <3C721ABC.4030605@spamm.me.l8s.co.uk> Markus Pietrek wrote: > Just did this on the NetARM based on pid7t. Took me some time to get all > running but buying the expensive BSP porting kit was not necessary. > You should have some knowledge of vxWorks before starting :-) That is for wimps..... Start with a processor you don't know the instruction set of, dodgy development hardware that only partially works, and management who want it working yesterday. I found that most of the device drivers source code isn't any use. Especially if your hardware doesn't exactly match what it was written for. Try getting the ATA/IDE disk driver to talk to 2 pcmcia cards and 2 compact flash card in 'true IDE' mode at random IO address and non-standard port spacing (its a 32 bit bus...). The driver will only work with PC hardware..... Quicker to write one from scratch that sort out the HW dependencies. David --------------------------- Newsgroups: comp.os.vxworks Subject: Re: MUX Load failed Date: Tue, 19 Feb 2002 11:15:48 -0500 From: david lindauer Message-ID: <3C727A34.5CA515CA@notifier-is.net> References: <3C720BFE.8F260574@icn.siemens.de> hi, you want to look in \tornado\target\config\pcpentium\sysnetif.c to see what drivers are supported. There are several in there. The one using the intel chipset that they support is 'Intel Ethernet Express Pro100B lan'; if you don't have that kind of card expect to have to modify the driver (as per their comments in the file). I don't know how to figure out the exact names they use but I would hazard a guess that in the boot string for this one you would use 'FEIx' where x is the unit number. David Franz Ettengruber wrote: > Hello! > We are doing our first steps with VxWorks and try to Boot it on a > PC-Pentium Target. > We are using a pre-built Boot Rom that should support pci network cards > with Intel chipset, but we tried several different (all?) of them and > always get this error message: "MUX Load failed". > In our opinion it must have something to do with networking. > So our question is where can we find information about what types of > network cards VxWorks does support and what parameter to specify for > "boot device", when booting VxWorks. > Any help would be very appreciated. > Regards, > Franz --------------------------- Newsgroups: comp.os.vxworks Subject: Why do I need to do eieio behavior after sysInByte() and sysOutByte()? Date: Tue, 19 Feb 2002 09:49:41 -0700 From: "Liang Yang" Organization: RTS-LAB Message-ID: Reply-To: "Liang Yang" Hi, I need to use both sysInByte() and sysOutLib(), but just like vxImmrGet(), I can not find any description about these two functions in any vxWorks manual. Where can I find the help of them? BTW, someone told me I should use eieio behavior to sychronize memory after sysInByte() and sysOutByte(). What's reason to use eieio here? Thank you. Liang --------------------------- Newsgroups: comp.os.vxworks Subject: IDE (ATA) interface on an MPC8260-based target Date: 19 Feb 2002 08:58:58 -0800 From: falk_beni@rada.com (Beni Falk) Organization: http://groups.google.com/ Message-ID: Hello, all We are designing a proprietary CPU board based on the MPC8260. We are considering to implement an IDE interface on the board. The interface will be implemented using one of the MPC8260 universal programmable machines. We plan to use an ATA flash card with dosFs(2). Does anybody have experience with interfacing an IDE disk to a PowerPC target, using a similar (or other) hardware interface? How easy was it to customize the driver? According to its header comments, it seems that ATA driver (ataDrv.c) supports big-endian hosts, but our FAE is not sure. What about ideDrv? Does anybody have any benchmark data (actual sustained write throughput) under similar conditions? How about CPU loading? Are there any other problems that we need to think about? Thank you in advance. Beni Falk --------------------------- Newsgroups: comp.os.vxworks Subject: Re: anticipating machine check error... Date: 19 Feb 2002 09:11:16 -0800 From: kannan@comneti.com (kannan) Organization: http://groups.google.com/ Message-ID: <2ede7703.0202190911.17af22aa@posting.google.com> References: <8fe86b76.0202131506.3c00702d@posting.google.com> <8fe86b76.0202181259.e622c15@posting.google.com> <2ede7703.0202190158.1e19c22a@posting.google.com> Michael/David, It even works with signal handlers.Simply you have to add "signal (SIGBUS,myHandler)" instead of excVecSet(I agree your point Mr.Michael). vxMemProbe is also a another good option. Thanks and Regards, Kannan --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Writing a BSP from scratch Date: Tue, 19 Feb 2002 17:21:22 +0000 (UTC) From: David Laight Organization: BT Openworld Message-ID: <3C7289E4.70201@spamm.me.l8s.co.uk> References: <3C721ABC.4030605@spamm.me.l8s.co.uk> Bill Pringlemeir wrote: > A great deal of kernel functions will call > a BSP hook (in varying ways). This is always a sneaky thing when > writing a BSP and that is why existing source is a good place to > start. I also seem to recall some things going via 'architecture' dependant code - which you dont get the source for. A pain since you need to fiddle with some of it (eg the interrupt entry code). David --------------------------- Newsgroups: comp.os.vxworks Subject: Re: IDE (ATA) interface on an MPC8260-based target Date: Tue, 19 Feb 2002 17:24:55 +0000 (UTC) From: David Laight Organization: BT Openworld Message-ID: <3C728AC2.8050601@spamm.me.l8s.co.uk> References: Beni Falk wrote: > Hello, all > > We are designing a proprietary CPU board based on the MPC8260. We are > considering to implement an IDE interface on the board. The interface > will > be implemented using one of the MPC8260 universal programmable > machines. > We plan to use an ATA flash card with dosFs(2). > > Does anybody have experience with interfacing an IDE disk to a PowerPC > target, using a similar (or other) hardware interface? How easy was it > to > customize the driver? > > According to its header comments, it seems that ATA driver (ataDrv.c) > supports big-endian hosts, but our FAE is not sure. What about ideDrv? > > Does anybody have any benchmark data (actual sustained write > throughput) under similar conditions? How about CPU loading? > > Are there any other problems that we need to think about? I would write one from scratch, you (probably) only need PIO data transfers, and can to them in the task that calls the driver. it's not hard ;-) IIRC neither ataDrv nor ideDrv will work on anything except an x86 PC without more hackong than writing a working driver from scratch. David --------------------------- Newsgroups: comp.os.vxworks Subject: Re: IDE (ATA) interface on an MPC8260-based target Date: Tue, 19 Feb 2002 19:22:43 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: Hi Benny, will try to answer some of your questions: On the first thought, it seems simple to implement an IDE interface, because all it really is is a bus-to-bus interface, as the ATA interface has different timing requirements then the system bus. All the command, status and data registers are internal to the ATA disk drive, so as long as the IDE interface handles timing, register addressing and interrupts, you got yourself a reasonable IDE. It does need a programmable timing registers so the driver can negotiate the fastest speed (PIO mode) with the disk and set the hardware parameters accordingly. But when considering CPU load, the basic modus operandi of ATA is PIO )Programmed I/O) where the CPU transfers disk data by reading or writing the data register which is 16-bits wide. These operations must be then delayed by the IDE interface to be adopt to the ATA disk timing, and that could slow down the CPU quite a bit. An advanced IDE interface thus has a DMA, preferrably a BusMaster DMA, which does multi-sector data transfer between the ATA data registers and memory without CPU intervention. Depending on your application, if heavy use of the disk is expected, then the CPU load resulting from lack of DMA could be rather significant. Hope this helps, - - Leonid "Beni Falk" wrote in message news:fe5d1c3c.0202190858.39d41947@posting.google.com... > Hello, all > > We are designing a proprietary CPU board based on the MPC8260. We are > considering to implement an IDE interface on the board. The interface > will > be implemented using one of the MPC8260 universal programmable > machines. > We plan to use an ATA flash card with dosFs(2). > > Does anybody have experience with interfacing an IDE disk to a PowerPC > target, using a similar (or other) hardware interface? How easy was it > to > customize the driver? > > According to its header comments, it seems that ATA driver (ataDrv.c) > supports big-endian hosts, but our FAE is not sure. What about ideDrv? > > Does anybody have any benchmark data (actual sustained write > throughput) under similar conditions? How about CPU loading? > > Are there any other problems that we need to think about? > > Thank you in advance. > > Beni Falk --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Why do I need to do eieio behavior after sysInByte() and sysOutByte()? Date: Tue, 19 Feb 2002 17:29:39 +0000 (UTC) From: David Laight Organization: BT Openworld Message-ID: <3C728BDC.8070407@spamm.me.l8s.co.uk> References: Liang Yang wrote: > Hi, > > I need to use both sysInByte() and sysOutLib(), but just like > vxImmrGet(), I can not find any description about these two functions > in any vxWorks manual. Where can I find the help of them? They are based on x86 inb and outb instructions.... > > BTW, someone told me I should use eieio behavior to sychronize memory > after sysInByte() and sysOutByte(). What's reason to use eieio here? The eieio ensures that the PPC doesn't reorder any memory cycles. I don't know the exact rules, but you probably need to issue an eieio instruction between access to hardware if the order of accesses matters. For safety the sysInByte() routine should probably issue eieio before and after the requested transfer. Also remember that a write could still be buffered (eg in a PCI bridge) even after the eieio completes. Reading the same location back is necessary in order to flush the written data. David --------------------------- Newsgroups: comp.os.vxworks,gnu.gcc.help Subject: Re: prjObjs.lst - error 0x80 - solution Date: Tue, 19 Feb 2002 15:15:38 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: "Ofer Goren" wrote in message news:newscache$pv3srg$ma6$1@lnews.actcom.co.il... > Well, the problem is in the combination w/ cygwin, god knows why. > > If cygwin was installed before Torando, removing cygwin after the torando > installation will hard Tornado. > > The solution is to re-install cygwin, and update the window's PATH variable > to the location of the cygwin's bin directory. You must never have two different versions of the cygwin dll active at the same time. Also, if you have multiple versions of cygwin dll in your system, even though you may be careful to only use one at a time, they both share the same set of registry entries relating to *nix-mode mountpoints etc. So maybe uninstalling Cygwin removed the registry keys and wiped out stuff that the Tornado-cygwin needed in order to find files on disk? It would certainly be a good idea for WRS' documentation to go into a bit greater depth about the fact that Tornado is a largely cygwin-based application and the possible interactions between the two. DaveK - -- moderator of alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! Proud Member of the Exclusive "I have been plonked by Davee because he thinks I'm interesting" List Member # Master of Many Meowing Minions Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage above and beyond the call of hilarity. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: IDE (ATA) interface on an MPC8260-based target Date: Tue, 19 Feb 2002 13:14:16 -0500 From: david lindauer Message-ID: <3C7295F8.BAE1782A@notifier-is.net> References: Leonid Rosenboim wrote: > But when considering CPU load, the basic modus operandi of ATA > is PIO )Programmed I/O) where the CPU transfers disk data by reading > or writing the data register which is 16-bits wide. These operations > must be then delayed by the IDE interface to be adopt to the ATA disk > timing, > and that could slow down the CPU quite a bit. I'm not sure what the throughput issue is... AFAIK when you write out a sector it just gets buffered in the drive's ram until the drive is ready to schedule a write to the disk for it. So the actual write should be rather quick... it is just when the drive's ram is full you have to wait until it is ready to accept more data. Theoretically the driver should be pending on a semaphore at that point to let other things happen. I *believe* we used the ATA driver on the MPC860 during prototyping. I know for sure we were using the FDC driver, with only very minor modifications to the header files to make it compile. Usually the vxworks drivers eventually call the BSP to resolve endianness issues, actually do I/O, etc... David --------------------------- Newsgroups: comp.os.vxworks Subject: Re: how to release a mutex when the process is killed Date: Tue, 19 Feb 2002 10:40:06 -0800 From: "Mark Galecki" Message-ID: <3c729ba9$0$5158$724ebb72@reader3.ash.ops.us.uu.net> References: <3c719169$0$5154$724ebb72@reader3.ash.ops.us.uu.net> Thank you all for very helpful comments. I think I will try to use taskDeleteHookDelete() together with a global variable that stores the TCB pointer of the task holding the semaphore. The act of taking and releasing the semaphore together with the update of that variable will be guarded by taskSafe(). Once I am deleted, in the hook I will check that this is the task that is holding the semaphore, then release it. I can store the semaphore ID in another global, or, if I want to be a real stud, I can use one of the spare fields in the TCB. Does anybody see any problems with that before I go and do it?? Thank you again. "Mark Galecki" wrote in message news:3c719169$0$5154$724ebb72@reader3.ash.ops.us.uu.net... > Hello, > > I have what must be a common issue when writing a communication device > driver and surprisingly I don't find an easy solution in VxWorks manuals. I > am experienced with writing drivers for other OSs, but not VxWorks, and > probably I don't see an easy solution, so please help... > ---- > > Description of problem: > To prevent interleaving under multiple threads, the driver read() handler > first grabs a read mutex and it releases it when it returns. In this way > only one read handler is executing at any given time. This is probably the > way all drivers like that are written. > > Now, inside the handler I have to wait for data. This is a blocking read. > So I wait on a semaphore, and when the data arrives, the read ISR signals me > and I can read the data. This also is probably very common. > > The problem is this: the user starts a read app and it hangs, waiting for > data. Then the user notices something is wrong and no data will be coming, > so they kill the user app. I do it with the Tornado debugger by > interrupting it and/or stopping it, but there are other ways, such as > sending a kill from command line. > > Then the read mutex which was taken, is not released, so upon fixing the > problem, if one tries to start the read app again, the read handler won't > even enter. > > I was expecting that, upon a kill, first the semaphore wait would exit with > an ERROR so I can clean up in the handler but that does not happen. I was > also hoping that maybe the close handler would get called so I could > attempt a cleanup based on a task ID, but no, the close handler does not get > called either. > > So I have to reboot the machine if no data is coming - hardly a satisfactory > solution. > > Thank you, > Mark Galecki > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VirtuallAlloc failed error on compilation Date: Tue, 19 Feb 2002 13:12:48 -0600 From: Andrew Johnson Organization: APS, Argonne National Laboratory Message-ID: <3C72A3B0.A5912971@aps.anl.gov> References: <89441eba.0202141011.1e4c5bd@posting.google.com> Dave Korn wrote: > > As you've realised, WRS' code does not correspond > to any official fsf release version: 2.96 is a pretty good description of it > though, since it brings to mind the brouhaha surrounding Red Hat's release > of a buggy intermediate gcc version with RH 7.0 (or was it 7.1?) ISTR a WRS salesman saying that AE shipped with cross-compilers based on the same 2.96 gcc version as RH 7.whatever anyway. Red Hat = Cygnus, who WRS always used to contract out compiler development to and I assume they now do to RH. - - Andrew - -- Perfection is reached, not when there is no longer anything to add, but when there is no longer anything to take away. - - Antoine de Saint-Exupery --------------------------- Newsgroups: comp.arch.embedded,comp.os.qnx,comp.os.vxworks,comp.realtime Subject: RTOS Feedback? Date: Tue, 19 Feb 2002 17:04:12 GMT From: Dan Murphy Organization: Murphstein Et Filles Message-ID: I'm in the process of selecting an RTOS for a new project, and the sales people are more than willing to give me references that'll provide good reviews. I'm more interested in hearing from knowledgeable, experienced people with sound technical reasons to regret having chosen their current RTOS. I'm looking for a few senior (5+ years) embedded system SW engineers who would be willing to share (confidentially) their current or recent experiences with: - - developing complex high-availability (five nines) systems - - using a loosely-coupled, distributed-processor architecture - - with a few hard R/T functions (deadlines of ~0.0001 sec) - - using one of {QNX-6, OSE-Delta, VxWorks-AE, Integrity} - - on an embedded 32-bit uC with MMU and 2 Fast Ethernets - - with extensive reliance on software OS or target simulation Please reply via email to the following address, after de-spamming it: nospam murphstein at nospam sbcglobal net Thanks in advance, Dan Murphy --------------------------- Newsgroups: comp.os.vxworks Subject: Re: IDE (ATA) interface on an MPC8260-based target Date: Tue, 19 Feb 2002 21:48:26 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <3C7295F8.BAE1782A@notifier-is.net> "david lindauer" wrote in message news:3C7295F8.BAE1782A@notifier-is.net... > > > Leonid Rosenboim wrote: > > > But when considering CPU load, the basic modus operandi of ATA > > is PIO )Programmed I/O) where the CPU transfers disk data by reading > > or writing the data register which is 16-bits wide. These operations > > must be then delayed by the IDE interface to be adopt to the ATA disk > > timing, > > and that could slow down the CPU quite a bit. > > I'm not sure what the throughput issue is... AFAIK when you write out a > sector it just gets buffered in the drive's ram until the drive is ready to > schedule a write to the disk for it. So the actual write should be rather > quick... it is just when the drive's ram is full you have to wait until it is > ready to accept more data. Theoretically the driver should be pending on a > semaphore at that point to let other things happen. > Used to be true, but now the ATA bus could be many times slower then the system bus and the CPU, understandably, the IDE bus and ATA device are not expected to sustain any transfer rate the CPU could generate. For example, if the ATA max trpt at a certain PIO level is 2MBs, it would take 0.5 millisecond to transfer a 512 byte sector, no matter how fast is your CPU. Moreover, if the disk is of the solid-state variety, and you are mostly reading, one could easily hog the CPU to 100% just by reading data constantly from the disk, thereby utilizing 100% of the disk throughput. Many people I met find it difficult to understand the effect of various I/O onto the performance of the CPU. - - Leonid --------------------------- Newsgroups: comp.os.vxworks Subject: Re: how to release a mutex when the process is killed Date: Tue, 19 Feb 2002 15:13:35 -0500 From: Keith Arner Organization: Marconi Message-ID: References: <3c719169$0$5154$724ebb72@reader3.ash.ops.us.uu.net> Reply-To: Keith Arner On Mon, 18 Feb 2002, Mark Galecki wrote: > Description of problem: > To prevent interleaving under multiple threads, the driver read() handler > first grabs a read mutex and it releases it when it returns. In this way > only one read handler is executing at any given time. This is probably the > way all drivers like that are written. > > Now, inside the handler I have to wait for data. This is a blocking read. > So I wait on a semaphore, and when the data arrives, the read ISR signals me > and I can read the data. This also is probably very common. > > The problem is this: the user starts a read app and it hangs, waiting for > data. Then the user notices something is wrong and no data will be coming, > so they kill the user app. I do it with the Tornado debugger by > interrupting it and/or stopping it, but there are other ways, such as > sending a kill from command line. As has been posted before, when the task goes away, there is no automatic cleanup of its "resources", whether they be semaphores, file descriptors, malloc()'d memory, or what have you. If the semaphore you are dealing with is a mutex semaphore, you have the option of making the semaphore deletion safe, which will prevent the task from going away while it still holds the mutex. However, I would warn you away from that path. The problem is that your task is safe while it holds that semaphore, but not necessarily any other semaphore. What happens if your task gets deleted while holding the semaphore that protects the heap, or the semaphore that protects the fd table, or *insert your favorite semaphore here*. Unless all resources are protected by deletion safe mutex semaphores (and not binary semaphores initialized to full, or any of a number of other mechanisms), there is always a chance that you will delete the task at just the wrong moment, and disaster will ensue. When people ask me the question "How should I delete a task?" I tell them not to delete the task at all. Rather tell the task that it is time to go away, and allow it to exit normally out of its entry function (or possibly call exit(), if you really know what you are doing). This means that taskDelete(), taskDestroy(), kill(), signal() and any other function that allows one task to force another to go away should be considered off limits for anything other than debugging. I realize that this pretty much means that you cannot do what you want: provide the user with the ability to kill off the task even if it is stuck in some I/O code. But no matter how you spin it, there will eventually be some timing issue where killing the task will spell disaster. Keith --------------------------- Newsgroups: comp.arch.embedded,comp.os.qnx,comp.os.vxworks,comp.realtime Subject: Re: RTOS Feedback? Date: 19 Feb 2002 20:29:22 GMT From: "Mark A. Odell" Organization: Embedded Firmware Design, Inc. Message-ID: References: Followup-To: comp.arch.embedded,comp.realtime Dan Murphy wrote in news:gAvc8.787$gw4.235809918@newssvr13.news.prodigy.com: > I'm in the process of selecting an RTOS for a new project, and the sales > people are more than willing to give me references that'll provide good > reviews. > > I'm more interested in hearing from knowledgeable, experienced people > with sound technical reasons to regret having chosen their current > RTOS. I'm looking for a few senior (5+ years) embedded system SW > engineers who would be willing to share (confidentially) their current > or recent experiences with: > > - developing complex high-availability (five nines) systems > - using a loosely-coupled, distributed-processor architecture > - with a few hard R/T functions (deadlines of ~0.0001 sec) > - using one of {QNX-6, OSE-Delta, VxWorks-AE, Integrity} I liked OSE-Delta except the build config/process was in need of some automation. - -- - - Mark A. Odell - - Embedded Firmware Design, Inc. - - http://www.embeddedfw.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: PCI card Date: Tue, 19 Feb 2002 20:32:06 GMT From: "Michael B." Organization: [Infostrada] Message-ID: References: <3C71317A.A31D4A45@notifier-is.net> Thank you David, but you've discovered the biggest of my problems: everything I know about my card is what VxWorks can say through PCI functions and registers' offsets on the board, starting from an address that actually I can't reach. Once examinated other drivers I can see that BAR0 is used for I/O mapping and BAR1 for memory mapping. Memory mapping requires sysMmuMapAdd for BAR1 but I/O mapping takes BAR0 and symply add it to CPU_PCI_IO_ADRS=0 ...something wrong...why 0? Would you like to help me to write few code lines to access to a simbolic offset OFF in my card? Should I use 'sysInXXX' or 'sysOutXXX' as mentioned in x86 BSP Ref. of VxWorks Prog.'s Guide? Thanks, again! Mic. "david lindauer" ha scritto nel messaggio news:3C71317A.A31D4A45@notifier-is.net... > if the card is designed to have both memory and I/O ports associated with > it, generally that will be setable in the PCI configuration space and then > you just access the card directly and ignore PCI. This configuration will > be card dependent. For a simple ATA PCI controller I set BAR0 and BAR1 to > the two ATA addresses to be used, and PCI_CFG_DEV_INT_LINE to the interrupt > to use, and it worked ok. > > If you have DOS/BIOS you can sometimes get a driver to configure your > device, then read the PCI configuration registers to determine what to do > next. > > David > > "Michael B." wrote: > > > Hi! > > I'm working on a I/O PCI card and using pciHeaderShow I obtained a base > > address 0 (BAR0) equal to 0xe4000000 and some other items (are they > > important?). > > How can I use BAR0 to access to card's register? > > What about I/O mapped and/or memory mapped devices? > > My BSP is a default PcPentium. > > > > Thanks. > > Mic. > --------------------------- Newsgroups: comp.os.vxworks Subject: PowerPC860 power-up reset Date: 19 Feb 2002 13:09:44 -0800 From: brightpolestar@yahoo.com (Daniel) Organization: http://groups.google.com/ Message-ID: <28ee7c97.0202191309.23df317d@posting.google.com> Hi, When powerpc 860 powers-up, it goes to 0x100 interrupt vector, from there, it goes to ROM to fetch instructions. My question is that 0x100 is in RAM, and there is nothing or just garbage in RAM at system initialization time, how does the system know to go to ROM? Thanks a lot. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Writing a BSP from scratch Date: Tue, 19 Feb 2002 16:46:14 -0500 From: "Harjinder Chayra" Organization: Storm Internet Services Message-ID: References: Sender: terayon@gw-terayon.storm.ca Thanks for all the input, looks like I am in for a *busy* time ahead of me. Harjinder PS no doubt I will be runing into issues in due course so the group will be hearing from me again! "Harjinder Chayra" wrote in message news:a4rmso$l9l$1@news.storm.ca... > Hello all, > I am planning to write a BSP for the ARM CPU from scratch with > the help of the BSP porting kit from WRS. > If you have been involved in this level of development I would > be interested in learning, > > how helpful (or not) the BSP porting kit was > how long it took to get the BSP working > any recommendations e.g. base the new BSP on an exisiting working one. > > Looking forward to your comments! > > Harjinder > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: fatal txStall error with motFccEnd drv wrSbc8260 bsp ver 33?? Date: Tue, 19 Feb 2002 17:47:13 -0500 From: Babu Yama Organization: Lucent Technologies Message-ID: <3C72D5F1.251C8FF3@lucent.com> Hi, I am seeing that after couple of hours of socket communication (10 UDP and 10 TCP, transmitting every 15 sec or so) between two of my targets, motFccEnd driver is going into txStall condition and NEVER recovering. The receives are being processed by the ISR and tNetTask, but transmission never happens. What might be causing this? and how I can fix this? Should I be looking at Why we are getting into the stall condition? Or, why the driver is not clearing the stall condition? Any help will be greatly appreciated. Thanks, Babu The details are: Number of tx Buffer Descriptors : 64 in DPRAM Number of rx Buffer Descriptors : 64 in DPRAM the actual buffers are in SDRAM or local bus. FCC1 and FCC3 are being used for ethernet using motFccEnd driver. Tornado version 2 BSP: wrSbc8260 version 1.1.2/33 version 33 for MPC8260 When the problem happens the pDrvCtrl->stats are: tbdNum=64 tbdBase=0xff002800 tbdIndex=7 usedTbdIndex=15 cleanTbdNum=6 txStall=1 numStallsEntered=1 (PROBLEM) numStallsCleared=0 numIsrEntries=53148 numIsrExits=53148 numZcopySends=22481 numNonZcopySends=0 numTXBInts=22444 numRXFInts=30757 numrecoveryeasy=0 numRXFramesLostToBufferReuse=0 numRXFHandlerEntries=30341 numRXFHandlerExits=30341 numRXFHandlerErrQuits=0 numRXFHandlerEmptyLoopBreaks=0 numRXFHandlerFramesProcessed=30757 numRXFHandlerFramesRejected=0 --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Writing a BSP from scratch Date: Tue, 19 Feb 2002 16:58:54 -0600 From: Dinker Charak Organization: Fermi national Accelerator Lab Message-ID: References: <3C721ABC.4030605@spamm.me.l8s.co.uk> Curious ... What is this BSP porting kit everyone is refering to? Is it provided by WRS? Part of Tornado? Or provided by the board manufacturer? Is it usually provided for all boards? Dinker --------------------------- Newsgroups: comp.os.vxworks Subject: Re: What means this sentens...? Date: Wed, 20 Feb 2002 09:17:21 +0900 From: "kimp" Organization: Korea Research Environment Open Network Message-ID: References: Thanks for your kind :-) "Gary M" wrote in message news:vrnc8.85647$fK1.4222437@rwcrnsc54... > SYPCR is the System Protection Control Register, residing in the System > Interface Unit (SIU). The instruction you specifically refer to stores the > contents of R5 (0xFFFFFF03) into this register. R4 points to the base of the > internal memory map. Note that this register is write-once, meaning that it > can be written to only one time after reset. > > The value of 0xFFFFFF03 will disable the Bus Monitor and the Software > Watchdog. > > Refer to section 11.4.3, page 11-9, in the MPC860 User Manual. > > "kimp" wrote in message > news:a4su05$mh$1@news.kreonet.re.kr... > > Hi, > > > > I look into the BSP for MPC860. > > But I couldn't undrestand the meaning of following sentence in romInit.s > of > > BSP. > > Please let me know what these means. > > > > ---------------------------------------- > > > > /* Setup SYPCR for normal and proper background mode operation */ > > > > lis r5,0xffff /* ffffff88*/ > > ori r5,r5,0xff03 > > stw r5,SYPCR(0)(r4) ==> I couldn't understand this line. > > > > --------------------------------------- > > > > Thanks to your attention :-) > > > > > > > > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Just a test(NULL!) Date: Wed, 20 Feb 2002 09:26:51 +0800 From: "Hill_liu" Organization: Changzhou Telecom Message-ID: --------------------------- Newsgroups: comp.os.vxworks Subject: Re: anticipating machine check error... Date: 19 Feb 2002 18:30:45 -0800 From: dwh8@yahoo.com (David) Organization: http://groups.google.com/ Message-ID: <8fe86b76.0202191830.1dd3febb@posting.google.com> References: <8fe86b76.0202131506.3c00702d@posting.google.com> <8fe86b76.0202181259.e622c15@posting.google.com> <2ede7703.0202190158.1e19c22a@posting.google.com> Alright! Thanks to everybody for your input. To clarify; the code I posted worked on my system if I removed any pend calls (msgQRecieve), but the loop would hog the CPU making it hard to exit. I had origionally tried taskDelay() instead of the msgQ, but that also crashed the system. Anyway, I had used a method I had found posted on windsurf (http://web2.windriver.com/cgi-bin/windsurf/techtips/public/viewSum.cgi?93 code posted below for your enjoyment) that showed how to use excVecSet similarly to how kannan had described, except that the tech tip manually incremented the PC counter and used eieio after the exception generator. Someone earlier (first reply?) had mentioned that incrementing the PC was not a solid solution, so I believe I will try to implement it instead with setjmp/longjump as kannan described. This should also alleviate the need to do an eieio;sync after each potential machine check error. Thank again for your help. David - ----begin code--------------- #include "vxWorks.h" #include "stdio.h" #include "excLib.h" #include "taskLib.h" #include "sysLib.h" #include "esf.h" #ifndef EIEIO_SYNC #define EIEIO_SYNC __asm__ volatile ("eieio;sync") #endif int daCount = 0; static int daIsr ( ESFPPC * pEsf /* pointer to exception stack frame */ ) { pEsf->regSet.pc += 4; daCount++; return 0; } void daTest ( int count ) { int i = 0; volatile int value; printf ("Start test with loop count of %d. ", count); excVecSet ((FUNCPTR *) _EXC_OFF_DATA, FUNCREF(daIsr)); for (i=0; i < count; i++) { value = *((volatile int *) 0x40500000); /* cause exception */ EIEIO_SYNC; taskDelay (sysClkRateGet() / 10); } printf ("DA EXC called %d times. ", daCount); } - ----------------------------------- kannan@comneti.com (kannan) wrote in message news:<2ede7703.0202190158.1e19c22a@posting.google.com>... > dwh8@yahoo.com (David) wrote in message news:<8fe86b76.0202181259.e622c15@posting.google.com>... > > Michael, > > > > Thanks for the info. I was able to write a little test program which > > works, but there is a problem. > > > > The code section that will be generating the exceptions also relies on > > a pending call (msgQReceive). In my test code I reduplicated that > > effect. When the pending call is reached, the system completely fails > > and must be rebooted. > > > > Any Ideas on how to resolve this? > > > > Below is the test code I wrote: > > ------------------------------------------------------------------- > > jmp_buf exceptionPoint; > > > > /* handle exception */ > > void expHandler ( int task, int vecNum, ESFPPC *pEsf) > > { > > printf("\nTask %i caused vector # %i error", task, vecNum); > > longjmp(exceptionPoint, 1); > > } > > > > int gen(int memLocation) > > { > > ULONG * buf = (ULONG *) memLocation; // 0xffffffff causes > > exception > > char cbuf[10]; > > > > MSG_Q_ID msgQ = msgQCreate(10,10,0); > > > > excHookAdd((FUNCPTR)expHandler); > > > > for(;;){ > > > > msgQReceive(msgQ,cbuf,10,20); > > > > if (setjmp(exceptionPoint) == 0){ > > printf("\ntask %i ready to print\n",taskIdSelf()); > > printf("\n%x",*buf); // generate exception > > }else{ > > printf("\ntask %i resumed\n",taskIdSelf()); > > } > > > > } // end of for(;;) > > } // end of gen() > > > > ------------------------------------------------------------- > > > > "Michael Lawnick" wrote in message news:... > > > Hi David, > > > > > > if your problem is only 'first access', your function is vxMemProbe(). > > > If you fear card fail while operating, use excHookAdd() and setjmp/longjmp. > > > If you don't like the logMsg of the vxWorks excHandler, you will need > > > excVecSet(). Use setjmp/longjmp here too. > > > > > > Incrementing PC would be for 'transparent operation fail', i.e. if failing > > > task shouldn't see its fail. IMO this is a solution for players, not a solid > > > one. > > > > > > -- > > > Mit freundlichen Grüßen, > > > Michael Lawnick > > > ============================================== > > > SOFTEC GmbH Tel +49-731-96600-0 > > > Promenade 17 Fax +49-731-96600-23 > > > D-89073 Ulm Michael Lawnick > > > Germany lawnick@softec.de > > > ============================================== > > > > > > "David" schrieb im Newsbeitrag > > > news:8fe86b76.0202131506.3c00702d@posting.google.com... > > > > I was hoping somebody could verify this for me: > > > > > > > > I am porting an application from PSOS to VxWorks 5.4 on a Power PC > > > > platform (mv5100). > > > > > > > > The existing application has a set of exception handling routines that > > > > are installed (in the PSOS assembly code) to handle machine check > > > > errors. The errors in question occur when card signal processing is > > > > attempted on a card that is missing from the chassis. We don't wan't > > > > the system to halt if a card is missing or not working correctly. > > > > > > > > From reading previous posts, it looks like I will have to use the > > > > excVecSet() function, but those posts also mention manually > > > > incrementing the Program Counter which sounds a bit hairy. My current > > > > handlers get passed: > > > > > > > > vector offset for the exception, > > > > addr - next program location when the exception occurred > > > > ExceptionFramePtr - exception frame pointer > > > > > > > > but the only thing I do with this information is display it in the > > > > error log. > > > > > > > > Any ideas on how to proceed would be greatly appreciated, > > > > > > > > David > > > Hi Gurus, > > The above code i tried working with excVecSet it works fine. > The code is changed as follows: > ------------------------------------------------------------------- > > jmp_buf exceptionPoint; > > > #if 0 /*commented this will not work*/ > > /* handle exception */ > > void expHandler ( int task, int vecNum, ESFPPC *pEsf) > > { > > printf("\nTask %i caused vector # %i error", task, vecNum); > > longjmp(exceptionPoint, 1); > > } > #endif > > /* handle exception */ > static int machineCheckIsr (ESFPPC * pEsf ) > { > logMsg("Machine Check\n",0,0,0,0,0,0); > longjmp(exceptionPoints, 1); > } > > > > > int gen(int memLocation) > > { > > ULONG * buf = (ULONG *) memLocation; // 0xffffffff causes > > exception > > char cbuf[10]; > > > > MSG_Q_ID msgQ = msgQCreate(10,10,0); > > > >/* excHookAdd((FUNCPTR)expHandler);*//*commented this will not > work*/ > excVecSet ((FUNCPTR *)_EXC_OFF_MACH, FUNCREF > (machineCheckIsr));/*Do this*/ > > > > > for(;;){ > > > > msgQReceive(msgQ,cbuf,10,20); > > > > if (setjmp(exceptionPoint) == 0){ > > printf("\ntask %i ready to print\n",taskIdSelf()); > > printf("\n%x",*buf); // generate exception > > }else{ > > printf("\ntask %i resumed\n",taskIdSelf()); > > } > > > > } // end of for(;;) > > } // end of gen() > > > > ------------------------------------------------------------- > what i have posted previously is correct the longjmp( ) function > executes correctly in the contexts of signal handlers ,But it will not > work in > the case of Interrupt Handlers and some Exception > handlers.(e.g.excHookAdd > function) > > Thanks and Regards > kannan --------------------------- Newsgroups: comp.arch.embedded,comp.os.qnx,comp.os.vxworks,comp.realtime Subject: Re: RTOS Feedback? Date: Wed, 20 Feb 2002 02:46:48 GMT From: dan@nospam.com (Dan) Organization: Customer of Connect.com.au Pty. Ltd. Message-ID: <3c730dda.3781586353@news.per.connect.com.au> References: On Tue, 19 Feb 2002 17:04:12 GMT, Dan Murphy wrote: >willing to share (confidentially) their current >or recent experiences with: >Please reply via email to the following address, after de-spamming it: Why not share with the rest of us? Dan --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VirtuallAlloc failed error on compilation Date: Tue, 19 Feb 2002 19:16:23 -0800 From: Lori Fraleigh Organization: Real-Time Innovations Message-ID: <3C731507.12900AEA@rti.com> References: <89441eba.0202141011.1e4c5bd@posting.google.com> <3C72A3B0.A5912971@aps.anl.gov> Here's the official story on Wind's gcc 2.96: "The Tornado AE GNU compiler distribution is called 2.96. It is not related to the Red Hat compiler distributed with Red Hat Linux 7.0. The identical name is an unfortuante coincidence." >From my understanding, Wind's gcc 2.96 is a modified gcc 2.95.2. lori Andrew Johnson wrote: > > Dave Korn wrote: > > > > As you've realised, WRS' code does not correspond > > to any official fsf release version: 2.96 is a pretty good description of it > > though, since it brings to mind the brouhaha surrounding Red Hat's release > > of a buggy intermediate gcc version with RH 7.0 (or was it 7.1?) > > ISTR a WRS salesman saying that AE shipped with cross-compilers based on > the same 2.96 gcc version as RH 7.whatever anyway. Red Hat = Cygnus, who > WRS always used to contract out compiler development to and I assume they > now do to RH. > > - Andrew > -- > Perfection is reached, not when there is no longer anything to add, > but when there is no longer anything to take away. > - Antoine de Saint-Exupery - -- Lori Fraleigh Director of Engineering Real-Time Innovations, Inc. http://www.rti.com Phone: 408-734-4200 x121 Fax: 408-734-5009 --------------------------- Newsgroups: comp.os.vxworks Subject: Help from the Networking Guru's of VxWorks. Date: 19 Feb 2002 20:36:18 -0800 From: ramit.bhalla@wipro.com (Ramit Bhalla) Organization: http://groups.google.com/ Message-ID: <78ccdd99.0202192036.2055eec1@posting.google.com> Hi, I was hoping to get some answers from the Networking Guru's here. We can create an clBlk and Cluster combination with a user defined free routine function by using NetClBlkJoin. Here we supply a pre-allocated Cluster and clBlk along with a User Defined free routine. Here's the question : 1. How does the OS free up the Cluster and clBlk and return them to the pool ? 2. If we give give a NULL in the free routine pointer then what happned when the combo is freed ? Does the OS put it's own free routine into the pointer and does it free it up or not ? 3. If we supply a free routine, now in the routine how do I free the cluster/clBlk combination ??? 4. And if I dont' free up the combination will the OS automatically free it up after I return from the call ?? Any help here would be appreaciated, It's almost like a black box in this Area, Thanks a lot Ramit. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Is there a free database can be used on vxworks Date: 19 Feb 2002 21:59:52 -0800 From: chwin@china.com (Adom) Organization: http://groups.google.com/ Message-ID: <92dbccc1.0202192159.34231c59@posting.google.com> References: <92dbccc1.0202182218.72990392@posting.google.com> <3C7272AD.8C931D1A@notifier-is.net> david lindauer wrote in message news:<3C7272AD.8C931D1A@notifier-is.net>... > http://www.sleepycat.com/ > Thank you very much , david! i have downloaded and study this database. Adom Chan 2002-2-20 --------------------------- Newsgroups: comp.os.vxworks Subject: structure sigwait not found in any header file Date: 19 Feb 2002 23:25:14 -0800 From: narendra_upadhyay@delhi.tcs.co.in (Narendra Upadhyay) Organization: http://groups.google.com/ Message-ID: <8f91e136.0202192325.252cdcf1@posting.google.com> Hi, The structure sigwait is being used in the definition of sigtcs structure in the file tornado\target\h\private\sigLibP.h as given below: struct sigtcb /* (Not posix) */ { struct sigaction sigt_vec[_NSIGS + 1]; struct sigq sigt_qhead[_NSIGS + 1]; struct sigstack sigt_stack; struct sigwait *sigt_wait; sigset_t sigt_kilsigs; sigset_t sigt_blocked; sigset_t sigt_pending; sigset_t sigt_ignored; }; But the definition of the structure sigwait is not available in any of the header files in tornado\target\h directory. Could any one please tell me from where to take the definition of sigwait structure? Thanks You. Narendra Upadhyay --------------------------- Newsgroups: comp.os.vxworks Subject: MPC860 with Vision ICE Date: 20 Feb 2002 00:41:14 -0800 From: nick_nw@yahoo.co.uk (Nick) Organization: http://groups.google.com/ Message-ID: <2413908f.0202200041.37818aa5@posting.google.com> All, Whilst developing for a BlueWave Systems CPCI6400 (MPC860) card, we have come across a nasty bug. We still don't know what the bug is, but during 48 hour loaf testing our system would reset. We bought a Vision ICE to trap the problem and to debug our code. Problem is that with the ICE connected via the JTAG interface, the system never crashes. The ICE itself makes our system reliable, but never reports any trapped errors, etc. So, why is an alledgedly non-intrusive device altering our system? What is the card doing differently with the ICE. Cheers for any ideas, N. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: PCI card Date: Wed, 20 Feb 2002 09:57:43 +0100 From: "Werner Schiendl" Message-ID: <3c7364f4$1@brateggebdc5.br-automation.co.at> References: <3C71317A.A31D4A45@notifier-is.net> Hi Mic, which of the BARx entries contains which base address is completely up to your card. You wrote, BAR0 contains 0xe4000000? Do you work on an x86 system? One with a BIOS? Then this BAR0 is most certainly a memory base address. You can find out the exact type of the memory mapped from the last couple bits of the BAR (the _least_ significant bits). It's quite some time since I read the docs, but I think if bit 0 was 0, then it is a memory mapping. Further, I/O mappings are always < 0xFFFF (at least for the systems that I have seen). So if you've sysMmuMapAdd() 'ed the entry properly (at system init time, in sysLib.c), then you should be able to access that memory. If, on the other hand you have IO mapped addresses, you need to access them with sys(In|Out)(Byte|Word|Long) on x86 CPU: The use _different_ instructions to use IO address space vs. memory address space. In case you need more help, post the output of pciHeaderShow. Try to get some docs on your hardware, and try to get your hands on the PCI specification. hth Werner "Michael B." wrote in message news:aDyc8.1870$Zf3.49218@twister2.libero.it... > Thank you David, > but you've discovered the biggest of my problems: everything I know about my > card is what VxWorks can say through PCI functions and registers' offsets on > the board, starting from an address that actually I can't reach. > Once examinated other drivers I can see that BAR0 is used for I/O mapping > and BAR1 for memory mapping. Memory mapping requires sysMmuMapAdd for BAR1 > but I/O mapping takes BAR0 and symply add it to CPU_PCI_IO_ADRS=0 > ...something wrong...why 0? > Would you like to help me to write few code lines to access to a simbolic > offset > OFF in my card? Should I use 'sysInXXX' or 'sysOutXXX' as mentioned in x86 > BSP Ref. of VxWorks Prog.'s Guide? > > Thanks, again! Mic. > > "david lindauer" ha scritto nel messaggio > news:3C71317A.A31D4A45@notifier-is.net... > > if the card is designed to have both memory and I/O ports associated with > > it, generally that will be setable in the PCI configuration space and then > > you just access the card directly and ignore PCI. This configuration will > > be card dependent. For a simple ATA PCI controller I set BAR0 and BAR1 to > > the two ATA addresses to be used, and PCI_CFG_DEV_INT_LINE to the > interrupt > > to use, and it worked ok. > > > > If you have DOS/BIOS you can sometimes get a driver to configure your > > device, then read the PCI configuration registers to determine what to do > > next. > > > > David > > > > "Michael B." wrote: > > > > > Hi! > > > I'm working on a I/O PCI card and using pciHeaderShow I obtained a base > > > address 0 (BAR0) equal to 0xe4000000 and some other items (are they > > > important?). > > > How can I use BAR0 to access to card's register? > > > What about I/O mapped and/or memory mapped devices? > > > My BSP is a default PcPentium. > > > > > > Thanks. > > > Mic. > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Wind Web Server memory problem Date: Wed, 20 Feb 2002 11:08:20 +0200 From: "Yaron Cohen" Message-ID: <3c736849$1@news.012.net.il> Hi, I would like to ask for help regarding a WInd Web Server problem. I think that I have a memory problem but I don't understand well the Wind Web memory constants. I wrote the following SSI HTML page: - ---------------------------------------------------------------------------- - ----------------------------------------------------------------------- Untitled Document
And more.... - ---------------------------------------------------------------------------- - ----------------------------------------------------------------- The problem is that when I added more functions to the Java Script section the page was partly loaded. I.e. I got a blank page and when I looked at the source code I only saw: Untitled Document > > > > > name="pageID"> > >
>
>
>
>
> > > > And more.... > > ---------------------------------------------------------------------------- > ----------------------------------------------------------------- > > The problem is that when I added more functions to the Java Script section > the page was partly loaded. > I.e. I got a blank page and when I looked at the source code I only saw: > > > > Untitled Document > > > > > > > name="pageID"> > >
>
>
>
>
> > > > And more.... > > -------------------------------------------------------------------------- - -- > ----------------------------------------------------------------- > > The problem is that when I added more functions to the Java Script section > the page was partly loaded. > I.e. I got a blank page and when I looked at the source code I only saw: > > > > Untitled Document > >

J1H"[&R./QK'[:Y\>(IM6-X5+G\B"C?WW:Y'(\ X MOW7"^F21PYY>RU(K.G]\U.J=3Y&%V$5(MLK"R$ZS6H=.X(OZ=)_ZK$J9,J<2 M;NR?+QZZ[E]+J:"#Z%(.Z1"A'RV8Z-GX\)D!$/[82^F3(Z6X)QTO"P8 M=ZH-2+'N/U 0*(;^T)8:*] (C"W]]:B\FY(D*9&WN@\FC*2Q((YO(&@FZZ^2AM*!1Y>=4N=5.:UOY77,B7,2<7P"M-Z&E@.0057> M-O9>*/-(0U4K>HM\ZNK"KHAO!%P#(GT1S')>U&0'FI_-8NGP-L4X*F*&M=_I M5XV!<1H$3=0^RZ=>4=;X$N,81J\!D9IVL$<9#-*]S.SH[=92'@R-GKHJW?,:?LY^KU M1LT:1*,#$@=Z1W-)VJ&UZW$E0GUI[OORN::N5]Z&#-2#`6J>-IL*8268/&N: M_MY<* M>:%"QA1D`-Y6HM=H?24+G%HI2T T]N=P/N:9: \T4E(_\_]_IF%M8V*0G*&C MLC=25U8<:QLAMLECN)<$7))E.')&#;V9W;'?71E:8I0"# 6Y\.-#6A[;0,MV M7 N)Y$8ZDE.\O/J9;#1<'A_N3VYQ_=RG GRK1.3;M=J8V*@8\ MYLUXQ]U%%@#T*<["SD%$:4FH-4B'FX];]U_#WD/=PXHGP5BKE7>%5N,J'>T2 MB9Q,BV,;YHE9\3U\@A\5FPI73$RN^OK@-?I++7(Q.IO[(:71(D-O`*:30=MD MV!>:5A==<%"U?U6G>CSRMW+;P4SW.W=VJQZR16QBYH&JQP8QA=B)N0-],7!@ MQI5,_8WA8F??$B*=]FA6]PD1Z)7WZLM ML""SESG(7RAA2#?M4.G0XA1J!6K MJ1A*K)!8CX\^_7;\>.K0L>Y<4'6VL-X'[27%BN0&NG_:C]P+^*U;"V.6R $+ MP/3^PI@GSCH=IZL8AX5V&W53@"-[]@.DG[319W@X^=)K\0MR2P0AC0*!DTV; MZ5N3=N+ZV#J9)90`FDT\#]S"M$!I%+7OE0#L+ED3-EHRH: DOJA8" 9;44)5->F/.4\/W/.Y"E! M<;-E^/V?`UD!V T.9.-]*.T*U 8X_8Y$E!E9*W.@A6@8B$R][KS%7:X`=E-: M"D^Y=JH/J!2DY[!V1!;L6 3/$5IQ-W^3'=2_@:*QZE()W$F`S/:HI[A@2+H")8PYT+YD5K>.G), M,@LB5PTH$.":)-I@(\ 9X2TT1\S3*K>.H=68%P1JQ;5Y(H&A2S0VABE1.=MNQL5*1:>WO78'A#Y_[@TQR2VJ)NNG7+V6+E<8D_- M0D75B+U7LB_K^.;9S^01A]GN(U"BYF!#V7;I:Z:PK6)&UOL^SO]A57E4Y[8S M[2A;00,4A&73G:\VG+3Y=XH?HHA_#2(K4:5-B Z1UC&ZB;]!G!\6\YI[-NOT MI?93B\XXZ#,FHDYK'Q1:2O5VB+;6>??%F\PXO!_+B%"$,'_BFM1-6">.L$PA&,S8SV MUKY6N,4;=([ M1_[:$3BGJ:DN?Q0!2\,-]8-A@80-O H5&7U;3@C[$.N/MJ\(YOZ7"PJQU^/6 M.A*F"K$_AL7YYAJ:=Z4Z,,@K?9HK&Z M/9.S>6QF^. OPZR9[:E((2;-7C"HE IMC]VZY*G!(7?J$!Q>B>4D`DFC^08Y M39\`89\_^U)3*SXDI+>P$WX*JYNQ#7V<2> ME4B9(&?5ZBB>0_R2#I,7%$+KF2QS`Z]1B.0B(&E9J_N&_0]HQ_(^^.Y8.!R? MBK)/FH_+PSH\\EW8"Y2U^G]TL91>,A,3AT^>:9F!/?0=V?X',*@*!ESJZ6W- MG*GCJQ#[OET(UTNBLTO''9L.L1.)>=/W8B-CY'8%&A!''RA3=2[\L.WRO,[; M]V_5K6MVHA;8- \[!.QW6]&3T@(4-R-7!*'M@5B>67L\L#R MJ4I96[%:3QB[V!/G;;>%W7]+MLF-R<\TC%8CDC >7R?K`M[X)0`DX,+'>[!9 M2'9I(V+Z%1;RU\Y;^>KDXS_A>GF?2U:ZJ23&-T#?]9O MQ+>GQW@G'*)?ZC, MX9F?^'N*IB*\!,C08Y.ZC5#(WV<+!H5A!>M&S&$J!(\!K-@8.M)7; 1:5CY) MNAXSZTNP?8,E/_E586D!?C-N%7ELV-31?G" MDZF_$..7IDHNP*I"RJJ4,+7^:BW0-# @:@U(VQ\]/ T6MNJ2I9+>6WP4K?[W MQJ]2R*'59CM"Y2,1$V*4[C7!?R'#>,+E_KM)/NU"Z?W DX07&W%`&[R#< C3^K?-EB\2%W)N<73WU7X MB7.Z%WS2-&A"Y6(2,?;+K0::;X=_26S_ MZP[[5(GE0JUA+=VZU+3LU:K%L>^?FC*?MR)NZ SU'W<.N-'"&K2 9.&ZK,S_ MR8G"G"V%Q *4-H@DZP8N"[V,]H/>Z1^!6G\KB/A"-U;7O." X6II&]&Q4)2Z M_U+>%E11]#BV"B4P"A#M70;&J&/@68FS$]#$MF$:MGTJ/7S"Y#MCZLN^: DW M48J'TB!EA=12\,2+8_4/,FAQ.3447QE1XL?DZL0UD#*QJ\C*OR\$1SS>=7<2 M,..$Y3ZI.;C=32%KV=FQ#U/$A(75:]@=Y>S;9$$L8A$BEC&(9@BNT@6G&ZT' MAI0VLL.;&!Q7J`-EM*)7EU;Z1S$;D M/&-DJ3D9-M-SZ?-=,S9>+U4#9?Z]TRC?TP+-JK%GEJ',:\N>!5_R7 3I;E;J MYJC)/6<>6KA#F^[H-\/8I%EB]<;0RO&ISLU]W&)69(4ZM]G6`^S\JN8/4PL27D [ MIS#3Y03E1Y[;%*,/ASZ.'<:JD2DP5Y7A(L'+?#BH+&E6J@0UYH=QT! 5>9MP MW(#5VK$W/Y0:C,4Y&&5B]A\Y6<#]B#OBTOH=[HED$P0&0@//.R8_:^L7/ T3 M!&3+FV[PH& @QY+:,??@?5E<3D?W60A884^$J$H!B54]$S46P:<#J,$D%?:ZMTO9IT=_J>X IA"EW^P?!" M-7MPJ$HR/CB1N01-F$]V3O1#SY9:[.:YUW%=P[% MIQ?Z/MMK)(YTM).*0<;-Q3SL-:E3J6J_I=_N[\R\\V8U=+?('[W&DH*BS'(#,0FY3#7M2_KLZJ;A_*;"'"/+&&D]LQ6A->UNJ%' M3!V"P0/B^WV$\R' O4%"V&Y6CCK_)W+:NK=@<)2!&5HE5OV>W:NJ"@`B:_<%TE0(TS>=TZ".:+&;Z+T&9R) M7 Q#;H>Z%$>Z$&&])30:O9S^5W2!#@SB7R#RS7H[R-O'J[A#0S R<'C6C3K5 MAY->DV:7?)YK!FMZ%QQ-J;4?H"8<<)L%$$W\+-I47/MB8)2?&P9AV.*?NB"S MUPRM3J\BZ#7.Z(@UXWREBFQJL28[:(!;F+>"N[5E[!%7S;#)XCV$;BSUOPJC M(&\OX'0L,_EN0!7OV[SH^[-\`8[.MX.L.P![=CQ90NSIRYYQO$@C$0,[D'^M M-=R,-A_36FS@24V'N&1 KXV)%OHFEWQ,BG6*U^+S]\:DTV%N88#O%KF!<4^/ M$M?@4R+MYYGIY91OF4!%WTJ9N+_"1)PZ#J@%BP<=M! MA//0N[VO>P9SY(GI0T\!/V.G>"W8Y M(,+EFK"?T.1OT4NN7[2W&O/8S@E_M9'\L[F]PX*I]*X0+J5!JAK^'M/NHW( (5A1!/SAR,]P7 MFCL>#P,$47]A.;S.:7+HZ$F-2@R'6650238H]!W'B$JUNIL2M\44-T2'W."" MSSD__%_,GO+=@I9\#DCZVE.F0M1L[C3368R-"\'Y;B@3[1*\YT8[&LE###[[ MTN,HGHIH7O!;VY,\ER0_F.DL#W=4X'[<^H5<"8/]JT''=64`Y[S M8_?++)YY0C21&_TTQ0) 9RQ47,7XF>M/@)W!D]Y@(7BP MIQ 4#!ED9/ [1!W[<_9V%J7V*2AC$X+UX_1(-T,)=V7&M&$D%)"0^#HW M<@E?S^KWYM*DRP3CBKU-0GD[`JWNL?4!H5]DA=\G]MG7&!.,/ AJ\?8[[NJM M\Z ZT5,PP[9[&\4PMUSK&;PDW8I&B1#S<%,`KLK']9Y'"/6+3*YU.:R]P^A: M* F4U^&^(%0<6.*)Y*Q$C")%=,M*L!3:K3 -NAW!F(Z4[.]S@[]&SC_KE_'] M%!V*"9F!V_78T#U+*CWZR(:B.^TME4)W%RJ($,5QOX+$[WO.;!AT\_G0H]7=\+&9*G<&3^F[?B[Y--IGN& MPB7:E]!C)284`2A88*\N5_U;3%"ATV+X=P,8PKE_\$'0M5,?6S *[+7(CL"( M'0S/5]C9GV!6LIA/Q?.]HM!D\A"H^RBEY)=9-X@13RQ8P"; WG+Z1,N\OE%+ MG8"AL5GO"TCP(: Y'$XXC6<7`<1[8CKAXL+OUGDCPE-5BN(`R4]9\HVV@-R8 M=]_X=D\B?>$'SQZ$2E*4YOMPY!VB=%W&)'LF(\C&)9X/2DS4D-?YX64PNU*W MD$0FVG_K[VRDTZ&5#N758W.CW),K\Q)<5[DK?\-[!!8M(]03O@:#5H]R\1MI MFNV"/6>E7,X2$&I(WEXU!JABE^;W2*?H65Z+C'T&?Q4V],I4V-U%#\R#:(*@ M(',(.QM:WUB(K295`>]C773I1KD*4L22;0*]'3WKTT%A4=5,;^U2MR^Q,YD# MYHHGR(0:;F!I-*#N]$P%"9]:/AQAPLV?9<]>[B8*9@%5'1$:OHYI*80^MVU4 M[0\KCP8K_FVS+(?)/K],%<+GP(V+];Y$V)2&@SF&B?ZC>5"!I#R=*[&PFAW0PCRP('>R?V0$^Z+O$U9!?V8N3QOC`SM" M]Q;[>N;>8BM]G3#3-*QCM3SD=3D/)]"&);G4_MCDG%3.N6\TMR)ZYGU;*7.# MMCTDMF?#-*&0`2EGEYCIB.2I 8*^ZS]J0 M[G>"318`5D 56MB7,]UICL!"?3 9^5G*[ZHD8K;+M_Y>#W)_YHP#`D!!A2T) MG7-J?G,+36)?EEH]M)Q@@>2:CMJF_"["V4W9"#H2\MJT4H=D# M8S7,<#LTDF/B;KQQ/&+IM10P-N(,7]W9:C&\PK[IB8(JOYS4]#"[XLD";7PP MRKBH&&\L.MWL2M74[ XLLV]]5Q(ID$5KMQ$XJ(%6CFFE)&2'LXF0I(QDV(A2 M>$^K0X0B]9PV`'=EO4&["1ZN%%4*WTL,C&Y+;_.&+.WQB!J$#_NDZ M!P3KKG4$0BB074G8C_Y5W$M2LN2@0;V'`)+&!U#$=J1!0"2_I()!#+7MF&]P=$Q U%M)XVB+`8>9 M+K4=5P])F336W+&YH5]QA+ZQ#:-4[1MG6O8]P[#9\ =J.D&PV2G:E+,N'D;CJ=*.V\>KY:K.BS'D.+3L\28H/F$NN8PR&>;+X3'SE.: M),P$:DS0/?'X2=Y3^0EMDQ+-+X7=QLBIU]'(8=;H_4SGOZ>2@+JWVW3M*G!0SJ`A?/$8 GJ6O"*[&0ZQ!5)$];D:!AZ=O=DR6?**I4>,J&S5YO4@)C."5X*$%/%CV M.@8V.F>5($(>F00Q-* ET:N'P2J?UHZZCD?Q\_R0U44#G"K(N?;A8 XO1DV> M1NTWC:+?(C0S=%F-OC[U(3!F>^A$3]YBG<>1(VN,*C9B?G717GW06V<@E5YT M2-O;#+8M8W;M=E,J;WH02F\_'&* 9&Y_KT_T4'*JM+37'A#_*@RH( MR7Q^&MCBN7OC(+ 8VH*8MZ"^('/W^H-K6^EAZ]0/Y7R/\:Z&LUDN"[).W=", M&OL7.A.$':Y#`1)%&/F;7+2J.$!Q"KEYQV6B$E^2$&]N(L8Q%V;2D$(F,=/9 M2["EF/=/':U+`,0>F"0'7>JK2Y0I'=!3V/K&3BM?2C\3.9V7R#7Q5W MON,*HRSD#?&[O46G+XZ.TM\FRCQ?Q[Q:%T9E?8JA7S=V+1QM&+)O_%_;2[R8 M!\\_[!O/E/42(?&_YL JPSTBL8*O:QL&H!GK"GBV/[](9"Z)KRD,D68BEW?# M/9M07IN@-NR"VDR]FOC/6")I&?KV%M4]W]UFSZMSY7"_T?AM>Q@Y7 M3OCF)A ,7TP8F$8B>G(Q\&;)\GH>.UQ#D_81PM5E`4/[5%8)K[/68O?G9'V7 M?N?17MJ1O;3MCI8U@"P7@0>W\N"^KRX/<[#4U]6[K31 MB-UJ3[=,D$LW-3DUHID([H LI/\&":.-"](4*W+"=;CZ+["XC1-*1 /5*=C" M[>[6SQX><53BBNAT]B"[M'*2V/;W"1HJ[J?0?%["_,!IX:)6(T?3>OLJBEQS ME;]%4&'<"JWYU[9Q/3$0:,U%BOTU8O!#B'>,S+'C+&!1$(9I%%6.87]J4QKI M,LZ>0Q@#L!V[#?%AVIN(]>S[OY+FRH.CK(I:H##"_4BWL0<:!7B_N26)X8YJ M@\+_R^;S?2(Y%JGN-!:3M8>^R^V'\'UQ>)=>K>,M`@`[N49?.1 _#9_WB$(I M+Q9@H2$Z]E$+$R$5)IB:<62K%?D_4B4C>UCWR0_:`!C,6Q,)%UJKWX4STC(/ MQZ/C9?S./E7YT)7VC&6D+(>.KF_WZA71DM/ZIW8QS$*VQ6BZM\+KD#=-?ZLV M+OWW(0U:*(=F9]PND1P0[(<]?Z8;6[,51R4IF8!NQ970>T/UEO5U6^EG'13% MR6*<=8X3TX>"F-!BU@\,E=R!`P/'ZVQFW8,8G_P44U7@_^R6`S.GW,\@R\ K M%!R-(E*YP;IJ1H//,7L=5,V2Q%)4A_NC/(]E+*))8-&7=DH3Q%LTHV"YQ%#7FW8-W*XNR&DGI-<]$D58P3DWV M4BLO!B,<.R!,-YGZBU@)R8*R#%;V_/#PB IGU>W+I.#?"=MX(_Y[5V&A(486X4HC8!8I^ M?=H61M)X"EZ4)IZ%V'M4!X0Y'_N2SA3>WXQ!W3R_^PT"W?5]G,2> -^W]5H[C<%Y9/I;':WWD[P62H'HCQA#J_[N$(T]F8&QU4_ MI0?H-JKI>C4^JU2UP6"E%/8T*BFW3_BH<8A6^BFIC^[*SB/2/)D" M:_M8N!+H,[QCK=JP2<)19AT7N$AR'QEV4TO1%Y^%`KO6D;D78P#-BN.;$=&G MVI7D6VMN\ L[,&)M<]?E=*R$[,H'4S.4^N;DIZB>>5TE+E%3R"DK$20,?3CM M%'I::1RIL;1&*SI9K^D@GW/]&^H6+,#YG]1'*/^/#8_'E=KC)]\S^WG0A6[7 M'0WP(,D08L"RU8UOU.?@QI/;9R]9"[B$+C\-&\2@\NOIH27T=RJ[%Y2*1[)L MF1U0*\;+4=4W;*LG1SV8-G%:F+XJA)J9N9"^@UM!WUP:KV@GC0!Q6XT=A..$ M(,$-=HCNW"HMU-6WDE'M2MUG$^4]LH6D1,YQ61.IW-)3'( *6*A] MSZ):.>K8L /H+:;@4X252A87X3V[)?GJZ3@Q`2. /HS:/5/-X)AGV_^I*-@= M!XKFK]5#=;53N Y"F;-70E9/[P$:HHP?S,*'+-K*R0CQA!2Z%=:.WN4C]A>4 MGQ=KQCX:8YUH5JP85$9I1IQ>-9I$5D.2,W=6KH;II%#K)@HUN^F\D-F"SC,C MFLS$L8O26QAA$HHQP/QCLP39`.+Y6ZQUAPC.GZ/,7\6BT:6^.TB$BQUYOA.> MZ"07P'E*0D0-4Z8I#;FZTSZ'#?0Q-E.F# RS&Z)7A3JQ#^\TF_9\N+57"3BB MK$OW8 ED6*,;BF+%.7D4Q/%_>J,1L3M74HX].\G4S4U"8@-=R&U@)G"$D9SZ M-Q+BUV))_NSJ?E)"AS"/@KF6SF4_<,$=4P^ T_E/IP7#B3!!=0]JR%;SCE M%*,,OV_*Z4@EQ;1F?:(^E4+9I^+;R\H!0?+BO)O2]=6!LQ\EMF7 YU8KK3TJ M&!D!J+[+K%S_8BGR]6 MA%T/@:=]2EO:=8XI#_W2CI,[L[K0`G:A_UK#7VAR(S!AJF>1)E6UNXKY<*[9 MW2,*&WKA$K!38PV I@=T^3U1XG)"^<5#X-O!A]HU;U'+KL' );:=,B[ 4+AR M+WZUEY\]^85[V3 -=O]54XLD*)X/\NFQ/W)7B_M0BK(`'43?F=..)&U')6H\ M%8E'16IVXPF)XC$A]L]4I(EF]9^6.]]75O(!NO;A/@*SX5LX:E(D\)7W[G0] M;Z(OQ\\*@VBS`)-),@,J)[&LN5_^[7Z:'6Y[C"5^G58J@OC.S_Q\TM&?[ 3' MV1JN$+35A&2EX1BHJX)+5^7B76W8^[,\4+F=_W_&A),^F$4Q%MTF.D+VM3Y2 M"5%BGNTIM?C>.4)+$U"68H)\?OKDT)_]0$C5@W*E$"@Z7(LIKO30>$[!;,[C MS285G&I*R#NK$B'UL[$('2(;FZTY9/BZ2:/"#8]3W1A;JOI\>VGNB.*S0CK& MONHCP-T,=?W(OZ':P;^?W3[/.H1R0*E$G/F/!HB=FP@89$$/88_1YGSF+?<^ M, A:3_[1N,S IAY,;._C5\(8>A&K0)$83XH(M@L7GF/-@R7]/V9I?^1HNOL* MTHYC0%M)N8Z3TH4^..N9J0MIT,37U]>/MF$.*J)I0-([0/]HF1@2`G%T1,*1 ME9/P)1MK)A(CN]I.VQ)[T<;"?,24K([F(ANU.$;4@Z\!%LFWFP0(PKPO#>0. M;^6)V'VKZ@S5V<=]B7#Q;!R?\&!.<1N"QTG_Z"3V^/M.VB3%GQL\/NNCTV5H M5D,O`KY3/J- Y;O?C)+7[/DFM6PS>KU.*$-:9Q5BQSP0E6BCVT.O6L&.U/FL M5-9)EBL5>BM*7#QX6VMP0?YJL4=G"'8G/C3&DD-A/-EO$28,+2PP`7(_A\_C MC;$^TCY/EPCS?_QUFYCZYR6\4J,F343;.+E@?:_8MUOM<1'QUO04Z($7'8(C MX2Q@V:M G=8'L0"/NN2P=5&NI%5DB:XX3V,,[">I"U*4[C8H1A6WF;0+8^6R MD0-6!':/N_/WV7-!KV%I&RJ<535M5^,\.W2X_0^HEAOL ;*-;IS3:XO8\:T3%*I9@) MVMX_9(9+(VNN*5'B%K_7K%8/ZF#@I&#RCI[$%-R9-C2@*+/!9,@R?=;M2]B' MG3*>4P.W9N4*E^)?0O:G7TXF$ FM687[Z/;&]/ENP0!>[1X<)]PT6^1;'+7E M%M#Z^1;#M^S(\*-O'GC G3&?,Z'Q6S*I6NG37S,00*K6BU+41?9"+]Z.X(!4 M&!4TVM"D1 F%B/%!JS&?&X\J:0B.B-XVJDJQ M,D$16(YA[<7T+ M7?+G>F 5OCIY(A#L:62C(PW5/2C:UV!WP--!I[FM*20`N.OY!]E'\^=[LLLH M@K.0A1.[$WQ6_A^EK?YOBU=F`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`TT6"-N-+(*XIWQRBVE]N*JP?-L+8;V)5DW=$CXN`T&,E( *)5C/ MC8"E$#5#2".83_"XN^Q [ES\:R?L=I%"ZP`/NAQUO>B0--Z;#E(R8/NQ+7_D M+WF5/] EBZ2CU[NK-E3.-$IGSZ?ULA_!-4IQ6HJLM[80K[XF_8^!^29!'T.N MX^)_2/C4J8[QK!/"0=9J+<;1KU3QD]D'&5B#S6LJWTNPT,==W]:<#GG'?W:8 M9%@"9&$@)!#T"9'LE^I/? M#65N9'-T"[(C_OX2KP:.[FE8+T9Z%2T`1PI?R?9[ M\D,?'4)$C4KI2M$OV+!8FTJ2?X"M4=7\9/IXU=I(D@O$B3FGYF+F MLC7V7A['QMBW<6XKV)[!8/.:H/R2E)_&%K'&UGHT:?P`NS#6^L'*_M_;B$O0 MS4@"(KE[/'NQ_V;W@^:;6#X+S=[>=:V;Q'$2].8O(ZPZ]PW)6QLCSWLN%<$Q MFUQ/4(<>>.7$Q0.23L2-M4TRT<8-TZ\1WYMWZB(,QGEGA:Y.M/P6H:3VID\* M+@Q433;@E(,>L6R@;?H@0>EO#_GX_$;A.QDQR'_8#1SMPZ0`F.L2]#^WG[Y' M*S^\]_\DHY@VTJ8+W-;XG!>\IJKR&ZNO\QJ';S1@A.QA,BR;?UW8M#+/++UI MT1:=*BCO8C[\+:4+==>-YDNH`10V?S$)29 7P@(?@/OWCQNWV6/SY<0#IGD[ MCQR*P VJ^(#!&]&)W'_58(1)<;<.)#8"AL8IY9%FAVM]PO[/:<%K&(R(ML?P M*WJ^X5]]4]AE]G\0/P^BDBI:W_3$N6[&+S_"RD1 M5JU_H E'0Q!\^DU.4MVT+B; C-KY[C]C8R6:LJ@G\X3<-?BPN4S&=3LR6X!. M9&\#F%BG/8CQJO@DI[8Y)&L".<[N\AX3`G$LLY94#3'P3G=,O/PCZZ82\O[/ MD\EYN0*\MVK!4-'7^2(E\9]MOJ"PB6@CW\^]AL>PHV?[-.'Q_\"R%[%V>3B; M\\93"./&+_JDN418&T($!] NU7E!4J%\705;,!F8D;[`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`A>ZBM(;3C-% MQKC4V&':4G(*38PP-?]K1M*\6-0`YDDMF1Y&.5* Z;I&]JVAS_ M;S]Y;[X*"X2IA\A507<,&G"6(\V0&O54T]E.D^,0>\+!&V'E+0(,H3:W^_Q< M/ZIDALR;OP8'-$M@>Y$TTC+"*DJC)B?T\0/SSQEY'\7>#65N9'-T7!E("]086=E Organization: Bell Sympatico Message-ID: References: <3C580FC8.8DAA7C55@sepro-robotique.com> <3C5901E5.B7FC9082@sepro-robotique.com> <3C5A4706.4C766158@sepro-robotique.com> Sender: bpringlemeir@DeadDuck >>>l> Bill Pringlemeir wrote: >> Apparently you like to ignore free advice. Perhaps you can give >> me the address of your company so that I can come and work for you >> for free? >>>>> "EH" == Emmanuel Herbreteau writes: EH> Hi Bill, Thanks for your "free advice". No, I don't need EH> your help to get an UART driver working. I was just asking EH> if a different driver can be find (I don't want to do EH> something that could already be a part of Tornado). EH> The latest news are : EH> - "ns16550sio.c" doesn't use FIFO and is the driver I have to change EH> - "st16552sio.c" use FIFO and it is the driver I need (june 1999 EH> revision) EH> This was the answer I was looking for. I guess you have grepped for this line? /* Reset/Enable the FIFOs */ ST16552_REG_WRITE(pChan, FCR, RxCLEAR | TxCLEAR | FIFO_ENABLE); I hope at some point you read about the Fifo control register (FCR). Bit 4, 5: FCR4 to FCR5 are reserved for future use. Bit 6, 7: FCR6 and FCR7 are used to set the trigger level forthe RCVR FIFO interrupt. 7 6 RCVR FIFO Trigger Level (Bytes) 0 0 01 0 1 04 1 0 08 1 1 14 The patch I sent in the original message set the bit 7, to use an eight byte FIFO. These are defined in `h/drv/sio/ns16552Sio.h' #define FCR_RXTRIG_L 0x40 /* rcvr fifo trigger lvl low */ #define FCR_RXTRIG_H 0x80 /* rcvr fifo trigger lvl high */ I don't think that the `st16552sio.c' is _EXACTLY_ what you want, but that is for you to decide... fwiw, Bill Pringlemeir. - -- When you said ``HEAVILY FORESTED'' it reminded me of an overdue CLEANING BILL.. Don't you SEE? O'Grogan SWALLOWED a VALUABLE COIN COLLECTION and HAD to murder the ONLY MAN who KNEW!! vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado2 and PVCS Date: Fri, 01 Feb 2002 00:27:04 +0000 From: molochai Message-ID: References: <3c587646$1@brateggebdc5.br-automation.co.at> <9ec53d11.0201310148.3c625ef5@posting.google.com> Ofer Goren wrote: > Hi. > Let me rephrase my question: > Fact 1: We use Tornado 2.0.2 > Fact 2: WE USE PVCS. > I use Clearcase (and previously RCS) but the same principles apply. > I was wonder what is the best way to work with this version control and > Tornado: > like, do you use archives, or flat trees? > If you use archives, do you put them also under PVCS? Archives as in libraries? Yes. Lots of them. Each has an associated Tornado II project file. The .a files are currently version controlled. If you go down this route, you definately need a top-level makefile to build everything which includes checkin and checkout rules. If you don't, things get very messy if there are any dependancies between the libraries. If you're planning to do development work on the libraries as you go along, you might want to consider arranging your makefiles so that you can take libraries from a stable release area (which is version controlled) or from a development area (which shouldn't be) - or from a mixture of both. You then use a release script to 'release' stable versions. We had nasty version control problems because we weren't doing this when using RCS. (under Clearcase it's all handled very nicely with dynamic views). > Do you put your project files under PVCS? Definately. Project files are highly important to version control if you're using TornadoII's makefile generation. > Do you put your makefiles, prjParams, prgConfig etc under PVCS? I currently do version control the prjParams and prjConfig files. However as these are objects derived from the project file, there is an argument that they should be generated as part of the build process [1] As for Tornado generated makefiles, these are always generated in my production builds from the project file. The other particular reason for for not Version Controlling derived objects such as these is hastle of having to check them out to perform any kind of build. > what do you do with you .out files? your .o files? your .exe files? you .o: Intermediate derived objects - absolutely no point in version controlling these. .out/.exe these are the files you release to customers. You definately version control these. > vxWorks files in case of bootable application? likewise. > Do you compile your archives with your own makefiles, or use the Tornado > GUI? I have a toplevel makefile to build all of the libraries and two of the .outs. It generates makefiles from .prj files, performs clean builds in both optimized and non-optimized forms (for debugging), checks-in and labels ::every:: source file. > All those question effect the way one work with PVCS, and the number of > files under the control of the source control utility. > FWTW We even version control the entire Tornado II release as well, and run a network-installed T2 seats. Absolutely ::everything:: that goes into the build, or is used to make it can be traced. Further more, I can confidently get a clean PC with nothing but clearcase and a stock network TII installation on it and am guaranteed to be able to build ::any:: version. > Thanks, > [1] There is some TCL out there somewhere that can do this for project makefiles. Presumably something similar could be arranged for VxWorks auto-generated files --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado2 and PVCS Date: Fri, 01 Feb 2002 00:51:31 +0000 From: molochai Message-ID: References: <3c587646$1@brateggebdc5.br-automation.co.at> <3c591cdc$1@brateggebdc5.br-automation.co.at> Werner Schiendl wrote: > Hi, > > "Martin Raabe" wrote in message > news:a3asq2$j3r$07$1@news.t-online.com... >> Hello Werner, >> I agree with you, >> when you are talking about pure Versioning. >> But then I think MS Source Safe also isn't for free! > > Yes, of course. > > PVCS is not free, too (IIRC). > > The thing is not if it is free, but if one has to spend money for it. > If you have the tool already, you can of course use it. > >> >> For pure versioning RCS or cvs is a good approach without any payment >> in Version Control Software to a vendor. > > rcs is probably a bit out-dated (if you mean that old unix command line > thing I have in mind). > > cvs is probably a very good choice, given the huge number of open source > projects which successfully use it. If you're going down the route of open-source tools, and pure version-control is sufficient (it is in many cases) I'd go for CVS. Any tool that has exclusive locking (RCS, and by the sounds of it many others too) scales badly for teams larger than about 2 or 3 people needing access to the same files. >> I experienced that many projects need more than that, which is > COnfiguration management. >> MS Source Safe lacks in offering sub-version numbering, which means, the >> version number is >> only one number like 1 or 534 or 3456. >> RCS, PVCS, CVS or others provide a better numbering scheme to reflect >> what happens in your project. >> like: 1.2.1 or 1.4.2 or 3.6.5.9.3 > > Probably a philosophic question, VSS supports labels, which you will use > for releases, patches, etc. > If it's a CM tool you're after, there's only one serious contender: Clearcase. It's expensive, but has paid itself off in my team several times over already. > I'm not a SourceSafe advocate for reasons (poor performance in WAN w/ > limited bandwith, exclusive checkout only, less-than-optimal command line > support). > But it has good features as well (nicely integrated with MS dev tools, > nice project explorer, easy to teach and use) > One tool to definately avoid is Component Software RCS. This is truly a dog. It has a GUI which basically imposes a project structure over GNU RCS in a totally limiting way. They also do a similar product based on CVS. Christ knows why anyone would pay for that when there's a good (and probably much better) freeware GUI for CVS available. M --------------------------- Newsgroups: comp.os.vxworks Subject: Re: login issue Date: 1 Feb 2002 12:58:18 -0800 From: les_news@att.net (Les Hammer) Organization: http://groups.google.com/ Message-ID: <9bbf9668.0202011258.78502c43@posting.google.com> References: <26ca8a8f.0201300030.47a0fa90@posting.google.com> <3C58628B.50702@notifier-is.net> <26ca8a8f.0201302203.893596c@posting.google.com> <3C5ABA6F.E5D4982B@notifier-is.net> david lindauer wrote in message news:<3C5ABA6F.E5D4982B@notifier-is.net>... > hi, > > we found out that the encrypted password 'LOGIN_PASSWORD' does NOT match > the password given by vxencrypt. We cannot add new users/passwords, but > we can successfully delete and readd the default password, but only if we > use the LOGIN_PASSWORD version rather than the encrypted given by > vxencrypt. Trying to delete and readd the default using the vxencrypt > version of the password does not work. Have you upgraded versions in the past year? About a year ago, I encrypted our project password and got one string of characters. Last September, my coding partner upgraded her VxWorks version and encrypted the same password -- and got a different result! It looks like the password encryption routine has been changed sometime. Les --------------------------- Newsgroups: comp.os.vxworks Subject: RTOS Evaluation Comments Request Date: Fri, 01 Feb 2002 21:56:58 +0000 From: NoSpam@SpamOff.com (Steve Huggins) Organization: Renishaw plc Message-ID: We are considering which RTOS to use for an x86 platform, developing on a Windows host. A paper evaluation is underway. If you have any comments that you would like to make about: VxWorks the tools the technical support they would be gratefully received and treated in the strictest confidence. Please email (or post here if you are happy for your comments to be public). Steve Huggins Principal Software Engineer Renishaw plc Great Britain Email (remove spam control): steve dot huggins at renishaw dot com --------------------------- Newsgroups: comp.os.vxworks Subject: +++ C-Designer (sehr flexibel) für Realtime/C/Systeme sucht neues Projekt +++ Date: Fri, 1 Feb 2002 23:11:49 +0100 From: "M.Thoemel" Organization: T-Online Message-ID: Guten Tag, ich bin freier Sofwareentwickler mit einem sehr breiten Spektrum. News: Aktuell habe ich einen Kurs über UML und Rational Rose RT absolviert und werde diese SOftware jetzt zwei Monate für ein VxWorks Target einsetzen. Ab dem April 2002 stehe ich wieder zur Verfügung. Meine Schwerpunkte sind: - - C (aber auch andere Sprachen!) - - Realtimesysteme - - systemnahe Programmierung - - spezielle Hardware-Schnittstellen - - Netzwerke - - Treiber - - Streamsdriver - - Grafische Datenauswertung/Darstellung - - Software-Messungen - - Gruppenleitung - - Projektierung - - Projektleitung etc... Sie können sich im Detail informieren unter: http://matthias.thoemel.bei.t-online.de Ich würde mich über einen Kontakt mit Ihnen freuen. Ihr Matthias Thömel (matthias.thoemel@t-online.de) --------------------------- Newsgroups: comp.os.vxworks Subject: vxEieio Date: 1 Feb 2002 14:19:17 -0800 From: vloscomp@yahoo.com (Vinh) Organization: http://groups.google.com/ Message-ID: Hi, I found a usefull routine 'vxEieio' in 'LibPPC604gnuvx.a' but there are no prototypes for it in any of vxWorks header file. I think the routine is a macro with one assembly instruction 'eieio', but i can't verified this with the 'objdumpppc'. I would like to know if any one has try using this routine, and if so how can i ensured that is it just a macro 'eieio'. I can always defined my own, but i'm not in a habit of re-engineering:) Vinh Lam --------------------------- Newsgroups: comp.os.vxworks Subject: Re: RTOS Evaluation Comments Request Date: 01 Feb 2002 17:32:23 -0500 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: References: Sender: bpringlemeir@DeadDuck >>>>> "Steve" == Steve Huggins writes: Steve> We are considering which RTOS to use for an x86 platform, Steve> developing on a Windows host. A paper evaluation is underway. Steve> If you have any comments that you would like to make about: VxWorks is a throughly debugged OS. If you don't go of the beaten path, then it is highly reliable. DosFs seems to be a sore point. I have been interviewing candidates for jobs and this is what they noted as being most problematic. I would also say that WindML and Zinc are not as reliable; just because they don't have as much use. I don't know about WinPoet, etc. The TCP/IP stack seems to function well for light loads. People have complained that it doesn't scale well [router traffic]. I haven't experienced this myself. The most loaded system I had was an 802.11 access point, but we had a custom routing layer. The GNU tools are usually pretty reliable. Using the GUI debugger often causes exceptions, especially if you are watching variables. It seems like this happens when NT is running low on memory. Our system is loads about 50 different modules, so it maybe specific to us. I don't think it is as I remember this from another vxWorks project I worked on... Most things works fine. WindView is a very nice tool. However, there are a few things that don't work and I have just `learned' to ignore them. The technical support has been improving. FAE support is highly variable. If you have a good FAE, then you are probably in heaven. I have had some that were not responsive at all. I don't know if this was because my company was not paying support bills or what. However, they didn't mention this. My current FAE is very helpful. Since you are using the x86 as a target, you have many options open. One of the main wins for vxWorks/Tornado is the number of platforms supported. If you always works within x86 confines, you may have better solutions available... but it would depend on the overall system requirements, where the code could be ported, etc. hth, Bill Pringlemeir. - -- I can't believe how unpleasant you are. I firmly believe that most disgusting people have operations . vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" --------------------------- Newsgroups: comp.os.vxworks,comp.protocols.snmp Subject: Re: can't compile a simple mib Date: Fri, 01 Feb 2002 23:44:54 +0100 From: ilyas G Organization: Wanadoo, l'internet avec France Telecom Message-ID: <3C5B1A66.7020302@no-spam.guennoun.org> References: <1012578939.423733@salvator.ibr.cs.tu-bs.de> should i compile the header file and the skeletons (.h and .c) in a specific order ? is there any other file i should compile with them ? Juergen Schoenwaelder wrote: > In comp.protocols.snmp Ilyas G wrote: > > >>i have the fellowing ASN1 declaration >> > > I am not sure why this specific compiler fails, but I strongly suggest > to use SMIv2 rather than SMIv1. As a quick start, here is your module > in SMIv2 format: > > MYPERSONAL-MIB DEFINITIONS ::= BEGIN > > IMPORTS OBJECT-TYPE, MODULE-IDENTITY, experimental > FROM SNMPv2-SMI; > > myDir MODULE-IDENTITY > LAST-UPDATED "200202010000Z" > ORGANIZATION > "..." > CONTACT-INFO > "Ilyas G " > DESCRIPTION > "..." > ::= { experimental 17 } > > myObject OBJECT-TYPE > SYNTAX OCTET STRING (SIZE (0..255)) > MAX-ACCESS read-write > STATUS current > DESCRIPTION > "Un object string cree par moi meme a des fin de test." > ::= { myDir 1 } > > END > > You can easily turn that into an SMIv1 module by using smidump or > other MIB compilers - if you really need to do that. Note that SMIv2 > is out since 1994 and a full Standard since 1999. Note also that > using { experimental 17 } is only allowed if you never use it > anywhere outside of your lab. > > /js > > --------------------------- Newsgroups: comp.os.vxworks Subject: What's usage of ccppc? Date: Fri, 1 Feb 2002 18:05:08 -0700 From: "Liang Yang" Organization: Arizona State University Message-ID: Hi, Which company developed ccppc used in Tornado? Where can I get detailed manul about the usage of ccppc? Liang --------------------------- Newsgroups: comp.os.vxworks Subject: Re: What's usage of ccppc? Date: 01 Feb 2002 20:34:18 -0500 From: "Paul D. Smith" Organization: Nortel Networks, Inc. / Billerica, MA Message-ID: References: Reply-To: pausmith@nortelnetworks.com %% "Liang Yang" writes: ly> Which company developed ccppc used in Tornado? Where can I get ly> detailed manul about the usage of ccppc? ccppc is the C/C++ compiler from GCC (the GNU Compiler Collection). Probably you have an older version, known as EGCS. I don't know if WRS ships any of the newer versions yet. You can find documentation on the newer versions at http://gcc.gnu.org Most of the command line options you find there will also work on your version. If you want documentation for your exact version, WRS should be able to give it to you. - -- - ------------------------------------------------------------------------------- Paul D. Smith HASMAT--HA Software Mthds & Tools "Please remain calm...I may be mad, but I am a professional." --Mad Scientist - ------------------------------------------------------------------------------- These are my opinions---Nortel Networks takes no responsibility for them. --------------------------- Newsgroups: comp.os.vxworks Subject: Tornado-2.1/mips and cp7000 BSP (Ocelot MIPS board) Date: Sat, 02 Feb 2002 02:09:24 GMT From: Andray Kaganovsky Organization: Primus Canada Message-ID: Greetings, I am using Tornado-2.1/mips and cp7000 BSP (Ocelot board by Momentum Computers). The original version of this BSP that came with the board was for Tornado-2.0x; I got T-2.1 compatible version of this BSP directly from WindRiver. Here is my problems: -> printf "%d %d %d %d %d %d %d %d\n",1,2,3,4,5,6,7,8 1 2 3 4 5 6 7 8 -> sp printf,"%d %d %d %d %d %d %d %d\n",1,2,3,4,5,6,7,8 1 2 3 0 0 536912869 -33554961 60401228 If I execute this command on any of my PowerPC or Pentium targets, the output is of cause "1 2 3 4 5 6 7 8". It looks like WindSh's 'sp' is broken on this Ocelot MIPS board - it is impossible to pass more then 3 arguements to spawned task. My question is, does anybody encountered this problem with other MIPS boards and Tornado-2.1/mips. I am trying to figure out if it's BSP problem, or there is a bug in core OS. Any help is greatly appreciated. Thank in advance, Andray --------------------------- Newsgroups: comp.os.vxworks Subject: Freebie VxWorks Tornado? Date: 1 Feb 2002 19:27:33 -0800 From: embedded771@yahoo.com (Embed77) Organization: http://groups.google.com/ Message-ID: <57d4545a.0202011927.77254ba7@posting.google.com> I am interesting in moving from BSD development towards VxWorks. Is there a non commercial development enviroment available for download? Is there a x86 version of VxWorks OS that is downloadable? If not , what is the cheapest way to get into VxWorks development. Thanks Sam. --------------------------- Newsgroups: comp.os.vxworks Subject: This is Cool 90548 Date: Saturday, 02 Feb 2002 01:42:54 -0600 From: paul_golik@hotmail.com Organization: Percussionist Paul Message-ID: <02020201.4254@hotmail.com> Reply-To: paul_golik@hotmail.com Hey. What's Up? Not much here, I Just found a really cool site on Mp3.com . This guy made some really cool Dance/Breakdance/Hip-Hop Beats! The URL is http://mp3.com/precussionist_paul I bought his CD (since it's so cheap) and all my friends like it to. You should go to his site and listen to his samples, then buy his CD and then tell all your friends! So go Here: http://mp3.com/precussionist_paul --------------------------- Newsgroups: comp.os.vxworks Subject: Returning memory Date: 1 Feb 2002 20:09:11 -0800 From: vividh@hotmail.com (rasayani) Organization: http://groups.google.com/ Message-ID: <28809481.0202012009.10cf7545@posting.google.com> If a task allocates memory using malloc/calloc and is deleted, is the memory released to the free memory pool? - -vs --------------------------- Newsgroups: comp.os.vxworks Subject: MPC8260, vector table @ 0xfff00000 Date: 2 Feb 2002 01:06:52 -0800 From: pgrangier@hotmail.com (pgrangier) Organization: http://groups.google.com/ Message-ID: The vector table for an MPC8260 can be mapped, either to 0x0000_0000 or to 0xfff0_0000. I need to set it to 0xfff0_0000.I've seen with a disassembly view, that the intVecBaseSet method seems to be designed only for the 0x0000_0000 case. On newsgroup, I've read something about the excVecBaseSet. I've tried it without success. So my questions are: Does vxWorks really support the vector table at 0xfff0_0000? If it works with excVecBaseSet, what are the side effects? Thanks in advance. Pascal --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Returning memory Date: Sat, 2 Feb 2002 10:53:08 +0000 (UTC) From: David Laight Organization: BT Openworld Message-ID: <3C5BC595.9C91A04C@l8s.co.uk> References: <28809481.0202012009.10cf7545@posting.google.com> rasayani wrote: > > If a task allocates memory using malloc/calloc and is deleted, is the > memory released to the free memory pool? No: the task doesn't 'own' any resources - so they can't be automatically freed (more particularly the system can't tell the differenct between your code calling malloc and system functions calling malloc (to add data to internal lists) while called from your task). Maybe AE does something different.... Davud --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Sat Feb 2 23:11:21 2002 From: redfortune Date: Sat Feb 2 23:11:23 PST 2002 Subject: shared memory of vxworks dear friends, Now I'm implementing a shared memory backplane using vxworks smEnd.c and four ppc8260 connected with powspan and PCI, but I do not know how the sm backplane abstraction routines in smUtil.h call the 8260 BSP routines. and what BSP routines need, sysmailboxint,sysmailboxenable, sysmailboxdisable, systas, systasclear, and what? anothor question, how one CPU gets other CPUs SM_CPU_DESC? thanks! Caifu Zhu redfortune@sina.com ______________________________________ =================================================================== ĐÂŔËĂâˇŃľç×ÓÓĘĎä (http://mail.sina.com.cn) ĐÂŔˡÖŔŕĐĹϢŁşÇáËÉśŠÔÄŁŹÁżÉ휨ÖĆ,şĂĐĹϢŔ´ŐŇÄă! (http://classad.sina.com.cn/) ĐÂŔËÉÁ˸śĚĐĹÉÁÁÁľÇłĄ ´ŤÇéÎŢĎŢ (http://sms.sina.com.cn/cgi-bin/sms/smspic.cgi?spec=11&type=0) From vxwexplo-errs@csg.lbl.gov Sun Feb 3 04:03:07 2002 From: Vxworks Exploder Date: Sun Feb 3 04:03:09 PST 2002 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sun Feb 3 04:03:04 PST 2002 Subject: Re: can't compile a simple mib Subject: How to change shell's default directory? ------------------------------------------------------- Newsgroups: comp.os.vxworks,comp.protocols.snmp Subject: Re: can't compile a simple mib Date: 2 Feb 2002 13:05:03 GMT From: Juergen Schoenwaelder Message-ID: <1012655102.816965@salvator.ibr.cs.tu-bs.de> References: <1012578939.423733@salvator.ibr.cs.tu-bs.de> <3C5B1A66.7020302@no-spam.guennoun.org> In comp.protocols.snmp ilyas G wrote: > should i compile the header file and the skeletons (.h and .c) in a > specific order ? > is there any other file i should compile with them ? I have no clue. Check the documentation of the MIB compiler you are using. /js - -- Juergen Schoenwaelder University of Osnabrueck Dept. of Mathematics and Computer Science Phone: +49 541 969 2483 Albrechtstr. 28, 49069 Osnabrueck, Germany Fax: +49 541 969 2770 --------------------------- Newsgroups: comp.os.vxworks Subject: How to change shell's default directory? Date: 2 Feb 2002 06:31:25 -0800 From: igtorque@eliop.es (=?ISO-8859-1?Q?Ignacio_Gonz=E1lez?=) Organization: http://groups.google.com/ Message-ID: <9ec53d11.0202020631.21afbef6@posting.google.com> Tornado 2.0.2 for Windows NT. When I run a local shell from the Tornado IDE, I usually perform some setting activities (like ?shConfig ... etc.) I've learned to perform these things writing text scripts and typing Date: Mon Feb 4 04:03:14 PST 2002 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Mon Feb 4 04:03:08 PST 2002 Subject: Re: Returning memory Subject: Re: Still "dosFsDevInit failed". Subject: Re: Returning memory Subject: Real Time OS for VME Subject: Re: Tornado2 and PVCS Subject: Re: i have downloaded the tornado 2.0 protoyper, but the key as not been emailed to me Subject: Returning data from S-func in RTW Subject: Thank You!!! Subject: Thank you!!! Subject: Re: transition from bootloader to vxworks Subject: Intermediate Driver ? Subject: Re: Returning memory Subject: Re: Tornado2 and PVCS Subject: Re: vxWorks and vxWorks.st Subject: Re: Returning data from S-func in RTW Subject: Re: How to change shell's default directory? Subject: Re: Default scheduler in VxSim ? Subject: Re: Tornado2 and PVCS Subject: Re: Tornado-2.1/mips and cp7000 BSP (Ocelot MIPS board) Subject: Re: frequent rebooting for an ethernet interface looses connectivity?? ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Returning memory Date: 3 Feb 2002 05:30:16 GMT From: Denis Perelyubskiy Message-ID: References: <28809481.0202012009.10cf7545@posting.google.com> * dale [02-Sat-02 20:04 -0800]: > > In article REPLY@finch.lecs.cs.ucla.edu>, denis_newsREMOVE-TO-REPLY@hotmail.com > says... > > * rasayani [02-Fri-02 20:09 -0800]: > > > > > > If a task allocates memory using malloc/calloc and is deleted, is the > > > memory released to the free memory pool? > > > > as others have stated "no", but i think there is a > > taskDeleteHookAdd() or something of that nature. you could > > play with it, to make that function deallocate your memory > > if you can't do it before deleting the task. > > > > i am not sure if that fn is called every time the task goes > > away for whatever reason, or only on taskDelete() called by > > a user, but it is something to look at > > > > denis > > > > > Good suggestion, Denis. But, the task delete hook is > performed as an extension of the windkernel and is not > itself a task, whereas "free" requires a task context. > What the hook routine could do is send the free > request(s) to the excTask, via excJobAdd, which will > perform the frees within a task context (at priority 0) > after the hook is completed. ah, that's very interesting. i had no idea that was the case do you, by any chance know, why free() requires the task context ? i could somehow see this restriction on the malloc, since it probably isn't atomic, and the task may be put to sleep when asking for memory... would there be some similar reason with a free(), or am i waaay off on this one? > The Task Delete Hook is called for every task termination > and deletion. that's good to know, thanks. denis - -- if you want to reply by mail, please tweak my address in a reasonable way. sorry for the inconvenience. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Still "dosFsDevInit failed". Date: Sun, 3 Feb 2002 15:05:51 +0100 From: "Charly" Organization: Guest of ProXad - France Message-ID: <3c5d420f$0$23515$626a54ce@news.free.fr> References: <2002130-143411-880483@foorum.com> <200221-134637-250830@foorum.com> <3c5aa480$1@brateggebdc5.br-automation.co.at> Reply-To: "Charly" Dear Werner, I did that from both sides (target and shell). I will check your tests next week. I will check also if the problem is "side effect". in reducing the application. Thanks a lot Charly "Werner Schiendl" a écrit dans le message de news: 3c5aa480$1@brateggebdc5.br-automation.co.at... > Hi Charly, > > As I see from your shell commands' output, you are trying those in WindShell > (Tornado II host shell). > > This can cause confusion, because devs lists the devices on the target, > while ls lists the directory of your Windows Box. > Type @ls to execute the command on your target. > Try @pwd as well, if the above does not help. > > If you do not have those included in your image, you can try ioDefPathGet(), > alternatively. > Something like: > > bufx=malloc(1000) > ioDefPathGet(bufx) > printf(bufx) > > in the WindShell. > > You need to find out, what your target tries to access, when you get that > error message. > It's probably using some 'relative' path. > > Did you try to access a file on your FTP server with it's absolute path? > > like (in WindShell) > > fdx = open("host:/test.txt", 0) > read(fdx, bufx, 1000) > d bufx > > Does this work? (note the _forward_ slash) > > Report your findings. > > hth > Werner > > "Charly" wrote in message > news:200221-134637-250830@foorum.com... > > > > Thanks for your help, > > The problem is still there "dosFsDevInit failed". > > > > When I start a big application on my PCPentium target, the shell display > > "dosFsDevInit failed". Then, the target seems dead, no Shell and no target > > server connection. This application works fine on an other BSP > (Lp-elektronik). > > Do you have an idea where I could look? VxWorks configuration? > > All my files are on a FTP server. > > I use also a RAMDISK. > > > > Here are the result of some shell command. > > -> devs > > drv name > > 0 /null > > 1 /tyCo/0 > > 1 /tyCo/1 > > 2 /pcConsole/0 > > 2 /pcConsole/1 > > 6 host: > > 7 /vio > > value = 0 = 0x0 > > > > -> ls > > > > Volume in drive D has no label > > Volume Serial Number is 12E3-344A > > Directory of D:\kern\Awd_num\T2mak\rtl > > > > [.] [..] RTL.WPJ USRAPP~1.C PRJCOMPS.H > > PRJPAR~1.H LINKSYMS.C PRJCON~1.C MAKEFILE [DEFAULT] > > PRJOBJS.LST ADD_VXW.C USRAPP~2.C [DEBUG] > > 10 file(s) 98,644 bytes > > 4 dir(s) 1,170,317,312 bytes free > > value = 0 = 0x0 > > > > -> version > > VxWorks (for PC PENTIUM) version 5.4.2. > > Kernel: WIND version 2.5 > > Made on Feb 1 2002, 13:09:39. > > Boot line: > > ene(0,0)host:vxWorks e=192.0.1.2:ffffff00 h=192.0.1.1 u=target pw=vxworks > f=0x8 > > tn=target > > o=ene. > > value = 0 = 0x0 > > > > > > -- > > Use our news server 'news.foorum.com' from anywhere. > > More details at: http://nnrpinfo.go.foorum.com/ > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Returning memory Date: Sun, 03 Feb 2002 04:04:57 GMT From: dale Message-ID: References: <28809481.0202012009.10cf7545@posting.google.com> In article , denis_newsREMOVE-TO-REPLY@hotmail.com says... > * rasayani [02-Fri-02 20:09 -0800]: > > > > If a task allocates memory using malloc/calloc and is deleted, is the > > memory released to the free memory pool? > > as others have stated "no", but i think there is a > taskDeleteHookAdd() or something of that nature. you could > play with it, to make that function deallocate your memory > if you can't do it before deleting the task. > > i am not sure if that fn is called every time the task goes > away for whatever reason, or only on taskDelete() called by > a user, but it is something to look at > > denis > > Good suggestion, Denis. But, the task delete hook is performed as an extension of the windkernel and is not itself a task, whereas "free" requires a task context. What the hook routine could do is send the free request(s) to the excTask, via excJobAdd, which will perform the frees within a task context (at priority 0) after the hook is completed. The Task Delete Hook is called for every task termination and deletion. - ---- Dale --------------------------- Newsgroups: comp.os.vxworks Subject: Real Time OS for VME Date: 2 Feb 2002 23:22:38 -0800 From: nsaeidi@nri.ac.ir (N. Saeidi) Organization: http://groups.google.com/ Message-ID: <1d934551.0202022322.4a335f9f@posting.google.com> Hi everybody We are designing a modular system based on VME backplane. We design I/O modules powerful modules which use DSP) by own, but we are going to buy CPU module and its real time OS. In this regard I have gathered some information about OS9 , QNX, Linux,VxWorks, RT Kernel,MicroC OS and some other RTOSs. Now I really need recommendation for choosing best OS for our system and these are important things that we should consider : >> Cost >> Easy device driver developing (for our I/O modules ) >> LAN interface support So I would appreciate any recommendation and suggestion. Thanks for any response or opinion, N. Saeidi nsaeidi@nri.ac.ir --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado2 and PVCS Date: Sun, 3 Feb 2002 10:41:32 +0200 From: "Ofer Goren" Organization: Verio Message-ID: References: good one!!! :) - -- Ofer Goren Iamba Technologies "molochai" wrote in message news:g36a3a.na3.ln@vapour-trail.demon.co.uk... > Martin Raabe wrote: > > > Hello Ofer, > > I recommend to use SNiFF+ PRO as a WindPowerTool on top of Tornado. > > For more derails look at: > > www.windriver.com for a free eval download. > > > > Didn't I read somewhere (probably on here in fact) that WRS employees are > strictly prohibitted by WRS from posting to comp.os.vxworks? > > I'm sure that I'm not the only one to be getting totally pissed off with > your repeated attempts to make sales in responce to questions. > > M > > > Ciao > > > > Martin Raabe > > > > "Ofer Goren" schrieb im Newsbeitrag > > news:newscache$ku9rqg$lck$1@lnews.actcom.co.il... > >> Hi. > >> Does anyone work with this combination, and if so, any recommendations > > about > >> the best way to work? > >> > >> Tornado2.02 for win2kSp2 > >> PVCS for windows, version 6.7.11 > >> > >> Best reagrds, > >> > >> -- > >> Ofer Goren > >> Iamba Technologies > >> > >> > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: i have downloaded the tornado 2.0 protoyper, but the key as not been emailed to me Date: Sun, 3 Feb 2002 10:23:04 -0700 From: "Liang Yang" Organization: Arizona State University Message-ID: References: You should click Installation Keys on http://web1.windriver.com/windsurf/prototyper/ then choose your platform. "Mirak" wrote in message news:a3hmpt$faa$1@wanadoo.fr... > Do you have a key ? > Or how can i obtain it ? > It should be emailed to me but it isn't. > > --------------------------- Newsgroups: comp.soft-sys.matlab,comp.os.vxworks Subject: Returning data from S-func in RTW Date: Sun, 03 Feb 2002 21:53:32 GMT From: "Ole-Johan Ellingsen" Organization: Tele2 Norway AS Public Access Message-ID: Hi. I am using VxWorks on a target system and use Simulink/RTW to generate code. I have several IO modules that are interfaced using C S-functions. It works great, but I am having some trouble returning variables from the S-function. The problem occurs when I return signed real values. This is an example on how I do it now (example from the C S-func.): - -------- real_T* parrOutput = ssGetOutputPortRealSignal(S,0); int_T iCompX, iCompY, iCompZ, iRot; iCompX = (0xff00 & (gbl_pCompRotCanClientInfo->szRxBuffer[0]<<8)) | (0x00ff & gbl_pCompRotCanClientInfo->szRxBuffer[1]); iCompY = (0xff00 & (gbl_pCompRotCanClientInfo->szRxBuffer[2]<<8)) | (0x00ff & gbl_pCompRotCanClientInfo->szRxBuffer[3]); iCompZ = (0xff00 & (gbl_pCompRotCanClientInfo->szRxBuffer[4]<<8)) | (0x00ff & gbl_pCompRotCanClientInfo->szRxBuffer[5]); iRot = (0xff00 & (gbl_pCompRotCanClientInfo->szRxBuffer[6]<<8)) | (0x00ff & gbl_pCompRotCanClientInfo->szRxBuffer[7]); printf("X=%16.16X Y=%16.16X Z=%16.16X Rot=%16.16X\n", iCompX, iCompY, iCompZ, iRot); /* Compass values are signed 16-bit words. */ if (iCompX > 0x7fff) iCompX = - (0xffff-iCompX); if (iCompY > 0x7fff) iCompY = - (0xffff-iCompY); if (iCompZ > 0x7fff) iCompZ = - (0xffff-iCompZ); printf("X=%16.16d Y=%16.16d Z=%16.16d Rot=%16.16d\n", iCompX, iCompY, iCompZ, iRot); *(parrOutput+0) = -1; /* iCompX; */ *(parrOutput+1) = -2; /* (double)iCompY; */ *(parrOutput+2) = (real_T)iCompZ; *(parrOutput+3) = (real_T)iRot; printf("X=%16.16f Y=%16.16f Z=%16.16f Rot=%16.16f\n", *(parrOutput+0), *(parrOutput+1), *(parrOutput+2), *(parrOutput+3)); - -------- The first two printf's output the data correctly. Also, the first two test-numbers where I input -1 and -2 hardcode, works. But the two last values, typecast'ed to (real_T), are given extremely high, positive values. Does anyone know how I do this correctly? Best regards, Ole-Johan Ellingsen --------------------------- Newsgroups: comp.os.vxworks Subject: Thank You!!! Date: 3 Feb 2002 14:11:13 -0800 From: arun4vxworks@indiatimes.com (Arun Prasad) Organization: http://groups.google.com/ Message-ID: Hi All, We have delivered our first cut release of our first project/product on vxworks last week, ofcourse untested executable and source. Thank you so much for all your kind help. I need to mention so many people those who helped with their answers during this time. Now, its time for us to test & debug our code. Hoping to get your support/help once again. Have a nice time. Regards Arun Arun Prasad Software Engineer Adaptec India Pvt. Limited (A wholly owned subsidiary of Adaptec Inc., USA) 6-3-1086, IV Floor, Vista Grand Towers, Rajbhavan Road, Somajiguda, Hydereabad - 500082 Phone: (O):91-40-6661555/56/57/58 (Ext):234 Email: arun_prasad@adaptecin.com --------------------------- Newsgroups: comp.os.vxworks Subject: Thank you!!! Date: 3 Feb 2002 16:29:59 -0800 From: arun4vxworks@indiatimes.com (Arun Prasad) Organization: http://groups.google.com/ Message-ID: Hi All, We have delivered our first product on VxWorks OS (compilbale source code, untested one) to our customer last week for review. Thank you so much for all your clarifications/suggestions/replies during these time. It helped us a lot. I need to mention several of your names. Now, its time for us to test and debug what we have done so far. We may require further help on debugging too, as we are new to VxWorks. Thanks again for all your good work done at your part. Have a nice time. Regards Arun - ----------- Arun Prasad Software Engineer Adaptec India Pvt. Limited (A wholly owned subsidiary of Adaptec Inc., USA) 6-3-1086, IV Floor, Vista Grand Towers, Rajbhavan Road, Somajiguda, Hydereabad - 500082 Phone: (O):91-40-6661555/56/57/58 (Ext):234 Email: arun_prasad@adaptecin.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: transition from bootloader to vxworks Date: 3 Feb 2002 21:20:35 -0800 From: ramit.bhalla@wipro.com (Ramit Bhalla) Organization: http://groups.google.com/ Message-ID: <78ccdd99.0202032120.7728b6eb@posting.google.com> References: <78ccdd99.0202010454.6cabbf7d@posting.google.com> <3c5a97c1$1@brateggebdc5.br-automation.co.at> Thanks, got it :) Ramit. "Werner Schiendl" wrote in message news:<3c5a97c1$1@brateggebdc5.br-automation.co.at>... > Hi, > > VxWorks uses only the bootline from the bootrom image, and initializes > everything from scratch. > At least for pc486 and pcPentium BSP. > > This is true for both, Tornado 1.0.1 and Tornado 2.0.x. > Unless the address, where the bootline is stored, has changes since 1.0.1 - > there is no immediate need to compile a new bootrom, AFAIK. > > The biggest transition work is to change your config.h modifications into a > working VxWorks bootable project. > You can search for the macro names in the VxWorks configuration tool, if you > can get all your changes to config.h and friends from source control (or > memory ;-) ) > > Your application should run without changes, unless you hacked something in > VxWorks 5.3's sources. > > hth > Werner > > "Ramit Bhalla" wrote in message > news:78ccdd99.0202010454.6cabbf7d@posting.google.com... > > Hi, > > > > Hoping some of the experts could give me some tips. > > > > When using Tornado 1.0.1 ot Tornado 2.0.2 what is the transition > > scheme like. > > > > When the bootloader loads, it initialized the ethernet interface to > > download the image. > > > > When the image is loaded and starts execution does it reuse the same > > network interface or does it reinitialize the whole network again ??? > > > > Thanks, > > Ramit. --------------------------- Newsgroups: comp.os.vxworks Subject: Intermediate Driver ? Date: 3 Feb 2002 21:36:24 -0800 From: ramit.bhalla@wipro.com (Ramit Bhalla) Organization: http://groups.google.com/ Message-ID: <78ccdd99.0202032136.5dc2093d@posting.google.com> Hi, I was wondering if it's possible to introduce a intermediate driver between the IGMP/IP layer and the MUX layer ?? Is there any known/documented way to achieve this or would one need to hack to get this done ? Thanks, Ramit. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Returning memory Date: Sun, 3 Feb 2002 22:56:30 -0800 From: "GaryM" Message-ID: References: <28809481.0202012009.10cf7545@posting.google.com> "dale" wrote in message news:MPG.16c67317e8b30d8498969e@news-server.kscable.com... > In article REPLY@finch.lecs.cs.ucla.edu>, denis_newsREMOVE-TO-REPLY@hotmail.com > says... > > * rasayani [02-Fri-02 20:09 -0800]: > > > > > > If a task allocates memory using malloc/calloc and is deleted, is the > > > memory released to the free memory pool? > > > > as others have stated "no", but i think there is a > > taskDeleteHookAdd() or something of that nature. you could > > play with it, to make that function deallocate your memory > > if you can't do it before deleting the task. > > > Good suggestion, Denis. But, the task delete hook is performed as an > extension of the windkernel and is not itself a task, whereas "free" > requires a task context. What the hook routine could do is send the free > request(s) to the excTask, via excJobAdd, which will perform the frees > within a task context (at priority 0) after the hook is completed. > (Disclaimer) Without knowing about the rationale behind the original question... taskDelete() is a tool that is useful for preemptively shutting down tasks when the caller is well aware of any and all potential side effects. The most obvious use is for a shutdown task which attempts to stop all activity before placing all hardware in a quiescent state. Under normal circumstances, I would always recommend that tasks clean up their own data structures and exit on their own. It's easy enough to signal a task to shutdown, and then allow it enough time to recognize the signal and run down its resources. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado2 and PVCS Date: 4 Feb 2002 08:54:28 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: <3c587646$1@brateggebdc5.br-automation.co.at> <9ec53d11.0201310148.3c625ef5@posting.google.com> <3c59cb5b$1@brateggebdc5.br-automation.co.at> <9ec53d11.0202010640.5085a2bc@posting.google.com> Bill Pringlemeir wrote: > Well, there is one problem. The "C:/Tornado/target". I want to make > the build drive independent (note, this is a deficiency in Windoze) > and the WPJ files and other include absolute information and drive > letters. We put the complete Tornado tree into CVS, so including host, docs, etc. The only file to change after install is torVars.bat, you have to put the right paths in there, but after that no problems. But to be sure you hav to run Tornado installer after extracting the complete tree for the first time, but not a complete install, but point it to the directory you just installed Tornado in. Groeten, Johan - -- o o o o o o o . . . _____________________________ o _____ || Johan Borkhuis | .][__n_n_|DD[ ====_____ | borkhuis@agere.com | >(________|__|_[_________]_|__________________________| _/oo OOOOO oo` ooo ooo 'o!o!o o!o!o` === VxWorks FAQ: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html === --------------------------- Newsgroups: comp.os.vxworks Subject: Re: vxWorks and vxWorks.st Date: Mon, 4 Feb 2002 10:08:20 +0100 From: "Werner Schiendl" Message-ID: <3c5e4f86$1@brateggebdc5.br-automation.co.at> References: <78ccdd99.0201312251.54553c53@posting.google.com> <78ccdd99.0202032118.3a385e64@posting.google.com> Hi, "Ramit Bhalla" wrote in message news:78ccdd99.0202032118.3a385e64@posting.google.com... > Thanks for the tip. > I got the network up manually but I still can't get the target server > to attach with the standalone image(.st), but I can get to attach with > the vxWorks image ?? > > Any clue ?? You need to include and configure the wdb target agent. I think the corresponding macro is INCLUDE_WDB_NETWORK. Don't know exactly how to attach and start it, but this can be found from the startup files with some grep'ing (or file search on Windows) I guess. hth Werner --------------------------- Newsgroups: comp.soft-sys.matlab,comp.os.vxworks Subject: Re: Returning data from S-func in RTW Date: Mon, 4 Feb 2002 09:46:24 +0100 From: "Ole-Johan Ellingsen" Organization: Norwegian university of science and technology Message-ID: References: Hi again. Sorry, but the problem seems to have been obvious. :) Using parrOutput[2] instead of *(parrOutput+2) did the trick. Ole-Johan "Ole-Johan Ellingsen" wrote in message news:wji78.1037$3c.56056@juliett.dax.net... > Hi. > > I am using VxWorks on a target system and use Simulink/RTW to generate code. > I have several IO modules that are interfaced using C S-functions. It works > great, but I am having some trouble returning variables from the S-function. > The problem occurs when I return signed real values. > > This is an example on how I do it now (example from the C S-func.): > > -------- > real_T* parrOutput = ssGetOutputPortRealSignal(S,0); > int_T iCompX, iCompY, iCompZ, iRot; > > iCompX = (0xff00 & (gbl_pCompRotCanClientInfo->szRxBuffer[0]<<8)) | > (0x00ff & gbl_pCompRotCanClientInfo->szRxBuffer[1]); > iCompY = (0xff00 & (gbl_pCompRotCanClientInfo->szRxBuffer[2]<<8)) | > (0x00ff & gbl_pCompRotCanClientInfo->szRxBuffer[3]); > iCompZ = (0xff00 & (gbl_pCompRotCanClientInfo->szRxBuffer[4]<<8)) | > (0x00ff & gbl_pCompRotCanClientInfo->szRxBuffer[5]); > iRot = (0xff00 & (gbl_pCompRotCanClientInfo->szRxBuffer[6]<<8)) | > (0x00ff & gbl_pCompRotCanClientInfo->szRxBuffer[7]); > printf("X=%16.16X Y=%16.16X Z=%16.16X Rot=%16.16X\n", iCompX, iCompY, > iCompZ, iRot); > > /* Compass values are signed 16-bit words. */ > if (iCompX > 0x7fff) iCompX = - (0xffff-iCompX); > if (iCompY > 0x7fff) iCompY = - (0xffff-iCompY); > if (iCompZ > 0x7fff) iCompZ = - (0xffff-iCompZ); > printf("X=%16.16d Y=%16.16d Z=%16.16d Rot=%16.16d\n", iCompX, iCompY, > iCompZ, iRot); > > *(parrOutput+0) = -1; /* iCompX; */ > *(parrOutput+1) = -2; /* (double)iCompY; */ > *(parrOutput+2) = (real_T)iCompZ; > *(parrOutput+3) = (real_T)iRot; > printf("X=%16.16f Y=%16.16f Z=%16.16f Rot=%16.16f\n", *(parrOutput+0), > *(parrOutput+1), *(parrOutput+2), *(parrOutput+3)); > -------- > > The first two printf's output the data correctly. Also, the first two > test-numbers where I input -1 and -2 hardcode, works. But the two last > values, typecast'ed to (real_T), are given extremely high, positive values. > > Does anyone know how I do this correctly? > > Best regards, > Ole-Johan Ellingsen > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to change shell's default directory? Date: 4 Feb 2002 01:04:30 -0800 From: igtorque@eliop.es (=?ISO-8859-1?Q?Ignacio_Gonz=E1lez?=) Organization: http://groups.google.com/ Message-ID: <9ec53d11.0202040104.1835464e@posting.google.com> References: <9ec53d11.0202020631.21afbef6@posting.google.com> El 4 Feb 2002, a las 12:18, RajaSekhar escribió: > Hi, > > can you please tell me how to debug the application if you download your > applicatoin thru host shell (windsh). > After loading it, go to the debugger and voilá > Also can you please tell me by running the script file if any one > application fails in the script, how to proceed > with the next, as the windsh hangs there. Fix the application that fails. > > Thanks in advance for any help > > Raja Sekhar > I don't use the scripts to make automatic debugging, of course. I want to use them to automate regression tests, i.e.: tests that must pass or, if they don't pass, it's enough for me to know this and the offending function. So, I don't expect the functions to be called in the script to fail. If they fail, I enter the debugger, fix the function (or the test, if it was buggy) and re-run the regression tests. By the way, please Raja, use the newsgroup instead of direct e-mail. In this way, everyone will have the chance to learn (or unlearn.) --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Default scheduler in VxSim ? Date: 4 Feb 2002 01:05:08 -0800 From: deepaks@sasken.com (Deepak Shetty) Organization: http://groups.google.com/ Message-ID: <421ece65.0202040105.3d9c4bcf@posting.google.com> References: <421ece65.0201300525.73bf4466@posting.google.com> <3c581398$1@brateggebdc5.br-automation.co.at> <421ece65.0201302142.39ce2dba@posting.google.com> <3c5919cb@brateggebdc5.br-automation.co.at> <421ece65.0201312204.46ebe764@posting.google.com> <3c5a8d33$1@brateggebdc5.br-automation.co.at> > > Can u elaborate on this more ? any links to info where its explained > > as to why POSIX is not suitable for realtime apps ? > > > > I did NOT say it is not suitable, just not particularly designed for this > purpose. > Therefore, some idioms are not ideally adapted to VxWorks. > > AFAIK most posix APIs are designed on top of the VxWorks native APIs. > So if size does matter for your target image, you may be able to safe some > kB using VxWorks' native APIs. > > If you need to have the software as portable as possible, POSIX usually is a > good choice. yeah , got it, thanx, its more a matter of size & necessity than anything else :) > > > > > To make sure everything works as expected, you usually have > watchdogs > > > > > guarding the software. > > > > > There are software watchdogs (like wdLib in VxWorks) and hardware > > > watchdogs. > > > > If wdLib is for software watchdog, what is the equivalent of hardware > > watchdog ? > > > > This is completely hardware dependent. > > Most hardware watchdogs are simply triggered by writing to a particular > memory or IO address. > Maybe you need to alter the value written to this location between to > subsequent triggers. > > However, hardware watchdogs and how to use them must be documented with your > hardware. I will check them out in the manuals provided > > also if wdLib is software watchdog, where is > > hardware watch dog documented ? how do i access it, in case i need to > > ? In the doc. of wdLib, there is an example in which he uses this > > software watchdog to guard a routine, but from ur point i deduce that > > one needs to use hardware watchdog for the same ? I am bit confused in > > this hardware Vs software watchdog, are they the same, and their names > > ( h/w or s/w ) varies depending on the context in which it is used, > > like if used for guarding some routine, its called h/w wd else if used > > as a timer its called s/w wd ? > > > > Hardware watchdogs are very reliable, compared to software watchdogs. > > Do you know a bit of x86 ASM? > > how about > > CLI ; lock all interrupts > JMP $ ; spin forever > > Some code like this one can lock your system completely, a software watchdog > cannot become effective - since interrupts are locked. > A hardware watchdog, however, is independent of the CPU to go off and reset > the CPU (and hardware) into a safe state. > > The difference between software and hardware watchdogs is, that hardware > watchdogs are solded onto the CPU board (and therefore specific to the > target in use, the VxWorks simulator has _NO_ hardware watchdogs), while > software watchdogs are a software concept. > While you can have only a very limited number of hardware watchdogs (because > of board size, cost, etc.), software watchdogs are basically limited by > system memory and performance (eventually) of the system. > > Does this clarify the difference between the two concepts? Yup, thanx a LOT :) for ur time & effort deepak --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado2 and PVCS Date: Mon, 4 Feb 2002 09:56:23 +0100 From: "Werner Schiendl" Message-ID: <3c5e4cb9$1@brateggebdc5.br-automation.co.at> References: <3c587646$1@brateggebdc5.br-automation.co.at> <9ec53d11.0201310148.3c625ef5@posting.google.com> <3c59cb5b$1@brateggebdc5.br-automation.co.at> <9ec53d11.0202010640.5085a2bc@posting.google.com> Hi, "Bill Pringlemeir" wrote in message news:ulmedgllf.fsf@yahoo.com... > > >>>> Bill Pringlemeir writes: > > >> I have been toying with the idea of putting the Tornado headers > >> under source control so that every developer doesn't have to > >> install every patch that WRS has ever produced. > > >>>>> "Ignacio" == Ignacio González writes: > > Ignacio> We put under version control every Tornado file (.c or .h) > Ignacio> that has to be changed to support one of our products. For > Ignacio> instance, we have put files like: > > Ignacio> C:\Tornado\target\config\comps\src\romStart.c > [snip] > > Well, there is one problem. The "C:/Tornado/target". I want to make > the build drive independent (note, this is a deficiency in Windoze) > and the WPJ files and other include absolute information and drive > letters. This is true - however, only for files that are not within the project directory. If you cannot agree on a common drive for shared files in the projects, you can 'post-process' the .wpj file before checking 'em in. It's plain text files (with Unix file ending) and thus easily manageable with any serious scripting language. Environment variables should be fine, since VxWorks uses them itself (e. g. PRJ_DIR). > > In another post, David Laight gave the step of installing the OS. > This in itself _COULD_ cause a difference in generated code. I > suppose that that is the key point. You should have everything under > version control that is needed to replicate the output [vxWorks > images]. > This probably implies to check in the tools that come with tornado as well. For the GNU toolchain, its probably safe to just check it out from source control and use it (but never tried it). Other compilers may - or may not - work depending on how they expect to be installed (win registry!). > I feel much more confident in recommending that the Tornado files are > put under revision control. I may change the location of the files to > be something other than $(WIND_BASE). People often run out of space > on the "C:" drive. Especially if this is where the "Windows" > directory lies... `mount' is a beautiful thing! > Agreed. regards Werner --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado-2.1/mips and cp7000 BSP (Ocelot MIPS board) Date: 4 Feb 2002 09:05:29 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: Andray Kaganovsky wrote: > Greetings, > > I am using Tornado-2.1/mips and cp7000 BSP (Ocelot board by > Momentum Computers). The original version of this BSP that > came with the board was for Tornado-2.0x; I got T-2.1 compatible > version of this BSP directly from WindRiver. > > Here is my problems: > > -> printf "%d %d %d %d %d %d %d %d\n",1,2,3,4,5,6,7,8 > 1 2 3 4 5 6 7 8 > > -> sp printf,"%d %d %d %d %d %d %d %d\n",1,2,3,4,5,6,7,8 > 1 2 3 0 0 536912869 -33554961 60401228 I tried it on a IDT board using MIPS32sfgnu library and running the commands using the target shell, and I got the right output. Did you try it using the target shell or the host shell? Groeten, Johan - -- o o o o o o o . . . _____________________________ o _____ || Johan Borkhuis | .][__n_n_|DD[ ====_____ | borkhuis@agere.com | >(________|__|_[_________]_|__________________________| _/oo OOOOO oo` ooo ooo 'o!o!o o!o!o` === VxWorks FAQ: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html === --------------------------- Newsgroups: comp.os.vxworks Subject: Re: frequent rebooting for an ethernet interface looses connectivity?? Date: 4 Feb 2002 01:18:18 -0800 From: igtorque@eliop.es (=?ISO-8859-1?Q?Ignacio_Gonz=E1lez?=) Organization: http://groups.google.com/ Message-ID: <9ec53d11.0202040118.514e0357@posting.google.com> References: <3C5DF5A5.1F6CFFCC@lucent.com> Babu Yama wrote in message news:<3C5DF5A5.1F6CFFCC@lucent.com>... > Hi, > My ethernet device driver runs on a target that reboots "often". When the target > reboots, sometimes, immediately following the reboot, the target cann't be > pinged from the network. Then, the subsequent reboot restores the network > connectivity to the target. This happens intermittently. > > Any ideas as to what might be causing this. Is ARP playing a role here? Some > thing I need to do when I bring up the interface each time? > Apprecaite your help, > > Thanks in advance, > Babu The description of the problem is too vague for me. Which version of VxWorks do you have ? BSP ? Kind of target ? I had a problem similar to yours: after a reboot, my target sometimes refused to send messages through Ethernet (as a response to ping or when taking the initiative of sending messages). It was caused by a bug in the Ethernet driver that I managed to fix (later, the driver's supplier learned of it and also fixed the driver.) --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Mon Feb 4 04:17:32 2002 From: "praba_ganesh" Date: Mon Feb 4 04:17:34 PST 2002 Subject: File system - reg --=_MAILER_ATTACH_BOUNDARY1_20022411744381931706506 Content-Type: text/plain; charset=us-ascii Hi Friends, We are assigned the task of developing a Flash driver & integrate the same with a File system. The system is an ARM7TDMI based board. We have to use VxWorks for this purpose. Now we are standing in the design phase. The chip we are planning to use is MBM29-Fujitsu flash. The whole system needs to boot from the FLASH. So the Flash would hold the BOOT image as well as the data area( need to store HTML pages). The problem is that, we can't use the TrueFFS since it is an optional component. The data area of the flash has to store the native system configurations while power down. So the flash would operate in both read/write mode. However in Tornado help manual it is specified that the , Flash that stores the VxWorks image should only be a ROM & shouldn't be Written. If the flash(with image) would undergo write cycles, then the whole image needs to be transferred to RAM at the strt-up. Is my above understanding correct. Next regarding the file system: Can I implement the dos file system on the Falsh which stores the image & the data memory. The dos FS would have a FAT table. Will the FAT be a part of Data area. My primary worry is , Can I have a dos FS or RawFS over the flash?. I am not bothered about the Falsh wear cycles at this stage. In case , if the above said file systems can't be used with flash, what would be proper alternative. Above the file system layer there would be the FTP & HTTP servers running. However there would be frequent write cycles into the flash. Next problem is, The user should be able to update the VxWorks boot image. Perhaps if the user dumps a corrupted image , then how to bring back the system to live condition. Should I need to go for a seperate boot loader which can point to both old & new image. My last query is , Can a NAND flash device be made bootable, where the flash contains both Image & data. Thanx & Regards Prabavathy Ganesh Get Your Private, Free E-mail from Indiatimes at http://email.indiatimes.com Buy Music, Video, CD-ROM, Audio-Books and Music Accessories from http://www.planetm.co.in --=_MAILER_ATTACH_BOUNDARY1_20022411744381931706506 Content-Type: text/html; charset=us-ascii