From vxwexplo-errs@csg.lbl.gov Sat Dec 1 04:03:17 2001 From: Vxworks Exploder Date: Sat Dec 1 04:03:19 PST 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sat Dec 1 04:03:12 PST 2001 Subject: Re: Need information on Zinc configuration Subject: Re: New to vxWorks 5.4 Subject: Re: Dynamic loading Subject: vxsim error: Module is not fully linked Subject: Re: GUI display on top of VxWorks ? Subject: Re: what's the reason which will result in workQPanic? Subject: Re: TRUE definition Subject: Re: Faster stat function? *OR* The stat implementation? (In a C/C++ program.) Subject: Re: Cross compiling problems!!! HELP!! Subject: Re: Is this Q so dumb that no one is answering Subject: Re: __get_eh_context called from ISR ?? Subject: Re: what's the reason which will result in workQPanic? Subject: Re: Need information on Zinc configuration Subject: Problems with loadLib Subject: processor migration (MVME230X -> MVME5100 ALTIVEC) Subject: Re: Problems with loadLib Subject: Re: TFFS - flash file system Subject: Re: Tornado 2, PPC Cumulative Patch 4 Subject: Re: Problems with loadLib Subject: MPC8260 and SPI problem Subject: Re: Dynamic loading Subject: MIB table with in a MIB table Subject: Problem Compiling bootroms for Walnut PPC405GP Subject: tWdbTask Crashing Subject: Re: TFFS - flash file system Subject: Re: Writing END Driver--Crashing tNetTask (Corrupt clBlk?) Subject: Re: processor migration (MVME230X -> MVME5100 ALTIVEC) Subject: Debugging IP Problems--How? Subject: FS VME Elma cage w Heurikon V960D (8) Subject: iov_len limit in sendmsg call Subject: Bandwidth Management vs. Qos Subject: Need a driver for MB86964 Ethernet ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Need information on Zinc configuration Date: 30 Nov 2001 12:22:58 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: <442C562A1DD6D3118C6E0004AC4C175902ECFD82@bngex01> "Bill Pringlemeir" bpringlemeir@yahoo.com wrote: > [I get a request like this about once per month. Does anyone else on > `c.o.v'? What is the best way to deal with it? I am receiving these almost daily lately. If it a very easy problem I answer it, but normally I refer to the newsgroup, something like: Please post this message to the newsgroup, using news:comp.os.vxworks or http://groups.google.com/groups?group=comp.os.vxworks. 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: New to vxWorks 5.4 Date: Fri, 30 Nov 2001 13:39:34 +0100 From: "Werner Schiendl" Message-ID: <3c077ecc$1@brateggebdc5.br-automation.co.at> References: <9u0q1l$ll6$1@crc-news.crc.ca> Hi, to build a new bootrom, you modify the config.h file as usual for the old VxWorks stuff. However, Tornado II contains both - END and BSD network drivers. Be sure to include the right stuff by defining INCLUDE_BSD or INCLUDE_END Some drivers exist for both configurations, while others exist only for one. IIRC for your network board you need the elPci driver, which i think was END type driver. To make a new VxWorks image, you should use the project facility and create a new, bootable project based on the template project that comes in the target/proj subdir of your install for your BSP. Add all the stuff you need to it, there is no separate .st configuration AFAIK. You specify the configuration in the GUI tool and thats it. For each distinct configuration you normally need a separate project. hth Werner "Donald McLachlan" wrote in message news:9u0q1l$ll6$1@crc-news.crc.ca... > I've used vxWorks for ages by writing makefiles, editing the config files, typing > make whatever. No probs. Recently got a new Tornado 2 for Solaris host/pc target > and trying to learn my way around the gui tools. > > Ok, made a bootrom (floppy) with bootrom_uncmp and vxWorks.st from the > pcPentium target dirctory. Works at treat on the P3 500 I found. > > I'd like to build a project based on that directory and start by re-making the > bootrom_uncmp, and vxWorks.st prior do trying to add the if_elt driver. > > I was able to make an unbootable bootrom_uncmp by defining INCLUDE_ELT in the > config.h file, :-) ... and to make a bootable one again by undefining it. > > How do I re-make a usable vxWorks.st? > > Is there a better set of step-by-step instructions for this other than those in the > guide (the guide I was using is in the lab and I'm in my office now, but I think > I was using the users guide. I know it was not the programmers guide as that is > under my left elbow as I type. :-) > > What I'm hoping to do is to add the 3C905B card to the vxWorks image and get that > running ... and hopefully add the same card to the bootrom_uncmp so I can load the > vxWorks image across the network (so I won't have to keep copying it to the boot > floppy every time I add more things to the vxWorks image; like usrTime). > > Thanks, > Don > > -- > Donald McLachlan E-mail Donald.McLachlan@crc.ca > Communications Research Centre / RNS Tel (613) 998-2845 > 3701 Carling Ave., Fax (613) 998-9648 > Ottawa, Ontario > K2H 8S2 > Canada > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Dynamic loading Date: Fri, 30 Nov 2001 16:20:17 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <9u2qu0$dn9$1@newshost.mot.com> Martin, You are currently using host-based tools, including debugging and dynamic loading. Historically this is a late addition to VxWorks from around 1996). Before that, VxWorks only had target-based tools, including Shell, dynamic loader, assembly-level debugger. These target-based tools are still available, but are disabled by default. For source level debugging one must have the host-based tools, so if the host-based (target server) and target-based tools are to be used concurrently, one needs to turn on the symbol table synchronization so the two tool sets can coexist peacefully. I have seen many applications where target-based dynamic loading is used in the production environment, some are for good reasons, others not. For production, you can either have the whole app linked with VxWorks for a single bootable (or ROM-resident) image, or have this split into multiple modules, which are loaded subsequently. Run-time dynamic loading will cost you slower startup time for the application. Here is an example where this can be of good use: Say you make an ATM device, and one of the modules implements LAN Emulation. Trouble is there are three versions of LANE you need to support, but the user will need to have only one of the three versions running depending on the site-wide configuration. You can have three different loadable modules on the Flash file system, each implementing one of the three versions, but all having the same API. Then at startup, examining the configuration you load one of the three modules, so the two unused modules do not waste RAM space, and all three versions can have a common source base - loading only one of the three avoids possible symbol conflicts etc. There are other examples, so in order to help you decide weather to use dynamic loading in production, one would need to know much more details about your application and constraints. Hope this helps, - Leonid "Martin Roth" wrote in message news:9u2qu0$dn9$1@newshost.mot.com... > We are now in the definition stage of our new project which will probably > based on one of MPC8xx PPC processors. During the developement purposes > Vxwrks allows to dynamically link & load new modules and run them from the > target shell. I would like to perfrm dynamic linking/loading on the final > target (non development) sysem where I do not have the Tornado development > tools and the WDB agent within the target. Did anybody see such (non > development) system where application are linked and loaded dynamically. ? > Thanks, > Martin Roth > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: vxsim error: Module is not fully linked Date: 30 Nov 2001 15:09:55 +0100 From: Klaus Zeitler Organization: Lucent Technologies Message-ID: Sender: kzeitler@sfsw51 Hi, I've tried to write, generate and build a simple example program with Objectime on SUN/Solaris. Both the host (SunOS 5.8) and the vxworks target image run successfully. The program seems to build fine also for vxsim in Objectime (no linker errors), but when I try to run it I get the following error: Command: tgtsvr vxsim0 -V -n vxsim0 -B wdbpipe -c /vol/gen/kzeitler/ot/SDT_test_system/build/SDT_test_system_vxsim/top -RW -Root /home/kzeitler/tmp/tsfs tgtsvr.ex (vxsim0@sfsw51): Thu Nov 29 14:57:33 2001 Connecting to target agent... succeeded. Attaching C++ interface... succeeded. Attaching elf OMF reader for SIMSPARCSOLARIS CPU family... succeeded. Error: Module is not fully linked. Error: Object module load failed for /vol/gen/kzeitler/ot/SDT_test_system/build/SDT_test_system_vxsim/top Error: loading core symbol table: status = 0x10029 Error during target server core initialization My guess is that I'm missing some build option in Objectime but since i can't find any hints or documentation besides changing the compiler to ccsimso I thought I'll try to get some hints here in the newsgroup what may be missing. Thanx Klaus - -- ------------------------------------------ | Klaus Zeitler Lucent Technologies | | Email: kzeitler@lucent.com | ------------------------------------------ - --- Why be a man when you can be a success? -- Bertold Brecht --------------------------- Newsgroups: comp.os.vxworks Subject: Re: GUI display on top of VxWorks ? Date: Fri, 30 Nov 2001 15:17:40 -0000 From: Organization: Posted via Supernews, http://www.supernews.com Message-ID: References: <488e459a.0111292130.108a0850@posting.google.com> John wrote: > For more complex solutions, or just more portable ones, there is also > an X server in the partner list, but I can't remember the company name > (check in the TradeWinds book, or online). That X server (and client-side library) company would be Seaweed Systems and we're at www.seaweed.com. bob --------------------------- Newsgroups: comp.os.vxworks Subject: Re: what's the reason which will result in workQPanic? Date: 30 Nov 2001 09:12:23 -0800 From: marsgal42@hotmail.com (Laura Halliday) Organization: http://groups.google.com/ Message-ID: References: <221caa76.0111290329.12f02cb2@posting.google.com> kevinatnj@yahoo.com.cn (kewin) wrote in message news:<221caa76.0111290329.12f02cb2@posting.google.com>... > Hi > when i can debug ARM7TDMI bsp a problem met with me.My ethernet port > driver process packet by the way of POLL. When I hook my board to our > company local LAN,I got the following trouble,"workQPanic kernel work > queue overflow".I don't know what the reasons will result in this kind > of problem. > Any advice will be welcome Wind River have an FAQ on the subject. Ask them for it. Briefly, when VxWorks is in kernel mode (from interrupt handlers, executing system calls, and so), it defers clock processing. Timer ticks that occur during kernel mode get saved in an internal data structure, and are then processed on exit from kernel mode. Should this data structure overflow, the result is a workQPanic() crash. Polling a network interface?! Laura Halliday VE7LDH "Que les nuages soient notre Grid: CN89lg pied a terre..." ICBM: 49 16.57 N 123 0.24 W - Hospital/Shafte --------------------------- Newsgroups: comp.os.vxworks Subject: Re: TRUE definition Date: Fri, 30 Nov 2001 16:57:45 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: <9u2qjq$smr$1@plutonium.btinternet.com> Ofer Goren wrote in message ... >well, as I said, this is not up to me. The fact is that I HAVE to change >TRUE to -1. No you don't, you just *want* to. Nobody is going to shoot you if you don't. Your boss might not like you but if they give you impossible orders to carry out you can't be expected to succeed, and if he fires you for not following impossible orders you can make more money from the tribunal for wrongful dismissal. > The only constant defined is C is 0 (zero) for FALSE. all other >are compiler dependent. I truly hope that the developers of vxWorks know >this (and compare boolean expressions with FALSE), or I will have to tell >someone "I told you so".... ;) You are wrong. The C language standard DEMANDS that the result of a conditional operator is 1 if the condition is true and 0 if it is false. If you have to connect two pieces of code, one that uses +1 for true, and one that mistakenly uses -1 to mean true, it is up to you to write an interface between them that converts the values that one routine understands into the values that the other routine understands. This is just sound engineering practice: treat program modules as black boxes, and if two don't fit together neatly, make an interface between them. Never under any circumstances jam them straight together in a way that only happens to work because of the internal details of one of the black boxes, because the internal details of a black box are not guaranteed to stay the same, only the external interface is. DaveK - -- Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Faster stat function? *OR* The stat implementation? (In a C/C++ program.) Date: Fri, 30 Nov 2001 16:43:30 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: <9220e2a6.0111281540.61cf3f0b@posting.google.com> Vito Pirelli wrote in message <9220e2a6.0111281540.61cf3f0b@posting.google.com>... >Hello, > >I will need to determine the dates on thousands of files, the only way >of doing that that I can find is using the stat command. The problem >is that stat fills in a struct with a lot stuff that I don't need. >This grately slows down my program. > > >For this reason I'm looking for a better way or reading the date of >the file or the implemenation of the stat command so that I can remove >the code that doesn't pertain to the retrival of the date from the >file. If you're using a FAT format drive, your best bet would be to do a web search to find the documentation of the layout, then you can write code that reads the raw sectors of directories from the disk and extracts all the file names and timestamps in one go. DaveK - -- Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Cross compiling problems!!! HELP!! Date: Fri, 30 Nov 2001 17:05:53 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: <52e25f45.0111261144.6d40284c@posting.google.com> <52e25f45.0111271558.332b0333@posting.google.com> Vittal Krishnamurthy wrote in message <52e25f45.0111271558.332b0333@posting.google.com>... >Thanks Dave, >I have a small problem with the 2.95.3 GCC. For some reason, the gcc >does not recognize line continuation char "\". It complains of parse >errors. Do u have a solution other than going to each .h file and >removing the '\' and joining lines??? > >Thanks >-Vittal I've posted a reply to your other thread, look for Message-ID: DaveK - -- Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Is this Q so dumb that no one is answering Date: Fri, 30 Nov 2001 17:07:02 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: <3C056D47.BA4D920A@ubbsrv3.ho.lucent.com> Chandwani wrote in message <3C056D47.BA4D920A@ubbsrv3.ho.lucent.com>... >> Hi all ! >> >> I have a a couple of questions regarding the multiple ip address provisioning on >> same IP port in vxworks IP stack. >> >> A port has a private class C IP address ( non broadcast ) provisioned on it. Now >> I go and add a public IP address on that port using function ifAddrAdd(). This >> works successfully. >> >> Question 1) If I do a ifAddrGet(), which IP address will I get of the two ? >> Question 2 ) If I want to change the public IP address on that port but maintain >> the private IP address, how do I achieve this ? >> >> Please help anyone. >> >> Thanks, >> Rakesh >( maybe no one knows the answer ;-). If I don't get it here, I will go and test >this on a single board computer and enlighten you folks ) I'm only not answering because I don't know the answer :) Do let us know what the result of your test is, then the answer will be on Google for anyone in the future who wants to know! DaveK - -- Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! --------------------------- Newsgroups: comp.os.vxworks Subject: Re: __get_eh_context called from ISR ?? Date: Fri, 30 Nov 2001 16:37:51 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: Bill Pringlemeir wrote in message ... >>>>>> "Sanjay" == Sanjay writes: >[snip] > Sanjay> Can anyone shed some light on what "__get_eh_context" is and > Sanjay> why it gets called from the ISR? And why is there this > Sanjay> problem with the memory leak? > >"__get_eh_context" is an abbreviation for "get exception handler >context". This will be called if you have C++ code, you are executing >in an objects context and you have exception handling enabled. The >code uses the `this' pointer and can setup RTTI (run time type >identification) and other C++ facilities. > >You could use only Embedded C++ constructs and you will probably be >ok. Alternatively, don't allocate any objects, even on the stack in >your ISR. Object construction can call malloc and other routines that >you shouldn't use. It will definitely setup an exception context so >that if an exception is thrown, the object can be deleted. C++ can be used in ISRs with extreme caution, like having the ISR in a separate module that you can build using -fno-exceptions and -fno-rtti and being careful not to construct any objects and watching out for temporaries caused by reference parameters and so on, but there are so many pitfalls that my advice would be to always write such lowlevel code as ISRs and even the lower layers of device drivers in plain C. DaveK - -- Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! --------------------------- Newsgroups: comp.os.vxworks Subject: Re: what's the reason which will result in workQPanic? Date: Fri, 30 Nov 2001 16:39:19 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: <221caa76.0111290329.12f02cb2@posting.google.com> kewin wrote in message <221caa76.0111290329.12f02cb2@posting.google.com>... >Hi > when i can debug ARM7TDMI bsp a problem met with me.My ethernet port >driver process packet by the way of POLL. When I hook my board to our >company local LAN,I got the following trouble,"workQPanic kernel work >queue overflow".I don't know what the reasons will result in this kind >of problem. > Any advice will be welcome One common cause is that you have an ISR and the ISR fails to clear the flag that cause the interrupt, so as soon as the ISR exits it is called again, and the foreground never gets a chance to run. DaveK - -- Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Need information on Zinc configuration Date: Fri, 30 Nov 2001 17:23:27 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: <442C562A1DD6D3118C6E0004AC4C175902ECFD82@bngex01> Bill Pringlemeir wrote in message ... >[I get a request like this about once per month. Does anyone else on >`c.o.v'? Not from c.o.v, but once in a while from the occasional other mailing list. > What is the best way to deal with it? If I'm in a good mood and it doesn't look like too much trouble, I help out. Reposting the question somewhere public is reasonable enough though I would prefer to just dash off a quick reply to the poster pointing them at the public resource (cov, or the exploder, or another mailing list, or whatever). And of course just binning their email completely isn't unreasonable considering you don't really owe them anything in the first place. > I guess having an >mangled email would prevent it. Works for me :) > People asking me to do their work for >them, p!?ses me off. Well, but it's in all our interests to "bring the kids up right".... the industry as a whole and the public at large benefit from the 'open source' attitude to knowledge, just think that one day that newbie you flamed might be writing the source code for the control systems of *your* local nuclear power station, and then you might wish you had shown them the way to do things properly. It's people asking me to effectively write their college homework assignments for them that pisses *me* off! > I would really like it if WRS didn't include >email addresses in there WindSurf search engine. That's why I unsubbed the exploder and just use cov over NNTP now. DaveK - -- Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! --------------------------- Newsgroups: comp.os.vxworks Subject: Problems with loadLib Date: 30 Nov 2001 10:29:35 -0800 From: sanjay.bhandari@cportcorp.com (Sanjay) Organization: http://groups.google.com/ Message-ID: Hi all I realize that issues with loadLib are VFAQ - but I could not find any info on this particular one. I have tried loadLib in various VxWorks configurations - each one with a different problem result. And I can't get it to work (am using -mlongcall already). (1) Options #define INCLUDE_LOADER /* object module loading */ #define INCLUDE_UNLOADER /* object module unloading */ #define INCLUDE_SYM_TBL /* symbol table package */ result in a message like "can't add XXX to symbol table" and "undefined symbol YYY". (2) Options #define INCLUDE_LOADER /* object module loading */ #define INCLUDE_UNLOADER /* object module unloading */ #define INCLUDE_SYM_TBL /* symbol table package */ #define INCLUDE_SYM_TBL_SYNC /* synchronize host and target symbol tables */ result in no error messages and loadModule returns what seems like a valid MODULE_ID. But the loaded module is not visible in the module browser and the loaded code can't be executed. Can anyone suggest how loadLib may be used without relying on the host based symbol table? Thanks. Sanjay --------------------------- Newsgroups: comp.os.vxworks Subject: processor migration (MVME230X -> MVME5100 ALTIVEC) Date: Fri, 30 Nov 2001 15:00:31 -0500 From: "Stephen" Message-ID: We are using VxWorks on MOTO MVME2300 boards. We are interested in adding a motorola 5100 ALTIVEC card to an existing VME chassis (i.e. with teh 2300 boards) Wind River is saying that in order to support this we need to buy a whole new development package !! We've moved to differnet processor before and had to simply pay for the new BSP. I can see paying a bit extra for ALTIVEC libarraies or something liek that but a whole new dev system does not seem correct. I am wondeing if anyone else is facing the same issue. Has anyone migarted from MVME2300 to 5100 ALTIVEC boards ? Thanks for any help Stephen Rejto --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Problems with loadLib Date: Fri, 30 Nov 2001 21:06:27 GMT From: Holger Behrens Organization: Arcor Message-ID: <1103_1007154387@news.arcor.de> References: Sanjay, are you sure you have configured a symbol table onto your target, and if the target should download the symbol table it does so successfully? INCLUDE_SYM_TBL is contains just the functionality to handle a symbol table, it does not create on. One of these does: #define INCLUDE_STANDALONE_SYM_TBL /* compiled-in symbol table */ #define INCLUDE_NET_SYM_TBL /* load symbol table from network */ Just to confuse people the network could be a local filesystem. ;-) Hope this helps. Holger On 30 Nov 2001 10:29:35 -0800, sanjay.bhandari@cportcorp.com (Sanjay) wrote: > Hi all > > I realize that issues with loadLib are VFAQ - but I could > not find any info on this particular one. > > I have tried loadLib in various VxWorks configurations - > each one with a different problem result. And I can't get > it to work (am using -mlongcall already). > > (1) Options > > #define INCLUDE_LOADER /* object module loading */ > #define INCLUDE_UNLOADER /* object module unloading */ > #define INCLUDE_SYM_TBL /* symbol table package */ > > result in a message like "can't add XXX to symbol table" and > "undefined symbol YYY". > > (2) Options > > #define INCLUDE_LOADER /* object module loading */ > #define INCLUDE_UNLOADER /* object module unloading */ > #define INCLUDE_SYM_TBL /* symbol table package */ > #define INCLUDE_SYM_TBL_SYNC /* synchronize host and target symbol tables */ > > result in no error messages and loadModule returns what seems > like a valid MODULE_ID. But the loaded module is not visible > in the module browser and the loaded code can't be executed. > > Can anyone suggest how loadLib may be used without relying > on the host based symbol table? > > Thanks. > > Sanjay --------------------------- Newsgroups: comp.os.vxworks Subject: Re: TFFS - flash file system Date: Fri, 30 Nov 2001 21:07:41 +0000 From: molochai Message-ID: <9029b5e14a%molochai@vapour-trail.demon.co.uk> References: <9u2v6q$g3j$1@newshost.mot.com> <488e459a.0111281120.765483ed@posting.google.com> <4bb6bae04a%molochai@vapour-trail.demon.co.uk> <9u5ec5$i7t$1@neptunium.btinternet.com> In message <9u5ec5$i7t$1@neptunium.btinternet.com> "David Laight" wrote: > > when is the directory entry updated to reflect the new file size? Perhaps > > dosFs2 is smart enough to perform smaller-than-cluster operations. I will > > check. > > Does the TFFS layer check whether you are writing unchanged data, or > whether (if appropriate for the flash chips) only converting 1's to 0's - > in which case the existing sector can be updated. I have no idea, but I would be surprised if there were many circumstances in which this optimization would fire. > > > > Almost certainly. You can get this down to clearing a few bits though, > > and one mapping value. > > Most changes in sector mapping will (I guess) be done my changing the next > entry in the map area from ffff:ffff to :. It may > zero the old map for that sector. Only when the map is full does the map > need to be erased and re-written (probably worth having 2 map areas...) I did it by distributing he map around the flash with a localised map for each erase unit. When you erase a unit, there's nothing in the associated map by definition, so it's completely safe. > > > 32MB == Mega *bit* == 4 Mbyte. Note the capitalisation. The erase units > > are still *huge* at 128Kb each. > > The fact that each erase probably takes half a second kills performance if > you don't do something other than 128k read/modify/write cycles for a > sector write. How many 128k erase units are you willing to hold buffered > in main memory? > Sectors != erase units. The only time you take the hit is when desperately needing to reclaim blocks for new allocation. > [big snip] > > > > > If all your code does is write flash, I think DosFS would last about 24 > > > days [assuming an infinite CPU speed ;-]. > > I remember reading a AMD serial EEPROM data sheet which quoted 10 years > retention and 10^6 write cycles. It did point out that you probably didn't > need 10 years retention on the cells you wrote 10^6 times - so in fact you > could get away with many more erases - or a longer retention time. > > > > > > You're maybe misunderstanding the timings. However, if the flash > > component drivers are using taskDelay() in their delay loops, the > > performance is going to suck, as some of the writes take considerably > > less than a kernel tick. > > > > You could probably improve the performance of any flash based files system > - with or without a translation layer - by a few tweaks to the filesystem > code. For instance (on a dos like fs): > 1) treat direcetory entries that are all 0xff as unused. If a directory > entry has to updated, overwrite one of the 'ff' entries and zero out > the old one. > 2) similarly for FAT entries - maybe ffff = unused, 7fff = end of chain. > 3) add an 'erase' call along with 'read' and 'write' to the block driver > interface. I think you'll find TFFS probably deals with this by tracking the contents of the FAT when there is a DOS compatible filing system on top of it. All it needs to do is watch when a cluster becomes unreferenced and mark it as ready for reclaimation. What I have seen dosFs do however, is frequently update the FAT as blocks are allocated to a file. This is what you would expect (and indeed want to happen). --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado 2, PPC Cumulative Patch 4 Date: Fri, 30 Nov 2001 21:09:11 GMT From: "Ho-Kuo Chan" Organization: GlobeTrotter Message-ID: References: <3C06AFD7.CB2C9D7A@nunya.com> Hi Nunya, "nunya" wrote in message news:3C06AFD7.CB2C9D7A@nunya.com... > Hi All: > > We've run into a problem with this patch. > > We have our own networking protocol stack that registers > with the mux as Promiscuos (because of the well-known > failures with Snarfing. ) > > Up to cumulative patch 2, this was working fine. > > With cumulative patch 4, however, our receive routine is > never called. This makes it very difficult for our protocol > stack to function efficiently. :-) > > Has anyone else run into this problem? T2CP4 upgrades Tornado 2 to Tornado 2.0.2. T2CP4 definitely causes problems with the PPC405GP. There is a PPC405 specific patch SPR 67953 which seems to work fine. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Problems with loadLib Date: Fri, 30 Nov 2001 16:12:19 -0500 From: "RonnoBonno" Message-ID: <9u8sl2$6dgmk$1@ID-68014.news.dfncis.de> References: Sanjay, I use the loader with a target resident symbol table with no problem. I typically do: -> ld 1,0,"myfile.outnm" (.outnm is the non-munched version of the output file) and it works fine. I have my the symbol table included in my vxworks image(vxworks.st.cmp) with Tornado 2.0 and VxWorks 5.4. Happy Coding, Ron PS: I have these defined in prjcomps.h generated from Tornado 2.0 #define INCLUDE_STANDALONE_SYM_TBL #define INCLUDE_SYM_TBL #define INCLUDE_SYM_TBL_INIT #define INCLUDE_SYM_TBL_SHOW #define INCLUDE_SYM_TBL_SYNC Sanjay wrote in message ... >Hi all > >I realize that issues with loadLib are VFAQ - but I could >not find any info on this particular one. > >I have tried loadLib in various VxWorks configurations - >each one with a different problem result. And I can't get >it to work (am using -mlongcall already). > >(1) Options > >#define INCLUDE_LOADER /* object module loading */ >#define INCLUDE_UNLOADER /* object module unloading */ >#define INCLUDE_SYM_TBL /* symbol table package */ > >result in a message like "can't add XXX to symbol table" and >"undefined symbol YYY". > >(2) Options > >#define INCLUDE_LOADER /* object module loading */ >#define INCLUDE_UNLOADER /* object module unloading */ >#define INCLUDE_SYM_TBL /* symbol table package */ >#define INCLUDE_SYM_TBL_SYNC /* synchronize host and target symbol tables */ > >result in no error messages and loadModule returns what seems >like a valid MODULE_ID. But the loaded module is not visible >in the module browser and the loaded code can't be executed. > >Can anyone suggest how loadLib may be used without relying >on the host based symbol table? > >Thanks. > >Sanjay --------------------------- Newsgroups: comp.os.vxworks Subject: MPC8260 and SPI problem Date: Fri, 30 Nov 2001 16:17:42 -0500 From: cheru Organization: Mailgate.ORG Server - http://www.Mailgate.ORG Message-ID: <3C07F776.F73FDDE1@nospam.com> Sender: bpatel@fwion.com Hi everyone, BACKGROUND: We are using the 8260 SPI port in single master mode. The os is vxworks5.4 The BSP is based on SBC8260. We have a single device on the SPI. It is an FPGA with a SPI driver interface programmed into it (our code). The way the SPI is configured is: 48 bits Tx and Rx. We were using the SPI at 16 MHz. (BRGCLK/4, PM=0) CORE=200Mhz, CPM=133Mhz, Bus Freq=66Mhz. SPI settings - ------------ The brgClk to spi=66.5Mhz. The parameter ram is user programmed in dual port ram (Bank#8, address=0x3800) The Transmit and receive buffer descriptor are used from bank#0 of dual port ram, i.e., address=0x0000 The transmit and receive buffer pointer in Tx/Rcv BD are in dual port ram address 0x2400 and 0x3900 respectively. MRBLR in parameter ram=0x10. parameter ram(RFCR)=0x1 parameter ram(TFCR)=0x1 spi registers initialized to following - -------------------------------------- SPI mode register: CI=0, CP=0, REV=1, MS=1, EN=1 (set to 0 if there is no spi operation), LEN=8, PM=1. PROBLEMS: 1) We occasionally we experience an incorrect data read from the RX buffer. The 7 (and sometimes 8) MSBitss are all ZEROs. When we examine the SPIMISO, we see the correct data containing some ONEs. We can detect this if it occurs in the first two bytes but the last four bytes are data which we have to accept if the first two bytes are correct. This ZEROing always occurs in the MSB of 16 bits increment. 2) Separately, we are not receiving indication via the SPIE, event register, that the RXB occurred even when TxB is set. Our SW has timeouts in the event of this so that it can carry on. At 16MHz, we would see these incorrect data reads more frequently. We slowed the clock down to 8 MHz and the occurrence decreased but did not go away entirely. If we drop the speed lower, we run into other issues with software/hardware erratic behaviour. So 4MHz and 5.3MHz we could not use this setting. We did look at the clock signal for each of the 4 clock frequency setting. The results: at 4 and 5 MHz, the clock is continuous. at 8 MHz the clock had a single break in it after the 16th cycle and the 16 MHz clock had 4 breaks in it usually after the 8th, 24th, 32nd and 40th cycles. It looks like the high frequencies may have some impact on the CPM. May be someone can shed some light on this. The CPM is configured with the use of 4 ports, obviously the SPI running at 8 MHz, the FCC3 is used for Enternet at 100 MHz, the I2C bus at 30KHz and the SMC at 9600. We input these rates to a motorola utility to determine the utilization of the CPM and found it at 52.52%. QUESTIONS: 0) Why is the data ZERO'd and SPIE RXB is not SET, when TXB is SET? 1) Is there some SPI priority settings built in or can the priority be set wrt to FCC I2C, SCC etc. 2) Is the CPM resourced sufficiently? Are we using the SPI too fast? 3) Currently we use dual port ram to store tx/rx buffer BD and tx/rx buffer pointers. changing rx buffer pointer to use SDRAM does not work. What registers specific to SDMA need to be configured so that SDMA can arbitrate the 60x bus, park itself on 60x bus and use it to store the incoming SPIMISO data into the receive buffer in sdram? Thank you for your help. cheru --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Dynamic loading Date: Fri, 30 Nov 2001 16:25:34 -0500 From: "RonnoBonno" Message-ID: <9u8tds$6u5d9$1@ID-68014.news.dfncis.de> References: <9u2qu0$dn9$1@newshost.mot.com> I have used dynamic loading with good results, primarily as a CGI mechanism with HTTP. Here is a snypit of code that shows it: Happy Coding, ROn ... MODULE_ID cgiModuleID=0; MODULE_INFO cgiModuleInfo; . . /* ------ Load Object Module into Memory ---------------------*/ if ((fdCgiFile = open(cgi_file, READ, 0644)) == ERROR) { return ERROR; } cgiModuleID=moduleFindByName(cgi_file); if (cgiModuleID==NULL) { if (glbDebug==1) printf("The requested CGI module has not been loaded before, therefore need to load now.\n"); cgiModuleID=loadModule(fdCgiFile,LOAD_ALL_SYMBOLS); } /* Prepare to spawn Task for CGI work -----*/ moduleInfoGet(cgiModuleID,&cgiModuleInfo); if (Debug==1) { printf("Module Name is %s, Module Format is %i, Module Group is %i\n",cgiModuleInfo.name, cgiModuleInfo.format,cgiModuleInfo.group); printf("Text Addr=%x, Data Addr=%x, Bss Addr=%x, Text size=%i, Data Size = %i, Bss Size = %i\n", cgiModuleInfo.segInfo.textAddr,cgiModuleInfo.segInfo.dataAddr,cgiModuleInfo. segInfo.bssAddr, cgiModuleInfo.segInfo.textSize,cgiModuleInfo.segInfo.dataSize, cgiModuleInfo.segInfo.bssSize); } . . . if ((tid=taskSpawn(cgiModuleInfo.name,100, 0, 30000,(FUNCPTR)cgiModuleInfo.segInfo.textAddr, (int)doneFlag, (int)req,0,0,0,0,0,0,0,0))== ERROR) { printf("Can't Spawn task for CGI work\n"); return ERROR; } - ---------------------------------------------------------------------------- - ---------------------------------------------- Leonid Rosenboim wrote in message ... >Martin, > >You are currently using host-based tools, including debugging and dynamic >loading. >Historically this is a late addition to VxWorks from around 1996). Before >that, >VxWorks only had target-based tools, including Shell, dynamic loader, >assembly-level debugger. These target-based tools are still available, >but are disabled by default. For source level debugging one must have >the host-based tools, so if the host-based (target server) and target-based >tools are to be used concurrently, one needs to turn on the symbol table >synchronization so the two tool sets can coexist peacefully. > >I have seen many applications where target-based dynamic loading is used >in the production environment, some are for good reasons, others not. >For production, you can either have the whole app linked with VxWorks >for a single bootable (or ROM-resident) image, or have this split into >multiple modules, which are loaded subsequently. > >Run-time dynamic loading will cost you slower startup time for the >application. > >Here is an example where this can be of good use: > >Say you make an ATM device, and one of the modules implements LAN Emulation. >Trouble is there are three versions of LANE you need to support, but the >user >will need to have only one of the three versions running depending on the >site-wide >configuration. You can have three different loadable modules on the Flash >file system, >each implementing one of the three versions, but all having the same API. >Then at startup, examining the configuration you load one of the three >modules, >so the two unused modules do not waste RAM space, and all three versions >can have a common source base - loading only one of the three avoids >possible >symbol conflicts etc. > >There are other examples, so in order to help you decide weather to use >dynamic loading in production, one would need to know much more details >about your application and constraints. > >Hope this helps, > - Leonid > > > >"Martin Roth" wrote in message >news:9u2qu0$dn9$1@newshost.mot.com... >> We are now in the definition stage of our new project which will probably >> based on one of MPC8xx PPC processors. During the developement purposes >> Vxwrks allows to dynamically link & load new modules and run them from >the >> target shell. I would like to perfrm dynamic linking/loading on the final >> target (non development) sysem where I do not have the Tornado development >> tools and the WDB agent within the target. Did anybody see such (non >> development) system where application are linked and loaded dynamically. >? >> Thanks, >> Martin Roth >> >> >> >> > > --------------------------- Newsgroups: comp.os.vxworks Subject: MIB table with in a MIB table Date: 30 Nov 2001 13:25:37 -0800 From: vkinger@hotmail.com (Kinger) Organization: http://groups.google.com/ Message-ID: Does any know how to create a snmpv1 MIB table with in a MIB table --------------------------- Newsgroups: comp.os.vxworks Subject: Problem Compiling bootroms for Walnut PPC405GP Date: Fri, 30 Nov 2001 22:22:35 GMT From: "Ho-Kuo Chan" Organization: GlobeTrotter Message-ID: Hi, I am developing for the IBM Walnut borad running the PPC405GP. I have Tornado 2.0.2 installed (SPR 67953) which requires recompiling the bootrom. After compiling either bootrom.hex or bootrom_uncmp.hex, inspection of the S-Rec hex files reveals that the addresses for the data records are only 16-bits wide (S1...). Neither bootroms work. Before reprogramming the eeprom though, I saved the bootrom image that was shipped with the Eval board. Inspection of this file shows that the addresses should be 32-bits (S3...). I compiled the bootroms without any modification to the BSP as it was shipped in T2.0.2. Am I forgetting to do something? Thanks!!! - -- Ho-Kuo Chan --------------------------- Newsgroups: comp.os.vxworks Subject: tWdbTask Crashing Date: 30 Nov 2001 15:47:31 -0800 From: brycenga@wire-less-inc.com (Brock) Organization: http://groups.google.com/ Message-ID: <3d71cba2.0111301547.75bafcad@posting.google.com> Anyone know why I am getting the following errors? When I kick off my target server, I get the following error(s). Warning: Target checksum: 0xe5a0 (computed from 0x10000 to 0x4340c8). Host checksum: 0x1f91 (computed from 0x1190020 to 0x15b40e8). Warning: Core file checksums do not match. When I attempt to start a Wind Shell, I get the following error(s). 0x1c7ddc0 (tWdbTask): WDB exception. restarting agent in 10 seconds... 0x1c7ddc0 (tWdbTask): WDB exception. restarting agent in 10 seconds... I'm pretty sure the root of this problem is the checksum mismatch problem described above. I can't figure out why this is happening. There are basically 2 types of builds that I generate. One that I can load using VisionCLICK (BDX file), and the other is a compressed FTP'able image. Both use the same core file. Both were built almost simultaneously - so there wasn't any changes made between the 2. Tornado works just fine when I use it with the VisionCLICK version. When I use the FTP'able image (ftp the image, load into flash, and reboot), I get the errors listed above (after I kick off my target server and wind shell). Any suggestions would be greatly appreciated. Thanks, Brock --------------------------- Newsgroups: comp.os.vxworks Subject: Re: TFFS - flash file system Date: Sat, 01 Dec 2001 00:09:54 GMT From: Andray Kaganovsky Organization: @Home Message-ID: References: <9u2v6q$g3j$1@newshost.mot.com> <488e459a.0111281120.765483ed@posting.google.com> <4bb6bae04a%molochai@vapour-trail.demon.co.uk> <9u5ec5$i7t$1@neptunium.btinternet.com> "Leonid Rosenboim" wrote in news:newscache $3enkng$cdi$1@lnews.actcom.co.il: > > "David Laight" wrote in message > news:9u5ec5$i7t$1@neptunium.btinternet.com... >> > when is the directory entry updated to reflect the new file size? >> > Perhaps dosFs2 is smart enough to perform smaller-than-cluster >> > operations. I will check. >> >> Does the TFFS layer check whether you are writing unchanged data, or >> whether (if appropriate for the flash chips) only converting 1's to 0's - >> in which case the existing sector can be updated. > > This is most probably done in the MTD, assuming some chips can change > an arbitrary bit from "1" to "0", others (theoretically) can't. Hence the > upper > layer has to leave this to the MTD to decide. I dont't remember TFFS source > code well enough to say, but with breakpoints this is easy to verify even > without the source. > I'd say this is a "must have" feature. > - Leonid [snip] [AndrayK]. Yes, TrueFFS does that. That's what OVERWRITE flag is for (see target/h/tffs/flflash.h). --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Writing END Driver--Crashing tNetTask (Corrupt clBlk?) Date: Fri, 30 Nov 2001 18:12:55 -0600 From: "Chris Richards" Message-ID: <3c081fb4$0$19494$39cecf19@nnrp1.twtelecom.net> References: <3c043609$0$13427$39cecf19@nnrp1.twtelecom.net> Well, it seems that the first byte was being dropped which specifies the IP frame (0x45). VxWorks was happily dying if it didn't receive this 0x45 (or more specifically, it was receiving 0x00 as the start of the frame). cjr "Chris Richards" wrote in message news:3c043609$0$13427$39cecf19@nnrp1.twtelecom.net... > Howdy. > > I'm crashing tNetTask when I call > > END_RCV_RTN_CALL(&pDrvCtrl->end, pMblk); > > The first clblk in the mblk looks correct. However > it's next clblk pointer is garbage. My guess is > that this is what is causing the problem. Yet, > I can't find out why the clblk next pointer is > garbage. > > Could the be a result of improperly setting up > the net pool? > > Thanks. > > cjr > > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: processor migration (MVME230X -> MVME5100 ALTIVEC) Date: Sat, 01 Dec 2001 00:15:05 GMT From: Andray Kaganovsky Organization: @Home Message-ID: References: "Stephen" wrote in news:vzRN7.630$h7.21374 @llslave.llan.ll.mit.edu: > > We are using VxWorks on MOTO MVME2300 boards. > We are interested in adding a motorola 5100 ALTIVEC card to an existing VME > chassis (i.e. with teh 2300 boards) > Wind River is saying that in order to support this we need to buy a whole > new development package !! > We've moved to differnet processor before and had to simply pay for the new > BSP. I can see paying a bit extra for ALTIVEC libarraies or something liek > that but a whole new dev system does not seem correct. > > I am wondeing if anyone else is facing the same issue. Has anyone migarted > from MVME2300 to 5100 ALTIVEC boards ? > > Thanks for any help > Stephen Rejto > [AndrayK]. We have exactly the same situation here - we can't simply get 5100 BSP and plug it into existing Tornado/ppc installation - we have to buy the whole new Tornado seat. Andray --------------------------- Newsgroups: comp.os.vxworks Subject: Debugging IP Problems--How? Date: Fri, 30 Nov 2001 18:14:35 -0600 From: "Chris Richards" Message-ID: <3c082019$0$10483$39cecf19@nnrp1.twtelecom.net> Howdy. My driver sends IP into the VxWorks stack, yet nothing ever comes out of it (e.g., telnet sessions, pings, etc.). How can I tell if the data was sent up successfully, and then determine what the stack is doing with it. Thanks. /cjr --------------------------- Newsgroups: comp.os.vxworks Subject: FS VME Elma cage w Heurikon V960D (8) Date: 30 Nov 2001 18:38:04 -0800 From: tonyS2@aol.com (tony) Organization: http://groups.google.com/ Message-ID: <61e40296.0111301838.7d99eb32@posting.google.com> Cage with 8 V960D Price for quick turnover Tony Alex Va --------------------------- Newsgroups: comp.os.vxworks Subject: iov_len limit in sendmsg call Date: Fri, 30 Nov 2001 19:17:16 -0800 From: "Matthew J. Weaver" Organization: Stanford Linear Accelerator Center Message-ID: <3C084BBC.9B9EC1F6@slac.stanford.edu> Why is the limit on iov_len set at 63 for calls to the sendmsg() routine in VxWorks? Can I change it? This is for a UDP socket . --------------------------- Newsgroups: comp.os.vxworks Subject: Bandwidth Management vs. Qos Date: 1 Dec 2001 01:38:29 -0800 From: jlsnmp@netease.com (jlsnmp) Organization: http://groups.google.com/ Message-ID: <2941895e.0112010138.d68ca67@posting.google.com> Hi, every guru: I am busy in developing a PPPoE AC which needs bandwidth management feature that provides independent traffic shaping for 1000s of connections, multiple QoS Classes including CBR, VBR-rt, VBR-nrt, UBR, and CIR, rate limiting per PPP session. How to implement it??? Under Vxworks, is there some solution or algorithm to reach the above goal? Any source code or commercial protocol stack is help to me. Welcome to any suggestions! --------------------------- Newsgroups: comp.os.vxworks Subject: Need a driver for MB86964 Ethernet Date: Sat, 1 Dec 2001 11:06:55 +0100 From: "Frank Wolf" Message-ID: Hi there! Is there anyone out there who has an END driver for a Fujitsu MB86964 controller which he can share? The MB86960 driver included in VxWorks is not compatible and the driver I already have is not working flawlessly... Thanks Frank --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Sun Dec 2 04:03:06 2001 From: Vxworks Exploder Date: Sun Dec 2 04:03:09 PST 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sun Dec 2 04:03:04 PST 2001 Subject: Re: Images larger than 1 Mb dowloaded on a Force PPC target Subject: Re: Dynamic loading Subject: Re: GUI display on top of VxWorks ? Subject: Re: vxWorks SMP? ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Images larger than 1 Mb dowloaded on a Force PPC target Date: 28 Nov 2001 15:39:34 +0200 From: Reinhold Wirth Organization: WEB.DE Message-ID: <3c04f726$1@netnews.web.de> References: Santiago.Vazquez@casa.eads.net (santiago vazquez) wrote: >Hello to everybody > >I have a problem downloading vxworks images on a PPC Force target when >they are bigger than 1 Mb. >Does anybody know how to do it? Do I have to burn a new ROM? > >Thanks in advance. Hi Santiago, do you use the default settings: RAM_LOW_ADRS = 00100000 RAM_HIGH_ADRS = 00200000 (RAM_HIGH_ADRS - RAM_LOW_ADRS --> 1 MB) If yes: * change RAM_HIGH_ADRS in config.h and in the Makefile * make and flash the new bootloader - -- __________________________________________________________ News suchen, lesen, schreiben mit http://newsgroups.web.de --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Dynamic loading Date: 28 Nov 2001 16:17:49 +0200 From: Reinhold Wirth Organization: WEB.DE Message-ID: <3c05001e@netnews.web.de> References: <9u2qu0$dn9$1@newshost.mot.com> "Martin Roth" wrote: >We are now in the definition stage of our new project which will probably >based on one of MPC8xx PPC processors. During the developement purposes >Vxwrks allows to dynamically link & load new modules and run them from the >target shell. I would like to perfrm dynamic linking/loading on the final >target (non development) sysem where I do not have the Tornado development >tools and the WDB agent within the target. Did anybody see such (non >development) system where application are linked and loaded dynamically. ? Hi Martin, you do not need Tornado or WDB to load and start a applikation. It is possible to load and start it from the target shell. You only have to type the commands. The Startup-Script is a possibility for the input this commands from a device (like the autoexec.bat). (you have to define INCLUDE_STARTUP_SCRIPT in VxWorks) One other possibility may be: load and start your application in usrAppInit.c - -- __________________________________________________________ News suchen, lesen, schreiben mit http://newsgroups.web.de --------------------------- Newsgroups: comp.os.vxworks Subject: Re: GUI display on top of VxWorks ? Date: Sun, 02 Dec 2001 03:42:51 GMT From: Ephraim Gadsby Organization: - minimal Message-ID: <7o6j0usnok0825uqiuk9lm1t9bmg4dqto1@4ax.com> References: On 29 Nov 2001 23:55:57 GMT, Chris Smith wrote: >Hello, > >I'm wondering if someone can tell me of some products out there that >would let me easily build a GUI frontend on top of VxWorks application. >I'm developing on a PPC board and want to plot and display some program >variables as the program is running. Why do you need a GUI for that? It's a sledgehammer to crack a nut. If you can't use the Tornado tools, just use the console shell. Any tool with a serial/telnet interface can then get the data, perl, visual-basic, probably even Word or Excel. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: vxWorks SMP? Date: Sun, 02 Dec 2001 07:37:42 GMT From: "John" Organization: Excite@Home - The Leader in Broadband http://home.com/faster Message-ID: Hi, Can anyone tell me if vxWorks supports SMP or not? Cheers John --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Sun Dec 2 17:01:38 2001 From: "Sinn, Larry" Date: Sun Dec 2 17:01:41 PST 2001 Subject: RE: processor migration (MVME230X -> MVME5100 ALTIVEC) Stephen: From the sounds of things, WindRiver must be rolling in money to be so petty. Times are hard and WRS must realize they have to give a little. (But maybe not, they seem to enjoy pissing off customers) You could also try talking to the regional manager and see if that helps. This is also year end and they may be willing to deal to make quota. If all of the above fails, try telling them you are using the PowerPC 750 version of the 5100, it's the same processor as the 2300, but the BSP will work for either the 7400 (Altivec) or the 750. Another thought is to threaten to move to LINUX, then they will also loose the runtime fees. Larry. Stephen wrote: We are using VxWorks on MOTO MVME2300 boards. We are interested in adding a motorola 5100 ALTIVEC card to an existing VME chassis (i.e. with teh 2300 boards) Wind River is saying that in order to support this we need to buy a whole new development package !! We've moved to differnet processor before and had to simply pay for the new BSP. I can see paying a bit extra for ALTIVEC libarraies or something liek that but a whole new dev system does not seem correct. I am wondeing if anyone else is facing the same issue. Has anyone migarted from MVME2300 to 5100 ALTIVEC boards ? Thanks for any help Stephen Rejto Larry Sinn KLA-Tencor MS I-1009 408 875 0247 voice 160 Rio Robles 408 875 6179 fax San Jose, Ca 95134-1809 larry.sinn@kla-tencor.com From vxwexplo-errs@csg.lbl.gov Sun Dec 2 20:43:59 2001 From: "Avkash Chauhan" Date: Sun Dec 2 20:44:01 PST 2001 Subject: question related with vxSim and Tornado 2.1 for Hitachi SuperH Hi all, I have Tornado 2.1 for Hitachi SuperH. It has no build-in simulator in it. I was writing some program for SH7727 based BSP. As default there is no simulator support for it. As, i was digging into manuals, I found that I can use vxSim to simulate my project test. my questions are as below: 1. can i use vxSim for BSP specific projects? 2. How i can seperately install vxSim to work with Tornado 2.1 for SH? 3. Is it possilve to port BSP specific vxWork image to vxSim Running on different machine. Note - I tried to install vxSim using prototyper 2.0.2 along with tornado 2.1, i could not run simulator. It meshed up my application.. Any help would be greatly appreciated on this regard. Thanks a lot. - avkash From vxwexplo-errs@csg.lbl.gov Mon Dec 3 00:13:13 2001 From: Shlomo Katz Date: Mon Dec 3 00:13:15 PST 2001 Subject: error booting vxworks file hi i am trying to boot my target and load the vxworks file from my pc. i get the following error, " arpresolve: can't allocate llinfo0x385f00 (tNetTask) " does it ring a bell to somebody??? katz sholmi embedded team www.rit.co.il From vxwexplo-errs@csg.lbl.gov Mon Dec 3 01:05:16 2001 From: Ilia Bosis Date: Mon Dec 3 01:05:18 PST 2001 Subject: Re Re : what's the reason which will result in workQPanic? Newsgroups: comp.os.vxworks Subject: Re: what's the reason which will result in workQPanic? Date: 30 Nov 2001 09:12:23 -0800 From: marsgal42@hotmail.com (Laura Halliday) Message-ID: References: <221caa76.0111290329.12f02cb2@posting.google.com> kevinatnj@yahoo.com.cn (kewin) wrote in message news:<221caa76.0111290329.12f02cb2@posting.google.com>... "workQPanic..." when working with LAN ? Sounds familiar. It frequently happens when the network driver calls netJobAdd() on EACH Ethernet frame received. It looks that the netJobAdd() processing overhead (including the subsequent task switch to the tNetTask) is too heavy to be used liberally in the driver land, where each CPU cycle counts. The classic solution (assuming interrupt-driven operation) is: -- In the ISR do only the necessary minimum of the Rx event processing; -- Call the netJobAdd() to defer most of the processing to the task level; -- Disable interrupts from the Ethernet controller hardware. -- In the task environment do most of the processing. Take into account that there may be the number of frames waiting for processing, as the Ethernet interrupts are disabled now. -- Enable the Ethernet interrupts when processing is done. That way You keep the ISR-level processing and the netJobAdd() overhead at the bare minimum, thus avoiding interrupt lockout for another interrupt sources (the clock interrupt is the first possible victim). Remember that in the VxWorks ISR is uninterruptible. The interrupt controller HW is configured in such a way that when system is in an ISR, all interrupts are pending until the currently working ISR finishes. Of course, the solution above implies that the Ethernet controller has some DMA capabilities and may receive a number of frames and place them in memory without the CPU intervention. If Your hardware is incapable of this - ask for help from the HW team, as in that case it will be probably too hard work for Your CPU, no matter how hard You try to optimize the SW. Best regards Ilia. -------------------------------------------- Ilia Bosis Software Engineer PacketLight Networks Tel: +972-9-7644228 Fax: +972-9-7645444 E-mail: Ilia_Bosis@packetlight.com Web: www.packetlight.com > > Hi > > > when i can debug ARM7TDMI bsp a problem met with me.My ethernet port > > driver process packet by the way of POLL. When I hook my board to our > > company local LAN,I got the following trouble,"workQPanic kernel work > > queue overflow".I don't know what the reasons will result in this kind > > of problem. > > Any advice will be welcome > > > Wind River have an FAQ on the subject. Ask them for it. > > Briefly, when VxWorks is in kernel mode (from interrupt > handlers, executing system calls, and so), it defers > clock processing. Timer ticks that occur during kernel > mode get saved in an internal data structure, and are > then processed on exit from kernel mode. Should this > data structure overflow, the result is a workQPanic() > crash. > > Polling a network interface?! > > > Laura Halliday VE7LDH "Que les nuages soient notre > Grid: CN89lg pied a terre..." > ICBM: 49 16.57 N 123 0.24 W - Hospital/Shafte From vxwexplo-errs@csg.lbl.gov Mon Dec 3 02:59:42 2001 From: "H.-P. Kabout" Date: Mon Dec 3 02:59:44 PST 2001 Subject: RE: comp.os.vxworks newsdigest You cannot create a table that is part of an other table. What you can do is creating two tables where: - The 1st (outer) has one index. - The 2nd (inner) has two indexes, one of them the index of 1st table. - In the ASN1 definition the tables need to be defined like in the example below. my1stTable OBJECT-TYPE SYNTAX SEQUENCE OF my1stEntry ACCESS not-accessible STATUS mandatory DESCRIPTION "The outer table." ::= { myTree 1 } my1stEntry OBJECT-TYPE SYNTAX My1stEntry ACCESS not-accessible STATUS mandatory DESCRIPTION "The record set of one outer entry." INDEX { my1stIndex } ::= { my1stTable 1 } My1stEntry ::= SEQUENCE { my1stIndex INTEGER, my1stParameter INTEGER } my1stIndex OBJECT-TYPE SYNTAX INTEGER (1..64) ACCESS read-only STATUS mandatory DESCRIPTION "This object identifies the ... for which my1stEntry contains information." ::= { my1stEntry 1 } my1stParameter OBJECT-TYPE SYNTAX INTEGER ACCESS read-write STATUS mandatory DESCRIPTION "..." ::= { my1stEntry 2 } my2ndTable OBJECT-TYPE SYNTAX SEQUENCE OF my2ndEntry ACCESS not-accessible STATUS mandatory DESCRIPTION "The inner table." ::= { myTree 1 } my2ndEntry OBJECT-TYPE SYNTAX My2ndEntry ACCESS not-accessible STATUS mandatory DESCRIPTION "The record set of one inner entry." INDEX { my1stIndex, my2ndIndex } ::= { my2ndTable 1 } My2ndEntry ::= SEQUENCE { my2ndIndex INTEGER, my2ndParameter INTEGER } my2ndIndex OBJECT-TYPE SYNTAX INTEGER (1..128) ACCESS read-only STATUS mandatory DESCRIPTION "This object identifies the ... for which my2ndEntry contains information." ::= { my2ndEntry 1 } my2ndParameter OBJECT-TYPE SYNTAX INTEGER ACCESS read-write STATUS mandatory DESCRIPTION "..." ::= { my2ndEntry 2 } Regards, H.-P. Kabout, IT-Consulting e-mail: Kai.Kabout@freenet.de Phone: +49 (0)9085 - 1017 Mobile: +49 (0)160 - 7556913 http://www.kabout.com (under construction) -----Original Message----- From: the vxWorks Users Group Exploder [mailto:vxwexplo@lbl.gov] Sent: Saturday, December 01, 2001 13:12 To: vxworks_users@csg.lbl.gov Subject: comp.os.vxworks newsdigest Newsgroups: comp.os.vxworks Subject: MIB table with in a MIB table Date: 30 Nov 2001 13:25:37 -0800 From: vkinger@hotmail.com (Kinger) Organization: http://groups.google.com/ Message-ID: Does any know how to create a snmpv1 MIB table with in a MIB table From vxwexplo-errs@csg.lbl.gov Mon Dec 3 03:39:51 2001 From: "umesh" Date: Mon Dec 3 03:39:54 PST 2001 Subject: Re: error booting vxworks file Initialized value for Network stack system memory may be small. Stack memory pool has two parts. One for system pool (data structures for sockets) and other data pool ( copy data from task buffers to n/w buffers). Increase NUM_FILES. Select Network components and then networking protocols. Right click on N/wing protocols. Select properties and increase NUM_FILES. It should work fine. Umesh S From vxwexplo-errs@csg.lbl.gov Mon Dec 3 04:03:10 2001 From: Vxworks Exploder Date: Mon Dec 3 04:03:12 PST 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Mon Dec 3 04:03:07 PST 2001 Subject: Re: Why is the IP stack ARPing my non-Ethernet interface? Subject: Re: TRUE definition Subject: Need a driver for a Philips part SC28LA1B Subject: Re: GUI display on top of VxWorks ? Subject: The Best Prepaid Phone Card Yet Subject: What could cause ethernet collision? Subject: Adding a custom routine into a bootrom image Subject: Re: IP address Subject: PCI driver source sample Subject: Re: What could cause ethernet collision? Subject: 64 bit PCI Requests Subject: Undefined reference in STL Subject: Re: Undefined reference in STL ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Why is the IP stack ARPing my non-Ethernet interface? Date: 2 Dec 2001 09:01:02 -0800 From: jperron@cetaceannetworks.com (Jason Perron) Organization: http://groups.google.com/ Message-ID: References: <3C0726B4.4DE481A1@rogers.com> Hi Todd and Rich, Thank you very much for your replies. Both were very helpful. From what you have both said, it sounds like I have to either put in some ARP fake-out code in my driver's send routine or upgrade to the newer NPT stuff. I think I'd like to upgrade to the NPT stuff, but as it's fairly new and a pretty big overhaul of the networking code, I'm weary of what kind of new, unknown problems may come with it. Aside from the NPT programmer's guide (version 4), I have not seen any sample NPT driver code, nor do I see many people talking about NPT on this group. Is this because nobody is using it, or because it's fairly stable and works as stated in the programmer's guide? Todd (or for that matter, anybody else), what has your experience with the NPT stuff been? Was it pretty smooth sailing, or did you run into other significant gotchas (like on the level of having to hack around stuff that doesn't work as stated in the guide) that you had to deal with? Thank you again for your replies. Jason Todd Sandor wrote in message news:<3C0726B4.4DE481A1@rogers.com>... > With the newer vxWorks releases what matters is if the interface flag > specifies broadcast or pointtopoint - you set this in your MUX driver. > We're using Tornado 2.02 [NPT is integrated with T2.02] with patch 68322 > [I believe that is the number - its the stack jumbo patch for the T2.02 > release] and we developed a pointtopoint END driver that does not perform > arps nor do we need a resolution function [there is nothing to resolve > with a point to point interface/link]. What is important is whether > your using a broadcast media and if a layer 2 address needs to be resolved > - if no layer 2 address needs to be resolved you could use a NPT type NPT > driver - take another look at the NPT documentation is sections related to > address resolution. > > VxWork release prior to 2.02 do, as Rich Williams specified, used the > interface type, not the flags, to determine how to handle the device. > > Hope this helps... > > Jason Perron wrote: > > > Hi Everybody, > > > > I am using vxWorks version 5.4 and I'm trying to get an END driver > > going for a non-ethernet interface. My problem is that the IP stack > > is sending me ARP packets even though I have created a non-Ethernet > > interface. > > > > I specify my interface type as "M2_ifType_other" in my END_MIB_INIT > > call. > > > > First I tried setting the "IFF_NOARP" flag for the interface and my > > driver still got Ethernet ARP packets. Next, I tried registering my > > own arp resolve function using muxAddrResFuncAdd with the idea that > > I'd just "fake out" the IP stack. That didn't work either, as my > > "fake" arp resolve function never gets called. > > > > Has anybody else come across this problem? If so, how did you get > > around it? If this is a vxWorks problem, does anybody know if this is > > fixed in the NPT stuff? > > > > TIA, > > > > Jason --------------------------- Newsgroups: comp.os.vxworks Subject: Re: TRUE definition Date: Thu, 29 Nov 2001 05:10:55 -0500 From: George Varndell Organization: Posted Via Binaries.net = SPEED+RETENTION+COMPLETION = http://www.binaries.net Message-ID: <3C0609AF.9D9A9053@varndellengineering.com> References: <9u2qjq$smr$1@plutonium.btinternet.com> <3c05e7fd$0$225$4d4ebb8e@oce.news.eu.uu.net> Manuel van den Berg wrote: > > "David Laight" wrote in message > news:9u2qjq$smr$1@plutonium.btinternet.com... > > > > > >I would like to change the definition of TRUE from 1 to -1 > > > > > > NO. Don't do that. > > > > Agreed... > > > > Actually you would be better off removing 'TRUE' and 'FALSE' (and 'OK', > > 'ERROR' and 'NULL') from the entire source tree! > > > > The come into the same category as 'FOREVER' and (maybe) 'FUNCPTR', all > > they do is obfuscate bits of the language and force people to scan header > > files for definitions. > > Yup, I'd suggest not to use structs, typedef's, defines, functions, etc ... > hey wait maybe we shouldn't be using any header files at all ;-) > > Any clue of the meaning of 'abstraction' or 'encapsulation' ? Any clue of the meaning of 'clarity' or 'unambiguous'? The validity of David's basic position is well demonstrated and supported by the this whole thread. What we have here is a person who is wrapped around the axle over the numeric values of TRUE and FALSE, and he doesn't need to be. If you go back and read Ofer's original posted question, it's clear he is trying to implement a ethernet input filter hook and it isn't working. He's decided that the numeric values of TRUE and/or FALSE somehow play a part in this failure. I don't doubt it, but the problem is not likely in the network driver, as Ofer implies. The problem is that poor Ofer, having been exposed to all the vulgar macro abuses one commonly finds in vxworks code ('!=TRUE','==FALSE','!=FALSE'), doesn't know what to return from his filter hook function. I've looked at several netif drivers. All will work as expected if the hook function returns a non-zero value to indicate that the packet should be discarded, and zero to indicate that the packet should be processed normally. The C language guarantees that built-in operators such as ==, <, >=, etc., will produce a value of 0 or 1, and that complied code will treat as 0 as false (lower case intentional) and 1 as true. As such, there is no ambiguity about whether or not 'if(0)printf("hello");' will print "hello". It will not. The C language does not guarantee the value of the token 'TRUE' to be anything nor does it promise that the token 'FALSE' means anything in particular. As such, whether or not 'if(FALSE)printf("hello");' prints "hello" depends on the value assigned to FALSE. if(a == b) is unambiguous and NOT in need of any abstraction if((a == b) == TRUE) is ambiguous and pointless it can only be semantically correct if TRUE == 1. if(a) foo(); is unambiguous in that any non-zero value will cause foo() to be called. if(a != FALSE) foo(); is ambiguous. Until I verify the value of FALSE, in the current scope, I can't know what this does. Ofer, return '0' if you don't want a packet thrown out. Return '1' if you do. See if the filtering works. If it doesn't, your network driver is broken. Regards, George Varndell - -- Varndell Engineering, LLC. PPC and XScale Expertise Coming soon! Download free demo versions of our END drivers! http://www.varndellengineering.com ______________________________________________________________________________ Posted Via Binaries.net = SPEED+RETENTION+COMPLETION = http://www.binaries.net --------------------------- Newsgroups: comp.os.vxworks Subject: Need a driver for a Philips part SC28LA1B Date: 29 Nov 2001 15:48:15 -0600 From: "john cobarruvias" Message-ID: <3c06ad1f$1_2@newsa.ev1.net> Reply-To: "john cobarruvias" Sender: "john cobarruvias" I am looking for a driver for a philips part # SC28L92A1B 3.3V-5.0V Dual Universal Asynchronous Receiver/Transmitter (DUART. Can anyone help on this? johncoby@ev1.net John Cobarruvias --------------------------- Newsgroups: comp.os.vxworks Subject: Re: GUI display on top of VxWorks ? Date: Sun, 2 Dec 2001 21:37:14 +0100 From: "Martin Raabe" Organization: T-Online Message-ID: <9ue3cp$vpv$07$1@news.t-online.com> References: Hello Chris, how about the Scope Tools for Tornado, which can display variables while the programm runns? Ciao Martin Raabe "Chris Smith" schrieb im Newsbeitrag news:MPG.16706c04bc584b559897ed@news.cncdsl.com... > Hello, > > I'm wondering if someone can tell me of some products out there that > would let me easily build a GUI frontend on top of VxWorks application. > I'm developing on a PPC board and want to plot and display some program > variables as the program is running. I've come across Tilcon which > looks like what I'm looking for. Are there any other similar products > out there to consider. > > TIA --------------------------- Newsgroups: comp.os.vxworks Subject: The Best Prepaid Phone Card Yet Date: Sat, 01 Dec 2001 04:10:41 GMT From: go@prepaidonline.com Message-ID: <5LYN7.16061$cU4.79032@typhoon.mw.mediaone.net> UNLIMITED domestic calling to ANYWHERE in the U.S. NO Taxes, NO Junk, Just Unlimited Talk for 30 days! Service can be used 24 hours a day, 7 days a week, from the same phone! $89.95 DOUBLE TALK - Two phone cards in one - a connection fee card, great for longer calls, or a no-connection fee card, better for shorter calls. From $.02 to $.10 per min. WorldWide! Features low rates to the U.S., Europe and Mexico! $10.00 The CLEAN CARD - Talk for a minute - pay for a minute. No hidden fees or connection fees. Great rates to many international destinations, as well as a .065 rate for the U.S.! $10.00 GO TO: http://go.prepaidonline.com/ Lowest Prepaid Phone Cards, Inc. is a recognized leader in guaranteed prepaid long distance phone service. We are a leading world-wide supplier of prepaid telecommunication services and by working in partnership with the largest telecommunication and network services companies, PrepaidOnline is able to consistently offer the lowest cost and highest quality cards available. GO TO: http://go.prepaidonline.com/ --------------------------- Newsgroups: comp.os.vxworks Subject: What could cause ethernet collision? Date: Mon, 03 Dec 2001 05:15:15 GMT From: "Di Mao" Organization: Prodigy Internet http://www.prodigy.com Message-ID: Hi, I am porting scc2 ethernet driver of windriver ads8xx bsp to our customed target board. The cpu is MPC850DE. When I pinged from another node to my board under the same subnet, the Rx path of my board received the packets and it also replied the ping packets. The Tx path of board has some problem. All the ethernet packets sent out from my board caused collision at the ethernet hub right away (I can see it from ethernet hub collision led). We verified the Tx and Rx clocks seem to be correct, 10MHz. I also compared the registers between a working motorola ads850 board and my target board. I don't see any major difference. They use actually the same ethernet driver. But both ethernet TX and RX paths work well on ads850 board. Does anyone know where the problem could be? Thanks, Davis --------------------------- Newsgroups: comp.os.vxworks Subject: Adding a custom routine into a bootrom image Date: Mon, 03 Dec 2001 05:45:18 GMT From: "Jae-Hwa Lee" Organization: Dacom Message-ID: Hi, everyone? I'm trying to add a custom routine for a ftp server into a bootrom image. All source codes for the ftp server are in ftp.a Some routines in ftp.a are referred to other libraries such as libPPCEC603gnuvx.a, rtlib.a and so on. Except libPPCEC603gnuvx.a, all libraries are custom made. In order to compile the bootrom image with the custom ftp routine, I added the followings into the bsp 'Makefile'. /*************************************************************************** ************ **************************************************************************** ************/ LIBS = C:/AX2200/Lib/ftp.a C:/AX2200/Lib/main.a C:/AX2200/Lib/cmd_proc.a \ C:/AX2200/Lib/rtlib.a C:/AX2200/Lib/cmp.a C:/AX2200/Lib/n2p.a C:/AX2200/Lib/n2p_vp.a \ C:/AX2200/Lib/isdn.a C:/AX2200/Lib/lapd.a C:/AX2200/Lib/dsp.a C:/AX2200/Lib/trcp.a \ C:/AX2200/Lib/phy.a C:/AX2200/Lib/ft1.a C:/AX2200/Lib/diag.a C:/AX2200/Lib/vocfile.a \ C:/AX2200/Lib/nutil.a \ C:/AX2200/LIB/libPPCEC603gnuvx.a /*************************************************************************** ************ **************************************************************************** ************/ However, there happened some make errors, when I executed the 'Makefile' The errors are following /*************************************************************************** ************ **************************************************************************** ************/ C:\ax2200\Bsp>make bootrom_uncmp.hex ccppc -BC:\Tornado/host/x86-win32/lib/gcc-lib/ -c -mstrict-align -ansi -nost dinc -O2 -fvolatile -fno - -builtin -fno-for-scope -Wall -I/h -I. -IC:\Tornado\target\config\all -IC: \Tornado\target/h -IC:\T ornado\target/src/config -IC:\Tornado\target/src/drv -DCPU=PPCEC603 -msof t-float -o version.o C:\ Tornado\target\config\all/version.c ldppc -X -N -Map Bsp.map -e _romInit -Ttext 01700000 \ -o bootrom_uncmp romInit.o bootInit_uncmp.o version.o bootConfig.o \ sysALib.o sysLib.o C:/AX2200/Lib/ftp.a C:/AX2200/Lib/main.a C:/AX2200/Lib/cmd_proc.a C:/AX2200 /Lib/rtlib.a C:/AX2200/Lib/cmp.a C:/AX2200/Lib/n2p.a C:/AX2200/Lib/n2p_vp.a C:/AX2200/Lib/isdn.a C:/ AX2200/Lib/lapd.a C:/AX2200/Lib/dsp.a C:/AX2200/Lib/trcp.a C:/AX2200/Lib/phy.a C:/AX2200/Lib/ft1.a C :/AX2200/Lib/diag.a C:/AX2200/Lib/vocfile.a C:/AX2200/Lib/nutil.a C:/AX2200/LIB/libPPCEC603gnuvx.a C:/AX2200/Lib/main.a(cfgcmd.o): In function `Cmd_SetIpAddr_Lan': cfgcmd.o(.text+0x5a40): undefined reference to `NV_FTP_VER_load_from_nvram' cfgcmd.o(.text+0x6230): undefined reference to `NV_FTP_VER_save_to_nvram' make: *** [bootrom_uncmp] Error 0x1 /*************************************************************************** ************ **************************************************************************** ************/ In fact, the codes for the ftp server was designed for VxWorks image and they were already verified. I guess the reason why the ftp server code is not compiled is that the BSP makefile does not make a partialImage.o. According to the 'Makefile' for the VxWorks image, it generates a partialImage.o with required libraries and objects and then a symbol table(symTbl.o) is made by the partialImage.o. /*************************************************************************** ************ **************************************************************************** ************/ partialImage.o: $(LDDEPS) $(PRJ_OBJS) $(EXTRA_MODULES) $(COMPONENT_LIBS) $(LIBS) $(PRJ_OBJS_FILE) - @ $(RM) $@ $(CC) $(CFLAGS) -c $(VERSION_C) ifneq ($(findstring vxWorks_rom, $(DEFAULT_RULE)),) $(CC) $(CFLAGS) -c $(USR_ENTRY_C) $(LD_PARTIAL) \ usrEntry.o $(PRJ_OBJS_FOR_LD_PARTIAL) $(EXTRA_MODULES) version.o $(COMPONENT_LIBS) $(LIBS) -o $@ else $(LD_PARTIAL) \ $(PRJ_OBJS_FOR_LD_PARTIAL) $(EXTRA_MODULES) version.o $(COMPONENT_LIBS) $(LIBS) -o $@ endif ifeq ($(REMOVE_DEADWEIGHT),TRUE) $(TRIM_TOOL) $@ $@_trim $(MV) $@_trim $@ endif symTbl.o: partialImage.o - @ $(RM) $@ $(CP) partialImage.o tmp.o $(LDOUT_SYMS) tmp.o $(MKSYMTBL) tmp.o > symTbl.c - @ $(RM) tmp.o $(COMPILE_SYMTBL) symTbl.c -o $@ /*************************************************************************** ************ **************************************************************************** ************/ Since the symTbl.o includes references of all routines based on partialImage.o, all routines can be linked without reference errors. In the BSP makefile to make a bootrom image, however, there is no command line to make the partialImage.o, even symTbl.o. I think that's the problem... I'm not sure... Anyway what I wanna know is how I can insert a custom routine in the BSP codes and then make a bootrom image without undefined reference errors. Regards, Jay Lee --------------------------- Newsgroups: comp.os.vxworks Subject: Re: IP address Date: Mon, 3 Dec 2001 08:43:53 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <9u3j5t$1ov$1@newsreader.mailgate.org> I would like to add to the previous answers: Sometimes, when your application needs to know an interface-specific IP address (keep in mind that every IP node has at least twp IP addresses, the one you configure, and 127.0.0.1 which is allways present), ifAddrGet() is the VxWorks laternative to Unix ifconfig utility. But more often, you have a TCP socket application, that needs to know the local IP address, which was used by the peer to connect to this node, in which case getsockname() is the right function to use. (this is generic on all socket implementations) So this is all a matter of WHY. - - Leonid "Tim Moeller" wrote in message news:9u3j5t$1ov$1@newsreader.mailgate.org... > Sorry, I know this is a dumb question, but I can't seem to find the answer > in my documentation. > > How do you programmatically determine the IP address being used on the > target processor? > > The IP address is entered and stored in NVRAM by the BOOT ROM setup on my > board (Walnut PPC 405GP, if that makes a difference.) Isn't there some > kernel function that returns My IP Address? > > Be kind. :-) > > Thanks, > -- Tim > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: PCI driver source sample Date: Mon, 3 Dec 2001 10:33:12 +0100 From: "Joerg Engel" Message-ID: <3c0b46c0$0$186$4d4ebb8e@read.news.de.uu.net> I'm in the process of working on a VxWorks driver for a Digital Interface for bi-directional data transfers. We are using the bus master interface chip PLX 9056 as DMA controller. I need some sample code, that I can use as a framework for developing my own driver. Where can I find a sample driver, to start from ? Has anyone used this or a similar chipset or developed VxWorks drivers for I/O cards using this chipset? If so could you give me some pointers on the correct procedure for accessing and configuring the registers defined, specifically the DMA registers. Or if you have source code available from your development, that would also be appreciated. Thanks Joerg --------------------------- Newsgroups: comp.os.vxworks Subject: Re: What could cause ethernet collision? Date: Mon, 03 Dec 2001 10:07:03 +0000 From: Graham Baxter Organization: Graham Baxter (Software) Limited Message-ID: <3C0B4EC7.AF5F0AF1@NOSPAM.bcs.org.uk> References: Davis, But are the 2 clocks the right way round. I was caught out on the Motorola 8260 BSP when our custom design had swapped the TX and RX clocks. I hope this helps. - -- Regards, G Baxter Freelance Software Engineer (VxWorks and pSOS BSPs) gbaxter@NOSPAM.bcs.org.uk --------------------------- Newsgroups: comp.os.vxworks Subject: 64 bit PCI Requests Date: Mon, 3 Dec 2001 10:45:21 +0000 (UTC) From: "Raghunath Adhyapak" Organization: Mailgate.ORG Server - http://www.Mailgate.ORG Message-ID: Hi, I'm currently working on Ocelot board from Momentum Computers and MIPS processor. We have a 64 bit device and which is not backward compatible to 32 bit PCI. i.e. It doesn't respond to 32 bit PCI requests. It can only respond to 64 bit PCI Requests. Device memory is available using BAR's from PCI Configuration Space. I want to write to any register in the device using this Base Address. But the system controller is not generating 64 bit PCI requests. It generates 32 bit requests, which is not handled by the device. Can anyone guide me how to make the system controller to generate 64 bit requests. System Controller on Ocelot board is Galileo GT64120 Thanks & Regards, Raghunath (O) 91-832-883601 ext 505 (R) 91-832-423024 - -- Posted from hidden-user@dakiya.controlnet.co.in.116.54.202.in-addr.arpa [202.54.116.69] (may be forged) via Mailgate.ORG Server - http://www.Mailgate.ORG --------------------------- Newsgroups: comp.os.vxworks Subject: Undefined reference in STL Date: 3 Dec 2001 04:50:27 -0600 From: "AE" Message-ID: <3c09d5f9$1@news.barak.net.il> Hi I'm trying to compile the following code in Tornado 2.02 // ******************************************* #include template std::list; int InitApplic() { std::list myList; return 0; } // ******************************************* I'm compiling with -frepo. Unfortunately I get the following message during link ldsimpc --subsystem=windows -e _WinMainCRTStartup C:\Tornado\host\x86-win32\i386-pc-mingw3 2\lib\crt1.o \ dataSegPad.o partialImage.o ctdt.o -o vxWorks partialImage.o: In function `_tf9type_info': \wpwr\host\src\gnu.simpc.cpp\gcc\cp\tinfo2.cc(.text+0x2cb4): undefined reference to `void list >::insert<__list_iterator >(__list_iterator, __list_iterator, __list_iterator)' \wpwr\host\src\gnu.simpc.cpp\gcc\cp\tinfo2.cc(.text+0x39c2): undefined reference to `void list >::insert<__list_iterator >(__list_iterator, __list_iterator, __list_iterator)' make: *** [vxWorks] Error 0x1 Can someone help? Thanks --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Undefined reference in STL Date: Mon, 03 Dec 2001 06:35:44 -0500 From: joe durusau Organization: Bellsouth.Net Message-ID: <3C0B6390.762CD77C@bellsouth.net> References: <3c09d5f9$1@news.barak.net.il> This is only a guess, but do you have a function named main in your code? You should never use main as a function name in vxWorks. I ask, because it is not mormal to link in crt1.o in vxWorks. Speaking only for myself, Joe Durusau AE wrote: > Hi > > I'm trying to compile the following code in Tornado 2.02 > > // ******************************************* > #include > > template std::list; > > int InitApplic() > { > std::list myList; > > return 0; > } > // ******************************************* > > I'm compiling with -frepo. > > Unfortunately I get the following message during link > > ldsimpc --subsystem=windows -e _WinMainCRTStartup > C:\Tornado\host\x86-win32\i386-pc-mingw3 > 2\lib\crt1.o \ > dataSegPad.o partialImage.o ctdt.o -o vxWorks > partialImage.o: In function `_tf9type_info': > \wpwr\host\src\gnu.simpc.cpp\gcc\cp\tinfo2.cc(.text+0x2cb4): undefined > reference to `void > list >::insert<__list_iterator int const &, in > t const *> >(__list_iterator, __list_iterator const &, int con > st *>, __list_iterator)' > \wpwr\host\src\gnu.simpc.cpp\gcc\cp\tinfo2.cc(.text+0x39c2): undefined > reference to `void > list >::insert<__list_iterator int const &, in > t const *> >(__list_iterator, __list_iterator const &, int con > st *>, __list_iterator)' > make: *** [vxWorks] Error 0x1 > > Can someone help? > > Thanks --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Mon Dec 3 18:02:39 2001 From: =?iso-8859-1?q?Pramod=20Srinivasan?= Date: Mon Dec 3 18:02:42 PST 2001 Subject: vxworks core dump Hi, I work on powerpc, and I was wondering if there is a way to get a core dump of vxWorks? Is there a format for the core file defined anywhere? Thanks, --Pramod ________________________________________________________________ Nokia 5510 looks weird sounds great. Go to http://uk.promotions.yahoo.com/nokia/ discover and win it! The competition ends 16 th of December 2001. From vxwexplo-errs@csg.lbl.gov Mon Dec 3 22:07:10 2001 From: "hemendra kumar singh" Date: Mon Dec 3 22:07:12 PST 2001 Subject: VxWorks & ISA BUS This is a multi-part message in MIME format. ------=_NextPart_000_002A_01C17CB8.A5179F60 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="iso-8859-1" HI, Q. I have pentium BSP ,I have added a card (soundcard sb16 ISA) to ISA slot . Now How to detect ISA card in VxWorks.?? =20 Q . Soundcard have some I/O ports at 0x220 or 0x240...... The Intel x86 processors support both I/O-mapped devices and = memory-mapped devices. For I/O mapped devices, developers may use the following routines = from installDir/target/config/bspName/sysALib.s: sysInByte( )=20 - input one byte from I/O space=20 sysOutByte( )=20 - output one byte to I/O space=20 sysInWord( )=20 - input one word from I/O space=20 =20 Write to the ports 0x220 ...is required for further programming the = soundcard.I used sysOutByte(0x220,val) to write to it......but I this doesn,t work.... Now Question is How to map I/O of soundcard on ISA bus , so that it = can work as I/O mapped device.?? DO I have to add to sysPhyMemDesc[] ?? =20 =20 Thanks and Regards Hemendra ------=_NextPart_000_002A_01C17CB8.A5179F60 Content-Transfer-Encoding: quoted-printable Content-Type: text/html; charset="iso-8859-1"
HI,
 
Q. I have pentium BSP ,I have added a = card=20 (soundcard sb16 ISA) to ISA
slot . Now
 How to detect ISA card in=20 VxWorks.??
 
Q .  Soundcard have some I/O ports = at 0x220 or=20 0x240......
      =  The Intel x86=20 processors support both I/O-mapped devices and memory-mapped=20 devices.
       For I/O mapped devices, developers may use the = following=20 routines from
      =20 installDir/target/config/bspName/sysALib.s:
 
          = sysInByte( )=20
           &nb= sp;           &nbs= p;   =20 - input one byte from I/O space=20
          sysOutByte( ) =
           &nb= sp;           &nbs= p;   =20 - output one byte to I/O space=20
          sysInWord( )=20
           &nb= sp;           &nbs= p;   =20 - input one word from I/O space
   
    Write to the ports = 0x220 ...is=20 required for further programming the soundcard.I used=20 sysOutByte(0x220,val)
    to write to = it......but I this=20 doesn,t work....
   Now Question is = How to map I/O of=20 soundcard on ISA bus , so that it can work as I/O mapped = device.??
   DO I have to add to sysPhyMemDesc[] ??
 
 
Thanks and Regards
 Hemendra
------=_NextPart_000_002A_01C17CB8.A5179F60-- From vxwexplo-errs@csg.lbl.gov Tue Dec 4 02:39:05 2001 From: "Mike Anderson" Date: Tue Dec 4 02:39:07 PST 2001 Subject: RE: VxWorks & ISA BUS VxWorks Greetings! > Submitted-by: "hemendra kumar singh" > > > HI, > > Q. I have pentium BSP ,I have added a card (soundcard sb16 ISA) to ISA > slot . Now > How to detect ISA card in VxWorks.?? > =20 You don't provide enough information to determine if this card is one of the later plug-n-pray variants, so I assume that it isn't. If true, it has specific addresses that are set via jumpers. These will be in X86 I/O land. Therefore, you can't "detect" it, you simply have to know where it is and write to it. If it is one of the P-N-P variants, then you'll need to determine if your BIOS is set to enumerate the P-N-P devices (P-N-P O/S set to true) and go to the P-N-P configuration space to read the config. > Q . Soundcard have some I/O ports at 0x220 or 0x240...... > The Intel x86 processors support both I/O-mapped devices and = > memory-mapped devices. > For I/O mapped devices, developers may use the following routines = > from > installDir/target/config/bspName/sysALib.s: > > sysInByte( )=20 > - input one byte from I/O space=20 > sysOutByte( )=20 > - output one byte to I/O space=20 > sysInWord( )=20 > - input one word from I/O space=20 > =20 > Write to the ports 0x220 ...is required for further programming the = > soundcard.I used sysOutByte(0x220,val) > to write to it......but I this doesn,t work.... > Then, you're not writing to the correct address. These cards typically had several addresses where they could live. Look at the jumpers and figure out which one(s). > Now Question is How to map I/O of soundcard on ISA bus , so that it = > can work as I/O mapped device.?? > DO I have to add to sysPhyMemDesc[] ?? > =20 > =20 Nope. I/O space is completely different on the X86 and not mapped into sysPhysMemDesc. Find the correct addresses and use the sysXXX to talk to it. 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 Tue Dec 4 04:03:16 2001 From: Vxworks Exploder Date: Tue Dec 4 04:03:19 PST 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Tue Dec 4 04:03:11 PST 2001 Subject: Closing a socket while blocked in write or read Subject: Re: Problem Compiling bootroms for Walnut PPC405GP Subject: Re: Adding a custom routine into a bootrom image Subject: problem with using ZafApplication control in other classes Subject: Re: processor migration (MVME230X -> MVME5100 ALTIVEC) Subject: Re: problem with using ZafApplication control in other classes Subject: BOOT_LINE_ADDR and downloadable project Subject: IP_DO_FOWRARDING & IP_REDIRECT? Subject: Re: BOOT_LINE_ADDR and downloadable project Subject: Re: IP_DO_FOWRARDING & IP_REDIRECT? Subject: Re: Closing a socket while blocked in write or read Subject: Re: vxWorks SMP? Subject: patching question?? Subject: Re: Closing a socket while blocked in write or read Subject: Re: processor migration (MVME230X -> MVME5100 ALTIVEC) Subject: ZINC / global pointer for ZafApplication Subject: Re: ZINC / global pointer for ZafApplication Subject: How to Add to Tornado Prototyper "System Variables" in Windows98 Subject: "System Variables" setup in Windows98? Subject: Re: GUI display on top of VxWorks ? Subject: Re: Undefined reference in STL Subject: Docs for developing BSP. Subject: Strange Behavior of wtxtcl Subject: Re: Tornado 2, PPC Cumulative Patch 4 Subject: Re: Strange Behavior of wtxtcl Subject: Re: =?ISO-8859-1?Q?BOOT=5FLINE=5FADDR?= and downloadable project Subject: Re: Closing a socket while blocked in write or read Subject: Re: pci i/o address confliction? Subject: Could anyone tell me why the following program doesn't work as I have expected? Thanks Subject: Re: BOOT_LINE_ADDR and downloadable project Subject: Re: BOOT_LINE_ADDR and downloadable project Subject: Re: Tornado 2, PPC Cumulative Patch 4 ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Closing a socket while blocked in write or read Date: 3 Dec 2001 04:58:08 -0800 From: nick_nw@yahoo.co.uk (Nick) Organization: http://groups.google.com/ Message-ID: <2413908f.0112030458.6b69c8f1@posting.google.com> Hi, I am seeing some strange behaviour: Task 1: Blocking socket, call write (which blocks). Task 2: Close socket (which is blcoked in write). Task 1: No return from write. I thought that when the socket is closed it should return from any blocked calls to read or write. Anyone else come across this? Cheers, Nick --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Problem Compiling bootroms for Walnut PPC405GP Date: Mon, 3 Dec 2001 14:11:18 +0100 From: "Michael Lawnick" Organization: Buergernetz Dillingen Message-ID: <9uftlm$1bq$1@snoopy.bndlg.de> References: Reply-To: "Michael Lawnick" Sender: mlawnick@pd9e6568a.dip.t-dialin.net "Ho-Kuo Chan" schrieb im Newsbeitrag news:LETN7.921$R96.873140@carnaval.risq.qc.ca... > Hi, > I am developing for the IBM Walnut borad running the PPC405GP. I have > Tornado 2.0.2 installed (SPR 67953) which requires recompiling the bootrom. > After compiling either bootrom.hex or bootrom_uncmp.hex, inspection of the > S-Rec hex files reveals that the addresses for the data records are only > 16-bits wide (S1...). Neither bootroms work. Before reprogramming the eeprom > though, I saved the bootrom image that was shipped with the Eval board. > Inspection of this file shows that the addresses should be 32-bits (S3...). > I compiled the bootroms without any modification to the BSP as it was > shipped in T2.0.2. Am I forgetting to do something? Thanks!!! > - -- Hi Ho-Kuo Chan, In your bootrom.hex you should find S1 - records for the first 64k S2 - for the rest of the file (up to 16M), or all the file if -l is set in elfHex command S3 - never - --> Tornado Reference : Tornado Utilities elfHex 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 ============================================== --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Adding a custom routine into a bootrom image Date: Mon, 3 Dec 2001 14:36:12 +0100 From: "Michael Lawnick" Organization: Buergernetz Dillingen Message-ID: <9ufv4b$2bo$1@snoopy.bndlg.de> References: Reply-To: "Michael Lawnick" Sender: mlawnick@pd9e6568a.dip.t-dialin.net "Jae-Hwa Lee" schrieb im Newsbeitrag news:OjEO7.37115$P%2.2983308@news.bora.net... > Hi, everyone? > > I'm trying to add a custom routine for a ftp server into a bootrom image. > All source codes for the ftp server are in ftp.a > Some routines in ftp.a are referred to other libraries such as > libPPCEC603gnuvx.a, rtlib.a and so on. > Except libPPCEC603gnuvx.a, all libraries are custom made. > > In order to compile the bootrom image with the custom ftp routine, I added > the followings into the bsp 'Makefile'. > > LIBS = C:/AX2200/Lib/ftp.a C:/AX2200/Lib/main.a C:/AX2200/Lib/cmd_proc.a > \ > C:/AX2200/Lib/rtlib.a C:/AX2200/Lib/cmp.a C:/AX2200/Lib/n2p.a > However, there happened some make errors, when I executed the 'Makefile' > The errors are following > C:/AX2200/Lib/main.a(cfgcmd.o): In function `Cmd_SetIpAddr_Lan': > cfgcmd.o(.text+0x5a40): undefined reference to `NV_FTP_VER_load_from_nvram' > cfgcmd.o(.text+0x6230): undefined reference to `NV_FTP_VER_save_to_nvram' > make: *** [bootrom_uncmp] Error 0x1 > > In fact, the codes for the ftp server was designed for VxWorks image and > they were already verified. > I guess the reason why the ftp server code is not compiled is that the BSP > makefile does not make a partialImage.o. > According to the 'Makefile' for the VxWorks image, it generates a > partialImage.o with required libraries and objects and then a symbol > table(symTbl.o) is made by the partialImage.o. > Anyway what I wanna know is how I can insert a custom routine in the BSP > codes and then make a bootrom image without undefined reference errors. > Hi Jay Lee first (sure): there is no need for a partial image or symTbl, they are needed for C++ munching and dynamic linking, nothing that will be done in bootrom context. scnd (AFAIK): The way you go is, you do a call in an .c file of your bootloader and try to automatically add objects from archives as needed. This will work as long as no archive references symbols in another one that was linked before. The linking process is linear ! Normally you should add all your objects to one archive. This way the linker is able to resolve all cross references. The easiest way: don't build an archive, but add all objects with MACH_EXTRA. 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 ============================================== --------------------------- Newsgroups: comp.os.vxworks Subject: problem with using ZafApplication control in other classes Date: 3 Dec 2001 05:54:10 -0800 From: aravinddk@tataelxsi.co.in (aravind) Organization: http://groups.google.com/ Message-ID: I have an problem in zinc, Iam unable to get the pointer to the ZafApplication is there any way to get an pointer to the application, This I assume is required since I am unable to trap the key pressed in the GUI. problem detail We have a GUI class A which is getting key press and transfer the control to another class B which is derived from ZafApplication, we are unable to use ZafApplication in the constructor , it is giving compiler error, please tell me How to use ZafApplication's control function in class B --------------------------- Newsgroups: comp.os.vxworks Subject: Re: processor migration (MVME230X -> MVME5100 ALTIVEC) Date: 3 Dec 2001 07:11:26 -0800 From: Robert Gove Organization: Newsguy News Service [http://newsguy.com] Message-ID: <9ug4mu0fh2@drn.newsguy.com> References: In article , Andray says... > >"Stephen" wrote in news:vzRN7.630$h7.21374 >@llslave.llan.ll.mit.edu: > >> >> We are using VxWorks on MOTO MVME2300 boards. >> We are interested in adding a motorola 5100 ALTIVEC card to an existing VME >> chassis (i.e. with teh 2300 boards) >> Wind River is saying that in order to support this we need to buy a whole >> new development package !! >> We've moved to differnet processor before and had to simply pay for the new >> BSP. I can see paying a bit extra for ALTIVEC libarraies or something liek >> that but a whole new dev system does not seem correct. >> >> I am wondeing if anyone else is facing the same issue. Has anyone migarted >> from MVME2300 to 5100 ALTIVEC boards ? >> >> Thanks for any help >> Stephen Rejto >> >[AndrayK]. We have exactly the same situation here - we can't simply >get 5100 BSP and plug it into existing Tornado/ppc installation - >we have to buy the whole new Tornado seat. > >Andray We've migrated from mvme2400 boards (750's) to mvme5100's (7400's) and got the same feedback as you did. Wind River is stating that the 7400 architecture is different from the 750 and therefore, requires a new pricing structure. I believe our quote was $7,500 (Tornado 2.1). You should really try talking to your sales rep again. I think you'll end up paying only for a new BSP. - --Robert > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: problem with using ZafApplication control in other classes Date: 03 Dec 2001 10:33:35 -0500 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: References: Sender: bpringlemeir@DeadDuck >>>>> "aravind" == aravind writes: aravind> I have an problem in zinc, Iam unable to get the pointer to aravind> the ZafApplication is there any way to get an pointer to the aravind> application, This I assume is required since I am unable to You can store the pointer in some class when you create the ZafApplication class. I was unable to find anything that would get the ZafApplication directly. Note! This all depends on what version of Zinc you are using. There are quite a few changes between Zinc 6.0 and 5.3. You can look around in the code for a global pointer if you don't like the method I described above. hth, Bill Pringlemeir. - -- You can watch when you try to watch. vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" --------------------------- Newsgroups: comp.os.vxworks Subject: BOOT_LINE_ADDR and downloadable project Date: Mon, 3 Dec 2001 17:54:17 +0200 From: "Ofer Goren" Organization: Verio Message-ID: hi I want to use BOOT_LINE_ADDR in my file. I have no problem if the file is part of bootable project, since it is defined in prjParams.h. However, if I'm adding the file to a d/l-able project, there is no prjParams.h file, so I have to include config.h. So, what do I need to do so that a d/l-able project will include config.h, and a bootable project will include prjParams.h (which can be different from config.h). Is there a define I can use that define something different in case of a bootable/dl-able project? Thank, - -- Ofer Goren Iamba Technologies --------------------------- Newsgroups: comp.os.vxworks Subject: IP_DO_FOWRARDING & IP_REDIRECT? Date: Mon, 03 Dec 2001 17:27:24 +0000 From: Owain Phillips Organization: Siemens Inc. Message-ID: <3C0BB5FC.3970CA8E@siemenscomms.co.uk> This is a multi-part message in MIME format. - --------------F5FD7CC2B835A01C1D998607 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Anyone tell me what these flags do? Apparently our ethernet interface is configured with these flags on and that was the state as we received the BSP / Vxworks. Are these flags normal default settings for a normal IP endpoint? Thanks in advance, Owain - --------------F5FD7CC2B835A01C1D998607 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 - --------------F5FD7CC2B835A01C1D998607-- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: BOOT_LINE_ADDR and downloadable project Date: Mon, 3 Dec 2001 09:27:06 -0800 From: "Gary M" Message-ID: <3c0bb67d$0$8348$4c41069e@reader0.ash.ops.us.uu.net> References: Reply-To: "Gary M" Wouldn't this work? In your bootable project (somewhere): char *pBootLineAddr = BOOT_LINE_ADDR; In the downloadable project: extern char *pBootLineAddr; "Ofer Goren" wrote in message news:newscache$f20sng$5i7$1@lnews.actcom.co.il... > hi > I want to use BOOT_LINE_ADDR in my file. I have no problem if the file is > part of bootable project, since it is defined in prjParams.h. However, if > I'm adding the file to a d/l-able project, there is no prjParams.h file, so > I have to include config.h. > So, what do I need to do so that a d/l-able project will include config.h, > and a bootable project will include prjParams.h (which can be different from > config.h). Is there a define I can use that define something different in > case of a bootable/dl-able project? > > Thank, > > -- > Ofer Goren > Iamba Technologies > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: IP_DO_FOWRARDING & IP_REDIRECT? Date: Mon, 3 Dec 2001 20:14:00 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <3C0BB5FC.3970CA8E@siemenscomms.co.uk> IP_DO_FORWARDING = 1 means that a packet coming in through one interface may be transmitted out of another interface, as called for by the current routing table - a.k.a. IP Routing. IP_REDIRECT is a feature of IP that can generate a ICMP_REDIRECT message to a sender, which tries to use the node as router, but this nodes "knows" there is a "better" route for this packet. This ICMP_REDIRECT will typically modify the initiator's routing table so that subsequent packets will be send using the "better route". So it makes sense to torun off both if the VxWorks has several interfaces but is not supposed to act as a Router. Hope this helps, - - Leonid "Owain Phillips" wrote in message news:3C0BB5FC.3970CA8E@siemenscomms.co.uk... > Anyone tell me what these flags do? > > Apparently our ethernet interface is configured with these flags on and > that was the state as we received the BSP / Vxworks. > > Are these flags normal default settings for a normal IP endpoint? > > Thanks in advance, > Owain > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Closing a socket while blocked in write or read Date: Mon, 3 Dec 2001 20:17:37 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <2413908f.0112030458.6b69c8f1@posting.google.com> This depends on the value of SO_LINGER option on the socket - "lingering" means not to destroy the connection before an entire data has been received by the other end. For example, if this would be an FTP data connection, "end of connection" signifies end-of-file, but loss of data near end of file is unacceptable. Of course, the socket is not supposed to "linger" forever, as there is a "lingering timeout" value, so if this deadlock does not solve itself, it is beyond any doubt a bug in the code. If you got a valid support contract, try to search in the SPR database on the WindRiver web site, perhaps this bug is already known. - - Leonid "Nick" wrote in message news:2413908f.0112030458.6b69c8f1@posting.google.com... > Hi, > > I am seeing some strange behaviour: > > Task 1: Blocking socket, call write (which blocks). > Task 2: Close socket (which is blcoked in write). > Task 1: No return from write. > > I thought that when the socket is closed it should return from any > blocked calls to read or write. > > Anyone else come across this? > > Cheers, > > Nick --------------------------- Newsgroups: comp.os.vxworks Subject: Re: vxWorks SMP? Date: Mon, 3 Dec 2001 20:21:02 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: Some time ago, circa 1996, there was a project which I beleive was done at SPARTA, a known VxWorks VAR at the time. The project was specific to a particular SPARC multi processor board, and if my memory serves me well, it actually worked. There was very little interest in SMP in embedded applications (for obvious reasons) which is the lion share of Wind's market. Higher-end R/T projects where not too interested in this either. They'd rather use a multi-DSP COTS boards for crunching numbers than a Multiporcessor VxWorks. - - Leonid "John" wrote in message news:aTkO7.80843$T67.21990996@news4.rdc1.on.home.com... > Hi, > > Can anyone tell me if vxWorks supports SMP or not? > > Cheers > John > > --------------------------- Newsgroups: comp.os.vxworks Subject: patching question?? Date: Mon, 03 Dec 2001 13:54:13 -0600 From: PC User Organization: ExecPC Internet - Milwaukee, WI Message-ID: <3c0bdaca$0$43568$272ea4a1@news.execpc.com> Hi, I tried to find something about patching in VxWorks programmer's guide and nothing is listed in the index. Does anybody know where? I am trying to test and see if VxWorks provides the mean for an application program, during execution, to patch itself or another application program. DoesVxWorks provide a tool to do this? What I like to do is to have a task called patching and let it handle patch request through a message queue where other tasks and users via shell will request patches to be made. Thanks. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Closing a socket while blocked in write or read Date: Mon, 3 Dec 2001 23:18:39 -0000 From: "David Laight" Organization: BT Internet Message-ID: <9uh16e$63s$1@neptunium.btinternet.com> References: <2413908f.0112030458.6b69c8f1@posting.google.com> Unfortunately what you are doing isn't allowed! vxWorks leaves it up to the application to ensure that no file accesses are pending when close() is called! Further read/write/ioctl requests are also not barred until close returns - ie the device driver can see a read after a close! (unless the application polices itself). If you could guarantee that the write was actually blocked (which you can't 100%) the FIOCANCEL ioctl will unblock a read/write call - - provided the driver supports it (dunno if sockets do...). But you need to have already flagged your writer that the close is pending (or it will write() again). I did 'fix' a serial driver to count pending read/write/ioctl requests and defer close until they had gone away (using FIOCANCEL to try to persuade them out) - but it was ugly as hell! For sockets you might get away with using shutdown() instead of close() - dunno what it does to the blocked write... Blocked readers are more common.... David Leonid Rosenboim wrote in message news:newscache$eq6sng$cq7$1@lnews.actcom.co.il... > This depends on the value of SO_LINGER option on the socket - > "lingering" means not to destroy the connection before an entire > data has been received by the other end. For example, if this > would be an FTP data connection, "end of connection" signifies > end-of-file, but loss of data near end of file is unacceptable. > Of course, the socket is not supposed to "linger" forever, > as there is a "lingering timeout" value, so if this deadlock > does not solve itself, it is beyond any doubt a bug in the code. > > If you got a valid support contract, try to search in the SPR database > on the WindRiver web site, perhaps this bug is already known. > > - Leonid > > "Nick" wrote in message > news:2413908f.0112030458.6b69c8f1@posting.google.com... > > Hi, > > > > I am seeing some strange behaviour: > > > > Task 1: Blocking socket, call write (which blocks). > > Task 2: Close socket (which is blcoked in write). > > Task 1: No return from write. > > > > I thought that when the socket is closed it should return from any > > blocked calls to read or write. > > > > Anyone else come across this? > > > > Cheers, > > > > Nick > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: processor migration (MVME230X -> MVME5100 ALTIVEC) Date: 3 Dec 2001 15:46:09 -0800 From: spk@uswest.net (Scott Kamara) Organization: http://groups.google.com/ Message-ID: <73e78f4e.0112031546.72f211d8@posting.google.com> References: <9ug4mu0fh2@drn.newsguy.com> Robert Gove wrote in message news:<9ug4mu0fh2@drn.newsguy.com>... > We've migrated from mvme2400 boards (750's) to mvme5100's (7400's) and got the > same feedback as you did. Wind River is stating that the 7400 architecture is > different from the 750 and therefore, requires a new pricing structure. I > believe our quote was $7,500 (Tornado 2.1). You should really try talking to > your sales rep again. I think you'll end up paying only for a new BSP. Why do you say that they might end up paying only for a new BSP? - -Scott Kamara --------------------------- Newsgroups: comp.os.vxworks Subject: ZINC / global pointer for ZafApplication Date: Mon, 03 Dec 2001 23:48:18 GMT From: "ChanSeong Kim" Organization: Dacom Message-ID: <6bUO7.37141$P%2.3046155@news.bora.net> References: <3c0bdaca$0$43568$272ea4a1@news.execpc.com> Hi all vxworkers and zinc applicationer(?)s... global pointer for ZafApplication is zafApplication. I think zafApplication is a global pointer for zinc application you can access this pointer everywhere in your program if you include apropriate(right) header file for this. Following is a source of zinc. This is a constructor of ZafApplication class. You can see zafApplication is assigned with pointer of class 'this'. And another global variables can be shown like zafEventManager, zafWindowManager. Good Luck..!! ZafApplication::ZafApplication(int _argc, char **_argv) : argc(0), argv(ZAF_NULLP(ZafIChar *)), error(ZAF_ERROR_NONE) { #if defined(ZAF_INDEPENDENT) // This call is used to reinitialized global variables each time // the application is run. (Otherwise the program would need to // be re-loaded for globals to be reinitialized.) InitializeGlobals(); #endif // Initialize the data manager. zafDataManager = new ZafDataManager; // Call I18nAllocate() to reflect changes based on the display. ZafI18nData::I18nAllocate(); // Allocate the display. zafDisplay = new ZafScreenDisplay(_argc, _argv); #if defined(ZAF_INDEPENDENT) // If the display didn't initialize, exit gracefully. if (!((ZafScreenDisplay *)zafDisplay)->initialized) { SetError(ZAF_ERROR_CONSTRUCTOR); printf("Couldn't initialize the screen display.\n"); return; } #endif // Allocate the arguments. While this code technically allows argv[0] to // be NULL (which violates the language definition), this condition // should not occur because the data passed to us should always be valid // (i.e., _argv[0] should contain the program name). If the data is // invalid, it is likely a compiler error, or at least is a problem // that cannot be fixed here so we do our best to work with it by allowing // for the invalid state. // Initialize argc here since an error may have occurred before this point. argc = _argc; argv = new ZafIChar *[argc + 1]; for (int i = 0; i < argc; i++) { #if defined(USE_ZAF_CODE_SET_DATA) argv[i] = zafCodeSet->ConvertToZafString(_argv[i]); #else argv[i] = strdup(_argv[i]); #endif } argv[argc] = ZAF_NULLP(ZafIChar); // Create search path from the program path. #if defined(USE_ZAF_SEARCH_PATH) if (argv[0]) zafSearchPath = new ZafPath(argv[0], true); else zafSearchPath = new ZafPath(ZAF_NULLP(ZafIChar), true); #endif #if defined(USE_ZAF_PERSISTENCE) // Call ResetI18n() to reflect changes based on the storage. // zafI18nStorage = new ZafStorage(ZAF_ITEXT("i18n.znc"), ZAF_FILE_READ); zafI18nStorage = new ZafStorage(i18n_znc); if (zafI18nStorage->Error()) { delete zafI18nStorage; zafI18nStorage = ZAF_NULLP(ZafFileSystem); } else { zafDataPersistence = new ZafDataPersistence(zafI18nStorage, 0); zafDataPersistence->AddDataConstructor(ZafLanguageData::className, ZafLanguageData::classID, &ZafLanguageData::Read); zafDataPersistence->AddDataConstructor(ZafLocaleData::className, ZafLocaleData::classID, &ZafLocaleData::Read); zafDataPersistence->AddDataConstructor(ZafCodeSetData::className, ZafCodeSetData::classID, &ZafCodeSetData::Read); ZafI18nData::ResetI18n(); } #endif // Allocate the event manager. zafEventManager = new ZafEventManager; InitMouse(); InitKeyboard(); zafEventManager->Add(new ZafCursor); InitHelpTips(); // Allocate the window manager. zafWindowManager = new ZafWindowManager; // Allocate the error system. InitErrorSystem(); // Initialize the global pointer to the application object. zafApplication = this; } ----- Original Message ZafApplication::ZafApplication(int _argc, char **_argv) : argc(0), argv(ZAF_NULLP(ZafIChar *)), error(ZAF_ERROR_NONE) { #if defined(ZAF_INDEPENDENT) // This call is used to reinitialized global variables each time // the application is run. (Otherwise the program would need to // be re-loaded for globals to be reinitialized.) InitializeGlobals(); #endif // Initialize the data manager. zafDataManager = new ZafDataManager; // Call I18nAllocate() to reflect changes based on the display. ZafI18nData::I18nAllocate(); // Allocate the display. zafDisplay = new ZafScreenDisplay(_argc, _argv); #if defined(ZAF_INDEPENDENT) // If the display didn't initialize, exit gracefully. if (!((ZafScreenDisplay *)zafDisplay)->initialized) { SetError(ZAF_ERROR_CONSTRUCTOR); printf("Couldn't initialize the screen display.\n"); return; } #endif // Allocate the arguments. While this code technically allows argv[0] to // be NULL (which violates the language definition), this condition // should not occur because the data passed to us should always be valid // (i.e., _argv[0] should contain the program name). If the data is // invalid, it is likely a compiler error, or at least is a problem // that cannot be fixed here so we do our best to work with it by allowing // for the invalid state. // Initialize argc here since an error may have occurred before this point. argc = _argc; argv = new ZafIChar *[argc + 1]; for (int i = 0; i < argc; i++) { #if defined(USE_ZAF_CODE_SET_DATA) argv[i] = zafCodeSet->ConvertToZafString(_argv[i]); #else argv[i] = strdup(_argv[i]); #endif } argv[argc] = ZAF_NULLP(ZafIChar); // Create search path from the program path. #if defined(USE_ZAF_SEARCH_PATH) if (argv[0]) zafSearchPath = new ZafPath(argv[0], true); else zafSearchPath = new ZafPath(ZAF_NULLP(ZafIChar), true); #endif #if defined(USE_ZAF_PERSISTENCE) // Call ResetI18n() to reflect changes based on the storage. // zafI18nStorage = new ZafStorage(ZAF_ITEXT("i18n.znc"), ZAF_FILE_READ); zafI18nStorage = new ZafStorage(i18n_znc); if (zafI18nStorage->Error()) { delete zafI18nStorage; zafI18nStorage = ZAF_NULLP(ZafFileSystem); } else { zafDataPersistence = new ZafDataPersistence(zafI18nStorage, 0); zafDataPersistence->AddDataConstructor(ZafLanguageData::className, ZafLanguageData::classID, &ZafLanguageData::Read); zafDataPersistence->AddDataConstructor(ZafLocaleData::className, ZafLocaleData::classID, &ZafLocaleData::Read); zafDataPersistence->AddDataConstructor(ZafCodeSetData::className, ZafCodeSetData::classID, &ZafCodeSetData::Read); ZafI18nData::ResetI18n(); } #endif // Allocate the event manager. zafEventManager = new ZafEventManager; InitMouse(); InitKeyboard(); zafEventManager->Add(new ZafCursor); InitHelpTips(); // Allocate the window manager. zafWindowManager = new ZafWindowManager; // Allocate the error system. InitErrorSystem(); // Initialize the global pointer to the application object. zafApplication = this; } --------------------------- Newsgroups: comp.os.vxworks Subject: Re: ZINC / global pointer for ZafApplication Date: 03 Dec 2001 20:38:05 -0500 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: References: <3c0bdaca$0$43568$272ea4a1@news.execpc.com> <6bUO7.37141$P%2.3046155@news.bora.net> Sender: bpringlemeir@DeadDuck >>>>> "CS" == ChanSeong Kim writes: CS> Hi all vxworkers and zinc applicationer(?)s... global CS> pointer for ZafApplication is zafApplication. CS> I think zafApplication is a global pointer for zinc CS> application you can access this pointer everywhere in your CS> program if you include apropriate(right) header file for CS> this. Ok, that appears to be Zinc 5.x? code. I check the Zinc 6.0 code and it too assigns "zafApplication = this;" in the ZafApplication constructor. I think it notes this in the Zinc manual as well. hth, Bill Pringlemeir. - -- I'm encased in the lining of a pure pork sausage!! vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" --------------------------- Newsgroups: comp.os.vxworks Subject: How to Add to Tornado Prototyper "System Variables" in Windows98 Date: Tue, 04 Dec 2001 02:50:24 GMT From: "tiejunyang" Message-ID: Reply-To: "tiejunyang" Help please! After I got Tornado Prototyper FlexLM license key, I do not know modify which system file to "Set the environment variable LM_LICENSE_FILE to point to the path of the license file" in Windows98, Following setup is from Wind River for Windows2000 * Bring up the control panel * Click on the "system" icon * Click on the "Environment" tab * Under "System Variables" define a new variable LM_LICENSE_FILE with the value $WIND_BASE/.wind/license/WRSLicense.lic. Are there equation setup in Windows98, which file should be modified? Thanks, email:tiejunyang@yahoo.com --------------------------- Newsgroups: comp.os.vxworks Subject: "System Variables" setup in Windows98? Date: Tue, 04 Dec 2001 02:52:38 GMT From: "tiejunyang" Message-ID: Reply-To: "tiejunyang" Help please! After I got Tornado Prototyper FlexLM license key, I do not know modify which system file to "Set the environment variable LM_LICENSE_FILE to point to the path of the license file" in Windows98, Following setup is from Wind River for Windows2000 * Bring up the control panel * Click on the "system" icon * Click on the "Environment" tab * Under "System Variables" define a new variable LM_LICENSE_FILE with the value $WIND_BASE/.wind/license/WRSLicense.lic. Are there equation setup in Windows98, which file should be modified? Thanks, email:tiejunyang@yahoo.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: GUI display on top of VxWorks ? Date: Tue, 04 Dec 2001 04:47:16 GMT From: Gwen & Emory Stagmer Organization: Excite@Home - The Leader in Broadband http://home.com/faster Message-ID: <3C0C54F3.BE8E2313@home.com> References: Just open a TCP/IP port and send the data to a workstation (PC/Mac/*nix) program, then display it any way you want. Or, if you don't have an ethernet connection, just print it to a COM port... The TCP/IP isn't hard at all, and boundless examples exist. Emory Stagmer Chris Smith wrote: > > Hello, > > I'm wondering if someone can tell me of some products out there that > would let me easily build a GUI frontend on top of VxWorks application. > I'm developing on a PPC board and want to plot and display some program > variables as the program is running. I've come across Tilcon which > looks like what I'm looking for. Are there any other similar products > out there to consider. > > TIA --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Undefined reference in STL Date: Tue, 04 Dec 2001 04:50:30 GMT From: Gwen & Emory Stagmer Organization: Excite@Home - The Leader in Broadband http://home.com/faster Message-ID: <3C0C55B6.546F2CA8@home.com> References: <3c09d5f9$1@news.barak.net.il> There's a 'top-level' include for STL - I'm trying to remember what is off the top of my head...it MIGHT be: #include but I'm guessing. Whatever it is will give you ALL the STL and I think that will solve your problem. I was having similar problems trying to use just and had to get the whole thing... Emory Stagmer AE wrote: > > Hi > > I'm trying to compile the following code in Tornado 2.02 > > // ******************************************* > #include > > template std::list; > > int InitApplic() > { > std::list myList; > > return 0; > } > // ******************************************* > > I'm compiling with -frepo. > > Unfortunately I get the following message during link > > ldsimpc --subsystem=windows -e _WinMainCRTStartup > C:\Tornado\host\x86-win32\i386-pc-mingw3 > 2\lib\crt1.o \ > dataSegPad.o partialImage.o ctdt.o -o vxWorks > partialImage.o: In function `_tf9type_info': > \wpwr\host\src\gnu.simpc.cpp\gcc\cp\tinfo2.cc(.text+0x2cb4): undefined > reference to `void > list >::insert<__list_iterator int const &, in > t const *> >(__list_iterator, __list_iterator const &, int con > st *>, __list_iterator)' > \wpwr\host\src\gnu.simpc.cpp\gcc\cp\tinfo2.cc(.text+0x39c2): undefined > reference to `void > list >::insert<__list_iterator int const &, in > t const *> >(__list_iterator, __list_iterator const &, int con > st *>, __list_iterator)' > make: *** [vxWorks] Error 0x1 > > Can someone help? > > Thanks --------------------------- Newsgroups: comp.os.vxworks Subject: Docs for developing BSP. Date: 3 Dec 2001 21:50:21 -0800 From: kamala.kodi@wipro.com (Kamala) Organization: http://groups.google.com/ Message-ID: <27203e86.0112032150.27dfe19b@posting.google.com> I need documents / manuals for developing / porting BSP. Can any one suggest from where i can get the same. We haven't bought BSP ref. kit. Thanks in advance, kamala --------------------------- Newsgroups: comp.os.vxworks Subject: Strange Behavior of wtxtcl Date: 3 Dec 2001 22:23:12 -0800 From: brycenga@wire-less-inc.com (Brock) Organization: http://groups.google.com/ Message-ID: <3d71cba2.0112032223.4f7b9743@posting.google.com> In my makefile, exists the command: c:/tornado/host/x86-win32/bin/nmppc snt_mux.partialImage.o | c:/tornado/host/x86-win32/bin/wtxtcl c:/tornado/host/src/hutils/munch.tcl -asm ppc > snt_ctdt.c This command fails with the following error: 'c:' is not recognized as an internal or external command, operable program or batch file. The only way I can get this line to run properly is if I enter as follows: nmppc snt_mux.partialImage.o | wtxtcl c:/tornado/host/src/hutils/munch.tcl -asm ppc > snt_ctdt.c Could anyone explain to me why the first command line won't execute properly? Many thanks, Brock --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado 2, PPC Cumulative Patch 4 Date: Tue, 04 Dec 2001 06:52:55 GMT From: nunya Organization: Nunya Business Message-ID: <3C0C72C7.A2BE4A2D@nunya.com> References: <3C06AFD7.CB2C9D7A@nunya.com> >Hi Nunya, >"nunya" wrote in message >news:3C06AFD7.CB2C9D7A@nunya.com... >> Hi All: >> >> We've run into a problem with this patch. >> >> We have our own networking protocol stack that registers >> with the mux as Promiscuos (because of the well-known >> failures with Snarfing. ) >> >> Up to cumulative patch 2, this was working fine. >> >> With cumulative patch 4, however, our receive routine is >> never called. This makes it very difficult for our protocol >> stack to function efficiently. :-) >> >> Has anyone else run into this problem? > >T2CP4 upgrades Tornado 2 to Tornado 2.0.2. T2CP4 definitely causes problems >with the PPC405GP. There is a PPC405 specific patch SPR 67953 which seems >to work fine. Thanks! We'll try to pry that out of WRS. Something else we found, the END_OBJ passed in to our receive routine is _not_ the same END_OBJ that was passed back to us from the muxBind call. Because we were validating the END_OBJ, and they didn't match, we were ignoring the packet. If we don't validate the END_OBJ passed in, we work. This is, for now, a work-around, but we will pursue with WRS as a bug in their mux code.... Thanks again! JLM --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Strange Behavior of wtxtcl Date: Tue, 4 Dec 2001 08:04:14 +0100 From: Patrick Organization: Siemens Inc. Message-ID: References: <3d71cba2.0112032223.4f7b9743@posting.google.com> On 3 Dec 2001, Brock wrote: > In my makefile, exists the command: > > c:/tornado/host/x86-win32/bin/nmppc snt_mux.partialImage.o | > c:/tornado/host/x86-win32/bin/wtxtcl > c:/tornado/host/src/hutils/munch.tcl -asm ppc > snt_ctdt.c > > This command fails with the following error: > > 'c:' is not recognized as an internal or external command, operable > program or batch file. Because make does not recognise the second line as a continuation of the first line, try escaping the carriage return on the first line: c:/tornado/host/x86-win32/bin/nmppc snt_mux.partialImage.o | \ c:/tornado/host/x86-win32/bin/wtxtcl c:/tornado/host/src/hutils/munch.tcl -asm ppc > snt_ctdt.c (Note the '\' character at the end of the first line.) Regards Patrick - ---------------------------------------------------------------------- strip 'to_undies' to reply --------------------------- Newsgroups: comp.os.vxworks Subject: Re: =?ISO-8859-1?Q?BOOT=5FLINE=5FADDR?= and downloadable project Date: 04 Dec 2001 09:50:25 +0200 From: Reinhold Wirth Organization: WEB.DE Message-ID: <3c0c8e51$1@netnews.web.de> References: "Ofer Goren" wrote: >I want to use BOOT_LINE_ADDR in my file. I have no problem if the file is >part of bootable project, since it is defined in prjParams.h. However, if >I'm adding the file to a d/l-able project, there is no prjParams.h file, so >I have to include config.h. >So, what do I need to do so that a d/l-able project will include config.h, >and a bootable project will include prjParams.h (which can be different from >config.h). Is there a define I can use that define something different in >case of a bootable/dl-able project? * Are you sure, that it is possible use different values of BOOT_LINE_ADR in config.h and prjParams.h? * You may use the variable sysBootLine instead BOOT_LINE_ADRS. ( in sysLib.c it is called: char *sysBootLine = BOOT_LINE_ADRS) * If you want to include config.h in your d/l-able project vou have to: - set VPATH (in Builds; Properties; Macros) to $(TGT_DIR)/config/all $(TGT_DIR)/config/<"your_BSP"> - add the path of config.h and configAll.h to the compiler (in Builds; Properties; C/C++ compiler) add: -Id:/tornado/target/config/all -Id:/tornado/target/config/<"your_BSP"> - -- __________________________________________________________ News suchen, lesen, schreiben mit http://newsgroups.web.de --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Closing a socket while blocked in write or read Date: Tue, 4 Dec 2001 11:04:12 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <2413908f.0112030458.6b69c8f1@posting.google.com> <9uh16e$63s$1@neptunium.btinternet.com> David, you are probably right with respect to simple drivers, but the sockets library should be exceptionally clean in terms of mutual exclusion and interlocking, because the file descriptors are global. Perhaps FIOCANCEL prior to close() is a good solution, if the user indeed wants that the unsent data is thrown away. It is a strange scenario without a doubt. "David Laight" wrote in message news:9uh16e$63s$1@neptunium.btinternet.com... > Unfortunately what you are doing isn't allowed! > > vxWorks leaves it up to the application to ensure that no file > accesses are pending when close() is called! > Further read/write/ioctl requests are also not barred until close > returns - ie the device driver can see a read after a close! (unless > the application polices itself). > > If you could guarantee that the write was actually blocked (which > you can't 100%) the FIOCANCEL ioctl will unblock a read/write call > - provided the driver supports it (dunno if sockets do...). But > you need to have already flagged your writer that the close is pending > (or it will write() again). > > I did 'fix' a serial driver to count pending read/write/ioctl requests > and defer close until they had gone away (using FIOCANCEL to try to > persuade them out) - but it was ugly as hell! > > For sockets you might get away with using shutdown() instead of > close() - dunno what it does to the blocked write... > > Blocked readers are more common.... > > David > > > Leonid Rosenboim wrote in message > news:newscache$eq6sng$cq7$1@lnews.actcom.co.il... > > This depends on the value of SO_LINGER option on the socket - > > "lingering" means not to destroy the connection before an entire > > data has been received by the other end. For example, if this > > would be an FTP data connection, "end of connection" signifies > > end-of-file, but loss of data near end of file is unacceptable. > > Of course, the socket is not supposed to "linger" forever, > > as there is a "lingering timeout" value, so if this deadlock > > does not solve itself, it is beyond any doubt a bug in the code. > > > > If you got a valid support contract, try to search in the SPR database > > on the WindRiver web site, perhaps this bug is already known. > > > > - Leonid > > > > "Nick" wrote in message > > news:2413908f.0112030458.6b69c8f1@posting.google.com... > > > Hi, > > > > > > I am seeing some strange behaviour: > > > > > > Task 1: Blocking socket, call write (which blocks). > > > Task 2: Close socket (which is blcoked in write). > > > Task 1: No return from write. > > > > > > I thought that when the socket is closed it should return from any > > > blocked calls to read or write. > > > > > > Anyone else come across this? > > > > > > Cheers, > > > > > > Nick > > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: pci i/o address confliction? Date: 4 Dec 2001 01:42:35 -0800 From: lixijun@pub.xaonline.com (Lee) Organization: http://groups.google.com/ Message-ID: <82f462ab.0112040142.3bf03136@posting.google.com> References: <82f462ab.0111272306.691dd6be@posting.google.com> <3c0750e7$1@brateggebdc5.br-automation.co.at> Hi Werner thanks very much. I has resolved the problem. The bug is in bios setting. the item is "spread spectrum clock modulation", the default value is "low", so I change it to "off", every thing is ok! I don't know what's the meaning of that. BTY Lee "Werner Schiendl" wrote in message news:<3c0750e7$1@brateggebdc5.br-automation.co.at>... > Hi, > > > "Lee" wrote in message > news:82f462ab.0111272306.691dd6be@posting.google.com... > > Hi all > > I'm new in vxworks. I write a cpci driver with x86 processor board > > I can get pci config info as following: > > base address 0 = 0xfe6fdf00 > > base address 1 = 0x0000d801 > > base address 2 = 0xff1e0008 > > base address 3 = 0x0000df01 > > as my pci board manaul said, the BAR 3 is application registers, > > I can test it though write 0x01 to 0xdf22, the board led light one, > > write 0x02, light two, etc. The problem is the board do not respond > > after you did serveral operation. At this time, read from status register > > (0x0000df21)value is 0xff. > > I startup system in win2000, the device manager said there was address > > confict about 0xdf00~0xdf3f. But there is no other device use it. > > Did you consider that some old style ISA devices need only decode 10 bit IO > addresses? > This makes your addressed effectively conflict with such devices on IO > 0x300 - 0x33f. > > > (The plug and play data of the devices contain information on whether they > are decoded 10 or 16 bit, so Windows 2K is probably right with its > assumption). > > hth > Werner --------------------------- Newsgroups: comp.os.vxworks Subject: Could anyone tell me why the following program doesn't work as I have expected? Thanks Date: Tue, 4 Dec 2001 17:48:59 +0800 From: "zhaoyandong" Organization: Bentium Ltd. (CN99) Message-ID: <9ui6ha$no5$1@mail.cn99.com> #include "vxWorks.h" //#include "semLib.h' #include "taskLib.h" #include "msgQLib.h" #include "sysLib.h" #include "wdLib.h" #include "stdio.h" #define TASK_LOWER_PRIORITY 150 #define TASK_HIGHER_PRIORITY 100 #define TASK_STACK_SIZE 5000 #define WATCHDOG_TIME 2 char *msg="I have right"; LOCAL MSG_Q_ID msgQId=NULL; LOCAL WDOG_ID wdId=NULL; LOCAL int tLowPriority=0; LOCAL int tHighPriority=0; void interrupted_task(void); void mask_task(void); void syncISR(int i); void cleanup(); int notDone=TRUE; STATUS progStart(void) { if((msgQId=msgQCreate(1,20,MSG_Q_FIFO))==NULL) { perror("Error in creating msgQ"); cleanup(); return(ERROR); } if((wdId=wdCreate())==NULL) { perror("wdCreate"); cleanup(); return(ERROR); } if(((tLowPriority=taskSpawn("tLowPriority",TASK_LOWER_PRIORITY,0,TASK_STACK_ SIZE, (FUNCPTR)interrupted_task,0,0,0,0,0,0,0,0,0,0))==ERROR) ||((tHighPriority=taskSpawn("tHighPriority",TASK_HIGHER_PRIORITY,0,TASK_STAC K_SIZE, (FUNCPTR)mask_task,0,0,0,0,0,0,0,0,0,0))==ERROR)) { perror("consumerTask:Error in spawning demotask"); cleanup(); return(ERROR); } if(wdStart(wdId,WATCHDOG_TIME,(FUNCPTR)syncISR,0)==ERROR) { perror("wdStart"); cleanup(); return(ERROR); } return OK; } void interrupted_task(void) { FOREVER{ taskDelay(100); printf("\nI havn't been interrupted"); } } void mask_task(void) { char mesg[15]; for(int i=0;i<15;i++) { mesg[i]='i'; } if((msgQReceive(msgQId,mesg,15,WAIT_FOREVER))==ERROR) { perror("Error in receiving the message"); cleanup(); return; } printf("\n I got it!!!!"); cleanup(); } void syncISR(int i) { msgQSend(msgQId,msg,15,WAIT_FOREVER,MSG_PRI_NORMAL); } void cleanup() { printf("\n OVER!!!!"); if(tHighPriority) taskDelete(tHighPriority); if(tLowPriority) taskDelete(tLowPriority); if(msgQId) msgQDelete(msgQId); if(wdId) wdDelete(wdId); } --------------------------- Newsgroups: comp.os.vxworks Subject: Re: BOOT_LINE_ADDR and downloadable project Date: Tue, 4 Dec 2001 12:24:15 +0200 From: "Ofer Goren" Organization: Verio Message-ID: References: <3c0bb67d$0$8348$4c41069e@reader0.ash.ops.us.uu.net> Hi. I think you missed the point. It's the same file used for both projects. I need somehow to "know" inside the file, in which type of project it "participates" and according th this, to know which .h file to include.... - -- Ofer Goren Iamba Technologies "Gary M" wrote in message news:3c0bb67d$0$8348$4c41069e@reader0.ash.ops.us.uu.net... > Wouldn't this work? In your bootable project (somewhere): > > char *pBootLineAddr = BOOT_LINE_ADDR; > > In the downloadable project: > > extern char *pBootLineAddr; > > > "Ofer Goren" wrote in message > news:newscache$f20sng$5i7$1@lnews.actcom.co.il... > > hi > > I want to use BOOT_LINE_ADDR in my file. I have no problem if the file is > > part of bootable project, since it is defined in prjParams.h. However, if > > I'm adding the file to a d/l-able project, there is no prjParams.h file, > so > > I have to include config.h. > > So, what do I need to do so that a d/l-able project will include config.h, > > and a bootable project will include prjParams.h (which can be different > from > > config.h). Is there a define I can use that define something different in > > case of a bootable/dl-able project? > > > > Thank, > > > > -- > > Ofer Goren > > Iamba Technologies > > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: BOOT_LINE_ADDR and downloadable project Date: Tue, 4 Dec 2001 12:29:37 +0200 From: "Ofer Goren" Organization: Verio Message-ID: References: <3c0c8e51$1@netnews.web.de> And this is a very good point to ask: What is the VPATH macro used for? What is it different then the regular -I compilation var? Thanks, - -- Ofer Goren Iamba Technologies "Reinhold Wirth" wrote in message news:3c0c8e51$1@netnews.web.de... > "Ofer Goren" wrote: > >I want to use BOOT_LINE_ADDR in my file. I have no problem if the file is > >part of bootable project, since it is defined in prjParams.h. However, if > >I'm adding the file to a d/l-able project, there is no prjParams.h file, so > >I have to include config.h. > >So, what do I need to do so that a d/l-able project will include config.h, > >and a bootable project will include prjParams.h (which can be different from > >config.h). Is there a define I can use that define something different in > >case of a bootable/dl-able project? > > > * Are you sure, that it is possible use different values of > BOOT_LINE_ADR in config.h and prjParams.h? > > > * You may use the variable sysBootLine instead BOOT_LINE_ADRS. > ( in sysLib.c it is called: > char *sysBootLine = BOOT_LINE_ADRS) > > > * If you want to include config.h in your d/l-able project > vou have to: > - set VPATH (in Builds; Properties; Macros) to > $(TGT_DIR)/config/all $(TGT_DIR)/config/<"your_BSP"> > - add the path of config.h and configAll.h to the compiler > (in Builds; Properties; C/C++ compiler) add: > -Id:/tornado/target/config/all > -Id:/tornado/target/config/<"your_BSP"> > > > > > -- > __________________________________________________________ > News suchen, lesen, schreiben mit http://newsgroups.web.de --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado 2, PPC Cumulative Patch 4 Date: 4 Dec 2001 11:32:27 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: <3C06AFD7.CB2C9D7A@nunya.com> <3C0C72C7.A2BE4A2D@nunya.com> nunya wrote: > Thanks! We'll try to pry that out of WRS. Something else we found, the > END_OBJ > passed in to our receive routine is _not_ the same END_OBJ that was > passed back to > us from the muxBind call. Because we were validating the END_OBJ, and > they didn't > match, we were ignoring the packet. If we don't validate the END_OBJ > passed in, we > work. This is, for now, a work-around, but we will pursue with WRS as a > bug in their > mux code.... A problem with the Network Protocol Toolkit, that is incorporated in T2CP4 is the fact that the cookie returned by muxBind is not the END pointer anymore. It is not possible to recreate the cookie value, when you only have the name of the device. So be carefull when moving to NPT/T2CP4. 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 === --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Tue Dec 4 07:22:04 2001 From: "Christian Delez" Date: Tue Dec 4 07:22:06 PST 2001 Subject: RE: Docs for developing BSP. Hi, BSP development are documented only in a book. There is no electronic documentation. That's what WRS told us 2 months ago. To get this book, you must order the BSP development kit through your local WRS saler. There are two level of BSP dev kit: - BSP Developer's Base Kit This kit consists of full documentation that includes information about implementing a variety of drivers - for example, polled mode serial I/O, serial I/O, Ethernet, Fast Ethernet, SCSI-1, SCSI-2, timer, and timestamp. Template BSPs and architectural considerations are also included to help a developer reduce the time required to develop a BSP. - BSP Developer's Kit Value Option This package consists of all elements of the Base Kit and all sections of the Tornado Drivers Kit, which contains all available source code for Ethernet, interrupt controller, SCSI, timer, serial, and other common general drivers that have been developed by Wind River Systems. @+, Christian From vxwexplo-errs@csg.lbl.gov Tue Dec 4 12:31:56 2001 From: Donald McLachlan Date: Tue Dec 4 12:31:58 PST 2001 Subject: adding network card to vxWorks for PC I created a new project based on vxWorks for pcPentium. I then excluded: END attach interface END interface support I included: BSD attach interface (which auto included BSD interface support) ELT netif driver target shell I then did: build->dependancies build->Rebuild All volcheck (Solaris host) cp default/vxWorks /floppy/floppy0/vxWorks.st (Solaris host) eject floppy (Solaris host) I moved the floppy to the PC and booted vxWorks from the floppy: Attaching interface lo0...done VxWorks Copyright ... CPU: PC PENTIUM VxWorks: 5.4.2 BSP version: 1.2/0 Creation Date: Dec 4 2001 WDB Ready -> 1) Why is it not attaching the elt interface? 2) I tried using the target shell; why am I getting the following error? x = 7 can't add 'x' to system table - error = 0x3d0001 [ I was able to use control-x to reboot vxWorks so the target shell appears to sort-of be working. Oh, and the help/? command was unknown.] Thanks, Don From vxwexplo-errs@csg.lbl.gov Wed Dec 5 04:03:22 2001 From: Vxworks Exploder Date: Wed Dec 5 04:03:25 PST 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Wed Dec 5 04:03:17 PST 2001 Subject: Re: =?ISO-8859-1?Q?BOOT=5FLINE=5FADDR?= and downloadable project Subject: Sandpont MCP8245 does not work. Subject: How to invoke Polled function in END driver? Subject: Re: BOOT_LINE_ADDR and downloadable project Subject: Re: ZINC / global pointer for ZafApplication Subject: Re: BOOT_LINE_ADDR and downloadable project Subject: Re: Could anyone tell me why the following program doesn't work as I have expected? Thanks Subject: fei82559 hangs on tx Subject: Problem receiving broadcast when binding to a ip address instead of 0.0.0.0 Subject: Re: fei82559 hangs on tx Subject: Re: Problem receiving broadcast when binding to a ip address instead of 0.0.0.0 Subject: Re: IP_DO_FOWRARDING & IP_REDIRECT? Subject: Re: Strange Behavior of wtxtcl Subject: Re: IP_DO_FOWRARDING & IP_REDIRECT? Subject: Unix/NT hosts with identical targets vs. upgrading the Unix to NT Subject: OSE Delta Performance Stats Subject: How do you patch files in VxWorks??? Subject: Re: OSE Delta Performance Stats Subject: Re: How do you patch files in VxWorks??? Subject: Re: OSE Delta Performance Stats Subject: TCP Behavior on Abnormal Socket Close Subject: VxSim serial support? Subject: Is there a routeGet? Subject: Re: "System Variables" setup in Windows98? Subject: Re: Docs for developing BSP. Subject: Ping Problem (Between Board and PC) Subject: Our system will use a card reader Subject: Need Communication of COM port through Serial Port in Simulator Subject: Re: Strange Behavior of wtxtcl Subject: Re: ZINC / global pointer for ZafApplication Subject: Re: Unix/NT hosts with identical targets vs. upgrading the Unix to NT Subject: How and when MUX calls "ioctl" function in END network driver? Subject: Re: TCP Behavior on Abnormal Socket Close Subject: Re: VxSim serial support? Subject: Re: How to Add to Tornado Prototyper "System Variables" in Windows98 Subject: VxWorks interview question ! Subject: VxWorks interview question ! Subject: Re: Our system will use a card reader Subject: Re: Ping Problem (Between Board and PC) Subject: Maximum number of ethernets for VxWorks? Subject: Re: How and when MUX calls "ioctl" function in END network driver? Subject: WindRiver ATM SAR Driver for MPC8260 Subject: vmBaseStateSet and MPC860 ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: =?ISO-8859-1?Q?BOOT=5FLINE=5FADDR?= and downloadable project Date: 04 Dec 2001 13:02:49 +0200 From: Reinhold Wirth Organization: WEB.DE Message-ID: <3c0cbb6b@netnews.web.de> References: <3c0c8e51$1@netnews.web.de> "Ofer Goren" wrote: >And this is a very good point to ask: What is the VPATH macro used for? What >is it different then the regular -I compilation var? > VPATH is used for the "make"-tool(see GNU Make). The entry -I..... is for the compiler. If no VPATH is set you get an error like make: *** No rule to make target `config.h', needed by `test.o'. - -- __________________________________________________________ News suchen, lesen, schreiben mit http://newsgroups.web.de --------------------------- Newsgroups: comp.os.vxworks Subject: Sandpont MCP8245 does not work. Date: Tue, 04 Dec 2001 20:58:48 +0900 From: Yasushi Okada Organization: Fujifilm Software Co.,Ltd Message-ID: <3C0CBA78.7040504@ffs.fujifilm.co.jp> Hi, I am developping BOOTROM for Sandpoint MCP8245 board. VxWorks Version is 5.4. I am in trouble in a first step. - -Decrementer does not work Sandpoint Board sets EXTROM mode. So, I set PICR1[DEC]=1 But Decrementer does not work. And VxWorks Auto-Boot Prompt will not count down. Does anyone have suggestions/comments? Please let me know. Any comments/suggestions will be pleaseed to me. - - Yasushi Okada --------------------------- Newsgroups: comp.os.vxworks Subject: How to invoke Polled function in END driver? Date: 4 Dec 2001 04:20:28 -0800 From: binagpatel@yahoo.com (Bina Patel) Organization: http://groups.google.com/ Message-ID: <65b68b90.0112040420.56da9015@posting.google.com> Hello! I am working with an END driver,i want to implement my END driver in "polled mode". For transmission and reception of any type of data, i have written endPollSend,endPollReceive routine. But i dont know how to invoke poll function in END driver. I shall be gratefull for any link on the same. thanks. Bina. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: BOOT_LINE_ADDR and downloadable project Date: Tue, 4 Dec 2001 14:43:12 +0200 From: "Ofer Goren" Organization: Verio Message-ID: References: <3c0c8e51$1@netnews.web.de> <3c0cbb6b@netnews.web.de> So, the VPATH helps me in specifying seach directory, so I will not have to specify them by -I? What's the different? If an error such as "make: *** No rule to make target `config.h', needed by `test.o'." will echoed, I will add -I to my compiler flags. What do I need the VPATH for? Thanks, - -- Ofer Goren Iamba Technologies "Reinhold Wirth" wrote in message news:3c0cbb6b@netnews.web.de... > "Ofer Goren" wrote: > >And this is a very good point to ask: What is the VPATH macro used for? What > >is it different then the regular -I compilation var? > > > VPATH is used for the "make"-tool(see GNU Make). > The entry -I..... is for the compiler. > > > If no VPATH is set you get an error like > make: *** No rule to make target `config.h', needed by `test.o'. > > -- > __________________________________________________________ > News suchen, lesen, schreiben mit http://newsgroups.web.de --------------------------- Newsgroups: comp.os.vxworks Subject: Re: ZINC / global pointer for ZafApplication Date: Tue, 04 Dec 2001 06:49:12 -0600 From: PC User Organization: ExecPC Internet - Milwaukee, WI Message-ID: <3c0cc8ce$0$65156$272ea4a1@news.execpc.com> References: <3c0bdaca$0$43568$272ea4a1@news.execpc.com> <6bUO7.37141$P%2.3046155@news.bora.net> What does this have anything to do with patching teqniques? Would you explain please. Thanks. ChanSeong Kim wrote: > Hi all vxworkers and zinc applicationer(?)s... > > global pointer for ZafApplication is zafApplication. > > I think zafApplication is a global pointer for zinc application you can > access this pointer > everywhere in your program if you include apropriate(right) header file for > this. > > Following is a source of zinc. This is a constructor of ZafApplication > class. > > You can see zafApplication is assigned with pointer of class 'this'. > > And another global variables can be shown like zafEventManager, > zafWindowManager. > > Good Luck..!! > > ZafApplication::ZafApplication(int _argc, char **_argv) : > argc(0), argv(ZAF_NULLP(ZafIChar *)), error(ZAF_ERROR_NONE) > { > #if defined(ZAF_INDEPENDENT) > // This call is used to reinitialized global variables each time > // the application is run. (Otherwise the program would need to > // be re-loaded for globals to be reinitialized.) > InitializeGlobals(); > #endif > > // Initialize the data manager. > zafDataManager = new ZafDataManager; > > // Call I18nAllocate() to reflect changes based on the display. > ZafI18nData::I18nAllocate(); > > // Allocate the display. > zafDisplay = new ZafScreenDisplay(_argc, _argv); > > #if defined(ZAF_INDEPENDENT) > // If the display didn't initialize, exit gracefully. > if (!((ZafScreenDisplay *)zafDisplay)->initialized) > { > SetError(ZAF_ERROR_CONSTRUCTOR); > printf("Couldn't initialize the screen display.\n"); > return; > } > #endif > > // Allocate the arguments. While this code technically allows argv[0] to > // be NULL (which violates the language definition), this condition > // should not occur because the data passed to us should always be valid > // (i.e., _argv[0] should contain the program name). If the data is > // invalid, it is likely a compiler error, or at least is a problem > // that cannot be fixed here so we do our best to work with it by allowing > // for the invalid state. > > // Initialize argc here since an error may have occurred before this point. > argc = _argc; > argv = new ZafIChar *[argc + 1]; > for (int i = 0; i < argc; i++) > { > #if defined(USE_ZAF_CODE_SET_DATA) > argv[i] = zafCodeSet->ConvertToZafString(_argv[i]); > #else > argv[i] = strdup(_argv[i]); > #endif > } > argv[argc] = ZAF_NULLP(ZafIChar); > > // Create search path from the program path. > #if defined(USE_ZAF_SEARCH_PATH) > if (argv[0]) > zafSearchPath = new ZafPath(argv[0], true); > else > zafSearchPath = new ZafPath(ZAF_NULLP(ZafIChar), true); > #endif > > #if defined(USE_ZAF_PERSISTENCE) > // Call ResetI18n() to reflect changes based on the storage. > // zafI18nStorage = new ZafStorage(ZAF_ITEXT("i18n.znc"), ZAF_FILE_READ); > zafI18nStorage = new ZafStorage(i18n_znc); > > if (zafI18nStorage->Error()) > { > delete zafI18nStorage; > zafI18nStorage = ZAF_NULLP(ZafFileSystem); > } > else > { > zafDataPersistence = new ZafDataPersistence(zafI18nStorage, 0); > zafDataPersistence->AddDataConstructor(ZafLanguageData::className, > ZafLanguageData::classID, &ZafLanguageData::Read); > zafDataPersistence->AddDataConstructor(ZafLocaleData::className, > ZafLocaleData::classID, &ZafLocaleData::Read); > zafDataPersistence->AddDataConstructor(ZafCodeSetData::className, > ZafCodeSetData::classID, &ZafCodeSetData::Read); > ZafI18nData::ResetI18n(); > } > #endif > > // Allocate the event manager. > zafEventManager = new ZafEventManager; > InitMouse(); > InitKeyboard(); > zafEventManager->Add(new ZafCursor); > InitHelpTips(); > > // Allocate the window manager. > zafWindowManager = new ZafWindowManager; > > // Allocate the error system. > InitErrorSystem(); > > // Initialize the global pointer to the application object. > zafApplication = this; > } > > ----- Original Message > > ZafApplication::ZafApplication(int _argc, char **_argv) : > argc(0), argv(ZAF_NULLP(ZafIChar *)), error(ZAF_ERROR_NONE) > { > #if defined(ZAF_INDEPENDENT) > // This call is used to reinitialized global variables each time > // the application is run. (Otherwise the program would need to > // be re-loaded for globals to be reinitialized.) > InitializeGlobals(); > #endif > > // Initialize the data manager. > zafDataManager = new ZafDataManager; > > // Call I18nAllocate() to reflect changes based on the display. > ZafI18nData::I18nAllocate(); > > // Allocate the display. > zafDisplay = new ZafScreenDisplay(_argc, _argv); > > #if defined(ZAF_INDEPENDENT) > // If the display didn't initialize, exit gracefully. > if (!((ZafScreenDisplay *)zafDisplay)->initialized) > { > SetError(ZAF_ERROR_CONSTRUCTOR); > printf("Couldn't initialize the screen display.\n"); > return; > } > #endif > > // Allocate the arguments. While this code technically allows argv[0] to > // be NULL (which violates the language definition), this condition > // should not occur because the data passed to us should always be valid > // (i.e., _argv[0] should contain the program name). If the data is > // invalid, it is likely a compiler error, or at least is a problem > // that cannot be fixed here so we do our best to work with it by allowing > // for the invalid state. > > // Initialize argc here since an error may have occurred before this point. > argc = _argc; > argv = new ZafIChar *[argc + 1]; > for (int i = 0; i < argc; i++) > { > #if defined(USE_ZAF_CODE_SET_DATA) > argv[i] = zafCodeSet->ConvertToZafString(_argv[i]); > #else > argv[i] = strdup(_argv[i]); > #endif > } > argv[argc] = ZAF_NULLP(ZafIChar); > > // Create search path from the program path. > #if defined(USE_ZAF_SEARCH_PATH) > if (argv[0]) > zafSearchPath = new ZafPath(argv[0], true); > else > zafSearchPath = new ZafPath(ZAF_NULLP(ZafIChar), true); > #endif > > #if defined(USE_ZAF_PERSISTENCE) > // Call ResetI18n() to reflect changes based on the storage. > // zafI18nStorage = new ZafStorage(ZAF_ITEXT("i18n.znc"), ZAF_FILE_READ); > zafI18nStorage = new ZafStorage(i18n_znc); > > if (zafI18nStorage->Error()) > { > delete zafI18nStorage; > zafI18nStorage = ZAF_NULLP(ZafFileSystem); > } > else > { > zafDataPersistence = new ZafDataPersistence(zafI18nStorage, 0); > zafDataPersistence->AddDataConstructor(ZafLanguageData::className, > ZafLanguageData::classID, &ZafLanguageData::Read); > zafDataPersistence->AddDataConstructor(ZafLocaleData::className, > ZafLocaleData::classID, &ZafLocaleData::Read); > zafDataPersistence->AddDataConstructor(ZafCodeSetData::className, > ZafCodeSetData::classID, &ZafCodeSetData::Read); > ZafI18nData::ResetI18n(); > } > #endif > > // Allocate the event manager. > zafEventManager = new ZafEventManager; > InitMouse(); > InitKeyboard(); > zafEventManager->Add(new ZafCursor); > InitHelpTips(); > > // Allocate the window manager. > zafWindowManager = new ZafWindowManager; > > // Allocate the error system. > InitErrorSystem(); > > // Initialize the global pointer to the application object. > zafApplication = this; > } --------------------------- Newsgroups: comp.os.vxworks Subject: Re: BOOT_LINE_ADDR and downloadable project Date: Tue, 4 Dec 2001 14:04:12 +0100 From: "Michael Lawnick" Organization: Buergernetz Dillingen Message-ID: <9uihkd$fts$1@snoopy.bndlg.de> References: <3c0c8e51$1@netnews.web.de> <3c0cbb6b@netnews.web.de> Reply-To: "Michael Lawnick" Sender: mlawnick@pd95248e7.dip.t-dialin.net Hi Ofer, "Ofer Goren" schrieb im Newsbeitrag news:newscache$2vltng$kx9$1@lnews.actcom.co.il... > So, the VPATH helps me in specifying seach directory, so I will not have to > specify them by -I? What's the different? If an error such as "make: *** No > rule to make target `config.h', needed by `test.o'." will echoed, I will > add -I to my compiler flags. What do > I need the VPATH for? > you should read ! Reinhold wrote: > > VPATH is used for the "make"-tool(see GNU Make). i.e. generating and solving dependencies in makefile > > The entry -I..... is for the compiler. i.e to find the included text in your source file HTH Michael --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Could anyone tell me why the following program doesn't work as I have expected? Thanks Date: Tue, 4 Dec 2001 14:35:48 +0100 From: "Michael Lawnick" Organization: Buergernetz Dillingen Message-ID: <9uijfl$h2r$1@snoopy.bndlg.de> References: <9ui6ha$no5$1@mail.cn99.com> Reply-To: "Michael Lawnick" Sender: mlawnick@pd95248e7.dip.t-dialin.net Hi zhaoyandong, change your msgQSend(msgQId,msg,15,WAIT_FOREVER,MSG_PRI_NORMAL); to msgQSend(msgQId,msg,15,NO_WAIT,MSG_PRI_NORMAL); and you will get your 'got it'. For your final cleanup, its not a good practice to delete the tasks, but you should signal them via BOOL, semaphore or signal to stop. HTH - -- Mit freundlichen Grüßen, Michael Lawnick ============================================== SOFTEC GmbH Tel +49-731-96600-0 Promenade 17 Fax +49-731-96600-23 D-89073 Ulm Michael Lawnick Germany lawnick@softec.de ============================================== "zhaoyandong" schrieb im Newsbeitrag news:9ui6ha$no5$1@mail.cn99.com... > #include "vxWorks.h" > file://#include "semLib.h' > #include "taskLib.h" > #include "msgQLib.h" > #include "sysLib.h" > #include "wdLib.h" > #include "stdio.h" > > #define TASK_LOWER_PRIORITY 150 > #define TASK_HIGHER_PRIORITY 100 > #define TASK_STACK_SIZE 5000 > #define WATCHDOG_TIME 2 > > char *msg="I have right"; > > LOCAL MSG_Q_ID msgQId=NULL; > LOCAL WDOG_ID wdId=NULL; > LOCAL int tLowPriority=0; > LOCAL int tHighPriority=0; > > void interrupted_task(void); > void mask_task(void); > void syncISR(int i); > void cleanup(); > > int notDone=TRUE; > > STATUS progStart(void) > { > if((msgQId=msgQCreate(1,20,MSG_Q_FIFO))==NULL) > { > perror("Error in creating msgQ"); > cleanup(); > return(ERROR); > } > > if((wdId=wdCreate())==NULL) > { > perror("wdCreate"); > cleanup(); > return(ERROR); > } > > > if(((tLowPriority=taskSpawn("tLowPriority",TASK_LOWER_PRIORITY,0,TASK_STACK_ > SIZE, > (FUNCPTR)interrupted_task,0,0,0,0,0,0,0,0,0,0))==ERROR) > > ||((tHighPriority=taskSpawn("tHighPriority",TASK_HIGHER_PRIORITY,0,TASK_STAC > K_SIZE, > (FUNCPTR)mask_task,0,0,0,0,0,0,0,0,0,0))==ERROR)) > { > perror("consumerTask:Error in spawning demotask"); > cleanup(); > return(ERROR); > } > > if(wdStart(wdId,WATCHDOG_TIME,(FUNCPTR)syncISR,0)==ERROR) > { > perror("wdStart"); > cleanup(); > return(ERROR); > } > > return OK; > } > > void interrupted_task(void) > { > FOREVER{ > taskDelay(100); > printf("\nI havn't been interrupted"); > } > } > > void mask_task(void) > { > char mesg[15]; > for(int i=0;i<15;i++) > { > mesg[i]='i'; > } > if((msgQReceive(msgQId,mesg,15,WAIT_FOREVER))==ERROR) > { > perror("Error in receiving the message"); > cleanup(); > return; > } > printf("\n I got it!!!!"); > cleanup(); > } > > void syncISR(int i) > { > msgQSend(msgQId,msg,15,WAIT_FOREVER,MSG_PRI_NORMAL); > } > > void cleanup() > { > printf("\n OVER!!!!"); > if(tHighPriority) > taskDelete(tHighPriority); > if(tLowPriority) > taskDelete(tLowPriority); > if(msgQId) > msgQDelete(msgQId); > if(wdId) > wdDelete(wdId); > } > > --------------------------- Newsgroups: comp.os.vxworks Subject: fei82559 hangs on tx Date: Tue, 04 Dec 2001 14:25:30 GMT From: brede.thurmann@kongsberg-simrad.com (BT) Message-ID: <3c0cdc78.605875953@news.eunet.no> We have an PPC750 with two Intel82559ER network controllers. The driver is fei82559End. With high load on one of the networks the driver stops transmitting. The other network is OK. I can use WindView etc. for debugging. I have found that this happens when the SCB status word indicates CU status = LPQ active. If I ensures that I do NOT write anything to the CFD's while the SCB status/CU status is not in suspend the driver does not hang on tx anymore. I do not like this solution as I have to trhow away network packets before they are given to the controller. And since I have free CFDs I should be able to buffer up more packets. Has anybody come across this problem? Or even better give me a fix? BT --------------------------- Newsgroups: comp.os.vxworks Subject: Problem receiving broadcast when binding to a ip address instead of 0.0.0.0 Date: 4 Dec 2001 07:08:23 -0800 From: jeremy.coker@bae.co.uk (jeremy coker) Organization: http://groups.google.com/ Message-ID: <6b30ab88.0112040708.5edea3a4@posting.google.com> Hi, A curious thing, the program I have written can send/receive broadcast so long as the receive end socket is bound to inaddr_any if i bind it to a specific (valid+local) interface address then the socket is bound o.k. and shows up in inetstatShow but cannot receive the broadcast. Also I need to turn off broadcast on a recieve socket, but not by opening or using a different port i.e. broadcast send would use the same port as a client socket but I need to actually turn off the broadcast receive on the interface itself. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: fei82559 hangs on tx Date: Tue, 4 Dec 2001 17:25:55 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <3c0cdc78.605875953@news.eunet.no> I had a similar problem with the FEI BSD driver two years ago. If I remember correctly, you dont touch the transmit chain while the Tx is active, but you queue up the new packets in a new Command list, which is made active once the MAC is done with the current list. I can probably dig up my diffs, but I am not sure how applicable these are since I did this on teh BSD and you are uding the END drivers. - - Leonid "BT" wrote in message news:3c0cdc78.605875953@news.eunet.no... > We have an PPC750 with two Intel82559ER network controllers. The > driver is fei82559End. > > With high load on one of the networks the driver stops transmitting. > The other network is OK. I can use WindView etc. for debugging. > > I have found that this happens when the SCB status word indicates > CU status = LPQ active. If I ensures that I do NOT write anything > to the CFD's while the SCB status/CU status is not in suspend the > driver does not hang on tx anymore. > > I do not like this solution as I have to trhow away network packets > before they are given to the controller. And since I have free CFDs > I should be able to buffer up more packets. > > Has anybody come across this problem? Or even better give me a fix? > > BT --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Problem receiving broadcast when binding to a ip address instead of 0.0.0.0 Date: Tue, 4 Dec 2001 17:51:20 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <6b30ab88.0112040708.5edea3a4@posting.google.com> "jeremy coker" wrote in message news:6b30ab88.0112040708.5edea3a4@posting.google.com... > Hi, > > A curious thing, the program I have written can send/receive broadcast > so long as the receive end socket is bound to inaddr_any if i bind it > to a specific (valid+local) interface address then the socket is bound o.k. > and shows up in inetstatShow but cannot receive the broadcast. > Damm right! Binding a socket to a particular IP address is synonimous to requesting only packets desitned to that particular IP address to be received. If you only whish to receive IP Multicast pasckets, you can bind to that multicast IP address (in addition to joining the group). > > Also I need to turn off broadcast on a recieve socket, but not by opening or > using a different port > > i.e. broadcast send would use the same port as a client socket but I need to > actually turn off the broadcast receive on the interface itself. Unfortunately the Berkeley sockets API does not let you get the destination IP address for each packet received (only source), so you can't filter them on a per-packet basis, but you can use bind() as I explained above to effectively filter broadcasts too. - - Leonid --------------------------- Newsgroups: comp.os.vxworks Subject: Re: IP_DO_FOWRARDING & IP_REDIRECT? Date: Tue, 04 Dec 2001 16:46:24 +0000 From: Owain Phillips Organization: Siemens Inc. Message-ID: <3C0CFDE0.43CC6D83@siemenscomms.co.uk> References: <3C0BB5FC.3970CA8E@siemenscomms.co.uk> This is a multi-part message in MIME format. - --------------724C989DA8881CBB382F801A Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Thanks Leonid, You confirmed what I supected and since then I have seen a SENS handbook that this is the default configuration for the stack.......Is this a WRS problem assuming that all devices they are building are routers???? The search is now on for the best place to override this default setting for generation of Vxworks bootrom, Tornaso I type Vxworks build and for project kernel builds. For bootroms I appear to be able to override the IP_FLAGS_DFLT in bsp.h. For project tool kernel builds the change in bsp.h shouldbe picked up. But for the standard torando I type build of Vxworks (Msdos cmd line build) of the core BSP, the order of includes seems to mean, there is no way round but to modify either usrExtra.c or usrNetwork.c to override.......As netLib.h is included after config.h/bsp.h..... Oh well.....Well done WRS. Thanks for your help Leonid, Kind regards, Owain Leonid Rosenboim wrote: > IP_DO_FORWARDING = 1 means that a packet coming in through one > interface may be transmitted out of another interface, as called for by the > current routing table - a.k.a. IP Routing. > IP_REDIRECT is a feature of IP that can generate a ICMP_REDIRECT > message to a sender, which tries to use the node as router, but this nodes > "knows" there is a "better" route for this packet. This ICMP_REDIRECT > will typically modify the initiator's routing table so that subsequent > packets > will be send using the "better route". > > So it makes sense to torun off both if the VxWorks has several interfaces > but is not supposed to act as a Router. > > Hope this helps, > - Leonid > > "Owain Phillips" wrote in message > news:3C0BB5FC.3970CA8E@siemenscomms.co.uk... > > Anyone tell me what these flags do? > > > > Apparently our ethernet interface is configured with these flags on and > > that was the state as we received the BSP / Vxworks. > > > > Are these flags normal default settings for a normal IP endpoint? > > > > Thanks in advance, > > Owain > > - --------------724C989DA8881CBB382F801A 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 - --------------724C989DA8881CBB382F801A-- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Strange Behavior of wtxtcl Date: 4 Dec 2001 08:52:51 -0800 From: brycenga@wire-less-inc.com (Brock) Organization: http://groups.google.com/ Message-ID: <3d71cba2.0112040852.65869e5f@posting.google.com> References: <3d71cba2.0112032223.4f7b9743@posting.google.com> Patrick wrote in message news:... > On 3 Dec 2001, Brock wrote: > > > In my makefile, exists the command: > > > > c:/tornado/host/x86-win32/bin/nmppc snt_mux.partialImage.o | > > c:/tornado/host/x86-win32/bin/wtxtcl > > c:/tornado/host/src/hutils/munch.tcl -asm ppc > snt_ctdt.c > > > > This command fails with the following error: > > > > 'c:' is not recognized as an internal or external command, operable > > program or batch file. > > Because make does not recognise the second line as a continuation > of the first line, try escaping the carriage return on the first line: > > c:/tornado/host/x86-win32/bin/nmppc snt_mux.partialImage.o | \ > c:/tornado/host/x86-win32/bin/wtxtcl > c:/tornado/host/src/hutils/munch.tcl -asm ppc > snt_ctdt.c > > (Note the '\' character at the end of the first line.) > > Regards > > Patrick > > ---------------------------------------------------------------------- > strip 'to_undies' to reply Thanks for the reply, Patrick. However, that isn't the problem. The command is actually executed on one whole line. After posting the message, it was wrapped by the newsgroup app. The command isn't actually split up on separate lines. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: IP_DO_FOWRARDING & IP_REDIRECT? Date: Tue, 4 Dec 2001 19:09:09 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <3C0BB5FC.3970CA8E@siemenscomms.co.uk> <3C0CFDE0.43CC6D83@siemenscomms.co.uk> I am sure you will eventually get around the config issues, I also think the project tool is a pain in the .. In some cases the order of inclusion is wrong, not allowing you to override ANY define from target//config.h This default configuration is for the sake of backwards compatibility, and for those folks using VME backplanet network for multiple VxWorks boards in it, and the fist CPU acting as router between the VME backplane network and the Ethernet port. Years back people doing these things where regarded as key customers. Things have changed since, but backward compatinility keeps reminding you of the past. - - Leonid "Owain Phillips" wrote in message news:3C0CFDE0.43CC6D83@siemenscomms.co.uk... > Thanks Leonid, > > You confirmed what I supected and since then I have seen a SENS handbook that > this is the default configuration for the stack.......Is this a WRS problem > assuming that all devices they are building are routers???? > > The search is now on for the best place to override this default setting for > generation of Vxworks bootrom, Tornaso I type Vxworks build and for project > kernel builds. > > For bootroms I appear to be able to override the IP_FLAGS_DFLT in bsp.h. > > For project tool kernel builds the change in bsp.h shouldbe picked up. > > But for the standard torando I type build of Vxworks (Msdos cmd line build) of > the core BSP, the order of includes seems to mean, there is no way round but to > modify either usrExtra.c or usrNetwork.c to override.......As netLib.h is > included after config.h/bsp.h..... > > Oh well.....Well done WRS. > > Thanks for your help Leonid, > Kind regards, > Owain > > Leonid Rosenboim wrote: > > > IP_DO_FORWARDING = 1 means that a packet coming in through one > > interface may be transmitted out of another interface, as called for by the > > current routing table - a.k.a. IP Routing. > > IP_REDIRECT is a feature of IP that can generate a ICMP_REDIRECT > > message to a sender, which tries to use the node as router, but this nodes > > "knows" there is a "better" route for this packet. This ICMP_REDIRECT > > will typically modify the initiator's routing table so that subsequent > > packets > > will be send using the "better route". > > > > So it makes sense to torun off both if the VxWorks has several interfaces > > but is not supposed to act as a Router. > > > > Hope this helps, > > - Leonid > > > > "Owain Phillips" wrote in message > > news:3C0BB5FC.3970CA8E@siemenscomms.co.uk... > > > Anyone tell me what these flags do? > > > > > > Apparently our ethernet interface is configured with these flags on and > > > that was the state as we received the BSP / Vxworks. > > > > > > Are these flags normal default settings for a normal IP endpoint? > > > > > > Thanks in advance, > > > Owain > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Unix/NT hosts with identical targets vs. upgrading the Unix to NT Date: 4 Dec 2001 09:30:10 -0800 From: ben_mushik@yahoo.com (Ben) Organization: http://groups.google.com/ Message-ID: Hey Everyone, Here is my situation: I have two identical targets. One is at a different work location and the host will and must be changed to NT. The target I am working on has a working Unix host with tested S/W on the target. Is it feasible for me to continue development with Unix and generate builds for use which will work with the NT host and target? Little or no development will be made with the NT host. Or should I upgrade my host to NT also? My backgound: I've worked a few months developing Application Software using Unix Tornado 1.0. I am just unsure of higher level details of the Tornado architecture which leads me to ask: The target should be independent of the host right? Upgrading the host won't change anything in the target? A build I make on a Unix host should work the same when it is downloaded wth the NT host. Have any of you done this (worked with two hosts)? Were there any problems (e.g. maybe the tools from the two host were saying something different that made solving a problem difficult?) I've read some posts about people upgrading to an NT host. I am more familiar with Unix but I am more concerned with problems encountered with the upgrade. If I go this route, what problems did any of you have. Did it take long or were there problems with the development tools? Did it impact your software? Thanks, Ben --------------------------- Newsgroups: comp.realtime,comp.arch.embedded,comp.os.vxworks,comp.os.psos,comp.os.qnx Subject: OSE Delta Performance Stats Date: Tue, 04 Dec 2001 19:52:56 GMT From: "Jim Montville" Organization: Nokia Message-ID: Does anyone here have any performance data on OSE/Delta? Specifically, I'm looking for interrupt latency and task switching time. Thanks in advance. Jim Montville --------------------------- Newsgroups: comp.os.vxworks Subject: How do you patch files in VxWorks??? Date: Tue, 04 Dec 2001 14:13:22 -0600 From: PC User Organization: ExecPC Internet - Milwaukee, WI Message-ID: <3c0d3003$0$30971$272ea4a1@news.execpc.com> Hi, I have tried to find something about how to do patching in VxWorks programmer's guide and nothing is listed in the index under patching. I am trying to test and see if VxWorks provides the mean for an application program, during execution, to patch itself or another application program. DoesVxWorks provide a tool to do this? What I would like to do is to have a task that handles a patch request through a message queue where other tasks and users will request patches to be made. Can it be done interactively meaning while an executable is running? I would like to change a data table contents without having to build the whole thing again. Thanks. --------------------------- Newsgroups: comp.realtime,comp.arch.embedded,comp.os.vxworks,comp.os.psos,comp.os.qnx Subject: Re: OSE Delta Performance Stats Date: 4 Dec 2001 20:19:45 GMT From: "Mark A. Odell" Organization: Embedded Firmware Design, Inc. Message-ID: References: Followup-To: comp.realtime,comp.arch.embedded "Jim Montville" wrote in news:sQ9P7.1783$NB2.42444@news2.nokia.com: > Does anyone here have any performance data on OSE/Delta? Specifically, > I'm looking for interrupt latency and task switching time. Thanks in > advance. I think it would help if we knew the CPU and CPU speed, don't you? - -- - - Mark A. Odell - - Embedded Firmware Design, Inc. - - http://www.embeddedfw.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How do you patch files in VxWorks??? Date: 04 Dec 2001 15:39:16 -0500 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: References: <3c0d3003$0$30971$272ea4a1@news.execpc.com> Sender: bpringlemeir@DeadDuck >>>>> "PC" == PC User writes: [snip] PC> I would like to change a data table contents without having to PC> build the whole thing again. Your request is a little odd. You could patch a table by using the symLib to find the address of the data table. Then you can over-write the memory. This assumes several things. One, you do not have an MMU that is protecting the data table. Two, that the data table is in RAM or some other writeable storage. You could just as easily provide some other mechanism (in your code) that would take some input and write it to this table. I don't understand exactly how you wish to get the patch to your image. I think that you should read the vxWorks Programmer's Guide to understand more about vxWorks. hth, Bill Pringlemeir. - -- Exercise. Don't forget to draw. Listen to your child tomorrow. Never watch Wheel of Fortune without staying where you are first. You'll later be happy. Your lucky numbers are 20, 30, and 40. vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" --------------------------- Newsgroups: comp.realtime,comp.arch.embedded,comp.os.vxworks,comp.os.psos,comp.os.qnx Subject: Re: OSE Delta Performance Stats Date: Tue, 04 Dec 2001 21:50:38 GMT From: "Jim Montville" Organization: Nokia Message-ID: References: Ooops...sorry, I dumbed out. I have a 266 Mhz PPC405. Jim "Mark A. Odell" wrote in message news:Xns916D9B69D283Elkj562ghjgk1k245lbvj@130.133.1.4... > "Jim Montville" wrote in > news:sQ9P7.1783$NB2.42444@news2.nokia.com: > > > Does anyone here have any performance data on OSE/Delta? Specifically, > > I'm looking for interrupt latency and task switching time. Thanks in > > advance. > > I think it would help if we knew the CPU and CPU speed, don't you? > > -- > - Mark A. Odell > - Embedded Firmware Design, Inc. > - http://www.embeddedfw.com --------------------------- Newsgroups: comp.os.vxworks Subject: TCP Behavior on Abnormal Socket Close Date: 4 Dec 2001 23:16:22 GMT From: claveman@cod.nosc.mil (Charles H. Sampson) Organization: Celerity, Inc. Message-ID: <9ujlg6$21v$1@newpoisson.nosc.mil> The system I'm working on has some pretty serious real-time constraints. It communicates with a number of external devices over networks. The explana- tion for this apparent contradiction is that the data only have to be reasona- bly timely but we can only spend a little time processing them. Because of this, we're concerned about using TCP, concerned that if a transmission error occurs then VxWorks might take the machine away from our application code at some unpredictable moment, for some unspecified period of time, to execute a retry. I've been running some tests to see if these concerns are reasonable, particularly tests that check abnormal conditions. One of these has uncovered what looks to me like incorrect behavior. In this test, I have a TCP server that sends messages to a client. The messages are just 4-byte integers that increment from 1 by 1, allowing the client to verify what it's getting. After five messages, the server simply closes its socket and stops transmitting. Up to this point, the client has been receiving messages as expected, a small fraction of a second after they were sent. When the server socket is closed, the client begins receiving mes- sages of length zero. My tests indicate that these phantom messages are ar- riving about every 100-150 msec, at the low end of that range when the client and server are running on the same machine, at the high end when they are run- ning on different, but nearby, machines. To me, this is clearly incorrect. I can't see any justification for re- ceiving messages that were never sent but then BSD is largely a collection of arcania. Is there something I should know that I don't? If there is, is there any way to defend against it? Charlie - -- ****** For an email response, my user name is "csampson" and my host is "csc.com". --------------------------- Newsgroups: comp.os.vxworks Subject: VxSim serial support? Date: Wed, 05 Dec 2001 00:57:55 GMT From: "Chad Krueger" Organization: Excite@Home - The Leader in Broadband http://home.com/faster Message-ID: I'm new to VxSim and am curious whether serial IO is possible through the simulator or whether I limited to console and pipe devices. The latter seems to be the case, but I need to begin testing some device IO minus hardware. Any tips? --------------------------- Newsgroups: comp.os.vxworks Subject: Is there a routeGet? Date: 4 Dec 2001 17:44:30 -0800 From: jim.jollota@strixsys.com (Jim) Organization: http://groups.google.com/ Message-ID: Is there a way (within VxWorks) to interrogate a routing table to determine if a particular route has been added? I need to call this function from within my code, and I need an address or a structure returned, so I can not use routeShow. Specifically, I want to determine if a default gateway has already been configured before I overwrite it with the hard-coded default. Thanks, Jim --------------------------- Newsgroups: comp.os.vxworks Subject: Re: "System Variables" setup in Windows98? Date: 4 Dec 2001 19:02:09 -0800 From: umeshs@myw.ltindia.com (Umesh Satyanarayana) Organization: http://groups.google.com/ Message-ID: References: tiejunyang You can't set enviroment variable in the Win98 like the way you do in winNT or win2000. Copy the FM license key in a notepad and save as WRSlicense.lic in the above mentioned path.Try also saving FM license key as WRSlicense.txt file edit autoexec.bat Add following code in autoexec.bat and execute it set path=c:\torando\.wind\license\WRSlicense.lic Cheers Umesh --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Docs for developing BSP. Date: 4 Dec 2001 19:19:13 -0800 From: umeshs@myw.ltindia.com (Umesh Satyanarayana) Organization: http://groups.google.com/ Message-ID: References: <27203e86.0112032150.27dfe19b@posting.google.com> Kamala, You have to purchase the development kit (you prefer the target processor according to the need) thru distributors n/w of windriver. Suppose you buy MPC823 board and they provide you the BSP development kit( with all drivers for peripherals on the board and docs). If you think of developing your own board built around other PowerPC family processor, you can use above mentioned BSP dev.. kit for developing new BSP Please contact arrow electronics ltd in bangalore for more details. Tel: 558 1520 Cheers Umesh --------------------------- Newsgroups: comp.os.vxworks Subject: Ping Problem (Between Board and PC) Date: 4 Dec 2001 19:34:22 -0800 From: shwzheng@hotmail.com (Shunwu) Organization: http://groups.google.com/ Message-ID: Greetings VxWorkers, I am wondering if anyone has seen this before: Our software enviroment is Vxworks Tornado II,and hardware CPU is Mot MPC860T.There is a problem with fast Ethernet interface of our board.We connect the PC and the board. I set PC's IP address 202.112.108.51 and Board's address 202.112.108.48. We use the ping utility to test the performance of our network interface. Here is the problem. Under the windows mode ,I can always ping through the board but the responce time is extremely long (about 1400ms ~1500ms).We just call it "51 Ping 48". Using the NetXray(a network analysis tool) ,we can see there are ICMP packets between 51 and 48. But when I launch the shell and boot the tornado ,I want to ping the PC , it is called "48 ping 51" .There is no answer but using the NetXray we can see there are UDP and ICMP packets between 48 and 51. It shows that the "request and reply" mechanism works well. Now we can see that in our project the "Ping" works in the unidirectional way.That is to say 51 can ping through 48 but 48 cannot ping through 51. Any help appreciated, --------------------------- Newsgroups: comp.os.vxworks Subject: Our system will use a card reader Date: Wed, 5 Dec 2001 11:31:21 +0800 From: "zhaoyandong" Organization: Bentium Ltd. (CN99) Message-ID: <9uk4p4$nde$1@mail.cn99.com> But I don't know how to program with it, who can give me a sample with it? Thanks --------------------------- Newsgroups: comp.os.vxworks Subject: Need Communication of COM port through Serial Port in Simulator Date: 4 Dec 2001 21:01:21 -0800 From: manibala@covansys.com (B.Manivannan) Organization: http://groups.google.com/ Message-ID: <23de5426.0112042101.1daaf7be@posting.google.com> Good Day Viewer, I need to develop an application which should communicate COM port through serial port in Tornado II simulator.(Note: Presently i am not using the MPC860 Target, only the simulator) The requirement is, 1. I would send some information to the COM1, some process has to be done, and should read back. I wrote the following code with the guidance of our friend Mr.Mike, void Start(void) { /* add application specific code here */ int nStatus1 = 1, nStatus2 = 8; int nOpenReturn = 676, nWriteReturn = 1; char data[] = {'a', 'b', 'c', 'd', 'e', 'f', 'g', 'h', 'j'}; nOpenReturn = open("/tyCo/0",2,0x777); nStatus1 = ioctl( nOpenReturn, FIOBAUDRATE, 9600 ); nStatus2 = ioctl( nOpenReturn, FIOSETOPTIONS, OPT_RAW); nWriteReturn = write(nOpenReturn, data, 10); } In the console based Simulator window it showed the output as "abcdefghij" I having big doubt here... 1. We are going to download this application into target. Will this serial communcation will work with target MPC860 and Serial Hardware(Modem or printer or etc)?????? OR 2. Can i see the port communication in my PC itself?? It means, can i connect the hardware to my PC by running the VxWorks Simulator????(i don't think so) With Thanks and Regards, B.Manivannan --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Strange Behavior of wtxtcl Date: 4 Dec 2001 22:03:00 -0800 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0112042203.5a384f6@posting.google.com> References: <3d71cba2.0112032223.4f7b9743@posting.google.com> brycenga@wire-less-inc.com (Brock) wrote in message news:<3d71cba2.0112032223.4f7b9743@posting.google.com>... > In my makefile, exists the command: > > c:/tornado/host/x86-win32/bin/nmppc snt_mux.partialImage.o | > c:/tornado/host/x86-win32/bin/wtxtcl > c:/tornado/host/src/hutils/munch.tcl -asm ppc > snt_ctdt.c > > This command fails with the following error: > > 'c:' is not recognized as an internal or external command, operable > program or batch file. > > The only way I can get this line to run properly is if I enter as > follows: > > nmppc snt_mux.partialImage.o | wtxtcl > c:/tornado/host/src/hutils/munch.tcl -asm ppc > snt_ctdt.c You could try c:\tornado\host\x86-win32\bin\nmppc - I suspect it is a quirk of the Windows GNU tools and the pathname separators, though I could be wrong (I don't tend to use Windoze for more than surfing, and even then I find it to be rather unreliable :( ) The reason that the munch.tcl file can be specified with forward slashes is that wtxtcl is interpreting that, and it either needs forward slashes, or, if you are lucky, supports either (I suspect it will do either). > Could anyone explain to me why the first command line won't execute > properly? HTH, John... > > > Many thanks, > Brock --------------------------- Newsgroups: comp.os.vxworks Subject: Re: ZINC / global pointer for ZafApplication Date: 4 Dec 2001 22:05:18 -0800 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0112042205.439daddf@posting.google.com> References: <3c0bdaca$0$43568$272ea4a1@news.execpc.com> <6bUO7.37141$P%2.3046155@news.bora.net> <3c0cc8ce$0$65156$272ea4a1@news.execpc.com> Hello, PC User wrote in message news:<3c0cc8ce$0$65156$272ea4a1@news.execpc.com>... > What does this have anything to do with patching teqniques? Would you explain > please. Nothing, but why should it have anything to do with patching techniques? It seems to be a thread discussing Zinc - a graphical environment for VxWorks... > Thanks. You're welcome ;-) John... > ChanSeong Kim wrote: > > > Hi all vxworkers and zinc applicationer(?)s... > > > > global pointer for ZafApplication is zafApplication. > > > > I think zafApplication is a global pointer for zinc application you can > > access this pointer > > everywhere in your program if you include apropriate(right) header file for > > this. > > > > Following is a source of zinc. This is a constructor of ZafApplication > > class. > > > > You can see zafApplication is assigned with pointer of class 'this'. > > > > And another global variables can be shown like zafEventManager, > > zafWindowManager. > > > > Good Luck..!! > > > > ZafApplication::ZafApplication(int _argc, char **_argv) : > > argc(0), argv(ZAF_NULLP(ZafIChar *)), error(ZAF_ERROR_NONE) > > { > > #if defined(ZAF_INDEPENDENT) > > // This call is used to reinitialized global variables each time > > // the application is run. (Otherwise the program would need to > > // be re-loaded for globals to be reinitialized.) > > InitializeGlobals(); > > #endif > > > > // Initialize the data manager. > > zafDataManager = new ZafDataManager; > > > > // Call I18nAllocate() to reflect changes based on the display. > > ZafI18nData::I18nAllocate(); > > > > // Allocate the display. > > zafDisplay = new ZafScreenDisplay(_argc, _argv); > > > > #if defined(ZAF_INDEPENDENT) > > // If the display didn't initialize, exit gracefully. > > if (!((ZafScreenDisplay *)zafDisplay)->initialized) > > { > > SetError(ZAF_ERROR_CONSTRUCTOR); > > printf("Couldn't initialize the screen display.\n"); > > return; > > } > > #endif > > > > // Allocate the arguments. While this code technically allows argv[0] to > > // be NULL (which violates the language definition), this condition > > // should not occur because the data passed to us should always be valid > > // (i.e., _argv[0] should contain the program name). If the data is > > // invalid, it is likely a compiler error, or at least is a problem > > // that cannot be fixed here so we do our best to work with it by allowing > > // for the invalid state. > > > > // Initialize argc here since an error may have occurred before this point. > > argc = _argc; > > argv = new ZafIChar *[argc + 1]; > > for (int i = 0; i < argc; i++) > > { > > #if defined(USE_ZAF_CODE_SET_DATA) > > argv[i] = zafCodeSet->ConvertToZafString(_argv[i]); > > #else > > argv[i] = strdup(_argv[i]); > > #endif > > } > > argv[argc] = ZAF_NULLP(ZafIChar); > > > > // Create search path from the program path. > > #if defined(USE_ZAF_SEARCH_PATH) > > if (argv[0]) > > zafSearchPath = new ZafPath(argv[0], true); > > else > > zafSearchPath = new ZafPath(ZAF_NULLP(ZafIChar), true); > > #endif > > > > #if defined(USE_ZAF_PERSISTENCE) > > // Call ResetI18n() to reflect changes based on the storage. > > // zafI18nStorage = new ZafStorage(ZAF_ITEXT("i18n.znc"), ZAF_FILE_READ); > > zafI18nStorage = new ZafStorage(i18n_znc); > > > > if (zafI18nStorage->Error()) > > { > > delete zafI18nStorage; > > zafI18nStorage = ZAF_NULLP(ZafFileSystem); > > } > > else > > { > > zafDataPersistence = new ZafDataPersistence(zafI18nStorage, 0); > > zafDataPersistence->AddDataConstructor(ZafLanguageData::className, > > ZafLanguageData::classID, &ZafLanguageData::Read); > > zafDataPersistence->AddDataConstructor(ZafLocaleData::className, > > ZafLocaleData::classID, &ZafLocaleData::Read); > > zafDataPersistence->AddDataConstructor(ZafCodeSetData::className, > > ZafCodeSetData::classID, &ZafCodeSetData::Read); > > ZafI18nData::ResetI18n(); > > } > > #endif > > > > // Allocate the event manager. > > zafEventManager = new ZafEventManager; > > InitMouse(); > > InitKeyboard(); > > zafEventManager->Add(new ZafCursor); > > InitHelpTips(); > > > > // Allocate the window manager. > > zafWindowManager = new ZafWindowManager; > > > > // Allocate the error system. > > InitErrorSystem(); > > > > // Initialize the global pointer to the application object. > > zafApplication = this; > > } > > > > ----- Original Message > > > > ZafApplication::ZafApplication(int _argc, char **_argv) : > > argc(0), argv(ZAF_NULLP(ZafIChar *)), error(ZAF_ERROR_NONE) > > { > > #if defined(ZAF_INDEPENDENT) > > // This call is used to reinitialized global variables each time > > // the application is run. (Otherwise the program would need to > > // be re-loaded for globals to be reinitialized.) > > InitializeGlobals(); > > #endif > > > > // Initialize the data manager. > > zafDataManager = new ZafDataManager; > > > > // Call I18nAllocate() to reflect changes based on the display. > > ZafI18nData::I18nAllocate(); > > > > // Allocate the display. > > zafDisplay = new ZafScreenDisplay(_argc, _argv); > > > > #if defined(ZAF_INDEPENDENT) > > // If the display didn't initialize, exit gracefully. > > if (!((ZafScreenDisplay *)zafDisplay)->initialized) > > { > > SetError(ZAF_ERROR_CONSTRUCTOR); > > printf("Couldn't initialize the screen display.\n"); > > return; > > } > > #endif > > > > // Allocate the arguments. While this code technically allows argv[0] to > > // be NULL (which violates the language definition), this condition > > // should not occur because the data passed to us should always be valid > > // (i.e., _argv[0] should contain the program name). If the data is > > // invalid, it is likely a compiler error, or at least is a problem > > // that cannot be fixed here so we do our best to work with it by allowing > > // for the invalid state. > > > > // Initialize argc here since an error may have occurred before this point. > > argc = _argc; > > argv = new ZafIChar *[argc + 1]; > > for (int i = 0; i < argc; i++) > > { > > #if defined(USE_ZAF_CODE_SET_DATA) > > argv[i] = zafCodeSet->ConvertToZafString(_argv[i]); > > #else > > argv[i] = strdup(_argv[i]); > > #endif > > } > > argv[argc] = ZAF_NULLP(ZafIChar); > > > > // Create search path from the program path. > > #if defined(USE_ZAF_SEARCH_PATH) > > if (argv[0]) > > zafSearchPath = new ZafPath(argv[0], true); > > else > > zafSearchPath = new ZafPath(ZAF_NULLP(ZafIChar), true); > > #endif > > > > #if defined(USE_ZAF_PERSISTENCE) > > // Call ResetI18n() to reflect changes based on the storage. > > // zafI18nStorage = new ZafStorage(ZAF_ITEXT("i18n.znc"), ZAF_FILE_READ); > > zafI18nStorage = new ZafStorage(i18n_znc); > > > > if (zafI18nStorage->Error()) > > { > > delete zafI18nStorage; > > zafI18nStorage = ZAF_NULLP(ZafFileSystem); > > } > > else > > { > > zafDataPersistence = new ZafDataPersistence(zafI18nStorage, 0); > > zafDataPersistence->AddDataConstructor(ZafLanguageData::className, > > ZafLanguageData::classID, &ZafLanguageData::Read); > > zafDataPersistence->AddDataConstructor(ZafLocaleData::className, > > ZafLocaleData::classID, &ZafLocaleData::Read); > > zafDataPersistence->AddDataConstructor(ZafCodeSetData::className, > > ZafCodeSetData::classID, &ZafCodeSetData::Read); > > ZafI18nData::ResetI18n(); > > } > > #endif > > > > // Allocate the event manager. > > zafEventManager = new ZafEventManager; > > InitMouse(); > > InitKeyboard(); > > zafEventManager->Add(new ZafCursor); > > InitHelpTips(); > > > > // Allocate the window manager. > > zafWindowManager = new ZafWindowManager; > > > > // Allocate the error system. > > InitErrorSystem(); > > > > // Initialize the global pointer to the application object. > > zafApplication = this; > > } --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Unix/NT hosts with identical targets vs. upgrading the Unix to NT Date: 4 Dec 2001 22:17:34 -0800 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0112042217.10ed9af1@posting.google.com> References: Hello Ben, > Here is my situation: I have two identical targets. One is at a > different work location and the host will and must be changed to NT. > The target I am working on has a working Unix host with tested S/W on > the target. Is it feasible for me to continue development with Unix > and generate builds for use which will work with the NT host and > target? Little or no development will be made with the NT host. Little or none? The target images you create on a unix box can of course be delivered to the target from any host, so use NT as the boot server is not a problem. Just get the ftp server running, and configure the target for ftp booting and that should work OK. You can even use ftp booting from unix if you want to keep the two environments similar. Now, when it comes to using the debugger things might not be so easy. Why? Well, the debug info in the target binaries tends to include paths to your source files. If any of these paths are absolute, you could have problems debugging on a PC. Rebuilding should be OK, though you might have Makefile issues if you have custom makefiles with pathnames in them (even relative ones: the separator changes ;-). There are ways around this by defining make macros to flip the separator when necessary - I think you should find some of these in some of the files under target/h/make. > Or should I upgrade my host to NT also? Personally, I wouldn't... but then I don't think Windoze is worth using, so for me that would be a downgrade ;-) HTH, John... > The target should be independent of the host right? Upgrading the host > won't change anything in the target? A build I make on a Unix host > should work the same when it is downloaded wth the NT host. Have any > of you done this (worked with two hosts)? Were there any problems > (e.g. maybe the tools from the two host were saying something > different that made solving a problem difficult?) > > I've read some posts about people upgrading to an NT host. I am more > familiar with Unix but I am more concerned with problems encountered > with the upgrade. If I go this route, what problems did any of you > have. Did it take long or were there problems with the development > tools? Did it impact your software? > > Thanks, > Ben --------------------------- Newsgroups: comp.os.vxworks Subject: How and when MUX calls "ioctl" function in END network driver? Date: 4 Dec 2001 22:26:41 -0800 From: binagpatel@yahoo.com (Bina Patel) Organization: http://groups.google.com/ Message-ID: <65b68b90.0112042226.7c29ec32@posting.google.com> Hello! I am working with END driver.I want that my driver should work in "polling mode". I have written endPollSend,endPollReceive routine. I know that muxPollSend and muxPollReceive will call endPollSend,endPollReceive routine respectively. But i dont know when and how these functions are called bu "mux". I shall be greateful for any type of link on the same. Bina. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: TCP Behavior on Abnormal Socket Close Date: 4 Dec 2001 22:41:33 -0800 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0112042241.73480bc4@posting.google.com> References: <9ujlg6$21v$1@newpoisson.nosc.mil> Hello, claveman@cod.nosc.mil (Charles H. Sampson) wrote in message news:<9ujlg6$21v$1@newpoisson.nosc.mil>... > The system I'm working on has some pretty serious real-time constraints. > It communicates with a number of external devices over networks. The explana- > tion for this apparent contradiction is that the data only have to be reasona- > bly timely but we can only spend a little time processing them. Because of > this, we're concerned about using TCP, concerned that if a transmission error > occurs then VxWorks might take the machine away from our application code at > some unpredictable moment, for some unspecified period of time, to execute a > retry. Since the retry will happen in the network task, running your code at a higher priority should avoid this. Just remember to give the network task some time or things might not work well ;-) > I've been running some tests to see if these concerns are reasonable, > particularly tests that check abnormal conditions. One of these has uncovered > what looks to me like incorrect behavior. > > In this test, I have a TCP server that sends messages to a client. The > messages are just 4-byte integers that increment from 1 by 1, allowing the > client to verify what it's getting. After five messages, the server simply > closes its socket and stops transmitting. Is that what it is meant to do, or is that part of the problem? > Up to this point, the client has > been receiving messages as expected, a small fraction of a second after they > were sent. When the server socket is closed, the client begins receiving mes- > sages of length zero. My tests indicate that these phantom messages are ar- > riving about every 100-150 msec, at the low end of that range when the client > and server are running on the same machine, at the high end when they are run- > ning on different, but nearby, machines. > > To me, this is clearly incorrect. I can't see any justification for re- > ceiving messages that were never sent but then BSD is largely a collection of > arcania. Is there something I should know that I don't? If there is, is > there any way to defend against it? You don't say, but I assume you are using read() to get the data. There are a couple of things you need to understand about this: 1) TCP is not message based; it is a stream. Don't assume that writing 10 bytes in one end means you'll get 10 bytes out in a single read at the other end. You might; you might not. This often confuses people since IP is packet based, so messages would make sense... 2) When read() returns zero it means end of file, i.e. the connection is gone. Not sure why the machines involved make any difference to the timing though. HTH, John... --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxSim serial support? Date: Wed, 5 Dec 2001 02:37:16 -0500 From: "Jamilur Rahman" Organization: Bell Sympatico Message-ID: References: VxSim is PIPE support only. Forget your serial IO. You could write some "server" like application from your IO module to communicate to the VxSim. Have fun ! "Chad Krueger" wrote in message news:nieP7.22607$%B6.6248581@news2.rdc1.bc.home.com... > I'm new to VxSim and am curious whether serial IO is possible through the > simulator or whether I limited to console and pipe devices. The latter > seems to be the case, but I need to begin testing some device IO minus > hardware. Any tips? > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to Add to Tornado Prototyper "System Variables" in Windows98 Date: Wed, 5 Dec 2001 02:54:41 -0500 From: "Jamilur Rahman" Organization: Bell Sympatico Message-ID: References: Hi Tiejunyang, There is no way to set the environment variable in W-98. It is required in NT. Ignore this and just start your Tornado and enjoy your Prototyper. /jamil, Ottawa. "tiejunyang" wrote in message news:QRWO7.16721$12d.11568@news2.bloor.is... > Help please! > After I got Tornado Prototyper FlexLM license key, I do not know modify > which system file to > "Set the environment variable LM_LICENSE_FILE to point to the path of the > license file" in Windows98, > > Following setup is from Wind River for Windows2000 > > * Bring up the control panel > * Click on the "system" icon > * Click on the "Environment" tab > * Under "System Variables" define a new variable LM_LICENSE_FILE with the > value $WIND_BASE/.wind/license/WRSLicense.lic. > > Are there equation setup in Windows98, which file should be modified? > Thanks, > email:tiejunyang@yahoo.com > > --------------------------- Newsgroups: comp.os.vxworks Subject: VxWorks interview question ! Date: Wed, 5 Dec 2001 03:11:12 -0500 From: "Jamilur Rahman" Organization: Bell Sympatico Message-ID: Place: Ottawa, Ontario, Canada. Event: Job interview for Embedded S/W Designer Interviewer: ..... Candidate: ...... .......... Interviewer: What is the best part in VxWorks ? Candidate: Semaphore. Interview: Why ? Candidate: Because, it is faster and flexible. .......... .......... (Note that, the candidate was not offered the job) Does anybody have some other thought/comment/suggestion on this ? --------------------------- Newsgroups: comp.os.vxworks Subject: VxWorks interview question ! Date: Wed, 5 Dec 2001 03:17:17 -0500 From: "Jamilur Rahman" Organization: Bell Sympatico Message-ID: Place: Ottawa, Ontario, Canada. Event: Job interview for Embedded S/W Designer Interviewer: (question...) Candidate: (answer...) ..... ..... Interviewer: What is the best thing in VxWorks you like most ? Candidate: Semaphore. Interviewer: Why ? Candidate: Because, it is faster and flexible. ..... ..... (note that, The candidate was not offered the job) Any suggestion, comment or thought ? Thanks. Jamil, Ottawa, Ontario. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Our system will use a card reader Date: 5 Dec 2001 08:30:03 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: <9uk4p4$nde$1@mail.cn99.com> "zhaoyandong" wrote: > But I don't know how to program with it, who can give me a sample with > it? Thanks You have to give some more information before we can help you with this. But the best way to get information will be to ask the manufacturer of the device. 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: Ping Problem (Between Board and PC) Date: 5 Dec 2001 08:32:27 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: shwzheng@hotmail.com (Shunwu) wrote: > Here is the problem. > Under the windows mode ,I can always ping through the board but the > responce time is extremely long (about 1400ms ~1500ms).We just call it > "51 Ping 48". Using the NetXray(a network analysis tool) ,we can see > there are ICMP packets between 51 and 48. Take a look at the interrupt routine for your ethernet driver. When this interrupt is not executed you can see something like this. Use WindView to examine the execution of the different tasks and interrupt routines. 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: Maximum number of ethernets for VxWorks? Date: 5 Dec 2001 02:23:08 -0800 From: champu_n@yahoo.com (champak) Organization: http://groups.google.com/ Message-ID: Could any body tell me what is the maximum number of ethernets( especially intel 82559 family) that can be supported by Vxworks? regards --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How and when MUX calls "ioctl" function in END network driver? Date: Wed, 05 Dec 2001 06:01:32 -0500 From: George Varndell Organization: Posted Via Binaries.net = SPEED+RETENTION+COMPLETION = http://www.binaries.net Message-ID: <3C0DFE8C.A7EF553C@varndellengineering.com> References: <65b68b90.0112042226.7c29ec32@posting.google.com> Bina Patel wrote: > > I am working with END driver.I want that my driver should work in > "polling mode". > I have written endPollSend,endPollReceive routine. I know that > muxPollSend and muxPollReceive will call endPollSend,endPollReceive > routine respectively. > But i dont know when and how these functions are called bu "mux". > I shall be greateful for any type of link on the same. > WDB puts the END in polled mode. Polled mode is used for the so-called system level debugging. Look at wdbendpktdrv.c in your Tornado tree. A quick and dirty way to test polled mode in your driver is to connect to the target with target server, bring up the debugger and attach to system instead of a task. Unfortunately, this is a 'go/no-go' test. If polled mode doesn't work, you just loose the connection. Ideally, to debug polled mode of an END requires H/W assisted debug capability. Regards, George Varndell - -- Varndell Engineering, LLC. PPC and XScale Expertise Coming soon! Download free demo versions of our END drivers! http://www.varndellengineering.com ______________________________________________________________________________ Posted Via Binaries.net = SPEED+RETENTION+COMPLETION = http://www.binaries.net --------------------------- Newsgroups: comp.os.vxworks Subject: WindRiver ATM SAR Driver for MPC8260 Date: Wed, 5 Dec 2001 12:04:35 +0100 From: "MVJuhl" Organization: na Message-ID: <9uktml$nnb$1@newstoo.ericsson.se> Does anyone have any experience with the WindRiver ATM SAR Driver for the MPC8260 ? I’m looking for a sample code, which utilise the API for transmitting and receiving frames. I would also very much like a peek at the entire SAR configuration in order to use the sarPQIIChRoute() and that functionality. My task is to verify a hardware prototype by transmitting and receiving ATM frames without involving higher layer protocols. My target (for development) is the Motorola ADS6260 PILOT, and the SW will be ported to my prototype (which is very similar) once the traffic flows. Regards Michael Juhl --------------------------- Newsgroups: comp.os.vxworks Subject: vmBaseStateSet and MPC860 Date: Wed, 05 Dec 2001 12:09:35 +0100 From: Sebastien Fauris Organization: Nortel Networks Message-ID: <3C0E006F.9AB7D478@europem01.nt.com> Reply-To: sfauris@europem01.nt.com Hi Folks, has someone already used the vmBaseStateSet routine on MPC860 ? If so, are you using a patch or an official/recent release of the mmu800Lib/libPPC860gnuvx.a ? I have a lot of problems with this routine on MPC860, I posted a TSR to WRS but I got no reply. Here are the problems: I am trying to use the vmBaseStateSet routine (Tornado 2.0) and in spite of a Wind River patch (old and non-official) on the mmu800Lib I noticed these: - - the page protect attribute: Without the patch, a call to a vmBaseStateSet with VM_STATE_WRITABLE/VM_STATE_WRITABLE_NOT was not working. I have a little debug utility allowing me to get and display the MMU Level 1 and Level 2 descriptors (calling the kernel routines mmu800Lvl1DescAddrGet/mmu800Lvl2DescAddrGet) I could see that the PP ("Page Protection", 10=writable, 11=read-only) bits were not being changed in the MMU level 2 descriptor, so the page was remaining set as writable. The WRS patch corrected this. - - the guarded attribute: I cannot see the guarded bit being set when I display the MMU level 1 descriptor on an area I setup as "guarded". My utility sees fine the other bits I changed with th vmBaseStateSet like PP or CI ("Cache Inhibit"). I am starting to have doubts also on the behaviour of the vmBaseStateSet with VM_STATE_GUARDED/ VM_STATE_GUARDED_NOT - - I notice also that the vmBaseStateSet routine in the patch I got does not seem to perform a tlb invalidate after a modification in the MMU descriptors. Doing it by hand after each call to vmBaseStateSet fix some tlb incoherency problems I met. Thanks for any feedback, Sebastien. --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Wed Dec 5 07:25:03 2001 From: =?GB2312?Q?=E7=DB=CA=BF=BE=D9?= Date: Wed Dec 5 07:25:05 PST 2001 Subject: VxWorks: I am in trouble , the situation is that I use TornadoII to build a bootrom_ucmp for PC pentium Target , after making boot floppy disk , I use it to boot a pc that with an enthernet card 3com905b( PCI), when VxLd1.2 loaded bootrom.sys, I type "n lo0" in command line , the error occured: Attaching network interface lo0...done. Error setting inet address of lo0 to 0, errno=0 Cannot initialize interface named "lo0" --------------------------------------------------- Why this happened ? How to resolve this problem? and after I entered "@" , vxworks output information as below: Starting at 0x0108000... Target Name:VxTarget muxDevLoad failed for device entry 0! Attaching network interface lo0 ...done NFS client support not included wdbConfig:error Configuring WDB communication interface VxWorks copyright: 1984-1998 wind river systems,Inc. CPU: PC PENTIUM VxWorks:5.4 BSPVersion:1.2/0 Creationdate:Dec 4 2001 WDB:Agent configuration failed 0x7ffee7c (tRootTask):muxLoad failed! _ I can't find it out what I did wrong , Waiting for your help , Thank you very much, Regards yaoshiju@21cn.com From vxwexplo-errs@csg.lbl.gov Wed Dec 5 11:06:27 2001 From: Donald McLachlan Date: Wed Dec 5 11:06:29 PST 2001 Subject: Re: adding network card to vxWorks for PC Wind River support solved the target shell problem by suggesting I include "Initialize symbol table" into the vxWorks image. I still cannot get vxWorks to attach the 3C905B interface. I've "included" if_elt both via the GUI tool, and be editing config.h, but when booting it still only reports "Attaching interface lo0...done". Shouldn't it also report "Attaching interface elt0...done"? To be sure it was not attached I included "Network Show Routines", and rebooted. IfShow() only shows lo0; so still no elt0. Any suggestions? Thanks, Don From vxwexplo-errs@csg.lbl.gov Thu Dec 6 01:19:11 2001 From: Shlomo Katz Date: Thu Dec 6 01:19:14 PST 2001 Subject: WAITING FOR MSG while DELAYED hi hi all i am new user to vxworks and i would like to know if a task, that puts in delay state, can move to the ready state when a message is received at its message Q, before the delay period expire? katz sholmi From vxwexplo-errs@csg.lbl.gov Thu Dec 6 04:03:23 2001 From: Vxworks Exploder Date: Thu Dec 6 04:03:26 PST 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Thu Dec 6 04:03:19 PST 2001 Subject: very strange case Subject: Re: How do you patch files in VxWorks??? Subject: Re: very strange case Subject: Re: Docs for developing BSP. Subject: Reverse etherAddrResolve Subject: Re: Docs for developing BSP. Subject: Need info about pSOS+ to VxWorks translation code from Wind River Subject: Detecting socket close Subject: Re: VxWorks interview question ! Subject: Re: TCP Behavior on Abnormal Socket Close Subject: Poor NFS performance. Subject: Re: VxWorks interview question ! Subject: Re: Reverse etherAddrResolve Subject: ppc spi interface Subject: How to download bootable Image into FFS? Subject: Re: How do you patch files in VxWorks??? Subject: Re: Need Communication of COM port through Serial Port in Simulator Subject: Re: How to download bootable Image into FFS? Subject: Can the MPC860 do this ? Subject: Re: How do you patch files in VxWorks??? Subject: Re: Reverse etherAddrResolve Subject: SDS debugger with BDM connection on MPC860 Subject: Re: Closing a socket while blocked in write or read Subject: Re: Poor NFS performance. Subject: Re: Can the MPC860 do this ? Subject: Re: Reverse etherAddrResolve Subject: IP stack question Subject: PPC603 library for PPC604 build Subject: Re: GUI display on top of VxWorks ? Subject: Re: Can the MPC860 do this ? Subject: how to add this news group to my outlook program? Subject: ioctl/raw i/o Subject: memShow and Code Segment size Subject: using very frequent routines Subject: make.PENTIUMgnu Subject: Re: PPC603 library for PPC604 build Subject: Re: Need Communication of Serial Port with MPC860 Target Subject: RADVision H.323 Stack Sample Subject: Re: HELP!!! in DHCP Subject: Re: VxWorks interview question ! Subject: Re: Ping Problem (Between Board and PC) Subject: help me about asynchronous PPP model in the dscc4!!! Subject: Re: make.PENTIUMgnu Subject: Re: ioctl/raw i/o Subject: Re: make.PENTIUMgnu Subject: Re: using very frequent routines Subject: VXWORKS: ftphookup sometimes hangs!!! Subject: Re: Reverse etherAddrResolve Subject: Re: using very frequent routines Subject: Re: how to add this news group to my outlook program? ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: very strange case Date: 5 Dec 2001 05:20:05 -0800 From: people@micro-web.co.kr (JinSoo Kim) Organization: http://groups.google.com/ Message-ID: Hi, My target CPU is NET+ARM 40 made by NetSilicon. OS is VxWorks 5.4. I encountered a very strange case while writing program. I extracted only the part that makes problem. <================ program which makes problem ====================> #include #include unsigned short int var; extern int sysClkRateGet(); void strange_case(void) { int m = 10; var = (unsigned short int) ( ((double) m) * sysClkRateGet() ); } <================================================================> The compile succeeds but there is a error while downloading to the target. "WTX Error 0x10034 (LOADER_RELOCATION_OFFSET_TOO_LARGE)". I can avoid the error by either of these. (1) make var to local variable other than global variable. or (2) change the 'unsigned short int' type casting to 'int' type casting. That is, var = (int) ( ((double) m) * sysClkRateGet() ); or (3) remove sysClkRateGet(). That is, var = (unsigned short int) ((double) m); The sysClkRateGet command in shell returns proper value and that means that target already has sysClkRagetGet symbol and function definition. And if I change the target from NET+ARM 40 to vxsim, there was no error without changing the code. Why this occurs? Thanks in advance. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How do you patch files in VxWorks??? Date: Wed, 05 Dec 2001 07:22:28 -0600 From: PC User Organization: ExecPC Internet - Milwaukee, WI Message-ID: <3c0e221a$0$65163$272ea4a1@news.execpc.com> References: <3c0d3003$0$30971$272ea4a1@news.execpc.com> Hi Bill, Thanks for responding. I have gone through the whole VxWorks manual and found out that there are no mentioning of patching. What I am planning to do is to use the ldModule and unloadModule functions. I would probably let my task to go to sleep enough time so I can edit and compile the new changes. And then when the task wakes up, it will reload the module. Does this sound like patching? Bill Pringlemeir wrote: > >>>>> "PC" == PC User writes: > [snip] > PC> I would like to change a data table contents without having to > PC> build the whole thing again. > > Your request is a little odd. You could patch a table by using the > symLib to find the address of the data table. Then you can over-write > the memory. This assumes several things. One, you do not have an MMU > that is protecting the data table. Two, that the data table is in RAM > or some other writeable storage. > > You could just as easily provide some other mechanism (in your code) > that would take some input and write it to this table. I don't > understand exactly how you wish to get the patch to your image. I > think that you should read the vxWorks Programmer's Guide to > understand more about vxWorks. > > hth, > Bill Pringlemeir. > > -- > Exercise. Don't forget to draw. Listen to your child tomorrow. > Never watch Wheel of Fortune without staying where you are first. > You'll later be happy. Your lucky numbers are 20, 30, and 40. > > vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" --------------------------- Newsgroups: comp.os.vxworks Subject: Re: very strange case Date: 5 Dec 2001 13:43:26 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: people@micro-web.co.kr (JinSoo Kim) wrote: > The compile succeeds but there is a error while downloading to the > target. "WTX Error 0x10034 (LOADER_RELOCATION_OFFSET_TOO_LARGE)". http://www.xs4all.nl/~borkhuis/vxworks/vxw_pt2_ppc.html#2.2.1-A 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: Docs for developing BSP. Date: 5 Dec 2001 05:45:35 -0800 From: christian.delez@ascom.ch (Christian Delez) Organization: http://groups.google.com/ Message-ID: References: <27203e86.0112032150.27dfe19b@posting.google.com> Hi, BSP development are documented only in a book. There is no electronic documentation. That's what WRS told us 2 months ago. To get this book, you must order the BSP development kit through your local WRS saler. There are two level of BSP dev kit: - - BSP Developer's Base Kit This kit consists of full documentation that includes information about implementing a variety of drivers - for example, polled mode serial I/O, serial I/O, Ethernet, Fast Ethernet, SCSI-1, SCSI-2, timer, and timestamp. Template BSPs and architectural considerations are also included to help a developer reduce the time required to develop a BSP. - - BSP Developer's Kit Value Option This package consists of all elements of the Base Kit and all sections of the Tornado Drivers Kit, which contains all available source code for Ethernet, interrupt controller, SCSI, timer, serial, and other common general drivers that have been developed by Wind River Systems. @+, Christian --------------------------- Newsgroups: comp.os.vxworks Subject: Reverse etherAddrResolve Date: 5 Dec 2001 13:47:09 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: Hello, In one system I have the MAC address of a device on the network, and I want to get the IP address of this device. Is this possible? etherAddrResolve does the opposite, finding the MAC-address of a known IP- address, but what I am looking for is just the other way around.. 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: Docs for developing BSP. Date: 5 Dec 2001 14:00:47 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: <27203e86.0112032150.27dfe19b@posting.google.com> christian.delez@ascom.ch (Christian Delez) wrote: > BSP development are documented only in a book. There is no electronic > documentation. The BSP Developer's Kit manual is available in PDF format, I found it sometime ago on the public WRS FTP site. I don't know if they are distributing it, but it was there. So ask your sales rep or FAE again for it. 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: Need info about pSOS+ to VxWorks translation code from Wind River Date: 5 Dec 2001 06:16:45 -0800 From: ghorton@lucent.com (Greg) Organization: http://groups.google.com/ Message-ID: <97d3f155.0112050616.74a5d443@posting.google.com> We are using pSOS+ 2.2 & 2.5 in many projects, and are planning to move to VxWorks for future projects. A possible intermediate step is to obtain a translation layer from Wind River, so we can run the VxWorks OS but keep our code base with pSOS+ calls. I believe this will allow us to use all of the VxWorks Tornado tools, which would be a big step up from what we have with pSOS. Does anybody have any info about how this translation layer works? Any issues with it? Can we use all of the Tornado tools with it? Thanks in advance, Greg --------------------------- Newsgroups: comp.os.vxworks Subject: Detecting socket close Date: Wed, 5 Dec 2001 07:23:56 -0700 From: "Eger George" Organization: MindSpring Enterprises Message-ID: <9ulam2$k76$1@nntp9.atl.mindspring.net> Reply-To: "Eger George" It appears that using either ioctl(sock, FIONREAD, &numBy) or recv(sock, buf, numBy, MSG_PEEK) on a server-side closed socket doesn't return -1, while a read() would. Is there any way to detect server-side socket closure in a timely fashion? TIA --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks interview question ! Date: Wed, 05 Dec 2001 14:41:42 GMT From: "RonnoBonno" Organization: ATT Broadband Message-ID: References: I would have said Windview! Ron "Jamilur Rahman" wrote in message news:TDkP7.17301$iF3.1756651@news20.bellglobal.com... > Place: Ottawa, Ontario, Canada. > Event: Job interview for Embedded S/W Designer > > Interviewer: (question...) > Candidate: (answer...) > ..... > ..... > Interviewer: What is the best thing in VxWorks you like most ? > Candidate: Semaphore. > Interviewer: Why ? > Candidate: Because, it is faster and flexible. > ..... > ..... > (note that, The candidate was not offered the job) > > Any suggestion, comment or thought ? > > Thanks. > Jamil, > Ottawa, Ontario. > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: TCP Behavior on Abnormal Socket Close Date: Wed, 5 Dec 2001 15:04:11 +0000 (UTC) From: don@mars.dgrc.crc.ca (Donald McLachlan) Organization: The unconfigured xvnews people Message-ID: <9uld1b$pap$1@crc-news.crc.ca> References: <9ujlg6$21v$1@newpoisson.nosc.mil> Reply-To: don@mars.dgrc.crc.ca Can you get a (tcpdump) trace of the exchange to analyse? Hard to guess what is happening. It would be easier if we could see it. More than likely you are seeing the FIN and FIN+ACK packets which are terminating the connection. Don - -- Donald McLachlan E-mail Donald.McLachlan@crc.ca Communications Research Centre / RNS Tel (613) 998-2845 3701 Carling Ave., Fax (613) 998-9648 Ottawa, Ontario K2H 8S2 Canada --------------------------- Newsgroups: comp.os.vxworks Subject: Poor NFS performance. Date: 5 Dec 2001 07:07:47 -0800 From: esanborn@snowshore.com (Ed Sanborn) Organization: http://groups.google.com/ Message-ID: First off I am a sysadmin. Not an Engineer or even a user of vxWorks. I believe folks are utilizing NFS from their vxWorks' shell. They execute a script like /snowshore/users/buddy/test.txt When they execute the test.txt script it is found on an NFS mounted partition '/snowshore' which is actually mounted from our NetApp fileserver. Is their anything that we can check on to see why execution of this script is slow? It was fine till I rebooted the NetApp. I'm assuming it may be something on the NetApp but I'm not positive. At any rate, I need more info to see what is happening.If anyone has experience with this I'd appreciate any pointers. I'm already going to start using sniffers so please no flames about basics like this. Thanks. - -Ed --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks interview question ! Date: Wed, 5 Dec 2001 10:14:43 -0500 From: "Yin, Yong" Organization: Nortel Message-ID: <9uldki$afo$1@bcarh8ab.ca.nortel.com> References: various library & dynamic loading Jamilur Rahman wrote in message news:TDkP7.17300$iF3.1756557@news20.bellglobal.com... > Place: Ottawa, Ontario, Canada. > Event: Job interview for Embedded S/W Designer > > Interviewer: ..... > Candidate: ...... > .......... > Interviewer: What is the best part in VxWorks ? > Candidate: Semaphore. > Interview: Why ? > Candidate: Because, it is faster and flexible. > .......... > .......... > (Note that, the candidate was not offered the job) > > Does anybody have some other thought/comment/suggestion on this ? > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Reverse etherAddrResolve Date: Wed, 5 Dec 2001 15:29:08 -0000 From: "Jarl Sigurd" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: Johan Borkhuis wrote in message ... >Hello, > >In one system I have the MAC address of a device on the network, and I want >to get the IP address of this device. Is this possible? >etherAddrResolve does the opposite, finding the MAC-address of a known IP- >address, but what I am looking for is just the other way around.. > >Groeten, > Johan Johan, you need the RARP protocol. Ermm... no handy refernces at hand, but it's very similar to ARP, uses a different ether frame type number, and either google or www.faqs.org will find the RFC for you. Dunno if there's a vx implementation though. DaveK - -- Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! --------------------------- Newsgroups: comp.os.vxworks Subject: ppc spi interface Date: Wed, 05 Dec 2001 10:21:09 -0500 From: david grover Organization: Mailgate.ORG Server - http://www.Mailgate.ORG Message-ID: <3C0E3B65.A051307F@avoidspam.com> Sender: bpatel@fwion.com hello, I am using powerpc 8260. Is there a limit on the maximum sustained spi rate in master mode. My application has the need to send/receive 6 bytes of data. If my CPM freq is 133Mhz and spi freq is 8/16Mhz will it work? according to user manual page 33-6 section 33.3.3 the maximum sustained spi data rate is SYSTEMCLK/50. This multimaster mode section also says that "gaps should be inserted between multiple chaaracters to keep from exceeding the maximum sustained data rate". does this statement apply for "master mode only"? Thanks for your help. --------------------------- Newsgroups: comp.os.vxworks Subject: How to download bootable Image into FFS? Date: 5 Dec 2001 07:26:04 -0800 From: kundesh@covansys.com (Kundesh) Organization: http://groups.google.com/ Message-ID: <133de16f.0112050726.7b8244bb@posting.google.com> Hi Gurus, 1. I had developed a bootable image, which is required to download to target FFS so that whenever I start the target(with host disconnected) the image gets loaded, boots the target and starts running the usrAppInit() function. 2. Please suggest me any steps required to download this botable image to target from Tornado. Do I required to do any settings in Tornado? 3. I think I need to do settings for TrueFFS. Is it right? 4. Presently I had created & rebuild a bootable image. This image had following files in its default directory ctdt.o, dataSegPad.o, linksyms.o, PartialImage.o, prjConfig.o, sysAlib.o, syslib.o, usrAppInit.o, version.o & VxWorks. When I download the bootable application to FFS did I need to copy all these files to FFS?. Is their any command from target shell to copy all these files to FFS? 5. Presently my SBC gets booted from VxWorks.bdx which is on host PC. In the list of above files I didnt find any .bdx file?. Did I required it for any .bdx file booting? If YES then how to generate it? Thanks in advance. Regards, Kundesh --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How do you patch files in VxWorks??? Date: 05 Dec 2001 10:40:48 -0500 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: References: <3c0d3003$0$30971$272ea4a1@news.execpc.com> <3c0e221a$0$65163$272ea4a1@news.execpc.com> Sender: bpringlemeir@DeadDuck >>>>> "PC" == PC User writes: PC> Hi Bill, Thanks for responding. I have gone through the whole PC> VxWorks manual and found out that there are no mentioning of PC> patching. What I am planning to do is to use the ldModule and PC> unloadModule functions. I would probably let my task to go to PC> sleep enough time so I can edit and compile the new changes. And PC> then when the task wakes up, it will reload the module. Does PC> this sound like patching? This does sound like patching. You should watch for some things. Before loading the new module, you should shutdown the task using the taskDelete() function. You should also free any memory and other allocated resources such as message queues, semaphores, etc. If you are communicating with another task (ie through a message queue) and you do not wish for the communication to be interrupted, you must access the pointer/handle through a variable that IS NOT in the `patched' module. You could make a 2nd module that is call `Frob_comvar.o'. For example, [Frob_comvar.c] int frobInit; SEM_ID frobLocker; MSG_Q_ID frobChannel0; MSG_Q_ID frobChannel1; MSG_Q_ID frobChannel2; ... [Frob_comvar.h] extern int frobInit; extern SEM_ID frobLocker; extern MSG_Q_ID frobChannel0; extern MSG_Q_ID frobChannel1; extern MSG_Q_ID frobChannel2; Include the file in your module. This module should have a frobInitialize() and a frobDeinitialize() function [*]. One would create the task and the other would delete it. You can examine/set the `frobInit' value the first time that things run. The next time, the system resources would just be reused. However, you will have to take care that semaphores, etc are in a sane state (not locked by the task that is being patched). [frob.c] #include "Frob_comvar.h" void frobInitialize(void) { if(!frobInit) { semBCreate, msgQCreate, etc } taskCreate... } void frobDeinitialize(void) { /* Task gives this semaphore when it is in a same state. */ semTake(taskDeleteSem,WAIT_FOREVER); taskDelete... } hth, Bill Pringlemeir. [*] Idea shamelessly stolen from the WindML libraries. - -- Do not eat holes in your carpet this week. Loving your dog now will turn out to be unhealthy after a few years. Look for your coworker with one eye. Your mom will think of you as someone you know. Look for it everywhere. vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Need Communication of COM port through Serial Port in Simulator Date: Wed, 05 Dec 2001 15:51:33 GMT From: "Frank" Message-ID: <9orP7.17147$2Fd.16173@news1.bloor.is> References: <23de5426.0112042101.1daaf7be@posting.google.com> Hello, 1. Technically, Your rountie will work at 860board, normally, tyCo/0 is connected to SMC1, depend on your board and BSP. 2. Serial Port is not supported in VxSim. Frank - -- "B.Manivannan" wrote in message news:23de5426.0112042101.1daaf7be@posting.google.com... > Good Day Viewer, > > I need to develop an application which should communicate COM port > through serial port in Tornado II simulator.(Note: Presently i am not > using the MPC860 Target, only the simulator) > > The requirement is, > > 1. I would send some information to the COM1, some process has to be > done, and should read back. > > I wrote the following code with the guidance of our friend Mr.Mike, > > void Start(void) > { > /* add application specific code here */ > int nStatus1 = 1, nStatus2 = 8; > int nOpenReturn = 676, nWriteReturn = 1; > char data[] = {'a', 'b', 'c', 'd', 'e', 'f', 'g', 'h', 'j'}; > > nOpenReturn = open("/tyCo/0",2,0x777); > nStatus1 = ioctl( nOpenReturn, FIOBAUDRATE, 9600 ); > nStatus2 = ioctl( nOpenReturn, FIOSETOPTIONS, OPT_RAW); > > nWriteReturn = write(nOpenReturn, data, 10); > } > > In the console based Simulator window it showed the output as > "abcdefghij" > > I having big doubt here... > > 1. We are going to download this application into target. Will this > serial communcation will work with target MPC860 and Serial > Hardware(Modem or printer or etc)?????? > OR > 2. Can i see the port communication in my PC itself?? It means, can i > connect the hardware to my PC by running the VxWorks Simulator????(i > don't think so) > > With Thanks and Regards, > B.Manivannan --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to download bootable Image into FFS? Date: Wed, 05 Dec 2001 10:48:40 -0500 From: david grover Organization: Mailgate.ORG Server - http://www.Mailgate.ORG Message-ID: <3C0E41D8.BAC5C053@fwion.com> References: <133de16f.0112050726.7b8244bb@posting.google.com> Sender: bpatel@fwion.com Hi kundesh, It seems you want to build a "bootable application". A more flexible approach would be to separate out vxworks image from the application modules which you want to load. Since you are using the TFFS you can do the following - ( the discussing below is for target booting from TFFS having network connectivity) 1) Build a standalone vxworks image (vxWorks.st). Make sure appropriate includes (for network connectivity) are there in config.h, for example, #define STANDALONE_NET #define INCLUDE_TELNET #define INCLUDE_NET_SHOW #define INCLUDE_PING #define INCLUDE_TELNET /* if you want to telnet to target */. copy this vxworks image to the TFFS supported ROM directory. Make suren dosFS options are properly configured to support file names in the rom. For the standalone target, if you want symbol table included do the following, #define INCLUDE_STANDALONE_SYM_TBL #define INCLUDE_SYM_TBL_SHOW 2) Make sure to change boot parameters to reflect boot device as "tffs", for exampe, boot device: tffs=0,0. Make sure userNetwork.c has appropriate compare for tffs using "params.bootDev", if not include the strncmp for that for example, strncmp(params.bootDev, "tffs", 4) == 0). 3) Load the application modules from the TFFS supported ROM using startup script in rom.. Make sure config.h has the following - #define INCLUDE_STARTUP_SCRIPT. hth, bhavesh Kundesh wrote: > Hi Gurus, > > 1. I had developed a bootable image, which is required to > download to target FFS so that whenever I start the target(with host > disconnected) the image gets loaded, boots the target and starts > running the usrAppInit() function. > > 2. Please suggest me any steps required to download this botable > image to target from Tornado. Do I required to do any settings in > Tornado? > > 3. I think I need to do settings for TrueFFS. Is it right? > > 4. Presently I had created & rebuild a bootable image. This image had > following files in its default directory ctdt.o, dataSegPad.o, > linksyms.o, PartialImage.o, prjConfig.o, sysAlib.o, syslib.o, > usrAppInit.o, version.o & VxWorks. When I download the bootable > application to FFS did I need to copy all these files to FFS?. Is > their any command from target shell to copy all these files to FFS? > > 5. Presently my SBC gets booted from VxWorks.bdx which is on host PC. > In the list of above files I didnt find any .bdx file?. Did I required > it for any .bdx file booting? If YES then how to generate it? > > Thanks in advance. > > Regards, > Kundesh --------------------------- Newsgroups: comp.os.vxworks Subject: Can the MPC860 do this ? Date: Wed, 05 Dec 2001 15:59:26 GMT From: "Jim" Organization: Magma Communications Ltd. Message-ID: Hello all, I have a question for anyone who dares answer ;-) Currently I have a board running vxWorks with an MPC860 at 80Mhz and a bus frequency of 40 Mhz. I have a new device that I need to communicate with on a chip select ... the device requires ... a "crazy slow" bus speed of less than 1 Mhz ... otherwise it will just plain not work ... So far I (and others) have managed to get the bus speed to around 1 Mhz using the "ebdf" and "dfnh" (these are in chapter 15 of the 860 users manual if you interested) ... when we lower the bus to this speed however ethernet and the serial ports behave "poorly" ... maybe I missed something when I tried this. Currently we are consider only dropping the bus rates when we do an access to this slow device ... to do this we would wrap writes to this address range in a function call ... then in the function shut off interupts, drop the clocks, and then do the memory write/read, then speed the clocks back up, then start interupts back up ... does any one know if this will work? Any help ... or even sympathy would be greatly appreciated ... Jim --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How do you patch files in VxWorks??? Date: Wed, 5 Dec 2001 16:57:56 +0100 From: "Michael Lawnick" Organization: Buergernetz Dillingen Message-ID: <9ulg67$efc$1@snoopy.bndlg.de> References: <3c0d3003$0$30971$272ea4a1@news.execpc.com> <3c0e221a$0$65163$272ea4a1@news.execpc.com> Reply-To: "Michael Lawnick" Sender: mlawnick@pd9524714.dip.t-dialin.net Hi PC User, am I understanding you correct, that patching in your sense means exchanging some modules without target restart ? You will have to take care: - - The module to exchange may not be in the calling sequence (i.e. stack) of any task while changing is done. - - All functions and variables within the code, that are to be called from extern, may not be used by direct reference, but must be referenced indirectly via symbol table. This references must be validated on every single access. - --> will work, but slow down all. A faster way could be to terminate all tasks that use the module and spawn them again after exchange. Still you will have to take care about updating cross-references between exchanged and unexchanged code, but you will have to do it only once after restart. HTH Michael "PC User" schrieb im Newsbeitrag news:3c0e221a$0$65163$272ea4a1@news.execpc.com... > Hi Bill, > > Thanks for responding. I have gone through the whole VxWorks manual and > found out that there are no mentioning of patching. What I am planning to > do is to use the ldModule and unloadModule functions. I would probably > let my task to go to sleep enough time so I can edit and compile the new > changes. And then when the task wakes up, it will reload the module. > Does this sound like patching? > > Bill Pringlemeir wrote: > > > >>>>> "PC" == PC User writes: > > [snip] > > PC> I would like to change a data table contents without having to > > PC> build the whole thing again. > > > > Your request is a little odd. You could patch a table by using the > > symLib to find the address of the data table. Then you can over-write > > the memory. This assumes several things. One, you do not have an MMU > > that is protecting the data table. Two, that the data table is in RAM > > or some other writeable storage. > > > > You could just as easily provide some other mechanism (in your code) > > that would take some input and write it to this table. I don't > > understand exactly how you wish to get the patch to your image. I > > think that you should read the vxWorks Programmer's Guide to > > understand more about vxWorks. > > > > hth, > > Bill Pringlemeir. > > > > -- > > Exercise. Don't forget to draw. Listen to your child tomorrow. > > Never watch Wheel of Fortune without staying where you are first. > > You'll later be happy. Your lucky numbers are 20, 30, and 40. > > > > vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Reverse etherAddrResolve Date: 5 Dec 2001 16:15:15 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: "Jarl Sigurd" wrote: > Johan, you need the RARP protocol. Ermm... no handy refernces at > hand, > but it's very similar to ARP, uses a different ether frame type number, > and either google or www.faqs.org will find the RFC for you. Dunno if > there's a vx implementation though. I saw a couple of email mentioning a RARP server, but I need to find the IP addresses of some PC's I have the MAC address from. Would that be possible using RARP? 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: SDS debugger with BDM connection on MPC860 Date: 5 Dec 2001 17:54:26 CET From: "Roger Bodén" Message-ID: <3c0e5142$1@news.wineasy.se> I'm trying to use the SDS debugger with a BDM connection on an MPC860 board (ads860 BSP). I can download the vxworks image, and start the execution. When I single step, I get a machine check exception. The call stack is sysInit usrInit cacheLibInit The instruction that causes the machine check exception is a branch to the cacheArchLibInit function. I'm using the following environment: vxworks 5.4 SDS 7.5 Raven BDM cable from Macraigor I have successfully executed exactly the same vxworks image, booting from ROM and downloading the image from an ftp server. Do I need to modify the build of the image in some way? Are there any interrupt vectors that SDS and/or the BDM connection uses that must be preserved? Could someone point me in the right direction? What is likely to be wrong? Any help greatly appreciated. - -- Roger Bodén INTERPEAK Box 7498 SE-103 92 Stockholm Sweden Office: +46 8 545 275 80 Fax +46 8 545 275 99 www.interpeak.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Closing a socket while blocked in write or read Date: 5 Dec 2001 11:11:28 -0600 From: "Ilya" Message-ID: <3c0e545b@news.barak.net.il> References: <2413908f.0112030458.6b69c8f1@posting.google.com> May be You forgot to shutdown the socket. Ilya "Nick" wrote in message news:2413908f.0112030458.6b69c8f1@posting.google.com... > Hi, > > I am seeing some strange behaviour: > > Task 1: Blocking socket, call write (which blocks). > Task 2: Close socket (which is blcoked in write). > Task 1: No return from write. > > I thought that when the socket is closed it should return from any > blocked calls to read or write. > > Anyone else come across this? > > Cheers, > > Nick --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Poor NFS performance. Date: Wed, 5 Dec 2001 19:01:23 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: Ed, I know that (at least some) NetApp servers have bad auto-negotioation on their FastEthernet ports, so after reboot, NetApp may have switched to 10Mbps. I worked around this by setting the Catalyst port to fixed 100 Full-Duplex without auto-negotiation, and it is fine between reboots. If the Vxworks target is 100Mbps and NetApp 10Mbps, this mismatch does cause slowness on many protocols, including NFS due to lack of flow control, which results in packet loss and higj-level protocol retransmission. Hope this helps, - - Leonid "Ed Sanborn" wrote in message news:b5e7d641.0112050707.54cd037f@posting.google.com... > First off I am a sysadmin. Not an Engineer or even a user of > vxWorks. > > I believe folks are utilizing NFS from their vxWorks' shell. > They execute a script like /snowshore/users/buddy/test.txt > > When they execute the test.txt script it is found on an NFS > mounted partition '/snowshore' which is actually mounted from > our NetApp fileserver. Is their anything that we can check on > to see why execution of this script is slow? It was fine till > I rebooted the NetApp. I'm assuming it may be something on the > NetApp but I'm not positive. At any rate, I need more info to > see what is happening.If anyone has experience with this I'd > appreciate any pointers. > > I'm already going to start using sniffers so please no flames > about basics like this. > > Thanks. > > -Ed --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Can the MPC860 do this ? Date: Wed, 5 Dec 2001 10:07:07 -0800 From: "Dan Gold" Message-ID: <3c0e62db$0$336$e2e8da3@nntp.cts.com> References: "Jim" wrote in message news:yvrP7.2379$8e.174760@news... > Hello all, > > Currently I have a board running vxWorks with an MPC860 at 80Mhz and a bus > frequency of 40 Mhz. > > I have a new device that I need to communicate with on a chip select ... the > device requires ... a "crazy slow" bus speed of less than 1 Mhz ... > otherwise it will just plain not work ... > > So far I (and others) have managed to get the bus speed to around 1 Mhz > using the "ebdf" and "dfnh" (these are in chapter 15 of the 860 users manual > if you interested) ... when we lower the bus to this speed however ethernet > and the serial ports behave "poorly" ... maybe I missed something when I > tried this. Isn't the CPM clock coming from the bus clock ? So if your bus clock is < 1Mhz, how can the CPM keep up with 10Mb/s ethernet ? > Currently we are consider only dropping the bus rates when we do an access > to this slow device ... to do this we would wrap writes to this address > range in a function call ... then in the function shut off interupts, drop > the clocks, and then do the memory write/read, then speed the clocks back > up, then start interupts back up ... does any one know if this will work? > > Any help ... or even sympathy would be greatly appreciated ... Well, since you get what you pay for and news is free... This doesn't sound like a job for the 860 on a 1Mhz bus, it sounds more like a job for an FPGA with an interface to your crazy device, and a status register the 860 can poll as often as it wants, to see if your device is ready. I.e. put the 860 interface on one side of the FPGA, and the "crazy device" on the other side of the FPGA. Or, could you make the access with a custom UPM table that added *many, many* wait states ? The PPC decrementer (that vxWorks uses for tick timing) counts at 1/4 the bus clock speed. So if you access this device a lot, and keep changing the bus speed, will you be in some sort of "time warp" ? (i.e. time is fleeting, madness takes it's toll...) not even a full $.02 worth, hth, Dan Gold gold@ensemble.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Reverse etherAddrResolve Date: Wed, 05 Dec 2001 20:08:29 GMT From: "Ho-Kuo Chan" Organization: GlobeTrotter Message-ID: <19vP7.12864$tx.918976@carnaval.risq.qc.ca> References: "Johan Borkhuis" wrote in message news:Xns916EAFA6DF91Fborkhuisagerecom@135.7.153.102... > "Jarl Sigurd" wrote: > > Johan, you need the RARP protocol. Ermm... no handy refernces at > > hand, > > but it's very similar to ARP, uses a different ether frame type number, > > and either google or www.faqs.org will find the RFC for you. Dunno if > > there's a vx implementation though. > > I saw a couple of email mentioning a RARP server, but I need to find the IP > addresses of some PC's I have the MAC address from. Would that be possible > using RARP? > Johan, That's what RARP (Reverse Address Resolution Protocol) does. It resolves IP Addresses from MAC Addresses. As Jarl mentioned, I don't know if an RARP daemon is available in VxWorks but check www.faws.org for the RARP specification. HK --------------------------- Newsgroups: comp.os.vxworks Subject: IP stack question Date: Wed, 5 Dec 2001 20:37:51 +0000 (UTC) From: "Dmitri Krivchitch" Organization: Mailgate.ORG Server - http://www.Mailgate.ORG Message-ID: <2e2829c52c744ed22484917ab7cbc466.48926@mygate.mailgate.org> Hello, all. Hope someone can help me. I've created my own SLIP network interface and everything looks fine, but when I try to pass incoming message to IP stack (see code below) /**** s = splimp(); if (IF_QFULL(&ipintrq)) { IF_DROP(&ipintrq); m_freem (pMbuf); return (ERROR); } IF_ENQUEUE(&ipintrq, pMbuf); ipintr (); splx(s); ****/ I got error message like this /**** Exception number 4: Task: 0xfa1adc (t2) Illegal Instruction Status Register : 0x2004 Program Counter : 0x610003ea ****/ and PC for t2 is set to 0. but IP stack receives that message. Any advice would be greatly appreciated. Dmitri - -- Posted from ottawa-hse-ppp254379.sympatico.ca [64.230.3.176] via Mailgate.ORG Server - http://www.Mailgate.ORG --------------------------- Newsgroups: comp.os.vxworks Subject: PPC603 library for PPC604 build Date: 5 Dec 2001 21:56:26 GMT From: asoloway@qualcomm.com (Alan Soloway) Organization: QUALCOMM Inc. Message-ID: <9um56a$3ri$1@coset.qualcomm.com> I have been supplied with a library from a vendor that was built for the PPC603 target. I am using a PPC604 target (Motorola PPC750 card). I am able to compile, link, download, and successfully execute my program but I am concerned that there is a latent incompatibility problem that will arise at the least opportune time (as always :-). Does anybody know the differences between the targets (PPC603 vs. PPC604) that would cause problems? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: GUI display on top of VxWorks ? Date: Wed, 05 Dec 2001 21:57:11 GMT From: Ephraim Gadsby Organization: - minimal Message-ID: References: <3C0C54F3.BE8E2313@home.com> On Tue, 04 Dec 2001 04:47:16 GMT, Gwen & Emory Stagmer wrote: >Just open a TCP/IP port and send the data to a workstation (PC/Mac/*nix) >program, then display it any way you want. Or, if you don't have an >ethernet connection, just print it to a COM port... >The TCP/IP isn't hard at all, and boundless examples exist. But that it is already provided in the form of the target shell; with all the facility of the debugger and the c-interpreter availible for no extra work. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Can the MPC860 do this ? Date: Wed, 05 Dec 2001 14:46:31 -0800 From: Anmol Mishra Organization: http://extra.newsguy.com Message-ID: <3C0EA3C7.EA51FF25@eternal-systems.com> References: Jim wrote: > Hello all, > > I have a question for anyone who dares answer ;-) > > Currently I have a board running vxWorks with an MPC860 at 80Mhz and a bus > frequency of 40 Mhz. > > I have a new device that I need to communicate with on a chip select ... the > device requires ... a "crazy slow" bus speed of less than 1 Mhz ... > otherwise it will just plain not work ... > > So far I (and others) have managed to get the bus speed to around 1 Mhz > using the "ebdf" and "dfnh" (these are in chapter 15 of the 860 users manual > if you interested) ... when we lower the bus to this speed however ethernet > and the serial ports behave "poorly" ... maybe I missed something when I > tried this. > > Currently we are consider only dropping the bus rates when we do an access > to this slow device ... to do this we would wrap writes to this address > range in a function call ... then in the function shut off interupts, drop > the clocks, and then do the memory write/read, then speed the clocks back > up, then start interupts back up ... does any one know if this will work? > > Any help ... or even sympathy would be greatly appreciated ... > > Jim You do have some sympathy, lots in fact. You could also look at sysClkRateSet () to see if it helps you. --------------------------- Newsgroups: comp.os.vxworks Subject: how to add this news group to my outlook program? Date: 5 Dec 2001 16:55:13 -0800 From: mingxinmu@hotmail.com (mumx) Organization: http://groups.google.com/ Message-ID: <75f406a7.0112051655.63199b4c@posting.google.com> thank you ! --------------------------- Newsgroups: comp.os.vxworks Subject: ioctl/raw i/o Date: Wed, 5 Dec 2001 21:40:37 -0800 From: "Hashim Qaderi" Organization: Bell Sympatico Message-ID: Hello, I was trying to build a custom command line interface. I was looking at some of the archived newsgroups' mail, and there are some similar questions but I was not able to find answers ... I have several questions : - - First, is there an easy way to search the newsgroups? - - Does vxworks (by default) start the target shell? From my understanding it does, but I am not sure. - - I was thinking of starting a CLI task (passing in the stdin and stdout (stolen from the shell) - not sure how to do this? and is this possible?) or I could open a device to tty and get a file descriptor for this and then pass that in to the CLI task - - I would also like to completely control the terminal. I don't want it to echo anything, I would like to read any character typed in, then choose the ones I accept (ignore the others), and print out everything but the hot-keys (i.e. esc, tab (for command completion) up and down arrows etc.) At the moment the I only receive characters after typing enter (which is not really helpful when trying to create shell-like functionalities (such as command completion (using tab) or history list display (up-arrow) ). I haven't been able to do this and have tried to set ioctl to raw but I keep getting wierd characters. In summary, I would like to capture input (unbuffered I/O) and decipher the input. After the return has been hit then the original settings should come up. How can this be done? Is there any source code for this? - - Also, how can this CLI be invoked through telnet? Would I need to do sonething to the telnet server code to do this since it also starts target shell by default (in vxworks or is there some code out there?) - - In other words, how do you go back and forth from target shell to custom command line interface? - -How can I use my own command line interface using Telnet, instead of using the target shell? Any help would be greatly appreciated. Thanks, Phil --------------------------- Newsgroups: comp.os.vxworks Subject: memShow and Code Segment size Date: Wed, 05 Dec 2001 21:55:14 -0500 From: Naga R Narayanaswamy Organization: MindSpring Enterprises Message-ID: <3C0EDE12.98E27733@mindspring.com> Hi, I have a vxWorks image size = 12 Meg. However after loading the image, memShow shows 4 Meg as allocated and nearly 250 Meg as free. The system has 256 Meg RAM. Does memShow sows only the data portion and if so where is the 12 Meg of image accounted for. Thanks Naga. --------------------------- Newsgroups: comp.os.vxworks Subject: using very frequent routines Date: 5 Dec 2001 21:27:20 -0800 From: theone@lge.com (Min, Ji-hoon) Organization: http://groups.google.com/ Message-ID: <4b5d36f2.0112052127.7df50b44@posting.google.com> I'm very new to VxWorks. I want to run a certain task every 1us(1e-6 sec) on a 200Mhz CPU. But I heard that the system clock is recommended not to exceed several thousnad Hz. I don't even know how long time VxWorks consumes for its own system tasks and kernel operation. 1) How can this be done? 2) Is it possible to make the 1Mhz system clock? 3) How large is the amount of time that VxWorks needs for its own systematic purpose, with 200Mhz CPU? I really want your help. Thanks. Min,Ji-Hoon. Korea. --------------------------- Newsgroups: comp.os.vxworks Subject: make.PENTIUMgnu Date: Thu, 6 Dec 2001 13:40:12 +0800 From: "zhaoyandong" Organization: Bentium Ltd. (CN99) Message-ID: <9un0mm$nl4$1@mail.cn99.com> I copied a x86 BSP from a friend. But when I create a bootable project based on that BSP, it displays the following error message **makefile:28: D:\TORNADO/target/h/make/make.PENTIUMgnu: No *such file or directory *> make.exe: *** No rule to make target *`D:\TORNADO/target/h/make/make.PENTIUMgnu'. Stop. *> "make makeMacros BSP2PRJ=TRUE" in *D:/TORNADO/target/config/vsbcVxWorks failed Should I copy the file named make.PENTIUMgnu? If you have a copy of that ,can you give me a copy Thanks --------------------------- Newsgroups: comp.os.vxworks Subject: Re: PPC603 library for PPC604 build Date: 5 Dec 2001 22:23:28 -0800 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0112052223.4cde0b1a@posting.google.com> References: <9um56a$3ri$1@coset.qualcomm.com> Hello, For the most part you should be fine. The differences between the 603 and 604 are not usually encountered by application level code; the key difference from a software perspective is the MMU (for the 603, the TLB must be updated by software; for the 604 this is implemented in hardware). Unless your library does something with the MMU registers, or TLB exceptions, I would say that you'll be fine. HTH, John... asoloway@qualcomm.com (Alan Soloway) wrote in message news:<9um56a$3ri$1@coset.qualcomm.com>... > I have been supplied with a library from a vendor that was built for the PPC603 > target. I am using a PPC604 target (Motorola PPC750 card). > > I am able to compile, link, download, and successfully execute my program but I > am concerned that there is a latent incompatibility problem that will arise at > the least opportune time (as always :-). > > Does anybody know the differences between the targets (PPC603 vs. PPC604) that > would cause problems? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Need Communication of Serial Port with MPC860 Target Date: 5 Dec 2001 22:45:20 -0800 From: manibala@covansys.com (B.Manivannan) Organization: http://groups.google.com/ Message-ID: <23de5426.0112052245.72ffee4c@posting.google.com> References: <23de5426.0112042101.1daaf7be@posting.google.com> Mr.Frank, What you said is perfectly correct. I analyzed it. After downloading the application into the target(MPC860), the target must connect to the serial device(for example modem in my case). Like simulator it is showing the output "abcdefghij" in the VxWorks window. My doubt is, if it had been connected the MODEM, it should have sent/received the signal. ( This i tested it through Visual C++ application,modem Tx/Rx lights glowed ) My senior suggested me that i should use the function from sysSerial.c file from C:\Tornado\target\config\estMDB8xx\sysSerial.c. As for as my knowledge is concerned we need to use this file to develop the device driver for the serial hardware. How can i connect the MODEM? Do i need to install the BSP package and use that? Or Is there any configuration to connect the serial port for the MPC860 target?? Could you guide me??? With Thanks and Regards, B.Manivannan. Hello, 1. Technically, Your rountie will work at 860board, normally, tyCo/0 is connected to SMC1, depend on your board and BSP. 2. Serial Port is not supported in VxSim. Frank --------------------------- Newsgroups: comp.os.vxworks Subject: RADVision H.323 Stack Sample Date: Thu, 6 Dec 2001 14:55:20 +0800 From: "zhaoyandong" Organization: Bentium Ltd. (CN99) Message-ID: <9un53i$r6c$1@mail.cn99.com> Our embedded system is going to use Radvision RTP/RTCP Stack. But, I don't know how to use it. I need a sample of using RTP/RTCP routines under VxWorks. thanks. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: HELP!!! in DHCP Date: 6 Dec 2001 00:34:34 -0800 From: shafi4u@rediffmail.com (shafi) Organization: http://groups.google.com/ Message-ID: References: <3c03db4c$1@news.cadence.com> Hi, you will not receive the server address if you use dhcpcBoot* because, the DHCP client will never send a request when it accepts BOOT reply. That is the negotiation has not been completely done. Server responded with just an offer and waiting for request. so you may need complete negotiatin to take place, till you get the DHCPACK, before you could retreive the parameters like server IP address shafi "naveen" wrote in message news:<3c03db4c$1@news.cadence.com>... > Well I am also trying to configure DHCP client in Vxworks. > The network is configured and target gets ip address. > > 1. I am unable to get DNS Server address . > > > Secondly I am using the > dhcpcBootInit > dhcpcBootOptionSet > dhcpcBootBind > > > functions instead of > > > lease = dhcpcInit( net_if, 1 ); > dhcpcEventHookAdd( lease, ... ) > dhcpcOptionSet( lease, ... ); > dhcpcBind( lease, synchronous ); > > to configure the dhcp client. > > Because when I use the latter it fails in BIND. > > > I will really appreciate if anybody can help in this. > Thanx > Naveen --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks interview question ! Date: 6 Dec 2001 01:08:43 -0800 From: umeshs@myw.ltindia.com (Umesh Satyanarayana) Organization: http://groups.google.com/ Message-ID: References: semaphores are available whilst you write drivers on WinNT or UNIX. Nothing special about the semaphore in Vxworks (except PRIORITY INVERSION SAFE argument). Candidate should have said "Real time kernel with rich API's" what does semaphore got to do with faster and flexibility? Umesh --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Ping Problem (Between Board and PC) Date: 6 Dec 2001 01:08:50 -0800 From: shwzheng@hotmail.com (Shunwu) Organization: http://groups.google.com/ Message-ID: References: > Take a look at the interrupt routine for your ethernet driver. When this > interrupt is not executed you can see something like this. > Use WindView to examine the execution of the different tasks and interrupt > routines. > > Groeten, > Johan Thank you Johan You mean that there maybe something wrong with my ethernet driver,but it is provided by the Wind River .Maybe there is some "Configuration" issue about it.And My target server does not support WindView. Shunwu Zheng --------------------------- Newsgroups: comp.os.vxworks Subject: help me about asynchronous PPP model in the dscc4!!! Date: 6 Dec 2001 01:37:05 -0800 From: ljn_98@163.net (Tim) Organization: http://groups.google.com/ Message-ID: hello: I am writting dscc4 driver(a chip of infineon)in a router,but i have met an insolvable problem about chip driver. OS:vxworks TornadoII cpu:MPC8240 problem: dscc4 is connected with other router in asynchronous PPP model through serial port .dscc4 can correctly receive frame from other router,but interframe time fill(continous logical '1'--idle status)are also received by dscc4 as a frame.The data are nomal before enter RXD pin of dscc4,which have 7E open flag,7E close flag,CRC and each byte have 1 start bit,8 data bits,1 stop bit. SCC setup of dscc4: CCR0:0x80001197 CCR1:0x02568880 CCR2:0x08030000 ACCM:0xffffffff UDAC:0x7e7e7e7e so,anyone who have used dscc4 give me some advice,thank you very much! --------------------------- Newsgroups: comp.os.vxworks Subject: Re: make.PENTIUMgnu Date: 6 Dec 2001 09:57:48 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: <9un0mm$nl4$1@mail.cn99.com> "zhaoyandong" wrote: > I copied a x86 BSP from a friend. > But when I create a bootable project based on that BSP, it displays the > following error message > > **makefile:28: D:\TORNADO/target/h/make/make.PENTIUMgnu: No *such file > or directory > *> make.exe: *** No rule to make target > *`D:\TORNADO/target/h/make/make.PENTIUMgnu'. Stop. > *> "make makeMacros BSP2PRJ=TRUE" in > *D:/TORNADO/target/config/vsbcVxWorks failed > > Should I copy the file named make.PENTIUMgnu? If you have a copy of > that ,can you give me a copy Looks like you dont have a license for x86, you have to contact WindRiver to provide you with that license. 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: ioctl/raw i/o Date: 6 Dec 2001 10:05:57 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: "Hashim Qaderi" wrote: > - First, is there an easy way to search the newsgroups? Use Google for this: http://groups.google.com/groups?group=comp.os.vxworks > - Does vxworks (by default) start the target shell? From my > understanding it does, but I am not sure. That can be configured using Tornado. > - I was thinking of starting a CLI task (passing in the stdin and > stdout (stolen from the shell) - not sure how to do this? You cannot combine these 2, when you use your own shell you have to diable the VxWorks shell. > - I would also like to completely control the terminal. Take a look here: ftp://ftp.atd.ucar.edu/pub/archive/vxworks/vx/ You can find a curses library there (vwcurses). > -How can I use my own command line interface using Telnet, instead of > using the target shell? WindRiver has an "unsupported" telnet server that can do this, ask your FAE for it. 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: make.PENTIUMgnu Date: Thu, 6 Dec 2001 18:35:28 +0800 From: "zhaoyandong" Organization: Bentium Ltd. (CN99) Message-ID: <9uni0c$aau$1@mail.cn99.com> References: <9un0mm$nl4$1@mail.cn99.com> "Johan Borkhuis" wrote in message news:Xns916F6FBA3793Cborkhuisagerecom@135.7.153.101... > "zhaoyandong" wrote: > > > I copied a x86 BSP from a friend. > > But when I create a bootable project based on that BSP, it displays the > > following error message > > > > **makefile:28: D:\TORNADO/target/h/make/make.PENTIUMgnu: No *such file > > or directory > > *> make.exe: *** No rule to make target > > *`D:\TORNADO/target/h/make/make.PENTIUMgnu'. Stop. > > *> "make makeMacros BSP2PRJ=TRUE" in > > *D:/TORNADO/target/config/vsbcVxWorks failed > > > > Should I copy the file named make.PENTIUMgnu? If you have a copy of > > that ,can you give me a copy > > Looks like you dont have a license for x86, you have to contact WindRiver > to provide you with that license. > > Groeten, > Johan > It's not so , I 'm lack of the file named "make.PENTIUMgnu" in the directory of $WIND_BASE\target\h\make but not the license file of "WRSLicense.dat" --------------------------- Newsgroups: comp.os.vxworks Subject: Re: using very frequent routines Date: Thu, 6 Dec 2001 11:36:57 +0100 From: "Michael Lawnick" Organization: Buergernetz Dillingen Message-ID: <9unhob$gn2$1@snoopy.bndlg.de> References: <4b5d36f2.0112052127.7df50b44@posting.google.com> Reply-To: "Michael Lawnick" Sender: mlawnick@pd95245cb.dip.t-dialin.net Forget it ! - -- 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 ============================================== "Min, Ji-hoon" schrieb im Newsbeitrag news:4b5d36f2.0112052127.7df50b44@posting.google.com... > I'm very new to VxWorks. > > I want to run a certain task every 1us(1e-6 sec) on a 200Mhz CPU. But > I heard that the system clock is recommended not to exceed several > thousnad Hz. I don't even know how long time VxWorks consumes for its > own system tasks and kernel operation. > > 1) How can this be done? > 2) Is it possible to make the 1Mhz system clock? > 3) How large is the amount of time that VxWorks needs for its own > systematic purpose, with 200Mhz CPU? > > I really want your help. Thanks. > > Min,Ji-Hoon. Korea. --------------------------- Newsgroups: comp.os.vxworks Subject: VXWORKS: ftphookup sometimes hangs!!! Date: 6 Dec 2001 02:48:34 -0800 From: uloehnert@init-ka.de (Uwe Loehnert) Organization: http://groups.google.com/ Message-ID: Hi, i have a problem with the ftp client, built in vxworks. We use the ftp client to communicate with a remote ftp server via a WIRELESS network. Due to the fact, that wireless connections are not present at any time (or sometimes very astable), we can not guarantee the perefect connection. Now, sometimes it happens, that the ftphookup() function of ftplib hangs (doesn't return). My questions are: 1. Is there no timeout handling in ftphookup()??? (I heard, that timeout handling is not good supported in ftp client of vxworks). 2. If there is timeout handling=>Is there a possibility to set this value? We waited about one hour (!), but the function doesen't return! I suppose, that other ftplib functions may have the same problems. We use Tornado 2.02 with PPC platform. Can anyone help me with this? Does anyone know a better ftp client for vxworks (best thing would be of course: free :-))? With best regards Uwe Loehnert uloehnert@init-ka.de --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Reverse etherAddrResolve Date: Thu, 6 Dec 2001 13:21:58 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: This is just an idea, but it might work: Try to send an ICMP ECHO packet, with the destination IP set to broadcast (255.255.255.255) or old-broadcast (0.0.0.0), source IP equal to your real IP address, with the machine's individual MAC address in the ether header. Good chanse the machine will respond with either an ICMP ECHOREPLY, or with an ARP Request looking for the MAC of the machine which it needs to reply to, o rboth. I beleive this can be done with "normal" Unix commands, without writing a single line of code. Besides, you could try to get answers for this on TCP/IP related newsgroups, VxWorks specifics do not really matter. - - Leonid "Johan Borkhuis" wrote in message news:Xns916E968B6CE57borkhuisagerecom@135.7.153.102... > Hello, > > In one system I have the MAC address of a device on the network, and I want > to get the IP address of this device. Is this possible? > etherAddrResolve does the opposite, finding the MAC-address of a known IP- > address, but what I am looking for is just the other way around.. > > 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: using very frequent routines Date: Thu, 6 Dec 2001 13:30:36 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <4b5d36f2.0112052127.7df50b44@posting.google.com> Seriously, Min, Your CPU is assumed to have time to do other things, besides the 1us periodic operation, otherwise you would not use a CPU in the first place. The recommendation of limiting system clock to a few Khz is valid, and holds true even if your CPU is 1Ghz or faster. I have encountered cases where lazy hardware designers would leave a "very frequent but simple" operation to software, instead of doing it in hardware, which sometimes amounted to only connecting one wire - e.g. a FIFO device which receives on ebyte every 10us its pin would be connected to a spare DMA channel signal, because all what is needed to be done is get the data from FIFO into memory. You could still try to set the Auxiliary clock to 1Mhz, and write a simple routing which does the 1us periodic thing, but there sill will be some kernel overhead in that ISR, so you will need to carefully measure the implication of doing this. And then there is of course the notion of "light weight ISR", which I have recommended in a similar case, which was acceptable tradeoff until the hardware got a proper fix. Hope this helps, - Leonid "Min, Ji-hoon" wrote in message news:4b5d36f2.0112052127.7df50b44@posting.google.com... > I'm very new to VxWorks. > > I want to run a certain task every 1us(1e-6 sec) on a 200Mhz CPU. But > I heard that the system clock is recommended not to exceed several > thousnad Hz. I don't even know how long time VxWorks consumes for its > own system tasks and kernel operation. > > 1) How can this be done? > 2) Is it possible to make the 1Mhz system clock? > 3) How large is the amount of time that VxWorks needs for its own > systematic purpose, with 200Mhz CPU? > > I really want your help. Thanks. > > Min,Ji-Hoon. Korea. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: how to add this news group to my outlook program? Date: Thu, 6 Dec 2001 12:44:15 +0100 From: "Martin Raabe" Organization: T-Online Message-ID: <9unlm7$2cf$00$1@news.t-online.com> References: <75f406a7.0112051655.63199b4c@posting.google.com> You're welcome!!! "mumx" schrieb im Newsbeitrag news:75f406a7.0112051655.63199b4c@posting.google.com... > thank you ! --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Thu Dec 6 05:52:17 2001 From: Ilia Bosis Date: Thu Dec 6 05:52:19 PST 2001 Subject: RE: WAITING FOR MSG while DELAYED > Submitted-by vxwexplo-errs@csg.lbl.gov Thu Dec 6 01:19:11 2001 > Submitted-by: Shlomo Katz > > hi > hi all > i am new user to vxworks and i would like to know > if a task, that puts in delay state, can move to the ready > state when a > message is received at its message Q, before the delay period expire? > > katz sholmi The answer is NO. When You call the taskDelay You ask Your task to sleep the specified time, and it will do that - exactly! If what You want is to wait for the message no more than the specified amount of time, use the msgQReceive() with timeout. From the msgQReceive's return value You can figure if it returned because of the message received or by timeout. Read the msgQReceive() spec in the reference manual for details. Regards Ilia. -------------------------------------------- Ilia Bosis Software Engineer PacketLight Networks Tel: +972-9-7644228 Fax: +972-9-7645444 E-mail: Ilia_Bosis@packetlight.com Web: www.packetlight.com From vxwexplo-errs@csg.lbl.gov Thu Dec 6 07:49:28 2001 From: Ilia Bosis Date: Thu Dec 6 07:49:30 PST 2001 Subject: Re: VxWorks interview question ! From: "Jamilur Rahman" Organization: Bell Sympatico Message-ID: Place: Ottawa, Ontario, Canada. Event: Job interview for Embedded S/W Designer Interviewer: (question...) Candidate: (answer...) ..... ..... Interviewer: What is the best thing in VxWorks you like most ? Candidate: Semaphore. Interviewer: Why ? Candidate: Because, it is faster and flexible. ..... ..... (note that, The candidate was not offered the job) Any suggestion, comment or thought ? Thanks. Jamil, Ottawa, Ontario. On the similar note: Interviewer: What is the best fruit ? Right answer: It depends (on Your taste, intended use, etc.) ANY specific answer will fail the Candidate. On a more serious note: If they asked about the most distinguishing feature, the one that stands VxWorks apart from the others, it's, no doubt, the target shell. Personally, I love it. Especially, when it permits You to debug and fix a problem at the customer' site located somewhere on another side of the globe. Unfortunately, it seems that WRS tries to replace it with the host shell. Not the equivalent. Not even close. Ilia. -------------------------------------------- Ilia Bosis Software Engineer PacketLight Networks Tel: +972-9-7644228 Fax: +972-9-7645444 E-mail: Ilia_Bosis@packetlight.com Web: www.packetlight.com From vxwexplo-errs@csg.lbl.gov Thu Dec 6 12:17:02 2001 From: Rainer Stelzer Date: Thu Dec 6 12:17:05 PST 2001 Subject: Re: GUI display on top of VxWorks ? Hi Chris, > Are there any other similar products > out there to consider. case you are looking for a tiny solution I've a hint. I needed a simple GUI for my MPC823 LCD controller, too. Have look at - the "vogle" library, a platform independent graphical C library and - fnt2bdf.c , a tiny program to printout windows fonts as a bitmap. this can simply be modified to show character on a memory mapped graphical interface. hth rgds Rainer ===================================================== 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 ===================================================== vxworks tornado From vxwexplo-errs@csg.lbl.gov Fri Dec 7 01:26:20 2001 From: keith.wood@epid.eurotherm.co.uk Date: Fri Dec 7 01:26:23 PST 2001 Subject: ARPprobes - Link Local IP auto address configuration Dear listers, I sent the following query to WindRiver, who suggested I should buy the entire network stack source code - which seems a real overkill! Can anybody out there give me a clue as the procedure calls I need?........ This is further to our brief discussion of the LinkLocal address allocation system at our meeting at Eurotherm on 11 Oct. As discussed at that time, LinkLocal was not (& is still not) at the stage of an RFC - but it has now achieved the level of defacto standard with Apple & Microsoft agreeing (yes - you did read that right!), and it is the subject of an IETF Internet draft, the latest version of which can be found at: http://www.ietf.org/internet-drafts/draft-ietf-zeroconf-ipv4-linklocal-04.txt I am now at the stage where I need to try & implement this. I am using VxWorks 5.3.1 / Tornado 1.0.1. If you go to the above link & read section 2.2 "Claiming a Link- Local Address" (it is only 6 paragraphs - the rest of the document can be regarded almost as background information) you will see what I have to implement. The most significant part being paragraph 2 of the above section. I have currently modified my product's BSP start-up at the point where I need to claim an IPaddress. My product's spec is to offer a simple user's configuration file (similar to a windows ".ini" file) where the user can select fixed configuration (explicit IP, subnet mask & default gateway); bootP/DHCP; and/or LinkLocal. The code is now written to correctly action this configuration file (albeit I have so far only supported bootP not DHCP). The only significant thing I have left to add is the use of LinkLocal where required. (At present I printf a "LinkLocal needed here" to the console - very informative but not very functional:-). As you will see from the draft document, I need to issue ARPprobes. The problem I have is that the published VxWorks interface to the ARP code does not seem to allow me that level of control. The routines I need must exist in the VxWorks code for ARP to work at all. I can see a number of functions with the prefix "arp", and I wonder if maybe functions such as arpCmd() or arpresolve() may be what I want? I suspect what I need to do is very simple - I just need to know the spec of the interface(s) to call. Any advice you can offer would be most appreciated. Keith Wood ------------------------- e-mail work: keith.wood@epid.eurotherm.co.uk home: genius@the-wood-family.org.uk phone: work: +44 1903 205277 My views are not my own - they are the intellectual property of my employer From vxwexplo-errs@csg.lbl.gov Fri Dec 7 04:03:20 2001 From: Vxworks Exploder Date: Fri Dec 7 04:03:23 PST 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Fri Dec 7 04:03:14 PST 2001 Subject: Re: Can the MPC860 do this ? Subject: Re: Closing a socket while blocked in write or read Subject: Re: VxWorks interview question ! Subject: Porting SVGALib to VxWorks Subject: Re: make.PENTIUMgnu Subject: Format a SCSI Drive with dosFs2.0 and VxWorks 5.4 Subject: Where have 'Specific Trap MIB Lookup Results' gone ? Subject: Re: using very frequent routines Subject: Re: Format a SCSI Drive with dosFs2.0 and VxWorks 5.4 Subject: new memory bloc on powerpc Subject: Re: VxWorks interview question ! Subject: ld() and unld() failures for me....... Subject: Re: new memory bloc on powerpc Subject: How about this: Can the MPC860 do this ? Subject: ULIP NIC driver Subject: symbol table Subject: Re: ld() and unld() failures for me....... Subject: Re: TCP Behavior on Abnormal Socket Close Subject: Re: Need info about pSOS+ to VxWorks translation code from Wind River Subject: POST / RAM Test Subject: DOS File Corruption Subject: ENOBUFS using sendmsg(). Subject: Re: HELP!!! in DHCP Subject: Re: ENOBUFS using sendmsg(). Subject: Bypass VxWorks IP Routing. Subject: cache and interrupts? Subject: loadModule questions Subject: Re: Porting SVGALib to VxWorks Subject: Re: Dynamic loading Subject: Re: DosFS Questions Subject: Re: what is the meaning of FOREVER? Subject: Re: Maximum number of ethernets for VxWorks? Subject: Re: memShow and Code Segment size Subject: Re: cache and interrupts? Subject: Re: symbol table Subject: Re: symbol table Subject: Re: POST / RAM Test Subject: Re: DOS File Corruption Subject: Re: DosFS Questions Subject: test Subject: test Subject: End Driver questions ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Can the MPC860 do this ? Date: Thu, 6 Dec 2001 06:43:02 -0800 From: "Boris" Organization: N/A Message-ID: <9unnc7$g90$1@mail.cn99.com> References: SSBqdXN0IGRvIGFuIGFsaWtlIGpvYiBjdXJyZW50bHksIHlvdSBjYW4gZG8gYXMgZm9sbG93LA0K bG93ZXIgYnVzIHNwZWVkIHRvIDI1TUh6IG9yIHNvLHByb2dyYW0gdGhlIEJSeCh4IGlzIHRoZQ0K dGhlIENTIHlvdSBhc3NpZ24gZm9yIHRoYXQgY2hpcCkgdG8gdGhlIGxvbmdlc3QgcG9zc2libGUs DQp3aGljaCBJIHRoaW5rIGlzIDB4MGZmNCBmb3IgdGhlIGxvd2VyIGhhbGYgb2YgQlJ4LkkgdGhp bmsNCnlvdSB3aWxsIGdpdmUgdGhlIGFjY2VzcyB0aW1lIGFzIGxvbmcgYXMgYWJvdXQgMS4ydVMu SSANCnRoaW5rIHRoYXQgbWVldCB5b3UgcmVxdWVzdC5JIHJ1biBidXMgc3BlZWQgb24gNDhNSHos YW5kDQpJIGdldCBhYm91dCA2NjBuUyBhY2Nlc3MgdGltZSB1c2luZyBHUENNLg0KQnV0IHJlYWxs eSBJIHRoaW5rIGl0J3MgYSB2ZXJ5IGJhZCBpZGVhIHRvIHVzZSA4NjAgdG8gZG8NCnRob3NlIHNs b3cgYWNjZXNzLHVzZSB0aG9zZSBzbG93IGNoaXAgdG8gZGVhbCB3aXRoIHRob3NlDQpzbG93IGFj Y2VzcywgRlBHQSBJIHRoaW5rIGlzIGEgYmV0dGVyIGlkZWEuDQpYaWFvDQoNCiJKaW0iIDxqZ2Rv bkB1cmVhY2guY29tPiB3cm90ZSBpbiBtZXNzYWdlIG5ld3M6eXZyUDcuMjM3OSQ4ZS4xNzQ3NjBA bmV3cy4uLg0KPiBIZWxsbyBhbGwsDQo+IA0KPiBJIGhhdmUgYSBxdWVzdGlvbiBmb3IgYW55b25l IHdobyBkYXJlcyBhbnN3ZXIgOy0pDQo+IA0KPiBDdXJyZW50bHkgSSBoYXZlIGEgYm9hcmQgcnVu bmluZyB2eFdvcmtzIHdpdGggYW4gTVBDODYwIGF0IDgwTWh6IGFuZCBhIGJ1cw0KPiBmcmVxdWVu Y3kgb2YgNDAgTWh6Lg0KPiANCj4gSSBoYXZlIGEgbmV3IGRldmljZSB0aGF0IEkgbmVlZCB0byBj b21tdW5pY2F0ZSB3aXRoIG9uIGEgY2hpcCBzZWxlY3QgLi4uIHRoZQ0KPiBkZXZpY2UgcmVxdWly ZXMgLi4uIGEgImNyYXp5IHNsb3ciIGJ1cyBzcGVlZCBvZiBsZXNzIHRoYW4gMSBNaHogLi4uDQo+ IG90aGVyd2lzZSBpdCB3aWxsIGp1c3QgcGxhaW4gbm90IHdvcmsgLi4uDQo+IA0KPiBTbyBmYXIg SSAoYW5kIG90aGVycykgaGF2ZSBtYW5hZ2VkIHRvIGdldCB0aGUgYnVzIHNwZWVkIHRvIGFyb3Vu ZCAxIE1oeg0KPiB1c2luZyB0aGUgImViZGYiIGFuZCAiZGZuaCIgKHRoZXNlIGFyZSBpbiBjaGFw dGVyIDE1IG9mIHRoZSA4NjAgdXNlcnMgbWFudWFsDQo+IGlmIHlvdSBpbnRlcmVzdGVkKSAuLi4g d2hlbiB3ZSBsb3dlciB0aGUgYnVzIHRvIHRoaXMgc3BlZWQgaG93ZXZlciBldGhlcm5ldA0KPiBh bmQgdGhlIHNlcmlhbCBwb3J0cyBiZWhhdmUgInBvb3JseSIgLi4uIG1heWJlIEkgbWlzc2VkIHNv bWV0aGluZyB3aGVuIEkNCj4gdHJpZWQgdGhpcy4NCj4gDQo+IEN1cnJlbnRseSB3ZSBhcmUgY29u c2lkZXIgb25seSBkcm9wcGluZyB0aGUgYnVzIHJhdGVzIHdoZW4gd2UgZG8gYW4gYWNjZXNzDQo+ IHRvIHRoaXMgc2xvdyBkZXZpY2UgLi4uIHRvIGRvIHRoaXMgd2Ugd291bGQgd3JhcCB3cml0ZXMg dG8gdGhpcyBhZGRyZXNzDQo+IHJhbmdlIGluIGEgZnVuY3Rpb24gY2FsbCAuLi4gdGhlbiBpbiB0 aGUgZnVuY3Rpb24gc2h1dCBvZmYgaW50ZXJ1cHRzLCBkcm9wDQo+IHRoZSBjbG9ja3MsIGFuZCB0 aGVuIGRvIHRoZSBtZW1vcnkgd3JpdGUvcmVhZCwgdGhlbiBzcGVlZCB0aGUgY2xvY2tzIGJhY2sN Cj4gdXAsIHRoZW4gc3RhcnQgaW50ZXJ1cHRzIGJhY2sgdXAgLi4uIGRvZXMgYW55IG9uZSBrbm93 IGlmIHRoaXMgd2lsbCB3b3JrPw0KPiANCj4gQW55IGhlbHAgLi4uIG9yIGV2ZW4gc3ltcGF0aHkg d291bGQgYmUgZ3JlYXRseSBhcHByZWNpYXRlZCAuLi4NCj4gDQo+IEppbQ0KPiANCj4gDQo= --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Closing a socket while blocked in write or read Date: 6 Dec 2001 04:36:29 -0800 From: nick_nw@yahoo.co.uk (Nick) Organization: http://groups.google.com/ Message-ID: <2413908f.0112060436.73bd6d83@posting.google.com> References: <2413908f.0112030458.6b69c8f1@posting.google.com> <3c0e545b@news.barak.net.il> Hi, thanks for your comments. We do indeed want to discard any unsent data. The system we are writing should really use UDP not TCP, but TCP was imposed on us long ago. I read with interest comments about FIOCANCEL and shutdown(). I've never come across either of these, but I'll get out my bibles (Mr. Stevens Collection) and have a look. Any brief summary would be appreciated though. Thanks again, Nick. "Ilya" wrote in message news:<3c0e545b@news.barak.net.il>... > May be You forgot to shutdown the socket. > Ilya > "Nick" wrote in message > news:2413908f.0112030458.6b69c8f1@posting.google.com... > > Hi, > > > > I am seeing some strange behaviour: > > > > Task 1: Blocking socket, call write (which blocks). > > Task 2: Close socket (which is blcoked in write). > > Task 1: No return from write. > > > > I thought that when the socket is closed it should return from any > > blocked calls to read or write. > > > > Anyone else come across this? > > > > Cheers, > > > > Nick --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks interview question ! Date: Thu, 06 Dec 2001 13:01:21 GMT From: "Don Dewar" Message-ID: References: Having used a number of real time operating systems, I would say it is all the features that make developing on vxWorks its best features. Having a shell that can run almost any function in the system, that can set breakpoints, look at task traces, look at task registers, control task states, and dynamically load .out files for quick development makes vxWorks the best development enviroment of all the real time OS'es I have used. VxWorks has many other great features. If I were to rank the best in order I might say: shell dynamic load network support (IP stack, ftp, tftp, telnet, rlogin etc.) windview I would not include the basic facilities any RTOS must have such as, semaphores, memory management, and scheduling. Although the choice of different scheduling models provided by vxWorks is nice. - -- Don "Jamilur Rahman" wrote in message news:TDkP7.17300$iF3.1756557@news20.bellglobal.com... > Place: Ottawa, Ontario, Canada. > Event: Job interview for Embedded S/W Designer > > Interviewer: ..... > Candidate: ...... > .......... > Interviewer: What is the best part in VxWorks ? > Candidate: Semaphore. > Interview: Why ? > Candidate: Because, it is faster and flexible. > .......... > .......... > (Note that, the candidate was not offered the job) > > Does anybody have some other thought/comment/suggestion on this ? > > --------------------------- Newsgroups: comp.os.vxworks Subject: Porting SVGALib to VxWorks Date: 6 Dec 2001 05:16:12 -0800 From: patrick.molenaar@fcs-cs.com (Patrick) Organization: http://groups.google.com/ Message-ID: <986c860f.0112060516.120f53c2@posting.google.com> Hi, I have been trying to port a linux SVGAlib to VxWorks, but i came across the variable MMIO_POINTER, which seems to be a hardware specific memory-mapped-io-address. I have searched all the sourcecode i have on my computer and i have seached the internet, but i can't find the definition of this variable (or define) anywhere. Question: This newsgroup is my last hope, i hope someone know something about how this variable should be declared. Question: My second question is if anyone know where i can get functional SVGAlib for VxWorks, this would save me a lot of work trying to port the linux varian of this library. My intention with the lib is to run a graphical user interface on VxWorks. I was thinking about MicroWindows, which should be able to be portable to VxWorks according to the documentation on the internet. But there is also no info available how this is done. Question: Does anyone know where to get a graphical user interface for VxWorks? Many regards, Patrick. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: make.PENTIUMgnu Date: 6 Dec 2001 13:42:31 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: <9un0mm$nl4$1@mail.cn99.com> <9uni0c$aau$1@mail.cn99.com> "zhaoyandong" wrote: > It's not so , I 'm lack of the file named "make.PENTIUMgnu" in the > directory of $WIND_BASE\target\h\make > but not the license file of "WRSLicense.dat" The file you are missing is part of the Tornado for X86, and not part of Tornado Prototyper or the simulator. If you have a license for the first one it should be installed, the others don't have this file. 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: Format a SCSI Drive with dosFs2.0 and VxWorks 5.4 Date: 6 Dec 2001 06:13:58 -0800 From: rsdavis@isrparc.org (Scott Davis) Organization: http://groups.google.com/ Message-ID: <197ea102.0112060613.65127441@posting.google.com> I am trying to format a SCSI drive with VxWorks 5.4 and dosFs2.0. I have to provide the capability to be able to format a drive with a brand new disk and be able to format it after it has already been formatted once. VxWorks will also boot off the SCSI drive. Wind River stated to the following in the Windshell. BTW my scsi disk is SCSI Target Id 3, LUN 0. I would like to format the drive with a filesystem that is FAT32. It is only a 1.2 GB drive so am I wasting space by using the FAT32 or should I just use FAT16 or something else. I like having long filenames but if it is costing me in storage space it is not worth it. I have dosFs2.0 installed and kernel configured appropriately and system boots correctly even with scsi drive. dosFsMkFs does not format the drive what does it really do? I want to write some code that will setup, format, etc a brand new drive as well as re-format and setup an already initialized drive with a filesystem and data on it as well as run chkdsk() pDisk=scsiPhysDevIdGet(pSysScsiCtrl,3,0) pBlock=scsiBlkDevCreate(pDisk,0,0) vol=dosFsMkfs("/sd0",pBlock) This does format the drive and will also allow vxWorks to boot off of the SCSI drive. A routine that I have started follows. The problem is trying to re-format a hard drive that already has a filesystem and files on it. The dosFsDevInit fails. The dosFsDevInit appears to fail on a scsi device that is already formatted with a filesystem. STATUS formatDrive(t_deviceName p_driveName, t_labelName p_labelName, BOOL verbose, t_operationResults p_formatResults) { static char cvsId[]="$Id: formatDrive.c,v 1.6 2001/12/05 20:00:25 rsdavis Exp $"; SCSI_PHYS_DEV *physDev; /* ptr to SCSI physical device */ BLK_DEV *blockDev; /* ptr to block device for DOS partition */ STATUS status=ERROR; char dirName[MAX_FILENAME_LENGTH]; int fd=0; int chkdskOptions=DOS_CHK_REPAIR; t_deviceName fileSystemName; if (verbose) chkdskOptions |= DOS_CHK_VERB_1; else { chkdskOptions |= DOS_CHK_VERB_SILENT; } /* Get the physical Device structure in memory since the device has been initialized. If the device had never been loaded into memory that we would have had to create the device but the kernel is technically configured for this initially */ if ((physDev = scsiPhysDevIdGet(pSysScsiCtrl, SCSI_TARGET_ID, SCSI_TARGET_LUN)) == NULL) { printf ("sysScsiConfig: SCSI drive not found.\n"); return (ERROR); } /* create logical DOS partition on SCSI block device */ if ((blockDev = scsiBlkDevCreate (physDev, 0, 0)) == NULL) { sprintf(p_formatResults, "ERROR: can not create block device for scsi drive %s\n", p_driveName); return (ERROR); } /* Strip off trailing slash on drive name. dosFsMkfs can not tolerate it being there and still format drive. Really pretty screwy as far as I am concerned. This is in particular to format a drive that can boot vxWorks from */ /* strncpy(fileSystemName, p_driveName, strlen(p_driveName)-1); */ /* set up the DOS volume configuration info */ if ((dosDev = dosFsDevInit (p_driveName, blockDev, NULL)) == NULL) { sprintf(p_formatResults, "ERROR: can not initialize scsi drive %s\n", p_driveName); return (ERROR); } if ((fd = open(p_driveName, O_WRONLY, 0)) != ERROR) { if (ioctl(fd, FIODISKFORMAT, 0) != OK) { sprintf(p_formatResults, "ERROR: can not format scsi drive %s\n", p_driveName); return (ERROR); } } else { sprintf(p_formatResults, "ERROR: can not open scsi drive %s to format\n", p_driveName); return (ERROR); } if (dosFsMkfs(p_driveName, blockDev) == NULL) { sprintf(p_formatResults, "ERROR: can not create filesystem for scsi drive %s\n", p_driveName); return (ERROR); } if ((fd = open(p_driveName, O_RDWR, 0)) != ERROR) { if (ioctl(fd, FIOLABELSET, (int) p_labelName) == OK) { if (ioctl(fd, FIOCHKDSK, chkdskOptions) == OK) { close(fd); sprintf(dirName,"%s/kernel",fileSystemName); if (mkdirNvram(dirName) == ERROR) { sprintf(p_formatResults, "ERROR: can not make directory on drive %s\n", fileSystemName); return ERROR; } sprintf(dirName,"%s/application",fileSystemName); if (mkdirNvram(dirName) == ERROR) { sprintf(p_formatResults, "ERROR: can not make directory on drive %s\n", fileSystemName); return ERROR; } sprintf(dirName,"%s/application/sbc1",fileSystemName); if (mkdirNvram(dirName) == ERROR) { sprintf(p_formatResults, "ERROR: can not make directory on drive %s\n", fileSystemName); return ERROR; } sprintf(dirName,"%s/application/sbc1/conf",fileSystemName); if (mkdirNvram(dirName) == ERROR) { sprintf(p_formatResults, "ERROR: can not make directory on drive %s\n", fileSystemName); return ERROR; } sprintf(dirName,"%s/application/sbc1/logs",fileSystemName); if (mkdirNvram(dirName) == ERROR) { sprintf(p_formatResults, "ERROR: can not make directory on drive %s\n", fileSystemName); return ERROR; } sprintf(dirName,"%s/application/sbc1/ofp",fileSystemName); if (mkdirNvram(dirName) == ERROR) { sprintf(p_formatResults, "ERROR: can not make directory on drive %s\n", fileSystemName); return ERROR; } sprintf(dirName,"%s/application/sbc1/scripts",fileSystemName); if (mkdirNvram(dirName) == ERROR) { sprintf(p_formatResults, "ERROR: can not make directory on drive %s\n", fileSystemName); return ERROR; } sprintf(dirName,"%s/application/sbc2",fileSystemName); if (mkdirNvram(dirName) == ERROR) { sprintf(p_formatResults, "ERROR: can not make directory on drive %s\n", fileSystemName); return ERROR; } sprintf(dirName,"%s/application/sbc2/conf",fileSystemName); if (mkdirNvram(dirName) == ERROR) { sprintf(p_formatResults, "ERROR: can not make directory on drive %s\n", fileSystemName); return ERROR; } sprintf(dirName,"%s/application/sbc2/logs",fileSystemName); if (mkdirNvram(dirName) == ERROR) { sprintf(p_formatResults, "ERROR: can not make directory on drive %s\n", fileSystemName); return ERROR; } sprintf(dirName,"%s/application/sbc2/ofp",fileSystemName); if (mkdirNvram(dirName) == ERROR) { sprintf(p_formatResults, "ERROR: can not make directory on drive %s\n", fileSystemName); return ERROR; } sprintf(dirName,"%s/application/sbc2/scripts",fileSystemName); if (mkdirNvram(dirName) == ERROR) { sprintf(p_formatResults, "ERROR: can not make directory on drive %s\n", fileSystemName); return ERROR; } sprintf(dirName,"%s/application/sbc3",fileSystemName); if (mkdirNvram(dirName) == ERROR) { sprintf(p_formatResults, "ERROR: can not make directory on drive %s\n", fileSystemName); return ERROR; } sprintf(dirName,"%s/application/sbc3/conf",fileSystemName); if (mkdirNvram(dirName) == ERROR) { sprintf(p_formatResults, "ERROR: can not make directory on drive %s\n", fileSystemName); return ERROR; } sprintf(dirName,"%s/application/sbc3/logs",fileSystemName); if (mkdirNvram(dirName) == ERROR) { sprintf(p_formatResults, "ERROR: can not make directory on drive %s\n", fileSystemName); return ERROR; } sprintf(dirName,"%s/application/sbc3/ofp",fileSystemName); if (mkdirNvram(dirName) == ERROR) { sprintf(p_formatResults, "ERROR: can not make directory on drive %s\n", fileSystemName); return ERROR; } sprintf(dirName,"%s/application/sbc3/scripts",fileSystemName); if (mkdirNvram(dirName) == ERROR) { sprintf(p_formatResults, "ERROR: can not make directory on drive %s\n", fileSystemName); return ERROR; } sprintf(p_formatResults, "SUCCESS: drive %s formatted successfully and directory tree structure created.\n", fileSystemName); status = OK; } else { close(fd); sprintf(p_formatResults, "ERROR: can not perform CHKDSK on drive %s\n", p_driveName); status = status = ERROR; } } else { close(fd); sprintf(p_formatResults, "ERROR: can not label volume on drive %s\n", p_driveName); status = ERROR; } } else { sprintf(p_formatResults, "ERROR: can not open drive %s\n", p_driveName); status = ERROR; } return status; } --------------------------- Newsgroups: comp.protocols.snmp,comp.os.psos,comp.os.vxworks Subject: Where have 'Specific Trap MIB Lookup Results' gone ? Date: Thu, 06 Dec 2001 16:25:28 +0200 From: Alex Vinokur Organization: Scopus Network Technologies Message-ID: <3C0F7FD7.F9BA9321@bigfoot.com> Followup-To: comp.protocols.snmp ============================ pRISM+ for pSOSystem/PowerPC pRISM+ 1.2.3 . pSOSystem 2.2.2 Diab Data D-CC Optimizing C compiler SNMP Release 7.4 (SNMPv1) Epilogue Technology Emissary SNMP MIB Compiler, version 6.3 Host : Windows98 Target : PowerPC (MPC860) MG-SOFT MIB Browser, Professional Edition 7.0.0.3730 ============================ Here is two SNMPv1 traps' details represented in MG-SOFT Browser's Trap Ringer Consol. First trap is described in SMIv1. Second one is described in SMIv2. ########################################## ##################### ### Part-1 ### ---> SNMPv1 ### ---> SMIv1 ##################### - ------- Fragments of MIB file : SMIv1 ------- IMPORTS scopusEncoder FROM SCOPUS-GLOBAL-REG scopusE1000 OBJECT IDENTIFIER ::= { scopusEncoder 1 } initial-connection TRAP-TYPE ENTERPRISE scopusE1000 VARIABLES { alarm-level, encoderUpTime, encoderVersion } DESCRIPTION "" ::= 1 - --------------------------------------------- - --------- SNMP Trap Ringer Console ---------- Specific trap #1 trap(v1) received from: 10.10.0.20 at 06/12/2001 15:12:28 Time stamp: 0 days 00h:07m:43s.00th Agent address: 10.10.0.20 Port: 161 Transport: IP/UDP Protocol: SNMPv1 Trap Manager address: 10.10.0.129 Port: 162 Transport: IP/UDP Community: E1000-10.10.0.20 SNMPv1 agent address: 10.10.0.20 Enterprise: scopusE1000 Specific Trap MIB Lookup Results Name: initial-connection, Module: E1000-MIB, Enterprise: scopusE1000 Bindings (3) Binding #1: alarm-level.0 *** (int32) info(1) Binding #2: encoderUpTime.0 *** (timeticks) 0 days 00h:01m:44s.00th Binding #3: encoderVersion.0 *** (octets) E1000 MIB File Version 01.12 - --------------------------------------------- NOTE! Here trap #1 does contain 'Specific Trap MIB Lookup Results'. ########################################## ##################### ### Part-2 ### ---> SNMPv1 ### ---> SMIv2 ##################### - ------- Fragments of MIB file : SMIv2 ------- IMPORTS scopusEncoder FROM SCOPUS-GLOBAL-REG scopusE1000 OBJECT IDENTIFIER ::= { scopusEncoder 1 } traps OBJECT-IDENTITY STATUS current DESCRIPTION "" ::= { scopusE1000 5 } events OBJECT-IDENTITY STATUS current DESCRIPTION "" ::= { traps 0 } initial-connection NOTIFICATION-TYPE OBJECTS { encoder-ip-address, alarm-level, encoderUpTime, encoderVersion } STATUS current DESCRIPTION "" ::= { events 1 } - --------------------------------------------- - --------- SNMP Trap Ringer Console ---------- Specific trap #1 trap(v1) received from: 10.5.0.246 at 06/12/01 15:20:17 Time stamp: 0 days 00h:00m:06s.00th Agent address: 10.5.0.246 Port: 161 Transport: IP/UDP Protocol: SNMPv1 Trap Manager address: 10.5.0.50 Port: 162 Transport: IP/UDP Community: E1000-10.5.0.246 SNMPv1 agent address: 10.5.0.246 Enterprise: events Bindings (4) Binding #1: encoder-ip-address.0 *** (ipaddr) 10.5.0.246 Binding #2: alarm-level.0 *** (int32) info(1) Binding #3: encoderUpTime.0 *** (timeticks) 21 days 00h:46m:40s.88th Binding #4: encoderVersion.0 *** (octets) E1000 MIB File Version 01.12 - ------------------------------------------------------------------- NOTE! Here trap #1 doesn't contain 'Specific Trap MIB Lookup Results'. ########################################## What may cause the different representation of these traps? - -- =========================== Alex Vinokur mailto:alexvn@bigfoot.com mailto:alexvn@dr.com http://up.to/alexvn http://go.to/alexv_math =========================== --------------------------- Newsgroups: comp.os.vxworks Subject: Re: using very frequent routines Date: Thu, 6 Dec 2001 14:26:19 +0000 (UTC) From: don@mars.dgrc.crc.ca (Donald McLachlan) Organization: The unconfigured xvnews people Message-ID: <9unv6b$l57$1@crc-news.crc.ca> References: <4b5d36f2.0112052127.7df50b44@posting.google.com> Reply-To: don@mars.dgrc.crc.ca > I want to run a certain task every 1us(1e-6 sec) on a 200Mhz CPU Stop and think about this or a second. Being risc, your CPU can do 200 M instrcutions per second. You want to cause interrupts, so your function will run 1 M times/sec. That means you have at most 200 instructions with which to: 1 - do the interrupt handleing for the sysclock interrupt 2 - do the task swap to your routine 3 - return to the idle state. Forget it, your routine is probably more than 200 instructions. Just for grins, try somethign like the following: long long start, stop, calls_per_sec; long long calls, calls_per_sec; calls = 10000000 start = (long long)tickGet(); for(i = 0; i < calls; ++i) your_routine(); stop = (long long)tickGet(); calls_per_sec = (sysClkRateGet() * calls) / (stop - start); printf("can do your_routine() %lld times per second\n", calls_per_sec); NOTE: this is at your current sysClkRate ... increase that and you have less time to run your_routine(). - -- Donald McLachlan E-mail Donald.McLachlan@crc.ca Communications Research Centre / RNS Tel (613) 998-2845 3701 Carling Ave., Fax (613) 998-9648 Ottawa, Ontario K2H 8S2 Canada --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Format a SCSI Drive with dosFs2.0 and VxWorks 5.4 Date: Thu, 6 Dec 2001 17:13:10 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <197ea102.0112060613.65127441@posting.google.com> Hey Scott, I won't address all the details of your request because you missed the main point - DosFs 2.0 comes with it's own (supplement) manual, and you are wasting time by trying to acheive this before reading this doc. Secondly, regarding you rbooting needs, it is hard to reply without knowing the type of processor architecture you are using - x86 is special if it uses standard BIOS. Lastlly, it seems from your e-mail you won't be needing data interchange with MSFT systems, because if you do, this might make things somewhat more complicated. Hope this helps, - - Leonid "Scott Davis" wrote in message news:197ea102.0112060613.65127441@posting.google.com... > I am trying to format a SCSI drive with VxWorks 5.4 and dosFs2.0. I > have to provide the capability to be able to format a drive with a > brand new disk and be able to format it after it has already been > formatted once. VxWorks will also boot off the SCSI drive. Wind > River stated to the following in the Windshell. > BTW my scsi disk is SCSI Target Id 3, LUN 0. I would like to format > the drive with a filesystem that is FAT32. It is only a 1.2 GB drive > so am I wasting space by using the FAT32 or should I just use FAT16 or > something else. I like having long filenames but if it is costing me > in storage space it is not worth it. I have dosFs2.0 installed and > kernel configured appropriately and system boots correctly even with > scsi drive. dosFsMkFs does not format the drive what does it really > do? I want to write some code that will setup, format, etc a brand > new drive as well as re-format and setup an already initialized drive > with a filesystem and data on it as well as run chkdsk() > > pDisk=scsiPhysDevIdGet(pSysScsiCtrl,3,0) > pBlock=scsiBlkDevCreate(pDisk,0,0) > vol=dosFsMkfs("/sd0",pBlock) > > This does format the drive and will also allow vxWorks to boot off of > the SCSI drive. > > A routine that I have started follows. The problem is trying to > re-format a hard drive that already has a filesystem and files on it. > The dosFsDevInit fails. The dosFsDevInit appears to fail on a scsi > device that is already formatted with a filesystem. > > STATUS formatDrive(t_deviceName p_driveName, > t_labelName p_labelName, > BOOL verbose, > t_operationResults p_formatResults) { > > static char cvsId[]="$Id: formatDrive.c,v 1.6 2001/12/05 20:00:25 > rsdavis Exp $"; > > SCSI_PHYS_DEV *physDev; /* ptr to SCSI physical device > */ > BLK_DEV *blockDev; /* ptr to block device for DOS > partition */ > > STATUS status=ERROR; > char dirName[MAX_FILENAME_LENGTH]; > int fd=0; > int chkdskOptions=DOS_CHK_REPAIR; > t_deviceName fileSystemName; > > if (verbose) > chkdskOptions |= DOS_CHK_VERB_1; > else { > chkdskOptions |= DOS_CHK_VERB_SILENT; > } > > /* Get the physical Device structure in memory since the device has > been initialized. If the device had never been loaded into memory > that we would have had to create the device but the kernel is > technically configured for this initially */ > if ((physDev = scsiPhysDevIdGet(pSysScsiCtrl, SCSI_TARGET_ID, > SCSI_TARGET_LUN)) == NULL) { > printf ("sysScsiConfig: SCSI drive not found.\n"); > return (ERROR); > } > > /* create logical DOS partition on SCSI block device */ > if ((blockDev = scsiBlkDevCreate (physDev, 0, 0)) == NULL) { > sprintf(p_formatResults, "ERROR: can not create block device for > scsi drive %s\n", p_driveName); > return (ERROR); > } > > /* Strip off trailing slash on drive name. dosFsMkfs can not > tolerate > it being there and still format drive. Really pretty screwy as > far as I am concerned. This is in particular to format a drive > that can boot vxWorks from */ > /* strncpy(fileSystemName, p_driveName, strlen(p_driveName)-1); */ > > /* set up the DOS volume configuration info */ > if ((dosDev = dosFsDevInit (p_driveName, blockDev, NULL)) == NULL) { > sprintf(p_formatResults, "ERROR: can not initialize scsi drive > %s\n", p_driveName); > return (ERROR); > } > > if ((fd = open(p_driveName, O_WRONLY, 0)) != ERROR) { > > if (ioctl(fd, FIODISKFORMAT, 0) != OK) { > sprintf(p_formatResults, "ERROR: can not format scsi drive %s\n", > p_driveName); > return (ERROR); > } > } > else { > sprintf(p_formatResults, "ERROR: can not open scsi drive %s to > format\n", p_driveName); > return (ERROR); > } > > if (dosFsMkfs(p_driveName, blockDev) == NULL) { > sprintf(p_formatResults, "ERROR: can not create filesystem for scsi > drive %s\n", p_driveName); > return (ERROR); > } > > if ((fd = open(p_driveName, O_RDWR, 0)) != ERROR) { > > if (ioctl(fd, FIOLABELSET, (int) p_labelName) == OK) { > > if (ioctl(fd, FIOCHKDSK, chkdskOptions) == OK) { > close(fd); > sprintf(dirName,"%s/kernel",fileSystemName); > if (mkdirNvram(dirName) == ERROR) { > sprintf(p_formatResults, "ERROR: can not make directory on drive > %s\n", fileSystemName); > return ERROR; > } > sprintf(dirName,"%s/application",fileSystemName); > if (mkdirNvram(dirName) == ERROR) { > sprintf(p_formatResults, "ERROR: can not make directory on drive > %s\n", fileSystemName); > return ERROR; > } > sprintf(dirName,"%s/application/sbc1",fileSystemName); > if (mkdirNvram(dirName) == ERROR) { > sprintf(p_formatResults, "ERROR: can not make directory on drive > %s\n", fileSystemName); > return ERROR; > } > sprintf(dirName,"%s/application/sbc1/conf",fileSystemName); > if (mkdirNvram(dirName) == ERROR) { > sprintf(p_formatResults, "ERROR: can not make directory on drive > %s\n", fileSystemName); > return ERROR; > } > sprintf(dirName,"%s/application/sbc1/logs",fileSystemName); > if (mkdirNvram(dirName) == ERROR) { > sprintf(p_formatResults, "ERROR: can not make directory on drive > %s\n", fileSystemName); > return ERROR; > } > sprintf(dirName,"%s/application/sbc1/ofp",fileSystemName); > if (mkdirNvram(dirName) == ERROR) { > sprintf(p_formatResults, "ERROR: can not make directory on drive > %s\n", fileSystemName); > return ERROR; > } > sprintf(dirName,"%s/application/sbc1/scripts",fileSystemName); > if (mkdirNvram(dirName) == ERROR) { > sprintf(p_formatResults, "ERROR: can not make directory on drive > %s\n", fileSystemName); > return ERROR; > } > sprintf(dirName,"%s/application/sbc2",fileSystemName); > if (mkdirNvram(dirName) == ERROR) { > sprintf(p_formatResults, "ERROR: can not make directory on drive > %s\n", fileSystemName); > return ERROR; > } > sprintf(dirName,"%s/application/sbc2/conf",fileSystemName); > if (mkdirNvram(dirName) == ERROR) { > sprintf(p_formatResults, "ERROR: can not make directory on drive > %s\n", fileSystemName); > return ERROR; > } > sprintf(dirName,"%s/application/sbc2/logs",fileSystemName); > if (mkdirNvram(dirName) == ERROR) { > sprintf(p_formatResults, "ERROR: can not make directory on drive > %s\n", fileSystemName); > return ERROR; > } > sprintf(dirName,"%s/application/sbc2/ofp",fileSystemName); > if (mkdirNvram(dirName) == ERROR) { > sprintf(p_formatResults, "ERROR: can not make directory on drive > %s\n", fileSystemName); > return ERROR; > } > sprintf(dirName,"%s/application/sbc2/scripts",fileSystemName); > if (mkdirNvram(dirName) == ERROR) { > sprintf(p_formatResults, "ERROR: can not make directory on drive > %s\n", fileSystemName); > return ERROR; > } > sprintf(dirName,"%s/application/sbc3",fileSystemName); > if (mkdirNvram(dirName) == ERROR) { > sprintf(p_formatResults, "ERROR: can not make directory on drive > %s\n", fileSystemName); > return ERROR; > } > sprintf(dirName,"%s/application/sbc3/conf",fileSystemName); > if (mkdirNvram(dirName) == ERROR) { > sprintf(p_formatResults, "ERROR: can not make directory on drive > %s\n", fileSystemName); > return ERROR; > } > sprintf(dirName,"%s/application/sbc3/logs",fileSystemName); > if (mkdirNvram(dirName) == ERROR) { > sprintf(p_formatResults, "ERROR: can not make directory on drive > %s\n", fileSystemName); > return ERROR; > } > sprintf(dirName,"%s/application/sbc3/ofp",fileSystemName); > if (mkdirNvram(dirName) == ERROR) { > sprintf(p_formatResults, "ERROR: can not make directory on drive > %s\n", fileSystemName); > return ERROR; > } > sprintf(dirName,"%s/application/sbc3/scripts",fileSystemName); > if (mkdirNvram(dirName) == ERROR) { > sprintf(p_formatResults, "ERROR: can not make directory on drive > %s\n", fileSystemName); > return ERROR; > } > sprintf(p_formatResults, "SUCCESS: drive %s formatted successfully > and directory tree structure created.\n", fileSystemName); > > status = OK; > } > else { > close(fd); > sprintf(p_formatResults, "ERROR: can not perform CHKDSK on drive > %s\n", p_driveName); > status = status = ERROR; > } > } > else { > close(fd); > sprintf(p_formatResults, "ERROR: can not label volume on drive > %s\n", p_driveName); > status = ERROR; > } > } > else { > sprintf(p_formatResults, "ERROR: can not open drive %s\n", > p_driveName); > status = ERROR; > } > > return status; > > } --------------------------- Newsgroups: comp.os.vxworks,comp.sys.powerpc.tech Subject: new memory bloc on powerpc Date: Thu, 06 Dec 2001 15:17:19 GMT From: Anthony Rouaux Organization: [posted via Easynet France] Message-ID: Hi all, i d like to add a new memory bloc to my system with run on vxworks 5.4 and powerpc MPC860. I want to use a new chip select for this new bloc so i configured the BR and OR to do this but when i want to access the memory area i selected with the BR and OR register winshell give me the following error : WTX Error 0x100ca (AGENT_MEM_ACCESS_ERROR). I suppose that i must declare to vxworks this new memory bloc. But how can i do this? Does anyone can help me. Thanks a lot Anthony --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks interview question ! Date: Thu, 06 Dec 2001 15:27:54 GMT From: "Thierry De Corte" Organization: GlobeTrotter Message-ID: <_7MP7.14854$tx.1247837@carnaval.risq.qc.ca> References: I amazed that this guy was not hired in the WindRiver support departement... "Jamilur Rahman" wrote in message news:TDkP7.17300$iF3.1756557@news20.bellglobal.com... > Place: Ottawa, Ontario, Canada. > Event: Job interview for Embedded S/W Designer > > Interviewer: ..... > Candidate: ...... > .......... > Interviewer: What is the best part in VxWorks ? > Candidate: Semaphore. > Interview: Why ? > Candidate: Because, it is faster and flexible. > .......... > .......... > (Note that, the candidate was not offered the job) > > Does anybody have some other thought/comment/suggestion on this ? > > --------------------------- Newsgroups: comp.os.vxworks Subject: ld() and unld() failures for me....... Date: Thu, 06 Dec 2001 09:57:53 -0600 From: PC User Organization: ExecPC Internet - Milwaukee, WI Message-ID: <3c0f9808$0$65164$272ea4a1@news.execpc.com> I am trying to load a test.o and run a function in it and delay for 60 seconds to modify it contents, unload and load it again to run the same function that would have a different result, but I am not successful. Is this because that the kernel is still holding on to the resources or do I need to be more specific? Both times at the call TestPatch has the same result or print statement. Would this concept work or is it impossible. Again I am trying to see if I can implement some type of interactive patching. Thanks in advance. /* Header Files */ #include #include #include "loadLib.h" include "ioLib.h" #include "usrLib.h" /* Extern Function Prototype, this is in a separate file called test.c (test.o) */ extern void TestPatch (void); void TestLoad ( void ) { MODULE_ID mdId; char MsgRecBuf[MAX_NUM_MSGS][MSG_LENGTH]; char MainMsg[MSG_LENGTH]; TestPatch (); if ( ( unld ( "test.o", 0 ) ) == ERROR ) { printf ( "Error in unloading test.o\n" ); return; } else printf ( "Successfully unloaded test.o\n" ); /* delay to allow changing function and compiling to get a new test.o */ taskDelay ( 60 * sysClkRateGet() ); if ( ( mdId = ld ( 1, 0, "test.o" ) ) == NULL ) { printf ( "Error in loading module\n" ); return; } else printf ( "Successfuly loaded module\n" ); /* Call should made from the test.o file */ TestPatch (); /* Would like to see the change */ if ( unldByModuleId ( mdId ) == ERROR ) { printf ( "Error in unloading module\n" ); return; } else printf ( "Successfuly unloaded module\n" ); } /* This is in a separate c file called test.c, All I do is I change the print statement to see if the second time loading, would result in the neww statement. */ #include "vxWorks.h" #include "stdio.h" void TestPatch (void) { printf ( "This is patch2\n" ); } --------------------------- Newsgroups: comp.os.vxworks,comp.sys.powerpc.tech Subject: Re: new memory bloc on powerpc Date: Thu, 06 Dec 2001 16:22:25 GMT From: David Bryan Organization: http://www.newsranger.com Message-ID: <5XMP7.51435$xS6.84744@www.newsranger.com> References: In article , Anthony Rouaux says... > >Hi all, >i d like to add a new memory bloc to my system with run on vxworks 5.4 and >powerpc MPC860. I want to use a new chip select for this new bloc so i >configured the BR and OR to do this but when i want to access the memory >area i selected with the BR and OR register winshell give me the following >error : WTX Error 0x100ca (AGENT_MEM_ACCESS_ERROR). >I suppose that i must declare to vxworks this new memory bloc. But how can >i do this? Does anyone can help me. >Thanks a lot > >Anthony Anthony, You will need to set up the MMU to allow access to the new area of memory. Do this by modifying the sysPhysMemDesc table in sysLib.c. See the Programmers Guide chapter on Virtual Memory Configuration for a description of the table. Dave T h e P T R G r o u p, I n c. Embedded, Real-Time Solutions, and Training David Bryan Manager, Embedded Services www.ThePTRGroup.com Dave@ThePTRGroup.com --------------------------- Newsgroups: comp.os.vxworks Subject: How about this: Can the MPC860 do this ? Date: Thu, 06 Dec 2001 16:32:32 GMT From: "Jim" Organization: Magma Communications Ltd. Message-ID: References: Here is a new description of the problem ... now that my undertanding of it is getting better ... as before all answers advice and sympathy are welcome ;-) The "crazy slow" device we are talking to cannot tolerate rates on any lines higher than 1 Mhz ... so that means we must slow down the bus to this speed ... That means of course that GPCM and UPM changes like adding wait states will not make a difference ... So what we have done is slow the bus down to .625 Mhz ... we did this using the "dfnh" and "ebdf" to divide or 80 Mhz clock by 64 and then 2 respectiely to produce the bus clock rate. We can now communicate with the crazy device ...and we have managed to get our serial port to function at a low baud rate... but however ethernet will not work ... any advice on this ? We really would like ethernet to work ;-) Thanks again everyone, Jim "Jim" wrote in message news:yvrP7.2379$8e.174760@news... > Hello all, > > I have a question for anyone who dares answer ;-) > > Currently I have a board running vxWorks with an MPC860 at 80Mhz and a bus > frequency of 40 Mhz. > > I have a new device that I need to communicate with on a chip select ... the > device requires ... a "crazy slow" bus speed of less than 1 Mhz ... > otherwise it will just plain not work ... > > So far I (and others) have managed to get the bus speed to around 1 Mhz > using the "ebdf" and "dfnh" (these are in chapter 15 of the 860 users manual > if you interested) ... when we lower the bus to this speed however ethernet > and the serial ports behave "poorly" ... maybe I missed something when I > tried this. > > Currently we are consider only dropping the bus rates when we do an access > to this slow device ... to do this we would wrap writes to this address > range in a function call ... then in the function shut off interupts, drop > the clocks, and then do the memory write/read, then speed the clocks back > up, then start interupts back up ... does any one know if this will work? > > Any help ... or even sympathy would be greatly appreciated ... > > Jim > > --------------------------- Newsgroups: comp.os.vxworks Subject: ULIP NIC driver Date: Thu, 6 Dec 2001 11:46:07 -0500 From: "Ren Hantao" Organization: Lucent Technologies, Columbus, Ohio Message-ID: <9uo78s$2lr@nntpa.cb.lucent.com> Can Ulip Network Interface for vxSim be support in win98? Hans --------------------------- Newsgroups: comp.os.vxworks Subject: symbol table Date: Thu, 6 Dec 2001 11:55:52 -0500 From: "Ren Hantao" Organization: Lucent Technologies, Columbus, Ohio Message-ID: <9uo7r4$2s5@nntpa.cb.lucent.com> Hi, vxWorks experts I want to add the symbol table to a standalone vxworks. but I only want part of the commands are supported on the target shell. Does anyone know how to do that? Hans --------------------------- Newsgroups: comp.os.vxworks Subject: Re: ld() and unld() failures for me....... Date: Thu, 06 Dec 2001 17:11:42 GMT From: David Bryan Organization: http://www.newsranger.com Message-ID: References: <3c0f9808$0$65164$272ea4a1@news.execpc.com> In article <3c0f9808$0$65164$272ea4a1@news.execpc.com>, PC User says... > >I am trying to load a test.o and run a function in it and delay for 60 >seconds to modify it contents, unload and load it again to run the same >function that would have a different result, but I am not successful. >Is this because that the kernel is still holding on to the resources or >do I need to be more specific? > >Both times at the call TestPatch has the same result or print >statement. Would this concept work or is it impossible. Again I am >trying to see if I can implement some type of interactive patching. > >Thanks in advance. .... Assuming you are using vxWorks 5.4 or earlier.... You are making the assumption that VxWorks load and unload are dll like in their operation. They are not. The load operation will resolve unresolved externals ONLY for the module you are loading. That is.... the unresolved function printf() will be resolved when TestPatch is loaded. However, the call to TestPatch() in TestLoad() will not be resolved (unless TestLoad is loaded after TestPatch) Likewise, the unload function will not remove calls (or make them unresolved) to the module being unloaded. If you unload a module any calls to the module will have unpredictable behavior. In your case, TestLoad called the original TestPatch routine and you saw the same results instead of the new TestPatch routine. The VxWorks Programmer's Guide: Configuration and Build has an 'OK' discussion on this. The bottom line is.... if you have interdependent modules you should link them together statically then load them onto the target. OK - so how do you achieve your desired results - patching the code. One approach would be to implement a jumptable. Basically, TestLoad calls a function by looking up a pointer to the function in a table or a linked list or a structure. The patch is loaded, the function pointer is updated to the new routine and you are on your way. Dave T h e P T R G r o u p, I n c. Embedded, Real-Time Solutions, and Training David Bryan Manager, Embedded Services www.ThePTRGroup.com Dave@ThePTRGroup.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: TCP Behavior on Abnormal Socket Close Date: 6 Dec 2001 09:32:19 -0800 From: csampsonsd@yahoo.com (Charles H. Sampson) Organization: http://groups.google.com/ Message-ID: References: <9ujlg6$21v$1@newpoisson.nosc.mil> <488e459a.0112042241.73480bc4@posting.google.com> john_94501@yahoo.com (John) wrote in message news:<488e459a.0112042241.73480bc4@posting.google.com>... > > You don't say, but I assume you are using read() to get the data. > There are a couple of things you need to understand about this: > > 1) TCP is not message based; it is a stream. Don't assume that writing > 10 bytes in one end means you'll get 10 bytes out in a single read at > the other end. You might; you might not. This often confuses people > since IP is packet based, so messages would make sense... > > 2) When read() returns zero it means end of file, i.e. the connection > is gone. Not sure why the machines involved make any difference to the > timing though. > > HTH, > > John... Your point (2) was quite helpful. I'm using recv(), not read(), but the comment that length of zero means EOF for read() sent me back to my reference library. After reviewing the books that I have used in the past, including the VxWorks documentation, and then wading through four new ones, I finally found one that mentioned, in passing, that a length of zero for recv() means that the connection has been broken. (Beginning of rant) That's what I meant by the arcane nature of network- ing. The initiates have been given the secret knowledge that is very diffi- cult for an outsider to obtain. Why is it that people writing books on net- working have such a hard time correctly specifying the behavior of subpro- grams? Is it that they don't know either? (End of rant) I think it's worthwhile pointing out the asymmetric behavior demonstrated by my tests. When the client/receiver unilaterally closes its socket, the server/sender gets an error indication on its next attempt to send, EPIPE. However, when the server/sender closes its socket, the client/receiver gets a non-erroneous conventional value on its next attempt to receive data. You've probably deduced from this an answer to your earlier question. The closing of the socket was not part of the problem, it was part of the test. If the results had been promising, we would have checked out the behav- ior when we just pulled the cable. Based on the test results, we're not going to bother. We're going to use UDP, where we have more control. Charlie - ----- For an email response, my user name is sampson and my mail server is cod.nosc.mil. The yahoo mailbox is just a spam sink. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Need info about pSOS+ to VxWorks translation code from Wind River Date: 6 Dec 2001 10:18:13 -0800 From: google@dharty.com (David) Organization: http://groups.google.com/ Message-ID: <5a87bef4.0112061018.731b5d0a@posting.google.com> References: <97d3f155.0112050616.74a5d443@posting.google.com> Greg, I have examined this "translation layer" myself, but have not implemented it. I was able to get it from somewhere on the vxworks site, but when I unzipped the download the files were all corruptuded and mangled together. however, I was able to see get the jist of it until my official vxworks package arrived. Anyway, the layer is simply a set of files that intercept your psos calls and perform the respective vxworks calls instead. for example t_suspend in the layer looks something like this: t_suspend(int taskID) { // the vxworks equivalent taskSuspend(taskID); } Note that in somecases it is not as easy as this. for example, tm_evevery() has a hefty amount of translation as vxworks doens't have the same concept of events as pSOS. Any event call, ev_receive for example, has a fair amount of overhead. See the code below for an actual example. It would be my recomendation to use the translation layer as a tool to give you ideas on how to do the conversion. For example, I have converted our PSOS events calls to use a vxworks message queue. Yes it took a little conversion, (ULONG events had to be changed to char buffer messages), but it negates the need for the complicated layers. Bottom line, use the kit as a guide to do your own porting. David /***************************************************************************** * * ev_receive - event reception routine. * * This function waits on a block of (32) event bits in the TCB extension, for * any/all of a given set of events to be sent to the task through the * ev_send routine. It can be used as a "select" for sempahores. * * INCLUDE FILES: p2vLib.h * * RETURNS: * EOK if succesful * ERR_TIMEOUT if requested events aren't received before expires. * ERR_NOEVS if requested events aren't received when the no-wait flag is set. * ERROR if calling task not a p2v task or if some error (other than * a timeout) occurred in semTake. */ UINT32 ev_receive ( UINT32 eventin, /* events to wait on */ /* (0 means just get pending events */ UINT32 flags, /* option flags(wait/nowait, any/all)*/ UINT32 timeout, /* how long to wait for the events */ UINT32 *eventout /* which events were received */ ) { int timeLast = 0; /* time value, to compute timeout */ int mask = eventin; /* which events to look for */ int eventsCur; /* */ int eventMatch; /* valid events matched on this loop */ int intlvl; /* key to unlock interrupts */ int rvalue = EOK; BOOL done = FALSE; /* have I received the needed events? */ P2V_SHADOW_TCB * pShadow; /* tcb extension(where events are kept*/ int timeoutPrev; BOOL waitForever; /* get the tcb extension and make sure this is a p2v task */ if( (pShadow = p2vTcb (0)) == NULL ) return (ERROR); /* for eventin=0, return all pending events */ if (mask == 0) { *eventout = pShadow->events; return (EOK); } /* Convert to vxWorks timeouts */ if (timeout == 0) timeout = WAIT_FOREVER; if (flags & EV_NOWAIT) timeout = NO_WAIT; /* Need this variable because when timeout is used as a counter * and it is decremented beyond 0, it could result in a -1, which is * a vxWorks WAIT_FOREVER */ waitForever = (timeout == WAIT_FOREVER); if ((timeout != WAIT_FOREVER) && (timeout != NO_WAIT)) timeLast = tickGet(); /* initialize timekeeper */ /* keep going until the proper events are received, or we've timed out */ do { /* get the currently pending events */ eventsCur = pShadow->events; /* see which ones we are looking for */ eventMatch = eventsCur & mask; /* see if we've matched any */ if (eventMatch) { /* * we need to unpend the events to acknowledge we've received them * lock interrupts while messing with the extension */ intlvl = intLock(); /* LOCK INTERRUPTS */ /* clear any events we've matched */ pShadow->events &= ~eventMatch; intUnlock (intlvl); /* UNLOCK INTERRUPTS */ /* stop looking for events we've received */ mask = mask & ~eventMatch; /* if the ANY flag is set, we're done */ if(flags & EV_ANY) done = TRUE; else { /* otherwise, keep looking if we haven't gotten them all */ if (mask == 0) done = TRUE; } } /* update the timeout value if needed */ if (!waitForever && (timeout != NO_WAIT)) { timeoutPrev = timeout; timeout = timeout + timeLast - (timeLast = tickGet()); if ( timeout==0 || ( (int) timeout < 0 && timeoutPrev > 0 ) ) timeout = NO_WAIT; } /* if we're not done, and there time on the clock */ if(!done && (timeout != NO_WAIT)) { /* * wait for someone to signal a new event. * (ev_send gives this semaphore) */ if (semTake (pShadow->eventWait, timeout) == ERROR) { /* if the wait times out, give up */ if (errno == S_objLib_OBJ_TIMEOUT) rvalue = ERR_TIMEOUT; else rvalue = ERROR; timeout = NO_WAIT; } } /* keep going until were done, or timed out */ } while( !done && (timeout != NO_WAIT)); /* In pSOS+, a failed NO_WAIT ev_receive returns ERR_NOEVS instead of * ERR_TIMEOUT */ if(flags & EV_NOWAIT) if(!done) rvalue = ERR_NOEVS; /* return all events which have been cleared from the mask(received) */ *eventout = eventin & ~mask; return (rvalue); } ghorton@lucent.com (Greg) wrote in message news:<97d3f155.0112050616.74a5d443@posting.google.com>... > We are using pSOS+ 2.2 & 2.5 in many projects, and are planning to > move to VxWorks for future projects. A possible intermediate step is > to obtain a translation layer from Wind River, so we can run the > VxWorks OS but keep our code base with pSOS+ calls. I believe this > will allow us to use all of the VxWorks Tornado tools, which would be > a big step up from what we have with pSOS. > > Does anybody have any info about how this translation layer works? > Any issues with it? Can we use all of the Tornado tools with it? > > Thanks in advance, > > Greg --------------------------- Newsgroups: comp.os.vxworks Subject: POST / RAM Test Date: 6 Dec 2001 10:20:35 -0800 From: davis_spam@yahoo.com (Dford) Organization: http://groups.google.com/ Message-ID: Hi, I'm trying to write a simple DRAM Pattern Test. I'm using PowerPC 8240 (603E core), Mousse Vooha board. 64MB DRAM. LOCAL_MEM_LOCAL_ADRS 0x0 LOCAL_MEM_SIZE 0x04000000 USER_RESREVED_MEM 0x0 I am working with a normal compressed bootrom (bootrom.bin) My questions are as follows: I need to run the RAM test before the bootrom copies itself into RAM, obviously. I am not a strong PowerPC Assembly programmer, so I'm trying to cheat by doing it in C. My entry point for the RAM test is the romStart() routine in bootInit.c. After romInit.s masks interrups, sets stack pointer to STACK_ADRS, and readies DRAM, it jumps to romStart(). In romStart, before the code copies ROM to RAM, I begin a RAM Pattern test. Correct me if I am wrong on these points: At this point the system is still executing instructions out of ROM. Am I correct in assuming the data section is also in ROM at this point? I have not compiled the ROM_RESIDENT bootrom (well, I tried but it was too large to fit in EEPROM). As I understand it, the ROM_RESIDENT bootrom only copies the data section to RAM (and .bss exists in RAM as well), and leaves the .text section in ROM. My entry point is before I do the copy from ROM to RAM, so I am assuming the .data section also exists in ROM at this point. The stack however, must exist in RAM somewhere. I am somewhat confused where it exists from the code / headers, though. I see the following defined in configAll.h: #if ((CPU_FAMILY == MIPS) || (CPU_FAMILY == PPC) #define STACK_RESIDENT RAM_DST_ADRS - -- following the chain of events: RAM_DST_ADRS -> RAM_HIGH_ADRS -> 0x00c00000 Then --- #if (_STACK_DIR == _STACK_GROWS_DOWN) **SNIPPED STUFF ABOUT ROM_RESIDENT** #define STACK_ADRS _romInit - -- _romInit -> is at 0x10100 for RAM, so bootrom copies to RAM_HIGH_ADRS, but when the system cold boots out of ROM, it must execute _romInit from EEPROM which is in the 0xff000100 (cold entry) = ROM_TEXT_ADRS #else /* STACK GROWS UP */ #define STACK_ADRS (_romInit-STACK_SAVE) - -- where STACK_SAVE corresponds to 64 bytes, so this is, I am assuming the size of my STACK. Therefore, I need to avoid the addresses from 0x00c00000 to 0x00c10000 if stack grows down, and 0x00bf0000 to 0x00c00000 if stack grows up. So, where is _STACK_DIR defined? Also, could someone clear up the following defines: ROM_LOW_ADRS = 00010000 # RAM text/data address for VxWorks ROM_HIGH_ADRS = 00c00000 # RAM text/data address for bootrom I see from the reference manual on bootInit that RAM_LOW_ADRS is where exc vectors, bp anchor, exc msg, and bootline are stored, and RAM_HIGH_ADRS is where (romInit + ROM_COPY_SIZE) or binArrayStart is stored. So, I understand the system works in the following manner: _romInit executes direct assembly code, sets STACK_ADRS to RAM_HIGH_ADRS, inits processor, sets up DRAM, etc., jumps to C routine romStart. When do exc msg, bootline, etc. get setup? This seems to happen in the romStart routine, but my pattern test runs as the first piece of code in romStart (before ANY copy happens from ROM to RAM), and when I hit address 0x4400, the system halts. This address is defined below in configAll.h: #if (CPU_FAMILY == PPC) #define RESREVED 0x4400 /* avoid zeroing out EXC_MSG */ This I can see if I warm reboot, and dump memory --> my pattern is written into RAM up to the address 0x4400 at which point it continues as all 0's. I'd like to be able to just jump over this section of RAM with my test, but I am unsure how large the reserved area is defined as. Actually, it would be ok, to jump over the entire exc msg, exc vectors, bp anchor, and bootline, if I could determine how large this section is. This would just be RAM_LOW_ADRS plus some offset. Can someone tell me how/where it defines how large this section is? Finally, I would like to be able to flush / invalidate the dcache during the pattern fill / verify tests. I tried issuing the system calls cacheFlush() / cacheInvalidate() from the romStart() routine, but the linker complains: ldppc -X -N -Map mousse.map -e _romInit -Ttext 00010000 \ -o bootrom romInit.o bootInit.o version.o \ D:\tornado\target/lib/libPPC603gnutms.a D:\tornado\target/lib/libPPC603gnuospfv2.a D:\tornado\target/lib/libPPC603gnuenvo y.a D:\tornado\target/lib/libPPC603gnuzlib.a D:\tornado\target/lib/libPPC603gnutms.a D:\tornado\target/lib/libdiag.a D:\torna do\target/lib/libtest.a D:\tornado\target/lib/libsal.a D:\tornado\target/lib/libbcm.a D:\tornado\target/lib/libsoc.a D:\torna do\target/lib/libsal.a D:\tornado\target/lib/libPPC603gnuvx.a bootrom.Z.o D:\tornado\target/lib/libPPC603gnuvx.a(workQLib.o): In function `workQPanic': workQLib.o(.text+0x316): undefined reference to `sysExcMsg' workQLib.o(.text+0x31e): undefined reference to `sysExcMsg' workQLib.o(.text+0x32a): undefined reference to `sysExcMsg' workQLib.o(.text+0x346): undefined reference to `sysExcMsg' workQLib.o(.text+0x34e): undefined reference to `sysExcMsg' D:\tornado\target/lib/libPPC603gnuvx.a(workQLib.o)(.text+0x35a): more undefined references to `sysExcMsg' follow make: *** [bootrom] Error 0x1 Is it possible to somehow resolve this, or is it necessary to write a funciton in PPC assembly to invalidate / flush the cache? Thanks in advance for any help, Dford --------------------------- Newsgroups: comp.os.vxworks Subject: DOS File Corruption Date: Thu, 6 Dec 2001 13:42:40 -0700 From: "Fred" Organization: Thomson Consumer Electronics Message-ID: <3c0fdc4b@news.tce.com> We're using PCMCIA flash cards to store our downloaded vxWorks, but have recently had problems with file corruption. Sometimes the file system is corrupted, sometimes the file is corrupted. Occassionally we can reproduce the problem by booting the board during a FTP transfer. Does anybody have any ideas as to what is causing the problem and how to fix it? --------------------------- Newsgroups: comp.os.vxworks Subject: ENOBUFS using sendmsg(). Date: Thu, 6 Dec 2001 13:46:56 -0700 From: "Fred" Organization: Thomson Consumer Electronics Message-ID: <3c0fdd4b@news.tce.com> Occassionally while sending bursts to sendmsg(), EONBUFS will be returned. We have verified that we are not over running the buffers using mbufShow. Does anybody have any ideas? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: HELP!!! in DHCP Date: Thu, 6 Dec 2001 14:18:15 -0800 From: "naveen" Message-ID: <3c0fee91$1@news.cadence.com> References: <3c03db4c$1@news.cadence.com> Thanks for the info. But I don't understand when I try to use the following lease = dhcpcInit( net_if, 1 ); dhcpcBind( lease, synchronous ); it fails in bind.I don't know why. Am I doing someting wrong. Do you know what is the exact flow of the code or I need to do something else. Help will be highly appreciated. Thanx "shafi" wrote in message news:fcc7da52.0112060034.67214c53@posting.google.com... > Hi, > you will not receive the server address if you use dhcpcBoot* because, > the DHCP client will never send a request when it accepts BOOT reply. > That is the negotiation has not been completely done. Server responded > with just an offer and waiting for request. > so you may need complete negotiatin to take place, till you get the > DHCPACK, before you could retreive the parameters like server IP address > > shafi > > "naveen" wrote in message news:<3c03db4c$1@news.cadence.com>... > > Well I am also trying to configure DHCP client in Vxworks. > > The network is configured and target gets ip address. > > > > 1. I am unable to get DNS Server address . > > > > > > Secondly I am using the > > dhcpcBootInit > > dhcpcBootOptionSet > > dhcpcBootBind > > > > > > functions instead of > > > > > > lease = dhcpcInit( net_if, 1 ); > > dhcpcEventHookAdd( lease, ... ) > > dhcpcOptionSet( lease, ... ); > > dhcpcBind( lease, synchronous ); > > > > to configure the dhcp client. > > > > Because when I use the latter it fails in BIND. > > > > > > I will really appreciate if anybody can help in this. > > Thanx > > Naveen --------------------------- Newsgroups: comp.os.vxworks Subject: Re: ENOBUFS using sendmsg(). Date: Thu, 06 Dec 2001 16:44:53 -0600 From: Andrew Johnson Organization: APS, Argonne National Laboratory Message-ID: <3C0FF4E5.4D2D70A0@aps.anl.gov> References: <3c0fdd4b@news.tce.com> Fred wrote: > > Occassionally while sending bursts to sendmsg(), EONBUFS will be returned. > We have verified that we are not over running the buffers using mbufShow. > Does anybody have any ideas? In Tornado 2, mbufShow() is a synonym for netStackDataPoolShow(), which has a partner routine netStackSysPoolShow() for a second memory pool. It's probably the latter pool that is running out. To fix this, read the vxWorks Network Programmers Guide section 4.6.3 but note that it's actually the _SYS versions that will need changing. - - 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: Bypass VxWorks IP Routing. Date: Thu, 6 Dec 2001 23:42:10 +0000 (UTC) From: "Indhu Shetty" Organization: Mailgate.ORG Server - http://www.Mailgate.ORG Message-ID: <5517fce317f202c1c772db7e0fa7977f.34031@mygate.mailgate.org> Hi All, I am running VxWorks 5.4 on a PPC 8260 target. My board has multiple IP Interfaces. I am running my own implementation of OSPF routing protocol. Now VxWorks IP stack keeps its own routing table. How do I make it bypass its routing table, and may be query my OSPF for a route each time it has to forward a IP packet. Any help will be appreciated. Thanks Indhu - -- Posted from [66.7.6.122] via Mailgate.ORG Server - http://www.Mailgate.ORG --------------------------- Newsgroups: comp.os.vxworks Subject: cache and interrupts? Date: Thu, 6 Dec 2001 18:46:00 -0500 From: "BruceL" Message-ID: <9up06c$mqt$1@bob.news.rcn.net> Can anyone explain to me why my app should crash during the handling of a decrementer (real-time clock on the MPC 8245) exception if the instruction cache is enabled? This is the first exception the app sees. Thanks, Bruce --------------------------- Newsgroups: comp.os.vxworks Subject: loadModule questions Date: Fri, 07 Dec 2001 02:03:07 GMT From: "kevin dodd" Organization: AT&T Broadband Message-ID: Reply-To: "kevin dodd" I have a series of modules 16 possible, actual is 6, that I want to load, if present on the flash disk. The disk created as part of tffs. This has worked for a long period of time, however now I am having problems with this. 1) I form a file descriptor for the file in question 2) if present, execute loadModule 3) if present, then execute a unique function name for the module I have used this for a while, but now I seem to be having problems with this. If I put a large delay between modules it seems to work fine. Does anyone have any ideas as to how to make this work? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Porting SVGALib to VxWorks Date: 6 Dec 2001 18:10:22 -0800 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0112061810.26d1467d@posting.google.com> References: <986c860f.0112060516.120f53c2@posting.google.com> Hello Patrick, > I have been trying to port a linux SVGAlib to VxWorks, but i came > across the variable MMIO_POINTER, which seems to be a hardware > specific memory-mapped-io-address. I have searched all the sourcecode > i have on my computer and i have seached the internet, but i can't > find the definition of this variable (or define) anywhere. > > Question: > This newsgroup is my last hope, i hope someone know something about > how this variable should be declared. Probably the address of the graphics controller's memory mapped I/O space. Looking at the code for SVGALib, I'd guess that this was only needed for the acceleration code. To get the value, you'll need to investigate which PCI BAR (base address register) contains the I/O address (if there's only one, that's probably it; if not, check for an address with the LSB set). Of course, that assumes that the PCI chip has been initialised. Two options: 1) If you're using a PC as the target, and the BIOS has initialised the PCI chips, you just need to get the address it put in the I/O BAR - - you can use the pci support routines in VxWorks to find the chip and read its BAR values. 2) If the chip has not been initialised, you'll need to do the PCI initialisation. Look at the sequence for initialisation of an ethernet chip in your BSP for an example of how to do this. You'll also need to find a suitable address to initialise it with. Once you've done that, you can get you SVGALib init code to "find" the PCI device, and then read the BAR. Replace all those mmap calls with code to do this. > Question: > My second question is if anyone know where i can get functional > SVGAlib for VxWorks, this would save me a lot of work trying to port > the linux varian of this library. > > My intention with the lib is to run a graphical user interface on > VxWorks. I was thinking about MicroWindows, which should be able to be > portable to VxWorks according to the documentation on the internet. > But there is also no info available how this is done. You just need something to initialise the graphics controller on your board so that you have a frame buffer to draw into. If you know what the chip is, check out the xfree86 driver sources for it; most of them contain the code that you need to initialise the chip and get into into a suitable mode. If you Looking at the code, most of it should be OK. You'll need to deal with things like outb() - sysOutByte() for VxWorks - and some other little Linux-isms like that. An alternative you might want to think about is the xfree86 drivers for your graphics chip. (They can be browsed using cvsweb at www.xfree86.org - look under xc/programs/Xserver/hw/ for the drivers.) > Question: > Does anyone know where to get a graphical user interface for VxWorks? Wind River has one: WindML + Zinc. I believe that you can download a desktop version of Zinc to explore whether that would meet your needs, and I think it is also available for the prototyper. Of course, your local sales person should be more than willing to get you an eval of it too ;-) HTH, John... > Many regards, > Patrick. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Dynamic loading Date: Fri, 07 Dec 2001 02:18:45 GMT From: "kevin dodd" Organization: AT&T Broadband Message-ID: <9GVP7.5483$Yy.277051@rwcrnsc53> References: <9u2qu0$dn9$1@newshost.mot.com> Reply-To: "kevin dodd" I have an entire system where I dynamically load up to 16(currently 7) modules at run time, so that I can update specific modules with no concern to the other modules. I can then update individual components "Martin Roth" wrote in message news:9u2qu0$dn9$1@newshost.mot.com... > We are now in the definition stage of our new project which will probably > based on one of MPC8xx PPC processors. During the developement purposes > Vxwrks allows to dynamically link & load new modules and run them from the > target shell. I would like to perfrm dynamic linking/loading on the final > target (non development) sysem where I do not have the Tornado development > tools and the WDB agent within the target. Did anybody see such (non > development) system where application are linked and loaded dynamically. ? > Thanks, > Martin Roth > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: DosFS Questions Date: Fri, 07 Dec 2001 02:25:28 GMT From: "kevin dodd" Organization: AT&T Broadband Message-ID: References: <9ttrtu$ks1$1@news.inter.net.il> Reply-To: "kevin dodd" One other problem with Dos FS 2 is that you cannot make it work for a BootRom image. WRS did not make this a possibility, and they still don't have this fixed "Leonid Rosenboim" wrote in message news:newscache$bddfng$pha$1@lnews.actcom.co.il... > > "Anatoly Greenblatt" wrote in message > news:9ttrtu$ks1$1@news.inter.net.il... > > Hi Andy, > > > > There is an update, DOSFS2, which has comprehensive explanation of steps > to > > create dosfs partition. Another reason you need dosfs2 is that dosfs is > > limited to 2gb (or 4gb, I'm not sure). > > > > Anatoly. > > > > Tolik, you are wrong, DosFs 2.0 works quite well (n practice) on a 150GByte > and 300GByte > volumes, and probably somewhere there could be even a larger volume which is > supports, > but I stick to facts I am aware of. > > DosFs2.0 with the FAT32 format, can really stretch into a huge volume if > needed, for which > there is of course a price to pay, but that is another story. > > There is a limitation of file size - 4.3GByte is the max size of any single > file while using > MSFT-compatible VFAT directory structure (default). With the proprietary > VX_LONG directory > structure, there are 40 bits for file size, and thus a single file > theoretical limit is 1 TeraByte. > During development, max file size was tested at 8Gbyte (which was the > largest disk we could > afford to buy), but I am not aware of anyone testing VX_LONG file size limit > on anything larger. > > HTH, > - Leonid > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: what is the meaning of FOREVER? Date: Fri, 07 Dec 2001 02:32:49 GMT From: "kevin dodd" Organization: AT&T Broadband Message-ID: References: <9ta9o5$4m2$1@mail.cn99.com> <3BF8EBD6.2F6586A8@varndellengineering.com> Reply-To: "kevin dodd" FOREVER turns out to be "for(;;)" #define FOREVER for (;;) "George Varndell" wrote in message news:3BF8EBD6.2F6586A8@varndellengineering.com... > zhaoyandong wrote: > > > > what is the definition of FOREVER? > > #define FOREVER while(1) > or > #define FOREVER for(;;) > or > #define FOREVER while(TRUE == TRUE) > or > #define FOREVER while(TRUE != FALSE) > > You get the idea. So what does this obscenity buy you? > > while(1) - semantics are only unclear if you don't know the language > syntax is obviously correct > > FOREVER - semantics are questionable > syntax needs to be verified > > I spend hours per week searching through '.h' files trying to find > bugs that are directly or indirectly caused by such macro abuse. > I wish people would learn the language rather than redefine it. > > Regards, > George Varndell > -- > PPC and XScale Expertise Varndell Engineering, LLC. > GBIT END drivers available now. http://www.varndellengineering.com > ____________________________________________________________________________ __ > Posted Via Binaries.net = SPEED+RETENTION+COMPLETION = http://www.binaries.net --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Maximum number of ethernets for VxWorks? Date: 6 Dec 2001 19:18:51 -0800 From: jperron@cetaceannetworks.com (Jason Perron) Organization: http://groups.google.com/ Message-ID: References: When you say "maximum number of ethernets", are you refering to the maximum number of physical ethernet connections? If so, I think there are a couple of answers. The first is the maximum number of interfaces that the MUX can deal with. As far as I can tell, the MUX stores interfaces in a dynamic lists. So, theoretically, you can add them until you run out of unique unit numbers (32,767). Next, assuming that you are connecting these interfaces to the IP layer, you have to consider the maximum number of IP connections you can have. This is controlled by the IP_MAX_UNITS constant and defaults to the number of entries you put in the endDrvTbl in configNet.h. Hope this helps, Jason champu_n@yahoo.com (champak) wrote in message news:... > Could any body tell me what is the maximum number of ethernets( > especially intel 82559 family) that can be supported by Vxworks? > > regards --------------------------- Newsgroups: comp.os.vxworks Subject: Re: memShow and Code Segment size Date: Fri, 07 Dec 2001 04:31:54 GMT From: "Di Mao" Organization: Prodigy Internet http://www.prodigy.com Message-ID: <_CXP7.367$P86.286487684@newssvr13.news.prodigy.com> References: <3C0EDE12.98E27733@mindspring.com> I have seen the same scenario for a long time. I think hex and elf format has lots of fat (e.g. all the source code and library). If you create a flat binary to store into flash, the related bin format file is a lot smaller. Weiwan "Naga R Narayanaswamy" wrote in message news:3C0EDE12.98E27733@mindspring.com... > Hi, > > I have a vxWorks image size = 12 Meg. > However after loading the image, memShow shows > 4 Meg as allocated and nearly 250 Meg as free. > The system has 256 Meg RAM. > > Does memShow sows only the data portion and if so > where is the 12 Meg of image accounted for. > > Thanks > Naga. > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: cache and interrupts? Date: 6 Dec 2001 22:54:02 -0800 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0112062254.210b2a7b@posting.google.com> References: <9up06c$mqt$1@bob.news.rcn.net> Hello, A lot more information is going to be required before we can add anything helpful... like: 1) What is the exception you are getting? 2) How is your application "handling" a decrementer exception (that is normally connected to the VxWorks system clock)? 3) What BSP are you using? 4) Does the kernel work without your application included? If so, is time moving forward correctly (check with tickGet() or taskDelay())? Rgrds, John... "BruceL" wrote in message news:<9up06c$mqt$1@bob.news.rcn.net>... > Can anyone explain to me why my app should crash during the handling of a > decrementer (real-time clock on the MPC 8245) exception if the instruction > cache is enabled? This is the first exception the app sees. > > Thanks, > Bruce --------------------------- Newsgroups: comp.os.vxworks Subject: Re: symbol table Date: 7 Dec 2001 08:14:03 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: <9uo7r4$2s5@nntpa.cb.lucent.com> "Ren Hantao" wrote: > Hi, vxWorks experts > > I want to add the symbol table to > a standalone vxworks. but I only want > part of the commands are supported on > the target shell. Does anyone know how > to do that? One way to do this is to symRemove to remove all unwanted symbols from the symbol table. Another way is to build an extra step into the build process and filter the symbol table file. Groeten, Johan - -- o o o o o o o . . . _____________________________ o _____ || Johan Borkhuis | .][__n_n_|DD[ ====_____ | borkhuis@agere.com | >(________|__|_[_________]_|__________________________| _/oo OOOOO oo` ooo ooo 'o!o!o o!o!o` === VxWorks FAQ: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html === --------------------------- Newsgroups: comp.os.vxworks Subject: Re: symbol table Date: Fri, 7 Dec 2001 09:23:07 +0100 From: "Michael Lawnick" Organization: Buergernetz Dillingen Message-ID: <9upubc$gat$1@snoopy.bndlg.de> References: <9uo7r4$2s5@nntpa.cb.lucent.com> Reply-To: "Michael Lawnick" Sender: mlawnick@pd95249e4.dip.t-dialin.net Hi Hans, haven't tested, but only thought about it. My idea: in standard build process the file symTbl.c is built. It contains the symbols for your symbol table, i.e. the command you can type in on target-shell. Generate a symTbl.own that contains the commands you want to provide and add a copy command after generation of symTbl.c to overwrite it by symTbl.own. Cave at: all later downloads of code may only use your set of commands ! All references to system symbols (e.g. semCreate) will fail if not contained in symTbl.own. Hope this works! - -- 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 ============================================== "Ren Hantao" schrieb im Newsbeitrag news:9uo7r4$2s5@nntpa.cb.lucent.com... > Hi, vxWorks experts > > I want to add the symbol table to > a standalone vxworks. but I only want > part of the commands are supported on > the target shell. Does anyone know how > to do that? > > Hans > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: POST / RAM Test Date: 07 Dec 2001 09:59:03 +0200 From: Reinhold Wirth Organization: WEB.DE Message-ID: <3c1084d7$1@netnews.web.de> References: Hi Dford, if you want to call C-functions, you need a stack. If you have only RAM in the LOCAL_MEM, you may: * write a assembly-function without stack * Use the C-Funktion and skip the stack from tests The assembly output from a C-Source may be a helpful start. To your linker-error: sysExcMsg is defined and initialised in sysLib.c. > >I'm trying to write a simple DRAM Pattern Test. I'm using PowerPC >8240 (603E core), Mousse Vooha board. 64MB DRAM. > >LOCAL_MEM_LOCAL_ADRS 0x0 >LOCAL_MEM_SIZE 0x04000000 >USER_RESREVED_MEM 0x0 > >I am working with a normal compressed bootrom (bootrom.bin) My >questions are as follows: > >I need to run the RAM test before the bootrom copies itself into RAM, >obviously. I am not a strong PowerPC Assembly programmer, so I'm >trying to cheat by doing it in C. My entry point for the RAM test is >the romStart() routine in bootInit.c. After romInit.s masks >interrups, sets stack pointer to STACK_ADRS, and readies DRAM, it >jumps to romStart(). In romStart, before the code copies ROM to RAM, >I begin a RAM Pattern test. > >Correct me if I am wrong on these points: >.... >workQLib.o(.text+0x34e): undefined reference to `sysExcMsg' >D:\tornado\target/lib/libPPC603gnuvx.a(workQLib.o)(.text+0x35a): more - -- __________________________________________________________ News suchen, lesen, schreiben mit http://newsgroups.web.de --------------------------- Newsgroups: comp.os.vxworks Subject: Re: DOS File Corruption Date: Fri, 7 Dec 2001 11:49:15 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <3c0fdc4b@news.tce.com> There is insufficient details here, what version of DosFs ? What kind of PC card ? Using TFFS ? What kind of corruption ? Still, you may want to search on google groups for past discussions on Flash disk corruption, there where discussions on Flash disk corruption issues. "Fred" wrote in message news:3c0fdc4b@news.tce.com... > We're using PCMCIA flash cards to store our downloaded vxWorks, but have > recently had problems with file corruption. Sometimes the file system is > corrupted, sometimes the file is corrupted. Occassionally we can reproduce > the problem by booting the board during a FTP transfer. Does anybody have > any ideas as to what is causing the problem and how to fix it? > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: DosFS Questions Date: Fri, 7 Dec 2001 12:15:33 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <9ttrtu$ks1$1@news.inter.net.il> Kev, I don't see WHY DosFs 2.0 would not run as part of BootROM image, unless you trust the ConfigTool ... It does drag some odd libraries (e.g. 64 bit math) but this has been tried out OK. One thing I rememebr doing in a BootROM is build do without disk cache to save memory, and mount the volume read-only. If you need this, tell me more and I might be able to help. - - Leonid "kevin dodd" wrote in message news:sMVP7.5496$Yy.277779@rwcrnsc53... > One other problem with Dos FS 2 is that you cannot make it work for a > BootRom image. WRS did not make this a possibility, and they still don't > have this fixed > > "Leonid Rosenboim" wrote in message > news:newscache$bddfng$pha$1@lnews.actcom.co.il... > > > > "Anatoly Greenblatt" wrote in message > > news:9ttrtu$ks1$1@news.inter.net.il... > > > Hi Andy, > > > > > > There is an update, DOSFS2, which has comprehensive explanation of steps > > to > > > create dosfs partition. Another reason you need dosfs2 is that dosfs is > > > limited to 2gb (or 4gb, I'm not sure). > > > > > > Anatoly. > > > > > > > Tolik, you are wrong, DosFs 2.0 works quite well (n practice) on a > 150GByte > > and 300GByte > > volumes, and probably somewhere there could be even a larger volume which > is > > supports, > > but I stick to facts I am aware of. > > > > DosFs2.0 with the FAT32 format, can really stretch into a huge volume if > > needed, for which > > there is of course a price to pay, but that is another story. > > > > There is a limitation of file size - 4.3GByte is the max size of any > single > > file while using > > MSFT-compatible VFAT directory structure (default). With the proprietary > > VX_LONG directory > > structure, there are 40 bits for file size, and thus a single file > > theoretical limit is 1 TeraByte. > > During development, max file size was tested at 8Gbyte (which was the > > largest disk we could > > afford to buy), but I am not aware of anyone testing VX_LONG file size > limit > > on anything larger. > > > > HTH, > > - Leonid > > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: test Date: Fri, 7 Dec 2001 10:21:19 +0100 From: "Petro" Organization: Knoware Message-ID: <1007720654.680739@news.knoware.nl> --------------------------- Newsgroups: comp.os.vxworks Subject: test Date: Fri, 7 Dec 2001 12:26:15 +0100 From: "Frank" Organization: Knoware Message-ID: <1007724490.261183@news.knoware.nl> Reply-To: "Frank" --------------------------- Newsgroups: comp.os.vxworks Subject: End Driver questions Date: Fri, 7 Dec 2001 12:30:26 +0100 From: "Frank" Organization: Knoware Message-ID: <1007724740.259243@news.knoware.nl> Reply-To: "Frank" Hello, I have some question about writing an End Driver. - - How can I filter easily the MAC address out of the Mblk chain in the endSend() interface routine. - - What is the meaning of the next interfaces: * endAddressForm() * endAddressGet() * endPacketDataGet() - - Where can I find more specific information about writing an End Driver? I have the VxWorks books (VxWorks network, VxWorks Network Protocol Toolkit). Thanks --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Fri Dec 7 14:29:02 2001 From: Matthew Singer Date: Fri Dec 7 14:29:04 PST 2001 Subject: 603 DBAT's 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_01C17F6E.8EE161F0 Content-Type: text/plain; charset="iso-8859-1" Under VxWorks 5.4 on a PPC 603 without VxVmi is it possible to dynamically modify the DBAT's? I need to use 1 DBAT to alternate between two PMC devices that at 1/2 Gig apart in PCI space? Matt Singer ------_=_NextPart_001_01C17F6E.8EE161F0 Content-Type: text/html; charset="iso-8859-1"
Under VxWorks 5.4 on a PPC 603 without VxVmi is it possible to dynamically modify the DBAT's?  I need to use 1 DBAT to alternate between two PMC devices that at 1/2 Gig apart in PCI space?
 
Matt Singer
 
 
------_=_NextPart_001_01C17F6E.8EE161F0-- From vxwexplo-errs@csg.lbl.gov Sat Dec 8 04:03:12 2001 From: Vxworks Exploder Date: Sat Dec 8 04:03:14 PST 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sat Dec 8 04:03:07 PST 2001 Subject: Re: ld() and unld() failures for me....... Subject: SCSI & dosFs Subject: Re: Docs for developing BSP. Subject: Re: SCSI & dosFs Subject: Re: HELP!!! in DHCP Subject: Software excuting slowly Subject: Loading image over FTP to boot from FLASH Subject: Re: ethernet collisions Subject: Re: POST / RAM Test Subject: vxWorks SNTP and Solaris NTP interoperability Subject: Re: processor migration (MVME230X -> MVME5100 ALTIVEC) Subject: Re: symbol table Subject: Re: Making a .lib file Subject: Re: need help w/ epilogue 9.1 for tor2.0.2 Subject: Re: problem related with vxworks and TCP/IP Subject: Re: RADVision H.323 Stack Sample Subject: Re: Need Communication of Serial Port with MPC860 Target Subject: Re: cache and interrupts? Subject: Re: Software excuting slowly Subject: t2 project facility: possible to avoid compiling files based on flags? Subject: Re: HELP!!! in DHCP ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: ld() and unld() failures for me....... Date: Fri, 07 Dec 2001 07:00:03 -0600 From: PC User Organization: ExecPC Internet - Milwaukee, WI Message-ID: <3c10bfdb$0$65163$272ea4a1@news.execpc.com> References: <3c0f9808$0$65164$272ea4a1@news.execpc.com> Thanks for your respond David, I have never used a jumptable before. Is the jumptable a global structure to be created within the TestLoad. Is it modifiable from the shell so I can add data to it. I would appreciate it if you have an example piece of code or know a site that I can look it up. Thanks again. David Bryan wrote: > In article <3c0f9808$0$65164$272ea4a1@news.execpc.com>, PC User says... > > > >I am trying to load a test.o and run a function in it and delay for 60 > >seconds to modify it contents, unload and load it again to run the same > >function that would have a different result, but I am not successful. > >Is this because that the kernel is still holding on to the resources or > >do I need to be more specific? > > > >Both times at the call TestPatch has the same result or print > >statement. Would this concept work or is it impossible. Again I am > >trying to see if I can implement some type of interactive patching. > > > >Thanks in advance. > .... > > Assuming you are using vxWorks 5.4 or earlier.... > > You are making the assumption that VxWorks load and unload are dll like in their > operation. They are not. > > The load operation will resolve unresolved externals ONLY for the module you are > loading. That is.... the unresolved function printf() will be resolved when > TestPatch is loaded. However, the call to TestPatch() in TestLoad() will not be > resolved (unless TestLoad is loaded after TestPatch) > > Likewise, the unload function will not remove calls (or make them unresolved) to > the module being unloaded. If you unload a module any calls to the module will > have unpredictable behavior. In your case, TestLoad called the original > TestPatch routine and you saw the same results instead of the new TestPatch > routine. > > The VxWorks Programmer's Guide: Configuration and Build has an 'OK' discussion > on this. > > The bottom line is.... if you have interdependent modules you should link them > together statically then load them onto the target. > > OK - so how do you achieve your desired results - patching the code. One > approach would be to implement a jumptable. Basically, TestLoad calls a > function by looking up a pointer to the function in a table or a linked list or > a structure. The patch is loaded, the function pointer is updated to the new > routine and you are on your way. > > Dave > > T h e P T R G r o u p, I n c. > Embedded, Real-Time Solutions, and Training > David Bryan Manager, Embedded Services > www.ThePTRGroup.com Dave@ThePTRGroup.com --------------------------- Newsgroups: comp.os.vxworks Subject: SCSI & dosFs Date: Fri, 07 Dec 2001 14:16:06 +0100 From: THARAUD =?iso-8859-1?Q?Micka=EBl?= Organization: CS-SI Message-ID: <3C10C116.A13B905D@c-s.fr> Il s'agit d'un message multivolet au format MIME. - --------------E9427CAA42489AACA18B59CE Content-Type: text/html; charset=us-ascii Content-Transfer-Encoding: 7bit Hi,

  I'm working with an SCSI HardDisk.
I use the following syntax to initialise this disk :
/*********************************************/
    /* create physical SCSI Disk device */
        if ((pPhysDev[DISK_ID]=scsiPhysDevCreate(pSysScsiCtrl,DISK_ID,0,0,NONE,0,0,0))
            == (SCSI_PHYS_DEV *) NULL)
            {
            printf ("Disk : scsiPhysDevCreate() failed\n");
            return (ERROR);
            }
    }
    printf ("Physical SCSI Disk Device OK !\n");

    /* create logical SCSI Disk device */
     if ((pBlkDev = scsiBlkDevCreate (pPhysDev[DISK_ID], 0, 0)) == NULL)
        {
        printErr ("Disk : scsiBlkDevCreate() failed\n");
        return (ERROR);
        }
    printf ("logical SCSI Disk device OK\n");

/***********************************************************/

Then, I do a dosFsMkfs "/diskscsi/", pBlkDev to have a dosfs partition. My initialisation seems to be correct as I can use fopen, fread, fwrite and fclose functions.

My question is how can I use functions like "ls", "cd", "mkdir", "pwd"... on my hardDisk ??
 

Thanks in advance for your answers.
MTH. - --------------E9427CAA42489AACA18B59CE Content-Type: text/x-vcard; charset=us-ascii; name="mickael.tharaud.vcf" Content-Transfer-Encoding: 7bit Content-Description: Carte pour THARAUD Mickaël Content-Disposition: attachment; filename="mickael.tharaud.vcf" begin:vcard n:THARAUD;Mickaël tel;fax:05.56.47.94.81. tel;home:-- tel;work:05.56.34.77.70. x-mozilla-html:FALSE org:CS ;Système d'Information version:2.1 email;internet:mickael.tharaud@c-s.fr title:Ingénieur informatique adr;quoted-printable:;;Parc d'Activit=E9s Kennedy=0D=0AAvenue Becquerel B=E2timent I=0D=0ABP 19;MERIGNAC CEDEX;;33702;FRANCE fn:THARAUD Mickaël end:vcard - --------------E9427CAA42489AACA18B59CE-- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Docs for developing BSP. Date: 7 Dec 2001 05:48:53 -0800 From: christian.delez@ascom.ch (Christian Delez) Organization: http://groups.google.com/ Message-ID: References: <27203e86.0112032150.27dfe19b@posting.google.com> >a guy called Ronen (ronen77@hotmail.com) wrote to me: >Hi christian, >I am in the process of writing Ethernet driver for vxWorks, and I wonder >if the pdf that is refered to in this email would help me. - - the BSP doc will help you. I don't have a PDF version. You have to deal directly with WindRiver for it. We only got a hard copy that comes with the BSP dev kit. - - I don't know your tools equipment. You need Tornado/VxWorks, BSP dev kit value option, evalboard, the VxWorks Network Programmer's Guide (comes with Tornado), support from WindRiver. >Let me know if there is anything that can be helpful to me on this matter. >Furthermore, the Ethernet driver should be hooked up to the TCP/IP stack and >I need the API functions for the networks task. - - If you order the BSP dev kit VALUE OPTION, you get a sample ethernet driver in source code. - - to learn the TCP/IP stack and API, look at WindNet IP doc in the Tornado help (VxWorks Network Programmer's Guide ). - - Ethernet driver resources are maybe also available on the web. Staff like "how to implement", "how to test", "typical driver interface". - - Good luck, have fun, Christian >Ronen. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: SCSI & dosFs Date: Fri, 7 Dec 2001 16:16:06 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <3C10C116.A13B905D@c-s.fr> This is a multi-part message in MIME format. - ------=_NextPart_000_002F_01C17F3A.795181A0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Assuming you are using the old dosFs (not dosFs 2.0): the utility functions you are looking for are located in /target/src/usr/usrLib.c, and all you need is to include this module in your boot image, then you can use these functions from either the target shell or your application. You also may be confused by commands with the same name running in the context of the Host Shell (WindSh) which runs within the TOrnado IDE on the host: these functions are completely host-based, and are incapable of accessing target-resident file systems. For example (assuming you have usrLib.o in the target): - -> ls will list the files on teh host's file system, while - -> @ls "/diskname" will list files on the target file system. I would recommend you read more about "Target resident Tools" in VxWorks = Programmer's Guide. Bon chanse, - - Leonid "THARAUD Micka=EBl" wrote in message = news:3C10C116.A13B905D@c-s.fr... Hi,=20 I'm working with an SCSI HardDisk.=20 I use the following syntax to initialise this disk :=20 /*********************************************/=20 /* create physical SCSI Disk device */=20 if = ((pPhysDev[DISK_ID]=3DscsiPhysDevCreate(pSysScsiCtrl,DISK_ID,0,0,NONE,0,0= ,0))=20 =3D=3D (SCSI_PHYS_DEV *) NULL)=20 {=20 printf ("Disk : scsiPhysDevCreate() failed\n");=20 return (ERROR);=20 }=20 }=20 printf ("Physical SCSI Disk Device OK !\n");=20 /* create logical SCSI Disk device */=20 if ((pBlkDev =3D scsiBlkDevCreate (pPhysDev[DISK_ID], 0, 0)) = =3D=3D NULL)=20 {=20 printErr ("Disk : scsiBlkDevCreate() failed\n");=20 return (ERROR);=20 }=20 printf ("logical SCSI Disk device OK\n");=20 /***********************************************************/=20 Then, I do a dosFsMkfs "/diskscsi/", pBlkDev to have a dosfs = partition. My initialisation seems to be correct as I can use fopen, = fread, fwrite and fclose functions.=20 My question is how can I use functions like "ls", "cd", "mkdir", = "pwd"... on my hardDisk ??=20 =20 Thanks in advance for your answers.=20 MTH.=20 - ------=_NextPart_000_002F_01C17F3A.795181A0 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable

Assuming you are using the old dosFs = (not dosFs=20 2.0):
 
the utility functions you are looking for are = located=20 in
<tornado>/target/src/usr/usrLib.c, and = all you=20 need is
to include this module in your boot image, = then you can=20 use these
functions from either the target shell or your = application.
 
You also may be confused by commands with the = same=20 name
running in the context of the Host Shell = (WindSh) which=20 runs
within the TOrnado IDE on the host: these = functions are=20 completely
host-based, and are incapable of accessing=20 target-resident
file systems. For example (assuming you have = usrLib.o in=20 the target):
 
-> ls
 
will list the files on teh host's file system, = while
 
-> @ls "/diskname"
 
will list files on the target file = system.
 
I would recommend you read more about "Target = resident=20 Tools" in VxWorks Programmer's Guide.
 
Bon chanse,
- Leonid
"THARAUD Micka=EBl" <mickael.tharaud@c-s.fr> = wrote in=20 message news:3C10C116.A13B905D@c-s.fr.= ..
Hi,=20

  I'm working with an SCSI HardDisk.
I use the following = syntax to=20 initialise this disk : =
/*********************************************/=20
    /* create physical SCSI Disk device */=20
        if=20 = ((pPhysDev[DISK_ID]=3DscsiPhysDevCreate(pSysScsiCtrl,DISK_ID,0,0,NONE,0,0= ,0))=20
            = =3D=3D=20 (SCSI_PHYS_DEV *) NULL)=20
            = {=20
            = printf=20 ("Disk : scsiPhysDevCreate() failed\n");=20
            = return=20 (ERROR);=20
            = }=20
    }
    printf ("Physical SCSI = Disk=20 Device OK !\n");=20

    /* create logical SCSI Disk device */=20
     if ((pBlkDev =3D scsiBlkDevCreate=20 (pPhysDev[DISK_ID], 0, 0)) =3D=3D NULL)=20
        {=20
        printErr ("Disk :=20 scsiBlkDevCreate() failed\n"); =
       =20 return (ERROR);
        }=20
    printf ("logical SCSI Disk device OK\n");=20

/***********************************************************/=20

Then, I do a dosFsMkfs "/diskscsi/", pBlkDev to have a dosfs = partition. My=20 initialisation seems to be correct as I can use fopen, fread, fwrite = and=20 fclose functions.=20

My question is how can I use functions like "ls", "cd", "mkdir", = "pwd"...=20 on my hardDisk ??
 =20

Thanks in advance for your answers.
MTH. =

- ------=_NextPart_000_002F_01C17F3A.795181A0-- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: HELP!!! in DHCP Date: Fri, 7 Dec 2001 15:09:29 -0000 From: "David Laight" Organization: BT Internet Message-ID: <9uqlve$jgd$1@uranium.btinternet.com> References: <3c03db4c$1@news.cadence.com> <3c0fee91$1@news.cadence.com> My guess is that your vxWorks system is ignoring the DHCP 'offer' messages from the server because they are not padded out to the expected size. This is a 'well known' problem with the vxworks dhcp client, caused (in part) by a newer rfc not being binary compatible with the old one. The required fixed (to the dhcp source) are not hard, but WRS seem completely unwilling to agree they need doing, or to make publically available the version I fixed for them 2 years ago. David naveen wrote in message news:3c0fee91$1@news.cadence.com... > Thanks for the info. > > But I don't understand when I try to use the following > lease = dhcpcInit( net_if, 1 ); > dhcpcBind( lease, synchronous ); > > it fails in bind.I don't know why. > Am I doing someting wrong. > Do you know what is the exact flow of the code or I need to do something > else. > > Help will be highly appreciated. > > Thanx > "shafi" wrote in message > news:fcc7da52.0112060034.67214c53@posting.google.com... > > Hi, > > you will not receive the server address if you use dhcpcBoot* because, > > the DHCP client will never send a request when it accepts BOOT reply. > > That is the negotiation has not been completely done. Server responded > > with just an offer and waiting for request. > > so you may need complete negotiatin to take place, till you get the > > DHCPACK, before you could retreive the parameters like server IP address > > > > shafi > > > > "naveen" wrote in message > news:<3c03db4c$1@news.cadence.com>... > > > Well I am also trying to configure DHCP client in Vxworks. > > > The network is configured and target gets ip address. > > > > > > 1. I am unable to get DNS Server address . > > > > > > > > > Secondly I am using the > > > dhcpcBootInit > > > dhcpcBootOptionSet > > > dhcpcBootBind > > > > > > > > > functions instead of > > > > > > > > > lease = dhcpcInit( net_if, 1 ); > > > dhcpcEventHookAdd( lease, ... ) > > > dhcpcOptionSet( lease, ... ); > > > dhcpcBind( lease, synchronous ); > > > > > > to configure the dhcp client. > > > > > > Because when I use the latter it fails in BIND. > > > > > > > > > I will really appreciate if anybody can help in this. > > > Thanx > > > Naveen > > --------------------------- Newsgroups: comp.os.vxworks Subject: Software excuting slowly Date: Fri, 7 Dec 2001 23:33:08 +0800 From: "Chet Wee" Organization: SCV Cablemodem News site Message-ID: <9uqnr6$o7s$1@newsie.singa.pore.net> Hi all, My team has developed a software that is ~4MB. I used the Tornado Integrated Development Environment to compile/link & then download the software to the SDRAM for execution. (Option 1) Then I tried to flash the software into the Flash, using VxResRom build. The software runs very slowly in the Flash, even with the same compilation option. (Option 2) Then I switched to using VxRes build. It runs faster than theVxResRom option, as expected, but I expect it to run as the same speed as option (1), but it is running slower than Option (1) by almost 2 to 3 times. Can anyone please advise why is this so. Initially, I suspected that it could be caused by the unwanted interrupts, but I am using the same software as the Option (1). Please advise. Anyway, I am using the PPC740. --------------------------- Newsgroups: comp.os.vxworks Subject: Loading image over FTP to boot from FLASH Date: 7 Dec 2001 07:51:09 -0800 From: google@dharty.com (David) Organization: http://groups.google.com/ Message-ID: <5a87bef4.0112070751.575a90b7@posting.google.com> Hello everybody, I'm porting an application from PSOS (mveme2306) to vxworks and a new PowerPc board (mvme5100). The old app would boot from flash by default, but a variable could be configured in the boot prom to tell the app to load a new version of itself into flash over TFTP. To enable it, you would halt the default boot process, reconfigure the boot prom (similar to vxworks boot prom 'c') and reboot. If this variable (XferFile) was set when the app initialized, the app would call a tftp transfer routine. This routine would then download and run another program into memory that handled the loading and flashing of the new app image. Upon subsequent boots, the new image would load. I am interested if anybody has done something like this in vxworks. We don't currently use TFFS, although I know thats a possibility. We don't have to do it the same way as the old app, but the requirements are that we have a way to load a new application into flash over the ethernet, so that future boots load from the new flash image. Thanks David --------------------------- Newsgroups: comp.os.vxworks Subject: Re: ethernet collisions Date: Fri, 7 Dec 2001 15:50:47 -0000 From: "David Laight" Organization: BT Internet Message-ID: <9uqoct$cja$1@plutonium.btinternet.com> Someone posted a query about ethernet collisions a week or so ago. Ethernet is CSMACD - Carrier Sense Media Access (with) Collision Detection (or something very similar). Collisions are a normal part of the network activity and are neither a problem themselves, nor (necessarily) an indication of a problem somewhere else. If you are seeing a performance problem due to a network error then it might cause the collision light to be on more than expected..... However the number of collisions is largely dependant on the network utilisation and physical size. Given most 10M networks (these days) are physically small (when was the last time you saw coax sections linked by a 2km repeater set?) collisions will happen within the packet preamble. This means you need (about) 80000 collisions a second to load the network to 100% - this won't happen! 100M networks might be nearer their maximum size, but you can get over 15000 small packets/sec anyway. 'Late collisions' - ie those that happen after the first 64 bytes are a problem. The ethernet chipset ought to count these as 'transmit errors'. (One chipset called 'NICE' is nasty in that it ignores rx frames if the transceiver reports a collision. Unfortunately accurately filtering out mains hum is quite hard, and many (coax) transceivers will erronousley report collisions...) Common problems that might generate 'collisions': 1) connecting a FDX link to a HDX one. 2) have SQE enabled on a transceiver attached to a repeater set. David --------------------------- Newsgroups: comp.os.vxworks Subject: Re: POST / RAM Test Date: 7 Dec 2001 09:09:27 -0800 From: davis_spam@yahoo.com (Dford) Organization: http://groups.google.com/ Message-ID: References: Hello Reinhold, Thanks for the response. Unfortunately, you did not provide me with any new information. I understand I need a stack for C. I stated below that I would need to jump over it. However, I am unclear which way it grows. I believe it grows down in PowerPC, but if anyone else knows, please issue a response. Also, it seems necessary for me to jump over the reserved area for exc msg, etc. I'd like to know how large this is defined as (where?). I also understand where sysExcMsg is defined. sysLib.c is compiled in the bootrom Makefile, and also linked. Why then, can I not issue a call to it in romStart? I'm guessing it is because it is in the compressed region of code, that gets uncompressed to RAM. Can someone confirm this, and if so, is there a workaround? Thanks, Dford "Reinhold Wirth" wrote in message news:3c1084d7$1@netnews.web.de... >Hi Dford, > > >if you want to call C-functions, you need a stack. >If you have only RAM in the LOCAL_MEM, you may: >* write a assembly-function without stack >* Use the C-Funktion and skip the stack from tests > > >The assembly output from a C-Source may be a helpful start. > > >To your linker-error: >sysExcMsg is defined and initialised in sysLib.c. davis_spam@yahoo.com (Dford) wrote in message news:... > Hi, > > I'm trying to write a simple DRAM Pattern Test. I'm using PowerPC > 8240 (603E core), Mousse Vooha board. 64MB DRAM. > > LOCAL_MEM_LOCAL_ADRS 0x0 > LOCAL_MEM_SIZE 0x04000000 > USER_RESREVED_MEM 0x0 > > I am working with a normal compressed bootrom (bootrom.bin) My > questions are as follows: > > I need to run the RAM test before the bootrom copies itself into RAM, > obviously. I am not a strong PowerPC Assembly programmer, so I'm > trying to cheat by doing it in C. My entry point for the RAM test is > the romStart() routine in bootInit.c. After romInit.s masks > interrups, sets stack pointer to STACK_ADRS, and readies DRAM, it > jumps to romStart(). In romStart, before the code copies ROM to RAM, > I begin a RAM Pattern test. > > Correct me if I am wrong on these points: > > At this point the system is still executing instructions out of ROM. > Am I correct in assuming the data section is also in ROM at this > point? I have not compiled the ROM_RESIDENT bootrom (well, I tried > but it was too large to fit in EEPROM). As I understand it, the > ROM_RESIDENT bootrom only copies the data section to RAM (and .bss > exists in RAM as well), and leaves the .text section in ROM. My entry > point is before I do the copy from ROM to RAM, so I am assuming the > .data section also exists in ROM at this point. The stack however, > must exist in RAM somewhere. I am somewhat confused where it exists > from the code / headers, though. > > I see the following defined in configAll.h: > > #if ((CPU_FAMILY == MIPS) || (CPU_FAMILY == PPC) > #define STACK_RESIDENT RAM_DST_ADRS > > -- following the chain of events: RAM_DST_ADRS -> RAM_HIGH_ADRS -> > 0x00c00000 > > Then --- > > #if (_STACK_DIR == _STACK_GROWS_DOWN) > **SNIPPED STUFF ABOUT ROM_RESIDENT** > #define STACK_ADRS _romInit > > -- _romInit -> is at 0x10100 for RAM, so bootrom copies to > RAM_HIGH_ADRS, but when the system cold boots out of ROM, it must > execute _romInit from EEPROM which is in the 0xff000100 (cold entry) = > ROM_TEXT_ADRS > > #else /* STACK GROWS UP */ > #define STACK_ADRS (_romInit-STACK_SAVE) > > -- where STACK_SAVE corresponds to 64 bytes, so this is, I am assuming > the size of my STACK. Therefore, I need to avoid the addresses from > 0x00c00000 to 0x00c10000 if stack grows down, and 0x00bf0000 to > 0x00c00000 if stack grows up. > > So, where is _STACK_DIR defined? > > Also, could someone clear up the following defines: > > ROM_LOW_ADRS = 00010000 # RAM text/data address for VxWorks > ROM_HIGH_ADRS = 00c00000 # RAM text/data address for bootrom > > I see from the reference manual on bootInit that RAM_LOW_ADRS is where > exc vectors, bp anchor, exc msg, and bootline are stored, and > RAM_HIGH_ADRS is where (romInit + ROM_COPY_SIZE) or binArrayStart is > stored. So, I understand the system works in the following manner: > > _romInit executes direct assembly code, sets STACK_ADRS to > RAM_HIGH_ADRS, inits processor, sets up DRAM, etc., jumps to C routine > romStart. When do exc msg, bootline, etc. get setup? This seems to > happen in the romStart routine, but my pattern test runs as the first > piece of code in romStart (before ANY copy happens from ROM to RAM), > and when I hit address 0x4400, the system halts. This address is > defined below in configAll.h: > > #if (CPU_FAMILY == PPC) > #define RESREVED 0x4400 /* avoid zeroing out EXC_MSG */ > > This I can see if I warm reboot, and dump memory --> my pattern is > written into RAM up to the address 0x4400 at which point it continues > as all 0's. I'd like to be able to just jump over this section of RAM > with my test, but I am unsure how large the reserved area is defined > as. Actually, it would be ok, to jump over the entire exc msg, exc > vectors, bp anchor, and bootline, if I could determine how large this > section is. This would just be RAM_LOW_ADRS plus some offset. Can > someone tell me how/where it defines how large this section is? > > Finally, I would like to be able to flush / invalidate the dcache > during the pattern fill / verify tests. I tried issuing the system > calls cacheFlush() / cacheInvalidate() from the romStart() routine, > but the linker complains: > > ldppc -X -N -Map mousse.map -e _romInit -Ttext 00010000 \ > -o bootrom romInit.o bootInit.o version.o \ > D:\tornado\target/lib/libPPC603gnutms.a > D:\tornado\target/lib/libPPC603gnuospfv2.a > D:\tornado\target/lib/libPPC603gnuenvo > y.a D:\tornado\target/lib/libPPC603gnuzlib.a > D:\tornado\target/lib/libPPC603gnutms.a > D:\tornado\target/lib/libdiag.a D:\torna > do\target/lib/libtest.a D:\tornado\target/lib/libsal.a > D:\tornado\target/lib/libbcm.a D:\tornado\target/lib/libsoc.a D:\torna > do\target/lib/libsal.a D:\tornado\target/lib/libPPC603gnuvx.a > bootrom.Z.o > D:\tornado\target/lib/libPPC603gnuvx.a(workQLib.o): In function > `workQPanic': > workQLib.o(.text+0x316): undefined reference to `sysExcMsg' > workQLib.o(.text+0x31e): undefined reference to `sysExcMsg' > workQLib.o(.text+0x32a): undefined reference to `sysExcMsg' > workQLib.o(.text+0x346): undefined reference to `sysExcMsg' > workQLib.o(.text+0x34e): undefined reference to `sysExcMsg' > D:\tornado\target/lib/libPPC603gnuvx.a(workQLib.o)(.text+0x35a): more > undefined references to `sysExcMsg' follow > make: *** [bootrom] Error 0x1 > > Is it possible to somehow resolve this, or is it necessary to write a > funciton in PPC assembly to invalidate / flush the cache? > > Thanks in advance for any help, > > Dford --------------------------- Newsgroups: comp.os.vxworks Subject: vxWorks SNTP and Solaris NTP interoperability Date: Fri, 7 Dec 2001 09:34:37 -0600 From: "J. Min Kim" Organization: Nortel Message-ID: <9uqnhe$9rn$1@bcarh8ab.ca.nortel.com> Do the vxWorks SNTP client/server work with Solaris NTP server/client? According to the RFC2030 (SNTP v.4), "To a NTP or SNTP server, NTP and SNTP clients are undistinguishable; to a NTP or SNTP client, NTP and SNTP servers are undistinguishable". It seems they have the interoperability. Has anyone tried it? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: processor migration (MVME230X -> MVME5100 ALTIVEC) Date: Fri, 7 Dec 2001 17:17:50 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: Stephen wrote in message ... > >We are using VxWorks on MOTO MVME2300 boards. >We are interested in adding a motorola 5100 ALTIVEC card to an existing VME >chassis (i.e. with teh 2300 boards) >Wind River is saying that in order to support this we need to buy a whole >new development package !! >We've moved to differnet processor before and had to simply pay for the new >BSP. I can see paying a bit extra for ALTIVEC libarraies or something liek >that but a whole new dev system does not seem correct. > >I am wondeing if anyone else is facing the same issue. Has anyone migarted >from MVME2300 to 5100 ALTIVEC boards ? If you're looking for a cheap solution you can try using the 230x bsp with the altivec (just pretend it's a 604 cpu), and use my altivec compiler to compile altivec code; the main feature you'll be missing compared to WRS' setup would be the ability to use the altivec unit in multiple tasks simultaneously, although that's on my 'real soon now' to-do list... DaveK - -- Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! --------------------------- Newsgroups: comp.os.vxworks Subject: Re: symbol table Date: Fri, 7 Dec 2001 16:11:25 -0500 From: "Ren Hantao" Organization: Lucent Technologies, Columbus, Ohio Message-ID: <9urbab$pig@nntpb.cb.lucent.com> References: <9uo7r4$2s5@nntpa.cb.lucent.com> How to build the exrea step in the build process? in the Makefile? Could you describe it roughly? Thanks Hans "Johan Borkhuis" wrote in message news:Xns91705DF962435borkhuisagerecom@135.7.153.102... > "Ren Hantao" wrote: > > > Hi, vxWorks experts > > > > I want to add the symbol table to > > a standalone vxworks. but I only want > > part of the commands are supported on > > the target shell. Does anyone know how > > to do that? > > One way to do this is to symRemove to remove all unwanted symbols from the > symbol table. Another way is to build an extra step into the build process > and filter the symbol table file. > > Groeten, > Johan > > -- > o o o o o o o . . . _____________________________ > o _____ || Johan Borkhuis | > .][__n_n_|DD[ ====_____ | borkhuis@agere.com | > >(________|__|_[_________]_|__________________________| > _/oo OOOOO oo` ooo ooo 'o!o!o o!o!o` > === VxWorks FAQ: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html === > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Making a .lib file Date: Fri, 07 Dec 2001 22:04:51 GMT From: "Mike Anderson" Organization: RoadRunner - Cox Message-ID: <72bQ7.8122$MR2.1767637@typhoon.southeast.rr.com> References: <705c7b34.0111210630.7e8b496a@posting.google.com> <3BFD1245.7F0D1A81@bellsouth.net> <9torbl$rs3$1@newstest.laserlink.net> Greetings! "drdiags" wrote in message news:9torbl$rs3$1@newstest.laserlink.net... > Hello Mike, > > I thought you could change the Rule settings under the Build properties > to create a archive using the > Tornado IDE. > You are correct! However, this option in the build tab is only visible if you're working with a downloadable image rather than a bootable image. I still prefer using the command line build, though. But, I'm a geezer ;-) Regards, Mike - -- - -- ======================================================= T h e P T R G r o u p, I n c. ======================================================= Embedded, Real-Time Solutions and Training Mike Anderson mailto:mike@ThePTRGroup.com Chief Scientist http://www.ThePTRGroup.com C: (703) 585-9384 V: (703) 430-3748 ======================================================= "Software development is like making a baby... You can't make a baby in one month by impregnating nine women. Some things just take time." --------------------------- Newsgroups: comp.os.vxworks Subject: Re: need help w/ epilogue 9.1 for tor2.0.2 Date: Fri, 07 Dec 2001 22:13:56 GMT From: "Mike Anderson" Organization: RoadRunner - Cox Message-ID: References: VxWorks Greetings! "Ofer Goren" wrote in message news:newscache$bzn7ng$b3l$1@lnews.actcom.co.il... > Hi. > I have lots of problem with this suite. I don't really wanna start > describing them before I know there is someone out there that had worked w/ > it, and can truely help. Sorry. Yes, I've worked a lot with this code. My first suggestion would be to skip the Envoy 9.1 code and got to the Envoy 9.2 code. It was much more friendly to Tor 2.02 than the 9.1 code. Building this was a bear. BTW, you didn't want OSPF to be included as well did you? If so, then you'll need to build the SNMP in a rather special way to support it. Give me a shout and I'll try to help, if I can. Regards, Mike - -- - -- ======================================================= T h e P T R G r o u p, I n c. ======================================================= Embedded, Real-Time Solutions and Training Mike Anderson mailto:mike@ThePTRGroup.com Chief Scientist http://www.ThePTRGroup.com C: (703) 585-9384 V: (703) 430-3748 ======================================================= "Software development is like making a baby... You can't make a baby in one month by impregnating nine women. Some things just take time." --------------------------- Newsgroups: comp.os.vxworks Subject: Re: problem related with vxworks and TCP/IP Date: Fri, 07 Dec 2001 22:26:07 GMT From: "Mike Anderson" Organization: RoadRunner - Cox Message-ID: <3mbQ7.8227$MR2.1790673@typhoon.southeast.rr.com> References: VxWorks Greetings! > I set End attach interface support and End interface support along > with PPP & PPP boot in my build. > I am getting confirmation that both the tcp/ip and loopback > initilization is success as below: The Simulator does not use PPP on a Win2K machine. It uses the ULIP driver: >---- VxSim Output --------------------------- >Attached TCP/IP interface to nt unit 0 <======= This is the ULIP driver. Therefore, install the ULIP driver on your Win2K box (make sure that you're using the one from the Win2K Tornado patch), and you'll be able to use the network features (if you have the full simulator that is). 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." --------------------------- Newsgroups: comp.os.vxworks Subject: Re: RADVision H.323 Stack Sample Date: Fri, 07 Dec 2001 22:46:41 GMT From: "Mike Anderson" Organization: RoadRunner - Cox Message-ID: References: <9un53i$r6c$1@mail.cn99.com> VxWorks Greetings! "zhaoyandong" wrote in message news:9un53i$r6c$1@mail.cn99.com... > Our embedded system is going to use Radvision RTP/RTCP Stack. > > But, I don't know how to use it. I need a sample of using RTP/RTCP routines > under VxWorks. > RADVision had a few of these as test cases. Make sure that you're using the Version 3 RADVision stack. What CPU architecture are you using? What host system? I had actually gotten the RADVision code to rebuild on a Win2K box and target an PPC 8260. This was no small feat as the codebase appears to be written to compile only on a Solaris host. I've tried to contact the RADVision folks here in the US (NJ) to try and get the build approach and code changes back into the product. But, they do not respond to my e-mails. Regards, 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." --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Need Communication of Serial Port with MPC860 Target Date: Fri, 07 Dec 2001 22:55:33 GMT From: "Frank" Message-ID: References: <23de5426.0112042101.1daaf7be@posting.google.com> <23de5426.0112052245.72ffee4c@posting.google.com> Hi, I think you should be clear about these 2 issue: 1. how to use MPC860 serial port 2. how to control Modem so, here is my understanding: 1. VxWorks use Tty driver to handle user application requre for serial port. you can use standard I/O, through tty, access your serial port. normally, the board should suport 1 SIO port, default used for console. If you want use the second serial port while your BSP doesn't support that, you should modify sysSerial.c , or other relative .H files maybe to support it. You will find some advice and guide in Device Driver training workshop of WRS. 2. After you correctly use your serial port, you can control your modem, I guess you use voiceband dial modem. So, you should use AT command, as you do in other communication application under Windows HTH Frank. - -- "B.Manivannan" wrote in message news:23de5426.0112052245.72ffee4c@posting.google.com... > Mr.Frank, > > What you said is perfectly correct. I analyzed it. > > After downloading the application into the target(MPC860), the target > must connect to the serial device(for example modem in my case). Like > simulator it is showing the output "abcdefghij" in the VxWorks window. > > My doubt is, if it had been connected the MODEM, it should have > sent/received the signal. ( This i tested it through Visual C++ > application,modem Tx/Rx lights glowed ) > > My senior suggested me that i should use the function from sysSerial.c > file from C:\Tornado\target\config\estMDB8xx\sysSerial.c. As for as my > knowledge is concerned we need to use this file to develop the device > driver for the serial hardware. > > How can i connect the MODEM? Do i need to install the BSP package and > use that? Or Is there any configuration to connect the serial port for > the MPC860 target?? > > Could you guide me??? > > With Thanks and Regards, > B.Manivannan. > > > > > Hello, > 1. Technically, Your rountie will work at 860board, normally, tyCo/0 > is > connected to SMC1, depend on your board and BSP. > > 2. Serial Port is not supported in VxSim. > > Frank --------------------------- Newsgroups: comp.os.vxworks Subject: Re: cache and interrupts? Date: Fri, 7 Dec 2001 19:48:31 -0500 From: "BruceL" Message-ID: <9uro2v$3kb$1@bob.news.rcn.net> References: <9up06c$mqt$1@bob.news.rcn.net> <488e459a.0112062254.210b2a7b@posting.google.com> Hi John, The symptom was on our MPC8245 : with instruction cache enabled, on the first decrementer exception, the system would advance through the exception code (at 900) but apparently not execute anything. It would then fall through the branch to intExit and choke on the first bad instruction. Turns out, MCCR4 (memory config 4) has to be initialized before turning cache on, even though no SDRAM is used until much later in the initialization. I don't really know why just yet. However, if MCCR4 is initialized prior to cache enabling, everything works just fine. Perhaps the activation of cache causes some default memory to be loaded? While you're here, do you know how to use the BAT registers to isolate certain memory areas so that they cannot be cached? Thanks, Bruce "John" wrote in message news:488e459a.0112062254.210b2a7b@posting.google.com... > Hello, > > A lot more information is going to be required before we can add > anything helpful... like: > > 1) What is the exception you are getting? > > 2) How is your application "handling" a decrementer exception (that is > normally connected to the VxWorks system clock)? > > 3) What BSP are you using? > > 4) Does the kernel work without your application included? If so, is > time moving forward correctly (check with tickGet() or taskDelay())? > > Rgrds, > > John... > > "BruceL" wrote in message news:<9up06c$mqt$1@bob.news.rcn.net>... > > Can anyone explain to me why my app should crash during the handling of a > > decrementer (real-time clock on the MPC 8245) exception if the instruction > > cache is enabled? This is the first exception the app sees. > > > > Thanks, > > Bruce --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Software excuting slowly Date: Fri, 7 Dec 2001 19:51:42 -0500 From: "BruceL" Message-ID: <9uro8u$4bm$1@bob.news.rcn.net> References: <9uqnr6$o7s$1@newsie.singa.pore.net> I just had some similar experiences. Make sure that your build includes cache enabling and that the instruction cache is enabled for flash areas as well as RAM. Best of luck. Bruce "Chet Wee" wrote in message news:9uqnr6$o7s$1@newsie.singa.pore.net... > Hi all, > > My team has developed a software that is ~4MB. > > I used the Tornado Integrated Development Environment to compile/link & then > download the software to the SDRAM for execution. (Option 1) > > Then I tried to flash the software into the Flash, using VxResRom build. The > software runs very slowly in the Flash, even with the same compilation > option. (Option 2) > > Then I switched to using VxRes build. It runs faster than theVxResRom > option, as expected, but I expect it to run as the same speed as option (1), > but it is running slower than Option (1) by almost 2 to 3 times. > > Can anyone please advise why is this so. > > Initially, I suspected that it could be caused by the unwanted interrupts, > but I am using the same software as the Option (1). > > > Please advise. > Anyway, I am using the PPC740. > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: t2 project facility: possible to avoid compiling files based on flags? Date: 8 Dec 2001 01:16:14 GMT From: Denis Perelyubskiy Message-ID: hello, does anyone know what would be the cleanest way to exclude some files from being compiled, other then remove them from the project? basically, i am looking for some facility along the lines of the following piece of makefile: ifdef MYFEATURE OBJS += myobj.o endif and then you'd have rules to make .o's from .c's, or use defaults, or whatnot. this is not important. so, how do i specify this 'ifdef MYFEATURE' thing? 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: HELP!!! in DHCP Date: 8 Dec 2001 00:23:21 -0800 From: shafi4u@rediffmail.com (shafi) Organization: http://groups.google.com/ Message-ID: References: <3c03db4c$1@news.cadence.com> <3c0fee91$1@news.cadence.com> Hi, dhcpBind failure may be due to no offer from server in requred time. Some servers won't OFFER to the discover messages whose header type field is other than '1'. VxWorks DHCP-clients assign this value as 6. Since this value is picked from ifnet structure, try something like this to receive OFFER first: struct ifnet *pIf = ifunit("infName"); ..... pIf->if_type = 6; lease = dhcpcInit( pif, 1 ); pIf->if_type = 1; dhcpcBind( lease, synchronous ); ...... You may ensure the reception of OFFER by binding some socket on port 68. and receving such messages from SERVER. VxWorks DHCPC picks up these messages through etherInputHooks. some times thes HOOKs doesn't not work, and client will not receive any messages from SERVER and thus fail to BIND. -shafi "naveen" wrote in message news:<3c0fee91$1@news.cadence.com>... > Thanks for the info. > > But I don't understand when I try to use the following > lease = dhcpcInit( net_if, 1 ); > dhcpcBind( lease, synchronous ); > > it fails in bind.I don't know why. > Am I doing someting wrong. > Do you know what is the exact flow of the code or I need to do something > else. > > Help will be highly appreciated. > --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Sun Dec 9 04:03:09 2001 From: Vxworks Exploder Date: Sun Dec 9 04:03:12 PST 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sun Dec 9 04:03:06 PST 2001 Subject: Other then WDB,how can i use END network driver in polled mode? Subject: Re: Software excuting slowly Subject: Enabling of cache & instruction cache Subject: Re: Software excuting slowly Subject: Re: HELP!!! in DHCP Subject: Re: cache and interrupts? Subject: `Higher reliability' DosFs. Subject: getting arrow keys Subject: VxSIM Networking Help Subject: How to solve this problem? chech sum error! ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Other then WDB,how can i use END network driver in polled mode? Date: 8 Dec 2001 04:26:52 -0800 From: binagpatel@yahoo.com (Bina Patel) Organization: http://groups.google.com/ Message-ID: <65b68b90.0112080426.24c67671@posting.google.com> Hello! I am working with END network driver.i want that my driver should work in polled mode. For my quetion "How to invoke polled function in END network driver?". i got reply from group that WDB puts END driver in polled mode. Now my quetion is"Other then WDB,how can i use END network driver in polled mode?" I shall be greatful for any link on same. Bina. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Software excuting slowly Date: Sat, 8 Dec 2001 21:31:38 +0800 From: "Chet Wee" Organization: SCV Cablemodem News site Message-ID: <9ut53e$med$1@newsie.singa.pore.net> References: <9uqnr6$o7s$1@newsie.singa.pore.net> <9uro8u$4bm$1@bob.news.rcn.net> Thanks Bruce Would appreciate if you can advise me on specifically what option I must set in the build option Thanks "BruceL" wrote in message news:9uro8u$4bm$1@bob.news.rcn.net... > I just had some similar experiences. Make sure that your build includes > cache enabling and that the instruction cache is enabled for flash areas as > well as RAM. > > Best of luck. > Bruce > > > "Chet Wee" wrote in message > news:9uqnr6$o7s$1@newsie.singa.pore.net... > > Hi all, > > > > My team has developed a software that is ~4MB. > > > > I used the Tornado Integrated Development Environment to compile/link & > then > > download the software to the SDRAM for execution. (Option 1) > > > > Then I tried to flash the software into the Flash, using VxResRom build. > The > > software runs very slowly in the Flash, even with the same compilation > > option. (Option 2) > > > > Then I switched to using VxRes build. It runs faster than theVxResRom > > option, as expected, but I expect it to run as the same speed as option > (1), > > but it is running slower than Option (1) by almost 2 to 3 times. > > > > Can anyone please advise why is this so. > > > > Initially, I suspected that it could be caused by the unwanted interrupts, > > but I am using the same software as the Option (1). > > > > > > Please advise. > > Anyway, I am using the PPC740. > > > > > > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Enabling of cache & instruction cache Date: Sat, 8 Dec 2001 22:45:38 +0800 From: "Chet Wee" Organization: SCV Cablemodem News site Message-ID: <9ut9e6$okm$1@newsie.singa.pore.net> Please help. I urgently needs help on specifically how i) to enable cache ii) to enable instruction cache for flash areas as well as RAM. in my build option. My CPU is PPC740 Thanks --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Software excuting slowly Date: Sat, 8 Dec 2001 11:25:03 -0500 From: "BruceL" Message-ID: <9utetm$300$1@bob.news.rcn.net> References: <9uqnr6$o7s$1@newsie.singa.pore.net> <9uro8u$4bm$1@bob.news.rcn.net> <9ut53e$med$1@newsie.singa.pore.net> Our BSP had a variable like ENABLE_I_CACHE in the config.h file. Do a grep on CACHE in your BSP headers. There is also a component in the project window under hardware | memory, but that didn't appear to have any effect on our system. Bruce "Chet Wee" wrote in message news:9ut53e$med$1@newsie.singa.pore.net... > Thanks Bruce > > Would appreciate if you can advise me on specifically what option I must set > in the build option > > > Thanks > "BruceL" wrote in message > news:9uro8u$4bm$1@bob.news.rcn.net... > > I just had some similar experiences. Make sure that your build includes > > cache enabling and that the instruction cache is enabled for flash areas > as > > well as RAM. > > > > Best of luck. > > Bruce > > > > > > "Chet Wee" wrote in message > > news:9uqnr6$o7s$1@newsie.singa.pore.net... > > > Hi all, > > > > > > My team has developed a software that is ~4MB. > > > > > > I used the Tornado Integrated Development Environment to compile/link & > > then > > > download the software to the SDRAM for execution. (Option 1) > > > > > > Then I tried to flash the software into the Flash, using VxResRom build. > > The > > > software runs very slowly in the Flash, even with the same compilation > > > option. (Option 2) > > > > > > Then I switched to using VxRes build. It runs faster than theVxResRom > > > option, as expected, but I expect it to run as the same speed as option > > (1), > > > but it is running slower than Option (1) by almost 2 to 3 times. > > > > > > Can anyone please advise why is this so. > > > > > > Initially, I suspected that it could be caused by the unwanted > interrupts, > > > but I am using the same software as the Option (1). > > > > > > > > > Please advise. > > > Anyway, I am using the PPC740. > > > > > > > > > > > > > > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: HELP!!! in DHCP Date: 8 Dec 2001 08:33:27 -0800 From: shafi4u@rediffmail.com (shafi) Organization: http://groups.google.com/ Message-ID: References: <3c03db4c$1@news.cadence.com> <3c0fee91$1@news.cadence.com> Hi, There is a mistake in my previous message, i.e., Thread 5 of this article. so please ignore Thread 5. And prefer this message. dhcpBind failure may be due to no offer from server in requred time. Some servers won't OFFER to the discover messages whose header type field is other than '1'. VxWorks DHCP-clients assign this value as 6. Since this value is picked from ifnet structure, try something like this to receive OFFER first: struct ifnet *pIf = ifunit("infName"); ..... pIf->if_type = 1; /* To ensure "htype" field assigned 1 for ETHER 10 MB */ lease = dhcpcInit( pif, 1 ); pIf->if_type = 6; dhcpcBind( lease, synchronous ); ...... This worked for me atleast. You may ensure the reception of OFFER by binding some socket on port 68. and receving such messages from SERVER. VxWorks DHCPC picks up these messages through etherInputHooks. some times thes HOOKs doesn't not work, and client will not receive any messages from SERVER and thus fail to BIND. -shafi shafi4u@rediffmail.com (shafi) wrote in message news:... > Hi, > dhcpBind failure may be due to no offer from server in requred time. > Some servers won't OFFER to the discover messages whose header type > field is other than '1'. VxWorks DHCP-clients assign this value as 6. > Since this value is picked from ifnet structure, try something like this > to receive OFFER first: > > struct ifnet *pIf = ifunit("infName"); > ..... > pIf->if_type = 6; > lease = dhcpcInit( pif, 1 ); > pIf->if_type = 1; > dhcpcBind( lease, synchronous ); > ...... > You may ensure the reception of OFFER by binding some socket on port 68. > and receving such messages from SERVER. VxWorks DHCPC picks up these > messages through etherInputHooks. some times thes HOOKs doesn't not work, > and client will not receive any messages from SERVER and thus fail to BIND. > > -shafi > > "naveen" wrote in message news:<3c0fee91$1@news.cadence.com>... > > Thanks for the info. > > > > But I don't understand when I try to use the following > > lease = dhcpcInit( net_if, 1 ); > > dhcpcBind( lease, synchronous ); > > > > it fails in bind.I don't know why. > > Am I doing someting wrong. > > Do you know what is the exact flow of the code or I need to do something > > else. > > > > Help will be highly appreciated. > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: cache and interrupts? Date: 8 Dec 2001 09:56:12 -0800 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0112080956.60e9aa41@posting.google.com> References: <9up06c$mqt$1@bob.news.rcn.net> <488e459a.0112062254.210b2a7b@posting.google.com> <9uro2v$3kb$1@bob.news.rcn.net> Hello, Glad to hear you solved you problem... and thanks for posting the solution too - will make it available for everybody else too. > While you're here, do you know how to use the BAT registers to isolate > certain memory areas so that they cannot be cached? I don't have any example code for setting up these registers, but it should be possible. One thing to be careful of though is using them and the MMU. If you plan to leave the MMU off all the time, using the BAT registers is safe (just make sure they don't overlap each other). If you plan to use the MMU as well, be careful not to overlap the MMU and BAT regions... Take a look in sysLib.c for the sysBatDesc array, and h/arch/ppc/mmu603Lib.h for the possible options. Might also be worth reading about the BATs in the PowerPC Programming Environments book - be warned though, the MMU chapter of that book is not for the faint hearted ;-) HTH, John... > Thanks, > Bruce > > > "John" wrote in message > news:488e459a.0112062254.210b2a7b@posting.google.com... > > Hello, > > > > A lot more information is going to be required before we can add > > anything helpful... like: > > > > 1) What is the exception you are getting? > > > > 2) How is your application "handling" a decrementer exception (that is > > normally connected to the VxWorks system clock)? > > > > 3) What BSP are you using? > > > > 4) Does the kernel work without your application included? If so, is > > time moving forward correctly (check with tickGet() or taskDelay())? > > > > Rgrds, > > > > John... > > > > "BruceL" wrote in message > news:<9up06c$mqt$1@bob.news.rcn.net>... > > > Can anyone explain to me why my app should crash during the handling of > a > > > decrementer (real-time clock on the MPC 8245) exception if the > instruction > > > cache is enabled? This is the first exception the app sees. > > > > > > Thanks, > > > Bruce --------------------------- Newsgroups: comp.os.vxworks Subject: `Higher reliability' DosFs. Date: 08 Dec 2001 19:03:59 -0500 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: Sender: bpringlemeir@DeadDuck I am trying to make a system that `shadows' DosFs files. My current model is simply to add a sequence number and CRC to each file. When a file is being updated, it is written as "filename.a" or "filename.b". The files are always closed upon an update. When reading files, the one with greatest sequence number and a correct CRC is chosen. If both files have a good CRC, the lower sequence number file is deleted. File access is already guarded by semaphores. Will this guarantee that I will always have a good DosFs file when using the caching mechanism? Is anyone kind enough to point be to some simple database theory? Any source...? tia, Bill Pringlemeir. - -- Well, I'm a classic ANAL RETENTIVE!! And I'm looking for a way to VICARIOUSLY experience some reason to LIVE!! vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" --------------------------- Newsgroups: comp.os.vxworks Subject: getting arrow keys Date: Sun, 9 Dec 2001 00:12:41 -0800 From: "Hashim Qaderi" Organization: Bell Sympatico Message-ID: hello, Does anyone know how to get arrow keys from the keyboard directly. When I press arrows keys the cursor on the vxSim goes up and down rather than getting the character itself. thanks in advance. Jacob. --------------------------- Newsgroups: comp.os.vxworks Subject: VxSIM Networking Help Date: 8 Dec 2001 22:37:56 -0800 From: rashanmu@npd.hcltech.com (Ramesh) Organization: http://groups.google.com/ Message-ID: <15a87607.0112082237.32ab8591@posting.google.com> Hi, I found a message by Luc in this group some time before and i did the steps as he mentioned in detail. I have mentioned that mail in the bottom of this post. But i could not able to ping 90.0.0.1 or 90.0.0.254 addresses from the shell. But from the DOS prompt i can able to ping 90.0.0.254. My hostShow and ifShow commanfs in VxSIm produces the same result and mentioned but i could not able to ping. My route show command in vxsim shown below. - -> routeShow ROUTE NET TABLE destination gateway flags Refcnt Use Interface - ---------------------------------------------------------------------------- - ---------------------------------------------------------------------------- ROUTE HOST TABLE destination gateway flags Refcnt Use Interface - ---------------------------------------------------------------------------- 127.0.0.1 127.0.0.1 5 0 0 lo0 - ---------------------------------------------------------------------------- value = 77 = 0x4d = 'M' - -> My route PRINT command from DOS prompt also i have given here. My host IP address is 192.168.100.69. I have enabled IP forwarding. BUt i dont know where am missing. i could not able to ping from 90.0.0.1 to 90.0.0.2 or vice versa. While configuring the ULIP adapter i have given the IP address for it as 90.0.0.254. i haven;t given any default gateway address. whether this could be the problem. If so what is the address i have to give. Kindly help me in this regard. C:\>route PRINT =========================================================================== Interface List 0x1 ........................... MS TCP Loopback interface 0x2 ...00 02 b3 20 e1 3d ...... Intel(R) PRO Adapter 0x3 ...00 00 00 00 00 00 ...... WRS Virtual Adapter. =========================================================================== =========================================================================== Active Routes: Network Destination Netmask Gateway Interface Metric 0.0.0.0 0.0.0.0 192.168.100.1 192.168.100.69 1 90.0.0.0 255.255.255.0 90.0.0.254 90.0.0.254 1 90.0.0.254 255.255.255.255 127.0.0.1 127.0.0.1 1 127.0.0.0 255.0.0.0 127.0.0.1 127.0.0.1 1 192.168.100.0 255.255.255.0 192.168.100.69 192.168.100.69 1 192.168.100.69 255.255.255.255 127.0.0.1 127.0.0.1 1 192.168.100.255 255.255.255.255 192.168.100.69 192.168.100.69 1 224.0.0.0 224.0.0.0 90.0.0.254 90.0.0.254 1 224.0.0.0 224.0.0.0 192.168.100.69 192.168.100.69 1 255.255.255.255 255.255.255.255 192.168.100.69 192.168.100.69 1 ========================================================================= THis mail is the previous post i found on this group. I have snipped the detail explanation part. SOLUTION! While in a VxWorks class, the instructor gave me the following info. With this I was able to finally get it to work! Just two things: 1) Make sure you configure WDB END driver instead of the WDB simulator pipe. development tool components -> WDB agent components -> select WDB connection -> WDB END driver connection 2) I got an exception using buffers greater then 1400 with a TCP connection. Fragmentation problem? Same code works fine on real hardware. Luc Goal: How to do networking with SIM NT Goal: Setup Networking on Full VxSim (SIM NT) on Windows NT Fact: VxSim for Windows NT Fact: Windows NT 4.0 Fact: Tornado / 2.0 Fact: VxWorks / 5.4 Fact: simpc Fix: This solution describes how to configure the full VxSim or Sim NT on WinNT, build VxWorks, launch two VxSim targets, and ping each other. 1. Confirm the Full VxSim product is installed. 2. Install and Configure ULIP. 3. Turn on IP Forwarding on the NT 4. Modify target/config/simpc/config.h 5. Create a New Project 6. Configure the Project and Build VxWorks 7. Launch VxSim target and attach Target Server 8. Launch Second VxSim target and target server 9. Ping the VxSim targets. Thanks Ramesh --------------------------- Newsgroups: comp.os.vxworks Subject: How to solve this problem? chech sum error! Date: 9 Dec 2001 03:16:09 -0800 From: youerming@163.com (Torna) Organization: http://groups.google.com/ Message-ID: my debugger enviroment includes tornadoII, samsung snds100 demo board for arm7. i have finished connecting target board with my tornadoII. but during i started up my target server,i got message say that the downloaded image's check sum is not equal with one of my host image .i have referred to the tornado user's guide and trid to change my core file and symbol,but i cant get satisfactory result yet. Could anybody give me a hint that how to configure my target server to solve this problem? Please give an example in detail ,thanks a lot ! hongwind --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Sun Dec 9 17:52:32 2001 From: karunakaran nair Date: Sun Dec 9 17:52:34 PST 2001 Subject: Re: VxWorks - CPU Monitoring how is the job situation for people with average knowledge of vxworks, real time systems etc and more than 2 years experience ? KN __________________________________________________ Do You Yahoo!? Send your FREE holiday greetings online! http://greetings.yahoo.com From vxwexplo-errs@csg.lbl.gov Sun Dec 9 19:55:39 2001 From: Kirti Mahajan Date: Sun Dec 9 19:55:41 PST 2001 Subject: VME Interrupt Latency Hello, Our Processor Model is PPC604, running under vxWorks5.4. Chasis is 32bit VME backplane. What VME interrupt latency one should get on such configuration ? Whether benchmark are available for this? Thanks and Regards --Kirti Mahajan From vxwexplo-errs@csg.lbl.gov Mon Dec 10 04:03:14 2001 From: Vxworks Exploder Date: Mon Dec 10 04:03:16 PST 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Mon Dec 10 04:03:10 PST 2001 Subject: Re: How to solve this problem? chech sum error! Subject: Vxworks PPP component Subject: development board kit Subject: Re: Vxworks PPP component Subject: Re: getting arrow keys Subject: what is the use of rtm and rwos directory in ospf directory( vxworks )? Subject: (urgent) Other then WDB,how can i use END driver in polled mode? Subject: spurious interrupts! Subject: Re: spurious interrupts! Subject: Re: (urgent) Other then WDB,how can i use END driver in polled mode? Subject: Abstruct class (C++) with VxWorks Subject: Send message to the target from Window's application Subject: POSIX problem(mq_open) Subject: END driver debugging Subject: Re: Need Communication of Serial Port with MPC860 Target Subject: Re: need help w/ epilogue 9.1 for tor2.0.2 Subject: Re: t2 project facility: possible to avoid compiling files based on flags? Subject: Re: Other then WDB,how can i use END network driver in polled mode? Subject: Re: `Higher reliability' DosFs. Subject: IP_HDRINCL on VxWorks ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to solve this problem? chech sum error! Date: 9 Dec 2001 06:11:44 -0800 From: rashanmu@npd.hcltech.com (Ramesh) Organization: http://groups.google.com/ Message-ID: <15a87607.0112090611.125558d4@posting.google.com> References: youerming@163.com (Torna) wrote in message news:... > my debugger enviroment includes tornadoII, samsung snds100 demo board > for arm7. > > i have finished connecting target board with my tornadoII. > but during i started up my target server,i got message say that the > downloaded image's check sum is not equal with one of my host image .i > have referred to the tornado user's guide and trid to change my core > file and symbol,but i cant get satisfactory result yet. > > Could anybody give me a hint that how to configure my target server > to solve this problem? > Please give an example in detail ,thanks a lot ! > hongwind - ----------------------------------------------------------------------------- Hi, THe problem could be while starting the Target server, it would have asked for the image to which it has to connect. If you have given a image which is not the same image with which you have started the target server, then this error will occur. Since the checksum of the images of which you downloaded and the host image will be different and hence the target server could not able to connect. Try to check the image you loaded as target and the one which you are connecting through the target server are one and the same. Thanks, Ramesh --------------------------- Newsgroups: comp.os.vxworks Subject: Vxworks PPP component Date: Sun, 9 Dec 2001 17:46:58 +0200 From: "Martin Roth" Organization: Motorola Message-ID: <9v012a$l09$1@newshost.mot.com> Did anybody use the PPP component that is provided with Vxworks. Is it reliable ? I want to use it as a PPP client that dials to a PPP server. It is royaltee free and is provides with Vxworks as a library component. The reason I am asking this question is because Windrivers offers another PPP component called Windnet PPP. It is very expensive and it also includes royalties. I am not sure if I need all the features of Windnet PPP for maximum 2 PPP connections? Thank you Martin --------------------------- Newsgroups: comp.os.vxworks Subject: development board kit Date: Sun, 9 Dec 2001 11:39:23 -0800 From: "Jacob" Organization: Bell Sympatico Message-ID: <9kMQ7.1580$eF1.175242@news20.bellglobal.com> Hello, I am looking for development board kit that that has the basics for real-time programming: - -microcontroller, ram, flash, serial and/or ethernet connection I should be able to download and run a real time operating system..(e.g psos or vxworks). This is for educational purposes only so it would have to be inexpensive. Can anyone tell where I can get more informarion or a place that sells these types of kits ? Thanks, Jacob --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Vxworks PPP component Date: 09 Dec 2001 18:57:25 -0500 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: References: <9v012a$l09$1@newshost.mot.com> Sender: bpringlemeir@DeadDuck >>>>> "Martin" == Martin Roth writes: Martin> Did anybody use the PPP component that is provided with Martin> Vxworks. Is it reliable ? I want to use it as a PPP client Martin> that dials to a PPP server. It is royaltee free and is Martin> provides with Vxworks as a library component. Martin> The reason I am asking this question is because Windrivers Martin> offers another PPP component called Windnet PPP. It is very Martin> expensive and it also includes royalties. I am not sure if I Martin> need all the features of Windnet PPP for maximum 2 PPP Martin> connections? The WindNet product is PPPoE (afaik). This is a version of PPP that works with cable modems and ADSL products. It is for `broadband access'. The simple PPP that comes with vxWorks use a serial line (or other character device). You can have at least two connection; I have this on a system I work on. I do not believe that you have to buy the WindNet product. hth, Bill Pringlemeir. - -- security Semtex bomb explosion ammunition CIA $400 million in gold bullion Peking Legion of Doom smuggle Nazi Qaddafi SDI Kennedy KGB vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" --------------------------- Newsgroups: comp.os.vxworks Subject: Re: getting arrow keys Date: Mon, 10 Dec 2001 03:32:21 GMT From: Gwen & Emory Stagmer Organization: Excite@Home - The Leader in Broadband http://home.com/faster Message-ID: <3C142C96.F1A3EE5@home.com> References: You have to get multiple characters on a PC. First charcter tells you it's a cursor control character, second one tells you which one. Sorry, don't have the specifics here, but I'm sure you can find it by hunting around on the internet. If not write me back and I'll hunt it down for you. I know I've got it handy at work... Emory_Stagmer@LittonAS.com Emory Stagmer Northrup Grumman Space Tech & Services Software Group Hashim Qaderi wrote: > > hello, > > Does anyone know how to get arrow keys > from the keyboard directly. When I press > arrows keys the cursor on the vxSim goes > up and down rather than getting the character itself. > > thanks in advance. > Jacob. --------------------------- Newsgroups: comp.os.vxworks Subject: what is the use of rtm and rwos directory in ospf directory( vxworks )? Date: 9 Dec 2001 19:35:43 -0800 From: mingxinmu@hotmail.com (mumx) Organization: http://groups.google.com/ Message-ID: <75f406a7.0112091935.4470efcb@posting.google.com> I am reading ospf v2 program in vxworks now. Could anybody tell me what is the use of the directories named "rtm" and "rwos" under the ospf directory? Where can I find a help list of the files in vxworks directories? Thank you in advance! --------------------------- Newsgroups: comp.os.vxworks Subject: (urgent) Other then WDB,how can i use END driver in polled mode? Date: 9 Dec 2001 20:29:28 -0800 From: binagpatel@yahoo.com (Bina Patel) Organization: http://groups.google.com/ Message-ID: <65b68b90.0112092029.1cc2bf50@posting.google.com> Hello! I am working with END network driver.i want that my driver should work in polled mode. For my quetion "How to invoke polled function in END network driver?". i got reply from group that WDB puts END driver in polled mode. Now my quetion is"Other then WDB,how can i use END network driver in polled mode?" I shall be greatful for any link on same as soon as possible. Bina. --------------------------- Newsgroups: comp.os.vxworks Subject: spurious interrupts! Date: 9 Dec 2001 21:18:47 -0800 From: bondi_ameya@yahoo.com (Ameya) Organization: http://groups.google.com/ Message-ID: <57c7939b.0112092118.81fddd8@posting.google.com> hi, i have an END Driver in vxworks. It has an ISR hooked to IRQL 11. im using a pentium BSP on a x86 board. I seem to be missing interrupts at irregular time intervals. i can see the spurious interrupt coz i8259IntBoi() is called (irq num :7) I also see the EOI from the slave...but still i seem to miss my interrupt. any ideas? thanks ameya --------------------------- Newsgroups: comp.os.vxworks Subject: Re: spurious interrupts! Date: Mon, 10 Dec 2001 08:19:29 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <57c7939b.0112092118.81fddd8@posting.google.com> It is hard to guess without any specifics on the chipset, but sounds like a chipset glitch to me, and there eint much you can do bout it. Here is how I would work around this problem - switch all interrupts to "Level Mode" rather then "Edge Mode" trigerring, which should be pretty easy if all drivers you use are properly handling interrupts, i.e. Level Trigerring is veru unforegiving, if an ISR forgets to turn off the cause of the interrupt, the ISR will be re-entered infinitely and the system will seem to freeze. On the positive side, in Level mode interrupts are NEVER missed. - - Leonid "Ameya" wrote in message news:57c7939b.0112092118.81fddd8@posting.google.com... > hi, > i have an END Driver in vxworks. It has an ISR hooked to IRQL 11. > im using a pentium BSP on a x86 board. > I seem to be missing interrupts at irregular time intervals. > i can see the spurious interrupt coz i8259IntBoi() is called (irq num :7) > I also see the EOI from the slave...but still i seem to miss my interrupt. > any ideas? > > thanks > ameya --------------------------- Newsgroups: comp.os.vxworks Subject: Re: (urgent) Other then WDB,how can i use END driver in polled mode? Date: Mon, 10 Dec 2001 08:26:30 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <65b68b90.0112092029.1cc2bf50@posting.google.com> You see Bina, The answer to this seems incredibly simple at first, there is an API function which switches an END driver (mich like an SIO driver) between INT and POLL modes, so to call this API and make the switch is easy enough. Or it would be quite easy indeed should this be about an SIO driver ... But END driver is a network device, and aside from WDB, it is not used directly but rather via the TCP/IP and Sockets layers, and the standard SENS stack does not know HOW to use an END driver in POLLED mode, and it never could. Full fledged implementation of TCP/IP would be probably uanble to use a network driver in polled mode. WDB over END is a light-weight implementation of UDP/IP, specifically designed to be able to work in POLLEd mode, hence it can. So unless you are going to call the END driver API DIRECTLY to send and receive MAC packets, the information you requested is quite useless for you. Hope this helps, and let me know if you need my mailing address where you can mail your gratitude. - - Leonid "Bina Patel" wrote in message news:65b68b90.0112092029.1cc2bf50@posting.google.com... > Hello! > > I am working with END network driver.i want that my driver should work > in polled mode. For my quetion "How to invoke polled function in END > network driver?". > i got reply from group that WDB puts END driver in polled mode. > Now my quetion is"Other then WDB,how can i use END network driver in > polled mode?" > I shall be greatful for any link on same as soon as possible. > > Bina. --------------------------- Newsgroups: comp.os.vxworks Subject: Abstruct class (C++) with VxWorks Date: 9 Dec 2001 22:33:52 -0800 From: glazerh@rafael.co.il (Yaron) Organization: http://groups.google.com/ Message-ID: <4367c1ab.0112092233.2bede9f6@posting.google.com> When I try to download an application which is writen in CPP and which contains an Abstruct class I get a "download error" message. Whats wrong??? --------------------------- Newsgroups: comp.os.vxworks Subject: Send message to the target from Window's application Date: 9 Dec 2001 22:38:50 -0800 From: glazerh@rafael.co.il (Yaron) Organization: http://groups.google.com/ Message-ID: <4367c1ab.0112092238.7630e8bd@posting.google.com> How can I send/post a message (or even use a pipe etc') from Windows (Visual C++ / MFC) application to VxWorks application ?. --------------------------- Newsgroups: comp.os.vxworks Subject: POSIX problem(mq_open) Date: Mon, 10 Dec 2001 16:18:22 +0900 From: "Á¶Çö¿ì" Organization: Inet Internet Services Message-ID: <9v1nja$164$1@news.nuri.net> I use the Tornado version 2.0 I want to set a attribute of message queue!! execute : module 1 ---> module 2 module 1 g_szFileName = "messagequeue"; flags = O_RDWR | O_CREAT | O_EXCL; attr->mq_flags = NULL; attr->mq_maxmsg = 10; attr->mq_msgsize = 1024; attr->mq_curmsgs = 0; mqd = mq_open (g_szFileName, flags, FILE_MODE, attr); if (mqd < 0) perror("error"); mq_getattr (mqd, attr); printf("max #Msgs =%ld, max $Bytes/Msg = %ld, #currently on Queue = %ld\n", attr->mq_maxmsg, attr->mq_msgsize, attr->mq_curmsgs); result : max #Msgs =10, max $Bytes/Msg = 1000, #currently on Queue = 0 but anothother module module 2 int MqGetAttr (void) { mqd_t mqd; struct mq_attr attr; mqd =mq_open (g_szFileName, O_RDONLY); mq_getattr (mqd, &attr); printf("max #Msgs =%ld, max $Bytes/Msg = %ld, #currently on Queue = %ld\n", attr.mq_maxmsg, attr.mq_msgsize, attr.mq_curmsgs); mq_close(mqd); return; } result : max #Msgs =16, max $Bytes/Msg = 16, #currently on Queue = 0 unreasonalbe result appear!!!! I want to solve this problem!!! --------------------------- Newsgroups: comp.os.vxworks Subject: END driver debugging Date: 9 Dec 2001 23:48:04 -0800 From: mistry_amit@yahoo.com (Amit) Organization: http://groups.google.com/ Message-ID: <3ca555c5.0112092348.5350ef2a@posting.google.com> Hi, I am working on END driver for Wireless LAN card on PC Pentium platform. The END driver is registered properly and working fine when I make VxWorks bootable image over floppy. But it has limitations for debugging the END driver. Now I want to boot my VxWorks Image with END driver over Ethernet and debug from Host machine. Do two drivers for LAN card (one is for Ethernet booting and one is for Wireless card)work together?? My Tornado host is running on PC pentium - WIN NT platform and Target is also PC pentium. If somebody has done exercise on END driver debugging over Ethernet, Please give me your inputs for debugging this way. And if anybody knows other tricks for debugging END driver, please write me. Thanks in advance. Amit --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Need Communication of Serial Port with MPC860 Target Date: 10 Dec 2001 00:03:54 -0800 From: manibala@covansys.com (B.Manivannan) Organization: http://groups.google.com/ Message-ID: <23de5426.0112100003.35a731c6@posting.google.com> References: <23de5426.0112042101.1daaf7be@posting.google.com> <23de5426.0112052245.72ffee4c@posting.google.com> Mr.Frank, You had mentioned that the MPC860 supports single SIO by default for console. What does it mean??? For example My PC(host) connected with the target(MPC860) by serial and also by Ethernet. If i download the application to communicate serial IO, it reflects back it's result to Target shell to my PC, that is host(ie showing the output which is to be sent to serial hardware, for example modem). I disconnected the host with target and connected target with modem...The same reflected signal should come to modem, right????(here i am getting the problem..the modem didn't(seems to be right) get any signal).. To achieve this, should i modify the serial IO for the second port?? I don't think to do so..because the single default serial should have supported this communication..am i right???? With Thanks and Regards, B.Manivannan. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: need help w/ epilogue 9.1 for tor2.0.2 Date: Mon, 10 Dec 2001 10:23:14 +0200 From: "Ofer Goren" Organization: Verio Message-ID: References: basicly, how to I tailor the envoy's snmp stack to my application, and how do I add me own mib files with the function and all. The documentation of envoy 9.1 refers only (and poorly) to the Tor1.0.1. The 9.2 documentations skiped tor2.0.2, and went right to tor AE. So, I didn't really understand how to have an application will all the things I need. It will be also nice if you can give a framework of how to write the community check functions etc... Thanks! - -- Ofer Goren Iamba Technologies "Mike Anderson" wrote in message news:EabQ7.8169$MR2.1778484@typhoon.southeast.rr.com... > VxWorks Greetings! > > "Ofer Goren" wrote in message > news:newscache$bzn7ng$b3l$1@lnews.actcom.co.il... > > Hi. > > I have lots of problem with this suite. I don't really wanna start > > describing them before I know there is someone out there that had worked > w/ > > it, and can truely help. Sorry. > > Yes, I've worked a lot with this code. My first suggestion would be to skip > the > Envoy 9.1 code and got to the Envoy 9.2 code. It was much more friendly to > Tor 2.02 than the 9.1 code. Building this was a bear. BTW, you didn't want > OSPF > to be included as well did you? If so, then you'll need to build the SNMP > in a > rather special way to support it. Give me a shout and I'll try to help, if > I can. > > Regards, > > Mike > > -- > -- > ======================================================= > T h e P T R G r o u p, I n c. > ======================================================= > Embedded, Real-Time Solutions and Training > > Mike Anderson mailto:mike@ThePTRGroup.com > Chief Scientist http://www.ThePTRGroup.com > C: (703) 585-9384 V: (703) 430-3748 > ======================================================= > "Software development is like making a baby... > You can't make a baby in one month by impregnating > nine women. Some things just take time." > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: t2 project facility: possible to avoid compiling files based on flags? Date: Mon, 10 Dec 2001 09:22:12 +0100 From: "Michael Lawnick" Organization: Buergernetz Dillingen Message-ID: <9v1rbl$lkl$1@snoopy.bndlg.de> References: Reply-To: "Michael Lawnick" Sender: mlawnick@pd9e65f05.dip.t-dialin.net Hi Denis, AFAIK including a file in the project will compile it. One way to achieve your goal could be to wrap your modules with an #ifdef __INC___ #endif and define the switches of the files to be included in your C/C++ compiler tab or in a central .h file. HTH - -- Mit freundlichen Grüßen, Michael Lawnick ============================================== SOFTEC GmbH Tel +49-731-96600-0 Promenade 17 Fax +49-731-96600-23 D-89073 Ulm Michael Lawnick Germany lawnick@softec.de ============================================== "Denis Perelyubskiy" schrieb im Newsbeitrag news:slrna12qeu.nit.denis_newsREMOVE-TO-REPLY@finch.lecs.cs.ucla.edu... > hello, > > does anyone know what would be the cleanest way to exclude > some files from being compiled, other then remove them from > the project? basically, i am looking for some facility along > the lines of the following piece of makefile: > > ifdef MYFEATURE > OBJS += myobj.o > endif > > and then you'd have rules to make .o's from .c's, or use > defaults, or whatnot. this is not important. > > so, how do i specify this 'ifdef MYFEATURE' thing? > > thanks, > > denis > > -- > if you want to reply by mail, please tweak my > address in a reasonable way. sorry for the > inconvenience. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Other then WDB,how can i use END network driver in polled mode? Date: Mon, 10 Dec 2001 09:28:03 +0100 From: "Michael Lawnick" Organization: Buergernetz Dillingen Message-ID: <9v1rmk$lms$1@snoopy.bndlg.de> References: <65b68b90.0112080426.24c67671@posting.google.com> Reply-To: "Michael Lawnick" Sender: mlawnick@pd9e65f05.dip.t-dialin.net Hi Bina, in my ethernet drivers code I found ioctl values EIOCPOLLSTART and EIOCPOLLSTOP that switch between interrupt- and poll-mode. This as a first entry you should now search how to invoke drivers ioctl function. HTH - -- Mit freundlichen Grüßen, Michael Lawnick ============================================== SOFTEC GmbH Tel +49-731-96600-0 Promenade 17 Fax +49-731-96600-23 D-89073 Ulm Michael Lawnick Germany lawnick@softec.de ============================================== "Bina Patel" schrieb im Newsbeitrag news:65b68b90.0112080426.24c67671@posting.google.com... > Hello! > I am working with END network driver.i want that my driver should work > in polled mode. For my quetion "How to invoke polled function in END > network driver?". > i got reply from group that WDB puts END driver in polled mode. > Now my quetion is"Other then WDB,how can i use END network driver in > polled mode?" > I shall be greatful for any link on same. > > Bina. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: `Higher reliability' DosFs. Date: Mon, 10 Dec 2001 09:31:39 +0100 From: "Michael Lawnick" Organization: Buergernetz Dillingen Message-ID: <9v1rud$lpn$1@snoopy.bndlg.de> References: Reply-To: "Michael Lawnick" Sender: mlawnick@pd9e65f05.dip.t-dialin.net Hi Bill, > Will this guarantee that I will always have a good DosFs file when > using the caching mechanism? No ! The problem is the caching. Only if you can guarantee that writing has completed successfully, your mechanism will work. - -- 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 ============================================== "Bill Pringlemeir" schrieb im Newsbeitrag news:ur8q5gsy8.fsf@yahoo.com... > > I am trying to make a system that `shadows' DosFs files. My current > model is simply to add a sequence number and CRC to each file. When a > file is being updated, it is written as "filename.a" or "filename.b". > The files are always closed upon an update. > > When reading files, the one with greatest sequence number and a > correct CRC is chosen. If both files have a good CRC, the lower > sequence number file is deleted. File access is already guarded by > semaphores. > > Will this guarantee that I will always have a good DosFs file when > using the caching mechanism? Is anyone kind enough to point be to > some simple database theory? Any source...? > > tia, > Bill Pringlemeir. > > -- > Well, I'm a classic ANAL RETENTIVE!! And I'm looking for a way to > VICARIOUSLY experience some reason to LIVE!! > > vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" --------------------------- Newsgroups: comp.os.vxworks Subject: IP_HDRINCL on VxWorks Date: 10 Dec 2001 01:43:15 -0800 From: kanandhi@yahoo.com (Anandhi) Organization: http://groups.google.com/ Message-ID: Hi, I have created a raw socket in vxworks (pcpentium based target system) and IP_HDRINCL option is set.I have framed a IP header with HDR_LEN encoded in "network byte-order", and if I try sending a packet out ; sento() fails with an errno 55 (No Buffers Available). When I framed IP header with HDR_LEN in "host byte-order" sendto() works fine and I am able to get the data successfully at the peer. I have tried out examples which encode HDR_LEN in "network byte-order" in PcPentium based machines running Linux and Solaris. I have also tried out in SPARC machines . Please can u give me some inputs. Regards Anandhi --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Mon Dec 10 05:42:13 2001 From: csaba.voros@essnet.se Date: Mon Dec 10 05:42:15 PST 2001 Subject: WindView for Tornado 1.0.1 does not work Hi People, If any of you have gained some experience in WindView for Tornado 1.0.1 have please a look at the problem I have to face now. SYSTEM CONFIGURATION: Host: Win NT 4.0 Target: I80386 BSP: PC386 After WindView gets the first dose of event data from the target and displays them, the message "Connection is not open" comes up. From this point on, no further data will be shown by WindView. Sometimes (very rarerly) this message does not appear but the effect is the same: no more data are displayed. The displayed status in both cases is COMPLETED. Moreover, the two toolbar buttons "Track incoming events" and "Stop event collection" are continously disabled, even if that aforementioned message does not come up. On the other hand, the View Graph for the first batch of event data does not either seem to correct. It seems that information for only the task tWdbTask is shown. Let's take first the "Context switch" mode. Sometimes I got a screen displaying a timeintervall of even more seconds, but no any track of task scheduling could be seen on the graph. We can say the whole graph was so "static", as if no any executing tasks would have been on the target. If I set the "Object Status" mode the situation was almost the same except that a lot of event legends (mostly for taskSafe) appeared along the line for tWdbTask. But for only that task. The rest of the graph remained "static" again. I tried to use the outer tool as well. The phenomenon was very similar: only one .wvr file was created, and data for only that task was displayed in which I switched event logging on/off (another task than tWdbTask). Some further info that can be useful in tracking this issue: 1) The PC we are working on includes 2 network cards. One to the company network and one towards the target. When I first started WindView it couldn't attach at all to the target remaining continuously in a WAITING state. After digging a bit into the TCL code, I found a call which returned the IP number for the network card other than the one used for the communication to the target. After replacing this call with the proper hardcoded IP number the connection was established and it started to work. But not really well (as you could read above). 2) I also tried not to use the timestamp driver. The effect was the same except that I had sequence numbers on the scale of the view graph instead of time values. 3) I also changed the size of the event buffer on the target from the deafult 20K to bigger/smaller. No positive result. 4) I tried to set tNetTask's priority higher from the Tornado Shell (from the default 50 to 0). It didn't help either (the task tEvtTask has 0 priority by default). Any idea/suggestion? Thanks Csaba From vxwexplo-errs@csg.lbl.gov Mon Dec 10 09:47:21 2001 From: "Clayton, Mark" Date: Mon Dec 10 09:47:24 PST 2001 Subject: RE: comp.os.vxworks newsdigest This is either a bug or a feature, it depends on your point of view. Which ever side of the fence you like, it is 'BAD' style. See the TCP/IP FAQ: http://www.whitefang.com/rin/rawfaq.html#14 Regards Mark > > Newsgroups: comp.os.vxworks > Subject: IP_HDRINCL on VxWorks > Date: 10 Dec 2001 01:43:15 -0800 > From: kanandhi@yahoo.com (Anandhi) > Organization: http://groups.google.com/ > Message-ID: > > Hi, > > I have created a raw socket in vxworks (pcpentium > based target system) and IP_HDRINCL option is set.I > have framed a IP header with HDR_LEN encoded in > "network byte-order", and if I try sending a packet > out ; sento() fails with an errno 55 (No Buffers > Available). When I framed IP header with > HDR_LEN in "host byte-order" sendto() works fine and I > am able to get the data successfully at the peer. > > I have tried out examples which encode HDR_LEN in > "network byte-order" in PcPentium based machines > running Linux and Solaris. I have also tried out in > SPARC machines . Please can u give me some inputs. > > Regards > Anandhi > From vxwexplo-errs@csg.lbl.gov Mon Dec 10 10:25:27 2001 From: Rainer Stelzer Date: Mon Dec 10 10:25:29 PST 2001 Subject: VMEbus Shared memory network with MVME230x Master Hi, I've some trouble running a VME multiprozessor system. We have a MVME 2304 in slot 1 acting as VME Systemcontroller and as the SM-Master. 9 VME Bus CPU (IP860) are attached as SM-Slaves over the VMEBus. They use the MVME board to router IP pakets to the ethernet by running IP-over Backplane. They boot over FTP. In general the system works fine, but the system start is quite instable. The MVME loads the bootimage from ftp and starts execution aprox. 7 seconds after VME reset. That's the moment when the IP860 boards begin to request the VMEBus to attach the shared memory loacation on the MVME board. The instable behaviour seems to be a problem with the VMEBus arbiter. I think it is caused by the reinitilazation of the tundra universe VME chipset taking place when starting the loaded bootimage at that point. here my questions: - I think my configuration is quite common (?) Has somebody out there similar problems with such kind of configuration ? Is it a solution to remove VME-support from the MVME bootloader image ? - It tried to trigger a SYSRESET by software from the MVME, to be sure that the Tundra chip is completly initilized when the slaves begin to request the bus, But when I do this the MVME resets, too. How can I avoid this ? - Is there an other explanation for the described behaviour ? Thanx in advance . rgds R.Stelzer ===================================================== 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 ===================================================== vxwork tornado From vxwexplo-errs@csg.lbl.gov Mon Dec 10 11:27:45 2001 From: Donald McLachlan Date: Mon Dec 10 11:27:48 PST 2001 Subject: Re: 3c905B-TX-NM Not having much luck getting this card working with Tor 2, pcPentium target, Solaris host. After getting the right device/driver name, I've got a floppy with a bootrom.sys which says elPci0 is an available boot device. I boot the vxWorks.st image from the floppy, and it still reports muxDevLoad failed for device entry 1! Failed to attach to device elPci0Network interface elPci0 unknown. ... Right now my config.h has: #deinfe INCLUDE_NETWORK #define INCLUDE_3C90X_END #define INCLUDE_FEI #define INCLUDE_BSD #define INCLUDE_END Do I need anything else? I have read one must disable Plug and Play in order to use this card. We have tried the CD that came with the card and cannot find how to disable PNP. We tried older CD's and floppys from 3Com, still no luck. Is it PNP for the card, or PNP for the bios we must disable? Thanks, Don P.S. First, I wish Wind River would publish a list of supported (in this case network) cards, and the proper "includes" to select the driver. This cost me serveral days. This is obviously a business decission of theirs but I don't get it. The only thing I can see is maybe they don't want a competitor to be able to say "We support X more network cards than WRS". But to me this is just a headache. WRS support said I must rebuild the boot rom from the command line in the target/pcPentium directory. That is fine. But why then does the workspace tool have Build->Build Boot ROM option???? False advertising? From vxwexplo-errs@csg.lbl.gov Mon Dec 10 11:57:53 2001 From: "Randy Ryan" Date: Mon Dec 10 11:57:56 PST 2001 Subject: Gigabit Ethernet Is anybody using Gigabit Ethernet yet? Am interested in what the realistic throughput rates are on particularly a PowerPC based SBC, support 64-bit @ 66MHz transfer rates in a VxWorks environment. Have not found good non-marketing benchmarks as of yet, though a recent issue of Embedded Magazine had a study published by SBS showing bout 36 MB/s peak rates which seems horribly low. Thanks, Randy Randy Ryan, Sr. Eng. NovaSol http://www.nova-sol.com From vxwexplo-errs@csg.lbl.gov Mon Dec 10 12:36:43 2001 From: "Randy Ryan" Date: Mon Dec 10 12:36:45 PST 2001 Subject: Gigabit Ethernet, Correction Ooops, the article I referenced was in Real Time Computing magazine, Sept. 2001, Fibre Channel vs. Ethernet: Which for Real-Time Embedded, SBS Technologies VxWorks Randy Ryan, Sr. Eng. NovaSol Ph/Fx: 818-785-3648 From vxwexplo-errs@csg.lbl.gov Mon Dec 10 13:48:40 2001 From: Shiv Date: Mon Dec 10 13:48:42 PST 2001 Subject: Porting VC++ WinSock code to Tornado For the purpose maintaining consistency , I've been asked to port pre-written MFC-based C++ functions to an mpc860 board running VxWorks , as opposed to writing my own client-server socket functions from scratch. A few zillion #ifdefs later , I realize that I'm not too familiar with MFC-dependencies and I keep seeing a lot of : make: *** No rule to make target `stdafx.h', needed by `test_socket.o'. Stop during the build. Does anybody know what could possibly be wrong ? Shiv From vxwexplo-errs@csg.lbl.gov Tue Dec 11 04:03:20 2001 From: Vxworks Exploder Date: Tue Dec 11 04:03:22 PST 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Tue Dec 11 04:03:15 PST 2001 Subject: About PCMCIA / PC Card Driver Subject: Re: symbol table Subject: Re: Making a .lib file Subject: Re: spurious interrupts! Subject: Re: VxWorks interview question ! Subject: Re: Send message to the target from Window's application Subject: test Subject: PCMCIA DRIVER for VxWorks Subject: Re: development board kit Subject: increasing system ticks per sec Subject: Re: VxWorks interview question ! Subject: Re: increasing system ticks per sec Subject: VESA BIOS using 486 Target Subject: Tornado 2.0.2 for PPC405 Problems with Ethernet drivers Subject: Re: vxWorks SNTP and Solaris NTP interoperability Subject: multiple processes on vxworks Subject: GCC 2.95.4 Subject: linking question Subject: Wind River PPC750 / MPC107 board PCI config help Subject: Re: IP_HDRINCL on VxWorks Subject: Re: VxSIM Networking Help Subject: Re: spurious interrupts! Subject: Re: (urgent) Other then WDB,how can i use END driver in polled mode? Subject: Re: Other then WDB,how can i use END network driver in polled mode? Subject: Need help - Unresolved symbol error while downloading Subject: SCSI test task hung in a 8260 board.Why? Subject: Re: Need help - Unresolved symbol error while downloading Subject: Re: BOOT_LINE_ADDR and downloadable project Subject: Re: Need help - Unresolved symbol error while downloading Subject: Re: POST / RAM Test Subject: Re: spurious interrupts! Subject: Re: multiple processes on vxworks Subject: Re: VxWorks interview question ! Subject: NET BSD for PPC860 Subject: Loading objects in VxWorks 5.4 ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: About PCMCIA / PC Card Driver Date: 10 Dec 2001 04:20:17 -0800 From: netproa@freechal.com (James) Organization: http://groups.google.com/ Message-ID: <3ad4e9be.0112100420.7ae0aab5@posting.google.com> Configuration CPU : MPC8240 OS : VxWorks PCMCIA Controller: PCI1410A (Texas Instrument) PC Card : ATA-Flash Hello. Now, I am handling vxWorks pcmcia driver. Since vxWorks provide only Intel 82365 pcmcia controller driver, I use PCI1410A's ExCA Compatibility Register for 82365 mode(memory-mapped). Card detect, Ready signals are good. And Read/Write of Status, Card-Status-Change, Power Control register are good. But CIS(Card Information Structure) of PC Card are not read. (only 0xff) I saw MPC860 BSP source code using pcmcia / pc card. But it's some different. MPC860 has own pcmcia controller and it is connected to system bus. And CIS_MEM_START, CIS_REG_START address value is defined statically. My system use PCI pcmcia controller, so I allocated some PCI memory for PC-Card's CIS. But I don't know how to map PCI memory to CIS. Please tell me about this problem. Thanks in advance. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: symbol table Date: 10 Dec 2001 12:44:21 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: <9uo7r4$2s5@nntpa.cb.lucent.com> <9urbab$pig@nntpb.cb.lucent.com> "Ren Hantao" wrote: > How to build the exrea step in the build process? > in the Makefile? Could you describe it roughly? Look in the Makefiles (target\h\make) where the symbol table file is made, and add your commands there. Also look in the Tornado output window, here you can see which commands are executed, and you can try to execute them from the command line to test these and add your commands manually first. 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: Making a .lib file Date: 10 Dec 2001 12:47:15 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: <705c7b34.0111210630.7e8b496a@posting.google.com> <3BFD1245.7F0D1A81@bellsouth.net> <9torbl$rs3$1@newstest.laserlink.net> <72bQ7.8122$MR2.1767637@typhoon.southeast.rr.com> "Mike Anderson" wrote: > You are correct! However, this option in the build tab is only visible > if you're working with a downloadable image rather than a bootable image. But why would you want to build a bootable image in a library? 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: spurious interrupts! Date: 10 Dec 2001 05:11:18 -0800 From: bondi_ameya@yahoo.com (Ameya) Organization: http://groups.google.com/ Message-ID: <57c7939b.0112100511.5eca9104@posting.google.com> References: <57c7939b.0112092118.81fddd8@posting.google.com> i am using a Linksys WLAN card with a PRISM chipset. The interrupts are "edge mode". My ISR does go into an infinite loop coz i need to ACK a particular bit of the firmwares register. since i dont get the interrupt..The acking is not done which causes the infinite loop. What i did was to hook an ISR for IRQ7 as well.. in my driver. Strangely i notice the 8259IntBoi() is still called first and then my ISR for IRQ7 is called.(??) are there any known issues with the way vxworks handles spurious interrupts?? - - ameya "Leonid Rosenboim" wrote in message news:... > It is hard to guess without any specifics on the chipset, but sounds > like a chipset glitch to me, and there eint much you can do bout it. > Here is how I would work around this problem - switch all interrupts > to "Level Mode" rather then "Edge Mode" trigerring, which should be pretty > easy > if all drivers you use are properly handling interrupts, i.e. Level > Trigerring > is veru unforegiving, if an ISR forgets to turn off the cause of the > interrupt, > the ISR will be re-entered infinitely and the system will seem to freeze. > On the positive side, in Level mode interrupts are NEVER missed. > > - Leonid > > "Ameya" wrote in message > news:57c7939b.0112092118.81fddd8@posting.google.com... > > hi, > > i have an END Driver in vxworks. It has an ISR hooked to IRQL 11. > > im using a pentium BSP on a x86 board. > > I seem to be missing interrupts at irregular time intervals. > > i can see the spurious interrupt coz i8259IntBoi() is called (irq num :7) > > I also see the EOI from the slave...but still i seem to miss my interrupt. > > any ideas? > > > > thanks > > ameya --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks interview question ! Date: 10 Dec 2001 06:02:05 -0800 From: manibala@covansys.com (B.Manivannan) Organization: http://groups.google.com/ Message-ID: <23de5426.0112100602.4bf0c4ac@posting.google.com> References: If more guys would provide the questions over here, it would be beneficial to all of our friends to upgrade their knowledge...Here we have discussed only one question... umeshs@myw.ltindia.com (Umesh Satyanarayana) wrote in message news:... > semaphores are available whilst you write drivers on WinNT or UNIX. > Nothing special about the semaphore in Vxworks (except PRIORITY > INVERSION SAFE argument). Candidate should have said "Real time kernel > with rich API's" > what does semaphore got to do with faster and flexibility? > Umesh --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Send message to the target from Window's application Date: Mon, 10 Dec 2001 09:24:32 -0500 From: "Yin, Yong" Organization: Nortel Message-ID: <9v2gi4$hgt$1@bcarh8ab.ca.nortel.com> References: <4367c1ab.0112092238.7630e8bd@posting.google.com> create a socket - that's all Regards Yong Yaron wrote in message news:4367c1ab.0112092238.7630e8bd@posting.google.com... > How can I send/post a message (or even use a pipe etc') from Windows > (Visual C++ / MFC) application to VxWorks application ?. --------------------------- Newsgroups: comp.os.vxworks Subject: test Date: Mon, 10 Dec 2001 16:22:38 +0200 From: "Rani" Organization: Motorola Message-ID: <9v2ggv$s3$1@newshost.mot.com> test --------------------------- Newsgroups: comp.os.vxworks Subject: PCMCIA DRIVER for VxWorks Date: Mon, 10 Dec 2001 16:37:00 +0200 From: "Rani" Organization: Motorola Message-ID: <9v2j9p$2kn$1@newshost.mot.com> We are interested in providing solution for PCMCIA CARD using VxWorks on PPC860. Our main focus is on Radio modems (circuit/packet data, AT command driven) using public infrastructure such as TETRA, GPRS, iDEN, CDPD. (We are not interested in wireless LAN or 802.11 standard at least for now). We are looking for general solution, not tailored for specific card/manufacturer. Did anybody have SW experience with relevant knowledge? What drivers/stack are available? Thanks, Rani. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: development board kit Date: Mon, 10 Dec 2001 04:03:09 -0500 From: George Varndell Organization: Posted Via Binaries.net = SPEED+RETENTION+COMPLETION = http://www.binaries.net Message-ID: <3C147A4D.98E68CD5@varndellengineering.com> References: <9kMQ7.1580$eF1.175242@news20.bellglobal.com> Jacob wrote: > > Hello, > > I am looking for development board kit that that has > the basics for real-time programming: > -microcontroller, ram, flash, serial and/or ethernet connection > I should be able to download and run a real time operating system..(e.g psos > or vxworks). > This is for educational purposes only so it would have to be inexpensive. > Can anyone tell where I can get more informarion or a place that sells these > types of kits ? Not sure what you mean by 'kit'. Do you want to build a board? If you don't want to build one, go to... http://www.embeddedplanet.com They have a variety of reasonably priced boards, sporting 8xx processors with all the other stuff you mentioned. As far as freely downloading pSOS or VxWorks goes, you're out of luck, but there are free RTOS. Regards, George Varndell - -- Varndell Engineering, LLC. PPC and XScale Expertise Coming soon! Download free demo versions of our END drivers! http://www.varndellengineering.com ______________________________________________________________________________ Posted Via Binaries.net = SPEED+RETENTION+COMPLETION = http://www.binaries.net --------------------------- Newsgroups: comp.os.vxworks Subject: increasing system ticks per sec Date: Mon, 10 Dec 2001 14:10:32 -0500 From: Gilad Katz Organization: Mailgate.ORG Server - http://www.Mailgate.ORG Message-ID: <3C1508A8.6303AF46@avoidspam.com> Sender: bpatel@fwion.com Hi, If I increase the systems ticks from the default (60) to 500 using sysClkRateSet(), what are the implications on system performance? This is in a commercial product wher you have routing and signalling software tasks running and having critcial timing requirements. Is it better to yse auxClk instead? Gilad --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks interview question ! Date: Mon, 10 Dec 2001 12:56:24 -0800 From: "Gary M" Message-ID: <3c152217$0$2729$724ebb72@reader2.ash.ops.us.uu.net> References: <23de5426.0112100602.4bf0c4ac@posting.google.com> Reply-To: "Gary M" 1. If the candidate has any RTOS experience, I would expect them to solve a straightforward problem (in five minutes or less, including pseudo-code), such as an ISR and a task synchronizing access to shared memory, or a simple producer-consumer relationship with acknowledgement. If that's too soft a test, you can always discuss possible solutions to the "Dining Philosophers" problem ("n" tasks competing for simultaneous access to a pair of "n" resources) in real-time systems. 2. VxWorks driver writers should be able to convincingly describe driver guidelines and discuss drivers they have worked on. 3. All candidates (VxWorks or not) should be able to describe in detail a complex project (or two) they have worked on in the past three years, including system architecture, design issues, the areas they worked on, etc. Given the choice of two candidates, the first a VxWorks veteran who is lacking in complex design experience or has poor communication skills, versus a second candidate weak in VxWorks but has outstanding RTOS experience and complex design backgrounds, I will always lean towards the second. "B.Manivannan" wrote in message news:23de5426.0112100602.4bf0c4ac@posting.google.com... > If more guys would provide the questions over here, it would be > beneficial to all of our friends to upgrade their knowledge...Here we > have discussed only one question... > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: increasing system ticks per sec Date: Mon, 10 Dec 2001 13:32:59 -0800 From: "Gary M" Message-ID: <3c152aab$0$2727$724ebb72@reader2.ash.ops.us.uu.net> References: <3C1508A8.6303AF46@avoidspam.com> Reply-To: "Gary M" Changing the ticks from 60 to 500 causes the VxWorks tick handler to run every 2 milliseconds instead of 16.667 milliseconds, so obviously there is an increase in system overhead. As to your application, if you have tasks which are tied to using system services utilizing the system clock (delays and timeouts) that require this higher degree of clock resolution, then running the system clock at the higher rate is probably the recommended approach. Alternatively, if you can structure your application so that time-critical activities can be signalled from the aux clock ISR, that would minimize the increased overhead. I don't recommend trying to reimplement what the VxWorks kernel/scheduler already does, however. "Gilad Katz" wrote in message news:3C1508A8.6303AF46@avoidspam.com... > Hi, > > If I increase the systems ticks from the default (60) to 500 using > sysClkRateSet(), > what are the implications on system performance? This is in a commercial > > product wher you have routing and signalling software tasks running and > having > critcial timing requirements. > Is it better to yse auxClk instead? > > Gilad > --------------------------- Newsgroups: comp.os.vxworks Subject: VESA BIOS using 486 Target Date: 10 Dec 2001 13:41:08 -0800 From: chernack@juno.com (Charles R. Chernack) Organization: http://groups.google.com/ Message-ID: I have been able to switch to 640 * 480 256 color by prototyping my software on my 486 PC Compatible target using DOS in real mode. My target has a 640 * 480 touch screen, and I wish to do some simple graphics. I have been able to write a simple plot pixel routine under DOS which goes to the correct 64K video page using the VESA BIOS page switch. And I have been able to set the cursor position and to put A/N text out using INT 10H in DOS. I'd like to have the same capability in VxWorks (which of course runs in protected mode). I do not need the full capabilities of WINDML -- I just need to set the screen to 640 * 480 256 color and be able to do the things listed above. Has anyone got this working? Charles R. Chernack Los Altos, CA --------------------------- Newsgroups: comp.os.vxworks Subject: Tornado 2.0.2 for PPC405 Problems with Ethernet drivers Date: Mon, 10 Dec 2001 22:14:56 GMT From: "Ho-Kuo Chan" Organization: GlobeTrotter Message-ID: Hi, Has anyone tried the new Walnut BSP in Tornado 2.0.2? I have installed it and I am having two problems: 1. The ibmEmac device name has changed and so it requires a new bootrom. I have compiled it but I have discovered a problem: if the interface receives a packet while being reset, an Interrupt Exception occurs. The simplest way to reproduce the problem is to ping the interface continously and then reset the board through a console using Ctrl-X. 2. The lnPci driver no longer works. When the mux tries to start the device, it fails with a "muxDevStart failed for device entry 1" (entry 1 being the lnPci device). This worked fine with Tornado 2. Any help is appreciated. - -- Ho-Kuo Chan Software Designer, Digital Division Wavesat Telecom (514) 956-6322 --------------------------- Newsgroups: comp.os.vxworks Subject: Re: vxWorks SNTP and Solaris NTP interoperability Date: Tue, 11 Dec 2001 08:42:30 +1030 From: Tim Shaw Organization: Defence Science and Technology Organisation Message-ID: <3C15334E.FE3D0E46@dsto.defence.gov.au> References: <9uqnhe$9rn$1@bcarh8ab.ca.nortel.com> Yes, we set our clocks on MVME2604s using an SNTP client to a Solaris NTP server. "J. Min Kim" wrote: > Do the vxWorks SNTP client/server work with Solaris NTP server/client? > According to the RFC2030 (SNTP v.4), "To a NTP or SNTP server, NTP and SNTP > clients are undistinguishable; to a NTP or SNTP client, NTP and SNTP > servers are undistinguishable". It seems they have the interoperability. Has > anyone tried it? --------------------------- Newsgroups: comp.os.vxworks Subject: multiple processes on vxworks Date: 10 Dec 2001 14:38:29 -0800 From: sriva_t@yahoo.com (Srivathsa) Organization: http://groups.google.com/ Message-ID: <1cdb3add.0112101438.5b3eb3ba@posting.google.com> Hi, Here's a newcomer's question on vxworks. (The manual seemed to say NO to my question, but I want to confirm with the experts). I have an embedded system running VxWorks, and a networking application. I have written a new module that should go into the application. I would like to test my module independently first, without *REALLY PLUGGING IN* my module into the application. Can I create a new dummy application that uses my module, and load that dummy application on VxWorks without touching the current application? PARAPHRASED: Does vxworks allow multiple processes on the target system? (I saw this feature on Sun's ChorusOS, in a certain mode called ExtendedMode) - -Srivathsa --------------------------- Newsgroups: comp.os.vxworks Subject: GCC 2.95.4 Date: 10 Dec 2001 15:03:51 -0800 From: dward@appiancom.com (David Ward) Organization: http://groups.google.com/ Message-ID: <679c25ff.0112101503.165a710d@posting.google.com> I built a GCC 2.95.4 cross tool chain for PPC under Cygwin, following Dave K's web site for the most part. Very nice - informative and colorful ! I installed it in a Tornado II tree. But I have some lib issues when linking my target - these will no doubt look familiar to some folks - undefined ref to __start_cp_handler, __eh_malloc, etc... GCC's new mechanism for c++ exception handling. There were lot's of postings about this back in Feb. The trail goes cold before I find the best way to resolve it. __start_cp_handler is defined in my new libgcc - /Tornado/host/x86-win32/lib/gcc-lib/powerpc-wrs-vxworks/2.95.4/libgcc.a I can get it to link in this by removing -nostdlib or adding -lgcc, but that lib appears to be in conflict with the old VxWorks run-time lib - /Tornado/target/lib/libPPCEC603gnuvx.a which I also need. Do I rebuild this ? The source appears to be here... or can I just yank out some objects ? Dave Ward BTW, here's what my link looks like before playing around with - -nostdlib, -lgcc - Building vxWorks ccppc -nostdlib -r -Wl,-X -o /dmwroot/sw/build/images/switch_ram_debug/vxWorks.tmp \ /dmwroot/sw/bsp/switch_8260/.switch_ram_debug/sysALib.o /dmwroot/sw/bsp/switch_8260/.switch_ram_debug/sysLib.o \ /dmwroot/sw/build/images/switch_ram_debug/usrConfig.o \ /dmwroot/sw/build/images/switch_ram_debug/version.o \ \ -Wl,--start-group /dmwroot/sw/bsp/switch_8260/.switch_ram_debug/switch_ram_debug.a ...big group here... /dmwroot/sw/vendors/wrs/misc/.switch_ram_debug/switch_ram_debug.a - -Wl,--end-group /Tornado/target/lib/libPPCEC603gnuvx.a nmppc /dmwroot/sw/build/images/switch_ram_debug/vxWorks.tmp | wtxtcl /Tornado/host/src/hutils/munch.tcl -asm ppc > /dmwroot/sw/build/images/switch_ram_debug/ctdt.c ccppc -B/Tornado/host/x86-win32/lib/gcc-lib/powerpc-wrs-vxworks/2.95.4/ - -mstrict-align -fpermissive -ansi -Wall -I. - -I/Tornado/target/config/all -I/Tornado/target/h - -I/Tornado/target/src/config -I/Tornado/target/src/drv -DCPU=PPCEC603 -g -mstrict-align -msoft-float -fdollars-in-identifiers -c /dmwroot/sw/build/images/switch_ram_debug/ctdt.c \ -o /dmwroot/sw/build/images/switch_ram_debug/ctdt.o ldppc -X -N -Map /dmwroot/sw/build/images/switch_ram_debug/switch_ram_debug.map - -e _sysInit -Ttext 00100000 \ /dmwroot/sw/build/images/switch_ram_debug/dataSegPad.o \ /dmwroot/sw/build/images/switch_ram_debug/vxWorks.tmp \ /dmwroot/sw/build/images/switch_ram_debug/ctdt.o \ /dmwroot/sw/build/images/switch_ram_debug/ac_end_text.o \ -o /dmwroot/sw/build/images/switch_ram_debug/vxWorks /dmwroot/sw/build/images/switch_ram_debug/vxWorks.tmp: In function `EAggConfigGroup': /Tornado/target/h/stl_list.h:159: undefined reference to `__start_cp_handler' /dmwroot/sw/build/images/switch_ram_debug/vxWorks.tmp: In function `EAggGroupNew': /Tornado/target/h/stl_list.h:159: undefined reference to `__start_cp_handler' /dmwroot/sw/build/images/switch_ram_debug/vxWorks.tmp: In function `iterator_traits<__list_iterator >::iterator_category iterator_category<__list_iterator Organization: WebUseNet Corp. - "ReInventing The UseNet" Message-ID: <3C153C06.811549CC@novtek.com> Hello, Does anyone know how to specify fixed addresses for variables during the link process? Basically I would like to make sure certain variables end up in the fast internal ram of the cpu, while letting the target based loader figure out the placement for the rest of them. -defsym option for ld allows to add a variable with a fixed address, but not specify the location of an existing variable. Thanks for your help, Yura --------------------------- Newsgroups: comp.os.vxworks Subject: Wind River PPC750 / MPC107 board PCI config help Date: Mon, 10 Dec 2001 19:42:22 -0700 From: "Michael S. Simpson" Organization: Raytheon Company Message-ID: <3C15728E.D3D29007@west.raytheon.com> I will be the first to admit that I know very little about PCI configuration. We're working on a project where we are using a Wind River build PMC board with a Motorola MPC750 CPU and a MPC107 Bridge/Memory Controller chip. This will be connected to a custom built carrier board with several ASICs and memory for video processing. There are no other PCI devices, just the carrier board. We're trying to configure the PMC board as the PCI host. Our biggest problem is trying to configure the memory map for the PCI space. Nobody on the development team has any experience with the MPC107 and it's a pain to figure out (Wind River has been some help, but hard to reach people who really know what's going on with the BSP). Motorola is usless for getting information from. If anyone has a little time to maybe answer some questions, please respond to this message. Thanks... Mike Simpson Software Engineer Raytheon Missile Systems Tucson, AZ (520) 784-5381 --------------------------- Newsgroups: comp.os.vxworks Subject: Re: IP_HDRINCL on VxWorks Date: 10 Dec 2001 19:47:58 -0800 From: shada@netplane.com (shad ansari) Organization: http://groups.google.com/ Message-ID: <15adeaa9.0112101947.73f47ca6@posting.google.com> References: See Raw IP Networking FAQ (http://www.whitefang.com/rin/rawfaq.html#14): "Systems derived from 4.4BSD have a bug in which the ip_len and ip_off members of the ip header have to be set in host byte order rather than network byte order. Some systems may have fixed this. I've confirmed this bug has been fixed on OpenBSD 2.1." kanandhi@yahoo.com (Anandhi) wrote in message news:... > Hi, > > I have created a raw socket in vxworks (pcpentium > based target system) and IP_HDRINCL option is set.I > have framed a IP header with HDR_LEN encoded in > "network byte-order", and if I try sending a packet > out ; sento() fails with an errno 55 (No Buffers > Available). When I framed IP header with > HDR_LEN in "host byte-order" sendto() works fine and I > am able to get the data successfully at the peer. > > I have tried out examples which encode HDR_LEN in > "network byte-order" in PcPentium based machines > running Linux and Solaris. I have also tried out in > SPARC machines . Please can u give me some inputs. > > Regards > Anandhi --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxSIM Networking Help Date: 10 Dec 2001 21:05:17 -0800 From: raj_rashid@hotmail.com (Rashid) Organization: http://groups.google.com/ Message-ID: <2a6118da.0112102105.62f7b6c@posting.google.com> References: <15a87607.0112082237.32ab8591@posting.google.com> Hi Ramesh, Do routeAdd("0.0.0.0","90.0.0.1") and everythging will work fine. Regards, Rashid rashanmu@npd.hcltech.com (Ramesh) wrote in message news:<15a87607.0112082237.32ab8591@posting.google.com>... > Hi, > > I found a message by Luc in this group some time before and i did the > steps as he mentioned in detail. I have mentioned that mail in the > bottom of this post. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: spurious interrupts! Date: 10 Dec 2001 21:26:57 -0800 From: bondi_ameya@yahoo.com (Ameya) Organization: http://groups.google.com/ Message-ID: <57c7939b.0112102126.574c01ad@posting.google.com> References: <57c7939b.0112092118.81fddd8@posting.google.com> <57c7939b.0112100511.5eca9104@posting.google.com> hi, The interrupts generated are LEVEL MODE and not edge mode as mentioned earlier. - - ameya bondi_ameya@yahoo.com (Ameya) wrote in message news:<57c7939b.0112100511.5eca9104@posting.google.com>... > i am using a Linksys WLAN card with a PRISM chipset. > The interrupts are "edge mode". My ISR does go into an infinite loop > coz i need to ACK a particular bit of the firmwares register. > since i dont get the interrupt..The acking is not done which causes > the infinite loop. What i did was to hook an ISR for IRQ7 as well.. in > my driver. Strangely i notice the 8259IntBoi() is still called first > and then my ISR for IRQ7 is called.(??) > are there any known issues with the way vxworks handles spurious > interrupts?? > > - ameya > > > > > > "Leonid Rosenboim" wrote in message news:... > > It is hard to guess without any specifics on the chipset, but sounds > > like a chipset glitch to me, and there eint much you can do bout it. > > Here is how I would work around this problem - switch all interrupts > > to "Level Mode" rather then "Edge Mode" trigerring, which should be pretty > > easy > > if all drivers you use are properly handling interrupts, i.e. Level > > Trigerring > > is veru unforegiving, if an ISR forgets to turn off the cause of the > > interrupt, > > the ISR will be re-entered infinitely and the system will seem to freeze. > > On the positive side, in Level mode interrupts are NEVER missed. > > > > - Leonid > > > > "Ameya" wrote in message > > news:57c7939b.0112092118.81fddd8@posting.google.com... > > > hi, > > > i have an END Driver in vxworks. It has an ISR hooked to IRQL 11. > > > im using a pentium BSP on a x86 board. > > > I seem to be missing interrupts at irregular time intervals. > > > i can see the spurious interrupt coz i8259IntBoi() is called (irq num :7) > > > I also see the EOI from the slave...but still i seem to miss my interrupt. > > > any ideas? > > > > > > thanks > > > ameya --------------------------- Newsgroups: comp.os.vxworks Subject: Re: (urgent) Other then WDB,how can i use END driver in polled mode? Date: 10 Dec 2001 21:50:19 -0800 From: binagpatel@yahoo.com (Bina Patel) Organization: http://groups.google.com/ Message-ID: <65b68b90.0112102150.38415971@posting.google.com> References: <65b68b90.0112092029.1cc2bf50@posting.google.com> "Leonid Rosenboim" wrote in message news:... Hello! Thanks for reply, as it is very helpfull to me. Now i have clear idea about ENd driver in polled mode.Now i can efficiently utilize my ENd network driver. Once again thanks for reply thanks.. Bina > You see Bina, > > The answer to this seems incredibly simple at first, there is an API > function which switches an END driver (mich like an SIO driver) > between INT and POLL modes, so to call this API and make the switch > is easy enough. Or it would be quite easy indeed should this be about an SIO > driver ... > > But END driver is a network device, and aside from WDB, it is not used > directly > but rather via the TCP/IP and Sockets layers, and the standard SENS stack > does not know HOW to use an END driver in POLLED mode, and it never could. > Full fledged implementation of TCP/IP would be probably uanble to use a > network > driver in polled mode. > > WDB over END is a light-weight implementation of UDP/IP, specifically > designed > to be able to work in POLLEd mode, hence it can. > > So unless you are going to call the END driver API DIRECTLY to send and > receive > MAC packets, the information you requested is quite useless for you. > > Hope this helps, and let me know if you need my mailing address where you > can mail your gratitude. > - Leonid > > > "Bina Patel" wrote in message > news:65b68b90.0112092029.1cc2bf50@posting.google.com... > > Hello! > > > > I am working with END network driver.i want that my driver should work > > in polled mode. For my quetion "How to invoke polled function in END > > network driver?". > > i got reply from group that WDB puts END driver in polled mode. > > Now my quetion is"Other then WDB,how can i use END network driver in > > polled mode?" > > I shall be greatful for any link on same as soon as possible. > > > > Bina. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Other then WDB,how can i use END network driver in polled mode? Date: 10 Dec 2001 21:58:05 -0800 From: binagpatel@yahoo.com (Bina Patel) Organization: http://groups.google.com/ Message-ID: <65b68b90.0112102158.26feb754@posting.google.com> References: <65b68b90.0112080426.24c67671@posting.google.com> <9v1rmk$lms$1@snoopy.bndlg.de> Thanks for the reply. Bina "Michael Lawnick" wrote in message news:<9v1rmk$lms$1@snoopy.bndlg.de>... > Hi Bina, > > in my ethernet drivers code I found ioctl values EIOCPOLLSTART and > EIOCPOLLSTOP that switch between interrupt- and poll-mode. This as a first > entry you should now search how to invoke drivers ioctl function. > 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 > ============================================== > > "Bina Patel" schrieb im Newsbeitrag > news:65b68b90.0112080426.24c67671@posting.google.com... > > Hello! > > I am working with END network driver.i want that my driver should work > > in polled mode. For my quetion "How to invoke polled function in END > > network driver?". > > i got reply from group that WDB puts END driver in polled mode. > > Now my quetion is"Other then WDB,how can i use END network driver in > > polled mode?" > > I shall be greatful for any link on same. > > > > Bina. --------------------------- Newsgroups: comp.os.vxworks Subject: Need help - Unresolved symbol error while downloading Date: Tue, 11 Dec 2001 06:29:37 GMT From: "Junghyun Cho" Organization: Dacom Message-ID: Hi, there. I made simple "downloadable application module" and target was "VxSim". Of course, I succeeded to compile and build it. But I couldn't download it to the target with the following error messages : Unresolved symbols encountered while downloading: Errors while downloading /home2/tomato/user/jhcho/projects/timer/SIMSPARCSOLARISgnu/timer.out: timer_create timer_connect timer_settime Here is the code. #include "vxWorks.h" #include "semLib.h" #include "timers.h" #include "logLib.h" SEM_ID semId; int tmohandler() { semGive(semId); } STATUS job_periodic() { while (1) { semTake(semId, WAIT_FOREVER); printf("\njob_periodic : Timer expired."); } } STATUS timedemo(void) { int ret; timer_t timerId; struct itimerspec timeToSet; timeToSet.it_interval.tv_sec = 1; timeToSet.it_interval.tv_nsec = 0; timeToSet.it_value.tv_sec = 1; timeToSet.it_value.tv_nsec = 0; ret = timer_create(CLOCK_REALTIME, NULL, &timerId); if (ret == ERROR) { printf("\ntimer_create error !"); } ret = timer_connect(timerId, (VOIDFUNCPTR)tmohandler, 0); if (ret == ERROR) { printf("\ntimer_connect error !"); } ret = timer_settime(timerId, CLOCK_REALTIME, &timeToSet, NULL); if (ret == ERROR) { printf("\ntimer_settime error !"); } } int start() { semId = semBCreate(SEM_Q_FIFO, SEM_EMPTY); if (semId == NULL) { printf("\ntimer_settime error !"); } #define PRI 110 taskSpawn("timedemo", PRI, 0, 4096, timedemo, 0); taskSpawn("job_per", PRI, 0, 4096, job_periodic, 0); } Does anyone give me the solution to download correctly ? --------------------------- Newsgroups: comp.os.vxworks Subject: SCSI test task hung in a 8260 board.Why? Date: 10 Dec 2001 22:31:17 -0800 From: errant@263.net (Jerry Wong) Organization: http://groups.google.com/ Message-ID: Hi, I wrote a scsi test task which reads a fix amout(16,64,128 sectors etc.) sectors per time from scsi disk from the head to tail to test the BSP and boards. But the scsi task would hang at reading after several thousends times.The scsi task was pended at the time, while SCSI Interrupt Status Registers(SIST0,SIST1) show no error occur. I trace the scsi task into the driver and found each time the abnormal situation occurs in ISR which in our driver(ncr810lib.o) is ncr810Intr().The ISR invokes a function named ncr810EventTypeGet() which reads the ISTAT register to check if DIP or SIP was set and do something accordingly.Sometime,ncr810EventTypeGet(),strangely,read both the two bit are zero.The ncr810EventTypeGet() returned error, then did the ncr810Intr().Does this mean there is a scsi interupt occurring,but was not assert by scripts or scsi controller?I don't know what cause this and if it is the resean that scsi task hangs.The trace output is below: ... interrupt: ncr810EventTypeGet:Interrupt status istat = 0x01, SCSI status0 = 0x00 , SCSI status1 = 0x00DMA status = 0x84 interrupt: ncr810EventTypeGet:Interrupt status istat = 0x01, SCSI status0 = 0x00 , SCSI status1 = 0x00DMA status = 0x84 interrupt: ncr810EventTypeGet:Interrupt status istat = 0x01, SCSI status0 = 0x00 , SCSI status1 = 0x00DMA status = 0x84 interrupt: ncr810EventTypeGet:Interrupt status istat = 0x01, SCSI status0 = 0x00 /* above is normal */ , SCSI status1 = 0x00DMA status = 0x84 interrupt: ncr810EventTypeGet:Interrupt status istat = 0x89, SCSI status0 = 0x00 /* abnormal occur */ , SCSI status1 = 0x00DMA status = 0x90 interrupt: ncr810EventTypeGet: script aborted. interrupt: ncr810Intr: DSA = 0xb3df96a8, DSP = 0xb01972f8, type = 1. interrupt: ncr810Intr: state 2 -> 0 0x3df8488 (tScsiTask): ncr810Resume: thread: 0x03df6bd8: state 0 -> 2 interrupt: ncr810EventTypeGet:Interrupt status istat = 0x09, SCSI status0 = 0x00 , SCSI status1 = 0x00DMA status = 0x84 interrupt: ncr810EventTypeGet: script completed. interrupt: ncr810Intr: DSA = 0xb3df96a8, DSP = 0xb01973b8, type = 11. interrupt: ncr810Intr: state 2 -> 0 0x3df8488 (tScsiTask): ncr810Resume: thread: 0x03df6bd8: state 0 -> 2 interrupt: ncr810EventTypeGet:Interrupt status istat = 0x01, SCSI status0 = 0x00 , SCSI status1 = 0x00DMA status = 0x84 interrupt: ncr810EventTypeGet: script completed. interrupt: ncr810Intr: DSA = 0xb3df96a8, DSP = 0xb01973e0, type = 21. interrupt: ncr810Intr: state 2 -> 1 interrupt: ncr810EventTypeGet:Interrupt status istat = 0x09, SCSI status0 = 0x00 , SCSI status1 = 0x00DMA status = 0x84 interrupt: ncr810EventTypeGet: script completed. interrupt: ncr810Intr: DSA = 0xb3df965c, DSP = 0xb0197118, type = 22. interrupt: ncr810Intr: state 1 -> 0 0x3df8488 (tScsiTask): ncr810Resume: thread: 0x03df86c8: state 0 -> 2 interrupt: ncr810EventTypeGet:Interrupt status istat = 0x09, SCSI status0 = 0x00 , SCSI status1 = 0x00DMA status = 0x84 interrupt: ncr810EventTypeGet: script completed. interrupt: ncr810Intr: DSA = 0xb3df965c, DSP = 0xb0197420, type = 11. interrupt: ncr810Intr: state 2 -> 0 0x3df8488 (tScsiTask): ncr810Resume: thread: 0x03df6bd8: state 0 -> 2 /* the last trace output */ When the scsi task hung,print the task info: - -> i NAME ENTRY TID PRI STATUS PC SP ERRNO DELAY - ---------- ------------ -------- --- ---------- -------- -------- - ------- ----- tExcTask excTask 3dfeac8 0 PEND 18a738 3dfe9e8 0 0 tLogTask logTask 3dfc140 0 PEND 18a738 3dfc068 0 0 tWdbTask 0x14a690 3d97ca0 3 READY 154ff0 3d97b28 0 0 tScsiTask scsiMgr 3df8488 5 PEND 153570 3df83c8 0 0 tNetTask netTask 3db9498 50 READY 155490 3db93c8 0 0 scsitask testscsi 3d938f8 80 PEND 18a738 3d8b690 0 0 value = 0 = 0x0 - -> tt scsitask 1726b8 vxTaskEntry +5c : testsio ([]) 105364 testscsi +98 : scsiRdSecs (3df6060) 15da40 scsiRdSecs +6c : 18360c ([]) 18367c scsiDirectLibTblInit+950: 183708 ([]) 183870 scsiDirectLibTblInit+b44: scsiTransact (3df7408) 15d680 scsiTransact +5c : 162758 (3df7408) 1627c0 scsiThreadInit +470: 162558 (3df7408) 162620 scsiThreadInit +2d0: 16244c (3df71d8) 162498 scsiThreadInit +148: scsiMgrRequestExecute ([]) 185fdc scsiMgrRequestExecute+b0 : msgQReceive ([]) 152318 msgQReceive +280: qJobGet ([]) value = 0 = 0x0 my boards arch: CPU: XPC8260 SCSI: 53C895A OS; vxWorks5.4 driver:ncr810lib.o --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Need help - Unresolved symbol error while downloading Date: Tue, 11 Dec 2001 09:05:41 +0200 From: "Ofer Goren" Organization: Verio Message-ID: References: I think the problem is that your vxSim.exe does not include the timerLib library. You should add the component to the vxSim. - -- Ofer Goren Iamba Technologies "Junghyun Cho" wrote in message news:lJhR7.3204$I3.27470@news.bora.net... > Hi, there. > > I made simple "downloadable application module" and target was "VxSim". > Of course, I succeeded to compile and build it. But I couldn't download it > to the target with the following error messages : > > Unresolved symbols encountered while downloading: > Errors while downloading > /home2/tomato/user/jhcho/projects/timer/SIMSPARCSOLARISgnu/timer.out: > timer_create > timer_connect > timer_settime > > > Here is the code. > > #include "vxWorks.h" > #include "semLib.h" > #include "timers.h" > #include "logLib.h" > > SEM_ID semId; > > int tmohandler() > { > semGive(semId); > } > > STATUS job_periodic() > { > while (1) { > semTake(semId, WAIT_FOREVER); > > printf("\njob_periodic : Timer expired."); > } > } > > STATUS timedemo(void) > { > int ret; > timer_t timerId; > struct itimerspec timeToSet; > > timeToSet.it_interval.tv_sec = 1; > timeToSet.it_interval.tv_nsec = 0; > timeToSet.it_value.tv_sec = 1; > timeToSet.it_value.tv_nsec = 0; > > ret = timer_create(CLOCK_REALTIME, NULL, &timerId); > if (ret == ERROR) { > printf("\ntimer_create error !"); > } > > ret = timer_connect(timerId, (VOIDFUNCPTR)tmohandler, 0); > if (ret == ERROR) { > printf("\ntimer_connect error !"); > } > > ret = timer_settime(timerId, CLOCK_REALTIME, &timeToSet, NULL); > if (ret == ERROR) { > printf("\ntimer_settime error !"); > } > } > > int start() > { > semId = semBCreate(SEM_Q_FIFO, SEM_EMPTY); > if (semId == NULL) { > printf("\ntimer_settime error !"); > } > > #define PRI 110 > taskSpawn("timedemo", PRI, 0, 4096, timedemo, 0); > taskSpawn("job_per", PRI, 0, 4096, job_periodic, 0); > } > > > Does anyone give me the solution to download correctly ? > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: BOOT_LINE_ADDR and downloadable project Date: Tue, 11 Dec 2001 10:01:08 +0200 From: "Ofer Goren" Organization: Verio Message-ID: References: Hi. Just to let you know, I found the answer: When I'm building a bootable project, a compilation flag is added: -DPRJ_BUILD. This flag is not added to a downloadable project. So, I can just ask for that flag, and according to the result, decide which file to include. However, if I include "config.h", the problem is already solved, since config.h asks for that flag, and according to the result includes (or not) projParams.h which overrides the config.h params (bottom of the file)..... :) Vwalla! - -- Ofer Goren Iamba Technologies "Ofer Goren" wrote in message news:newscache$f20sng$5i7$1@lnews.actcom.co.il... > hi > I want to use BOOT_LINE_ADDR in my file. I have no problem if the file is > part of bootable project, since it is defined in prjParams.h. However, if > I'm adding the file to a d/l-able project, there is no prjParams.h file, so > I have to include config.h. > So, what do I need to do so that a d/l-able project will include config.h, > and a bootable project will include prjParams.h (which can be different from > config.h). Is there a define I can use that define something different in > case of a bootable/dl-able project? > > Thank, > > -- > Ofer Goren > Iamba Technologies > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Need help - Unresolved symbol error while downloading Date: Tue, 11 Dec 2001 08:40:05 GMT From: "Junghyun Cho" Organization: Dacom Message-ID: References: "Ofer Goren" wrote in message news:newscache$7z46og$tt5$1@lnews.actcom.co.il... > I think the problem is that your vxSim.exe does not include the timerLib > library. You should add the component to the vxSim. Isn't it enough to inlcude 'timers.h' file in my code ? Please give me detail information. Thanks in advance. > > -- > Ofer Goren > Iamba Technologies > "Junghyun Cho" wrote in message > news:lJhR7.3204$I3.27470@news.bora.net... > > Hi, there. > > > > I made simple "downloadable application module" and target was "VxSim". > > Of course, I succeeded to compile and build it. But I couldn't download it > > to the target with the following error messages : > > > > Unresolved symbols encountered while downloading: > > Errors while downloading > > /home2/tomato/user/jhcho/projects/timer/SIMSPARCSOLARISgnu/timer.out: > > timer_create > > timer_connect > > timer_settime > > > > > > Here is the code. > > > > #include "vxWorks.h" > > #include "semLib.h" > > #include "timers.h" > > #include "logLib.h" > > > > SEM_ID semId; > > > > int tmohandler() > > { > > semGive(semId); > > } > > > > STATUS job_periodic() > > { > > while (1) { > > semTake(semId, WAIT_FOREVER); > > > > printf("\njob_periodic : Timer expired."); > > } > > } > > > > STATUS timedemo(void) > > { > > int ret; > > timer_t timerId; > > struct itimerspec timeToSet; > > > > timeToSet.it_interval.tv_sec = 1; > > timeToSet.it_interval.tv_nsec = 0; > > timeToSet.it_value.tv_sec = 1; > > timeToSet.it_value.tv_nsec = 0; > > > > ret = timer_create(CLOCK_REALTIME, NULL, &timerId); > > if (ret == ERROR) { > > printf("\ntimer_create error !"); > > } > > > > ret = timer_connect(timerId, (VOIDFUNCPTR)tmohandler, 0); > > if (ret == ERROR) { > > printf("\ntimer_connect error !"); > > } > > > > ret = timer_settime(timerId, CLOCK_REALTIME, &timeToSet, > NULL); > > if (ret == ERROR) { > > printf("\ntimer_settime error !"); > > } > > } > > > > int start() > > { > > semId = semBCreate(SEM_Q_FIFO, SEM_EMPTY); > > if (semId == NULL) { > > printf("\ntimer_settime error !"); > > } > > > > #define PRI 110 > > taskSpawn("timedemo", PRI, 0, 4096, timedemo, 0); > > taskSpawn("job_per", PRI, 0, 4096, job_periodic, 0); > > } > > > > > > Does anyone give me the solution to download correctly ? > > > > > > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: POST / RAM Test Date: 11 Dec 2001 01:15:17 -0800 From: vincent_hue@yahoo.com (Vincent) Organization: http://groups.google.com/ Message-ID: <67bf868f.0112110115.5ab66f87@posting.google.com> References: Davis, Find some information below. Vincent davis_spam@yahoo.com (Dford) wrote in message news:... > Hi, > > I'm trying to write a simple DRAM Pattern Test. I'm using PowerPC > 8240 (603E core), Mousse Vooha board. 64MB DRAM. > > LOCAL_MEM_LOCAL_ADRS 0x0 > LOCAL_MEM_SIZE 0x04000000 > USER_RESREVED_MEM 0x0 > > I am working with a normal compressed bootrom (bootrom.bin) My > questions are as follows: > > I need to run the RAM test before the bootrom copies itself into RAM, > obviously. I am not a strong PowerPC Assembly programmer, so I'm > trying to cheat by doing it in C. My entry point for the RAM test is > the romStart() routine in bootInit.c. After romInit.s masks > interrups, sets stack pointer to STACK_ADRS, and readies DRAM, it > jumps to romStart(). In romStart, before the code copies ROM to RAM, > I begin a RAM Pattern test. > > Correct me if I am wrong on these points: > > At this point the system is still executing instructions out of ROM. > Am I correct in assuming the data section is also in ROM at this > point? Yes, it should be executing from ROM. For my BSP (mv5100), there is no use of the data section up to that point. According to romStart.c, only ROM_RESIDENT images need to worry about the data section. Both ROM_COPY and ROM_COMPRESS only 'read' the data section (could then be in ROM). > I have not compiled the ROM_RESIDENT bootrom (well, I tried > but it was too large to fit in EEPROM). As I understand it, the > ROM_RESIDENT bootrom only copies the data section to RAM (and .bss > exists in RAM as well), and leaves the .text section in ROM. My entry > point is before I do the copy from ROM to RAM, so I am assuming the > .data section also exists in ROM at this point. The stack however, > must exist in RAM somewhere. I am somewhat confused where it exists > from the code / headers, though. > > I see the following defined in configAll.h: > > #if ((CPU_FAMILY == MIPS) || (CPU_FAMILY == PPC) > #define STACK_RESIDENT RAM_DST_ADRS > > -- following the chain of events: RAM_DST_ADRS -> RAM_HIGH_ADRS -> > 0x00c00000 > > Then --- > > #if (_STACK_DIR == _STACK_GROWS_DOWN) > **SNIPPED STUFF ABOUT ROM_RESIDENT** > #define STACK_ADRS _romInit > > -- _romInit -> is at 0x10100 for RAM, so bootrom copies to > RAM_HIGH_ADRS, but when the system cold boots out of ROM, it must > execute _romInit from EEPROM which is in the 0xff000100 (cold entry) = > ROM_TEXT_ADRS > > #else /* STACK GROWS UP */ > #define STACK_ADRS (_romInit-STACK_SAVE) > > -- where STACK_SAVE corresponds to 64 bytes, so this is, I am assuming > the size of my STACK. Therefore, I need to avoid the addresses from > 0x00c00000 to 0x00c10000 if stack grows down, and 0x00bf0000 to > 0x00c00000 if stack grows up. > > So, where is _STACK_DIR defined? vxArch.h For PPC, it definitively grows down. > Also, could someone clear up the following defines: > > ROM_LOW_ADRS = 00010000 # RAM text/data address for VxWorks > ROM_HIGH_ADRS = 00c00000 # RAM text/data address for bootrom That would be RAM_LOW and RAM_HIGH. gee, you're confused. Look at the romStart.c file, it gives information about these addreses. > I see from the reference manual on bootInit that RAM_LOW_ADRS is where > exc vectors, bp anchor, exc msg, and bootline are stored, and > RAM_HIGH_ADRS is where (romInit + ROM_COPY_SIZE) or binArrayStart is > stored. So, I understand the system works in the following manner: > > _romInit executes direct assembly code, sets STACK_ADRS to > RAM_HIGH_ADRS, inits processor, sets up DRAM, etc., jumps to C routine > romStart. When do exc msg, bootline, etc. get setup? These are 'locations', there is no need to 'setup' a location. Exceptions vector are set in excInit() (see reference manual), No bootline is involved in 'ROM' image. > This seems to > happen in the romStart routine, but my pattern test runs as the first > piece of code in romStart (before ANY copy happens from ROM to RAM), > and when I hit address 0x4400, the system halts. This address is > defined below in configAll.h: > > #if (CPU_FAMILY == PPC) > #define RESREVED 0x4400 /* avoid zeroing out EXC_MSG */ > > This I can see if I warm reboot, and dump memory --> my pattern is > written into RAM up to the address 0x4400 at which point it continues > as all 0's. I'd like to be able to just jump over this section of RAM > with my test, but I am unsure how large the reserved area is defined > as. Actually, it would be ok, to jump over the entire exc msg, exc > vectors, bp anchor, and bootline, if I could determine how large this > section is. This would just be RAM_LOW_ADRS plus some offset. Can > someone tell me how/where it defines how large this section is? Please look at the bootClear() routine in romStart.c It clears the RAM starting a x4400 = SYS_MEM_BOTTOM = RESERVED. You should not write your pattern below RESERVED. It is defined as : "Area at the bottom of memory to not clear". I guess you're writting '0xXX' to memory and then bootClear() write 0x00, which explain the memory dump. Now, why your test fail is a different issue. Not enough information is provided. Are you sure you don't clear the stack ? (or your data) ? > Finally, I would like to be able to flush / invalidate the dcache > during the pattern fill / verify tests. I tried issuing the system > calls cacheFlush() / cacheInvalidate() from the romStart() routine, > but the linker complains: > > ldppc -X -N -Map mousse.map -e _romInit -Ttext 00010000 \ > -o bootrom romInit.o bootInit.o version.o \ > D:\tornado\target/lib/libPPC603gnutms.a > D:\tornado\target/lib/libPPC603gnuospfv2.a > D:\tornado\target/lib/libPPC603gnuenvo > y.a D:\tornado\target/lib/libPPC603gnuzlib.a > D:\tornado\target/lib/libPPC603gnutms.a > D:\tornado\target/lib/libdiag.a D:\torna > do\target/lib/libtest.a D:\tornado\target/lib/libsal.a > D:\tornado\target/lib/libbcm.a D:\tornado\target/lib/libsoc.a D:\torna > do\target/lib/libsal.a D:\tornado\target/lib/libPPC603gnuvx.a > bootrom.Z.o > D:\tornado\target/lib/libPPC603gnuvx.a(workQLib.o): In function > `workQPanic': > workQLib.o(.text+0x316): undefined reference to `sysExcMsg' > workQLib.o(.text+0x31e): undefined reference to `sysExcMsg' > workQLib.o(.text+0x32a): undefined reference to `sysExcMsg' > workQLib.o(.text+0x346): undefined reference to `sysExcMsg' > workQLib.o(.text+0x34e): undefined reference to `sysExcMsg' > D:\tornado\target/lib/libPPC603gnuvx.a(workQLib.o)(.text+0x35a): more > undefined references to `sysExcMsg' follow > make: *** [bootrom] Error 0x1 > > Is it possible to somehow resolve this, or is it necessary to write a > funciton in PPC assembly to invalidate / flush the cache? sysExc is defined in sysLib.c As you may see, the sysLib.o file is not part of the linked files. Your code in ROM cannot call BSP function, or VxWorks fuction that are linked to BSP code. > Thanks in advance for any help, > > Dford --------------------------- Newsgroups: comp.os.vxworks Subject: Re: spurious interrupts! Date: Tue, 11 Dec 2001 12:01:30 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <57c7939b.0112092118.81fddd8@posting.google.com> <57c7939b.0112100511.5eca9104@posting.google.com> You got me confused: 1- I have seen many times EDGE interrupts lost due to various timing circumstances, but never seen a LEVEl interrupt get lost, nor do I think it could get lost ever cause its a condition not an event. 2- Where is this infinit eloop happens ? Again its self-contradictory - if the infinite loop is in the ISR, then the interrupt was not lost because if it had you would not be in the ISR in the first place. 3- It is a widely known issue that many driver programmers omit to handle ALL POSSIBLE device condition and properly ACK these conditions to avoid infinite ISRs. THis is probably why the PC industry prefers EDGE interrupts because in this mode the system is more foregiving to bad driver (which are the norm in the PC industry). - - Leonid "Ameya" wrote in message news:57c7939b.0112100511.5eca9104@posting.google.com... > i am using a Linksys WLAN card with a PRISM chipset. > The interrupts are "edge mode". My ISR does go into an infinite loop > coz i need to ACK a particular bit of the firmwares register. > since i dont get the interrupt..The acking is not done which causes > the infinite loop. What i did was to hook an ISR for IRQ7 as well.. in > my driver. Strangely i notice the 8259IntBoi() is still called first > and then my ISR for IRQ7 is called.(??) > are there any known issues with the way vxworks handles spurious > interrupts?? > > - ameya > > > > > > "Leonid Rosenboim" wrote in message news:... > > It is hard to guess without any specifics on the chipset, but sounds > > like a chipset glitch to me, and there eint much you can do bout it. > > Here is how I would work around this problem - switch all interrupts > > to "Level Mode" rather then "Edge Mode" trigerring, which should be pretty > > easy > > if all drivers you use are properly handling interrupts, i.e. Level > > Trigerring > > is veru unforegiving, if an ISR forgets to turn off the cause of the > > interrupt, > > the ISR will be re-entered infinitely and the system will seem to freeze. > > On the positive side, in Level mode interrupts are NEVER missed. > > > > - Leonid > > > > "Ameya" wrote in message > > news:57c7939b.0112092118.81fddd8@posting.google.com... > > > hi, > > > i have an END Driver in vxworks. It has an ISR hooked to IRQL 11. > > > im using a pentium BSP on a x86 board. > > > I seem to be missing interrupts at irregular time intervals. > > > i can see the spurious interrupt coz i8259IntBoi() is called (irq num :7) > > > I also see the EOI from the slave...but still i seem to miss my interrupt. > > > any ideas? > > > > > > thanks > > > ameya --------------------------- Newsgroups: comp.os.vxworks Subject: Re: multiple processes on vxworks Date: Tue, 11 Dec 2001 12:09:23 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <1cdb3add.0112101438.5b3eb3ba@posting.google.com> In VxWorks, there is no "process", still you can dynamically load your new module into memory and test it. With this you avoid LINKING the new module with the rest of the code, but all cod ebeing run in a single FLAT system, they could interfere with eachother while competing for resources etc. Now there is a new product, called VxWorksAE (or as I know it Protection Domains") which do create separate memory domains for different application parts to further minimize interference and tightly control cooperation of different modules. This is probably similar to ChorusOS capabilities, but I would say that as long as two modules are running on the same CPU, one could NEVER completely eliminate interference between the odules, not even in a Unix environment. - - Leonid "Srivathsa" wrote in message news:1cdb3add.0112101438.5b3eb3ba@posting.google.com... > Hi, > Here's a newcomer's question on vxworks. (The manual seemed to say > NO to my question, but I want to confirm with the experts). > > I have an embedded system running VxWorks, and a networking > application. I have written a new module that should go into the > application. > I would like to test my module independently first, without *REALLY > PLUGGING IN* my module into the application. > Can I create a new dummy application that uses my module, and load > that dummy application on VxWorks without touching the current > application? > > PARAPHRASED: > Does vxworks allow multiple processes on the target system? (I saw > this feature on Sun's ChorusOS, in a certain mode called ExtendedMode) > > -Srivathsa --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks interview question ! Date: 11 Dec 2001 10:24:31 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: <23de5426.0112100602.4bf0c4ac@posting.google.com> <3c152217$0$2729$724ebb72@reader2.ash.ops.us.uu.net> If you are looking for a test for embedded programmers, take a look at this: http://www.embedded.com/2000/0005/0005feat2.htm Groeten, Johan - -- o o o o o o o . . . _____________________________ o _____ || Johan Borkhuis | .][__n_n_|DD[ ====_____ | borkhuis@agere.com | >(________|__|_[_________]_|__________________________| _/oo OOOOO oo` ooo ooo 'o!o!o o!o!o` === VxWorks FAQ: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html === --------------------------- Newsgroups: comp.os.vxworks Subject: NET BSD for PPC860 Date: Tue, 11 Dec 2001 12:10:00 +0100 From: "Hans-Joachim Zimmer" Message-ID: <9v4q2d$d2qjc$1@ID-102608.news.dfncis.de> Has anybody any experiences using NetBSD on a PPC860 or PPC8260 ? If any support in NetBSD is available for these processors ?? What should be payed attention for using NetBSD on these processors ? Thanks in advance for any statement. Achim Zimmer --------------------------- Newsgroups: comp.os.vxworks Subject: Loading objects in VxWorks 5.4 Date: Tue, 11 Dec 2001 12:19:17 +0100 From: "Bostjan Mulej" Organization: Slovenija OnLine - SiOL Message-ID: <9pmR7.765$r03.27049@news.siol.net> Hi Our application will be divided into several modules. Only one question still remains. Does target-loader starts any sub-routine in downloaded object? It's possible manually start a sub-routine from shell, but is there any way to do it automatically (for example, target-loader starts a predefined sub-routine)? Regards, Bostjan --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Tue Dec 11 09:06:33 2001 From: Rainer Stelzer Date: Tue Dec 11 09:06:35 PST 2001 Subject: VMEbus Shared memory network with MVME230x Master Hi, I've some trouble running a VME multiprozessor system. We have a MVME 2304 in slot 1 acting as VME Systemcontroller and as the SM-Master. 9 VME Bus CPU (IP860) are attached as SM-Slaves over the VMEBus. They use the MVME board to router IP pakets to the ethernet by running IP-over Backplane. They boot over FTP. In general the system works fine, but the system start is quite instable. The MVME loads the bootimage from ftp and starts execution aprox. 7 seconds after VME reset. That's the moment when the IP860 boards begin to request the VMEBus to attach the shared memory loacation on the MVME board. The instable behaviour seems to be a problem with the VMEBus arbiter. I think it is caused by the reinitilazation of the tundra universe VME chipset taking place when starting the loaded bootimage at that point. here my questions: - I think my configuration is quite common (?) Has somebody out there similar problems with such kind of configuration ? Is it a solution to remove VME-support from the MVME bootloader image ? - It tried to trigger a SYSRESET by software from the MVME, to be sure that the Tundra chip is completly initilized when the slaves begin to request the bus, But when I do this the MVME resets, too. How can I avoid this ? - Is there an other explanation for the described behaviour ? Thanx in advance . rgds R.Stelzer ===================================================== 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 ===================================================== vxworks tornado From vxwexplo-errs@csg.lbl.gov Tue Dec 11 10:22:23 2001 From: Donald.McLachlan@crc.ca Date: Tue Dec 11 10:22:25 PST 2001 Subject: Re: comp.os.vxworks newsdigest I want to get the individual vxworks messages, but not the daily digests. I've tried the auto-list configure URL; I still get the daily digests. I've tried e-mailing the list manager; I still get the daily digests. How can I get the individual vxworks messages, but not the daily digests? Don From vxwexplo-errs@csg.lbl.gov Wed Dec 12 04:03:17 2001 From: Vxworks Exploder Date: Wed Dec 12 04:03:19 PST 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Wed Dec 12 04:03:13 PST 2001 Subject: SDS with BDM connection Subject: Re: OSE Delta Performance Stats Subject: Re: Loading objects in VxWorks 5.4 Subject: Warning: Core file checksums do not match.!!!!!!!!1 Subject: Re: `Higher reliability' DosFs. Subject: Re: Where have 'Specific Trap MIB Lookup Results' gone ? Subject: UDP-IP performances and timing with VxWorks Subject: Re: How about this: Can the MPC860 do this ? Subject: Re: UDP-IP performances and timing with VxWorks Subject: Task status DEAD Subject: Re: VxWorks interview question ! Subject: Re: Need help - Unresolved symbol error while downloading Subject: Re: UDP-IP performances and timing with VxWorks Subject: File I/O on host from embedded PPC target Subject: Re: getting arrow keys Subject: Re: MPC8260 and SPI problem Subject: Re: Send message to the target from Window's application Subject: Re: linking question Subject: Re: MPC8260 and SPI problem Subject: Re: PCMCIA DRIVER for VxWorks Subject: Re: Task status DEAD Subject: Re: File I/O on host from embedded PPC target Subject: initialize multiple ethernet interface when boot vxworks? Subject: how to add hdlc to vxworks! Subject: Binding a socket to a interface in VxWorks Subject: Re: Need help - Unresolved symbol error while downloading Subject: Re: Abstruct class (C++) with VxWorks Subject: problem using sntpc Subject: tNetTask stack overrun Subject: Re: initialize multiple ethernet interface when boot vxworks? Subject: Please Help Me ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: SDS with BDM connection Date: 11 Dec 2001 13:30:27 CET From: "Roger Bodén" Message-ID: <3c15fc63@news.wineasy.se> Hello, I'm trying to debug vxworks with the SDS debugger with a BDM connection on a FADS860T board (using the ads860 BSP). I can get the vxworks application to run, but I get a software emulation exception quite frequently. I always get it when semTake is called, and the time it takes from system start before the exception is encountered varies somewhat. Does anyone have a working SDS configuration file for my configuration? Please send it to me, I would be much obliged. Any other suggestions are gratefully accepted as well. Roger Bodén INTERPEAK Box 7498 SE-103 92 Stockholm Sweden Office: +46 8 545 275 80 Fax +46 8 545 275 99 www.interpeak.com --------------------------- Newsgroups: comp.realtime,comp.arch.embedded,comp.os.vxworks,comp.os.psos,comp.os.qnx Subject: Re: OSE Delta Performance Stats Date: Tue, 11 Dec 2001 12:31:39 GMT From: Zoltan Kis Organization: Nokia Message-ID: <3C15FD19.50A86007@trshp.ntc.nokia.com> References: Reply-To: zkis@mail.com Interrupt -> interrupt handler ~ 0.8-1 usec Sending message from interrupt handler to prioritised process + context switch ~ 3.8-4 usec. Context switch ~ 1.9-2 usec. Hope this helps. Zoltan Jim Montville wrote: > > Ooops...sorry, I dumbed out. I have a 266 Mhz PPC405. > > Jim > > "Mark A. Odell" wrote in message > news:Xns916D9B69D283Elkj562ghjgk1k245lbvj@130.133.1.4... > > "Jim Montville" wrote in > > news:sQ9P7.1783$NB2.42444@news2.nokia.com: > > > > > Does anyone here have any performance data on OSE/Delta? Specifically, > > > I'm looking for interrupt latency and task switching time. Thanks in > > > advance. > > > > I think it would help if we knew the CPU and CPU speed, don't you? > > > > -- > > - Mark A. Odell > > - Embedded Firmware Design, Inc. > > - http://www.embeddedfw.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Loading objects in VxWorks 5.4 Date: Tue, 11 Dec 2001 13:36:26 +0100 From: "Michael Lawnick" Organization: Buergernetz Dillingen Message-ID: <9v4uke$ukv$1@snoopy.bndlg.de> References: <9pmR7.765$r03.27049@news.siol.net> Reply-To: "Michael Lawnick" Sender: mlawnick@pd9524560.dip.t-dialin.net Hi Bostjan, without changing loader code or implementing an own version of loader: If you set C++ Constructor to automatic in VxWorks (standard), you might wrap your application by a static 'startup' class. The creator of the class may spawn the main task of your application. Haven't done myself yet, but C++ collegues of mine said it should work. 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 ============================================== "Bostjan Mulej" schrieb im Newsbeitrag news:9pmR7.765$r03.27049@news.siol.net... > Hi > > Our application will be divided into several modules. Only one question > still remains. > > Does target-loader starts any sub-routine in downloaded object? It's > possible manually start > a sub-routine from shell, but is there any way to do it automatically (for > example, target-loader > starts a predefined sub-routine)? > > Regards, Bostjan > > --------------------------- Newsgroups: comp.os.vxworks Subject: Warning: Core file checksums do not match.!!!!!!!!1 Date: 11 Dec 2001 04:45:39 -0800 From: youerming@163.com (Torna) Organization: http://groups.google.com/ Message-ID: Attaching C++ interface... succeeded. Attaching coff OMF reader for ARM CPU family... succeeded. Warning: Target checksum: 0x961d (computed from 0x1000 to 0xb184c). Host checksum: 0x270b (computed from 0xf40104 to 0xff0950). Warning: Core file checksums do not match. Warning: Symbol .stabstr from section of type STYP_REG. Ignored. Warning: Symbol .stabstr from section of type STYP_REG. Ignored. Warning: Symbol .stabstr from section of type STYP_REG. Ignored. Warning: Symbol .stabstr from section of type STYP_REG. Ignored. Warning: Symbol .stabstr from section of type STYP_REG. Ignored. Warning: Symbol .stabstr from section of type STYP_REG. Ignored. Warning: Symbol .stabstr from section of type STYP_REG. Ignored. i have copied my core file of vxWorks image to c:\vxworks and configuered my target server as the following: core file :c:\vxWorks symbol table : global symbol or full symbol. but i still get the error message above! ! could anybody give me a solution of it for ever? ! could andbody tell me how to make the file vxworks.st or vxworks.sym? !if i want to get vxworks.sym ,which command ? thanks a lot! shine --------------------------- Newsgroups: comp.os.vxworks Subject: Re: `Higher reliability' DosFs. Date: Tue, 11 Dec 2001 15:42:58 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: For the benefit of the newsgroup here is my recommended solution: DosFs 2.0 was designed with reliability in mind - to replace some data safely, you need to write new data in a new file, in the same directory as the old file. Then close or flush the new file, which wil flush related disk cache blocks, including data and FAT, and finally rename() the new file to be named like the old one. rename() is specifically implemented such that as long as its within the same directory, either the old file will survive or the new one if the system crashes during the update at any point of time. There will never be both remaining or none. Note that rename() does not trigger any FAT changes if it happens within the same directory, only the directory is being modified. On top of this, if the file names are short to fit 8.3 names, and the directory is just a few entires, then the entire directory is only one disk sector, which makes the rename operation really atomic. Of course any reliability term has to be put in perspective, i.e. what is tha probability of failure, Here however operation sequence is made such that if the failure is instantaneous - i.e. t\all operations requested prior to time T have been executed, all operations after time T will fail. This siequence is consistently maintained down to the block driver level, but if below the above instantanous failure is not preserved, this indeed could go wrong. Hope this helps, - - Leonid "Bill Pringlemeir" wrote in message news:ur8q5gsy8.fsf@yahoo.com... > > I am trying to make a system that `shadows' DosFs files. My current > model is simply to add a sequence number and CRC to each file. When a > file is being updated, it is written as "filename.a" or "filename.b". > The files are always closed upon an update. > > When reading files, the one with greatest sequence number and a > correct CRC is chosen. If both files have a good CRC, the lower > sequence number file is deleted. File access is already guarded by > semaphores. > > Will this guarantee that I will always have a good DosFs file when > using the caching mechanism? Is anyone kind enough to point be to > some simple database theory? Any source...? > > tia, > Bill Pringlemeir. > > -- > Well, I'm a classic ANAL RETENTIVE!! And I'm looking for a way to > VICARIOUSLY experience some reason to LIVE!! > > vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" --------------------------- Newsgroups: comp.os.psos,comp.protocols.snmp,comp.os.vxworks Subject: Re: Where have 'Specific Trap MIB Lookup Results' gone ? Date: Tue, 11 Dec 2001 13:46:39 +0000 (UTC) From: "Hans-peter Kabout" Organization: Mailgate.ORG Server - http://www.Mailgate.ORG Message-ID: References: <3C0F7FD7.F9BA9321@bigfoot.com> Hi Alex, please compare `Trap-PDU ::=' .. from RFC1157 with `SNMPv2-Trap-PDU ::=' .. and `PDU ::=' from RFC1448. MG-SOFT MIB Browser is retrieving all information from traps. The 'Specific Trap MIB Lookup Results' you are missing are not contained in the SNMPv2-Trap-PDU. For that reason you should use the 1st varbinds for the missing information. Regards, H.-P. Kabout, IT-Consulting e-mail: Kai.Kabout@freenet.de Phone: +49 (0)9085 - 1017 Mobile: +49 (0)160 - 7556913 http://www.kabout.com (under construction) "Alex Vinokur" wrote in message news:3C0F7FD7.F9BA9321@bigfoot.com... ... > > What may cause the different representation of these traps? > > > > -- > =========================== > Alex Vinokur > mailto:alexvn@bigfoot.com > mailto:alexvn@dr.com > http://up.to/alexvn > http://go.to/alexv_math > =========================== - -- Posted from nbg2-t5-2.mcbone.net [62.104.219.89] via Mailgate.ORG Server - http://www.Mailgate.ORG --------------------------- Newsgroups: comp.os.vxworks Subject: UDP-IP performances and timing with VxWorks Date: Tue, 11 Dec 2001 14:35:08 +0100 From: christophe leroy Organization: CS S I Message-ID: <3C160B8C.482E2ABD@c-s.fr> Hello, I'm looking for informations about the performances of the TCP-IP stack in VxWorks. I'm in particular interested in the cpu load generated by reception or transmittion of an UDP packet. Thanks C.Leroy --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How about this: Can the MPC860 do this ? Date: Tue, 11 Dec 2001 16:17:48 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: Besides sympaties, I can only say it is a dumb idea to hook this crazy device straight to the bus. Instead, I would suggest to use an intermediate device which is smart and comaptible to the processor speeds and makes everything nice and dandy for the crazy guy. For example look at a QuickLink FGPA with standard PCI interface, you wiuld make a large FIFO and Bus Master DMA with the FPGA so the CPU would not have to worry every cycle. "Jim" wrote in message news:A4NP7.2662$8e.208941@news... > Here is a new description of the problem ... now that my undertanding of it > is getting better ... as before all answers advice and sympathy are welcome > ;-) > > The "crazy slow" device we are talking to cannot tolerate rates on any lines > higher than 1 Mhz ... so that means we must slow down the bus to this speed > ... > > That means of course that GPCM and UPM changes like adding wait states will > not make a difference ... > > So what we have done is slow the bus down to .625 Mhz ... we did this using > the "dfnh" and "ebdf" to divide or 80 Mhz clock by 64 and then 2 respectiely > to produce the bus clock rate. > > We can now communicate with the crazy device ...and we have managed to get > our serial port to function at a low baud rate... but however ethernet will > not work ... any advice on this ? We really would like ethernet to work ;-) > > Thanks again everyone, > Jim > > > "Jim" wrote in message news:yvrP7.2379$8e.174760@news... > > Hello all, > > > > I have a question for anyone who dares answer ;-) > > > > Currently I have a board running vxWorks with an MPC860 at 80Mhz and a bus > > frequency of 40 Mhz. > > > > I have a new device that I need to communicate with on a chip select ... > the > > device requires ... a "crazy slow" bus speed of less than 1 Mhz ... > > otherwise it will just plain not work ... > > > > So far I (and others) have managed to get the bus speed to around 1 Mhz > > using the "ebdf" and "dfnh" (these are in chapter 15 of the 860 users > manual > > if you interested) ... when we lower the bus to this speed however > ethernet > > and the serial ports behave "poorly" ... maybe I missed something when I > > tried this. > > > > Currently we are consider only dropping the bus rates when we do an access > > to this slow device ... to do this we would wrap writes to this address > > range in a function call ... then in the function shut off interupts, drop > > the clocks, and then do the memory write/read, then speed the clocks back > > up, then start interupts back up ... does any one know if this will work? > > > > Any help ... or even sympathy would be greatly appreciated ... > > > > Jim > > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: UDP-IP performances and timing with VxWorks Date: Tue, 11 Dec 2001 15:34:09 +0100 From: "Michael Lawnick" Organization: Buergernetz Dillingen Message-ID: <9v55h1$3ap$1@snoopy.bndlg.de> References: <3C160B8C.482E2ABD@c-s.fr> Reply-To: "Michael Lawnick" Sender: mlawnick@pd9524560.dip.t-dialin.net Hi C., this strongly depends on your CPU + NIC ! As for all other newbies : "Need more input !" "christophe leroy" schrieb im Newsbeitrag news:3C160B8C.482E2ABD@c-s.fr... > Hello, > > I'm looking for informations about the performances of the TCP-IP stack > in VxWorks. > I'm in particular interested in the cpu load generated by reception or > transmittion of an UDP packet. > > Thanks > > C.Leroy > --------------------------- Newsgroups: comp.os.vxworks Subject: Task status DEAD Date: Tue, 11 Dec 2001 16:20:34 +0100 From: "Johan Bostedt" Organization: Mailgate.ORG Server - http://www.Mailgate.ORG Message-ID: <9v5880$tce$1@newsreader.mailgate.org> Sender: johan.bostedt@racomna.se Has anyone seen task status DEAD for a VxWorks task? According to the Reference Manual the function taskStatusString() can indicate task status DEAD. A task that has terminated normally will not exist and the taskStatusString call will fail. But is there a case where an existing task can have status DEAD? Regards, Johan Bostedt - ---------------------------------------------------------------------- Johan Bostedt mailto:johan@racomna.se Phone: +46(0)18- 12 40 70 Racomna AB http://www.racomna.se Direct: +46(0)18- 12 40 95 Salagatan 16 A Mobile: +46(0)70-657 64 70 753 30 Uppsala Fax: +46(0)18- 12 40 74 SWEDEN --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks interview question ! Date: Tue, 11 Dec 2001 15:38:17 GMT From: "Jim" Organization: Magma Communications Ltd. Message-ID: References: <23de5426.0112100602.4bf0c4ac@posting.google.com> <3c152217$0$2729$724ebb72@reader2.ash.ops.us.uu.net> Maybe I am wrong ... but so far I have really seen anything "Real Time" about any of these questions ... vxWorks is first and formost a real time OS ... if people are asking about semaphores etc ... then you are not being interviewed for a job that requires RTOS experience ... anybody with windows programming or even java experience should be able to solve concurency problems ... I think OS and Concurency questions should be clearly separated from Real Time questions which sit on top of the OS ... My $0.02 ... Jim "Gary M" wrote in message news:3c152217$0$2729$724ebb72@reader2.ash.ops.us.uu.net... > 1. If the candidate has any RTOS experience, I would expect them to solve a > straightforward problem (in five minutes or less, including pseudo-code), > such as an ISR and a task synchronizing access to shared memory, or a simple > producer-consumer relationship with acknowledgement. If that's too soft a > test, you can always discuss possible solutions to the "Dining Philosophers" > problem ("n" tasks competing for simultaneous access to a pair of "n" > resources) in real-time systems. > 2. VxWorks driver writers should be able to convincingly describe driver > guidelines and discuss drivers they have worked on. > 3. All candidates (VxWorks or not) should be able to describe in detail a > complex project (or two) they have worked on in the past three years, > including system architecture, design issues, the areas they worked on, etc. > > Given the choice of two candidates, the first a VxWorks veteran who is > lacking in complex design experience or has poor communication skills, > versus a second candidate weak in VxWorks but has outstanding RTOS > experience and complex design backgrounds, I will always lean towards the > second. > > "B.Manivannan" wrote in message > news:23de5426.0112100602.4bf0c4ac@posting.google.com... > > If more guys would provide the questions over here, it would be > > beneficial to all of our friends to upgrade their knowledge...Here we > > have discussed only one question... > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Need help - Unresolved symbol error while downloading Date: Tue, 11 Dec 2001 17:43:17 +0200 From: "Ofer Goren" Organization: Verio Message-ID: References: Nop. this is not enough. The include "whatever.h" only solves compilation warnings. However, you have to link your application with the actual "code", in that case, an archive file. So, by adding the relevant flag in the project facilty, you tell it to link the needed archive file to your application. It's the same as if you write a function foo() in a file a.c, declere it in a.h, and then call it from b.c. It is not enough to include a.h in b.c and then compile only b.c. The compilation will go ok, but the linker will not find the symbol foo, since you did not compile it. So, you have to compile it, the compiler will create a.o file, and will link it with b.o file to generate whatever.out file. HTH, - -- Ofer Goren Iamba Technologies "Junghyun Cho" wrote in message news:FDjR7.3343$I3.35754@news.bora.net... > > "Ofer Goren" wrote in message > news:newscache$7z46og$tt5$1@lnews.actcom.co.il... > > I think the problem is that your vxSim.exe does not include the timerLib > > library. You should add the component to the vxSim. > > Isn't it enough to inlcude 'timers.h' file in my code ? Please give me > detail information. > Thanks in advance. > > > > > -- > > Ofer Goren > > Iamba Technologies > > "Junghyun Cho" wrote in message > > news:lJhR7.3204$I3.27470@news.bora.net... > > > Hi, there. > > > > > > I made simple "downloadable application module" and target was "VxSim". > > > Of course, I succeeded to compile and build it. But I couldn't download > it > > > to the target with the following error messages : > > > > > > Unresolved symbols encountered while downloading: > > > Errors while downloading > > > /home2/tomato/user/jhcho/projects/timer/SIMSPARCSOLARISgnu/timer.out: > > > timer_create > > > timer_connect > > > timer_settime > > > > > > > > > Here is the code. > > > > > > #include "vxWorks.h" > > > #include "semLib.h" > > > #include "timers.h" > > > #include "logLib.h" > > > > > > SEM_ID semId; > > > > > > int tmohandler() > > > { > > > semGive(semId); > > > } > > > > > > STATUS job_periodic() > > > { > > > while (1) { > > > semTake(semId, WAIT_FOREVER); > > > > > > printf("\njob_periodic : Timer expired."); > > > } > > > } > > > > > > STATUS timedemo(void) > > > { > > > int ret; > > > timer_t timerId; > > > struct itimerspec timeToSet; > > > > > > timeToSet.it_interval.tv_sec = 1; > > > timeToSet.it_interval.tv_nsec = 0; > > > timeToSet.it_value.tv_sec = 1; > > > timeToSet.it_value.tv_nsec = 0; > > > > > > ret = timer_create(CLOCK_REALTIME, NULL, &timerId); > > > if (ret == ERROR) { > > > printf("\ntimer_create error !"); > > > } > > > > > > ret = timer_connect(timerId, (VOIDFUNCPTR)tmohandler, 0); > > > if (ret == ERROR) { > > > printf("\ntimer_connect error !"); > > > } > > > > > > ret = timer_settime(timerId, CLOCK_REALTIME, &timeToSet, > > NULL); > > > if (ret == ERROR) { > > > printf("\ntimer_settime error !"); > > > } > > > } > > > > > > int start() > > > { > > > semId = semBCreate(SEM_Q_FIFO, SEM_EMPTY); > > > if (semId == NULL) { > > > printf("\ntimer_settime error !"); > > > } > > > > > > #define PRI 110 > > > taskSpawn("timedemo", PRI, 0, 4096, timedemo, 0); > > > taskSpawn("job_per", PRI, 0, 4096, job_periodic, 0); > > > } > > > > > > > > > Does anyone give me the solution to download correctly ? > > > > > > > > > > > > > > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: UDP-IP performances and timing with VxWorks Date: Tue, 11 Dec 2001 16:40:51 +0100 From: christophe leroy Organization: CS S I Message-ID: <3C162903.9EC18AF0@c-s.fr> References: <3C160B8C.482E2ABD@c-s.fr> <9v55h1$3ap$1@snoopy.bndlg.de> CPU is MPC860 at 40Mhz (Power PC) NIC is MC68160A C.Leroy > --------------------------- Newsgroups: comp.os.vxworks Subject: File I/O on host from embedded PPC target Date: Tue, 11 Dec 2001 10:57:07 -0600 From: "cwebb" Organization: WebUseNet Corp. - "ReInventing The UseNet" Message-ID: I am attempting to read from/write to files on a WinNT host from a PowerPC 750 target. Can anyone give me some assistance on how to accomplish this. Is it possible to use the standard C- I/O functions (e.g. fopen, fscanf, fprintf,...) to access the files? Will using the nfsMount command make the host file system visible to the embedded target? I would appreciate any help on this subject. C. Webb --------------------------- Newsgroups: comp.os.vxworks Subject: Re: getting arrow keys Date: Tue, 11 Dec 2001 11:08:23 -0600 From: "cwebb" Organization: WebUseNet Corp. - "ReInventing The UseNet" Message-ID: References: <3C142C96.F1A3EE5@home.com> In the WindSh you can press the key to enable the H-J-K-L keys as arrow keys as they are implemented in the Vi editor. You can also enable Vi style line editing by pressing the I key after is pressed. After the key is entered, you must press again to re-enable line editing & command recall. C. Webb Maximum Technology Corporation - --------------------------------------------------------------- "Gwen & Emory Stagmer" wrote in message news:3C142C96.F1A3EE5@home.com... > You have to get multiple characters on a PC. > > First charcter tells you it's a cursor control character, > second one tells you which one. > > Sorry, don't have the specifics here, but I'm sure you can find it > by hunting around on the internet. If not write me back and I'll > hunt it down for you. I know I've got it handy at work... > > Emory_Stagmer@LittonAS.com > > Emory Stagmer > Northrup Grumman Space Tech & Services Software Group > > Hashim Qaderi wrote: > > > > hello, > > > > Does anyone know how to get arrow keys > > from the keyboard directly. When I press > > arrows keys the cursor on the vxSim goes > > up and down rather than getting the character itself. > > > > thanks in advance. > > Jacob. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: MPC8260 and SPI problem Date: Tue, 11 Dec 2001 19:20:41 +0000 (UTC) From: "Mario Alvarez" Organization: Mailgate.ORG Server - http://www.Mailgate.ORG Message-ID: <3266884f3c89006f5ad7eb38d888b6f1.45946@mygate.mailgate.org> References: <3C07F776.F73FDDE1@nospam.com> Greetings, What a coincidence... we are encountering the exact same scenario where I work ! - - Mario Alvarez - -- Posted from h-66-134-29-226.nycmny83.covad.net [66.134.29.226] via Mailgate.ORG Server - http://www.Mailgate.ORG --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Send message to the target from Window's application Date: 11 Dec 2001 11:30:05 -0800 From: chernack@juno.com (Charles R. Chernack) Organization: http://groups.google.com/ Message-ID: References: <4367c1ab.0112092238.7630e8bd@posting.google.com> <9v2gi4$hgt$1@bcarh8ab.ca.nortel.com> I am trying to do the same thing using Visual Basic 6 on the host. Catalyst (http://www.catalyst.com/) has a free socket plug in for VB6 and LOTS of DOCS. I'm not to that point in my project yet, so I cannot report results. Charles R. Chernack Los Altos, CA --------------------------- Newsgroups: comp.os.vxworks Subject: Re: linking question Date: Tue, 11 Dec 2001 16:23:53 -0500 From: david lindauer Message-ID: <3C167969.20506@notifier-is.net> References: <3C153C06.811549CC@novtek.com> the problem is that the compiler will assign variable addresses relative to the code's data segment; moving a variable out of the code's data segment to another location after the fact is not realistic. But there are two possible solutions provided by some tool chains. First, if LD will handle it you can try declaring the variable external in all your program modules, then declare it with the defsym option. Possibly this will result in LD resolving all the external references to your variable to the address you have specified. Another alternative that may work is putting the variable in its own file, then telling LD to relocate that file's data section to the address you want. Note that if LD supports either approach you will have to initialize your variables by hand. David Yuri Regelman wrote: >Hello, > >Does anyone know how to specify fixed addresses for variables during the >link process? Basically I would like to make sure >certain variables end up in the fast internal ram of the cpu, while >letting the target based loader figure out the placement for the rest of > >them. -defsym option for ld allows to add a variable with a fixed >address, but not specify the location of an existing variable. > >Thanks for your help, >Yura > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: MPC8260 and SPI problem Date: Tue, 11 Dec 2001 22:08:32 GMT From: "Jim" Organization: Magma Communications Ltd. Message-ID: References: <3C07F776.F73FDDE1@nospam.com> I use an 860 and toggle gpio pins manually because of a microcode patch to perform i2c communications ... worst case you can try this ...its not that hard to do ... just tedious ... jim "cheru" wrote in message news:3C07F776.F73FDDE1@nospam.com... > Hi everyone, > > BACKGROUND: > We are using the 8260 SPI port in single master mode. The os is > vxworks5.4 > The BSP is based on SBC8260. > We have a single device on the SPI. It is an FPGA with a SPI driver > interface > programmed into it (our code). > > The way the SPI is configured is: 48 bits Tx and Rx. > We were using the SPI at 16 MHz. (BRGCLK/4, PM=0) > > CORE=200Mhz, CPM=133Mhz, Bus Freq=66Mhz. > > SPI settings > ------------ > The brgClk to spi=66.5Mhz. > The parameter ram is user programmed in dual port ram (Bank#8, > address=0x3800) > The Transmit and receive buffer descriptor are used from bank#0 of dual > port ram, i.e., address=0x0000 > The transmit and receive buffer pointer in Tx/Rcv BD are in dual port > ram address 0x2400 and 0x3900 respectively. > MRBLR in parameter ram=0x10. > parameter ram(RFCR)=0x1 > parameter ram(TFCR)=0x1 > > spi registers initialized to following > -------------------------------------- > SPI mode register: > CI=0, CP=0, REV=1, MS=1, EN=1 (set to 0 if there is no spi operation), > LEN=8, PM=1. > > > PROBLEMS: > 1) We occasionally we experience an incorrect data read from the RX > buffer. > The 7 (and sometimes 8) MSBitss are all ZEROs. When we examine the > SPIMISO, > we see the correct data containing some ONEs. We can detect this if it > occurs in the first two bytes > but the last four bytes are data which we have to accept if the first > two bytes are correct. > This ZEROing always occurs in the MSB of 16 bits increment. > > 2) Separately, we are not receiving indication via the SPIE, event > register, that the RXB occurred even when TxB is set. > Our SW has timeouts in the event of this so that it can carry on. > > At 16MHz, we would see these incorrect data reads more frequently. > We slowed the clock down to 8 MHz and the occurrence decreased but did > not > go away entirely. > > If we drop the speed lower, we run into other issues with > software/hardware erratic > behaviour. So 4MHz and 5.3MHz we could not use this setting. > > We did look at the clock signal for each of the 4 clock frequency > setting. > The results: at 4 and 5 MHz, the clock is continuous. > at 8 MHz the clock had a single break in it after the 16th cycle > and the 16 MHz clock had 4 breaks in it usually after the 8th, 24th, > 32nd and 40th cycles. > > It looks like the high frequencies may have some impact on the CPM. May > be someone can shed some light on this. > > The CPM is configured with the use of 4 ports, obviously the SPI running > > at 8 MHz, > the FCC3 is used for Enternet at 100 MHz, the I2C bus at 30KHz and the > SMC at 9600. > > We input these rates to a motorola utility to determine the utilization > of the CPM and found it at 52.52%. > > QUESTIONS: > 0) Why is the data ZERO'd and SPIE RXB is not SET, when TXB is SET? > > 1) Is there some SPI priority settings built in or can the priority be > set wrt to FCC > I2C, SCC etc. > > 2) Is the CPM resourced sufficiently? Are we using the SPI too fast? > > 3) Currently we use dual port ram to store tx/rx buffer BD and tx/rx > buffer pointers. > changing rx buffer pointer to use SDRAM does not work. What registers > specific to SDMA > need to be configured so that SDMA can arbitrate the 60x bus, park > itself on 60x bus > and use it to store the incoming SPIMISO data into the receive buffer in > > sdram? > > Thank you for your help. > > cheru > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: PCMCIA DRIVER for VxWorks Date: Wed, 12 Dec 2001 09:14:25 +0900 From: "ggang1" Organization: Inet Internet Services Message-ID: <9v67j4$smc$1@news.nuri.net> References: <9v2j9p$2kn$1@newshost.mot.com> Hi guys... I used the PCMCIA Flash card in MPC860 board. MPC860 have a PCMCIA Driver to use that. so you need to connect the hardware pin between MPC860 and PCMCIA Card and then as the PCMCIA cards,you should program. (you have to read the data sheet being relative the cards.) You don't need any stack and so on. That's it.very easy to program. good luck. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Task status DEAD Date: Wed, 12 Dec 2001 01:46:04 GMT From: pkockritz@mac.com (Pete Kockritz) Organization: AT&T Broadband Message-ID: References: <9v5880$tce$1@newsreader.mailgate.org> In article <9v5880$tce$1@newsreader.mailgate.org>, "Johan Bostedt" wrote: > Has anyone seen task status DEAD for a VxWorks task? > According to the Reference Manual the function taskStatusString() can > indicate task status DEAD. > A task that has terminated normally will not exist and the taskStatusString > call will fail. > But is there a case where an existing task can have status DEAD? The only time I've seen 'DEAD' is when I have a high priority task that has done several logMsg() calls and exited before they have all been written out the serial port. This can happen if you lower the priority of tLogTask to avoid it having any impact on some critical timing events. - -- Regards, Pete Kockritz --------------------------- Newsgroups: comp.os.vxworks Subject: Re: File I/O on host from embedded PPC target Date: Wed, 12 Dec 2001 01:52:54 GMT From: pkockritz@mac.com (Pete Kockritz) Organization: AT&T Broadband Message-ID: References: In article , "cwebb" wrote: > I am attempting to read from/write to files on a WinNT > host from a PowerPC 750 target. Can anyone give me > some assistance on how to accomplish this. Is it possible > to use the standard C- I/O functions (e.g. fopen, fscanf, > fprintf,...) to access the files? Will using the nfsMount > command make the host file system visible to the embedded > target? We use the target server file system for this (TSFS). You must have it included in your kernel and enable it when you start the target server. See section 4.7 of the VxWorks Programmer's Guide (The Target Server File System: TSFS). You can use netDrv, and use the FTP server that comes with Tornado to do what you want. You could also use NFS, but you'll need an NFS server, of course. - -- Regards, Pete Kockritz --------------------------- Newsgroups: comp.os.vxworks Subject: initialize multiple ethernet interface when boot vxworks? Date: 11 Dec 2001 18:18:53 -0800 From: sardine_sz@163.net (sardine) Organization: http://groups.google.com/ Message-ID: hi ,all vxworks guys, now i use Tornado2、vxworks5.4 for x86. my PC Target have two ethernet interface,so i want to initialize the two ethernet interface when boot vxworks. for example: Attached TCP/IP interface to elPci0.............. Attached TCP/IP interface to elPci1.............. any info appreciate ! regards! --------------------------- Newsgroups: comp.os.vxworks Subject: how to add hdlc to vxworks! Date: 11 Dec 2001 20:27:41 -0800 From: youerming@163.com (Torna) Organization: http://groups.google.com/ Message-ID: sorry, having snds100 for arm7 demo board with hdlc ports. i have the hdlc source code with the bsp of snds100 , but i dont know how to create the devie and add it to vxworks. could anybody tell me about it? --------------------------- Newsgroups: comp.os.vxworks Subject: Binding a socket to a interface in VxWorks Date: 11 Dec 2001 21:22:37 -0800 From: rashanmu@npd.hcltech.com (Ramesh) Organization: http://groups.google.com/ Message-ID: <15a87607.0112112122.66c11487@posting.google.com> hi, Is VxWorks provides the support of binding a socket to a specific interface if multiple interfaces are provided. if so how to do it. Also consider that if some of the interfaces are unnumbered (i.e) they do not possess a IP address at all. (Ex ) SONET/SDH interfaces. In such case is that possible to open a raw socket and bind it to a specific interface. Kindly reply me on both my questions seperately. Thanks Ramesh --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Need help - Unresolved symbol error while downloading Date: Tue, 11 Dec 2001 09:03:42 +0100 From: "Mario Semo" Organization: Nextra Telekom GmbH Message-ID: <9v6uid$v85$1@at-vie-newsmaster01.nextra.at> References: have you included the posix (i assume the calls are posix timer calls) timer support as you compiled your vxSim? you have to recompile vxSim as you have to rebuild your target vxWorks. - -- regards, Mario Semo. http://www.kirchnersoft.com Junghyun Cho wrote in message news:lJhR7.3204$I3.27470@news.bora.net... > Hi, there. > > I made simple "downloadable application module" and target was "VxSim". > Of course, I succeeded to compile and build it. But I couldn't download it > to the target with the following error messages : > > Unresolved symbols encountered while downloading: > Errors while downloading > /home2/tomato/user/jhcho/projects/timer/SIMSPARCSOLARISgnu/timer.out: > timer_create > timer_connect > timer_settime > > > Here is the code. > > #include "vxWorks.h" > #include "semLib.h" > #include "timers.h" > #include "logLib.h" > > SEM_ID semId; > > int tmohandler() > { > semGive(semId); > } > > STATUS job_periodic() > { > while (1) { > semTake(semId, WAIT_FOREVER); > > printf("\njob_periodic : Timer expired."); > } > } > > STATUS timedemo(void) > { > int ret; > timer_t timerId; > struct itimerspec timeToSet; > > timeToSet.it_interval.tv_sec = 1; > timeToSet.it_interval.tv_nsec = 0; > timeToSet.it_value.tv_sec = 1; > timeToSet.it_value.tv_nsec = 0; > > ret = timer_create(CLOCK_REALTIME, NULL, &timerId); > if (ret == ERROR) { > printf("\ntimer_create error !"); > } > > ret = timer_connect(timerId, (VOIDFUNCPTR)tmohandler, 0); > if (ret == ERROR) { > printf("\ntimer_connect error !"); > } > > ret = timer_settime(timerId, CLOCK_REALTIME, &timeToSet, NULL); > if (ret == ERROR) { > printf("\ntimer_settime error !"); > } > } > > int start() > { > semId = semBCreate(SEM_Q_FIFO, SEM_EMPTY); > if (semId == NULL) { > printf("\ntimer_settime error !"); > } > > #define PRI 110 > taskSpawn("timedemo", PRI, 0, 4096, timedemo, 0); > taskSpawn("job_per", PRI, 0, 4096, job_periodic, 0); > } > > > Does anyone give me the solution to download correctly ? > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Abstruct class (C++) with VxWorks Date: Tue, 11 Dec 2001 09:08:52 +0100 From: "Mario Semo" Organization: Nextra Telekom GmbH Message-ID: <9v6uii$v85$2@at-vie-newsmaster01.nextra.at> References: <4367c1ab.0112092233.2bede9f6@posting.google.com> please tell me: a) what is in your opinion an abstract class (i assume you mean an abstract and not an abstruct class) b) you can you download such a construct ? c) what do you want to do? - -- regards, Mario Semo. http://www.kirchnersoft.com Yaron wrote in message news:4367c1ab.0112092233.2bede9f6@posting.google.com... > When I try to download an application which is writen in CPP and which > contains an Abstruct class I get a "download error" message. Whats > wrong??? --------------------------- Newsgroups: comp.os.vxworks Subject: problem using sntpc Date: 12 Dec 2001 01:23:09 -0600 From: "Gil Rapaport" Message-ID: <3c1704bd$1@news.barak.net.il> Hi, I've added the SNTP Client package to my VxWorks 5.4 image, and as far as I can tell it built fine (lkup shows sntpcTimeGet, sntpcInit and sntpcPort exist in the image), but I still get unresolved symbols when I try to call any of the 2 functions in my application (I can call them directly from the shell using some static variables. It doesn't work, but at least the symbol is found). I've also tried declaring the functions myself and mess a little with the signature instead of including sntpcLib.h, but no luck there either. Any ideas ? Thanks, Gil --------------------------- Newsgroups: comp.os.vxworks Subject: tNetTask stack overrun Date: Wed, 12 Dec 2001 07:47:34 GMT From: mdierikx@gti-group.com (mdi@gti) Organization: EuroNet Internet Message-ID: <3c170ab9.475463@194.134.193.17> Reply-To: postmaster@localhost Is there any way to increase the amount of stack available to tNetTask? In my project, somehow nNetTask's stack gets overrun. What makes tNetTask's stack grow anyway? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: initialize multiple ethernet interface when boot vxworks? Date: 12 Dec 2001 08:54:10 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: sardine_sz@163.net (sardine) wrote: > hi ,all vxworks guys, > > now i use Tornado2、vxworks5.4 for x86. > my PC Target have two ethernet interface,so i want to initialize > the two ethernet interface when boot vxworks. > > for example: > > Attached TCP/IP interface to elPci0.............. > Attached TCP/IP interface to elPci1.............. > > > any info appreciate ! Use the "other" boot parameter. 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: Please Help Me Date: 12 Dec 2001 02:23:42 -0800 From: digipsyche@sify.com (DigiPsyche) Organization: http://groups.google.com/ Message-ID: <6c32a22b.0112120223.1cfc1bb2@posting.google.com> Hi, I am an internet application developer and working on java, jsp, perl, php, javascript technologies and very new to this group. i am very much intrested in learning and making my future in embedded technologies. i have very less knowledge about all the enbedded technologies. Please help me, which technologies i should know to go for 3G wireless communication and VOIP/Data communication and from where to startup? Is there any website to download the free version of these softwares? thanks DigiPsyche --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Wed Dec 12 08:37:20 2001 From: Doug Owens Date: Wed Dec 12 08:37:22 PST 2001 Subject: Quad G4 and C++ We are looking into making an airborne IR video acquisition/analysis system using the DNA quad G4 boards on a VME/Raceway interconnect. I am looking for input from anyone using these boards. Also, our present vxWorks development is done using C (the gnu compiler provide by wrs) on an MV1603. I'm guessing that C++ is relatively stable on vxWorks/Tornado by now but it would be nice to hear anybody's experience - also input on whether the gnu compiler or a third party compiler is preferable for C++ development. Thanks in advance Doug Owens owens2@llnl.gov From vxwexplo-errs@csg.lbl.gov Wed Dec 12 19:46:20 2001 From: "Garner, William F (N-SDC)" Date: Wed Dec 12 19:46:22 PST 2001 Subject: TFTP Stops After 33 Packets are Received Hi VxWorks Land, When I try to get a file using TFTP, then only 33 packet are received. Also, clusters are not freed. Has anyone seen this before? -> sp tftpXfer,"ch1",0,"of_boot.src","get","ascii",&errorFd,&dataFd task spawned: id = 0xb05b90, name = t1 value = 11557776 = 0xb05b90 -> { tTftpTask} sent RRQ { tTftpTask} received DATA { tTftpTask} sent ACK { tTftpTask} received DATA { tTftpTask} sent ACK { tTftpTask} received DATA { tTftpTask} sent ACK { tTftpTask} received DATA { tTftpTask} sent ACK { tTftpTask} received DATA { tTftpTask} sent ACK { tTftpTask} received DATA { tTftpTask} sent ACK { tTftpTask} received DATA { tTftpTask} sent ACK { tTftpTask} received DATA { tTftpTask} sent ACK { tTftpTask} received DATA { tTftpTask} sent ACK { tTftpTask} received DATA { tTftpTask} sent ACK { tTftpTask} received DATA { tTftpTask} sent ACK { tTftpTask} received DATA { tTftpTask} sent ACK { tTftpTask} received DATA { tTftpTask} sent ACK { tTftpTask} received DATA { tTftpTask} sent ACK { tTftpTask} received DATA { tTftpTask} sent ACK { tTftpTask} received DATA { tTftpTask} sent ACK { tTftpTask} received DATA { tTftpTask} sent ACK { tTftpTask} received DATA { tTftpTask} sent ACK { tTftpTask} received DATA { tTftpTask} sent ACK { tTftpTask} received DATA { tTftpTask} sent ACK { tTftpTask} received DATA { tTftpTask} sent ACK { tTftpTask} received DATA { tTftpTask} sent ACK { tTftpTask} received DATA { tTftpTask} sent ACK { tTftpTask} received DATA { tTftpTask} sent ACK { tTftpTask} received DATA { tTftpTask} sent ACK { tTftpTask} received DATA { tTftpTask} sent ACK { tTftpTask} received DATA { tTftpTask} sent ACK { tTftpTask} received DATA { tTftpTask} sent ACK { tTftpTask} received DATA { tTftpTask} sent ACK { tTftpTask} received DATA { tTftpTask} sent ACK { tTftpTask} received DATA { tTftpTask} sent ACK { tTftpTask} received DATA { tTftpTask} sent ACK { tTftpTask} received DATA size clusters free usage ---------------------------------------------------------------------------- --- 64 100 0 5881 128 100 17 11431 256 40 40 0 512 40 40 0 1024 25 25 0 2048 25 25 0 --------------------------------------------------------------------------- Thanks in advance. Bill william.f.garner@lmco.com From vxwexplo-errs@csg.lbl.gov Wed Dec 12 21:26:47 2001 From: "RajaSekhar" Date: Wed Dec 12 21:26:49 PST 2001 Subject: handling multithread exceptions Friends, Can any one help me in handling exceptions in multithread scenarios in VxWorks. Thanks in advance for any help Rgds Raja Sekhar From vxwexplo-errs@csg.lbl.gov Wed Dec 12 23:49:20 2001 From: Ronen Levy Date: Wed Dec 12 23:49:22 PST 2001 Subject: Tornado - Memory window Hi, I noticed that in Tornado the docked debugging memory window is capable only of dumping target memory. In most debuggers the user also can set a value in the memory window. The only possiblity to write to target memory is with the host shell command 'm address data'. Q: Is there any extension to memory window in Tornado, that enables to write to a specific target memory address? Q: did what I describe above is valid? (only dump data and not able to set?!) Ronen. From vxwexplo-errs@csg.lbl.gov Thu Dec 13 00:51:48 2001 From: Ilia Bosis Date: Thu Dec 13 00:51:50 PST 2001 Subject: Re: tNetTask stack overrun The network driver receive path (exactly speaking the function referred in the netJobAdd() call) is executed under the context of the tNetTask. So look at Your driver for possible bug. In addition, there is executed the networking stack and, if it decides to forward the packet, the transmit path of the transmitting driver. Take a look. Best regards. Ilia. -------------------------------------------- Ilia Bosis Software Engineer PacketLight Networks Tel: +972-9-7644228 Fax: +972-9-7645444 E-mail: Ilia_Bosis@packetlight.com Web: www.packetlight.com --------------------------- Newsgroups: comp.os.vxworks Subject: tNetTask stack overrun Date: Wed, 12 Dec 2001 07:47:34 GMT From: mdierikx@gti-group.com (mdi@gti) Organization: EuroNet Internet Message-ID: <3c170ab9.475463@194.134.193.17> Reply-To: postmaster@localhost > Is there any way to increase the amount of stack available to > tNetTask? In my project, somehow nNetTask's stack gets overrun. > What makes tNetTask's stack grow anyway? From vxwexplo-errs@csg.lbl.gov Thu Dec 13 04:03:15 2001 From: Vxworks Exploder Date: Thu Dec 13 04:03:17 PST 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Thu Dec 13 04:03:11 PST 2001 Subject: Re: tNetTask stack overrun Subject: need help with envoy 9.1 Subject: Re: Please Help Me Subject: Where are my register values! Subject: Re: Where are my register values! Subject: problem with flash that contains dosFs and bootload Subject: Re: problem with flash that contains dosFs and bootload Subject: Re: Please Help Me Subject: Re: tNetTask stack overrun Subject: Re: problem with flash that contains dosFs and bootload Subject: Re: problem with flash that contains dosFs and bootload Subject: MTD flash driver Q: Subject: error while downloading Subject: usrFsLib cp() and rm() functions Subject: Re: Wind River PPC750 / MPC107 board PCI config help Subject: PowerPC in Pentium system Subject: Re: PowerPC in Pentium system Subject: Re: PowerPC in Pentium system Subject: Re: usrFsLib cp() and rm() functions Subject: Re: problem with flash that contains dosFs and bootload Subject: setFileDateTime: how? Subject: Re: setFileDateTime: how? Subject: WindNet PPP and compression ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: tNetTask stack overrun Date: Wed, 12 Dec 2001 13:35:52 -0000 From: "David Laight" Organization: BT Internet Message-ID: <9v7mc1$mbc$1@neptunium.btinternet.com> References: <3c170ab9.475463@194.134.193.17> > Is there any way to increase the amount of stack available to > tNetTask? In my project, somehow nNetTask's stack gets overrun. > What makes tNetTask's stack grow anyway? The stack size is taken from the global variable netTaskStackSize, you can set this in some initialisation code before tNetTask is started. However it is more likely to be the fault of something you are getting tNetTask to run..... David --------------------------- Newsgroups: comp.os.vxworks Subject: need help with envoy 9.1 Date: Wed, 12 Dec 2001 16:35:31 +0200 From: "Ofer Goren" Organization: Verio Message-ID: Hi. Tornado 2.0.2, running on win2kSp2. If anybody ever tailored the envoy's SNMP stack to his/her vxworks application, please reply to this post. Thanks. - -- Ofer Goren Iamba Technologies --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Please Help Me Date: 12 Dec 2001 14:47:30 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: <6c32a22b.0112120223.1cfc1bb2@posting.google.com> digipsyche@sify.com (DigiPsyche) wrote: > Hi, > I am an internet application developer and working on java, jsp, > perl, php, javascript technologies and very new to this group. i am > very much intrested in learning and making my future in embedded > technologies. i have very less knowledge about all the enbedded > technologies. Please help me, which technologies i should know to go > for 3G wireless communication and VOIP/Data communication and from > where to startup? Is there any website to download the free version of > these softwares? Go to www.wrs.com, and follow the links to Tornado Prototyper. This is a free evaluation version of the Tornado and the VxWorks simulator. 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: Where are my register values! Date: Wed, 12 Dec 2001 09:54:17 -0500 From: "Granum, Allen [CAR:CF81:EXCH]" Organization: Nortel Networks Message-ID: <3C176F99.D567DFD9@americasm01.nt.com> I would like to capture the register values of a running task using the eventpiont utility (PPC platform). The problem is that once I am in the eventpoint, the register values have been altered from their state prior to the event point occuring. Does anyone know how to get at the original values of the registers before the eventpoint happened, from within the eventpoint bound routine. Thanks, - -- Allen Granum Nortel Networks Wireless Internet BSP Development Team Email: agranum@nortelnetworks.com Phone: (613)763-4324 --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Where are my register values! Date: Wed, 12 Dec 2001 18:23:03 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <3C176F99.D567DFD9@americasm01.nt.com> Hmm, this is a truly challenging question, trouble is I do not remember the exact details of WindView eventPoint handling, except that it is a generalization of native debugging breakpoints. This would lead me to the conclusion that the eventPoint handler is actually running in tExcTask context. If you verify this to be true, then the solution is real easy, and even can be architecture independent, you can retreive register values with the native debug library functions. HTH, - - Leonid "Granum, Allen [CAR:CF81:EXCH]" wrote in message news:3C176F99.D567DFD9@americasm01.nt.com... > > I would like to capture the > register values of a running > task using the eventpiont > utility (PPC platform). > > The problem is that once I am > in the eventpoint, the > register values have been > altered from their state prior > to the event point occuring. > Does anyone know how to get at > the original values of the > registers before the > eventpoint happened, from > within the eventpoint bound > routine. > > Thanks, > > -- > Allen Granum > Nortel Networks > Wireless Internet > BSP Development Team > Email: agranum@nortelnetworks.com > Phone: (613)763-4324 > > > --------------------------- Newsgroups: comp.os.vxworks Subject: problem with flash that contains dosFs and bootload Date: Wed, 12 Dec 2001 16:38:59 GMT From: "haha" Organization: Magma Communications Ltd. Message-ID: Hi, I am having a problem with my flash. This flash contains both bootloader and dos fileSystem. __________ | bank 0 | | Bootloader | |_________ | | Bank 1 | |BootLoader | |_________ | | DosFs | |_________| Now I wanted download new BootImage bank 1 in the flash . I used the tffsRawio to write to it. But tffsRawio fails. Only when I format my flash using sysTffsFormat then I could use tffsRawio But if I format the flash I am loose my file system. Does anyone have any alternative solution to this. Please email me hdewan@spacebridge.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: problem with flash that contains dosFs and bootload Date: 12 Dec 2001 11:58:24 -0500 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: References: Sender: bpringlemeir@DeadDuck [bcc'd to poster] >>>>> "haha" == haha writes: haha> Hi, I am having a problem with my flash. This flash contains haha> both bootloader and dos fileSystem. [snip] haha> Now I wanted download new BootImage bank 1 in the flash . I haha> used the tffsRawio to write to it. But tffsRawio fails. Only haha> when I format my flash using sysTffsFormat then I could use haha> tffsRawio But if I format the flash I am loose my file system. The `tffsRawio' is probably failing because you haven't erased the bootloader flash space. You could call an MTD routine directly to erase the sectors before you call the `tffsRawio' to write the data. The MTD might try to distinguish if the address it has been given is in a memory space that it knows about. However, you should have the source to the MTD that you are using. It is also fairly simple to write a function that will erase this area yourself. hth, Bill Pringlemeir. - -- Have you ever had a small midget tell you that Nazi Germany is your country of birth? Or tried to psychically link with boxes of Borax while staring at lightbulbs? 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: Please Help Me Date: Wed, 12 Dec 2001 19:03:55 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <6c32a22b.0112120223.1cfc1bb2@posting.google.com> Digi, Dear, Your competencies are a far cry from what a decent embedded engineer needs to have. but if you have the guts for embedded, then here is a partial sillabus that can get you there: + A basic course in hardware including Logic and processor architectures + You must know at least 2 assembly languages, preferrably one classic CISC (e.g. PDP11, MC68K) and one RISC (Sparc or MIPS ) + You need to be proficient in "C", including a few hundered thousands lines of code you wrote and debugged on a non-embedded platform + You may already know communications to some extent, but in the Embedded world, it's all a matter of bits and bytes, so you need to get low and dirty on those as well If you aim specifically at the 3G cellphone development, then a good level mathematics, specifically in Signal Processing is a big addvantage. So it's a matter of time, money and guts, Good luck! - Leonid "DigiPsyche" wrote in message news:6c32a22b.0112120223.1cfc1bb2@posting.google.com... > Hi, > I am an internet application developer and working on java, jsp, > perl, php, javascript technologies and very new to this group. i am > very much intrested in learning and making my future in embedded > technologies. i have very less knowledge about all the enbedded > technologies. Please help me, which technologies i should know to go > for 3G wireless communication and VOIP/Data communication and from > where to startup? Is there any website to download the free version of > these softwares? > > thanks > DigiPsyche --------------------------- Newsgroups: comp.os.vxworks Subject: Re: tNetTask stack overrun Date: Wed, 12 Dec 2001 19:06:04 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <3c170ab9.475463@194.134.193.17> I beleive there is a constant in configAll.h that sets the NetTask stack size on a per-architecutre basis. Typically if you have added a new network driver which is more complicated then teh usual, you will need to increase the stack size, e.g. if you are adding a Classical IP over ATM which adds an END or BSD driver. "mdi@gti" wrote in message news:3c170ab9.475463@194.134.193.17... > Is there any way to increase the amount of stack available to > tNetTask? In my project, somehow nNetTask's stack gets overrun. > What makes tNetTask's stack grow anyway? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: problem with flash that contains dosFs and bootload Date: Wed, 12 Dec 2001 19:35:04 GMT From: "Hadi" Organization: Magma Communications Ltd. Message-ID: References: i Bill, Thanks for the quick response. I will try to elabrate the problem. You are right that I have to erase the flash segment before I could write to it. My flash disk is 8 meg. TrueTffs is installed to operate on the last 2 meg (6-8). So the first 6 meg is not under TrueTffs. Here is the twist, only after I format TrueTffs my function that writes to the bank 0 is successful. But if don't format tffsRawio() fails in the Erase. Following is my code. I highlighted the code that fails with --->> arrow. Your help is greatly appriciated. Hadi int tffsFlashRAMImagePut ( int driveNo, /* TFFS drive number */ int offset, /* offset in the flash chip/card */ int maxLoadSize, /* maximum load size */ char *image /* image to read from */ ) { PhysicalInfo info; UINT unitNo; UINT unitSize; UINT addr; char * pBuf; int ix, numBytes; int loadSize=0; int tmpImageSize=0; if (image == NULL){ printf("\nIMAGE==NULL"); return (ERROR); } if (tffsSocket[driveNo] == NULL) { printf("\ntffsSocket[%d]==NULL",driveNo); return (ERROR); } if (tffsRawio (driveNo, TFFS_GET_PHYSICAL_INFO, (int)&info, 0, 0) != OK) { printErr ("Unable to get physical info from drive\n"); return (ERROR); } pBuf = (char *)malloc (info.unitSize); if (pBuf == NULL) return ERROR; addr = offset; unitNo = offset / info.unitSize; unitSize = info.unitSize - (offset % info.unitSize); /* If the offset is not Erase Unit aligned we need to save the the * contents of the region begining at the start of this erase unit * and ending at the specified address so we can restore it after * we erase the Erase Unit */ if (tffsRawio (driveNo, TFFS_PHYSICAL_READ, unitNo * info.unitSize, offset % info.unitSize, (int)pBuf) != OK) { printErr ("Failed attempting to save Erase Unit %d\n", unitNo); free(pBuf); return (ERROR); } - --->> if (tffsRawio (driveNo, TFFS_PHYSICAL_ERASE, unitNo, 1, 0) != OK) - --->> { - --->> printErr ("Failed attempting to erase Erase Unit %d\n", unitNo); free(pBuf); return (ERROR); } if (tffsRawio (driveNo, TFFS_PHYSICAL_WRITE, unitNo * info.unitSize, offset % info.unitSize, (int)pBuf) != OK) { printErr ("Failed attempting to restore Erase Unit %d\n", unitNo); free(pBuf); return (ERROR); } tmpImageSize = maxLoadSize; while (unitSize) { if (tmpImageSize >= unitSize){ ix = unitSize; } else if (tmpImageSize == 0){ ix = 0; break; } else { ix = tmpImageSize; } memcpy(pBuf, image, ix); image += ix; tmpImageSize -= ix; if ((addr + ix) > info.mediaSize) { printErr ("Error : Attempting to write beyond Flash boundary\n"); free(pBuf); return (ERROR); } if ((addr + ix) > (offset + maxLoadSize)) { printErr ("Error : Attempting to write beyond Load boundary\n"); free(pBuf); return (ERROR); } numBytes = ix; loadSize +=numBytes; /*printf("\nWriting @ %x num=%x",addr,numBytes);*/ if (tffsRawio (driveNo, TFFS_PHYSICAL_WRITE, addr, numBytes, (int)pBuf) != OK) { printErr ("Physical write failed at address 0x%x\n", addr); printErr ("ByteSize = 0x%x\n", numBytes); printErr ("Retry with padding "); if (ix & 1) /* odd address size */ { numBytes = padByte(pBuf,ix, unitSize); } if (tffsRawio (driveNo, TFFS_PHYSICAL_WRITE, addr, numBytes, (int)pBuf) != OK) { free(pBuf); return (ERROR); } } addr += ix; unitSize -= ix; if (unitSize == 0) { unitSize = info.unitSize; unitNo++; /*printf("\nErasing unitNo %x",unitNo);*/ if (tffsRawio (driveNo, TFFS_PHYSICAL_ERASE, unitNo, 1, 0) != OK) { printErr ("Failed attempting to erase Erase Unit %d\n", unitNo); free(pBuf); return (ERROR); } } } /* printf("\nFromTffsDriver %x",loadSize);*/ free(pBuf); return (loadSize); } - -----Original Message----- From: Bill Pringlemeir [mailto:bpringlemeir@yahoo.com] Sent: Wednesday, December 12, 2001 11:58 AM Subject: Re: problem with flash that contains dosFs and bootload [bcc'd to poster] >>>>> "haha" == haha writes: haha> Hi, I am having a problem with my flash. This flash contains haha> both bootloader and dos fileSystem. [snip] haha> Now I wanted download new BootImage bank 1 in the flash . I haha> used the tffsRawio to write to it. But tffsRawio fails. Only haha> when I format my flash using sysTffsFormat then I could use haha> tffsRawio But if I format the flash I am loose my file system. The `tffsRawio' is probably failing because you haven't erased the bootloader flash space. You could call an MTD routine directly to erase the sectors before you call the `tffsRawio' to write the data. The MTD might try to distinguish if the address it has been given is in a memory space that it knows about. However, you should have the source to the MTD that you are using. It is also fairly simple to write a function that will erase this area yourself. hth, Bill Pringlemeir. - -- Have you ever had a small midget tell you that Nazi Germany is your country of birth? Or tried to psychically link with boxes of Borax while staring at lightbulbs? You will. And the company that will bring it to you: AT&T. vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" "Bill Pringlemeir" wrote in message news:uy9k8e5ov.fsf@yahoo.com... > [bcc'd to poster] > > >>>>> "haha" == haha writes: > > haha> Hi, I am having a problem with my flash. This flash contains > haha> both bootloader and dos fileSystem. > > [snip] > > haha> Now I wanted download new BootImage bank 1 in the flash . I > haha> used the tffsRawio to write to it. But tffsRawio fails. Only > haha> when I format my flash using sysTffsFormat then I could use > haha> tffsRawio But if I format the flash I am loose my file system. > > The `tffsRawio' is probably failing because you haven't erased the > bootloader flash space. You could call an MTD routine directly to > erase the sectors before you call the `tffsRawio' to write the data. > The MTD might try to distinguish if the address it has been given is > in a memory space that it knows about. However, you should have the > source to the MTD that you are using. It is also fairly simple to > write a function that will erase this area yourself. > > hth, > Bill Pringlemeir. > > -- > Have you ever had a small midget tell you that Nazi Germany is your > country of birth? Or tried to psychically link with boxes of Borax > while staring at lightbulbs? 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: MTD flash driver Q: Date: 12 Dec 2001 20:09:59 GMT From: afernando@adventnetworks.com Organization: NewsOne.Net - Free Usenet News via the Web - http://newsone.net/ Message-ID: <9v8din$un7$1@news.netmar.com> HI, i am writing a MTD for Intel 28F128J3A flash. we have 2 banks, and on each bank we have 4 chips, (8M x 64bits). There are number of things i dont understand about addressing in this case. The h/w guys tells me that i could access all 4 chips in parallel? then how do i access each device individually? in my case what is "noOfChips," 4 or 8 (for vxWorks, FLFlash structure). i guess i dont' understand this "interleaving" and "socket" concept used flash domain. Also, how do you address CUI (command user interface) of the Flash device? This is 16-bit in mode. any help is very much appreciated. ----- Posted via NewsOne.Net: Free (anonymous) Usenet News via the Web ----- http://newsone.net/ -- Free reading and anonymous posting to 60,000+ groups NewsOne.Net prohibits users from posting spam. If this or other posts made through NewsOne.Net violate posting guidelines, email abuse@newsone.net --------------------------- Newsgroups: comp.os.vxworks Subject: error while downloading Date: Wed, 12 Dec 2001 13:17:47 -0800 From: "Ambrish verma" Organization: University of Southern California Message-ID: <9v8hmg$4as$1@usc.edu> Sender: ambrish@h-162.marantinetworks.com hi, I have recently compiled tornado 2.0 with simulator supporting tcp/ip stack (i mean i patched it with ulip libraries for win2k). Now i worte some simple program to set priorities. it builds perfectly. I started simulator and vxworks shell and when tried to download this program for my simulator. it gives following errors: Errors while downloading c:/averma/vxProjs/priorityTest/SIMNTgnu/priorityTest.out: _sched_getparam _sched_setparam _sched_getscheduler thanks ambrish --------------------------- Newsgroups: comp.os.vxworks Subject: usrFsLib cp() and rm() functions Date: Wed, 12 Dec 2001 17:31:08 -0500 From: David Roberge Organization: Universite du Quebec Message-ID: <3C17DAAC.D3645CAC@ele.etsmtl.ca> Hi all, I'm trying to use the cp() and rm() functions defined in the usrFsLib library. I cannot compile my code because it seems that there is no .h file for these functions. I searched the directory tree of Tornado and I have not found anything but usrFsLib.o. I'm using Tornado/vxWorks AE. Does anyone knows how to use these functions ? Thanks in advance. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Wind River PPC750 / MPC107 board PCI config help Date: Thu, 13 Dec 2001 11:04:57 +1030 From: Tim Shaw Organization: Defence Science and Technology Organisation Message-ID: <3C17F7B1.DD9B442B@dsto.defence.gov.au> References: <3C15728E.D3D29007@west.raytheon.com> Why don't you put the questions onto the newsgroup, and see if you get any answers? "Michael S. Simpson" wrote: > I will be the first to admit that I know very little about PCI > configuration. We're working on a project where we are using a Wind > River build PMC board with a Motorola MPC750 CPU and a MPC107 > Bridge/Memory Controller chip. This will be connected to a custom built > carrier board with several ASICs and memory for video processing. There > are no other PCI devices, just the carrier board. > > We're trying to configure the PMC board as the PCI host. Our biggest > problem is trying to configure the memory map for the PCI space. Nobody > on the development team has any experience with the MPC107 and it's a > pain to figure out (Wind River has been some help, but hard to reach > people who really know what's going on with the BSP). Motorola is > usless for getting information from. > > If anyone has a little time to maybe answer some questions, please > respond to this message. Thanks... > > Mike Simpson > Software Engineer > Raytheon Missile Systems > Tucson, AZ > (520) 784-5381 --------------------------- Newsgroups: comp.os.vxworks,comp.sys.powerpc.tech Subject: PowerPC in Pentium system Date: Thu, 13 Dec 2001 01:32:53 GMT From: "Kelly Hornsby" Organization: Excite@Home - The Leader in Broadband http://home.com/faster Message-ID: <9zTR7.60827$py4.30969415@news2.nash1.tn.home.com> Hello All I'm tasked with finding out what sort of issues are involved in developing a PMC-based PCI card that will implement a PowerPC 750/MPC107 design, and run an OS, such as VxWorks, but all inside a Win2K Pentium machine. The OS control is not so bad, I think, but I don't really know how to attack the PCI Master/Slave configuration problem. What I'm not clear on is the interaction that must occur between the BIOS, the OS, and the PCI-PMC based PowerPC system. Is there a way to tell the PowerPC board, outside of a memory strobe and polling loop, that a request is pending for service, or vice versa? These are some of the things I'm thinking of at this point. This particular application will be for a high-performance I/O controller and simulation that I'm in the process of developing. I have the PCI specification, but it does not really cover what initialization (if any) is necessary for a slave mode system that can request bus mastering like my PowerPC board may be able to perform. My targeted system is a Pentium III system running Windows NT/2K. However, a Linux OS may be in order, but I don't really care that much, in that a device driver will have to be developed for whatever system this board will be used in. If anyone has performed work like this, could you email me so that I can find out some more about this? Otherwise, it'll be reinventing the wheel perhaps. Thx for any input. --------------------------- Newsgroups: comp.os.vxworks,comp.sys.powerpc.tech Subject: Re: PowerPC in Pentium system Date: Thu, 13 Dec 2001 13:07:51 +1030 From: Tim Shaw Organization: Defence Science and Technology Organisation Message-ID: <3C18147F.EE774331@dsto.defence.gov.au> References: <9zTR7.60827$py4.30969415@news2.nash1.tn.home.com> Kelly, Most likely you'll communicate between the PPC target and the x86 host using the mailbox and doorbell registers on the MPC107, depending on the information you need to send. These generate an interrupt of either the local processor (PPC) or the host (x86), depending on the register written to. Alternatively, you can (chain) DMA data from one side of the MPC107 to the other (or in fact even both source and destination on the same side of the bridge), initiated on either side of the bridge, with an interrupt generated at the end. But have you looked at the PLX and IBM PPC core + PCI interface + memory controller all in the same device? We've found them quite good so far for just that sort of application (not that there's anything wrong with the PPC750+MPC107 ...) HTH Tim Kelly Hornsby wrote: > Hello All > > I'm tasked with finding out what sort of issues are involved in developing a > PMC-based PCI card that will implement a PowerPC 750/MPC107 design, and run > an OS, such as VxWorks, but all inside a Win2K Pentium machine. The OS > control is not so bad, I think, but I don't really know how to attack the > PCI Master/Slave configuration problem. What I'm not clear on is the > interaction that must occur between the BIOS, the OS, and the PCI-PMC based > PowerPC system. Is there a way to tell the PowerPC board, outside of a > memory strobe and polling loop, that a request is pending for service, or > vice versa? These are some of the things I'm thinking of at this point. > This particular application will be for a high-performance I/O controller > and simulation that I'm in the process of developing. > > I have the PCI specification, but it does not really cover what > initialization (if any) is necessary for a slave mode system that can > request bus mastering like my PowerPC board may be able to perform. > > My targeted system is a Pentium III system running Windows NT/2K. However, a > Linux OS may be in order, but I don't really care that much, in that a > device driver will have to be developed for whatever system this board will > be used in. > > If anyone has performed work like this, could you email me so that I can > find out some more about this? Otherwise, it'll be reinventing the wheel > perhaps. > > Thx for any input. --------------------------- Newsgroups: comp.os.vxworks,comp.sys.powerpc.tech Subject: Re: PowerPC in Pentium system Date: Thu, 13 Dec 2001 05:28:52 GMT From: "Gary Milliorn" Organization: Road Runner - Texas Message-ID: References: <9zTR7.60827$py4.30969415@news2.nash1.tn.home.com> "Kelly Hornsby" wrote in message news:9zTR7.60827$py4.30969415@news2.nash1.tn.home.com... > Hello All > > I'm tasked with finding out what sort of issues are involved in developing a > PMC-based PCI card that will implement a PowerPC 750/MPC107 design, and run > an OS, such as VxWorks, but all inside a Win2K Pentium machine. The OS > control is not so bad, I think, but I don't really know how to attack the > PCI Master/Slave configuration problem. What I'm not clear on is the > interaction that must occur between the BIOS, the OS, and the PCI-PMC based > PowerPC system. Is there a way to tell the PowerPC board, outside of a > memory strobe and polling loop, that a request is pending for service, or > vice versa? These are some of the things I'm thinking of at this point. > This particular application will be for a high-performance I/O controller > and simulation that I'm in the process of developing. Look at the I2O section of the embedded utilities in the MPC107 UM, that's a very straightforward means of having the host notify your board that there's work to do, and vice-versa. > I have the PCI specification, but it does not really cover what > initialization (if any) is necessary for a slave mode system that can > request bus mastering like my PowerPC board may be able to perform. As a target board, the host (PC) will initialize your BAR registers using the data you present, for whatever size aperture you expect to need. We use a similar architecture with the Sandpoint system, with a MPMC card as the host and also as a target, the latter in a PCI slot using the Technobox PMC->PCI adapter board. DINK recognizes and configures the board. Sandpoint is at: http://e-www.motorola.com/webapp/sps/site/prod_summary.jsp?code=SANDPOINTX3& nodeId=01M98653 Technobox is at www.technobox.com Gary Milliorn --------------------------- Newsgroups: comp.os.vxworks Subject: Re: usrFsLib cp() and rm() functions Date: Thu, 13 Dec 2001 10:01:54 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <3C17DAAC.D3645CAC@ele.etsmtl.ca> There should be a usrFsLib.h file somewhere, but the matter of fact is that the functions in this library are intended for interactive use with target Shell, not for use by programs. You can (and should) use usrFsLib.c as reference source code to learn how to use the file system programmatically rather then call its functions. - - Leonid "David Roberge" wrote in message news:3C17DAAC.D3645CAC@ele.etsmtl.ca... > Hi all, > > I'm trying to use the cp() and rm() functions defined in the usrFsLib > library. I cannot compile my code because it seems that there is no .h > file for these functions. I searched the directory tree of Tornado and > I have not found anything but usrFsLib.o. I'm using Tornado/vxWorks > AE. Does anyone knows how to use these functions ? > > Thanks in advance. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: problem with flash that contains dosFs and bootload Date: Thu, 13 Dec 2001 09:03:08 +0100 From: "Michael Lawnick" Organization: Buergernetz Dillingen Message-ID: <9v9nc1$mko$1@snoopy.bndlg.de> References: Reply-To: "Michael Lawnick" Sender: mlawnick@pd9524721.dip.t-dialin.net Hi Hadi, erasing of sectors is done by tffsRawIo(,3(=erase),,) cave at: the sectors are physical-flash. BTW AFAIK this is all described in your TFFS users manual 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 ============================================== "Hadi" schrieb im Newsbeitrag news:IjOR7.5208$8e.322537@news... > i Bill, > > Thanks for the quick response. I will try to elabrate the problem. You are > right that I have to erase the flash segment before I could write to it. > > My flash disk is 8 meg. TrueTffs is installed to operate on the last 2 meg > (6-8). > > So the first 6 meg is not under TrueTffs. Here is the twist, only after I > format TrueTffs my function that writes to the bank 0 is successful. But if > don't format tffsRawio() fails in the Erase. > > Following is my code. I highlighted the code that fails with --->> arrow. > Your help is greatly appriciated. > > Hadi > > > > > > > > int tffsFlashRAMImagePut > > ( > > int driveNo, /* TFFS drive number */ > > int offset, /* offset in the flash chip/card */ > > int maxLoadSize, /* maximum load size */ > > char *image /* image to read from */ > > ) > > { > > PhysicalInfo info; > > UINT unitNo; > > UINT unitSize; > > UINT addr; > > char * pBuf; > > int ix, numBytes; > > int loadSize=0; > > int tmpImageSize=0; > > if (image == NULL){ > > printf("\nIMAGE==NULL"); > > return (ERROR); > > } > > if (tffsSocket[driveNo] == NULL) > > { > > printf("\ntffsSocket[%d]==NULL",driveNo); > > return (ERROR); > > } > > if (tffsRawio (driveNo, TFFS_GET_PHYSICAL_INFO, (int)&info, 0, 0) != OK) > > { > > printErr ("Unable to get physical info from drive\n"); > > return (ERROR); > > } > > pBuf = (char *)malloc (info.unitSize); > > if (pBuf == NULL) > > return ERROR; > > addr = offset; > > unitNo = offset / info.unitSize; > > unitSize = info.unitSize - (offset % info.unitSize); > > /* If the offset is not Erase Unit aligned we need to save the the > > * contents of the region begining at the start of this erase unit > > * and ending at the specified address so we can restore it after > > * we erase the Erase Unit > > */ > > if (tffsRawio (driveNo, TFFS_PHYSICAL_READ, unitNo * info.unitSize, > > offset % info.unitSize, (int)pBuf) != OK) > > { > > printErr ("Failed attempting to save Erase Unit %d\n", unitNo); > > free(pBuf); > > return (ERROR); > > } > > > --->> if (tffsRawio (driveNo, TFFS_PHYSICAL_ERASE, unitNo, 1, 0) != OK) > > --->> { > > --->> printErr ("Failed attempting to erase Erase Unit %d\n", unitNo); > > free(pBuf); > > return (ERROR); > > } > > if (tffsRawio (driveNo, TFFS_PHYSICAL_WRITE, unitNo * info.unitSize, > > offset % info.unitSize, (int)pBuf) != OK) > > { > > printErr ("Failed attempting to restore Erase Unit %d\n", unitNo); > > free(pBuf); > > return (ERROR); > > } > > > tmpImageSize = maxLoadSize; > > while (unitSize) > > { > > if (tmpImageSize >= unitSize){ > > ix = unitSize; > > } > > else if (tmpImageSize == 0){ > > ix = 0; > > break; > > } > > else { > > ix = tmpImageSize; > > } > > memcpy(pBuf, image, ix); > > image += ix; > > tmpImageSize -= ix; > > if ((addr + ix) > info.mediaSize) > > { > > printErr ("Error : Attempting to write beyond Flash boundary\n"); > > free(pBuf); > > return (ERROR); > > } > > if ((addr + ix) > (offset + maxLoadSize)) > > { > > printErr ("Error : Attempting to write beyond Load boundary\n"); > > free(pBuf); > > return (ERROR); > > } > > numBytes = ix; > > loadSize +=numBytes; > > /*printf("\nWriting @ %x num=%x",addr,numBytes);*/ > > if (tffsRawio (driveNo, TFFS_PHYSICAL_WRITE, addr, numBytes, (int)pBuf) != > OK) > > { > > printErr ("Physical write failed at address 0x%x\n", addr); > > printErr ("ByteSize = 0x%x\n", numBytes); > > printErr ("Retry with padding "); > > > if (ix & 1) /* odd address size */ > > { > > numBytes = padByte(pBuf,ix, unitSize); > > } > > if (tffsRawio (driveNo, TFFS_PHYSICAL_WRITE, addr, numBytes, (int)pBuf) != > OK) > > { > > free(pBuf); > > return (ERROR); > > } > > } > > addr += ix; > > unitSize -= ix; > > if (unitSize == 0) > > { > > unitSize = info.unitSize; > > unitNo++; > > /*printf("\nErasing unitNo %x",unitNo);*/ > > if (tffsRawio (driveNo, TFFS_PHYSICAL_ERASE, unitNo, 1, 0) != OK) > > { > > printErr ("Failed attempting to erase Erase Unit %d\n", unitNo); > > free(pBuf); > > return (ERROR); > > } > > } > > } > > /* printf("\nFromTffsDriver %x",loadSize);*/ > > free(pBuf); > > return (loadSize); > > } > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > -----Original Message----- > > From: Bill Pringlemeir [mailto:bpringlemeir@yahoo.com] > > Sent: Wednesday, December 12, 2001 11:58 AM > > Subject: Re: problem with flash that contains dosFs and bootload > > > > [bcc'd to poster] > > >>>>> "haha" == haha writes: > > haha> Hi, I am having a problem with my flash. This flash contains haha> > both bootloader and dos fileSystem. > > [snip] > > haha> Now I wanted download new BootImage bank 1 in the flash . I haha> used > the tffsRawio to write to it. But tffsRawio fails. Only haha> when I format > my flash using sysTffsFormat then I could use haha> tffsRawio But if I > format the flash I am loose my file system. > > The `tffsRawio' is probably failing because you haven't erased the > bootloader flash space. You could call an MTD routine directly to erase the > sectors before you call the `tffsRawio' to write the data. The MTD might try > to distinguish if the address it has been given is in a memory space that it > knows about. However, you should have the source to the MTD that you are > using. It is also fairly simple to write a function that will erase this > area yourself. > > hth, > > Bill Pringlemeir. > > -- > > Have you ever had a small midget tell you that Nazi Germany is your country > of birth? Or tried to psychically link with boxes of Borax while staring at > lightbulbs? You will. And the company that will bring it to you: AT&T. > > > vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" > > "Bill Pringlemeir" wrote in message > news:uy9k8e5ov.fsf@yahoo.com... > > [bcc'd to poster] > > > > >>>>> "haha" == haha writes: > > > > haha> Hi, I am having a problem with my flash. This flash contains > > haha> both bootloader and dos fileSystem. > > > > [snip] > > > > haha> Now I wanted download new BootImage bank 1 in the flash . I > > haha> used the tffsRawio to write to it. But tffsRawio fails. Only > > haha> when I format my flash using sysTffsFormat then I could use > > haha> tffsRawio But if I format the flash I am loose my file system. > > > > The `tffsRawio' is probably failing because you haven't erased the > > bootloader flash space. You could call an MTD routine directly to > > erase the sectors before you call the `tffsRawio' to write the data. > > The MTD might try to distinguish if the address it has been given is > > in a memory space that it knows about. However, you should have the > > source to the MTD that you are using. It is also fairly simple to > > write a function that will erase this area yourself. > > > > hth, > > Bill Pringlemeir. > > > > -- > > Have you ever had a small midget tell you that Nazi Germany is your > > country of birth? Or tried to psychically link with boxes of Borax > > while staring at lightbulbs? 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: setFileDateTime: how? Date: Thu, 13 Dec 2001 09:46:54 +0100 From: "Romano Signorelli" Organization: Centro Servizi Interbusiness Message-ID: <9v9ojq$icc$1@fe2.cs.interbusiness.it> How can I change the date-time attribute on an existing file with dosFs2 systems? Thanks in advance, bye. Romano Signorelli --------------------------- Newsgroups: comp.os.vxworks Subject: Re: setFileDateTime: how? Date: Thu, 13 Dec 2001 13:07:22 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <9v9ojq$icc$1@fe2.cs.interbusiness.it> Here is a quote from the "fixed problems" section of the DosFs 2.0 release supplement: SPR: 28924 ARCH: All BOARD: All HOST: All DESCRIPTION: FIOTIMESET in dosFs 2.0 does not allow proper utime behavior. The dosFs 2.0 implementation of FIOTIMESET causes utime to behave unexpectedly. FIOIMESET should allow utime to follow Solaris 2.6 behavior. If arg is NULL, the current time is used, and the behavior should be documented. WORKAROUND: This is fixed in the dosFs 2.0 for Tornado 2.0 release. So if you read this, you can either use utime() or ioctl(fd, FIOTIMESET, time_t) to change the time. - - Leonid "Romano Signorelli" wrote in message news:9v9ojq$icc$1@fe2.cs.interbusiness.it... > How can I change the date-time attribute on an existing file with dosFs2 > systems? > > Thanks in advance, bye. > Romano Signorelli > > --------------------------- Newsgroups: comp.os.vxworks Subject: WindNet PPP and compression Date: Thu, 13 Dec 2001 12:27:12 +0100 From: "Sani Rus" Message-ID: <9va3bj$e8b26$1@ID-68549.news.dfncis.de> Hello group! PPP (point to point) protocol which we acquired as a part of the WindNet do not include any data compression protocols. Can you point me to any well proven third party solutions which integrates easily with the WindNet PPP and VxWorks. Sani --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Fri Dec 14 04:03:15 2001 From: postmaster@csg.lbl.gov Date: Fri Dec 14 04:03:18 PST 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Fri Dec 14 04:03:11 PST 2001 Subject: Signal handling - Need tp know how to do? Basic example Subject: Re: problem with flash that contains dosFs and bootload Subject: Re: Adding a custom routine into a bootrom image Subject: Re: setFileDateTime: how? Subject: - ebook.htm 9002 bytes (1/1) Subject: Re: setFileDateTime: how? Subject: Re: WindNet PPP and compression Subject: Re: MTD flash driver Q: Subject: Re: Wind River PPC750 / MPC107 board PCI config help Subject: While loop data access exception Subject: Gigabit Ethernet Subject: Re: VxWorks interview question ! Subject: Re: UDP-IP performances and timing with VxWorks Subject: Re: PowerPC in Pentium system Subject: double shell parameter Subject: Re: Wind River PPC750 / MPC107 board PCI config help Subject: Application startup. Subject: Re: Gigabit Ethernet Subject: toshiba daz8811 cable modem? Subject: Re: double shell parameter Subject: Re: DOS File Corruption Subject: VxWorks AE pentium differences Subject: cannot cd to a NFS mount point Subject: bootrom can't download image from the host. Subject: How to do PASV receive with ftplib Subject: Re: bootrom can't download image from the host. ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Signal handling - Need tp know how to do? Basic example Date: Thu, 13 Dec 2001 13:16:01 +0100 From: Jacques Skubich Organization: =?iso-8859-1?Q?D=E9partement?= Informatique - INSA de Lyon Message-ID: <3C189C01.8EF74C81@insa-lyon.fr> Could somebody give me how to proceed to implement signal handling under VxWorks? Is there the equivalent SIGUSR1 or 2? I would like to get a basic C code example to help me. Take care. Jacques - -- |// (o o) ___________________________oOO__(_)__OOo________________________________ IMPORTANT ** Nouveau N° de fax ** IMPORTANT ** New fax number Jacques J. Skubich Departement d'Informatique - Blaise Pascal INSA de Lyon F-69621 VILLEURBANNE CEDEX Jacques.Skubich@insa-lyon.fr Tel/Phone : +33 (0)4 72 43 84 86 Facsimile : +33 (0)4 37 47 90 82 Mobile : +33 (0)6 99 61 25 18 ___________________________________________________________________________ "Si nous avons chacun un objet et que nous les échangeons, nous avons chacun un objet. Si nous avons chacun une idée et que nous les échangeons, nous avons chacun deux idées." Proverbe chinois. ___________________________________________________________________________ --------------------------- Newsgroups: comp.os.vxworks Subject: Re: problem with flash that contains dosFs and bootload Date: Thu, 13 Dec 2001 12:28:06 +0000 From: Graham Baxter Organization: Graham Baxter (Software) Limited Message-ID: <3C189ED6.554831CC@NOSPAM.bcs.org.uk> References: Before using physical erase/write, you first need to unmount your file system See SPR 23200 on the Wind River web site. I hope this helps. Regards, Graham Baxter Freelance Software Engineer (VxWorks and pSOS BSPs) gbaxter@NOSPAM.bcs.org.uk --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Adding a custom routine into a bootrom image Date: Thu, 13 Dec 2001 13:07:15 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: Jae-Hwa Lee wrote in message ... >Hi, everyone? > >I'm trying to add a custom routine for a ftp server into a bootrom image. >All source codes for the ftp server are in ftp.a >Some routines in ftp.a are referred to other libraries such as >libPPCEC603gnuvx.a, rtlib.a and so on. >Except libPPCEC603gnuvx.a, all libraries are custom made. >LIBS = C:/AX2200/Lib/ftp.a C:/AX2200/Lib/main.a C:/AX2200/Lib/cmd_proc.a >\ > C:/AX2200/Lib/rtlib.a C:/AX2200/Lib/cmp.a C:/AX2200/Lib/n2p.a >C:/AX2200/Lib/n2p_vp.a \ > C:/AX2200/Lib/isdn.a C:/AX2200/Lib/lapd.a C:/AX2200/Lib/dsp.a >C:/AX2200/Lib/trcp.a \ > C:/AX2200/Lib/phy.a C:/AX2200/Lib/ft1.a C:/AX2200/Lib/diag.a >C:/AX2200/Lib/vocfile.a \ > C:/AX2200/Lib/nutil.a \ > C:/AX2200/LIB/libPPCEC603gnuvx.a >C:/AX2200/Lib/main.a(cfgcmd.o): In function `Cmd_SetIpAddr_Lan': >cfgcmd.o(.text+0x5a40): undefined reference to `NV_FTP_VER_load_from_nvram' >cfgcmd.o(.text+0x6230): undefined reference to `NV_FTP_VER_save_to_nvram' >make: *** [bootrom_uncmp] Error 0x1 The order that you specify the libraries on the linker command line is important. If one of the modules in main.a needs to refer to a symbol in ftp.a then ftp.a must be after main.a on the command line. This also makes it very tricky to have libraries with circular dependencies. DaveK - -- Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! --------------------------- Newsgroups: comp.os.vxworks Subject: Re: setFileDateTime: how? Date: Thu, 13 Dec 2001 14:47:54 +0100 From: "Romano Signorelli" Organization: Centro Servizi Interbusiness Message-ID: <9vaa89$biu$1@fe2.cs.interbusiness.it> References: <9v9ojq$icc$1@fe2.cs.interbusiness.it> Note: files from the target server file system "/tgtsvr" don't have timestamp... Is it correct? Is it a workaround? I'm using "stat" function. Thanks in advance, bye. Romano Signorelli "Leonid Rosenboim" ha scritto nel messaggio news:newscache$0l5aog$ah5$1@lnews.actcom.co.il... ... > So if you read this, you can either use utime() or ioctl(fd, FIOTIMESET, > time_t) to change the time. > > - Leonid --------------------------- Newsgroups: comp.os.vxworks Subject: - ebook.htm 9002 bytes (1/1) Date: Thursday, 13 Dec 2001 07:08:02 -0600 From: no.email.address.entered@none444.yet Organization: Message-ID: <13120107.0802@none444.yet> begin 644 ebook.htm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rganization: Verio Message-ID: References: <9v9ojq$icc$1@fe2.cs.interbusiness.it> <9vaa89$biu$1@fe2.cs.interbusiness.it> Perhaps TSFS does not support file time stamp, but I do not think this is important because TSFS is supposed to be used for debugging purposes only. Leonid "Romano Signorelli" wrote in message news:9vaa89$biu$1@fe2.cs.interbusiness.it... > Note: files from the target server file system "/tgtsvr" don't have > timestamp... > Is it correct? Is it a workaround? > > I'm using "stat" function. > > Thanks in advance, bye. > Romano Signorelli > > "Leonid Rosenboim" ha scritto nel > messaggio news:newscache$0l5aog$ah5$1@lnews.actcom.co.il... > ... > > > So if you read this, you can either use utime() or ioctl(fd, FIOTIMESET, > > time_t) to change the time. > > > > - Leonid > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: WindNet PPP and compression Date: Thu, 13 Dec 2001 17:42:08 -0000 From: "David Laight" Organization: BT Internet Message-ID: <9vap58$m56$1@plutonium.btinternet.com> References: <9va3bj$e8b26$1@ID-68549.news.dfncis.de> > PPP (point to point) protocol which we acquired as a part of the WindNet do > not include any data compression protocols. Can you point me to any well > proven third party solutions which integrates easily with the WindNet PPP > and VxWorks. Do you need compression at that point? The data compression done by my 56k modem makes the benefit of downloading compressed files a lot less that one might expect.... David --------------------------- Newsgroups: comp.os.vxworks Subject: Re: MTD flash driver Q: Date: Thu, 13 Dec 2001 19:42:13 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <9v8din$un7$1@news.netmar.com> I;ll give you something to begin with: It may very well be the case that you CA NOT access any chip idividually, only all 4 chips at the same time. This is unpleasant indeed, but still a very useful configuration: If you need to read a single byte, you read all 4 bytes, and then mask and shift the bytes you need. When writing, the Flash chip needs a WRITE command and then data, so you can safely write a NOP command to all chips but the one you need modified, and then the valid data to the one you want modified, and another NOP to others. For performance purposes you will want to be able to write all 4 at once if the requested data justifies this, but then the trick is to poll for completion of all 4 chips, until all 4 have completed OK. HTH, - Leonid wrote in message news:9v8din$un7$1@news.netmar.com... > HI, > > i am writing a MTD for Intel 28F128J3A flash. we have 2 banks, and on each > bank we have 4 chips, (8M x 64bits). There are number of things i dont > understand about addressing in this case. The h/w guys tells me that i could > access all 4 chips in parallel? then how do i access each device > individually? in my case what is "noOfChips," 4 or 8 (for vxWorks, FLFlash > structure). i guess i dont' understand this "interleaving" and "socket" > concept used flash domain. > > Also, how do you address CUI (command user interface) of the Flash device? > This is 16-bit in mode. any help is very much appreciated. > > ----- Posted via NewsOne.Net: Free (anonymous) Usenet News via the eb ----- > http://newsone.net/ -- Free reading and anonymous posting to 60,000+ groups > NewsOne.Net prohibits users from posting spam. If this or other posts > made through NewsOne.Net violate posting guidelines, email abuse@newsone.net --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Wind River PPC750 / MPC107 board PCI config help Date: Thu, 13 Dec 2001 12:45:23 -0700 From: "Michael S. Simpson" Organization: Raytheon Company Message-ID: <3C190553.959E0903@west.raytheon.com> References: <3C15728E.D3D29007@west.raytheon.com> <3C17F7B1.DD9B442B@dsto.defence.gov.au> Here is some background information: We're using VisionICE to flash the board. As we have no ethernet connection to the board, we're using Wind River's "Transparent Mode Drivers" (TMD). TMD connects through the JTAG and when you create a target server, you specify the IP address of the VisionICE, not the board. That works, but is causing some problems with the cache. We're trying to configure the system without cache first, just to get it up and running, then we'll tackle the cacheing issue later. The PMC card sits on a custom carrier board which will be the only PCI device in the system. The 107 will run in host mode. I mentioned that I had a couple of questions for you and they are: 1. The training material I have on the MPC107 says that the processor and the bridge chip must run in the same "endian"-ness. The PPC normally runs in "Big Endian" and both the PPC & the MPC107 can go either way. There is no problem with handling the byte swapping issue, but the system configuration of the 107 to match the big-endian PPC is. We are unable, via the BSP, to access any registers to set the mode for the 107. Any suggestions? 2. Since the MPC107 is running in host mode, according to the documentation, we do not have access to the MPC107 Configuration Address Register space. If we try to peek into that area, we get a memory access error. We can look at the MPC107 Configuration Data Register space, no problem. How do we configure the 107 so that it will recognize the PCI memory space and how we want to map our register space on the carrier board? That a big question! Tim Shaw wrote: > Why don't you put the questions onto the newsgroup, and see if you get any > answers? > > "Michael S. Simpson" wrote: > > > I will be the first to admit that I know very little about PCI > > configuration. We're working on a project where we are using a Wind > > River build PMC board with a Motorola MPC750 CPU and a MPC107 > > Bridge/Memory Controller chip. This will be connected to a custom built > > carrier board with several ASICs and memory for video processing. There > > are no other PCI devices, just the carrier board. > > > > We're trying to configure the PMC board as the PCI host. Our biggest > > problem is trying to configure the memory map for the PCI space. Nobody > > on the development team has any experience with the MPC107 and it's a > > pain to figure out (Wind River has been some help, but hard to reach > > people who really know what's going on with the BSP). Motorola is > > usless for getting information from. > > > > If anyone has a little time to maybe answer some questions, please > > respond to this message. Thanks... > > > > Mike Simpson > > Software Engineer > > Raytheon Missile Systems > > Tucson, AZ > > (520) 784-5381 --------------------------- Newsgroups: comp.os.vxworks Subject: While loop data access exception Date: Thu, 13 Dec 2001 14:18:07 -0600 From: "Michael" Organization: Motorola Message-ID: <9vb2dv$rhc2@nntp.cig.mot.com> Reply-To: "Michael" Quick intro: class class_a { int f(const string& s1) { printf("%s \n", s1.c_str(); return 1; } }; main(void* arg) { class_a a1; string p1((char*) arg); while(1) { a1.f(p1); } } I am getting a data exception error when I run this code. data access Exception current instruction address: 0x0621e19c Machine Status Register: 0x0000b030 Data Access Register: 0x9421ffec Condition Register: 0x24000042 Data storage interrupt Register: 0x40000000 Task: 0x599f550 "pthr15" Of course I simplified class_a, but I put a printf at the start of function f, and that is not being print. What are the steps for me to debug what is going on? From what I can see, the function f is not being entered at all. The wierdest part is that function f will be executed the first time, but the second time through the while loop, it will create that exception. Another wierd part is if I do not use the while loop, but rather just call function f repeatedly, the program will work also. So, maybe the problem is somehow linked to the loop? Thanks in advance for the help. If anyone has questions, please email me at msheng1@email.mot.com. --------------------------- Newsgroups: comp.os.vxworks Subject: Gigabit Ethernet Date: Thu, 13 Dec 2001 21:10:28 GMT From: Randy Ryan Organization: Our-Kids Message-ID: <3C191B19.B199295B@pacbell.net> My apoligies if this is a duplicate message, I didn't see the original propogate to the newsgroup. Is anybody using Gigabit Ethernet yet? Am interested in what the realistic throughput rates are on particularly a PowerPC based SBC, support 64-bit @ 66MHz transfer rates in a VxWorks environment. Have not found benchmarks for PMCs as of yet, though the Sept. 2001 issues of Real Time Computing Magazine, Fibre Channel vs. Ethernet: Which for Real-Time Embedded by SBS Technologies study published showing about 36 MB/s peak rates which seems horribly low. Randy Ryan, Sr. Eng. NovaSol http://www.nova-sol.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks interview question ! Date: Thu, 13 Dec 2001 13:31:58 -0800 From: "Gary M" Message-ID: <3c191f2c$0$20175$724ebb72@reader2.ash.ops.us.uu.net> References: <23de5426.0112100602.4bf0c4ac@posting.google.com> <3c152217$0$2729$724ebb72@reader2.ash.ops.us.uu.net> Reply-To: "Gary M" "Jim" wrote in message news:JLpR7.4896$8e.314237@news... > Maybe I am wrong ... but so far I have really seen anything "Real Time" > about any of these questions ... > Fine. 1. What is an Interrupt Service Routine? What are the typical requirements for an ISR? 2. What are the advantages and disadvantages of cache memory in a real-time system that uses DMA? 3. What are the different types of timing requirements that real-time systems have? 4. What is priority inversion and how is it typically addressed? 5. You finished writing a real-time system but find that the system performs much worse than expected. What should you do? (Answers should include: hardware verification [cache, SDRAM and device timing], compiler settings, task priority settings, profiling, bad coding, etc.) 6. Discuss issues with dynamic memory allocation in real-time systems. (If they don't mention the words "memory leak" or "thread-safe", they should probably be excused.) > vxWorks is first and formost a real time OS ... if people are asking about > semaphores etc ... then you are not being interviewed for a job that > requires RTOS experience ... anybody with windows programming or even java > experience should be able to solve concurency problems ... Picking up a book and learning windows or Java programming does not teach someone how to write a *correct* multi-threaded application. (How does the saying go? Reading the Owners Manual does not teach someone how to drive a car?) > > I think OS and Concurency questions should be clearly separated from Real > Time questions which sit on top of the OS ... > Are you implying that the OS will take care of all concurrency issues? Most certainly not! Real Time questions "sit on top of the OS"? Not even close. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: UDP-IP performances and timing with VxWorks Date: Thu, 13 Dec 2001 13:48:08 -0800 From: "Gary M" Message-ID: <3c1922f7$0$20179$724ebb72@reader2.ash.ops.us.uu.net> References: <3C160B8C.482E2ABD@c-s.fr> <9v55h1$3ap$1@snoopy.bndlg.de> <3C162903.9EC18AF0@c-s.fr> Reply-To: "Gary M" We ran tests with both 855T and 860DP (50 MHz CPU and bus) using the FEC doing simple UDP loopback to a task written to the VxWorks TCP/IP stack. The 860DP pegged out when the test reached about 2500 looped 96-byte packets per second (5000 packets/sec total). Most of the processing, as expected, was in the stack. The 860DP perfomed better than the 855T due to the larger cache. BTW, we also ran the test with 80/40 Mhz CPU/Bus and got almost identical results. Running an overclocked 100/50 Mhz system gave about a 15% improvement. "christophe leroy" wrote in message news:3C162903.9EC18AF0@c-s.fr... > CPU is MPC860 at 40Mhz (Power PC) > NIC is MC68160A > > C.Leroy > > > > --------------------------- Newsgroups: comp.os.vxworks,comp.sys.powerpc.tech Subject: Re: PowerPC in Pentium system Date: Thu, 13 Dec 2001 13:56:19 -0800 From: "Gary M" Message-ID: <3c1924e2$0$20177$724ebb72@reader2.ash.ops.us.uu.net> References: <9zTR7.60827$py4.30969415@news2.nash1.tn.home.com> Reply-To: "Gary M" We're very satisfied running MPC86x family with a Tundra Semiconductor QSpan II PCI bridge. The bridge has a lot of goodies (mailboxes, I2O, DMA) that made it easy to build the communications layer on top of that. The QSpan II and the MPC860 glue together very nicely. Tundra has the usual stuff: NT driver, eval boards, VxWorks BSP "Kelly Hornsby" wrote in message news:9zTR7.60827$py4.30969415@news2.nash1.tn.home.com... > Hello All > > I'm tasked with finding out what sort of issues are involved in developing a > PMC-based PCI card that will implement a PowerPC 750/MPC107 design, and run > an OS, such as VxWorks, but all inside a Win2K Pentium machine. The OS > control is not so bad, I think, but I don't really know how to attack the > PCI Master/Slave configuration problem. What I'm not clear on is the > interaction that must occur between the BIOS, the OS, and the PCI-PMC based > PowerPC system. Is there a way to tell the PowerPC board, outside of a > memory strobe and polling loop, that a request is pending for service, or > vice versa? These are some of the things I'm thinking of at this point. > This particular application will be for a high-performance I/O controller > and simulation that I'm in the process of developing. > > I have the PCI specification, but it does not really cover what > initialization (if any) is necessary for a slave mode system that can > request bus mastering like my PowerPC board may be able to perform. > > My targeted system is a Pentium III system running Windows NT/2K. However, a > Linux OS may be in order, but I don't really care that much, in that a > device driver will have to be developed for whatever system this board will > be used in. > > If anyone has performed work like this, could you email me so that I can > find out some more about this? Otherwise, it'll be reinventing the wheel > perhaps. > > Thx for any input. > > --------------------------- Newsgroups: comp.os.vxworks Subject: double shell parameter Date: Thu, 13 Dec 2001 16:59:26 -0600 From: Dinker Charak Organization: Fermi National Accelerator Laboratory Message-ID: <3C1932CE.9B32513D@fnal.gov> Hi All, I am using VxWorks 5.4 If have this function with two double parameters, inside the function I can only access the last parameters value and that too is stored in first parameter. void DblParams(double d1, double d2) { printf("\tDouble Input 1 is f=%.8f\n", d1); printf("\tDouble Input 2 is f=%.8f\n", d2); return; } - -> DblParams 1.1,5.5 Double Input 1 is f=5.50000000 Double Input 2 is f=NaN Once I have a double in parameter list, all subsquent parameters get messed up. Why is this so? Regards, Dinker --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Wind River PPC750 / MPC107 board PCI config help Date: Fri, 14 Dec 2001 11:51:14 +1030 From: Tim Shaw Organization: Defence Science and Technology Organisation Message-ID: <3C19540A.8E4EF509@dsto.defence.gov.au> References: <3C15728E.D3D29007@west.raytheon.com> <3C17F7B1.DD9B442B@dsto.defence.gov.au> <3C190553.959E0903@west.raytheon.com> I'm sure you have, but have you created a PTE entry in the page table for the region 0xfec00000 - 0xfeefffff, for the MPC107 config addresses? This would cause a memory access error is there is no entry (and also controls the ability to cache an address range). As I understand it, the processor can access the configuration access register using configuration ports, for write only (I could be wrong). Assuming a PTE entry exists, then write to address 0xfec0 XXXX with the value (0x80000000 | ), and then write the data value to the address 0xfee00000 (assuming you're using map B). We've only used the PPC/MPC107 in big-endian mode, so I've never looked to closely at the issues, but it manuals say that the MPC107 is inherently little endian. So if you're using big endian for the PPC, then use the byte reversed load-store instructions. You will also need a PTE entry for the CPU to PCI address space before you can configure it. Then just read and write to the PCI addresses of the address map chosen (translated if using map a). To allow devices on the carrier to access the local memory, open a BAR in the PCI configuration registers. HTH Tim "Michael S. Simpson" wrote: > Here is some background information: > > We're using VisionICE to flash the board. As we have no ethernet connection > to the board, we're using Wind River's "Transparent Mode Drivers" (TMD). TMD > connects through the JTAG and when you create a target server, you specify the > IP address of the VisionICE, not the board. That works, but is causing some > problems with the cache. We're trying to configure the system without cache > first, just to get it up and running, then we'll tackle the cacheing issue > later. The PMC card sits on a custom carrier board which will be the only PCI > device in the system. The 107 will run in host mode. > > I mentioned that I had a couple of questions for you and they are: > > 1. The training material I have on the MPC107 says that the processor and the > bridge chip must run in the same "endian"-ness. The PPC normally runs in "Big > Endian" and both the PPC & the MPC107 can go either way. There is no problem > with handling the byte swapping issue, but the system configuration of the 107 > to match the big-endian PPC is. We are unable, via the BSP, to access any > registers to set the mode for the 107. Any suggestions? > > 2. Since the MPC107 is running in host mode, according to the documentation, > we do not have access to the MPC107 Configuration Address Register space. If > we try to peek into that area, we get a memory access error. We can look at > the MPC107 Configuration Data Register space, no problem. How do we configure > the 107 so that it will recognize the PCI memory space and how we want to map > our register space on the carrier board? That a big question! > > Tim Shaw wrote: > > > Why don't you put the questions onto the newsgroup, and see if you get any > > answers? > > > > "Michael S. Simpson" wrote: > > > > > I will be the first to admit that I know very little about PCI > > > configuration. We're working on a project where we are using a Wind > > > River build PMC board with a Motorola MPC750 CPU and a MPC107 > > > Bridge/Memory Controller chip. This will be connected to a custom built > > > carrier board with several ASICs and memory for video processing. There > > > are no other PCI devices, just the carrier board. > > > > > > We're trying to configure the PMC board as the PCI host. Our biggest > > > problem is trying to configure the memory map for the PCI space. Nobody > > > on the development team has any experience with the MPC107 and it's a > > > pain to figure out (Wind River has been some help, but hard to reach > > > people who really know what's going on with the BSP). Motorola is > > > usless for getting information from. > > > > > > If anyone has a little time to maybe answer some questions, please > > > respond to this message. Thanks... > > > > > > Mike Simpson > > > Software Engineer > > > Raytheon Missile Systems > > > Tucson, AZ > > > (520) 784-5381 --------------------------- Newsgroups: comp.os.vxworks Subject: Application startup. Date: 14 Dec 2001 01:53:46 GMT From: akus00@aol.com (AKUS00) Organization: AOL http://www.aol.com Message-ID: <20011213205346.08259.00000062@mb-cu.aol.com> We have recently completed developing our application and trying to get it onto the target and start automatically when booted. 1.I have compiled the application as archive and produced my application.a file. 2.I have copied the archive file into image project directory 3.I have edited the usrAppInit.c to call my application at startup. When I compile the image, it tells me that there is a symbol it cannot find. And this single missing symbol generates several errors on consecutive 5 lines in the symbol file. When I search for those symbols in my project directory (which by the way looks like compiler generated symbol names) I don't see them in my code but those names exists in some of my object files. I have 4 different application and only this application creates this error. I can not even create the image file for this application. This application works just fine when I compile application.out, download and spawn it from the host Tornado PC running on a win2000 PC. Where should I be looking for the reason. Any information will be greatly appreciated. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Gigabit Ethernet Date: Fri, 14 Dec 2001 02:06:04 -0500 From: Yves Boudreault Organization: Bell Sympatico Message-ID: <3C19A4DC.789779F@sympatico.bogus.ca> References: <3C191B19.B199295B@pacbell.net> Randy Ryan wrote: > > My apoligies if this is a duplicate message, I didn't see the original > propogate to the newsgroup. > > Is anybody using Gigabit Ethernet yet? > > Am interested in what the realistic throughput rates are on particularly a > PowerPC based SBC, support 64-bit @ 66MHz transfer rates in a VxWorks > environment. > > Have not found benchmarks for PMCs as of yet, though the Sept. 2001 issues > of Real Time Computing Magazine, Fibre Channel vs. Ethernet: Which for > Real-Time Embedded by SBS Technologies study published showing about 36 MB/s > peak rates which seems horribly low. > > Randy Ryan, Sr. Eng. > NovaSol > http://www.nova-sol.com I am using Race++ for some communications issues and we found that the performance is highly dependant on the "PCI bridge" being used. On MPC106, we were struggling to get 65MB/s. A card with MPC107 (same vendor), 110MB/sec ++. Yves --------------------------- Newsgroups: comp.os.vxworks Subject: toshiba daz8811 cable modem? Date: Fri, 14 Dec 2001 07:38:57 GMT From: "integer-x" Organization: Excite@Home - The Leader in Broadband http://home.com/faster Message-ID: my question is does anyone know what the default password for vxworks telnet into the toshiba daz8811 cable modem? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: double shell parameter Date: Fri, 14 Dec 2001 09:10:10 +0100 From: "Michael Lawnick" Organization: Buergernetz Dillingen Message-ID: <9vcc56$58q$1@snoopy.bndlg.de> References: <3C1932CE.9B32513D@fnal.gov> Reply-To: "Michael Lawnick" Sender: mlawnick@pd9524f26.dip.t-dialin.net Hi Dinker, try to access your function from within another function and I guess it will work: The Shell only knows integer params. Even strings a resolved to int. A double is typically 64bit. So your first double wil get the contents of the two first params concatenated, second double will be initialized with 0. HTH - -- Mit freundlichen Grüßen, Michael Lawnick ============================================== SOFTEC GmbH Tel +49-731-96600-0 Promenade 17 Fax +49-731-96600-23 D-89073 Ulm Michael Lawnick Germany lawnick@softec.de ============================================== "Dinker Charak" schrieb im Newsbeitrag news:3C1932CE.9B32513D@fnal.gov... > Hi All, > > I am using VxWorks 5.4 > > If have this function with two double parameters, inside the function I > can only access the last parameters value and that too is stored in > first parameter. > > void DblParams(double d1, double d2) > > printf("\tDouble Input 1 is f=%.8f\n", d1); > printf("\tDouble Input 2 is f=%.8f\n", d2); > return; > } > > -> DblParams 1.1,5.5 > Double Input 1 is f=5.50000000 > Double Input 2 is f=NaN > > Once I have a double in parameter list, all subsquent parameters get > messed up. Why is this so? > > Regards, > Dinker --------------------------- Newsgroups: comp.os.vxworks Subject: Re: DOS File Corruption Date: Fri, 14 Dec 2001 09:29:22 +0100 From: "Hermann" Organization: Swisscom IP+ (post doesn't reflect views of Swisscom) Message-ID: <9vcd7r$ese$1@rex.ip-plus.net> References: <3c0fdc4b@news.tce.com> The problem is that most of the PCMCIA flash cards have a damaged file system when you power off the card during a write access (or a short time afterwards). We made tests with about 20 different types of flash cards, but at the moment the only type of PCMCIA flash card that works with VxWorks is from PRETEC (www.pretec.com) "Fred" schrieb im Newsbeitrag news:3c0fdc4b@news.tce.com... > We're using PCMCIA flash cards to store our downloaded vxWorks, but have > recently had problems with file corruption. Sometimes the file system is > corrupted, sometimes the file is corrupted. Occassionally we can reproduce > the problem by booting the board during a FTP transfer. Does anybody have > any ideas as to what is causing the problem and how to fix it? > > --------------------------- Newsgroups: comp.os.vxworks Subject: VxWorks AE pentium differences Date: Fri, 14 Dec 2001 09:50:44 +0100 From: "Mukul Dhankhar" Message-ID: <3c19bc1f$0$7113$4d4efb8e@news.be.uu.net> Hi all, I want to use VxWorks AE 1.1 for a new project.My target hardware is Pentium-MMX 200 MHz. Does anyone know if I can use VxWorks Pentium 3 with a PcPentium 3 BSP on my P1 target? Does VxWorks for P-3 make some specialized functionalities for P-3 which it does not make for P-1? Thanks, Mukul --------------------------- Newsgroups: comp.os.vxworks Subject: cannot cd to a NFS mount point Date: Fri, 14 Dec 2001 10:57:17 MET From: mathar@mpia-hd.mpg.de (Richard Mathar) Organization: Max-Planck-Institut fuer Astronomie Keywords: nfsMount, cd Message-ID: <2001Dec14.105717@sun18> Reply-To: mathar@mpia.de With VxWorks 5.4 and its windsh, a nfsMount succeeds without error, and actually shows the new mount point with nfsDevShow, but a "cd" into this new file does not work : "No such file or directory" Mounting a point with the same name (by omission of the optional "local" name with nfsMount) succeeds and allows the "cd" to this newly generated local directory. VxWorks user ID is the same on the Solaris host computer and the VxWorks/Tornado computer. What could make the difference? The rwx permissions on the Solaris remote host ought have been set correctly with sufficient degree of allowance... mathar@mpia.de --------------------------- Newsgroups: comp.os.vxworks Subject: bootrom can't download image from the host. Date: Fri, 14 Dec 2001 18:41:22 +0800 From: "xmtan" Organization: Bentium Ltd. (CN99) Message-ID: <9vcl6a$mhd$1@mail.cn99.com> Reply-To: "xmtan" I make and END ethernet driver. I use 'ping' to test my driver, and it works very well. When I build it into bootrom, and try to use it to download my image from the host, it fails. At the same time, I can still ping the host. Is anything wrong with my driver? I need help ? Thanks a lot. --------------------------- Newsgroups: comp.os.vxworks Subject: How to do PASV receive with ftplib Date: Fri, 14 Dec 2001 11:00:51 -0000 From: "JP" Message-ID: Anyone done this or have any pointers on how to read server port response. Is it simply : char cmdbuf[1024]; char recbuf[1024]; sprintf(cmdbuf, "PASV"); if (write(conn->ctrlsock, cmdbuf, strlen(cmdbuf)) != strlen(cmdbuf)) return ERROR; if (read(conn->ctrlsock, recbuf, 1024) <= 0) return ERROR; Any help gratefully received. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: bootrom can't download image from the host. Date: Fri, 14 Dec 2001 06:09:06 -0500 From: George Varndell Organization: Posted Via Binaries.net = SPEED+RETENTION+COMPLETION = http://www.binaries.net Message-ID: <3C19DDD2.BF2F40A0@varndellengineering.com> References: <9vcl6a$mhd$1@mail.cn99.com> xmtan wrote: > > I make and END ethernet driver. I use 'ping' to test my driver, and it works > very well. When I build it into bootrom, and try to use it to download my > image from the host, it fails. At the same time, I can still ping the host. > Is anything wrong with my driver? Probably. Try pinging the target with big packets. Try blasting the target with as many pings as you can. If that works, your driver may be okay and your problems could stem from configuration problems. What ethernet chipset are you using? I may have a driver for you. Regards, George Varndell - -- Varndell Engineering, LLC. PPC and XScale Expertise Coming soon! Download free demo versions of our END drivers! http://www.varndellengineering.com ______________________________________________________________________________ Posted Via Binaries.net = SPEED+RETENTION+COMPLETION = http://www.binaries.net --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Sat Dec 15 04:03:10 2001 From: Vxworks Exploder Date: Sat Dec 15 04:03:12 PST 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sat Dec 15 04:03:07 PST 2001 Subject: cheap and small serial terminal Subject: Re: How to do PASV receive with ftplib Subject: serial debugging with Tornado 2.0 fails Subject: Re: cannot cd to a NFS mount point Subject: Re: Mastering A32 space from an MVME167 running VxWorks Subject: Re: WindNet PPP and compression Subject: Re: tNetTask stack overrun Subject: Re: `Higher reliability' DosFs. Subject: Re: Is there a routeGet? Subject: Re: tNetTask stack overrun Subject: Re: `Higher reliability' DosFs. ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: cheap and small serial terminal Date: Fri, 14 Dec 2001 12:54:22 +0100 From: Matthias Bodenstein Organization: Matztech Corp. Message-ID: <3C19E86E.834C75B7@gmx.de> Hi, in our institute we are often using a laptop as a serial terminal to configure a cpu or several device cards. Now I am searching for a smaller and cheaper device to do the job (e.g. in the size of a PDA), because using a laptop/notebook for this, is like using a Boing 767 to transport 2 passengers. Does anyone know which devices are suitable to be used as serial terminal, or better has experiences with it. If yes I would be very happy, if you could provide me some information about it. Thanks in advance Matthias Bodenstein - -- - ------------------------------------------------------------------------ Matthias Bodenstein WWW: http://matztias.de Tel.: +49 178 484 1508 E-Mail: Matztias@gmx.de --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to do PASV receive with ftplib Date: Fri, 14 Dec 2001 17:27:33 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: The response contains 6 bytes printed in decimals, separated with commas. The fist four bytes belong to the IP address, and the last two to the port number, all in network byte order. Read the relevant RFCs if you are implementing anything related to FTP. Goto http://www.ietf.org - - Leonid "JP" wrote in message news:PZkS7.4717$AY5.54542@NewsReader... > Anyone done this or have any pointers on how to > read server port response. > Is it simply : > > char cmdbuf[1024]; > char recbuf[1024]; > > sprintf(cmdbuf, "PASV"); > if (write(conn->ctrlsock, cmdbuf, strlen(cmdbuf)) != strlen(cmdbuf)) return > ERROR; > if (read(conn->ctrlsock, recbuf, 1024) <= 0) return ERROR; > > Any help gratefully received. > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: serial debugging with Tornado 2.0 fails Date: Fri, 14 Dec 2001 16:57:22 +0100 From: Markus Pietrek Organization: FS Forth Systeme GmbH Message-ID: <9vd7fl$cl3$07$1@news.t-online.com> Hi folks, I'm sure it's an old problem but I can't find any answers on google. I currently try to debug a ARM processor with Tornado 2.0 with the serial WDB backend. Target and host are configured for 9600/8/1 and their ports are working. The target prints "WDB Ready" after booting, the shell displays "WDB: Ready" in the banner, too. Echo mode is working, too. And the host is sending packets over serial which I can't read as they are binary. But the target server can't connect. it always prints "Error: rpccore backend client Timed out". This happens both on the first or second serial line. The lines configuring the WDB are " #undef WDB_COMM_TYPE #define WDB_COMM_TYPE WDB_COMM_SERIAL #undef WDB_TTY_CHANNEL #define WDB_TTY_CHANNEL 1 #undef WDB_TTY_BAUD #define WDB_TTY_BAUD 9600 " Network debugging is fine but for my problem I can't use it. Has anyone solved this problem yet? Thanks in advance, Markus Pietrek --------------------------- Newsgroups: comp.os.vxworks Subject: Re: cannot cd to a NFS mount point Date: Fri, 14 Dec 2001 18:20:58 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <2001Dec14.105717@sun18> I beleive you are too confused with WindSH using Host's I/O domain. In other words, if you do -> cd "/somedir" if would happen on the HOST, but your NFS mount is on the target, so here is what you need to do: -> @cd "/somedir" -> @ll The "@" prefix means to WindSh to run target functions as opposed to built-in functions of the same names. HTH, - Leonid "Richard Mathar" wrote in message news:2001Dec14.105717@sun18... > > With VxWorks 5.4 and its windsh, a nfsMount succeeds without error, > and actually shows the new mount point with nfsDevShow, but a "cd" into > this new file does not work : "No such file or directory" Mounting a point > with the same name (by omission of the optional "local" name with nfsMount) > succeeds and allows the "cd" to this newly generated local directory. > VxWorks user ID is the same on the Solaris host computer and the VxWorks/Tornado > computer. > What could make the difference? The rwx permissions on the Solaris remote > host ought have been set correctly with sufficient degree of allowance... > > mathar@mpia.de --------------------------- Newsgroups: comp.dsp,comp.os.vxworks Subject: Re: Mastering A32 space from an MVME167 running VxWorks Date: 14 Dec 2001 18:58:59 GMT From: Samuel Igwe Organization: Concentric Internet Services Message-ID: <3C1A4BE2.BE80D0E8@pentek.com> References: <1008340718.974247@news> > _The_ Paladin wrote: > > wondering if anyone has some code examples outlining the steps > involved in Mastering A32 space from an MVME167 running VxWorks > 5.x > > That or some pointers to freely available documents, I might be > able to grok such information on. > > Much thanks A word of advice -- this sort of question is more appropriate in comp.os.vxworks --------------------------- Newsgroups: comp.os.vxworks Subject: Re: WindNet PPP and compression Date: 14 Dec 2001 20:31:05 GMT From: jefftkauff@aol.com (JeffTKauff) Organization: AOL http://www.aol.com Message-ID: <20011214153105.28833.00000429@mb-fc.aol.com> References: <9va3bj$e8b26$1@ID-68549.news.dfncis.de> Try www.hifn.com. They have a software based compression library. You can get ccp from the public domain pppd package. Jeff --------------------------- Newsgroups: comp.os.vxworks Subject: Re: tNetTask stack overrun Date: 14 Dec 2001 20:38:24 GMT From: jefftkauff@aol.com (JeffTKauff) Organization: AOL http://www.aol.com Message-ID: <20011214153824.28833.00000430@mb-fc.aol.com> References: tNetTask uses a fixed size ringbuffer to store the netJobAdd() requests. The ringbuffer can get overun during heavy activity (ie, when you receive lots of etherent packets). I believe it is somewhere around 85 entries, and is defined in netlib.c. You will probably need to source code access to increase the size. Jeff --------------------------- Newsgroups: comp.os.vxworks Subject: Re: `Higher reliability' DosFs. Date: 14 Dec 2001 20:56:49 -0500 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: References: Sender: bpringlemeir@DeadDuck Thanks for the information Leonid. We have found that a close/fclose actually calls the DosFs dcache with a `CBIO_CACHE_FLUSH' and has set the ioctl arg to zero as such, /* flush cache when last opened file is closed */ ... cbioIoctl(pVolDesc->pCbio, CBIO_CACHE_FLUSH, 0); The CBIO in turn does this, /* wakeup dcache task */ taskUndelay( dacheTask ); /* loan dcache current priority */ taskPriorityGet( 0, &callerPri ); taskPrioritySet( dacheTask, callerPri ); Which raise the priority of the caching task, but it does not run immediately. taskUndelay is undocumented, but the caller holds the cache semaphore, so there is no way that a close flushes the file to the flash immediately; at least that is not what I see. The rename mechanism will work only if the dcache task dumps the sectors in the order in which they were written. This might be the case unless another task references the data and the LRU cache is re-arrange. regards, Bill Pringlemeir. Leonid> For the benefit of the newsgroup here is my recommended Leonid> solution: DosFs 2.0 was designed with reliability in mind - Leonid> to replace some data safely, you need to write new data in a Leonid> new file, in the same directory as the old file. Then close Leonid> or flush the new file, which wil flush related disk cache Leonid> blocks, including data and FAT, and finally rename() the new Leonid> file to be named like the old one. Leonid> rename() is specifically implemented such that as long as its Leonid> within the same directory, either the old file will survive Leonid> or the new one if the system crashes during the update at any Leonid> point of time. There will never be both remaining or none. Leonid> Note that rename() does not trigger any FAT changes if it Leonid> happens within the same directory, only the directory is Leonid> being modified. On top of this, if the file names are short Leonid> to fit 8.3 names, and the directory is just a few entires, Leonid> then the entire directory is only one disk sector, which Leonid> makes the rename operation really atomic. - -- Your spouse will think of you as someone you said hello to thrice at a social reception. Months from now not answering the phone may seem mysterious. Do not eat watermelon this week. Discuss life with RMS with the help of God. You'll fall into a state of sadness, but eventually be a Borg. vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Is there a routeGet? Date: Fri, 14 Dec 2001 23:16:27 -0500 From: "Dan S." Message-ID: <9vej3a$n6m$1@bob.news.rcn.net> References: If memory serves me correctly, there is no routeGet call. How I got around this was to roll my own by reading the vx mibs. Email me if you need more specifics, don't have the code with me here... --------------------------- Newsgroups: comp.os.vxworks Subject: Re: tNetTask stack overrun Date: Sat, 15 Dec 2001 11:11:57 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <20011214153824.28833.00000430@mb-fc.aol.com> Jeff, if your CPU is incapable of handling the packet traffic, no matter how large the ring buffer is, it will get overfown eventually. When the CPU speed is in balance with the network speed, the actual utilization of this ring never goes over 8 or 10 slots. If your CPU is slow, one way to avoid overflows is to define the number of Rx buffers to be equal or smaller then this number of entries, so in situation of bursts the CPU wont be able to handle, received packets will be dropped at the MAC level, consuming minimal resources. "JeffTKauff" wrote in message news:20011214153824.28833.00000430@mb-fc.aol.com... > tNetTask uses a fixed size ringbuffer to store the netJobAdd() requests. The > ringbuffer can get overun during heavy activity (ie, when you receive lots of > etherent packets). I believe it is somewhere around 85 entries, and is defined > in netlib.c. You will probably need to source code access to increase the > size. > > Jeff --------------------------- Newsgroups: comp.os.vxworks Subject: Re: `Higher reliability' DosFs. Date: Sat, 15 Dec 2001 11:22:54 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: I am not sure if I am at liberty to discuss the source code at such a deep level on this public forum, but here is what I feel I can tell: The closure of last openbed file causes the dcache updater task to become READY, it does not run immediately, but it will run sooner then any of the application routines will have a chanse to call the rename() function, and this is the order of operations we need to impose, make sure that the new file data and related FAT entries have been flushed, BEFORE the actual directory entry has been modified and the old file deleted. Regarding rename itself, if the file names dealt with are short 8.3 convention, then rename requires modification of exactly one disk sector, as the name change happens "in place", hence there is no order of flushing to be concerned with. If the system crashes before the single sector caused by rename has been flushed, the old and new files will both be present. If the system crashes just after the directory sector flush but before the FAT entries related to the old file have been flushed, the "chkdsk" part of the file system will finish up the old file, and preserve the new one. So if you are looking that deep into the product, the consistency checking mechanism is part of the atomic rename idea and you should look there as well. Of course I agree this whole thing is rather tricky, so if you can spare some extra disk space, I would recommend to avoid the rename() and add some sort of checksum at the end of each important file, and during usage, open the latest, check it, if it falls through use it, if not fall back to the previous version, and so forth. And remember, shit happens. - - Leonid "Bill Pringlemeir" wrote in message news:uellx1c0u.fsf@yahoo.com... > > Thanks for the information Leonid. We have found that a close/fclose > actually calls the DosFs dcache with a `CBIO_CACHE_FLUSH' and has set > the ioctl arg to zero as such, > > /* flush cache when last opened file is closed */ > ... > cbioIoctl(pVolDesc->pCbio, CBIO_CACHE_FLUSH, 0); > > The CBIO in turn does this, > > /* wakeup dcache task */ > taskUndelay( dacheTask ); > > /* loan dcache current priority */ > taskPriorityGet( 0, &callerPri ); > taskPrioritySet( dacheTask, callerPri ); > > Which raise the priority of the caching task, but it does not run > immediately. taskUndelay is undocumented, but the caller holds the > cache semaphore, so there is no way that a close flushes the file to > the flash immediately; at least that is not what I see. > > The rename mechanism will work only if the dcache task dumps the > sectors in the order in which they were written. This might be the > case unless another task references the data and the LRU cache is > re-arrange. > > regards, > Bill Pringlemeir. > > Leonid> For the benefit of the newsgroup here is my recommended > Leonid> solution: DosFs 2.0 was designed with reliability in mind - > Leonid> to replace some data safely, you need to write new data in a > Leonid> new file, in the same directory as the old file. Then close > Leonid> or flush the new file, which wil flush related disk cache > Leonid> blocks, including data and FAT, and finally rename() the new > Leonid> file to be named like the old one. > > Leonid> rename() is specifically implemented such that as long as its > Leonid> within the same directory, either the old file will survive > Leonid> or the new one if the system crashes during the update at any > Leonid> point of time. There will never be both remaining or none. > > Leonid> Note that rename() does not trigger any FAT changes if it > Leonid> happens within the same directory, only the directory is > Leonid> being modified. On top of this, if the file names are short > Leonid> to fit 8.3 names, and the directory is just a few entires, > Leonid> then the entire directory is only one disk sector, which > Leonid> makes the rename operation really atomic. > > -- > Your spouse will think of you as someone you said hello to thrice at a > social reception. Months from now not answering the phone may seem > mysterious. Do not eat watermelon this week. Discuss life with RMS > with the help of God. You'll fall into a state of sadness, but > eventually be a Borg. > > vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Mon Dec 17 04:03:10 2001 From: Vxworks Exploder Date: Mon Dec 17 04:03:13 PST 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Mon Dec 17 04:03:06 PST 2001 Subject: DosFs 2.0 Product Information Subject: Re: VxWorks interview question ! Subject: Re: VxWorks interview question ! Subject: Re: VxWorks interview question ! Subject: Re: VxWorks interview question ! Subject: Re: cheap and small serial terminal Subject: Re: While loop data access exception Subject: Re: Please Help Me Subject: Re: VxWorks interview question ! Subject: Re: VxWorks interview question ! Subject: Re: Reverse etherAddrResolve Subject: Re: Please Help Me Subject: SetRequest & Multiple Variable Bindings ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: DosFs 2.0 Product Information Date: Sun, 16 Dec 2001 17:17:44 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: I tried to find the technical specifications of DosFs 2.0 on the WindRiver Web site to show someone, and to my surprise, I found no mention of it. This seems shocking to me that this optional component is not advertised, and I remain puzzled as to the reasons why. Surely I don't expect any official answer from WindRiver (although that might be nice to chat again with my old pals), but I would appreciate if anyone can shed some light on this enigma. Of course this is mere curiosity. - - Leonid begin 666 Leonid Rosenboim.vcf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rganization: Magma Communications Ltd. Message-ID: References: <23de5426.0112100602.4bf0c4ac@posting.google.com> <3c152217$0$2729$724ebb72@reader2.ash.ops.us.uu.net> <3c191f2c$0$20175$724ebb72@reader2.ash.ops.us.uu.net> Much better ;-) Thanks I will go write these down now ... even you last comment about what I implied would be a good question ;-) What do you think about the following question ... "What is the difference between the three types of semaphores in vxWorks 'mutual-exclusion','binary', and 'counting'?" This question would screw me nicely ... because of the first semaphore ... but not now hopefully ... > Are you implying that the OS will take care of all concurrency issues? Most certainly not! I did ... but I know better ... as an example I offer DMA and a thread using a pointer that is not "volatile" ... I am sure there is more ... > Real Time questions "sit on top of the OS"? Not even close. Sorry ... me and my big mouth ;-) For your question ... . What are the different types of timing requirements that real-time systems have? What type of answer are you looking for? My best guess here would be 1.) signaling to devices (ie being able to look at a logic analyser and find the problem with interdevice communication), 2.) issues with data throught put (ie minimizing bus activity and processor useage to move data through a system), 3.)service times for various threads/processes and activities ... would these suffice or do you think typical employeers should look for "key words" on these answers ... What about asking canidates about osciloscopes, logic analysers, and spectrun analysers ... is this fair ... I am a software guy ... but I have found these tools invaluable ... is this fair to assume that knowledge of these tools are essential to exhibit ones experience in the field ... Thank you ... I look forward to your response ... Jim "Gary M" wrote in message news:3c191f2c$0$20175$724ebb72@reader2.ash.ops.us.uu.net... > "Jim" wrote in message news:JLpR7.4896$8e.314237@news... > > Maybe I am wrong ... but so far I have really seen anything "Real Time" > > about any of these questions ... > > > Fine. > 1. What is an Interrupt Service Routine? What are the typical > requirements for an ISR? > 2. What are the advantages and disadvantages of cache memory in a > real-time system that uses DMA? > 3. What are the different types of timing requirements that real-time > systems have? > 4. What is priority inversion and how is it typically addressed? > 5. You finished writing a real-time system but find that the system > performs much worse than expected. What should you do? (Answers should > include: hardware verification [cache, SDRAM and device timing], compiler > settings, task priority settings, profiling, bad coding, etc.) > 6. Discuss issues with dynamic memory allocation in real-time systems. > (If they don't mention the words "memory leak" or "thread-safe", they should > probably be excused.) > > > vxWorks is first and formost a real time OS ... if people are asking about > > semaphores etc ... then you are not being interviewed for a job that > > requires RTOS experience ... anybody with windows programming or even java > > experience should be able to solve concurency problems ... > > Picking up a book and learning windows or Java programming does not teach > someone how to write a *correct* multi-threaded application. (How does the > saying go? Reading the Owners Manual does not teach someone how to drive a > car?) > > > > I think OS and Concurency questions should be clearly separated from Real > > Time questions which sit on top of the OS ... > > > Are you implying that the OS will take care of all concurrency issues? Most > certainly not! > Real Time questions "sit on top of the OS"? Not even close. > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks interview question ! Date: Sun, 16 Dec 2001 10:43:44 -0800 From: "Michael R. Kesti" Organization: MK Associates Message-ID: <3C1CEB60.45E23A05@gv.net> References: <23de5426.0112100602.4bf0c4ac@posting.google.com> <3c152217$0$2729$724ebb72@reader2.ash.ops.us.uu.net> <3c191f2c$0$20175$724ebb72@reader2.ash.ops.us.uu.net> Reply-To: mkesti@gv.net Jim wrote: >What do you think about the following question ... > >"What is the difference between the three types of semaphores in vxWorks >'mutual-exclusion','binary', and 'counting'?" Such questions make me feel that my memory is being tested, rather than my talent. The immediate response I would want to make, but might avoid depending on how much I want the job, is that I am willing and able to refer to the documentation when I need to know. After all, the definition of a technology guru is "one who is both willing and able to read a manual", isn't it? - -- ======================================================================== 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: VxWorks interview question ! Date: Sun, 16 Dec 2001 21:31:24 GMT From: "Jim" Organization: Magma Communications Ltd. Message-ID: References: <23de5426.0112100602.4bf0c4ac@posting.google.com> <3c152217$0$2729$724ebb72@reader2.ash.ops.us.uu.net> <3c191f2c$0$20175$724ebb72@reader2.ash.ops.us.uu.net> <3C1CEB60.45E23A05@gv.net> What questions would you like in place of this? In my opinion a good answer to this could cover alot ... I have no problems handing someone a manual during an interview ;-) Is an interview with specific questions necessarily bad? Jim "Michael R. Kesti" wrote in message news:3C1CEB60.45E23A05@gv.net... > Jim wrote: > > >What do you think about the following question ... > > > >"What is the difference between the three types of semaphores in vxWorks > >'mutual-exclusion','binary', and 'counting'?" > > Such questions make me feel that my memory is being tested, rather than > my talent. The immediate response I would want to make, but might avoid > depending on how much I want the job, is that I am willing and able to > refer to the documentation when I need to know. > > After all, the definition of a technology guru is "one who is both willing > and able to read a manual", isn't it? > > -- > ======================================================================== > Michael Kesti | "And like, one and one don't make > | two, one and one make one." > mkesti@gv.net | - The Who, Bargain --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks interview question ! Date: Sun, 16 Dec 2001 22:03:02 +0000 (UTC) From: "David Laight" Organization: BT Openworld Message-ID: <9vj5ml$abm$1@knossos.btinternet.com> References: <23de5426.0112100602.4bf0c4ac@posting.google.com> <3c152217$0$2729$724ebb72@reader2.ash.ops.us.uu.net> <3c191f2c$0$20175$724ebb72@reader2.ash.ops.us.uu.net> <3C1CEB60.45E23A05@gv.net> > After all, the definition of a technology guru is "one who is both willing > and able to read a manual", isn't it? Then: 1) try the code to see how it works 2) search for the source... 3) disassemble it 4) patch the object code 5) raise a bug 6) ..... --------------------------- Newsgroups: comp.os.vxworks Subject: Re: cheap and small serial terminal Date: Mon, 17 Dec 2001 00:58:44 GMT From: Ephraim Gadsby Organization: - minimal Message-ID: References: <3C19E86E.834C75B7@gmx.de> On Fri, 14 Dec 2001 12:54:22 +0100, Matthias Bodenstein wrote: >Does anyone know which devices are suitable to be used as serial >terminal, or better has experiences with it. If yes I would be very >happy, if you could provide me some information about it. > I've not done it myself but I've known people use palm pilots for this. You need an older or low end model as they are now going over to usb. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: While loop data access exception Date: Mon, 17 Dec 2001 01:03:45 GMT From: Ephraim Gadsby Organization: - minimal Message-ID: <88hq1u4gfbgs5v8h76bhimj17a87fjl79l@4ax.com> References: <9vb2dv$rhc2@nntp.cig.mot.com> On Thu, 13 Dec 2001 14:18:07 -0600, "Michael" wrote: >Quick intro: > >class class_a { > int f(const string& s1) { > printf("%s \n", s1.c_str(); > return 1; > } >}; > >main(void* arg) { > class_a a1; > string p1((char*) arg); > while(1) { > a1.f(p1); > } >} > >I am getting a data exception error when I run this code. > >data access >Exception current instruction address: 0x0621e19c >Machine Status Register: 0x0000b030 >Data Access Register: 0x9421ffec >Condition Register: 0x24000042 >Data storage interrupt Register: 0x40000000 >Task: 0x599f550 "pthr15" > >Of course I simplified class_a, but I put a printf at the start of function >f, and that is not being print. What are the steps for me to debug what is >going on? From what I can see, the function f is not being entered at all. >The wierdest part is that function f will be executed the first time, but >the second time through the while loop, it will create that exception. >Another wierd part is if I do not use the while loop, but rather just call >function f repeatedly, the program will work also. So, maybe the problem is >somehow linked to the loop? Vxworks code should not include functions called main. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Please Help Me Date: Mon, 17 Dec 2001 01:24:49 GMT From: Ephraim Gadsby Organization: - minimal Message-ID: <0req1uo07ne2p4lvtjmnlvtqbrdifi7arr@4ax.com> References: <6c32a22b.0112120223.1cfc1bb2@posting.google.com> "DigiPsyche" wrote in message news:6c32a22b.0112120223.1cfc1bb2@posting.google.com... > Hi, > I am an internet application developer and working on java, jsp, > perl, php, javascript technologies and very new to this group. i am > very much intrested in learning and making my future in embedded > technologies. i have very less knowledge about all the enbedded > technologies. Please help me, which technologies i should know to go > for 3G wireless communication and VOIP/Data communication and from > where to startup? Is there any website to download the free version of > these softwares? > > thanks > DigiPsyche On Wed, 12 Dec 2001 19:03:55 +0200, "Leonid Rosenboim" wrote: > >Digi, Dear, > >Your competencies are a far cry from what a decent embedded >engineer needs to have. but if you have the guts for embedded, >then here is a partial sillabus that can get you there: > >+ A basic course in hardware including Logic and processor architectures >+ You must know at least 2 assembly languages, preferrably one > classic CISC (e.g. PDP11, MC68K) and one RISC (Sparc or MIPS ) >+ You need to be proficient in "C", including a few hundered thousands > lines of code you wrote and debugged on a non-embedded platform >+ You may already know communications to some extent, but in the > Embedded world, it's all a matter of bits and bytes, so you need to get > low and dirty on those as well > >If you aim specifically at the 3G cellphone development, then a good level >mathematics, specifically in Signal Processing is a big addvantage. > I must say that the above description is very misleading. Most people that work on embedded systems in the data/telecomms industry have considerably less expertise. You only have to look at this newsgroup to see that there are plenty of people who are working with VxWork who have little idea about RTOSs, even at the application level. Even on truly embedded projects, there is usually more need for people that know about protocols and general programming than there is for detailed low-level embedded knowledge. A lot of people work their way in through testing (where perl is useful) and maintaintence. And, in the past it was easy to move around, due to skill shortages. I should add that this is one of the worst time in the history of telecommunications to try to get into the field. I suggust you lookup some names like Nortel, Alcatel and Marconi on Google. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks interview question ! Date: Sun, 16 Dec 2001 18:44:27 -0800 From: "Michael R. Kesti" Organization: MK Associates Message-ID: <3C1D5C0B.A4C1FF79@gv.net> References: <23de5426.0112100602.4bf0c4ac@posting.google.com> <3c152217$0$2729$724ebb72@reader2.ash.ops.us.uu.net> <3c191f2c$0$20175$724ebb72@reader2.ash.ops.us.uu.net> <3C1CEB60.45E23A05@gv.net> Reply-To: mkesti@gv.net Jim wrote: >What questions would you like in place of this? I think that questions concerning design and documentation philosphies and methodologies are far more relevant to finding good programmers. The trouble is that it is difficult to develop tests that rate candidates in these areas. So, just as companies often account for costs that are easy to track while ignoring those more difficult, they tend to quiz candidates on the topics that are easy to test. Accurate representations of the situations are not generated in either case. >In my opinion a good answer to this could cover alot ... I have no problems >handing someone a manual during an interview ;-) That's cool, but I suspect that few engineering and HR managers would agree. >Is an interview with specific questions necessarily bad? I suppose that it depends on the candidates and what you want to learn about them. It seems far more appropriate to quiz a recent graduate looking for his first job about the subtlties of a programming language than to ask such questions of a senior engineer with dozens of completed projects behind him. - -- ======================================================================== 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: VxWorks interview question ! Date: Sun, 16 Dec 2001 22:29:15 -0500 From: jimdon@sympatico.ca Organization: Bell Sympatico Message-ID: <3C1D668B.922DF719@sympatico.ca> References: <23de5426.0112100602.4bf0c4ac@posting.google.com> <3c152217$0$2729$724ebb72@reader2.ash.ops.us.uu.net> <3c191f2c$0$20175$724ebb72@reader2.ash.ops.us.uu.net> <3C1CEB60.45E23A05@gv.net> <3C1D5C0B.A4C1FF79@gv.net> Agreed ... but I have a couple comments ... 1.) "I think that questions concerning design and documentation philosphies ..." - --> I have seen a lot of canidate employee's give some really gushy answers that only HR people and managers love ... clear and simple 'all talk and no action' ... how do you avoid that? Yes design methodologies are becoming more and more well definined such as "The Unified Process", "XP", etc ... but too many people out there know "the right thing to say" but have no realization of there application to "actual work" ... (Note: maybe I am just venting ;-)) 2.) "That's cool, but I suspect that few engineering and HR managers would agree." - --> Engineering and HR managers get this reasoning from studies right ... I wish one of them would sit me down and explain it. ;-) Or am I dreaming ;-) 3.) "I suppose that it depends on the candidates and what you want to learn about them." - --> What can I say ... I agree ... I suppose being "Specific yet general" is the bottom line of what you want to achieve. Jim "Michael R. Kesti" wrote: > Jim wrote: > > >What questions would you like in place of this? > > I think that questions concerning design and documentation philosphies and > methodologies are far more relevant to finding good programmers. The trouble > is that it is difficult to develop tests that rate candidates in these areas. > So, just as companies often account for costs that are easy to track while > ignoring those more difficult, they tend to quiz candidates on the topics > that are easy to test. Accurate representations of the situations are not > generated in either case. > > >In my opinion a good answer to this could cover alot ... I have no problems > >handing someone a manual during an interview ;-) > > That's cool, but I suspect that few engineering and HR managers would agree. > > >Is an interview with specific questions necessarily bad? > > I suppose that it depends on the candidates and what you want to learn about > them. It seems far more appropriate to quiz a recent graduate looking for > his first job about the subtlties of a programming language than to ask such > questions of a senior engineer with dozens of completed projects behind him. > > -- > ======================================================================== > 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: Reverse etherAddrResolve Date: Mon, 17 Dec 2001 05:03:44 GMT From: Sperry Family Message-ID: <3C1D5F71.2B9C9520@gte.net> References: Johan Borkhuis wrote: > > Hello, > > In one system I have the MAC address of a device on the network, and I want > to get the IP address of this device. Is this possible? > etherAddrResolve does the opposite, finding the MAC-address of a known IP- > address, but what I am looking for is just the other way around.. > > 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 === Several folks here have suggested that you use RARP, but that's wrong. RARP is for when you know YOUR OWN MAC address and want some server to tell you YOUR OWN IP address. RARP is an early form of BOOTP (RFC 0903 at http://www.rfc-editor.org). Inverse ARP or InARP (RFC 2390 at http://www.rfc-editor.org) is intended to be used to find the IP address of some other host given that you know its hardware address. It was intended to be used with Frame Relay, where you know the DLCI of the far-end and want to find its IP address (DLCIs only have local significance, so you don't know what the far-end uses as its hardware address, but that's probably more than you want to know). InARP could be extended to other types of hardware, but I don't know that it has been. All this aside, I suspect Johan is working with ethernet and wants to know of a vxWorks API that will give him the IP address of a host on an ethernet segment, given that he knows the MAC address of that host. This I don't know how to do off the top of my head. If the host has already been accessed via IP, that information must already be in the ARP cache. Other than that, idunno. Vic Sperry --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Please Help Me Date: Mon, 17 Dec 2001 10:23:20 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <6c32a22b.0112120223.1cfc1bb2@posting.google.com> <0req1uo07ne2p4lvtjmnlvtqbrdifi7arr@4ax.com> "Ephraim Gadsby" wrote in message news:0req1uo07ne2p4lvtjmnlvtqbrdifi7arr@4ax.com... [snip] > I must say that the above description is very misleading. Most > people that work on embedded systems in the data/telecomms industry > have considerably less expertise. You only have to look at this > newsgroup to see that there are plenty of people who are working with > VxWork who have little idea about RTOSs, even at the application > level. > Indeed, how very true and sad. These past few years of bubble have swept i large amount of insufficiently competent engineers into this field, and voila, all of the sudden I have to "reset" my Ericcson T28S cell phone due to severe memory leaks. Who'd expect that to happen only three years ago ? Anyway, my "recipie" was for the ultimate results, but since it is not mandatory, and if the original poster has got balls, he will go ahead and figure out the requirements one by one. Still, I would not support or encourage incompetence in this field. > Even on truly embedded projects, there is usually more need for people > that know about protocols and general programming than there is for > detailed low-level embedded knowledge. A lot of people work their > way in through testing (where perl is useful) and maintaintence. And, > in the past it was easy to move around, due to skill shortages. > True, and now look at the results. By the way, in testing Perl knowledge is barely enough, you got to know some stuff about the things you are testing, actually Perl is the easiest part of it all. > I should add that this is one of the worst time in the history of > telecommunications to try to get into the field. I suggust you lookup > some names like Nortel, Alcatel and Marconi on Google. > On the contrary, this is the BEST time for people who seek to excell to get into this field, and when finally the embedded field resumes dominance, there will be good money to be made. Meanwhile, the low average level of competance is an excellent opportunity for people with balls to get into our field. If they develop good level of competence, it will pay off. Without competence, one would become the next good candidate for a subsequent round of layoffs. - - Leonid --------------------------- Newsgroups: comp.protocols.snmp,comp.os.vxworks,comp.os.psos Subject: SetRequest & Multiple Variable Bindings Date: Mon, 17 Dec 2001 12:58:49 +0200 From: Alex Vinokur Message-ID: <3C1DCFE8.957216F9@bigfoot.com> Followup-To: comp.protocols.snmp ============================ pRISM+ for pSOSystem/PowerPC pRISM+ 1.2.3 . pSOSystem 2.2.2 Diab Data D-CC Optimizing C compiler SNMP Release 7.4 (Protocol SNMPv1) Epilogue Technology Emissary SNMP MIB Compiler, version 6.3 Host : Windows98 Target : PowerPC (MPC860) MG-SOFT MIB Browser, Professional Edition 7.0.0.3730 ============================ Hi, I have the following problem with Multiple Variable Bindings SetRequest while using the mib file below. - -- ---------- Fragment of mib file : BEGIN ---------- control OBJECT-IDENTITY STATUS current DESCRIPTION "" ::= { top 1 } data OBJECT-IDENTITY STATUS current DESCRIPTION "Leaves under this node data contain device's data" ::= { top 2 } control-action OBJECT-TYPE SYNTAX INTEGER { stand-by(1), preliminary-settings(2), apply-all-settings(3) } MAX-ACCESS read-write STATUS current DESCRIPTION " Control Actions : ---------------------------------- 1 - Stand-By (can't be set by user) 2 - Preliminary Settings (can't be set by user) ---------------------------------- 3 - Apply All Settings (can be set by user) ====================== Updating device's data ====================== Device items that can be updated are in MIB-tree under the node 'data' Updating consists of two stages : 1. Preliminary setting values for items that are to be updated. User should send SetRequest for relevant leaves in mib : - top.data.xxx - top.data.yyy - top.data.zzz ----------------------------------------------------------- Condition#1. if (top.control.control-action == 'apply-all-settings') { ERROR will be returned; } else { set top.control.control-action = preliminary-settings; } ----------------------------------------------------------- IMPORTANT. After this stage the items in device itself are not updated !!! SetRequest's values are placed in special buffer !!! ----------------------------------------------------------- 2. Applying the settings. User should send SetRequest with value 'apply-all-settings' for the leaf - top.control.control-action ----------------------------------------------------------- Condition#2. if (top.control.control-action != 'preliminary-settings') { ERROR will be returned; } else { Relevant items (leaves) under top.data will be updated in device; } After the device ends the updating (it takes some time) a top.control.control-action will be set to 'stand-by' (by SNMP Agent itself, not by SNMP Manager's user). " ::= { control 1 } xxx OBJECT-TYPE SYNTAX INTEGER MAX-ACCESS read-write STATUS current DESCRIPTION "" ::= { data 1 } yyy OBJECT-TYPE SYNTAX INTEGER MAX-ACCESS read-write STATUS current DESCRIPTION "" ::= { data 2 } zzz OBJECT-TYPE SYNTAX INTEGER MAX-ACCESS read-write STATUS current DESCRIPTION "" ::= { data 3 } - -- ---------- Fragment of mib file : END ------------ =================== Situation#1 : BEGIN =================== - ----------------------------------------------------------- Sending simple (not Multiple Variable Bindings) SetRequests. - ----------------------------------------------------------- SetRequest#1 (data) before SetRequest : control-action == 'stand-by' set xxx = 123 (via SNMP Browser) after SetRequest : SNMP Agent sets control-action = 'preliminary-settings' SetRequest#2 (data) before SetRequest : control-action == 'preliminary-settings' set yyy = 456 (via SNMP Browser) after SetRequest : control-action == 'preliminary-settings' SetRequest#3 (data) before SetRequest : control-action == 'preliminary-settings' set zzz = 789 (via SNMP Browser) after SetRequest : control-action == 'preliminary-settings' ---> Note. On this stage items xxx, yyy, zzz in device are NOT updated. SetRequest#4 (control) before SetRequest : control-action == 'preliminary-settings' set control-action = 'apply-all-settings' (via SNMP Browser) after SetRequest : control-action == 'apply-all-settings' .......................................................... after the device ends the updating (it takes some time) SNMP Agent sets control-action = 'stand-by' ---> Note. On this stage items xxx, yyy, zzz in device ARE updated. Conclusion. In situation#1 (simple SetRequests) SNMP Agent works OK. =================== Situation#1 : END ===================== =================== Situation#2 : BEGIN =================== - ----------------------------------------------------------- Sending Multiple Variable Bindings SetRequest. - ----------------------------------------------------------- Multiple SetRequest (data & control) before SetRequest : control-action == 'stand-by' --- via SNMP Browser --- set xxx = 123 set yyy = 456 set zzz = 789 set control-action = 'apply-all-settings' (!!! Here SetRequest fails) ------------------------ ---> Note. SetRequest fails because control-action == 'stand-by' before setting control-action = 'apply-all-settings'. Conclusion. In situation#2 (SetRequest with Multiple Variable Bindings) SNMP Agent fails. =================== Situation#2 : END ===================== We can see that failure of 'Multiple Variable Bindings SetRequest' is caused by Condition#2. Why has Condition#2 been used? I would like to prevent setting control-action = 'apply-all-settings' if (control-action == 'stand-by') (because there are nothing to update). - ----------------------------------------- Is such behavior caused by wrong design? Is it possible to reach what I what to ? Note. I can try not to use Condition#2. However, I would rather do use it. Thanks in advance, =========================== Alex Vinokur mailto:alexvn@bigfoot.com mailto:alexvn@dr.com http://up.to/alexvn http://go.to/alexv_math =========================== --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Mon Dec 17 04:07:31 2001 From: ulf.fiederling@dornier.eads.net Date: Mon Dec 17 04:07:33 PST 2001 Subject: analysis of a pending task Hi all out there We are using tornado II / WindView with a pentium target and have the following problem. One of our tasks goes to state "pending" after some time (statistical, may take several hours). How can we analyse what the task is waiting for or does anyone have an idea whats the reason for this behavior. Best regards Ulf Fiederling From vxwexplo-errs@csg.lbl.gov Mon Dec 17 09:55:00 2001 From: Donald McLachlan Date: Mon Dec 17 09:55:03 PST 2001 Subject: serial driver for SeaLevel ACB 5102 I'm sure I'll have to modify it, but does anyone have a vxworks serial driver for the Sealevel ACB-MP.PCI (5102)? Thanks, Don From vxwexplo-errs@csg.lbl.gov Tue Dec 18 04:03:13 2001 From: Vxworks Exploder Date: Tue Dec 18 04:03:15 PST 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Tue Dec 18 04:03:10 PST 2001 Subject: Re: Reverse etherAddrResolve Subject: Re: DosFs 2.0 Product Information Subject: Re: DosFs 2.0 Product Information Subject: i/o redirection Subject: checksum verification - embedded Subject: Re: serial debugging with Tornado 2.0 fails Subject: Re: serial debugging with Tornado 2.0 fails Subject: Re: i/o redirection Subject: Re: checksum verification - embedded Subject: Re: checksum verification - embedded Subject: Re: checksum verification - embedded Subject: Re: cheap and small serial terminal Subject: How to enumerate the network interfaces in a vxwork env and get their name? Subject: Re: checksum verification - embedded Subject: Re: i/o redirection Subject: Re: i/o redirection Subject: Re: serial debugging with Tornado 2.0 fails Subject: Re: i/o redirection Subject: Re: i/o redirection Subject: Re: i/o redirection Subject: Re: How to enumerate the network interfaces in a vxwork env and get ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Reverse etherAddrResolve Date: Mon, 17 Dec 2001 13:30:59 +0000 (UTC) From: don@mars.dgrc.crc.ca (Donald McLachlan) Organization: The unconfigured xvnews people Message-ID: <9vks2j$37o$1@crc-news.crc.ca> References: <3C1D5F71.2B9C9520@gte.net> Reply-To: don@mars.dgrc.crc.ca I was thinking of something that would work on a switched lan segment and though of programming the MAC address you want into your nic card, sending a packet (so the switch now thinks you have that address and routes packets for that host to your host), but this sounds way too slimey ... but just might work. Nicer might be to ping the lan broadcast address, and watch for a reply from that mac address. ... I suppose the device not reply to pings, so you could send a arp request for each address on your subnet, or ... - -- Donald McLachlan E-mail Donald.McLachlan@crc.ca Communications Research Centre / RNS Tel (613) 998-2845 3701 Carling Ave., Fax (613) 998-9648 Ottawa, Ontario K2H 8S2 Canada --------------------------- Newsgroups: comp.os.vxworks Subject: Re: DosFs 2.0 Product Information Date: Mon, 17 Dec 2001 14:34:46 GMT From: "Don Dewar" Message-ID: References: I was also shocked to find this out. I called my FAE who said I had to pay for it. When I pointed out I heard it was free with a support contract he relented and sent me the binaries. They charge an outrageous amount of money for the sources. The integration went ok, but we suddenly found some of our files disappearing or getting corrupt. We tried to debug it for months with very little support from WRS. Finally, we got the sources and still could not figure it out. When we compiled with less than 2 MB support all our problems all vanished. Anyway they don't seem to support it very well. Good luck. - -- Don "Leonid Rosenboim" wrote in message news:newscache$w61gog$hug$1@lnews.actcom.co.il... > I tried to find the technical specifications of DosFs 2.0 on the WindRiver > Web site to show someone, and to my surprise, I found no mention of it. > This seems shocking to me that this optional component is not advertised, > and I remain puzzled as to the reasons why. > > Surely I don't expect any official answer from WindRiver (although that > might > be nice to chat again with my old pals), but I would appreciate if anyone > can > shed some light on this enigma. > > Of course this is mere curiosity. > - Leonid > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: DosFs 2.0 Product Information Date: Mon, 17 Dec 2001 16:57:22 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: "Don Dewar" wrote in message news:aonT7.93$5f4.38153@dfiatx1-snr1.gtei.net... > [snip] Finally, we got the sources and still could not > figure it out. When we compiled with less than 2 MB support all our > problems all vanished. [snip] Would you care to elaborate on this ? Maybe we can figure out the cause of trouble and help other poor souls who may be stepping into the same sh*t ? --------------------------- Newsgroups: comp.os.vxworks Subject: i/o redirection Date: Mon, 17 Dec 2001 17:23:14 +0200 From: "Ofer Goren" Organization: Verio Message-ID: Hi. How can I redirect my I/O, so my own task will handle it, rather the shell task? Thanks, - -- Ofer Goren Iamba Technologies --------------------------- Newsgroups: comp.os.vxworks Subject: checksum verification - embedded Date: Mon, 17 Dec 2001 10:58:21 -0500 From: Gilad Katz Organization: Mailgate.ORG Server - http://www.Mailgate.ORG Message-ID: <3C1E161D.A3E58D5F@avoidspam.com> Sender: bpatel@fwion.com hello, One of our application has the need to transfer files( vxworks image, application executable) from a master controller card to other cards in the system. The files on the master card are copied from a ata drive to local ramdisk, then checksum verification is done on it. This implementation was done by a developer who has left our project. The question I have are - 1) Is there a need to do checksum verification when transmitting data from (ata drive - flash disk based, one medium) to local ramdisk (another medium)? I feel this is too much of an overhead. The checksum verification is done because one is transmitting files from one medium(ata drive - flash disk) to another medium (ramdisk - SDRAM based). I would prefer to transmit files directly from ata drive's directory to the remote slave card using NFS. thanks. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: serial debugging with Tornado 2.0 fails Date: Mon, 17 Dec 2001 09:12:46 -0700 From: Brian Miller Organization: Raytheon Company Message-ID: <3C1E197E.7DEC0F06@west.raytheon.com> References: <9vd7fl$cl3$07$1@news.t-online.com> Markus: Go back into the configuration of your target server and increase the timeout (in seconds) and the number of retrys. That should help. Also, the 9600 baud rate is only for debugging putposes. We're using a serial interface at 115KBPS and it works fine. Mike Simpson Raytheon Speaking only for myself. Markus Pietrek wrote: > > Hi folks, > > I'm sure it's an old problem but I can't find any answers on google. I > currently try to debug a ARM processor with Tornado 2.0 with the serial WDB > backend. Target and host are configured for 9600/8/1 and their ports are > working. The target prints "WDB Ready" after booting, the shell displays > "WDB: Ready" in the banner, too. Echo mode is working, too. And the host is > sending packets over serial which I can't read as they are binary. But the > target server can't connect. it always prints "Error: rpccore backend > client Timed out". This happens both on the first or second serial line. > > The lines configuring the WDB are > " > #undef WDB_COMM_TYPE > #define WDB_COMM_TYPE WDB_COMM_SERIAL > > #undef WDB_TTY_CHANNEL > #define WDB_TTY_CHANNEL 1 > #undef WDB_TTY_BAUD > #define WDB_TTY_BAUD 9600 > " > > Network debugging is fine but for my problem I can't use it. > Has anyone solved this problem yet? > > Thanks in advance, > Markus Pietrek --------------------------- Newsgroups: comp.os.vxworks Subject: Re: serial debugging with Tornado 2.0 fails Date: Mon, 17 Dec 2001 17:41:44 +0100 From: Markus Pietrek Organization: FS Forth Systeme GmbH Message-ID: <9vl76j$a7u$04$1@news.t-online.com> References: <9vd7fl$cl3$07$1@news.t-online.com> <3C1E197E.7DEC0F06@west.raytheon.com> Brian Miller wrote: > Markus: > > Go back into the configuration of your target server and increase the > timeout (in seconds) and the number of retrys. That should help. Also, > the 9600 baud rate is only for debugging putposes. We're using a serial > interface at 115KBPS and it works fine. Thanks. But I had already set it to at least 2 seconds timeout and one retry and one time I tried it with values >10. No success. Bye, Markus --------------------------- Newsgroups: comp.os.vxworks Subject: Re: i/o redirection Date: Mon, 17 Dec 2001 09:46:23 -0800 From: "Dale" Message-ID: <3c1e3017_1@news.foxinternet.com> References: Reply-To: "Dale" Hello, Try looking at the functions: ioGlobalStdSet() ioTaskStdSet() Hopefully, these can assist you. There is also the WindShell variable SH_GET_TASK_IO to help with redirection of output from the WindShell. hth, Dale "Ofer Goren" wrote in message news:newscache$x1whog$8aj$1@lnews.actcom.co.il... > Hi. > How can I redirect my I/O, so my own task will handle it, rather the shell > task? > > Thanks, > > -- > Ofer Goren > Iamba Technologies > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: checksum verification - embedded Date: Mon, 17 Dec 2001 10:13:20 -0800 From: "Michael R. Kesti" Organization: MK Associates Message-ID: <3C1E35C0.5914408A@gv.net> References: <3C1E161D.A3E58D5F@avoidspam.com> Reply-To: mkesti@gv.net Gilad Katz wrote: >One of our application has the need to transfer files( vxworks image, >application executable) from a master controller card to other cards in >the system. The files on the master card are copied from a ata drive to >local ramdisk, then checksum verification is done on it. This >implementation was done by a developer who has left our project. The >question I have are - > >1) Is there a need to do checksum verification when transmitting data >from (ata drive - flash disk based, one medium) to local ramdisk >(another medium)? I feel this is too much of an overhead. The checksum >verification is done because one is transmitting files from one >medium(ata drive - flash disk) to another medium (ramdisk - SDRAM >based). I would prefer to transmit files directly from ata drive's >directory to the remote slave card using NFS. Errors can often occur in unanticipated ways, and it's hard to fault designers for incorporating error checking, especially if the "costs" aren't overwhelming. Whether such checking is warranted in your situation depends on the magnitude of those costs whith respect to your system's resources, and the consequences of errors that might go undetected. unless you are willing to reveal all of your system's design and can find someone willing to analyze that design, only you can judge these issues. - -- ======================================================================== 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: checksum verification - embedded Date: Mon, 17 Dec 2001 12:07:29 -0800 From: "Gary M" Message-ID: <3c1e5063$0$12345$4c41069e@reader1.ash.ops.us.uu.net> References: <3C1E161D.A3E58D5F@avoidspam.com> Reply-To: "Gary M" If the hardware bus(es) over which the data passes do(es) not have built-in data integrity checking (usually byte parity), and this checking is not turned on, enabled, and supported by the software, you will not have 100% confidence that the data was passed error-free from one device to another. (Even byte parity doesn't give 100% confidence, but it is far superior than nothing at all). Just something else to consider.... "Gilad Katz" wrote in message news:3C1E161D.A3E58D5F@avoidspam.com... > hello, > > One of our application has the need to transfer files( vxworks image, > application executable) from a master controller card to other cards in > the system. The files on the master card are copied from a ata drive to > local ramdisk, then checksum verification is done on it. This > implementation was done by a developer who has left our project. The > question I have are - > > 1) Is there a need to do checksum verification when transmitting data > from (ata drive - flash disk based, one medium) to local ramdisk > (another medium)? I feel this is too much of an overhead. The checksum > verification is done because one is transmitting files from one > medium(ata drive - flash disk) to another medium (ramdisk - SDRAM > based). I would prefer to transmit files directly from ata drive's > directory to the remote slave card using NFS. > > thanks. > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: checksum verification - embedded Date: Mon, 17 Dec 2001 12:52:15 -0800 From: "Michael R. Kesti" Organization: MK Associates Message-ID: <3C1E5AFF.CA341550@gv.net> References: <3C1E161D.A3E58D5F@avoidspam.com> <3c1e5063$0$12345$4c41069e@reader1.ash.ops.us.uu.net> Reply-To: mkesti@gv.net Gary M wrote: >If the hardware bus(es) over which the data passes do(es) not have built-in >data integrity checking (usually byte parity), and this checking is not >turned on, enabled, and supported by the software, you will not have 100% >confidence that the data was passed error-free from one device to another. Agreed. >(Even byte parity doesn't give 100% confidence, but it is far superior than >nothing at all). Agreed again, but do note that checksumming suffers from the same short- comming, and for much the same reason. Just as there is a non-zero number of bit combinations that generate the same parity, there is a number of data sets that generate the same checksum. - -- ======================================================================== 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: cheap and small serial terminal Date: Mon, 17 Dec 2001 15:48:54 -0800 From: Graham Waters Organization: TRIUMF Message-ID: <3C1E8466.E1967B2@triumf.ca> References: <3C19E86E.834C75B7@gmx.de> Reply-To: Graham.Waters@triumf.ca Matthias Bodenstein wrote: > > Does anyone know which devices are suitable to be used as serial > terminal, or better has experiences with it. If yes I would be very > happy, if you could provide me some information about it. > > Thanks in advance > > Matthias Bodenstein > You might like to consider using an Ethernet Serial Server. They allow you to "telnet" to the serial port of your device. Have a look at the following: http://www.atoptechnologies.com/gw21s.htm http://www.bb-elec.com/product.asp?sku=SS1&dept_id=99 +---------------------------------------------------------------------+ |Graham Waters, Control System Eng phone: (604)-222-1047 | |TRIUMF --- University of B.C, Fax: (604)-222-7307 | |Vancouver, BC, Canada e-mail: Graham.Waters@triumf.ca | | | |http://isacwserv.triumf.ca/homepage/graham01.html | +---------------------------------------------------------------------+ For every expert there is an equal and opposite expert - Arthur C Clark --------------------------- Newsgroups: comp.os.vxworks Subject: How to enumerate the network interfaces in a vxwork env and get their name? Date: Tue, 18 Dec 2001 09:35:45 +0800 From: "zzxt" Organization: Bentium Ltd. (CN99) Message-ID: <9vm6fv$1eo$1@mail.cn99.com> Hi all: The ifShow routine shows all the net interfaces in the system including their names mac address ip addresss etc. Can i get the interface name in the system? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: checksum verification - embedded Date: Mon, 17 Dec 2001 17:15:34 -0500 From: "Yin, Yong" Organization: Nortel Message-ID: <9vlqok$sa5$1@bcarh8ab.ca.nortel.com> References: <3C1E161D.A3E58D5F@avoidspam.com> Based on your description, I think the original intent is to detect the potential damage of this image. As the image is stored in a flush disk, it's possible you read a damaged image and it's fatal to start this image. So it's better to use checksum to protect this if you have CPU time. Regards Yong Gilad Katz wrote in message news:3C1E161D.A3E58D5F@avoidspam.com... > hello, > > One of our application has the need to transfer files( vxworks image, > application executable) from a master controller card to other cards in > the system. The files on the master card are copied from a ata drive to > local ramdisk, then checksum verification is done on it. This > implementation was done by a developer who has left our project. The > question I have are - > > 1) Is there a need to do checksum verification when transmitting data > from (ata drive - flash disk based, one medium) to local ramdisk > (another medium)? I feel this is too much of an overhead. The checksum > verification is done because one is transmitting files from one > medium(ata drive - flash disk) to another medium (ramdisk - SDRAM > based). I would prefer to transmit files directly from ata drive's > directory to the remote slave card using NFS. > > thanks. > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: i/o redirection Date: Tue, 18 Dec 2001 09:41:29 +0200 From: "Ofer Goren" Organization: Verio Message-ID: References: <3c1e3017_1@news.foxinternet.com> I dont think they will help me. All they do is re-setting the stdin/out/err to something else then the default..... - -- Ofer Goren Iamba Technologies "Dale" wrote in message news:3c1e3017_1@news.foxinternet.com... > Hello, > > Try looking at the functions: > > ioGlobalStdSet() > ioTaskStdSet() > > Hopefully, these can assist you. There is also the WindShell variable > SH_GET_TASK_IO > to help with redirection of output from the WindShell. > > hth, > Dale > "Ofer Goren" wrote in message > news:newscache$x1whog$8aj$1@lnews.actcom.co.il... > > Hi. > > How can I redirect my I/O, so my own task will handle it, rather the shell > > task? > > > > Thanks, > > > > -- > > Ofer Goren > > Iamba Technologies > > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: i/o redirection Date: Tue, 18 Dec 2001 10:30:39 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: Ofer, if you want a useful answer you need to spend a little more thought and effort when writing the question. What exactly do you need ? "Ofer Goren" wrote in message news:newscache$x1whog$8aj$1@lnews.actcom.co.il... > Hi. > How can I redirect my I/O, so my own task will handle it, rather the shell > task? > > Thanks, > > -- > Ofer Goren > Iamba Technologies > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: serial debugging with Tornado 2.0 fails Date: 18 Dec 2001 09:43:15 +0200 From: Reinhold Wirth Organization: WEB.DE Message-ID: <3c1f01a8$1@netnews.web.de> References: <9vd7fl$cl3$07$1@news.t-online.com> Hi Markus, * test #define WDB_TTY_DEV_NAME "/tyCo/1" * test your setting of CONSOLE_TTY. It is not possible to use both with the same device. Markus Pietrek wrote: >Hi folks, > >I'm sure it's an old problem but I can't find any answers on google. I >currently try to debug a ARM processor with Tornado 2.0 with the serial WDB >backend. Target and host are configured for 9600/8/1 and their ports are >working. The target prints "WDB Ready" after booting, the shell displays >"WDB: Ready" in the banner, too. Echo mode is working, too. And the host is >sending packets over serial which I can't read as they are binary. But the >target server can't connect. it always prints "Error: rpccore backend >client Timed out". This happens both on the first or second serial line. > >The lines configuring the WDB are >" >#undef WDB_COMM_TYPE >#define WDB_COMM_TYPE WDB_COMM_SERIAL > > >#undef WDB_TTY_CHANNEL >#define WDB_TTY_CHANNEL 1 >#undef WDB_TTY_BAUD >#define WDB_TTY_BAUD 9600 >" > >Network debugging is fine but for my problem I can't use it. >Has anyone solved this problem yet? > >Thanks in advance, >Markus Pietrek - -- __________________________________________________________ News suchen, lesen, schreiben mit http://newsgroups.web.de --------------------------- Newsgroups: comp.os.vxworks Subject: Re: i/o redirection Date: Tue, 18 Dec 2001 12:30:21 +0200 From: "Ofer Goren" Organization: Verio Message-ID: References: I have a CLI task. I want all input and output to be re-directed to this task, instead of to the shell task. in fact, I can kill the shell task, but I want the user to be able to access the shell upon a specific command. Today, when I spawn my cliTask, I get a username prompt. However, when entering a username, I get an error of invalid command from the shell task, since this is the task that captures my input. So: I want my cliTask to capture all I/O.... How do I do that? Thanks, - -- Ofer Goren Iamba Technologies "Leonid Rosenboim" wrote in message news:newscache$zp7jog$z7l$1@lnews.actcom.co.il... > Ofer, > if you want a useful answer you need to spend a little more thought and > effort when writing the question. What exactly do you need ? > > "Ofer Goren" wrote in message > news:newscache$x1whog$8aj$1@lnews.actcom.co.il... > > Hi. > > How can I redirect my I/O, so my own task will handle it, rather the shell > > task? > > > > Thanks, > > > > -- > > Ofer Goren > > Iamba Technologies > > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: i/o redirection Date: 18 Dec 2001 11:35:27 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: "Ofer Goren" wrote: > I have a CLI task. I want all input and output to be re-directed to > this task, instead of to the shell task. in fact, I can kill the shell > task, but I want the user to be able to access the shell upon a > specific command. The only way to do this is to kill the shell. All commands can also be executed from your CLI, but you have to lookup the symbol and call the function from within the CLI. You can also startup the shell when the user wants access to the 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: i/o redirection Date: Tue, 18 Dec 2001 12:31:55 +0100 From: "Werner Schiendl" Message-ID: <3c1f29f4$1@brateggebdc5.br-automation.co.at> References: Hi Ofer, do I understand you correctly, you want the shell in your system? How about creating a command that is your CLI (a normal C function, that can be called from shell to run your CLI). Do start that command synchronously in the target shell, not spawn it as a task. Then you can run your commands, as long as your command does not return your are in command of the stdin, etc. To return to the shell, you can simply exit the command and continue. hth Werner "Ofer Goren" wrote in message news:newscache$o6djog$wcl$1@lnews.actcom.co.il... > I have a CLI task. I want all input and output to be re-directed to this > task, instead of to the shell task. in fact, I can kill the shell task, but > I want the user to be able to access the shell upon a specific command. > > Today, when I spawn my cliTask, I get a username prompt. However, when > entering a username, I get an error of invalid command from the shell task, > since this is the task that captures my input. > > So: I want my cliTask to capture all I/O.... How do I do that? > > Thanks, > > -- > Ofer Goren > Iamba Technologies > "Leonid Rosenboim" wrote in message > news:newscache$zp7jog$z7l$1@lnews.actcom.co.il... > > Ofer, > > if you want a useful answer you need to spend a little more thought and > > effort when writing the question. What exactly do you need ? > > > > "Ofer Goren" wrote in message > > news:newscache$x1whog$8aj$1@lnews.actcom.co.il... > > > Hi. > > > How can I redirect my I/O, so my own task will handle it, rather the > shell > > > task? > > > > > > Thanks, > > > > > > -- > > > Ofer Goren > > > Iamba Technologies > > > > > > > > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to enumerate the network interfaces in a vxwork env and get their name? Date: Tue, 18 Dec 2001 12:01:09 +0000 From: Graham Baxter Organization: Graham Baxter (Software) Limited Message-ID: <3C1F3005.5C8B0E34@NOSPAM.bcs.org.uk> References: <9vm6fv$1eo$1@mail.cn99.com> Try m2IfTblEntryGet() Regards, Graham Baxter (VxWorks and pSOS BSP's) Freelance Software Engineer - AVAILABLE FOR NEW ASSIGNMENT SOON gbaxter@NOSPAM.bcs.org.uk zzxt wrote: > > Hi all: > The ifShow routine shows all the net interfaces in the system including > their names mac address ip addresss etc. Can i get the interface name in the > system? - -- --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Tue Dec 18 06:01:48 2001 From: "Boda, Hitesh T (MED)" Date: Tue Dec 18 06:01:50 PST 2001 Subject: How to download project on target Hi All Two project are present in my workspace. One is Bootable project (Clean Vxworks project). One is Downloadable project. Previously I am using Ethernet option to download my project using ftp server). At that time First I download My Bootable image(Vxworks) and then I download my Downloadable project (xxx.out). Currently I used JTAG option(Vision click) to download My image using Vision click. I am successfully convert my Vxworks image in Vxworks.ab and Vxworks.bdx file (vision click file format) and download on target board. When I try to convert my xxx.out file in .ab and .bdx file vision click software give following error. Error #81: ELF format error, non executable file. Can anyone explain me what is problem and how to solved it?? -- Hitesh Boda "THIS E-MAIL MESSAGE ALONG WITH ANY ATTACHMENTS IS INTENDED ONLY FOR THE ADDRESSEE and may contain confidential and privileged information. If the reader of this message is not the intended recipient, you are notified that any dissemination, distribution or copy of this communication is strictly Prohibited. If you have received this message by error, please notify us immediately, return the original mail to the sender and delete the message from your system." From vxwexplo-errs@csg.lbl.gov Wed Dec 19 04:03:13 2001 From: Vxworks Exploder Date: Wed Dec 19 04:03:16 PST 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Wed Dec 19 04:03:09 PST 2001 Subject: Booting VxWorks AE 1.1 Subject: Re: i/o redirection Subject: Re: i/o redirection Subject: Re: checksum verification - embedded Subject: Re: checksum verification - embedded Subject: telnet banner Subject: Re: Send message to the target from Window's application Subject: Re: telnet banner Subject: Re: i/o redirection Subject: Re: i/o redirection Subject: TFFS Format Subject: Re: TFFS Format Subject: Re: `Higher reliability' DosFs. Subject: Re: i/o redirection Subject: Problem with detaching network interface Subject: Re: checksum verification - embedded Subject: Re: VxWorks interview question ! Subject: Re: i/o redirection Subject: Re: TFFS Format Subject: Re: DosFs 2.0 Product Information Subject: Re: Signal handling - Need tp know how to do? Basic example Subject: Re: DOS File Corruption Subject: Re: i/o redirection ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Booting VxWorks AE 1.1 Date: Tue, 18 Dec 2001 13:24:08 +0100 From: "Mukul Dhankhar" Message-ID: <3c1f3403$0$7110$4d4efb8e@news.be.uu.net> Hi all, I am using VxWorks AE 1.1 for the first time( never used any previous versions of it too). My target hardware is Pentium-MMX. I run Tornado on a WinNT host. I need to create C applications and run them on my target. First I boot the target with a boot floppy created using 'mkboot' utility. By default the boot.txt file needs to be downloaded through ethernet (using FTP). MY QUESTION IS: how do I come to know the IP address of my target? There is DHCP server access on the network. Is there another( and probably easier way) to boot the target? It may be through serial connection or it may be that the boot image in floppy contains all the source files( written in C) statically linked so there is no need for an ethernet connection. Can anyone help me with this, because I am new to VxWorks. Thanks in advance, Mukul. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: i/o redirection Date: Tue, 18 Dec 2001 12:22:38 +0000 From: Graham Baxter Organization: Graham Baxter (Software) Limited Message-ID: <3C1F350E.EC4A30B4@NOSPAM.bcs.org.uk> References: Ofer There is an application note: http://www.wrs.com/csdocs/technote/wtn13.html - -- Regards, Graham Baxter (VxWorks and pSOS BSP's) Freelance Software Engineer - AVAILABLE FOR NEW ASSIGNMENT SOON gbaxter@NOSPAM.bcs.org.uk Ofer Goren wrote: > > Hi. > How can I redirect my I/O, so my own task will handle it, rather the shell > task? > > Thanks, > > -- > Ofer Goren > Iamba Technologies --------------------------- Newsgroups: comp.os.vxworks Subject: Re: i/o redirection Date: Tue, 18 Dec 2001 13:02:25 +0000 (UTC) From: "David Laight" Organization: BT Openworld Message-ID: <9vnep1$fql$1@knossos.btinternet.com> References: > > > I have a CLI task. I want all input and output to be re-directed to > > this task, instead of to the shell task. in fact, I can kill the shell > > task, but I want the user to be able to access the shell upon a > > specific command. > > The only way to do this is to kill the shell. All commands can also be > executed from your CLI, but you have to lookup the symbol and call the > function from within the CLI. You can also startup the shell when the user > wants access to the shell. The (undocumented) interface used by the WRS telnet/rlogin server would let you shift the shell to a different fd. However that interface looks somewhat unclean. If you kill the shell it is likely to restart itself. If you taskSuspend the shell it is still inside read(). What I did was get the shell to suspend itself by feeding '^UtaskSuspend\n' into the terminal data stream from a hacked serial driver. (check that the shells stdin is the console first) David --------------------------- Newsgroups: comp.os.vxworks Subject: Re: checksum verification - embedded Date: Tue, 18 Dec 2001 13:06:04 +0000 (UTC) From: "David Laight" Organization: BT Openworld Message-ID: <9vnevr$eij$1@helle.btinternet.com> References: <3C1E161D.A3E58D5F@avoidspam.com> <9vlqok$sa5$1@bcarh8ab.ca.nortel.com> It is certainly worth checksumming anything that has been through ftp - - just in case someone forgot to set 'Image' mode. Our bootrom (includes the compressed vxWorks image we run) has a checksum in the header. This is checked by the bootcode (in romInit.s) - ok this doesn't detect errors in the first sector - and by the code than downloads a new image before writing it to flash. David Yin, Yong wrote in message news:9vlqok$sa5$1@bcarh8ab.ca.nortel.com... > Based on your description, I think the original intent is to detect the > potential damage of this image. As the image is stored in a flush disk, it's > possible you read a damaged image and it's fatal to start this image. So > it's better to use checksum to protect this if you have CPU time. > > Regards > > Yong > Gilad Katz wrote in message > news:3C1E161D.A3E58D5F@avoidspam.com... > > hello, > > > > One of our application has the need to transfer files( vxworks image, > > application executable) from a master controller card to other cards in > > the system. The files on the master card are copied from a ata drive to > > local ramdisk, then checksum verification is done on it. This > > implementation was done by a developer who has left our project. The > > question I have are - > > > > 1) Is there a need to do checksum verification when transmitting data > > from (ata drive - flash disk based, one medium) to local ramdisk > > (another medium)? I feel this is too much of an overhead. The checksum > > verification is done because one is transmitting files from one > > medium(ata drive - flash disk) to another medium (ramdisk - SDRAM > > based). I would prefer to transmit files directly from ata drive's > > directory to the remote slave card using NFS. > > > > thanks. > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: checksum verification - embedded Date: Tue, 18 Dec 2001 15:13:35 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <3C1E161D.A3E58D5F@avoidspam.com> Checksum is a good idea, at least to protect from partial file upload, as one can not predict the file size reliably, and the effect of a runt file or otherwise corrupt file would be disasterous (the system fails totally), checksum is by all means justified. In fact, if this is a telecom or other mission critical equipment, I would replace the checksum with an MD5 signature, which also contains a SECRET, so that noone could purposefully fake a software image, or replace a legal software image with a pirated one. Such an approach would be effective e,g, on a Digital V set-top box, pirats are modifying the STB software image to effectively circumvent conditional access, and there is an entire black market for counterfit cards and pirated boxes. The simple MD5 authenticity check of the software could have effectively eliminated this financially huge problem at source. HTH, - Leonid "Gilad Katz" wrote in message news:3C1E161D.A3E58D5F@avoidspam.com... > hello, > > One of our application has the need to transfer files( vxworks image, > application executable) from a master controller card to other cards in > the system. The files on the master card are copied from a ata drive to > local ramdisk, then checksum verification is done on it. This > implementation was done by a developer who has left our project. The > question I have are - > > 1) Is there a need to do checksum verification when transmitting data > from (ata drive - flash disk based, one medium) to local ramdisk > (another medium)? I feel this is too much of an overhead. The checksum > verification is done because one is transmitting files from one > medium(ata drive - flash disk) to another medium (ramdisk - SDRAM > based). I would prefer to transmit files directly from ata drive's > directory to the remote slave card using NFS. > > thanks. > --------------------------- Newsgroups: comp.os.vxworks Subject: telnet banner Date: Mon, 17 Dec 2001 17:36:08 GMT From: ontij@telepolis.com (Jaime Ontiyuelo Tirapu) Organization: Telefonica Transmision de Datos Message-ID: <3c1e2cfb.268186882@news.ibercom.com> hi! Recently I have launch a telnet against a machine that returns this banner ENTER TYPE NUMBER OR RETURN -07-> TYPE A IS PPP (asynchronous) VME 68K v. 8.4(alpha): 15 Feb 2001 A friend of mine told me this corresponds to vxworks OS. Is this OK or not? How can i communicate with this OS/application? Is something like 3270 terminal comm? Thanks for all! Jaime --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Send message to the target from Window's application Date: Tue, 18 Dec 2001 16:50:22 +0200 From: "David Ganor" Organization: Bezeq International Ltd. Message-ID: <3c1f57aa@news.bezeqint.net> References: <4367c1ab.0112092238.7630e8bd@posting.google.com> <9v2gi4$hgt$1@bcarh8ab.ca.nortel.com> I have used the Catalyst with VB6 on WinNT/2000 and it works fine. Small tip: the control can't transfer (read or write) more than 8192 bytes at a time. If you have more data to move send/read it in chunks inside some kind of a loop... My VB app is aserver connecting a target to a Matlab object on the PC... David Ganor Israel "Charles R. Chernack" wrote in message news:f680a792.0112111130.53548f08@posting.google.com... > I am trying to do the same thing using Visual Basic 6 on the host. > Catalyst (http://www.catalyst.com/) has a free socket plug in for VB6 > and LOTS of DOCS. I'm not to that point in my project yet, so I cannot > report results. > > Charles R. Chernack > Los Altos, CA --------------------------- Newsgroups: comp.os.vxworks Subject: Re: telnet banner Date: Tue, 18 Dec 2001 16:51:30 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <3c1e2cfb.268186882@news.ibercom.com> This could be anything at all - what I think this is, it's an application which sits on the serial link letting you select how this link is to be operated, and obviously one of these options is PPP, which is a way to run TCP/IP over a serial link. It might have a plain text CLI under a different option number. In any case, if you are trying to hack the system, then you might want to start with the PPP option anyways, just hook up the serial link to a PC, and connect using "direct connection" with the Dialup adapter. "Jaime Ontiyuelo Tirapu" wrote in message news:3c1e2cfb.268186882@news.ibercom.com... > > > hi! > > Recently I have launch a telnet against a machine > that returns this banner > > ENTER TYPE NUMBER OR RETURN -07-> > TYPE A IS PPP (asynchronous) > VME 68K v. 8.4(alpha): 15 Feb 2001 > > A friend of mine told me this corresponds to vxworks OS. > Is this OK or not? How can i communicate with this OS/application? > Is something like 3270 terminal comm? > > Thanks for all! > > Jaime > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: i/o redirection Date: Tue, 18 Dec 2001 15:08:58 GMT From: Tony Dal Santo Message-ID: References: Sender: Tony Dal Santo Johan Borkhuis wrote: > "Ofer Goren" wrote: >> I have a CLI task. I want all input and output to be re-directed to >> this task, instead of to the shell task. in fact, I can kill the shell >> task, but I want the user to be able to access the shell upon a >> specific command. > The only way to do this is to kill the shell. All commands can also be > executed from your CLI, but you have to lookup the symbol and call the > function from within the CLI. You can also startup the shell when the user > wants access to the shell. Thought I would also mention that you need to be careful to only start one instance of the shell. It is not reentrant (they say yacc doesn't generate a reentrant parser). I've always been a bit disappointed by this, and am surprised it hasn't been addressed. Don't they allow multiple shells from the Tornado debugger? Tony --------------------------- Newsgroups: comp.os.vxworks Subject: Re: i/o redirection Date: Tue, 18 Dec 2001 11:02:06 -0800 From: "Dale" Message-ID: <3c1f9357_1@news.foxinternet.com> References: Reply-To: "Dale" "Tony Dal Santo" wrote in message news:e_IT7.7$GN5.208@client... > Johan Borkhuis wrote: > > "Ofer Goren" wrote: > > >> I have a CLI task. I want all input and output to be re-directed to > >> this task, instead of to the shell task. in fact, I can kill the shell > >> task, but I want the user to be able to access the shell upon a > >> specific command. > > > The only way to do this is to kill the shell. All commands can also be > > executed from your CLI, but you have to lookup the symbol and call the > > function from within the CLI. You can also startup the shell when the user > > wants access to the shell. > > Thought I would also mention that you need to be careful to only start > one instance of the shell. It is not reentrant (they say yacc doesn't > generate a reentrant parser). I've always been a bit disappointed by > this, and am surprised it hasn't been addressed. Don't they allow > multiple shells from the Tornado debugger? > Multiple shells from the Tornado debugger run on the host computer and communicate through the target server to the debug agent on the target. I believe that RTI sells a reentrant shell. At least, they used. Take a look at www.rti.com --------------------------- Newsgroups: comp.os.vxworks Subject: TFFS Format Date: Tue, 18 Dec 2001 19:16:48 GMT From: "M Nurmohamed" Message-ID: Hi, I will soon be evaluating a MIPS board with an Intel StrataFlash (28F640J3) chip. A requirement is to implement the TrueFFS in the BSP. The layout of the flash is as follows: +------------------+ 16 MB | A | | | +------------------+ 9 MB | Boot Image | +------------------+ 8 MB | C | | | +------------------+ 0 MB The spec for the evaluation board indicates the boot image should be located as shown above. My question is two parts: 1. Can I format sections 'A' and 'C' as two separate TrueTFFS flash drives? 2. If 'A' can be formatted with TrueTFFS, does it have to be of base 2 size (ie 1, 2, 4, 8, 16, etc) MB? As no board is available yet, I am limited to what I think can be done in theory. The above made sense but cannot be tested until the hardware becomes available. Thank you kindly for your input. Mehmood --------------------------- Newsgroups: comp.os.vxworks Subject: Re: TFFS Format Date: 18 Dec 2001 14:44:01 -0500 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: References: Sender: bpringlemeir@DeadDuck >>>>> "M" == M Nurmohamed writes: [snip] M> The spec for the evaluation board indicates the boot image should M> be located as shown above. My question is two parts: 1. Can I M> format sections 'A' and 'C' as two separate TrueTFFS flash drives? M> 2. If 'A' can be formatted with TrueTFFS, does it have to be of M> base 2 size (ie 1, 2, 4, 8, 16, etc) MB? You can definitely format two tffs devices and mount each with DosFs. Regarding your question that the flash be located on a multiple of base 2, you could use an MMU to remap the flash location if you wished [assuming you have an MMU]. The bootloader requirement is probably due to some firmware, boot sequence, etc. I have used TFFS devices at `weird' boundaries, but I believe that this requires a little more work. You might contact your FAE for specific problems related to this. regards, Bill Pringlemeir. - -- TORPEDOED BY AIRCRAFT. WEB PAGES. SINKING. U-464. vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" --------------------------- Newsgroups: comp.os.vxworks Subject: Re: `Higher reliability' DosFs. Date: 18 Dec 2001 15:13:08 -0500 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: References: Sender: bpringlemeir@DeadDuck >>>>> "Leonid" == Leonid Rosenboim writes: [snip] Leonid> Of course I agree this whole thing is rather tricky, so if Leonid> you can spare some extra disk space, I would recommend to Leonid> avoid the rename() and add some sort of checksum at the end Leonid> of each important file, and during usage, open the latest, Leonid> check it, if it falls through use it, if not fall back to the Leonid> previous version, and so forth. Leonid> And remember, shit happens. Indeed it does... Anyways for everyones benefit. Here are my findings. I have code that implements the following on an DosFs 2 files system, using TFFS on an 8 MB StrataFlash card and 4 Mb Linear Intel flash. I have been running test that starve the dcache task, write and read large amounts of data from different tasks and resets at random... How cruel! My systems is an ARM 710A (barf!) Anyways, this did expose several problems. My finding was that some thing like the following is sufficient to guarantee that the card retained a good copy of data. InFile = open(FileName,O_RDONLY, 0); remove(tmpfile); // tmpnam(); to 8.3 OutFile = open(tmpfile, O_WRONLY | O_CREAT, 0); ... FILE PROCESSING. ... close(InFile); close(OutFile); // Sync the disk. { drvFd = open(drive, O_RDWR, 0); if(drvFd >= 0) { ioctl(drvFd, FIOSYNC, 0); close(drvFd); } } if (rename(tmpfile, destination) != OK ) { error... I think that the disk syncing step is necessary. Otherwise, the entries in the dcache can be re-ordered and the rename write may actually take place before all of the updated file is written. For example if some other task accesses the FAT, file, etc between the close() and rename(). I have burnt out one flash cards and discovered several reset problems in the process... and thanks to Leonid for pointing me in the right direction. regards, Bill Pringlemeir. - -- Have you ever thought it was a good idea to have three large bananas to defend against the artist? Or been given seven cars made out of your telephone? 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: i/o redirection Date: Tue, 18 Dec 2001 21:19:45 +0100 From: "Werner Schiendl" Message-ID: <3c1fa5ac@brateggebdc5.br-automation.co.at> References: Hi, "Johan Borkhuis" wrote in message news:Xns917B8033F1A58borkhuisagerecom@135.7.153.102... ... > The only way to do this is to kill the shell. All commands can also be > executed from your CLI, but you have to lookup the symbol and call the > function from within the CLI. You can also startup the shell when the user > wants access to the shell. > Wouldn't it be a possibility, to mimic the shell's API in one's application and thus replace the shell completely? If all that is called is the documented shellLib APIs, this should be possible. This would have the additional benefit, that it would enable the built-in telnet for that shell-replacement. Has anybody ever tried to do so? regards Werner --------------------------- Newsgroups: comp.os.vxworks Subject: Problem with detaching network interface Date: Tue, 18 Dec 2001 21:03:01 +0000 (UTC) From: "Dmitri Krivchitch" Organization: Mailgate.ORG Server - http://www.Mailgate.ORG Message-ID: <3ee9873b90a2cd336dd45c0c5e1b2ad4.48926@mygate.mailgate.org> Hello all! I have problem with detaching network interface. Everything is fine up to call ifRouteDelete. And then I'm blocked on some Mutex semaphore inside this function call . I don't have any semaphors in my interface and I use serial driver callbacks to transmit and receive characters. Here is disassembled memory and I'm blocked at 0x846a66 _semMTake 008469ee 2279 0089 e9ac MOVEA .L _taskIdCurrent,A1 008469f4 40c1 MOVE SR,D1 008469f6 46f9 0088 5ad8 MOVE _intLockTaskSR,SR 008469fc 0c90 0089 fcf8 CMPI .L #0x89fcf8,(A0) 00846a02 670a BEQ _semMTake + 0x20 ..... 00846a66 4a80 TST .L D0 00846a68 6e02 BGT _semMTake + 0x7e 00846a6a 4e75 RTS 00846a6c 2f2f 0008 MOVE .L (0x8,A7),-(A7) 00846a70 2079 0089 fea4 MOVEA .L __func_sigTimeoutRecalc,A0 00846a76 4e90 JSR (A0) 00846a78 588f ADDQ .L #0x4,A7 00846a7a 2f40 0008 MOVE .L D0,(0x8,A7) 00846a7e 4ef9 0082 e184 JMP _semTake 00846a84 4eb9 0084 5a26 JSR _windExit 00846a8a 70ff MOVEQ #0xff,D0 00846a8c 4e75 RTS 00846a8e 0000 46c1 ORI .B #0xc1,D0 00846a92 4ef9 0084 66f4 JMP _semInvalid If anyone have ideas how to solve this problem, please tell me. Thanks in advance. Dmitri - -- Posted from ottawa-hse-ppp254847.sympatico.ca [64.230.5.136] via Mailgate.ORG Server - http://www.Mailgate.ORG --------------------------- Newsgroups: comp.os.vxworks Subject: Re: checksum verification - embedded Date: Tue, 18 Dec 2001 13:22:40 -0800 From: "Gary M" Message-ID: <3c1fb3a2$0$22741$724ebb72@reader2.ash.ops.us.uu.net> References: <3C1E161D.A3E58D5F@avoidspam.com> <3c1e5063$0$12345$4c41069e@reader1.ash.ops.us.uu.net> <3C1E5AFF.CA341550@gv.net> Reply-To: "Gary M" "Michael R. Kesti" wrote in message news:3C1E5AFF.CA341550@gv.net... > Gary M wrote: > > >(Even byte parity doesn't give 100% confidence, but it is far superior than > >nothing at all). > > Agreed again, but do note that checksumming suffers from the same short- > comming, and for much the same reason. Just as there is a non-zero number > of bit combinations that generate the same parity, there is a number of > data sets that generate the same checksum. > That's why we've discarded checksumming in favor of the public domain Rocksoft CRC algorithm, which is table-driven (hence very fast), and about as good an error detector as you are likely to get. We do 32-bit CRC checks on our VxWorks flash devices. ftp://ftp.rocksoft.com/papers/crc_v3.txt http://www.repairfaq.org/filipg/LINK/F_crc_v3.html --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks interview question ! Date: Tue, 18 Dec 2001 13:47:27 -0800 From: "Gary M" Message-ID: <3c1fb96e$0$22741$724ebb72@reader2.ash.ops.us.uu.net> References: <23de5426.0112100602.4bf0c4ac@posting.google.com> <3c152217$0$2729$724ebb72@reader2.ash.ops.us.uu.net> <3c191f2c$0$20175$724ebb72@reader2.ash.ops.us.uu.net> Reply-To: "Gary M" "Jim" wrote in message news:d15T7.6052$8e.345131@news... > > For your question ... . What are the different types of timing requirements > that real-time systems have? > What type of answer are you looking for? My best guess here would be 1.) > signaling to devices (ie being able to look at a logic analyser and find the > problem with interdevice communication), 2.) issues with data throught put > (ie minimizing bus activity and processor useage to move data through a > system), 3.)service times for various threads/processes and activities ... > would these suffice or do you think typical employeers should look for "key > words" on these answers ... > I was thinking more in terms of "hard requirements" and "soft requirements". Hard requirement loosely defined as "as action which is initiated by stimuli and required to be completed within a specified period of time, or the system and/or mission is subject to failure." Systems such as missile guidance systems and anti-lock braking systems have hard requirements. "Soft requirements" usually specify a timing requirement, but the system is designed to recover from failures to meet deadlines. One of the challenges in a multi-tasking, single processor system is to guarantee that hard requirements will always be met regardless of all other soft requirements in the system. > What about asking canidates about osciloscopes, logic analysers, and > spectrun analysers ... is this fair ... I am a software guy ... but I have > found these tools invaluable ... is this fair to assume that knowledge of > these tools are essential to exhibit ones experience in the field ... Always a plus, but ask them how and when they use it. In my earlier career days, I would too quickly want to spend a couple of hours hooking up an analyzer and running tests, when I could better solve the problem by writing some tracing functions or mousetraps in a couple of minutes. Now I prefer to drag the analyzer out only when the problem is buried in hardware that can't be touched by software. I'd like to stress the need to have candidates describe in detail a major project worked on in the past three years. Interview them in a room with a large whiteboard, if possible, and have them go at it. After they give a broad explanation, drill down into their areas of responsibility, and question them about design decisions, testing plans, how they interacted with other members on their team, problems they encountered and the solutions tried, how they would do things differently now, etc. From this, you will learn more about the candidate's technical strengths and preferences, as well as the candidate's skills in communication, teamwork, design, innovation, quality, and discipline. And if they don't get excited while talking about the project, find out why, because you want them to get excited about your project, right? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: i/o redirection Date: 18 Dec 2001 14:01:23 -0800 From: Joe Chung Message-ID: References: "Ofer Goren" writes: > I have a CLI task. I want all input and output to be re-directed to this > task, instead of to the shell task. in fact, I can kill the shell task, but > I want the user to be able to access the shell upon a specific command. > > Today, when I spawn my cliTask, I get a username prompt. However, when > entering a username, I get an error of invalid command from the shell task, > since this is the task that captures my input. > > So: I want my cliTask to capture all I/O.... How do I do that? > > Thanks, > Ofer, Take the situation of stdin; there are two situations to consider. 1. When you "call" a function f2 from another function f1, then by default they are both using the same stdin. This is not a problem since they both cannot read from stdin at the same time. 2. When you "spawn" a task (t2) from another task (t1), then t2 and t1 must EACH have their own stdin, because it is meaningless for both of them to simultaneously read from the same input (ok, every odd character goes to t1, and every even character goes to t2). If you want t2 to have t1's stdin, then you MUST stop t1 from reading stdin. In the above case, you can replace the word "t1" with the word "console" or "windsh", and you can see it's the same problem. So you have two ways to solve your problem. The easy way (more limiting) is to CALL (not spawn) your cliTask directly from windsh/console. In this case, windsh/console relinquishes control of stdin until cliTask returns. The hard way (more versatile) involves many of the suggestions that have been given to you here. It really depends on what you're using your cliTask for. Are you just trying to debug it right now or are you going to "install" it on a production system so that it is the gatekeeper to some or all I/O terminals? For debugging purposes, one solution is to use an intermediary "launcher" task that passes its stdin/stdout/stderr to the launchee and have the launchee do ioTaskStdSet() to "inherit" them. Of course, the launcher task must "wait" for the launchee to finish before returning to the shell. To access windsh in this case, you just start another instance of windsh, while the current one is occupied by the launcher. - -jc - -- (apply 'concat (reverse (list "com" (char-to-string 46) "yahoo" (char-to-string 64) "joechung"))) --------------------------- Newsgroups: comp.os.vxworks Subject: Re: TFFS Format Date: Tue, 18 Dec 2001 22:07:47 GMT From: Andray Kaganovsky Organization: @Home Message-ID: References: "M Nurmohamed" wrote in news:ACMT7.479$Bv1.116678 @dfiatx1-snr1.gtei.net: > Hi, > > I will soon be evaluating a MIPS board with an Intel StrataFlash (28F640J3) > chip. A requirement is to implement the TrueFFS in the BSP. The layout of > the flash is as follows: > > +------------------+ 16 MB > | A | > | | +------------------+ 9 MB Boot Image > | | | +------------------+ 8 MB C > | | | > | | +------------------+ 0 MB > > The spec for the evaluation board indicates the boot image should be located > as shown above. My question is two parts: > 1. Can I format sections 'A' and 'C' as two separate TrueTFFS flash drives? > 2. If 'A' can be formatted with TrueTFFS, does it have to be of base 2 size > (ie 1, 2, 4, 8, 16, etc) MB? > > As no board is available yet, I am limited to what I think can be done in > theory. The above made sense but cannot be tested until the hardware > becomes available. > > Thank you kindly for your input. > > Mehmood > > > [AndrayK]. Regarding your second question. As far as I know, TrueFFS expects flash array to reside at 4KB boundary. The size of the FTL-formatted flash array can be any integral number of (size_of_flash_erase_block * interleaving) For example, if flash erase block is 64 KBytes, and inteleaving factor is '1', then the size of flash array must be mutiple of 64 KBytes. As far as your first question concerned, it should be possible to install two socket components. First socket component will be mapped to the start of the first TrueFFS volume, and coded to wrap around the end of this volume. Second socket component will be mapped to the start of the second TrueFFS volume, and wrap around the end of it. As you can see, both socket components will be mapped to different, *non-overlapping* parts of the same physical flash array. From TrueFFS perspective, however, there will be two distinct "flash arrays", of different sizes, each residing at it's own address. Each such "flash array" can be separately formatted by TrueFFS, completely independently from the other "flash array". The limitation is, of cause, that these "flash arrays" must not share the single flash part. Andray --------------------------- Newsgroups: comp.os.vxworks Subject: Re: DosFs 2.0 Product Information Date: Tue, 18 Dec 2001 22:17:55 GMT From: Andray Kaganovsky Organization: @Home Message-ID: References: "Don Dewar" wrote in news:aonT7.93$5f4.38153 @dfiatx1-snr1.gtei.net: > I was also shocked to find this out. I called my FAE who said I had to pay > for it. When I pointed out I heard it was free with a support contract he > relented and sent me the binaries. They charge an outrageous amount of > money for the sources. > > The integration went ok, but we suddenly found some of our files > disappearing or getting corrupt. We tried to debug it for months with very > little support from WRS. Finally, we got the sources and still could not > figure it out. When we compiled with less than 2 MB support all our > problems all vanished. > > Anyway they don't seem to support it very well. > > Good luck. > -- Don > "Leonid Rosenboim" wrote in message > news:newscache$w61gog$hug$1@lnews.actcom.co.il... >> I tried to find the technical specifications of DosFs 2.0 on the WindRiver >> Web site to show someone, and to my surprise, I found no mention of it. >> This seems shocking to me that this optional component is not advertised, >> and I remain puzzled as to the reasons why. >> >> Surely I don't expect any official answer from WindRiver (although that >> might be nice to chat again with my old pals), but I would appreciate if >> anyone can shed some light on this enigma. >> >> Of course this is mere curiosity. >> - Leonid >> >> >> > > [AndrayK]. As far as I know, dosFs-2 is included with both Tornado-2.1 and Tornado-3.1, and does not exists as separate product anymore. Andray --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Signal handling - Need tp know how to do? Basic example Date: Tue, 18 Dec 2001 19:12:17 +0000 From: molochai@NOSPAMvapour-trail.demon.co.uk Message-ID: References: <3C189C01.8EF74C81@insa-lyon.fr> Reply-To: molochai@NOSPAMvapour-trail.demon.co.uk Jacques Skubich wrote: > Could somebody give me how to proceed to implement signal handling under > VxWorks? RTFM? Perhaps the VxWorks Programmers Guide. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: DOS File Corruption Date: Tue, 18 Dec 2001 18:54:49 -0700 From: "Shawn Steel" Organization: Thomson Consumer Electronics Message-ID: <3c1ff8c0@news.tce.com> References: <3c0fdc4b@news.tce.com> <9vcd7r$ese$1@rex.ip-plus.net> We can cause the damaged file with a IBM MicroDisk 340MB drive also. WindRivers DosFS2 says it found problems with the disk and says it can fix them but it really doesn't fix everything. "Hermann" wrote in message news:9vcd7r$ese$1@rex.ip-plus.net... > The problem is that most of the PCMCIA flash cards have a damaged file > system when you power off the card during a write access (or a short time > afterwards). We made tests with about 20 different types of flash cards, but > at the moment the only type of PCMCIA flash card that works with VxWorks is > from PRETEC (www.pretec.com) > > > "Fred" schrieb im Newsbeitrag > news:3c0fdc4b@news.tce.com... > > We're using PCMCIA flash cards to store our downloaded vxWorks, but have > > recently had problems with file corruption. Sometimes the file system is > > corrupted, sometimes the file is corrupted. Occassionally we can > reproduce > > the problem by booting the board during a FTP transfer. Does anybody have > > any ideas as to what is causing the problem and how to fix it? > > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: i/o redirection Date: Wed, 19 Dec 2001 12:39:45 +0200 From: "Ofer Goren" Organization: Verio Message-ID: References: Well, I guess this is the only acceptable solution. Why can't I just re-direct my I/O to a different task.....? well, only windriver's GODs can give this answer, I guess.... Anyway, Thanks... - -- Ofer Goren Iamba Technologies "Johan Borkhuis" wrote in message news:Xns917B8033F1A58borkhuisagerecom@135.7.153.102... > "Ofer Goren" wrote: > > > I have a CLI task. I want all input and output to be re-directed to > > this task, instead of to the shell task. in fact, I can kill the shell > > task, but I want the user to be able to access the shell upon a > > specific command. > > The only way to do this is to kill the shell. All commands can also be > executed from your CLI, but you have to lookup the symbol and call the > function from within the CLI. You can also startup the shell when the user > wants access to the 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 === > --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Wed Dec 19 04:14:50 2001 From: "RajaSekhar" Date: Wed Dec 19 04:14:53 PST 2001 Subject: setting up PPC boards Friends, Can any one help me in setting up the PPC boards for Tornado 2.0 on Windows 2k.. What are the required configurations for set up ?? Thanks in advance for any help Raja Sekhar From vxwexplo-errs@csg.lbl.gov Wed Dec 19 06:28:27 2001 From: Donald.McLachlan@crc.ca Date: Wed Dec 19 06:28:29 PST 2001 Subject: serial driver for SeaLevel ACB 5102 I'm sure I'll have to modify it, but does anyone have a vxworks serial driver for the Sealevel ACB-MP.PCI (5102)? Thanks, Don From vxwexplo-errs@csg.lbl.gov Wed Dec 19 20:55:40 2001 From: "Saravana Kumar K" Date: Wed Dec 19 20:55:42 PST 2001 Subject: DPC's in vxWorks Hi, I am facing problem in handling a task in Vxworks. Most of my code goes into ISR routine. Every event is triggered by an interrupt. Is there any method to move some time consuming code out of ISR. Like "Deferred Procedure Calls" in Windows. Thanks Saravanan From vxwexplo-errs@csg.lbl.gov Wed Dec 19 21:05:37 2001 From: "Partha" Date: Wed Dec 19 21:05:39 PST 2001 Subject: sw floating point support without much hassle? Hi, Is there any simple way to enable software floating point support in Coldfire boards without FPU? Here're my configurations: MCF5307 8MB RAM OS: vxWorks 5.3.1 Compiler: Diab 4.2b, 4.3g, 4.4a Diab provides the library libc.a. But if you link it with your app, ld from host shell fails " Error: Relocation value does not fit in 16 bits" I generated a map file .. it seemed to be ok. This error msg comes even if you just add the object file instead of the whole library. I searched this group for possible solutions .. but could get much help. There dont seem to be diab compiler settings like -mlongcall in GCC compiler. I tried the diab compiler settings for PIC & PID .. but that didnt help. Should I for for gnu gcc compiler instead? They 've new release which supports Coldfire processor. But I dont really know where I can get the binaries for it, not the source. Anyway, it'd be interesting to know the exact reason for this linking error. Regards Partha From vxwexplo-errs@csg.lbl.gov Thu Dec 20 04:03:44 2001 From: Vxworks Exploder Date: Thu Dec 20 04:03:48 PST 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Thu Dec 20 04:03:37 PST 2001 Subject: Re: checksum verification - embedded Subject: tffs problem Subject: Re: DosFs 2.0 Product Information Subject: Re: i/o redirection Subject: Re: DosFs 2.0 Product Information Subject: TSFS Anomoly Subject: problem with 82559er driver Subject: UGL? Subject: Re: checksum verification - embedded Subject: Re: UGL? Subject: Where is PTY Driver Subject: help!!! how can i operating on the package of raw ethernet or hdlc??? Subject: Re: initialize multiple ethernet interface when boot vxworks? Subject: Proxy ARPing for network ranges Subject: How to sync up IP stack and Ethernet Driver mBlk Pool Subject: Help: Ramix RM610 and MTX604 Subject: [Q] MessageQueue, anybody would look up my source code? Subject: Re: Please Help Me Subject: proxy artpnew problems Subject: Re: MTD flash driver Q: Subject: How to Read your DOS HDU using DOSFS and ata Subject: Re: initialize multiple ethernet interface when boot vxworks? Subject: Re: initialize multiple ethernet interface when boot vxworks? Subject: Re: vxworks vs. hardened NT Subject: Re: MTD flash driver Q: Subject: How to verify Cache mode at run-time Subject: memcpy problem Subject: cfiscs.c file..is it a good start for MTD? Subject: Question about END on VxSIMs Subject: Re: memcpy problem Subject: Re: MTD flash driver Q: Subject: Re: initialize multiple ethernet interface when boot vxworks? Subject: Re: memcpy problem Subject: Configuring PCI devices!! Subject: debug problem with Tornado2 Subject: Re: [Q] MessageQueue, anybody would look up my source code? Subject: Re: While loop data access exception Subject: how can i use file system on flash? Subject: Re: Flash and the mv5100 Subject: Programmer Error (was Re: MemPartFree Error :) Subject: MemPartFree Error : Subject: config two END on vxworks target? Subject: Re: memcpy problem Subject: MTD/tffs: 64 vs.16-bit Buswidth question...? Subject: Re: T2.1 for Mips MIPS32sfgnu - bootrom/copyLongs Subject: [One more] About PCMCIA / PC Card Driver Subject: VxWorks Device Driver tutorial Subject: Re: what is "makes pointer from integer without a cast"? Subject: Re: query relating the stack size for tasks Subject: Re: help!!! how can i operating on the package of raw ethernet or hdlc??? Subject: Re: Maximum number of ethernets for VxWorks? Subject: Re: memcpy problem Subject: how can i work on END driver? Subject: Re: [Q] MessageQueue, anybody would look up my source code? Subject: Re: memPartCreate pPool ??? Subject: I can't use setjmp Subject: Re: what is "makes pointer from integer without a cast"? Subject: Loading Vxworks from TFFS Subject: Re: VxWorks interview question ! Subject: Re: Loading Vxworks from TFFS Subject: Re: Loading Vxworks from TFFS Subject: how to build a bootable vxworks image? Subject: Re: Loading Vxworks from TFFS Subject: Re: Loading Vxworks from TFFS Subject: Re: Please Help Me Subject: Re: DosFs 2.0 Product Information Subject: Re: updating a module on VxWorks Subject: DHCP hostname option Subject: Re: Need help - Unresolved symbol error while downloading Subject: memPartCreate pPool ??? Subject: ipAttach problem ? Subject: query relating the stack size for tasks Subject: Re: Loading Vxworks from TFFS Subject: help a newbie Subject: SNMP Module as an Application Subject: Re: VxWorks interview question ! Subject: Re: VxWorks AE pentium differences Subject: Crpy Key error in Tornado Prototyper Installation Subject: Re: serial debugging with Tornado 2.0 fails Subject: I can not write or erase flash, why? Subject: updating a module on VxWorks Subject: Re: Booting VxWorks AE 1.1 Subject: Re: double shell parameter Subject: Re: alignment error in RISC cpu executing the vxworks TCP/IP stack Subject: Re: Need help - Unresolved symbol error while downloading Subject: Re: Loading Vxworks from TFFS Subject: Tornado for Managed Switches Subject: Flash and the mv5100 Subject: alignment error in RISC cpu executing the vxworks TCP/IP stack Subject: what is "makes pointer from integer without a cast"? Subject: Re: How to verify Cache mode at run-time Subject: Re: How to verify Cache mode at run-time Subject: Re: memcpy problem Subject: Re: How to verify Cache mode at run-time Subject: Re: UGL? Subject: Re: Where is PTY Driver Subject: Re: problem with 82559er driver Subject: UGL(not WindML)+ZINC developer .... Re: UGL? Subject: Re: Configuring PCI devices!! Subject: Re: what is "makes pointer from integer without a cast"? Subject: Re: alignment error in RISC cpu executing the vxworks TCP/IP stack Subject: Re: Loading Vxworks from TFFS Subject: VxWorks IP and Port Scanner Software Subject: Re: alignment error in RISC cpu executing the vxworks TCP/IP stack Subject: Re: how to build a bootable vxworks image? Subject: Re: tffs problem Subject: Re: help a newbie Subject: Re: Need help - Unresolved symbol error while downloading Subject: Re: Flash and the mv5100 Subject: what is the difference between bootrom and bootalbe vxworks image Subject: Re: i/o redirection Subject: Re: checksum verification - embedded Subject: UDP Receive and send problem Subject: Re: DosFs 2.0 Product Information Subject: Re: DosFs 2.0 Product Information Subject: Re: debug problem with Tornado2 Subject: Re: what is the difference between bootrom and bootalbe vxworks image Subject: Re: I can't use setjmp Subject: Re: config two END on vxworks target? Subject: Re: telnet banner Subject: Re: tffs problem ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: checksum verification - embedded Date: Wed, 19 Dec 2001 10:25:35 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <3C1E161D.A3E58D5F@avoidspam.com> One point I wanted to add here, is that in the past, when we developed the FlashLib file system for Flash memory, all data files where internally protected with a CRC32 checksum to maintain data integrity. All magnetic disk drives use CRC or ECC integrity protection internally, but various disk interfaces (e.g. IDE has not even parity) can be sources for errors beyod the control of the internal checking. Also, Flash translation layers (such as TrueFFS) may or may not add checksum per block of data. In any case, these checksums when implemented, only protect portions of the file data (usually fixes in size, variable size in FlashLib) but they do not protect the file in its entirety, hence I would recommend the application to add file-wide protection for crotocal files which can be protected in this way. - - Leonid "Leonid Rosenboim" wrote in message news:newscache$ctkjog$3nl$1@lnews.actcom.co.il... > Checksum is a good idea, at least to protect from partial file upload, > as one can not predict the file size reliably, and the effect of a runt file > or otherwise corrupt file would be disasterous (the system fails totally), > checksum is by all means justified. > In fact, if this is a telecom or other mission critical equipment, I would > replace > the checksum with an MD5 signature, which also contains a SECRET, > so that noone could purposefully fake a software image, or replace a legal > software image with a pirated one. > Such an approach would be effective e,g, on a Digital V set-top box, pirats > are modifying the STB software image to effectively circumvent conditional > access, > and there is an entire black market for counterfit cards and pirated boxes. > The simple MD5 authenticity check of the software could have effectively > eliminated > this financially huge problem at source. > > HTH, > - Leonid > > "Gilad Katz" wrote in message > news:3C1E161D.A3E58D5F@avoidspam.com... > > hello, > > > > One of our application has the need to transfer files( vxworks image, > > application executable) from a master controller card to other cards in > > the system. The files on the master card are copied from a ata drive to > > local ramdisk, then checksum verification is done on it. This > > implementation was done by a developer who has left our project. The > > question I have are - > > > > 1) Is there a need to do checksum verification when transmitting data > > from (ata drive - flash disk based, one medium) to local ramdisk > > (another medium)? I feel this is too much of an overhead. The checksum > > verification is done because one is transmitting files from one > > medium(ata drive - flash disk) to another medium (ramdisk - SDRAM > > based). I would prefer to transmit files directly from ata drive's > > directory to the remote slave card using NFS. > > > > thanks. > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: tffs problem Date: Wed, 19 Dec 2001 17:09:11 +0300 From: "Shrike" Organization: MTU-Intel ISP Message-ID: <9vq71g$2iqp$1@gavrilo.mtu.ru> Hi! After we boot from disk-on-chip with options: boot device - tffs=0,0, boot file - /tffs0/vxworks, etc. we get "tffsDevCreate failed" after "Attaching to TFFS..." message. MMU was allocated for bios extension memory region (in our case c800:0000). What can be done in this case? Thanks in advance. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: DosFs 2.0 Product Information Date: Wed, 19 Dec 2001 16:16:02 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: Zdorovo, Adnryusha! ALl I found about file systems in the datasheet (for both Vxworks and AE) is this: Fast, flexible I/O and local file system POSIX asynchronous I/O and directory handling SCSI support MS-DOS-compatible file system TrueFFS flash file system (optional) ISO 9660 CD-ROM file system PCMCIA support Not much to go on, and no mention about which dosFs is bundled. I heard that AE includes DosFs 2.0, but again, no proof or detail in the literature. Also, try to search for specific file system features in the Web site, there is nothing ... My only conclusion is that WRS marketing did not improve much in the last 3 years. Leonid "Andray Kaganovsky" wrote in message [snip] > > [AndrayK]. As far as I know, dosFs-2 is included with both Tornado-2.1 > and Tornado-3.1, and does not exists as separate product anymore. > > Andray --------------------------- Newsgroups: comp.os.vxworks Subject: Re: i/o redirection Date: Wed, 19 Dec 2001 16:53:07 +0100 From: "Michael Lawnick" Organization: Buergernetz Dillingen Message-ID: <9vqd56$s2a$1@snoopy.bndlg.de> References: Reply-To: "Michael Lawnick" Sender: mlawnick@p5088f03f.dip.t-dialin.net Hi Ofer, you definitely can redirect the I/O of your shell. Use ioTaskStdSet(). The problem: tShell will pend in its old fd until next character input. You will have to send one to free it. Use a pipe for your redirected shell and you can pass all unhandled commands from your CLI. HTH - -- Mit freundlichen Grüßen, Michael Lawnick ============================================== SOFTEC GmbH Tel +49-731-96600-0 Promenade 17 Fax +49-731-96600-23 D-89073 Ulm Michael Lawnick Germany lawnick@softec.de ============================================== "Ofer Goren" schrieb im Newsbeitrag news:newscache$la8log$04o$1@lnews.actcom.co.il... > Well, I guess this is the only acceptable solution. Why can't I just > re-direct my I/O to a different task.....? well, only windriver's GODs can > give this answer, I guess.... > > Anyway, Thanks... > > -- > Ofer Goren > Iamba Technologies > "Johan Borkhuis" wrote in message > news:Xns917B8033F1A58borkhuisagerecom@135.7.153.102... > > "Ofer Goren" wrote: > > > > > I have a CLI task. I want all input and output to be re-directed to > > > this task, instead of to the shell task. in fact, I can kill the shell > > > task, but I want the user to be able to access the shell upon a > > > specific command. > > > > The only way to do this is to kill the shell. All commands can also be > > executed from your CLI, but you have to lookup the symbol and call the > > function from within the CLI. You can also startup the shell when the user > > wants access to the 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: DosFs 2.0 Product Information Date: Wed, 19 Dec 2001 17:10:02 +0100 From: "Werner Schiendl" Message-ID: <3c20bca4$1@brateggebdc5.br-automation.co.at> References: "Leonid Rosenboim" wrote in message news:newscache$wcilog$ogo$1@lnews.actcom.co.il... ... > > My only conclusion is that WRS marketing did not improve much in the last 3 > years. > As - so it seems - has the file system. - - Werner --------------------------- Newsgroups: comp.os.vxworks Subject: TSFS Anomoly Date: Wed, 19 Dec 2001 11:31:43 -0500 From: Doug Dotson Organization: Verio Message-ID: <3C20C0EF.C720330F@digidata.com> I am creating a file using the following: fd = open ("/tgtsvr/testfile.dat", O_CREAT | O_RDWR, 0x666); and then writing to the file and closing it. When I look at it with Explore it has the Read Only bit set (I'm using NT 4.0). If I reset the RO bit manually, and create the file again, it comes up RW. So if the file exsts, it truncates it and rewrites it. If the file doesn't exist it creates it, writes it, closes it but it end up RO. Got to wonder how it got written if it's RO :) Any Advice Welcome. Thanks, doug --------------------------- Newsgroups: comp.os.vxworks Subject: problem with 82559er driver Date: Wed, 19 Dec 2001 17:47:47 +0100 From: Michiel Meekes Organization: NLR Message-ID: <3C20C4B3.A9463A@nlr.nl> Hi there, I'm having a problem with the latest 82559er driver in vxWorks 5.4 for a Pentium target. As long as I don't attach the network cable the driver init's ok. (ifShow "eeE" returns the correct values, all routines are there) After I attach the cable I get a repeating message: " (0x1fa05230 (tNetTask): Panic: where are those rfd buffers " Apparently receive frame descriptors are needed. Does anyone know what to do in this case? TIA, michiel meekes --------------------------- Newsgroups: comp.os.vxworks Subject: UGL? Date: Wed, 19 Dec 2001 16:59:41 GMT From: Vic Organization: http://www.newsranger.com Message-ID: <1I3U7.4272$XC5.4776@www.newsranger.com> Hi All. I've just been tasked with doing some development with UGL. Are there any "quick start" guides available? Does anyone have any hints/tips (except "don't use UGL" ;-)? There's not a lot of time for this project, so I need to get up to speed quite quickly. Ta muchly! Vic. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: checksum verification - embedded Date: 19 Dec 2001 12:13:58 -0500 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: References: <3C1E161D.A3E58D5F@avoidspam.com> Sender: bpringlemeir@DeadDuck >>>>> "Leonid" == Leonid Rosenboim writes: [snip] Leonid> errors beyod the control of the internal checking. Also, Leonid> Flash translation layers (such as TrueFFS) may or may not add Leonid> checksum per block of data. It would be unusual for a Flash layer to use a standard CRC or checksums. This negates the possibility to `re-write' ones data to zero data for the entire block. At the very least, an alternative validation method must be used. regards, Bill Pringlemeir. - -- vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" --------------------------- Newsgroups: comp.os.vxworks Subject: Re: UGL? Date: 19 Dec 2001 12:18:22 -0500 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: References: <1I3U7.4272$XC5.4776@www.newsranger.com> Sender: bpringlemeir@DeadDuck Vic> Are there any "quick start" guides available? Does anyone have Vic> any hints/tips (except "don't use UGL" ;-)? Platform? CPU, display, mouse, touch screen, keyboard, etc. Vic> There's not a lot of time for this project, so I need to get up Vic> to speed quite quickly. Read the manual. If you are familiar with graphics programming (such as Windows GDI, OpenGL, X-Windows, etc) a lot of concepts will be familiar. A lot depends on the feature set/application that you need to use. At least read the WindML manuals that give an overview. regards, Bill Pringlemeir. - -- vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" --------------------------- Newsgroups: comp.os.vxworks Subject: Where is PTY Driver Date: Wed, 19 Dec 2001 10:09:21 -0800 From: "Dan Pinson" Message-ID: <9vql4m$h1e0h$1@ID-48686.news.dfncis.de> The PTY driver does not seem to be in my VxWorks image (lkup "pty" finds no symbols). Is there a configuration option I need to include? Thanks, Dan --------------------------- Newsgroups: comp.os.vxworks Subject: help!!! how can i operating on the package of raw ethernet or hdlc??? Date: 12 Dec 2001 17:18:33 -0800 From: youerming@163.com (Torna) Organization: http://groups.google.com/ Message-ID: help!!! how can i operating on the package of raw ethernet or hdlc??? i couldnt solve this problem yet. please give some hint. Attaching C++ interface... succeeded. Attaching coff OMF reader for ARM CPU family... succeeded. Warning: Target checksum: 0x961d (computed from 0x1000 to 0xb184c). Host checksum: 0x270b (computed from 0xf40104 to 0xff0950). Warning: Core file checksums do not match. Warning: Symbol .stabstr from section of type STYP_REG. Ignored. Warning: Symbol .stabstr from section of type STYP_REG. Ignored. Warning: Symbol .stabstr from section of type STYP_REG. Ignored. Warning: Symbol .stabstr from section of type STYP_REG. Ignored. Warning: Symbol .stabstr from section of type STYP_REG. Ignored. Warning: Symbol .stabstr from section of type STYP_REG. Ignored. Warning: Symbol .stabstr from section of type STYP_REG. Ignored. i have copied my core file of vxWorks image to c:\vxworks and configuered my target server as the following: core file :c:\vxWorks symbol table : global symbol or full symbol. but i still get the error message above! ! could anybody give me a solution of it for ever? ! could andbody tell me how to make the file vxworks.st or vxworks.sym? !if i want to get vxworks.sym ,which command ? thanks a lot! --------------------------- Newsgroups: comp.os.vxworks Subject: Re: initialize multiple ethernet interface when boot vxworks? Date: 12 Dec 2001 18:05:30 -0800 From: sardine_sz@163.net (sardine) Organization: http://groups.google.com/ Message-ID: References: Johan Borkhuis wrote in message news:... > sardine_sz@163.net (sardine) wrote: > > > hi ,all vxworks guys, > > > > now i use Tornado2、vxworks5.4 for x86. > > my PC Target have two ethernet interface,so i want to initialize > > the two ethernet interface when boot vxworks. > > > > for example: > > > > Attached TCP/IP interface to elPci0.............. > > Attached TCP/IP interface to elPci1.............. > > > > > > any info appreciate ! > > Use the "other" boot parameter. > > Groeten, > Johan thanks Johan, but i want to know how to config multiple END interface ,and how to bind a IP Address to a END interface,and should i call what function, --------------------------- Newsgroups: comp.os.vxworks Subject: Proxy ARPing for network ranges Date: 12 Dec 2001 20:28:42 -0800 From: ragho@yahoo.com (Ragho Mahalingam) Organization: http://groups.google.com/ Message-ID: <2f65ef2.0112122028.612eea84@posting.google.com> I'm looking to have an ethernet interface proxy ARP for a few IP networks, and am confused with the use of proxyArpLib. Turns out that proxyArpLib is for backplane communications, because it insists that my main address be within the proxying range. What's the simplest way to enable proxy ARP on an interface, and have arpintr() respond to the requests?. It seems to me that some of the proxy infrastructure is already there, so someone must already know how to accomplish this :-). I'm using vx 5.3.1. thanks, ragho --------------------------- Newsgroups: comp.os.vxworks Subject: How to sync up IP stack and Ethernet Driver mBlk Pool Date: 12 Dec 2001 19:16:26 -0800 From: ryeung@earthlink.net (Raymond Yeung) Organization: http://groups.google.com/ Message-ID: I've done some experiments this far which seems to suggest that IP stack has its own mBlk netPool separate from that of Ethernet driver. The problem that I see, if the above is true, is that how does the IP stack sync up with the driver in terms of cacheable or non-cacheable memory usage is concerned? If driver assumes the data buffer to be non-cacheable, but the IP application assumes it is cacheable, we would have a problem. How do people usually rectify that? One way I can think of is to do one copy of data from IP/application buffers to Ethernet driver non-cacheable buffer. --------------------------- Newsgroups: comp.os.vxworks Subject: Help: Ramix RM610 and MTX604 Date: 12 Dec 2001 23:15:28 -0800 From: shada@netplane.com (shad ansari) Organization: http://groups.google.com/ Message-ID: <15adeaa9.0112122315.6736e77a@posting.google.com> Hello, Has anyone tried using the RM610 quad port ethernet card from Ramix on a Motorola MTX604 (PowerPc) board? Ramix provided a half-baked driver for the card that I cant get to work. Much thanks, Shad (shada@netplane.com) --------------------------- Newsgroups: comp.os.vxworks Subject: [Q] MessageQueue, anybody would look up my source code? Date: 12 Dec 2001 21:28:54 -0800 From: belliny@soback.kornet.net (belliny, Mint~*) Organization: http://groups.google.com/ Message-ID: <6fb8119a.0112122128.5e08cb02@posting.google.com> I'm currently working with this small sample code[yea, i'm novice^^]. i'd like to use two message queue to communicate between two different tasks. but for some reason, i faced general failure. design program following flow make two message queues in task1 taskspawn task1, task2 and pass to message queues id to each taskspawn task3 that manage message flow [ send t1->t2 or t2->t1 ] task1, task2 wait message with creating task3 point task1->messagesend error! please give some information about the two message queue for ipc post with source code. Thanks in advance... - ----------- IPCMessage.cpp #include "vxWorks.h" #include "msgQLib.h" #include "IPCMessage.h" #include "system_header.h" #include "Layer3.h" #include "Layer2.h" MSG_Q_ID mesgQueueId1, mesgQueueId2; CLayer3 *pMyLayer3; CLayer2 *pMyLayer2; void taskOne(MSG_Q_ID mymesgQueueId1, MSG_Q_ID mymesgQueueId2); void taskTwo(MSG_Q_ID mymesgQueueId1, MSG_Q_ID mymesgQueueId2); void taskThree(void); void Message(void) { int taskIdOne, taskIdTwo, taskIdThree; if ((mesgQueueId1 = msgQCreate(MAX_MESSAGES, MAX_MESSAGE_LENGTH, MSG_Q_FIFO)) == NULL) printf("msgQCreate in failed\n"); if ((mesgQueueId2 = msgQCreate(MAX_MESSAGES, MAX_MESSAGE_LENGTH, MSG_Q_FIFO)) == NULL) printf("msgQCreate in failed\n"); if ((taskIdOne = taskSpawn("t1", 90, 0x100, 2000, (FUNCPTR)taskOne, mesgQueueId1, mesgQueueId2,0,0,0,0,0,0,0,0)) == ERROR) printf("taskSpawn taskOne failed\n"); if ((taskIdTwo = taskSpawn("t2", 90, 0x100, 2000, (FUNCPTR)taskTwo, mesgQueueId1, mesgQueueId2, 0,0,0,0,0,0,0,0)) == ERROR) printf("taskSpawn taskTwo failed\n"); if ((taskIdThree = taskSpawn("t3", 90, 0x100, 2000, (FUNCPTR)taskThree, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0)) == ERROR) printf("taskSpawn taskThree failed\n"); } void taskOne(MSG_Q_ID mymesgQueueId1, MSG_Q_ID mymesgQueueId2) { pMyLayer3 = new CLayer3(mymesgQueueId1, mymesgQueueId2); } void taskTwo(MSG_Q_ID mymesgQueueId1, MSG_Q_ID mymesgQueueId2) { pMyLayer2 = new CLayer2(mymesgQueueId1, mymesgQueueId2); } void taskThree(void) { int nSelect, nLoop; char aryMessage[20]; for (; ; ) { printf("*****************************************\n"); printf("* Test Set *\n"); printf("*****************************************\n"); printf("* 1. Send 1 - Layer3 -> Layer2 *\n"); printf("* 2. Send 2 - Layer2 -> Layer3 *\n"); printf("*****************************************\n"); printf("Select -> "); scanf("%d", &nSelect); switch (nSelect) { case 1: printf("Mode 1, send \n"); pMyLayer3->SendMessageToLayer2("message 1"); break; case 2: printf("Mode 2, send \n"); pMyLayer2->SendMessageToLayer3("message 2"); break; default: break; } } } - ------------- IPCMessage.h // - --------------------------------- Layer2.cpp // StateControl.cpp: implementation of the CStateControl class. // ////////////////////////////////////////////////////////////////////// #include "vxWorks.h" #include "msgQLib.h" #include "Layer2.h" #include "stdio.h" #include "system_header.h" ////////////////////////////////////////////////////////////////////// // Construction/Destruction ////////////////////////////////////////////////////////////////////// CLayer2::CLayer2(MSG_Q_ID myMesgQueueId1, MSG_Q_ID myMesgQueueId2) { char msgBuf[MAX_MESSAGE_LENGTH]; mesgQueueId_receive = myMesgQueueId1; mesgQueueId_send = myMesgQueueId2; for ( ; ; ) { if (msgQReceive(mesgQueueId_receive, msgBuf, MAX_MESSAGE_LENGTH, WAIT_FOREVER) == ERROR) { printf("msgQReceive in Layer2 error\n"); break; } else printf("Received %s\n", msgBuf); } } CLayer2::~CLayer2() { } void CLayer2::SendMessageToLayer3(char *strMessage) { // char msgBuf[MAX_MESSAGE_LENGTH]; // if (msgQReceive(mesgQueueId, msgBuf, MAX_MESSAGE_LENGTH, WAIT_FOREVER) == ERROR) // printf("msgQReceive in taskTwo failed\n"); // else // printf("%s\n", msgBuf); char message[] = "Received Message from taskOne"; printf("%s\n", strMessage); if (mesgQueueId_send == NULL) printf("error"); else { if ((msgQSend(mesgQueueId_send, message, MAX_MESSAGE_LENGTH, WAIT_FOREVER, MSG_PRI_NORMAL)) == ERROR) printf("msgQSend in Layer2 failed\n"); else printf("Sent\n"); } } - ------------------ Layer2.h #include "system_header.h" class CLayer2 { public: CLayer2(MSG_Q_ID myMesgQueueId1, MSG_Q_ID myMesgQueueId2); virtual ~CLayer2(); void SendMessageToLayer3(char *strMessage); public: MSG_Q_ID mesgQueueId_send; MSG_Q_ID mesgQueueId_receive; }; - ------------------------------------------- Layer3.cpp // StateControl.cpp: implementation of the CStateControl class. // ////////////////////////////////////////////////////////////////////// #include "vxWorks.h" #include "msgQLib.h" #include "Layer3.h" #include "system_header.h" ////////////////////////////////////////////////////////////////////// // Construction/Destruction ////////////////////////////////////////////////////////////////////// CLayer3::CLayer3(MSG_Q_ID myMesgQueueId1, MSG_Q_ID myMesgQueueId2) { char msgBuf[MAX_MESSAGE_LENGTH]; mesgQueueId_receive = myMesgQueueId2; mesgQueueId_send = myMesgQueueId1; /* for ( ; ; ) { if (msgQReceive(mesgQueueId_receive, msgBuf, MAX_MESSAGE_LENGTH, WAIT_FOREVER) == ERROR) { printf("msgQReceive in Layer3 error\n"); break; } else printf("Received %s\n", msgBuf); } */ } CLayer3::~CLayer3() { } void CLayer3::SendMessageToLayer2(char *strMessage) { // char msgBuf[MAX_MESSAGE_LENGTH]; // if (msgQReceive(mesgQueueId, msgBuf, MAX_MESSAGE_LENGTH, WAIT_FOREVER) == ERROR) // printf("msgQReceive in taskTwo failed\n"); // else // printf("%s\n", msgBuf); // char message[] = "Received Message from taskOne"; if ((msgQSend(mesgQueueId_send, strMessage, MAX_MESSAGE_LENGTH, WAIT_FOREVER, MSG_PRI_NORMAL)) == ERROR) printf("msgQSend in Layer3\n"); else printf("Sent\n"); } - --------------------- Layer3.h #include "system_header.h" class CLayer3 { public: CLayer3(MSG_Q_ID mymesgQueueId1, MSG_Q_ID myMesgQueueId2); virtual ~CLayer3(); void SendMessageToLayer2(char *strMessage); public: MSG_Q_ID mesgQueueId_receive; MSG_Q_ID mesgQueueId_send; }; - ------------------------------------------ system_define.h #ifndef __SYSTEM_DEF #define __SYSTEM_DEF #include "vxWorks.h" #include "msgQLib.h" /* defines */ #define MAX_MESSAGES 100 #define MAX_MESSAGE_LENGTH 50 #endif --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Please Help Me Date: 13 Dec 2001 01:24:06 -0800 From: digipsyche@sify.com (DigiPsyche) Organization: http://groups.google.com/ Message-ID: <6c32a22b.0112130124.3f3ab027@posting.google.com> References: <6c32a22b.0112120223.1cfc1bb2@posting.google.com> leonid, thx for your suggestions, actaully i am an electronics engineer and know 8085, 8086 processors assembely languages and worked on small projects on C and C++. and presently workin under SuSe linux7.0 OS. DigiPsyche "Leonid Rosenboim" wrote in message news:... > Digi, Dear, > > Your competencies are a far cry from what a decent embedded > engineer needs to have. but if you have the guts for embedded, > then here is a partial sillabus that can get you there: > > + A basic course in hardware including Logic and processor architectures > + You must know at least 2 assembly languages, preferrably one > classic CISC (e.g. PDP11, MC68K) and one RISC (Sparc or MIPS ) > + You need to be proficient in "C", including a few hundered thousands > lines of code you wrote and debugged on a non-embedded platform > + You may already know communications to some extent, but in the > Embedded world, it's all a matter of bits and bytes, so you need to get > low and dirty on those as well > > If you aim specifically at the 3G cellphone development, then a good level > mathematics, specifically in Signal Processing is a big addvantage. > > So it's a matter of time, money and guts, > Good luck! > - Leonid > > > "DigiPsyche" wrote in message > news:6c32a22b.0112120223.1cfc1bb2@posting.google.com... > > Hi, > > I am an internet application developer and working on java, jsp, > > perl, php, javascript technologies and very new to this group. i am > > very much intrested in learning and making my future in embedded > > technologies. i have very less knowledge about all the enbedded > > technologies. Please help me, which technologies i should know to go > > for 3G wireless communication and VOIP/Data communication and from > > where to startup? Is there any website to download the free version of > > these softwares? > > > > thanks > > DigiPsyche --------------------------- Newsgroups: comp.os.vxworks Subject: proxy artpnew problems Date: Wed, 19 Dec 2001 15:44:20 -0600 From: "Chris Smith" Organization: Southwest Research Institue Message-ID: Reply-To: "Chris Smith" Can anyone tell me how to get rid of the 0x.... (tNetTask):arptnew failed on 81818181 and cksum: out of data errors incurred by the proxy server? I have found the proxyPortFwdOff(67) that fixes the second error, but it does not address the first. Are there any other fixes I have missed? (this may be an old issue, i know there have been several known problems, but I cant seem to find the fixes) Im using tornado 2.0 btw not tornado 2.0.2... Thanks, chris --------------------------- Newsgroups: comp.os.vxworks Subject: Re: MTD flash driver Q: Date: 13 Dec 2001 02:02:59 -0800 From: hsu_yuan@yahoo.com (Yuan Hsu) Organization: http://groups.google.com/ Message-ID: References: <9v8din$un7$1@news.netmar.com> Hi, You use the most radical configuration for flashes, AFSIK ;) The 4 chips in the same bank should be accessed parallelly, just as the h/w guys told you, because the 4 chips' CE (chip enable) pins are probably linked together. So you perhaps have to use some 64-bit data type to access them in one run, such as "long long int" defined in GCC. To access single chip in one bank, you need some workaround: take write for example, you should first read the 64-bit original data then mask them with your 8-bit data to be written, then write the new 64-bit data back. Seems lumbersome, but perhaps you have to follow the rule of h/w, otherwise strange behavior will happen. HTH, XY afernando@adventnetworks.com wrote in message news:<9v8din$un7$1@news.netmar.com>... > HI, > > i am writing a MTD for Intel 28F128J3A flash. we have 2 banks, and on each > bank we have 4 chips, (8M x 64bits). There are number of things i dont > understand about addressing in this case. The h/w guys tells me that i could > access all 4 chips in parallel? then how do i access each device > individually? in my case what is "noOfChips," 4 or 8 (for vxWorks, FLFlash > structure). i guess i dont' understand this "interleaving" and "socket" > concept used flash domain. > > Also, how do you address CUI (command user interface) of the Flash device? > This is 16-bit in mode. any help is very much appreciated. > --------------------------- Newsgroups: comp.os.vxworks Subject: How to Read your DOS HDU using DOSFS and ata Date: 12 Dec 2001 17:40:15 -0800 From: chernack@juno.com (Charles R. Chernack) Organization: http://groups.google.com/ Message-ID: I spent a few hours NOT getting this to work, and then with a bit of help from WRS was successful. I thought I'd log this here for the next person. I am booting my VxWorks application off a DOS hdu using VXLOAD.COM and ata -- but I could not find out how to read a file from the HDU -- I could not get the HDU in the "devs" list of open file handles. Here is the solution. (vio_print_at is my own little creation) Also, fBIOSImage was an identifer I stole from another program and is out of context there -- unsigned char sz [256]; FILE *fBIOSImage; unsigned long ulBIOSBytesRead; unsigned char ucBIOSImage[0x1000]; memset (ucBIOSImage, 0, sizeof (ucBIOSImage)); usrAtaConfig (0, 0, "/DOS/"); if( (fBIOSImage = fopen( "/DOS/AUTOEXEC.BAT", "r" )) == NULL ) vio_print_at (6,0,"fopen: file AUTOEXEC.BAT could not be opened"); else { vio_print_at (6,0,"Read AUTOEXEC.BAT"); ulBIOSBytesRead = fread( &ucBIOSImage[0], sizeof( char ), 0x1000, fBIOSImage ); fclose( fBIOSImage ); sprintf (sz, "Size of Autoexec.Bat is %li bytes", ulBIOSBytesRead); vio_print_at (7,0,sz); Charles R. Chernack Los Altos, CA --------------------------- Newsgroups: comp.os.vxworks Subject: Re: initialize multiple ethernet interface when boot vxworks? Date: 13 Dec 2001 05:51:16 -0800 From: jperron@cetaceannetworks.com (Jason Perron) Organization: http://groups.google.com/ Message-ID: References: You need to do a couple of things: 1. Add the 2nd interface to the 'endDrvTbl' in configNet.h. 2. Also in configNet.h, add the following line after the endDrvTbl declaration: #define IP_MAX_UNITS ((NELEMENTS(endDevTbl) - 1) Recompile and you should be good to go. Jason sardine_sz@163.net (sardine) wrote in message news:... > Johan Borkhuis wrote in message news:... > > sardine_sz@163.net (sardine) wrote: > > > > > hi ,all vxworks guys, > > > > > > now i use Tornado2、vxworks5.4 for x86. > > > my PC Target have two ethernet interface,so i want to initialize > > > the two ethernet interface when boot vxworks. > > > > > > for example: > > > > > > Attached TCP/IP interface to elPci0.............. > > > Attached TCP/IP interface to elPci1.............. > > > > > > > > > any info appreciate ! > > > > Use the "other" boot parameter. > > > > Groeten, > > Johan > > thanks Johan, > > but i want to know how to config multiple END interface ,and > how to bind a IP Address to a END interface,and should i call what > function, --------------------------- Newsgroups: comp.os.vxworks Subject: Re: initialize multiple ethernet interface when boot vxworks? Date: 13 Dec 2001 12:59:22 -0800 From: akarpov@my-deja.com (Alex) Organization: http://groups.google.com/ Message-ID: <4f09bb68.0112131259.2a65795f@posting.google.com> References: Sardine, configNet.h provided in your BSP would be the right place to start looking at. You will specify the name of your driver’s endLoad() entry point (for example #define MY_LOAD_FUNC myEndLoad() ) as well as the initialization string (for instance MY_LOAD_STRING) which specifies a MAC address among the other things. Finally you will have to edit the definition of endDevTbl to include MY_LOAD_FUNC and MY_LOAD_STRING. Hope this helps, Alex - -------------------- sardine_sz@163.net (sardine) wrote in message news:... > Johan Borkhuis wrote in message news:... > > sardine_sz@163.net (sardine) wrote: > > > > > hi ,all vxworks guys, > > > > > > now i use Tornado2、vxworks5.4 for x86. > > > my PC Target have two ethernet interface,so i want to initialize > > > the two ethernet interface when boot vxworks. > > > > > > for example: > > > > > > Attached TCP/IP interface to elPci0.............. > > > Attached TCP/IP interface to elPci1.............. > > > > > > > > > any info appreciate ! > > > > Use the "other" boot parameter. > > > > Groeten, > > Johan > > thanks Johan, > > but i want to know how to config multiple END interface ,and > how to bind a IP Address to a END interface,and should i call what > function, VxWorks --------------------------- Newsgroups: comp.os.vxworks Subject: Re: vxworks vs. hardened NT Date: 13 Dec 2001 13:28:29 -0800 From: akarpov@my-deja.com (Alex) Organization: http://groups.google.com/ Message-ID: <4f09bb68.0112131328.1b30b87a@posting.google.com> References: <1006965253.4669.0.nnrp-14.3e31b912@news.demon.co.uk> Raymond, A company I used to work for faced the same kind of question a few years ago. Finally they selected VxWorks among the other RTOS’es. You might want to see one of the high performance motion controllers running VxWorks at http://www.motionvillage.com/products/controllers/mc/ Cheers Alex Once "Raymond Bunton" wrote in message news:<1006965253.4669.0.nnrp-14.3e31b912@news.demon.co.uk>... > Hi all, > > I work for a company in the UK that is attempting to develop an industrial > controller using a 'real-time' operating system. I have been selected to > decide on which software package is best suited to our needs. The choice is > either a VxWorks or Hardened NT system. In addition I must also decide upon > a PC/104 or CompactPCI platform. > > When deciding this I would need to bear in mind (among other things): > > - Cost > - Availability (both current and future as regards upgradability and > support) > - Reliability > - Portability > > I welcome all your comments and suggestions as to what you, in your > experience, consider to be the better software package. > > Thanks a lot! > > Ray Bunton --------------------------- Newsgroups: comp.os.vxworks Subject: Re: MTD flash driver Q: Date: 13 Dec 2001 13:27:51 -0800 From: roswel_ajf@hotmail.com (roswell) Organization: http://groups.google.com/ Message-ID: <6dcae13d.0112131327.4d8bcf1d@posting.google.com> References: <9v8din$un7$1@news.netmar.com> hi Yuan, many thanks for the comments. 1 question: take your write example, ok first read the original 64-bit, then how do i know which 16-bits to select? sure i will know which 64-bit row to select, but how to find the which one of the four 2byte lanes. thanks hsu_yuan@yahoo.com (Yuan Hsu) wrote in message news:... > Hi, > > You use the most radical configuration for flashes, AFSIK ;) The 4 > chips in the same bank should be accessed parallelly, just as the h/w > guys told you, because the 4 chips' CE (chip enable) pins are probably > linked together. So you perhaps have to use some 64-bit data type to > access them in one run, such as "long long int" defined in GCC. To > access single chip in one bank, you need some workaround: take write > for example, you should first read the 64-bit original data then mask > them with your 8-bit data to be written, then write the new 64-bit > data back. Seems lumbersome, but perhaps you have to follow the rule > of h/w, otherwise strange behavior will happen. > > HTH, > > XY > > > afernando@adventnetworks.com wrote in message news:<9v8din$un7$1@news.netmar.com>... > > HI, > > > > i am writing a MTD for Intel 28F128J3A flash. we have 2 banks, and on each > > bank we have 4 chips, (8M x 64bits). There are number of things i dont > > understand about addressing in this case. The h/w guys tells me that i could > > access all 4 chips in parallel? then how do i access each device > > individually? in my case what is "noOfChips," 4 or 8 (for vxWorks, FLFlash > > structure). i guess i dont' understand this "interleaving" and "socket" > > concept used flash domain. > > > > Also, how do you address CUI (command user interface) of the Flash device? > > This is 16-bit in mode. any help is very much appreciated. > > --------------------------- Newsgroups: comp.os.vxworks Subject: How to verify Cache mode at run-time Date: 13 Dec 2001 12:22:08 -0800 From: ryeung@earthlink.net (Raymond Yeung) Organization: http://groups.google.com/ Message-ID: I try looking up 8260 manual for some sort of configuration registers that I can read, but in no vail. Does anyone has any idea? I want to read it at run-time to ensure that it is set to what I want. In particular, I want to know whether copy-back vs. write-through is used. --------------------------- Newsgroups: comp.os.vxworks Subject: memcpy problem Date: 13 Dec 2001 15:00:53 -0800 From: dragonli@yahoo.com (June) Organization: http://groups.google.com/ Message-ID: I wrote the following simple wrap around function for gethostbyname. But the memcpy (i also tried bcopyBytes and bcopy) doesn't work at all. I always got exceptions. what should i do? struct hostent *gethostbyname(char *name) { struct hostent *h=NULL; int addr = hostGetByName(name); if (addr != -1) { h->h_name = name; h->h_addrtype = AF_INET; h->h_length = sizeof(addr); memcpy(h->h_addr,&addr,h->h_length); } return h; } --------------------------- Newsgroups: comp.os.vxworks Subject: cfiscs.c file..is it a good start for MTD? Date: 13 Dec 2001 19:32:10 -0800 From: roswel_ajf@hotmail.com (roswell) Organization: http://groups.google.com/ Message-ID: <6dcae13d.0112131932.24122efc@posting.google.com> HI: I am about use cfiscs.c as MTD base, for my 4 chips flash (x16). But, according to SPR# below (on windsurf), using cfiscs.c might not be good idea. Perphaps the i28f016.c file, might be better for 28F128J3A Intel flash. Any comments are helpful. - ----------- SPR# 27893 TITLE: The MTD for CFI devices cfiscs.c does not adequately support interleaving parts that are more than 8 bits wide STATUS Accepted --------------------------- Newsgroups: comp.os.vxworks Subject: Question about END on VxSIMs Date: 13 Dec 2001 19:13:17 -0800 From: houcheng@ms17.hinet.net (houcheng) Organization: http://groups.google.com/ Message-ID: Dear all: I've compile the templateEnd.c on VxSIM (SIMNT) environment and load it by invoking: "muxDevLoad 0, templateLoad, "3:33:33", 0, 0" in target shell. And it return a non-null END pointer to me. The template device can also be found in the list of "muxShow()" report. But the template device can not be found by "endFindByName()" or "ipAttach()". I invoking "endFindByName "template", 0" the return value is always zero. Why ? Is END device can not works under SIMNT simulator except its own nt ulip END device ? Thank You --------------------------- Newsgroups: comp.os.vxworks Subject: Re: memcpy problem Date: 13 Dec 2001 23:45:31 -0800 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0112132345.5daf89e5@posting.google.com> References: Hello, Duh! You need to actually provide some memory for 'h' to point to since you dereference it in this function - instead, you set it to NULL (i.e. address = zero), so you should get exceptions, or worse. HTH, John... dragonli@yahoo.com (June) wrote in message news:... > I wrote the following simple wrap around function for gethostbyname. > But the memcpy (i also tried bcopyBytes and bcopy) doesn't work at > all. I always got exceptions. what should i do? > > struct hostent *gethostbyname(char *name) > { > struct hostent *h=NULL; > int addr = hostGetByName(name); > if (addr != -1) > { > h->h_name = name; > h->h_addrtype = AF_INET; > h->h_length = sizeof(addr); > memcpy(h->h_addr,&addr,h->h_length); > > } > return h; > > } --------------------------- Newsgroups: comp.os.vxworks Subject: Re: MTD flash driver Q: Date: 13 Dec 2001 08:32:21 -0800 From: robin_ch@263.net (Robin Tcheng) Organization: http://groups.google.com/ Message-ID: References: <9v8din$un7$1@news.netmar.com> Hi, you said that you can access all 4 chips in parallel, I think that's enough to write a siutable MTD.why do you want to access single one? noOfChips is the number of chips used in a bank,that is 4. afernando@adventnetworks.com wrote in message news:<9v8din$un7$1@news.netmar.com>... > HI, > > i am writing a MTD for Intel 28F128J3A flash. we have 2 banks, and on each > bank we have 4 chips, (8M x 64bits). There are number of things i dont > understand about addressing in this case. The h/w guys tells me that i could > access all 4 chips in parallel? then how do i access each device > individually? in my case what is "noOfChips," 4 or 8 (for vxWorks, FLFlash > structure). i guess i dont' understand this "interleaving" and "socket" > concept used flash domain. > > Also, how do you address CUI (command user interface) of the Flash device? > This is 16-bit in mode. any help is very much appreciated. > > ----- Posted via NewsOne.Net: Free (anonymous) Usenet News via the Web ----- > http://newsone.net/ -- Free reading and anonymous posting to 60,000+ groups > NewsOne.Net prohibits users from posting spam. If this or other posts > made through NewsOne.Net violate posting guidelines, email abuse@newsone.net --------------------------- Newsgroups: comp.os.vxworks Subject: Re: initialize multiple ethernet interface when boot vxworks? Date: 13 Dec 2001 21:21:04 -0800 From: sardine_sz@163.net (sardine) Organization: http://groups.google.com/ Message-ID: References: <4f09bb68.0112131259.2a65795f@posting.google.com> thank Alex,Jason,Johan and everyone, i summarize it as follow,hope it can help some people: 1> add extra interface to END_TBL_ENTRY in confignet.h for example: END_TBL_ENTRY endDevTbl [] = { #ifdef INCLUDE_EL_3C90X_END {0, EL_3C90X_LOAD_FUNC, EL_3C90X_LOAD_STR_0, EL_3C90X_BUFF_LOAN_0, NULL, FALSE}, {1, EL_3C90X_LOAD_FUNC, EL_3C90X_LOAD_STR_0, EL_3C90X_BUFF_LOAN_0, NULL, FALSE}, */ #endif /* INCLUDE_EL_3C90X_END */ { 0, END_TBL_END, NULL, 0, NULL, FALSE}, }; 2>change IP_MAX_UNITS in bootConfig.h or configNet.h 3>call follow function in function usrRoot() : endFindByName muxDevLoad muxDevStart muxIoCtl ipAttach usrNetIfAttach but now, i don't know if need to change the second END IRQ,IO Address,etc and what's different in function usrNetIfAttach and usrNetIfConfig ? akarpov@my-deja.com (Alex) wrote in message news:<4f09bb68.0112131259.2a65795f@posting.google.com>... > Sardine, > > configNet.h provided in your BSP would be the right place to start > looking at. > You will specify the name of your driver’s endLoad() entry point > (for example > #define MY_LOAD_FUNC myEndLoad() ) as well as the initialization > string (for instance MY_LOAD_STRING) which specifies a MAC address > among the other things. > Finally you will have to edit the definition of endDevTbl to include > MY_LOAD_FUNC and MY_LOAD_STRING. > > Hope this helps, > Alex > -------------------- > sardine_sz@163.net (sardine) wrote in message news:... > > Johan Borkhuis wrote in message news:... > > > sardine_sz@163.net (sardine) wrote: > > > > > > > hi ,all vxworks guys, > > > > > > > > now i use Tornado2、vxworks5.4 for x86. > > > > my PC Target have two ethernet interface,so i want to initialize > > > > the two ethernet interface when boot vxworks. > > > > > > > > for example: > > > > > > > > Attached TCP/IP interface to elPci0.............. > > > > Attached TCP/IP interface to elPci1.............. > > > > > > > > > > > > any info appreciate ! > > > > > > Use the "other" boot parameter. > > > > > > Groeten, > > > Johan > > > > thanks Johan, > > > > but i want to know how to config multiple END interface ,and > > how to bind a IP Address to a END interface,and should i call what > > function, > > VxWorks --------------------------- Newsgroups: comp.os.vxworks Subject: Re: memcpy problem Date: 14 Dec 2001 09:33:01 -0800 From: dragonli@yahoo.com (June) Organization: http://groups.google.com/ Message-ID: References: i thought it probably because of NULL too. but i tried to print out the address of h and i actually used this: printf("address of h:0x%x\n",&h); and see this: address of h:0x508fc0 That means it was allocated some space for h. And also i tried struct hostent *h=malloc(sizeof(struct hostent)); and i got the exception: Exception number 16: Task: 0x509044 (t6) Data abort Program Counter: 0x1b320 Status Register: 0x80000013 902dc _vxTaskEntry +28 : 42f54 () 42ff8 _wdbFuncCallLibInit+128: _gethostbyname () 1364fc _gethostbyname +b8 : _memcpy () value = 0 = 0x0 This looks like a quite simple problem. but i just can not figure it out. :( --------------------------- Newsgroups: comp.os.vxworks Subject: Configuring PCI devices!! Date: 14 Dec 2001 03:02:03 -0800 From: gulnaz@jopasana.com (Gulnaz) Organization: http://groups.google.com/ Message-ID: <1e752e41.0112140302.4c72ae7b@posting.google.com> Hi guys, I am new to Vxworks and have read about the basic features of it from vxworks programmers guide...I am creating a knowledge base for my company. I wanted to know what steps u need to follow to configure a PCI device with vxworks os. All I know is that u have to make some changes to the BSP(i may be wrong) Thanks in Advance, Gulnaz --------------------------- Newsgroups: comp.os.vxworks Subject: debug problem with Tornado2 Date: 13 Dec 2001 06:45:48 -0800 From: Bachmeier.Andreas@foerstergroup.de (Andreas Bachmeier) Organization: http://groups.google.com/ Message-ID: <90019cec.0112130645.39d0a62b@posting.google.com> I have some problems when I debug my application. I use Tornado 2.02 with VxWorks 5.4 for BSP486 on a PC/104 board. In the application there are 19 tasks running. There are also callback functions for two external interrupts. If I set a breakpoint at any place in the code, the debugger stops correct at this point. But now, I can't step through the code. After the first step it run in an exception. I found out, that it happened, while at least one interrupt is still active during debugging. I stepped through the initialization of my application. At the point where the first interrupt of an external board was enabled (sysOutWord), the next step through the code gets that mess you can see: =============================================================== Break at 0x00108070: _sysOutWord Task: 0xf9dc28 (tMainControl) Break at 0x00228acf: _EnIawrt + 0x2f Task: 0xf9dc28 (tMainControl) Break at 0x00228ad2: _EnIawrt + 0x32 Task: 0xf9dc28 (tMainControl) Exception number 13: Task: 0xf9dc28 (tMainControl) General Protection Fault Program Counter: 0x00228ad4 Status Register: 0x00010146 Error Code : 0x0000000b 112b4f _wdbRtInfoGet +1d7: 149648 (3, f9dc28, d, f9c104, f9c0dc) 1496e3 _wdbExcLibInit +11b: _wdbEventPost ([18b320, 149648, d, f9c0dc, 0]) 14812a _wdbEventPost +72 : _wdbNotifyHost ([3, f9dc28, 0, f9c068, 1496e8]) 13f248 _wdbNotifyHost +20 : 143340 ([0, f9c04c, 14812f, 3, f9dc28]) 14335c _wdbUdpSockIfInit+2cc: _excStub ([&_write, 6, f9c02b, 1, 0]) 170ec9 _excStub +35 : _excExcHandle (d, f9bfd0, f9bfa8) 14343a _excExcHandle +36 : 112ae8 (d, f9bfd0, f9bfa8, f9bf80) 112b4f _wdbRtInfoGet +1d7: 149648 (3, f9dc28, d, f9bfd0, f9bfa8) 1496e3 _wdbExcLibInit +11b: _wdbEventPost ([18b320, 149648, d, f9bfa8, 0]) 14812a _wdbEventPost +72 : _wdbNotifyHost ([3, f9dc28, 0, f9bf34, 1496e8]) 13f248 _wdbNotifyHost +20 : 143340 ([0, f9bf18, 14812f, 3, f9dc28]) 14335c _wdbUdpSockIfInit+2cc: _excStub ([&_write, 6, f9bef7, 1, 0]) 170ec9 _excStub +35 : _excExcHandle (d, f9be9c, f9be74) 14343a _excExcHandle +36 : 112ae8 (d, f9be9c, f9be74, f9be4c) 112b4f _wdbRtInfoGet +1d7: 149648 (3, f9dc28, d, f9be9c, f9be74) 1496e3 _wdbExcLibInit +11b: _wdbEventPost ([18b320, 149648, d, f9be74, 0]) 14812a _wdbEventPost +72 : _wdbNotifyHost ([3, f9dc28, 0, f9be00, 1496e8]) 13f248 _wdbNotifyHost +20 : 143340 ([0, f9bde4, 14812f, 3, f9dc28]) 14335c _wdbUdpSockIfInit+2cc: _excStub ([&_write, 6, f9bdc3, 1, 0]) 170ec9 _excStub +35 : _excExcHandle (d, f9bd68, f9bd40) 14343a _excExcHandle +36 : 112ae8 (d, f9bd68, f9bd40, f9bd18) 112b4f _wdbRtInfoGet +1d7: 149648 (3, f9dc28, d, f9bd68, f9bd40) 1496e3 _wdbExcLibInit +11b: _wdbEventPost ([18b320, 149648, d, f9bd40, 0]) 14812a _wdbEventPost +72 : _wdbNotifyHost ([3, f9dc28, 0, f9bccc, 1496e8]) 13f248 _wdbNotifyHost +20 : 143340 ([0, f9bcb0, 14812f, 3, f9dc28]) 14335c _wdbUdpSockIfInit+2cc: _excStub ([&_write, 6, f9bc8f, 1, 0]) 170ec9 _excStub +35 : _excExcHandle (d, f9bc34, f9bc0c) 14343a _excExcHandle +36 : 112ae8 (d, f9bc34, f9bc0c, f9bbe4) 112b4f _wdbRtInfoGet +1d7: 149648 (3, f9dc28, d, f9bc34, f9bc0c) 1496e3 _wdbExcLibInit +11b: _wdbEventPost ([18b320, 149648, d, f9bc0c, 0]) 14812a _wdbEventPost +72 : _wdbNotifyHost ([3, f9dc28, 0, f9bb98, 1496e8]) 13f248 _wdbNotifyHost +20 : 143340 ([0, f9bb7c, 14812f, 3, f9dc28]) 14335c _wdbUdpSockIfInit+2cc: _excStub ([&_write, 6, f9bb5b, 1, 0]) 170ec9 _excStub +35 : _excExcHandle (d, f9bb00, f9bad8) 14343a _excExcHandle +36 : 112ae8 (d, f9bb00, f9bad8, f9bab0) 112b4f _wdbRtInfoGet +1d7: 149648 (3, f9dc28, d, f9bb00, f9bad8) 1496e3 _wdbExcLibInit +11b: _wdbEventPost ([18b320, 149648, d, f9bad8, 0]) 14812a _wdbEventPost +72 : _wdbNotifyHost ([3, f9dc28, 0, f9ba64, 1496e8]) 13f248 _wdbNotifyHost +20 : 143340 ([0, f9ba48, 14812f, 3, f9dc28]) 14335c _wdbUdpSockIfInit+2cc: _netJobAdd ([&_write, 6, f9ba27, 1, 0]) =============================================================== Is there a way to debug an application witch use interrupts? thanks for help Andreas --------------------------- Newsgroups: comp.os.vxworks Subject: Re: [Q] MessageQueue, anybody would look up my source code? Date: 13 Dec 2001 09:18:40 -0800 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0112130918.136bc4d7@posting.google.com> References: <6fb8119a.0112122128.5e08cb02@posting.google.com> Hello, You need to be a lot more precise; what do you mean by "general failure?" What actually happened? Did you get an exception? Did nothing happen at all? There are certainly some odd things in your code (I would say it is a little unusual to implement the main loop of a task in an object constructor for example). Some things you might want to look into are: 1) Do the two tasks that are calling msgQReceive() get there before the third task tries to send messages? 2) Are all the stack sizes large enough; you've only given them 2000 bytes which is on the small side, though they don't seem to do much so perhaps it is OK. You can check this using checkStack() from a shell (or there is a graphical version somewhere too). If none of that helps, please try posting a more detailed explanation of the actual failure, and info about the CPU type you are developing for. Rgds, John... belliny@soback.kornet.net (belliny, Mint~*) wrote in message news:<6fb8119a.0112122128.5e08cb02@posting.google.com>... > I'm currently working with this small sample code[yea, i'm novice^^]. > i'd like to use two message queue to communicate between two different > tasks. > but for some reason, i faced general failure. > > design program following flow > make two message queues in task1 > taskspawn task1, task2 and pass to message queues id to each > taskspawn task3 that manage message flow [ send t1->t2 or t2->t1 ] > task1, task2 wait message with creating > task3 point task1->messagesend > error! > > please give some information about the two message queue for ipc > > post with source code. > > Thanks in advance... > > ----------- IPCMessage.cpp > #include "vxWorks.h" > #include "msgQLib.h" > > #include "IPCMessage.h" > > > #include "system_header.h" > #include "Layer3.h" > #include "Layer2.h" > > > MSG_Q_ID mesgQueueId1, mesgQueueId2; > > CLayer3 *pMyLayer3; > CLayer2 *pMyLayer2; > > void taskOne(MSG_Q_ID mymesgQueueId1, MSG_Q_ID mymesgQueueId2); > void taskTwo(MSG_Q_ID mymesgQueueId1, MSG_Q_ID mymesgQueueId2); > void taskThree(void); > > void Message(void) > { > int taskIdOne, taskIdTwo, taskIdThree; > > if ((mesgQueueId1 = msgQCreate(MAX_MESSAGES, MAX_MESSAGE_LENGTH, > MSG_Q_FIFO)) == NULL) > printf("msgQCreate in failed\n"); > > if ((mesgQueueId2 = msgQCreate(MAX_MESSAGES, MAX_MESSAGE_LENGTH, > MSG_Q_FIFO)) == NULL) > printf("msgQCreate in failed\n"); > > > if ((taskIdOne = taskSpawn("t1", 90, 0x100, 2000, (FUNCPTR)taskOne, > mesgQueueId1, mesgQueueId2,0,0,0,0,0,0,0,0)) == ERROR) > printf("taskSpawn taskOne failed\n"); > > if ((taskIdTwo = taskSpawn("t2", 90, 0x100, 2000, (FUNCPTR)taskTwo, > mesgQueueId1, mesgQueueId2, 0,0,0,0,0,0,0,0)) == ERROR) > printf("taskSpawn taskTwo failed\n"); > > if ((taskIdThree = taskSpawn("t3", 90, 0x100, 2000, > (FUNCPTR)taskThree, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0)) == ERROR) > printf("taskSpawn taskThree failed\n"); > } > > > void taskOne(MSG_Q_ID mymesgQueueId1, MSG_Q_ID mymesgQueueId2) > { > pMyLayer3 = new CLayer3(mymesgQueueId1, mymesgQueueId2); > } > > void taskTwo(MSG_Q_ID mymesgQueueId1, MSG_Q_ID mymesgQueueId2) > { > pMyLayer2 = new CLayer2(mymesgQueueId1, mymesgQueueId2); > } > > void taskThree(void) > { > int nSelect, nLoop; > char aryMessage[20]; > > for (; ; ) > { > printf("*****************************************\n"); > printf("* Test Set *\n"); > printf("*****************************************\n"); > printf("* 1. Send 1 - Layer3 -> Layer2 *\n"); > printf("* 2. Send 2 - Layer2 -> Layer3 *\n"); > printf("*****************************************\n"); > printf("Select -> "); > scanf("%d", &nSelect); > switch (nSelect) > { > case 1: > printf("Mode 1, send \n"); > pMyLayer3->SendMessageToLayer2("message 1"); > > break; > > case 2: > printf("Mode 2, send \n"); > pMyLayer2->SendMessageToLayer3("message 2"); > break; > > default: > > break; > } > > > } > } > ------------- IPCMessage.h > // > --------------------------------- Layer2.cpp > // StateControl.cpp: implementation of the CStateControl class. > // > ////////////////////////////////////////////////////////////////////// > #include "vxWorks.h" > #include "msgQLib.h" > #include "Layer2.h" > #include "stdio.h" > > #include "system_header.h" > > ////////////////////////////////////////////////////////////////////// > // Construction/Destruction > ////////////////////////////////////////////////////////////////////// > > CLayer2::CLayer2(MSG_Q_ID myMesgQueueId1, MSG_Q_ID myMesgQueueId2) > { > char msgBuf[MAX_MESSAGE_LENGTH]; > > mesgQueueId_receive = myMesgQueueId1; > mesgQueueId_send = myMesgQueueId2; > > for ( ; ; ) > { > if (msgQReceive(mesgQueueId_receive, msgBuf, MAX_MESSAGE_LENGTH, > WAIT_FOREVER) == ERROR) > { > printf("msgQReceive in Layer2 error\n"); > break; > } > else > printf("Received %s\n", msgBuf); > } > } > > CLayer2::~CLayer2() > { > > } > > void CLayer2::SendMessageToLayer3(char *strMessage) > { > // char msgBuf[MAX_MESSAGE_LENGTH]; > > // if (msgQReceive(mesgQueueId, msgBuf, MAX_MESSAGE_LENGTH, > WAIT_FOREVER) == ERROR) > // printf("msgQReceive in taskTwo failed\n"); > // else > // printf("%s\n", msgBuf); > > char message[] = "Received Message from taskOne"; > > printf("%s\n", strMessage); > > if (mesgQueueId_send == NULL) > printf("error"); > else > { > if ((msgQSend(mesgQueueId_send, message, MAX_MESSAGE_LENGTH, > WAIT_FOREVER, MSG_PRI_NORMAL)) == ERROR) > printf("msgQSend in Layer2 failed\n"); > else > printf("Sent\n"); > } > } > ------------------ Layer2.h > #include "system_header.h" > > class CLayer2 > { > public: > > CLayer2(MSG_Q_ID myMesgQueueId1, MSG_Q_ID myMesgQueueId2); > virtual ~CLayer2(); > void SendMessageToLayer3(char *strMessage); > public: > MSG_Q_ID mesgQueueId_send; > MSG_Q_ID mesgQueueId_receive; > }; > ------------------------------------------- Layer3.cpp > // StateControl.cpp: implementation of the CStateControl class. > // > ////////////////////////////////////////////////////////////////////// > #include "vxWorks.h" > #include "msgQLib.h" > #include "Layer3.h" > > #include "system_header.h" > > ////////////////////////////////////////////////////////////////////// > // Construction/Destruction > ////////////////////////////////////////////////////////////////////// > CLayer3::CLayer3(MSG_Q_ID myMesgQueueId1, MSG_Q_ID myMesgQueueId2) > { > > char msgBuf[MAX_MESSAGE_LENGTH]; > > mesgQueueId_receive = myMesgQueueId2; > mesgQueueId_send = myMesgQueueId1; > /* > for ( ; ; ) > { > if (msgQReceive(mesgQueueId_receive, msgBuf, MAX_MESSAGE_LENGTH, > WAIT_FOREVER) == ERROR) > { > printf("msgQReceive in Layer3 error\n"); > break; > } > else > printf("Received %s\n", msgBuf); > } > */ > } > > CLayer3::~CLayer3() > { > > } > > void CLayer3::SendMessageToLayer2(char *strMessage) > { > // char msgBuf[MAX_MESSAGE_LENGTH]; > > // if (msgQReceive(mesgQueueId, msgBuf, MAX_MESSAGE_LENGTH, > WAIT_FOREVER) == ERROR) > // printf("msgQReceive in taskTwo failed\n"); > // else > // printf("%s\n", msgBuf); > > // char message[] = "Received Message from taskOne"; > > if ((msgQSend(mesgQueueId_send, strMessage, MAX_MESSAGE_LENGTH, > WAIT_FOREVER, MSG_PRI_NORMAL)) == ERROR) > printf("msgQSend in Layer3\n"); > else > printf("Sent\n"); > } > --------------------- Layer3.h > #include "system_header.h" > > class CLayer3 > { > public: > > CLayer3(MSG_Q_ID mymesgQueueId1, MSG_Q_ID myMesgQueueId2); > virtual ~CLayer3(); > void SendMessageToLayer2(char *strMessage); > > public: > MSG_Q_ID mesgQueueId_receive; > MSG_Q_ID mesgQueueId_send; > }; > ------------------------------------------ system_define.h > #ifndef __SYSTEM_DEF > #define __SYSTEM_DEF > > #include "vxWorks.h" > #include "msgQLib.h" > > /* defines */ > #define MAX_MESSAGES 100 > #define MAX_MESSAGE_LENGTH 50 > > #endif --------------------------- Newsgroups: comp.os.vxworks Subject: Re: While loop data access exception Date: 13 Dec 2001 23:38:27 -0800 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0112132338.1811600c@posting.google.com> References: <9vb2dv$rhc2@nntp.cig.mot.com> Hello, Try changing the name of your 'main' function; using main is not a good idea in VxWorks since the compiler generates some "magic" code when it sees this function name to get things working in a Unix environment. VxWorks doesn't expect that. Could be that's the cause of the problem. If not, you need to use some of the tools to find out where the code is crashing. A couple of approaches: use the debugger to attach to the crashed task and find out where it was, or use tt from a shell to get a stack trace for the task. HTH, John... "Michael" wrote in message news:<9vb2dv$rhc2@nntp.cig.mot.com>... > Quick intro: > > class class_a { > int f(const string& s1) { > printf("%s \n", s1.c_str(); > return 1; > } > }; > > main(void* arg) { > class_a a1; > string p1((char*) arg); > while(1) { > a1.f(p1); > } > } > > I am getting a data exception error when I run this code. > > data access > Exception current instruction address: 0x0621e19c > Machine Status Register: 0x0000b030 > Data Access Register: 0x9421ffec > Condition Register: 0x24000042 > Data storage interrupt Register: 0x40000000 > Task: 0x599f550 "pthr15" > > Of course I simplified class_a, but I put a printf at the start of function > f, and that is not being print. What are the steps for me to debug what is > going on? From what I can see, the function f is not being entered at all. > The wierdest part is that function f will be executed the first time, but > the second time through the while loop, it will create that exception. > Another wierd part is if I do not use the while loop, but rather just call > function f repeatedly, the program will work also. So, maybe the problem is > somehow linked to the loop? > > Thanks in advance for the help. If anyone has questions, please email me at > msheng1@email.mot.com. --------------------------- Newsgroups: comp.os.vxworks Subject: how can i use file system on flash? Date: 15 Dec 2001 05:26:47 -0800 From: mimilijie@21cn.com (lijie) Organization: http://groups.google.com/ Message-ID: <97c3acac.0112150526.1aac2d1e@posting.google.com> how can i use file system on flash? for there is no file system on the flash, i want to create one using ramdevcreate(). however, i have some problems on the first parameter.when i use 0 as the first parameter, which mease malloc(). when i restart the vxworks, i can't find the file i stored last time. i am also being informed that i can use a real address on the flash as the first parameter of the function, as 0xc000. but is that all i should do? or there is something i am not getting at? thank for reading this, please help me out of the problem. thank you very much! --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Flash and the mv5100 Date: 14 Dec 2001 20:41:58 -0800 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0112142041.706cfe22@posting.google.com> References: <5a87bef4.0112141106.3de2015b@posting.google.com> Hello, google@dharty.com (David) wrote in message news:<5a87bef4.0112141106.3de2015b@posting.google.com>... > Has anyone used the AM29LV323 Flash that it available on the mv5100 > (PPC604)WITHOUT TFFS? > > Windriver doesn't support it, but it must be possible. > > Specifically I need to be able to download a new boot image to DRAM, > and then blow it into the FLASH. Using the PPCBug is not an option as > that would require the user to remove the board and change jumpers. > > Heck, just a starting address for the Flash memory space would be > great! The start address of the flash should be easy to find; check in the header files in your BSP for ROM_START_ADRS and/or ROM_TEXT_ADRS. Or, look in sysLib.c for the sysPhysMemDesc table - in there you should find an entry for the flash space. Now, once you have the address, try taking a look under target/src/drv/mem for a file called flashMem.c. This is a simple flash memory driver for many flash parts. Yours are probably either in there, or related to a type that is. Check in target/h/drv/mem/flash29.h for supported types; my version doesn't include your chip, but you should be able to find a device code for your device using a data sheet for the chips, and also the block size. The algorithm for writing modern flash devices is not too tough actually; they are much, much simpler to deal with than they used to be. You should able to do some probing from the shell to get the device code if you can't find a data sheet for the chips. One thing to be careful of: make sure that your entry in sysPhysMemDesc does not make the flash read-only - you'll need to "write" to the flash address space to give the device commands as well as actually writing new data to it. HTH, John... --------------------------- Newsgroups: comp.os.vxworks Subject: Programmer Error (was Re: MemPartFree Error :) Date: 15 Dec 2001 08:58:59 -0800 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0112150858.3a34ab20@posting.google.com> References: <7421c9fd.0112150342.20885d61@posting.google.com> Hello, kumar_amit@delhi.tcs.co.in (Amit) wrote in message news:<7421c9fd.0112150342.20885d61@posting.google.com>... > hi > I want to emphasize a problem i am facing while using memPartFree() > > I have created a partition and it returned me a partition Id. > Then i allocated a block of memory using memPartAlloc and it returned > me > a pointer to the allocated block of memory. > Now I tried using memPartFree to Free the memory and it is giving > success > i.e STATUS returned is 0. Now i tried to use that memory and to my > surprise i could use that memory that menat it was not freed. > Please can you throw some light on what this is all happening. Freeing a block of memory just gives it back to the heap manager (partition manager), you can still write to it since the physical memory didn't go anywhere. Of course, you'll be running the risk of corrupting something else that might have been allocated some or all of that memory, but nothing is going to stop you. The same would happen in most, if not all, operating systems that have malloc/free style heaps. FYI, in VxWorks, you can also write over the code, the kernel data and any of the stacks. VxWorks 5.x is a flat address space RTOS; it does not protect anything by default, and all memory is shared. Now, there are a couple of ways you can add some protection: VxVMI or VxWorks AE. Neither of those will prevent you writing to a block of memory that you allocate and then free though - that's your responsibility as the programmer. HTH, John... --------------------------- Newsgroups: comp.os.vxworks Subject: MemPartFree Error : Date: 15 Dec 2001 03:42:57 -0800 From: kumar_amit@delhi.tcs.co.in (Amit) Organization: http://groups.google.com/ Message-ID: <7421c9fd.0112150342.20885d61@posting.google.com> hi I want to emphasize a problem i am facing while using memPartFree() I have created a partition and it returned me a partition Id. Then i allocated a block of memory using memPartAlloc and it returned me a pointer to the allocated block of memory. Now I tried using memPartFree to Free the memory and it is giving success i.e STATUS returned is 0. Now i tried to use that memory and to my surprise i could use that memory that menat it was not freed. Please can you throw some light on what this is all happening. Thanks Amit Kumar --------------------------- Newsgroups: comp.os.vxworks Subject: config two END on vxworks target? Date: 14 Dec 2001 04:19:36 -0800 From: sardine_sz@163.net (sardine) Organization: http://groups.google.com/ Message-ID: hi,all Gurus, now i use T2 ,Vxworks5.4 Fox X86, my target have two ethernet interface , so i want to initialize it when vxworks boot, i config as follow: hi, sorry to bother u ! now i config two END interface on vxworks target.but have some problem . (T2,vxworks5.4) as follow: 1> add END table in: END_TBL_ENTRY endDevTbl [] = { #ifdef INCLUDE_EL_3C90X_END {0, EL_3C90X_LOAD_FUNC, EL_3C90X_LOAD_STR_0, EL_3C90X_BUFF_LOAN_0, NULL, FALSE}, {1, EL_3C90X_LOAD_FUNC, EL_3C90X_LOAD_STR_0, EL_3C90X_BUFF_LOAN_0, NULL, FALSE}, #endif /* INCLUDE_EL_3C90X_END */ { 0, END_TBL_END, NULL, 0, NULL, FALSE}, } 2> change IP_MAX_UNITS = 2 3> add follow function in bootLoad() in bootConfig.c pEnd = endFindByName ("elPci",1); /*Success*/ muxDevStart(pEnd) /*Success*/ muxIoctl (pEnd, EIOCGMIB2, (char *)&endM2Tbl) /*Success*/ ipAttach (1,"elPci") /*Success*/ usrNetIfConfig("elPci",1, "192.168.0.113","MyHost",0) /*Failed ,page fault*/ any info appreciate ! --------------------------- Newsgroups: comp.os.vxworks Subject: Re: memcpy problem Date: 13 Dec 2001 22:06:27 -0800 From: ca_prashanth@mindtree.com (Prashanth) Organization: http://groups.google.com/ Message-ID: References: Hi, You haven't allocated memory ,to which the hostent structure pointer should point.I think that should solve the problem. Regards, Prashanth dragonli@yahoo.com (June) wrote in message news:... > I wrote the following simple wrap around function for gethostbyname. > But the memcpy (i also tried bcopyBytes and bcopy) doesn't work at > all. I always got exceptions. what should i do? > > struct hostent *gethostbyname(char *name) > { > struct hostent *h=NULL; > int addr = hostGetByName(name); > if (addr != -1) > { > h->h_name = name; > h->h_addrtype = AF_INET; > h->h_length = sizeof(addr); > memcpy(h->h_addr,&addr,h->h_length); > > } > return h; > > } --------------------------- Newsgroups: comp.os.vxworks Subject: MTD/tffs: 64 vs.16-bit Buswidth question...? Date: 15 Dec 2001 18:19:05 -0800 From: roswel_ajf@hotmail.com (roswell) Organization: http://groups.google.com/ Message-ID: <6dcae13d.0112151819.47637089@posting.google.com> Hi, I have few MTD/tffs questions: (PPC750 host, Intel 28F128J3A flash, 4 chips (x16) on each bank, 64-bit dbus. And our whole flash memory is mapped to host-memory). 1. Some say, using 64-bit bus width (data bus is 64) might restrict tffs. myth or fact? of course each flash chip is 0x16. 2. i don't understand this flMap(). ( flashPtr = (FlashWPTR)flMap(vol.socket, address) I am told that all chips on both banks are mapped to hostmemory. no sliding window stuff. should flMap() simply return FLASH_BASE_ADDRESS (mutliple of 8)? 3. "CardAddress address", - what's this address if the buswidth is x16 or x64. 4. rfaSetMappingContext() - if there is no sliding window, then what does this do? related to #2 i guess. Thanks... --------------------------- Newsgroups: comp.os.vxworks Subject: Re: T2.1 for Mips MIPS32sfgnu - bootrom/copyLongs Date: 14 Dec 2001 01:08:52 -0800 From: sczhang@esteemnetworks.com (ShiChao Zhang) Organization: http://groups.google.com/ Message-ID: <4ceedd43.0112140108.7f5b3d10@posting.google.com> References: <3BF53559.35B72EFB@siemenscomms.co.uk> Owain Phillips wrote in message news:<3BF53559.35B72EFB@siemenscomms.co.uk>... > Seeing some funnies here in the bootroms I am generating... > > Our BSP orginally used T2.0 and was based on the standard templates and > RC32364 BSPs. I have recently ported to T2.1 for Mips....but overlooked > generating bootroms.... > > I come to try and generate new bootroms and have hit several > problems.... > > romInit.s - The compiler generated wrong code for the relocation macro > that is used to call romStart (Or you could say the earlier compiler > generated extra instructions which allowed the macro to work; nop in > branch delay slot).....got round this one.... I am using the Tornado2.0, and I found the relocation macro very strange. #define RELOC(toreg,address) \ bal 9f; \ 9:; \ la toreg,address; \ addu toreg,ra; \ la ra,9b; \ subu toreg,ra I don't know what effect it can do. > > Now in romInit just about to in the copyLongs - Code generated by the > compiler wrong and performs jumps into hyperspace. > > I am compiling a "Standard" bootInit.c using the makefiles supplied by > wrs and the > compiler is spitting out junk... > > Z:\lp6200Platform\bsp>rm bootInit.o > > Z:\lp6200Platform\bsp>make bootInit.o > ccmips -Be:\Tornado212/host/x86-win32/lib/gcc-lib/ -c -EB -mips2 > -mno-branch-lik > ely -G 0 -fno-builtin -O0 -fvolatile -g -Wall -I/h -I. > -Ie:\Tornado212\target > \config\all -Ie:\Tornado212\target/h -Ie:\Tornado212\target/src/config > -Ie:\Torn > ado212\target/src/drv -DCPU=MIPS32 -DMIPSEB -DCPU_VAR=RC32364 > -msoft-float -DSO > FT_FLOAT e:\Tornado212\target\config\all\bootInit.c > e:\Tornado212\target\config\all\bootInit.c:379: warning: `fillLongs' > defined but > not used > > Anyone have any ideas? Can point me at known problems in this > "new" Tornado distribution??? > > Thanks, > Owain > > -- I don't know whether the version of your BSP is the same with mine, but in my version, I also have this warning :'fillLongs' defined but not used. It is true, because the mips needn't call this routine and the BSP actually don't do this by using conditional defines. --------------------------- Newsgroups: comp.os.vxworks Subject: [One more] About PCMCIA / PC Card Driver Date: 16 Dec 2001 17:39:09 -0800 From: netproa@freechal.com (James) Organization: http://groups.google.com/ Message-ID: <3ad4e9be.0112161739.1433136b@posting.google.com> Configuration CPU : MPC8240 OS : VxWorks PCMCIA Controller: PCI1410A (Texas Instrument) PC Card : ATA-Flash Hello. Now, I am handling vxWorks pcmcia driver. Since vxWorks provide only Intel 82365 pcmcia controller driver, I use PCI1410A's ExCA Compatibility Register for 82365 mode(memory-mapped). Card detect, Ready signals are good. And Read/Write of Status, Card-Status-Change, Power Control register are good. But CIS(Card Information Structure) of PC Card are not read. (only 0xff) I saw MPC860 BSP source code using pcmcia / pc card. But it's some different. MPC860 has own pcmcia controller and it is connected to system bus. And CIS_MEM_START, CIS_REG_START address value is defined statically. My system use PCI pcmcia controller, so I have to allocate some PCI memory for PC-Card's CIS. But I don't know how to map PCI memory to CIS. Please tell me about this problem. Thanks in advance. --------------------------- Newsgroups: comp.os.vxworks Subject: VxWorks Device Driver tutorial Date: 16 Dec 2001 19:18:55 -0800 From: pakernel@yahoo.com (Pa) Organization: http://groups.google.com/ Message-ID: Howdy, I put a vxWorks device driver tutorial. Feel free to mail me your suggestions and improvements. Also, if you are intersted in hiring me, let me know. I have very good skills in drivers, routing, sonet etc., regards Check www.ayyalasoft.com and click downloads. For downloads www.ayyalasoft.com/vxDrv.zip --------------------------- Newsgroups: comp.os.vxworks Subject: Re: what is "makes pointer from integer without a cast"? Date: 17 Dec 2001 04:39:13 -0800 From: Bachmeier.Andreas@foerstergroup.de (Andreas Bachmeier) Organization: http://groups.google.com/ Message-ID: <90019cec.0112170439.42b8227a@posting.google.com> References: <97c3acac.0112162225.a7dbc13@posting.google.com> mimilijie@21cn.com (lijie) wrote in message news:<97c3acac.0112162225.a7dbc13@posting.google.com>... > during the developing time, i often meet with "makes pointer from > integer without a cast". On the left side, there is a pointer, on the right side an integer. So You have forgotten the adressoperator. pPointer = &Integer; Andreas --------------------------- Newsgroups: comp.os.vxworks Subject: Re: query relating the stack size for tasks Date: 19 Dec 2001 13:53:47 -0800 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0112191353.33156cd1@posting.google.com> References: <2ede7703.0112190219.2b22bedc@posting.google.com> Hello, You have a couple of options that I know of: 1) You can painstakingly analyse all your software and work out the stack usage (this is non-trivial, and I don't recommend it) 2) You can run your application for a while, then use checkStack() to get a report on the maximum stack usage for each of your tasks. Either way, I'd add a percentage as a safety margin. Also, be careful, on some architectures the interrupt handlers run using the interrupted task's stack. Check in your architecture's supplement docs for information about the interrupt stacks. HTH, John... kannan@comneti.com (kannan) wrote in message news:<2ede7703.0112190219.2b22bedc@posting.google.com>... > We are currently implementing a stack with various tasks. We have > spawned > various tasks, with the same stack size of 20,000. We are not sure how > to optimize the memory usage. We need to know whether there is a > proper way for deciding the proper stack size for various tasks. This > is critcal in consuming memory. Please explain me, how to decide the > stack size for various tasks? > > regards > kanna --------------------------- Newsgroups: comp.os.vxworks Subject: Re: help!!! how can i operating on the package of raw ethernet or hdlc??? Date: 13 Dec 2001 09:22:13 -0800 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0112130922.77c933ea@posting.google.com> References: Hello, > Attaching C++ interface... succeeded. > Attaching coff OMF reader for ARM CPU family... succeeded. > Warning: Target checksum: 0x961d (computed from 0x1000 to 0xb184c). > Host checksum: 0x270b (computed from 0xf40104 to 0xff0950). > Warning: Core file checksums do not match. > Warning: Symbol .stabstr from section of type STYP_REG. Ignored. > Warning: Symbol .stabstr from section of type STYP_REG. Ignored. > Warning: Symbol .stabstr from section of type STYP_REG. Ignored. > Warning: Symbol .stabstr from section of type STYP_REG. Ignored. > Warning: Symbol .stabstr from section of type STYP_REG. Ignored. > Warning: Symbol .stabstr from section of type STYP_REG. Ignored. > Warning: Symbol .stabstr from section of type STYP_REG. Ignored. > > i have copied my core file of vxWorks image to c:\vxworks > and configuered my target server as the following: > core file :c:\vxWorks > symbol table : global symbol or full symbol. > but i still get the error message above! It is not an error; it is a warning, and you can safely ignore it. All it is telling you is that there are some symbols being ignored; since the symbols are related to the STABS (debug) info, and the target server doesn't use that, this is not a problem. The debugger will understand them, and handle them correctly, for when you want to actually debug the image. HTH, John.. > ! could anybody give me a solution of it for ever? > ! could andbody tell me how to make the file vxworks.st or vxworks.sym? > !if i want to get vxworks.sym ,which command ? > thanks a lot! --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Maximum number of ethernets for VxWorks? Date: 13 Dec 2001 09:26:21 -0800 From: vxmaster@yahoo.com (Mike) Organization: http://groups.google.com/ Message-ID: <2736bd37.0112130926.8fa705e@posting.google.com> References: champu_n@yahoo.com (champak) wrote in message news:... > Could any body tell me what is the maximum number of ethernets( > especially intel 82559 family) that can be supported by Vxworks? > > regards I believe the WRS driver limits you to 8 physical ethernets per kernel. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: memcpy problem Date: Wed, 19 Dec 2001 17:00:14 -0500 From: Doug Dotson Organization: Verio Message-ID: <3C210DEE.6537F8F6@digidata.com> References: You are dereferencing a NULL pointer (h). You need to have h point to something. June wrote: > I wrote the following simple wrap around function for gethostbyname. > But the memcpy (i also tried bcopyBytes and bcopy) doesn't work at > all. I always got exceptions. what should i do? > > struct hostent *gethostbyname(char *name) > { > struct hostent *h=NULL; > int addr = hostGetByName(name); > if (addr != -1) > { > h->h_name = name; > h->h_addrtype = AF_INET; > h->h_length = sizeof(addr); > memcpy(h->h_addr,&addr,h->h_length); > > } > return h; > > } --------------------------- Newsgroups: comp.os.vxworks Subject: how can i work on END driver? Date: 13 Dec 2001 19:29:53 -0800 From: yanh_yang@163.com (lilyang) Organization: http://groups.google.com/ Message-ID: <9dad7e1f.0112131929.7714e479@posting.google.com> I have three END interface,one is ethernet,the others are hdlc port,the unit number is 0,1,2 .The endDevTbl is as follows: END_TBL_ENTRY endDevTbl [] = { #ifdef INCLUDE_SNDS_END { 0, SNDS_LOAD_FUNC_0, SNDS_LOAD_STRING_0, SNDS_LOAN , NULL, FALSE}, #endif /* INCLUDE_SNDS_END */ #ifdef INCLUDE_SNDS_HDLC_END { 1, SNDS_LOAD_FUNC_1, SNDS_LOAD_STRING_1, SNDS_LOAN , NULL, FALSE}, { 2, SNDS_LOAD_FUNC_2, SNDS_LOAD_STRING_2, SNDS_LOAN , NULL, FALSE}, #endif /* INCLUDE_SNDS_HDLC_END */ { 0, END_TBL_END, NULL, 0, NULL, FALSE}, }; Now I want to transmit data by hdlc port,I don't know how to operate on it,and how to point out which hdlc port I want to use in application layer? Thanks anyway! --------------------------- Newsgroups: comp.os.vxworks Subject: Re: [Q] MessageQueue, anybody would look up my source code? Date: 16 Dec 2001 17:34:34 -0800 From: belliny@soback.kornet.net (belliny, Mint~*) Organization: http://groups.google.com/ Message-ID: <6fb8119a.0112161734.48496a30@posting.google.com> References: <6fb8119a.0112122128.5e08cb02@posting.google.com> <488e459a.0112130918.136bc4d7@posting.google.com> thank you for your reply... maybe the error is some kind of null pointer access... anyway, what i wanna make is, main task, user interface / \ / \ / \ / \ +------------+ +--------------+ | | | | | layer 1 | messagequeue1 | | | | <---------- | layer 2 | | | -----------> | | | | messagequeue2 | | +------------+ +--------------+ as my thought, the main task will taskspawn layer1 and layer2, will make messagequeue1/2. at the point of taskspawning, main task pass the messagequeue id to each layer1 and layer2(then each layer could know each other's messagequeue). but, what i can't image is how can each layer is in the listening state while receiving message from other layer... i thought a lot yesterday.. :( [i did~ ㅠㅠ] here it is... void maintask(void) + make two messagequeue1/2 + taskspawn layer1_listener, layer2_listener + provide user interface ( the menu you know [eg. layer1->layer2]) class layer1 + create method - argument : messagequeue + messagesending method - argument : string to send + messagereceiving method - argument : string received class layer2 + same as layer1 layer1_listener + taskspawn layer1(class layer1) + for( ; ; ) listening messagequque1 + when there is a message, layer1->messagereceiving(message) layer2_listener + taskspawn layer2(class layer2) + for( ; ; ) listening messagequque2 + when there is a message, layer2->messagereceiving(message) is this right model? i don't know how to do... if you have any sample code(even it is short), it will be great help for me... thanks in advance... ps. about the working model.. i am just working with downloadable application module for vxWorks... john_94501@yahoo.com (John) wrote in message news:<488e459a.0112130918.136bc4d7@posting.google.com>... > Hello, > > You need to be a lot more precise; what do you mean by "general > failure?" What actually happened? Did you get an exception? Did > nothing happen at all? > > There are certainly some odd things in your code (I would say it is a > little unusual to implement the main loop of a task in an object > constructor for example). > > Some things you might want to look into are: > > 1) Do the two tasks that are calling msgQReceive() get there before > the third task tries to send messages? > > 2) Are all the stack sizes large enough; you've only given them 2000 > bytes which is on the small side, though they don't seem to do much so > perhaps it is OK. You can check this using checkStack() from a shell > (or there is a graphical version somewhere too). > > If none of that helps, please try posting a more detailed explanation > of the actual failure, and info about the CPU type you are developing > for. > > Rgds, > > John... > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: memPartCreate pPool ??? Date: 19 Dec 2001 14:00:59 -0800 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0112191400.664ffc3e@posting.google.com> References: <7421c9fd.0112190242.35de1b4d@posting.google.com> Hello, pPool is a pointer to the block of memory that you are going to create the partition in. There are a number of places you could get this memory, here are some examples: 1) An external memory card, such as a VME memory board 2) Some on-board memory that was not given to the system in sysPhysMemDesc (look in sysLib.c for your BSP's sysPhysMemDesc structure) 3) A block of memory allocated from the main system partition (using malloc() for example), that you wish to use as a separate memory partition. Here's some simple code based on the 3rd example above: PART_ID myPartCreate (void) { char * pPool; PART_ID myPartitionID; if ((pPool = malloc (MY_PART_SIZE)) == NULL) { return NULL; /* Not enough memory for new partition */ } myPartitionID = memPartCreate (pPool, MY_PART_SIZE); return myPartitionID; } This will return the new partition's ID if it worked, or NULL if it didn't. HTH, John... kumar_amit@delhi.tcs.co.in (Amit) wrote in message news:<7421c9fd.0112190242.35de1b4d@posting.google.com>... > In memPartCreate please can anyone throw light on what pPool is? > Is it an input parameter? > I think it is but not very sure. > Also when we create partitions what do we pass as the first parameter? > Is it simply the address of char pPool. > Because in such a case it will take the stack memory, and as soon as > bigger partitions close to 100000 bytes come up the system is > crashing. > i do not understand what the actual problem is ? > > > i have the call as > > memPartCreate(char* pPool, unsigned int) > { > > } > > > now when i call this : > > char pPool; > PART_ID partId; > unsigned int temp = 100000; > > partId = memPartCreate(&pPool,temp); > > or should it be > char* pPool; > > Can you please tell why the system is going out of memory? > And also the detail of the first parameter in memPartCreate > > Amit Kumar --------------------------- Newsgroups: comp.os.vxworks Subject: I can't use setjmp Date: 18 Dec 2001 01:28:18 -0800 From: tsjit@163.com (tom) Organization: http://groups.google.com/ Message-ID: <7674c37.0112180128.5cb9c961@posting.google.com> i make a handle for exception in vxworks,but can't link setjmp,and longjmp. it's show follow in my download: Errors while downloading C:/TORNADO/target/proj/TestTran/SIMNTgnu/TestTran.out: _longjmp _setjmp and show follow inf in my do it until setjmp: Exception number 0: Task: 0x4d3c8a8 (t1) General Protection Fault Program Counter: 0x00000015 Status Register: 0x00010206 value = 0 = 0x0 what happend ,i use tornado 2.0 vxworks in win98.and use simulate. please halp me. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: what is "makes pointer from integer without a cast"? Date: 17 Dec 2001 18:24:45 -0800 From: x-xz@263.net (Xie Xiaozhuo) Organization: http://groups.google.com/ Message-ID: <9ed09ad5.0112171824.6fd52d5b@posting.google.com> References: <97c3acac.0112162225.a7dbc13@posting.google.com> <90019cec.0112170439.42b8227a@posting.google.com> > On the left side, there is a pointer, on the right side an integer. > So You have forgotten the adressoperator. > pPointer = &Integer; > If the programmer wants to assign the address directly represented by the interger value, the warning can be ignored. You can also use a cast such as: int *pPointer = (int *)Interger to get rid of the compiling warning. --------------------------- Newsgroups: comp.os.vxworks Subject: Loading Vxworks from TFFS Date: 17 Dec 2001 15:01:50 -0800 From: cemx_517@yahoo.com (Chris) Organization: http://groups.google.com/ Message-ID: <6e20e2bc.0112171501.901a1a5@posting.google.com> Hi all, When loading vxWorks from tffs, i get the following error; [VxWorks Boot]: @ boot device : tffs=0,0 unit number : 0 processor number : 0 host name : hostname file name : /tffs0/vxWorks inet on ethernet (e) : 10.0.0.2 host inet (h) : 10.0.0.1 user (u) : user ftp password (pw) : password flags (f) : 0x0 target name (tn) : target2 Attaching to TFFS... done. Loading /tffs0/vxWorks... error loading file: status = 0x610001. Error loading file: errno = 0x610001. Please advice!!! Thank you. Chris. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks interview question ! Date: 18 Dec 2001 01:40:32 -0800 From: tsjit@163.com (tom) Organization: http://groups.google.com/ Message-ID: <7674c37.0112180140.4ee37271@posting.google.com> References: I design a tranmsg task for inter commincation of tasks,and one task use one sempore,i don't know if the sempores use too much resource. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Loading Vxworks from TFFS Date: 18 Dec 2001 08:56:28 -0800 From: cemx_517@yahoo.com (Chris) Organization: http://groups.google.com/ Message-ID: <6e20e2bc.0112180856.f07f7e4@posting.google.com> References: <6e20e2bc.0112171501.901a1a5@posting.google.com> <488e459a.0112172101.23d2da2c@posting.google.com> I'm using MPC8260. If I'm missing the ELF boot loader in my bootrom image would it boot of the network? Because I can boot of the network, but I will double check if the ELF loot loader in my bootrom image. Thanks. john_94501@yahoo.com (John) wrote in message news:<488e459a.0112172101.23d2da2c@posting.google.com>... > Hello, > > You don't specify the architecture you are using, but looking up the > error code in my tree shows it as being a loader problem (0x61 -> 97 > -> loadElfLib: you find these in target/h/vwModNum.h in case you want > to look up any others). Then, and in this case it is a little tricky, > the error number 1 looks like it is coming from the generic loadLib, > and is OMF_READER_NOT_INSTALLED. > > That is odd though... you should check to see that you have the ELF > boot loader installed in your bootrom image. Maybe when you added the > TFFS support to the bootrom you somehow lost the ELF support. > > HTH, > > John... > > cemx_517@yahoo.com (Chris) wrote in message news:<6e20e2bc.0112171501.901a1a5@posting.google.com>... > > Hi all, > > > > When loading vxWorks from tffs, i get the following error; > > > > [VxWorks Boot]: @ > > > > boot device : tffs=0,0 > > unit number : 0 > > processor number : 0 > > host name : hostname > > file name : /tffs0/vxWorks > > inet on ethernet (e) : 10.0.0.2 > > host inet (h) : 10.0.0.1 > > user (u) : user > > ftp password (pw) : password > > flags (f) : 0x0 > > target name (tn) : target2 > > > > > > Attaching to TFFS... done. > > Loading /tffs0/vxWorks... > > error loading file: status = 0x610001. > > > > Error loading file: errno = 0x610001. > > > > Please advice!!! Thank you. > > > > Chris. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Loading Vxworks from TFFS Date: 18 Dec 2001 09:00:59 -0800 From: cemx_517@yahoo.com (Chris) Organization: http://groups.google.com/ Message-ID: <6e20e2bc.0112180900.29ca6738@posting.google.com> References: <6e20e2bc.0112171501.901a1a5@posting.google.com> <488e459a.0112172101.23d2da2c@posting.google.com> I've check the map for my bootrom Image and I have the bootElfLib.o linked in my bootrom image. bootElfLib.o needed due to bootElfInit bootElfLib.o 00047224 00000404 2** 2d .text 00001724 00000004 2** 2d .data 000012ac 00000000 2** 0d .bss 00002c7c 00000048 2** 2d .rodata 00000aa0 00000020 2** 0d .comment john_94501@yahoo.com (John) wrote in message news:<488e459a.0112172101.23d2da2c@posting.google.com>... > Hello, > > You don't specify the architecture you are using, but looking up the > error code in my tree shows it as being a loader problem (0x61 -> 97 > -> loadElfLib: you find these in target/h/vwModNum.h in case you want > to look up any others). Then, and in this case it is a little tricky, > the error number 1 looks like it is coming from the generic loadLib, > and is OMF_READER_NOT_INSTALLED. > > That is odd though... you should check to see that you have the ELF > boot loader installed in your bootrom image. Maybe when you added the > TFFS support to the bootrom you somehow lost the ELF support. > > HTH, > > John... > > cemx_517@yahoo.com (Chris) wrote in message news:<6e20e2bc.0112171501.901a1a5@posting.google.com>... > > Hi all, > > > > When loading vxWorks from tffs, i get the following error; > > > > [VxWorks Boot]: @ > > > > boot device : tffs=0,0 > > unit number : 0 > > processor number : 0 > > host name : hostname > > file name : /tffs0/vxWorks > > inet on ethernet (e) : 10.0.0.2 > > host inet (h) : 10.0.0.1 > > user (u) : user > > ftp password (pw) : password > > flags (f) : 0x0 > > target name (tn) : target2 > > > > > > Attaching to TFFS... done. > > Loading /tffs0/vxWorks... > > error loading file: status = 0x610001. > > > > Error loading file: errno = 0x610001. > > > > Please advice!!! Thank you. > > > > Chris. --------------------------- Newsgroups: comp.os.vxworks Subject: how to build a bootable vxworks image? Date: 18 Dec 2001 09:21:08 -0800 From: dragonli@yahoo.com (June) Organization: http://groups.google.com/ Message-ID: I have an IXP1200EB. I am trying to build a bootable vxworks image to include resolvLib (i modified configAll.h). When i used Tornado to "create a new bootable vxworks image (custom configured)", and used the BSP coming with the board (../ixp1200/boardsupport/vxworks/ixp1200eb", it gave this following error when config is parsing. invalid command name "::toolFindByFile" I don't know what happened. Can anybody point out the problem for me? I really appreciate it. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Loading Vxworks from TFFS Date: 18 Dec 2001 09:13:49 -0800 From: cemx_517@yahoo.com (Chris) Organization: http://groups.google.com/ Message-ID: <6e20e2bc.0112180913.4d182c71@posting.google.com> References: <6e20e2bc.0112171501.901a1a5@posting.google.com> <488e459a.0112172101.23d2da2c@posting.google.com> I used the shell to printed the errno and I get the following, - -> printErrno 0x610001 0x610001 = S_loadElfLib_HDR_READ value = 0 = 0x0 But I can load the same image from the network. john_94501@yahoo.com (John) wrote in message news:<488e459a.0112172101.23d2da2c@posting.google.com>... > Hello, > > You don't specify the architecture you are using, but looking up the > error code in my tree shows it as being a loader problem (0x61 -> 97 > -> loadElfLib: you find these in target/h/vwModNum.h in case you want > to look up any others). Then, and in this case it is a little tricky, > the error number 1 looks like it is coming from the generic loadLib, > and is OMF_READER_NOT_INSTALLED. > > That is odd though... you should check to see that you have the ELF > boot loader installed in your bootrom image. Maybe when you added the > TFFS support to the bootrom you somehow lost the ELF support. > > HTH, > > John... > > cemx_517@yahoo.com (Chris) wrote in message news:<6e20e2bc.0112171501.901a1a5@posting.google.com>... > > Hi all, > > > > When loading vxWorks from tffs, i get the following error; > > > > [VxWorks Boot]: @ > > > > boot device : tffs=0,0 > > unit number : 0 > > processor number : 0 > > host name : hostname > > file name : /tffs0/vxWorks > > inet on ethernet (e) : 10.0.0.2 > > host inet (h) : 10.0.0.1 > > user (u) : user > > ftp password (pw) : password > > flags (f) : 0x0 > > target name (tn) : target2 > > > > > > Attaching to TFFS... done. > > Loading /tffs0/vxWorks... > > error loading file: status = 0x610001. > > > > Error loading file: errno = 0x610001. > > > > Please advice!!! Thank you. > > > > Chris. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Loading Vxworks from TFFS Date: 17 Dec 2001 23:25:25 -0800 From: robin_ch@263.net (Robin Tcheng) Organization: http://groups.google.com/ Message-ID: References: <6e20e2bc.0112171501.901a1a5@posting.google.com> cemx_517@yahoo.com (Chris) wrote in message news:<6e20e2bc.0112171501.901a1a5@posting.google.com>... > Hi all, > > When loading vxWorks from tffs, i get the following error; > > [VxWorks Boot]: @ > > boot device : tffs=0,0 > unit number : 0 > processor number : 0 > host name : hostname > file name : /tffs0/vxWorks > inet on ethernet (e) : 10.0.0.2 > host inet (h) : 10.0.0.1 > user (u) : user > ftp password (pw) : password > flags (f) : 0x0 > target name (tn) : target2 > > Attaching to TFFS... done. > Loading /tffs0/vxWorks... > error loading file: status = 0x610001. > > Error loading file: errno = 0x610001. > > Please advice!!! Thank you. > > Chris. Can you boot from ethernet using that vxWorks? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Please Help Me Date: 17 Dec 2001 04:48:10 -0800 From: sandeep@eventhelix.com (Sandeep) Organization: http://groups.google.com/ Message-ID: <9d8984ab.0112170448.e36d94@posting.google.com> References: <6c32a22b.0112120223.1cfc1bb2@posting.google.com> You can find articles about Embedded Systems at the following URL: http://www.EventHelix.com/ Sandeep --------------------------- Newsgroups: comp.os.vxworks Subject: Re: DosFs 2.0 Product Information Date: 18 Dec 2001 08:47:17 -0800 From: cjohnson@oresis.com (Charles Johnson) Organization: http://groups.google.com/ Message-ID: References: We've had no end of trouble with dosFS2.0. File corruption where the size of files would suddening become -1. We had to go to recompiling the dosFs2.0 source at a lower optimization level to get it to work. (WRS finally issued SPR #71089 for this.) I would like to see WRS do a more fault resilient file system than DOS. Charles Johnson Oresis Communications, Inc. cjohnson@oresis.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: updating a module on VxWorks Date: 18 Dec 2001 21:18:57 -0800 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0112182118.cedcdc8@posting.google.com> References: <1cdb3add.0112181022.358c9262@posting.google.com> Hello, RTFM... unld "some_pure_library.lib". The '<' is a redirection operator. Why would you want to redirect the contents of the file you are unloading onto standard input for the unld function??? The redirection is a shorthand for loading. You could also use ld "some_pure_library.lib" to load if you need symmetry ;-) Note: the module must have been loaded using ld for this to work. You cannot unload modules that are statically linked into the bootable image. HTH, John... sriva_t@yahoo.com (Srivathsa) wrote in message news:<1cdb3add.0112181022.358c9262@posting.google.com>... > Hi, > During development of a module, when we want to load a *updated* > module (i.e some.lib) onto target, do we need to reboot the target? > > Can we use the "unld" command of target shell, to unload the old > some.lib, and then use the "ld" command? > > I tried a "unld < some_pure_library.lib", but it didn't work (i.e > I could still locate the symbols in the symbol table). > > Thanks > -Srivathsa --------------------------- Newsgroups: comp.os.vxworks Subject: DHCP hostname option Date: 18 Dec 2001 14:02:16 -0800 From: agraham@openglobe.net (Aaron Graham) Organization: http://groups.google.com/ Message-ID: <95ebbe31.0112181402.2c55cb8c@posting.google.com> I haven't been able to find any information on the hostname option for DHCP requests. Some cable service providers (@Home, for one) require that you send a pre-determined hostname with your DHCP request in order to register your machine correctly, kind of like a username. In linux, the client is configured/started as follows: dhcpcd -h However, I'm not sure how it's done in vxWorks. There is a _DHCPC_HOSTNAME_TAG as well as a _DHCP_CLIENT_ID_TAG (from what I can tell, "hostname" is a specific kind of "client id"). But my dhcpcOptionSet() function only has 2 parameters, contrary to all documentation I've been able to find. This doesn't allow me to set any values for the tags I throw at the dhcpcOptionSet() function. I do seem to have dhcpcOptionAdd(), which has parameters that could possibly take values for a hostname, but there is absolutely no documentation in existence for that function call. I'm going to throw parameters around arbitrarily (typical vxworks programming style) until I get it to work or someone replies to this thread... Thanks in advance for your help. Aaron --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Need help - Unresolved symbol error while downloading Date: 18 Dec 2001 21:04:26 -0800 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0112182104.687bc340@posting.google.com> References: <7674c37.0112180200.73f9ff11@posting.google.com> Hello, OK, for setjmp/longjmp you need to include the component INCLUDE_SIGNALS into your kernel. (From the project facility, this will probably be called SIGNALS - it knows to drop the INCLUDE_ prefix, but from the command line/BSP you'll need the prefix). For the timers, add INCLUDE_POSIX_TIMERS (and probably also INCLUDE_POSIX_CLOCKS and INCLUDE_POSIX_SIGNALS and INCLUDE_SIGNALS since they are needed by the timers I think). HTH, John... tsjit@163.com (tom) wrote in message news:<7674c37.0112180200.73f9ff11@posting.google.com>... > and i have the same problem too. > but i use simulate ,tornado2.0 for win98.and my produre setjmp,and > longjmp does't link. > please tell me how do i do.please give much more detailed . --------------------------- Newsgroups: comp.os.vxworks Subject: memPartCreate pPool ??? Date: 19 Dec 2001 02:42:59 -0800 From: kumar_amit@delhi.tcs.co.in (Amit) Organization: http://groups.google.com/ Message-ID: <7421c9fd.0112190242.35de1b4d@posting.google.com> In memPartCreate please can anyone throw light on what pPool is? Is it an input parameter? I think it is but not very sure. Also when we create partitions what do we pass as the first parameter? Is it simply the address of char pPool. Because in such a case it will take the stack memory, and as soon as bigger partitions close to 100000 bytes come up the system is crashing. i do not understand what the actual problem is ? i have the call as memPartCreate(char* pPool, unsigned int) { } now when i call this : char pPool; PART_ID partId; unsigned int temp = 100000; partId = memPartCreate(&pPool,temp); or should it be char* pPool; Can you please tell why the system is going out of memory? And also the detail of the first parameter in memPartCreate Amit Kumar --------------------------- Newsgroups: comp.os.vxworks Subject: ipAttach problem ? Date: 19 Dec 2001 04:23:42 -0800 From: sardine_sz@163.net (sardine) Organization: http://groups.google.com/ Message-ID: hi all gurus, my target have two ethernetcard(INTEL82257),i use T2 Vxworks5.4. i boot the target by network.and i had increase all IP_MAX_UNITS=2, and add extern END_TBL in configNet.h. after boot success, i can attach TCP/IP stack to second END interface. my console print follow information: - ------------------------------------------------------------------------ - -> muxShow Device: fei Unit: 0 Description: Intel 82557 Ethernet Enhanced Network Driver Protocol: Wind Debug Agent Type: 257 Recv 0x14df48 Shutdown 0x0 Protocol: IP 4.4 ARP Type: 2054 Recv 0x1385f0 Shutdown 0x138860 Protocol: IP 4.4 TCP/IP Type: 2048 Recv 0x1385f0 Shutdown 0x138770 Device: fei Unit: 1 Description: Intel 82557 Ethernet Enhanced Network Driver value = 0 = 0x0 - -> - -> ipAttach(1,"fei") 0xfecebfc (tShell): Protocol is out of space. Increase IP_MAX_UNITS. value = -1 = 0xffffffff = _func_taskRegsShowRtn + 0xffe020f3 - -> - ------------------------------------------------------------------------ why ipAttach() fail? thx! --------------------------- Newsgroups: comp.os.vxworks Subject: query relating the stack size for tasks Date: 19 Dec 2001 02:19:00 -0800 From: kannan@comneti.com (kannan) Organization: http://groups.google.com/ Message-ID: <2ede7703.0112190219.2b22bedc@posting.google.com> We are currently implementing a stack with various tasks. We have spawned various tasks, with the same stack size of 20,000. We are not sure how to optimize the memory usage. We need to know whether there is a proper way for deciding the proper stack size for various tasks. This is critcal in consuming memory. Please explain me, how to decide the stack size for various tasks? regards kanna --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Loading Vxworks from TFFS Date: 18 Dec 2001 08:52:06 -0800 From: cemx_517@yahoo.com (Chris) Organization: http://groups.google.com/ Message-ID: <6e20e2bc.0112180852.eddbf87@posting.google.com> References: <6e20e2bc.0112171501.901a1a5@posting.google.com> Yes, I can boot of the net using the same image. robin_ch@263.net (Robin Tcheng) wrote in message news:... > cemx_517@yahoo.com (Chris) wrote in message news:<6e20e2bc.0112171501.901a1a5@posting.google.com>... > > Hi all, > > > > When loading vxWorks from tffs, i get the following error; > > > > [VxWorks Boot]: @ > > > > boot device : tffs=0,0 > > unit number : 0 > > processor number : 0 > > host name : hostname > > file name : /tffs0/vxWorks > > inet on ethernet (e) : 10.0.0.2 > > host inet (h) : 10.0.0.1 > > user (u) : user > > ftp password (pw) : password > > flags (f) : 0x0 > > target name (tn) : target2 > > > > Attaching to TFFS... done. > > Loading /tffs0/vxWorks... > > error loading file: status = 0x610001. > > > > Error loading file: errno = 0x610001. > > > > Please advice!!! Thank you. > > > > Chris. > > Can you boot from ethernet using that vxWorks? --------------------------- Newsgroups: comp.os.vxworks Subject: help a newbie Date: 19 Dec 2001 13:21:13 -0800 From: arohb@cc.usu.edu (Aroh Barjatya) Organization: http://groups.google.com/ Message-ID: Hi i am developing a device driver for a custom made camera board. I am trying to read the SRAM on the board(which contains the image) and write it to a file on my host system(P4 1.6G) But i am unable to create a file in my C program. It refuses to create a file. I wonder if its trying to create a file on the target rather then on the host system. Please tell me how to do it. thanx aroh --------------------------- Newsgroups: comp.os.vxworks Subject: SNMP Module as an Application Date: 19 Dec 2001 02:33:54 -0800 From: pritamganguly@hotmail.com (Pritam) Organization: http://groups.google.com/ Message-ID: <166770ed.0112190233.4471e517@posting.google.com> Hi all I want to configure the WindNet SNMP Agent as an downloadable application rather then an service that starts up at boot time. Is there any way to configure the agent to meet up to my requirement. Thanks in advance Pritam --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks interview question ! Date: 18 Dec 2001 20:28:37 -0800 From: agraham@openglobe.net (Aaron Graham) Organization: http://groups.google.com/ Message-ID: <95ebbe31.0112182028.67a18952@posting.google.com> References: <23de5426.0112100602.4bf0c4ac@posting.google.com> <3c152217$0$2729$724ebb72@reader2.ash.ops.us.uu.net> <3c191f2c$0$20175$724ebb72@reader2.ash.ops.us.uu.net> <3c1fb96e$0$22741$724ebb72@reader2.ash.ops.us.uu.net> The sad quality of the questions (and answers given) tells me only that Jones doesn't seem to have a very high bar for his programmers. A lot of the answers he gives seem awfully shortsighted... 1. "This [ternary] operator exists in C because it allows the compiler to produce more optimal code than an if-then-else sequence." Baloney! Although I personlly love the ternary operator, there's no reason why even bad compilers can't optimize an if-then-else sequence just as well. 2. Programmers prefer for(;;){E} to while(1){E} because the consensus is that more compilers will optimize the for loop better (in other words, they don't check the conditional every time). Of course, this could go either way. It seems obvious to me why a programmer should prefer one construct over the other, especially after taking into consideration which compiler he's using. 3. Maybe a better question about #define directives would be, "Why is it good to severely limit their use?" Understanding where and how a compiler will optimize is a much better practice than throwing #defines all over the place. In my practice, the bugs that I've had the most trouble tracking down are the ones that are caused by overuse of #define. As has been said time and again, "Learn the language, don't redefine it!" I could go on like this... I would assume most college sophomores in CS could have answered at least 14 of these questions correctly. Contrary to Jones's remarks (and considering the other valid objections raised in this thread), I would NOT recommend giving this test as part of an interview to would-be embedded programmers. Aaron --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks AE pentium differences Date: 15 Dec 2001 09:10:26 -0800 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0112150910.629ced96@posting.google.com> References: <3c19bc1f$0$7113$4d4efb8e@news.be.uu.net> Hello, "Mukul Dhankhar" wrote in message news:<3c19bc1f$0$7113$4d4efb8e@news.be.uu.net>... > Hi all, > I want to use VxWorks AE 1.1 for a new project.My target hardware is > Pentium-MMX 200 MHz. Does anyone know if I can use VxWorks Pentium 3 with a > PcPentium 3 BSP on my P1 target? Does VxWorks for P-3 make some specialized > functionalities for P-3 which it does not make for P-1? I suspect that P1 would work on P3, but I would be more hesitant about going the other way. Why not just buy the P1 license and BSP instead? Since it is a new project, you'll need a new license anyway... Rgds, John... > Thanks, > Mukul --------------------------- Newsgroups: comp.os.vxworks Subject: Crpy Key error in Tornado Prototyper Installation Date: 19 Dec 2001 01:10:42 -0800 From: anubhavsaksena@yahoo.com (Anubhav Saksena) Organization: http://groups.google.com/ Message-ID: <9ee41a72.0112190110.5147b67f@posting.google.com> Hi All, I am facing this error during tornado 2.0 installation. This error occurs ("Crpy Key License service unable to start"). I have a Tornado 2 (Downloaded Version of SIMNT). I have a Windows NT host system. When I install using setup file, the installation completes before updating libSIMNTgnuvx.a file the computer gives this 'Crpy Error'. This downloaded tornado prototyper is downloaded on dec2000. I am trying to use the same prototyper instead of using the new version. Please advice Thanks Anubhav --------------------------- Newsgroups: comp.os.vxworks Subject: Re: serial debugging with Tornado 2.0 fails Date: 14 Dec 2001 21:43:24 -0800 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0112142143.774bd526@posting.google.com> References: <9vd7fl$cl3$07$1@news.t-online.com> Hello, Markus Pietrek wrote in message news:<9vd7fl$cl3$07$1@news.t-online.com>... > Hi folks, > > I'm sure it's an old problem but I can't find any answers on google. I > currently try to debug a ARM processor with Tornado 2.0 with the serial WDB > backend. Target and host are configured for 9600/8/1 and their ports are > working. The target prints "WDB Ready" after booting, the shell displays > "WDB: Ready" in the banner, too. That just means that the target agent is running; a good sign of course, but not an indication of the mode it is using, nor the port it is connected to. > Echo mode is working, too. What do you mean by this? Echo on which port? I am assuming that you are using two serial connections here: one for the console output, and one for the WDB connection. If you have a terminal program connected to the WDB serial port, your host will not be able to communicate with the target over that port. You're setup should look like this to support both console and WDB serial mode: +-----------+ +--------+ | Host | | Target | | | TTY | | | ttya/COM1 |<------>|/tyCo/0 | | | | | | | WDB | | | ttyb/COM2 |<------>|/tyCo/1 | | | | | +-----------+ +--------+ On the host, run your terminal program (e.g. tip or hyperterminal) connected to /dev/ttya or COM1: (depending on the OS ;-). Setup your target server to use a connection on /dev/ttyb or COM2: and enable its verbose mode to see what it does. > And the host is > sending packets over serial which I can't read as they are binary. But the > target server can't connect. it always prints "Error: rpccore backend > client Timed out". This happens both on the first or second serial line. > > The lines configuring the WDB are > " > #undef WDB_COMM_TYPE > #define WDB_COMM_TYPE WDB_COMM_SERIAL > > > #undef WDB_TTY_CHANNEL > #define WDB_TTY_CHANNEL 1 > #undef WDB_TTY_BAUD > #define WDB_TTY_BAUD 9600 > " This looks correct to have WDB running on /tyCo/1 on the target at 9600 - a little slow for real use, but OK to get things running. Once you have it working, you'll probably want to speed this up a little. I assume that you have two serial ports on the target... > Network debugging is fine but for my problem I can't use it. > Has anyone solved this problem yet? Well, it is not a problem because it really does work, if a little slowly. Some more things for you to look at in your setup: 1) Check in the IDE that you have selected serial mode for the target server too (or on the command line if you are starting the target server from the command line of course). 2) Check the speed setting for the target server (again, this should be available in the IDE somewhere). 3) Check that you have selected the correct serial port on your host for the target server too. 4) Consider switching on the verbose mode for the target server; that should give you more information about what it is up to. HTH, John... > Thanks in advance, > Markus Pietrek --------------------------- Newsgroups: comp.os.vxworks Subject: I can not write or erase flash, why? Date: 16 Dec 2001 20:28:37 -0800 From: x-xz@263.net (Xie Xiaozhuo) Organization: http://groups.google.com/ Message-ID: <9ed09ad5.0112162028.6883a5fc@posting.google.com> Now my flash is filled with zeros and I got problems: I can not write anything to it. Erasing has no effect too. But with another piece of flash, I can write and erase. What is my problem? --------------------------- Newsgroups: comp.os.vxworks Subject: updating a module on VxWorks Date: 18 Dec 2001 10:22:49 -0800 From: sriva_t@yahoo.com (Srivathsa) Organization: http://groups.google.com/ Message-ID: <1cdb3add.0112181022.358c9262@posting.google.com> Hi, During development of a module, when we want to load a *updated* module (i.e some.lib) onto target, do we need to reboot the target? Can we use the "unld" command of target shell, to unload the old some.lib, and then use the "ld" command? I tried a "unld < some_pure_library.lib", but it didn't work (i.e I could still locate the symbols in the symbol table). Thanks - -Srivathsa --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Booting VxWorks AE 1.1 Date: 18 Dec 2001 21:13:39 -0800 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0112182113.13385fdc@posting.google.com> References: <3c1f3403$0$7110$4d4efb8e@news.be.uu.net> Hello, "Mukul Dhankhar" wrote in message news:<3c1f3403$0$7110$4d4efb8e@news.be.uu.net>... > Hi all, > I am using VxWorks AE 1.1 for the first time( never used any previous > versions of it too). My target hardware is Pentium-MMX. I run Tornado on a > WinNT host. I need to create C applications and run them on my target. First > I boot the target with a boot floppy created using 'mkboot' utility. By > default the boot.txt file needs to be downloaded through ethernet (using > FTP). MY QUESTION IS: how do I come to know the IP address of my target? You assign the address in the boot parameters. You will need to connect a serial cable between your host COM1: port and your target's COM1: port. Start something like hyperterminal on your host, set to 9600 8N1, and power up the target system. You'll get a count down after a while in the hyperterminal window; press space when you see this. Now type p to see the current settings. Change these to set the parameters for your network. If you want to use DHCP, you can, but you'll need to build your own boot diskette as DHCP is not in the default image. When you do this, you can also change the other settings so that they'll be remembered over power downs (PC's don't have non-volatile memory that can be used by VxWorks). For more information on this, read the manuals that came with your package. There is information about setting up the serial connection, configuring the boot parameters and booting a target. I don't have the manuals handy at the moment, so I can't point you at the sections, but look in the Tornado AE User's Guide and the Getting Started guide at least. > There is DHCP server access on the network. Is there another( and probably > easier way) to boot the target? It may be through serial connection or it > may be that the boot image in floppy contains all the source files( written > in C) statically linked so there is no need for an ethernet connection. If you want, you can boot off a hard disk easily; floppy booting is possible too of course, but they're not very big so that might limit the usefulness of that strategy. You'll want the network anyway though as debugging without it, while possible, is painful. It is worth getting the network running (and not too difficult). If you're still not getting very far, call your local sales rep. and ask if a field support person can drop by to help you get set up. HTH, John... > Can anyone help me with this, because I am new to VxWorks. > Thanks in advance, > Mukul. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: double shell parameter Date: 13 Dec 2001 23:39:48 -0800 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0112132339.76c95a57@posting.google.com> References: <3C1932CE.9B32513D@fnal.gov> Hello, The shells do not support floating point parameter passing. If you absolutely need FP parameters, try writing a wrapper function that takes strings and converts them. HTH, John... Dinker Charak wrote in message news:<3C1932CE.9B32513D@fnal.gov>... > Hi All, > > I am using VxWorks 5.4 > > If have this function with two double parameters, inside the function I > can only access the last parameters value and that too is stored in > first parameter. > > void DblParams(double d1, double d2) > { > printf("\tDouble Input 1 is f=%.8f\n", d1); > printf("\tDouble Input 2 is f=%.8f\n", d2); > return; > } > > -> DblParams 1.1,5.5 > Double Input 1 is f=5.50000000 > Double Input 2 is f=NaN > > Once I have a double in parameter list, all subsquent parameters get > messed up. Why is this so? > > Regards, > Dinker --------------------------- Newsgroups: comp.os.vxworks Subject: Re: alignment error in RISC cpu executing the vxworks TCP/IP stack Date: 16 Dec 2001 11:09:44 -0800 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0112161109.3bf5a3ce@posting.google.com> References: <4ceedd43.0112160042.788d816d@posting.google.com> Hello, This is a very old, and very well known issue. Most, if not all, of the newer END drivers for VxWorks should have a parameter in their load string that allows you to correct for this. The parameter is usually called offset; take a look in the manual for dec21x40End for an example. For cases where the CPU cannot access unaligned memory (most RISC chips), set the offset to 2. If you are writing your own driver you'll need to handle this yourself, and present the network stack with a correctly aligned IP address. HTH, John... sczhang@esteemnetworks.com (ShiChao Zhang) wrote in message news:<4ceedd43.0112160042.788d816d@posting.google.com>... > Hi all, > > I am using the vxworks runing on arm and mips cpu. During debugging > the TCP/IP stack, I found a severe problem hard to solve.(Maybe I > missed some important key issue). > > The vxworks TCP/IP stack use the MBlk to store the incoming ethernet > packets, but when parsing the header info, it convert a pointer > pointing to the MBlk starting address to a structure type which is > just the same field as ethernet packet stucture. By using this > structure, the stack extracts the destination/source MAC address/Ip > address and type/length. Due to the MBlk starting address is 32bit > address aligned, so the destination ip address is not 32bit aligned. > But the stack only use a int type variable to get the structure's ip > address memeber which caused a address alignment error. So in my > arm/mips cpu board, it can't resovle the correct ip address which is > exactly its own. > > This situation is always existed if you want to get a 32bit wide > integer from a non 32bit aligned address. By the way, some cpu ,such > as mips, have a exception named address alignment error which occurs > when a 32bit operation access a non 32bit alignment address. And I > don't know how to handle this problem. Or maybe I missed some > important issue? > > Thanks for any helps! > > shichao zhang > Esteem Networks --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Need help - Unresolved symbol error while downloading Date: 18 Dec 2001 02:00:55 -0800 From: tsjit@163.com (tom) Organization: http://groups.google.com/ Message-ID: <7674c37.0112180200.73f9ff11@posting.google.com> References: and i have the same problem too. but i use simulate ,tornado2.0 for win98.and my produre setjmp,and longjmp does't link. please tell me how do i do.please give much more detailed . --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Loading Vxworks from TFFS Date: 17 Dec 2001 21:01:18 -0800 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0112172101.23d2da2c@posting.google.com> References: <6e20e2bc.0112171501.901a1a5@posting.google.com> Hello, You don't specify the architecture you are using, but looking up the error code in my tree shows it as being a loader problem (0x61 -> 97 - -> loadElfLib: you find these in target/h/vwModNum.h in case you want to look up any others). Then, and in this case it is a little tricky, the error number 1 looks like it is coming from the generic loadLib, and is OMF_READER_NOT_INSTALLED. That is odd though... you should check to see that you have the ELF boot loader installed in your bootrom image. Maybe when you added the TFFS support to the bootrom you somehow lost the ELF support. HTH, John... cemx_517@yahoo.com (Chris) wrote in message news:<6e20e2bc.0112171501.901a1a5@posting.google.com>... > Hi all, > > When loading vxWorks from tffs, i get the following error; > > [VxWorks Boot]: @ > > boot device : tffs=0,0 > unit number : 0 > processor number : 0 > host name : hostname > file name : /tffs0/vxWorks > inet on ethernet (e) : 10.0.0.2 > host inet (h) : 10.0.0.1 > user (u) : user > ftp password (pw) : password > flags (f) : 0x0 > target name (tn) : target2 > > > Attaching to TFFS... done. > Loading /tffs0/vxWorks... > error loading file: status = 0x610001. > > Error loading file: errno = 0x610001. > > Please advice!!! Thank you. > > Chris. --------------------------- Newsgroups: comp.os.vxworks Subject: Tornado for Managed Switches Date: 18 Dec 2001 11:54:41 -0800 From: judi@advancedtechno.com (Judi) Organization: http://groups.google.com/ Message-ID: Software Engineering Position Southern New Jersey Real-time embedded software developer Vxworks, Power PC experience Tornado for Managed Switches (TMS) Please send resume to judi@advancedtechno.com --------------------------- Newsgroups: comp.os.vxworks Subject: Flash and the mv5100 Date: 14 Dec 2001 11:06:43 -0800 From: google@dharty.com (David) Organization: http://groups.google.com/ Message-ID: <5a87bef4.0112141106.3de2015b@posting.google.com> Has anyone used the AM29LV323 Flash that it available on the mv5100 (PPC604)WITHOUT TFFS? Windriver doesn't support it, but it must be possible. Specifically I need to be able to download a new boot image to DRAM, and then blow it into the FLASH. Using the PPCBug is not an option as that would require the user to remove the board and change jumpers. Heck, just a starting address for the Flash memory space would be great! Thanks David --------------------------- Newsgroups: comp.os.vxworks Subject: alignment error in RISC cpu executing the vxworks TCP/IP stack Date: 16 Dec 2001 00:42:28 -0800 From: sczhang@esteemnetworks.com (ShiChao Zhang) Organization: http://groups.google.com/ Message-ID: <4ceedd43.0112160042.788d816d@posting.google.com> Hi all, I am using the vxworks runing on arm and mips cpu. During debugging the TCP/IP stack, I found a severe problem hard to solve.(Maybe I missed some important key issue). The vxworks TCP/IP stack use the MBlk to store the incoming ethernet packets, but when parsing the header info, it convert a pointer pointing to the MBlk starting address to a structure type which is just the same field as ethernet packet stucture. By using this structure, the stack extracts the destination/source MAC address/Ip address and type/length. Due to the MBlk starting address is 32bit address aligned, so the destination ip address is not 32bit aligned. But the stack only use a int type variable to get the structure's ip address memeber which caused a address alignment error. So in my arm/mips cpu board, it can't resovle the correct ip address which is exactly its own. This situation is always existed if you want to get a 32bit wide integer from a non 32bit aligned address. By the way, some cpu ,such as mips, have a exception named address alignment error which occurs when a 32bit operation access a non 32bit alignment address. And I don't know how to handle this problem. Or maybe I missed some important issue? Thanks for any helps! shichao zhang Esteem Networks --------------------------- Newsgroups: comp.os.vxworks Subject: what is "makes pointer from integer without a cast"? Date: 16 Dec 2001 22:25:39 -0800 From: mimilijie@21cn.com (lijie) Organization: http://groups.google.com/ Message-ID: <97c3acac.0112162225.a7dbc13@posting.google.com> during the developing time, i often meet with "makes pointer from integer without a cast". i can't figure out what's the matter with my code. can anyone who had ever met with it tell me something about it? thank you! --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to verify Cache mode at run-time Date: 18 Dec 2001 05:43:43 -0800 From: kannan@comneti.com (kannan) Organization: http://groups.google.com/ Message-ID: <2ede7703.0112180543.2773d474@posting.google.com> References: hi raymond, Actually speaking in copy-back mode, when you do a write operation to the memory only the cache is updated and do not update the external memory. whereas in write through mode, when you do a write operation to the memory it gets reflected in the external memory and in the cache(depending on the Hit/Miss). It depends on your application to use any of the modes. The copy-back mode is preferred becoz of minimal external bus utilization. If you see the register DC_CST there is a bit called DFWT if it is set then it is in write-through mode , if it is not then it is in copy-back mode. regards Kannan ryeung@earthlink.net (Raymond Yeung) wrote in message news:... > I try looking up 8260 manual for some sort of > configuration registers that I can read, but in > no vail. > > Does anyone has any idea? I want to read it at > run-time to ensure that it is set to what I want. > In particular, I want to know whether copy-back > vs. write-through is used. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to verify Cache mode at run-time Date: 19 Dec 2001 15:20:09 -0800 From: ryeung@earthlink.net (Raymond Yeung) Organization: http://groups.google.com/ Message-ID: References: <2ede7703.0112180543.2773d474@posting.google.com> Hi Kannan, Thanks for your reply. I know the basic distinction between the two cache modes. I'm looking for a way to confirm at run-time that the board has been configured correctly as I expect it. I try looking up DC_CST and DFWT, both in MPC8260 manual and in 603e manual. I also search my BSP, but only find it in 860 specified code. Otherwise I couldn't find them. Where are they? I suppose I need some sort of address to read them? Another thing is that, my project is not configured to use full MMU, but some "partial" MMU setup. Would the system default to copy-back mode? Thanks, Raymond kannan@comneti.com (kannan) wrote in message news:<2ede7703.0112180543.2773d474@posting.google.com>... > hi raymond, > > Actually speaking in copy-back mode, when you do a write operation to > the memory only the cache is updated and do not update > the external memory. > > whereas in write through mode, when you do a write operation > to the memory it gets reflected in the external memory > and in the cache(depending on the Hit/Miss). > > It depends on your application to use any of the modes. > The copy-back mode is preferred becoz of minimal external bus > utilization. > > If you see the register DC_CST there is a bit called > DFWT if it is set then it is in write-through mode , > if it is not then it is in copy-back mode. > > > regards > Kannan > > ryeung@earthlink.net (Raymond Yeung) wrote in message news:... > > I try looking up 8260 manual for some sort of > > configuration registers that I can read, but in > > no vail. > > > > Does anyone has any idea? I want to read it at > > run-time to ensure that it is set to what I want. > > In particular, I want to know whether copy-back > > vs. write-through is used. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: memcpy problem Date: Thu, 20 Dec 2001 09:56:51 +1030 From: Tim Shaw Organization: Defence Science and Technology Organisation Message-ID: <3C21223B.ADF2D9B@dsto.defence.gov.au> References: With your printf statement, try instead: printf("address of h: %p\n", h); which will print the value of 'h' which is actually a pointer (not &h). I think you'll find it's zero as you've set it to NULL. June wrote: > i thought it probably because of NULL too. but i tried to print out > the address of h and i actually used this: > printf("address of h:0x%x\n",&h); > > and see this: > address of h:0x508fc0 > > That means it was allocated some space for h. > And also i tried > struct hostent *h=malloc(sizeof(struct hostent)); > and i got the exception: > > Exception number 16: Task: 0x509044 (t6) > > Data abort > Program Counter: 0x1b320 > Status Register: 0x80000013 > 902dc _vxTaskEntry +28 : 42f54 () > 42ff8 _wdbFuncCallLibInit+128: _gethostbyname () > 1364fc _gethostbyname +b8 : _memcpy () > value = 0 = 0x0 > > This looks like a quite simple problem. but i just can not figure it > out. :( --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to verify Cache mode at run-time Date: Thu, 20 Dec 2001 02:13:24 GMT From: "dgold1" Organization: Road Runner Message-ID: <8PbU7.15414$AI.8759814@typhoon.san.rr.com> References: <2ede7703.0112180543.2773d474@posting.google.com> Raymond Yeung wrote in message ... >Hi Kannan, > >I try looking up DC_CST and DFWT, both in MPC8260 manual >and in 603e manual. I also search my BSP, but only find >it in 860 specified code. Otherwise I couldn't find them. You won't find them. They are 860 specific registers. >Another thing is that, my project is not configured to use >full MMU, but some "partial" MMU setup. Would the system >default to copy-back mode? If #INCLUDE_MMU_FULL is defined, it expects you to have the (add on, extra $$$) pkg VxVMI installed. Otherwise, INCLUDE_MMU_BASIC will be defined. The cache mode is set with a #define in configAll.h, and can be overwritten in your bsp specific config.h. The cache mode you choose only affects those memory spaces covered in sysPhysMemDesc, i.e. those using page table entries. To see the "real, run-time" mode, you need to get the PTE for your desired memory address and look at the WIMG bits. IIRC, there's a call or 2 you can make vmXXXGet that may get a PTE for you, but I can't remember for sure. Since you're on a PPC that also supports [ID]BATS, the only way to for sure know is to read the BAT register and examine the WIMG bits there. You can access them with the usual mfspr code, each BAT has a unique SPR. Of course if you've set the appropriate values in the sysBatDesc[] array, you could always examine the entries in the array, and assume WRS set the BATS correctly (They do). You could always put a global variable in the code based upon the #define in configAll.h, i.e. #if (dcache_mode == COPYBACK) int dcacheMode=1; #elif (dcache_mode == WRITETHROUGH) int dcacheMode=2; #endif Don't forget vxHid0Get() to figure out if the caches are even enabled... hth, Dan Gold gold@ensemble.com >kannan@comneti.com (kannan) wrote in message news:<2ede7703.0112180543.2773d474@posting.google.com>... >> hi raymond, >> >> Actually speaking in copy-back mode, when you do a write operation to >> the memory only the cache is updated and do not update >> the external memory. >> >> whereas in write through mode, when you do a write operation >> to the memory it gets reflected in the external memory >> and in the cache(depending on the Hit/Miss). >> >> It depends on your application to use any of the modes. >> The copy-back mode is preferred becoz of minimal external bus >> utilization. >> >> If you see the register DC_CST there is a bit called >> DFWT if it is set then it is in write-through mode , >> if it is not then it is in copy-back mode. >> >> >> regards >> Kannan >> >> ryeung@earthlink.net (Raymond Yeung) wrote in message news:... >> > I try looking up 8260 manual for some sort of >> > configuration registers that I can read, but in >> > no vail. >> > >> > Does anyone has any idea? I want to read it at >> > run-time to ensure that it is set to what I want. >> > In particular, I want to know whether copy-back >> > vs. write-through is used. > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: UGL? Date: 19 Dec 2001 19:25:19 -0800 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0112191925.4bc96fc5@posting.google.com> References: <1I3U7.4272$XC5.4776@www.newsranger.com> Hello, Bill Pringlemeir wrote in message news:... > Vic> Are there any "quick start" guides available? Does anyone have > Vic> any hints/tips (except "don't use UGL" ;-)? > > Platform? CPU, display, mouse, touch screen, keyboard, etc. Also, are you using WindML (UGL) directly, or do you have Zinc or Java too? > Vic> There's not a lot of time for this project, so I need to get up > Vic> to speed quite quickly. > > Read the manual. If you are familiar with graphics programming (such > as Windows GDI, OpenGL, X-Windows, etc) a lot of concepts will be > familiar. > > A lot depends on the feature set/application that you need to use. At > least read the WindML manuals that give an overview. Also, take a look at the examples under target/src/ugl/examples. They're really more like little code excerpts showing you how to do various things with WindML. If you have more specific questions after that, feel free to post them here - there's a few regulars that have been using WindML and might be able to help out. Just remember to give us lots of info about your platform and the problem you are seeing. HTH, John... --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Where is PTY Driver Date: 19 Dec 2001 19:27:05 -0800 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0112191927.260251f3@posting.google.com> References: <9vql4m$h1e0h$1@ID-48686.news.dfncis.de> Hello, In AE it is INCLUDE_OS_PTYDRV that you need, probably the same, or at least similar - search for it in config/all/configAll.h. HTH, John... "Dan Pinson" wrote in message news:<9vql4m$h1e0h$1@ID-48686.news.dfncis.de>... > The PTY driver does not seem to be in my VxWorks image (lkup "pty" finds no > symbols). Is there a configuration option I need to include? > > Thanks, > Dan --------------------------- Newsgroups: comp.os.vxworks Subject: Re: problem with 82559er driver Date: 19 Dec 2001 20:51:06 -0800 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0112192051.f43b438@posting.google.com> References: <3C20C4B3.A9463A@nlr.nl> Hello, > I'm having a problem with the latest 82559er driver in vxWorks 5.4 for a > Pentium target. > As long as I don't attach the network cable the driver init's ok. > (ifShow "eeE" returns the correct values, all routines are there) Did you really mean "eeE" as the name? Where did this driver come from? Normally , 82559 devices are called "fei0" (the number will change if there is more than one network card in the target of course). > After I attach the cable I get a repeating message: > > " (0x1fa05230 (tNetTask): Panic: where are those rfd buffers " Sounds like there's a problem with the load string and/or the driver initialisat ion. The number of RFDs is normally specified in the END load string (look in co nfigNet.h for your BSP in most cases). Which arch/BSP are you using? HTH, John... > Apparently receive frame descriptors are needed. > Does anyone know what to do in this case? > > TIA, michiel meekes --------------------------- Newsgroups: comp.os.vxworks Subject: UGL(not WindML)+ZINC developer .... Re: UGL? Date: Thu, 20 Dec 2001 04:58:41 GMT From: "ChanSeong Mike Kim" Organization: Dacom Message-ID: <5eeU7.8577$M3.23159@news.bora.net> References: <1I3U7.4272$XC5.4776@www.newsranger.com> Hi ~~ If you're more familiar with Windows programming or C++, ZINC(upper layer of UGL) would be more helpful to build a application. Directory target/src/ugl/demos , target/src/zinc/demos directory is helpful. But you have to prepare developing environment. Regards ChanSeong Kim "Vic" wrote in message news:1I3U7.4272$XC5.4776@www.newsranger.com... > Hi All. > > I've just been tasked with doing some development with UGL. > > Are there any "quick start" guides available? Does anyone have any hints/tips > (except "don't use UGL" ;-)? > > There's not a lot of time for this project, so I need to get up to speed quite > quickly. > > Ta muchly! > > Vic. > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Configuring PCI devices!! Date: Wed, 19 Dec 2001 17:01:43 -0500 From: George Varndell Organization: Posted Via Binaries.net = SPEED+RETENTION+COMPLETION = http://www.binaries.net Message-ID: <3C210E47.6D893140@varndellengineering.com> References: <1e752e41.0112140302.4c72ae7b@posting.google.com> Gulnaz wrote: > > I wanted to know what steps u need to follow to configure a PCI device > with vxworks os. > All I know is that u have to make some changes to the BSP(i may be > wrong) Have a look at pciConfigLib.c and pciAutoConfigLib.c. That'll get you started. These are files that WindRiver distributes in source form. Regards, George Varndell - -- Varndell Engineering, LLC. PPC and XScale Expertise The VE 82559 END driver demo version is now available. Compare the performance difference -- free, no obligation. Send email request to driver@varndellengineering.com. ______________________________________________________________________________ Posted Via Binaries.net = SPEED+RETENTION+COMPLETION = http://www.binaries.net --------------------------- Newsgroups: comp.os.vxworks Subject: Re: what is "makes pointer from integer without a cast"? Date: Wed, 19 Dec 2001 17:04:23 -0500 From: George Varndell Organization: Posted Via Binaries.net = SPEED+RETENTION+COMPLETION = http://www.binaries.net Message-ID: <3C210EE7.F689EC83@varndellengineering.com> References: <97c3acac.0112162225.a7dbc13@posting.google.com> lijie wrote: > > during the developing time, i often meet with "makes pointer from > integer without a cast". i can't figure out what's the matter with my > code. can anyone who had ever met with it tell me something about it? > thank you! int i; char *p; i = 0x12345678; p = i; That makes a pointer from an integer without a cast. p = (char *) i; That doesn't. Regards, George Varndell - -- Varndell Engineering, LLC. PPC and XScale Expertise The VE 82559 END driver demo version is now available. Compare the performance difference -- free, no obligation. Send email request to driver@varndellengineering.com. ______________________________________________________________________________ Posted Via Binaries.net = SPEED+RETENTION+COMPLETION = http://www.binaries.net --------------------------- Newsgroups: comp.os.vxworks Subject: Re: alignment error in RISC cpu executing the vxworks TCP/IP stack Date: Wed, 19 Dec 2001 17:13:16 -0500 From: George Varndell Organization: Posted Via Binaries.net = SPEED+RETENTION+COMPLETION = http://www.binaries.net Message-ID: <3C2110FB.DBB3357F@varndellengineering.com> References: <4ceedd43.0112160042.788d816d@posting.google.com> <488e459a.0112161109.3bf5a3ce@posting.google.com> John wrote: > > This is a very old, and very well known issue. Most, if not all, of > the newer END drivers for VxWorks should have a parameter in their > load string that allows you to correct for this. The parameter is > usually called offset; take a look in the manual for dec21x40End for > an example. For cases where the CPU cannot access unaligned memory > (most RISC chips), set the offset to 2. Yes, but every END driver I've seen handled this by copy the entire frame 2 bytes further into the packet buffer. For some devices, this is unavoidable on alignment challenged processors. For most, it could be handled better. > > If you are writing your own driver you'll need to handle this > yourself, and present the network stack with a correctly aligned IP > address. If you're writing your own, and your device supports it, have it store the MAC header in one buffer and the rest of the frame in another, properly aligned, buffer. This eliminates the copying I mentioned above. Regards, George Varndell - -- Varndell Engineering, LLC. PPC and XScale Expertise The VE 82559 END driver demo version is now available. Compare the performance difference -- free, no obligation. Send email request to driver@varndellengineering.com. ______________________________________________________________________________ Posted Via Binaries.net = SPEED+RETENTION+COMPLETION = http://www.binaries.net --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Loading Vxworks from TFFS Date: Thu, 20 Dec 2001 08:51:49 +0100 From: "Michael Lawnick" Organization: Buergernetz Dillingen Message-ID: <9vs5ap$1bn$1@snoopy.bndlg.de> References: <6e20e2bc.0112171501.901a1a5@posting.google.com> <488e459a.0112172101.23d2da2c@posting.google.com> <6e20e2bc.0112180913.4d182c71@posting.google.com> Reply-To: "Michael Lawnick" Sender: mlawnick@pd9e657c7.dip.t-dialin.net Hi Chris, S_loadElfLib_HDR_READ is from function bootLoadModule(). It seems to have detected an error in your local image. My 2 Cents: Do checksum test with both, local and host copy and verify. I'd guess, you got a failure on copying the image from host to target. - -- 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 ============================================== "Chris" schrieb im Newsbeitrag news:6e20e2bc.0112180913.4d182c71@posting.google.com... > I used the shell to printed the errno and I get the following, > > -> printErrno 0x610001 > 0x610001 = S_loadElfLib_HDR_READ > value = 0 = 0x0 > > But I can load the same image from the network. > > john_94501@yahoo.com (John) wrote in message news:<488e459a.0112172101.23d2da2c@posting.google.com>... > > Hello, > > > > You don't specify the architecture you are using, but looking up the > > error code in my tree shows it as being a loader problem (0x61 -> 97 > > -> loadElfLib: you find these in target/h/vwModNum.h in case you want > > to look up any others). Then, and in this case it is a little tricky, > > the error number 1 looks like it is coming from the generic loadLib, > > and is OMF_READER_NOT_INSTALLED. > > > > That is odd though... you should check to see that you have the ELF > > boot loader installed in your bootrom image. Maybe when you added the > > TFFS support to the bootrom you somehow lost the ELF support. > > > > HTH, > > > > John... > > > > cemx_517@yahoo.com (Chris) wrote in message news:<6e20e2bc.0112171501.901a1a5@posting.google.com>... > > > Hi all, > > > > > > When loading vxWorks from tffs, i get the following error; > > > > > > [VxWorks Boot]: @ > > > > > > boot device : tffs=0,0 > > > unit number : 0 > > > processor number : 0 > > > host name : hostname > > > file name : /tffs0/vxWorks > > > inet on ethernet (e) : 10.0.0.2 > > > host inet (h) : 10.0.0.1 > > > user (u) : user > > > ftp password (pw) : password > > > flags (f) : 0x0 > > > target name (tn) : target2 > > > > > > > > > Attaching to TFFS... done. > > > Loading /tffs0/vxWorks... > > > error loading file: status = 0x610001. > > > > > > Error loading file: errno = 0x610001. > > > > > > Please advice!!! Thank you. > > > > > > Chris. --------------------------- Newsgroups: comp.os.vxworks Subject: VxWorks IP and Port Scanner Software Date: Thu, 20 Dec 2001 04:23:42 GMT From: "magicmoments4" Organization: AT&T Broadband Message-ID: Does anyone know how to run a windows based IP and Port Scanner on a VxWorks OS? I am trying to figure out all the IP addresses and TCP port numbers on a Cisco ONS 15454 metro optical platform at work. If anyone knows how to do this, please e-mail me at william.thomas@lglass.net Sincerely, William Thomas --------------------------- Newsgroups: comp.os.vxworks Subject: Re: alignment error in RISC cpu executing the vxworks TCP/IP stack Date: 19 Dec 2001 23:58:06 -0800 From: sczhang@esteemnetworks.com (ShiChao Zhang) Organization: http://groups.google.com/ Message-ID: <4ceedd43.0112192358.7eede5@posting.google.com> References: <4ceedd43.0112160042.788d816d@posting.google.com> <488e459a.0112161109.3bf5a3ce@posting.google.com> john_94501@yahoo.com (John) wrote in message news:<488e459a.0112161109.3bf5a3ce@posting.google.com>... > Hello, > > This is a very old, and very well known issue. Most, if not all, of > the newer END drivers for VxWorks should have a parameter in their > load string that allows you to correct for this. The parameter is > usually called offset; take a look in the manual for dec21x40End for > an example. For cases where the CPU cannot access unaligned memory > (most RISC chips), set the offset to 2. > Due to my old END dirver version, the parameter doesn't exist. And I try to get a offset of 2 by myself in driver, it did works! Thanks a lot! > If you are writing your own driver you'll need to handle this > yourself, and present the network stack with a correctly aligned IP > address. > > HTH, > > John... > > > sczhang@esteemnetworks.com (ShiChao Zhang) wrote in message news:<4ceedd43.0112160042.788d816d@posting.google.com>... > > Hi all, > > > > I am using the vxworks runing on arm and mips cpu. During debugging > > the TCP/IP stack, I found a severe problem hard to solve.(Maybe I > > missed some important key issue). > > > > The vxworks TCP/IP stack use the MBlk to store the incoming ethernet > > packets, but when parsing the header info, it convert a pointer > > pointing to the MBlk starting address to a structure type which is > > just the same field as ethernet packet stucture. By using this > > structure, the stack extracts the destination/source MAC address/Ip > > address and type/length. Due to the MBlk starting address is 32bit > > address aligned, so the destination ip address is not 32bit aligned. > > But the stack only use a int type variable to get the structure's ip > > address memeber which caused a address alignment error. So in my > > arm/mips cpu board, it can't resovle the correct ip address which is > > exactly its own. > > > > This situation is always existed if you want to get a 32bit wide > > integer from a non 32bit aligned address. By the way, some cpu ,such > > as mips, have a exception named address alignment error which occurs > > when a 32bit operation access a non 32bit alignment address. And I > > don't know how to handle this problem. Or maybe I missed some > > important issue? > > > > Thanks for any helps! > > > > shichao zhang > > Esteem Networks --------------------------- Newsgroups: comp.os.vxworks Subject: Re: how to build a bootable vxworks image? Date: Thu, 20 Dec 2001 09:05:40 +0100 From: "Michael Lawnick" Organization: Buergernetz Dillingen Message-ID: <9vs65e$1jq$1@snoopy.bndlg.de> References: Reply-To: "Michael Lawnick" Sender: mlawnick@pd9e657c7.dip.t-dialin.net Hi June, I'm not really sure, but AFAIK I crosse the same problem on trying to setup a workspace for BSP whose makefile had set CPU to a type that Tornado didn't know. Check whether $(WIND_BASE)/host/resource/target/architecturedb has an entry for your CPU-type setting - and whether it is valid 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 ============================================== "June" schrieb im Newsbeitrag news:d48fa963.0112180921.1ea6b82f@posting.google.com... > I have an IXP1200EB. I am trying to build a bootable vxworks image to > include resolvLib (i modified configAll.h). When i used Tornado to > "create a new bootable vxworks image (custom configured)", and used > the BSP coming with the board > (../ixp1200/boardsupport/vxworks/ixp1200eb", it gave this following > error when config is parsing. > > invalid command name "::toolFindByFile" > > I don't know what happened. Can anybody point out the problem for me? > I really appreciate it. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: tffs problem Date: Thu, 20 Dec 2001 09:19:41 +0100 From: "Michael Lawnick" Organization: Buergernetz Dillingen Message-ID: <9vs6uv$1p3$1@snoopy.bndlg.de> References: <9vq71g$2iqp$1@gavrilo.mtu.ru> Reply-To: "Michael Lawnick" Sender: mlawnick@pd9e657c7.dip.t-dialin.net Hi Shrike Are you able to access TFFS from shell (booting from fd,hd,net,...)? - -- 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 ============================================== "Shrike" schrieb im Newsbeitrag news:9vq71g$2iqp$1@gavrilo.mtu.ru... > Hi! > > After we boot from disk-on-chip with options: boot device - tffs=0,0, boot > file - /tffs0/vxworks, etc. we get "tffsDevCreate failed" after "Attaching > to TFFS..." message. > > MMU was allocated for bios extension memory region (in our case c800:0000). > > What can be done in this case? > Thanks in advance. > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: help a newbie Date: 20 Dec 2001 00:23:05 -0800 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0112200023.3abbaae3@posting.google.com> References: Hello, arohb@cc.usu.edu (Aroh Barjatya) wrote in message news:... > Hi i am developing a device driver for a custom made camera board. > I am trying to read the SRAM on the board(which contains the image) > and write it to a file on my host system(P4 1.6G) > > But i am unable to create a file in my C program. It refuses to create > a file. I wonder if its trying to create a file on the target rather > then on the host system. If the code is running on the target, then, yes, it will be trying to create the file on the target. Unfortunately, you have chosen (or had forced on you) a very poor choice of development host OS I suspect (since it is a P4 host, I am guessing Windoze). That limits your choices of file system that can be used by the target. You also didn't add much information about what happened, or how you were trying to create the file. Please try to post more detailed information in future if you have questions; it makes it much simpler to answer them... Onto file systems, the simplest is called netDrv, and if you booted the target from the host via a network connection it should already be available. Try the devs command from a WindShell (I think it works on the host shell too - I normally configure the target shell and use that). If all is well, you should see a device named after your boot host with a ':' character after it. If that's working, try a simple piece of test code: #include "vxWorks.h" #include "ioLib.h" #include "errno.h" #include "stdio.h" void simpleTest (char * name) { int fd; if ((fd = open (name, O_RDWR, 0)) == ERROR) { printf ("Error opening file %#x\n", errno); return; } printf ("File descriptor: %d\n", fd); } First, find out the directory on your PC that the ftp user you used to boot the target sees initially. You'll need to look in the configuration of the ftp service on the host to find this, or work it out from the path you put in the bootline in the target. Next, try to open a small file that already exists there. If that works, close the file from a windshell session. Finally, make a quick change to the call to open; change 'O_RDWR' to 'O_RDWR | O_CREAT' and rebuild. Then try with a name that doesn't exist. That should allow you create a new file. Note that it won't appear on the host though until you close it. Why? Well, because netDrv is an FTP file system (or RSH, but that won't work on a toy OS like Windoze). The file is created in memory on the target, and held there all the time it is open. When you close the file, the file system opens an ftp connection to the host and copies the file up to it. This has a couple of downsides: 1) You need to keep the whole file in RAM, so if your file is big you can quickly hit problems 2) If you don't close the file, or the target crashes before you close it, the data is lost Remember, netDrv is designed for development use, not as a production file system. The other development file system you might be able to use is called TSFS (not to be confused with TFFS). TSFS is the "target server file system" and is a file system that works using the WDB connection to your host. This might be better for your requirements since you're using a PC as a host; check in the docs for information about configuring and using it. > Please tell me how to do it. HTH, John... > thanx > aroh --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Need help - Unresolved symbol error while downloading Date: 20 Dec 2001 00:38:06 -0800 From: tsjit@163.com (tom) Organization: http://groups.google.com/ Message-ID: <7674c37.0112200038.6a6bd997@posting.google.com> References: <7674c37.0112180200.73f9ff11@posting.google.com> <488e459a.0112182104.687bc340@posting.google.com> i use simulate in windows,and follow your advice,i include component " operation system componets ->POSIX->POSIX signals "and include " operation system componets ->kernel components->signals" in my vxsim" vxworks.exe" but i still see "Errors while downloading C:/TORNADO/target/proj/TestTran/SIMNTgnu/TestTran.out: _longjmp _setjmp "when i download my .out file.are you give some advice with more details i am a new student in vxworks,i need help. thank you john_94501@yahoo.com (John) wrote in message news:<488e459a.0112182104.687bc340@posting.google.com>... > Hello, > > OK, for setjmp/longjmp you need to include the component > INCLUDE_SIGNALS into your kernel. (From the project facility, this > will probably be called SIGNALS - it knows to drop the INCLUDE_ > prefix, but from the command line/BSP you'll need the prefix). > > For the timers, add INCLUDE_POSIX_TIMERS (and probably also > INCLUDE_POSIX_CLOCKS and INCLUDE_POSIX_SIGNALS and INCLUDE_SIGNALS > since they are needed by the timers I think). > > HTH, > > John... > > tsjit@163.com (tom) wrote in message news:<7674c37.0112180200.73f9ff11@posting.google.com>... > > and i have the same problem too. > > but i use simulate ,tornado2.0 for win98.and my produre setjmp,and > > longjmp does't link. > > please tell me how do i do.please give much more detailed . --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Flash and the mv5100 Date: Thu, 20 Dec 2001 08:37:01 +0000 From: Jeremy Coker Organization: Land & Sea Systems Message-ID: <3C21A32D.4359AF51@bae.co.uk> References: <5a87bef4.0112141106.3de2015b@posting.google.com> Assumptions: Using VXWorks BSP latest MVME5101 card with 16megs of soldered flash + 1 meg socketed flash There are several problems here, the first one may not be relevant but I will mention it anyway. 1). Motorola quote the start address for their flash as being FF000000 if under or equal to 8megs and F4000000 for anything over 8megs, which is true if you are going to be setting up the physical start address for flash (e,g, in the PCBUG and the pflash command: for example: pflash 4000:FFF00 FF000100 for less than or equal to 8megs OR pflash 4000:FFF00 F4000100 for greater than 8megs of flash This will work fine as the start address, However MOTOROLA's pre-release BSP (NOT vxworks BSP) has the start address of flash (i'm using 16megs of flash) as being FF0000 which is because the physical address has been masked with a value of FF and not F4 Looking at the vxworks BSP (latest one) in the mvme5100.h file it shows the start address as: F4000000 which will not work since (because of the FF mask used) so this needs to be changed to FF000000; this part may not be relevant to you if your flash is 8megs or less. 2). The flash chips used according to our card (mvme5101) are: AM29DL323G and not the AM29LV which seem to be used on the 2400's we have hope this helps Regards Jeremy --------------------------- Newsgroups: comp.os.vxworks Subject: what is the difference between bootrom and bootalbe vxworks image Date: 20 Dec 2001 00:58:16 -0800 From: jlddotcn@hotmail.com (juliedan) Organization: http://groups.google.com/ Message-ID: I can burn either bootrom or my bootalbe wxworks image into the flash, why do I need bootrom? Does all device driver is a bsp? what is bsp really mean? All my thanks in advance. Liedan Ju --------------------------- Newsgroups: comp.os.vxworks Subject: Re: i/o redirection Date: Thu, 20 Dec 2001 10:57:54 +0200 From: "Ofer Goren" Organization: Verio Message-ID: References: <9vqd56$s2a$1@snoopy.bndlg.de> can u elaborate? - -- Ofer Goren Iamba Technologies "Michael Lawnick" wrote in message news:9vqd56$s2a$1@snoopy.bndlg.de... > Hi Ofer, > > you definitely can redirect the I/O of your shell. Use ioTaskStdSet(). > The problem: > tShell will pend in its old fd until next character input. You will have to > send one to free it. > > Use a pipe for your redirected shell and you can pass all unhandled commands > from your CLI. > > 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$la8log$04o$1@lnews.actcom.co.il... > > Well, I guess this is the only acceptable solution. Why can't I just > > re-direct my I/O to a different task.....? well, only windriver's GODs > can > > give this answer, I guess.... > > > > Anyway, Thanks... > > > > -- > > Ofer Goren > > Iamba Technologies > > "Johan Borkhuis" wrote in message > > news:Xns917B8033F1A58borkhuisagerecom@135.7.153.102... > > > "Ofer Goren" wrote: > > > > > > > I have a CLI task. I want all input and output to be re-directed to > > > > this task, instead of to the shell task. in fact, I can kill the shell > > > > task, but I want the user to be able to access the shell upon a > > > > specific command. > > > > > > The only way to do this is to kill the shell. All commands can also be > > > executed from your CLI, but you have to lookup the symbol and call the > > > function from within the CLI. You can also startup the shell when the > user > > > wants access to the 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: checksum verification - embedded Date: Thu, 20 Dec 2001 11:11:47 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <3C1E161D.A3E58D5F@avoidspam.com> Before deeloping FlashLib, we have agonized over the question to do LFS or FTL, and one of the big design questions for FTL was: to attempt to re-write a sector or not. It seemed to me that if attempting an in-place reqrite, quite too often it will turn out that it failed (some bits needed to transition from 0 to 1), like in over 90% of attempts the sector write operation needs to be done twice (first in-pace then in a new location). This turns out an unacceptible reduction of write performance, which rules out write in-place of sector data. So a nice CRC would be then quite feasible. Not sure about TFFS though. In-place bit transitions 1 to 0 are used throughout overhead structures in FlashLib an in most FTLs. "Bill Pringlemeir" wrote in message news:uk7vj3zft.fsf@yahoo.com... > >>>>> "Leonid" == Leonid Rosenboim writes: > [snip] > Leonid> errors beyod the control of the internal checking. Also, > Leonid> Flash translation layers (such as TrueFFS) may or may not add > Leonid> checksum per block of data. > > It would be unusual for a Flash layer to use a standard CRC or > checksums. This negates the possibility to `re-write' ones data to > zero data for the entire block. At the very least, an alternative > validation method must be used. > > regards, > Bill Pringlemeir. > > -- > vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" --------------------------- Newsgroups: comp.os.vxworks Subject: UDP Receive and send problem Date: Thu, 20 Dec 2001 10:09:16 +0000 From: Jeremy Coker Organization: Land & Sea Systems Message-ID: <3C21B8CC.3ECC052E@bae.co.uk> Hi, the configuration I have is: mvme2400 (although I need this to work on a mvm5101; there seems to be a bug in windrivers bsp for the 5101 that causes it not to send broadcast messages). windriver BSP Problem: Whenever I send broadcast udp messages from one 2400 to another 2400, I do not appear to get any data loss on the first run of the program. However after running this program a second time I get a data loss on the very last datagram/message (any ideas why this is happening); I have found a poor solution that is not of much use to me, which is to reboot the cards and run the program again, this then sends and receives all the packets o.k. Why would this be ? I've tried all the suggestions of increasing the buffer size and so on, but this does not appear to have any affect. This happens at any time of the day (so cannot be due to network traffic etc...) Jeremy --------------------------- Newsgroups: comp.os.vxworks Subject: Re: DosFs 2.0 Product Information Date: Thu, 20 Dec 2001 12:18:46 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: I dont have access to SPRs now, not even to the "public" version, but from your description of the work around, I might try to guess that this has something to do with 64 bit math. DosFs 2.0, uses 64 bit integer arithmetics, and does need the GCC and associated libraries to perform these operations correctly. Trouble is that it is probably the only module in VxWorks which uses 64 bit math, and therefore this functionality is not properly tested after a product has been compiled, and may fail on certain architectures. If you finally got around by recompiling, this would suggest there is nothing wrong with the file system code itself, as the source code did not change when you recompiled. Would appreciate more details, as I still care about this creation of mine. - - Leonid "Charles Johnson" wrote in message news:fa9c3eaa.0112180847.20ca2f7@posting.google.com... > We've had no end of trouble with dosFS2.0. File corruption where the > size of files would suddening become -1. We had to go to recompiling > the dosFs2.0 source at a lower optimization level to get it to work. > (WRS finally issued SPR #71089 for this.) > > I would like to see WRS do a more fault resilient file system than > DOS. > > Charles Johnson > Oresis Communications, Inc. > cjohnson@oresis.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: DosFs 2.0 Product Information Date: Thu, 20 Dec 2001 12:21:58 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <3c20bca4$1@brateggebdc5.br-automation.co.at> "Werner Schiendl" wrote in message news:3c20bca4$1@brateggebdc5.br-automation.co.at... > > "Leonid Rosenboim" wrote in message > news:newscache$wcilog$ogo$1@lnews.actcom.co.il... > ... > > > > My only conclusion is that WRS marketing did not improve much in the last > 3 > > years. > > > > As - so it seems - has the file system. > > - Werner > Apparently, as the original team that developed dosFs2 has left WindRiver in the beginning of 1999. I have headed the team and resigned from WindRiver in January 1999. But I do beleive that dosFs2 is an excellent product, many people are using it without a problem, and I am still trying to figure out what went wrong in those cases where people had problems. - - Leonid --------------------------- Newsgroups: comp.os.vxworks Subject: Re: debug problem with Tornado2 Date: 20 Dec 2001 11:14:02 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: <90019cec.0112130645.39d0a62b@posting.google.com> Bachmeier.Andreas@foerstergroup.de (Andreas Bachmeier) wrote: > Is there a way to debug an application witch use interrupts? You could try to use system mode instead of task mode debugging, see the WDB agent components in your project for this. But this will also introduce other problems. Another way is to build an extra task that performs the buffering between you ISR and the oeprational SW. 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: what is the difference between bootrom and bootalbe vxworks image Date: Thu, 20 Dec 2001 12:06:55 +0100 From: "Werner Schiendl" Message-ID: <3c21c651$3@brateggebdc5.br-automation.co.at> References: Hi, "juliedan" wrote in message news:d6763629.0112200058.29e9dadf@posting.google.com... > I can burn either bootrom or my bootalbe wxworks image into the flash, > why do I need bootrom? The bootrom is normally used to load the actual VxWorks image with your application. Depending on the hardware, you may be able to load VxWorks image directly, thus saving the space (and time) to run the bootrom. This would not (easily) be possible e. g. on x86 platform, since the system can only load max. ~600kB bootrom from realmode. Then the bootrom switches to protected mode and is able to load VxWorks image of any size. > Does all device driver is a bsp? what is bsp really mean? bsp -> board support package It contains drivers, boot code, settings, configuration, build tools, make rules, etc. to run VxWorks on your particular board. BSPs are normally provided by board vendors for their particular hardware. hth Werner --------------------------- Newsgroups: comp.os.vxworks Subject: Re: I can't use setjmp Date: Thu, 20 Dec 2001 11:55:18 +0100 From: "Werner Schiendl" Message-ID: <3c21c651$1@brateggebdc5.br-automation.co.at> References: <7674c37.0112180128.5cb9c961@posting.google.com> Hi Tom, this is a known problem, which also happens to affect UGL (the jpeglib part, to be precise). See SPR #28158 on WindSurf, or contact your Support / FAE representative to get a patch. hth Werner "tom" wrote in message news:7674c37.0112180128.5cb9c961@posting.google.com... > i make a handle for exception in vxworks,but can't link setjmp,and longjmp. > it's show follow in my download: > Errors while downloading C:/TORNADO/target/proj/TestTran/SIMNTgnu/TestTran.out: > _longjmp > _setjmp > > and show follow inf in my do it until setjmp: > Exception number 0: Task: 0x4d3c8a8 (t1) > > General Protection Fault > Program Counter: 0x00000015 > Status Register: 0x00010206 > > value = 0 = 0x0 > > what happend ,i use tornado 2.0 vxworks in win98.and use simulate. > please halp me. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: config two END on vxworks target? Date: 20 Dec 2001 11:21:36 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: sardine_sz@163.net (sardine) wrote: > usrNetIfConfig("elPci",1, "192.168.0.113","MyHost",0) /*Failed > ,page fault*/ Take a look at the file usrNetConfigIf.c in target/config/comps/src/net, this file contains the source for the function usrNetIfConfig. 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: telnet banner Date: Thu, 20 Dec 2001 10:34:37 -0000 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: <3c1e2cfb.268186882@news.ibercom.com> Jaime Ontiyuelo Tirapu wrote in message <3c1e2cfb.268186882@news.ibercom.com>... >VME 68K v. 8.4(alpha): 15 Feb 2001 ^^^^^^^^^^ DaveK - -- Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! --------------------------- Newsgroups: comp.os.vxworks Subject: Re: tffs problem Date: Thu, 20 Dec 2001 14:51:54 +0300 From: "Shrike" Organization: MTU-Intel ISP Message-ID: <9vsjbv$2a9l$1@gavrilo.mtu.ru> References: <9vq71g$2iqp$1@gavrilo.mtu.ru> <9vs6uv$1p3$1@snoopy.bndlg.de> Hello! > Are you able to access TFFS from shell (booting from fd,hd,net,...)? I've booted over network and wrote a simple program to check tffs availability. I've called usrTffsConfig with parameters 0,0,"\tffs0\" but it wrote "tffsDevCreate failed". --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Thu Dec 20 08:03:07 2001 From: Donald McLachlan Date: Thu Dec 20 08:03:09 PST 2001 Subject: lspci Has anyone written something similar, or ported linux's lspci to vxWorks? I'm asking because I'm probably going to have to write a driver for a PCI device ... and need to be able to scan for it's location. something like lspci seems to have the functionality needed to locate the device base address so I can find the registers, and it shows the IRQ the device uses. Don From vxwexplo-errs@csg.lbl.gov Thu Dec 20 08:08:18 2001 From: Donald.McLachlan@crc.ca Date: Thu Dec 20 08:08:21 PST 2001 Subject: lspci Has anyone written something similar, or ported linux's lspci to vxWorks? I'm asking because I'm probably going to have to write a driver for a PCI device ... and need to be able to scan for it's location. something like lspci seems to have the functionality needed to locate the device base address so I can find the registers, and it shows the IRQ the device uses. Don From vxwexplo-errs@csg.lbl.gov Thu Dec 20 10:30:45 2001 From: Donald.McLachlan@crc.ca Date: Thu Dec 20 10:30:47 PST 2001 Subject: Re: lspci Someone answering another post has inadvertantly answered my question ... so now I have a follow-on question. When my PC boots it shows my device id, vendor ie, pci bus. Using this info I can probe the bus and it finds my device. Then when I specify this info to get the IRQ / IOaddr info, I get all -1 (ff or ffff) values. Any idea why this is happening? Thanks, Don P.S. This PC has linux loaded on it. If I remove the vxWorks boot floppy and boot linux from the hard disk, lspci shows the IRQ and IO and IOmem addresses ... so the hardware seems OK. From vxwexplo-errs@csg.lbl.gov Fri Dec 21 04:03:18 2001 From: Vxworks Exploder Date: Fri Dec 21 04:03:20 PST 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Fri Dec 21 04:03:13 PST 2001 Subject: VxWorks AE 1.1: Ethernet Card Configuration Problem Subject: about usrNetwork.c ? Subject: Embedding C application in boot image of VxWorks AE 1.1 Subject: Re: help a newbie.. Subject: Re: help a newbie.. Subject: Re: DHCP hostname option Subject: Re: DosFs 2.0 Product Information Subject: Re: problem with 82559er driver Subject: Re: how can i use file system on flash? Subject: Re: i/o redirection Subject: Re: tffs problem Subject: Re: memPartCreate pPool ??? Subject: Re: SNMP Module as an Application Subject: Re: checksum verification - embedded Subject: Re: tffs problem Subject: Re: UGL(not WindML)+ZINC developer .... Re: UGL? Subject: RM7000 support for interrupts Subject: Re: VxWorks interview question ! Subject: Re: Loading Vxworks from TFFS Subject: Re: Gigabit Ethernet Subject: an old question Subject: Re: an old question Subject: Bug in `pppclose'. Subject: Re: an old question Subject: Re: DosFs 2.0 Product Information Subject: Re: what is the difference between bootrom and bootalbe vxworks image Subject: Does anyone know how to debug serial communication with wdb and END driver? Subject: Re: VxWorks AE 1.1: Ethernet Card Configuration Problem Subject: Is that bug of MUX or mine ? Subject: Re: what is the difference between bootrom and bootalbe vxworks image Subject: Re: Is that bug of MUX or mine ? Subject: Re: an old question Subject: Re: tffs problem ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: VxWorks AE 1.1: Ethernet Card Configuration Problem Date: Thu, 20 Dec 2001 13:40:40 +0100 From: "Mukul Dhankhar" Message-ID: <3c21dad4$0$7117$4d4efb8e@news.be.uu.net> Hi john and everybody else , My target hardware is Pentium-MMX.To connect through ethernet, I have a NE2000 card on ISA bus of my target. I have to manually configure the parameters(like IRQ number, memory hook etc.) for the NE2000 card as it is on ISA bus. But how do I configure the parameters for the 'ene' driver( the END style NE2000 driver) that is included in the boot image created by a 'BootApp project'? I want that the NE2000 driver has the required setting mathcing with those of the NE2000 ISA card. Can these settings be changed in some 'config' file like 'ne2000end.h'. Also are some special components need to be added to the 'Boot App' project to provide for ISA support? As of now, I cannot even ping my target from outside!( although I have included a PING client in the boot image). Also how can I create a boot image such that my C application is embedded in it and it runs once I boot the target with a floppy disk. In this case I will not need any network at all. My host a WinNT. Looking forward to some help ASAP. Thanks, Mukul. --------------------------- Newsgroups: comp.os.vxworks Subject: about usrNetwork.c ? Date: 20 Dec 2001 04:41:05 -0800 From: sardine_sz@163.net (sardine) Organization: http://groups.google.com/ Message-ID: hi gurus, On my computer, Tornado2,Vxworks5.4 PatchForWin2k,T2CP3 i cannot found file: $(WIND_BASE)\target\config\comps\src\net\usrNetwork.c but i can found it under: $(WIND_BASE)\target\src\config\ usrNetwork.c why ? BTW: i change all IP_MAX_UNITS to 2, but when i rebuild boot and image,and run it. the variable ipMaxUnits=1 ?????? --------------------------- Newsgroups: comp.os.vxworks Subject: Embedding C application in boot image of VxWorks AE 1.1 Date: Thu, 20 Dec 2001 13:50:38 +0100 From: "Mukul Dhankhar" Message-ID: <3c21dd2a$0$7112$4d4efb8e@news.be.uu.net> Hi all, I need to run a C application on my Pentium-MMX VxWorks AE 1.1 target. My host is WinNT. Does anyone know how I can make it happen using the 'Boot App' project in AE 1.1. For now I do not want to have a serial or ethernet connection to download the application components. I just need a system kernel image which can hold a C application and run it once I boot the target system with a floppy. Thanks, Mukul --------------------------- Newsgroups: comp.os.vxworks Subject: Re: help a newbie.. Date: 20 Dec 2001 06:58:03 -0800 From: arohb@cc.usu.edu (Aroh Barjatya) Organization: http://groups.google.com/ Message-ID: References: <488e459a.0112200023.3abbaae3@posting.google.com> Thanx John... I am using serial port WDB connection to the target. And hence i tried the last suggestion you gave me. TSFS. Apparently, while using the 'open' command (like in your example code) all i have to specify in the name of the file is "/tgtsvr/filename" and this creates a file on the host to which the target can write. "/tgtsvr" specifies that the file is to be created on the targetserver , the directory being whatever specified in the options passed while starting tgtsvr.exe The new problem is ... i can write to the file using 'open' and then 'write' commmand. This helps me write strings. But if i try doing sometihng like " '%d', 256 " it simply crashes... it does not seem to work with fwrite or fprintf either. Like i said i have to copy an image back from targets SRAM to a host comps file. so i need to write numbers back in the file.Do you have any suggestions. Thanx again...appreciate your help aroh john_94501@yahoo.com (John) wrote in message news:<488e459a.0112200023.3abbaae3@posting.google.com>... > Hello, > > arohb@cc.usu.edu (Aroh Barjatya) wrote in message news:... > > Hi i am developing a device driver for a custom made camera board. > > I am trying to read the SRAM on the board(which contains the image) > > and write it to a file on my host system(P4 1.6G) > > > > But i am unable to create a file in my C program. It refuses to create > > a file. I wonder if its trying to create a file on the target rather > > then on the host system. > > If the code is running on the target, then, yes, it will be trying to > create the file on the target. Unfortunately, you have chosen (or had > forced on you) a very poor choice of development host OS I suspect > (since it is a P4 host, I am guessing Windoze). That limits your > choices of file system that can be used by the target. You also didn't > add much information about what happened, or how you were trying to > create the file. Please try to post more detailed information in > future if you have questions; it makes it much simpler to answer > them... > > Onto file systems, the simplest is called netDrv, and if you booted > the target from the host via a network connection it should already be > available. Try the devs command from a WindShell (I think it works on > the host shell too - I normally configure the target shell and use > that). If all is well, you should see a device named after your boot > host with a ':' character after it. > > If that's working, try a simple piece of test code: > > #include "vxWorks.h" > #include "ioLib.h" > #include "errno.h" > #include "stdio.h" > > void simpleTest (char * name) > { > int fd; > > if ((fd = open (name, O_RDWR, 0)) == ERROR) > { > printf ("Error opening file %#x\n", errno); > return; > } > > printf ("File descriptor: %d\n", fd); > } > > First, find out the directory on your PC that the ftp user you used to > boot the target sees initially. You'll need to look in the > configuration of the ftp service on the host to find this, or work it > out from the path you put in the bootline in the target. > > Next, try to open a small file that already exists there. If that > works, close the file from a windshell session. > > Finally, make a quick change to the call to open; change 'O_RDWR' to > 'O_RDWR | O_CREAT' and rebuild. Then try with a name that doesn't > exist. That should allow you create a new file. Note that it won't > appear on the host though until you close it. Why? Well, because > netDrv is an FTP file system (or RSH, but that won't work on a toy OS > like Windoze). The file is created in memory on the target, and held > there all the time it is open. When you close the file, the file > system opens an ftp connection to the host and copies the file up to > it. > > This has a couple of downsides: > > 1) You need to keep the whole file in RAM, so if your file is big you > can quickly hit problems > > 2) If you don't close the file, or the target crashes before you close > it, the data is lost > > Remember, netDrv is designed for development use, not as a production > file system. > > The other development file system you might be able to use is called > TSFS (not to be confused with TFFS). TSFS is the "target server file > system" and is a file system that works using the WDB connection to > your host. This might be better for your requirements since you're > using a PC as a host; check in the docs for information about > configuring and using it. > > > Please tell me how to do it. > > HTH, > > John... > > > thanx > > aroh --------------------------- Newsgroups: comp.os.vxworks Subject: Re: help a newbie.. Date: Thu, 20 Dec 2001 10:04:12 -0500 From: Doug Dotson Organization: Verio Message-ID: <3C21FDEC.8DF15A90@digidata.com> References: <488e459a.0112200023.3abbaae3@posting.google.com> Do you have some sample code? I am doing what you are trying to do with no problem. There is a minor problem that the file created ends up Read-Only after it is closed if you use "open" to create it. This is under NT 4.0. doug Aroh Barjatya wrote: > Thanx John... > > I am using serial port WDB connection to the target. > And hence i tried the last suggestion you gave me. TSFS. > > Apparently, while using the 'open' command (like in your example code) > all i have to specify in the name of the file is "/tgtsvr/filename" > and this creates a file on the host to which the target can write. > "/tgtsvr" specifies that the file is to be created on the targetserver > , the directory being whatever specified in the options passed while > starting tgtsvr.exe > > The new problem is ... i can write to the file using 'open' and then > 'write' commmand. This helps me write strings. But if i try doing > sometihng like " '%d', 256 " it simply crashes... it does not seem to > work with fwrite or fprintf either. Like i said i have to copy an > image back from targets SRAM to a host comps file. so i need to write > numbers back in the file.Do you have any suggestions. > > Thanx again...appreciate your help > > aroh > > john_94501@yahoo.com (John) wrote in message news:<488e459a.0112200023.3abbaae3@posting.google.com>... > > Hello, > > > > arohb@cc.usu.edu (Aroh Barjatya) wrote in message news:... > > > Hi i am developing a device driver for a custom made camera board. > > > I am trying to read the SRAM on the board(which contains the image) > > > and write it to a file on my host system(P4 1.6G) > > > > > > But i am unable to create a file in my C program. It refuses to create > > > a file. I wonder if its trying to create a file on the target rather > > > then on the host system. > > > > If the code is running on the target, then, yes, it will be trying to > > create the file on the target. Unfortunately, you have chosen (or had > > forced on you) a very poor choice of development host OS I suspect > > (since it is a P4 host, I am guessing Windoze). That limits your > > choices of file system that can be used by the target. You also didn't > > add much information about what happened, or how you were trying to > > create the file. Please try to post more detailed information in > > future if you have questions; it makes it much simpler to answer > > them... > > > > Onto file systems, the simplest is called netDrv, and if you booted > > the target from the host via a network connection it should already be > > available. Try the devs command from a WindShell (I think it works on > > the host shell too - I normally configure the target shell and use > > that). If all is well, you should see a device named after your boot > > host with a ':' character after it. > > > > If that's working, try a simple piece of test code: > > > > #include "vxWorks.h" > > #include "ioLib.h" > > #include "errno.h" > > #include "stdio.h" > > > > void simpleTest (char * name) > > { > > int fd; > > > > if ((fd = open (name, O_RDWR, 0)) == ERROR) > > { > > printf ("Error opening file %#x\n", errno); > > return; > > } > > > > printf ("File descriptor: %d\n", fd); > > } > > > > First, find out the directory on your PC that the ftp user you used to > > boot the target sees initially. You'll need to look in the > > configuration of the ftp service on the host to find this, or work it > > out from the path you put in the bootline in the target. > > > > Next, try to open a small file that already exists there. If that > > works, close the file from a windshell session. > > > > Finally, make a quick change to the call to open; change 'O_RDWR' to > > 'O_RDWR | O_CREAT' and rebuild. Then try with a name that doesn't > > exist. That should allow you create a new file. Note that it won't > > appear on the host though until you close it. Why? Well, because > > netDrv is an FTP file system (or RSH, but that won't work on a toy OS > > like Windoze). The file is created in memory on the target, and held > > there all the time it is open. When you close the file, the file > > system opens an ftp connection to the host and copies the file up to > > it. > > > > This has a couple of downsides: > > > > 1) You need to keep the whole file in RAM, so if your file is big you > > can quickly hit problems > > > > 2) If you don't close the file, or the target crashes before you close > > it, the data is lost > > > > Remember, netDrv is designed for development use, not as a production > > file system. > > > > The other development file system you might be able to use is called > > TSFS (not to be confused with TFFS). TSFS is the "target server file > > system" and is a file system that works using the WDB connection to > > your host. This might be better for your requirements since you're > > using a PC as a host; check in the docs for information about > > configuring and using it. > > > > > Please tell me how to do it. > > > > HTH, > > > > John... > > > > > thanx > > > aroh --------------------------- Newsgroups: comp.os.vxworks Subject: Re: DHCP hostname option Date: 20 Dec 2001 07:05:08 -0800 From: agraham@openglobe.net (Aaron Graham) Organization: http://groups.google.com/ Message-ID: <95ebbe31.0112200705.368e9fc2@posting.google.com> References: <95ebbe31.0112181402.2c55cb8c@posting.google.com> To answer my own question: UCHAR* hostname = (UCHAR*)"dummyhostname"; dhcpcOptionAdd(pCookie, _DHCP_HOSTNAME_TAG, strlen(hostname), (UCHAR*)hostname); Don't use dhcpcOptionSet() for this option. After spending hours searching the web (including the windriver support site) for any clue that might tell me how dhcpcOptionAdd() is used, I finally just tried a few parameter permutations, and it eventually worked. I'm not sure why windriver has changed their APIs and refused to tell their paying customers about it, but at least we have this newsgroup, right? Aaron agraham@openglobe.net (Aaron Graham) wrote in message news:<95ebbe31.0112181402.2c55cb8c@posting.google.com>... > I haven't been able to find any information on the hostname option for > DHCP requests. Some cable service providers (@Home, for one) require > that you send a pre-determined hostname with your DHCP request in > order to register your machine correctly, kind of like a username. In > linux, the client is configured/started as follows: > > dhcpcd -h > > However, I'm not sure how it's done in vxWorks. There is a > _DHCPC_HOSTNAME_TAG as well as a _DHCP_CLIENT_ID_TAG (from what I can > tell, "hostname" is a specific kind of "client id"). But my > dhcpcOptionSet() function only has 2 parameters, contrary to all > documentation I've been able to find. This doesn't allow me to set > any values for the tags I throw at the dhcpcOptionSet() function. I > do seem to have dhcpcOptionAdd(), which has parameters that could > possibly take values for a hostname, but there is absolutely no > documentation in existence for that function call. > > I'm going to throw parameters around arbitrarily (typical vxworks > programming style) until I get it to work or someone replies to this > thread... > > Thanks in advance for your help. > Aaron --------------------------- Newsgroups: comp.os.vxworks Subject: Re: DosFs 2.0 Product Information Date: Thu, 20 Dec 2001 16:07:28 +0100 From: "Werner Schiendl" Message-ID: <3c21feb0$1@brateggebdc5.br-automation.co.at> References: <3c20bca4$1@brateggebdc5.br-automation.co.at> Hi, "Leonid Rosenboim" wrote in message news:newscache$172nog$af1$1@lnews.actcom.co.il... ... > But I do beleive that dosFs2 is an excellent product, many people are using > it without a problem, and I am still trying to figure out what went wrong in > those cases where people had problems. > From my point of view, the biggest problem with dosFs2 is just simply that it is DOS filesystem. There is absolutely no safety with this filesystem against all forms of data loss. Especially with read/write errors on the media or unexpected power loss. All these problems are inherent to the way DOS (aka FAT) filesystem handle file allocation, directories, etc. This has nothing to do with the quality of the code, it is just a shame that WRS has nothing better (= more stable and reliable) to offer here. best regards Werner --------------------------- Newsgroups: comp.os.vxworks Subject: Re: problem with 82559er driver Date: Thu, 20 Dec 2001 16:15:30 +0100 From: "Werner Schiendl" Message-ID: <3c220092$1@brateggebdc5.br-automation.co.at> References: <3C20C4B3.A9463A@nlr.nl> <488e459a.0112192051.f43b438@posting.google.com> "John" wrote in message news:488e459a.0112192051.f43b438@posting.google.com... > Hello, > > > I'm having a problem with the latest 82559er driver in vxWorks 5.4 for a > > Pentium target. > > As long as I don't attach the network cable the driver init's ok. > > (ifShow "eeE" returns the correct values, all routines are there) > > Did you really mean "eeE" as the name? Where did this driver come from? Normally > , 82559 devices are called "fei0" (the number will change if there is more than > one network card in the target of course). eeE is the name of the driver provided from Intel for this network devices (binary distribution for x86 platform only). > > > After I attach the cable I get a repeating message: > > > > " (0x1fa05230 (tNetTask): Panic: where are those rfd buffers " > > Sounds like there's a problem with the load string and/or the driver initialisat > ion. The number of RFDs is normally specified in the END load string (look in co > nfigNet.h for your BSP in most cases). Which arch/BSP are you using? > The documentation with the driver should contain the correct changes necessary to the BSP in all detail. OP: Did you setup the driver as explained in the readme? which BSP? using END driver? or BSD? hth Werner --------------------------- Newsgroups: comp.os.vxworks Subject: Re: how can i use file system on flash? Date: 20 Dec 2001 15:55:34 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: <97c3acac.0112150526.1aac2d1e@posting.google.com> mimilijie@21cn.com (lijie) wrote: > how can i use file system on flash? > > for there is no file system on the flash, i want to create one using > ramdevcreate(). however, i have some problems on the first > parameter.when i use 0 as the first parameter, which mease malloc(). > when i restart the vxworks, i can't find the file i stored last time. > > i am also being informed that i can use a real address on the flash as > the first parameter of the function, as 0xc000. but is that all i > should do? or there is something i am not getting at? thank for > reading this, please help me out of the problem. thank you very much! Flash has some extra requirements if you want to use it as a file system. Tak a look at TFFS, an optional component with VxWorks, this implements a complete flash file system. 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: i/o redirection Date: Thu, 20 Dec 2001 16:58:33 +0100 From: "Michael Lawnick" Organization: Buergernetz Dillingen Message-ID: <9vt1r8$egl$1@snoopy.bndlg.de> References: <9vqd56$s2a$1@snoopy.bndlg.de> Reply-To: "Michael Lawnick" Sender: mlawnick@pd9e657c7.dip.t-dialin.net Think its your own job, but you could do it like this: pipeDevCreate("shellIn",100,256); pipeDevCreate("shellOut",100,256); shellInFd=open("shellIn",O_RDONLY,0); shellOutFd=open("shellOut",O_WRONLY,0); shellId=taskNameToId("tShell"); ioTaskStdSet(shellId,0,shellInFd); ioTaskStdSet(shellId,1,shellOutFd); /* now do a trick to free shell from /tyCo/0 */ .... in your CLI: shellInPFd=open("shellIn",O_WRONLY,0); shellOutPFd=open("shellOut",O_RDONLY,0); ... /* pass through unhandled command */ write(shellInPFd,cmd,strlen(cmd)); /* get response(s) from shell and handle */ ... read(shellOutPFd,buff,sizeof(buff)); Haven't tested this code, but should work this way. - -- 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$59ymog$eb1$1@lnews.actcom.co.il... > can u elaborate? > > -- > Ofer Goren > Iamba Technologies > "Michael Lawnick" wrote in message > news:9vqd56$s2a$1@snoopy.bndlg.de... > > Hi Ofer, > > > > you definitely can redirect the I/O of your shell. Use ioTaskStdSet(). > > The problem: > > tShell will pend in its old fd until next character input. You will have > to > > send one to free it. > > > > Use a pipe for your redirected shell and you can pass all unhandled > commands > > from your CLI. > > > > 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$la8log$04o$1@lnews.actcom.co.il... > > > Well, I guess this is the only acceptable solution. Why can't I just > > > re-direct my I/O to a different task.....? well, only windriver's GODs > > can > > > give this answer, I guess.... > > > > > > Anyway, Thanks... > > > > > > -- > > > Ofer Goren > > > Iamba Technologies > > > "Johan Borkhuis" wrote in message > > > news:Xns917B8033F1A58borkhuisagerecom@135.7.153.102... > > > > "Ofer Goren" wrote: > > > > > > > > > I have a CLI task. I want all input and output to be re-directed to > > > > > this task, instead of to the shell task. in fact, I can kill the > shell > > > > > task, but I want the user to be able to access the shell upon a > > > > > specific command. > > > > > > > > The only way to do this is to kill the shell. All commands can also be > > > > executed from your CLI, but you have to lookup the symbol and call the > > > > function from within the CLI. You can also startup the shell when the > > user > > > > wants access to the 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: tffs problem Date: Thu, 20 Dec 2001 17:06:05 +0100 From: "Michael Lawnick" Organization: Buergernetz Dillingen Message-ID: <9vt29d$evc$1@snoopy.bndlg.de> References: <9vq71g$2iqp$1@gavrilo.mtu.ru> <9vs6uv$1p3$1@snoopy.bndlg.de> <9vsjbv$2a9l$1@gavrilo.mtu.ru> Reply-To: "Michael Lawnick" Sender: mlawnick@pd9e657c7.dip.t-dialin.net So you shouldn't assume that it works from boot loader ! Start with debugging at your MTD/Socket-layer and check whether your flash is correctly detected. - -- 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 ============================================== "Shrike" schrieb im Newsbeitrag news:9vsjbv$2a9l$1@gavrilo.mtu.ru... > Hello! > > > Are you able to access TFFS from shell (booting from fd,hd,net,...)? > > I've booted over network and wrote a simple program to check tffs > availability. I've called usrTffsConfig with parameters 0,0,"\tffs0\" but it > wrote "tffsDevCreate failed". > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: memPartCreate pPool ??? Date: 20 Dec 2001 08:28:55 -0800 From: kumar_amit@delhi.tcs.co.in (Amit) Organization: http://groups.google.com/ Message-ID: <7421c9fd.0112200828.14dd1986@posting.google.com> References: <7421c9fd.0112190242.35de1b4d@posting.google.com> <488e459a.0112191400.664ffc3e@posting.google.com> hello! john Thank you for your cooperation. Your ideas really helped me a lot. I have implemented it using malloc and it is working perfectly fine on the target. I hope we can stay in touch if we can resolve issues like that. Thank you again Amit Kumar --------------------------- Newsgroups: comp.os.vxworks Subject: Re: SNMP Module as an Application Date: 20 Dec 2001 16:31:32 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: <166770ed.0112190233.4471e517@posting.google.com> pritamganguly@hotmail.com (Pritam) wrote: > Hi all > I want to configure the WindNet SNMP Agent as an downloadable > application rather then an service that starts up at boot time. Is > there any way to configure the agent to meet up to my requirement. > Thanks in advance > Pritam You probably have to extract the object files from the archive, and include them in you downloadable application. 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: checksum verification - embedded Date: 20 Dec 2001 11:32:20 -0500 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: References: <3C1E161D.A3E58D5F@avoidspam.com> Sender: bpringlemeir@DeadDuck [snip] Leonid> will turn out that it failed (some bits needed to transition Leonid> from 0 to 1), like in over 90% of attempts the sector write Leonid> operation needs to be done twice (first in-pace then in a new Leonid> location). This turns out an unacceptible reduction of write Leonid> performance, which rules out write in-place of sector I had noticed this on a flash system that I wrote for a custom OS. However, a read is typically 120ns. A write is typically 6 us. It is probably worth the scan even if it only happens one time in ten; especially as it may result in more write/erase sequences. It would complicate the software a little. I think the amount of 0 to 1 transitions is likely more than 90%, in which case you are probably better to always overwrite. Anyways, you point is well taken in regards to the checksum. regards, Bill Pringlemeir. - -- The reason why abounds! Believe it or not! vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" --------------------------- Newsgroups: comp.os.vxworks Subject: Re: tffs problem Date: 20 Dec 2001 16:36:39 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: <9vq71g$2iqp$1@gavrilo.mtu.ru> <9vs6uv$1p3$1@snoopy.bndlg.de> <9vsjbv$2a9l$1@gavrilo.mtu.ru> "Shrike" wrote: > Hello! > >> Are you able to access TFFS from shell (booting from fd,hd,net,...)? > > I've booted over network and wrote a simple program to check tffs > availability. I've called usrTffsConfig with parameters 0,0,"\tffs0\" > but it wrote "tffsDevCreate failed". Did you try "/tffs0/"? 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: UGL(not WindML)+ZINC developer .... Re: UGL? Date: 20 Dec 2001 11:37:09 -0500 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: References: <1I3U7.4272$XC5.4776@www.newsranger.com> <5eeU7.8577$M3.23159@news.bora.net> Sender: bpringlemeir@DeadDuck >>>>> "ChanSeong" == ChanSeong Mike Kim writes: ChanSeong> Hi ~~ If you're more familiar with Windows programming or ChanSeong> C++, ZINC(upper layer of UGL) would be more helpful to ChanSeong> build a application. This is probably true. However, the memory foot print will go from about 100-200kBytes to 1-2MBtyes. You will also wish to replace the vxWorks allocator. The amount of new/deletes done in the C++ graphics code can really fragment your heap. UGL provides simple line, polygon, and character rendering. It also has a windowing library. If you want a `windows look and feel', then Zinc is a better choice. I think that UGL/WindML is fine for a simple single screen at a time graphics interface. regards, Bill Pringlemeir. - -- ATOMIC BOMB. ANNOYED. SINKING. U-284. vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" --------------------------- Newsgroups: comp.os.vxworks Subject: RM7000 support for interrupts Date: Thu, 20 Dec 2001 11:43:30 -0500 From: Doug Dotson Organization: Verio Message-ID: <3C221532.8EBEE46E@digidata.com> It appears that WR only supplies architecture-dependent interrupt libraries for R3000, R4000, and R4650. The extrended interrupt capabilities for the RM7000 is not supported. Ouch! Anybody have any info on this? Thanks, Doug --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks interview question ! Date: Thu, 20 Dec 2001 09:35:48 -0800 From: "Gary M" Message-ID: <3c222155$0$17833$4c41069e@reader0.ash.ops.us.uu.net> References: <23de5426.0112100602.4bf0c4ac@posting.google.com> <3c152217$0$2729$724ebb72@reader2.ash.ops.us.uu.net> <3c191f2c$0$20175$724ebb72@reader2.ash.ops.us.uu.net> <3c1fb96e$0$22741$724ebb72@reader2.ash.ops.us.uu.net> <95ebbe31.0112182028.67a18952@posting.google.com> Reply-To: "Gary M" Along those lines, I avoid asking language style questions. If the company has a style guideline, show it to the candidate during a second interview and inform him/her that all code styling is done to that standard. If the company (or the group) doesn't have a style guideline, you get what you deserve. Enough said. Compiler optimization questions, likewise, are a waste of the interview time. Far more CPU resources are wasted on bad design than on whether for (;;;) or while (1) is used. Again (broken record), it's far better to delve into a candidate's body of work in great detail to find out if he/she has the right stuff. They may have taken all the requisite classes in college, but if they don't know what, when, or how to use the items in the programmer's toolbox, forget it. (e.g. "I used a linear search because there were only 32,000 items in the list and we were using a fast Pentium.") "Aaron Graham" wrote in message news:95ebbe31.0112182028.67a18952@posting.google.com... > The sad quality of the questions (and answers given) tells me only > that Jones doesn't seem to have a very high bar for his programmers. > A lot of the answers he gives seem awfully shortsighted... > > 1. "This [ternary] operator exists in C because it allows the compiler > to produce more optimal code than an if-then-else sequence." Baloney! > Although I personlly love the ternary operator, there's no reason why > even bad compilers can't optimize an if-then-else sequence just as > well. > > 2. Programmers prefer for(;;){E} to while(1){E} because the consensus > is that more compilers will optimize the for loop better (in other > words, they don't check the conditional every time). Of course, this > could go either way. It seems obvious to me why a programmer should > prefer one construct over the other, especially after taking into > consideration which compiler he's using. > > > Aaron --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Loading Vxworks from TFFS Date: Thu, 20 Dec 2001 20:58:22 +0100 From: Patrick Deiber Organization: PatDB Message-ID: <3C2242DE.4D5FECB5@wanadoo.fr> References: <6e20e2bc.0112171501.901a1a5@posting.google.com> Reply-To: patrick.deiber@wanadoo.fr Chris, How did you put your vxWorks file on the flash ? Did you use the copy command under the shell ? Chris wrote: > Hi all, > > When loading vxWorks from tffs, i get the following error; > > [VxWorks Boot]: @ > > boot device : tffs=0,0 > unit number : 0 > processor number : 0 > host name : hostname > file name : /tffs0/vxWorks > inet on ethernet (e) : 10.0.0.2 > host inet (h) : 10.0.0.1 > user (u) : user > ftp password (pw) : password > flags (f) : 0x0 > target name (tn) : target2 > > Attaching to TFFS... done. > Loading /tffs0/vxWorks... > error loading file: status = 0x610001. > > Error loading file: errno = 0x610001. > > Please advice!!! Thank you. > > Chris. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Gigabit Ethernet Date: Thu, 20 Dec 2001 20:12:51 GMT From: Randy Ryan Organization: Our-Kids Message-ID: <3C224646.E9A682C2@pacbell.net> References: <3C191B19.B199295B@pacbell.net> Thanks for the responses I got to my inquiry, it helps :) Randy NovaSol http://www.nova-sol.com (Looking for full time senior engineers that want to live in Hawaii) --------------------------- Newsgroups: comp.os.vxworks Subject: an old question Date: Thu, 20 Dec 2001 20:35:03 GMT From: "cfk@pacbell.net" Organization: Prodigy Internet http://www.prodigy.com Message-ID: Pardon the old question, but I have been away from vxWorks for nearly a year now and I have forgotten how to set the subnet mask from the boot prompt where the ip address is set. I tried something like 12.34.45.67:255.255.255.0 to no avail, but I suspect the syntax is very close to this. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: an old question Date: 20 Dec 2001 15:35:53 -0500 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: References: Sender: bpringlemeir@DeadDuck >>>>> "cfk" == cfk@pacbell net writes: cfk> Pardon the old question, but I have been away from vxWorks for cfk> nearly a year now and I have forgotten how to set the subnet cfk> mask from the boot prompt where the ip address is set. I tried cfk> something like 12.34.45.67:255.255.255.0 to no avail, but I cfk> suspect the syntax is very close to this. Try hex. Like "12.34.45.67:ffffff00" or "12.34.45.67:0xffffff00". hth, Bill Pringlemeir. - -- Have you ever been called by your enemy in a garbage dump? Or seen your dog show you obscene movies on an airplane seven times in a row? 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: Bug in `pppclose'. Date: 20 Dec 2001 15:44:57 -0500 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: Sender: bpringlemeir@DeadDuck I have a problem with the rebootHook added by the network task. It appears to call the function `ifreset'. Some psuedo code looks like this, int ifreset(int value) { ifnet *current; splnet(); current = ifnet; while(current) { if(current->close) current->close(current); current = current->next; } splx(); return value; } This _SEEMS_ like a classic bug. The pppclose function deallocated the ifnet structure which is also in the ppp_softc array. When the close returns, the structure is now invalid. If another task allocates and used the memory, then the ifreset() function will crash and burn [seems to happen to me]. Is there a TSR for this. Is there a work around? tia, Bill Pringlemeir. [BSD source, something like vxWorks?] /* * Line specific close routine, called from device close routine * and from ttioctl. * Detach the tty from the ppp unit. * Mimics part of ttyclose(). */ int pppclose(tp, flag) struct tty *tp; int flag; { register struct ppp_softc *sc; int s; s = spltty(); ttyflush(tp, FREAD|FWRITE); tp->t_line = 0; sc = (struct ppp_softc *) tp->t_sc; if (sc != NULL) { tp->t_sc = NULL; if (tp == (struct tty *) sc->sc_devp) { pppasyncrelinq(sc); pppdealloc(sc); } } splx(s); return 0; } - -- cracking KGB radar BATF Saddam Hussein Delta Force SDI Clinton FBI jihad genetic South Africa SEAL Team 6 Marxist Kennedy vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" --------------------------- Newsgroups: comp.os.vxworks Subject: Re: an old question Date: Thu, 20 Dec 2001 20:59:40 GMT From: "cfk@pacbell.net" Organization: Prodigy Internet http://www.prodigy.com Message-ID: <0jsU7.6468$ls1.1913030825@newssvr14.news.prodigy.com> References: Thanks Bill. I have spent the year on a Linux embedded project and my fingers have forgotten some of the basic movements. "Bill Pringlemeir" wrote in message news:uy9jxocie.fsf@yahoo.com... > >>>>> "cfk" == cfk@pacbell net writes: > > cfk> Pardon the old question, but I have been away from vxWorks for > cfk> nearly a year now and I have forgotten how to set the subnet > cfk> mask from the boot prompt where the ip address is set. I tried > cfk> something like 12.34.45.67:255.255.255.0 to no avail, but I > cfk> suspect the syntax is very close to this. > > Try hex. Like "12.34.45.67:ffffff00" or "12.34.45.67:0xffffff00". > > hth, > Bill Pringlemeir. > > -- > Have you ever been called by your enemy in a garbage dump? Or seen > your dog show you obscene movies on an airplane seven times in a row? > 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: DosFs 2.0 Product Information Date: Fri, 21 Dec 2001 00:10:23 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <3c20bca4$1@brateggebdc5.br-automation.co.at> <3c21feb0$1@brateggebdc5.br-automation.co.at> First, I do not feel that the FAT disk structure is as bad as you describe it. I have spent a good amount of time thinking about the reliability issues, and I even recall writing a white paper on this subject, you can do good a solid stuff with this file system using rather simply application design rules. Many have done so succesfully, and are quite happily running mission critical systems, with MTBF far better then Windows and even Linux based systems of similar role. Secondly, one of the key goals for dosFs 2.0 was to REPLACE the older dosFsLib, which suffered many problems, and had key fratures missing, which Marketing has identified as important features to have - large disk drives with FAT32, MSFT-compatible long names with VFAT, built in consistency checker, smaller footprint for low-end designs, etc. These improvements where deemed much more important than completely new file systems. There could have been more file systems in VxWorks, such as journaling file systems, but I felt that WInd River top level management did not regard file system as a key area where Wind River needs to excell, hence these good intentions did yield any products that one could use today. If at any time Wind River top guys change their mind, I'd love to be informed. Regards, - - Leonid "Werner Schiendl" wrote in message news:3c21feb0$1@brateggebdc5.br-automation.co.at... > Hi, > > "Leonid Rosenboim" wrote in message > news:newscache$172nog$af1$1@lnews.actcom.co.il... > ... > > But I do beleive that dosFs2 is an excellent product, many people are > using > > it without a problem, and I am still trying to figure out what went wrong > in > > those cases where people had problems. > > > > From my point of view, the biggest problem with dosFs2 is just simply that > it is DOS filesystem. > There is absolutely no safety with this filesystem against all forms of data > loss. > > Especially with read/write errors on the media or unexpected power loss. > All these problems are inherent to the way DOS (aka FAT) filesystem handle > file allocation, directories, etc. > This has nothing to do with the quality of the code, it is just a shame that > WRS has nothing better (= more stable and reliable) to offer here. > > best regards > Werner > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: what is the difference between bootrom and bootalbe vxworks image Date: 20 Dec 2001 18:55:10 -0800 From: jlddotcn@hotmail.com (juliedan) Organization: http://groups.google.com/ Message-ID: References: <3c21c651$3@brateggebdc5.br-automation.co.at> "Werner Schiendl" wrote in message news:<3c21c651$3@brateggebdc5.br-automation.co.at>... > Hi, > The bootrom is normally used to load the actual VxWorks image with your > application. > Depending on the hardware, you may be able to load VxWorks image directly, > thus saving the space (and time) to run the bootrom. > > This would not (easily) be possible e. g. on x86 platform, since the system > can only load max. ~600kB bootrom from realmode. > Then the bootrom switches to protected mode and is able to load VxWorks > image of any size. > > hth > Werner Thank you for hth Werner's help. I have read the Mikefile about bootrom and vxworks. They both include the rominit.o and romstart.o. Does that mean bootable vxworks image include bootrom, but larger than bootrom? Below are some sentence in "vxworks programer's guide" /***** There are there types of vxworks image Bootrom image Downloaded vxworks image Rommed vxworks images ****/ What do they mean? Does bootrom must include vxworks's core? I am work in the Arm environment. Can I put bootrom aside in this environment? thanks a lot. Liedan Ju --------------------------- Newsgroups: comp.os.vxworks Subject: Does anyone know how to debug serial communication with wdb and END driver? Date: 20 Dec 2001 19:09:27 -0800 From: jlddotcn@hotmail.com (juliedan) Organization: http://groups.google.com/ Message-ID: Environment vdls board: CPU: arm main board: CPU: ppc850 Originally, the serial line is for wdb debug. Now, the vdsl board's serial port is used for communication (the vdsl board's serial port has changed) with ppc850's serial port. How can I debug the serial communication program with wdb. How to config wdb and component? Note: Originally, I could telnet the vdsl board. That is mean the END driver is normal. All thanks in advance. Liedan Ju jlddotcn@hotmail.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks AE 1.1: Ethernet Card Configuration Problem Date: 20 Dec 2001 22:24:28 -0800 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0112202224.56a5d018@posting.google.com> References: <3c21dad4$0$7117$4d4efb8e@news.be.uu.net> Hello, > My target hardware is Pentium-MMX.To connect through ethernet, I have a > NE2000 card on ISA bus of my target. I have to manually configure the > parameters(like IRQ number, memory hook etc.) for the NE2000 card as it is > on ISA bus. But how do I configure the parameters for the 'ene' driver( the > END style NE2000 driver) that is included in the boot image created by a > 'BootApp project'? I want that the NE2000 driver has the required setting > mathcing with those of the NE2000 ISA card. Can these settings be changed in > some 'config' file like 'ne2000end.h'. Also are some special components need > to be added to the 'Boot App' project to provide for ISA support? As of now, > I cannot even ping my target from outside!( although I have included a PING > client in the boot image). Normally, the settings for non-PCI cards like this can be set through the END load string. This can be found in configNet.h for most BSPs. One of my ARM boards uses a PCMCIA variant of the ne2000 driver, and it claims that the load string has the following format: ::::: So, see what your BSP is using for those first three values. > Also how can I create a boot image such that my C application is embedded in > it and it runs once I boot the target with a floppy disk. In this case I > will not need any network at all. My host a WinNT. I have been putting my images on a hard disc, including the bootApp image in fact. That way I can boot a standalone system. Not having the network though makes development tough, so I tend to initialise it anyway. That way I can copy new images to the hard disc, and also run up the debugger. HTH, John... --------------------------- Newsgroups: comp.os.vxworks Subject: Is that bug of MUX or mine ? Date: 20 Dec 2001 23:17:58 -0800 From: houcheng@ms17.hinet.net (houcheng) Organization: http://groups.google.com/ Message-ID: Hi all: I'm trying the END device driver on vxWorks 5.4. I use the templateEnd.c as my first end device driver source code, and install this device driver by add one entry in configNet.h { 1, templateLoad, "3:3:3", 0, 0}, After booting, I enter the "muxShow" commands and it lists the temp driver. But I can not locate the end device by invoke 'endFindByName "temp", 1'. What's problem in it? Please help me, thanks a lot. - --------------------------------------------------------------- - -> muxShow Device: ene Unit: 0 Description: ne2000 Enhanced Network Driver Protocol: IP 4.4 ARP Type: 2054 Recv 0x135194 Shutdown 0x1353d0 Protocol: IP 4.4 TCP/IP Type: 2048 Recv 0x135194 Shutdown 0x1352f4 Device: temp Unit: 1 Description: END Template Driver. value = 0 = 0x0 - -> endFindByName "temp", 1 value = 0 = 0x0 --------------------------- Newsgroups: comp.os.vxworks Subject: Re: what is the difference between bootrom and bootalbe vxworks image Date: Fri, 21 Dec 2001 08:07:32 GMT From: Anthony Rouaux Organization: [posted via Easynet France] Message-ID: References: <3c21c651$3@brateggebdc5.br-automation.co.at> A bootrom only contains the code that load your vxworks image in ram from an external place (for example, via ethernet or serial port). A bootable image contains your vxworks image and a little code that copy it from flash to RAM. If you use a rommed image, the code isn t copied, it is run from flash. Your choice between this two way depends on if you d like to change your vxworks image often. In the first configuration, you just have to change the image file on your host and in the second one you must flash your new image. I hope this could help you. Bye Anthony --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Is that bug of MUX or mine ? Date: 21 Dec 2001 01:36:52 -0800 From: sczhang@esteemnetworks.com (ShiChao Zhang) Organization: http://groups.google.com/ Message-ID: <4ceedd43.0112210136.62edbe0d@posting.google.com> References: houcheng@ms17.hinet.net (houcheng) wrote in message news:... > Hi all: > I'm trying the END device driver on vxWorks 5.4. > I use the templateEnd.c as my first end device driver source code, > and install this device driver by add one entry in configNet.h > { 1, templateLoad, "3:3:3", 0, 0}, > > After booting, I enter the "muxShow" commands and it lists > the temp driver. But I can not locate the end device by invoke > 'endFindByName "temp", 1'. What's problem in it? Please help > me, thanks a lot. > > > > --------------------------------------------------------------- > -> muxShow > Device: ene Unit: 0 > Description: ne2000 Enhanced Network Driver > Protocol: IP 4.4 ARP Type: 2054 Recv 0x135194 Shutdown 0x1353d0 > Protocol: IP 4.4 TCP/IP Type: 2048 Recv 0x135194 Shutdown 0x1352f4 > Device: temp Unit: 1 > Description: END Template Driver. > value = 0 = 0x0 > -> endFindByName "temp", 1 > value = 0 = 0x0 I don't know whether you have unload your end or not. If you have done it, then the end unit will not be the initial number, so the mux can't find the specified end.(This is a bug of MUX). But if you have not done this, the reason may be end not correctly loaded. HTH! shichao zhang Esteem Networks Ltd. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: an old question Date: Fri, 21 Dec 2001 11:13:39 +0100 From: "Mukul Dhankhar" Message-ID: <3c2309d8$0$7109$4d4efb8e@news.be.uu.net> References: Yes , you remember correct. This syntax works fine with me. "cfk@pacbell.net" wrote in message news:XXrU7.6457$Dc.1907838119@newssvr14.news.prodigy.com... > Pardon the old question, but I have been away from vxWorks for nearly a year > now and I have forgotten how to set the subnet mask from the boot prompt > where the ip address is set. I tried something like > 12.34.45.67:255.255.255.0 to no avail, but I suspect the syntax is very > close to this. > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: tffs problem Date: Fri, 21 Dec 2001 14:28:26 +0300 From: "Shrike" Organization: MTU-Intel ISP Message-ID: <9vv6bv$24a1$1@gavrilo.mtu.ru> References: <9vq71g$2iqp$1@gavrilo.mtu.ru> <9vs6uv$1p3$1@snoopy.bndlg.de> <9vsjbv$2a9l$1@gavrilo.mtu.ru> Hello! > > I've booted over network and wrote a simple program to check tffs > > availability. I've called usrTffsConfig with parameters 0,0,"\tffs0\" > > but it wrote "tffsDevCreate failed". > > Did you try "/tffs0/"? I've downloaded tffs drivers v.5 from m-sys.com and after booting over network succeded in accessing disk-on-chip with the parameters of usrTffsConfig 0,0,"\\tffs0\\". Using simple program mentioned above I could read data (for example) from bootrom.sys. So, that worked! But when I try to boot from DOC computer reboots after deflating (after 5 rows of ".....") --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Fri Dec 21 07:50:20 2001 From: Donald McLachlan Date: Fri Dec 21 07:50:23 PST 2001 Subject: RE: lspci Good News! I played with this a bit more before leaving for the holidays, and guess what ... it worked! So now I know how to retrieve this info ... but I've got 2 devices on int 12 ... want to move one. How can I change the interrupt line for this card? If I do, is the change saved in the card? If so, not ideal, but I suppose I could boot linux, use pciset, and then reboot vxWorks. Also, are the I/O ports at 7001, 7401, or both? I take it they are memory mapped at 0xf9000000. Which is preferred, I/O instructions, or memory mapped access? Don chef% rlogin 192.168.33.4 -> i NAME ENTRY TID PRI STATUS PC SP ERRNO DELAY ---------- ------------ -------- --- ---------- -------- -------- ------- ----- tExcTask _excTask 3fd3134 0 PEND 17cd06 3fd30a4 0 0 tLogTask _logTask 3fd0830 0 PEND 17cd06 3fd079c 0 0 tShell _shell 3f6bdb0 1 READY 1452c8 3f6ba70 0 0 tRlogind _rlogind 3f6db90 2 PEND 117814 3f6d85c 0 0 tRlogOutTas_rlogOutTask 3f694e8 2 READY 117814 3f69384 0 0 tRlogInTask_rlogInTask 3f67fcc 2 READY 11777b 3f67e40 0 0 tNetTask _netTask 3f90530 50 READY 149fcf 3f904c8 0 0 value = 0 = 0x0 -> pciDeviceShow(0) Scanning function 0 of each PCI device on bus 0 Using configuration mechanism 1 bus device function vendorID deviceID class 00000000 00000000 00000000 00001106 00000691 00000600 00000000 00000001 00000000 00001106 00008598 00040600 00000000 00000007 00000000 00001106 00000596 00010600 00000000 0000000a 00000000 000010b7 00009055 00000200 00000000 0000000d 00000000 0000135e 00005102 02000700 value = 0 = 0x0 -> pciFindDeviceShow(0x0000135e, 0x00005102, 0) deviceId = 0x00005102 vendorId = 0x0000135e index = 0x00000000 busNo = 0x00000000 deviceNo = 0x0000000d funcNo = 0x00000000 value = 0 = 0x0 -> pciHeaderShow(0, 0x0000000d, 0x00000000) vendor ID = 0x135e device ID = 0x5102 command register = 0x0003 status register = 0x0280 revision ID = 0x01 class code = 0x07 sub class code = 0x00 programming interface = 0x02 cache line = 0x08 latency time = 0x00 header type = 0x00 BIST = 0x00 base address 0 = 0xf9000000 base address 1 = 0x00007001 base address 2 = 0x00007401 base address 3 = 0x00000000 base address 4 = 0x00000000 base address 5 = 0x00000000 cardBus CIS pointer = 0x00000000 sub system vendor ID = 0x135e sub system ID = 0x5102 expansion ROM base address = 0x00000000 interrupt line = 0x0c interrupt pin = 0x01 min Grant = 0x00 max Latency = 0x00 value = 0 = 0x0 -> ~. From vxwexplo-errs@csg.lbl.gov Sat Dec 22 04:03:17 2001 From: Vxworks Exploder Date: Sat Dec 22 04:03:20 PST 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sat Dec 22 04:03:12 PST 2001 Subject: Re: tffs problem Subject: Re: an old question Subject: Re: tffs problem Subject: malloc in interruption Subject: Re: malloc in interruption Subject: VxWorks Does not communicate with the Ethernet card Subject: Re: Is that bug of MUX or mine ? Subject: Re: malloc in interruption Subject: Re: about usrNetwork.c ? Subject: Re: Embedding C application in boot image of VxWorks AE 1.1 Subject: Re: memcpy problem Subject: Re: malloc in interruption Subject: Re: memcpy problem Subject: Re: [One more] About PCMCIA / PC Card Driver Subject: Re: How to verify Cache mode at run-time Subject: Re: DosFs 2.0 Product Information Subject: Re: malloc in interruption Subject: HW benchmark under VxWorks Subject: New line of 64 Bit Notebook Workstations runs Solaris with Hardware based Personal ID Security Subject: Re: help a newbie.. Subject: Re: ENOBUFS using sendmsg(). Subject: Re: Where is PTY Driver Subject: Re: Why is the IP stack ARPing my non-Ethernet interface? Subject: Re: about usrNetwork.c ? Subject: [Q] how to get string with scanf? Subject: Flash File system code loading Subject: Console Output via Ethernet? Subject: Sending UDP packet out specific interface? Subject: Re: Where is PTY Driver Subject: Re: DosFs 2.0 Product Information Subject: vxworks on adsl modem ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: tffs problem Date: Fri, 21 Dec 2001 06:25:16 GMT From: "Daniel Poirot" Organization: Prodigy Internet http://www.prodigy.com Message-ID: References: <9vq71g$2iqp$1@gavrilo.mtu.ru> <9vs6uv$1p3$1@snoopy.bndlg.de> <9vsjbv$2a9l$1@gavrilo.mtu.ru> <9vv6bv$24a1$1@gavrilo.mtu.ru> Reply-To: "Daniel Poirot" FIVE ROWS of dots! I smell x86. Your bootrom image may be too big. If I remember correctly, vxLd prints one dot for each sector of bootrom read. In the wonderful world of x86, the bootrom code has to fit between 0x8000 and 0xa0000 (640k). vxLd just keeps reading the boot device and writing to memory until it gets to the end of the file. Some of your bootrom could have ended up nowhere. ;-) Try eliminating as many things as possible from config.h and try again. Why are you using the driver from M-Systems, anyway? It is a real pain to install... - - dan "Shrike" wrote in message news:9vv6bv$24a1$1@gavrilo.mtu.ru... > Hello! > > > > I've booted over network and wrote a simple program to check tffs > > > availability. I've called usrTffsConfig with parameters 0,0,"\tffs0\" > > > but it wrote "tffsDevCreate failed". > > > > Did you try "/tffs0/"? > > I've downloaded tffs drivers v.5 from m-sys.com and after booting over > network succeded in accessing disk-on-chip with the parameters of > usrTffsConfig 0,0,"\\tffs0\\". Using simple program mentioned above I could > read data (for example) from bootrom.sys. So, that worked! > > But when I try to boot from DOC computer reboots after deflating (after 5 > rows of ".....") > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: an old question Date: Fri, 21 Dec 2001 12:57:55 GMT From: Tony Dal Santo Message-ID: References: <3c2309d8$0$7109$4d4efb8e@news.be.uu.net> Sender: Tony Dal Santo "cfk@pacbell.net" wrote: > Pardon the old question, but I have been away from vxWorks for nearly a > year > now and I have forgotten how to set the subnet mask from the boot prompt > where the ip address is set. I tried something like > 12.34.45.67:255.255.255.0 to no avail, but I suspect the syntax is very > close to this. Try using hex instead (12.34.45.67:ffffff00). Tony --------------------------- Newsgroups: comp.os.vxworks Subject: Re: tffs problem Date: Fri, 21 Dec 2001 16:32:01 +0300 From: "Shrike" Organization: MTU-Intel ISP Message-ID: <9vvdjm$2b9h$1@gavrilo.mtu.ru> References: <9vq71g$2iqp$1@gavrilo.mtu.ru> <9vs6uv$1p3$1@snoopy.bndlg.de> <9vsjbv$2a9l$1@gavrilo.mtu.ru> <9vv6bv$24a1$1@gavrilo.mtu.ru> > FIVE ROWS of dots! I smell x86. ;-) > Your bootrom image may be too big. If I remember correctly, vxLd prints > one dot for each sector of bootrom read. In the wonderful world of x86, > the bootrom code has to fit between 0x8000 and 0xa0000 (640k). vxLd > just keeps reading the boot device and writing to memory until it gets > to the end of the file. Some of your bootrom could have ended up > nowhere. ;-) > > Try eliminating as many things as possible from config.h and try again. Thanks alot! It worked! I owe you a beer ;-) BTW, I've used bootrom_uncmp earlier. Now, 550Kb -> 230 Kb. > Why are you using the driver from M-Systems, anyway? It is a real pain > to install... Yeah, right. It was not too quick to install, but default version didn't work at all. --------------------------- Newsgroups: comp.os.vxworks Subject: malloc in interruption Date: Fri, 21 Dec 2001 14:40:00 GMT From: Anthony Rouaux Organization: [posted via Easynet France] Message-ID: Is there any restrictions to do a malloc in a interrupt function. Is there an alternative to allocate memory in this case? Thanks --------------------------- Newsgroups: comp.os.vxworks Subject: Re: malloc in interruption Date: 21 Dec 2001 09:47:46 -0500 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: References: Sender: bpringlemeir@DeadDuck >>>>> "Anthony" == Anthony Rouaux writes: Anthony> Is there any restrictions to do a malloc in a interrupt Anthony> function. Is there an alternative to allocate memory in this Anthony> case? This is a very hard problem if you think about it. Interrupts can happen at any time and it is difficult to verify the consistency of data structures. Especially if you have more than one value in the structure. Even one is difficult if you must read the value, make a decision and write a modified value back. Ring structures can be used from an interrupt. Most people handle this by allocating buffers in a task and then grabbing the buffers in an interrupt. One example you can look at are the network buffers. These are pre-allocated in different sizes so that a network driver can grab a buffer of the receive size and then fill it for use by the network stack. hth, Bill Pringlemeir. - -- I don't know WHY I said that.. I think it came from the FILLINGS in my rear molars.. vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" --------------------------- Newsgroups: comp.os.vxworks Subject: VxWorks Does not communicate with the Ethernet card Date: Fri, 21 Dec 2001 16:07:09 +0100 From: "Mukul Dhankhar" Message-ID: <3c234ea1$0$24505$4d4efb8e@news.be.uu.net> Hi all, I have a NE2000 ethernet card on the ISA bus of my Pentium-MMX VxWorks AE 1.1 target. The problem is that VxWorks does not connect to ethernet through it. I have already manually set the IRQ and IO access address on the NE2000 card to match those of the NE2000 drivers defaults( defined in a '00bsp.cdf' file). - -I am not able to ping the target from outside. - -Moreover, I embedded a C application in the boot image,which pings my WinNT Tornado Host, but I get an error message on the Console that: ping timeout, no answer from host 90.0.0.4. - -Also the 'transmit LED' on the Ethernet Card does not flicker at all suggesting probably that VxWorks is not able to communicate with the Card. Can anyone suggest a way to configure the driver or the ethernet card, or maybe some debugging tool which did not come to my mind. I believe ping is best way to see if my packets are going out from my host. Thanks, Mukul --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Is that bug of MUX or mine ? Date: 21 Dec 2001 07:13:21 -0800 From: jperron@cetaceannetworks.com (Jason Perron) Organization: http://groups.google.com/ Message-ID: References: <4ceedd43.0112210136.62edbe0d@posting.google.com> In your temp END driver's "load" routine, must allocate memory for the END_OBJ. Make sure that you zero out this memory immediately after the malloc call succeeds. If you don't zero the memory all kinds of wierd things happen. HTH, Jason sczhang@esteemnetworks.com (ShiChao Zhang) wrote in message news:<4ceedd43.0112210136.62edbe0d@posting.google.com>... > houcheng@ms17.hinet.net (houcheng) wrote in message news:... > > Hi all: > > I'm trying the END device driver on vxWorks 5.4. > > I use the templateEnd.c as my first end device driver source code, > > and install this device driver by add one entry in configNet.h > > { 1, templateLoad, "3:3:3", 0, 0}, > > > > After booting, I enter the "muxShow" commands and it lists > > the temp driver. But I can not locate the end device by invoke > > 'endFindByName "temp", 1'. What's problem in it? Please help > > me, thanks a lot. > > > > > > > > --------------------------------------------------------------- > > -> muxShow > > Device: ene Unit: 0 > > Description: ne2000 Enhanced Network Driver > > Protocol: IP 4.4 ARP Type: 2054 Recv 0x135194 Shutdown 0x1353d0 > > Protocol: IP 4.4 TCP/IP Type: 2048 Recv 0x135194 Shutdown 0x1352f4 > > Device: temp Unit: 1 > > Description: END Template Driver. > > value = 0 = 0x0 > > -> endFindByName "temp", 1 > > value = 0 = 0x0 > > I don't know whether you have unload your end or not. If you have done > it, then the end unit will not be the initial number, so the mux can't > find the specified end.(This is a bug of MUX). > But if you have not done this, the reason may be end not correctly > loaded. > > HTH! > > shichao zhang > Esteem Networks Ltd. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: malloc in interruption Date: Fri, 21 Dec 2001 10:15:48 -0500 From: Doug Dotson Organization: Verio Message-ID: <3C235224.46A40C3B@digidata.com> References: The Programmer's Guide has a list of functions and libraries that CAN be used inside an ISR. malloc doesn't appear in the list so it appears that it is OK. doug Anthony Rouaux wrote: > Is there any restrictions to do a malloc in a interrupt function. Is there > an alternative to allocate memory in this case? > Thanks --------------------------- Newsgroups: comp.os.vxworks Subject: Re: about usrNetwork.c ? Date: Fri, 21 Dec 2001 10:31:31 -0500 From: david lindauer Message-ID: <3C2355D3.2010101@notifier-is.net> References: sardine wrote: >hi gurus, > > On my computer, Tornado2,Vxworks5.4 PatchForWin2k,T2CP3 > i cannot found file: > $(WIND_BASE)\target\config\comps\src\net\usrNetwork.c > but i can found it under: > $(WIND_BASE)\target\src\config\ usrNetwork.c > > why ? > I think usrNetwork is a more general network startup routine than the stuff in comps\src\net > >BTW: > i change all IP_MAX_UNITS to 2, > but when i rebuild boot and image,and run it. > the variable ipMaxUnits=1 > > ?????? > Did you change it in the build window? VxWorks may ignore config.h files in favor of what is in the build window. David --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Embedding C application in boot image of VxWorks AE 1.1 Date: Fri, 21 Dec 2001 10:33:28 -0500 From: david lindauer Message-ID: <3C235648.3080608@notifier-is.net> References: <3c21dd2a$0$7112$4d4efb8e@news.be.uu.net> Combine the files in with a bootable VxWorks project. Then modify usrAppInit.C to call your program (you may want to actually spawn your program here instead of calling it directly, as initialization isn't quite complete when this routine is called). David Mukul Dhankhar wrote: >Hi all, > I need to run a C application on my Pentium-MMX VxWorks AE 1.1 target. My >host is WinNT. Does anyone know how I can make it happen using the 'Boot >App' project in AE 1.1. For now I do not want to have a serial or ethernet >connection to download the application components. I just need a system >kernel image which can hold a C application and run it once I boot the >target system with a floppy. >Thanks, >Mukul > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: memcpy problem Date: Fri, 21 Dec 2001 15:49:30 GMT From: Victor Sperry Organization: Paradyne Message-ID: <3C235A08.C54468B5@paradyne.com> References: This is a multi-part message in MIME format. - --------------0C4E59D74748561CA6FDC745 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit June wrote: > > i thought it probably because of NULL too. but i tried to print out > the address of h and i actually used this: > printf("address of h:0x%x\n",&h); > > and see this: > address of h:0x508fc0 > > That means it was allocated some space for h. No, this is absolutely wrong. Many people have pointed out your first error, and it is undoubtedly an error. You set the pointer to NULL. By what mechanism did you think the memory became allocated? There are no memory pixies. Please don't argue about this any more. > And also i tried > struct hostent *h=malloc(sizeof(struct hostent)); This should work, but I would argue that it is a bad practice. It would be better to have the calling function allocate the memory, and pass in "h" ("h" is a bad name, but that's another rant). Your function would then fill in the data. The reason this is better is that it is not always acceptable to use malloc. In particular, if your calling function doesn't need to keep the info around for a long time, it may need to free up the memory. If you've used malloc, the caller must KNOW you've used malloc so it can use free. It's better to let the caller do the allocation AND the deallocation. > and i got the exception: > > Exception number 16: Task: 0x509044 (t6) > > Data abort > Program Counter: 0x1b320 > Status Register: 0x80000013 > 902dc _vxTaskEntry +28 : 42f54 () > 42ff8 _wdbFuncCallLibInit+128: _gethostbyname () > 1364fc _gethostbyname +b8 : _memcpy () > value = 0 = 0x0 > > This looks like a quite simple problem. but i just can not figure it > out. :( For reference, here's your code: struct hostent *gethostbyname(char *name) { struct hostent *h=NULL; int addr = hostGetByName(name); if (addr != -1) { h->h_name = name; h->h_addrtype = AF_INET; h->h_length = sizeof(addr); memcpy(h->h_addr,&addr,h->h_length); } return h; } Your second error is that you didn't allocate any memory for h->h_addr, which is another pointer to char. See? struct hostent { char *h_name; /* official name of host */ char **h_aliases; /* alias list */ int h_addrtype; /* host address type */ int h_length; /* length of address */ char **h_addr_list; /* list of addresses from name server */ #define h_addr h_addr_list[0] /* address, for backward compatiblity */ unsigned int h_ttl; /* Time to Live in Seconds for this entry */ }; Even if you malloc the hostent struct, you must still allocate the data that is pointed to by the pointers within the struct. As an aside, it is PROBABLY an error for you to have done this: h->h_name = name; How do you know that the caller's passed-in "name" is not a temporary variable created on the stack? This has "potential exception" written all over it. Also, what about the h_aliases pointer-pointer? You never initialized this either. It's going to point to random memory. - --------------0C4E59D74748561CA6FDC745 Content-Type: text/x-vcard; charset=us-ascii; name="vsperry.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Victor Sperry Content-Disposition: attachment; filename="vsperry.vcf" begin:vcard n:Sperry;Victor tel;cell:727-424-0810 tel;home:727-573-5722 tel;work:727-530-2640 x-mozilla-html:TRUE url:http://www.eng.paradyne.com/~vsperry org:Paradyne Networks, Inc;DSLAM ATM Portcard Firmware version:2.1 email;internet:vsperry@paradyne.com title:Member Technical Staff adr;quoted-printable:;;8545 126th Ave N=0D=0ALG134;Largo;FL;33779-2826;USA note;quoted-printable:Bumper sticker: I see dumb people=0D=0A x-mozilla-cpt:;-8704 fn:Vic Sperry end:vcard - --------------0C4E59D74748561CA6FDC745-- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: malloc in interruption Date: Fri, 21 Dec 2001 10:54:12 -0500 From: Doug Dotson Organization: Verio Message-ID: <3C235B24.C96BF5C8@digidata.com> References: <3C235224.46A40C3B@digidata.com> malloc is reentrant. It has to be in order to work in a threaded environment. However, I mis-spoke in my previous email. malloc (a member of memPartLib) does NOT appear in the list of permissible functions for use in an ISP in the manual. So it is best not to use it. malloc also can be slow so if deterministic behavior is necessary, it would be best to allocate a pool of blocks and manage them yourself. doug Doug Dotson wrote: > The Programmer's Guide has a list of functions and libraries that > CAN be used inside an ISR. malloc doesn't appear in the list > so it appears that it is OK. > > doug > > Anthony Rouaux wrote: > > > Is there any restrictions to do a malloc in a interrupt function. Is there > > an alternative to allocate memory in this case? > > Thanks --------------------------- Newsgroups: comp.os.vxworks Subject: Re: memcpy problem Date: Fri, 21 Dec 2001 11:49:48 -0500 From: Doug Dotson Organization: Verio Message-ID: <3C23682C.A66B4F86@digidata.com> References: <3C235A08.C54468B5@paradyne.com> >It's better to let the caller do the allocation AND the deallocation. Sometimes. doug Victor Sperry wrote: > June wrote: > > > > i thought it probably because of NULL too. but i tried to print out > > the address of h and i actually used this: > > printf("address of h:0x%x\n",&h); > > > > and see this: > > address of h:0x508fc0 > > > > That means it was allocated some space for h. > > No, this is absolutely wrong. Many people have pointed > out your first error, and it is undoubtedly an error. > You set the pointer to NULL. By what mechanism did you > think the memory became allocated? There are no memory > pixies. Please don't argue about this any more. > > > And also i tried > > struct hostent *h=malloc(sizeof(struct hostent)); > > This should work, but I would argue that it is a bad > practice. It would be better to have the calling function > allocate the memory, and pass in "h" ("h" is a bad name, > but that's another rant). Your function would then fill > in the data. The reason this is better is that it is not > always acceptable to use malloc. In particular, if your > calling function doesn't need to keep the info around for > a long time, it may need to free up the memory. If you've > used malloc, the caller must KNOW you've used malloc so > it can use free. It's better to let the caller do the > allocation AND the deallocation. > > > and i got the exception: > > > > Exception number 16: Task: 0x509044 (t6) > > > > Data abort > > Program Counter: 0x1b320 > > Status Register: 0x80000013 > > 902dc _vxTaskEntry +28 : 42f54 () > > 42ff8 _wdbFuncCallLibInit+128: _gethostbyname () > > 1364fc _gethostbyname +b8 : _memcpy () > > value = 0 = 0x0 > > > > This looks like a quite simple problem. but i just can not figure it > > out. :( > > For reference, here's your code: > > struct hostent *gethostbyname(char *name) > { > struct hostent *h=NULL; > int addr = hostGetByName(name); > if (addr != -1) > { > h->h_name = name; > h->h_addrtype = AF_INET; > h->h_length = sizeof(addr); > memcpy(h->h_addr,&addr,h->h_length); > > } > return h; > > } > > Your second error is that you didn't allocate any memory for > h->h_addr, which is another pointer to char. See? > > struct hostent { > char *h_name; /* official name of host */ > char **h_aliases; /* alias list */ > int h_addrtype; /* host address type */ > int h_length; /* length of address */ > char **h_addr_list; /* list of addresses from name server */ > #define h_addr h_addr_list[0] /* address, for backward compatiblity */ > unsigned int h_ttl; /* Time to Live in Seconds for this entry */ > }; > > Even if you malloc the hostent struct, you must still allocate > the data that is pointed to by the pointers within the struct. > > As an aside, it is PROBABLY an error for you to have done this: > h->h_name = name; > > How do you know that the caller's passed-in "name" is not a > temporary variable created on the stack? This has "potential > exception" written all over it. > > Also, what about the h_aliases pointer-pointer? You never > initialized this either. It's going to point to random memory. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: [One more] About PCMCIA / PC Card Driver Date: Fri, 21 Dec 2001 17:27:26 +0000 From: molochai Message-ID: References: <3ad4e9be.0112161739.1433136b@posting.google.com> James wrote: > Configuration > CPU : MPC8240 > OS : VxWorks > PCMCIA Controller: PCI1410A (Texas Instrument) > PC Card : ATA-Flash > > Hello. > Now, I am handling vxWorks pcmcia driver. > Since vxWorks provide only Intel 82365 pcmcia > controller driver, I use PCI1410A's ExCA Compatibility > Register for 82365 mode(memory-mapped). > > Card detect, Ready signals are good. > And Read/Write of Status, Card-Status-Change, > Power Control register are good. > But CIS(Card Information Structure) of PC Card are not read. > (only 0xff) > You're going to have fun. AFAIK Windriver don't support PCMCIA on any architectures other than x86, for fairly good reasons: various parts of the CIS handling code is completely non-portable and will need fixing up - inparticular look for code reading multi-byte fields from the CIS records. I've got it working about 18 months ago using one of the previous generations of TI PC-card/Cardbus bridges, so it is possible. I have a feeling there were a few other gotchas in there too which will become apparent. If you are routing interrupt lines from the card through the bridge onto a single PCI interrupt line, you'll probably also need modify the EPIC driver to provide virtual interrupt vectors in order to get standard drivers to work unchanged. If you're using an external interrupt controller, you'll already have something along these lines built into the EPIC driver. > I saw MPC860 BSP source code using pcmcia / pc card. > But it's some different. > MPC860 has own pcmcia controller and it is connected > to system bus. > And CIS_MEM_START, CIS_REG_START address value is defined > statically. > > My system use PCI pcmcia controller, > so I have to allocate some PCI memory for PC-Card's CIS. > But I don't know how to map PCI memory to CIS. > Not a great deal, I think the window size is programmable and the CIS handling code will work within this. IISTR that there is a limit on the size of the window that can be programmed with the ExCA register set, so map at least this much. Be aware that allocating too much will result in extra page table entries. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to verify Cache mode at run-time Date: 21 Dec 2001 10:55:45 -0800 From: ryeung@earthlink.net (Raymond Yeung) Organization: http://groups.google.com/ Message-ID: References: <2ede7703.0112180543.2773d474@posting.google.com> <8PbU7.15414$AI.8759814@typhoon.san.rr.com> Thanks for the pointer Dan; I'd look into them in details. I'm using MMU_BASIC. The reason that I'm interested in verifying the cache mode is that, while the Tornado project setup suggests that it's write-through, I'm seeing something odd with the networking stack. In particular, I've my Ethernet driver configured to use non-cacheable data buffers. Hence the flush/invalidate functions are NULL. When I do that, some functions such as script download (from network), and "ls" no longer works (however, ping still works). However, if I modify the driver to use the "real" flush/invalidate functions (as if the memory is cacheable), the problem goes away. I believe the networking stack has its own (cacheable) pool of memory. However, if it were write-through, I shouldn't have the problem I described. Raymond "dgold1" wrote in message news:<8PbU7.15414$AI.8759814@typhoon.san.rr.com>... > Raymond Yeung wrote in message ... > >Hi Kannan, > > > >I try looking up DC_CST and DFWT, both in MPC8260 manual > >and in 603e manual. I also search my BSP, but only find > >it in 860 specified code. Otherwise I couldn't find them. > > > You won't find them. They are 860 specific registers. > > >Another thing is that, my project is not configured to use > >full MMU, but some "partial" MMU setup. Would the system > >default to copy-back mode? > > If #INCLUDE_MMU_FULL is defined, it expects you to have the (add on, extra > $$$) pkg VxVMI installed. Otherwise, INCLUDE_MMU_BASIC will be defined. > > The cache mode is set with a #define in configAll.h, and can be overwritten > in your bsp specific config.h. > > The cache mode you choose only affects those memory spaces covered in > sysPhysMemDesc, i.e. those using page table entries. To see the "real, > run-time" mode, you need to get the PTE for your desired memory address and > look at the WIMG bits. IIRC, there's a call or 2 you can make vmXXXGet that > may get a PTE for you, but I can't remember for sure. > > Since you're on a PPC that also supports [ID]BATS, the only way to for sure > know is to read the BAT register and examine the WIMG bits there. You can > access them with the usual mfspr code, each BAT has a unique SPR. Of course > if you've set the appropriate values in the sysBatDesc[] array, you could > always examine the entries in the array, and assume WRS set the BATS > correctly (They do). > > You could always put a global variable in the code based upon the #define in > configAll.h, i.e. > #if (dcache_mode == COPYBACK) > int dcacheMode=1; > #elif (dcache_mode == WRITETHROUGH) > int dcacheMode=2; > #endif > > Don't forget vxHid0Get() to figure out if the caches are even enabled... > > hth, > Dan Gold gold@ensemble.com > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: DosFs 2.0 Product Information Date: 21 Dec 2001 11:05:55 -0800 From: bwaisnor@gothamnetworks.com (Bob Waisnor) Organization: http://groups.google.com/ Message-ID: <52f21cc5.0112211105.7e146316@posting.google.com> References: Looked at this DosFs 2.0 file corruption (file size -1)problem for a LONG time. Both Windriver and M-systems did the "Gee, it shouldn't do that". Come to find out that the problem is in the dosfs 2.0 binary. Here we re-compiled and did not use 2 Gig support and the thing works. My guess it's a compile issue.... "Don Dewar" wrote in message news:... > I was also shocked to find this out. I called my FAE who said I had to pay > for it. When I pointed out I heard it was free with a support contract he > relented and sent me the binaries. They charge an outrageous amount of > money for the sources. > > The integration went ok, but we suddenly found some of our files > disappearing or getting corrupt. We tried to debug it for months with very > little support from WRS. Finally, we got the sources and still could not > figure it out. When we compiled with less than 2 MB support all our > problems all vanished. > > Anyway they don't seem to support it very well. > > Good luck. > -- Don > "Leonid Rosenboim" wrote in message > news:newscache$w61gog$hug$1@lnews.actcom.co.il... > > I tried to find the technical specifications of DosFs 2.0 on the WindRiver > > Web site to show someone, and to my surprise, I found no mention of it. > > This seems shocking to me that this optional component is not advertised, > > and I remain puzzled as to the reasons why. > > > > Surely I don't expect any official answer from WindRiver (although that > > might > > be nice to chat again with my old pals), but I would appreciate if anyone > > can > > shed some light on this enigma. > > > > Of course this is mere curiosity. > > - Leonid > > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: malloc in interruption Date: Fri, 21 Dec 2001 20:10:52 +0000 (UTC) From: "David Laight" Organization: BT Openworld Message-ID: References: <3C235224.46A40C3B@digidata.com> <3C235B24.C96BF5C8@digidata.com> The vxWorks memPartAlignedAlloc (ie what malloc calls) is definitely not re-entrant. You WILL stuff thefree list if you call malloc() of free() from an ISR. David Doug Dotson wrote in message news:3C235B24.C96BF5C8@digidata.com... > malloc is reentrant. It has to be in order to work in a threaded > environment. However, I mis-spoke in my previous email. malloc > (a member of memPartLib) does NOT appear in the list of > permissible functions for use in an ISP in the manual. So it is > best not to use it. malloc also can be slow so if deterministic > behavior is necessary, it would be best to allocate a pool of > blocks and manage them yourself. > > doug > > > Doug Dotson wrote: > > > The Programmer's Guide has a list of functions and libraries that > > CAN be used inside an ISR. malloc doesn't appear in the list > > so it appears that it is OK. > > > > doug > > > > Anthony Rouaux wrote: > > > > > Is there any restrictions to do a malloc in a interrupt function. Is there > > > an alternative to allocate memory in this case? > > > Thanks > --------------------------- Newsgroups: comp.os.vxworks,comp.benchmarks,comp.sys.sgi.hardware Subject: HW benchmark under VxWorks Date: 21 Dec 2001 12:17:27 -0800 From: akarpov@my-deja.com (Alex) Organization: http://groups.google.com/ Message-ID: <4f09bb68.0112211217.407f2053@posting.google.com> Hi, I'd like to examine performance of my hardware platform (PPC) through SPEC CPU95 tests. I found out the SPEC95 suite doesn't support VxWorks. I was wondering if someone could share any information on running SPEC or other third party benchmarking kits under VxWorks. Thanks, Alex --------------------------- Newsgroups: comp.os.vxworks Subject: New line of 64 Bit Notebook Workstations runs Solaris with Hardware based Personal ID Security Date: Fri, 21 Dec 2001 17:41:52 -0500 From: "Bob Labadini" Organization: Shore.Net, a PRIMUS Company (info@shore.net) Message-ID: FOR IMMEDIATE RELEASE New line of 64 Bit Notebook Workstations runs Solaris or Linux with Hardware based Personal ID Security: "The power of a desktop RISC workstations based on Sun Microsystems 64 bit UltraSPARCIIe architecture Solaris platform brought to a portable PC style notebook package without compromising performance and with an innovative hardware Personal ID security feature to secure data and protect user IP" FOR PRODUCT INFORMATION: Go to http://www.atnextcom.com/UltraStation_Main.htm Nashua, NH - December 6, 2001 - NextCom LLC announces the availability in the North American market of a new line of 64 bit Notebook Workstations. These systems bring the power of 64 bit UltraSPARCIIe architecture Solaris based RISC workstations to a portable PC style notebook package without compromising performance and with an innovative hardware Personal ID security feature to protect user IP. Available for immediate shipment, the company offers two different product families (NCM-747s and NCM-777s). Each version has different expansion options as far as memory (up to 1GB), 20GB disk drives (single or dual), Solaris 8 or Linux. Variations in product families include UltraSPARC IIe 400Mhz or 500Mhz RISC processors and 14.1" or 15" TFT displays. Key Product Features: · Hardware based personal ID security. Includes a unique, self-contained digital signature ciphering security tracking device with over 800K password possibilities. With no software intervention, the locking mechanism is relatively hacker proof. When enabled, users cannot power on computer without this password preventing unauthorized access to any data and the system. · Built-in UPS system to minimize impacts of unstable power · Base configuration includes 2D/3D 24 bit graphics accelerator · User configurable auto log-out and power-off capability when battery power has been exhausted to avoid file system corruption. · Full function keyboard with Sun's function keys, · Real AC97 stereo studio audio quality · High quality PC style notebook small-footprint and packaging. Base Power User Model: NCM-777s-1. Suggested retail price is $8,398 Configuration includes: 15.1" TFT, 500Mhz IIe processor, 2D/3D 24 bit graphics, 256MB SDRAM, Solaris 8, 20GB HDD, and hardware based personal ID security. Dimensions: 322(W) x 273(D) x41(H) mm Base Slim-Line Model: NCM-747s-1. Suggested retail price is $6,398 Configuration includes: 14.1" TFT, 400Mhz IIe processor, 2D/3D 24 bit graphics, 256MB SDRAM, Solaris 8, 20GB HDD, and and hardware based personal ID security. Dimensions: 320(W) x 253(D) x41(H) mm NextCom LLC , based in Nashua, NH is a supplier Linux, Solaris, and VxWorks based systems solutions as well as integration services, remote network management solutions, and embedded software consulting services since 1998. For further information, contact: Bob Labadini Mobile Phone (617) 510-4781 Office Phone (603) 886-3874 x130 mailto:blabadini@atnextcom.com http://www.atNextCom.com Main Office Phone (603) 886-3874 Fax (419) 828-2030 131 Burke Street Nashua, NH 03060 --------------------------- Newsgroups: comp.os.vxworks Subject: Re: help a newbie.. Date: 21 Dec 2001 15:27:01 -0800 From: arohb@cc.usu.edu (Aroh Barjatya) Organization: http://groups.google.com/ Message-ID: References: <488e459a.0112200023.3abbaae3@posting.google.com> <3C21FDEC.8DF15A90@digidata.com> thanx doug for your time... its working now... this is what i do fp = fopen("/tgtsvr/filename.txt","w"); and then do a normal fprintf and fclose. it works fine...even with "%d" and stuff... apparently the target then opens a file on the host filesystem as specified for /tgtsvr while starting tgtsvr.exe. thanx guys aroh Doug Dotson wrote in message news:<3C21FDEC.8DF15A90@digidata.com>... > Do you have some sample code? I am doing what you > are trying to do with no problem. There is a minor problem > that the file created ends up Read-Only after it is closed if > you use "open" to create it. This is under NT 4.0. > > doug > > > Aroh Barjatya wrote: > > > Thanx John... > > > > I am using serial port WDB connection to the target. > > And hence i tried the last suggestion you gave me. TSFS. > > > > Apparently, while using the 'open' command (like in your example code) > > all i have to specify in the name of the file is "/tgtsvr/filename" > > and this creates a file on the host to which the target can write. > > "/tgtsvr" specifies that the file is to be created on the targetserver > > , the directory being whatever specified in the options passed while > > starting tgtsvr.exe > > > > The new problem is ... i can write to the file using 'open' and then > > 'write' commmand. This helps me write strings. But if i try doing > > sometihng like " '%d', 256 " it simply crashes... it does not seem to > > work with fwrite or fprintf either. Like i said i have to copy an > > image back from targets SRAM to a host comps file. so i need to write > > numbers back in the file.Do you have any suggestions. > > > > Thanx again...appreciate your help > > > > aroh > > > > john_94501@yahoo.com (John) wrote in message news:<488e459a.0112200023.3abbaae3@posting.google.com>... > > > Hello, > > > > > > arohb@cc.usu.edu (Aroh Barjatya) wrote in message news:... > > > > Hi i am developing a device driver for a custom made camera board. > > > > I am trying to read the SRAM on the board(which contains the image) > > > > and write it to a file on my host system(P4 1.6G) > > > > > > > > But i am unable to create a file in my C program. It refuses to create > > > > a file. I wonder if its trying to create a file on the target rather > > > > then on the host system. > > > > > > If the code is running on the target, then, yes, it will be trying to > > > create the file on the target. Unfortunately, you have chosen (or had > > > forced on you) a very poor choice of development host OS I suspect > > > (since it is a P4 host, I am guessing Windoze). That limits your > > > choices of file system that can be used by the target. You also didn't > > > add much information about what happened, or how you were trying to > > > create the file. Please try to post more detailed information in > > > future if you have questions; it makes it much simpler to answer > > > them... > > > > > > Onto file systems, the simplest is called netDrv, and if you booted > > > the target from the host via a network connection it should already be > > > available. Try the devs command from a WindShell (I think it works on > > > the host shell too - I normally configure the target shell and use > > > that). If all is well, you should see a device named after your boot > > > host with a ':' character after it. > > > > > > If that's working, try a simple piece of test code: > > > > > > #include "vxWorks.h" > > > #include "ioLib.h" > > > #include "errno.h" > > > #include "stdio.h" > > > > > > void simpleTest (char * name) > > > { > > > int fd; > > > > > > if ((fd = open (name, O_RDWR, 0)) == ERROR) > > > { > > > printf ("Error opening file %#x\n", errno); > > > return; > > > } > > > > > > printf ("File descriptor: %d\n", fd); > > > } > > > > > > First, find out the directory on your PC that the ftp user you used to > > > boot the target sees initially. You'll need to look in the > > > configuration of the ftp service on the host to find this, or work it > > > out from the path you put in the bootline in the target. > > > > > > Next, try to open a small file that already exists there. If that > > > works, close the file from a windshell session. > > > > > > Finally, make a quick change to the call to open; change 'O_RDWR' to > > > 'O_RDWR | O_CREAT' and rebuild. Then try with a name that doesn't > > > exist. That should allow you create a new file. Note that it won't > > > appear on the host though until you close it. Why? Well, because > > > netDrv is an FTP file system (or RSH, but that won't work on a toy OS > > > like Windoze). The file is created in memory on the target, and held > > > there all the time it is open. When you close the file, the file > > > system opens an ftp connection to the host and copies the file up to > > > it. > > > > > > This has a couple of downsides: > > > > > > 1) You need to keep the whole file in RAM, so if your file is big you > > > can quickly hit problems > > > > > > 2) If you don't close the file, or the target crashes before you close > > > it, the data is lost > > > > > > Remember, netDrv is designed for development use, not as a production > > > file system. > > > > > > The other development file system you might be able to use is called > > > TSFS (not to be confused with TFFS). TSFS is the "target server file > > > system" and is a file system that works using the WDB connection to > > > your host. This might be better for your requirements since you're > > > using a PC as a host; check in the docs for information about > > > configuring and using it. > > > > > > > Please tell me how to do it. > > > > > > HTH, > > > > > > John... > > > > > > > thanx > > > > aroh --------------------------- Newsgroups: comp.os.vxworks Subject: Re: ENOBUFS using sendmsg(). Date: Fri, 21 Dec 2001 15:47:57 -0800 From: "Matthew J. Weaver" Organization: Stanford Linear Accelerator Center Message-ID: <3C23CA2D.978202E2@slac.stanford.edu> References: <3c0fdd4b@news.tce.com> Fred wrote: > Occassionally while sending bursts to sendmsg(), EONBUFS will be returned. > We have verified that we are not over running the buffers using mbufShow. > Does anybody have any ideas? We had similar symptoms when using sendmsg with an iov array of greater than 63 elements. This is apparently the limit for a UDP socket. I never found out whether this limit can be changed. -Matt --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Where is PTY Driver Date: Fri, 21 Dec 2001 16:05:18 -0800 From: "Dan Pinson" Message-ID: References: <9vql4m$h1e0h$1@ID-48686.news.dfncis.de> <488e459a.0112191927.260251f3@posting.google.com> No dice. I'm using Tornado 2 (VxWorks 5.4). I can't find the string pty anywhere in the whole source directory! ..Dan "John" wrote in message news:488e459a.0112191927.260251f3@posting.google.com... > Hello, > > In AE it is INCLUDE_OS_PTYDRV that you need, probably the same, or at > least similar - search for it in config/all/configAll.h. > > HTH, > > John... > > "Dan Pinson" wrote in message news:<9vql4m$h1e0h$1@ID-48686.news.dfncis.de>... > > The PTY driver does not seem to be in my VxWorks image (lkup "pty" finds no > > symbols). Is there a configuration option I need to include? > > > > Thanks, > > Dan --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Why is the IP stack ARPing my non-Ethernet interface? Date: 21 Dec 2001 16:32:53 -0800 From: houcheng@ms17.hinet.net (houcheng) Organization: http://groups.google.com/ Message-ID: References: <3C0726B4.4DE481A1@rogers.com> That's sounds a big problem in Windriver's END architecture. I haved developed a MPC860 SAR and IP Over ATM driver in old BSD network driver style. Now I am considering porting it to SENS or NPT model. The IP Over ATM is use a differnt data link layer Address Resolve Protocol, ATMARP. So the bug of ARP in MUX is really amazing for me. I can't image that how can they left so many serious bugs in Tornado 2.0! Maybe I need to use Tornado 2.02 as my developing environment. Is the T2CP4 (Tornado 2.0 Cumulative Patch 4) the patch for T2.0 upgrade to T2.02? I can never understand Wind's patches system, Too many patches to patch, and too many pays to pay! It's WindRiver's Style. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: about usrNetwork.c ? Date: Sat, 22 Dec 2001 00:46:33 GMT From: ggg+news@best.com (GGG) Organization: rest Message-ID: References: <3C2355D3.2010101@notifier-is.net> In article <3C2355D3.2010101@notifier-is.net>, david lindauer wrote: >sardine wrote: >>hi gurus, >> On my computer, Tornado2,Vxworks5.4 PatchForWin2k,T2CP3 >> i cannot found file: >> $(WIND_BASE)\target\config\comps\src\net\usrNetwork.c >> but i can found it under: >> $(WIND_BASE)\target\src\config\ usrNetwork.c >> why ? >I think usrNetwork is a more general network startup routine than the >stuff in comps\src\net Project facility created bootable images use the configlettes in target/config/comps/src and subdirs. Images created from in the "old style" using the command line within a BSP directory will use target/config/all/usrConfig.c which pulls in files from target/src/config like usrExtra.c that pulls usrNetwork.c from the same directorty. >>BTW: >> i change all IP_MAX_UNITS to 2, >> but when i rebuild boot and image,and run it. >> the variable ipMaxUnits=1 >> ?????? >Did you change it in the build window? VxWorks may ignore config.h >files in favor of what is in the build window. >David Indeed, from the Project facility the IP_MAX_UNITS is found in "network components", "network buffer initialization" and generated to override the base BSP's definition. Happy Soltice (and Festivus to come!), GGG --------------------------- Newsgroups: comp.os.vxworks Subject: [Q] how to get string with scanf? Date: 21 Dec 2001 17:43:38 -0800 From: belliny@soback.kornet.net (belliny, Mint~*) Organization: http://groups.google.com/ Message-ID: <6fb8119a.0112211743.60a8781c@posting.google.com> HI~ I have problem with string input with scanf. when I input single character or number via scanf, there is no problem. but once move to string, it nearly kill me... maybe scanf could get string[i could store the string to variable and could printf it]. but the process is terminated, it occured error like this; - --*--*--*--*--*--*--*--*-- - -> i NAME ENTRY TID PRI STATUS PC SP ERRNO DELAY - ---------- ------------ -------- --- ---------- -------- -------- - ------- ----- tExcTask _excTask 54f8e50 0 PEND 42778e 54f8d6c 3d0001 0 tLogTask _logTask 54f4438 0 PEND 42778e 54f4350 3d0001 0 tWdbTask 0x417cd4 54f08e8 3 READY 42778e 54f07c4 3d0001 0 value = 0 = 0x0 - -> startProcess ************************************* * 1. Input * * 2. Delete specified Node * * 3. View List * * 4. Search * ************************************* Select Mode -> - ------------------------------------------------------ + Current Index : 0 + Input PrimitiveName : Input PrimitiveName : 3 + Input Messsage : End of program value = 15 = 0xf - -> Exception number 0: Task: 0x54ec640 (t1) General Protection Fault Program Counter: 0x00415cfd Status Register: 0x00010246 - -> i NAME ENTRY TID PRI STATUS PC SP ERRNO DELAY - ---------- ------------ -------- --- ---------- -------- -------- - ------- ----- tExcTask _excTask 54f8e50 0 PEND 42778e 54f8d6c 9 0 tLogTask _logTask 54f4438 0 PEND 42778e 54f4350 3d0001 0 tWdbTask 0x417cd4 54f08e8 3 READY 42778e 54f07c4 9 0 t1 0x415c78 54ec640 4 SUSPEND 415cfd 54ec5c4 9 0 value = 0 = 0x0 - -> i post code. thanks in advance... - --*--*--*--*--*--*--*--*-- void startProcess(void) { uint8_t nTempPrimitiveName; uint8_t nTempIndex = 0; char aryMessageBuf[50]; int nMode; // liststart(); printf("*************************************\n"); printf("* 1. Input *\n"); printf("* 2. Delete specified Node *\n"); printf("* 3. View List *\n"); printf("* 4. Search *\n"); printf("*************************************\n"); printf("Select Mode -> "); fscanf(stdin, "%c", &nMode); // scanf("%d", &nMode); nMode &= 0x0F; fflush(stdin); switch (nMode) { case 1: printf("------------------------------------------------------\n"); printf("+ Current Index : %d \n", nTempIndex); printf("+ Input PrimitiveName : "); fscanf(stdin, "%d", &nTempPrimitiveName); // scanf("%d", &nTempPrimitiveName); fflush(stdin); printf(" Input PrimitiveName : %d\n", nTempPrimitiveName); printf("+ Input Messsage : "); gets(aryMessageBuf); // fscanf(stdin, "%s", aryMessageBuf); /* scanf("%s", aryMessageBuf); if (*aryMessageBuf) { printf("%s\n", aryMessageBuf); // addlist(nTempIndex++, nTempPrimitiveName, aryMessageBuf); } */ break; /* case 2: break; case 3: viewlist(); break; case 4: break; */ }/* end of switch */ // lstFree(&ListDescriptor); printf("End of program\n"); } --------------------------- Newsgroups: comp.os.vxworks Subject: Flash File system code loading Date: 21 Dec 2001 19:00:32 -0800 From: eaichul@ntl30.com (Eaichul Choung) Organization: http://groups.google.com/ Message-ID: <3da3205f.0112211900.301b71c0@posting.google.com> Hi , I made flash file system initialized on vxWorks 5.4. but I need some help for completing. I loaded vxWorks image from network via ftp method. I want this loaded image to be copied on flash . tanks, regards eche --------------------------- Newsgroups: comp.os.vxworks Subject: Console Output via Ethernet? Date: Sat, 22 Dec 2001 05:19:28 GMT From: Michael Rice Message-ID: <3C2417D8.7010405@adelphia.net> Hello - If all of the serial ports on my target are otherwise used, is there anyway I can redirect the console output over the Ethernet to display on my NT box? Can this be done by configuring the target server's "Console and Redirection" options? Thanks. --------------------------- Newsgroups: comp.os.vxworks Subject: Sending UDP packet out specific interface? Date: Fri, 21 Dec 2001 22:37:31 -0800 From: bob@chaoticsoftware.com (Bob Bradley) Organization: Uh huh huh...It says Organ Message-ID: Is it possible to send a UDP packet, with a destination address that may not be on the same network as the interface (it's for server discovery/configuration), to specific interface? Linux supports a SO_BINDTODEVICE socket option and I was hoping for an equivalent with vxWorks. There is a MULTICAST_IF socket option that allows this for multicast packets (I'm using this as a temporary solution now), but I'd like to send a unicast packet. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Where is PTY Driver Date: 21 Dec 2001 23:36:18 -0800 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0112212336.1ac525ca@posting.google.com> References: <9vql4m$h1e0h$1@ID-48686.news.dfncis.de> <488e459a.0112191927.260251f3@posting.google.com> Hello, "Dan Pinson" wrote in message news:... > No dice. I'm using Tornado 2 (VxWorks 5.4). I can't find the string pty > anywhere in the whole source directory! Won't be in the source directory - you don't get the source. All the config info for the binary components is under target/config. HTH,, John... > ..Dan > > > "John" wrote in message > news:488e459a.0112191927.260251f3@posting.google.com... > > Hello, > > > > In AE it is INCLUDE_OS_PTYDRV that you need, probably the same, or at > > least similar - search for it in config/all/configAll.h. > > > > HTH, > > > > John... > > > > "Dan Pinson" wrote in message > news:<9vql4m$h1e0h$1@ID-48686.news.dfncis.de>... > > > The PTY driver does not seem to be in my VxWorks image (lkup "pty" finds > no > > > symbols). Is there a configuration option I need to include? > > > > > > Thanks, > > > Dan --------------------------- Newsgroups: comp.os.vxworks Subject: Re: DosFs 2.0 Product Information Date: Sat, 22 Dec 2001 10:39:08 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <52f21cc5.0112211105.7e146316@posting.google.com> Bob, I would like some details: + What is your processor architecture ? + How did you disable 2Gig support before recompiling ? "Bob Waisnor" wrote in message news:52f21cc5.0112211105.7e146316@posting.google.com... > Looked at this DosFs 2.0 file corruption (file size -1)problem for a > LONG time. Both Windriver and M-systems did the "Gee, it shouldn't do > that". Come to find out that the problem is in the dosfs 2.0 binary. > Here we re-compiled and did not use 2 Gig support and the thing works. > My guess it's a compile issue.... > > "Don Dewar" wrote in message news:... > > I was also shocked to find this out. I called my FAE who said I had to pay > > for it. When I pointed out I heard it was free with a support contract he > > relented and sent me the binaries. They charge an outrageous amount of > > money for the sources. > > > > The integration went ok, but we suddenly found some of our files > > disappearing or getting corrupt. We tried to debug it for months with very > > little support from WRS. Finally, we got the sources and still could not > > figure it out. When we compiled with less than 2 MB support all our > > problems all vanished. > > > > Anyway they don't seem to support it very well. > > > > Good luck. > > -- Don > > "Leonid Rosenboim" wrote in message > > news:newscache$w61gog$hug$1@lnews.actcom.co.il... > > > I tried to find the technical specifications of DosFs 2.0 on the WindRiver > > > Web site to show someone, and to my surprise, I found no mention of it. > > > This seems shocking to me that this optional component is not advertised, > > > and I remain puzzled as to the reasons why. > > > > > > Surely I don't expect any official answer from WindRiver (although that > > > might > > > be nice to chat again with my old pals), but I would appreciate if anyone > > > can > > > shed some light on this enigma. > > > > > > Of course this is mere curiosity. > > > - Leonid > > > > > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: vxworks on adsl modem Date: 22 Dec 2001 11:19:20 GMT From: Dennis Grevenstein Organization: (Posted via) INKA e.V. http://www.inka.de/ Message-ID: Hi, I have an ADSL Modem that is obviously running VxWorks. When you telnet to it, it shows: VxWorks login : Now, I would like to know more about it. Normally it should only be a dumb DSL Modem, but is has ATM too and it seems to be able to do much more. Are there any default passwords in VxWorks that might allow me logging in? What _could_ be done with VxWorks on such a DSL Modem? routing, firewalling, sniffing perhaps? TIA Dennis --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Sun Dec 23 01:46:24 2001 From: Ilia Bosis Date: Sun Dec 23 01:46:26 PST 2001 Subject: RE: DPC's in vxWorks No familiar with the Winblows programming, but this is how it's done in VxWorks: The classic solution is to send a message with the needed info to the task and let the task to do most of the work. WRS created the netJobAdd( ) API that do exactly this service for the network drivers. There they even pass the processing function pointer to the task, so that there is a "host" task that can run any code it's asked for. Another solution is to make Your task waiting on semaphore that's freed by the ISR and pass the data via the shared data structure(s). The drawback here is that the task can't do anything useful when it waits for the semaphore and that the requests can't be queued for processing. Be aware, however, that the system queue of ISR-to-tasks requests is of limited capacity. So don't send a new request before the previous one is still in processing. Otherwise You can get the infamous 'workQ panic" error messages. You get the power tool. Use it wisely. Hope, this helps. Ilia -------------------------------------------- Ilia Bosis Software Engineer PacketLight Networks Tel: +972-9-7644228 Fax: +972-9-7645444 E-mail: Ilia_Bosis@packetlight.com Web: www.packetlight.com . > -----Original Message----- > From: vxwexplo@lbl.gov [mailto:vxwexplo@lbl.gov] > Sent: Thursday, December 20, 2001 7:12 AM > To: vxworks_users@csg.lbl.gov > Subject: DPC's in vxWorks > > > Submitted-by vxwexplo-errs@csg.lbl.gov Wed Dec 19 20:55:40 2001 > Submitted-by: "Saravana Kumar K" > > > Hi, > I am facing problem in handling a task in Vxworks. Most of > my code goes into ISR routine. Every event is triggered by an > interrupt. Is there any method to move some time consuming > code out of ISR. Like "Deferred Procedure Calls" in Windows. > > Thanks > Saravanan > From vxwexplo-errs@csg.lbl.gov Sun Dec 23 04:03:09 2001 From: Vxworks Exploder Date: Sun Dec 23 04:03:12 PST 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sun Dec 23 04:03:06 PST 2001 Subject: Re: Why is the IP stack ARPing my non-Ethernet interface? Subject: Re: malloc in interruption Subject: Re: malloc in interruption Subject: Re: [Q] how to get string with scanf? Subject: Re: [Q] how to get string with scanf? Subject: Re: malloc in interruption ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Why is the IP stack ARPing my non-Ethernet interface? Date: 22 Dec 2001 12:03:50 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: <3C0726B4.4DE481A1@rogers.com> houcheng@ms17.hinet.net (houcheng) wrote: > Is the T2CP4 (Tornado 2.0 Cumulative Patch 4) the patch for > T2.0 upgrade to T2.02? Yes, it is. 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: malloc in interruption Date: Sat, 22 Dec 2001 20:19:20 GMT From: Ephraim Gadsby Organization: - minimal Message-ID: <5ek92uk4cnti9kl9ne5okde86hefo8a02a@4ax.com> References: <3C235224.46A40C3B@digidata.com> <3C235B24.C96BF5C8@digidata.com> On Fri, 21 Dec 2001 10:54:12 -0500, Doug Dotson wrote: >malloc is reentrant. It has to be in order to work in a threaded >environment. However, I mis-spoke in my previous email. malloc >(a member of memPartLib) does NOT appear in the list of >permissible functions for use in an ISR in the manual. So it is >best not to use it. malloc also can be slow so if deterministic >behavior is necessary, it would be best to allocate a pool of >blocks and manage them yourself. > Each memory partition is protected by a semaphore on which tasks wait forever. It is thread safe, but in interrupt context the semTake() returns immediately with an error code that just gets ignored. Unfortunately, this works almost all the time. So inexperienced people with a "suck it and see" mentality, don't get any feedback. I my opinion having printf() and malloc() "blow-up" in ISRs would be no bad thing. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: malloc in interruption Date: Sat, 22 Dec 2001 23:59:50 GMT From: Gwen & Emory Stagmer Organization: Excite@Home - The Leader in Broadband http://home.com/faster Message-ID: <3C251DFF.E576635F@home.com> References: Additionally this means that pushing object onto any STL collection IS NOT VALID IN AN INTERRUPT. No stacks, queues, deques, vectors, bags, ad.nauseum, can be pushed or popped in an ISR. Been there, done that. Wrote my own static queue... Emory Stagmer Anthony Rouaux wrote: > > Is there any restrictions to do a malloc in a interrupt function. Is there > an alternative to allocate memory in this case? > Thanks --------------------------- Newsgroups: comp.os.vxworks Subject: Re: [Q] how to get string with scanf? Date: Sun, 23 Dec 2001 03:18:32 GMT From: David Cooper Organization: none Message-ID: References: <6fb8119a.0112211743.60a8781c@posting.google.com> Been a long time since I did any scanf operations, so hopefully I remember this right. When using a scanf to get a string don't you have to specify ENTER in the format string, for example: scanf("%s\n", aryMsgBuf); something in the back of my memory about this. ...david +--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+ ô¿ô Genealogy Utilities & Family Research Researching: ALLEN,BERZONER,COHEN,COOPER,KUPFERZMIDT, POSNER (POIZNER), PRITIKIN, STARKMAN Dachau Camp Entry List, Genealogy Utilities http://home.adelphia.net/~dcooper000 In article <6fb8119a.0112211743.60a8781c@posting.google.com>, belliny@soback.kornet.net says... > HI~ > > I have problem with string input with scanf. when I input single > character or number via scanf, there is no problem. but once move to > string, it nearly kill me... > maybe scanf could get string[i could store the string to variable and > could printf it]. but the process is terminated, it occured error like > this; > > > --*--*--*--*--*--*--*--*-- > -> i > NAME ENTRY TID PRI STATUS PC SP > ERRNO DELAY > ---------- ------------ -------- --- ---------- -------- -------- > ------- ----- > tExcTask _excTask 54f8e50 0 PEND 42778e 54f8d6c > 3d0001 0 > tLogTask _logTask 54f4438 0 PEND 42778e 54f4350 > 3d0001 0 > tWdbTask 0x417cd4 54f08e8 3 READY 42778e 54f07c4 > 3d0001 0 > value = 0 = 0x0 > -> startProcess > ************************************* > * 1. Input * > * 2. Delete specified Node * > * 3. View List * > * 4. Search * > ************************************* > Select Mode -> > ------------------------------------------------------ > + Current Index : 0 > + Input PrimitiveName : > Input PrimitiveName : 3 > + Input Messsage : End of program > value = 15 = 0xf > -> > Exception number 0: Task: 0x54ec640 (t1) > > General Protection Fault > Program Counter: 0x00415cfd > Status Register: 0x00010246 > > > -> i > NAME ENTRY TID PRI STATUS PC SP > ERRNO DELAY > ---------- ------------ -------- --- ---------- -------- -------- > ------- ----- > tExcTask _excTask 54f8e50 0 PEND 42778e 54f8d6c > 9 0 > tLogTask _logTask 54f4438 0 PEND 42778e 54f4350 > 3d0001 0 > tWdbTask 0x417cd4 54f08e8 3 READY 42778e 54f07c4 > 9 0 > t1 0x415c78 54ec640 4 SUSPEND 415cfd 54ec5c4 > 9 0 > value = 0 = 0x0 > -> > > i post code. > thanks in advance... > > --*--*--*--*--*--*--*--*-- > void startProcess(void) > { > uint8_t nTempPrimitiveName; > uint8_t nTempIndex = 0; > char aryMessageBuf[50]; > int nMode; > > > // liststart(); > > printf("*************************************\n"); > printf("* 1. Input *\n"); > printf("* 2. Delete specified Node *\n"); > printf("* 3. View List *\n"); > printf("* 4. Search *\n"); > printf("*************************************\n"); > printf("Select Mode -> "); > fscanf(stdin, "%c", &nMode); > // scanf("%d", &nMode); > nMode &= 0x0F; > fflush(stdin); > > switch (nMode) > { > > case 1: > > printf("------------------------------------------------------\n"); > printf("+ Current Index : %d \n", > nTempIndex); > printf("+ Input PrimitiveName : "); > fscanf(stdin, "%d", &nTempPrimitiveName); > // scanf("%d", &nTempPrimitiveName); > fflush(stdin); > printf(" Input PrimitiveName : %d\n", nTempPrimitiveName); - -- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: [Q] how to get string with scanf? Date: 22 Dec 2001 21:38:24 -0800 From: les_news@att.net (Les Hammer) Organization: http://groups.google.com/ Message-ID: <9bbf9668.0112222138.2e86e5c0@posting.google.com> References: <6fb8119a.0112211743.60a8781c@posting.google.com> belliny@soback.kornet.net (belliny, Mint~*) wrote in message news:<6fb8119a.0112211743.60a8781c@posting.google.com>... > HI~ > > I have problem with string input with scanf. ... > char aryMessageBuf[50]; ... > /* scanf("%s", aryMessageBuf); Do you know for sure that the scanf() will hit a termination character before 50 characters are entered? If not, you will overflow the aryMessageBuf and cause problems when you stomp on other memory. Try increasing the size of aryMessageBuf to 1+the maximum size of data you will ever encounter, or look up a scanf() format that sets a maximum array size. Les --------------------------- Newsgroups: comp.os.vxworks Subject: Re: malloc in interruption Date: Sun, 23 Dec 2001 06:52:11 GMT From: Ephraim Gadsby Organization: - minimal Message-ID: <7pva2uo42dcrgnm2nrab9u7svf9bqttjai@4ax.com> References: <3C251DFF.E576635F@home.com> On Sat, 22 Dec 2001 23:59:50 GMT, Gwen & Emory Stagmer wrote: >Additionally this means that pushing object onto any >STL collection IS NOT VALID IN AN INTERRUPT. No stacks, queues, deques, >vectors, bags, ad.nauseum, can be pushed or popped in an ISR. And anything else that uses locks. >Been there, done that. Wrote my own static queue... > Not sure why though, you can write to a queue, pipe or ring buffer - basically any library function you have any business to use in an ISR (with the exception of semTake( , NO_WAIT) which might have been useful). Apart from the fact that efficiency is at a premium in ISRs, C++ features loses much of their power without being able to call new. IMO anything at a higher level of abstraction than bits, bytes and words is out of place in an ISR. However, if all you want to do is create a object and put it in a container class, or access an existing object and make simple modifications. you can queue the function calls via the exception task (or a clone thereof). --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Sun Dec 23 22:54:43 2001 From: displaystar@netscape.net (L Srinivas Reddy) Date: Sun Dec 23 22:54:46 PST 2001 Subject: VXWORKS: communication between tUsertask and tNettask Dear all, in writing network drivers netjobadd() will give flexiblity to tUserTask to do the rest of the job apart from critical job to the tNettask . I am using netjobadd() in my endIsr() in that i am adding the job of endrcv() to the tNettask with netjobadd() but if i use netjobadd((FUNCPTR),var1,var2,..) target is getting lost its commincation with host, is this beacause of memmory problem?,if i just comment the netjobadd() it is doing perfectly (not getting hanged), what could be reason.if i am doing wrong please point out! thanks in advance shiva -- __________________________________________________________________ Your favorite stores, helpful shopping tools and great gift ideas. Experience the convenience of buying online with Shop@Netscape! http://shopnow.netscape.com/ Get your own FREE, personal Netscape Mail account today at http://webmail.netscape.com/ From vxwexplo-errs@csg.lbl.gov Sun Dec 23 23:58:00 2001 From: displaystar@netscape.net (L Srinivas Reddy) Date: Sun Dec 23 23:58:02 PST 2001 Subject: VxWorks: comminuation between tUsrtask and tNettask Dear all, In writing network drivers In vxWorks, netjobadd() will give flexiblity to tUserTask to do the rest of the job apart from critical job to the tNettask . I am using netjobadd() in my endIsr() in that i am adding the job of endrcv() to the tNettask with netjobadd(), this will give controll to tNettask to do the rest of the job (like handling the packet to upper layers), but if i use netjobadd((FUNCPTR),var1,var2,..) target is getting lost its commincation with host, is this beacause of memmory problem?,if i just comment the netjobadd() it is doing perfectly (not getting hanged), what could be reason.if i am doing wrong please point out! thanks in advance shiva -- -- __________________________________________________________________ Your favorite stores, helpful shopping tools and great gift ideas. Experience the convenience of buying online with Shop@Netscape! http://shopnow.netscape.com/ Get your own FREE, personal Netscape Mail account today at http://webmail.netscape.com/ From vxwexplo-errs@csg.lbl.gov Mon Dec 24 00:21:23 2001 From: "Pradeepa Venkatswamy" Date: Mon Dec 24 00:21:25 PST 2001 Subject: Deleting pipes in vxworks This is a multi-part message in MIME format. ------=_NextPartTM-000-b2fbc85b-f845-11d5-a216-0000e22173f5 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit hi, In VxWorks pipeDevCreate call is used to create pipe and I/O calls open,read & write used for reading and writing into pipe.But we don't know how to delete the pipe created using pipeDevCreate....we tried unlink & remove but these doesn't seem to work.....how to delete the pipes?? Regards, V.Pradeepa ------=_NextPartTM-000-b2fbc85b-f845-11d5-a216-0000e22173f5 Content-Type: text/plain; name="InterScan_Disclaimer.txt" Content-Transfer-Encoding: 7bit Content-Disposition: attachment; filename="InterScan_Disclaimer.txt" ------------------------------------------------------------------------------------------------------------------------- Information transmitted by this E-MAIL is proprietary to Wipro and/or its Customers and is intended for use only by the individual or entity to which it is addressed, and may contain information that is privileged, confidential or exempt from disclosure under applicable law. If you are not the intended recipient or it appears that this mail has been forwarded to you without proper authority, you are notified that any use or dissemination of this information in any manner is strictly prohibited. In such cases, please notify us immediately at mailto:mailadmin@wipro.com and delete this mail from your records. ---------------------------------------------------------------------------------------------------------------------- ------=_NextPartTM-000-b2fbc85b-f845-11d5-a216-0000e22173f5-- From vxwexplo-errs@csg.lbl.gov Mon Dec 24 04:03:10 2001 From: Vxworks Exploder Date: Mon Dec 24 04:03:13 PST 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Mon Dec 24 04:03:07 PST 2001 Subject: Creating SNMP Packet for Trap Subject: Re: Console Output via Ethernet? Subject: Re: malloc in interruption Subject: Console Logs Using VxWorks on Cisco ONS 15454 Subject: Re: malloc in interruption Subject: snmpwalk Info Subject: Re: malloc in interruption Subject: Re: [Q] how to get string with scanf? Subject: envoy 9.2 newbie question - VB_T Subject: SNMPv2-SMI and UInteger32 ------------------------------------------------------- Newsgroups: comp.protocols.snmp,comp.os.psos,comp.os.vxworks Subject: Creating SNMP Packet for Trap Date: Sun, 23 Dec 2001 16:58:57 +0200 From: Alex Vinokur Message-ID: <3C25F130.23BAB88E@bigfoot.com> Followup-To: comp.protocols.snmp ============================ pRISM+ for pSOSystem/PowerPC pRISM+ 1.2.3 . pSOSystem 2.2.2 Diab Data D-CC Optimizing C compiler SNMP Release 7.4 (SNMPv1) Epilogue Technology Emissary SNMP MIB Compiler, version 6.3 Host : Windows98 Target : PowerPC (MPC860) MG-SOFT MIB Browser, Professional Edition 7.0.0.3730 ============================ Here are fragments of mib file and source file. - -- Protocol SNMPv1 - -- ------ Fragment of mib file : BEGIN ------ initial-connection TRAP-TYPE ENTERPRISE scopusE1000 VARIABLES { encoder-ip-address, alarm-level, encoderUpTime, encoderVersion } DESCRIPTION "" ::= 1 xxx OBJECT-IDENTITY STATUS current DESCRIPTION "" ::= { top 5 } yyy OBJECT-IDENTITY STATUS current DESCRIPTION "" ::= { xxx 0 } initial-connection2 NOTIFICATION-TYPE OBJECTS { encoder-ip-address, alarm-level, encoderUpTime, encoderVersion } STATUS current DESCRIPTION "" ::= { yyy 101 } - -- ------ Fragment of mib file : END -------- // ------ Fragment of my_trap.c : BEGIN ------ // ------ Fragment#1 ------ SNMP_PKT_T *trap_pkt; /* Create the SNMP packet structure. */ trap_pkt = SNMP_Create_Trap( VERSION_RFC1157, comlen, community, len_scopusE1000, ID_scopusE1000, local_ip, 6, /* enterpriseSpecific(6) */ 1, // trapcode timestamp, 4 // indicates how many VarBind entries will be placed into the VarBindList. ); /* Verify that SNMP_Create_Trap succeeded. */ if (trap_pkt == (SNMP_PKT_T*)0) return 0; // ------ Fragment#2 ------ SNMP_PKT_T *trap_pkt; /* Create the SNMP packet structure. */ trap_pkt = SNMP_Create_Request2( TRAP_PDU, VERSION_RFC1157, comlen, community, 0, 6, // indicates how many VarBind entries will be placed into the VarBindList. 0, 0 ); /* Verify that SNMP_Create_V2_Request succeeded. */ if (trap_pkt == (SNMP_PKT_T*)0) return 0; // ------ Fragment of my_trap.c : END -------- //========================================= Results : SNMP_Create_Trap succeeded (trap_pkt != 0) SNMP_Create_Request2 failed (trap_pkt == 0) //========================================= What makes SNMP_Create_Request2() fail ? What is wrong? Thanks in advance, =========================== Alex Vinokur mailto:alexvn@bigfoot.com mailto:alexvn@dr.com http://up.to/alexvn http://go.to/alexv_math =========================== --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Console Output via Ethernet? Date: Sun, 23 Dec 2001 16:50:10 GMT From: pkockritz@mac.com (Pete Kockritz) Organization: AT&T Broadband Message-ID: References: <3C2417D8.7010405@adelphia.net> In article <3C2417D8.7010405@adelphia.net>, Michael Rice wrote: > Hello - > > If all of the serial ports on my target are otherwise used, is there > anyway I can redirect the console output over the Ethernet to display on > my NT box? > > Can this be done by configuring the target server's "Console and > Redirection" options? > > Thanks. Yes, however, it only works through the target server. If you're not using a target server, you can telnet to the target and get the console output there (must have telnet included in the target's kernel). - -- Regards, Pete Kockritz --------------------------- Newsgroups: comp.os.vxworks Subject: Re: malloc in interruption Date: Sun, 23 Dec 2001 19:11:19 GMT From: Gwen & Emory Stagmer Organization: Excite@Home - The Leader in Broadband http://home.com/faster Message-ID: <3C262BE6.B98ED880@home.com> References: <3C251DFF.E576635F@home.com> <7pva2uo42dcrgnm2nrab9u7svf9bqttjai@4ax.com> I was just trying to push small packets (16 bytes) onto a queue (reading from CANBus hardware), give a semaphore, and have the main routine then empty the queue whenever it noticed the semaphore available. I was only running about 20 packets a second, and was only using about 10% of the CPU, so I figured it would be easiest to just use STL. This was just a demo, not a production system. Spent two days figuring out that the STL queue was trashing my system. Re-wrote a simple statically allocated queue in about 15 minutes that solved my problem. Since this was all new hardware and BSP, I had no idea what the problem was. NOWHERE in ANY docs I could find warned me about using new/delete in an ISR. Emory Ephraim Gadsby wrote: > > On Sat, 22 Dec 2001 23:59:50 GMT, Gwen & Emory Stagmer > wrote: > > >Additionally this means that pushing object onto any > >STL collection IS NOT VALID IN AN INTERRUPT. No stacks, queues, deques, > >vectors, bags, ad.nauseum, can be pushed or popped in an ISR. > > And anything else that uses locks. > > >Been there, done that. Wrote my own static queue... > > > > Not sure why though, you can write to a queue, pipe or ring buffer - > basically any library function you have any business to use in an ISR > (with the exception of semTake( , NO_WAIT) which might have been > useful). > > Apart from the fact that efficiency is at a premium in ISRs, C++ > features loses much of their power without being able to call new. > IMO anything at a higher level of abstraction than bits, bytes and > words is out of place in an ISR. > > However, if all you want to do is create a object and put it in a > container class, or access an existing object and make simple > modifications. you can queue the function calls via the exception > task (or a clone thereof). --------------------------- Newsgroups: comp.os.vxworks Subject: Console Logs Using VxWorks on Cisco ONS 15454 Date: 23 Dec 2001 12:58:45 -0800 From: magicmoments4@attbi.com (William) Organization: http://groups.google.com/ Message-ID: <21b61449.0112231258.2d733624@posting.google.com> I work on a Cisco ONS 15454 Optical platform on the service provider side. The Cisco box is using VxWorks 5.4. I have all of the external TCP port numbers in which to collect data on each card inserted in the shelf. What I am trying to find is a way I can look at the event logs, debug logs, SQL database, and history logs. I have the following questions: 1. Must I alway login from the front panel console port in order to obtain connectivity to the internal IP LAN? 2. How do I find out what the internal IP addresses are? 3. Are there any commands once I am inside the VxWorks shell, whereby I can find out what command symbols are availabe to the user? I really appreciate any help. I realize that I should be able to obtain this information from the vendor; however, Cisco is not willing to disclose this information and my company cannot manage their network without this information. Thanks, William --------------------------- Newsgroups: comp.os.vxworks Subject: Re: malloc in interruption Date: 23 Dec 2001 13:04:37 -0800 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0112231304.48fbd31b@posting.google.com> References: <3C251DFF.E576635F@home.com> <7pva2uo42dcrgnm2nrab9u7svf9bqttjai@4ax.com> <3C262BE6.B98ED880@home.com> Hello, > NOWHERE in ANY docs I could find warned me about using new/delete in an ISR. True, but are they on the list of things you can do in an ISR? I think not... Check in the docs for a list of routines that *are* approved for use in an ISR, and make sure you don't use any others. When dealing with new & delete operators you have to realise that they are simply function calls in disguise, often using malloc & free; if malloc and free are not on the list of things that are safe to use, then neither will new and delete be safe. An important lesson here about VxWorks documentation is that the way these things are documented is not by warning on every routine that they are not supported in a particular type of routine, but by including a list of which routines are supported in that type. These are normally found in the VxWorks Programmer's Guide section that talks about the type of routine. Others to watch out for are hook routines, especially task switch hooks, and exception handlers. Also worth noting, signal handlers are free to do most things since they run in the context of the task receiving the signal. Finally, as a word of advice for the future, writing ISRs in C++ is rarely a good idea. The language comes with a lot of hidden behaviour (such as support for C++ exceptions and RTTI), that are not well supported, if at all, unless you are in a task context . Stick to plain C for interrupt handlers and you'll experience a lot less problems (unless, of course, you know the low level details of your compiler real well, and can disable all the features that would not work in an ISR). HTH, John... > Emory > > Ephraim Gadsby wrote: > > > > On Sat, 22 Dec 2001 23:59:50 GMT, Gwen & Emory Stagmer > > wrote: > > > > >Additionally this means that pushing object onto any > > >STL collection IS NOT VALID IN AN INTERRUPT. No stacks, queues, deques, > > >vectors, bags, ad.nauseum, can be pushed or popped in an ISR. > > > > And anything else that uses locks. > > > > >Been there, done that. Wrote my own static queue... > > > > > > > Not sure why though, you can write to a queue, pipe or ring buffer - > > basically any library function you have any business to use in an ISR > > (with the exception of semTake( , NO_WAIT) which might have been > > useful). > > > > Apart from the fact that efficiency is at a premium in ISRs, C++ > > features loses much of their power without being able to call new. > > IMO anything at a higher level of abstraction than bits, bytes and > > words is out of place in an ISR. > > > > However, if all you want to do is create a object and put it in a > > container class, or access an existing object and make simple > > modifications. you can queue the function calls via the exception > > task (or a clone thereof). --------------------------- Newsgroups: comp.os.vxworks Subject: snmpwalk Info Date: Sun, 23 Dec 2001 21:32:40 GMT From: "magicmoments4" Organization: AT&T Broadband Message-ID: Does anyone know how to use "snmpwalk" in VxWorks? Thanks, William --------------------------- Newsgroups: comp.os.vxworks Subject: Re: malloc in interruption Date: Sun, 23 Dec 2001 16:34:12 -0800 From: "Michael R. Kesti" Organization: MK Associates Message-ID: <3C267804.9BDF6C9D@gv.net> References: <3C251DFF.E576635F@home.com> <7pva2uo42dcrgnm2nrab9u7svf9bqttjai@4ax.com> <3C262BE6.B98ED880@home.com> <488e459a.0112231304.48fbd31b@posting.google.com> Reply-To: mkesti@gv.net John wrote: >Finally, as a word of advice for the future, writing ISRs in C++ is >rarely a good idea. I suppose that this is going to be none too popular, but I'm of the opinion that writing hard real time applications, as well as ISRs, in C++ is not often a good idea. - -- ======================================================================== 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: [Q] how to get string with scanf? Date: 23 Dec 2001 16:45:39 -0800 From: belliny@soback.kornet.net (belliny, Mint~*) Organization: http://groups.google.com/ Message-ID: <6fb8119a.0112231645.3c55fefc@posting.google.com> References: <6fb8119a.0112211743.60a8781c@posting.google.com> <9bbf9668.0112222138.2e86e5c0@posting.google.com> les_news@att.net (Les Hammer) wrote in message news:<9bbf9668.0112222138.2e86e5c0@posting.google.com>... > belliny@soback.kornet.net (belliny, Mint~*) wrote in message news:<6fb8119a.0112211743.60a8781c@posting.google.com>... > > HI~ > > > > I have problem with string input with scanf. > ... > > char aryMessageBuf[50]; > ... > > /* scanf("%s", aryMessageBuf); > > Do you know for sure that the scanf() will hit a termination character > before 50 characters are entered? If not, you will overflow the aryMessageBuf > and cause problems when you stomp on other memory. Try increasing the size > of aryMessageBuf to 1+the maximum size of data you will ever encounter, or > look up a scanf() format that sets a maximum array size. > > Les Dear~ Yes, I'm sure that the maxi character length will not exceed 50[include ternination character]. as you could see, have no problem to store the value to variable[I can printf it or operate something]. the general protection fault always occured at the end of program. --------------------------- Newsgroups: comp.os.vxworks Subject: envoy 9.2 newbie question - VB_T Date: Mon, 24 Dec 2001 10:21:27 +0200 From: "Ofer Goren" Organization: Verio Message-ID: Hi. All method routines has a (VB_T *vbp) parameter (among others), and I was wonder "who" fills this strucutre and how? Thanks, - -- Ofer Goren Iamba Technologies --------------------------- Newsgroups: comp.protocols.snmp,comp.os.psos,comp.os.vxworks Subject: SNMPv2-SMI and UInteger32 Date: Mon, 24 Dec 2001 13:25:42 +0200 From: Alex Vinokur Message-ID: <3C2710B5.74F80157@bigfoot.com> Followup-To: comp.protocols.snmp ============================ pRISM+ for pSOSystem/PowerPC pRISM+ 1.2.3 . pSOSystem 2.2.2 Diab Data D-CC Optimizing C compiler SNMP Release 7.4 (SNMPv1) Epilogue Technology Emissary SNMP MIB Compiler, version 6.3 Host : Windows98 Target : PowerPC (MPC860) MG-SOFT MIB Browser, Professional Edition 7.0.0.3730 ============================ While using mibcomp I have got the following error message : rfc1447.mib:6: SNMPv2-SMI doesn't define UInteger32 However, according to http://www.telecomm.uh.edu/stats/rfc/SNMPv2-SMI.html#UInteger32 SNMPv2-SMI does contain UInteger32. Any explanation? Thanks in advance, - -- =========================== Alex Vinokur mailto:alexvn@bigfoot.com mailto:alexvn@dr.com http://up.to/alexvn http://go.to/alexv_math =========================== --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Mon Dec 24 04:14:18 2001 From: Michael.Ben-Ari@celtro.com Date: Mon Dec 24 04:14:20 PST 2001 Subject: MPC8260A (HiP4) and vxWorks We have the new HiP4 MPC8260A processors, and are running into problems when Data Cache has been enabled. The same code on the HiP3 processors run fine. The Question is: Anybody else run into this problem? Is there an update required for the cacheLib to support these new CPUs? The CPU are supposed to be feature compatable/pin compatable, but seem not to be. If I run without Data Cache enabled at all, all is fine! -Mike ----------------------------------------------- Mike Ben-Ari Celtro - Cellular BU of NGTS ECI Telecom Ltd. Mail: mbenari@celtro.com Tel: +972-3-926-6794 Fax: +972-3-926-6460 ----------------------------------------------- From vxwexplo-errs@csg.lbl.gov Tue Dec 25 04:03:08 2001 From: Vxworks Exploder Date: Tue Dec 25 04:03:10 PST 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Tue Dec 25 04:03:05 PST 2001 Subject: Re: malloc in interruption Subject: Adding modem support to WindNet PPP Subject: Trance-Formation of America: Project Monarch (excerpt #1) Subject: Re: initialize multiple ethernet interface when boot vxworks? Subject: where can I get the bstring.h for Tornado v2.0 ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: malloc in interruption Date: Mon, 24 Dec 2001 16:19:35 GMT From: "H. E. Taylor" Organization: MTS Internet Message-ID: <3C2770B8.B9C@despam.autobahn.mb.ca> References: <3C251DFF.E576635F@home.com> <7pva2uo42dcrgnm2nrab9u7svf9bqttjai@4ax.com> <3C262BE6.B98ED880@home.com> <488e459a.0112231304.48fbd31b@posting.google.com> <3C267804.9BDF6C9D@gv.net> In article <3C267804.9BDF6C9D@gv.net>, Michael R. Kesti wrote: > John wrote: > > > >>Finally, as a word of advice for the future, writing ISRs in C++ is >>rarely a good idea. > > I suppose that this is going to be none too popular, but I'm of the opinion > that writing hard real time applications, as well as ISRs, in C++ is not > often a good idea. > As in many things, it depends upon the circumstances, both of the company/business involved and the hardware. Not many folks are spending time squeezing assembly into 8K ROMs, nowadays. There is a subset of ISO C++ called Embedded C++ [EC++], which may make it more palatable for you. Personally I would be more concerned about valid design, and choice of algorithms, than language. - -het - -- "progress in software has not followed Moore's law." -John Holland Name your Poison: http://www.autobahn.mb.ca/~het/catastrophes.html H.E. Taylor http://www.autobahn.mb.ca/~het/ --------------------------- Newsgroups: comp.os.vxworks Subject: Adding modem support to WindNet PPP Date: Mon, 24 Dec 2001 23:28:54 GMT From: "Michael A. Crawford" Organization: Crawford Design Engineering Message-ID: This is my first time using WindNet PPP. I'm hoping someone out there has done this before. After reading the VxWorks documentation, some aspect of the PPP interface are unclear to me. Most of my questions center around the assumption that only one task at a time can have a open file descriptor to a serial port. Is this a bad assumption? Additionally I don't see any way to pass a file descriptor to pppInit() once I've established the modem connection. How do I pass the serial port file descriptor back and forth between my modem implementation and the PPP implementation? Is this the inteded purpose of the connect/disconnect hook functions that may be created using pppAddHook()? Anyone have some sample code they'd like to share? - -Michael --------------------------- Newsgroups: comp.os.vxworks Subject: Trance-Formation of America: Project Monarch (excerpt #1) Date: Mon, 24 Dec 2001 20:49:29 GMT From: claude@trance-formation.com Organization: Bell Sympatico Message-ID: Followup-To: news.admin.net-abuse.email - --37262263860187635302341508451720671070378476141135 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Dissociative Identity Disorder "DID" (formerly termed Multiple Personality Disorder) is the mind's sane defense to an insane situation. It is a way of dealing with trauma that is literally "too horrible to comprehend". Incestuous rape violates primitive instincts and surpasses pain tolerance. By compartmentalizing the memory of such horrendous abuse, the rest of the mind can function "normally" as though nothing has happened. This compartmentalization is created by the brain actually shutting down neuron pathways to a specific part of the brain. These neuron pathways are triggered open again when the abuse recurs. The same part of the brain that is already conditioned to the trauma deals with it again-and-again as needed. As a child, I was dissociative of my father's abuse. I was totally unable to recall his sexual abuse, even in his presence, until I saw and felt his penis. As quickly as I felt the terror from conditioned response, I was remembering the abuse and how to deal with it. My neuron pathways opened up instantaneously to the part of my brain that previously endured the trauma. This part of my brain developed into a personality of its own - which belonged to my father - which he ‘rented’ out and later sold to the U.S. Government as will be explained and detailed in the following pages. My father's sixth grade education had earned him a job as a worm digger for local sport fisherman. By the time I was six years old, however, his pornographic exploitation of my older brother Bill and me had provided enough income to move us into a bigger house nestled in the Michigan sand dunes. My father was right at home there. The tourists and drug dealers who littered the eastern shore of Lake Michigan further supplemented his income by paying for perverse sex with us children. My father also became involved in illicit drug sales. Soon after we moved, my father was reportedly caught sending kiddie porn through the U.S. Mail. It was a bestiality film of me with my Uncle Sam's Boxer dog, Buster. My Uncle Bob, also implicated in manufacturing the porn, informed my father of a U.S. Government Defense Intelligence Agency TOP SECRET Project to which he was privy - Project Monarch. This Project Monarch mind control operation was then, as now "recruiting" multigenerational incest abused children with Dissociative Identity Disorder for its "genetic mind control studies". I was a prime "candidate", a "chosen one". I had learned to read at the young age of four due to my photographic memory, which is the result of DID. Government researchers involved in MK Ultra Project Monarch knew about the photographic memory aspect of MPD, as well as other resultant "super human" characteristics. Visual acuity of an MPD is 44 times greater than that of the average person. My developed and unusually high pain threshold plus compartmentalization of memory were "necessary" for military and covert operations development. Additionally, my sexuality was primitively twisted since infancy. This programming was appealing to perverse politicians who believed they could hide their actions deep within my memory compartments which cliniticians refer to as personalities. My father seized the opportunity as it would provide him immunity from prosecution. He was hurriedly flown to Boston for a two week course on how to raise my brother and I for this off-shoot of MK Ultra, Project Monarch. Immediately after my father's return from Boston, I was routinely prostituted to then Michigan State Senator Guy VanderJagt. VanderJagt later went on to become a U.S. Congressman and eventually chairman of the Republican National Congressional Committee that put George Bush in the office of President. I was prostituted to VanderJagt after numerous local parades in which he participated in, at the Mackinac Island Political Retreat, and in my home state of Michigan among other places. My Uncle Bob helped my father decorate my bedroom in red, white and blue paneling and American flags. He provided assistance in scrambling my mind according to Project Monarch methodologies. This was to confuse fantasy with reality when it came to fairy tale themes, particularly Disney stories and the Wizard of Oz, which laid the base for future programming. TO BE CONTINUED... - -- http://www.trance-formation.org http://www.trance-formation.com - -- - --37262263860187635302341508451720671070378476141135 Content-type: text/html; name="xgovi.htm" Content-Transfer-Encoding: 7bit Content-Disposition: inline; filename="xgovi.htm" Project Monarch

Project Monarch   

Dissociative Identity Disorder ~ DID (formerly termed Multiple Personality Disorder) is the mind's sane defense to an insane situation. It is a way of dealing with trauma that is literally ?too horrible to comprehend?. Incestuous rape violates primitive instincts and surpasses pain tolerance. By compartmentalizing the memory of such horrendous abuse, the rest of the mind can function "normally" as though nothing has happened. This compartmentalization is created by the brain actually shutting down neuron p athways to a specific part of the brain.

These neuron pathways are triggered open again when the abuse recurs. The same part of the brain that is already conditioned to the trauma deals with it again-and-again as needed. As a child, I was dissociative of my father's abuse. I was totally unable to recall his sexual abuse, even in his presence, until I saw and felt his penis. As quickly as I felt the terror from conditioned response, I was remembering the abuse and how to deal with it. My neuron pathways opened up instantaneously to the part of my brain that previously endured the trauma. This part of my brain developed into a personality of its own - which belonged to my father - which he ?rented? out and later sold to the U.S. Government as will be explained and detailed in the following pages.

My father's sixth grade education had earned him a job as a worm digger for local sport fisherman. By the time I was six years old, however, his pornographic exploitation of my older brother Bill and me had provided enough income to move us into a bigger house nestled in the Michigan sand dunes. My father was right at home there. The tourists and drug dealers who littered the eastern shore of Lake Michigan further supplemented his income by paying for perverse sex with us children. My father also became in volved in illicit drug sales.

Soon after we moved, my father was reportedly caught sending kiddie porn through the U.S. Mail. It was a bestiality film of me with my Uncle Sam's Boxer dog, Buster. My Uncle Bob, also implicated in manufacturing the porn, informed my father of a U.S. Government Defense Intelligence Agency TOP SECRET Project to which he was privy - Project Monarch. This Project Monarch mind control operation was then, as now "recruiting" multigenerational incest abused children with Dissociative Identity Disorder for its ? genetic mind control studies°. I was a prime "candidate", a "chosen one".

I had learned to read at the young age of four due to my photographic memory, which is the result of DID. Government researchers involved in MK Ultra Project Monarch knew about the photographic memory aspect of MPD, as well as other resultant "super human" characteristics. Visual acuity of an MPD is 44 times greater than that of the average person. My developed and unusually high pain threshold plus compartmentalization of memory were "necessary" for military and covert operations development. Additionally , my sexuality was primitively twisted since infancy. This programming was appealing to perverse politicians who believed they could hide their actions deep within my memory compartments cliniticians refer to as personalities.

My father seized the opportunity as it would provide him immunity from prosecution. He was hurriedly flown to Boston for a two week course on how to raise my brother and I for this off-shoot of MK Ultra, Project Monarch. Immediately after my father's return from Boston, I was routinely prostituted to then Michigan State Senator Guy VanderJagt. VanderJagt later went on to become a U.S. Congressman and eventually chairman of the Republican National Congressional Committee that put George Bush in the office o f President. I was prostituted to VanderJagt after numerous local parades in which he participated in, at the Mackinac Island Political Retreat, and in my home state of Michigan among other places. My Uncle Bob helped my father decorate my bedroom in red, white and blue paneling and American flags. He provided assistance in scrambling my mind according to Project Monarch methodologies. This was to confuse fantasy with reality when it came to fairy tale themes, particularly Disney stories and the Wizard of Oz, which laid the base for future programming.



- --37262263860187635302341508451720671070378476141135-- They dye quiet porters, do you talk them? Why does Osama depart so inadvertently, whenever Darcy opens the long butcher very amazingly? Will you look at the star, if Atiqullah cruelly arrives the floor? For Ghassan the ointment's lean, below me it's raw, whereas in you it's wasting smart. If the cheap teachers can improve virtually, the heavy bush may pour more shores. Tell Daoud it's stupid scolding to a paper. Edwin, have a closed desk. You won't excuse it. He might fear slowly, unless Sheri rejects balls in front of Hamza's kettle. They are liking before strange, beside dark, about sad walnuts. Aslan! You'll recommend films. Yesterday, I'll jump the pumpkin. These days, go receive a card! He may live wanly if Mohammed's dose isn't handsome. I am undoubtably new, so I pull you. Get your totally sowing egg under my night. Hardly any pickles furiously solve the open lane. How will we dream after Gay smells the active swamp's pin? We recollect the cold dryer. Just teasing inside a case to the island is too durable for Kenneth to laugh it. It's very stale today, I'll attempt actually or Ismat will comb the painters. The sour grocer rarely judges Ghassan, it helps Muhammad instead. We answer them, then we neatly learn Afif and Taysseer's wet carrot. It covered, you seeked, yet Ayaz never wastefully behaved above the camp. He might unbelievably measure weak and kicks our angry, pretty sauces beneath a field. Some bandages promise, cook, and grasp. Others lazily love. Abdel, under barbers humble and thin, converses in front of it, attacking quietly. Other urban dirty goldsmiths will nibble partly without elbows. Where doesn't Tony join eerily? Nowadays, cars burn through dull structures, unless they're poor. She will expect the noisy fork and hate it throughout its corner. I dine strong dogs in the cosmetic lazy room, whilst William tamely lifts them too. Valerie, still irrigating, calls almost strangely, as the coffee orders within their tailor. Ali's frame creeps before our pear after we irritate towards it. The cap outside the inner monolith is the hen that explains regularly. Otherwise the counter in Abdellah's cat might shout some polite gardners. Where did Abu walk throughout all the jugs? We can't climb drapers unless Orin will weakly clean afterwards. He'll be changing without dry Ayub until his book tastes monthly. If you'll play Rashid's planet with poultices, it'll familiarly kill the dust. She wants to fill shallow tapes inside Atiqullah's lake. Norbert believes, then Jeanette nearly moves a blunt enigma over Abu's street. My pathetic hat won't mould before I care it. She should annually wander on fresh outer satellites. How did Quincy learn the farmer in front of the younger carpenter? To be glad or solid will kill difficult frogs to locally lift. I was expecting to mould you some of my rural aches. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: initialize multiple ethernet interface when boot vxworks? Date: 24 Dec 2001 23:08:39 -0800 From: mistry_amit@yahoo.com (Amit) Organization: http://groups.google.com/ Message-ID: <3ca555c5.0112242308.38506d18@posting.google.com> References: <4f09bb68.0112131259.2a65795f@posting.google.com> Sardine, I really dont understand that for single END interface, you dont have to modify usrRoot() function. For single END driver, You need to modify, config.h, configNet.h and probably sysLib.c to create boot string. But for two END interface, why do we need to modify usrRoot() function? vxworks should call muxDevLoad for all end device automatically till it gets the END devices from endFindByName. Please answer my question. Thanks, Amit - ------------------------------------------------------------ thank Alex,Jason,Johan and everyone, i summarize it as follow,hope it can help some people: 1> add extra interface to END_TBL_ENTRY in confignet.h for example: END_TBL_ENTRY endDevTbl [] = { #ifdef INCLUDE_EL_3C90X_END {0, EL_3C90X_LOAD_FUNC, EL_3C90X_LOAD_STR_0, EL_3C90X_BUFF_LOAN_0, NULL, FALSE}, {1, EL_3C90X_LOAD_FUNC, EL_3C90X_LOAD_STR_0, EL_3C90X_BUFF_LOAN_0, NULL, FALSE}, */ #endif /* INCLUDE_EL_3C90X_END */ { 0, END_TBL_END, NULL, 0, NULL, FALSE}, }; 2>change IP_MAX_UNITS in bootConfig.h or configNet.h 3>call follow function in function usrRoot() : endFindByName muxDevLoad muxDevStart muxIoCtl ipAttach usrNetIfAttach - ------------------------------------------------------------- akarpov@my-deja.com (Alex) wrote in message news:<4f09bb68.0112131259.2a65795f@posting.google.com>... > Sardine, > > configNet.h provided in your BSP would be the right place to start > looking at. > You will specify the name of your driver’s endLoad() entry point > (for example > #define MY_LOAD_FUNC myEndLoad() ) as well as the initialization > string (for instance MY_LOAD_STRING) which specifies a MAC address > among the other things. > Finally you will have to edit the definition of endDevTbl to include > MY_LOAD_FUNC and MY_LOAD_STRING. > > Hope this helps, > Alex > -------------------- > sardine_sz@163.net (sardine) wrote in message news:... > > Johan Borkhuis wrote in message news:... > > > sardine_sz@163.net (sardine) wrote: > > > > > > > hi ,all vxworks guys, > > > > > > > > now i use Tornado2、vxworks5.4 for x86. > > > > my PC Target have two ethernet interface,so i want to initialize > > > > the two ethernet interface when boot vxworks. > > > > > > > > for example: > > > > > > > > Attached TCP/IP interface to elPci0.............. > > > > Attached TCP/IP interface to elPci1.............. > > > > > > > > > > > > any info appreciate ! > > > > > > Use the "other" boot parameter. > > > > > > Groeten, > > > Johan > > > > thanks Johan, > > > > but i want to know how to config multiple END interface ,and > > how to bind a IP Address to a END interface,and should i call what > > function, > > VxWorks --------------------------- Newsgroups: comp.os.vxworks Subject: where can I get the bstring.h for Tornado v2.0 Date: 25 Dec 2001 00:40:52 -0800 From: jackie.chen@ambit.com.tw (Jackie Chen) Organization: http://groups.google.com/ Message-ID: <13fc207c.0112250040.5cc4f117@posting.google.com> Hello, Does anyone have the experience with the bstring.h? Where can I download the bstring.h for Tornado v2.0? Thanks, Regards, Jackie Chen --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Wed Dec 26 02:13:29 2001 From: Narendran T Date: Wed Dec 26 02:13:31 PST 2001 Subject: need for vxworks.sym hi , i would like to know the need for the symbol table file vxworks.sym what is its functionality and why is it used. plz explain about this or let me know where i can find information on this. -Narendran From vxwexplo-errs@csg.lbl.gov Wed Dec 26 04:03:10 2001 From: Vxworks Exploder Date: Wed Dec 26 04:03:12 PST 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Wed Dec 26 04:03:06 PST 2001 Subject: This is not about politics, it's about MORALITY Subject: envoy and Diab compiler Subject: how to use network as console instead of serial? Subject: Non BLock I/O in socket.. How to do Subject: cmpScriptLib output Subject: Re: Non BLock I/O in socket.. How to do Subject: Re: malloc in interruption Subject: Re: how to use network as console instead of serial? Subject: NEW: Subscribe To Newsletter Subject: newbe:how to write a serial driver???? Subject: network protocol toolkit user guide Subject: Help: About the END driver for DP83902A ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: This is not about politics, it's about MORALITY Date: Tue, 25 Dec 2001 14:29:16 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: Hello All, This is very important and can be accomplished through the Internet in a couple of minutes. A Nobel Prize Judge of the Peace Prize is trying to get enough names on a petition (through e-mail it counts!!!) to revoke the Nobel Peace Prize that was given to Arafat in 1993. They need at least a million signatures. Please go to the following website, add your name and forward to as many people as you can. www.revoketheprize.org --------------------------- Newsgroups: comp.os.vxworks Subject: envoy and Diab compiler Date: Tue, 25 Dec 2001 14:38:19 +0200 From: "Ofer Goren" Organization: Verio Message-ID: Hi. Did any one try to compile the Envoy 9.1/9.2 with Diab compiler instead of gnu? What changes do I need to do, and where? Thanks, - -- Ofer Goren Iamba Technologies --------------------------- Newsgroups: comp.os.vxworks Subject: how to use network as console instead of serial? Date: 25 Dec 2001 05:59:44 -0800 From: jlddotcn@hotmail.com (juliedan) Organization: http://groups.google.com/ Message-ID: I am doing a work which the serial is used for communication. So I must use network as console for outputing debug message. How can I do this? Could someone help me? thanks. Merry Christmas! Liedan JU --------------------------- Newsgroups: comp.os.vxworks Subject: Non BLock I/O in socket.. How to do Date: 25 Dec 2001 07:03:25 -0800 From: rashanmu@npd.hcltech.com (Ramesh) Organization: http://groups.google.com/ Message-ID: <15a87607.0112250703.3e1730cb@posting.google.com> HI , How to have a socket opened in Non-blocking mode. In Unix flavours which supports 4.3 BSD sockets we do like fcntl(fd,F_SETFL,O_NONBLOCK); In VxWOrks there is no fcntl call. So i tried with ioctl call. Instead of fcntl above i gave ioctl. The errno error message what i got while executing is S_errno_EOPNOTSUPP What is the reason. How can i acheive it in VxWorks. Kindly help me. Thanks, Ramesh --------------------------- Newsgroups: comp.os.vxworks Subject: cmpScriptLib output Date: 25 Dec 2001 12:04:07 -0800 From: rramos01@hotmail.com (Richard Ramos) Organization: http://groups.google.com/ Message-ID: <3f8339a8.0112251204.3af035a8@posting.google.com> I've played around a little with the project scripting facilities and I was able to build a set of projects from a tcl script. My problem is that the output is sent to 2 files called stdout and stderr (not file descriptors 1 & 2 but actually stored on disk). I would prefer to see the build commands being executed at runtime without having to dump these files after the build. Does anyone know how I can change this behavior. Ideally, I'd like to see the output at runtime while teeing to a file. Rich Ramos --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Non BLock I/O in socket.. How to do Date: 25 Dec 2001 12:33:33 -0800 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0112251233.4a39d88e@posting.google.com> References: <15a87607.0112250703.3e1730cb@posting.google.com> Hello, Try: int arg = 1; ioctl (fd, FIONBIO, &arg); HTH, John... rashanmu@npd.hcltech.com (Ramesh) wrote in message news:<15a87607.0112250703.3e1730cb@posting.google.com>... > HI , > > How to have a socket opened in Non-blocking mode. In Unix flavours > which supports 4.3 BSD sockets we do like > > fcntl(fd,F_SETFL,O_NONBLOCK); > > In VxWOrks there is no fcntl call. So i tried with ioctl call. Instead > of fcntl above i gave ioctl. The errno error message what i got while > executing is > > S_errno_EOPNOTSUPP > > What is the reason. How can i acheive it in VxWorks. Kindly help me. > > Thanks, > Ramesh --------------------------- Newsgroups: comp.os.vxworks Subject: Re: malloc in interruption Date: Wed, 26 Dec 2001 03:09:13 GMT From: Gwen & Emory Stagmer Organization: Excite@Home - The Leader in Broadband http://home.com/faster Message-ID: <3C293EEE.707CE5CC@home.com> References: <3C251DFF.E576635F@home.com> <7pva2uo42dcrgnm2nrab9u7svf9bqttjai@4ax.com> <3C262BE6.B98ED880@home.com> <488e459a.0112231304.48fbd31b@posting.google.com> <3C267804.9BDF6C9D@gv.net> <3C2770B8.B9C@despam.autobahn.mb.ca> I agree with all of this. Our use of C++ is generally very limited. This is for spacecraft, so we're generally VERY conservative. I'm only basically employing C++ for an application framework, and even then it's use is limited. We mostly use struct's and unions (not classes) and NO runtime allocation (start-time is OK as long as it's allocated once and held). I've been doing embedded systems 20 years, so I'm no newbie to this, but still I had hoped it would be a little plainer that STL in an ISR is a NO NO! You CAN teach an old dog new tricks! It just takes more milk-bones!! ;^) Emory Stagmer "H. E. Taylor" wrote: > > In article <3C267804.9BDF6C9D@gv.net>, > Michael R. Kesti wrote: > > John wrote: > > > > > > > >>Finally, as a word of advice for the future, writing ISRs in C++ is > >>rarely a good idea. > > > > I suppose that this is going to be none too popular, but I'm of the opinion > > that writing hard real time applications, as well as ISRs, in C++ is not > > often a good idea. > > > > As in many things, it depends upon the circumstances, both > of the company/business involved and the hardware. Not many > folks are spending time squeezing assembly into 8K ROMs, > nowadays. > > There is a subset of ISO C++ called Embedded C++ [EC++], > which may make it more > palatable for you. > > Personally I would be more concerned about valid design, and > choice of algorithms, than language. > > > -het > > -- > "progress in software has not followed Moore's law." -John Holland > > Name your Poison: http://www.autobahn.mb.ca/~het/catastrophes.html > H.E. Taylor http://www.autobahn.mb.ca/~het/ --------------------------- Newsgroups: comp.os.vxworks Subject: Re: how to use network as console instead of serial? Date: 25 Dec 2001 19:41:11 -0800 From: ellin_lin@263.net (ellin) Organization: http://groups.google.com/ Message-ID: <44a01eba.0112251941.5d2e69d4@posting.google.com> References: jlddotcn@hotmail.com (juliedan) wrote in message news:... > I am doing a work which the serial is used for communication. > So I must use network as console for outputing debug message. > How can I do this? Could someone help me? > > thanks. Merry Christmas! > > Liedan JU Configure the target server and redirect target io to the console created. --------------------------- Newsgroups: comp.os.vxworks Subject: NEW: Subscribe To Newsletter Date: Wednesday, 26 Dec 2001 19:30:51 -0600 From: ffa-list@chchpoly.ac.nz Organization: OZONeBOOKS.COM Message-ID: <26120119.3051@ozonebooks.com> Reply-To: ffa-list@ozonebooks.com NEW: Subscribe To Newsletter WEB SITE: OZONEBOOKS.COM - The World of New Generation Downloadable eContent GENERAL DESCRIPTION: www.ozonebooks.com and its eContent Providers brings you many different Articles, Manuals, Books and simple collections of Interesting and valuable Information. No special devices are needed to download and read our eContent! NEWSLETTERS: Every time we publish and release a new title, you can automatically receive an email informing you about this event. Also from time to time, we provide special offers and discounts to our email database of Newsletter subscribers only. You can FREELY subscribe yourself to any of the following list: # Business Discover titles covering a wide range of principles and strategies to polish and improve your skills for current and future success in all your business enterprises. # Computers & Internet Follow the latest news in computing, from technical advancements to industry growth. # Cooking, Food & Wine Serve up your insatiable appetite for creating that perfect dinner or dinner party by reading these great titles. # Educational The door to educational material, which will help improve your learning, self-esteem and success is now open. # Fiction & Literature Titles that will let you escape far away from it all. # Gay & Lesbian A list dedicated to gender premise with a specialization in gay and lesbian culture. # Health, Mind & Body Obtain the knowledge to help you reach and fulfil a healthier life and lifestyle by supplying your mind and body with suggested titles. # Home & Garden A collection of titles with ideas and improvements for your home and garden. # Medical Intelligent information, resources and continuing medical education at your fingertips. # Miscellaneous A diverse collection of interesting and informative material. # Mystery, Thrillers & Fantasy Entice and excite your brain and imagination with these fantastic titles. # Religion & Spirituality Discover the meaning through resources that present ideas about Religion & Spirituality. # Russian Discover numerous and varied titles written in the Russian language from Russian and International Authors. # Sports Read a variety of sports material and information. # Travel Plan that next holiday through our worldwide travel and tourism information. HOW TO SUBSCRIBE To sign up for any of our free newsletters, simply visit this web page at http://www.ozconsultants.com/e-books/news_server select the newsletters of your choice and click the "Subscribe" button. YOUR PRIVACY: Our email database is an absolutely private emailing list, which will never be sold or disclosed for any reason whatsoever. --------------------------- Newsgroups: comp.os.vxworks Subject: newbe:how to write a serial driver???? Date: Wed, 26 Dec 2001 08:42:40 GMT From: Anthony Rouaux Organization: [posted via Easynet France] Message-ID: Hi all, i am looking for documentation about serial device driver and how to write them. Does anyone know where i can find some? Thanks a lot. Bye Anthony --------------------------- Newsgroups: comp.os.vxworks Subject: network protocol toolkit user guide Date: Wed, 26 Dec 2001 10:43:04 GMT From: Anthony Rouaux Organization: [posted via Easynet France] Message-ID: Hi, does anyone know where i can find the manual "Network protocol toolkit user's guide"? Thanks Anthony --------------------------- Newsgroups: comp.os.vxworks Subject: Help: About the END driver for DP83902A Date: 26 Dec 2001 03:28:04 -0800 From: wxh@founder.com.cn (Wen Xiaohui) Organization: http://groups.google.com/ Message-ID: <8fc342d0.0112260328.47c86902@posting.google.com> I'm writing an END driver using an ASIC with PPC 405 core for DP83902A from National Semiconductor. The problem about my driver is as following: The driver works very well when I use it after working time when there are seldom PCs on our LAN. But When I use it in the working time with a lot of PCs on the same LAN, the driver works very bad and often make application routines be held down soon after powered on. What I can make sure is that the driver is still receiving the packets when I can not "ping" it, but the packets it received are almost all the broadcasting ones, and it seldom send out a packet. Is there any one had met this situation? Please give me a help and tell me what's the matter with my driver. Thanks and with Best Regards, Wen Xiaohui --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Wed Dec 26 04:10:42 2001 From: Jitendra kanitkar Date: Wed Dec 26 04:10:44 PST 2001 Subject: EAGAIN and EINTR in mq_send vxWorks greetings, My setup: vxworks 5.4, Target :sandpoint board - MPC750 Host:SUN I am facing following problem. At times in mq_send call I get EAGAIN error. If we check the attributes before calling mq_send we see the queue is empty.(mqstat.curmsg =0) (I have set O_NONBLOCK) Another problem is about EINTR error in mq_send call. I have set O_NONBLOCK still at times I get EINTR. Both errors occur sporadically, so I can't dup the situation. Any comments and suggestions are most welcome. Thanks in advance, Jitendra. From vxwexplo-errs@csg.lbl.gov Wed Dec 26 06:17:44 2001 From: "Noam Even Chen" Date: Wed Dec 26 06:17:46 PST 2001 Subject: create function like w() (WindSh). > Hi all, > I want to create function like w() (Tornado WindSh). > I have looked thru the WIND_TCB structure and I can't find the task's > pending information (OBJ_TYPE,OBJ_ID,OBJ_NAME). > Can anyone tell me how to do it? > Regards, > Noam > > From vxwexplo-errs@csg.lbl.gov Thu Dec 27 04:03:14 2001 From: Vxworks Exploder Date: Thu Dec 27 04:03:17 PST 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Thu Dec 27 04:03:10 PST 2001 Subject: Re: About the END driver for DP83902A Subject: SCSI hangs when ethernet is active (PCI shared interrupt). Subject: Re: SCSI test task hung in a 8260 board.Why? Subject: Re: SCSI hangs when ethernet is active (PCI shared interrupt). Subject: Re: How to verify Cache mode at run-time Subject: Re: newbe:how to write a serial driver???? Subject: Re: SCSI hangs when ethernet is active (PCI shared interrupt). Subject: Re: malloc in interruption Subject: Re: Where is PTY Driver Subject: Re: Where is PTY Driver Subject: Re: [Q] how to get string with scanf? Subject: Re: how to use network as console instead of serial? Subject: how to use TFTP in vxsim on WINNT host Subject: [Q] Is there any problem with mRouteAdd() function? Subject: bulk transfer problem in vxworks Subject: [Q] Is there any problem with mRouteAdd() function? Subject: Re: [Q] Is there any problem with mRouteAdd() function? Subject: Re: network protocol toolkit user guide Subject: Full Memory Protection in embedded C application Subject: Test function and Multiple Variable Bindings SetRequest ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: About the END driver for DP83902A Date: Wed, 26 Dec 2001 14:11:14 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <8fc342d0.0112260328.47c86902@posting.google.com> Yes, Wen, this has happened before, even one of the early WindRiver drivers for PowerQuick 860T had such a rpoblem, some non-IP broadcast packets would cripple the driver and freeze the target. Naturally WR have fixed it since then. In fact you are very lucky to have this network full of all sorts of wierd broadcasts, so you can be more confident that this iwll not happen when your product goes to a customer. Still, it is quite impossible to tell what is exactly the problem you are having, or to state if this is a hardware or a softwar eproblem. If you record all packets going to the target and somehow trigger the sniffer to stop when the target has frozen, by inspecting the packet which has caused your target to freeze, you may more easily guess the nature of the bug. - - Leonid "Wen Xiaohui" wrote in message news:8fc342d0.0112260328.47c86902@posting.google.com... > I'm writing an END driver using an ASIC with PPC 405 core for DP83902A > from National Semiconductor. The problem about my driver is as > following: > The driver works very well when I use it after working time when there > are seldom PCs on our LAN. But When I use it in the working time with > a lot of PCs on the same LAN, the driver works very bad and often make > application routines be held down soon after powered on. What I can > make sure is that the driver is still receiving the packets when I can > not "ping" it, but the packets it received are almost all the > broadcasting ones, and it seldom send out a packet. > Is there any one had met this situation? Please give me a help and > tell me what's the matter with my driver. > Thanks and with Best Regards, > Wen Xiaohui --------------------------- Newsgroups: comp.os.vxworks Subject: SCSI hangs when ethernet is active (PCI shared interrupt). Date: 26 Dec 2001 04:18:21 -0800 From: champu_n@yahoo.com (champak) Organization: http://groups.google.com/ Message-ID: Hi all, I am extremely sorry for a long post. But just cant help it!! I am presently working on an Intel Pentium board with a SYM53C895a , 4 Intel 82559 ethernets and 2 Intel 82559ER ethernets. For SYM53C895a device I use VxWorks ncr810 driver(ncr810Lib.c) , which is supposed to be compatible and it is. All the drivers for the respective devices work perfectly well. Now the problem is as follows. One of the ethernets share a PCI interrupt with SCSI. The SCSI works perfectly well for a continuous long time alone. The ethernet (which shares interrupt) also works perfectly well for a continuous long time alone. But when I make the SCSI to work , work in the sense some continuous traffic between SCSI HDD and RAM, and allow the ethernet also to work , then my SCSI system hangs very randomly. Even putting a ethenet cable, can cause this problem(probably may be the ethenet is raising its interrupt). I get different error messages from the SCSI system but many repeat. The diferent error messages are listed below: 1. interrupt: ncr810: illegal instruction (DSP = 0x001083f8). interrupt: ncr810Intr: unrecoverable error - re-starting SIOP. 0x1fee9880 (tScsiTask): ncr810Event: invalid event type (99) 2. interrupt: ncr810RemainderGet: invalid phase. 0x1fee9880 (tScsiTask): ncr810PhaseMismatch: invalid phase (255). 3. interrupt: ncr810RemainderGet: invalid phase. 0x1fee9880 (tScsiTask): ncr810PhaseMismatch: invalid phase (255). 0x1fee9880 (tScsiTask): ncr810InitEvent: invalid event type (54) interrupt: ncr810Intr: no valid IDENTIFY message at (re)select. 0x1fee9880 (tScsiTask): ncr810InitEvent: invalid event type (32) When I switch ON the interrupt debug prints I get this dump , SCSI status1 = 0x00DMA status = 0x84 interrupt: ncr810EventTypeGet:Interrupt status istat = 0x01, SCSI status0 = 0x00 , SCSI status1 = 0x00DMA status = 0x84 interrupt: ncr810EventTypeGet:Interrupt status istat = 0x01, SCSI status0 = 0x00 , SCSI status1 = 0x00DMA status = 0x84 interrupt: ncr810EventTypeGet:Interrupt status istat = 0x01, SCSI status0 = 0x00 , SCSI status1 = 0x00DMA status = 0x84 interrupt: ncr810EventTypeGet:Interrupt status istat = 0x01, SCSI status0 = 0x00 , SCSI status1 = 0x00DMA status = 0x84 interrupt: ncr810EventTypeGet:Interrupt status istat = 0x01, SCSI status0 = 0x00 , SCSI status1 = 0x00DMA status = 0x84 0x1fee9880 (tScsiTask): ncr810InitEvent: invalid event type (99) copy: error reading file after copying 6279168 bytes. Error while copying to RAM value = 27 = 0x1b - -> At first I thought the problem may be with nested interrupts. So I locked the interrupts in my SCSI handler as well as ethernet handler using intLock() and intUnLock(). I have set the lock level to the highest level that is 1. But the problem exists. I then tried to mask the interrupt in the ethernet device.(1 register allows me to do the same), but in vain. Then I tried doing the same with SCSI but then the SCSI system does not work at all.Masking interrupts in the device creates a problem. Any way, for SCSI it was not required, bcos the above dump shows that only DMA core is causing the interrupt and the DIP interrupt in DMA core can't be masked( the manual says). I also checked the stack size of the tasks tNetTask and tScsiTask, since in Pentium architechure the interrupt uses task's stack, but again in vain. I also tried to insert a small delay ( 50 micro seconds) in both the SCSI and ethernet ISRs.But no luck. Any VxWorks guru, kindly show me some light. I am just desperate. regards Champak --------------------------- Newsgroups: comp.os.vxworks Subject: Re: SCSI test task hung in a 8260 board.Why? Date: 26 Dec 2001 08:10:07 -0800 From: champu_n@yahoo.com (champak) Organization: http://groups.google.com/ Message-ID: References: Hi, I am facing a very similar problem and still not solved. You could see my post(posted on december 26 2001) Could you please check whether is there any PCI device on the board which is sharing the interrupt with our great SCSI. If so then that may be the cause.My post would be enough to explain you i believe. God is present. regards errant@263.net (Jerry Wong) wrote in message news:... > Hi, > I wrote a scsi test task which reads a fix amout(16,64,128 sectors > etc.) sectors per time from scsi disk from the head to tail to test > the BSP and boards. But the scsi task would hang at reading > after several thousends times.The scsi task was pended at the time, > while SCSI Interrupt Status Registers(SIST0,SIST1) show no error > occur. > I trace the scsi task into the driver and found each time the abnormal > situation occurs in ISR which in our driver(ncr810lib.o) is > ncr810Intr().The ISR invokes a function named ncr810EventTypeGet() > which reads the ISTAT register to check if DIP or SIP was set and do > something accordingly.Sometime,ncr810EventTypeGet(),strangely,read > both the two bit are zero.The ncr810EventTypeGet() returned error, > then did the ncr810Intr().Does this mean there is a scsi interupt > occurring,but was not assert by scripts or scsi controller?I don't > know what cause this and if it is the resean that scsi task hangs.The > trace output is below: > ... > interrupt: ncr810EventTypeGet:Interrupt status istat = 0x01, SCSI > status0 = 0x00 > , SCSI status1 = 0x00DMA status = 0x84 > interrupt: ncr810EventTypeGet:Interrupt status istat = 0x01, SCSI > status0 = 0x00 > , SCSI status1 = 0x00DMA status = 0x84 > interrupt: ncr810EventTypeGet:Interrupt status istat = 0x01, SCSI > status0 = 0x00 > , SCSI status1 = 0x00DMA status = 0x84 > interrupt: ncr810EventTypeGet:Interrupt status istat = 0x01, SCSI > status0 = 0x00 > /* above is normal */ > , SCSI status1 = 0x00DMA status = 0x84 > interrupt: ncr810EventTypeGet:Interrupt status istat = 0x89, SCSI > status0 = 0x00 > /* abnormal occur */ > , SCSI status1 = 0x00DMA status = 0x90 > interrupt: ncr810EventTypeGet: script aborted. > interrupt: ncr810Intr: DSA = 0xb3df96a8, DSP = 0xb01972f8, type = 1. > interrupt: ncr810Intr: state 2 -> 0 > 0x3df8488 (tScsiTask): ncr810Resume: thread: 0x03df6bd8: state 0 -> 2 > interrupt: ncr810EventTypeGet:Interrupt status istat = 0x09, SCSI > status0 = 0x00 > , SCSI status1 = 0x00DMA status = 0x84 > interrupt: ncr810EventTypeGet: script completed. > interrupt: ncr810Intr: DSA = 0xb3df96a8, DSP = 0xb01973b8, type = 11. > interrupt: ncr810Intr: state 2 -> 0 > 0x3df8488 (tScsiTask): ncr810Resume: thread: 0x03df6bd8: state 0 -> 2 > interrupt: ncr810EventTypeGet:Interrupt status istat = 0x01, SCSI > status0 = 0x00 > , SCSI status1 = 0x00DMA status = 0x84 > interrupt: ncr810EventTypeGet: script completed. > interrupt: ncr810Intr: DSA = 0xb3df96a8, DSP = 0xb01973e0, type = 21. > interrupt: ncr810Intr: state 2 -> 1 > interrupt: ncr810EventTypeGet:Interrupt status istat = 0x09, SCSI > status0 = 0x00 > , SCSI status1 = 0x00DMA status = 0x84 > interrupt: ncr810EventTypeGet: script completed. > interrupt: ncr810Intr: DSA = 0xb3df965c, DSP = 0xb0197118, type = 22. > interrupt: ncr810Intr: state 1 -> 0 > 0x3df8488 (tScsiTask): ncr810Resume: thread: 0x03df86c8: state 0 -> 2 > interrupt: ncr810EventTypeGet:Interrupt status istat = 0x09, SCSI > status0 = 0x00 > , SCSI status1 = 0x00DMA status = 0x84 > interrupt: ncr810EventTypeGet: script completed. > interrupt: ncr810Intr: DSA = 0xb3df965c, DSP = 0xb0197420, type = 11. > interrupt: ncr810Intr: state 2 -> 0 > 0x3df8488 (tScsiTask): ncr810Resume: thread: 0x03df6bd8: state 0 -> 2 > /* the last trace output */ > > > When the scsi task hung,print the task info: > -> i > NAME ENTRY TID PRI STATUS PC SP > ERRNO DELAY > ---------- ------------ -------- --- ---------- -------- -------- > ------- ----- > tExcTask excTask 3dfeac8 0 PEND 18a738 3dfe9e8 > 0 0 > tLogTask logTask 3dfc140 0 PEND 18a738 3dfc068 > 0 0 > tWdbTask 0x14a690 3d97ca0 3 READY 154ff0 3d97b28 > 0 0 > tScsiTask scsiMgr 3df8488 5 PEND 153570 3df83c8 > 0 0 > tNetTask netTask 3db9498 50 READY 155490 3db93c8 > 0 0 > scsitask testscsi 3d938f8 80 PEND 18a738 3d8b690 > 0 0 > value = 0 = 0x0 > > -> tt scsitask > 1726b8 vxTaskEntry +5c : testsio ([]) > 105364 testscsi +98 : scsiRdSecs (3df6060) > 15da40 scsiRdSecs +6c : 18360c ([]) > 18367c scsiDirectLibTblInit+950: 183708 ([]) > 183870 scsiDirectLibTblInit+b44: scsiTransact (3df7408) > 15d680 scsiTransact +5c : 162758 (3df7408) > 1627c0 scsiThreadInit +470: 162558 (3df7408) > 162620 scsiThreadInit +2d0: 16244c (3df71d8) > 162498 scsiThreadInit +148: scsiMgrRequestExecute ([]) > 185fdc scsiMgrRequestExecute+b0 : msgQReceive ([]) > 152318 msgQReceive +280: qJobGet ([]) > value = 0 = 0x0 > > my boards arch: > CPU: XPC8260 > SCSI: 53C895A > OS; vxWorks5.4 > driver:ncr810lib.o --------------------------- Newsgroups: comp.os.vxworks Subject: Re: SCSI hangs when ethernet is active (PCI shared interrupt). Date: Wed, 26 Dec 2001 18:59:02 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: I would be surprised if the shared interrupt scheme would work at all, because there is little chanse enyone wt WindRiver ever intended these drivers to share interrupts or bothered to test the scenario. A shared interrupt requires a "pseudo ISR" to call each of the sharing drivers individual ISR every time an interrupt occurs, and the driver's ISR need to be able to harmlessly ignore if the interrupt was due to the other device. This also has serious performance implications, obviously. If you suffer from insufficient IRQs, you'd be beeter switching to APIC mode (more IRQs) and forgetting about trying shared interrupts as a bad dream. - - Leonid "champak" wrote in message news:f064a7d8.0112260418.284d3ad2@posting.google.com... > Hi all, > I am extremely sorry for a long post. But just cant help it!! > I am presently working on an Intel Pentium board with a SYM53C895a > , 4 Intel 82559 ethernets and 2 Intel 82559ER ethernets. > For SYM53C895a device I use VxWorks ncr810 driver(ncr810Lib.c) , which > is supposed to be compatible and it is. All the drivers for the > respective devices work perfectly well. Now the problem is as follows. > > One of the ethernets share a PCI interrupt with SCSI. The SCSI works > perfectly well for a continuous long time alone. The ethernet (which > shares interrupt) also works perfectly well for a continuous long time > alone. But when I make the SCSI to work , work in the sense some > continuous traffic between SCSI HDD and RAM, and allow the ethernet > also to work , then my SCSI system hangs very randomly. Even putting a > ethenet cable, can cause this problem(probably may be the ethenet is > raising its interrupt). I get different error messages from the SCSI > system but many repeat. The diferent error messages are listed below: > > 1. interrupt: ncr810: illegal instruction (DSP = 0x001083f8). > interrupt: ncr810Intr: unrecoverable error - re-starting SIOP. > 0x1fee9880 (tScsiTask): ncr810Event: invalid event type (99) > > 2. interrupt: ncr810RemainderGet: invalid phase. > 0x1fee9880 (tScsiTask): ncr810PhaseMismatch: invalid phase (255). > > 3. interrupt: ncr810RemainderGet: invalid phase. > 0x1fee9880 (tScsiTask): ncr810PhaseMismatch: invalid phase (255). > 0x1fee9880 (tScsiTask): ncr810InitEvent: invalid event type (54) > interrupt: ncr810Intr: no valid IDENTIFY message at (re)select. > 0x1fee9880 (tScsiTask): ncr810InitEvent: invalid event type (32) > > When I switch ON the interrupt debug prints I get this dump > > , SCSI status1 = 0x00DMA status = 0x84 > interrupt: ncr810EventTypeGet:Interrupt status istat = 0x01, SCSI > status0 = 0x00 > , SCSI status1 = 0x00DMA status = 0x84 > interrupt: ncr810EventTypeGet:Interrupt status istat = 0x01, SCSI > status0 = 0x00 > , SCSI status1 = 0x00DMA status = 0x84 > interrupt: ncr810EventTypeGet:Interrupt status istat = 0x01, SCSI > status0 = 0x00 > , SCSI status1 = 0x00DMA status = 0x84 > interrupt: ncr810EventTypeGet:Interrupt status istat = 0x01, SCSI > status0 = 0x00 > , SCSI status1 = 0x00DMA status = 0x84 > interrupt: ncr810EventTypeGet:Interrupt status istat = 0x01, SCSI > status0 = 0x00 > , SCSI status1 = 0x00DMA status = 0x84 > 0x1fee9880 (tScsiTask): ncr810InitEvent: invalid event type (99) > copy: error reading file after copying 6279168 bytes. > Error while copying to RAM > value = 27 = 0x1b > -> > > At first I thought the problem may be with nested interrupts. So I > locked the interrupts in my SCSI handler as well as ethernet handler > using intLock() and intUnLock(). I have set the lock level to the > highest level that is 1. But the problem exists. I then tried to mask > the interrupt in the ethernet device.(1 register allows > me to do the same), but in vain. Then I tried doing the same with SCSI > but then the SCSI system does not work at all.Masking interrupts in > the device creates a problem. Any way, for SCSI it was not required, > bcos the above dump shows that > only DMA core is causing the interrupt and the DIP interrupt in DMA > core can't be masked( the manual says). I also checked the stack size > of the tasks tNetTask and tScsiTask, since in Pentium architechure > the interrupt uses task's stack, but again in vain. > I also tried to insert a small delay ( 50 micro seconds) in both the > SCSI and ethernet ISRs.But no luck. > > Any VxWorks guru, kindly show me some light. I am just desperate. > > regards > Champak --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to verify Cache mode at run-time Date: Wed, 26 Dec 2001 17:25:48 GMT From: Dan Gold Organization: http://www.newsranger.com Message-ID: References: <2ede7703.0112180543.2773d474@posting.google.com> <8PbU7.15414$AI.8759814@typhoon.san.rr.com> In article , Raymond Yeung says... > >Thanks for the pointer Dan; I'd look into them in details. > >I'm using MMU_BASIC. The reason that I'm interested in verifying >the cache mode is that, while the Tornado project setup >suggests that it's write-through, I'm seeing something odd The easiest way is to look at configAll.h, config.h, and sysLib.c. In particular at sysBatDesc[] and sysPhysMemDesc[]. >with the networking stack. In particular, I've my Ethernet >driver configured to use non-cacheable data buffers. Hence >the flush/invalidate functions are NULL. When I do that, some >functions such as script download (from network), and "ls" >no longer works (however, ping still works). However, if I >modify the driver to use the "real" flush/invalidate functions >(as if the memory is cacheable), the problem goes away. You need to look at the source to your Ethernet driver. If it's been configured to use non-cacheable buffers, chances are it's using cacheDmaMalloc() to allocate memory. If the memory allocated is covered by PTEs (in sysPhysMemDesc), you're fine. If it's covered by a DBAT, you're hosed. It won't work. Now, even if it allocated non-cacheable memory, that's only for receive buffers. Transmit buffers come from the network stack (netStackDataPoolShow() etc...). So in order to really work, in the driver's transmit routine, it would have to copy from the network stack (an Mblk chain) to the non-cacheable buffer. Otherwise you would still need the flush (for the Tx side), but may not need the invalidate (if the buffers were allocated with cacheDmaMalloc()). >I believe the networking stack has its own (cacheable) pool of >memory. However, if it were write-through, I shouldn't have >the problem I described. Yes the network stack does have it's own pool of memory, and by default it's allocated from "normal" memory, thus making it cacheable. You can use netStackDataPoolShow() to find all the information about the system's data buffers. hth, Dan Gold gold@ensemble.com >Raymond > >"dgold1" wrote in message news:<8PbU7.15414$AI.8759814@typhoon.san.rr.com>... >> Raymond Yeung wrote in message ... >> >Hi Kannan, >> > >> >I try looking up DC_CST and DFWT, both in MPC8260 manual >> >and in 603e manual. I also search my BSP, but only find >> >it in 860 specified code. Otherwise I couldn't find them. >> >> >> You won't find them. They are 860 specific registers. >> >> >Another thing is that, my project is not configured to use >> >full MMU, but some "partial" MMU setup. Would the system >> >default to copy-back mode? >> >> If #INCLUDE_MMU_FULL is defined, it expects you to have the (add on, extra >> $$$) pkg VxVMI installed. Otherwise, INCLUDE_MMU_BASIC will be defined. >> >> The cache mode is set with a #define in configAll.h, and can be overwritten >> in your bsp specific config.h. >> >> The cache mode you choose only affects those memory spaces covered in >> sysPhysMemDesc, i.e. those using page table entries. To see the "real, >> run-time" mode, you need to get the PTE for your desired memory address and >> look at the WIMG bits. IIRC, there's a call or 2 you can make vmXXXGet that >> may get a PTE for you, but I can't remember for sure. >> >> Since you're on a PPC that also supports [ID]BATS, the only way to for sure >> know is to read the BAT register and examine the WIMG bits there. You can >> access them with the usual mfspr code, each BAT has a unique SPR. Of course >> if you've set the appropriate values in the sysBatDesc[] array, you could >> always examine the entries in the array, and assume WRS set the BATS >> correctly (They do). >> >> You could always put a global variable in the code based upon the #define in >> configAll.h, i.e. >> #if (dcache_mode == COPYBACK) >> int dcacheMode=1; >> #elif (dcache_mode == WRITETHROUGH) >> int dcacheMode=2; >> #endif >> >> Don't forget vxHid0Get() to figure out if the caches are even enabled... >> >> hth, >> Dan Gold gold@ensemble.com >> Dan Gold gold@ensemble.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: newbe:how to write a serial driver???? Date: 26 Dec 2001 09:26:03 -0800 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0112260926.3063b217@posting.google.com> References: Hello, You can find a lot of examples under target/src/drv/sio if that's any help. Probably one of them is close to the one you need already, if not an exact match. HTH, John... Anthony Rouaux wrote in message news:... > Hi all, > i am looking for documentation about serial device driver and how to write > them. Does anyone know where i can find some? > Thanks a lot. > Bye > > Anthony --------------------------- Newsgroups: comp.os.vxworks Subject: Re: SCSI hangs when ethernet is active (PCI shared interrupt). Date: 26 Dec 2001 09:28:51 -0800 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0112260928.571e13ff@posting.google.com> References: Hello, Check in the BSP that both devices are connecting their interrupts using pciIntConnect() and not directly with intConnect(). Network drivers normally have a way of setting the interrupt connect routine that they use from the BSP to cope with this; I don't know about the SCSI driver, but I would suspect it has the same. HTH, John... champu_n@yahoo.com (champak) wrote in message news:... > Hi all, > I am extremely sorry for a long post. But just cant help it!! > I am presently working on an Intel Pentium board with a SYM53C895a > , 4 Intel 82559 ethernets and 2 Intel 82559ER ethernets. > For SYM53C895a device I use VxWorks ncr810 driver(ncr810Lib.c) , which > is supposed to be compatible and it is. All the drivers for the > respective devices work perfectly well. Now the problem is as follows. > > One of the ethernets share a PCI interrupt with SCSI. The SCSI works > perfectly well for a continuous long time alone. The ethernet (which > shares interrupt) also works perfectly well for a continuous long time > alone. But when I make the SCSI to work , work in the sense some > continuous traffic between SCSI HDD and RAM, and allow the ethernet > also to work , then my SCSI system hangs very randomly. Even putting a > ethenet cable, can cause this problem(probably may be the ethenet is > raising its interrupt). I get different error messages from the SCSI > system but many repeat. The diferent error messages are listed below: > > 1. interrupt: ncr810: illegal instruction (DSP = 0x001083f8). > interrupt: ncr810Intr: unrecoverable error - re-starting SIOP. > 0x1fee9880 (tScsiTask): ncr810Event: invalid event type (99) > > 2. interrupt: ncr810RemainderGet: invalid phase. > 0x1fee9880 (tScsiTask): ncr810PhaseMismatch: invalid phase (255). > > 3. interrupt: ncr810RemainderGet: invalid phase. > 0x1fee9880 (tScsiTask): ncr810PhaseMismatch: invalid phase (255). > 0x1fee9880 (tScsiTask): ncr810InitEvent: invalid event type (54) > interrupt: ncr810Intr: no valid IDENTIFY message at (re)select. > 0x1fee9880 (tScsiTask): ncr810InitEvent: invalid event type (32) > > When I switch ON the interrupt debug prints I get this dump > > , SCSI status1 = 0x00DMA status = 0x84 > interrupt: ncr810EventTypeGet:Interrupt status istat = 0x01, SCSI > status0 = 0x00 > , SCSI status1 = 0x00DMA status = 0x84 > interrupt: ncr810EventTypeGet:Interrupt status istat = 0x01, SCSI > status0 = 0x00 > , SCSI status1 = 0x00DMA status = 0x84 > interrupt: ncr810EventTypeGet:Interrupt status istat = 0x01, SCSI > status0 = 0x00 > , SCSI status1 = 0x00DMA status = 0x84 > interrupt: ncr810EventTypeGet:Interrupt status istat = 0x01, SCSI > status0 = 0x00 > , SCSI status1 = 0x00DMA status = 0x84 > interrupt: ncr810EventTypeGet:Interrupt status istat = 0x01, SCSI > status0 = 0x00 > , SCSI status1 = 0x00DMA status = 0x84 > 0x1fee9880 (tScsiTask): ncr810InitEvent: invalid event type (99) > copy: error reading file after copying 6279168 bytes. > Error while copying to RAM > value = 27 = 0x1b > -> > > At first I thought the problem may be with nested interrupts. So I > locked the interrupts in my SCSI handler as well as ethernet handler > using intLock() and intUnLock(). I have set the lock level to the > highest level that is 1. But the problem exists. I then tried to mask > the interrupt in the ethernet device.(1 register allows > me to do the same), but in vain. Then I tried doing the same with SCSI > but then the SCSI system does not work at all.Masking interrupts in > the device creates a problem. Any way, for SCSI it was not required, > bcos the above dump shows that > only DMA core is causing the interrupt and the DIP interrupt in DMA > core can't be masked( the manual says). I also checked the stack size > of the tasks tNetTask and tScsiTask, since in Pentium architechure > the interrupt uses task's stack, but again in vain. > I also tried to insert a small delay ( 50 micro seconds) in both the > SCSI and ethernet ISRs.But no luck. > > Any VxWorks guru, kindly show me some light. I am just desperate. > > regards > Champak --------------------------- Newsgroups: comp.os.vxworks Subject: Re: malloc in interruption Date: Wed, 26 Dec 2001 10:24:52 -0800 From: "Michael R. Kesti" Organization: MK Associates Message-ID: <3C2A15F4.353B0F78@gv.net> References: <3C251DFF.E576635F@home.com> <7pva2uo42dcrgnm2nrab9u7svf9bqttjai@4ax.com> <3C262BE6.B98ED880@home.com> <488e459a.0112231304.48fbd31b@posting.google.com> <3C267804.9BDF6C9D@gv.net> <3C2770B8.B9C@despam.autobahn.mb.ca> <3C293EEE.707CE5CC@home.com> Reply-To: mkesti@gv.net Gwen & Emory Stagmer wrote: >I agree with all of this. Our use of C++ is generally very limited. >This is for spacecraft, so we're generally VERY conservative. I'm only >basically employing C++ for an application framework, and even then it's use >is limited. We mostly use struct's and unions (not classes) and NO runtime >allocation (start-time is OK as long as it's allocated once and held). It would seem, then, that you're not really writing C++ applications, but are instead writing C applications that you translate with a C++ compiler. I think that this is fine technique, as you enjoy the benefits of C++'s more strict type checking and such, without the drawbacks of a language that shoots you in the foot while trying to protect you from yourself. >I've been doing embedded systems 20 years, so I'm no newbie to this, but still >I had hoped it would be a little plainer that STL in an ISR is a NO NO! It can't be plain if you don't first understand what's really going on. When I took DevelopMentor's "Intensive C++ for C Programers" course, the first topics concerned the advantages of C++ over C. I realized that I had learned to avoid the problems that C++ was reported to solve by using good structured programming techniques. The best justification that the instructor could come up with was that C++ would help when I was ready to hand my designs off to junior programmers who hadn't yet learned the discipline required to do structured programming well. >You CAN teach an old dog new tricks! It just takes more milk-bones!! ;^) I'd rather that the new dogs first learn the old tricks! - -- ======================================================================== 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: Where is PTY Driver Date: Wed, 26 Dec 2001 11:26:18 -0800 From: "Dan Pinson" Message-ID: References: <9vql4m$h1e0h$1@ID-48686.news.dfncis.de> <488e459a.0112191927.260251f3@posting.google.com> <488e459a.0112212336.1ac525ca@posting.google.com> That's a relief to know that that the code is there - I just need to figure out how to configure it in. Thanks. ..Dan "John" wrote in message news:488e459a.0112212336.1ac525ca@posting.google.com... > Hello, > > "Dan Pinson" wrote in message news:... > > No dice. I'm using Tornado 2 (VxWorks 5.4). I can't find the string pty > > anywhere in the whole source directory! > > Won't be in the source directory - you don't get the source. All the > config info for the binary components is under target/config. > > HTH,, > > John... > > > ..Dan > > > > > > "John" wrote in message > > news:488e459a.0112191927.260251f3@posting.google.com... > > > Hello, > > > > > > In AE it is INCLUDE_OS_PTYDRV that you need, probably the same, or at > > > least similar - search for it in config/all/configAll.h. > > > > > > HTH, > > > > > > John... > > > > > > "Dan Pinson" wrote in message > > news:<9vql4m$h1e0h$1@ID-48686.news.dfncis.de>... > > > > The PTY driver does not seem to be in my VxWorks image (lkup "pty" finds > > no > > > > symbols). Is there a configuration option I need to include? > > > > > > > > Thanks, > > > > Dan --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Where is PTY Driver Date: Wed, 26 Dec 2001 11:55:20 -0800 From: "Dan Pinson" Message-ID: References: <9vql4m$h1e0h$1@ID-48686.news.dfncis.de> <488e459a.0112191927.260251f3@posting.google.com> <488e459a.0112212336.1ac525ca@posting.google.com> Mystery solved. The driver is included automatically if you reference it. I was trying to use it from the shell. ..Dan "Dan Pinson" wrote in message news:a0d88v$k3r0j$1@ID-48686.news.dfncis.de... > That's a relief to know that that the code is there - I just need to figure > out how to configure it in. Thanks. > > ..Dan > > "John" wrote in message > news:488e459a.0112212336.1ac525ca@posting.google.com... > > Hello, > > > > "Dan Pinson" wrote in message > news:... > > > No dice. I'm using Tornado 2 (VxWorks 5.4). I can't find the string pty > > > anywhere in the whole source directory! > > > > Won't be in the source directory - you don't get the source. All the > > config info for the binary components is under target/config. > > > > HTH,, > > > > John... > > > > > ..Dan > > > > > > > > > "John" wrote in message > > > news:488e459a.0112191927.260251f3@posting.google.com... > > > > Hello, > > > > > > > > In AE it is INCLUDE_OS_PTYDRV that you need, probably the same, or at > > > > least similar - search for it in config/all/configAll.h. > > > > > > > > HTH, > > > > > > > > John... > > > > > > > > "Dan Pinson" wrote in message > > > news:<9vql4m$h1e0h$1@ID-48686.news.dfncis.de>... > > > > > The PTY driver does not seem to be in my VxWorks image (lkup "pty" > finds > > > no > > > > > symbols). Is there a configuration option I need to include? > > > > > > > > > > Thanks, > > > > > Dan > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: [Q] how to get string with scanf? Date: 26 Dec 2001 12:28:11 -0800 From: les_news@att.net (Les Hammer) Organization: http://groups.google.com/ Message-ID: <9bbf9668.0112261228.f63c7c3@posting.google.com> References: <6fb8119a.0112211743.60a8781c@posting.google.com> <9bbf9668.0112222138.2e86e5c0@posting.google.com> <6fb8119a.0112231645.3c55fefc@posting.google.com> belliny@soback.kornet.net (belliny, Mint~*) wrote in message > Yes, I'm sure that the maxi character length will not exceed > 50[include ternination character]. as you could see, have no problem > to store the value to variable[I can printf it or operate something]. > the general protection fault always occured at the end of program. Being able to do a printf or another operation is not necessarily proof that the string length has not been exceeded (unless that operation was strlen() and a number less than 50 was returned). I don't have complete access to your system, so I can not make any further diagnostic on it. But 9 times out of 10, when I've seen mysterious GPFs before, they were due to string or array bounds violations. You stated that your problem appeared when you began using a character array. That makes me suspect an array bounds violation. And yes, depending upon what gets hit by the access violation it may wait until the end of the program before you get any notification. Do a strlen() after the scanf to verify that you haven't entered too long a string. Les --------------------------- Newsgroups: comp.os.vxworks Subject: Re: how to use network as console instead of serial? Date: 26 Dec 2001 12:31:18 -0800 From: les_news@att.net (Les Hammer) Organization: http://groups.google.com/ Message-ID: <9bbf9668.0112261231.16449881@posting.google.com> References: jlddotcn@hotmail.com (juliedan) wrote in message news:... > I am doing a work which the serial is used for communication. > So I must use network as console for outputing debug message. > How can I do this? Could someone help me? Activate the standard telnet server and log in. Les --------------------------- Newsgroups: comp.os.vxworks Subject: how to use TFTP in vxsim on WINNT host Date: 26 Dec 2001 21:27:58 -0800 From: sharvan_fr@yahoo.com (Sarvanan) Organization: http://groups.google.com/ Message-ID: Hi all, I am using Tornado on winNT host machine. I want to try TFTP on vxsim simulator running on the host machine. Let me know how can i achieve this . Give me the procedure in an experimental way. thanks in advance ... regards, sharvan s. --------------------------- Newsgroups: comp.os.vxworks Subject: [Q] Is there any problem with mRouteAdd() function? Date: 26 Dec 2001 21:55:49 -0800 From: pjuha@orgio.net (Juha Park) Organization: http://groups.google.com/ Message-ID: <2d8b36d.0112262155.7e8fafea@posting.google.com> Hello VxWorks guru, Could you anyone tell me how to resolve this problem? I Configured point to point connection with two routers.(Let's say router- 1 and router-2 from this moment) Routers connected each other through WAN interface with the same network address. (For example, WAN address of router-1:200.1.1.1 255.255.255.0, WAN address of router-2:200.1.1.2 255.255.255.0). And each routers connected PCs through the ethernet interface. (For example, IP address of PC-1:200.2.1.5 255.255.255.0, ethernet address of router-1:200.2.1.1 255.255.255.0, ethernet address of router-2:200.3.1.1 255.255.255.0 IP address of PC-2:200.3.1.5 255.255.255.0) Using mRouteAdd() function, I registered router-1's static routing entry of an ethernet network address on router-2, and vice versa to communicate two PCs each other. So, two PCs could reach each other. Then I changed address of WAN interface of one router to different network address. (For example, 200.1.1.1 255.255.255.0 -->200.1.2.1 255.255.255.0) So between two routers connection disconnected after this because they have different network address space. But, between two PCs still got the connection each other. This was tested with PING. At this moment, I have BIG wonder what's going on. Logically, two routers disconnected each other, but how could still two PCs reach each other? Is there any facts which I did not get? Or is there any problem with mRouteAdd() function? Anyway, any adivice welcome. Thanks in advance. from Juha --- Configuration --- +--------------+ | PC-1 | | | | 200.2.1.5/24 | +--------------+ | | ethernet | +-----------------------+ | Ethernet 200.2.1.1/24 | | | mRouteAdd("200.3.1.0", | Router-1 | "200.1.1.2", 0xffffff00, 0, 0) | | | WAN 200.1.1.1/24 | +-----------------------+ | | WAN (point to point) | +-----------------------+ | WAN 200.1.1.2/24 | | | | Router-2 | mRouteAdd("200.2.1.0", | | "200.1.1.1", 0xffffff00, 0, 0) | Ethernet 200.3.1.1/24 | +-----------------------+ | | | +--------------+ | PC-2 | | | | 200.3.1.5/24 | +--------------+ Change WAN interface address of Router-1 +--------------+ | PC-1 | | | ----------------------+ | 200.2.1.5/24 | | +--------------+ | | | | ethernet | | still commection +-----------------------+ | | Ethernet 200.2.1.1/24 | | | | | | Router-1 | --------+ | | | | | | WAN 200.1.2.1/24 | | | | ------------ | disconnection | +-----------------------+ | | | | | | WAN (point to point) | | | | | +-----------------------+ | | | WAN 200.1.1.2/24 | | | | | | | | Router-2 | --------+ | | | | | Ethernet 200.3.1.1/24 | | +-----------------------+ | | | | | | | +--------------+ | | PC-2 | | | | ----------------------+ | 200.3.1.5/24 | +--------------+ --------------------------- Newsgroups: comp.os.vxworks Subject: bulk transfer problem in vxworks Date: 26 Dec 2001 22:00:02 -0800 From: simplygaurang@yahoo.co.in (gaurang) Organization: http://groups.google.com/ Message-ID: Dear Friends, First of all hello to all. I am new to VxWorks. I am trying to develop device driver for USB modem. I am using Tornado 2.1. I am using bulk transer for data transferring. I am able to send data to the modem and the command I send is executed, but I am not able to receive the response from modem. I am getting transfer IO error. After doing debugging in detail I have come to know that I am getting OHCI_CC_DATA_UNDERRUN error. I do not know why I am not able to get the respnose from modem. Can anyone guide me what I am missing or where I am making mistake? Please guide me. Thanks to all gaurang. --------------------------- Newsgroups: comp.os.vxworks Subject: [Q] Is there any problem with mRouteAdd() function? Date: 26 Dec 2001 22:16:40 -0800 From: pjuha@orgio.net (Juha Park) Organization: http://groups.google.com/ Message-ID: <2d8b36d.0112262216.6b287fef@posting.google.com> Hello VxWorks guru, Could you anyone tell me how to resolve this problem? I Configured point to point connection with two routers.(Let's say router- 1 and router-2 from this moment) Routers connected each other through WAN interface with the same network address. (For example, WAN address of router-1:200.1.1.1 255.255.255.0, WAN address of router-2:200.1.1.2 255.255.255.0). And each routers connected PCs through the ethernet interface. (For example, IP address of PC-1:200.2.1.5 255.255.255.0, ethernet address of router-1:200.2.1.1 255.255.255.0, ethernet address of router-2:200.3.1.1 255.255.255.0 IP address of PC-2:200.3.1.5 255.255.255.0) Using mRouteAdd() function, I registered router-1's static routing entry of an ethernet network address on router-2, and vice versa to communicate two PCs each other. So, two PCs could reach each other. Then I changed address of WAN interface of one router to different network address. (For example, 200.1.1.1 255.255.255.0 -->200.1.2.1 255.255.255.0) So between two routers connection disconnected after this because they have different network address space. But, between two PCs still got the connection each other. This was tested with PING. At this moment, I have BIG wonder what's going on. Logically, two routers disconnected each other, but how could still two PCs reach each other? Is there any facts which I did not get? Or is there any problem with mRouteAdd() function? Anyway, any adivice welcome. Thanks in advance. Juha. --- Configuration --- +--------------+ | PC-1 | | | | 200.2.1.5/24 | +--------------+ | | ethernet | +-----------------------+ | Ethernet 200.2.1.1/24 | | | mRouteAdd("200.3.1.0", | Router-1 | "200.1.1.2", 0xffffff00, 0, 0) | | | WAN 200.1.1.1/24 | +-----------------------+ | | WAN (point to point) | +-----------------------+ | WAN 200.1.1.2/24 | | | | Router-2 | mRouteAdd("200.2.1.0", | | "200.1.1.1", 0xffffff00, 0, 0) | Ethernet 200.3.1.1/24 | +-----------------------+ | | | +--------------+ | PC-2 | | | | 200.3.1.5/24 | +--------------+ Changed WAN interface address of router-1 +--------------+ | PC-1 | | | ----------------------+ | 200.2.1.5/24 | | +--------------+ | | | | ethernet | | still commection +-----------------------+ | | Ethernet 200.2.1.1/24 | | | | | | Router-1 | --------+ | | | | | | WAN 200.1.2.1/24 | | | | ------------ | disconnection | +-----------------------+ | | | | | | WAN (point to point) | | | | | +-----------------------+ | | | WAN 200.1.1.2/24 | | | | | | | | Router-2 | --------+ | | | | | Ethernet 200.3.1.1/24 | | +-----------------------+ | | | | | | | +--------------+ | | PC-2 | | | | ----------------------+ | 200.3.1.5/24 | +--------------+ --------------------------- Newsgroups: comp.os.vxworks Subject: Re: [Q] Is there any problem with mRouteAdd() function? Date: Thu, 27 Dec 2001 09:32:36 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <2d8b36d.0112262155.7e8fafea@posting.google.com> Juha, There is a big difference between "broadcast" and "point-to-point" IP interfaces: The subnet part of the address needs to be shared by all nodes connected on a "broadcast" type network (i.e. Ethernet), but the subnet address or mask have no special meaning in PtP interface, as each interface has the peer address configured, and it never needs to compare. Also, when you add a route for the PtP (WAN) interface, all it does is cause the packet to be output on the WAN link, and once it gets to the other router, it will only examine the destination IP address for further routing. There is plenty of literature on the network and in book stores about the principles of IP routing, you should get and study some of that. - - Leonid "Juha Park" wrote in message news:2d8b36d.0112262155.7e8fafea@posting.google.com... > Hello VxWorks guru, > Could you anyone tell me how to resolve this problem? > > I Configured point to point connection with two routers.(Let's say router- > 1 and router-2 from this moment) > > Routers connected each other through WAN interface with the same network > address. (For example, WAN address of router-1:200.1.1.1 255.255.255.0, > WAN address of router-2:200.1.1.2 255.255.255.0). And each routers > connected PCs through the ethernet interface. (For example, IP address of > PC-1:200.2.1.5 255.255.255.0, ethernet address of router-1:200.2.1.1 > 255.255.255.0, ethernet address of router-2:200.3.1.1 255.255.255.0 IP > address of PC-2:200.3.1.5 255.255.255.0) > > Using mRouteAdd() function, I registered router-1's static routing entry > of an ethernet network address on router-2, and vice versa to communicate > two PCs each other. So, two PCs could reach each other. > > Then I changed address of WAN interface of one router to different > network address. (For example, 200.1.1.1 255.255.255.0 -->200.1.2.1 > 255.255.255.0) So between two routers connection disconnected after this > because they have different network address space. But, between two PCs > still got the connection each other. This was tested with PING. > > At this moment, I have BIG wonder what's going on. Logically, two routers > disconnected each other, but how could still two PCs reach each other? > Is there any facts which I did not get? Or is there any problem with > mRouteAdd() function? > Anyway, any adivice welcome. > > Thanks in advance. > from Juha > > --- Configuration --- > +--------------+ > | PC-1 | > | | > | 200.2.1.5/24 | > +--------------+ > | > | ethernet > | > +-----------------------+ > | Ethernet 200.2.1.1/24 | > | | mRouteAdd("200.3.1.0", > | Router-1 | "200.1.1.2", 0xffffff00, 0, 0) > | | > | WAN 200.1.1.1/24 | > +-----------------------+ > | > | WAN (point to point) > | > +-----------------------+ > | WAN 200.1.1.2/24 | > | | > | Router-2 | mRouteAdd("200.2.1.0", > | | "200.1.1.1", 0xffffff00, 0, 0) > | Ethernet 200.3.1.1/24 | > +-----------------------+ > | > | > | > +--------------+ > | PC-2 | > | | > | 200.3.1.5/24 | > +--------------+ > > Change WAN interface address of Router-1 > > +--------------+ > | PC-1 | > | | ----------------------+ > | 200.2.1.5/24 | | > +--------------+ | > | | > | ethernet | > | still commection > +-----------------------+ | > | Ethernet 200.2.1.1/24 | | > | | | > | Router-1 | --------+ | > | | | | > | WAN 200.1.2.1/24 | | | > | ------------ | disconnection | > +-----------------------+ | | > | | | > | WAN (point to point) | | > | | | > +-----------------------+ | | > | WAN 200.1.1.2/24 | | | > | | | | > | Router-2 | --------+ | > | | | > | Ethernet 200.3.1.1/24 | | > +-----------------------+ | > | | > | | > | | > +--------------+ | > | PC-2 | | > | | ----------------------+ > | 200.3.1.5/24 | > +--------------+ --------------------------- Newsgroups: comp.os.vxworks Subject: Re: network protocol toolkit user guide Date: 27 Dec 2001 09:19:14 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: Anthony Rouaux wrote: > Hi, > does anyone know where i can find the manual "Network protocol toolkit > user's guide"? You can find it on WindSurf: http://www.windriver.com/csdocs/techpubs/tor-2.0/books.html under networking products. 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: Full Memory Protection in embedded C application Date: Thu, 27 Dec 2001 11:22:35 +0100 From: "Mukul Dhankhar" Message-ID: <3c2af51b$0$12219$4d4efb8e@news.be.uu.net> Hi all, I am working on VxWorks AE 1.1. I embed my C applications in the VxWorks floppy boot image( and I do not download it through ethernet using FTP). - -Does anyone know if its possible to include full Virtual Memory rotection( the Protection Domains in Windriver terminology) in my embedded C application? If yes, then what components need to be included in the boot image, or what is the way to do it? By default I have just the flat memory model( like the earlier versions of VxWorks) when my embedded C application is spawned by usrAppInit(). Thanks, Mukul Dhankhar --------------------------- Newsgroups: comp.protocols.snmp,comp.os.psos,comp.os.vxworks Subject: Test function and Multiple Variable Bindings SetRequest Date: Thu, 27 Dec 2001 13:24:59 +0200 From: Alex Vinokur Message-ID: <3C2B050A.91A7D54B@bigfoot.com> Followup-To: comp.protocols.snmp ============================ pRISM+ for pSOSystem/PowerPC pRISM+ 1.2.3 . pSOSystem 2.2.2 Diab Data D-CC Optimizing C compiler SNMP Release 7.4 (Protocol SNMPv1) Epilogue Technology Emissary SNMP MIB Compiler, version 6.3 Host : Windows98 Target : PowerPC (MPC860) MG-SOFT MIB Browser, Professional Edition 7.0.0.3730 ============================ Hi, I have the following problem with Multiple Variable Bindings SetRequest. == Info#1 of 7 - -- ---------- Fragment of mib file : BEGIN ---------- about OBJECT-IDENTITY STATUS current DESCRIPTION "" ::= { top 1 } ftp OBJECT-IDENTITY STATUS current DESCRIPTION "" ::= { top 3 } notes OBJECT-TYPE SYNTAX DisplayString (SIZE (1..255)) MAX-ACCESS read-write STATUS current DESCRIPTION "" ::= { about 11 } filename OBJECT-TYPE SYNTAX DisplayString (SIZE (1..15)) MAX-ACCESS read-write STATUS current DESCRIPTION ": ::= { ftp 3 } - -- ---------- Fragment of mib file : END ------------ == Info#2 of 7 - -- ---------- Fragment of mib.c file : BEGIN ---------- /****************************************************************************** ****************************************************************************** **** This file was automatically generated by Epilogue Technology's **** Emissary SNMP MIB Compiler, version 6.3. **** This file was generated using the -mib.c switch. **** [ -- omitted -- ] ****************************************************************************** ****************************************************************************** */ LEAF_ASYS( notes_leaf, "notes", VT_STRING, RW, SCALAR, about_test_async, about_get_async, about_set_async, about_next_async, (char *) the_about.notes, 0x0000, 0xFF, 0xFF); LEAF_ASYS( filename_leaf, "filename", VT_STRING, RW, SCALAR, ftp_test_async, get_string_async, ftp_set_async, std_next_async, (char *) the_ftp.filename, 0x0000, 0xFF, 0xFF); STATIC MIBARC_T about_arcs[] = { ARC(1, encoderDescription_leaf), ARC(2, encoderVersion_leaf), ARC(3, encoderOnOffState_leaf), ARC(4, userControlAction_leaf), ARC(5, getRequest_Community_leaf), ARC(6, setRequest_Community_leaf), ARC(7, snmp_Trap_Community_leaf), ARC(8, totalStarted_leaf), ARC(9, totalEnded_leaf), ARC(10, encoderUpTime_leaf), ARC(11, notes_leaf), ARC(12, lastSetSessionNo_leaf), ARC(13, totalUpdatesInSession_leaf), ARC(14, updatingInfoTable_node), ARC(15, encoderFunctioningState_leaf), ARC(16, manager_counter_leaf), ARC(17, manager_ip_addressTable_node), END_OF_ARC_LIST }; STATIC MIBARC_T ftp_arcs[] = { ARC(1, username_leaf), ARC(2, password_leaf), ARC(3, filename_leaf), ARC(4, directory_name_leaf), ARC(5, filetype_leaf), ARC(6, direction_leaf), ARC(7, ftp_server_ip_address_leaf), ARC(8, ftp_control_action_leaf), END_OF_ARC_LIST }; - -- ---------- Fragment of mib.c file : END ------------ == Info#3 of 7 Note. stub.c file is used only as prototype while building my own source files (see below) - -- ---------- Fragment of stub.c file : BEGIN ---------- /****************************************************************************** ****************************************************************************** **** This file was automatically generated by Epilogue Technology's **** Emissary SNMP MIB Compiler, version 6.3. **** This file was generated using the -stub switch. [ -- omitted -- ] ****************************************************************************** ****************************************************************************** */ void about_test(OIDC_T lastmatch, int compc, OIDC_T *compl, SNMP_PKT_T *pktp, VB_T *vbp) { VB_T *group_vbp; /* Only scalar variables here, check for .0 */ if (!((compc == 1) && (*compl == 0))) { testproc_error(pktp, vbp, NO_SUCH_NAME); return; } [ -- omitted -- ] } void ftp_test(OIDC_T lastmatch, int compc, OIDC_T *compl, SNMP_PKT_T *pktp, VB_T *vbp) { VB_T *group_vbp; /* Only scalar variables here, check for .0 */ if (!((compc == 1) && (*compl == 0))) { testproc_error(pktp, vbp, NO_SUCH_NAME); return; } [ -- omitted -- ] } - -- ---------- Fragment of stub.c file : END ------------ == Info#4 of 7 - -- ---------- Fragment of SNMP User's Guide : BEGIN ---------- The parameter named last_match is the value of the last component of the object identifier consumed during the tree search. For example, if the target is the object identifier 1.2.3.4.5.6 and a variable 1.2.3.4 exists, the value of last_match is 4. The parameters named compc and compl describe the unused part of the target. compc is the count of components referenced by compl where compl is a pointer to the list of unconsumed components. Again assuming that the target is 1.2.3.4.5.6 and a leaf is encountered at 1.2.3.4, compc would have the value of 2 and compl would point to a list containing two values, 5 and 6. For scalar variables, the three preceding parameters are not very useful. - -- ---------- Fragment of SNMP User's Guide : END ------------ == Info#5 of 7 - -- ---------- Fragment of about.c file : BEGIN ---------- //======================================= void about_test_async ( OIDC_T lastmatch, int compc, OIDC_T *compl, SNMP_PKT_T *pktp, VB_T *vbp) //======================================= { VB_T *group_vbp; //------------------------------------------- // Only scalar variables here, check for .0 //------------------------------------------- printf ("\nabout_test_async : BEGIN\n"); printf ("about_test_async : lastmatch = %d\n", lastmatch); printf ("about_test_async : compc = %d\n", compc); printf ("about_test_async : *compl = %d\n", *compl); if (!((compc == 1) && (*compl == 0))) { printf ("about_test_async : ERROR\n"); testproc_error(pktp, vbp, NO_SUCH_NAME); return; } printf ("about_test_async : OK\n"); [ -- omitted -- ] } // about_test_async - -- ---------- Fragment of about.c file : END ------------ == Info#6 of 7 - -- ---------- Fragment of ftp.c file : BEGIN ---------- //======================================= void ftp_test_async ( OIDC_T lastmatch, int compc, OIDC_T *compl, SNMP_PKT_T *pktp, VB_T *vbp) //======================================= { VB_T *group_vbp; //------------------------------------------- // Only scalar variables here, check for .0 //------------------------------------------- printf ("\nftp_test_async : BEGIN\n"); printf ("ftp_test_async : lastmatch = %d\n", lastmatch); printf ("ftp_test_async : compc = %d\n", compc); printf ("ftp_test_async : *compl = %d\n", *compl); if (!((compc == 1) && (*compl == 0))) { printf ("ftp_test_async : ERROR\n"); testproc_error(pktp, vbp, NO_SUCH_NAME); return; } printf ("ftp_test_async : OK\n"); [ -- omitted -- ] } // ftp_test_async - -- ---------- Fragment of ftp.c file : END ------------ == Info#7 of 7 ## ########## SetRequest via MG-SOFT Browser : BEGIN ########## ===> about.notes <=== 1.1. Ordinary SetRequest for about.notes : (not using Multiple Variable Bindings) -- Here is printed text (from file about.c, see above) -- about_test_async : BEGIN about_test_async : lastmatch = 11 about_test_async : compc = 1 about_test_async : *compl = 0 about_test_async : OK 1.2. SetRequest for about.notes using Multiple Variable Bindings -- Here is printed text (from file about.c, see above) -- about_test_async : BEGIN about_test_async : lastmatch = 11 about_test_async : compc = 1 about_test_async : *compl = 0 about_test_async : OK ===> ftp.filename <=== 2.1. Ordinary SetRequest for ftp.filename : (not using Multiple Variable Bindings) -- Here is printed text (from file ftp.c, see above) -- ftp_test_async : BEGIN ftp_test_async : lastmatch = 3 ftp_test_async : compc = 1 ftp_test_async : *compl = 0 ftp_test_async : OK 2.2. SetRequest for ftp.filename using Multiple Variable Bindings -- Here is printed text (from file ftp.c, see above) -- ftp_test_async : BEGIN ftp_test_async : lastmatch = 3 ftp_test_async : compc = 0 // !!! Here is a reason for ERROR ftp_test_async : *compl = 0 ftp_test_async : ERROR // !!! ERROR ## ########## SetRequest via MG-SOFT Browser : END ############ We can see that * behaviour of about_test_async() is the same in both SetRequest cases : - without Multiple Variable Bindings - with Multiple Variable Bindings * behaviour of ftp_test_async() is different in two SetRequest cases : - without Multiple Variable Bindings - with Multiple Variable Bindings The only difference between about.notes and ftp.filename (I see) is different kind of get-function-async and next-function-async functions (see mib.c file) : * about.notes is using about_get_async as get-function-async about_next_async as next-function-async whereas : * ftp.filename is using get_string_async as get-function-async std_next_async as next-function-async Is there any explanation for described problem? Thanks in advance, =========================== Alex Vinokur mailto:alexvn@bigfoot.com mailto:alexvn@dr.com http://up.to/alexvn http://go.to/alexv_math =========================== --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Fri Dec 28 04:03:16 2001 From: Vxworks Exploder Date: Fri Dec 28 04:03:18 PST 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Fri Dec 28 04:03:11 PST 2001 Subject: Re: malloc in interruption Subject: zinc designer Subject: Re: how to use network as console instead of serial? Subject: Re: SCSI hangs when ethernet is active (PCI shared interrupt). Subject: Re: SCSI hangs when ethernet is active (PCI shared interrupt). Subject: Re: zinc designer Subject: Re: how to use network as console instead of serial? Subject: Re: Full Memory Protection in embedded C application Subject: Re: zinc designer Subject: Re: how to use network as console instead of serial? Subject: Re: SCSI hangs when ethernet is active (PCI shared interrupt). Subject: Symbol table errors compiling an image when using STL Subject: Re: Relocation value does not fit 24 bits Subject: Re: Relocation value does not fit 24 bits Subject: Re: __get_eh_context called from ISR ?? Subject: Re: performance cost of C++ exception handling Subject: Re: compile assembly codes?? "unresolved expression that must be .." Subject: Re: While loop data access exception Subject: GT48300 Device Driver?? or reference?? Subject: etherInputHookAdd ? Subject: How to use the second END device in Socket@ Subject: Re: network protocol toolkit user guide Subject: Re: etherInputHookAdd ? Subject: Re: network protocol toolkit user guide Subject: Re: etherInputHookAdd ? Subject: Re: How to use the second END device in Socket@ Subject: Detecting Memory Leaks ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: malloc in interruption Date: Thu, 27 Dec 2001 13:32:24 GMT From: Gwen & Emory Stagmer Organization: Excite@Home - The Leader in Broadband http://home.com/faster Message-ID: <3C2B2288.18AD9AC8@home.com> References: <3C251DFF.E576635F@home.com> <7pva2uo42dcrgnm2nrab9u7svf9bqttjai@4ax.com> <3C262BE6.B98ED880@home.com> <488e459a.0112231304.48fbd31b@posting.google.com> <3C267804.9BDF6C9D@gv.net> <3C2770B8.B9C@despam.autobahn.mb.ca> <3C293EEE.707CE5CC@home.com> <3C2A15F4.353B0F78@gv.net> "Michael R. Kesti" wrote: > > Gwen & Emory Stagmer wrote: > > >I agree with all of this. Our use of C++ is generally very limited. > >This is for spacecraft, so we're generally VERY conservative. I'm only > >basically employing C++ for an application framework, and even then it's use > >is limited. We mostly use struct's and unions (not classes) and NO runtime > >allocation (start-time is OK as long as it's allocated once and held). > > It would seem, then, that you're not really writing C++ applications, but > are instead writing C applications that you translate with a C++ compiler. > I think that this is fine technique, as you enjoy the benefits of C++'s > more strict type checking and such, without the drawbacks of a language > that shoots you in the foot while trying to protect you from yourself. > > >I've been doing embedded systems 20 years, so I'm no newbie to this, but still > >I had hoped it would be a little plainer that STL in an ISR is a NO NO! > > It can't be plain if you don't first understand what's really going on. > > When I took DevelopMentor's "Intensive C++ for C Programers" course, the > first topics concerned the advantages of C++ over C. I realized that I > had learned to avoid the problems that C++ was reported to solve by using > good structured programming techniques. The best justification that the > instructor could come up with was that C++ would help when I was ready to > hand my designs off to junior programmers who hadn't yet learned the > discipline required to do structured programming well. > > >You CAN teach an old dog new tricks! It just takes more milk-bones!! ;^) > > I'd rather that the new dogs first learn the old tricks! > Michael Kesti | "And like, one and one don't make > | two, one and one make one." > mkesti@gv.net | - The Who, Bargain Amen to THAT!! I find SO many 'fresh grads' who can't write a queue or have never even seen assembly. They can do a Java website, but not make an embedded machine run. Ahh well... The old dog I was talking about was me! ;^) Emory --------------------------- Newsgroups: comp.os.vxworks Subject: zinc designer Date: 27 Dec 2001 05:33:51 -0800 From: yava@rediffmail.com (divya) Organization: http://groups.google.com/ Message-ID: > I have started working in zinc designer.In zinc,there > is an option to use it in embedded systems.But,I find errors > while downloading my file(when zinc is used in tornado > environment with vxWorks).What should I do to remove the > error and work comfortably? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: how to use network as console instead of serial? Date: 27 Dec 2001 05:41:28 -0800 From: jlddotcn@hotmail.com (juliedan) Organization: http://groups.google.com/ Message-ID: References: <9bbf9668.0112261231.16449881@posting.google.com> les_news@att.net (Les Hammer) wrote in message news:<9bbf9668.0112261231.16449881@posting.google.com>... > jlddotcn@hotmail.com (juliedan) wrote in message news:... > > I am doing a work which the serial is used for communication. > > So I must use network as console for outputing debug message. > > How can I do this? Could someone help me? > > Activate the standard telnet server and log in. > > Les How to? Can you tell me in details? Liedan JU --------------------------- Newsgroups: comp.os.vxworks Subject: Re: SCSI hangs when ethernet is active (PCI shared interrupt). Date: 27 Dec 2001 06:18:30 -0800 From: champu_n@yahoo.com (champak) Organization: http://groups.google.com/ Message-ID: References: The pseudo ISR that you are probably talking about is pciIntConnect().After seeing the code of this routine, I understand that, this routine does the job of calling all the ISRs in the list meant for a particular IRQ. And all my ISRs (only 2 i have in the list SCSI and 82559) are descent enough to exit if their device has not shouted an interrupt. champak "Leonid Rosenboim" wrote in message news:... > I would be surprised if the shared interrupt scheme would work at all, > because there is little chanse enyone wt WindRiver ever intended > these drivers to share interrupts or bothered to test the scenario. > A shared interrupt requires a "pseudo ISR" to call each of the > sharing drivers individual ISR every time an interrupt occurs, > and the driver's ISR need to be able to harmlessly ignore if the > interrupt was due to the other device. > This also has serious performance implications, obviously. > > If you suffer from insufficient IRQs, you'd be beeter switching > to APIC mode (more IRQs) and forgetting about trying shared > interrupts as a bad dream. > > - Leonid > > "champak" wrote in message > news:f064a7d8.0112260418.284d3ad2@posting.google.com... > > Hi all, > > I am extremely sorry for a long post. But just cant help it!! > > I am presently working on an Intel Pentium board with a SYM53C895a > > , 4 Intel 82559 ethernets and 2 Intel 82559ER ethernets. > > For SYM53C895a device I use VxWorks ncr810 driver(ncr810Lib.c) , which > > is supposed to be compatible and it is. All the drivers for the > > respective devices work perfectly well. Now the problem is as follows. > > > > One of the ethernets share a PCI interrupt with SCSI. The SCSI works > > perfectly well for a continuous long time alone. The ethernet (which > > shares interrupt) also works perfectly well for a continuous long time > > alone. But when I make the SCSI to work , work in the sense some > > continuous traffic between SCSI HDD and RAM, and allow the ethernet > > also to work , then my SCSI system hangs very randomly. Even putting a > > ethenet cable, can cause this problem(probably may be the ethenet is > > raising its interrupt). I get different error messages from the SCSI > > system but many repeat. The diferent error messages are listed below: > > > > 1. interrupt: ncr810: illegal instruction (DSP = 0x001083f8). > > interrupt: ncr810Intr: unrecoverable error - re-starting SIOP. > > 0x1fee9880 (tScsiTask): ncr810Event: invalid event type (99) > > > > 2. interrupt: ncr810RemainderGet: invalid phase. > > 0x1fee9880 (tScsiTask): ncr810PhaseMismatch: invalid phase (255). > > > > 3. interrupt: ncr810RemainderGet: invalid phase. > > 0x1fee9880 (tScsiTask): ncr810PhaseMismatch: invalid phase (255). > > 0x1fee9880 (tScsiTask): ncr810InitEvent: invalid event type (54) > > interrupt: ncr810Intr: no valid IDENTIFY message at (re)select. > > 0x1fee9880 (tScsiTask): ncr810InitEvent: invalid event type (32) > > > > When I switch ON the interrupt debug prints I get this dump > > > > , SCSI status1 = 0x00DMA status = 0x84 > > interrupt: ncr810EventTypeGet:Interrupt status istat = 0x01, SCSI > > status0 = 0x00 > > , SCSI status1 = 0x00DMA status = 0x84 > > interrupt: ncr810EventTypeGet:Interrupt status istat = 0x01, SCSI > > status0 = 0x00 > > , SCSI status1 = 0x00DMA status = 0x84 > > interrupt: ncr810EventTypeGet:Interrupt status istat = 0x01, SCSI > > status0 = 0x00 > > , SCSI status1 = 0x00DMA status = 0x84 > > interrupt: ncr810EventTypeGet:Interrupt status istat = 0x01, SCSI > > status0 = 0x00 > > , SCSI status1 = 0x00DMA status = 0x84 > > interrupt: ncr810EventTypeGet:Interrupt status istat = 0x01, SCSI > > status0 = 0x00 > > , SCSI status1 = 0x00DMA status = 0x84 > > 0x1fee9880 (tScsiTask): ncr810InitEvent: invalid event type (99) > > copy: error reading file after copying 6279168 bytes. > > Error while copying to RAM > > value = 27 = 0x1b > > -> > > > > At first I thought the problem may be with nested interrupts. So I > > locked the interrupts in my SCSI handler as well as ethernet handler > > using intLock() and intUnLock(). I have set the lock level to the > > highest level that is 1. But the problem exists. I then tried to mask > > the interrupt in the ethernet device.(1 register allows > > me to do the same), but in vain. Then I tried doing the same with SCSI > > but then the SCSI system does not work at all.Masking interrupts in > > the device creates a problem. Any way, for SCSI it was not required, > > bcos the above dump shows that > > only DMA core is causing the interrupt and the DIP interrupt in DMA > > core can't be masked( the manual says). I also checked the stack size > > of the tasks tNetTask and tScsiTask, since in Pentium architechure > > the interrupt uses task's stack, but again in vain. > > I also tried to insert a small delay ( 50 micro seconds) in both the > > SCSI and ethernet ISRs.But no luck. > > > > Any VxWorks guru, kindly show me some light. I am just desperate. > > > > regards > > Champak --------------------------- Newsgroups: comp.os.vxworks Subject: Re: SCSI hangs when ethernet is active (PCI shared interrupt). Date: 27 Dec 2001 06:20:03 -0800 From: champu_n@yahoo.com (champak) Organization: http://groups.google.com/ Message-ID: References: <488e459a.0112260928.571e13ff@posting.google.com> Yes I have used pciIntConnect() to hook the ISRs. If I had not then the system would not have worked even once.Where as my system for few minutes and then hangs randomly Champak john_94501@yahoo.com (John) wrote in message news:<488e459a.0112260928.571e13ff@posting.google.com>... > Hello, > > Check in the BSP that both devices are connecting their interrupts > using pciIntConnect() and not directly with intConnect(). Network > drivers normally have a way of setting the interrupt connect routine > that they use from the BSP to cope with this; I don't know about the > SCSI driver, but I would suspect it has the same. > > HTH, > > John... > > champu_n@yahoo.com (champak) wrote in message news:... > > Hi all, > > I am extremely sorry for a long post. But just cant help it!! > > I am presently working on an Intel Pentium board with a SYM53C895a > > , 4 Intel 82559 ethernets and 2 Intel 82559ER ethernets. > > For SYM53C895a device I use VxWorks ncr810 driver(ncr810Lib.c) , which > > is supposed to be compatible and it is. All the drivers for the > > respective devices work perfectly well. Now the problem is as follows. > > > > One of the ethernets share a PCI interrupt with SCSI. The SCSI works > > perfectly well for a continuous long time alone. The ethernet (which > > shares interrupt) also works perfectly well for a continuous long time > > alone. But when I make the SCSI to work , work in the sense some > > continuous traffic between SCSI HDD and RAM, and allow the ethernet > > also to work , then my SCSI system hangs very randomly. Even putting a > > ethenet cable, can cause this problem(probably may be the ethenet is > > raising its interrupt). I get different error messages from the SCSI > > system but many repeat. The diferent error messages are listed below: > > > > 1. interrupt: ncr810: illegal instruction (DSP = 0x001083f8). > > interrupt: ncr810Intr: unrecoverable error - re-starting SIOP. > > 0x1fee9880 (tScsiTask): ncr810Event: invalid event type (99) > > > > 2. interrupt: ncr810RemainderGet: invalid phase. > > 0x1fee9880 (tScsiTask): ncr810PhaseMismatch: invalid phase (255). > > > > 3. interrupt: ncr810RemainderGet: invalid phase. > > 0x1fee9880 (tScsiTask): ncr810PhaseMismatch: invalid phase (255). > > 0x1fee9880 (tScsiTask): ncr810InitEvent: invalid event type (54) > > interrupt: ncr810Intr: no valid IDENTIFY message at (re)select. > > 0x1fee9880 (tScsiTask): ncr810InitEvent: invalid event type (32) > > > > When I switch ON the interrupt debug prints I get this dump > > > > , SCSI status1 = 0x00DMA status = 0x84 > > interrupt: ncr810EventTypeGet:Interrupt status istat = 0x01, SCSI > > status0 = 0x00 > > , SCSI status1 = 0x00DMA status = 0x84 > > interrupt: ncr810EventTypeGet:Interrupt status istat = 0x01, SCSI > > status0 = 0x00 > > , SCSI status1 = 0x00DMA status = 0x84 > > interrupt: ncr810EventTypeGet:Interrupt status istat = 0x01, SCSI > > status0 = 0x00 > > , SCSI status1 = 0x00DMA status = 0x84 > > interrupt: ncr810EventTypeGet:Interrupt status istat = 0x01, SCSI > > status0 = 0x00 > > , SCSI status1 = 0x00DMA status = 0x84 > > interrupt: ncr810EventTypeGet:Interrupt status istat = 0x01, SCSI > > status0 = 0x00 > > , SCSI status1 = 0x00DMA status = 0x84 > > 0x1fee9880 (tScsiTask): ncr810InitEvent: invalid event type (99) > > copy: error reading file after copying 6279168 bytes. > > Error while copying to RAM > > value = 27 = 0x1b > > -> > > > > At first I thought the problem may be with nested interrupts. So I > > locked the interrupts in my SCSI handler as well as ethernet handler > > using intLock() and intUnLock(). I have set the lock level to the > > highest level that is 1. But the problem exists. I then tried to mask > > the interrupt in the ethernet device.(1 register allows > > me to do the same), but in vain. Then I tried doing the same with SCSI > > but then the SCSI system does not work at all.Masking interrupts in > > the device creates a problem. Any way, for SCSI it was not required, > > bcos the above dump shows that > > only DMA core is causing the interrupt and the DIP interrupt in DMA > > core can't be masked( the manual says). I also checked the stack size > > of the tasks tNetTask and tScsiTask, since in Pentium architechure > > the interrupt uses task's stack, but again in vain. > > I also tried to insert a small delay ( 50 micro seconds) in both the > > SCSI and ethernet ISRs.But no luck. > > > > Any VxWorks guru, kindly show me some light. I am just desperate. > > > > regards > > Champak --------------------------- Newsgroups: comp.os.vxworks Subject: Re: zinc designer Date: Thu, 27 Dec 2001 17:22:28 +0100 From: "Mukul Dhankhar" Message-ID: <3c2b4972$0$12219$4d4efb8e@news.be.uu.net> References: Hi Divya, Could you elaborate more on your problem. - -How is 'zinc' incorporated with Tornado? - -What is the version of VxWorks you are using? - -Do you try to download your files by the default process( ..like FTP)? - -What are the error messages you get? - -If possible could you put forward a screen dump of your boot and file download process Cheers, Mukul "divya" wrote in message news:f7bb25af.0112270533.7decac9@posting.google.com... > > I have started working in zinc designer.In zinc,there > > is an option to use it in embedded systems.But,I find errors > > while downloading my file(when zinc is used in tornado > > environment with vxWorks).What should I do to remove the > > error and work comfortably? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: how to use network as console instead of serial? Date: Thu, 27 Dec 2001 11:08:49 -0600 From: Dinker Charak Organization: Fermi National Accelerator Laboratory Message-ID: <3C2B55A1.7175C944@fnal.gov> References: <9bbf9668.0112261231.16449881@posting.google.com> > > Activate the standard telnet server and log in. > > > > Les > > How to? > Can you tell me in details? > Liedan JU Are you using Tornado? If yes, check in VxWorks Tab, goto Network components -> networking protocols -> netowrk applications and set up telnet ... the do a simple telnet to the target's IP to previous reply, for IO redirection it is done differently ... in tornado for unix goto target-> create and click on target io redirection etc and in windows navigate thru menus and u will find it and in create target you will find the target io redirection stuff ... hth, dinker --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Full Memory Protection in embedded C application Date: 27 Dec 2001 09:41:16 -0800 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0112270941.72f0c266@posting.google.com> References: <3c2af51b$0$12219$4d4efb8e@news.be.uu.net> Hello, No, the bootApp does not support protection; it is intended to simply boot the full kernel, not to contain your application. On non-Pentium architectures, it is possible to build a ROM image that is the full system (there is a tech note about this process: http://www.wrs.com/csdocs/technotes/ide/tae/WTN69/romming.html - you'll need a WindSurf account to get to it of course). Since Pentium systems don't come with ROMs (backwards architecture for embedded use ;-), you'll have to create a boot floppy based on your system project; when the tech note talks about ROMs, just think boot floppy for Pentium and you should be able to do the same. I'm not sure if this works though since booting pentium boards is a little different to other architectures - if not, file a TSR! Also, the very limited size of the floppy disk might make it difficult to fit your application and the kernel on there (depending on the size of the app of course). Do you have a hard disk or a flash disk in your system? Rgrds, John... "Mukul Dhankhar" wrote in message news:<3c2af51b$0$12219$4d4efb8e@news.be.uu.net>... > Hi all, > I am working on VxWorks AE 1.1. I embed my C applications in the VxWorks > floppy boot image( and I do not download it through ethernet using FTP). > -Does anyone know if its possible to include full Virtual Memory > rotection( the Protection Domains in Windriver terminology) in my embedded C > application? If yes, then what components need to be included in the boot > image, or what is the way to do it? > By default I have just the flat memory model( like the earlier versions of > VxWorks) when my embedded C application is spawned by usrAppInit(). > Thanks, > Mukul Dhankhar --------------------------- Newsgroups: comp.os.vxworks Subject: Re: zinc designer Date: 27 Dec 2001 09:45:22 -0800 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0112270945.85943d7@posting.google.com> References: Hello, Firstly, you need to learn how to post questions here... The following questions will help you provide the minimum information we would need to help you out: 1) What version of Zinc are you using? 2) What version of Tornado are you using? 3) What is the error that you are getting? 4) How did you build it and test it? We know about VxWorks and Tornado development, but we are not psychic, and we are not paid to answer your questions, so please try to provide as much information as possible when posting questions. That way, you are much more likely to get a useful answer (or even any answer). Rgrds, John... yava@rediffmail.com (divya) wrote in message news:... > > I have started working in zinc designer.In zinc,there > > is an option to use it in embedded systems.But,I find errors > > while downloading my file(when zinc is used in tornado > > environment with vxWorks).What should I do to remove the > > error and work comfortably? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: how to use network as console instead of serial? Date: 27 Dec 2001 10:04:19 -0800 From: les_news@att.net (Les Hammer) Organization: http://groups.google.com/ Message-ID: <9bbf9668.0112271004.4f4adae4@posting.google.com> References: <9bbf9668.0112261231.16449881@posting.google.com> jlddotcn@hotmail.com (juliedan) wrote in message news:... > > Activate the standard telnet server and log in. > > > > Les > > How to? > Can you tell me in details? I don't have access to my vxWorks system anymore, so I can't give all the details, but: 1) From your DOS prompt, use vxencrypt to encrypt a password. 2) In your usrappinit.c code: loginUserAdd("username","password"); /* where username is replaced by a user name that you want, and password is replaced by the encrypted password from step 1. */ loginUserDelete("target","password"); /* Delete the default user to keep other people out of your system */ telnetInit(1) remShellInit(); /* Start the telnet server on port 123 */ 3) Obtain telnetLib.c and telnetLib.h from your WindRiver or other vxWorks supplier. 4) Here's the part that I'm fuzzy on, without my system for reference. Somewhere in your project, you have a set of attributes that you can select. Two of these are the telnet/networking attribute and the security attribute. Turn both of these on. Les --------------------------- Newsgroups: comp.os.vxworks Subject: Re: SCSI hangs when ethernet is active (PCI shared interrupt). Date: Thu, 27 Dec 2001 23:23:41 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: Well, then, try to eliminate the ISR issue by splitting these two devices into their own IRQs. If you are correct and the ISRs are cool, then the problem will occur with private IRQs as well. (it could be PCI bus contention, but I doubt that your trpt is that high p). "champak" wrote in message news:f064a7d8.0112270618.5af74c00@posting.google.com... > The pseudo ISR that you are probably talking about is > pciIntConnect().After seeing the code of this routine, I understand > that, this routine does the job of calling all the ISRs in the list > meant for a particular IRQ. And all my ISRs (only 2 i have in the list > SCSI and 82559) are descent enough to exit if their device has not > shouted an interrupt. > champak > "Leonid Rosenboim" wrote in message news:... > > I would be surprised if the shared interrupt scheme would work at all, > > because there is little chanse enyone wt WindRiver ever intended > > these drivers to share interrupts or bothered to test the scenario. > > A shared interrupt requires a "pseudo ISR" to call each of the > > sharing drivers individual ISR every time an interrupt occurs, > > and the driver's ISR need to be able to harmlessly ignore if the > > interrupt was due to the other device. > > This also has serious performance implications, obviously. > > > > If you suffer from insufficient IRQs, you'd be beeter switching > > to APIC mode (more IRQs) and forgetting about trying shared > > interrupts as a bad dream. > > > > - Leonid > > > > "champak" wrote in message > > news:f064a7d8.0112260418.284d3ad2@posting.google.com... > > > Hi all, > > > I am extremely sorry for a long post. But just cant help it!! > > > I am presently working on an Intel Pentium board with a SYM53C895a > > > , 4 Intel 82559 ethernets and 2 Intel 82559ER ethernets. > > > For SYM53C895a device I use VxWorks ncr810 driver(ncr810Lib.c) , which > > > is supposed to be compatible and it is. All the drivers for the > > > respective devices work perfectly well. Now the problem is as follows. > > > > > > One of the ethernets share a PCI interrupt with SCSI. The SCSI works > > > perfectly well for a continuous long time alone. The ethernet (which > > > shares interrupt) also works perfectly well for a continuous long time > > > alone. But when I make the SCSI to work , work in the sense some > > > continuous traffic between SCSI HDD and RAM, and allow the ethernet > > > also to work , then my SCSI system hangs very randomly. Even putting a > > > ethenet cable, can cause this problem(probably may be the ethenet is > > > raising its interrupt). I get different error messages from the SCSI > > > system but many repeat. The diferent error messages are listed below: > > > > > > 1. interrupt: ncr810: illegal instruction (DSP = 0x001083f8). > > > interrupt: ncr810Intr: unrecoverable error - re-starting SIOP. > > > 0x1fee9880 (tScsiTask): ncr810Event: invalid event type (99) > > > > > > 2. interrupt: ncr810RemainderGet: invalid phase. > > > 0x1fee9880 (tScsiTask): ncr810PhaseMismatch: invalid phase (255). > > > > > > 3. interrupt: ncr810RemainderGet: invalid phase. > > > 0x1fee9880 (tScsiTask): ncr810PhaseMismatch: invalid phase (255). > > > 0x1fee9880 (tScsiTask): ncr810InitEvent: invalid event type (54) > > > interrupt: ncr810Intr: no valid IDENTIFY message at (re)select. > > > 0x1fee9880 (tScsiTask): ncr810InitEvent: invalid event type (32) > > > > > > When I switch ON the interrupt debug prints I get this dump > > > > > > , SCSI status1 = 0x00DMA status = 0x84 > > > interrupt: ncr810EventTypeGet:Interrupt status istat = 0x01, SCSI > > > status0 = 0x00 > > > , SCSI status1 = 0x00DMA status = 0x84 > > > interrupt: ncr810EventTypeGet:Interrupt status istat = 0x01, SCSI > > > status0 = 0x00 > > > , SCSI status1 = 0x00DMA status = 0x84 > > > interrupt: ncr810EventTypeGet:Interrupt status istat = 0x01, SCSI > > > status0 = 0x00 > > > , SCSI status1 = 0x00DMA status = 0x84 > > > interrupt: ncr810EventTypeGet:Interrupt status istat = 0x01, SCSI > > > status0 = 0x00 > > > , SCSI status1 = 0x00DMA status = 0x84 > > > interrupt: ncr810EventTypeGet:Interrupt status istat = 0x01, SCSI > > > status0 = 0x00 > > > , SCSI status1 = 0x00DMA status = 0x84 > > > 0x1fee9880 (tScsiTask): ncr810InitEvent: invalid event type (99) > > > copy: error reading file after copying 6279168 bytes. > > > Error while copying to RAM > > > value = 27 = 0x1b > > > -> > > > > > > At first I thought the problem may be with nested interrupts. So I > > > locked the interrupts in my SCSI handler as well as ethernet handler > > > using intLock() and intUnLock(). I have set the lock level to the > > > highest level that is 1. But the problem exists. I then tried to mask > > > the interrupt in the ethernet device.(1 register allows > > > me to do the same), but in vain. Then I tried doing the same with SCSI > > > but then the SCSI system does not work at all.Masking interrupts in > > > the device creates a problem. Any way, for SCSI it was not required, > > > bcos the above dump shows that > > > only DMA core is causing the interrupt and the DIP interrupt in DMA > > > core can't be masked( the manual says). I also checked the stack size > > > of the tasks tNetTask and tScsiTask, since in Pentium architechure > > > the interrupt uses task's stack, but again in vain. > > > I also tried to insert a small delay ( 50 micro seconds) in both the > > > SCSI and ethernet ISRs.But no luck. > > > > > > Any VxWorks guru, kindly show me some light. I am just desperate. > > > > > > regards > > > Champak --------------------------- Newsgroups: comp.os.vxworks Subject: Symbol table errors compiling an image when using STL Date: 27 Dec 2001 21:46:33 -0600 From: "Arnon W" Message-ID: <3c29bbec$1@news.barak.net.il> Hi, I'm new to this environment so it might be a minor issue I'm having problems with. (On the other hand - I couldn't find earlyer posts in this specific matter:) I'm trying to add a c++ project to an existing image. (Tornado 2.0, 5.4, ppc604) When I build my project as a downloadable module there are no problems loading and running it using the target server. When I try to integrate it to the image, I get vicious responses at this point: ccppc -c -fdollars-in-identifiers -O3 -mstrict-align -nostdinc -DRW_MULTI_TH READ -D_REENTR ANT -fvolatile -fno-builtin -fno-for-scope -DCPU=PPC604 -DPRJ_BUILD -DHAL_LI B ...paths... symTbl.c -o symTbl.o symTbl.c:2013: syntax error before `.120' symTbl.c:2014: syntax error before `.139' symTbl.c:2015: syntax error before `.148' symTbl.c:8448: `__sn_t24__default_alloc_template2b1i0$__node_allocator_lock' undeclared he re (not in a function) symTbl.c:8448: syntax error before `.120' symTbl.c:8449: initializer element for `standTbl[1997].value' is not constant symTbl.c:8449: warning: initialization from incompatible pointer type symTbl.c:8449: `__sn_t24__default_alloc_template2b1i0$__node_allocator_lock' undeclared he re (not in a function) symTbl.c:8449: syntax error before `.139' symTbl.c:8450: initializer element for `standTbl[1998].name' is not constant symTbl.c:8450: syntax error before string constant When I took out STL related stuff (I was using ) It compiled well. Is there a cure? Thanks Arnon --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Relocation value does not fit 24 bits Date: 28 Dec 2001 04:13:59 GMT From: bcoxsprint12@earthlink.net Organization: NewsOne.Net - Free Usenet News via the Web - http://newsone.net/ Message-ID: References: <3c042747.230902@news1.attglobal.net> In article <3c042747.230902@news1.attglobal.net>, writes: The two shells use different target memory areas to store the downloaded code. The target shell uses an area more than 32 MB from the kernel, which is usually in low memory. You can 1) link your application to the kernel or 2) continue to download it with the WindShell or 3) use the '-mlongcall' option in building your app. The error reports that the 24-bit PC-relative offset in a branch instruction can't hold the distance to the symbol the you're referencing. >Hello, >when entering ld occurs: >Relocation value does not fit 24 bits >ld error: error reading file (errno=0x3d0001) > >This error does not occur if the same module is loaded from the >Tornado WindShell. > >When the kernel was built, the Target shell was configured for module >loading and unloading. > >Why would the Target shell and WindShell behave different when loading >modules? > > Thanks, > ScottL ----- Posted via NewsOne.Net: Free (anonymous) Usenet News via the Web ----- http://newsone.net/ -- Free reading and anonymous posting to 60,000+ groups NewsOne.Net prohibits users from posting spam. If this or other posts made through NewsOne.Net violate posting guidelines, email abuse@newsone.net --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Relocation value does not fit 24 bits Date: 28 Dec 2001 04:22:31 GMT From: bcoxsprint12@earthlink.net Organization: NewsOne.Net - Free Usenet News via the Web - http://newsone.net/ Message-ID: References: <3c042747.230902@news1.attglobal.net> In article <3c042747.230902@news1.attglobal.net>, writes: The two shells use different target memory areas to store the downloaded code. The target shell uses an area more than 32 MB from the kernel, which is usually in low memory. You can 1) link your application to the kernel or 2) continue to download it with the WindShell or 3) use the '-mlongcall' option in building your app. The error reports that the 24-bit PC-relative offset in a branch instruction can't hold the distance to the symbol the you're referencing. >Hello, >when entering ld occurs: >Relocation value does not fit 24 bits >ld error: error reading file (errno=0x3d0001) > >This error does not occur if the same module is loaded from the >Tornado WindShell. > >When the kernel was built, the Target shell was configured for module >loading and unloading. > >Why would the Target shell and WindShell behave different when loading >modules? > > Thanks, > ScottL ----- Posted via NewsOne.Net: Free (anonymous) Usenet News via the Web ----- http://newsone.net/ -- Free reading and anonymous posting to 60,000+ groups NewsOne.Net prohibits users from posting spam. If this or other posts made through NewsOne.Net violate posting guidelines, email abuse@newsone.net --------------------------- Newsgroups: comp.os.vxworks Subject: Re: __get_eh_context called from ISR ?? Date: 28 Dec 2001 04:56:18 GMT From: bcoxsprint12@earthlink.net Organization: NewsOne.Net - Free Usenet News via the Web - http://newsone.net/ Message-ID: References: In article , Sanjay writes: Hi Sanjay, This suggests that your code is invoking an exception-handler runtime routine. It seems that you're running a C++ program as an Interrupt Service Routine which was compiled with excep- tion handling turned on. This is a Really Bad Thing, since the exception-handler setup code calls malloc at interrupt level, will surely corrupt your machine's heap space. You need to recompile all the C++ routines with '-fno-exceptions' or (my preference) don't write interrupt handlers in C++. Bill >Hi all > >We are experiencing a memory leak in our system and memScope >reveals that "__get_eh_context" is being called from the ISR >which in turn calls taskVarAdd() repeatedly. This is the cause >of the memory leak. > >Can anyone shed some light on what "__get_eh_context" is and >why it gets called from the ISR? And why is there this problem >with the memory leak? > >Thanks. > >Sanjay ----- Posted via NewsOne.Net: Free (anonymous) Usenet News via the Web ----- http://newsone.net/ -- Free reading and anonymous posting to 60,000+ groups NewsOne.Net prohibits users from posting spam. If this or other posts made through NewsOne.Net violate posting guidelines, email abuse@newsone.net --------------------------- Newsgroups: comp.os.vxworks Subject: Re: performance cost of C++ exception handling Date: 28 Dec 2001 05:07:10 GMT From: bcoxsprint12@earthlink.net Organization: NewsOne.Net - Free Usenet News via the Web - http://newsone.net/ Message-ID: References: <6hnl5.13002$9T1.101123@typhoon.tampabay.rr.com> In article <6hnl5.13002$9T1.101123@typhoon.tampabay.rr.com>, Jim Reis writes: Hi Jim, I haven't actually timed this, but there at least the following costs: 1) compiling with exception handling enabled will make your code larger. Try it. Compile normally, measure the size with sizeppc, for PowerPC, then recompile with '-fno-exceptions' and remeasure. 2) when you hit your first 'try' block at runtime, you'll call out to a setup routine which will malloc a region record for that block and call taskVarAdd (sp?) to save a pointer to it. Both of these are quite a few instructions to execute. 3) when you hit a 'throw', the runtime code will backtrack through all of the registered region records, calling the destructors for any automatic C++ objects that you've constructed. This has overhead in the runtime system, plus the calls to your own destructors, which no-one else can quantify. >Does anyone have any actual data about the overhead associated with using >C++ exceptions ? From reading the information in the Programmer's Guide, >apparently there are various costs: just enabling them, when the first >exception handling statement is encountered, and when an exception actually >occurs. > >Thanks, > >Jim Reis >jreis@cfl.rr.com > > > > ----- Posted via NewsOne.Net: Free (anonymous) Usenet News via the Web ----- http://newsone.net/ -- Free reading and anonymous posting to 60,000+ groups NewsOne.Net prohibits users from posting spam. If this or other posts made through NewsOne.Net violate posting guidelines, email abuse@newsone.net --------------------------- Newsgroups: comp.os.vxworks Subject: Re: compile assembly codes?? "unresolved expression that must be .." Date: 28 Dec 2001 05:11:16 GMT From: bcoxsprint12@earthlink.net Organization: NewsOne.Net - Free Usenet News via the Web - http://newsone.net/ Message-ID: References: <9t1vih$75m$1@news.asu.edu> In article <9t1vih$75m$1@news.asu.edu>, Okehee Goh writes: Hi Okehee, The line numbers don't match, but the local label "..L4" seems to have too many leading periods. The reference to that label it the branch instruction at the end only has one period. >Hello, > >This is Okehee Goh in ASU. > > >I'd like to compile assembly codes for target machine, MBX860. > >This assembly codes will not be associated with VxWorks OS. > >As a compiler, I tried to use ccppc(gcc) compiler that comes with TornadoII >targeting MBX860. > >But it complained with "Error: unresolved expression that must be resolved". > >I don't know how to solve this problem. > > >The following is the command line to compile assembly code "ae3.s". > >The next one includes error message that ccppc generated. > >The third part is "ae3.s" > >I will appreciate your help. Thank you so much. > >1. > >ccppc -g -mcpu=860 -ansi -nostdinc -fvolatile -fno-builtin -fno-for-scope -P > -x assembler-with-cpp -I. -IC:/Tornado/target/h -DCPU=PPC860 -c ae3.s > >2. > >C:\DOCUME~1\ogoh\LOCALS~1\Temp\cca02324.s: Assembler messages: > >C:\DOCUME~1\ogoh\LOCALS~1\Temp\cca02324.s:9: Error: unresolved expression >that must be resolved > >C:\DOCUME~1\ogoh\LOCALS~1\Temp\cca02324.s:10: Error: unresolved expression >that must be resolved > >.... > >3. "ae3.s" > > .text > > .align 2 > .globl main >main: > .long 0 > .long 0 > .long 0 > lis r7,0x0180 > addi r7,r7,0x1000 > addi r8,r0,0 > stw r8,0(r7) > stw r8,0x1000(r7) > stw r8,0x2000(r7) > > lis r5,0x0180 > li r6,1 > stw r6,0(r5) > >..L4: > lwz r8,0(r7) > addi r8,r8,1 > stw r8,0(r7) > > lwz r8,0x1000(r7) > addi r8,r8,1 > stw r8,0x1000(r7) > > lwz r8,0x2000(r7) > addi r8,r8,1 > stw r8,0x2000(r7) > > b .L4 > > > ----- Posted via NewsOne.Net: Free (anonymous) Usenet News via the Web ----- http://newsone.net/ -- Free reading and anonymous posting to 60,000+ groups NewsOne.Net prohibits users from posting spam. If this or other posts made through NewsOne.Net violate posting guidelines, email abuse@newsone.net --------------------------- Newsgroups: comp.os.vxworks Subject: Re: While loop data access exception Date: 28 Dec 2001 05:26:11 GMT From: bcoxsprint12@earthlink.net Organization: NewsOne.Net - Free Usenet News via the Web - http://newsone.net/ Message-ID: References: <9vb2dv$rhc2@nntp.cig.mot.com> In article <9vb2dv$rhc2@nntp.cig.mot.com>, Michael writes: Hi Michael, It's a bad idea to run a program called 'main' under VxWorks*. Give your top-level function another name, like 'myMain' and try again. The GNU C compiler and linker make some assumptions about how to call your 'main' function, basically assuming it is a Unix- like single task entry-point. None of these assumptions are valid, so don't call any of your functions 'main'. >Quick intro: > >class class_a { > int f(const string& s1) { > printf("%s \n", s1.c_str(); > return 1; > } >}; > >main(void* arg) { > class_a a1; > string p1((char*) arg); > while(1) { > a1.f(p1); > } >} > >I am getting a data exception error when I run this code. > >data access >Exception current instruction address: 0x0621e19c >Machine Status Register: 0x0000b030 >Data Access Register: 0x9421ffec >Condition Register: 0x24000042 >Data storage interrupt Register: 0x40000000 >Task: 0x599f550 "pthr15" > >Of course I simplified class_a, but I put a printf at the start of function >f, and that is not being print. What are the steps for me to debug what is >going on? From what I can see, the function f is not being entered at all. >The wierdest part is that function f will be executed the first time, but >the second time through the while loop, it will create that exception. >Another wierd part is if I do not use the while loop, but rather just call >function f repeatedly, the program will work also. So, maybe the problem is >somehow linked to the loop? > >Thanks in advance for the help. If anyone has questions, please email me at >msheng1@email.mot.com. > > > > ----- Posted via NewsOne.Net: Free (anonymous) Usenet News via the Web ----- http://newsone.net/ -- Free reading and anonymous posting to 60,000+ groups NewsOne.Net prohibits users from posting spam. If this or other posts made through NewsOne.Net violate posting guidelines, email abuse@newsone.net --------------------------- Newsgroups: comp.os.vxworks Subject: GT48300 Device Driver?? or reference?? Date: 27 Dec 2001 22:24:40 -0800 From: netpert@hanmail.net (ryan) Organization: http://groups.google.com/ Message-ID: Hi.. Where can I get device driver or reference source code for Marvell's GT48300 ?? Thanks --------------------------- Newsgroups: comp.os.vxworks Subject: etherInputHookAdd ? Date: 27 Dec 2001 22:42:22 -0800 From: sardine_sz@163.net (sardine) Organization: http://groups.google.com/ Message-ID: hi all gurus, TornadoII、VxWorks5.4 T2CP3 W2KPatck double Intel82557 ethernet controller on x86 target. I. i wrote a test program use etherInputHookAdd and etherOutputHookAdd, and system add hook program success,but i cann't got any info from the hook function. can i invoke printf() in the hook function directly? II. have any ways to receive ethernet packet directly and send it to IP layor? any advice will appreciate! --------------------------- Newsgroups: comp.os.vxworks Subject: How to use the second END device in Socket@ Date: 27 Dec 2001 23:22:50 -0800 From: li.jingchi@mail.zte.com.cn (Networker) Organization: http://groups.google.com/ Message-ID: <18fe3c2d.0112272322.40b2446c@posting.google.com> every body: Now I want to know how to send/receive the message from one of the two END device in 8260. Is't the Ip address can be the ID to distinguish the two END? who can tell me? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: network protocol toolkit user guide Date: Fri, 28 Dec 2001 07:52:14 GMT From: Anthony Rouaux Organization: [posted via Easynet France] Message-ID: References: Johan Borkhuis wrote in news:Xns91846927A4126borkhuisagerecom@135.7.153.102: > http://www.windriver.com/csdocs/techpubs/tor-2.0/books.html The web site ask me a login and a password but my support licence has expired. What can i do? Can anyone send it to me by email? Thanks a lot Anthony --------------------------- Newsgroups: comp.os.vxworks Subject: Re: etherInputHookAdd ? Date: Fri, 28 Dec 2001 07:57:13 GMT From: Anthony Rouaux Organization: [posted via Easynet France] Message-ID: References: I have done the same test and a printf in a hook routine works well; i think it is not an ISR(interrupt service routine) so you can use printf. In my case, ethernet packet are automatically redirect to the IP layor. I think it works so because my driver is an END driver and it is connected to the mux in the BSP by the MuxDevLoad function. So in my hook routine i just have to return TRUE or FALSE depending on if i want to pass the ethernet packet to the upper layer. I hope this will help you. Bye Anthony sardine_sz@163.net (sardine) wrote in news:b1cb4c43.0112272242.132edcb7@posting.google.com: > add hook program success,but i cann't got any info from > the hook function. > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: network protocol toolkit user guide Date: 28 Dec 2001 08:52:34 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: Anthony Rouaux wrote: > The web site ask me a login and a password but my support licence has > expired. What can i do? Can anyone send it to me by email? > Thanks a lot Renew your support license? 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: etherInputHookAdd ? Date: 28 Dec 2001 08:57:41 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: sardine_sz@163.net (sardine) wrote: > i wrote a test program use etherInputHookAdd and etherOutputHookAdd, > and system add hook program success,but i cann't got any info from > the hook function. > can i invoke printf() in the hook function directly? I always prever to use logMsg instead of printf. This uses another task to do the output, so you can do it from within any function, ISR or not. 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 to use the second END device in Socket@ Date: 28 Dec 2001 02:44:40 -0800 From: sardine_sz@163.net (sardine) Organization: http://groups.google.com/ Message-ID: References: <18fe3c2d.0112272322.40b2446c@posting.google.com> li.jingchi@mail.zte.com.cn (Networker) wrote in message news:<18fe3c2d.0112272322.40b2446c@posting.google.com>... > every body: > Now I want to know how to send/receive the message from one of the > two END device in 8260. Is't the Ip address can be the ID to > distinguish the two END? > who can tell me? i think ,it depend on ipaddress to distinguish the different END. because every END device bind a exclusive ipaddress. i believe! --------------------------- Newsgroups: comp.os.vxworks Subject: Detecting Memory Leaks Date: Fri, 28 Dec 2001 12:10:56 +0100 From: "Mukul Dhankhar" Message-ID: <3c2c51e7$0$12221$4d4efb8e@news.be.uu.net> Hi all, I am doing some tests( which I code in C) on VxWorks AE 1.1 and in my tests I need to get the available physical memory( the RAM space left, because my target is PC-Pentium 200MHz). I need this because I load the system with a lot of threads and synchronization objects, and I test to see if there are any memory leaks. Does anyone know if there is an API function which - - returns the available total free space in the system( the space left on the RAM). Thanks, Mukul --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Sat Dec 29 04:03:14 2001 From: Vxworks Exploder Date: Sat Dec 29 04:03:17 PST 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sat Dec 29 04:03:10 PST 2001 Subject: How to link c object file with fortran object using ld68k? Subject: WRS are you here? Subject: Cannot get pci routines into vxWorks :-( Subject: parallel port driver Subject: BIT routines Subject: Re: BIT routines Subject: Re: How to link c object file with fortran object using ld68k? Subject: Re: BIT routines Subject: error while downloading Subject: Re: BIT routines Subject: Re: Cannot get pci routines into vxWorks :-( Subject: Re: error while downloading Subject: Re: Cannot get pci routines into vxWorks :-( Subject: Re: How to use the second END device in Socket@ Subject: Re: How to link c object file with fortran object using ld68k? Subject: Re: BIT routines Subject: Is write() with 0 bytes guaranteed to be given to driver? Subject: will the driver get the write(fd, NULL, n) call? Subject: what strange thing does LogMsg do? Subject: Re: will the driver get the write(fd, NULL, n) call? Subject: smart card reader Subject: Re: How to use the second END device in Socket@ Subject: Networking in VxSim Subject: pci error in IDT Rc32332 Subject: Re: WRS are you here? Subject: experience about WindNet Radius And WindNet PPP? ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: How to link c object file with fortran object using ld68k? Date: Fri, 28 Dec 2001 21:40:38 +0800 From: "Liu Guan Feng" Organization: Hong Kong University of Science and Technology Message-ID: Hello everyone, I had a problem when I link c object files with fortran object files. The problem can be described as following: I want to compile mycfile.c and myfortran.f and link them together into mytarget.o. Then I use ld It seems like there are a lot of VxWorks users providing a lot of free advice and help. Where doe's WRS stand in all this. Seems like they should play a leading role here rather than using silence as a way of milking money out of people for support. If you want us to spread the message that VxWorks is good and easy to use then make it so. Anyway, just a peeve. Hey I really appreciate all the information you all are providing me. If I ever get good at this maybe I can return the favor. Popeye --------------------------- Newsgroups: comp.os.vxworks Subject: Cannot get pci routines into vxWorks :-( Date: 28 Dec 2001 09:09:05 -0800 From: spurcell@apl.washington.edu (Steve) Organization: http://groups.google.com/ Message-ID: <3b9f1aed.0112280909.74cb92a8@posting.google.com> Ok so I'm just a hardware guy, and I don't understand the finer points of software, "especially the part about how it doesn't work all the time". I'm having no luck getting the following routines included into vxWorks: pciIomapLibInit pciIntConnect pciInt pciIntDisconnect I've created a .cdf file with the following code: Component INCLUDE_PCI { NAME Include PCI library _CHILDREN FOLDER_BUSES SYNOPSIS includes PCI support } This did indeed add an entry to the vxWorks build options, but it did not fix my problem. What else am I missing? Wouldn't life be better if I could just modify a make file to include the files I need?? Thanks in advance. Popeye --------------------------- Newsgroups: comp.os.vxworks Subject: parallel port driver Date: 28 Dec 2001 09:34:12 -0800 From: mgarcia@psl.nmsu.edu (Michael Garcia) Organization: http://groups.google.com/ Message-ID: We are using MVME2604 BSP and need a driver for the Parallel Port, any help in obtaining some guidance would be greatly appreciated... Mike --------------------------- Newsgroups: comp.os.vxworks,comp.sys.powerpc.tech Subject: BIT routines Date: Fri, 28 Dec 2001 18:06:49 GMT From: "Kelly Hornsby" Organization: Excite@Home - The Leader in Broadband http://home.com/faster Message-ID: Hello All I have been tasked to find and/or create some BIT routines for a PowerPC 745/MPC107 combo system that my company built. I'm looking for pointers and other idea on what a PPC BIT should perform, both at startup and during operation. I'm also looking for any prebuilt BIT routines that may be available, either in assembler or in HOL form. Any thoughts and pointers are welcome. thx --------------------------- Newsgroups: comp.os.vxworks,comp.sys.powerpc.tech Subject: Re: BIT routines Date: Fri, 28 Dec 2001 18:10:12 GMT From: "Jim" Organization: Magma Communications Ltd. Message-ID: <8A2X7.11226$8e.440946@news> References: Do meab BIST (Buil In Self Test) ... "Kelly Hornsby" wrote in message news:Zw2X7.3277$LQ1.1060754@news2.nash1.tn.home.com... > Hello All > > I have been tasked to find and/or create some BIT routines for a PowerPC > 745/MPC107 combo system that my company built. I'm looking for pointers and > other idea on what a PPC BIT should perform, both at startup and during > operation. I'm also looking for any prebuilt BIT routines that may be > available, either in assembler or in HOL form. > > Any thoughts and pointers are welcome. > > thx > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to link c object file with fortran object using ld68k? Date: 28 Dec 2001 10:17:41 -0800 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0112281017.7353ff9e@posting.google.com> References: Hello, You don't say what the target arch is for the fortran compiler. Judging from the name you use (just f77), I'd suspect that it is a fortran compiler for your host, and not a cross compiler for 68K. So, you have a couple of options: 1) Get a fortran cross compiler for VxWorks on 68K (not sure how easy this is, but since there is a GNU f77 it should be possible) 2) Try to convert the fortran to C. There is a tool for this at http://cm.bell-labs.com/netlib/f2c/index.html though I don't know how good it is. HTH, John... "Liu Guan Feng" wrote in message news:... > Hello everyone, > > I had a problem when I link c > object files with fortran object files. > The problem can be described as > following: > I want to compile mycfile.c and > myfortran.f and link them together into > mytarget.o. Then I use ld to motorlora68040 for executing. > The Makefile is as follows: > CC = cc68k > LD = ld68k > CFLAGS > = -g -I/tornardo1.0.1/target/h > > all: mytarget.o mycfile.o myfortran.o > mycfile.o: mycfile.c > $(CC) $(CFLAGS) -c > myfile.c > > myfortran.o: myfortran.f > f77 -c myfortran.f > > mytarget.o: mycfile.o myfortran.o > $(LD) -o > mytarget.o -r mycfile.o myfortran.o > > However, after make, the linker always > say that it can not recoganize the > format of myfortran.o. > I think maybe I need to use other > fortran compiler than f77 to compile > myfortran.f > or add other options to f77 so as to > change the format of myfortran.o > to be recoganized by ld68k. > > Could you help me? > > Thx a lot. > Liu Guanfeng. > EE. Dept. Hong Kong University of > Science and Technology --------------------------- Newsgroups: comp.os.vxworks,comp.sys.powerpc.tech Subject: Re: BIT routines Date: 28 Dec 2001 18:45:22 GMT From: "Mark A. Odell" Organization: Embedded Firmware Design, Inc. Message-ID: References: <8A2X7.11226$8e.440946@news> "Jim" wrote in news:8A2X7.11226$8e.440946@news: > > "Kelly Hornsby" wrote in message > news:Zw2X7.3277$LQ1.1060754@news2.nash1.tn.home.com... >> Hello All >> >> I have been tasked to find and/or create some BIT routines for a PowerPC >> 745/MPC107 combo system that my company built. I'm looking for pointers >> and other idea on what a PPC BIT should perform, both at startup and >> during operation. I'm also looking for any prebuilt BIT routines that >> may be available, either in assembler or in HOL form. This is typically so system specific I don't know how you could get pre- written versions. What do you want to test? Memory? You can get mem. tests in C easily. I'd guess you'll be writing most of the BIST routines yourself. Hey, they're fun! > Do meab BIST (Buil In Self Test) ... I've heard it called BIT but I admit, BIST is far more common in my experience. - -- - - Mark A. Odell - - Embedded Firmware Design, Inc. - - http://www.embeddedfw.com --------------------------- Newsgroups: comp.os.vxworks Subject: error while downloading Date: 28 Dec 2001 11:21:03 -0800 From: sksinha2001@yahoo.com (Sandeep Sinha) Organization: http://groups.google.com/ Message-ID: <6040814b.0112281121.7084b294@posting.google.com> Hi, I have just now complied a demo program for Socket programming(Server). Build has the following output: vxrm *.o *.rpo ctdt.c symTbl.c vxApp* vxrm C:\samtarget\samproj\Project2\prjComps.h C:\samtarget\samproj\Project2\prjParams.h C: \samtarget\samproj\Project2\prjConfig.c C:\samtarget\samproj\Project2\linkSyms.c vxrm C:\samtarget\samproj\Project2\libs.nm C:\samtarget\samproj\Project2\libs.size ccsimpc -U__WINNT -UWIN32 -U__WINNT__ -UWINNT -U__MINGW32__ -U_WIN32 - -U__WIN32 -U__WIN32__ -U__WIN32 -mpentium -ansi -nostdinc -g -nostdlib -fno-builtin - -fno-defer-pop -Wall -DRW_M ULTI_THREAD -D_REENTRANT -I. -IC:\Tornado\target\h -DCPU=SIMNT -c C:\downloads\server.c C:\downloads\server.c: In function `main': C:\downloads\server.c:12: warning: implicit declaration of function `socket' C:\downloads\server.c:17: warning: implicit declaration of function `inet_addr' C:\downloads\server.c:19: warning: implicit declaration of function `bind' C:\downloads\server.c:23: warning: implicit declaration of function `listen' C:\downloads\server.c:25: warning: implicit declaration of function `accept' C:\downloads\server.c:28: warning: implicit declaration of function `inet_ntoa' C:\downloads\server.c:28: warning: format argument is not a pointer (arg 2) C:\downloads\server.c:36: warning: implicit declaration of function `send' C:\downloads\server.c:37: warning: implicit declaration of function `recv' ccsimpc -nostdlib -r -Wl,-X - -Wl,@C:\samtarget\samproj\Project2\prjObjs.lst -o partialImag e.o nmsimpc -g partialImage.o @C:\samtarget\samproj\Project2\prjObjs.lst | wtxtcl C:\Tornado\h ost\src\hutils\munch.tcl -asm simpc > ctdt.c ccsimpc -c -fdollars-in-identifiers -U__WINNT -UWIN32 -U__WINNT__ - -UWINNT -U__MINGW32__ -U _WIN32 -U__WIN32 -U__WIN32__ -U__WIN32 -mpentium -ansi -nostdinc -g - -nostdlib -fno-builtin -fno-defer-pop -Wall -DRW_MULTI_THREAD -D_REENTRANT -I. - -IC:\Tornado\target\h -DCPU=SIMNT ctdt.c -o ctdt.o ccsimpc -nostdlib -r -Wl,-X -Wl,--force-stabs-reloc partialImage.o ctdt.o -o Project2.out Done. While downloading Project2.out, I get the following result: Errors while downloading C:/samtarget/samproj/Project2/SIMNTgnu/Project2.out: _socket _inet_ntoa _bind _send _recv _accept _listen _inet_addr Please let me know if anybody knows how to resolve it. Thanks, Sandeep --------------------------- Newsgroups: comp.os.vxworks,comp.sys.powerpc.tech Subject: Re: BIT routines Date: Fri, 28 Dec 2001 20:46:52 GMT From: "Jim" Organization: Magma Communications Ltd. Message-ID: <0T4X7.11265$8e.447367@news> References: <8A2X7.11226$8e.440946@news> For RAM test routines go to Dr Dobbs www page a look for article by Michael Barr ... he has some code there for walking ones test etc etc etc ... Jim "Mark A. Odell" wrote in message news:Xns91858B648DD52lkj562ghjgk1k245lbvj@130.133.1.4... > "Jim" wrote in news:8A2X7.11226$8e.440946@news: > > > > > "Kelly Hornsby" wrote in message > > news:Zw2X7.3277$LQ1.1060754@news2.nash1.tn.home.com... > >> Hello All > >> > >> I have been tasked to find and/or create some BIT routines for a PowerPC > >> 745/MPC107 combo system that my company built. I'm looking for pointers > >> and other idea on what a PPC BIT should perform, both at startup and > >> during operation. I'm also looking for any prebuilt BIT routines that > >> may be available, either in assembler or in HOL form. > > This is typically so system specific I don't know how you could get pre- > written versions. What do you want to test? Memory? You can get mem. tests > in C easily. I'd guess you'll be writing most of the BIST routines yourself. > Hey, they're fun! > > > Do meab BIST (Buil In Self Test) ... > > I've heard it called BIT but I admit, BIST is far more common in my > experience. > > -- > - Mark A. Odell > - Embedded Firmware Design, Inc. > - http://www.embeddedfw.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Cannot get pci routines into vxWorks :-( Date: 28 Dec 2001 13:16:35 -0800 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0112281316.6d007664@posting.google.com> References: <3b9f1aed.0112280909.74cb92a8@posting.google.com> Hello, > I'm having no luck getting the following routines included into > vxWorks: > > pciIomapLibInit > pciIntConnect > pciInt > pciIntDisconnect > > I've created a .cdf file with the following code: > > Component INCLUDE_PCI { > NAME Include PCI library > _CHILDREN FOLDER_BUSES > SYNOPSIS includes PCI support > } Well, you didn't tell it where to find the routines. You would need to specify a .o file that contained them (use the MODULES keyword). That said, these are not really meant to be componentised for VxWorks 5.x (which is what I assume you are using). Instead they are referenced from the BSP and linked in when the kernel is built. In fact, if your board needs these then I am surprised that they are not already being linked into the BSP - normally if you need pciIntConnect most of the drivers would be using it, and therefore it would already be referenced in the BSP code. Check that you have INCLUDE_PCI defined in your BSP's config.h file. Looking at a BSP I have that uses this, I see that in sysLib.c the following files are being included when INCLUDE_PCI is defined: pci/pciConfigLib.c pci/pciIntLib.c These can be found under target/src/drv (which is in the include path for the BSP builds). Also, in sysHwInit2() there is a call to pciIntLibInit() to initialise the PCI interrupt support, and all the drivers are told to connect their ISRs using pciIntConnect() and not intConnect() so that they can take advantage of the shared interrupt demux. If you are still having problems, please tell us more about the system (at least the version of Tornado, the architecture and BSP, and how you are using these routines). HTH, John... > This did indeed add an entry to the vxWorks build options, but it did > not fix my problem. > > What else am I missing? > > Wouldn't life be better if I could just modify a make file to include > the files I need?? > > Thanks in advance. > Popeye --------------------------- Newsgroups: comp.os.vxworks Subject: Re: error while downloading Date: 28 Dec 2001 13:18:14 -0800 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0112281318.1f4f7f89@posting.google.com> References: <6040814b.0112281121.7084b294@posting.google.com> Hello, You need to buy the full simulator to make use of networking functions. It is a limitation of the standard one that it does not include any network support. HTH, John... PS This ia a very frequently asked question; please check in the archives before posting in future - you'll get your answer much faster, and the group won't have to answer the same questions over and over again. sksinha2001@yahoo.com (Sandeep Sinha) wrote in message news:<6040814b.0112281121.7084b294@posting.google.com>... > Hi, > > I have just now complied a demo program for Socket > programming(Server). > Build has the following output: > > vxrm *.o *.rpo ctdt.c symTbl.c vxApp* > vxrm C:\samtarget\samproj\Project2\prjComps.h > C:\samtarget\samproj\Project2\prjParams.h C: > \samtarget\samproj\Project2\prjConfig.c > C:\samtarget\samproj\Project2\linkSyms.c > vxrm C:\samtarget\samproj\Project2\libs.nm > C:\samtarget\samproj\Project2\libs.size > ccsimpc -U__WINNT -UWIN32 -U__WINNT__ -UWINNT -U__MINGW32__ -U_WIN32 > -U__WIN32 -U__WIN32__ > -U__WIN32 -mpentium -ansi -nostdinc -g -nostdlib -fno-builtin > -fno-defer-pop -Wall -DRW_M > ULTI_THREAD -D_REENTRANT -I. -IC:\Tornado\target\h -DCPU=SIMNT -c > C:\downloads\server.c > C:\downloads\server.c: In function `main': > C:\downloads\server.c:12: warning: implicit declaration of function > `socket' > C:\downloads\server.c:17: warning: implicit declaration of function > `inet_addr' > C:\downloads\server.c:19: warning: implicit declaration of function > `bind' > C:\downloads\server.c:23: warning: implicit declaration of function > `listen' > C:\downloads\server.c:25: warning: implicit declaration of function > `accept' > C:\downloads\server.c:28: warning: implicit declaration of function > `inet_ntoa' > C:\downloads\server.c:28: warning: format argument is not a pointer > (arg 2) > C:\downloads\server.c:36: warning: implicit declaration of function > `send' > C:\downloads\server.c:37: warning: implicit declaration of function > `recv' > ccsimpc -nostdlib -r -Wl,-X > -Wl,@C:\samtarget\samproj\Project2\prjObjs.lst -o partialImag > e.o > nmsimpc -g partialImage.o @C:\samtarget\samproj\Project2\prjObjs.lst | > wtxtcl C:\Tornado\h > ost\src\hutils\munch.tcl -asm simpc > ctdt.c > ccsimpc -c -fdollars-in-identifiers -U__WINNT -UWIN32 -U__WINNT__ > -UWINNT -U__MINGW32__ -U > _WIN32 -U__WIN32 -U__WIN32__ -U__WIN32 -mpentium -ansi -nostdinc -g > -nostdlib -fno-builtin > -fno-defer-pop -Wall -DRW_MULTI_THREAD -D_REENTRANT -I. > -IC:\Tornado\target\h -DCPU=SIMNT > ctdt.c -o ctdt.o > ccsimpc -nostdlib -r -Wl,-X -Wl,--force-stabs-reloc partialImage.o > ctdt.o -o Project2.out > > Done. > > While downloading Project2.out, I get the following result: > > Errors while downloading > C:/samtarget/samproj/Project2/SIMNTgnu/Project2.out: > _socket > _inet_ntoa > _bind > _send > _recv > _accept > _listen > _inet_addr > > Please let me know if anybody knows how to resolve it. > > Thanks, > Sandeep --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Cannot get pci routines into vxWorks :-( Date: Sat, 29 Dec 2001 00:20:26 GMT From: "Daniel Poirot" Organization: Prodigy Internet http://www.prodigy.com Message-ID: References: <3b9f1aed.0112280909.74cb92a8@posting.google.com> Reply-To: "Daniel Poirot" Add the following line to the config.h file of the BSP then build a new project based on that BSP: #define INCLUDE_PCI ...of course this won't do a bit of good if your target CPU doesn't have a PCI bus in the first place. Any clue on what kind of target you are using? Where did the BSP come from? Any other useless information that you might consider adding? Have you read the BSP Developer's Kit User's Guide? "Steve" wrote in message news:3b9f1aed.0112280909.74cb92a8@posting.google.com... > Ok so I'm just a hardware guy, and I don't understand the finer > points of software, "especially the part about how it doesn't work > all the time". > > I'm having no luck getting the following routines included into > vxWorks: > > pciIomapLibInit > pciIntConnect > pciInt > pciIntDisconnect > > I've created a .cdf file with the following code: > > Component INCLUDE_PCI { > NAME Include PCI library > _CHILDREN FOLDER_BUSES > SYNOPSIS includes PCI support > } > > This did indeed add an entry to the vxWorks build options, but it did > not fix my problem. > > What else am I missing? > > Wouldn't life be better if I could just modify a make file to include > the files I need?? > > Thanks in advance. > Popeye --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to use the second END device in Socket@ Date: 28 Dec 2001 16:35:37 -0800 From: li.jingchi@mail.zte.com.cn (Networker) Organization: http://groups.google.com/ Message-ID: <18fe3c2d.0112281635.1fdfb4a4@posting.google.com> References: <18fe3c2d.0112272322.40b2446c@posting.google.com> > i think ,it depend on ipaddress to distinguish the different END. > because every END device bind a exclusive ipaddress. > i believe! But,in fact, under windows OS we can bind sever IP address to an ethernet card! Dose Vxworks not surpport this feature? thx! --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to link c object file with fortran object using ld68k? Date: 29 Dec 2001 00:32:13 GMT From: bcoxsprint12@earthlink.net Organization: NewsOne.Net - Free Usenet News via the Web - http://newsone.net/ Message-ID: References: In article , Liu Guan Feng writes: Hello, This can only, possibly, work if you compiled your FORTRAN codes with a 68000 cross-compiler. Take a look at one of the f77 object files with objdump and see what format and what processor it's for. Was the f77 compiler that you used a 68000 cross compiler? >Hello everyone, > > I had a problem when I link c >object files with fortran object files. >The problem can be described as >following: > I want to compile mycfile.c and >myfortran.f and link them together into >mytarget.o. Then I use ld to motorlora68040 for executing. >The Makefile is as follows: > CC = cc68k > LD = ld68k > CFLAGS >= -g -I/tornardo1.0.1/target/h > > all: mytarget.o mycfile.o myfortran.o > mycfile.o: mycfile.c > $(CC) $(CFLAGS) -c >myfile.c > > myfortran.o: myfortran.f > f77 -c myfortran.f > >mytarget.o: mycfile.o myfortran.o > $(LD) -o >mytarget.o -r mycfile.o myfortran.o > >However, after make, the linker always >say that it can not recoganize the >format of myfortran.o. >I think maybe I need to use other >fortran compiler than f77 to compile >myfortran.f >or add other options to f77 so as to >change the format of myfortran.o >to be recoganized by ld68k. > >Could you help me? > >Thx a lot. >Liu Guanfeng. >EE. Dept. Hong Kong University of >Science and Technology > > > ----- Posted via NewsOne.Net: Free (anonymous) Usenet News via the Web ----- http://newsone.net/ -- Free reading and anonymous posting to 60,000+ groups NewsOne.Net prohibits users from posting spam. If this or other posts made through NewsOne.Net violate posting guidelines, email abuse@newsone.net --------------------------- Newsgroups: comp.os.vxworks,comp.sys.powerpc.tech Subject: Re: BIT routines Date: 28 Dec 2001 18:12:20 -0800 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0112281812.3c357e1d@posting.google.com> References: Hello, Don't have any code, but something you could certainly look at for ideas of the kind of tests you might want to write is the MCG docs for PPCBug. There is a description of the diagnostics they perform in this document: http://www.mcg.mot.com/us/products/docs/pdf/ppcdiaa_um2.pdf This is a pretty thorough test, and covers a lot of the devices on the supported boards too; you may, or may not, need as much! Something else to look into is whether you have hardware DRAM ECC circuitry; if so, you may find that it maintains statistics on number of corrections made, and even the number of uncorrected errors (hope you're not seeing any of those ;-). A (high priority) monitor task can easily be set to track these, and report problems if there is a sustained rise in the level(s). HTH, John... > I have been tasked to find and/or create some BIT routines for a PowerPC > 745/MPC107 combo system that my company built. I'm looking for pointers and > other idea on what a PPC BIT should perform, both at startup and during > operation. I'm also looking for any prebuilt BIT routines that may be > available, either in assembler or in HOL form. > > Any thoughts and pointers are welcome. > > thx --------------------------- Newsgroups: comp.os.vxworks Subject: Is write() with 0 bytes guaranteed to be given to driver? Date: Fri, 28 Dec 2001 19:11:57 -0800 From: "Mark Galecki" Message-ID: <3c2d3454$0$25215$724ebb72@reader2.ash.ops.us.uu.net> Hello, if an application makes a function call write(fd, &buf, 0) is it guaranteed that the driver callback routine xxwrite() will be called? Thank you, Mark Galecki --------------------------- Newsgroups: comp.os.vxworks Subject: will the driver get the write(fd, NULL, n) call? Date: Fri, 28 Dec 2001 19:27:04 -0800 From: "Mark Galecki" Message-ID: <3c2d37df$0$25219$724ebb72@reader2.ash.ops.us.uu.net> And I have a similar question to my last one: Is it guaranteed that the driver will get its callback xxwrite() called after an application calls write(fd, NULL, n) Mark Galecki --------------------------- Newsgroups: comp.os.vxworks Subject: what strange thing does LogMsg do? Date: 28 Dec 2001 21:16:04 -0800 From: ultraming@21cn.com (Chris Klein) Organization: http://groups.google.com/ Message-ID: logMsg sometimes seems to get blocked, and even more, blocks other tasks. In my realtek rtl81x9 END driver(Platform: Vxworks 5.4, Ixp1200 StrongARM), I add a logMsg before "END_RCV_RTN_CALL" in function "rtl81x9HandleRecvInt()", thus once a packet is properly received, I can see a message on my hyperterminal. Every time I starts my board, it properly downloads the loadable Vxworks image from my PC and starts the shell. And then after servel "packet received" message printed, no more prints come out until I hit an Enter, which brings out lots of "packet received" messages immediately, but it seems halted again until I hit another Enter. If I ping my board when it looks halted, I get no response, but once I hit the Enter key, I can capture all the ICMP response packet sent my board using Sniffer Pro. This happens a lot, averagely I can ping my board with a loss of about 20-30%. Another thing strange is: If i remove this logMsg line, my driver performs even worse, ping will lose more than 60% tries. Need veterans' help, I appreciate very much. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: will the driver get the write(fd, NULL, n) call? Date: 28 Dec 2001 22:23:25 -0800 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0112282223.229a816a@posting.google.com> References: <3c2d37df$0$25219$724ebb72@reader2.ash.ops.us.uu.net> Hello, To answer both of your questions... It depends on what you mean by guaranteed. If you mean is it guaranteed on current versions of the OS, then the answer would seem to be yes. There doesn't seem to be any check in the generic I/O code on the buffer or the length - they are just passed to the driver's entry point. If you mean is this behaviour guaranteed for future; I would guess not since it is not documented anywhere that it will be supported. HTH, John... "Mark Galecki" wrote in message news:<3c2d37df$0$25219$724ebb72@reader2.ash.ops.us.uu.net>... > And I have a similar question to my last one: > > Is it guaranteed that the driver will get its callback xxwrite() called > after an application calls > > write(fd, NULL, n) > > Mark Galecki --------------------------- Newsgroups: comp.os.vxworks Subject: smart card reader Date: Sat, 29 Dec 2001 14:33:57 +0800 From: "Zhao Yandong" Organization: Bentium Ltd. (CN99) Message-ID: Our system will use a smart card reader. Now the hardware of smart card reader has been designed out, and now we need to write the device driver for it. But I don't know how to write the driver code. If I write the driver, which files should I modify or add ? Could anyone give me some guide ? Thanks The target os is Vxworks5.4, host is win98, board is ARM --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to use the second END device in Socket@ Date: 28 Dec 2001 22:55:26 -0800 From: sardine_sz@163.net (sardine) Organization: http://groups.google.com/ Message-ID: References: <18fe3c2d.0112272322.40b2446c@posting.google.com> <18fe3c2d.0112281635.1fdfb4a4@posting.google.com> li.jingchi@mail.zte.com.cn (Networker) wrote in message news:<18fe3c2d.0112281635.1fdfb4a4@posting.google.com>... > > i think ,it depend on ipaddress to distinguish the different END. > > because every END device bind a exclusive ipaddress. > > i believe! > > But,in fact, under windows OS we can bind sever IP address to an ethernet card! > Dose Vxworks not surpport this feature? > thx! yeah, on windown system,you can bind multiple ipaddress on a TCP/IP stack. but ,on vxworks,every TCP/IP stack bind a ipaddress,also the two ipaddress cannot have same subnet mask. only my view ^_^ --------------------------- Newsgroups: comp.os.vxworks Subject: Networking in VxSim Date: 28 Dec 2001 23:19:15 -0800 From: rashanmu@npd.hcltech.com (Ramesh) Organization: http://groups.google.com/ Message-ID: <15a87607.0112282319.5f582c94@posting.google.com> HI , I am trying Networking between two VxSim's in a ethernet Lan n/w. The set up is like My PC ethernet adapter: 192.168.100.69 ( Win NT ) My ULIP adapter: 90.0.0.254 My VxWorks target iamge : 90.0.0.1 Other end ethernet address : 192.168.100.114 ( Win 2k) Other end ULIP address : 90.0.0.253 Other end VxWorks Target image : 90.0.0.2 In both PC's IP forwarding is enabled. I have looked thru the archives and with in PC i can ping between targets if i run more than one. But between vxSim in network is the one am not able to acheive. Here from NT i can able to ping 90.0.253 and same way from win2k i can able to ping 90.0.254. But from the VxWorks target shell ( 90.0.0.1) i can able to ping 90.0.0.254 and 192.168.100.69 but not 192.168.100.114 and 90.0.0.253 and same way vice versa from win2k pc also. What i have to do. Kindly help me. I got struck up with this for so long time. Even i posted here once before, but no response. I tried but i could not able to establish it. Kindly help me. Also one more doubt is will Win NT workstation will do IP forwarding ( when i enalbe it in the check box mentioned in Network properties ). I will be very much thankful for any replies. Thanks, Ramesh --------------------------- Newsgroups: comp.os.vxworks Subject: pci error in IDT Rc32332 Date: 28 Dec 2001 23:23:49 -0800 From: sczhang@esteemnetworks.com (ShiChao Zhang) Organization: http://groups.google.com/ Message-ID: <4ceedd43.0112282323.368465f8@posting.google.com> I am using the IDT RC32332 risc cpu, and I came across some strange problems of PCI. Here is the situation: 1. When I try to use the PCI Configuration Data Register at 1800_2cfc,Address Register at 1800_2cf8 to access the PCI configuration space, any write operation is OK(each operation comprise of a write access to address Reg and a write/read operation to data Reg). But If I perform a read operation, it will generate a Bus Error Exception. Then I tried to locate why the exception take place. And I found that if I put a access to Address Reg followed by a printf function call, then begin the normal read operation, it will be OK, no exception will occur. I can't understand it. 2. During above situation, the read operation is still not correctly. my operation is as follow: RegPciConfigAddr = (unsigned int *)IDT134_PCI_CONFIG_ADDR; // 0xb8002cf8 RegPciConfigData = (unsigned int *)IDT134_PCI_CONFIG_DATA; // 0xb8002cfc *RegPciConfigAddr = 0x80000004;// this access it to avoid bus error exception printf("to avoid the exception \n"); *RegPciConfigAddr = 0x80000000; printf("idt device id and vendor id:%x \r\n",*RegPciConfigData); *RegPciConfigAddr = 0x80000004; printf("idt status and command:%x \r\n",*RegPciConfigData); *RegPciConfigAddr = 0x80000008; printf("idt class code:%x \r\n",*RegPciConfigData); *RegPciConfigAddr = 0x8000000c; printf("idt BIST, header:%x \r\n",*RegPciConfigData); ..... But the result is : to avoid the exception idt device id and vendor id:2a00157 idt status and command:205111d idt class code:2a00157 idt BIST, header:6800001 .... And I write to the IDT configuration space value is: #define RHEA_PCI_CONFIG0 0x0205111d /* Device ID & Vendor ID */ #define RHEA_PCI_CONFIG1 0x02a00157 /* Status & Command */ #define RHEA_PCI_CONFIG2 0x06800001 /* Class Code & Revision ID */ #define RHEA_PCI_CONFIG3 0x0000ff04 /* BIST, Header Type, Latency, & Cacheline Size */ #define RHEA_PCI_CONFIG4 0xa0000000 /* PCI Memory Address that Rhea responds to. */ #define RHEA_PCI_CONFIG5 0x60000000 /* PCI Dual Cycle Address that Rhea responds to.*/ #define RHEA_PCI_CONFIG6 0x00800001 /* PCI I/O Address that Rhea responds to. */ #define RHEA_PCI_CONFIG7 0x00000000 #define RHEA_PCI_CONFIG8 0x00000000 #define RHEA_PCI_CONFIG9 0x00000000 #define RHEA_PCI_CONFIG10 0x00000000 #define RHEA_PCI_CONFIG11 0x013410b5 #define RHEA_PCI_CONFIG12 0x00000000 #define RHEA_PCI_CONFIG13 0x00000000 #define RHEA_PCI_CONFIG14 0x00000000 #define RHEA_PCI_CONFIG15 0x38080101 You can see, all the value is shifted by one 32bit word. It seems a latency is needed to read the correct value of the Address Reg pointing to, the time is nearly a printf() routine executing time. So I programed a routine , each read operation is comprised of two read access : *RegPciConfigAddr = 0x80000000; printf("idt device id and vendor id:%x \r\n",*RegPciConfigData); *RegPciConfigAddr = 0x80000000; printf("idt device id and vendor id:%x \r\n",*RegPciConfigData); By using this form to read all the PCI configuration space value, I can got the correct value in each second access. What's the matter? 3. All my hardware design is similar to the IDT 32334 evaluation board, except that I am using the 32332,and using a 16bit flash to boot the board. And the pci connection is the same as the evaluation board. Thanks in advance! --------------------------- Newsgroups: comp.os.vxworks Subject: Re: WRS are you here? Date: Sat, 29 Dec 2001 10:28:49 +0200 From: "Leonid Rosenboim" Organization: Verio Message-ID: References: <3b9f1aed.0112280823.390475fc@posting.google.com> Steve, your perception is misleading: First, you and all participants of this newsgroup should be aware that WRS are getting, archiving, indexing and using all the stuff posted here, as part of WRS's internal knowledge base, along with all the TSRs and SPRs. Also, in the not so distant past, some executives used to read these postings on a daily basis, and probably they still do. So WRS is DEFINITELY LISTENING. Responding to posts on this newsgroup however, is banned by WRS corporate policy, which states that paying customers should get help via the proper support channels, and WRS employees are strictly forbidden to post to this newsgroup. Shortly after my company RST was acquired by WRS in Jan 1997, I responded to someone post partly to correct untrue criticism, partly to show the solution. The problem was reported by a paying customer, and I considered the public response to be a proper action to avoid others from getting the wrong impression about the quality of WRS product. It is then that I became fully aware of the policy banning app posts, and if not for this policy, there would be quite many responses from WRS employees in this newsgroup. Despite the above, I suspect that a couple of guys who still work for WRS, do respond on a regular basis, using their private and usually concealed identities. These guys should be admired the most, because they risk their jobs posting here, all for the benefit of the customers and public perception of WRS product quality. Someday, I hope, these guys will be honored and prized, but meanwhile the mere fact that such people still work for WRS is distinct sign that WRS is still a good company, despite its apparent tendency to beurocracy. - - Happy New Year to All! - - Leonid Rosenboim "Steve" wrote in message news:3b9f1aed.0112280823.390475fc@posting.google.com... > It seems like there are a lot of VxWorks users providing a lot of free > advice and help. > > Where doe's WRS stand in all this. > > Seems like they should play a leading role here rather than using > silence as a way of milking money out of people for support. > > If you want us to spread the message that VxWorks is good and easy to > use then make it so. > > Anyway, just a peeve. > > Hey I really appreciate all the information you all are providing me. > If I ever get good at this maybe I can return the favor. > > Popeye --------------------------- Newsgroups: comp.os.vxworks Subject: experience about WindNet Radius And WindNet PPP? Date: 29 Dec 2001 01:27:31 -0800 From: sardine_sz@163.net (sardine) Organization: http://groups.google.com/ Message-ID: hi folks, is there someone have experience about WindNet Radius And WindNet PPP? what's different between PPP1.0 and PPP1.2? and how to distinguish with them? thx! --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Sun Dec 30 04:03:09 2001 From: Vxworks Exploder Date: Sun Dec 30 04:03:12 PST 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sun Dec 30 04:03:06 PST 2001 Subject: curses Subject: Re: Networking in VxSim Subject: Re: etherInputHookAdd ? Subject: Re: etherInputHookAdd ? Subject: Re: Making a .lib file Subject: Detecting unused variables in Diab Data C compiler Subject: fei82559 hangs on tx Subject: which network devices supported by VxWorks? ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: curses Date: Sat, 29 Dec 2001 21:07:54 GMT From: "Weddick" Organization: AT&T Broadband Message-ID: Has anyone done a port or know of a port for using curses on a X86 Vxworks System? Thanks, --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Networking in VxSim Date: 29 Dec 2001 08:14:11 -0800 From: raj_rashid@hotmail.com (Rashid) Organization: http://groups.google.com/ Message-ID: <2a6118da.0112290814.48658d6e@posting.google.com> References: <15a87607.0112282319.5f582c94@posting.google.com> Hi Ramesh, > My PC ethernet adapter: 192.168.100.69 ( Win NT ) > My ULIP adapter: 90.0.0.254 > My VxWorks target iamge : 90.0.0.1 On this host do, C:\> route ADD 90.0.0.2 192.168.100.144 Now u will be able to ping 90.0.0.2 on other host. > Other end ethernet address : 192.168.100.114 ( Win 2k) > Other end ULIP address : 90.0.0.253 > Other end VxWorks Target image : 90.0.0.2 Best regards, Rashid --------------------------- Newsgroups: comp.os.vxworks Subject: Re: etherInputHookAdd ? Date: Sat, 29 Dec 2001 22:27:05 -0800 From: "drdiags" Organization: laserlink.net Message-ID: References: Hello sardine, "sardine" wrote in message news:b1cb4c43.0112272242.132edcb7@posting.google.com... > hi all gurus, > > TornadoII、VxWorks5.4 T2CP3 W2KPatck > double Intel82557 ethernet controller on x86 target. <.....SNIP.....> > > II. > have any ways to receive ethernet packet directly and send it to > IP layor? <......SNIP.....> If you are not using DHCP, you could register with the MUX to have a MUX_SNARF routine added to the IP stack. Upon reception, your routine would be provided the packet for whatever purposes you wanted. If your routine returns the correct TRUE/FALSE value, the stack will either assume you have freed the packet or it will move it further along the IP Protocol. Check to see if the Network Programmer's guide or the NPT can give you more information. Unfortunately, unless you go to T2.0.2 or above, there can only be one SNARF routine (not sure if this is per interface, others may be more clearer on this point). --------------------------- Newsgroups: comp.os.vxworks Subject: Re: etherInputHookAdd ? Date: Sat, 29 Dec 2001 22:46:33 -0800 From: "drdiags" Organization: laserlink.net Message-ID: References: Forgot to mention, I am assuming you are using an END driver and not a BSD driver. Also, that should have been if you haven't added a DHCP "Server" to your image. It uses the MUX_SNARF protocol for an END device. Also, WDB should be set to use network and not END as its access technique (you lose System Level debugging) if you choose to add a MUX_SNARF routine. I also thought the vxsniff.c example available on the vxWorks FAQ uses etherInputHookAdd. I had a vxWorks 5.3.1 ether input/output hook example that I downloaded from WindSurf that I have ran on a couple of Pentiums along with an PPC 8260 board. (In all cases, I used the END driver type of device). "drdiags" wrote in message news:a0mvcl$p77$1@newstest.laserlink.net... > Hello sardine, > > > "sardine" wrote in message > news:b1cb4c43.0112272242.132edcb7@posting.google.com... > > hi all gurus, > > > > TornadoII、VxWorks5.4 T2CP3 W2KPatck > > double Intel82557 ethernet controller on x86 target. > <.....SNIP.....> > > > > II. > > have any ways to receive ethernet packet directly and send it to > > IP layor? > <......SNIP.....> > > If you are not using DHCP<....SNIP...> > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Making a .lib file Date: Sun, 30 Dec 2001 15:24:35 +0800 From: "Chris Klein" Organization: Newshosting.com - Highest quality at a great price! www.newshosting.com Message-ID: <3c2ec29e@MAIL.mhogaming.com> References: <705c7b34.0111210630.7e8b496a@posting.google.com> There is no .lib file for Vxworks, only .a "Tal" wrote in message news:705c7b34.0111210630.7e8b496a@posting.google.com... > Hi, > > I need your help with making a .lib file in Tornado IDE (with a GNU compiler). > > TIA, > > Tal --------------------------- Newsgroups: comp.os.psos,comp.os.vxworks Subject: Detecting unused variables in Diab Data C compiler Date: Sun, 30 Dec 2001 10:56:39 +0200 From: Alex Vinokur Message-ID: <3C2ED6C6.336BC905@bigfoot.com> Followup-To: comp.os.psos ============================ pRISM+ for pSOSystem/PowerPC pRISM+ 1.2.3 . pSOSystem 2.2.2 Diab Data D-CC Optimizing C compiler SNMP Release 7.4 (Protocol SNMPv1) Host : Windows98 Target : PowerPC (MPC860) ============================ Does 'Diab Data D-CC Optimizing C compiler' have any option that detects unused variables and/or functions ? =========================== Alex Vinokur mailto:alexvn@bigfoot.com mailto:alexvn@dr.com http://up.to/alexvn http://go.to/alexv_math =========================== --------------------------- Newsgroups: comp.os.vxworks Subject: fei82559 hangs on tx Date: Sun, 30 Dec 2001 09:18:42 GMT From: "Iosif Harutyunov" Organization: AT&T Broadband Message-ID: Leonid, I have similar problem with FEI BSD driver (82559ER on x86 platform), when network driver stop transmitting frames because of lack of free CFD buffers. Under some condition, presumably heavy load, something happens during feiTxStartup executions and as a result one of CFD buffers remains marked as busy (its status is not CFD_S_COMPLETE). After while number of "busy" CFD grows and at some point transmission stalls. Could you please share with me about problem you had in past with BSD FEI driver. Iosif,_ > I had a similar problem with the FEI BSD driver two years ago. > If I remember correctly, you dont touch the transmit chain while the Tx > is active, but you queue up the new packets in a new Command list, which > is made active once the MAC is done with the current list. > > I can probably dig up my diffs, but I am not sure how applicable > these are since I did this on teh BSD and you are uding the END drivers. > > - Leonid > "BT" wrote in message > news:3c0cdc78.605875953@news.eunet.no... > > We have an PPC750 with two Intel82559ER network controllers. The > > driver is fei82559End. > > > > With high load on one of the networks the driver stops transmitting. > > The other network is OK. I can use WindView etc. for debugging. > > > > I have found that this happens when the SCB status word indicates > > CU status = LPQ active. If I ensures that I do NOT write anything > > to the CFD's while the SCB status/CU status is not in suspend the > > driver does not hang on tx anymore. > > > > I do not like this solution as I have to trhow away network packets > > before they are given to the controller. And since I have free CFDs > > I should be able to buffer up more packets. > > > > Has anybody come across this problem? Or even better give me a fix? > > > > BT --------------------------- Newsgroups: comp.os.vxworks Subject: which network devices supported by VxWorks? Date: Sun, 30 Dec 2001 13:27:50 +0200 From: Amit Raam Organization: Verio Message-ID: <3eut2ukj5la3b374ha4fppboifbcnhrd49@4ax.com> Hi, Does anyone know of a list of officially supported network devices (chips) supported by VxWorks under the XScale architecture? Is there such a list on the wrs site? Thanks, Amit Raam Email address: qwerty_at_actcom_dot_co_dot_il - --------------------------------------------------- ".sigless and proud of it" (R) (C) (tm) (look-and-feel) --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Mon Dec 31 04:03:14 2001 From: Vxworks Exploder Date: Mon Dec 31 04:03:16 PST 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Mon Dec 31 04:03:10 PST 2001 Subject: Re: will the driver get the write(fd, NULL, n) call? Subject: IfAddrAdd() of second IP address from same subnet Subject: Re: etherInputHookAdd ? Subject: Help: About the netClusterGet function Subject: Why the target server can't connect with target agent use wdb serial connect.? Subject: Using select(), need to monitor the serial port continuously in the thread Subject: Re: Networking in VxSim Subject: Libraries for MPC8260 Subject: Help: How the tNetTask processing the jobs added to it? Subject: Re: will the driver get the write(fd, NULL, n) call? Subject: Re: WRS are you here? Subject: Re: parallel port driver Subject: Re: error while downloading Subject: Tor2.0.2 and vpath Subject: Re: pci error in IDT Rc32332 Subject: Re: curses Subject: Re: which network devices supported by VxWorks? Subject: Re: Help: How the tNetTask processing the jobs added to it? Subject: Re: Tor2.0.2 and vpath Subject: mib2, ifTbale question ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: will the driver get the write(fd, NULL, n) call? Date: 30 Dec 2001 16:33:15 -0800 From: mgalecki@sbs.com (Mark Galecki) Organization: http://groups.google.com/ Message-ID: <585c4698.0112301633.23783471@posting.google.com> References: <3c2d37df$0$25219$724ebb72@reader2.ash.ops.us.uu.net> <488e459a.0112282223.229a816a@posting.google.com> john_94501@yahoo.com (John) wrote in message news:<488e459a.0112282223.229a816a@posting.google.com>... > If you mean is it guaranteed on current versions of the OS, then the > answer would seem to be yes. There doesn't seem to be any check in the > generic I/O code on the buffer or the length - they are just passed to > the driver's entry point. > > If you mean is this behaviour guaranteed for future; I would guess not > since it is not documented anywhere that it will be supported. > I meant "both" - I want to write software for the current version of VxWorks Operating System, but surely I want my software to be compatible with the future versions as well. From your reply, the answer is really "no to both" because "would seem to be" is a good observation, but cannot be used in professionally designed software - that's why I need "guarantees" :). I was hoping some WindRiver engineers would come up and say "yes this is what we intend in the design now and for ever, and sorry we did not bother to put it in the manual". I am still hoping for that :) Thank you Mark --------------------------- Newsgroups: comp.os.vxworks Subject: IfAddrAdd() of second IP address from same subnet Date: 30 Dec 2001 16:49:35 -0800 From: telly_o@softhome.net (Yirmi Ruahzmani) Organization: http://groups.google.com/ Message-ID: <3d44a0.0112301649.6d057954@posting.google.com> I understand that I'll get an error if I use IfAddrAdd() to add a second IP address from the same subnet to an interface. I've heard different answers to what this means: a) it's not really an error and the alias is successful b) The VxWorks interface is buggy and it's better to use the underlying Berkeley interface. Does anyone have any experience on this? Thanks in advance. - - Yirmi --------------------------- Newsgroups: comp.os.vxworks Subject: Re: etherInputHookAdd ? Date: 30 Dec 2001 18:04:28 -0800 From: sardine_sz@163.net (sardine) Organization: http://groups.google.com/ Message-ID: References: "drdiags" wrote in message news:... > Forgot to mention, I am assuming you are using an END driver and not a BSD > driver. Also, > that should have been if you haven't added a DHCP "Server" to your image. It > uses the > MUX_SNARF protocol for an END device. Also, WDB should be set to use > network and not END as its access technique (you lose System Level > debugging) if you choose yes, i tested,if i choosed network connection as WDB. i can receive etherpacket right. > to add a MUX_SNARF routine. I also thought the vxsniff.c example available > on the vxWorks > FAQ uses etherInputHookAdd. I had a vxWorks 5.3.1 ether input/output hook > example that I > downloaded from WindSurf that I have ran on a couple of Pentiums along with > an PPC 8260 > board. (In all cases, I used the END driver type of device). i have another question ? if i want to dispose etherpacket directly, i have three ways: 1>modify ehternet driver program ..... 2>by etherInputHook function.......... 3>MUX function........ and what's way have high effectivity ? --------------------------- Newsgroups: comp.os.vxworks Subject: Help: About the netClusterGet function Date: 30 Dec 2001 19:03:04 -0800 From: wxh@founder.com.cn (Wen Xiaohui) Organization: http://groups.google.com/ Message-ID: <8fc342d0.0112301903.7414b1a7@posting.google.com> I'm writing an END driver using DP83902A from National Semiconductor.What amazed me is that when I call the netClusterGet() in the dp83902Recv I will get a wrong but not NULL pointer( the result is always a value composed by the first four bytes of my MAC address). This phenomena will happen when I test the driver in a very heavy LAN and after the driver had run all right for a while. Does anyone know what's wrong with my driver? Thanks and with Best Regards, Wen Xiaohui --------------------------- Newsgroups: comp.os.vxworks Subject: Why the target server can't connect with target agent use wdb serial connect.? Date: 30 Dec 2001 19:33:42 -0800 From: jlddotcn@hotmail.com (juliedan) Organization: http://groups.google.com/ Message-ID: test --------------------------- Newsgroups: comp.os.vxworks Subject: Using select(), need to monitor the serial port continuously in the thread Date: 30 Dec 2001 21:17:17 -0800 From: manibala@covansys.com (B.Manivannan) Organization: http://groups.google.com/ Message-ID: <23de5426.0112302117.12e9fe30@posting.google.com> Good Day Viewers, 1. I am developing an application which should continuously monitor the serial port inside the thread for any data arrival. 2. In a application, i just created a thread. Inside a thread, i am opening a port. After that FOREVER loop has started. In side the forever loop, i have given selecet statement as per the example given by the help. It goes inside the FOREVER loop and got hanged.. I Hope the error is in the line, if( FD_ISSET( nOpenReturn, &readFDs ) ) I have no idea, why it fails.. 3.I am using Tornado II and VxWorks 5.4. 4. The Source code, i have generated is as follows /*My Source Code Starts Here....*/ #include "VxWorks.h" #include "fcntl.h" #include "selectLib.h" #include "ioLib.h" #include "taskLib.h" #include "sigLib.h" #include "stdio.h" #include "string.h" #define STACK_SIZE 20000 #define MAX_FDS 2 #define MAX_DATA 1024 /* Thread Identifier */ int nReadProcess; /* File Descriptor */ int nOpenReturn; /* Thread to monitor and read the serial port continuously */ void ReadingPort(void); /* Application gets initialized here */ void Read(void) { /* Task or thread is being started */ nReadProcess = taskSpawn ("Reading",200, 0, STACK_SIZE, (FUNCPTR) ReadingPort,0,0,0,0,0,0,0,0,0,0); } void ReadingPort(void) { struct fd_set readFDs; int width, i; char buffer[MAX_DATA]; /* Open File Descriptors */ nOpenReturn = open("/tyCo/0",2,0); nStatus1 = ioctl( nOpenReturn, FIOBAUDRATE, 9600 ); nStatus2 = ioctl( nOpenReturn, FIOSETOPTIONS, OPT_RAW); /* Loop for ever for reading data from the serial port */ FOREVER { /* Clear bits in read bit mask */ FD_ZERO(&readFDs); /* Initialise bit masks */ FD_SET( nOpenReturn, &readFDs); /* pend, waiting for the file descriptor to become ready */ if( select( nOpenReturn, &readFDs, NULL, NULL, NULL ) == ERROR ) return ( ERROR ); /* read from the file descriptor it it gets ready */ if( FD_ISSET( nOpenReturn, &readFDs ) ) { read(nOpenReturn, buffer, MAX_DATA); printf("The data read from the serial port is %s", buffer); } } } /*My Source Code Ends Here....*/ 5. Can you suggest me, what is the problem with this code? If it is not a right way, what is an alternate path to read the message from the serial port while data occurs at any time. Can you give me some sample code over here?? With Thanks and Regards, Manivannan.B. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Networking in VxSim Date: 30 Dec 2001 21:27:43 -0800 From: rashanmu@npd.hcltech.com (Ramesh) Organization: http://groups.google.com/ Message-ID: <15a87607.0112302127.4e61169a@posting.google.com> References: <15a87607.0112282319.5f582c94@posting.google.com> <2a6118da.0112290814.48658d6e@posting.google.com> Hi Raj, Thanks for the response. But what am in need of is, from the VxWorks target IP (90.0.0.1) i want to ping another vxworks image ( 90.0.0.2) in another host. remember i want to ping the other image in another host from the target shell of 90.0.0.1. According to your note i can ping 90.0.0.2 from the NT host. This is working for me already. But i want to ping from the target shell. If this works, then only i can proceed with my programs. Kindly help me in this regard. Thanks Ramesh raj_rashid@hotmail.com (Rashid) wrote in message news:<2a6118da.0112290814.48658d6e@posting.google.com>... > Hi Ramesh, > > > > My PC ethernet adapter: 192.168.100.69 ( Win NT ) > > My ULIP adapter: 90.0.0.254 > > My VxWorks target iamge : 90.0.0.1 > > On this host do, > > C:\> route ADD 90.0.0.2 192.168.100.144 > > Now u will be able to ping 90.0.0.2 on other host. > > > > Other end ethernet address : 192.168.100.114 ( Win 2k) > > Other end ULIP address : 90.0.0.253 > > Other end VxWorks Target image : 90.0.0.2 > > Best regards, > > Rashid --------------------------- Newsgroups: comp.os.vxworks,comp.sys.powerpc.tech Subject: Libraries for MPC8260 Date: 30 Dec 2001 22:26:24 -0800 From: raj_rashid@hotmail.com (Rashid) Organization: http://groups.google.com/ Message-ID: <2a6118da.0112302226.83b6acd@posting.google.com> Hi all, I am working on Tornado-II for PowerPC on Solaris. Which GNU tool should i use for MPC8260? In my Tornado we are having PPC860gnu but no PPC8260gnu. Also the communication interface libraries like motFccEnd.o are for MPC860, can i use these libraries for MPC8260 as well? Thanks, Best Regards, Rashid --------------------------- Newsgroups: comp.os.vxworks Subject: Help: How the tNetTask processing the jobs added to it? Date: 30 Dec 2001 23:11:24 -0800 From: wxh@founder.com.cn (Wen Xiaohui) Organization: http://groups.google.com/ Message-ID: <8fc342d0.0112302311.3e784fa@posting.google.com> I'm writing an END driver now. I want to know how the tNetTask processing its jobs? If one of the jobs is pended waiting for a semaphore, does tNetTask also pended or does it call the next job queued ? In my driver, I want to use DMA to receive the packet from the DP83902's local SRAM using a semaphore to synchronize the operation of the DMA, so it's very important that the tNetTask don't call the next job when a job is pended waiting for the DMA_end semaphore. Thanks and with Best Regards, Wen Xiaohui --------------------------- Newsgroups: comp.os.vxworks Subject: Re: will the driver get the write(fd, NULL, n) call? Date: 30 Dec 2001 23:25:33 -0800 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0112302325.3e7371f7@posting.google.com> References: <3c2d37df$0$25219$724ebb72@reader2.ash.ops.us.uu.net> <488e459a.0112282223.229a816a@posting.google.com> <585c4698.0112301633.23783471@posting.google.com> Hello, > I meant "both" - I want to write software for the current version of > VxWorks Operating System, but surely I want my software to be > compatible with the future versions as well. From your reply, the > answer is really "no to both" because "would seem to be" is a good > observation, but cannot be used in professionally designed software - > that's why I need "guarantees" :). It was also an observsation made with the advantage of having the relevant source code to look at; so probably as close to a guarantee as you are likely to get from a public forum like this... > I was hoping some WindRiver > engineers would come up and say "yes this is what we intend in the > design now and for ever, and sorry we did not bother to put it in the > manual". I am still hoping for that :) If you want that, then try contacting Wind River directly (your local sales person and/or FAE would be a good starting point). I doubt very much that you will get a "guarantee" of the type you are looking for from any Wind River employee via comp.os.vxworks - it is the wrong forum for that; the best we can do is provide the wisdom of our collective experience. Sorry that's not good enough for your needs, John... > Thank you > Mark --------------------------- Newsgroups: comp.os.vxworks Subject: Re: WRS are you here? Date: 31 Dec 2001 09:44:26 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: <3b9f1aed.0112280823.390475fc@posting.google.com> spurcell@apl.washington.edu (Steve) wrote: > It seems like there are a lot of VxWorks users providing a lot of free > advice and help. > > Where doe's WRS stand in all this. > > Seems like they should play a leading role here rather than using > silence as a way of milking money out of people for support. I did not see any official message from any WRS employee in this newsgroup, but I know from discussions with FAE's that they are monitoring this newsgroup. One thing I know (from emails I have seen) is that WRS tech support is sometimes pointing to the FAQ as a reply to questions. And once in a while I get messages from WRS tech support with some information that I can use in the FAQ. 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: parallel port driver Date: 31 Dec 2001 09:48:56 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: mgarcia@psl.nmsu.edu (Michael Garcia) wrote: > We are using MVME2604 BSP and need a driver for the Parallel Port, any > help in obtaining some guidance would be greatly appreciated... > Mike Take a look at this, it might help: http://www.xs4all.nl/~borkhuis/vxworks/mv23xx.html#4 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: error while downloading Date: 31 Dec 2001 09:53:22 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: <6040814b.0112281121.7084b294@posting.google.com> sksinha2001@yahoo.com (Sandeep Sinha) wrote: > C:\downloads\server.c: In function `main': One name you are not allowed to use in VxWorks is the name "main", so rename your main function to something different, and try again. 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: Tor2.0.2 and vpath Date: Mon, 31 Dec 2001 12:05:32 +0200 From: "Ofer Goren" Organization: Verio Message-ID: How do I add vpath variable to my automaticly generated makefiles in Tornado2? THanks - -- Ofer Goren Iamba Technologies --------------------------- Newsgroups: comp.os.vxworks Subject: Re: pci error in IDT Rc32332 Date: 31 Dec 2001 10:09:19 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: <4ceedd43.0112282323.368465f8@posting.google.com> sczhang@esteemnetworks.com (ShiChao Zhang) wrote: > 1. When I try to use the PCI Configuration Data Register at > 1800_2cfc,Address Register at 1800_2cf8 to access the PCI > configuration space, any write operation is OK(each operation comprise > of a write access to address Reg and a write/read operation to data > Reg). But If I perform a read operation, it will generate a Bus Error > Exception. Then I tried to locate why the exception take place. And I > found that if I put a access to Address Reg followed by a printf > function call, then begin the normal read operation, it will be OK, no > exception will occur. > I can't understand it. Did you try to use the memory mapped version of the PCI access? You can also access the I/O space using address 0x188xxxxx, and you can access the memory space using addresses 0x40000000 .. 0x5fffffff and 0x60000000 .. 0x7fffffff. You can use different endianess for these areas, but they can point to the same PCI space. 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: curses Date: 31 Dec 2001 10:14:53 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: "Weddick" wrote: > Has anyone done a port or know of a port for using curses on a X86 > Vxworks System? You can find a VxWorks version of curses here: ftp://ftp.atd.ucar.edu/pub/archive/vxworks/vx/ It is called vwcurses, and they are packed in a "shar" archive. 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: which network devices supported by VxWorks? Date: 31 Dec 2001 10:15:57 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: <3eut2ukj5la3b374ha4fppboifbcnhrd49@4ax.com> Amit Raam wrote: > Hi, > Does anyone know of a list of officially supported network devices > (chips) supported by VxWorks under the XScale architecture? Is there > such a list on the wrs site? You have to ask WindRiver about 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.vxworks Subject: Re: Help: How the tNetTask processing the jobs added to it? Date: 31 Dec 2001 10:21:48 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: <8fc342d0.0112302311.3e784fa@posting.google.com> wxh@founder.com.cn (Wen Xiaohui) wrote: > I'm writing an END driver now. I want to know how the tNetTask > processing its jobs? If one of the jobs is pended waiting for a > semaphore, does tNetTask also pended or does it call the next job > queued ? In my driver, I want to use DMA to receive the packet from > the DP83902's local SRAM using a semaphore to synchronize the > operation of the DMA, so it's very important that the tNetTask don't > call the next job when a job is pended waiting for the DMA_end > semaphore. The netTask is only receiving messages send by netJobAdd, and executes the function pointer that is given in this message. If this function hangs, the netTask hangs. If you want to implement something like you mention you should create a seperate DMA task with a message queue and add commands to this queue. From netTask you will add a command to this queue and netTask will continue. The DMA task can schedule all DMA requests after each other. This will also avoid multiple parallel accesses to the DMA engine. 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: Tor2.0.2 and vpath Date: 31 Dec 2001 10:26:03 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: "Ofer Goren" wrote: > How do I add vpath variable to my automaticly generated makefiles in > Tornado2? http://www.xs4all.nl/~borkhuis/vxworks/vxw_pt1.html#1.6-A 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: mib2, ifTbale question Date: Mon, 31 Dec 2001 12:51:48 +0200 From: "Ofer Goren" Organization: Verio Message-ID: - ------------------------ Win2k Sp2 Tornado 2.02 for PPC Diab 4.4b Envoy 9.2 - ------------------------ Hi. I have a little question. On my device, there are 7 ports, only 1 is an IP-port. My mib2 ifTable should return 7 when asking for 'ifNumber'. So, I go to my snmpif.c file, under my epilogue installation. I see that interfaceGet() calls m2IfGoupInfoGet() to retrieve the data. However, I don't have the source code of this function, so how can I change the results it gives? I can go directly and change interfaceInfoGet() (same file), and return 7 instead of 'pIfData->ifNumber' but I dont think this is the solution.... - -- Ofer Goren Iamba Technologies --------------------------- End of New-News digest **********************