From vxwexplo-errs@csg.lbl.gov Tue Aug 1 02:17:15 2000 From: "Palaniappan" Date: Tue Aug 1 02:17:17 PDT 2000 Subject: Re: TrueFFS mountFTL Hi, We too had the same problem in building the TrueFFS system. Basically, VxWorks deals with three types of Flash Translation Layers, They are NOR-FLASH, NAND-FLASH and SDFDC FLASH. For all these three type of flash translations layers, there will be a .o file which is archieved into PPC860gnuvx.a for each translation layer. For example, for a NAND-FLASH, there will be a .o file called "nftllite.o" where as nftl for "NAND Flash" and lite stands for "the file is from M-Systems FLite". If you see the documentation on TrueFFS, you can see that the "TrueFFS for Tornado is a VxWorks-compatible implementation of M-Systems FLite, Version 2.0" in the introduction itself. Likewise, if there will be a file called "ftllite.o" for NOR-FLASH Translation Layers. TrueFFS's special feature says that it can transact with the flash technology using wear-level algorithm, erase-after-write algorithms. To provide these special algorithms, M-Systems FLite provides a flash translation layer files, which could possibly be a licensed version from WRS, or otherwise you could approach M-Systems FLite. I had a thorough search on M-Systems FLite web site and found no downloads for NOR-FLASH Arrays. For the sake of understanding the problem, we just extracted the ftllite.o file from PPCCEC603gnuvx.a and added into PPC860gnuvx.a and the vxWorks image was successfully built, but when we attempt to format the TrueFFS, it returns the value -1. And hence, I concluded that we may need to know whether it's a licensed version or not. Any help would be appreciated. ARP Newsgroups: comp.os.vxworks Subject: TrueFFS mountFTL Date: Sun, 30 Jul 2000 01:30:32 GMT From: fywang@my-deja.com Organization: Deja.com - Before you buy. Message-ID: <8m00fo$6on$1@nnrp1.deja.com> Hi all, I am trying to build a TrueFFS on my 2 AM29LV017D chips. Since AM29LV017D is a NOR-based flash, I should define INCLUDE_TL_FTL in my sysTffs.c file. In target/src/drv/tffs/tffsConfig.c: TLentry tlTable[] = /* TL tables */ { ... #ifdef INCLUDE_TL_FTL #ifdef FORMAT_VOLUME {mountFTL, formatFTL}, #else mountFTL, #endif /* FORMAT_VOLUME */ #endif /* INCLUDE_TL_FTL */ ... } I got a compile error when building my BSP: ... C:\Tornado\target\proj\gold860 \default\C:\Tornado\target\src\drv\intrCtl\ppc860Intr.c:224: undefined reference to `mountFTL' C:\Tornado\target\proj\gold860 \default\C:\Tornado\target\src\drv\intrCtl\ppc860Intr.c:224: undefined reference to `formatFTL' make: *** [vxWorks] Error 0x1 Searching "mountFTL" shows it's referenced by h/tffs/stdcomp.h & src/drv/tffs, no definition. The only occurence in lib is libPPCEC603gnuvx.a, but I am working on PPC860. Is there something wrong with my Tornado2/VxWorks? Can I work around this problem? Regards, Tom Wang Sent via Deja.com http://www.deja.com/ Before you buy. From vxwexplo-errs@csg.lbl.gov Tue Aug 1 04:03:20 2000 From: Vxworks Exploder Date: Tue Aug 1 04:03:23 PDT 2000 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Tue Aug 1 04:03:16 PDT 2000 Subject: Re: Changing the value of LOCAL_MEM_SIZE from 0x0100 0000 to 0x0200 0000 results in Boot-up failure Subject: Re: FAQs are where? Subject: intercommunication between processors Subject: RE: verbosity macro Subject: Command line compiling in Tornado 2 Subject: Re: Raw IP Subject: MVME2431 Boot problem Subject: Re: Changing the value of LOCAL_MEM_SIZE from 0x0100 0000 to 0x0200 0000 Subject: VMEbus reset on boot? Subject: Re: sntpclib, localtime() and TIMEZONE Subject: vxWorks Route Table Subject: ftpLib Subject: Looking for help with MPCxxx and NEC SDRAM Subject: Re: VMEbus reset on boot? Subject: Re: Tornado Shell: d = 123 Subject: mbufs and sockets Subject: Re: Tornado Shell: d = 123 Subject: Re: sntpclib, localtime() and TIMEZONE Subject: Re: MVME2431 Boot problem Subject: Re: Web Server under VxWorks Subject: Re: Tornado Shell: d = 123 Subject: Re: JWorks and JNI Subject: Re: Command line compiling in Tornado 2 Subject: TCP Dump for VxWorks? Subject: File truncation. Subject: Re: VMEbus reset on boot? Subject: Re: VMEbus reset on boot? Subject: WIN2000 SP1 & TORNADO II Subject: Re: VxWorks ROM image and application. Subject: Linked lists Subject: Loading new application Subject: PCI Configuration Subject: Re: PCI Configuration Subject: Re: PCI Configuration Subject: raw keyboard input to Tornado shell process? Subject: Re: Changing the value of LOCAL_MEM_SIZE from 0x0100 0000 to 0x0200 0000 results in Boot-up failure Subject: PMCspan expansion board setup Subject: Re: VxWorks & ARM Dev Board Subject: Re: Tornado Shell: d = 123 Subject: Re: Tornado Shell: d = 123 Subject: Re: Tornado Shell: d = 123 Subject: Re: Tornado Shell: d = 123 Subject: Compressed Image Subject: Re: mbufs and sockets Subject: Re: problem using FEC on 860T Subject: Re: Linked lists Subject: Re: 64-bit division Subject: Re: Bothered by DHCP Subject: Can I pass the structures through the Message Queues! Subject: Wind River support experience? Subject: re: TrueFFS mountFTL Subject: TrueFFS mountFTL Subject: Configure BSP fails ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Changing the value of LOCAL_MEM_SIZE from 0x0100 0000 to 0x0200 0000 results in Boot-up failure Date: 31 Jul 2000 13:17:10 GMT From: Geir Frode Raanes Organization: UNINETT news service Message-ID: <8m3u8m$k93$1@snipp.uninett.no> References: <01BFFA56.92683EA0.tewarics@future.futsoft.com> Sender: Geir Frode Raanes Tewari C wrote: : Hi, : The DRAM used on the EBSA285 was 16MB. However our board as 32MB DRAM i.e. : 4 chips X 16bit wide each. Changes have been made in the rominit.s to : change the DRAM initialization for full 32MB DRAM and LOCAL_MEM_SIZE has : been redefined to 0x02000000. The Tornado project configuration (the hardware tab) overrides the default settings in the BSP source files. : When Bootrom is used to Boot-up the process hangs at the point before : loading the VxWorks image via FTP to the DRAM. That is because the bootrom is what configures the DRAM controller hardware, not the kernel. Consequently, you need to rebuild the bootrom with the same memory configuration you have informed the kernel readily available. - -- ****************************************************** Never ever underestimate the power of human stupidity. -Robert Anson Heinlein GeirFRS@invalid.and.so.forth ****************************************************** --------------------------- Newsgroups: comp.os.vxworks Subject: Re: FAQs are where? Date: Mon, 31 Jul 2000 15:24:35 +0200 From: "Martin Raabe" Organization: T-Online Message-ID: <8m3uqs$ksf$15$1@news.t-online.com> References: <8ls5o8$hjs$1@nnrp1.deja.com> <8lss1n$li32@nntp.cig.mot.com> <8lt0ka$7i2$1@nnrp1.deja.com> Hello Pat, there are also FAQs fo rsome tools. E.g. SNiFF+ has the FAQs located at: W: www.takefive.co.at/faq schrieb in im Newsbeitrag: 8lt0ka$7i2$1@nnrp1.deja.com... > > > In article <8lss1n$li32@nntp.cig.mot.com>, > > "Derek Opitz" wrote: > > > There are some FAQ pages at www.wrs.com under support, and then > > windsurf. > > Ahhh yes thank you. Notably: > http://www.wrs.com/csdocs/sortedfaqs/index.shtml > and: > > http://www.wrs.com/csdocs/sortedfaqs/summariesTools/htmlfiles/shell.html > explaining such things don't work as: > printf("\x41\n"); > > Anyone got more FAQs? > > > Sent via Deja.com http://www.deja.com/ > Before you buy. --------------------------- Newsgroups: comp.os.vxworks Subject: intercommunication between processors Date: Mon, 31 Jul 2000 13:36:08 GMT From: Carlos Prendes Organization: SAES Message-ID: <3985801D.4BA4D598@terra.es> I´d like to know the real utility of VxMP for helping in programming for intertasks comunication between processors. Anyone have experience in this subject and the different options for this issue? I am a newcomer and I´m now reading the manual and I want to decide if we should buy the library but this decission must be done without previous experience (nobody here has done a program in VxWorks). Any help would be greatly thanked. Sorry for my english and for the disturbs. Carlos Prendes --------------------------- Newsgroups: comp.os.vxworks Subject: RE: verbosity macro Date: Mon, 31 Jul 2000 15:21:40 +0200 From: Diego Novellon Organization: Deja.com - Before you buy Message-ID: <01BFFB03.072370A0.novellon@teltronic.es> > Subject: verbosity macro > From: Sebastien Fauris (sfauris@club-internet.fr) > > Hi folks, > > I have a macro very convenient to debug my code. > For that I compile my code in debug mode (flag DRV_DEBUG_MODE defined), > and I use the following macro to display my debug messages with a > certain level > of verbosity that I can change (by a global variable): > > Macro: > #ifdef DRV_DEBUG_MODE > #define VERB_DRV(level,a1,a2,a3,a4,a5,a6,a7) \ > if (level <=3D drv_dbg) \ > {logMsg(a1,a2,a3,a4,a5,a6,a7);} > #else > #define VERB_DRV(level,a1,a2,a3,a4,a5,a6,a7) {} > #endif > > and use it in my code: > > #ifdef DRV_DEBUG_MODE > int drv_dbg =3D 0; /* default 0, but can be changed to a higher verbosit= y > level after */ > #endif > > int toto (void) > { > > VERB_DRV (2,"my debug message displayed only in debug mode when > drv_dbg global variable set higher or equal to > 2\n",0,0,0,0,0,0); > } > > I want to do the same macro, but using printf instead of logMsg. > How can I do to pass the arguments to the printf (variable number of > arguments) > while logMsg arguments is fixed to 6 and this in a macro definition ? > > Thanks a lot, > > Sebastien. Hi Sebastien, I think there's some info on writing macros with a variable number of=20 arguments in the GNU Toolkit User's Guide. IIRC, it's something like #define MY_PRINTF(format, args...) printf(format , ## args) I'm not totally sure, but it's worth doing a text search of=20 docs/gnu/toolkit/gcc.html. Regards, =09Diego Novell=F3n Sent via Deja.com http://www.deja.com/ Before you buy. --------------------------- Newsgroups: comp.os.vxworks Subject: Command line compiling in Tornado 2 Date: Mon, 31 Jul 2000 09:48:30 -0400 From: "Joe Frankovich" Organization: Naval Research Laboratory, Washington, DC Message-ID: <8m403e$kh5$1@ra.nrl.navy.mil> Hi, I have Tornado 2.0, and a Blue Wave System Quad TMS320C6701 board with a MPC860 on board processor. I am using a Sun Solaris 8 workstation as my host environment. Blue Wave has provided me with a BSP, but has informed me that I cannot use the project facility. This means I must do everything at the command line. I am a Windows programmer and until recently have never seen a makefile. Can anyone give me any guidance as to command line methods? I've taken the WRS training course on Tornado 2.0 and feel very comfortable with the tools as they are very similar to the visual environments I used with windows, but I am lost at the command line. Thanks, Joe --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Raw IP Date: 31 Jul 2000 14:13:41 GMT From: cedilnik andrej Organization: University of Maryland, Baltimore County Message-ID: <8m41il$ccf$1@news.umbc.edu> References: <8la2v8$k4v$1@news.umbc.edu> Edt wrote: > You should be able to open a raw socket and receive all the IP packets for a > particular address family. The etherHookAdd function mentioned in another > message will get ethernet formatted packets from a particular interface, but > not from any interface like a raw socket will, and the ethernet header will > still be in the packet data. Ok, here is a code: cm->ifd = socket(AF_INET, SOCK_RAW, IPPROTO_RAW); if (cm->ifd < 0){ printf("Errno: %d\n", errno); return FALSE; } while(1){ printf("Wait for package\n"); size= read_packet(cm->ifd, buff, MAX_BUFF); printf("Got package\n"); } } This haings on line size= read_packet(cm->ifd, buff, MAX_BUFF); btw, read_packet: int read_packet(int ifd, unsigned char *buf, int size){ return recv(ifd, buf, size, 0); } So what am I doing wrong? Andy --------------------------- Newsgroups: comp.os.vxworks Subject: MVME2431 Boot problem Date: Mon, 31 Jul 2000 09:26:30 -0500 From: "Ken Parrish" Organization: WebUseNet Corp http://www.usenetserver.com - Home of the fastest NNTP servers on the Net. Message-ID: <5_fh5.9609$mv.341518@news-east.usenetserver.com> Can anyone give me all the default usernames, passwords, IP addresses, filenames etc. that vxworks uses as default values in their BSP's and operating system software. I compiled the BSP image, used the PPC4-Bug commands niop to tftp it to the MVME2431 board and used pflash to load the image in rom. When I power down the board, change the jumper J8 to pins 1 and 2 and power up the board I get the [VxWorks Boot]: prompt, however when I type @ and return I get this message: attached TCP/IP interface to dc0. attaching network interface lo0 ... done. Loading ... 0x1fbb260 (tNetTask): dc0 - restarting device. attached TCP/IP interface to dc0. attaching network interface lo0 ... done. Loading ... 0x1fbb260 (tNetTask): dc0 - restarting device. and it continues to repeat. Does anyone have any ideas, I'm very open to suggestions now? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Changing the value of LOCAL_MEM_SIZE from 0x0100 0000 to 0x0200 0000 results in Boot-up failure Date: Mon, 31 Jul 2000 10:23:15 -0400 From: Al Johnston Organization: MindSpring Enterprises Message-ID: <39858BD3.654B8F6D@mindspring.com> References: <01BFFA56.92683EA0.tewarics@future.futsoft.com> <8m3u8m$k93$1@snipp.uninett.no> > hardware, not the kernel. Consequently, you need to rebuild the > bootrom with the same memory configuration you have informed the Sorry I did not make this clear. I did rebuild the bootroom, and flashed it into eeprom. --------------------------- Newsgroups: comp.os.vxworks Subject: VMEbus reset on boot? Date: Mon, 31 Jul 2000 09:54:44 -0500 From: "Mark P. Fishman" Message-ID: We would like to have all cards on the VME bus reset whenever we (re)boot vxWorks. Is there a way to do this in software? either with a target-shell command from a startup script or from a vxWorks function call? Thanks for any assistance -- Mark F. - -- reply to: mfishman at el el dot em eye tee dot ee dee ewe OPINIONS EXPRESSED ARE MY OWN AND DO NOT REFLECT THOSE OF MY EMPLOYER OR ANY OTHER ENTITY IN THE KNOWN WORLD. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: sntpclib, localtime() and TIMEZONE Date: Mon, 31 Jul 2000 10:43:44 -0400 From: Al Johnston Organization: MindSpring Enterprises Message-ID: <398590A0.B4B5B07D@mindspring.com> References: <3983D15C.1233E21E@mindspring.com> <8m20l4$fto$1@nnrp1.deja.com> Looked at you "bad" method.. checked __loctime... the strings it points to dont look right, so i think i will do this the "supported" way... My initial objection (besides building vxworks with a capability I dont otherwise need) was that the envlib might require building with the internal shell as well. thanks. - -al --------------------------- Newsgroups: comp.os.vxworks Subject: vxWorks Route Table Date: Mon, 31 Jul 2000 07:37:57 -0700 From: mpiuze Organization: http://www.remarq.com: The World's Usenet/Discussions Start Here Message-ID: <0221c012.41e189e2@usw-ex0108-062.remarq.com> Hi, I'm trying to include the capability in our vxWorks system of changing IP addresses on the fly after the image is loaded. I have been able so far to stop an interface and reload it with a new IP, but I need a way to delete all the routes on a given interface. This product is not intended to have routing functionality, so RIP and OSPF are not being used. The only way I can find right now is to use routeDelete(), but routeDelete does not allow the deletion of more than one entry with each call, and I see no route function that allows me to specify the interface. There must be a way to associate interfaces with the routes, if for no other reason than that routeShow() does it, ( plus, a route table is pretty useless without it... :) I also would need to find a way to walk through the route table, but I cannot find a global symbol for this table. Is the table a static variable for the routeLib module? I know that is the way the endList the muxLib module uses does things, so I gather it must be similar. Stepping through the routeShow() function, I see that eventually rn_walktree () is called, which makes an indirect function call to rtEntryPrint(). Beyond that I haven't been able to decipher much on how the interfaces are associated in the route entries. Thanks for the help, Matt * Sent from AltaVista http://www.altavista.com Where you can also find related Web Pages, Images, Audios, Videos, News, and Shopping. Smart is Beautiful --------------------------- Newsgroups: comp.os.vxworks Subject: ftpLib Date: 31 Jul 2000 18:26:03 +0200 From: aviad Message-ID: <39859a8b$1@neutron.ecitele.com> Hi, We use ftpLib and ftpdLib to download files between CPUs working with vxWorks. There are possible situation of disconect, ( for example: the server side CPU may have reset). at such situation the client side is hang-up (sometimes it is waiting forever on semaphore). We are looking for a way to detect such situation and abort the ftp command at the client side. Is there a configurable (or not) timeout in vxWorks for the ftp commands ? thanks in advance, aviad. --------------------------- Newsgroups: comp.arch.embedded,comp.os.vxworks,comp.sys.powerpc.tech Subject: Looking for help with MPCxxx and NEC SDRAM Date: Mon, 31 Jul 2000 16:58:30 +0100 From: "Jeremy" Organization: Customer of Energis Squared Message-ID: <8m47mq$63o$1@newsreaderg1.core.theplanet.net> Has anyone any experience of SDRAM (NEC 4564163) interfacing to MPC8xx. This is very important. Any information will be very gratefully recieved. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VMEbus reset on boot? Date: Mon, 31 Jul 2000 16:43:31 GMT From: "news1.van.metronet.ca" Organization: MetroNet Communications Group Inc. Message-ID: References: Mark: We added a "sysReset" command to our BSP. By typing the sysReset command at the shell, it would cause a reset to all VME boards on the backplain. Besides that, there is no other reset command that I know of. Eric Jennings "Mark P. Fishman" wrote in message news:Pine.LNX.4.05.10007310952260.15350-100000@shmulin.llan.ll.mit.edu... > > We would like to have all cards on the VME bus reset whenever we > (re)boot vxWorks. Is there a way to do this in software? either > with a target-shell command from a startup script or from a > vxWorks function call? > > Thanks for any assistance -- Mark F. > > > -- > reply to: mfishman at el el dot em eye tee dot ee dee ewe > > OPINIONS EXPRESSED ARE MY OWN AND DO NOT REFLECT THOSE OF > MY EMPLOYER OR ANY OTHER ENTITY IN THE KNOWN WORLD. > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado Shell: d = 123 Date: Mon, 31 Jul 2000 16:53:57 GMT From: p_lavarre@my-deja.com Organization: Deja.com - Before you buy. Message-ID: <8m4av4$48g$1@nnrp1.deja.com> References: <8ls5l2$ham$1@nnrp1.deja.com> <3981AF4D.216FABDD@swri.edu> <8lsl2v$ubu$1@nnrp1.deja.com> <8lsmd4$msm@niven.ksc.nasa.gov> <8lsrgb$3gg$1@nnrp1.deja.com> <8lv4la$o7j$1@news1.xs4all.nl> > In article <8lv4la$o7j$1@news1.xs4all.nl>, > johan.borkhuis@peektraffic.nl wrote: > > > > > ... WDB Agent across serial ... > > > > : I just tried downloading an app that included the C lines: > > : extern int d = 123; > > : Downloading such an app makes the @ notation work great. > > : @d = 567 > Yes, that is right, but now you redefined the symbol d in your > application. You cannot use the function d anymore from within > the shell. Ouch, ouch, ouch. I thought this was NOT true? Am I completely failing to understand? There are TWO separate and independent symbol tables involved, yes? The "d" of the app is in the app table, visible to `lkup`. The "d" of the shell is hidden somewhere else? For example, after I download: extern int d = 123; I find here I can, for example, enter the commands: lkup "^d$" @d = 0x12345678 d &@d,0,1 In the last command, I'm using both "d" and "@d", simultaneously. That command "d"umps the memory at the app symbol "d". My CPU is Motorola- endian, so I see: 0014a800: 12 34 56 78 ... 0014a810: ... ... x0014A800 is just an example here - for this run of this experiment, x0014A800 happens to be the "data" address than `lkup "^d$` reports as the address of app symbol "d". Thanks again. Pat LaVarre p.lavarre@ieee.org http://members.aol.com/ppaatt Sent via Deja.com http://www.deja.com/ Before you buy. --------------------------- Newsgroups: comp.os.vxworks Subject: mbufs and sockets Date: Mon, 31 Jul 2000 13:34:30 -0400 From: Dmitriy Zavin Message-ID: <3985B8A6.E2CF02B5@txc.com> Hi. Anyone know how I can pass data to a socket that is allready in an mbuf?
I don't want to copy it out of an mbuf chain into a flat buffer just so the socket layer will copy that buffer into a new chain.. I want to skip that whole process and just hand an open socket some data in an allready preallocated mbuf chain.

Thanks,

--Dima
  --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado Shell: d = 123 Date: Mon, 31 Jul 2000 18:36:10 GMT From: p_lavarre@my-deja.com Organization: Deja.com - Before you buy. Message-ID: <8m4gup$9dq$1@nnrp1.deja.com> References: <8ls5l2$ham$1@nnrp1.deja.com> <3981AF4D.216FABDD@swri.edu> <8lsl2v$ubu$1@nnrp1.deja.com> <8lsmd4$msm@niven.ksc.nasa.gov> <8lsrgb$3gg$1@nnrp1.deja.com> > > ... add ... symbols of my choice to the table that `lkup` shows? I > > was surprised to find that a form like: > > @d = 890 > > does not work ... > > Do I really have to edit, compile, and download an app, > In article , > pkockritz@home.com (Pete Kockritz) wrote: > symAdd() ... in the vxWorks manuals for symLib ... Ah, fun, thank you. I regret to say I find the symbols I add via symAdd are NOT visible to `lkup` ... nor can I use them on the command line? I guess I'm off to read up the TCL portion of the Tornado Shell manual - somebody tell me if I'll be reading the wrong stuff? > I regret to say I find the symbols I add via symAdd are NOT visible Searching for symAdd in: file:///C:/Tornado/docs/vxworks/rtnIndex.html leads me to: file:///C:/Tornado/docs/vxworks/ref/symLib.html `lkup "SymT"` tells me I have a "sysSymTbl" and "statSymTbl" - resident in the "bss" of my target. After downloading some helper code, symEach lets me dump the content of both - including the "type" and "group" of each symbol. My "statSymTbl" is empty, but my "sysSymTbl" has xB50 entries, including "sysSymTbl" and "statSymTbl". Using "statSymTbl" as an example, I find I am able to symAdd other names. What I add is visible to symEach ... ... but what I add is NOT visible to `lkup`. For example: - -> /* allocate some working storage */ - -> ii = 123 new symbol "ii" added to symbol table. ii = 0x14a820: value = 123 = 0x7b = '{' - -> jj = 123 new symbol "jj" added to symbol table. jj = 0x14a810: value = 123 = 0x7b = '{' - -> /* add */ - -> /* 7 == N_DATA | EXT by target/h/a.out_h */ - -> symAdd(sysSymTbl, "mySym", &ii, 7, 0) value = 0 = 0x0 - -> /* confirm added */ - -> symFindByName(sysSymTbl,"mySym", &ii, &jj) value = 0 = 0x0 - -> ii ii = 0x14a820: value = 1353760 = 0x14a820 = ii - -> (char) jj jj = 0x14a810: value = 7 = 0x7 - -> /* confirm added invisibly */ - -> mySym undefined symbol: mySym - -> lkup "mySym" value = 0 = 0x0 More later, thanks again. Pat LaVarre p.lavarre@ieee.org http://members.aol.com/ppaatt/ Sent via Deja.com http://www.deja.com/ Before you buy. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: sntpclib, localtime() and TIMEZONE Date: Mon, 31 Jul 2000 19:04:57 GMT From: john_94501@my-deja.com Organization: Deja.com - Before you buy. Message-ID: <8m4ikk$are$1@nnrp1.deja.com> References: <3983D15C.1233E21E@mindspring.com> <8m20l4$fto$1@nnrp1.deja.com> <398590A0.B4B5B07D@mindspring.com> Hi, In article <398590A0.B4B5B07D@mindspring.com>, Al Johnston wrote: > Looked at you "bad" method.. checked __loctime... the strings it points > to > dont look right, so i think i will do this the "supported" way... They are in fact logical, but set up for UTC. I can send you the structure in source form if you want to see it... > My > initial > objection (besides building vxworks with a capability I dont otherwise > need) > was that the envlib might require building with the internal shell as > well. There aren't many things that need the target shell. It is in fact a pretty independent component. The reverse is not true though, the target shell needs quite a lot of support components to do its thing. TTFN, John... Sent via Deja.com http://www.deja.com/ Before you buy. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: MVME2431 Boot problem Date: Mon, 31 Jul 2000 19:08:29 GMT From: john_94501@my-deja.com Organization: Deja.com - Before you buy. Message-ID: <8m4ir8$av5$1@nnrp1.deja.com> References: <5_fh5.9609$mv.341518@news-east.usenetserver.com> Hello, Section 2.6 of the Tornado User's guide will explain what all the bootrom options are. You need to use the 'c' command at the boot prompt to set the IP addresses and boot method. HTH, John... In article <5_fh5.9609$mv.341518@news-east.usenetserver.com>, "Ken Parrish" wrote: > Can anyone give me all the default usernames, passwords, IP addresses, > filenames etc. that vxworks uses as default values in their BSP's and > operating system software. > > I compiled the BSP image, used the PPC4-Bug commands niop to tftp it to the > MVME2431 board and used pflash to load the image in rom. When I power down > the board, change the jumper J8 to pins 1 and 2 and power up the board I get > the [VxWorks Boot]: prompt, however when I type @ and return I get this > message: > > attached TCP/IP interface to dc0. > attaching network interface lo0 ... done. > > Loading ... 0x1fbb260 (tNetTask): dc0 - restarting device. > > attached TCP/IP interface to dc0. > attaching network interface lo0 ... done. > > Loading ... 0x1fbb260 (tNetTask): dc0 - restarting device. > > and it continues to repeat. Does anyone have any ideas, I'm very open > to suggestions now? > > Sent via Deja.com http://www.deja.com/ Before you buy. --------------------------- Newsgroups: comp.os.vxworks,comp.realtime Subject: Re: Web Server under VxWorks Date: Mon, 31 Jul 2000 13:12:58 -0500 From: John McClenny Organization: Posted via Supernews, http://www.supernews.com Message-ID: References: <397D1FD2.2D51DB48@yahoo.com> <3980A162.29461E3E@odn.de> In article <3980A162.29461E3E@odn.de>, ensslin@odn.de says... > Shelly HUANG wrote: > > > > Hi, > > > > Is any web server running under VxWorks? > > > > thank you for any information. > > > > best regards, > > Shelly > > Check out http://www.quiotix.com/products.html for a pretty full featured embedded server. john --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado Shell: d = 123 Date: Mon, 31 Jul 2000 19:00:36 GMT From: john_94501@my-deja.com Organization: Deja.com - Before you buy. Message-ID: <8m4icg$ai1$1@nnrp1.deja.com> References: <8ls5l2$ham$1@nnrp1.deja.com> <3981AF4D.216FABDD@swri.edu> <8lsl2v$ubu$1@nnrp1.deja.com> <8lsmd4$msm@niven.ksc.nasa.gov> <8lsrgb$3gg$1@nnrp1.deja.com> <8lv4la$o7j$1@news1.xs4all.nl> <8m4av4$48g$1@nnrp1.deja.com> Hi > Ouch, ouch, ouch. I thought this was NOT true? Am I completely > failing to understand? There are TWO separate and independent symbol > tables involved, yes? The "d" of the app is in the app table, visible > to `lkup`. The "d" of the shell is hidden somewhere else? There are not really two symbol tables involved here. You have to remember there's some history to the wind shell. It is meant to behave as closely as possible to the original target based shell. On the target shell there is one symbol table, and if you were to do lkup "d" you would see the function d() that performs the memory dump (assuming you had added that feature to the kernel). Now, the goal of the wind shell was to minimise the impact on the target's memory footprint by moving these sorts of functions to the host. They are implemented by the host shell and therefore are not shown by lkup which is still telling you about symbols on the target. The parser however is the same. When you type d at the target shell prompt the host shell interprets it as a reference to the function d(). Were you to be using the target shell, typing d=123 would give you an error since the shell knows 'd' refers to a function. The host shell therefore behaves the same way. If it sees 'd' it assumes you mean the pseudo-function d(); if you are trying to assign a value to it, then it declares that an error. If you are using the target based tools, and download a module onto the target that has a global symbol 'd' then there will be two entries for 'd' in the target's symbol table. You will probably find that the d() function no longer works. As you have found, the host shell overcomes this by letting you prefix the symbol with '@' when you want it to use the target's symbol table and not check against its builtin functions, but it doesn't allow you to create new symbols this way. That's just the way it is. Also, bear in mind that the shell is a development tool. You are not intended to create lots of global variables from it, merely have the ability to test out function calls in your code quickly. The price you pay for this feature is that some symbols are already taken by the tools. I find that double characters are more likely to work than singles so I use things like 'dd', 'xx' and 'yy'. Hope that makes things a little clearer, John... > For example, after I download: > extern int d = 123; > I find here I can, for example, enter the commands: > lkup "^d$" > @d = 0x12345678 > d &@d,0,1 > > In the last command, I'm using both "d" and "@d", simultaneously. That > command "d"umps the memory at the app symbol "d". My CPU is Motorola- > endian, so I see: > 0014a800: 12 34 56 78 ... > 0014a810: ... > ... > x0014A800 is just an example here - for this run of this experiment, > x0014A800 happens to be the "data" address than `lkup "^d$` reports as > the address of app symbol "d". > > Thanks again. Pat LaVarre p.lavarre@ieee.org > http://members.aol.com/ppaatt > > Sent via Deja.com http://www.deja.com/ > Before you buy. > Sent via Deja.com http://www.deja.com/ Before you buy. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: JWorks and JNI Date: Mon, 31 Jul 2000 13:34:43 -0500 From: Gretchen Patti Organization: ASIC Designs, Inc. Message-ID: <3985C6C3.4E06FA1F@adichi.com> References: <397C6BDC.9F552364@adichi.com> <39803829.8050D85E@irisa.fr> Gilbert Cabillic wrote: > > Gretchen Patti wrote: > > > > Is anybody using JNI under Personal JWorks? > > > > I'm creating a Java application for an embedded system under vxWorks. > > We're using Personal JWorks 3.0. I'm trying to use JNI to call a > > native C routine. The System.loadLibrary call works OK, but when I > > call the native method I get an UnsatisfiedLinkError. Help!! > > > > Hi. > > Jni calls on JWorks are supported and we had no problems to use them. > I can recommend you to test a basic 'hello world' Java application using > a jni call for printing on the screen to verify that the configuration > of Jworks is OK. > > Gilbert. Gilbert -- The problem was indeed in my configuration, but what a wild, complex configuration it is!! Once I finally located "symbol table -- initialization -- select", I changed it to "Built-In", and then everything worked fine. Thanks for your feedback! :) - -- Gretchen ======================================================================== Sent or forwarded by Gretchen Patti gretchen@adichi.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Command line compiling in Tornado 2 Date: Mon, 31 Jul 2000 12:35:06 -0700 From: Eric Monsler Organization: The ISP formerly known as Best Message-ID: <3985D4EA.BC079DD1@radixtek.com> References: <8m403e$kh5$1@ra.nrl.navy.mil> Joe Frankovich wrote: (snip) > I am a Windows programmer and until recently have > never seen a makefile. Can anyone give me any > guidance as to command line methods? (snip) Get a book on make, or else use the GNU make manual, available online: http://www.delorie.com/gnu/docs/make/make_toc.html In fact, you can also look at the gcc manual, and most or all of the other GNU manuals, from: http://www.delorie.com/gnu/docs/ If you've really never used a command line at all, you may want to get a book describing the shell you are using (csh, bash, etc.). You may even find that you like the command line, although learning a really powerful editor like Emacs which interoperates with the command line is even better. Eric --------------------------- Newsgroups: comp.os.vxworks Subject: TCP Dump for VxWorks? Date: 31 Jul 2000 19:40:46 GMT From: cedilnik andrej Organization: University of Maryland, Baltimore County Message-ID: <8m4knu$dih$1@news.umbc.edu> Is there any simple way of performing packet snooping in VxWorks? I tried socket(AF_INET, SOCK_RAW, IPPROTO_RAW) and it does not read anything. If I say socket(AF_INET, SOCK_RAW, xxx), where xxx is some number that system does not know, it works. How would I get all packages no meter what they are? Andy --------------------------- Newsgroups: comp.os.vxworks Subject: File truncation. Date: Mon, 31 Jul 2000 16:10:37 -0400 From: viv Organization: Lucent Technologies Message-ID: <3985DD3D.BB52887@lucent.com> How can I truncate a file if I know the FILE pointer? Isn't there a facility like in Unix (trunc, truncate ) to truncate a file?

Thanks,
viv --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VMEbus reset on boot? Date: Mon, 31 Jul 2000 10:36:15 -0500 From: Bob Irwin Organization: SIS Development Message-ID: <39859CEF.2CD97821@systemsinterface.com> References: "Mark P. Fishman" wrote: > > We would like to have all cards on the VME bus reset whenever we > (re)boot vxWorks. Is there a way to do this in software? either > with a target-shell command from a startup script or from a > vxWorks function call? This is very BSP-dependent. For example in the Synergy VGMD bsp I'm using, a call to sysReset() generates the VME bus reset signal. search thru your bsp code and use the lkup target shell command for clues: example: dstore1-> lkup "Reset" scsiTargetReset 0x000a174c text (vxWorks.sym) pciAutoDevReset 0x00030368 text (vxWorks.sym) sysReset 0x000403c4 text (vxWorks.sym) <<<<----- - -- Bob Irwin | bobi@sis-dev.com | SIS Development | 281-484-6661 | Houston --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VMEbus reset on boot? Date: Mon, 31 Jul 2000 20:41:34 GMT From: lee.deraud@boeing.com (Lee DeRaud) Organization: None to speak of Message-ID: <3985e2b5.28408359@news.boeing.com> References: Sender: nntp@news.boeing.com (Boeing NNTP News Access) On Mon, 31 Jul 2000 09:54:44 -0500, "Mark P. Fishman" wrote: >We would like to have all cards on the VME bus reset whenever we >(re)boot vxWorks. Is there a way to do this in software? either >with a target-shell command from a startup script or from a >vxWorks function call? This sounds a bit backwards: usually you're (re)booting VxWorks as a *result* of a VME bus reset, not the other way around. Check the documentation for the VME interface chipset used by your processor: most of them have a way to strobe the SYSRESET signal. Lee --------------------------- Newsgroups: comp.os.vxworks Subject: WIN2000 SP1 & TORNADO II Date: Mon, 31 Jul 2000 16:37:50 -0500 From: "Michael Murphy" Organization: Imaging Business Machines, LLC Message-ID: Installed Windows 2000 SP1 today. When running Tornado II I get a bunch of the following errors: "RPCExit: extra call" Tornado II seems to run fine after I accept all these errors (4 to 5 of them) Anybody have any clues? Thanx, Michael --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks ROM image and application. Date: Mon, 31 Jul 2000 20:51:02 GMT From: ddinker@my-deja.com Organization: Deja.com - Before you buy. Message-ID: <8m4ork$fpt$1@nnrp1.deja.com> References: <397E43D4.3445446D@lucent.com> <8lpdif$g2h$1@nnrp1.deja.com> <39819165.77A03D37@lucent.com> > How do you decide on the size of your mymodule.o? From where do you start > the VxWorks image? do you use fpromread and fpromwrite to the flash? > The size of mymodule.o is fixed as it is a piece of asmebly code. The VxWorks imgae start from a #defined value. We download the images through the emulator one after the other. mymodule.o is run first and after doinga thing or two it sets some addresses (start of ROM and RRAM) and hands control to VxWorks image. HTH, Dinker Sent via Deja.com http://www.deja.com/ Before you buy. --------------------------- Newsgroups: comp.os.vxworks Subject: Linked lists Date: Mon, 31 Jul 2000 18:01:52 -0400 From: viv Organization: Lucent Technologies Message-ID: <3985F750.8D4956BB@lucent.com> Has any one developed a wrapper class for the lstLib library of managing doubly linked lists?

Thanks,
viv --------------------------- Newsgroups: comp.os.vxworks Subject: Loading new application Date: Mon, 31 Jul 2000 21:26:35 GMT From: "Joseph Sebastian" Organization: Lawrence Berkeley National Laboratory Message-ID: <8m4tp6$lt0$1@overload.lbl.gov> Hi all, I like to incorporate the feature to upgrade my vxWorks application software in the system. I want to load a latest version of my software and replace the existing one, the command for this will be given from an SNMP manager. The module to be loaded is kept in a TFTP server, download into the RAM and do a version checking on it. If the module is acceptable the software should be loaded from this location. Any help for this will be well appreciated. Joseph KS. ________________________________________________________________________ Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com --------------------------- Newsgroups: comp.lang.c-programming,comp.os.vxworks,comp.sys.powerpc,comp.sys.powerpc.tech Subject: PCI Configuration Date: Mon, 31 Jul 2000 20:35:07 GMT From: "Rob Targosz" Organization: Excite@Home - The Leader in Broadband Message-ID: <%plh5.237979$7o1.6352532@news2.rdc1.on.home.com> I am trying to initialize a PCI video card under VxWorks on a PowerPC 604 architecture. So far, I have initialized the configuration header BARs and Command register as shown below. Unfortunately, after configuration, I am unable to write to memory using sysOutXXXX(). Is this the proper way to do this? What am I missing? Any help is appreciated! Thanks Rob Targosz robtargosz@hotmail.com #includes <...> #define PCI_CACHE_LINE_SIZE 0x20 #define PCI_LATENCY_TIMER 0xff #define PCI_CMD_MASK 0xffff0000 STATUS v3(void) { PCI_HEADER_DEVICE headerDevice; PCI_HEADER_DEVICE *pD = &headerDevice; UINT32 v3Base0 = NULL; UINT32 v3Base1 = NULL; UINT32 v3Base2 = NULL; int vendorId = VENDOR_ID, deviceId = DEVICE_ID, index = 0; int busNo, devNo, funcNo; /* Locate Device */ if (pciFindDevice (vendorId, deviceId, index, &busNo, &devNo, &funcNo) == ERROR) { DEBUGMSG(("Voodoo3 device not found!\n")); return ERROR; } /* Device Located - Disable */ pciConfigOutLong(busNo, devNo, funcNo, PCI_CFG_COMMAND, 0x0); /* Initialize and Configure BARs */ pciConfigOutLong(0,11,0,16,0xd2000000); pciConfigOutLong(0,11,0,20,0xd4000008); pciConfigOutLong(0,11,0,24,0xe0002001); /* BAR 0 - Memory Address */ pciConfigInLong(busNo, devNo, funcNo, PCI_CFG_BASE_ADDRESS_0, &v3Base0); v3Base0 = v3Base0 & ~0xF; /* BAR 0 - Memory Address */ pciConfigInLong(busNo, devNo, funcNo, PCI_CFG_BASE_ADDRESS_1, &v3Base1); v3Base1 = v3Base1 & ~0xF; /* BAR 0 - IO Address */ pciConfigInLong(busNo, devNo, funcNo, PCI_CFG_BASE_ADDRESS_2, &v3Base2); v3Base2 = v3Base2 & ~0xF; /* Configure PCI Registers */ pciConfigOutByte(busNo, devNo, funcNo, PCI_CFG_CACHE_LINE_SIZE, PCI_CACHE_LINE_SIZE); pciConfigOutByte(busNo, devNo, funcNo, PCI_CFG_LATENCY_TIMER, PCI_LATENCY_TIMER); pciConfigOutLong(busNo, devNo, funcNo, PCI_CFG_COMMAND, (PCI_CMD_MASK | PCI_CMD_MASTER_ENABLE | PCI_CMD_IO_ENABLE | PCI_CMD_MEM_ENABLE)); /* Send Voodoo3 Initialization Sequence */ sysOutLong(v3Base0 + 0x03C3, 0x001); /* This write doesn't work */ return OK; } --------------------------- Newsgroups: comp.lang.c-programming,comp.os.vxworks,comp.sys.powerpc,comp.sys.powerpc.tech Subject: Re: PCI Configuration Date: Mon, 31 Jul 2000 22:47:10 GMT From: John Fusco Organization: Deja.com - Before you buy. Message-ID: <8m4vld$l8a$1@nnrp1.deja.com> References: <%plh5.237979$7o1.6352532@news2.rdc1.on.home.com> Is the cache enabled? If so, the memory range of the BAT registers has to have page table entries (see sysPhysMemDesc[] in sysLib.c) or you need to program one or more of the BAT registers. There are comments in sysLib.c to show you how to do this. In article <%plh5.237979$7o1.6352532@news2.rdc1.on.home.com>, "Rob Targosz" wrote: > I am trying to initialize a PCI video card under VxWorks on a PowerPC 604 > architecture. So far, I have initialized the configuration header BARs and > Command register as shown below. Unfortunately, after configuration, I am > unable to write to memory using sysOutXXXX(). > > Is this the proper way to do this? What am I missing? > > Any help is appreciated! > > Thanks > Rob Targosz > robtargosz@hotmail.com > > #includes <...> > #define PCI_CACHE_LINE_SIZE 0x20 > #define PCI_LATENCY_TIMER 0xff > #define PCI_CMD_MASK 0xffff0000 > > STATUS v3(void) > { > PCI_HEADER_DEVICE headerDevice; > PCI_HEADER_DEVICE *pD = &headerDevice; > UINT32 v3Base0 = NULL; > UINT32 v3Base1 = NULL; > UINT32 v3Base2 = NULL; > > int vendorId = VENDOR_ID, deviceId = DEVICE_ID, index = 0; > int busNo, devNo, funcNo; > > /* Locate Device */ > if (pciFindDevice (vendorId, deviceId, index, &busNo, &devNo, &funcNo) == > ERROR) > { > DEBUGMSG(("Voodoo3 device not found!\n")); > return ERROR; > } > > /* Device Located - Disable */ > pciConfigOutLong(busNo, devNo, funcNo, PCI_CFG_COMMAND, 0x0); > > /* Initialize and Configure BARs */ > pciConfigOutLong(0,11,0,16,0xd2000000); > pciConfigOutLong(0,11,0,20,0xd4000008); > pciConfigOutLong(0,11,0,24,0xe0002001); > > /* BAR 0 - Memory Address */ > pciConfigInLong(busNo, devNo, funcNo, PCI_CFG_BASE_ADDRESS_0, &v3Base0); > v3Base0 = v3Base0 & ~0xF; > > /* BAR 0 - Memory Address */ > pciConfigInLong(busNo, devNo, funcNo, PCI_CFG_BASE_ADDRESS_1, &v3Base1); > v3Base1 = v3Base1 & ~0xF; > > /* BAR 0 - IO Address */ > pciConfigInLong(busNo, devNo, funcNo, PCI_CFG_BASE_ADDRESS_2, &v3Base2); > v3Base2 = v3Base2 & ~0xF; > > /* Configure PCI Registers */ > pciConfigOutByte(busNo, devNo, funcNo, PCI_CFG_CACHE_LINE_SIZE, > PCI_CACHE_LINE_SIZE); > pciConfigOutByte(busNo, devNo, funcNo, PCI_CFG_LATENCY_TIMER, > PCI_LATENCY_TIMER); > pciConfigOutLong(busNo, devNo, funcNo, PCI_CFG_COMMAND, (PCI_CMD_MASK | > PCI_CMD_MASTER_ENABLE | PCI_CMD_IO_ENABLE | PCI_CMD_MEM_ENABLE)); > > /* Send Voodoo3 Initialization Sequence */ > sysOutLong(v3Base0 + 0x03C3, 0x001); /* This write doesn't work */ > > return OK; > } > > Sent via Deja.com http://www.deja.com/ Before you buy. --------------------------- Newsgroups: comp.lang.c-programming,comp.os.vxworks,comp.sys.powerpc,comp.sys.powerpc.tech Subject: Re: PCI Configuration Date: Mon, 31 Jul 2000 22:50:30 GMT From: John Fusco Organization: Deja.com - Before you buy. Message-ID: <8m4vrk$lar$1@nnrp1.deja.com> References: <%plh5.237979$7o1.6352532@news2.rdc1.on.home.com> <8m4vld$l8a$1@nnrp1.deja.com> In article <8m4vld$l8a$1@nnrp1.deja.com>, John Fusco wrote: > Is the cache enabled? > > If so, the memory range of the BAT registers has to have page table Make that, "the memory range programmed into the base address registers" Sorry Sent via Deja.com http://www.deja.com/ Before you buy. --------------------------- Newsgroups: comp.os.vxworks Subject: raw keyboard input to Tornado shell process? Date: Tue, 01 Aug 2000 00:00:41 GMT From: p_lavarre@my-deja.com Organization: Deja.com - Before you buy. Message-ID: <8m53v7$oco$1@nnrp1.deja.com> How can a process run from the Tornado shell ask to get one keystroke at a time from the shell input device? I'm running `windsh` to talk to a target server, thence via wdb serial to my target itself ... I see I can run `windsh` from Tornado or from an MS-Dos Prompt. I see that turning on Ansi.sys at the MS-Dos Prompt enables 2D coloured character graphics, for example: printf("\033[2J"); clears the screen, \033[y;xH moves the cursor arbitrarily, etc. But if I try: getchar that routine doesn't return until I press Enter. How do I ask to turn the input echo & line-buffering off? I just want to see each keystroke. I went looking for - no joy. I went looking for raw() and cbreak() - no joy. Maybe I'm looking for some IOCTL? Or is it TCL code on the host that sees each keystroke - where do I go to find that code so I can patch it, say, to end the line after every keystroke? Thanks in advance. Pat LaVarre p.lavarre@ieee.org http://members.aol.com/ppaatt/ Sent via Deja.com http://www.deja.com/ Before you buy. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Changing the value of LOCAL_MEM_SIZE from 0x0100 0000 to 0x0200 0000 results in Boot-up failure Date: Tue, 01 Aug 2000 00:49:47 GMT From: Hosehead Jones 2nd-1/2 Organization: United Garage Mechanics and Fruitbats, Ltd. Message-ID: References: <01BFFA56.92683EA0.tewarics@future.futsoft.com> <8m3u8m$k93$1@snipp.uninett.no> <39858BD3.654B8F6D@mindspring.com> Al Johnston took time out of his/her busy schedule to pen: > >> hardware, not the kernel. Consequently, you need to rebuild the >> bootrom with the same memory configuration you have informed the > >Sorry I did not make this clear. I did rebuild the bootroom, and >flashed it into eeprom. I believe that LOCAL_MEM_SIZE is one of those variables that's defined in three separate places, and all three have to match or else lots of lovely errors result. (I don't have access to Tornado right now, so I can't double-check that this is the case.) One of them will be in config.h; the second will be in the Makefile for your BSP (in the Tornado\target\\config directory); the third will be in the project for the VxWorks kernel. If you switch to the "VxWorks configuration" tab of the workspace window and do a Find Object (right click on something), search for the variable LOCAL_MEM_SIZE; it should point you to where you need to change it. If it's not there, the other place it may be in the project is in the build settings. Again, I don't have my Tornado with me here, so I can't verify this. --------------------------- Newsgroups: comp.os.vxworks Subject: PMCspan expansion board setup Date: Mon, 31 Jul 2000 21:10:09 -0400 From: "Minbo Shim" Message-ID: Hi, I am trying to setup a PMC card in a PMC adapter carrier board from Motorola (PMCspan-002) which is attached to a MVME2431. The PMC card is in the slot 1 (device # 2) of the expansion board with interrupt level 3 (PMC_INT_LVL3). The DEC21150 PCI-to-PCI bridge chip provides the interface between the primary PCI bus and the secondary PCI bus through Type 0 and Type 1 configuration cycle. I am using CHRP Extended VME memory map for this configuration. My questions are: 1) can PMC_SPAN work with the CHRP memory model? 2) can I use some known pci setup functions like pciDevConfig(), pciConfigOutLong(), etc? 3) what is the physical address (hardware unmapped) of the PMC card as seen from the processor? 4) what is the virtual address (MMU mapped)? Any help/comments on this would be greatly appreciated. - -Minbo GDRS --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks & ARM Dev Board Date: Mon, 31 Jul 2000 18:17:14 -0700 From: Mohit Organization: http://www.remarq.com: The World's Usenet/Discussions Start Here Message-ID: <207a7325.b630c794@usw-ex0106-045.remarq.com> References: <03db9226.c358038d@usw-ex0105-038.remarq.com> <8lu0p4$tpb$1@nnrp1.deja.com> <1b8aeca0.856a27a7@usw-ex0105-038.remarq.com> <8m1s3d$cvh$1@nnrp1.deja.com> Hello John: Thanks a lot for your detailed reply. It sure got me oriented in the correct direction! After following your advice [and the manual], we've been able to get moving a wee bit more. [1] The VxWorks boot PROM is in place on the board. [2] The environment is set up to talk using Com2. Angel is out. [3] I tried to start up the target server to talk to the port using the serial BUT we ran into a license problem. Apparently the host ID of the server that runs the license manager has changed since we requested for the license :( So, we're contacting wrs again. [4] While we wait, I'm just wonderng if anyone knows the answer to this one question: does the VxWorks boot PROM have support for a link using the serial cable? Or do I need to make the boot image again? It's not like it's a problem... but I'm just wondering. Thanks a lot, Mohit. - ----------------------------------------------------------- Got questions? Get answers over the phone at Keen.com. Up to 100 minutes free! http://www.keen.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado Shell: d = 123 Date: Tue, 01 Aug 2000 03:41:48 GMT From: "Paddy" Organization: Time Warner Cable of San Diego, CA Message-ID: <0Grh5.2496$wn.36645@typhoon.san.rr.com> References: <8ls5l2$ham$1@nnrp1.deja.com> <3981AF4D.216FABDD@swri.edu> <8lsl2v$ubu$1@nnrp1.deja.com> <8lss52$li33@nntp.cig.mot.com> <8lt1jn$8b4$1@nnrp1.deja.com> <39822B25.5B6ED9AC@bellsouth.net> Try helloThere=malloc(4) to add a 32 bit integer to the symbol table. Patrick "Joe Durusau" wrote in message news:39822B25.5B6ED9AC@bellsouth.net... > Hmmmm... I've never heard of such a thing, but > you could always write a program that declared global > symbols of your choice and load the thing. That should do it. > Don't know of a way to get at it directly from the shell, though. > > Speaking only for myself, > > Joe Durusau > > > p_lavarre@my-deja.com wrote: > > > > Isn't there some syntax - some call - I can use to just add a symbol of > > my choice to the `lkup` table? > > > > > In article <8lss52$li33@nntp.cig.mot.com>, > > > "Derek Opitz" wrote: > > > you can declare any variable you want at the shell, > > > as long as it is not used already. d() is a function > > > that dumps memory. So, therefore, you > > > can't use d as a variable name. HELLO!!!!!!! > > > > What can I say. Certainly I don't mean to be slow & obscure. > > > > There are TWO symbol tables involved here - yes? One table visible to > > "lkup". And a separate, second table hidden away in the Shell, but > > visible to Ctrl+D and Ctrl+W. For example, if I try: > > lkup "d" > > lkup "tt" > > I get no hits - even though "d" and "tt" are Shell commands. Part of > > how I know they are Shell commands is that typing them alone on a > > command line, then striking the spacebar, then pressing Ctrl+D gives me > > a synopsis of their purpose & arguments. > > > > I understand that because "d" is a Shell command, an entry like: > > d = 123 > > is considered to mean something like: > > d(= 123); > > and therefore reported as a: > > syntax error > > > > What's got me is that the commands: > > @d = 123 > > @d > > sometimes work and sometimes don't. > > > > If I have a C program loaded that declared: > > extern int d = 456; > > then these commands do just plain work. > > > > Otherwise, entering these two commands always causes the Shell to > > report: > > new symbol "@d" added to symbol table > > > > If I review the effect of these commands via: > > lkup "@d" > > then I can confirm that I'm creating a whole series of symbols with the > > same name but pointing to different locations. Naturally enough, > > creating an "@d" symbol make syntaxes like: > > @@d = 0x12345678 > > @@d > > d &@@d > > work, keeping in mind that each of these only refer to the very last > > definition created for the symbol "@d". > > > > What I don't get is why it isn't easy to make > > @d = 123 > > @d > > work as if I had loaded a C program to declare: > > extern int d = 456; > > > > Isn't there some syntax - some call - I can use to just add a symbol of > > my choice to the `lkup` table? > > > > Thanks again in advance. Pat LaVarre p.lavarre@ieee.org > > http://members.aol.com/ppaatt > > > > Sent via Deja.com http://www.deja.com/ > > Before you buy. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado Shell: d = 123 Date: Tue, 01 Aug 2000 04:17:03 GMT From: pkockritz@home.com (Pete Kockritz) Organization: @Home Network Message-ID: References: <8ls5l2$ham$1@nnrp1.deja.com> <3981AF4D.216FABDD@swri.edu> <8lsl2v$ubu$1@nnrp1.deja.com> <8lsmd4$msm@niven.ksc.nasa.gov> <8lsrgb$3gg$1@nnrp1.deja.com> <8m4gup$9dq$1@nnrp1.deja.com> In article <8m4gup$9dq$1@nnrp1.deja.com>, p_lavarre@my-deja.com wrote: > > I regret to say I find the symbols I add via symAdd are NOT visible Because the symbols added with symAdd are added to the target symbol table and not in the symbol table on the host. (Yes, there really are two symbol tables, or at least, there _can_ be two symbol tables. Acutally, there can be any number of symbol tables on the target, but only one system symbol table). The host symbol table contains the symbols that are in the 'core' file (vxWorks file) that is known by the target server and any symbols added by doing a 'ld' in the target shell or loaded by the Tornado download command. There is a BSP configuration option to allow the two symbol tables to be synchronized. There are some issues with that have been discussed in other threads in this newsgroup. I guess it depends on what you're trying to do with the symbols that you want to add to the symbol table. If you really must have a 'd' in your application, you could exclude usrLib from your vxWorks build. Then you should be able to have a target symbol 'd' and execute it with '@d' (assuming it is a function). Not sure what you could do with it if it is just a variable. From the windsh man page: > Built-in Routines > > The Tornado shell includes a set of built-in routines. These > routines are executed on the host, not in the context of the remote > run-time. They are available from both the C interpreter and the Tcl > interpreter. > The most important built-in routines are i( ), ti( ), d( ), l( ), > ts( ), tr( ), td( ), ld( ). When you type a command in the shell, the shell first checks to see if it is a built-in routine. If not, it attempts to execute it as a target function (I assume by looking it up in the host resident symbol table to see if it is present). The '@' is provided to force the function to be executed on the target, even if it is present as a shell command. But I assume (perhaps a large assumption) that the windsh would use the symbol table on the host to get the address of the target function to be evaluated. So if you had a function named 'd' that was part of the vxWorks kernel, and then did a symAdd to override the target symbol table's value of 'd', the windsh would still use the value of 'd' that was in the host's symbol table. (Confusing). Regards, Pete - -- +-----------------------------------------------------------------+ | Pete Kockritz mailto:pkockritz@home.com | | This space for rent. | +-----------------------------------------------------------------+ --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado Shell: d = 123 Date: Tue, 01 Aug 2000 04:30:53 GMT From: pkockritz@home.com (Pete Kockritz) Organization: @Home Network Message-ID: References: <8ls5l2$ham$1@nnrp1.deja.com> <3981AF4D.216FABDD@swri.edu> <8lsl2v$ubu$1@nnrp1.deja.com> <8lsmd4$msm@niven.ksc.nasa.gov> <8lsrgb$3gg$1@nnrp1.deja.com> <8lv4la$o7j$1@news1.xs4all.nl> <8m4av4$48g$1@nnrp1.deja.com> <8m4icg$ai1$1@nnrp1.deja.com> In article <8m4icg$ai1$1@nnrp1.deja.com>, john_94501@my-deja.com wrote: > There are not really two symbol tables involved here. You have to > remember there's some history to the wind shell. It is meant to behave > as closely as possible to the original target based shell. On the target > shell there is one symbol table, and if you were to do lkup "d" you > would see the function d() that performs the memory dump (assuming you > had added that feature to the kernel). If there are not two symbol tables, why does this exist? > symSyncLib > > NAME > symSyncLib - host/target symbol table synchronization > ROUTINES > symSyncLibInit( ) - initialize host/target symbol table synchronization > symSyncTimeoutSet( ) - set WTX timeout > DESCRIPTION > This module provides host/target symbol table synchronization. With > synchronization, every module or symbol added to the run-time > system from either the target or host side can be seen by facilities > on both the target and the host. Symbol-table > synchronization makes it possible to use host tools to debug > application modules loaded with the target loader or from a target > file system. Regards, Pete - -- +-----------------------------------------------------------------+ | Pete Kockritz mailto:pkockritz@home.com | | This space for rent. | +-----------------------------------------------------------------+ --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado Shell: d = 123 Date: Tue, 01 Aug 2000 05:28:02 GMT From: john_94501@my-deja.com Organization: Deja.com - Before you buy. Message-ID: <8m5n50$5l9$1@nnrp1.deja.com> References: <8ls5l2$ham$1@nnrp1.deja.com> <3981AF4D.216FABDD@swri.edu> <8lsl2v$ubu$1@nnrp1.deja.com> <8lsmd4$msm@niven.ksc.nasa.gov> <8lsrgb$3gg$1@nnrp1.deja.com> <8lv4la$o7j$1@news1.xs4all.nl> <8m4av4$48g$1@nnrp1.deja.com> <8m4icg$ai1$1@nnrp1.deja.com> Hi, In article , pkockritz@home.com (Pete Kockritz) wrote: > In article <8m4icg$ai1$1@nnrp1.deja.com>, john_94501@my-deja.com wrote: > > There are not really two symbol tables involved here. You have to > > remember there's some history to the wind shell. It is meant to behave > > as closely as possible to the original target based shell. On the target > > shell there is one symbol table, and if you were to do lkup "d" you > > would see the function d() that performs the memory dump (assuming you > > had added that feature to the kernel). > > If there are not two symbol tables, why does this exist? There are only two symbol tables when you have *BOTH* host and target tools running (e.g. you install the target shell, but still use the host based shell). This is not the same issue as the host shell not accepting d() because it is a reserved symbol. TTFN, John... > > symSyncLib > > > > NAME > > symSyncLib - host/target symbol table synchronization > > ROUTINES > > symSyncLibInit( ) - initialize host/target symbol table synchronization > > symSyncTimeoutSet( ) - set WTX timeout > > DESCRIPTION > > This module provides host/target symbol table synchronization. With > > synchronization, every module or symbol added to the run-time > > system from either the target or host side can be seen by facilities > > on both the target and the host. Symbol-table > > synchronization makes it possible to use host tools to debug > > application modules loaded with the target loader or from a target > > file system. > > Regards, > Pete > -- > +-----------------------------------------------------------------+ > | Pete Kockritz mailto:pkockritz@home.com | > | This space for rent. | > +-----------------------------------------------------------------+ > Sent via Deja.com http://www.deja.com/ Before you buy. --------------------------- Newsgroups: comp.os.vxworks Subject: Compressed Image Date: Mon, 31 Jul 2000 21:42:41 GMT From: ddinker@my-deja.com Organization: Deja.com - Before you buy. Message-ID: <8m4rsg$icc$1@nnrp1.deja.com> Hi, Am back with the question again. This time more general. Has anyone compressed a VxWorks image using the vxworks.st_rom rule? We need to download the image to ROM (compresssion will make the download fast) which will on reboot inflate it self and copy to RAM and start to run. Any leads will be helpful. Regards, Dinker Sent via Deja.com http://www.deja.com/ Before you buy. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: mbufs and sockets Date: Tue, 01 Aug 2000 09:11:27 +0200 From: Leonid Rosenboim Organization: BitBand Technologies Ltd. http://www.bitband.com Message-ID: <3986781F.7AB07480@bitband.com> References: <3985B8A6.E2CF02B5@txc.com> Reply-To: leonid@bitband.com Dimka, youy should spend some time reading the fine manual about Zero-Copy TCP, or ZBuf API, it is exactly what you are looking for. Dmitriy Zavin wrote: > Hi. Anyone know how I can pass data to a socket that is allready in an > mbuf? > I don't want to copy it out of an mbuf chain into a flat buffer just > so the socket layer will copy that buffer into a new chain.. I want to > skip that whole process and just hand an open socket some data in an > allready preallocated mbuf chain. > > Thanks, > > --Dima > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: problem using FEC on 860T Date: Mon, 31 Jul 2000 21:56:34 GMT From: mahmood@alum.rpi.edu Organization: Deja.com - Before you buy. Message-ID: <8m4smg$iu8$1@nnrp1.deja.com> References: <964519072.363104@newsmaster-04.atnet.at> config.h add this after line 56 : #define FADS_860T The "motfec" is initialized but I had other problems after that. The windriver customer support sucks. I am still looking. Let me know if this works (or how your worked). In article <964519072.363104@newsmaster-04.atnet.at>, "Martin Walser" wrote: > Hi, > > does anyone know how to initialise the MotFec for MPC860T using tornado2. > I don't know which components are necessary. > I tried to do this, but i received an exception and the system seems to > crash. > > Thanks > > Martin > > Sent via Deja.com http://www.deja.com/ Before you buy. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Linked lists Date: Tue, 01 Aug 2000 09:26:55 +0200 From: Daniel Schnell Organization: Siemens AG Message-ID: <39867BBF.20CBB6AF@icn.siemens.de> References: <3985F750.8D4956BB@lucent.com> If you want to use c++ then you can use the stl list template, if memory size is not a problem. Daniel. viv wrote: > > Has any one developed a wrapper class for the lstLib library of > managing doubly linked lists? > > Thanks, > viv --------------------------- Newsgroups: comp.os.vxworks Subject: Re: 64-bit division Date: Tue, 01 Aug 2000 01:23:20 -0700 From: hruesing Organization: http://www.remarq.com: The World's Usenet/Discussions Start Here Message-ID: <0eacb984.41606c9b@usw-ex0107-055.remarq.com> References: The original GNU-Compiler does support the data type "long long" which is on my machine (i486) 64 bit. May be it works for your system - ----------------------------------------------------------- Got questions? Get answers over the phone at Keen.com. Up to 100 minutes free! http://www.keen.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Bothered by DHCP Date: Tue, 01 Aug 2000 08:38:13 GMT From: ellin_lin@my-deja.com Organization: Deja.com - Before you buy. Message-ID: <8m629l$di4$1@nnrp1.deja.com> References: <8lhhgl$l6o$1@nnrp1.deja.com> <3985168C.B68B169F@lucent.com> In article <3985168C.B68B169F@lucent.com>, Yusuf Motiwala wrote: > DHCP server IP pool is also defined in usrNetDhcpsCfg.c. So probably you > want to make sure that it picks up from the correct file. > > Yusuf > > ellin_lin@my-deja.com wrote: > > > I'm a tenderfoot of vxWorks and Tornado. And I > > met some problems when I tried to configured a > > DHCP server. I modified usrnetwork.c setting the > > IP range as written in user manual, and test this > > DHCP server with a DHCP client of Win98, it did > > nothing. Can someone give me some guidance? > > > > Sent via Deja.com http://www.deja.com/ > > Before you buy. > > Sent via Deja.com http://www.deja.com/ Before you buy. --------------------------- Newsgroups: comp.os.vxworks Subject: Can I pass the structures through the Message Queues! Date: Tue, 1 Aug 2000 11:13:55 -0500 From: "Keith McHugh" Organization: Esat Net Customer Message-ID: <3986a137.0@seat.s3two.ie> Reply-To: "Keith McHugh" --------------------------- Newsgroups: comp.os.vxworks Subject: Wind River support experience? Date: Tue, 01 Aug 2000 03:04:34 -0700 From: Nick Hoogland Organization: http://www.remarq.com: The World's Usenet/Discussions Start Here Message-ID: <03009a08.5b942430@usw-ex0105-036.remarq.com> Hi, Who can share his/her experience with Wind River support (both positive and negative). For reasons of evaluation of the suitebility of the VxWorks, also related to available tools and cost, I am trying to get some answers on basic questions as foot prints and boot time (information that should have been readily available within Wind River) for three weeks now. I posted my questions at the Wind River site and mailed them directly to (sales) contact persons of Wind River. The answers I received consisted of glossary's that only contained some very global VxWorks and Tools information and which did not address any of my answers even remotely. Selecting a company with such poor pre-sales support for something as essential as an OS is questionable, so how is the support after being committed to Wind River? Thanks, Nick - ----------------------------------------------------------- Got questions? Get answers over the phone at Keen.com. Up to 100 minutes free! http://www.keen.com --------------------------- Newsgroups: comp.os.vxworks Subject: re: TrueFFS mountFTL Date: Tue, 1 Aug 2000 14:47:04 +0530 From: "Palaniappan" Organization: Lawrence Berkeley National Laboratory Message-ID: <8m67v8$734$1@overload.lbl.gov> Hi, We too had the same problem in building the TrueFFS system. Basically, VxWorks deals with three types of Flash Translation Layers, They are NOR-FLASH, NAND-FLASH and SDFDC FLASH. For all these three type of flash translations layers, there will be a .o file which is archieved into PPC860gnuvx.a for each translation layer. For example, for a NAND-FLASH, there will be a .o file called "nftllite.o" where as nftl for "NAND Flash" and lite stands for "the file is from M-Systems FLite". If you see the documentation on TrueFFS, you can see that the "TrueFFS for Tornado is a VxWorks-compatible implementation of M-Systems FLite, Version 2.0" in the introduction itself. Likewise, if there will be a file called "ftllite.o" for NOR-FLASH Translation Layers. TrueFFS's special feature says that it can transact with the flash technology using wear-level algorithm, erase-after-write algorithms. To provide these special algorithms, M-Systems FLite provides a flash translation layer files, which could possibly be a licensed version from WRS, or otherwise you could approach M-Systems FLite. I had a thorough search on M-Systems FLite web site and found no downloads for NOR-FLASH Arrays. For the sake of understanding the problem, we just extracted the ftllite.o file from PPCCEC603gnuvx.a and added into PPC860gnuvx.a and the vxWorks image was successfully built, but when we attempt to format the TrueFFS, it returns the value -1. And hence, I concluded that we may need to know whether it's a licensed version or not. Any help would be appreciated. ARP Newsgroups: comp.os.vxworks Subject: TrueFFS mountFTL Date: Sun, 30 Jul 2000 01:30:32 GMT From: fywang@my-deja.com Organization: Deja.com - Before you buy. Message-ID: <8m00fo$6on$1@nnrp1.deja.com> Hi all, I am trying to build a TrueFFS on my 2 AM29LV017D chips. Since AM29LV017D is a NOR-based flash, I should define INCLUDE_TL_FTL in my sysTffs.c file. In target/src/drv/tffs/tffsConfig.c: TLentry tlTable[] = /* TL tables */ { ... #ifdef INCLUDE_TL_FTL #ifdef FORMAT_VOLUME {mountFTL, formatFTL}, #else mountFTL, #endif /* FORMAT_VOLUME */ #endif /* INCLUDE_TL_FTL */ ... } I got a compile error when building my BSP: ... C:\Tornado\target\proj\gold860 \default\C:\Tornado\target\src\drv\intrCtl\ppc860Intr.c:224: undefined reference to `mountFTL' C:\Tornado\target\proj\gold860 \default\C:\Tornado\target\src\drv\intrCtl\ppc860Intr.c:224: undefined reference to `formatFTL' make: *** [vxWorks] Error 0x1 Searching "mountFTL" shows it's referenced by h/tffs/stdcomp.h & src/drv/tffs, no definition. The only occurence in lib is libPPCEC603gnuvx.a, but I am working on PPC860. Is there something wrong with my Tornado2/VxWorks? Can I work around this problem? Regards, Tom Wang Sent via Deja.com http://www.deja.com/ Before you buy. --------------------------- Newsgroups: comp.os.vxworks Subject: Configure BSP fails Date: Tue, 1 Aug 2000 12:30:51 +0100 From: "Henrik Husted" Organization: Customer at Tele Danmark Erhverv Message-ID: <8m68pc$m4t$1@news1.tele.dk> Hi. When I select 'Project -> Configure BSP' I get the following error message: "WindCFG was unable to run the C preprocessor and read the default configuration. Check your environment or reinstall Tornado" Hmm...I running from within the IDE (on a WinNT 4 sp6) and the documentation states that I do not need to be concerned about setting up the environment variables. Reinstalling Tornado didn't help much. I'm a bit new to VxWorks but I need to build a BSP for my special in-house hardware (based on a SH7410), right? TIA - Henrik --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Tue Aug 1 09:32:40 2000 From: David Laight Date: Tue Aug 1 09:32:42 PDT 2000 Subject: RE: verbosity macro Sebastien was having problem making a 'varags' version of: > I have a macro very convenient to debug my code. > For that I compile my code in debug mode (flag DRV_DEBUG_MODE defined), > and I use the following macro to display my debug messages with a > certain level > of verbosity that I can change (by a global variable): > > Macro: > #ifdef DRV_DEBUG_MODE > #define VERB_DRV(level,a1,a2,a3,a4,a5,a6,a7) \ > if (level <=3D drv_dbg) \ > {logMsg(a1,a2,a3,a4,a5,a6,a7);} > #else > #define VERB_DRV(level,a1,a2,a3,a4,a5,a6,a7) {} > #endif With a standard C preprocessor you have no choice but to add another set of () making the calls VERB_DRV( level, ( arg1, arg2, arg3 ) ) Also note that to avoid some control flow pitfalls you need to make the macro body a C statement, typically: #ifdef DEBUG #define drv_debug( lvl, args ) \ do { if ((lvl) & drv_debug) printf args; } while (0) #else #define drv_debug( lvl, args ) #endif Now the code: if (some_error_condition) drv_debug( 10, ( "it's all gone horribly wrong at line %d in file " __FILE__ "\n", __LINE__ )); else xyz(); will work with whether or not DEBUG is defined. Remember 'if () {}; else' isn't valid C. I can't remember whether vxWorks has vsprintf (or better vnsprintf) - which you need to do any clever formatting of the output. David ---------------------------------------------------------------- David Laight email: dsl@tadpole.co.uk Tadpole Technology plc phone: +44 1223 428 232 Cambridge, UK fax: +44 1223 428 201 From vxwexplo-errs@csg.lbl.gov Tue Aug 1 10:50:24 2000 From: David Laight Date: Tue Aug 1 10:50:26 PDT 2000 Subject: re: vxWorks Route Table I tend to agree with Matt's summary of the vxWorks route table. > I'm trying to include the capability in our vxWorks system > of changing IP addresses on the fly after the image is > loaded. I have been able so far to stop an interface and > reload it with a new IP, but I need a way to delete all the > routes on a given interface. This product is not intended > to have routing functionality, so RIP and OSPF are not > being used. The only way I can find right now is to use > routeDelete(), but routeDelete does not allow the deletion > of more than one entry with each call, and I see no route > function that allows me to specify the interface. There > must be a way to associate interfaces with the routes, if > for no other reason than that routeShow() does it, ( plus, > a route table is pretty useless without it... :) I also > would need to find a way to walk through the route table, > but I cannot find a global symbol for this table. Is the > table a static variable for the routeLib module? I know > that is the way the endList the muxLib module uses does > things, so I gather it must be similar. Stepping through > the routeShow() function, I see that eventually rn_walktree > () is called, which makes an indirect function call to > rtEntryPrint(). Beyond that I haven't been able to > decipher much on how the interfaces are associated in the > route entries. IMHO it is rather excessive for a default route table for a simple system. (It is probably fine if you are an ISP!) I also found that the code isn't very robust, attempting to delete and/or create routes can cause the table to get completely stuffed. This is particularly true when using the ppp and dhcpc code to assign dynamic addresses. Anyway changing the local IP address is relatively straightforward, ALL you need to do to disconnect IP from an interface is: ifAddrGet( "xyz0", my_old_address ); inet_netof_string( my_old_address, my_network ); arpFlush(); arpDelete( my_old_address ); hostDelete( sysBootParams.targetName, my_old_address ); ipDetach( 0, "xyz" ); routeDelete( my_network, my_old_address ); ifRouteDelete( "xyz", 0 ); If sysBootParams.gad is defined you may also need to explicitly delete the relevant route. Bringing it up with a new address is just a matter of calling: ipAttach( 0, "xyz" ); usrNetIfConfig( "xyz", 0, my_new_address, sysBootParams.targetName, my_new_netmask ); possibly followed by a routeAdd() for sysBootParams.gad. If you are using dhcp it is all more complicated. David ---------------------------------------------------------------- David Laight email: dsl@tadpole.co.uk Tadpole Technology plc phone: +44 1223 428 232 Cambridge, UK fax: +44 1223 428 201 From vxwexplo-errs@csg.lbl.gov Wed Aug 2 04:03:10 2000 From: Vxworks Exploder Date: Wed Aug 2 04:03:12 PDT 2000 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Wed Aug 2 04:03:05 PDT 2000 Subject: TFTP server problem on DOS FS 2.0 Subject: Re: VMEbus reset on boot? Subject: Libpcap vxWorks??? BPF?? Subject: multiple unit END Subject: Re: PCI Configuration Subject: Re: Can I pass the structures through the Message Queues! Subject: Re: Can I pass the structures through the Message Queues! Subject: Re: BSP for the Mot 8260VADS? Subject: Re: Wind River support experience? Subject: Re: raw keyboard input to Tornado shell process? Subject: Re: Subsystem Enable Subject: END in polled mode Subject: Re: mbufs and sockets Subject: RE: verbosity macro Subject: Re: Wind River support experience? Subject: Re: PMCspan expansion board setup Subject: Re: building GCC for powerpc-wrs-vxworks on a solaris Subject: Re: verbosity macro Subject: re: vxWorks Route Table Subject: interrupt: dc0 - no carrier Subject: Re: Wind River support experience? Subject: Re: interrupt: dc0 - no carrier Subject: shared memory network establishment Subject: SENS Stack. Subject: Re: Can I pass the structures through the Message Queues! Subject: Re: TrueFFS mountFTL Subject: Re: Wind River support experience? Subject: SNTP server ACTIVE mode Subject: Re: Wind River support experience? Subject: ATA/CompactFlash on a PPC860 Subject: 8260 MCC Initialization. Subject: Re: BSP for the Mot 8260VADS? Subject: Re: PCI Configuration Subject: Subsystem Enable Subject: Re: light-weight processes in VxWorks Subject: light-weight processes in VxWorks Subject: Re: Bothered by DHCP Subject: Re: raw keyboard input to Tornado shell process? Subject: Re: Bothered by DHCP Subject: Re: interrupt: dc0 - no carrier Subject: Re: Looking for help with MPCxxx and NEC SDRAM Subject: Re: Bothered by DHCP Subject: Re: UDP problem Subject: Re: Bothered by DHCP Subject: Re: multiple unit END Subject: Re: [BSP] question ? Subject: Re: VxWorks & ARM Dev Board Subject: Dynamicaly loading an application symbols Subject: File system ioctls Subject: Re: ATA/CompactFlash on a PPC860 ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: TFTP server problem on DOS FS 2.0 Date: Tue, 01 Aug 2000 13:33:16 +0200 From: Uwe Fuhr Organization: Ericsson Eurolab Deutschland GmbH Message-ID: <3986B57C.F7BA34E9@eede.ericsson.se> Reply-To: Uwe.Fuhr@eed.ericsson.se Hi, if anybody can help, feel free. System setup: - ------------- 1) Board A with TFFS and DOSFS2.0, the DOSFS device is called "/RFA/", - sub directory "/RFA/dir_1/dir_2/" contains a number (approx. 8) of binary files and an ASCII file containing the file names of the the binary files. 2) Board B (up to 6) requesting download of binary files. The flow is like that: The mentioned files are ftp'ed from a Sun workstation onto Board A in the named directory. Board A sends a message to all connected boards B (via a socket) containing the complete directory name and the name of the ASCII file in that directory. Each of the boards B open a TFTP session and request the binary files. Sketch: - ------- --------- --------- Board A ------------------- Board B1 --------- | --------- | --------- --------- Board B2 | --------- | --------- --------- Board B3 | --------- | --------- --------- Board Bn (n max. = 6) --------- Problem: - -------- Board A runs a TFTP server, allowing simultaneous connections. If more than one B-Board connects to the TFTP server in parallel an error occurs. Information from the target shell: - ---------------------------------- - -> i ... tTftpRRQ tftpdFileRea f2e198 100 SUSPEND 33a020 f2e128 d0003 0 tTftpRRQ tftpdFileRea f30ab0 100 SUSPEND 33a020 f30a40 c0006 0 tTftpRRQ tftpdFileRea f2b880 100 SUSPEND 33a020 f2b810 d0003 0 ... - -> printErrno 0xc0006 0xc0006 = S_ioLib_DISK_NOT_PRESENT - -> printErrno 0xd0003 0xd0003 = S_iosLib_INVALID_FILE_DESCRIPTOR ___________________________________________________________________________ _/ _/ _/ Uwe Fuhr EED/E/D/U Access Design _/ _/ _/ Ericsson Eurolab Deutschland phone : +49-5121 - 707104 _/ _/ _/ Daimlerring 9 FAX : +49-5121 - 707314 _/ _/ _/ 31135 Hildesheim - Germany mobile : +49-173 - 5430771 __________________________________ mailto:Uwe.Fuhr@eed.ericsson.se ____ \___________________________________/ --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VMEbus reset on boot? Date: Tue, 1 Aug 2000 07:53:36 -0500 From: "Mark P. Fishman" Message-ID: References: <3985e2b5.28408359@news.boeing.com> comments below... On Mon, 31 Jul 2000, Lee DeRaud wrote: > On Mon, 31 Jul 2000 09:54:44 -0500, "Mark P. Fishman" > wrote: > >We would like to have all cards on the VME bus reset whenever we > >(re)boot vxWorks. Is there a way to do this in software? either > >with a target-shell command from a startup script or from a > >vxWorks function call? > This sounds a bit backwards: usually you're (re)booting VxWorks as a > *result* of a VME bus reset, not the other way around. Check the > documentation for the VME interface chipset used by your processor: > most of them have a way to strobe the SYSRESET signal. I know it sounds backwards; the problem is that we want to ensure that all the other hardware is reset to a known state when we so a warm reboot -- we're still in development. A bus reset (without rebooting the CPU board or cycling power) seems like the way to do this. Cheers -- Mark F. - -- reply to: mfishman at el el dot em eye tee dot ee dee ewe OPINIONS EXPRESSED ARE MY OWN AND DO NOT REFLECT THOSE OF MY EMPLOYER OR ANY OTHER ENTITY IN THE KNOWN WORLD. --------------------------- Newsgroups: comp.os.vxworks Subject: Libpcap vxWorks??? BPF?? Date: 1 Aug 2000 13:26:48 GMT From: cedilnik andrej Organization: University of Maryland, Baltimore County Message-ID: <8m6j6o$f5l$1@news.umbc.edu> Hi! So, VxWorks is supposed to have BSD 4.4 network. BSD uses BPF for accessing stuff on the data link layer. Well, in order to access BPF, we need to open file /dev/bpf???? How des that apply to VxWorks, which does not have a filesystem? Andy - -- - ------------------------------------------------------------------- )\._.,--....,'``. Andrej Cedilnik (acedil1@cs.umbc.edu) /, _.. \ _\ (`._ ,. URL: http://www.cs.umbc.edu/~acedil1 `._.-(,_..'--(,_..'`-.;.' TEL: (301)725-5724 ICQ: 27889915 - ------------------------------------------------------------------- --------------------------- Newsgroups: comp.os.vxworks Subject: multiple unit END Date: Tue, 01 Aug 2000 09:28:10 -0400 From: "Leung, Spencer [SKY:1E13:EXCH]" Organization: Nortel Networks Message-ID: <3986D06A.3796EBBE@americasm01.nt.com> Hi, I'm trying to write an END for a 4 ports Ethernet chipset (MMC EPIF50). The templateEnd.c only support single unit. Does anyone have a version of templateEnd.c for multiple units or have an END written for the EPIF50 or 200? Thanks. Spencer - -- - ---------------------------------------------------------------------- NORTEL SPENCER LEUNG leungsp@nortelnetworks.com NETWORKS Data Systems Engineer Tel 613 765-4790 1E13 Fax 613 768-4904 www.nortelnetworks.com Systems Development ESN 395 GMS 045/4P/E9 - ---------------------------------------------------------------------- --------------------------- Newsgroups: comp.lang.c-programming,comp.os.vxworks,comp.sys.powerpc,comp.sys.powerpc.tech Subject: Re: PCI Configuration Date: 1 Aug 2000 06:43:33 -0700 From: andrew@user2.teleport.com (Andrew Klossner) Message-ID: <8m6k65$ocr$1@user2.teleport.com> References: <%plh5.237979$7o1.6352532@news2.rdc1.on.home.com> <8m4vld$l8a$1@nnrp1.deja.com> >> Command register as shown below. Unfortunately, after >> configuration, I am unable to write to memory using sysOutXXXX(). Don't keep it a secret. What happens when you try to write to memory? Does it silently do nothing? Do you take a data access exception? Do the fans on your power supply spin down? My guess is that you're running the 604 in big-endian mode, and you need to convert big- to little-endian when accessing I/O registers on the PCI bus. > Is the cache enabled? A better question is "Is the MMU enabled?" You're inventing three addresses (d2000000, d4000000, and e0020000). If the MMU is enabled, how are you mapping those locations into your virtual space? Some nits -- three comments describe "BAR 0". The second and third should instead mention "BAR 1" and "BAR 2". -=- Andrew Klossner (andrew@teleport.com) --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Can I pass the structures through the Message Queues! Date: Tue, 1 Aug 2000 15:53:09 -0500 From: "Keith McHugh" Organization: Esat Net Customer Message-ID: <3986e2aa.0@seat.s3two.ie> References: <3986a137.0@seat.s3two.ie> Reply-To: "Keith McHugh" You don't've to answer this question! I found it that I can pass the strutures (as a attached buffer) with the message queues. Many Thanks --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Can I pass the structures through the Message Queues! Date: Tue, 01 Aug 2000 11:16:29 -0700 From: Joe Durusau Organization: Lockheed Martin Corporation Message-ID: <398713FD.DB10F1BE@delphi.com> References: <3986a137.0@seat.s3two.ie> Not without making a mess. You can. however pass pointers through them without too much trouble. Speaking only for myself, Joe Durusau Keith McHugh wrote: --------------------------- Newsgroups: comp.os.vxworks Subject: Re: BSP for the Mot 8260VADS? Date: Tue, 1 Aug 2000 10:42:29 -0500 From: "Derek Opitz" Organization: Motorola CIG Message-ID: <8m6r2u$5bg3@nntp.cig.mot.com> References: <3987adf6.20551791@24.0.240.30> Thanks for the info, we were wondering why we couldnt get our serial driver to work for anything over 9600 baud. Is it just a problem with the serial driver? Gary Beck wrote in message news:3987adf6.20551791@24.0.240.30... > Yes, there is a BSP for the Motorola VADS8260. The first verion is > titled VADS820 and creates a drectory of the same name under > \tornado\target\config directory. The second version is titled > ADS8260. Be careful with their serial driver. It does not work above > 9600 baud. There are also other bugs with this BSP. The BSP name > that WRS gave you is incorrect. > > > On Sun, 30 Jul 2000 15:43:43 GMT, clayton@shore.net (Mark Clayton) > wrote: > > > > >Is there a BSP for the 8260VADS board. it sounds like it > >from heading this group. What is it called? WindRivers > >says to use the 860FADS BSP, but that seems incorrect. Are > >they right!?! > > > >Thanks, > >Mark > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Wind River support experience? Date: Tue, 1 Aug 2000 10:34:04 -0500 From: "Derek Opitz" Organization: Motorola CIG Message-ID: <8m6qj5$c6u4@nntp.cig.mot.com> References: <03009a08.5b942430@usw-ex0105-036.remarq.com> About 5 years ago, the support was pretty good. Now it is bad. If you call in and file a TSR, it takes them 3 or 4 days to get back with you, and they never have the answers, so it takes even longer, sometimes months, to get back to you. If you send email, you will get a follow up email in a few days, but help is almost non-existent via email. If you call your local sales guy and complain, he can usually get you some answers in decent time, and this is the approach we have been using for a while. It seems to work pretty good. But going through normal tech support is pointless. Obviously they think people developing products for htem don't have deadlines. Nick Hoogland wrote in message news:03009a08.5b942430@usw-ex0105-036.remarq.com... > Hi, > > Who can share his/her experience with Wind River support (both > positive and negative). > > For reasons of evaluation of the suitebility of the VxWorks, > also related to available tools and cost, I am trying to get > some answers on basic questions as foot prints and boot time > (information that should have been readily available within Wind > River) for three weeks now. I posted my questions at the Wind > River site and mailed them directly to (sales) contact persons > of Wind River. The answers I received consisted of glossary's > that only contained some very global VxWorks and Tools > information and which did not address any of my answers even > remotely. > > Selecting a company with such poor pre-sales support for > something as essential as an OS is questionable, so how is the > support after being committed to Wind River? > > Thanks, > > Nick > > > ----------------------------------------------------------- > > Got questions? Get answers over the phone at Keen.com. > Up to 100 minutes free! > http://www.keen.com > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: raw keyboard input to Tornado shell process? Date: Tue, 01 Aug 2000 09:13:42 -0700 From: Ken Norlin Organization: NASA Dryden FRC Message-ID: <3986F736.3F61C561@dfrc.nasa.gov> References: <8m53v7$oco$1@nnrp1.deja.com> p_lavarre@my-deja.com wrote: > > How can a process run from the Tornado shell ask to get one keystroke > at a time from the shell input device? > Hi, We've used VxCurses at my site for a few years now -- it allows you to get a character (or keystroke I think) at a time. You could check that code to see how this is done. (It's been years now since I grabbed a copy -- I think it was over on the Wind River ftp site). Ken N. - -- Ken Norlin --- Simulation Engineer NASA Dryden Flight Research Center ken.norlin@dfrc.nasa.gov (661) 276-2046 FAX (661) 276-2792 --------------------------- Newsgroups: comp.lang.c-programming,comp.os.vxworks,comp.sys.powerpc,comp.sys.powerpc.tech Subject: Re: Subsystem Enable Date: Tue, 01 Aug 2000 11:41:21 -0500 From: Gary Milliorn Organization: Motorola Semiconductor Products Sector Message-ID: <3986FDB1.D4359BC9@motorola.com> References: Rob Targosz wrote: > > Does anybody know what the equivalents for the following IBM PS/2 Model 70 > architecture registers are on PowerPC 604? > > 1. 0x03C3 - Motherboard Enable VGA Video Subsystem > 2. 0x4AE8 - Motherboard Enable 8514/A Video Subsystem > 3. 0x0102 - Video Subsystem General Enable NOP. By which I mean, those are system-architecture specific registers. The MPC604 doesn't care one way or another. If there's something on your '604 system that responds to those ISA I/O addresses, fine; otherwise, you shouldn't need worry. > Thanks, > Rob > robtargosz@hotmail.com - -- Gary Milliorn | Motorola PowerPC Reference Platforms | Fax: 512-895-2638 Personal Computer Systems Division | email: Gary.Milliorn@Motorola.com --------------------------- Newsgroups: comp.os.vxworks Subject: END in polled mode Date: Tue, 01 Aug 2000 13:04:56 -0400 From: "Leung, Spencer [SKY:1E13:EXCH]" Organization: Nortel Networks Message-ID: <39870338.5156DDD0@americasm01.nt.com> Hi, Is it possible to force an END to switch between interrupt mode and polled mode by invoking templatePollStart and templatePollStop manually? Is there a way to load the END with polled mode as the default mode? Thanks in advance for any help. Regard, Spencer - -- - ---------------------------------------------------------------------- NORTEL SPENCER LEUNG leungsp@nortelnetworks.com NETWORKS Data Systems Engineer Tel 613 765-4790 1E13 Fax 613 768-4904 www.nortelnetworks.com Systems Development ESN 395 GMS 045/4P/E9 - ---------------------------------------------------------------------- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: mbufs and sockets Date: Tue, 01 Aug 2000 13:14:53 -0400 From: Dmitriy Zavin Message-ID: <3987058D.4BA799A4@txc.com> References: <3985B8A6.E2CF02B5@txc.com> <3986781F.7AB07480@bitband.com> I've looked for it in the manuals, does it require that my driver follow the END driver interface?
I am using the BSD4.4 interface because its simpler and provides more flexibility for me currently.
Which manual should i look at to find information about it.

Thanks,

--Dima

Leonid Rosenboim wrote:

Dimka,

youy should spend some time reading the fine manual about Zero-Copy TCP,
or ZBuf API, it is exactly what you are looking for.

Dmitriy Zavin wrote:

> Hi. Anyone know how I can pass data to a socket that is allready in an
> mbuf?
> I don't want to copy it out of an mbuf chain into a flat buffer just
> so the socket layer will copy that buffer into a new chain.. I want to
> skip that whole process and just hand an open socket some data in an
> allready preallocated mbuf chain.
>
> Thanks,
>
> --Dima
>

-- 
/----------------------------------/
Dmitriy Zavin
Software Engineering Intern
TranSwitch Corporation
email: dzavin@txc.com
Ph: 203-929-8810 ext. 2624
/----------------------------------/
  --------------------------- Newsgroups: comp.os.vxworks Subject: RE: verbosity macro Date: Tue, 1 Aug 2000 17:32:47 +0100 (BST) From: David Laight Organization: Lawrence Berkeley National Laboratory Message-ID: <200008011632.RAA04904@dsl-2.tadpole.co.uk> Sebastien was having problem making a 'varags' version of: > I have a macro very convenient to debug my code. > For that I compile my code in debug mode (flag DRV_DEBUG_MODE defined), > and I use the following macro to display my debug messages with a > certain level > of verbosity that I can change (by a global variable): > > Macro: > #ifdef DRV_DEBUG_MODE > #define VERB_DRV(level,a1,a2,a3,a4,a5,a6,a7) \ > if (level <=3D drv_dbg) \ > {logMsg(a1,a2,a3,a4,a5,a6,a7);} > #else > #define VERB_DRV(level,a1,a2,a3,a4,a5,a6,a7) {} > #endif With a standard C preprocessor you have no choice but to add another set of () making the calls VERB_DRV( level, ( arg1, arg2, arg3 ) ) Also note that to avoid some control flow pitfalls you need to make the macro body a C statement, typically: #ifdef DEBUG #define drv_debug( lvl, args ) \ do { if ((lvl) & drv_debug) printf args; } while (0) #else #define drv_debug( lvl, args ) #endif Now the code: if (some_error_condition) drv_debug( 10, ( "it's all gone horribly wrong at line %d in file " __FILE__ "\n", __LINE__ )); else xyz(); will work with whether or not DEBUG is defined. Remember 'if () {}; else' isn't valid C. I can't remember whether vxWorks has vsprintf (or better vnsprintf) - which you need to do any clever formatting of the output. David - ---------------------------------------------------------------- David Laight email: dsl@tadpole.co.uk Tadpole Technology plc phone: +44 1223 428 232 Cambridge, UK fax: +44 1223 428 201 --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Wind River support experience? Date: Tue, 01 Aug 2000 13:19:20 -0400 From: Dmitriy Zavin Message-ID: <39870698.3D8AC78A@txc.com> References: <03009a08.5b942430@usw-ex0105-036.remarq.com> Actually, I have had a different experience. I called up the FAE near us and talked to him. He was not very helpfull since I had a pretty in depth question. But he directed me to a tech support person (direct phone number) which, after I tried a couple of times, actually picked up the phone and helped me out. I was pretty surprised. So I guess it depends if your FAE wants to give you direct phone numbers for tech support. But i do have to agree thatthe email support system is ABSOLUTELY USELESS.  It is very sad that such an advanced (but slow :) ) OS has such bad support.  Give RT-Linux a little time, and all these no-support companies are going DOWN! :)

Hope this helped.

--Dima

Nick Hoogland wrote:

Hi,

Who can share his/her experience with Wind River support (both
positive and negative).

For reasons of evaluation of the suitebility of the VxWorks,
also related to available tools and cost, I am trying to get
some answers on basic questions as foot prints and boot time
(information that should have been readily available within Wind
River) for three weeks now. I posted my questions at the Wind
River site and mailed them directly to (sales) contact persons
of Wind River. The answers I received consisted of glossary's
that only contained some very global VxWorks and Tools
information and which did not address any of my answers even
remotely.

Selecting a company with such poor pre-sales support for
something as essential as an OS is questionable, so how is the
support after being committed to Wind River?

Thanks,

Nick

-----------------------------------------------------------

Got questions?  Get answers over the phone at Keen.com.
Up to 100 minutes free!
http://www.keen.com

-- 
/----------------------------------/
Dmitriy Zavin
Software Engineering Intern
TranSwitch Corporation
email: dzavin@txc.com
Ph: 203-929-8810 ext. 2624
/----------------------------------/
  --------------------------- Newsgroups: comp.os.vxworks Subject: Re: PMCspan expansion board setup Date: Tue, 01 Aug 2000 19:19:21 +0200 From: Sebastien Fauris Organization: Club-Internet (France) Message-ID: <39870699.82EB6124@club-internet.fr> References: Hi, I have used a Ramix PMC-237 expander on a PowerCore750 motherboard and also a Motorola PMC-Span on a MVME2300. Minbo Shim wrote: > Hi, > > I am trying to setup a PMC card in a PMC adapter carrier > board from Motorola (PMCspan-002) which is attached to a > MVME2431. The PMC card is in the slot 1 (device # 2) of > the expansion board with interrupt level 3 (PMC_INT_LVL3). > The DEC21150 PCI-to-PCI bridge chip provides the interface > between the primary PCI bus and the secondary PCI bus > through Type 0 and Type 1 configuration cycle. I am using > CHRP Extended VME memory map for this configuration. > > My questions are: > > 1) can PMC_SPAN work with the CHRP memory model? There is no reason why it wouldn't work. It depends only of the addresses used to set up the PCI to PCI bridge on the PMC Span. On MVME2300, addresses used for the PMC Span can be found in mv2600.h. > > > 2) can I use some known pci setup functions like > pciDevConfig(), pciConfigOutLong(), etc? > absolutely, it is completely transparent to you the fact that the PCI devices are on the primary PCI bus (bus #0) or secondary PCI bus (bus#1). > > 3) what is the physical address (hardware unmapped) of the > PMC card as seen from the processor? > > 4) what is the virtual address (MMU mapped)? > > Any help/comments on this would be greatly appreciated. > > -Minbo > GDRS Same thing, this can be seen on a MVME2300 in mv2600.h. You need to translate a PCI address to a CPU address and vice-versa. You need also to swap the bytes (PPC is big endian, PCI is little endian). Best of luck, Sebastien. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: building GCC for powerpc-wrs-vxworks on a solaris Date: Tue, 1 Aug 2000 18:19:36 +0100 From: "Dave Korn" Organization: (Posted via) U-NET Internet Ltd. Message-ID: References: Xenos wrote in message ... >Can anyone please help me? I've trying to build GCC on a solaris 2.5.1 >for powerpc-wrs-vxworks. It keeps getting parse error building >iovfprintf.c on all lines containing a reference to va_arg. It >includes and can find stdarg.h but it seems that va_arg is not >getting defined (but va_list is). Has anyone seem this?? > >I was able to successful build the very first time, but I accidentally >cobbered it while trying to tar it up for safe keeping!! I have yet >to be able to duplicate my initial success! > >I able able to build the bin utils and a native compile, no problem. The problem is to do with the target include headers. When you use configure to set up the target type, you must also use the option - --with-headers= where is the path to the Tornado target headers installed on your system, i.e. something/tornado/target/h or similar. DaveK - -- "Reality is whatever doesn't go away after you stop believing in it." -- Philip K. Dick --------------------------- Newsgroups: comp.os.vxworks Subject: Re: verbosity macro Date: Tue, 01 Aug 2000 20:03:54 +0200 From: Sebastien Fauris Organization: Club-Internet (France) Message-ID: <3987110A.F04218F7@club-internet.fr> References: <200008011632.RAA04904@dsl-2.tadpole.co.uk> Hi David, I found out about my variable arguments problem. Here is the solution: #define VERB_PNP(level, format, args...) \ if (level <= pnp_dbg) \ {printf (format, ## args);} #else #define VERB_PNP(level, format, args...) {} Cheers, Sebastien. --------------------------- Newsgroups: comp.os.vxworks Subject: re: vxWorks Route Table Date: Tue, 1 Aug 2000 18:50:32 +0100 (BST) From: David Laight Organization: Lawrence Berkeley National Laboratory Message-ID: <200008011750.SAA05662@dsl-2.tadpole.co.uk> I tend to agree with Matt's summary of the vxWorks route table. > I'm trying to include the capability in our vxWorks system > of changing IP addresses on the fly after the image is > loaded. I have been able so far to stop an interface and > reload it with a new IP, but I need a way to delete all the > routes on a given interface. This product is not intended > to have routing functionality, so RIP and OSPF are not > being used. The only way I can find right now is to use > routeDelete(), but routeDelete does not allow the deletion > of more than one entry with each call, and I see no route > function that allows me to specify the interface. There > must be a way to associate interfaces with the routes, if > for no other reason than that routeShow() does it, ( plus, > a route table is pretty useless without it... :) I also > would need to find a way to walk through the route table, > but I cannot find a global symbol for this table. Is the > table a static variable for the routeLib module? I know > that is the way the endList the muxLib module uses does > things, so I gather it must be similar. Stepping through > the routeShow() function, I see that eventually rn_walktree > () is called, which makes an indirect function call to > rtEntryPrint(). Beyond that I haven't been able to > decipher much on how the interfaces are associated in the > route entries. IMHO it is rather excessive for a default route table for a simple system. (It is probably fine if you are an ISP!) I also found that the code isn't very robust, attempting to delete and/or create routes can cause the table to get completely stuffed. This is particularly true when using the ppp and dhcpc code to assign dynamic addresses. Anyway changing the local IP address is relatively straightforward, ALL you need to do to disconnect IP from an interface is: ifAddrGet( "xyz0", my_old_address ); inet_netof_string( my_old_address, my_network ); arpFlush(); arpDelete( my_old_address ); hostDelete( sysBootParams.targetName, my_old_address ); ipDetach( 0, "xyz" ); routeDelete( my_network, my_old_address ); ifRouteDelete( "xyz", 0 ); If sysBootParams.gad is defined you may also need to explicitly delete the relevant route. Bringing it up with a new address is just a matter of calling: ipAttach( 0, "xyz" ); usrNetIfConfig( "xyz", 0, my_new_address, sysBootParams.targetName, my_new_netmask ); possibly followed by a routeAdd() for sysBootParams.gad. If you are using dhcp it is all more complicated. David - ---------------------------------------------------------------- David Laight email: dsl@tadpole.co.uk Tadpole Technology plc phone: +44 1223 428 232 Cambridge, UK fax: +44 1223 428 201 --------------------------- Newsgroups: comp.os.vxworks Subject: interrupt: dc0 - no carrier Date: Tue, 01 Aug 2000 13:22:16 -0500 From: Greg Willden Organization: Southwest Research Institute Message-ID: <39871557.240F928A@swri.edu> I am using a Motorola MVME2307 card. As the board boots up the following line periodically gets output through the serial line. Also, after the board boots up and I telnet to it and the same line is repeatedly sent out to the screen. However if I attach a target server/shell to it in Tornado I do not see the error. The error message is: interrupt: dc0 - no carrier In the telnet window I do have a prompt and it is possible to type in a command like "i" and receive the results. The prompt and results just scroll away quickly due to this error. Anyway the point is: Has anyone seen this before? I think that dc0 is the ethernet interface but I haven't been able to find in the source code where this error might be comming from. Any Ideas? Thanks in advance Greg --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Wind River support experience? Date: 1 Aug 2000 18:32:26 GMT From: Tom.Gee@nospamdciem.dnd.ca (Tom Gee) Organization: DCIEM Message-ID: References: <03009a08.5b942430@usw-ex0105-036.remarq.com> <39870698.3D8AC78A@txc.com> Reply-To: Tom.Gee@dciemDOTdndDOTca I've had two significant incidents in which I've involved WRS support. The first, which was pretty simple, was resolved in a day or two. The second very serious issue is still open. I'm trying to get Solaris-based 'wdbserial' to work with a certain target. This target works fine with NT-based wdbserial, but not at all with the Solaris one. With such a specific problem, I thought the resolution would not take long. It's been almost a month since I opened the TSR, and it takes me many phone calls to get anyone to call me back about this problem. So far, no progress at all. It is unlikely we will be renewing our support contract. Unless we can purchase software maintenance and access to WindSurf separately. Tom - --- Smart Aircrew Integrated Life Support System -- software team lead Aerospace Life Support / DCIEM / Defence R&D Canada (www.dciem.dnd.ca) --------------------------- Newsgroups: comp.os.vxworks Subject: Re: interrupt: dc0 - no carrier Date: Tue, 01 Aug 2000 14:20:19 -0500 From: Andrew Johnson Organization: APS, Argonne National Laboratory Message-ID: <398722F3.5800D758@aps.anl.gov> References: <39871557.240F928A@swri.edu> Greg Willden wrote: > > The error message is: > > interrupt: dc0 - no carrier We used to get this I think with the combination of Tornado 1.0.1, the SENS IP stack and a 100Mbps full duplex connection. Change any one of those variables and it goes away, ie: Reinstall Tornado 1.0.1 without installing the SENS IP stack, or install Tornado 2.0, or force your network hub to half-duplex for this port. There may also be a fix available from WRS but I didn't bother asking myself... - - Andrew - -- Sharks kill 10 people each year. People kill 60,000,000 sharks each year. --------------------------- Newsgroups: comp.os.vxworks Subject: shared memory network establishment Date: Tue, 01 Aug 2000 20:20:03 GMT From: Dan Yang Organization: Magma Communications Ltd. Message-ID: <39873051.7725E00F@drs.ca> Does anybody have the experience of building a shared memory network on the fly, that means, after the boards are booted via flash file system for example. I attempted to use usrNetInit("sm=0xf0004100(0,1) b=161.10.1.2 g=161.10.1.1"); but it failed. Any hints would be appreciated. Dan --------------------------- Newsgroups: comp.os.vxworks Subject: SENS Stack. Date: Tue, 1 Aug 2000 13:50:05 -0700 From: "SCR" Message-ID: <39873324$0$26899@wodc7nh1.news.uu.net> SENS Stack Users, Has anyone faced faced problems after porting SENS Stack. Basically my question is with regard to the Stability of SENS stack under extreme load conditions. thanks, SCR. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Can I pass the structures through the Message Queues! Date: Tue, 01 Aug 2000 15:34:33 -0500 From: CRichards Organization: Nortel Networks Message-ID: <39873459.CF8C4B74@hotmail.com> References: <3986a137.0@seat.s3two.ie> <3986e2aa.0@seat.s3two.ie> You can... but it's probably not the best thing to do: it makes the code extremely difficult to port to new HW architectures (padding, etc). Best to use bytes streams. JMO. Chris. Keith McHugh wrote: > You don't've to answer this question! I found it that I can pass the > strutures (as a attached buffer) with the message queues. > > Many Thanks --------------------------- Newsgroups: comp.os.vxworks Subject: Re: TrueFFS mountFTL Date: 1 Aug 2000 20:56:29 GMT From: johan@borksoft.xs4all.nl Organization: XS4ALL Internet BV Message-ID: <8m7dht$4b6$1@news1.xs4all.nl> References: <8m67v8$734$1@overload.lbl.gov> Palaniappan wrote: : Hi, : To provide these special algorithms, M-Systems : FLite provides a flash translation layer files, which could possibly be a : licensed version from WRS, or otherwise you could approach M-Systems FLite. For the VxWorks version of TFFS you have to contact WindRiver. I contacted M-Systems once and they said that I had to contact WRS. : I had a thorough search on M-Systems FLite web site and found no downloads : for NOR-FLASH Arrays. For the sake of understanding the problem, we just : extracted the ftllite.o file from PPCCEC603gnuvx.a and added into : PPC860gnuvx.a and the vxWorks image was successfully built, but when we : attempt to format the TrueFFS, it returns the value -1. And hence, I : concluded that we may need to know whether it's a licensed version or not. When you want to use TFFS you have to purchase it from WRS. The version in the EC603 library is a "mistake", but you could use it of course. But there is a seperate license policy for TFFS. See also my website for info about this. Groeten, Johan - -- o o o o o o o . . . ___________________________________ o _____ || Johan Borkhuis | .][__n_n_|DD[ ====_____ | johan@borksoft.xs4all.nl | >(________|__|_[_________]_|________________________________| _/oo OOOOO oo` ooo ooo 'o!o!o o!o!o` === VxWorks page: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html === --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Wind River support experience? Date: 1 Aug 2000 21:11:40 GMT From: johan.borkhuis@peektraffic.nl Organization: Peek Traffic Message-ID: <8m7eec$4b6$2@news1.xs4all.nl> References: <03009a08.5b942430@usw-ex0105-036.remarq.com> Nick Hoogland wrote: : Hi, : Who can share his/her experience with Wind River support (both : positive and negative). My experience with the FAE's are very positive. Once we even had our HW shipped to one of the engineers to find a nasty problem, and got it back very soon with a solution. The email is a little bit less positive. It takes a lot of time before you get a usefull answer. But my experience is not as bad as some others said. What I is more of a problem is that if you have a problem and your TSR is being "promoted" to an SPR this SPR is not published but stays inside the WRS database. Groeten, Johan - -- o o o o o o o . . . ___________________________________ o _____ || Johan Borkhuis | .][__n_n_|DD[ ====_____ | johan.borkhuis@peektraffic.nl | >(________|__|_[_________]_|________________________________| _/oo OOOOO oo` ooo ooo 'o!o!o o!o!o` === VxWorks page: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html === --------------------------- Newsgroups: comp.os.vxworks Subject: SNTP server ACTIVE mode Date: Tue, 01 Aug 2000 14:31:00 -0700 From: Terry Duncan Message-ID: <39874194.D69F7E0C@cintacom.com> Has anyone been successfull in getting the SNTP server to function in active mode? In this mode the server is supposed to transmit the timestamp to SNTPS_DSTADDR every 64 seconds by default. Its not happening for me. I have no problem with passive mode in which the timestamp is transmitted only upon request from a client. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Wind River support experience? Date: Tue, 01 Aug 2000 14:43:57 -0700 From: Robert Organization: http://www.remarq.com: The World's Usenet/Discussions Start Here Message-ID: <045ac7d4.90b477d0@usw-ex0101-007.remarq.com> References: <03009a08.5b942430@usw-ex0105-036.remarq.com> <8m7eec$4b6$2@news1.xs4all.nl> You can probably gauge the effectiveness of Wind River support by the high activity on this newsgroup. I suspect that this newsgroup would be a lot less active if the Wind River support team were up to snuff. Rob't Blumberg Embedded Software Developer Anything below is tacked on by Remarq! - ----------------------------------------------------------- Got questions? Get answers over the phone at Keen.com. Up to 100 minutes free! http://www.keen.com --------------------------- Newsgroups: comp.os.vxworks Subject: ATA/CompactFlash on a PPC860 Date: Tue, 1 Aug 2000 15:16:35 -0700 From: "Steve Jones" Organization: CTSnet Internet Services Message-ID: <8m7htm$1pml$1@thoth.cts.com> I am trying to get the ATA driver up and running on my PPC860. I am setting up the PCMCIA interface in the PPC, and then initializing the ATA driver over the top of it. The reference I have is a little different it is using the Winbond chip over a PCI bus. I have made the changes to the driver but am a bit confused on the addressing to the register interface of an ATA device. Is there any one out their that has added a CompactFlash card to there PPC project? --------------------------- Newsgroups: comp.os.vxworks Subject: 8260 MCC Initialization. Date: Tue, 1 Aug 2000 18:19:56 -0400 From: "Teik-Ming Goh" Organization: Nortel Message-ID: <8m7hv4$igf$1@bcarh8ab.ca.nortel.com> Greeting, I have written an initialization routine for MCC2 channel 128 & 129 which are connected to TDM A2 for PPC 8260 and I have tested the code in EST SBC 8260 evaluation board with VxWorks as RTOS. The buffer descriptor status didn't change and the buffer was empty. Can anyone tell me how I can debug this problem? Thank you a lot. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: BSP for the Mot 8260VADS? Date: Tue, 01 Aug 2000 22:29:06 GMT From: nospam@home.net (Gary Beck) Organization: @Home Network Message-ID: <39874dc0.685435@24.0.240.30> References: <3987adf6.20551791@24.0.240.30> <8m6r2u$5bg3@nntp.cig.mot.com> There are basic fundamental problems with their driver, so we ended up writing our own. I couldn't convince WRS support of any serial driver problems. Also, if you are using the Motorola VADS8260, you should look at ADS8260.H and CONFIG.H. The addresses for the IMMR and HW control registers are incorrect. Also, you might want to confirm the bit definitions for the HW control register. On Tue, 1 Aug 2000 10:42:29 -0500, "Derek Opitz" wrote: >Thanks for the info, we were wondering why we couldnt get our serial driver >to work for anything over 9600 baud. > >Is it just a problem with the serial driver? > >Gary Beck wrote in message >news:3987adf6.20551791@24.0.240.30... >> Yes, there is a BSP for the Motorola VADS8260. The first verion is >> titled VADS820 and creates a drectory of the same name under >> \tornado\target\config directory. The second version is titled >> ADS8260. Be careful with their serial driver. It does not work above >> 9600 baud. There are also other bugs with this BSP. The BSP name >> that WRS gave you is incorrect. >> >> >> On Sun, 30 Jul 2000 15:43:43 GMT, clayton@shore.net (Mark Clayton) >> wrote: >> >> > >> >Is there a BSP for the 8260VADS board. it sounds like it >> >from heading this group. What is it called? WindRivers >> >says to use the 860FADS BSP, but that seems incorrect. Are >> >they right!?! >> > >> >Thanks, >> >Mark >> > --------------------------- Newsgroups: comp.lang.c-programming,comp.os.vxworks,comp.sys.powerpc,comp.sys.powerpc.tech Subject: Re: PCI Configuration Date: Tue, 01 Aug 2000 14:02:22 GMT From: "Rob Targosz" Organization: Excite@Home - The Leader in Broadband Message-ID: References: <%plh5.237979$7o1.6352532@news2.rdc1.on.home.com> <8m4vld$l8a$1@nnrp1.deja.com> <8m6k65$ocr$1@user2.teleport.com> "Andrew Klossner" wrote in message news:8m6k65$ocr$1@user2.teleport.com... > >> Command register as shown below. Unfortunately, after > >> configuration, I am unable to write to memory using sysOutXXXX(). > > Don't keep it a secret. What happens when you try to write to memory? > Does it silently do nothing? Do you take a data access exception? Do > the fans on your power supply spin down? > The write doesn't happen; the values are the same before and after writing to the address using sysOutXXXX(). > > My guess is that you're running the 604 in big-endian mode, and you > need to convert big- to little-endian when accessing I/O registers on > the PCI bus. > Thanks, I'll try that. > > > Is the cache enabled? > > A better question is "Is the MMU enabled?" You're inventing three > addresses (d2000000, d4000000, and e0020000). If the MMU is enabled, > how are you mapping those locations into your virtual space? > Actually, I should have deleted that code before posting. I'm simply using that to avoid having to reboot every time I mess up my BARs. The POST is assigning those values. > > Some nits -- three comments describe "BAR 0". The second and third > should instead mention "BAR 1" and "BAR 2". > That's the joy of copy and paste! Thanks. > > -=- Andrew Klossner (andrew@teleport.com) --------------------------- Newsgroups: comp.lang.c-programming,comp.os.vxworks,comp.sys.powerpc,comp.sys.powerpc.tech Subject: Subsystem Enable Date: Tue, 01 Aug 2000 14:21:04 GMT From: "Rob Targosz" Organization: Excite@Home - The Leader in Broadband Message-ID: Does anybody know what the equivalents for the following IBM PS/2 Model 70 architecture registers are on PowerPC 604? 1. 0x03C3 - Motherboard Enable VGA Video Subsystem 2. 0x4AE8 - Motherboard Enable 8514/A Video Subsystem 3. 0x0102 - Video Subsystem General Enable Thanks, Rob robtargosz@hotmail.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: light-weight processes in VxWorks Date: Tue, 01 Aug 2000 16:14:51 -0700 From: Ken Norlin Organization: NASA Dryden FRC Message-ID: <398759EB.73B27DB0@dfrc.nasa.gov> References: <8m7eok$fs7$1@nnrp1.deja.com> syee@connex.com wrote: > > I wanted to know if vxWorks has functions to create and execute LWPs. > If not doesn anyone know a way of doing this in vxWorks? > Try taskSpawn -- to make it a light-weight process, just reduce the stacksize (& maybe the priority), eh? - -- Ken Norlin --- Simulation Engineer NASA Dryden Flight Research Center ken.norlin@dfrc.nasa.gov (661) 276-2046 FAX (661) 276-2792 --------------------------- Newsgroups: comp.os.vxworks Subject: light-weight processes in VxWorks Date: Tue, 01 Aug 2000 21:17:14 GMT From: syee@connex.com Organization: Deja.com - Before you buy. Message-ID: <8m7eok$fs7$1@nnrp1.deja.com> I wanted to know if vxWorks has functions to create and execute LWPs. If not doesn anyone know a way of doing this in vxWorks? Sent via Deja.com http://www.deja.com/ Before you buy. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Bothered by DHCP Date: Wed, 02 Aug 2000 01:01:22 GMT From: ellin_lin@my-deja.com Organization: Deja.com - Before you buy. Message-ID: <8m7rt0$pf9$1@nnrp1.deja.com> References: <8lhhgl$l6o$1@nnrp1.deja.com> <3985168C.B68B169F@lucent.com> <8m629l$di4$1@nnrp1.deja.com> I modified both usrNetwork.c and usrNetDhcpsCfg.c adding entries for DHCP as below: DHCPS_LEASE_DESC dhcpsLeaseTbl [] = {... {"myentry","192.168.1.10","192.168.1.11","snmk=255.255.255.0"} } The IP of the server is "192.168.1.123". When the server started-up, Message "DHCP Server started" was given, so was "DHCP Server read 3 entries from the database". But the Win98 dhcp client's IP is "169.254.XX.XX", "XX" is random. Is there anyting else I need to do? Best regards. In article <8m629l$di4$1@nnrp1.deja.com>, ellin_lin@my-deja.com wrote: > In article <3985168C.B68B169F@lucent.com>, > Yusuf Motiwala wrote: > > DHCP server IP pool is also defined in usrNetDhcpsCfg.c. So probably > you > > want to make sure that it picks up from the correct file. > > > > Yusuf > > > > ellin_lin@my-deja.com wrote: > > > > > I'm a tenderfoot of vxWorks and Tornado. And I > > > met some problems when I tried to configured a > > > DHCP server. I modified usrnetwork.c setting the > > > IP range as written in user manual, and test this > > > DHCP server with a DHCP client of Win98, it did > > > nothing. Can someone give me some guidance? > > > > > > Sent via Deja.com http://www.deja.com/ > > > Before you buy. > > > > > > Sent via Deja.com http://www.deja.com/ > Before you buy. > Sent via Deja.com http://www.deja.com/ Before you buy. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: raw keyboard input to Tornado shell process? Date: Tue, 01 Aug 2000 18:11:22 -0700 From: "Michael R. Kesti" Organization: MK Associates Message-ID: <3987753A.88BA58F3@gv.net> References: <8m53v7$oco$1@nnrp1.deja.com> <3986F736.3F61C561@dfrc.nasa.gov> Reply-To: mkesti@gv.net Ken Norlin wrote: >We've used VxCurses at my site for a few years now -- it allows you to >get a character (or keystroke I think) at a time. You could check that >code to see how this is done. (It's been years now since I grabbed a copy -- >I think it was over on the Wind River ftp site). It can be found at ftp://ftp.atd.ucar.edu/pub/vxworks/vx/ - -- ======================================================================== Michael Kesti | "And like, one and one don't make | two, one and one make one." mkesti@gv.net | - The Who, Bargain --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Bothered by DHCP Date: Wed, 02 Aug 2000 01:14:40 GMT From: ellin_lin@my-deja.com Organization: Deja.com - Before you buy. Message-ID: <8m7slt$pu0$1@nnrp1.deja.com> References: <8lhhgl$l6o$1@nnrp1.deja.com> <3985168C.B68B169F@lucent.com> I modified both usrNetwork.c and usrNetDhcpsCfg.c adding entries for DHCP as below: DHCPS_LEASE_DESC dhcpsLeaseTbl [] = {... {"myentry","192.168.1.10","192.168.1.11","snmk=255.255.255.0"} } The IP of the server is "192.168.1.123". When the server started-up, Message "DHCP Server started" was given, so was "DHCP Server read 3 entries from the database". But the Win98 dhcp client's IP is "169.254.XX.XX", "XX" is random. Is there anyting else I need to do? Best regards. In article <3985168C.B68B169F@lucent.com>, Yusuf Motiwala wrote: > DHCP server IP pool is also defined in usrNetDhcpsCfg.c. So probably you > want to make sure that it picks up from the correct file. > > Yusuf > > ellin_lin@my-deja.com wrote: > > > I'm a tenderfoot of vxWorks and Tornado. And I > > met some problems when I tried to configured a > > DHCP server. I modified usrnetwork.c setting the > > IP range as written in user manual, and test this > > DHCP server with a DHCP client of Win98, it did > > nothing. Can someone give me some guidance? > > > > Sent via Deja.com http://www.deja.com/ > > Before you buy. > > Sent via Deja.com http://www.deja.com/ Before you buy. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: interrupt: dc0 - no carrier Date: Wed, 02 Aug 2000 02:04:19 GMT From: paulvroberts@my-deja.com Organization: Deja.com - Before you buy. Message-ID: <8m7vj1$rto$1@nnrp1.deja.com> References: <39871557.240F928A@swri.edu> In article <39871557.240F928A@swri.edu>, Greg Willden wrote: > I am using a Motorola MVME2307 card. As the board boots up the > following line periodically gets output through the serial line. Also, > after the board boots up and I telnet to it and the same line is > repeatedly sent out to the screen. However if I attach a target > server/shell to it in Tornado I do not see the error. > The error message is: > > interrupt: dc0 - no carrier > If you are running Tornado 2, then the following patch fixed "Dec 21x40 Auto negotiation" for us: SPR28573 Paul Roberts Sent via Deja.com http://www.deja.com/ Before you buy. --------------------------- Newsgroups: comp.arch.embedded,comp.os.vxworks,comp.sys.powerpc.tech Subject: Re: Looking for help with MPCxxx and NEC SDRAM Date: Wed, 02 Aug 2000 03:31:58 GMT From: Richard Hendricks Organization: Road Runner - Texas Message-ID: <398797A0.4CAE31DD@austin.rr.com> References: <8m47mq$63o$1@newsreaderg1.core.theplanet.net> Reply-To: rahendric@hotmail.com Try looking for the Interfacing to SDRAM PDF file I posted here a while back. If you can't find it, let me know and I'll repost it. Jeremy wrote: > > Has anyone any experience of SDRAM (NEC 4564163) > interfacing to MPC8xx. > This is very important. > Any information will be very gratefully recieved. - -- "Never apply a Star Trek solution to a Babylon 5 problem." --FermiLab --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Bothered by DHCP Date: Wed, 02 Aug 2000 10:52:15 +0530 From: Vijay Kumar Peshkar Organization: Ye 'Ol Disorganized NNTPCache groupie Message-ID: <3987B007.8DCD41AD@cisco.com> References: <8lhhgl$l6o$1@nnrp1.deja.com> <3985168C.B68B169F@lucent.com> <8m7slt$pu0$1@nnrp1.deja.com> Reply-To: vpeshkar@cisco.com This is a multi-part message in MIME format. - --------------7501B7503D36DA809779381B Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit I had once tried to do this setup...using my own net with vxworks floppy booted server and a 98 client. I don't remember clearly now, but think it has something to do with the initial handshake. Why not try building a floppy bootable client?. That way, you may be able to debug at each stage... good luck, Vijay ellin_lin@my-deja.com wrote: > I modified both usrNetwork.c and usrNetDhcpsCfg.c adding entries for > DHCP as below: > DHCPS_LEASE_DESC dhcpsLeaseTbl [] = > {... > {"myentry","192.168.1.10","192.168.1.11","snmk=255.255.255.0"} > } > The IP of the server is "192.168.1.123". > When the server started-up, Message "DHCP Server started" was given, so > was "DHCP Server read 3 entries from the database". > But the Win98 dhcp client's IP is "169.254.XX.XX", "XX" is random. > Is there anyting else I need to do? > > Best regards. > > In article <3985168C.B68B169F@lucent.com>, > Yusuf Motiwala wrote: > > DHCP server IP pool is also defined in usrNetDhcpsCfg.c. So probably > you > > want to make sure that it picks up from the correct file. > > > > Yusuf > > > > ellin_lin@my-deja.com wrote: > > > > > I'm a tenderfoot of vxWorks and Tornado. And I > > > met some problems when I tried to configured a > > > DHCP server. I modified usrnetwork.c setting the > > > IP range as written in user manual, and test this > > > DHCP server with a DHCP client of Win98, it did > > > nothing. Can someone give me some guidance? > > > > > > Sent via Deja.com http://www.deja.com/ > > > Before you buy. > > > > > > Sent via Deja.com http://www.deja.com/ > Before you buy. - --------------7501B7503D36DA809779381B Content-Type: text/x-vcard; charset=us-ascii; name="vpeshkar.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Vijay Kumar Peshkar Content-Disposition: attachment; filename="vpeshkar.vcf" begin:vcard n:Kumar.P;Vijay tel;work:080-5530035/36 Extn:3464 x-mozilla-html:FALSE org:Wipro Technologies, K2, Bangalore.;Wishing you blue skies, a couple of daisies, some fresh air and a smile! adr:;;;;;; version:2.1 email;internet:vpeshkar@cisco.com title:Senior Software Engineer note;quoted-printable: o o o o o o o . . . _____________________________________________=0D=0A o _____ || Vijay Kumar .P =20 |=0D=0A .] [__n_n_| DD[ =3D=3D=3D=3D_____ | vpeshkar@cisco.com, vijay_peshkar@hotmail.com | =0D=0A >(________|__|_[_________]_ |___________________________________________|=0D=0A _/oo OOOOO oo` ooo ooo 'o!o!o =20 o!o!o` fn:Vijay Kumar.P end:vcard - --------------7501B7503D36DA809779381B-- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: UDP problem Date: Wed, 02 Aug 2000 11:03:53 +0530 From: Vijay Kumar Peshkar Organization: Ye 'Ol Disorganized NNTPCache groupie Message-ID: <3987B2C1.80231233@cisco.com> References: <39760EC0.1EAEF7BF@139.85.103.45> <300DA502.63410AB6@mediaone.net> Reply-To: vpeshkar@cisco.com This is a multi-part message in MIME format. - --------------9194F70ECF25733334EF4B32 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit > Are you bench marking for performance reasons ?? If so there are other > means to send IP packets > through the physical layer other than using sockets. The VxWorks socket > layer is pretty slow. If you are interested I can tell how to go about > bypassing the vxWorks netWork layer for processing IP packets at a much > faster rate. > > --Steve Steve, Could you tell us more?. I am interested to know how? Thanks, Vijay - --------------9194F70ECF25733334EF4B32 Content-Type: text/x-vcard; charset=us-ascii; name="vpeshkar.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Vijay Kumar Peshkar Content-Disposition: attachment; filename="vpeshkar.vcf" begin:vcard n:Kumar.P;Vijay tel;work:080-5530035/36 Extn:3464 x-mozilla-html:FALSE org:Wipro Technologies, K2, Bangalore.;Wishing you blue skies, a couple of daisies, some fresh air and a smile! adr:;;;;;; version:2.1 email;internet:vpeshkar@cisco.com title:Senior Software Engineer note;quoted-printable: o o o o o o o . . . _____________________________________________=0D=0A o _____ || Vijay Kumar .P =20 |=0D=0A .] [__n_n_| DD[ =3D=3D=3D=3D_____ | vpeshkar@cisco.com, vijay_peshkar@hotmail.com | =0D=0A >(________|__|_[_________]_ |___________________________________________|=0D=0A _/oo OOOOO oo` ooo ooo 'o!o!o =20 o!o!o` fn:Vijay Kumar.P end:vcard - --------------9194F70ECF25733334EF4B32-- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Bothered by DHCP Date: Wed, 02 Aug 2000 11:42:13 +0530 From: Yusuf Motiwala Organization: Lucent Technologies, Columbus, Ohio Message-ID: <3987BBBD.DCF35DD1@lucent.com> References: <8lhhgl$l6o$1@nnrp1.deja.com> <3985168C.B68B169F@lucent.com> <8m7slt$pu0$1@nnrp1.deja.com> It means your DHCP server not responding. By any chance, you are using WDB task attached to ethernet? If yes, use serial because both (DHCP server as well as WDB) uses SNARF which is not allowed in VxWorks. First try removing WDB from VxWorks and then if it works switch to serial. Yusuf ellin_lin@my-deja.com wrote: > I modified both usrNetwork.c and usrNetDhcpsCfg.c adding entries for > DHCP as below: > DHCPS_LEASE_DESC dhcpsLeaseTbl [] = > {... > {"myentry","192.168.1.10","192.168.1.11","snmk=255.255.255.0"} > } > The IP of the server is "192.168.1.123". > When the server started-up, Message "DHCP Server started" was given, so > was "DHCP Server read 3 entries from the database". > But the Win98 dhcp client's IP is "169.254.XX.XX", "XX" is random. > Is there anyting else I need to do? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: multiple unit END Date: Wed, 02 Aug 2000 11:52:32 +0530 From: Yusuf Motiwala Organization: Lucent Technologies, Columbus, Ohio Message-ID: <3987BE27.D0DCD73F@lucent.com> References: <3986D06A.3796EBBE@americasm01.nt.com> Are you looking for multiple units or multiple ports support . Mutiple units can simply be supported by allocating new structure for every unit. However, for multiple ports within a single device, you need to modify template slightly or need to put a wrapper around it. Some other implementation for mutiple ports used fixed location in cluster to put the port information but it depend from one system to another. Yusuf "Leung, Spencer [SKY:1E13:EXCH]" wrote: > Hi, > > I'm trying to write an END for a 4 ports Ethernet chipset (MMC EPIF50). > The templateEnd.c only support single unit. Does anyone have a version > of templateEnd.c for multiple units or have an END written for the > EPIF50 or 200? Thanks. > > Spencer > > -- > ---------------------------------------------------------------------- > NORTEL SPENCER LEUNG leungsp@nortelnetworks.com > NETWORKS Data Systems Engineer Tel 613 765-4790 > 1E13 Fax 613 768-4904 > www.nortelnetworks.com Systems Development ESN 395 GMS 045/4P/E9 > ---------------------------------------------------------------------- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: [BSP] question ? Date: Wed, 02 Aug 2000 12:16:54 +0530 From: Vijay Kumar Peshkar Organization: Ye 'Ol Disorganized NNTPCache groupie Message-ID: <3987C3DE.31C04400@cisco.com> References: <8ku3pl$j9n@netnews.hinet.net> Reply-To: vpeshkar@cisco.com This is a multi-part message in MIME format. - --------------9A04CCC038DE55B2E595794B Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hi, If U are using Tornado 2, should update the path directory using the GUI...and add the new .c files name to the EXTRA_MODULES definition( assuming you have already compiled the rest of your files). good luck, Vijay kathy wrote: > If I want to change target directory to another, > How to do it to let compiler compile my new directory ? - --------------9A04CCC038DE55B2E595794B Content-Type: text/x-vcard; charset=us-ascii; name="vpeshkar.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Vijay Kumar Peshkar Content-Disposition: attachment; filename="vpeshkar.vcf" begin:vcard n:Kumar.P;Vijay tel;work:080-5530035/36 Extn:3464 x-mozilla-html:FALSE org:Wipro Technologies, K2, Bangalore.;Wishing you blue skies, a couple of daisies, some fresh air and a smile! adr:;;;;;; version:2.1 email;internet:vpeshkar@cisco.com title:Senior Software Engineer note;quoted-printable: o o o o o o o . . . _____________________________________________=0D=0A o _____ || Vijay Kumar .P =20 |=0D=0A .] [__n_n_| DD[ =3D=3D=3D=3D_____ | vpeshkar@cisco.com, vijay_peshkar@hotmail.com | =0D=0A >(________|__|_[_________]_ |___________________________________________|=0D=0A _/oo OOOOO oo` ooo ooo 'o!o!o =20 o!o!o` fn:Vijay Kumar.P end:vcard - --------------9A04CCC038DE55B2E595794B-- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks & ARM Dev Board Date: Wed, 02 Aug 2000 07:43:47 GMT From: john_94501@my-deja.com Organization: Deja.com - Before you buy. Message-ID: <8m8jfi$ai2$1@nnrp1.deja.com> References: <03db9226.c358038d@usw-ex0105-038.remarq.com> <8lu0p4$tpb$1@nnrp1.deja.com> <1b8aeca0.856a27a7@usw-ex0105-038.remarq.com> <8m1s3d$cvh$1@nnrp1.deja.com> <207a7325.b630c794@usw-ex0106-045.remarq.com> Hi, > [4] While we wait, I'm just wonderng if anyone knows the answer > to this one question: does the VxWorks boot PROM have support for > a link using the serial cable? Or do I need to make the boot > image again? It's not like it's a problem... but I'm just > wondering. I seem to remember that you can boot from TSFS (the target server file system), but you'll probably have to rebuild the bootrom since it will default to WDB_COMM_END (i.e. network), and doesn't include the TSFS boot support stuff by default. Take a look in the manuals for information about configuring a bootrom for this mode, and what to set the bootline to be (you just need the boot device as tsfs and the filename, all the rest can be ignored). HTH, John... Sent via Deja.com http://www.deja.com/ Before you buy. --------------------------- Newsgroups: comp.os.vxworks Subject: Dynamicaly loading an application symbols Date: Wed, 2 Aug 2000 09:11:40 +0200 From: "Arik Halperin" Organization: NetVision Israel Message-ID: <8m8e5k$3b9$1@news.netvision.net.il> Hi, I'm using the dynamic load library functions(loadLib) to load object files into my target using the following piece of code: mFd = open(moduleName,O_RDONLY,0644); if (mFd == ERROR) { logMsg("\ncannot open module file (line # %d)",lineNumber,0,0,0,0,0); fclose(lFp); return(ERROR); } pText = pData = pBss = LD_NO_ADDRESS; mid = loadModuleAt(mFd, LOAD_ALL_SYMBOLS, &pText, &pData, &pBss); if (mid == NULL) { logMsg("\ncannot load module (line # %d)",lineNumber,0,0,0,0,0); close(mFd); fclose(lFp); return(ERROR); } close(mFd); The loadModuleAt routine performs OK , But the symbols cannot be found using lkup or symFindByCName (They are not entered into the symbol table). I'm using Tornado 2.0 with an EBSA285 BSP(Arm 110 processor) The symbol table macros: INCLUDE_SYM_TBL,INCLUDE_SYM_TBL_INIT,INCLUDE_STANDALONE_SYM_TBL and SYM_TBL_HASH_SIZE_LOG is 9. When using the ld command from the shell everything works fine,but the code above does'nt! What to do? Regards, Arik Halperin --------------------------- Newsgroups: comp.os.vxworks Subject: File system ioctls Date: 31 Jul 2000 06:25:31 GMT From: dre052@preston.cmst.csiro.au (Ashley Dreier) Organization: CSIRO Manufacturing Science and Technology, Preston, Victoria, Australia Message-ID: <8m364r$ikb$1@commsun.its.csiro.au> Followup-To: comp.os.vxworks Hi there. I'm trying to implement a file system under VxWorks 5.1.1. This is to bypass some problems associated with the Dos file system, etc, as well as speed up the file writing. I've been searching for references to the ioctl commands, in particular to describe the expected return values and the argument types. I've not had much luck so far. I can write test routines which exercise the ioctl commands as I have defined them, but I have encountered problems when system functions call these ioctls (e.g. lseek). I'm bascially trying to support all of the ioctl commands that the Dos file system supports. If anyone can give me references to the arguments for ioctls and associated parameters (int vs long vs int*, etc) it would be very much appreciated. Email or news posting aare both okay. Thanks very much. Regards, Ash. - -- Ashley J. Dreier. Email: Ashley.Dreier@cmst.csiro.au CSIRO Manufacturing Science and Tech, Phone: +61 3 9662 7799 Locked Bag No. 9, Fax: +61 3 9662 7851 Preston, Vic, 3072, Australia. URL: http://www.msa.cmst.csiro.au/ourstaff/AshleyDreier/AshleyDreier.html --------------------------- Newsgroups: comp.os.vxworks Subject: Re: ATA/CompactFlash on a PPC860 Date: 2 Aug 2000 10:07:48 GMT From: Geir Frode Raanes Organization: UNINETT news service Message-ID: <8m8rtk$p19$1@snipp.uninett.no> References: <8m7htm$1pml$1@thoth.cts.com> Sender: Geir Frode Raanes Steve Jones wrote: : I am trying to get the ATA driver up and running on my PPC860. I am setting : up the PCMCIA interface in the PPC, and then initializing the ATA driver : over the top of it. The reference I have is a little different it is using : the Winbond chip over a PCI bus. I have made the changes to the driver but : am a bit confused on the addressing to the register interface of an ATA : device. Is there any one out their that has added a CompactFlash card to : there PPC project? I specifically asked Windriver about this. The MBX860 BSP only supports SRAM devices on the PCMCIA interface. That is, *not* ATA deices. Thus, if you want ATA on MBX860 and do not want to write or solder anything yourself, you have to use the standard version of the board and have fun with the ISA bus bridge and the 8259 interrupt controllers. BTW, there is a bug in the memory mapping of the ISA bridge that prohibits the ATA drive from working out of the box. Don't really know what the bug was, though. Basically, the MBX860 BSP lack the card services required for anything but default SRAM devices. - -- ****************************************************** Never ever underestimate the power of human stupidity. -Robert Anson Heinlein GeirFRS@invalid.and.so.forth ****************************************************** --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Wed Aug 2 08:50:09 2000 From: "Sinn, Larry" Date: Wed Aug 2 08:50:11 PDT 2000 Subject: Re: VMEbus reset on boot? We're using a mvme2304 and when ever we do a VME Reset thru the Universe Chip, vxWorks reboots. The only way we could get vxWorks to issue a reset was to detect a cold boot, do the VME Reset, and then the warm reboot would not do the VME Reset. Larry. comments below... On Mon, 31 Jul 2000, Lee DeRaud wrote: > On Mon, 31 Jul 2000 09:54:44 -0500, "Mark P. Fishman" > wrote: > >We would like to have all cards on the VME bus reset whenever we > >(re)boot vxWorks. Is there a way to do this in software? either > >with a target-shell command from a startup script or from a > >vxWorks function call? > This sounds a bit backwards: usually you're (re)booting VxWorks as a > *result* of a VME bus reset, not the other way around. Check the > documentation for the VME interface chipset used by your processor: > most of them have a way to strobe the SYSRESET signal. I know it sounds backwards; the problem is that we want to ensure that all the other hardware is reset to a known state when we so a warm reboot -- we're still in development. A bus reset (without rebooting the CPU board or cycling power) seems like the way to do this. Cheers -- Mark F. - 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 Wed Aug 2 08:59:35 2000 From: David Anderson Date: Wed Aug 2 08:59:37 PDT 2000 Subject: Re: TFTP server problem on DOS FS 2.0 vxworks vxworks Tornado >Hi, > >if anybody can help, feel free. > > >System setup: >- ------------- >1) Board A with TFFS and DOSFS2.0, the DOSFS device is called "/RFA/", > - sub directory "/RFA/dir_1/dir_2/" contains a number (approx. 8) of > binary files and an ASCII file containing the file names of the > the binary files. > >2) Board B (up to 6) requesting download of binary files. > The flow is like that: The mentioned files are ftp'ed from a Sun workstation > onto Board A in the named directory. Board A sends a message to all > connected boards B (via a socket) containing the complete directory name > and the name of the ASCII file in that directory. Each of the boards B > open a TFTP session and request the binary files. X-------------X >Problem: >- -------- >Board A runs a TFTP server, allowing simultaneous connections. >If more than one B-Board connects to the TFTP server in parallel >an error occurs. > >Information from the target shell: >- ---------------------------------- >- -> i >... >tTftpRRQ tftpdFileRea f2e198 100 SUSPEND 33a020 f2e128 d0003 0 >tTftpRRQ tftpdFileRea f30ab0 100 SUSPEND 33a020 f30a40 c0006 0 >tTftpRRQ tftpdFileRea f2b880 100 SUSPEND 33a020 f2b810 d0003 0 >... I think you have a problem with opening a single file simultaneously with multiple opens(...) in the tftp server. I am assuming here that all boards are booting the same file. Possibly could also occur if accessing same directory but this would be very unusual and would suggest something very strange in the O/S. If the tftp server opens a file for read and write ( rather than just for read ) then the filing system will NOT let the file be opened more than once. Try ensuring that each board is trying to load a unique file from the tftp server and see if that works OK. Your ascii file seems to be opened/loaded by multiple boards which could be the problem. You will either need to do the above or fix the tftp server code. The errors above suggest that the tftp server does not correctly check that a file can be opened. Probably only checks that it exists e.g. if ( file requested exists ) start data tranfer start data transfer open file ................ succeeds first time only, tranfer data close file David Anderson. mailto:danderson@dtrack.demon.co.uk ----------------------------------------------------- Who said "There's no life here in outer space." ----------------------------------------------------- Disclaimer ---------- This email and any files transmitted with it are confidential and intended solely for the use of the individual or entity to whom they are addressed. If you have received this email in error please destroy and notify Data Track Technology Plc +44 1425 271900. ------------------------------------------------------- From vxwexplo-errs@csg.lbl.gov Wed Aug 2 15:21:11 2000 From: "Jason Ahmad" Date: Wed Aug 2 15:21:13 PDT 2000 Subject: dos fs access and ct69000 drivers Hi, I'm currently working on a project which requires MS-DOS fs disk access under vxworks and have run into a few problems setting things up. As it stands right now, my x86 target boots off ide0 (with the boot file specified as "/ata0/vxWorks") but no FD is created to access the drive once everything is up and running. Both ATA and MS-DOS drivers are compiled into the image, and a quick check of iosDrvShow confirms the dosFs drivers are up and running. In an attempt to setup an FD, I ran ataDevCreate(0,0,0,0); and used the resulting address in dosFsDevInit("cf", blkdev, 0); however, dosFsDevInit consistantly returns 0. I cannot seem to get past this step. The one thing I can think of is I may be using the wrong value for the final parameter of ataDevCreate. (The programmers guide mentions that you need to specify a block offset for dos disks with partitions, but I have no idea what that offset should be for my disk.) Does anyone have any insight as to how to figure out what the correct offset should be and or get the disk mounted with a Dev handle? In addition, I was wondering about the availability of Chips and Technology 69000 video drivers. I heard a rumor that the 65555 drivers included will work, but I have had no luck with them. Thanks in advance, Jason Ahmad From vxwexplo-errs@csg.lbl.gov Wed Aug 2 19:55:11 2000 From: "G. Nagaraj" Date: Wed Aug 2 19:55:13 PDT 2000 Subject: RE: Using Single Serial Line with VxWorks Thanks Bo. I have tried even this. There is one more place in usrConfig.c, where the check is made for CONSOLE_TTY and WDB_TTY_CHANNEL being same. I tried defining CONSOLE_TTY and WDB_TTY_CHANNEL the same (zero) and tried disabling this check in usrConfig.c. I get console printing until the "wdbConfig" is called after which nothing is getting printed. All printings start only after connecting the target server and then opening the host shell and using the virtual i/o. Nagaraj -----Original Message----- From: Yu Bo-BOYU1 [mailto:Bo_Yu-BOYU1@email.mot.com] Sent: Thursday, August 03, 2000 1:10 AM To: 'g.nagaraj@wipro.com' Subject: Re: Using Single Serial Line with VxWorks Nagaraj, INCLUDE_WDB_BANNER are disabled in bootConfig.c if you define #if ((WDB_COMM_TYPE == WDB_COMM_SERIAL) && \ (CONSOLE_TTY == NONE || CONSOLE_TTY == WDB_TTY_CHANNEL)) Apparently none of Tornado manuals explain that clearly. It gave me a lot of trouble when I tried to make WDB_SERIAL work on a single serial port. Also, when you start the target server, the virtual i/o redirection needs to be checked. Bo G.N.Raj wrote in message news:<8lrgdp$2kc$1@nnrp1.deja.com>... > Yes. I selected WDB_BANNER > > Nagaraj > > In article <8lrf2f$qdk$1@news1.xs4all.nl>, > johan.borkhuis@peektraffic.nl wrote: > > g.nagaraj@wipro.com wrote: > > : I am using EV386EX BSP with Tornado II. I have my custom board. I > have > > : successfully ported the EV386EX BSP to my custom board. I have > modified > > : the "i8250Sio.c" for my board and I can see the printing on my > serial > > : console also. But when I configure the WDB on the serial using > > : WDB_SERIAL, I can't see any message on the serial console. As per > the > > : manual, I am supposed to see the WDB Banner, after which I can > > : disconnect the serial console and attach the WDB target server and > work > > : with the host shell. But, the WDB banner does not appear on the > > : console. But the strange thing is that, if I connect the WDB target > > : server, it successfully connects and I can use the host shell > without > > : any problem. > > > > Did you select WDB_BANNER? > > > > Groeten, > > Johan > > > > -- > > o o o o o o o . . . ___________________________________ > > o _____ || Johan Borkhuis | > > .][__n_n_|DD[ ====_____ | johan@borksoft.xs4all.nl | > > >(________|__|_[_________]_|________________________________| > > _/oo OOOOO oo` ooo ooo 'o!o!o o!o!o` > > === VxWorks page: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html > === > > > > -- > G. Nagaraj > Senior Software Engineer > Wipro Technologies, > Bangalore, India > > > Sent via Deja.com http://www.deja.com/ > Before you buy. From vxwexplo-errs@csg.lbl.gov Thu Aug 3 04:03:08 2000 From: Vxworks Exploder Date: Thu Aug 3 04:03:11 PDT 2000 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Thu Aug 3 04:03:04 PDT 2000 Subject: Hdlc driver Subject: Writing Device Drivers Subject: Re: interrupt: dc0 - no carrier Subject: Test Subject: Re: Dynamicaly loading an application symbols Subject: test Subject: Re: Dynamicaly loading an application symbols Subject: Frequency of TimeStampCounter (TSC)? Subject: Re: PMCspan expansion board setup Subject: Problems w/ select() and timeout on mvme2700 Subject: Software archive Subject: Re: VxWorks 5.3.1 Subject: re: VMEbus reset on boot? Subject: re: TFTP server problem on DOS FS 2.0 Subject: Re: Writing Device Drivers Subject: gcc-2.95.2-2, VxWorks, CygWin: Can't build it. Subject: Re: Libpcap vxWorks??? BPF?? Subject: Get rid of physical console terminal Subject: Re: Get rid of physical console terminal Subject: What is the best way to orgenize a new project? Subject: Re: TrueFFS mountFTL Subject: trcStack function Subject: Mounting and/or booting Zip Drives Subject: POSIX vs. native VxWorks queues Subject: Re: verbosity macro Subject: why stdin only sometimes has a `vi` interface Subject: dos fs access and ct69000 drivers Subject: BSP from Tornado 1.01 to 2.0 Subject: Re: Configure BSP fails Subject: Re: m Enter Ctrl+D in Tornado Shell Subject: Re: Get rid of physical console terminal Subject: Re: raw keyboard input to Tornado shell process? Subject: Re: verbosity macro Subject: Re: Software archive Subject: RE: Using Single Serial Line with VxWorks Subject: Re: Using Single Serial Line with VxWorks Subject: Re: why stdin only sometimes has a `vi` interface Subject: Re: gcc-2.95.2-2, VxWorks, CygWin: Can't build it. Subject: Re: PCI Configuration Subject: Re: What is the best way to orgenize a new project? Subject: Re: Bothered by DHCP Subject: Re: Get rid of physical console terminal Subject: Re: TFTP server problem on DOS FS 2.0 Subject: Re: BSP from Tornado 1.01 to 2.0 Subject: Problem in booting from network Subject: alignment & endian Subject: Re: raw keyboard input to Tornado shell process? ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Hdlc driver Date: Wed, 2 Aug 2000 14:02:36 +0200 From: "davidz" Organization: NetVision Israel Message-ID: <8m8v0a$97q$1@news.netvision.net.il> Hi ! Can any body have a good driver for hdlc based on PowerPC ! Best Regards David --------------------------- Newsgroups: comp.os.vxworks Subject: Writing Device Drivers Date: Mon, 31 Jul 2000 08:51:58 -0400 From: "Tim Stanka" Organization: Eastman Kodak Company Message-ID: <8m95or$720$1@news.news.kodak.com> I want to write my first VXWORKS device driver. I have a fair top-level understanding of the task but I find the two books that discuss drivers (VXWORKS Reference Manual and BSP Developer Manual) don't have alot of detail. Does anyone know or have documentation that would help me out. Thanks --------------------------- Newsgroups: comp.os.vxworks Subject: Re: interrupt: dc0 - no carrier Date: Wed, 02 Aug 2000 08:00:27 -0500 From: Greg Willden Organization: Posted via Supernews, http://www.supernews.com Message-ID: <39881B6B.35D850EE@swri.edu> References: <39871557.240F928A@swri.edu> <398722F3.5800D758@aps.anl.gov> I forced the port to half duplex and it did the trick. Thank you for your help Greg Andrew Johnson wrote: > Greg Willden wrote: > > > > The error message is: > > > > interrupt: dc0 - no carrier > > We used to get this I think with the combination of Tornado 1.0.1, the > SENS IP stack and a 100Mbps full duplex connection. Change any one of > those variables and it goes away, ie: Reinstall Tornado 1.0.1 without > installing the SENS IP stack, or install Tornado 2.0, or force your > network hub to half-duplex for this port. There may also be a fix > available from WRS but I didn't bother asking myself... > > - Andrew > -- > Sharks kill 10 people each year. People kill 60,000,000 sharks each year. --------------------------- Newsgroups: comp.os.vxworks Subject: Test Date: Wed, 2 Aug 2000 14:22:49 +0100 From: "Dougie Coull" Message-ID: <39881ff9@pull.gecm.com> References: <8m7htm$1pml$1@thoth.cts.com> <8m8rtk$p19$1@snipp.uninett.no> Test --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Dynamicaly loading an application symbols Date: Wed, 2 Aug 2000 13:26:44 GMT From: lee.deraud@boeing.com (Lee DeRaud) Organization: None to speak of Message-ID: <39881f17.1153759@news.boeing.com> References: <8m8e5k$3b9$1@news.netvision.net.il> Sender: nntp@news.boeing.com (Boeing NNTP News Access) On Wed, 2 Aug 2000 09:11:40 +0200, "Arik Halperin" wrote: >Hi, > I'm using the dynamic load library functions(loadLib) to load object >files into my target using the following piece of code: > > mFd = open(moduleName,O_RDONLY,0644); > if (mFd == ERROR) { > logMsg("\ncannot open module file (line # %d)",lineNumber,0,0,0,0,0); > fclose(lFp); > return(ERROR); > } > > pText = pData = pBss = LD_NO_ADDRESS; > mid = loadModuleAt(mFd, LOAD_ALL_SYMBOLS, &pText, &pData, &pBss); > if (mid == NULL) { > logMsg("\ncannot load module (line # %d)",lineNumber,0,0,0,0,0); > close(mFd); > fclose(lFp); > return(ERROR); > } > close(mFd); > >The loadModuleAt routine performs OK , But the symbols cannot be found using >lkup or symFindByCName >(They are not entered into the symbol table). > >I'm using Tornado 2.0 with an EBSA285 BSP(Arm 110 processor) The symbol >table macros: >INCLUDE_SYM_TBL,INCLUDE_SYM_TBL_INIT,INCLUDE_STANDALONE_SYM_TBL >and SYM_TBL_HASH_SIZE_LOG is 9. > >When using the ld command from the shell everything works fine,but the code >above does'nt! > >What to do? Looks like you're doing the right stuff, although the INCLUDE_SYM_* macros only affect the generation of the kernel (i.e. they're not needed when you compile moduleName or its components). Assuming you're linking moduleName correctly (which you must be or ld wouldn't work from the shell), how about just calling "ld 1,1,moduleName" and then use symFindByName to get the segment addresses? (They'll be named "moduleName_text", etc) Lee --------------------------- Newsgroups: comp.os.vxworks Subject: test Date: Wed, 2 Aug 2000 14:44:28 +0100 From: "Dougie Coull" Message-ID: <3988250b$1@pull.gecm.com> test --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Dynamicaly loading an application symbols Date: 2 Aug 2000 14:14:47 GMT From: johan@borksoft.xs4all.nl Organization: XS4ALL Internet BV Message-ID: <8m9acn$1q7$1@news1.xs4all.nl> References: <8m8e5k$3b9$1@news.netvision.net.il> Arik Halperin wrote: : The loadModuleAt routine performs OK , But the symbols cannot be found using : lkup or symFindByCName : (They are not entered into the symbol table). Are you using the host-shell or the target shell to execute lkup? If you are using the host shell try using the "@lkup" command (so add a @ in front of it). When using the host shell the symbol table on the host is used. When you execute your program the symbol table on the target gets modified. You can avoid this problem by using the "Synchronize host and target symbol tables" option in your BSP configuration. Groeten, Johan - -- o o o o o o o . . . ___________________________________ o _____ || Johan Borkhuis | .][__n_n_|DD[ ====_____ | johan@borksoft.xs4all.nl | >(________|__|_[_________]_|________________________________| _/oo OOOOO oo` ooo ooo 'o!o!o o!o!o` === VxWorks page: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html === --------------------------- Newsgroups: comp.os.vxworks Subject: Frequency of TimeStampCounter (TSC)? Date: Wed, 02 Aug 2000 16:28:30 +0200 From: Albert Wijnja Organization: MECO Equipment Engineers BV Message-ID: <3988300E.4DE80833@meco.nl> Reply-To: Albert.Wijnja@meco.nl Hello, In my application I use the pentiumTscGet64() routine to get the current counter value of the TSC. Is there anyway to determine the frequency of this counter? I use this routine for timestamping some events in my application. Eventually I want to represent the events in (micro)-seconds. Thanks Albert Wijnja --------------------------- Newsgroups: comp.os.vxworks Subject: Re: PMCspan expansion board setup Date: Wed, 2 Aug 2000 11:04:51 -0400 From: Minbo Shim Organization: GDRS Message-ID: <965228898.88514121@news.qis.net> References: <39870699.82EB6124@club-internet.fr> Followup-To: comp.os.vxworks Sebastien, Thanks for your reply. I have defined INCLUDE_PMC_SPAN in config.h in the BSP. I wonder if you have changed something else other than this. Some folks said that I need to change P2P_IO_BASE which is 0x01060000 as default. That's because I am using EXTENDED_VME and the address can not be mapped in the PCI IO space (0xFE000000 - 0xFE7FFFFF). So I tried SPAN_IO_DEV_SPACE for P2P_IO_BASE, which is 0x80000 in mv2400.h. But none of those changes have affected positively to the PMC cards (I/O based devices) in the extensioin board. I just wonder if there is something that I have to do on the PCI- to-PCI bridge configuration. The following is the current type 1 pci configuration setup that is supposed to initialize the DEC 21150 bridge chip for memory, io and configuration forwarding: vendor ID = 0x1011 device ID = 0x0022 command register = 0x0007 status register = 0x0290 revision ID = 0x04 class code = 0x06 sub class code = 0x04 programming interface = 0x00 cache line = 0x08 latency time = 0xff header type = 0x01 BIST = 0x00 base address 0 = 0x00000000 base address 1 = 0x00000000 primary bus number = 0x00 secondary bus number = 0x01 subordinate bus number = 0x01 secondary latency timer = 0xff IO base = 0x51 IO limit = 0x51 secondary status = 0x2280 memory base = 0xfd10 memory limit = 0xfd10 prefetch memory base = 0xfd91 prefetch memory limit = 0xfd91 prefetch memory base upper = 0x00000000 prefetch memory limit upper = 0x00000000 IO base upper 16 bits = 0x0000 IO limit upper 16 bits = 0x0000 expansion ROM base address = 0x00000000 interrupt line = 0x00 interrupt pin = 0x00 bridge control = 0x0000 Any suspicions on this? Thanks, Minbo On Tue, 01 Aug 2000, Sebastien Fauris wrote: >Hi, > >I have used a Ramix PMC-237 expander on a PowerCore750 >motherboard and also a Motorola PMC-Span on a MVME2300. > >Minbo Shim wrote: > >> Hi, >> >> I am trying to setup a PMC card in a PMC adapter carrier >> board from Motorola (PMCspan-002) which is attached to a >> MVME2431. The PMC card is in the slot 1 (device # 2) of >> the expansion board with interrupt level 3 (PMC_INT_LVL3). >> The DEC21150 PCI-to-PCI bridge chip provides the interface >> between the primary PCI bus and the secondary PCI bus >> through Type 0 and Type 1 configuration cycle. I am using >> CHRP Extended VME memory map for this configuration. >> >> My questions are: >> >> 1) can PMC_SPAN work with the CHRP memory model? > >There is no reason why it wouldn't work. It depends only of the >addresses used to set up the PCI to PCI bridge on the PMC Span. >On MVME2300, addresses used for the PMC Span can be found in mv2600.h. > >> >> >> 2) can I use some known pci setup functions like >> pciDevConfig(), pciConfigOutLong(), etc? >> > > absolutely, it is completely transparent to you the fact that the PCI >devices are on the primary PCI bus (bus #0) or secondary PCI bus >(bus#1). > >> >> 3) what is the physical address (hardware unmapped) of the >> PMC card as seen from the processor? > >> >> 4) what is the virtual address (MMU mapped)? >> >> Any help/comments on this would be greatly appreciated. >> >> -Minbo >> GDRS > >Same thing, this can be seen on a MVME2300 in mv2600.h. >You need to translate a PCI address to a CPU address and vice-versa. >You need also to swap the bytes (PPC is big endian, PCI is little >endian). > >Best of luck, > >Sebastien. --------------------------- Newsgroups: comp.os.vxworks Subject: Problems w/ select() and timeout on mvme2700 Date: Wed, 2 Aug 2000 10:55:37 -0400 From: "Jim Pisano" Organization: NRAO Message-ID: <8m9d1u$h91$1@newsfeed.cv.nrao.edu> I'm having a problem with select() on a TCP socket and timeouts. When I remove the timeout (i.e., set the pointer to the timeval struct to be 0)everything works correctly, but when I insert a non-null timeout into select(), it always returns 0, i.e., a timeout occurred w/out any file descriptors ready. I have set the timeout value to many seconds to ensure that the select() function waits long enough for data from the client app. I'm running VxWorks 5.4 on Tornado II hosted on Win NT. The target is an mvme-2700. here's a snippet of code: // create & open server socket (all details left out except for "accept()" call... m_ACC_ClientSocketFD = accept (m_listenSocket, (struct sockaddr *) &clientAddr, &sockAddrSize); // set up file descriptor set structure for select() call int m_width; fd_set m_fdSet; int m_ACC_ClientSocketFD; FD_ZERO( &m_fdSet ); FD_SET( m_ACC_ClientSocketFD, &m_fdSet ); m_width = m_ACC_ClientSocketFD + 1; // set up time out for 10 ms struct timeval timeout = {0,10000}; int selectVal = select( m_width, &m_fdSet, NULL, NULL, &timeout ); if( selectVal > 0 ) { bzero(buffer,100); if( read( m_ACC_ClientSocketFD, buffer, 100) > 0 ) { // process buffer ... } } Thanks. - -Jim --------------------------- Newsgroups: comp.os.vxworks Subject: Software archive Date: 2 Aug 2000 16:00:42 GMT From: cedilnik andrej Organization: University of Maryland, Baltimore County Message-ID: <8m9gja$t78$1@news.umbc.edu> Is there a software archive for VxWorks? I am looking for network software. For example is there a NTP server and client for VxWorks? Andy --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks 5.3.1 Date: Wed, 02 Aug 2000 11:03:22 -0500 From: Jimmie Marks Organization: Raytheon Company Message-ID: <3988464A.EC93014E@ftw.rsc.raytheon.com> References: <10yf5.545$_B6.49113@news1.onlynews.com> I beleive the defaults are : username: target password: password Poof wrote: > Anybody here know if there is a backdoor type username/password for VxWorks > 5.3.1? On the rLogin(514) port? > > Thanks! > > Me > (please reply even if you don't know.. GRIN thanks..) --------------------------- Newsgroups: comp.os.vxworks Subject: re: VMEbus reset on boot? Date: Wed, 2 Aug 2000 08:50:04 -0700 From: "Sinn, Larry" Organization: Lawrence Berkeley National Laboratory Message-ID: <8m9he6$g4q$1@overload.lbl.gov> We're using a mvme2304 and when ever we do a VME Reset thru the Universe Chip, vxWorks reboots. The only way we could get vxWorks to issue a reset was to detect a cold boot, do the VME Reset, and then the warm reboot would not do the VME Reset. Larry. comments below... On Mon, 31 Jul 2000, Lee DeRaud wrote: > On Mon, 31 Jul 2000 09:54:44 -0500, "Mark P. Fishman" > wrote: > >We would like to have all cards on the VME bus reset whenever we > >(re)boot vxWorks. Is there a way to do this in software? either > >with a target-shell command from a startup script or from a > >vxWorks function call? > This sounds a bit backwards: usually you're (re)booting VxWorks as a > *result* of a VME bus reset, not the other way around. Check the > documentation for the VME interface chipset used by your processor: > most of them have a way to strobe the SYSRESET signal. I know it sounds backwards; the problem is that we want to ensure that all the other hardware is reset to a known state when we so a warm reboot -- we're still in development. A bus reset (without rebooting the CPU board or cycling power) seems like the way to do this. Cheers -- Mark F. - - Larry Sinn KLA-Tencor MS I-1009 408 875 0247 voice 160 Rio Robles 408 875 6179 fax San Jose, Ca 95134-1809 larry.sinn@kla-tencor.com --------------------------- Newsgroups: comp.os.vxworks Subject: re: TFTP server problem on DOS FS 2.0 Date: Wed, 2 Aug 2000 16:50:24 +0100 From: David Anderson Organization: Lawrence Berkeley National Laboratory Message-ID: <8m9he6$g4r$1@overload.lbl.gov> vxworks vxworks Tornado >Hi, > >if anybody can help, feel free. > > >System setup: >- ------------- >1) Board A with TFFS and DOSFS2.0, the DOSFS device is called "/RFA/", > - sub directory "/RFA/dir_1/dir_2/" contains a number (approx. 8) of > binary files and an ASCII file containing the file names of the > the binary files. > >2) Board B (up to 6) requesting download of binary files. > The flow is like that: The mentioned files are ftp'ed from a Sun workstation > onto Board A in the named directory. Board A sends a message to all > connected boards B (via a socket) containing the complete directory name > and the name of the ASCII file in that directory. Each of the boards B > open a TFTP session and request the binary files. X-------------X >Problem: >- -------- >Board A runs a TFTP server, allowing simultaneous connections. >If more than one B-Board connects to the TFTP server in parallel >an error occurs. > >Information from the target shell: >- ---------------------------------- >- -> i >... >tTftpRRQ tftpdFileRea f2e198 100 SUSPEND 33a020 f2e128 d0003 0 >tTftpRRQ tftpdFileRea f30ab0 100 SUSPEND 33a020 f30a40 c0006 0 >tTftpRRQ tftpdFileRea f2b880 100 SUSPEND 33a020 f2b810 d0003 0 >... I think you have a problem with opening a single file simultaneously with multiple opens(...) in the tftp server. I am assuming here that all boards are booting the same file. Possibly could also occur if accessing same directory but this would be very unusual and would suggest something very strange in the O/S. If the tftp server opens a file for read and write ( rather than just for read ) then the filing system will NOT let the file be opened more than once. Try ensuring that each board is trying to load a unique file from the tftp server and see if that works OK. Your ascii file seems to be opened/loaded by multiple boards which could be the problem. You will either need to do the above or fix the tftp server code. The errors above suggest that the tftp server does not correctly check that a file can be opened. Probably only checks that it exists e.g. if ( file requested exists ) start data tranfer start data transfer open file ................ succeeds first time only, tranfer data close file David Anderson. mailto:danderson@dtrack.demon.co.uk - ----------------------------------------------------- Who said "There's no life here in outer space." - ----------------------------------------------------- Disclaimer - ---------- This email and any files transmitted with it are confidential and intended solely for the use of the individual or entity to whom they are addressed. If you have received this email in error please destroy and notify Data Track Technology Plc +44 1425 271900. - ------------------------------------------------------- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Writing Device Drivers Date: Wed, 02 Aug 2000 12:02:41 -0500 From: Bob Irwin Organization: SIS Development Message-ID: <39885431.FD9B36CC@systemsinterface.com> References: <8m95or$720$1@news.news.kodak.com> Tim Stanka wrote: > > I want to write my first VXWORKS device driver. I have a fair top-level > understanding of the task but > I find the two books that discuss drivers (VXWORKS Reference Manual and BSP > Developer Manual) > don't have alot of detail. Does anyone know or have documentation that would > help me out. > > Thanks vxWorks Programmer's Guide - -- Bob Irwin | bobi@sis-dev.com | SIS Development | 281-484-6661 | Houston --------------------------- Newsgroups: comp.os.vxworks Subject: gcc-2.95.2-2, VxWorks, CygWin: Can't build it. Date: Wed, 2 Aug 2000 16:38:50 +0100 From: "Dave Korn" Organization: (Posted via) U-NET Internet Ltd. Message-ID: <%BYh5.1300$Vc.882168@newsr2.u-net.net> Hi everyone, I'm being driven slowly out of my mind. I am trying to build gcc-2.95.2-2 for the powerpc-wrs-vxworks target on a PC using NT4.0sp6 and a fairly recent (1.1.2) download of cygwin. It goes dreadfully wrong, no matter what I try. I've searched various mailing list archives (gcc-help, crossgcc, cygwin), and found that lots of people have been having trouble with it under all sorts of platforms, but no mention of cygwin. Has *anybody* managed to build this target under cygwin? Anyone at all? My main problem at the moment is to do with errors compiling choose-temp.c in libiberty which is suffering some sort of header conflict that prevents it knowing the definition of a va_list: /usr/home/dkorn/gcc-2.95.2-2/libiberty/../include/libiberty.h:193: parse error before `va_list' ...despite having successfully completed the fixincs step. And that's not all... /usr/home/dkorn/gcc-2.95.2-2/libiberty/choose-temp.c: In function `try': /usr/home/dkorn/gcc-2.95.2-2/libiberty/choose-temp.c:89: warning: implicit declaration of function `access' /usr/home/dkorn/gcc-2.95.2-2/libiberty/choose-temp.c: In function `choose_temp_base': /usr/home/dkorn/gcc-2.95.2-2/libiberty/choose-temp.c:111: warning: implicit declaration of function `getenv' /usr/home/dkorn/gcc-2.95.2-2/libiberty/choose-temp.c:138: warning: implicit declaration of function `mktemp' /usr/home/dkorn/gcc-2.95.2-2/libiberty/choose-temp.c: In function `make_temp_file': /usr/home/dkorn/gcc-2.95.2-2/libiberty/choose-temp.c:157: warning: function `getenv' was previously declared within a block /usr/home/dkorn/gcc-2.95.2-2/libiberty/choose-temp.c:192: warning: implicit declaration of function `strcat' /usr/home/dkorn/gcc-2.95.2-2/libiberty/choose-temp.c:200: warning: implicit declaration of function `close' I won't go into further detail right now, as I'm sure most people aren't interested. But if *anyone* out there has built this version, can you please drop me a line and tell me whether you had similar trouble, and what kind of configure options you used ? Thanks for your time, DaveK - -- What part of "Gestalt" don't you understand? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Libpcap vxWorks??? BPF?? Date: Wed, 02 Aug 2000 10:19:25 -0700 From: Eric Monsler Organization: The ISP formerly known as Best Message-ID: <3988581D.A7C1940C@radixtek.com> References: <8m6j6o$f5l$1@news.umbc.edu> cedilnik andrej wrote: > > Hi! > > So, VxWorks is supposed to have BSD 4.4 network. BSD uses BPF for > accessing stuff on the data link layer. > Well, in order to access BPF, we need to open file /dev/bpf???? > How des that apply to VxWorks, which does not have a filesystem? > > Andy > Andy, In general, when something (VxWorks) has libFOO (BSD socket) support, and the standard FOO library on UNIX uses BAR, I do not think that you can assume BAR support. Eric --------------------------- Newsgroups: comp.os.vxworks Subject: Get rid of physical console terminal Date: Wed, 02 Aug 2000 13:31:57 -0400 From: "Connolly, F. T. (Francis)" Organization: Ford Motor Company Message-ID: <39885B0D.19EE7E83@mail.ford.com> Hi, Does anyone using a Sun/Solaris (ultra 60/2.6) host (or perhaps more generically, any unix) know to use a Sun serial port (and an xterm window?) as the vxWorks console terminal (mvme167 and niCpu030)? I searched around a bit but did not find any straightforward info on how this might happen (I'm an amateur sys admin--only by necessity, so I don't really want to learn too much about solaris/unix serial ports). Most of the (straightforward) info I saw seems to be on how to connect an ascii terminal to a serial port, not the other way around--make the serial port look like an ascii terminal. Maybe it's pretty simple; I hope so. Thanks for any info/help. - -- Francis Connolly Technical Specialist FMC-FAO-PD-R&VT-P&AE-CAPE-APPTE-CEFET-P/T ET 313-594-4081 office (1115 GTL) fconnoll@ford.com 313-317-7404 fax CDS ID: fconnoll --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Get rid of physical console terminal Date: Wed, 02 Aug 2000 11:31:47 -0700 From: Ken Norlin Organization: NASA Dryden FRC Message-ID: <39886913.CE73E646@dfrc.nasa.gov> References: <39885B0D.19EE7E83@mail.ford.com> "Connolly, F. T. (Francis)" wrote: > > Does anyone using a Sun/Solaris (ultra 60/2.6) host (or perhaps more > generically, any unix) know to use a Sun serial port (and an xterm window?) as > the vxWorks console terminal (mvme167 and niCpu030)? Hi, I think what you are looking for is 'tip'. Ken N. Ken Norlin --- Simulation Engineer NASA Dryden Flight Research Center ken.norlin@dfrc.nasa.gov (661) 276-2046 FAX (661) 276-2792 --------------------------- Newsgroups: comp.os.vxworks Subject: What is the best way to orgenize a new project? Date: Wed, 02 Aug 2000 21:46:05 +0200 From: Ezra Koper Organization: Bezeq International Ltd. Message-ID: <39887A7D.89C0D6C0@nevision.net.il> Hello All. We are starting to convert an embedded equ using embeded-NT to VxWorks. the NT software use several task's and approximetly 40 files ( for example 5 files per each task). My question is what will be the best way to orgenaze the files under Tornado-2 work space windo? will it be a single project ? or multiple projects when one is bootble and the other downloadble ? flat file system with singe project? The idea was that each task will look as a folder hiding inside its relevent files. Any replay will be appreciated!! --------------------------- Newsgroups: comp.os.vxworks Subject: Re: TrueFFS mountFTL Date: Wed, 02 Aug 2000 19:03:42 GMT From: w_rick_pierpont@ms.bd.com Organization: Deja.com - Before you buy. Message-ID: <8m9raa$8gi$1@nnrp1.deja.com> References: <8m67v8$734$1@overload.lbl.gov> <8m7dht$4b6$1@news1.xs4all.nl> I was told by my FAE that TFFS is free when used with M-Systems DOC 2000. I'm still waiting for the ftllite.o file, but hopefully it will come soon. Rick Pierpont Software Engineer Becton Dickinson Biosciences In article <8m7dht$4b6$1@news1.xs4all.nl>, johan@borksoft.xs4all.nl wrote: > Palaniappan wrote: > : Hi, > : To provide these special algorithms, M-Systems > : FLite provides a flash translation layer files, which could possibly be a > : licensed version from WRS, or otherwise you could approach M- Systems FLite. > > For the VxWorks version of TFFS you have to contact WindRiver. I contacted > M-Systems once and they said that I had to contact WRS. > > : I had a thorough search on M-Systems FLite web site and found no downloads > : for NOR-FLASH Arrays. For the sake of understanding the problem, we just > : extracted the ftllite.o file from PPCCEC603gnuvx.a and added into > : PPC860gnuvx.a and the vxWorks image was successfully built, but when we > : attempt to format the TrueFFS, it returns the value -1. And hence, I > : concluded that we may need to know whether it's a licensed version or not. > > When you want to use TFFS you have to purchase it from WRS. The version in > the EC603 library is a "mistake", but you could use it of course. But > there is a seperate license policy for TFFS. > See also my website for info about this. > > Groeten, > Johan > > -- > o o o o o o o . . . ___________________________________ > o _____ || Johan Borkhuis | > .][__n_n_|DD[ ====_____ | johan@borksoft.xs4all.nl | > >(________|__|_[_________]_|________________________________| > _/oo OOOOO oo` ooo ooo 'o!o!o o!o!o` > === VxWorks page: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html === > Sent via Deja.com http://www.deja.com/ Before you buy. --------------------------- Newsgroups: comp.os.vxworks Subject: trcStack function Date: Wed, 2 Aug 2000 15:09:44 -0400 From: "news.idt.net" Organization: IDT (Best News In The World) Message-ID: <8m9rlu$cki@nnrp2.farm.idt.net> Hi: We would like to unwind the stack frame under exception conditions. While calling trcStack we are getting error "-> trcStack aborted: error in top frame". Any details regarding this would be appreciated. The code we are using for trc stack... /*************************************************************************** **** * * xxxTracePrint - stack trace print function * * RETURNS: OK or ERROR */ static void xxxTracePrint ( INSTR *caller, int func, int nargs, int *args ) { char buf [250]; int ix; int len = 0; len += sprintf (&buf [len], "%#10x: %#10x (", (int)caller, func); for (ix = 0; ix < nargs; ix++) { if (ix != 0) len += sprintf (&buf [len], ", "); len += sprintf (&buf [len], "%#x", args [ix]); } len += sprintf (&buf [len], ")\n"); LogMessage(buf); } /*************************************************************************** **** * * xxxTrace - stack trace * * RETURNS: OK or ERROR */ int xxxTrace ( int tId ) { REG_SET regs; if(taskIdVerify(tId) == ERROR) return(ERROR); int prevOSint = intLock( ); taskRegsGet (tId, ®s); trcStack (®s, (FUNCPTR) xxxTracePrint, tId); intUnlock(prevOSint); return (OK); } Thanks, Hitesh Patel --------------------------- Newsgroups: comp.os.vxworks Subject: Mounting and/or booting Zip Drives Date: Wed, 02 Aug 2000 19:35:52 GMT From: w_rick_pierpont@ms.bd.com Organization: Deja.com - Before you buy. Message-ID: <8m9t6o$a34$1@nnrp1.deja.com> Hi, I'm in the process of getting my hardware up and running and I've run into a few problems (surprised?). I'm using T2/NT/x86 I have no problem getting to an ATA hard drive or a floppy disk but getting to an internal ATA Iomega 250MB Zip drive has not been working. Using usrAtaConfig() hangs the process somewhere down in usrAtaPartition() so that is useless. I think that is related to the fact that a Zip disk is really more like a floppy disk. When I use ataDevCreate(0,1,0,0), I get a BLK_DEV * that _appears_ to be valid (although I don't agree with bd_nBlocks), but a subsequent call to dosFsDevInit("zip:", pBlkDev, NULL) hangs! The devs command in WindSh reports a zip: device but it doesn't work. Supposedly VxWorks supports Zip drives but I haven't seen it work yet. I'm also trying to boot VxWorks from this disk. I can get the VxWorks boot sector on there with the vxSys program but when it boots it reports that it can't find bootrom.sys (which is there AND contiguous). I know it can work because I took a homegrown utility to steal a boot sector off of a floppy disk and write it to the Zip then copied bootrom.sys to the Zip -- and it worked! This isn't a valid approach since floppies use FAT12 but it proved feasibility. Any ideas for either problem? Thanks in advance, Rick Pierpont Software Engineer Becton Dickinson Biosciences Sent via Deja.com http://www.deja.com/ Before you buy. --------------------------- Newsgroups: comp.os.vxworks Subject: POSIX vs. native VxWorks queues Date: Wed, 02 Aug 2000 16:15:35 -0500 From: Saurabh Tewari Organization: Alcatel U.S.A. Message-ID: <39888F77.2E777C37@usa.alcatel.com> Hi All, Does anyone have an idea of any performance hit I'd take if I used POSIX message queueing vs. VxWorks message queueing? Are there numbers available somewhere? Also, has anyone had experience putting 20,000 messages of 500 bytes each thru VxWorks (or POSIX) queues on VxWorks? Does the performance go down significantly? Thanks, Saurabh --------------------------- Newsgroups: comp.os.vxworks Subject: Re: verbosity macro Date: Wed, 02 Aug 2000 21:15:45 GMT From: John Fusco Organization: Deja.com - Before you buy. Message-ID: <8ma31t$epg$1@nnrp1.deja.com> References: <200008011632.RAA04904@dsl-2.tadpole.co.uk> <3987110A.F04218F7@club-internet.fr> FYI. The "..." in the #define looks like a GNU extension. It's a nice feature, but it's not standard C. It doesn't work on my Solaris compiler. In article <3987110A.F04218F7@club-internet.fr>, Sebastien Fauris wrote: > Hi David, > > I found out about my variable arguments problem. > Here is the solution: > > #define VERB_PNP(level, format, args...) \ > if (level <= pnp_dbg) \ > {printf (format, ## args);} > #else > #define VERB_PNP(level, format, args...) {} > > Cheers, > > Sebastien. > > Sent via Deja.com http://www.deja.com/ Before you buy. --------------------------- Newsgroups: comp.os.vxworks Subject: why stdin only sometimes has a `vi` interface Date: Wed, 02 Aug 2000 22:55:45 GMT From: p_lavarre@my-deja.com Organization: Deja.com - Before you buy. Message-ID: <8ma8te$jde$1@nnrp1.deja.com> I see that I get `vi`-style editing (e.g. Escape 0 $ moves the cursor back and forth) at the: -> prompt and at the: tcl> prompt and during TCL use of stdin (ended with Ctrl+D): ->?read stdin ... but I get no `vi` support during C use of stdin: ->getchar Why is that? I'm using VxWorks 2.0 with a host shell talking through a target server then via wdbserial to a target. Thanks in advance. Pat LaVarre p.lavarre@ieee.org http://members.aol.com/ppaatt/ Sent via Deja.com http://www.deja.com/ Before you buy. --------------------------- Newsgroups: comp.os.vxworks Subject: dos fs access and ct69000 drivers Date: Wed, 2 Aug 2000 15:21:37 -0700 From: "Jason Ahmad" Organization: Lawrence Berkeley National Laboratory Message-ID: <006901bffcd0$0761a4a0$41f122aa@muc> Hi, I'm currently working on a project which requires MS-DOS fs disk access under vxworks and have run into a few problems setting things up. As it stands right now, my x86 target boots off ide0 (with the boot file specified as "/ata0/vxWorks") but no FD is created to access the drive once everything is up and running. Both ATA and MS-DOS drivers are compiled into the image, and a quick check of iosDrvShow confirms the dosFs drivers are up and running. In an attempt to setup an FD, I ran ataDevCreate(0,0,0,0); and used the resulting address in dosFsDevInit("cf", blkdev, 0); however, dosFsDevInit consistantly returns 0. I cannot seem to get past this step. The one thing I can think of is I may be using the wrong value for the final parameter of ataDevCreate. (The programmers guide mentions that you need to specify a block offset for dos disks with partitions, but I have no idea what that offset should be for my disk.) Does anyone have any insight as to how to figure out what the correct offset should be and or get the disk mounted with a Dev handle? In addition, I was wondering about the availability of Chips and Technology 69000 video drivers. I heard a rumor that the 65555 drivers included will work, but I have had no luck with them. Thanks in advance, Jason Ahmad --------------------------- Newsgroups: comp.os.vxworks Subject: BSP from Tornado 1.01 to 2.0 Date: Wed, 2 Aug 2000 23:48:58 +0200 From: "Jörg Krein" Organization: Customer of UUNET Deutschland GmbH Message-ID: <398896f5$0$16020@personalnews.de.uu.net> Hello, cause I'm new with Tornado/VxWorks I have the folling question: I have to use a BSP made with Tornado 1.0.1. As Tornado 2.0 doesn't use usrConfig.c anymore I have to put my own hardware initialisiation somewere else. Does anyone know where? Is it possible to add own parts in the project window like WindRivers own parts? Thank you! --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Configure BSP fails Date: Wed, 2 Aug 2000 17:48:31 -0400 From: "John Turner" Organization: UUNET WorldCom server (post doesn't reflect views of UUNET WorldCom Message-ID: <8ma4v6$b9g$1@soap.pipex.net> References: <8m68pc$m4t$1@news1.tele.dk> "Henrik Husted" wrote in message news:8m68pc$m4t$1@news1.tele.dk... > Hi. > > When I select 'Project -> Configure BSP' I get the following error message: > > "WindCFG was unable to run the C preprocessor > and read the default configuration. > Check your environment or reinstall > Tornado" > > Hmm...I running from within the IDE (on a WinNT 4 sp6) and the documentation > states that I do not need to be concerned about setting up the environment > variables. Reinstalling Tornado didn't help much. > I'm a bit new to VxWorks but I need to build a BSP for my special in-house > hardware (based on a SH7410), right? > > TIA > - Henrik > I would call WRS or email support@windriver.com. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: m Enter Ctrl+D in Tornado Shell Date: Wed, 02 Aug 2000 23:26:41 GMT From: p_lavarre@my-deja.com Organization: Deja.com - Before you buy. Message-ID: <8maanb$ko5$1@nnrp1.deja.com> References: <8lsmkb$vh7$1@nnrp1.deja.com> In article <8lsmkb$vh7$1@nnrp1.deja.com>, p_lavarre@my-deja.com wrote: > Here, if at the Shell prompt I strike the three keys: > m Enter Ctrl+D > I get an Unhandled EXCEPTION_ACCESS_VIOLATION in TCL.DLL ... > > Is this interesting? Does it suggest I should muck around differently > with how I have std i/o or the target agent configured? Or is this > just life as normal with a Windows Millenium GUI? Ahhh. I grow positively dangerous. shellproc m appears in the Tornado/host/resource/tcl/ShellMemCmd.tcl source file. Your input reaches shellproc m via the Tcl expression [gets stdin]. That tells me `vi` editing is available, by the way, and suggests the experiment: ?gets stdin Enter Ctrl+D Indeed this likewise crashes Windsh.exe by way of Tcl.dll. I still have no idea why. I think I'll go ask in comp.lang.tcl too. Sent via Deja.com http://www.deja.com/ Before you buy. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Get rid of physical console terminal Date: Wed, 02 Aug 2000 17:20:21 -0500 From: Bob Irwin Organization: SIS Development Message-ID: <39889EA5.5EA17248@systemsinterface.com> References: <39885B0D.19EE7E83@mail.ford.com> <39886913.CE73E646@dfrc.nasa.gov> Connolly, F. T. (Francis)" wrote: > > Does anyone using a Sun/Solaris (ultra 60/2.6) host (or perhaps more > generically, any unix) know to use a Sun serial port (and an xterm window?) as > the vxWorks console terminal (mvme167 and niCpu030)? script to execute tip in an xterm window. ( use ~. to exit): xterm -bg seashell -geometry 80x24+0+0 -fn 7x13 -sb -sl 5000 -title 'tipterm, port A, 9600 baud' -e tip hardwire_a & and the hardwire_a entry in /etc/remote: hardwire_a:\ :dv=/dev/term/a:br#9600:el=^C^S^Q^U^D:ie=%$:oe=^D: - -- Bob Irwin | bobi@sis-dev.com | SIS Development | 281-484-6661 | Houston --------------------------- Newsgroups: comp.os.vxworks Subject: Re: raw keyboard input to Tornado shell process? Date: Wed, 02 Aug 2000 22:53:17 GMT From: p_lavarre@my-deja.com Organization: Deja.com - Before you buy. Message-ID: <8ma8or$jal$1@nnrp1.deja.com> References: <8m53v7$oco$1@nnrp1.deja.com> <3986F736.3F61C561@dfrc.nasa.gov> <3987753A.88BA58F3@gv.net> > > > raw keyboard input ... Tornado shell ... wdbserial target server > > Ken Norlin ... > > We've used VxCurses at my site for a few years now -- > > it allows you to get a character (or keystroke I think) at a time. > > You could check that code to see how this is done. > Michael Kesti ... > It can be found at ftp://ftp.atd.ucar.edu/pub/vxworks/vx/ Thank you folk for the pointer ... I dug a little, I'm now thinking @dfrc.nasa.gov is uses a target shell? Yes? If I'm wrong, then you have my answer, please tell me. If I'm right, then I need to keep looking: I don't have a target shell, I have only the host shell. So WHY am I thinking @dfrc.nasa.gov solves this by using a target shell? Well, what I found at the ftp:// link was vwcurses01, vwcurses02, ... vwcurses08. These are eight shell archives - containing more or less readable C source. In there, I see functions like raw and cbreak are built up as macros using stty. But here, lacking a target shell, I see: - -> lkup "stty" value = 0 = 0x0 - -> ? tcl> stty invalid command name "stty" tcl> Is there some way I can reach stty for the input to `windsh`? I mean, where is the code that supplies the `vi`-style editing for the command line? That code must have `stty` privileges. Is it a binary- code-only part of `windsh` or somewhere more accessible? Thanks again in advance. Pat LaVarre p.lavarre@ieee.org http://members.aol.com/ppaatt/ Sent via Deja.com http://www.deja.com/ Before you buy. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: verbosity macro Date: Wed, 2 Aug 2000 23:26:28 GMT From: hoh@invalid.invalid (Goran Larsson) Organization: [1] + 5934 done /bin/rm -rf ~/ & Message-ID: <2000Aug2.232628.10933@lorelei.approve.se> References: <200008011632.RAA04904@dsl-2.tadpole.co.uk> <3987110A.F04218F7@club-internet.fr> <8ma31t$epg$1@nnrp1.deja.com> Sender: hoh@lorelei.approve.se.NO_JUNK_EMAIL (Goran Larsson) In article <8ma31t$epg$1@nnrp1.deja.com>, John Fusco wrote: > The "..." in the #define looks like a GNU extension. It's a nice > feature, but it's not standard C. It doesn't work on my Solaris > compiler. True, in standard C it should be: #define VERB_PNP(level, ...) \ if (level <= pnp_dbg) \ {printf (__VA_ARGS__);} - -- Göran Larsson Senior Systems Analyst hoh AT approve DOT se --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Software archive Date: Wed, 02 Aug 2000 20:42:22 -0400 From: viv Organization: Lucent Technologies Message-ID: <3988BFEE.E5CA137F@lucent.com> References: <8m9gja$t78$1@news.umbc.edu> try ftp.atd.ucar.edu...also the faq has some more information...

cedilnik andrej wrote:

Is there a software archive for VxWorks?

I am looking for network software.
For example is there a NTP server and client for VxWorks?

                        Andy

--------------------------- Newsgroups: comp.os.vxworks Subject: RE: Using Single Serial Line with VxWorks Date: Thu, 3 Aug 2000 08:33:44 +0530 From: "G. Nagaraj" Organization: Lawrence Berkeley National Laboratory Message-ID: <8mao3l$bel$1@overload.lbl.gov> Thanks Bo. I have tried even this. There is one more place in usrConfig.c, where the check is made for CONSOLE_TTY and WDB_TTY_CHANNEL being same. I tried defining CONSOLE_TTY and WDB_TTY_CHANNEL the same (zero) and tried disabling this check in usrConfig.c. I get console printing until the "wdbConfig" is called after which nothing is getting printed. All printings start only after connecting the target server and then opening the host shell and using the virtual i/o. Nagaraj - -----Original Message----- From: Yu Bo-BOYU1 [mailto:Bo_Yu-BOYU1@email.mot.com] Sent: Thursday, August 03, 2000 1:10 AM To: 'g.nagaraj@wipro.com' Subject: Re: Using Single Serial Line with VxWorks Nagaraj, INCLUDE_WDB_BANNER are disabled in bootConfig.c if you define #if ((WDB_COMM_TYPE == WDB_COMM_SERIAL) && \ (CONSOLE_TTY == NONE || CONSOLE_TTY == WDB_TTY_CHANNEL)) Apparently none of Tornado manuals explain that clearly. It gave me a lot of trouble when I tried to make WDB_SERIAL work on a single serial port. Also, when you start the target server, the virtual i/o redirection needs to be checked. Bo G.N.Raj wrote in message news:<8lrgdp$2kc$1@nnrp1.deja.com>... > Yes. I selected WDB_BANNER > > Nagaraj > > In article <8lrf2f$qdk$1@news1.xs4all.nl>, > johan.borkhuis@peektraffic.nl wrote: > > g.nagaraj@wipro.com wrote: > > : I am using EV386EX BSP with Tornado II. I have my custom board. I > have > > : successfully ported the EV386EX BSP to my custom board. I have > modified > > : the "i8250Sio.c" for my board and I can see the printing on my > serial > > : console also. But when I configure the WDB on the serial using > > : WDB_SERIAL, I can't see any message on the serial console. As per > the > > : manual, I am supposed to see the WDB Banner, after which I can > > : disconnect the serial console and attach the WDB target server and > work > > : with the host shell. But, the WDB banner does not appear on the > > : console. But the strange thing is that, if I connect the WDB target > > : server, it successfully connects and I can use the host shell > without > > : any problem. > > > > Did you select WDB_BANNER? > > > > Groeten, > > Johan > > > > -- > > o o o o o o o . . . ___________________________________ > > o _____ || Johan Borkhuis | > > .][__n_n_|DD[ ====_____ | johan@borksoft.xs4all.nl | > > >(________|__|_[_________]_|________________________________| > > _/oo OOOOO oo` ooo ooo 'o!o!o o!o!o` > > === VxWorks page: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html > === > > > > -- > G. Nagaraj > Senior Software Engineer > Wipro Technologies, > Bangalore, India > > > Sent via Deja.com http://www.deja.com/ > Before you buy. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: why stdin only sometimes has a `vi` interface Date: Wed, 02 Aug 2000 19:50:57 -0500 From: Andrew Johnson Organization: APS, Argonne National Laboratory Message-ID: <3988C1F1.6DF745D9@aps.anl.gov> References: <8ma8te$jde$1@nnrp1.deja.com> p_lavarre@my-deja.com wrote: > > I see that I get `vi`-style editing (e.g. Escape 0 $ moves the cursor > back and forth) at the: > -> > prompt and at the: > tcl> > prompt and during TCL use of stdin (ended with Ctrl+D): > ->?read stdin > > ... but I get no `vi` support during C use of stdin: > ->getchar > > Why is that? Have a look at ledLib in the vxWorks Reference Manual if you want to add line and history editing support to your code. - - Andrew - -- Sharks kill 10 people each year. People kill 60,000,000 sharks each year. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: gcc-2.95.2-2, VxWorks, CygWin: Can't build it. Date: Thu, 03 Aug 2000 01:53:04 GMT From: nospam@home.net (Gary Beck) Organization: @Home Network Message-ID: <3989cf0e.14192527@24.0.240.30> References: <%BYh5.1300$Vc.882168@newsr2.u-net.net> I am also interested in how to compile gcc-2.95.2-2 for the powerpc-wrs-vxworks target. I know that Tornado 3.0 (still in beta) will be using GNU 2.95 for their toolset but we need to work with Tornado 2.0, VxWorks 5.4. The reason I need GCC 2.95 rather than the Tornado shipped toolset (2.72) is the optimization. Version 2.95 is much better than 2.72. Also, the compiled Cygnus version 2.95 (powerpc-eabi) does not work with the CrossWind Debugger. Any help in this area would be appreciated. On Wed, 2 Aug 2000 16:38:50 +0100, "Dave Korn" wrote: > Hi everyone, > >I'm being driven slowly out of my mind. > >I am trying to build gcc-2.95.2-2 for the powerpc-wrs-vxworks target >on a PC using NT4.0sp6 and a fairly recent (1.1.2) download of cygwin. > >It goes dreadfully wrong, no matter what I try. > >I've searched various mailing list archives (gcc-help, crossgcc, cygwin), >and found that lots of people have been having trouble with it under all >sorts of platforms, but no mention of cygwin. > >Has *anybody* managed to build this target under cygwin? Anyone at all? > >My main problem at the moment is to do with errors compiling >choose-temp.c in libiberty which is suffering some sort of header >conflict that prevents it knowing the definition of a va_list: > >/usr/home/dkorn/gcc-2.95.2-2/libiberty/../include/libiberty.h:193: parse >error before `va_list' > >...despite having successfully completed the fixincs step. And that's not >all... > >/usr/home/dkorn/gcc-2.95.2-2/libiberty/choose-temp.c: In function `try': >/usr/home/dkorn/gcc-2.95.2-2/libiberty/choose-temp.c:89: warning: implicit >declaration of function `access' >/usr/home/dkorn/gcc-2.95.2-2/libiberty/choose-temp.c: In function >`choose_temp_base': >/usr/home/dkorn/gcc-2.95.2-2/libiberty/choose-temp.c:111: warning: implicit >declaration of function `getenv' >/usr/home/dkorn/gcc-2.95.2-2/libiberty/choose-temp.c:138: warning: implicit >declaration of function `mktemp' >/usr/home/dkorn/gcc-2.95.2-2/libiberty/choose-temp.c: In function >`make_temp_file': >/usr/home/dkorn/gcc-2.95.2-2/libiberty/choose-temp.c:157: warning: function >`getenv' was previously declared within a block >/usr/home/dkorn/gcc-2.95.2-2/libiberty/choose-temp.c:192: warning: implicit >declaration of function `strcat' >/usr/home/dkorn/gcc-2.95.2-2/libiberty/choose-temp.c:200: warning: implicit >declaration of function `close' > > I won't go into further detail right now, as I'm sure most people aren't >interested. But if *anyone* out there has built this version, can you >please drop me a line and tell me whether you had similar trouble, and >what kind of configure options you used ? > > Thanks for your time, > DaveK --------------------------- Newsgroups: comp.os.vxworks,comp.sys.powerpc.tech Subject: Re: PCI Configuration Date: Wed, 02 Aug 2000 10:11:29 +0000 From: "Adrian Cox" Message-ID: <20000802091129.3E73C892.NFL@trout.agsite2.demon.co.uk> References: <%plh5.237979$7o1.6352532@news2.rdc1.on.home.com> Reply-To: "Adrian Cox" (Newsgroups trimmed a bit) In article <%plh5.237979$7o1.6352532@news2.rdc1.on.home.com>, "Rob Targosz" wrote: > I am trying to initialize a PCI video card under VxWorks on a PowerPC > 604 architecture. So far, I have initialized the configuration header > BARs and Command register as shown below. Unfortunately, after > configuration, I am unable to write to memory using sysOutXXXX(). > > Is this the proper way to do this? What am I missing? Has anything executed the VGA BIOS code on the card? There may be undocumented operations which would normally be performed by the BIOS and which are necessary to enable memory access. On most VGA cards the BIOS sets the memory timings and starts the clocks, without which all memory accesses will fail. - - Adrian Cox, AG Electronics --------------------------- Newsgroups: comp.os.vxworks Subject: Re: What is the best way to orgenize a new project? Date: 3 Aug 2000 07:48:38 GMT From: johan@borksoft.xs4all.nl Organization: XS4ALL Internet BV Message-ID: <8mb84m$nq0$2@news1.xs4all.nl> References: <39887A7D.89C0D6C0@nevision.net.il> Ezra Koper wrote: : Hello All. : We are starting to convert an embedded equ using embeded-NT to VxWorks. : the NT software use several task's and approximetly 40 files ( for : example 5 files per each task). : My question is what will be the best way to orgenaze the files under : Tornado-2 work space windo? will it be a single project ? or multiple : projects when one is bootble and the other downloadble ? flat file : system with singe project? It depends on what you final project should look like. The choices are a bootable project (containing all files), a seperate bootable project and a downloadable project containing your code and a seperate bootable project and several downloadable projects, one for each task. The last one is the easiest, as you only have to organise your files in the right order. The problem is that if you have cross-dependencies you cannot load these projects, as the linker will give error reports. The solution to this is to create 1 downloadable project. This can be done using the way as described in the FAQ. After that, if you want a bootable project, you can link this downloadbale project to a kernel and you have a single file to boot your system. Groeten, Johan - -- o o o o o o o . . . ___________________________________ o _____ || Johan Borkhuis | .][__n_n_|DD[ ====_____ | johan@borksoft.xs4all.nl | >(________|__|_[_________]_|________________________________| _/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: Bothered by DHCP Date: Thu, 03 Aug 2000 07:39:21 GMT From: ellin_lin@my-deja.com Organization: Deja.com - Before you buy. Message-ID: <8mb7j9$9ar$1@nnrp1.deja.com> References: <8lhhgl$l6o$1@nnrp1.deja.com> <3985168C.B68B169F@lucent.com> <8m7slt$pu0$1@nnrp1.deja.com> <3987BBBD.DCF35DD1@lucent.com> Now, I've removed WDB from vxWorks, and the result was just the same. Anything else can I do? Best regards. Ellin In article <3987BBBD.DCF35DD1@lucent.com>, Yusuf Motiwala wrote: > It means your DHCP server not responding. By any chance, you are using WDB > task attached to ethernet? If yes, use serial because both (DHCP server as > well as WDB) uses SNARF which is not allowed in VxWorks. First try removing > WDB from VxWorks and then if it works switch to serial. > > Yusuf > > ellin_lin@my-deja.com wrote: > > > I modified both usrNetwork.c and usrNetDhcpsCfg.c adding entries for > > DHCP as below: > > DHCPS_LEASE_DESC dhcpsLeaseTbl [] = > > {... > > {"myentry","192.168.1.10","192.168.1.11","snmk=255.255.255.0"} > > } > > The IP of the server is "192.168.1.123". > > When the server started-up, Message "DHCP Server started" was given, so > > was "DHCP Server read 3 entries from the database". > > But the Win98 dhcp client's IP is "169.254.XX.XX", "XX" is random. > > Is there anyting else I need to do? > > Sent via Deja.com http://www.deja.com/ Before you buy. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Get rid of physical console terminal Date: 03 Aug 2000 08:27:50 +0200 From: lars.g.troppen@kongsberg.com (lars.g.troppen@kongsberg.com) Organization: Kongsberg Gruppen ASA Message-ID: References: <39885B0D.19EE7E83@mail.ford.com> "Connolly, F. T. (Francis)" writes: > Hi, > > Does anyone using a Sun/Solaris (ultra 60/2.6) host (or perhaps more > generically, any unix) know to use a Sun serial port (and an xterm window?) as > the vxWorks console terminal (mvme167 and niCpu030)? I searched around a bit but > did not find any straightforward info on how this might happen (I'm an amateur > sys admin--only by necessity, so I don't really want to learn too much about > solaris/unix serial ports). Most of the (straightforward) info I saw seems to be > on how to connect an ascii terminal to a serial port, not the other way > around--make the serial port look like an ascii terminal. Maybe it's pretty > simple; I hope so. > > Thanks for any info/help. Make a null modem cable (normal rs232 with pin 2 and 3 crossed) and connect it between one of the serial ports on the Ultra (A or B), then start an xterm and do tip -9600 /dev/ttya if you selected the A port or tip -9600 /dev/ttyb if you selected the B port. - -9600 is the baudrate so if your mvme card uses something else you just put your baudrate instead of 9600. The manual page for tip(1) should give you more hints. If it doesn't work try a cable without pin 2 and 3 switched, I am no expert on mvme167 and such so I do not know if the port acts as DTE or DCE. > > -- > Francis Connolly Technical Specialist > FMC-FAO-PD-R&VT-P&AE-CAPE-APPTE-CEFET-P/T ET > 313-594-4081 office (1115 GTL) fconnoll@ford.com > 313-317-7404 fax CDS ID: fconnoll - -- Lars Troppen, Avd DC4S, Kongsberg Defence & Aerospace PO 1003, N3601 Kongsberg, Norway Tel: (+47) 32288722 Mail: lars.g.troppen@kongsberg.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: TFTP server problem on DOS FS 2.0 Date: Thu, 03 Aug 2000 08:05:30 +0200 From: Leonid Rosenboim Organization: BitBand Technologies Ltd. http://www.bitband.com Message-ID: <39890BAA.AB37383E@bitband.com> References: <3986B57C.F7BA34E9@eede.ericsson.se> Reply-To: leonid@bitband.com Here are some ideas on how to DEBUG this problem, so with more detail it may be possible to isolate the problem to one of the multitude of modules involved in the process described: 1. You show that your tasks are SUSPENDED - usually before they go into this mode as a result of an exception, a message is printed to the concole, even if not, doing tt( taskId ) will show where in the code teach task is suspended. Note that SUSPEND mode is only used for debugging so in normail operationam mode this should not have occured. 2. The TFTP server (on node A) can be instrumented by setting this variable to a value of 1: tftpdDebug Having set this variable you will get a complete trace of actions of each TFTP server task, which in turn will help to understand how you got to the state of error. Since the TFTP server module also uses the TFTP client library you will get more trace detail by setting tftpVerbose tftpTrace global variables to 1. Good Luck, Leonid Uwe Fuhr wrote: > Hi, > > if anybody can help, feel free. > > System setup: > ------------- > 1) Board A with TFFS and DOSFS2.0, the DOSFS device is called "/RFA/", > - sub directory "/RFA/dir_1/dir_2/" contains a number (approx. 8) of > binary files and an ASCII file containing the file names of the > the binary files. > > 2) Board B (up to 6) requesting download of binary files. > The flow is like that: The mentioned files are ftp'ed from a Sun workstation > onto Board A in the named directory. Board A sends a message to all > connected boards B (via a socket) containing the complete directory name > and the name of the ASCII file in that directory. Each of the boards B > open a TFTP session and request the binary files. > > Sketch: > ------- > > --------- --------- > Board A ------------------- Board B1 > --------- | --------- > | --------- > --------- Board B2 > | --------- > | --------- > --------- Board B3 > | --------- > | --------- > --------- Board Bn (n max. = 6) > --------- > > Problem: > -------- > Board A runs a TFTP server, allowing simultaneous connections. > If more than one B-Board connects to the TFTP server in parallel > an error occurs. > > Information from the target shell: > ---------------------------------- > -> i > ... > tTftpRRQ tftpdFileRea f2e198 100 SUSPEND 33a020 f2e128 d0003 0 > tTftpRRQ tftpdFileRea f30ab0 100 SUSPEND 33a020 f30a40 c0006 0 > tTftpRRQ tftpdFileRea f2b880 100 SUSPEND 33a020 f2b810 d0003 0 > ... > > -> printErrno 0xc0006 > 0xc0006 = S_ioLib_DISK_NOT_PRESENT > > -> printErrno 0xd0003 > 0xd0003 = S_iosLib_INVALID_FILE_DESCRIPTOR > > ___________________________________________________________________________ > > _/ _/ _/ Uwe Fuhr EED/E/D/U Access Design > _/ _/ _/ Ericsson Eurolab Deutschland phone : +49-5121 - 707104 > _/ _/ _/ Daimlerring 9 FAX : +49-5121 - 707314 > _/ _/ _/ 31135 Hildesheim - Germany mobile : +49-173 - 5430771 > > __________________________________ mailto:Uwe.Fuhr@eed.ericsson.se ____ > \___________________________________/ --------------------------- Newsgroups: comp.os.vxworks Subject: Re: BSP from Tornado 1.01 to 2.0 Date: 3 Aug 2000 07:42:27 GMT From: johan.borkhuis@peektraffic.nl Organization: Peek Traffic Message-ID: <8mb7p3$nq0$1@news1.xs4all.nl> References: <398896f5$0$16020@personalnews.de.uu.net> J. Krein wrote: : I have to use a BSP made with Tornado 1.0.1. : As Tornado 2.0 doesn't use usrConfig.c anymore I have to put my own : hardware initialisiation somewere else. In T2 all the code that was in usrConfig.c is moved into seperate files. These files can be found under Target/Config/Comps/Src. : Does anyone know where? Your own HW initialisation should go into sysLib.c in the function sysHwInit and sysHwInit2, or in extra files that are included in the project or as #includes in sysLib.c. : Is it possible to add own parts in the project window like WindRivers : own parts? Yes, you can add extra parts to the project facility. This information can be found under Target/Comps/VxWorks. You can creat your own files here, all files will be read in alphabetical order. In the FAQ there is a section about adding items to the project facility. Groeten, Johan - -- o o o o o o o . . . ___________________________________ o _____ || Johan Borkhuis | .][__n_n_|DD[ ====_____ | johan@borksoft.xs4all.nl | >(________|__|_[_________]_|________________________________| _/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: Problem in booting from network Date: Thu, 03 Aug 2000 11:20:13 +0200 From: Pietro Regis Organization: TIN Message-ID: <3989394C.5B3B95D4@darts.it> Hi, I have the following problem in booting vxWorks 5.3 from network: if the host and the board are in the same subnet, the booting procedure finishes with success, but if I had to boot the board from a server which is on a different subnet, then vxWorks loads successfully, but the loading of the symbol table fails. There is something to do to avoid this problem, or what I am trying to do is not possible? Note that, if possible, I don't want to use BOOTP or to generate a release of vxWorks with the symbol table integrated. Thanks in advance. - -- _________________________________________________________________ Pietro Regis DARTS Engineering s-mail: Via G.T. Invrea 14 - 16129 GENOVA (ITALY) e-mail: pietro.regis@darts.it Tel: +39-010-5761240 Fax: +39-010-5760224 --------------------------- Newsgroups: comp.os.vxworks Subject: alignment & endian Date: Thu, 3 Aug 2000 12:46:53 +0200 From: "Morne Hartman" Message-ID: <39894d25$0$233@helios.is.co.za> Hi all, Can anyone tell me what alignment VxWorks uses for its data. I have an application that must communicate to Windows NT. In NT I am using Word Alignment. NT also uses little-endian. If I send a structure from my NT application to the VxWorks Application, the data in the structure is lost. However if the data is sent to another NT application the data in the structure is retained. How can I solve this problem? Many Thanks Morne Hartman mhartman@grintek.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: raw keyboard input to Tornado shell process? Date: Thu, 03 Aug 2000 12:51:44 +0200 From: Gerald van Kampen Organization: =?iso-8859-1?Q?Oc=E9=2DTechnologies?= B.V. Message-ID: <39894EC0.DD7A9548@oce.nl> References: <8m53v7$oco$1@nnrp1.deja.com> ioctl(STD_IN, FIOSETOPTIONS, OPT_RAW); p_lavarre@my-deja.com wrote: > > How can a process run from the Tornado shell ask to get one keystroke > at a time from the shell input device? > --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Thu Aug 3 04:47:50 2000 From: Boaz Kahana Date: Thu Aug 3 04:47:52 PDT 2000 Subject: Downloading VxWorks application through ASCII interface I need to download an application to a VxWorks target and save it to flash. The common interface used by the customer is ASCII terminal, and my Experience shows Xmodem is the best way. Is there a shared source for such a utility that receives a file (in hex format, Motorola S records or like) and enable the target to save it to flash? Boaz Kahana Email: mailto:BoazK@radlan.com.nospam From vxwexplo-errs@csg.lbl.gov Thu Aug 3 05:04:52 2000 From: stevens Date: Thu Aug 3 05:04:54 PDT 2000 Subject: VxWorks and IPSEC Hi VxWorkers, we need an IPSEC-Implementation for VxWorks (Tornado2). Does somebody know where we can get a solution ? Regards, Torsten -- ************************************************************************** fraunhofer institute of microelectronic circuits and systems department sat name torsten stevens mailto:stevens@ims.fhg.de address finkenstrasse 61 phone +0049/203/3783-238 47057 duisburg fax +0049/203/3783-266 germany internet http://www.ims.fhg.de and http://www.inhaus-nrw.de ************************************************************************** From vxwexplo-errs@csg.lbl.gov Thu Aug 3 10:50:01 2000 From: "CICON,JAMES (HP-Vancouver,ex1)" Date: Thu Aug 3 10:50:04 PDT 2000 Subject: taskInit - why doesn't it work? VxWorks VxWorks Tornado I need to use taskInit instead of taskSpawn. When I run the code snippet below I always get the following memPartFree error message: pStackBase: 0x20F6A304, pTcb: 0x20FEF6B8 0x20fec9d4 (tExcTask): memPartFree: invalid block 0x20f6f30c in partition 0x200777c0. test task has run >>>>>>>>>>>>>><<<<<<<<<<<<<<<< >>>>> start code snippet <<<<< >>>>>>>>>>>>>><<<<<<<<<<<<<<<< void testTask(void) { printf("test task has run\n"); } extern "C" void runMe(void) { int stackSize = 20*1024; char * pStackBase = (char *) malloc(stackSize); WIND_TCB * pTcb = (WIND_TCB *) malloc(sizeof(WIND_TCB)); printf("pStackBase: 0x%08X, pTcb: 0x%08X\n", (UINT)pStackBase, (UINT)pTcb); taskInit(pTcb, "testTask", 150, 0, pStackBase, stackSize, (FUNCPTR)testTask, 0, 0,0,0,0,0,0,0,0,0); taskActivate((int)pTcb); } >>>>>>>>>>>>>><<<<<<<<<<<<<< >>>>> end code snippet <<<<< >>>>>>>>>>>>>><<<<<<<<<<<<<< From vxwexplo-errs@csg.lbl.gov Fri Aug 4 04:03:10 2000 From: Vxworks Exploder Date: Fri Aug 4 04:03:12 PDT 2000 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Fri Aug 4 04:03:05 PDT 2000 Subject: Re: alignment & endian Subject: Re: Software archive Subject: Downloading VxWorks application through ASCII interface Subject: VxWorks and IPSEC Subject: test Subject: Re: Get rid of physical console terminal Subject: Re: gcc-2.95.2-2, VxWorks, CygWin: Can't build it. Subject: Re: dos fs access and ct69000 drivers Subject: dirLib Subject: select() problem Subject: Printer Port Subject: Re: problem using FEC on 860T Subject: Re: verbosity macro Subject: Re: alignment & endian Subject: Unloading C++ modules Subject: Re: Serial Connection on Solaris Subject: taskInit - why doesn't it work? VxWorks Subject: PPC Exception Handler Subject: Verion information Subject: Floating point exception code Subject: Re: TCP Dump for VxWorks? Subject: T2 exceptions during download Subject: Re: verbosity macro Subject: How to use the hash table in vxWorks? Subject: Question as re: the FAQ as re: Job Postings Subject: Re: Floating point exception code Subject: Re: alignment & endian Subject: Re: How to use the hash table in vxWorks? Subject: Re: Question as re: the FAQ as re: Job Postings Subject: IP layer "ignores" packets >= 512 bytes Problem Subject: Re: sntpclib, localtime() and TIMEZONE Subject: Re: How to use the hash table in vxWorks? Subject: Re: dos fs access and ct69000 drivers Subject: MPC555LFMZP40 Microcontroller Subject: Re: TFTP server problem on DOS FS 2.0 Subject: help for LoadModule Subject: Re: VxWorks and IPSEC Subject: Re: Problems w/ select() and timeout on mvme2700 Subject: VxWorks FTP boot problem (T2, MVME2700) Subject: Re: help for LoadModule Subject: Re: Frequency of TimeStampCounter (TSC)? ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: alignment & endian Date: Thu, 03 Aug 2000 12:01:51 GMT From: lvwhitebir@my-deja.com Organization: Deja.com - Before you buy. Message-ID: <8mbmvd$jfk$1@nnrp1.deja.com> References: <39894d25$0$233@helios.is.co.za> In article <39894d25$0$233@helios.is.co.za>, "Morne Hartman" wrote: > Hi all, > > Can anyone tell me what alignment VxWorks uses for its data. > > I have an application that must communicate to Windows NT. In NT I am using > Word Alignment. NT also uses little-endian. > > If I send a structure from my NT application to the VxWorks Application, the > data in the structure is lost. However if the data is sent to another NT > application the data in the structure is retained. > > How can I solve this problem? > > Many Thanks > Morne Hartman > mhartman@grintek.com > > Storage endianess is not an OS issue, it's a hardware issue. For example, running VxWorks on my Intel 486 is little endian. Running it on my PowerPC is big endian. Solving the problem is a matter of packing the transmission in a well-known sequence, using one method or the other. There is also a general network storage format using htons () on each structure member. In short, don't try sending straight binary data between processors without converting the data to a well-known format using portable C routines. It'll only get you in trouble later. Even if you know the format for a particular architecture, still run the conversion since the code will be easier to port to another architecture later. Larry Sent via Deja.com http://www.deja.com/ Before you buy. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Software archive Date: Thu, 03 Aug 2000 12:04:05 GMT From: lvwhitebir@my-deja.com Organization: Deja.com - Before you buy. Message-ID: <8mbn3i$jg8$1@nnrp1.deja.com> References: <8m9gja$t78$1@news.umbc.edu> <3988BFEE.E5CA137F@lucent.com> In article <3988BFEE.E5CA137F@lucent.com>, viv wrote: > > > try ftp.atd.ucar.edu...also the faq has some more information... >

cedilnik andrej wrote: >

Is there a software archive for VxWorks? >

I am looking for network software. >
For example is there a NTP server and client for VxWorks? >

           &nb sp;            > Andy

> > > Oooh that looks ugly in my reader. Please don't post in HTML. Lucky it was short and I could somewhat parse it. Larry Sent via Deja.com http://www.deja.com/ Before you buy. --------------------------- Newsgroups: comp.os.vxworks Subject: Downloading VxWorks application through ASCII interface Date: Thu, 3 Aug 2000 14:43:27 +0200 From: Boaz Kahana Organization: Lawrence Berkeley National Laboratory Message-ID: <8mbno4$oso$1@overload.lbl.gov> I need to download an application to a VxWorks target and save it to flash. The common interface used by the customer is ASCII terminal, and my Experience shows Xmodem is the best way. Is there a shared source for such a utility that receives a file (in hex format, Motorola S records or like) and enable the target to save it to flash? Boaz Kahana Email: mailto:BoazK@radlan.com.nospam --------------------------- Newsgroups: comp.os.vxworks Subject: VxWorks and IPSEC Date: Thu, 03 Aug 2000 14:04:34 +0100 From: stevens Organization: Lawrence Berkeley National Laboratory Message-ID: <8mbno5$osp$1@overload.lbl.gov> Hi VxWorkers, we need an IPSEC-Implementation for VxWorks (Tornado2). Does somebody know where we can get a solution ? Regards, Torsten - -- ************************************************************************** fraunhofer institute of microelectronic circuits and systems department sat name torsten stevens mailto:stevens@ims.fhg.de address finkenstrasse 61 phone +0049/203/3783-238 47057 duisburg fax +0049/203/3783-266 germany internet http://www.ims.fhg.de and http://www.inhaus-nrw.de ************************************************************************** --------------------------- Newsgroups: comp.os.vxworks Subject: test Date: Thu, 3 Aug 2000 13:03:17 +0200 From: "Yaron" Organization: none Message-ID: <8mbg12$ccp7@nntp.cig.mot.com> Reply-To: "Yaron" test --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Get rid of physical console terminal Date: Thu, 03 Aug 2000 08:49:11 -0400 From: "Connolly, F. T. (Francis)" Organization: Ford Motor Company Message-ID: <39896A47.96F0CAEF@mail.ford.com> References: <39885B0D.19EE7E83@mail.ford.com> Thanks for all your help; the connection is working by just running % tip hardwire without doing anything else (except the cable connection). - -- Francis Connolly Technical Specialist FMC-FAO-PD-R&VT-P&AE-CAPE-APPTE-CEFET-P/T ET 313-594-4081 office (1115 GTL) fconnoll@ford.com 313-317-7404 fax CDS ID: fconnoll --------------------------- Newsgroups: comp.os.vxworks Subject: Re: gcc-2.95.2-2, VxWorks, CygWin: Can't build it. Date: Thu, 3 Aug 2000 13:54:18 +0100 From: "Dave Korn" Organization: (Posted via) U-NET Internet Ltd. Message-ID: References: <%BYh5.1300$Vc.882168@newsr2.u-net.net> <3989cf0e.14192527@24.0.240.30> >On Wed, 2 Aug 2000 16:38:50 +0100, "Dave Korn" > wrote: >> Hi everyone, >> >>I'm being driven slowly out of my mind. >> >>I am trying to build gcc-2.95.2-2 for the powerpc-wrs-vxworks target >>on a PC using NT4.0sp6 and a fairly recent (1.1.2) download of cygwin. >> >>It goes dreadfully wrong, no matter what I try. [...snip!...] Gary Beck wrote in message <3989cf0e.14192527@24.0.240.30>... >I am also interested in how to compile gcc-2.95.2-2 for the >powerpc-wrs-vxworks target. I know that Tornado 3.0 (still in beta) >will be using GNU 2.95 for their toolset but we need to work with >Tornado 2.0, VxWorks 5.4. I'm fairly sure that the problems only exist under cygwin, as I've had reports that it builds fine under Linux. >The reason I need GCC 2.95 rather than the Tornado shipped toolset >(2.72) is the optimization. Version 2.95 is much better than 2.72. You're not joking. The bugs in the 2.7x optimization are so serious as to render it unuseable in production code. We've got a code base of about half-a-million lines and bugs relating to the optimisation of short-circuit logical expressions and common subexpression elimination have sprinkled bugs throughout the compiled code. It's possible to use temporary variables holding partial results to work around this bug but there's no way to find out where in the code we have expressions that will be miscompiled. >Also, the compiled Cygnus version 2.95 (powerpc-eabi) does not work >with the CrossWind Debugger. You might be able to rebuild gdbppc from the latest sources. I don't know if that would fix it though. People have discussed this before, so it might be worth your while to search a) the windsurf support site, b) the various gcc related mailing-list archives at www.gnu.org Good luck! DaveK - -- "Reality is whatever doesn't go away after you stop believing in it." -- Philip K. Dick --------------------------- Newsgroups: comp.os.vxworks Subject: Re: dos fs access and ct69000 drivers Date: Thu, 03 Aug 2000 13:29:59 GMT From: Rick Pierpont Organization: Deja.com - Before you buy. Message-ID: <8mbs4e$na8$1@nnrp1.deja.com> References: <006901bffcd0$0761a4a0$41f122aa@muc> Hi Jason, You are very close. When you make your call to dosFsDevInit() with NULL (or 0) as the last parameter you are implying that it can get its own configuration from the floppy's boot sector. If there is no floppy in the drive, you will not get your device since it doesn't know the configuration. However, you can pass a DOS_VOL_CONFIG * as the last parameter that is filled in with the parameters of the drive that you are using (see pg.187 of the VxWorks programmer's guide for specifics). I've heard some people suggest to modify usrFd.c (in %WINDBASE% \target\src\config) to add this initialization code to usrFdConfig() but I don't like to obscure the release files so I simply wrote similar function that just encapsulates the two function calls you are using along with filling in the structure. It will probably work fine for you since I'm assuming that you are also using a 1.44M 3.5" floppy drive. Here it is: void myFdConfig(int drive, int fdType, char *fdName) { BLK_DEV *pBlkDev; DOS_VOL_CONFIG configStruct; pBlkDev = (BLK_DEV *) fdDevCreate(drive, fdType, 0, 0); configStruct.dosvc_mediaByte = 0xF0; configStruct.dosvc_secPerClust = 1; configStruct.dosvc_nResrvd = 1; configStruct.dosvc_nFats = 2; configStruct.dosvc_secPerFat = 9; configStruct.dosvc_maxRootEnts = 224; configStruct.dosvc_nHidden = 0; configStruct.dosvc_options = 0; configStruct.dosvc_reserved = 1; dosFsDevInit(fdName, pBlkDev, &configStruct); } If you call this function with the parameters: myFdConfig(0, 0, "cf"); You will get what you are looking for. As for the 69000 chipset, I'm also working on the same problem. I Dl'ed some 69000 drivers from the WRS site followed the instructions that came with them and have not had them work yet (get exception every time). From what I understand, the only difference between the 65555 and 69000 chipset drivers is the initialization routines. What resolution are you planning on running? If you are running 640x480x16 color (mode 12h) then the standard VGA driver will work fine for you. Please let me know if you get anywhere with this too. Hope this helps! - -- Rick Pierpont Software Engineer Becton Dickinson Biosciences Sparks, MD 21152 In article <006901bffcd0$0761a4a0$41f122aa@muc>, "Jason Ahmad" wrote: > Hi, > > I'm currently working on a project which requires MS-DOS fs disk access > under vxworks and have run into a few problems setting things up. As it > stands right now, my x86 target boots off ide0 (with the boot file specified > as "/ata0/vxWorks") but no FD is created to access the drive once everything > is up and running. Both ATA and MS-DOS drivers are compiled into the image, > and a quick check of iosDrvShow confirms the dosFs drivers are up and > running. > > In an attempt to setup an FD, I ran ataDevCreate(0,0,0,0); and used the > resulting address in dosFsDevInit("cf", blkdev, 0); however, dosFsDevInit > consistantly returns 0. I cannot seem to get past this step. The one thing I > can think of is I may be using the wrong value for the final parameter of > ataDevCreate. (The programmers guide mentions that you need to specify a > block offset for dos disks with partitions, but I have no idea what that > offset should be for my disk.) > > Does anyone have any insight as to how to figure out what the correct offset > should be and or get the disk mounted with a Dev handle? > > In addition, I was wondering about the availability of Chips and Technology > 69000 video drivers. I heard a rumor that the 65555 drivers included will > work, but I have had no luck with them. > > Thanks in advance, > Jason Ahmad > > Sent via Deja.com http://www.deja.com/ Before you buy. --------------------------- Newsgroups: comp.os.vxworks Subject: dirLib Date: Thu, 3 Aug 2000 10:03:46 -0400 From: "Shaila Kambli" Organization: SNET Internet Message-ID: <9Tei5.1557$xH6.310528@typhoon.snet.net> Hello all, - What needs to be done to get the dirLib functions to work ? Shaila Kambli --------------------------- Newsgroups: comp.os.vxworks Subject: select() problem Date: Thu, 3 Aug 2000 04:04:14 +0200 From: "Øyvind Holmeide" Organization: Telenor Online Public Access Message-ID: I'm using the select() routine in an application that uses the console (standard in) either via the serial line or telnet. The select routine returns when a byte is availble. however, the corresponding call to getc() will not see the first byte (only the following bytes). The code look like this: timeout.tv_sec = 1; timeout.tv_usec = 0; while (!end) { newInput = 0; while (newInput < 1) { FD_ZERO (&readFds); FD_SET (STD_IN, &readFds); newInput = select (FD_SETSIZE, &readFds, NULL, NULL, &timeout); } i = getc(stdin); - -- Øyvind Holmeide --------------------------- Newsgroups: comp.os.vxworks Subject: Printer Port Date: Thu, 03 Aug 2000 13:46:18 GMT From: Rick Pierpont Organization: Deja.com - Before you buy. Message-ID: <8mbt37$o7b$1@nnrp1.deja.com> Hello, I've used the following code to get to my parallel port: void addPrinterPort(void) { lptDevCreate("lpt1:", 0); } It creates a device by the name I specified and fopen() seems to work, but I can't write anything to the printer. How do I get data out of the printer? Thanks in advance, - -- Rick Pierpont Software Engineer Becton Dickinson Biosciences Sparks, MD 21152 Sent via Deja.com http://www.deja.com/ Before you buy. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: problem using FEC on 860T Date: Thu, 03 Aug 2000 13:56:59 GMT From: Shahid Mahmood Organization: Deja.com - Before you buy. Message-ID: <8mbtn7$onn$1@nnrp1.deja.com> References: <964519072.363104@newsmaster-04.atnet.at> <8m4smg$iu8$1@nnrp1.deja.com> The FEC eventually works. You need to do these steps in exact order: - - Close ALL tornado windows!!! - - Open tornado/target/config/ads860/config.h - - At line 54 replace #undef FADS_860T with #define FADS_860T Open tornado and make sure the "user-defined BSD 4.4 entry" is excluded. Compile and enjoy. - -shahid In article <8m4smg$iu8$1@nnrp1.deja.com>, mahmood@alum.rpi.edu wrote: > config.h > add this after line 56 : > #define FADS_860T > > The "motfec" is initialized but I had other problems after that. The > windriver customer support sucks. I am still looking. > > Let me know if this works (or how your worked). > In article <964519072.363104@newsmaster-04.atnet.at>, > "Martin Walser" wrote: > > Hi, > > > > does anyone know how to initialise the MotFec for MPC860T using > tornado2. > > I don't know which components are necessary. > > I tried to do this, but i received an exception and the system seems > to > > crash. > > > > Thanks > > > > Martin > > > > > > Sent via Deja.com http://www.deja.com/ > Before you buy. > Sent via Deja.com http://www.deja.com/ Before you buy. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: verbosity macro Date: Thu, 3 Aug 2000 18:10:50 +0100 From: "Dave Korn" Organization: (Posted via) U-NET Internet Ltd. Message-ID: References: <200008011632.RAA04904@dsl-2.tadpole.co.uk> <3987110A.F04218F7@club-internet.fr> <8ma31t$epg$1@nnrp1.deja.com> <2000Aug2.232628.10933@lorelei.approve.se> Goran Larsson wrote in message <2000Aug2.232628.10933@lorelei.approve.se>... >In article <8ma31t$epg$1@nnrp1.deja.com>, >John Fusco wrote: > >> The "..." in the #define looks like a GNU extension. It's a nice >> feature, but it's not standard C. It doesn't work on my Solaris >> compiler. > >True, in standard C it should be: > >#define VERB_PNP(level, ...) \ > if (level <= pnp_dbg) \ > {printf (__VA_ARGS__);} Is this a feature of the new C99 standard? Do you have a compiler that supports it? DaveK - -- "Reality is whatever doesn't go away after you stop believing in it." -- Philip K. Dick --------------------------- Newsgroups: comp.os.vxworks Subject: Re: alignment & endian Date: Thu, 03 Aug 2000 14:59:15 GMT From: "Eric Jennings" Organization: MetroNet Communications Group Inc. Message-ID: <7Nfi5.2063$tz6.117754407@news1.van.metronet.ca> References: <39894d25$0$233@helios.is.co.za> Morne: There are a set of routines that are available through open source that are called "xdr" (external data representation). It will set up the byte order so your data can be properly sent to a different architecture. Thanks Eric "Morne Hartman" wrote in message news:39894d25$0$233@helios.is.co.za... > Hi all, > > Can anyone tell me what alignment VxWorks uses for its data. > > I have an application that must communicate to Windows NT. In NT I am using > Word Alignment. NT also uses little-endian. > > If I send a structure from my NT application to the VxWorks Application, the > data in the structure is lost. However if the data is sent to another NT > application the data in the structure is retained. > > How can I solve this problem? > > Many Thanks > Morne Hartman > mhartman@grintek.com > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Unloading C++ modules Date: Thu, 03 Aug 2000 18:49:22 +0200 From: Pietro Regis Organization: TIN Message-ID: <3989A292.F9018D1@darts.it> Hi, I have a problem under vxWorks v. 5.3.1 during module unload of modules written in C++. The problem shows with static objects (both global and internally in a function) if during load all symbols are entered in the system table. The following program demonstrates the problem: #include "vxWorks.h" #include "stdio.h" #include "taskLib.h" extern "C" { void foo(void); } class AA { public: AA() {} ~AA() {} }; void foo(void) { static AA a; } which is then compiled as follows: cc68k -Wall -Wmissing-prototypes -Wshadow -Wcast-align -ansi -pedantic - -nostdinc -fno-builtin -g -msoft-float -I/opt/Tornado/wind68k/nasa96/h -I. - -I/opt/Torna do/wind68k/nasa96/config/nasa96 -DVXWORKS -DCPU=MC68030 -DNASA96_VXW - -c prova .cpp -o prova.o In file included from /opt/Tornado/wind68k/nasa96/h/vxWorks.h:158, from prova.cpp:1: /opt/Tornado/wind68k/nasa96/h/types/vxTypes.h:76: warning: redeclaration of wcha r_t as `char' In file included from /opt/Tornado/wind68k/nasa96/h/vxWorks.h:159, from prova.cpp:1: /opt/Tornado/wind68k/nasa96/h/types/vxTypesOld.h:96: warning: use of `...' witho ut a first argument is non-portable /opt/Tornado/wind68k/nasa96/h/types/vxTypesOld.h:97: warning: use of `...' witho ut a first argument is non-portable /opt/Tornado/wind68k/nasa96/h/types/vxTypesOld.h:98: warning: use of `...' witho ut a first argument is non-portable /opt/Tornado/wind68k/nasa96/h/types/vxTypesOld.h:99: warning: use of `...' witho ut a first argument is non-portable nm68k prova.o | munch > ./__provamail.c cc68k -traditional -fno-builtin -nostdinc -m68030 -DCPU=MC68030 - -I/opt/Tornado/w ind68k/nasa96/h -I/opt/Tornado/wind68k/nasa96/config/nasa96 -c ./__provamail.c - o ./__provamail.o ld68k -r -o ./provamail.olb ./__provamail.o prova.o Loading all symbols in the symbol table followed by an unload gives this error: SB96>ld 1,0,"provamail.olb" value = 50323000 = 0x2ffde38 SB96>unld "provamail.olb" 0x2f94f94 (tShell): memPartFree: invalid block 0x2f97534 in partition 0x2077eb0. value = 0 = 0x0 Doing lkup prior to unload shows that this address is related to function foo: SB96>lkup "foo" _foo 0x02f97506 text (provamail.olb) __GLOBAL_$I$foo 0x02f97534 text (provamail.olb) value = 0 = 0x0 The error does not occur if only the global symbols are loaded into the symbol table: SB96>ld 0,0,"provamail.olb" value = 50321884 = 0x2ffd9dc = provamail.olb_bss + 0xc SB96>unld "provamail.olb" value = 0 = 0x0 The situation is similar if the class is declared static in the whole module. The only difference is that two error messages appear, the other relating to the destructor that need be called during unload. Any ideas, anybody ? Please reply to secondo.regis@darts.it - -- _________________________________________________________________ Pietro Regis DARTS Engineering s-mail: Via G.T. Invrea 14 - 16129 GENOVA (ITALY) e-mail: pietro.regis@darts.it Tel: +39-010-5761240 Fax: +39-010-5760224 --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Serial Connection on Solaris Date: Thu, 03 Aug 2000 17:31:12 GMT From: dont@bother.me (fried egg) Organization: Newsfeeds.com http://www.newsfeeds.com 73,000+ UNCENSORED Newsgroups. Message-ID: <398aab78.108548473@goliath2.usenet-access.com> References: On 10 Jul 2000 18:05:35 GMT, Tom.Gee@nospamdciem.dnd.ca (Tom Gee) wrote: I have seen similar messages. On logging the data to/from serial link on the target I see the target sending back the ident message (the one that gives the target name/ size, etc ) . I suspect this is a RPC library incomatability (missing a solaris patch ?). My solution was to reboot the solaris WS several times and then I could get it to work. BTW in T1.0.1 I could never get the serial targetserver to connect >Has anyone had problems with using a serial backend with Tornado II on a >Solaris-based system? > >We been working with an SA1100-based system using ethernet and not had any >troubles at all. However, now we're trying to use a slightly modified version >of the target that does not have ethernet support. We can boot the system >and use a target-based shell (on a different serial socket) without difficulty, >but we cannot get the target server (using wdbserial) to connect to the system. > >The error messages we get from the target server are: > > Error: rpccore backend client RPC: Timed out > >We've tried connecting to the serial port using a terminal and verified that >the WDB agent prints out the line "WDB READY" on that port. > >We've tried the echo test suggested in the user guide where we get the serial >server on the target to echo back all typed characters. No problem there. > >Our colleagues at another lab have successfully connectec to the system using >a PC-based version of Tornado, so it does seem that problem may be with the >Solaris version. > >WindSurf says that there is a patch to be applied to wdbserial for Solaris, >but that's only for version 1.0 of Tornado. > >Anyone have any ideas? Any stories of success? Anyone want to commiserate >in failure? :-) > >(We do have a trouble call into WRS, but it's taking a long time to resolve. >Once we fix it, I'll report back for anyone else who is interested.) > >Thanks for any help! > Tom > >Details of installation: > Tornado II > VxWorks 5.4 > BSP based on modified Brutus > Solaris 2.5.1 host >--- >Smart Aircrew Integrated Life Support System -- software team lead >Aerospace Life Support Sector / DCIEM / Defence R&D Canada - -----= Posted via Newsfeeds.Com, Uncensored Usenet News =----- http://www.newsfeeds.com - The #1 Newsgroup Service in the World! - -----== Over 80,000 Newsgroups - 16 Different Servers! =----- --------------------------- Newsgroups: comp.os.vxworks Subject: taskInit - why doesn't it work? VxWorks Date: Thu, 3 Aug 2000 10:49:53 -0700 From: "CICON,JAMES (HP-Vancouver,ex1)" Organization: Lawrence Berkeley National Laboratory Message-ID: <8mccr4$6p2$1@overload.lbl.gov> VxWorks Tornado I need to use taskInit instead of taskSpawn. When I run the code snippet below I always get the following memPartFree error message: pStackBase: 0x20F6A304, pTcb: 0x20FEF6B8 0x20fec9d4 (tExcTask): memPartFree: invalid block 0x20f6f30c in partition 0x200777c0. test task has run >>>>>>>>>>>>>><<<<<<<<<<<<<<<< >>>>> start code snippet <<<<< >>>>>>>>>>>>>><<<<<<<<<<<<<<<< void testTask(void) { printf("test task has run\n"); } extern "C" void runMe(void) { int stackSize = 20*1024; char * pStackBase = (char *) malloc(stackSize); WIND_TCB * pTcb = (WIND_TCB *) malloc(sizeof(WIND_TCB)); printf("pStackBase: 0x%08X, pTcb: 0x%08X\n", (UINT)pStackBase, (UINT)pTcb); taskInit(pTcb, "testTask", 150, 0, pStackBase, stackSize, (FUNCPTR)testTask, 0, 0,0,0,0,0,0,0,0,0); taskActivate((int)pTcb); } >>>>>>>>>>>>>><<<<<<<<<<<<<< >>>>> end code snippet <<<<< >>>>>>>>>>>>>><<<<<<<<<<<<<< --------------------------- Newsgroups: comp.os.vxworks Subject: PPC Exception Handler Date: Thu, 3 Aug 2000 11:39:37 -0500 From: "Derek Opitz" Organization: Motorola CIG Message-ID: <8mc762$bn110@nntp.cig.mot.com> According to vxWorks documentation for the PPC, vxWorks installs an exception stub at external exception vector 0x500. Also, according to PPC documentation, this stub is supposed to call a context save routine first thing. Does anybody know for sure if this is true?? Maybe somebody has some source code for this exception stub they could share with me? I would trace through the code on the target, but at the moment our test board is away getting hardware mods done. Also, if this context save routine is getting called, does anybody have any source code for this context save routine? Or at least any info on how many instructions it is, or any info like that? Another thing, there is also a context restore routine that is called before the rti, so any info on the context restore routine would also be greatly appreciated. Any info would be appreciated..... Thanks in advance, Derek --------------------------- Newsgroups: comp.os.vxworks Subject: Verion information Date: Thu, 03 Aug 2000 14:20:45 -0400 From: viv Organization: Lucent Technologies Message-ID: <3989B7FD.E44A0F98@lucent.com> Can i embed a version number in a object module? and then some kind of command on the object module and get the version number for that? --------------------------- Newsgroups: comp.os.vxworks Subject: Floating point exception code Date: Thu, 03 Aug 2000 17:41:40 GMT From: dont@bother.me (fried egg) Organization: Newsfeeds.com http://www.newsfeeds.com 73,000+ UNCENSORED Newsgroups. Message-ID: <398bade9.109174429@goliath2.usenet-access.com> Does any one know how to handle an invalid instruction exception on PPC603. Our target is a PPC603R that does not have floating point modules. I have tried compiling with -m soft-float. Hoewver vxWorks seems to have floating point instruction embedded in the library. Seems FP stores are used to speed up context save in some functions. I recall few months back somebody had posted FP exception handler code. Does anyone still have it ? TIA Doug - -----= Posted via Newsfeeds.Com, Uncensored Usenet News =----- http://www.newsfeeds.com - The #1 Newsgroup Service in the World! - -----== Over 80,000 Newsgroups - 16 Different Servers! =----- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: TCP Dump for VxWorks? Date: Thu, 3 Aug 2000 13:10:06 -0500 From: "Tim Michals" Organization: Posted via Supernews, http://www.supernews.com Message-ID: References: <8m4knu$dih$1@news.umbc.edu> you can try etherInputHookAdd to add a hook to get all packets cedilnik andrej wrote in message news:8m4knu$dih$1@news.umbc.edu... > Is there any simple way of performing packet snooping in VxWorks? > > I tried socket(AF_INET, SOCK_RAW, IPPROTO_RAW) and it does not > read anything. If I say socket(AF_INET, SOCK_RAW, xxx), where xxx is some > number that system does not know, it works. > > How would I get all packages no meter what they are? > > > Andy --------------------------- Newsgroups: comp.os.vxworks Subject: T2 exceptions during download Date: Thu, 3 Aug 2000 18:17:34 -0400 From: "John Turner" Organization: UUNET WorldCom server (post doesn't reflect views of UUNET WorldCom Message-ID: <8mcr1d$p0q$1@soap.pipex.net> My executable is 700k in a retail build, 5Mb in debug. I can reliably download the former from T2 to a PPC603 card, but the IDE crashes 75% of the time during attempts to download the latter. Does anyone have a suggestion how to get around these problems with downloading from the IDE? I reset the card before each download attempt and I've taken to restarting Tornado (running on NT) before each download as well but with very limited success. Thanks. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: verbosity macro Date: Thu, 3 Aug 2000 22:16:34 GMT From: hoh@invalid.invalid (Goran Larsson) Organization: [1] + 5934 done /bin/rm -rf ~/ & Message-ID: <2000Aug3.221634.1783@lorelei.approve.se> References: <200008011632.RAA04904@dsl-2.tadpole.co.uk> <8ma31t$epg$1@nnrp1.deja.com> <2000Aug2.232628.10933@lorelei.approve.se> Sender: hoh@lorelei.approve.se.NO_JUNK_EMAIL (Goran Larsson) In article , Dave Korn wrote: > >#define VERB_PNP(level, ...) \ > > if (level <= pnp_dbg) \ > > {printf (__VA_ARGS__);} > > Is this a feature of the new C99 standard? Yes. > Do you have a compiler > that supports it? Unfortunately no. I haven't heard of any C99 compiler yet. I hope they start adding things like this to gcc soon. - -- Göran Larsson Senior Systems Analyst hoh AT approve DOT se --------------------------- Newsgroups: comp.os.vxworks Subject: How to use the hash table in vxWorks? Date: Thu, 3 Aug 2000 11:50:55 +0800 From: "spine zhang" Organization: The news server of Changzhou TeleCom Message-ID: <8maq99$pgk$1@news.cz.js.cn> The vxWorks has an internal library hashLib, who knows how to use it, could you give a simple demo? Thanks very much! Spine.zhang --------------------------- Newsgroups: comp.os.vxworks Subject: Question as re: the FAQ as re: Job Postings Date: Thu, 3 Aug 2000 17:25:26 -0400 From: "Jeff Newman" Organization: Posted via Supernews, http://www.supernews.com Message-ID: Hello! My name is Jeff Newman with recruitDYNAMICS.com. I posted a job here awhile back, and somehow it was deleted!!!! I checked the FAQ and there is NOTHING that prohibits job postings. So, I have a few questions: a. Why would my post be deleted, unless someone did it on purpose? b. Where should I post very specific VxWorks jobs, if not here? All suggestions are appreciated!!!!!!!!!!!!!!!!!!!!!!!! Jeff Newman Technical Recruiter recruitDYNAMICS.com www.eclaro.com jbn@recruitdynamics.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Floating point exception code Date: Fri, 04 Aug 2000 01:28:47 GMT From: pkockritz@home.com (Pete Kockritz) Organization: @Home Network Message-ID: References: <398bade9.109174429@goliath2.usenet-access.com> In article <398bade9.109174429@goliath2.usenet-access.com>, dont@bother.me (fried egg) wrote: > Does any one know how to handle an invalid instruction exception on > PPC603. Our target is a PPC603R that does not have floating point > modules. I have tried compiling with -m soft-float. Hoewver vxWorks > seems to have floating point instruction embedded in the library. > Seems FP stores are used to speed up context save in some functions. Should you be using the EC603e libraries instead of the 603e? Regards, Pete - -- +-----------------------------------------------------------------+ | Pete Kockritz mailto:pkockritz@home.com | | This space for rent. | +-----------------------------------------------------------------+ --------------------------- Newsgroups: comp.os.vxworks Subject: Re: alignment & endian Date: Fri, 04 Aug 2000 01:31:25 GMT From: pkockritz@home.com (Pete Kockritz) Organization: @Home Network Message-ID: References: <39894d25$0$233@helios.is.co.za> In article <39894d25$0$233@helios.is.co.za>, "Morne Hartman" wrote: > Hi all, > > Can anyone tell me what alignment VxWorks uses for its data. > > I have an application that must communicate to Windows NT. In NT I am using > Word Alignment. NT also uses little-endian. > > If I send a structure from my NT application to the VxWorks Application, the > data in the structure is lost. However if the data is sent to another NT > application the data in the structure is retained. > > How can I solve this problem? > > Many Thanks > Morne Hartman > mhartman@grintek.com Depends on the platform, the following is for PPC. From the PowerPC Embedded Application Binary Interface > Aggregates and Unions > Unlike the SVR4 ABI, an array, structure or union containing a long > double shall start aligned on an 8 byte boundary. However, as in the > SVR4 ABI, a long double member within a structure or union shall start > at the lowest available offset aligned on a 16 byte boundary, and the > size of a structure or union with a long double member shall be a > multiple of 16 bytes. Gcc has some options to change the alignment for structs, but be careful using them. Especially if you use 3rd party libraries or drivers that you obtain in binary form. But alignment might not be your only problem. There's the endian problem. You say 'NT also uses little-endian'. You're probably running NT on an Intel processor, which is little-endian. But are you running vxWorks on Intel or PPC? If your vxWorks is on PPC (which is big-endian) you'll have endian problems. If vxWorks is on Intel you shouldn't have an endian problem, but you could have others (what does your compiler on NT use for a 'short'? 16 bits?, 32 bits?). One thing I've found convenient for data that goes between two platforms is to use the network-to-host and host-to-network macros, found in the target/h/netinet/in.h file, (htons - host to network short, htonl, ntohs, ntohl). They allow you to write code that doesn't care whether the platform is big or little endian. The only platform specific things you have to have are 8, 16 and 32 bit variable types (like UINT8, UINT16, UINT32, etc in vxWorks). Those can be adjusted on a particular platform to insure that both platforms agree on the number of bits in the variable types. There are other ways to handle this, sometimes called 'serialization' or 'byte-stream' which basically convert a struct (or other data structure) into a bunch of bytes are sent between platforms that can be decoded on the receiving end using the agreed upon message format. And then there are even more generic ways, like the XDR form, used by RPC. Or even by converting the data to ASCII and then parsing it on the receiving end. Regards, Pete - -- +-----------------------------------------------------------------+ | Pete Kockritz mailto:pkockritz@home.com | | This space for rent. | +-----------------------------------------------------------------+ --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to use the hash table in vxWorks? Date: Thu, 03 Aug 2000 20:21:07 -0400 From: viv Organization: Lucent Technologies Message-ID: <398A0C73.B11D3AB2@lucent.com> References: <8maq99$pgk$1@news.cz.js.cn> There seems to be no hasbLib library. I saw it in the documentation for VxWorks 5.4....Which one are u talking about? spine zhang wrote: > > The vxWorks has an internal library hashLib, who knows how to use it, could > you give a simple demo? > > Thanks very much! > > Spine.zhang --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Question as re: the FAQ as re: Job Postings Date: Fri, 04 Aug 2000 09:40:07 +0530 From: Vijay Kumar Peshkar Organization: Ye 'Ol Disorganized NNTPCache groupie Message-ID: <398A421F.81E84AB5@cisco.com> References: Reply-To: vpeshkar@cisco.com This is a multi-part message in MIME format. - --------------BE1310CC34B393BB15AF060B Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Jeff, The hectic activity in this news group is to SOLVE the vxWorks probs. we face. The group is already receiving lots of mails- though specific to PROBS. So I request you and any other recruiter to STOP mailing ANYTHING but genuine vxWorks doubts. Yes going by the book there's nothing to stop you. But guess, anyone posting such mails will be irritating the subscribers and gain nothing else. I don't know the answer to b)...we already have so many genuine vxWorks probs to fix..... Sorry pal. How about picking the mail ids from newsgroup and trying UR luck individually? good luck, Vijay Jeff Newman wrote: > Hello! My name is Jeff Newman with recruitDYNAMICS.com. I posted a job here > awhile back, and somehow it was deleted!!!! I checked the FAQ and there is > NOTHING that prohibits job postings. So, I have a few questions: > > a. Why would my post be deleted, unless someone did it on purpose? > b. Where should I post very specific VxWorks jobs, if not here? > > All suggestions are appreciated!!!!!!!!!!!!!!!!!!!!!!!! > > Jeff Newman > Technical Recruiter > recruitDYNAMICS.com > www.eclaro.com > jbn@recruitdynamics.com - --------------BE1310CC34B393BB15AF060B Content-Type: text/x-vcard; charset=us-ascii; name="vpeshkar.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Vijay Kumar Peshkar Content-Disposition: attachment; filename="vpeshkar.vcf" begin:vcard n:Kumar.P;Vijay tel;work:080-5530035/36 Extn:3464 x-mozilla-html:FALSE org:Wipro Technologies, K2, Bangalore.;Wishing you blue skies, a couple of daisies, some fresh air and a smile! adr:;;;;;; version:2.1 email;internet:vpeshkar@cisco.com title:Senior Software Engineer note;quoted-printable: o o o o o o o . . . _____________________________________________=0D=0A o _____ || Vijay Kumar .P =20 |=0D=0A .] [__n_n_| DD[ =3D=3D=3D=3D_____ | vpeshkar@cisco.com, vijay_peshkar@hotmail.com | =0D=0A >(________|__|_[_________]_ |___________________________________________|=0D=0A _/oo OOOOO oo` ooo ooo 'o!o!o =20 o!o!o` fn:Vijay Kumar.P end:vcard - --------------BE1310CC34B393BB15AF060B-- --------------------------- Newsgroups: comp.os.vxworks Subject: IP layer "ignores" packets >= 512 bytes Problem Date: Thu, 03 Aug 2000 20:22:36 -0400 From: Dmitriy Zavin Message-ID: <398A0CCC.5A999354@txc.com> Hi guys,

This might sound real weird, and I can't figure it out. I am using the CS8900A ethernet chip. The driver processes the frame, and passes it up to the IP layer by doiing

      do_protocol(pEtherHeader, pMbufChain, &pCS->ArpCom, DataLength);

Now the weird thing is that if DataLength is greater than  or equal to 512,
the IP layer acts as if nothing happened. I can't trace it further because I don't have the source for that function. I checked and on call all values are what they should be. Here is how i am testing this:

( board (192.168.1.3) is the address of the dev board's ethernet card. MTU on both ends is 1500 )

-[~]$ sudo ping -s 483 board -c 2
PING board (192.168.1.3): 483 data bytes
491 bytes from 192.168.1.3: icmp_seq=0 ttl=64 time=93.1 ms
491 bytes from 192.168.1.3: icmp_seq=1 ttl=64 time=91.6 ms

--- board ping statistics ---
2 packets transmitted, 2 packets received, 0% packet loss
round-trip min/avg/max = 91.6/92.3/93.1 ms

This works (483 ping data + 8 ICMP header + 20 IP header = 511 bytes and it still works).

But doiing this:

-[~]$ sudo ping -s 484 board -c 5
PING board (192.168.1.3): 484 data bytes
--- board ping statistics ---
5 packets transmitted, 0 packets received, 100% packet loss

This does not. If you do the calculation, 484+8+20 = 512.

I do not know why 512 is significant, but its such a "round" number that it cannot be a coincidence :)
Has anyone else encountered this???

By the way,  its not just ICMP. Sending a UDP packet which results in a 512 IP packet does not work either. How can I attempt to fix something like this?

Thanks a lot in advance,  any help would be greatly appreciated.

--Dima 




-- 
/----------------------------------/
Dmitriy Zavin
Software Engineering Intern
TranSwitch Corporation
email: dzavin@txc.com
Ph: 203-929-8810 ext. 2624
/----------------------------------/

 


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: sntpclib, localtime() and TIMEZONE
Date: Fri, 04 Aug 2000 01:15:41 -0400
From: Al Johnston 
Organization: MindSpring Enterprises
Message-ID: <398A517D.AFECBABB@mindspring.com>
References: <3983D15C.1233E21E@mindspring.com> <8m20l4$fto$1@nnrp1.deja.com> <398590A0.B4B5B07D@mindspring.com> <8m4ikk$are$1@nnrp1.deja.com>

> They are in fact logical, but set up for UTC. I can send you the
> structure in source form if you want to see it...

No need.  Things are working quit nicely now.  thanks.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: How to use the hash table in vxWorks?
Date: Fri, 4 Aug 2000 09:41:38 +0800
From: "spine zhang" 
Organization: The news server of Changzhou TeleCom
Message-ID: <8md73b$27ro$1@news.cz.js.cn>
References: <8maq99$pgk$1@news.cz.js.cn> <398A0C73.B11D3AB2@lucent.com>

Which documentation? I saw the definition in hashLib.h, it is not explained
in the help file. Please give me advice how to use it. Thanks!

Spine Zhang

viv  wrote in message
news:398A0C73.B11D3AB2@lucent.com...
> There seems to be no hasbLib library. I saw it in the documentation for
> VxWorks 5.4....Which one are u talking about?
>
> spine zhang wrote:
> >
> > The vxWorks has an internal library hashLib, who knows how to use it,
could
> > you give a simple demo?
> >
> > Thanks very much!
> >
> > Spine.zhang



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: dos fs access and ct69000 drivers
Date: Fri, 04 Aug 2000 08:16:19 +0200
From: Ole Asbjorn Fadum 
Organization: Telenor Online Public Access
Message-ID: <398A5FB2.A6656EFA@scanmar.no>
References: <006901bffcd0$0761a4a0$41f122aa@muc> <8mbs4e$na8$1@nnrp1.deja.com>

Hi,

It's been a long time I did it but the 65555 and the 69000 is mapped into
different memory ranges. You have to make/change a mapentry in
%WINDBASE%\target\config\pcPentium\sysLib.c in the array PHYS_MEM_DESC
sysPhysMemDesc [] =

I have the following entry:

#if 0
    /* 65550 screen buffer memory map*/
    {
      (void *) 0xFF000000,/* Physical mapping */
      (void *) 0xFF000000,/* virtual mapping */
      0x200000,           /* size of segment */
      VM_STATE_MASK_FOR_ALL,
      VM_STATE_FOR_IO
    },
#else
    /* 69000 screen buffer memory map*/
    {
      (void *) 0xE0000000,/* Physical mapping */
      (void *) 0xE0000000,/* virtual mapping */
      0x200000,           /* size of segment */
      VM_STATE_MASK_FOR_ALL,
      VM_STATE_FOR_IO
    },
#endif

According to my notes I used the 65555 driver on a 69000 chip. After that
we have got a driver with 1024x768x16 resolution, done by WRS in Sweden.


Rick Pierpont wrote:

> Hi Jason,
>
> You are very close.  When you make your call to dosFsDevInit() with
> NULL (or 0) as the last parameter you are implying that it can get its
> own configuration from the floppy's boot sector.  If there is no floppy
> in the drive, you will not get your device since it doesn't know the
> configuration.  However, you can pass a DOS_VOL_CONFIG * as the last
> parameter that is filled in with the parameters of the drive that you
> are using (see pg.187 of the VxWorks programmer's guide for specifics).
>
> I've heard some people suggest to modify usrFd.c (in %WINDBASE%
> \target\src\config) to add this initialization code to usrFdConfig()
> but I don't like to obscure the release files so I simply wrote similar
> function that just encapsulates the two function calls you are using
> along with filling in the structure.  It will probably work fine for
> you since I'm assuming that you are also using a 1.44M 3.5" floppy
> drive.  Here it is:
>
> void myFdConfig(int drive, int fdType, char *fdName)
> {
>         BLK_DEV *pBlkDev;
>         DOS_VOL_CONFIG configStruct;
>
>         pBlkDev = (BLK_DEV *) fdDevCreate(drive, fdType, 0, 0);
>         configStruct.dosvc_mediaByte = 0xF0;
>         configStruct.dosvc_secPerClust = 1;
>         configStruct.dosvc_nResrvd = 1;
>         configStruct.dosvc_nFats = 2;
>         configStruct.dosvc_secPerFat = 9;
>         configStruct.dosvc_maxRootEnts = 224;
>         configStruct.dosvc_nHidden = 0;
>         configStruct.dosvc_options = 0;
>         configStruct.dosvc_reserved = 1;
>
>         dosFsDevInit(fdName, pBlkDev, &configStruct);
> }
>
> If you call this function with the parameters:
>
> myFdConfig(0, 0, "cf");
>
> You will get what you are looking for.
>
> As for the 69000 chipset, I'm also working on the same problem.  I
> Dl'ed some 69000 drivers from the WRS site followed the instructions
> that came with them and have not had them work yet  (get exception
> every time).  From what I understand, the only difference between the
> 65555 and 69000 chipset drivers is the initialization routines.  What
> resolution are you planning on running?  If you are running 640x480x16
> color (mode 12h) then the standard VGA driver will work fine for you.
> Please let me know if you get anywhere with this too.
>
> Hope this helps!
>
> --
> Rick Pierpont
> Software Engineer
> Becton Dickinson Biosciences
> Sparks, MD 21152
>
> In article <006901bffcd0$0761a4a0$41f122aa@muc>,
>   "Jason Ahmad"  wrote:
> > Hi,
> >
> > I'm currently working on a project which requires MS-DOS fs disk
> access
> > under vxworks and have run into a few problems setting things up. As
> it
> > stands right now, my x86 target boots off ide0 (with the boot file
> specified
> > as "/ata0/vxWorks") but no FD is created to access the drive once
> everything
> > is up and running. Both ATA and MS-DOS drivers are compiled into the
> image,
> > and a quick check of iosDrvShow confirms the dosFs drivers are up and
> > running.
> >
> > In an attempt to setup an FD, I ran ataDevCreate(0,0,0,0); and used
> the
> > resulting address in dosFsDevInit("cf", blkdev, 0); however,
> dosFsDevInit
> > consistantly returns 0. I cannot seem to get past this step. The one
> thing I
> > can think of is I may be using the wrong value for the final
> parameter of
> > ataDevCreate. (The programmers guide mentions that you need to
> specify a
> > block offset for dos disks with partitions, but I have no idea what
> that
> > offset should be for my disk.)
> >
> > Does anyone have any insight as to how to figure out what the correct
> offset
> > should be and or get the disk mounted with a Dev handle?
> >
> > In addition, I was wondering about the availability of Chips and
> Technology
> > 69000 video drivers. I heard a rumor that the 65555 drivers included
> will
> > work, but I have had no luck with them.
> >
> > Thanks in advance,
> > Jason Ahmad
> >
> >
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.


---------------------------

Newsgroups: comp.os.vxworks
Subject: MPC555LFMZP40 Microcontroller
Date: Fri, 04 Aug 2000 02:29:42 GMT
From: Joe Grosholz 
Message-ID: <398A2C1A.2AA191BE@bellatlantic.net>

I am looking for someone who has any MPC555LFMZP40 microcontrollers they

are willing to part with.  I am willing to pay top dollar for up to 10
pieces that can be delivered to me by Wednesday August 9, 2000.  Please
get in touch if you have any of these parts.

Joe


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: TFTP server problem on DOS FS 2.0
Date: Fri, 04 Aug 2000 08:49:10 +0200
From: Uwe Fuhr 
Organization: Ericsson Eurolab Deutschland GmbH
Message-ID: <398A6766.13D2E514@eede.ericsson.se>
References: <3986B57C.F7BA34E9@eede.ericsson.se> <39890BAA.AB37383E@bitband.com>
Reply-To: Uwe.Fuhr@eed.ericsson.se

Hi Leonid,


> tftpdDebug = 1
> tftpVerbose = 1
> tftpTrace = 1

using this variables brougth nice and usefull printouts.

Are there any other values possible when using the variables
to get more detailed information?

Thank you ...

/Uwe
- -- 
___________________________________________________________________________

     _/  _/  _/  Uwe Fuhr                       EED/E/D/U    Access Design    
    _/  _/  _/   Ericsson Eurolab Deutschland   phone  : +49-5121 - 707104
   _/  _/  _/    Daimlerring 9                  FAX    : +49-5121 - 707314
  _/  _/  _/     31135 Hildesheim - Germany     mobile : +49-173 - 5430771  

__________________________________   mailto:Uwe.Fuhr@eed.ericsson.se   ____
                                  \___________________________________/

---------------------------

Newsgroups: comp.os.vxworks
Subject: help for LoadModule
Date: Fri, 4 Aug 2000 10:35:51 +0800
From: "tanghaoyu" 
Organization: The news server of Changzhou TeleCom
Message-ID: <8mda0f$29di$1@news.cz.js.cn>

Hello All,

When I load an object file by calling LoadModule() or ld (1)

Organization: HACOM Internet - Amersfoort - The Netherlands. (033-455611)
Message-ID: <965373049.235344@big.hacom.nl>
References: <8mbno5$osp$1@overload.lbl.gov>

http://www.windriver.com/cgi-bin/corporate/tradeWinds/viewProd.cgi?id=323

if this is not what you're looking for, just go to the windriversite
(http://www.wrs.com/) and search for ipsec


Frank


"stevens"  wrote in message
news:8mbno5$osp$1@overload.lbl.gov...
> Hi VxWorkers,
>
> we need an IPSEC-Implementation for VxWorks (Tornado2).
> Does somebody know where we can get a solution ?
>
> Regards,
> Torsten
>
> --
>
> **************************************************************************
> fraunhofer institute of microelectronic circuits and systems
> department sat
> name       torsten stevens           mailto:stevens@ims.fhg.de
> address    finkenstrasse 61          phone      +0049/203/3783-238
>            47057 duisburg            fax        +0049/203/3783-266
>            germany
> internet   http://www.ims.fhg.de and http://www.inhaus-nrw.de
> **************************************************************************
>
>



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Problems w/ select() and timeout on mvme2700
Date: Fri, 04 Aug 2000 02:58:56 GMT
From: paulvroberts@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8mdbhf$sns$1@nnrp1.deja.com>
References: <8m9d1u$h91$1@newsfeed.cv.nrao.edu>

Hi Jim,

This is a bit of a guess, but here goes.

From the code you have below, it looks like the select timeout is set to
10000nS, ie 10 mS, which is less than the granularity of the system
clock (1/60 sec by default on MVME2700). Perhaps Vxworks is seeing this
value as "0", and thereby returning straight away?

Paul Roberts.

In article <8m9d1u$h91$1@newsfeed.cv.nrao.edu>,
  "Jim Pisano"  wrote:
> I'm having a problem with select() on a TCP socket and timeouts.
> When I remove the timeout (i.e., set the pointer to the timeval struct
to be
> 0)everything works correctly, but when I insert a non-null timeout
into
> select(), it always returns 0, i.e., a timeout occurred w/out any file
> descriptors ready.
>
> I have set the timeout value to many seconds to ensure that the
select()
> function waits long enough for data from the client app.
>
> I'm running VxWorks 5.4 on Tornado II hosted on Win NT. The target is
an
> mvme-2700.
>
> here's a snippet of code:
>
> // create & open server socket (all details left out except for
"accept()"
> call...
>     m_ACC_ClientSocketFD = accept (m_listenSocket, (struct sockaddr *)
> &clientAddr, &sockAddrSize);
>
> // set up file descriptor set structure for select() call
>     int m_width;
>     fd_set m_fdSet;
>     int m_ACC_ClientSocketFD;
>
>     FD_ZERO( &m_fdSet );
>     FD_SET( m_ACC_ClientSocketFD, &m_fdSet );
>     m_width = m_ACC_ClientSocketFD + 1;
>
> // set up time out for 10 ms
> struct timeval timeout = {0,10000};
>
>  int selectVal = select( m_width, &m_fdSet, NULL, NULL, &timeout );
>  if( selectVal > 0 )
>  {
>      bzero(buffer,100);
>      if( read( m_ACC_ClientSocketFD, buffer, 100) > 0 )
>      {
>        // process buffer ...
>      }
>  }
>
> Thanks.
> -Jim
>
>


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: VxWorks FTP boot problem (T2, MVME2700)
Date: Fri, 04 Aug 2000 03:16:33 GMT
From: paulvroberts@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8mdcif$tgl$1@nnrp1.deja.com>

Hi,

We are having trouble with ftp boot using the VxWorks boot rom loader,
and I'm wondering if anyone else has seen this before or has any
suggestions or troubleshooting hints.

The problem occurs when the ftp server used is the one on Red Hat Linux
6.2 [uu ftpd Version wu-2.6.0(1) Mon Feb 28
10:30:36 EST 2000)]. It doesn't seem to happen with Red Hat Linux 6.0
[uu ftp Version wu-2.4.2-VR17(1) Mon Apr 19 09:21:53 EDT 1999] or an old
Windows ftp server (WFTPD from 1995).

What happens is that the VxWorks boot loader appears to load
the image ok, but doesn't continue the boot process - what we see is it
saying "loading....", followed by load numbers increasing, then a new
line, and then it just sits there.

A "netstat" on the ftp server end revealed that the ftp data connections
are in the FIN_WAIT2 state - from what I know of TCP (only pieces!), I
think this means that the ftp server has closed its end, and is waiting
for VXWorks to finish the connection.

I suspect that for some reason, the VxWorks ftp boot loader is not
handling the case where the ftp server closes the connection before it
does (TCP half close?).

We are using Tornado 2, VxWorks 5.4 and the MVME 2700 BSP. It doesn't
make any difference if I am booting a VxWorks image
with or without a downloadable symbol table.

Now it might be ok to live with this at the moment (only use Red Hat
6.0), but I am concerned that this could be a problem in future years
when that restriction is no longer appropriate.

Any ideas? Thanks for any help,

Paul Roberts


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: help for LoadModule
Date: Fri, 04 Aug 2000 07:48:27 GMT
From: Urban Lindberg 
Organization: Deja.com - Before you buy.
Message-ID: <8mdsgb$8vn$1@nnrp1.deja.com>
References: <8mda0f$29di$1@news.cz.js.cn>

Hi!

You need to add the -mlongcall option to your compilation.

HTH
Urban

In article <8mda0f$29di$1@news.cz.js.cn>,
  "tanghaoyu"  wrote:
> Hello All,
>
> When I load an object file by calling LoadModule() or ld (1)
> 
> Relocation value does not fit in 24 bits.
> Relocation value does not fit in 24 bits
> ......
>
> But if I download this file by Tornado II,it is ok.
> Who can help me?
>
> Tang Haoyu
>
>


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Frequency of TimeStampCounter (TSC)?
Date: Fri, 04 Aug 2000 16:00:39 +0530
From: Yusuf Motiwala 
Organization: Lucent Technologies, Columbus, Ohio
Message-ID: <398A9B4F.F46EF7B4@lucent.com>
References: <3988300E.4DE80833@meco.nl>

It's CPU Freq but to get it you need to calibrate the TSC. However, make
sure you make some dummy call to get TSC before actually using it because
initial calls to TSC does not reflect proper values.

Yusuf

Albert Wijnja wrote:

> Hello,
>
> In my application I use the pentiumTscGet64() routine to get the current
> counter value of the TSC. Is there anyway to determine the frequency of
> this counter? I use this routine for timestamping some events in my
> application. Eventually I want to represent the events in
> (micro)-seconds.
>
> Thanks
>
> Albert Wijnja


---------------------------

End of New-News digest
**********************


From vxwexplo-errs@csg.lbl.gov  Fri Aug  4 07:02:03 2000
From: stevens 
Date: Fri Aug  4 07:02:05 PDT 2000
Subject: Re: VxWorks and IPSEC

Thanks a lot Frank,

this is a good starting point for us.
I didn't saw it on the web page and our customer didn't tell.
The main problem is that most IPSEC-Implementation need sourcecode from the IP-Stack.
Maybe they found a solution with Windriver or they implemented it a different way.
If there is someone out there who uses the stack for VxWorks, please give me a signal.

Torsten


http://www.windriver.com/cgi-bin/corporate/tradeWinds/viewProd.cgi?id=323

if this is not what you're looking for, just go to the windriversite
(http://www.wrs.com/) and search for ipsec


Frank

> Hi VxWorkers,
>
> we need an IPSEC-Implementation for VxWorks (Tornado2).
> Does somebody know where we can get a solution ?
>
> Regards,
> Torsten

--

**************************************************************************
fraunhofer institute of microelectronic circuits and systems
department sat
name       torsten stevens           mailto:stevens@ims.fhg.de
address    finkenstrasse 61          phone      +0049/203/3783-238
           47057 duisburg            fax        +0049/203/3783-266
           germany
internet   http://www.ims.fhg.de and http://www.inhaus-nrw.de
**************************************************************************




From vxwexplo-errs@csg.lbl.gov  Sat Aug  5 04:03:08 2000
From: Vxworks Exploder 
Date: Sat Aug  5 04:03:11 PDT 2000
Subject: comp.os.vxworks newsdigest

Comp.Os.Vxworks Daily Digest    Sat Aug  5 04:03:04 PDT 2000

        Subject: Re: gcc-2.95.2-2, VxWorks, CygWin: Can't build it.
        Subject: Re: TCP Dump for VxWorks?
        Subject: Re: VxWorks FTP boot problem (T2, MVME2700)
        Subject: Re: Problem in booting from network
        Subject: re: VxWorks and IPSEC
        Subject: Re: Hdlc driver
        Subject: Re: Question as re: the FAQ as re: Job Postings
        Subject: shared memory: 2 vxWorks builds?
        Subject: Re: IP layer "ignores" packets >= 512 bytes Problem
        Subject: Re: IP layer "ignores" packets >= 512 bytes Problem
        Subject: Re: TCP Dump for VxWorks?
        Subject: PPP termination
        Subject: Re: gcc-2.95.2-2, VxWorks, CygWin: Can't build it.
        Subject: Re: Floating point exception code
        Subject: Re: shared memory: 2 vxWorks builds?
        Subject: VxWorks 5.4 and DNS
        Subject: fei82557End.c File
        Subject: Re: shared memory: 2 vxWorks builds?
        Subject: JWorks - mouse and font questions????
        Subject: Re: shared memory: 2 vxWorks builds?
        Subject: Arcom MediaGX 233 Board + CT69000 Chipset - need ~Help!!!
        Subject: Some Quesions about Mux
        Subject: Some Quesions about Mux
        Subject: "AGENT_RT_ERROR" when using End driver
        Subject: mv2700 and "initial stack" memory region.
        Subject: Re: shared memory: 2 vxWorks builds?
        Subject: how does purify work?

-------------------------------------------------------

Newsgroups: comp.os.vxworks
Subject: Re: gcc-2.95.2-2, VxWorks, CygWin: Can't build it.
Date: Fri, 04 Aug 2000 12:29:45 GMT
From: big_dope@bigfoot.com (Ami Ghua)
Organization: Newsfeeds.com http://www.newsfeeds.com 80,000+ UNCENSORED Newsgroups.
Message-ID: <398bb68d.176932209@news.usenet-access.com>
References: <%BYh5.1300$Vc.882168@newsr2.u-net.net> <3989cf0e.14192527@24.0.240.30> 

This is very scary. I plan to use T2 for a production PPC cpu board. 
Does there exist any safe optimization flags that will avoid all these
problem optimizations? 

Ami

On Thu, 3 Aug 2000 13:54:18 +0100, "Dave Korn"
 wrote:

>>On Wed, 2 Aug 2000 16:38:50 +0100, "Dave Korn"
>> wrote:
>>>      Hi everyone,
>>>
>>>I'm being driven slowly out of my mind.
>>>
>>>I am trying to build gcc-2.95.2-2 for the powerpc-wrs-vxworks target
>>>on a PC using NT4.0sp6 and a fairly recent (1.1.2) download of cygwin.
>>>
>>>It goes dreadfully wrong, no matter what I try.   [...snip!...]
>
>Gary Beck wrote in message <3989cf0e.14192527@24.0.240.30>...
>>I am also interested in how to compile gcc-2.95.2-2 for the
>>powerpc-wrs-vxworks target.  I know that Tornado 3.0 (still in beta)
>>will be using GNU 2.95 for their toolset but we need to work with
>>Tornado 2.0, VxWorks 5.4.
>
>  I'm fairly sure that the problems only exist under cygwin, as I've had
>reports that it builds fine under Linux.
>
>>The reason I need GCC 2.95 rather than the Tornado shipped toolset
>>(2.72) is the optimization.  Version 2.95 is much better than 2.72.
>
>  You're not joking. The bugs in the 2.7x optimization are so serious as to
>render it unuseable in production code. We've got a code base of about
>half-a-million lines and bugs relating to the optimisation of short-circuit
>logical expressions and common subexpression elimination have sprinkled
>bugs throughout the compiled code. It's possible to use temporary variables
>holding partial results to work around this bug but there's no way to find
>out where in the code we have expressions that will be miscompiled.
>
>>Also, the compiled Cygnus version 2.95 (powerpc-eabi) does not work
>>with the CrossWind Debugger.
>
>  You might be able to rebuild gdbppc from the latest sources. I don't
>know if that would fix it though. People have discussed this before, so it
>might be worth your while to search a) the windsurf support site, b) the
>various gcc related mailing-list archives at www.gnu.org
>
>      Good luck!
>                     DaveK



- -----= Posted via Newsfeeds.Com, Uncensored Usenet News =-----
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
- -----==  Over 80,000 Newsgroups - 16 Different Servers! =-----

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: TCP Dump for VxWorks?
Date: 4 Aug 2000 12:40:24 GMT
From: cedilnik andrej 
Organization: University of Maryland, Baltimore County
Message-ID: <8medjo$vqp$1@news.umbc.edu>
References: <8m4knu$dih$1@news.umbc.edu> 

Here is my hook:

BOOL 
myinputHook
(
        struct ifnet *pIf,    /* interface packet was received on */
        char         *buffer, /* received packet */
        int          length   /* length of received packet */
        )
{
        printf("#\n");
        return false;
}


This is the registration of the hook:

       char dev[]="ln";
        var=etherInputHookAdd(
                        myinputHook, 
                        dev,
                        0
                        );
 
This does not do anything no matter how much trafic goes
into VxWorks box.

What am I doing wrong?

btw. I have VxWorks 5.4
				Andy

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: VxWorks FTP boot problem (T2, MVME2700)
Date: Fri, 04 Aug 2000 14:07:19 +0400
From: "Arthur V. Grigorian" 
Organization: Auriga, Inc.
Message-ID: <398A95D7.A8BBCE60@auriga.ru>
References: <8mdcif$tgl$1@nnrp1.deja.com>

Hello,

   We had  the same problem with that  FTP server and  the solution was
   to downgrade it.

   Actually your target doesn't hang during  the FTP transfer - it just
   doesn't think it logged out successfully, so if you wait 900 seconds
   (default FTP timeout) it will run the loaded image.

   AFAIR the problem is in some unexpected messages from the FTP server
   like:
	221-You have transferred X bytes in X files.
	221-Total traffic for this session was X bytes in X transfers.
	221-Thank you for using the FTP service on ftp.server.name.
	221 Goodbye.
   instead of just:
	221 Goodbye.

   You can try to hack it yourself - see target/config/all/bootCongig.c
   file, netLoad() function, around ftpCommand(errFd, "QUIT",...) call.

Hope that helps,
__
R2


> Hi,
> 
> We are having trouble with ftp boot using the VxWorks boot rom loader,
> and I'm wondering if anyone else has seen this before or has any
> suggestions or troubleshooting hints.
> 
> The problem occurs when the ftp server used is the one on Red Hat Linux
> 6.2 [uu ftpd Version wu-2.6.0(1) Mon Feb 28
> 10:30:36 EST 2000)]. It doesn't seem to happen with Red Hat Linux 6.0
> [uu ftp Version wu-2.4.2-VR17(1) Mon Apr 19 09:21:53 EDT 1999] or an old
> Windows ftp server (WFTPD from 1995).
> 
> What happens is that the VxWorks boot loader appears to load
> the image ok, but doesn't continue the boot process - what we see is it
> saying "loading....", followed by load numbers increasing, then a new
> line, and then it just sits there.
> 
> A "netstat" on the ftp server end revealed that the ftp data connections
> are in the FIN_WAIT2 state - from what I know of TCP (only pieces!), I
> think this means that the ftp server has closed its end, and is waiting
> for VXWorks to finish the connection.
> 
> I suspect that for some reason, the VxWorks ftp boot loader is not
> handling the case where the ftp server closes the connection before it
> does (TCP half close?).
> 
> We are using Tornado 2, VxWorks 5.4 and the MVME 2700 BSP. It doesn't
> make any difference if I am booting a VxWorks image
> with or without a downloadable symbol table.
> 
> Now it might be ok to live with this at the moment (only use Red Hat
> 6.0), but I am concerned that this could be a problem in future years
> when that restriction is no longer appropriate.
> 
> Any ideas? Thanks for any help,
> 
> Paul Roberts

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Problem in booting from network
Date: Fri, 04 Aug 2000 14:49:54 +0400
From: "Arthur V. Grigorian" 
Organization: Auriga, Inc.
Message-ID: <398A9FD2.7EFAAD16@auriga.ru>
References: <3989394C.5B3B95D4@darts.it>

Hello,

   It looks  like your target  doesn't have  the  corresponding
   gateway in its routing table. See routeShow() and routeAdd()

Regards,
__
R2

>Pietro Regis wrote:
> 
> Hi,
> I have the following problem in booting vxWorks 5.3 from network:
> if the host and the board are in the same subnet, the booting procedure
> finishes with success, but if I had to boot the board from a server
> which is on a different subnet, then vxWorks loads successfully, but the
> loading of the symbol table fails.
> There is something to do to avoid this problem, or what I am trying to
> do is not possible?
> 
> Note that, if possible, I don't want to use BOOTP or to generate a
> release of vxWorks with the symbol table integrated.
> 
> Thanks in advance.
> 
> --
> _________________________________________________________________
> Pietro Regis
> DARTS Engineering
> s-mail: Via G.T. Invrea 14 - 16129 GENOVA (ITALY)
> e-mail: pietro.regis@darts.it
> Tel: +39-010-5761240  Fax: +39-010-5760224

---------------------------

Newsgroups: comp.os.vxworks
Subject: re: VxWorks and IPSEC
Date: Fri, 04 Aug 2000 16:01:53 +0100
From: stevens 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <8mej55$gvn$1@overload.lbl.gov>

Thanks a lot Frank,

this is a good starting point for us.
I didn't saw it on the web page and our customer didn't tell.
The main problem is that most IPSEC-Implementation need sourcecode from the IP-Stack.
Maybe they found a solution with Windriver or they implemented it a different way.
If there is someone out there who uses the stack for VxWorks, please give me a signal.

Torsten


http://www.windriver.com/cgi-bin/corporate/tradeWinds/viewProd.cgi?id=323

if this is not what you're looking for, just go to the windriversite
(http://www.wrs.com/) and search for ipsec


Frank

> Hi VxWorkers,
>
> we need an IPSEC-Implementation for VxWorks (Tornado2).
> Does somebody know where we can get a solution ?
>
> Regards,
> Torsten

- --

**************************************************************************
fraunhofer institute of microelectronic circuits and systems
department sat
name       torsten stevens           mailto:stevens@ims.fhg.de
address    finkenstrasse 61          phone      +0049/203/3783-238
           47057 duisburg            fax        +0049/203/3783-266
           germany
internet   http://www.ims.fhg.de and http://www.inhaus-nrw.de
**************************************************************************



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Hdlc driver
Date: Fri, 4 Aug 2000 10:08:06 -0500
From: "Tim Michals" 
Organization: Posted via Supernews, http://www.supernews.com
Message-ID: 
References: <8m8v0a$97q$1@news.netvision.net.il>

The Motorola web site has some good example code for 860 850 etc
davidz  wrote in message
news:8m8v0a$97q$1@news.netvision.net.il...
> Hi !
> Can any body have a good driver for hdlc based on PowerPC !
> Best Regards
>
> David
>
>



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Question as re: the FAQ as re: Job Postings
Date: Fri, 4 Aug 2000 09:03:37 -0400
From: "Jeff Newman" 
Organization: Posted via Supernews, http://www.supernews.com
Message-ID: 
References:  <398A421F.81E84AB5@cisco.com>

Vijay,

I understnad what you are saying, but I think that would be even WORSE spam
(To email individually). The average person switches jobs 7 times in their
life. I can't even guess for a consultant!!! (I handle consulting,
primarily) I am a senior recruiter, and I am not a "Database filler" I don't
have time to waste on BS like that. If I say I have a Job, I have one! The
only reason I don't mention the company, or get more specific, is because of
how competitive our business is. I don't want to irritate anyone! I do know
that VxWorks jobs are not as common as a Java Developer, and people may want
to hear what is going on..... Perhaps we can stat a comp.os.vxworks.jobs
group!

Jeff


"Vijay Kumar Peshkar"  wrote in message
news:398A421F.81E84AB5@cisco.com...
> Jeff,
>
> The hectic activity in this news group is to SOLVE the vxWorks probs. we
face.
> The group is
> already receiving lots of mails- though specific to PROBS. So I request
you and
> any other
> recruiter to STOP mailing ANYTHING but genuine vxWorks doubts. Yes going
by
> the book there's nothing to stop you. But guess, anyone posting such mails
will
> be
> irritating the subscribers and gain nothing else.
> I don't know the answer to b)...we already  have so many genuine vxWorks
probs
> to fix.....
> Sorry pal. How about picking the mail ids from newsgroup and trying UR
luck
> individually?
>
> good luck,
> Vijay
>
>
>
> Jeff Newman wrote:
>
> > Hello! My name is Jeff Newman with recruitDYNAMICS.com. I posted a job
here
> > awhile back, and somehow it was deleted!!!! I checked the FAQ and there
is
> > NOTHING that prohibits job postings. So, I have a few questions:
> >
> > a. Why would my post be deleted, unless someone did it on purpose?
> > b. Where should I post very specific VxWorks jobs, if not here?
> >
> > All suggestions are appreciated!!!!!!!!!!!!!!!!!!!!!!!!
> >
> > Jeff Newman
> > Technical Recruiter
> > recruitDYNAMICS.com
> > www.eclaro.com
> > jbn@recruitdynamics.com
>



---------------------------

Newsgroups: comp.os.vxworks
Subject: shared memory: 2 vxWorks builds?
Date: Fri, 04 Aug 2000 12:15:34 -0400
From: Al Johnston 
Organization: MindSpring Enterprises
Message-ID: <398AEC26.7CD4F004@mindspring.com>

The way I read the books, if you wish to uses vxw's "shared memory",
and you want to use the duel ported ram from cpu0, then you must
build 2 different "configurations" of vxworks. one with SM_OFF_BOARD
false, and one with it true.

Is this really necessary?  Sure seems down right stupid to me.

thanks.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: IP layer "ignores" packets >= 512 bytes Problem
Date: Fri, 04 Aug 2000 16:02:45 +0200
From: Heribert Volkhausen 
Organization: T-Online
Message-ID: <398ACD05.1ADF007E@t-online.de>
References: <398A0CCC.5A999354@txc.com>

Hi,

which version of Tornado/VxWorks do you use?

Regards
Heribert

Dmitriy Zavin wrote:
> 
> Hi guys,
> 
> This might sound real weird, and I can't figure it out. I am using the
> CS8900A ethernet chip. The driver processes the frame, and passes it
> up to the IP layer by doiing
> 
>       do_protocol(pEtherHeader, pMbufChain, &pCS->ArpCom, DataLength);
> 
> Now the weird thing is that if DataLength is greater than  or equal to
> 512,
> the IP layer acts as if nothing happened. I can't trace it further
> because I don't have the source for that function. I checked and on
> call all values are what they should be. Here is how i am testing
> this:
> 
> ( board (192.168.1.3) is the address of the dev board's ethernet card.
> MTU on both ends is 1500 )
> 
> -[~]$ sudo ping -s 483 board -c 2
> PING board (192.168.1.3): 483 data bytes
> 491 bytes from 192.168.1.3: icmp_seq=0 ttl=64 time=93.1 ms
> 491 bytes from 192.168.1.3: icmp_seq=1 ttl=64 time=91.6 ms
> 
> --- board ping statistics ---
> 2 packets transmitted, 2 packets received, 0% packet loss
> round-trip min/avg/max = 91.6/92.3/93.1 ms
> 
> This works (483 ping data + 8 ICMP header + 20 IP header = 511 bytes
> and it still works).
> 
> But doiing this:
> 
> -[~]$ sudo ping -s 484 board -c 5
> PING board (192.168.1.3): 484 data bytes
> --- board ping statistics ---
> 5 packets transmitted, 0 packets received, 100% packet loss
> 
> This does not. If you do the calculation, 484+8+20 = 512.
> 
> I do not know why 512 is significant, but its such a "round" number
> that it cannot be a coincidence :)
> Has anyone else encountered this???
> 
> By the way,  its not just ICMP. Sending a UDP packet which results in
> a 512 IP packet does not work either. How can I attempt to fix
> something like this?
> 
> Thanks a lot in advance,  any help would be greatly appreciated.
> 
> --Dima
> 
> --
> /----------------------------------/
> Dmitriy Zavin
> Software Engineering Intern
> TranSwitch Corporation
> email: dzavin@txc.com
> Ph: 203-929-8810 ext. 2624
> /----------------------------------/
> 
>

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: IP layer "ignores" packets >= 512 bytes Problem
Date: Fri, 04 Aug 2000 10:21:29 -0400
From: Dmitriy Zavin 
Message-ID: <398AD169.19C3FD69@txc.com>
References: <398A0CCC.5A999354@txc.com> <398ACD05.1ADF007E@t-online.de>



Thanks for replying,

VxWorks version 5.4

 ]]]]]]]]]]]]]]]]]]]]]]]]]]      
KERNEL: WIND version 2.5

 ]]]]]]]]]]]]]]]]]]]]]]]]]      
Copyright Wind River Systems, Inc., 1984-1999

Memory Size: 0x200000.  BSP version 1.2/0.

Thats what i have..

Thanks.

Heribert Volkhausen wrote:

Hi,

which version of Tornado/VxWorks do you use?

Regards

Heribert

Dmitriy Zavin wrote:

>

> Hi guys,

>

> This might sound real weird, and I can't figure it out. I am using
the

> CS8900A ethernet chip. The driver processes the frame, and passes
it

> up to the IP layer by doiing

>

>       do_protocol(pEtherHeader, pMbufChain,
&pCS->ArpCom, DataLength);

>

> Now the weird thing is that if DataLength is greater than  or
equal to

> 512,

> the IP layer acts as if nothing happened. I can't trace it further

> because I don't have the source for that function. I checked and
on

> call all values are what they should be. Here is how i am testing

> this:

>

> ( board (192.168.1.3) is the address of the dev board's ethernet
card.

> MTU on both ends is 1500 )

>

> -[~]$ sudo ping -s 483 board -c 2

> PING board (192.168.1.3): 483 data bytes

> 491 bytes from 192.168.1.3: icmp_seq=0 ttl=64 time=93.1 ms

> 491 bytes from 192.168.1.3: icmp_seq=1 ttl=64 time=91.6 ms

>

> --- board ping statistics ---

> 2 packets transmitted, 2 packets received, 0% packet loss

> round-trip min/avg/max = 91.6/92.3/93.1 ms

>

> This works (483 ping data + 8 ICMP header + 20 IP header = 511 bytes

> and it still works).

>

> But doiing this:

>

> -[~]$ sudo ping -s 484 board -c 5

> PING board (192.168.1.3): 484 data bytes

> --- board ping statistics ---

> 5 packets transmitted, 0 packets received, 100% packet loss

>

> This does not. If you do the calculation, 484+8+20 = 512.

>

> I do not know why 512 is significant, but its such a "round" number

> that it cannot be a coincidence :)

> Has anyone else encountered this???

>

> By the way,  its not just ICMP. Sending a UDP packet which results
in

> a 512 IP packet does not work either. How can I attempt to fix

> something like this?

>

> Thanks a lot in advance,  any help would be greatly appreciated.

>

> --Dima

>

> --

> /----------------------------------/

> Dmitriy Zavin

> Software Engineering Intern

> TranSwitch Corporation

> email: dzavin@txc.com

> Ph: 203-929-8810 ext. 2624

> /----------------------------------/

>

>



-- 
/----------------------------------/
Dmitriy Zavin
Software Engineering Intern
TranSwitch Corporation
email: dzavin@txc.com
Ph: 203-929-8810 ext. 2624
/----------------------------------/

 


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: TCP Dump for VxWorks?
Date: Fri, 4 Aug 2000 10:06:47 -0500
From: "Tim Michals" 
Organization: Posted via Supernews, http://www.supernews.com
Message-ID: 
References: <8m4knu$dih$1@news.umbc.edu>  <8medjo$vqp$1@news.umbc.edu>

I see the problem...... please see below

cedilnik andrej  wrote in message
news:8medjo$vqp$1@news.umbc.edu...
> Here is my hook:
>
> BOOL
> myinputHook
> (
>         struct ifnet *pIf,    /* interface packet was received on */
>         char         *buffer, /* received packet */
>         int          length   /* length of received packet */
>         )
> {
>         printf("#\n");
>         return false;
> }
>
>
> This is the registration of the hook:
>
please use showInfo to show all interfaces in the box at the command line
you will notice that #  the # is the instance so ln0 would
be the proper name.  So you are not hooking the ln0 driver!  Please let me
know if this works! Thanks!


>        char dev[]="ln";
>         var=etherInputHookAdd(
>                         myinputHook,
>                         dev,
>                         0
>                         );
>
> This does not do anything no matter how much trafic goes
> into VxWorks box.
>
> What am I doing wrong?
>
> btw. I have VxWorks 5.4
> Andy



---------------------------

Newsgroups: comp.os.vxworks
Subject: PPP termination
Date: 4 Aug 2000 15:06:48 GMT
From: johan@borksoft.xs4all.nl
Organization: XS4ALL Internet BV
Message-ID: <8mem68$p28$1@news1.xs4all.nl>

Hello,

I am working with a PPP link between NT and VxWorks, with VxWorks as
server. I am using LCP echoing to detect a disconnect, and I get to see
this with the debug information. When I look with pppInfoShow I can see
that the state of the LCP-phase has gone from ESTABLISH (before the
connection) to NETWORK (when the connection is established) to TERMINATE.

Is it possible to have the the PPP server restarting itself, so I can
connect again without having to delete the PPP connection and reopening
it?
If this is not possible, is it possible to get the phase information from
the PPP connection? I can find the state info in the PPP_INFO struct, but
not the phase info of the LCP. Or is there another way to detect this
phase without looking at the internals of the PPP-server?

Groeten,
	Johan

- -- 
     o o o o o o o . . .   ___________________________________
    o      _____           ||    Johan Borkhuis              |
  .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
 >(________|__|_[_________]_|________________________________|
 _/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: gcc-2.95.2-2, VxWorks, CygWin: Can't build it.
Date: Fri, 4 Aug 2000 10:10:39 -0500
From: "Tim Michals" 
Organization: Posted via Supernews, http://www.supernews.com
Message-ID: 
References: <%BYh5.1300$Vc.882168@newsr2.u-net.net> <3989cf0e.14192527@24.0.240.30>  <398bb68d.176932209@news.usenet-access.com>

Ecos has directions on build a gcc for the PPC www.redhat.com and it is
under ECOS.
Ami Ghua  wrote in message
news:398bb68d.176932209@news.usenet-access.com...
> This is very scary. I plan to use T2 for a production PPC cpu board.
> Does there exist any safe optimization flags that will avoid all these
> problem optimizations?
>
> Ami
>
> On Thu, 3 Aug 2000 13:54:18 +0100, "Dave Korn"
>  wrote:
>
> >>On Wed, 2 Aug 2000 16:38:50 +0100, "Dave Korn"
> >> wrote:
> >>>      Hi everyone,
> >>>
> >>>I'm being driven slowly out of my mind.
> >>>
> >>>I am trying to build gcc-2.95.2-2 for the powerpc-wrs-vxworks target
> >>>on a PC using NT4.0sp6 and a fairly recent (1.1.2) download of cygwin.
> >>>
> >>>It goes dreadfully wrong, no matter what I try.   [...snip!...]
> >
> >Gary Beck wrote in message <3989cf0e.14192527@24.0.240.30>...
> >>I am also interested in how to compile gcc-2.95.2-2 for the
> >>powerpc-wrs-vxworks target.  I know that Tornado 3.0 (still in beta)
> >>will be using GNU 2.95 for their toolset but we need to work with
> >>Tornado 2.0, VxWorks 5.4.
> >
> >  I'm fairly sure that the problems only exist under cygwin, as I've had
> >reports that it builds fine under Linux.
> >
> >>The reason I need GCC 2.95 rather than the Tornado shipped toolset
> >>(2.72) is the optimization.  Version 2.95 is much better than 2.72.
> >
> >  You're not joking. The bugs in the 2.7x optimization are so serious as
to
> >render it unuseable in production code. We've got a code base of about
> >half-a-million lines and bugs relating to the optimisation of
short-circuit
> >logical expressions and common subexpression elimination have sprinkled
> >bugs throughout the compiled code. It's possible to use temporary
variables
> >holding partial results to work around this bug but there's no way to
find
> >out where in the code we have expressions that will be miscompiled.
> >
> >>Also, the compiled Cygnus version 2.95 (powerpc-eabi) does not work
> >>with the CrossWind Debugger.
> >
> >  You might be able to rebuild gdbppc from the latest sources. I don't
> >know if that would fix it though. People have discussed this before, so
it
> >might be worth your while to search a) the windsurf support site, b) the
> >various gcc related mailing-list archives at www.gnu.org
> >
> >      Good luck!
> >                     DaveK
>
>
>
> -----= Posted via Newsfeeds.Com, Uncensored Usenet News =-----
> http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
> -----==  Over 80,000 Newsgroups - 16 Different Servers! =-----



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Floating point exception code
Date: Fri, 04 Aug 2000 15:08:13 GMT
From: gold@sdsu.edu
Organization: Deja.com - Before you buy.
Message-ID: <8mem8o$rd8$1@nnrp1.deja.com>
References: <398bade9.109174429@goliath2.usenet-access.com> 

In article ,
  pkockritz@home.com (Pete Kockritz) wrote:
> In article <398bade9.109174429@goliath2.usenet-access.com>,
dont@bother.me
> (fried egg) wrote:
>
> > Does any one know how to handle an invalid instruction exception on
> > PPC603. Our target is a PPC603R that does not have floating point
> > modules. I have tried compiling with -m soft-float. Hoewver vxWorks
> > seems to have floating point instruction embedded in the library.
> > Seems FP stores are used to speed up context save in some functions.
>
> Should you be using the EC603e libraries instead of the 603e?
>
And shouldn't you also be doing a
#undef INCLUDE_HW_FP
#define INCLUDE_SW_FP

in config.h or configAll.h

  hth,
   Dan Gold  gold@sdsu.edu
> Pete
> --
> +-----------------------------------------------------------------+
> |       Pete Kockritz                  mailto:pkockritz@home.com  |
> |                       This space for rent.                      |
> +-----------------------------------------------------------------+
>


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: shared memory: 2 vxWorks builds?
Date: Fri, 4 Aug 2000 17:51:43 GMT
From: lee.deraud@boeing.com (Lee DeRaud)
Organization: None to speak of
Message-ID: <398b01dc.17208704@news.boeing.com>
References: <398AEC26.7CD4F004@mindspring.com>
Sender: nntp@news.boeing.com (Boeing NNTP News Access)

On Fri, 04 Aug 2000 12:15:34 -0400, Al Johnston
 wrote:

>The way I read the books, if you wish to uses vxw's "shared memory",
>and you want to use the duel ported ram from cpu0, then you must
>build 2 different "configurations" of vxworks. one with SM_OFF_BOARD
>false, and one with it true.
>
>Is this really necessary?  Sure seems down right stupid to me.

I think that parameter only affects the SM master (cpu 0): it defines
whether to allocate onboard memory for the SM network or use a fied
memory location, presumably on an external board. The clients get the
VME address of that memory from the anchor block on the master...they
neither know (nor care) where the memory physically resides.

Lee

---------------------------

Newsgroups: comp.os.vxworks
Subject: VxWorks 5.4 and DNS
Date: Fri, 4 Aug 2000 12:52:07 -0700
From: "Gregor" 
Organization: SSO-IT, Hewlett-Packard Co.
Message-ID: <8mf6tr$2jv$1@web1.cup.hp.com>

I am using a PowerPC target (MPC860) with VxWorks 5.4. Is there a
possibility to make DNS work? If WindRiver does not provide anything for
DNS, are there reasonable third-party companies that offer DNS solutions for
VxWorks 5.4 running on a PowerPC target?

Thanks,

Gregor




---------------------------

Newsgroups: comp.os.vxworks
Subject: fei82557End.c File
Date: Fri, 04 Aug 2000 10:22:28 -0700
From: asher 
Organization: SBC Internet Services
Message-ID: <398AFBD4.95CF6CBA@lanterncom.com>

Hi,

I'm looking for an up-to-date (working) fei82557End.c and
sysFei82557End.c files for a PPC750 BSP. Can anybody please send it to
me?

emaitol: rapidnet@pacbell.net

Thanks a lot!

A.G.


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: shared memory: 2 vxWorks builds?
Date: Fri, 04 Aug 2000 12:46:59 -0500
From: Andrew Johnson 
Organization: APS, Argonne National Laboratory
Message-ID: <398B0192.5E0223E7@aps.anl.gov>
References: <398AEC26.7CD4F004@mindspring.com>

Hi Al,

Al Johnston wrote:
> 
> The way I read the books, if you wish to uses vxw's "shared memory",
> and you want to use the duel ported ram from cpu0, then you must
> build 2 different "configurations" of vxworks. one with SM_OFF_BOARD
> false, and one with it true.
> 
> Is this really necessary?  Sure seems down right stupid to me.

No, the main purpose of SM_OFF_BOARD seems to be for proc #0, where the
CPU might need to know at compile-time whether the SM network buffers will
be in local RAM or not.  Other processors will always expect the shared
memory to be off-board (ie on cpu#0). You can change the compile-time
switch to a run-time one if you wish, possibly based on the proc number,
although in my sysBusTas() routine I just test the location against the
VME master window registers (or maybe the constants that they are loaded
with) to see whether it's an on-board address or not.

The only thing I had to do to be able to use the same boot file for all
proc #s was to disable a warning message in target/src/config/usrNetwork.c
associated with the Proxy ARP master, which would otherwise have confused
other developers when they used the boot file in a 1-cpu configuration.

- - Andrew
- -- 
Sharks kill 10 people each year.  People kill 60,000,000 sharks each year.

---------------------------

Newsgroups: comp.os.vxworks
Subject: JWorks - mouse and font questions????
Date: Fri, 4 Aug 2000 18:55:23 +0100
From: "Peter Deed" 
Organization: Infocell Telecom
Message-ID: <8mf016$47m$1@taliesin2.netcom.net.uk>

Hi

I am developing a JWorks application which displays images to the screen. I
want to stop the mouse cursor appearing on the screen.

Does anyone know how to do this ? I am using the standard sun.touchable
toolkit for my graphics interface.


Also..has anyone had any luck displaying text to the screen in Jworks with
anything but the standard system font and size.  I think that the
implementation of Personal Jworks I am using does not use the underlying UGL
fonts ..


Any help would be appreciated ..

Regards

Peter Deed
Infocell Telecom





---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: shared memory: 2 vxWorks builds?
Date: Fri, 04 Aug 2000 14:10:05 -0400
From: Al Johnston 
Organization: MindSpring Enterprises
Message-ID: <398B06FD.A980BFD2@mindspring.com>
References: <398AEC26.7CD4F004@mindspring.com> <398B0192.5E0223E7@aps.anl.gov>


Your back!  Thanks for the answer.  I was hoping this was the case.

Understanding "sm" is giving me a bit of trouble.  Partially because
they cant make up their mind as to what "share memory" is... That
term is so overloaded in vxworks I feel like I am using ada! (g)

Best I can figure, using the default values for SM_xxx_ADRS/SIZE,
(INCLUDE_ing the vxMP and NET, but not booting using the sm net)
the area 0x1,0000-2,ffff is reserved for use by one "flavor"
of sm (network booting), the area 0x3,0000-3,ffff is reserved for 
use by vxMP (part is available to user via smMemMalloc()).

I have not been able to account for the dual ported area from
0x4c00 to 0xffff.  this (like all of the vx sm area) starts out
as "reserved" for the "initial stack" (I am told) but once
the system is up, it becomes available... right?  Is vxworks
"loosing" this area? (by this I mean its not using it, and it
will not give it you the user via any xxAlloc() call).


If that is the case, I can use it.  I need 0x1000 byte area in 
d-ported ram.  This seems like an easy way to get it.  Is this
okay?

- -al

---------------------------

Newsgroups: comp.os.vxworks
Subject: Arcom MediaGX 233 Board + CT69000 Chipset - need ~Help!!!
Date: Fri, 4 Aug 2000 19:17:51 +0100
From: "Peter Deed" 
Organization: Infocell Telecom
Message-ID: <8mf1ba$4aa$1@taliesin2.netcom.net.uk>

Hi

I'm tring to run my grahical application on the MediaGX Board which runs a
Cyrix 233 Mhz chip and contains an embedded CT69000 graphics chip.

I have downloaded the driver for this board from Wind River's site and
managed to include it within my build.

It seems to cause a page fault when I call uglInitialize() - specifically
within the set_display_mode() function.

Am I supposed to make an entry within the syslib.c source just like one
would with a CT65555 entry ?


Has anyone ever got this working ??

Thanks in advance

Peter Deed
Infocell Telecom



---------------------------

Newsgroups: comp.os.vxworks
Subject: Some Quesions about Mux
Date: Fri, 4 Aug 2000 12:38:15 +0800
From: "Ren Hantao" 
Organization: The news server of Changzhou TeleCom
Message-ID: <8mdhmg$2fa0$1@news.cz.js.cn>

Hi,

  I am confronted with many difficulties in writing MUX based protocols,
Your help will be appreciated very much,some examples in detail are
prefered.

1. First of all,
  I am not clear about priority mechanism.Chapter 4 in <> says that protoclol type supplied to the bind
function
is used by MUX to prioritize the protocols,MUX_PROTO_SNARF has the highest
priority otherwise MUX_PROTO_PROMISC the lowest.

They are defined in muxLib.h as following,

#define MUX_PROTO_PROMISC 0x100
#define MUX_PROTO_SNARF   0x101
#define MUX_PROTO_OUTPUT  0x102

They are in the ethernet size range(marked as Experimental in rfc 1700)
Initially "muxShow" routing displays the following,
- ----------------------------------------------------------------------------
- --
- -> muxShow
Device: dc Unit: 0
Description: dec21x40 Enhanced Network Driver
Protocol: Wind Debug Agent Type: 257 Recv 0x13ede0 Shutdown 0x0
Protocol: IP 4.4 ARP Type: 2054 Recv 0x1556d4 Shutdown 0x155998
Protocol: IP 4.4 TCP/IP Type: 2048 Recv 0x1556d4 Shutdown 0x1558d0
value = 0 = 0x0
- ->
- ----------------------------------------------------------------------------
- --
Now I want to write a protocol which has higher priority than "IP 4.4 ARP"
and "IP 4.4 TCP" do.
I am confused by which type should I use for my protocol.
I has tested a lot of types such as 2049,259,258 etc,however they can not
get
any packet even if I Shundown the "IP 4,4 xx" protocol with "ipDetach()"
routing.
I can not muxbind() my protocol with type of MUX_PROTO_SNARF because WDB has
used
this type. Protocol with type of MUX_PROTO_PROMISC can get packet until "IP
4.4 xxx"
has been shutdowned.

2. Secondly,
  I want to make use of "IP 4.4 ARP" protocol as my address resolution
function.
but the "arpresolve" function is so complicated.
  arpresolve (struct arpcom *, struct rtentry *, struct mbuf *, struct
sockaddr *, u_char *);
  Lack of document about it make thing worse. Please give me some example
about the use of
arp functions.

3.
   Manual says Routing muxAddressForm() returns the head of an mBLK chain
containing the data
form pMBLK with a prepended link-level header.  When I use it, it really
returns a pMBLK with
a link-level header.  The link-level header (14 byte)is as following,

   00 10 5a 5e d3 bf 08 00 3e 2f 2f 5e 00 00
  ~~~~~~~~~~~~~~~~~  ~~~~~~~~~~~~~~~~~ ~~~~~
  Destionation MAC     Source MAC      Protocol

  Oh,Protocol filed is always 00 00,but I need other types such as 08
00(IP),08 06(ARP)
  I can only chang this filed manually.
  What makes me confused is that transmit functions in wdbEndPktDrv.c  use
only the
muxAddressForm(),but get "08 00" type filed.  Is there any relationgship
between type
in muxBind and type in the link-level formed by muxAddressForm?

4.
 netMblkOffsetToBufCopy() or netMblkBufCopy() can copy a chains of data to a
buffer.
 If I copy data from a long buffer to a mBlk chains ,are there any functions
I can
 use to facilitate the opeation without take the size of cluster into
consideration,
 which can chains the mBlk automatically?

 I also need some help about how to use mbuf sturcture, Is mbuf the same as
mblk?..
 Examples are prefered.


You help is very much appreciated!


Ren Hantao
2000/8/4




---------------------------

Newsgroups: comp.os.vxworks
Subject: Some Quesions about Mux
Date: Fri, 4 Aug 2000 12:42:47 +0800
From: "Ren Hantao" 
Organization: The news server of Changzhou TeleCom
Message-ID: <8mdhnd$2fa0$2@news.cz.js.cn>

Hi,

  I am confronted with many difficulties in writing MUX based protocols,
Your help will be appreciated very much,some examples in detail are
prefered.

1. First of all,
  I am not clear about priority mechanism.Chapter 4 in <> says that protoclol type supplied to the bind
function
is used by MUX to prioritize the protocols,MUX_PROTO_SNARF has the highest
priority otherwise MUX_PROTO_PROMISC the lowest.

They are defined in muxLib.h as following,

#define MUX_PROTO_PROMISC 0x100
#define MUX_PROTO_SNARF   0x101
#define MUX_PROTO_OUTPUT  0x102

They are in the ethernet size range(marked as Experimental in rfc 1700)
Initially "muxShow" routing displays the following,
- ----------------------------------------------------------------------------
- --
- -> muxShow
Device: dc Unit: 0
Description: dec21x40 Enhanced Network Driver
Protocol: Wind Debug Agent Type: 257 Recv 0x13ede0 Shutdown 0x0
Protocol: IP 4.4 ARP Type: 2054 Recv 0x1556d4 Shutdown 0x155998
Protocol: IP 4.4 TCP/IP Type: 2048 Recv 0x1556d4 Shutdown 0x1558d0
value = 0 = 0x0
- ->
- ----------------------------------------------------------------------------
- --
Now I want to write a protocol which has higher priority than "IP 4.4 ARP"
and "IP 4.4 TCP" do.
I am confused by which type should I use for my protocol.
I has tested a lot of types such as 2049,259,258 etc,however they can not
get
any packet even if I Shundown the "IP 4,4 xx" protocol with "ipDetach()"
routing.
I can not muxbind() my protocol with type of MUX_PROTO_SNARF because WDB has
used
this type. Protocol with type of MUX_PROTO_PROMISC can get packet until "IP
4.4 xxx"
has been shutdowned.

2. Secondly,
  I want to make use of "IP 4.4 ARP" protocol as my address resolution
function.
but the "arpresolve" function is so complicated.
  arpresolve (struct arpcom *, struct rtentry *, struct mbuf *, struct
sockaddr *, u_char *);
  Lack of document about it make thing worse. Please give me some example
about the use of
arp functions.

3.
   Manual says Routing muxAddressForm() returns the head of an mBLK chain
containing the data
form pMBLK with a prepended link-level header.  When I use it, it really
returns a pMBLK with
a link-level header.  The link-level header (14 byte)is as following,

   00 10 5a 5e d3 bf 08 00 3e 2f 2f 5e 00 00
  ~~~~~~~~~~~~~~~~~  ~~~~~~~~~~~~~~~~~ ~~~~~
  Destionation MAC     Source MAC      Protocol

  Oh,Protocol filed is always 00 00,but I need other types such as 08
00(IP),08 06(ARP)
  I can only chang this filed manually.
  What makes me confused is that transmit functions in wdbEndPktDrv.c  use
only the
muxAddressForm(),but get "08 00" type filed.  Is there any relationgship
between type
in muxBind and type in the link-level formed by muxAddressForm?

4.
 netMblkOffsetToBufCopy() or netMblkBufCopy() can copy a chains of data to a
buffer.
 If I copy data from a long buffer to a mBlk chains ,are there any functions
I can
 use to facilitate the opeation without take the size of cluster into
consideration,
 which can chains the mBlk automatically?

 I also need some help about how to use mbuf sturcture, Is mbuf the same as
mblk?..
 Examples are prefered.


You help is very much appreciated!


Ren Hantao
2000/8/4






---------------------------

Newsgroups: comp.os.vxworks
Subject: "AGENT_RT_ERROR" when using End driver
Date: Fri, 4 Aug 2000 12:27:37 +0800
From: "Ren Hantao" 
Organization: The news server of Changzhou TeleCom
Message-ID: <8mdgp5$2eto$1@news.cz.js.cn>

I replaced common ethernet driver with END driver
on the MPC860 board yesterday. But when I connect
to the target and lunch the shell, there are always
some errors when I use some commands, like "ping",
"semCCreate", etc. It report "AGENT_RT_ERROR" in the
shell. What does this mean? Does anybody know where
is wrong?





---------------------------

Newsgroups: comp.os.vxworks
Subject: mv2700 and "initial stack" memory region.
Date: Fri, 04 Aug 2000 16:48:35 -0400
From: Al Johnston 
Organization: MindSpring Enterprises
Message-ID: <398B2C23.33F473A9@mindspring.com>

The memory map for this bsp labels the area from 0xfc00 to
"bsp dependent value" (which is 0x10,0000 for the bsp) as
"initial stack".  My understanding is that this area is used
as a stack during the startup of vxworks.  Once the root task
has started, this area's use as a stack is ended.

Is that correct?

Next, during startup, part of this "initial stack" area is 
used by various parts of vxw's "shared memory" thingie.  For
this bsp (with default settings for SM address/sizes and with
bp and vxMP INCLUDE's included) part of this "initial stack"
area is put back into service.  The back in service parts are
from address 0x1,0000 to 0x3,ffff.  

This leaves the areas 0x4c00 -> 0xffff and 0x4,0000 -> 0xf,ffff
still not being used by the os or being made available by the
os to the applications.

Is this really the case?  If so I would like to steal a few kbyts
in the 0x4c00:0xffff "not in service" area.

Anybody know if these two "not in service" areas are really unused?

thanks.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: shared memory: 2 vxWorks builds?
Date: Fri, 04 Aug 2000 16:50:43 -0400
From: Al Johnston 
Organization: MindSpring Enterprises
Message-ID: <398B2CA3.995AE431@mindspring.com>
References: <398AEC26.7CD4F004@mindspring.com> <398b01dc.17208704@news.boeing.com>


> I think that parameter only affects the SM master (cpu 0): it defines

Thats the way I remembered it being back in our hkv30 days (6-7 yrs
ago).  I was hoping it was still the case.  But the doc sure seems
to say otherwise... in several places!

thanks for the answer... esp. since it was the answer I wanted -g-.

---------------------------

Newsgroups: comp.os.vxworks,comp.compilers,comp.programming
Subject: how does purify work?
Date: 4 Aug 2000 21:40:24 -0400
From: toddhoff@my-deja.com
Organization: Deja.com - Before you buy.
Keywords: debug, tools, question, comment
Message-ID: <00-08-027@comp.compilers>

Once upon a time Purify did work for vxworks.  It would be a very nice
tool to have and it seems very odd that there's no product with
equivalent functionality.

If one wanted to build a Purify like tool, does anyone know how it
would be done? I don't really understand what it does and haven't
found much related info.

thanx
[Doesn't it stick extra code into your program to audit all of the
allocation, freeing, and pointer references?  Dunno if it hacks the
source or rewrites the object. -John]

---------------------------

End of New-News digest
**********************


From vxwexplo-errs@csg.lbl.gov  Sat Aug  5 13:05:52 2000
From: "Bridget OConnor" 
Date: Sat Aug  5 13:05:54 PDT 2000
Subject: VxWorks programmers needed in Boca Raton, FL

This is a multi-part message in MIME format.

------=_NextPart_000_01C7_01BFFEF6.F501A980
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

Hello All,

Below is a description of a terrific opportunity with an exciting (and
already successful) Boca Raton company preparing to IPO soon.  If you're =
not
interested in a new position, please skip this post or pass it along to =
a
friend.  You'd be doing them a favor as it really is a great company =
with
phenomenal potential.  Thanks a bunch for your time!

C/C++ Programmers with VxWorks / Unix Experience

Our client's Chairman started this pre-IPO Internet Services company two
years ago. His first start-up was listed on the New York Stock Exchange
until recently merging with another large company. All indications are =
that
this endeavor will be even more successful than his first. The company =
has
nearly 200 employees now and is preparing itself to go public in the =
very
near future. The company by the way will go public making a profit which =
is
virtually unheard of in an internet start-up.  This envious financial
position will allow them to time their IPO for the best market =
conditions.

This Information Technology Solutions company has built one of the =
largest
and most sophisticated computer systems in the world with over 30 =
TERAbytes
of marketing data and growing fast. Despite the enormous size of the =
data
warehouse, queried information is returned to the user virtually
instantaneously via the internet. This proprietary system simply has to =
be
seen to be believed.

Companies integrating with our client's access engine are able to =
improve
the effectiveness of their Customer Relationship Management (CRM) system
providing a competitive advantage in the market place.

Our client is not resting on its laurels however.  A subsidiary of the
company is working on creating in simplistic terms, a super computer.  =
To
that end, they are in need of some brilliant programmers with the =
following
characteristics:

- Recent graduate with Master's degree in Computer Science, or with
Bachelor's degree and professional C++ programming experience.
- Solid knowledge of application programming under Unix and VxWorks.
- Experience with parallel programming, Sun/Solaris, Windows NT, and
knowledge of Graph Theory a BIG plus.

Our client wants the best and is willing to pay for them.  Those meeting
their rigorous requirements will enjoy all the benefits of the parent
company: excellent salaries (the majority of experienced candidates I've
placed with the company have been over 100k), pre-IPO stock options,
terrific benefits, and relocation assistance.  This isn't your typical
start-up that is low on funds.  The company has all of the benefits of a
start-up including exciting, non-bureaucratic leading edge development =
but all of the comforts of a larger established company including =
financial stability, excellent bennies, state of the art equipment and =
facility, and much more.  It could be the last job you'll ever need.

If you're interested in this position, please forward your resume with
"VxWorks" in the subject line to resume@bkoenterprises.com .  =
Alternatively,
you may fax us at 800-842-7861.  For more information on the position, =
you
may also call or e-mail:

Bridget O'Connor
800-385-1914
bko@bkoenterprises.com

Thanks again for your time!










------=_NextPart_000_01C7_01BFFEF6.F501A980
Content-Type: text/html;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable









Hello All,

Below is a =
description of a=20
terrific opportunity with an exciting (and
already successful) Boca =
Raton=20
company preparing to IPO soon.  If you're not
interested in a =
new=20
position, please skip this post or pass it along to a
friend.  =
You'd be=20
doing them a favor as it really is a great company with
phenomenal=20
potential.  Thanks a bunch for your time!

C/C++ Programmers =
with=20
VxWorks / Unix Experience

Our client's Chairman started this =
pre-IPO=20
Internet Services company two
years ago. His first start-up was =
listed on the=20
New York Stock Exchange
until recently merging with another large =
company.=20
All indications are that
this endeavor will be even more successful =
than his=20
first. The company has
nearly 200 employees now and is preparing =
itself to go=20
public in the very
near future. The company by the way will go public =
making=20
a profit which is
virtually unheard of in an internet start-up.  =
This=20
envious financial
position will allow them to time their IPO for the =
best=20
market conditions.

This Information Technology Solutions company =
has=20
built one of the largest
and most sophisticated computer systems in =
the world=20
with over 30 TERAbytes
of marketing data and growing fast. Despite =
the=20
enormous size of the data
warehouse, queried information is returned =
to the=20
user virtually
instantaneously via the internet. This proprietary =
system=20
simply has to be
seen to be believed.

Companies integrating =
with our=20
client's access engine are able to improve
the effectiveness of their =

Customer Relationship Management (CRM) system
providing a competitive =

advantage in the market place.

Our client is not resting on its =
laurels=20
however.  A subsidiary of the
company is working on creating in=20
simplistic terms, a super computer.  To
that end, they are in =
need of=20
some brilliant programmers with the =
following
characteristics:

-=20
Recent graduate with Master's degree in Computer Science, or =
with
Bachelor's=20
degree and professional C++ programming experience.
- Solid knowledge =
of=20
application programming under Unix and VxWorks.
- Experience with =
parallel=20
programming, Sun/Solaris, Windows NT, and
knowledge of Graph Theory a =
BIG=20
plus.

Our client wants the best and is willing to pay for =
them. =20
Those meeting
their rigorous requirements will enjoy all the benefits =
of the=20
parent
company: excellent salaries (the majority of experienced =
candidates=20
I've
placed with the company have been over 100k), pre-IPO stock=20
options,
terrific benefits, and relocation assistance.  This =
isn't your=20
typical
start-up that is low on funds.  The company has all of =
the=20
benefits of a
start-up including exciting, non-bureaucratic leading =
edge=20
development but all of the comforts of a larger established company =

including financial stability, excellent bennies, state of the art =
equipment and=20
facility, and much more.  It could be the last job you'll ever =

need.

If you're interested in this position, please forward your =
resume=20
with
"VxWorks" in the subject line to resume@bkoenterprises.com =
. =20
Alternatively,
you may fax us at 800-842-7861.  For more =
information on=20
the position, you
may also call or e-mail:

Bridget=20
O'Connor
800-385-1914
bko@bkoenterprises.com

=
Thanks=20
again for your=20
time!










------=_NextPart_000_01C7_01BFFEF6.F501A980--



From vxwexplo-errs@csg.lbl.gov  Sun Aug  6 04:03:06 2000
From: Vxworks Exploder 
Date: Sun Aug  6 04:03:08 PDT 2000
Subject: comp.os.vxworks newsdigest

Comp.Os.Vxworks Daily Digest    Sun Aug  6 04:03:03 PDT 2000

        Subject: WindWeb Server & NFS file system
        Subject: Re: bc635PCI and Linux
        Subject: Boca Raton, FL: Need VxWorks Programmers - Profitable Pre-IPO company...
        Subject: VxWorks programmers needed in Boca Raton, FL
        Subject: Re: WindWeb Server & NFS file system
        Subject: Re: mv2700 and "initial stack" memory region.
        Subject: Re: fei82557End.c File
        Subject: Re: how does purify work?
        Subject: Re: how does purify work?
        Subject: Re: how does purify work?
        Subject: Re: how does purify work?
        Subject: Re: how does purify work?

-------------------------------------------------------

Newsgroups: comp.os.vxworks
Subject: WindWeb Server & NFS file system
Date: Sat, 05 Aug 2000 13:49:31 +0200
From: Bernhard Staudacher 
Message-ID: <398BFF4B.12211077@gmx.de>

Hello,

I integrated the WindWeb server (version 2.0) to our PPC-target. As our
flash-file system is not available yet, I tried to use a NFS-share on my
PC. The NFS-connection seems to work, because I can dump files from the
VxWorks target.
Unfortunately, the WindWeb server is not able to access files on the NFS
share.
My network monitor grabs some NFS-requests ("Look Up File"), that are
acknowledged positively (so the FS-RPM seems to be OK). But instead of
trying to get the file content by sending a NFS request "Read from File"
(proc #6), the web server sends a negative response (404 not found) to
the client.
Has anybody any idea, what's wrong in this scenario?

Thanks,
    Bernhard


---------------------------

Newsgroups: comp.os.vxworks,comp.realtime
Subject: Re: bc635PCI and Linux
Date: Sat, 05 Aug 2000 14:20:28 +0200
From: "Robert H. de Vries" 
Organization: Cistron Internet Services B.V.
Message-ID: <398C068C.C57D9497@cistron.nl>
References: <3981C16F.D33435D3@raf.atd.ucar.edu>


- --------------2D8212216BA01706786C0641
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Chris Webster wrote:

> Anyone have a driver/program for the Datum bc635PCI IRIG card running
> under Linux?
>

I wrote a driver for IRIX. The ftp site of datum has sample source code
for a driver for SunOS/Solaris
(ftp://ftp.datum.com/BusProducts/PCIbus/bc635PCI/Solaris).
You can get my driver as a starting point as well.

    Robert

- --
Robert de Vries
rhdv@rhdv.cistron.nl



- --------------2D8212216BA01706786C0641
Content-Type: text/html; charset=us-ascii
Content-Transfer-Encoding: 7bit



Chris Webster wrote:

Anyone have a driver/program for the Datum bc635PCI
IRIG card running

under Linux?

 

I wrote a driver for IRIX. The ftp site of datum has sample source code
for a driver for SunOS/Solaris (ftp://ftp.datum.com/BusProducts/PCIbus/bc635PCI/Solaris).

You can get my driver as a starting point as well.

    Robert

-- 
Robert de Vries
rhdv@rhdv.cistron.nl

 

- --------------2D8212216BA01706786C0641--


---------------------------

Newsgroups: comp.os.vxworks
Subject: Boca Raton, FL: Need VxWorks Programmers - Profitable Pre-IPO company...
Date: 05 Aug 2000 15:47:28 GMT
From: "Bridget OConnor" 
Organization: Concentric Internet Services
Message-ID: <8mhcug$e3u@dispatch.concentric.net>

Hello All,

Below is a description of a terrific opportunity with an exciting (and
already successful) Boca Raton company preparing to IPO soon.  If you're not
interested in a new position, please skip this post or pass it along to a
friend.  You'd be doing them a favor as it really is a great company with
phenomenal potential.  Thanks a bunch for your time!

C/C++ Programmers with VxWorks / Unix Experience

Our client's Chairman started this pre-IPO Internet Services company two
years ago. His first start-up was listed on the New York Stock Exchange
until recently merging with another large company. All indications are that
this endeavor will be even more successful than his first. The company has
nearly 200 employees now and is preparing itself to go public in the very
near future. The company by the way will go public making a profit which is
virtually unheard of in an internet start-up.  This envious financial
position will allow them to time their IPO for the best market conditions.

This Information Technology Solutions company has built one of the largest
and most sophisticated computer systems in the world with over 30 TERAbytes
of marketing data and growing fast. Despite the enormous size of the data
warehouse, queried information is returned to the user virtually
instantaneously via the internet. This proprietary system simply has to be
seen to be believed.

Companies integrating with our client's access engine are able to improve
the effectiveness of their Customer Relationship Management (CRM) system
providing a competitive advantage in the market place.

Our client is not resting on its laurels however.  A subsidiary of the
company is working on creating in simplistic terms, a super computer.  To
that end, they are in need of some brilliant programmers with the following
characteristics:

- - Recent graduate with Master's degree in Computer Science, or with
Bachelor's degree and professional C++ programming experience.
- - Solid knowledge of application programming under Unix and VxWorks.
- - Experience with parallel programming, Sun/Solaris, Windows NT, and
knowledge of Graph Theory a BIG plus.

Our client wants the best and is willing to pay for them.  Those meeting
their rigorous requirements will enjoy all the benefits of the parent
company: excellent salaries (the majority of experienced candidates I've
placed with the company have been over 100k), pre-IPO stock options,
terrific benefits, and relocation assistance.  This isn't your typical
start-up that is low on funds.  The company has all of the benefits of a
start-up including pre-IPO options but all of the comforts of a larger
established company including financial stability, excellent bennies, state
of the art equipment and facility,  and much more.  It could be the last job
you'll ever need.

If you're interested in this position, please forward your resume with
"VxWorks" in the subject line to resume@bkoenterprises.com .  Alternatively,
you may fax us at 800-842-7861.  For more information on the position, you
may also call or e-mail:

Bridget O'Connor
800-385-1914
bko@bkoenterprises.com

Thanks again for your time!










---------------------------

Newsgroups: comp.os.vxworks
Subject: VxWorks programmers needed in Boca Raton, FL
Date: Sat, 5 Aug 2000 16:05:21 -0400
From: "Bridget OConnor" 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <8mhsk3$o1j$1@overload.lbl.gov>

This is a multi-part message in MIME format.

- ------=_NextPart_000_01C7_01BFFEF6.F501A980
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

Hello All,

Below is a description of a terrific opportunity with an exciting (and
already successful) Boca Raton company preparing to IPO soon.  If you're =
not
interested in a new position, please skip this post or pass it along to =
a
friend.  You'd be doing them a favor as it really is a great company =
with
phenomenal potential.  Thanks a bunch for your time!

C/C++ Programmers with VxWorks / Unix Experience

Our client's Chairman started this pre-IPO Internet Services company two
years ago. His first start-up was listed on the New York Stock Exchange
until recently merging with another large company. All indications are =
that
this endeavor will be even more successful than his first. The company =
has
nearly 200 employees now and is preparing itself to go public in the =
very
near future. The company by the way will go public making a profit which =
is
virtually unheard of in an internet start-up.  This envious financial
position will allow them to time their IPO for the best market =
conditions.

This Information Technology Solutions company has built one of the =
largest
and most sophisticated computer systems in the world with over 30 =
TERAbytes
of marketing data and growing fast. Despite the enormous size of the =
data
warehouse, queried information is returned to the user virtually
instantaneously via the internet. This proprietary system simply has to =
be
seen to be believed.

Companies integrating with our client's access engine are able to =
improve
the effectiveness of their Customer Relationship Management (CRM) system
providing a competitive advantage in the market place.

Our client is not resting on its laurels however.  A subsidiary of the
company is working on creating in simplistic terms, a super computer.  =
To
that end, they are in need of some brilliant programmers with the =
following
characteristics:

- - Recent graduate with Master's degree in Computer Science, or with
Bachelor's degree and professional C++ programming experience.
- - Solid knowledge of application programming under Unix and VxWorks.
- - Experience with parallel programming, Sun/Solaris, Windows NT, and
knowledge of Graph Theory a BIG plus.

Our client wants the best and is willing to pay for them.  Those meeting
their rigorous requirements will enjoy all the benefits of the parent
company: excellent salaries (the majority of experienced candidates I've
placed with the company have been over 100k), pre-IPO stock options,
terrific benefits, and relocation assistance.  This isn't your typical
start-up that is low on funds.  The company has all of the benefits of a
start-up including exciting, non-bureaucratic leading edge development =
but all of the comforts of a larger established company including =
financial stability, excellent bennies, state of the art equipment and =
facility, and much more.  It could be the last job you'll ever need.

If you're interested in this position, please forward your resume with
"VxWorks" in the subject line to resume@bkoenterprises.com .  =
Alternatively,
you may fax us at 800-842-7861.  For more information on the position, =
you
may also call or e-mail:

Bridget O'Connor
800-385-1914
bko@bkoenterprises.com

Thanks again for your time!










- ------=_NextPart_000_01C7_01BFFEF6.F501A980
Content-Type: text/html;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable









Hello All,

Below is a =
description of a=20
terrific opportunity with an exciting (and
already successful) Boca =
Raton=20
company preparing to IPO soon.  If you're not
interested in a =
new=20
position, please skip this post or pass it along to a
friend.  =
You'd be=20
doing them a favor as it really is a great company with
phenomenal=20
potential.  Thanks a bunch for your time!

C/C++ Programmers =
with=20
VxWorks / Unix Experience

Our client's Chairman started this =
pre-IPO=20
Internet Services company two
years ago. His first start-up was =
listed on the=20
New York Stock Exchange
until recently merging with another large =
company.=20
All indications are that
this endeavor will be even more successful =
than his=20
first. The company has
nearly 200 employees now and is preparing =
itself to go=20
public in the very
near future. The company by the way will go public =
making=20
a profit which is
virtually unheard of in an internet start-up.  =
This=20
envious financial
position will allow them to time their IPO for the =
best=20
market conditions.

This Information Technology Solutions company =
has=20
built one of the largest
and most sophisticated computer systems in =
the world=20
with over 30 TERAbytes
of marketing data and growing fast. Despite =
the=20
enormous size of the data
warehouse, queried information is returned =
to the=20
user virtually
instantaneously via the internet. This proprietary =
system=20
simply has to be
seen to be believed.

Companies integrating =
with our=20
client's access engine are able to improve
the effectiveness of their =

Customer Relationship Management (CRM) system
providing a competitive =

advantage in the market place.

Our client is not resting on its =
laurels=20
however.  A subsidiary of the
company is working on creating in=20
simplistic terms, a super computer.  To
that end, they are in =
need of=20
some brilliant programmers with the =
following
characteristics:

-=20
Recent graduate with Master's degree in Computer Science, or =
with
Bachelor's=20
degree and professional C++ programming experience.
- Solid knowledge =
of=20
application programming under Unix and VxWorks.
- Experience with =
parallel=20
programming, Sun/Solaris, Windows NT, and
knowledge of Graph Theory a =
BIG=20
plus.

Our client wants the best and is willing to pay for =
them. =20
Those meeting
their rigorous requirements will enjoy all the benefits =
of the=20
parent
company: excellent salaries (the majority of experienced =
candidates=20
I've
placed with the company have been over 100k), pre-IPO stock=20
options,
terrific benefits, and relocation assistance.  This =
isn't your=20
typical
start-up that is low on funds.  The company has all of =
the=20
benefits of a
start-up including exciting, non-bureaucratic leading =
edge=20
development but all of the comforts of a larger established company =

including financial stability, excellent bennies, state of the art =
equipment and=20
facility, and much more.  It could be the last job you'll ever =

need.

If you're interested in this position, please forward your =
resume=20
with
"VxWorks" in the subject line to resume@bkoenterprises.com =
. =20
Alternatively,
you may fax us at 800-842-7861.  For more =
information on=20
the position, you
may also call or e-mail:

Bridget=20
O'Connor
800-385-1914
bko@bkoenterprises.com

=
Thanks=20
again for your=20
time!










- ------=_NextPart_000_01C7_01BFFEF6.F501A980--


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: WindWeb Server & NFS file system
Date: Sat, 05 Aug 2000 22:11:03 GMT
From: john_94501@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8mi3dm$5cq$1@nnrp1.deja.com>
References: <398BFF4B.12211077@gmx.de>

Hello,

Have you "mounted" the NFS file system into the web server's file
systems?

    httpFsConfInit (MAX_FILE_MOUNTPOINTS);
    httpRpmConfAdd (HTTP_M_GET, "/", httpRpmFs);
    httpFsConfAdd  (            "/", "/nfs");

I do this as the last thing in my HTTP setup sequence (httpconf.c) so
that anything not caught by earlier RPMS will be passed to the
filesystem. Since my target has no local FS, I use an NFS file system
mounted on /nfs.

Another thing that could be catching you is having one of the earlier
RPMs match your URL and therefore mask the file system. I had this for a
while, and it took me a long time to work out what was going on (it's
amazing how the simple mistakes can take so long to find!).

HTH,

John...

In article <398BFF4B.12211077@gmx.de>,
  Bernhard Staudacher  wrote:
> Hello,
>
> I integrated the WindWeb server (version 2.0) to our PPC-target. As
our
> flash-file system is not available yet, I tried to use a NFS-share on
my
> PC. The NFS-connection seems to work, because I can dump files from
the
> VxWorks target.
> Unfortunately, the WindWeb server is not able to access files on the
NFS
> share.
> My network monitor grabs some NFS-requests ("Look Up File"), that are
> acknowledged positively (so the FS-RPM seems to be OK). But instead of
> trying to get the file content by sending a NFS request "Read from
File"
> (proc #6), the web server sends a negative response (404 not found) to
> the client.
> Has anybody any idea, what's wrong in this scenario?
>
> Thanks,
>     Bernhard
>
>


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: mv2700 and "initial stack" memory region.
Date: Sat, 05 Aug 2000 22:16:36 GMT
From: john_94501@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8mi3o3$5m6$1@nnrp1.deja.com>
References: <398B2C23.33F473A9@mindspring.com>

Hello,

In article <398B2C23.33F473A9@mindspring.com>,
  Al Johnston  wrote:
> The memory map for this bsp labels the area from 0xfc00 to
> "bsp dependent value" (which is 0x10,0000 for the bsp) as
> "initial stack".  My understanding is that this area is used
> as a stack during the startup of vxworks.  Once the root task
> has started, this area's use as a stack is ended.
>
> Is that correct?

Partly...

> Next, during startup, part of this "initial stack" area is
> used by various parts of vxw's "shared memory" thingie.  For
> this bsp (with default settings for SM address/sizes and with
> bp and vxMP INCLUDE's included) part of this "initial stack"
> area is put back into service.  The back in service parts are
> from address 0x1,0000 to 0x3,ffff.
>
> This leaves the areas 0x4c00 -> 0xffff

I have a feeling that some of the network driver's buffer memory comes
from here... it did on some of the earlier boards. Check in the BSP's
config files for addresses in this area!

HTH,

John...


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: fei82557End.c File
Date: Sat, 05 Aug 2000 22:19:23 GMT
From: john_94501@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8mi3ta$5mu$1@nnrp1.deja.com>
References: <398AFBD4.95CF6CBA@lanterncom.com>

Hi,

In article <398AFBD4.95CF6CBA@lanterncom.com>,
  asher  wrote:
> Hi,
>
> I'm looking for an up-to-date (working) fei82557End.c and
> sysFei82557End.c files for a PPC750 BSP. Can anybody please send it to
> me?

I hope not - it is proprietary source and to send it would be a clear
violation of copyright & the source license agreement it comes with.

If you want this source then contact your WR sales rep., but you're
going to have to buy it (sorry).

Regards,

John...


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks,comp.compilers,comp.programming
Subject: Re: how does purify work?
Date: 5 Aug 2000 20:39:49 -0400
From: Ben Pfaff 
Organization: Michigan State University
Keywords: tools, storage
Message-ID: <00-08-029@comp.compilers>
References: <00-08-027@comp.compilers>

toddhoff@my-deja.com writes:

> If one wanted to build a Purify like tool, does anyone know how it
> would be done? I don't really understand what it does and haven't
> found much related info.

Have a look at GNU Checker, available from
	ftp://alpha.gnu.org/gnu
or *much* newer snapshots via CVS at cvs.gnu.org.  It doesn't
work in the same way as Purify, as I understand it, but it does
the same job.  Older versions of Checker worked by using a
modified version of the GNU assembler that inserted calls to
check every memory reference; newer versions of Checker work by
using a modified version of the GNU C compiler that does the same
thing.

(FWIW, I am co-maintainer of Checker, but not the author.)

---------------------------

Newsgroups: comp.os.vxworks,comp.compilers,comp.programming
Subject: Re: how does purify work?
Date: 5 Aug 2000 21:31:26 -0400
From: Chris F Clark 
Organization: The World Public Access UNIX, Brookline, MA
Keywords: debug
Message-ID: <00-08-033@comp.compilers>
References: <00-08-027@comp.compilers>

Someone asked:
> Once upon a time Purify did work for vxworks.  It would be a very nice
> tool to have and it seems very odd that there's no product with
> equivalent functionality.
>
> If one wanted to build a Purify like tool, does anyone know how it
> would be done? I don't really understand what it does and haven't
> found much related info.
>
> thanx
> [Doesn't it stick extra code into your program to audit all of the
> allocation, freeing, and pointer references?  Dunno if it hacks the
> source or rewrites the object. -John]

There are (at least) 5 ways to implement Purify like tools.

1) Purify (and Third Degree--the tool I maintained for Compaq for a
   while) work by taking each object file (either the .obj or the
   .exe) apart sticking in extra instructions that monitor when the
   program does each "interesting" thing (allocate, deallocate,
   reference through a pointer, initialize a variable, whatever).

   The extra instructions then store bits of information away in the
   tool's database (essentially a big array of bits denoting the
   status of each memory location as allocated and/or initialized) and
   check that database to report problems.

2) Insight++ does roughly the same thing but takes the source code
   apart and inserts source code to do the same thing.

3) Centerline does roughly the same thing, but implements it by
   interpreting the source code.

4) It can also be done by interpreting the instruction stream.  Sun's
   tool Shade does something like this (though not to implement
   Purify functionality).

5) It should also be possible to do it via a "pointer swizzling"
   technique, where only the libraries are replaced.  This would be
   like the way ObjectStore implements object-oriented databases.  In
   this case, only the run-time library routines like malloc and free
   are modified.  Instead of returning real pointers, they return
   "fake" values that are guaranteed to cause "segmentation faults" or
   "access violations" (or some other trappable signal).  The tool
   also inserts its own signal handlers (possibly by modifying the
   signal and program startup runtime routines).  The signal handlers
   wait for the trappable signals and then use them to detect the
   memory accesses to be watched and use that to perform the
   annotation and also substitute the correct action.

My guess is that the above are ordered approximately by how fast they
would execute--technique 1 being fastest and either 4 or 5 being
slowest.  (The speed of 5 would depend on how often the faults
occured.  If seldom, it might be fast.)

Either technique 2 or 3 is probably the simplest to implement if you
have only 1 source language to deal with.

Note, if you use technique 1, you have to be careful as the Purify
folks staked out the territory pretty well with patents which they
have pursued.

Hope this helps,
- -Chris

*****************************************************************************
Chris Clark                    Internet   :  compres@world.std.com
Compiler Resources, Inc.       Web Site   :  http://world.std.com/~compres
3 Proctor Street               voice      :  (508) 435-5016
Hopkinton, MA  01748  USA      fax        :  (508) 435-4847  (24 hours)
- ------------------------------------------------------------------------------

---------------------------

Newsgroups: comp.os.vxworks,comp.compilers,comp.programming
Subject: Re: how does purify work?
Date: 5 Aug 2000 21:31:47 -0400
From: plakal@cs.wisc.edu
Organization: University of WI, Madison -- Computer Sciences Dept.
Keywords: debug
Message-ID: <00-08-034@comp.compilers>
References: <00-08-027@comp.compilers>

In comp.compilers toddhoff@my-deja.com wrote:
> If one wanted to build a Purify like tool, does anyone know how it
> would be done? I don't really understand what it does and haven't
> found much related info.
>
> [Doesn't it stick extra code into your program to audit all of the
> allocation, freeing, and pointer references?  Dunno if it hacks the
> source or rewrites the object. -John]

	There's a USENIX'92 paper that describes Purify:
        Reed Hastings and Bob Joyce. Purify: Fast Detection of
        Memory Leaks and Access Errors. Winter USENIX Conference '92.

        You might also want to look at a paper by Todd Austin & co
        in PLDI'94 which describes Purify and other run-time memory
        checkers in its related work section.

	Manoj

---------------------------

Newsgroups: comp.os.vxworks,comp.compilers,comp.programming
Subject: Re: how does purify work?
Date: 5 Aug 2000 21:35:24 -0400
From: "Martin Raabe" 
Organization: T-Online
Keywords: debug
Message-ID: <00-08-037@comp.compilers>
References: <00-08-027@comp.compilers>

Maybe you want to take a look at  the RTA Suite of the DiabData Compiler of
WindRiver.
Ciao
- --
Martin Raabe

 schrieb in im Newsbeitrag:
> If one wanted to build a Purify like tool, does anyone know how it
> would be done? I don't really understand what it does and haven't
> found much related info.

---------------------------

Newsgroups: comp.os.vxworks,comp.compilers,comp.programming
Subject: Re: how does purify work?
Date: 5 Aug 2000 21:35:51 -0400
From: nsrcccw@leonis.nus.edu.sg (Chea Chee Weng)
Organization: National University of Singapore
Keywords: debug
Message-ID: <00-08-038@comp.compilers>
References: <00-08-027@comp.compilers>
Followup-To: comp.os.vxworks,comp.compilers,comp.programming

toddhoff@my-deja.com wrote:
: Once upon a time Purify did work for vxworks.  It would be a very nice
: tool to have and it seems very odd that there's no product with
: equivalent functionality.

: If one wanted to build a Purify like tool, does anyone know how it
: would be done? I don't really understand what it does and haven't
: found much related info.

You can find this using a patent search - use either the one at ibm or
the www.uspto.gov. The "purify" method was patented but at this point
i am unable to remember the "keywords" to find it. Anyway you may find
a bunch of ways people did this either search "parasoft" for their
insure++ tools. (for memory overrun and others). or search for
"Detecting concurrency errors in multi-threaded programs" 6,009,269 to
see how one method to find race condition detecting is made.

Roughly, all will augment/instrument the source or object code to do
this (otherwise some will have a modified VM like the JavaVM concept
to monitor violation).

For instrumenting code (or injecting more functions) to an existing
binary (without the source code), you could see one Microsoft tool
with source (Gee can't remember now but it is in microsoft research
and free to download or try "DynInst" that works for non-Intel
machines (See for it - is was part of Paradyn tools)).

With all these tools and info, it does not seem hard to make your own
but i have a hunch that there may be "special" condition cases in all
the above methods (patents) that one has to hack to get things working

---------------------------

End of New-News digest
**********************


From vxwexplo-errs@csg.lbl.gov  Mon Aug  7 04:03:06 2000
From: Vxworks Exploder 
Date: Mon Aug  7 04:03:08 PDT 2000
Subject: comp.os.vxworks newsdigest

Comp.Os.Vxworks Daily Digest    Mon Aug  7 04:03:03 PDT 2000

        Subject: Re: Frequency of TimeStampCounter (TSC)?
        Subject: Re: taskInit - why doesn't it work? VxWorks
        Subject: VxWorks for StrongArm
        Subject: “printf” in romStart()/bootInit.c?
        Subject: Re: VxWorks for StrongArm
        Subject: Re: "printf" in romStart()/bootInit.c?
        Subject: Re: “printf” in romStart()/bootInit.c?
        Subject: Re: VxWorks FTP boot problem (T2, MVME2700)
        Subject: Re: WindWeb Server & NFS file system
        Subject: Re: Problems w/ select() and timeout on mvme2700
        Subject: How can i enable target server?
        Subject: How can i enable target server?

-------------------------------------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Frequency of TimeStampCounter (TSC)?
Date: Sun, 06 Aug 2000 11:26:14 GMT
From: dixon@crd.ge.com
Organization: Deja.com - Before you buy.
Message-ID: <8mji0k$324$1@nnrp1.deja.com>
References: <3988300E.4DE80833@meco.nl>

In article <3988300E.4DE80833@meco.nl>,
  Albert.Wijnja@meco.nl wrote:
> Hello,
>
> In my application I use the pentiumTscGet64() routine to get the
current
> counter value of the TSC. Is there anyway to determine the frequency
of
> this counter? I use this routine for timestamping some events in my
> application. Eventually I want to represent the events in
> (micro)-seconds.
>
> Thanks
>
> Albert Wijnja
>
>
Read counter(t0)
sleep(1000ms)
Read counter(t1)
t1-t0 are approximate # cycles per second
Of course a 1sec sleep isn't necessarily 1sec.
You could also read the rtc
disable interrupts
read cycle counter
read rtc
enable interrupts
sleep
disable interrupts
read cycle counter
read rtc
enable interrupts
the difference in rtc readings give the elapsed
time with the accuracy of the rtc.

HTH. Std disclaimers apply


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: taskInit - why doesn't it work? VxWorks
Date: Sun, 06 Aug 2000 18:37:00 GMT
From: pwhicker@home.com (Paul Whicker)
Organization: @Home Network
Message-ID: 
References: <8mccr4$6p2$1@overload.lbl.gov>

Just some silly questions. 

Did you check the return values from malloc()?

Does the TCB need to be aligned? How about memPartAlignedAlloc (I think), same 
with the stack I suppose..

Paul

In article <8mccr4$6p2$1@overload.lbl.gov>, "CICON,JAMES (HP-Vancouver,ex1)" 
 wrote:
>VxWorks Tornado 
>
>I need to use taskInit instead of taskSpawn.  When I run the code snippet below
>I always get the following memPartFree error message:
>   pStackBase: 0x20F6A304, pTcb: 0x20FEF6B8
>   0x20fec9d4 (tExcTask): memPartFree: invalid block 0x20f6f30c in partition
>0x200777c0.
>   test task has run
>
>>>>>>>>>>>>>>><<<<<<<<<<<<<<<<
>>>>>> start code snippet <<<<<
>>>>>>>>>>>>>>><<<<<<<<<<<<<<<<
>void testTask(void)
>{
>        printf("test task has run\n");
>}
>
>extern "C" void runMe(void)
>{       
>        int stackSize = 20*1024;
>        char * pStackBase = (char *) malloc(stackSize);
>        WIND_TCB * pTcb = (WIND_TCB *) malloc(sizeof(WIND_TCB));
>
>        printf("pStackBase: 0x%08X, pTcb: 0x%08X\n", (UINT)pStackBase,
>(UINT)pTcb);
>
>        taskInit(pTcb, "testTask", 150, 0, pStackBase, stackSize, 
>              (FUNCPTR)testTask, 0, 0,0,0,0,0,0,0,0,0);
>
>        taskActivate((int)pTcb);
>}
>>>>>>>>>>>>>>><<<<<<<<<<<<<<
>>>>>> end code snippet <<<<<
>>>>>>>>>>>>>>><<<<<<<<<<<<<<

---------------------------

Newsgroups: comp.os.vxworks
Subject: VxWorks for StrongArm
Date: Sun, 06 Aug 2000 21:42:53 GMT
From: yarpro@onramp.net
Organization: Verio
Message-ID: <398ddb82.332021100@news.verio.net>

Does anybody know if VxWorks for the StrongArm (SA-110) is targeted
for big or little endian?

TIA,

marcus yarbrough
yarpro@onramp.net

---------------------------

Newsgroups: comp.os.vxworks
Subject: “printf” in romStart()/bootInit.c?
Date: Sun, 06 Aug 2000 22:58:10 GMT
From: toomuchham@yahoo.com
Organization: Deja.com - Before you buy.
Message-ID: <8mkqi1$tdt$1@nnrp1.deja.com>

     Hi all!
     Was wondering if there’s a way to write to the serial port in
romStart() in file bootInit.c before the ROM copies itself into RAM?
     In one of the Power-On Self Test routines I’m creating, I want to
test the integrity of the RAM datapath before the VxWorks boot ROM
dumps itself into RAM and continues execution.  I’d like to write a
message to the serial port reporting the test’s success or failure.
Unfortunately, my target board doesn’t have any LEDs (yet) I can
trigger to signal the test’s result.
     I can’t call sysSerialHwInit1/2 in sysSerial.c since these
functions allocate buffers and create interrupt tables in RAM (which
would fail if the RAM datapath was corrupt).
     My target processor is a PPC 850.
     Thanks in advance!
     David Downey
     Software Engineering Consultant


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: VxWorks for StrongArm
Date: Sun, 06 Aug 2000 23:06:26 GMT
From: john_94501@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8mkr1g$tqm$1@nnrp1.deja.com>
References: <398ddb82.332021100@news.verio.net>

Hi,

In article <398ddb82.332021100@news.verio.net>,
  yarpro@onramp.net wrote:
> Does anybody know if VxWorks for the StrongArm (SA-110) is targeted
> for big or little endian?

Little endian.

John...


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: "printf" in romStart()/bootInit.c?
Date: Mon, 7 Aug 2000 07:41:15 +0530
From: "G.N.Raj" 
Organization: Wipro
Message-ID: 
References: <8mkqi1$tdt$1@nnrp1.deja.com>
Reply-To: "G.N.Raj" 

I have worked on ev386ex BSP. I also had similar problems. I wrote serial
initialization and printing routines myself with the drivers in
target/src/drv/sio/*.c as the reference. You could do the same.

Raj

 wrote in message news:8mkqi1$tdt$1@nnrp1.deja.com...
>      Hi all!
>      Was wondering if there's a way to write to the serial port in
> romStart() in file bootInit.c before the ROM copies itself into RAM?
>      In one of the Power-On Self Test routines I'm creating, I want to
> test the integrity of the RAM datapath before the VxWorks boot ROM
> dumps itself into RAM and continues execution.  I'd like to write a
> message to the serial port reporting the test's success or failure.
> Unfortunately, my target board doesn't have any LEDs (yet) I can
> trigger to signal the test's result.
>      I can't call sysSerialHwInit1/2 in sysSerial.c since these
> functions allocate buffers and create interrupt tables in RAM (which
> would fail if the RAM datapath was corrupt).
>      My target processor is a PPC 850.
>      Thanks in advance!
>      David Downey
>      Software Engineering Consultant
>
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.





---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: “printf” in romStart()/bootInit.c?
Date: Mon, 07 Aug 2000 02:01:49 GMT
From: pkockritz@home.com (Pete Kockritz)
Organization: @Home Network
Message-ID: 
References: <8mkqi1$tdt$1@nnrp1.deja.com>

In article <8mkqi1$tdt$1@nnrp1.deja.com>, toomuchham@yahoo.com wrote:

>      Hi all!
>      Was wondering if there’s a way to write to the serial port in
> romStart() in file bootInit.c before the ROM copies itself into RAM?
>      In one of the Power-On Self Test routines I’m creating, I want to
> test the integrity of the RAM datapath before the VxWorks boot ROM
> dumps itself into RAM and continues execution.  I’d like to write a
> message to the serial port reporting the test’s success or failure.
> Unfortunately, my target board doesn’t have any LEDs (yet) I can
> trigger to signal the test’s result.
>      I can’t call sysSerialHwInit1/2 in sysSerial.c since these
> functions allocate buffers and create interrupt tables in RAM (which
> would fail if the RAM datapath was corrupt).
>      My target processor is a PPC 850.
>      Thanks in advance!
>      David Downey
>      Software Engineering Consultant


If all you want to do is write to the serial port (no reading) you can
setup the CPM to do this fairly simply. I did that for pSOS, I started with
some example code from the Motorola web site. I set up SMC2 for
polling to see when there was room for more bytes in the output buffer,
I believe I just wrote one byte at a time. You could do it more efficiently,
but I just took the simple path to allow error messages to be written out
during boot.

I didn't implement 'printf', just 'puts'. That way I didn't have to have
the standard I/O library to do the formating as part of the boot image.

Regards,
Pete
- --
+-----------------------------------------------------------------+
|       Pete Kockritz                  mailto:pkockritz@home.com  |
|                       This space for rent.                      |
+-----------------------------------------------------------------+

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: VxWorks FTP boot problem (T2, MVME2700)
Date: Mon, 07 Aug 2000 08:30:06 GMT
From: paulvroberts@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8mls2f$keu$1@nnrp1.deja.com>
References: <8mdcif$tgl$1@nnrp1.deja.com> <398A95D7.A8BBCE60@auriga.ru>

Hi there,

Thanks very much for your reply. You are definitely right about the
wait - I waited a lot longer and VxWorks did eventually boot. I can
certainly downgrade our ftp server in the short-term. However, I am
also concerned for the long-term. Both ftp servers give out the same
messages.

I had a look at the code in bootconfig. I could not see an awful lot
that I could fiddle with. I found the comment "Empty the Data Socket
before close. PC FTP server hangs otherwise" suspicious, so I removed
the following line:

        while ((read (fd, command, sizeof (command))) > 0);

But this made the system not able to boot from anything.
Hmmm....that makes me even more suspicious but not at all sure what to
try next.

Paul Roberts.

In article <398A95D7.A8BBCE60@auriga.ru>,
  "Arthur V. Grigorian"  wrote:
> Hello,
>
>    We had  the same problem with that  FTP server and  the solution
was
>    to downgrade it.
>
>    Actually your target doesn't hang during  the FTP transfer - it
just
>    doesn't think it logged out successfully, so if you wait 900
seconds
>    (default FTP timeout) it will run the loaded image.
>
>    AFAIR the problem is in some unexpected messages from the FTP
server
>    like:
> 	221-You have transferred X bytes in X files.
> 	221-Total traffic for this session was X bytes in X transfers.
> 	221-Thank you for using the FTP service on ftp.server.name.
> 	221 Goodbye.
>    instead of just:
> 	221 Goodbye.
>
>    You can try to hack it yourself - see
target/config/all/bootCongig.c
>    file, netLoad() function, around ftpCommand(errFd, "QUIT",...)
call.
>
> Hope that helps,
> __
> R2
>
> > Hi,
> >
> > We are having trouble with ftp boot using the VxWorks boot rom
loader,
> > and I'm wondering if anyone else has seen this before or has any
> > suggestions or troubleshooting hints.
> >
> > The problem occurs when the ftp server used is the one on Red Hat
Linux
> > 6.2 [uu ftpd Version wu-2.6.0(1) Mon Feb 28
> > 10:30:36 EST 2000)]. It doesn't seem to happen with Red Hat Linux
6.0
> > [uu ftp Version wu-2.4.2-VR17(1) Mon Apr 19 09:21:53 EDT 1999] or
an old
> > Windows ftp server (WFTPD from 1995).
> >
> > What happens is that the VxWorks boot loader appears to load
> > the image ok, but doesn't continue the boot process - what we see
is it
> > saying "loading....", followed by load numbers increasing, then a
new
> > line, and then it just sits there.
> >
> > A "netstat" on the ftp server end revealed that the ftp data
connections
> > are in the FIN_WAIT2 state - from what I know of TCP (only
pieces!), I
> > think this means that the ftp server has closed its end, and is
waiting
> > for VXWorks to finish the connection.
> >
> > I suspect that for some reason, the VxWorks ftp boot loader is not
> > handling the case where the ftp server closes the connection before
it
> > does (TCP half close?).
> >
> > We are using Tornado 2, VxWorks 5.4 and the MVME 2700 BSP. It
doesn't
> > make any difference if I am booting a VxWorks image
> > with or without a downloadable symbol table.
> >
> > Now it might be ok to live with this at the moment (only use Red Hat
> > 6.0), but I am concerned that this could be a problem in future
years
> > when that restriction is no longer appropriate.
> >
> > Any ideas? Thanks for any help,
> >
> > Paul Roberts
>


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: WindWeb Server & NFS file system
Date: Mon, 7 Aug 2000 10:54:40 +0200
From: "Bernhard Staudacher" 
Organization: Siemens Inc.
Message-ID: <8mlth6$drq$1@news.mch.sbs.de>
References: <398BFF4B.12211077@gmx.de> <8mi3dm$5cq$1@nnrp1.deja.com>

Hello John,

I think the NFS file system is "mounted" correctly. There are some
NFS-packets ("Look up file") transmitted via the network. The problem is,
that after the file is found (i.e. the "Look up File" request returns
successfully), the Web server does not try to read its content but sends a
negative response to the HTML-browser.
I tried it with different NFS-servers. First I used a freeware NFS-server
(TrueGrid) for WIN-NT. Since it did not work I also mounted a LINUX-share.
In both cases the result was the same.
Are there any incompatibilities between the vxworks-based NFS-client and a
LINUX-server?

Regards, Bernhard



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Problems w/ select() and timeout on mvme2700
Date: Mon, 07 Aug 2000 08:53:25 GMT
From: paulvroberts@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8mlte3$l8m$1@nnrp1.deja.com>
References: <8m9d1u$h91$1@newsfeed.cv.nrao.edu> <8mdbhf$sns$1@nnrp1.deja.com>

Jim,

We use select in our 2700 based product for sockets and it works fine
for us. There's a local copy made of the timeout, but I'm not sure what
else is different - there's lots of missing "const"s in the VxWorks API
even in new code such as dosFs2.

Anyway, here's an extract of our code, which you can hopefully use:

	fd_set         rset        ;
	int            selectResult;
	struct timeval tv          ;

	FD_ZERO(&rset);
	FD_SET(m_fd, &rset);

	// make a copy here because we pass a non const pointer to the
timout
	// value when we call select... ANYTHING could happen
	tv = m_rcvTimeout;

	selectResult = ::select(m_fd+1, &rset, NULL, NULL, &tv);

	TRACE(TR_IGNORE, "Select returned %d, %d\n", selectResult,
FD_ISSET(m_fd, &rset));

	if((selectResult == 1) && (FD_ISSET(m_fd, &rset) != 0))
	{
		return(OK);
	}

	return(ERROR);

Hope this helps,

Paul.

On Fri, 4 Aug 2000 08:49:19    Jim Pisano wrote:
>Thanks for your reply.
>
>You are correct about the 10ms issue, I later changed it to 100ms. I
also played around w/
>the value making it many seconds (by setting timeout.secs = 10) w/ no
improvement.
>
>Thanks.
>-Jim

In article <8mdbhf$sns$1@nnrp1.deja.com>,
  paulvroberts@my-deja.com wrote:
> Hi Jim,
>
> This is a bit of a guess, but here goes.
>
> From the code you have below, it looks like the select timeout is set
to
> 10000nS, ie 10 mS, which is less than the granularity of the system
> clock (1/60 sec by default on MVME2700). Perhaps Vxworks is seeing
this
> value as "0", and thereby returning straight away?
>
> Paul Roberts.
>
> In article <8m9d1u$h91$1@newsfeed.cv.nrao.edu>,
>   "Jim Pisano"  wrote:
> > I'm having a problem with select() on a TCP socket and timeouts.
> > When I remove the timeout (i.e., set the pointer to the timeval
struct
> to be
> > 0)everything works correctly, but when I insert a non-null timeout
> into
> > select(), it always returns 0, i.e., a timeout occurred w/out any
file
> > descriptors ready.
> >
> > I have set the timeout value to many seconds to ensure that the
> select()
> > function waits long enough for data from the client app.
> >
> > I'm running VxWorks 5.4 on Tornado II hosted on Win NT. The target
is
> an
> > mvme-2700.
> >
> > here's a snippet of code:
> >
> > // create & open server socket (all details left out except for
> "accept()"
> > call...
> >     m_ACC_ClientSocketFD = accept (m_listenSocket, (struct sockaddr
*)
> > &clientAddr, &sockAddrSize);
> >
> > // set up file descriptor set structure for select() call
> >     int m_width;
> >     fd_set m_fdSet;
> >     int m_ACC_ClientSocketFD;
> >
> >     FD_ZERO( &m_fdSet );
> >     FD_SET( m_ACC_ClientSocketFD, &m_fdSet );
> >     m_width = m_ACC_ClientSocketFD + 1;
> >
> > // set up time out for 10 ms
> > struct timeval timeout = {0,10000};
> >
> >  int selectVal = select( m_width, &m_fdSet, NULL, NULL, &timeout );
> >  if( selectVal > 0 )
> >  {
> >      bzero(buffer,100);
> >      if( read( m_ACC_ClientSocketFD, buffer, 100) > 0 )
> >      {
> >        // process buffer ...
> >      }
> >  }
> >
> > Thanks.
> > -Jim
> >


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: How can i enable target server?
Date: Mon, 07 Aug 2000 09:27:56 GMT
From: "KBS" 
Organization: Dacom
Message-ID: 

I'm using VxWork 5.4 for Windows NT version. But the target server is only
30 day version and now expired.
How can i enable the server at NT server? At the starting target server, it
ask the "site key" value for given "site code". How can i get it? Somebody
help me!

Thanks for your help.
Bongsoo Kang



---------------------------

Newsgroups: comp.os.vxworks
Subject: How can i enable target server?
Date: Mon, 07 Aug 2000 09:27:56 GMT
From: "KBS" 
Organization: Dacom
Message-ID: 

I'm using VxWork 5.4 for Windows NT version. But the target server is only
30 day version and now expired.
How can i enable the server at NT server? At the starting target server, it
ask the "site key" value for given "site code". How can i get it? Somebody
help me!

Thanks for your help.
Bongsoo Kang





---------------------------

End of New-News digest
**********************


From vxwexplo-errs@csg.lbl.gov  Tue Aug  8 00:35:49 2000
From: "R.Chandrasekar" 
Date: Tue Aug  8 00:35:51 PDT 2000
Subject: Re: Vx Exploder Digest

Hi KBS,

You need to get the site Key from WRS or your local representative. This
again depends on the number of installations licensed along with your
tornado software.

You need to send the site code and IP address of your machine. Then you will
get a site key which will make you have unlimited license for that project
development.

Alternately you can transfer the site key from existing licensed machine to
the new one. Your old machine will then have  no more license

Hope u r clear.

Bye
Chandrasekar.R
Engineer R &D
CMC Limited
Real Time Systems
Hyderabad, India

>Newsgroups: comp.os.vxworks
>Subject: How can i enable target server?
>Date: Mon, 07 Aug 2000 09:27:56 GMT
>From: "KBS" 
>Organization: Dacom
>Message-ID: 
>
>I'm using VxWork 5.4 for Windows NT version. But the target server is only
>30 day version and now expired.
>How can i enable the server at NT server? At the starting target server, it
>ask the "site key" value for given "site code". How can i get it? Somebody
>help me!
>
>Thanks for your help.
>Bongsoo Kang




From vxwexplo-errs@csg.lbl.gov  Tue Aug  8 04:03:07 2000
From: Vxworks Exploder 
Date: Tue Aug  8 04:03:09 PDT 2000
Subject: comp.os.vxworks newsdigest

Comp.Os.Vxworks Daily Digest    Tue Aug  8 04:03:04 PDT 2000

        Subject: Re: How can i enable target server?
        Subject: running out of buffers
        Subject: Re: running out of buffers
        Subject: Floating Point Problems
        Subject: Replacing global new and delete
        Subject: Re: raw keyboard input to Tornado shell process?
        Subject: Re: why stdin only sometimes has a `vi` interface
        Subject: Re: Question as re: the FAQ as re: Job Postings
        Subject: Re: gcc-2.95.2-2, VxWorks, CygWin: Can't build it.
        Subject: help: porting chaivm to vxworks/ppc
        Subject: Re: why stdin only sometimes has a `vi` interface
        Subject: Enet Boot using MUX Driver
        Subject: IP layer API?
        Subject: dec21x4xEnd driver
        Subject: VxWorks Network doesnt work on PC target platform
        Subject: Ctrl+D synopsis for my own code?
        Subject: Re: Ctrl+D synopsis for my own code?
        Subject: Re: VxWorks Network doesnt work on PC target platform
        Subject: Writing Device Drivers
        Subject: re: Vx Exploder Digest
        Subject: Tornado Launch

-------------------------------------------------------

Newsgroups: comp.os.vxworks
Subject: Re: How can i enable target server?
Date: Mon, 07 Aug 2000 06:52:22 -0700
From: Joe Durusau 
Organization: Lockheed Martin Corporation
Message-ID: <398EBF16.A9C3C8A9@delphi.com>
References: 

	Assuming that you have paid your bill, contact WRS support.  The key
comes
after you have paid them, usually by email.

Speaking only for myself,

Joe Durusau


KBS wrote:
> 
> I'm using VxWork 5.4 for Windows NT version. But the target server is only
> 30 day version and now expired.
> How can i enable the server at NT server? At the starting target server, it
> ask the "site key" value for given "site code". How can i get it? Somebody
> help me!
> 
> Thanks for your help.
> Bongsoo Kang

---------------------------

Newsgroups: comp.os.vxworks
Subject: running out of buffers
Date: Mon, 7 Aug 2000 18:26:00 +0530
From: "Thunderbird" 
Organization: Motorola
Message-ID: <8mmbl7$r5t$1@schbbs.mot.com>

hi vxWorkers,


 I am facing some problem in reading and writing of data into some four
sockets then close these sockets. this is done repeatedly. then after some
30 such cycles, the network buffers are getting exhausted.(u can see this by
command "netStackDataPoolShow" on target shell). Once this happens tNetTask
shows error number 0x73000a and my socket read/write task is showing error
number 0x37.

My NUM_FILES is set to 300 and all parameters like NUM_64.....for data pool
and NUM_SYS_64... for system pool are configured according to faq posted on
WRS web page.

Does anybody know how to get out of this ?


regds,
venkat





---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: running out of buffers
Date: Mon, 07 Aug 2000 10:21:51 -0700
From: Joe Durusau 
Organization: Lockheed Martin Corporation
Message-ID: <398EF02F.FEA95AEB@delphi.com>
References: <8mmbl7$r5t$1@schbbs.mot.com>

	No solution, but a suggestion on where to look.  Check the status of
the 
sockets.  They are probably in close-wait, at least if they are TCP
sockets.
It takes a fairly long time for a TCP conneciton to close.  If your
cycle time is 
faster than the time required for a connection to complete the close
process, 
you will run out sooner or later.

Speaking only for myself,

Joe Durusau


Thunderbird wrote:
> 
> hi vxWorkers,
> 
>  I am facing some problem in reading and writing of data into some four
> sockets then close these sockets. this is done repeatedly. then after some
> 30 such cycles, the network buffers are getting exhausted.(u can see this by
> command "netStackDataPoolShow" on target shell). Once this happens tNetTask
> shows error number 0x73000a and my socket read/write task is showing error
> number 0x37.
> 
> My NUM_FILES is set to 300 and all parameters like NUM_64.....for data pool
> and NUM_SYS_64... for system pool are configured according to faq posted on
> WRS web page.
> 
> Does anybody know how to get out of this ?
> 
> regds,
> venkat

---------------------------

Newsgroups: comp.os.vxworks
Subject: Floating Point Problems
Date: Mon, 07 Aug 2000 14:53:04 GMT
From: "Kent Nilsson" 
Organization: Telia Internet
Message-ID: 

We are running VxWorks 5.3.1 on a Pentium processor. Okay, properly speaking
the Pentium is not supported by this VxWorks version, but nevertheless it
functions. Normally. Now we have tried to replace the 200 MHz Pentium with a
500 MHz Celeron processor, and then something happens with the floating
point calculations. They take about double the time on the Celeron compared
to the Pentium! Integer calculations are much faster on Celeron, but since
we have a lot of heavy floating point calculations the total speed is
considerably lower. Perhaps something fails in the co-operation between the
main processor and the math processor. Has anyone tackled this problem, or a
similar, before?



---------------------------

Newsgroups: comp.os.vxworks
Subject: Replacing global new and delete
Date: Mon, 7 Aug 2000 09:15:14 -0600
From: "Jim Anderson" 
Organization: SSO-IT, Hewlett-Packard Co.
Message-ID: <8mmjsr$n39$1@web1.cup.hp.com>

Due to some code re-use from a large tornado1 project (and a bug in the
2.95.1 compiler), we are trying to replace the global new and delete
operators.  I've worked out the syntax and have this working fine in VxSim.
The problem, of course, arrives when I try to build a rom image.  My
versions of the operators are in libmynew.a which is the first lib in the
LIBS macro.  VxWorks' version is in libARM7TDMIgnuvx.a, which is listed
second on the LIBS macro.
When I build a rom image, I get errors for each of the news and deletes
(normal and vector) I get multiple definition errors like such:

C:\Tornado\target\lib\libARM7TDMIgnuvx.a(opvdel.o)(.text+0x0):fake: multiple
definition of  `__builtin_vec_delete'
c:\myproj\myproj\lib\arm\libmynew.a
(spNewDel.o):C:\myproj\configs\svc\arm\c:\myproj\sp\src\spNewDel.cpp:61:
first defined here

Has anyone successfully done this before?  Any clues as to what I'm doing
wrong here?

I know that I can hand modify the arm7 lib to remove the .o files which are
in conflict, but I am hoping for something a bit cleaner than that.

Thanks!
jim anderson
jim_anderson@hp.com



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: raw keyboard input to Tornado shell process?
Date: Mon, 07 Aug 2000 16:23:46 GMT
From: p_lavarre@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8mmnqc$87m$1@nnrp1.deja.com>
References: <8m53v7$oco$1@nnrp1.deja.com> <39894EC0.DD7A9548@oce.nl>


> > p_lavarre@my-deja.com ...
> > How can a process run from the Tornado shell
> > ask to get one keystroke at a time from the shell input device?
> > I'm running `windsh` to talk to a target server,
> > thence via wdb serial to my target itself ...

> article <39894EC0.DD7A9548@oce.nl> ...
> Gerald van Kampen  ...
> ioctl(STD_IN, FIOSETOPTIONS, OPT_RAW);

Thanks for the pointer - via Tornado shell Ctrl+w that led to:
        Tornado/docs/vxworks/ref/ioLib.html#ioctl
hence the following fun experiments.

Please clarify.  Am I correct to imagine: running your code on the
target IS the answer IF I have a target shell.  However ...

While running a HOST SHELL, to react to a single keystroke, I need a
different answer?

Helped here, I have now found:
        ioctl(STD_IN, FIOGETOPTIONS, 0);
For me, BY DEFAULT, this returns 0.  The target/h/ioLib.h source tells
me this means OPT_RAW.  And I did try contrasting:
        ioctl(STD_IN, FIOSETOPTIONS, OPT_RAW);
        ioctl(STD_IN, FIOSETOPTIONS, OPT_TERMINAL);

I do see some small differences.

For example, running `windsh` directly, with OPT_RAW, I see only one
echo of each input line.  By contrast, with OPT_TERMINAL, I see two
echoes of each input line.  The power-up default is one echo i.e.
OPT_RAW.

Running `windsh` in Tornado, I see zero or one echoes, with OPT_RAW and
OPT_TERMINAL respectively.  Again the power-up default is OPT_RAW.

No matter my choice, I see ...

1) ... a call to:
        getchar()
does not return to until I press Ctrl+J or Ctrl+M.

2) ... everything I type is echoed.  (I say "everything", actually, I
have tried only Ctrl+A thru Ctrl+Z and "abc" and "Q".  Some of the Ctrl
keys are eaten by Windows unless I run `windsh` in a Dos box.)

3) ... Ctrl+C kills my task.

4) ... getchar never returns Ctrl+M as input: getchar returns Ctrl+J
instead.

Ctrl+D is fun to play with.  With OPT_RAW, it is just another input
character.  Even with OPT_TERMINAL, pressing Ctrl+D doesn't make
getchar return - you still have to press Ctrl+M.  But with
OPT_TERMINAL, Ctrl+D does temporarily kill off stdin.  getchar starts
returning EOF until I exit the process and start another.

Cheers.    Pat LaVarre
p.lavarre@ieee.org    http://members.aol.com/ppaatt/


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: why stdin only sometimes has a `vi` interface
Date: Mon, 07 Aug 2000 16:39:59 GMT
From: p_lavarre@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8mmoou$8rm$1@nnrp1.deja.com>
References: <8ma8te$jde$1@nnrp1.deja.com> <3988C1F1.6DF745D9@aps.anl.gov>


> > ... but I get no `vi` support during C use of stdin:
> >         ->getchar
> > ...
> > I'm using VxWorks 2.0 with a
> > host shell talking through a target server
> > then via wdbserial to a target.

> <3988C1F1.6DF745D9@aps.anl.gov>
> Andrew Johnson  ...

> Have a look at ledLib in the vxWorks Reference Manual
> if you want to add line and history editing support to your code.

Please elaborate - I'm thinking
        Tornado/docs/vxworks/ref/ledLib.html
helps only if I have first achieved raw access to a tty?

It seems raw access to a tty is not readily available in the host shell?

I'm trying to understand where it is in Tornado that the distinction is
made between the line-edited:
        -> ?gets stdin
and the not line-edited but still line-buffered:
        -> getchar
not to mention the line-editing available at:
        ->
        tcl>

I think the code that gives me line-buffering & line-editing is running
somewhere on the host?

I'm starting to think the line buffering/editing code CANNOT be Tcl
code - I'm starting to think that you cannot write a line-editor in
Tcl.  You have to use Tk - and Tk isn't in vxWorks, for example, I see:

tcl> button .hello -text "Hello, World!" -command { exit } ; pack .hello
invalid command name "button"

So I conclude raw access to the tty of the host shell is flatly not
available in vxWorks?  I can't get at it with C code on my target.  I
can't get at it with Tcl code on my host.  I just flatly can't get
there?

No?

Thanks again in advance.    Pat LaVarre
p.lavarre@ieee.org   http://members.aol.com/ppaatt/


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Question as re: the FAQ as re: Job Postings
Date: Mon, 7 Aug 2000 18:08:15 +0100
From: "Dave Korn" 
Organization: (Posted via) U-NET Internet Ltd.
Message-ID: 
References: 

Jeff Newman wrote in message ...
>Hello! My name is Jeff Newman with recruitDYNAMICS.com. I posted a job here
>awhile back, and somehow it was deleted!!!! I checked the FAQ and there is
>NOTHING that prohibits job postings. So, I have a few questions:
>
>a. Why would my post be deleted, unless someone did it on purpose?

  All posts expire after a while. Nothing sinister about it.

>b. Where should I post very specific VxWorks jobs, if not here?
>
>All suggestions are appreciated!!!!!!!!!!!!!!!!!!!!!!!!

  Do not, as one poster has suggested, gather email addresses from the
group! That would be very poorly received. Instead, read the FAQ
about how to advertise on usenet:-

"FAQ: Advertising on Usenet: How To Do It, How Not To Do It", by Furr


  In order to avoid excessive posting at a level that would be cancellable
as spam, please refer to the Usenet Current spam thresholds FAQ at



  Basically, the guidelines would suggest that there's no good reason
why you shouldn't post job adverts here, considering that VxWorks
is a speciality, and this is the only ng relating to it. Just remember to
keep them occasional, brief and to the point!

        hth,
           DaveK
- --
"Reality is whatever doesn't go away after you stop believing in it."
                                   -- Philip K. Dick



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: gcc-2.95.2-2, VxWorks, CygWin: Can't build it.
Date: Mon, 7 Aug 2000 18:08:28 +0100
From: "Dave Korn" 
Organization: (Posted via) U-NET Internet Ltd.
Message-ID: <66Cj5.1512$Vc.1115786@newsr2.u-net.net>
References: <%BYh5.1300$Vc.882168@newsr2.u-net.net> <3989cf0e.14192527@24.0.240.30>  <398bb68d.176932209@news.usenet-access.com>

Ami Ghua wrote in message <398bb68d.176932209@news.usenet-access.com>...
>On Thu, 3 Aug 2000 13:54:18 +0100, "Dave Korn"
> wrote:
>
>>Gary Beck wrote in message <3989cf0e.14192527@24.0.240.30>...
>>
>>>The reason I need GCC 2.95 rather than the Tornado shipped toolset
>>>(2.72) is the optimization.  Version 2.95 is much better than 2.72.
>>
>>  You're not joking. The bugs in the 2.7x optimization are so serious as
to
>>render it unuseable in production code. We've got a code base of about
>>half-a-million lines and bugs relating to the optimisation of
short-circuit
>>logical expressions and common subexpression elimination have sprinkled
>>bugs throughout the compiled code. It's possible to use temporary
variables
>>holding partial results to work around this bug but there's no way to find
>>out where in the code we have expressions that will be miscompiled.
>>
>
>This is very scary. I plan to use T2 for a production PPC cpu board.
>Does there exist any safe optimization flags that will avoid all these
>problem optimizations?
>
>Ami

  I haven't had any trouble with -O0. I have had trouble with -O1.
YMMV.

                 DaveK
- --
"Reality is whatever doesn't go away after you stop believing in it."
                                   -- Philip K. Dick



---------------------------

Newsgroups: comp.os.vxworks
Subject: help: porting chaivm to vxworks/ppc
Date: Mon, 07 Aug 2000 17:16:45 GMT
From: young_choi@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8mmqtn$aov$1@nnrp1.deja.com>

Hi,

I am porting ChaiVM(4.1) to vxWorks(5.4)/PowerPC(ppcec603) platform,
using gcc version cygnus-2.7.2-960126.
After I modifyed/added vm_config.pl, gcc.config, machine.hh,
main.cc..etc, as porting guide procedure,
I am getting following compiler error message. Please let know what is
the problem.

Thank you.
Young


/* following is captured error message */

Reading in classes
Sorting classes
Linking classes
Romster-izing
- ----- E:/jvm/CHAI/Multi_Target/CHAIVM/obj/linker.o -----
        ccppc -Wall -W -Werror -Wno-uninitialized -Wno-unused -Wpointer-
arith -DCPU=PPCEC603 -D
__vxworks -DRW_MULTI_THREAD -D_REENTRANT -DLANG=1  -DUTIL=1  -DIO=1  -
DNET=1  -DREFLECTION=1  -
DSERIALIZATION=1  -DZIP=1  -DZIPPCL=1  -DJDKSPEC=11  -DCLASS_GC=1  -
DCLASS_LOADER=1  -DNDEBUG=1
 -I E:\jvm\Tornado/target/config/all -I E:\jvm\Tornado/target/h -
IE:/jvm/Chai/Multi_Target/Chai
VM/src/jvm -IE:/jvm/Chai/Multi_Target/ChaiVM/src/jvm/jniapi  -
IE:/jvm/CHAI/Multi_Target/CHAIVM/
obj -mcpu=ppcec603 -fno-implicit-templates  -fcheck-new -fno-builtin -
fvolatile -ansi -nostdinc
  -O2 -fnonnull-objects -fomit-frame-pointer  -DDO_INLINE  -o
E:/jvm/CHAI/Multi_Target/CHAIVM/o
bj/linker.o -c
E:/jvm/Chai/Multi_Target/ChaiVM/src/jvm/classes/linker.cc
E:/jvm/Chai/Multi_Target/ChaiVM/src/jvm/classes/linker.cc:101: no
matching template for `fixed_
array16::array_copy_function(omt_segment *, const
omt_segment &)' found
E:/jvm/Chai/Multi_Target/ChaiVM/src/jvm/classes/linker.cc:102: no
matching template for `fixed_
array16::array_copy_function(const
linked_method_info **, const lin
ked_method_info *const &)' found
E:/jvm/Chai/Multi_Target/ChaiVM/src/jvm/classes/linker.cc:124: no
matching template for `fixed_
array::array_copy_function(class_runtime_state
**, class_runtime_state *
const &)' found
E:/jvm/Chai/Multi_Target/ChaiVM/src/jvm/classes/linker.cc:142: no
matching template for `fixed_
array::array_copy_function(class_name_space **,
class_name_space *const &)'
 found
cc1plus.exe: warnings being treated as errors
E:/jvm/Chai/Multi_Target/ChaiVM/src/jvm/foundation/array.hh: In method
`fixed_array16::fixed_array16(const class fixed_array16 &)':
E:/jvm/Chai/Multi_Target/ChaiVM/src/jvm/foundation/array.hh:424:
warning: base class `class arr
ay_base16' should be explicitly initialized in the copy
constructor
E:/jvm/Chai/Multi_Target/ChaiVM/src/jvm/foundation/array.hh: In method
`fixed_array16::fixed_array16(const class fixed_array16 &)':
E:/jvm/Chai/Multi_Target/ChaiVM/src/jvm/foundation/array.hh:424:
warning: base class `class arr
ay_base16' should be explicitly initialized
in the copy constructor

NMAKE : fatal error U1077: 'ccppc' : return code '0x1'
Stop.

E:\jvm\CHAI\Multi_Target\CHAIVM\obj>
E:\jvm\CHAI\Multi_Target\CHAIVM\obj>gcc -v
Reading specs from E:\jvm\Tornado\host\x86-win32\lib\gcc-lib\powerpc-
wrs-vxworks\cygnus-2.7.2-960126\specs
gcc driver version cygnus-2.7.2-960126 egcs-971225 tornado 2.0
executing gcc version cygnus-2.7.2-960126

E:\jvm\CHAI\Multi_Target\CHAIVM\obj>nmake info
Microsoft (R) Program Maintenance Utility   Version 6.00.8168.0
Copyright (C) Microsoft Corp 1988-1998. All rights reserved.
"E:/jvm/Chai/Multi_Target/ChaiVM/src/build_tools/config_vm.pl -
arch=powerpc -comp=gcc -threads=
VxWorks -jdkspec=11 -vm_internals -notextpkg -nomath -novm_internals -
normi -noi18n -nosql"
" -arch=powerpc -compiler=gcc -threads=VxWorks -dyncload -classgc -
jdkspec=11 -release -inline=1 -rom"
"libs: lang util io net reflection serialization zip zippcl "
E:\jvm\CHAI\Multi_Target\CHAIVM\obj>


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: why stdin only sometimes has a `vi` interface
Date: Mon, 07 Aug 2000 13:39:15 -0500
From: Andrew Johnson 
Organization: APS, Argonne National Laboratory
Message-ID: <398F0253.CCA2C14@aps.anl.gov>
References: <8ma8te$jde$1@nnrp1.deja.com> <3988C1F1.6DF745D9@aps.anl.gov> <8mmoou$8rm$1@nnrp1.deja.com>

p_lavarre@my-deja.com wrote:
> 
> Andrew Johnson  wrote:
> 
> > Have a look at ledLib in the vxWorks Reference Manual
> > if you want to add line and history editing support to your code.
> 
> Please elaborate - I'm thinking
>         Tornado/docs/vxworks/ref/ledLib.html
> helps only if I have first achieved raw access to a tty?
> 
> It seems raw access to a tty is not readily available in the host shell?

Oops, sorry I didn't see your "host shell" part, ledLib is for the target
shell (which we use exclusively).  You're probably right that you don't
have raw tty access through the host shell - you may be able to create
some kind of host-end component to handle this side of things for you, but
I have no experience of that.

- - Andrew
- -- 
Sharks kill 10 people each year.  People kill 60,000,000 sharks each year.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Enet Boot using MUX Driver
Date: Mon, 07 Aug 2000 19:03:33 GMT
From: maitra_nc_rtp@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8mn160$g4j$1@nnrp1.deja.com>

Hello,
Has anyone tried to boot using the motFecEnd.* driver
instead of the "cpm0" net-if driver?  Platform is PPC860, VxWorks 5.4.

I hope to find someone who has replaced a net-if driver in FLASH
and is using an END driver in its place -- right before
the boot (just before BootP requests are issued from FLASH on the net).

The flow in question is usrInit() -> usrRoot() ->
Task-bootCmdLoop() -> autoBoot() -> bootLoad() ->
(bootpGet() and then  [netLoad() -> ftpXfer()] )  in  bootConfig.c.

Thanks.

- --
pradip dot maitra at ericsson dot company
Views expressed are not necessarily shared by Ericsson, Inc.


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: IP layer API?
Date: 7 Aug 2000 19:14:17 GMT
From: Jeff Rugen 
Organization: PrimeNet
Message-ID: <8mn1q9$r6c$1@nnrp02.primenet.com>

I've just moved to a project where we are moving an embedded network
system to VxWorks.  The original system runs TP4 over IP (instead of TCP).
I'm new to VxWorks, and my first goal is to find an existing IP stack to
use.  Now, the first thing that comes to mind is to use the IP stack that
comes with VxWorks.  However, I can't find any information on interfacing
directly with the IP layer, and porting our TP4 implementation to run on
top of that.  Is this possible to do?  If so, where can I find information
on the API I'd be writing to (we are using VxWorks 5.4).  Right now I only
have the WindRiver VxWorks programmers guide, and I don't know what other
documentation is available.

If I can't do this easily, is there a good implementation of the IP stack
available that we can port our TP4 implementation to (or is there a TP4/IP
stack) that already works on VxWorks?  If I remember correctly, we will be
supporting UDP as well as TP4.  Thanks for any info.

- -- 
- ----------------------------------------------------------------------------
Jeff Rugen                      jrugen@primenet.com

When I said "we", officer, I was referring to myself, the four young ladies, 
and, of course, the goat.

---------------------------

Newsgroups: comp.os.vxworks
Subject: dec21x4xEnd driver
Date: Mon, 07 Aug 2000 20:11:28 GMT
From: stourangeau@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8mn55b$j7p$1@nnrp1.deja.com>

I am having problems setting up the dec21x4xEnd driver. I dont have the
source file so I am only playing with the configuration of the driver.

After lots of work, I was able to make the driver attach to the mux and
attach protocols to it.

I had lots of problem with getting the link to be indicated on the
hub... I had to play around with a mediaSelect routine..

Problem is, the hub sees the link but the dec21x doesnt seem to see the
hub. By looking at the status registers, I see link failures and
autonegociation that is not completed...

Need some help please...


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: VxWorks Network doesnt work on PC target platform
Date: Mon, 07 Aug 2000 23:17:23 +0200
From: Ezra Koper 
Organization: Bezeq International Ltd.
Message-ID: <398F2763.6E6DBB45@nevision.net.il>

Hello All.

We are using in first stage Pentium2 PC with Intel 100PRO Ethernet card
as a target platform.
I was able to build a boot loader & bootable VxWorks that is downloaded
to the target by FTP.
No problem to activate the target server & shell window over the
Ethernet Network connection.
My problem is that if i try any network command as socket() or changing
interface IP address, the shell window or my evaluation software
hang-up. It seem as if i can't do any network commands at all

Any idea what stage i have skipped or what has to be done ?

Note - i am planning to use the Network interface for debug & my real
application at the same time.

Thanks in advanced!!!!


---------------------------

Newsgroups: comp.os.vxworks
Subject: Ctrl+D synopsis for my own code?
Date: Mon, 07 Aug 2000 23:43:38 GMT
From: p_lavarre@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8mnhja$spi$1@nnrp1.deja.com>

How do I make Ctrl+D and Ctrl+W work for my own symbols?

I'm using a HOST shell.  I have found the customization examples:
        Tornado/docs/tornado/api/c-extend.html
        Tornado/docs/tornado/api/c-adding.html
These examples don't seem to discuss this - if I worked through them in
detail, would I discover different?  Is teaching new words to Ctrl+D
somehow a function of `refgen`?

I'm looking for the LEGITIMATE technique.  I have already stumbled
across an illegitimate technique.  A `grep` through *.* including
binaries led me to realise that a quick, simple way that works is to
edit:

        host/resource/synopsis/windsh.syn

Adding:

        host/resource/synopsis/LaVarre.syn

does NOT work.  But what does work is to insert text into the
windsh.sym file.  I find I can I insert my own symbols there, in sorted
order, using LF as the EOL and precisely duplicating the syntax of
other entries ... this works.  For example, I can add:

- -----
vi() - style of command-line editing

    Ctrl+U  erase line and start over
    k       move back one command
    /       move back to a command containing string (not regexp)
    j       move forward one command

After adding this, pressing v i blank Ctrl+D at the Tornado shell
prompt reminds me how to use the `vi` style command editing.  If I
actually add "vi" to the `lkup` symbol table, then the sequence v i
Ctrl+D Ctrl+D works.

But what is the LEGITIMATE technique for adding to the text that Ctrl+D
can find in a HOST shell?

Thanks in advance.    Pat LaVarre
p.lavarre@ieee.org    http://members.aol.com/ppaatt/


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Ctrl+D synopsis for my own code?
Date: Tue, 08 Aug 2000 00:28:20 GMT
From: p_lavarre@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8mnk70$uq5$1@nnrp1.deja.com>
References: <8mnhja$spi$1@nnrp1.deja.com>


> How do I make Ctrl+D and Ctrl+W work for my own symbols?

Search for the string ".syn" (i.e. regexp "\.syn") suggests that a tool
named `syngen` might be relevant, running in place of `refgen`.  The
source for the `syngen` tool seems to appear at:

        Tornado/host/resource/doctools/syngen.tcl

Perhaps my Ctrl+D question is actually a way of asking WHERE IS AN
EXAMPLE of using this tool to add the synopses of my own Tcl shellprocs
and my own C routines?

Thanks again in advance.    Pat LaVarre
p.lavarre@ieee.org    http://members.aol.com/ppaatt/


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: VxWorks Network doesnt work on PC target platform
Date: Tue, 08 Aug 2000 01:00:32 GMT
From: john_94501@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8mnm3e$3c$1@nnrp1.deja.com>
References: <398F2763.6E6DBB45@nevision.net.il>

Hi,

If you're using the network for the host tools connection I don't think
changing its IP address from the host shell (or in fact from anywhere)
would be a smart move... the existing connections between the host and
target for the tools are not going to work well like that.

Calls to socket() should work though...

HTH,

John...

In article <398F2763.6E6DBB45@nevision.net.il>,
  Ezra Koper  wrote:
> Hello All.
>
> We are using in first stage Pentium2 PC with Intel 100PRO Ethernet
card
> as a target platform.
> I was able to build a boot loader & bootable VxWorks that is
downloaded
> to the target by FTP.
> No problem to activate the target server & shell window over the
> Ethernet Network connection.
> My problem is that if i try any network command as socket() or
changing
> interface IP address, the shell window or my evaluation software
> hang-up. It seem as if i can't do any network commands at all
>
> Any idea what stage i have skipped or what has to be done ?
>
> Note - i am planning to use the Network interface for debug & my real
> application at the same time.
>
> Thanks in advanced!!!!
>
>


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Writing Device Drivers
Date: Mon, 7 Aug 2000 12:30:32 +0200
From: "Ramakrishnan,CN" 
Organization: UUNET-NL (http://www.nl.uu.net)
Message-ID: 
References: <8m95or$720$1@news.news.kodak.com>

What is the device for which you wish to develop the driver. I have made
a small write up on serial device driver. I will send it to you.
Ramakrishnan.C.N
Peek Traffic B.V.


> ----------
> From: 	Tim Stanka[SMTP:stanka@mojo.kodak.com]
> Posted At: 	maandag 31 juli 2000 14:51
> Posted To: 	vxworks
> Conversation: 	Writing Device Drivers
> Subject: 	Writing Device Drivers
> 
> I want to write my first VXWORKS device driver. I have a fair
> top-level
> understanding of the task but
> I find the two books that discuss drivers (VXWORKS Reference Manual
> and BSP
> Developer Manual)
> don't have alot of detail. Does anyone know or have documentation that
> would
> help me out.
> 
> Thanks
> 
> 


---------------------------

Newsgroups: comp.os.vxworks
Subject: re: Vx Exploder Digest
Date: Tue, 8 Aug 2000 12:59:12 +0530
From: "R.Chandrasekar" 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <8mofi2$bt1$1@overload.lbl.gov>

Hi KBS,

You need to get the site Key from WRS or your local representative. This
again depends on the number of installations licensed along with your
tornado software.

You need to send the site code and IP address of your machine. Then you will
get a site key which will make you have unlimited license for that project
development.

Alternately you can transfer the site key from existing licensed machine to
the new one. Your old machine will then have  no more license

Hope u r clear.

Bye
Chandrasekar.R
Engineer R &D
CMC Limited
Real Time Systems
Hyderabad, India

>Newsgroups: comp.os.vxworks
>Subject: How can i enable target server?
>Date: Mon, 07 Aug 2000 09:27:56 GMT
>From: "KBS" 
>Organization: Dacom
>Message-ID: 
>
>I'm using VxWork 5.4 for Windows NT version. But the target server is only
>30 day version and now expired.
>How can i enable the server at NT server? At the starting target server, it
>ask the "site key" value for given "site code". How can i get it? Somebody
>help me!
>
>Thanks for your help.
>Bongsoo Kang



---------------------------

Newsgroups: comp.os.vxworks
Subject: Tornado Launch
Date: Tue, 08 Aug 2000 11:07:05 +0100
From: Martin Hewitson 
Organization: Universitaet Hannover
Message-ID: <398FDBC8.2EBB9582@astro.gla.ac.uk>

Hell out there,

Maybe someone can help. I am using the Tornado Launch program to
(primarily) boot and control three data servers that have vxWorks
installed on them. On occasion, the launch program decides it can't
connect to the servers and starts complaining about the license server
being down??!!? Wind River gave me instructions as to how to install a
new license and this seems to go some way to solve the problem. However,
at the end of the key installation, the install program says that the
license server is down. When we use launch to connect to the servers
now, it works.  Anybody any idea why it is going wrong in the first plae
and why a failed attempt to re-install the license key should correct
the problem?

Any suggestions would be gratefully received,

thankjs

M


---------------------------

End of New-News digest
**********************


From vxwexplo-errs@csg.lbl.gov  Tue Aug  8 08:05:19 2000
From: Jim Way 
Date: Tue Aug  8 08:05:22 PDT 2000
Subject: Optimization bugs

RE: Discussion on optimization

I'm using VxWorks 5.4 / Tornado II / gcc 2.72

In general, I'm compiling with -O2 but I have a couple files that I must
compile with -O0 (zero optimization). My specific clue is the error message:
STATUS_ACCESS_VIOLATION although I am not certain exactly *what* that means.
I just know it goes away when I switch to -O0.

Other than the compile-time problem, I haven't seen anything that I can
specifically point to as an optimization error. As we all know, the absence
of evidence does not prove the absence of bugs. But my product is still new.
What types of things might I be seeing?

Jim
-----------------------------------------
Jim Way, Software Engineer
Datum Austin (Austron Inc.)
voice: 512.721.4170
fax  : 512.990.9712
email: jwayATdatumDOTcom (no spam please)
-----------------------------------------


From vxwexplo-errs@csg.lbl.gov  Tue Aug  8 08:05:30 2000
From: David Laight 
Date: Tue Aug  8 08:05:32 PDT 2000
Subject: re: Replacing global new and delete

Jim Anderson wrote:

> Due to some code re-use from a large tornado1 project (and a bug in the
> 2.95.1 compiler), we are trying to replace the global new and delete
> operators.  I've worked out the syntax and have this working fine in VxSim.
> The problem, of course, arrives when I try to build a rom image.  My
> versions of the operators are in libmynew.a which is the first lib in the
> LIBS macro.  VxWorks' version is in libARM7TDMIgnuvx.a, which is listed
> second on the LIBS macro.
> When I build a rom image, I get errors for each of the news and deletes
> (normal and vector) I get multiple definition errors like such:

> C:\Tornado\target\lib\libARM7TDMIgnuvx.a(opvdel.o)(.text+0x0):fake: multiple
> definition of  `__builtin_vec_delete'
> c:\myproj\myproj\lib\arm\libmynew.a
> (spNewDel.o):C:\myproj\configs\svc\arm\c:\myproj\sp\src\spNewDel.cpp:61:
> first defined here

If you are getting multiple definitions, then the WRS `__builtin_vec_delete' 
function must be linked with some other symbols that you need - which pulls it 
in - rather than being a separate entity in the .a file.  It is somewhat 
unfortunate that WRS seem to do 'ld -r' links of sets of files before putting 
them into the .a (I also suspect there are rather too many #include  
requests as well.

So you won't manage to remove the single definition from the arm7 library.

> I know that I can hand modify the arm7 lib to remove the .o files which are
> in conflict, but I am hoping for something a bit cleaner than that.

What I do to include a replacement copy of a routine is to overwrite the first 
instruction of the WRS routine with a jump to be replacement one.  This allows 
you to re-engineer a single routine from a large source file (although finding 
the address of local routines can be taxing).  If the change only involves 
fixing the argument, an assembler fix can jump back to the second instruction.

For example (for ARM):

	extern int ping, fix_ping;
	/* fix_ping changes ping "host",0 to ping "host",5 */
	ping = 0xea000000 | ((&fix_ping - &ping - 2) & 0xffffff);
	cacheTextUpdate( &ping, 4 );

        .global _fix_ping
_fix_ping:
        cmp     r1,#0
        moveq   r1,#5
        mov     r12,r13		/* first instruction of ping */
        b       _ping+4

(Note that defining the code addresses as 'int' makes the source simpler)

I currently also patch:

1) open( "//name", ... ) to open( "/name", ... )
2) the body of mountproc_mnt_1 to call an alternate nfsExportFindByName that
   allows /dosfs/ to be exported and mounted as /dosfs or /dosfs/
   (without the patch it will mount as "/dosfs//")
3) make check_access null (part of ppp) to remove the warning about access 
   to the ppp secrets file
4) the body of ppp_wrt_task to ignore 'T' (ie the echo of AT isn't the end
   of CLIENT)
5) memPartAlignedAlloc to use a best fit algorithm
6) nfsproc_create_2 so that is doesn't return an errno value instead of a
   structure
7) setUglColorModel so that it allocates 6 shades of each of red, green and
   blue
8) Java_sun_awt_image_ImageRepresentation_setIntPixels so that it doesn't skip
   the dither code when the input colour hasn't changed
9) uglColorFind to a copy that works about 7 times faster (and is still
   the most used routine)
10) endFindByName to support multiple drivers
11) _sun_awt_motif_MToolkit_run so back to back calls don't lock up.

Most of the above haven't made it past a TSR into an SPR.

	David

----------------------------------------------------------------
David Laight                          email: dsl@tadpole.co.uk
Tadpole Technology plc                phone: +44 1223 428 232
Cambridge, UK                         fax:   +44 1223 428 201



From vxwexplo-errs@csg.lbl.gov  Tue Aug  8 12:28:29 2000
From: "Tim Allen" 
Date: Tue Aug  8 12:28:31 PDT 2000
Subject: Re: Arcom MediaGX 233 Board + CT69000 Chipset - need ~Help!!!

This has been delayed because I omitted to say VxWorks!

>I'm tring to run my grahical application on the MediaGX Board which runs a
>Cyrix 233 Mhz chip and contains an embedded CT69000 graphics chip.
>
>I have downloaded the driver for this board from Wind River's site and
>managed to include it within my build.
>
>It seems to cause a page fault when I call uglInitialize() - specifically
>within the set_display_mode() function.
>
>Am I supposed to make an entry within the syslib.c source just like one
>would with a CT65555 entry ?

If you mean the MMU setup, then the answers yes, as someone indicated a day
or two ago on the exploder. This probably would explain the page fault.

>Has anyone ever got this working ??

Although I did this some time ago, and the driver may have changed, I had
lots of trouble getting this driver going.  The 69000 "driver" was basically
a dump of the intial values taken from a specific hardware platform.  Only
by going through the registers value by value did I find that some of the
settings were not suitable for our hardware platform, and also that some of
the registers set, did not even exist on the 69000. Unfortunately I don't
have access to the source code so I can't help with the detail!

Having worked out the correct register settings I remember using the updated
settings with the 65555 driver since it came with more supporting functions.
This then worked on the Arcom board as well as some other boards with the
69000 chip.

Regards,
Tim.
--------------------
Tim Allen Ltd
Real Time Embedded Software Consultancy




From vxwexplo-errs@csg.lbl.gov  Tue Aug  8 21:32:54 2000
From: "Chandrashekar Nagaraj" 
Date: Tue Aug  8 21:32:56 PDT 2000
Subject: Same Error Status Strings........

Hi,
    As everyone know that each header file in vxWorks will have a set of Error Status Values defined in them, in the following manner

#define S_loadLib_FILE_READ_ERROR               (M_loadPecoffLib | 1)
#define S_loadLib_REALLOC_ERROR                 (M_loadPecoffLib | 2)
#define S_loadLib_JMPADDR_ERROR                 (M_loadPecoffLib | 3)
#define S_loadLib_FILE_ENDIAN_ERROR             (M_loadPecoffLib | 8)
#define S_loadLib_FILE_READ_ERROR               (M_loadCoffLib | 1)
#define S_loadLib_REALLOC_ERROR                 (M_loadCoffLib | 2)
#define S_loadLib_JMPADDR_ERROR                 (M_loadCoffLib | 3)
#define S_loadLib_FILE_ENDIAN_ERROR             (M_loadCoffLib | 8)

But if  u take a clear notice of these Error Status values in the loadPecoffLib.h and loadCoffLib.h we see that some of the Error Status are same( i mean
the error status strings). This does not cause any problem for program execution because the error status values are distinct for each of them but the compiler will give
warnings because of redundancy. Is there any way to remove these warnings,
we have tried changing the preprocessor's flag values, commenting it but none of
them works.

The program environment is something like this.

We create statTbl of vxWorks using makestatTbl.
Then we try to build our project files with statTbl.c being
one among the files. And we have /host/target/h as one of the
dependency.

Thank You,
Chandra.



From vxwexplo-errs@csg.lbl.gov  Wed Aug  9 01:56:14 2000
From: David Anderson 
Date: Wed Aug  9 01:56:16 PDT 2000
Subject: RE: Same Error Status Strings........

On 09 August 2000 05:52, the vxWorks Users Group Exploder 
[SMTP:vxwexplo@lbl.gov] wrote:
> Submitted-by vxwexplo-errs@csg.lbl.gov  Tue Aug  8 21:32:54 2000
> Submitted-by: "Chandrashekar Nagaraj" 
>
> Hi,
>     As everyone know that each header file in vxWorks will have a set of 
Error Status Values defined in them, in the following manner
>
> #define S_loadLib_FILE_READ_ERROR               (M_loadPecoffLib | 1)
> #define S_loadLib_REALLOC_ERROR                 (M_loadPecoffLib | 2)
> #define S_loadLib_JMPADDR_ERROR                 (M_loadPecoffLib | 3)
> #define S_loadLib_FILE_ENDIAN_ERROR             (M_loadPecoffLib | 8)

how about adding #ifndef / #endif conditional compilation around both in 
both header files.
You would not need to change the compiler flags which tend to have a global 
affect on everything.

#ifndef S_loadLib_FILE_READ_ERROR
#define S_loadLib_FILE_READ_ERROR               (M_loadCoffLib | 1)
#endif

#ifndef S_loadLib_REALLOC_ERROR
#define S_loadLib_REALLOC_ERROR                 (M_loadCoffLib | 2)
#endif

#ifndef S_loadlib_JMPADDR_ERROR
#define S_loadLib_JMPADDR_ERROR                 (M_loadCoffLib | 3)
#endif

#ifndfe S_loadlib_FILE_ENDIAN_ERROR
#define S_loadLib_FILE_ENDIAN_ERROR             (M_loadCoffLib | 8)
#endif

>
> But if  u take a clear notice of these Error Status values in the 
loadPecoffLib.h and loadCoffLib.h we see that some of the Error Status are 
same( i mean
> the error status strings). This does not cause any problem for program 
execution because the error status values are distinct for each of them but 
the compiler will give
> warnings because of redundancy. Is there any way to remove these 
warnings,
> we have tried changing the preprocessor's flag values, commenting it but 
none of
> them works.
>

One thing that I've come across is that .h files compiled into .cpp source 
produce different
warnings than if they are compiled into .c source. This arises from the 
differences in the error
and warning checking. You do not say but I would not be surprised to find 
that a .c source did
not produce your warnings.


David Anderson.  mailto:danderson@dtrack.demon.co.uk

-----------------------------------------------------
Who said "There's no life here in outer space."
-----------------------------------------------------

Disclaimer
----------
This email and any files transmitted with it are confidential and intended 
solely for the use of the individual or entity to whom they are addressed.

If you have received this email in error please destroy and notify Data 
Track Technology Plc +44 1425 271900.
-------------------------------------------------------




From vxwexplo-errs@csg.lbl.gov  Wed Aug  9 04:03:08 2000
From: Vxworks Exploder 
Date: Wed Aug  9 04:03:10 PDT 2000
Subject: comp.os.vxworks newsdigest

Comp.Os.Vxworks Daily Digest    Wed Aug  9 04:03:05 PDT 2000

        Subject: Does anyone use VxWorks on a standard PC?
        Subject: Re: UDP problem
        Subject: Re: Tornado Launch
        Subject: Optimization bugs
        Subject: re: Replacing global new and delete
        Subject: Re: Question as re: the FAQ as re: Job Postings
        Subject: MUX ReceiveRtn
        Subject: NTP client on VxWorks
        Subject: How to Support Dynamic memory sizes
        Subject: anyone used diab and RTA under vxworks?
        Subject: re: Arcom MediaGX 233 Board + CT69000 Chipset - need ~Help!!!
        Subject: Re: Failure to load application module with T2
        Subject: Re: Does anyone use VxWorks on a standard PC?
        Subject: mv2700 guru: mbx/lm - pci io reg's vs univ chip
        Subject: Re: Does anyone use VxWorks on a standard PC?
        Subject: Is there any on-line trainning course of VxWorks?
        Subject: Re: MUX ReceiveRtn
        Subject: Same Error Status Strings........
        Subject: RE: Same Error Status Strings........
        Subject: Re: VxWorks FTP boot problem (T2, MVME2700)

-------------------------------------------------------

Newsgroups: comp.os.vxworks
Subject: Does anyone use VxWorks on a standard PC?
Date: Tue, 08 Aug 2000 11:17:10 GMT
From: "Charles Magne" 
Organization: Guest of ProXad - France
Message-ID: 

If you are using that,
I d like to know if it is necessary use an additionnal card?
and where to find the BSP associated.

Thanks a lot.
 Charles



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: UDP problem
Date: Sat, 29 Jul 2000 17:38:56 -0500
From: "VIJAY" 
Organization: MOTOROLA
Message-ID: <8lvme2$cuj$1@schbbs.mot.com>
References: <39760EC0.1EAEF7BF@139.85.103.45> <39761F27.62260CB0@bellsouth.net>

Adding to Joe's reply, run ->netStackDataPoolShow command in the tornado
shell and check for the free UDP buffers.  They should not be zero's.

You can increase the number of those buffers using TORNADO2 GUI while
building your vxworks image.

Hope it helps.

vijay kumar
(vijaykulal@hotmail.com)


Joe Durusau  wrote in message
news:39761F27.62260CB0@bellsouth.net...
> You are probably exceeding the capacity of some buffer
> in the system, on one end or the other.  Running snoop or some such
> would identify which end.  It's possible that crosswind could give
> a clue to the problem if it's on the receive end, but of course
> it will add some overhead to your system.  I suggest that you need
> to remember that BY DEFINITION, udp does NOT guarantee transmission,
> since the higher level protocols are expected to make arrangements
> for recover of lost datagrams.
>
> Speaking only for myself,
>
> Joe Durusau
>
>
> ico wrote:
> >
> > Hi all,
> >           We are trying to run UDP bench mark tests for VxWorks on a
> > 100baseT ethernet. We have only 2 machines on the network. The client
> > sends a burst of 10 packets (500 bytes each) to a server, then waits
> > for  20-50ms and sends another burst. On the server side there are four
> > tasks(of same priority) running. The packets are transferred from one
> > task to another with each task doing some processing on the packet
> > before sending it out to a subsequent task. One of the tasks breaks up
> > the 500 byte packet into 100 byte packets. The packets are then sent
> > back to the client. The client measures the time it takes to get back
> > the 500 byte packets(approx 1ms per 500 bytes). If we sent out 10,000
> > packets or so then we seem to be losing packets towards the end. Sending
> > 100 packets doesn't  seem to give any problems.   Any help would be
> > greatly appreciated.
> >
> > Thanks
> >
> > naveen



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Tornado Launch
Date: Tue, 08 Aug 2000 06:59:56 -0700
From: Joe Durusau 
Organization: Lockheed Martin Corporation
Message-ID: <3990125C.1E3D9C0F@delphi.com>
References: <398FDBC8.2EBB9582@astro.gla.ac.uk>

	Which version of Tornado, what OS?  On my system, running Tornado
1.0.1,
Solaris 2.6, there is a file in /etc/rc2.d that starts lmgrd with a
parameter
pointing to the license file.  whatever equiv. you have may be damaged,
or you
may need to restart the server.  AFAIK, anybody can start it.  Try
ps -e | grep  to see if the server is
in fact 
running.   If it is, send it a hangup and maybe it will reread the
license file.
If not , run the command that starts it.

HTH.

Speaking only for myself,

Joe Durusau
 

Martin Hewitson wrote:
> 
> Hell out there,
> 
> Maybe someone can help. I am using the Tornado Launch program to
> (primarily) boot and control three data servers that have vxWorks
> installed on them. On occasion, the launch program decides it can't
> connect to the servers and starts complaining about the license server
> being down??!!? Wind River gave me instructions as to how to install a
> new license and this seems to go some way to solve the problem. However,
> at the end of the key installation, the install program says that the
> license server is down. When we use launch to connect to the servers
> now, it works.  Anybody any idea why it is going wrong in the first plae
> and why a failed attempt to re-install the license key should correct
> the problem?
> 
> Any suggestions would be gratefully received,
> 
> thankjs
> 
> M

---------------------------

Newsgroups: comp.os.vxworks
Subject: Optimization bugs
Date: Tue, 8 Aug 2000 08:03:04 -0700
From: Jim Way 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <8mp85h$q2j$1@overload.lbl.gov>

RE: Discussion on optimization

I'm using VxWorks 5.4 / Tornado II / gcc 2.72

In general, I'm compiling with -O2 but I have a couple files that I must
compile with -O0 (zero optimization). My specific clue is the error message:
STATUS_ACCESS_VIOLATION although I am not certain exactly *what* that means.
I just know it goes away when I switch to -O0.

Other than the compile-time problem, I haven't seen anything that I can
specifically point to as an optimization error. As we all know, the absence
of evidence does not prove the absence of bugs. But my product is still new.
What types of things might I be seeing?

Jim
- -----------------------------------------
Jim Way, Software Engineer
Datum Austin (Austron Inc.)
voice: 512.721.4170
fax  : 512.990.9712
email: jwayATdatumDOTcom (no spam please)
- -----------------------------------------

---------------------------

Newsgroups: comp.os.vxworks
Subject: re: Replacing global new and delete
Date: Tue, 8 Aug 2000 16:05:39 +0100 (BST)
From: David Laight 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <200008081505.QAA03918@dsl-2.tadpole.co.uk>

Jim Anderson wrote:

> Due to some code re-use from a large tornado1 project (and a bug in the
> 2.95.1 compiler), we are trying to replace the global new and delete
> operators.  I've worked out the syntax and have this working fine in VxSim.
> The problem, of course, arrives when I try to build a rom image.  My
> versions of the operators are in libmynew.a which is the first lib in the
> LIBS macro.  VxWorks' version is in libARM7TDMIgnuvx.a, which is listed
> second on the LIBS macro.
> When I build a rom image, I get errors for each of the news and deletes
> (normal and vector) I get multiple definition errors like such:

> C:\Tornado\target\lib\libARM7TDMIgnuvx.a(opvdel.o)(.text+0x0):fake: multiple
> definition of  `__builtin_vec_delete'
> c:\myproj\myproj\lib\arm\libmynew.a
> (spNewDel.o):C:\myproj\configs\svc\arm\c:\myproj\sp\src\spNewDel.cpp:61:
> first defined here

If you are getting multiple definitions, then the WRS `__builtin_vec_delete' 
function must be linked with some other symbols that you need - which pulls it 
in - rather than being a separate entity in the .a file.  It is somewhat 
unfortunate that WRS seem to do 'ld -r' links of sets of files before putting 
them into the .a (I also suspect there are rather too many #include  
requests as well.

So you won't manage to remove the single definition from the arm7 library.

> I know that I can hand modify the arm7 lib to remove the .o files which are
> in conflict, but I am hoping for something a bit cleaner than that.

What I do to include a replacement copy of a routine is to overwrite the first 
instruction of the WRS routine with a jump to be replacement one.  This allows 
you to re-engineer a single routine from a large source file (although finding 
the address of local routines can be taxing).  If the change only involves 
fixing the argument, an assembler fix can jump back to the second instruction.

For example (for ARM):

	extern int ping, fix_ping;
	/* fix_ping changes ping "host",0 to ping "host",5 */
	ping = 0xea000000 | ((&fix_ping - &ping - 2) & 0xffffff);
	cacheTextUpdate( &ping, 4 );

        .global _fix_ping
_fix_ping:
        cmp     r1,#0
        moveq   r1,#5
        mov     r12,r13		/* first instruction of ping */
        b       _ping+4

(Note that defining the code addresses as 'int' makes the source simpler)

I currently also patch:

1) open( "//name", ... ) to open( "/name", ... )
2) the body of mountproc_mnt_1 to call an alternate nfsExportFindByName that
   allows /dosfs/ to be exported and mounted as /dosfs or /dosfs/
   (without the patch it will mount as "/dosfs//")
3) make check_access null (part of ppp) to remove the warning about access 
   to the ppp secrets file
4) the body of ppp_wrt_task to ignore 'T' (ie the echo of AT isn't the end
   of CLIENT)
5) memPartAlignedAlloc to use a best fit algorithm
6) nfsproc_create_2 so that is doesn't return an errno value instead of a
   structure
7) setUglColorModel so that it allocates 6 shades of each of red, green and
   blue
8) Java_sun_awt_image_ImageRepresentation_setIntPixels so that it doesn't skip
   the dither code when the input colour hasn't changed
9) uglColorFind to a copy that works about 7 times faster (and is still
   the most used routine)
10) endFindByName to support multiple drivers
11) _sun_awt_motif_MToolkit_run so back to back calls don't lock up.

Most of the above haven't made it past a TSR into an SPR.

	David

- ----------------------------------------------------------------
David Laight                          email: dsl@tadpole.co.uk
Tadpole Technology plc                phone: +44 1223 428 232
Cambridge, UK                         fax:   +44 1223 428 201


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Question as re: the FAQ as re: Job Postings
Date: Tue, 8 Aug 2000 12:18:45 -0400
From: "Jeff Newman" 
Organization: Posted via Supernews, http://www.supernews.com
Message-ID: 
References:  

Dave,

I know things expire after awhile, but there are posts, on my server at
least, that are older then the post that I had made, which didn't make sense
to me. Also, the email "theft" would be bad! I post so that people can reply
if they are interested! Thanks for your feedback!

Jeff

"Dave Korn"  wrote in message
news:X5Cj5.1511$Vc.1115816@newsr2.u-net.net...
> Jeff Newman wrote in message ...
> >Hello! My name is Jeff Newman with recruitDYNAMICS.com. I posted a job
here
> >awhile back, and somehow it was deleted!!!! I checked the FAQ and there
is
> >NOTHING that prohibits job postings. So, I have a few questions:
> >
> >a. Why would my post be deleted, unless someone did it on purpose?
>
>   All posts expire after a while. Nothing sinister about it.
>
> >b. Where should I post very specific VxWorks jobs, if not here?
> >
> >All suggestions are appreciated!!!!!!!!!!!!!!!!!!!!!!!!
>
>   Do not, as one poster has suggested, gather email addresses from the
> group! That would be very poorly received. Instead, read the FAQ
> about how to advertise on usenet:-
>
> "FAQ: Advertising on Usenet: How To Do It, How Not To Do It", by Furr
> 
>
>   In order to avoid excessive posting at a level that would be cancellable
> as spam, please refer to the Usenet Current spam thresholds FAQ at
>
> 
>
>   Basically, the guidelines would suggest that there's no good reason
> why you shouldn't post job adverts here, considering that VxWorks
> is a speciality, and this is the only ng relating to it. Just remember to
> keep them occasional, brief and to the point!
>
>         hth,
>            DaveK
> --
> "Reality is whatever doesn't go away after you stop believing in it."
>                                    -- Philip K. Dick
>
>



---------------------------

Newsgroups: comp.os.vxworks
Subject: MUX ReceiveRtn
Date: Tue, 08 Aug 2000 14:26:57 GMT
From: Kans 
Organization: Deja.com - Before you buy.
Message-ID: <8mp5bd$16c$1@nnrp1.deja.com>

Hi,
I was reading up on SENS.. and on the driver section it says,
receiveRtn is actually a pointer in end_object and mux supplies the
pointer after muxDevLoad(). Can I have my own demultiplexing routine
instead of using muxReceiveRtn(). I have both the protocol and the
driver sources, and I intend to use the Protocol Type field for my own
purposes (not the rfc1700 assigned protocol number), that is why I
intend to write my own demultiplexing routine .. Or can I still
continue and use the muxReceieveRtn for my own purposes?.

Hope someone can help me out of this.
Kans.


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: NTP client on VxWorks
Date: Tue, 08 Aug 2000 07:44:05 -0700
From: Warren.Brown 
Organization: http://www.remarq.com: The World's Usenet/Discussions Start Here
Message-ID: <311c2be6.cfb476ec@usw-ex0106-045.remarq.com>

Hello
Has anyone added the NTP (Network Time Protocol) client
services to a VxWorks embedded target ?
Any hints as to the right steps to take ?

W


- -----------------------------------------------------------

Got questions?  Get answers over the phone at Keen.com.
Up to 100 minutes free!
http://www.keen.com


---------------------------

Newsgroups: comp.os.vxworks
Subject: How to Support Dynamic memory sizes
Date: Tue, 8 Aug 2000 13:38:55 -0500
From: "Jack" 
Organization: Motorola
Message-ID: <8mpk3v$d8c$1@schbbs.mot.com>

Anyone know how to support field upgrades of SIMM sizes in VxWorks 5.4?

Jack
fjk003@email.mot.com



---------------------------

Newsgroups: comp.os.vxworks
Subject: anyone used diab and RTA under vxworks?
Date: Tue, 08 Aug 2000 15:38:42 GMT
From: toddhoff@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8mp9i2$3tp$1@nnrp1.deja.com>



There's not a lot of documentation on rta on the windriver
site or related articles in dejanews. I was wondering if anyone
has used the diab/rta product under vxworks and would like to
share their experiences? I'm particularly interested in
the "purify" like aspects of rta.

thanx


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: re: Arcom MediaGX 233 Board + CT69000 Chipset - need ~Help!!!
Date: Tue, 8 Aug 2000 20:26:48 +0100
From: "Tim Allen" 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <8mppo4$72v$1@overload.lbl.gov>

This has been delayed because I omitted to say VxWorks!

>I'm tring to run my grahical application on the MediaGX Board which runs a
>Cyrix 233 Mhz chip and contains an embedded CT69000 graphics chip.
>
>I have downloaded the driver for this board from Wind River's site and
>managed to include it within my build.
>
>It seems to cause a page fault when I call uglInitialize() - specifically
>within the set_display_mode() function.
>
>Am I supposed to make an entry within the syslib.c source just like one
>would with a CT65555 entry ?

If you mean the MMU setup, then the answers yes, as someone indicated a day
or two ago on the exploder. This probably would explain the page fault.

>Has anyone ever got this working ??

Although I did this some time ago, and the driver may have changed, I had
lots of trouble getting this driver going.  The 69000 "driver" was basically
a dump of the intial values taken from a specific hardware platform.  Only
by going through the registers value by value did I find that some of the
settings were not suitable for our hardware platform, and also that some of
the registers set, did not even exist on the 69000. Unfortunately I don't
have access to the source code so I can't help with the detail!

Having worked out the correct register settings I remember using the updated
settings with the 65555 driver since it came with more supporting functions.
This then worked on the Arcom board as well as some other boards with the
69000 chip.

Regards,
Tim.
- --------------------
Tim Allen Ltd
Real Time Embedded Software Consultancy



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Failure to load application module with T2
Date: Tue, 08 Aug 2000 16:39:06 GMT
From: fywang@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8mpd3a$6br$1@nnrp1.deja.com>
References: 

Hi,

Did you see the LIST BOX in your tool bar of T2?
After you Launch your target server, you should use this LIST BOX to
choose a target server before downloading.

Wish this help,
Tom Wang

In article ,
  "Lou Kale"  wrote:
> I am having trouble loading an application module (.out) using
Tornado 2.
>
> I can boot the vxWorks, the vxWorks image loads over the ethernet
ok.  I try
> to download an application image via Tornado but this fails with a
> communication timeout.  I tried a direct connection between the host
and
> target so network traffic is not a problem.
>
> I believe that the vxWorks image is loaded using FTP and that Tornado
then
> tries to download the application using RPC.  I tried loading the same
> module using the console command "ld think that is done via FTP also.
>
> The application that I fail to load is about 200K bytes.  The Host is
NT.
> The Target is a MIPS.
>
> The target server log always has the following data in it:
>
> tgtsvr (192.168.0.25@carlson2): Fri Jul 14 16:33:17 2000
>     Wind River Systems Target Server: NT/Win95 version
>     Connecting to target agent... succeeded.
>     Attaching C++ interface... succeeded.
>     Attaching elf OMF reader for MIPS CPU family... succeeded.
>
> tgtsvr (192.168.0.25@carlson2): Fri Jul 14 16:34:15 2000
>     Error: rpccore backend client Timed out
>     Error: Unrecoverable trouble while loading object module.
>     Error: Object module load failed for
> D:/SWPROJECT/EvLineCard/Application/RC4650/KubeLcRoot.out
>
> --
> Lou Kale
> KeNETec Inc.
> 115 Hurley Road
> Building 1B
> Oxford, CT 06478
> Tel (203)-264-1888 x130
> Fax (203)-264-6724
> lou.kale@kenetec.com
>
>


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Does anyone use VxWorks on a standard PC?
Date: Wed, 09 Aug 2000 01:01:04 GMT
From: nobody@nowhere.com (coder)
Organization: SBC Internet Services
Message-ID: <3990ad22.10229036@news.swbell.net>
References: 

In comp.os.vxworks
"Charles Magne"  wrote:

>If you are using that,
>I d like to know if it is necessary use an additionnal card?
>and where to find the BSP associated.

An additional card for what?  The BSP is PC486 and you find it in your
target directory if you bought and installed it.


---------------------------

Newsgroups: comp.os.vxworks
Subject: mv2700 guru: mbx/lm - pci io reg's vs univ chip
Date: Tue, 08 Aug 2000 20:58:48 -0400
From: Al Johnston 
Organization: MindSpring Enterprises
Message-ID: <3990ACC8.6CE17548@mindspring.com>


Can someone explain the relationship between the 2700 pci io registers
at 0x1000 (sig/lm ctrl reg, sig/lm stat reg, ... pg 1-40 of prog.
ref.) and the similar components on the universe chip (sect 2.2.3.7,
2.7.3.5, 2.7.3.6, 2.9.4 of Univ. manual)?

These two parts of the card both provide location monitor/mailbox
type capabilities.  Does the former (pci io registers) use the
latter (Univ. registers) to provide this capability, or are the
two methods of doing mb/lm independent of each other?

I understand that the universe would be used to map memory/io
space in ether case, so I am not asking about that.  I am 
asking about the universe's "interrupt" capability -vs- the
board level capability provided via the pci io registers.  

thanks

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Does anyone use VxWorks on a standard PC?
Date: Wed, 09 Aug 2000 09:40:35 +0530
From: Vijay Kumar Peshkar 
Organization: Ye 'Ol Disorganized NNTPCache groupie
Message-ID: <3990D9BA.FDE4F4B7@cisco.com>
References:  <3990ad22.10229036@news.swbell.net>
Reply-To: vpeshkar@cisco.com

This is a multi-part message in MIME format.
- --------------1FE500E1F3CA96BA044649BF
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

And you also have BSP for petiumpro!!!.

coder wrote:

> In comp.os.vxworks
> "Charles Magne"  wrote:
>
> >If you are using that,
> >I d like to know if it is necessary use an additionnal card?
> >and where to find the BSP associated.
>
> An additional card for what?  The BSP is PC486 and you find it in your
> target directory if you bought and installed it.

- --------------1FE500E1F3CA96BA044649BF
Content-Type: text/x-vcard; charset=us-ascii;
 name="vpeshkar.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Card for Vijay Kumar Peshkar
Content-Disposition: attachment;
 filename="vpeshkar.vcf"

begin:vcard 
n:Kumar.P;Vijay
tel;work:080-5530035/36 Extn:3464
x-mozilla-html:FALSE
org:Wipro Technologies, K2, Bangalore.;Wishing you blue skies, a couple of daisies, some fresh air and a smile!
adr:;;;;;;
version:2.1
email;internet:vpeshkar@cisco.com
title:Senior Software Engineer
note;quoted-printable:      o   o   o   o   o   o   o   .  .  . _____________________________________________=0D=0A    o           _____                      ||  Vijay Kumar .P         =20                                                    |=0D=0A   .] [__n_n_| DD[  =3D=3D=3D=3D_____   | vpeshkar@cisco.com, vijay_peshkar@hotmail.com     | =0D=0A >(________|__|_[_________]_ |___________________________________________|=0D=0A _/oo OOOOO oo`  ooo   ooo     'o!o!o                                 =20           
                     o!o!o`
fn:Vijay Kumar.P
end:vcard

- --------------1FE500E1F3CA96BA044649BF--


---------------------------

Newsgroups: comp.os.vxworks
Subject: Is there any on-line trainning course of VxWorks?
Date: Wed, 9 Aug 2000 08:57:28 +0900
From: "Jeongsu Byun" 
Organization: KREONet news service
Message-ID: <8mq6m5$2h6$1@usenet.kreonet.re.kr>

I am at beginner of VxWorks.
but, I have no time to subscribe offline tutrial,

Is there any online trainning course or tutorial course of
VxWorks?



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: MUX ReceiveRtn
Date: Wed, 09 Aug 2000 10:08:55 +0530
From: Yusuf Motiwala 
Organization: Lucent Technologies, Columbus, Ohio
Message-ID: <3990E05F.AD1540EC@lucent.com>
References: <8mp5bd$16c$1@nnrp1.deja.com>

You can do it without touching the source code. After loading the driver,
save the pointer to origional receive routine in END_OBJ and replace the
orig receiveRtn with your function with same prototype.

    origRcvRtn = pEnd->receiveRtn;
     pEnd->receiveRtn = (FUNCPTR) demultiplexing;

The origional routine then can be used from demultiplexing routine:

    origRcvRtn(pEnd, pMblk);

Save END_OBJ  if you want to use origional routine from somewhere else
also.

Regards,
Yusuf

Kans wrote:

> Hi,
> I was reading up on SENS.. and on the driver section it says,
> receiveRtn is actually a pointer in end_object and mux supplies the
> pointer after muxDevLoad(). Can I have my own demultiplexing routine
> instead of using muxReceiveRtn(). I have both the protocol and the
> driver sources, and I intend to use the Protocol Type field for my own
> purposes (not the rfc1700 assigned protocol number), that is why I
> intend to write my own demultiplexing routine .. Or can I still
> continue and use the muxReceieveRtn for my own purposes?.


---------------------------

Newsgroups: comp.os.vxworks
Subject: Same Error Status Strings........
Date: Wed, 09 Aug 2000 10:01:27 +0530
From: "Chandrashekar Nagaraj" 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <8mqpcj$pej$1@overload.lbl.gov>

Hi,
    As everyone know that each header file in vxWorks will have a set of Error Status Values defined in them, in the following manner

#define S_loadLib_FILE_READ_ERROR               (M_loadPecoffLib | 1)
#define S_loadLib_REALLOC_ERROR                 (M_loadPecoffLib | 2)
#define S_loadLib_JMPADDR_ERROR                 (M_loadPecoffLib | 3)
#define S_loadLib_FILE_ENDIAN_ERROR             (M_loadPecoffLib | 8)
#define S_loadLib_FILE_READ_ERROR               (M_loadCoffLib | 1)
#define S_loadLib_REALLOC_ERROR                 (M_loadCoffLib | 2)
#define S_loadLib_JMPADDR_ERROR                 (M_loadCoffLib | 3)
#define S_loadLib_FILE_ENDIAN_ERROR             (M_loadCoffLib | 8)

But if  u take a clear notice of these Error Status values in the loadPecoffLib.h and loadCoffLib.h we see that some of the Error Status are same( i mean
the error status strings). This does not cause any problem for program execution because the error status values are distinct for each of them but the compiler will give
warnings because of redundancy. Is there any way to remove these warnings,
we have tried changing the preprocessor's flag values, commenting it but none of
them works.

The program environment is something like this.

We create statTbl of vxWorks using makestatTbl.
Then we try to build our project files with statTbl.c being
one among the files. And we have /host/target/h as one of the
dependency.

Thank You,
Chandra.


---------------------------

Newsgroups: comp.os.vxworks
Subject: RE: Same Error Status Strings........
Date: Wed, 9 Aug 2000 09:46:09 +0100
From: David Anderson 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <8mr7ej$441$1@overload.lbl.gov>

On 09 August 2000 05:52, the vxWorks Users Group Exploder 
[SMTP:vxwexplo@lbl.gov] wrote:
> Submitted-by vxwexplo-errs@csg.lbl.gov  Tue Aug  8 21:32:54 2000
> Submitted-by: "Chandrashekar Nagaraj" 
>
> Hi,
>     As everyone know that each header file in vxWorks will have a set of 
Error Status Values defined in them, in the following manner
>
> #define S_loadLib_FILE_READ_ERROR               (M_loadPecoffLib | 1)
> #define S_loadLib_REALLOC_ERROR                 (M_loadPecoffLib | 2)
> #define S_loadLib_JMPADDR_ERROR                 (M_loadPecoffLib | 3)
> #define S_loadLib_FILE_ENDIAN_ERROR             (M_loadPecoffLib | 8)

how about adding #ifndef / #endif conditional compilation around both in 
both header files.
You would not need to change the compiler flags which tend to have a global 
affect on everything.

#ifndef S_loadLib_FILE_READ_ERROR
#define S_loadLib_FILE_READ_ERROR               (M_loadCoffLib | 1)
#endif

#ifndef S_loadLib_REALLOC_ERROR
#define S_loadLib_REALLOC_ERROR                 (M_loadCoffLib | 2)
#endif

#ifndef S_loadlib_JMPADDR_ERROR
#define S_loadLib_JMPADDR_ERROR                 (M_loadCoffLib | 3)
#endif

#ifndfe S_loadlib_FILE_ENDIAN_ERROR
#define S_loadLib_FILE_ENDIAN_ERROR             (M_loadCoffLib | 8)
#endif

>
> But if  u take a clear notice of these Error Status values in the 
loadPecoffLib.h and loadCoffLib.h we see that some of the Error Status are 
same( i mean
> the error status strings). This does not cause any problem for program 
execution because the error status values are distinct for each of them but 
the compiler will give
> warnings because of redundancy. Is there any way to remove these 
warnings,
> we have tried changing the preprocessor's flag values, commenting it but 
none of
> them works.
>

One thing that I've come across is that .h files compiled into .cpp source 
produce different
warnings than if they are compiled into .c source. This arises from the 
differences in the error
and warning checking. You do not say but I would not be surprised to find 
that a .c source did
not produce your warnings.


David Anderson.  mailto:danderson@dtrack.demon.co.uk

- -----------------------------------------------------
Who said "There's no life here in outer space."
- -----------------------------------------------------

Disclaimer
- ----------
This email and any files transmitted with it are confidential and intended 
solely for the use of the individual or entity to whom they are addressed.

If you have received this email in error please destroy and notify Data 
Track Technology Plc +44 1425 271900.
- -------------------------------------------------------



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: VxWorks FTP boot problem (T2, MVME2700)
Date: Wed, 09 Aug 2000 13:22:37 +0400
From: "Arthur V. Grigorian" 
Organization: Auriga, Inc.
Message-ID: <399122DD.25CBAA52@auriga.ru>
References: <8mdcif$tgl$1@nnrp1.deja.com> <398A95D7.A8BBCE60@auriga.ru> <8mls2f$keu$1@nnrp1.deja.com>

Hi again,

   Actually  I was talking  about  the following block:
***cut***
   if(bootFtp)
     (void) ftpCommand(errFd, "QUIT",0,0,0,0,0,0);
***cut***
   in  the  $(WIND_BASE)/target/config/all/bootCongig.c
   file, netLoad()  function.  If  you  change  it  for
   example to the following block:
***cut***
   if(bootFtp) {
     printf(">>>DEBUG>>> about to ftpCommand(QUIT)\n");
     (void) ftpCommand(errFd, "QUIT",0,0,0,0,0,0);
     printf(">>>DEBUG>>> finished ftpCommand(QUIT)\n");
   }
***cut***
   you'll see that the problem is somewhere inside this
   ftpCommand  call.  I think it's not hard  to replace
   this call by your own function which  won't wait for
   a reply from  the FTP server but  I didn't write it,
   (we decided to simply downgrade)  so it's up to you.

Regards,
__
R2

> Hi there,
> 
> Thanks very much for your reply. You are definitely right about the
> wait - I waited a lot longer and VxWorks did eventually boot. I can
> certainly downgrade our ftp server in the short-term. However, I am
> also concerned for the long-term. Both ftp servers give out the same
> messages.
> 
> I had a look at the code in bootconfig. I could not see an awful lot
> that I could fiddle with. I found the comment "Empty the Data Socket
> before close. PC FTP server hangs otherwise" suspicious, so I removed
> the following line:
> 
>         while ((read (fd, command, sizeof (command))) > 0);
> 
> But this made the system not able to boot from anything.
> Hmmm....that makes me even more suspicious but not at all sure what to
> try next.
> 
> Paul Roberts.

---------------------------

End of New-News digest
**********************


From vxwexplo-errs@csg.lbl.gov  Thu Aug 10 04:03:09 2000
From: Vxworks Exploder 
Date: Thu Aug 10 04:03:12 PDT 2000
Subject: comp.os.vxworks newsdigest

Comp.Os.Vxworks Daily Digest    Thu Aug 10 04:03:04 PDT 2000

        Subject: Re: dec21x4xEnd driver
        Subject: Re: dec21x4xEnd driver
        Subject: select() problem
        Subject: Re: Is there any on-line trainning course of VxWorks?
        Subject: c++ comments
        Subject: Re: help for LoadModule
        Subject: Re: c++ comments
        Subject: Re: c++ comments
        Subject: writing device drivers
        Subject: mv2700 duplicate int! why?
        Subject: Re: mv2700 duplicate int! why?
        Subject: Display of Windows NT Windows on a VXWorks Power PC
        Subject: Re: NTP client on VxWorks
        Subject: Socket buffer limit?
        Subject: Re: c++ comments
        Subject: Re: help for LoadModule
        Subject: Re: c++ comments
        Subject: Re: Socket buffer limit?
        Subject: Re: NTP client on VxWorks
        Subject: tcp retransmit delays
        Subject: Re: NTP client on VxWorks
        Subject: Re: c++ comments
        Subject: Re: dec21x4xEnd driver
        Subject: Re: mv2700 duplicate int! why?
        Subject: Condition variables in VxWorks?
        Subject: MVME2604 Memory maps
        Subject: MVME2604 Memory maps
        Subject: Network File system for vxWorks other than NFS
        Subject: Re: Tornado Launch
        Subject: Testing END Driver Mulitcast
        Subject: Re: NTP client on VxWorks
        Subject: Re: how does purify work?
        Subject: Debugging variables defined in nested blocks
        Subject: Re: tcp retransmit delays
        Subject: Re: Debugging variables defined in nested blocks
        Subject: Replacement for Altera ByteBlaster & ByteBlasterMV
        Subject: WRS white paper on vxWorks timing
        Subject: Tornado 2: How do I redirct the Virtual I/O Output???
        Subject: Access Network from VxSim ?
        Subject: Re: NTP client on VxWorks
        Subject: Re: Tornado 2: How do I redirct the Virtual I/O Output???

-------------------------------------------------------

Newsgroups: comp.os.vxworks
Subject: Re: dec21x4xEnd driver
Date: Wed, 9 Aug 2000 13:06:48 +0200
From: Jan Koeppen 
Organization: Robert Bosch GmbH
Message-ID: <8mre08$t92$2@proxy2.fe.internet.bosch.de>
References: <8mn55b$j7p$1@nnrp1.deja.com>

We had the problem that this driver seems not to be working on a 100MBit Hub
:-((.
We solved it by buying old 10Mbit hubs that work fine (but is not
satisfactory.)

regards
Jan

stourangeau@my-deja.com schrieb in Nachricht
<8mn55b$j7p$1@nnrp1.deja.com>...
>I am having problems setting up the dec21x4xEnd driver. I dont have the
>source file so I am only playing with the configuration of the driver.
>
>After lots of work, I was able to make the driver attach to the mux and
>attach protocols to it.
>
>I had lots of problem with getting the link to be indicated on the
>hub... I had to play around with a mediaSelect routine..
>
>Problem is, the hub sees the link but the dec21x doesnt seem to see the
>hub. By looking at the status registers, I see link failures and
>autonegociation that is not completed...
>
>Need some help please...
>
>
>Sent via Deja.com http://www.deja.com/
>Before you buy.



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: dec21x4xEnd driver
Date: Wed, 9 Aug 2000 13:08:12 +0200
From: Jan Koeppen 
Organization: Robert Bosch GmbH
Message-ID: <8mre2t$77$2@proxy2.fe.internet.bosch.de>
References: <8mn55b$j7p$1@nnrp1.deja.com>

We had the problem that this driver seems not to be working on a 100MBit Hub
:-((.
We solved it by buying old 10Mbit hubs that work fine (but is not
satisfactory.)

regards
Jan

stourangeau@my-deja.com schrieb in Nachricht
<8mn55b$j7p$1@nnrp1.deja.com>...
>I am having problems setting up the dec21x4xEnd driver. I dont have the
>source file so I am only playing with the configuration of the driver.
>
>After lots of work, I was able to make the driver attach to the mux and
>attach protocols to it.
>
>I had lots of problem with getting the link to be indicated on the
>hub... I had to play around with a mediaSelect routine..
>
>Problem is, the hub sees the link but the dec21x doesnt seem to see the
>hub. By looking at the status registers, I see link failures and
>autonegociation that is not completed...
>
>Need some help please...
>
>
>Sent via Deja.com http://www.deja.com/
>Before you buy.



---------------------------

Newsgroups: comp.os.vxworks
Subject: select() problem
Date: Thu, 10 Aug 2000 02:04:02 +0200
From: "Øyvind Holmeide" 
Organization: Telenor Online Public Access
Message-ID: 

I'm using the select() routine in an application that uses the console
(standard in) either via the serial line or telnet.  The select routine
returns when a byte is available.  However, the corresponding call to getc()
will not see the first byte (only the following bytes).  "Syntax error" is
also written to the console after the select() returns with 1.  The code
look like
this:

 timeout.tv_sec = 1;
 timeout.tv_usec = 0;

 while (!end) {
    newInput = 0;
    while (newInput < 1) {
    FD_ZERO (&readFds);
    FD_SET (STD_IN, &readFds);
    newInput = select (FD_SETSIZE, &readFds, NULL, NULL, &timeout);
}
  i = getc(stdin);

- --

Øyvind Holmeide






---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Is there any on-line trainning course of VxWorks?
Date: 9 Aug 2000 12:13:48 GMT
From: johan@borksoft.xs4all.nl
Organization: Borksoft
Message-ID: <8mrhts$hfk$1@news1.xs4all.nl>
References: <8mq6m5$2h6$1@usenet.kreonet.re.kr>

Jeongsu Byun  wrote:
: I am at beginner of VxWorks.
: but, I have no time to subscribe offline tutrial,

: Is there any online trainning course or tutorial course of
: VxWorks?

Take a look at :
http://www.smartmodulartech.com/systems/library/docs/vxworks/vxintro/VXINTRO.HTM

Groeten,
	Johan

- -- 
     o o o o o o o . . .   ___________________________________
    o      _____           ||    Johan Borkhuis              |
  .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
 >(________|__|_[_________]_|________________________________|
 _/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: c++ comments
Date: Wed, 09 Aug 2000 08:49:28 -0400
From: Paul Oliveira 
Organization: A thing of the past
Message-ID: <39915358.F4E8D94F@yahoo.com>

Hi

We are switching to T2 and from solaris to NT. I got the lovely job of
doing the port.  When I compile using ccmips on solaris (outside of T1),
no problems (because somewhere in my environment the problem is taken
care of).  When I compile on NT with T2 I get compile errors on the //
(c++ comment) lines.  What the heck am I missing to solve this.

Thanks
Paul


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: help for LoadModule
Date: Wed, 09 Aug 2000 15:21:47 +0200
From: Uwe Fuhr 
Organization: Ericsson Eurolab Deutschland GmbH
Message-ID: <39915AEB.EF71C6EC@eede.ericsson.se>
References: <8mda0f$29di$1@news.cz.js.cn> <8mdsgb$8vn$1@nnrp1.deja.com>
Reply-To: Uwe.Fuhr@eed.ericsson.se

Hi Urban,

I run into the same problem as Tang, while extending the RAM from 32
to 64 MB of RAM, and then loading objects via ld() to the new 64 MB vxWorks
load module.

We are using the gcc compiler but I could not find the mentioned
option -mlongcall in the manual.

What about this.

Best regards

/Uwe

- -- 
___________________________________________________________________________

     _/  _/  _/  Uwe Fuhr                       EED/E/D/U    Access Design    
    _/  _/  _/   Ericsson Eurolab Deutschland   phone  : +49-5121 - 707104
   _/  _/  _/    Daimlerring 9                  FAX    : +49-5121 - 707314
  _/  _/  _/     31135 Hildesheim - Germany     mobile : +49-173 - 5430771  

__________________________________   mailto:Uwe.Fuhr@eed.ericsson.se   ____
                                  \___________________________________/



> Hi!
> 
> You need to add the -mlongcall option to your compilation.
> 
> HTH
> Urban
> 
> In article <8mda0f$29di$1@news.cz.js.cn>,
>   "tanghaoyu"  wrote:
> > Hello All,
> >
> > When I load an object file by calling LoadModule() or ld (1)
> >  >
> > Relocation value does not fit in 24 bits.
> > Relocation value does not fit in 24 bits
> > ......

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: c++ comments
Date: Wed, 09 Aug 2000 14:23:35 +0100
From: Hugh Duggan 
Organization: Lucent Technologies, Chippenham, United Kingdom
Message-ID: <39915B57.E30E6F2F@lucent.com>
References: <39915358.F4E8D94F@yahoo.com>

Paul Oliveira wrote:
> 
> Hi
> 
> We are switching to T2 and from solaris to NT. I got the lovely job of
> doing the port.  When I compile using ccmips on solaris (outside of T1),
> no problems (because somewhere in my environment the problem is taken
> care of).  When I compile on NT with T2 I get compile errors on the //
> (c++ comment) lines.  What the heck am I missing to solve this.
> 
> Thanks
> Paul

Sounds to me like you are using a .c extension and an ANSI
compiler flag.  //-style comments are illegal in ANSI C, but many
modern compilers will allow them unless told to be strict ANSI.

Suggested solutions: either use a .cc extension, or remove the -ansi ?


    Hugh
 
- -----------------------------------------------------------
Hugh Duggan                 | Phone   : +44 1666 832118
Lucent Technologies         | FAX     : +44 1666 832925
Building 7,                 |
Greenways Business Park,    | E-mail  : hduggan@lucent.com
Bellinger Road,             |
Chippenham,                 |
UNITED KINGDOM              |

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: c++ comments
Date: Wed, 09 Aug 2000 08:39:59 -0500
From: Greg Willden 
Organization: Posted via Supernews, http://www.supernews.com
Message-ID: <39915F2F.8B7184CE@swri.edu>
References: <39915358.F4E8D94F@yahoo.com>

Paul Oliveira wrote:

> Hi
>
> We are switching to T2 and from solaris to NT. I got the lovely job of
> doing the port.  When I compile using ccmips on solaris (outside of T1),
> no problems (because somewhere in my environment the problem is taken
> care of).  When I compile on NT with T2 I get compile errors on the //
> (c++ comment) lines.  What the heck am I missing to solve this.
>
> Thanks
> Paul

I am assuming that you are compiling a file with a .c extension.  There is a
switch in the compiler settings that says -ansi.  If you remove it the
compiler will allow the C++ comments in a C file.



---------------------------

Newsgroups: comp.os.vxworks
Subject: writing device drivers
Date: Wed, 9 Aug 2000 10:06:24 +0200
From: "Ramakrishnan,CN" 
Organization: UUNET-NL (http://www.nl.uu.net)
Message-ID: 
References: <8m95or$720$1@news.news.kodak.com>

Hai,
Thank you for all your mails. I am attaching my write up along with
this. It is a word document.
Regards
Ramakrishnan.C.N
Peek Traffic B.V.
Basicweg 16,
3821 BR AMERSFOORT 
P.O. Box 2542,
3800 GB AMERSFOORT
The Netherlands 
telephone:  +31 33 454 1777
fax: 	     +31 33 454 1850
Internet: www.peektraffic.nl 
	
e-mail: C.N.Ramakrishnan@peektraffic.nl


>  <> 

begin 600 driverguide.doc
MT,\1X*&Q&N$`````````````````````/@`#`/[_"0`&```````````````!
M`````@``````````$```+@````$```#^____``````,```#_____________
M____________________________________________________________
M____________________________________________________________
M____________________________________________________________
M____________________________________________________________
M____________________________________________________________
M____________________________________________________________
M____________________________________________________________
M____________________________________________________________
M____________________________________________________________
M______________________]2`&\`;P!T`"``10!N`'0`<@!Y````````````
M````````````````````````````````````````````````%@`%`?______
M____`0`````)`@``````P````````$8`````0%E<6HCMOP$@A60MV`'``04`
M``#``P```````%<`;P!R`&0`1`!O`&,`=0!M`&4`;@!T````````````````
M```````````````````````````````````````:``(!`@````,```#_____
M````````````````````````````````````````````````(````$I@````
M`````0!#`&\`;0!P`$\`8@!J````````````````````````````````````
M`````````````````````````````!(``@'_______________\`````````
M````````````````````````````````````````````:@`````````%`%,`
M=0!M`&T`80!R`'D`20!N`&8`;P!R`&T`80!T`&D`;P!N````````````````
M````````````````````*``"`?____\$````_____P``````````````````
M``````````````````````````````(```#D`0```````#````#]________
M_______^____!````/__________"0````H````+````#`````T````.````
M#P```!`````1````$@```!,````4````%0```!8````7````&````!D````:
M````&P```!P````=````'@```!\````R````(0```"(````C````)````"4`
M```O````___________________________________________^____,0``
M`/[___\(````,P```#0````U````10```/___________O______________
M_____________________________________________________T8```!'
M````2````$D```!*````2P```$P```!-````3@```$\```!0````.````/__
M____________________________________________________________
M____________________________________________________________
M____________________________________________________________
M____________________________________________________________
M________4@!O`&\`=``@`$4`;@!T`'(`>0``````````````````````````
M`````````````````````````````````!8`!0'__________P$`````"0(`
M`````,````````!&`````$!97%J([;\!((5D+=@!P`$%````P`,```````!7
M`&\`<@!D`$0`;P!C`'4`;0!E`&X`=```````````````````````````````
M````````````````````````&@`"`0(````#````_____P``````````````
M``````````````````````````````````8```!*8`````````$`0P!O`&T`
M<`!/`&(`:@``````````````````````````````````````````````````
M```````````````2``(!________________````````````````````````
M`````````````````````````````&H`````````!0!3`'4`;0!M`&$`<@!Y
M`$D`;@!F`&\`<@!M`&$`=`!I`&\`;@``````````````````````````````
M`````"@``@'_____!````/____\`````````````````````````````````
M```````````````"````Y`$```````#__________S````#]_____O___P0`
M```'````)@````D````*````"P````P````-````#@````\````0````$0``
M`!(````3````%````!4````6````%P```!@````9````&@```!L````<````
M'0```!X````?````+````/_______________________________R<````H
M````*0```"H````K````"````"T````V````_O_________^____________
M__________________\W````.0```/[___\Z````.P```#P````]````/@``
M`#\```!`````00```$(```!#````1````#@```#_____________________
M____________________________________________________________
M____________________________________________________________
M____________________________________________________________
M____________________________________________________________
M_____________________________________________________W)D(&9O
M````$@```%!%14L@5')A9F9I
M8R!"+E8N`````P````"D```#````;@````,````:````"P`````````+````
M``````P0```"````'@```"X```!7SV3````"T`0``$@````$```"8`````@```*`````#````V`````0```#D
M````!0```/0````&``````$```<````,`0``"````"`!```)````,`$``!(`
M```\`0``"@```&0!```+````<`$```P```!\`0``#0```(@!```.````E`$`
M``\```"<`0``$````*0!```3````K`$```(```#D!```'@```"X```!7
M`````0`````````>````!0```%!40E8``&<`'@````$`````0D0`'@````$`
M````````'@````L```!.;W)M86PN9&]T```>````!0```%!40E8``$``'@``
M``,````Q.0``'@```!X```!-:6-R;W-O9G0@5V_E3```L````%50``$H```#54P```````-53````````U5,```````#54P``
M`````&19````````K5D```````"&60``)P```(99````````````````````
M````````````````````````````````````````````````````````````
M``````````````````````````````````````````!7$%R9R!A;F0@
M<'5T4F-V07)G(&%R92!A&%M<&QE
M(&EN=&5R2P@;6]D92!O9B!O<&5R871I;VX@*"!03TQ,140@;W(@24Y415)255!4*2!E
M=&,N+B!"=70@:70@:7,@8F5T=&5R('1O(&QO;VL@:6YT;R!T:&4@2XN*2X@5&AI6]U6]U(&-A;B!D969I;F4@86YO=&AE7!E
M9&5F('-T2!A;'-O(&)E(&%D=F%N=&%G96]U'1E'1E'0@2!I="!S96QF(&EN:71I86QI
M7-3
M97)I86Q(=TEN:70H*2!A;F0@7-(=TEN
M:70H*2!A;F0@2!C'0@9G5N8W1I;VX@:7,@=&5M
M<&QA=&5$979);FET,B!A;F0@:7,@8V%L;&5D(&9R;VT@2!A('-E2!D97!E;F1E;G0@;VX@=&AE(&1E=FEC92!A;F0@:&5N8V4@8V%N
M;F]T(&)E(&1I&%C="!I;7!L96UE;G1A=&EO;B!T:&4@:6YT97)R=7!T
M(&AA;F1L:6YG+B`-268@=&AE('-E2!I;7!R;W9E('1H92!E9F9I8VEE;F-Y+@U4:&5R92!A%-T87)T=7`@9G5N8W1I;VX@86-T=6%L;'D@2!M96%N
M'1E2!A;'-O(&)E('5S969U;"!T;R!S96YD(&UO
M2!O;F4@8VAA2!T:&4@54%25"`H(&EN(&-A6]U('5S92!O;F4I+"!I="!S96YD
M'0@-#D@8VAA&-E
M<'0@9F]R('1H92!C:&%N9V4@;V8@9G5N8W1I;VX@;F%M92X-#4UO2!F;W(@82!P2!F=6YC=&EO;G,N($]T:&5R=VES
M92!P;VQL26YP=70@9G5N8W1I;VX@=VEL;"!S8V%N('1H92!A<'!R;W!R:6%T
M92!R96=I2!B
M>2!T:&4@8V%L;&EN9R!F=6YC=&EO;BP@8GD@;65A;G,@;V8@=&AE(')E='5R
M;B!V86QU97,L('=H971H97(@:70@:&%S('-E;F0@82!C:&%R86-T97(L('=H
M971H97(@=&AE2!D979I8V4@97)R;W(@;W(@:68@=&AE(&]U
M='!U="!B=69F97(@:7,@9G5L;"X-1F]R('-O;64@<')O8V5S&-E<'0@
M9F]R(&EN=&5R2!I='-E;&8N(`U)9B!T:&4@9')I=F5R('-U<'!O6YA;6EC
M86QL>2P@=&AE(&EN9F]R;6%T:6]N(&]F('1H92!M;V1E(&]F(&]P97)A=&EO
M;B!O9B!T:&4@9&5V:6-E('-H;W5L9"!B92!K97!T(&EN('1H92!414U03$%4
M15]#2$%.('-T2!TFEN9R!O;F4@8VAA;FYE;"!A;F0@9&\@=&AE('1E2!B
M92!A(&=O;V0@97AE2!W2!B92!G;V]D
M('1O(&-R96%T92!T:&4@9&5V:6-E+"!O<&5N(&ET(&%N9"!S96YD(&$@8VAA
M2!C;VUP;&5T960N("!)9B!Y;W4@:&%V92!M;W)E('1H86X@;VYE
M(&-H86YN96PL(&ET(&ES(&%LF5R('-H;W5L9"!B92!D;VYE(&%T(&QA2!I2!T:&4@;W!E
M2!L:71T
M;&4@86)O=70@:70N($)U="!A(&)A9"!L871E;F-Y('-H;W=S('1H870@=&AE
MF5D
M('1O(&EM<')O=F4@<&5R9F]R;6%N8V4N(%=I;F12:79E````N5P`
M`%@````170``20```'!:```8`@```````````````````````+I2````````
M*U@``````````"$`(@`!``8`*U@````````K6```````````````````````
M```````K6````````"M8````````<%H```````!!6````````+I2````````
MNE(````````K6``````````````````````````````66```%0```$%8````
M````05@```````!!6````````"M8```6````NE(````````K6````````+I2
M````````*U@```````!H6````````````````````,PGMU8`P`'.4@``+```
M`/I2``!*````NE(```````"Z4@```````+I2````````NE(````````K6```
M`````&A8````````05@``"<```!!6```````````````````````````````
M````````````````````````````````````````````````````````````
M````````````````````````````````5W)I=&EN9R!A('-E3H@4F%M86MR
M:7-H;F%N+D,N3B`@("`@("`@("`@("`@("`@("`@("`@("`@("`@("`@("`@
M("`@("`@("`@("`@("`@("`@("`@("`@("`@("`@("`@("`@("`@("`@("`@
M("!$871E.B`R,"TP-RTR,#`P#0U4:&ES(&QI=&5R871U2X@2&]W
M979E2X@66]U(&-A;B!M86ME('EO=7(@
M9FEL97,@8GD@8V]P>6EN9R!T:&5M(&%N9"!N86UI;F<@=&AE;2!F;W(@>6]U
M2!3
M24\@9&5V:6-E(&ES(&-O;G1R;VQL960@8GD@86X@4TE/7T-(04X@2!F;VQL;W=E9"X@2&]W979E&-L=7-I=F4@9W5I9&4N
M(%1H:7,@=W)I=&4@=7`@9&]E2!I;F%D97%U871E+B!"=70@
M<&5R:&%P6]U2!C;W!Y:6YG('1H96T@
M86YD(&YA;6EN9R!T:&5M(&9O7-I2!A;B!324]?0TA!3B!S=')U8W1U2!E>'!A;F1I
M;F<@=&AE(%-)3U]#2$%.('-T6]U(&QO
M;VL@870@=&AE(%1%35!,051%7T-(04X@%=O2!S
M=6=G97-T:6]N6]U(&%L;"!T:&4@8F5S=(4N
M#0T-#0T-#0DN(`T-%0"D@BZEQD&F"`>G"`>HH`6IH`6J``!);B!T:&4@9&5C
M;&%R871I;VYS+"!A#0UA;'-O(`T-#0T-#2`5`*2"+J7&0:8(!Z<(!ZB@!:F@
M!:H``&DN92YI;FET:6%L:7IA=&EO;FEN:71I86QI>F5S:6YI=&EA;&EZ871I
M;VYI;FET:6%L:7IE9````````P``+@,``+`#``"9!@``K`8``-T'``#>!P``
MX`<``/('``#S!P``]0<``-0)``#5"0``V`D``.P)``#N"0``[PD``!4*```6
M"@``.PH``$,*``")"@``E@H``*(*``"C"@``7!D``&P9``#E-0``]C4``'1!
M``"+00``H4$``*-!``"H00``J4$``*I!``"K00``K4$``*Y!``"O00``QD$`
M`,I!``#800``XT$``/%!``#\00```%(```I2```+4@``NE(``+M2``#%4@``
MQE(``,=2```84P``&5,``!M3``"(4P``J%,``*E3``"]4P``OE,``-53``#^
M`/KU^@#S]?KS^@#S]?H`^@#Z]?KU^@#Z]?KU^O'Z^OKZ^O7Z^OKQ^OKZ^OKQ
M^OKZ^O7Z^OKZ^OKZ^OKZ\0``````````````````````````````````````
M````````````````````````````````````````````````````````````
M````````````````````````````````````````````````````````````
M````````````````````````````````````````````````````````````
M``````)U`0`#8Q@`"%6!70(`8Q@```9=`@!C&````EX!/@$```#^____`P``
M``0````%````!@````<````(````"0```/[___\+````#`````T````.````
M_O__________________________________________________________
M____________________________________________________________
M____________________________________________________________
M____________________________________________________________
M____________________________________________________________
M____________________________________________________________
M____________________________________________________________
M____________________________________________________________
M____________________________________________________________
M____________________________________________________________
M________:&5I7-T96T@86QW87ES(&=E
M=',@=&AE('!O:6YT97(@=&\@=&AE(&1E=FEC92!D%=O2!S=6=G97-T:6]NG"`>HH`6IH`6J
M``!);B!T:&4@9&5C;&%R871I;VYS+"!A#0UA;'-O(`T-#0T-#2`5`*2"+J7&
M0:8(!Z<(!ZB@!:F@!:H``&DN92YI;FET:6%L:7IA=&EO;FEN:71I86QI>F5S
M:6YI=&EA;&EZ871I;VYI;FET:6%L:7IE9````````P``+@,``+`#``"9!@``
MK`8``-T'``#>!P``X`<``/('``#S!P``]0<``-0)``#5"0``V`D``.P)``#N
M"0``[PD``!4*```6"@``.PH``$,*``")"@``E@H``*(*``"C"@``7!D``&P9
M``#E-0``]C4``'1!``"+00``H4$``*-!``"H00``J4$``*I!``"K00``K4$`
M`*Y!``"O00``QD$``,I!``#800``XT$``/%!``#\00```%(```I2```+4@``
MNE(``/X`^O7Z`//U^O/Z`//U^@#Z`/KU^O7Z`/KU^O7Z\?KZ^OKZ]?KZ^O'Z
M^OKZ^O'Z^OH`````````````````````````````````````````````````
M````````````````````````````````````````````````````````````
M````````````````````````````````````````````````````````````
M````````````````````````````````````````````````````````````
M````````````````````````````````````````````````````````````
M``````````````````````)U`0`#8Q@`"%6!70(`8Q@```9=`@!C&````EX!
M,0`#```N`P``+P,``*\#``"P`P```P0``)@&``"9!@``K08``-X'``#?!P``
M]`<``/4'``#5"0``U@D``.X)``#Y``%R(%`!]P`!#```8@X``.$.``#Y#@``$0\``",/```V#P``-P\``)00``"5$```
MX1```!H2```O$@``/!(``+L2``!`$P``01,``%$3``!Q$P``B1,``*`3``"W
M$P``N!,``,P3``#-$P``$!0``"H4```K%```+!0``&44``!F%```H!0``-H`
M`7(@(`':``%R("`!V@`"!P``WP<``/0'``#U!P``U0D`
M`-8)``#N"0``^0`!&-E;&QE;G0@;6%T97)I
M86PN(%1H92!P6]U('=I;&P@87!P6]U1$``+P1``#6$0``UQ$``-@1```1$@``
M$A(``$P2``".$@``QA(``,<2``#S%```]10``'(6``!S%@``!A<```<7```9
M%P``&A<``!P8```=&```>QD``'P9```5&@``%AH``.D;``#J&P``%!X``),?
M```2(@``$R(``!0F``"1)@``DB8``-2!$;V-U
M;65N='-<9')I=F5R9W5I9&4N9&]C_T!(4"!,87-E0L``'T+```>$P``
M+!,``%\3``!J$P``Y1,``/,3```:%P``,!<``,(8``#-&```FAD``*09``"E
M,P``IS,``#,Y```X.0``5#D``%XY``!?.0``8#D``.,[``#D.P``YCL``.P[
M``#M.P``[CL``$(\``!#/```]3T``/8]````/@```3X```(^``!3/@``5#X`
M`%8^``##/@``XSX``.0^``#X/@``^3X``/H^``#[/@``_#X```4_```&/P``
M00```P````!``"T#```*`4``+@,```H#00`O`P```0!``($#```!`$``KP,`
M``H#00`+4@````!``-T'```*`T$`QD$`````0`#1#0``"@-!`,I!`````$$`
M@!4```H#00#800````!!`+X5```*`T$`XT$`````0`!'%@``"@-!`(M!````
M`$``;QD```H#00#Q00````!```P;```*`T$``%(`````0`#C&P``"@-``*%!
M`````$``Y#4```H#00"C00````!``'`[```*`T$`C#L```$`0`"H00````!`
M`)8[```#`$$`ESL```H#0`"I00````!``!H^```*`T$`'#X```$`0`"J00``
M``!``"(^```%`$$`(SX```H#00`*4@````!``'<^```*`T``NE(`````00"[
M4@````!``,52`````$``QE(`````00#'4@````!!`!A3`````$$`&5,`````
M00`;4P````!!`(A3`````$$`J%,`````00"I4P````!!`+U3`````$``JT$`
M``<`0`"L00````!``"E````*`T``:D$```H#0`!S00``"@,Q`!46D`$``%1I
M;65S($YE=R!2;VUA;@`,%I`!`@!3>6UB;VP`"R:0`0``07)I86P`(@`$`'$(
MB!@``-`"``!H`0````!@F@DI(9@`````3`(&``````\`90$```T)``"7,P``!P`:````!`"#$&X`````````````
4``<``0````$`````````(0,````=
`
end


---------------------------

Newsgroups: comp.os.vxworks
Subject: mv2700 duplicate int! why?
Date: Wed, 09 Aug 2000 12:37:05 -0400
From: Al Johnston 
Organization: MindSpring Enterprises
Message-ID: <399188B1.C9AA0763@mindspring.com>

In the Prog Ref for the 2700 a group of interrupt sources are shown
as being intputs to the MPIC (p/o RAVEN) and the PIB (p/o w83c553f).

In this group are:
  ethernet
  scsi
  VME INT0/LINT0#
  LM/SIG Int0
  LM/SIG Int1

The doc says that these are on the PIB for "backward compatibility".

Ignoring the compatibility rational, does anyone understand what
it is they are doing here?  On it face it sound like the interrupt
source is being connected in two places... But that would mean the
int sig would go to the mpic TWICE, once via the "all ISA line"
(IRQ0) and also from the individual line (one of IRQ2, IRQ3, IRQ5,
IRQ9, IRQ13, IRQ14).  This really seems like a strange way to set
thing up.

Anybody know what is actually being done here?

thanks

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: mv2700 duplicate int! why?
Date: Wed, 09 Aug 2000 12:58:11 -0500
From: Andrew Johnson 
Organization: APS, Argonne National Laboratory
Message-ID: <39919BB3.8D0848AB@aps.anl.gov>
References: <399188B1.C9AA0763@mindspring.com>

Hi Al,

Al Johnston wrote:
> 
...
> The doc says that these are on the PIB for "backward compatibility".
> 
> Ignoring the compatibility rational, does anyone understand what
> it is they are doing here?  On it face it sound like the interrupt
> source is being connected in two places... But that would mean the
> int sig would go to the mpic TWICE, once via the "all ISA line"
> (IRQ0) and also from the individual line (one of IRQ2, IRQ3, IRQ5,
> IRQ9, IRQ13, IRQ14).  This really seems like a strange way to set
> thing up.

I think your "ignoring the compatibility rational"  is the key point - if
you look at the BSP, from memory I think it either initializes one chip or
the other depending on a compile-time switch somewhere.  I assume that the
PIB hardware is only there for older software (maybe it's needed for CHRP
compatibility?).

- - Andrew
- -- 
Sharks kill 10 people each year.  People kill 60,000,000 sharks each year.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Display of Windows NT Windows on a VXWorks Power PC
Date: Wed, 09 Aug 2000 18:32:38 GMT
From: lpeck@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8ms846$380$1@nnrp1.deja.com>

I have a configuration with a dumb  Power PC
display server running VXworks that supports X
Displays. Currently we are able to display
windows from a Unix window on this server by
simply setting the display variable to the
display server. However, we would like to replace
the Unix machine with a PC running Windows NT. Is
there an application or a setting which will
allow me to run my existing Windows NT apps on my
Windows NT and display the windows for the app on
the dumb display server?


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: NTP client on VxWorks
Date: Wed, 09 Aug 2000 08:50:57 -0400
From: "Bryan D. Lee" 
Organization: Sanders-A Lockheed Martin Company
Message-ID: <399153B1.C906C5EE@lmco.com>
References: <311c2be6.cfb476ec@usw-ex0106-045.remarq.com>

Yes and no,

I think if you started from the NTP v3 sources, you'd be OK.  If you
want v4 (current release) then either go at it yourself, or wait till
I'm done.  I'm trying to port v4 now.  If you want to go at it
yourself, you can start with the vxworks port information in the NTP
distribution, but there are a significant number of small changes that
have to be made.  Additionally, code has to be added to the kernel to
support NTP; specifically a PLL/FLL as shown in the kernel and
nanokernel tar files on UDel.

For more information, try comp.protocols.time.ntp,
http://www.eecis.udel.edu/~ntp/, and/or the 'hackers' and 'bugs' mail
lists at majordomo@ntp.org

If you're serious about starting immediately, then I'll send you a
list of changes to make from the 4.0.99j distribution but it's still a
work-in-progress.

Bryan


"Warren.Brown" wrote:
> 
> Hello
> Has anyone added the NTP (Network Time Protocol) client
> services to a VxWorks embedded target ?
> Any hints as to the right steps to take ?
> 
> W
> 
> -----------------------------------------------------------
> 
> Got questions?  Get answers over the phone at Keen.com.
> Up to 100 minutes free!
> http://www.keen.com

- -- 
Bryan D. Lee                             Phone: 603/885-7025
Software Engineer & ELDP                 Fax  : 603/885-3016
Sanders-A Lockheed Martin Company        Pager: 603/423-5565
PO Box 868, PTP1-2361, Nashua, NH 03061  Email: bryan.d.lee@lmco.com

---------------------------

Newsgroups: comp.os.vxworks
Subject: Socket buffer limit?
Date: Wed, 09 Aug 2000 11:10:06 -0400
From: Steve Rottinger 
Organization: Pentek, Inc.
Message-ID: <3991744D.82206FCE@pentek.com>

Hi,

Does anyone know if there is a hard limit to the socket buffer size that
you
can specify under vxWorks.  I'm using setsockopt() to adjust the
SO_SNDBUF,
and SO_RCVBUF.   Things seem to work fine with a 64K buffer, but a 128K
buffer results in a crash.

Thanks in advance,

- -Steve



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: c++ comments
Date: Wed, 09 Aug 2000 11:00:55 -0400
From: Paul Oliveira 
Organization: A thing of the past
Message-ID: <39917227.C889147F@yahoo.com>
References: <39915358.F4E8D94F@yahoo.com> <39915B57.E30E6F2F@lucent.com>

Thankx Hugh and Greg

Removing the -ansi worked

Paul

Hugh Duggan wrote:

> Paul Oliveira wrote:
> >
> > Hi
> >
> > We are switching to T2 and from solaris to NT. I got the lovely job of
> > doing the port.  When I compile using ccmips on solaris (outside of T1),
> > no problems (because somewhere in my environment the problem is taken
> > care of).  When I compile on NT with T2 I get compile errors on the //
> > (c++ comment) lines.  What the heck am I missing to solve this.
> >
> > Thanks
> > Paul
>
> Sounds to me like you are using a .c extension and an ANSI
> compiler flag.  //-style comments are illegal in ANSI C, but many
> modern compilers will allow them unless told to be strict ANSI.
>
> Suggested solutions: either use a .cc extension, or remove the -ansi ?
>
>     Hugh
>
> -----------------------------------------------------------
> Hugh Duggan                 | Phone   : +44 1666 832118
> Lucent Technologies         | FAX     : +44 1666 832925
> Building 7,                 |
> Greenways Business Park,    | E-mail  : hduggan@lucent.com
> Bellinger Road,             |
> Chippenham,                 |
> UNITED KINGDOM              |


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: help for LoadModule
Date: Wed, 09 Aug 2000 15:22:40 GMT
From: Urban Lindberg 
Organization: Deja.com - Before you buy.
Message-ID: <8mrsvn$ptp$1@nnrp1.deja.com>
References: <8mda0f$29di$1@news.cz.js.cn> <8mdsgb$8vn$1@nnrp1.deja.com> <39915AEB.EF71C6EC@eede.ericsson.se>

Hi Uwe!

I don't use the printed manuals, but the online manuals and there you
will find it in: GNU Toolkit User's Guide/The C/C++ Compiler/IBM RS/6000
and PowerPC Options. It's the last option mentioned in that section.
And, yes the error message is erroneous, it should be 26 bits :-)

HTH
Urban

In article <39915AEB.EF71C6EC@eede.ericsson.se>,
  Uwe.Fuhr@eed.ericsson.se wrote:
> Hi Urban,
>
> I run into the same problem as Tang, while extending the RAM from 32
> to 64 MB of RAM, and then loading objects via ld() to the new 64 MB
vxWorks
> load module.
>
> We are using the gcc compiler but I could not find the mentioned
> option -mlongcall in the manual.
>
> What about this.
>
> Best regards
>
> /Uwe
>
> --
>
___________________________________________________________________________
>
>      _/  _/  _/  Uwe Fuhr                       EED/E/D/U    Access
Design
>     _/  _/  _/   Ericsson Eurolab Deutschland   phone  : +49-5121 -
707104
>    _/  _/  _/    Daimlerring 9                  FAX    : +49-5121 -
707314
>   _/  _/  _/     31135 Hildesheim - Germany     mobile : +49-173 -
5430771
>
> __________________________________   mailto:Uwe.Fuhr@eed.ericsson.se
____
>
\___________________________________/
>
> > Hi!
> >
> > You need to add the -mlongcall option to your compilation.
> >
> > HTH
> > Urban
> >
> > In article <8mda0f$29di$1@news.cz.js.cn>,
> >   "tanghaoyu"  wrote:
> > > Hello All,
> > >
> > > When I load an object file by calling LoadModule() or ld (1)
> > >  > >
> > > Relocation value does not fit in 24 bits.
> > > Relocation value does not fit in 24 bits
> > > ......
>


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: c++ comments
Date: Wed, 09 Aug 2000 10:22:08 -0500
From: cbsmith 
Organization: Soutwest Research Institute
Message-ID: <39917720.BC9A8DA7@here.org>
References: <39915358.F4E8D94F@yahoo.com>
Reply-To: cbsmithATswriDOTorg

take the -ansi compiler flag out of you makefile or project c/c++ settings.

chris

Paul Oliveira wrote:

> Hi
>
> We are switching to T2 and from solaris to NT. I got the lovely job of
> doing the port.  When I compile using ccmips on solaris (outside of T1),
> no problems (because somewhere in my environment the problem is taken
> care of).  When I compile on NT with T2 I get compile errors on the //
> (c++ comment) lines.  What the heck am I missing to solve this.
>
> Thanks
> Paul


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Socket buffer limit?
Date: Wed, 09 Aug 2000 10:46:22 -0500
From: cbsmith 
Organization: Soutwest Research Institute
Message-ID: <39917CCE.C6D3066F@here.org>
References: <3991744D.82206FCE@pentek.com>
Reply-To: cbsmithATswriDOTorg

These "bufers" have different meanings if you are using tcp or udp. For tcp
they are used in negotiating a data transfer window with the other end of a
connection. The field used in this negotiation is an unsigned short -- so
64K is the max. VxWorks does implement a window scaling algorithm which is
designed to get around this limit. (try looking in netLib.h if you need
this)

For udp this limit does not apply. I think I remember from somewhere that
the current absolute max limit is around 256K.

chris

Steve Rottinger wrote:

> Hi,
>
> Does anyone know if there is a hard limit to the socket buffer size that
> you
> can specify under vxWorks.  I'm using setsockopt() to adjust the
> SO_SNDBUF,
> and SO_RCVBUF.   Things seem to work fine with a 64K buffer, but a 128K
> buffer results in a crash.
>
> Thanks in advance,
>
> -Steve


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: NTP client on VxWorks
Date: 9 Aug 2000 19:07:30 GMT
From: litwin@izzy.jpl.nasa.gov (Todd Litwin)
Organization: jpl.nasa.gov
Message-ID: <8msa5i$it2$1@nntp1.jpl.nasa.gov>
References: <311c2be6.cfb476ec@usw-ex0106-045.remarq.com> <399153B1.C906C5EE@lmco.com>

Bryan D. Lee  wrote:

>If you want v4 (current release) then either go at it yourself, or wait till
>I'm done.  I'm trying to port v4 now.

I also am interested in having an NTP-client under VxWorks. From your posting,
I'm assuming that you will be making the results of your efforts public. I'll
be looking forward to it. I suspect many others will, too.

When do you think that you will have a version to release? How will you
announce its availability, and where will you be posting it?

Thank you.

- -- 
		Todd Litwin
		Jet Propulsion Laboratory
		(818) 354-5028
		Todd.E.Litwin@jpl.nasa.gov

---------------------------

Newsgroups: comp.os.vxworks
Subject: tcp retransmit delays
Date: Wed, 09 Aug 2000 11:42:46 -0500
From: Chris Smith 
Organization: Soutwest Research Institute
Message-ID: <39918A06.50372148@here.org>
Reply-To: cbsmithATswriDOTorg

Has anyone figured out how to get around, or scale down the time it
takes a tcp packet to be retransmitted?

I have several vxworks processors conencted with ethernet based tcp
socket connections. Occasionally I see several packets dropped from the
network. I am seeing a ~1.5 second pause when a packet needs to be
retransmitted.

Using tcpDebugShow this pause corresponds to a SLOWTIMO event.

I have seen a couple of packet trace scenarios in which the slow
retransmit is putting a long pause in my data stream (under different
network environments). In my latest packet traces it appears that the
"fast retransmit algorithm" is working and retransmitting the missed
packets in a resonable amount of time (<100 ms) but there appears to be
a building effect until a packet times out. In previous traces the fast
retransmit does not appear to work at all.

I have tried playing with the KEEP_ALIVE parameters that others have
suggested, and using TCP_NODELAY, but they do not help. (they actually
should not help except through side effects)

chris


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: NTP client on VxWorks
Date: Wed, 09 Aug 2000 20:47:19 GMT
From: john_94501@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8msg0l$9ci$1@nnrp1.deja.com>
References: <311c2be6.cfb476ec@usw-ex0106-045.remarq.com>

Add the INCLUDE_SNTPC component to the kernel, then do something like
this in your application:

    struct timespec     tm;

    /* Make sure we have a default route set */

    routeAdd ("0", DEFAULT_GATEWAY);

    /* Set the time */

    for (i=0; i < NUM_RETRIES; i++)
        {
        if (sntpcTimeGet (TIME_SERVER, sysClkRateGet() * 5, &tm) == OK)
            {
            break;
            }
        }

That gets you a timespec structure from the server. After that you'll
need to cope with time zones etc to get the VxWorks clock set correctly.

HTH,

John...

In article <311c2be6.cfb476ec@usw-ex0106-045.remarq.com>,
  Warren.Brown  wrote:
> Hello
> Has anyone added the NTP (Network Time Protocol) client
> services to a VxWorks embedded target ?
> Any hints as to the right steps to take ?
>
> W
>
> -----------------------------------------------------------
>
> Got questions?  Get answers over the phone at Keen.com.
> Up to 100 minutes free!
> http://www.keen.com
>
>


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: c++ comments
Date: Wed, 09 Aug 2000 10:23:39 -0700
From: Eric Monsler 
Organization: The ISP formerly known as Best
Message-ID: <3991939B.822C0029@radixtek.com>
References: <39915358.F4E8D94F@yahoo.com>

Paul,

How much code are we talking about here?

How was it compiled before, and how much use and debugging has the
application had?

I would be worried about changing compiler options on stable, debugged
code, especially if it was a large, complex codebase that would be
difficult to regression test.  If the code had previously always been
run with a nominally ANSI compiler that happened to accept C++ comments,
I would want to compile it again as ANSI.  Replacing the "//" with "/*"
and placing a "*/" at the end of that line would be a solution unlikely
to introduce errors as subtle as ansi vs. non-ansi would.  A fairly
simple script ought to be able to do this, maybe with awk (any
text-processing gurus out there?) or with perl.

If the compilation was previously non-ansi, then removing the -ansi
option is definitely the better solution.

My $/50

Eric Monsler

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: dec21x4xEnd driver
Date: Wed, 09 Aug 2000 21:05:59 GMT
From: john_94501@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8msh3j$a9k$1@nnrp1.deja.com>
References: <8mn55b$j7p$1@nnrp1.deja.com> <8mre08$t92$2@proxy2.fe.internet.bosch.de>

Hi,

You are correct to switch; the driver was a 10Mbps only configuration
(although you have the media select function in the BSP so you should be
able to update it if you so choose).

A better solution would be to ask WR customer support if they have the
dec21x40End driver available for your version of the OS and
architecture. I am assuming that you're using ARM, in which case you
need some mods to the 40 driver to make it work (ARM-specific alignment
issues). The newest 40 driver also supports some of the newer 21143
devices properly.

Hope that helps,

John...

In article <8mre08$t92$2@proxy2.fe.internet.bosch.de>,
  Jan Koeppen  wrote:
> We had the problem that this driver seems not to be working on a
100MBit Hub
> :-((.
> We solved it by buying old 10Mbit hubs that work fine (but is not
> satisfactory.)
>
> regards
> Jan
>
> stourangeau@my-deja.com schrieb in Nachricht
> <8mn55b$j7p$1@nnrp1.deja.com>...
> >I am having problems setting up the dec21x4xEnd driver. I dont have
the
> >source file so I am only playing with the configuration of the
driver.
> >
> >After lots of work, I was able to make the driver attach to the mux
and
> >attach protocols to it.
> >
> >I had lots of problem with getting the link to be indicated on the
> >hub... I had to play around with a mediaSelect routine..
> >
> >Problem is, the hub sees the link but the dec21x doesnt seem to see
the
> >hub. By looking at the status registers, I see link failures and
> >autonegociation that is not completed...
> >
> >Need some help please...
> >
> >
> >Sent via Deja.com http://www.deja.com/
> >Before you buy.
>
>


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: mv2700 duplicate int! why?
Date: Wed, 09 Aug 2000 14:17:18 -0400
From: Al Johnston 
Organization: MindSpring Enterprises
Message-ID: <3991A02E.868A848C@mindspring.com>
References: <399188B1.C9AA0763@mindspring.com> <39919BB3.8D0848AB@aps.anl.gov>

I sure am glad you are a nice guy... thanks for answering.

> I think your "ignoring the compatibility rational"  is the key point - if
> you look at the BSP, from memory I think it either initializes one chip or
> the other depending on a compile-time switch somewhere.  I assume that the

I was kind of expecting that, and best I can tell it is not there.  what
I
have found is that it inits the PIB on the winbond chip
(sl82565/w83c553f).
It then if you have an include set (INCLUDE_MPIC) - which we do -
it will initializes the MPIC on the Raven.

So, best I can tell they are both being set up...and best I can tell,
for 
at least some of the lines they are active on both chips.

I will work through it again, maybe I missed something....

---------------------------

Newsgroups: comp.os.vxworks
Subject: Condition variables in VxWorks?
Date: Wed, 09 Aug 2000 22:44:34 GMT
From: James Sheppard 
Organization: Time Warner Cable of San Diego, CA
Message-ID: 

I'm just starting to learn multi-threaded programming, so bear with me.  I
was started off with pthreads and now I'm learning VxWorks.  I was wondering
if there was something similiar to the condition variables found in pthreads
that is in VxWorks.

I was trying to use binary semaphores, but that doesn't seem to work the way I
thought they did.

Thanks,

James

reply to warblade@san.rr.com 
if responding by e-mail



---------------------------

Newsgroups: comp.os.vxworks
Subject: MVME2604 Memory maps
Date: Wed, 9 Aug 2000 18:53:06 GMT
From: Joe Andrews 
Organization: Vehicle Managememt Systems
Message-ID: <3991A892.81F1DBB8@mw.boeing.com>
Reply-To: joe.andrews@mw.boeing.com
Sender: nntp@news.boeing.com (Boeing NNTP News Access)

I hope this isn't too basic, but I'm brand new to the VxWorks and VME
world and need to get a couple of MVME2604 boards working. I can load
VxWorks boot image and flash stuff fine, but I can for the life of me
figure out how to set up the memory mapping on the board. All I want to
do is map the board's memory into the A32 address space at
0x30000000-0x31fffff. Can anybody shed some light on what I need to do
to make this happen?

Thanks,

Joe Andrews

---------------------------

Newsgroups: comp.os.vxworks
Subject: MVME2604 Memory maps
Date: Wed, 9 Aug 2000 18:50:06 GMT
From: Joe Andrews 
Organization: Vehicle Managememt Systems
Message-ID: <3991A7DE.C76A4225@mw.boeing.com>
Reply-To: joe.andrews@mw.boeing.com
Sender: nntp@news.boeing.com (Boeing NNTP News Access)

I hope this isn't too basic, but I'm brand new to the VxWorks and VME
world and need to get a couple of MVME2604 boards working. I can load
VxWorks boot image and flash stuff fine, but I can for the life of me
figure out how to set up the memory mapping on the board. All I want to
do is map the board's memory into the A32 address space at
0x30000000-0x31fffff. Can anybody shed some light on what I need to do
to make this happen?

Thanks,

Joe Andrews

---------------------------

Newsgroups: comp.os.vxworks
Subject: Network File system for vxWorks other than NFS
Date: Wed, 09 Aug 2000 16:04:11 -0700
From: Waqar 
Organization: http://www.remarq.com: The World's Usenet/Discussions Start Here
Message-ID: <1978aa00.54f0c802@usw-ex0105-040.remarq.com>

We are using vxWorks v5.31 and use NFS to share files between
our target boards. We find NFS to be slow and would like to find
an alternative for that (CFS maybe). Does anyone out there have
any information on where one can get/buy such a file system that
can work with vxWorks or can be ported to vxWorks.

Waqar


- -----------------------------------------------------------

Got questions?  Get answers over the phone at Keen.com.
Up to 100 minutes free!
http://www.keen.com


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Tornado Launch
Date: Wed, 09 Aug 2000 21:15:29 GMT
From: john_94501@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8mshlc$aoj$1@nnrp1.deja.com>
References: <398FDBC8.2EBB9582@astro.gla.ac.uk> <3990125C.1E3D9C0F@delphi.com>

Hi,

I seem to remember that in 1.0.1 there was a license manager bug that
resulted in corruption of the key file sometimes when two or more
products tried to update it together.

Speak to customer support about this as I think there was a patch from
Elan that fixed the bug in their license manager s/w.

Hope that helps,

John...

In article <3990125C.1E3D9C0F@delphi.com>,
  Joe Durusau  wrote:
> 	Which version of Tornado, what OS?  On my system, running Tornado
> 1.0.1,
> Solaris 2.6, there is a file in /etc/rc2.d that starts lmgrd with a
> parameter
> pointing to the license file.  whatever equiv. you have may be
damaged,
> or you
> may need to restart the server.  AFAIK, anybody can start it.  Try
> ps -e | grep  to see if the server
is
> in fact
> running.   If it is, send it a hangup and maybe it will reread the
> license file.
> If not , run the command that starts it.
>
> HTH.
>
> Speaking only for myself,
>
> Joe Durusau
>
> Martin Hewitson wrote:
> >
> > Hell out there,
> >
> > Maybe someone can help. I am using the Tornado Launch program to
> > (primarily) boot and control three data servers that have vxWorks
> > installed on them. On occasion, the launch program decides it can't
> > connect to the servers and starts complaining about the license
server
> > being down??!!? Wind River gave me instructions as to how to install
a
> > new license and this seems to go some way to solve the problem.
However,
> > at the end of the key installation, the install program says that
the
> > license server is down. When we use launch to connect to the servers
> > now, it works.  Anybody any idea why it is going wrong in the first
plae
> > and why a failed attempt to re-install the license key should
correct
> > the problem?
> >
> > Any suggestions would be gratefully received,
> >
> > thankjs
> >
> > M
>


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Testing END Driver Mulitcast
Date: Wed, 9 Aug 2000 14:43:18 -0700
From: "AG" 
Organization: SBC Internet Services
Message-ID: 

Hi,

I'm trying to setup a UDP socket for multicast and I'm using the following
send function:

sendto (sockDesc, bufPtr, len, 0, (struct sockaddr *) &, len);

but it returns an errno =>   0x41 = S_errno_EHOSTUNREACH.

Can anyone point out what the problem could be? Also, is there another
simple way to test a driver+IP stack multicast capability?
If I use "ifShow" the driver multicast flag is enabled.

Thanks so much,

AG



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: NTP client on VxWorks
Date: Wed, 09 Aug 2000 16:44:42 -0400
From: "Bryan D. Lee" 
Organization: Sanders-A Lockheed Martin Company
Message-ID: <3991C2BA.4FAB2A86@lmco.com>
References: <311c2be6.cfb476ec@usw-ex0106-045.remarq.com> <399153B1.C906C5EE@lmco.com> <8msa5i$it2$1@nntp1.jpl.nasa.gov>

If you're looking for a basic NTP client (no PPS support) then I can
tell you now what to do to get one working.  I can't offer patches
since the computer system I'm working on is on an isolated network. 
Hence, I'm going to go through the trouble of sneaker-netting this
once and once only; when I'm finally done.

I'll post info here when I'm done and the changes will be submitted to
the main NTP CVS tree on UDel.

At the moment, the Phase Locked Loop and Freq Locked Loop for slaving
the time within the kernel is driving me up a wall, so it'll be a
while on that score.  And you/I would need this for PPS
synchronization or to discipline the clock to a tighter tolerance.  It
touches on the stuff I've already modified, hence why I haven't
released the stuff I've done so far.

Bryan


Todd Litwin wrote:
> 
> Bryan D. Lee  wrote:
> 
> >If you want v4 (current release) then either go at it yourself, or wait till
> >I'm done.  I'm trying to port v4 now.
> 
> I also am interested in having an NTP-client under VxWorks. From your posting,
> I'm assuming that you will be making the results of your efforts public. I'll
> be looking forward to it. I suspect many others will, too.
> 
> When do you think that you will have a version to release? How will you
> announce its availability, and where will you be posting it?
> 
> Thank you.
> 
> --
>                 Todd Litwin
>                 Jet Propulsion Laboratory
>                 (818) 354-5028
>                 Todd.E.Litwin@jpl.nasa.gov

- -- 
Bryan D. Lee                             Phone: 603/885-7025
Software Engineer & ELDP                 Fax  : 603/885-3016
Sanders-A Lockheed Martin Company        Pager: 603/423-5565
PO Box 868, PTP1-2361, Nashua, NH 03061  Email: bryan.d.lee@lmco.com

---------------------------

Newsgroups: comp.os.vxworks,comp.compilers,comp.programming
Subject: Re: how does purify work?
Date: 10 Aug 2000 00:04:52 -0400
From: ralph@inputplus.demon.co.uk (Ralph Corderoy)
Organization: InputPlus Ltd.
Keywords: debug, tools
Message-ID: <00-08-049@comp.compilers>
References: <00-08-027@comp.compilers>

Hi,

> If one wanted to build a Purify like tool, does anyone know how it
> would be done? I don't really understand what it does and haven't
> found much related info.

You might want to take a look at Aprobe from OC Systems.
www.aprobe.com.  Follow the Papers link to the _Aprobe Whitepaper_.  It
contains background on how Aprobe works.


Ralph.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Debugging variables defined in nested blocks
Date: Wed, 09 Aug 2000 22:50:14 GMT
From: Edward Bishop 
Organization: Deja.com - Before you buy.
Message-ID: <8msn75$em6$1@nnrp1.deja.com>

I would like to be able to be able to view the contents of the
variable 'status' in code like this:

int bar(void)
{
    int blub;

    while(1)
    {
        int status = foo();
        // break here and check status
        ...
    }
}

but "p status" yields 'No symbol "status" in current context'.  There
is no problem viewing variables such as blub.  I am using
Tornado2 on a MIPS 5400 processor.  Are there any debugger or
compiler options (besides -g of course) that I should be using?
Thanks in advance for comments or advice.

- --
edward.bishop@am.sony.!spam.com


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: tcp retransmit delays
Date: 9 Aug 2000 23:22:40 GMT
From: kml@baygate.bayarea.net (Kevin Lahey)
Organization: Bayarea Internet Solutions
Message-ID: <8msp40$qr$1@news.bayarea.net>
References: <39918A06.50372148@here.org>

In article <39918A06.50372148@here.org>,
Chris Smith   wrote:
>Has anyone figured out how to get around, or scale down the time it
>takes a tcp packet to be retransmitted?
>
>I have several vxworks processors conencted with ethernet based tcp
>socket connections. Occasionally I see several packets dropped from the
>network. I am seeing a ~1.5 second pause when a packet needs to be
>retransmitted.

A 1.5 second pause for retransmission is pretty reasonable, according
to the RFCs.

>I have seen a couple of packet trace scenarios in which the slow
>retransmit is putting a long pause in my data stream (under different
>network environments).

I've noticed this, too, with delays of up to 20 seconds, which is
pretty much ridiculous.  I don't know what to suggest;  I plan to
do some more testing on a controlled network, then submit a bug
report.

Kevin
kml@bayarea.net

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Debugging variables defined in nested blocks
Date: Thu, 10 Aug 2000 07:19:15 +0200
From: "Serge Wenger" 
Organization: Swisscom IP+ (post doesn't reflect views of Swisscom)
Message-ID: <8mtdpu$oj$1@pollux.ip-plus.net>
References: <8msn75$em6$1@nnrp1.deja.com>

Try to remove optimisation with -O0

- --
/Serge Wenger

PS: in my e-mail replace nobody by matisa

"Edward Bishop"  a écrit dans le message news:
8msn75$em6$1@nnrp1.deja.com...
> I would like to be able to be able to view the contents of the
> variable 'status' in code like this:
>
> int bar(void)
> {
>     int blub;
>
>     while(1)
>     {
>         int status = foo();
>         // break here and check status
>         ...
>     }
> }
>
> but "p status" yields 'No symbol "status" in current context'.  There
> is no problem viewing variables such as blub.  I am using
> Tornado2 on a MIPS 5400 processor.  Are there any debugger or
> compiler options (besides -g of course) that I should be using?
> Thanks in advance for comments or advice.
>
> --
> edward.bishop@am.sony.!spam.com
>
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.



---------------------------

Newsgroups: comp.os.vxworks
Subject: Replacement for Altera ByteBlaster & ByteBlasterMV
Date: Thu, 10 Aug 2000 01:18:37 GMT
From: bingoeugene@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8msvtb$ki0$1@nnrp1.deja.com>

Reliable and cost effective replacement solution for the Altera
Corporation ByteBlaster and ByteBlasterMV download cable. Downloads
data from the MAX+PLUS ® II development software.Programs MAX ® 9000,
7000 in-system. Configures FLEX ® 20K, 10K, FLEX 8K, FLEX 6K devices

AnyVoltage VCC ( from 1.8 V - 5.5 V)

AnyVoltage FPGA / EPLD I/O ( 1.8 V - 5.5 V)

3' Cable version available

You will never need another FPGA downloader.


Please visit:


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: WRS white paper on vxWorks timing
Date: Wed, 09 Aug 2000 19:28:29 -0700
From: Luciana Messina 
Organization: Software to Spec, Inc.
Message-ID: <3992134D.60275B98@ricochet.net>

WRS used to provide a white paper on vxWorks timing. Does anyone have a
copy they wouldn't mind sharing on the Web (I've misplaced my copy,
dated 1997) ?  I'll type it up and post it on my site if someone can
provide...

Luciana




---------------------------

Newsgroups: comp.os.vxworks
Subject: Tornado 2: How do I redirct the Virtual I/O Output???
Date: Wed, 9 Aug 2000 22:30:32 -0400
From: "Paul Worley" 
Organization: WebUseNet Corp  http://www.usenetserver.com - Home of the fastest NNTP servers on the Net.
Message-ID: <4rok5.3159$Lw4.134757@news-west.usenetserver.com>
Reply-To: "Paul Worley" 

I am running VxWorks on an embedded system and connecting to it with Tornado
2 from Windows NT.  I have large amounts of data directed to stdout that is
being redirect to the Virtual I/O Console window.  I need to capture this
data into a file on the NT machine.  How do I redirect the data from virtual
I/O to a file on the NT machine?

Thanks,
Paul





---------------------------

Newsgroups: comp.os.vxworks
Subject: Access Network from VxSim ?
Date: 10 Aug 2000 02:48:15 GMT
From: drennall@chat.carleton.ca (David Rennalls)
Organization: Carleton University, Ottawa, Canada
Message-ID: <8mt55f$pb1$1@bertrand.ccs.carleton.ca>

I have a VxSim(Full) running on Windows NT with the Ulip driver installed.
Although I can ping the VxSim from another machine on the network I can't
seem to ping any external machines from a shell on the simulator. For e.g.

NT Box1 - 192.168.0.20
VxSim0 - 90.0.0.1 (on Nt Box1)
From VxSim0 I can ping the Nt Box1 but no other 192.168.0.x hosts

NT Box2 - 192.168.0.21
If a do a "route add 90.0.0.1 192.168.0.20" on Box2 and I can ping VxSim0

routeShow on VxSim0 gives the following:

ROUTE NET TABLE
destination      gateway          flags  Refcnt  Use   Interface
- ------------------------------------------------------------------
0.0.0.0          90.0.0.1         101    0       0      nt0

ROUTE HOST TABLE
- ------------------------------------------------------------------
127.0.0.1        127.0.0.1        5      0       0      lo0

hostShow gives:
hostname         inet address       aliases
- --------         ------------       -------
vxTarget         90.0.0.1           
localhost        127.0.0.1          
host             90.0.0.254    

I tried changing the default gateway from 90.0.0.1 to 90.0.0.254 but with
the same result. Does anyone know why I can connect into the sim but not
out past the Nt Host it runs on ? Any help would be greatly appreciated.


- ----------------------------->
David Rennalls  
drennall@chat.carleton.ca




---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: NTP client on VxWorks
Date: Wed, 09 Aug 2000 17:51:07 -0400
From: "Bryan D. Lee" 
Organization: Sanders-A Lockheed Martin Company
Message-ID: <3991D24B.97FCA644@lmco.com>
References: <311c2be6.cfb476ec@usw-ex0106-045.remarq.com> <8msg0l$9ci$1@nnrp1.deja.com>

Mind if I ask what version of vxworks this is?  I don't have it in the
kernel and I don't see it in the 5.3 docs.

Bryan

john_94501@my-deja.com wrote:
> 
> Add the INCLUDE_SNTPC component to the kernel, then do something like
> this in your application:
> 
>     struct timespec     tm;
> 
>     /* Make sure we have a default route set */
> 
>     routeAdd ("0", DEFAULT_GATEWAY);
> 
>     /* Set the time */
> 
>     for (i=0; i < NUM_RETRIES; i++)
>         {
>         if (sntpcTimeGet (TIME_SERVER, sysClkRateGet() * 5, &tm) == OK)
>             {
>             break;
>             }
>         }
> 
> That gets you a timespec structure from the server. After that you'll
> need to cope with time zones etc to get the VxWorks clock set correctly.
> 
> HTH,
> 
> John...
> 
> In article <311c2be6.cfb476ec@usw-ex0106-045.remarq.com>,
>   Warren.Brown  wrote:
> > Hello
> > Has anyone added the NTP (Network Time Protocol) client
> > services to a VxWorks embedded target ?
> > Any hints as to the right steps to take ?
> >
> > W
> >
> > -----------------------------------------------------------
> >
> > Got questions?  Get answers over the phone at Keen.com.
> > Up to 100 minutes free!
> > http://www.keen.com
> >
> >
> 
> Sent via Deja.com http://www.deja.com/
> Before you buy.

- -- 
Bryan D. Lee                             Phone: 603/885-7025
Software Engineer & ELDP                 Fax  : 603/885-3016
Sanders-A Lockheed Martin Company        Pager: 603/423-5565
PO Box 868, PTP1-2361, Nashua, NH 03061  Email: bryan.d.lee@lmco.com

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Tornado 2: How do I redirct the Virtual I/O Output???
Date: 10 Aug 2000 10:42:32 GMT
From: johan@borksoft.xs4all.nl
Organization: Borksoft
Message-ID: <8mu0uo$jvk$1@news1.xs4all.nl>
References: <4rok5.3159$Lw4.134757@news-west.usenetserver.com>

Paul Worley  wrote:
: I am running VxWorks on an embedded system and connecting to it with Tornado
: 2 from Windows NT.  I have large amounts of data directed to stdout that is
: being redirect to the Virtual I/O Console window.  I need to capture this
: data into a file on the NT machine.  How do I redirect the data from virtual
: I/O to a file on the NT machine?

Take a look am my VxWorks page, question 1.11.1

Groeten,
	Johan

- -- 
     o o o o o o o . . .   ___________________________________
    o      _____           ||    Johan Borkhuis              |
  .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
 >(________|__|_[_________]_|________________________________|
 _/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  Thu Aug 10 21:33:27 2000
From: "Krishna Kishore Reddy Sureddy" 
Date: Thu Aug 10 21:33:29 PDT 2000
Subject: vxWorks task suspend

Hi,
   Can a task in VXWORKS get suspended other than the conditions
preseted below.
   If yes please send them across as soon  as possible.

   1. Task causing exception(task level exception).
   2. Suspending a task from shell by ts( )  or taskSuspend( ) .
   3. Attaching a task to Debugger.
   4. Keeping taskSuspend(0) in the code of the task.
   5. Keeping a break point in the code of the task.

   The below case i am not sure and correct me if i am wrong
   6. When you are trying to free/delete the memory which is already
deleted.

Thanks and Regards,
S.K.K.Reddy.






From vxwexplo-errs@csg.lbl.gov  Fri Aug 11 04:03:09 2000
From: Vxworks Exploder 
Date: Fri Aug 11 04:03:11 PDT 2000
Subject: comp.os.vxworks newsdigest

Comp.Os.Vxworks Daily Digest    Fri Aug 11 04:03:04 PDT 2000

        Subject: Use the cbio cache
        Subject: Re: dosFs 2.0 and NFS
        Subject: Select() problems when TOS is set
        Subject: UDP Sockets crashes if data >1540 bytes
        Subject: Re: UDP Sockets crashes if data >1540 bytes
        Subject: Re: help for LoadModule
        Subject: P3 Interrupt Latency Time
        Subject: mkbootAta question
        Subject: Re: tcp retransmit delays
        Subject: Re: tcp retransmit delays
        Subject: Re: help for LoadModule
        Subject: Re: Condition variables in VxWorks?
        Subject: Re: tcp retransmit delays
        Subject: Re: Does anyone use VxWorks on a standard PC?
        Subject: Re: tcp retransmit delays
        Subject: AMD 79C973 driver (BSP Janich Klass ISA p7/103)
        Subject: IP broadcasting to 255.255.255.255 ??
        Subject: Re: IP broadcasting to 255.255.255.255 ??
        Subject: Re: help for LoadModule
        Subject: which function can I use to get host ipaddr?
        Subject: Re: WIN2000 SP1 & TORNADO II
        Subject: Re: Display of Windows NT Windows on a VXWorks Power PC
        Subject: Re: NTP client on VxWorks
        Subject: Re: MVME2604 Memory maps
        Subject: Re: MVME2604 Memory maps
        Subject: Re: Testing END Driver Mulitcast
        Subject: Need a better FTP Server soft for VxWorks
        Subject: Re: OT BSP question
        Subject: Re: AMD 79C973 driver (BSP Janich Klass ISA p7/103)
        Subject: Virtual memory exhausted error
        Subject: Re: Need a better FTP Server soft for VxWorks
        Subject: makefile problems
        Subject: Tundra PowerSpan Evaluation Board
        Subject: Re: Testing END Driver Mulitcast
        Subject: Re: Condition variables in VxWorks?
        Subject: Re: OT BSP question
        Subject: Re: mkbootAta question
        Subject: vxWorks task suspend
        Subject: Re: tcp retransmit delays
        Subject: BSP for mvme374
        Subject: Re: makefile problems
        Subject: Problem white read of A24 (MVME2305)
        Subject: mvme1604 to 2604 or 2700
        Subject: Re: Does anyone use VxWorks on a standard PC?

-------------------------------------------------------

Newsgroups: comp.os.vxworks
Subject: Use the cbio cache
Date: Thu, 10 Aug 2000 14:31:10 +0200
From: "Kilbo, Daniel" 
Keywords: dosFs,NFS
Message-ID: 
References: <395771CC.1208329E@wmdata.com> <39584590.B3935F5B@bitband.com>
Sender: "Kilbo, Daniel" 

I have now found a solution on the problem with the NFS and dosFs2.0.

The problem occures when the dosFS is runned without the new software
cache cbio. Since I don't need a cache it would only cost performance
for me to use it. But when the cache is removed NFS totaly freakes out.

If you are going to use NFS with dosFs2.0 use the cbio cache!
dcacheDevCreate(...)
 
Regards
/Daniel Kilbo
WM-data eTechnology
 
- -----Original Message-----
From: Leonid Rosenboim
Posted At: 2000-06-27 08:11
Conversation: dosFs 2.0 and NFS
Subject: Re: dosFs 2.0 and NFS

Daniel,

DosFs 1.0 had some built-in features to support NFS, some of which had
caused great problems. For example, for the sake of NFS server, dosFs
1.0
had a list of all directory entries (files and dirs) in RAM. This made
NFS
server handle-based searches much faster, but caused RAM consumption fo
explode when large disk drives with a lot of files where used, while the
"typical" disk drive size has grown several orders of magnitude sicne
VxWorks 5.1.1 which is when the NFS server was introduced. Also, dosFs
1.0
pact with NFS server depended on the 16-bit FAT entry size, which FAT32
breaks naturally.

So in dosFs 2.0, a new NFS server was implemented as a completely
layered
module on top of dosFs 2.0, without any intimate relations between the
two.

I do not know much details about the new NFS server as it was outside
the
scope of the team I was heading, but my best guess is that the new NFS
server implementation could be improved to enhance performance. For you
I
would recommend to see if you can use the FTP server on the VxWorks side
instead, just a wild guess.

Leonid

Daniel Kilbo wrote:

> Hi!
>
> I got a problem with a SCSI-disk on a Power PC card with dosFs 2.0 on
> it. The disk is mounted on a unix system. The transfer rate between
the
> disk and the unix-system is about 60 kBytes/s. Very low! With old
dosFs
> it was about 10 times better.
>
> Does anybody know why dosFs 2.0 and NFS is slow? Or how it could be
> speeded up.
>
> Regards
> /Daniel Kilbo


---------------------------

Newsgroups: comp.os.vxworks
Subject: Select() problems when TOS is set
Date: Thu, 10 Aug 2000 15:37:12 GMT
From: khanh_nguyen44@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8mui78$oh8$1@nnrp1.deja.com>

I call select() waiting only for readability on
TCP sockets. When the Type of Service (TOS) is
set in the IP header, select() does NOT return.
The TOS option is set as follows:

int tos = IPTOS_LOWDELAY | IPTOS_PREC_CRITIC_ECP;
if ( setsockopt (socketDescriptor, IPPROTO_IP,
IP_TOS, (char *)&tos, sizeof(int)) == ERROR)
    perror("Error setting socket option\n");
else
    printf ("setsockopt OK\n");

By using a sniffer, I can see that the exchange
of packets for a TCP connection is correct
between client and server. But for some unknown
reason, the packets are NOT sent up to the server
application from the TCP layer.

Thanks in advance.

- ---
Khanh Nguyen
System IP Access Interfaces
Mitel Corporation
613-592-2122 x1746



Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: UDP Sockets crashes if data >1540 bytes
Date: Thu, 10 Aug 2000 13:39:07 +0200
From: Florian-A Schindler 
Organization: Siemens AG,(Hofmannstr)  Munich-Germany-Europe.
Message-ID: <3992945B.8CA86EC5@icn.siemens.de>

Hello!

I try to send more than 1540 bytes in a UDP packet (datagram) and the
"sendto" call crashes with
"Exception !
 Vector 13 : General Protection Fault
  Program Counter : 0x00000000
 Status Register : 0x00010246"
(Smaller size such about 1400 bytes is OK!)

I use a WinNT simulator (with networking support) and I think all
buffers are sufficient (default values are used).
Also the setsockopt( )/SO_SNDBUF should be sufficient (default value is
9216 bytes).
Does anybody know, if this is a socket problem or is the WinNT simulator
the reason for the crash (if the latter case, then I can hope it will
run at the real target)?

Thanks and best regards,
  Florian


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: UDP Sockets crashes if data >1540 bytes
Date: Thu, 10 Aug 2000 09:33:28 -0700
From: Eric Monsler 
Organization: The ISP formerly known as Best
Message-ID: <3992D958.6CD3D0F6@radixtek.com>
References: <3992945B.8CA86EC5@icn.siemens.de>

Florian-A Schindler wrote:
> 
> Hello!
> 
> I try to send more than 1540 bytes in a UDP packet (datagram) and the
> "sendto" call crashes with
> "Exception !
>  Vector 13 : General Protection Fault
>   Program Counter : 0x00000000
>  Status Register : 0x00010246"
> (Smaller size such about 1400 bytes is OK!)
> 
> I use a WinNT simulator (with networking support) and I think all
> buffers are sufficient (default values are used).
> Also the setsockopt( )/SO_SNDBUF should be sufficient (default value is
> 9216 bytes).
> Does anybody know, if this is a socket problem or is the WinNT simulator
> the reason for the crash (if the latter case, then I can hope it will
> run at the real target)?
> 
> Thanks and best regards,
>   Florian


Are you sending this UDP packet over ethernet?  You're just over the
ethernet frame size, which might be a clue.  If it works at 1492 but not
1493, or at 1518 but not 1519, then that is definitely related to the
problem.

But, my understanding is that a UDP packet should be broken into two IP
datagrams, each in their own ethernet frame, if the protocol stacks all
behave as standards dictate.

Sounda like a broken socket/UDP/IP implementation.

Eric Monsler

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: help for LoadModule
Date: Thu, 10 Aug 2000 08:50:02 +0200
From: Uwe Fuhr 
Organization: Ericsson Eurolab Deutschland GmbH
Message-ID: <3992509A.BC802787@eede.ericsson.se>
References: <8mda0f$29di$1@news.cz.js.cn> <8mdsgb$8vn$1@nnrp1.deja.com> <39915AEB.EF71C6EC@eede.ericsson.se> <8mrsvn$ptp$1@nnrp1.deja.com>
Reply-To: Uwe.Fuhr@eed.ericsson.se

Hi Urban,

I am not sure if the error message 

> Relocation value does not fit in 24 bits.

is really erroneous.

One think I can imagine is, that the code which is loaded via 
the ld() command can not be located/linked into the area of 24 Bits.
This means that code must be placed within the space of 16 MB.
If this is the fact, then I should have had the same problem
before, when using 32 MB of RAM.

Do you mean the -G option? And why do you think that helps?

Regards

/Uwe
- -- 


Urban Lindberg wrote:
> 
> Hi Uwe!
> 
> I don't use the printed manuals, but the online manuals and there you
> will find it in: GNU Toolkit User's Guide/The C/C++ Compiler/IBM RS/6000
> and PowerPC Options. It's the last option mentioned in that section.
> And, yes the error message is erroneous, it should be 26 bits :-)
> 
> HTH
> Urban
> 
> In article <39915AEB.EF71C6EC@eede.ericsson.se>,
>   Uwe.Fuhr@eed.ericsson.se wrote:
> > Hi Urban,
> >
> > I run into the same problem as Tang, while extending the RAM from 32
> > to 64 MB of RAM, and then loading objects via ld() to the new 64 MB
> vxWorks
> > load module.
> >
> > We are using the gcc compiler but I could not find the mentioned
> > option -mlongcall in the manual.
> >
> > What about this.
> >
> > Best regards
> >
> > /Uwe
> >
> > --
> >
> ___________________________________________________________________________
> >
> >      _/  _/  _/  Uwe Fuhr                       EED/E/D/U    Access
> Design
> >     _/  _/  _/   Ericsson Eurolab Deutschland   phone  : +49-5121 -
> 707104
> >    _/  _/  _/    Daimlerring 9                  FAX    : +49-5121 -
> 707314
> >   _/  _/  _/     31135 Hildesheim - Germany     mobile : +49-173 -
> 5430771
> >
> > __________________________________   mailto:Uwe.Fuhr@eed.ericsson.se
> ____
> >
> \___________________________________/
> >
> > > Hi!
> > >
> > > You need to add the -mlongcall option to your compilation.
> > >
> > > HTH
> > > Urban
> > >
> > > In article <8mda0f$29di$1@news.cz.js.cn>,
> > >   "tanghaoyu"  wrote:
> > > > Hello All,
> > > >
> > > > When I load an object file by calling LoadModule() or ld (1)
> > > >  > > >
> > > > Relocation value does not fit in 24 bits.
> > > > Relocation value does not fit in 24 bits
> > > > ......
> >
> 
> Sent via Deja.com http://www.deja.com/
> Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: P3 Interrupt Latency Time
Date: Thu, 10 Aug 2000 16:52:26 GMT
From: cferrell6421@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8mumk7$s4c$1@nnrp1.deja.com>

Does anyone have ballpark Interrupt Latency timing numbers for an Intel P3
800 MHz w/440BX Chip Set.  Operating under VxWorks 5.4. I could extrapolate
numbers for other x86 processors if you have them.  My WRS FAE is on vacation
and the 1-800 number for WRS support can't help.

Thanks in advance.....



Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: mkbootAta question
Date: Thu, 10 Aug 2000 09:55:16 -0400
From: "Mike Kordik" 
Message-ID: <1uyk5.30$2S1.72@client>

I am trying to get my IDE device to boot vxworks. I am booting from a floppy
right now. I can init the IDE device and I can write files to it that can be
loaded from the bootloader but I cannot make it boot the bootloader. When I
run mkbootAta I get S_dosFsLib_NO_CONFIG_SPACE error. When mkbootAta is done
running I get a /vxsysdos/ device. That device has the bootrom.sys file but
it is zero bytes in length.

Does anyone know what I am doing wrong?

Thanks,
Mike



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: tcp retransmit delays
Date: Thu, 10 Aug 2000 09:20:37 -0500
From: Chris Smith 
Organization: Soutwest Research Institute
Message-ID: <3992BA35.3D769225@here.org>
References: <39918A06.50372148@here.org> <8msp40$qr$1@news.bayarea.net>
Reply-To: cbsmithATswriDoTorg



Kevin Lahey wrote:

> In article <39918A06.50372148@here.org>,
> Chris Smith   wrote:
> >Has anyone figured out how to get around, or scale down the time it
> >takes a tcp packet to be retransmitted?
> >
> >I have several vxworks processors conencted with ethernet based tcp
> >socket connections. Occasionally I see several packets dropped from the
> >network. I am seeing a ~1.5 second pause when a packet needs to be
> >retransmitted.
>
> A 1.5 second pause for retransmission is pretty reasonable, according
> to the RFCs.
>

I can understand a 1.5 second timeout period before retransmitting (although
this seems long for many network environments) but not an actual pause in
the data flow from the vxworks board. I would have thought that even though
a packet may need retransmitting (and take 1.5 seconds for it to be
retransmitted) the packet stream should not be interrupted. I can only see
this interruption occurring if the sender or receiver is running out of
buffer space. I have alloced more than enough mbuf buffers (clusters) to
handle this amount of data (actually over 51 seconds of data).

The connection window is negotiated to 61K. Do you know if the receiver puts
data into mbufs or does it just use the socket receive buffer (the 61K
window)? It is possible that the receiver is running out of space, but I do
not see a change in the window size (like I would expect if it ran out of
room). The 61K window is 7 seconds of data -- that is many packets for it to
be behind.


>
> >I have seen a couple of packet trace scenarios in which the slow
> >retransmit is putting a long pause in my data stream (under different
> >network environments).
>
> I've noticed this, too, with delays of up to 20 seconds, which is
> pretty much ridiculous.  I don't know what to suggest;  I plan to
> do some more testing on a controlled network, then submit a bug
> report.
>

Note that a carefully controlled network would not have any packets dropped
so you would not see the problem (theoretically). I have noticed that some
switches actually increase the number of packets dropped -- just something
to be aware of in your testing.

>
> Kevin
> kml@bayarea.net

chris


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: tcp retransmit delays
Date: 10 Aug 2000 14:44:51 GMT
From: kml@baygate.bayarea.net (Kevin Lahey)
Organization: Bayarea Internet Solutions
Message-ID: <8muf53$huv$1@news.bayarea.net>
References: <39918A06.50372148@here.org> <8msp40$qr$1@news.bayarea.net> <3992BA35.3D769225@here.org>

In article <3992BA35.3D769225@here.org>,
Chris Smith   wrote:
>
>Kevin Lahey wrote:
>
>> In article <39918A06.50372148@here.org>,
>> Chris Smith   wrote:
>> >I have several vxworks processors conencted with ethernet based tcp
>> >socket connections. Occasionally I see several packets dropped from the
>> >network. I am seeing a ~1.5 second pause when a packet needs to be
>> >retransmitted.
>>
>> A 1.5 second pause for retransmission is pretty reasonable, according
>> to the RFCs.
>>
>
>I can understand a 1.5 second timeout period before retransmitting (although
>this seems long for many network environments) but not an actual pause in
>the data flow from the vxworks board. I would have thought that even though
>a packet may need retransmitting (and take 1.5 seconds for it to be
>retransmitted) the packet stream should not be interrupted. I can only see
>this interruption occurring if the sender or receiver is running out of
>buffer space. I have alloced more than enough mbuf buffers (clusters) to
>handle this amount of data (actually over 51 seconds of data).

Hmmm.  Maybe I'm misunderstanding your situation.  The way TCP usually
works is that it sends out more data as it receives acknowledgements for
the old data being received by the other side.  If it stops getting ACKs,
it should stop sending.  It'll then time out, and retransmit the packet.  
While it is timing out, it shouldn't be sending packets.[*]

>The connection window is negotiated to 61K. Do you know if the receiver puts
>data into mbufs or does it just use the socket receive buffer (the 61K
>window)? It is possible that the receiver is running out of space, but I do
>not see a change in the window size (like I would expect if it ran out of
>room). The 61K window is 7 seconds of data -- that is many packets for it to
>be behind.

I'm relatively new to VxWorks, so I'm still just figuring out this
stuff, too.  As far as I can tell, it VxWorks doesn't ever change the
advertised window as a receiver.  I don't know how to find out if it
is running out of all memory, although perhaps you'd get some kinda
log message to the console?

Good luck,

Kevin
kml@bayarea.net

[*]  This is a simplification -- Stevens' __TCP/IP Illustrated, Vol. 1__
has all the ugly details...

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: help for LoadModule
Date: Thu, 10 Aug 2000 13:26:14 -0500
From: Andrew Johnson 
Organization: APS, Argonne National Laboratory
Message-ID: <3992F3C6.18F9CEF7@aps.anl.gov>
References: <8mda0f$29di$1@news.cz.js.cn> <8mdsgb$8vn$1@nnrp1.deja.com> <39915AEB.EF71C6EC@eede.ericsson.se> <8mrsvn$ptp$1@nnrp1.deja.com> <3992509A.BC802787@eede.ericsson.se>

Uwe Fuhr wrote:
> 
> I am not sure if the error message
> 
> > Relocation value does not fit in 24 bits.
> 
> is really erroneous.

The message is correct, but misleading.  In PPC machine language the
offset of a long branch is specified as a 24-bit number, but that refers
to instructions not bytes (the CPU can only execute 4-byte aligned
instructions so the bottom 2 bits would always be 00).  The maximum branch
offset is thus effectively a 26-bit signed number, ie +/- 32Mb.  The
reason this shows up is that vxWorks loads itself at the bottom of memory
on bootup, but then ld puts code at the top of memory so with more than
32Mb of RAM it can't link it properly.

If you need a version of gcc that supports -mlongcall you should be able
to get WRS to provide you with one.  There are also other ways to get
around the problem, generally by reserving RAM from the top of memory (see
the comments in your BSP's sysMemTop() routine) and adding it to the
system memory pool after all your code has been loaded.

- - Andrew
- -- 
Sharks kill 10 people each year.  People kill 60,000,000 sharks each year.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Condition variables in VxWorks?
Date: Thu, 10 Aug 2000 13:28:53 -0500
From: Andrew Johnson 
Organization: APS, Argonne National Laboratory
Message-ID: <3992F465.6935FC1E@aps.anl.gov>
References: 

James Sheppard wrote:
> 
> I'm just starting to learn multi-threaded programming, so bear with me.  I
> was started off with pthreads and now I'm learning VxWorks.  I was wondering
> if there was something similiar to the condition variables found in pthreads
> that is in VxWorks.
> 
> I was trying to use binary semaphores, but that doesn't seem to work the way I
> thought they did.

Read the section on Semaphores in the vxWorks programmers guide, it covers
the different types and their uses pretty well.

- - Andrew
- -- 
Sharks kill 10 people each year.  People kill 60,000,000 sharks each year.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: tcp retransmit delays
Date: Thu, 10 Aug 2000 10:50:22 -0400
From: Dmitriy Zavin 
Message-ID: <3992C12E.6C36F033@txc.com>
References: <39918A06.50372148@here.org> <8msp40$qr$1@news.bayarea.net> <3992BA35.3D769225@here.org> <8muf53$huv$1@news.bayarea.net>



To find out what you receive window is, throw a packet sniffer on the LAN where
your vxworks machine is sitting. In the TCP packets, it has an RWIN field.
I use ethereal for linux and it shows what the windows are. Try it out.

--Dima

Kevin Lahey wrote:

In article <3992BA35.3D769225@here.org>,

Chris Smith  <cbsmithATswriDoTorg> wrote:

>

>Kevin Lahey wrote:

>

>> In article <39918A06.50372148@here.org>,

>> Chris Smith  <cbsmithATswriDOTorg> wrote:

>> >I have several vxworks processors conencted with ethernet based
tcp

>> >socket connections. Occasionally I see several packets dropped
from the

>> >network. I am seeing a ~1.5 second pause when a packet needs to
be

>> >retransmitted.

>>

>> A 1.5 second pause for retransmission is pretty reasonable, according

>> to the RFCs.

>>

>

>I can understand a 1.5 second timeout period before retransmitting
(although

>this seems long for many network environments) but not an actual pause
in

>the data flow from the vxworks board. I would have thought that even
though

>a packet may need retransmitting (and take 1.5 seconds for it to be

>retransmitted) the packet stream should not be interrupted. I can
only see

>this interruption occurring if the sender or receiver is running out
of

>buffer space. I have alloced more than enough mbuf buffers (clusters)
to

>handle this amount of data (actually over 51 seconds of data).

Hmmm.  Maybe I'm misunderstanding your situation.  The way
TCP usually

works is that it sends out more data as it receives acknowledgements
for

the old data being received by the other side.  If it stops getting
ACKs,

it should stop sending.  It'll then time out, and retransmit the
packet.

While it is timing out, it shouldn't be sending packets.[*]

>The connection window is negotiated to 61K. Do you know if the receiver
puts

>data into mbufs or does it just use the socket receive buffer (the
61K

>window)? It is possible that the receiver is running out of space,
but I do

>not see a change in the window size (like I would expect if it ran
out of

>room). The 61K window is 7 seconds of data -- that is many packets
for it to

>be behind.

I'm relatively new to VxWorks, so I'm still just figuring out this

stuff, too.  As far as I can tell, it VxWorks doesn't ever change
the

advertised window as a receiver.  I don't know how to find out
if it

is running out of all memory, although perhaps you'd get some kinda

log message to the console?

Good luck,

Kevin

kml@bayarea.net

[*]  This is a simplification -- Stevens' __TCP/IP Illustrated,
Vol. 1__

has all the ugly details...



-- 
/----------------------------------/
Dmitriy Zavin
Software Engineering Intern
TranSwitch Corporation
/----------------------------------/

 


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Does anyone use VxWorks on a standard PC?
Date: Thu, 10 Aug 2000 14:50:00 GMT
From: "Charles Magne" 
Organization: Guest of ProXad - France
Message-ID: 
References:  <3990ad22.10229036@news.swbell.net>



With the PC486  BSP , is it possible to use all the connected interfaces:
Floppy
Hard disk
Screen,
USB,
Ethernet card
CD ROM.

I d like to use also some WINDOWS PC applications in the same time.

Thanks
 Charles



coder  a écrit dans le message :
3990ad22.10229036@news.swbell.net...
> In comp.os.vxworks
> "Charles Magne"  wrote:
>
> >If you are using that,
> >I d like to know if it is necessary use an additionnal card?
> >and where to find the BSP associated.
>
> An additional card for what?  The BSP is PC486 and you find it in your
> target directory if you bought and installed it.
>
>



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: tcp retransmit delays
Date: Thu, 10 Aug 2000 13:43:57 -0500
From: Chris Smith 
Organization: Soutwest Research Institute
Message-ID: <3992F7ED.4FCED82B@here.org>
References: <39918A06.50372148@here.org> <8msp40$qr$1@news.bayarea.net> <3992BA35.3D769225@here.org> <8muf53$huv$1@news.bayarea.net> <3992C12E.6C36F033@txc.com>
Reply-To: cbsmithATswriDoTorg


- --------------644AB522BA66B40BFFFA24AB
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

I'll reply to both of you at once...  thanks for both of your comments.

Dmitriy Zavin wrote:

> To find out what you receive window is, throw a packet sniffer on the
> LAN where your vxworks machine is sitting. In the TCP packets, it has
> an RWIN field. I use ethereal for linux and it shows what the windows
> are. Try it out.

yes I am using etherpeek on a separate nt machine, and coordinating that
with the tcpDebugShow packets on the vxworks boards. The "RWIN" field is
not changing from the 61K it is initially using.

>
>
> --Dima
>
> Kevin Lahey wrote:
>
>> In article <3992BA35.3D769225@here.org>,
>> Chris Smith   wrote:
>> >
>> >Kevin Lahey wrote:
>> >
>> >> In article <39918A06.50372148@here.org>,
>> >> Chris Smith   wrote:
>> >> >I have several vxworks processors conencted with ethernet based
>> tcp
>> >> >socket connections. Occasionally I see several packets dropped
>> from the
>> >> >network. I am seeing a ~1.5 second pause when a packet needs to
>> be
>> >> >retransmitted.
>> >>
>> >> A 1.5 second pause for retransmission is pretty reasonable,
>> according
>> >> to the RFCs.
>> >>
>> >
>> >I can understand a 1.5 second timeout period before retransmitting
>> (although
>> >this seems long for many network environments) but not an actual
>> pause in
>> >the data flow from the vxworks board. I would have thought that
>> even though
>> >a packet may need retransmitting (and take 1.5 seconds for it to be
>>
>> >retransmitted) the packet stream should not be interrupted. I can
>> only see
>> >this interruption occurring if the sender or receiver is running
>> out of
>> >buffer space. I have alloced more than enough mbuf buffers
>> (clusters) to
>> >handle this amount of data (actually over 51 seconds of data).
>>
>> Hmmm.  Maybe I'm misunderstanding your situation.  The way TCP
>> usually
>> works is that it sends out more data as it receives acknowledgements
>> for
>> the old data being received by the other side.  If it stops getting
>> ACKs,
>> it should stop sending.  It'll then time out, and retransmit the
>> packet.
>> While it is timing out, it shouldn't be sending packets.[*]
>

The sender does not wait for an ack from the receiver before
transmitting the next packet.

Each packet that is received generates an ACK. If a packet is dropped
the next packet that arrives at the receiver generates a "bad" ack -- it
only acknowledges up to the missed packet but it still sends an ack
because it did get a packet. Once vxworks gets a configurable number of
"bad" acks it is supposed to trigger a fast retransmit. Or if a packet
has not been acknowledged after a certain period of time it is supposed
to retransmit it (the "slow" version). But I do not think that is
supposed to stop the data stream unless it runs out of buffer room. I
could be wrong on this point but I do not see a reason for it to
actually pause.


>>
>>
>> >The connection window is negotiated to 61K. Do you know if the
>> receiver puts
>> >data into mbufs or does it just use the socket receive buffer (the
>> 61K
>> >window)? It is possible that the receiver is running out of space,
>> but I do
>> >not see a change in the window size (like I would expect if it ran
>> out of
>> >room). The 61K window is 7 seconds of data -- that is many packets
>> for it to
>> >be behind.
>>
>> I'm relatively new to VxWorks, so I'm still just figuring out this
>> stuff, too.  As far as I can tell, it VxWorks doesn't ever change
>> the
>> advertised window as a receiver.  I don't know how to find out if it
>>
>> is running out of all memory, although perhaps you'd get some kinda
>> log message to the console?
>
I thought the window was essantially saying it is safe to send this much
data, but now that I look at it I don't think it ever changes. hmmm...

>>
>> Good luck,
>>
>> Kevin
>> kml@bayarea.net
>>
>> [*]  This is a simplification -- Stevens' __TCP/IP Illustrated, Vol.
>> 1__
>> has all the ugly details...
>
> --
> /----------------------------------/
> Dmitriy Zavin
> Software Engineering Intern
> TranSwitch Corporation
> /----------------------------------/
>
>

- --------------644AB522BA66B40BFFFA24AB
Content-Type: text/html; charset=us-ascii
Content-Transfer-Encoding: 7bit



I'll reply to both of you at once...  thanks for both of your comments.

Dmitriy Zavin wrote:

To find out what you receive window is, throw a packet
sniffer on the LAN where your vxworks machine is sitting. In the TCP packets,
it has an RWIN field. I use ethereal for linux and it shows what the windows
are. Try it out.

yes I am using etherpeek on a separate nt machine, and coordinating that
with the tcpDebugShow packets on the vxworks boards. The "RWIN" field is
not changing from the 61K it is initially using.

 

--Dima

Kevin Lahey wrote:

In article <3992BA35.3D769225@here.org>,

Chris Smith  <cbsmithATswriDoTorg> wrote:

>

>Kevin Lahey wrote:

>

>> In article <39918A06.50372148@here.org>,

>> Chris Smith  <cbsmithATswriDOTorg> wrote:

>> >I have several vxworks processors conencted with ethernet based
tcp

>> >socket connections. Occasionally I see several packets dropped
from the

>> >network. I am seeing a ~1.5 second pause when a packet needs to
be

>> >retransmitted.

>>

>> A 1.5 second pause for retransmission is pretty reasonable, according

>> to the RFCs.

>>

>

>I can understand a 1.5 second timeout period before retransmitting
(although

>this seems long for many network environments) but not an actual pause
in

>the data flow from the vxworks board. I would have thought that even
though

>a packet may need retransmitting (and take 1.5 seconds for it to be

>retransmitted) the packet stream should not be interrupted. I can
only see

>this interruption occurring if the sender or receiver is running out
of

>buffer space. I have alloced more than enough mbuf buffers (clusters)
to

>handle this amount of data (actually over 51 seconds of data).

Hmmm.  Maybe I'm misunderstanding your situation.  The way
TCP usually

works is that it sends out more data as it receives acknowledgements
for

the old data being received by the other side.  If it stops getting
ACKs,

it should stop sending.  It'll then time out, and retransmit the
packet.

While it is timing out, it shouldn't be sending packets.[*]






The sender does not wait for an ack from the receiver before transmitting
the next packet.

Each packet that is received generates an ACK. If a packet is dropped
the next packet that arrives at the receiver generates a "bad" ack -- it
only acknowledges up to the missed packet but it still sends an ack because
it did get a packet. Once vxworks gets a configurable number of "bad" acks
it is supposed to trigger a fast retransmit. Or if a packet has not been
acknowledged after a certain period of time it is supposed to retransmit
it (the "slow" version). But I do not think that is supposed to stop the
data stream unless it runs out of buffer room. I could be wrong on this
point but I do not see a reason for it to actually pause.

 



 

>The connection window is negotiated to 61K. Do you know if the receiver
puts

>data into mbufs or does it just use the socket receive buffer (the
61K

>window)? It is possible that the receiver is running out of space,
but I do

>not see a change in the window size (like I would expect if it ran
out of

>room). The 61K window is 7 seconds of data -- that is many packets
for it to

>be behind.

I'm relatively new to VxWorks, so I'm still just figuring out this

stuff, too.  As far as I can tell, it VxWorks doesn't ever change
the

advertised window as a receiver.  I don't know how to find out
if it

is running out of all memory, although perhaps you'd get some kinda

log message to the console?



I thought the window was essantially saying it is safe to send this much
data, but now that I look at it I don't think it ever changes. hmmm...



 

Good luck,

Kevin

kml@bayarea.net

[*]  This is a simplification -- Stevens' __TCP/IP Illustrated,
Vol. 1__

has all the ugly details...



-- 
/----------------------------------/
Dmitriy Zavin
Software Engineering Intern
TranSwitch Corporation
/----------------------------------/

 



- --------------644AB522BA66B40BFFFA24AB--


---------------------------

Newsgroups: comp.os.vxworks
Subject: AMD 79C973 driver (BSP Janich Klass ISA p7/103)
Date: Thu, 10 Aug 2000 15:30:09 GMT
From: juno@fokkercontrol.com
Organization: FCS
Message-ID: <8muhpo$o3h$1@nnrp1.deja.com>

I'm trying to build a BSP for the Janich Klass AT96 p7/103
board.
Adapting the pentium BSP works, but including my network
device results in some trouble.
The chipset on the AT96 board is an AMD 79C973.

I use the default "LN netif driver" from the workspace,
but lacks the corresponding parameter values.

Could somebody give me some hints...

Best Regards,
FOKKER CONTROL SYSTEMS

Juno Kampstra
Sofware Design Engineer

- -----------------------------------------------------------------------
mail address         visit/shipping address      phone: 31 20 605 2567
P.O. Box 7600        Fokkerweg 300               fax:   31 20 605 2524
1117 ZJ Schiphol     1438 AN Oude Meer           email:
The Netherlands      The Netherlands             juno@fokkercontrol.com
- -----------------------------------------------------------------------
For  WORLDWIDE SIMULATION SOLUTIONS  and the latest company information
visit our web site: http://www.fokkercontrol.com
- -----------------------------------------------------------------------
The information contained in this message is intended for the addressee
only and may contain confidential and/or privileged information. If you
are not the addressee,  or receiving this message by an error,   please
delete this message and notify the sender;
you should not copy or distribute this message or disclose its contents
to anyone.
- -----------------------------------------------------------------------



Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: IP broadcasting to 255.255.255.255 ??
Date: Thu, 10 Aug 2000 15:52:22 GMT
From: "Victor Rosu" 
Organization: Excite@Home - The Leader in Broadband
Message-ID: 

Does anyowne know how to setup an Ethernet adapter or how to use the socket
API to broadcast UDP packets with a destination address of 255.255.255.255?

If I try sendto() with the destination address set to 255.255.255.255 it
performs the operation but actually nothing comes out on the network.

ifShow displays:
ene (unit number 2):
     Flags: (0x8063) UP BROADCAST MULTICAST ARP RUNNING
     Type: ETHERNET_CSMACD
     Internet address: 20.20.20.20
     Broadcast address: 20.255.255.255
     Netmask 0xff000000 Subnetmask 0xff000000
........

What bugs me is the notion of "Broadcast address" of an adaptor which is the
result of its IP address and its netmask (in this case 20.20.20.20 and
0xff000000). Of course it is a broadcast address and it actually works but
only for the subnet you're in. What if you want to broadcast to machines
that are physically on the same subnet, but they don't have an IP address
assigned yet and they won't listen to packets sent to 20.255.255.255 ???

Apart from intercepting the ethernet frames just before they go out with a
destination address of 20.255.255.255 and changing that on-the-fly to
255.255.255.255, what else could I do???

Thanks,
    Victor.




---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: IP broadcasting to 255.255.255.255 ??
Date: Thu, 10 Aug 2000 16:16:25 GMT
From: "Victor Rosu" 
Organization: Excite@Home - The Leader in Broadband
Message-ID: 
References: 

OK, I found it on wrs.com support...

How to send broadcast datagram to 255.255.255.255
Solution ID: 16.0.638438.2492251
Solution Type: Product Solution
Solution Status: Confirmed

Goal:  How to send broadcast datagram to 255.255.255.255
Fact:  Tornado 2.0
Fact:  SENS
Fact:  VxWorks / 5.4
Fact:  Tornado / 2.0 / FCS
Symptom:  Packet sent to 255.255.255.255 is actually sent to subnet
broadcast address
Fix:  Below is some information from SPR# 8019 (which was rejected) that
explains why 255.255.255.255 gets translated to the subnet broadcast
address:

The following code fragment will send a broadcast packet with a destination
address of 255.255.255.255. It requires the creation of a routing entry
which matches that destination address. The gateway for the route entry
determines the interface that sends the broadcast. The following example
creates the necessary
entry.

- ->; mRouteEntryAdd (0xffffffff, 0x930b7497, 0xffffffff, 0, 0, 1, 0)
value = 0 = 0x0
- ->; routeShow

ROUTE NET TABLE
destination gateway flags Refcnt Use Interface
- ----------------------------------------------------------------------------
147.11.116.0 147.11.116.151 101 0 0 ln0
255.255.255.255 147.11.116.151 3 0 0 ln0
- ----------------------------------------------------------------------------

ROUTE HOST TABLE
destination gateway flags Refcnt Use Interface
- ----------------------------------------------------------------------------
127.0.0.1 127.0.0.1 5 1 4 lo0
- ----------------------------------------------------------------------------
value = 77 = 0x4d = 'M'

The entry can be removed by:

mRouteEntryDelete (0xffffffff, 0x930b7497, 0xffffffff, 0, 0, 1)

when the broadcasts are completed.

The route entry allows the route lookup for the 255.255.255.255 destination
to
succeed so that the packet is sent. Otherwise, the sendto() call fails with
EHOSTUNREACH (0x41). A default route may also be created instead of the
explicit destination shown above.

Normally, the socket layer transmission calls such as sendto() replace the
all-ones
broadcast address with the locally configured broadcast address of the
primary
(i.e - first attached) interface, which has the form <;net number>;, <;-1>;.
This
behavior occurs for convenience so that users do not need to know the
assigned
address to send broadcasts. It also enhances security since use of the
all-ones
broadcast is limited to prevent network flooding. That broadcast address
should
only be used by uninitialized interfaces using protocols like DHCP or BOOTP
to
obtain configuration information. In other cases, the standard subnet local
broadcast which is normally produced is appropriate.

The following code should only be used by customers attempting to write a
DHCP
or BOOTP client. It replaces the broadcast address of the primary interface
with
the all-ones broadcast so that the substition performed by the socket layer
still
produces the desired result. The actual interface which transmits the
broadcast
is determined by the gateway of the routing entry, but the destination
address for
that packet is always set to the broadcast address of the primary interface.

The broadcast address must be changed by directly issuing the appropriate
low-level ioctl() call because the inet_addr() conversion performed when
processing
the higher-level ifBroadcastSet() routine produces a value of -1 (ERROR)
when the
all-ones broadcast is specified. For the same reason, the mRouteEntryAdd()
routine
must be used when creating a route entry to the 255.255.255.255 address. If
a
default route is created instead, any of the routing API routines may be
used.

To completely mimic certain parts of a DHCP/BOOTP client, the IP address of
the
sending interface must be changed to 0.0.0.0 (using ifAddrSet or the
corresponding
ioctl call) at some point before the sendto() call. That step is not shown
here, but
was also tested successfully.

#include "vxWorks.h"
#include "stdio.h"
#include "sockLib.h"
#include "netinet/in.h"
#include "sys/ioctl.h"
#include "logLib.h"

#define SOCK 7000

int client_test()
{ 
char buf[512]; 
int bufLen = 512; 
struct sockaddr_in name; 
int status; 
struct ifreq ifr; 
int optval = 1; 

int fd = socket (AF_INET, SOCK_DGRAM, 0); 
if (fd == ERROR) 
{ 
perror("no socket "); 
return (ERROR); 
} 
memset (&;name, 0, sizeof (name)); 
name.sin_family = AF_INET; 
name.sin_port = htons (SOCK); 
name.sin_addr.s_addr = 0xffffffff; 
/* 
* Replace the broadcast address to prevent substitution 
* of the network number (e.g. 138.11.255.255) 
*/ 

strcpy (ifr.ifr_name, "dc0"); 
bzero ((caddr_t) &;ifr.ifr_addr, sizeof (ifr.ifr_addr)); 
ifr.ifr_addr.sa_len = sizeof (struct sockaddr_in); 
ifr.ifr_addr.sa_family = AF_INET; 
((struct sockaddr_in *) &;ifr.ifr_addr)->;sin_addr.s_addr = 0xffffffff; 
status = ioctl (fd, SIOCSIFBRDADDR, &;ifr); 
if (status != 0) 
{ 
if (status != ERROR) /* iosIoctl() can return ERROR */ 
(void)errnoSet (status); 

perror ("ioctl failed"); 
close (fd); 
return (ERROR); 
} 
setsockopt (fd, SOL_SOCKET, SO_BROADCAST, (char *)&;optval, sizeof (optval)); 

if (sendto (fd, buf, bufLen, 0, (struct sockaddr *)&;name, sizeof (name))==ERROR) 
{ 
perror(" sendto failed "); 
close (fd); 
return (ERROR); 
} 
close (fd); 
return (OK); 
} 


With the route entry and the code above the packet goes out with 255.255.255.255 as the IP destination address. However, due to the route entry the destination MAC address is the target's Ethernet address, not the Ethernet broadcast address. 

The following code can be used to insert the Ethernet broadcast address: 

#include "vxWorks.h" 
#include "logLib.h" 
#include "ifLib.h" 

BOOL myOutputHook (struct ifnet *pIf, char *buffer, int length) 
{ 
char *enet = buffer; 
char *ipaddress = buffer+30; /* Ethernet header is 14 bytes 
* Ip dest field is located 16 bytes 
* after that 
*/ 
unsigned char etbcastaddr [6] = { 0xff, 0xff, 0xff, 0xff, 0xff, 0xff }; 
logMsg("Output hook called\n", 0,0
,0,0,0,0);

if (ipaddress[0] == 0xff &;&; ipaddress[1] == 0xff
&;&; ipaddress[2] == 0xff &;&; ipaddress[3] == 0xff)
{ 
logMsg(" Destination Ethernet address is %02x:%02x:%02x:%02x:%02x:%02x\n", 
enet [0], enet [1], enet [2], enet [3], enet [4], enet[5]); 
bcopy(etbcastaddr, enet, 6); 
logMsg(" Destination Ethernet address is %02x:%02x:%02x:%02x:%02x:%02x\n", 
enet [0], enet [1], enet [2], enet [3], enet [4], enet[5]); 
} 
return (FALSE); 
}  

 
 
[Next Doc] 





---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: help for LoadModule
Date: Thu, 10 Aug 2000 16:30:19 GMT
From: Urban Lindberg 
Organization: Deja.com - Before you buy.
Message-ID: <8mular$r3l$1@nnrp1.deja.com>
References: <8mda0f$29di$1@news.cz.js.cn> <8mdsgb$8vn$1@nnrp1.deja.com> <39915AEB.EF71C6EC@eede.ericsson.se> <8mrsvn$ptp$1@nnrp1.deja.com> <3992509A.BC802787@eede.ericsson.se>

Hi Uwe!

I guess you have Tornado 1, I use Tornado 2 and the -mlongcall is in my
online manuals. Don't know how to solve this in Tornado 1 :-(
Maybe WRS have som info on this?

Regards
Urban

In article <3992509A.BC802787@eede.ericsson.se>,
  Uwe.Fuhr@eed.ericsson.se wrote:
> Hi Urban,
>
> I am not sure if the error message
>
> > Relocation value does not fit in 24 bits.
>
> is really erroneous.
>
> One think I can imagine is, that the code which is loaded via
> the ld() command can not be located/linked into the area of 24 Bits.
> This means that code must be placed within the space of 16 MB.
> If this is the fact, then I should have had the same problem
> before, when using 32 MB of RAM.
>
> Do you mean the -G option? And why do you think that helps?
>
> Regards
>
> /Uwe
> --
>
> Urban Lindberg wrote:
> >
> > Hi Uwe!
> >
> > I don't use the printed manuals, but the online manuals and there
you
> > will find it in: GNU Toolkit User's Guide/The C/C++ Compiler/IBM
RS/6000
> > and PowerPC Options. It's the last option mentioned in that section.
> > And, yes the error message is erroneous, it should be 26 bits :-)
> >
> > HTH
> > Urban
> >
> > In article <39915AEB.EF71C6EC@eede.ericsson.se>,
> >   Uwe.Fuhr@eed.ericsson.se wrote:
> > > Hi Urban,
> > >
> > > I run into the same problem as Tang, while extending the RAM from
32
> > > to 64 MB of RAM, and then loading objects via ld() to the new 64
MB
> > vxWorks
> > > load module.
> > >
> > > We are using the gcc compiler but I could not find the mentioned
> > > option -mlongcall in the manual.
> > >
> > > What about this.
> > >
> > > Best regards
> > >
> > > /Uwe
> > >
> > > --
> > >
> >
___________________________________________________________________________
> > >
> > >      _/  _/  _/  Uwe Fuhr                       EED/E/D/U
Access
> > Design
> > >     _/  _/  _/   Ericsson Eurolab Deutschland   phone  : +49-5121
- -
> > 707104
> > >    _/  _/  _/    Daimlerring 9                  FAX    : +49-5121
- -
> > 707314
> > >   _/  _/  _/     31135 Hildesheim - Germany     mobile : +49-173 -
> > 5430771
> > >
> > > __________________________________
mailto:Uwe.Fuhr@eed.ericsson.se
> > ____
> > >
> > \___________________________________/
> > >
> > > > Hi!
> > > >
> > > > You need to add the -mlongcall option to your compilation.
> > > >
> > > > HTH
> > > > Urban
> > > >
> > > > In article <8mda0f$29di$1@news.cz.js.cn>,
> > > >   "tanghaoyu"  wrote:
> > > > > Hello All,
> > > > >
> > > > > When I load an object file by calling LoadModule() or ld (1)
> > > > >  > > > >
> > > > > Relocation value does not fit in 24 bits.
> > > > > Relocation value does not fit in 24 bits
> > > > > ......
> > >
> >
> > Sent via Deja.com http://www.deja.com/
> > Before you buy.
>


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: which function can I use to get host ipaddr?
Date: Thu, 10 Aug 2000 20:40:24 GMT
From: lx1975@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8mv3vn$7ai$1@nnrp1.deja.com>

Same as subject.


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: WIN2000 SP1 & TORNADO II
Date: Thu, 10 Aug 2000 17:38:00 GMT
From: venezim@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8mup9n$udn$1@nnrp1.deja.com>
References: 

Hi Michael,

I have no clue what these errors are but I am
experiencing them as well.  I'm only just
starting to use Tornado II so I'm glad to hear
that the program seems to work fine after
accepting the error messages.

Mark V.

In article ,
  "Michael Murphy"  wrote:
> Installed Windows 2000 SP1 today. When running
Tornado II I get a bunch of
> the following errors:
>
> "RPCExit: extra call"
>
> Tornado II seems to run fine after I accept all
these errors (4 to 5 of
> them)
>
> Anybody have any clues?
>
> Thanx,
>
> Michael
>
>



Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Display of Windows NT Windows on a VXWorks Power PC
Date: Thu, 10 Aug 2000 12:54:56 -0500
From: Andrew Johnson 
Organization: APS, Argonne National Laboratory
Message-ID: <3992EC70.AB84236D@aps.anl.gov>
References: <8ms846$380$1@nnrp1.deja.com>

lpeck@my-deja.com wrote:
> 
> I have a configuration with a dumb  Power PC
> display server running VXworks that supports X
> Displays. Currently we are able to display
> windows from a Unix window on this server by
> simply setting the display variable to the
> display server. However, we would like to replace
> the Unix machine with a PC running Windows NT. Is
> there an application or a setting which will
> allow me to run my existing Windows NT apps on my
> Windows NT and display the windows for the app on
> the dumb display server?

Hmm, interesting little problem.  I'm not aware of any applications that
export a Windows frame buffer through X, although there may be one.  I
would try building an X11 VNC client for vxWorks which ought to do the
trick.  Do a web search for VNC to find it.  Of course if you can still
have the Unix machine around then you could run the VNC client on Unix,
with its display pointing to the vxWorks box - probably Ok for an interim
solution, but longer-term I'd port the client.

- - Andrew
- -- 
Sharks kill 10 people each year.  People kill 60,000,000 sharks each year.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: NTP client on VxWorks
Date: Thu, 10 Aug 2000 13:02:07 -0500
From: Andrew Johnson 
Organization: APS, Argonne National Laboratory
Message-ID: <3992EE1F.DBF8E127@aps.anl.gov>
References: <311c2be6.cfb476ec@usw-ex0106-045.remarq.com> <8msg0l$9ci$1@nnrp1.deja.com> <3991D24B.97FCA644@lmco.com>

"Bryan D. Lee" wrote:
> 
> Mind if I ask what version of vxworks this is?  I don't have it in the
> kernel and I don't see it in the 5.3 docs.
> 
> Bryan
> 
> john_94501@my-deja.com wrote:
> >
> > Add the INCLUDE_SNTPC component to the kernel, then do something like
> > this in your application:

Bryan,

Tornado 2.0 (vxWorks 5.4) includes a Simple NTP Client, but it only seems
to provide one-shot "what time is it now" or "come back when you next see
an NTP broadcast" functionality and AFAIK doesn't try to correct for
network delays or interface to the vxWorks clock.  It sounds like the
solution you're working in is a lot more comprehensive.

Wasn't there a port of xntp available for vxWorks several years ago
though?

- - Andrew
- -- 
Sharks kill 10 people each year.  People kill 60,000,000 sharks each year.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: MVME2604 Memory maps
Date: Thu, 10 Aug 2000 13:11:59 -0500
From: Andrew Johnson 
Organization: APS, Argonne National Laboratory
Message-ID: <3992F06F.92C55D19@aps.anl.gov>
References: <3991A892.81F1DBB8@mw.boeing.com>

Joe Andrews wrote:
> 
> I hope this isn't too basic, but I'm brand new to the VxWorks and VME
> world and need to get a couple of MVME2604 boards working. I can load
> VxWorks boot image and flash stuff fine, but I can for the life of me
> figure out how to set up the memory mapping on the board. All I want to
> do is map the board's memory into the A32 address space at
> 0x30000000-0x31fffff. Can anybody shed some light on what I need to do
> to make this happen?

Have a look at the config.h file in your target/config/mv26xx directory. 
You'll find some #defines in there starting VME_A32_SLV_ which define
whereabouts the local memory (for cpu#0) appears on the VMEbus.  Change
the VME_A32_SLV_BUS setting to 0x30000000 and rebuild the boot file -
voila!  Note though that locally the memory still appears at address 0. 
If your second 2604 board is in the same chassis and supposed to be able
to see the memory of the first you'll have to do some more work to make
this visible though.

- - Andrew
- -- 
Sharks kill 10 people each year.  People kill 60,000,000 sharks each year.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: MVME2604 Memory maps
Date: Thu, 10 Aug 2000 12:42:13 -0500
From: John Hagen 
Organization: Rockwell International
Message-ID: <3992E975.8D777479@collinsNOSPAM.rockwell.com>
References: <3991A7DE.C76A4225@mw.boeing.com>



Joe Andrews wrote:

> I hope this isn't too basic, but I'm brand new to the VxWorks and VME
> world and need to get a couple of MVME2604 boards working. I can load
> VxWorks boot image and flash stuff fine, but I can for the life of me
> figure out how to set up the memory mapping on the board. All I want to
> do is map the board's memory into the A32 address space at
> 0x30000000-0x31fffff. Can anybody shed some light on what I need to do
> to make this happen?
>
> Thanks,
>
> Joe Andrews

Hi Joe,

Take a look in mv2600.h for "VME MASTER WINDOW FOR A32 SPACE".
I believe what you are after is the #define VME_A32_MSTR_LOCAL 0x10000000.

I'm not sure what impact this will have with the config files and all the
macros. Remember according to WRS "Macros in mv2600.h should not be changed
by the user." So you will have to check that your windows for A16/A24/A32
are separated etc...

Side note. I'm using Tornado I here so I touch my config files by hand. I
can't tell you much if your using T2. I just assume the defines would be
kept mostly the same.

Hope that gives you a direction to search in,
John


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Testing END Driver Mulitcast
Date: Thu, 10 Aug 2000 13:11:04 -0500
From: John Hagen 
Organization: Rockwell International
Message-ID: <3992F038.3C05BB5B@collinsNOSPAM.rockwell.com>
References: 

I'll take a guess. It sounds like the routing table isn't set up correctly.
IIRC try routeShow() to find out what your table looks like. I don't know much
about multicast except the address range is 224.0.0.0- 239.255.255.255. I
assume your using a valid address and
have 'joined' the multicast group in question.

sendto (sockDesc, bufPtr, len, 0, (struct sockaddr *) &, len);

^^ is this a typo????

Routing problem is my humble guess at where to look.
John

AG wrote:

> Hi,
>
> I'm trying to setup a UDP socket for multicast and I'm using the following
> send function:
>
> sendto (sockDesc, bufPtr, len, 0, (struct sockaddr *) &, len);
>
> but it returns an errno =>   0x41 = S_errno_EHOSTUNREACH.
>
> Can anyone point out what the problem could be? Also, is there another
> simple way to test a driver+IP stack multicast capability?
> If I use "ifShow" the driver multicast flag is enabled.
>
> Thanks so much,
>
> AG


---------------------------

Newsgroups: comp.os.vxworks
Subject: Need a better FTP Server soft for VxWorks
Date: Thu, 10 Aug 2000 22:04:42 +0200
From: Ezra Koper 
Organization: Bezeq International Ltd.
Message-ID: <39930AD9.2F74F81A@nevision.net.il>

Hello All

Does anyone know were i can find a better VxWorks FTP Server software?

The FTP server software that is part of VxWorks is very basic.
It doesn't  allow to define various restrictions based on user&pass.
At the moment you log in you can do everything & access any folder.


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: OT BSP question
Date: Thu, 10 Aug 2000 19:17:09 GMT
From: richard.lee@smsc.com
Organization: Deja.com - Before you buy.
Message-ID: <8muv3g$39q$1@nnrp1.deja.com>
References: <8mumk7$s4c$1@nnrp1.deja.com>

I apologize for not having your answer, but could you please
tell me what BSP you are using? I want to purchase vxworks, but
neither the salesman nor FAE can tell me if the Generic x86 BSP
supports the 440BX chipset. Obviously yours does. I would be very
grateful for the info. Thanks.

In article <8mumk7$s4c$1@nnrp1.deja.com>,
  cferrell6421@my-deja.com wrote:
> Does anyone have ballpark Interrupt Latency timing numbers for an
Intel P3
> 800 MHz w/440BX Chip Set.  Operating under VxWorks 5.4. I could
extrapolate
> numbers for other x86 processors if you have them.  My WRS FAE is on
vacation
> and the 1-800 number for WRS support can't help.
>
> Thanks in advance.....
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.
>


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: AMD 79C973 driver (BSP Janich Klass ISA p7/103)
Date: Thu, 10 Aug 2000 19:57:01 GMT
From: Rick Pierpont 
Organization: Deja.com - Before you buy.
Message-ID: <8mv1ea$56p$1@nnrp1.deja.com>
References: <8muhpo$o3h$1@nnrp1.deja.com>

Juno,

Since that chipset isn't directly supported by the GUI project facility
(without a custom configlette) you need to make this modification in
your BSPs config.h file.

1) Exclude any END drivers in the project facility.
2) By default INCLUDE_FEI is defined in config.h.  You need to #undef
that line and #define INCLUDE_LN_97X_END instead.

Now when you build your bootrom image and your VxWorks you will have
network support for that chipset.

This is exactly the steps I had to take in order to use the same
chipset on my Pentium-based Jumptech/Adastra VNS-786 board.

Hope this helps,

- --
Rick Pierpont
Software Engineer
Becton Dickinson Biosciences
Sparks, MD 21152

In article <8muhpo$o3h$1@nnrp1.deja.com>,
  juno@fokkercontrol.com wrote:
> I'm trying to build a BSP for the Janich Klass AT96 p7/103
> board.
> Adapting the pentium BSP works, but including my network
> device results in some trouble.
> The chipset on the AT96 board is an AMD 79C973.
>
> I use the default "LN netif driver" from the workspace,
> but lacks the corresponding parameter values.
>
> Could somebody give me some hints...
>
> Best Regards,
> FOKKER CONTROL SYSTEMS
>
> Juno Kampstra
> Sofware Design Engineer
>
> ----------------------------------------------------------------------
- -
> mail address         visit/shipping address      phone: 31 20 605 2567
> P.O. Box 7600        Fokkerweg 300               fax:   31 20 605 2524
> 1117 ZJ Schiphol     1438 AN Oude Meer           email:
> The Netherlands      The Netherlands
juno@fokkercontrol.com
> ----------------------------------------------------------------------
- -
> For  WORLDWIDE SIMULATION SOLUTIONS  and the latest company
information
> visit our web site: http://www.fokkercontrol.com
> ----------------------------------------------------------------------
- -
> The information contained in this message is intended for the
addressee
> only and may contain confidential and/or privileged information. If
you
> are not the addressee,  or receiving this message by an error,
please
> delete this message and notify the sender;
> you should not copy or distribute this message or disclose its
contents
> to anyone.
> ----------------------------------------------------------------------
- -
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.
>


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Virtual memory exhausted error
Date: Thu, 10 Aug 2000 20:17:35 GMT
From: Joe Andrews 
Organization: Vehicle Managememt Systems
Message-ID: <39930DDF.BBAC8DDA@mw.boeing.com>
Reply-To: joe.andrews@mw.boeing.com
Sender: nntp@news.boeing.com (Boeing NNTP News Access)

Thanks for the help on my question about memory mapping the 2604. Now I
have another question:

I have some code that was generated from f2c that I have to compile. All
seems to work well for the most part, but on one particularly large file
with a lot of initialized arrays I get a 'virtual memory exhausted'
error from the compiler. Reading the GNU book seems to indicate that I
need to relint GCC with gmalloc.o. But it doesn't say anywhere how to do
that. My development environment is NT, and I'm using Tornado 1.01 with
PowerPC architecture. Any clues?

Thanks in advance....

Joe

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Need a better FTP Server soft for VxWorks
Date: 10 Aug 2000 21:04:00 GMT
From: johan@borksoft.xs4all.nl
Organization: Borksoft
Message-ID: <8mv5c0$2fa$1@news1.xs4all.nl>
References: <39930AD9.2F74F81A@nevision.net.il>

Ezra Koper  wrote:
: Hello All

: Does anyone know were i can find a better VxWorks FTP Server software?

: The FTP server software that is part of VxWorks is very basic.
: It doesn't  allow to define various restrictions based on user&pass.
: At the moment you log in you can do everything & access any folder.

Try the FTP version that is in the unsupported directory. This is
completely source code, with various levels. It is shipped with the DosFS2
update.

Groeten,
	Johan

- -- 
     o o o o o o o . . .   ___________________________________
    o      _____           ||    Johan Borkhuis              |
  .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
 >(________|__|_[_________]_|________________________________|
 _/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: makefile problems
Date: Thu, 10 Aug 2000 17:37:41 -0400
From: "Phillip Wang" 
Message-ID: <8mv7cn$56o$1@bob.news.rcn.net>

Hi,

I'm running into some pretty weird problems with gnu make's ifeq.  When I
run this piece of the test makefile, it goesn't get into the 'worked'
sections.  I'm using gnu make v3.76 on Win2000.  Can somebody help me?
Thanks

p.w.
phillip_wang@hotmail.com


SUBSYSTEM = $(strip $(basename $(@F)))

sds.lib: sds.h
 @echo "1: $(SUBSYSTEM)"
 @echo "2: $@"
      ifeq ($@, sds.lib)
#     ifeq ($(strip $@), sds.lib)
 @echo haha "$(SUBSYSTEM2)" - worked
      else
 @echo haha "$@" - not worked
      endif
# $(AR) $^ -o $@

      ifeq ($(SUBSYSTEM), sds)
 @echo haha "$(SUBSYSTEM2)" - worked
      else
 @echo haha "$@" - not worked
      endif

clean:
 -$(RM) *.tco
 -$(RM) sds.lib





---------------------------

Newsgroups: comp.os.vxworks
Subject: Tundra PowerSpan Evaluation Board
Date: Thu, 10 Aug 2000 14:38:40 -0700
From: Eric Monsler 
Organization: The ISP formerly known as Best
Message-ID: <399320E0.3E1F5B9A@radixtek.com>

All,

I am wondering about the PowerSpan Evaluation Board from Tundra.  They
claim that their software development environment for the board is
"competely compatible with VxWorks and Tornado."

Is anyone out there using this board with VxWorks?  Does anyone have any
experience with this board at all?

If so, I'd like to ask how difficult it was to get VxWorks up as if with
a full BSP?  Were there any hardware difficulties or anomolies with bus
contention?  What configuration are you running it in, 8260 core off or
on?

TIA,

Eric Monsler

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Testing END Driver Mulitcast
Date: Fri, 11 Aug 2000 12:21:54 +0900
From: "Sung Won Ho" 
Organization: ELIMnet Internet News
Message-ID: <8mvrop$fpl$1@news.elim.net>
References: 

Hi,

Why don't you try following statement in you code.

"toAddr" structure must be configured to your multicast group.

    #define MCASTADDR       "234.5.6.7"
    #define MCASTPORT       25000
    ....
    ....
    struct  sockaddr_in toAddr;
    int     toAddrLen;
    ....
    ....
    toAddrLen = sizeof(struct sockaddr_in);
    toAddr.sin_len = (u_char) toAddrLen;
    toAddr.sin_family = AF_INET;

    /* initialize the address to the send */
    toAddr.sin_addr.s_addr = inet_addr (MCASTADDR);
    /* initialize the port to send */
    toAddr.sin_port = htons(MCASTPORT);
    ....
    ....
    sendto (sockDesc, bufPtr, sendLen, 0, (struct sockaddr *)&toAddr,
toAddrLen);

Good luck

Sung.


"AG"  wrote in message
news:uhkk5.83$v65.71038@news.pacbell.net...
> Hi,
>
> I'm trying to setup a UDP socket for multicast and I'm using the following
> send function:
>
> sendto (sockDesc, bufPtr, len, 0, (struct sockaddr *) &, len);
>
> but it returns an errno =>   0x41 = S_errno_EHOSTUNREACH.
>
> Can anyone point out what the problem could be? Also, is there another
> simple way to test a driver+IP stack multicast capability?
> If I use "ifShow" the driver multicast flag is enabled.
>
> Thanks so much,
>
> AG
>
>



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Condition variables in VxWorks?
Date: Thu, 10 Aug 2000 23:01:17 GMT
From: john_94501@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8mvc7s$do9$1@nnrp1.deja.com>
References: 

Hi,

There are no condition variables in VxWorks by default, but there is a
WindLink partner who supplies a pthreads package for VxWorks that adds
them - check the WR web site for info on partner products...

HTH,

John...

In article ,
  James Sheppard  wrote:
> I'm just starting to learn multi-threaded programming, so bear with
me.  I
> was started off with pthreads and now I'm learning VxWorks.  I was
wondering
> if there was something similiar to the condition variables found in
pthreads
> that is in VxWorks.
>
> I was trying to use binary semaphores, but that doesn't seem to work
the way I
> thought they did.
>
> Thanks,
>
> James
>
> reply to warblade@san.rr.com
> if responding by e-mail
>
>


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: OT BSP question
Date: Thu, 10 Aug 2000 23:14:30 GMT
From: cferrell6421@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8mvd0j$e74$1@nnrp1.deja.com>
References: <8mumk7$s4c$1@nnrp1.deja.com> <8muv3g$39q$1@nnrp1.deja.com>

This will be the first time we try to use an x86 processor in a 'generic
PC' format.  We are currently have BSPs for Force6750 (PPC) and Tadpole
TP40 and TP41 (68040).  We were assured that "If it run's Windows it
will run the Generic PC/AT with a P6".  We have not ordered it yet and I
do feel a little unsure about the whole thing...



In article <8muv3g$39q$1@nnrp1.deja.com>,
  richard.lee@smsc.com wrote:
> I apologize for not having your answer, but could you please
> tell me what BSP you are using? I want to purchase vxworks, but
> neither the salesman nor FAE can tell me if the Generic x86 BSP
> supports the 440BX chipset. Obviously yours does. I would be very
> grateful for the info. Thanks.
>
> In article <8mumk7$s4c$1@nnrp1.deja.com>,
>   cferrell6421@my-deja.com wrote:
> > Does anyone have ballpark Interrupt Latency timing numbers for an
> Intel P3
> > 800 MHz w/440BX Chip Set.  Operating under VxWorks 5.4. I could
> extrapolate
> > numbers for other x86 processors if you have them.  My WRS FAE is on
> vacation
> > and the 1-800 number for WRS support can't help.
> >
> > Thanks in advance.....
> >
> > Sent via Deja.com http://www.deja.com/
> > Before you buy.
> >
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.
>


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: mkbootAta question
Date: Fri, 11 Aug 2000 10:20:43 +0530
From: Vijay Kumar Peshkar 
Organization: Ye 'Ol Disorganized NNTPCache groupie
Message-ID: <39938623.698169BB@cisco.com>
References: <1uyk5.30$2S1.72@client>
Reply-To: vpeshkar@cisco.com

This is a multi-part message in MIME format.
- --------------738E3DDC1BD3D95649A8241E
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Mike,

If you are able to see the contents of the device...then you should be able to
see other files
as well. try copying some junk files into the IDE. Then reboot the system on
VxWorks and see
if those files are visible...name, size and all. Try playing around those files
etc.,
You can also try booting from floppy and loading the image from the IDE...the
IDE being DOS
formatted and the image being copied there before hand.



cheers,
Vijay

Mike Kordik wrote:

> I am trying to get my IDE device to boot vxworks. I am booting from a floppy
> right now. I can init the IDE device and I can write files to it that can be
> loaded from the bootloader but I cannot make it boot the bootloader. When I
> run mkbootAta I get S_dosFsLib_NO_CONFIG_SPACE error. When mkbootAta is done
> running I get a /vxsysdos/ device. That device has the bootrom.sys file but
> it is zero bytes in length.
>
> Does anyone know what I am doing wrong?
>
> Thanks,
> Mike

- --------------738E3DDC1BD3D95649A8241E
Content-Type: text/x-vcard; charset=us-ascii;
 name="vpeshkar.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Card for Vijay Kumar Peshkar
Content-Disposition: attachment;
 filename="vpeshkar.vcf"

begin:vcard 
n:Kumar.P;Vijay
tel;work:080-5530035/36 Extn:3464
x-mozilla-html:FALSE
org:Wipro Technologies, K2, Bangalore.;Wishing you blue skies, a couple of daisies, some fresh air and a smile!
adr:;;;;;;
version:2.1
email;internet:vpeshkar@cisco.com
title:Senior Software Engineer
note;quoted-printable:      o   o   o   o   o   o   o   .  .  . _____________________________________________=0D=0A    o           _____                      ||  Vijay Kumar .P         =20                                                    |=0D=0A   .] [__n_n_| DD[  =3D=3D=3D=3D_____   | vpeshkar@cisco.com, vijay_peshkar@hotmail.com     | =0D=0A >(________|__|_[_________]_ |___________________________________________|=0D=0A _/oo OOOOO oo`  ooo   ooo     'o!o!o                                 =20           
                     o!o!o`
fn:Vijay Kumar.P
end:vcard

- --------------738E3DDC1BD3D95649A8241E--


---------------------------

Newsgroups: comp.os.vxworks
Subject: vxWorks task suspend
Date: Fri, 11 Aug 2000 10:04:38 +0530
From: "Krishna Kishore Reddy Sureddy" 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <8n024i$p7e$1@overload.lbl.gov>

Hi,
   Can a task in VXWORKS get suspended other than the conditions
preseted below.
   If yes please send them across as soon  as possible.

   1. Task causing exception(task level exception).
   2. Suspending a task from shell by ts( )  or taskSuspend( ) .
   3. Attaching a task to Debugger.
   4. Keeping taskSuspend(0) in the code of the task.
   5. Keeping a break point in the code of the task.

   The below case i am not sure and correct me if i am wrong
   6. When you are trying to free/delete the memory which is already
deleted.

Thanks and Regards,
S.K.K.Reddy.





---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: tcp retransmit delays
Date: 11 Aug 2000 00:14:35 GMT
From: kml@baygate.bayarea.net (Kevin Lahey)
Organization: Bayarea Internet Solutions
Message-ID: <8mvghb$sbt$1@news.bayarea.net>
References: <39918A06.50372148@here.org> <8muf53$huv$1@news.bayarea.net> <3992C12E.6C36F033@txc.com> <3992F7ED.4FCED82B@here.org>

In article <3992F7ED.4FCED82B@here.org>,
Chris Smith   wrote:
>yes I am using etherpeek on a separate nt machine, and coordinating that
>with the tcpDebugShow packets on the vxworks boards. The "RWIN" field is
>not changing from the 61K it is initially using.

Yah, tcpdump shows me the same thing.  It seems odd.

>> Hmmm.  Maybe I'm misunderstanding your situation.  The way TCP
>> usually works is that it sends out more data as it receives
>> acknowledgements for the old data being received by the other side.
>> If it stops getting ACKs, it should stop sending.  It'll then time
>> out, and retransmit the packet.  While it is timing out, it shouldn't
>> be sending packets.[*]
>
>The sender does not wait for an ack from the receiver before
>transmitting the next packet.
>
>Each packet that is received generates an ACK. If a packet is dropped
>the next packet that arrives at the receiver generates a "bad" ack -- it
>only acknowledges up to the missed packet but it still sends an ack
>because it did get a packet. Once vxworks gets a configurable number of
>"bad" acks it is supposed to trigger a fast retransmit. 

Duplicate ACK.  When you get three, you are supposed to go into
fast retransmit.  See:

	http://www.ietf.org/rfc/rfc2581.txt

for a complete explanation.

>Or if a packet
>has not been acknowledged after a certain period of time it is supposed
>to retransmit it (the "slow" version). But I do not think that is
>supposed to stop the data stream unless it runs out of buffer room. I
>could be wrong on this point but I do not see a reason for it to
>actually pause.

The TCP stack is supposed to maintain a separate "congestion window"
which is an estimate of how much data can be supported by the particular
connection.  If the connection window runs out, sends should stop,
even if the advertised window is still sufficient.

Also, if you are in fast retransmit (you've already seen three dup acks),
the congestion window gets shrunk pretty agressively, so you won't see
a whole lot of transmissions during a timeout after a fast 
retransmission.

Good luck,

Kevin
kml@bayarea.net

---------------------------

Newsgroups: comp.os.vxworks
Subject: BSP for mvme374
Date: Fri, 11 Aug 2000 09:53:57 +0200
From: Signaal Hamburg 
Organization: Hollandse Signaalapparaten GmbH (Hamburg)
Message-ID: <3993B115.8B5605AC@CompuServe.com>

Hi outthere,

has enyone already written an BSP for the Motorola VME board
MVME374 ? It's yust an old board, MC68020 based and has an
AM7990 Ethernet controller.

Any comment would be bery useful.

Thanx
Roger


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: makefile problems
Date: 11 Aug 2000 08:33:49 GMT
From: johan@borksoft.xs4all.nl
Organization: Borksoft
Message-ID: <8n0dpd$jv$2@news1.xs4all.nl>
References: <8mv7cn$56o$1@bob.news.rcn.net>

Phillip Wang  wrote:
: I'm running into some pretty weird problems with gnu make's ifeq.  When I
: run this piece of the test makefile, it goesn't get into the 'worked'
: sections.  I'm using gnu make v3.76 on Win2000.  Can somebody help me?

: sds.lib: sds.h
:  @echo "1: $(SUBSYSTEM)"
:  @echo "2: $@"
:       ifeq ($@, sds.lib)

The problem is that the ifeq is evaluated when the Makefile is read, and
not when it is executed. It is similar to #if-statement in a C-file: you
cannot use variable is C preprocessor statements

Groeten,
	Johan

- -- 
     o o o o o o o . . .   ___________________________________
    o      _____           ||    Johan Borkhuis              |
  .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
 >(________|__|_[_________]_|________________________________|
 _/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: Problem white read of A24 (MVME2305)
Date: Fri, 11 Aug 2000 10:52:12 +0200
From: "T. Casteleijn" 
Organization: NLR
Message-ID: <3993BEBC.D293CE4A@nlr.nl>
References: <8kj2fm$4la$1@news.monmouth.com> <8kmom3$2ple$1@news.cz.js.cn>
Reply-To: castelyn@nlr.nl

We have the following problem if we want to read from the A24 VME space the
following problem occurs, the PPC board hang's and only a hard reset can wake
the PPC up.

We use the following hard and software;

- -    processor board MVME2305 of motorola (PPC)
- -    DIO XVME-244 of XYCOM
- -    OS VxWorks version 5.4

The problem also occurs with other salve board's.


---------------------------

Newsgroups: comp.os.vxworks
Subject: mvme1604 to 2604 or 2700
Date: Thu, 10 Aug 2000 22:20:49 -0400 (EDT)
From: pdirienzo@webtv.net (Patrick DiRienzo)
Organization: WebTV Subscriber
Message-ID: <13315-39936301-103@storefull-284.iap.bryant.webtv.net>

Has anyone gone from a mvme1604 windriver bsp to a mvme2604 or mvme2700
bsp?  Has there been problems or much work involved?


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Does anyone use VxWorks on a standard PC?
Date: Fri, 11 Aug 2000 09:02:29 +0530
From: Vijay Kumar Peshkar 
Organization: Ye 'Ol Disorganized NNTPCache groupie
Message-ID: <399373CD.8A80CCD0@cisco.com>
References:  <3990ad22.10229036@news.swbell.net> 
Reply-To: vpeshkar@cisco.com

This is a multi-part message in MIME format.
- --------------126A5F80C231DEB22170A9DB
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit

I have been using Floppy, Hard disk,Screen,Ethernet Card..for long time now.
Never needed CD or
USB. They work fine. The PC boots up with vxworks image...and you cannot run
Windows from that...
(think that's what you meant...right?). But yes a standard PC can be used as
target. In my setup..target
boots from floppy...load the image either from floopy, or Host(thru serial
port or thru ethernet)...or load an image lying on my target hard disk.

Hope this helps.

bye,
Vijay

Charles Magne wrote:

> With the PC486  BSP , is it possible to use all the connected interfaces:
> Floppy
> Hard disk
> Screen,
> USB,
> Ethernet card
> CD ROM.
>
> I d like to use also some WINDOWS PC applications in the same time.
>
> Thanks
>  Charles
>
> coder  a écrit dans le message :
> 3990ad22.10229036@news.swbell.net...
> > In comp.os.vxworks
> > "Charles Magne"  wrote:
> >
> > >If you are using that,
> > >I d like to know if it is necessary use an additionnal card?
> > >and where to find the BSP associated.
> >
> > An additional card for what?  The BSP is PC486 and you find it in your
> > target directory if you bought and installed it.
> >
> >

- --------------126A5F80C231DEB22170A9DB
Content-Type: text/x-vcard; charset=us-ascii;
 name="vpeshkar.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Card for Vijay Kumar Peshkar
Content-Disposition: attachment;
 filename="vpeshkar.vcf"

begin:vcard 
n:Kumar.P;Vijay
tel;work:080-5530035/36 Extn:3464
x-mozilla-html:FALSE
org:Wipro Technologies, K2, Bangalore.;Wishing you blue skies, a couple of daisies, some fresh air and a smile!
adr:;;;;;;
version:2.1
email;internet:vpeshkar@cisco.com
title:Senior Software Engineer
note;quoted-printable:      o   o   o   o   o   o   o   .  .  . _____________________________________________=0D=0A    o           _____                      ||  Vijay Kumar .P         =20                                                    |=0D=0A   .] [__n_n_| DD[  =3D=3D=3D=3D_____   | vpeshkar@cisco.com, vijay_peshkar@hotmail.com     | =0D=0A >(________|__|_[_________]_ |___________________________________________|=0D=0A _/oo OOOOO oo`  ooo   ooo     'o!o!o                                 =20           
                     o!o!o`
fn:Vijay Kumar.P
end:vcard

- --------------126A5F80C231DEB22170A9DB--


---------------------------

End of New-News digest
**********************


From vxwexplo-errs@csg.lbl.gov  Fri Aug 11 05:06:03 2000
From: "Keith W. Muzzioli" 
Date: Fri Aug 11 05:06:05 PDT 2000
Subject: serial communication

We are having a problem with some serial communications running either VxWorks 
5.1.1 or VxWorks 5.3, using a DY4 DMV-163 68040 board with a DMAX227 (Raytheon 
version of the 220) mezzanine card.

We are sending packets of data from the PC to the DMV-163 card
of varying lengths, and started noticing errors with large packets.  If
the packet size is greater than 640 Bytes  (for example 740 in our
case), data gets "jumbled".  By jumbled, I mean that the first 640 bytes
come through just fine, but then the next 48 bytes are skipped, and the
LAST 64 bytes are sent next, followed by the skipped 48 bytes.  No
data is actually LOST, just mixed up a little...

Sent =>  < 740 Bytes >
Rcvd => < First 640 Bytes> < Last 64 Bytes > < Middle 48 Bytes >

Some comments and observations:

1) We have verified that the order is correct coming out of the PC.
2) The PC writes one byte at a time.
3) The code on the 163 card reads one byte at a time (with calls
that use the read() function).
4) As required, the memory buffer that we submit to the read() function
is on the DMAX card.
5) The failure is not 100% of the time.  It may or may not happen.  Not a
hard fault, but very frequent.
6) We have extended the memory buffer size to > our maximum data size.
7) In fact, if we pause the reading until after all writing has been completed,
we are able to read the data with no errors, so we know we are not over-
writing a maximum buffer size.
8) Also, if we slow down the rate at which data is written, it does not
fail.
9) ie: Only when writing and reading are simultaneous does it break.
10) This happens on more than one PC and 163 combination.
11) We do not use timeouts.  This is asynchronous, thus the read()
call is pending on each byte request.
12) Baud rate is 19200.

Has anyone seen behavior like this before??? Any suggestions would be greatly 
appreciated.

TIA,
Keith



From vxwexplo-errs@csg.lbl.gov  Fri Aug 11 07:21:49 2000
From: Jim Way 
Date: Fri Aug 11 07:21:51 PDT 2000
Subject: RE: which function can I use to get host ipaddr?

lx1975,

There's probably a number of ways to do it. One way is to examine the
sysBootLine string. Here's a shell example:

	-> printf sysBootLine
	cpm(0,0)jway:vxWorks e=172.16.18.23:ffff0000 h=172.16.16.166
g=172.16.16.1 u=remoteftp pw= 
	tomahawk f=0x8 tn=rot21value = 113 = 0x71 = 'q'

The e= value is the target ip and mask; the h= value is the host.

In code, this snippet seems to work ok:

	extern char *sysBootLine;
	BOOT_PARAMS BootParams;
	if (*bootStringToStruct (sysBootLine, &BootParams) == 0)
	{
		// Get my IP address. (see documentation for field name
associated with host IP address).
		sscanf (BootParams.ead, "%d.%d.%d.%d", &l1, &l2, &l3, &l4);
	}

Hope this helps.

Jim

>	Newsgroups: comp.os.vxworks
>	Subject: which function can I use to get host ipaddr?
>	Date: Thu, 10 Aug 2000 20:40:24 GMT
>	From: lx1975@my-deja.com
>	Organization: Deja.com - Before you buy.
>	Message-ID: <8mv3vn$7ai$1@nnrp1.deja.com>

-----------------------------------------
Jim Way, Software Engineer
Datum Austin (Austron Inc.)
voice: 512.721.4170
fax  : 512.990.9712
email: jwayATdatumDOTcom (no spam please)
-----------------------------------------


From vxwexplo-errs@csg.lbl.gov  Fri Aug 11 08:25:09 2000
From: "Mike Anderson" 
Date: Fri Aug 11 08:25:11 PDT 2000
Subject: RE: serial communication

VxWorks Greetings!

> 
> We are having a problem with some serial communications running 
> either VxWorks 
> 5.1.1 or VxWorks 5.3, using a DY4 DMV-163 68040 board with a 
> DMAX227 (Raytheon 
> version of the 220) mezzanine card.
> 


> data is actually LOST, just mixed up a little...
> 
> Sent =>  < 740 Bytes >
> Rcvd => < First 640 Bytes> < Last 64 Bytes > < Middle 48 Bytes >
> 


> 9) ie: Only when writing and reading are simultaneous does it break.
> 10) This happens on more than one PC and 163 combination.
> 11) We do not use timeouts.  This is asynchronous, thus the read()
> call is pending on each byte request.
> 12) Baud rate is 19200.
> 
> Has anyone seen behavior like this before??? Any suggestions 
> would be greatly 
> appreciated.

It sounds like there is some sort of preemption problem going
on here.  What are the relative priorities of the tasks doing 
the reading/writing?  Are you using read/write or fread/fwrite?
How did you open the serial port?  Are you using the standard
TyCo driver or one of your own making?

Inquiring minds want to know...

Regards,

Mike Anderson
The PTR Group, Inc.
BSPs, Device Drivers, Systems Architecture, custom training
http://www.theptrgroup.com
mailto:mike@theptrgroup.com


From vxwexplo-errs@csg.lbl.gov  Fri Aug 11 08:54:17 2000
From: David Laight 
Date: Fri Aug 11 08:54:20 PDT 2000
Subject: NFS problems with windows

Some of our customers are seeing problems using the OnmiNFS windows client (on 
NT4) talking to a vxWorks server (in this case vxWorks 5.3.1).

I was wondering if anyone else has seen anything similar - or even who knows 
enough about the NFS protocol to know who is at fault.

NFS is being used to update the filestore held on the vxWorks in flash.  (This 
isn't TFFS, but has a caching layer to avoid the works effects of the 128k 
read-modify-write sequence required to modify a sector.)

Attempts to do recursive copies apparantly lock up and corrupt the vxWorks 
filesystem.  I see no such problems with my systems that use a solaris host - so 
am tempted to blame the PC NFS client.

For a small file snoop reports the following sequence:

123   0.00025 windows -> vxWorks NFS C CREATE2 FH=E53F 1.jpg
124   0.00871 vxWorks -> windows NFS R CREATE2 OK FH=E53E
125   0.00045 windows -> vxWorks NFS C READDIR2 FH=E53F Cookie=0
126   0.00677 vxWorks -> windows NFS R READDIR2 OK 3 entries (No more)
127   0.00020 windows -> vxWorks NFS C LOOKUP2 FH=E53F 1.jpg
128   0.00635 vxWorks -> windows NFS R LOOKUP2 OK FH=E53E
129   0.00019 windows -> vxWorks NFS C LOOKUP2 FH=E53F 1.jpg
130   0.00836 vxWorks -> windows NFS R LOOKUP2 OK FH=E53E
131   0.00020 windows -> vxWorks NFS C READ2 FH=E53E at 0 for 4096
132   0.00614 vxWorks -> windows NFS R READ2 OK (0 bytes)
133   0.00148 windows -> vxWorks NFS C WRITE2 FH=E53E at 0 for 2195
134   0.00072 windows -> vxWorks UDP continuation ID=53091
135   0.49994 windows -> vxWorks NFS C WRITE2 FH=E53E at 0 for 2048
136   0.00061 windows -> vxWorks UDP continuation ID=53347
137   0.30471 vxWorks -> windows NFS R WRITE2 OK
138   0.00418 vxWorks -> windows NFS R WRITE2 OK
139   0.00032 windows -> vxWorks NFS C WRITE2 FH=E53E at 2048 for 147
140   0.00735 vxWorks -> windows NFS R WRITE2 OK
141   0.00025 windows -> vxWorks NFS C LOOKUP2 FH=E53F 1.jpg
142   0.00609 vxWorks -> windows NFS R LOOKUP2 OK FH=E53E
143   0.00019 windows -> vxWorks NFS C LOOKUP2 FH=E53F 1.jpg
144   0.00637 vxWorks -> windows NFS R LOOKUP2 OK FH=E53E
145   0.00022 windows -> vxWorks NFS C SETATTR2 FH=E53E
146   0.00689 vxWorks -> windows NFS R SETATTR2 OK
147   0.00018 windows -> vxWorks NFS C GETATTR2 FH=E53E
148   0.00559 vxWorks -> windows NFS R GETATTR2 OK
149   0.02517 windows -> vxWorks NFS C READDIR2 FH=F2EA Cookie=0
150   0.00754 vxWorks -> windows NFS R READDIR2 OK 6 entries (No more)

Note that the write (frame 133) is timed out and retransmitted (frame 135) after 
a 1/2 second (or so) timeout.  The vxWorks system is probably erasing a flash 
sector.  I'm surprised this timeout is quite that small.

There is also a read request (frame 131), inspection of larger files indicates 
this is an attempt to read the last 4k block of the file.

Frame 145 sets the file owner - or rather tries to as the vxWorks filesystem 
doesn't support the relevant attributes. This is rather a good job as what it 
actually sets is:

NFS:  Proc = 2 (Set file attributes)
NFS:  File handle = CE2B3CC100000000027615A200000000
NFS:                00000000000000000000000000000000
NFS:  Mode = 037777777777
NFS:   Type = ?
NFS:   Setuid = 1, Setgid = 1, Sticky = 1
NFS:   Owner's permissions = rwx
NFS:   Group's permissions = rwx
NFS:   Other's permissions = rwx
NFS:  UID = -1
NFS:  GID = -1
NFS:  Size = 2195
NFS:  Access time       = 28-Jun-00 17:01:54.000000 GMT
NFS:  Modification time = 28-Jun-00 17:01:54.000000 GMT

I'm sure that the setuid, setgid and sticky bits should not be set.

Anyway sequences like the one above work.

The problems arise with much larger files.  What we then get is:

1909   0.00027 windows -> vxWorks NFS C LOOKUP2 FH=E6DB UserGuide.pdf
1910   0.00647 vxWorks -> windows NFS R LOOKUP2 OK FH=E6DA
1911   0.00021 windows -> vxWorks NFS C LOOKUP2 FH=E6DB UserGuide.pdf
1912   0.00645 vxWorks -> windows NFS R LOOKUP2 OK FH=E6DA
1913   0.00134 windows -> vxWorks NFS C WRITE2 FH=E6DA at 0 for 4096
1914   0.00122 windows -> vxWorks UDP continuation ID=7527
1915   0.00106 windows -> vxWorks UDP continuation ID=7527
1916   0.03302 vxWorks -> windows NFS R WRITE2 OK
(load of writes)
2623   0.00125 windows -> vxWorks NFS C WRITE2 FH=E6DA at 323584 for 4096
2624   0.00123 windows -> vxWorks UDP continuation ID=19816
2625   0.00106 windows -> vxWorks UDP continuation ID=19816
2626   0.02605 vxWorks -> windows NFS R WRITE2 OK
2627   0.00095 windows -> vxWorks NFS C READ2 FH=E6DA at 1351680 for 4096
2628   0.00605 vxWorks -> windows NFS R READ2 OK (0 bytes)
2629   0.00182 windows -> vxWorks NFS C WRITE2 FH=E6DA at 327680 for 4096
(loads more writes)
4412   0.00134 windows -> vxWorks NFS C WRITE2 FH=E6DA at 1310720 for 4096
4413   0.00123 windows -> vxWorks UDP continuation ID=1131
4414   0.00106 windows -> vxWorks UDP continuation ID=1131
4415   0.00014 windows -> vxWorks NFS C READDIR2 FH=E6DB Cookie=0
4416   0.02707 vxWorks -> windows NFS R WRITE2 OK
4417   0.00129 windows -> vxWorks NFS C WRITE2 FH=E6DA at 1314816 for 4096
4418   0.00123 windows -> vxWorks UDP continuation ID=1643
4419   0.00106 windows -> vxWorks UDP continuation ID=1643
4420   0.00084 vxWorks -> windows NFS R READDIR2 OK 3 entries (No more)
4421   0.00021 windows -> vxWorks NFS C LOOKUP2 FH=E6DB UserGuide.pdf
4422   0.02604 vxWorks -> windows NFS R WRITE2 OK
4423   0.00129 windows -> vxWorks NFS C WRITE2 FH=E6DA at 1318912 for 4096
4424   0.00123 windows -> vxWorks UDP continuation ID=2155
4425   0.00106 windows -> vxWorks UDP continuation ID=2155
4426   0.00084 vxWorks -> windows NFS R LOOKUP2 OK FH=E6DA
4427   0.00021 windows -> vxWorks NFS C LOOKUP2 FH=E6DB UserGuide.pdf
4428   0.02591 vxWorks -> windows NFS R WRITE2 OK
4429   0.00129 windows -> vxWorks NFS C WRITE2 FH=E6DA at 1323008 for 4096
4430   0.00123 windows -> vxWorks UDP continuation ID=2667
4431   0.00106 windows -> vxWorks UDP continuation ID=2667
4432   0.00086 vxWorks -> windows NFS R LOOKUP2 OK FH=E6DA
4433   0.00022 windows -> vxWorks NFS C SETATTR2 FH=E6DA
4434   0.05144 vxWorks -> windows NFS R WRITE2 OK
4435   0.00129 windows -> vxWorks NFS C WRITE2 FH=E6DA at 1327104 for 4096
4436   0.00123 windows -> vxWorks UDP continuation ID=3179
4437   0.00108 windows -> vxWorks UDP continuation ID=3179
4438   0.00008 vxWorks -> windows NFS R SETATTR2 I/O error
4439   0.02607 vxWorks -> windows NFS R WRITE2 OK
(6 more writes)

Here the read (frame 2627) and readdir-lookup-setattr sequence (frames 4415, 
4421, 4423 and 4433) are all done DURING the sequence of writes.  The setattr is 
rejected, maybe because it specifies the full file size but the actual file 
isn't that big (yet).

This looks to me very much as though the PC has two control threads running in 
parallel (or is not allowing for the asynchronous nature of the writes) and so 
is generating requests out of sequence.

Has anyone seen anything similar?

(Note this is not dosFs 2.  dosfs 2 is not available for vxWorks 5.3.1, also I 
don't know how its cache layer would interact with the one we have to have to 
get non-zero performance from flash memory.  I'm also wary of dosfs2's ability 
to allocate NFS file handles that are preserved across system reloads (dosfs 1 
saves these in the directory entry).

	David

----------------------------------------------------------------
David Laight                          email: dsl@tadpole.co.uk
Tadpole Technology plc                phone: +44 1223 428 232
Cambridge, UK                         fax:   +44 1223 428 201



From vxwexplo-errs@csg.lbl.gov  Fri Aug 11 09:51:40 2000
From: David Laight 
Date: Fri Aug 11 09:51:42 PDT 2000
Subject: RE: which function can I use to get host ipaddr?

> >	Subject: which function can I use to get host ipaddr?

If you know the interface name just call ifAddrGet().

Getting the list of interface names may be more difficult.

The vxWorks ifShow() routines checks each entry in the ifnet table.

	David

----------------------------------------------------------------
David Laight                          email: dsl@tadpole.co.uk
Tadpole Technology plc                phone: +44 1223 428 232
Cambridge, UK                         fax:   +44 1223 428 201



From vxwexplo-errs@csg.lbl.gov  Fri Aug 11 10:35:49 2000
From: David Laight 
Date: Fri Aug 11 10:35:52 PDT 2000
Subject: re: TCP retransmittions and lost packets

> >Has anyone figured out how to get around, or scale down the time it
> >takes a tcp packet to be retransmitted?
> >
> >I have several vxworks processors conencted with ethernet based tcp
> >socket connections. Occasionally I see several packets dropped from the
> >network. I am seeing a ~1.5 second pause when a packet needs to be
> >retransmitted.
>

1.5 seconds is (as other people have said) quite normal, as is the reported 
recovery sequence.  The fact you see a gap in in the transmittion may be because 
the sending TCP stack doesn't have any more data buffered (I have a feeling that 
vxworks has a send-buffer size which can be configured/set).  On the Unix 
Streams based TCP stacks I am more familiar with the sending side always has 
plenty of data available.

What you really need to look at is why packets are being dropped.

Although TCP does recover from lost packets, on a local network it shouldn't 
have to.  Indeed NFS works only because packets are not lost (It actually works 
faster because it uses UDP not TCP because it then has absolute control of when 
acknowlegements are sent).

Places packets can go amiss:

1) ethernet cards with insufficient receive buffer space
2) ethernet cards with a slow bus interface
3) badly terminated and/or incorrectly grounded coax cable
4) AUI connectors that are not firmly home
5) ethernet repeater sets connecter to transceivers with SQE enabled
6) store-forward switches (and routers) with insufficient buffer space
7) cut though switches that discard packets if the destination segment is busy
8) links set full duplex at one end
9) overlong cable segments
10) any cards with the fujitsu 'NICE' chipset
11) thin ethernet cabling in general (bad joints etc)
12) broken device drivers

ALL ISA cards fall foul of either 1 or 2.
Sbus bus master cards fail at (2) on multi-cpu systems
PCI fast ethernet cards require the host (north) bridge to merge requests
A surprising number of early ethernet switches suffer from 7
Full duplex is best only used if enabled by auto-negotiation, even then the 
device driver and transceiver have to keep in sync.

	David


----------------------------------------------------------------
David Laight                          email: dsl@tadpole.co.uk
Tadpole Technology plc                phone: +44 1223 428 232
Cambridge, UK                         fax:   +44 1223 428 201



From vxwexplo-errs@csg.lbl.gov  Fri Aug 11 11:30:32 2000
From: "Keith W. Muzzioli" 
Date: Fri Aug 11 11:30:34 PDT 2000
Subject: RE: serial communication

> 
> VxWorks Greetings!
> 
> > 
> > We are having a problem with some serial communications running 
> > either VxWorks 
> > 5.1.1 or VxWorks 5.3, using a DY4 DMV-163 68040 board with a 
> > DMAX227 (Raytheon 
> > version of the 220) mezzanine card.
> > 
> 
> 
> > data is actually LOST, just mixed up a little...
> > 
> > Sent =>  < 740 Bytes >
> > Rcvd => < First 640 Bytes> < Last 64 Bytes > < Middle 48 Bytes >
> > 
> 
> 
> > 9) ie: Only when writing and reading are simultaneous does it break.
> > 10) This happens on more than one PC and 163 combination.
> > 11) We do not use timeouts.  This is asynchronous, thus the read()
> > call is pending on each byte request.
> > 12) Baud rate is 19200.
> > 
> > Has anyone seen behavior like this before??? Any suggestions 
> > would be greatly 
> > appreciated.
> 
> It sounds like there is some sort of preemption problem going
> on here.  What are the relative priorities of the tasks doing 
> the reading/writing?  Are you using read/write or fread/fwrite?
> How did you open the serial port?  Are you using the standard
> TyCo driver or one of your own making?
> 
> Inquiring minds want to know...

Thanks for your inquiring mind, let me try to clarify our situation...

The sending task is running on a PC that is running LYNX 3.0.1. We are using the 
OS select and write to write the data to the serial line.

The receiving task is running on the DY4 DMV-163 68040 board. There are 2 serial 
ports directly on the DMV-163 that are using the standard TyCo driver. We are 
NOT using these ports for this serial communication. There are four additional 
serial ports on the DMAX-227 daughter card, it is one of these that we are 
using. These four ports are using the DLCS (Data Link Communications Suite) 
device driver provided with our BSP from DY4. We initialized and opened the 
serial ports according to the documentation provided with the device driver. We 
are reading with just read, NOT fread.

I don't know how important the priorities are since both tasks are running under 
different OS, but the receiving task has a priority of 100 and is the only 
application task running at the time. I don't know the priority of the sending 
task but it is also the only application task running at the time.

The device driver documentation states the following:
The driver conforms to the conventional VxWorks I/O system interface and that 
the driver is accessed using the basic VxWorks I/O calls, open(), close(), 
read(), write(), and ioctl(). There are additional DLCS specific routines used 
in setting up the ports, etc, which are not passed to the I/O system. The driver 
is generic and has no knowledge of any device firmware specifics. It uses the 
interface provided by the VxCCB module to access the device firmware specific 
interfaces.

Thanks again,
Keith



From vxwexplo-errs@csg.lbl.gov  Fri Aug 11 11:35:47 2000
From: "Martin Mayer" 
Date: Fri Aug 11 11:35:49 PDT 2000
Subject: Tornado project facility

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_01C003C1.C1E0CAC0
Content-Type: text/plain;
	charset="iso-8859-1"

Hi,

I have modified my config.h file to force the default boot line and
configured the IP address that vxWorks will use. I have then rebuilt the
BSP.

This works fine when I create a new project based on this BSP. The issue I
am having is that existing projects do not pick up changes I make in the
BSP. I need to delete the project I have and create a new one. This is bad
because I have to reselect all of the vxWorks components everytime I modify
the config.h file and rebuild the BSP...

Does anyone have an idea about how I can make existing projects pickup the
changes in config.h?

Thanks,
Martin Mayer
IP Comms Software Designer
Nortel Networks


------_=_NextPart_001_01C003C1.C1E0CAC0
Content-Type: text/html;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable






Tornado project facility




Hi,




I have modified my config.h file to =
force the default boot line and configured the IP address that vxWorks =
will use. I have then rebuilt the BSP.



This works fine when I create a new =
project based on this BSP. The issue I am having is that existing =
projects do not pick up changes I make in the BSP. I need to delete the =
project I have and create a new one. This is bad because I have to =
reselect all of the vxWorks components everytime I modify the config.h =
file and rebuild the BSP...



Does anyone have an idea about how I =
can make existing projects pickup the changes in config.h?




Thanks,

Martin Mayer

IP Comms Software Designer

Nortel Networks





------_=_NextPart_001_01C003C1.C1E0CAC0--


From vxwexplo-errs@csg.lbl.gov  Sat Aug 12 03:26:34 2000
From: Thomas Martin 
Date: Sat Aug 12 03:26:37 PDT 2000
Subject: Re: vxWorks task suspend

Dies ist eine mehrteilige Nachricht im MIME-Format.
--------------2F5B8847A7A14E2D0DD26929
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Hi,

the vxWorks Users Group Exploder wrote:
> 
> Submitted-by vxwexplo-errs@csg.lbl.gov  Thu Aug 10 21:33:27 2000
> Submitted-by: "Krishna Kishore Reddy Sureddy" 
> 
> Hi,
>    Can a task in VXWORKS get suspended other than the conditions
> preseted below.
>    If yes please send them across as soon  as possible.
> 
>    1. Task causing exception(task level exception).
>    2. Suspending a task from shell by ts( )  or taskSuspend( ) .
>    3. Attaching a task to Debugger.
>    4. Keeping taskSuspend(0) in the code of the task.
>    5. Keeping a break point in the code of the task.
> 
>    The below case i am not sure and correct me if i am wrong
>    6. When you are trying to free/delete the memory which is already
> deleted.
You are right!
free() and malloc() have default debugging options. The default for
free() is (if a block is corrupted) suspend the task and log an
error message. Options can be:
MEM_ALLOC_ERROR_LOG_FLAG
MEM_ALLOC_ERROR_SUSPEND_FLAG
MEM_BLOCK_CHECK
MEM_BLOCK_CHECK_LOG_FLAG
MEM_BLOCK_CHECK_SUSPEND_FLAG
These option can be set by memOptionsSet() and can be bitwise ORed.

HTH
Thomas


-- 
MfG / kind regards

Thomas MARTIN 
*******************************************************************
* Ingenieurbuero MARTIN            * Zwei Dinge sind unendlich:   *
* Schieszstaettstrasse 5a          * das Universum und die mensch-*
* D-82515 Wolfratshausen - Germany * liche Dummheit, aber bei dem *
* Phone/FAX: +49(0)8171/28445      * Universum bin ich mir noch   *
* mobile: +49(0)171/3454864        * nicht ganz sicher.           *
* email:ib-martin@topmail.de       *                              *
* http://ib-martin.ilo.de          *             [Albert Einstein]*
*******************************************************************
--------------2F5B8847A7A14E2D0DD26929
Content-Type: text/x-vcard; charset=us-ascii;
 name="ing.buero.martin.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Karte für Thomas Martin
Content-Disposition: attachment;
 filename="ing.buero.martin.vcf"

begin:vcard 
n:Martin;Thomas
tel;pager:+49(0)171/3454864
tel;cell:+49(0)171/3454864
tel;fax:+49(0)8171/28445
tel;home:+49(0)8171/28445
tel;work:+49(0)8171/28445
x-mozilla-html:FALSE
url:http://ib-martin.ilo.de/
org:Ingenieurbuero Martin;F&E
adr:;;Schiessstaettstrasse 5a;Wolfratshausen;Bavaria;D-82515;Germany
version:2.1
email;internet:ib-martin@topmail.de
title:Dipl.-Ing.
x-mozilla-cpt:;8128
fn:Thomas Martin
end:vcard

--------------2F5B8847A7A14E2D0DD26929--





From vxwexplo-errs@csg.lbl.gov  Sat Aug 12 04:03:10 2000
From: Vxworks Exploder 
Date: Sat Aug 12 04:03:12 PDT 2000
Subject: comp.os.vxworks newsdigest

Comp.Os.Vxworks Daily Digest    Sat Aug 12 04:03:05 PDT 2000

        Subject: Re: muxReceive
        Subject: serial communication
        Subject: Problems with VxWorks FTP server. 
        Subject: Re: OT BSP question
        Subject: Re: vxWorks task suspend
        Subject: mv2700 LM/SIG and LM address
        Subject: Re: Problems with VxWorks FTP server.
        Subject: muxReceive
        Subject: Re: NTP client on VxWorks
        Subject: RE: which function can I use to get host ipaddr?
        Subject: RE: serial communication
        Subject: NFS problems with windows
        Subject: Re: Need a better FTP Server soft for VxWorks
        Subject: RE: which function can I use to get host ipaddr?
        Subject: Re: dec21x4xEnd driver
        Subject: re: TCP retransmittions and lost packets
        Subject: MVME2305 will not read proper from A24 slave
        Subject: Re: OT BSP question
        Subject: Re: MVME2305 will not read proper from A24 slave
        Subject: Re: tcp retransmit delays
        Subject: Re: NTP client on VxWorks
        Subject: RE: serial communication
        Subject: Tornado project facility
        Subject: Raw IP sockets and port number
        Subject: Re: NTP client on VxWorks
        Subject: Re: tcp retransmit delays
        Subject: Re: Display of Windows NT Windows on a VXWorks Power PC
        Subject: Re: Display of Windows NT Windows on a VXWorks Power PC
        Subject: Re: select() problem
        Subject: Re: WIN2000 SP1 & TORNADO II
        Subject: Re: TCP retransmittions and lost packets -- New Question
        Subject: Re: Tornado project facility
        Subject: END Driver In Promiscuous Mode
        Subject: Re: select() problem

-------------------------------------------------------

Newsgroups: comp.os.vxworks
Subject: Re: muxReceive
Date: Fri, 11 Aug 2000 16:32:01 +0530
From: Yusuf Motiwala 
Organization: Lucent Technologies, Columbus, Ohio
Message-ID: <3993DD29.F65B8B8E@lucent.com>
References: <8n073i$lpn$1@news.elim.net>

Yes you can but you should provide valid END_OBJ

Yusuf


Sung Won Ho wrote:

> My concern is whether I can use muxReceive() in user application not in
> driver level.


---------------------------

Newsgroups: comp.os.vxworks
Subject: serial communication
Date: Fri, 11 Aug 2000 08:05:58 -0400 (EDT)
From: "Keith W. Muzzioli" 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <200008111205.IAA17628@stickleback.ssd.ray.com>

We are having a problem with some serial communications running either VxWorks 
5.1.1 or VxWorks 5.3, using a DY4 DMV-163 68040 board with a DMAX227 (Raytheon 
version of the 220) mezzanine card.

We are sending packets of data from the PC to the DMV-163 card
of varying lengths, and started noticing errors with large packets.  If
the packet size is greater than 640 Bytes  (for example 740 in our
case), data gets "jumbled".  By jumbled, I mean that the first 640 bytes
come through just fine, but then the next 48 bytes are skipped, and the
LAST 64 bytes are sent next, followed by the skipped 48 bytes.  No
data is actually LOST, just mixed up a little...

Sent =>  < 740 Bytes >
Rcvd => < First 640 Bytes> < Last 64 Bytes > < Middle 48 Bytes >

Some comments and observations:

1) We have verified that the order is correct coming out of the PC.
2) The PC writes one byte at a time.
3) The code on the 163 card reads one byte at a time (with calls
that use the read() function).
4) As required, the memory buffer that we submit to the read() function
is on the DMAX card.
5) The failure is not 100% of the time.  It may or may not happen.  Not a
hard fault, but very frequent.
6) We have extended the memory buffer size to > our maximum data size.
7) In fact, if we pause the reading until after all writing has been completed,
we are able to read the data with no errors, so we know we are not over-
writing a maximum buffer size.
8) Also, if we slow down the rate at which data is written, it does not
fail.
9) ie: Only when writing and reading are simultaneous does it break.
10) This happens on more than one PC and 163 combination.
11) We do not use timeouts.  This is asynchronous, thus the read()
call is pending on each byte request.
12) Baud rate is 19200.

Has anyone seen behavior like this before??? Any suggestions would be greatly 
appreciated.

TIA,
Keith


---------------------------

Newsgroups: comp.os.vxworks
Subject: Problems with VxWorks FTP server. 
Date: Fri, 11 Aug 2000 15:11:53 +0200
From: Ezra Koper 
Organization: Bezeq International Ltd.
Message-ID: <3993FB99.6EBEE096@nevision.net.il>

Hello all.

I have a problem with VxWorks FTP server. I have no problem to activate
the FTP server on the target, and i can login to it. After i login , i
cant see any files , and cant upload any file to the FTP server.
Does anyone know what might be the reason that i can only login to the
FTP but no other activity can be done?


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: OT BSP question
Date: Fri, 11 Aug 2000 15:08:40 +0200
From: Ezra Koper 
Organization: Bezeq International Ltd.
Message-ID: <3993FAD8.A9621254@nevision.net.il>
References: <8mumk7$s4c$1@nnrp1.deja.com> <8muv3g$39q$1@nnrp1.deja.com>

Hello Richard

We also started to use the VxWorks & Tornado-2 fiew weeks ago. We have a
platform which is P2-333Mhz. But many times i actualy use a standart PC
with Intel Pro100 Ethernet card as my target HW. Best to my knowladge
there is no problem to use it on machines having 440BX chipset ( i am
using such a machine). I have no problem to Boot the target ( which is a
standart PC or my proprietary HW = 90% standart PC) by using the Boot
Loader for Pentium Pro that they provide.

If you need a detailed "geting started" document on how to run VxWorks
over X86 processors please e-mail me to   ezra_k@arelnet.com and i will be
happy to send yoy the document i have prepared. If wind river had such a
socument it would same new users a great deal of time.

richard.lee@smsc.com wrote:

> I apologize for not having your answer, but could you please
> tell me what BSP you are using? I want to purchase vxworks, but
> neither the salesman nor FAE can tell me if the Generic x86 BSP
> supports the 440BX chipset. Obviously yours does. I would be very
> grateful for the info. Thanks.
>
> In article <8mumk7$s4c$1@nnrp1.deja.com>,
>   cferrell6421@my-deja.com wrote:
> > Does anyone have ballpark Interrupt Latency timing numbers for an
> Intel P3
> > 800 MHz w/440BX Chip Set.  Operating under VxWorks 5.4. I could
> extrapolate
> > numbers for other x86 processors if you have them.  My WRS FAE is on
> vacation
> > and the 1-800 number for WRS support can't help.
> >
> > Thanks in advance.....
> >
> > Sent via Deja.com http://www.deja.com/
> > Before you buy.
> >
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: vxWorks task suspend
Date: Fri, 11 Aug 2000 11:57:45 +0530
From: Vijay Kumar Peshkar 
Organization: Ye 'Ol Disorganized NNTPCache groupie
Message-ID: <39939CE1.967368D4@cisco.com>
References: <8n024i$p7e$1@overload.lbl.gov>
Reply-To: vpeshkar@cisco.com

This is a multi-part message in MIME format.
- --------------0620FFF8E2223430B5015D2E
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Hi,

condition 6  leads to condition 1.
Can you please explain condition 3?

I am not recollecting any other condition now.

regds,
Vijay

Krishna Kishore Reddy Sureddy wrote:

> Hi,
>    Can a task in VXWORKS get suspended other than the conditions
> preseted below.
>    If yes please send them across as soon  as possible.
>
>    1. Task causing exception(task level exception).
>    2. Suspending a task from shell by ts( )  or taskSuspend( ) .
>    3. Attaching a task to Debugger.
>    4. Keeping taskSuspend(0) in the code of the task.
>    5. Keeping a break point in the code of the task.
>
>    The below case i am not sure and correct me if i am wrong
>    6. When you are trying to free/delete the memory which is already
> deleted.
>
> Thanks and Regards,
> S.K.K.Reddy.

- --------------0620FFF8E2223430B5015D2E
Content-Type: text/x-vcard; charset=us-ascii;
 name="vpeshkar.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Card for Vijay Kumar Peshkar
Content-Disposition: attachment;
 filename="vpeshkar.vcf"

begin:vcard 
n:Kumar.P;Vijay
tel;work:080-5530035/36 Extn:3464
x-mozilla-html:FALSE
org:Wipro Technologies, K2, Bangalore.;Wishing you blue skies, a couple of daisies, some fresh air and a smile!
adr:;;;;;;
version:2.1
email;internet:vijay.peshkar@wipro.com
title:Senior Software Engineer
note;quoted-printable:      o   o   o   o   o   o   o   .  .  . _____________________________________________=0D=0A    o           _____                      ||  Vijay Kumar .P         =20                                                    |=0D=0A   .] [__n_n_| DD[  =3D=3D=3D=3D_____   | vpeshkar@cisco.com, vijay_peshkar@hotmail.com     | =0D=0A >(________|__|_[_________]_ |___________________________________________|=0D=0A _/oo OOOOO oo`  ooo   ooo     'o!o!o                                 =20           
                     o!o!o`
fn:Vijay Kumar.P
end:vcard

- --------------0620FFF8E2223430B5015D2E--


---------------------------

Newsgroups: comp.os.vxworks
Subject: mv2700 LM/SIG and LM address
Date: Fri, 11 Aug 2000 03:01:10 -0400
From: Al Johnston 
Organization: MindSpring Enterprises
Message-ID: <3993A4B6.4815D66A@mindspring.com>

I have been puzzling over how to use the sig/lm registers on the
mv2700 card.  The registers I am asking about are (in pci/io space):

  0000,1000 - sig/lm ctrl reg
  0000,1001 - sig/lm stat reg
  0000,1002 - vmebus location monitor upper base address
  0000,1003 - vmebus location monitor lower base address

I assume that the "base address" is the location in vme space
that we want to monitor, but I cant find a clue in the book
as to what "space" (a16/24/32) this is in or how to control it.

Can anyone fill in the missing pieces for me?  coding examples
would be most welcome.

I ask my FAE about this, but its been a week now... and I am running
out of other things to do.

Note that I am not asking about vxworks "mailbox" stuff, as I 
need to roll my own.

tnx.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Problems with VxWorks FTP server.
Date: Fri, 11 Aug 2000 17:23:42 +0400
From: "Arthur V. Grigorian" 
Organization: Auriga, Inc.
Message-ID: <3993FE5E.E1849ECB@auriga.ru>
References: <3993FB99.6EBEE096@nevision.net.il>

Hi,

   Try to create ramFS on your vxWorks target for example
   by following commands:
      pBlkDev = ramDevCreate(0, 512, 10000, 10000, 0)
      dosFsMkfs("/ftproot", pBlkDev)
   and  repeat  your  upload  attempt.  (Don't forget  to
   `cd /ftproot`  after  you  log into  the  FTP  server)
   Probably you have  some permission problems  with  the
   file system on your FTP server.

Regards,
__
R2

>Ezra Koper wrote:
> 
> Hello all.
> 
> I have a problem with VxWorks FTP server. I have no problem to activate
> the FTP server on the target, and i can login to it. After i login , i
> cant see any files , and cant upload any file to the FTP server.
> Does anyone know what might be the reason that i can only login to the
> FTP but no other activity can be done?

---------------------------

Newsgroups: comp.os.vxworks
Subject: muxReceive
Date: Fri, 11 Aug 2000 15:35:20 +0900
From: "Sung Won Ho" 
Organization: ELIMnet Internet News
Message-ID: <8n073i$lpn$1@news.elim.net>

Hi All,

Can I use muxReceive() function in user application?
I want to test an UDP socket which is properly set , without real network
frame but
user genarated frame.

Of course, before calling muxReceive(), I configured mBlk/clBlk/Cluster
chain with
test data. I think that the two input parameters of muxReceive(), (END_OBJ
*) and M_BLK_ID,
are proper.

My concern is whether I can use muxReceive() in user application not in
driver level.

Thanks in advance.
Sung.




---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: NTP client on VxWorks
Date: 11 Aug 2000 08:22:52 GMT
From: johan@borksoft.xs4all.nl
Organization: Borksoft
Message-ID: <8n0d4s$jv$1@news1.xs4all.nl>
References: <311c2be6.cfb476ec@usw-ex0106-045.remarq.com> <8msg0l$9ci$1@nnrp1.deja.com> <3991D24B.97FCA644@lmco.com> <3992EE1F.DBF8E127@aps.anl.gov>

Andrew Johnson  wrote:

: Wasn't there a port of xntp available for vxWorks several years ago
: though?

There is a file called xntp-3.4y.tar.Z available on
ftp://ftp.atd.ucar.edu/pub/vxworks/vx/

Groeten,
	Johan

- -- 
     o o o o o o o . . .   ___________________________________
    o      _____           ||    Johan Borkhuis              |
  .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
 >(________|__|_[_________]_|________________________________|
 _/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 function can I use to get host ipaddr?
Date: Fri, 11 Aug 2000 07:19:26 -0700
From: Jim Way 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <8n159i$aj6$1@overload.lbl.gov>

lx1975,

There's probably a number of ways to do it. One way is to examine the
sysBootLine string. Here's a shell example:

	-> printf sysBootLine
	cpm(0,0)jway:vxWorks e=172.16.18.23:ffff0000 h=172.16.16.166
g=172.16.16.1 u=remoteftp pw= 
	tomahawk f=0x8 tn=rot21value = 113 = 0x71 = 'q'

The e= value is the target ip and mask; the h= value is the host.

In code, this snippet seems to work ok:

	extern char *sysBootLine;
	BOOT_PARAMS BootParams;
	if (*bootStringToStruct (sysBootLine, &BootParams) == 0)
	{
		// Get my IP address. (see documentation for field name
associated with host IP address).
		sscanf (BootParams.ead, "%d.%d.%d.%d", &l1, &l2, &l3, &l4);
	}

Hope this helps.

Jim

>	Newsgroups: comp.os.vxworks
>	Subject: which function can I use to get host ipaddr?
>	Date: Thu, 10 Aug 2000 20:40:24 GMT
>	From: lx1975@my-deja.com
>	Organization: Deja.com - Before you buy.
>	Message-ID: <8mv3vn$7ai$1@nnrp1.deja.com>

- -----------------------------------------
Jim Way, Software Engineer
Datum Austin (Austron Inc.)
voice: 512.721.4170
fax  : 512.990.9712
email: jwayATdatumDOTcom (no spam please)
- -----------------------------------------

---------------------------

Newsgroups: comp.os.vxworks
Subject: RE: serial communication
Date: Fri, 11 Aug 2000 11:30:26 -0400
From: "Mike Anderson" 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <8n18q1$bp0$1@overload.lbl.gov>

VxWorks Greetings!

> 
> We are having a problem with some serial communications running 
> either VxWorks 
> 5.1.1 or VxWorks 5.3, using a DY4 DMV-163 68040 board with a 
> DMAX227 (Raytheon 
> version of the 220) mezzanine card.
> 


> data is actually LOST, just mixed up a little...
> 
> Sent =>  < 740 Bytes >
> Rcvd => < First 640 Bytes> < Last 64 Bytes > < Middle 48 Bytes >
> 


> 9) ie: Only when writing and reading are simultaneous does it break.
> 10) This happens on more than one PC and 163 combination.
> 11) We do not use timeouts.  This is asynchronous, thus the read()
> call is pending on each byte request.
> 12) Baud rate is 19200.
> 
> Has anyone seen behavior like this before??? Any suggestions 
> would be greatly 
> appreciated.

It sounds like there is some sort of preemption problem going
on here.  What are the relative priorities of the tasks doing 
the reading/writing?  Are you using read/write or fread/fwrite?
How did you open the serial port?  Are you using the standard
TyCo driver or one of your own making?

Inquiring minds want to know...

Regards,

Mike Anderson
The PTR Group, Inc.
BSPs, Device Drivers, Systems Architecture, custom training
http://www.theptrgroup.com
mailto:mike@theptrgroup.com

---------------------------

Newsgroups: comp.os.vxworks
Subject: NFS problems with windows
Date: Fri, 11 Aug 2000 16:54:23 +0100 (BST)
From: David Laight 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <200008111554.QAA11804@dsl-2.tadpole.co.uk>

Some of our customers are seeing problems using the OnmiNFS windows client (on 
NT4) talking to a vxWorks server (in this case vxWorks 5.3.1).

I was wondering if anyone else has seen anything similar - or even who knows 
enough about the NFS protocol to know who is at fault.

NFS is being used to update the filestore held on the vxWorks in flash.  (This 
isn't TFFS, but has a caching layer to avoid the works effects of the 128k 
read-modify-write sequence required to modify a sector.)

Attempts to do recursive copies apparantly lock up and corrupt the vxWorks 
filesystem.  I see no such problems with my systems that use a solaris host - so 
am tempted to blame the PC NFS client.

For a small file snoop reports the following sequence:

123   0.00025 windows -> vxWorks NFS C CREATE2 FH=E53F 1.jpg
124   0.00871 vxWorks -> windows NFS R CREATE2 OK FH=E53E
125   0.00045 windows -> vxWorks NFS C READDIR2 FH=E53F Cookie=0
126   0.00677 vxWorks -> windows NFS R READDIR2 OK 3 entries (No more)
127   0.00020 windows -> vxWorks NFS C LOOKUP2 FH=E53F 1.jpg
128   0.00635 vxWorks -> windows NFS R LOOKUP2 OK FH=E53E
129   0.00019 windows -> vxWorks NFS C LOOKUP2 FH=E53F 1.jpg
130   0.00836 vxWorks -> windows NFS R LOOKUP2 OK FH=E53E
131   0.00020 windows -> vxWorks NFS C READ2 FH=E53E at 0 for 4096
132   0.00614 vxWorks -> windows NFS R READ2 OK (0 bytes)
133   0.00148 windows -> vxWorks NFS C WRITE2 FH=E53E at 0 for 2195
134   0.00072 windows -> vxWorks UDP continuation ID=53091
135   0.49994 windows -> vxWorks NFS C WRITE2 FH=E53E at 0 for 2048
136   0.00061 windows -> vxWorks UDP continuation ID=53347
137   0.30471 vxWorks -> windows NFS R WRITE2 OK
138   0.00418 vxWorks -> windows NFS R WRITE2 OK
139   0.00032 windows -> vxWorks NFS C WRITE2 FH=E53E at 2048 for 147
140   0.00735 vxWorks -> windows NFS R WRITE2 OK
141   0.00025 windows -> vxWorks NFS C LOOKUP2 FH=E53F 1.jpg
142   0.00609 vxWorks -> windows NFS R LOOKUP2 OK FH=E53E
143   0.00019 windows -> vxWorks NFS C LOOKUP2 FH=E53F 1.jpg
144   0.00637 vxWorks -> windows NFS R LOOKUP2 OK FH=E53E
145   0.00022 windows -> vxWorks NFS C SETATTR2 FH=E53E
146   0.00689 vxWorks -> windows NFS R SETATTR2 OK
147   0.00018 windows -> vxWorks NFS C GETATTR2 FH=E53E
148   0.00559 vxWorks -> windows NFS R GETATTR2 OK
149   0.02517 windows -> vxWorks NFS C READDIR2 FH=F2EA Cookie=0
150   0.00754 vxWorks -> windows NFS R READDIR2 OK 6 entries (No more)

Note that the write (frame 133) is timed out and retransmitted (frame 135) after 
a 1/2 second (or so) timeout.  The vxWorks system is probably erasing a flash 
sector.  I'm surprised this timeout is quite that small.

There is also a read request (frame 131), inspection of larger files indicates 
this is an attempt to read the last 4k block of the file.

Frame 145 sets the file owner - or rather tries to as the vxWorks filesystem 
doesn't support the relevant attributes. This is rather a good job as what it 
actually sets is:

NFS:  Proc = 2 (Set file attributes)
NFS:  File handle = CE2B3CC100000000027615A200000000
NFS:                00000000000000000000000000000000
NFS:  Mode = 037777777777
NFS:   Type = ?
NFS:   Setuid = 1, Setgid = 1, Sticky = 1
NFS:   Owner's permissions = rwx
NFS:   Group's permissions = rwx
NFS:   Other's permissions = rwx
NFS:  UID = -1
NFS:  GID = -1
NFS:  Size = 2195
NFS:  Access time       = 28-Jun-00 17:01:54.000000 GMT
NFS:  Modification time = 28-Jun-00 17:01:54.000000 GMT

I'm sure that the setuid, setgid and sticky bits should not be set.

Anyway sequences like the one above work.

The problems arise with much larger files.  What we then get is:

1909   0.00027 windows -> vxWorks NFS C LOOKUP2 FH=E6DB UserGuide.pdf
1910   0.00647 vxWorks -> windows NFS R LOOKUP2 OK FH=E6DA
1911   0.00021 windows -> vxWorks NFS C LOOKUP2 FH=E6DB UserGuide.pdf
1912   0.00645 vxWorks -> windows NFS R LOOKUP2 OK FH=E6DA
1913   0.00134 windows -> vxWorks NFS C WRITE2 FH=E6DA at 0 for 4096
1914   0.00122 windows -> vxWorks UDP continuation ID=7527
1915   0.00106 windows -> vxWorks UDP continuation ID=7527
1916   0.03302 vxWorks -> windows NFS R WRITE2 OK
(load of writes)
2623   0.00125 windows -> vxWorks NFS C WRITE2 FH=E6DA at 323584 for 4096
2624   0.00123 windows -> vxWorks UDP continuation ID=19816
2625   0.00106 windows -> vxWorks UDP continuation ID=19816
2626   0.02605 vxWorks -> windows NFS R WRITE2 OK
2627   0.00095 windows -> vxWorks NFS C READ2 FH=E6DA at 1351680 for 4096
2628   0.00605 vxWorks -> windows NFS R READ2 OK (0 bytes)
2629   0.00182 windows -> vxWorks NFS C WRITE2 FH=E6DA at 327680 for 4096
(loads more writes)
4412   0.00134 windows -> vxWorks NFS C WRITE2 FH=E6DA at 1310720 for 4096
4413   0.00123 windows -> vxWorks UDP continuation ID=1131
4414   0.00106 windows -> vxWorks UDP continuation ID=1131
4415   0.00014 windows -> vxWorks NFS C READDIR2 FH=E6DB Cookie=0
4416   0.02707 vxWorks -> windows NFS R WRITE2 OK
4417   0.00129 windows -> vxWorks NFS C WRITE2 FH=E6DA at 1314816 for 4096
4418   0.00123 windows -> vxWorks UDP continuation ID=1643
4419   0.00106 windows -> vxWorks UDP continuation ID=1643
4420   0.00084 vxWorks -> windows NFS R READDIR2 OK 3 entries (No more)
4421   0.00021 windows -> vxWorks NFS C LOOKUP2 FH=E6DB UserGuide.pdf
4422   0.02604 vxWorks -> windows NFS R WRITE2 OK
4423   0.00129 windows -> vxWorks NFS C WRITE2 FH=E6DA at 1318912 for 4096
4424   0.00123 windows -> vxWorks UDP continuation ID=2155
4425   0.00106 windows -> vxWorks UDP continuation ID=2155
4426   0.00084 vxWorks -> windows NFS R LOOKUP2 OK FH=E6DA
4427   0.00021 windows -> vxWorks NFS C LOOKUP2 FH=E6DB UserGuide.pdf
4428   0.02591 vxWorks -> windows NFS R WRITE2 OK
4429   0.00129 windows -> vxWorks NFS C WRITE2 FH=E6DA at 1323008 for 4096
4430   0.00123 windows -> vxWorks UDP continuation ID=2667
4431   0.00106 windows -> vxWorks UDP continuation ID=2667
4432   0.00086 vxWorks -> windows NFS R LOOKUP2 OK FH=E6DA
4433   0.00022 windows -> vxWorks NFS C SETATTR2 FH=E6DA
4434   0.05144 vxWorks -> windows NFS R WRITE2 OK
4435   0.00129 windows -> vxWorks NFS C WRITE2 FH=E6DA at 1327104 for 4096
4436   0.00123 windows -> vxWorks UDP continuation ID=3179
4437   0.00108 windows -> vxWorks UDP continuation ID=3179
4438   0.00008 vxWorks -> windows NFS R SETATTR2 I/O error
4439   0.02607 vxWorks -> windows NFS R WRITE2 OK
(6 more writes)

Here the read (frame 2627) and readdir-lookup-setattr sequence (frames 4415, 
4421, 4423 and 4433) are all done DURING the sequence of writes.  The setattr is 
rejected, maybe because it specifies the full file size but the actual file 
isn't that big (yet).

This looks to me very much as though the PC has two control threads running in 
parallel (or is not allowing for the asynchronous nature of the writes) and so 
is generating requests out of sequence.

Has anyone seen anything similar?

(Note this is not dosFs 2.  dosfs 2 is not available for vxWorks 5.3.1, also I 
don't know how its cache layer would interact with the one we have to have to 
get non-zero performance from flash memory.  I'm also wary of dosfs2's ability 
to allocate NFS file handles that are preserved across system reloads (dosfs 1 
saves these in the directory entry).

	David

- ----------------------------------------------------------------
David Laight                          email: dsl@tadpole.co.uk
Tadpole Technology plc                phone: +44 1223 428 232
Cambridge, UK                         fax:   +44 1223 428 201


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Need a better FTP Server soft for VxWorks
Date: Fri, 11 Aug 2000 08:26:50 -0500
From: Greg Willden 
Organization: Posted via Supernews, http://www.supernews.com
Message-ID: <3993FF1A.957C88A4@swri.edu>
References: <39930AD9.2F74F81A@nevision.net.il> <8mv5c0$2fa$1@news1.xs4all.nl>

Or you can find a lot of good free FTP server packages on various software
sites.  We found one that we like that plays different sounds for a file get
and a file put.  I has been a nice audible way of knowing when things are
occuring.

johan@borksoft.xs4all.nl wrote:

> Ezra Koper  wrote:
> : Hello All
>
> : Does anyone know were i can find a better VxWorks FTP Server software?
>
> : The FTP server software that is part of VxWorks is very basic.
> : It doesn't  allow to define various restrictions based on user&pass.
> : At the moment you log in you can do everything & access any folder.
>
> Try the FTP version that is in the unsupported directory. This is
> completely source code, with various levels. It is shipped with the DosFS2
> update.
>
> Groeten,
>         Johan
>
> --
>      o o o o o o o . . .   ___________________________________
>     o      _____           ||    Johan Borkhuis              |
>   .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
>  >(________|__|_[_________]_|________________________________|
>  _/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 function can I use to get host ipaddr?
Date: Fri, 11 Aug 2000 17:51:48 +0100 (BST)
From: David Laight 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <200008111651.RAA11833@dsl-2.tadpole.co.uk>

> >	Subject: which function can I use to get host ipaddr?

If you know the interface name just call ifAddrGet().

Getting the list of interface names may be more difficult.

The vxWorks ifShow() routines checks each entry in the ifnet table.

	David

- ----------------------------------------------------------------
David Laight                          email: dsl@tadpole.co.uk
Tadpole Technology plc                phone: +44 1223 428 232
Cambridge, UK                         fax:   +44 1223 428 201


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: dec21x4xEnd driver
Date: Fri, 11 Aug 2000 13:35:50 GMT
From: stourangeau@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8n0vfm$hjb$1@nnrp1.deja.com>
References: <8mn55b$j7p$1@nnrp1.deja.com> <8mre08$t92$2@proxy2.fe.internet.bosch.de> <8msh3j$a9k$1@nnrp1.deja.com>

FINALLY !!!!!!!!!!!!!!!!!

I was able to make it work this morning...

By playing around with the registers, I figured out that in order to
correct the media select routine, we had to do
a two step programming of the csr6 register.

Unfortunately, the driver plays around with this register and it was
not possible to configure
it from the media select routine.

I decided to put the final configuration of csr6 in the sysLanIntEnable
routine...
That did it !!!!!!

Here are my mediaSelect and sysLanIntEnable routines...

/***********************************************************************
********
*
* sysDec21143MediaSelect - hook routine for dec21x4xStart
*
* This is hook routine for dec21x4xStart(). This routine selects a
media for
* dec21143 board.
*
* RETURNS: OK, always.
*/

STATUS sysDec21143MediaSelect
    (
    DEC21X4X_DRV_CTRL *      pDrvCtrl,   /* Driver control */
    UINT *                   pCsr6Val    /* CSR6 return value */
    )
    {
    ULONG *     csrReg;
    ULONG       csrData;

    /* set CSR6 value */

    *pCsr6Val |= CSR6_2114X_MB1;

    /* Led/Control selection */

    csrReg   = (ULONG *)(pDrvCtrl->devAdrs + (CSR15 *
DECPCI_REG_OFFSET));
    csrData  = (CSR15_21143_CWE | CSR15_21143_LG3 | CSR15_21143_LG1
							 |
CSR15_MD_MSK);
    *csrReg  = (csrData);

    csrData = (CSR15_21143_LG3 | CSR15_21143_LG1 | CSR15_MODE_10);
    *csrReg  = (csrData);

    /* force 10Base-T halh duplex configuration */

    /* set SIA registers */

    csrReg   = (ULONG *)(pDrvCtrl->devAdrs + (CSR14 *
DECPCI_REG_OFFSET));
    csrData  = (*csrReg);
    csrData &= ~(CSR14_21143_T4 | CSR14_21143_TXF| CSR14_21143_TXH |
                        CSR14_21143_TAS | CSR14_21143_TH |
CSR14_21143_ANE);
    *csrReg = (csrData);

    csrReg   = (ULONG *)(pDrvCtrl->devAdrs + (CSR13 *
DECPCI_REG_OFFSET));
    csrData  = (*csrReg);
    csrData |= CSR13_SRL_SIA;
    *csrReg  = (csrData);

    return (OK);
    }

/***********************************************************************
********
*
* sysLanIntEnable - enable dec21X4X interrupts
*
* This routine enables dec21X4X interrupts.  This may involve
operations on
* interrupt control hardware.
*
* RETURNS: OK or ERROR for invalid arguments.
*/

STATUS sysLanIntEnable
    (
    int	level		/* level number */
    )
    {

    ULONG *     csrReg;
    ULONG       csrData;

	/* Complete the MII port configuration
    /* set CSR6 value */
    csrReg   = (ULONG *)(dcDevAdrs + (CSR6 * DECPCI_REG_OFFSET));
	csrData  = *csrReg;
    csrData  |= CSR6_2114X_PS;
    *csrReg  = (csrData);

    return (sysIntEnablePIC(level));
    }


- --------------------------------------------------------
In article <8msh3j$a9k$1@nnrp1.deja.com>,
  john_94501@my-deja.com wrote:
> Hi,
>
> You are correct to switch; the driver was a 10Mbps only configuration
> (although you have the media select function in the BSP so you should
be
> able to update it if you so choose).
>
> A better solution would be to ask WR customer support if they have the
> dec21x40End driver available for your version of the OS and
> architecture. I am assuming that you're using ARM, in which case you
> need some mods to the 40 driver to make it work (ARM-specific
alignment
> issues). The newest 40 driver also supports some of the newer 21143
> devices properly.
>
> Hope that helps,
>
> John...
>
> In article <8mre08$t92$2@proxy2.fe.internet.bosch.de>,
>   Jan Koeppen  wrote:
> > We had the problem that this driver seems not to be working on a
> 100MBit Hub
> > :-((.
> > We solved it by buying old 10Mbit hubs that work fine (but is not
> > satisfactory.)
> >
> > regards
> > Jan
> >
> > stourangeau@my-deja.com schrieb in Nachricht
> > <8mn55b$j7p$1@nnrp1.deja.com>...
> > >I am having problems setting up the dec21x4xEnd driver. I dont have
> the
> > >source file so I am only playing with the configuration of the
> driver.
> > >
> > >After lots of work, I was able to make the driver attach to the mux
> and
> > >attach protocols to it.
> > >
> > >I had lots of problem with getting the link to be indicated on the
> > >hub... I had to play around with a mediaSelect routine..
> > >
> > >Problem is, the hub sees the link but the dec21x doesnt seem to see
> the
> > >hub. By looking at the status registers, I see link failures and
> > >autonegociation that is not completed...
> > >
> > >Need some help please...
> > >
> > >
> > >Sent via Deja.com http://www.deja.com/
> > >Before you buy.
> >
> >
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.
>


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: re: TCP retransmittions and lost packets
Date: Fri, 11 Aug 2000 18:35:58 +0100 (BST)
From: David Laight 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <200008111735.SAA11853@dsl-2.tadpole.co.uk>

> >Has anyone figured out how to get around, or scale down the time it
> >takes a tcp packet to be retransmitted?
> >
> >I have several vxworks processors conencted with ethernet based tcp
> >socket connections. Occasionally I see several packets dropped from the
> >network. I am seeing a ~1.5 second pause when a packet needs to be
> >retransmitted.
>

1.5 seconds is (as other people have said) quite normal, as is the reported 
recovery sequence.  The fact you see a gap in in the transmittion may be because 
the sending TCP stack doesn't have any more data buffered (I have a feeling that 
vxworks has a send-buffer size which can be configured/set).  On the Unix 
Streams based TCP stacks I am more familiar with the sending side always has 
plenty of data available.

What you really need to look at is why packets are being dropped.

Although TCP does recover from lost packets, on a local network it shouldn't 
have to.  Indeed NFS works only because packets are not lost (It actually works 
faster because it uses UDP not TCP because it then has absolute control of when 
acknowlegements are sent).

Places packets can go amiss:

1) ethernet cards with insufficient receive buffer space
2) ethernet cards with a slow bus interface
3) badly terminated and/or incorrectly grounded coax cable
4) AUI connectors that are not firmly home
5) ethernet repeater sets connecter to transceivers with SQE enabled
6) store-forward switches (and routers) with insufficient buffer space
7) cut though switches that discard packets if the destination segment is busy
8) links set full duplex at one end
9) overlong cable segments
10) any cards with the fujitsu 'NICE' chipset
11) thin ethernet cabling in general (bad joints etc)
12) broken device drivers

ALL ISA cards fall foul of either 1 or 2.
Sbus bus master cards fail at (2) on multi-cpu systems
PCI fast ethernet cards require the host (north) bridge to merge requests
A surprising number of early ethernet switches suffer from 7
Full duplex is best only used if enabled by auto-negotiation, even then the 
device driver and transceiver have to keep in sync.

	David


- ----------------------------------------------------------------
David Laight                          email: dsl@tadpole.co.uk
Tadpole Technology plc                phone: +44 1223 428 232
Cambridge, UK                         fax:   +44 1223 428 201


---------------------------

Newsgroups: comp.os.vxworks
Subject: MVME2305 will not read proper from A24 slave
Date: Fri, 11 Aug 2000 15:59:11 +0200
From: "T. Casteleijn" 
Organization: NLR
Message-ID: <399406AE.A1AA70E5@nlr.nl>
Reply-To: castelyn@nlr.nl

We have the following problem if we want to read with the PPC from the
A24 VME space the
following problem occurs, the PPC board hang's and only a hard reset can
wake
the PPC up.

We use the following hard and software;

- -    processor board MVME2305 of motorola (PPC)
- -    DIO XVME-244 of XYCOM
- -    OS VxWorks version 5.4

The problem also occurs with other salve board's.




---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: OT BSP question
Date: Fri, 11 Aug 2000 14:01:11 GMT
From: Richard Lee 
Organization: Deja.com - Before you buy.
Message-ID: <8n10v3$isk$1@nnrp1.deja.com>
References: <8mumk7$s4c$1@nnrp1.deja.com> <8muv3g$39q$1@nnrp1.deja.com> <3993FAD8.A9621254@nevision.net.il>

Thanks to both of you for the info. I saw the same info re the P6 or
Pentium Pro, but the P6 came out in 1995 so it doesn't say much about
the supported chipset. I'm a little afraid that after I get the BSP I
will find that only some of the functionality is implemented. For
example, is UDMA-ATA33 or USB on the south bridge supported? It really
is quite frustrating getting answers from WRS. I had no problems with
QNX. They were very helpful with their answers. What bugs me most is
that the answer is easy to find. I worked with VxWorks on the Intel i960
so I know how easy it would be to look through the source files for the
BSP. On the WRS page for supported BSP's, the latest x86 BSP is the P55
or the Pentium MMX which came out early in 1997. That seems to say that
the  440BX chipset is NOT explicitly supported. Of course it would still
work if the newer stuff is not used. Strangely, pSOS (owned by WRS) has
a BSP for the P3 which would suggest full 440BX support.


In article <3993FAD8.A9621254@nevision.net.il>,
  Ezra Koper  wrote:
> Hello Richard
>
> We also started to use the VxWorks & Tornado-2 fiew weeks ago. We have
a
> platform which is P2-333Mhz. But many times i actualy use a standart
PC
> with Intel Pro100 Ethernet card as my target HW. Best to my knowladge
> there is no problem to use it on machines having 440BX chipset ( i am
> using such a machine). I have no problem to Boot the target ( which is
a
> standart PC or my proprietary HW = 90% standart PC) by using the Boot
> Loader for Pentium Pro that they provide.
>
> If you need a detailed "geting started" document on how to run VxWorks
> over X86 processors please e-mail me to   ezra_k@arelnet.com and i
will be
> happy to send yoy the document i have prepared. If wind river had such
a
> socument it would same new users a great deal of time.
>
> richard.lee@smsc.com wrote:
>
> > I apologize for not having your answer, but could you please
> > tell me what BSP you are using? I want to purchase vxworks, but
> > neither the salesman nor FAE can tell me if the Generic x86 BSP
> > supports the 440BX chipset. Obviously yours does. I would be very
> > grateful for the info. Thanks.
> >
> > In article <8mumk7$s4c$1@nnrp1.deja.com>,
> >   cferrell6421@my-deja.com wrote:
> > > Does anyone have ballpark Interrupt Latency timing numbers for an
> > Intel P3
> > > 800 MHz w/440BX Chip Set.  Operating under VxWorks 5.4. I could
> > extrapolate
> > > numbers for other x86 processors if you have them.  My WRS FAE is
on
> > vacation
> > > and the 1-800 number for WRS support can't help.
> > >
> > > Thanks in advance.....
> > >
> > > Sent via Deja.com http://www.deja.com/
> > > Before you buy.
> > >
> >
> > Sent via Deja.com http://www.deja.com/
> > Before you buy.
>
>


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: MVME2305 will not read proper from A24 slave
Date: Fri, 11 Aug 2000 09:36:59 -0500
From: Greg Willden 
Organization: Posted via Supernews, http://www.supernews.com
Message-ID: <39940F8B.8F21AEF7@swri.edu>
References: <399406AE.A1AA70E5@nlr.nl>

"T. Casteleijn" wrote:

> We have the following problem if we want to read with the PPC from the
> A24 VME space the
> following problem occurs, the PPC board hang's and only a hard reset can
> wake
> the PPC up.
>

Are you sure that the VME A24 address you are reading from is the same VME
A24 address as on the slave board.  i.e.  Are you sure you're reading from
the right address.  The board may be trying to read from an invalid
address.  I seem to recall that our processor board would hang if we were
reading from an address at which no slave board was responding.


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: tcp retransmit delays
Date: Fri, 11 Aug 2000 09:35:38 -0500
From: Chris Smith 
Organization: Soutwest Research Institute
Message-ID: <39940F39.75AA0619@here.org>
References: <39918A06.50372148@here.org> <8muf53$huv$1@news.bayarea.net> <3992C12E.6C36F033@txc.com> <3992F7ED.4FCED82B@here.org> <8mvghb$sbt$1@news.bayarea.net>
Reply-To: cbsmithATswriDoTorg

Do you know if vxworks has implemented the NewReno modification to tcp? It
would be interesting to see if this algorithm would help my current scenario.

Kevin Lahey wrote:

> In article <3992F7ED.4FCED82B@here.org>,
> Chris Smith   wrote:
> >yes I am using etherpeek on a separate nt machine, and coordinating that
> >with the tcpDebugShow packets on the vxworks boards. The "RWIN" field is
> >not changing from the 61K it is initially using.
>
> Yah, tcpdump shows me the same thing.  It seems odd.
>
> >> Hmmm.  Maybe I'm misunderstanding your situation.  The way TCP
> >> usually works is that it sends out more data as it receives
> >> acknowledgements for the old data being received by the other side.
> >> If it stops getting ACKs, it should stop sending.  It'll then time
> >> out, and retransmit the packet.  While it is timing out, it shouldn't
> >> be sending packets.[*]
> >
> >The sender does not wait for an ack from the receiver before
> >transmitting the next packet.
> >
> >Each packet that is received generates an ACK. If a packet is dropped
> >the next packet that arrives at the receiver generates a "bad" ack -- it
> >only acknowledges up to the missed packet but it still sends an ack
> >because it did get a packet. Once vxworks gets a configurable number of
> >"bad" acks it is supposed to trigger a fast retransmit.
>
> Duplicate ACK.  When you get three, you are supposed to go into
> fast retransmit.  See:
>
>         http://www.ietf.org/rfc/rfc2581.txt
>
> for a complete explanation.
>
> >Or if a packet
> >has not been acknowledged after a certain period of time it is supposed
> >to retransmit it (the "slow" version). But I do not think that is
> >supposed to stop the data stream unless it runs out of buffer room. I
> >could be wrong on this point but I do not see a reason for it to
> >actually pause.
>
> The TCP stack is supposed to maintain a separate "congestion window"
> which is an estimate of how much data can be supported by the particular
> connection.  If the connection window runs out, sends should stop,
> even if the advertised window is still sufficient.
>
> Also, if you are in fast retransmit (you've already seen three dup acks),
> the congestion window gets shrunk pretty agressively, so you won't see
> a whole lot of transmissions during a timeout after a fast
> retransmission.
>
> Good luck,
>
> Kevin
> kml@bayarea.net


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: NTP client on VxWorks
Date: 11 Aug 2000 14:42:32 GMT
From: cedilnik andrej 
Organization: University of Maryland, Baltimore County
Message-ID: <8n13co$ses$1@news.umbc.edu>
References: <311c2be6.cfb476ec@usw-ex0106-045.remarq.com>

Warren.Brown  wrote:
> Hello
> Has anyone added the NTP (Network Time Protocol) client
> services to a VxWorks embedded target ?
> Any hints as to the right steps to take ?

Here you go. Sync("18.72.0.3");

struct {
    unsigned char livnmode, stratum, poll, prec;
    unsigned long delay, disp, refid;
    unsigned long reftime[2], origtime[2], rcvtime[2], txtime[2];
} pkt;

#define PORT 123
#define MY_OFFSET 2208988800L


/* Sync the time */
int 
Sync(char *hostname)
{
    int i, fd = -1, len, n, zonemin;
    struct sockaddr_in saddr, myaddr;
    struct hostent *hp;
        char answer[2000];
    unsigned long myt;
    long offset;
        fd_set readfds;
        struct timeval tv;

    /* Get server name */

    n = strlen(hostname);
    if (n <= 0) {
                perror("hostname");
                return 0;
    }
        sprintf(answer, "Connecting to %s\n", hostname);
        printf(answer);

    /* Resolve server name */

    bzero((char *)&saddr, sizeof saddr);
    saddr.sin_family = AF_INET;
    saddr.sin_port = htons(PORT);
    if ((saddr.sin_addr.s_addr = inet_addr(hostname)) == ERROR) {
                perror("unknown server");
                return 0;
    }
        sprintf(answer, "Hostname resolved, getting the socket\n");
        printf(answer);

    /* Get a socket */
        
    if ((fd = socket(AF_INET, SOCK_DGRAM, 0)) < 0) {
                perror("socket");
                return 0;
    }
        
        sprintf(answer, "Got socket, bind to port\n");
        printf(answer);
    /* Bind it to some port */

    bzero((char *)&myaddr, sizeof myaddr);
    myaddr.sin_family = AF_INET;
    myaddr.sin_port = htons(3030);
    myaddr.sin_addr.s_addr = INADDR_ANY;

    if (bind(fd, (struct sockaddr *) &myaddr, sizeof myaddr) < 0) {
                perror("bind error");
                close(fd);
                return 0;
    }

    /* Fill in the request packet */

    bzero((char *)&pkt, sizeof pkt);
    pkt.livnmode = 0x0b;        /* version 1, mode 3 (client) */

        sprintf(answer, "Send packet\n");
        printf(answer);
    /* Send packet to server */
        
    if ((n = sendto(fd, 
                                        (char *)&pkt, sizeof pkt, 0, 
                                        (struct sockaddr *) &saddr, 
                                        sizeof saddr)) != sizeof pkt) {
                sprintf(answer, "sendto %d err %d", n, errno);
                perror(answer);
                close(fd);
                return 0;
    }

        sprintf(answer, "Wait for reply\n");
        printf(answer);
    /* Read reply packet from server */
len = sizeof pkt;
        FD_SET(fd, &readfds);
        tv.tv_sec = 5;
        tv.tv_usec = 0;
        if(select(fd+1, &readfds, NULL, NULL, &tv)){
          if ((n = recvfrom(fd, 
                                                (char *)&pkt, sizeof pkt, 0, 
                                                (struct sockaddr *) &saddr, 
                                                &len)) != sizeof pkt) {
                sprintf(answer, "recvfrom %d err %d", n, errno);
                perror(answer);
                close(fd);
                return 0;
          }
        }
        else {
          perror("Timeout");
          close(fd);
          return 0;
        }
        sprintf(answer, "Sanitu check\n");
        printf(answer);
    /* Sanity checks */

    if ((pkt.livnmode & 0xc0) == 0xc0 || !pkt.stratum || !pkt.txtime[0]) {
                perror("server not synchronized");
                close(fd);
                return 0;
    }

    /*
     * NTP timestamps are represented as a 64-bit unsigned fixed-
     * point number, in seconds relative to 0h on 1 January 1900. The integer
     * part is in the first 32 bits and the fraction part in the last 32 bits.
     */

        myt = ntohl(pkt.txtime[0]) - MY_OFFSET;
        offset = myt - time(NULL);
        
    sprintf(answer, "Time: %s\nstratum %d offset %ld\n", 
                        ctime(&myt), (int) pkt.stratum, offset);
        
    if (fd >= 0)
                close(fd);
        printf(answer);
        return 1;
}


				Andy

- -- 
- -------------------------------------------------------------------
  )\._.,--....,'``.           Andrej Cedilnik (acedil1@cs.umbc.edu)
 /,   _.. \   _\  (`._ ,.     URL: http://www.cs.umbc.edu/~acedil1
`._.-(,_..'--(,_..'`-.;.'     TEL: (301)725-5724  ICQ: 27889915
- -------------------------------------------------------------------

---------------------------

Newsgroups: comp.os.vxworks
Subject: RE: serial communication
Date: Fri, 11 Aug 2000 14:30:23 -0400 (EDT)
From: "Keith W. Muzzioli" 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <200008111830.OAA13105@stickleback.ssd.ray.com>

> 
> VxWorks Greetings!
> 
> > 
> > We are having a problem with some serial communications running 
> > either VxWorks 
> > 5.1.1 or VxWorks 5.3, using a DY4 DMV-163 68040 board with a 
> > DMAX227 (Raytheon 
> > version of the 220) mezzanine card.
> > 
> 
> 
> > data is actually LOST, just mixed up a little...
> > 
> > Sent =>  < 740 Bytes >
> > Rcvd => < First 640 Bytes> < Last 64 Bytes > < Middle 48 Bytes >
> > 
> 
> 
> > 9) ie: Only when writing and reading are simultaneous does it break.
> > 10) This happens on more than one PC and 163 combination.
> > 11) We do not use timeouts.  This is asynchronous, thus the read()
> > call is pending on each byte request.
> > 12) Baud rate is 19200.
> > 
> > Has anyone seen behavior like this before??? Any suggestions 
> > would be greatly 
> > appreciated.
> 
> It sounds like there is some sort of preemption problem going
> on here.  What are the relative priorities of the tasks doing 
> the reading/writing?  Are you using read/write or fread/fwrite?
> How did you open the serial port?  Are you using the standard
> TyCo driver or one of your own making?
> 
> Inquiring minds want to know...

Thanks for your inquiring mind, let me try to clarify our situation...

The sending task is running on a PC that is running LYNX 3.0.1. We are using the 
OS select and write to write the data to the serial line.

The receiving task is running on the DY4 DMV-163 68040 board. There are 2 serial 
ports directly on the DMV-163 that are using the standard TyCo driver. We are 
NOT using these ports for this serial communication. There are four additional 
serial ports on the DMAX-227 daughter card, it is one of these that we are 
using. These four ports are using the DLCS (Data Link Communications Suite) 
device driver provided with our BSP from DY4. We initialized and opened the 
serial ports according to the documentation provided with the device driver. We 
are reading with just read, NOT fread.

I don't know how important the priorities are since both tasks are running under 
different OS, but the receiving task has a priority of 100 and is the only 
application task running at the time. I don't know the priority of the sending 
task but it is also the only application task running at the time.

The device driver documentation states the following:
The driver conforms to the conventional VxWorks I/O system interface and that 
the driver is accessed using the basic VxWorks I/O calls, open(), close(), 
read(), write(), and ioctl(). There are additional DLCS specific routines used 
in setting up the ports, etc, which are not passed to the I/O system. The driver 
is generic and has no knowledge of any device firmware specifics. It uses the 
interface provided by the VxCCB module to access the device firmware specific 
interfaces.

Thanks again,
Keith


---------------------------

Newsgroups: comp.os.vxworks
Subject: Tornado project facility
Date: Fri, 11 Aug 2000 13:27:08 -0500
From: "Martin Mayer" 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <8n1jbh$ht5$1@overload.lbl.gov>

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_01C003C1.C1E0CAC0
Content-Type: text/plain;
	charset="iso-8859-1"

Hi,

I have modified my config.h file to force the default boot line and
configured the IP address that vxWorks will use. I have then rebuilt the
BSP.

This works fine when I create a new project based on this BSP. The issue I
am having is that existing projects do not pick up changes I make in the
BSP. I need to delete the project I have and create a new one. This is bad
because I have to reselect all of the vxWorks components everytime I modify
the config.h file and rebuild the BSP...

Does anyone have an idea about how I can make existing projects pickup the
changes in config.h?

Thanks,
Martin Mayer
IP Comms Software Designer
Nortel Networks


- ------_=_NextPart_001_01C003C1.C1E0CAC0
Content-Type: text/html;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable






Tornado project facility




Hi,




I have modified my config.h file to =
force the default boot line and configured the IP address that vxWorks =
will use. I have then rebuilt the BSP.



This works fine when I create a new =
project based on this BSP. The issue I am having is that existing =
projects do not pick up changes I make in the BSP. I need to delete the =
project I have and create a new one. This is bad because I have to =
reselect all of the vxWorks components everytime I modify the config.h =
file and rebuild the BSP...



Does anyone have an idea about how I =
can make existing projects pickup the changes in config.h?




Thanks,

Martin Mayer

IP Comms Software Designer

Nortel Networks





- ------_=_NextPart_001_01C003C1.C1E0CAC0--

---------------------------

Newsgroups: comp.os.vxworks
Subject: Raw IP sockets and port number
Date: Fri, 11 Aug 2000 10:04:29 -0500
From: Michael White 
Organization: Motorola
Message-ID: <399415FD.D2A33025@ftw.mot.com>

Hi,

I want to use a Raw IP socket (type AF_INET) to send GRE messages.
A couple of quick questions:

   1)  How do I set the TOS in the IP header?
   2)  Is the port field in the sockaddr used on a Raw IP socket?
   3)  When sending a Raw IP message, what is the format of the data
       that needs to be passed to sendto()?  Does it include the IP
       header or just the GRE header and body?

Any help would be greatly appreciated.  If it's somewhere in the VxWorks
manuals, 
that'd be helpful, too.

Thanks in advance.
- ---
Michael White, qa1402@ftw.mot.com
Cellular Infrastructure Group
Ft. Worth, Texas    (817)245-6418

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: NTP client on VxWorks
Date: Thu, 10 Aug 2000 16:22:25 -0400
From: "Bryan D. Lee" 
Organization: Sanders-A Lockheed Martin Company
Message-ID: <39930F01.3B2F3533@lmco.com>
References: <311c2be6.cfb476ec@usw-ex0106-045.remarq.com> <8msg0l$9ci$1@nnrp1.deja.com> <3991D24B.97FCA644@lmco.com> <3992EE1F.DBF8E127@aps.anl.gov>

Thanks Andrew,

It looked from the code fragment that it was a one-shot deal.  And
what I'm working on now ought to allow continuous fine adjustments
with or without a 1-PPS signal as if the board is another computer NTP
client/server.  Besides we didn't have any intention of upgrading to
v5.4 any time soon and this ought to work on earlier vxworks
versions.  I'm not doing anything Tornado specific.

And yes there was/is an xntpd port from v3 sources, I believe.

For better or worse I started from the v4 sources.

Bryan


Andrew Johnson wrote:
> 
> "Bryan D. Lee" wrote:
> >
> > Mind if I ask what version of vxworks this is?  I don't have it in the
> > kernel and I don't see it in the 5.3 docs.
> >
> > Bryan
> >
> > john_94501@my-deja.com wrote:
> > >
> > > Add the INCLUDE_SNTPC component to the kernel, then do something like
> > > this in your application:
> 
> Bryan,
> 
> Tornado 2.0 (vxWorks 5.4) includes a Simple NTP Client, but it only seems
> to provide one-shot "what time is it now" or "come back when you next see
> an NTP broadcast" functionality and AFAIK doesn't try to correct for
> network delays or interface to the vxWorks clock.  It sounds like the
> solution you're working in is a lot more comprehensive.
> 
> Wasn't there a port of xntp available for vxWorks several years ago
> though?
> 
> - Andrew
> --
> Sharks kill 10 people each year.  People kill 60,000,000 sharks each year.

- -- 
Bryan D. Lee                             Phone: 603/885-7025
Software Engineer & ELDP                 Fax  : 603/885-3016
Sanders-A Lockheed Martin Company        Pager: 603/423-5565
PO Box 868, PTP1-2361, Nashua, NH 03061  Email: bryan.d.lee@lmco.com

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: tcp retransmit delays
Date: 11 Aug 2000 16:21:52 GMT
From: kml@baygate.bayarea.net (Kevin Lahey)
Organization: Bayarea Internet Solutions
Message-ID: <8n1970$em4$1@news.bayarea.net>
References: <39918A06.50372148@here.org> <3992F7ED.4FCED82B@here.org> <8mvghb$sbt$1@news.bayarea.net> <39940F39.75AA0619@here.org>

In article <39940F39.75AA0619@here.org>,
Chris Smith   wrote:
>Do you know if vxworks has implemented the NewReno modification to tcp? 
>It would be interesting to see if this algorithm would help my current 
>scenario.

I doubt that VxWorks has it. 

My impression so far (based on just a couple weeks of messing around
with VxWorks) is that the TCP stack is about three or four years out of 
date when compared to other modern stacks (Linux, Solaris, IRIX, *BSD).
It doesn't seem to have working TCP timestamps (they are advertised
in the SYN packet, but disappear after that), path MTU discovery,
Selective Acknowledgements, or defenses against some common DoS attacks.

While I'm using VxWorks 5.4, perhaps I have an old stack, or the wrong
stack, or something.  I'd be delighted if someone on this newsgroup
could point me at something better!

Thanks,

Kevin
kml@bayarea.net

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Display of Windows NT Windows on a VXWorks Power PC
Date: Fri, 11 Aug 2000 11:43:08 -0400
From: Charles H. Chapman 
Organization: Dynacs Engineering Company
Message-ID: <8n1702$9ci@niven.ksc.nasa.gov>
References: <8ms846$380$1@nnrp1.deja.com> <3992EC70.AB84236D@aps.anl.gov>
Reply-To: chc@nasa2.ksc.nasa.gov

Andrew Johnson wrote:
> lpeck@my-deja.com wrote:
> > 
> > I have a configuration with a dumb  Power PC
> > display server running VXworks that supports X
> > Displays. Currently we are able to display
> > windows from a Unix window on this server by
> > simply setting the display variable to the
> > display server. However, we would like to replace
> > the Unix machine with a PC running Windows NT. Is
> > there an application or a setting which will
> > allow me to run my existing Windows NT apps on my
> > Windows NT and display the windows for the app on
> > the dumb display server?
> 
> Hmm, interesting little problem.  I'm not aware of any applications that
> export a Windows frame buffer through X, although there may be one.  I
> would try building an X11 VNC client for vxWorks which ought to do the
> trick.  Do a web search for VNC to find it.  Of course if you can still
> have the Unix machine around then you could run the VNC client on Unix,
> with its display pointing to the vxWorks box - probably Ok for an interim
> solution, but longer-term I'd port the client.

Another somewhat convoluted solution would be to install Linux on the
Windows NT machine and then run Windows NT under Linux using VMWare (see
http://www.vmware.com) and redirect the output of VMWare to your display
server.

Chuck

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Display of Windows NT Windows on a VXWorks Power PC
Date: Fri, 11 Aug 2000 10:48:02 -0500
From: "Matthew Majka" 
Organization: NASA, Kennedy Space Center
Message-ID: <8n178g$9kk@niven.ksc.nasa.gov>
References: <8ms846$380$1@nnrp1.deja.com> <3992EC70.AB84236D@aps.anl.gov> <8n1702$9ci@niven.ksc.nasa.gov>


Charles H. Chapman  wrote in message
news:8n1702$9ci@niven.ksc.nasa.gov...
> Andrew Johnson wrote:
> > lpeck@my-deja.com wrote:
> > >
> > > I have a configuration with a dumb  Power PC
> > > display server running VXworks that supports X
> > > Displays. Currently we are able to display
> > > windows from a Unix window on this server by
> > > simply setting the display variable to the
> > > display server. However, we would like to replace
> > > the Unix machine with a PC running Windows NT. Is
> > > there an application or a setting which will
> > > allow me to run my existing Windows NT apps on my
> > > Windows NT and display the windows for the app on
> > > the dumb display server?
> >
> > Hmm, interesting little problem.  I'm not aware of any applications that
> > export a Windows frame buffer through X, although there may be one.  I
> > would try building an X11 VNC client for vxWorks which ought to do the
> > trick.  Do a web search for VNC to find it.  Of course if you can still
> > have the Unix machine around then you could run the VNC client on Unix,
> > with its display pointing to the vxWorks box - probably Ok for an
interim
> > solution, but longer-term I'd port the client.

You could install Linux on the PC, get the WinFrame client
for Linux and set the display to your VxWorks system.  Of
course you'd have to get a WinNT server to run the WinFrame
hosting software...

http://www.citrix.com/products/winframe/




---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: select() problem
Date: Fri, 11 Aug 2000 16:45:08 GMT
From: rhunnicutt@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8n1aij$qno$1@nnrp1.deja.com>
References: 

When I ran this, I would get the first byte but not the rest. On the
next iteration, I would get the remaining buffer and the first byte of
the next iteration. Not sure why this is happening but suspect select()
was not intended to be use like this. I added a do/while loop to finish
the read (or a read() call works great too) and it works fine.

  while (!end) {
    i = 0;
    newInput = 0;
    while (newInput < 1) {
      FD_ZERO (&readFds);
      FD_SET (STD_IN, &readFds);
      newInput = select (FD_SETSIZE, &readFds, NULL, NULL, &timeout);
    }
    do {
      i = getc(stdin);
      printf("%c", i);
    } while (i);
  }

In article ,
  "Øyvind Holmeide"  wrote:
> I'm using the select() routine in an application that uses the console
> (standard in) either via the serial line or telnet.  The select
routine
> returns when a byte is available.  However, the corresponding call to
getc()
> will not see the first byte (only the following bytes).  "Syntax
error" is
> also written to the console after the select() returns with 1.  The
code
> look like
> this:
>
>  timeout.tv_sec = 1;
>  timeout.tv_usec = 0;
>
>  while (!end) {
>     newInput = 0;
>     while (newInput < 1) {
>     FD_ZERO (&readFds);
>     FD_SET (STD_IN, &readFds);
>     newInput = select (FD_SETSIZE, &readFds, NULL, NULL, &timeout);
> }
>   i = getc(stdin);
>
> --
>
> Øyvind Holmeide
>
>


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: WIN2000 SP1 & TORNADO II
Date: Fri, 11 Aug 2000 12:07:58 -0500
From: "Michael Murphy" 
Organization: Imaging Business Machines, LLC
Message-ID: 
References:  <8mup9n$udn$1@nnrp1.deja.com>

I have been in touch with WRS support about this problem. They have no clue
either. There answer as of now is "Tornado II does not support Windows
2000".

Michael

 wrote in message news:8mup9n$udn$1@nnrp1.deja.com...
> Hi Michael,
>
> I have no clue what these errors are but I am
> experiencing them as well.  I'm only just
> starting to use Tornado II so I'm glad to hear
> that the program seems to work fine after
> accepting the error messages.
>
> Mark V.
>
> In article  east.usenetserver.com>,
>   "Michael Murphy"  wrote:
> > Installed Windows 2000 SP1 today. When running
> Tornado II I get a bunch of
> > the following errors:
> >
> > "RPCExit: extra call"
> >
> > Tornado II seems to run fine after I accept all
> these errors (4 to 5 of
> > them)
> >
> > Anybody have any clues?
> >
> > Thanx,
> >
> > Michael
> >
> >
>
>
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: TCP retransmittions and lost packets -- New Question
Date: Fri, 11 Aug 2000 17:02:21 -0500
From: Chris Smith 
Organization: Soutwest Research Institute
Message-ID: <399477ED.58C95E0@here.org>
References: <200008111735.SAA11853@dsl-2.tadpole.co.uk>
Reply-To: cbsmithATswriDoTorg

You can add setting socket send and receive buffers too large as another cause of
missing packets under vxworks -- at least under some networking conditions.

I have been running with SO_RCVBUF and SO_SNDBUF at 61440 (an "optimization" for
another embedded os). After reducing the buffers to 8192 -- the default -- I am no
longer seeing the 1.5 pauses. 16k also appears to be a safe buffer size, but 32k is
too big. It appears that I am no longer seeing the lost packets on the network.

I have used this same code on the same hardware under vxworks with a slightly
different networking environment (a different hub) and not had any problems. Can
anyone explain why this would suddenly cause lost packets? My current "network" is
two Synergy ppc sbc's and an NT laptop ethernet sniffer connected to a hub. The
hub's uplink is connected to a switch and the switch is on our general network.
Nothing on the general network is interacting with the hub.

What ramifications are there to changing the send and receive buffers? I know they
change the maximum allowable window to be used for the tcp connection, but what
other side effects does this have? Would external network components (the hub or
switch) affect what the window should be? Is there an RFC that addresses this?

chris


David Laight wrote:

> > >Has anyone figured out how to get around, or scale down the time it
> > >takes a tcp packet to be retransmitted?
> > >
> > >I have several vxworks processors conencted with ethernet based tcp
> > >socket connections. Occasionally I see several packets dropped from the
> > >network. I am seeing a ~1.5 second pause when a packet needs to be
> > >retransmitted.
> >
>
> 1.5 seconds is (as other people have said) quite normal, as is the reported
> recovery sequence.  The fact you see a gap in in the transmittion may be because
> the sending TCP stack doesn't have any more data buffered (I have a feeling that
> vxworks has a send-buffer size which can be configured/set).  On the Unix
> Streams based TCP stacks I am more familiar with the sending side always has
> plenty of data available.
>
> What you really need to look at is why packets are being dropped.
>
> Although TCP does recover from lost packets, on a local network it shouldn't
> have to.  Indeed NFS works only because packets are not lost (It actually works
> faster because it uses UDP not TCP because it then has absolute control of when
> acknowlegements are sent).
>
> Places packets can go amiss:
>
> 1) ethernet cards with insufficient receive buffer space
> 2) ethernet cards with a slow bus interface
> 3) badly terminated and/or incorrectly grounded coax cable
> 4) AUI connectors that are not firmly home
> 5) ethernet repeater sets connecter to transceivers with SQE enabled
> 6) store-forward switches (and routers) with insufficient buffer space
> 7) cut though switches that discard packets if the destination segment is busy
> 8) links set full duplex at one end
> 9) overlong cable segments
> 10) any cards with the fujitsu 'NICE' chipset
> 11) thin ethernet cabling in general (bad joints etc)
> 12) broken device drivers
>
> ALL ISA cards fall foul of either 1 or 2.
> Sbus bus master cards fail at (2) on multi-cpu systems
> PCI fast ethernet cards require the host (north) bridge to merge requests
> A surprising number of early ethernet switches suffer from 7
> Full duplex is best only used if enabled by auto-negotiation, even then the
> device driver and transceiver have to keep in sync.
>
>         David
>
> ----------------------------------------------------------------
> David Laight                          email: dsl@tadpole.co.uk
> Tadpole Technology plc                phone: +44 1223 428 232
> Cambridge, UK                         fax:   +44 1223 428 201


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Tornado project facility
Date: 11 Aug 2000 20:17:29 GMT
From: johan@borksoft.xs4all.nl
Organization: Borksoft
Message-ID: <8n1n0p$iee$1@news1.xs4all.nl>
References: <8n1jbh$ht5$1@overload.lbl.gov>

Martin Mayer  wrote:
: I have modified my config.h file to force the default boot line and
: configured the IP address that vxWorks will use. I have then rebuilt the
: BSP.

: This works fine when I create a new project based on this BSP. The issue I
: am having is that existing projects do not pick up changes I make in the
: BSP. I need to delete the project I have and create a new one. This is bad
: because I have to reselect all of the vxWorks components everytime I modify
: the config.h file and rebuild the BSP...

: Does anyone have an idea about how I can make existing projects pickup the
: changes in config.h?

When you want to make changes to a project you have to do that in the
project configuration and *NOT* in the config.h. Almost all the defines
you make in the config.h are redefined in the project files.
When you create a new project it is based on the config.h in the BSP
directory. But after that it is seperate from the config.h. So the
config.h only provides the defaults for a project.

Groeten,
	Johan

- -- 
     o o o o o o o . . .   ___________________________________
    o      _____           ||    Johan Borkhuis              |
  .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
 >(________|__|_[_________]_|________________________________|
 _/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: END Driver In Promiscuous Mode
Date: Fri, 11 Aug 2000 14:18:48 -0700
From: "AG" 
Organization: SBC Internet Services
Message-ID: 

I can't seem to be able to force my 82557 END driver in promiscuous mode
(windows T2). Has anybody tried that? Any input at all?

Thanks a lot,

AG.



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: select() problem
Date: Sat, 12 Aug 2000 02:42:21 GMT
From: pkockritz@home.com (Pete Kockritz)
Organization: @Home Network
Message-ID: 
References:  <8n1aij$qno$1@nnrp1.deja.com>

In article <8n1aij$qno$1@nnrp1.deja.com>, rhunnicutt@my-deja.com wrote:

> When I ran this, I would get the first byte but not the rest. On the
> next iteration, I would get the remaining buffer and the first byte of
> the next iteration. Not sure why this is happening but suspect select()
> was not intended to be use like this. I added a do/while loop to finish
> the read (or a read() call works great too) and it works fine.
> 
>   while (!end) {
>     i = 0;
>     newInput = 0;
>     while (newInput < 1) {
>       FD_ZERO (&readFds);
>       FD_SET (STD_IN, &readFds);
>       newInput = select (FD_SETSIZE, &readFds, NULL, NULL, &timeout);
>     }
>     do {
>       i = getc(stdin);
>       printf("%c", i);
>     } while (i);
>   }


Remember, on a serial port the chars are coming in one at a time. At 9600
baud, that means ~1 msec per char. Depending on your serial port hardware,
you could get an interrupt on the first char and the other chars have not
yet been received. Of course, if your serial port has a receive FIFO, you
will usually get the chars in groups.

Regards,
Pete
- --
+-----------------------------------------------------------------+
|       Pete Kockritz                  mailto:pkockritz@home.com  |
|                       This space for rent.                      |
+-----------------------------------------------------------------+

---------------------------

End of New-News digest
**********************


From vxwexplo-errs@csg.lbl.gov  Sun Aug 13 04:03:06 2000
From: Vxworks Exploder 
Date: Sun Aug 13 04:03:08 PDT 2000
Subject: comp.os.vxworks newsdigest

Comp.Os.Vxworks Daily Digest    Sun Aug 13 04:03:03 PDT 2000

        Subject: re: vxWorks task suspend
        Subject: Re: vxWorks task suspend
        Subject: Re: Web Server under VxWorks
        Subject: Re: Web Server under VxWorks
        Subject: performance cost of C++ exception handling
        Subject: Re: TCP retransmittions and lost packets -- New Question
        Subject: Ethernet sniffing
        Subject: Re: Need a better FTP Server soft for VxWorks

-------------------------------------------------------

Newsgroups: comp.os.vxworks
Subject: re: vxWorks task suspend
Date: Sat, 12 Aug 2000 10:14:34 +0200
From: Thomas Martin 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <8n3bjg$gut$1@overload.lbl.gov>

Dies ist eine mehrteilige Nachricht im MIME-Format.
- --------------2F5B8847A7A14E2D0DD26929
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Hi,

the vxWorks Users Group Exploder wrote:
> 
> Submitted-by vxwexplo-errs@csg.lbl.gov  Thu Aug 10 21:33:27 2000
> Submitted-by: "Krishna Kishore Reddy Sureddy" 
> 
> Hi,
>    Can a task in VXWORKS get suspended other than the conditions
> preseted below.
>    If yes please send them across as soon  as possible.
> 
>    1. Task causing exception(task level exception).
>    2. Suspending a task from shell by ts( )  or taskSuspend( ) .
>    3. Attaching a task to Debugger.
>    4. Keeping taskSuspend(0) in the code of the task.
>    5. Keeping a break point in the code of the task.
> 
>    The below case i am not sure and correct me if i am wrong
>    6. When you are trying to free/delete the memory which is already
> deleted.
You are right!
free() and malloc() have default debugging options. The default for
free() is (if a block is corrupted) suspend the task and log an
error message. Options can be:
MEM_ALLOC_ERROR_LOG_FLAG
MEM_ALLOC_ERROR_SUSPEND_FLAG
MEM_BLOCK_CHECK
MEM_BLOCK_CHECK_LOG_FLAG
MEM_BLOCK_CHECK_SUSPEND_FLAG
These option can be set by memOptionsSet() and can be bitwise ORed.

HTH
Thomas


- -- 
MfG / kind regards

Thomas MARTIN 
*******************************************************************
* Ingenieurbuero MARTIN            * Zwei Dinge sind unendlich:   *
* Schieszstaettstrasse 5a          * das Universum und die mensch-*
* D-82515 Wolfratshausen - Germany * liche Dummheit, aber bei dem *
* Phone/FAX: +49(0)8171/28445      * Universum bin ich mir noch   *
* mobile: +49(0)171/3454864        * nicht ganz sicher.           *
* email:ib-martin@topmail.de       *                              *
* http://ib-martin.ilo.de          *             [Albert Einstein]*
*******************************************************************
- --------------2F5B8847A7A14E2D0DD26929
Content-Type: text/x-vcard; charset=us-ascii;
 name="ing.buero.martin.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Karte für Thomas Martin
Content-Disposition: attachment;
 filename="ing.buero.martin.vcf"

begin:vcard 
n:Martin;Thomas
tel;pager:+49(0)171/3454864
tel;cell:+49(0)171/3454864
tel;fax:+49(0)8171/28445
tel;home:+49(0)8171/28445
tel;work:+49(0)8171/28445
x-mozilla-html:FALSE
url:http://ib-martin.ilo.de/
org:Ingenieurbuero Martin;F&E
adr:;;Schiessstaettstrasse 5a;Wolfratshausen;Bavaria;D-82515;Germany
version:2.1
email;internet:ib-martin@topmail.de
title:Dipl.-Ing.
x-mozilla-cpt:;8128
fn:Thomas Martin
end:vcard

- --------------2F5B8847A7A14E2D0DD26929--




---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: vxWorks task suspend
Date: Sat, 12 Aug 2000 13:45:29 GMT
From: "Dan Perlberger" 
Organization: Road Runner - EC
Message-ID: 
References: <8n024i$p7e$1@overload.lbl.gov>

7.  Another task calls taskSuspend on the task.
8.  The task itself calls taskSuspend on itself (instead of taskSuspend(0),
it calls taskSuspend(taskNameToId("theTaskName")).



                Dan




"Krishna Kishore Reddy Sureddy"  wrote in
message news:8n024i$p7e$1@overload.lbl.gov...
> Hi,
>    Can a task in VXWORKS get suspended other than the conditions
> preseted below.
>    If yes please send them across as soon  as possible.
>
>    1. Task causing exception(task level exception).
>    2. Suspending a task from shell by ts( )  or taskSuspend( ) .
>    3. Attaching a task to Debugger.
>    4. Keeping taskSuspend(0) in the code of the task.
>    5. Keeping a break point in the code of the task.
>
>    The below case i am not sure and correct me if i am wrong
>    6. When you are trying to free/delete the memory which is already
> deleted.
>
> Thanks and Regards,
> S.K.K.Reddy.
>
>
>
>



---------------------------

Newsgroups: comp.os.vxworks,comp.realtime
Subject: Re: Web Server under VxWorks
Date: Sat, 12 Aug 2000 20:21:45 GMT
From: "connie kephart" 
Organization: EarthLink Inc. -- http://www.EarthLink.net
Message-ID: 
References: <397D1FD2.2D51DB48@yahoo.com> <397d085a.952203437@news.netis.com>
Reply-To: "connie kephart" 

Eric,

You suggested GoAhead's web server, which happens to be what we're planning
to use. Have you tried to run the web server on the full-blown VxSim? I can
build the server for my target with no problem, but can't get external
references to unld (from unldLib.h) and loadModule (from loadLib.h)
resolved. Any hints?

TIA,
Connie Kephart

"RKT Technologies, Inc."  wrote in message
news:397d085a.952203437@news.netis.com...
> On Mon, 24 Jul 2000 22:04:19 -0700, Shelly HUANG 
> wrote:
>
>
> >Is any web server running under VxWorks?
>
> Yes, look at http://www.goahead.com.
>
> >-----------------------------------------------------------------<
> Eric Poole                 Specializing in Software Development for
> RKT Technologies, Inc.       Medical Devices / Regulated Industries
> PO Box 1100                ISO 9001 / FDA Quality System Regulation
> Londonderry, NH 03053
> Phone:   603-437-1811      Fax:    603-425-6475
> >-----------------------------------------------------------------<
> E-mail: epooleNO@UCEnetis.com             (Remove "NO" and "UCE" to
>                       reply by e-mail; use epoole-at-netis-dot-com)
> >-----------------------------------------------------------------<
>   Qualifications and Background: see http://www.netis.com/rkt
> >-----------------------------------------------------------------<
>



---------------------------

Newsgroups: comp.os.vxworks,comp.realtime
Subject: Re: Web Server under VxWorks
Date: Sat, 12 Aug 2000 22:06:01 GMT
From: epoole@scoot.netis.com (RKT Technologies, Inc.)
Organization: Software Consultants to Medical Device Manufacturers
Message-ID: <3995c9c9.673513133@news.netis.com>
References: <397D1FD2.2D51DB48@yahoo.com> <397d085a.952203437@news.netis.com> 

On Sat, 12 Aug 2000 20:21:45 GMT, "connie kephart"
 wrote:

>You suggested GoAhead's web server, which happens to be what we're planning
>to use. Have you tried to run the web server on the full-blown VxSim?

No, we've used the Linux version, and the VxWorks version installed
and running on a PowerPC.

> I can
>build the server for my target with no problem, but can't get external
>references to unld (from unldLib.h) and loadModule (from loadLib.h)
>resolved. Any hints?

I don't know.  I'm the one that initially got it up and running but
I'm not the one that's been working with it since.  I'll ask one of
the other engineers on Monday and see if I can get some info.

Eric

>-----------------------------------------------------------------<
Eric Poole                 Specializing in Software Development for
RKT Technologies, Inc.       Medical Devices / Regulated Industries
PO Box 1100                ISO 9001 / FDA Quality System Regulation
Londonderry, NH 03053
Phone:   603-437-1811      Fax:    603-425-6475
>-----------------------------------------------------------------<
E-mail: epooleNO@UCEnetis.com             (Remove "NO" and "UCE" to
                      reply by e-mail; use epoole-at-netis-dot-com)
>-----------------------------------------------------------------<
  Qualifications and Background: see http://www.netis.com/rkt
>-----------------------------------------------------------------<

---------------------------

Newsgroups: comp.os.vxworks
Subject: performance cost of C++ exception handling
Date: Sun, 13 Aug 2000 01:58:26 GMT
From: "Jim Reis" 
Organization: RoadRunner - Central Florida
Message-ID: <6hnl5.13002$9T1.101123@typhoon.tampabay.rr.com>

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





---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: TCP retransmittions and lost packets -- New Question
Date: Sat, 12 Aug 2000 21:25:08 -0700
From: Hwajin Bae 
Organization: http://www.remarq.com: The World's Usenet/Discussions Start Here
Message-ID: <0eda9159.dc1845f1@usw-ex0104-033.remarq.com>
References: <200008111735.SAA11853@dsl-2.tadpole.co.uk> <399477ED.58C95E0@here.org>

The send side socket buffer size is used by the TCP/IP stack to
throttle the data output.  The user data is copied into mbufs at
the socket level and put into socket level buffers, which are
limited at the size specified.  The mbufs for each socket is not
preallocated; they are allocated as needed (as more data is
copied into mbufs from user buffers).  The send side socket
level buffer size is used by the TCP/IP stack to limit this
function (allocation and copying into mbufs), pending the task
that is sending data when limit is reached.  Pending is done by
taking a semaphore which will be given later by the stack when
the queued data is actually sent over the TCP to the peer and
ACK'ed and mbuf is freed, thereby creating more room in the send
side buffer area of a given socket.

The receive side works similarly.  One additional thing that
happens is that the actual size of the receive side socket level
buffer at a given time is used by TCP code to indicate the
receive window size to the peer.  For example, if the data
arrives from the peer and gets queued up into socket level
receive queue (of mbufs), but the reader (the task that is to
call read/recv routine) is not reading the data into task's
buffer from mbufs, then the TCP window will close down.
Sometimes this can happen due to programming errors.  Sometimes,
due to unique runtime behavior of VxWorks networking software,
this can also be caused by various other factors.  For example,
your reader task may not get to run due to low task priority,
while the netTask keeps queueing the input data in the socket
queue.  Some systems can exhibit stop-and-go type of behavior,
no matter how big the window size is.

Usually, programming TCP/IP application on VxWorks requires fine
tuning, especially if you have strict requirements for
throughput and latency.

Good luck.



- -----------------------------------------------------------

Got questions?  Get answers over the phone at Keen.com.
Up to 100 minutes free!
http://www.keen.com


---------------------------

Newsgroups: comp.os.vxworks
Subject: Ethernet sniffing
Date: Sun, 13 Aug 2000 11:04:26 +0200
From: "Arik Halperin" 
Organization: NetVision Israel	
Message-ID: <8n5kt6$n72$1@news.netvision.net.il>

Hello,
    Does anyone know the best way to sniff ethernet packets in VxWorks?

Regards,
    Arik Halperin



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Need a better FTP Server soft for VxWorks
Date: Sun, 13 Aug 2000 11:21:55 +0200
From: Leonid Rosenboim 
Organization: BitBand Technologies Ltd. http://www.bitband.com
Message-ID: <399668B2.5135F2FD@bitband.com>
References: <39930AD9.2F74F81A@nevision.net.il>
Reply-To: leonid@bitband.com

Ezra,

If you run FTP server on the target, it means you have (or should have)
installed DosFs 2.0,
thence you will find an enhanced FTP server in the "unsupported"
directory, in source so you
can enhance it more if you like to.

L'hitraot,
Leonid

Ezra Koper wrote:

> Hello All
>
> Does anyone know were i can find a better VxWorks FTP Server software?
>
> The FTP server software that is part of VxWorks is very basic.
> It doesn't  allow to define various restrictions based on user&pass.
> At the moment you log in you can do everything & access any folder.


---------------------------

End of New-News digest
**********************


From vxwexplo-errs@csg.lbl.gov  Sun Aug 13 10:38:52 2000
From: Fred Roeber 
Date: Sun Aug 13 10:38:54 PDT 2000
Subject: Re: vxWorks task suspend

On Sat, 12 Aug 2000, Thomas Martin  verified item
6 in the list below that was sent by Krishna Kishore Reddy
Sureddy . I just wanted to mention the
additional case that a task can be suspended if something like a bus
error happens while the task is active. This is sort of like item 1
in the list except that it can happen even if the task in question
does nothing "wrong". I am talking about the case where a device driver,
for instance, does a hardware access that generates an error. In such
a case the current task is normally suspended. There are lots of potential
variations on this scheme including bus errors and access violations due
to task level code, driver code, interrupt handlers...   Fred

> >    Can a task in VXWORKS get suspended other than the conditions
> > preseted below.
> >    If yes please send them across as soon  as possible.
> > 
> >    1. Task causing exception(task level exception).
> >    2. Suspending a task from shell by ts( )  or taskSuspend( ) .
> >    3. Attaching a task to Debugger.
> >    4. Keeping taskSuspend(0) in the code of the task.
> >    5. Keeping a break point in the code of the task.
> > 
> >    The below case i am not sure and correct me if i am wrong
> >    6. When you are trying to free/delete the memory which is already
> > deleted.

> You are right!
> free() and malloc() have default debugging options. The default for
> free() is (if a block is corrupted) suspend the task and log an
> error message. Options can be:
> MEM_ALLOC_ERROR_LOG_FLAG
> MEM_ALLOC_ERROR_SUSPEND_FLAG
> MEM_BLOCK_CHECK
> MEM_BLOCK_CHECK_LOG_FLAG
> MEM_BLOCK_CHECK_SUSPEND_FLAG
> These option can be set by memOptionsSet() and can be bitwise ORed.
> 
> HTH
> Thomas

   |      Fred J Roeber,  BBN Systems & Technologies      |
   |    4 John Clarke Road   Middletown, RI  02842-5202   |
   |          froeber@bbn.com     401-848-3548            |
   |    Division Scientist, High Performance Computing    |



From vxwexplo-errs@csg.lbl.gov  Sun Aug 13 12:44:27 2000
From: Fred Roeber 
Date: Sun Aug 13 12:44:29 PDT 2000
Subject: Re: serial communication

On Sat, 12 Aug 2000, "Keith W. Muzzioli"  wrote:

> We are having a problem with some serial communications running either
> VxWorks 5.1.1 or VxWorks 5.3, using a DY4 DMV-163 68040 board with a
> DMAX227 (Raytheon version of the 220) mezzanine card.
> 
> We are sending packets of data from the PC to the DMV-163 card
> of varying lengths, and started noticing errors with large packets.  If
> the packet size is greater than 640 Bytes  (for example 740 in our
> case), data gets "jumbled".  By jumbled, I mean that the first 640 bytes
> come through just fine, but then the next 48 bytes are skipped, and the
> LAST 64 bytes are sent next, followed by the skipped 48 bytes.  No
> data is actually LOST, just mixed up a little...

<>

> Has anyone seen behavior like this before??? Any suggestions would be greatly 
> appreciated.

Keith
We had a similar problem with the DY4 MAX 220 interfaces a few years
back.  Unfortunately, I tend to be in a position where I deal with BSP
issues for a lot of different vendors boards and run into a lot of
problems like this so the details sometimes become a little fuzzy. As I
remember, we got a listing of the DLCS software for the MAX 220 card
and found that there were some errors in it. The DLCS software is code
that runs on the MAX 220 card that was produced by a third party and
licensed by DY4. It provides a high level "command packet" driven
interface to the card and then implements all the device driver details
internally. There were some race conditions in the interface packet
handling code that could cause data to get misordered. This software
problem was correctable through some changes to the DLCS software. As I
remember (this is the fuzzy part) we fixed the problem and turned the
changes back over to DY4. You should probably try asking them about
this. Good luck.   Fred

   |      Fred J Roeber,  BBN Systems & Technologies      |
   |    4 John Clarke Road   Middletown, RI  02842-5202   |
   |          froeber@bbn.com     401-848-3548            |
   |    Division Scientist, High Performance Computing    |



From vxwexplo-errs@csg.lbl.gov  Sun Aug 13 13:14:46 2000
From: Fred Roeber 
Date: Sun Aug 13 13:14:48 PDT 2000
Subject: Re: select problem

On Sat, 12 Aug 2000, "Øyvind Holmeide"  wrote:

> I'm using the select() routine in an application that uses the console
> (standard in) either via the serial line or telnet.  The select routine
> returns when a byte is available.  However, the corresponding call to
> getc() will not see the first byte (only the following bytes).  "Syntax
> error" is also written to the console after the select() returns with 1.
> The code look like this:
>
>  timeout.tv_sec = 1;
>  timeout.tv_usec = 0;
>
>  while (!end) {
>     newInput = 0;
>     while (newInput < 1) {
>       FD_ZERO (&readFds);
>       FD_SET (STD_IN, &readFds);
>       newInput = select (FD_SETSIZE, &readFds, NULL, NULL, &timeout);
>     }
>   i = getc(stdin);

There is a "problem" with the select function that isn't very well
documented. That is the fact that when data is available on an interface,
the select function wakes up all tasks waiting for data on the interface
and lets them fight for the available data. Thus, if multiple tasks are
trying to get data from the same interface, when any data becomes available,
they will all be made runnable. The first one that runs (due to priority)
will do a read that will get the first available data. Based on how the
different tasks are written, other tasks may then be enabled to try to
read the same data.

At one point in the early 90s I had an email discussion on the issue
with Dave Wilner of WRS on the subject because I thought the VxWorks
code was broken with respect to the "standard" BSD implementation. He
convinced me that this VxWorks implementation was correct and even the
BSD implementation worked this way. The real problem back then, was
that there wasn't any support for non blocking terminal IO.

In the case of this example, if the target processor is running the
VxWorks "shell" that task (running at priority 1) will also be trying
to read from STDIN. It will wake up first and read the first character.
It will probably try to echo that character out (depending on terminal
mode) and block at that point. This will allow the user task to run and
get other characters. Needless to say, things can get messy due to
multi task interactions. My suggestion is to make sure that the user
task is running at higher priority than the shell and to make sure that
the standard input is configured to be in raw mode.

The details of the test above can vary based on exactly where STDIN is
coming from (i.e. a terminal line or an rlogin line). I think the main
thing to remember is that a select call doesn't guarantee that there
will be data to read on the indicated line and that, therefor, the
application can block if the input "channel" isn't set up to be non
blocking.

   |      Fred J Roeber,  BBN Systems & Technologies      |
   |    4 John Clarke Road   Middletown, RI  02842-5202   |
   |          froeber@bbn.com     401-848-3548            |
   |    Division Scientist, High Performance Computing    |



From vxwexplo-errs@csg.lbl.gov  Sun Aug 13 18:43:27 2000
From: Fred Roeber 
Date: Sun Aug 13 18:43:29 PDT 2000
Subject: Re: Debugging variables defined in nested blocks

On Thu, 10 Aug 2000, Edward Bishop  wrote:

> I would like to be able to be able to view the contents of the
> variable 'status' in code like this:
> 
> int bar(void)
> {
>     int blub;
> 
>     while(1)
>     {
>         int status = foo();
>         // break here and check status
>         ...
>     }
> }
> 
> but "p status" yields 'No symbol "status" in current context'.  There
> is no problem viewing variables such as blub.  I am using
> Tornado2 on a MIPS 5400 processor.  Are there any debugger or
> compiler options (besides -g of course) that I should be using?

Edward, The problem is that the variable status isn't used anywhere
in the code so the compiler is free to optimize it out of existence.
Even if it did leave the variable around, it would be free to stick
it in a register rather than a memory location. Some solutions
to get the compiler not to perform this optimization are to:
1) Turn off optimization of the file (-O0)
2) Declare the variable "volatile" so that the compiler will be forced
   to preserve it
	volatile int status = foo();
3) Declare the variable "static" so that it will be preserved between
   different executions of the function (but will render the code
   non reentrant)
	static int status = foo();

I think the second choice is usually your best bet.  Fred

   |      Fred J Roeber,  BBN Systems & Technologies      |
   |    4 John Clarke Road   Middletown, RI  02842-5202   |
   |          froeber@bbn.com     401-848-3548            |
   |    Division Scientist, High Performance Computing    |



From vxwexplo-errs@csg.lbl.gov  Mon Aug 14 03:31:24 2000
From: David Laight 
Date: Mon Aug 14 03:31:26 PDT 2000
Subject: Re: TCP retransmittions and lost packets -- New Question

Chris wrote:

> I have used this same code on the same hardware under vxworks with a slightly
> different networking environment (a different hub) and not had any problems.
> Can anyone explain why this would suddenly cause lost packets? My current
> "network" is two Synergy ppc sbc's and an NT laptop ethernet sniffer connected
> to a hub. The hub's uplink is connected to a switch and the switch is on our
> general network.
> Nothing on the general network is interacting with the hub.

Sounds to me like there is a 'fault' in your network.  These can be hard to 
track down!

I would simplify things further:
1) disconnect from the general network
2) try different cables (just in case one is duff)
3) use a different hub, make sure it is a simple one - they have less scope for 
introducing errors.
4) make ABSOLUTELY sure that nothing anywhere on the network your side of a 
store and forward switch (these usually route as well) has full duplex enabled.

> What ramifications are there to changing the send and receive buffers? I know 
> they change the maximum allowable window to be used for the tcp connection,
> but what other side effects does this have? Would external network components
> (the hub or switch) affect what the window should be? Is there an RFC that
> addresses this?

My guess is that enabling a large transmit window, together with the large 
receive window changes the data pattern on the LAN sufficiently to cause 
transmit bacxkoff (and/or collisions).  These are completely benign on a 
properly built network, but quickly show up any problems that do exist.

I don't know the ack strategy of the vxWorks TCP, but it is quite likely that is 
will ack every other received frame (this is typical of older implementations).  
With a window size of 60k and fast systems the network is likely to show the 
following traffic pattern:

	-> 1500 byte data frame
	-> 1500 byte data frame
		(target TCP sends an ack, but the LAN is busy)
	-> 1500 byte data frame
	-> 1500 byte data frame
		(another ack is generated, queued to be transmitted behind
		 the first one)
	-> 36 more data frames
		(generating 18 more acks, all get queued)
	(slight pause until target samples an idle LAN)
	<- 20 back to back ethernet frames each acking 2 more data packets
	(slight delay while all 20 acks are processes)
	
The sequence then repeats from the top.

Now if the collision detect and/or backoff algorithm isn't working (because of a 
broken cable or duff hub/switch) then frames will get lost.

Also an ethernet driver with inadequate transmit buffering will also drop 
packets.

If you unplug your uplink you should be able to see which packets are missing.

	David


(The above sequence is REAL - I saw it a few years ago using relatively slow (by 
current standards) hardware.  It can be avoided by putting two network cards in 
each system and cabling two equivatent LANs and getting different systems to 
route packets to different lans!)

----------------------------------------------------------------
David Laight                          email: dsl@tadpole.co.uk
Tadpole Technology plc                phone: +44 1223 428 232
Cambridge, UK                         fax:   +44 1223 428 201



From vxwexplo-errs@csg.lbl.gov  Mon Aug 14 04:03:06 2000
From: Vxworks Exploder 
Date: Mon Aug 14 04:03:08 PDT 2000
Subject: comp.os.vxworks newsdigest

Comp.Os.Vxworks Daily Digest    Mon Aug 14 04:03:03 PDT 2000

        Subject: Re: Tornado project facility
        Subject: Adaptec 2940U2W on VxWorks, anybody?
        Subject: Re: Anyone using the LSI sym53C895 SCSI controller on VxWorks
        Subject: re: vxWorks task suspend
        Subject: re: serial communication
        Subject: re: select problem
        Subject: re: Debugging variables defined in nested blocks
        Subject: Re: VxWorks FTP boot problem (T2, MVME2700)
        Subject: Re: Ethernet sniffing
        Subject: How does Insure++6.0 work?

-------------------------------------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Tornado project facility
Date: Sun, 13 Aug 2000 14:58:42 GMT
From: John Fusco 
Organization: Deja.com - Before you buy.
Message-ID: <8n6d31$6nk$1@nnrp1.deja.com>
References: <8n1jbh$ht5$1@overload.lbl.gov> <8n1n0p$iee$1@news1.xs4all.nl>

I would call that a bug.

Since the you can't build a bootrom in a project, you have to modify
config.h if you need to modify your bootrom (e.g. to make your kernel
bigger).  There is no sanity checking in Tornado to verify that the
changes you made to the bootrom are still compatible with your
project.  There is no tool in Tornado to tell you if your project is
still compatible with your bootrom.

You can create a new project based on the new bootrom to ensure that
they will be compatible, but there is no tool in Tornado that will
simply copy all the components from the old project.  You have to do
that by hand!  In essense you have to start all over.

Yes, I know the other alternative is not to use a bootrom.

Sorry for the rant, but this has been a thorn in my side for some
time.  Sloppy configuration management is such a big part of Tornado,
it's almost a feature.  There has been no thought put into
configuration management in Tornado.


> When you want to make changes to a project you have to do that in the
> project configuration and *NOT* in the config.h. Almost all the
defines
> you make in the config.h are redefined in the project files.
> When you create a new project it is based on the config.h in the BSP
> directory. But after that it is seperate from the config.h. So the
> config.h only provides the defaults for a project.
>
> Groeten,
> 	Johan


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Adaptec 2940U2W on VxWorks, anybody?
Date: Sun, 13 Aug 2000 17:29:11 +0300
From: Amit Margalit 
Organization: Conexant Systems
Message-ID: <3996B0B7.CE09F3@conexant.com>

This is a multi-part message in MIME format.
- --------------97012195570ADCC41F62C122
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Hello!

I have a Adaptec 2940U2W (AIC-7890) card, and would like to get VxWorks
to see it. I use Tornado 2, and VxWorks 5.4 on a x86 target.

Anyone has experience with this?

I was able to run a AIC-7870 card with the AIC-7880 driver, with minimal
effort (just forced sysScsi.c to search its Vendor ID and Device ID). But
the same work doesn't help with AIC-7890 - it gets hung in
aic7880CtrlCreate() for some reason.

I'd appreciate direct emails to:

Amit.Margalit@conexant.com

Thanks,

	Amit

- -- 
#include 
- -----
I had a feeling once about mathematics -- that I saw it all.  Depth beyond
depth was revealed to me -- the Byss and the Abyss. I saw -- as one might
see the transit of Venus or even the Lord Mayor's Show -- a quantity passing
through infinity and changing its sign from plus to minus.  I saw exactly
why it happened and why tergiversation was inevitable -- but it was after
dinner and I let it go.
		-- Winston Churchill
- --------------97012195570ADCC41F62C122
Content-Type: text/x-vcard; charset=iso-8859-1;
 name="Amit.Margalit.vcf"
Content-Transfer-Encoding: base64
Content-Description: Card for Amit Margalit
Content-Disposition: attachment;
 filename="Amit.Margalit.vcf"

YmVnaW46dmNhcmQgCm46TWFyZ2FsaXQ7QW1pdAp0ZWw7ZmF4OisoOTcyKS05LTk1NzM3MzIK
dGVsO3dvcms6Kyg5NzIpLTktOTUyNDA5Mwp4LW1vemlsbGEtaHRtbDpGQUxTRQpvcmc6Q29u
ZXhhbnQgSXNyYWVsIEx0ZC47TkFECnZlcnNpb246Mi4xCmVtYWlsO2ludGVybmV0OkFtaXQu
TWFyZ2FsaXRAY29uZXhhbnQuY29tCnRpdGxlOkRyaXZlciBMZWFkCmFkcjtxdW90ZWQtcHJp
bnRhYmxlOjs7MTEgR2FsZ2FsZXkgSGFwbGFkYSBzdC49MEQ9MEFQLk8uQm94IDEyNjYwO0hl
cnpsaWE7OzQ2NzMzO0lTUkFFTAp4LW1vemlsbGEtY3B0OjstOTA0MApmbjpBbWl0IE1hcmdh
bGl0CmVuZDp2Y2FyZAo=
- --------------97012195570ADCC41F62C122--


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Anyone using the LSI sym53C895 SCSI controller on VxWorks
Date: Sun, 13 Aug 2000 17:31:48 +0300
From: Amit Margalit 
Organization: Conexant Systems
Message-ID: <3996B154.5F998B3A@conexant.com>
References: <8lqegd$ap0$1@nnrp1.deja.com>

This is a multi-part message in MIME format.
- --------------BB815363C553F9B16B2F5290
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Hi!

No, but I want to. Am waiting to get it. I have some experience with the
53c810 driver, though.

	Amit

syee@connex.com wrote:
> 
> Has anyone using the LSI sym53C895 SCSI controller got it to work on
> VxWorks?
> 
> Sent via Deja.com http://www.deja.com/
> Before you buy.

- -- 
#include 
- -----
I had a feeling once about mathematics -- that I saw it all.  Depth beyond
depth was revealed to me -- the Byss and the Abyss. I saw -- as one might
see the transit of Venus or even the Lord Mayor's Show -- a quantity passing
through infinity and changing its sign from plus to minus.  I saw exactly
why it happened and why tergiversation was inevitable -- but it was after
dinner and I let it go.
		-- Winston Churchill
- --------------BB815363C553F9B16B2F5290
Content-Type: text/x-vcard; charset=iso-8859-1;
 name="Amit.Margalit.vcf"
Content-Transfer-Encoding: base64
Content-Description: Card for Amit Margalit
Content-Disposition: attachment;
 filename="Amit.Margalit.vcf"

YmVnaW46dmNhcmQgCm46TWFyZ2FsaXQ7QW1pdAp0ZWw7ZmF4OisoOTcyKS05LTk1NzM3MzIK
dGVsO3dvcms6Kyg5NzIpLTktOTUyNDA5Mwp4LW1vemlsbGEtaHRtbDpGQUxTRQpvcmc6Q29u
ZXhhbnQgSXNyYWVsIEx0ZC47TkFECnZlcnNpb246Mi4xCmVtYWlsO2ludGVybmV0OkFtaXQu
TWFyZ2FsaXRAY29uZXhhbnQuY29tCnRpdGxlOkRyaXZlciBMZWFkCmFkcjtxdW90ZWQtcHJp
bnRhYmxlOjs7MTEgR2FsZ2FsZXkgSGFwbGFkYSBzdC49MEQ9MEFQLk8uQm94IDEyNjYwO0hl
cnpsaWE7OzQ2NzMzO0lTUkFFTAp4LW1vemlsbGEtY3B0OjstOTA0MApmbjpBbWl0IE1hcmdh
bGl0CmVuZDp2Y2FyZAo=
- --------------BB815363C553F9B16B2F5290--


---------------------------

Newsgroups: comp.os.vxworks
Subject: re: vxWorks task suspend
Date: Sun, 13 Aug 2000 13:32:43 -0400 (EDT)
From: Fred Roeber 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <8n6oj1$he4$1@overload.lbl.gov>

On Sat, 12 Aug 2000, Thomas Martin  verified item
6 in the list below that was sent by Krishna Kishore Reddy
Sureddy . I just wanted to mention the
additional case that a task can be suspended if something like a bus
error happens while the task is active. This is sort of like item 1
in the list except that it can happen even if the task in question
does nothing "wrong". I am talking about the case where a device driver,
for instance, does a hardware access that generates an error. In such
a case the current task is normally suspended. There are lots of potential
variations on this scheme including bus errors and access violations due
to task level code, driver code, interrupt handlers...   Fred

> >    Can a task in VXWORKS get suspended other than the conditions
> > preseted below.
> >    If yes please send them across as soon  as possible.
> > 
> >    1. Task causing exception(task level exception).
> >    2. Suspending a task from shell by ts( )  or taskSuspend( ) .
> >    3. Attaching a task to Debugger.
> >    4. Keeping taskSuspend(0) in the code of the task.
> >    5. Keeping a break point in the code of the task.
> > 
> >    The below case i am not sure and correct me if i am wrong
> >    6. When you are trying to free/delete the memory which is already
> > deleted.

> You are right!
> free() and malloc() have default debugging options. The default for
> free() is (if a block is corrupted) suspend the task and log an
> error message. Options can be:
> MEM_ALLOC_ERROR_LOG_FLAG
> MEM_ALLOC_ERROR_SUSPEND_FLAG
> MEM_BLOCK_CHECK
> MEM_BLOCK_CHECK_LOG_FLAG
> MEM_BLOCK_CHECK_SUSPEND_FLAG
> These option can be set by memOptionsSet() and can be bitwise ORed.
> 
> HTH
> Thomas

   |      Fred J Roeber,  BBN Systems & Technologies      |
   |    4 John Clarke Road   Middletown, RI  02842-5202   |
   |          froeber@bbn.com     401-848-3548            |
   |    Division Scientist, High Performance Computing    |


---------------------------

Newsgroups: comp.os.vxworks
Subject: re: serial communication
Date: Sun, 13 Aug 2000 15:38:18 -0400 (EDT)
From: Fred Roeber 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <8n6vk0$jso$1@overload.lbl.gov>

On Sat, 12 Aug 2000, "Keith W. Muzzioli"  wrote:

> We are having a problem with some serial communications running either
> VxWorks 5.1.1 or VxWorks 5.3, using a DY4 DMV-163 68040 board with a
> DMAX227 (Raytheon version of the 220) mezzanine card.
> 
> We are sending packets of data from the PC to the DMV-163 card
> of varying lengths, and started noticing errors with large packets.  If
> the packet size is greater than 640 Bytes  (for example 740 in our
> case), data gets "jumbled".  By jumbled, I mean that the first 640 bytes
> come through just fine, but then the next 48 bytes are skipped, and the
> LAST 64 bytes are sent next, followed by the skipped 48 bytes.  No
> data is actually LOST, just mixed up a little...

<>

> Has anyone seen behavior like this before??? Any suggestions would be greatly 
> appreciated.

Keith
We had a similar problem with the DY4 MAX 220 interfaces a few years
back.  Unfortunately, I tend to be in a position where I deal with BSP
issues for a lot of different vendors boards and run into a lot of
problems like this so the details sometimes become a little fuzzy. As I
remember, we got a listing of the DLCS software for the MAX 220 card
and found that there were some errors in it. The DLCS software is code
that runs on the MAX 220 card that was produced by a third party and
licensed by DY4. It provides a high level "command packet" driven
interface to the card and then implements all the device driver details
internally. There were some race conditions in the interface packet
handling code that could cause data to get misordered. This software
problem was correctable through some changes to the DLCS software. As I
remember (this is the fuzzy part) we fixed the problem and turned the
changes back over to DY4. You should probably try asking them about
this. Good luck.   Fred

   |      Fred J Roeber,  BBN Systems & Technologies      |
   |    4 John Clarke Road   Middletown, RI  02842-5202   |
   |          froeber@bbn.com     401-848-3548            |
   |    Division Scientist, High Performance Computing    |


---------------------------

Newsgroups: comp.os.vxworks
Subject: re: select problem
Date: Sun, 13 Aug 2000 16:08:38 -0400 (EDT)
From: Fred Roeber 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <8n6vk1$jsp$1@overload.lbl.gov>

On Sat, 12 Aug 2000, "Øyvind Holmeide"  wrote:

> I'm using the select() routine in an application that uses the console
> (standard in) either via the serial line or telnet.  The select routine
> returns when a byte is available.  However, the corresponding call to
> getc() will not see the first byte (only the following bytes).  "Syntax
> error" is also written to the console after the select() returns with 1.
> The code look like this:
>
>  timeout.tv_sec = 1;
>  timeout.tv_usec = 0;
>
>  while (!end) {
>     newInput = 0;
>     while (newInput < 1) {
>       FD_ZERO (&readFds);
>       FD_SET (STD_IN, &readFds);
>       newInput = select (FD_SETSIZE, &readFds, NULL, NULL, &timeout);
>     }
>   i = getc(stdin);

There is a "problem" with the select function that isn't very well
documented. That is the fact that when data is available on an interface,
the select function wakes up all tasks waiting for data on the interface
and lets them fight for the available data. Thus, if multiple tasks are
trying to get data from the same interface, when any data becomes available,
they will all be made runnable. The first one that runs (due to priority)
will do a read that will get the first available data. Based on how the
different tasks are written, other tasks may then be enabled to try to
read the same data.

At one point in the early 90s I had an email discussion on the issue
with Dave Wilner of WRS on the subject because I thought the VxWorks
code was broken with respect to the "standard" BSD implementation. He
convinced me that this VxWorks implementation was correct and even the
BSD implementation worked this way. The real problem back then, was
that there wasn't any support for non blocking terminal IO.

In the case of this example, if the target processor is running the
VxWorks "shell" that task (running at priority 1) will also be trying
to read from STDIN. It will wake up first and read the first character.
It will probably try to echo that character out (depending on terminal
mode) and block at that point. This will allow the user task to run and
get other characters. Needless to say, things can get messy due to
multi task interactions. My suggestion is to make sure that the user
task is running at higher priority than the shell and to make sure that
the standard input is configured to be in raw mode.

The details of the test above can vary based on exactly where STDIN is
coming from (i.e. a terminal line or an rlogin line). I think the main
thing to remember is that a select call doesn't guarantee that there
will be data to read on the indicated line and that, therefor, the
application can block if the input "channel" isn't set up to be non
blocking.

   |      Fred J Roeber,  BBN Systems & Technologies      |
   |    4 John Clarke Road   Middletown, RI  02842-5202   |
   |          froeber@bbn.com     401-848-3548            |
   |    Division Scientist, High Performance Computing    |


---------------------------

Newsgroups: comp.os.vxworks
Subject: re: Debugging variables defined in nested blocks
Date: Sun, 13 Aug 2000 21:37:17 -0400 (EDT)
From: Fred Roeber 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <8n7kn1$r4b$1@overload.lbl.gov>

On Thu, 10 Aug 2000, Edward Bishop  wrote:

> I would like to be able to be able to view the contents of the
> variable 'status' in code like this:
> 
> int bar(void)
> {
>     int blub;
> 
>     while(1)
>     {
>         int status = foo();
>         // break here and check status
>         ...
>     }
> }
> 
> but "p status" yields 'No symbol "status" in current context'.  There
> is no problem viewing variables such as blub.  I am using
> Tornado2 on a MIPS 5400 processor.  Are there any debugger or
> compiler options (besides -g of course) that I should be using?

Edward, The problem is that the variable status isn't used anywhere
in the code so the compiler is free to optimize it out of existence.
Even if it did leave the variable around, it would be free to stick
it in a register rather than a memory location. Some solutions
to get the compiler not to perform this optimization are to:
1) Turn off optimization of the file (-O0)
2) Declare the variable "volatile" so that the compiler will be forced
   to preserve it
	volatile int status = foo();
3) Declare the variable "static" so that it will be preserved between
   different executions of the function (but will render the code
   non reentrant)
	static int status = foo();

I think the second choice is usually your best bet.  Fred

   |      Fred J Roeber,  BBN Systems & Technologies      |
   |    4 John Clarke Road   Middletown, RI  02842-5202   |
   |          froeber@bbn.com     401-848-3548            |
   |    Division Scientist, High Performance Computing    |


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: VxWorks FTP boot problem (T2, MVME2700)
Date: Mon, 14 Aug 2000 03:30:59 GMT
From: paulvroberts@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8n7p5j$48b$1@nnrp1.deja.com>
References: <8mdcif$tgl$1@nnrp1.deja.com> <398A95D7.A8BBCE60@auriga.ru> <8mls2f$keu$1@nnrp1.deja.com> <399122DD.25CBAA52@auriga.ru>

For those of you who are interested:

Some more research turned up a plausible explanation of the problem on
the wu ftp web site (the Red Hat Linux ftp server) and the fix was
pretty simple.

Basically, the ftp server has been made more RFC "insistent" and this
caused the VxWorks boot code to break (it also broke Lynx,
Midnight Commander and various other programs so WRS were not alone!).

The patch is to close the data connection immediately after
the read returns EOF (ie the other end has signalled a close). This is
consistent with the FTP RFC 959.

The patch should be safe (the data connection is no longer used), and it
works for me with the ftp daemon (wu ftpd) on Linux Red Hat 6.0, 6.2, as
well as the ftp supplied fom WRS for Windows - wftpd version 2.03.

So, in bootConfig.c, in the 2 places where the connection is terminated
(normal case and error case). Move the close(fd) up after the read. That
is:

    /* Empty the Data Socket before close. PC FTP server hangs otherwise
*/

    while ((read (fd, command, sizeof (command))) > 0);
    /* Close data connection now
    close (fd);

See http://www.wu-ftpd.org/broken-clients.html for more info.

Arthur: Thanks for the help,


Paul Roberts.


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Ethernet sniffing
Date: Sun, 13 Aug 2000 21:16:28 -0700
From: Hwajin Bae 
Organization: http://www.remarq.com: The World's Usenet/Discussions Start Here
Message-ID: <1f35b05a.5175dbbd@usw-ex0104-033.remarq.com>
References: <8n5kt6$n72$1@news.netvision.net.il>

I don't know of the "best" way, but here is a quick and dirty
way.

/*
 * quick and dirty ethernet sniffer for vxworks
 */

#include "vxWorks.h"
#include "sys/types.h"
#include "ioLib.h"
#include "iosLib.h"
#include "fioLib.h"
#include "stdio.h"
#include "unistd.h"
#include "string.h"
#include "usrLib.h"
#include "errnoLib.h"
#include "hostLib.h"
#include "sockLib.h"
#include "socket.h"
#include "inetLib.h"
#include "in.h"
#include "selectLib.h"
#include "taskLib.h"
#include "dosFsLib.h"
#include "etherLib.h"
#include "systm.h"
#include "mbuf.h"
#include "domain.h"
#include "tickLib.h"
#include "protosw.h"
#include "socket.h"
#include "errno.h"
#include "if.h"
#include "route.h"
#include "in.h"
#include "in_pcb.h"
#include "in_systm.h"
#include "in_var.h"
#include "ip.h"
#include "ip_var.h"
#include "ip_icmp.h"
#include "udp.h"
#include "tcp.h"
#include "if_ether.h"
#include "systm.h"
#include "inetLib.h"

char vxSniffLog[200][200];
int vxSniffLogIdx= 0;

STATUS
vxSniffHook(struct ifnet *ifp, char *buf, int len)
{
	struct ip *ip;
	struct  udphdr *udp;
	struct tcphdr *tcp;
	unsigned long ticks;
	char *proto;
	int optlen;
	char srcAddr[20], dstAddr[20];
	char protoxxx[10];

	ip = (struct ip *)(buf + 14);
	inet_ntoa_b(ip->ip_src, srcAddr);
	inet_ntoa_b(ip->ip_dst, dstAddr);
	switch (ip->ip_p) {
	case IPPROTO_TCP:
		proto = "TCP";
		break;
	case IPPROTO_UDP:
		proto = "UDP";
		break;
	case IPPROTO_ICMP:
		proto = "ICMP";
		break;
	default:
		sprintf(protoxxx, "?0x%x", ip->ip_p);
		proto = protoxxx;
		break;
	}

	ticks = tickGet ();
	
	sprintf(vxSniffLog[vxSniffLogIdx++],
		"0x%x:IP(p:%s,v:%d,h:%d,l:%d) %s -> %s\n",
		ticks,
		proto,
		(ip->ip_v_hl>>4)&0xf,
		(ip->ip_v_hl & 0xf)<<2, ip->ip_len,
		srcAddr, dstAddr);

	if (vxSniffLogIdx == 200)
		vxSniffLogIdx = 0;

	optlen = ((ip->ip_v_hl & 0xf)<<2) - sizeof( struct ip );
	if (ip->ip_p == IPPROTO_UDP) {
		udp = (struct udphdr *)((char *)(ip + 1) +
optlen);
		sprintf(vxSniffLog[vxSniffLogIdx++],
			"    UDP(s:%d,d:%d,l:%d)\n",
			udp->uh_sport, udp->uh_dport,
			udp->uh_ulen);
	} else if (ip->ip_p == IPPROTO_TCP) {
		tcp = (struct tcphdr *)((char *)(ip + 1) +
optlen);
		sprintf(vxSniffLog[vxSniffLogIdx++],
			"
TCP(s:%d,d:%d,q:%d,a:%d,f:0x%x,w:%d)\n",
			tcp->th_sport, tcp->th_dport,
			tcp->th_seq, tcp->th_ack, tcp->th_flags,
			tcp->th_win);
	}

	if (vxSniffLogIdx == 200)
		vxSniffLogIdx = 0;

	return FALSE;
}

int vxSniffStartIdx = 0;

void
vxSniffStart()
{
	vxSniffStartIdx = vxSniffLogIdx;
}

void
vxSniffShow()
{
	int end = vxSniffLogIdx;
	int i;

	if (end < vxSniffStartIdx) {
		for (i = vxSniffStartIdx; i < 200; i++)
			printf("%s", vxSniffLog[i]);
		for (i = 0; i < end; i++)
			printf("%s", vxSniffLog[i]);
	} else {
		for (i = vxSniffStartIdx; i < end; i++)
			printf("%s", vxSniffLog[i]);
	}
}

void
vxSniffInit()
{
	/* XXX do something here to put ethernet driver in
promiscuous mode */

	etherInputHookAdd(vxSniffHook);
}



- -----------------------------------------------------------

Got questions?  Get answers over the phone at Keen.com.
Up to 100 minutes free!
http://www.keen.com


---------------------------

Newsgroups: comp.os.vxworks
Subject: How does Insure++6.0 work?
Date: Mon, 14 Aug 2000 08:54:15 GMT
From: Harvey Tang 
Organization: Deja.com - Before you buy.
Message-ID: <8n8c3m$hd0$1@nnrp1.deja.com>

Hello,

 Who had even used Insure++6.0 to analyze the C code for VxWorks
PowerPC) in Windows NT 4.0(X86 and Service Pack 6) and Tornado 2.0?
How to modify the configuration of Insure++6.0?

Thanx.

Harvey


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

End of New-News digest
**********************


From vxwexplo-errs@csg.lbl.gov  Mon Aug 14 07:57:22 2000
From: Bill Brown 
Date: Mon Aug 14 07:57:24 PDT 2000
Subject: Re: vxWorks task suspend

At 10:52 AM 8/13/00 -0700, you wrote:
>Submitted-by vxwexplo-errs@csg.lbl.gov  Sun Aug 13 10:38:52 2000
>Submitted-by: Fred Roeber 
>
>On Sat, 12 Aug 2000, Thomas Martin  verified item
>6 in the list below that was sent by Krishna Kishore Reddy
>Sureddy . I just wanted to mention the
>additional case that a task can be suspended if something like a bus
>error happens while the task is active. This is sort of like item 1
>in the list except that it can happen even if the task in question
>does nothing "wrong". I am talking about the case where a device driver,
>for instance, does a hardware access that generates an error. In such
>a case the current task is normally suspended. There are lots of potential
>variations on this scheme including bus errors and access violations due
>to task level code, driver code, interrupt handlers...   Fred
>

Not to mention a bus error caused by a DMA cycle which was set up by
a totally different task.  That can really cause confusion!  vxWorks
makes the assumption that the error was caused by the currrently
executing task, which is about all it can do.

-bill


From vxwexplo-errs@csg.lbl.gov  Tue Aug 15 03:57:08 2000
From: Henrik Bergstrom 
Date: Tue Aug 15 03:57:10 PDT 2000
Subject: Multiple instances of same (modern) NIC in an x86 target?

Dear VxWorks-users,

I am having trouble finding a modern, commercially available, NIC which
works with several instances of the same card in an x86 target. Some
drivers are prepared for multiple instances, but don't work in
practice.

I am in need of a END (MUX) driver and the only one my local
vxw-representative knew of was Intel's 82559, but I don't have access
to that NIC.

What would make me really happy are END/MUX-drivers with support for
multiple instances of the same card for any of the following:

  o 3com 3c905C-TX "Fast EtherLink XL PCI"
    (This has "theoretical support" for multiple instances, but not
    fully so in practice.)
  o SMC 9432TX "EtherPower II 10/100"
  o any other modern, commercially available, card

Any help, tips or recommendations is appreciated.


Best regards,

Henrik Bergstrom

- --
Henrik Bergstrom  +46-8-628 28 28  henrik.bergstrom@intertex.se
Intertex Data AB  (fax) 628 64 14




From vxwexplo-errs@csg.lbl.gov  Tue Aug 15 04:03:08 2000
From: Vxworks Exploder 
Date: Tue Aug 15 04:03:10 PDT 2000
Subject: comp.os.vxworks newsdigest

Comp.Os.Vxworks Daily Digest    Tue Aug 15 04:03:04 PDT 2000

        Subject: Re: Tornado project facility
        Subject: re: TCP retransmittions and lost packets -- New Question
        Subject: Re: Ethernet sniffing
        Subject: New link layer
        Subject: ARCnet NPT driver
        Subject: Implement password masking on user input ?
        Subject: re: vxWorks task suspend
        Subject: Target Server Connect Problem?
        Subject: Ethernet questions
        Subject: Re: TCP retransmittions and lost packets -- New Question
        Subject: Re: TCP retransmittions and lost packets
        Subject: DosFs 2.0 in vxWorks (ATA disks)..
        Subject: VxWorks contractors needed in Chicago
        Subject: signal problem
        Subject: motFecPollSend
        Subject: How to porting TERMIOS or SGTTY?
        Subject: Re: Debugging variables defined in nested blocks
        Subject: Re: mv2700 duplicate int! why?
        Subject: Tornado project facility
        Subject: WindView file format
        Subject: Re: Implement password masking on user input ?
        Subject: Re: Target Server Connect Problem?
        Subject: Re: DosFs 2.0 in vxWorks (ATA disks)..
        Subject: [Q] Fatal WTX (0x10129) Synchronization Error

-------------------------------------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Tornado project facility
Date: 14 Aug 2000 11:09:19 GMT
From: johan@borksoft.xs4all.nl
Organization: Borksoft
Message-ID: <8n8k0v$pvm$1@news1.xs4all.nl>
References: <8n1jbh$ht5$1@overload.lbl.gov> <8n1n0p$iee$1@news1.xs4all.nl> <8n6d31$6nk$1@nnrp1.deja.com>

John Fusco  wrote:
: Since the you can't build a bootrom in a project, you have to modify
: config.h if you need to modify your bootrom (e.g. to make your kernel
: bigger).  There is no sanity checking in Tornado to verify that the
: changes you made to the bootrom are still compatible with your
: project.  There is no tool in Tornado to tell you if your project is
: still compatible with your bootrom.

Your bootable project does not have to be compatible with the bootrom, it
has to be compatible with the HW. After the bootrom is finished it starts
the VxWorks system from the start-vector. From here on all HW will be
initialised, even if it was used in the bootstrap.

If you want to know if you downloadbale application can be loaded on your
system you have to download it and check if all symbols can be resolved.
The only other way is to link your downloadable application against the
bootable kernel you are using. As you can download an application to any
system with the same processor architecture there is no standard check to
see if it is compatible with you kernel.

If you need a bootable image (not the bootstrap the boots another
application, but a booting VxWorks image) you can select for example a
vxworks_romResident. This will boot-up your system, and you can configure
it using the GUI.

: You can create a new project based on the new bootrom to ensure that
: they will be compatible, but there is no tool in Tornado that will
: simply copy all the components from the old project.  You have to do
: that by hand!  In essense you have to start all over.

You can also create a new project based on an existing project. If you
select Create a bootable  VxWorks Image and in step 2 you can select  to
base your new project on an existing project. You can select the project
from a list (these are the projects that in Target\Proj) or select one by
hand by clicking on the "..."-button nect to the directory path.


Groeten,
	Johan

- -- 
     o o o o o o o . . .   ___________________________________
    o      _____           ||    Johan Borkhuis              |
  .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
 >(________|__|_[_________]_|________________________________|
 _/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: TCP retransmittions and lost packets -- New Question
Date: Mon, 14 Aug 2000 11:31:34 +0100 (BST)
From: David Laight 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <200008141031.LAA15514@dsl-2.tadpole.co.uk>

Chris wrote:

> I have used this same code on the same hardware under vxworks with a slightly
> different networking environment (a different hub) and not had any problems.
> Can anyone explain why this would suddenly cause lost packets? My current
> "network" is two Synergy ppc sbc's and an NT laptop ethernet sniffer connected
> to a hub. The hub's uplink is connected to a switch and the switch is on our
> general network.
> Nothing on the general network is interacting with the hub.

Sounds to me like there is a 'fault' in your network.  These can be hard to 
track down!

I would simplify things further:
1) disconnect from the general network
2) try different cables (just in case one is duff)
3) use a different hub, make sure it is a simple one - they have less scope for 
introducing errors.
4) make ABSOLUTELY sure that nothing anywhere on the network your side of a 
store and forward switch (these usually route as well) has full duplex enabled.

> What ramifications are there to changing the send and receive buffers? I know 
> they change the maximum allowable window to be used for the tcp connection,
> but what other side effects does this have? Would external network components
> (the hub or switch) affect what the window should be? Is there an RFC that
> addresses this?

My guess is that enabling a large transmit window, together with the large 
receive window changes the data pattern on the LAN sufficiently to cause 
transmit bacxkoff (and/or collisions).  These are completely benign on a 
properly built network, but quickly show up any problems that do exist.

I don't know the ack strategy of the vxWorks TCP, but it is quite likely that is 
will ack every other received frame (this is typical of older implementations).  
With a window size of 60k and fast systems the network is likely to show the 
following traffic pattern:

	-> 1500 byte data frame
	-> 1500 byte data frame
		(target TCP sends an ack, but the LAN is busy)
	-> 1500 byte data frame
	-> 1500 byte data frame
		(another ack is generated, queued to be transmitted behind
		 the first one)
	-> 36 more data frames
		(generating 18 more acks, all get queued)
	(slight pause until target samples an idle LAN)
	<- 20 back to back ethernet frames each acking 2 more data packets
	(slight delay while all 20 acks are processes)
	
The sequence then repeats from the top.

Now if the collision detect and/or backoff algorithm isn't working (because of a 
broken cable or duff hub/switch) then frames will get lost.

Also an ethernet driver with inadequate transmit buffering will also drop 
packets.

If you unplug your uplink you should be able to see which packets are missing.

	David


(The above sequence is REAL - I saw it a few years ago using relatively slow (by 
current standards) hardware.  It can be avoided by putting two network cards in 
each system and cabling two equivatent LANs and getting different systems to 
route packets to different lans!)

- ----------------------------------------------------------------
David Laight                          email: dsl@tadpole.co.uk
Tadpole Technology plc                phone: +44 1223 428 232
Cambridge, UK                         fax:   +44 1223 428 201


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Ethernet sniffing
Date: 14 Aug 2000 12:47:19 GMT
From: cedilnik andrej 
Organization: University of Maryland, Baltimore County
Message-ID: <8n8pon$mmm$1@news.umbc.edu>
References: <8n5kt6$n72$1@news.netvision.net.il> <1f35b05a.5175dbbd@usw-ex0104-033.remarq.com>

> I don't know of the "best" way, but here is a quick and dirty
> way.

> void
> vxSniffInit()
> {
> 	/* XXX do something here to put ethernet driver in
> promiscuous mode */

> 	etherInputHookAdd(vxSniffHook);
> }

Are you sure that this works?
(it does not work for me: t2 and VxWorks 5.4).
The prototype of etherInputHookAdd()
looks like this:

STATUS etherInputHookAdd
    (
    FUNCPTR inputHook, /* routine to receive Ethernet input */
    char*   pName,     /* name of device if MUX/END is being used */
    int     unit       /* unit of device if MUX/END is being used */
    )

What is pName and unit?
I put pName "ln" and unit 0. If you do it your way, it 
returns -1. If you do it my way it returns 0 and still 
does not work.

				Andy



---------------------------

Newsgroups: comp.os.vxworks
Subject: New link layer
Date: 14 Aug 2000 12:50:15 GMT
From: cedilnik andrej 
Organization: University of Maryland, Baltimore County
Message-ID: <8n8pu7$mmm$2@news.umbc.edu>

I am developing new link layer for VxWorks. It will be below 
IP and it will carry IP packages.
Any simple idea to interface it with VxWorks stack? 
I would like to use TCP/IP from VxWorks distribution.

Also, is there any simple open source TCP/IP stack, that 
would work on VxWorks?

				Andy

---------------------------

Newsgroups: comp.os.vxworks
Subject: ARCnet NPT driver
Date: Mon, 14 Aug 2000 13:10:05 GMT
From: janrouvillain@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8n8r3b$r9u$1@nnrp1.deja.com>

Hi,

I am looking for a arcnet npt driver for Tornado 2.0.1. It must support
tcp/ip (ATA/ansi standard 878.2) and raw arcnet (ATA/ansi standard
878.1).

If you have interest in a arcnet npt driver or has programmed an arcnet
driver, then I will be glad to hear from you.

Regards
Jan Rouvillain, Cotas Computer Technology A/S


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Implement password masking on user input ?
Date: Mon, 14 Aug 2000 12:31:57 -0700
From: "Jeremy Pye" 
Organization: Customer of Energis Squared
Message-ID: 

I am writing an application for embedded system that will
allow configuration through the serial port using a terminal
emulator at the PC end.

Similar if you like to configuring a router.

Has anyone got any ideas on how to program the passwd
entry so that the user input is not echoed to the terminal
screen.

i.e passwd: *******






---------------------------

Newsgroups: comp.os.vxworks
Subject: re: vxWorks task suspend
Date: Mon, 14 Aug 2000 07:57:20 -0700
From: Bill Brown 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <4.1.20000814075325.01d07e00@imap4.lbl.gov>

At 10:52 AM 8/13/00 -0700, you wrote:
>Submitted-by vxwexplo-errs@csg.lbl.gov  Sun Aug 13 10:38:52 2000
>Submitted-by: Fred Roeber 
>
>On Sat, 12 Aug 2000, Thomas Martin  verified item
>6 in the list below that was sent by Krishna Kishore Reddy
>Sureddy . I just wanted to mention the
>additional case that a task can be suspended if something like a bus
>error happens while the task is active. This is sort of like item 1
>in the list except that it can happen even if the task in question
>does nothing "wrong". I am talking about the case where a device driver,
>for instance, does a hardware access that generates an error. In such
>a case the current task is normally suspended. There are lots of potential
>variations on this scheme including bus errors and access violations due
>to task level code, driver code, interrupt handlers...   Fred
>

Not to mention a bus error caused by a DMA cycle which was set up by
a totally different task.  That can really cause confusion!  vxWorks
makes the assumption that the error was caused by the currrently
executing task, which is about all it can do.

- -bill

---------------------------

Newsgroups: comp.os.vxworks
Subject: Target Server Connect Problem?
Date: Mon, 14 Aug 2000 10:31:34 -0400
From: "Mike Kordik" 
Message-ID: <8oTl5.128$Vi5.400@client>

I am new to VxWorks and I cannot get the target server to work with my
target. I have the WDB components included in my VxWorks image and
networking is working. I can FTP, WebServe, telnet...  When I start the
target server and try the "i" command in the target shell I get the
following error:

WTX Error 0x100d4 (AGENT_NO_AGENT_PROC)
value = -1 = 0xffffffff

What am I doing wrong?

Thanks,
Mike



---------------------------

Newsgroups: comp.os.vxworks
Subject: Ethernet questions
Date: Mon, 14 Aug 2000 18:01:20 +0200
From: "Arik Halperin" 
Organization: NetVision Israel	
Message-ID: <8n91n1$8jn$1@news.netvision.net.il>

1. What is the best way to perform Ethernet sniffing?(Using an end device)

2. How do you make sure the device driver for the Ethernet card(DEC21140 in
our case) works in promiscuous mode?

3. Is there a way to use a snarf protocol without disabling debugging
capabilities?

I managed to perform sniffing by changing the end debug agent's code,but
some how I feel this is not very elegant. I'll be glad to know if there is a
better way.

Regards,

    Arik Halperin








---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: TCP retransmittions and lost packets -- New Question
Date: Mon, 14 Aug 2000 10:03:18 -0500
From: Chris Smith 
Organization: Soutwest Research Institute
Message-ID: <39980A36.2620C00A@here.org>
References: <200008111735.SAA11853@dsl-2.tadpole.co.uk> <399477ED.58C95E0@here.org> <0eda9159.dc1845f1@usw-ex0104-033.remarq.com>
Reply-To: cbsmithATswriDoTorg



Hwajin Bae wrote:

> The send side socket buffer size is used by the TCP/IP stack to
> throttle the data output.  The user data is copied into mbufs at
> the socket level and put into socket level buffers, which are
> limited at the size specified.  The mbufs for each socket is not
> preallocated; they are allocated as needed (as more data is
> copied into mbufs from user buffers).  The send side socket
> level buffer size is used by the TCP/IP stack to limit this
> function (allocation and copying into mbufs), pending the task
> that is sending data when limit is reached.  Pending is done by
> taking a semaphore which will be given later by the stack when
> the queued data is actually sent over the TCP to the peer and
> ACK'ed and mbuf is freed, thereby creating more room in the send
> side buffer area of a given socket.
>
> The receive side works similarly.  One additional thing that
> happens is that the actual size of the receive side socket level
> buffer at a given time is used by TCP code to indicate the
> receive window size to the peer.  For example, if the data
> arrives from the peer and gets queued up into socket level
> receive queue (of mbufs), but the reader (the task that is to
> call read/recv routine) is not reading the data into task's
> buffer from mbufs, then the TCP window will close down.

Do you mean the RWIN field (in each netwrok packet) will actually change
as a receiver gets more data backed up? I initially thought that was how
the window 'mechanism' is supposed to work, but it does not actually
appear to change once a connection is established -- as seen from a
packet sniffer. I now think the window fields are a starting point for
the "congestion window" estimator on the sending side. see rfc for more
info:
http://www.ietf.org/rfc/rfc2581.txt

Actually the windows also do not appear to synchronize between the two
ends of a connection like I though they would. They each appear to just
continually send their initial values.

If you (or anyone else) know how this is actually supposed to work let
me know, I am just learning the subtleties of this protocol.


>
> Sometimes this can happen due to programming errors.  Sometimes,
> due to unique runtime behavior of VxWorks networking software,
> this can also be caused by various other factors.  For example,
> your reader task may not get to run due to low task priority,
> while the netTask keeps queueing the input data in the socket
> queue.  Some systems can exhibit stop-and-go type of behavior,
> no matter how big the window size is.
>
> Usually, programming TCP/IP application on VxWorks requires fine
> tuning, especially if you have strict requirements for
> throughput and latency.

Yeah, I've heard this alot. Do you know of a comprehensive list of what
can be tweeked -- or at least an attempt at a collection? There appear
to be several global variables under vxworks and a number of the
statndard socket and ip options. Is there a good reference that examines
the ramifications of the various 'tweeks'?

>
>
> Good luck.
>
> -----------------------------------------------------------
>
> Got questions?  Get answers over the phone at Keen.com.
> Up to 100 minutes free!
> http://www.keen.com


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: TCP retransmittions and lost packets
Date: Mon, 14 Aug 2000 10:34:22 -0500
From: Chris Smith 
Organization: Soutwest Research Institute
Message-ID: <3998117E.E6274FB9@here.org>
References: <200008141031.LAA15514@dsl-2.tadpole.co.uk>
Reply-To: cbsmithATswriDoTorg



David Laight wrote:

> Chris wrote:
>
> > I have used this same code on the same hardware under vxworks with a slightly
> > different networking environment (a different hub) and not had any problems.
> > Can anyone explain why this would suddenly cause lost packets? My current
> > "network" is two Synergy ppc sbc's and an NT laptop ethernet sniffer connected
> > to a hub. The hub's uplink is connected to a switch and the switch is on our
> > general network.
> > Nothing on the general network is interacting with the hub.
>
> Sounds to me like there is a 'fault' in your network.  These can be hard to
> track down!
>
> I would simplify things further:
> 1) disconnect from the general network
> 2) try different cables (just in case one is duff)

>
> 3) use a different hub, make sure it is a simple one - they have less scope for
> introducing errors.

I actually have a hand full of cables and processing boards and hubs that I have
tried. I would have expected some change but they all behaved the same.

> 4) make ABSOLUTELY sure that nothing anywhere on the network your side of a
> store and forward switch (these usually route as well) has full duplex enabled.

I have read this before and tried it when we were having packet dropping problems
with a cisco 3500 -- locking the switches ports to half duplex did not help. We
actually had to put a hub between the vxworks boards and the switch. Is there a way
to lock this down under vxworks on an autonegotiating port (short of digging into a
driver)? Why does this seem to cause such a problem?


>
>
> > What ramifications are there to changing the send and receive buffers? I know
> > they change the maximum allowable window to be used for the tcp connection,
> > but what other side effects does this have? Would external network components
> > (the hub or switch) affect what the window should be? Is there an RFC that
> > addresses this?
>
> My guess is that enabling a large transmit window, together with the large
> receive window changes the data pattern on the LAN sufficiently to cause
> transmit bacxkoff (and/or collisions).  These are completely benign on a
> properly built network, but quickly show up any problems that do exist.
>
> I don't know the ack strategy of the vxWorks TCP, but it is quite likely that is
> will ack every other received frame (this is typical of older implementations).

I do not think that is the implementation here. I have TCP_NODELAY enabled -- which
changes things of course -- and it sends an ack as soon as it gets a packet (almost
anyway).

>
> With a window size of 60k and fast systems the network is likely to show the
> following traffic pattern:
>
>         -> 1500 byte data frame
>         -> 1500 byte data frame
>                 (target TCP sends an ack, but the LAN is busy)
>         -> 1500 byte data frame
>         -> 1500 byte data frame
>                 (another ack is generated, queued to be transmitted behind
>                  the first one)

What do you mean by queued? Put in an mbuf or some other low level buffer?

>
>         -> 36 more data frames
>                 (generating 18 more acks, all get queued)
>         (slight pause until target samples an idle LAN)
>         <- 20 back to back ethernet frames each acking 2 more data packets
>         (slight delay while all 20 acks are processes)
>
> The sequence then repeats from the top.
>
> Now if the collision detect and/or backoff algorithm isn't working (because of a
> broken cable or duff hub/switch) then frames will get lost.
>
> Also an ethernet driver with inadequate transmit buffering will also drop
> packets.

Any suggestions on how do I check this?

>
>
> If you unplug your uplink you should be able to see which packets are missing.
>
>         David
>
> (The above sequence is REAL - I saw it a few years ago using relatively slow (by
> current standards) hardware.  It can be avoided by putting two network cards in
> each system and cabling two equivatent LANs and getting different systems to
> route packets to different lans!)

I am not seeing the same pattern but it seems to be very similar to the one you
describe. This has been helpfull, thanks for replying -- same to all the others who
have put in their 2 cents so far.


>
>
> ----------------------------------------------------------------
> David Laight                          email: dsl@tadpole.co.uk
> Tadpole Technology plc                phone: +44 1223 428 232
> Cambridge, UK                         fax:   +44 1223 428 201


---------------------------

Newsgroups: comp.os.vxworks
Subject: DosFs 2.0 in vxWorks (ATA disks)..
Date: Mon, 14 Aug 2000 19:26:07 GMT
From: venkat_gl@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8n9h47$dkg$1@nnrp1.deja.com>

Hello

I am working with an ATA disk which has to respond to dosFs (am using
dosFs 2.0) and Windows. (ie. there must be two partitions one
responding to dosFs and another in which Windows can perform
reads/writes.). My target board is a MIPS board to which the disk is
connected.

Situation:
The disk is partitioned into two partitions and formatted using
fdisk.exe in windows.
When I now connect this disk to my target board and enable dosFs, the
following happpens:

When I type "devs" in shell prompt, I can see both the partitions
installed.
- -> devs
drv name
4   /ata0a
4   /ata1a
value = 0 = 0x0

- -> @cd "/ata0a"
value = 0 = 0x0

- ->@ls "/ata0a"
Can't open "/ata0a"
value = -1 = 0xffffffff

The following output spews on the hyper terminal:

Malformed boot sector. Offset 0 Value 62
Try to accept other boot block copy...
Malformed boot sector. Offset 0 Value 255
Malformed boot sector. Offset 0 Value 0
....
....
.... (the above repeats)

These log messages seem to be coming from somewhere in dosFsLib.c.
Could somebody guide me as to what is happening??? I hope so it is not
something to do with endianness problem...

Thanks for any help in advance,,

Venkat..


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: VxWorks contractors needed in Chicago
Date: Mon, 14 Aug 2000 14:33:28 -0400
From: "Colleen Nielsen" 
Organization: EnterAct Corp.
Message-ID: <8n9i25$1o9q$1@news.enteract.com>

If anyone would be interested in doing a long term contract in the Chicago
area, please e-mail me for more details.  Strong Embedded skills with
VxWorks experience is required.

Thanks
Colleen Nielsen
Senior Regional Recruiter
Technisource
(800)330-3308
www.Technisource.com




---------------------------

Newsgroups: comp.os.vxworks
Subject: signal problem
Date: Mon, 14 Aug 2000 18:46:46 +0200
From: aalimari 
Organization: PMSN
Message-ID: <39982275.E312031@best.ms.philips.com>


- --------------30EDBC1DC227D5A868CB1266
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Hello

Being a new to VxWorks I am having the following problem.
The attached program crashes with general protection fault. I run vxsig
in the SIMNT on intel x86 (Tornado 2.0). The crash happens when I send a
signal to another task. When signal is sent to self the problem does not
occur.

Any help is appreciated
Thanks in advance,
Artem Alimarine


The program
- -------------------------------------------------
#include 
#include 
#include 
#include 
#include 

static bool raised = false;

void handler(int)
{
 raised = true;
 printf("HANDLER\n");
}


void task(int)
{
 signal(SIGUSR1, handler);

 //sigset_t sigs = 0xFFFFFFFF;
 //int s = sigtimedwait(&sigs, NULL, NULL);

 //printf("awaited %d\n", s);
 //raise(SIGUSR1);
 //kill(taskIdSelf(), SIGUSR1);

 while(!raised) taskDelay(1);

 printf("raised\n");
}


void vxsig()
{
    int tid;


    tid = taskSpawn(
     "test",
     0,
     VX_FP_TASK | VX_STDIO,
     32 * 1024,
     (FUNCPTR)task,
     0,0,0,0,0, 0,0,0,0,0);

    printf("created\n");
    taskShow(taskIdSelf(), 2);

    taskDelay(50);

    printf("cancel\n");
    kill(tid, SIGUSR1);
    printf("canceled\n");
    taskShow(taskIdSelf(), 2);

    while(!raised) taskDelay(1);
    printf("OK\n");
}

The simulator window displays:
- --------------------------------
Exception !
 Vector 13 : General Protection Fault
  Program Counter : 0x0041d8e1
 Status Register : 0x00010202

The shell window:
- --------------------
- -> vxsig
created

  NAME        ENTRY       TID    PRI   STATUS      PC       SP
ERRNO  DELAY
- ---------- ------------ -------- --- ---------- -------- --------
- ------- -----
tExcTask   41ecf8        4b88e58   0 PEND         430906  4b88d74
d0003     0
tLogTask   412814        4b84440   0 PEND         430906  4b84358
d0003     0
test       49aeec4       4b753f8   0 SUSPEND      41d8e1  4b75226
d0003     0
test       49aeec4       4b63ea8   0 DELAY        430906  4b63e10
d0003     1
tWdbTask   40b960        4b80c20   3 PEND+T       430906  4b80b04
d0003   180
t1         414c58        4b7c6a0   4 DELAY        430906  4b7c5d0
d0003     1
t2         414c58        4b6b150   4 READY        4166e8  4b6a244
d0003     0
cancel
canceled

  NAME        ENTRY       TID    PRI   STATUS      PC       SP
ERRNO  DELAY
- ---------- ------------ -------- --- ---------- -------- --------
- ------- -----
tExcTask   41ecf8        4b88e58   0 PEND         430906  4b88d74
d0003     0
tLogTask   412814        4b84440   0 PEND         430906  4b84358
d0003     0
test       49aeec4       4b753f8   0 SUSPEND      41d8e1  4b75226
d0003     0
test       49aeec4       4b63ea8   0 SUSPEND      41d8e1  4b63cd6
d0003     0
tWdbTask   40b9
Exception number 0: Task: 0x4b63ea8 (test)

General Protection Fault
Program Counter:            0x0041d8e1
Status Register:            0x00010202

60        4b80c20   3 PEND+T       430906  4b80b04   d0003   180
t1         414c58        4b7c6a0   4 DELAY        430906  4b7c5d0
d0003     1
t2         414c58        4b6b150   4 READY        4166e8  4b6a25c
d0003     0



- --------------30EDBC1DC227D5A868CB1266
Content-Type: text/html; charset=us-ascii
Content-Transfer-Encoding: 7bit



Hello

Being a new to VxWorks I am having the following problem.

The attached program crashes with general protection fault. I run vxsig
in the SIMNT on intel x86 (Tornado 2.0). The crash happens when I send
a signal to another task. When signal is sent to self the problem does
not occur.

Any help is appreciated

Thanks in advance,

Artem Alimarine

 

The program

-------------------------------------------------

#include <vxWorks.h>

#include <taskLib.h>

#include <sigLib.h>

#include <signal.h>

#include <stdio.h>

static bool raised = false;

void handler(int)

{

 raised = true;

 printf("HANDLER\n");

}

 

void task(int)

{

 signal(SIGUSR1, handler);

 //sigset_t sigs = 0xFFFFFFFF;

 //int s = sigtimedwait(&sigs, NULL, NULL);

 //printf("awaited %d\n", s);

 //raise(SIGUSR1);

 //kill(taskIdSelf(), SIGUSR1);

 while(!raised) taskDelay(1);

 printf("raised\n");

}

 

void vxsig()

{

    int tid;

 

    tid = taskSpawn(

     "test",

     0,

     VX_FP_TASK | VX_STDIO,

     32 * 1024,

     (FUNCPTR)task,

     0,0,0,0,0, 0,0,0,0,0);

    printf("created\n");

    taskShow(taskIdSelf(), 2);

 

    taskDelay(50);

 

    printf("cancel\n");

    kill(tid, SIGUSR1);

    printf("canceled\n");

    taskShow(taskIdSelf(), 2);

    while(!raised) taskDelay(1);

    printf("OK\n");

}

The simulator window displays:

--------------------------------

Exception !

 Vector 13 : General Protection Fault

  Program Counter : 0x0041d8e1

 Status Register : 0x00010202

The shell window:

--------------------

-> vxsig

created

  NAME        ENTRY      
TID    PRI   STATUS     
PC       SP     ERRNO 
DELAY

---------- ------------ -------- --- ---------- -------- --------
- ------- -----

tExcTask   41ecf8       
4b88e58   0 PEND        
430906  4b88d74   d0003     0

tLogTask   412814       
4b84440   0 PEND        
430906  4b84358   d0003     0

test       49aeec4      
4b753f8   0 SUSPEND      41d8e1 
4b75226   d0003     0

test       49aeec4      
4b63ea8   0 DELAY        430906 
4b63e10   d0003     1

tWdbTask   40b960       
4b80c20   3 PEND+T       430906 
4b80b04   d0003   180

t1         414c58       
4b7c6a0   4 DELAY        430906 
4b7c5d0   d0003     1

t2         414c58       
4b6b150   4 READY        4166e8 
4b6a244   d0003     0

cancel

canceled

  NAME        ENTRY      
TID    PRI   STATUS     
PC       SP     ERRNO 
DELAY

---------- ------------ -------- --- ---------- -------- --------
- ------- -----

tExcTask   41ecf8       
4b88e58   0 PEND        
430906  4b88d74   d0003     0

tLogTask   412814       
4b84440   0 PEND        
430906  4b84358   d0003     0

test       49aeec4      
4b753f8   0 SUSPEND      41d8e1 
4b75226   d0003     0

test       49aeec4      
4b63ea8   0 SUSPEND      41d8e1 
4b63cd6   d0003     0

tWdbTask   40b9

Exception number 0: Task: 0x4b63ea8 (test)

General Protection Fault

Program Counter:           
0x0041d8e1

Status Register:           
0x00010202

60        4b80c20  
3 PEND+T       430906  4b80b04  
d0003   180

t1         414c58       
4b7c6a0   4 DELAY        430906 
4b7c5d0   d0003     1

t2         414c58       
4b6b150   4 READY        4166e8 
4b6a25c   d0003     0

 

 

- --------------30EDBC1DC227D5A868CB1266--


---------------------------

Newsgroups: comp.os.vxworks
Subject: motFecPollSend
Date: Mon, 14 Aug 2000 17:30:13 -0400
From: Sunil Joseph 
Organization: Lucent Technologies, Columbus, Ohio
Message-ID: <399864E5.9FD1F083@lucent.com>

Folks,
	I am testing a Fast ethernet controller on a PPC855T(same as ppc860T). I am
trying to use the motFecPollSend to send a packet. Has anybody done this before?
Please let me know how you could use it to send a packet. 
Thanks

---------------------------

Newsgroups: comp.os.vxworks
Subject: How to porting TERMIOS or SGTTY?
Date: Mon, 14 Aug 2000 18:39:18 GMT
From: aw99_1999@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8n9eck$b7j$1@nnrp1.deja.com>

I'm working on a project to port a UNIX application to vxWorks. It
requires TERMIOS or SGTTY library and vxWorks doesn't have either.
There is only tyLib in vxWorks. Does anyone know how to deal with this
problem? Any help is appreciated!

- ----- Steve Cheng


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Debugging variables defined in nested blocks
Date: Mon, 14 Aug 2000 20:42:55 GMT
From: Edward Bishop 
Organization: Deja.com - Before you buy.
Message-ID: <8n9lke$h7c$1@nnrp1.deja.com>
References: <8msn75$em6$1@nnrp1.deja.com> <8mtdpu$oj$1@pollux.ip-plus.net>

- -O0 worked.  Many thanks to all who replied.


In article <8mtdpu$oj$1@pollux.ip-plus.net>,
  "Serge Wenger"  wrote:
> Try to remove optimisation with -O0
>
> --
> /Serge Wenger
>
> PS: in my e-mail replace nobody by matisa
>
> "Edward Bishop"  a écrit dans le message
news:
> 8msn75$em6$1@nnrp1.deja.com...
> > I would like to be able to be able to view the contents of the
> > variable 'status' in code like this:
> >
> > int bar(void)
> > {
> >     int blub;
> >
> >     while(1)
> >     {
> >         int status = foo();
> >         // break here and check status
> >         ...
> >     }
> > }
> >
> > but "p status" yields 'No symbol "status" in current context'.
There
> > is no problem viewing variables such as blub.  I am using
> > Tornado2 on a MIPS 5400 processor.  Are there any debugger or
> > compiler options (besides -g of course) that I should be using?
> > Thanks in advance for comments or advice.
> >
> > --
> > edward.bishop@am.sony.!spam.com
> >
> >
> > Sent via Deja.com http://www.deja.com/
> > Before you buy.
>
>

- --
edward.bishop@am.sony.!spam.com


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: mv2700 duplicate int! why?
Date: Mon, 14 Aug 2000 17:13:49 -0400
From: Al Johnston 
Organization: MindSpring Enterprises
Message-ID: <3998610D.62030A38@mindspring.com>
References: <399188B1.C9AA0763@mindspring.com> <39919BB3.8D0848AB@aps.anl.gov> <3991A02E.868A848C@mindspring.com>


> > I think your "ignoring the compatibility rational"  is the key point - if
> > you look at the BSP, from memory I think it either initializes one chip or
> > the other depending on a compile-time switch somewhere.  I assume that the

Finally worked my way through enough of this to figure it out.  the
default
mode (include_mpic) uses the MPIC (raven) and the PIB
(winbond/sl82565/w83c553f).
but the drivers for BOTH are in the ravenMpic.c;  the sl82565IntCtrl.c
driver is
NOT used to setup/control the sl82565!!!!!  the online version of the 
target.nr file tells us this by saying that the "sl82565IntCtrl.c OR
ravenMpic.c" driver is used.  the version in the bsp itself leaves out
the "or"!... that and the identical naming of most of the routines made
this one very hard to figure out.

Interestingly, the connections of the ISA interrupts to the 82565 which
the
doc says is done for compatibility, is the set of the irq that are used.
the identical set of isa irq's that go to the raven mpic are never
"un masked"... so, by default we are using the connections that are
only there for compatibility with non-ravan cards...  

this sure confused the !@# out of an already confusing issue (how many
different things on the card can we name either "level" or "vector"?!!!
- --- g ---)

Man I am glad we only change cards every 5-8 years!!!

- -al

---------------------------

Newsgroups: comp.os.vxworks
Subject: Tornado project facility
Date: Mon, 14 Aug 2000 13:38:18 +0200
From: "Ramakrishnan,CN" 
Organization: UUNET-NL (http://www.nl.uu.net)
Message-ID: 
References: <8n1jbh$ht5$1@overload.lbl.gov> <8n1n0p$iee$1@news1.xs4all.nl> <8n6d31$6nk$1@nnrp1.deja.com>

Hello,
I agree that the configuration in Tornado II which is now spread into 3
places, namely project, config.h and configAll.h should be done away
with. Though i won't dare it to call it a bug, it causes lot problems.
Also in the project facility, you have to be very careful while enabling
a feature. When you enable TELNET server, PPP_BOOT is also enabled which
may be undesirable in many situation( like one i had ). Enabling the
dependencies has been a good improvement compared to Tornado I,but i
think there is scope for improvement. 
Ramakrishnan.C.N
Peek Traffic B.V.
Basicweg 16,
3821 BR AMERSFOORT 
P.O. Box 2542,
3800 GB AMERSFOORT
The Netherlands 
telephone:  +31 33 454 1777
fax: 	     +31 33 454 1850
Internet: www.peektraffic.nl 
	
e-mail: C.N.Ramakrishnan@peektraffic.nl


> ----------
> From: 	John Fusco[SMTP:john_fusco@my-deja.com]
> Posted At: 	zondag 13 augustus 2000 16:58
> Posted To: 	vxworks
> Conversation: 	Tornado project facility
> Subject: 	Re: Tornado project facility
> 
> I would call that a bug.
> 
> Since the you can't build a bootrom in a project, you have to modify
> config.h if you need to modify your bootrom (e.g. to make your kernel
> bigger).  There is no sanity checking in Tornado to verify that the
> changes you made to the bootrom are still compatible with your
> project.  There is no tool in Tornado to tell you if your project is
> still compatible with your bootrom.
> 
> You can create a new project based on the new bootrom to ensure that
> they will be compatible, but there is no tool in Tornado that will
> simply copy all the components from the old project.  You have to do
> that by hand!  In essense you have to start all over.
> 
> Yes, I know the other alternative is not to use a bootrom.
> 
> Sorry for the rant, but this has been a thorn in my side for some
> time.  Sloppy configuration management is such a big part of Tornado,
> it's almost a feature.  There has been no thought put into
> configuration management in Tornado.
> 
> 
> > When you want to make changes to a project you have to do that in
> the
> > project configuration and *NOT* in the config.h. Almost all the
> defines
> > you make in the config.h are redefined in the project files.
> > When you create a new project it is based on the config.h in the BSP
> > directory. But after that it is seperate from the config.h. So the
> > config.h only provides the defaults for a project.
> >
> > Groeten,
> > 	Johan
> 
> 
> Sent via Deja.com http://www.deja.com/
> Before you buy.
> 


---------------------------

Newsgroups: comp.os.vxworks
Subject: WindView file format
Date: Mon, 14 Aug 2000 14:56:14 +0200
From: "Dr. Peter Salewsky" 
Message-ID: <3997EC6E.8BDC4372@resolut.de>

Hi all,

I'm performing time measurements of an application
using WindView. Unfortunately, the WindView GUI allows
statistical analyses only for task durations but not
for time intervals between user events.

Therefore I'd like to convert the WindView logfile
into e.g. ASCII to postprocess it with other tools.

1. Is a filter to convert WindView logfiles into ASCII
   available?

2. If not, could someone point me to a description of
   the WindView logfile format so that I can write such
   a filter myself?

Thanks,

Peter

- -- 
Dr. Peter Salewsky                  phone +49 3493 743-20
resolut GmbH                        fax   +49 3493 743-25
Antonienstr. 3                      dialog@resolut.de
D-06749 Bitterfeld, Germany         www.resolut.de

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Implement password masking on user input ?
Date: 15 Aug 2000 08:32:58 GMT
From: johan@borksoft.xs4all.nl
Organization: Borksoft
Message-ID: <8nav7q$hvf$2@news1.xs4all.nl>
References: 

Jeremy Pye  wrote:
: Has anyone got any ideas on how to program the passwd
: entry so that the user input is not echoed to the terminal
: screen.

: i.e passwd: *******

Did you look at loginLib?

Otherwise you can disable echoing using an ioctl call.

Groeten,
	Johan

- -- 
     o o o o o o o . . .   ___________________________________
    o      _____           ||    Johan Borkhuis              |
  .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
 >(________|__|_[_________]_|________________________________|
 _/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: Target Server Connect Problem?
Date: 15 Aug 2000 08:35:30 GMT
From: johan@borksoft.xs4all.nl
Organization: Borksoft
Message-ID: <8navci$hvf$3@news1.xs4all.nl>
References: <8oTl5.128$Vi5.400@client>

Mike Kordik  wrote:
: WTX Error 0x100d4 (AGENT_NO_AGENT_PROC)
: value = -1 = 0xffffffff

What does the targetserver itself report? If you look in the Log-console
of the Target-server you will probably see some kind of error.

Groeten,
	Johan

- -- 
     o o o o o o o . . .   ___________________________________
    o      _____           ||    Johan Borkhuis              |
  .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
 >(________|__|_[_________]_|________________________________|
 _/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 in vxWorks (ATA disks)..
Date: 15 Aug 2000 08:30:38 GMT
From: johan@borksoft.xs4all.nl
Organization: Borksoft
Message-ID: <8nav3e$hvf$1@news1.xs4all.nl>
References: <8n9h47$dkg$1@nnrp1.deja.com>

venkat_gl@my-deja.com wrote:
: Situation:
: The disk is partitioned into two partitions and formatted using
: fdisk.exe in windows.
: When I now connect this disk to my target board and enable dosFs, the
: following happpens:

Looks like the disk is not formatted. Can you access both partitions from
Windows?

Groeten,
	Johan

- -- 
     o o o o o o o . . .   ___________________________________
    o      _____           ||    Johan Borkhuis              |
  .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
 >(________|__|_[_________]_|________________________________|
 _/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: [Q] Fatal WTX (0x10129) Synchronization Error
Date: Mon, 14 Aug 2000 23:24:37 -0700
From: "T. Wright" 
Organization: ExecPC Internet - Milwaukee, WI
Message-ID: <3998c5fa$0$60930$392904a7@news.execpc.com>

Has anybody encountered an error like this one?  Better yet, if so, what did
you do to resolve this.

From Tornado II, I am attaching a target server to a Motorola board
(PPC604).  When I attach to the
target server, I always get this error. The result is that the symbol table
on the Motorola board is
will not be loaded when I download my .out file.

Can anybody inform me as to what is going on and how I can possibly solve
this problem.  Greatly appreciated!

Signed,
"Documenting until resolved" - Tom




---------------------------

End of New-News digest
**********************


From vxwexplo-errs@csg.lbl.gov  Tue Aug 15 08:24:40 2000
From: "Crozier, Bruce" 
Date: Tue Aug 15 08:24:43 PDT 2000
Subject: problems with vxWorks bspVal on Intel architecture 


Dear vxWorks Gurus,

I am attempting to run the WRS standard BSP validation suite (bspVal) on 
an Intel architecture board, and have a couple of questions I hope someone 
will be able to help me with.  I am running Tornado 1.0.1, from cdrom ZC-08
(with Cumulative Patch 1),  on a Solaris host.

One of the 'bootline' tests uses the 'm' command to write the value 0x55aa
into memory at address 0x4864, and then uses 'd 0x4864,1' to print it out,
and checks the result looking for 0x55aa.  The code for both the 'm' and 'd'
commands are WRS written, located in
/target/config/all/bootConfig.c.

On my board, these commands return the following:

   [VxWorks Boot]: m 0x4864
   004864:  0000-a55aa
   004866:  0000-.
   [vxWorks Boot]: d 0x4864,1
   004860:            aa55

I'm not very familiar with little-endian architecture boards, but since 'm' 
works with 16 bit (short) values, and 'd' prints out one byte at a time, 
unless someone does a byteswap somewhere, it appears that this is what you
get, 
and the bspVal suite test fails.

Question: I don't see how the BSP is involved in this behaviour.  Any
suggestions
about how to get this bspVal test to pass ?  Is this an indication of some
error in
the BSP ?

---------------

My other problem with bspVal is in the 'error2' test that forces a system
reboot by causing a busError while an interrupt handler is running.  My
board
gets its bootrom code from an IDE flash disk.  Under error2 test conditions,
it is trying to execute sysToMonitor() at interrupt level, and the
dosFsDevInit()
that is called by usrAtaConfig() fails under these circumstances with error
code
0x430001: NOT_ISR_CALLABLE.

Question: Is there a simple way to get my bootrom code out of the Ata disk
without
calling usrAtaConfig() ?  Or, is there something I could/should be doing in 
sysToMonitor() that would tell the OS to ignore the not_isr_callable test in

this case - since I'm going to re-boot anyway.

---------------

Any help appreciated.

Regards,
Bruce Crozier


From vxwexplo-errs@csg.lbl.gov  Tue Aug 15 11:14:20 2000
From: David Laight 
Date: Tue Aug 15 11:14:22 PDT 2000
Subject: Re: TCP retransmittions and lost packets -- New Question

Chris Smith said:

> Do you mean the RWIN field (in each netwrok packet) will actually change
> as a receiver gets more data backed up? I initially thought that was how
> the window 'mechanism' is supposed to work, but it does not actually
> appear to change once a connection is established -- as seen from a
> packet sniffer

That is what you SHOULD see, however you may need a test program (or a slow 
system) to see this in action.

> Actually the windows also do not appear to synchronize between the two
> ends of a connection like I though they would. They each appear to just
> continually send their initial values.

The sequence numbers and window sizes are entirely separate for the two ends of 
a connection (ie the two directions of data flow).

A TCP header looks like:

	Source-port	|  dest- port
	Seq. no first byte in this packet
	seq. no of last data byte accepted
	off|rsvd|flags	|   window
	Checksum	| urgent ptr
	options		| padding
	user data
	
	off is data-offset (4 bits) - I'm not sure what its for!
	rsvd is 6 bits
	flags are	urgent
			ack	ack of syn or fin
			push
			reset	abortive release/connect reject
			syn	connect request
			fin	orderly reslease
	urget ptr is only present if the urgent flag is set
	options are TLV and normally absent
	the user data is always 4-byte aligned

Each data byte (in a direction) has a sequence number.

The data receiver gives two values - the sequence number of the last byte 
accepted (the sender can then discard the relevant data as it won't need to 
resend it) and the window (ie the number of bytes it is willing to accept.

It is valid to send data with a sequence number(s) between the last byte 
accepted and the advertised receive window.

If data is backing up in the receiving system it (ie its protocol stack) would 
normally ack the data but not increase the window.  When the appication finally 
reads out some data another message is sent increasing the window.  Better 
performance is had by delaying the ack slightly hoping the application will 
consume the data.  You might be able to see this by using taskSuspend (or a 
breakpoint) to stop the receiving vxWorks task.

Similarly the system can be optimistic with its window size.  If too much data 
suddenly arrives leading to a buffer crisis it can ignore the received data (as 
if the packet got lost on the lan) - indeed I believe it can even advertise a 
smaller window (stopping the sender repeatedly retrying).  The standard retry 
sequence will let data through when the crisis is over.

The sequence numbers for each direction are randomly selected during connection 
setup.  This makes it slightly more difficult for a malitious agent to insert 
data into a connection.

Any clearer?

	David


----------------------------------------------------------------
David Laight                          email: dsl@tadpole.co.uk
Tadpole Technology plc                phone: +44 1223 428 232
Cambridge, UK                         fax:   +44 1223 428 201



From vxwexplo-errs@csg.lbl.gov  Tue Aug 15 13:19:42 2000
From: Leah Real 
Date: Tue Aug 15 13:19:44 PDT 2000
Subject: SSH for VxWorks

Hello. 

Has anybody ported the SSH (the one that uses RSA) to VxWorks?  Are there
any 3rd party vendor out there who has done this?

I searched the archives on the VxWorks list with a string "SSH" and "secure
shell", but no relevant or current messages were posted.  I found some posts
dated 1996 and 1999, but did not really talk about the SSH I am looking for.


Any help will be much appreciated.  Thanks.
Leah



From vxwexplo-errs@csg.lbl.gov  Wed Aug 16 01:23:20 2000
From: David Laight 
Date: Wed Aug 16 01:23:23 PDT 2000
Subject: Re: TCP retransmittions and lost packet

Chris wrote:

> > 4) make ABSOLUTELY sure that nothing anywhere on the network your side of a
> > store and forward switch (these usually route as well) has full duplex
> > enabled.

> I have read this before and tried it when we were having packet dropping 
> problems with a cisco 3500 -- locking the switches ports to half duplex did
> not help. We actually had to put a hub between the vxworks boards and the
> switch. Is there a way to lock this down under vxworks on an autonegotiating
> port (short of digging into a driver)? Why does this seem to cause such a
> problem?

The BIG problem with full duplex is that the ethernet device driver and 
transceivers at both ends of the link have to agree on using it.

The transceivers use information encoded on the link test pulses to determine 
the 'best' transfer speed.  Two mechanisms are used, auto-detect and 
auto-negotiation.

A 10Mb transceiver (on a network card or in a hub) sends out a single link test 
pulse at a fixed frequency - these light the 'link test' led on the receiving 
tranceiver.

A 100Mb transceiver sends out a short burst of link test pulses at a fixed 
frequency.  A 10M only tranceiver will think this is a valid link test pulse.

A 10/100 autodetecting tranceiver can differentiate between the above data 
streams and will transmit at the fastest speed.

A transceiver that support auto-negotiation encodes its capabilities (10M, 
100T4, 100TX, 10FDX, 100FDX etc) into the link test data stream.  The two 
transceivers then agree on the fastest available transfer rate.  This is likely 
to be full duplex if supported by both transceivers,

Ethernet CSMACD (Carrier Sense Media Access with Collison Detect) works on 
twisted pair by disallowing transmit if there is a receive signal (this works 
much much better than on coax where an average absolute cabe voltage has to be 
detected in the presence of 50/60Hz hum).

If the two ends of a cable disagree over full duplex then two packets will be 
transmitted at the same time and one gets lost.  (maybe a 'late collision' 
error).

If the ethernet driver believes the link is half duplex, but the transceiver 
thinks it is full duplex, then the driver is likely to report 'no carrier' 
during transmit.  (This is the absense of the receive enable caused by its own 
transmittions - as is seen on an AUI interface to a coax segment).

The problem is that the transceiver can flip between full and half duplex at 
almost any time (eg plug into a different remote system).  The device driver has 
to notice this change and change its own behaviour to match.  However the 
interface between the driver and transceiver (the MII interface) doen't make 
this easy.  Indeed within the standard transceiver registers it isn't even 
possible to determine the speed!

I actually suspect that many of the vxworks drivers get this completely wrong!
My driver (for the Xircom pcmcia cards - written under an NDA) does:

	/* Change Auto negotiate advertise register so we only advertise
           supported options (in particular exclude FDX) */
        td_write_mii( eth, MII_ADDR_REG( 0, MII_ANAR ),
                td_read_mii( eth, MII_ADDR_REG( 0, MII_ANAR ) ) & ~0x1f40  );

to ensure the transceiver doesn't enable full duplex.

If you get this far (without suffering brain overload) also:

> What do you mean by queued? Put in an mbuf or some other low level buffer?

Could do anything depending on the card hardware and the amount of queued data.
Typically the hardware will be given the first few packets (or until its FIFO is 
full), the rest have to be queued within the driver (or an appropriate response 
given to the protocol stack so that it will queue the packet).

It is IMPERITIVE that transmit packets are not discarded just because the LAN is 
busy (or the mbuf list (or whatever) has lots of fragments).

> Any suggestions on how do I check this?

>From the network trace it should be possible to see whether the sender or 
receiver are losing packets.

	David


----------------------------------------------------------------
David Laight                          email: dsl@tadpole.co.uk
Tadpole Technology plc                phone: +44 1223 428 232
Cambridge, UK                         fax:   +44 1223 428 201



From vxwexplo-errs@csg.lbl.gov  Wed Aug 16 04:03:08 2000
From: Vxworks Exploder 
Date: Wed Aug 16 04:03:10 PDT 2000
Subject: comp.os.vxworks newsdigest

Comp.Os.Vxworks Daily Digest    Wed Aug 16 04:03:04 PDT 2000

        Subject: Multiple instances of same (modern) NIC in an x86 target?
        Subject: ioctl and pipes
        Subject: Re: Target Server Connect Problem?
        Subject: Re: How to adust telnet priority
        Subject: Re: How to adust telnet priority
        Subject: How to adust telnet priority
        Subject: Re: Target Server Connect Problem?
        Subject: problems with vxWorks bspVal on Intel architecture
        Subject: memory mapping
        Subject: Re: Target Server Connect Problem?
        Subject: re: TCP retransmittions and lost packets -- New Question
        Subject: Real time clock
        Subject: Re: Ethernet questions
        Subject: Re: [Q] Fatal WTX (0x10129) Synchronization Error
        Subject: Adventnet SNMP agent for vxworks
        Subject: SSH for VxWorks
        Subject: cannot build bootrom_res.hex
        Subject: BSP for Motorola CPV5350
        Subject: Re: Target Server Connect Problem?
        Subject: BSD4.4 Required Driver Entry point and structures.
        Subject: Re: mv2700: where is "shared" memory in mpu space?
        Subject: Re: How to adust telnet priority
        Subject: Re: mv2700: where is "shared" memory in mpu space?
        Subject: Re: Display of Windows NT Windows on a VXWorks Power PC
        Subject: mv2700: where is "shared" memory in mpu space?
        Subject: Re: Real time clock
        Subject: $$$$$$  Free Music!!!  $$$$$$$$$!
        Subject: BitBand Hiring (ISRAEL)
        Subject: re: TCP retransmittions and lost packet

-------------------------------------------------------

Newsgroups: comp.os.vxworks
Subject: Multiple instances of same (modern) NIC in an x86 target?
Date: Tue, 15 Aug 2000 12:57:19 +0200 (CEST)
From: Henrik Bergstrom 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <8nb8ng$ghd$1@overload.lbl.gov>

Dear VxWorks-users,

I am having trouble finding a modern, commercially available, NIC which
works with several instances of the same card in an x86 target. Some
drivers are prepared for multiple instances, but don't work in
practice.

I am in need of a END (MUX) driver and the only one my local
vxw-representative knew of was Intel's 82559, but I don't have access
to that NIC.

What would make me really happy are END/MUX-drivers with support for
multiple instances of the same card for any of the following:

  o 3com 3c905C-TX "Fast EtherLink XL PCI"
    (This has "theoretical support" for multiple instances, but not
    fully so in practice.)
  o SMC 9432TX "EtherPower II 10/100"
  o any other modern, commercially available, card

Any help, tips or recommendations is appreciated.


Best regards,

Henrik Bergstrom

- - --
Henrik Bergstrom  +46-8-628 28 28  henrik.bergstrom@intertex.se
Intertex Data AB  (fax) 628 64 14



---------------------------

Newsgroups: comp.os.vxworks
Subject: ioctl and pipes
Date: Tue, 15 Aug 2000 13:45:45 +0200
From: Gunter Ritter 
Organization: Tenovis GmbH u. Co. KG
Message-ID: <8nbabm$aen$1@proxy2.fe.internet.bosch.de>

Hello!

I think I have some problems using ioctl and pipes.

After writing some messages into different pipes, I try to find out how many
messages are in one of these pipes. I use ioctl and FIONSMGS. But instead of
the number of messages in the pipe I always find 0 (zero) in the 3rd
argument of ioctl.

What´s going wrong. Is there anybody who has experience with ioctl and
pipes.

Thanks.

Gunter



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Target Server Connect Problem?
Date: Tue, 15 Aug 2000 08:22:02 -0400
From: "Mike Kordik" 
Message-ID: 
References: <8oTl5.128$Vi5.400@client> <8navci$hvf$3@news1.xs4all.nl>

Johan,
   When you say log-console, are you referring to the window you see when
you select Tool->Target Server->Manage...? If so, there are no errors. This
is what it says:

Name   : vxTarget@art_vandelay
Version: 2.0
Status : unreserved
Runtime: VxWorks 5.4
Agent  : 2.0
CPU    : PENTIUM
BSP    : PC PENTIUM
Memory : 0x1fff000
Link   : WDB Agent RPC
User   : kordikmj
Start  : Tue Aug 15 08:18:08 2000
Last   : Tue Aug 15 08:18:31 2000
Attached Tools :
 symTblsSync (target@vxTarget)
 windsh (kordikmj@art_vandelay)

The target server seems to run without problems. It is when I run the shell
or the browser I get errors.

Thanks,
Mike

 wrote in message
news:8navci$hvf$3@news1.xs4all.nl...
> Mike Kordik  wrote:
> : WTX Error 0x100d4 (AGENT_NO_AGENT_PROC)
> : value = -1 = 0xffffffff
>
> What does the targetserver itself report? If you look in the Log-console
> of the Target-server you will probably see some kind of error.
>
> Groeten,
> Johan
>
> --
>      o o o o o o o . . .   ___________________________________
>     o      _____           ||    Johan Borkhuis              |
>   .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
>  >(________|__|_[_________]_|________________________________|
>  _/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 adust telnet priority
Date: 15 Aug 2000 13:36:38 GMT
From: johan@borksoft.xs4all.nl
Organization: Borksoft
Message-ID: <8nbh16$2e5$1@news1.xs4all.nl>
References: <3999017A.123B1758@dit.siemens.de>

Uwe Foss  wrote:
: The standard telnet priority(2) doesn't fit to my application.
: How is it possible to adjust the priorities of telnetInTask and
: telnetOutTask ?

taskPrioritySet

Groeten,
	Johan

- -- 
     o o o o o o o . . .   ___________________________________
    o      _____           ||    Johan Borkhuis              |
  .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
 >(________|__|_[_________]_|________________________________|
 _/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 adust telnet priority
Date: Tue, 15 Aug 2000 15:40:08 +0200
From: Uwe Foss 
Organization: Siemens AG
Message-ID: <39994838.1F4FE042@dit.siemens.de>
References: <3999017A.123B1758@dit.siemens.de> <8nbh16$2e5$1@news1.xs4all.nl>

But the problem is that the telnetXXTasks are spawned by the deamon
dynamically.
The application doesn't know when a user is contacting the system via
telnet.
Uwe

johan@borksoft.xs4all.nl schrieb:

> Uwe Foss  wrote:
> : The standard telnet priority(2) doesn't fit to my application.
> : How is it possible to adjust the priorities of telnetInTask and
> : telnetOutTask ?
>
> taskPrioritySet
>
> Groeten,
>         Johan
>
> --
>      o o o o o o o . . .   ___________________________________
>     o      _____           ||    Johan Borkhuis              |
>   .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
>  >(________|__|_[_________]_|________________________________|
>  _/oo OOOOO oo`  ooo   ooo  'o!o!o                      o!o!o`
> === VxWorks FAQ: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html ===

- --
- ----------------------------
SIEMENS AG
ATD TD 23 STG
Uwe Foss
Siemenstr. 33
71254 Ditzigen

Fon:  0711-137-6082
Fax:  0711-137-6071
Mail: uwe.foss@dit.siemens.de
- -----------------------------



---------------------------

Newsgroups: comp.os.vxworks
Subject: How to adust telnet priority
Date: Tue, 15 Aug 2000 10:38:18 +0200
From: Uwe Foss 
Organization: Siemens AG
Message-ID: <3999017A.123B1758@dit.siemens.de>

The standard telnet priority(2) doesn't fit to my application.
How is it possible to adjust the priorities of telnetInTask and
telnetOutTask ?

Uwe Foß



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Target Server Connect Problem?
Date: Tue, 15 Aug 2000 11:11:41 -0400
From: "Mike Kordik" 
Message-ID: 
References: <8oTl5.128$Vi5.400@client> <8navci$hvf$3@news1.xs4all.nl>  <8nbhef$4ro$1@news1.xs4all.nl>

I am getting a bunch of these messages when I start the target server:

tgtsvr (vxTarget@art_vandelay): Tue Aug 15 10:40:58 2000
    Error: rpccore backend client Procedure unavailable

Do you know what this means?

I was getting a cache full erorr but after I increased the cache size that
error went away but I still get the above error.

Thanks,
Mike

 wrote in message
news:8nbhef$4ro$1@news1.xs4all.nl...
> Mike Kordik  wrote:
> : Johan,
> :    When you say log-console, are you referring to the window you see
when
> : you select Tool->Target Server->Manage...? If so, there are no errors.
This
> : is what it says:
>
> No, I mean the control window of the target server itself. If you are
> running Windows there is a Target server icon in lower-right corner, next
> to the clock. If you double-click on this you get a Log-console. I don't
> know how this works on other platforms.
>
>
> Groeten,
> Johan
>
> --
>      o o o o o o o . . .   ___________________________________
>     o      _____           ||    Johan Borkhuis              |
>   .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
>  >(________|__|_[_________]_|________________________________|
>  _/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: problems with vxWorks bspVal on Intel architecture
Date: Tue, 15 Aug 2000 08:16:31 -0700
From: "Crozier, Bruce" 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <8nbqa0$2bm$1@overload.lbl.gov>


Dear vxWorks Gurus,

I am attempting to run the WRS standard BSP validation suite (bspVal) on 
an Intel architecture board, and have a couple of questions I hope someone 
will be able to help me with.  I am running Tornado 1.0.1, from cdrom ZC-08
(with Cumulative Patch 1),  on a Solaris host.

One of the 'bootline' tests uses the 'm' command to write the value 0x55aa
into memory at address 0x4864, and then uses 'd 0x4864,1' to print it out,
and checks the result looking for 0x55aa.  The code for both the 'm' and 'd'
commands are WRS written, located in
/target/config/all/bootConfig.c.

On my board, these commands return the following:

   [VxWorks Boot]: m 0x4864
   004864:  0000-a55aa
   004866:  0000-.
   [vxWorks Boot]: d 0x4864,1
   004860:            aa55

I'm not very familiar with little-endian architecture boards, but since 'm' 
works with 16 bit (short) values, and 'd' prints out one byte at a time, 
unless someone does a byteswap somewhere, it appears that this is what you
get, 
and the bspVal suite test fails.

Question: I don't see how the BSP is involved in this behaviour.  Any
suggestions
about how to get this bspVal test to pass ?  Is this an indication of some
error in
the BSP ?

- ---------------

My other problem with bspVal is in the 'error2' test that forces a system
reboot by causing a busError while an interrupt handler is running.  My
board
gets its bootrom code from an IDE flash disk.  Under error2 test conditions,
it is trying to execute sysToMonitor() at interrupt level, and the
dosFsDevInit()
that is called by usrAtaConfig() fails under these circumstances with error
code
0x430001: NOT_ISR_CALLABLE.

Question: Is there a simple way to get my bootrom code out of the Ata disk
without
calling usrAtaConfig() ?  Or, is there something I could/should be doing in 
sysToMonitor() that would tell the OS to ignore the not_isr_callable test in

this case - since I'm going to re-boot anyway.

- ---------------

Any help appreciated.

Regards,
Bruce Crozier

---------------------------

Newsgroups: comp.os.vxworks
Subject: memory mapping
Date: Tue, 15 Aug 2000 13:57:31 +0100
From: "Kai Mattern" 
Organization: Heidelberger Druckmaschinen AG
Message-ID: <8nbb40$gcl$1@news.hdpp.de>

Hi,

like many others im new at VxWorks and have to write a driver for a printer.
The printer will be controlled over a PCI card. This card has 256MB memory
on board. I was told i have to reprogramm the raven chipset on my mtx603
board to map the pci memory. But i dunno how this works. I read all source i
have...but im more confused then before.

The Board Description tells me that if i use the "PREP Memory Map Model"
the processor Address 8000 0000 to BFFF FFFF is maped to the PCI I/O Space
and C000 0000 to FCFF FFFF is maped to PCI memory space (point of view from
processor).
On the other side there is the point of view of PCI bus. Manual says PCI
memory space is maped from 0200 0000 to 7FFF FFFF.

I took a look in the source code from the raven shipset and found this in
the ravenpic.c file


1. For the standard vxWorks configuration the  MPC to
      PCI mapping registers are initialized to the PReP
      model with some additions:

   MPC Address Range        PCI Address Range          Definition
   -----------------                -----------------               --------
- ---------
   80000000 BF7FFFFF      00000000 3F7FFFFF      ISA/PCI I/O space
   C0000000 FCFFFFFF     00000000 3CFFFFFF     ISA/PCI Mem space w/MPIC
   FD000000 FDFFFFFF    00000000 00FFFFFF      ISA/PCI Mem space
   FE000000 FE7FFFFF     00000000 007FFFFF       ISA/PCI I/O space



i thought great so i only need to change the PCI Address Range to my memory
range and thats it.......but i cant find them in the source.
There seams to be 4 "windows" inside the raven chipset who provide the
memory mapping for the pci bus. The PPC Slave Adress Registers.
I can enter there Start and End adresses for the PPC bus which will be used
to access PCI bus resources. But how works the connect ??

does anybody out there ever did something like this and is able to help me ?
or just know how this memory mapping with the raven chipset works ? Maybe i
solve the problem if i understand how it work.

Thanks a lot

Kai
Kai_Mattern@web.de







---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Target Server Connect Problem?
Date: 15 Aug 2000 13:43:43 GMT
From: johan@borksoft.xs4all.nl
Organization: Borksoft
Message-ID: <8nbhef$4ro$1@news1.xs4all.nl>
References: <8oTl5.128$Vi5.400@client> <8navci$hvf$3@news1.xs4all.nl> 

Mike Kordik  wrote:
: Johan,
:    When you say log-console, are you referring to the window you see when
: you select Tool->Target Server->Manage...? If so, there are no errors. This
: is what it says:

No, I mean the control window of the target server itself. If you are
running Windows there is a Target server icon in lower-right corner, next
to the clock. If you double-click on this you get a Log-console. I don't
know how this works on other platforms.


Groeten,
	Johan

- -- 
     o o o o o o o . . .   ___________________________________
    o      _____           ||    Johan Borkhuis              |
  .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
 >(________|__|_[_________]_|________________________________|
 _/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: TCP retransmittions and lost packets -- New Question
Date: Tue, 15 Aug 2000 19:14:22 +0100 (BST)
From: David Laight 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <200008151814.TAA20497@dsl-2.tadpole.co.uk>

Chris Smith said:

> Do you mean the RWIN field (in each netwrok packet) will actually change
> as a receiver gets more data backed up? I initially thought that was how
> the window 'mechanism' is supposed to work, but it does not actually
> appear to change once a connection is established -- as seen from a
> packet sniffer

That is what you SHOULD see, however you may need a test program (or a slow 
system) to see this in action.

> Actually the windows also do not appear to synchronize between the two
> ends of a connection like I though they would. They each appear to just
> continually send their initial values.

The sequence numbers and window sizes are entirely separate for the two ends of 
a connection (ie the two directions of data flow).

A TCP header looks like:

	Source-port	|  dest- port
	Seq. no first byte in this packet
	seq. no of last data byte accepted
	off|rsvd|flags	|   window
	Checksum	| urgent ptr
	options		| padding
	user data
	
	off is data-offset (4 bits) - I'm not sure what its for!
	rsvd is 6 bits
	flags are	urgent
			ack	ack of syn or fin
			push
			reset	abortive release/connect reject
			syn	connect request
			fin	orderly reslease
	urget ptr is only present if the urgent flag is set
	options are TLV and normally absent
	the user data is always 4-byte aligned

Each data byte (in a direction) has a sequence number.

The data receiver gives two values - the sequence number of the last byte 
accepted (the sender can then discard the relevant data as it won't need to 
resend it) and the window (ie the number of bytes it is willing to accept.

It is valid to send data with a sequence number(s) between the last byte 
accepted and the advertised receive window.

If data is backing up in the receiving system it (ie its protocol stack) would 
normally ack the data but not increase the window.  When the appication finally 
reads out some data another message is sent increasing the window.  Better 
performance is had by delaying the ack slightly hoping the application will 
consume the data.  You might be able to see this by using taskSuspend (or a 
breakpoint) to stop the receiving vxWorks task.

Similarly the system can be optimistic with its window size.  If too much data 
suddenly arrives leading to a buffer crisis it can ignore the received data (as 
if the packet got lost on the lan) - indeed I believe it can even advertise a 
smaller window (stopping the sender repeatedly retrying).  The standard retry 
sequence will let data through when the crisis is over.

The sequence numbers for each direction are randomly selected during connection 
setup.  This makes it slightly more difficult for a malitious agent to insert 
data into a connection.

Any clearer?

	David


- ----------------------------------------------------------------
David Laight                          email: dsl@tadpole.co.uk
Tadpole Technology plc                phone: +44 1223 428 232
Cambridge, UK                         fax:   +44 1223 428 201


---------------------------

Newsgroups: comp.os.vxworks
Subject: Real time clock
Date: Tue, 15 Aug 2000 17:05:55 +0200
From: "Achim Zimmer" 
Organization: Colt Telecom GmbH
Message-ID: <8nbm79$i4m$1@crusher.de.colt.net>

has anybody programmed a clock including day, month and year based on RTC ?

Is there any free source for calculating date and time from a second counter
noting leapyears aso ?

Thanks
Achim



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Ethernet questions
Date: Tue, 15 Aug 2000 11:35:58 -0700
From: Joe Durusau 
Organization: Lockheed Martin Corporation
Message-ID: <39998D8E.612FF07B@delphi.com>
References: <8n91n1$8jn$1@news.netvision.net.il>

	Do you really need to do this with vxworks???  There are host tools
provided
for sniffing that just work, without any effort on your part.  For
nstance,
on Unix-like hosts, snoop usually exists.

Speaking only for myself,

Joe Durusau



Arik Halperin wrote:
> 
> 1. What is the best way to perform Ethernet sniffing?(Using an end device)
> 
> 2. How do you make sure the device driver for the Ethernet card(DEC21140 in
> our case) works in promiscuous mode?
> 
> 3. Is there a way to use a snarf protocol without disabling debugging
> capabilities?
> 
> I managed to perform sniffing by changing the end debug agent's code,but
> some how I feel this is not very elegant. I'll be glad to know if there is a
> better way.
> 
> Regards,
> 
>     Arik Halperin

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: [Q] Fatal WTX (0x10129) Synchronization Error
Date: Tue, 15 Aug 2000 09:19:34 -0700
From: Lori Fraleigh 
Organization: Real-Time Innovations
Message-ID: <39996D96.920729E7@rti.com>
References: <3998c5fa$0$60930$392904a7@news.execpc.com>

Tom -

It is possible that symbol table synchronization is not
working properly due to a a bug in Tornado.

I'm assuming you have a Windows host running Tornado.

Bring up the Tornado II IDE, Select Tools->Options.  Scroll 
over to the right most tab, Tornado Registry.  Select remote 
registry and enter the name of the machine running the registry.  
If you select local registry, it will not work.

Basically, "localhost" gets stored in the Windows registry
for the name of the machine running the Tornado registry.
The WTX tools on the host resolve localhost to the ip address
127.0.0.1 and pass this down to the target.  The target then
tries to communicate with 127.0.0.1 to enable the symbol
table synchronization but ends up talking to itself and
not the Windows host.  Neat "feature".

If your Windows machine has multiple ethernet interfaces and
the target is located on a non-primary interface, this still
won't be enough for you.  The machine name will get resolved
to an ip address that the target doesn't know about.  You will
need to execute a routeNetAdd on the target before connecting
the target server in order to get symbol table synchronization
to work or enter the IP address of the non-primary interface
that the target knows about for WIND_REGISTRY.

Regards,
Lori

"T. Wright" wrote:
> 
> Has anybody encountered an error like this one?  Better yet, if so, what did
> you do to resolve this.
> 
> From Tornado II, I am attaching a target server to a Motorola board
> (PPC604).  When I attach to the
> target server, I always get this error. The result is that the symbol table
> on the Motorola board is
> will not be loaded when I download my .out file.
> 
> Can anybody inform me as to what is going on and how I can possibly solve
> this problem.  Greatly appreciated!
> 
> Signed,
> "Documenting until resolved" - Tom

- -- 
Lori Fraleigh                                Real-Time Innovations
Senior Software Engineer                     ScopeTools Manager
Phone: 408-734-4200 x121  Fax: 408-734-5009  http://www.rti.com
Please send/cc all support related email to support@rti.com

---------------------------

Newsgroups: comp.os.vxworks
Subject: Adventnet SNMP agent for vxworks
Date: Tue, 15 Aug 2000 16:56:09 -0400
From: Bhavesh Patel 
Organization: Posted via Supernews, http://www.supernews.com
Message-ID: <3999AE69.AF526782@yahoo.com>

Hi,

Has anybody used adventnet embedded/C snmp agent for vxworks platform? I
would like to know their experiences.  Adventnet website does not have
example/demo for vxworks platform.

Thanks for any help.

Bhavesh Patel


---------------------------

Newsgroups: comp.os.vxworks
Subject: SSH for VxWorks
Date: Tue, 15 Aug 2000 13:24:10 -0700
From: Leah Real 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <8ncbsg$d0t$1@overload.lbl.gov>

Hello. 

Has anybody ported the SSH (the one that uses RSA) to VxWorks?  Are there
any 3rd party vendor out there who has done this?

I searched the archives on the VxWorks list with a string "SSH" and "secure
shell", but no relevant or current messages were posted.  I found some posts
dated 1996 and 1999, but did not really talk about the SSH I am looking for.


Any help will be much appreciated.  Thanks.
Leah


---------------------------

Newsgroups: comp.os.vxworks
Subject: cannot build bootrom_res.hex
Date: Tue, 15 Aug 2000 16:05:50 -0500
From: Bo Yu 
Organization: Motorola CIG
Message-ID: <3999B0AE.2FBA7E6@email.mot.com>


- --------------C933732A2026E3D3B1AB016F
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

I cannot build rom resident image for PowerPC target:

error message:

ldppc -X -N  -e _romInit \
    -Ttext FFF00000  -Tdata 1000000  -o bootrom_res romInit_res.o
bootInit_res.o \
    version.o bootConfig.o sysALib.o sysLib.o end82xx.o membankc.o
memtest.o TestMenu.o Generator.o Dw_82xx.o System.o ports.o brgs.o
fcc1.o fcc3.o scc1.o SMC1.o Timer1.o dw_mmu.o Services.o board.o
Dw_err.o mem_heap.o mem_buff.o
/usr/vob/sc/xc/psi.orig/src/os/tornado2/target/lib/libPPCEC603gnuvx.a

romInit_res.o(.text+0x1ce): undefined reference to `RAM_DST_ADRS'
romInit_res.o(.text+0x1d2): undefined reference to `RAM_DST_ADRS'
make: *** [bootrom_res] Error 1

I can build bootrom.hex and bootrom_uncmp.hex in the same BSP directory.
Why cannot I
build bootrom_res.hex ?

Thanks.


- --
Bo Yu



- --------------C933732A2026E3D3B1AB016F
Content-Type: text/html; charset=us-ascii
Content-Transfer-Encoding: 7bit



I cannot build rom resident image for PowerPC target:

error message:

ldppc -X -N  -e _romInit \

    -Ttext FFF00000  -Tdata 1000000  -o bootrom_res
romInit_res.o bootInit_res.o \

    version.o bootConfig.o sysALib.o sysLib.o end82xx.o
membankc.o memtest.o TestMenu.o Generator.o Dw_82xx.o System.o ports.o
brgs.o fcc1.o fcc3.o scc1.o SMC1.o Timer1.o dw_mmu.o Services.o board.o
Dw_err.o mem_heap.o mem_buff.o    /usr/vob/sc/xc/psi.orig/src/os/tornado2/target/lib/libPPCEC603gnuvx.a

romInit_res.o(.text+0x1ce): undefined reference to `RAM_DST_ADRS'

romInit_res.o(.text+0x1d2): undefined reference to `RAM_DST_ADRS'

make: *** [bootrom_res] Error 1

I can build bootrom.hex and bootrom_uncmp.hex in the same BSP directory.
Why cannot I

build bootrom_res.hex ?

Thanks.

 

-- 
Bo Yu

 

- --------------C933732A2026E3D3B1AB016F--


---------------------------

Newsgroups: comp.os.vxworks
Subject: BSP for Motorola CPV5350
Date: Tue, 15 Aug 2000 14:26:43 -0700
From: Ray Sprangers 
Organization: BCnet
Message-ID: 

Hi,

Does anyone know if there is a Vxworks BSP for the Motorola CPV5350?

Thanks,
Ray


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Target Server Connect Problem?
Date: 15 Aug 2000 18:35:24 GMT
From: johan@borksoft.xs4all.nl
Organization: Borksoft
Message-ID: <8nc2hc$94q$1@news1.xs4all.nl>
References: <8oTl5.128$Vi5.400@client> <8navci$hvf$3@news1.xs4all.nl>  <8nbhef$4ro$1@news1.xs4all.nl> 

Mike Kordik  wrote:
: I am getting a bunch of these messages when I start the target server:

: tgtsvr (vxTarget@art_vandelay): Tue Aug 15 10:40:58 2000
:     Error: rpccore backend client Procedure unavailable

: Do you know what this means?

This probably means that the target server is not working properly. There
are several possible solutions, but it might be a problem with the license
manager. Is your license manager running? If so is it running on the same
machine as your target server? In this case be sure not to use "localhost"
as address but use the IP address of your machine.
The other thing is to check the configuration of your target server and
your WDB on your target, but you probably did that already.

Groeten,
	Johan

- -- 
     o o o o o o o . . .   ___________________________________
    o      _____           ||    Johan Borkhuis              |
  .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
 >(________|__|_[_________]_|________________________________|
 _/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: BSD4.4 Required Driver Entry point and structures.
Date: Tue, 15 Aug 2000 15:01:30 -0400
From: myhandle@lucent.com
Organization: Lucent Technologies
Message-ID: <3999938A.50D6CDF2@lucent.com>

Hi Group,
I need documentation which specifies the BSD4.4 network interface driver format,
specially the required driver entry points and the details of the various
interface data structures that are involved. If One of U can give me a pointer,
it will be appreciated...
thanx
Jish

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: mv2700: where is "shared" memory in mpu space?
Date: Tue, 15 Aug 2000 19:21:05 -0400
From: Al Johnston 
Organization: MindSpring Enterprises
Message-ID: <3999D061.FDDFB30A@mindspring.com>
References: <3999BA42.A32F485E@mindspring.com>

Playing around a bit more i found that I could see a change made to
cpu-0's address 0x4100 by looking at cpu-1's 0x1000,4100 so the 
shared memory is there after all.

But why can't I see it from cpu-0's 0x1000,4100? I "know" I read
somewhere
that this should be possible on this card... can someone please
shed some light on this issue?

thanks

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: How to adust telnet priority
Date: 15 Aug 2000 18:50:56 GMT
From: johan@borksoft.xs4all.nl
Organization: Borksoft
Message-ID: <8nc3eg$cbj$1@news1.xs4all.nl>
References: <3999017A.123B1758@dit.siemens.de> <8nbh16$2e5$1@news1.xs4all.nl> <39994838.1F4FE042@dit.siemens.de>

Uwe Foss  wrote:
: But the problem is that the telnetXXTasks are spawned by the deamon
: dynamically.
: The application doesn't know when a user is contacting the system via
: telnet.

OK, that was an answer that was to quick.

I checked the file telnetLib.o in the library. There is a symbol in there
called telnetTaskPriority. Check if this is what you need. I cannot see if
it is a function or a variable (strings does not provide that
informmation and I don't have nm available), but my guess is that it is a
variable.

Hope this helps a little bit more.

Groeten,
	Johan

- -- 
     o o o o o o o . . .   ___________________________________
    o      _____           ||    Johan Borkhuis              |
  .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
 >(________|__|_[_________]_|________________________________|
 _/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: mv2700: where is "shared" memory in mpu space?
Date: Tue, 15 Aug 2000 18:55:08 -0500
From: Andrew Johnson 
Organization: APS, Argonne National Laboratory
Message-ID: <3999D85C.B9CBDCA4@aps.anl.gov>
References: <3999BA42.A32F485E@mindspring.com> <3999D061.FDDFB30A@mindspring.com>

Hi Al,

Al Johnston wrote:
> 
> Playing around a bit more i found that I could see a change made to
> cpu-0's address 0x4100 by looking at cpu-1's 0x1000,4100 so the
> shared memory is there after all.
> 
> But why can't I see it from cpu-0's 0x1000,4100? I "know" I read
> somewhere
> that this should be possible on this card... can someone please
> shed some light on this issue?

Are you saying you read somewhere that the Universe chip supports
simultaneous VMEbus Master & Slave accesses?  I very much doubt it,
because that would mean both the CPU and the Universe chip using the local
bus simultaneously.

CPU #0 will only see its own memory at 0x00000000, never through the
Universe VME Master window at 0x10000000.  You need to use the
sysBusToLocalAdrs() and sysLocalToBusAdrs() routines to do address
conversions, but beware that the WRS sysBusToLocalAdrs() is not designed
to handle loopback addresses properly (ie when you give it a bus address
that points within a slave window on the local CPU).  I made the necessary
modifications so it did, thus one can always give mine a reachable bus
address and be sure the converted pointer will work on all CPUs.

- - Andrew
- -- 
Sharks kill 10 people each year.  People kill 60,000,000 sharks each year.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Display of Windows NT Windows on a VXWorks Power PC
Date: Tue, 15 Aug 2000 19:32:58 -0500
From: Edsel Harrell 
Organization: Raytheon Company
Message-ID: <3999E13A.748F6C3B@raytheon.com>
References: <8ms846$380$1@nnrp1.deja.com>

This is a multi-part message in MIME format.
- --------------D068453E9F41434C153668B9
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

I have used a product called "ntrigue" from a unix workstation to access

MS applications on a PC box via X ... 3 or 4 years ago...
It was suprisingly fast... I was using large MS eXcel spreadsheets and
navigation was not a problem.
I don't remember vendor, a search on Yahoo poped up several hits...

lpeck@my-deja.com wrote:

> I have a configuration with a dumb  Power PC
> display server running VXworks that supports X
> Displays. Currently we are able to display
> windows from a Unix window on this server by
> simply setting the display variable to the
> display server. However, we would like to replace
> the Unix machine with a PC running Windows NT. Is
> there an application or a setting which will
> allow me to run my existing Windows NT apps on my
> Windows NT and display the windows for the app on
> the dumb display server?
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.

- --
Edsel A. Harrell, Jr              (voice) (972) 575-6468   (email)
eah@raytheon.com
Sr Principal Systems Eng/honors   (pager) (972) 597-8874
Raytheon Systems Company          (snail) 6600 Chase Oaks Blvd, MS 8454,
Plano, Tx 75023
- ------------------------------------------------------------------------------------

"Proofread Carefully to See if You Any Words Out"

DISCLAIMER: I speak only for myself, not my employer.
   EXCEPTION: If any of my ideas are worth more than $0.02, then they
are
              property of my employer.


- --------------D068453E9F41434C153668B9
Content-Type: text/x-vcard; charset=us-ascii;
 name="eah.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Card for Edsel Harrell
Content-Disposition: attachment;
 filename="eah.vcf"

begin:vcard 
n:A. Harrell, Jr;Edsel
tel;pager:972-597-8874
tel;home:817-430-3903
tel;work:972-575-6468
x-mozilla-html:FALSE
org:Raytheon Systems Company;Advanced Airborne EO
version:2.1
email;internet:eah@raytheon.com
title:Sr Principal Systems Engineer/w Honors
adr;quoted-printable:;;6600 Chase Oaks Blvd, MS 8454 (shipping)=0D=0Aor=0D=0AP.O.Box 660246, MS 8454 (mailing);Plano;TX;75066;
x-mozilla-cpt:;-23744
fn:Edsel
end:vcard

- --------------D068453E9F41434C153668B9--


---------------------------

Newsgroups: comp.os.vxworks
Subject: mv2700: where is "shared" memory in mpu space?
Date: Tue, 15 Aug 2000 17:46:42 -0400
From: Al Johnston 
Organization: MindSpring Enterprises
Message-ID: <3999BA42.A32F485E@mindspring.com>

I am running vxworks 5.4;  vxWorks is being built pretty much in the "as
delivered" configuration.  I have two mv2700 cards in a vme chassis. 
Each
of these cards have 64m of dram.

I have left all of the address space and ram size stuff set to the
configuration that WRS ships.

Having spent many hours researching the memory maps for vxw on this
card (I was waiting for my 2nd cpu to arrive) I thought I knew that
the first cpu's dram should show up on all cpus at 0x1000,0000.  That
does not seem to be the case.

On cpu "0" I see nothing but "FFFF"'s at 0x1000,0000; the m() command
does not result in an error, but a subsequent d() shows that the changes
are not made.

On cpu "1" I see what seems to be random data.  the m() command here
seems to work - subsequent d() show the change.

Sooooo, where should this memory be seen?

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Real time clock
Date: Tue, 15 Aug 2000 17:48:41 -0400
From: "Bryan D. Lee" 
Organization: Sanders-A Lockheed Martin Company
Message-ID: <3999BAB9.DEB3AAC8@lmco.com>
References: <8nbm79$i4m$1@crusher.de.colt.net>

I don't know if this is close to what you are looking for but Bob
Herlien wrote a clock interface based off of the on board RTC in
ftp://ftp.atd.ucar.edu/pub/vxworks/vx/usrTime.shar.

It includes a julian date function which you might be able to reverse
to get the date.  But if you already have an RTC, why bother with
absolute seconds?

Bryan


Achim Zimmer wrote:
> 
> has anybody programmed a clock including day, month and year based on RTC ?
> 
> Is there any free source for calculating date and time from a second counter
> noting leapyears aso ?
> 
> Thanks
> Achim

- -- 
Bryan D. Lee                             Phone: 603/885-7025
Software Engineer & ELDP                 Fax  : 603/885-3016
Sanders-A Lockheed Martin Company        Pager: 603/423-5565
PO Box 868, PTP1-2361, Nashua, NH 03061  Email: bryan.d.lee@lmco.com

---------------------------

Newsgroups: comp.os.vxworks
Subject: $$$$$$  Free Music!!!  $$$$$$$$$!
Date: Wed, 16 Aug 2000 4:19:1 GMT
From: aaa@aaa.com
Organization: Hong Kong Boardband Network Ltd.
Message-ID: <8nd4kg$ebq$1221@news.ctimail.com>






Please click here if you Want to Listen to Free Musics!!!!



---------------------------

Newsgroups: comp.os.vxworks
Subject: BitBand Hiring (ISRAEL)
Date: Wed, 16 Aug 2000 10:36:25 +0200
From: Leonid Rosenboim 
Organization: BitBand Technologies Ltd. http://www.bitband.com
Message-ID: <399A5289.8721F35B@bitband.com>
Reply-To: leonid@bitband.com

BitBand has a number of open positions for highly-skilled engineers,
location is Tel-Aviv Israel.
For details please clock here:
http://www.bitband.com/references/newsgroup.html
or send your CV to 



---------------------------

Newsgroups: comp.os.vxworks
Subject: re: TCP retransmittions and lost packet
Date: Wed, 16 Aug 2000 09:23:31 +0100 (BST)
From: David Laight 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <200008160823.JAA21380@dsl-2.tadpole.co.uk>

Chris wrote:

> > 4) make ABSOLUTELY sure that nothing anywhere on the network your side of a
> > store and forward switch (these usually route as well) has full duplex
> > enabled.

> I have read this before and tried it when we were having packet dropping 
> problems with a cisco 3500 -- locking the switches ports to half duplex did
> not help. We actually had to put a hub between the vxworks boards and the
> switch. Is there a way to lock this down under vxworks on an autonegotiating
> port (short of digging into a driver)? Why does this seem to cause such a
> problem?

The BIG problem with full duplex is that the ethernet device driver and 
transceivers at both ends of the link have to agree on using it.

The transceivers use information encoded on the link test pulses to determine 
the 'best' transfer speed.  Two mechanisms are used, auto-detect and 
auto-negotiation.

A 10Mb transceiver (on a network card or in a hub) sends out a single link test 
pulse at a fixed frequency - these light the 'link test' led on the receiving 
tranceiver.

A 100Mb transceiver sends out a short burst of link test pulses at a fixed 
frequency.  A 10M only tranceiver will think this is a valid link test pulse.

A 10/100 autodetecting tranceiver can differentiate between the above data 
streams and will transmit at the fastest speed.

A transceiver that support auto-negotiation encodes its capabilities (10M, 
100T4, 100TX, 10FDX, 100FDX etc) into the link test data stream.  The two 
transceivers then agree on the fastest available transfer rate.  This is likely 
to be full duplex if supported by both transceivers,

Ethernet CSMACD (Carrier Sense Media Access with Collison Detect) works on 
twisted pair by disallowing transmit if there is a receive signal (this works 
much much better than on coax where an average absolute cabe voltage has to be 
detected in the presence of 50/60Hz hum).

If the two ends of a cable disagree over full duplex then two packets will be 
transmitted at the same time and one gets lost.  (maybe a 'late collision' 
error).

If the ethernet driver believes the link is half duplex, but the transceiver 
thinks it is full duplex, then the driver is likely to report 'no carrier' 
during transmit.  (This is the absense of the receive enable caused by its own 
transmittions - as is seen on an AUI interface to a coax segment).

The problem is that the transceiver can flip between full and half duplex at 
almost any time (eg plug into a different remote system).  The device driver has 
to notice this change and change its own behaviour to match.  However the 
interface between the driver and transceiver (the MII interface) doen't make 
this easy.  Indeed within the standard transceiver registers it isn't even 
possible to determine the speed!

I actually suspect that many of the vxworks drivers get this completely wrong!
My driver (for the Xircom pcmcia cards - written under an NDA) does:

	/* Change Auto negotiate advertise register so we only advertise
           supported options (in particular exclude FDX) */
        td_write_mii( eth, MII_ADDR_REG( 0, MII_ANAR ),
                td_read_mii( eth, MII_ADDR_REG( 0, MII_ANAR ) ) & ~0x1f40  );

to ensure the transceiver doesn't enable full duplex.

If you get this far (without suffering brain overload) also:

> What do you mean by queued? Put in an mbuf or some other low level buffer?

Could do anything depending on the card hardware and the amount of queued data.
Typically the hardware will be given the first few packets (or until its FIFO is 
full), the rest have to be queued within the driver (or an appropriate response 
given to the protocol stack so that it will queue the packet).

It is IMPERITIVE that transmit packets are not discarded just because the LAN is 
busy (or the mbuf list (or whatever) has lots of fragments).

> Any suggestions on how do I check this?

>From the network trace it should be possible to see whether the sender or 
receiver are losing packets.

	David


- ----------------------------------------------------------------
David Laight                          email: dsl@tadpole.co.uk
Tadpole Technology plc                phone: +44 1223 428 232
Cambridge, UK                         fax:   +44 1223 428 201


---------------------------

End of New-News digest
**********************


From vxwexplo-errs@csg.lbl.gov  Thu Aug 17 10:23:08 2000
From: Henrik Bergstrom 
Date: Thu Aug 17 10:23:12 PDT 2000
Subject: MUX: mNextPkt does not work!?

VxWorks users,

I would like to pass several linked network packets to the MUX by
utilising the mNextPkt pointer in the mBlkHdr but this does not seem to
work.

Is there something I have to do except setting the mNextPkt pointer?

/Henrik

- --
Henrik Bergstrom  +46-8-628 28 28  henrik.bergstrom@intertex.se
Intertex Data AB  (fax) 628 64 14



From vxwexplo-errs@csg.lbl.gov  Thu Aug 17 10:23:09 2000
From: David Anderson 
Date: Thu Aug 17 10:23:12 PDT 2000
Subject: dosFs 1.0 

vxWorks Tornado
-------------------------

Can anyone point me at the correct way of verifying a ram block device 
contains a valid dos volume before calling
dosFsInit(...) which appears to crash with bus exceptions if the BLOCK_DEV 
has not previously been formatted.
( the manual actually states that it should return NULL volume 
descriptor.... )

e.g.

      pBlkDev = ramDevCreate( address , 512 , 400 , 400 , 0 );
      pVolDesc= dosFsDevInit("DEV1:",pBlkDev,NULL )
      if ( pVolDesc == NULL )
      {
              pVolDesc = dosFsMkFs("DEV1:",pBlkDev);
      }

does not work - dosFsDevInit() fails with an address exception ( attempting 
to perform a bcopy(...) ).

p.s. we are talking dosfs 1.0 and vxworks 5.3.1 here - old stuff so there 
must be a guru for it by now !.


David Anderson.  mailto:danderson@dtrack.demon.co.uk

p.s. It is a bright sunny 26 centigrade here in the UK. It does not always 
rain !.

-----------------------------------------------------
Who said "There's no life here in outer space."
-----------------------------------------------------

Disclaimer
----------
This email and any files transmitted with it are confidential and intended 
solely for the use of the individual or entity to whom they are addressed.

If you have received this email in error please destroy and notify Data 
Track Technology Plc +44 1425 271900.
-------------------------------------------------------



From vxwexplo-errs@csg.lbl.gov  Thu Aug 17 10:56:38 2000
From: Vxworks Exploder 
Date: Thu Aug 17 10:56:40 PDT 2000
Subject: comp.os.vxworks newsdigest

Comp.Os.Vxworks Daily Digest    Thu Aug 17 10:12:17 PDT 2000

        Subject: Fundamental Questions on VxWorks!
        Subject: How to ifconfig a private address
        Subject: mv2700: pci io space - Rav. MPIC Reg -vs- "vme reg"
        Subject: Re: struct Alignment
        Subject: 1 Need RTOS, File System and Database developers
        Subject: Re: Target Server Connect Problem?
        Subject: Re: Target Server Connect Problem?
        Subject: 1c Need Developers
        Subject: Re: signal problem
        Subject: Re: Fundamental Questions on VxWorks!
        Subject: Re: 1c Need Developers
        Subject: Re: Bothered by DHCP
        Subject: struct Alignment
        Subject: Re: task name garbled - newbie q
        Subject: Re: ioctl and pipes
        Subject: Re: task name garbled - newbie q
        Subject: Re: ioctl and pipes
        Subject: DosFS2.0 and D drives
        Subject: Re: Raw IP sockets and port number
        Subject: Re: mv2700: pci io space - Rav. MPIC Reg -vs- "vme reg"
        Subject: Re: cannot build bootrom_res.hex
        Subject: comp.os.qnx; comp.os.vxworks; comp.realtime
        Subject: task name garbled - newbie q
        Subject: Need Database, RTOS and File System developers
        Subject: Re: Adventnet SNMP agent for vxworks
        Subject: [Q] IPoA solution under VxWorks??
        Subject: Re: Fundamental Questions on VxWorks!
        Subject: Watchdogs
        Subject: IVR consultants/contract in So CA
        Subject: Missing mfree(), m_freem(), mbstat(), and m_more() from 5.3 to 5.4
        Subject: how to install ATA flash disk
        Subject: ZafText --> MultiLine?
        Subject: Win2K port
        Subject: Re: Looking for help with MPCxxx and NEC SDRAM
        Subject: Re: Watchdogs
        Subject: arpShow does not show interface
        Subject: connect() returns EADDRINUSE
        Subject: Re: arpShow does not show interface
        Subject: Problems with method routines in WindNet SNMP
        Subject: Re: task name garbled - newbie q
        Subject: Memory problem, help needed, pls !
        Subject: Re: mv2700: pci io space - Rav. MPIC Reg -vs- "vme reg"
        Subject: Re: Need Database, RTOS and File System developers

-------------------------------------------------------

Newsgroups: comp.os.vxworks
Subject: Fundamental Questions on VxWorks!
Date: Wed, 16 Aug 2000 10:21:17 GMT
From: joriveek@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8ndpuj$a1g$1@nnrp1.deja.com>

Hi all,

I'll be very much thankful if any1 on this board could give some clue
on the following questions.

i) What is the difference between "sp task1" and "task1" from the shell.

I've written a very small program which is suppose to take a character
from the shell (prompt) and print the same again. When I was invoking
as a "task1" from the shell, it is working fine...as it is looking for
the character and when I give the character, it is printing the same.
Alternatively,
when I invoke the task from the shell as "sp task1" then, it is pended
as it is waiting for the character to be given by me. In that case, how
to give the character?

ii) I've the Tornado on the PowerPC, now I'm looking for the small
program which can do the following: When I press a character from the
keyboard on the shell/hyper terminal, it should able to generate an
interrupt and run the ISR. Now for this should have to use SCC in UART
mode or SMC in UART mode. Any Clue???
Thanks for your help in advance/
Jrk


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: How to ifconfig a private address
Date: Wed, 16 Aug 2000 18:29:14 +0200
From: Theo =?iso-8859-1?Q?=D6jerteg?= 
Organization: A Customer of Tele2
Message-ID: <399AC15A.DDED5899@student.liu.se>
Reply-To: theoj055@student.liu.se


Does anyone know how to add a private ip address,
that is, one that is not published on the internet,
with ifAddrAdd?
I want to send ip packets on a raw socket
with an ip address different from the one
currently published on my network interface.

*thanx*
/Theo


---------------------------

Newsgroups: comp.os.vxworks
Subject: mv2700: pci io space - Rav. MPIC Reg -vs- "vme reg"
Date: Wed, 16 Aug 2000 12:55:03 -0400
From: Al Johnston 
Organization: MindSpring Enterprises
Message-ID: <399AC767.4BEE7767@mindspring.com>

I am still trying to figure out the memory spaces on this card.
Looks to me like there are two devices occupying the same space
so I must be missing something... Can someone enlighten me?

The things called "vme registers" (LM/SIG/SEM stuff) is located
at address 0000,1000 through 0000,1006 in pci io space. (2700
sbc ref. guide pg 1-40).

the vxw bsp routine sysPciExtRavenInit() (p/o sysLib) says the following
  
  pciConfigOutLong(busNo,deviceNo,funcNo,PCI_CFG_BASE_ADDRESS_0
    0x00000000);

and has the comment "init raven's MPIC control register access address
in I/O
and memory spaces.  IOSPACE - 0x00000000 [no access]"

Seems to me that this is putting the MPIC registers at the same place in
pci io space that the vme registers are already at...  What am I not
understanding please (I no absolutely nothing of pciConfig stuff).

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: struct Alignment
Date: Wed, 16 Aug 2000 12:38:18 -0500
From: Andrew Johnson 
Organization: APS, Argonne National Laboratory
Message-ID: <399AD18A.F1A3A8BD@aps.anl.gov>
References: <399AAB64.A6BB9291@lucent.com>

Sunil Joseph wrote:
> 
>         Does anybody know how to align structures at 16 byte boundaries using the gnu
> compiler. Having used other compilers I tried using  the #pragma but doesnt seem
> to work.

Look up __attribute__ in the gcc info files or the relevent chapter of the
WRS Gnu Toolkit User's Guide.

- - Andrew
- -- 
Sharks kill 10 people each year.  People kill 60,000,000 sharks each year.

---------------------------

Newsgroups: comp.os.qnx,comp.os.vxworks,comp.realtime
Subject: 1 Need RTOS, File System and Database developers
Date: Wed, 16 Aug 2000 17:41:26 GMT
From: "Scott A. Vigil" 
Organization: Centura Solutions
Message-ID: 

1
Hello Developers,

We have several consulting positions available for people interested in
helping our customers put databases in their embedded systems.  If
you have experience in one or more of the following and can handle up
to 25% travel, we may be interested in you.

o embedded C or C++
o real-time operating systems
o file systems
o databases

Centura Solutions is a consulting firm which specializes in providing
solutions to users of database, connectivity and integrated development
products.  We have implemented embedded solutions for companies such as
Lucent, Nortel and Hughes Networks.  Centura has the stability of an
established company and the excitement of an internet startup.

If such a faced paced, performance oriented environment sounds good to
you, we may have your dream job.  Send us your resume.

The opportunity is in Seattle.  However, if you really think you're hot
stuff but must stay in California or on the East coast, send in your resume
anyways.  We might be able to make a deal.

Send resume to David.Figge@CenturaSoft.com.  Please do _not_ hit .
Also, please do not post reply to news group.

Centura Solutions
1111 3rd Avenue, twenty-eighth floor
Seattle, WA 98101
http://www.centurasolutions.com/solutions/whatwedo.asp
http://www.centurasoft.com




---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Target Server Connect Problem?
Date: Wed, 16 Aug 2000 13:42:03 -0400
From: "Mike Kordik" 
Message-ID: 
References: <8oTl5.128$Vi5.400@client> <8navci$hvf$3@news1.xs4all.nl>  <8nbhef$4ro$1@news1.xs4all.nl>  <399A93EA.572@worldnet.att.net>

Mary,
   Thanks for the info. Unfortunately I still get the same error messages.

Thanks,
Mike

"Mary Frantz"  wrote in message
news:399A93EA.572@worldnet.att.net...
> Mike Kordik wrote:
> >
> > I am getting a bunch of these messages when I start the target server:
> >
> > tgtsvr (vxTarget@art_vandelay): Tue Aug 15 10:40:58 2000
> >     Error: rpccore backend client Procedure unavailable
> >
> > Do you know what this means?
> >
> > I was getting a cache full erorr but after I increased the cache size
that
> > error went away but I still get the above error.
> >
> > Thanks,
> > Mike
> >
>
> I had a similar problem.  Try this:
>
> Select Tools->Target Server->Configure from the menu ... a dialog box
> will come up.
>
> Under "Target Server Properies" select "Back End" from the drop down.
> Select "wdbrpc" from "Available Back Ends"
> Set the "Timeout" (I am using 10)
> Set the "Re-try" (I am using 2)
> (With these settings the virtual console takes awhile to appear.)
> Set the "Target name/IP address" to the actual IP address of your
> target.
>
> This fixed it for me.
>
> Mary Frantz



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Target Server Connect Problem?
Date: Wed, 16 Aug 2000 13:18:46 GMT
From: Mary Frantz 
Organization: AT&T Worldnet
Message-ID: <399A93EA.572@worldnet.att.net>
References: <8oTl5.128$Vi5.400@client> <8navci$hvf$3@news1.xs4all.nl>  <8nbhef$4ro$1@news1.xs4all.nl> 

Mike Kordik wrote:
> 
> I am getting a bunch of these messages when I start the target server:
> 
> tgtsvr (vxTarget@art_vandelay): Tue Aug 15 10:40:58 2000
>     Error: rpccore backend client Procedure unavailable
> 
> Do you know what this means?
> 
> I was getting a cache full erorr but after I increased the cache size that
> error went away but I still get the above error.
> 
> Thanks,
> Mike
> 

I had a similar problem.  Try this:

Select Tools->Target Server->Configure from the menu ... a dialog box
will come up.

Under "Target Server Properies" select "Back End" from the drop down.
Select "wdbrpc" from "Available Back Ends"
Set the "Timeout" (I am using 10)
Set the "Re-try" (I am using 2)
(With these settings the virtual console takes awhile to appear.)
Set the "Target name/IP address" to the actual IP address of your
target.

This fixed it for me.

Mary Frantz

---------------------------

Newsgroups: comp.os.vxworks,comp.realtime
Subject: 1c Need Developers
Date: Wed, 16 Aug 2000 17:52:59 GMT
From: "Scott A. Vigil" 
Organization: Centura Solutions
Message-ID: <%xAm5.1315$3T4.258898@sea-read.news.verio.net>

1c

Hello Developers,

We have several consulting positions available for people interested in
helping our customers put databases in their embedded systems.  If
you have experience in one or more of the following and can handle up
to 25% travel, we may be interested in you.

o embedded C or C++
o real-time operating systems
o file systems
o databases

Centura Solutions is a consulting firm which specializes in providing
solutions to users of database, connectivity and integrated development
products.  We have implemented embedded solutions for companies such as
Lucent, Nortel and Hughes Networks.  Centura has the stability of an
established company and the excitement of an internet startup.

If such a faced paced, performance oriented environment sounds good to
you, we may have your dream job.  Send us your resume.

The opportunity is in Seattle.  However, if you really think you're hot
stuff but must stay in California or on the East coast, send in your resume
anyways.  We might be able to make a deal.

Send resume to David.Figge@CenturaSoft.com.  Please do _not_ hit .
Also, please do not post reply to news group.

Centura Solutions
1111 3rd Avenue, twenty-eighth floor
Seattle, WA 98101
http://www.centurasolutions.com/solutions/whatwedo.asp
http://www.centurasoft.com




---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: signal problem
Date: Wed, 16 Aug 2000 13:05:50 -0500
From: "M. Kotiaho" 
Organization: Lockheed Martin Missiles and Fire Control - Dallas
Message-ID: <399AD7FE.CA5396E9@NULL.lmco.com>
References: <39982275.E312031@best.ms.philips.com>

Well, I don't know if this helps, but I was able to run your program on
a PowerPC board with no problem!  Perhaps there is something wrong in
your BSP ...

- --
MK

aalimari wrote:

> Hello
>
> Being a new to VxWorks I am having the following problem.
> The attached program crashes with general protection fault. I run
> vxsig in the SIMNT on intel x86 (Tornado 2.0). The crash happens when
> I send a signal to another task. When signal is sent to self the
> problem does not occur.
>
> Any help is appreciated
> Thanks in advance,
> Artem Alimarine
>
>
> The program
> -------------------------------------------------
> #include 
> #include 
> #include 
> #include 
> #include 
>
> static bool raised = false;
>
> void handler(int)
> {
>  raised = true;
>  printf("HANDLER\n");
> }
>
>
> void task(int)
> {
>  signal(SIGUSR1, handler);
>
>  //sigset_t sigs = 0xFFFFFFFF;
>  //int s = sigtimedwait(&sigs, NULL, NULL);
>
>  //printf("awaited %d\n", s);
>  //raise(SIGUSR1);
>  //kill(taskIdSelf(), SIGUSR1);
>
>  while(!raised) taskDelay(1);
>
>  printf("raised\n");
> }
>
>
> void vxsig()
> {
>     int tid;
>
>
>     tid = taskSpawn(
>      "test",
>      0,
>      VX_FP_TASK | VX_STDIO,
>      32 * 1024,
>      (FUNCPTR)task,
>      0,0,0,0,0, 0,0,0,0,0);
>
>     printf("created\n");
>     taskShow(taskIdSelf(), 2);
>
>     taskDelay(50);
>
>     printf("cancel\n");
>     kill(tid, SIGUSR1);
>     printf("canceled\n");
>     taskShow(taskIdSelf(), 2);
>
>     while(!raised) taskDelay(1);
>     printf("OK\n");
> }
>
> The simulator window displays:
> --------------------------------
> Exception !
>  Vector 13 : General Protection Fault
>   Program Counter : 0x0041d8e1
>  Status Register : 0x00010202
>
> The shell window:
> --------------------
> -> vxsig
> created
>
>   NAME        ENTRY       TID    PRI   STATUS      PC       SP
> ERRNO  DELAY
> ---------- ------------ -------- --- ---------- -------- --------
> ------- -----
> tExcTask   41ecf8        4b88e58   0 PEND         430906  4b88d74
> d0003     0
> tLogTask   412814        4b84440   0 PEND         430906  4b84358
> d0003     0
> test       49aeec4       4b753f8   0 SUSPEND      41d8e1  4b75226
> d0003     0
> test       49aeec4       4b63ea8   0 DELAY        430906  4b63e10
> d0003     1
> tWdbTask   40b960        4b80c20   3 PEND+T       430906  4b80b04
> d0003   180
> t1         414c58        4b7c6a0   4 DELAY        430906  4b7c5d0
> d0003     1
> t2         414c58        4b6b150   4 READY        4166e8  4b6a244
> d0003     0
> cancel
> canceled
>
>   NAME        ENTRY       TID    PRI   STATUS      PC       SP
> ERRNO  DELAY
> ---------- ------------ -------- --- ---------- -------- --------
> ------- -----
> tExcTask   41ecf8        4b88e58   0 PEND         430906  4b88d74
> d0003     0
> tLogTask   412814        4b84440   0 PEND         430906  4b84358
> d0003     0
> test       49aeec4       4b753f8   0 SUSPEND      41d8e1  4b75226
> d0003     0
> test       49aeec4       4b63ea8   0 SUSPEND      41d8e1  4b63cd6
> d0003     0
> tWdbTask   40b9
> Exception number 0: Task: 0x4b63ea8 (test)
>
> General Protection Fault
> Program Counter:            0x0041d8e1
> Status Register:            0x00010202
>
> 60        4b80c20   3 PEND+T       430906  4b80b04   d0003   180
> t1         414c58        4b7c6a0   4 DELAY        430906  4b7c5d0
> d0003     1
> t2         414c58        4b6b150   4 READY        4166e8  4b6a25c
> d0003     0
>
>


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Fundamental Questions on VxWorks!
Date: Wed, 16 Aug 2000 13:39:53 GMT
From: Mary Frantz 
Organization: AT&T Worldnet
Message-ID: <399A98DD.C73@worldnet.att.net>
References: <8ndpuj$a1g$1@nnrp1.deja.com>

joriveek@my-deja.com wrote:
> 
> Hi all,
> 
> I'll be very much thankful if any1 on this board could give some clue
> on the following questions.
> 
> i) What is the difference between "sp task1" and "task1" from the shell.
> 
> I've written a very small program which is suppose to take a character
> from the shell (prompt) and print the same again. When I was invoking
> as a "task1" from the shell, it is working fine...as it is looking for
> the character and when I give the character, it is printing the same.
> Alternatively,
> when I invoke the task from the shell as "sp task1" then, it is pended
> as it is waiting for the character to be given by me. In that case, how
> to give the character?

I assume you are using standard input/output to get and print this
character (getchar(), or scanf(), and printf()).  When you run directly
from the shell ("task1") the standard input and output are redirected to
the shell.  When you spawn it as a task ("sp task1"), then standard
input/output goes the the console.  Run hyperterminal with the COM port
connected to the target console port.  Type the character there. If your
target server is running a virtual console, the standard output will
appear on the virtual console.

You can also force the standard i/o to the console device within your
code. I use the following code to start a utility that uses
Hyperterminal for all its input and output, no matter how I launch it.



extern int consoleFd;	/* global fd of initial console device */


/*** Set standard i/o to console ***/
	
    ioTaskStdSet (0, STD_IN,  consoleFd);
    ioTaskStdSet (0, STD_OUT, consoleFd);
    ioTaskStdSet (0, STD_ERR, consoleFd);

		/* See Programmer's Guide, pg 132 sec 3.7.1 */

    ioctl(consoleFd, FIOSETOPTIONS, OPT_TERMINAL & ~OPT_MON_TRAP);

	/* Wait for the user to type a letter */
    printf("\n\nEnter any letter to begin.\n");		
    scanf("%c", &cmd);
	
Note that with OPT_TERMINAL set you have to hit the Enter key before the
application sees any characters.

Hope this helps.
Mary Frantz

---------------------------

Newsgroups: comp.os.vxworks,comp.realtime
Subject: Re: 1c Need Developers
Date: Wed, 16 Aug 2000 13:40:15 -0500
From: "Derek Opitz" 
Organization: Motorola CIG
Message-ID: <8nen4b$92o1@nntp.cig.mot.com>
References: <%xAm5.1315$3T4.258898@sea-read.news.verio.net>

Dude,  one job spam I can live with, but 4?  I, and I think most people,
come to this newsgroup looking for info on vxWorks implementation, not to
read spam about job positions.

Scott A. Vigil  wrote in message
news:%xAm5.1315$3T4.258898@sea-read.news.verio.net...
> 1c
>
> Hello Developers,
>
> We have several consulting positions available for people interested in
> helping our customers put databases in their embedded systems.  If
> you have experience in one or more of the following and can handle up
> to 25% travel, we may be interested in you.
>
> o embedded C or C++
> o real-time operating systems
> o file systems
> o databases
>
> Centura Solutions is a consulting firm which specializes in providing
> solutions to users of database, connectivity and integrated development
> products.  We have implemented embedded solutions for companies such as
> Lucent, Nortel and Hughes Networks.  Centura has the stability of an
> established company and the excitement of an internet startup.
>
> If such a faced paced, performance oriented environment sounds good to
> you, we may have your dream job.  Send us your resume.
>
> The opportunity is in Seattle.  However, if you really think you're hot
> stuff but must stay in California or on the East coast, send in your
resume
> anyways.  We might be able to make a deal.
>
> Send resume to David.Figge@CenturaSoft.com.  Please do _not_ hit .
> Also, please do not post reply to news group.
>
> Centura Solutions
> 1111 3rd Avenue, twenty-eighth floor
> Seattle, WA 98101
> http://www.centurasolutions.com/solutions/whatwedo.asp
> http://www.centurasoft.com
>
>
>



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Bothered by DHCP
Date: Wed, 16 Aug 2000 14:47:55 GMT
From: mchowdry@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8ne9ii$s24$1@nnrp1.deja.com>
References: <8lhhgl$l6o$1@nnrp1.deja.com>

DHCP has been a pain in my backside for some time too. I've been trying
to get a DHCP server running off my vxWorks image for the MCP750.
Although I have the Server running prom a pool table, it's not what i'm
looking for. I simply can't (even after liasing with the WindRiver
support) get it to work from a non-pool table based set-up. If anyone
out there has managed this, I'd be interested to hear.
Ellin, after speaking with WindRiver's support, it appears the order in
which you define the entries can affect how DHCP works. Play around
with the order of the parameters for each entry and see if it makes any
difference


In article <8lhhgl$l6o$1@nnrp1.deja.com>,
  ellin_lin@my-deja.com wrote:
> I'm a tenderfoot of vxWorks and Tornado. And I
> met some problems when I tried to configured a
> DHCP server. I modified usrnetwork.c setting the
> IP range as written in user manual, and test this
> DHCP server with a DHCP client of Win98, it did
> nothing. Can someone give me some guidance?
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.
>


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: struct Alignment
Date: Wed, 16 Aug 2000 10:55:32 -0400
From: Sunil Joseph 
Organization: Lucent Technologies, Columbus, Ohio
Message-ID: <399AAB64.A6BB9291@lucent.com>

Folks,
	Does anybody know how to align structures at 16 byte boundaries using the gnu
compiler. Having used other compilers I tried using  the #pragma but doesnt seem
to work. 
Sunil

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: task name garbled - newbie q
Date: Wed, 16 Aug 2000 20:22:25 GMT
From: john_94501@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8net59$kmh$1@nnrp1.deja.com>
References: <8nei63$70i$1@nnrp1.deja.com>

Hi,

This probably means that you are running off the bottom of the task
stack (the name is held there). Try checkStack to see if any of the task
stacks have overflowed; if they have, allocate some more memory to the
stack when you spawn the task.

HTH,

John...

In article <8nei63$70i$1@nnrp1.deja.com>,
  vasu_nukala@my-deja.com wrote:
> Whenever I start a specific task using taskSpawn
> in my code and I look at it in the shell using
> i the name is not the usual t# or the name I give
> it but a funny characters ?
>
> Is that normal ?
>
> thanx
> nd
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.
>


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: ioctl and pipes
Date: Wed, 16 Aug 2000 17:50:26 +0200
From: Theo =?iso-8859-1?Q?=D6jerteg?= 
Organization: A Customer of Tele2
Message-ID: <399AB842.D90A3FEE@student.liu.se>
References: <8nbabm$aen$1@proxy2.fe.internet.bosch.de>
Reply-To: theoj055@student.liu.se

I have the same problem as well and would be glad to hear if anyone has
some hints to share.
/Theo

Gunter Ritter wrote:

> Hello!
>
> I think I have some problems using ioctl and pipes.
>
> After writing some messages into different pipes, I try to find out how many
> messages are in one of these pipes. I use ioctl and FIONSMGS. But instead of
> the number of messages in the pipe I always find 0 (zero) in the 3rd
> argument of ioctl.
>
> What´s going wrong. Is there anybody who has experience with ioctl and
> pipes.
>
> Thanks.
>
> Gunter


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: task name garbled - newbie q
Date: Wed, 16 Aug 2000 15:40:24 -0500
From: Greg Willden 
Organization: Posted via Supernews, http://www.supernews.com
Message-ID: <399AFC38.3C64AB87@swri.edu>
References: <8nei63$70i$1@nnrp1.deja.com>

vasu_nukala@my-deja.com wrote:

> Whenever I start a specific task using taskSpawn
> in my code and I look at it in the shell using
> i the name is not the usual t# or the name I give
> it but a funny characters ?
>
> Is that normal ?

No that is not normal.  My guess is that you have a memory overwrite in
your code.  For example writing past the end of an array might cause a
problem like that.
    Good luck.


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: ioctl and pipes
Date: Wed, 16 Aug 2000 17:58:02 +0200
From: Theo =?iso-8859-1?Q?=D6jerteg?= 
Organization: A Customer of Tele2
Message-ID: <399ABA0A.B228E3E6@student.liu.se>
References: <8nbabm$aen$1@proxy2.fe.internet.bosch.de>
Reply-To: theoj055@student.liu.se


I also have problem setting the pipe to non-blocking mode
with ioctl and FIONBM.... it seems to be possible
I simply do not want the write() vall to block if the pipe is full.
Any hints?
/Theo



Gunter Ritter wrote:

> Hello!
>
> I think I have some problems using ioctl and pipes.
>
> After writing some messages into different pipes, I try to find out how many
> messages are in one of these pipes. I use ioctl and FIONSMGS. But instead of
> the number of messages in the pipe I always find 0 (zero) in the 3rd
> argument of ioctl.
>
> What´s going wrong. Is there anybody who has experience with ioctl and
> pipes.
>
> Thanks.
>
> Gunter


---------------------------

Newsgroups: comp.os.vxworks
Subject: DosFS2.0 and D drives
Date: Wed, 16 Aug 2000 11:52:12 -0400
From: Susan LoVerso 
Message-ID: <399AB8AB.A55FDBCE@loverso.southborough.ma.us>

I am using DosFS 2.0 and the x86 BSP.  I have 2 disks that are set up as
follows:
Disk 0:  6Gb drive
partition 1 4Gb   Windows98 (active partition) C:
partition 2 2Gb   Non-DOS partition

Disk 1: 4Gb drive
partition 1 2Gb  Windows98  D:
partition 2 1Gb   Non-DOS partition
(and 1Gb not yet partitioned)

Both of my Windows partitions were created with fdisk, under windows.
The D: partition
is "exclusively for VxWorks use".  I can access both C: and D: under
Windows.  I put
2 small text files onto D: under Windows so that I can attempt to see
them under VxWorks.

When I boot VxWorks, I can see the files that are normally C:, but not
D:.
- -> devs
drv name
  4 /ata0a
  4 /ata1a

- -> @ls "/ata0a"
[... list shortened for brevity]
/ata0a/My Documents
/ata0a/Program Files
/ata0a/win98
/ata0a/W9x_552
/ata0a/SCANDISK.LOG
value = 0 = 0x0
- -> @ls "/ata1a"
Can't open "/ata1a".
value = -1 = 0xffffffff

There are no messages on the console from that 'ls' command.

When that didn't work, I tried the steps in the DosFS 2.0 Release Notes:

- ->  < /vxtmp/vxtmp/d1.script

#
# Create atadevice on drive 1
pblk1 = ataDevCreate(0,1,0,0)
new symbol "pblk1" added to symbol table.
pblk1 = 0x599760: value = 67100180 = 0x3ffde14
#
# Set up cache
pcbio1 = dcacheDevCreate(pblk1, 0, 0x40000, "disk")
new symbol "pcbio1" added to symbol table.
pcbio1 = 0x599750: value = 67099916 = 0x3ffdd0c
#
# mount partitions
pcbiop1 = dpartDevCreate(pcbio1, 2, usrFdiskPartRead)
new symbol "pcbiop1" added to symbol table.
pcbiop1 = 0x599748: value = 67096384 = 0x3ffcf40
#
# create dosfs device on partition 0
dosFsDevCreate("/disk1",dpartPartGet(pcbiop1,0), 0x100, 0)
value = 0 = 0x0
- -> devs
drv name
  4 /ata0a
  4 /ata1a
  4 /disk1
value = 0 = 0x0
- -> @ls "/disk1"
Can't open "/disk1".
value = -1 = 0xffffffff

Again, there are no error messages associated with this ls command
either.

I also cannot do a 'dosFsShow' on the D drive either:
- -> dosFsShow "/ata0a"

volume descriptor ptr (pVolDesc):       0x3f8c898
cache block I/O descriptor ptr (pCbio): 0x3f8cb64
auto disk check on mount:               DOS_CHK_REPAIR | DOS_CHK_VERB_2
max # of simultaneously open files:     22
file descriptors in use:                0
# of different files in use:            0
# of descriptors for deleted files:     0
# of  obsolete descriptors:             0

current volume configuration:
 - volume label:        NO NAME ; (in boot sector:      NO NAME    )
 - volume Id:           0x1f2707de
 - total number of sectors:     8,193,087
 - bytes per sector:            512
 - # of sectors per cluster:    8
 - # of reserved sectors:       32
 - FAT entry size:              FAT32
 - # of sectors per FAT copy:   7,994
 - # of FAT table copies:       2
 - # of hidden sectors:         63
 - first cluster is in sector # 16,020
 - directory structure:         VFAT
 - root dir start cluster:      2

FAT handler information:
- ------------------------
 - allocation group size:       103 clusters
 - free space on volume:        3,780,792,320 bytes
value = 0 = 0x0
- -> dosFsShow "/ata1a"

volume descriptor ptr (pVolDesc):       0x3f6aa78
cache block I/O descriptor ptr (pCbio): 0x3f6ad44
auto disk check on mount:               DOS_CHK_REPAIR | DOS_CHK_VERB_2
max # of simultaneously open files:     22
file descriptors in use:                0
# of different files in use:            0
# of descriptors for deleted files:     0
# of  obsolete descriptors:             0

can't mount volume

value = 0 = 0x0

- --------------------------
Any suggestions on what I am doing wrong?  The first drive works just
fine without me
having to do anything.  They are similarly partitioned and accessible on
Windows.
Thanks for any help



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Raw IP sockets and port number
Date: Wed, 16 Aug 2000 18:03:03 +0200
From: Theo =?iso-8859-1?Q?=D6jerteg?= 
Organization: A Customer of Tele2
Message-ID: <399ABB37.62164158@student.liu.se>
References: <399415FD.D2A33025@ftw.mot.com>
Reply-To: theoj055@student.liu.se

Have a look in th e Unix Network programmers guide to see how this is
done...
There are som source code from the book (which is very good) to download
from
http://www.kohala.com/start/unpv12e.html

Basically, to send IP data on a raw socket, you just create the socket and
start sending data. specifying
If you specify IP_HDR_INCLUDE the the pointer to the data, used when
sending, should point to the very first byte in the IP header otherwise it
should be a pointer to the data in the ip packet.

good luck
/Theo


Michael White wrote:

> Hi,
>
> I want to use a Raw IP socket (type AF_INET) to send GRE messages.
> A couple of quick questions:
>
>    1)  How do I set the TOS in the IP header?
>    2)  Is the port field in the sockaddr used on a Raw IP socket?
>    3)  When sending a Raw IP message, what is the format of the data
>        that needs to be passed to sendto()?  Does it include the IP
>        header or just the GRE header and body?
>
> Any help would be greatly appreciated.  If it's somewhere in the VxWorks
> manuals,
> that'd be helpful, too.
>
> Thanks in advance.
> ---
> Michael White, qa1402@ftw.mot.com
> Cellular Infrastructure Group
> Ft. Worth, Texas    (817)245-6418


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: mv2700: pci io space - Rav. MPIC Reg -vs- "vme reg"
Date: Wed, 16 Aug 2000 21:36:30 GMT
From: rjdejoie@kmrmail.kmr.ll.mit.edu (Bob DeJoie)
Message-ID: <399b0863.1705161051@news>
References: <399AC767.4BEE7767@mindspring.com>

On Wed, 16 Aug 2000 12:55:03 -0400, Al Johnston
 wrote:

The last write in the sysPciExtRavenInit()  routine actually
is the one that enables the address decoders. This writes
a 0x6 to the PCI command register which enables bus
mastership and memory space accesses only. I/O space accesses
are not enabled so what the address decoder is set to is 
actually a NOP.


>I am still trying to figure out the memory spaces on this card.
>Looks to me like there are two devices occupying the same space
>so I must be missing something... Can someone enlighten me?
>
>The things called "vme registers" (LM/SIG/SEM stuff) is located
>at address 0000,1000 through 0000,1006 in pci io space. (2700
>sbc ref. guide pg 1-40).
>
>the vxw bsp routine sysPciExtRavenInit() (p/o sysLib) says the following
>  
>  pciConfigOutLong(busNo,deviceNo,funcNo,PCI_CFG_BASE_ADDRESS_0
>    0x00000000);
>
>and has the comment "init raven's MPIC control register access address
>in I/O
>and memory spaces.  IOSPACE - 0x00000000 [no access]"
>
>Seems to me that this is putting the MPIC registers at the same place in
>pci io space that the vme registers are already at...  What am I not
>understanding please (I no absolutely nothing of pciConfig stuff).

================================================
Bob DeJoie      
Kwajalein Missile Range
Republic of the Marshall Islands
email: rjdejoie@kmrmail.kmr.ll.mit.edu    
phone: 805-355-5855   fax: 805-355-3833
Time Zone: GMT +12 hours                     
================================================

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: cannot build bootrom_res.hex
Date: Wed, 16 Aug 2000 14:40:36 -0700
From: James Chen 
Organization: Globix Corp.
Message-ID: <399B0A54.56078C4E@gigabitwireless.com>
References: <3999B0AE.2FBA7E6@email.mot.com>

This is a multi-part message in MIME format.
- --------------8427BFCA08248A6D2361E5E9
Content-Type: multipart/alternative;
 boundary="------------AB53E02AE2816CBABF748140"


- --------------AB53E02AE2816CBABF748140
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

That's because in your romInit.s RAM_DST_ADRS is referenced but is not
defined in anywhere, or least it is not defined in a file that is on the
search path.

Bo Yu wrote:

> I cannot build rom resident image for PowerPC target:
>
> error message:
>
> ldppc -X -N  -e _romInit \
>     -Ttext FFF00000  -Tdata 1000000  -o bootrom_res romInit_res.o
> bootInit_res.o \
>     version.o bootConfig.o sysALib.o sysLib.o end82xx.o membankc.o
> memtest.o TestMenu.o Generator.o Dw_82xx.o System.o ports.o brgs.o
> fcc1.o fcc3.o scc1.o SMC1.o Timer1.o dw_mmu.o Services.o board.o
> Dw_err.o mem_heap.o mem_buff.o
> /usr/vob/sc/xc/psi.orig/src/os/tornado2/target/lib/libPPCEC603gnuvx.a
>
> romInit_res.o(.text+0x1ce): undefined reference to `RAM_DST_ADRS'
> romInit_res.o(.text+0x1d2): undefined reference to `RAM_DST_ADRS'
> make: *** [bootrom_res] Error 1
>
> I can build bootrom.hex and bootrom_uncmp.hex in the same BSP
> directory. Why cannot I
> build bootrom_res.hex ?
>
> Thanks.
>
>
> --
> Bo Yu
>
>

- --------------AB53E02AE2816CBABF748140
Content-Type: text/html; charset=us-ascii
Content-Transfer-Encoding: 7bit



That's because in your romInit.s RAM_DST_ADRS is referenced but is not
defined in anywhere, or least it is not defined in a file that is on the
search path.

Bo Yu wrote:

I cannot build rom resident image for PowerPC target:

error message:

ldppc -X -N  -e _romInit \

    -Ttext FFF00000  -Tdata 1000000  -o bootrom_res
romInit_res.o bootInit_res.o \

    version.o bootConfig.o sysALib.o sysLib.o end82xx.o
membankc.o memtest.o TestMenu.o Generator.o Dw_82xx.o System.o ports.o
brgs.o fcc1.o fcc3.o scc1.o SMC1.o Timer1.o dw_mmu.o Services.o board.o
Dw_err.o mem_heap.o mem_buff.o    /usr/vob/sc/xc/psi.orig/src/os/tornado2/target/lib/libPPCEC603gnuvx.a

romInit_res.o(.text+0x1ce): undefined reference to `RAM_DST_ADRS'

romInit_res.o(.text+0x1d2): undefined reference to `RAM_DST_ADRS'

make: *** [bootrom_res] Error 1

I can build bootrom.hex and bootrom_uncmp.hex in the same BSP directory.
Why cannot I

build bootrom_res.hex ?

Thanks.

 

-- 
Bo Yu

 



- --------------AB53E02AE2816CBABF748140--

- --------------8427BFCA08248A6D2361E5E9
Content-Type: text/x-vcard; charset=us-ascii;
 name="jchen.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Card for James Chen
Content-Disposition: attachment;
 filename="jchen.vcf"

begin:vcard 
n:Chen;James
tel;fax:408.577.0700
tel;work:408.577.7411
x-mozilla-html:FALSE
url:www.gigabitwireless.com
org:Gigabit Wireless, Inc;BTS Software Development
version:2.1
email;internet:jchen@gigabitwireless.com
adr;quoted-printable:;;3099 N. 1st Street,=0D=0A;San Jose;California;95134;
fn:James Chen
end:vcard

- --------------8427BFCA08248A6D2361E5E9--


---------------------------

Newsgroups: comp.os.qnx,comp.os.vxworks,comp.realtime
Subject: comp.os.qnx; comp.os.vxworks; comp.realtime
Date: Wed, 16 Aug 2000 17:40:27 GMT
From: "Scott A. Vigil" 
Organization: Centura Solutions
Message-ID: 

1
Hello Developers,

We have several consulting positions available for people interested in
helping our customers put databases in their embedded systems.  If
you have experience in one or more of the following and can handle up
to 25% travel, we may be interested in you.

o embedded C or C++
o real-time operating systems
o file systems
o databases

Centura Solutions is a consulting firm which specializes in providing
solutions to users of database, connectivity and integrated development
products.  We have implemented embedded solutions for companies such as
Lucent, Nortel and Hughes Networks.  Centura has the stability of an
established company and the excitement of an internet startup.

If such a faced paced, performance oriented environment sounds good to
you, we may have your dream job.  Send us your resume.

The opportunity is in Seattle.  However, if you really think you're hot
stuff but must stay in California or on the East coast, send in your resume
anyways.  We might be able to make a deal.

Send resume to David.Figge@CenturaSoft.com.  Please do _not_ hit .
Also, please do not post reply to news group.

Centura Solutions
1111 3rd Avenue, twenty-eighth floor
Seattle, WA 98101
http://www.centurasolutions.com/solutions/whatwedo.asp
http://www.centurasoft.com




---------------------------

Newsgroups: comp.os.vxworks
Subject: task name garbled - newbie q
Date: Wed, 16 Aug 2000 17:15:05 GMT
From: vasu_nukala@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8nei63$70i$1@nnrp1.deja.com>

Whenever I start a specific task using taskSpawn
in my code and I look at it in the shell using
i the name is not the usual t# or the name I give
it but a funny characters ?

Is that normal ?

thanx
nd


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.qnx,comp.os.vxworks,comp.realtime
Subject: Need Database, RTOS and File System developers
Date: Wed, 16 Aug 2000 17:47:08 GMT
From: "Scott A. Vigil" 
Organization: Centura Solutions
Message-ID: 

1

Hello Developers,

We have several consulting positions available for people interested in
helping our customers put databases in their embedded systems.  If
you have experience in one or more of the following and can handle up
to 25% travel, we may be interested in you.

o embedded C or C++
o real-time operating systems
o file systems
o databases

Centura Solutions is a consulting firm which specializes in providing
solutions to users of database, connectivity and integrated development
products.  We have implemented embedded solutions for companies such as
Lucent, Nortel and Hughes Networks.  Centura has the stability of an
established company and the excitement of an internet startup.

If such a faced paced, performance oriented environment sounds good to
you, we may have your dream job.  Send us your resume.

The opportunity is in Seattle.  However, if you really think you're hot
stuff but must stay in California or on the East coast, send in your resume
anyways.  We might be able to make a deal.

Send resume to David.Figge@CenturaSoft.com.  Please do _not_ hit .
Also, please do not post reply to news group.

Centura Solutions
1111 3rd Avenue, twenty-eighth floor
Seattle, WA 98101
http://www.centurasolutions.com/solutions/whatwedo.asp
http://www.centurasoft.com




---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Adventnet SNMP agent for vxworks
Date: Wed, 16 Aug 2000 18:30:34 GMT
From: randyj5573@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8nemkb$cl5$1@nnrp1.deja.com>
References: <3999AE69.AF526782@yahoo.com>

Why don't you use Wind River's SNMP agent?

Randy Jones

In article <3999AE69.AF526782@yahoo.com>,
  Bhavesh Patel  wrote:
> Hi,
>
> Has anybody used adventnet embedded/C snmp agent for vxworks
platform? I
> would like to know their experiences.  Adventnet website does not have
> example/demo for vxworks platform.
>
> Thanks for any help.
>
> Bhavesh Patel
>
>


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: [Q] IPoA solution under VxWorks??
Date: Thu, 17 Aug 2000 09:33:39 +0900
From: "???" 
Organization: Electronics and Telecommunications Research Institute (ETRI)
Message-ID: <8nfbqe$dfu$1@news.etri.re.kr>

Hello, folks

I'm looking for IPoA solution for following embedded board.
If you have or know anything, could you let me know a.s.a.p?

CPU Core: MPC750
SAR: NEC uPD98405
RTOS: VxWorks 5.x

Thank you for reading. Bye.

Jerry
- --
***************************************************
Jehun Rhee (Jerry)
- ---------------------------------------------------
IMT-2000 System Development Dept.
Radio & Broadcasting Technology Lab., ETRI
161 Kajong Dong, Yusong Gu, Taejon, 305-350, Korea
- ---------------------------------------------------
Voice Phone        : +82-42-860-3820 (office)
Fax                    : +82-42-860-6500 (office)
Email                 : jhrhee@etri.re.kr
- ---------------------------------------------------




---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Fundamental Questions on VxWorks!
Date: Wed, 16 Aug 2000 13:42:16 -0500
From: "Derek Opitz" 
Organization: Motorola CIG
Message-ID: <8nen84$hrv2@nntp.cig.mot.com>
References: <8ndpuj$a1g$1@nnrp1.deja.com>

sp task1 spawns a task called task1, task1 executes task1 as a function
running in the shell task.

 wrote in message news:8ndpuj$a1g$1@nnrp1.deja.com...
> Hi all,
>
> I'll be very much thankful if any1 on this board could give some clue
> on the following questions.
>
> i) What is the difference between "sp task1" and "task1" from the shell.
>
> I've written a very small program which is suppose to take a character
> from the shell (prompt) and print the same again. When I was invoking
> as a "task1" from the shell, it is working fine...as it is looking for
> the character and when I give the character, it is printing the same.
> Alternatively,
> when I invoke the task from the shell as "sp task1" then, it is pended
> as it is waiting for the character to be given by me. In that case, how
> to give the character?
>
> ii) I've the Tornado on the PowerPC, now I'm looking for the small
> program which can do the following: When I press a character from the
> keyboard on the shell/hyper terminal, it should able to generate an
> interrupt and run the ISR. Now for this should have to use SCC in UART
> mode or SMC in UART mode. Any Clue???
> Thanks for your help in advance/
> Jrk
>
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.



---------------------------

Newsgroups: comp.os.vxworks
Subject: Watchdogs
Date: Wed, 16 Aug 2000 20:07:09 GMT
From: "Gagnon, Sylvain" 
Organization: Sympatico
Message-ID: 
Sender: "Gagnon, Sylvain" 

Hi all,

I'm currently using a watchdog timer to schedule the call of a C
function. I wonder if this watchdog timer will delete itself after
completion of my C function. If not, can I delete it in C function?

Thanks,
Sylvain Gagnon


---------------------------

Newsgroups: comp.os.vxworks
Subject: IVR consultants/contract in So CA
Date: Wed, 16 Aug 2000 19:58:01 GMT
From: manojkn@my-deja.com
Organization: Priority Search Partners
Message-ID: <8nerns$j2e$1@nnrp1.deja.com>
Reply-To: manoj@prioritysp.com

3IVR consultants for troubleshooting/maintenance/implementation. The key
word here is Voicetek IVR. Alternative IVRs are Aspect and/or
Periphone. 1 year of IVR maintenance/implementation will do. Additional
preferences are CTI and ACDs. 4 to 6 years of total experience. 6
months contract minimum in a huge communiation company. So CA. rate is
DOE but not more than $70.00/hr. on a Corp.-to-Corp. or $60.00 on a W2
hourly rate. This is immediate.


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Missing mfree(), m_freem(), mbstat(), and m_more() from 5.3 to 5.4
Date: Wed, 16 Aug 2000 18:29:43 -0400
From: Chris Williamson 
Message-ID: <399B15D6.AEA273A9@erols.com>

We've got code that uses these functions, and we discovered that they
aren't in VxWorks 5.4. Does anyone know how to get them into the kernel?
Are they gone for good? The code is in a FDDI driver.

Thanks.
Chris Williamson


---------------------------

Newsgroups: comp.os.vxworks
Subject: how to install ATA flash disk
Date: Wed, 16 Aug 2000 21:57:36 GMT
From: jun_li@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8nf2o6$rgt$1@nnrp1.deja.com>

Can anybody tell me what I need to do
if I want to make a ATA flash disk work for vxworks?

Thanks in advance,

Jun Li


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: ZafText --> MultiLine?
Date: Thu, 17 Aug 2000 08:45:43 +0200
From: "Morne Hartman" 
Message-ID: <399b8976$0$226@hades.is.co.za>

Hi all,

I would like to have multiple lines of text placed into my ZafText box. I am
placing text messages into the box from code, and I do not allow the user to
edit the text. However if a new message is to be displayed, it overwrites
the existing message instead of appearing in the next line below it.

How can I remedy this problem.

Many Thanks
Morne Hartman
mhartman@grintek.com




---------------------------

Newsgroups: comp.os.vxworks
Subject: Win2K port
Date: Thu, 17 Aug 2000 08:45:48 +0200
From: "zbody" 
Organization: Slovenija OnLine - SiOL
Message-ID: 

Hello!

I would like to run Tornado 2.0 under Win2K and I have heard that some
software exists which can turn Tornado2 into such a mode that it works under
Win2K (because originaly WindRiver doesn't support Tornado2 under Win2K
yet).
Please, if somebodi knows the name of it and maybe where can I find it let
me know!!

Thanks!

Regards!

Darko



---------------------------

Newsgroups: comp.arch.embedded,comp.os.vxworks,comp.sys.powerpc.tech
Subject: Re: Looking for help with MPCxxx and NEC SDRAM
Date: Thu, 17 Aug 2000 08:02:31 GMT
From: dksee@hotmail.com
Organization: Deja.com - Before you buy.
Message-ID: <8ng66h$3a9$1@nnrp1.deja.com>
References: <8m47mq$63o$1@newsreaderg1.core.theplanet.net> <398797A0.4CAE31DD@austin.rr.com>

In article <398797A0.4CAE31DD@austin.rr.com>, Richard Hendricks wrote:
> Try looking for the Interfacing to SDRAM PDF file I posted
> here a while back.  If you can't find it, let me know and
> I'll repost it.

I've looked, but I can't find it. Has it been canceled from
the archives???

Can you repost it, or post the URL?

Thanks!
David


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Watchdogs
Date: 17 Aug 2000 09:30:58 GMT
From: johan@borksoft.xs4all.nl
Organization: Borksoft
Message-ID: <8ngbci$dc7$1@news1.xs4all.nl>
References: 

Gagnon, Sylvain  wrote:
: Hi all,

: I'm currently using a watchdog timer to schedule the call of a C
: function. I wonder if this watchdog timer will delete itself after
: completion of my C function. If not, can I delete it in C function?

When this timer expires is stops, but is not deleted. The watchdog is
executed as an interrupt, so be carefull with what you are doing in this
routine.
When the watchdog-routine is finished the system continues with what it
was doing before, there is no special action required to finish the action
of a watchdog, only when you want thet timer te restart you have to do
this in the routine.

Groeten,
	Johan

- -- 
     o o o o o o o . . .   ___________________________________
    o      _____           ||    Johan Borkhuis              |
  .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
 >(________|__|_[_________]_|________________________________|
 _/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: arpShow does not show interface
Date: Wed, 16 Aug 2000 23:17:56 -0700
From: "Sandhya Potluri" 
Organization: Monmouth Internet
Message-ID: <8nflcc$i0l$1@news.monmouth.com>

Folks,
    I am trying to debug a Ethernet interface. I made all the necessary
calls to attach the interface,
associate an ip address/mask  with the interface, and  none of the calls
returned an error. muxShow indicates  that the
interface is present, similarly ifShow indicates the  interface/ip address/
mac address /mask... However when I
did arpShow the interface info does not show up.I also  have a second
Ethernet interface on the board which I used for boot etc, which is up and
running and obviously this shows up on arpShow. Is there a problem having
two interfaces?
Or does it point to other problems. Should I remove the existing interface
before I attach the new interface ?
I checked the Ethernet controller  the MII interface   they seems to work
fine, I was also able to write/read raw Ethernet packets without interfacing
with the vxworks TCP Stack( raw reads/writes to registers / interrupt
handler). This points to some
problem in my interface with the stack. Any help to untangle this
mystery/(misery):=) is appreciated.
Thanks
Sandy



---------------------------

Newsgroups: comp.os.vxworks
Subject: connect() returns EADDRINUSE
Date: Thu, 17 Aug 2000 04:31:04 GMT
From: chinmay@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8nfpq9$leb$1@nnrp1.deja.com>

Hello:

   I am running vxWorks on a Motorola processor. The version is 5.4 and
the host os is Solaris.

   I am getting intermittent failures when I call connect() system
call. The error number is 48 (EADDRINUSE). I dont call bind() before I
call connect(). The socket is set to be non blocking.

   Has anyone experienced any similar problem.

Regards,

Chinmay



Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: arpShow does not show interface
Date: Thu, 17 Aug 2000 11:31:43 +0530
From: Venkatesh Prakash 
Organization: Lucent Technologies
Message-ID: <399B7FC7.4040100@lucent.com>
References: <8nflcc$i0l$1@news.monmouth.com>

Hi,
It is not necessary that arpShow should display that interface. arpShow 
will display all the ip address to MAC address mapping learnt on that 
interface. If nothing is learnt on a particular interface no entry will 
be made for that interface.

One can attach as many interfaces as he can. There is no restriction on 
that. If you can send me ifShow,routeShow and arpShow dumps I may help you.

Venki

Sandhya Potluri wrote:

> Folks,
>     I am trying to debug a Ethernet interface. I made all the necessary
> calls to attach the interface,
> associate an ip address/mask  with the interface, and  none of the calls
> returned an error. muxShow indicates  that the
> interface is present, similarly ifShow indicates the  interface/ip address/
> mac address /mask... However when I
> did arpShow the interface info does not show up.I also  have a second
> Ethernet interface on the board which I used for boot etc, which is up and
> running and obviously this shows up on arpShow. Is there a problem having
> two interfaces?
> Or does it point to other problems. Should I remove the existing interface
> before I attach the new interface ?
> I checked the Ethernet controller  the MII interface   they seems to work
> fine, I was also able to write/read raw Ethernet packets without interfacing
> with the vxworks TCP Stack( raw reads/writes to registers / interrupt
> handler). This points to some
> problem in my interface with the stack. Any help to untangle this
> mystery/(misery):=) is appreciated.
> Thanks
> Sandy
> 
> 
> 


---------------------------

Newsgroups: comp.os.vxworks,comp.protocols.snmp
Subject: Problems with method routines in WindNet SNMP
Date: Thu, 17 Aug 2000 09:56:46 +0200
From: "S. Schmidt" 
Organization: roka GmbH
Message-ID: <8ng5pe$6u2$1@news10.roka.net>

Hello,

I have problems with the generated code of the mib-compiler if I use a MIB
with tabular variables. I do not understand how I can accomodate the
next-function to my application.

Could anyone help me with a sample-code?

Thanks



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: task name garbled - newbie q
Date: Thu, 17 Aug 2000 14:29:00 +0530
From: Venkatesh Prakash 
Organization: Lucent Technologies
Message-ID: <399BA954.3010904@lucent.com>
References: <8nei63$70i$1@nnrp1.deja.com>

Hi,
I also faced the same problem and finaly I found that this will be due 
to stack overflow. You can also check for stack overflow.

Venki 


vasu_nukala@my-deja.com wrote:

> Whenever I start a specific task using taskSpawn
> in my code and I look at it in the shell using
> i the name is not the usual t# or the name I give
> it but a funny characters ?
> 
> Is that normal ?
> 
> thanx
> nd
> 
> 
> Sent via Deja.com http://www.deja.com/
> Before you buy.
> 


---------------------------

Newsgroups: comp.os.vxworks
Subject: Memory problem, help needed, pls !
Date: Thu, 17 Aug 2000 09:45:26 GMT
From: Arcady Gilinsky 
Organization: Deja.com - Before you buy.
Message-ID: <8ngc7j$9hi$1@nnrp1.deja.com>

Hi VxWorkers !
I'm  just the beginner.
I'm using the Tornado 2 for NT platform and can't adjust memory for
target server. I looking in the Tornado Registry there is only 2.00 Mb
of memory and when I try to allocate f.e. 4 Mb of memory, while load my
*.OUT file, I get the following error message in "VxWorks Simulator for
Windows" window:
"0x6b40c08 (tWdbTask): memPartAlloc: block too big - 4718592 in
partition 0x446298."
errno = 0x110001
I tried to adjust the USER_RESERVED_MEMORY, but seems like this doesn't
affects on memory size :(

Help me pls!
P.S.
Sorry for my English.


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: mv2700: pci io space - Rav. MPIC Reg -vs- "vme reg"
Date: Thu, 17 Aug 2000 12:07:33 -0400
From: Al Johnston 
Organization: MindSpring Enterprises
Message-ID: <399C0DC5.CE970007@mindspring.com>
References: <399AC767.4BEE7767@mindspring.com> <399b0863.1705161051@news>


> The last write in the sysPciExtRavenInit()  routine actually

I figured it was something like that... but I couldnt find it.
Thanks!

- -al

---------------------------

Newsgroups: comp.os.qnx,comp.os.vxworks,comp.realtime
Subject: Re: Need Database, RTOS and File System developers
Date: Thu, 17 Aug 2000 16:21:49 GMT
From: "Scott A. Vigil" 
Organization: Centura Solutions
Message-ID: 
References: 

Sorry for the redundant post.
News server error messages made it appear my posts were being rejected.

Scott Vigil
Lead, Embedded Software Team
Centura Solutions
1111 3rd Avenue, twenty-eighth floor
Seattle, WA 98101
http://www.centurasolutions.com/solutions/whatwedo.asp
http://www.centurasoft.com




---------------------------

End of New-News digest
**********************


From vxwexplo-errs@csg.lbl.gov  Fri Aug 18 00:17:57 2000
From: "Krishna Kishore Reddy Sureddy" 
Date: Fri Aug 18 00:17:59 PDT 2000
Subject: vxWorks taskStatusString

Hi,
      In vxWorks Reference Manual (Page Number 2-886)
     Under "taskStatusString" call its given that
        String                Meaning
       SUSPEND        Task is unavailable for execution(but not
suspended,delayed or pended).

      What does the "but not suspended" in above sentence mean?

  Thanks and Regards,
   S.K.K.Reddy.





From vxwexplo-errs@csg.lbl.gov  Fri Aug 18 04:03:07 2000
From: Vxworks Exploder 
Date: Fri Aug 18 04:03:10 PDT 2000
Subject: comp.os.vxworks newsdigest

Comp.Os.Vxworks Daily Digest    Fri Aug 18 04:03:04 PDT 2000

        Subject: SMC9432 driver?
        Subject: powerpc-wrs-vxworks cross compiler
        Subject: MUX: mNextPkt does not work!?
        Subject: dosFs 1.0
        Subject: Does anyone have experience with VNC on VxWorks
        Subject: Re: Problems with method routines in WindNet SNMP
        Subject: Flash File System
        Subject: Bootable VxWorks Image problem
        Subject: libdbm
        Subject: RMI on vxWorks?
        Subject: Re: Bootable VxWorks Image problem
        Subject: Re: Missing mfree(), m_freem(), mbstat(), and m_more() from 5.3 to 5.4
        Subject: Network File system
        Subject: Re: comp.os.qnx; comp.os.vxworks; comp.realtime
        Subject: Re: 1c Need Developers
        Subject: Re: Win2K port
        Subject: Re: Flash File System
        Subject: RTOS Comparison
        Subject: Re: how to install ATA flash disk
        Subject: MediaGX and 8139 END
        Subject: Re: Win2K port
        Subject: Problem creating files on host
        Subject: Watchdog interrupt level
        Subject: Hello !!
        Subject: Re: How to adust telnet priority
        Subject: vxWorks taskStatusString
        Subject: Re: Bootable VxWorks Image problem
        Subject: Re: Problems with method routines in WindNet SNMP
        Subject: Memory r/w problems
        Subject: SCSI driver for sym53C860
        Subject: Simultaneous use of BSD and END interface
        Subject: Recommendation needed for a shell with history, etc

-------------------------------------------------------

Newsgroups: comp.os.vxworks
Subject: SMC9432 driver?
Date: Thu, 17 Aug 2000 19:08:52 +0200
From: Zdenek Sebek 
Organization: Czech Technical University
Message-ID: <399C1C21.51471AEC@fel.cvut.cz>

hi
has anyone used or developed a driver for  SMC9432 PCI NIC (based on
83C171 chip) under VxWorks?  I have to run vxworks on several PCs and I
would like to use PCI NICs there in order to save ISA slots. But manual
says the only PCI NIC supported is Intel E100.
Any help welcomed.




---------------------------

Newsgroups: comp.os.vxworks
Subject: powerpc-wrs-vxworks cross compiler
Date: Thu, 17 Aug 2000 17:38:53 GMT
From: "Rob Targosz" 
Organization: Excite@Home - The Leader in Broadband
Message-ID: 

Anybody had any luck building the following for an i686-linux-pc-gnu host to
powerpc-wrs-vxworks target:

binutils-2.10
gcc-2.95.2
gdb-5.0
newlib-1.8.2

I'm using the following configuration:
Redhat Linux 2.6 on a Pentium III 550
Built and installed above packages as local tools.

1. Build binutils-2.01 (works):
configure --target=powerpc-wrs-vxworks --prefix=/opt/gnu
make all install

2. Build gcc-2.95.2 (doesn't work):
configure --target=powerpc-wrs-vxworks --prefix=/opt/gnu --with-headers=/usr
/local/powerpc-wrs-vxworks/include
make all install

Gives me an error:
../../../gcc-2.95.2/libiberty/choose-temp.c:38:67: sys/file.h no such file
or directory
make: *** [all-target-libiberty] Error 2

I have used the "patch" info from Brad Garcia to get this far:
added #define SIZE_TYPE "unsigned int" to vxi386.h and vx960-coff.h
removed operator== from tinfo.cc
removed printf function define from libio/gen-params

Also made minor changes to vxCpu.h (Tornado headers) to define my CPU PPC604
and CPU_FAMILY PPC and fixed some checks that were WIN32 specific (that's
where I got the headers from).

Any help at all is appreciated!

Thanks,
Rob Targosz
robtargosz@hotmail.com



---------------------------

Newsgroups: comp.os.vxworks
Subject: MUX: mNextPkt does not work!?
Date: Thu, 17 Aug 2000 15:12:17 +0200 (CEST)
From: Henrik Bergstrom 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <8nha2v$955$1@overload.lbl.gov>

VxWorks users,

I would like to pass several linked network packets to the MUX by
utilising the mNextPkt pointer in the mBlkHdr but this does not seem to
work.

Is there something I have to do except setting the mNextPkt pointer?

/Henrik

- - --
Henrik Bergstrom  +46-8-628 28 28  henrik.bergstrom@intertex.se
Intertex Data AB  (fax) 628 64 14


---------------------------

Newsgroups: comp.os.vxworks
Subject: dosFs 1.0
Date: Thu, 17 Aug 2000 14:02:32 +0100
From: David Anderson 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <8nha2v$956$1@overload.lbl.gov>

vxWorks Tornado
- -------------------------

Can anyone point me at the correct way of verifying a ram block device 
contains a valid dos volume before calling
dosFsInit(...) which appears to crash with bus exceptions if the BLOCK_DEV 
has not previously been formatted.
( the manual actually states that it should return NULL volume 
descriptor.... )

e.g.

      pBlkDev = ramDevCreate( address , 512 , 400 , 400 , 0 );
      pVolDesc= dosFsDevInit("DEV1:",pBlkDev,NULL )
      if ( pVolDesc == NULL )
      {
              pVolDesc = dosFsMkFs("DEV1:",pBlkDev);
      }

does not work - dosFsDevInit() fails with an address exception ( attempting 
to perform a bcopy(...) ).

p.s. we are talking dosfs 1.0 and vxworks 5.3.1 here - old stuff so there 
must be a guru for it by now !.


David Anderson.  mailto:danderson@dtrack.demon.co.uk

p.s. It is a bright sunny 26 centigrade here in the UK. It does not always 
rain !.

- -----------------------------------------------------
Who said "There's no life here in outer space."
- -----------------------------------------------------

Disclaimer
- ----------
This email and any files transmitted with it are confidential and intended 
solely for the use of the individual or entity to whom they are addressed.

If you have received this email in error please destroy and notify Data 
Track Technology Plc +44 1425 271900.
- -------------------------------------------------------


---------------------------

Newsgroups: comp.os.vxworks
Subject: Does anyone have experience with VNC on VxWorks
Date: Thu, 17 Aug 2000 18:17:26 GMT
From: lpeck@sd.inri.com
Organization: Deja.com - Before you buy.
Message-ID: <8nha70$d7e$1@nnrp1.deja.com>

I am investigating using VNCViewer on a PowerPC running VxWorks.  As
far as I know there is not a version of VNCViewer readily available,
but I do have the Unix source code. Has anyone attempted/ had success
in porting the VNCViewer to VxWorks?


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks,comp.protocols.snmp
Subject: Re: Problems with method routines in WindNet SNMP
Date: Thu, 17 Aug 2000 18:21:56 GMT
From: randyj5573@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8nhafc$dp2$1@nnrp1.deja.com>
References: <8ng5pe$6u2$1@news10.roka.net>

In article <8ng5pe$6u2$1@news10.roka.net>,
  "S. Schmidt"  wrote:
> Hello,
>
> I have problems with the generated code of the mib-compiler if I use
a MIB
> with tabular variables. I do not understand how I can accomodate the
> next-function to my application.
>
> Could anyone help me with a sample-code?
>
> Thanks
>
>

Hello Sven,

Wind River has some example code for using tables on their web site.
You need to contact customer service and they will tell you how to
access it.

Otherwise, the MIB compiler just generates a skeleton and you have to
fill in or add to it to make it work. The next-function uses the table
index that you defined in your MIB. The example code that Wind River
has shows how to use an Integer as the table index, but the RFC's allow
you to use an IP Address, Octet String and others. You need to give
more info.

Regards,
Randy


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Flash File System
Date: Thu, 17 Aug 2000 14:34:57 -0400
From: Henry 
Organization: UUNET Canada News Reader Service
Message-ID: <399C3051.5583033E@mail.com>

This is a multi-part message in MIME format.
- --------------D8565D9744D21977B54B5BD9
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Hi, All,

Does VxWorks support such system natively or I need to get 3rd party
supports (if so, who makes such offer))?

Thanks.

Henry

- --------------D8565D9744D21977B54B5BD9
Content-Type: text/x-vcard; charset=us-ascii;
 name="hkhoo.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Card for Henry
Content-Disposition: attachment;
 filename="hkhoo.vcf"

begin:vcard 
n:Khoo;Henry 
x-mozilla-html:FALSE
version:2.1
email;internet:henry@solinetsystems.com
fn:Henry Khoo
end:vcard

- --------------D8565D9744D21977B54B5BD9--


---------------------------

Newsgroups: comp.os.vxworks
Subject: Bootable VxWorks Image problem
Date: Thu, 17 Aug 2000 19:19:50 +0530
From: slkumar 
Organization: Motorola
Message-ID: <399BED7D.E034873@miel.mot.com>

Hello

We have MCP750 target. We created a new bootable project based on the
MCP750 BSP. We built the project without any customization. We tried
booting the Target with the bootable VxWorks image. The target
successfully downloaded the image but it hanged thereafter.

Any idea how to get over this problem ?

Regards
Shailendra


---------------------------

Newsgroups: comp.os.vxworks
Subject: libdbm
Date: Thu, 17 Aug 2000 16:53:32 +0200
From: Andy Winton 
Organization: Swisscom IP+ (post doesn't reflect views of Swisscom)
Message-ID: <399BFC6C.5D62E718@stest.ch>

hi,
	Is there a port of the UNIX libdbm for
	VxWorks?

	You know the one with functions -
	
		dbm_open, dbm_store, dbm_find, etc.

	cheers,

andy

- --
Andy Winton.

---------------------------

Newsgroups: comp.os.vxworks
Subject: RMI on vxWorks?
Date: Thu, 17 Aug 2000 20:22:29 GMT
From: "Dennis Wang" 
Organization: Magma Communications Ltd.
Message-ID: <9QXm5.335$Td.2608@news.magma.ca>

Hi vxWorkers:

Does anyone have any information on using Java RMI on vxWorks?

Regards


Dongsheng Wang



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Bootable VxWorks Image problem
Date: Thu, 17 Aug 2000 16:53:17 +0200
From: Sebastien Mallet 
Organization: Nortel Networks
Message-ID: <399BFC5D.3E10238A@nortelnetworks.com>
References: <399BED7D.E034873@miel.mot.com>

Hi,
 The image will not work without modification. The
  user has to invoke project facility and change the 'default' build
  specification to use optimization level other than 'none' under the
  'C/C++ compiler' tab. Please refer to SPR #27046 for more details.
Sebastien Mallet





---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Missing mfree(), m_freem(), mbstat(), and m_more() from 5.3 to 5.4
Date: Thu, 17 Aug 2000 16:01:05 GMT
From: John Efstathiades 
Organization: Deja.com - Before you buy.
Message-ID: <8nh27l$3ao$1@nnrp1.deja.com>
References: <399B15D6.AEA273A9@erols.com>

In article <399B15D6.AEA273A9@erols.com>,
  Chris Williamson  wrote:
> We've got code that uses these functions, and we discovered that they
> aren't in VxWorks 5.4. Does anyone know how to get them into the
kernel?
> Are they gone for good? The code is in a FDDI driver.

Chris,

I suspect that your FDDI driver is a BSD4.3 driver. You will have to
upgrade it to BSD4.4 before it will run under VxWorks 5.4. The VxWorks
Network Programmer's Guide has some information on how to do this.

HTH,
John.

John Efstathiades
Telelogic UK


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Network File system
Date: Thu, 17 Aug 2000 15:16:53 GMT
From: waqarahmad@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8ngvkf$vrg$1@nnrp1.deja.com>

We are using vxWorks v5.31 and use NFS to share files between
our target boards. We find NFS to be slow and would like to find an
alternative for that (CFS maybe). Does anyone out there have
any information on where one can get/buy such a file system that can
work with vxWorks or can be ported to vxWorks.

Waqar




Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.qnx,comp.os.vxworks,comp.realtime
Subject: Re: comp.os.qnx; comp.os.vxworks; comp.realtime
Date: Thu, 17 Aug 2000 16:35:05 GMT
From: "Scott A. Vigil" 
Organization: Centura Solutions
Message-ID: 
References: 

Sorry for the redundant post.
News server error messages made it appear my posts were being rejected.

Scott Vigil
Lead, Embedded Software Team
Centura Solutions
1111 3rd Avenue, twenty-eighth floor
Seattle, WA 98101
http://www.centurasolutions.com/solutions/whatwedo.asp
http://www.centurasoft.com




---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: 1c Need Developers
Date: Thu, 17 Aug 2000 16:36:46 GMT
From: "Scott A. Vigil" 
Organization: Centura Solutions
Message-ID: 
References: <%xAm5.1315$3T4.258898@sea-read.news.verio.net> <8nen4b$92o1@nntp.cig.mot.com>

Sorry for the redundant post.
News server error messages made it appear my posts were being rejected.

Scott Vigil
Lead, Embedded Software Team
Centura Solutions
1111 3rd Avenue, twenty-eighth floor
Seattle, WA 98101
http://www.centurasolutions.com/solutions/whatwedo.asp
http://www.centurasoft.com




---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Win2K port
Date: Thu, 17 Aug 2000 13:15:34 -0400
From: "Sean Evoy" 
Organization: UUNET Canada News Reader Service
Message-ID: <05Vm5.8102$Z2.126579@nnrp1.uunet.ca>
References: 

Darko,
You need to use the APCOMPAT.EXE utility that is available on your Win2K CD.
If you don't have the CD, go to Microsoft's website and download the
application compatibility toolkit to get it.Next, run apcompat and fiddle
the settings until you get them right. Hope this helps,

Sean Evoy
Rational Software


"zbody"  wrote in message
news:iSLm5.446$UZ2.12147@news.siol.net...
> Hello!
>
> I would like to run Tornado 2.0 under Win2K and I have heard that some
> software exists which can turn Tornado2 into such a mode that it works
under
> Win2K (because originaly WindRiver doesn't support Tornado2 under Win2K
> yet).
> Please, if somebodi knows the name of it and maybe where can I find it let
> me know!!
>
> Thanks!
>
> Regards!
>
> Darko
>
>



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Flash File System
Date: Thu, 17 Aug 2000 22:29:55 GMT
From: waqarahmad@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8nhp0d$vfr$1@nnrp1.deja.com>
References: <399C3051.5583033E@mail.com>



WindRiver has True FFS which they released not too long ago that works
with vxWorks. You should visit www.wrs.com and get the contact info.

Waqar



In article <399C3051.5583033E@mail.com>,
  Henry  wrote:
> This is a multi-part message in MIME format.
> --------------D8565D9744D21977B54B5BD9
> Content-Type: text/plain; charset=us-ascii
> Content-Transfer-Encoding: 7bit
>
> Hi, All,
>
> Does VxWorks support such system natively or I need to get 3rd party
> supports (if so, who makes such offer))?
>
> Thanks.
>
> Henry
>
> --------------D8565D9744D21977B54B5BD9
> Content-Type: text/x-vcard; charset=us-ascii;
>  name="hkhoo.vcf"
> Content-Transfer-Encoding: 7bit
> Content-Description: Card for Henry
> Content-Disposition: attachment;
>  filename="hkhoo.vcf"
>
> begin:vcard
> n:Khoo;Henry
> x-mozilla-html:FALSE
> version:2.1
> email;internet:henry@solinetsystems.com
> fn:Henry Khoo
> end:vcard
>
> --------------D8565D9744D21977B54B5BD9--
>
>


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: RTOS Comparison
Date: Fri, 18 Aug 2000 02:01:14 GMT
From: "Casey Ballentine" 
Organization: @Home Network
Message-ID: 

Hello,

I'm looking for an article or URL that compares several RTOS products.  I've
found some by searching industry rags, but the articles have all been
industry specific (i.e. RTOSes in the Automotive Industry).  What I'd like
is a more general comparison of RTOSes using several benchmarks (ROM and RAM
use, CPU usage, interrupt latency, etc.)  Does anyone have a good reference?

Thanks



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: how to install ATA flash disk
Date: Thu, 17 Aug 2000 17:40:50 -0400
From: "Mike Kordik" 
Message-ID: 
References: <8nf2o6$rgt$1@nnrp1.deja.com>

This is what I did to init and mount an ATA DiskOnModule:

pb = ataDevCreate(0,0,0,0)
dosFsMkfsOptionsSet(4)   // (4) for long filename support
dosFsMkfs("/ata0/",pb)
diskInit("/ata0/")

Hope this helps!

- -Mike

 wrote in message news:8nf2o6$rgt$1@nnrp1.deja.com...
> Can anybody tell me what I need to do
> if I want to make a ATA flash disk work for vxworks?
>
> Thanks in advance,
>
> Jun Li
>
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.



---------------------------

Newsgroups: comp.os.vxworks
Subject: MediaGX and 8139 END
Date: Thu, 17 Aug 2000 17:42:42 -0400
From: "Mike Kordik" 
Message-ID: 

Is anybody using this combination? If so are you able to debug using WDG and
the Target Server?

I get "Segment Error" when I try to debug either through the END or serial
driver if I have the END driver compiled and attached in my vxworks image.

Thanks,
Mike



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Win2K port
Date: 17 Aug 2000 09:37:21 GMT
From: johan@borksoft.xs4all.nl
Organization: Borksoft
Message-ID: <8ngboh$dc7$2@news1.xs4all.nl>
References: 

zbody  wrote:
: Hello!

: I would like to run Tornado 2.0 under Win2K and I have heard that some
: software exists which can turn Tornado2 into such a mode that it works under
: Win2K (because originaly WindRiver doesn't support Tornado2 under Win2K
: yet).
: Please, if somebodi knows the name of it and maybe where can I find it let
: me know!!

See http://www.xs4all.nl/~borkhuis/vxworks/vxw_pt1.html#1.11.2-D

Groeten,
	Johan

- -- 
     o o o o o o o . . .   ___________________________________
    o      _____           ||    Johan Borkhuis              |
  .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
 >(________|__|_[_________]_|________________________________|
 _/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: Problem creating files on host
Date: Fri, 18 Aug 2000 06:18:06 GMT
From: gardhab 
Organization: Deja.com - Before you buy.
Message-ID: <8nikek$t6t$1@nnrp1.deja.com>

hi all ,

i am trying to read a script file "test.txt" on the host (which is win
NT) .
i use open("host:G:/test.txt",0x2,0644).
this command returns a valid fd, the test file is opened,i'm able to
read the data in the file and parse the commands within.

but when i need to write the results of the test into a new file i call
open ( "host:G:/results.txt", 0x2|0x0200 , 0644)  ( with the O_CREAT
option)
this command returns ok with an fd , but there is no file created.
also using the fd , i'm able to write data to this file ( wherever it
is) , but when i read the data back , it is just junk.
i tried using the create call but this too gives the same result .

could someone tell me what i'm doing wrong here and how i can fix this
problem ?
thanks in advance  ,
karthik




Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Watchdog interrupt level
Date: Fri, 18 Aug 2000 06:22:04 GMT
From: gardhab 
Organization: Deja.com - Before you buy.
Message-ID: <8nikm1$tgc$1@nnrp1.deja.com>

hi ,
i'd like to know at what interrupt level a routine associated with
a Watchdog timer executes.
supppose after expiration, a WD timer startsup the associated routine
and during this execution,another WD timer expires.
does the first routine get preempted ? is there some way to achieve
this?
or is it necessary that 2nd WD timer routine waits for the 1st routine
to exit ?
is it possible that the 2nd execution does not happen as the interrupt
is lost ?
thanks in advance ,
karthik


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Hello !!
Date: Thu, 17 Aug 2000 19:16:08 -0400
From: Sunil Joseph 
Organization: Lucent Technologies, Columbus, Ohio
Message-ID: <399C7238.78E7F4A3@lucent.com>

Gopal,
	This is a test mail, I have been sending this for sometime now. Always get an
error.
Sunil

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: How to adust telnet priority
Date: Fri, 18 Aug 2000 09:09:15 +0200
From: Uwe Foss 
Organization: Siemens AG
Message-ID: <399CE11B.4D97191C@dit.siemens.de>
References: <3999017A.123B1758@dit.siemens.de> <8nbh16$2e5$1@news1.xs4all.nl> <39994838.1F4FE042@dit.siemens.de> <8nc3eg$cbj$1@news1.xs4all.nl>

You were right


- -----------usrNetwork.c-----------
telnetTaskPriority= xxx;
telnetInit ();
- -----------------------------------

was the solution.

Uwe

johan@borksoft.xs4all.nl schrieb:

> Uwe Foss  wrote:
> : But the problem is that the telnetXXTasks are spawned by the deamon
> : dynamically.
> : The application doesn't know when a user is contacting the system via
> : telnet.
>
> OK, that was an answer that was to quick.
>
> I checked the file telnetLib.o in the library. There is a symbol in there
> called telnetTaskPriority. Check if this is what you need. I cannot see if
> it is a function or a variable (strings does not provide that
> informmation and I don't have nm available), but my guess is that it is a
> variable.
>
> Hope this helps a little bit more.
>
> Groeten,
>         Johan
>
> --
>      o o o o o o o . . .   ___________________________________
>     o      _____           ||    Johan Borkhuis              |
>   .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
>  >(________|__|_[_________]_|________________________________|
>  _/oo OOOOO oo`  ooo   ooo  'o!o!o                      o!o!o`
> === VxWorks FAQ: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html ===

- --
- ----------------------------
SIEMENS AG
ATD TD 23 STG
Uwe Foss
Siemenstr. 33
71254 Ditzigen

Fon:  0711-137-6082
Fax:  0711-137-6071
Mail: uwe.foss@dit.siemens.de
- -----------------------------



---------------------------

Newsgroups: comp.os.vxworks
Subject: vxWorks taskStatusString
Date: Fri, 18 Aug 2000 12:49:15 +0530
From: "Krishna Kishore Reddy Sureddy" 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <8nir9v$8id$1@overload.lbl.gov>

Hi,
      In vxWorks Reference Manual (Page Number 2-886)
     Under "taskStatusString" call its given that
        String                Meaning
       SUSPEND        Task is unavailable for execution(but not
suspended,delayed or pended).

      What does the "but not suspended" in above sentence mean?

  Thanks and Regards,
   S.K.K.Reddy.




---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Bootable VxWorks Image problem
Date: Fri, 18 Aug 2000 12:26:51 +0400
From: "Arthur V. Grigorian" 
Organization: Auriga, Inc.
Message-ID: <399CF34B.E07B5537@auriga.ru>
References: <399BED7D.E034873@miel.mot.com> <399BFC5D.3E10238A@nortelnetworks.com> <399CD5FC.B5CD2A0C@miel.mot.com>

Hello,

   Try to disable ATA disk support in the HardWare configuration
   section of your project  (probably  the default configuration
   doesn't match your system)

Hope that helps,
__
R2

>slkumar wrote:
> 
> Hi Sebastien
> 
> Thanks for the information.
> I built the image with optimization level 2 (as recommended by the SPR).
> The target is still not coming up ...
> 
> Regards
> Shailendra
> 
> Sebastien Mallet wrote:
> 
> > Hi,
> >  The image will not work without modification. The
> >   user has to invoke project facility and change the 'default' build
> >   specification to use optimization level other than 'none' under the
> >   'C/C++ compiler' tab. Please refer to SPR #27046 for more details.
> > Sebastien Mallet

---------------------------

Newsgroups: comp.os.vxworks,comp.protocols.snmp
Subject: Re: Problems with method routines in WindNet SNMP
Date: Fri, 18 Aug 2000 00:53:53 GMT
From: Les Cargill 
Organization: AT&T Worldnet
Message-ID: <399CA61F.8D3B686C@worldnet.att.net>
References: <8ng5pe$6u2$1@news10.roka.net>



"S. Schmidt" wrote:
> 
> Hello,
> 
> I have problems with the generated code of the mib-compiler if I use a MIB
> with tabular variables. I do not understand how I can accomodate the
> next-function to my application.
> 
> Could anyone help me with a sample-code?
> 
> Thanks

The MIB compiler should be able to generate a default (  but broken ) set
of method routines for you. You'll have to modify them but it'll get you
started.

I forget the option. Check the manual. 

- --
http://home.att.net/~lcargill

---------------------------

Newsgroups: comp.os.vxworks
Subject: Memory r/w problems
Date: Fri, 18 Aug 2000 10:10:51 +0200
From: =?iso-8859-1?Q?Torbj=F6rn?= Frostemark 
Organization: FOA
Message-ID: <399CEF8B.4CDF6906@lin.foa.se>

This is a multi-part message in MIME format.
- --------------5537F42B539911E79F14315A
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

I have a MVME204-2F memory VME module with 2 MB starting at 0x0fe00000.
Reading and writing to it with a MVME147 is no problem, but with a
MVME2700 I always get errors, both from VxWorks applications and from
the Motorola BUG. I have tried different starting addresses, always the
same result. What am I doing wrong?

/Torbjorn

- --------------5537F42B539911E79F14315A
Content-Type: text/x-vcard; charset=us-ascii;
 name="torfr.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Card for Torbjörn Frostemark
Content-Disposition: attachment;
 filename="torfr.vcf"

begin:vcard 
n:Frostemark;Torbjörn
tel;cell:+46-(0)70-249 76 27
tel;fax:+46-(0)13-37 80 09
tel;work:+46-(0)13-37 80 14
x-mozilla-html:TRUE
org:FOA;Electronic Warfare Systems
adr:;;Box 1165;SE-581 11  Linköping;;;SWEDEN
version:2.1
email;internet:torfr@lin.foa.se
title:MSc
fn:Torbjörn Frostemark
end:vcard

- --------------5537F42B539911E79F14315A--


---------------------------

Newsgroups: comp.os.vxworks
Subject: SCSI driver for sym53C860
Date: Thu, 17 Aug 2000 18:34:30 -0700
From: "Sheela Kashyap" 
Organization: Posted via Supernews, http://www.supernews.com
Message-ID: 

Hi,

We have an x86 board for which I need to get the SCSI driver working. I am
trying to use the ncr810 driver for the 860. The sysScsi.c file I am using
is very similar to the one under config\templateX86. I find the memory
mapped address of the device by reading the PCI config space. I  built the
bootrom_uncmp binary without the SCSI driver and built the vxWorks image
with the SCSI driver. When I get to the bootrom prompt, I am able to dump
the memory mapped space (0xfc102000) for the SCSI, however while trying to
load the kernel, it page faults when reading the memory mapped space. I have
verified that the memory mapped address range has been added to the
sysPhysMemDesc table. I am able to dump the address space for the ethernet
which is at 0xfc100000 once the kernel boots. I am not sure why the OS is
not letting me touch the address space at 0xfc102000.

I have verified that the SCSI hard drive works. I am using Tornado 2.0 and
vxWorks 5.4. I would appreciate it greatly if somebody could throw some
light on this problem.

Thanks in advance,
Sheela.



---------------------------

Newsgroups: comp.os.vxworks
Subject: Simultaneous use of BSD and END interface
Date: Thu, 17 Aug 2000 03:28:56 +0900
From: "Sung Won Ho" 
Organization: ELIMnet Internet News
Message-ID: <8nems3$ikc$1@news.elim.net>

Hi all,

I want use two kind of NICs in one MIPS based VxWorks system.
One for BSD 4.4 interface and one for END interface.

No matter how I tried to find the way, I couldn't.

Does anyone can help me?

Sung.




---------------------------

Newsgroups: comp.os.vxworks
Subject: Recommendation needed for a shell with history, etc
Date: Fri, 18 Aug 2000 10:56:42 GMT
From: David Hoerl 
Organization: @Home Network
Message-ID: 

I am looking for a shell which would mimic the Cisco shell - ie, command 
line completion, history etc. RapidLogic is overkill for my application.

Any recommendation would be appreciated for a PD or otherwise product.

David Hoerl
dfh "at" home dot com   (reply address has EAT_SPAM_ spam filter)

---------------------------

End of New-News digest
**********************


From vxwexplo-errs@csg.lbl.gov  Fri Aug 18 04:13:38 2000
From: "Chandrashekar Nagaraj" 
Date: Fri Aug 18 04:13:40 PDT 2000
Subject: Task Suspend

Hi,
    I am hereby giving some cases for task suspension in vxWorks,

Are there are any cases other than these ?..........

1. vxWorks calls
   a. ts()
   b. taskSuspend()
   c. sigsuspend()
   d. pause()
   e. nanosleep()
   f. aio_suspend
   g. sigtimedwait()
   h. sigwaitinfo()

2. Task level exception

3. Breakpoint in task

4. Attaching task to Debugger

5. CS_MSEC_DELAY(msec), a vxWorks Macro

6. Error in Allocating/Freeing Memory

7. When a pipe is full and task attempts to write to a pipe

8. One task stack overflows and corrupts other tasks stack

9. ioctl() call causes non-recoverable error

10. If an invalid device descriptor in the valid range of device
descriptors
    numbers is passed to an I/O function

11. Some errors caused in the system even though the task itself
    may not have caused that error such as
       a. Device driver doing a hardware access that generates an error
       b. Bus error caused by a DMA cycle which was set up by a totally
different task
       c. Bus errors and access violations due to task level code,
driver code, interrupt handlers


Thanks,
KK...



From vxwexplo-errs@csg.lbl.gov  Sat Aug 19 03:53:00 2000
From: "V.Suraj" 
Date: Sat Aug 19 03:53:03 PDT 2000
Subject: RE: comp.os.vxworks newsdigest


David,

Try this.

BLK_DEV *pBlkDev;
DOS_VOL_DESC *pVolDesc;
pBlkDev = ramDevCreate (0, 512, 400, 400, 0);
pVolDesc = dosFsMkfs ("DEV1:", pBlkDev);

Once the ramDrive has been created, you can use dosFsDevInit.
dosFsDevInit() must be used only if the file system already exists on the
disk and should not be reinitialized.

Also ensure that the ramAdd is valid. (first parameter in the ramDevCreate
function.). Better yet, make it 0, and let system allocate memory for your
disk.

--------------------------------

Can anyone point me at the correct way of verifying a ram block device
contains a valid dos volume before calling
dosFsInit(...) which appears to crash with bus exceptions if the BLOCK_DEV
has not previously been formatted.
( the manual actually states that it should return NULL volume
descriptor.... )

e.g.

      pBlkDev = ramDevCreate( address , 512 , 400 , 400 , 0 );
      pVolDesc= dosFsDevInit("DEV1:",pBlkDev,NULL )
      if ( pVolDesc == NULL )
      {
              pVolDesc = dosFsMkFs("DEV1:",pBlkDev);
      }

does not work - dosFsDevInit() fails with an address exception ( attempting
to perform a bcopy(...) ).

p.s. we are talking dosfs 1.0 and vxworks 5.3.1 here - old stuff so there
must be a guru for it by now !.


David Anderson.  mailto:danderson@dtrack.demon.co.uk




Thanks and Regards,

V.Suraj
Future Software
0091 - 44 - 4330550 x 302




From vxwexplo-errs@csg.lbl.gov  Sat Aug 19 04:03:07 2000
From: Vxworks Exploder 
Date: Sat Aug 19 04:03:10 PDT 2000
Subject: comp.os.vxworks newsdigest

Comp.Os.Vxworks Daily Digest    Sat Aug 19 04:03:04 PDT 2000

        Subject: Task Suspend
        Subject: MPC860 and real time clock
        Subject: Re: Hello !!
        Subject: Re: Problem creating files on host
        Subject: T2 dosFs2 file date/time problem
        Subject: Pentium SIMD instructions
        Subject: Re: Bootable VxWorks Image problem
        Subject: Re: RTOS Comparison
        Subject: Re: Adventnet SNMP agent for vxworks
        Subject: Re: Fundamental Questions on VxWorks!
        Subject: Re: using fopen crashes my program
        Subject: Re: IVR consultants/contract in So CA
        Subject: Protocols in VxWorks!
        Subject: Re: T2 dosFs2 file date/time problem
        Subject: Re: vxWorks taskStatusString
        Subject: Re: Real time clock
        Subject: using fopen crashes my program
        Subject: available driver
        Subject: 2700 univ implementation of location monitor?
        Subject: Error  on building vxWorks.res_rom.hex!
        Subject: Re:Bootable VxWorks Image problem
        Subject: FTP Problem with Solaris and Linux FTP Servers
        Subject: Re: Watchdog interrupt level
        Subject: mv2700 - dram behind a16 space
        Subject: muxBind() and simulator
        Subject: Re: mv2700 - dram behind a16 space
        Subject: lost bytes on rs232
        Subject: how to enable floating point in C/C++
        Subject: Re: Error  on building vxWorks.res_rom.hex!

-------------------------------------------------------

Newsgroups: comp.os.vxworks
Subject: Task Suspend
Date: Fri, 18 Aug 2000 16:39:54 +0530
From: "Chandrashekar Nagaraj" 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <8nj5rf$c70$1@overload.lbl.gov>

Hi,
    I am hereby giving some cases for task suspension in vxWorks,

Are there are any cases other than these ?..........

1. vxWorks calls
   a. ts()
   b. taskSuspend()
   c. sigsuspend()
   d. pause()
   e. nanosleep()
   f. aio_suspend
   g. sigtimedwait()
   h. sigwaitinfo()

2. Task level exception

3. Breakpoint in task

4. Attaching task to Debugger

5. CS_MSEC_DELAY(msec), a vxWorks Macro

6. Error in Allocating/Freeing Memory

7. When a pipe is full and task attempts to write to a pipe

8. One task stack overflows and corrupts other tasks stack

9. ioctl() call causes non-recoverable error

10. If an invalid device descriptor in the valid range of device
descriptors
    numbers is passed to an I/O function

11. Some errors caused in the system even though the task itself
    may not have caused that error such as
       a. Device driver doing a hardware access that generates an error
       b. Bus error caused by a DMA cycle which was set up by a totally
different task
       c. Bus errors and access violations due to task level code,
driver code, interrupt handlers


Thanks,
KK...


---------------------------

Newsgroups: comp.os.vxworks
Subject: MPC860 and real time clock
Date: Fri, 18 Aug 2000 13:27:45 +0200
From: "Achim Zimmer" 
Organization: Colt Telecom GmbH
Message-ID: <8nj6ib$390$1@crusher.de.colt.net>

we have connected an external oszillator to the MPC
and have "configured" the RTC so that we can read the SecondsCounter
(register RTC). This works fine.

Our problem is, if we disconnect the power supply, the value in this counter
is still present, but in the time with no power supply (only the battery) it
didnt continuing counting.

Please any hints for configuring registers or so ????

Thanks Achim



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Hello !!
Date: Fri, 18 Aug 2000 12:53:36 +0100
From: "Dave Korn" 
Organization: (Posted via) U-NET Internet Ltd.
Message-ID: 
References: <399C7238.78E7F4A3@lucent.com>

Sunil Joseph wrote in message <399C7238.78E7F4A3@lucent.com>...
>Gopal,
> This is a test mail, I have been sending this for sometime now. Always get
an
>error.
>Sunil

  The error is so serious that your mail has ended up in a newsgroup!


     hth,
        DaveK
- --
"Reality is whatever doesn't go away after you stop believing in it."
                                   -- Philip K. Dick





---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Problem creating files on host
Date: Fri, 18 Aug 2000 12:16:21 GMT
From: lvwhitebir@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8nj9e9$k44$1@nnrp1.deja.com>
References: <8nikek$t6t$1@nnrp1.deja.com>

In article <8nikek$t6t$1@nnrp1.deja.com>,
  gardhab  wrote:
> hi all ,
>
> i am trying to read a script file "test.txt" on the host (which is win
> NT) .
> i use open("host:G:/test.txt",0x2,0644).
> this command returns a valid fd, the test file is opened,i'm able to
> read the data in the file and parse the commands within.
>
> but when i need to write the results of the test into a new file i
call
> open ( "host:G:/results.txt", 0x2|0x0200 , 0644)  ( with the O_CREAT
> option)
> this command returns ok with an fd , but there is no file created.
> also using the fd , i'm able to write data to this file ( wherever it
> is) , but when i read the data back , it is just junk.
> i tried using the create call but this too gives the same result .
>
> could someone tell me what i'm doing wrong here and how i can fix this
> problem ?
> thanks in advance  ,
> karthik
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.
>

If you can post some code that shows the problem you are having
we might get a better understanding of your problem.  One quick
suggestion for readability is to use the constants in fcntl.h
for open() such as O_CREAT.

One question, when you check to see if the file was created, do
you close() it first?  NT may be buffering the file in memory and
not actually writing it to disk until the close() (don't know NT
too well to tell you for sure).

Larry


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: T2 dosFs2 file date/time problem
Date: Wed, 16 Aug 2000 12:24:54 -0500
From: Bob Irwin 
Organization: SIS Development
Message-ID: <399ACE66.5689DDCB@systemsinterface.com>

I've created an 18 GB dosFs2 scsi drive /sd1/.  I find that when I copy
data to the disk with utility commands like cp, copy, etc  the file
time/date is correct.

however, when I creat(), write(), and close() a file from
my app, the time/date is bogus.  The utilities in usrFsLib.c use the
same calls.  any ideas what might be going on?

Any help is welcome.

console capture:

vxworks-> ll "/sd1/data"

Listing Directory /sd1/data:
drwxrwxrwx  1 0       0            16384 Nov 22  2002 ./
drwxrwxrwx  1 0       0            16384 Aug 15 17:15 ../
- -rwxrwxrwx  1 0       0        104857600 Nov 22  2002 testdata.000
- -rwxrwxrwx  1 0       0        104857600 Nov 22  2002 testdata.001
- -rwxrwxrwx  1 0       0           121011 Aug 15 17:09 dsmain

vxworks-> creat "/sd1/data/fubar", 1
value = 6 = 0x6
vxworks-> close 6
value = 0 = 0x0
vxworks-> ll "/sd1/data"

Listing Directory /sd1/data:
drwxrwxrwx  1 0       0            16384 Aug 15 17:22 ./
drwxrwxrwx  1 0       0            16384 Aug 15 17:15 ../
- -rwxrwxrwx  1 0       0        104857600 Nov 22  2002 testdata.000
- -rwxrwxrwx  1 0       0        104857600 Nov 22  2002 testdata.001
- -rwxrwxrwx  1 0       0           121011 Aug 15 17:09 dsmain
- -rwxrwxrwx  1 0       0                0 Aug 15 17:22 fubar


Thanks,
   bob

- -- 
Bob Irwin | bobi@sis-dev.com | SIS Development | 281-484-6661 | Houston

---------------------------

Newsgroups: comp.os.vxworks
Subject: Pentium SIMD instructions
Date: Fri, 18 Aug 2000 13:26:05 GMT
From: lourens@mecalc.co.za (Lourens Geldenhuys)
Organization: Mecalc (Pty) Ltd
Message-ID: <399d3906.864921651@modem.mecalc>
Reply-To: lourens@mecalc.co.za

Does anyone know if the GNU development tools that ship with Tornado
2.0 support the SIMD (Single Instruction Multiple Data - I believe)
instructions?

Regards,
  Lourens


================================================================
Lourens Geldenhuys                Tel:        +27 (0)12 665 1480
Project Engineer                  Fax:        +27 (0)12 665 1495
Mecalc (Pty) Ltd                  e-mail:   lourens@mecalc.co.za
86 Oak Avenue, Highveld Technopark, Centurion, South Africa
================================================================

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Bootable VxWorks Image problem
Date: Fri, 18 Aug 2000 11:51:48 +0530
From: slkumar 
Organization: Motorola
Message-ID: <399CD5FC.B5CD2A0C@miel.mot.com>
References: <399BED7D.E034873@miel.mot.com> <399BFC5D.3E10238A@nortelnetworks.com>

Hi Sebastien

Thanks for the information.
I built the image with optimization level 2 (as recommended by the SPR).
The target is still not coming up ...

Regards
Shailendra

Sebastien Mallet wrote:

> Hi,
>  The image will not work without modification. The
>   user has to invoke project facility and change the 'default' build
>   specification to use optimization level other than 'none' under the
>   'C/C++ compiler' tab. Please refer to SPR #27046 for more details.
> Sebastien Mallet


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: RTOS Comparison
Date: Fri, 18 Aug 2000 08:55:04 +0200
From: Joern Tietjen 
Organization: =?iso-8859-1?Q?Dr=E4ger?= Medizintechnik GmbH
Message-ID: <399CDDC8.6BF7371F@draeger.com>
References: 

Casey Ballentine schrieb:
> 
> Hello,
> 
> I'm looking for an article or URL that compares several RTOS products.  I've
> found some by searching industry rags, but the articles have all been
> industry specific (i.e. RTOSes in the Automotive Industry).  What I'd like
> is a more general comparison of RTOSes using several benchmarks (ROM and RAM
> use, CPU usage, interrupt latency, etc.)  Does anyone have a good reference?
> 

you may try
http://www.dedicated-systems.com:6001/encyc/market/rtos/eval_roadmap.htm
to find some comparison info. however, not all of it is for free (since
evaluation is NOT performed by RTOS vendors).

anyhow, i suggest that you should take a close look to some non-tech
features when comparing RTOSes:
- - tool chain (completeness, all tools supported by one vendor ?)
- - market share (big figures here mean availability of tools, drivers,
BSPs, etc.)
- - quality of support by vendor (because you _will_ need some advice
[sometime,somewhere])
- - number of supported platforms (host and target)

HTH,

Joern

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Adventnet SNMP agent for vxworks
Date: Fri, 18 Aug 2000 10:06:04 -0400
From: Bhavesh Patel 
Organization: Posted via Supernews, http://www.supernews.com
Message-ID: <399D42CC.23F50FAF@yahoo.com>
References: <3999AE69.AF526782@yahoo.com> <8nemkb$cl5$1@nnrp1.deja.com>

we plan to use both jmx agent and snmp embedded C agent, this is the reason
the architects here choose adventnet snmp toolkit. How are your experiences
with wind river's snmp agent?
Thanks for your response.

bhavesh
randyj5573@my-deja.com wrote:

> Why don't you use Wind River's SNMP agent?
>
> Randy Jones
>
> In article <3999AE69.AF526782@yahoo.com>,
>   Bhavesh Patel  wrote:
> > Hi,
> >
> > Has anybody used adventnet embedded/C snmp agent for vxworks
> platform? I
> > would like to know their experiences.  Adventnet website does not have
> > example/demo for vxworks platform.
> >
> > Thanks for any help.
> >
> > Bhavesh Patel
> >
> >
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Fundamental Questions on VxWorks!
Date: Fri, 18 Aug 2000 18:34:42 +0200
From: "Heiko Elger" 
Organization: LF.net GmbH, Stuttgart, Germany
Message-ID: <8njofb$2k9m$1@news.lf.net>
References: <8ndpuj$a1g$1@nnrp1.deja.com>

Hi
> i) What is the difference between "sp task1" and "task1" from the shell.

another difference is the priority:
"sp task1" --> spawning tasks form the shell - all tasks will have the same
priority 100
"task"       --> starting functions form the shell all functions will have
the prioriy of the shell and I think it is 4

regards

Heiko





---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: using fopen crashes my program
Date: Fri, 18 Aug 2000 17:20:05 GMT
From: lvwhitebir@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8njr7h$9qg$1@nnrp1.deja.com>
References: <399D46FA.BF32B403@yahoo.com>

In article <399D46FA.BF32B403@yahoo.com>,
  Bhavesh Patel  wrote:
> Hi all,
>
> I am trying to use the vxsim for windows to test my program. The code
> for this program was generated on linux and later ftped and build on
> tornado2 for NT using their project facility. I am facing two
problems -
>
> 1) when I download the application to the target simulator, I get
> unresolved symbols for  bind(),socket(),sendto() etc. Since I am using
> full scale version of the target simulator (which supports networking)
> I should not get these errrors. Any clue to why these unresolved
> symbols?
>
> 2)  The generated code logs error/debug/trade msgs to a log file. As
> part of this it makes a call to fopen() with fileName =  "D:
\agent.log".
> However when I  spawn this program it crashes. Further investigation
> revealed that fopen failed . The way I used fopen() is -   fd =
> fopen(fileName, "w").
>

The filename is incorrect.  You should _always_ use a forward slash
instead of a backslash in filenames.  The correct filename should be
"D:/agent.log".  Even DOS supports forward slashes.

HTH, Larry


> Any help is appreciated.
> Thanks.
>
> bhavesh
>
>


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: IVR consultants/contract in So CA
Date: Fri, 18 Aug 2000 12:54:57 +0100
From: "Dave Korn" 
Organization: (Posted via) U-NET Internet Ltd.
Message-ID: 
References: <8nerns$j2e$1@nnrp1.deja.com>

manojkn@my-deja.com wrote in message <8nerns$j2e$1@nnrp1.deja.com>...
>3IVR consultants for troubleshooting/maintenance/implementation. The key
>word here is Voicetek IVR.[..SMACK!..]

    No, the keyword *here* is VxWorks, fuckwit.

     HTH HAND
              DaveK
- --
"Reality is whatever doesn't go away after you stop believing in it."
                                   -- Philip K. Dick



---------------------------

Newsgroups: comp.os.vxworks
Subject: Protocols in VxWorks!
Date: Fri, 18 Aug 2000 13:49:48 -0500
From: "Keith McHugh" 
Organization: Esat Net Customer
Message-ID: <399d3056.0@seat.s3two.ie>
Reply-To: "Keith McHugh" 

What is the use of protocols in the real-time application development? Where
do we use that kind in the applications? Any Clue???




---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: T2 dosFs2 file date/time problem
Date: 18 Aug 2000 12:46:49 GMT
From: johan@borksoft.xs4all.nl
Organization: Borksoft
Message-ID: <8njb7p$2uu$1@news1.xs4all.nl>
References: <399ACE66.5689DDCB@systemsinterface.com>

Bob Irwin  wrote:
: I've created an 18 GB dosFs2 scsi drive /sd1/.  I find that when I copy
: data to the disk with utility commands like cp, copy, etc  the file
: time/date is correct.

: however, when I creat(), write(), and close() a file from
: my app, the time/date is bogus.  The utilities in usrFsLib.c use the
: same calls.  any ideas what might be going on?

Can you explain what the problem exactly is, because when I look at the
output the file created with creat looks to have the right time, while the
other files have a timestamp of 2002.

Groeten,
	Johan

- -- 
     o o o o o o o . . .   ___________________________________
    o      _____           ||    Johan Borkhuis              |
  .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
 >(________|__|_[_________]_|________________________________|
 _/oo OOOOO oo`  ooo   ooo  'o!o!o                      o!o!o`
=== VxWorks FAQ: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html ===

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: vxWorks taskStatusString
Date: Fri, 18 Aug 2000 09:59:23 -0400
From: Charles H. Chapman 
Organization: Dynacs Engineering Company
Message-ID: <8njfiv$qp2@niven.ksc.nasa.gov>
References: <8nir9v$8id$1@overload.lbl.gov>
Reply-To: chc@nasa2.ksc.nasa.gov

Krishna Kishore Reddy Sureddy wrote:
> Hi,
>       In vxWorks Reference Manual (Page Number 2-886)
>      Under "taskStatusString" call its given that
>         String                Meaning
>        SUSPEND        Task is unavailable for execution(but not
> suspended,delayed or pended).
> 
>       What does the "but not suspended" in above sentence mean?

I think it was a typo.  I think they meant to say "but not delayed
or pended".

Chuck

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Real time clock
Date: 18 Aug 2000 10:11:37 -0400
From: Eric L. Schott 
Organization: The Pennsylvania State University, Univeristy Park, PA
Message-ID: 
References: <8nbm79$i4m$1@crusher.de.colt.net>
Sender: els@picard.arl.psu.edu

"Achim Zimmer"  writes:

> has anybody programmed a clock including day, month and year based on RTC ?

This actually depends upon the board you have.

> Is there any free source for calculating date and time from a second counter
> noting leapyears aso ?

What about using "localtime," an ANSI function included in the kernel?

---------------------------

Newsgroups: comp.os.vxworks
Subject: using fopen crashes my program
Date: Fri, 18 Aug 2000 10:23:54 -0400
From: Bhavesh Patel 
Organization: Posted via Supernews, http://www.supernews.com
Message-ID: <399D46FA.BF32B403@yahoo.com>

Hi all,

I am trying to use the vxsim for windows to test my program. The code
for this program was generated on linux and later ftped and build on
tornado2 for NT using their project facility. I am facing two problems -

1) when I download the application to the target simulator, I get
unresolved symbols for  bind(),socket(),sendto() etc. Since I am using
full scale version of the target simulator (which supports networking)
I should not get these errrors. Any clue to why these unresolved
symbols?

2)  The generated code logs error/debug/trade msgs to a log file. As
part of this it makes a call to fopen() with fileName =  "D:\agent.log".
However when I  spawn this program it crashes. Further investigation
revealed that fopen failed . The way I used fopen() is -   fd =
fopen(fileName, "w").

Any help is appreciated.
Thanks.

bhavesh


---------------------------

Newsgroups: comp.os.vxworks
Subject: available driver
Date: Fri, 18 Aug 2000 14:57:16 GMT
From: lei_sun@hotmail.com
Organization: Deja.com - Before you buy.
Message-ID: <8njis0$vcm$1@nnrp1.deja.com>

Hi all:
   Does anyone know how to find the vxWorks device driver? Especialy,
we want to put the HomePNA card working in vxWorks, due to the timing
issue, it would be better if anybody have the driver source code, which
can give me some clue to write it quickly.

Thank you!
Regards
lei


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: 2700 univ implementation of location monitor?
Date: Fri, 18 Aug 2000 11:15:29 -0400
From: Al Johnston 
Organization: MindSpring Enterprises
Message-ID: <399D5311.C523EE6@mindspring.com>

Does anyone have an implementation of a "location monitor" utilizing
the 2700 card's universe hardware?  Apparently the "vme register"
implementation forces one to use a16 space, and that is a problem
for us.

thanks.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Error  on building vxWorks.res_rom.hex!
Date: Fri, 18 Aug 2000 23:42:21 +0800
From: "kvd2" 
Organization: DCI HiNet
Message-ID: <8njm10$355@netnews.hinet.net>

I got an error message in building vxWorks.res_rom.hex.
What I supposed to modify?


- ----------------------------------------------------------------------------
- ---------------------------------
ccppc -BC:\Tornado/host/x86-win32/lib/gcc-lib/ -c -mcpu=860 -ansi -nostdinc 
- -O2 -fvolatile
 -fno-builtin -fno-for-scope -Wall -I/h   -I. -IC:\Tornado\target\config\all
 -IC:\Tornado\
target/h -IC:\Tornado\target/src/config -IC:\Tornado\target/src/drv -DCPU=PP
C860      symT
bl.c
ldppc -o vxWorks.res_rom -X -N  -e _romInit \
    -Ttext 02800100  -Tdata 00010000  romInit_res.o bootInit_res.o symTbl.o
\
    tmp.o  C:\Tornado\target/lib/libPPC860gnuvx.a  ctdt.o
romInit_res.o(.text+0x4ae): undefined reference to `RAM_DST_ADRS'
romInit_res.o(.text+0x4b2): undefined reference to `RAM_DST_ADRS'
make.exe: *** [vxWorks.res_rom] Error 0x1

Done.





---------------------------

Newsgroups: comp.os.vxworks
Subject: Re:Bootable VxWorks Image problem
Date: Fri, 18 Aug 2000 14:35:07 +0200
From: "Ramakrishnan,CN" 
Organization: UUNET-NL (http://www.nl.uu.net)
Message-ID: 
References: <399BED7D.E034873@miel.mot.com> <399BFC5D.3E10238A@nortelnetworks.com> <399CD5FC.B5CD2A0C@miel.mot.com>

Hello,
I have had problems with building the image with optimisation level 2 at
the beginning. So it always best to introduce optimisation when you are
on firmer grounds.
Ramakrishnan.C.N
Peek Traffic B.V.
Basicweg 16,
3821 BR AMERSFOORT 
P.O. Box 2542,
3800 GB AMERSFOORT
The Netherlands 
telephone:  +31 33 454 1777
fax: 	     +31 33 454 1850
Internet: www.peektraffic.nl 
	
e-mail: C.N.Ramakrishnan@peektraffic.nl


> ----------
> From: 	slkumar[SMTP:slkumar@miel.mot.com]
> Posted At: 	vrijdag 18 augustus 2000 8:21
> Posted To: 	vxworks
> Conversation: 	Bootable VxWorks Image problem
> Subject: 	Re: Bootable VxWorks Image problem
> 
> Hi Sebastien
> 
> Thanks for the information.
> I built the image with optimization level 2 (as recommended by the
> SPR).
> The target is still not coming up ...
> 
> Regards
> Shailendra
> 
> Sebastien Mallet wrote:
> 
> > Hi,
> >  The image will not work without modification. The
> >   user has to invoke project facility and change the 'default' build
> >   specification to use optimization level other than 'none' under
> the
> >   'C/C++ compiler' tab. Please refer to SPR #27046 for more details.
> > Sebastien Mallet
> 


---------------------------

Newsgroups: comp.os.vxworks
Subject: FTP Problem with Solaris and Linux FTP Servers
Date: Fri, 18 Aug 2000 14:34:07 +0200
From: "Andrea Bigagli" 
Organization: Simtel
Message-ID: <399d2d49.0@news.dada.it>
Reply-To: "Andrea Bigagli" 

Hi VxWorkers,
we're having the following problem with an i960 target running VxWorks 5.4
booting from a Solaris or Linux host:
after issuing the RETRIEVE command (e.g. at boot or during fread operation
using netDrv) the target hangs waiting a very long time (around 1-2
minutes).
The strange thing is that this doesn't happen if the same long time has
elapsed since the last RETRIEVE command issued.
To clarify: we run a simple program which after the boot reads a file,
updates it (with a counter value to count iterations) and then reboots.
Well, at first iteration the boot happens quickly and then the reading of
the file takes a very long time. Then also all other boots take the same
long time and everything goes this slowly...
If we stop the target for a minute or two and start it again the first boot
runs quickly and so on...
Note that this doesn't happen with TFTP servers under Solaris or Linux and
with FTP servers under WinNT.
Unfortunately we really can't use TFTP and WinNT on the field, so if someone
can point me in the right direction about this issue I'd greatly
appreciate...
Many thanks in advance,

Regards, Andrea Bigagli.

Simtel Group
abigagli@simtel.it



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Watchdog interrupt level
Date: Fri, 18 Aug 2000 09:31:55 +0100
From: molochai 
Message-ID: <4feae8ef49%molochai@vapour-trail.demon.co.uk>
References: <8nikm1$tgc$1@nnrp1.deja.com>

In message <8nikm1$tgc$1@nnrp1.deja.com>
          gardhab  wrote:

> hi , i'd like to know at what interrupt level a routine associated with a
> Watchdog timer executes. supppose after expiration, a WD timer startsup the
> associated routine and during this execution,another WD timer expires. 
> does the first routine get preempted ? is there some way to achieve this?
> or is it necessary that 2nd WD timer routine waits for the 1st routine to
> exit ? is it possible that the 2nd execution does not happen as the
> interrupt is lost ? thanks in advance , karthik
> 

Depends on what you mean by 'startsup the associated routine'. If by this you
mean the processing is carried out by the function called by the watchdog,
then no. Watchdogs functions are called by off back of the kernel timer
interrupt. (see the VxWorks Programmers' Guide).   It therefore follows that
the functions run in interrupt mode, and that the timer interrupt is disabled
for the duration.

If the processing you are intending to perform on the watchdog timer is
taking so long that another timer interrupt can occur (eg. a kernel tick) you
shouldn't be doing it at interrupt level - use high-priority tasks pending on
a semaphore given in the watchdog.

In the other hand you are scheduling tasks from the watchdogs, normal kernel
scheduling rules apply.  Use a counted semaphore if you want to guarentee
getting every watchdog events.


> Sent via Deja.com http://www.deja.com/
> Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: mv2700 - dram behind a16 space
Date: Fri, 18 Aug 2000 15:40:40 -0400
From: Al Johnston 
Organization: MindSpring Enterprises
Message-ID: <399D9138.AEB7675C@mindspring.com>

I am trying to get 64k of cpu0 dram to be mapped to vme16 space.

My cards have 64Meg so I wanted to move the top 64k (0x3ff,0000-
0x3ff,ffff) from a32 to a16.

Most of the mapping is already done in the default (as shipped)
configuration.  All I thought I would have to do is:

- ------------- mv2600.h -----------------

replaced the line
  #define VAL_VSI5_BD     (VAL_VSI1_BS + VME_A32_SLV_SIZE)
with
  #define VAL_VSI5_BD     (VAL_VSI1_BS + VME_A32_SLV_SIZE - 0x00010000)

and add

  /* VME/A16 (0000)0000 - (0000)FFFF => PCI/Mem 03EF,0000 - 03EF,FFFF */
  #define VAL_VSI3_BS     0x00000000
  #define VAL_VSI3_BD     (VAL_VSI3_BS + 0x0000FFFF)
  #define VAL_VSI3_TO     0x03FF0000
  #define VAL_VSI3_CTL    ( VSI_CTL_EN       | VSI_CTL_PREN    | \
                            VSI_CTL_AM_DATA  | VSI_CTL_AM_PGM  | \
                            VSI_CTL_AM_SUPER | VSI_CTL_AM_USER | \
                            VSI_CTL_VAS_A16  | VSI_CTL_LAS_MEM | \
                            VSI_CTL_LD64EN   | VSI_CTL_PWEN



- ---------- universe.c -----------

add the flaged lines in the cpu0 only section

    if (procNum == 0)
        {
        /* setup the A32 window */

        UNIV_OUT_LONG(UNIVERSE_VSI1_BS,  VAL_VSI1_BS);
        UNIV_OUT_LONG(UNIVERSE_VSI1_BD,  VAL_VSI1_BD);
        UNIV_OUT_LONG(UNIVERSE_VSI1_TO,  VAL_VSI1_TO);

        UNIV_OUT_LONG(UNIVERSE_VSI3_BS,  VAL_VSI3_BS); /* flagged */
        UNIV_OUT_LONG(UNIVERSE_VSI3_BD,  VAL_VSI3_BD); /* flagged */
        UNIV_OUT_LONG(UNIVERSE_VSI3_TO,  VAL_VSI3_TO); /* flagged */

        if (pciToVmeDev == UNIVERSE_I)
- ---------- end of changes

This resulted in the ram disappearing from a32 space but it did not
appear in a16 space...

Anybody know what I am missing/doing wrong?

thanks.

---------------------------

Newsgroups: comp.os.vxworks
Subject: muxBind() and simulator
Date: Fri, 18 Aug 2000 15:18:44 -0500
From: Juana Nakfour 
Organization: Motorola CIG
Message-ID: <399D9A24.A8D8F43E@email.mot.com>

Hi everyone,

    Does anyone know if muxBind() works with Vxworks full simulator. I
have been trying to get it to work with the following piece of code. The
cookie returned by muxBind() is NULL, does that mean it is not
working??? I couldn't find any documentation that explains what
muxBind() would return if it failed. Any pointers would be appreciated.
Thanks in advance !!!
__________________________________________
void *gMyMuxCookie;
void *pMyMuxSpare;

STATUS muxStartup()
{

    logFdSet(STD_OUT);
    LOG("myMux starting up...\n");

    gMyMuxCookie = muxBind("dc",0,(FUNCPTR) &myStackRcvRtn,(FUNCPTR)
&myStackShutdownRtn,
                      (FUNCPTR) &myStackTxRestartRtn, &myStackErrorRtn,
                      MUX_PROTO_SNARF, "ANY", pMyMuxSpare);

    return OK;

}
______________________________________________



Juana Nakfour


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: mv2700 - dram behind a16 space
Date: Fri, 18 Aug 2000 18:06:33 -0400
From: Al Johnston 
Organization: MindSpring Enterprises
Message-ID: <399DB369.53DA5EC@mindspring.com>
References: <399D9138.AEB7675C@mindspring.com>

I made an error in the first post... left out CTL, should
have read:

> ---------- universe.c -----------
> 
> add the flaged lines in the cpu0 only section
> 
>     if (procNum == 0)
>         {
>         /* setup the A32 window */
> 
>         UNIV_OUT_LONG(UNIVERSE_VSI1_BS,  VAL_VSI1_BS);
>         UNIV_OUT_LONG(UNIVERSE_VSI1_BD,  VAL_VSI1_BD);
>         UNIV_OUT_LONG(UNIVERSE_VSI1_TO,  VAL_VSI1_TO);
> 
>         UNIV_OUT_LONG(UNIVERSE_VSI3_BS,  VAL_VSI3_BS); /* flagged */
>         UNIV_OUT_LONG(UNIVERSE_VSI3_BD,  VAL_VSI3_BD); /* flagged */
>         UNIV_OUT_LONG(UNIVERSE_VSI3_TO,  VAL_VSI3_TO); /* flagged */
          UNIV_OUT_LONG(UNIVERSE_VSI3_CTL, VAL_VSI3_CTL); /* flagged */
> 
>         if (pciToVmeDev == UNIVERSE_I)
> ---------- end of changes
> 
> This resulted in the ram disappearing from a32 space but it did not
> appear in a16 space...
> 
> Anybody know what I am missing/doing wrong?
> 
> thanks.

---------------------------

Newsgroups: comp.os.vxworks
Subject: lost bytes on rs232
Date: Sat, 19 Aug 2000 00:57:51 -0700
From: noway 
Organization: WebUseNet Corp  http://www.usenetserver.com - Home of the fastest NNTP servers on the Net.
Message-ID: 

I would appreciate any hints on what might be the reason i'm getting
lost rs232 data when I start increasing the demand of the cpu.  I have
the following configuration:

  1. Tornado II
  2. 486DX4-100 Ampro pc104
  3. 16550 chip on the Ampro board
  4. Using the standard vxworks I8250sio.c driver

My highest priority task is a 10 and does nothing but hang on a read
on the rs232 port and does semgive's to lower priority tasks to handle
the messages comming in on that port.  I noticed I was getting lost
bytes from the comm port when the lower priority tasks got VERY busy.
The int3 driver no matter how much loading should never loose bytes
because it has the highiest priority over my tasks.  I also new that
missing bytes were not because I exceeded the drivers buffer.  I
finally went into the interrupt routine in I8250sio.c and the first
line of the routine put a intLock() and at the end of the service
routine put a intUnlock().  This FIXED!!! the problem..... Why?  How
could my task loading screw something up in the driver???

Thanks




---------------------------

Newsgroups: comp.os.vxworks
Subject: how to enable floating point in C/C++
Date: Sat, 19 Aug 2000 02:02:21 GMT
From: mei2468@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8nkpr6$dca$1@nnrp1.deja.com>

I use floating point operation in my C code. It doesn't work. See the
following code:

void test (void)
{
    double d1, d2;
    d1 = 1.0;
    d2 = 2.0;
    d1 +=d2;       //here, I got message: "floating point disabled"
    ...
}

I'm wantering if I need to include some library or something else in
vxWorks?

Thanks

Mei


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Error  on building vxWorks.res_rom.hex!
Date: Fri, 18 Aug 2000 20:19:13 -0700
From: James Chen 
Organization: Globix Corp.
Message-ID: <399DFCB1.AB281D6D@gigabitwireless.com>
References: <8njm10$355@netnews.hinet.net>

This is a multi-part message in MIME format.
- --------------F1DDA1E2BB4C75CFF9F6A6B4
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Your problem is that when you build a ROM resident vxWorks image, the makefile
activated the compile switch "ROM_RESIDENT" which gets the code in bootInit.c to
reference the #define "RESIDENT_DATA" macro.

You could put the line #define RAM_DST_ADRS xxxxxxxx in the header file
config.h. It can be the same as RAM_HIGH_ADRS or RAM_LOW_ADRS depending on where
in the DRAM you want vxWorks to run off.

kvd2 wrote:

> I got an error message in building vxWorks.res_rom.hex.
> What I supposed to modify?
>
> ----------------------------------------------------------------------------
> ---------------------------------
> ccppc -BC:\Tornado/host/x86-win32/lib/gcc-lib/ -c -mcpu=860 -ansi -nostdinc
> -O2 -fvolatile
>  -fno-builtin -fno-for-scope -Wall -I/h   -I. -IC:\Tornado\target\config\all
>  -IC:\Tornado\
> target/h -IC:\Tornado\target/src/config -IC:\Tornado\target/src/drv -DCPU=PP
> C860      symT
> bl.c
> ldppc -o vxWorks.res_rom -X -N  -e _romInit \
>     -Ttext 02800100  -Tdata 00010000  romInit_res.o bootInit_res.o symTbl.o
> \
>     tmp.o  C:\Tornado\target/lib/libPPC860gnuvx.a  ctdt.o
> romInit_res.o(.text+0x4ae): undefined reference to `RAM_DST_ADRS'
> romInit_res.o(.text+0x4b2): undefined reference to `RAM_DST_ADRS'
> make.exe: *** [vxWorks.res_rom] Error 0x1
>
> Done.

- --------------F1DDA1E2BB4C75CFF9F6A6B4
Content-Type: text/x-vcard; charset=us-ascii;
 name="jchen.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Card for James Chen
Content-Disposition: attachment;
 filename="jchen.vcf"

begin:vcard 
n:Chen;James
tel;fax:408.577.0700
tel;work:408.577.7411
x-mozilla-html:FALSE
url:www.gigabitwireless.com
org:Gigabit Wireless, Inc;BTS Software Development
version:2.1
email;internet:jchen@gigabitwireless.com
adr;quoted-printable:;;3099 N. 1st Street,=0D=0A;San Jose;California;95134;
fn:James Chen
end:vcard

- --------------F1DDA1E2BB4C75CFF9F6A6B4--


---------------------------

End of New-News digest
**********************


From vxwexplo-errs@csg.lbl.gov  Sun Aug 20 04:03:06 2000
From: Vxworks Exploder 
Date: Sun Aug 20 04:03:08 PDT 2000
Subject: comp.os.vxworks newsdigest

Comp.Os.Vxworks Daily Digest    Sun Aug 20 04:03:03 PDT 2000

        Subject: RE: comp.os.vxworks newsdigest
        Subject: ifShow gives wrong ethernet address
        Subject: Re: FTP Problem with Solaris and Linux FTP Servers
        Subject: How to "link" some objs into sigle obj

-------------------------------------------------------

Newsgroups: comp.os.vxworks
Subject: RE: comp.os.vxworks newsdigest
Date: Sat, 19 Aug 2000 16:22:07 +0530
From: "V.Suraj" 
Organization: Lawrence Berkeley National Laboratory
Message-ID: 


David,

Try this.

BLK_DEV *pBlkDev;
DOS_VOL_DESC *pVolDesc;
pBlkDev = ramDevCreate (0, 512, 400, 400, 0);
pVolDesc = dosFsMkfs ("DEV1:", pBlkDev);

Once the ramDrive has been created, you can use dosFsDevInit.
dosFsDevInit() must be used only if the file system already exists on the
disk and should not be reinitialized.

Also ensure that the ramAdd is valid. (first parameter in the ramDevCreate
function.). Better yet, make it 0, and let system allocate memory for your
disk.

- --------------------------------

Can anyone point me at the correct way of verifying a ram block device
contains a valid dos volume before calling
dosFsInit(...) which appears to crash with bus exceptions if the BLOCK_DEV
has not previously been formatted.
( the manual actually states that it should return NULL volume
descriptor.... )

e.g.

      pBlkDev = ramDevCreate( address , 512 , 400 , 400 , 0 );
      pVolDesc= dosFsDevInit("DEV1:",pBlkDev,NULL )
      if ( pVolDesc == NULL )
      {
              pVolDesc = dosFsMkFs("DEV1:",pBlkDev);
      }

does not work - dosFsDevInit() fails with an address exception ( attempting
to perform a bcopy(...) ).

p.s. we are talking dosfs 1.0 and vxworks 5.3.1 here - old stuff so there
must be a guru for it by now !.


David Anderson.  mailto:danderson@dtrack.demon.co.uk




Thanks and Regards,

V.Suraj
Future Software
0091 - 44 - 4330550 x 302



---------------------------

Newsgroups: comp.os.vxworks
Subject: ifShow gives wrong ethernet address
Date: Sat, 19 Aug 2000 18:24:03 +0530
From: slkumar 
Organization: Motorola
Message-ID: <399E836B.827CB7DB@miel.mot.com>

Our target is an MCPN750 board and when we run ifShow on the target it
provides incorrect ethernet address for the device "dc0" !!


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: FTP Problem with Solaris and Linux FTP Servers
Date: Sat, 19 Aug 2000 10:18:23 -0500
From: Bob Irwin 
Organization: SIS Development
Message-ID: <399EA53F.21859C9E@systemsinterface.com>
References: <399d2d49.0@news.dada.it>

Andrea Bigagli wrote:
> 
> Hi VxWorkers,
> we're having the following problem with an i960 target running VxWorks 5.4
> booting from a Solaris or Linux host:
> after issuing the RETRIEVE command (e.g. at boot or during fread operation
> using netDrv) the target hangs waiting a very long time (around 1-2
> minutes).
> abigagli@simtel.it
 
In Solaris refer to the 'ndd' man page. for sol2.6, the param is
tcp_close_wait_interval. it can be shortened  with:

      /usr/sbin/ndd -set /dev/tcp/ tcp_close_wait_interval 1000

In sol2.7 the param is tcp_time_wait_interval & can be shortened with

      /usr/sbin/ndd -set /dev/tcp/ tcp_time_wait_interval 1000

I put a script in /etc/rc2.d to automate this on each reboot:

#!/sbin/sh
#   for Solaris 2.7
# refer to ndd man page

case "$1" in
'start')
   if [ -f /usr/sbin/ndd  ]; then
      echo 'shortening tcp_close_wait_interval to 1000'
      /usr/sbin/ndd -set /dev/tcp/ tcp_close_wait_interval 1000
   fi
   ;;

'stop')
   ;;

*)
        echo "Usage: $0 { start | stop }"
        exit 1
        ;;
esac
exit 0


- -- 
Bob Irwin | bobi@sis-dev.com | SIS Development | 281-484-6661 | Houston

---------------------------

Newsgroups: comp.os.vxworks
Subject: How to "link" some objs into sigle obj
Date: Sun, 20 Aug 2000 15:40:35 +0800
From: "kvd2" 
Organization: DCI HiNet
Message-ID: <8no2d3$p7k@netnews.hinet.net>

Dear all,

Anyone knows how to "link", or some process else, some obj file into single
object? My complier suit is diab data.

Thanks!







---------------------------

End of New-News digest
**********************


From vxwexplo-errs@csg.lbl.gov  Mon Aug 21 02:08:23 2000
From: David Anderson 
Date: Mon Aug 21 02:08:25 PDT 2000
Subject: RE: comp.os.vxworks newsdigest

> Submitted-by: "V.Suraj" 
>
>
> David,
>
> Try this.
>
> BLK_DEV *pBlkDev;
> DOS_VOL_DESC *pVolDesc;
> pBlkDev = ramDevCreate (0, 512, 400, 400, 0);
> pVolDesc = dosFsMkfs ("DEV1:", pBlkDev);
>
> Once the ramDrive has been created, you can use dosFsDevInit.

So if I power down the system and then power up I can use dosFsDevInit for
the second and subsequent cycles ( providing the volume has not been
corrupted ).

> dosFsDevInit() must be used only if the file system already exists on the
> disk and should not be reinitialized.

That's what I said ... how do I check that it is valid BEFORE calling 
dosFsInit !

>
> Also ensure that the ramAdd is valid. (first parameter in the 
ramDevCreate
> function.). Better yet, make it 0, and let system allocate memory for 
your
> disk.

Humm... If I use zero then the volume is allocated from malloc() memory 
space. So
how does this ensure that I can re-launch the same volume after a power 
fail/power-up.

I use a definite address because the volume resides in non-volatile ram. 
This means
it should retain all data through a power up/down.

What I really need is a vxWorks equivalent of 'scandisk' for block devices. 
This seems
quite complicated so I now add a CRC to the memory allocated to the disk 
and re-check
this on power-up. I assume the volume is corrupt if it does not match.

In psuedo-psuedo code:-
----------------------------------
If ( disk_corrupt ) use dosFsMkFs() /* Volume is corrupt */

else use dosFsDevInit()   /* volume is valid */
---------------------------------

Regards & thanks

David.

>
> --------------------------------
>
> Can anyone point me at the correct way of verifying a ram block device
> contains a valid dos volume before calling
> dosFsInit(...) which appears to crash with bus exceptions if the 
BLOCK_DEV
> has not previously been formatted.
> ( the manual actually states that it should return NULL volume
> descriptor.... )
>
> e.g.
>
>       pBlkDev = ramDevCreate( address , 512 , 400 , 400 , 0 );
>       pVolDesc= dosFsDevInit("DEV1:",pBlkDev,NULL )
>       if ( pVolDesc == NULL )
>       {
>               pVolDesc = dosFsMkFs("DEV1:",pBlkDev);
>       }
>
> does not work - dosFsDevInit() fails with an address exception ( 
attempting
> to perform a bcopy(...) ).
>
> p.s. we are talking dosfs 1.0 and vxworks 5.3.1 here - old stuff so there
> must be a guru for it by now !.
>
>
> David Anderson.  mailto:danderson@dtrack.demon.co.uk
>
>
>
>
> Thanks and Regards,
>
> V.Suraj


David Anderson.  mailto:danderson@dtrack.demon.co.uk

-----------------------------------------------------
Who said "There's no life here in outer space."
-----------------------------------------------------

Disclaimer
----------
This email and any files transmitted with it are confidential and intended 
solely for the use of the individual or entity to whom they are addressed.

If you have received this email in error please destroy and notify Data 
Track Technology Plc +44 1425 271900.
-------------------------------------------------------




From vxwexplo-errs@csg.lbl.gov  Mon Aug 21 04:03:06 2000
From: Vxworks Exploder 
Date: Mon Aug 21 04:03:09 PDT 2000
Subject: comp.os.vxworks newsdigest

Comp.Os.Vxworks Daily Digest    Mon Aug 21 04:03:03 PDT 2000

        Subject: Re: How to "link" some objs into sigle obj
        Subject: Re: ifShow gives wrong ethernet address
        Subject: Re: Protocols in VxWorks!
        Subject: Re: Real time clock
        Subject: Performance measurement tools & facilities
        Subject: RE: comp.os.vxworks newsdigest

-------------------------------------------------------

Newsgroups: comp.os.vxworks
Subject: Re: How to "link" some objs into sigle obj
Date: 20 Aug 2000 09:11:48 -0700
From: Matt Wette 
Organization: Jet Propulsion Laboratory
Message-ID: <7kvgwvx67v.fsf@mr-ed.jpl.nasa.gov>
References: <8no2d3$p7k@netnews.hinet.net>

"kvd2"  writes:
> Anyone knows how to "link", or some process else, some obj file into single
> object? My complier suit is diab data.

I think this is mostly a object-format issue.  My guess is that if
you use the GCC "ld" that comes w/ VxWorks that should work.  Also,
the Diab ld might work the same way.   The following example assumes
that you are using PowerPC on VxWorks on Unix host.  It combines
the three objects obj1.o obj2.o obj3.o into the single object allobjs.o.

  $ ldpcc -r -o allobjs.o obj1.o obj2.o obj3.o

Matt
- -- 
Matthew.R.Wette at jpl.nasa.gov -- I speak for myself, not for JPL.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: ifShow gives wrong ethernet address
Date: Sun, 20 Aug 2000 20:39:15 GMT
From: "telia" 
Organization: Enitel Internet Public Access
Message-ID: 
References: <399E836B.827CB7DB@miel.mot.com>

I have seen this for  TII / x86 pentium also
IfShow displays a pointer (can't remember the name ) and the 2 first octets
of the e-address

slkumar  wrote in message
news:399E836B.827CB7DB@miel.mot.com...
> Our target is an MCPN750 board and when we run ifShow on the target it
> provides incorrect ethernet address for the device "dc0" !!
>



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Protocols in VxWorks!
Date: Sun, 20 Aug 2000 20:44:41 GMT
From: "telia" 
Organization: Enitel Internet Public Access
Message-ID: 
References: <399d3056.0@seat.s3two.ie>

A protocol is a set of of rules that describes the format and meaning of
information exchanged between { programs | humans | computers | companies |
.....  }. If your application have any kind of input and/or output, the
information must be according to some protocol in order to be generated or
interpreted by an external entity.



Keith McHugh  wrote in message
news:399d3056.0@seat.s3two.ie...
> What is the use of protocols in the real-time application development?
Where
> do we use that kind in the applications? Any Clue???
>
>
>



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Real time clock
Date: Mon, 21 Aug 2000 08:28:11 +0200
From: "Achim Zimmer" 
Organization: Colt Telecom GmbH
Message-ID: <8nqi4k$dcv$1@crusher.de.colt.net>
References: <8nbm79$i4m$1@crusher.de.colt.net> 

Yes, that is it.


Eric L. Schott  schrieb in im Newsbeitrag:
woitsyoe06.fsf@picard.arl.psu.edu...
> "Achim Zimmer"  writes:
>
> > has anybody programmed a clock including day, month and year based on
RTC ?
>
> This actually depends upon the board you have.
>
> > Is there any free source for calculating date and time from a second
counter
> > noting leapyears aso ?
>
> What about using "localtime," an ANSI function included in the kernel?



---------------------------

Newsgroups: comp.os.vxworks
Subject: Performance measurement tools & facilities
Date: Mon, 21 Aug 2000 10:53:54 +0200
From: Carsten Boeckmann 
Organization: Ericsson Eurolab Deutschland GmbH
Message-ID: <39A0EE22.E43AFA56@eed.ericsson.se>

Hi, 

we are looking for the best possible method on how to measure the
performance of specific functions within our software to identify
bottlenecks.
We are interested in experiences with ProfileScope (Tornado Dev. Tools)
and in any other methods or VxWorks provided facilities of how to
measure a software functions' performance at highest accuracy. 

Looking forward for your comments.

Best regards,
Carsten 

- -- 
 ____________________________________________
   -= Carsten.Boeckmann@eede.ericsson.se =-

---------------------------

Newsgroups: comp.os.vxworks
Subject: RE: comp.os.vxworks newsdigest
Date: Mon, 21 Aug 2000 09:56:39 +0100
From: David Anderson 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <8nqruc$1fq$1@overload.lbl.gov>

> Submitted-by: "V.Suraj" 
>
>
> David,
>
> Try this.
>
> BLK_DEV *pBlkDev;
> DOS_VOL_DESC *pVolDesc;
> pBlkDev = ramDevCreate (0, 512, 400, 400, 0);
> pVolDesc = dosFsMkfs ("DEV1:", pBlkDev);
>
> Once the ramDrive has been created, you can use dosFsDevInit.

So if I power down the system and then power up I can use dosFsDevInit for
the second and subsequent cycles ( providing the volume has not been
corrupted ).

> dosFsDevInit() must be used only if the file system already exists on the
> disk and should not be reinitialized.

That's what I said ... how do I check that it is valid BEFORE calling 
dosFsInit !

>
> Also ensure that the ramAdd is valid. (first parameter in the 
ramDevCreate
> function.). Better yet, make it 0, and let system allocate memory for 
your
> disk.

Humm... If I use zero then the volume is allocated from malloc() memory 
space. So
how does this ensure that I can re-launch the same volume after a power 
fail/power-up.

I use a definite address because the volume resides in non-volatile ram. 
This means
it should retain all data through a power up/down.

What I really need is a vxWorks equivalent of 'scandisk' for block devices. 
This seems
quite complicated so I now add a CRC to the memory allocated to the disk 
and re-check
this on power-up. I assume the volume is corrupt if it does not match.

In psuedo-psuedo code:-
- ----------------------------------
If ( disk_corrupt ) use dosFsMkFs() /* Volume is corrupt */

else use dosFsDevInit()   /* volume is valid */
- ---------------------------------

Regards & thanks

David.

>
> --------------------------------
>
> Can anyone point me at the correct way of verifying a ram block device
> contains a valid dos volume before calling
> dosFsInit(...) which appears to crash with bus exceptions if the 
BLOCK_DEV
> has not previously been formatted.
> ( the manual actually states that it should return NULL volume
> descriptor.... )
>
> e.g.
>
>       pBlkDev = ramDevCreate( address , 512 , 400 , 400 , 0 );
>       pVolDesc= dosFsDevInit("DEV1:",pBlkDev,NULL )
>       if ( pVolDesc == NULL )
>       {
>               pVolDesc = dosFsMkFs("DEV1:",pBlkDev);
>       }
>
> does not work - dosFsDevInit() fails with an address exception ( 
attempting
> to perform a bcopy(...) ).
>
> p.s. we are talking dosfs 1.0 and vxworks 5.3.1 here - old stuff so there
> must be a guru for it by now !.
>
>
> David Anderson.  mailto:danderson@dtrack.demon.co.uk
>
>
>
>
> Thanks and Regards,
>
> V.Suraj


David Anderson.  mailto:danderson@dtrack.demon.co.uk

- -----------------------------------------------------
Who said "There's no life here in outer space."
- -----------------------------------------------------

Disclaimer
- ----------
This email and any files transmitted with it are confidential and intended 
solely for the use of the individual or entity to whom they are addressed.

If you have received this email in error please destroy and notify Data 
Track Technology Plc +44 1425 271900.
- -------------------------------------------------------



---------------------------

End of New-News digest
**********************


From vxwexplo-errs@csg.lbl.gov  Mon Aug 21 08:49:25 2000
From: David Laight 
Date: Mon Aug 21 08:49:28 PDT 2000
Subject: Re: Performance measurement tools & facilities

Carsten,

I had a similar problem - trying to find why the vxWorks 2.0.1 UGL stack is so 
much slower than any of its predecessors.

WRS only bundle the 'browser' - which only tells you which task is busy - not 
which code.

Much the same scheme - only sampling the program counter instead of the taskId - 
can be used to generate a histogram of which routine is busy.  It is more useful 
to have a histogram bucket for each code symbol (rather than linearly spaced 
buckets).

Such code isn't hard to implement, I use the FIRQ interrupt on my ARM system 
(not used by vxWorks) to generate about 14000 interrupts a second and search the 
address/count histogram using a binary chop (ensure there is a 0 one end and 
0xffffffff the other and you can fall off).  Even with 15000 code symbols the 
asm ISR doesn't take long.

Of course you do need a test that runs long enough to take the samples!

Dumping the table every (say) 10 seconds and a program to format the output (add 
the symbol names and ignore infrequently used code and you end up with output 
like:

   total      1      2      3      4 ===columns===
     354    142    106    106    106 ===time===
     310    120     90    100    109 _best_memPartAlignedAlloc
    2691    990    817    884    865 _colour_find
     158     56     49     53     47 _HashToKey
     365    128    126    111    107 _invokeJavaMethod
    5174    759   2187   2228   1707 _ExecuteJava
     296    105     97     94     98 _invokers_adcon1
     187     65     60     62     82 _pinObj
     317    130     87    100    113 _jni_Invoke
     192     73     59     60     51 _createMonitor
    2955   1094    936    925    997
    			_Java_sun_awt_image_GifImageDecoder_parseImage
    2954   1135    925    894   1013
    			_Java_sun_awt_image_ImageRepresentation_setBytePixels
     308    122    108     78    104 _jhpiInvokeNative
     198     75     57     66     68 _jhpiMonitorEnter
     456    335     66     55    149 _memmove
   16693   5564   5581   5548   5612 idleloop
     324    113    106    105    107 _semMGive
     379    163    113    103    102 _semMTake
     758    283    245    230    256 _uglGeneric8BitBitmapWrite
     906    311    312    283    324 _uglGeneric8BitBitmapWriteMono
     480    172    151    157    157 _uglGeneric8BitBlitMonoToColor
     166     69     46     51     53 _uglGenericClipBitmapBlit
   41129  13685  13721  13723  13714 ===total===
   
Which clearly shows where to put any effort.
(I've done considerable work to get the above!)

Given you are typically looking for loops, another scheme that works well with 
micros that have no I-cache is to monitor the data bus with a logic analiser 
saving code fetches (or code fetches where the low address bits are fixed to 
increase the length of a sample).   Run in continuous mode displaying the output 
as a chart (address on Y, samples on X).  Any code loops appear as strong 
horizontal lines.

	David


> Subject: Performance measurement tools & facilities
> Date: Mon, 21 Aug 2000 10:53:54 +0200
> From: Carsten Boeckmann 
> Organization: Ericsson Eurolab Deutschland GmbH
> Message-ID: <39A0EE22.E43AFA56@eed.ericsson.se>
> 
> Hi, 
> 
> we are looking for the best possible method on how to measure the
> performance of specific functions within our software to identify
> bottlenecks.
> We are interested in experiences with ProfileScope (Tornado Dev. Tools)
> and in any other methods or VxWorks provided facilities of how to
> measure a software functions' performance at highest accuracy. 
> 
> Looking forward for your comments.
> 
> Best regards,
> Carsten 
> 
> 


----------------------------------------------------------------
David Laight                          email: dsl@tadpole.co.uk
Tadpole Technology plc                phone: +44 1223 428 232
Cambridge, UK                         fax:   +44 1223 428 201



From vxwexplo-errs@csg.lbl.gov  Mon Aug 21 16:55:56 2000
From: "Garner, William F (N-SDC)" 
Date: Mon Aug 21 16:55:58 PDT 2000
Subject: VxSim and a PCI Card

VxWorks/Tornado users-

Has anyone tried to communicate with a PCI Card using VxSim?

Putting "INCLUDE_PCI" in the BSP does not add any functions (like
pciConfigInLong).

Also, tried to include PCI with Tornado.  Still no functions.

Then tried to include MMU with Tornado and got the following error:


   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 `winIdeNotify':
   C:\Tornado\target\config\comps\src\usrWdbCore.c:76: 
      undefined reference to `sysPhysMemDesc 
   NumEnt'
   C:\Tornado\target\config\comps\src\usrWdbCore.c:76:
       undefined reference to `sysPhysMemDesc 
   '
   make: *** [vxWorks] Error 0x1

How can you look at memory outside of the system?

Thanks for any help,
william.f.garner@lmco.com



From vxwexplo-errs@csg.lbl.gov  Tue Aug 22 04:03:08 2000
From: Vxworks Exploder 
Date: Tue Aug 22 04:03:10 PDT 2000
Subject: comp.os.vxworks newsdigest

Comp.Os.Vxworks Daily Digest    Tue Aug 22 04:03:04 PDT 2000

        Subject: PCI Standard Under VxWorks
        Subject: Re: SCSI driver for sym53C860
        Subject: Linker placement problem?
        Subject: Re: how to enable floating point in C/C++
        Subject: Re: ifShow gives wrong ethernet address
        Subject: Re: T2 dosFs2 file date/time problem
        Subject: Re: comp.os.vxworks newsdigest
        Subject: Re: 1c Need Developers
        Subject: Re: DosFS2.0 and D drives
        Subject: re: Performance measurement tools & facilities
        Subject: Re: Adventnet SNMP agent for vxworks
        Subject: Download location
        Subject: Floating point exception trap
        Subject: customize tornado function keys?
        Subject: Re: PCI Standard Under VxWorks
        Subject: Re: Tornado 2.0 adding componets
        Subject: Re: PCI Standard Under VxWorks
        Subject: Tornado 2.0 adding componets
        Subject: Accessing Packet Header+Data in M_BLK
        Subject: VxSim and a PCI Card
        Subject: Re: PCI Standard Under VxWorks
        Subject: C&T 69000
        Subject: Re: WIN2000 SP1 & TORNADO II
        Subject: Re: T2 dosFs2 file date/time problem
        Subject: GDB and Monitoring varible Format
        Subject: touching a file date and time
        Subject: Re: PCI Standard Under VxWorks
        Subject: R: PCI Standard Under VxWorks
        Subject: Re: Watchdog interrupt level
        Subject: Re: Download location
        Subject: Re: Tornado 2.0 adding componets
        Subject: Re: customize tornado function keys?

-------------------------------------------------------

Newsgroups: comp.os.vxworks
Subject: PCI Standard Under VxWorks
Date: Mon, 21 Aug 2000 12:44:05 GMT
From: "Shawn Girard" 
Organization: Excite@Home - The Leader in Broadband
Message-ID: 

Guys, I have some question in reference PCI standard and VxWorks.  I am
currently writing a driver for PCI target device.  I am little bit confused
I how the bars are used and configured.

Questions:
1) Thinking in sections, who has the responsible to configure the Bars
address in the Configuration Header?  Is it the Kernel, Driver, or the
application?
2) I been seeing documentation that talks about memory addresses from
different point of views.  In this case Host and PCI memory.  Could someone
explain this to me?
3) The address that is assign to the BAR location of the configuration
header, which point of view is that, Host or PCI?

Thanks in Advance
Shawn Girard



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: SCSI driver for sym53C860
Date: Mon, 21 Aug 2000 09:55:08 +0300
From: Amit Margalit 
Organization: Conexant Systems
Message-ID: <39A0D24C.4FC14FE4@conexant.com>
References: 

This is a multi-part message in MIME format.
- --------------C13CC0D79427C02370957061
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Hi!

Sorry I can't help. I have had limited experience with the ncr810. I am
interested to hear if you got it working, though. I have a 53c1510 (RAID)
thingy, which I would like to try any kind of driver on.

I'd be grateful if you could let me know if you succeed.

Thanks,

	Amit

Sheela Kashyap wrote:
> 
> Hi,
> 
> We have an x86 board for which I need to get the SCSI driver working. I am
> trying to use the ncr810 driver for the 860. The sysScsi.c file I am using
> is very similar to the one under config\templateX86. I find the memory
> mapped address of the device by reading the PCI config space. I  built the
> bootrom_uncmp binary without the SCSI driver and built the vxWorks image
> with the SCSI driver. When I get to the bootrom prompt, I am able to dump
> the memory mapped space (0xfc102000) for the SCSI, however while trying to
> load the kernel, it page faults when reading the memory mapped space. I have
> verified that the memory mapped address range has been added to the
> sysPhysMemDesc table. I am able to dump the address space for the ethernet
> which is at 0xfc100000 once the kernel boots. I am not sure why the OS is
> not letting me touch the address space at 0xfc102000.
> 
> I have verified that the SCSI hard drive works. I am using Tornado 2.0 and
> vxWorks 5.4. I would appreciate it greatly if somebody could throw some
> light on this problem.
> 
> Thanks in advance,
> Sheela.

- -- 
#include 
- -----
The characteristic property of hallucinogens, to suspend the boundaries
between
the experiencing self and the outer world in an ecstatic, emotional
experience,
makes it posible with their help, and after suitable internal and external
perparation...to evoke a mystical experience according to plan, so to
speak...
I see the true importance of LSD in the possibility of providing materail
aid 
to meditation aimed at the mystical experience of a deeper, comprehensive
reality.  Such a use accords entirely with the essence and working character
of LSD as a sacred drug.
- - Dr. Albert Hoffman, the discoverer of LSD
- --------------C13CC0D79427C02370957061
Content-Type: text/x-vcard; charset=iso-8859-1;
 name="Amit.Margalit.vcf"
Content-Transfer-Encoding: base64
Content-Description: Card for Amit Margalit
Content-Disposition: attachment;
 filename="Amit.Margalit.vcf"

YmVnaW46dmNhcmQgCm46TWFyZ2FsaXQ7QW1pdAp0ZWw7ZmF4OisoOTcyKS05LTk1NzM3MzIK
dGVsO3dvcms6Kyg5NzIpLTktOTUyNDA5Mwp4LW1vemlsbGEtaHRtbDpGQUxTRQpvcmc6Q29u
ZXhhbnQgSXNyYWVsIEx0ZC47TkFECnZlcnNpb246Mi4xCmVtYWlsO2ludGVybmV0OkFtaXQu
TWFyZ2FsaXRAY29uZXhhbnQuY29tCnRpdGxlOkRyaXZlciBMZWFkCmFkcjtxdW90ZWQtcHJp
bnRhYmxlOjs7MTEgR2FsZ2FsZXkgSGFwbGFkYSBzdC49MEQ9MEFQLk8uQm94IDEyNjYwO0hl
cnpsaWE7OzQ2NzMzO0lTUkFFTAp4LW1vemlsbGEtY3B0OjstOTA0MApmbjpBbWl0IE1hcmdh
bGl0CmVuZDp2Y2FyZAo=
- --------------C13CC0D79427C02370957061--


---------------------------

Newsgroups: comp.os.vxworks
Subject: Linker placement problem?
Date: Mon, 21 Aug 2000 15:15:30 +0200
From: Matthias Gottwald 
Organization: Siemens AG
Message-ID: <39A12B72.10211D8A@icn.siemens.de>

Hi folks,

I have a problem concerning the placement of
memory sections in a linker command file.
I use ldppc (for Motorola PowerPC CPU) and
I can not locate a specific region in a memory - what's
the problem:

assembler file (test.s):
...
.text
...
.section .exc_tbl
code to be done
.text

linker command file (lnk_cmd.lk):
MEMORY
{     ...
     flash    : ORIGIN = 0xFF000000, LENGTH = 0x00F00000
    exc_vt  : ORIGIN = 0xFFF00000, LENGTH = 0x00020000
}

SECTIONS
{
    .text 0xFF000000:
    {
         ...
    }> flash

    .exc_tbl 0xFFF00000 :
    {
        romInit              = 0x00000100;
        exceptionMCE  = 0x00000200;
    } > exc_vt

linker-call:

ldppc -X -N -Map linkmap.map  -e romInit  -T lnk_cmd.lk  \
    -o test test.o PPC -b 01000000

Question: Why does this combination not work properly (The exc_tbl-array
is linked to address 0x00000000 - seen with tool objdumpppc)?
Can anybody explain the correct use of linker command file
(GROUP-assignment not supported?)?

Thanks very much

M.Gottwald


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: how to enable floating point in C/C++
Date: Mon, 21 Aug 2000 13:25:11 GMT
From: esan96@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8nraj3$9s5$1@nnrp1.deja.com>
References: <8nkpr6$dca$1@nnrp1.deja.com>

two things:
1) your BSP must be compiled with either hardware or software floating-
point processing enabled.

2) you need to launch your task with VX_FP_TASK option.

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Edward Sánchez
Software Engineer
esanchez@OnBoard-Software.com
(210) 925-7698x3381
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
OnBoard Software, Inc.
www.OnBoard-Software.com
404 Greig St. Bldg. 178
Kelly AFB, TX 78241
FAX: (210) 925-2144

In article <8nkpr6$dca$1@nnrp1.deja.com>,
  mei2468@my-deja.com wrote:
> I use floating point operation in my C code. It doesn't work. See the
> following code:
>
> void test (void)
> {
>     double d1, d2;
>     d1 = 1.0;
>     d2 = 2.0;
>     d1 +=d2;       //here, I got message: "floating point disabled"
>     ...
> }
>
> I'm wantering if I need to include some library or something else in
> vxWorks?
>
> Thanks
>
> Mei
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.
>


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: ifShow gives wrong ethernet address
Date: Mon, 21 Aug 2000 13:40:32 GMT
From: ricerockut@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8nrbgc$aum$1@nnrp1.deja.com>
References: <399E836B.827CB7DB@miel.mot.com>

This was fixed in T2CP2 (Cummulative Patch #2).  You can download it
from windsurf.

In article <399E836B.827CB7DB@miel.mot.com>,
  slkumar  wrote:
> Our target is an MCPN750 board and when we run ifShow on the target it
> provides incorrect ethernet address for the device "dc0" !!
>
>


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: T2 dosFs2 file date/time problem
Date: Mon, 21 Aug 2000 17:21:47 +0200
From: Leonid Rosenboim 
Organization: BitBand Technologies Ltd. http://www.bitband.com
Message-ID: <39A1490B.8B47890E@bitband.com>
References: <399ACE66.5689DDCB@systemsinterface.com>
Reply-To: leonid@bitband.com

Bon Irwin, your name sounds familiar... did I sell you the CD-ROM file
system some years back ?

Regarding your DosFs 2.0 time issue, my best guess is that the application
changes
the VxWorks time of day or the system clock frequency somehow. Your test
case clearly
shows the problem, is not with the creat() or close() calls either.

Leronid

Bob Irwin wrote:

> I've created an 18 GB dosFs2 scsi drive /sd1/.  I find that when I copy
> data to the disk with utility commands like cp, copy, etc  the file
> time/date is correct.
>
> however, when I creat(), write(), and close() a file from
> my app, the time/date is bogus.  The utilities in usrFsLib.c use the
> same calls.  any ideas what might be going on?
>
> Any help is welcome.
>
> console capture:
>
> vxworks-> ll "/sd1/data"
>
> Listing Directory /sd1/data:
> drwxrwxrwx  1 0       0            16384 Nov 22  2002 ./
> drwxrwxrwx  1 0       0            16384 Aug 15 17:15 ../
> -rwxrwxrwx  1 0       0        104857600 Nov 22  2002 testdata.000
> -rwxrwxrwx  1 0       0        104857600 Nov 22  2002 testdata.001
> -rwxrwxrwx  1 0       0           121011 Aug 15 17:09 dsmain
>
> vxworks-> creat "/sd1/data/fubar", 1
> value = 6 = 0x6
> vxworks-> close 6
> value = 0 = 0x0
> vxworks-> ll "/sd1/data"
>
> Listing Directory /sd1/data:
> drwxrwxrwx  1 0       0            16384 Aug 15 17:22 ./
> drwxrwxrwx  1 0       0            16384 Aug 15 17:15 ../
> -rwxrwxrwx  1 0       0        104857600 Nov 22  2002 testdata.000
> -rwxrwxrwx  1 0       0        104857600 Nov 22  2002 testdata.001
> -rwxrwxrwx  1 0       0           121011 Aug 15 17:09 dsmain
> -rwxrwxrwx  1 0       0                0 Aug 15 17:22 fubar
>
> Thanks,
>    bob
>
> --
> Bob Irwin | bobi@sis-dev.com | SIS Development | 281-484-6661 | Houston


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: comp.os.vxworks newsdigest
Date: 21 Aug 2000 14:49:45 GMT
From: johan@borksoft.xs4all.nl
Organization: Borksoft
Message-ID: <8nrfi9$sib$1@news1.xs4all.nl>
References: <8nqruc$1fq$1@overload.lbl.gov>

David Anderson  wrote:
: I use a definite address because the volume resides in non-volatile ram. 
: This means it should retain all data through a power up/down.

What I do is open the file system, and if this open returns error the
device is initialised, otherwise I have to call dosFsVolFormat.
The order I use is:

ramDiskDevCreate
dosFsDevCreate(ramdevice, ...)
open(ramdevice, ...)
if -1 
   dosFsVolFormat
else
   close (...)

Groeten,
	Johan

- -- 
     o o o o o o o . . .   ___________________________________
    o      _____           ||    Johan Borkhuis              |
  .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
 >(________|__|_[_________]_|________________________________|
 _/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: 1c Need Developers
Date: Mon, 21 Aug 2000 09:35:26 -0500
From: "Derek Opitz" 
Organization: Motorola CIG
Message-ID: <8nre8c$8et2@nntp.cig.mot.com>
References: <%xAm5.1315$3T4.258898@sea-read.news.verio.net> <8nen4b$92o1@nntp.cig.mot.com> 

Ah, sorry, didnt realize all 4 were the same post.  Apologies.

Scott A. Vigil  wrote in message
news:ywUm5.35$Z96.4297@sea-read.news.verio.net...
> Sorry for the redundant post.
> News server error messages made it appear my posts were being rejected.
>
> Scott Vigil
> Lead, Embedded Software Team
> Centura Solutions
> 1111 3rd Avenue, twenty-eighth floor
> Seattle, WA 98101
> http://www.centurasolutions.com/solutions/whatwedo.asp
> http://www.centurasoft.com
>
>
>



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: DosFS2.0 and D drives
Date: Mon, 21 Aug 2000 17:13:22 +0200
From: Leonid Rosenboim 
Organization: BitBand Technologies Ltd. http://www.bitband.com
Message-ID: <39A14712.66AC0BA@bitband.com>
References: <399AB8AB.A55FDBCE@loverso.southborough.ma.us>
Reply-To: leonid@bitband.com

You are trying to create "Disk 1" as a Slave on Primary IDE - is this really
the hardware setup ?
or perhaps the second disk is Master on Secondary IDE ?

Also, when a function returns -1 in VxWorks, you should inspect errno for
details on the cause of error.

There is also an optional ataShow() function which you can use to check the
IDE driver configuration.

Leonid

Susan LoVerso wrote:

> I am using DosFS 2.0 and the x86 BSP.  I have 2 disks that are set up as
> follows:
> Disk 0:  6Gb drive
> partition 1 4Gb   Windows98 (active partition) C:
> partition 2 2Gb   Non-DOS partition
>
> Disk 1: 4Gb drive
> partition 1 2Gb  Windows98  D:
> partition 2 1Gb   Non-DOS partition
> (and 1Gb not yet partitioned)
>
> Both of my Windows partitions were created with fdisk, under windows.
> The D: partition
> is "exclusively for VxWorks use".  I can access both C: and D: under
> Windows.  I put
> 2 small text files onto D: under Windows so that I can attempt to see
> them under VxWorks.
>
> When I boot VxWorks, I can see the files that are normally C:, but not
> D:.
> -> devs
> drv name
>   4 /ata0a
>   4 /ata1a
>
> -> @ls "/ata0a"
> [... list shortened for brevity]
> /ata0a/My Documents
> /ata0a/Program Files
> /ata0a/win98
> /ata0a/W9x_552
> /ata0a/SCANDISK.LOG
> value = 0 = 0x0
> -> @ls "/ata1a"
> Can't open "/ata1a".
> value = -1 = 0xffffffff
>
> There are no messages on the console from that 'ls' command.
>
> When that didn't work, I tried the steps in the DosFS 2.0 Release Notes:
>
> ->  < /vxtmp/vxtmp/d1.script
>
> #
> # Create atadevice on drive 1
> pblk1 = ataDevCreate(0,1,0,0)
> new symbol "pblk1" added to symbol table.
> pblk1 = 0x599760: value = 67100180 = 0x3ffde14
> #
> # Set up cache
> pcbio1 = dcacheDevCreate(pblk1, 0, 0x40000, "disk")
> new symbol "pcbio1" added to symbol table.
> pcbio1 = 0x599750: value = 67099916 = 0x3ffdd0c
> #
> # mount partitions
> pcbiop1 = dpartDevCreate(pcbio1, 2, usrFdiskPartRead)
> new symbol "pcbiop1" added to symbol table.
> pcbiop1 = 0x599748: value = 67096384 = 0x3ffcf40
> #
> # create dosfs device on partition 0
> dosFsDevCreate("/disk1",dpartPartGet(pcbiop1,0), 0x100, 0)
> value = 0 = 0x0
> -> devs
> drv name
>   4 /ata0a
>   4 /ata1a
>   4 /disk1
> value = 0 = 0x0
> -> @ls "/disk1"
> Can't open "/disk1".
> value = -1 = 0xffffffff
>
> Again, there are no error messages associated with this ls command
> either.
>
> I also cannot do a 'dosFsShow' on the D drive either:
> -> dosFsShow "/ata0a"
>
> volume descriptor ptr (pVolDesc):       0x3f8c898
> cache block I/O descriptor ptr (pCbio): 0x3f8cb64
> auto disk check on mount:               DOS_CHK_REPAIR | DOS_CHK_VERB_2
> max # of simultaneously open files:     22
> file descriptors in use:                0
> # of different files in use:            0
> # of descriptors for deleted files:     0
> # of  obsolete descriptors:             0
>
> current volume configuration:
>  - volume label:        NO NAME ; (in boot sector:      NO NAME    )
>  - volume Id:           0x1f2707de
>  - total number of sectors:     8,193,087
>  - bytes per sector:            512
>  - # of sectors per cluster:    8
>  - # of reserved sectors:       32
>  - FAT entry size:              FAT32
>  - # of sectors per FAT copy:   7,994
>  - # of FAT table copies:       2
>  - # of hidden sectors:         63
>  - first cluster is in sector # 16,020
>  - directory structure:         VFAT
>  - root dir start cluster:      2
>
> FAT handler information:
> ------------------------
>  - allocation group size:       103 clusters
>  - free space on volume:        3,780,792,320 bytes
> value = 0 = 0x0
> -> dosFsShow "/ata1a"
>
> volume descriptor ptr (pVolDesc):       0x3f6aa78
> cache block I/O descriptor ptr (pCbio): 0x3f6ad44
> auto disk check on mount:               DOS_CHK_REPAIR | DOS_CHK_VERB_2
> max # of simultaneously open files:     22
> file descriptors in use:                0
> # of different files in use:            0
> # of descriptors for deleted files:     0
> # of  obsolete descriptors:             0
>
> can't mount volume
>
> value = 0 = 0x0
>
> --------------------------
> Any suggestions on what I am doing wrong?  The first drive works just
> fine without me
> having to do anything.  They are similarly partitioned and accessible on
> Windows.
> Thanks for any help


---------------------------

Newsgroups: comp.os.vxworks
Subject: re: Performance measurement tools & facilities
Date: Mon, 21 Aug 2000 16:49:37 +0100 (BST)
From: David Laight 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <200008211549.QAA29912@dsl-2.tadpole.co.uk>

Carsten,

I had a similar problem - trying to find why the vxWorks 2.0.1 UGL stack is so 
much slower than any of its predecessors.

WRS only bundle the 'browser' - which only tells you which task is busy - not 
which code.

Much the same scheme - only sampling the program counter instead of the taskId - 
can be used to generate a histogram of which routine is busy.  It is more useful 
to have a histogram bucket for each code symbol (rather than linearly spaced 
buckets).

Such code isn't hard to implement, I use the FIRQ interrupt on my ARM system 
(not used by vxWorks) to generate about 14000 interrupts a second and search the 
address/count histogram using a binary chop (ensure there is a 0 one end and 
0xffffffff the other and you can fall off).  Even with 15000 code symbols the 
asm ISR doesn't take long.

Of course you do need a test that runs long enough to take the samples!

Dumping the table every (say) 10 seconds and a program to format the output (add 
the symbol names and ignore infrequently used code and you end up with output 
like:

   total      1      2      3      4 ===columns===
     354    142    106    106    106 ===time===
     310    120     90    100    109 _best_memPartAlignedAlloc
    2691    990    817    884    865 _colour_find
     158     56     49     53     47 _HashToKey
     365    128    126    111    107 _invokeJavaMethod
    5174    759   2187   2228   1707 _ExecuteJava
     296    105     97     94     98 _invokers_adcon1
     187     65     60     62     82 _pinObj
     317    130     87    100    113 _jni_Invoke
     192     73     59     60     51 _createMonitor
    2955   1094    936    925    997
    			_Java_sun_awt_image_GifImageDecoder_parseImage
    2954   1135    925    894   1013
    			_Java_sun_awt_image_ImageRepresentation_setBytePixels
     308    122    108     78    104 _jhpiInvokeNative
     198     75     57     66     68 _jhpiMonitorEnter
     456    335     66     55    149 _memmove
   16693   5564   5581   5548   5612 idleloop
     324    113    106    105    107 _semMGive
     379    163    113    103    102 _semMTake
     758    283    245    230    256 _uglGeneric8BitBitmapWrite
     906    311    312    283    324 _uglGeneric8BitBitmapWriteMono
     480    172    151    157    157 _uglGeneric8BitBlitMonoToColor
     166     69     46     51     53 _uglGenericClipBitmapBlit
   41129  13685  13721  13723  13714 ===total===
   
Which clearly shows where to put any effort.
(I've done considerable work to get the above!)

Given you are typically looking for loops, another scheme that works well with 
micros that have no I-cache is to monitor the data bus with a logic analiser 
saving code fetches (or code fetches where the low address bits are fixed to 
increase the length of a sample).   Run in continuous mode displaying the output 
as a chart (address on Y, samples on X).  Any code loops appear as strong 
horizontal lines.

	David


> Subject: Performance measurement tools & facilities
> Date: Mon, 21 Aug 2000 10:53:54 +0200
> From: Carsten Boeckmann 
> Organization: Ericsson Eurolab Deutschland GmbH
> Message-ID: <39A0EE22.E43AFA56@eed.ericsson.se>
> 
> Hi, 
> 
> we are looking for the best possible method on how to measure the
> performance of specific functions within our software to identify
> bottlenecks.
> We are interested in experiences with ProfileScope (Tornado Dev. Tools)
> and in any other methods or VxWorks provided facilities of how to
> measure a software functions' performance at highest accuracy. 
> 
> Looking forward for your comments.
> 
> Best regards,
> Carsten 
> 
> 


- ----------------------------------------------------------------
David Laight                          email: dsl@tadpole.co.uk
Tadpole Technology plc                phone: +44 1223 428 232
Cambridge, UK                         fax:   +44 1223 428 201


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Adventnet SNMP agent for vxworks
Date: Mon, 21 Aug 2000 14:51:49 GMT
From: randyj5573@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8nrfls$fti$1@nnrp1.deja.com>
References: <3999AE69.AF526782@yahoo.com> <8nemkb$cl5$1@nnrp1.deja.com> <399D42CC.23F50FAF@yahoo.com>

My experiences with Wind River's SNMP agent has been extreemly
frustrating. I have found that WRS tech support in other area's of
Vxworks to be pretty good but their tech support for their SNMP agent
to be almost non-existant. It seems that WRS's general tech suport has
almost no experience with SNMP. They purchased their current agent from
a third party. WRS has since purchased the company that wrote their
agent but there has yet to be any transfer of knowledge out to the tech
support people. WRS has fairly good general documentation about their
SNMP agent but when it came time to write some code, I came up with
questions on how some of their functions worked, what some of their
internal structures looked like and questions about undocumented
functions and macros that their mib compiler would use its generated
code. I quickly came to the limit of what help their tech support could
provide and found myself teaching them so that maybe they could
research a question to provide an answer. One question on traps took
over three weeks to get an answer. The one thing that helped me was
finding some programming examples on their web site. I was able to
compile them and then step through them in order to figure out how they
work.

Well enough venting. This was my 1st time at trying to extend an SNMP
agent so I have no experience with other tool vendors to compare to.

Regards,
Randy Jones

In article <399D42CC.23F50FAF@yahoo.com>,
  Bhavesh Patel  wrote:
> we plan to use both jmx agent and snmp embedded C agent, this is the
reason
> the architects here choose adventnet snmp toolkit. How are your
experiences
> with wind river's snmp agent?
> Thanks for your response.
>
> bhavesh
> randyj5573@my-deja.com wrote:
>
> > Why don't you use Wind River's SNMP agent?
> >
> > Randy Jones
> >
> > In article <3999AE69.AF526782@yahoo.com>,
> >   Bhavesh Patel  wrote:
> > > Hi,
> > >
> > > Has anybody used adventnet embedded/C snmp agent for vxworks
> > platform? I
> > > would like to know their experiences.  Adventnet website does not
have
> > > example/demo for vxworks platform.
> > >
> > > Thanks for any help.
> > >
> > > Bhavesh Patel
> > >
> > >
> >
> > Sent via Deja.com http://www.deja.com/
> > Before you buy.
>
>


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Download location
Date: Mon, 21 Aug 2000 10:24:28 -0500
From: "Michael S." 
Organization: MediaOne Express -=- Central Region
Message-ID: 

Anyone know where I can download a beta or trial of the tornado developement
kit?

Sincerely,
Michael S.



---------------------------

Newsgroups: comp.os.vxworks
Subject: Floating point exception trap
Date: Mon, 21 Aug 2000 12:02:08 -0400
From: Martin Beaucage 
Message-ID: <39A15280.1573A376@tech-m4.com>

Hi,

I need to replace VxWorks floating point exception interrupt with my
own interrupt routine.  Up to now, every thing I tried doesn't work.

Can any one help me?

Thanks!

        Martin


---------------------------

Newsgroups: comp.os.vxworks
Subject: customize tornado function keys?
Date: Mon, 21 Aug 2000 09:39:41 -0700
From: "Ryan Tanner" 
Organization: Posted via Supernews, http://www.supernews.com
Message-ID: 

Is there a way to customize the accelerator key bindings in the Tornado GUI?
In specific, I'd like to remap the debug key bindings.



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: PCI Standard Under VxWorks
Date: Mon, 21 Aug 2000 18:55:37 +0200
From: Sebastien Fauris 
Organization: Club-Internet (France)
Message-ID: <39A15F09.33370212@club-internet.fr>
References: 

Hi Shawn,

Shawn Girard wrote:

> Guys, I have some question in reference PCI standard and VxWorks.  I am
> currently writing a driver for PCI target device.  I am little bit confused
> I how the bars are used and configured.
>
> Questions:
> 1) Thinking in sections, who has the responsible to configure the Bars
> address in the Configuration Header?  Is it the Kernel, Driver, or the
> application?

Definetly not the application, but 2 cases are possible:
- - your kernel has a 'plug an play configuration mechanism' implemented,
   then you develop a 'Plug and Play' PCI driver and then just use the
   ressources the system allocated to you (the driver reads back the BAR values

   /IRQ and use them).
   This is the way it is on a PC for example (the BIOS does the Plug and Play
   for you). This is the 'proper way' to do the things.

- - your kernel does not implement the 'Plug and Play', you have to set up
the BARs and get your IRQ number yourself in your driver then.

Usually VxWorks does not have a 'Plug and Play' mechanism implemented
(Tornado 1). I heard that it is provided in Tornado 2 but I don't know if it
is true. In my case, we developped ourselves our 'Plug and Play' layer.
Do a pciDeviceShow then a pciHeaderShow(bus#,device#,function#) in a shell
to display the PCI Config Space Header of your board after power-up to
know if its BARs and IRQ were set up or not.

>
> 2) I been seeing documentation that talks about memory addresses from
> different point of views.  In this case Host and PCI memory.  Could someone
> explain this to me?

The address for a register on a PCI board is not seen the same way from
the CPU bus or from the PCI bus. You need to translate the address depending
the place you use it (from CPU or from the PCI controller on the daughter
board).

>
> 3) The address that is assign to the BAR location of the configuration
> header, which point of view is that, Host or PCI?

The address in the BAR is stored as an address seen from the PCI bus.
You need to translate it into a CPU address before to perform an access to
your board from the CPU.

>
> Thanks in Advance
> Shawn Girard

Hope this helps,

Sebastien.


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Tornado 2.0 adding componets
Date: Mon, 21 Aug 2000 15:34:51 -0500
From: Greg Willden 
Organization: Posted via Supernews, http://www.supernews.com
Message-ID: <39A1926B.E3D7FA09@swri.edu>
References: <8nrr41$9b6pr$2@ID-2497.news.cis.dfn.de>

> Does anyone know how I can tell Tornado where it should look
> for my object files so that can be added using the project
> facility and will get linked with the main project?

Go to the Builds tab of the workspace window and right click on the specific
build that you are using and select Properties.  Go to the Macros tab of the
pop-up window and under PRJ_LIBS add the path and filename of any archive
files that you want to link to.  If you want to link in .o files you go to
the linker tab and specify the .o files.

>
>
> Second problem:
> During Tornado 1.0 time, we built and linked the modules using makefiles,
> this created one .O file for each module, which we then linked with the
> main project. Now I don't know the correct way to create these modules.
> If I create them by creating a "downloadable application module"with
> Tornado 2.0, I either have multiple .O files (one for each .C file) or
> an .a archive or an .out file, depending on the link rule. Neither of
> these seem to be the correct one. What can I do?
>

You want to create a downloadable application module that links to a .out
file.  This .out file consists of all the .o's generated in compilation, the
libs specified in the PRJ_LIBS and any other .o's that you specify in the
linker section.

Hope this helps



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: PCI Standard Under VxWorks
Date: Mon, 21 Aug 2000 17:41:03 -0400
From: "Tim Stanka" 
Organization: Eastman Kodak Company
Message-ID: <8ns6r0$g6s$1@news.news.kodak.com>
References: 

Shawn,
I have formulated alot of documentation for using PCI with VXWORKS and the
Motorola MCP750
CPU board. If this would be useful let me know. VXWORKS has PCI autoconfig
to configure your
PCI target board. As the other guy mentioned, there are functions to read
config space info. You can
read your BARS and add the appropriate offset since PCI memory and I/O space
is memory mapped.

Tim

Shawn Girard wrote in message ...
>Guys, I have some question in reference PCI standard and VxWorks.  I am
>currently writing a driver for PCI target device.  I am little bit confused
>I how the bars are used and configured.
>
>Questions:
>1) Thinking in sections, who has the responsible to configure the Bars
>address in the Configuration Header?  Is it the Kernel, Driver, or the
>application?
>2) I been seeing documentation that talks about memory addresses from
>different point of views.  In this case Host and PCI memory.  Could someone
>explain this to me?
>3) The address that is assign to the BAR location of the configuration
>header, which point of view is that, Host or PCI?
>
>Thanks in Advance
>Shawn Girard
>
>



---------------------------

Newsgroups: comp.os.vxworks
Subject: Tornado 2.0 adding componets
Date: Mon, 21 Aug 2000 20:11:19 +0200
From: Marcus Overhagen 
Organization: I'm on my own
Message-ID: <8nrr41$9b6pr$2@ID-2497.news.cis.dfn.de>

Hello,

I would like to add new components to my VxWorks configuration.

We currenlty have a number of sub-projects which we want to be able 
to add or remove to a main project. The most convenient way seems
to be managing them as modules (or components) using the VxWorks project
facility.

To get started, I created a new file in \target\config\coms\vxWorks
which I called 00toolkit.cdf. This file defines a new FOLDER_TOOLKIT
where I want to manage my components.
I also added this FOLDER_TOOLKIT to the FOLDER_ROOT folder in file
00vxWorks.cdf

If I create a new project inside Tornado, the new "Toolkit" folder
does appear, and this seems to be a nice way to manage componets.

But two problems appear.

If I want to include any of the new components, Tornado tells me that:
"Componet XYZ is not installed. You may need to purchase and/or install
it"

It looks like it does not find the neccessary object files, regardless 
where I place them.
I already tried:
\tornado\target\
\tornado\target\lib\objCPU32gnuvx
\tornado\target\config
\tornado\target\config\MY_BSP
\tornado\target\project\MY_PROJECT

All object files of components belonging to Tornado seem to be archived
in the file \tornado\target\lib\libCPU32gnuvx.a

Perhaps I could add my object files using the "AR" tool to this
archive, but this would be a very difficult operation since the modules
sometimes need to be modified and recompiled during the development
process 
of the main application.

Does anyone know how I can tell Tornado where it should look 
for my object files so that can be added using the project
facility and will get linked with the main project?

Second problem:
During Tornado 1.0 time, we built and linked the modules using makefiles,
this created one .O file for each module, which we then linked with the
main project. Now I don't know the correct way to create these modules.
If I create them by creating a "downloadable application module"with 
Tornado 2.0, I either have multiple .O files (one for each .C file) or
an .a archive or an .out file, depending on the link rule. Neither of 
these seem to be the correct one. What can I do?

Below is the .cdf file I created, the .o files listed as MODULES need to
be 
created and linked.

Regards,

Marcus Overhagen

FILE 00toolkit.cdf begin
=======================================================================
Folder	FOLDER_TOOLKIT {
	NAME		toolkit components
	SYNOPSIS	Komponenten des Toolkit
	CHILDREN	INCLUDE_TOOLKIT_SYSTEM_INITIALIZATION \
			INCLUDE_TOOLKIT_SYSTEM_MSGSCHEDULER \
			INCLUDE_TOOLKIT_MODULE_EIB112
	DEFAULTS	INCLUDE_TOOLKIT_SYSTEM_INITIALIZATION \
			INCLUDE_TOOLKIT_SYSTEM_MSGSCHEDULER
}

Component INCLUDE_TOOLKIT_SYSTEM_INITIALIZATION {
	NAME		Toolkit System Initialization 
	SYNOPSIS	Started den Message Scheduler und Protocolstacks
	MODULES		startup.o
}

Component INCLUDE_TOOLKIT_SYSTEM_MSGSCHEDULER {
	NAME		Toolkit System Message Scheduler 
	SYNOPSIS	Verteilt Messages zwischen den Protocolstacks
	REQUIRES	INCLUDE_TOOLKIT_SYSTEM_INITIALIZATION
	MODULES		msgsched.o
}

Component INCLUDE_TOOLKIT_MODULE_EIB112 {
	NAME		EIB BIM 112 Modul
	SYNOPSIS	EIB BIM 112 Protocolstack
	REQUIRES	INCLUDE_TOOLKIT_SYSTEM_MSGSCHEDULER
	MODULES		eib_stack.o
}
=======================================================================
FILE 00toolkit.cdf end

- -- 
If debugging is the process of removing bugs, then
programming must be the process of putting them in.


- -- 
begin 666 \con\con\vcard.vcf
Heute schon abgestürzt?
end

---------------------------

Newsgroups: comp.os.vxworks
Subject: Accessing Packet Header+Data in M_BLK
Date: Mon, 21 Aug 2000 11:38:52 -0700
From: "AG" 
Organization: SBC Internet Services
Message-ID: 

Is there a way to referrence an Ehternet header and/or data portion of the
packet within a M_BLK struct without the extra copying to a temp buf using
"netMblkToBufCopy"?

Thanks.



---------------------------

Newsgroups: comp.os.vxworks
Subject: VxSim and a PCI Card
Date: Mon, 21 Aug 2000 16:58:33 -0700
From: "Garner, William F (N-SDC)" 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <8nsglt$f5o$1@overload.lbl.gov>

VxWorks/Tornado users-

Has anyone tried to communicate with a PCI Card using VxSim?

Putting "INCLUDE_PCI" in the BSP does not add any functions (like
pciConfigInLong).

Also, tried to include PCI with Tornado.  Still no functions.

Then tried to include MMU with Tornado and got the following error:


   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 `winIdeNotify':
   C:\Tornado\target\config\comps\src\usrWdbCore.c:76: 
      undefined reference to `sysPhysMemDesc 
   NumEnt'
   C:\Tornado\target\config\comps\src\usrWdbCore.c:76:
       undefined reference to `sysPhysMemDesc 
   '
   make: *** [vxWorks] Error 0x1

How can you look at memory outside of the system?

Thanks for any help,
william.f.garner@lmco.com


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: PCI Standard Under VxWorks
Date: Mon, 21 Aug 2000 18:53:51 GMT
From: John Fusco 
Organization: Deja.com - Before you buy.
Message-ID: <8nrtrk$1ns$1@nnrp1.deja.com>
References: 

In article ,
  "Shawn Girard"  wrote:
> Guys, I have some question in reference PCI standard and VxWorks.  I
am
> currently writing a driver for PCI target device.  I am little bit
confused
> I how the bars are used and configured.
>
> Questions:
> 1) Thinking in sections, who has the responsible to configure the Bars
> address in the Configuration Header?  Is it the Kernel, Driver, or the
> application?

[SNIP]

FYI,

Tornado II has code for a PCI auto configuration library that is meant
to be called during sysHwInit (i.e. in the kernel).  This provides a
sort of plug-and-play configuration mechanism so that when your driver
gets to the device, you can assume that the BARs have already been
programmed.


SEE: $WIND_BASE/target/src/drv/pci


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: C&T 69000
Date: Mon, 21 Aug 2000 15:57:55 -0400
From: "Mike Kordik" 
Message-ID: <1Qfo5.172$o_3.1254@client>

I have a board that uses the Chips 69000 video controller. I received a
"driver" from WRS which initializes the chip. When I use this file, VxWorks
gives me a page fault when the init routine tries to clear the screen. Doing
"tt" shows that it is trying to clear at address 0xe0000000. I did notice in
the MMU map that something is trying to map at address 0xe2000000. I
manually, in syslib.c, mapped a window at 0xe0000000 and VxWorks now boots
without error. I am also able to run the ugldemo. The problem now is that
the video is offset from the top left corner of the LCD.

What is going on? How do I fix it?

Thanks,
Mike



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: WIN2000 SP1 & TORNADO II
Date: Mon, 21 Aug 2000 19:58:21 GMT
From: fuzzy_duck_54@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8ns1kh$6o2$1@nnrp1.deja.com>
References:  <8mup9n$udn$1@nnrp1.deja.com>

I discovered this too, but only with SP1.  Remove
SP1 and the errors go away.  I have found no
reference to this error in the Microsoft
knowledge base.
To VxWorks dudes: lets get suppport up for
Win2K.  It's time.

In article <8mup9n$udn$1@nnrp1.deja.com>,
  venezim@my-deja.com wrote:
> Hi Michael,
>
> I have no clue what these errors are but I am
> experiencing them as well.  I'm only just
> starting to use Tornado II so I'm glad to hear
> that the program seems to work fine after
> accepting the error messages.
>
> Mark V.
>
> In article  east.usenetserver.com>,
>   "Michael Murphy"  wrote:
> > Installed Windows 2000 SP1 today. When running
> Tornado II I get a bunch of
> > the following errors:
> >
> > "RPCExit: extra call"
> >
> > Tornado II seems to run fine after I accept
all
> these errors (4 to 5 of
> > them)
> >
> > Anybody have any clues?
> >
> > Thanx,
> >
> > Michael
> >
> >
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.
>



Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: T2 dosFs2 file date/time problem
Date: Mon, 21 Aug 2000 20:17:57 GMT
From: "Craig E. Johnston" 
Organization: The Boeing Company
Message-ID: <39A18E75.56378EE8@boeing.com>
References: <399ACE66.5689DDCB@systemsinterface.com>
Sender: nntp@news.boeing.com (Boeing NNTP News Access)

Bob,

I've seen a similar problem with my PPC board.  WRS sent me a patch for
the following SPR that fixed the problem.  This SPR does not show up in
the WRS support list, so you will have to get it from your WRS rep.

SPR# 30136

vxTicks value in clock_gettime not checked and can be interrupted (ie by
SCSI)

If clock_gettime() is interrupted between it's two reads of the system 
clock, it can return an invalid value.  The following are two examples,
both involve the use of SCSI, which can interrupt the two reads of the
system 
clock.

Craig

Bob Irwin wrote:
> 
> I've created an 18 GB dosFs2 scsi drive /sd1/.  I find that when I copy
> data to the disk with utility commands like cp, copy, etc  the file
> time/date is correct.
> 
> however, when I creat(), write(), and close() a file from
> my app, the time/date is bogus.  The utilities in usrFsLib.c use the
> same calls.  any ideas what might be going on?
> 
> Any help is welcome.
> 
> console capture:
> 
> vxworks-> ll "/sd1/data"
> 
> Listing Directory /sd1/data:
> drwxrwxrwx  1 0       0            16384 Nov 22  2002 ./
> drwxrwxrwx  1 0       0            16384 Aug 15 17:15 ../
> -rwxrwxrwx  1 0       0        104857600 Nov 22  2002 testdata.000
> -rwxrwxrwx  1 0       0        104857600 Nov 22  2002 testdata.001
> -rwxrwxrwx  1 0       0           121011 Aug 15 17:09 dsmain
> 
> vxworks-> creat "/sd1/data/fubar", 1
> value = 6 = 0x6
> vxworks-> close 6
> value = 0 = 0x0
> vxworks-> ll "/sd1/data"
> 
> Listing Directory /sd1/data:
> drwxrwxrwx  1 0       0            16384 Aug 15 17:22 ./
> drwxrwxrwx  1 0       0            16384 Aug 15 17:15 ../
> -rwxrwxrwx  1 0       0        104857600 Nov 22  2002 testdata.000
> -rwxrwxrwx  1 0       0        104857600 Nov 22  2002 testdata.001
> -rwxrwxrwx  1 0       0           121011 Aug 15 17:09 dsmain
> -rwxrwxrwx  1 0       0                0 Aug 15 17:22 fubar
> 
> Thanks,
>    bob
> 
> --
> Bob Irwin | bobi@sis-dev.com | SIS Development | 281-484-6661 | Houston

- --
Craig Johnston                                 
Craig.E.Johnston@boeing.com
                                                (253)773-4808

---------------------------

Newsgroups: comp.os.vxworks
Subject: GDB and Monitoring varible Format
Date: Tue, 22 Aug 2000 03:26:34 GMT
From: "Shawn Girard" 
Organization: Excite@Home - The Leader in Broadband
Message-ID: 

Is there way to have the Crosswind display nothing but hex for the p,p*,i
and i* buttons?

Thanks
Shawn



---------------------------

Newsgroups: comp.os.vxworks
Subject: touching a file date and time
Date: Mon, 21 Aug 2000 13:59:52 -0700
From: "Michael R. Kesti" 
Organization: MK Associates
Message-ID: <39A19848.A31DCCF5@gv.net>
Reply-To: mkesti@gv.net

I am trying to write a un*x-style touch shell command for dosFs files.  If
the file doesn't exist, I creat() and close() the file; no problem.  For the
case of the file already existing, though, I cannot seem to find a way to
set the file's date and time without modifying the file.  Can anybody help?

- -- 
========================================================================
          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: PCI Standard Under VxWorks
Date: Mon, 21 Aug 2000 22:06:09 GMT
From: "Shawn Girard" 
Organization: Excite@Home - The Leader in Broadband
Message-ID: 
References: 

Thanks guys, this information helps me out alot, it enforces what I was
seeing in the lab today.

Another question though, I assuming all this semi plugMplay is being doing
whenever the kernal is booted up.  Is there a way to active the auto
configuration without reboot the kernal each time?


"Shawn Girard"  wrote in message
news:pu9o5.17328$eR5.482441@news1.rdc1.on.wave.home.com...
> Guys, I have some question in reference PCI standard and VxWorks.  I am
> currently writing a driver for PCI target device.  I am little bit
confused
> I how the bars are used and configured.
>
> Questions:
> 1) Thinking in sections, who has the responsible to configure the Bars
> address in the Configuration Header?  Is it the Kernel, Driver, or the
> application?
> 2) I been seeing documentation that talks about memory addresses from
> different point of views.  In this case Host and PCI memory.  Could
someone
> explain this to me?
> 3) The address that is assign to the BAR location of the configuration
> header, which point of view is that, Host or PCI?
>
> Thanks in Advance
> Shawn Girard
>
>



---------------------------

Newsgroups: comp.os.vxworks
Subject: R: PCI Standard Under VxWorks
Date: Mon, 21 Aug 2000 23:15:29 +0100
From: "Merlino" 
Organization: TIN
Message-ID: <8ns6cp$hlm$1@nslave2.tin.it>
References: 
Reply-To: "Merlino" 


Shawn Girard  wrote in message
pu9o5.17328$eR5.482441@news1.rdc1.on.wave.home.com...
> Guys, I have some question in reference PCI standard and VxWorks.  I am
> currently writing a driver for PCI target device.  I am little bit
confused
> I how the bars are used and configured.
>
> Questions:
> 1) Thinking in sections, who has the responsible to configure the Bars
> address in the Configuration Header?  Is it the Kernel, Driver, or the
> application?

If you are running on a x86 PC it is the BIOS in charge to configure BARS
address before vxWorks be started.

> 2) I been seeing documentation that talks about memory addresses from
> different point of views.  In this case Host and PCI memory.  Could
someone
> explain this to me?

Generally speaking PCI address are, of course, physicall address while Host'
s are virtual

If you don ' t have the VMM boundled with Vx they are the same thing.

> 3) The address that is assign to the BAR location of the configuration
> header, which point of view is that, Host or PCI?
>
PCI, physical address

> Thanks in Advance
> Shawn Girard
>

Bye
Merlino



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Watchdog interrupt level
Date: Tue, 22 Aug 2000 06:08:17 GMT
From: gardhab 
Organization: Deja.com - Before you buy.
Message-ID: <8nt5c7$f70$1@nnrp1.deja.com>
References: <8nikm1$tgc$1@nnrp1.deja.com> <4feae8ef49%molochai@vapour-trail.demon.co.uk>


> Depends on what you mean by 'startsup the associated routine'. If by
this you
> mean the processing is carried out by the function called by the
watchdog,
> then no. Watchdogs functions are called by off back of the kernel
timer
> interrupt. (see the VxWorks Programmers' Guide).   It therefore
follows that
> the functions run in interrupt mode, and that the timer interrupt is
disabled
> for the duration.
>
> If the processing you are intending to perform on the watchdog timer
is
> taking so long that another timer interrupt can occur (eg. a kernel
tick) you
> shouldn't be doing it at interrupt level - use high-priority tasks
pending on
> a semaphore given in the watchdog.
>
> In the other hand you are scheduling tasks from the watchdogs, normal
kernel
> scheduling rules apply.  Use a counted semaphore if you want to
guarentee
> getting every watchdog events.

your undertanding of the statement "startsup  the associated routine"
was correct. the explanation provided was most helpful in clearing up
the concept that i was grappling with.
thank you.



Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Download location
Date: 22 Aug 2000 08:03:44 GMT
From: johan@borksoft.xs4all.nl
Organization: Borksoft
Message-ID: <8ntc50$qjd$2@news1.xs4all.nl>
References: 

Michael S.  wrote:
: Anyone know where I can download a beta or trial of the tornado developement
: kit?

There is no download, you can ask your local sales person for an
evaluation license.

Groeten,
	Johan

- -- 
     o o o o o o o . . .   ___________________________________
    o      _____           ||    Johan Borkhuis              |
  .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
 >(________|__|_[_________]_|________________________________|
 _/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: Tornado 2.0 adding componets
Date: 22 Aug 2000 08:00:35 GMT
From: johan@borksoft.xs4all.nl
Organization: Borksoft
Message-ID: <8ntbv3$qjd$1@news1.xs4all.nl>
References: <8nrr41$9b6pr$2@ID-2497.news.cis.dfn.de>

Marcus Overhagen  wrote:
: To get started, I created a new file in \target\config\coms\vxWorks
: which I called 00toolkit.cdf. This file defines a new FOLDER_TOOLKIT
: where I want to manage my components.
: I also added this FOLDER_TOOLKIT to the FOLDER_ROOT folder in file
: 00vxWorks.cdf

You don't have to change the 00vxWorks.cdf file. If you define the item
_CHILDREN in your cdf file and give this the value FOLDER_ROOT it will be
added to the root folder, and you don't get problems when WindRiver
releases a new version of the 00vxWorks.cdf.

: It looks like it does not find the neccessary object files, regardless 
: where I place them.
: I already tried:
: \tornado\target\
: \tornado\target\lib\objCPU32gnuvx
: \tornado\target\config
: \tornado\target\config\MY_BSP
: \tornado\target\project\MY_PROJECT

: All object files of components belonging to Tornado seem to be archived
: in the file \tornado\target\lib\libCPU32gnuvx.a

Add you components to a new archive that is company specific, for example
libCompanyCPU32.a. Then add this library to the LIBS macro in the build
specs.
If you want to use this library for all your projects add this library to
the LIB_EXTRA macro in the BSP Makefile.
It is possible that all other items in your configuration now become
unavailable. To avoid this you have to make some changes to the file
prj_vxworks.tcl. See
http://www.xs4all.nl/~borkhuis/vxworks/vxw_pt2.html#2.6.2 for more info on
this.

: During Tornado 1.0 time, we built and linked the modules using makefiles,
: this created one .O file for each module, which we then linked with the
: main project. Now I don't know the correct way to create these modules.
: If I create them by creating a "downloadable application module"with 
: Tornado 2.0, I either have multiple .O files (one for each .C file) or
: an .a archive or an .out file, depending on the link rule. Neither of 
: these seem to be the correct one. What can I do?

You can select the build-rule objects, this will only generate the objects
per C-file.

Groeten,
	Johan

- -- 
     o o o o o o o . . .   ___________________________________
    o      _____           ||    Johan Borkhuis              |
  .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
 >(________|__|_[_________]_|________________________________|
 _/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: customize tornado function keys?
Date: 22 Aug 2000 08:04:52 GMT
From: johan@borksoft.xs4all.nl
Organization: Borksoft
Message-ID: <8ntc74$qjd$3@news1.xs4all.nl>
References: 

Ryan Tanner  wrote:
: Is there a way to customize the accelerator key bindings in the Tornado GUI?
: In specific, I'd like to remap the debug key bindings.

You have to check the TCL code for that. I did not look into this, but the
menus are buildup using TCL, so it should be possible to reassign them.

Groeten,
	Johan

- -- 
     o o o o o o o . . .   ___________________________________
    o      _____           ||    Johan Borkhuis              |
  .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
 >(________|__|_[_________]_|________________________________|
 _/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 Aug 22 06:06:50 2000
From: Rainer Stelzer 
Date: Tue Aug 22 06:06:53 PDT 2000
Subject: How to configure a vxworks system to act like a gateway from ethernet to 

Hi vxworker's,

I have to establish a VxWorks-host as a gateway between two different
physical networks.
The main network is Ethernet. The other network consists of several
PPP-connections to further VxWorks-hosts.
The protocols on both networks is TCP/IP.
The PPP-Network has to be installed as a subnet of the main network.
I added the route by using the routeAdd() command.
At the moment it is possible to access the dedicated gateway-host over
Ethernet by a PC.
Also it is possible to access any VxWorks-hosts by the gateway-host over
the PPP-Network.
But it isn't possible to access any VxWorks-host beyond the gateway by
the PC.
The reason is that the ARP request of the PC isn't resolved, because the
destination
address isn't available on Ethernet.
How can I configure the gateway to achieve a network transparency for
both Networks?
The "VxWorks Network Programmer's Guide" processes a similar problem in
4.7.
There is a Proxy ARP Server established as a gateway between
Ethernet on the one hand and Shared-Memory-Network on the other hand.
It seems to me, that this solution only works with Shared-Memory and
doesn't with PPP.
Isn't it?

Any hint is welcome.

thanx in advance.

Rainer

vxworks tornado
=====================================================
            Gebrüder Stoye GmbH
department: traffic control systems
      name: Rainer Stelzer
       adr: Longericher Strasse 177
      city: Cologne, Germany
       zip: 50739
     phone: ++49/221/2616-551
    mailto: r.stelzer@stoye.de
       www: http://www.stoye.de
=====================================================




From vxwexplo-errs@csg.lbl.gov  Tue Aug 22 07:28:01 2000
From: "Thomas R. Walsh" 
Date: Tue Aug 22 07:28:03 PDT 2000
Subject: vxWorks test posting

test


From vxwexplo-errs@csg.lbl.gov  Tue Aug 22 17:41:21 2000
From: "Qin Ting" 
Date: Tue Aug 22 17:41:23 PDT 2000
Subject: why does d command showing different content

Hi, vxworks

   I finds a problem recently that when I use d command in target
shell(terminal) to display my device's register, it shows 0xffff02ff(this is
correct). But when in tornado shell, the d command shows 0xffff00ff(this is
incorrect). Then I use @d command in tornado shell, the display is
0xffff02ff(this is correct). My target is MPC860, and I use
WDB_COMM_NETWORK(10M Ethernet).
   Can anyone tell me WHY? Does this affect my application?
   Thanks.

Qin Ting




From vxwexplo-errs@csg.lbl.gov  Wed Aug 23 04:03:08 2000
From: Vxworks Exploder 
Date: Wed Aug 23 04:03:10 PDT 2000
Subject: comp.os.vxworks newsdigest

Comp.Os.Vxworks Daily Digest    Wed Aug 23 04:03:04 PDT 2000

        Subject: Changing / Deleting 2nd IP Address's on Interfaces
        Subject: Value of PCI Card
        Subject: How to configure a vxworks system to act like a gateway from ethernet to
        Subject: Using serial interface of MVME2308
        Subject: Re: Real time clock
        Subject: Re: customize tornado function keys?
        Subject: NY-White Plains - RealTime embedded, RTOS (QNX, VxWorks) DSP, C++, Windows NT, MFC
        Subject: vxWorks test posting
        Subject: Re: Floating point exception trap
        Subject: Re: Value of PCI Card
        Subject: Building custom bootable vxWorks image 
        Subject: Building custom bootable vxWorks image 
        Subject: Re: PPP on VxWorks
        Subject: Re: Building custom bootable vxWorks image
        Subject: Re: PPP on VxWorks
        Subject: Multiple Developers on same project (How to?)
        Subject: Re: touching a file date and time
        Subject: Re: How to configure a vxworks system to act like a gateway from    ethernet to
        Subject: IP Tunneling
        Subject: MOT FCC FEC
        Subject: Re: Building custom bootable vxWorks image
        Subject: regarding bind() fail
        Subject: Re: Value of PCI Card
        Subject: Re: MOT FCC FEC
        Subject: PPP on VxWorks
        Subject: Re: dosFs 1.0
        Subject: Hooking Ctrl-X?
        Subject: Re: ATA/CompactFlash on a PPC860
        Subject: HELP ME!!! THIS IS SERIOUS:(
        Subject: Re: Multicast on VxWorks
        Subject: how to realize TrueFFS using one erase unit
        Subject: Multicast on VxWorks
        Subject: Re: how to realize TrueFFS using one erase unit
        Subject: Re: Hooking Ctrl-X?

-------------------------------------------------------

Newsgroups: comp.os.vxworks
Subject: Changing / Deleting 2nd IP Address's on Interfaces
Date: Tue, 22 Aug 2000 12:30:50 GMT
From: erickcrowell@yahoo.com
Organization: Deja.com - Before you buy.
Message-ID: <8ntrpr$7vg$1@nnrp1.deja.com>

Hello,

  I have a few questions (I checked the list first!) concerning the
ability of vxWorks to add a 2nd IP Address to an existing interface..

1) How do I remove the 2nd IP Address from the interface after I
ifAddrAdd() it?

2) How do I verify if the interface already has the 2nd IP Address
after I assign it?  (The vxWorks Command ifAddrGet() only returns the
first IP Address on the interface).

3) How do I change the 2nd IP Address from the interface after I
ifAddrAdd() it?  (I would be content with the old "remove" and "add" if
I could answer question #1 above).

  ANSWERS to ANY of these questions would be much appreciated!!!

 Thank You in advance.

 Erick C.


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Value of PCI Card
Date: Tue, 22 Aug 2000 09:59:18 +0100
From: "Kai Mattern" 
Organization: Heidelberger Druckmaschinen AG
Message-ID: <8ntbpt$nbm$1@news.hdpp.de>

Good morning all,

i have a PCI card who wanna have 256MB ram mapped. After power up the EEPROM
writes a value into the PCI configuration header. There should (in this
case) a 0xF000 0000 inside the configuration header. But all time i read
them out they are 0 ( all 4 Base Adresses). I tried to read them out after
startup in syslib.c (function sysHwInit) but with no result (allways 0).
Card is running fine under WinNT.....

Anybody have a tip ?

Maybe im in the wrong function...wrong file...to late or something
else....without configuring the card i cant talk to it.

Thanks a lot
Kai




---------------------------

Newsgroups: comp.os.vxworks
Subject: How to configure a vxworks system to act like a gateway from ethernet to
Date: Tue, 22 Aug 2000 15:06:43 +0200
From: Rainer Stelzer 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <8ntucd$9jf$1@overload.lbl.gov>

Hi vxworker's,

I have to establish a VxWorks-host as a gateway between two different
physical networks.
The main network is Ethernet. The other network consists of several
PPP-connections to further VxWorks-hosts.
The protocols on both networks is TCP/IP.
The PPP-Network has to be installed as a subnet of the main network.
I added the route by using the routeAdd() command.
At the moment it is possible to access the dedicated gateway-host over
Ethernet by a PC.
Also it is possible to access any VxWorks-hosts by the gateway-host over
the PPP-Network.
But it isn't possible to access any VxWorks-host beyond the gateway by
the PC.
The reason is that the ARP request of the PC isn't resolved, because the
destination
address isn't available on Ethernet.
How can I configure the gateway to achieve a network transparency for
both Networks?
The "VxWorks Network Programmer's Guide" processes a similar problem in
4.7.
There is a Proxy ARP Server established as a gateway between
Ethernet on the one hand and Shared-Memory-Network on the other hand.
It seems to me, that this solution only works with Shared-Memory and
doesn't with PPP.
Isn't it?

Any hint is welcome.

thanx in advance.

Rainer

vxworks tornado
=====================================================
            Gebrüder Stoye GmbH
department: traffic control systems
      name: Rainer Stelzer
       adr: Longericher Strasse 177
      city: Cologne, Germany
       zip: 50739
     phone: ++49/221/2616-551
    mailto: r.stelzer@stoye.de
       www: http://www.stoye.de
=====================================================



---------------------------

Newsgroups: comp.os.vxworks
Subject: Using serial interface of MVME2308
Date: Tue, 22 Aug 2000 10:34:52 +0200
From: "Holger Roos" 
Organization: Daimler-Benz Aerospace
Message-ID: <8nte7l$d2j@newsserv.vs.dasa.de>

Hi,
I want to use the serial interface of a MVME 2308 for controll of an
external device.
I tried to connect stdin/stout/stderr and logfd (which is by default
connected to the serial interface) by:

#define COMFD  3  /* Serial Com File Descriptor value */
LOCAL int  NullFD;  /* NULL File Descriptor  */

NullFD = open("/null", O_RDWR|O_CREAT, 0); /* Open NULL Device  */
  logFdDelete(COMFD);   /* Remove terminal from LOG  */
  if (ioGlobalStdGet(0) == COMFD)  /* Set standard input to NULL  */
    ioGlobalStdSet(0, NullFD);   /*    */
  if (ioGlobalStdGet(1) == COMFD)  /* Set standard output to NULL */
    ioGlobalStdSet(1, NullFD);   /*    */
  if (ioGlobalStdGet(2) == COMFD)  /* Set standard error to NULL  */
    ioGlobalStdSet(2, NullFD);   /*    */

That doesn't seem to work, because I can't get any answers of the device.
Connecting with telnet to the board helps, because telnet explicitely steals
away stdin/stdout/stderr from the serial interface.
Does anyone have any suggestions what to do ?
Thanks,
Holger.



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Real time clock
Date: Tue, 22 Aug 2000 17:08:45 +0530
From: Vijay Kumar Peshkar 
Organization: Ye 'Ol Disorganized NNTPCache groupie
Message-ID: <39A26645.FCB38309@cisco.com>
References: <8nbm79$i4m$1@crusher.de.colt.net>
Reply-To: vpeshkar@cisco.com

This is a multi-part message in MIME format.
- --------------28403621D7C3E1B6DF67FDE4
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Achim,

I remember using the below functions to set/get the time/date etc.,
mktime,clock_settime,gmtime_r & localtime_r.
Think the last two were suggested with a couple of lines workaround code to get
time by...not sure a friend or a WRS support engineer.

Hope this helps,

regds,
Vijay

Achim Zimmer wrote:

> has anybody programmed a clock including day, month and year based on RTC ?
>
> Is there any free source for calculating date and time from a second counter
> noting leapyears aso ?
>
> Thanks
> Achim

- --------------28403621D7C3E1B6DF67FDE4
Content-Type: text/x-vcard; charset=us-ascii;
 name="vpeshkar.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Card for Vijay Kumar Peshkar
Content-Disposition: attachment;
 filename="vpeshkar.vcf"

begin:vcard 
n:Kumar.P;Vijay
tel;work:080-5530035/36 Extn:3464
x-mozilla-html:FALSE
org:Wipro Technologies, K2, Bangalore.;Wishing you blue skies, a couple of daisies, some fresh air and a smile!
adr:;;;;;;
version:2.1
email;internet:vijay.peshkar@wipro.com
title:Senior Software Engineer
note;quoted-printable:      o   o   o   o   o   o   o   .  .  . _____________________________________________=0D=0A    o           _____                      ||  Vijay Kumar .P         =20                                                    |=0D=0A   .] [__n_n_| DD[  =3D=3D=3D=3D_____   | vpeshkar@cisco.com, vijay_peshkar@hotmail.com     | =0D=0A >(________|__|_[_________]_ |___________________________________________|=0D=0A _/oo OOOOO oo`  ooo   ooo     'o!o!o                                 =20           
                     o!o!o`
fn:Vijay Kumar.P
end:vcard

- --------------28403621D7C3E1B6DF67FDE4--


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: customize tornado function keys?
Date: Tue, 22 Aug 2000 16:46:52 +0530
From: Vijay Kumar Peshkar 
Organization: Ye 'Ol Disorganized NNTPCache groupie
Message-ID: <39A26124.26545AC1@cisco.com>
References:  <8ntc74$qjd$3@news1.xs4all.nl>
Reply-To: vpeshkar@cisco.com

This is a multi-part message in MIME format.
- --------------AA5327D84500697379EE3BB0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Adding to that you must have also got a TCL manual in the box. It has some good
examples too to begin with...

have fun,
Vijay

johan@borksoft.xs4all.nl wrote:

> Ryan Tanner  wrote:
> : Is there a way to customize the accelerator key bindings in the Tornado GUI?
> : In specific, I'd like to remap the debug key bindings.
>
> You have to check the TCL code for that. I did not look into this, but the
> menus are buildup using TCL, so it should be possible to reassign them.
>
> Groeten,
>         Johan
>
> --
>      o o o o o o o . . .   ___________________________________
>     o      _____           ||    Johan Borkhuis              |
>   .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
>  >(________|__|_[_________]_|________________________________|
>  _/oo OOOOO oo`  ooo   ooo  'o!o!o                      o!o!o`
> === VxWorks FAQ: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html ===

- --------------AA5327D84500697379EE3BB0
Content-Type: text/x-vcard; charset=us-ascii;
 name="vpeshkar.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Card for Vijay Kumar Peshkar
Content-Disposition: attachment;
 filename="vpeshkar.vcf"

begin:vcard 
n:Kumar.P;Vijay
tel;work:080-5530035/36 Extn:3464
x-mozilla-html:FALSE
org:Wipro Technologies, K2, Bangalore.;Wishing you blue skies, a couple of daisies, some fresh air and a smile!
adr:;;;;;;
version:2.1
email;internet:vijay.peshkar@wipro.com
title:Senior Software Engineer
note;quoted-printable:      o   o   o   o   o   o   o   .  .  . _____________________________________________=0D=0A    o           _____                      ||  Vijay Kumar .P         =20                                                    |=0D=0A   .] [__n_n_| DD[  =3D=3D=3D=3D_____   | vpeshkar@cisco.com, vijay_peshkar@hotmail.com     | =0D=0A >(________|__|_[_________]_ |___________________________________________|=0D=0A _/oo OOOOO oo`  ooo   ooo     'o!o!o                                 =20           
                     o!o!o`
fn:Vijay Kumar.P
end:vcard

- --------------AA5327D84500697379EE3BB0--


---------------------------

Newsgroups: comp.os.vxworks
Subject: NY-White Plains - RealTime embedded, RTOS (QNX, VxWorks) DSP, C++, Windows NT, MFC
Date: Tue, 22 Aug 2000 11:53:07 GMT
From: vailj@ibm.net (John Vail)
Organization: Advantis
Message-ID: <39a569a2.4919335@news-s01.ny.us.ibm.net>
Reply-To: vailj@ibm.net


Project: C++, RTROS (QNX, VxWorks, pSOS, SPOX) RealTime programming
Job Location: White Plains , New York 
Area code: (914) 
Contact: John Vail 
email: vailj@ibm.net 
Skills desired: C++, VxWorks, RealTime programming, Unix, RISC
processor  knowledge, embedded OS, C,  C++, Windows NT, MFC
 

We  are looking  for a programmer with a strong background
in DSP Real Time Programming, an RTOS, C++, GUI and Windows. 
Knowledge of sockets / TCP/IP (Winsock) under MS Windows is a plus. 
Candidate will be involved in the development of data 
communications, and algorithms for voice / audio compression, and 
computer based telephony integration. Other experiences which 
would be highly desirable are: modem development, digital or 
analog telephony / switching subsystem development, and TAPI


If you wish to send a resume please emial  Mr. John Vail at 
vailj@ibm.net.



---------------------------

Newsgroups: comp.os.vxworks
Subject: vxWorks test posting
Date: Tue, 22 Aug 2000 07:27:55 -0700
From: "Thomas R. Walsh" 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <8nu5dc$e4q$1@overload.lbl.gov>

test

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Floating point exception trap
Date: Tue, 22 Aug 2000 16:44:39 +0530
From: Vijay Kumar Peshkar 
Organization: Ye 'Ol Disorganized NNTPCache groupie
Message-ID: <39A2609F.36FA2400@cisco.com>
References: <39A15280.1573A376@tech-m4.com>
Reply-To: vpeshkar@cisco.com

This is a multi-part message in MIME format.
- --------------52D3F0198C86235B5BFF229C
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Martin,

The syslib.c file has the below lines....
- -------------------------------------------------------
/* initialize floating pt unit */

    if (fppProbe () == OK)
           {
           fppInitialize ();
           intVecSet (INUM_TO_IVEC (IV_FPA_BASE_VEC), fpIntr);
           }

- --------------------------------------------------------

Here replace fpIntr with your routine.

cheers,
Vijay

Martin Beaucage wrote:

> Hi,
>
> I need to replace VxWorks floating point exception interrupt with my
> own interrupt routine.  Up to now, every thing I tried doesn't work.
>
> Can any one help me?
>
> Thanks!
>
>         Martin

- --------------52D3F0198C86235B5BFF229C
Content-Type: text/x-vcard; charset=us-ascii;
 name="vpeshkar.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Card for Vijay Kumar Peshkar
Content-Disposition: attachment;
 filename="vpeshkar.vcf"

begin:vcard 
n:Kumar.P;Vijay
tel;work:080-5530035/36 Extn:3464
x-mozilla-html:FALSE
org:Wipro Technologies, K2, Bangalore.;Wishing you blue skies, a couple of daisies, some fresh air and a smile!
adr:;;;;;;
version:2.1
email;internet:vijay.peshkar@wipro.com
title:Senior Software Engineer
note;quoted-printable:      o   o   o   o   o   o   o   .  .  . _____________________________________________=0D=0A    o           _____                      ||  Vijay Kumar .P         =20                                                    |=0D=0A   .] [__n_n_| DD[  =3D=3D=3D=3D_____   | vpeshkar@cisco.com, vijay_peshkar@hotmail.com     | =0D=0A >(________|__|_[_________]_ |___________________________________________|=0D=0A _/oo OOOOO oo`  ooo   ooo     'o!o!o                                 =20           
                     o!o!o`
fn:Vijay Kumar.P
end:vcard

- --------------52D3F0198C86235B5BFF229C--


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Value of PCI Card
Date: Tue, 22 Aug 2000 10:17:35 -0700
From: James Chen 
Organization: Globix Corp.
Message-ID: <39A2B5AF.8E668DFA@gigabitwireless.com>
References: <8ntbpt$nbm$1@news.hdpp.de>

This is a multi-part message in MIME format.
- --------------EA9F450B18D1B893947F9994
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

How did you "read" the PCI configuration register space? As you would do in
reading a memory location?  The configuration register space must be read using
PCI configuration transaction cycle. It is certainly different from a regular
memory read.

Kai Mattern wrote:

> Good morning all,
>
> i have a PCI card who wanna have 256MB ram mapped. After power up the EEPROM
> writes a value into the PCI configuration header. There should (in this
> case) a 0xF000 0000 inside the configuration header. But all time i read
> them out they are 0 ( all 4 Base Adresses). I tried to read them out after
> startup in syslib.c (function sysHwInit) but with no result (allways 0).
> Card is running fine under WinNT.....
>
> Anybody have a tip ?
>
> Maybe im in the wrong function...wrong file...to late or something
> else....without configuring the card i cant talk to it.
>
> Thanks a lot
> Kai

- --------------EA9F450B18D1B893947F9994
Content-Type: text/x-vcard; charset=us-ascii;
 name="jchen.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Card for James Chen
Content-Disposition: attachment;
 filename="jchen.vcf"

begin:vcard 
n:Chen;James
tel;fax:408.577.0700
tel;work:408.577.7411
x-mozilla-html:FALSE
url:www.gigabitwireless.com
org:Gigabit Wireless, Inc;BTS Software Development
version:2.1
email;internet:jchen@gigabitwireless.com
adr;quoted-printable:;;3099 N. 1st Street,=0D=0A;San Jose;California;95134;
fn:James Chen
end:vcard

- --------------EA9F450B18D1B893947F9994--


---------------------------

Newsgroups: comp.os.vxworks
Subject: Building custom bootable vxWorks image 
Date: Tue, 22 Aug 2000 13:49:56 -0400
From: Bhavesh Patel 
Organization: Posted via Supernews, http://www.supernews.com
Message-ID: <39A2BD44.42CB9D0D@yahoo.com>

Hi,

I am newbie to using Tornado-II project facility on NT. I am trying to
build a custom vxworks image so that it has the network components and
some i/o components ( the default vxworks image for the tornado target
simulator does not have that). The vxworks image is a bootable image for
Motorola SBC8240 (PPC603).The project facility builds a image called
"vxWorks", However when I launch vxSim to select the image named
"vxWorks", it gives the following error message -
"unable to launch vxWorks for the file specified".

I think vxSim is expecting an image with .exe suffix, however the
project facility builds the custom vxWorks image without any prefix.
Any help to resolve this problem is appreciated.
Thanks.

Bhavesh




---------------------------

Newsgroups: comp.os.vxworks
Subject: Building custom bootable vxWorks image 
Date: Tue, 22 Aug 2000 13:49:17 -0400
From: Bhavesh Patel 
Organization: Posted via Supernews, http://www.supernews.com
Message-ID: <39A2BD1D.30FC6D48@yahoo.com>

Hi,

I am newbie to using Tornado-II project facility on NT. I am trying to
build a custom vxworks image so that it has the network components and
some i/o components ( the default vxworks image for the tornado target
simulator does not have that). The vxworks image is a bootable image for
Motorola SBC8240 (PPC603).The project facility builds a image called
"vxWorks", However when I launch vxSim to select the image named
"vxWorks", it gives the following error message -
"unable to launch vxWorks for the file specified".

I think vxSim is expecting an image with .exe suffix, however the
project facility builds the custom vxWorks image without any prefix.
Any help to resolve this problem is appreciated.
Thanks.

Bhavesh




---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: PPP on VxWorks
Date: 22 Aug 2000 18:28:24 GMT
From: johan@borksoft.xs4all.nl
Organization: Borksoft
Message-ID: <8nugo8$o21$1@news1.xs4all.nl>
References: <8nuekf$1df$1@news.umbc.edu>

Andrej Cedilnik  wrote:
: Is there a free implementation of PPP that runs under VxWorks 5.4?

Yes, it is included with VxWorks. Select PPP when configuring your kernel.

Groeten,
	Johan

- -- 
     o o o o o o o . . .   ___________________________________
    o      _____           ||    Johan Borkhuis              |
  .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
 >(________|__|_[_________]_|________________________________|
 _/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: Building custom bootable vxWorks image
Date: 22 Aug 2000 18:33:30 GMT
From: johan@borksoft.xs4all.nl
Organization: Borksoft
Message-ID: <8nuh1q$o21$2@news1.xs4all.nl>
References: <39A2BD44.42CB9D0D@yahoo.com>

Bhavesh Patel  wrote:
: Hi,

: I am newbie to using Tornado-II project facility on NT. I am trying to
: build a custom vxworks image so that it has the network components and
: some i/o components ( the default vxworks image for the tornado target
: simulator does not have that). The vxworks image is a bootable image for
: Motorola SBC8240 (PPC603).The project facility builds a image called
: "vxWorks", However when I launch vxSim to select the image named
: "vxWorks", it gives the following error message -
: "unable to launch vxWorks for the file specified".

vxSim is not a simulator to run a vxworks image, but it is a vxworks image
itself. So if you want to run VxSim you have to build a new VxWorks
simulator with the options that you want.

In the Tornado Getting started guide an example is given on how to start
with VxWorks and how to build a custom configured simulator.

Groeten,
	Johan

- -- 
     o o o o o o o . . .   ___________________________________
    o      _____           ||    Johan Borkhuis              |
  .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
 >(________|__|_[_________]_|________________________________|
 _/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: PPP on VxWorks
Date: 22 Aug 2000 19:10:33 GMT
From: Andrej Cedilnik 
Organization: University of Maryland, Baltimore County
Message-ID: <8nuj79$84f$2@news.umbc.edu>
References: <8nuekf$1df$1@news.umbc.edu>

Andrej Cedilnik  wrote:
> Is there a free implementation of PPP that runs under VxWorks 5.4?

Let me refraze that. Is there a source for a free implementation of
PPP that runs under VxWorks 5.4?
I found the one that works on 5.2 but it does not compile under 5.4.

			Andy

---------------------------

Newsgroups: comp.os.vxworks
Subject: Multiple Developers on same project (How to?)
Date: Tue, 22 Aug 2000 15:22:07 -0400
From: "Joe Ott" 
Organization: Verio
Message-ID: 


Tornado is a new developement environment to us and we are in the beginning
phases of a project with multiple developers. The project
(Tornado-workspace) will be seperated into various sub-projects
(Tornado-projects). Our concerns are:

1) There will be multiple developers working on different pieces of the same
sub-project. The work of developer A will not interfer with the work of
developer B but the combined work will compile to a single task. What are
some methods that can be used to allow several developers to work on their
seperate areas and still allow them to compile and test without waiting for
the other developer to be finished?

2) The various sub-projects (Tornado-projects) reflect various SW layers. In
order to test the upper layers working copies of the lower layers are
needed. What is the best way to support the upper layers while modifying the
lower layers? In other words what is the best way to keep a copy of a
working downloadable image (*.out file) other can use while you are making
changes to the project?

3) We will be using source control. Beside our source code, what if any
Tornado files (workspace, project, makefile, etc) should be source
controlled?

One problem we have encounter with the Tornado project and workspace files
is that if there are two developers working and they both make changes to
the files (such as adding projects or adding files to the project) the last
one to save wins. That is the changes made by the first one who saved the
files are lost.

Has anyone worked in an environment where there were many developers and you
had to keep a working downloadable image for other to use while making
changes to your own project?

Any help or insight would be appreciated.

Joe Ott
VTG, an Intel Dialogic Company




---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: touching a file date and time
Date: Tue, 22 Aug 2000 09:11:32 -0400
From: Charles H. Chapman 
Organization: Dynacs Engineering Company
Message-ID: <8ntua4$47q@niven.ksc.nasa.gov>
References: <39A19848.A31DCCF5@gv.net>
Reply-To: chc@nasa2.ksc.nasa.gov

Michael R. Kesti wrote:
> I am trying to write a un*x-style touch shell command for dosFs files.  If
> the file doesn't exist, I creat() and close() the file; no problem.  For the
> case of the file already existing, though, I cannot seem to find a way to
> set the file's date and time without modifying the file.  Can anybody help?

Look at the function "utime".  BTW, there is no time-of-last-access
maintained for files in DOS file systems, only the time of creation or
last modification.

Chuck

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: How to configure a vxworks system to act like a gateway from    ethernet to
Date: Tue, 22 Aug 2000 08:50:02 -0500
From: Greg Willden 
Organization: Posted via Supernews, http://www.supernews.com
Message-ID: <39A2850A.FDAAAAED@swri.edu>
References: <8ntucd$9jf$1@overload.lbl.gov>

Rainer Stelzer wrote:

> At the moment it is possible to access the dedicated gateway-host over
> Ethernet by a PC.
> Also it is possible to access any VxWorks-hosts by the gateway-host over
> the PPP-Network.
> But it isn't possible to access any VxWorks-host beyond the gateway by
> the PC.
> The reason is that the ARP request of the PC isn't resolved, because the
> destination
> address isn't available on Ethernet.
> How can I configure the gateway to achieve a network transparency for
> both Networks?

Just some things to verify.
1. Have you added a route to the PC's routing table to go through the
VxWorks target to get to the PPP network? (using route add -p)
2. Have you setup the VxWorks target as the defualt gateway on the PPP
network?



---------------------------

Newsgroups: comp.os.vxworks
Subject: IP Tunneling
Date: 22 Aug 2000 13:47:43 GMT
From: Andrej Cedilnik 
Organization: University of Maryland, Baltimore County
Message-ID: <8nu09v$jsc$1@news.umbc.edu>

Hello!

Is there any free implementation of some tunneling protocol for VxWorks?
I would like to have network device that would tunnel all trafic through
a regular network device and maybe process it a little bit.
Does somebody have a source of something like this?

			Andy


---------------------------

Newsgroups: comp.os.vxworks
Subject: MOT FCC FEC
Date: Mon, 21 Aug 2000 09:04:38 +0100
From: Graham Baxter 
Organization: Graham Baxter (Software) Limited
Message-ID: <39A0E296.4B2C48D7@NOSPAM.bcs.org.uk>

VxWorkers,

We are using the Motorola Fast Ethernet Controler driver on the PPC
8260.

When we ping our target, about 4% of the pings are lost.

Does anyone have any advice? Should we increase the resources (buffers,
BDs etc)?

We have tried a direct target to host connection and get the same
results.


Thanks for any help.


Graham Baxter
gbaxter@NOSPAMbcs.org.uk



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Building custom bootable vxWorks image
Date: Tue, 22 Aug 2000 16:37:22 -0400
From: Bhavesh Patel 
Organization: Posted via Supernews, http://www.supernews.com
Message-ID: <39A2E482.7153FC51@yahoo.com>
References: <39A2BD44.42CB9D0D@yahoo.com> <8nuh1q$o21$2@news1.xs4all.nl>

johan@borksoft.xs4all.nl wrote:

> Bhavesh Patel  wrote:
> : Hi,
>
> : I am newbie to using Tornado-II project facility on NT. I am trying to
> : build a custom vxworks image so that it has the network components and
> : some i/o components ( the default vxworks image for the tornado target
> : simulator does not have that). The vxworks image is a bootable image for
> : Motorola SBC8240 (PPC603).The project facility builds a image called
> : "vxWorks", However when I launch vxSim to select the image named
> : "vxWorks", it gives the following error message -
> : "unable to launch vxWorks for the file specified".
>
> vxSim is not a simulator to run a vxworks image, but it is a vxworks image
> itself. So if you want to run VxSim you have to build a new VxWorks
> simulator with the options that you want.
>
> In the Tornado Getting started guide an example is given on how to start
> with VxWorks and how to build a custom configured simulator.
>
> Groeten,
>         Johan
>
> --
>      o o o o o o o . . .   ___________________________________
>     o      _____           ||    Johan Borkhuis              |
>   .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
>  >(________|__|_[_________]_|________________________________|
>  _/oo OOOOO oo`  ooo   ooo  'o!o!o                      o!o!o`
> === VxWorks FAQ: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html ===

Thanks for the help, I have figured out the solution.

bhavesh


---------------------------

Newsgroups: comp.os.vxworks
Subject: regarding bind() fail
Date: Tue, 22 Aug 2000 17:05:44 -0400
From: Bhavesh Patel 
Organization: Posted via Supernews, http://www.supernews.com
Message-ID: <39A2EB28.963FAD6C@yahoo.com>

Hi,

I am trying to use adventnet snmp agent on vxSim for NT. The agent as
part of its initialization tries to create a datagram socket at port
161. The socket() is successful, however the bind() call fails. To get
around this problem I use setsockopt with the option for
SO_REUSEADDR turned ON. However, setsockopt fails. below is the dump of
inetstatShow -

- ->inetstatShow
Active Internet connections (including servers)
PCB        Proto    Recv-Q    Send-Q    LocalAddr  ForeignAddr State
4b83744    UDP        0        0        0.0.0.0    0.0.0.0
4b836c0    UDP        0        0        0.0.0.0    0.0.0.0
4b8363c    UDP        0        0        0.0.0.9889    0.0.0.0
4b8334     UDP        0        0        0.0.0.0    0.0.0.0
value = 0 = 0x0

- -----------------------
Any help is appreciated.

bhavesh


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Value of PCI Card
Date: Tue, 22 Aug 2000 17:51:15 GMT
From: John Fusco 
Organization: Deja.com - Before you buy.
Message-ID: <8nuei9$vgk$1@nnrp1.deja.com>
References: <8ntbpt$nbm$1@news.hdpp.de>

Some more information would be helpful...
What BSP?
What Processor?
Are you using the autoconfig library?

I have 256MB PCI slave devices that are configured by the autoconfig
library.  By default, my BSP did not provide enough space for these
devices, so they were never configured.  I had to tweak the system
information that was given to the autoconfig routine so that it would
know what address space to use for the device.  This information is in
the PCI_SYSTEM structure that is passed to the pciAutoConfig() routine.

Of course none of this applies to you if you are not using autoconfig.

Good luck.

In article <8ntbpt$nbm$1@news.hdpp.de>,
  "Kai Mattern"  wrote:
> Good morning all,
>
> i have a PCI card who wanna have 256MB ram mapped. After power up the
EEPROM
> writes a value into the PCI configuration header. There should (in
this
> case) a 0xF000 0000 inside the configuration header. But all time i
read
> them out they are 0 ( all 4 Base Adresses). I tried to read them out
after
> startup in syslib.c (function sysHwInit) but with no result (allways
0).
> Card is running fine under WinNT.....
>
> Anybody have a tip ?
>
> Maybe im in the wrong function...wrong file...to late or something
> else....without configuring the card i cant talk to it.
>
> Thanks a lot
> Kai
>
>


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: MOT FCC FEC
Date: Tue, 22 Aug 2000 17:09:06 GMT
From: gold@sdsu.edu
Organization: Deja.com - Before you buy.
Message-ID: <8nuc2v$s5h$1@nnrp1.deja.com>
References: <39A0E296.4B2C48D7@NOSPAM.bcs.org.uk>

In article <39A0E296.4B2C48D7@NOSPAM.bcs.org.uk>,
  Graham Baxter  wrote:
> VxWorkers,
>
> We are using the Motorola Fast Ethernet Controler driver on the PPC
> 8260.
>
> When we ping our target, about 4% of the pings are lost.
>
> Does anyone have any advice? Should we increase the resources
(buffers,
> BDs etc)?
>
> We have tried a direct target to host connection and get the same
> results.
>
> Thanks for any help.

If you're just pinging and nothing else, and have reasonably allocated
resources, you shouldn't be running out, so increasing the number of
BDs and buffers may not help.  Have you looked at the statistics for
the driver, i.e. how many errors it's seen on Tx/Rx ?  Try doing a
motFccErrorShow or motFccDrvShow ?  If you have the source to the
driver you'll see they don't handle the 8260 errata for CPM#21
correctly, at least in my *very* old copy.  Maybe they've updated the
driver since I've seen it.  Are you using their driver or your own ?

Do you have a network analyzer so you can see whether you're losing
them on the Tx or the Rx side ?  Again, the driver stats will help
you.  Are you running half or full duplex ?  I assume 100Base-T since
you're on the FCC.  Since it's pings also try using icmpstatShow().

  hth,
   Dan Gold  gold@sdsu.edu


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: PPP on VxWorks
Date: 22 Aug 2000 17:52:15 GMT
From: Andrej Cedilnik 
Organization: University of Maryland, Baltimore County
Message-ID: <8nuekf$1df$1@news.umbc.edu>

Is there a free implementation of PPP that runs under VxWorks 5.4?

			Andy

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: dosFs 1.0
Date: 22 Aug 2000 13:28:19 -0500
From: Paul Banta 
Organization: ADC Telecommunications
Message-ID: 
References: <8nha2v$956$1@overload.lbl.gov>
Sender: pbanta@wst218

David Anderson  writes:

> vxWorks Tornado
> -------------------------
> 
> Can anyone point me at the correct way of verifying a ram block device 
> contains a valid dos volume before calling
> dosFsInit(...) which appears to crash with bus exceptions if the BLOCK_DEV 
> has not previously been formatted.
> ( the manual actually states that it should return NULL volume 
> descriptor.... )
> 
> e.g.
> 
>       pBlkDev = ramDevCreate( address , 512 , 400 , 400 , 0 );
>       pVolDesc= dosFsDevInit("DEV1:",pBlkDev,NULL )
>       if ( pVolDesc == NULL )
>       {
>               pVolDesc = dosFsMkFs("DEV1:",pBlkDev);
>       }
> 
> does not work - dosFsDevInit() fails with an address exception ( attempting 
> to perform a bcopy(...) ).
> 
> p.s. we are talking dosfs 1.0 and vxworks 5.3.1 here - old stuff so there 
> must be a guru for it by now !.
> 

How do you know that pBlkDev is non-null before you call
dosFsDevInit()?  What is the value of "address" before you call
ramDevCreate()? 

I have no problems with this code (which is nearly the same as what you
have):

  if ((pRdBlkDev  = ramDevCreate(RAM_DISK_BASE_ADDR, RAM_DISK_BYTES_IN_BLK,
				 RAM_DISK_BLOCKS, RAM_DISK_BLOCKS, 0)) == NULL)
    {
      returnCode = ERROR;
    }
  .
  .
  .
      if ((pRdVolDesc = dosFsDevInit(RAM_DISK_NAME, pRdBlkDev, NULL)) == NULL)
	{
	  returnCode = ERROR;
	}



- -- 
- --
Paul Banta
ADC Telecommunications
Broadband Access & Transport Group         email paul_banta@spamtrap.adc.com

---------------------------

Newsgroups: comp.os.vxworks
Subject: Hooking Ctrl-X?
Date: Wed, 23 Aug 2000 01:28:27 GMT
From: krhohio@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8nv9b8$vbu$1@nnrp1.deja.com>

Hi Tornado 1 and VxWorks 5.3.1 users,

I need to "hook" the Ctrl-X (boot-to-monitor ROM) serial port character
and do some application specific things before the reboot occurs.

Can this be done using ioctl(fd, FIOPROTOHOOK, pFunction)?
Do I also need to execute ioctl (fd, FIOPROTOARG, arg)?

Thanks,
Kent


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: ATA/CompactFlash on a PPC860
Date: 23 Aug 2000 06:12:08 GMT
From: houcheng@ms17.hinet.net ()
Organization: DCI HiNet
Message-ID: <8nvpvo$fp0@netnews.hinet.net>
References: <8m7htm$1pml$1@thoth.cts.com>

Steve Jones (sjones@logici.com) wrote:
: I am trying to get the ATA driver up and running on my PPC860.  I am setting
: up the PCMCIA interface in the PPC, and then initializing the ATA driver
: over the top of it.  The reference I have is a little different it is using
: the Winbond chip over a PCI bus.  I have made the changes to the driver but
: am a bit confused on the addressing to the register interface of an ATA
: device.  Is there any one out their that has added a CompactFlash card to
: there PPC project?



        I'm using Motorolla FADS860 Platofmr. It supports

PCMCIA2.1 complaint interface and CompactFlash card 

works well with FADS860 board. 

   You can try to access compactflash's

attribute/common memory space by ICD(In Circuit Debugger)

to ensure that your board support CompactFlash PCMCIA card.

After access att/comm memory space, you can try to 

send ATA command through common memory space. Hope this

helps.

                                    Lin Houcheng


---------------------------

Newsgroups: comp.os.vxworks
Subject: HELP ME!!! THIS IS SERIOUS:(
Date: Wed, 23 Aug 2000 06:17:10 GMT
From: fowcpq@nurb.com
Organization: Clientes_Teleline
Message-ID: 

My parents are going to lose their home if I cannot raise $1,000 by next week! I would appreciate if everyone could donate money to me by sending a donation through paypal to shocker1@nurb.com! If you could do this for me I would be so greatful! If you cant send me some money, atleast join up under shocker1@nurb.com at www.paypal.com and I'll get $5 for everyone who joins under me. Please help me save my family. If you do this I'll be eternally greatful and will return the favor once my family gets in bett
er financial condition!
P.S. If you send me any money through papyal, email me afterwards and when I get back on my feet I'll be glad to repay you what you lent me!




---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Multicast on VxWorks
Date: Wed, 23 Aug 2000 09:27:46 +0200
From: Albert Wijnja 
Organization: MECO Equipment Engineers BV
Message-ID: <39A37CF2.709EC383@meco.nl>
References: <8nv98c$coh$1@news.elim.net>
Reply-To: Albert.Wijnja@meco.nl

Sung,

I use multicasting for TAO. I had some problems in the beginning. At one
point I added a route and from then on multicasting ran smoothly. If the
IP-address of your card is e.g. 90.0.0.50 add a route with the following
command:

    routeAdd “224.0.0.0”, “90.0.0.50”

The multicasting IP address range runs from 224.0.0.0 to 239.255.255.255.
See "VxWorks Network Programmers Manual".

Albert


Sung Won Ho wrote:

> Hi all,
>
> Does anyone have an experience in using multicast feature in VxWorks?
> I use T2 and END driver.
> But it seemed that it can't process multicast packet.
>
> Do you have any comments?
>
> Sung.


---------------------------

Newsgroups: comp.os.vxworks
Subject: how to realize TrueFFS using one erase unit
Date: 23 Aug 2000 08:51:20 -0700
From: "sunny" 
Organization: The news server of Changzhou TeleCom
Message-ID: <01bfb178$b9b12750$8cc210ac@sunny>

Hi all,

Now I am engaged in filesystem using TrueFFS. But there have something
problems with me. 

We use atmel flash(AT49LV040) to realize filesystem. There is only chip
erase in AT49LV040,that is ,there is only one erase unit about AT49LV040.
Through introduction about TrueFFS from windriver, I think the garbage
collection and wear-leveling need serveral earse units to realize. 

So I want to know if it is possible that realize TrueFFS filesystem with
only one erase unit. if yes, which keys we should consider in details?

Thanks very much!

Regards,


---------------------------

Newsgroups: comp.os.vxworks
Subject: Multicast on VxWorks
Date: Wed, 23 Aug 2000 10:23:33 +0900
From: "Sung Won Ho" 
Organization: ELIMnet Internet News
Message-ID: <8nv98c$coh$1@news.elim.net>

Hi all,

Does anyone have an experience in using multicast feature in VxWorks?
I use T2 and END driver.
But it seemed that it can't process multicast packet.

Do you have any comments?

Sung.




---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: how to realize TrueFFS using one erase unit
Date: Wed, 23 Aug 2000 11:35:23 +0200
From: Leonid Rosenboim 
Organization: BitBand Technologies Ltd. http://www.bitband.com
Message-ID: <39A39ADB.7FD1F5FA@bitband.com>
References: <01bfb178$b9b12750$8cc210ac@sunny>
Reply-To: leonid@bitband.com

ANy Flash file system or translation layer needs multiple (at lease 8 ?)
erase units,
otherwise it wont be able to erase one block without th eothers, and it must
erase
in order to change bits from 0 to 1.

If your Flash chip is a bulk-erase type, i.e. the whole chip is a single
erase unit, but
you have many such chips in teh Flash array, at least 8, you can probably
tweak the
driver enough to use each separate chip as an erase unit. Remember that one
erase unit is kept as spare at all times, so the more erase units you got the
less overhead
it is to have this spare block.

sunny wrote:

> Hi all,
>
> Now I am engaged in filesystem using TrueFFS. But there have something
> problems with me.
>
> We use atmel flash(AT49LV040) to realize filesystem. There is only chip
> erase in AT49LV040,that is ,there is only one erase unit about AT49LV040.
> Through introduction about TrueFFS from windriver, I think the garbage
> collection and wear-leveling need serveral earse units to realize.
>
> So I want to know if it is possible that realize TrueFFS filesystem with
> only one erase unit. if yes, which keys we should consider in details?
>
> Thanks very much!
>
> Regards,


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Hooking Ctrl-X?
Date: Wed, 23 Aug 2000 03:15:06 GMT
From: pkockritz@home.com (Pete Kockritz)
Organization: @Home Network
Message-ID: 
References: <8nv9b8$vbu$1@nnrp1.deja.com>

In article <8nv9b8$vbu$1@nnrp1.deja.com>, krhohio@my-deja.com wrote:

> Hi Tornado 1 and VxWorks 5.3.1 users,
> 
> I need to "hook" the Ctrl-X (boot-to-monitor ROM) serial port character
> and do some application specific things before the reboot occurs.
> 
> Can this be done using ioctl(fd, FIOPROTOHOOK, pFunction)?
> Do I also need to execute ioctl (fd, FIOPROTOARG, arg)?
> 

You can put the code you desire to be executed in SysToMonitor(). That
should be in sysLib.c

Regards,
Pete
- --
+-----------------------------------------------------------------+
|       Pete Kockritz                  mailto:pkockritz@home.com  |
|                       This space for rent.                      |
+-----------------------------------------------------------------+

---------------------------

End of New-News digest
**********************


From vxwexplo-errs@csg.lbl.gov  Wed Aug 23 09:09:21 2000
From: "Straub, Robert" 
Date: Wed Aug 23 09:09:23 PDT 2000
Subject: VxWorks/Tornado shell checksum and ls utilities

Fellow VxWorkers,

I am trying to find some UNIX-like utilities that will run from the VxWorks
Tornado or target shell.

- One is a directory listing utility.  I want it to display a recursive
listing of files in a dosFs directory tree structure, in a similar way to
the UNIX command, "ls -lR".

- The other is a file checksum utility that will display checksums for one
or more files in a dosFs directory, in a similar way to the UNIX command
"sum *".  If this program has a "recursive" option, to sum all files in a
tree structure, so much the better.  It would also be helpful if there were
a companion program that would run the same algorithm on files in a W*****s
host.

The goal is to gain confidence that files which were downloaded to a dosFs
file system in flash memory (True FFS) have not become corrupted.

Thanks in advance for any help you can provide.

Bob Straub.

If this is a duplicate posting on this subject, I apologize.  I tried to
post though deja.com / comp.os/vxworks yesterday and had trouble (I think).

Robert J. Straub
Principal Software Engineer
Telephonics Corporation
815 Broad Hollow Road, Mail Stop 3
Farmingdale, NY 11735

516/631-755-7769 (Voice)
516/631-755-7642 (Fax)

straub@telephonics.com




From vxwexplo-errs@csg.lbl.gov  Thu Aug 24 04:03:12 2000
From: Vxworks Exploder 
Date: Thu Aug 24 04:03:14 PDT 2000
Subject: comp.os.vxworks newsdigest

Comp.Os.Vxworks Daily Digest    Thu Aug 24 04:03:07 PDT 2000

        Subject: The use of multiple NIC
        Subject: endPollSend, endPollReceive usage ?
        Subject: Re: Value of PCI Card
        Subject: Re: Multiple Developers on same project (How to?)
        Subject: Re: ATA/CompactFlash on a PPC860
        Subject: c&t 69030
        Subject: Re: PPC860 WatchDog Timer
        Subject: Re: endPollSend, endPollReceive usage ?
        Subject: Re: Value of PCI Card
        Subject: Search for LAN driver for SiS900 in a SiS630
        Subject: Re: Multicast on VxWorks
        Subject: VxWorks/Tornado shell checksum and ls utilities
        Subject: third party VxWorks shell / extensions
        Subject: TrueFFS
        Subject: mvme2700 rev level?
        Subject: compiler upgrade with Tornado
        Subject: Loading Object Module
        Subject: Pcmcia! How to start?
        Subject: Installation of PMC-Octal Pro-232 on MVME2306 
        Subject: Multiple symbol tables
        Subject: Nt build issue
        Subject: Re: Multicast on VxWorks
        Subject: Re: Pcmcia! How to start?
        Subject: Strange interrupt behavior noticed with WindView
        Subject: Re: Need Database, RTOS and File System developers
        Subject: Problem with ls120 Panasonic
        Subject: Bizarre page faults
        Subject: Pcmcia! How to start?
        Subject: security between 2 networks
        Subject: Re: Multiple Developers on same project (How to?)
        Subject: Re: Bizarre page faults
        Subject: boot networking  problems
        Subject: 16550 RS232 dirver for x86
        Subject: C++ Execptions in VxWorks
        Subject: Re: MOT FCC FEC
        Subject: Re: Adventnet SNMP agent for vxworks
        Subject: Re: Value of PCI Card
        Subject: Hi, LINUX HERE!!!!
        Subject: need help with project facility
        Subject: VxSim i/o performance unusably slow?
        Subject: Re: Loading Object Module
        Subject: Re: need help with project facility
        Subject: Re: Multiple symbol tables
        Subject: Help
        Subject: Q: How to use memPartCreate()
        Subject: target server/dual ethernet devices
        Subject: Re: Multiple symbol tables
        Subject: PCI driver guidelines?
        Subject: Problems with tabular index of WindNet SNMP Method Routines

-------------------------------------------------------

Newsgroups: comp.os.vxworks
Subject: The use of multiple NIC
Date: Wed, 23 Aug 2000 19:43:59 +0900
From: "Sung Won Ho" 
Organization: ELIMnet Internet News
Message-ID: <8o0a3d$vn4$1@news.elim.net>

Hi,

I want to use two NIC in VxWorks.
When I customize VxWorks image, I found that Tornado regenerates
projConfig.c

Looking into projConfig.c, it seems there is no way to initialize second
NIC.
It performs initialization only once for booting NIC.

Where can I insert attaching and initialization code such as IP address
configuration
for second NIC?
Does anyone have an idea?

Thanks,

Sung.





---------------------------

Newsgroups: comp.os.vxworks
Subject: endPollSend, endPollReceive usage ?
Date: Wed, 23 Aug 2000 00:27:55 -0700
From: "Sandhya Potluri" 
Organization: Monmouth Internet
Message-ID: <8nvjna$sf5$1@news.monmouth.com>

Hi,
    I am testing a ethernet driver for an ethernet driver on 860T. Has
anybody used the endPollSend and endPollReceive routine? Help appreciated.
Thanks
Sandy



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Value of PCI Card
Date: Wed, 23 Aug 2000 15:38:34 +0100
From: "Kai Mattern" 
Organization: Heidelberger Druckmaschinen AG
Message-ID: <8o0k24$um1$1@news.hdpp.de>
References: <8ntbpt$nbm$1@news.hdpp.de> <8nuei9$vgk$1@nnrp1.deja.com>

What target do you have ? I have a Motorola 603 board and thought autoconfig
only run with 750 boards ?



"John Fusco"  wrote in message
news:8nuei9$vgk$1@nnrp1.deja.com...
> Some more information would be helpful...
> What BSP?
> What Processor?
> Are you using the autoconfig library?
>
> I have 256MB PCI slave devices that are configured by the autoconfig
> library.  By default, my BSP did not provide enough space for these
> devices, so they were never configured.  I had to tweak the system
> information that was given to the autoconfig routine so that it would
> know what address space to use for the device.  This information is in
> the PCI_SYSTEM structure that is passed to the pciAutoConfig() routine.
>
> Of course none of this applies to you if you are not using autoconfig.
>
> Good luck.
>
> In article <8ntbpt$nbm$1@news.hdpp.de>,
>   "Kai Mattern"  wrote:
> > Good morning all,
> >
> > i have a PCI card who wanna have 256MB ram mapped. After power up the
> EEPROM
> > writes a value into the PCI configuration header. There should (in
> this
> > case) a 0xF000 0000 inside the configuration header. But all time i
> read
> > them out they are 0 ( all 4 Base Adresses). I tried to read them out
> after
> > startup in syslib.c (function sysHwInit) but with no result (allways
> 0).
> > Card is running fine under WinNT.....
> >
> > Anybody have a tip ?
> >
> > Maybe im in the wrong function...wrong file...to late or something
> > else....without configuring the card i cant talk to it.
> >
> > Thanks a lot
> > Kai
> >
> >
>
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Multiple Developers on same project (How to?)
Date: Wed, 23 Aug 2000 14:46 +0100 (BST)
From: rew@cp.cix.co.uk (Richard Wenham)
Organization: CIX - Compulink Information eXchange
Message-ID: 
References: 
Reply-To: rew@cp.cix.co.uk

This is how we manage the problem.  Everything required to build the 
system is kept under source control (we use SourceSafe), including the 
directory hierarchy.  Each developer gets a copy of all sources into their 
private area, we use local workstation disk, and builds their own copy of 
the complete system.

Before a chunk of development, a developer will get the latest version of 
everything from source control, check out the required files, do the work, 
and test it on his own target. The changed files are then checked in 
again, at which point text merging may be required if another developer 
has been working on the same files.

The cycle repeats.  Co-operation is required when getting the most recent 
source to avoid another developer's partial check in.

Richard Wenham
Cranfield Precision

In article , 
joe.ott@dialogic.com (Joe Ott) wrote:

> 
> Tornado is a new developement environment to us and we are in the 
> beginning
> phases of a project with multiple developers. The project
> (Tornado-workspace) will be seperated into various sub-projects
> (Tornado-projects). Our concerns are:
> 
> 1) There will be multiple developers working on different pieces of the 
> same
> sub-project. The work of developer A will not interfer with the work of
> developer B but the combined work will compile to a single task. What 
> are
> some methods that can be used to allow several developers to work on 
> their
> seperate areas and still allow them to compile and test without waiting 
> for
> the other developer to be finished?
> 
> 2) The various sub-projects (Tornado-projects) reflect various SW 
> layers. In
> order to test the upper layers working copies of the lower layers are
> needed. What is the best way to support the upper layers while 
> modifying the
> lower layers? In other words what is the best way to keep a copy of a
> working downloadable image (*.out file) other can use while you are 
> making
> changes to the project?
> 
> 3) We will be using source control. Beside our source code, what if any
> Tornado files (workspace, project, makefile, etc) should be source
> controlled?
> 
> One problem we have encounter with the Tornado project and workspace 
> files
> is that if there are two developers working and they both make changes 
> to
> the files (such as adding projects or adding files to the project) the 
> last
> one to save wins. That is the changes made by the first one who saved 
> the
> files are lost.
> 
> Has anyone worked in an environment where there were many developers 
> and you
> had to keep a working downloadable image for other to use while making
> changes to your own project?
> 
> Any help or insight would be appreciated.
> 
> Joe Ott
> VTG, an Intel Dialogic Company

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: ATA/CompactFlash on a PPC860
Date: 23 Aug 2000 06:13:12 GMT
From: houcheng@ms17.hinet.net ()
Organization: DCI HiNet
Message-ID: <8nvq1o$fp0@netnews.hinet.net>
References: <8m7htm$1pml$1@thoth.cts.com>

Steve Jones (sjones@logici.com) wrote:
: I am trying to get the ATA driver up and running on my PPC860.  I am setting
: up the PCMCIA interface in the PPC, and then initializing the ATA driver
: over the top of it.  The reference I have is a little different it is using
: the Winbond chip over a PCI bus.  I have made the changes to the driver but
: am a bit confused on the addressing to the register interface of an ATA
: device.  Is there any one out their that has added a CompactFlash card to
: there PPC project?




        I'm using Motorolla FADS860 Platofmr. It supports

PCMCIA2.1 complaint interface and CompactFlash card 

works well with FADS860 board. 

   You can try to access compactflash's

attribute/common memory space by ICD(In Circuit Debugger)

to ensure that your board support CompactFlash PCMCIA card.

After access att/comm memory space, you can try to 

send ATA command through common memory space. Hope this

helps.

                                    Lin Houcheng


---------------------------

Newsgroups: comp.os.vxworks
Subject: c&t 69030
Date: Wed, 23 Aug 2000 13:54:35 GMT
From: dpepper@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8o0l2h$fqu$1@nnrp1.deja.com>

Anybody used the 69030 with UGL?

thanks,

dp


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: PPC860 WatchDog Timer
Date: 23 Aug 2000 06:19:19 GMT
From: houcheng@ms17.hinet.net ()
Organization: DCI HiNet
Message-ID: <8nvqd7$fp0@netnews.hinet.net>
References: <8J8TP6$PV4$1@NEWS.NURI.NET>


À¯Ã¶±Ô (ckryu@woorigs.co.kr) wrote:
: Hi.
: I use PPC860 target board.
: I have problem that watchdog reset too quickly the board.
: I know that the longest time of 860 reset time is about 3 seconds.
: I use vxWorks.st_rom.hex, then during the time of decompression, board is
: reset.
: I look forward to response answer from expert.
: Thanks..

        A easy method: modify your Bootstrap code.

        Change your inflate() function to periodically

        Reset software watchdog timer.


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: endPollSend, endPollReceive usage ?
Date: Wed, 23 Aug 2000 11:39:36 +0200
From: Leonid Rosenboim 
Organization: BitBand Technologies Ltd. http://www.bitband.com
Message-ID: <39A39BD8.CD5CBB73@bitband.com>
References: <8nvjna$sf5$1@news.monmouth.com>
Reply-To: leonid@bitband.com

The polling mode routines are used when you attach WDB (the tornado debug
agent)
directly on top of END driver, and enter the target into system debug mode.
When happens in system (aka external) debug mode, is that when a breakpoint is
hit,
the entire system is stopped by means of disabling all interrupts and then WDB
enters
the polling mode to be able to continue communications with the Host's target
server.

You will find more details about System Debugging and particularly System
debug over
Ethernet in Tornado users manual.

HTH,
Leonid

Sandhya Potluri wrote:

> Hi,
>     I am testing a ethernet driver for an ethernet driver on 860T. Has
> anybody used the endPollSend and endPollReceive routine? Help appreciated.
> Thanks
> Sandy


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Value of PCI Card
Date: Wed, 23 Aug 2000 12:03:44 +0100
From: "Kai Mattern" 
Organization: Heidelberger Druckmaschinen AG
Message-ID: <8o07fa$tsl$1@news.hdpp.de>
References: <8ntbpt$nbm$1@news.hdpp.de> <39A2B5AF.8E668DFA@gigabitwireless.com>

I now got the value i expected. This is first time useing Pci bus and the
Developer of the Hardware told me the value is in the Register after
booting. After many hours i decided to take a look in the PCI Specification
myself. And there i found i have to write FFFF FFFF to the register and then
read it back. *bonks hadware man on head* :-) at the end its my fault.
But know there is the next problem. I have a mtx603 target and autoconfig
seams only to work with 750 target (maybe im wrong ?). So i have to do it
myself. I put this code into the sysHwInit.

Because i already know the values the wirst 2 lines for every configuration
is with no effekt ..i added them just for me to control the value

 (void)pciConfigOutLong(pciBusNo, pciDevNo, pciFuncNo, 0x10, 0xFFFFFFFF);
/* writing '1' to register */
 (void)pciConfigInLong(pciBusNo, pciDevNo, pciFuncNo, 0x10, & config_daten);
/* read it back*/
 (void)pciConfigOutLong(pciBusNo, pciDevNo, pciFuncNo, 0x10, 0x'startadress
in mem space');

 (void)pciConfigOutLong(pciBusNo, pciDevNo, pciFuncNo, 0x14, 0xFFFFFFFF);
/* writing '1' to register */
 (void)pciConfigInLong(pciBusNo, pciDevNo, pciFuncNo, 0x14, & config_daten);
/* read it back */
 (void)pciConfigOutLong(pciBusNo, pciDevNo, pciFuncNo, 0x14, 0x'startadress
in io space');

 (void)pciConfigOutLong(pciBusNo, pciDevNo, pciFuncNo, 0x18, 0xFFFFFFFF);
/* writing '1' to register */
 (void)pciConfigInLong(pciBusNo, pciDevNo, pciFuncNo, 0x18, & config_daten);
/* read it back */
 (void)pciConfigOutLong(pciBusNo, pciDevNo, pciFuncNo, 0x18, 0x'startadress
in mem space');   /* 256MB Ram */

 (void)pciConfigOutLong(pciBusNo, pciDevNo, pciFuncNo, 0x1C, 0xFFFFFFFF);
/* writing '1' to register */
 (void)pciConfigInLong(pciBusNo, pciDevNo, pciFuncNo, 0x1C, & config_daten);
/* read it back */
 (void)pciConfigOutLong(pciBusNo, pciDevNo, pciFuncNo, 0x1C, 0x'startadress
in mem space');  /*800 Byte of Configdata */

Because the Prep memory model from raven chip already maped 1GByte for io
and another 1GB for mem. i assigned startadresses.
But everytime i try to write in.......i get an exception.

Is there something else i need to configure to get my card running ?
This is a snip from the ravenpci.c

    MPC Address Range   PCI Address Range      Definition
   -----------------   -----------------   -----------------
   80000000 BF7FFFFF   00000000 3F7FFFFF   ISA/PCI I/O space
/*here i put in my io adresses .....range is 1GB*/
   C0000000 FCFFFFFF   00000000 3CFFFFFF   ISA/PCI Mem space w/MPIC    /*
here i put in my mem adresses range is 1GB - 64Mb*/
   FD000000 FDFFFFFF   00000000 00FFFFFF   ISA/PCI Mem space
/* range only 16MB.....maybe reconfigure ? */
   FE000000 FE7FFFFF   00000000 007FFFFF   ISA/PCI I/O space

What does this w/MPIC mean ? i think it could mean 'write over Motorola
Programable Interrupt Controller' ???
So maybe im in the wrong memspace.....?
I checked the other devices (IDE and SCSI controller and network
controller...i thought maybe i have a collision...but with no result.....

Do you have any clue ?
Thanks a lot in advance

Kai
Kai_Mattern@web.de

"James Chen"  wrote in message
news:39A2B5AF.8E668DFA@gigabitwireless.com...
> How did you "read" the PCI configuration register space? As you would do
in
> reading a memory location?  The configuration register space must be read
using
> PCI configuration transaction cycle. It is certainly different from a
regular
> memory read.
>
> Kai Mattern wrote:
>
> > Good morning all,
> >
> > i have a PCI card who wanna have 256MB ram mapped. After power up the
EEPROM
> > writes a value into the PCI configuration header. There should (in this
> > case) a 0xF000 0000 inside the configuration header. But all time i read
> > them out they are 0 ( all 4 Base Adresses). I tried to read them out
after
> > startup in syslib.c (function sysHwInit) but with no result (allways 0).
> > Card is running fine under WinNT.....
> >
> > Anybody have a tip ?
> >
> > Maybe im in the wrong function...wrong file...to late or something
> > else....without configuring the card i cant talk to it.
> >
> > Thanks a lot
> > Kai
>



---------------------------

Newsgroups: comp.os.vxworks
Subject: Search for LAN driver for SiS900 in a SiS630
Date: Wed, 23 Aug 2000 10:24:07 GMT
From: HAROLD STOUT 
Organization: @Home Network Member
Message-ID: <39A3A644.20E6DB0D@home.com>

Hi

Has anyone used or developed a driver for the Ethernet controller
embedded in SiS630 ,called SiS900, under VxWorks? 

The Manufacturer of the chip is Silicon Integrated Systems Corp.

We are looking at an Intel processor based board which uses this chip
and I am looking for an ethernet driver and any other information.

Any help welcomed.

Hal Stout

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Multicast on VxWorks
Date: Wed, 23 Aug 2000 19:33:28 +0900
From: "Sung Won Ho" 
Organization: ELIMnet Internet News
Message-ID: <8o09fm$vbp$1@news.elim.net>
References: <8nv98c$coh$1@news.elim.net> <39A37CF2.709EC383@meco.nl>

Thanks a lot, Albert!

Though not solved yet..
Your comment greatly helped me.

Best regards,

"Albert Wijnja"  wrote in message
news:39A37CF2.709EC383@meco.nl...
> Sung,
>
> I use multicasting for TAO. I had some problems in the beginning. At one
> point I added a route and from then on multicasting ran smoothly. If the
> IP-address of your card is e.g. 90.0.0.50 add a route with the following
> command:
>
>     routeAdd "224.0.0.0", "90.0.0.50"
>
> The multicasting IP address range runs from 224.0.0.0 to 239.255.255.255.
> See "VxWorks Network Programmers Manual".
>
> Albert
>
>
> Sung Won Ho wrote:
>
> > Hi all,
> >
> > Does anyone have an experience in using multicast feature in VxWorks?
> > I use T2 and END driver.
> > But it seemed that it can't process multicast packet.
> >
> > Do you have any comments?
> >
> > Sung.
>



---------------------------

Newsgroups: comp.os.vxworks
Subject: VxWorks/Tornado shell checksum and ls utilities
Date: Wed, 23 Aug 2000 12:08:44 -0400
From: "Straub, Robert" 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <8o0t9r$oth$1@overload.lbl.gov>

Fellow VxWorkers,

I am trying to find some UNIX-like utilities that will run from the VxWorks
Tornado or target shell.

- - One is a directory listing utility.  I want it to display a recursive
listing of files in a dosFs directory tree structure, in a similar way to
the UNIX command, "ls -lR".

- - The other is a file checksum utility that will display checksums for one
or more files in a dosFs directory, in a similar way to the UNIX command
"sum *".  If this program has a "recursive" option, to sum all files in a
tree structure, so much the better.  It would also be helpful if there were
a companion program that would run the same algorithm on files in a W*****s
host.

The goal is to gain confidence that files which were downloaded to a dosFs
file system in flash memory (True FFS) have not become corrupted.

Thanks in advance for any help you can provide.

Bob Straub.

If this is a duplicate posting on this subject, I apologize.  I tried to
post though deja.com / comp.os/vxworks yesterday and had trouble (I think).

Robert J. Straub
Principal Software Engineer
Telephonics Corporation
815 Broad Hollow Road, Mail Stop 3
Farmingdale, NY 11735

516/631-755-7769 (Voice)
516/631-755-7642 (Fax)

straub@telephonics.com



---------------------------

Newsgroups: comp.os.vxworks
Subject: third party VxWorks shell / extensions
Date: Wed, 23 Aug 2000 15:41:19 GMT
From: bob_straub@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8o0raq$nja$1@nnrp1.deja.com>

Does anyone know of a third-party replacement for
the tornado or target shells, or expanded command
sets for either?

I am specifically interested in two commands that
would work as in UNIX:

- - A utility which displays a recursive listing of
a complete directory hierarchy, working like "ls -
lR" in UNIX.

- - A utility which calculates checksums for all of
the files in a directory, working like "sum *" in
UNIX.  It would be nice for this one to have a "-
R" option, too.  It would also be nice to have a
W*****s utility to do the same on a host PC, or
at least to have the algorithm so that utility
could be written.

Thanks in advance for any help,

Bob Straub
Telephonics Corp.
Farmingdale, NY


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: TrueFFS
Date: Wed, 23 Aug 2000 19:55:34 +0200
From: "Y.H" 
Message-ID: <39a4022a@news.barak.net.il>

Does anybody knows if est860mdprpo board support TrueFFS. WindRiver support
, didn't come up with a clear answer .
Thanks

Yaron



---------------------------

Newsgroups: comp.os.vxworks
Subject: mvme2700 rev level?
Date: Wed, 23 Aug 2000 08:47:02 -0400
From: Al Johnston 
Organization: MindSpring Enterprises
Message-ID: <39A3C7C6.1AD41B3C@mindspring.com>

Anyone out there understand the defects in the universe and
falcon chips that are mentioned in the target.nr file for this
cards bsp?

Are there any newer versions of this card that use the universe IIB/
Falcon 3?  The cards I have received so far have been "11B" and
"11C" versions, both of which seem to have the universe II chip
on them rather than the IIB.

I am very concerned that the defects mentioned in the target.nr
file (mainly atomic access issues) are going to cause me 
problems.

Anyone have any information / experience they can share with
me?

Yes, I have ask moto... but the cards may be obsolete before
we get an answers -g-.

thanks.

---------------------------

Newsgroups: comp.os.vxworks
Subject: compiler upgrade with Tornado
Date: Wed, 23 Aug 2000 09:45:50 -0400
From: Yi Qin 
Organization: Storm Internet Services
Message-ID: <39A3D58E.FA65747C@ss8networks.com>
Sender: ss8@gw-ss8networks.storm.ca

I'm compiling some c++ code on VxWorks and I encountered
two problems. One is that the compiler does not support
name space, and the other is that template class pointer
is not handled properly. It's a lot of work to change all
the code to fix the two problems.

One simple solution is to upgrade the VxWorks compiler.
I have compiled the code successfully on Solaris using
gcc 2.95.2. I would like to know if there is a patch
to upgrade the compiler. If not is it possible to upgrade
the compiler by ourselves ?

We are using VxWorks 5.4 and the built in compiler is
gcc version egcs-2.90.29 tornado 2.0 980515 (egcs-1.0.3
release) for vxsim and gcc driver version cygnus-2.7.2-960126 for
ppc860.

I'll really appreciate your help.

Regards,

Yi Qin


---------------------------

Newsgroups: comp.os.vxworks
Subject: Loading Object Module
Date: Wed, 23 Aug 2000 18:42:33 GMT
From: hugoitt@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8o15uk$54r$1@nnrp1.deja.com>

I get the following error when I attempt to load an object module.

Undefined symbols:
__eabi
Warning: object module may not be usable because of undefined symbols.

Does anyone know what the symbol __eabi references?


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Pcmcia! How to start?
Date: Wed, 23 Aug 2000 12:19:28 +0800
From: "kvd2" 
Organization: DCI HiNet
Message-ID: <8o0n4h$l7o@netnews.hinet.net>

Dear all,

I got some problem..
1.How to enable the pcmcia driver?How to start?
2.How to read the data in CF card? set up file system and link with VxWorks?

My target is ppc860.

Kvd2










---------------------------

Newsgroups: comp.os.vxworks
Subject: Installation of PMC-Octal Pro-232 on MVME2306 
Date: Wed, 23 Aug 2000 15:01:11 -0400
From: kostizt 
Organization: Raytheon Company
Message-ID: <39A41F77.2631AE@res.raytheon.com>

Having trouble getting serial channels to work.  xpciShow yields BAR
contents that are essentially zeros.  "devs" shows all 8 octal
channels.  Has anyone done a similar installation?


---------------------------

Newsgroups: comp.os.vxworks
Subject: Multiple symbol tables
Date: Wed, 23 Aug 2000 18:57:43 GMT
From: gpbeck@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8o16qq$6g9$1@nnrp1.deja.com>

Does anyone know how to load more that one symbol table with the
Tornado debugger?  I have more than one object running, but the objects
were not loaded with the ld command.

I tried using the add-symbol-file command from the GDB command line,
but I get the error message...

Cannot file object module on target: LOADER_OBJ_MODULE_NOT_FOUND

Any ideas.  Thanks.


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Nt build issue
Date: Wed, 23 Aug 2000 10:15:52 -0400
From: Paul Oliveira 
Organization: A thing of the past
Message-ID: <39A3DC98.65BD05EF@yahoo.com>

Hi

We have just ported our compile environment from unix to nt.  When I
build the vxWorks.st image on NT it works fine as long as the depend.xxx
file is present.  If the file is missing (like after a make clean) the
build fails with the error

make: *** [depend.xxx] Error 0x21
make: *** Deleting file 'depend.xxx'

Any suggestions?  We are using tornado 1.0.1.  The build works fine on
unix either way.

Thanks
paul


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Multicast on VxWorks
Date: Wed, 23 Aug 2000 19:42:06 GMT
From: rjdejoie@kmrmail.kmr.ll.mit.edu (Bob DeJoie)
Message-ID: <39a428b9.2303323454@news>
References: <8nv98c$coh$1@news.elim.net>

When I converted an application that was using multicast from T1 to T2
it didn't work initialiiy because in T2 IGMP wasn't enabled. 

Do you explicitly include IGMP?



On Wed, 23 Aug 2000 10:23:33 +0900, "Sung Won Ho" 
wrote:

>Hi all,
>
>Does anyone have an experience in using multicast feature in VxWorks?
>I use T2 and END driver.
>But it seemed that it can't process multicast packet.
>
>Do you have any comments?
>
>Sung.
>
>
>

================================================
Bob DeJoie      
Kwajalein Missile Range
Republic of the Marshall Islands
email: rjdejoie@kmrmail.kmr.ll.mit.edu    
phone: 805-355-5855   fax: 805-355-3833
Time Zone: GMT +12 hours                     
================================================

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Pcmcia! How to start?
Date: 23 Aug 2000 14:51:58 GMT
From: Geir Frode Raanes 
Organization: UNINETT news service
Message-ID: <8o0oee$8q9$1@snipp.uninett.no>
References: <8o0n4i$l7o@netnews.hinet.net>
Sender: Geir Frode Raanes 

kvd2  wrote:
: Dear all,

: I got some problem..
: 1.How to enable the pcmcia driver?

You include it in the Tornado II kernel config tool.

: 2.How to read the data in CF card? 
: My target is ppc860.

You don't. 

The MBX860 BSP does _not_ support ATA drives
(wich  CF emulates) on the PCMCIA port. 
This BSP only supports SRAM devices. 

- --
  ******************************************************
  Never ever underestimate the power of human stupidity.
  -Robert Anson Heinlein

		GeirFRS@invalid.and.so.forth
  ******************************************************

---------------------------

Newsgroups: comp.os.vxworks
Subject: Strange interrupt behavior noticed with WindView
Date: Wed, 23 Aug 2000 10:34:25 -0400
From: Frederick Byrne 
Organization: Lockheed Martin - GES
Message-ID: <39A3E0F1.37C79E65@lmco.com>

We've noticed in a Windview display that occasionally an interrupt handler (it
appears to be a vxWorks handler and not one we've defined) will keep the CPU
until the OS clock tick has expired (16.6 mS).  The specific interrupt number
changes but the behavior is the same.  Does anyone have any ideas?  Thanks in
advance for any help.

- -- 
==============================================================
	Frederick "Rick" Byrne	frederick.j.byrne@lmco.com
	Lead Member of the Engineering Staff
	Lockheed Martin NE&SS, Moorestown NJ
	Phone: (856) 722-7344	Fax: (856) 722-4103

---------------------------

Newsgroups: comp.realtime,comp.os.qnx,comp.os.vxworks
Subject: Re: Need Database, RTOS and File System developers
Date: Wed, 23 Aug 2000 07:43:35 -0700
From: Henry5th 
Organization: http://www.remarq.com: The World's Usenet/Discussions Start Here
Message-ID: <11f733ec.1216ad93@usw-ex0108-061.remarq.com>
References:  

Do you need any embedded software engineers working on a
real time environment?


* Sent from AltaVista http://www.altavista.com Where you can also find related Web Pages, Images, Audios, Videos, News, and Shopping.  Smart is Beautiful

---------------------------

Newsgroups: comp.os.vxworks
Subject: Problem with ls120 Panasonic
Date: Wed, 23 Aug 2000 17:17:33 +0200
From: "ADL systeme" 
Organization: ADL systeme
Message-ID: <8o0q1h$3dk$2@wanadoo.fr>

Hello
I have a problem to build driver for ls120 panasonic in vxworks.
The command REQUEST SENSE (03H) does'nt generate an interrupt.
If you have source code about this subject or solution
thank you
Didier Police    email didier.police@wanadoo.fr






---------------------------

Newsgroups: comp.os.vxworks
Subject: Bizarre page faults
Date: Wed, 23 Aug 2000 15:30:19 GMT
From: tph@pcisys.no.spam.dammit.net (Tom "Tom" Harrington)
Organization: Tigers on the Roof, Inc.
Message-ID: 
Sender: Tom Harrington 

I've been getting some page faults on VxWorks that I haven't been able
to diagnose.  This particular system has a number of threads all doing
the same thing, handling an incoming network connection and then
sleeping until a new one appears (normally they wouldn't be doing
_exactly_ the same thing, but for debugging I can make sure that all
incoming requests are the same).

Every so often-- after hudreds of thousands of requests have been
served correctly-- one of them page faults.  The faults have a number
of unusual (in my experience) details:

* There's no stack trace.  Attach CrossWind and it shows the thread
  in a function called cachewait() (which is not in my code anywhere),
  and the backtrace window just says "0 - unknown".  Doing a ti() in
  Windshell gives a memory access error.
* Examining the thread in WindShell shows nothing at all unusual about
  the thread, except for one thing: The PC is obviously garbage, being
  some impossibly huge number.  Everything else (including, for
  example, the stack pointer) is typical of one of these threads.
* Unlike most of my page faults, this does not disrupt the rest of the
  system in any way at all.  The one thread stops running, but there
  is no other impact.

I thought maybe a stack overflow, but the fact that the threads are OK
better than 99% of the time seems to rule that out.  There is some
shared memory among these threads, but access is blocked by mutexes to
ensure consistency.

These threads can be fairly long-lived, although in the cases I've
seen it hasn't been more than 10 minutes or so before a fault occurs.

Any clues?  How could the PC get corrupted on VxWorks, anyway?  That
seems like it's got to be relevant, but I don't know how.

- -- 
Tom "Tom" Harrington ------------ Decode to email: tph (at) pcisys (dot) net
  "And you may ask yourself, well, how did I get there?" -- Talking Heads
        Watch this space, spiffy home page, "Toast Talk" coming soon

---------------------------

Newsgroups: comp.os.vxworks
Subject: Pcmcia! How to start?
Date: Wed, 23 Aug 2000 22:23:28 +0800
From: "kvd2" 
Organization: DCI HiNet
Message-ID: <8o0n4i$l7o@netnews.hinet.net>

Dear all,

I got some problem..
1.How to enable the pcmcia driver?How to start?
2.How to read the data in CF card? set up file system and link with VxWorks?

My target is ppc860.

Kvd2












---------------------------

Newsgroups: comp.os.vxworks
Subject: security between 2 networks
Date: Wed, 23 Aug 2000 18:18:38 +0200
From: Sebastien Fauris 
Organization: Club-Internet (France)
Message-ID: <39A3F95E.DF67574E@club-internet.fr>

Hi vxWorkers,

I have a vxWorks target with 2 ethernet interfaces, each one on a
different
network. My target can access both networks, but I don't want that a
machine on a network can access to a machine on the other network.
What can I do on my vxWorks target to prevent/block/secure this ?

Thanks,

Sebastien.


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Multiple Developers on same project (How to?)
Date: Wed, 23 Aug 2000 10:22:40 +0200
From: "Martin Raabe" 
Organization: T-Online
Message-ID: <8o01q7$m98$14$1@news.t-online.com>
References: 

Hello Joe,
this typical scenario is supported by a TorndaoII add on product.
It is called SNiFF+ for Tornado and can be purchased through the WindRiver
Sales
force.

To find out about it, you may refer to www.WindRiver.com and search for
"SNiFF+".
You end up at http://www.takefive.com/products/sniff+.html
You may download a free copy of SNiFF+ for evaluation purposes (40MB).
The download contains a license for up to 200 files in your project, so you
may
test your scenario as described in the Tutorial, which comes with the
product.

SNiFF+ is available for all TornadoII target CPUs and host platforms.

Ciao


- --
(Watch out, I moved to a new office!)

Martin Raabe
Project Manager
Integration of SNiFF+ in Tornado II and pRISM+
TakeFive Software GmbH ,  a Wind River Company
Carl-Zeiss-Strasse 41
55129 Mainz
T: (+49) 170 / 1838405
F: (+49) 6131 / 35840-99
M: (+49) 170 / 1838405
E: Martin.Raabe@WindRiver.com
W: www.WindRiver.com
W: www.takefive.co.at/faq

Joe Ott  schrieb in im Newsbeitrag:
VpAo5.16034$Y51.140387@iad-read.news.verio.net...
>
> Tornado is a new developement environment to us and we are in the
beginning
> phases of a project with multiple developers. The project
> (Tornado-workspace) will be seperated into various sub-projects
> (Tornado-projects). Our concerns are:
>
> 1) There will be multiple developers working on different pieces of the
same
> sub-project. The work of developer A will not interfer with the work of
> developer B but the combined work will compile to a single task. What are
> some methods that can be used to allow several developers to work on their
> seperate areas and still allow them to compile and test without waiting
for
> the other developer to be finished?
>
> 2) The various sub-projects (Tornado-projects) reflect various SW layers.
In
> order to test the upper layers working copies of the lower layers are
> needed. What is the best way to support the upper layers while modifying
the
> lower layers? In other words what is the best way to keep a copy of a
> working downloadable image (*.out file) other can use while you are making
> changes to the project?
>
> 3) We will be using source control. Beside our source code, what if any
> Tornado files (workspace, project, makefile, etc) should be source
> controlled?
>
> One problem we have encounter with the Tornado project and workspace files
> is that if there are two developers working and they both make changes to
> the files (such as adding projects or adding files to the project) the
last
> one to save wins. That is the changes made by the first one who saved the
> files are lost.
>
> Has anyone worked in an environment where there were many developers and
you
> had to keep a working downloadable image for other to use while making
> changes to your own project?
>
> Any help or insight would be appreciated.
>
> Joe Ott
> VTG, an Intel Dialogic Company
>
>
>







---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Bizarre page faults
Date: Wed, 23 Aug 2000 15:41:44 GMT
From: tph@pcisys.no.spam.dammit.net (Tom "Tom" Harrington)
Organization: Tigers on the Roof, Inc.
Message-ID: 
References: 
Sender: Tom Harrington 

From a remote bunker, Tom "Tom" Harrington issued the following manifesto:
> * There's no stack trace.  Attach CrossWind and it shows the thread
>   in a function called cachewait() (which is not in my code anywhere),
>   and the backtrace window just says "0 - unknown".  Doing a ti() in
>   Windshell gives a memory access error.

A correction-- I meant to say that tt() gives this error, i.e. that I
cannot get a stack trace in Windshell.  The ti() command works.

- -- 
Tom "Tom" Harrington ------------ Decode to email: tph (at) pcisys (dot) net
  "And you may ask yourself, well, how did I get there?" -- Talking Heads
        Watch this space, spiffy home page, "Toast Talk" coming soon

---------------------------

Newsgroups: comp.os.vxworks
Subject: boot networking  problems
Date: Wed, 23 Aug 2000 19:01:23 +0200
From: "zbody" 
Organization: Slovenija OnLine - SiOL
Message-ID: 

Hi!

I have error report by the booting of the vxWorks.

Everything went good until usrRoot() - (tRootTask) : muxLoad failed!

I made my own boot image and included all network facilities -
INCLUDE_NETWORK
INCLUDE_NET_INIT
INCLUDE_NET_SYM_TBL
INCLUDE_NFS
INCLUDE_RPC
I included the right one params (I think) for the network card but the error
mentioned above still persists!

Thanks for the answer!

zbody



---------------------------

Newsgroups: comp.os.vxworks
Subject: 16550 RS232 dirver for x86
Date: Wed, 23 Aug 2000 11:13:24 +0100
From: "Wade Oram" 
Organization: IFR Ltd., Stevenage, Herts, England
Message-ID: <8o07tf$7s2$1@ipgw.ifrinternational.co.uk>

Hello,

I am looking for a 16550 RS232 driver that I can use as a drop in
replacement for the 8250 driver supplied as part of the standard x86 BSP's.
I have tried to adapt the generic 16550 driver supplied for use in an x86
architecture (the generic driver assumes memory mapped ports) and have had a
degree of success in that I can output to RS232 but as yet I am unable in
input from RS232.

If anyone has any one has any ideas, or has solved this problem for
themselves I would like to here from them.

- --
Wade Oram



---------------------------

Newsgroups: comp.os.vxworks
Subject: C++ Execptions in VxWorks
Date: Wed, 23 Aug 2000 11:00:41 +0100
From: "Wade Oram" 
Organization: IFR Ltd., Stevenage, Herts, England
Message-ID: <8o075g$7iq$1@ipgw.ifrinternational.co.uk>

Hello,

I am currenly experiencing a problem with some code whereby a C++ exception
causes a General Protection Fault when optimisation (O1 or O2) is turned on.
When Optimisation is turned off, no problem is encountered.
However, it is only when execptions are caught in one particular case that
the problem occurs. When stepping through the code with the debugger, I can
observe the exception being thrown and a being caught. However, as the
functions return, thereafter, I get to a point where the this pointer is
corrupted. Performing a subsequent 'delete' operation on a member pointer
then causes the Protection Fault.

I thought at first that it would be a problem with the destructors of the
objects on the stack causing stack corruption so I replaced the try-catch
block that catches the execption with code of the following form:

try
{
    throw QueryError(...);
}
catch (Error& e)
{
    ...
}

where QueryError is derived from Error.

This code still exhibits the problem.

Elsewhere in our application, Errors (and QueryErrors) are thrown and caught
with no problems. Both Error and QueryError are composed of simple types
with no associated destructors.

If anyone has any ideas as to what might be going on I would be most
grateful for their contributions.

- --
Wade Oram



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: MOT FCC FEC
Date: 23 Aug 2000 13:26:57 -0500
From: vanbaren@falcon.si.com (vanbaren_gerald)
Organization: Smiths Industries
Message-ID: <39a40961@news.si.com>
References:  <39A0E296.4B2C48D7@NOSPAM.bcs.org.uk>

Graham Baxter  writes:

>VxWorkers,

>We are using the Motorola Fast Ethernet Controler driver on the PPC
>8260.

>When we ping our target, about 4% of the pings are lost.

>Does anyone have any advice? Should we increase the resources (buffers,
>BDs etc)?

>We have tried a direct target to host connection and get the same
>results.


>Thanks for any help.


>Graham Baxter
>gbaxter@NOSPAMbcs.org.uk


What CPM clock are you running?  According to the User's Manual, it
must be at least 100MHz to keep up with 100bT.  The EST SBC8260 board
ships with the CPM clock running at 66MHz (it may have changed in the
last couple of months, I've complained a couple of times but have not
bought a new board lately).

My experience was that the SBC8260 dropped packets frequently when it
was miss-clocked.  When I increased the CPM clock, the 8260 ran great.
It made a major difference in TFTP timing.  It made a less significant
difference in FTP timing since TCP had a quicker timeout/retransmission
than TFTP/UDP.  Running TFTP, I would seen very short bursts with 3
second dead times in between (as seen on the etherLEDs).

gvb

- --
+---------------------------------------------------------------------------+
| Jerry Van Baren / vanbaren_gerald@si.com / Grand Rapids Mi / 616-241-7973 |
|   My employer is a company.  Companies are artifacts of a legal system.   |
|________________Artifacts are incapable of having opinions.________________|

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Adventnet SNMP agent for vxworks
Date: Wed, 23 Aug 2000 23:15:40 GMT
From: pintu_12@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8o1lub$otq$1@nnrp1.deja.com>
References: <3999AE69.AF526782@yahoo.com> <8nemkb$cl5$1@nnrp1.deja.com> <399D42CC.23F50FAF@yahoo.com> <8nrfls$fti$1@nnrp1.deja.com>



 Hi,

   Apparently the advent net SNMP agent currently provides support
 only for SNMP group counters, support for the Mib-2 variables would
 be provided later.

  Also could any one of you let me know as to whether the Advent
 Net SNMP agent would be compatible with TCP/IP stack of VxWorks
 (as I am not very sure whether there SNMP agent could be integerated
 with the VxWorks TCP/IP stack)

  Randy, which SNMP agent of VxWorks did you try using - Envoy 8.0
 (better known as WindNet SNMP) or Envoy 9.1 ? An alternative to
 Envoy could be the emanate SNMP agent. You have said that the
 documentation for the wind river SNMP agent was good - which SNMP
 agent are you reffering to - the Envoy 8.0 or the Envoy 9.1. I am
 trying to use Envoy 9.1 and I am not finding the documentation to be
 sufficient. In case you feel that the documentation for Envoy 8.0 is
 good please let me know,

 Regards,
 Piyush


In article <8nrfls$fti$1@nnrp1.deja.com>,
  randyj5573@my-deja.com wrote:
> My experiences with Wind River's SNMP agent has been extreemly
> frustrating. I have found that WRS tech support in other area's of
> Vxworks to be pretty good but their tech support for their SNMP agent
> to be almost non-existant. It seems that WRS's general tech suport has
> almost no experience with SNMP. They purchased their current agent
from
> a third party. WRS has since purchased the company that wrote their
> agent but there has yet to be any transfer of knowledge out to the
tech
> support people. WRS has fairly good general documentation about their
> SNMP agent but when it came time to write some code, I came up with
> questions on how some of their functions worked, what some of their
> internal structures looked like and questions about undocumented
> functions and macros that their mib compiler would use its generated
> code. I quickly came to the limit of what help their tech support
could
> provide and found myself teaching them so that maybe they could
> research a question to provide an answer. One question on traps took
> over three weeks to get an answer. The one thing that helped me was
> finding some programming examples on their web site. I was able to
> compile them and then step through them in order to figure out how
they
> work.
>
> Well enough venting. This was my 1st time at trying to extend an SNMP
> agent so I have no experience with other tool vendors to compare to.
>
> Regards,
> Randy Jones
>
> In article <399D42CC.23F50FAF@yahoo.com>,
>   Bhavesh Patel  wrote:
> > we plan to use both jmx agent and snmp embedded C agent, this is the
> reason
> > the architects here choose adventnet snmp toolkit. How are your
> experiences
> > with wind river's snmp agent?
> > Thanks for your response.
> >
> > bhavesh
> > randyj5573@my-deja.com wrote:
> >
> > > Why don't you use Wind River's SNMP agent?
> > >
> > > Randy Jones
> > >
> > > In article <3999AE69.AF526782@yahoo.com>,
> > >   Bhavesh Patel  wrote:
> > > > Hi,
> > > >
> > > > Has anybody used adventnet embedded/C snmp agent for vxworks
> > > platform? I
> > > > would like to know their experiences.  Adventnet website does
not
> have
> > > > example/demo for vxworks platform.
> > > >
> > > > Thanks for any help.
> > > >
> > > > Bhavesh Patel
> > > >
> > > >
> > >
> > > Sent via Deja.com http://www.deja.com/
> > > Before you buy.
> >
> >
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.
>


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Value of PCI Card
Date: Wed, 23 Aug 2000 18:10:06 GMT
From: John Fusco 
Organization: Deja.com - Before you buy.
Message-ID: <8o141b$2pt$1@nnrp1.deja.com>
References: <8ntbpt$nbm$1@news.hdpp.de> <8nuei9$vgk$1@nnrp1.deja.com> <8o0k24$um1$1@news.hdpp.de>

In article <8o0k24$um1$1@news.hdpp.de>,
  "Kai Mattern"  wrote:
> What target do you have ? I have a Motorola 603 board and thought
autoconfig
> only run with 750 boards ?
>


The only BSP that I have that came configured with autoconfig is the
MVME-2100 BSP.  There is no reason why any BSP can't use it though, It's
a generic library.  Which is why WRS included it with the installation I
suppose. It has nothing to do with the processor.


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Hi, LINUX HERE!!!!
Date: Thu, 24 Aug 2000 02:27:24 GMT
From: phgsvp@e.com
Organization: Clientes_Teleline
Message-ID: 

Go to: http://www.lanzadera.com/revistapandora
zjkdrrpiygtjooupednnfbtxxgfrqlzldriupzgmonxbckbslngjtdixzrbykdu


---------------------------

Newsgroups: comp.os.vxworks
Subject: need help with project facility
Date: Wed, 23 Aug 2000 15:57:12 -0400
From: Bhavesh Patel 
Organization: Posted via Supernews, http://www.supernews.com
Message-ID: <39A42C98.B1E6E610@yahoo.com>

Hi all,

I have projectX in a workspaceA which is used for SIMNTgnu. Now I want
to build the same application for a target machine. So I copied the "C"
files from projectX to the new projectY. However when I look at the
properties for files in projectY they do NOT have the same custom
dependencies as for files in projectX (the dependencies are listed as
things like "-Iagent/include -Iagent/include/snmpv3 ..". Due to this the
build fails. I updated the build properties for projectX (C/C++
properties to have the above include files) even then the makefile is
not updated to take the above include file directories, as a result the
build fails since it does not find the include directory to look in. Any
solution for the above problem?
Do I need to copy files individually again from my the original source
directory instead from the projectX and update the C/C++ properties tab?



Thanks for any help.

bhavesh


---------------------------

Newsgroups: comp.os.vxworks
Subject: VxSim i/o performance unusably slow?
Date: 23 Aug 2000 12:40:19 -0700
From: thain@usr09.primenet.com (Gregory Thain)
Organization: Primenet (602)416-7000
Message-ID: 


We're trying to use VxSim on Windows NT, but find that the i/o
performance is unusably slow.  For example, a program that writes one
byte at a time through TSFS writes about 4 bytes/second.  When we
asked Wind River about this, their official answer was that "VxSim is
not designed for real time performance".

I don't need VxSim to have exactly the same performance characters
as when running real VxWorks on my board, but the i/o does need
to useable.  At this speed, there's really no point to using VxSim.

Does anyone else see this problem?  Is there some workaround?

Thanks,

- -greg
thain@primenet.com

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Loading Object Module
Date: 23 Aug 2000 20:21:38 GMT
From: johan@borksoft.xs4all.nl
Organization: Borksoft
Message-ID: <8o1boi$has$1@news1.xs4all.nl>
References: <8o15uk$54r$1@nnrp1.deja.com>

hugoitt@my-deja.com wrote:
: I get the following error when I attempt to load an object module.

: Undefined symbols:
: __eabi
: Warning: object module may not be usable because of undefined symbols.

: Does anyone know what the symbol __eabi references?

You probably have a function called main. rename this function to for
example myMainFunction, and try this again.

Groeten,
	Johan

- -- 
     o o o o o o o . . .   ___________________________________
    o      _____           ||    Johan Borkhuis              |
  .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
 >(________|__|_[_________]_|________________________________|
 _/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: need help with project facility
Date: Wed, 23 Aug 2000 22:21:13 +0200
From: "Martin Raabe" 
Organization: T-Online
Message-ID: <8o1bt7$rhu$16$1@news.t-online.com>
References: <39A42C98.B1E6E610@yahoo.com>

Hello Bhavesh,
why don't you just switch the buildspec?
Stay in the same workspace and project.

Ciao
- --
Martin Raabe




---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Multiple symbol tables
Date: 23 Aug 2000 20:23:17 GMT
From: johan@borksoft.xs4all.nl
Organization: Borksoft
Message-ID: <8o1brl$has$2@news1.xs4all.nl>
References: <8o16qq$6g9$1@nnrp1.deja.com>

gpbeck@my-deja.com wrote:
: Does anyone know how to load more that one symbol table with the
: Tornado debugger?  I have more than one object running, but the objects
: were not loaded with the ld command.

Try turning on the synchronise symbol table option in the Target server
and in your kernel configuration, or load the object using Tornado instead
of the target shell.

Groeten,
	Johan

- -- 
     o o o o o o o . . .   ___________________________________
    o      _____           ||    Johan Borkhuis              |
  .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
 >(________|__|_[_________]_|________________________________|
 _/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: Help
Date: Wed, 23 Aug 2000 16:41:56 -0500
From: "Michael S." 
Organization: MediaOne Express -=- Central Region
Message-ID: 

I cant figure out why there is no vxworks.st which the mkboot.bat bath file
is looking for. Also I am trying to figure out how to costumize your vxworks
os.Anyone have any ideas?

Sincerely,
Michael S - scikid@hotmail.com



---------------------------

Newsgroups: comp.os.vxworks
Subject: Q: How to use memPartCreate()
Date: Wed, 23 Aug 2000 22:15:08 GMT
From: wdewilde@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8o1icr$kp3$1@nnrp1.deja.com>

Hi,

I'm wondering how to use memPartCreate(), to setup a memory
partition, separate from the VxWorks system partition.
Should I configure the kernel to not know about a piece of memory,
and the later call memPartCreate() with the pointer (base address) and
size of that piece of memory ?
Or should I have system partition own all of memory, and during
initialization of the application code, malloc() a big chunk and then
create a partition out of that ?

Thanks
- - Wim


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: target server/dual ethernet devices
Date: Wed, 23 Aug 2000 22:31:38 GMT
From: eliasson@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8o1jcb$lrv$1@nnrp1.deja.com>

Hello,

I'm working with a board that has 2 ethernet devices built in.  I've got
it set up so that I can boot and get the 'WDB: Ready.' message from
one device or the other.  However, I can only get a target server to
connect to device 0 and bring up a shell.  When I try to get the target
server to connect to device 1, I get rpccore back end client timed out
messages.  I've double checked the IP addresses to see that they
match.  Any ideas/help will be appreciated.


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Multiple symbol tables
Date: Wed, 23 Aug 2000 22:33:15 GMT
From: gpbeck@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8o1jfa$lsu$1@nnrp1.deja.com>
References: <8o16qq$6g9$1@nnrp1.deja.com> <8o1brl$has$2@news1.xs4all.nl>

I've tried doing just that but it requires a network port for debug.  I
can only use a serial connection.  My hardware platform has two serial
ports, one for WDB and second for a User Interface.  I've tried the
command 'add-symbol-table' from the GDB command interface but I get an
error message.  I can't remember the exact message, but it said
something about the loader module not being installed.  But that is not
true.

Any other ideas?  Any input would be appreciated.

Thanks.

In article <8o1brl$has$2@news1.xs4all.nl>,
  johan@borksoft.xs4all.nl wrote:
> gpbeck@my-deja.com wrote:
> : Does anyone know how to load more that one symbol table with the
> : Tornado debugger?  I have more than one object running, but the
objects
> : were not loaded with the ld command.
>
> Try turning on the synchronise symbol table option in the Target
server
> and in your kernel configuration, or load the object using Tornado
instead
> of the target shell.
>
> Groeten,
> 	Johan
>
> --
>      o o o o o o o . . .   ___________________________________
>     o      _____           ||    Johan Borkhuis              |
>   .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
>  >(________|__|_[_________]_|________________________________|
>  _/oo OOOOO oo`  ooo   ooo  'o!o!o                      o!o!o`
> === VxWorks FAQ: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html
===
>


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: PCI driver guidelines?
Date: Wed, 23 Aug 2000 21:26:43 -0400
From: Bruce Chrustie 
Organization: Nortel
Message-ID: <39A479D3.CA30BBE5@root.com>

Hi gang,

Planning on writing a driver for the IDT ATM NIC for the x86. Anyone
have some insight as how to do this in VxWorks? Never wrote a PCI driver
before :)

Ist step would be to see if the PCI card is there
then look at the config reg
and so on....

Searched the net for guidelines and came up empty.

Bruce,

---------------------------

Newsgroups: comp.os.vxworks,comp.protocols.snmp
Subject: Problems with tabular index of WindNet SNMP Method Routines
Date: Thu, 24 Aug 2000 12:59:45 +0200
From: "S. Schmidt" 
Organization: roka GmbH
Message-ID: <8o2v3h$v3s$1@news10.roka.net>

Hello,

I have problems with the tabular index discovery of WindNet SNMP method
routines. For each leaf the number of indexes is correct identified and the
discovery is finished, but the index discovery of the last leaf is not
finished.
What can I do?

Thanks





---------------------------

End of New-News digest
**********************


From vxwexplo-errs@csg.lbl.gov  Thu Aug 24 07:23:15 2000
From: Rainer Stelzer 
Date: Thu Aug 24 07:23:17 PDT 2000
Subject: Project/linker question

Hi vxworkers,

I want to link my object to my standalone vxworks-image.

The default make rule in the project management includes the ctors/dtors
call
even if it's C-Code, so the bootrom generation fails, cause it's already
munched.

I expect the Tornado II  project management to provide a solution to
link
single project objects in one workspace object, that can be linked to
the rom image.
How to ? Or do I have to write my own makefile ?

thanx in advance

Rainer

=====================================================
            Gebrüder Stoye GmbH
department: traffic control systems
      name: Rainer Stelzer
       adr: Longericher Strasse 177
      city: Cologne, Germany
       zip: 50739
     phone: ++49/221/2616-551
    mailto: r.stelzer@stoye.de
       www: http://www.stoye.de
=====================================================




From vxwexplo-errs@csg.lbl.gov  Thu Aug 24 11:53:13 2000
From: Tom Miller 
Date: Thu Aug 24 11:53:15 PDT 2000
Subject: building vxWorks using Tornado II

Hello Group,

I am trying to build a bootable vxWorks image using Tornado II and I'm
having some conceptual difficulties. I understand that I can select
different components to include in the operating system image by selecting
them in the "Workspace" window, "vxWorks" tab (for example I want to select
from the folder tree: vxWorks/hardware/peripherals/trueFFS flash file
system).  I assume that what this does is select a pre-compiled component
that gets linked into my final vxWorks image.  What I don't understand is:

1) How do I know what library these components are coming from?    say:
\target\lib\objPENTIUMgnuvx  vs  \target\lib\obj80386gunvx?

2) How do I know which "C" source files are associated with a given
component that I'm including?

3) If I want to modify a given components source file to make it fit my
particular hardware, how do I specify to use my version of the file instead
of the one from the library?


Any and all insights would be greatly appreciated.


Tom Miller
Software Engineer
Motion Engineering Inc.
Santa Barbara Ca.







From vxwexplo-errs@csg.lbl.gov  Thu Aug 24 15:44:05 2000
From: venky 
Date: Thu Aug 24 15:44:09 PDT 2000
Subject: RE: building vxWorks using Tornado II

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_01C00E1B.9C1BBEC6
Content-Type: text/plain;
	charset="iso-8859-1"

:->1) How do I know what library these components are coming 
:->from?    say:
:->\target\lib\objPENTIUMgnuvx  vs  \target\lib\obj80386gunvx?

When a project is created, it is created based on a BSP. THis 
will decide which library to choose the components from.

:->2) How do I know which "C" source files are associated with a given
:->component that I'm including?

This will be a little difficult if you dont have the entire vxWOrks source
code.
even if you do have, a manual search in the source directory is the option.

:->3) If I want to modify a given components source file to 
:->make it fit my
:->particular hardware, how do I specify to use my version of 
:->the file instead
:->of the one from the library?

You can include that particular file in the project thereby overriding the 
library version of the file ...

HTH

V.

------_=_NextPart_001_01C00E1B.9C1BBEC6
Content-Type: text/html;
	charset="iso-8859-1"






RE: building vxWorks using Tornado II




:->1) How do I know what library these components are coming 

:->from?    say:

:->\target\lib\objPENTIUMgnuvx  vs  \target\lib\obj80386gunvx?




When a project is created, it is created based on a BSP. THis 

will decide which library to choose the components from.




:->2) How do I know which "C" source files are associated with a given

:->component that I'm including?




This will be a little difficult if you dont have the entire vxWOrks source code.

even if you do have, a manual search in the source directory is the option.




:->3) If I want to modify a given components source file to 

:->make it fit my

:->particular hardware, how do I specify to use my version of 

:->the file instead

:->of the one from the library?




You can include that particular file in the project thereby overriding the 

library version of the file ...




HTH




V.





------_=_NextPart_001_01C00E1B.9C1BBEC6--


From vxwexplo-errs@csg.lbl.gov  Thu Aug 24 18:34:51 2000
From: "Qin Ting" 
Date: Thu Aug 24 18:34:54 PDT 2000
Subject: Can I use more than 16 PPP interfaces in vxworks5.4

Hi, vxworks

   From VxWorks Network Programmer's Guide(5.4), I read that vxWorks can
support Up to 16 PPP interfaces at any one time.
   Now, I want to use more than 16 PPP interfaces at one time. What can I do
to achieve it? Can the vxWorks's PPP be modified?

  Anticipate your help!

Qin Ting



From vxwexplo-errs@csg.lbl.gov  Fri Aug 25 04:03:10 2000
From: Vxworks Exploder 
Date: Fri Aug 25 04:03:12 PDT 2000
Subject: comp.os.vxworks newsdigest

Comp.Os.Vxworks Daily Digest    Fri Aug 25 04:03:05 PDT 2000

        Subject: Re: Raw IP sockets and port number
        Subject: Where have the USB driver (or reference code )?
        Subject: Re: Pcmcia! How to start?
        Subject: symbol table problems
        Subject: Re: Raw IP sockets and port number
        Subject: Help Setting DMA timeout on Universe II
        Subject: How do I use op code (Assembly Language) Instructions in the PowerPC with the VxWorks development tools?
        Subject: Project/linker question
        Subject: Re: Help Setting DMA timeout on Universe II
        Subject: Re: Pcmcia! How to start?
        Subject: Re: Help
        Subject: symbol table problems
        Subject: Re: need help with project facility
        Subject: What is error S_objLib_OBJ_ID_ERROR ?
        Subject: building vxWorks using Tornado II
        Subject: Increasing SIMNT memory
        Subject: MCPN765 from Motorola
        Subject: Re: mvme2700 rev level?
        Subject: To C or C++?
        Subject: termios.h
        Subject: Re: MCPN765 from Motorola
        Subject: Re: compiler upgrade with Tornado
        Subject: Re: Adventnet SNMP agent for vxworks
        Subject: Re: How do I use op code (Assembly Language)...
        Subject: Have you tried VxVMI?
        Subject: Re: Project/linker question
        Subject: Re: mvme2700 rev level?
        Subject: Re: Project/linker question
        Subject: RE: building vxWorks using Tornado II
        Subject: Re: To C or C++?
        Subject: post again: Some Quesions on Mux
        Subject: Write Protection for Task Stacks
        Subject: Re: Strange interrupt behavior noticed with WindView
        Subject: Re: Configure BSP fails
        Subject: Re: Help
        Subject: Re: symbol table problems
        Subject: Re: To C or C++?
        Subject: I/O device read/write
        Subject: DHCP client configuration
        Subject: Re: Increasing SIMNT memory
        Subject: Re: Increasing SIMNT memory
        Subject: Re: I/O device read/write
        Subject: Re: VxWorks/Tornado shell checksum and ls utilities
        Subject: Re: Multiple symbol tables
        Subject: Re: Write Protection for Task Stacks

-------------------------------------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Raw IP sockets and port number
Date: 24 Aug 2000 02:24:07 GMT
From: Diamand@btinternet.com (Luke Diamand)
Organization: Customer of Energis Squared
Message-ID: 
References: <399415FD.D2A33025@ftw.mot.com> <399ABB37.62164158@student.liu.se>

I have a vague feeling that SOCK_RAW isn't supported under vxWorks. 
I could be wrong, though.

Luke Diamand

>Michael White wrote:
>
>> Hi,
>>
>> I want to use a Raw IP socket (type AF_INET) to send GRE messages.
>> A couple of quick questions:
>>
>>    1)  How do I set the TOS in the IP header?
>>    2)  Is the port field in the sockaddr used on a Raw IP socket?
>>    3)  When sending a Raw IP message, what is the format of the data
>>        that needs to be passed to sendto()?  Does it include the IP
>>        header or just the GRE header and body?
>>
>> Any help would be greatly appreciated.  If it's somewhere in the VxWorks
>> manuals,
>> that'd be helpful, too.
>>
>> Thanks in advance.
>> ---
>> Michael White, qa1402@ftw.mot.com
>> Cellular Infrastructure Group
>> Ft. Worth, Texas    (817)245-6418
>


- -- 

---------------------------

Newsgroups: comp.os.vxworks
Subject: Where have the USB driver (or reference code )?
Date: Thu, 24 Aug 2000 16:00:42 +0800
From: "kvd2" 
Organization: DCI HiNet
Message-ID: <8o337m$n0k@netnews.hinet.net>

Dear All,
I wanna to enable the usb port of mpc8xx,
If any one has the experience on that?

Kvd



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Pcmcia! How to start?
Date: Thu, 24 Aug 2000 15:42:45 +0800
From: "kvd2" 
Organization: DCI HiNet
Message-ID: <8o337k$n0k@netnews.hinet.net>
References: <8o0n4i$l7o@netnews.hinet.net> <8o0oee$8q9$1@snipp.uninett.no>

How about to write a new one?

Geir Frode Raanes  wrote in message
news:8o0oee$8q9$1@snipp.uninett.no...
> kvd2  wrote:
> : Dear all,
>
> : I got some problem..
> : 1.How to enable the pcmcia driver?
>
> You include it in the Tornado II kernel config tool.
>
> : 2.How to read the data in CF card?
> : My target is ppc860.
>
> You don't.
>
> The MBX860 BSP does _not_ support ATA drives
> (wich  CF emulates) on the PCMCIA port.
> This BSP only supports SRAM devices.
>
> --
>   ******************************************************
>   Never ever underestimate the power of human stupidity.
>   -Robert Anson Heinlein
>
> GeirFRS@invalid.and.so.forth
>   ******************************************************





---------------------------

Newsgroups: comp.os.vxworks
Subject: symbol table problems
Date: Thu, 24 Aug 2000 14:47:47 +0200
From: Samuel F. Romero =?iso-8859-1?Q?Garc=EDa?= 
Organization: INTA
Message-ID: <39A51973.58655C1B@inta.es>

Hi there!

I'm a newbi on VxWorks, and I'm trying to make use of the already
defined class string (by inserting in my code #include    and so
on).

The compiler goes OK, no complaint, but once I download it to the
target... it says:

Unresolved symbols encountered while downloading:

Errors while downloading
C:/TORNADO/target/proj/SimulaTrib/SIMNTgnu/SimulaTrib.out:
_length__t18string_char_traits1ZcPCc
_clone__Q2t12basic_string3ZcZt18string_char_traits1ZcZt24__default_alloc_template2b1i03Rep

__t12basic_string3ZcZt18string_char_traits1ZcZt24__default_alloc_template2b1i0$nilRep

___dl__Q2t12basic_string3ZcZt18string_char_traits1ZcZt24__default_alloc_template2b1i03RepPv

_replace__t12basic_string3ZcZt18string_char_traits1ZcZt24__default_alloc_template2b1i0UiUiPCcUi

I'm using Tornado II for Windows 95 (NT), and I use the simulator as
target (Pentium). Thanks for your help!
- --
Samuel F. Romero García    - romerogs@inta.es
Instituto Nacional de Técnica Aeroespacial - INTA
División de Sistemas Electrónicos - Área Ing. Software
Tel. 91 520 13 49 - Crtra. Ajalvir s/n Km 4.5
Torrejón de Ardoz, 28850 - MADRID
- --
If we knew what we were doing, it wouldn't be called research... - A.
Einstein



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Raw IP sockets and port number
Date: Thu, 24 Aug 2000 09:52:47 +0530
From: Vijay Kumar Peshkar 
Organization: Ye 'Ol Disorganized NNTPCache groupie
Message-ID: <39A4A317.9FA284EE@cisco.com>
References: <399415FD.D2A33025@ftw.mot.com> <399ABB37.62164158@student.liu.se> 
Reply-To: vpeshkar@cisco.com

Michael,

RAW sockets in general and RAW IP sockets in particular should not
be bound to any port.  there is no concept of a "port" at the IP
layer.  "port" is a transport layer thing (as in TCP and UDP).
You should just use sendto() and recvfrom() to send/receive
data immediately after opening the socket.  when an IP datagram
is received, you will get a copy of the data. In fact , if you have
multiple RAW IP applications running on your system, each will
receive a copy.  the data delivered thus is delivered to a "host"
whose address is specified by the IP address, not to the particular
port.

Hope this helps.

regards,
Vijay

Luke Diamand wrote:

> I have a vague feeling that SOCK_RAW isn't supported under vxWorks.
> I could be wrong, though.
>
> Luke Diamand
>
> >Michael White wrote:
> >
> >> Hi,
> >>
> >> I want to use a Raw IP socket (type AF_INET) to send GRE messages.
> >> A couple of quick questions:
> >>
> >>    1)  How do I set the TOS in the IP header?
> >>    2)  Is the port field in the sockaddr used on a Raw IP socket?
> >>    3)  When sending a Raw IP message, what is the format of the data
> >>        that needs to be passed to sendto()?  Does it include the IP
> >>        header or just the GRE header and body?
> >>
> >> Any help would be greatly appreciated.  If it's somewhere in the VxWorks
> >> manuals,
> >> that'd be helpful, too.
> >>
> >> Thanks in advance.
> >> ---
> >> Michael White, qa1402@ftw.mot.com
> >> Cellular Infrastructure Group
> >> Ft. Worth, Texas    (817)245-6418
> >
>
> --


---------------------------

Newsgroups: comp.os.vxworks
Subject: Help Setting DMA timeout on Universe II
Date: Thu, 24 Aug 2000 09:34:48 -0400
From: "Jim Pisano" 
Organization: NRAO
Message-ID: <8o38ir$9f0$1@newsfeed.cv.nrao.edu>

I have an MVME-2700 w/ a Universe II DMA controller & Tornado II.

I have a custom-built memory FIFO card on the VME bus which performs DMA
transfers to the 2700. Due to some modifications to this card, it can take
up to 1 ms to start transferring data across the VME bus. I also get a DMA
VME bus error which I think is a result of the DMA timing out waiting for
data.

I have read the Tundra Universe II manual regarding the DMA Controller
(section 2.8) , but I can't figure out how to adjust the timeout delay. The
only thing that I can see is the VON register which is set to 0 ("until
done")


I'm using the standard DMA driver from WRS which uses the direct-mode
operation, is polled & not interrupt driven (which is okay for now -- I'll
deal w/ an interrupt-driven version later).

Any help on this would be greatly appreciated.

- -Jim

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-==-=
Jim Pisano
Scientific Programming Analyst, ALMA Correlator Group
NRAO Central Development Lab
2015 Ivy Road                          Fax: 804-296-0324
Charlottesville VA, 22903           Office: 804-296-0351



---------------------------

Newsgroups: comp.os.vxworks
Subject: How do I use op code (Assembly Language) Instructions in the PowerPC with the VxWorks development tools?
Date: Thu, 24 Aug 2000 15:49:22 -0500
From: "Keith McHugh" 
Organization: Esat Net Customer
Message-ID: <39a5354c.0@seat.s3two.ie>
Reply-To: "Keith McHugh" 






---------------------------

Newsgroups: comp.os.vxworks
Subject: Project/linker question
Date: Thu, 24 Aug 2000 16:13:03 +0200
From: Rainer Stelzer 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <8o3e5c$bp2$1@overload.lbl.gov>

Hi vxworkers,

I want to link my object to my standalone vxworks-image.

The default make rule in the project management includes the ctors/dtors
call
even if it's C-Code, so the bootrom generation fails, cause it's already
munched.

I expect the Tornado II  project management to provide a solution to
link
single project objects in one workspace object, that can be linked to
the rom image.
How to ? Or do I have to write my own makefile ?

thanx in advance

Rainer

=====================================================
            Gebrüder Stoye GmbH
department: traffic control systems
      name: Rainer Stelzer
       adr: Longericher Strasse 177
      city: Cologne, Germany
       zip: 50739
     phone: ++49/221/2616-551
    mailto: r.stelzer@stoye.de
       www: http://www.stoye.de
=====================================================



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Help Setting DMA timeout on Universe II
Date: Thu, 24 Aug 2000 12:56:01 -0500
From: Greg Willden 
Organization: Posted via Supernews, http://www.supernews.com
Message-ID: <39A561B1.BCFD09A2@swri.edu>
References: <8o38ir$9f0$1@newsfeed.cv.nrao.edu>

Jim Pisano wrote:

> I have an MVME-2700 w/ a Universe II DMA controller & Tornado II.
>
> I have a custom-built memory FIFO card on the VME bus which performs DMA
> transfers to the 2700. Due to some modifications to this card, it can take
> up to 1 ms to start transferring data across the VME bus. I also get a DMA
> VME bus error which I think is a result of the DMA timing out waiting for
> data.

Under what circumstances do you get the error?
Are you able to read from the FIFO card using a normal memory read?
Just some ideas.

>
>
> I have read the Tundra Universe II manual regarding the DMA Controller
> (section 2.8) , but I can't figure out how to adjust the timeout delay. The
> only thing that I can see is the VON register which is set to 0 ("until
> done")
>

My understanding:  the VON value is the number of bytes that the DMA controller
will transfer before allowing itself to give up control of the VME bus.
Universe.c in your BSP contains information about the dma copy functions.
Just a few ideas.  HTH
Greg


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Pcmcia! How to start?
Date: 24 Aug 2000 12:17:06 GMT
From: Geir Frode Raanes 
Organization: UNINETT news service
Message-ID: <8o33o2$p57$1@snipp.uninett.no>
References: <8o0n4i$l7o@netnews.hinet.net> <8o0oee$8q9$1@snipp.uninett.no> <8o337k$n0k@netnews.hinet.net>
Sender: Geir Frode Raanes 

kvd2  wrote:
: How about to write a new one?

There is a reason why WindRiver left this out.
If you want to write or port a complete set of 
card services, feel free to do so. 
Me - not a chance.


- --
  ******************************************************
  Never ever underestimate the power of human stupidity.
  -Robert Anson Heinlein

		GeirFRS@invalid.and.so.forth
  ******************************************************

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Help
Date: Thu, 24 Aug 2000 08:44:36 -0700
From: Joe Durusau 
Organization: Lockheed Martin Corporation
Message-ID: <39A542E4.C9CCF4FC@delphi.com>
References: 

	At least in some versions, you have to request that file specifically.
In T1, you get it by 

make vxWorks.st

You customize vxWorks by changing the source code.  Most things can be
done 
by changing #define lines in configAll.h or some similar file.  I
suggest you
read the manuals with great care before changing much.

Speaking only for myself,

Joe Durusau


"Michael S." wrote:
> 
> I cant figure out why there is no vxworks.st which the mkboot.bat bath file
> is looking for. Also I am trying to figure out how to costumize your vxworks
> os.Anyone have any ideas?
> 
> Sincerely,
> Michael S - scikid@hotmail.com

---------------------------

Newsgroups: comp.os.vxworks
Subject: symbol table problems
Date: Thu, 24 Aug 2000 14:46:45 +0200
From: Samuel F. Romero =?iso-8859-1?Q?Garc=EDa?= 
Organization: INTA
Message-ID: <39A51933.D3E4AA1F@inta.es>

Hi there!

I'm a newbi on VxWorks, and I'm trying to make use of the already
defined class string (by inserting in my code #include    and so
on).

The compiler goes OK, no complaint, but once I download it to the
target... it says:

Unresolved symbols encountered while downloading:

Errors while downloading
C:/TORNADO/target/proj/SimulaTrib/SIMNTgnu/SimulaTrib.out:
_length__t18string_char_traits1ZcPCc
_clone__Q2t12basic_string3ZcZt18string_char_traits1ZcZt24__default_alloc_template2b1i03Rep

__t12basic_string3ZcZt18string_char_traits1ZcZt24__default_alloc_template2b1i0$nilRep

___dl__Q2t12basic_string3ZcZt18string_char_traits1ZcZt24__default_alloc_template2b1i03RepPv

_replace__t12basic_string3ZcZt18string_char_traits1ZcZt24__default_alloc_template2b1i0UiUiPCcUi

I'm using Tornado II for Windows 95 (NT), and I use the simulator as
target (Pentium). Thanks for your help!
- --
Samuel F. Romero García    - romerogs@inta.es
Instituto Nacional de Técnica Aeroespacial - INTA
División de Sistemas Electrónicos - Área Ing. Software
Tel. 91 520 13 49 - Crtra. Ajalvir s/n Km 4.5
Torrejón de Ardoz, 28850 - MADRID
- --
If we knew what we were doing, it wouldn't be called research... - A.
Einstein



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: need help with project facility
Date: Thu, 24 Aug 2000 09:03:23 -0400
From: Bhavesh Patel 
Organization: Posted via Supernews, http://www.supernews.com
Message-ID: <39A51D1B.8DA14319@yahoo.com>
References: <39A42C98.B1E6E610@yahoo.com> <8o1bt7$rhu$16$1@news.t-online.com>

Martin Raabe wrote:

> Hello Bhavesh,
> why don't you just switch the buildspec?
> Stay in the same workspace and project.
>
> Ciao
> --
> Martin Raabe

Thanks, you are suggestion helped to fix the problem.


---------------------------

Newsgroups: comp.os.vxworks
Subject: What is error S_objLib_OBJ_ID_ERROR ?
Date: Thu, 24 Aug 2000 14:39:44 -0400
From: "Jim Pisano" 
Organization: NRAO
Message-ID: <8o3qek$j44$1@newsfeed.cv.nrao.edu>

How does this error (S_objLib_OBJ_ID_ERROR) arise? The only reference that I
see to it is in the file target/h/objLib.h for a macro called
OBJ_VERIFY(objId, classId).

I get this error when I execute select() on the 2-nd serial port (COM2) of
an MVME-2700.  Here's a snippet of code (which works fine):

    int width;
    int status;

    int m_serialIO_fd = open("/tyCo/1", O_RDWR, 0 );

    fd_set m_serialSetFD;

    FD_ZERO(&m_serialSetFD);
    FD_SET(m_serialIO_fd, &m_serialSetFD );
    width = m_serialIO_fd + 1;

    struct timeval m_readTimeOut;

    m_readTimeOut.tv_sec = 0;
    m_readTimeOut.tv_usec = 100000; // wait 100 ms on select for slow serial
port response

    // read nBytes data into m_rawBytesBuffer array.
    int nBytes = 10;
    unsigned char m_rawBytesBuffer[100];

   // This generates the following error: S_objLib_OBJ_ID_ERROR for the
calling
   // task. Don't understand this. 8-24-00 JAP
   status = select( width, &m_serialSetFD, NULL, NULL, &m_readTimeOut);
   if( status > 0 )
   {
       // read response
       bytesRead = read(m_serialIO_fd, (char *)&m_rawBytesBuffer, nBytes);
   }
...



---------------------------

Newsgroups: comp.os.vxworks
Subject: building vxWorks using Tornado II
Date: Thu, 24 Aug 2000 11:55:43 -0700
From: Tom Miller 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <8o3s7c$3qm$1@overload.lbl.gov>

Hello Group,

I am trying to build a bootable vxWorks image using Tornado II and I'm
having some conceptual difficulties. I understand that I can select
different components to include in the operating system image by selecting
them in the "Workspace" window, "vxWorks" tab (for example I want to select
from the folder tree: vxWorks/hardware/peripherals/trueFFS flash file
system).  I assume that what this does is select a pre-compiled component
that gets linked into my final vxWorks image.  What I don't understand is:

1) How do I know what library these components are coming from?    say:
\target\lib\objPENTIUMgnuvx  vs  \target\lib\obj80386gunvx?

2) How do I know which "C" source files are associated with a given
component that I'm including?

3) If I want to modify a given components source file to make it fit my
particular hardware, how do I specify to use my version of the file instead
of the one from the library?


Any and all insights would be greatly appreciated.


Tom Miller
Software Engineer
Motion Engineering Inc.
Santa Barbara Ca.






---------------------------

Newsgroups: comp.os.vxworks
Subject: Increasing SIMNT memory
Date: Thu, 24 Aug 2000 14:42:06 GMT
From: "Said Abdosh" 
Message-ID: <2vap5.1225$R2.107383306@nr1.ottawa.istar.net>

Hi there,

When I am trying download to Target Server (SIMNT), the error says that the
file is to large to download.

I know that the default allocated memory is around 1.8M. But I would like to
increase by certain amount of MB. I used the following way from the command
line
D:\TORNADO\target\config\simpc>vxworks /r 12345 <- this in KB. But it did
not increased the memory.

Does anybody know how to increase the SIMNT (Target Server) memory?

Thank you

/Said




---------------------------

Newsgroups: comp.os.vxworks
Subject: MCPN765 from Motorola
Date: Thu, 24 Aug 2000 09:21:51 -0500
From: Rohit Sharma 
Organization: iDEN Systems Division, Motorola
Message-ID: <39A52F7F.E6B14C07@email.mot.com>

Is VxWorks supported on the MCPN765 from Motorola
with the MPC7400 processor ?
Are there any plans to do so ?
Rohit

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: mvme2700 rev level?
Date: 24 Aug 2000 15:43:38 GMT
From: bowtie@netcom.com (Bill Dennen)
Organization: MindSpring Enterprises
Message-ID: <8o3fra$glc$1@slb7.atl.mindspring.net>
References: <39A3C7C6.1AD41B3C@mindspring.com>

Al Johnston  writes:

>Anyone out there understand the defects in the universe and
>falcon chips that are mentioned in the target.nr file for this
>cards bsp?

>Are there any newer versions of this card that use the universe IIB/
>Falcon 3?  The cards I have received so far have been "11B" and
>"11C" versions, both of which seem to have the universe II chip
>on them rather than the IIB.

>I am very concerned that the defects mentioned in the target.nr
>file (mainly atomic access issues) are going to cause me 
>problems.

>Anyone have any information / experience they can share with
>me?

>Yes, I have ask moto... but the cards may be obsolete before
>we get an answers -g-.

>thanks.

The boards you appear to have are the latest revisions; both have Falcon 3
chips.  The Universe IIB was a drop in replacement and thus did not cause 
a rev change on the board.  You can determine if it is a II or IIB by looking
at the chip markings -- CA91C142B-33CE is a IIB while CA91C142-33CE or 
CA91C142-33CEZ1 are Universe II.

Referring to atomic access issues I suspect your concern is about RMW across
the VMEbus.  This is a "problem" with all PowerPC based SBCs (not just the 
ones from MCG).  The fundamental issue is that a PowerPC processor does not
have a TAS instruction (as do the 680x0 processors).  Thus, other mechanisms
must be invoked when an RMW operation is desired.  The procedures supplied
by VxWorks have proven to work quite well.

Regards

- --
>@<					bowtie@netcom.com
Bill Dennen				william.dennen@motorola.com   
	Motorola Computer Group, TOPAZ

---------------------------

Newsgroups: comp.os.vxworks
Subject: To C or C++?
Date: Thu, 24 Aug 2000 11:48:59 -0400
From: Henry 
Organization: UUNET Canada News Reader Service
Message-ID: <39A543EB.C9E19479@mail.com>

Hi, All,

This is not my intention to start a language war. I want to konw should
we C or C++, and if we use C++ under VxWork, do we pay penalty on code
sizes, performace etc.?

I personally prefer C++ for OO and its strong type checking, but I want
your opinions.

Thanks in advance.

Henry


---------------------------

Newsgroups: comp.os.vxworks
Subject: termios.h
Date: Thu, 24 Aug 2000 10:55:39 -0400
From: sunil 
Organization: mot
Message-ID: <39A5376B.CF2B886A@email.mot.com>

Hello VxWorkers.

I have a problem looking for  a file in vxworks which does same thing as
file "termios.h" does in unix. So basically I am looking for functions
defined in termios.h on vxworks.
Can any body help.

thanks
Sunil


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: MCPN765 from Motorola
Date: 24 Aug 2000 15:50:40 GMT
From: bowtie@netcom.com (Bill Dennen)
Organization: MindSpring Enterprises
Message-ID: <8o3g8g$764$1@slb6.atl.mindspring.net>
References: <39A52F7F.E6B14C07@email.mot.com>

Rohit Sharma  writes:

>Is VxWorks supported on the MCPN765 from Motorola
>with the MPC7400 processor ?
>Are there any plans to do so ?
>Rohit

An Early Access BSP for the MCPN765 is available from MCG.  Contact your
local technical support to obtain it.  The BSP requires Tornado II and is
SENS only.  There is also a patch that will be required if you're using an
MPC7400 2.9 (or earlier) processor.  Be aware that when Wind River issues its
BSP for the MCPN765 that all support for this Early Access BSP will terminate.

Regards
- --
>@<					bowtie@netcom.com
Bill Dennen				william.dennen@motorola.com   
	Motorola Computer Group, TOPAZ

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: compiler upgrade with Tornado
Date: 24 Aug 2000 19:50:26 GMT
From: Diamand@btinternet.com (Luke Diamand)
Organization: Customer of Energis Squared
Message-ID: 
References: <39A3D58E.FA65747C@ss8networks.com>

If you upgrade to the newer compiler you will struggle
to make use of any WRS supplied C++ libraries, due to 
changes in object layout (IIRC) and name mangling (IIRC).

But you can easily build the compiler - just grab the gcc
sources, configure and make.

HTH!
Luke Diamand

On Wed, 23 Aug 2000 09:45:50 -0400, Yi Qin  wrote:
>I'm compiling some c++ code on VxWorks and I encountered
>two problems. One is that the compiler does not support
>name space, and the other is that template class pointer
>is not handled properly. It's a lot of work to change all
>the code to fix the two problems.
>
>One simple solution is to upgrade the VxWorks compiler.
>I have compiled the code successfully on Solaris using
>gcc 2.95.2. I would like to know if there is a patch
>to upgrade the compiler. If not is it possible to upgrade
>the compiler by ourselves ?
>
>We are using VxWorks 5.4 and the built in compiler is
>gcc version egcs-2.90.29 tornado 2.0 980515 (egcs-1.0.3
>release) for vxsim and gcc driver version cygnus-2.7.2-960126 for
>ppc860.
>
>I'll really appreciate your help.
>
>Regards,
>
>Yi Qin
>


- -- 

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Adventnet SNMP agent for vxworks
Date: Thu, 24 Aug 2000 15:17:10 GMT
From: randyj5573@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8o3e8u$o76$1@nnrp1.deja.com>
References: <3999AE69.AF526782@yahoo.com> <8nemkb$cl5$1@nnrp1.deja.com> <399D42CC.23F50FAF@yahoo.com> <8nrfls$fti$1@nnrp1.deja.com> <8o1lub$otq$1@nnrp1.deja.com>

The WRS SNMP agent just listens to the SNMP port (161 I think) so any
SNMP implementation that can interface with Berkly sockets should work.

I am using WindNet SNMPv1/v2c. I do not know which version of Envoy
that would be.

I don't think that you understood what I said about the WRS
documentation of their agent. Please go back and read what I said. The
WRS SNMP documentation has a good general overview of their agent but
lacks the depth needed to actually use it to write code. I quickly came
up with more questions than the documentation or the WRS tech suport
could answer. The only way that I have been able to get my code working
is to compile the WRS code examples and step through them to see how
they work. The use of undocumented macros in the code generated by the
mib compiler is frustrating.

Regards,
Randy Jones

In article <8o1lub$otq$1@nnrp1.deja.com>,
  pintu_12@my-deja.com wrote:
>
>
>  Hi,
>
>    Apparently the advent net SNMP agent currently provides support
>  only for SNMP group counters, support for the Mib-2 variables would
>  be provided later.
>
>   Also could any one of you let me know as to whether the Advent
>  Net SNMP agent would be compatible with TCP/IP stack of VxWorks
>  (as I am not very sure whether there SNMP agent could be integerated
>  with the VxWorks TCP/IP stack)
>
>   Randy, which SNMP agent of VxWorks did you try using - Envoy 8.0
>  (better known as WindNet SNMP) or Envoy 9.1 ? An alternative to
>  Envoy could be the emanate SNMP agent. You have said that the
>  documentation for the wind river SNMP agent was good - which SNMP
>  agent are you reffering to - the Envoy 8.0 or the Envoy 9.1. I am
>  trying to use Envoy 9.1 and I am not finding the documentation to be
>  sufficient. In case you feel that the documentation for Envoy 8.0 is
>  good please let me know,
>
>  Regards,
>  Piyush


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: How do I use op code (Assembly Language)...
Date: 24 Aug 2000 15:59:50 GMT
From: Geir Frode Raanes 
Organization: UNINETT news service
Message-ID: <8o3gpm$rv9$1@snipp.uninett.no>
References: <39a5354c.0@seat.s3two.ie>
Sender: Geir Frode Raanes 

Keith McHugh  wrote:

>How do I use op code (Assembly Language) Instructions 
>in the PowerPC with the VxWorks development tools?

First of all - keep your subject line short and
place the actual question in the message body. 
Thank you.


Next; Book:		"GNU Toolkit User's guide"
      Section 1:	"Using GNU CC" 
      SubChapter 3.32:	"Assembler instructions with
			 C Expression Operands"

And after you realise that you can not make any sense 
out of this documentation, please RTFM again and search 
the net - more specifically the mail list achives at 
'sourceware.cygnus.com' before you ask here again.


I'll even give you a short form hint. Place the assembly
instruction(s) within the C attribute container structure. 

This structure is a gateway between assembly symbol space
and C symbol space. GCC will place C variables (symbols)
into the proper registers _for_ you. Thus do not use 
register names directly but instead use placeholder
variables (%0-9) that refers to the C type symbol you
intend to operate on. 

Then throw in the restrictions as to what kind of registers 
or even exactly what register can be used to exchange this 
symbol (data) together with the C symbol name trailing the 
assembly block. You will need to inform GCC what registers/
symbols are only read from (=input,) what registers/symbols 
are read/written to (=output) and what registers simply fall 
victim to drive-by shooting due to this assembly listing.
(The latter registers will be saved/restored before/after 
 the assembly list is executed.)

- -- 
  ******************************************************
  Never ever underestimate the power of human stupidity.
  -Robert Anson Heinlein

		GeirFRS@invalid.and.so.forth
  ******************************************************

---------------------------

Newsgroups: comp.os.vxworks
Subject: Have you tried VxVMI?
Date: Thu, 24 Aug 2000 16:20:19 -0400
From: Michael Simpson 
Organization: Solinet Systems
Message-ID: <39A58383.72B5914C@solinetsystems.com>

This is a multi-part message in MIME format.
- --------------3C75487CFC5D0CE223DDCC03
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Hi there,

Does anybody out there have any experience using VxVMI to protect text
and interrupt vector memory?  I have contacted WRS to find out more
about VxVMI as the documentation I have is thin.  Is there a run-time
cost associated with using this add-on?

Thx,

Mike Simpson
michael@solinetsystems.com

- --------------3C75487CFC5D0CE223DDCC03
Content-Type: text/x-vcard; charset=us-ascii;
 name="michael.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Card for Michael Simpson
Content-Disposition: attachment;
 filename="michael.vcf"

begin:vcard 
n:Simpson;Michael
x-mozilla-html:TRUE
org:Solinet Systems
adr:;;;;;;
version:2.1
email;internet:michael@solinetsystems.com
title:Software Engineer
fn:Michael Simpson
end:vcard

- --------------3C75487CFC5D0CE223DDCC03--


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Project/linker question
Date: 24 Aug 2000 20:53:13 GMT
From: johan@borksoft.xs4all.nl
Organization: Borksoft
Message-ID: <8o41vp$a5s$3@news1.xs4all.nl>
References: <8o3e5c$bp2$1@overload.lbl.gov>

Rainer Stelzer  wrote:
: Hi vxworkers,

: I want to link my object to my standalone vxworks-image.

: The default make rule in the project management includes the ctors/dtors
: call
: even if it's C-Code, so the bootrom generation fails, cause it's already
: munched.

: I expect the Tornado II  project management to provide a solution to
: link
: single project objects in one workspace object, that can be linked to
: the rom image.
: How to ? Or do I have to write my own makefile ?

You can use the file pasrtialImage.o. This is a temporary file, so you
have to modify the file rules.vxApp (I think). In this file the file
partialImage.o is generated but also removed. If you remove the RM command
you can use this file in your link.

Groeten,
	Johan

- -- 
     o o o o o o o . . .   ___________________________________
    o      _____           ||    Johan Borkhuis              |
  .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
 >(________|__|_[_________]_|________________________________|
 _/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: mvme2700 rev level?
Date: Thu, 24 Aug 2000 13:46:45 -0400
From: Al Johnston 
Organization: MindSpring Enterprises
Message-ID: <39A55F85.A4276ABF@mindspring.com>
References: <39A3C7C6.1AD41B3C@mindspring.com> <8o3fra$glc$1@slb7.atl.mindspring.net>


> The boards you appear to have are the latest revisions; both have Falcon 3

Glad to hear from Motorola, even if a "any views are my own" type
disclaimer
applies.  Have been trying to get this question answered by MCG for 10
weeks now... So I will take the info wherever I can get it!

> chips.  The Universe IIB was a drop in replacement and thus did not cause
> a rev change on the board.  You can determine if it is a II or IIB by looking
> at the chip markings -- CA91C142B-33CE is a IIB while CA91C142-33CE or
> CA91C142-33CEZ1 are Universe II.

The markings are "143-" (no B) but I was not sure if tundra printed the
"B"
on the IIB chips or not.  pci bug also says its a rev "1" (if I got the 
right item).  So we have a II.  Little bit surprised about the "no new
rev" 
but thanks for letting me know.

> Referring to atomic access issues I suspect your concern is about RMW across
> the VMEbus.  This is a "problem" with all PowerPC based SBCs (not just the

The case I am concerned with is where a semaphore is located in the OB 
DRAM on cpu0.  Code running on CPU0 can try to get the S4 while code
running on other cpus are also trying to get the S4 via the vme slave
window.  

My concern is driven by statements in wrs's documents as well as the UII
errata.  The concern is that these accesses my fail to be atomic during 
periods where both (local and vme) types of access are occurring.


My impression so far is that (contrary to what WRS says in their 
target.nr file) I can configure vxworks with
  ANY_BRDS_IN_CHASSIS_NOT_RMW undefined
  SM_OFF_BOARD defined as FALSE
and in the case in question, the sysBusTas routines will work as they
should.

Can you confirm this?

And now for an easy one... Is there an "errata" document for the board?
I have a lot of doc from the web site but I couldn't find that type
of document... maybe there are no errata to document!

Thanks for the help.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Project/linker question
Date: Thu, 24 Aug 2000 13:04:53 -0500
From: Greg Willden 
Organization: Posted via Supernews, http://www.supernews.com
Message-ID: <39A563C5.8EB73F5D@swri.edu>
References: <8o3e5c$bp2$1@overload.lbl.gov>

Rainer Stelzer wrote:

> Hi vxworkers,
>
> I want to link my object to my standalone vxworks-image.
>
> The default make rule in the project management includes the ctors/dtors
> call
> even if it's C-Code, so the bootrom generation fails, cause it's already
> munched.
>

I changed the build rule for my project to make an archive/library file
(projectx.a).  Then I added the projectx.a file to the PRJ_LIBS macro of the
build.  Also make sure that usrAppInit contains the entry point for your
application code and that usrAppInit is included in the kernel build.  (I
forget the INCLUDE_ )
Have fun.
    Greg


---------------------------

Newsgroups: comp.os.vxworks
Subject: RE: building vxWorks using Tornado II
Date: Thu, 24 Aug 2000 15:35:31 -0700
From: venky 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <8o4a9c$drf$1@overload.lbl.gov>

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_01C00E1B.9C1BBEC6
Content-Type: text/plain;
	charset="iso-8859-1"

:->1) How do I know what library these components are coming 
:->from?    say:
:->\target\lib\objPENTIUMgnuvx  vs  \target\lib\obj80386gunvx?

When a project is created, it is created based on a BSP. THis 
will decide which library to choose the components from.

:->2) How do I know which "C" source files are associated with a given
:->component that I'm including?

This will be a little difficult if you dont have the entire vxWOrks source
code.
even if you do have, a manual search in the source directory is the option.

:->3) If I want to modify a given components source file to 
:->make it fit my
:->particular hardware, how do I specify to use my version of 
:->the file instead
:->of the one from the library?

You can include that particular file in the project thereby overriding the 
library version of the file ...

HTH

V.

- ------_=_NextPart_001_01C00E1B.9C1BBEC6
Content-Type: text/html;
	charset="iso-8859-1"






RE: building vxWorks using Tornado II




:->1) How do I know what library these components are coming 

:->from?    say:

:->\target\lib\objPENTIUMgnuvx  vs  \target\lib\obj80386gunvx?




When a project is created, it is created based on a BSP. THis 

will decide which library to choose the components from.




:->2) How do I know which "C" source files are associated with a given

:->component that I'm including?




This will be a little difficult if you dont have the entire vxWOrks source code.

even if you do have, a manual search in the source directory is the option.




:->3) If I want to modify a given components source file to 

:->make it fit my

:->particular hardware, how do I specify to use my version of 

:->the file instead

:->of the one from the library?




You can include that particular file in the project thereby overriding the 

library version of the file ...




HTH




V.





- ------_=_NextPart_001_01C00E1B.9C1BBEC6--

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: To C or C++?
Date: Thu, 24 Aug 2000 18:17:07 -0500
From: Steven Deans 
Organization: Fujitsu Companies of North America
Message-ID: <39A5ACF3.DCCEC871@fnc.fujitsu.com>
References: <39A543EB.C9E19479@mail.com>

Hi Henry,

In general C++ will only be slightly larger and slightly slower than C. In
fact, well written C++ with a good optimizing C++ compiler can sometimes
execute faster than C because of inlined code and the compiler can
optimize better because of the stronger type checking. The size and speed
differences are small generally < 3%. If this difference is a big concern,
maybe you should consider writing in assembly. :) I have 5+ years writing
Real Time embedded OO C++ and  3+ years writing real time embedded C. I
recommend C++. It has been my experience that most people that prefer C
either have not used C++ or don't "get" OO.

The bottom line is the quality of your code and your compiler are going to
be  a bigger factor than C vs. C++ differences.

Steven


Henry wrote:

> Hi, All,
>
> This is not my intention to start a language war. I want to konw should
> we C or C++, and if we use C++ under VxWork, do we pay penalty on code
> sizes, performace etc.?
>
> I personally prefer C++ for OO and its strong type checking, but I want
> your opinions.
>
> Thanks in advance.
>
> Henry


---------------------------

Newsgroups: comp.os.vxworks
Subject: post again: Some Quesions on Mux
Date: Fri, 25 Aug 2000 09:27:46 +0800
From: "Ren Hantao" 
Organization: The news server of Changzhou TeleCom
Message-ID: <8o4i63$1j51$1@news.cz.js.cn>

  I have post this message several days ago.
since no one response the letter, I want to
post it again. Hope anybody who know this
issue can give me some advice.


Thanks

Ren Hantao
2000/8/25


Hi,

  I am confronted with many difficulties in writing MUX based protocols,
Your help will be appreciated very much,some examples in detail are
prefered.

1. First of all,
  I am not clear about priority mechanism.Chapter 4 in <> says that protoclol type supplied to the bind
function
is used by MUX to prioritize the protocols,MUX_PROTO_SNARF has the highest
priority otherwise MUX_PROTO_PROMISC the lowest.

They are defined in muxLib.h as following,

#define MUX_PROTO_PROMISC 0x100
#define MUX_PROTO_SNARF   0x101
#define MUX_PROTO_OUTPUT  0x102

They are in the ethernet size range(marked as Experimental in rfc 1700)
Initially "muxShow" routing displays the following,
- ----------------------------------------------------------------------------
- --
- -> muxShow
Device: dc Unit: 0
Description: dec21x40 Enhanced Network Driver
Protocol: Wind Debug Agent Type: 257 Recv 0x13ede0 Shutdown 0x0
Protocol: IP 4.4 ARP Type: 2054 Recv 0x1556d4 Shutdown 0x155998
Protocol: IP 4.4 TCP/IP Type: 2048 Recv 0x1556d4 Shutdown 0x1558d0
value = 0 = 0x0
- ->
- ----------------------------------------------------------------------------
- --
Now I want to write a protocol which has higher priority than "IP 4.4 ARP"
and "IP 4.4 TCP" do.
I am confused by which type should I use for my protocol.
I has tested a lot of types such as 2049,259,258 etc,however they can not
get
any packet even if I Shundown the "IP 4,4 xx" protocol with "ipDetach()"
routing.
I can not muxbind() my protocol with type of MUX_PROTO_SNARF because WDB has
used
this type. Protocol with type of MUX_PROTO_PROMISC can get packet until "IP
4.4 xxx"
has been shutdowned.

2. Secondly,
  I want to make use of "IP 4.4 ARP" protocol as my address resolution
function.
but the "arpresolve" function is so complicated.
  arpresolve (struct arpcom *, struct rtentry *, struct mbuf *, struct
sockaddr *, u_char *);
  Lack of document about it make thing worse. Please give me some example
about the use of
arp functions.

3.
   Manual says Routing muxAddressForm() returns the head of an mBLK chain
containing the data
form pMBLK with a prepended link-level header.  When I use it, it really
returns a pMBLK with
a link-level header.  The link-level header (14 byte)is as following,

   00 10 5a 5e d3 bf 08 00 3e 2f 2f 5e 00 00
  ~~~~~~~~~~~~~~~~~  ~~~~~~~~~~~~~~~~~ ~~~~~
  Destionation MAC     Source MAC      Protocol

  Oh,Protocol filed is always 00 00,but I need other types such as 08
00(IP),08 06(ARP)
  I can only chang this filed manually.
  What makes me confused is that transmit functions in wdbEndPktDrv.c  use
only the
muxAddressForm(),but get "08 00" type filed.  Is there any relationgship
between type
in muxBind and type in the link-level formed by muxAddressForm?

4.
 netMblkOffsetToBufCopy() or netMblkBufCopy() can copy a chains of data to a
buffer.
 If I copy data from a long buffer to a mBlk chains ,are there any functions
I can
 use to facilitate the opeation without take the size of cluster into
consideration,
 which can chains the mBlk automatically?

 I also need some help about how to use mbuf sturcture, Is mbuf the same as
mblk?..
 Examples are prefered.


You help is very much appreciated!


Ren Hantao
2000/8/4






---------------------------

Newsgroups: comp.os.vxworks
Subject: Write Protection for Task Stacks
Date: Fri, 25 Aug 2000 02:24:58 GMT
From: madowdy@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8o4ld8$6ia$1@nnrp1.deja.com>

We may be insane to try this but we have a highly reliable
application and we'd like to provide additional protection
from our own indiscretions by write protecting the stack
of all but the current task. We've taken care of allocating
and sizing the stacks to ensure that they are placed on page
boundaries and are an even number of pages long. In our
first cut at the switch hook, we write-enabled the stack
of pNewTcb and write-protected the stack of pOldTcb. This
failed (I guess I can see how the stack of the "old" task
could be used by the task switching code after the call
to a switch hook). In our next version of the switch hook,
we saved the TCB of the penultimate task (the one that ran
before "old"). If the TCB of the penultimate task did not
match the TCB of the "new" task, we write protected the
stack of the penultimate TCB. This too failed (i.e. caused
write protect violations).

We don't have a source license so it's hard to know what
WRS is doing around the calls to the task switch hooks.
Would someone with the source let me know what is going
on with the task stacks around the time the switch hooks
are called. Specifically, does WRS modify the stack of
the "new" task prior to the calls to the switch hooks.
Thanks.

Mark


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Strange interrupt behavior noticed with WindView
Date: Fri, 25 Aug 2000 09:10:51 +0530
From: Vijay Kumar Peshkar 
Organization: Ye 'Ol Disorganized NNTPCache groupie
Message-ID: <39A5EAC3.1CC2F519@cisco.com>
References: <39A3E0F1.37C79E65@lmco.com>
Reply-To: vpeshkar@cisco.com

This is a multi-part message in MIME format.
- --------------E6ACBA11092652BB33800A7B
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Frederick,

No clue.
You have to give me some description of h/w and s/w configuration at least to
guess.

regards
Vijay

Frederick Byrne wrote:

> We've noticed in a Windview display that occasionally an interrupt handler (it
> appears to be a vxWorks handler and not one we've defined) will keep the CPU
> until the OS clock tick has expired (16.6 mS).  The specific interrupt number
> changes but the behavior is the same.  Does anyone have any ideas?  Thanks in
> advance for any help.
>
> --
> ==============================================================
>         Frederick "Rick" Byrne  frederick.j.byrne@lmco.com
>         Lead Member of the Engineering Staff
>         Lockheed Martin NE&SS, Moorestown NJ
>         Phone: (856) 722-7344   Fax: (856) 722-4103

- --------------E6ACBA11092652BB33800A7B
Content-Type: text/x-vcard; charset=us-ascii;
 name="vpeshkar.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Card for Vijay Kumar Peshkar
Content-Disposition: attachment;
 filename="vpeshkar.vcf"

begin:vcard 
n:Peshkar;Vijay
x-mozilla-html:FALSE
org:Wipro Technologies,K2, Bangalore;Wishing you blue skies, a couple of daisies, some fresh air and a smile!
adr:;;;;;;
version:2.1
email;internet:vijay.peshkar@wipro.com
title:Senior Software Engineer
fn:Vijay Kumar Peshkar
end:vcard

- --------------E6ACBA11092652BB33800A7B--


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Configure BSP fails
Date: Fri, 25 Aug 2000 03:51:19 GMT
From: john.h.kennedy@worldnet.att.net (John Kennedy)
Organization: AT&T Worldnet
Message-ID: <39a5d4bb.5659686@netnews.worldnet.att.net>
References: <8m68pc$m4t$1@news1.tele.dk>


I'm also new to the product and had this same fault after installation
of the WRS SW ( T101 on WinNT4 SP6 ), but before installation of the
supplied BSP into the ../target/config/ directory.

Once I installed the BSP (by unzipping the ZIP file into the ../target
directory) the error message went away and I could see the default
VxWorks components to be loaded at build time.

You seem to indicate that you do not have a BSP yet.  This may be the
problem.  For further info check into the HELP that's available in the
under the BSP Config window.  There's a more exacting and decent
explaination of how WINDCFG works.

John




---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Help
Date: Fri, 25 Aug 2000 09:22:04 +0530
From: Vijay Kumar Peshkar 
Organization: Ye 'Ol Disorganized NNTPCache groupie
Message-ID: <39A5ED64.B3AC9D23@cisco.com>
References: 
Reply-To: vpeshkar@cisco.com

This is a multi-part message in MIME format.
- --------------545BFE74D3252792205F5A69
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Michael,

A wild guess.
You have loaded vxworks image into the target while your configuration in
config.h is asking mkboot.bat to look for vxworks.st.
Go to config.h and have a look at the line where 'DEFAULT_BOOT_LINE'
is defined. If my guess is right, the image name here is vxworks.st'.
If true, you can either build a vxworks.st and load it to the target. Else
specify the correct image name here and rebuild.

To customise, U need to go to 'tornado' and include/exclude the components
you want to go in your target image. Please have a look at the tornado user's
guide. Its all there ;)

regards,
Vijay

"Michael S." wrote:

> I cant figure out why there is no vxworks.st which the mkboot.bat bath file
> is looking for. Also I am trying to figure out how to costumize your vxworks
> os.Anyone have any ideas?
>
> Sincerely,
> Michael S - scikid@hotmail.com

- --------------545BFE74D3252792205F5A69
Content-Type: text/x-vcard; charset=us-ascii;
 name="vpeshkar.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Card for Vijay Kumar Peshkar
Content-Disposition: attachment;
 filename="vpeshkar.vcf"

begin:vcard 
n:Peshkar;Vijay
x-mozilla-html:FALSE
org:Wipro Technologies,K2, Bangalore;Wishing you blue skies, a couple of daisies, some fresh air and a smile!
adr:;;;;;;
version:2.1
email;internet:vijay.peshkar@wipro.com
title:Senior Software Engineer
fn:Vijay Kumar Peshkar
end:vcard

- --------------545BFE74D3252792205F5A69--


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: symbol table problems
Date: Fri, 25 Aug 2000 07:43:57 +0200
From: "Serge Wenger" 
Organization: Swisscom IP+ (post doesn't reflect views of Swisscom)
Message-ID: <8o50rp$pe5$1@pollux.ip-plus.net>
References: <39A51933.D3E4AA1F@inta.es>

Samuel,

You must include C++ String Type in your VxWorks. On the tab VxWorks, C++
Components, standard library, C++ string type. Rebuild your VxWorks and try
again.

I hope it will be OK.

/Serge Wenger

"Samuel F. Romero García"  a écrit dans le message news:
39A51933.D3E4AA1F@inta.es...
> Hi there!
>
> I'm a newbi on VxWorks, and I'm trying to make use of the already
> defined class string (by inserting in my code #include    and so
> on).
>
> The compiler goes OK, no complaint, but once I download it to the
> target... it says:
>
> Unresolved symbols encountered while downloading:
>
> Errors while downloading
> C:/TORNADO/target/proj/SimulaTrib/SIMNTgnu/SimulaTrib.out:
> _length__t18string_char_traits1ZcPCc
>
_clone__Q2t12basic_string3ZcZt18string_char_traits1ZcZt24__default_alloc_tem
plate2b1i03Rep
>
>
__t12basic_string3ZcZt18string_char_traits1ZcZt24__default_alloc_template2b1
i0$nilRep
>
>
___dl__Q2t12basic_string3ZcZt18string_char_traits1ZcZt24__default_alloc_temp
late2b1i03RepPv
>
>
_replace__t12basic_string3ZcZt18string_char_traits1ZcZt24__default_alloc_tem
plate2b1i0UiUiPCcUi
>
> I'm using Tornado II for Windows 95 (NT), and I use the simulator as
> target (Pentium). Thanks for your help!
> --
> Samuel F. Romero García    - romerogs@inta.es
> Instituto Nacional de Técnica Aeroespacial - INTA
> División de Sistemas Electrónicos - Área Ing. Software
> Tel. 91 520 13 49 - Crtra. Ajalvir s/n Km 4.5
> Torrejón de Ardoz, 28850 - MADRID
> --
> If we knew what we were doing, it wouldn't be called research... - A.
> Einstein
>
>



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: To C or C++?
Date: Fri, 25 Aug 2000 07:26:52 GMT
From: tangoNO_SPAM@pulp-fiction.com (Max)
Organization: Posted via Supernews, http://www.supernews.com
Message-ID: <39a61a22.375055037@news.pacificnet.net>
References: <39A543EB.C9E19479@mail.com>

On Thu, 24 Aug 2000 11:48:59 -0400, Henry  wrote:

>Hi, All,
>
>This is not my intention to start a language war. I want to konw should
>we C or C++, and if we use C++ under VxWork, do we pay penalty on code
>sizes, performace etc.?
>
>I personally prefer C++ for OO and its strong type checking, but I want
>your opinions.
>
>Thanks in advance.
>
>Henry
>

It also depends on the amount of work to be done.
Smaller projects tend to be easier to write and maintain in C, as OOP
always requires some overhead in system layout and class preparation.
However, in larger projects this overhead pays back handsomely.

Another issue of C++, especially for VxWorks, is the tremendous size
of the linked C++ support libraries. In my experience with ARM cores
the output binary with OO blew at least twice the size of a plain C
system, unless rigorously controlled. At the same time, the
performance degradation is much less noticeable.

Max.

Cambury Technologies, Inc
embedded wireless systems
http://www.cambury.com

---------------------------

Newsgroups: comp.os.vxworks
Subject: I/O device read/write
Date: Fri, 25 Aug 2000 07:24:37 GMT
From: mohankumarh@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8o56v9$qe6$1@nnrp1.deja.com>

What are the functions for writing/reading from I/O devices.
In unix we have outp/outw/outb/inp/inw/inb, I would like to know the
corresponding functions in vxWorks.

Thanks
Mohan


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: DHCP client configuration
Date: Fri, 25 Aug 2000 12:53:32 +0530
From: slkumar 
Organization: Motorola
Message-ID: <39A61EF4.434A006B@miel.mot.com>

We need to get the value of option-128 (site specific)  from  the DHCP
server. In the client application code we have call :

dhcpcOptionSet (pLeaseCookie, 128  , 0, 0, NULL);

This call returns ERROR at runtime.

Any idea how to obtain site specfic dhcp options (128-255) in the client
code ?

Regards
Shailendra


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Increasing SIMNT memory
Date: Fri, 25 Aug 2000 10:03:13 +0200
From: anybody@nowhere.com
Organization: Siemens AG,(Hofmannstr)  Munich-Germany-Europe.
Message-ID: <39A62841.6542DEE@nowhere.com>
References: <2vap5.1225$R2.107383306@nr1.ottawa.istar.net>

Hi!

You must specify the memory in bytes, so if you run
vxWorks.exe -r16777215
you run it with 16MB of RAM and this works (at least at our project).
This is the memory it may use, but it does allocate it at startup. The memory is
increased if you upload a large OUT-file.

It is also a good idea to increase the simulator memory: Tools/Target
Server/Configure/Target Server Properties/Memory Cache Size=... (this is in
kBytes)!

Said Abdosh wrote:

> Hi there,
>
> When I am trying download to Target Server (SIMNT), the error says that the
> file is to large to download.
>
> I know that the default allocated memory is around 1.8M. But I would like to
> increase by certain amount of MB. I used the following way from the command
> line
> D:\TORNADO\target\config\simpc>vxworks /r 12345 <- this in KB. But it did
> not increased the memory.
>
> Does anybody know how to increase the SIMNT (Target Server) memory?
>
> Thank you
>
> /Said


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Increasing SIMNT memory
Date: 25 Aug 2000 03:23:22 -0500
From: Bob Westbrook 
Message-ID: 
References: <2vap5.1225$R2.107383306@nr1.ottawa.istar.net> <8o41ql$a5s$2@news1.xs4all.nl>

On 24 Aug 2000 20:50:29 GMT, johan@borksoft.xs4all.nl wrote:

>Said Abdosh  wrote:
>: D:\TORNADO\target\config\simpc>vxworks /r 12345 <- this in KB. But it did
>: not increased the memory.
>
>The size is in bytes, not in kBytes. The manual is wrong here.
>
>Groeten,
>	Johan

... and the default Simulator doesn't support the /r switch.

You have to build a new Simulator with File->New Project->
Bootable vxWorks Image->Add to Current Workspace(Next)->
existing project based on simpc_vx.wpj->Finish
Then use it's vxWorks properties tab to change any components you
need.

Once it's built you get a new vxWorks.exe which you can use the /r
switch on to change the memory size. To use it choose Tools->Target
server from the Tornado menu, configure it with your options, then
select it.
If it's properly configured then when you click on the dropdown list
nex to the (i) (launch browser) button you should see your new target
server, be able to select it and then be able to download to it.

Hope this helps


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: I/O device read/write
Date: Fri, 25 Aug 2000 09:58:43 GMT
From: hruesing@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8o5g0d$3m4$1@nnrp1.deja.com>
References: <8o56v9$qe6$1@nnrp1.deja.com>

In article <8o56v9$qe6$1@nnrp1.deja.com>,
  mohankumarh@my-deja.com wrote:
> What are the functions for writing/reading from
I/O devices.
> In unix we have outp/outw/outb/inp/inw/inb, I
would like to know the
> corresponding functions in vxWorks.
>
> Thanks
> Mohan
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.
>

The functions are sysInByte etc. in syslib.c

Harud Rüsing


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: VxWorks/Tornado shell checksum and ls utilities
Date: 25 Aug 2000 10:11:14 GMT
From: johan@borksoft.xs4all.nl
Organization: Borksoft
Message-ID: <8o5go2$pir$1@news1.xs4all.nl>
References: <8o0t9r$oth$1@overload.lbl.gov>

Straub, Robert  wrote:
: - One is a directory listing utility.  I want it to display a recursive
: listing of files in a dosFs directory tree structure, in a similar way to
: the UNIX command, "ls -lR".

If you are using DosFs2: try the command lsr or llr

Groeten,
	Johan

- -- 
     o o o o o o o . . .   ___________________________________
    o      _____           ||    Johan Borkhuis              |
  .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
 >(________|__|_[_________]_|________________________________|
 _/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: Multiple symbol tables
Date: 25 Aug 2000 10:17:24 GMT
From: johan@borksoft.xs4all.nl
Organization: Borksoft
Message-ID: <8o5h3k$pir$2@news1.xs4all.nl>
References: <8o16qq$6g9$1@nnrp1.deja.com> <8o1brl$has$2@news1.xs4all.nl> <8o1jfa$lsu$1@nnrp1.deja.com>

gpbeck@my-deja.com wrote:
: I've tried doing just that but it requires a network port for debug.  I
: can only use a serial connection.  My hardware platform has two serial
: ports, one for WDB and second for a User Interface.  I've tried the
: command 'add-symbol-table' from the GDB command interface but I get an
: error message.  I can't remember the exact message, but it said
: something about the loader module not being installed.  But that is not
: true.

The problem you are facing is because of the difference between the host
and target environment. The only way to do that is to switch on the
symbol-table synchrionisation. As this uses the WDB link that you already
have it should not be a problem. Check the target server configuration and
your VxWorks configuration.

Groeten,
	Johan

- -- 
     o o o o o o o . . .   ___________________________________
    o      _____           ||    Johan Borkhuis              |
  .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
 >(________|__|_[_________]_|________________________________|
 _/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: Write Protection for Task Stacks
Date: Fri, 25 Aug 2000 10:38:33 GMT
From: hruesing@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8o5ib7$640$1@nnrp1.deja.com>
References: <8o4ld8$6ia$1@nnrp1.deja.com>

In article <8o4ld8$6ia$1@nnrp1.deja.com>,

When allocating the stack for a task also the TCB (and perhaps other
VxWorks items ar allocated).
Try to write protect only the stack (something like sizeof buffer -
sizeof(TCB)), where the TCB is at the start or the end of your buffer
dependingon your target architecture.

Harud Rüsing



Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

End of New-News digest
**********************


From vxwexplo-errs@csg.lbl.gov  Fri Aug 25 05:03:07 2000
From: Johan van der Werf 
Date: Fri Aug 25 05:03:09 PDT 2000
Subject: Canceling a select()


  Hi all,

 Question: what would be the best way to make a task return from a select()
           function call before the timeout is expired, or any data is sent?

 At first sight ioctl(fd, FIOUNSELECT, xxx) seemed a likely candidate, 
 unfortunately it can only be used from inside a driver, because the
 third argument must be a pointer to a SEL_WAKEUP_NODE. Please tell me if
 there is a standard way of doing this?

 We're using VxWorks 5.3.1 on M68K platform.

                        Regards,

-- 
|Johan van der Werf   |   tel: ++31 (0)40 2684280 |$B%O%s%9(B  |
|ASM Lithography      |   fax: ++31 (0)40 2682224 |$B%t%!%s(B  |
|Dept: Basic Software | email: ujw@asml.nl        |$B%G%k(B    |
|Room: 7A3.012        |        vdwerf@xs4all.nl   |$B%&%(%k%U(B|


From vxwexplo-errs@csg.lbl.gov  Fri Aug 25 07:33:44 2000
From: David Laight 
Date: Fri Aug 25 07:33:46 PDT 2000
Subject: Re: Canceling a select()

You could add an extra file descriptor into the select list and use that driver 
to generate a read event.  Maybe a writing a character to a pipe will work (does 
the pipe driver support select?).

Alternatively if one of your own drivers is in the select list you could make it 
wake up all the people waiting for data.  Either using an iocyt() or, since 
vxWorks has open plan access permissions, a direct call of a driver routine.

Indeed if you can find (or know) the address of the SEL_WAKEUP_NODE calling 
selWakeupAll() itself will do.

For a typical serial driver (ie something that uses tyLib):

#include 
kick_select( int fd )
{
 selWakeupAll( ((TY_DEV *)fdTable[ fd - 3 ].pDevHdr)->selWakeupList, SELREAD );
}

should have the desired effect - If you don't mind using some 'private' 
information.

	David
	
>  Question: what would be the best way to make a task return from a select()
>            function call before the timeout is expired, or any data is sent?
> 
>  At first sight ioctl(fd, FIOUNSELECT, xxx) seemed a likely candidate, 
>  unfortunately it can only be used from inside a driver, because the
>  third argument must be a pointer to a SEL_WAKEUP_NODE. Please tell me if
>  there is a standard way of doing this?
> 
>  We're using VxWorks 5.3.1 on M68K platform.
> 
>                         Regards,
> 


----------------------------------------------------------------
David Laight                          email: dsl@tadpole.co.uk
Tadpole Technology plc                phone: +44 1223 428 232
Cambridge, UK                         fax:   +44 1223 428 201



From vxwexplo-errs@csg.lbl.gov  Sat Aug 26 04:03:08 2000
From: Vxworks Exploder 
Date: Sat Aug 26 04:03:10 PDT 2000
Subject: comp.os.vxworks newsdigest

Comp.Os.Vxworks Daily Digest    Sat Aug 26 04:03:04 PDT 2000

        Subject: Re: Have you tried VxVMI?
        Subject: Re: symbol table problems
        Subject: Re: symbol table problems
        Subject: SCSI ON VxWorks
        Subject: Canceling a select()
        Subject: Re: Strange interrupt behavior noticed with WindView
        Subject: vxSim problem
        Subject: Have you used VxVMI?
        Subject: Loading application module during VxWorks.st boot
        Subject: Change # of fds and select limit
        Subject: More PCI Bus & VXWorks
        Subject: Shared Memory Network booting problems
        Subject: re: Canceling a select()
        Subject: Re: Shared Memory Network booting problems
        Subject: DDCMP for vxWorks
        Subject: Re: Shared Memory Network booting problems
        Subject: Re: Canceling a select()
        Subject: Re: What is error S_objLib_OBJ_ID_ERROR ?
        Subject: Re: Strange interrupt behavior noticed with WindView
        Subject: Re: Write Protection for Task Stacks
        Subject: Vxsim File I/O with NT
        Subject: mpc75x+mpc107 BSP yet?
        Subject: ppc: rework of intLock/intUnlock for nesting?

-------------------------------------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Have you tried VxVMI?
Date: Fri, 25 Aug 2000 10:56:30 GMT
From: hruesing@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8o5jcn$7b3$1@nnrp1.deja.com>
References: <39A58383.72B5914C@solinetsystems.com>

In article <39A58383.72B5914C@solinetsystems.com>,
  Michael Simpson  wrote:
> This is a multi-part message in MIME format.
> --------------3C75487CFC5D0CE223DDCC03
> Content-Type: text/plain; charset=us-ascii
> Content-Transfer-Encoding: 7bit
>
> Hi there,
>
> Does anybody out there have any experience using VxVMI to protect text
> and interrupt vector memory?  I have contacted WRS to find out more
> about VxVMI as the documentation I have is thin.  Is there a run-time
> cost associated with using this add-on?
>
> Thx,
>
> Mike Simpson
> michael@solinetsystems.com
>
We are also thinking to buy VxSim. As far as I know the run-time costs
are using the processors MMU (if you are using "basic MMU support"
there should be no overhead because the mmu is in use already). You are
not using the *target* loader/unloader I suppose. Then of course would
be overhead building the new page tables.
Another question is whether you need VxSim or if a "static" protection
in the final release version is enough. Then it could be archieved
with "basic MMU support".

Harud Ruesing



Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: symbol table problems
Date: Fri, 25 Aug 2000 13:22:26 +0200
From: Samuel F. Romero =?iso-8859-1?Q?Garc=EDa?= 
Organization: INTA
Message-ID: <39A656F1.B1FC562B@inta.es>
References: <39A51933.D3E4AA1F@inta.es> <8o50rp$pe5$1@pollux.ip-plus.net>

Thank you for your answer, Serge. It seems reasonable. I followed your
instructions, but it happens again.

When I created my project I did it as downloadable files, not bootable VxWorks
image, so no VxWorks components were included. Now, I have added VxWorks and the
string class, but I don´t know how to download both the files and the OS onto
the simulator.

I guess I need to connect both my files and the VxWorks to run on the simulator.
That may be the problem... what can I do?

Regards.

Serge Wenger wrote:

> Samuel,
>
> You must include C++ String Type in your VxWorks. On the tab VxWorks, C++
> Components, standard library, C++ string type. Rebuild your VxWorks and try
> again.
>
> I hope it will be OK.
>
> /Serge Wenger
>
> "Samuel F. Romero García"  a écrit dans le message news:
> 39A51933.D3E4AA1F@inta.es...
> > Hi there!
> >
> > I'm a newbi on VxWorks, and I'm trying to make use of the already
> > defined class string (by inserting in my code #include    and so
> > on).
> >
> > The compiler goes OK, no complaint, but once I download it to the
> > target... it says:
> >
> > Unresolved symbols encountered while downloading:
> >
> > Errors while downloading
> > C:/TORNADO/target/proj/SimulaTrib/SIMNTgnu/SimulaTrib.out:
> > _length__t18string_char_traits1ZcPCc
> >
> _clone__Q2t12basic_string3ZcZt18string_char_traits1ZcZt24__default_alloc_tem
> plate2b1i03Rep
> >
> >
> __t12basic_string3ZcZt18string_char_traits1ZcZt24__default_alloc_template2b1
> i0$nilRep
> >
> >
> ___dl__Q2t12basic_string3ZcZt18string_char_traits1ZcZt24__default_alloc_temp
> late2b1i03RepPv
> >
> >
> _replace__t12basic_string3ZcZt18string_char_traits1ZcZt24__default_alloc_tem
> plate2b1i0UiUiPCcUi
> >
> > I'm using Tornado II for Windows 95 (NT), and I use the simulator as
> > target (Pentium). Thanks for your help!
> > --
> > Samuel F. Romero García    - romerogs@inta.es
> > Instituto Nacional de Técnica Aeroespacial - INTA
> > División de Sistemas Electrónicos - Área Ing. Software
> > Tel. 91 520 13 49 - Crtra. Ajalvir s/n Km 4.5
> > Torrejón de Ardoz, 28850 - MADRID
> > --
> > If we knew what we were doing, it wouldn't be called research... - A.
> > Einstein
> >
> >

- --
Samuel F. Romero García    - romerogs@inta.es
Instituto Nacional de Técnica Aeroespacial - INTA
División de Sistemas Electrónicos - Área Ing. Software
Tel. 91 520 13 49 - Crtra. Ajalvir s/n Km 4.5
Torrejón de Ardoz, 28850 - MADRID
- --
If we knew what we were doing, it wouldn't be called research... - A.
Einstein



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: symbol table problems
Date: Fri, 25 Aug 2000 13:53:44 +0200
From: "Serge Wenger" 
Organization: Swisscom IP+ (post doesn't reflect views of Swisscom)
Message-ID: <8o5mh5$c1f$1@pollux.ip-plus.net>
References: <39A51933.D3E4AA1F@inta.es> <8o50rp$pe5$1@pollux.ip-plus.net> <39A656F1.B1FC562B@inta.es>

Samuel,

You must recreate your simulator (see your manuals) with the class string
(the simulator contain the VxWorks) and after load your downloadble files.

/Serge
"Samuel F. Romero García"  a écrit dans le message news:
39A656F1.B1FC562B@inta.es...
> Thank you for your answer, Serge. It seems reasonable. I followed your
> instructions, but it happens again.
>
> When I created my project I did it as downloadable files, not bootable
VxWorks
> image, so no VxWorks components were included. Now, I have added VxWorks
and the
> string class, but I don´t know how to download both the files and the OS
onto
> the simulator.
>
> I guess I need to connect both my files and the VxWorks to run on the
simulator.
> That may be the problem... what can I do?
>
> Regards.
>
> Serge Wenger wrote:
>
> > Samuel,
> >
> > You must include C++ String Type in your VxWorks. On the tab VxWorks,
C++
> > Components, standard library, C++ string type. Rebuild your VxWorks and
try
> > again.
> >
> > I hope it will be OK.
> >
> > /Serge Wenger
> >
> > "Samuel F. Romero García"  a écrit dans le message
news:
> > 39A51933.D3E4AA1F@inta.es...
> > > Hi there!
> > >
> > > I'm a newbi on VxWorks, and I'm trying to make use of the already
> > > defined class string (by inserting in my code #include    and
so
> > > on).
> > >
> > > The compiler goes OK, no complaint, but once I download it to the
> > > target... it says:
> > >
> > > Unresolved symbols encountered while downloading:
> > >
> > > Errors while downloading
> > > C:/TORNADO/target/proj/SimulaTrib/SIMNTgnu/SimulaTrib.out:
> > > _length__t18string_char_traits1ZcPCc
> > >
> >
_clone__Q2t12basic_string3ZcZt18string_char_traits1ZcZt24__default_alloc_tem
> > plate2b1i03Rep
> > >
> > >
> >
__t12basic_string3ZcZt18string_char_traits1ZcZt24__default_alloc_template2b1
> > i0$nilRep
> > >
> > >
> >
___dl__Q2t12basic_string3ZcZt18string_char_traits1ZcZt24__default_alloc_temp
> > late2b1i03RepPv
> > >
> > >
> >
_replace__t12basic_string3ZcZt18string_char_traits1ZcZt24__default_alloc_tem
> > plate2b1i0UiUiPCcUi
> > >
> > > I'm using Tornado II for Windows 95 (NT), and I use the simulator as
> > > target (Pentium). Thanks for your help!
> > > --
> > > Samuel F. Romero García    - romerogs@inta.es
> > > Instituto Nacional de Técnica Aeroespacial - INTA
> > > División de Sistemas Electrónicos - Área Ing. Software
> > > Tel. 91 520 13 49 - Crtra. Ajalvir s/n Km 4.5
> > > Torrejón de Ardoz, 28850 - MADRID
> > > --
> > > If we knew what we were doing, it wouldn't be called research... - A.
> > > Einstein
> > >
> > >
>
> --
> Samuel F. Romero García    - romerogs@inta.es
> Instituto Nacional de Técnica Aeroespacial - INTA
> División de Sistemas Electrónicos - Área Ing. Software
> Tel. 91 520 13 49 - Crtra. Ajalvir s/n Km 4.5
> Torrejón de Ardoz, 28850 - MADRID
> --
> If we knew what we were doing, it wouldn't be called research... - A.
> Einstein
>
>



---------------------------

Newsgroups: comp.os.vxworks
Subject: SCSI ON VxWorks
Date: Fri, 25 Aug 2000 11:58:02 GMT
From: ravi_dubey@techie.com
Organization: Deja.com - Before you buy.
Message-ID: <8o5n03$b2h$1@nnrp1.deja.com>

Hi All !!

I have to write a SCSI Driver for the SYMBIOS SYM53C810A card. Could
anyone suggest me some links or a book which will give me more
information about the same.

Thanx in anticipation
Ravi


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Canceling a select()
Date: Fri, 25 Aug 2000 14:02:59 +0200
From: Johan van der Werf 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <8o5nvs$b5n$1@overload.lbl.gov>


  Hi all,

 Question: what would be the best way to make a task return from a select()
           function call before the timeout is expired, or any data is sent?

 At first sight ioctl(fd, FIOUNSELECT, xxx) seemed a likely candidate, 
 unfortunately it can only be used from inside a driver, because the
 third argument must be a pointer to a SEL_WAKEUP_NODE. Please tell me if
 there is a standard way of doing this?

 We're using VxWorks 5.3.1 on M68K platform.

                        Regards,

- -- 
|Johan van der Werf   |   tel: ++31 (0)40 2684280 |$B%O%s%9(B  |
|ASM Lithography      |   fax: ++31 (0)40 2682224 |$B%t%!%s(B  |
|Dept: Basic Software | email: ujw@asml.nl        |$B%G%k(B    |
|Room: 7A3.012        |        vdwerf@xs4all.nl   |$B%&%(%k%U(B|

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Strange interrupt behavior noticed with WindView
Date: Fri, 25 Aug 2000 08:09:24 -0400
From: Frederick Byrne 
Organization: Lockheed Martin - GES
Message-ID: <39A661F4.E970927A@lmco.com>
References: <39A3E0F1.37C79E65@lmco.com> <39A5EAC3.1CC2F519@cisco.com>

Vijay Kumar Peshkar wrote:
> 
> Frederick,
> 
> No clue.
> You have to give me some description of h/w and s/w configuration at least to
> guess.
> 
> regards
> Vijay
> 
> Frederick Byrne wrote:
> 
> > We've noticed in a Windview display that occasionally an interrupt handler (it
> > appears to be a vxWorks handler and not one we've defined) will keep the CPU
> > until the OS clock tick has expired (16.6 mS).  The specific interrupt number
> > changes but the behavior is the same.  Does anyone have any ideas?  Thanks in
> > advance for any help.
> >
> > --
> > ==============================================================
> >         Frederick "Rick" Byrne  frederick.j.byrne@lmco.com
> >         Lead Member of the Engineering Staff
> >         Lockheed Martin NE&SS, Moorestown NJ
> >         Phone: (856) 722-7344   Fax: (856) 722-4103

We're using vxWorks 5.3.1, Torndado 1.1 for the PPC.  We're using a Synergy
VGM5 dual processor board - this is 2 300MHz PC 750.  I talked with Synergy's
BSP person and he said that it is a problem with Windview.  It does not handle
the multiprocessor situation.  Apparently this is not an accurate indication
of what is going on.   Thanks.

- -- 
==============================================================
	Frederick "Rick" Byrne	frederick.j.byrne@lmco.com
	Lead Member of the Engineering Staff
	Lockheed Martin NE&SS, Moorestown NJ
	Phone: (856) 722-7344	Fax: (856) 722-4103

---------------------------

Newsgroups: comp.os.vxworks
Subject: vxSim problem
Date: Fri, 25 Aug 2000 08:45:53 -0400
From: Bhavesh Patel 
Organization: Posted via Supernews, http://www.supernews.com
Message-ID: <39A66A81.D964EE16@yahoo.com>

Hi all,

I want to built a vxsim image so that I can do the following things -

1) set up udp sessions by making calls to socket(), bind(),..

Which network components do I need to include in the vxsim image so that
I can successfully setup udp session from my downloadable application.
Secondly, Do I need to use port numbers > 5000 for the datagram socket
I create in vxworks?
If I create a datagram socket with a port number 161 and 162 will it
fail at bind()?

Any help is appreciated.

bhavesh


---------------------------

Newsgroups: comp.os.vxworks
Subject: Have you used VxVMI?
Date: Fri, 25 Aug 2000 08:50:53 -0400
From: Michael Simpson 
Organization: UUNET Canada News Reader Service
Message-ID: <39A66BAD.9E65E47A@ddt.com>

Hi there,

Does anybody out there have any experience using VxVMI to protect text
and interrupt vector memory?  I have contacted WRS to find out more
about VxVMI as the documentation I have is thin.  Is there a run-time
cost associated with using this add-on?

Thx,

Mike


---------------------------

Newsgroups: comp.os.vxworks
Subject: Loading application module during VxWorks.st boot
Date: Fri, 25 Aug 2000 09:28:38 -0400
From: "Alexander Povolotsky" 
Organization: Lucent Technologies, Columbus, Ohio
Message-ID: <8o5scd$6fv@nntpb.cb.lucent.com>

Hello,

We are using VxWorks 5.3.1 and Tornado 1.0.1 on PPC604e (MTX bsp).
We are using loadModule (to load application object module nxRCP.o on to the
target) at the end of usrRoot in a vxWorks.st image -
could see this module loaded by executing moduleShow on the target
shell, then connect to the Tornado target server/Launcher and do not
see these symbols from nxRCP.o to be picked up by synching the host and
target symbol tables
- - therefore cannot debug with CrossWind.

1) Target Shell
RCP-0->moduleShow

MODULE NAME     MODULE ID  GROUP #    TEXT START DATA START  BSS START
- --------------- ---------- ---------- ---------- ----------
    nxRCP.o  0x3475148          1  0x2a2e000  0x2929b18  0x2921a90
    spfx0000.o  0x3ffc4f0       2  0x17e6000  0x17c8c38 0xffffffff
    bgpx0000.o  0x3ff6928       3  0x1656000  0x1635018 0xffffffff
    ripx0000.o  0x17dcb20       4  0x15d4000  0x15cf620 0xffffffff
    isis00000.o  0x162f078      5  0x14b4000  0x148e080  0x148a068
value = 0 = 0x0
RCP-0->

2) WindShell (and same is being seen by the CrossWind)
- -> moduleShow

MODULE NAME MODULE ID  GROUP #    TEXT START DATA START  BSS START
- --------------- ---------- ---------- ---------- ----------
NxRcp.st    0x8a398          1   0x100000   0x1db648   0x2024b4
spfx0000.o  0x9e610          2  0x17e6000  0x17c8c38
bgpx0000.o  0xb6b78          3  0x1656000  0x1635018
ripx0000.o  0xc21b0          4  0x15d4000  0x15cf620
isis00000.o 0xc3870          5  0x14b4000  0x148e080  0x148a068
value = 0 = 0x0
- ->
That what I see in the log (no explicit references to  NxRcp.st,  which I
see in Windsh->moduleShow
and in a CrossWind. Also it shows no references to nxRCP.o, which I see on
the targetShell->moduleShow )

tgtsvr (nx1@yarmouth): Tue Aug 22 11:06:31 2000
    License request... authorized on Local Host.
    Warning: Can't create the log file
    Warning: Continue without logging the requests
    Attaching backend... succeeded.
    Connecting to target agent... succeeded.
    Attaching C++ interface... succeeded.
    Attaching elf OMF reader... succeeded.
    Added target_modules         to target-server.....done
    Added spfx0000.o             to target-server.....done
    Added bgpx0000.o             to target-server.....done
    Added ripx0000.o             to target-server.....done
    Added isis00000.o            to target-server.....done

Does anyone know how to debug object files loaded directly by the
VxWorks initialization code before the Tornado target server has been
 connected (without breaking/interrupting/delaying this sequence of loading)
?.
What specifics brings to this picture the fact, that standalone
VxWorks.st image is used (it contains symbol table in it) - in
fact we are loading it onto the target over network from the Unix
Solaris host (sorry for such misuse ...).

Another related question - what "strip" does to the VxWorks.st
image which was compiled by gcc with "-g" flag ?



---------------------------

Newsgroups: comp.os.vxworks
Subject: Change # of fds and select limit
Date: Fri, 25 Aug 2000 09:24:37 -0400
From: Yi Qin 
Organization: Storm Internet Services
Message-ID: <39A67394.13D439EE@ss8networks.com>
Sender: ss8@gw-ss8networks.storm.ca

This is probably discussed before, any info on this will be very much
appreciated.

I want to increase the size of the default limit on number of files that

can be opened per vxworks node from 64 to 1024. What are the files
that I need to touch to modify the constants? Any performance issues
I should be noted?

Also, I want to have each select() call to cover upto 1024 fds. What
has been done by the community to go beyond the vxworks limit? I
think they use an unsigned char for the fd set. (Correct me if I am
wrong.)

Any suggestion will be greatly appreciated.


Yi Qin


---------------------------

Newsgroups: comp.os.vxworks
Subject: More PCI Bus & VXWorks
Date: Fri, 25 Aug 2000 13:32:23 GMT
From: "Shawn Girard" 
Organization: Excite@Home - The Leader in Broadband
Message-ID: 

Please bear with me if this question has already been asked, but I am new in
working with the PCI bus and I am having problems in accessing the device
through its assigned BARs.

I am currently writing a device driver for a PCI bridge chip, to be exact it
is the TI PCI2040 PCI-DSP bridge chip through VxWorks on a Radstone PPC4.

I ask some folks earlier in how are the Bars in the PCI Configuration Header
are assigned and they informed me it is done in the Auto Config  routine
that is called in sysHwInit routine of the Kernel.  Working off this, I can
successfully access the PCI Config Header through the Radstone's API, which
shown below.

- -> sysPciDevShow 0,6,0
0x00 DeviceID       0xac60104c Texas Instruments
0x04 Command        0x0142
0x06 Status         0x0210
0x08 RevisionID     0x00
0x09 ProgIF         0x00 SubClass 0x80 BaseClass 0x06
                    Bridge, Other
0x0c CacheLineSize  0x08
0x0d LatencyTimer   0x00
0x0e HeaderType     0x00
0x0f BIST           0x00
0x10 BaseAddresses  0x01048000 0x01040000 0x00000000
                    0x00000000 0x00000000 0x00000000
0x2c Sub-System ID  0x00000000
0x30 BaseAddressROM 0x00000000
0x3c InterruptLine  0x09
0x3d InterruptPin   0x01
0x3e Min_Gnt        0x00
0x3f Max_Lat        0x00
0x40 0x00000000 0x00001f00 0x00000000 0x0000c10f
0x50 0xfe115c01 0x00000000 0x00000000 0x00000006
0x60 0x00000000 0x00000000 0x00000000 0x00000000
0x70 0x00000000 0x00000000 0x00000000 0x00000000
0x80 0x00000000 0x00000000 0x00000000 0x00000000
0x90 0x00000000 0x00000000 0x00000000 0x00000000
0xa0 0x00000000 0x00000000 0x00000000 0x00000000
0xb0 0x00000000 0x00000000 0x00000000 0x00000000
0xc0 0x00000000 0x00000000 0x00000000 0x00000000
0xd0 0x00000000 0x00000000 0x00000000 0x00000000
0xe0 0x00000000 0x00000000 0x00000000 0x00000000
0xf0 0x00000000 0x00000000 0x00000000 0x00000000

Being aware that I have to translate the Bar address to the host address, I
apply the needed translatation for allocated PCI/IO Space of the host.  That
gives me the following host address for Bar_0 0x81048000.

Being aware that this is a window to the devices register, I try doing the
address dump to see what is there.  Please note, at this point I never
loaded in my device driver in yet.  This the result of my dump:

81048000:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
81048010:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
81048020:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
81048030:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
81048040:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
81048050:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
81048060:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
81048070:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
81048080:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
81048090:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
810480a0:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
810480b0:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
810480c0:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
810480d0:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
810480e0:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
810480f0:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
value = 0 = 0x0

Now for the million dollar question, why am I seeing nothing but ones?  Am I
accessing the memory location correctly?  The reason why I know that I am
accessing it wrong is that I am not seeing the default values for some of
the registers.

Is there something I should be doing with translating the PCI address to
Host address.  What I recall is that the PCI/IO Space in PCI address starts
at 0x1000000 and the host starts at 0x81000000.

Thanks in advance
Shawn Girard



---------------------------

Newsgroups: comp.os.vxworks
Subject: Shared Memory Network booting problems
Date: Fri, 25 Aug 2000 14:04:09 GMT
From: Joe Andrews 
Organization: Vehicle Managememt Systems
Message-ID: <39A67CD9.24482896@mw.boeing.com>
Reply-To: joe.andrews@mw.boeing.com
Sender: nntp@news.boeing.com (Boeing NNTP News Access)

I'm trying to set up a shared memory network between 2 MVME2604 cards
running VxWorks. I used the original BSP source files and compiled a
bootrom & Vxworks image and D/led to each card. (WRS said it would work
fine this way). The first card boots from the ethernet fine and
initializes the backplane anchor, heartbeat works. The second card,
specified to boot from sm, sees the backplane anchor at the proper
location and tries to boot, but can never load VxWorks. After a couple
of minutes, I get 'Error cannot load file   error=0xd0003'

I'm pretty sure I have the gateway set correctly on the first card, and
while the second is trying to boot I can ping it using the backplane IP
address. 

I'm not so sure my Windows NT host has routing set up correctly. Could
this be the problem? When I tried to add a route, it fails with an error
code of 87, whatever that means.

Thanks for any help...

---------------------------

Newsgroups: comp.os.vxworks
Subject: re: Canceling a select()
Date: Fri, 25 Aug 2000 15:33:53 +0100 (BST)
From: David Laight 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <200008251433.PAA11811@dsl-2.tadpole.co.uk>

You could add an extra file descriptor into the select list and use that driver 
to generate a read event.  Maybe a writing a character to a pipe will work (does 
the pipe driver support select?).

Alternatively if one of your own drivers is in the select list you could make it 
wake up all the people waiting for data.  Either using an iocyt() or, since 
vxWorks has open plan access permissions, a direct call of a driver routine.

Indeed if you can find (or know) the address of the SEL_WAKEUP_NODE calling 
selWakeupAll() itself will do.

For a typical serial driver (ie something that uses tyLib):

#include 
kick_select( int fd )
{
 selWakeupAll( ((TY_DEV *)fdTable[ fd - 3 ].pDevHdr)->selWakeupList, SELREAD );
}

should have the desired effect - If you don't mind using some 'private' 
information.

	David
	
>  Question: what would be the best way to make a task return from a select()
>            function call before the timeout is expired, or any data is sent?
> 
>  At first sight ioctl(fd, FIOUNSELECT, xxx) seemed a likely candidate, 
>  unfortunately it can only be used from inside a driver, because the
>  third argument must be a pointer to a SEL_WAKEUP_NODE. Please tell me if
>  there is a standard way of doing this?
> 
>  We're using VxWorks 5.3.1 on M68K platform.
> 
>                         Regards,
> 


- ----------------------------------------------------------------
David Laight                          email: dsl@tadpole.co.uk
Tadpole Technology plc                phone: +44 1223 428 232
Cambridge, UK                         fax:   +44 1223 428 201


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Shared Memory Network booting problems
Date: Fri, 25 Aug 2000 09:33:03 -0500
From: Greg Willden 
Organization: Posted via Supernews, http://www.supernews.com
Message-ID: <39A6839F.41D927BC@swri.edu>
References: <39A67CD9.24482896@mw.boeing.com>

> I'm not so sure my Windows NT host has routing set up correctly. Could
> this be the problem? When I tried to add a route, it fails with an error
> code of 87, whatever that means.

I say that is the answer.  Your NT machine needs to know who to send the
data back to so you need to add a route to the routing tables with route
add.  If you want it to be permanent you need to add the '-p' flag.
Something like this may help

route add  mask 255.255.255.0 

the mask may need to be 255.255.0.0 depending on your network configuration.

Good luck


---------------------------

Newsgroups: comp.os.vxworks
Subject: DDCMP for vxWorks
Date: Fri, 25 Aug 2000 14:47:00 GMT
From: jsisensors@on.aibn.com (Will MacGregor)
Organization: Sympatico
Message-ID: <39a685c2.346873786@news21.on.aibn.com>

Would anyone out there have any information on an implementation of
DDCMP (Digital Data Communications Message Protocol) for vxWorks?
Will MacGregor
JATOM SYSTEMS INC
Tel: 613 738-3349 x229
email: jsisensors@on.aibn.com

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Shared Memory Network booting problems
Date: Fri, 25 Aug 2000 12:44:38 -0500
From: Andrew Johnson 
Organization: APS, Argonne National Laboratory
Message-ID: <39A6B086.6C69CAC3@aps.anl.gov>
References: <39A67CD9.24482896@mw.boeing.com>

Joe Andrews wrote:
> 
> I'm not so sure my Windows NT host has routing set up correctly. Could
> this be the problem? When I tried to add a route, it fails with an error
> code of 87, whatever that means.

If you can I highly recommend that you build your vxWorks images with the
Proxy ARP stuff enabled - read the network programming section on this. 
Doing so you don't need to do anything special with routes and gateways,
the sm networking takes care of everything you need.  It is even possible
to create a single vxWorks image that will work for both proxy master
(cpu0) and slave (cpu1), although the WRS documentation states that you
have to #define either PROXY_ARP_MASTER or PROXY_ARP_SLAVE; that may only
be true if you're relying on the default boot parameters, which you don't
need to do on the MVME2603.

- - Andrew
- -- 
Every great idea appears crazy in the beginning.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Canceling a select()
Date: Fri, 25 Aug 2000 16:48:34 GMT
From: John Fusco 
Organization: Deja.com - Before you buy.
Message-ID: <8o680p$4r$1@nnrp1.deja.com>
References: <8o5nvs$b5n$1@overload.lbl.gov>

In Unix, a process returns from select() if it gets a signal that is not
blocked.  In this case select() returns -1 and errno==EINTR.  This
should work the same way in VxWorks, but I don't know for sure.

In article <8o5nvs$b5n$1@overload.lbl.gov>,
  Johan van der Werf  wrote:
>
>   Hi all,
>
>  Question: what would be the best way to make a task return from a
select()
>            function call before the timeout is expired, or any data is
sent?
>
>  At first sight ioctl(fd, FIOUNSELECT, xxx) seemed a likely candidate,
>  unfortunately it can only be used from inside a driver, because the
>  third argument must be a pointer to a SEL_WAKEUP_NODE. Please tell me
if
>  there is a standard way of doing this?
>
>  We're using VxWorks 5.3.1 on M68K platform.
>
>                         Regards,
>
> --
> |Johan van der Werf   |   tel: ++31 (0)40 2684280 |$B%O%s%9(B  |
> |ASM Lithography      |   fax: ++31 (0)40 2682224 |$B%t%!%s(B  |
> |Dept: Basic Software | email: ujw@asml.nl        |$B%G%k(B    |
> |Room: 7A3.012        |        vdwerf@xs4all.nl   |$B%&%(%k%U(B|
>


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: What is error S_objLib_OBJ_ID_ERROR ?
Date: Fri, 25 Aug 2000 19:01:02 GMT
From: john_94501@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8o6fp0$9uv$1@nnrp1.deja.com>
References: <8o3qek$j44$1@newsfeed.cv.nrao.edu>

Hi,

> I get this error when I execute select() on the 2-nd serial port
(COM2) of
> an MVME-2700.  Here's a snippet of code (which works fine):

This is a subject that has come up often, and must be in the FAQ by
now... Anyway, your last comment: "which works fine" is the clue. You
should not take any notice of errno unless a system call you make
explicitly tells you there was an error. The errno value is undefined in
all cases other than when a function (a) returns an error state *AND*
(b) states in the docs that it will set errno to indicate the cause of
the error.

Hope that helps,

John...



Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Strange interrupt behavior noticed with WindView
Date: Fri, 25 Aug 2000 20:05:14 GMT
From: Ioan.Marusca@indsys.ge.com
Organization: Deja.com - Before you buy.
Message-ID: <8o6jhb$eq2$1@nnrp1.deja.com>
References: <39A3E0F1.37C79E65@lmco.com> <39A5EAC3.1CC2F519@cisco.com> <39A661F4.E970927A@lmco.com>

Hi,

I've seen on my board (single CPU PPC860) same thing. We had many, many
problems with WindView and even after we fixed many BSP problems and
apply all known and discovered patches still .wv files are not showing
correctly. You have to use it only in an approximative way.

I'm curious to compare a *.wv file from your board. Can you send me a
file ? Thank you.

Ioan

In article <39A661F4.E970927A@lmco.com>,
  Frederick Byrne  wrote:
> Vijay Kumar Peshkar wrote:
> >
> > Frederick,
> >
> > No clue.
> > You have to give me some description of h/w and s/w configuration
at least to
> > guess.
> >
> > regards
> > Vijay
> >
> > Frederick Byrne wrote:
> >
> > > We've noticed in a Windview display that occasionally an
interrupt handler (it
> > > appears to be a vxWorks handler and not one we've defined) will
keep the CPU
> > > until the OS clock tick has expired (16.6 mS).  The specific
interrupt number
> > > changes but the behavior is the same.  Does anyone have any
ideas?  Thanks in
> > > advance for any help.
> > >
> > > --
> > > ==============================================================
> > >         Frederick "Rick" Byrne  frederick.j.byrne@lmco.com
> > >         Lead Member of the Engineering Staff
> > >         Lockheed Martin NE&SS, Moorestown NJ
> > >         Phone: (856) 722-7344   Fax: (856) 722-4103
>
> We're using vxWorks 5.3.1, Torndado 1.1 for the PPC.  We're using a
Synergy
> VGM5 dual processor board - this is 2 300MHz PC 750.  I talked with
Synergy's
> BSP person and he said that it is a problem with Windview.  It does
not handle
> the multiprocessor situation.  Apparently this is not an accurate
indication
> of what is going on.   Thanks.
>
> --
> ==============================================================
> 	Frederick "Rick" Byrne	frederick.j.byrne@lmco.com
> 	Lead Member of the Engineering Staff
> 	Lockheed Martin NE&SS, Moorestown NJ
> 	Phone: (856) 722-7344	Fax: (856) 722-4103
>


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Write Protection for Task Stacks
Date: Fri, 25 Aug 2000 18:57:04 GMT
From: john_94501@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8o6fhk$9q9$1@nnrp1.deja.com>
References: <8o4ld8$6ia$1@nnrp1.deja.com>

Hello,

This is impossible to do using switch hooks since the scheduling code is
executed using the "old" task stack (and that, by the way, includes
your switch hook function). By write protecting it you effectively
remove the stack from the hook routine and the scheduler. This is not
going to work well, as you have discovered.

The only way around this would be to buy the source for the scheduler
code and modify it such that it no longer needed the old task's stack.
This is not a trivial change, though it would be possible. It would also
need a fairly detailed understanding of how the scheduler works.

Regards,

John...

In article <8o4ld8$6ia$1@nnrp1.deja.com>,
  madowdy@my-deja.com wrote:
> We may be insane to try this but we have a highly reliable
> application and we'd like to provide additional protection
> from our own indiscretions by write protecting the stack
> of all but the current task. We've taken care of allocating
> and sizing the stacks to ensure that they are placed on page
> boundaries and are an even number of pages long. In our
> first cut at the switch hook, we write-enabled the stack
> of pNewTcb and write-protected the stack of pOldTcb. This
> failed (I guess I can see how the stack of the "old" task
> could be used by the task switching code after the call
> to a switch hook). In our next version of the switch hook,
> we saved the TCB of the penultimate task (the one that ran
> before "old"). If the TCB of the penultimate task did not
> match the TCB of the "new" task, we write protected the
> stack of the penultimate TCB. This too failed (i.e. caused
> write protect violations).
>
> We don't have a source license so it's hard to know what
> WRS is doing around the calls to the task switch hooks.
> Would someone with the source let me know what is going
> on with the task stacks around the time the switch hooks
> are called. Specifically, does WRS modify the stack of
> the "new" task prior to the calls to the switch hooks.
> Thanks.
>
> Mark
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.
>


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Vxsim File I/O with NT
Date: Fri, 25 Aug 2000 20:22:01 GMT
From: jim.themis@genband.com
Organization: Deja.com - Before you buy.
Message-ID: <8o6kgi$fvu$1@nnrp1.deja.com>

I reviewed the archives but I could find anyone with a similar problem.
I am running VxSim on an NT box.  I am trying to do file I/O via fopen,
fclose, fputs, etc.  I am able to read files, but I am unable to create
and write files.  I have tried adding "host:" to the beginning of my
path.  I have enabled tsfs with r/w.  I get no error when I perform
an fopen.  However, when I execute an fputs(), I get the following
error:
ntPassFsLib: unknown ioctl = 0x14

I looked at errno and the value showed Unknown Request.

Can anybody explain their set-up for reading and WRITE files under
VxSim.


Thanks,
Jim Themis


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: mpc75x+mpc107 BSP yet?
Date: Sat, 26 Aug 2000 00:01:22 GMT
From: hornsbyk@att.net
Organization: Deja.com - Before you buy.
Message-ID: <8o71c9$uoq$1@nnrp1.deja.com>

I'm looking for any information on VxWorks's support for the MPC75x and
the MPC107.

Evidently the MPC107 is not supported in any way.  Is there ANY BSP
that contains the support for the MPC107?

Any information would be GREATLY appreciated



Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: ppc: rework of intLock/intUnlock for nesting?
Date: Sat, 26 Aug 2000 00:26:25 -0400
From: Al Johnston 
Organization: MindSpring Enterprises
Message-ID: <39A746F1.7B16B2EF@mindspring.com>

Has anyone re-coded a version if the intLock() and intUnlock() routines
that allow nested calls to these routines without prematurely unlocking
interrupts?  Seems to me that this should be possible by using the
"key" parameter... 

I was surprised that vxw doesnt already do this.  Is there a problem
with doing this that I am missing?

I am hoping for a c routines (with asm in it); but I haven't gotten
to the point of knowing how to writing ppc asm that can be compiled by 
gcc.  Playing with interrupts seems like a terrible place to try to
learn!

pointers to "how to" pages on writing ppc asm within gcc c 
code would be welcome to.  I have the asm code (courtesy of wrs)
for the vxw supplied int routines.

as always, thanks guys.

---------------------------

End of New-News digest
**********************


From vxwexplo-errs@csg.lbl.gov  Sun Aug 27 04:03:06 2000
From: Vxworks Exploder 
Date: Sun Aug 27 04:03:09 PDT 2000
Subject: comp.os.vxworks newsdigest

Comp.Os.Vxworks Daily Digest    Sun Aug 27 04:03:03 PDT 2000

        Subject: Loading a device driver
        Subject: R: More PCI Bus & VXWorks
        Subject: Re: More PCI Bus & VXWorks
        Subject: loopback
        Subject: Re: Multiple Developers on same project (How to?)
        Subject: writing a device driver

-------------------------------------------------------

Newsgroups: comp.os.vxworks
Subject: Loading a device driver
Date: Sat, 26 Aug 2000 12:30:38 +0100
From: "Merlino" 
Organization: TIN
Message-ID: <8o86fn$qpk$1@nslave2.tin.it>
Reply-To: "Merlino" 

Hi guys,

i have an x86  board with a P2 BSP vxWorks 5.4;
i would like to run the system this way:
1- loading vxworks image ( without any device driver )
2- scanning the PCI bus for known device
3- loading modules with the suitables device drivers
4- activating device drivers

Is it possible ?
How ?

Thanks



---------------------------

Newsgroups: comp.os.vxworks
Subject: R: More PCI Bus & VXWorks
Date: Sat, 26 Aug 2000 15:34:44 +0100
From: "Merlino" 
Organization: TIN
Message-ID: <8o8h8t$6kg$1@nslave2.tin.it>
References: 
Reply-To: "Merlino" 

Pci registers are ALWAYS little endian ...
may be you do not convert from big to little endian when converting from
physical to virtual address ( host address )

I bet this is "the million dollar" answer.

Bye
Merlin

Shawn Girard  wrote in message
Hzup5.27054$eR5.885737@news1.rdc1.on.wave.home.com...
> Please bear with me if this question has already been asked, but I am new
in
> working with the PCI bus and I am having problems in accessing the device
> through its assigned BARs.
>
> I am currently writing a device driver for a PCI bridge chip, to be exact
it
> is the TI PCI2040 PCI-DSP bridge chip through VxWorks on a Radstone PPC4.
>
> I ask some folks earlier in how are the Bars in the PCI Configuration
Header
> are assigned and they informed me it is done in the Auto Config  routine
> that is called in sysHwInit routine of the Kernel.  Working off this, I
can
> successfully access the PCI Config Header through the Radstone's API,
which
> shown below.
>
> -> sysPciDevShow 0,6,0
> 0x00 DeviceID       0xac60104c Texas Instruments
> 0x04 Command        0x0142
> 0x06 Status         0x0210
> 0x08 RevisionID     0x00
> 0x09 ProgIF         0x00 SubClass 0x80 BaseClass 0x06
>                     Bridge, Other
> 0x0c CacheLineSize  0x08
> 0x0d LatencyTimer   0x00
> 0x0e HeaderType     0x00
> 0x0f BIST           0x00
> 0x10 BaseAddresses  0x01048000 0x01040000 0x00000000
>                     0x00000000 0x00000000 0x00000000
> 0x2c Sub-System ID  0x00000000
> 0x30 BaseAddressROM 0x00000000
> 0x3c InterruptLine  0x09
> 0x3d InterruptPin   0x01
> 0x3e Min_Gnt        0x00
> 0x3f Max_Lat        0x00
> 0x40 0x00000000 0x00001f00 0x00000000 0x0000c10f
> 0x50 0xfe115c01 0x00000000 0x00000000 0x00000006
> 0x60 0x00000000 0x00000000 0x00000000 0x00000000
> 0x70 0x00000000 0x00000000 0x00000000 0x00000000
> 0x80 0x00000000 0x00000000 0x00000000 0x00000000
> 0x90 0x00000000 0x00000000 0x00000000 0x00000000
> 0xa0 0x00000000 0x00000000 0x00000000 0x00000000
> 0xb0 0x00000000 0x00000000 0x00000000 0x00000000
> 0xc0 0x00000000 0x00000000 0x00000000 0x00000000
> 0xd0 0x00000000 0x00000000 0x00000000 0x00000000
> 0xe0 0x00000000 0x00000000 0x00000000 0x00000000
> 0xf0 0x00000000 0x00000000 0x00000000 0x00000000
>
> Being aware that I have to translate the Bar address to the host address,
I
> apply the needed translatation for allocated PCI/IO Space of the host.
That
> gives me the following host address for Bar_0 0x81048000.
>
> Being aware that this is a window to the devices register, I try doing the
> address dump to see what is there.  Please note, at this point I never
> loaded in my device driver in yet.  This the result of my dump:
>
> 81048000:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> 81048010:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> 81048020:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> 81048030:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> 81048040:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> 81048050:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> 81048060:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> 81048070:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> 81048080:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> 81048090:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> 810480a0:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> 810480b0:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> 810480c0:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> 810480d0:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> 810480e0:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> 810480f0:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> value = 0 = 0x0
>
> Now for the million dollar question, why am I seeing nothing but ones?  Am
I
> accessing the memory location correctly?  The reason why I know that I am
> accessing it wrong is that I am not seeing the default values for some of
> the registers.
>
> Is there something I should be doing with translating the PCI address to
> Host address.  What I recall is that the PCI/IO Space in PCI address
starts
> at 0x1000000 and the host starts at 0x81000000.
>
> Thanks in advance
> Shawn Girard
>
>



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: More PCI Bus & VXWorks
Date: Sat, 26 Aug 2000 17:06:35 GMT
From: "Shawn Girard" 
Organization: Excite@Home - The Leader in Broadband
Message-ID: 
References:  <8o8h8t$6kg$1@nslave2.tin.it>

That I thought, but the hardware engineer convice me that was not the case.
One would have thought the address would be corrected through the
sysPciDevShow function call.

Anyway, it be worth the try to switch the address around.  I just have to
check the host board memory mapping to determine that is a valid address.

I will keep you posit how it goes.

Thanks a lot
Shawn
"Merlino"  wrote in message
news:8o8h8t$6kg$1@nslave2.tin.it...
> Pci registers are ALWAYS little endian ...
> may be you do not convert from big to little endian when converting from
> physical to virtual address ( host address )
>
> I bet this is "the million dollar" answer.
>
> Bye
> Merlin
>
> Shawn Girard  wrote in message
> Hzup5.27054$eR5.885737@news1.rdc1.on.wave.home.com...
> > Please bear with me if this question has already been asked, but I am
new
> in
> > working with the PCI bus and I am having problems in accessing the
device
> > through its assigned BARs.
> >
> > I am currently writing a device driver for a PCI bridge chip, to be
exact
> it
> > is the TI PCI2040 PCI-DSP bridge chip through VxWorks on a Radstone
PPC4.
> >
> > I ask some folks earlier in how are the Bars in the PCI Configuration
> Header
> > are assigned and they informed me it is done in the Auto Config  routine
> > that is called in sysHwInit routine of the Kernel.  Working off this, I
> can
> > successfully access the PCI Config Header through the Radstone's API,
> which
> > shown below.
> >
> > -> sysPciDevShow 0,6,0
> > 0x00 DeviceID       0xac60104c Texas Instruments
> > 0x04 Command        0x0142
> > 0x06 Status         0x0210
> > 0x08 RevisionID     0x00
> > 0x09 ProgIF         0x00 SubClass 0x80 BaseClass 0x06
> >                     Bridge, Other
> > 0x0c CacheLineSize  0x08
> > 0x0d LatencyTimer   0x00
> > 0x0e HeaderType     0x00
> > 0x0f BIST           0x00
> > 0x10 BaseAddresses  0x01048000 0x01040000 0x00000000
> >                     0x00000000 0x00000000 0x00000000
> > 0x2c Sub-System ID  0x00000000
> > 0x30 BaseAddressROM 0x00000000
> > 0x3c InterruptLine  0x09
> > 0x3d InterruptPin   0x01
> > 0x3e Min_Gnt        0x00
> > 0x3f Max_Lat        0x00
> > 0x40 0x00000000 0x00001f00 0x00000000 0x0000c10f
> > 0x50 0xfe115c01 0x00000000 0x00000000 0x00000006
> > 0x60 0x00000000 0x00000000 0x00000000 0x00000000
> > 0x70 0x00000000 0x00000000 0x00000000 0x00000000
> > 0x80 0x00000000 0x00000000 0x00000000 0x00000000
> > 0x90 0x00000000 0x00000000 0x00000000 0x00000000
> > 0xa0 0x00000000 0x00000000 0x00000000 0x00000000
> > 0xb0 0x00000000 0x00000000 0x00000000 0x00000000
> > 0xc0 0x00000000 0x00000000 0x00000000 0x00000000
> > 0xd0 0x00000000 0x00000000 0x00000000 0x00000000
> > 0xe0 0x00000000 0x00000000 0x00000000 0x00000000
> > 0xf0 0x00000000 0x00000000 0x00000000 0x00000000
> >
> > Being aware that I have to translate the Bar address to the host
address,
> I
> > apply the needed translatation for allocated PCI/IO Space of the host.
> That
> > gives me the following host address for Bar_0 0x81048000.
> >
> > Being aware that this is a window to the devices register, I try doing
the
> > address dump to see what is there.  Please note, at this point I never
> > loaded in my device driver in yet.  This the result of my dump:
> >
> > 81048000:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> > 81048010:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> > 81048020:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> > 81048030:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> > 81048040:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> > 81048050:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> > 81048060:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> > 81048070:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> > 81048080:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> > 81048090:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> > 810480a0:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> > 810480b0:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> > 810480c0:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> > 810480d0:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> > 810480e0:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> > 810480f0:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> > value = 0 = 0x0
> >
> > Now for the million dollar question, why am I seeing nothing but ones?
Am
> I
> > accessing the memory location correctly?  The reason why I know that I
am
> > accessing it wrong is that I am not seeing the default values for some
of
> > the registers.
> >
> > Is there something I should be doing with translating the PCI address to
> > Host address.  What I recall is that the PCI/IO Space in PCI address
> starts
> > at 0x1000000 and the host starts at 0x81000000.
> >
> > Thanks in advance
> > Shawn Girard
> >
> >
>
>



---------------------------

Newsgroups: comp.os.vxworks
Subject: loopback
Date: Sat, 26 Aug 2000 23:05:38 -0700
From: "Sandhya Potluri" 
Organization: Monmouth Internet
Message-ID: <8oa0cv$2o3$1@news.monmouth.com>

Hi,
    Is there any specific inititlization I need to do to get loopback
working
on a ethernet interface?  Say I want to do a ping "127.0.0.1", on an
interface.
Sandy




---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Multiple Developers on same project (How to?)
Date: 27 Aug 2000 02:24:09 GMT
From: Diamand@btinternet.com (Luke Diamand)
Organization: Customer of Energis Squared
Message-ID: 
References: 

CVS is your friend. Get it from your nearest FSF mirror and all your
version control woes will fade away.

Luke Diamand

On Tue, 22 Aug 2000 15:22:07 -0400, Joe Ott  wrote:
>
>Tornado is a new developement environment to us and we are in the beginning
>phases of a project with multiple developers. The project
>(Tornado-workspace) will be seperated into various sub-projects
>(Tornado-projects). Our concerns are:
>
>1) There will be multiple developers working on different pieces of the same
>sub-project. The work of developer A will not interfer with the work of
>developer B but the combined work will compile to a single task. What are
>some methods that can be used to allow several developers to work on their
>seperate areas and still allow them to compile and test without waiting for
>the other developer to be finished?
>
>2) The various sub-projects (Tornado-projects) reflect various SW layers. In
>order to test the upper layers working copies of the lower layers are
>needed. What is the best way to support the upper layers while modifying the
>lower layers? In other words what is the best way to keep a copy of a
>working downloadable image (*.out file) other can use while you are making
>changes to the project?
>
>3) We will be using source control. Beside our source code, what if any
>Tornado files (workspace, project, makefile, etc) should be source
>controlled?
>
>One problem we have encounter with the Tornado project and workspace files
>is that if there are two developers working and they both make changes to
>the files (such as adding projects or adding files to the project) the last
>one to save wins. That is the changes made by the first one who saved the
>files are lost.
>
>Has anyone worked in an environment where there were many developers and you
>had to keep a working downloadable image for other to use while making
>changes to your own project?
>
>Any help or insight would be appreciated.
>
>Joe Ott
>VTG, an Intel Dialogic Company
>
>
>


- -- 

---------------------------

Newsgroups: comp.os.vxworks
Subject: writing a device driver
Date: Sun, 27 Aug 2000 05:36:55 GMT
From: ehudfl@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8oa9dm$ck8$1@nnrp1.deja.com>

Where can i find a web site who explain how to write device drivers in
vxworks and with source code who explain how to do that ?


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

End of New-News digest
**********************


From vxwexplo-errs@csg.lbl.gov  Mon Aug 28 04:03:08 2000
From: Vxworks Exploder 
Date: Mon Aug 28 04:03:11 PDT 2000
Subject: comp.os.vxworks newsdigest

Comp.Os.Vxworks Daily Digest    Mon Aug 28 04:03:04 PDT 2000

        Subject: Problems with vxworks dhcp server under load
        Subject: [q] Is gated working on Vxworks?
        Subject: Re: Project/linker question
        Subject: SNTP
        Subject: More PCI Bus & VXWorks
        Subject: Look! message: AGENT_EVENTPOINT_TABLE_FULL
        Subject: Memory window for PCMCIA
        Subject: memory window for accessing pcmcia!
        Subject: Re: SNTP

-------------------------------------------------------

Newsgroups: comp.os.vxworks
Subject: Problems with vxworks dhcp server under load
Date: Sun, 27 Aug 2000 21:03:00 GMT
From: mikecarlton@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8obvlp$5ck$1@nnrp1.deja.com>

I'm having lots of problems using the supported WIDE-based DHCP server under
vxworks. Under moderate load (300 active leases, 3 hour lease time) I see
lots of different problems.

The overall effect is that clients make requests and none get any response
(no offer) when the server gets in a funny state.

Has anyone seen behavior like this?  Suggestions on how to deal with it?

Also, can anyone recommend another DHCP server for Vxworks (commercial or
freeware would be ok).

Thanks,
- --Mike Carlton


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: [q] Is gated working on Vxworks?
Date: Mon, 28 Aug 2000 11:23:58 +0900
From: "Raphael Lee" 
Organization: Electronics and Telecommunications Research Institute (ETRI)
Message-ID: <8ocig0$qka$1@news.etri.re.kr>

hi..
Is gated program(router S/W) working on Vxworks?
If it has been to fixed , could I find it? where?
other route program working on?

see You

- --
- -----------------------------------------
Raphael Lee
Member of Research Staff
Wireless LAN Team
Router Technology Department
Switching & Transmission Technology Laboraory.
ETRI(Electronics and Telecommunications Research Institute)

Tel : +82 42 860 1205    Fax : +82 42 860 5213
Cellular Phone : +82 16 567 4545
E-mail : raphael@etri.re.kr
Homepage : http://routertech.etri.re.kr/~raphael/



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Project/linker question
Date: Mon, 28 Aug 2000 07:48:38 +0200
From: Rainer Stelzer 
Organization: Stoye GmbH, =?iso-8859-1?Q?K=F6ln?=
Message-ID: <39A9FD36.4338B748@stoye.de>
References: <8o3e5c$bp2$1@overload.lbl.gov> <8o41vp$a5s$3@news1.xs4all.nl>

Hi Johan,

heel bedankt for de informatie .

Daag.

Rainer

=====================================================
            Gebrüder Stoye GmbH
department: traffic control systems
      name: Rainer Stelzer
       adr: Longericher Strasse 177
      city: Cologne, Germany
       zip: 50739
     phone: ++49/221/2616-551
    mailto: r.stelzer@stoye.de
       www: http://www.stoye.de
=====================================================



---------------------------

Newsgroups: comp.os.vxworks
Subject: SNTP
Date: Mon, 28 Aug 2000 08:07:53 +0200
From: =?iso-8859-1?Q?J=FCrgen?= Braun 
Organization: T-Online
Message-ID: <39AA01B9.2F0B8A65@t-online.de>

I want to use SNTP under VxWorks. Does somebody has an example of a
server application and a client abblication for me?

Juergen


---------------------------

Newsgroups: comp.os.vxworks
Subject: More PCI Bus & VXWorks
Date: Mon, 28 Aug 2000 08:36:16 GMT
From: paul@mecalc.co.za (Paul Olivier)
Organization: The South African Internet Exchange
Message-ID: <39aa2445.6328319@modem.mecalc>

Hi

This solution worked for us:

Create a pointer *pMemBase which points
to a PCI memory space. Set pMemBase 
equal to the value of a memory space BAR.
Then perform the following calculation:

pMemBase |= CPU_PCI_ISA_MEM_ADRS;

CPU_PCI_ISA_MEM_ADRS is defined in rs60x.h.

This is the address you can use to access your
PCI memory space.

Paul Olivier 

---------------------------

Newsgroups: comp.os.vxworks
Subject: Look! message: AGENT_EVENTPOINT_TABLE_FULL
Date: Mon, 28 Aug 2000 11:07:05 +0200
From: Albert Wijnja 
Organization: MECO Equipment Engineers BV
Message-ID: <39AA2BB9.BE482C7E@meco.nl>
Reply-To: Albert.Wijnja@meco.nl

Hello,

When running my application via Look! I get an error message (about 1000
times)
    "Insert breakpoint fail: AGENT_EVENTPOINT_TABLE_FULL"
My question is: how do I increase the table (which table?) in order to
get rid of this message.
By the way, I am able to continue though.

Albert Wijnja



---------------------------

Newsgroups: comp.os.vxworks
Subject: Memory window for PCMCIA
Date: Mon, 28 Aug 2000 17:22:27 +0800
From: "kvd2" 
Organization: DCI HiNet
Message-ID: <8odbst$btm@netnews.hinet.net>

Dear All,

I wanna to open memory window for accessing pcmcia register and memory.
But the error message, as follow, is returned. What I have to do?
- -------------------------------------------------------------------------
WTX Error 0x100ca (AGENT_MEM_ACCESS_ERROR)
value = -1 = 0xffffffff
- -------------------------------------------------------------------------
My configuration
cpu:mpc8xx
pcmcia base address:0x04000000
                                   size =2k

I've add this to sysLib.c, but it won't work!
{
    (void *) PCMCIA_BASE_ADRS,
    (void *) PCMCIA_BASE_ADRS,
   PCMCIA_SIZE,    /* 4 k - DICE parameter */
   VM_STATE_MASK_VALID | VM_STATE_MASK_WRITABLE | VM_STATE_MASK_CACHEABLE,
    VM_STATE_VALID      | VM_STATE_WRITABLE      | VM_STATE_CACHEABLE_NOT
    },

kvd2
















---------------------------

Newsgroups: comp.os.vxworks
Subject: memory window for accessing pcmcia!
Date: Mon, 28 Aug 2000 17:31:12 +0800
From: "kvd2" 
Organization: DCI HiNet
Message-ID: <8odbtq$btm@netnews.hinet.net>

Dear All,

I wanna to open a memory window for accessing pcmcia register and common
memory. But the error message, as follow, is returned. What I have to
to?
- -------------------------------------------------------------------------
WTX Error 0x100ca (AGENT_MEM_ACCESS_ERROR)
value = -1 = 0xffffffff
- -------------------------------------------------------------------------
My configuration
cpu:mpc8xx
pcmcia base address:0x04000000
                                   size =2k

I've add this to sysLib.c, but it won't work!
{
    (void *) PCMCIA_BASE_ADRS,
    (void *) PCMCIA_BASE_ADRS,
   PCMCIA_SIZE,    /* 4 k - DICE parameter */
   VM_STATE_MASK_VALID | VM_STATE_MASK_WRITABLE | VM_STATE_MASK_CACHEABLE,
    VM_STATE_VALID      | VM_STATE_WRITABLE      | VM_STATE_CACHEABLE_NOT
    },

kvd2












---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: SNTP
Date: Mon, 28 Aug 2000 12:56:10 +0200
From: Joachim Pluschke 
Organization: Stoye GmbH
Message-ID: <39AA454A.F93E9CFE@stoye.de>
References: <39AA01B9.2F0B8A65@t-online.de>


- --------------224EBB7A49B7049626EF9925
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit



Jürgen Braun schrieb:

> I want to use SNTP under VxWorks. Does somebody has an example of a
> server application and a client abblication for me?
>
> Juergen

Hello Jürgen!

Just a few months ago I implemented a simple sntp client-server
connection between x86/Pentium based vx-targets using Tornado II.
I picked out some Information and code, the client is an uncompiled
sample showing it in general.

At least this is "easy" except a support-routine for the SNTP-server
which must be done by the user.

To Check independently whether it works or not, I used a "CareWare"-tool
that does work with the VxWorks SNTP components ( don't know why (the
author of the "CareWare"-tool has only made a fractional implementation
of sntp) but it runs!)
The Tool is: >About Time 4.8< (Win) to find at "www.arachnoid.com".
This tool is able to be a sever as well as a client.

Have a good luck, hope this helps you.

Joachim

- --
=====================================================
            Gebrüder Stoye GmbH
department: traffic control systems
      Name: Joachim Pluschke
       Adr: Longericher Strasse 177
      city: Cologne, Germany
       Zip: 50739
     phone: ++49/221/2616-534
    mailto: j.pluschke@stoye.de
       www: www.stoye.de
=====================================================

==========================
==========================
DESCRIPTION

Let's start at the beginning.....
===================
At first you have to include and configure the VxWorks - support of SNTP
in the target's bootimage.

With GUI configuration activate and configure(below there's something
about configuring)
/network components/networking protocols/networking applications/SNTP
 depending for which of the two services the target
will used.

In the Vx-Networking manual there's described how to use SNTP (see also
"RFC 1769"), but there was a lack of clear information how to do the
Clock-Hook routine needed by the SNTP server.
After several unsuccessful trys to correctly feed in a date into the
server I found a remark from windriver in "comp.os.vxworks" what was
really helpfull:

=======================================
>> Forum: comp.os.vxworks
 >> Thread: VxWorks SNTP server
 >> Message 2 of 3

 Subject:
       Re: VxWorks SNTP server
 Date:
       08/17/1999
 Author:
       Stephen Macmanus 


                                       << previous in thread  ·  next in
thread >>

 nogap@my-deja.com wrote:
 >
 > Hi,
 >
 > I am looking for some information or  example code for the hook
routine
 > that shall be suplied to the SNTP server of VxWorks.
 > This hook routine is in the following format:
 >  sntpsClockHook (int request, void *pBuffer);
 >

 This sample hook routine reports the time information based on
 the internal system clock:

 #include 
 #include 
 #include 

 #include "timers.h"

 #include 

 STATUS sampleClockHook
     (
     int request,
     void *buffer
     )
     {
     struct timespec timeval;
     long value;

     ULONG *pTime;

     if (request == SNTPS_ID)
         {
         strcpy ((char *)buffer, "pps");
         return (OK);
         }

     if (request == SNTPS_RESOLUTION)
         {
         if (clock_getres (CLOCK_REALTIME, &timeval) == -1)
             return (ERROR);

         value = timeval.tv_nsec;
         *((ULONG *)buffer) = (ULONG)value;
         }

     if (request == SNTPS_TIME)
         {
         if (clock_gettime (CLOCK_REALTIME, &timeval) == -1)
             return (ERROR);
         pTime = (ULONG *)buffer;

         /* Copy number of seconds since 1900 to timestamp. */

         *pTime++ = timeval.tv_sec + SNTP_UNIX_OFFSET;

         /* Convert nanoseconds to fractional part and copy to
timestamp. */

         *pTime = sntpsNsecToFraction (timeval.tv_nsec);
         }

     return (OK);
     }

 --
 ------------------
 Stephen Macmanus                         #include 
 Software Engineer
 Wind River Systems




=======================================
DESCRIPTION

My version is slightly different but has the same core-functionality
=======================================






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

**
**  includes
*/
#include 
#include 
#include 
#include 
#include 
#include 
#include 
#include 



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

**
**  typedefs
*/

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

**
**  local memory
*/

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

**
**  function prototypes
*/
STATUS sntpsTimeHook(int request, void *pBuffer);







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

**
**  name ...... :    sntpsClockHook()
**
**  description :  This is the >hook routine< that is called to access
**       the reference clock used by the SNTP Server.
**       The hook routine copys one of three settings
**       used by the server to construct outgoing NTP messages
**       into pBuffer according to the value of the request parameter.
**
**       request parameters:
**        SNTPS_ID to get the 32-bit reference identifier in the format
specified in RFC 1769  or
**        SNTPS_RESOLUTION to retrieve a 32-bit value containing the
clock resolution in nanoseconds.
**         That value will be used to determine the 8-bit signed integer
indicating the
**         clock precision (according to the format specified in RFC
1769) or
**        SNTPS_TIME to retrieve the current 64-bit NTP timestamp from
pBuffer in host byte order.
**
**
** >> THIS routine can be assigned directly by altering the value of
SNTPS_TIME_HOOK
** >> or can be installed by a call to the sntpsClockSet( ) routine.
** >> Until this routine is specified, the SNTP server will not provide
timestamp information!
**
**
**  returns ... :  OK if request setting available (or ERROR otherwise).

**
**
**  global var. :
*/


/*
**
**
Zeitformate:

C:\Tornado\target\h\types\vxTypesBase.h(84): #define
_TYPE_time_t           typedef unsigned long time_t

struct timespec
    {
    time_t tv_sec;  // seconds
    long tv_nsec;   // nanoseconds (0 - 1,000,000,000)
    };
**
mktime( ) - convert broken-down time into calendar time (ANSI)

SYNOPSIS

     time_t mktime
         (
         struct tm * timeptr // pointer to broken-down structure
         )
RETURNS

     The calendar time in seconds, or ERROR (-1) if calendar time cannot
be calculated.
**
**





SNTPS_ID:
=========
      1   pps       precision calibrated source, such as ATOM (atomic
                    clock), PPS (precision pulse-per-second source),
                    etc.
      1   service   generic time service other than NTP, such as ACTS
                    (Automated Computer Time Service), TIME (UDP/Time
                    Protocol), TSP (Unix Time Service Protocol), DTSS
                    (Digital Time Synchronization Service), etc.
      1   radio     Generic radio service, with callsigns such as CHU,
                    DCF77, MSF, TDF, WWV, WWVB, WWVH, etc.
      1   nav       radionavigation system, such as OMEG (OMEGA), LORC
                    (LORAN-C), etc.
      1   satellite generic satellite service, such as GOES
                    (Geostationary Orbit Environment Satellite, GPS
                    (Global Positioning Service), etc.
      2   address   secondary reference (four-octet Internet address of
                    the NTP server)





SNTPS_RESOLUTION:
=================

       0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |LI | VN  |Mode |    Stratum    |     Poll      |   Precision   |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+


 LI:   -

 VN:   3  includes all NTP versions

 MODE:   4/5 [server/broadcast]

 STRATUM:  0/1/2-15 [unspec./primary ref./sec.ref]

 POLL INTERVALL: 6..10 [64s...1024s]

 PRECISION: This is an eight-bit signed integer indicating the
   precision of the local clock, in seconds to the nearest power of two.

   The values that normally appear in this field range from -6 for
   mains-frequency clocks to -20 for microsecond clocks found in some
   workstations.




ZUSAETZLICH muss SNTPS_MODE gesetzt werden: SNTP_PASSIVE (0x2), the
server waits for requests from clients
                oder: SNTP_ACTIVE (0x1), the server transmits NTP
timestamp information at fixed intervals.

    When executing in active mode, the SNTP server uses the
SNTPS_DSTADDR and SNTPS_INTERVAL
    definitions to determine the target IP address and broadcast
interval.

- ->
#define SNTPS_MODE SNTP_PASSIVE

und ein initial-aufruf von
sntpsClockSet((FUNCPTR)sntpsClockHook);

>>>>>>>(oder das o.g. 'zusaetzliche' bei der BSP-Modulselektion &
Konfiguration ueber das GUI)<<<<<<<<

*/




//#define SNTP_UNIX_OFFSET /*already defined, but where???*/


STATUS sntpsTimeHook(int request, void *pBuffer)
{


 struct timespec systemZeit;
 long    value;     //hilfsvariable
 ULONG *   pTime;     //hilfspointer




 switch(request)
 {
  case SNTPS_ID: /* put the 32-bit reference identifier in the format
specified in RFC 1769 into pBuffer*/
  {
   printf("sntpsTimeHook(): req. SNTPS_ID recv'd.\n");
   strcpy((char*) pBuffer, "pps"); //pps = siehe oben!, STRING !!!!

  break;
  }

  case SNTPS_RESOLUTION: /*write a 32-bit value containing the clock
resolution in nanoseconds into pBuffer*/
  {
   printf("sntpsTimeHook(): req. SNTPS_RESOLUTION recv'd.\n");
   if(clock_getres(CLOCK_REALTIME,&systemZeit) == ERROR)
   {
    printf("sntpsTimeHook(): >> clock_getres() failed! <<\n");
    return(ERROR);
   }

   value = systemZeit.tv_nsec;
   *((ULONG*)pBuffer) = (ULONG)value;

  break;
  }

  case SNTPS_TIME: /*write the current 64-bit NTP timestamp in host byte
order into pBuffer*/
  {
   printf("sntpsTimeHook(): req. SNTPS_TIME recv'd.\n");


   /*get time from the servers system-time (in ANSI format)*/
   /*which is here defined as THE reference source)*/
   if (clock_gettime(CLOCK_REALTIME, &systemZeit) == ERROR)
   {
    printf("sntpsTimeHook(): >> clock_gettime() failed! <<\n");
    return(ERROR);
   }

   pTime = (ULONG*)pBuffer; //get pointer

   /*write number of seconds since 1900 to timestamp */
   *pTime++ = systemZeit.tv_sec + SNTP_UNIX_OFFSET;
   /*Convert nanoseconds to fractional part and copy to timestamp*/
   *pTime = sntpsNsecToFraction(systemZeit.tv_nsec);


   // ctime() converts time in seconds into a string
   printf("sntpsTimeHook(): req. SNTPS_TIME done at %s.\n",
ctime(&systemZeit.tv_sec));
   printf("sntpsTimeHook() value tv_sec  =
%012d\n",*(&systemZeit.tv_sec));
   printf("sntpsTimeHook() value tv_sec  =
0x%012x\n",*(&systemZeit.tv_sec));

  break;
  }

  default:  /*service NOT available*/
  {
   return(ERROR);
  }
 }
 return(OK); /*service available*/
}


/********************************* END OF MODULE
************************************/
=======================================
DESCRIPTION


For my small application it was ok only to add the function prototype to
userAppInit.c.
(The routine sntpsInit( ) is called automatically during system startup
when the SNTP server library is included in the VxWorks image an the
aviability of the hook-routine is assumed!)


=======================================
/* usrAppInit.c - stub application initialization routine */

/* Copyright 1984-1998 Wind River Systems, Inc. */

/*
modification history
- --------------------
*/
#include 
#include 
#include 
#include "intLib.h"
#include "../pc486/pc.h"

/*
DESCRIPTION
Initialize user application code.
*/
STATUS sntpsTimeHook(int request, void *pBuffer);

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

*
* usrAppInit - initialize the users application
*/
......



=======================================
DESCRIPTION

A file with this hook routine was added to the server's bootimage.
After this, the server should be able to answer SNTP requests (not
broadcasting!).

To implement a "little" requesting client is possible as follows:
(stand-alone task cycling periodically)
(Take in care that the values of broken-down time are Unix-like (i.e.
Year 2000 corresponds to "100")
and asctime() assumes this and adds an offset to display it. )

To connect to >About Time< start it at your PC and enable SNTP.
Use your PC's IP adress as "pServerAddr".

========================================

#include 
#include 
#include 
#include 
#include 
#include 
#include 
#include 
#include 


void SystimeSNTPC()
{
 int status1 = ERROR;  // sntpcTimeGet()
 int status2 = 0;   // gmtime_r()
 int status3 = ERROR;  // checkTimeRange()

 struct timespec   CurrTimeSecond; // format ist NICHT analog zu >struct
tm64
sekunden)
                         &CurrTimeSecond);

  //>>>>>>>> See >RFC 1769 (SNTP) March 1995<  <<<<<<<<<<<<<<<<
  //NTP data are specified as integer or fixed-point
    //quantities, with bits numbered in big-endian fashion from 0
starting
    //at the left, or high-order, position. Unless specified otherwise,
all
    //quantities are unsigned and may occupy the full field width with
an
    //implied 0 preceding bit 0.
  //  0                   1                   2                   3
  //  0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
  // +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
  // |                           Seconds                             |
  // +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
  // |                  Seconds Fraction (0-padded)                  |
  // +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
  //
  //NTP timestamps are represented as a 64-bit unsigned
    //fixed-point number, in seconds relative to 0h on 1 January 1900.
The
    //integer part is in the first 32 bits and the fraction part in the
    //last 32 bits. This format allows convenient multiple-
  //precision arithmetic and conversion to UDP/TIME representation
  //(seconds), but does complicate the conversion to ICMP Timestamp
  //message representation (milliseconds).
  //>>>>>>>> See >RFC 1769 (SNTP) March 1995<  <<<<<<<<<<<<<<<<

  /* moegiche fehlernummern in sntpcTimeGet()
  #define S_sntpcLib_INVALID_PARAMETER         (M_sntpcLib | 1)
//online-help, BSP
  #define S_sntpcLib_INVALID_ADDRESS           (M_sntpcLib | 2)
//online-help, BSP
  #define S_sntpcLib_TIMEOUT                   (M_sntpcLib | 3) //BSP
  #define S_sntpcLib_VERSION_UNSUPPORTED       (M_sntpcLib | 4) //BSP,
rec'vd NTP info
  #define S_sntpcLib_SERVER_UNSYNC             (M_sntpcLib | 5) //BSP,
rec'vd NTP info
  */


// ctime(const time_t * timer) converts >time in seconds< into a string
// and returns a pointer to the string
printf("\n  *\n****\n  *\nSystimeSNTPC: timestamp %s arrived.\n",
ctime(&CurrTimeSecond.tv_sec));


  // konvertiere nach "broken Down", = menschenlesbar;
  // NUTZT nur SEKUNDENinformation (UDP/TIME representation)
  status2 = gmtime_r(&CurrTimeSecond.tv_sec, pZeit);
  if (status2 != OK) printf("status2 gmtime_r() ERROR\n");

  // pruefe wertebereiche,  -> interne plausibilitaet!
  status3 = checkTimeRange(pZeit);
  if (status3 == ERROR) printf("status3 checkTimeRange() ERROR\n");


  if ( ((status2==OK)&&(status1!=ERROR)&&(status3!=ERROR)) )
  {

   /*Do something usefull with the received timestamp*/

     printf("SystimeSNTPC Output: calendar %s\n", asctime((struct
tm*)pZeit));
  }
  else
     printf("SystimeSNTPC Output:  ERROR\n");
  }

  taskDelay( FetchRepeatTime_in_ticks );    //(übliches delay : >64
Sekunden)
}
/************************************************************************

**
**  name ...... :    checkTimeRange()
**
**  description :    prueft Wertebereiche der broken-down Zeit (struct
tm)
**
**
**  returns ... :    ERROR, OK
**
**
**
**  global var. :  NONE
*/

/* macro to check that  is between  and  inclusive range */
#define CHECK_RANGE(v,l,h) (((v)>=(l))&&((v)<=(h)))

static int checkTimeRange(const struct tm* pZeit)
{

    struct tm t1 = * pZeit ;
    FAST struct tm *tm = &t1 ;  /* make a local copy of the argument */

    /* Check value ranges */

    if( !CHECK_RANGE( tm->tm_sec,  0, 59)) return ERROR ;
    if( !CHECK_RANGE( tm->tm_min,  0, 59)) return ERROR ;
    if( !CHECK_RANGE( tm->tm_hour, 0, 23)) return ERROR ;
    if( !CHECK_RANGE( tm->tm_mday, 1, 31)) return ERROR ;
    if( !CHECK_RANGE( tm->tm_mon , 0, 11)) return ERROR ;
    if( !CHECK_RANGE( tm->tm_wday, 0, 6 )) return ERROR ;

    /* correction - for y2k */
    if( tm->tm_year > 99 )
    tm->tm_year -= 100 ;

    if( !CHECK_RANGE( tm->tm_year, 0, 99)) return ERROR ;

return(OK);
}


===========================================

/*EOF*/


- --------------224EBB7A49B7049626EF9925
Content-Type: text/html; charset=us-ascii
Content-Transfer-Encoding: 7bit



 

Jürgen Braun schrieb:

I want to use SNTP under VxWorks. Does somebody
has an example of a

server application and a client abblication for me?

Juergen




Hello Jürgen!

Just a few months ago I implemented a simple sntp client-server
connection between x86/Pentium based vx-targets using Tornado II.

I picked out some Information and code, the client is an uncompiled
sample showing it in general.

At least this is "easy" except a support-routine for the SNTP-server
which must be done by the user.

To Check independently whether it works or not, I used a "CareWare"-tool
that does work with the VxWorks SNTP components ( don't know why (the author
of the "CareWare"-tool has only made a fractional implementation of sntp)
but it runs!)

The Tool is: >About Time 4.8< (Win) to find at "www.arachnoid.com".

This tool is able to be a sever as well as a client.

Have a good luck, hope this helps you.

Joachim

--

=====================================================

           
Gebrüder Stoye GmbH

department: traffic control systems

      Name: Joachim Pluschke

       Adr: Longericher Strasse 177

      city: Cologne, Germany

       Zip: 50739

     phone: ++49/221/2616-534

    mailto: j.pluschke@stoye.de

       www: www.stoye.de

=====================================================

==========================

==========================

DESCRIPTION

Let's start at the beginning.....

===================

At first you have to include and configure the VxWorks - support
of SNTP in the target's bootimage.

With GUI configuration activate and configure(below there's something
about configuring)

/network components/networking protocols/networking applications/SNTP
<client OR server> depending for which of the two services the target
will used.

In the Vx-Networking manual there's described how to use SNTP (see
also "RFC 1769"), but there was a lack of clear information how to do the
Clock-Hook routine needed by the SNTP server.

After several unsuccessful trys to correctly feed in a date into
the server I found a remark from windriver in "comp.os.vxworks" what was
really helpfull:

=======================================

>> Forum: comp.os.vxworks

 >> Thread: VxWorks SNTP server

 >> Message 2 of 3

 Subject:

       Re: VxWorks SNTP server

 Date:

       08/17/1999

 Author:

       Stephen Macmanus <stephenm@wrs.com>

 

                                      
<< previous in thread  ·  next in thread >>

 nogap@my-deja.com wrote:

 >

 > Hi,

 >

 > I am looking for some information or  example code
for the hook routine

 > that shall be suplied to the SNTP server of VxWorks.

 > This hook routine is in the following format:

 >  sntpsClockHook (int request, void *pBuffer);

 >

 This sample hook routine reports the time information based
on

 the internal system clock:

 #include <vxWorks.h>

 #include <stdio.h>

 #include <string.h>

 #include "timers.h"

 #include <sntpsLib.h>

 STATUS sampleClockHook

     (

     int request,

     void *buffer

     )

     {

     struct timespec timeval;

     long value;

     ULONG *pTime;

     if (request == SNTPS_ID)

         {

         strcpy ((char
*)buffer, "pps");

         return (OK);

         }

     if (request == SNTPS_RESOLUTION)

         {

         if (clock_getres
(CLOCK_REALTIME, &timeval) == -1)

            
return (ERROR);

         value = timeval.tv_nsec;

         *((ULONG *)buffer)
= (ULONG)value;

         }

     if (request == SNTPS_TIME)

         {

         if (clock_gettime
(CLOCK_REALTIME, &timeval) == -1)

            
return (ERROR);

         pTime = (ULONG
*)buffer;

         /* Copy number
of seconds since 1900 to timestamp. */

         *pTime++ = timeval.tv_sec
+ SNTP_UNIX_OFFSET;

         /* Convert nanoseconds
to fractional part and copy to timestamp. */

         *pTime = sntpsNsecToFraction
(timeval.tv_nsec);

         }

     return (OK);

     }

 --

 ------------------

 Stephen Macmanus                        
#include <std_disclaimer.h>

 Software Engineer

 Wind River Systems

 

 

 

=======================================

DESCRIPTION

My version is slightly different but has the same core-functionality

=======================================

 

 

 

 

 

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

**

**  includes

*/

#include <sys/types.h>

#include <stdio.h>

#include <stdlib.h>

#include <string.h>

#include <vxWorks.h>

#include <taskLib.h>

#include <time.h>

#include <sntpsLib.h>

 

 

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

**

**  typedefs

*/

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

**

**  local memory

*/

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

**

**  function prototypes

*/

STATUS sntpsTimeHook(int request, void *pBuffer);

 

 

 

 

 

 

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

**

**  name ...... :    sntpsClockHook()

**

**  description :  This is the >hook routine< that
is called to access

**       the reference clock used
by the SNTP Server.

**       The hook routine copys one
of three settings

**       used by the server to construct
outgoing NTP messages

**       into pBuffer according to
the value of the request parameter.

**

**       request parameters:

**        SNTPS_ID to get the
32-bit reference identifier in the format specified in RFC 1769  or

**        SNTPS_RESOLUTION to
retrieve a 32-bit value containing the clock resolution in nanoseconds.

**         That value will
be used to determine the 8-bit signed integer indicating the

**         clock precision
(according to the format specified in RFC 1769) or

**        SNTPS_TIME to retrieve
the current 64-bit NTP timestamp from pBuffer in host byte order.

**

**

** >> THIS routine can be assigned directly by altering the value
of SNTPS_TIME_HOOK

** >> or can be installed by a call to the sntpsClockSet( ) routine.

** >> Until this routine is specified, the SNTP server will not
provide timestamp information!

**

**

**  returns ... :  OK if request setting available (or
ERROR otherwise).

**

**

**  global var. :

*/

 

/*

**

**

Zeitformate:

C:\Tornado\target\h\types\vxTypesBase.h(84): #define _TYPE_time_t          
typedef unsigned long time_t

struct timespec

    {

    time_t tv_sec;  // seconds

    long tv_nsec;   // nanoseconds (0
- - 1,000,000,000)

    };

**

mktime( ) - convert broken-down time into calendar time (ANSI)

SYNOPSIS

     time_t mktime

         (

         struct tm * timeptr
// pointer to broken-down structure

         )

RETURNS

     The calendar time in seconds, or ERROR
(-1) if calendar time cannot be calculated.

**

**

 

 

 

 

SNTPS_ID:

=========

      1   pps      
precision calibrated source, such as ATOM (atomic

                   
clock), PPS (precision pulse-per-second source),

                   
etc.

      1   service  
generic time service other than NTP, such as ACTS

                   
(Automated Computer Time Service), TIME (UDP/Time

                   
Protocol), TSP (Unix Time Service Protocol), DTSS

                   
(Digital Time Synchronization Service), etc.

      1   radio    
Generic radio service, with callsigns such as CHU,

                   
DCF77, MSF, TDF, WWV, WWVB, WWVH, etc.

      1   nav      
radionavigation system, such as OMEG (OMEGA), LORC

                   
(LORAN-C), etc.

      1   satellite generic
satellite service, such as GOES

                   
(Geostationary Orbit Environment Satellite, GPS

                   
(Global Positioning Service), etc.

      2   address  
secondary reference (four-octet Internet address of

                   
the NTP server)

 

 

 

 

SNTPS_RESOLUTION:

=================

       0 1 2 3 4 5 6 7 8 9 0 1 2 3
4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1

      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

      |LI | VN  |Mode |   
Stratum    |     Poll     
|   Precision   |

      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

 

 LI:   -

 VN:   3  includes all NTP versions

 MODE:   4/5 [server/broadcast]

 STRATUM:  0/1/2-15 [unspec./primary ref./sec.ref]

 POLL INTERVALL: 6..10 [64s...1024s]

 PRECISION: This is an eight-bit signed integer indicating
the

   precision of the local clock, in seconds to the nearest
power of two.

   The values that normally appear in this field range
from -6 for

   mains-frequency clocks to -20 for microsecond clocks
found in some

   workstations.

 

 

 

ZUSAETZLICH muss SNTPS_MODE gesetzt werden: SNTP_PASSIVE (0x2),
the server waits for requests from clients

               
oder: SNTP_ACTIVE (0x1), the server transmits NTP timestamp information
at fixed intervals.

    When executing in active mode, the SNTP server
uses the SNTPS_DSTADDR and SNTPS_INTERVAL

    definitions to determine the target IP address
and broadcast interval.

->

#define SNTPS_MODE SNTP_PASSIVE

und ein initial-aufruf von

sntpsClockSet((FUNCPTR)sntpsClockHook);

>>>>>>>(oder das o.g. 'zusaetzliche' bei der BSP-Modulselektion
& Konfiguration ueber das GUI)<<<<<<<<

*/

 

 

 

//#define SNTP_UNIX_OFFSET /*already defined, but where???*/

 

STATUS sntpsTimeHook(int request, void *pBuffer)

{

 

 struct timespec systemZeit;

 long    value;     //hilfsvariable

 ULONG *   pTime;     //hilfspointer

 

 

 

 switch(request)

 {

  case SNTPS_ID: /* put the 32-bit reference identifier in
the format specified in RFC 1769 into pBuffer*/

  {

   printf("sntpsTimeHook(): req. SNTPS_ID recv'd.\n");

   strcpy((char*) pBuffer, "pps"); //pps = siehe oben!,
STRING !!!!

  break;

  }

  case SNTPS_RESOLUTION: /*write a 32-bit value containing
the clock resolution in nanoseconds into pBuffer*/

  {

   printf("sntpsTimeHook(): req. SNTPS_RESOLUTION recv'd.\n");

   if(clock_getres(CLOCK_REALTIME,&systemZeit) ==
ERROR)

   {

    printf("sntpsTimeHook(): >> clock_getres() failed!
<<\n");

    return(ERROR);

   }

   value = systemZeit.tv_nsec;

   *((ULONG*)pBuffer) = (ULONG)value;

  break;

  }

  case SNTPS_TIME: /*write the current 64-bit NTP timestamp
in host byte order into pBuffer*/

  {

   printf("sntpsTimeHook(): req. SNTPS_TIME recv'd.\n");

 

   /*get time from the servers system-time (in ANSI format)*/

   /*which is here defined as THE reference source)*/

   if (clock_gettime(CLOCK_REALTIME, &systemZeit)
== ERROR)

   {

    printf("sntpsTimeHook(): >> clock_gettime()
failed! <<\n");

    return(ERROR);

   }

   pTime = (ULONG*)pBuffer; //get pointer

   /*write number of seconds since 1900 to timestamp */

   *pTime++ = systemZeit.tv_sec + SNTP_UNIX_OFFSET;

   /*Convert nanoseconds to fractional part and copy
to timestamp*/

   *pTime = sntpsNsecToFraction(systemZeit.tv_nsec);

 

   // ctime() converts time in seconds into a string

   printf("sntpsTimeHook(): req. SNTPS_TIME done at %s.\n",
ctime(&systemZeit.tv_sec));

   printf("sntpsTimeHook() value tv_sec  =  
%012d\n",*(&systemZeit.tv_sec));

   printf("sntpsTimeHook() value tv_sec  = 0x%012x\n",*(&systemZeit.tv_sec));

  break;

  }

  default:  /*service NOT available*/

  {

   return(ERROR);

  }

 }

 return(OK); /*service available*/

}

 

/********************************* END OF MODULE ************************************/

=======================================

DESCRIPTION

 

For my small application it was ok only to add the function prototype
to userAppInit.c.

(The routine sntpsInit( ) is called automatically during system
startup when the SNTP server library is included in the VxWorks image an
the aviability of the hook-routine is assumed!)

 

=======================================

/* usrAppInit.c - stub application initialization routine */

/* Copyright 1984-1998 Wind River Systems, Inc. */

/*

modification history

--------------------

*/

#include <sys/types.h>

#include <vxWorks.h>

#include <time.h>

#include "intLib.h"

#include "../pc486/pc.h"

/*

DESCRIPTION

Initialize user application code.

*/

STATUS sntpsTimeHook(int request, void *pBuffer);

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

*

* usrAppInit - initialize the users application

*/

......

 

 

=======================================

DESCRIPTION

A file with this hook routine was added to the server's bootimage.

After this, the server should be able to answer SNTP requests (not
broadcasting!).

To implement a "little" requesting client is possible as follows:

(stand-alone task cycling periodically)

(Take in care that the values of broken-down time are Unix-like
(i.e. Year 2000 corresponds to "100")

and asctime() assumes this and adds an offset to display it. )

To connect to >About Time< start it at your PC and enable SNTP.

Use your PC's IP adress as "pServerAddr".

========================================

#include <stdio.h>

#include <stdlib.h>

#include <string.h>

#include <vxWorks.h>

#include <time.h>

#include <ioLib.h>

#include <taskLib.h>

#include <sntpcLib.h>

#include <hostLib.h>

 

void SystimeSNTPC()

{

 int status1 = ERROR;  // sntpcTimeGet()

 int status2 = 0;   // gmtime_r()

 int status3 = ERROR;  // checkTimeRange()

 struct timespec   CurrTimeSecond; // format ist
NICHT analog zu >struct tm<!

 struct tm*    pZeit;

 struct tm    Zeit;

 //Arbeitszeiger

 pZeit = (struct tm*)&Zeit;

 if (pZeit == NULL) return;

 CurrTimeSecond.tv_sec = 0;

 CurrTimeSecond.tv_nsec = 0;

 while(1)

 {

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

 /* retrieve the current time from a remote network source
(SNTP-Server) */

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

  //The time is converted to the elapsed time since January
1, 1970, 00:00, GMT,

  //which is the base value used by UNIX systems and stored
in the location indicated by pCurrTimeSecond.

  status1 = sntpcTimeGet(pServerAddr,        
//char* like "192.43.244.18"

                        
TimeoutTime_in_ticks,  //(übliches timeout: >64 sekunden)

                        
&CurrTimeSecond);

  //>>>>>>>> See >RFC 1769 (SNTP) March 1995<  <<<<<<<<<<<<<<<<

  //NTP data are specified as integer or fixed-point

    //quantities, with bits numbered in big-endian
fashion from 0 starting

    //at the left, or high-order, position. Unless
specified otherwise, all

    //quantities are unsigned and may occupy the
full field width with an

    //implied 0 preceding bit 0.

  //  0                  
1                  
2                  
3

  //  0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3
4 5 6 7 8 9 0 1

  // +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

  // |                          
Seconds                            
|

  // +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

  // |                 
Seconds Fraction (0-padded)                 
|

  // +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

  //

  //NTP timestamps are represented as a 64-bit unsigned

    //fixed-point number, in seconds relative to
0h on 1 January 1900. The

    //integer part is in the first 32 bits and the
fraction part in the

    //last 32 bits. This format allows convenient
multiple-

  //precision arithmetic and conversion to UDP/TIME representation

  //(seconds), but does complicate the conversion to ICMP
Timestamp

  //message representation (milliseconds).

  //>>>>>>>> See >RFC 1769 (SNTP) March 1995<  <<<<<<<<<<<<<<<<

  /* moegiche fehlernummern in sntpcTimeGet()

  #define S_sntpcLib_INVALID_PARAMETER        
(M_sntpcLib | 1) //online-help, BSP

  #define S_sntpcLib_INVALID_ADDRESS          
(M_sntpcLib | 2) //online-help, BSP

  #define S_sntpcLib_TIMEOUT                  
(M_sntpcLib | 3) //BSP

  #define S_sntpcLib_VERSION_UNSUPPORTED      
(M_sntpcLib | 4) //BSP, rec'vd NTP info

  #define S_sntpcLib_SERVER_UNSYNC            
(M_sntpcLib | 5) //BSP, rec'vd NTP info

  */

 

// ctime(const time_t * timer) converts >time in seconds< into
a string

// and returns a pointer to the string

printf("\n  *\n****\n  *\nSystimeSNTPC: timestamp %s
arrived.\n", ctime(&CurrTimeSecond.tv_sec));

 

  // konvertiere nach "broken Down", = menschenlesbar;

  // NUTZT nur SEKUNDENinformation (UDP/TIME representation)

  status2 = gmtime_r(&CurrTimeSecond.tv_sec, pZeit);

  if (status2 != OK) printf("status2 gmtime_r() ERROR\n");

  // pruefe wertebereiche,  -> interne plausibilitaet!

  status3 = checkTimeRange(pZeit);

  if (status3 == ERROR) printf("status3 checkTimeRange() ERROR\n");

 

  if ( ((status2==OK)&&(status1!=ERROR)&&(status3!=ERROR))
)

  {

   /*Do something usefull with the received timestamp*/

     printf("SystimeSNTPC Output: calendar %s\n",
asctime((struct tm*)pZeit));

  }

  else

     printf("SystimeSNTPC Output:  ERROR\n");

  }

  taskDelay( FetchRepeatTime_in_ticks );   
//(übliches delay : >64 Sekunden)

}

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

**

**  name ...... :    checkTimeRange()

**

**  description :    prueft Wertebereiche der
broken-down Zeit (struct tm)

**

**

**  returns ... :    ERROR, OK

**

**

**

**  global var. :  NONE

*/

/* macro to check that <v> is between <l> and <h> inclusive
range */

#define CHECK_RANGE(v,l,h) (((v)>=(l))&&((v)<=(h)))

static int checkTimeRange(const struct tm* pZeit)

{

    struct tm t1 = * pZeit ;

    FAST struct tm *tm = &t1 ;  /* make
a local copy of the argument */

    /* Check value ranges */

    if( !CHECK_RANGE( tm->tm_sec,  0, 59)) return
ERROR ;

    if( !CHECK_RANGE( tm->tm_min,  0, 59))
return ERROR ;

    if( !CHECK_RANGE( tm->tm_hour, 0, 23)) return
ERROR ;

    if( !CHECK_RANGE( tm->tm_mday, 1, 31)) return
ERROR ;

    if( !CHECK_RANGE( tm->tm_mon , 0, 11)) return
ERROR ;

    if( !CHECK_RANGE( tm->tm_wday, 0, 6 )) return
ERROR ;

    /* correction - for y2k */

    if( tm->tm_year > 99 )

    tm->tm_year -= 100 ;

    if( !CHECK_RANGE( tm->tm_year, 0, 99)) return
ERROR ;

return(OK);

}

 

===========================================

/*EOF*/

 

- --------------224EBB7A49B7049626EF9925--


---------------------------

End of New-News digest
**********************


From vxwexplo-errs@csg.lbl.gov  Mon Aug 28 11:05:45 2000
From: "Martin Mayer" 
Date: Mon Aug 28 11:05:47 PDT 2000
Subject: Networking question (MUX layer)

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_01C0111A.3A272640
Content-Type: text/plain;
	charset="iso-8859-1"

Hi,

I am testing the creation of a new "test" protocol and I need to attach to
the vxWorks MUX Layer.

For some reason, my muxBind call is failing. I suspect that the device name
I am using is incorrect. How do I get a list of devices that the MUX layer
knows about?

Thanks,
Martin Mayer
Nortel Networks 
Data Comms Software Designer
(613) 763-9547


------_=_NextPart_001_01C0111A.3A272640
Content-Type: text/html;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable






Networking question (MUX layer)




Hi,




I am testing the creation of a new =
"test" protocol and I need to attach to the vxWorks MUX =
Layer.




For some reason, my muxBind call is =
failing. I suspect that the device name I am using is incorrect. How do =
I get a list of devices that the MUX layer knows about?



Thanks,

Martin Mayer

Nortel Networks 

Data Comms Software Designer

(613) 763-9547





------_=_NextPart_001_01C0111A.3A272640--


From vxwexplo-errs@csg.lbl.gov  Mon Aug 28 13:54:21 2000
From: Jim Way 
Date: Mon Aug 28 13:54:23 PDT 2000
Subject: To C or C++, that is the question.

Henry,

I recently released my first VxWorks-based project. It was also my group's
first experience with PowerPC and with C++. Because of that, we may have
overspec'd the board a little. But I have not seen performance problems like
I had been told to expect. My code is typically using 2-3% of the available
CPU resources. I've found that there are some great design patterns that
simplified my design and eliminated lots of code that would have been
written had we continued with C.

I've seen and heard conflicting advice. Some people say C generates smaller
code. Others say that was based on older C++ compilers and that the new ones
are much improved. Other people say that C++ has you writing fewer lines of
code, so there is savings there. Yes, it can be a religious war .

Do you think you are likely to be in trouble? I'd ask myself questions like,
Can I increase my RAM and/or my processor performance? Which language do I
know better how to optimize? Which language has the better optimizing
compiler (so I don't have to do it myself)? What tools do I have or can
acquire to give me visibility into my internal details?

For us, time to market was THE issue. I think the C++ choice helped us get
done quicker and with a better product.

$0.02

Jim

-----------------------------------------
Jim Way, Software Engineer
Datum Austin (Austron Inc.)
voice: 512.721.4170
fax  : 512.990.9712
email: jwayATdatumDOTcom (no spam please)
-----------------------------------------


From vxwexplo-errs@csg.lbl.gov  Mon Aug 28 15:11:55 2000
From: "Martin Mayer" 
Date: Mon Aug 28 15:12:01 PDT 2000
Subject: What am I not getting? (MUX layer question)

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_01C0113C.B04F9790
Content-Type: text/plain;
	charset="iso-8859-1"

Hi,

(Please reply to all)

I am trying to create my own little protocol and am having trouble with the
vxWorks MUX layer. 
 
I am binding in very simple test routines. The bind works but I can't seem
to get any response from the MUX layer. 

Here is what I do...

I download the following code into the target (see lower)...
I type    bind "motfcc",0     and it works...
I try to ping the target's IP and expect to see that I "Received data"...
(Is this a good assumption?)
P.S. I also tried an ipDetach and got rid of the vxWorks IP stack... still
doesn't work.

Why is the MUX layer not calling my snarfRcvRtn function?

Thanks,
Martin Mayer
Comms software designer

int snarfBind(char* dev, int num)
{
   /* Bind this protocol to a device via the mux */
   if (!(IN_COOKIE[num] = (END_OBJ*) muxBind(dev, num, snarfRcvRtn,
            snarfShutdownRtn, snarfRestartRtn, snarfErrorRtn,
            MUX_PROTO_SNARF, "myIP", NULL))) {
      return ERROR;
   }
   return OK;
}

BOOL snarfRcvRtn(void* cookie, long type, M_BLK_ID netbuff,
                 LL_HDR_INFO* linkHdr, void* spare)
{
   printf("Received data\n");
   return 0;
}


STATUS snarfShutdownRtn(void* cookie, void* spare)
{
   printf("Shutting down snarfProto\n");
   return muxUnbind(cookie, MUX_PROTO_SNARF, snarfRcvRtn);
}


void snarfErrorRtn(END_OBJ* end, END_ERR* error, void* spare)
{
   printf("snarfError called\n");
}

STATUS snarfRestartRtn(void* cookie, void* spare)
{
   printf("snarfTxRestartRtn called\n");
   return OK;
}

STATUS unbind(void)
{
   return muxUnbind(IN_COOKIE[0], MUX_PROTO_SNARF, snarfRcvRtn);   
}

int bind(char* name, int i)
{
   /* Do network init for all ports on device - except Net Mgment Port */
   char addrBuffer[20];

   if (!muxDevExists(name,i))
      printf("%s%d DOES NOT exist   ",name,i);
   else
      printf("%s%d DOES exist   ",name,i);

   if( snarfBind(name, i) != OK)
      printf("Bind FAILED");
   else
      printf("Bind OK");

   printf("\n");
   return 1;
}


------_=_NextPart_001_01C0113C.B04F9790
Content-Type: text/html;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable






What am I not getting? (MUX layer question)




Hi,




(Please reply to all)




I am trying to create my own little =
protocol and am having trouble with the vxWorks MUX layer. 

 

I am binding in very simple test =
routines. The bind works but I can't seem to get any response from the =
MUX layer. 




Here is what I do...




I download the following code into the =
target (see lower)...

I type    bind =
"motfcc",0     and it works...

I try to ping the target's IP and =
expect to see that I "Received data"... (Is this a good =
assumption?)

P.S. I also tried an ipDetach and got =
rid of the vxWorks IP stack... still doesn't work.




Why is the MUX layer not calling my =
snarfRcvRtn function?




Thanks,

Martin Mayer

Comms software designer




int snarfBind(char* dev, int =
num)

{

   /* Bind this protocol to =
a device via the mux */

   if (!(IN_COOKIE[num] =3D =
(END_OBJ*) muxBind(dev, num, snarfRcvRtn,

         &nb=
sp;  snarfShutdownRtn, snarfRestartRtn, snarfErrorRtn,

         &nb=
sp;  MUX_PROTO_SNARF, "myIP", NULL))) {

      return =
ERROR;

   }

   return OK;

}




BOOL snarfRcvRtn(void* cookie, long =
type, M_BLK_ID netbuff,

         &nb=
sp;       LL_HDR_INFO* linkHdr, void* =
spare)

{

   printf("Received =
data\n");

   return 0;

}






STATUS snarfShutdownRtn(void* cookie, =
void* spare)

{

   printf("Shutting =
down snarfProto\n");

   return muxUnbind(cookie, =
MUX_PROTO_SNARF, snarfRcvRtn);

}






void snarfErrorRtn(END_OBJ* end, =
END_ERR* error, void* spare)

{

   printf("snarfError =
called\n");

}




STATUS snarfRestartRtn(void* cookie, =
void* spare)

{

   =
printf("snarfTxRestartRtn called\n");

   return OK;

}




STATUS unbind(void)

{

   return =
muxUnbind(IN_COOKIE[0], MUX_PROTO_SNARF, snarfRcvRtn);   =


}




int bind(char* name, int i)

{

   /* Do network init for =
all ports on device - except Net Mgment Port */

   char =
addrBuffer[20];




   if =
(!muxDevExists(name,i))

      =
printf("%s%d DOES NOT exist   ",name,i);

   else

      =
printf("%s%d DOES exist   ",name,i);




   if( snarfBind(name, i) =
!=3D OK)

      =
printf("Bind FAILED");

   else

      =
printf("Bind OK");




   =
printf("\n");

   return 1;

}





------_=_NextPart_001_01C0113C.B04F9790--


From vxwexplo-errs@csg.lbl.gov  Mon Aug 28 22:07:03 2000
From: "Bryan de Vera" 
Date: Mon Aug 28 22:07:05 PDT 2000
Subject: first timer

good day!

i'm a design engineer for a software/firmware company.

i'm been working with VxWorks for about a year now.

i was just wondering how i can join in on the daily digests?

thanks!

the bry

$B!{!|!{!|!{!|!{!|!{!|!{!|!{!|!{!|!{!|!{!|!{!|(B
THE BRY
  Bryan de Vera
  Design Engineer

  MANILA OFFICE:
  Millennium Software, Inc. (MSI)
  25th Flr, Robinson's-Equitable Bank Tower,
    4 ADB Ave. corner Poveda Road,
    Ortigas Center, Pasig City,
    1605, Philippines
  Tel   : +63 2 638-3070 to 72
  Fax   : +63 2 638-3079
  mailto:jdevera@tspi.com.ph

  HEAD OFFICE:
  Tsukiden Software Japan, Inc. (TSJI)
  68 Okidaka, Taiheiji
    Fukushima City, Fukushima Pref.
    960 Japan
  Tel: +81 245 39-8400
  Fax: +81 245 39-8399

  MANILA RESIDENCE:
  3rd Flr, 9624-E Pililla St.,
    Rizal Village, Makati City,
    1208, Philippines
  Tel   : +63 2 895-7204
  mailto:ayt_yall@bigfoot.com
$B!{!|!{!|!{!|!{!|!{!|!{!|!{!|!{!|!{!|!{!|!{!|(B



From vxwexplo-errs@csg.lbl.gov  Mon Aug 28 23:02:44 2000
From: "Rajendra Prasad" 
Date: Mon Aug 28 23:02:46 PDT 2000
Subject: compression & decompression in vxworks

Hi,

I require a good compression/decompression software that runs both on 
vxworks 5.4 and sunsolaris 2.6. Has anybody tried  Zlib ? Does it work on 
both vxworks 5.4 and sunsolaris 2.6 ?

thanks and regards,
rpraghuram.
_________________________________________________________________________
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com.

Share information about yourself, create your own public profile at 
http://profiles.msn.com.



From vxwexplo-errs@csg.lbl.gov  Tue Aug 29 01:49:57 2000
From: "Tim Allen" 
Date: Tue Aug 29 01:49:59 PDT 2000
Subject: Re:More PCI Bus & VXWorks

Your header dump shows:

>0x10 BaseAddresses  0x01048000 0x01040000 0x00000000
>                    0x00000000 0x00000000 0x00000000

and you said:

>Is there something I should be doing with translating the PCI address to
>Host address.  What I recall is that the PCI/IO Space in PCI address starts
>at 0x1000000 and the host starts at 0x81000000.

But your base addresses show the device is in PCI memory space, not I/O
space which is indicated by having bit 0 set. In this case you should try
accessing 0xc1048000.

Regards,
Tim.
--------------------
Tim Allen Ltd
Real Time Embedded Software Consultancy
vxworks



From vxwexplo-errs@csg.lbl.gov  Tue Aug 29 02:09:25 2000
From: David Laight 
Date: Tue Aug 29 02:09:27 PDT 2000
Subject: re: ppc: rework of intLock/intUnlock for nesting?


> Has anyone re-coded a version if the intLock() and intUnlock() routines
> that allow nested calls to these routines without prematurely unlocking
> interrupts?  Seems to me that this should be possible by using the
> "key" parameter... 

> I was surprised that vxw doesnt already do this.  Is there a problem
> with doing this that I am missing?

Don't know about ppc, but the arm ones are:

_intLock:
00126c24: e10f1000      mrs     r1,cpsr
00126c28: e2010080      and     r0,r1,#0x80
00126c2c: e3811080      orr     r1,r1,#0x80
00126c30: e129f001      msr     cpsr_cf,r1
00126c34: e1a0f00e      mov     pc,lr
_intUnlock:
00126c38: e10f1000      mrs     r1,cpsr
00126c3c: e3c11080      bic     r1,r1,#0x80
00126c40: e2000080      and     r0,r0,#0x80
00126c44: e1811000      orr     r1,r1,r0
00126c48: e129f001      msr     cpsr_cf,r1
00126c4c: e1a0f00e      mov     pc,lr

Which return/load bit 0x80 from/to the cpsr as expected.

These routines are also (usefully) callable from an ISR.  I have a horrid
feeling that this isn't necessarily the case.  As usual the vxworks 
documentation isn't clear.

	David

----------------------------------------------------------------
David Laight                          email: dsl@tadpole.co.uk
Tadpole Technology plc                phone: +44 1223 428 232
Cambridge, UK                         fax:   +44 1223 428 201



From vxwexplo-errs@csg.lbl.gov  Tue Aug 29 04:03:12 2000
From: Vxworks Exploder 
Date: Tue Aug 29 04:03:15 PDT 2000
Subject: comp.os.vxworks newsdigest

Comp.Os.Vxworks Daily Digest    Tue Aug 29 04:03:08 PDT 2000

        Subject: Re: writing a device driver
        Subject: Booting vxWorks with MCPN765
        Subject: Re: memory window for accessing pcmcia!
        Subject: long file names using DOSFS
        Subject: How to view / edit files on Target RAM disk
        Subject: Re: Canceling a select()
        Subject: Re: PCI driver guidelines?
        Subject: Re: More PCI Bus & VXWorks
        Subject: MPC860 Boot Code Memory Access
        Subject: netBufLib and END
        Subject: Re: writing a device driver
        Subject: vxworks and IP packet forwarding
        Subject: Re: netBufLib and END
        Subject: Re: To C or C++?
        Subject: Networking question (MUX layer)
        Subject: PCI mapping on MVME2700
        Subject: Advice on "global variable" technique
        Subject: PCI Configuration help for a novice
        Subject: Compilation problem
        Subject: Re: Booting vxWorks with MCPN765
        Subject: Re: Advice on "global variable" technique
        Subject: To C or C++, that is the question.
        Subject: Setting it up...
        Subject: Re: Advice on "global variable" technique
        Subject: Re: Download location
        Subject: What am I not getting? (MUX layer question)
        Subject: How to get NT server to recognize SNMP device on PPP serial port.
        Subject: Re: Booting vxWorks with MCPN765
        Subject: Re: Advice on "global variable" technique
        Subject: Re: Loading a device driver
        Subject: Control-C Handler w/ multiple ttys?
        Subject: Re: How do I use op code (Assembly Language) Instructions in the PowerPC with the VxWorks development tools?
        Subject: Re: To C or C++, that is the question.
        Subject: Re: ppc: rework of intLock/intUnlock for nesting?
        Subject: How can i Develop the driver for PCI Display
        Subject: first timer
        Subject: CompactFlash disk on PC.
        Subject: compression & decompression in vxworks
        Subject: Who uses TAO with VxWorks
        Subject: watch variables
        Subject: Re: To C or C++?
        Subject: using software floating point
        Subject: re:More PCI Bus & VXWorks
        Subject: re: ppc: rework of intLock/intUnlock for nesting?
        Subject: Re: Project/linker question = Super Projects
        Subject: Re: Advice on "global variable" technique
        Subject: pciDeviceShow

-------------------------------------------------------

Newsgroups: comp.os.vxworks
Subject: Re: writing a device driver
Date: Mon, 28 Aug 2000 12:16:52 -0500
From: "Keith McHugh" 
Organization: Esat Net Customer
Message-ID: <39aa4975.0@seat.s3two.ie>
References: <8oa9dm$ck8$1@nnrp1.deja.com>
Reply-To: "Keith McHugh" 

Try with Motorola web site???

 wrote in message news:8oa9dm$ck8$1@nnrp1.deja.com...
> Where can i find a web site who explain how to write device drivers in
> vxworks and with source code who explain how to do that ?
>
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.



---------------------------

Newsgroups: comp.os.vxworks
Subject: Booting vxWorks with MCPN765
Date: Mon, 28 Aug 2000 13:47:59 +0200
From: Marc Waber 
Organization: Swisscom IP+ (post doesn't reflect views of Swisscom)
Message-ID: <39AA516F.3A85F3D4@stest.ch>

Dear all!

I am trying to boot vxWorks on a MCPN765 over Ethernet with a PowerPC
7400 Processor in a slave configuration with a CPV5350.

When booting, I always get the following output:

[VxWorks Boot]: @

boot device          : dc
unit number          : 0 
processor number     : 0 
host name            : gemini
file name            : vxWorks.custom
inet on ethernet (e) : 141.169.76.51:ffffff00
host inet (h)        : 141.169.76.95
gateway inet (g)     : 141.169.76.1
user (u)             : 8610ng
flags (f)            : 0x0 
target name (tn)     : powerpc
other (o)            : dc0

Attached TCP/IP interface to dc0.
interrupt: dc0 - no carrier
Attaching network interface lo0... done.
Loading... interrupt: dc0 - no carrier
interrupt: dc0 - no carrier
interrupt: dc0 - no carrier
interrupt: dc0 - no carrier
interrupt: dc0 - no carrier
interrupt: dc0 - no carrier
interrupt: dc0 - no carrier
interrupt: dc0 - no carrier
interrupt: dc0 - no carrier
interrupt: dc0 - no carrier
interrupt: dc0 - no carrier
interrupt: dc0 - no carrier
interrupt: dc0 - no carrier
interrupt: dc0 - no carrier
interrupt: dc0 - no carrier
interrupt: dc0 - no carrier
interrupt: dc0 - no carrier
interrupt: dc0 - no carrier

Error loading file: errno = 0x3c.

Any ideas?

Marc


- --
Marc Waber				E-mail:    marc.waber@stest.ch
STS Switching Test Solutions AG		Phone:	   +41-76-355-6713
Foerrlibuckstr. 62 / Postfach 74	Fax:  	   +41-1-355-6605
CH-8037 Zurich				Homepage:  www.stest.com

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: memory window for accessing pcmcia!
Date: 28 Aug 2000 11:56:29 GMT
From: Geir Frode Raanes 
Organization: UNINETT news service
Message-ID: <8odk1d$a0p$1@snipp.uninett.no>
References: <8odbtq$btm@netnews.hinet.net>

kvd2  wrote:
: Dear All,

: I wanna to open a memory window for accessing pcmcia register and common
: memory. But the error message, as follow, is returned. What I have to

: I've add this to sysLib.c, but it won't work!
: {
:     (void *) PCMCIA_BASE_ADRS,
:     (void *) PCMCIA_BASE_ADRS,
:    PCMCIA_SIZE,    /* 4 k - DICE parameter */
:    VM_STATE_MASK_VALID | VM_STATE_MASK_WRITABLE | VM_STATE_MASK_CACHEABLE,
:     VM_STATE_VALID      | VM_STATE_WRITABLE      | VM_STATE_CACHEABLE_NOT
:     },

First of all I suggest you look up the "vmStateSet()" routine in your
VxWorks reference manual. I would also enable the 'guarded' VM feature
of the PowerPC through "VM_STATE_MASK_GUARDED" and then guard this region
through "VM_STATE_GUARDED." 

Next I did not really understand what you are attempting to do, but I 
suspect you are attempting to access the PCMCIA configuration registers.
These registers/memory resides in it's own _separate_ address space.
To access them you have to pull down the configuration space chip enable
line. This implies that you have to do some XPC860 PCMCIA PCMCIA control
register fiddling before you do the mem in/out action. I have not yet 
looked into exactly what this fiddeling constitutes of.

Conclusion - do orientate yourself about the inner working of the 
PCMCIA bus HW before you try to force VxWorks do magic for you.

- --
  ******************************************************
  Never ever underestimate the power of human stupidity.
  -Robert Anson Heinlein

		GeirFRS@invalid.and.so.forth
  ******************************************************












- -- 
	********************************************
	  E'kke greit å være geit i et sauesamfunn

		  GeirFRS@omegav.ntnu.no
	********************************************

---------------------------

Newsgroups: comp.os.vxworks
Subject: long file names using DOSFS
Date: Mon, 28 Aug 2000 15:30:41 +0200
From: "Ezra Koper" 
Organization: NetVision Israel	
Message-ID: <8odm14$lqr$1@news.netvision.net.il>

Hello All.

I am using VxWorks DOSFS ( not dosfs2) which best to my knowladge support
only FAT16 and is limited to 8 chars + 3 ext char file names.

I have seen that i can configure long file name option:

status = dosFsMkfsOptionsSet (DOS_OPT_LONGNAMES);
if (status != OK)
        printf("\n\r Error - unable to define long file name system!!");


When i run it it doesn't return with error, but i can't create file names
longer then 8 chars.
Do i need DOSFS2 in order to support it? or i am doing somthing wrong?





---------------------------

Newsgroups: comp.os.vxworks
Subject: How to view / edit files on Target RAM disk
Date: Mon, 28 Aug 2000 18:45:29 +0530
From: slkumar 
Organization: Motorola
Message-ID: <39AA65F1.B82C8E76@miel.mot.com>

Is there a tool to view and edit files on Target RAM Disk drive ?


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Canceling a select()
Date: Mon, 28 Aug 2000 15:44:41 +0200
From: Leif Thuresson 
Organization: TSL
Message-ID: <39AA6CC9.7E7301C9@tsl.uu.se>
References: <8o5nvs$b5n$1@overload.lbl.gov>
Reply-To: leif.thuresson@tsl.uu.se



Johan van der Werf wrote:

>   Hi all,
>
>  Question: what would be the best way to make a task return from a select()
>            function call before the timeout is expired, or any data is sent?
>
>  At first sight ioctl(fd, FIOUNSELECT, xxx) seemed a likely candidate,
>  unfortunately it can only be used from inside a driver, because the
>  third argument must be a pointer to a SEL_WAKEUP_NODE. Please tell me if
>  there is a standard way of doing this?
>
>  We're using VxWorks 5.3.1 on M68K platform.
>
>                         Regards,
>
> --
> |Johan van der Werf   |   tel: ++31 (0)40 2684280 |$B%O%s%9(B  |
> |ASM Lithography      |   fax: ++31 (0)40 2682224 |$B%t%!%s(B  |
> |Dept: Basic Software | email: ujw@asml.nl        |$B%G%k(B    |
> |Room: 7A3.012        |        vdwerf@xs4all.nl   |$B%&%(%k%U(B|

Another way is to use signals. Wake-up the task with a signal and then use
setjmp/longjmp to jump out of the signal handler

regards,
- --leif

- ---------------------------------------------------------
Leif Thuresson               Tel:    +46-18-4713860
Uppsala University           Fax:    +46-18-4713833
The Svedberg Laboratory      E-mail: leif.thuresson@tsl.uu.se
P.O.Box 533
S-751 21 Uppsala
Sweden



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: PCI driver guidelines?
Date: Mon, 28 Aug 2000 07:20:46 -0700
From: "Charles Krinke" 
Message-ID: <8uuq5.1$Yo.160@news.connectnet.com>
References: <39A479D3.CA30BBE5@root.com>

Look in the ..target\src\drv\pci directory and you will find a myriad of pci
routines. pciFindDevice(vendorID, deviceId, 0,&BusNo,&DeviceNo,&FuncNo)
should get you started. If you have the device on an attached pci bus,
pciFindDevice, given the vendor & device ID's, will tell you its bus number,
device number and function number. Given, this info, you can call
pciDevConfig and write control register space with memory BAR, interrupt
stuff, whatever (examine the pciLocalBus spec a bit, it will help). You can
look at all devices on the bus with pciHeaderShow(0,0..X,0). Once, you have
the device mapped into memory space, with a valid interrupt, you can
continue with traditional End device initialization.

"Bruce Chrustie"  wrote in message
news:39A479D3.CA30BBE5@root.com...
> Hi gang,
>
> Planning on writing a driver for the IDT ATM NIC for the x86. Anyone
> have some insight as how to do this in VxWorks? Never wrote a PCI driver
> before :)
>
> Ist step would be to see if the PCI card is there
> then look at the config reg
> and so on....
>
> Searched the net for guidelines and came up empty.
>
> Bruce,



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: More PCI Bus & VXWorks
Date: Mon, 28 Aug 2000 07:28:01 -0700
From: "Charles Krinke" 
Message-ID: 
References: 

Dear Shawn:
    I don't know about your target, but my ARM, does an *automagic*
translation of CPU addresses to pci addresses inside my intel 21285 bridge
chip. If you have any existing pci devices, look around sysPhysMemDesc. If,
not, look at the docs for your bridge chip if you have one. Hope this helps
a bit.

Charles

"Shawn Girard"  wrote in message
news:Hzup5.27054$eR5.885737@news1.rdc1.on.wave.home.com...
> Please bear with me if this question has already been asked, but I am new
in
> working with the PCI bus and I am having problems in accessing the device
> through its assigned BARs.
>
> I am currently writing a device driver for a PCI bridge chip, to be exact
it
> is the TI PCI2040 PCI-DSP bridge chip through VxWorks on a Radstone PPC4.
>
> I ask some folks earlier in how are the Bars in the PCI Configuration
Header
> are assigned and they informed me it is done in the Auto Config  routine
> that is called in sysHwInit routine of the Kernel.  Working off this, I
can
> successfully access the PCI Config Header through the Radstone's API,
which
> shown below.
>
> -> sysPciDevShow 0,6,0
> 0x00 DeviceID       0xac60104c Texas Instruments
> 0x04 Command        0x0142
> 0x06 Status         0x0210
> 0x08 RevisionID     0x00
> 0x09 ProgIF         0x00 SubClass 0x80 BaseClass 0x06
>                     Bridge, Other
> 0x0c CacheLineSize  0x08
> 0x0d LatencyTimer   0x00
> 0x0e HeaderType     0x00
> 0x0f BIST           0x00
> 0x10 BaseAddresses  0x01048000 0x01040000 0x00000000
>                     0x00000000 0x00000000 0x00000000
> 0x2c Sub-System ID  0x00000000
> 0x30 BaseAddressROM 0x00000000
> 0x3c InterruptLine  0x09
> 0x3d InterruptPin   0x01
> 0x3e Min_Gnt        0x00
> 0x3f Max_Lat        0x00
> 0x40 0x00000000 0x00001f00 0x00000000 0x0000c10f
> 0x50 0xfe115c01 0x00000000 0x00000000 0x00000006
> 0x60 0x00000000 0x00000000 0x00000000 0x00000000
> 0x70 0x00000000 0x00000000 0x00000000 0x00000000
> 0x80 0x00000000 0x00000000 0x00000000 0x00000000
> 0x90 0x00000000 0x00000000 0x00000000 0x00000000
> 0xa0 0x00000000 0x00000000 0x00000000 0x00000000
> 0xb0 0x00000000 0x00000000 0x00000000 0x00000000
> 0xc0 0x00000000 0x00000000 0x00000000 0x00000000
> 0xd0 0x00000000 0x00000000 0x00000000 0x00000000
> 0xe0 0x00000000 0x00000000 0x00000000 0x00000000
> 0xf0 0x00000000 0x00000000 0x00000000 0x00000000
>
> Being aware that I have to translate the Bar address to the host address,
I
> apply the needed translatation for allocated PCI/IO Space of the host.
That
> gives me the following host address for Bar_0 0x81048000.
>
> Being aware that this is a window to the devices register, I try doing the
> address dump to see what is there.  Please note, at this point I never
> loaded in my device driver in yet.  This the result of my dump:
>
> 81048000:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> 81048010:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> 81048020:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> 81048030:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> 81048040:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> 81048050:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> 81048060:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> 81048070:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> 81048080:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> 81048090:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> 810480a0:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> 810480b0:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> 810480c0:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> 810480d0:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> 810480e0:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> 810480f0:  ffff ffff ffff ffff ffff ffff ffff ffff   *................*
> value = 0 = 0x0
>
> Now for the million dollar question, why am I seeing nothing but ones?  Am
I
> accessing the memory location correctly?  The reason why I know that I am
> accessing it wrong is that I am not seeing the default values for some of
> the registers.
>
> Is there something I should be doing with translating the PCI address to
> Host address.  What I recall is that the PCI/IO Space in PCI address
starts
> at 0x1000000 and the host starts at 0x81000000.
>
> Thanks in advance
> Shawn Girard
>
>



---------------------------

Newsgroups: comp.os.vxworks
Subject: MPC860 Boot Code Memory Access
Date: Mon, 28 Aug 2000 09:51:36 -0400
From: jim 
Organization: University of Waterloo
Message-ID: 
References: <8no2d3$p7k@netnews.hinet.net>

Hey there,

I was hoping to find some help with a couple questions about MPC860
asmembler ... I am a bit of newbie so any help would be great ... thanks
in advance.

1.) First ... Is there any good URL's for MPC860 assembler? ... I have a
reference but it is not as descriptive as I would like.

2.) I am writing a little memory test in some boot-code ... the memory in
question is associated with BRx ... I assume that to access memory I need
only to get the first 17 bits of this register (the base address) and use
this plus an offset to access memory. Can I access BRx the same as any
other register? To obtain the first 17 bits should this be ok:

/* move BRx contents to Ry */
addi Ry BRx '0' 

/* create a mask to wipe out last 15 bits*/
lis	rz, HIADJ(0xffff 8000)
addi	rz, rz, LO(0xffff 8000)

/*set y to contain only the base address part of Brx and set rest all to
zero*/
andi Ry Ry Rz  

3.) Any recomendations for memory testing in boot code ?

Thanks

Jim



---------------------------

Newsgroups: comp.os.vxworks
Subject: netBufLib and END
Date: Mon, 28 Aug 2000 15:56:53 +0100
From: John N Gray 
Organization: Cadence Design Systems
Message-ID: <39AA7DB5.FCA5FED3@Cadence.COM>

Hi,

The MUX interface uses these elaborate mBlk/clBlk and cluster structures
to transfer
data up and down the networks layers. At driver level mBlks are passed
to and from
the driver.

...but the LANCE driver is very ambiguous when it comes to handling
these mBlks.
Should ethernet drivers support chained mBlks in both the sending and
receiving
operations, or is the MUX simply going to only give single cluster
(non-chained)
data to the driver?

Thanks in advance,

John


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: writing a device driver
Date: Mon, 28 Aug 2000 07:30:23 -0700
From: "Charles Krinke" 
Message-ID: 
References: <8oa9dm$ck8$1@nnrp1.deja.com>

VxWorks, contains many device driver template with a name of the form
templateEnd.c, templateScsi.c. These might be helpful to look at. Also, look
at the BSP Developer's Kit book. It is quite helpful on the details of
writing a device driver.
Charles
 wrote in message news:8oa9dm$ck8$1@nnrp1.deja.com...
> Where can i find a web site who explain how to write device drivers in
> vxworks and with source code who explain how to do that ?
>
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.



---------------------------

Newsgroups: comp.os.vxworks
Subject: vxworks and IP packet forwarding
Date: Mon, 28 Aug 2000 19:00:39 +0200
From: Sebastien Fauris 
Organization: Club-Internet (France)
Message-ID: <39AA9AB7.5B84AFD8@club-internet.fr>

Hi Folks,

I have a target with 2 ethernet interfaces.
Is there a way to disable the IP packet forwarding from an ethernet
interface
to the other when the packet is not for my vxWorks target ?
(i.e. to disable the routing feature so my target can not be used as a
gateway
from a network to another).

Thanks !

Sebastien.


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: netBufLib and END
Date: Mon, 28 Aug 2000 16:59:01 GMT
From: gold@sdsu.edu
Organization: Deja.com - Before you buy.
Message-ID: <8oe5o9$ik5$1@nnrp1.deja.com>
References: <39AA7DB5.FCA5FED3@Cadence.COM>

In article <39AA7DB5.FCA5FED3@Cadence.COM>,
  John N Gray  wrote:

> ...but the LANCE driver is very ambiguous when it comes to handling
> these mBlks.
> Should ethernet drivers support chained mBlks in both the sending and
> receiving
> operations, or is the MUX simply going to only give single cluster
> (non-chained)
> data to the driver?

Absolutely not.  The MUX can always give you an mblk chain to send.
Note that the chain the MUX sends down is not a chain of (say
Ethernet) "packets", but a chain of mblks/clusters designed to go out
as 1 ethernet packet.  That chain is linked via the 'mNext' field.  I
believe the mNextPkt field can be used to chain 'link layer' frames.  I
don't think ip_output() uses it though.  Why would you receive a chain
from the MUX ?  Let's say you want to send a 1500 byte packet, but your
system pool only has 512 byte clusters left, no bigger.  The MUX will
give you a chain of the mblk/clblk/clusters, each cluster being 512
bytes.

On the receive side, it's up to you whether you want to handle a chain,
depending on the driver architecture.  Most drivers have their own
cluster pool they use to allocate receive buffers for.  In that case
you can always just use 1500 byte clusters for ethernet (following
alignment restrictions for your hardware of course).  But if you're
sharing with the system pool and you don't have any MTU-size clusters
available, you need to handle a chain.

  hth,
   Dan Gold  gold@sdsu.edu


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: To C or C++?
Date: Mon, 28 Aug 2000 17:52:12 GMT
From: djhystad@msn.com
Organization: Deja.com - Before you buy.
Message-ID: <8oe8s1$mgj$1@nnrp1.deja.com>
References: <39A543EB.C9E19479@mail.com> <39A5ACF3.DCCEC871@fnc.fujitsu.com>

In article <39A5ACF3.DCCEC871@fnc.fujitsu.com>,
  Steven Deans  wrote:
Maybe they "get" OO, but are appalled at the horrible way that it is
supported by C++.  It was such an unfortunate day when the programming
community picked C++ over Objective C.  Oh well, it has always been
difficult for even a superior solution to compete against a large
installed base.

> Hi Henry,
>
> In general C++ will only be slightly larger and slightly slower than
C. In
> fact, well written C++ with a good optimizing C++ compiler can
sometimes
> execute faster than C because of inlined code and the compiler can
> optimize better because of the stronger type checking. The size and
speed
> differences are small generally < 3%. If this difference is a big
concern,
> maybe you should consider writing in assembly. :) I have 5+ years
writing
> Real Time embedded OO C++ and  3+ years writing real time embedded C.
I
> recommend C++. It has been my experience that most people that prefer
C
> either have not used C++ or don't "get" OO.
>
> The bottom line is the quality of your code and your compiler are
going to
> be  a bigger factor than C vs. C++ differences.
>
> Steven
>
> Henry wrote:
>
> > Hi, All,
> >
> > This is not my intention to start a language war. I want to konw
should
> > we C or C++, and if we use C++ under VxWork, do we pay penalty on
code
> > sizes, performace etc.?
> >
> > I personally prefer C++ for OO and its strong type checking, but I
want
> > your opinions.
> >
> > Thanks in advance.
> >
> > Henry
>
>


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Networking question (MUX layer)
Date: Mon, 28 Aug 2000 14:03:10 -0400
From: "Martin Mayer" 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <8oea6s$sjh$1@overload.lbl.gov>

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_01C0111A.3A272640
Content-Type: text/plain;
	charset="iso-8859-1"

Hi,

I am testing the creation of a new "test" protocol and I need to attach to
the vxWorks MUX Layer.

For some reason, my muxBind call is failing. I suspect that the device name
I am using is incorrect. How do I get a list of devices that the MUX layer
knows about?

Thanks,
Martin Mayer
Nortel Networks 
Data Comms Software Designer
(613) 763-9547


- ------_=_NextPart_001_01C0111A.3A272640
Content-Type: text/html;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable






Networking question (MUX layer)




Hi,




I am testing the creation of a new =
"test" protocol and I need to attach to the vxWorks MUX =
Layer.




For some reason, my muxBind call is =
failing. I suspect that the device name I am using is incorrect. How do =
I get a list of devices that the MUX layer knows about?



Thanks,

Martin Mayer

Nortel Networks 

Data Comms Software Designer

(613) 763-9547





- ------_=_NextPart_001_01C0111A.3A272640--

---------------------------

Newsgroups: comp.os.vxworks
Subject: PCI mapping on MVME2700
Date: Mon, 28 Aug 2000 20:35:55 +0200
From: =?iso-8859-1?Q?Torbj=F6rn?= Frostemark 
Organization: FOA
Message-ID: <39AAB10B.CC4BCBF1@lin.foa.se>

This is a multi-part message in MIME format.
- --------------C25FC130912148BDCA253E7B
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

I (well, my VSB PMC module) want to map a 128Mb window in PCI memory
space on a MVME2700 using extended VME. The module is only looking at
bit 27-31 in the VSB BAR, so I guess I am stuck trying to get a free
128Mb window somewere between 0x0800 0000 and 0xF800 0000. The question
is where and how...

Default this window starts at 0xFD000000 and is 16Mb, thus out of range
and too small. Looking at the Programmer's Ref. Guide a CHRP compliant
memory map can map PCI memory at 0x8000 0000. If I change
CPU_PCI_MEM_ADRS to 0x8000 0000 and size to 128Mb, should I get a 1:1
mapping CPU addresses to PCI (and VSB) addresses? Right now it does not
seem so... :(

/Tobbe



- --------------C25FC130912148BDCA253E7B
Content-Type: text/x-vcard; charset=us-ascii;
 name="torfr.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Card for Torbjörn Frostemark
Content-Disposition: attachment;
 filename="torfr.vcf"

begin:vcard 
n:Frostemark;Torbjörn
tel;cell:+46-(0)70-249 76 27
tel;fax:+46-(0)13-37 80 09
tel;work:+46-(0)13-37 80 14
x-mozilla-html:TRUE
org:FOA;Electronic Warfare Systems
adr:;;Box 1165;SE-581 11  Linköping;;;SWEDEN
version:2.1
email;internet:torfr@lin.foa.se
title:MSc
fn:Torbjörn Frostemark
end:vcard

- --------------C25FC130912148BDCA253E7B--


---------------------------

Newsgroups: comp.os.vxworks
Subject: Advice on "global variable" technique
Date: Mon, 28 Aug 2000 13:47:52 -0500
From: "Tim Berger" 
Organization: Motorola CIG
Message-ID: <8oec3r$fql5@nntp.cig.mot.com>

Hello,

I have a task that spawns two other tasks (sort of a parent-children type
relationship).  This group of tasks uses a bunch of common variables that
are accessed often (but never change after they get initialized -- e.g., a
socket integer; all three of these tasks use a single socket).  Naturally,
I'd want to put these variables into a global structure, but globals cause
problems, of course, if I run multiple instances of this group of tasks.
So, rather than passing everything around, or passing a pointer to a
structure containing all these variables around, I defined a single
structure containing all the "global" variables.  The parent task then
allocates memory for this structure, and then sets a global pointer to this
structure.  Finally, the task makes the global pointer a task variable.  The
other two tasks get the task id of the parent task and retrieve the global
pointer to the structure through a taskVarGet.  This works, and I can have
many instances of the group of tasks running without any race conditions.

My question: is there a better way to do this?  The method above seems a bit
hokey to me now (I made it without really thinking it through).  Can't I
simply allocate memory for a global structure in the parent task and pass
the pointer to the structure to the children (without using task variables)?
Or, should I use some other mechanism?

I'm looking for the most reliable method -- this thing has to be
bulletproof.  What worries me with the simple pointer-passing above is the
case when a "parent" task dies unexpectedly and the "child" tries to access
the pointer.  At least with task variables, the taskVarGet will simply fail
in this case.

Any advice would be greatly appreciated.  Thanks!

Tim




---------------------------

Newsgroups: comp.os.vxworks
Subject: PCI Configuration help for a novice
Date: Mon, 28 Aug 2000 15:38:34 -0400
From: kostizt 
Organization: Raytheon Company
Message-ID: <39AABFBA.C71B9657@res.raytheon.com>

Folks,

I must admit that I am new to the PCI bus and I am having a lot of
trouble getting a vendor supplied driver / serial device into my
system.  The symptom is that (a) the BARs don't make a lot of sense to
me and (b) the device (serial), doesn't seem to be mapped (i/o simply
posting in memory with no interrupts).

I have a 2304 BSP.  The vendor's documentation says "ensure that the BSP
itself performs configuration of all PCI devices supported by the PCI
HAL.  This includes configuring the PCI base address registers (BARs)
and the interrupt line request register in the PCI configuration space
of each device."

I thought I would find the code that does this in the BSP and go from
there.  The problem is that I can't find where any of this happens!

Does anyone know how the PCI devices get their configuration information
processed by the BSP?

If I invoke an xpciShow, it looks as follows:

- -> xpciShow
Scanning function 0 of each PCI device on bus 0 of card 0
bus slot vendor  dev IRQ    Bar0     Bar1     Bar2     Bar3     Bar4
Bar5
 0   00   1057  4801  00  00000001 fc000000 00000000 00000000 00000000
00000000
 0   11   10ad  0565  00  00000000 00000000 00000000 00000000 00000000
00000000
 0   13   10e3  0000  00  fd050000 00000001 00000000 00000000 00000000
00000000
 0   14   1011  0009  00  00020001 00000000 00000000 00000000 00000000
00000000
 0   17   124b  0001  00  00000000 00000000 00000000 00000000 00000000
00000000
value = 1 = 0x1

The information for bus/slot/vendor for slot 17 is the interesting
piece.  The vendor info is the PCI device of interest.  Does a BAR0 of 0
make sense?

Any help would be greatly appreciated.

Thanks in advance!

Tom Kostizak


---------------------------

Newsgroups: comp.os.vxworks
Subject: Compilation problem
Date: Mon, 28 Aug 2000 15:48:23 -0400
From: Sairam Yadlapati 
Organization: Lucent Technologies
Message-ID: <39AAC207.8978031D@lucent.com>
Reply-To: sairam@lucent.com

Hi,
	I am tryiong to compile the est8260 board support package with vxwork image on
my workstation. And  I was able top compile the image successfully for solaris,
for MPC8260 when I am compiling I am getting a linker error saying
"a.out-sunos-big". I have no clue to solve this problem. Can some body help me
on this?  Your help is very much appreciated.


My environment variables are:
=============================


# Command line build environments
WIND_HOST_TYPE=sun4-solaris2
export WIND_HOST_TYPE
WIND_BASE=/home/sairam/VXWORKS/
export WIND_BASE
WIND_REGISTRY=ubb34
export WIND_REGISTRY
PATH=$WIND_BASE/host/$WIND_HOST_TYPE/bin:$PATH
export PATH
if [ "$USER" = "" ] ; then
    USER=$LOGNAME
fi
export USER
if [ "$LD_LIBRARY_PATH" = "" ] ; then
   
LD_LIBRARY_PATH=$WIND_BASE/host/$WIND_HOST_TYPE/lib:/usr/dt/lib:/usr/openwin/lib
else
   
LD_LIBRARY_PATH=$LD_LIBRARY_PATH:$WIND_BASE/host/$WIND_HOST_TYPE/lib:/usr/dt/lib:/usr/openwin/lib
fi
export LD_LIBRARY_PATH

======the error log is attached===

ccppc -B/home/sairam/VXWORKS//host/sun4-solaris2/lib/gcc-lib/  -M -MG
- -mstrict-align -ansi -nostdinc -O2 -fvolatile -fno-builtin -fno-for-scope -Wall
- -I/h   -I. -I/home/sairam/VXWORKS//target/config/all
- -I/home/sairam/VXWORKS//target/h -I/home/sairam/VXWORKS//target/src/config
- -I/home/sairam/VXWORKS//target/src/drv -DCPU=PPCEC603  -DEST8260  -gdwarf -O0
- -DROM_BASED -msoft-float *.c /home/sairam/VXWORKS//target/config/all/*.c \
| sed -e "s:/home/sairam/VXWORKS//target/config/all:\\$\(CONFIG_ALL\):g" \
	-e "s:/home/sairam/VXWORKS//target:\\$\(TGT_DIR\):g" \
	-e "s:/home/sairam/VXWORKS/:\\$\(WIND_BASE\):g" > depend.est8260
ccppc -B/home/sairam/VXWORKS//host/sun4-solaris2/lib/gcc-lib/ -E -P -xc -M -E
- -I/h   -I. -I/home/sairam/VXWORKS//target/config/all
- -I/home/sairam/VXWORKS//target/h -I/home/sairam/VXWORKS//target/src/config
- -I/home/sairam/VXWORKS//target/src/drv -DCPU=PPCEC603  -DEST8260  romInit.s \
| sed -e "s:/home/sairam/VXWORKS//target/config/all:\\$\(CONFIG_ALL\):g" \
	-e "s:/home/sairam/VXWORKS//target:\\$\(TGT_DIR\):g" \
	-e "s:/home/sairam/VXWORKS/:\\$\(WIND_BASE\):g" >> depend.est8260
ccppc -B/home/sairam/VXWORKS//host/sun4-solaris2/lib/gcc-lib/ -E -P -xc -M -E
- -I/h   -I. -I/home/sairam/VXWORKS//target/config/all
- -I/home/sairam/VXWORKS//target/h -I/home/sairam/VXWORKS//target/src/config
- -I/home/sairam/VXWORKS//target/src/drv -DCPU=PPCEC603  -DEST8260  sysALib.s \
| sed -e "s:/home/sairam/VXWORKS//target/config/all:\\$\(CONFIG_ALL\):g" \
	-e "s:/home/sairam/VXWORKS//target:\\$\(TGT_DIR\):g" \
	-e "s:/home/sairam/VXWORKS/:\\$\(WIND_BASE\):g" >> depend.est8260
ccppc -B/home/sairam/VXWORKS//host/sun4-solaris2/lib/gcc-lib/ -c -mstrict-align
- -ansi -nostdinc -O2 -fvolatile -fno-builtin -fno-for-scope -Wall -I/h   -I.
- -I/home/sairam/VXWORKS//target/config/all -I/home/sairam/VXWORKS//target/h
- -I/home/sairam/VXWORKS//target/src/config -I/home/sairam/VXWORKS//target/src/drv
- -DCPU=PPCEC603  -DEST8260  -gdwarf -O0 -DROM_BASED -msoft-float usrConfig.c -o
usrConfig.o
ccppc -B/home/sairam/VXWORKS//host/sun4-solaris2/lib/gcc-lib/ -c -mstrict-align
- -ansi -nostdinc -O2 -fvolatile -fno-builtin -fno-for-scope -Wall -I/h   -I.
- -I/home/sairam/VXWORKS//target/config/all -I/home/sairam/VXWORKS//target/h
- -I/home/sairam/VXWORKS//target/src/config -I/home/sairam/VXWORKS//target/src/drv
- -DCPU=PPCEC603  -DEST8260  -gdwarf -O0 -DROM_BASED -msoft-float
/home/sairam/VXWORKS//target/config/all/dataSegPad.c -o dataSegPad.o
ccppc -B/home/sairam/VXWORKS//host/sun4-solaris2/lib/gcc-lib/ -mstrict-align
- -ansi -nostdinc -O2 -fvolatile -fno-builtin -fno-for-scope -I/h   -I.
- -I/home/sairam/VXWORKS//target/config/all -I/home/sairam/VXWORKS//target/h
- -I/home/sairam/VXWORKS//target/src/config -I/home/sairam/VXWORKS//target/src/drv
- -DCPU=PPCEC603  -DEST8260  -gdwarf -O0 -DROM_BASED -P -x assembler-with-cpp -c
- -o sysALib.o sysALib.s 
ccppc -B/home/sairam/VXWORKS//host/sun4-solaris2/lib/gcc-lib/ -mstrict-align
- -ansi -nostdinc -O2 -fvolatile -fno-builtin -fno-for-scope -Wall -I/h   -I.
- -I/home/sairam/VXWORKS//target/config/all -I/home/sairam/VXWORKS//target/h
- -I/home/sairam/VXWORKS//target/src/config -I/home/sairam/VXWORKS//target/src/drv
- -DCPU=PPCEC603  -DEST8260  -gdwarf -O0 -DROM_BASED -msoft-float -c sysLib.c
ccppc -B/home/sairam/VXWORKS//host/sun4-solaris2/lib/gcc-lib/ -c -mstrict-align
- -ansi -nostdinc -O2 -fvolatile -fno-builtin -fno-for-scope -Wall -I/h   -I.
- -I/home/sairam/VXWORKS//target/config/all -I/home/sairam/VXWORKS//target/h
- -I/home/sairam/VXWORKS//target/src/config -I/home/sairam/VXWORKS//target/src/drv
- -DCPU=PPCEC603  -DEST8260  -gdwarf -O0 -DROM_BASED -msoft-float -o version.o
/home/sairam/VXWORKS//target/config/all/version.c
ccppc -B/home/sairam/VXWORKS//host/sun4-solaris2/lib/gcc-lib/ -nostdlib -r
- -Wl,-X \
    -o vxWorks.tmp sysALib.o sysLib.o  motFccEnd.obj miiLib.obj m8260CpmEnd.obj 
usrConfig.o version.o  /home/sairam/VXWORKS//target/lib/libPPCEC603gnuvx.a
/home/sairam/VXWORKS//host/sun4-solaris2/lib/gcc-lib/powerpc-wrs-vxworks/cygnus-2.7.2-960126/../../../../powerpc-wrs-vxworks/bin/ld:
invalid BFD target `a.out-sunos-big'
collect2: ld returned 1 exit status
make: *** [vxWorks] Error 1

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Booting vxWorks with MCPN765
Date: 28 Aug 2000 20:03:49 GMT
From: johan@borksoft.xs4all.nl
Organization: Borksoft
Message-ID: <8oegj5$5f0$1@news1.xs4all.nl>
References: <39AA516F.3A85F3D4@stest.ch>

Marc Waber  wrote:
: Dear all!

: I am trying to boot vxWorks on a MCPN765 over Ethernet with a PowerPC
: 7400 Processor in a slave configuration with a CPV5350.

: When booting, I always get the following output:

: interrupt: dc0 - no carrier

According to some other mails in the group there are several possible
solutions. There is an SPR on a related problem (I don't know if it is
this problem). The number of theis SPR is 28573.
Someone else suggested putting the port in half duplex mode.
See the newsgroup archives on deja.com and search for "dc0 - no carrier",
and you might find some more articles.

Groeten,
	Johan

- -- 
     o o o o o o o . . .   ___________________________________
    o      _____           ||    Johan Borkhuis              |
  .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
 >(________|__|_[_________]_|________________________________|
 _/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: Advice on "global variable" technique
Date: 28 Aug 2000 20:09:43 GMT
From: johan@borksoft.xs4all.nl
Organization: Borksoft
Message-ID: <8oegu7$5f0$2@news1.xs4all.nl>
References: <8oec3r$fql5@nntp.cig.mot.com>

Another possibility is to have a variable in the main function of the
parent task. You can pass a pointer to this variable to the child task.
You can have as many parent task as you need, because the variables are
created on the stack.

Groeten,
	Johan

- -- 
     o o o o o o o . . .   ___________________________________
    o      _____           ||    Johan Borkhuis              |
  .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
 >(________|__|_[_________]_|________________________________|
 _/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: To C or C++, that is the question.
Date: Mon, 28 Aug 2000 13:51:00 -0700
From: Jim Way 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <8oekoa$80n$1@overload.lbl.gov>

Henry,

I recently released my first VxWorks-based project. It was also my group's
first experience with PowerPC and with C++. Because of that, we may have
overspec'd the board a little. But I have not seen performance problems like
I had been told to expect. My code is typically using 2-3% of the available
CPU resources. I've found that there are some great design patterns that
simplified my design and eliminated lots of code that would have been
written had we continued with C.

I've seen and heard conflicting advice. Some people say C generates smaller
code. Others say that was based on older C++ compilers and that the new ones
are much improved. Other people say that C++ has you writing fewer lines of
code, so there is savings there. Yes, it can be a religious war .

Do you think you are likely to be in trouble? I'd ask myself questions like,
Can I increase my RAM and/or my processor performance? Which language do I
know better how to optimize? Which language has the better optimizing
compiler (so I don't have to do it myself)? What tools do I have or can
acquire to give me visibility into my internal details?

For us, time to market was THE issue. I think the C++ choice helped us get
done quicker and with a better product.

$0.02

Jim

- -----------------------------------------
Jim Way, Software Engineer
Datum Austin (Austron Inc.)
voice: 512.721.4170
fax  : 512.990.9712
email: jwayATdatumDOTcom (no spam please)
- -----------------------------------------

---------------------------

Newsgroups: comp.os.vxworks
Subject: Setting it up...
Date: Mon, 28 Aug 2000 16:30:43 -0500
From: "Michael S." 
Organization: MediaOne Express -=- Central Region
Message-ID: <7KAq5.185$d4.36055@elnws01>

I am still unable to figure out how to build vxworks, can someone give me
some step-by-step instructions? It would be nice to have.

Thank you,
Michael S. - scikid@hotmail.com



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Advice on "global variable" technique
Date: Mon, 28 Aug 2000 16:53:43 -0500
From: Greg Willden 
Organization: Posted via Supernews, http://www.supernews.com
Message-ID: <39AADF66.13A07406@swri.edu>
References: <8oec3r$fql5@nntp.cig.mot.com>

Tim Berger wrote:

> Hello,
>
> I have a task that spawns two other tasks (sort of a parent-children type
> relationship).  This group of tasks uses a bunch of common variables that
> are accessed often (but never change after they get initialized -- e.g., a
> socket integer; all three of these tasks use a single socket).

If the data don't change after they are initialized why can't you just pass
those values to the spawned task.  The taskSpawn function allows 10 arguments to
be passed as parameters to the entry point of the spawned task.  This would
remove a tremendous amount of complexity.

- --
Greg Willden      gwillden@swri.edu
Engineer
Southwest Research Institute
www.swri.edu



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Download location
Date: Mon, 28 Aug 2000 13:43:47 +0800
From: "tanghaoyu" 
Organization: The news server of Changzhou TeleCom
Message-ID: <8ocu24$1t5j$1@news.cz.js.cn>
References:  <8ntc50$qjd$2@news1.xs4all.nl>

Now you can free download it from http://www.wrs.com/prototyper/index.html.


 wrote in message
news:8ntc50$qjd$2@news1.xs4all.nl...
> Michael S.  wrote:
> : Anyone know where I can download a beta or trial of the tornado
developement
> : kit?
>
> There is no download, you can ask your local sales person for an
> evaluation license.
>
> Groeten,
> Johan
>
> --
>      o o o o o o o . . .   ___________________________________
>     o      _____           ||    Johan Borkhuis              |
>   .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
>  >(________|__|_[_________]_|________________________________|
>  _/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: What am I not getting? (MUX layer question)
Date: Mon, 28 Aug 2000 18:09:52 -0400
From: "Martin Mayer" 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <8oeo8r$9ue$1@overload.lbl.gov>

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_01C0113C.B04F9790
Content-Type: text/plain;
	charset="iso-8859-1"

Hi,

(Please reply to all)

I am trying to create my own little protocol and am having trouble with the
vxWorks MUX layer. 
 
I am binding in very simple test routines. The bind works but I can't seem
to get any response from the MUX layer. 

Here is what I do...

I download the following code into the target (see lower)...
I type    bind "motfcc",0     and it works...
I try to ping the target's IP and expect to see that I "Received data"...
(Is this a good assumption?)
P.S. I also tried an ipDetach and got rid of the vxWorks IP stack... still
doesn't work.

Why is the MUX layer not calling my snarfRcvRtn function?

Thanks,
Martin Mayer
Comms software designer

int snarfBind(char* dev, int num)
{
   /* Bind this protocol to a device via the mux */
   if (!(IN_COOKIE[num] = (END_OBJ*) muxBind(dev, num, snarfRcvRtn,
            snarfShutdownRtn, snarfRestartRtn, snarfErrorRtn,
            MUX_PROTO_SNARF, "myIP", NULL))) {
      return ERROR;
   }
   return OK;
}

BOOL snarfRcvRtn(void* cookie, long type, M_BLK_ID netbuff,
                 LL_HDR_INFO* linkHdr, void* spare)
{
   printf("Received data\n");
   return 0;
}


STATUS snarfShutdownRtn(void* cookie, void* spare)
{
   printf("Shutting down snarfProto\n");
   return muxUnbind(cookie, MUX_PROTO_SNARF, snarfRcvRtn);
}


void snarfErrorRtn(END_OBJ* end, END_ERR* error, void* spare)
{
   printf("snarfError called\n");
}

STATUS snarfRestartRtn(void* cookie, void* spare)
{
   printf("snarfTxRestartRtn called\n");
   return OK;
}

STATUS unbind(void)
{
   return muxUnbind(IN_COOKIE[0], MUX_PROTO_SNARF, snarfRcvRtn);   
}

int bind(char* name, int i)
{
   /* Do network init for all ports on device - except Net Mgment Port */
   char addrBuffer[20];

   if (!muxDevExists(name,i))
      printf("%s%d DOES NOT exist   ",name,i);
   else
      printf("%s%d DOES exist   ",name,i);

   if( snarfBind(name, i) != OK)
      printf("Bind FAILED");
   else
      printf("Bind OK");

   printf("\n");
   return 1;
}


- ------_=_NextPart_001_01C0113C.B04F9790
Content-Type: text/html;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable






What am I not getting? (MUX layer question)




Hi,




(Please reply to all)




I am trying to create my own little =
protocol and am having trouble with the vxWorks MUX layer. 

 

I am binding in very simple test =
routines. The bind works but I can't seem to get any response from the =
MUX layer. 




Here is what I do...




I download the following code into the =
target (see lower)...

I type    bind =
"motfcc",0     and it works...

I try to ping the target's IP and =
expect to see that I "Received data"... (Is this a good =
assumption?)

P.S. I also tried an ipDetach and got =
rid of the vxWorks IP stack... still doesn't work.




Why is the MUX layer not calling my =
snarfRcvRtn function?




Thanks,

Martin Mayer

Comms software designer




int snarfBind(char* dev, int =
num)

{

   /* Bind this protocol to =
a device via the mux */

   if (!(IN_COOKIE[num] =3D =
(END_OBJ*) muxBind(dev, num, snarfRcvRtn,

         &nb=
sp;  snarfShutdownRtn, snarfRestartRtn, snarfErrorRtn,

         &nb=
sp;  MUX_PROTO_SNARF, "myIP", NULL))) {

      return =
ERROR;

   }

   return OK;

}




BOOL snarfRcvRtn(void* cookie, long =
type, M_BLK_ID netbuff,

         &nb=
sp;       LL_HDR_INFO* linkHdr, void* =
spare)

{

   printf("Received =
data\n");

   return 0;

}






STATUS snarfShutdownRtn(void* cookie, =
void* spare)

{

   printf("Shutting =
down snarfProto\n");

   return muxUnbind(cookie, =
MUX_PROTO_SNARF, snarfRcvRtn);

}






void snarfErrorRtn(END_OBJ* end, =
END_ERR* error, void* spare)

{

   printf("snarfError =
called\n");

}




STATUS snarfRestartRtn(void* cookie, =
void* spare)

{

   =
printf("snarfTxRestartRtn called\n");

   return OK;

}




STATUS unbind(void)

{

   return =
muxUnbind(IN_COOKIE[0], MUX_PROTO_SNARF, snarfRcvRtn);   =


}




int bind(char* name, int i)

{

   /* Do network init for =
all ports on device - except Net Mgment Port */

   char =
addrBuffer[20];




   if =
(!muxDevExists(name,i))

      =
printf("%s%d DOES NOT exist   ",name,i);

   else

      =
printf("%s%d DOES exist   ",name,i);




   if( snarfBind(name, i) =
!=3D OK)

      =
printf("Bind FAILED");

   else

      =
printf("Bind OK");




   =
printf("\n");

   return 1;

}





- ------_=_NextPart_001_01C0113C.B04F9790--

---------------------------

Newsgroups: comp.os.vxworks
Subject: How to get NT server to recognize SNMP device on PPP serial port.
Date: Mon, 28 Aug 2000 22:17:52 GMT
From: randyj5573@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8oeods$a76$1@nnrp1.deja.com>

I am designing a device which communicates to an NT server via a PPP
serial link over a serial cable. I can establish the PPP link by having
Vxworks initiate the connection with "NULL" as the remote and local IP
addresses. The NT server then tells Vxworks what IP address to use for
the local and remote address.

The problem is that I have an application which is performing a
broadcast to see which SNMP devices respond. This broadcast never makes
it out to the serial port. I have a protocol analyzer on the serial
port and I never see the broadcast. When the application sends a
message specificaly to the IP address of the serial device or when I
PING the serial device, I then see the request and response on the
protocol analyzer.

How do I get the NT server to recognize that the serial device exists?

Thanks,
Randy Jones


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Booting vxWorks with MCPN765
Date: 28 Aug 2000 22:58:45 GMT
From: bowtie@netcom.com (Bill Dennen)
Organization: MindSpring Enterprises
Message-ID: <8oeqr5$n80$1@slb2.atl.mindspring.net>
References: <39AA516F.3A85F3D4@stest.ch> <8oegj5$5f0$1@news1.xs4all.nl>

johan@borksoft.xs4all.nl writes:

If you're willing to look outside of VxWorks, select the board to boot Bug
rather than VxWorks and try nping.  (Your network addresses, set by niot,
were probably set up earlier when you burned in the kernel).  Format is
nping 0 0 board_ip host_ip packet_count

It takes a while so set packet_count low and read the return carefully (it's
packets LOST not packets returned).  The error code returned can be quite
helpful -- see Bug manual appendices at http://www.mcg.mot.com/literature.

No carrier implies the wire's not seen ...

Regards
>@<


>Marc Waber  wrote:
>: Dear all!

>: I am trying to boot vxWorks on a MCPN765 over Ethernet with a PowerPC
>: 7400 Processor in a slave configuration with a CPV5350.

>: When booting, I always get the following output:

>: interrupt: dc0 - no carrier

>According to some other mails in the group there are several possible
>solutions. There is an SPR on a related problem (I don't know if it is
>this problem). The number of theis SPR is 28573.
>Someone else suggested putting the port in half duplex mode.
>See the newsgroup archives on deja.com and search for "dc0 - no carrier",
>and you might find some more articles.

>Groeten,
>	Johan

>-- 
>     o o o o o o o . . .   ___________________________________
>    o      _____           ||    Johan Borkhuis              |
>  .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
> >(________|__|_[_________]_|________________________________|
> _/oo OOOOO oo`  ooo   ooo  'o!o!o                      o!o!o`
>=== VxWorks FAQ: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html ===
- --
>@<					bowtie@netcom.com
Bill Dennen				william.dennen@motorola.com   
	Motorola Computer Group, TOPAZ

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Advice on "global variable" technique
Date: Mon, 28 Aug 2000 23:03:22 GMT
From: "telia" 
Organization: Enitel Internet Public Access
Message-ID: <_cCq5.6637$Dxe.200009216@news.telia.no>
References: <8oec3r$fql5@nntp.cig.mot.com>


Tim Berger  wrote in message
news:8oec3r$fql5@nntp.cig.mot.com...
> Hello,
>
> relationship).  This group of tasks uses a bunch of common variables that
> are accessed often (but never change after they get initialized -- e.g., a
> socket integer; all three of these tasks use a single socket).  Naturally,
> I'd want to put these variables into a global structure, but globals cause
> problems, of course, if I run multiple instances of this group of tasks.
>
>

This is not a response to your question, just a warning !
Be aware that sockets calls may screw up if you perform concurrent
read/write operations on a single socket.





---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Loading a device driver
Date: Mon, 28 Aug 2000 23:06:55 GMT
From: "telia" 
Organization: Enitel Internet Public Access
Message-ID: 
References: <8o86fn$qpk$1@nslave2.tin.it>


Merlino  wrote in message
news:8o86fn$qpk$1@nslave2.tin.it...
> Hi guys,
>
> i have an x86  board with a P2 BSP vxWorks 5.4;
> i would like to run the system this way:
> 1- loading vxworks image ( without any device driver )
> 2- scanning the PCI bus for known device
> 3- loading modules with the suitables device drivers
> 4- activating device drivers
>
> Is it possible ?
> How ?
>
> Thanks
>
>
From where would you load the driver if you have no device-drivers loaded
???




---------------------------

Newsgroups: comp.os.vxworks
Subject: Control-C Handler w/ multiple ttys?
Date: Mon, 28 Aug 2000 23:09:37 GMT
From: Tony Dal Santo 
Message-ID: 

Does anyone have any suggestions on how to handle a user pressing
control-C (abort character) when there are multiple ttys?  Looks like
there is a single global function that gets called, and it doesn't get
passed any parameters!  This function can be set via tyAbortFuncSet.  If
I go this route, is there an easy way to figure out on which tty the
control-C was received?

I also noticed that tyLib supports a protocol hook (FIOPROTOHOOK).
Would this be better to use, and just add my own handler?  Would my
routine be called in interrupt context?

I'm guess that someone out there has multiple ttys connected to a
vxWorks system, and has already solved this.

Thanks,
Tony

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: How do I use op code (Assembly Language) Instructions in the PowerPC with the VxWorks development tools?
Date: Mon, 28 Aug 2000 23:15:40 GMT
From: "telia" 
Organization: Enitel Internet Public Access
Message-ID: 
References: <39a5354c.0@seat.s3two.ie>


Keith McHugh  wrote in message
news:39a5354c.0@seat.s3two.ie...
>
>
>
Just another hint.

Be careful about letting the compiler allocate registers. Some instructions
interprete register 0 as constant 0. It appears that the GNU  allocator
ignore
the restrictions these instructions place on the allowed register set.





---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: To C or C++, that is the question.
Date: Mon, 28 Aug 2000 23:12:59 GMT
From: djhystad@msn.com
Organization: Deja.com - Before you buy.
Message-ID: <8oerlc$dno$1@nnrp1.deja.com>
References: <8oekoa$80n$1@overload.lbl.gov>

I guess maybe the best way to answer the question is to take a good
look at your problem space, evaulate how each solution fits, and choose
the best for you.  Unlike Jim, I usually have a processor utilization
of around 50 - 70%.  I also find myself having to optimize pieces of
code in assembler after inspecting it with profiler tools.  In these
cases, I don't care how good the compiler was, C++ is usually not an
option.

It's all relative.  I thought I had hard real-time constraints.  Then
at an embedded developers conference I met someone who developed
software/hardware for instrumenting explosions.  The event they were
designed to monitor had a lifespan roughly equivalent to one sample
period for my controllers.

In article <8oekoa$80n$1@overload.lbl.gov>,
  Jim Way  wrote:
> Henry,
>
> I recently released my first VxWorks-based project. It was also my
group's
> first experience with PowerPC and with C++. Because of that, we may
have
> overspec'd the board a little. But I have not seen performance
problems like
> I had been told to expect. My code is typically using 2-3% of the
available
> CPU resources. I've found that there are some great design patterns
that
> simplified my design and eliminated lots of code that would have been
> written had we continued with C.
>
> I've seen and heard conflicting advice. Some people say C generates
smaller
> code. Others say that was based on older C++ compilers and that the
new ones
> are much improved. Other people say that C++ has you writing fewer
lines of
> code, so there is savings there. Yes, it can be a religious war
.
>
> Do you think you are likely to be in trouble? I'd ask myself
questions like,
> Can I increase my RAM and/or my processor performance? Which language
do I
> know better how to optimize? Which language has the better optimizing
> compiler (so I don't have to do it myself)? What tools do I have or
can
> acquire to give me visibility into my internal details?
>
> For us, time to market was THE issue. I think the C++ choice helped
us get
> done quicker and with a better product.
>
> $0.02
>
> Jim
>
> -----------------------------------------
> Jim Way, Software Engineer
> Datum Austin (Austron Inc.)
> voice: 512.721.4170
> fax  : 512.990.9712
> email: jwayATdatumDOTcom (no spam please)
> -----------------------------------------
>


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: ppc: rework of intLock/intUnlock for nesting?
Date: Mon, 28 Aug 2000 22:06:26 -0400
From: Al Johnston 
Organization: MindSpring Enterprises
Message-ID: <39AB1AA2.F4CF1F0E@mindspring.com>
References: <39A746F1.7B16B2EF@mindspring.com>



Al Johnston wrote:
> 
> Has anyone re-coded a version if the intLock() and intUnlock() routines
> that allow nested calls to these routines without prematurely unlocking
> interrupts?  Seems to me that this should be possible by using the

Looking at the asm src a little harder, It looks like the ppc/2700
version
of these routines ARE nestable.  Not used to things going "right" for
a change.

I still would like any sites/how-to's for using gcc's asm statement.
the stuff in the book does not cover the basics... like 

1) how the %0, %1, etc. stuff in the "opcode field" relates to the 
other two fields, and

2) how do you write asm statements containing multiple opcodes

tnx.

---------------------------

Newsgroups: comp.os.vxworks
Subject: How can i Develop the driver for PCI Display
Date: Mon, 28 Aug 2000 19:01:11 -0700
From: Wayne Hsu 
Organization: http://www.remarq.com: The World's Usenet/Discussions Start Here
Message-ID: <155fba88.26dfdba1@usw-ex0109-068.remarq.com>

I'm a beginer with the VX Works System,who can do the
favour to tell me how i can develope the driver for the PCI
bus w/ the resolution of 1280x1024 or higher.



* Sent from AltaVista http://www.altavista.com Where you can also find related Web Pages, Images, Audios, Videos, News, and Shopping.  Smart is Beautiful

---------------------------

Newsgroups: comp.os.vxworks
Subject: first timer
Date: Tue, 29 Aug 2000 13:12:32 +0800
From: "Bryan de Vera" 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <8ofgsa$qbp$1@overload.lbl.gov>

good day!

i'm a design engineer for a software/firmware company.

i'm been working with VxWorks for about a year now.

i was just wondering how i can join in on the daily digests?

thanks!

the bry

$B!{!|!{!|!{!|!{!|!{!|!{!|!{!|!{!|!{!|!{!|!{!|(B
THE BRY
  Bryan de Vera
  Design Engineer

  MANILA OFFICE:
  Millennium Software, Inc. (MSI)
  25th Flr, Robinson's-Equitable Bank Tower,
    4 ADB Ave. corner Poveda Road,
    Ortigas Center, Pasig City,
    1605, Philippines
  Tel   : +63 2 638-3070 to 72
  Fax   : +63 2 638-3079
  mailto:jdevera@tspi.com.ph

  HEAD OFFICE:
  Tsukiden Software Japan, Inc. (TSJI)
  68 Okidaka, Taiheiji
    Fukushima City, Fukushima Pref.
    960 Japan
  Tel: +81 245 39-8400
  Fax: +81 245 39-8399

  MANILA RESIDENCE:
  3rd Flr, 9624-E Pililla St.,
    Rizal Village, Makati City,
    1208, Philippines
  Tel   : +63 2 895-7204
  mailto:ayt_yall@bigfoot.com
$B!{!|!{!|!{!|!{!|!{!|!{!|!{!|!{!|!{!|!{!|!{!|(B


---------------------------

Newsgroups: comp.os.vxworks
Subject: CompactFlash disk on PC.
Date: Tue, 29 Aug 2000 07:45:05 +0200
From: Hank van Bekkem 
Organization: Oce Technologies B.V.
Message-ID: <39AB4DE1.56ECF231@oce.nl>

Hi,

I'm using:
- - PentiumPro BSP.
- - Tornado 2/VxWorks 5.4.
- - Motorola PATX3000 target.
- - Celeron 433MHz processor.
- - SanDisk 8MB CompactFlash disk.

The PATX3000 features one flash disk connector. It is wired to the
secondary IDE controller. We've installed jumper J13 to designate the
flash disk as master.

We've managed to get the flash disk working with the PATX3000 under
MS-DOS. Commands "format", "dir", "copy" and "del" work fine. The flash
disk is mapped as drive D:.

We are unable to get the flash disk working under VxWorks.

We expected "BlkDev = ataDevCreate(1,0,0,0)" to work, but it returns
NULL.

We also tried in sysLib.c ataResources[], the second entry
(for controller 1), to change the type from ATA_PCMCIA to IDE_LOCAL.
We've also tried to fill ataTypes[] with correct size info. Both didn't
help.

The flash disk should behave identical to an ATA harddisk, on hardware
level. So, VxWorks should access it as a regular ATA harddisk at
controller 1, drive 0. That's why I don't think we should use the
tffsDrv library. The flash disk behaves as a regular ATA harddisk at
hardware level.

So our question is: how can we get our flash disk to work ?

Can anybody help ?

Thanks in advance,
Hank van Bekkem

---------------------------

Newsgroups: comp.os.vxworks
Subject: compression & decompression in vxworks
Date: Tue, 29 Aug 2000 06:02:11 GMT
From: "Rajendra Prasad" 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <8ofkcp$s9g$1@overload.lbl.gov>

Hi,

I require a good compression/decompression software that runs both on 
vxworks 5.4 and sunsolaris 2.6. Has anybody tried  Zlib ? Does it work on 
both vxworks 5.4 and sunsolaris 2.6 ?

thanks and regards,
rpraghuram.
_________________________________________________________________________
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com.

Share information about yourself, create your own public profile at 
http://profiles.msn.com.


---------------------------

Newsgroups: comp.os.vxworks
Subject: Who uses TAO with VxWorks
Date: Tue, 29 Aug 2000 08:39:24 +0200
From: Albert Wijnja 
Organization: MECO Equipment Engineers BV
Message-ID: <39AB5A9B.5F9378D5@meco.nl>
Reply-To: Albert.Wijnja@meco.nl

Hello,

We have started a project with TAO and VxWorks. At the moment we are
running into some serious problems (mostly page faults deep in TAO).
- - Is anyone also using TAO on VxWorks?
- - Did you also encounter serious problems?
- - Were you able to solve them?

Please, any help or directions are appreciated!

I also asked several questions in the comp.soft-sys.ace newsgroup but
have had no response sofar :(

Albert Wijnja





---------------------------

Newsgroups: comp.os.vxworks
Subject: watch variables
Date: Tue, 29 Aug 2000 08:54:35 +0200
From: Samuel F. Romero =?iso-8859-1?Q?Garc=EDa?= 
Organization: INTA
Message-ID: <39AB5E2A.FFCF79A0@inta.es>

Hi there!

Do you know how can I keep the set of variables I'm watching during
debugging? Whenever I stop or restart step by step execution I lose them
all...

Thanks in advance!

- --
Samuel F. Romero García    - romerogs@inta.es
Instituto Nacional de Técnica Aeroespacial - INTA
División de Sistemas Electrónicos - Área Ing. Software
Tel. 91 520 13 49 - Crtra. Ajalvir s/n Km 4.5
Torrejón de Ardoz, 28850 - MADRID
- --
If we knew what we were doing, it wouldn't be called research... - A.
Einstein



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: To C or C++?
Date: Thu, 24 Aug 2000 19:34:46 GMT
From: djhystad@msn.com
Organization: Deja.com - Before you buy.
Message-ID: <8o3tcl$bbc$1@nnrp1.deja.com>
References: <39A543EB.C9E19479@mail.com>

You will pay exactly the same performance penalty that you pay for
programs written in a non-embedded environment.  Of course now time
(and maybe memory) are much more important.  It all depends on the
features that you choose.

Things to look out for:
1. Creation of temporary objects.  I would be careful about overriding
operators.

2. Virtual methods.  They are pretty efficient these days. Try to limit
their use in time critical sections.

3. Name mangling makes it almost impossible to use the target shell for
debugging.

In general, I get the impression that using C++ in VxWorks today is a
lot like using C++ on the desktop 5-10 years ago.  It works,  it's
cool, but there is some maturation required in the tools and libraries.

By the way.  You can do OO programing in C (or assembly for that
matter), and ANSII C has pretty good type checking these days.

In article <39A543EB.C9E19479@mail.com>,
  Henry  wrote:
> Hi, All,
>
> This is not my intention to start a language war. I want to konw
should
> we C or C++, and if we use C++ under VxWork, do we pay penalty on code
> sizes, performace etc.?
>
> I personally prefer C++ for OO and its strong type checking, but I
want
> your opinions.
>
> Thanks in advance.
>
> Henry
>
>


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: using software floating point
Date: Tue, 29 Aug 2000 07:44:29 GMT
From: mei2468@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8ofpkq$fki$1@nnrp1.deja.com>

Our hardware doesn't support floating point. So I have to use software
floating point by including compiler option "msoft-float". It works,
but still remain a problem, see the following code:

#include 
#include 

void test (void)
{
    float x = 1.0;
    int i = 2;

    x = x + i;    //got error, integer can not be casted to double
    ...
}

Does anybody know how to solve this problem except calling some
function like iround(), irintf()...?

Thanks

Mei



Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: re:More PCI Bus & VXWorks
Date: Tue, 29 Aug 2000 09:47:07 +0100
From: "Tim Allen" 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <8ofuua$5v0$1@overload.lbl.gov>

Your header dump shows:

>0x10 BaseAddresses  0x01048000 0x01040000 0x00000000
>                    0x00000000 0x00000000 0x00000000

and you said:

>Is there something I should be doing with translating the PCI address to
>Host address.  What I recall is that the PCI/IO Space in PCI address starts
>at 0x1000000 and the host starts at 0x81000000.

But your base addresses show the device is in PCI memory space, not I/O
space which is indicated by having bit 0 set. In this case you should try
accessing 0xc1048000.

Regards,
Tim.
- --------------------
Tim Allen Ltd
Real Time Embedded Software Consultancy
vxworks


---------------------------

Newsgroups: comp.os.vxworks
Subject: re: ppc: rework of intLock/intUnlock for nesting?
Date: Tue, 29 Aug 2000 10:09:37 +0100 (BST)
From: David Laight 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <200008290909.KAA17665@dsl-2.tadpole.co.uk>


> Has anyone re-coded a version if the intLock() and intUnlock() routines
> that allow nested calls to these routines without prematurely unlocking
> interrupts?  Seems to me that this should be possible by using the
> "key" parameter... 

> I was surprised that vxw doesnt already do this.  Is there a problem
> with doing this that I am missing?

Don't know about ppc, but the arm ones are:

_intLock:
00126c24: e10f1000      mrs     r1,cpsr
00126c28: e2010080      and     r0,r1,#0x80
00126c2c: e3811080      orr     r1,r1,#0x80
00126c30: e129f001      msr     cpsr_cf,r1
00126c34: e1a0f00e      mov     pc,lr
_intUnlock:
00126c38: e10f1000      mrs     r1,cpsr
00126c3c: e3c11080      bic     r1,r1,#0x80
00126c40: e2000080      and     r0,r0,#0x80
00126c44: e1811000      orr     r1,r1,r0
00126c48: e129f001      msr     cpsr_cf,r1
00126c4c: e1a0f00e      mov     pc,lr

Which return/load bit 0x80 from/to the cpsr as expected.

These routines are also (usefully) callable from an ISR.  I have a horrid
feeling that this isn't necessarily the case.  As usual the vxworks 
documentation isn't clear.

	David

- ----------------------------------------------------------------
David Laight                          email: dsl@tadpole.co.uk
Tadpole Technology plc                phone: +44 1223 428 232
Cambridge, UK                         fax:   +44 1223 428 201


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Project/linker question = Super Projects
Date: Tue, 29 Aug 2000 11:30:28 +0200
From: Rainer Stelzer 
Organization: Stoye GmbH, =?iso-8859-1?Q?K=F6ln?=
Message-ID: <39AB82B4.58058F9C@stoye.de>
References: <8o3e5c$bp2$1@overload.lbl.gov>

Hi vxworkers,

Regarding this issue of my initial email:

> I expect the Tornado II  project management to provide a solution to
> link
> single project objects in one workspace object, that can be linked to
> the rom image.

i got an interesting solution from wind river.
They call it SuperProject....


Multiple Projects
This is an instruction on how to overcome one of the shortcomings of the
TornadoII Project Facility:
How can I link multiple Projects together to make one single SuperProject.

The setup may seem like too much work, but once it is done, it does not have
to be changed unless another project or build.
The idea is to have a super project which includes other projects. The
project file structure is as follows:
\myProject\superProject /* the super project directory */
\myProject\projectN /* n project directories */
\myProject\super_link /* a magic file to make a super project */
\myProject\superProject\super.mk /* another super project magic file */
The directory and file names for superProject, projectN, and workspace may
be anything you want. However, whatever you change the name of superProject
to, the file and directory names must match. Also, the build specification
names must match for all projectN and superProject. The method for taking
existing files and adding them to the project is as
follows:

1. Fire up Tornado II
2. Create the workspace and superProject:

File
Open Workspace
New
Create downloadable application modules for VxWorks
Name: superProject
Location: \target\proj\superProject
Workspace
Add to a New or Existing workspace
\target\proj\MyWorkspace.wsp
A Toolchain: (select one)

In the Files tab of the workspace
Select superProject
Right click to Add Files
Add superProject.c (this file may be empty, but must exist)
Right click to calculate Dependencies
3. Create each projectN

File
New project
New
Create downloadable application modules for VxWorks
Name: projectN
Location: \target\proj\projectN
Workspace
Add to current workspace
A Toolchain: (select one)
4. Add files to each projectN

In the Files tab of the workspace
Select projectN
Right click to Add Files

When all of the files have been added, right click on projectN to calculate
Dependencies.

5. Modify build for superProject

In the Builds tab of the workspace double click on superProject Builds
In the Macros tab:
Change the name field to: SUB_PROJS
Change the value field to: project1 project2 ... projectN
Click Add/Set
Click Apply
In the Rules tab
Click New/Edit
Target: -include
Dependencies: ../super.mk
Uncheck Invoke this rule before building project
Click OK
Click New/Edit
Target: superProject.out
Dependencies: Super
Uncheck Invoke this rule before building project
Click OK
Make sure superProject.out is the visible Rule
Click Apply
Click OK

6. Modify each build for projectN

In the Builds tab of the workspace double click on projectN Builds
In the Macros tab change the name field to: include ../../super_link #
Erase the value field
Click Add/Set
Click Apply
Click OK

7. The Miracle
Copy the file super.mk into the superProject directory.
\myProject\superProject\super.mk /* another super project magic file */
Copy the file super_link into the directory above the projectN directories:
\myProject\super_link /* a magic file to make a super project */

8. Build

In the Files tab
Right click on each projectN (do these first)
Select Rebuild All
Right click on superProject (do this last)
Select Rebuild All

You can now download superProject.out.

Files may be added to any projectN or superProject without having to change
any of the build specifications. Building the superProject executes each
projectN make. This will recompile any projectN file that has been changed.

You can still build and download each projectN separately.

You can add additional builds to each projectN (for different architectures
or the simulator). You can also have more than one superProject build. The
only restriction is that there must be the same build name in each projectN
(e.g. SIMNTgnu). Remember to add the Macros and Rules for each build if it
is to be part of the superProject.
An example of this can be found here. It consists of 5 subprojects, each
containing one file containing one function. It's supposed to be logical:
project1 one contains file1, which contains a function called 'f1' printing
"Function 1". Sorry, I know it is not very highly sophisticated, but it
works. The superProject has the above described file 'superProject'
containing one function 'super', which calls the functions of project1 to
projet5. To control if it works, just compile all projects, then download
subproject.out and run sp super. Now you should see the printouts of the
functions of all subprojects.

regards

Rainer

=====================================================
            Gebrüder Stoye GmbH
department: traffic control systems
      name: Rainer Stelzer
       adr: Longericher Strasse 177
      city: Cologne, Germany
       zip: 50739
     phone: ++49/221/2616-551
    mailto: r.stelzer@stoye.de
       www: http://www.stoye.de
=====================================================



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Advice on "global variable" technique
Date: Tue, 29 Aug 2000 16:08:21 +0530
From: Vijay Kumar Peshkar 
Organization: Ye 'Ol Disorganized NNTPCache groupie
Message-ID: <39AB929D.8C09C398@cisco.com>
References: <8oec3r$fql5@nntp.cig.mot.com>
Reply-To: vpeshkar@cisco.com

This is a multi-part message in MIME format.
- --------------6B9E6BE610BD919278564819
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Tim,

Your idea to create only one task variable, viz., a global pointer to a private
(stack-based or dynamically-allocated) structure which contains (what were)
the global variables is really a good one.
You might also consider using the C preprocessor to map the original global
variable references into references through the new global/task
pointer.
But as another mail pointed out, please take care of what these variables
exactly
hold !!!

regards,
Vijay

Tim Berger wrote:

> Hello,
>
> I have a task that spawns two other tasks (sort of a parent-children type
> relationship).  This group of tasks uses a bunch of common variables that
> are accessed often (but never change after they get initialized -- e.g., a
> socket integer; all three of these tasks use a single socket).  Naturally,
> I'd want to put these variables into a global structure, but globals cause
> problems, of course, if I run multiple instances of this group of tasks.
> So, rather than passing everything around, or passing a pointer to a
> structure containing all these variables around, I defined a single
> structure containing all the "global" variables.  The parent task then
> allocates memory for this structure, and then sets a global pointer to this
> structure.  Finally, the task makes the global pointer a task variable.  The
> other two tasks get the task id of the parent task and retrieve the global
> pointer to the structure through a taskVarGet.  This works, and I can have
> many instances of the group of tasks running without any race conditions.
>
> My question: is there a better way to do this?  The method above seems a bit
> hokey to me now (I made it without really thinking it through).  Can't I
> simply allocate memory for a global structure in the parent task and pass
> the pointer to the structure to the children (without using task variables)?
> Or, should I use some other mechanism?
>
> I'm looking for the most reliable method -- this thing has to be
> bulletproof.  What worries me with the simple pointer-passing above is the
> case when a "parent" task dies unexpectedly and the "child" tries to access
> the pointer.  At least with task variables, the taskVarGet will simply fail
> in this case.
>
> Any advice would be greatly appreciated.  Thanks!
>
> Tim

- --------------6B9E6BE610BD919278564819
Content-Type: text/x-vcard; charset=us-ascii;
 name="vpeshkar.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Card for Vijay Kumar Peshkar
Content-Disposition: attachment;
 filename="vpeshkar.vcf"

begin:vcard 
n:Peshkar;Vijay
x-mozilla-html:FALSE
org:Wipro Technologies,K2, Bangalore;Wishing you blue skies, a couple of daisies, some fresh air and a smile!
adr:;;;;;;
version:2.1
email;internet:vijay.peshkar@wipro.com
title:Senior Software Engineer
fn:Vijay Kumar Peshkar
end:vcard

- --------------6B9E6BE610BD919278564819--


---------------------------

Newsgroups: comp.os.vxworks
Subject: pciDeviceShow
Date: Tue, 29 Aug 2000 12:43:06 +0100
From: "Bertram Tenner" 
Organization: Customer of UUNET Deutschland GmbH
Message-ID: <39ab93f4$0$24766@businessnews.de.uu.net>

Hello,

I'm quite new to vxWorks and have the following problem :
When I include the routine pciDeviceShow in my program, during download
the error "unresolved symbol .... _pciDeviceShow ..." occurs.
What settings in the vxWorks image or in certain other files are needed ?

Thanks
                  Bertram




---------------------------

End of New-News digest
**********************


From vxwexplo-errs@csg.lbl.gov  Tue Aug 29 09:30:28 2000
From: David Laight 
Date: Tue Aug 29 09:30:31 PDT 2000
Subject: re: vxworks and IP packet forwarding


> I have a target with 2 ethernet interfaces.
> Is there a way to disable the IP packet forwarding from an ethernet interface
> to the other when the packet is not for my vxWorks target ?

Edit target/h/netLib.h and change:

#define IP_FLAGS_DFLT           (IP_DO_FORWARDING | IP_DO_REDIRECT | \
                                 IP_DO_CHECKSUM_SND | IP_DO_CHECKSUM_RCV)
to:
#define IP_FLAGS_DFLT           (/* IP_DO_FORWARDING | */ IP_DO_REDIRECT | \
                                 IP_DO_CHECKSUM_SND | IP_DO_CHECKSUM_RCV)

(or do something else that has the same effect).

As to why this isn't either:
a) documented
b) in the standard config files
one would have to ask the vxWorks founding fathers!

	David

                            
----------------------------------------------------------------
David Laight                          email: dsl@tadpole.co.uk
Tadpole Technology plc                phone: +44 1223 428 232
Cambridge, UK                         fax:   +44 1223 428 201



From vxwexplo-errs@csg.lbl.gov  Wed Aug 30 04:03:10 2000
From: Vxworks Exploder 
Date: Wed Aug 30 04:03:12 PDT 2000
Subject: comp.os.vxworks newsdigest

Comp.Os.Vxworks Daily Digest    Wed Aug 30 04:03:06 PDT 2000

        Subject: Re: compression & decompression in vxworks
        Subject: Re: Booting vxWorks with MCPN765
        Subject: Re: ppc: rework of intLock/intUnlock for nesting?
        Subject: Re: pciDeviceShow
        Subject: *URGENT* Treatement of Packet Fragmentation
        Subject: Re: Multicast on VxWorks
        Subject: Re: Advice on "global variable" technique
        Subject: Re: Advice on "global variable" technique
        Subject: Re: long file names using DOSFS
        Subject: Re: What am I not getting? (MUX layer question)
        Subject: Re: To C or C++?
        Subject: re: vxworks and IP packet forwarding
        Subject: Spawning C++ task
        Subject: How to use the "gasp" in VxWorks on Windows?
        Subject: Green Hills Software
        Subject: Re: vxworks and IP packet forwarding
        Subject: Re: Project/linker question = Super Projects
        Subject: Memory Mapping on the MVME2604 again
        Subject: bootrom.hex fails to build
        Subject: T2 target shell problems
        Subject: Tornado 1.0.1 problems going from Unix to NT?
        Subject: R: Loading a device driver
        Subject: Re: Advice on "global variable" technique
        Subject: Re: Multiple symbol tables
        Subject: How to upgrade Tornado 1.0.1 to 2.0?
        Subject: MIPS requirements
        Subject: Re: Memory Mapping on the MVME2604 again
        Subject: Re: compression & decompression in vxworks
        Subject: Re: writing a device driver
        Subject: A Puzzle
        Subject: Memory Alignment Problems when Networking 
        Subject: bootrom customization
        Subject: Re: Memory Mapping on the MVME2604 again
        Subject: URGENT!Virtual Memory problem!
        Subject: Re: bootrom customization
        Subject: Re: How to upgrade Tornado 1.0.1 to 2.0?
        Subject: Intel PRO100+ Ethernet card on a PC
        Subject: How to work with PCI ?

-------------------------------------------------------

Newsgroups: comp.os.vxworks
Subject: Re: compression & decompression in vxworks
Date: Tue, 29 Aug 2000 16:30:48 +0530
From: Vijay Kumar Peshkar 
Organization: Ye 'Ol Disorganized NNTPCache groupie
Message-ID: <39AB97E0.1E98D316@cisco.com>
References: <8ofkcp$s9g$1@overload.lbl.gov>
Reply-To: vpeshkar@cisco.com

This is a multi-part message in MIME format.
- --------------3C408EB2BFC4A994B849ED6F
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Rajendra,

Why not use the one already present?. I am refering to the one used for
compressing
and decompressing the image to & from ROM.

regards,
Vijay

Rajendra Prasad wrote:

> Hi,
>
> I require a good compression/decompression software that runs both on
> vxworks 5.4 and sunsolaris 2.6. Has anybody tried  Zlib ? Does it work on
> both vxworks 5.4 and sunsolaris 2.6 ?
>
> thanks and regards,
> rpraghuram.
> _________________________________________________________________________
> Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com.
>
> Share information about yourself, create your own public profile at
> http://profiles.msn.com.

- --------------3C408EB2BFC4A994B849ED6F
Content-Type: text/x-vcard; charset=us-ascii;
 name="vpeshkar.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Card for Vijay Kumar Peshkar
Content-Disposition: attachment;
 filename="vpeshkar.vcf"

begin:vcard 
n:Peshkar;Vijay
x-mozilla-html:FALSE
org:Wipro Technologies,K2, Bangalore;Wishing you blue skies, a couple of daisies, some fresh air and a smile!
adr:;;;;;;
version:2.1
email;internet:vijay.peshkar@wipro.com
title:Senior Software Engineer
fn:Vijay Kumar Peshkar
end:vcard

- --------------3C408EB2BFC4A994B849ED6F--


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Booting vxWorks with MCPN765
Date: Tue, 29 Aug 2000 13:21:04 +0200
From: Marc Waber / 30060 
Organization: Swisscom IP+ (post doesn't reflect views of Swisscom)
Message-ID: <39AB9CA0.CFB713AE@stest.ch>
References: <39AA516F.3A85F3D4@stest.ch> <8oegj5$5f0$1@news1.xs4all.nl> <8oeqr5$n80$1@slb2.atl.mindspring.net>

Bill Dennen wrote:
> 
> johan@borksoft.xs4all.nl writes:
> 
> If you're willing to look outside of VxWorks, select the board to boot Bug
> rather than VxWorks and try nping.  (Your network addresses, set by niot,
> were probably set up earlier when you burned in the kernel).  Format is
> nping 0 0 board_ip host_ip packet_count
> 
> It takes a while so set packet_count low and read the return carefully (it's
> packets LOST not packets returned).  The error code returned can be quite
> helpful -- see Bug manual appendices at http://www.mcg.mot.com/literature.
> 
> No carrier implies the wire's not seen ...
> 
> Regards
> >@<
> 
> >Marc Waber  wrote:
> >: Dear all!
> 
> >: I am trying to boot vxWorks on a MCPN765 over Ethernet with a PowerPC
> >: 7400 Processor in a slave configuration with a CPV5350.
> 
> >: When booting, I always get the following output:
> 
> >: interrupt: dc0 - no carrier
> 
> >According to some other mails in the group there are several possible
> >solutions. There is an SPR on a related problem (I don't know if it is
> >this problem). The number of theis SPR is 28573.
> >Someone else suggested putting the port in half duplex mode.
> >See the newsgroup archives on deja.com and search for "dc0 - no carrier",
> >and you might find some more articles.
> >Groeten,
> >       Johan
> >@<                                     bowtie@netcom.com
> Bill Dennen                             william.dennen@motorola.com
>         Motorola Computer Group, TOPAZ

nping works ok. I can download files over the etheret with ppcbug. I
used it to flash the vxWorks boot kernel, so I think the problem is not
with a bad network connection.

Regards,
Marc
- -- 
Marc Waber				E-mail:    marc.waber@stest.ch
STS Switching Test Solutions AG		Phone:	   +41-76-355-6713
Foerrlibuckstr. 62 / Postfach 74	Fax:  	   +41-1-355-6605
CH-8037 Zurich				Homepage:  www.stest.com

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: ppc: rework of intLock/intUnlock for nesting?
Date: Tue, 29 Aug 2000 13:57:44 -0500
From: "Keith McHugh" 
Organization: Esat Net Customer
Message-ID: <39abb296.0@seat.s3two.ie>
References: <200008290909.KAA17665@dsl-2.tadpole.co.uk>
Reply-To: "Keith McHugh" 

could you pl. give me the reference where from you got all these op-code
instruction coding? I could not find in the power pc programming manual?!?!

"David Laight"  wrote in message
news:200008290909.KAA17665@dsl-2.tadpole.co.uk...
>
> > Has anyone re-coded a version if the intLock() and intUnlock() routines
> > that allow nested calls to these routines without prematurely unlocking
> > interrupts?  Seems to me that this should be possible by using the
> > "key" parameter...
>
> > I was surprised that vxw doesnt already do this.  Is there a problem
> > with doing this that I am missing?
>
> Don't know about ppc, but the arm ones are:
>
> _intLock:
> 00126c24: e10f1000      mrs     r1,cpsr
> 00126c28: e2010080      and     r0,r1,#0x80
> 00126c2c: e3811080      orr     r1,r1,#0x80
> 00126c30: e129f001      msr     cpsr_cf,r1
> 00126c34: e1a0f00e      mov     pc,lr
> _intUnlock:
> 00126c38: e10f1000      mrs     r1,cpsr
> 00126c3c: e3c11080      bic     r1,r1,#0x80
> 00126c40: e2000080      and     r0,r0,#0x80
> 00126c44: e1811000      orr     r1,r1,r0
> 00126c48: e129f001      msr     cpsr_cf,r1
> 00126c4c: e1a0f00e      mov     pc,lr
>
> Which return/load bit 0x80 from/to the cpsr as expected.
>
> These routines are also (usefully) callable from an ISR.  I have a horrid
> feeling that this isn't necessarily the case.  As usual the vxworks
> documentation isn't clear.
>
> David
>
> ----------------------------------------------------------------
> David Laight                          email: dsl@tadpole.co.uk
> Tadpole Technology plc                phone: +44 1223 428 232
> Cambridge, UK                         fax:   +44 1223 428 201
>



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: pciDeviceShow
Date: 29 Aug 2000 13:19:20 GMT
From: johan@borksoft.xs4all.nl
Organization: Borksoft
Message-ID: <8ogd8o$ftq$1@news1.xs4all.nl>
References: <39ab93f4$0$24766@businessnews.de.uu.net>

Bertram Tenner  wrote:
: Hello,

: I'm quite new to vxWorks and have the following problem :
: When I include the routine pciDeviceShow in my program, during download
: the error "unresolved symbol .... _pciDeviceShow ..." occurs.
: What settings in the vxWorks image or in certain other files are needed ?

You have to include the module in your VxWorks image. I just saw this
morning that there is an SPR out that includes this in the configuration,
so that you can include it using Tornado. The number of this SPR is 33972.

Groeten,
	Johan

- -- 
     o o o o o o o . . .   ___________________________________
    o      _____           ||    Johan Borkhuis              |
  .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
 >(________|__|_[_________]_|________________________________|
 _/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: *URGENT* Treatement of Packet Fragmentation
Date: Tue, 29 Aug 2000 22:56:34 +0900
From: "Á¤´ëÁø" 
Organization: System Engineering Research Institute (SERI)
Message-ID: <8ogf92$l35$1@green.kreonet.re.kr>

Hello. We've developed PPC860-based ethernet switching device and

wrote a new END driver for Ethernet CSMA/CD.

We executed "PING" program with 1473 bytes at end station ( desktop PC ).

  C:\ > ping 10.1.1.200 -l 1473

The end station sends two packets  to the device: one is 1514 bytes, and the
other 60 bytes.

Then the Ethernet END driver does not response to the end station, that is,
timed out.

We pings 1472 bytes as following.

  C:\ > ping 10.1.1.200 -l 1472

The end station sends one packet to the device, where length of packet is
1514 bytes.

At this time, the END driver responses to the end station so that the
timeout does not occur any more.

How can we treat the packet fragmentation?


  Some of initialization codes are as following.

   --------------------------------------------------

  #define LAN_LOAD_STRING
0xFF000000:0x34:-1:0x0:-1:-1:0x40:0x30:0x0:0xff:2:0x4

  END_TBL_ENTRY endDevTbl [] =
  {
        { 0, CPM_LOAD_FUNC, CPM_LOAD_STRING, 1, NULL,FALSE},
        { 1, LAN_LOAD_FUNC,  LAN_LOAD_STRING, 1, NULL, FALSE},
        { 0, END_TBL_END, NULL, 0, NULL, FALSE},
  };

  #define LAN_BUFSIZ      (ETHERMTU + ENET_HDR_REAL_SIZ + 6)

  static M_CL_CONFIG lan2MclConfig =  /* network mbuf configuration table */
  {
       /*
         no. mBlks        no. clBlks    memArea   memSize
        --------          ---------    -------     -------
       */
    NUM_NET_MBLKS,    NUM_NET_CLBLKS,    NULL,
(NUM_NET_MBLKS*(M_BLK_SZ+sizeof (long))+NUM_NET_CLBLKS*(CL_BLK_SZ +
sizeof(long)))


  static CL_DESC lan2ClDescTbl [] =  /* network cluster pool configuration
table */
  {
    /*
    clusterSize              num     memArea    memSize
    -----------            ----     --------     -------
    */
    { LAN_BUFSIZ,    NUM_NET_CLBLKS,      NULL,         ((LAN_BUFSIZ +
8)*NUM_NET_CLBLKS) }
  };

  END_OBJ* EndDrv_Load ( char *initString )
  {
    .........

       if  ( (END_OBJ_INIT (&gEndDrvTbl[cur_unit].end,
(DEV_OBJ*)&(gEndDrvTbl[cur_unit]),
                    gEndDrvTbl[cur_unit].szDevName,
gEndDrvTbl[cur_unit].unit,
                   &lanBridgeFuncTable, "EtherDrv" ) == ERROR) ||
             (END_MIB_INIT
 &(gEndDrvTbl[cur_unit].end),M2_ifType_ethernet_csmacd,
                   gEndDrvTbl[cur_unit].enetAddr, 6, ETHERMTU, LAN_SPEED )
== ERROR) )
       {
              printf("END Driver[%s] or MIB-II Structure Load Fail...\n",
szDevName );
              goto ERROR_EXIT;
      }

      END_OBJ_READY ( &(gEndDrvTbl[cur_unit].end),  IFF_UP | IFF_RUNNING
|IFF_NOTRAILERS |
                       IFF_BROADCAST | IFF_PROMISC | IFF_MULTICAST
|IFF_ALLMULTI | IFF_SIMPLEX | IFF_POINTOPOINT );

      .............

 ERROR_EXIT:

       ..............


   --------------------------------------------------

* ETHERMTU is 1500 as described in the VxWorks reference manual.

How can we know your contact points to technical support ?


Anyhelp will be appreciated.


- ---------------------------------
Chaewoo Rha, Research Engineer,
Research & Developement Centre
TELEWARE Network Systems Co.
cwrha@tware.co.kr


- ---------------------------------
Chaewoo Rha, Research Engineer,
Research & Developement Centre
TELEWARE Network Systems Co.

http://www.tware.co.kr
http://www.nettaxi.com/citizens/lalabs
lalabs@nettaxi.com
cwrha@tware.co.kr

Tel.             : 82-31-420-7947
Internet FAX : 82-2-6276-8484(PhoneID=03434207947)
C.P.            : 82-016-328-0275
- ---------------------------------




---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Multicast on VxWorks
Date: Thu, 24 Aug 2000 13:36:54 +0900
From: "Sung Won Ho" 
Organization: ELIMnet Internet News
Message-ID: <8o28vg$571$1@news.elim.net>
References: <8nv98c$coh$1@news.elim.net> <39a428b9.2303323454@news>

Hi Bob,

Yes, I've already included INCLUDE_IGMP.
There maybe another reason with my system.
But thanks anyway.

By the way, I heard that T1 doesn't support multicast.
Because multicast feature is only available in MUX interface driver called
END.
And T2 only supports END interface.
I heard that T1 only support BSD interface network driver.

Is it possoble for one to use multicast feature in T1?

Best regards,
Sung.


"Bob DeJoie"  wrote in message
news:39a428b9.2303323454@news...
> When I converted an application that was using multicast from T1 to T2
> it didn't work initialiiy because in T2 IGMP wasn't enabled.
>
> Do you explicitly include IGMP?
>
>
>
> On Wed, 23 Aug 2000 10:23:33 +0900, "Sung Won Ho" 
> wrote:
>
> >Hi all,
> >
> >Does anyone have an experience in using multicast feature in VxWorks?
> >I use T2 and END driver.
> >But it seemed that it can't process multicast packet.
> >
> >Do you have any comments?
> >
> >Sung.
> >
> >
> >
>
> ================================================
> Bob DeJoie
> Kwajalein Missile Range
> Republic of the Marshall Islands
> email: rjdejoie@kmrmail.kmr.ll.mit.edu    
> phone: 805-355-5855   fax: 805-355-3833
> Time Zone: GMT +12 hours                    ?
> ================================================



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Advice on "global variable" technique
Date: Tue, 29 Aug 2000 09:05:31 -0500
From: "Tim Berger" 
Organization: Motorola CIG
Message-ID: <8ogfuc$fkl1@nntp.cig.mot.com>
References: <8oec3r$fql5@nntp.cig.mot.com> <8oegu7$5f0$2@news1.xs4all.nl>

Yes, I've thought of that, but the problem is that the code is very
modularized, and almost every function uses one of these "global" variables.
I'd have to pass the pointer to every function... but I'm willing to do if
this is a more reliable and efficient way of doing it.

Thanks for your reply,

Tim

 wrote in message
news:8oegu7$5f0$2@news1.xs4all.nl...
> Another possibility is to have a variable in the main function of the
> parent task. You can pass a pointer to this variable to the child task.
> You can have as many parent task as you need, because the variables are
> created on the stack.
>
> Groeten,
> Johan




---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Advice on "global variable" technique
Date: Tue, 29 Aug 2000 09:26:27 -0500
From: "Tim Berger" 
Organization: Motorola CIG
Message-ID: <8ogh5k$bdd1@nntp.cig.mot.com>
References: <8oec3r$fql5@nntp.cig.mot.com> <_cCq5.6637$Dxe.200009216@news.telia.no>

I was under the impression that concurrent calls on a single socket were OK
in VxWorks (they would just get queued up)...

Where did you see this (is it in the vxworks docs somewhere?)

"telia"  wrote in message
news:_cCq5.6637$Dxe.200009216@news.telia.no...
>

> This is not a response to your question, just a warning !
> Be aware that sockets calls may screw up if you perform concurrent
> read/write operations on a single socket.




---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: long file names using DOSFS
Date: Tue, 29 Aug 2000 18:10:55 +0200
From: Leonid Rosenboim 
Organization: BitBand Technologies Ltd. http://www.bitband.com
Message-ID: <39ABE08F.840491FD@bitband.com>
References: <8odm14$lqr$1@news.netvision.net.il>
Reply-To: leonid@bitband.com

Ezra,

for the long names options to take effect, you must reformat the disk drive
using dosFsMkfs().

You will then be able to create long (up to 40 chars) file names, case
sensitive,
but this volume wont be compatible with Microsoft anymore. Beware of this
Long names option if you use TFFS.

DosFs2 supports VFAT style (MSFT compatinble) file names by default.

Ezra Koper wrote:

> Hello All.
>
> I am using VxWorks DOSFS ( not dosfs2) which best to my knowladge support
> only FAT16 and is limited to 8 chars + 3 ext char file names.
>
> I have seen that i can configure long file name option:
>
> status = dosFsMkfsOptionsSet (DOS_OPT_LONGNAMES);
> if (status != OK)
>         printf("\n\r Error - unable to define long file name system!!");
>
> When i run it it doesn't return with error, but i can't create file names
> longer then 8 chars.
> Do i need DOSFS2 in order to support it? or i am doing somthing wrong?


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: What am I not getting? (MUX layer question)
Date: Tue, 29 Aug 2000 11:39:21 -0500
From: Juana Nakfour 
Organization: Motorola CIG
Message-ID: <39ABE739.24E7B61E@email.mot.com>
References: <8oeo8r$9ue$1@overload.lbl.gov>

Martin,

 Try muxShow() to make sure :
1. The new protocol is bound to that specific interface
2. The new protocol has the highest TYPE (SNARF mode)

I have done the same thing and it worked !!
Hope this helps

Juana

Martin Mayer wrote:

> 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_01C0113C.B04F9790
> Content-Type: text/plain;
>         charset="iso-8859-1"
>
> Hi,
>
> (Please reply to all)
>
> I am trying to create my own little protocol and am having trouble with the
> vxWorks MUX layer.
>
> I am binding in very simple test routines. The bind works but I can't seem
> to get any response from the MUX layer.
>
> Here is what I do...
>
> I download the following code into the target (see lower)...
> I type    bind "motfcc",0     and it works...
> I try to ping the target's IP and expect to see that I "Received data"...
> (Is this a good assumption?)
> P.S. I also tried an ipDetach and got rid of the vxWorks IP stack... still
> doesn't work.
>
> Why is the MUX layer not calling my snarfRcvRtn function?
>
> Thanks,
> Martin Mayer
> Comms software designer
>
> int snarfBind(char* dev, int num)
> {
>    /* Bind this protocol to a device via the mux */
>    if (!(IN_COOKIE[num] = (END_OBJ*) muxBind(dev, num, snarfRcvRtn,
>             snarfShutdownRtn, snarfRestartRtn, snarfErrorRtn,
>             MUX_PROTO_SNARF, "myIP", NULL))) {
>       return ERROR;
>    }
>    return OK;
> }
>
> BOOL snarfRcvRtn(void* cookie, long type, M_BLK_ID netbuff,
>                  LL_HDR_INFO* linkHdr, void* spare)
> {
>    printf("Received data\n");
>    return 0;
> }
>
> STATUS snarfShutdownRtn(void* cookie, void* spare)
> {
>    printf("Shutting down snarfProto\n");
>    return muxUnbind(cookie, MUX_PROTO_SNARF, snarfRcvRtn);
> }
>
> void snarfErrorRtn(END_OBJ* end, END_ERR* error, void* spare)
> {
>    printf("snarfError called\n");
> }
>
> STATUS snarfRestartRtn(void* cookie, void* spare)
> {
>    printf("snarfTxRestartRtn called\n");
>    return OK;
> }
>
> STATUS unbind(void)
> {
>    return muxUnbind(IN_COOKIE[0], MUX_PROTO_SNARF, snarfRcvRtn);
> }
>
> int bind(char* name, int i)
> {
>    /* Do network init for all ports on device - except Net Mgment Port */
>    char addrBuffer[20];
>
>    if (!muxDevExists(name,i))
>       printf("%s%d DOES NOT exist   ",name,i);
>    else
>       printf("%s%d DOES exist   ",name,i);
>
>    if( snarfBind(name, i) != OK)
>       printf("Bind FAILED");
>    else
>       printf("Bind OK");
>
>    printf("\n");
>    return 1;
> }
>
> ------_=_NextPart_001_01C0113C.B04F9790
> Content-Type: text/html;
>         charset="iso-8859-1"
> Content-Transfer-Encoding: quoted-printable
>
> 
> 
> 
>  charset=3Diso-8859-1">
>  5.5.2652.35">
> What am I not getting? (MUX layer question)
> 
> 
>
> 
Hi,
> 

>
> 
(Please reply to all)
> 

>
> 
I am trying to create my own little =
> protocol and am having trouble with the vxWorks MUX layer. 
> 
 
> 
I am binding in very simple test =
> routines. The bind works but I can't seem to get any response from the =
> MUX layer. 
> 

>
> 
Here is what I do...
> 

>
> 
I download the following code into the =
> target (see lower)...
> 
I type    bind =
> "motfcc",0     and it works...
> 
I try to ping the target's IP and =
> expect to see that I "Received data"... (Is this a good =
> assumption?)
> 
P.S. I also tried an ipDetach and got =
> rid of the vxWorks IP stack... still doesn't work.
> 

>
> 
Why is the MUX layer not calling my =
> snarfRcvRtn function?
> 

>
> 
Thanks,
> 
Martin Mayer
> 
Comms software designer
> 

>
> 
int snarfBind(char* dev, int =
> num)
> 
{
> 
   /* Bind this protocol to =
> a device via the mux */
> 
   if (!(IN_COOKIE[num] =3D =
> (END_OBJ*) muxBind(dev, num, snarfRcvRtn,
> 
 FACE=3D"Arial">         &nb=
> sp;  snarfShutdownRtn, snarfRestartRtn, snarfErrorRtn,
> 
 FACE=3D"Arial">         &nb=
> sp;  MUX_PROTO_SNARF, "myIP", NULL))) {
> 
      return =
> ERROR;
> 
   }
> 
   return OK;
> 
}
> 

>
> 
BOOL snarfRcvRtn(void* cookie, long =
> type, M_BLK_ID netbuff,
> 
 FACE=3D"Arial">         &nb=
> sp;       LL_HDR_INFO* linkHdr, void* =
> spare)
> 
{
> 
   printf("Received =
> data\n");
> 
   return 0;
> 
}
> 

> 

>
> 
STATUS snarfShutdownRtn(void* cookie, =
> void* spare)
> 
{
> 
   printf("Shutting =
> down snarfProto\n");
> 
   return muxUnbind(cookie, =
> MUX_PROTO_SNARF, snarfRcvRtn);
> 
}
> 

> 

>
> 
void snarfErrorRtn(END_OBJ* end, =
> END_ERR* error, void* spare)
> 
{
> 
   printf("snarfError =
> called\n");
> 
}
> 

>
> 
STATUS snarfRestartRtn(void* cookie, =
> void* spare)
> 
{
> 
   =
> printf("snarfTxRestartRtn called\n");
> 
   return OK;
> 
}
> 

>
> 
STATUS unbind(void)
> 
{
> 
   return =
> muxUnbind(IN_COOKIE[0], MUX_PROTO_SNARF, snarfRcvRtn);   =
> 
> 
}
> 

>
> 
int bind(char* name, int i)
> 
{
> 
   /* Do network init for =
> all ports on device - except Net Mgment Port */
> 
   char =
> addrBuffer[20];
> 

>
> 
   if =
> (!muxDevExists(name,i))
> 
      =
> printf("%s%d DOES NOT exist   ",name,i);
> 
   else
> 
      =
> printf("%s%d DOES exist   ",name,i);
> 

>
> 
   if( snarfBind(name, i) =
> !=3D OK)
> 
      =
> printf("Bind FAILED");
> 
   else
> 
      =
> printf("Bind OK");
> 

>
> 
   =
> printf("\n");
> 
   return 1;
> 
}
> 

>
> 
> 
> ------_=_NextPart_001_01C0113C.B04F9790--


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: To C or C++?
Date: Tue, 29 Aug 2000 10:42:56 -0500
From: Steven Deans 
Organization: Fujitsu Companies of North America
Message-ID: <39ABD9FF.AED8C0E8@fnc.fujitsu.com>
References: <39A543EB.C9E19479@mail.com> <39a61a22.375055037@news.pacificnet.net>

Are the added libraries VxWorks libraries or g++ libraries.
There are changes, that I cannot go into, in the next release that could
make a difference in this area.

Please reply as this will help an ongoing evaluation of OSes we are under
going.

Thanks,
Steven


Max wrote:

> On Thu, 24 Aug 2000 11:48:59 -0400, Henry  wrote:
>
> >Hi, All,
> >
> >This is not my intention to start a language war. I want to konw should
> >we C or C++, and if we use C++ under VxWork, do we pay penalty on code
> >sizes, performace etc.?
> >
> >I personally prefer C++ for OO and its strong type checking, but I want
> >your opinions.
> >
> >Thanks in advance.
> >
> >Henry
> >
>
> It also depends on the amount of work to be done.
> Smaller projects tend to be easier to write and maintain in C, as OOP
> always requires some overhead in system layout and class preparation.
> However, in larger projects this overhead pays back handsomely.
>
> Another issue of C++, especially for VxWorks, is the tremendous size
> of the linked C++ support libraries. In my experience with ARM cores
> the output binary with OO blew at least twice the size of a plain C
> system, unless rigorously controlled. At the same time, the
> performance degradation is much less noticeable.
>
> Max.
>
> Cambury Technologies, Inc
> embedded wireless systems
> http://www.cambury.com


---------------------------

Newsgroups: comp.os.vxworks
Subject: re: vxworks and IP packet forwarding
Date: Tue, 29 Aug 2000 17:30:37 +0100 (BST)
From: David Laight 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <200008291630.RAA19650@dsl-2.tadpole.co.uk>


> I have a target with 2 ethernet interfaces.
> Is there a way to disable the IP packet forwarding from an ethernet interface
> to the other when the packet is not for my vxWorks target ?

Edit target/h/netLib.h and change:

#define IP_FLAGS_DFLT           (IP_DO_FORWARDING | IP_DO_REDIRECT | \
                                 IP_DO_CHECKSUM_SND | IP_DO_CHECKSUM_RCV)
to:
#define IP_FLAGS_DFLT           (/* IP_DO_FORWARDING | */ IP_DO_REDIRECT | \
                                 IP_DO_CHECKSUM_SND | IP_DO_CHECKSUM_RCV)

(or do something else that has the same effect).

As to why this isn't either:
a) documented
b) in the standard config files
one would have to ask the vxWorks founding fathers!

	David

                            
- ----------------------------------------------------------------
David Laight                          email: dsl@tadpole.co.uk
Tadpole Technology plc                phone: +44 1223 428 232
Cambridge, UK                         fax:   +44 1223 428 201


---------------------------

Newsgroups: comp.os.vxworks
Subject: Spawning C++ task
Date: Tue, 29 Aug 2000 17:23:46 GMT
From: "G Kraemer" 
Organization: @Home Network
Message-ID: 

VxWorks experts:

I have a MVME-2700 using vxWorks 5.4 (T2) and I have added a C function,
usrpgms() called from usrApplInit, which installs some of my own routines
(file system date/time stamping, printer driver install, clock
initializations, symFindByName(), and taskSpawn()).

I have been delivered several C++ modules which are compiled with the Green
Hills 1.8.9 C++ compiler.  This executable (65M) resides on an exported
network drive which is nfs mounted to the MVME-2700.

Here's my problem:  usrpgms() can load the C++ executable into memory but
when I attempt to spawn it, the C++ executable does not function correctly.
When I let usrpgms() load the C++ executable and taskSpawn from the target
shell it functions correctly.

Target shell execution:
taskSpawn("niu_main",140,0x08,165000,(FUNCTPTR)symaddr)

Can anyone provide me with sample code on how to spawn the C++ executable
from my C function?

Thanks in advance

Garry


int niu_main()     // does not appear to be within a class
{
// C++ code ....
}



---------------------------

Newsgroups: comp.os.vxworks
Subject: How to use the "gasp" in VxWorks on Windows?
Date: Tue, 29 Aug 2000 19:00:18 -0500
From: "Keith McHugh" 
Organization: Esat Net Customer
Message-ID: <39abf981.0@seat.s3two.ie>
Reply-To: "Keith McHugh" 





---------------------------

Newsgroups: comp.os.vxworks
Subject: Green Hills Software
Date: Tue, 29 Aug 2000 18:25:25 GMT
From: gpbeck@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8ogv5q$rpv$1@nnrp1.deja.com>

Is anyone using the Green Hills software to develop/debug their VxWorks
applications?

If so, what is your opinion of the their software?  Is it
reliable/stabile?  How is their support?

Any info would be appreciated.  Thanks.


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: vxworks and IP packet forwarding
Date: Tue, 29 Aug 2000 20:33:28 +0200
From: Sebastien Fauris 
Organization: Club-Internet (France)
Message-ID: <39AC01F8.3C2074AE@club-internet.fr>
References: <200008291630.RAA19650@dsl-2.tadpole.co.uk>

Thanks a lot David !

Cheers,

Sebastien.

David Laight wrote:

> > I have a target with 2 ethernet interfaces.
> > Is there a way to disable the IP packet forwarding from an ethernet interface
> > to the other when the packet is not for my vxWorks target ?
>
> Edit target/h/netLib.h and change:
>
> #define IP_FLAGS_DFLT           (IP_DO_FORWARDING | IP_DO_REDIRECT | \
>                                  IP_DO_CHECKSUM_SND | IP_DO_CHECKSUM_RCV)
> to:
> #define IP_FLAGS_DFLT           (/* IP_DO_FORWARDING | */ IP_DO_REDIRECT | \
>                                  IP_DO_CHECKSUM_SND | IP_DO_CHECKSUM_RCV)
>
> (or do something else that has the same effect).
>
> As to why this isn't either:
> a) documented
> b) in the standard config files
> one would have to ask the vxWorks founding fathers!
>
>         David
>
>
> ----------------------------------------------------------------
> David Laight                          email: dsl@tadpole.co.uk
> Tadpole Technology plc                phone: +44 1223 428 232
> Cambridge, UK                         fax:   +44 1223 428 201


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Project/linker question = Super Projects
Date: Tue, 29 Aug 2000 21:20:14 +0200
From: "Martin Raabe" 
Organization: T-Online
Message-ID: <8oh2k8$iml$11$1@news.t-online.com>
References: <8o3e5c$bp2$1@overload.lbl.gov> <39AB82B4.58058F9C@stoye.de>

Hello Rainer,
there is also another solution from WindRiver.
There is a SourceCode Browsing and Handling software called SNiFF+ for
TornadoII available.
There is also a download at: http://www.takefive.co.at
(37MB)
This is full functioning, but restricted for 200 files handled in the
evaluation process.
For a eval. license you may contact your local WindRiver Representative.

SNiFF+ can handle complex projectstructures also in a team or even
multi-team environment including
SourceCode Browsing/Editing, VersionControl, Configuration Management an
much more.

Give it a try!

Ciao


- --
Martin Raabe
Project Manager
Integration of SNiFF+ in Tornado II and pRISM+
TakeFive Software GmbH ,  a Wind River Company
Carl-Zeiss-Strasse 41
55129 Mainz
T: (+49) 170 / 1838405
F: (+49) 6131 / 35840-99
M: (+49) 170 / 1838405
E: Martin.Raabe@WindRiver.com
W: www.WindRiver.com
W: www.takefive.co.at/faq




---------------------------

Newsgroups: comp.os.vxworks
Subject: Memory Mapping on the MVME2604 again
Date: Tue, 29 Aug 2000 19:13:10 GMT
From: Joe Andrews 
Organization: Vehicle Managememt Systems
Message-ID: <39AC0B46.45B32B76@mw.boeing.com>
Reply-To: joe.andrews@mw.boeing.com
Sender: nntp@news.boeing.com (Boeing NNTP News Access)

I asked a question on this topic before and I wanted to thank you for
your help. It did get me a little farther down the road, but I've hit
another pothole...

I have two MVME2604 boards in a VXI chassis and want to map them to VME
addresses 0x30000000 and 0x32000000 (these are 32MB cards). We have a
program from National Instruments called VIC that allows us to read and
write to specific memory locations on the VME bus. With this program I
can read and write to memory on each card at the addresses mentioned
above. But, using the shell, when I try to read one of those addresses,
I get an access error. vxMemProbe returns a bus error, also.

Since I can read and write to the memory from the VIC software, I'm
assuming I have the memory window set up correctly, but there is
obviously still something wrong in my setup or in my understanding. 

What I did to get the memory visible at the above locations:
Change VME_A32_MSTR_BUS on card 1 to 0x30000000
Change VME_A32_MSTR_BUS on card 2 to 0x32000000
Change VME_A32_MSTR_SIZE to 0x02000000

Built bootrom and vxworks and loaded into cards.

Both processors are set up as processor 0 (problem?)

is there something else I need to do get get vxWorks to see these
addresses properly?

---------------------------

Newsgroups: comp.os.vxworks
Subject: bootrom.hex fails to build
Date: Tue, 29 Aug 2000 20:16:35 +0000
From: Brian Genisio 
Organization: Michigan State University
Message-ID: <39AC1A23.C81654E5@orion-sim.com>

Hello,
I have a system VME PPC 603 board with 64MB memory.  The app I am
running takes ~15 MB of memory.  I have successfully loaded the object
through the shell with development, but now I need to build it into the
VX object for bootup and execute.

So, I linked it in with the MACH_EXTRA define, changed my LOCAL_RAM_SIZE
to 0x04000000 and my RAM_HIGH_ADRS to 0x02000000 (in Makefile
andconfig.h) and built the image (make vxWorks).  Of course, this did
not work, because it was too large.  I read that I needed to rebuild the
bootrom.hex file (make bootrom.hex) but when I do, it fails to link
(undefined reference to `statSymTbl').

My question(s) : Am I setting up the preprocessor defines correctly?
When bootrom.hex is built, does that link into the vxWorks image that my
machine downloads from FTP?  How does bootrom.hex relate to vxWorks
object?  Any Ideas why making vxWorks finishes, but making bootrom.hex
does not?  I dont want to muck with theiir makefiles if it is not
necessary.

Thanks,
Brian


---------------------------

Newsgroups: comp.os.vxworks
Subject: T2 target shell problems
Date: 29 Aug 2000 20:54:02 GMT
From: unbelver@network.jpl.nasa.gov (Carlos Y. Villalpando)
Organization: Concurrent Processing Devices Group, Jet Propulsion Laboratory
Message-ID: <8oh7ta$aq1$1@nntp1.jpl.nasa.gov>

Hello all.

I've got a bit of a problem with the T2 target shell.

We've been developing under T1 for quite some time, and one of the
useful features we use is calling our module's functions from the
_target_ shell and from an init script.

However, having recently switched to T2 kernel and target shell (or at
least we're trying to) we've lost the ability to call functions that
are in C++ Modules.  The problem exists whether the modules are
compiled with either T1 or T2.

Under T1, everything works great.  Under T2 _host_ shell, everything
works great. Under T2 _target_ shell, calling C module functions
works, accessing global variables in C++ modules work, but calling
functions (not just object methods) in C++ modules does not work.  The
error is "Undefined symbol."

'lkup' and symFindByName work and show our symbols.  I can call the
function entry point manually (e.g 0x12345678 (5) ) But the target
shell can't see them.

Any ideas on what I did wrong?  As we are still deep in development,
putting the function in a usrAppInit() would still be a bad idea at
this point and we would really still like to use the target shell for
debugging and manual starts and/or control.

- --Carlos V.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Tornado 1.0.1 problems going from Unix to NT?
Date: Tue, 29 Aug 2000 16:10:05 -0500
From: Eric Lesch 
Organization: Lockheed Martin Missiles and Firing Control - Dallas
Message-ID: <39AC26AC.63FF7E8C@lmco.com>

Our subcontractor developed the VxWorks kernel on a Unix platform, and
we are developing applications for it on an NT platform.  All of our
applications will be downloaded to the target.  Are there any issues
related to the difference in platform?

Both of us are using Tornado 1.0.1.

I have seen file naming problems when going from NT to Unix due to case
sensitivity, but don't see any reason for problems when going from NT to
Unix.

Thanks for your help


---------------------------

Newsgroups: comp.os.vxworks
Subject: R: Loading a device driver
Date: Tue, 29 Aug 2000 23:10:15 +0100
From: "Merlino" 
Organization: TIN
Message-ID: <8oh932$2f2$1@nslave2.tin.it>
References: <8o86fn$qpk$1@nslave2.tin.it> 
Reply-To: "Merlino" 


telia  wrote in message
jgCq5.6638$Dxe.199355904@news.telia.no...
>
> Merlino  wrote in message
> news:8o86fn$qpk$1@nslave2.tin.it...
> > Hi guys,
> >
> > i have an x86  board with a P2 BSP vxWorks 5.4;
> > i would like to run the system this way:
> > 1- loading vxworks image ( without any device driver )
> > 2- scanning the PCI bus for known device
> > 3- loading modules with the suitables device drivers
> > 4- activating device drivers
> >
> > Is it possible ?
> > How ?
> >
> > Thanks
> >
> >
> From where would you load the driver if you have no device-drivers loaded
> ???

at first from bootHost: via FTP; later i will load  driver.o from EIDE;
obviously by BSP let me access a MSDOS5 filesystem.
> > 1- loading vxworks image ( without any device driver )
I meant SCC driver, PCMCIA ATA disk driver, Redundant ethernet driver ecc ..




---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Advice on "global variable" technique
Date: Tue, 29 Aug 2000 17:40:40 -0400
From: viv 
Organization: Lucent Technologies
Message-ID: <39AC2DD8.1A0E6043@lucent.com>
References: <8oec3r$fql5@nntp.cig.mot.com> <39AADF66.13A07406@swri.edu>

VxWorks has a flat mempry model. so if you have a global variable which
hold the socket value the other object files can extern it and use it...

Greg Willden wrote:
> 
> Tim Berger wrote:
> 
> > Hello,
> >
> > I have a task that spawns two other tasks (sort of a parent-children type
> > relationship).  This group of tasks uses a bunch of common variables that
> > are accessed often (but never change after they get initialized -- e.g., a
> > socket integer; all three of these tasks use a single socket).
> 
> If the data don't change after they are initialized why can't you just pass
> those values to the spawned task.  The taskSpawn function allows 10 arguments to
> be passed as parameters to the entry point of the spawned task.  This would
> remove a tremendous amount of complexity.
> 
> --
> Greg Willden      gwillden@swri.edu
> Engineer
> Southwest Research Institute
> www.swri.edu

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Multiple symbol tables
Date: Tue, 29 Aug 2000 17:46:37 -0400
From: viv 
Organization: Lucent Technologies
Message-ID: <39AC2F3D.95E19996@lucent.com>
References: <8o16qq$6g9$1@nnrp1.deja.com>

check ur path....

gpbeck@my-deja.com wrote:
> 
> Does anyone know how to load more that one symbol table with the
> Tornado debugger?  I have more than one object running, but the objects
> were not loaded with the ld command.
> 
> I tried using the add-symbol-file command from the GDB command line,
> but I get the error message...
> 
> Cannot file object module on target: LOADER_OBJ_MODULE_NOT_FOUND
> 
> Any ideas.  Thanks.
> 
> Sent via Deja.com http://www.deja.com/
> Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: How to upgrade Tornado 1.0.1 to 2.0?
Date: Tue, 29 Aug 2000 23:21:47 GMT
From: aw99_1999@yahoo.com
Organization: Deja.com - Before you buy.
Message-ID: <8ohghm$hls$1@nnrp1.deja.com>

We have an application developed on Tornado 1.0.1 and now want to
upgrade it to Tornado 2.0. But I could not find a porting guide for
Tornado 2.0 from Wind River. I know there is a similar one in Tornado
1.0.1. Could anyone give me some suggestions? Thanks a lot,

- ---- Steve


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: MIPS requirements
Date: Tue, 29 Aug 2000 16:48:03 -0700
From: "Eric McDaniel" 
Organization: Posted via Supernews, http://www.supernews.com
Message-ID: 


I am trying to size up the processor I need for a project.  It will be
something in the PPC family, running VxWorks, with a bunch of IP routing
protocols.

I would like to break the problem down into MIPS requirements for the base
O/S, the base TCP/IP stack, and the routing protocols for various levels of
network traffic. There must be some published information like this, right?

Thanks.

Eric McDaniel
ericm@vertical.com




---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Memory Mapping on the MVME2604 again
Date: Tue, 29 Aug 2000 20:14:11 -0500
From: Andrew Johnson 
Organization: APS, Argonne National Laboratory
Message-ID: <39AC5FE3.EA2FED26@aps.anl.gov>
References: <39AC0B46.45B32B76@mw.boeing.com>

Joe Andrews wrote:
> 
> I have two MVME2604 boards in a VXI chassis and want to map them to VME
> addresses 0x30000000 and 0x32000000 (these are 32MB cards). We have a
> program from National Instruments called VIC that allows us to read and
> write to specific memory locations on the VME bus. With this program I
> can read and write to memory on each card at the addresses mentioned
> above. But, using the shell, when I try to read one of those addresses,
> I get an access error. vxMemProbe returns a bus error, also.
> 
> Since I can read and write to the memory from the VIC software, I'm
> assuming I have the memory window set up correctly, but there is
> obviously still something wrong in my setup or in my understanding.
> 
> What I did to get the memory visible at the above locations:
> Change VME_A32_MSTR_BUS on card 1 to 0x30000000
> Change VME_A32_MSTR_BUS on card 2 to 0x32000000
> Change VME_A32_MSTR_SIZE to 0x02000000
> 
> Built bootrom and vxworks and loaded into cards.
> 
> Both processors are set up as processor 0 (problem?)
> 
> is there something else I need to do get get vxWorks to see these
> addresses properly?

Yes, look at the table called sysPhysMemDesc towards the beginning of
sysLib.c.  This defines how the MMU will be set up, and if your memory
range doesn't fall into one of the ones listed it won't allow accesses to
those addresses.  However if you put large chunks of memory into the
sysPhysMemDesc table it has to set up page table entries for it all, which
takes time at startup and uses up RAM too.  There's another table which is
appropriate for large blocks of memory called something like sysBatDesc -
this sets up some special MMU registers that can bypass the page tables;
you should be able to set one Data BAT to cover the range you want, they
will map up to 256MB each.

- - Andrew
- -- 
Every great idea appears crazy to start with.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: compression & decompression in vxworks
Date: Wed, 30 Aug 2000 03:41:57 GMT
From: pkockritz@home.com (Pete Kockritz)
Organization: @Home Network
Message-ID: 
References: <8ofkcp$s9g$1@overload.lbl.gov> <39AB97E0.1E98D316@cisco.com>

The vxWorks deflate/inflate work fine, as long as you don't need to 
deflate (compress) files or data on the target.

I've used zlib with pSOS and on Solaris (and other platforms). It should
work fine on vxWorks, after all, zlib is what the vxWorks inflate/deflate
are based on.

You could also use 'libbzip2' (I haven't used it myself, however). libbzip2
is said to do a somewhat better job of compressing.




In article <39AB97E0.1E98D316@cisco.com>, vpeshkar@cisco.com wrote:
> Rajendra,
> 
> Why not use the one already present?. I am refering to the one used for
> compressing and decompressing the image to & from ROM.
> 
> regards,
> Vijay
> 
> Rajendra Prasad wrote:
> 
> > Hi,
> >
> > I require a good compression/decompression software that runs both on
> > vxworks 5.4 and sunsolaris 2.6. Has anybody tried  Zlib ? Does it work on
> > both vxworks 5.4 and sunsolaris 2.6 ?
> >
> > thanks and regards,
> > rpraghuram.

Regards,
Pete
- --
+-----------------------------------------------------------------+
|       Pete Kockritz                  mailto:pkockritz@home.com  |
|                       This space for rent.                      |
+-----------------------------------------------------------------+

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: writing a device driver
Date: Wed, 30 Aug 2000 10:20:17 +0530
From: Vijay Kumar Peshkar 
Organization: Ye 'Ol Disorganized NNTPCache groupie
Message-ID: <39AC9288.3EB60152@cisco.com>
References: <8oa9dm$ck8$1@nnrp1.deja.com>
Reply-To: vpeshkar@cisco.com

This is a multi-part message in MIME format.
- --------------C4380B9AC66E8FD86D936115
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Here's a clipping from an FAQ I had downloaded long time back...

How do I add select support to my driver?

#include "selectLib.h"
...

xxx_init(...)
{
 ...
 selWakeupListInit(&xxxdev->selwakeupList);
 ...
}

xxx_ioctl(...)
{
 ...
 switch(request_type) {
 ...
 case FIOSELECT:
  selNodeAdd(&xxxdev->selwakeupList,
   (SEL_WAKEUP_NODE *)request_arg);
  if ((selWakeupType((SEL_WAKEUP_NODE*)request_arg) == SELREAD)
   && readable_condition_is_met)
   selWakeup((SEL_WAKEUP_NODE*)request_arg);
  if ((selWakeupType((SEL_WAKEUP_NODE*)request_arg) == SELWRITE)
   && writable_condition_is_met)
   selWakeup((SEL_WAKEUP_NODE*)request_arg);
  break;
 ...
 case FIOUNSELECT:
  selNodeDelete(&xxxdev->selWakeupList,
   (SEL_WAKEUP_NODE*)request_arg);
  break;
 ...
 }
}

And, add calls to selWakeup() as appropriate in your interrupt handlers
and read/write routines as selective conditions are toggled or satisfied.

Hope this helps.

have fun,
Vijay

ehudfl@my-deja.com wrote:

> Where can i find a web site who explain how to write device drivers in
> vxworks and with source code who explain how to do that ?
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.

- --------------C4380B9AC66E8FD86D936115
Content-Type: text/x-vcard; charset=us-ascii;
 name="vpeshkar.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Card for Vijay Kumar Peshkar
Content-Disposition: attachment;
 filename="vpeshkar.vcf"

begin:vcard 
n:Peshkar;Vijay
x-mozilla-html:FALSE
org:Wipro Technologies,K2, Bangalore;Wishing you blue skies, a couple of daisies, some fresh air and a smile!
adr:;;;;;;
version:2.1
email;internet:vijay.peshkar@wipro.com
title:Senior Software Engineer
fn:Vijay Kumar Peshkar
end:vcard

- --------------C4380B9AC66E8FD86D936115--


---------------------------

Newsgroups: comp.os.vxworks
Subject: A Puzzle
Date: Wed, 30 Aug 2000 05:13:48 GMT
From: ellin_lin@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8oi561$8t9$1@nnrp1.deja.com>

I created a bootable project based on BSP MCP750, then tried to build
it with rule vxworks_rom, and got error as below:
undefined reference to "sysGetDramSpd"
undefined reference to "sysGetBusSpd".
Does somebody know how to solve it?

ellin


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Memory Alignment Problems when Networking 
Date: Wed, 30 Aug 2000 05:29:31 GMT
From: john.h.kennedy@worldnet.att.net (John Kennedy)
Organization: AT&T Worldnet
Message-ID: <39ac1b51.1076743@netnews.worldnet.att.net>

Hi,

I've built a TCP/IP Server and UDP/IP Client under T101 and VxWorks
5.3.1 that's to run on a SPARC CPU.  The TCP Server Task collects data
and passes this off to the UDP Client Task. Everything built just fine
and the VxWorks tasks all start-up w/o problem. The UDP Clients run as
expected, but I'm running into problem with the integrated system.  

I'm getting an exception 52 (memory address not aligned) raised
whenever I try to establish a Client connection to the TCP/IP Server .


I've tried spinning this several ways, but always with the same
result.  The debugger has shown the exception to occur whenever I'm
using a resource such as a timer, semaphore or queue.

Has anyone seen this exception before or have any suggestions ?

Thanks in advance.

John


---------------------------

Newsgroups: comp.os.vxworks
Subject: bootrom customization
Date: Wed, 30 Aug 2000 06:08:28 GMT
From: egluska@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8oi8ch$c3o$1@nnrp1.deja.com>

Is there a way to customize vxWorks bootrom using Tornado 2 project
facility? It seems that the project facility allows generation and
customization of vxWorks images, but if you want bootrom, you have
to work in the BSP directory under target/config using the local
makefile, and customization of vxWorks components within the bootom
is done in the "old style" makefile INCLUDE statements.

Thanks

Eran Gluska
PacketLight Networks


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Memory Mapping on the MVME2604 again
Date: Wed, 30 Aug 2000 07:07:29 GMT
From: john_94501@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8oibr9$fsf$1@nnrp1.deja.com>
References: <39AC0B46.45B32B76@mw.boeing.com>

Hi Joe,

You might find that the following tech note is useful:

WTN47 - Navigating the MV2600 from the MPC to the VME and Back MV2603
and MV2604 Board Support Packages

The link is here (you'll need a WindSurf password to get to it):

http://www.wrs.com/csdocs/product/t2/technote/WTN47.pdf

Hope that helps,

John...


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: URGENT!Virtual Memory problem!
Date: Wed, 30 Aug 2000 15:32:36 +0800
From: "kvd2" 
Organization: DCI HiNet
Message-ID: <8oidt0$fmd@netnews.hinet.net>

Dear All,

I've add
{
    (void *) DICE_BASE_ADRS,
    (void *) DICE_BASE_ADRS,
   DICE_SIZE,
   VM_STATE_MASK_VALID | VM_STATE_MASK_WRITABLE | VM_STATE_MASK_CACHEABLE,
    VM_STATE_VALID     | VM_STATE_WRITABLE      | VM_STATE_CACHEABLE_NOT
    },

    {
    (void *) 0x10000000,
    (void *) 0x10000000,
   0x00001000,
   VM_STATE_MASK_VALID | VM_STATE_MASK_WRITABLE | VM_STATE_MASK_CACHEABLE,
    VM_STATE_VALID     | VM_STATE_WRITABLE      | VM_STATE_CACHEABLE_NOT
    }

to sysPhysMemDesc [] . The fisrt memory space is working well.
However, the following error message is returned if I tried to dump the
memory starting at
0x10000000, ie, the 2nd memory space.

WTX Error 0x100ca (AGENT_MEM_ACCESS_ERROR)
value = -1 = 0xffffffff

What's wrong?

kvd2
p.s. I don't have vxVMI( the opt. component of mmu)





---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: bootrom customization
Date: 30 Aug 2000 08:14:17 GMT
From: johan@borksoft.xs4all.nl
Organization: Borksoft
Message-ID: <8oifop$rho$1@news1.xs4all.nl>
References: <8oi8ch$c3o$1@nnrp1.deja.com>

egluska@my-deja.com wrote:
: Is there a way to customize vxWorks bootrom using Tornado 2 project
: facility? It seems that the project facility allows generation and
: customization of vxWorks images, but if you want bootrom, you have
: to work in the BSP directory under target/config using the local
: makefile, and customization of vxWorks components within the bootom
: is done in the "old style" makefile INCLUDE statements.

No, there is no way to configure the bootrom. The bootrom is configured
using the defaults in the config.h. You could use the project facility to
create a bootable project and build a romresident VxWorks image and
include the bootrom code in it.

Groeten,
	Johan

- -- 
     o o o o o o o . . .   ___________________________________
    o      _____           ||    Johan Borkhuis              |
  .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
 >(________|__|_[_________]_|________________________________|
 _/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 upgrade Tornado 1.0.1 to 2.0?
Date: 30 Aug 2000 08:16:27 GMT
From: johan@borksoft.xs4all.nl
Organization: Borksoft
Message-ID: <8oifsr$rho$2@news1.xs4all.nl>
References: <8ohghm$hls$1@nnrp1.deja.com>

aw99_1999@yahoo.com wrote:
: We have an application developed on Tornado 1.0.1 and now want to
: upgrade it to Tornado 2.0. But I could not find a porting guide for
: Tornado 2.0 from Wind River. I know there is a similar one in Tornado
: 1.0.1. Could anyone give me some suggestions? Thanks a lot,

What do you want to upgrade, the BSP or an application? The BSP is usually
about 1 to 2 weeks work, the most usefull resource for this is the BSP
development kit.
If you want to port an application just create a new project and include
the source files.

Groeten,
	Johan

- -- 
     o o o o o o o . . .   ___________________________________
    o      _____           ||    Johan Borkhuis              |
  .][__n_n_|DD[  ====_____  |    johan@borksoft.xs4all.nl    |
 >(________|__|_[_________]_|________________________________|
 _/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: Intel PRO100+ Ethernet card on a PC
Date: Wed, 30 Aug 2000 10:39:39 +0200
From: "Gijs Withagen" 
Organization: XS4ALL Internet BV
Message-ID: <8oigmj$tp$1@news1.xs4all.nl>

Hi All,

Recently we switched from a 3C905C-TX-M Etherlink 10/100 XL PCI ethernet
card to a Intel PRO100+ because the 3COM card leaded occassionally to a page
fault (with the EL_3C90X_END driver).

We managed to get the system running with the Intel PRO100+ (based on the
Intel 82559) with the FEI_END driver but are running into another problem
now.

When we try to FTP a larger file to the VxWorks target system the FTP
session hangs. After downloading some 700+ Kbytes the download stops and
after a while the FTP client displays the message "netout:Connection reset"
(this could be from our FTP client (standard NT4.0 ftp client program), I'm
not
sure). Also, the connection state in the inetstatShow overview shows "1536".

We tried disabling buffer loaning (netConfig.h) and increasing the number of
transmit/receive buffer descriptors (from 32 to 256), to no avail.

Has anyone discovered the same problems and/or solutions for them?

Gijs Withagen
Technolution BV







---------------------------

Newsgroups: comp.os.vxworks
Subject: How to work with PCI ?
Date: Wed, 30 Aug 2000 11:05:03 +0200
From: "Ezra Koper" 
Organization: NetVision Israel	
Message-ID: <8oif6g$5li$1@news.netvision.net.il>

Hello

1.
We are using a Pentium2 333Mhz PC platform. I need to be able to open memory
space (and later I/O space) with a PCI device connected to primary PCI Slot.
I have no experience with PCI & VxWorks, and would like to know how it can
be done.

For example address 0x800000 on my target has to be mapped to address
0x0000000 on primary PCI device.
So writing to address 0x800010 will actually generate writing to address
0x10 on the primary PCI device.

There is almost no ducementation on how to use PCI with VxWorks. Any idea
were i can find more information?

2.
I can't activate commands as sysPciDevShow. This command paper in Widriver
WindSurf WEB site for Technical Solution 130785.5
so what might be the problem?

Note - I have included the PCI support component at VxWorks.




---------------------------

End of New-News digest
**********************


From vxwexplo-errs@csg.lbl.gov  Wed Aug 30 08:11:18 2000
From: "Michael Lawnick" 
Date: Wed Aug 30 08:11:21 PDT 2000
Subject: TFFS on MV260x

Hi folks,

has anybody already integrated TFFS to on board flash of a MV230x, 
MV240x, MV260x or MV270x ?
Could you share your code changes or help me to get on the road ?
I think vxWorks documentation is as simple as usually.

Regards,

MfG
Lawnick, SOFTEC GmbH

==============================================
SOFTEC GmbH               Tel +49-731-96600-0
Promenade 17              Fax +49-731-96600-23
D-89073 Ulm               Michael Lawnick
Germany                   lawnick@softec.de
==============================================



From vxwexplo-errs@csg.lbl.gov  Wed Aug 30 09:22:33 2000
From: David Laight 
Date: Wed Aug 30 09:22:35 PDT 2000
Subject: re: compression & decompression in vxworks

Interestingly I found the full zlib object in my vxWorks image.

I think this arrives as part of the JVM support - but cannot be sure.
It is quite an old version though...  Tornado 2 contains version 1.0.4 compiled 
with z_ prepended to all its symbols.
I stole (then hacked into submission) a header file from a random Linux system - 
for version 1.1.3 and dated July 1998.

Seems to work fine for generating screen dumps in .png format.
(I had barfed at the idea of including SO much code - then found it was
already present!)

I don't think that zlib needs 'porting' as such, I don't think it contains ANY 
undefined references (no system/library calls at all!).

	David


---------------------------------------------------------------
David Laight                          email: dsl@tadpole.co.uk
Tadpole Technology plc                phone: +44 1223 428 232
Cambridge, UK                         fax:   +44 1223 428 201



From vxwexplo-errs@csg.lbl.gov  Wed Aug 30 21:55:01 2000
From: Shrishail 
Date: Wed Aug 30 21:55:03 PDT 2000
Subject: VxWorks on Cadre III board


------ =_NextPart_000_01C0126A.1EF9E1E0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit

Hi List,
   Excuse me if this topic has been discussed earlier; I am new to the list.
   I am using a Motorola 5307 eval board from cardre III. I am facing problems
   in running the BSP. I am using the ROM resident version of the same. 
   Anyone out there who faced similar problem / who was able to use the BSP
   as it came with Tornado (CF5307)?
   
   Any comments / suggestions?
Thanks,
Shrishail
   
------ =_NextPart_000_01C0126A.1EF9E1E0
Content-Type: application/ms-tnef
Content-Transfer-Encoding: base64

eJ8+IiQEAQaQCAAEAAAAAAABAAEAAQeQBgAIAAAA5AQAAAAAAADoAAEIgAcAGAAAAElQTS5NaWNy
b3NvZnQgTWFpbC5Ob3RlADEIAQ2ABAACAAAAAgACAAEEkAYAFAEAAAEAAAANAAAAAwAAMAIAAAAL
AA8OAAAAAAIB/w8BAAAAPwAAAAAAAACBKx+kvqMQGZ1uAN0BD1QCAAAAAHZ4d2V4cGxvQGxibC5n
b3YAU01UUAB2eHdleHBsb0BsYmwuZ292AAAeAAIwAQAAAAUAAABTTVRQAAAAAB4AAzABAAAAEQAA
AHZ4d2V4cGxvQGxibC5nb3YAAAAAAwAVDAEAAAADAP4PBgAAAB4AATABAAAAEwAAACd2eHdleHBs
b0BsYmwuZ292JwAAAgELMAEAAAAWAAAAU01UUDpWWFdFWFBMT0BMQkwuR09WAAAAAwAAOQAAAAAL
AEA6AQAAAAMAcToAAAAAAgH2DwEAAAAEAAAAAAAAAkcvAQSAAQAbAAAAVnhXb3JrcyBvbiBDYWRy
ZSBJSUkgYm9hcmQAAwkBBYADAA4AAADQBwgAHgAKAAcAHgADAC8BASCAAwAOAAAA0AcIAB4ACgAB
ADkAAwBEAQEJgAEAIQAAADNFQzUwQkY1NUI3RUQ0MTE4MDA5MDA1MEJBRDQzNjlBAB4HAQOQBgBg
AwAAFAAAAAsAIwAAAAAAAwAmAAAAAAALACkAAAAAAAMALgAAAAAAAwA2AAAAAABAADkAIADiATwS
wAEeAHAAAQAAABsAAABWeFdvcmtzIG9uIENhZHJlIElJSSBib2FyZAAAAgFxAAEAAAAWAAAAAcAS
PAHi9QvFP35bEdSACQBQutQ2mgAAHgAeDAEAAAAFAAAAU01UUAAAAAAeAB8MAQAAABkAAABzaHJp
c2hhaWxAaW9uaWNtaWNyby5jb20AAAAAAwAGENHnLJYDAAcQMAEAAB4ACBABAAAAZQAAAEhJTElT
VCxFWENVU0VNRUlGVEhJU1RPUElDSEFTQkVFTkRJU0NVU1NFREVBUkxJRVI7SUFNTkVXVE9USEVM
SVNUSUFNVVNJTkdBTU9UT1JPTEE1MzA3RVZBTEJPQVJERlJPTUMAAAAAAgEJEAEAAADOAQAAygEA
AJ0CAABMWkZ1kbWhvgMACgByY3BnMTI1/jIA/wIGAqQD5AXrAoMAUBMDVAIAY2gKwHNldP4yBgAG
wwKDDlAD1QcTAoD+fQqACM8J2QKACoEOcQtg4m4OEDAzMwr7E7IB0CAgSGkgTAQAdCxDCoUZwUV4
Y3USsCDFB4AgBpAgdGgEABrAtG9wDeAgEoAEIGIJ4RwgZAQAGiESsGQgZQUKwGwIkTsgSSBh+G0g
bgfRGyAawRpQHOC9GTAuGWgdQxowC4BnHVAdBdBvGyADYAtgIDUztDA3HKB2B0AbsG8LEfIgA1Ig
YwsRFZAdMCJQ2i4dNGYA0B+icANgAmAYZW1zGWgLgCBydY5uAwAfsR4CQlNQIoVjH4QeAlJPTSSg
B5Bpcw5wAjAgdgSQAJACICDWbxqyGlBzHWBlIoAZaPhBbnkCIBpQCGAFQB4BeSIhd2gd4CLxHIEA
kG2nAxAKwSNlIC8qc3cbkX8BoCOgHcIaMiU1GWgbkWmfBUAh4BpxA/Aa0CBUBbCIbmFkHeAoQ0Yg
kvwpPxloKNsh0ANwB4ACMEMEICwgc3VnZweQdBUnwXMwllQSgG5rc+kZVlNoBRBzEoADEDCvBRSx
ADbAAAADABAQAAAAAAMAERAAAAAAQAAHMID7OTs7EsABQAAIMID7OTs7EsABHgA9AAEAAAABAAAA
AAAAAAMADTT9NwAAxtg=

------ =_NextPart_000_01C0126A.1EF9E1E0--




From vxwexplo-errs@csg.lbl.gov  Thu Aug 31 04:03:08 2000
From: Vxworks Exploder 
Date: Thu Aug 31 04:03:10 PDT 2000
Subject: comp.os.vxworks newsdigest

Comp.Os.Vxworks Daily Digest    Thu Aug 31 04:03:04 PDT 2000

        Subject: C++ dynamic loading problem
        Subject: Creating a Project in T2 question
        Subject: TFFS on MV260x
        Subject: netTask -- what part of a network operation is performed by it?
        Subject: re: compression & decompression in vxworks
        Subject: Re: How to upgrade Tornado 1.0.1 to 2.0?
        Subject: Re: Tornado 1.0.1 problems going from Unix to NT?
        Subject: Re: *URGENT* Treatement of Packet Fragmentation
        Subject: Re: long file names using DOSFS
        Subject: network time protocol
        Subject: Problems writing with dosFs 2.0
        Subject: NFS woes
        Subject: Re: To C or C++?
        Subject: how to sychronize the target/host symbol?
        Subject: VxWorks on Cadre III board
        Subject: differences between vxworks and linux
        Subject: Re: Vxsim File I/O with NT
        Subject: pipes and ioctl
        Subject: VxWorks and dual processors
        Subject: ? Loading a device driver
        Subject: ? Loading a device driver
        Subject: Re: network time protocol
        Subject: Re: URGENT!Virtual Memory problem!

-------------------------------------------------------

Newsgroups: comp.os.vxworks
Subject: C++ dynamic loading problem
Date: Wed, 30 Aug 2000 13:19:21 GMT
From: hokie94@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8oj1k5$8c1$1@nnrp1.deja.com>

Hello,

I am trying to dynamically load a C++ module into a running system.  I
have munched this module to create a static object
constructor/deconstructor file (ctdt.c) and linked this into the
module.

When I load the module, the constructor functions are called
automatically, but then the system hangs.  It looks like the ctdt.c
constructor/deconstructor arrays are globals, and are not uniquely
named for the module, so they are conflicting with the root image ct/dt
arrays.  It looks like I might have to modify the munch.tcl file to
provide a unique module based array name for the ct/dt arrays?  Anyone
have any luck with this?

Thanks,

Jeff


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Creating a Project in T2 question
Date: Wed, 30 Aug 2000 10:33:17 -0400
From: "Mike Kordik" 
Message-ID: 

I have an existing project layout on my hard drive. I am trying to convert
this to a project in T2. I added all the source files from the various
directories to the project. I also had T2 generate the external dependency
list. Because the header files are in a different directory the dependency
scan knows they exist but does not know where to find them. When the
makefile is generated the correct dependencies are there but make does not
know how to find them. How do I get T2 to find these files and include the
path to the header files in the makefile?

If this is possible is it also possible to have T2 generate the path with a
macro so that I can easily transport this project to another computer with a
different base path? i.e. the makefiles for the tornado sample builds uses
WIND_BASE.

Thanks,
Mike



---------------------------

Newsgroups: comp.os.vxworks
Subject: TFFS on MV260x
Date: Wed, 30 Aug 2000 16:54:11 +0200
From: "Michael Lawnick" 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <200008301453.QAA01130@net.softec.de>

Hi folks,

has anybody already integrated TFFS to on board flash of a MV230x, 
MV240x, MV260x or MV270x ?
Could you share your code changes or help me to get on the road ?
I think vxWorks documentation is as simple as usually.

Regards,

MfG
Lawnick, SOFTEC GmbH

==============================================
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: netTask -- what part of a network operation is performed by it?
Date: 30 Aug 2000 15:53:14 GMT
From: Jeff Rugen 
Organization: PrimeNet
Message-ID: <8ojala$okk$1@nnrp03.primenet.com>

First of all, I'm developing for VxWorks 5.4 using Tornado II on a PPC
target.

I have access to the VxWorks networking source, but I think its from
Tornado I, so I don't know what's changed since then.

My basic question is, when I have my own task and I perform a network
operation (primarily send data and recieve data), what code is executed in
netTask, and what code is executed in my own task?

Take the following two cases (sending data and recieving):

myTask -> BSD sockets -> TCP -> IP -> MUX -> ethernet driver

ethernet -> MUX -> IP -> TCP -> BSD sockets -> myTask

Does netTask execute all the code in the TCP/IP/MUX sections?  Or does it
only execute them for recieve?  Or does it work differently? 

The reason I'm asking, is my goal is to add a couple more transport
protocols to the stack -- replacing TCP with TP4, SP3, etc.  I am using
pre-existing code that's seperated into tasks based on functionality (TP4
is a task, SP3 is a task, etc), but depending on how netTask is organized,
that model probably won't work.

Also, does anyone have any advice (or better, documentation) on adding new
transport protocols above the IP layer in VxWorks?  The Network Protocol
Toolkit doesn't help much, because it assumes you're replacing IP as well,
but I'm not doing that.  For example, if I update the inetsw[] array to
add the new protocols, will they automatically be usable by the BSD socket
interface?  The code appears that way, but I can't test it currently.

Thanks for any information

Jeff Rugen

---------------------------

Newsgroups: comp.os.vxworks
Subject: re: compression & decompression in vxworks
Date: Wed, 30 Aug 2000 17:22:46 +0100 (BST)
From: David Laight 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <200008301622.RAA28915@dsl-2.tadpole.co.uk>

Interestingly I found the full zlib object in my vxWorks image.

I think this arrives as part of the JVM support - but cannot be sure.
It is quite an old version though...  Tornado 2 contains version 1.0.4 compiled 
with z_ prepended to all its symbols.
I stole (then hacked into submission) a header file from a random Linux system - 
for version 1.1.3 and dated July 1998.

Seems to work fine for generating screen dumps in .png format.
(I had barfed at the idea of including SO much code - then found it was
already present!)

I don't think that zlib needs 'porting' as such, I don't think it contains ANY 
undefined references (no system/library calls at all!).

	David


- ---------------------------------------------------------------
David Laight                          email: dsl@tadpole.co.uk
Tadpole Technology plc                phone: +44 1223 428 232
Cambridge, UK                         fax:   +44 1223 428 201


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: How to upgrade Tornado 1.0.1 to 2.0?
Date: Wed, 30 Aug 2000 17:10:33 GMT
From: aw99_1999@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8ojf5l$pua$1@nnrp1.deja.com>
References: <8ohghm$hls$1@nnrp1.deja.com> <8oifsr$rho$2@news1.xs4all.nl>

Hi Johan,

We would like to upgrade BSP and kernel. Thank you very much for your
help.

- ----Steve Cheng


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Tornado 1.0.1 problems going from Unix to NT?
Date: Wed, 30 Aug 2000 17:47:09 GMT
From: Jerry Hudgins 
Organization: Western Digital Design, Marin County, California
Message-ID: <39AD4896.B86A8A2C@e-farm.com>
References: <39AC26AC.63FF7E8C@lmco.com>

Eric Lesch wrote:
> 
> Our subcontractor developed the VxWorks kernel on a Unix platform, and
> we are developing applications for it on an NT platform.  All of our
> applications will be downloaded to the target.  Are there any issues
> related to the difference in platform?

I've observed a difference with the behavior of startup scripts.  The ld
command's normal form:

  ld < foo

worked fine with a UNIX host, but failed on NT.  I eventually found that
the terminating  was being included in the file name read from the
script and passed back to the ftp server.  The workaround was to use an
alternate form of the ld command in startup scripts:

  ld 1,0,"foo"

This works fine with both UNIX and NT hosts.

- -jch

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: *URGENT* Treatement of Packet Fragmentation
Date: Wed, 30 Aug 2000 14:08:57 -0500
From: "Tim Michals" 
Organization: Posted via Supernews, http://www.supernews.com
Message-ID: 
References: <8ogf92$l35$1@green.kreonet.re.kr>

The fragmentation happens in the IP protocol stack not in END driver.
Verify the sizes for smaller packets using ping i.e. ping  -l 78 see if
all of the lengths works.


"Á¤´ëÁø"  wrote in message
news:8ogf92$l35$1@green.kreonet.re.kr...
> Hello. We've developed PPC860-based ethernet switching device and
>
> wrote a new END driver for Ethernet CSMA/CD.
>
> We executed "PING" program with 1473 bytes at end station ( desktop PC ).
>
>   C:\ > ping 10.1.1.200 -l 1473
>
> The end station sends two packets  to the device: one is 1514 bytes, and
the
> other 60 bytes.
>
> Then the Ethernet END driver does not response to the end station, that
is,
> timed out.
>
> We pings 1472 bytes as following.
>
>   C:\ > ping 10.1.1.200 -l 1472
>
> The end station sends one packet to the device, where length of packet is
> 1514 bytes.
>
> At this time, the END driver responses to the end station so that the
> timeout does not occur any more.
>
> How can we treat the packet fragmentation?
>
>
>   Some of initialization codes are as following.
>
>    --------------------------------------------------
>
>   #define LAN_LOAD_STRING
> 0xFF000000:0x34:-1:0x0:-1:-1:0x40:0x30:0x0:0xff:2:0x4
>
>   END_TBL_ENTRY endDevTbl [] =
>   {
>         { 0, CPM_LOAD_FUNC, CPM_LOAD_STRING, 1, NULL,FALSE},
>         { 1, LAN_LOAD_FUNC,  LAN_LOAD_STRING, 1, NULL, FALSE},
>         { 0, END_TBL_END, NULL, 0, NULL, FALSE},
>   };
>
>   #define LAN_BUFSIZ      (ETHERMTU + ENET_HDR_REAL_SIZ + 6)
>
>   static M_CL_CONFIG lan2MclConfig =  /* network mbuf configuration table
*/
>   {
>        /*
>          no. mBlks        no. clBlks    memArea   memSize
>         --------          ---------    -------     -------
>        */
>     NUM_NET_MBLKS,    NUM_NET_CLBLKS,    NULL,
> (NUM_NET_MBLKS*(M_BLK_SZ+sizeof (long))+NUM_NET_CLBLKS*(CL_BLK_SZ +
> sizeof(long)))
>
>
>   static CL_DESC lan2ClDescTbl [] =  /* network cluster pool configuration
> table */
>   {
>     /*
>     clusterSize              num     memArea    memSize
>     -----------            ----     --------     -------
>     */
>     { LAN_BUFSIZ,    NUM_NET_CLBLKS,      NULL,         ((LAN_BUFSIZ +
> 8)*NUM_NET_CLBLKS) }
>   };
>
>   END_OBJ* EndDrv_Load ( char *initString )
>   {
>     .........
>
>        if  ( (END_OBJ_INIT (&gEndDrvTbl[cur_unit].end,
> (DEV_OBJ*)&(gEndDrvTbl[cur_unit]),
>                     gEndDrvTbl[cur_unit].szDevName,
> gEndDrvTbl[cur_unit].unit,
>                    &lanBridgeFuncTable, "EtherDrv" ) == ERROR) ||
>              (END_MIB_INIT
>  &(gEndDrvTbl[cur_unit].end),M2_ifType_ethernet_csmacd,
>                    gEndDrvTbl[cur_unit].enetAddr, 6, ETHERMTU, LAN_SPEED )
> == ERROR) )
>        {
>               printf("END Driver[%s] or MIB-II Structure Load Fail...\n",
> szDevName );
>               goto ERROR_EXIT;
>       }
>
>       END_OBJ_READY ( &(gEndDrvTbl[cur_unit].end),  IFF_UP | IFF_RUNNING
> |IFF_NOTRAILERS |
>                        IFF_BROADCAST | IFF_PROMISC | IFF_MULTICAST
> |IFF_ALLMULTI | IFF_SIMPLEX | IFF_POINTOPOINT );
>
>       .............
>
>  ERROR_EXIT:
>
>        ..............
>
>
>    --------------------------------------------------
>
> * ETHERMTU is 1500 as described in the VxWorks reference manual.
>
> How can we know your contact points to technical support ?
>
>
> Anyhelp will be appreciated.
>
>
> ---------------------------------
> Chaewoo Rha, Research Engineer,
> Research & Developement Centre
> TELEWARE Network Systems Co.
> cwrha@tware.co.kr
>
>
> ---------------------------------
> Chaewoo Rha, Research Engineer,
> Research & Developement Centre
> TELEWARE Network Systems Co.
>
> http://www.tware.co.kr
> http://www.nettaxi.com/citizens/lalabs
> lalabs@nettaxi.com
> cwrha@tware.co.kr
>
> Tel.             : 82-31-420-7947
> Internet FAX : 82-2-6276-8484(PhoneID=03434207947)
> C.P.            : 82-016-328-0275
> ---------------------------------
>
>
>



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: long file names using DOSFS
Date: Wed, 30 Aug 2000 22:17:28 GMT
From: Tony Dal Santo 
Message-ID: 
References: <8odm14$lqr$1@news.netvision.net.il> <39ABE08F.840491FD@bitband.com>

Leonid Rosenboim  wrote:

> for the long names options to take effect, you must reformat the disk drive
> using dosFsMkfs().
>
> Beware of this Long names option if you use TFFS.

Could you please elaborate on what issues there are between TFFS and long
filenames?  I am aware of the comments at
http://www.xs4all.nl/~borkhuis/vxworks about TFFS messing with DOS FS
internals, and ask WRS.  The support person claimed there were no issues
with long file names and TFFS.

Do you have an SPR or TSR reference?

Thanks,
Tony

---------------------------

Newsgroups: comp.os.vxworks
Subject: network time protocol
Date: Thu, 31 Aug 2000 00:23:28 GMT
From: "Dragan Ostojic" 
Organization: Sympatico
Message-ID: <4Ahr5.207632$Gh.4061788@news20.bellglobal.com>

Hi,

I need information regarding the use of network time protocol (NTP). Is it
available for VxWorks and is it feasible using NTP to synchronize the
timeofday clock on the target board?

Regards,
Dragan.



---------------------------

Newsgroups: comp.os.vxworks
Subject: Problems writing with dosFs 2.0
Date: Thu, 31 Aug 2000 01:57:01 GMT
From: Hosehead Jones 2nd-1/2 
Organization: United Garage Mechanics and Fruitbats, Ltd.
Message-ID: 

We are experiencing an intermittent (but still serious) problem with
dosFs 2.0.  Our system is running on a MVME2700, connected to an 18 GB
SCSI drive divided into 4 partitions of 4095 MB each (so that Windows
NT can also see the drive).  The drive has been formatted using
Windows NT using FAT16.

One task in our application handles archiving; at a 100 Hz rate, this
task checks for new data to archive to disk, then archives it if new
data is available.  Each block of new data is on the order of 10 KB.

We check the current filesize (using fstat) after each new write to
verify that the data was archived properly.  Most of the time, the
data is written to disk correctly.  However, once in a while (say once
every few hours) the filesize jumps by a huge margin (at least 1 GB,
and usually about 4 GB), then all subsequent writes to this file don't
increase the filesize at all.  Once this happens, the data in the file
is corrupted beyond repair, and when the file is closed it will either
be truncated by dosFs to a 0-byte file, or it will be enormous and the
checkdisk utilities will report that the filesize is larger than its
clusters and delete it.

We've pretty well tracked this down to being a bug in dosFs; we've
verified that we call write() with a valid pointer and data size.  Has
anyone else come across something similar?

- -----------------------
Mark Morrison
Coherent Technologies, Inc.
markm@ctilidar.com


---------------------------

Newsgroups: comp.os.vxworks
Subject: NFS woes
Date: Thu, 31 Aug 2000 02:13:35 GMT
From: Hosehead Jones 2nd-1/2 
Organization: United Garage Mechanics and Fruitbats, Ltd.
Message-ID: <4perqs4m4ob1mt7hr2b4sjvh4mni4rdvgf@4ax.com>

All righty, then - the background.  We're running VxWorks 5.4 (Tornado
II) with dosFs 2.0 on a MV2700 board; our kernel includes both NFS
client and server.  The MV2700 has a SCSI hard drive attached to it
that's been divided up into four partitions.  The host platform is
Windows NT.

What we want to do is use NFS to export the local partitions in such a
way that the Windows box will be able to see the local drive and copy
files back and forth between the two systems.  Unfortunately, we've
only had marginal success doing this.  The Windows NFS client software
will usually see the VxWorks machine, but either timeout trying to
read the disk or not see all the files.

Since we can sometimes see the VxWorks partitions, I know we're
exporting the volumes correctly  (I can't recall the exact call to do
this - something like nfsExport).  NFS appears to be working in
general, because we also open files on the Wintel box from VxWorks (so
WinNT is the server and VxWorks the client), and this works with no
problems.  

At this point, I don't know if the problem lies with WRS's NFS server
software, or with the WinNT setup.  (We've tried a few different
Windows NFS packages - FrontierTech, Omni Enterprise, Hummingbird -
with no improvement.)  Has anyone else gotten this to work reliably?
Is NFS even the answer?

- -----------------------------
Mark Morrison
Coherent Technologies, Inc.
markm@ctilidar.com


---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: To C or C++?
Date: Thu, 31 Aug 2000 03:54:59 GMT
From: tangoNO_SPAM@pulp-fiction.com (Max)
Organization: Posted via Supernews, http://www.supernews.com
Message-ID: <39add683.189543854@news.pacificnet.net>
References: <39A543EB.C9E19479@mail.com> <39a61a22.375055037@news.pacificnet.net> <39ABD9FF.AED8C0E8@fnc.fujitsu.com>

On Tue, 29 Aug 2000 10:42:56 -0500, Steven Deans
 wrote:

>Are the added libraries VxWorks libraries or g++ libraries.
>There are changes, that I cannot go into, in the next release that could
>make a difference in this area.
>
>Please reply as this will help an ongoing evaluation of OSes we are under
>going.
>
>Thanks,
>Steven
>
>
>Max wrote:
>
>> On Thu, 24 Aug 2000 11:48:59 -0400, Henry  wrote:
>>
>> >Hi, All,
>> >
>> >This is not my intention to start a language war. I want to konw should
>> >we C or C++, and if we use C++ under VxWork, do we pay penalty on code
>> >sizes, performace etc.?
>> >
>> >I personally prefer C++ for OO and its strong type checking, but I want
>> >your opinions.
>> >
>> >Thanks in advance.
>> >
>> >Henry
>> >
>>
>> It also depends on the amount of work to be done.
>> Smaller projects tend to be easier to write and maintain in C, as OOP
>> always requires some overhead in system layout and class preparation.
>> However, in larger projects this overhead pays back handsomely.
>>
>> Another issue of C++, especially for VxWorks, is the tremendous size
>> of the linked C++ support libraries. In my experience with ARM cores
>> the output binary with OO blew at least twice the size of a plain C
>> system, unless rigorously controlled. At the same time, the
>> performance degradation is much less noticeable.
>>
>> Max.
>>
>> Cambury Technologies, Inc
>> embedded wireless systems
>> http://www.cambury.com
>


I believe that those are mostly VxWorks C++ support modules.
My VxWorks configuration manager (Tornado 2,  ARM) indicates that
adding the following C++ support increases the binary size by:

Configuration #1:
C++ runtime support (requires module manager),
plus C++ STL,
plus basic C++ iostreams ~ 100k

Configuration #2:
The above plus C++ kernel classes (OS C++ wrappers) ~ 250k

Configuration #3:
Full support for C++ (whatever it means) ~350k

Configuration #4:
Full support for C++ plus Rogue Wave's tools.h - over 500k

Max.

Cambury Technologies, Inc
embedded wireless systems
http://www.cambury.com

---------------------------

Newsgroups: comp.os.vxworks
Subject: how to sychronize the target/host symbol?
Date: Thu, 31 Aug 2000 12:16:21 +0800
From: "jeasun.li" 
Organization: The news server of Changzhou TeleCom
Message-ID: <8okm6q$1usf$1@news.cz.js.cn>

hi, guys,
 can  anyone  tell me how to sychronize the
target/host symbol?
 now , when I downlowd a function use the Tornado, I cann't run it on the
console (or telnet ), while load a function by using the
start script at boot time, I cann't run it
at wind shell, what can I do? thank you very much!!
 I had set the target server  options of
sychronize the target/host symbol?
and no matter whether downlable or build in symbol table when build the
system image . the phenomenons are the same.

Best regards!!

   jeasun
          jianchen.li@wacos.com



---------------------------

Newsgroups: comp.os.vxworks
Subject: VxWorks on Cadre III board
Date: Wed, 30 Aug 2000 10:07:30 +0530
From: Shrishail 
Organization: Lawrence Berkeley National Laboratory
Message-ID: <8okpk9$ct2$1@overload.lbl.gov>


- ------ =_NextPart_000_01C0126A.1EF9E1E0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit

Hi List,
   Excuse me if this topic has been discussed earlier; I am new to the list.
   I am using a Motorola 5307 eval board from cardre III. I am facing problems
   in running the BSP. I am using the ROM resident version of the same. 
   Anyone out there who faced similar problem / who was able to use the BSP
   as it came with Tornado (CF5307)?
   
   Any comments / suggestions?
Thanks,
Shrishail
   
- ------ =_NextPart_000_01C0126A.1EF9E1E0
Content-Type: application/ms-tnef
Content-Transfer-Encoding: base64
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- ------ =_NextPart_000_01C0126A.1EF9E1E0--



---------------------------

Newsgroups: comp.os.vxworks
Subject: differences between vxworks and linux
Date: Thu, 31 Aug 2000 05:32:22 GMT
From: ehudfl@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8okql6$dep$1@nnrp1.deja.com>

Hello
I need information about the differences between those o.s. especially
in the times of doing tasks, performing operations, system calls etc...



Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: Vxsim File I/O with NT
Date: Thu, 31 Aug 2000 06:18:03 GMT
From: hruesing@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8oktah$gdc$1@nnrp1.deja.com>
References: <8o6kgi$fvu$1@nnrp1.deja.com>

In article <8o6kgi$fvu$1@nnrp1.deja.com>,
  jim.themis@genband.com wrote:

I have the same problems. There is an SPR 27918 that shall fix the
problems. But after installing it I have noticed no change. Also opendir
() etc. fails! I have an TSR open and let you know the results.

Harud Rüsing

> I reviewed the archives but I could find anyone with a similar
problem.
> I am running VxSim on an NT box.  I am trying to do file I/O via
fopen,
> fclose, fputs, etc.  I am able to read files, but I am unable to
create
> and write files.  I have tried adding "host:" to the beginning of my
> path.  I have enabled tsfs with r/w.  I get no error when I perform
> an fopen.  However, when I execute an fputs(), I get the following
> error:
> ntPassFsLib: unknown ioctl = 0x14
>
> I looked at errno and the value showed Unknown Request.
>
> Can anybody explain their set-up for reading and WRITE files under
> VxSim.
>
> Thanks,
> Jim Themis
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.
>


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: pipes and ioctl
Date: Thu, 31 Aug 2000 06:56:51 GMT
From: zschub@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <8okvjd$it5$1@nnrp1.deja.com>

hello,
	is it possible to set vxWorks pipes to be blocking or nonblocking using
ioctl calls? the pipeDrv API suggests only the following ioctl functions
are implemented for pipes:

      FIOGETNAME
      FIONREAD
      FIONMSGS
      FIOFLUSH

and mucking around with others seems not to work.

any ideas?

thanks,

  matt brennan


Sent via Deja.com http://www.deja.com/
Before you buy.

---------------------------

Newsgroups: comp.os.vxworks
Subject: VxWorks and dual processors
Date: Thu, 31 Aug 2000 09:00:08 GMT
From: "Den" 
Organization: Excite@Home - The Leader in Broadband
Message-ID: 

Hi there, I have a question regarding VxWorks.

1.  Does VxWorks support dual processors?  Do any other embedded OS's
support dual or multiple processors?

2.  Will a future version of VxWorks support 64-bit processors?  Such as the
new Intel Itanium chip?

Thanks in advance,

- -Dennis




---------------------------

Newsgroups: comp.os.vxworks
Subject: ? Loading a device driver
Date: Wed, 30 Aug 2000 22:17:42 +0100
From: "Merlino" 
Organization: TIN
Message-ID: <8ojq92$bp7$1@nslave1.tin.it>
Reply-To: "Merlino" 

Hi guys,

i have an x86  board with a P2 BSP vxWorks 5.4;
i would like to run the system this way:

1- loading vxworks image ( without any device driver )

2- scanning the PCI bus for known device

3- loading modules with the suitables device drivers  at first from
bootHost: via FTP; later i will load  driver.o from EIDE;
obviously by BSP let me access a MSDOS5 filesystem.

4- activating device drivers

 Is it possible ?
 How ?

 Thanks








---------------------------

Newsgroups: comp.os.vxworks
Subject: ? Loading a device driver
Date: Wed, 30 Aug 2000 22:17:42 +0100
From: "Merlino" 
Organization: TIN
Message-ID: <8ojqfv$jme$1@nslave2.tin.it>
Reply-To: "Merlino" 

Hi guys,

i have an x86  board with a P2 BSP vxWorks 5.4;
i would like to run the system this way:

1- loading vxworks image ( without any device driver )

2- scanning the PCI bus for known device

3- loading modules with the suitables device drivers  at first from
bootHost: via FTP; later i will load  driver.o from EIDE;
obviously by BSP let me access a MSDOS5 filesystem.

4- activating device drivers

 Is it possible ?
 How ?

 Thanks








---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: network time protocol
Date: Thu, 31 Aug 2000 12:31:36 +0200
From: Joachim Pluschke 
Organization: Stoye GmbH
Message-ID: <39AE3407.F820901A@stoye.de>
References: <4Ahr5.207632$Gh.4061788@news20.bellglobal.com>



Dragan Ostojic schrieb:

> Hi,
>
> I need information regarding the use of network time protocol (NTP). Is it
> available for VxWorks and is it feasible using NTP to synchronize the
> timeofday clock on the target board?
>
> Regards,
> Dragan.

Hi,

I don't clearly know about a support of NTP but to synchronize the timeofday
clock on a target there might be a solution for you using the SNTP which is a
overhead-reduced ( slightly reduced in milliseconds-precision, see >RFC 1769
(SNTP) March 1995< ) subset of the NTP. It uses the same timestamp!

To a VxWorks-Bootimage there can be included a  _small_  SNTP-client or
- -server both work fine in my application.
(using TornadoII / VxWorks Pentium BSP)

With the clients timestamp you might set the VxWorks soft-clock using the
seconds without the fractional part (at an appropiate time, i.e. the
fractional part is near a full second, if this is necessary):
    [clock_settime( CLOCK_REALTIME, (time_t*)&Timestamp.tv_sec) ;]
and/or the Target-Boards RTC after a conversion to broken-down time
(and additional timezone-corrections if needed):
    [sysRtcSet((struct tm*)&tmTimestamp);]

Take care of the different formats of the timestamps  (struct timespec; struct
tm) and  NTP/SNTP is based on UTC (Timezone GMT+zero).

To have a SNTP source/sink for developing you could use
>About Time 4.8< (Windows9x/NT) to find at "www.arachnoid.com".

If you are interested in issues using the SNTP-server have a look on this
newsgroup, some days earlier I anwered a question from "Jürgen Braun"
regarding to this.


Regards,
Joachim


- --
=====================================================
            Gebrüder Stoye GmbH
department: traffic control systems
      Name: Joachim Pluschke
       Adr: Longericher Strasse 177
      city: Cologne, Germany
       Zip: 50739
     phone: ++49/221/2616-534
    mailto: j.pluschke@stoye.de
       www: www.stoye.de
=====================================================



---------------------------

Newsgroups: comp.os.vxworks
Subject: Re: URGENT!Virtual Memory problem!
Date: Thu, 31 Aug 2000 18:56:05 +0800
From: "kvd2" 
Organization: DCI HiNet
Message-ID: <8ole3l$e40@netnews.hinet.net>
References: <8oidt0$fmd@netnews.hinet.net>

Dear All,

This problem is solved! Thanks!





kvd2  wrote in message news:8oidt0$fmd@netnews.hinet.net...
> Dear All,
>
> I've add
> {
>     (void *) DICE_BASE_ADRS,
>     (void *) DICE_BASE_ADRS,
>    DICE_SIZE,
>    VM_STATE_MASK_VALID | VM_STATE_MASK_WRITABLE | VM_STATE_MASK_CACHEABLE,
>     VM_STATE_VALID     | VM_STATE_WRITABLE      | VM_STATE_CACHEABLE_NOT
>     },
>
>     {
>     (void *) 0x10000000,
>     (void *) 0x10000000,
>    0x00001000,
>    VM_STATE_MASK_VALID | VM_STATE_MASK_WRITABLE | VM_STATE_MASK_CACHEABLE,
>     VM_STATE_VALID     | VM_STATE_WRITABLE      | VM_STATE_CACHEABLE_NOT
>     }
>
> to sysPhysMemDesc [] . The fisrt memory space is working well.
> However, the following error message is returned if I tried to dump the
> memory starting at
> 0x10000000, ie, the 2nd memory space.
>
> WTX Error 0x100ca (AGENT_MEM_ACCESS_ERROR)
> value = -1 = 0xffffffff
>
> What's wrong?
>
> kvd2
> p.s. I don't have vxVMI( the opt. component of mmu)
>
>
>
>



---------------------------

End of New-News digest
**********************