From vxwexplo-errs@csg.lbl.gov Mon Apr 1 04:03:08 2002 From: Vxworks Exploder Date: Mon Apr 1 04:03:10 PST 2002 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Mon Apr 1 04:03:05 PST 2002 Subject: Shared Memory Queue Receive Subject: Does "Building Envoy"mean to make a VxWorks core file in a bootable project? Subject: questions about building envoy? Subject: Mount/unmout drive Subject: bind error(0xc0002) Subject: How to build envoy in the Tornado GUI? Subject: What's the REPEATED START in I2C? ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Shared Memory Queue Receive Date: Sun, 31 Mar 2002 15:02:51 GMT From: "Paul Van Bellinghen" Organization: AT&T Worldnet Message-ID: I just set up 2 Force PCore boards (PPC based) in a VME chassis as a shared memory network. I tried using shared memory queues to send and receive messages between the boards. However, I get a data access exception in bcopy of msgQSmReceive. Any ideads why?I am sending/receiving a 4 byte pointer. --------------------------- Newsgroups: comp.protocols.snmp,comp.os.vxworks Subject: Does "Building Envoy"mean to make a VxWorks core file in a bootable project? Date: 31 Mar 2002 17:42:57 -0800 From: xiangxinyan@yahoo.com.cn (Alice) Organization: http://groups.google.com/ Message-ID: 11.2 Building Envoy These instructions assume that you are building VxWorks from your BSP directory. (1) Copy the include files. (2) Copy vxworks/h/snmpdLib.h into the main include directory ($(WIND_BASE)/target/include) (3) Replace the directory $(WIND_BASE)/target/include/snmp with the directory snark/vxworks/h/snmp. (4) Edit the sample configuration file described above and build. This creates the module snmp in the snark/vxworks/agent directory. (5) Go to the directory in which you build your BSP. (6) In the file config.h, find the line that reads: #include "configAll.h" and insert the following three lines before this line: #define INCLUDE_MIB2_ALL #define INCLUDE_SNMPD #define INCLUDE_SHOW_ROUTINES INCLUDE_SHOW_ROUTINES is only necessary if you are building a subagent. (7) Edit the Makefile in their BSP directory and include the path to their snmp module on the line MACH_EXTRA =. For example MACH_EXTRA = /home/users/username/prologue/snark/vxworks/agent/snmp (8) Rebuild VxWorks in the BSP directory. The SNMP agent should be built into it automatically. Can I build envoy in the Tornado GNU without doing the steps above? Thanks in advance!!!!!!! --------------------------- Newsgroups: comp.protocols.snmp,comp.os.vxworks Subject: questions about building envoy? Date: 31 Mar 2002 18:30:46 -0800 From: xiangxinyan@yahoo.com.cn (Alice) Organization: http://groups.google.com/ Message-ID: question1:Are the following 2 parts the 2 ways to build the Envoy SNMP server? question2:In part2,do I need to make a bootrom with the VxWorks as core file,and then boot target? question3:I see a snmputil.h, what is it for?Does it have any relation with snmputil? question4:Can I query the target agent with Mib Browser?> Part 1 Ill assume that $SNMP is the installation directoy of the epilogue stuff. Some of what is listed below is identical to section 5.2 of the getting started guide. Copy the snark/vxworks/cfgs/wind.cfg to the $SNMP dir. Edit it to make the two key changes: The first loadconfig statement in the file loadconfig 'vxppc' ;my system The SRCDIR setting: equate SRCDIR 'yourdirectory to the SNMP dir' ;would be $SNMP Now, make sure you have copied $SNMP/snark/vxworks/h/snmpdLib.h to $WIND_BASE/target/h. Also, replace all contents of $WIND_BASE/target/h/snmp with $SNMP/snark/vxworks/h/snmp. Run makmak in the $SNMP directory specifying this cfg file. In the SNMP directory, type in "make all". If you get compile errors, try defining the following environment variables. I set them for my NT box with Tornado under C:\Tornado: WIND_BASE C:\Tornado WIND_BASE_TYPE x86-win32 WIND_HOST_TYPE x86-win32 Set these variables on an NT using the properties menu by right-mouse clicking over your 'my computer'. On this dialog use the environment tab. Open up a new DOS shell and try the make all again. Part 2 Make sure you go to your BSP directory and edit config.h as per section 5.2 to add the defines BEFORE the configAll include. Fire up tornado Start a new vxWorks bootable image project based on this BSP. Open up the build properties and click on the MACROS tab and find the EXTRA_MODULES. In the value field, add the full path and name of the snmp library. For some reason, it has no extension. Rebuild all. --------------------------- Newsgroups: comp.os.vxworks Subject: Mount/unmout drive Date: Mon, 1 Apr 2002 14:43:30 +0900 From: "Seunghoon Lee\(À̽ÂÈÆ\)" Message-ID: Smart media(and compact flash) is mounted using dosFSLib. When unmount, memory leak is occured. (about 500kbytes) How can I fix this problem? My code is below... STATUS smartConn() { void *cbio; BLK_DEV *pBlkDev; if( (pBlkDev = smartDevCreate (512, BLOCKS_PER_TRACK, TOTAL_BLOCKS, BLOCK_OFFSET)) == (BLK_DEV*)NULL ) { printf("Error during smartDevCreate\n"); return (ERROR); } printf("smartDevCreate() at block 0 OK\n"); if ( (cbio = dcacheDevCreate(pBlkDev, NULL,DOSFS_EAR_DCACHE_SIZE, "s:")) == NULL ) { printf("Error during dcacheDevCreate: %x\n", errno); return (ERROR); } printf("dcacheDevCreate() recognized disk formatted in VxWorks. OK\n"); /* create partition manager with FDISK style decoder, up to 3 parts */ if( (smartCbio = dpartDevCreate( cbio, 3, usrFdiskPartRead )) == NULL ) { printf("Error during dpartDevCreate: 0x%x\n", errno); return (ERROR); } printf("dpartDevCreate() recognized disk formatted in VxWorks. OK\n"); /* create the 1st file system, 8 simult. open files, with No CHKDSK */ if (dosFsDevCreate("s:",dpartPartGet(smartCbio,0), DOSFS_EAR_NUM_FILE_DESCRIPTOR, 0 /*DOS_CHK_REPAIR | DOS_CHK_SILENT */) == ERROR) { printf("Error during dosFsDevCreate: %x\n", errno); return (ERROR); } printf("dosFsDevCreate() recognized disk formatted in VxWorks. OK\n"); } STATUS smartDisconn() { DEV_HDR *pDevHdr; /*dosFsVolUnmount*/ if( smartCbio == NULL ) return (OK); if( dosFsVolUnmount(smartCbio) == (ERROR) ) { printf("Error during dosFsVolUnmount: 0x%x\n", errno); /*return (ERROR);*/ } smartCbio = NULL; if( (pDevHdr = iosDevFind("s:", NULL)) == NULL ) { printf("Error during iosDevFind: 0x%x\n", errno); return (ERROR); } iosDevDelete(pDevHdr); return (OK); } --------------------------- Newsgroups: comp.os.vxworks Subject: bind error(0xc0002) Date: 31 Mar 2002 23:25:02 -0800 From: essenlee@hanmail.net (LEE MYUNG SUK) Organization: http://groups.google.com/ Message-ID: server : solaris 7.0 vxWorks 5.4 tornado2.0 board : MPC860 (unix)/etc/tornado/wind/target/config/ntp860/> make vxworks.st I maked vxworks.st, and downloaded(vxWorks.st with application). so, It isn't bind error. but, I builded vxWorks at tornado 2.0 tool, and downloaded (vxWorks.st with application). and It is bind error. (at winShell) -> i errono = 0xc0002 I think that the problem is component(at workspace). Do you know this problem? please answer me a piece of advice. then thank you very much. --------------------------- Newsgroups: comp.os.vxworks Subject: How to build envoy in the Tornado GUI? Date: 31 Mar 2002 23:25:19 -0800 From: xiangxinyan@yahoo.com.cn (Alice) Organization: http://groups.google.com/ Message-ID: Can I build envoy in the Tornado GUI without doing the following steps ? Thanks in advance!!!!!!! 11.2 Building Envoy These instructions assume that you are building VxWorks from your BSP directory. (1) Copy the include files. (2) Copy vxworks/h/snmpdLib.h into the main include directory ($(WIND_BASE)/target/include) (3) Replace the directory $(WIND_BASE)/target/include/snmp with the directory snark/vxworks/h/snmp. (4) Edit the sample configuration file described above and build. This creates the module snmp in the snark/vxworks/agent directory. (5) Go to the directory in which you build your BSP. (6) In the file config.h, find the line that reads: #include "configAll.h" and insert the following three lines before this line: #define INCLUDE_MIB2_ALL #define INCLUDE_SNMPD #define INCLUDE_SHOW_ROUTINES INCLUDE_SHOW_ROUTINES is only necessary if you are building a subagent. (7) Edit the Makefile in their BSP directory and include the path to their snmp module on the line MACH_EXTRA =. For example MACH_EXTRA = /home/users/username/prologue/snark/vxworks/agent/snmp (8) Rebuild VxWorks in the BSP directory. The SNMP agent should be built into it automatically. --------------------------- Newsgroups: comp.os.vxworks Subject: What's the REPEATED START in I2C? Date: 1 Apr 2002 03:36:43 -0800 From: 10month@hanmail.net (Changwon Lee) Organization: http://groups.google.com/ Message-ID: Hi everyone. I am finding the way to generate repeated start condition in i2c... plz see bit stream below... in binary... 11101000 0 00001101 0 0 11101001 0 01101100 1 0 in hexa... 0xE8 [ACK] 0x0D [ACK] [Sr] 0xE9 [ACK] 0x6C [NACK] [STOP]. as you can see... this device(0xE8) has sub address... after sending write request and sub address, I NEED TO MAKE REPEATED START condition.... thanks for any comments.... bye --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Mon Apr 1 05:48:01 2002 From: Richard Garrick Date: Mon Apr 1 05:48:03 PST 2002 Subject: BSP Developer Job Opening FirstWave Intelligent Optical Networking, Inc. is looking for a BSP developer to join the Core Embedded Software team in Holmdel, NJ. FirstWave ION is developing leading edge Terahertz Metro Core All Optical Switching systems. The successful candidate will be well versed in distributed control system hardware and software. We are looking for a team player and an excellent communicator. Candidates for this position will be able to interact with other engineers to translate ideas and concepts into software designs, and then document the same for reviews and inspections. Our team is chartered with the development of software drivers and controller software to support the Digital and Fiber Optic engineering teams. Therefore, familiarity with A/D and D/A converters and PowerPC processor based distributed controllers is desired for this position. We are also looking for some experience with TCP/IP client/server systems and familiarity with PCI bus is a plus. Requirements: · BSEE desired, MSEE preferred · C and asssembly language programming experience · Real-time system design experience with vxWorks · Debugging skill using scopes and ICE (emulator/JTAG/BDM) · Understanding of memory, flash, communications and I/O subsystems · Motorola PowerPC 8260 experience is desired · PCI Bus software experience is a plus Applicants must have the right to work in the USA. No responses from agencies please. Rick Garrick FirstWave ION, Inc From vxwexplo-errs@csg.lbl.gov Mon Apr 1 21:39:28 2002 From: sgoswami@hss.hns.com Date: Mon Apr 1 21:39:31 PST 2002 Subject: Query Regarding Performance Measurement on PPC6603e vxworks hi all, I need a way to measure time on a Power PC system running VxWorks version 5.3.1 with a granularity/resolution of atleast 1microsecond. Is there any on board timer which can be used. Currently the tick rate can be set (using sysClkrateSet) only to a maximum of 5000 ticks/sec which gives me a resolution of 0.2milliseconds. regards, saurabh "DISCLAIMER: This message is proprietary to Hughes Software Systems Limited (HSS) and is intended solely for the use of the individual to whom it is addressed. It may contain privileged or confidential information and should not be circulated or used for any purpose other than for what it is intended. If you have received this message in error, please notify the originator immediately. If you are not the intended recipient, you are notified that you are strictly prohibited from using, copying, altering, or disclosing the contents of this message. HSS accepts no responsibility for loss or damage arising from the use of the information transmitted by this email including damage from virus." From vxwexplo-errs@csg.lbl.gov Tue Apr 2 04:03:15 2002 From: Vxworks Exploder Date: Tue Apr 2 04:03:17 PST 2002 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Tue Apr 2 04:03:11 PST 2002 Subject: [question]on the MPC107... Subject: singelStep debugger and semTake Subject: Task Question: Hows a task find its name and id ... Subject: Re: What's the REPEATED START in I2C? Subject: ~~~ Systemprogrammierer/-designer sucht neues Projekt ~~~ Subject: Re: Task Question: Hows a task find its name and id ... Subject: Re: Task Question: Hows a task find its name and id ... Subject: Re: Task Question: Hows a task find its name and id ... Subject: Re: Task Question: Hows a task find its name and id ... Subject: Endian Problem with DMA addresses? IBM PPC 405GP Subject: Re: FCC Tx underrun on 8260 for single buffer implementation?? Subject: Re: Strange /CS timing with GPCM on 8260 (SETA=1) Subject: Re: GoAhead Webserver Porting to VxWorks Porblems! Subject: DO I need run makmak to building envoy in TMS? Subject: Re: how to create tasks with desired status Subject: Tornado Developement system Subject: Re: rlogin client puts stdout to shell session Subject: Interrupt Nesting Level ? Subject: Install Tornado 2.0 on Win 2000 error, Help!!! Subject: Re: Pipes and Message Queues....blocking? Subject: Re: DosFs2 performance Subject: Re: Task Question: Hows a task find its name and id ... Subject: To know Heap Size Subject: Re: Interrupt Nesting Level ? ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: [question]on the MPC107... Date: 1 Apr 2002 04:28:58 -0800 From: emycall@dreamx.net (Ahn Jung Hoon) Organization: http://groups.google.com/ Message-ID: <2b0ae335.0204010428.753c2651@posting.google.com> Hi, all. I try to implement I2c function to read from EEPROM or to write to EEPROM on the MPC750.. and I found source code(mpc107I2c.h) in C:\Tornado2\target\h\drv\multi. As below, there are three function declarations.. But I could not find the source code to implement function. I want to know that Where these is.. please, help me.. - ------------------------------------------------------------------------ /* function declarations */ IMPORT VOID mpc107I2cInit (); IMPORT INT32 mpc107i2cRead (UINT32 deviceAddress, UINT32 numBytes, char *pBuf); IMPORT INT32 mpc107i2cWrite (UINT32 deviceAddress, UINT32 numBytes, char *pBuf); - ------------------------------------------------------------------------- --------------------------- Newsgroups: comp.os.vxworks Subject: singelStep debugger and semTake Date: Mon, 1 Apr 2002 15:43:01 +0200 From: "Ofer Goren" Organization: Verio Message-ID: - -- - ------------------------ Win2k Sp2 Tornado 2.02 for PPC Diab 4.4b Envoy 9.2 - ------------------------ Hi. I'm using Diab compiler, and singelstep as my debugger. Everytime I reach a semTake function with WAIT_FOREVER, the debugger is halt, I'm loosing connection with the target, and I have to reset. So I cannot debug any function calling semTake with WAIT_FOREVER. Any ideas why is that? I know the debugger does not stop all the other tasks, or I would not be able to use the target shell. However, other task should give the semaphore.... Thanks, Ofer, Israel. --------------------------- Newsgroups: comp.os.vxworks Subject: Task Question: Hows a task find its name and id ... Date: Mon, 01 Apr 2002 14:27:51 GMT From: "Jim" Organization: Magma Communications Ltd. Message-ID: How can a task in vxWorks find its name and id ... ? Does this need to be done in the task ahead of time ... --------------------------- Newsgroups: comp.os.vxworks Subject: Re: What's the REPEATED START in I2C? Date: Mon, 1 Apr 2002 11:41:06 -0600 From: "G. Smith" Message-ID: <3ca89a61$0$3689$39cecf19@nnrp1.twtelecom.net> References: Go read the excellent response to the same question on comp.sys.powerpc.tech. "Changwon Lee" <10month@hanmail.net> wrote in message news:e76ae87c.0204010336.291d6850@posting.google.com... > Hi everyone. > I am finding the way to generate repeated start condition in i2c... > plz see bit stream below... > > in binary... > 11101000 0 00001101 0 0 11101001 0 01101100 1 0 > > in hexa... > 0xE8 [ACK] 0x0D [ACK] [Sr] 0xE9 [ACK] 0x6C [NACK] [STOP]. > > as you can see... this device(0xE8) has sub address... > after sending write request and sub address, I NEED TO MAKE REPEATED > START condition.... > > thanks for any comments.... bye --------------------------- Newsgroups: comp.os.vxworks Subject: ~~~ Systemprogrammierer/-designer sucht neues Projekt ~~~ Date: Mon, 1 Apr 2002 19:44:27 +0200 From: "M.Thoemel" Organization: T-Online Message-ID: Guten Tag, ich bin freier Systemprogrammerer und C-Designer mit einem sehr breiten Spektrum. Ab dem Juli/2002 stehe ich wieder zur Verfügung. Meine Schwerpunkte sind: - - C (aber auch andere Sprachen!) - - Realtimesysteme - - systemnahe Programmierung - - UML (RoseRT) - - spezielle Hardware-Schnittstellen - - Netzwerke - - Treiber - - Streamsdriver - - Grafische Datenauswertung/Darstellung - - Software-Messungen - - Gruppenleitung - - Projektierung - - Projektleitung etc... Sie können sich im Detail informieren unter: http://matthias.thoemel.bei.t-online.de Ich würde mich über einen Kontakt mit Ihnen freuen. Ihr Matthias Thömel (matthias.thoemel@t-online.de) --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Task Question: Hows a task find its name and id ... Date: Mon, 01 Apr 2002 19:13:32 GMT From: "Verizon" Message-ID: References: Task ID : taskid = taskIdSelf(); Name : taskName( taskid ); "Jim" wrote in message news:H7_p8.9581$f5.684560@news... > How can a task in vxWorks find its name and id ... ? > > Does this need to be done in the task ahead of time ... > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Task Question: Hows a task find its name and id ... Date: 1 Apr 2002 11:40:24 -0800 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0204011140.290fbc43@posting.google.com> References: Hello, A simple look in the reference manual entries for taskLib and taskInfo would reveal the following two functions: taskIdSelf() - get the task ID of a running task taskName() - get the name associated with a task ID HTH, John... "Jim" wrote in message news:... > How can a task in vxWorks find its name and id ... ? > > Does this need to be done in the task ahead of time ... --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Task Question: Hows a task find its name and id ... Date: Mon, 01 Apr 2002 20:04:47 GMT From: "Jim" Organization: Magma Communications Ltd. Message-ID: References: Thanks ... I missed them in the manual ... oops ... "Verizon" wrote in message news:wj2q8.3602$AE1.2275@nwrddc03.gnilink.net... > Task ID : > > taskid = taskIdSelf(); > > Name : > > taskName( taskid ); > > "Jim" wrote in message > news:H7_p8.9581$f5.684560@news... > > How can a task in vxWorks find its name and id ... ? > > > > Does this need to be done in the task ahead of time ... > > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Task Question: Hows a task find its name and id ... Date: Mon, 01 Apr 2002 20:05:20 GMT From: "Jim" Organization: Magma Communications Ltd. Message-ID: <443q8.9628$f5.692090@news> References: <488e459a.0204011140.290fbc43@posting.google.com> thnx "John" wrote in message news:488e459a.0204011140.290fbc43@posting.google.com... > Hello, > > A simple look in the reference manual entries for taskLib and taskInfo > would reveal the following two functions: > > taskIdSelf() - get the task ID of a running task > taskName() - get the name associated with a task ID > > HTH, > > John... > > "Jim" wrote in message news:... > > How can a task in vxWorks find its name and id ... ? > > > > Does this need to be done in the task ahead of time ... --------------------------- Newsgroups: comp.os.vxworks Subject: Endian Problem with DMA addresses? IBM PPC 405GP Date: 1 Apr 2002 15:04:19 -0800 From: arun4vxworks@indiatimes.com (Arun Prasad) Organization: http://groups.google.com/ Message-ID: Hi, I'm in the process of porting a Linux NIC driver (PCI based) into VxWorks. I have a doubt with DMA address accessing. I use sysPciInLong() and sysPciOutLong() functions to access PCI I/O region. This takes care of the Endian Issue by using " It uses the load word byte-reversed instruction" and "It uses the store word byte-reversed instruction" as I've seen the code in sysLib.s > I don't find any problem with this. However, I'm having some problem in accessing DMA addresses as per our Card's report. Before I pass the DMA address (gained by cacheDmaMalloc() ), I do the LOCAL2PCI_MEM() conversion so that the card will be able to access this address. (Special thanks to Michael Lawnick in explaining this patiently). But we are not doing any endian adjustments (conversion). My question is do we need to byte swap the DMA addresses before giving it to the PCI based NIC card? and do the vice versa after reading it from the card? The question raised because, our card is PCI based and our CPU is PPC (big endian). Anything else I'm missing? Any reply would be of great help. And a supplementary question. For the linux's virt_to_bus() call in our software we are calling LOCAL2PCI_MEM(). What about bus_to_virt() implemention? What should we do with that? Any Linux cum VxWorks Gurus' suggestions? Note : We are not making use of virtual memory in VxWorks. Regards Arun --------------------------- Newsgroups: comp.sys.powerpc.misc,comp.os.vxworks Subject: Re: FCC Tx underrun on 8260 for single buffer implementation?? Date: 1 Apr 2002 15:41:13 -0800 From: ryeung@earthlink.net (Raymond Yeung) Organization: http://groups.google.com/ Message-ID: References: A lot of good questions, Dan. "Dan Gold" wrote in message news:... > "Raymond Yeung" wrote in message > news:d83bb8e3.0203291031.6747fea8@posting.google.com... > > I'm using 0K26N (ancient revision) of 8260. All > > frames are transmitted from big buffers (enough > > to hold maximum frame size). > > Some more info would be nice. I'd infer you're running 100Base-T on an FCC, Not quite. On target side, it's actually auto-negotiating with a peer running 100BaseT full duplex. Off the bat, I know this would cause problem w.r.t. duplex. But I'm not expecting an underrun though. > correct ? Which FCC ? Are you running (since it's an A.1 part) on an old > VADS (or EST) board ? Your own design ? What is the CPU and CPM speed ? > VxWorks, or your own stuff ? It's our own board modified from EST board design. CPU is at 200MHz and CPM at 133MHz. We're using VxWorks (SBC8260 BSP). > > > As I pump up (bi-directional) data rate, this > > underrun thing shows up. I'm really not expecting > > it. My understanding is that Tx underrun would only > > happen if I use multiple buffers per frame (and set > > the "R" bit too early). Scanning through Motorola's > > errata list, I couldn't find any logged issue directly > > related to this. > > > > Has anyone experienced this? What did you find? > > The Tx-underrun is usually because the FCCs Tx FIFO requested more data and > the CPM couldn't get data out of your buffers fast enough to fill the FIFO. > Are your buffers on the 60x or local bus ? What else is going on in the The buffers are on 60x bus (running at 33MHz). > 8260 (i.e. other peripherals) when this is happening ? Do you have someone Not much else, other than AAL5 on FCC1. Fast Ethernet is run on FCC2. > hogging the bus (or the CPM) for a long period of time, so it couldn't > service the FCC request or couldn't access the bus ? Have you looked at > your Tx/Rx clocks (from the PHY) to make sure they look ok ? Since you're > running full duplex, are you sure the LPB and FDE bits are set in the 8260, > and your PHY is in FDX mode ? Good questions. I don't have an answer to them yet, except the last one; it's probably not in FDX mode, due to misconfiguration (as I described above). Are you sure you're setting the "R" bit at > the appropriate time even for 1 buffer ? (i.e. after the length, L-bit, are > set). This one could be tricky. It appears that the "R" bit may be set before the L-bit. I need to study the code in a bit more details. That's a good point, Dan. Does this happen on a single board with the A1 or other boards (with > the A1) too ? Does it make a difference if you use small or large frames ? > At what rate do you see this "under-run thing" start to happen ? > > Dan Gold gold@ensemble.com --------------------------- Newsgroups: comp.sys.powerpc.tech,comp.sys.powerpc.misc,comp.os.vxworks Subject: Re: Strange /CS timing with GPCM on 8260 (SETA=1) Date: 1 Apr 2002 15:45:47 -0800 From: ryeung@earthlink.net (Raymond Yeung) Organization: http://groups.google.com/ Message-ID: References: It appears that the /PGTA assertion time is too long; 8260 not only terminates the current access, but also terminates th next access (once /PGTA assertion is sampled on 8260 rising clock edge, bus access is terminated in 2 clock cycles). I believe that's why the next /CS only last for 1 clock cycle. The resolution we tested out is to ensure /PGTA be just 1 clock cycle. Also, lining up /PGTA to center around rising edge of 8260 clock may help. ryeung@earthlink.net (Raymond Yeung) wrote in message news:... > I've an external device interfacing with 8260 via a CPLD. > The device is under GPCM control with SETA=1 (i.e. external > access termination). A READY signal is hooked up to /PGTA > to trigger the access termination. > > I'm seeing some strange timing on the /CS. After the first > successful access, the device asserts its READY, which gets > translated by CPLD to /PGTA. The /CS is deasserted shortly > after and stays that way for a few clock cycles. In the > mean time, /PGTA becomes deasserted. So far so good. > > Then /CS becomes asserted again, and lasts only for about > 1 clock cycle and deasserted itself. Why? There's data to > write to the device (so I do expect /CS to be asserted). But > /PGTA is not asserted, so what terminates the access? Any > other register setting or ext. signals I need to check to get > more insight? > > Furthermore, 8260 manual says that /GTA needs to be one clock > cycle long (see sec. 10.5.2). Is it harmful (other than slowing > things down) to have a longer /PGTA (say, a few clock cycles)? > > Thanks, > Raymond --------------------------- Newsgroups: comp.os.vxworks Subject: Re: GoAhead Webserver Porting to VxWorks Porblems! Date: 1 Apr 2002 18:50:06 -0800 From: Xiaohua.zhang@163.com (xiaohua) Organization: http://groups.google.com/ Message-ID: <49321d54.0204011850.26c84db@posting.google.com> References: <49321d54.0203301824.3b3e3a6b@posting.google.com> <488e459a.0203310318.3fd8cc9d@posting.google.com> ***************************************************** 1. Download GoAhead Web Server from Http://www.goahead.com 2. Use Visual C++ to compile webcomp.c and result the webcomp.exe 3. Make a filelist file,and run webcomp prefix file>webrom.c And at the end of the webrom.c, it shows: --------------------------------------------------------------- websRomPageIndexType websRomPageIndex[] = { { T("/about.htm"), page_0, 4353 }, { T("/addgroup.asp"), page_1, 933 }, { T("/addlimit.asp"), page_2, 897 }, { T("/adduser.asp"), page_3, 1077 }, { T("/asp.asp"), page_4, 621 }, { T("/contents.asp"), page_5, 2146 }, { T("/delgroup.asp"), page_6, 609 }, { T("/dellimit.asp"), page_7, 627 }, { T("/deluser.asp"), page_8, 619 }, { T("/docs.htm"), page_9, 1800 }, { T("/dspuser.asp"), page_10, 623 }, { T("/FAQ.htm"), page_11, 9408 }, { T("/forms.asp"), page_12, 678 }, { T("/home.asp"), page_13, 753 }, { T("/loadcfg.asp"), page_14, 597 }, { T("/overview.htm"), page_15, 3037 }, { T("/savecfg.asp"), page_16, 587 }, { T("/small.htm"), page_17, 887 }, { T("/tests.htm"), page_18, 506 }, { T("/title.htm"), page_19, 321 }, { T("/treeapp.asp"), page_20, 1089 }, { T("/um.htm"), page_21, 4091 }, { 0, 0, 0 }, }; ----------------------------------------------------------------- 4. Modify the /VxWorks/main.c in the webvxmain() function. I want to delay more time,so I add some delay. ------------------------------------------------------------------ while (!finished) { for(i=0;i<50000;i++) if (socketReady(-1) || socketSelect(-1, 2000)) { socketProcess(-1); } websCgiCleanup(); emfSchedProcess(); } ------------------------------------------------------------------- 5. Build a bootable project with the BSP. Add the Load/Unload Module in the components. 6. Build a downloadable project. Add the \webserver\*.c, but exclude the webSSL.c and webcomp.c. Of course it include the webrom.c Add the "-DWEBS -DUEFM -DVXWORKS -DWEB_PAGE_ROM" in the compiler flag Then I build the project. OK 7. In the Shell, I type "sp(webvxmain)" It return OK. 8. Using IE: "http://192.168.0.131" 9. It show in the Hyperterminal: ---------------------------------------------- Data abort Exception address: 0x0005e6c0 Current Processor Status Register: 0x20000013 Task: 0x346e7c "s1u0" ---------------------------------------------- 10. Using the Debugger, I find the error address is "_cksum" 11. I enlarge the TCP/IP pramaters ,but it still not work. Need I modify the "directory" in the main.c? Somebody tell me the "redirction: function has some bug,anybody knows? ***************************************************** john_94501@yahoo.com (John) wrote in message news:<488e459a.0203310318.3fd8cc9d@posting.google.com>... > Hello, > > [steps deleted] > > > 5.Visit the target board using the IE http://192.168.0.131 > > > > 6.Error: > > > > Data abort > > Exception address: 0x0005e6c0 > > Current Processor Status Register: 0x20000013 > > Task: 0x2c7de0 "webd" > > > > ********************* > > *What can I do next?* > > ********************* > > Try starting the debugger and find out what caused the exception... > since you have source code you should be able to work this out > relatively easily using the debugger. > > HTH, > > John... > > Hint: If you want more specific answers you need to post much more > detailed information. --------------------------- Newsgroups: comp.os.vxworks,comp.protocols.snmp Subject: DO I need run makmak to building envoy in TMS? Date: 1 Apr 2002 19:06:55 -0800 From: xiangxinyan@yahoo.com.cn (Alice) Organization: http://groups.google.com/ Message-ID: The following is from "Tornado for Managed Switches 2.0.1 Programmer’s Guide" Since TMS users are not required to (and should not) run makmak, the install.h file is pre-built and provided in the distribution. This file should be edited for uninstalling any options or installing new options. I am using TMS,don't I need to run makmak to make makefiles for building envoy? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: how to create tasks with desired status Date: 1 Apr 2002 21:00:25 -0800 From: senthilb@comneti.com (senthil) Organization: http://groups.google.com/ Message-ID: <26ca8a8f.0204012100.1810306d@posting.google.com> References: <26ca8a8f.0203290212.2761c2ea@posting.google.com> Thanks Keith. I intially planned with taskinit and some API functions. But then i decided to implement it using semaphores . Ur suggestions too includes the semaphore logic as i decided. Can you explain me how to use the intlock and unlock tricks. I believe it will be useful for me sometimes . Thanks for ur feedback. Cheers senthil --------------------------- Newsgroups: comp.os.vxworks Subject: Tornado Developement system Date: Tue, 2 Apr 2002 11:13:11 +0530 From: "Girish pathak" Message-ID: Hi , I have compiled program a in Tornado development system,which includes semaphore.h and uses semPxLib, while compiling it compiles fine but when downloading it gives error like Errors while downloading C:/Tornado/target/proj/Project1/SIMNTgnu/Project1.out: _sem_wait _sem_destroy _sem_init _sem_post Does any body know how to link semPxLib file while downloading to simulator Girish pathak_girish@yahoo.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: rlogin client puts stdout to shell session Date: Tue, 02 Apr 2002 07:05:31 GMT From: pjt@iki.fi Organization: Embassy of Elbonia, 42 Kûrvi-Tasch Avenue, Szohod Message-ID: References: <3CA2FDDC.8010007@spamm.me.l8s.co.uk> Sender: ptaipale@dslnm.trs.ntc.nokia.com "Hwa Jin Bae" writes: > The rlogin child task simply reads from the socket and write to stdout. But why is stdout the global one, and not that of the calling task (when stdin is that of the calling task, and not the global one?) - -- Pekka Taipale -- pjt@iki.fi -- http://www.iki.fi/pjt/ --------------------------- Newsgroups: comp.os.vxworks Subject: Interrupt Nesting Level ? Date: 2 Apr 2002 00:19:18 -0800 From: fxcrazy@yahoo.com (Tom) Organization: http://groups.google.com/ Message-ID: <633e72d2.0204020019.66a72c64@posting.google.com> Is there any constraints like maximum number of Interrupts being nested ?? - - Tom --------------------------- Newsgroups: comp.os.vxworks Subject: Install Tornado 2.0 on Win 2000 error, Help!!! Date: Tue, 2 Apr 2002 16:57:11 -0800 From: "Qian Ling" Organization: Lucent Technologies, Columbus, Ohio Message-ID: Dear all: When I install the Tornado 2.0 on my win 2000 machine after I input the serial number and press "next>" button a dialog jump out with "Can't initiate SETUP: ZIP_ERROR. please correct this problem then run setup again." However my colleagues have installed it on same machine successfully. I try to find out the reason, and tried: (1) close and uninstall Mcafee Antivirus ==> Same Error! (2) Install the tornado patch ==> Same Error! (3) trace the TCL script, but the error is in a external function call (setupInit, not a TCL script) ==> No way to solve! (4) Uninstall Vs.NET beta and install ==> Same Error! Can anyone tell me what wrong with my machine and/or tornado? I do not want to reinstall win2000 manu thanks --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Pipes and Message Queues....blocking? Date: Tue, 2 Apr 2002 11:48:07 +0200 From: "Michael Lawnick" Organization: Buergernetz Dillingen Message-ID: References: <6755aee5.0203291212.6c96bd2b@posting.google.com> Reply-To: "Michael Lawnick" Sender: mlawnick@p5088eae9.dip.t-dialin.net Hi Whacked, you could code if ((ioctl (fd, FIONMSGS, &nMessages)!=ERROR) && (nMessages == 0)) write (fd, message,50); but keep track of the time-leak between test and message drop. A mutex could make it more secure. Else check setting timeout for the fd. I'm not really sure, but I seem to remember, that it is possible to set the fd to non-blocking like a socket, just give it a try (an tell your experience). on = TRUE; status = ioctl (sFd, FIONBIO, &on); HTH Michael "Whacked" schrieb im Newsbeitrag news:6755aee5.0203291212.6c96bd2b@posting.google.com... > Hi, > > Here is the situation. Ive set up a pipe for IPC that can accept only > one message. Something like the following: > > pipDevCreate ("/pipe/demo",1,50) > > Then I write a message to the pipe (pseudo code) > > fd = open("/pipe/demo", 0_RDRW,0) > write (fd, msg,50) > close (fd) > > Now, what happens if I immediately write to the pipe again before any > process reads? > > fd = open("/pipe/demo", 0_RDRW,0) > write (fd, message,50) > close (fd) > > It appears to me that I will be blocked in the write call since there > is already a message in the pipe. This is not what I want to happen. I > would like for the 2nd message to simply be dropped. > > The pipe facility is based on message queues. If I was setting up a > queue, I could pass the it a NO_WAIT timeout parameter. So, I could: > > Q_ID = msgQCreate (1,50,MSG_Q_FIFO) > msgQSend (Q_ID, msg, 50, NO_WAIT, MSG_PRI_NORMAL) > > and then send the second message > > msgQSend (Q_ID, msg, 50, NO_WAIT, MSG_PRI_NORMAL) > > This message would get dropped on the floor if the first one had not > been read from the queue. > > I need to iplement this behavior with pipes. Since the pipes are based > on message queues in VxWorks, can I do it?? > > thanks, > > Whacked --------------------------- Newsgroups: comp.os.vxworks Subject: Re: DosFs2 performance Date: Tue, 2 Apr 2002 12:04:37 +0200 From: "Michael Lawnick" Organization: Buergernetz Dillingen Message-ID: References: <3ca612ce.1867405@netnews.attbi.com> Reply-To: "Michael Lawnick" Sender: mlawnick@p5088eae9.dip.t-dialin.net Hi David, have not worke with AE yet, but with dosFs2. First access to the device will do a rough check and fills buffers. So the first question is: Do you notice the delay on every open() or on the first open() (since reboot) associated to this device ? Michael "David H. Olson" schrieb im Newsbeitrag news:3ca612ce.1867405@netnews.attbi.com... > I'm posting this from home, so don't have all the details, but here > goes anyway. We're running under the latest non-AE vxWorks on a > PPC750 and using the DosFs2 file system over a flash drive. We've > noticed that opening a file takes a surprisingly long time (~100 ms). > The transfer speed, once the file is open, seems in line w/ what we > would predict given the specs of the various parts & busses involved. > So should we be surprised at how long it takes to open a file? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Task Question: Hows a task find its name and id ... Date: Tue, 02 Apr 2002 10:14:47 GMT From: pjt@iki.fi Organization: Embassy of Elbonia, 42 Kûrvi-Tasch Avenue, Szohod Message-ID: References: <488e459a.0204011140.290fbc43@posting.google.com> Sender: ptaipale@dslnm.trs.ntc.nokia.com john_94501@yahoo.com (John) writes: > A simple look in the reference manual entries for taskLib and taskInfo > would reveal the following two functions: Right, Jim's question was too easy, but how about this one: how do I find out the IP address at the other end of a connection which has been created with telnetCallAdd()? It seems that the task created by telnetCallAdd cannot use getpeername() to the fd found in its own TCB. I.e, how to read into a variable the information displayed by inetstatShow? I can think of ways to do it (set stdout to a file, call inetstatShow, look for the correct line in the file) but is there a cleaner way? - -- Pekka Taipale -- pjt@iki.fi -- http://www.iki.fi/pjt/ --------------------------- Newsgroups: comp.os.vxworks Subject: To know Heap Size Date: 2 Apr 2002 03:15:40 -0800 From: mvsenthilkumar@yahoo.com (M V Senthil Kumar) Organization: http://groups.google.com/ Message-ID: Is there any function which gives the Heap size available?I want to use this function/code before allocating memory sothat if the available heap size is less than the size of memory tobe allocated then malloc neednot be called. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Interrupt Nesting Level ? Date: Tue, 02 Apr 2002 06:32:04 -0500 From: joe durusau Organization: BELLSOUTH.net Message-ID: <3CA996B4.831EAD36@bellsouth.net> References: <633e72d2.0204020019.66a72c64@posting.google.com> Tom wrote: > Is there any constraints like maximum number of Interrupts being nested ?? > > - Tom It would be limited by available stack size and the size of the work queue. Having said that, I can't imagine having more than 1 or two nested if the ISRs are written decently and the system doesn't have some device sending entirely too many interrupts. Speaking only for myself, Joe Durusau --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Wed Apr 3 04:03:16 2002 From: Vxworks Exploder Date: Wed Apr 3 04:03:18 PST 2002 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Wed Apr 3 04:03:11 PST 2002 Subject: Re: Tornado Developement system Subject: Re: To know Heap Size Subject: Re: GoAhead Webserver Porting to VxWorks Porblems! Subject: Re: Pipes and Message Queues....blocking? Subject: Is there a vxworks kernel consisteny checker ? Subject: Lex & Yacc tools with VxWorks Subject: Re: How to Communicate between Two vxsim Targets in Windows 2000 Subject: Re: Lex & Yacc tools with VxWorks Subject: Re: How to Communicate between Two vxsim Targets in Windows 2000 Subject: Re: Establishing ppp Connection between vxWorks and windows2000 Subject: Distruction of static objects in function scope Subject: Startup scripts Subject: arpAdd Question Subject: Unsupported code? Subject: What is the relationship between .c file generatd by "mibcomp -stub"and by idbGen? Subject: Re: Startup scripts Subject: Re: DosFs2 performance Subject: Re: Lex & Yacc tools with VxWorks Subject: Re: DosFs2 performance Subject: console in TMS can't display in the right way Subject: Re: Establishing ppp Connection between vxWorks and windows2000 Subject: eieio and sync in sysInLong() and sysOutLong() Subject: Re: Startup scripts Subject: Re: Startup scripts Subject: Re: Lex & Yacc tools with VxWorks Subject: Re: Lex & Yacc tools with VxWorks Subject: Re: Distruction of static objects in function scope Subject: Re: Install Tornado 2.0 on Win 2000 error, Help!!! ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado Developement system Date: Tue, 2 Apr 2002 14:08:47 +0200 From: "Werner Schiendl" Message-ID: <3ca99f48@brateggebdc5.br-automation.co.at> References: Hi, you need to build a custom simulator image and include POSIX semaphore support. Alternatively, you can also switch to using VxWorks native semaphores (see semLib.h). To build your custom simulator, you can use the provided project simpc_vx. Alternatively, you can create a new bootable VxWorks project based on that project. Then add the above mentioned POSIX semaphores to the VxWorks configuration. The result of the build is a file VxWorks.exe that you then select as your custom simulator when launching the simulator. hth Werner "Girish pathak" wrote in message news:a8bghn$ph5qe$1@ID-120142.news.dfncis.de... > Hi , > I have compiled program a in Tornado development system,which includes > semaphore.h and uses semPxLib, > while compiling it compiles fine but when downloading it gives error like > > Errors while downloading > C:/Tornado/target/proj/Project1/SIMNTgnu/Project1.out: > _sem_wait > _sem_destroy > _sem_init > _sem_post > > Does any body know how to link semPxLib file while downloading to simulator > > Girish > pathak_girish@yahoo.com > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: To know Heap Size Date: Tue, 2 Apr 2002 14:11:31 +0200 From: "Werner Schiendl" Message-ID: <3ca99feb$1@brateggebdc5.br-automation.co.at> References: Hi, you problably want use memFindMax() It gives you the largest available memory block (in the system memory partition) hth Werner "M V Senthil Kumar" wrote in message news:c5908dd.0204020315.424faadb@posting.google.com... > Is there any function which gives the Heap size available?I want to > use this function/code before allocating memory sothat if the > available heap size is less than the size of memory tobe allocated > then malloc neednot be called. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: GoAhead Webserver Porting to VxWorks Porblems! Date: Tue, 2 Apr 2002 14:18:41 +0200 From: "Werner Schiendl" Message-ID: <3ca9a199$1@brateggebdc5.br-automation.co.at> References: <49321d54.0203301824.3b3e3a6b@posting.google.com> <488e459a.0203310318.3fd8cc9d@posting.google.com> <49321d54.0204011850.26c84db@posting.google.com> Hi, Did you check that you have enough stack? The error seems to result from access to non-existent memory (or otherwise not allowed) - maybe one of the ARM gurus here can tell you what exactly wents wrong. Did you check if it works without the modification? Did you check if it works with a 'proper' URL (one _including_ the filename to serve out)? hth Werner "xiaohua" wrote in message news:49321d54.0204011850.26c84db@posting.google.com... > ***************************************************** > 1. Download GoAhead Web Server from Http://www.goahead.com > 2. Use Visual C++ to compile webcomp.c and result the webcomp.exe > 3. Make a filelist file,and run webcomp prefix file>webrom.c > And at the end of the webrom.c, it shows: > --------------------------------------------------------------- > websRomPageIndexType websRomPageIndex[] = { > { T("/about.htm"), page_0, 4353 }, > { T("/addgroup.asp"), page_1, 933 }, > { T("/addlimit.asp"), page_2, 897 }, > { T("/adduser.asp"), page_3, 1077 }, > { T("/asp.asp"), page_4, 621 }, > { T("/contents.asp"), page_5, 2146 }, > { T("/delgroup.asp"), page_6, 609 }, > { T("/dellimit.asp"), page_7, 627 }, > { T("/deluser.asp"), page_8, 619 }, > { T("/docs.htm"), page_9, 1800 }, > { T("/dspuser.asp"), page_10, 623 }, > { T("/FAQ.htm"), page_11, 9408 }, > { T("/forms.asp"), page_12, 678 }, > { T("/home.asp"), page_13, 753 }, > { T("/loadcfg.asp"), page_14, 597 }, > { T("/overview.htm"), page_15, 3037 }, > { T("/savecfg.asp"), page_16, 587 }, > { T("/small.htm"), page_17, 887 }, > { T("/tests.htm"), page_18, 506 }, > { T("/title.htm"), page_19, 321 }, > { T("/treeapp.asp"), page_20, 1089 }, > { T("/um.htm"), page_21, 4091 }, > { 0, 0, 0 }, > }; > ----------------------------------------------------------------- > 4. Modify the /VxWorks/main.c in the webvxmain() function. > I want to delay more time,so I add some delay. > ------------------------------------------------------------------ > while (!finished) { > for(i=0;i<50000;i++) > if (socketReady(-1) || socketSelect(-1, 2000)) { > socketProcess(-1); > } > websCgiCleanup(); > emfSchedProcess(); > } > ------------------------------------------------------------------- > 5. Build a bootable project with the BSP. > Add the Load/Unload Module in the components. > 6. Build a downloadable project. > Add the \webserver\*.c, but exclude the webSSL.c and webcomp.c. > Of course it include the webrom.c > Add the "-DWEBS -DUEFM -DVXWORKS -DWEB_PAGE_ROM" in the compiler flag > Then I build the project. OK > 7. In the Shell, I type "sp(webvxmain)" > It return OK. > 8. Using IE: > "http://192.168.0.131" > 9. It show in the Hyperterminal: > ---------------------------------------------- > Data abort > > Exception address: 0x0005e6c0 > > Current Processor Status Register: 0x20000013 > > Task: 0x346e7c "s1u0" > ---------------------------------------------- > 10. Using the Debugger, I find the error address is "_cksum" > 11. I enlarge the TCP/IP pramaters ,but it still not work. > > Need I modify the "directory" in the main.c? > Somebody tell me the "redirction: function has some bug,anybody knows? > > > > > > > ***************************************************** > john_94501@yahoo.com (John) wrote in message news:<488e459a.0203310318.3fd8cc9d@posting.google.com>... > > Hello, > > > > [steps deleted] > > > > > 5.Visit the target board using the IE http://192.168.0.131 > > > > > > 6.Error: > > > > > > Data abort > > > Exception address: 0x0005e6c0 > > > Current Processor Status Register: 0x20000013 > > > Task: 0x2c7de0 "webd" > > > > > > ********************* > > > *What can I do next?* > > > ********************* > > > > Try starting the debugger and find out what caused the exception... > > since you have source code you should be able to work this out > > relatively easily using the debugger. > > > > HTH, > > > > John... > > > > Hint: If you want more specific answers you need to post much more > > detailed information. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Pipes and Message Queues....blocking? Date: 2 Apr 2002 05:09:31 -0800 From: kenb88@hotmail.com (Whacked) Organization: http://groups.google.com/ Message-ID: <6755aee5.0204020509.3507c244@posting.google.com> References: <6755aee5.0203291212.6c96bd2b@posting.google.com> "Alex Pogostin" wrote in message news:... > If the pipe is being written to from an ISR, the pipe driver will force the > NO_WAIT onto the lower level msgQSend. > > Alex Pogostin > Silicon Drivers > > "Whacked" wrote in message > news:6755aee5.0203291212.6c96bd2b@posting.google.com... > > Hi, > > > > Here is the situation. Ive set up a pipe for IPC that can accept only > > one message. Something like the following: > > > > pipDevCreate ("/pipe/demo",1,50) > > > > Then I write a message to the pipe (pseudo code) > > > > fd = open("/pipe/demo", 0_RDRW,0) > > write (fd, msg,50) > > close (fd) > > > > Now, what happens if I immediately write to the pipe again before any > > process reads? > > > > fd = open("/pipe/demo", 0_RDRW,0) > > write (fd, message,50) > > close (fd) > > > > It appears to me that I will be blocked in the write call since there > > is already a message in the pipe. This is not what I want to happen. I > > would like for the 2nd message to simply be dropped. > > > > The pipe facility is based on message queues. If I was setting up a > > queue, I could pass the it a NO_WAIT timeout parameter. So, I could: > > > > Q_ID = msgQCreate (1,50,MSG_Q_FIFO) > > msgQSend (Q_ID, msg, 50, NO_WAIT, MSG_PRI_NORMAL) > > > > and then send the second message > > > > msgQSend (Q_ID, msg, 50, NO_WAIT, MSG_PRI_NORMAL) > > > > This message would get dropped on the floor if the first one had not > > been read from the queue. > > > > I need to iplement this behavior with pipes. Since the pipes are based > > on message queues in VxWorks, can I do it?? > > > > thanks, > > > > Whacked Alex, What if it does not?? Anyway to force it? thanks, --------------------------- Newsgroups: comp.os.vxworks Subject: Is there a vxworks kernel consisteny checker ? Date: 2 Apr 2002 05:49:25 -0800 From: r.stelzer@stoye.de (Rainer Stelzer) Organization: http://groups.google.com/ Message-ID: <78a5d372.0204020549.106e50f2@posting.google.com> Hi vxworkers ! We do have a strange bug in our system that destroys a semaphore control structur by writing garbage into one two byte long location within this structur. When seeking the bug we asked us if there is a method to check if further queue's, semaphore, fd's s and other os resources are been hit by the bug. In other words: - - Is there a vxworks kernel consisteny checker ? or - - has vxworks a chance to check the consisteny of kernel system while running (rather than crashing the system when operating with damaged control structures ) thanx in advance rainer --------------------------- Newsgroups: comp.os.vxworks Subject: Lex & Yacc tools with VxWorks Date: 2 Apr 2002 09:14:42 -0800 From: venu_padakanti@yahoo.com (Venu Madhav Padakanti) Organization: http://groups.google.com/ Message-ID: <55aab3a9.0204020914.6d475cbd@posting.google.com> I have been told that Wind River Systems does not support the lex/yacc or the flex/bison GNU tools. Has any one ported them to VxWorks AE environment successfully? I went to the gnu.org site but and could find the tar file for bison but not for flex. Any help is appreciated Thanks ..venu --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to Communicate between Two vxsim Targets in Windows 2000 Date: Tue, 2 Apr 2002 18:50:44 +0100 From: "Dimitrios Siganos" Organization: Newport Networks Ltd. Message-ID: <1017769845.756174@azores.network-i.net> References: <538e14e.0203272137.50aedafc@posting.google.com> The fact that the ULIP driver is shown as disabled is not the problem. i have multiple simulators running in my system (win2000) and the driver is always shown as disabled. A mistake that I often make is to forget to add a "catch-all" route back to the ULIP driver. It sounds like you are having the same problem. The following command does that: routeAdd("0", "90.0.0.254"); assuming that the ulip addr is 90.0.0.254 Dimitris "Mohit" wrote in message news:538e14e.0203272137.50aedafc@posting.google.com... > I have followed the procedure mentione in the tutorial to create > multiple targets and ping them. I also Configured ULIP driver and > enabled IpForwarding. But still when i try to ping Target i.e. > 90.0.0.1 or 90.0.0.2 either from Tronado shell or from ms-Dos prompt > there is no reply.Target servers for both the targets are running > fine. I have observed one problem in ULIP driver, the connection for > ULIP driver is shown disabled when i try to enable it i get the Error > message "Connection Failed". Is it related to the problem i mentioned > above. If yes then could anybody plese tell how to enable the ULIP > connection.If no then please tell me what additional steps should be > taken to solve the problem. > > Thanks, --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Lex & Yacc tools with VxWorks Date: Tue, 2 Apr 2002 13:11:40 -0500 From: "doug dotson" Organization: Verio Message-ID: References: <55aab3a9.0204020914.6d475cbd@posting.google.com> Do you mean port the tools or modifying them so that the generated code runs under vxWorks? I've used code generated by flex and bison under vxWorks. It's not reentrant though so you can only have a single parser/scanner running on a box. Not sure about AE in that regard. doug "Venu Madhav Padakanti" wrote in message news:55aab3a9.0204020914.6d475cbd@posting.google.com... > I have been told that Wind River Systems does not support the lex/yacc > or the flex/bison GNU tools. > > Has any one ported them to VxWorks AE environment successfully? > > I went to the gnu.org site but and could find the tar file for > bison but not for flex. > > Any help is appreciated > > Thanks > ..venu --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to Communicate between Two vxsim Targets in Windows 2000 Date: Tue, 2 Apr 2002 13:14:35 -0500 From: "doug dotson" Organization: Verio Message-ID: <2wmq8.5636$T_.116458@iad-read.news.verio.net> References: <538e14e.0203272137.50aedafc@posting.google.com> <1017769845.756174@azores.network-i.net> I put this kludge into usrAppInit(): int procNum; char myIP [16]; procNum = sysProcNumGet (); sprintf (myIP, "90.0.0.%d", procNum+1); routeAdd ("0.0.0.0", myIP); doug "Dimitrios Siganos" wrote in message news:1017769845.756174@azores.network-i.net... > The fact that the ULIP driver is shown as disabled is not the problem. i > have multiple simulators running in my system (win2000) and the driver is > always shown as disabled. > > A mistake that I often make is to forget to add a "catch-all" route back to > the ULIP driver. It sounds like you are having the same problem. The > following command does that: > > routeAdd("0", "90.0.0.254"); > > assuming that the ulip addr is 90.0.0.254 > > Dimitris > > "Mohit" wrote in message > news:538e14e.0203272137.50aedafc@posting.google.com... > > I have followed the procedure mentione in the tutorial to create > > multiple targets and ping them. I also Configured ULIP driver and > > enabled IpForwarding. But still when i try to ping Target i.e. > > 90.0.0.1 or 90.0.0.2 either from Tronado shell or from ms-Dos prompt > > there is no reply.Target servers for both the targets are running > > fine. I have observed one problem in ULIP driver, the connection for > > ULIP driver is shown disabled when i try to enable it i get the Error > > message "Connection Failed". Is it related to the problem i mentioned > > above. If yes then could anybody plese tell how to enable the ULIP > > connection.If no then please tell me what additional steps should be > > taken to solve the problem. > > > > Thanks, > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Establishing ppp Connection between vxWorks and windows2000 Date: Tue, 02 Apr 2002 15:36:30 -0500 From: david lindauer Message-ID: <3CAA164E.220F9B7F@notifier-is.net> References: Michael Lawnick wrote: > "vinod" schrieb im Newsbeitrag > news:cb7218f2.0203280344.5fefa2f3@posting.google.com... > > hi > > Could someone please tell me the configurations required on the > > windows 2000 side in order to establish a ppp connection between > > vxworks and windows 2000 > > > > Thanx in advance, > > vinod > In the Control Panel: 'Network and Dial-up Connections'->'Make New > Connection' > In the Connection Wizard select the suitable mechansim and go through > installation process. I found it straight forward. > > Michael that works... but I couldn't get direct connection to work so I had to resort to using modems and a PBX simulator for testing... David --------------------------- Newsgroups: comp.os.vxworks Subject: Distruction of static objects in function scope Date: Tue, 02 Apr 2002 22:52:54 GMT From: "Matt Schuckmann" Message-ID: This may be some what of a C++ question but I believe that the standard leaves this type of stuff up to the particular implimentation of the language, so here goes. The question is, in VxWorks when do static objects that are declared/defined within a functions scope get destroyed? I would think that they would get destroyed when the module gets unloaded but from my testing that does not appear to be the case. Can anyone give me some clues on this? I can provide my test code for an example if would help. Thanks Matt Schuckmann matthew_schuckmann@amat.com --------------------------- Newsgroups: comp.os.vxworks Subject: Startup scripts Date: 2 Apr 2002 15:36:20 -0800 From: arcv0@lycos.com (Roger) Organization: http://groups.google.com/ Message-ID: <45dd6f79.0204021536.c706a48@posting.google.com> I am using Tornado 2.0. I am having problems using startup script. The startup scripts runs fine if I boot my target machine over the ethernet. But when it boots over the flash file system(Disk-on-chip) I get the messages "unable to open the startup script". It looks like the startup scripts are called before the Disk-on-chip is configured. I called the "usrTffsConfig(0,0,"/DOC") in the usrAppInit.c. Please let me know if anybody has encountered this problem before. Thanks in advance for the help Here are the boot parameters. boot device : tffs=0,0 unit number : 0 processor number : 0 file name : /tffs/vxworks flags (f) : 0x0 startup script (s) : /tffs0/startup.txt Attaching to TFFS... done. Loading /tffs/vxworks...1089412 Starting at 0x100000... Failed to attach to device tffs=0,0Attaching interface lo0...0x3fffdf8 (tRootTask): Can't attach unknown device tffs=0,0 (unit 0). done Adding 3401 symbols for standalone. Unable to open startup script /tffs0/startup.txt - -> --------------------------- Newsgroups: comp.os.vxworks Subject: arpAdd Question Date: Tue, 02 Apr 2002 23:52:42 GMT From: "spaceman" Organization: Prodigy Internet http://www.prodigy.com Message-ID: If the arp entry is already in the arp table, it looks like a call to arpAdd will fail. Has anyone seen a similar behavior? Thanks, Jim F. --------------------------- Newsgroups: comp.os.vxworks Subject: Unsupported code? Date: 02 Apr 2002 20:21:00 -0500 From: Bill Pringlemeir Organization: Factory of the mind Message-ID: Sender: bpringlemeir@DeadDuck My target/src/unsupported has the following directories, bedk, bootp2.1, config, cslip-2.7, demo, dhcp-1.3beta, emul frc, h, mkits, ppp-2.1.2, rpc4.0, rt11,src Are the directories like ppp-2.1.2 and dhcp-1.3beta the same as the source that was compiled for ppp and dhcp support in vxWorks? Or are these utilities for Linux, Solaris, and BSD users that were put on my NT machine by the install program? I don't see any relavent build files, so I am guessing that they are *nix source. How is one suppose to know what in "target/src/unsupported" is for the target? Regards, Bill Pringlemeir. - -- Zumfing abounds! vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" --------------------------- Newsgroups: comp.os.vxworks Subject: What is the relationship between .c file generatd by "mibcomp -stub"and by idbGen? Date: 2 Apr 2002 17:36:03 -0800 From: xiangxinyan@yahoo.com.cn (Alice) Organization: http://groups.google.com/ Message-ID: What is the relationship between .c file generatd by "mibcomp -stub" and the C file generated by idbGen in TMS?How do they communicate with each other? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Startup scripts Date: Wed, 3 Apr 2002 08:42:31 +0200 From: "Michael Lawnick" Organization: Buergernetz Dillingen Message-ID: References: <45dd6f79.0204021536.c706a48@posting.google.com> Reply-To: "Michael Lawnick" Sender: mlawnick@pd9e658e9.dip.t-dialin.net Hi Roger, you hit the point yourself: > I called the "usrTffsConfig(0,0,"/DOC") in the usrAppInit.c Of whatever reason, the startup-script is executed _before_ usrAppInit ! I solved the problem by modifying the *.cdf in its tffs section: Component INCLUDE_TFFS { INIT_RTN tffsDrv (); usrTffsConfig (0,0,"/tffs0"); } HTH Michael "Roger" schrieb im Newsbeitrag news:45dd6f79.0204021536.c706a48@posting.google.com... > I am using Tornado 2.0. I am having problems using startup script. The > startup scripts runs fine if I boot my target machine over the > ethernet. But when it boots over the flash file system(Disk-on-chip) I > get the messages "unable to open the startup script". It looks like > the startup scripts are called before the Disk-on-chip is configured. > I called the "usrTffsConfig(0,0,"/DOC") in the usrAppInit.c. Please > let me know if anybody has encountered this problem before. > > Thanks in advance for the help > > > Here are the boot parameters. > > boot device : tffs=0,0 > unit number : 0 > processor number : 0 > file name : /tffs/vxworks > flags (f) : 0x0 > startup script (s) : /tffs0/startup.txt > > Attaching to TFFS... done. > Loading /tffs/vxworks...1089412 > Starting at 0x100000... > > Failed to attach to device tffs=0,0Attaching interface lo0...0x3fffdf8 > (tRootTask): Can't attach unknown device tffs=0,0 (unit 0). > done > > Adding 3401 symbols for standalone. > Unable to open startup script /tffs0/startup.txt > -> --------------------------- Newsgroups: comp.os.vxworks Subject: Re: DosFs2 performance Date: Wed, 03 Apr 2002 07:01:02 GMT From: dcolson@ieee.org (David H. Olson) Organization: Oresis Communications, Inc. Message-ID: <3caaa780.48021461@netnews.attbi.com> References: <3ca612ce.1867405@netnews.attbi.com> Reply-To: dolson@oresis.com The lengthy open() occurs everytime. We're paranoid about messing up the entire file system if there's a power fail while a file's open (it's happened before!) so we tend to open a file, do a little bit w/ it and then close it. It wasn't a big deal before, but it's now become an issue so I'm starting to do a little research. Oh, and to be clear, we're running vxWorks 5.4, not AE. On Tue, 2 Apr 2002 12:04:37 +0200, "Michael Lawnick" wrote: >Hi David, > >have not worke with AE yet, but with dosFs2. >First access to the device will do a rough check and fills buffers. So the >first question is: >Do you notice the delay on every open() or on the first open() (since >reboot) associated to this device ? > >Michael >"David H. Olson" schrieb im Newsbeitrag >news:3ca612ce.1867405@netnews.attbi.com... >> I'm posting this from home, so don't have all the details, but here >> goes anyway. We're running under the latest non-AE vxWorks on a >> PPC750 and using the DosFs2 file system over a flash drive. We've >> noticed that opening a file takes a surprisingly long time (~100 ms). >> The transfer speed, once the file is open, seems in line w/ what we >> would predict given the specs of the various parts & busses involved. >> So should we be surprised at how long it takes to open a file? > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Lex & Yacc tools with VxWorks Date: 3 Apr 2002 07:20:07 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: <55aab3a9.0204020914.6d475cbd@posting.google.com> venu_padakanti@yahoo.com (Venu Madhav Padakanti) wrote: > I have been told that Wind River Systems does not support the lex/yacc > or the flex/bison GNU tools. > > Has any one ported them to VxWorks AE environment successfully? > > I went to the gnu.org site but and could find the tar file for > bison but not for flex. > > Any help is appreciated Take a look at the cygwin distribution (http://www.cygwin.com). This is a Unix emulator running under Windows (Tornado is using this to run the GNU tools). Groeten, Johan - -- o o o o o o o . . . _____________________________ o _____ || Johan Borkhuis | .][__n_n_|DD[ ====_____ | borkhuis@agere.com | >(________|__|_[_________]_|__________________________| _/oo OOOOO oo` ooo ooo 'o!o!o o!o!o` === VxWorks FAQ: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html === --------------------------- Newsgroups: comp.os.vxworks Subject: Re: DosFs2 performance Date: 3 Apr 2002 07:24:36 GMT From: Johan Borkhuis Organization: Agere Systems Message-ID: References: <3ca612ce.1867405@netnews.attbi.com> <3caaa780.48021461@netnews.attbi.com> dcolson@ieee.org (David H. Olson) wrote: > The lengthy open() occurs everytime. We're paranoid about messing up > the entire file system if there's a power fail while a file's open > (it's happened before!) so we tend to open a file, do a little bit w/ > it and then close it. It wasn't a big deal before, but it's now > become an issue so I'm starting to do a little research. What about flushing after every file access instead of opening/closing the file? Groeten, Johan - -- o o o o o o o . . . _____________________________ o _____ || Johan Borkhuis | .][__n_n_|DD[ ====_____ | borkhuis@agere.com | >(________|__|_[_________]_|__________________________| _/oo OOOOO oo` ooo ooo 'o!o!o o!o!o` === VxWorks FAQ: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html === --------------------------- Newsgroups: comp.os.vxworks,comp.protocols.snmp Subject: console in TMS can't display in the right way Date: 2 Apr 2002 23:36:22 -0800 From: xiangxinyan@yahoo.com.cn (Alice) Organization: http://groups.google.com/ Message-ID: My console in TMS can't display in the right way,it always displays something that no one knows. What is the trouble? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Establishing ppp Connection between vxWorks and windows2000 Date: Wed, 3 Apr 2002 11:52:50 +0200 From: "Michael Lawnick" Organization: Buergernetz Dillingen Message-ID: References: <3CAA164E.220F9B7F@notifier-is.net> Reply-To: "Michael Lawnick" Sender: mlawnick@pd9e658e9.dip.t-dialin.net Hi David, I found a strange behaviour with direct connection concerning CTS. Check your cabling, that it only builds a 3-wire connection, otherwise the login process might be stopped by W2K because of modem hang up. Michael "david lindauer" schrieb im Newsbeitrag news:3CAA164E.220F9B7F@notifier-is.net... > > > Michael Lawnick wrote: > > > "vinod" schrieb im Newsbeitrag > > news:cb7218f2.0203280344.5fefa2f3@posting.google.com... > > > hi > > > Could someone please tell me the configurations required on the > > > windows 2000 side in order to establish a ppp connection between > > > vxworks and windows 2000 > > > > > > Thanx in advance, > > > vinod > > In the Control Panel: 'Network and Dial-up Connections'->'Make New > > Connection' > > In the Connection Wizard select the suitable mechansim and go through > > installation process. I found it straight forward. > > > > Michael > > that works... but I couldn't get direct connection to work so I had to > resort to using modems and a PBX simulator for testing... > > David > --------------------------- Newsgroups: comp.os.vxworks Subject: eieio and sync in sysInLong() and sysOutLong() Date: Wed, 03 Apr 2002 11:18:29 GMT From: brede.thurmann@kongsberg-simrad.com (BT) Message-ID: <3caae4d0.98474498@news.eunet.no> I have a computer based on a PPC755 with an MPC107 connected to a PCI bus. I use the functions sysInLong() and sysoutLong() to access devices on the PCI bus. The functions from vxWorks disasemble like this: sysInLong: lwbrx r4,r0,r3 eieio sync or r3,r4,r4 bclr 20,0 sysOutLong: stwbrx r4,r0,r3 eieio sync bclr 20,0 Is it realy necessary to use both 'eieio' and 'sync'? I does not get any problems if I try to remove the 'eieio' but I think it is a bit difficult to fully understand when to use these functions. BT --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Startup scripts Date: 3 Apr 2002 13:23:05 +0100 From: Reinhold Wirth Organization: WEB.DE Message-ID: <3caae619@netnews.web.de> References: <45dd6f79.0204021536.c706a48@posting.google.com> Hi Roger, arcv0@lycos.com (Roger) wrote: >boot device : tffs=0,0 >unit number : 0 >Loading /tffs/vxworks...1089412 >Starting at 0x100000... > >Failed to attach to device tffs=0,0Attaching interface lo0...0x3fffdf8 >(tRootTask): Can't attach unknown device tffs=0,0 (unit 0). >done tffs=0,0 is used as a net device. Try to change "config/comps/src/net/usrNetBoot.c to: void usrNetDevNameGet (void) ...... if ( (strncmp (sysBootParams.bootDev, "scsi", 4) == 0) || (strncmp (sysBootParams.bootDev, "ide", 3) == 0) || (strncmp (sysBootParams.bootDev, "ata", 3) == 0) || (strncmp (sysBootParams.bootDev, "tffs", 4) == 0) || (strncmp (sysBootParams.bootDev, "fd", 2) == 0)) .... - -- __________________________________________________________ News suchen, lesen, schreiben mit http://newsgroups.web.de --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Startup scripts Date: 3 Apr 2002 13:23:18 +0100 From: Reinhold Wirth Organization: WEB.DE Message-ID: <3caae626$2@netnews.web.de> References: <45dd6f79.0204021536.c706a48@posting.google.com> Hi Roger, arcv0@lycos.com (Roger) wrote: >boot device : tffs=0,0 >unit number : 0 >Loading /tffs/vxworks...1089412 >Starting at 0x100000... > >Failed to attach to device tffs=0,0Attaching interface lo0...0x3fffdf8 >(tRootTask): Can't attach unknown device tffs=0,0 (unit 0). >done tffs=0,0 is used as a net device. Try to change "config/comps/src/net/usrNetBoot.c to: void usrNetDevNameGet (void) ...... if ( (strncmp (sysBootParams.bootDev, "scsi", 4) == 0) || (strncmp (sysBootParams.bootDev, "ide", 3) == 0) || (strncmp (sysBootParams.bootDev, "ata", 3) == 0) || (strncmp (sysBootParams.bootDev, "tffs", 4) == 0) || (strncmp (sysBootParams.bootDev, "fd", 2) == 0)) .... - -- __________________________________________________________ News suchen, lesen, schreiben mit http://newsgroups.web.de --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Lex & Yacc tools with VxWorks Date: Wed, 03 Apr 2002 11:54:14 GMT From: pjt@iki.fi Organization: Embassy of Elbonia, 42 Kûrvi-Tasch Avenue, Szohod Message-ID: References: <55aab3a9.0204020914.6d475cbd@posting.google.com> Sender: ptaipale@dslnm.trs.ntc.nokia.com venu_padakanti@yahoo.com (Venu Madhav Padakanti) writes: > Has any one ported them to VxWorks AE environment successfully? They don't need any porting. Code generated by yacc and bison, or bison and flex, cross-compiles and runs just fine for VxWorks. I've been trying both Berkeley yacc and bison 1.27 and flex 2.5.4. (Or did you mean running yacc and lex natively in VxWorks? Before we go to that, do you have a native C compiler that runs in VxWorks?) - -- Pekka Taipale -- pjt@iki.fi -- http://www.iki.fi/pjt/ --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Lex & Yacc tools with VxWorks Date: Wed, 03 Apr 2002 11:57:04 GMT From: pjt@iki.fi Organization: Embassy of Elbonia, 42 Kûrvi-Tasch Avenue, Szohod Message-ID: References: <55aab3a9.0204020914.6d475cbd@posting.google.com> Sender: ptaipale@dslnm.trs.ntc.nokia.com "doug dotson" writes: > I've used code generated by flex and > bison under vxWorks. It's not reentrant though so you can only have > a single parser/scanner running on a box. Actually, bison can generate re-entrant parsers. The input interface is just slightly different. Look for the documentation of %pure_parser declaration. (I haven't tried it myself, but I do trust the documentation). - -- Pekka Taipale -- pjt@iki.fi -- http://www.iki.fi/pjt/ --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Distruction of static objects in function scope Date: Wed, 3 Apr 2002 10:04:18 +0100 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: "Matt Schuckmann" wrote in message news:aDqq8.22783$9d3.19723@nwrddc01.gnilink.net... > This may be some what of a C++ question but I believe that the standard > leaves this type of stuff up to the particular implimentation of the > language, so here goes. > > The question is, in VxWorks when do static objects that are declared/defined > within a functions scope get destroyed? I would think that they would get > destroyed when the module gets unloaded but from my testing that does not > appear to be the case. I would have thought so too. Were they constructed when the module was loaded? Then they ought to be destructed when it gets unloaded. However this will only happen if you've properly MUNCHed the object code during compilation. You should also look at the documentation for cplusLib, in particular the routines cplusXtorSet, cplusCtors and cplusDtors. DaveK - -- moderator of alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! Proud Member of the Exclusive "I have been plonked by Davee because he thinks I'm interesting" List Member # Master of Many Meowing Minions Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage above and beyond the call of hilarity. > > Can anyone give me some clues on this? > > I can provide my test code for an example if would help. > > Thanks > Matt Schuckmann > matthew_schuckmann@amat.com > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Install Tornado 2.0 on Win 2000 error, Help!!! Date: Wed, 3 Apr 2002 10:00:31 +0100 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: "Qian Ling" wrote in message news:a8brst$7gf@nntpa.cb.lucent.com... > Dear all: > > When I install the Tornado 2.0 on my win 2000 machine > after I input the serial number and press "next>" button > a dialog jump out with > "Can't initiate SETUP: ZIP_ERROR. please correct this problem > then run setup again." > > However my colleagues have installed it on same machine > successfully. I try to find out the reason, and tried: > > (1) close and uninstall Mcafee Antivirus ==> Same Error! > (2) Install the tornado patch ==> Same Error! > (3) trace the TCL script, but the error is in a external > function call (setupInit, not a TCL script) ==> No way to solve! > (4) Uninstall Vs.NET beta and install ==> Same Error! > > Can anyone tell me what wrong with my machine and/or tornado? > I do not want to reinstall win2000 One possibility is that you have a stray copy of the cygwin dll somewhere in your $PATH setting and it is clashing with the version used by the tornado tools. Search your system for files called "cygwin1.dll" and if you find any either move them from their directories or remove those directories from the PATH, then try the installation again. DaveK - -- moderator of alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow Burn your ID card! http://www.optional-identity.org.uk/ Help support the campaign, copy this into your .sig! Proud Member of the Exclusive "I have been plonked by Davee because he thinks I'm interesting" List Member # Master of Many Meowing Minions Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage above and beyond the call of hilarity. --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Thu Apr 4 04:03:21 2002 From: Vxworks Exploder Date: Thu Apr 4 04:03:23 PST 2002 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Thu Apr 4 04:03:15 PST 2002 Subject: Re: arpAdd Question Subject: Is windML necessary? Subject: Re: Lex & Yacc tools with VxWorks Subject: The drivers for windML is for what purpose? Subject: Padding/alignment of structure Subject: Re: The drivers for windML is for what purpose? Subject: Re: Padding/alignment of structure Subject: GDB - How to debug large object file Subject: Re: Padding/alignment of structure Subject: Re: eieio and sync in sysInLong() and sysOutLong() Subject: Re: Padding/alignment of structure Subject: Re: Distruction of static objects in function scope Subject: Re: Startup scripts Subject: m()  modify memory Subject: Re: Padding/alignment of structure Subject: Re: Padding/alignment of structure Subject: BSP UART config. Subject: What are the common customization to default Exception Handler Subject: Re: BSP UART config. Subject: TCL/Gopher code? pPendQ is for what? Subject: Re: TCL/Gopher code? pPendQ is for what? Subject: Re: TCL/Gopher code? pPendQ is for what? Subject: Re: Does "Building Envoy"mean to make a VxWorks core file in a bootable Subject: Re: questions about building envoy? Subject: Re: DO I need run makmak to building envoy in TMS? Subject: Re: What is the relationship between .c file generatd by "mibcomp Subject: Re: Padding/alignment of structure Subject: Re: DosFs2 performance Subject: Re: Padding/alignment of structure Subject: Problem Subject: Re: stack checker Subject: Error : block too big Subject: Class C subnet mask on class B network Subject: Re: Lex & Yacc tools with VxWorks Subject: Re: Problem Subject: how to change radius client ip? (windNet Radius Client) Subject: Clearmake and Tornado II generated makefiles Subject: Re: m()  modify memory Subject: Re: Problem Subject: Re: Tornado Developement system Subject: Re: Class C subnet mask on class B network Subject: Re: DosFs2 performance ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: arpAdd Question Date: 3 Apr 2002 04:50:13 -0800 From: kannan@comneti.com (kannan) Organization: http://groups.google.com/ Message-ID: <2ede7703.0204030450.43c1126b@posting.google.com> References: The strange thing i found out is arpShow works in the shell.But arpAdd doesn't works in the shell.I tried checking the target\lib\objPPC860gnuvx there the arpLib.o is missing . But arpShow and arpAdd works in the simulator i tried checking the target\lib\objSIMNTgnuvx which contains arpLib.o. solution : In target\lib\libPPC860gnuvx.a i have extracted the ".o" files using the "ar" tool and manually load the arpLib.o into my target board through the ld command.Now i was able to do arpAdd in my target board. Thanks and Regards, Kannan --------------------------- Newsgroups: comp.os.vxworks Subject: Is windML necessary? Date: 3 Apr 2002 05:47:14 -0800 From: liandongzhang@hotmail.com (liandongzhang) Organization: http://groups.google.com/ Message-ID: <83f4aa95.0204030547.76dc4597@posting.google.com> I want to do some graphic manipulation, i have a pc/104 and lcd , and i want to disply certain kind of graphic file.Then whether the optional package windML is necessary or not? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Lex & Yacc tools with VxWorks Date: Wed, 3 Apr 2002 09:38:04 -0500 From: "doug dotson" Organization: Verio Message-ID: <4rEq8.5710$T_.119725@iad-read.news.verio.net> References: <55aab3a9.0204020914.6d475cbd@posting.google.com> You did have to port libflex didn't you. Or do the more recent versions of flex not require a library? doug wrote in message news:m3wuvpdkr3.fsf@dslnm.trs.ntc.nokia.com... > venu_padakanti@yahoo.com (Venu Madhav Padakanti) writes: > > Has any one ported them to VxWorks AE environment successfully? > > They don't need any porting. Code generated by yacc and bison, or > bison and flex, cross-compiles and runs just fine for VxWorks. > I've been trying both Berkeley yacc and bison 1.27 and flex 2.5.4. > > (Or did you mean running yacc and lex natively in VxWorks? Before we go to > that, do you have a native C compiler that runs in VxWorks?) > > -- > Pekka Taipale -- pjt@iki.fi -- http://www.iki.fi/pjt/ --------------------------- Newsgroups: comp.os.vxworks Subject: The drivers for windML is for what purpose? Date: 3 Apr 2002 07:16:16 -0800 From: liandongzhang@hotmail.com (liandongzhang) Organization: http://groups.google.com/ Message-ID: <83f4aa95.0204030716.324ac219@posting.google.com> May my question is sounded funny:( I think windML is on the upper level than VxWorks,and sorts of drivers is of parts of VxWorks,so driver should independed of WindML. I don't know clearly, above is just my idea in the first thought. Another question: what is the relation between UGL and windML and Zinc? I think Zinc is the upper level application of WindML,but where is UGL? If i want to display some graphic file other than GIF, should i must programme the functions that used to manipulate the specific graphic file? And what i must buy?WIndML UGL Zinc?? Thanks for any reply and it will be appreciated --------------------------- Newsgroups: comp.os.vxworks,comp.sys.arm Subject: Padding/alignment of structure Date: Wed, 3 Apr 2002 10:48:43 -0500 From: "Harjinder Chayra" Organization: Storm Internet Services Message-ID: Sender: terayon@gw-terayon.storm.ca I have a structure which represents interrrupt controller HW registers; it is being padded and is giving me grief because when I access the registers I get another registers values! How can I get rid of the padding? I have tried to use __attribute__ ((aligned(1)) on the structure elements but no joy. I am using the GNU toolkit for ARM9 Thanks in advance Harjinder --------------------------- Newsgroups: comp.os.vxworks Subject: Re: The drivers for windML is for what purpose? Date: 03 Apr 2002 11:34:12 -0500 From: Bill Pringlemeir Organization: Factory of the mind Message-ID: References: <83f4aa95.0204030716.324ac219@posting.google.com> Sender: bpringlemeir@DeadDuck >>>>> "ldz" == liandongzhang writes: ldz> May my question is sounded funny:( I think windML is [snip] ldz> Zinc is the upper level application of WindML,but where is UGL? Look at this picture. Study it carefully. "http://www.windriver.com/products/assets/images/zarch.gif" WindML == UGL 2.0 The old name of WindML was UGL. However, it seems pretty close to "UGLy" so they changed the name. There is only Zinc and WindML. The directory names in WindML 2.0 are still "UGL". [Note to WRS, please keep it this way!] ldz> If i want to display some graphic file other than GIF, should i ldz> must programme the functions that used to manipulate the ldz> specific graphic file? And what i must buy? WIndML UGL Zinc?? WindML understands JPG files as far as I understand. You must convert the images to JPG or re-write the uglxxxToDDBFromFile() function and replace references to uglJpgToDDBFromFile(). hth, Bill Pringlemeir. - -- ARPA rail gun counter terrorism condor Skipjack Comirex LABLINK propaganda sniper sweep AIMSX Area 51 Armani Fortezza secure vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" --------------------------- Newsgroups: comp.os.vxworks,comp.sys.arm Subject: Re: Padding/alignment of structure Date: Wed, 3 Apr 2002 11:50:56 -0500 From: "doug dotson" Organization: Verio Message-ID: References: Does your compiler support the -fpack-struct option? Otherwise you may have to do it the old-fashioned way. doug "Harjinder Chayra" wrote in message news:a8f8ad$pik$1@news.storm.ca... > I have a structure which represents interrrupt controller HW registers; it > is being padded and is giving me grief because when I access the registers > I get another registers values! > > How can I get rid of the padding? > > I have tried to use __attribute__ ((aligned(1)) on the structure elements > but no joy. > > I am using the GNU toolkit for ARM9 > > Thanks in advance > Harjinder > > > --------------------------- Newsgroups: comp.os.vxworks Subject: GDB - How to debug large object file Date: Wed, 3 Apr 2002 12:00:34 -0500 From: "dp" Organization: Nortel Message-ID: I have a very large object file to debug (>500MB). Obviously, it is too big for my board (Motorola mcpn750 card, 128 MB RAM). I used Tonado 2.1/vxWorks 5.4.2. What options should I use to debug it? - strip it using stripppc -S and load it to the target but the how can I syncronize it with its symbols? - Can't load all of it Do you know the best way of debug a very large file using GDB? Any help or suggestions should be very appreciated. Dat --------------------------- Newsgroups: comp.os.vxworks,comp.sys.arm Subject: Re: Padding/alignment of structure Date: Wed, 03 Apr 2002 12:03:05 -0800 From: Joe Durusau Organization: Lockheed Martin Corporation Message-ID: <3CAB5FF9.B71A2360@lmco.com> References: Try __attribute__ ((packed)) but note that it must appear on each and every line that you don't want padded. Also, make sure that you have the right sizes for the regs. On many platforms, you can NEVER have a misaligned data reference. I don't know about yours, but it's worth a second trip through the manual for your board. Speaking only for myself, Joe Durusau Harjinder Chayra wrote: > > I have a structure which represents interrrupt controller HW registers; it > is being padded and is giving me grief because when I access the registers > I get another registers values! > > How can I get rid of the padding? > > I have tried to use __attribute__ ((aligned(1)) on the structure elements > but no joy. > > I am using the GNU toolkit for ARM9 > > Thanks in advance > Harjinder --------------------------- Newsgroups: comp.os.vxworks Subject: Re: eieio and sync in sysInLong() and sysOutLong() Date: 3 Apr 2002 09:51:41 -0800 From: vloscomp@yahoo.com (Vinh) Organization: http://groups.google.com/ Message-ID: References: <3caae4d0.98474498@news.eunet.no> Hi, if you used 'sync' then 'eieio' is not needed. eieio enforced synchronization for data in pipes. sync enforced synchronization for data and instruction. This mean that sync takes longer. 1. You have to determined if you need data or instruction synchronization, or both. 2. You have to determined if the MPC107 needs synchronization for PCI bus. Furthermore, if the MPC107 supports eieio and sync instruction, then the buffering on the MPC107 will be synchronized with the CPU. Vinh Lam brede.thurmann@kongsberg-simrad.com (BT) wrote in message news:<3caae4d0.98474498@news.eunet.no>... > I have a computer based on a PPC755 with an MPC107 > connected to a PCI bus. I use the functions sysInLong() > and sysoutLong() to access devices on the PCI bus. > The functions from vxWorks disasemble like this: > > sysInLong: > lwbrx r4,r0,r3 > eieio > sync > or r3,r4,r4 > bclr 20,0 > > sysOutLong: > stwbrx r4,r0,r3 > eieio > sync > bclr 20,0 > > Is it realy necessary to use both 'eieio' and 'sync'? > I does not get any problems if I try to remove the 'eieio' > but I think it is a bit difficult to fully understand > when to use these functions. > > BT --------------------------- Newsgroups: comp.os.vxworks,comp.sys.arm Subject: Re: Padding/alignment of structure Date: 3 Apr 2002 11:05:01 -0800 From: yarramsetty@hotmail.com (Satya Yarramsetty) Organization: http://groups.google.com/ Message-ID: References: "Harjinder Chayra" wrote in message news:... > I have a structure which represents interrrupt controller HW registers; it > is being padded and is giving me grief because when I access the registers > I get another registers values! > > How can I get rid of the padding? > > I have tried to use __attribute__ ((aligned(1)) on the structure elements > but no joy. > > I am using the GNU toolkit for ARM9 > > Thanks in advance > Harjinder Try this.... typedef struct { UINT8 x; UINT8 y; UINT8 z; ... } __attribute__((packed)) XYZ_T; Hope this wud solve your problem SY --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Distruction of static objects in function scope Date: Wed, 03 Apr 2002 19:31:27 GMT From: "Matt Schuckmann" Message-ID: References: As a matter of fact they were not created when the module was loaded. They were not constructed until the function was called. Here is my test code: class cStaticTest { public: cStaticTest( char* str ) : m_str( str ) { cout << "Constructor for cStaticTest( " << m_str << ");" << endl; } ~cStaticTest() { cout << "Destructor for cStaticTest( " << m_str << ");" << endl; } char* m_str; }; static cStaticTest test0( "test0: Global static" ); static Ptr test1( new cStaticTest("test1: Global static smart pointer") ); extern "C" void staticTestFunct() { static cStaticTest test2( "test2: Function static" ); static Ptr test3( new cStaticTest("test3: Function static smart pointer") ); } When the module loads the following text is printed: Constructor for cStaticTest( test0: Global static); Constructor for cStaticTest( test1: Global static smart pointer); When I run staticTestFunct() the following text is printed: Constructor for cStaticTest( test2: Function static); Constructor for cStaticTest( test3: Function static smart pointer); When the module is unloaded the following text is printed: Destructor for cStaticTest( test1: Global static smart pointer); Destructor for cStaticTest( test0: Global static); Apparently trest2, and test3 are never destroyed. "Dave Korn" wrote in message news:iaCq8.542$7L2.853370@newsr2.u-net.net... > "Matt Schuckmann" wrote in message > news:aDqq8.22783$9d3.19723@nwrddc01.gnilink.net... > > This may be some what of a C++ question but I believe that the standard > > leaves this type of stuff up to the particular implimentation of the > > language, so here goes. > > > > The question is, in VxWorks when do static objects that are > declared/defined > > within a functions scope get destroyed? I would think that they would get > > destroyed when the module gets unloaded but from my testing that does not > > appear to be the case. > > I would have thought so too. Were they constructed when the module was > loaded? Then they ought to be destructed when it gets unloaded. However > this will only happen if you've properly MUNCHed the object code during > compilation. You should also look at the documentation for cplusLib, in > particular the routines cplusXtorSet, cplusCtors and cplusDtors. > > > DaveK > -- > moderator of > alt.talk.rec.soc.biz.news.comp.humanities.meow.misc.moderated.meow > Burn your ID card! http://www.optional-identity.org.uk/ > Help support the campaign, copy this into your .sig! > Proud Member of the Exclusive "I have been plonked by Davee because he > thinks I'm interesting" List Member # > Master of Many Meowing Minions > Holder of the exhalted PF Chang's Crab Wonton Award for kook spankage above > and beyond the call of hilarity. > > > > Can anyone give me some clues on this? > > > > I can provide my test code for an example if would help. > > > > Thanks > > Matt Schuckmann > > matthew_schuckmann@amat.com > > > > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Startup scripts Date: 3 Apr 2002 12:06:00 -0800 From: arcv0@lycos.com (Roger) Organization: http://groups.google.com/ Message-ID: <45dd6f79.0204031206.2ba2a057@posting.google.com> References: <45dd6f79.0204021536.c706a48@posting.google.com> Thank you everyone. My startup script runs fine now :) arcv0@lycos.com (Roger) wrote in message news:<45dd6f79.0204021536.c706a48@posting.google.com>... > I am using Tornado 2.0. I am having problems using startup script. The > startup scripts runs fine if I boot my target machine over the > ethernet. But when it boots over the flash file system(Disk-on-chip) I > get the messages "unable to open the startup script". It looks like > the startup scripts are called before the Disk-on-chip is configured. > I called the "usrTffsConfig(0,0,"/DOC") in the usrAppInit.c. Please > let me know if anybody has encountered this problem before. > > Thanks in advance for the help > > > Here are the boot parameters. > > boot device : tffs=0,0 > unit number : 0 > processor number : 0 > file name : /tffs/vxworks > flags (f) : 0x0 > startup script (s) : /tffs0/startup.txt > > Attaching to TFFS... done. > Loading /tffs/vxworks...1089412 > Starting at 0x100000... > > Failed to attach to device tffs=0,0Attaching interface lo0...0x3fffdf8 > (tRootTask): Can't attach unknown device tffs=0,0 (unit 0). > done > > Adding 3401 symbols for standalone. > Unable to open startup script /tffs0/startup.txt > -> --------------------------- Newsgroups: comp.os.vxworks Subject: m()  modify memory Date: Wed, 03 Apr 2002 14:15:55 -0600 From: Eric Geelen Organization: tellabs Message-ID: <3CAB62FB.E24D9B31@tellabs.com> This is a multi-part message in MIME format. - --------------743E30773075EDAE88561805 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit When using d() I can spec a range of memory that I would like to display, right? Is there something I can use like the modify command, m(), that allows me to write to more than one byte or word at a time? something like an m_blk() or so? Or should I just write something. Thanks for taking the time. Eric. - --------------743E30773075EDAE88561805 Content-Type: text/x-vcard; charset=us-ascii; name="Eric.Geelen.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Eric Geelen Content-Disposition: attachment; filename="Eric.Geelen.vcf" begin:vcard n:Geelen;Eric L. tel;work:630-512-7880 x-mozilla-html:FALSE url:http://intwww.hq.tellabs.com/~egeelen/ org:Tellabs Inc.;ONG Infrastructure Group adr:;;;;;; version:2.1 email;internet:eric.geelen@tellabs.com title:Firmware Developer note;quoted-printable:tel. 630-512-7880 =0D=0Afax. 630-512-7098=0D=0A x-mozilla-cpt:;9504 fn:Eric L. Geelen end:vcard - --------------743E30773075EDAE88561805-- --------------------------- Newsgroups: comp.os.vxworks,comp.sys.arm Subject: Re: Padding/alignment of structure Date: Wed, 3 Apr 2002 15:32:31 -0500 From: "Harjinder Chayra" Organization: Storm Internet Services Message-ID: References: Sender: terayon@gw-terayon.storm.ca I tried the -fpack-struct option but it had some side effects - still looking into it. Thanks for your help. Harjinder "Harjinder Chayra" wrote in message news:a8f8ad$pik$1@news.storm.ca... > I have a structure which represents interrrupt controller HW registers; it > is being padded and is giving me grief because when I access the registers > I get another registers values! > > How can I get rid of the padding? > > I have tried to use __attribute__ ((aligned(1)) on the structure elements > but no joy. > > I am using the GNU toolkit for ARM9 > > Thanks in advance > Harjinder > > > --------------------------- Newsgroups: comp.os.vxworks,comp.sys.arm Subject: Re: Padding/alignment of structure Date: Wed, 3 Apr 2002 15:38:08 -0500 From: "doug dotson" Organization: Verio Message-ID: References: I assume that will pack all your structs which is probably going to be a performance hit or result in unaligned data accesses depending upon your platform. But if you isolate the stuff that is doing this low-level access into a separate module and compile it only with the -fpack-struct option then that might do it. doug "Harjinder Chayra" wrote in message news:a8fouj$t0j$1@news.storm.ca... > I tried the -fpack-struct option but it had some side effects - still > looking into it. > > Thanks for your help. > > Harjinder > > "Harjinder Chayra" wrote in message > news:a8f8ad$pik$1@news.storm.ca... > > I have a structure which represents interrrupt controller HW registers; it > > is being padded and is giving me grief because when I access the > registers > > I get another registers values! > > > > How can I get rid of the padding? > > > > I have tried to use __attribute__ ((aligned(1)) on the structure elements > > but no joy. > > > > I am using the GNU toolkit for ARM9 > > > > Thanks in advance > > Harjinder > > > > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: BSP UART config. Date: Wed, 3 Apr 2002 15:51:43 -0500 From: "Harjinder Chayra" Organization: Storm Internet Services Message-ID: Sender: terayon@gw-terayon.storm.ca How do you inform VxWorks of the function to call to write out the first character when the output serial device and output ring are enpty? Thanks Harjinder --------------------------- Newsgroups: comp.os.vxworks Subject: What are the common customization to default Exception Handler Date: 3 Apr 2002 16:43:08 -0800 From: ryeung@earthlink.net (Raymond Yeung) Organization: http://groups.google.com/ Message-ID: All, VxWorks supplies a default handler that has the same behavior for all exceptions. I'm using PPC603e (MPC8260 to be precise), which has a short list of 10-15 exceptions. Note: I'm not talking about external interrupts here. What're some of the common exceptions people normally handle with their own handlers? How do they generate the conditions to test their handlers? And I suppose if one uses BAT, there'd be no need to handle TLB misses? Thanks, Raymond --------------------------- Newsgroups: comp.os.vxworks Subject: Re: BSP UART config. Date: Wed, 03 Apr 2002 18:06:15 -0800 From: "Michael R. Kesti" Organization: MK Associates Message-ID: <3CABB517.6803989A@gv.net> References: Reply-To: mkesti@gv.net Harjinder Chayra wrote: >How do you inform VxWorks of the function to call to write out the first >character when the >output serial device and output ring are enpty? You inform vxWorks indirectly, by properly installing a properly written and compiled device driver. Such a driver adds a name to a list of devices that application code may then access using vxWorks' open(), close(), read(), write(), and ioctl() functions. open(devName) returns fd, a file descriptor that is used as an arguement for the others. In the case of a UART driver, write(fd, ...) is used to request data transmission. When an application calls write(fd, ...), the OS calls the write function in the device driver associated with the file descriptor, and that function decides whether its transmit buffer is empty and a character should be written immediately to the UART. - -- ======================================================================== 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: TCL/Gopher code? pPendQ is for what? Date: 03 Apr 2002 21:33:07 -0500 From: Bill Pringlemeir Organization: Factory of the mind Message-ID: Sender: bpringlemeir@DeadDuck It is bugging the marshmallow puss out of me that I can not type `w' and `tw' from a shell over a tyCo. I have found the guilty TCL code. They are found in, $(WIND_BASE)/host/resource/tcl/taskWaitCore.tcl $(WIND_BASE)/host/resource/tcl/app-config/WindSh/01taskWaitShow.tcl I get as far as the TCL procedure `taskWaitGet' whence I encounter, set gopherPrg "$tid +$offset(WIND_TCB,pPendQ) *\ -$offset(SEMAPHORE,qHead) ! <*!>\ +$offset(SEMAPHORE,semType) @b" This is performing some sort of addressing to grab members of the WIND_TCB. It grabs pPendQ. Does it dereference this? The full TCL string translates to, "0x12345678 + 0x05c * - 0x008! <*!> + 0x004 @b" ^^^^^^^^^^ Task ID/WIND_TCB parameter. I guess that gopher is infix/postfix mix and it is grabing a byte (@b) at some address. I don't know what happens from the (WIND_TCB*)p->pPendQ to get to a SEMAPHORE type. The stuff in taskLibP.h doesn't help much either. I would like to re-implement this statement in `C'. tia, Bill Pringlemeir. - -- South Africa mania BROMURE Ft. Meade Crypto AG Exon Shell JSOFC3IP nuclear ARPA Mossad Rand Corporation INSCOM Ceridian CIDA bomb vxWorks FAQ, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" --------------------------- Newsgroups: comp.os.vxworks Subject: Re: TCL/Gopher code? pPendQ is for what? Date: 03 Apr 2002 22:18:42 -0500 From: Bill Pringlemeir Organization: Factory of the mind Message-ID: References: Sender: bpringlemeir@DeadDuck >>>>> "Bill" == Bill Pringlemeir writes: Bill> It is bugging the marshmallow puss out of me that I can not Bill> type `w' and `tw' from a shell over a tyCo. I have found the Bill> guilty TCL code. They are found in, Bill> $(WIND_BASE)/host/resource/tcl/taskWaitCore.tcl Bill> $(WIND_BASE)/host/resource/tcl/app-config/WindSh/01taskWaitShow.tcl Bill> I get as far as the TCL procedure `taskWaitGet' whence I Bill> encounter, Bill> set gopherPrg "$tid +$offset(WIND_TCB,pPendQ) *\ Bill> -$offset(SEMAPHORE,qHead) ! <*!>\ +$offset(SEMAPHORE,semType) Bill> @b" Bill> This is performing some sort of addressing to grab members of Bill> the WIND_TCB. It grabs pPendQ. Does it dereference this? The Bill> full TCL string translates to, Bill> "0x12345678 + 0x05c * - 0x008! <*!> + 0x004 @b" ^^^^^^^^^^ Bill> Task ID/WIND_TCB parameter. The answer is found in the following table, Item | Action - --------------+--------------------------------------------------------- integer | Sets the pointer to the value integer. | [+-] integer | Increments or decrements the pointer by integer. | ! | Writes the pointer to the tape. | * | Replaces the pointer with the value pointed to by the pointer. | @[bwl] | Writes the value pointed to by the pointer on the tape, and | advances the pointer. An optional flag may be specified | to indicate the bit-size that should be written. b | designates an 8-bit value, w a 16-bit value, and l a 32 | bit value. The pointer is advanced by the size of the | element written. If no flag is given, l (32-bits) is | assumed. | $ | Writes the string pointed to by the pointer on the tape. | {script} | Executes the Gopher script inside the braces repeatedly while | the pointer is not NULL. The string inside the braces | works with a local copy of the pointer, maintaining the | original, external pointer value. |