From vxwexplo-errs@csg.lbl.gov Fri Jun 1 04:03:31 2001 From: Vxworks Exploder Date: Fri Jun 1 04:03:33 PDT 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Fri Jun 1 04:03:28 PDT 2001 Subject: Vx_Memory_Partition Subject: SIOCGIFCONF ioctl Subject: dn_expand not found in vxWorks Subject: dn_expand not found in vxWorks Subject: Re: Shared Memory Network Subject: Re: undefined symbols Subject: Anyone want an Altivec-VxWorks Gcc compiler? Subject: Re: Threads and C++ destructors Subject: Re: floating point problems Subject: Writing in flash an tty confict. Subject: copyright_wrs.h Subject: Re: copyright_wrs.h Subject: Re: Threads and C++ destructors Subject: dosFs: how much memory is used by each dir entry Subject: Re: dosFs: how much memory is used by each dir entry Subject: Re: float parameter Subject: a test post 2001-06-01 Subject: Re: zip for vxwork Subject: test only Subject: Re: Is VxWorks my best choice? Subject: Re: CFI driver for TrueFFS? Subject: Re: PCI Problems on x86 ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Vx_Memory_Partition Date: 31 May 2001 12:25:06 GMT From: psenthilnathan@dsqsoft.com Organization: NewsOne.Net - Free Usenet News via the Web - http://newsone.net/ Message-ID: <9f5d72$tcp$1@news.netmar.com> Followup-To: comp.os.vxworks Hi, In the process of creating user memory partition using the simulator, memory which is getting allocated is less by some bytes which varies from 8 to 12 bytes for the allocation of block from a user Partition and 24 to 28 bytes or the creation of Partition . This,I,understand, is the overhead to maintain that block and partition details by the system. But the VXworks manual says this overhead is maintained using the system memory partition only. So, 1. How to handle this loss of memory during allocation process ? 2. Does this loss follow any pattern, like how much less will be allocated ? , so that I can workout accordingly..... Thanks in advance, Senthil. ----- Posted via NewsOne.Net: Free (anonymous) Usenet News via the Web ----- http://newsone.net/ -- Free reading and anonymous posting to 60,000+ groups NewsOne.Net prohibits users from posting spam. If this or other posts made through NewsOne.Net violate posting guidelines, email abuse@newsone.net --------------------------- Newsgroups: comp.os.vxworks Subject: SIOCGIFCONF ioctl Date: 31 May 2001 14:20:04 GMT From: bhttchry@buck.cig.mot.com (Soumyabrata Bhattacharya {SBHATTA1}) Organization: Cellular Infrastructure Group, Motorola Message-ID: <9f5juk$jsa5@nntp.cig.mot.com> Hello Readers, Is there a problem with the SIOCGIFCONF ioctl call on VxWorks 5.4? I am using it to obtain the IP address of the Ethernet interface of the board running VxWorks; but the ioctl keeps returning garbage. The same code works fine on SunOS 4.1.3/Solaris 2.6 (with appropriate changes to the Ethernet interface name, of course). Has anyone noticed any problem using SIOCGIFCONF icotl on VxWorks 5.4 trying to obtain IP address of an interface? Thanks Soumya - -- Soumyabrata Bhattacharya Motorola Inc Lead Software Engineer Email : Soumya.Bhattacharya@motorola.com 3227 N. Wilke Rd Phone : +1-847-632-7830 Arlington Hts, IL 60004 Fax : +1-847-632-2867 --------------------------- Newsgroups: comp.os.vxworks Subject: dn_expand not found in vxWorks Date: 31 May 2001 08:35:39 -0700 From: sarveshku@lycos.com (Sarvesh) Organization: http://groups.google.com/ Message-ID: I wanted to use DNS emplementation on VxWorks. There I have used dn_expand(). But while loading the image to VxWorks, it is giving error undefined symbol dn_expand(). Can anyone suggest me what step shall I take to resolv this. Shall I have to include any defines in Configall.h? Kindly suggest me the remedy. Thanks. Sarvesh --------------------------- Newsgroups: comp.os.vxworks Subject: dn_expand not found in vxWorks Date: 31 May 2001 08:39:24 -0700 From: sarveshku@lycos.com (Sarvesh) Organization: http://groups.google.com/ Message-ID: I wanted to use DNS emplementation on VxWorks. There I have used dn_expand(). But while loading the image to VxWorks, it is giving error undefined symbol dn_expand(). Can anyone suggest me what step shall I take to resolv this. Shall I have to include any defines in Configall.h? Kindly suggest me the remedy. Thanks. Sarvesh --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Shared Memory Network Date: 31 May 2001 09:05:08 -0700 From: Arthur.Quinto@gat.com (Arthur Quinto) Organization: http://groups.google.com/ Message-ID: <5e744e5d.0105310805.3744d35b@posting.google.com> References: <9f3u2g$l35$1@newsie.singa.pore.net> boot device : sm=0x2000600 > processor number : 1 > host name : eesu64 > file name : /robot/vx/config/mv166a/vxWorks > inet on backplane (b): 173.16.12.2:fffffc00 > host inet (h) : 143.89.44.64 > gateway inet (g) : 173.16.12.1 > user (u) : eerobot > ftp password (pw) : multirobot > flags (f) : 0x0 > target name (tn) : vxwork1 > Backplane anchor at 0x2000600... Attaching network interface sm0... > done. > smIsAlive: readyValue = 0x0, heartbeat = 0x0 > smIsAlive: readyValue = 0x87654321, heartbeat = 0x0 > Attaching network interface lo0... done. > Loading... > Error loading file: errno = 0x3c. > Can't load boot file!! I see that you are using FTP to retrieve the kernel. Is the FTP Server running on the host 143.89.44.64? Does the host 143.89.44.64 have a route to your target on the backplane? Is the username "eerobot" and password "multirobot" valid on the host 143.89.44.64? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: undefined symbols Date: Thu, 31 May 2001 18:38:30 +0200 From: "Marek Maleta" Organization: Agilent Technologies Message-ID: <991327115.609319@cswreg.cos.agilent.com> References: <991069142.983331@cswreg.cos.agilent.com> <9evcol$98k$1@rex.ip-plus.net> > > You must add to your VxWorks image: GNU C++ components-standard library- c++ > string > You're right, I haven't all libraries in VxWorks image. Thanks all! Marek --------------------------- Newsgroups: comp.os.vxworks Subject: Anyone want an Altivec-VxWorks Gcc compiler? Date: Thu, 31 May 2001 17:59:12 +0100 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: Hi everyone, I've been hard at work porting the gcc/altivec patchset that I got from those wonderful people at http://www.altivec.org/ into my latest version of the gcc compiler. It's now ready for release. I haven't given it any extensive testing but it compiles altivec functions and outputs altivec machine code, so it looks ok. Problem is that I've run out of space on my website. I want to avoid moving the whole thing across to one of those free servers that insist on plastering banner ads over the whole thing, so I thought I'd ask here if anyone could donate (approx) 45 MB of space on a web or ftp server, to host the source and binary archives for it ? Anyone feeling generous? If you are, please reply to my webmail account throwaway58@hotmail.com or use the "mail me" link on my website, since my newsserver seems to miss 3 out of 4 posts these days. DaveK - -- They laughed at Galileo. They laughed at Copernicus. They laughed at Columbus. But remember, they also laughed at Bozo the Clown. --------------------------- Newsgroups: comp.os.vxworks,comp.programming.threads Subject: Re: Threads and C++ destructors Date: Thu, 31 May 2001 17:53:32 GMT From: Joe Smith Organization: Kill Bastard Spammers Message-ID: <3B168544.7011D312@bogusaddress.com> References: <3AF8DAA6.68D791C9@someplace.com> <989407507.21537.0.nnrp-02.c1c34af9@news.demon.co.uk> <9dfu1l$60o$1@glue.ucr.edu> Reply-To: JoeSmith@bogusaddress.com How does one find out if it is OK to throw threads from a signal handler? (I am interested in the trappable signals like SEGV, BUS, etc.) I am on solaris. This to me seems to be the cleanest way to get all the destructors on the stack to kick in. thp@cs.ucr.edu wrote: > In comp.programming.threads Bill Klein wrote: > : I don't know if this will help you at all, but I've personally found that > : the cleanest way to do things is to *always* let threads finish naturally by > : returning from their starting function. This means that you'll need to come > : up with your own scheme to tell the thread that you want it to exit (though > : a condition variable, or just a flag, or whatever). Following this way of > : doing things guarantees cleanup through C++ destructors, etc... > > Per the C and C++ standards, a signal handler can do almost nothing > more than write to a volatile variable of type sig_atomic_t. In > particular, it invokes undefined behavior if it throws an exception. > But to cleanup you need the thread to process an exception. So, AFIK, > the thread must either respond to the setting of a variable of type > volatile sig_atomic_t or you must have some assurance from your > libraries and/or OS that signals can throw exceptions. > > Tom Payne --------------------------- Newsgroups: comp.os.vxworks Subject: Re: floating point problems Date: Thu, 31 May 2001 13:49:27 -0400 From: "Keith W. Muzzioli" Organization: Raytheon Company Message-ID: <3B168426.800546B8@ssd.ray.com> References: <9eh5s0$ks7$1@overload.lbl.gov> Dave, Thanks for your input! I have to set this task aside for a while but wanted to respond to your questions and remarks... > >-> l 0x000baa4c > > _mathFabsFunc: > >0baa4c 000a DC.W 0x000a > >0baa4e e322 ASL .B D1 , D2 > > Here we see you disassembling the value stored in the variable as if it > was code. Oops! That's the pointer: you want to try "l 0x000ae322". Let > us know what you get when you've looked at the two actual routines > themselves! I guess it's pretty obvious that I don't know assembly!!! After looking in the right location I've found that the code is in fact the same except when the opcode is BSR for the network version of the kernel the argument passed in is __l_sabs, but for the standalone version the argument is an actual address (I asssume one is just using a label and the other is not). > > > >Illegal Instruction > >Program Counter: 0x000710b8 > >Status Register: 0x3000 > >Task: 0x16014a8 > > Have you tried "lkAddr 0x000710b8" yet? To find out what actual routine > is crashing? None of the lookups you did were in that area. Here is the output from "lkAddr 0x000710b8" - -> lkAddr 0x000710b8 0x00070f76 _routestatShow text 0x00071128 _routeShow text 0x00071358 _hostShow text 0x00071410 _nfsDrv text 0x0007146c _nfsMount text 0x0007152c _nfsMountAll text 0x00071614 _nfsDevShow text 0x000716b8 _nfsUnmount text 0x00071ec8 _nfsDirMount text 0x00071f4a _nfsDirUnmount text 0x00071f7a _nfsMntUnmountAll text 0x00071fa8 _nfsMntDump text value = 0 = 0x0 > > General advice: > - Have you called the necessary routine to init the maths library? I believe so. It appears that the same init code is called regardless of whether I'm building the standalone or network version of the kernel. > > - Are you sure the kernel is properly configured for software/hardware FP > support? I believe the answer to this again is yes, however I can't help but think that I am overlooking some #ifdef somewhere and something I need is not being included (as I've mentioned before we had to modify an 5.3 BSP to run under VxWorks5.1). > > - Did you spawn your task with the VX_FP_TASK flag? Yes. > > - Have you been swindled by your supplier and given the cut-down 68EC040 > that has no builtin FPU when you were expecting the full 68040 that does > have an FPU ? No, because when we run the network version of the kernel everything works fine. > Thanks again for your insight, you have given me more things to follow up on. > > > DaveK > -- > They laughed at Galileo. They laughed at Copernicus. They laughed at > Columbus. But remember, they also laughed at Bozo the Clown. --------------------------- Newsgroups: comp.os.vxworks Subject: Writing in flash an tty confict. Date: Thu, 31 May 2001 20:44:38 +0200 From: "XAG" Message-ID: <9f63af$8g8$1@news5.isdnet.net> I have a vxWorks in FLASH (with a ColdFire) and a part of this flash is reserved for configuration (it is a 29F160). Writing to the flash sometimes locks my tty connection. I have locked tasking (taskLock) an ios (intLock) but the problem remains. The flash sector I write is 128K and it lasts few milli seconds. Before writing, I copy my write routine in ram. I don't know exactly what happens and how to solve this problem. Good reception. --------------------------- Newsgroups: comp.os.vxworks Subject: copyright_wrs.h Date: Thu, 31 May 2001 21:01:22 +0200 From: "Michel.Condemine" Organization: 4CE Industry Message-ID: <9f64b9$i6u$1@wanadoo.fr> Reply-To: "Michel.Condemine" Hello, I receive NO rules to make target copyright_wrs.h need by syslib.o ??? copyright_wrs.h is empty. what's this file ? Can you help me ? Michel --------------------------- Newsgroups: comp.os.vxworks Subject: Re: copyright_wrs.h Date: 31 May 2001 13:14:51 -0700 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0105311214.3ab236cd@posting.google.com> References: <9f64b9$i6u$1@wanadoo.fr> "Michel.Condemine" wrote in message news:<9f64b9$i6u$1@wanadoo.fr>... > Hello, > > I receive NO rules to make target copyright_wrs.h need by syslib.o ??? You don't need to make it, you just need to make sure that target/h is in your include path and it will be happy. Sounds like you need to clean build your BSP/project (you didn't say what you were trying to build) so you remove the dependency file. That has probably been corrupted somehow and does not have the full path to the copyright file in it. HTH, John... --------------------------- Newsgroups: comp.os.vxworks,comp.programming.threads Subject: Re: Threads and C++ destructors Date: Thu, 31 May 2001 23:03:10 +0100 From: "Bill Klein" Message-ID: <991346590.6713.0.nnrp-12.c1c34af9@news.demon.co.uk> References: <3AF8DAA6.68D791C9@someplace.com> <989407507.21537.0.nnrp-02.c1c34af9@news.demon.co.uk> <9dfu1l$60o$1@glue.ucr.edu> <3B168544.7011D312@bogusaddress.com> "Joe Smith" wrote in message news:3B168544.7011D312@bogusaddress.com... > How does one find out if it is OK to throw threads from a signal handler? (I am > interested in the trappable signals like SEGV, BUS, etc.) I am on solaris. This > to me seems to be the cleanest way to get all the destructors on the stack to kick > in. Absolutely, it would make life sooo much easier, but no, it can't be done in a platform-independent way (or even compiler-independent way). There may be some specific way to make it work in your situation, but I don't know for sure... After much trial and error and even more frustration I've come to the conclusion that the whole concept of signals is simply incompatible with the concept of multithreaded programming. :( --------------------------- Newsgroups: comp.os.vxworks Subject: dosFs: how much memory is used by each dir entry Date: Thu, 31 May 2001 15:12:35 -0700 From: Satheesha Rangegowda Organization: Agilent Technologies Message-ID: <3B16C1D3.6ADA650B@agilent.com> Does any one know how much RAM is used for each dir/file entry in dosFs. I have about 460 files in my hard disk & it is taking about 4Mb RAM. Is it true? Also, it is using 4Mb in lower 32Meg memory, is there any way to force this to any other memory location. We have PowerPC, so, we would like save this 4Mb for the text segment of the objectModule. Thanks, Satheesha. - -- - ------------------------------------------------------------- Satheesha Rangegowda satheesha_rangegowda@agilent.com R&D Engineer (509) 921 3960 Agilent Technologies Spokane, WA - 99019. - ------------------------------------------------------------- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: dosFs: how much memory is used by each dir entry Date: Thu, 31 May 2001 23:23:47 GMT From: andreyk@home.com (Andray Kaganovsky) Organization: @Home Message-ID: References: <3B16C1D3.6ADA650B@agilent.com> Satheesha Rangegowda wrote in news:3B16C1D3.6ADA650B@agilent.com: > > Does any one know how much RAM is used for each dir/file entry in dosFs. > > I have about 460 files in my hard disk & it is taking about 4Mb RAM. Is > it true? > > Also, it is using 4Mb in lower 32Meg memory, is there any way to force > this to any other memory location. > We have PowerPC, so, we would like save this 4Mb for the text segment of > the objectModule. > > Thanks, > > Satheesha. > > -- > ------------------------------------------------------------- > Satheesha Rangegowda satheesha_rangegowda@agilent.com > R&D Engineer (509) 921 3960 > Agilent Technologies > Spokane, WA - 99019. > ------------------------------------------------------------- > [AndrayK]. If you don't have better ideas, here is what you can try to do. Before you mounted dosFs volume, call malloc() to grab all memory minus 4MB. After you mounted dosFs volume, call free() to return the memory that you have previouslly allocated back to system memory pool. This will hopefully place dosFs data structures in upper memory. I would also check if amount of memory that dosFs allocates for it's internal purposes, is proportional to the size of FAT (File Allocation Table). For performance reasons, dosFs could try to cache the large sections of FAT, or even the entire FAT, in memory. Modern hard disk have capacities measured in GBs, so their FATs could be pretty large. If you happened to have dosFs-2 (which supports multiple disk partitions), you can try partitioning the hard disk, and mounting only those partitions that you actually need. Hope this helps. Andray --------------------------- Newsgroups: comp.os.vxworks Subject: Re: float parameter Date: Thu, 31 May 2001 16:54:33 -0700 From: "Joseph Hansen" Organization: Posted via Supernews, http://www.supernews.com Message-ID: References: <9etljn$1rtm$1@news5.isdnet.net> Jack, The shell does not do floating-point. When you spawn a task, you have to specify that it is a floating-point task (VX_FP_TASK is the parameter to pass in, I believe). But if you call a function from the shell, it is treated as an integer function. One of VxWorks' many limitations, I'm afraid. Joseph "JJS" wrote in message news:9etljn$1rtm$1@news5.isdnet.net... > I use Tornado 2 and VxWorks 5.4 on PPC603 > > A test with a simple function : > TstFloat(float x) > { > printf("x = %f\n",x); > } > > Under shell I launch the function : > TstFloat(10.2) > > The result is : > x = NaN > > What is wrong here? > Thanks > Jack > > --------------------------- Newsgroups: comp.os.vxworks Subject: a test post 2001-06-01 Date: 31 May 2001 18:00:25 -0700 From: zhengfish@21cn.com (rainboy) Organization: http://groups.google.com/ Message-ID: <68781475.0105311700.45c0d857@posting.google.com> a test --------------------------- Newsgroups: comp.os.vxworks Subject: Re: zip for vxwork Date: Fri, 1 Jun 2001 02:06:57 +0000 (UTC) From: vdalal@ipunity.com (Vasu Dalal) Organization: Mailgate.ORG Server - http://www.Mailgate.ORG Message-ID: <3B16F7ED.28C7C6C8@ipunity.com> References: Naveen: > Is there a zip/unzip utility that has been ported for vxWorks ? If not > is there a gzip/tar utility for vxWorks ? Look at deflate and inflate. Typically under :: $WIND_BASE/host/$WIND_HOST_TYPE/bin/deflate To deflate from the command line ( also can write a Makefile rule ):: $) deflate < vxWorks > vxWorks.compressed "inflate" is available and can be found in bootInit.c. We use the API in our boot code.eg. extern STATUS inflate(char *src, char *dst, int src_size); From the reference manual :: [snip] inflateLib NAME inflateLib - inflate code using public domain zlib functions ROUTINES inflate( ) - inflate compressed code DESCRIPTION This library is used to inflate a compressed data stream, primarily for boot ROM decompression. Compressed boot ROMs contain a compressed executable in the data segment between the symbols binArrayStart and binArrayEnd (the compressed data is generated by deflate( ) and binToAsm). The boot ROM startup code (in target/src/config/all/bootInit.c) calls inflate( ) to decompress the executable and then jump to it. This library is based on the public domain zlib code, which has been modified by Wind River Systems. For more information, see the zlib home page at http://quest.jpl.nasa.gov/zlib/. SEE ALSO inflateLib [snip] It works in-memory and not on a file pointer. If u port zlib or gzip to vxWorks, please let me know. HTH, vasu - -- Posted from [12.44.179.201] via Mailgate.ORG Server - http://www.Mailgate.ORG --------------------------- Newsgroups: comp.os.vxworks Subject: test only Date: Fri, 1 Jun 2001 11:42:34 +0800 From: "HuangXin" Organization: BST Message-ID: <9f72vb$2eeq$1@news.cz.js.cn> --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Is VxWorks my best choice? Date: Fri, 01 Jun 2001 05:10:50 GMT From: Ephraim Gadsby Organization: - minimal Message-ID: References: <9e1djv$cub$1@news.netvision.net.il> <3B0F7359.A349872C@gte.net> On Tue, 29 May 2001 09:15:59 GMT, "Vic Sperry" wrote: >"Bill Pringlemeir" wrote in message >news:uy9rha2vx.fsf@yahoo.com... >> >>>>> "Sperry" == Sperry Family writes: >> [snip] >> Sperry> 3. Run both stacks simultaneously. Perhaps designate one >> Sperry> interface as your debugging interface and attach it to your >> Sperry> vxWorks stack. Run everything else through your "real" >> Sperry> stack. I'm not sure this even _can_ work, as I think you'll >> Sperry> have name clashes. Both stacks will have the routine >> Sperry> "socket()", for example. But, this might be doable. >> >> 1. You can run WDB over a serial link. There is no need to have a >> network stack for this to work. > >That's a good idea. I was wrong when I said Tornado "requires" an IP >stack; it doesn't. > >Practically, though, I think a serial (RS-232) interface would be too slow >for every-day development. I've not used Tornado over a serial port, but I have used pSOS Prism and that was pretty usable - even on slow serial ports. Remember that when you are running between breakpoints, not much data needs to go over the link. When single stepping you need to be sparing about what you display eg individual elements instead of the entire array. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: CFI driver for TrueFFS? Date: Fri, 1 Jun 2001 08:57:33 +0200 From: "Jørgen Jakobsen" Organization: Posted through some European Outpost of TDC Internet A/S Message-ID: <9f7ejd$qeu$1@news.inet.tele.dk> References: <3B0A9F9F.35CDA599@digidata.com> Almost correct : It's called cfiscs.c and is found in target\src\drv\tffs /Jørgen "molochai" wrote in message news:ca51557f4a%molochai@vapour-trail.demon.co.uk... > In message <3B0A9F9F.35CDA599@digidata.com> > Doug Dotson wrote: > > > Is there a driver for TrueFFS that supports the new CFI mechanism? > > Yes.. The driver is called something like scscfi.c. > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: PCI Problems on x86 Date: 1 Jun 2001 01:52:20 -0700 From: naggys@hotmail.com (Arjun) Organization: http://groups.google.com/ Message-ID: <5a329084.0106010052.76c37784@posting.google.com> References: <88f5485d.0105300552.5b07275e@posting.google.com> Hi, After defining the macros, you can use pciFindDevice () function to find your device which you need to program. Using the BDF, which is returned by this function, you can proceed to do further setting and config setting required for your device. Hope this helps. - - Arjun eel3672@yahoo.com (Eric Lesch) wrote in message news:<88f5485d.0105300552.5b07275e@posting.google.com>... > Hello, > > I am using VxWorks 5.4 on a pentium PC. I have macros for including > PCI and its show routines defined. > > From DOS I can see all PCI devices, but from VxWorks none of the > devices seem to exist. I have tried both configuration mechanisms 1 > and 2 to see if they work, but they do not. I used the > pciDeviceShow() routine, but it doesn't show any devices. I have even > written and read directly from the PCI I/O ports, CAR and CDR, and > still cannot see any devices. > > Does anybody have any ideas? > > Thanks, > Eric Lesch --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Fri Jun 1 05:57:06 2001 From: vinod m Date: Fri Jun 1 05:57:08 PDT 2001 Subject: networking sir, i'm new to tornado. i've downloaded tornado with networking support. how can i transfer the processed data from the host system(winNT), in which tornado and VxSim are loaded, to another system in the LAN. should i install ULIP?, if so how should i configure it. please send a solution for this. vinod m. ____________________________________________________________________ Get free email and a permanent address at http://www.netaddress.com/?N=1 From vxwexplo-errs@csg.lbl.gov Fri Jun 1 06:18:38 2001 From: PhilP10280@aol.com Date: Fri Jun 1 06:18:40 PDT 2001 Subject: Can VxWorks Apps co-existence with Hard Real-Time Embedded Linux --part1_42.159de58c.2848feb7_boundary Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable Content-Language: en I'm beginning my search to migrateto Embedded Linux from VxWorks.=C2=A0 My g= oalis=20 to run applications with greater complexity, thus Linux.=C2=A0 I would like=20= to use=20 my existing VxWorksapplications as well.=C2=A0=C2=A0 There seems tobe few op= tions out=20 there at the moment when it comes to a hard real time Linuxsolution.=C2=A0 I= =E2=80=99ve=20 read about OnCoreSystems ( http://= www.oncoresystems.com).=C2=A0 Apparently, they=20 have what they sayis a real-time microkernel OS that is based on a Virtual=20 Memory architecture.=C2=A0 Using the processors MMU, their ad notedthat they= can=20 run both both VxWorks applications and a Hard Real-Time Embedded=20 Linuxsolution that can achieve sub 5 microsecond response times.=C2=A0=20 =20 I=E2=80=99ve also looked into using RTLinux=C2=A0 http://www.rtlinux.com/.=C2=A0 They=E2=80=99ve been=20 around a while, but its not clear if they support VxWorks and the difference= s=20 between both.=C2=A0=20 =20 Any suggestions or comments out there. =20 =20 Phil Peterson --part1_42.159de58c.2848feb7_boundary Content-Type: text/html; charset="UTF-8" Content-Transfer-Encoding: quoted-printable Content-Language: en I'm beginning my search t= o migrateto Embedded Linux from VxWorks.=C2=A0 My goalis=20
to run applications with greater complexity, thus Linux.=C2=A0 I would l= ike to use=20
my existing VxWorksapplications as well.=C2=A0=C2=A0 There seems tobe fe= w options out=20
there at the moment when it comes to a hard real time Linuxsolution.=C2= =A0 I=E2=80=99ve=20
read about OnCoreSystems ( htt= p://www.oncoresystems.com).=C2=A0 Apparently, they=20
have what they sayis a real-time microkernel OS that is based on a Virtu= al=20
Memory architecture.=C2=A0 Using the processors MMU, their ad notedthat=20= they can=20
run both both VxWorks applications and a Hard Real-Time Embedded=20
Linuxsolution that can achieve sub 5 microsecond response times.=C2=A0=20
=20
I=E2=80=99ve also looked into using RTLinux=C2=A0 http://www.rtlinux.com/.=C2=A0 They=E2=80=99ve been=20
around a while, but its not clear if they support VxWorks and the differ= ences=20
between both.=C2=A0=20
=20
Any suggestions or comments out there.  
=20
Phil Peterson
--part1_42.159de58c.2848feb7_boundary-- From vxwexplo-errs@csg.lbl.gov Fri Jun 1 07:13:18 2001 From: "Michael Baumann" Date: Fri Jun 1 07:13:21 PDT 2001 Subject: RE: Can VxWorks Apps co-existence with Hard Real-Time Embedded Linux Phil: Sorry I had to cut out your entire quote - but the encoding you used for your mail totally trashed the readability of your message. The gist, from what I can see, is that you want a hard real-time kernel, with linux, and able to run vxWorks tasks. In short - I think it can be done, but don't expect to be able to just copy over your vxWorks code. I am not familiar with OnCoreSystems, but the RT-Linux is an interesting solution, as it is actually a hard real-time microkernel that runs Linux as it's lowest priority task. There is also a group that is doing some work patching the scheduler to give a 'firm' real-time system - not hard, yet not as soft as the normal system. The research is interesting on that one. We are evaluating the possiblity of switching to Linux as well - in fact I've been tasked with porting one of our controllers to Linux as a proof of concept. This first attempt is on a system without hard real-time requirements, so that won't be an issue. Fortunately for us, the port has not been too difficult so far. Biggest difference comes in the device driver arena - as we have a lot of custom hardware, I've had to port our drivers from vxWorks to Linux. Some gotchas to be aware of: 1)VM and an MMU make for an interesting time doing memory-mapped IO. 2)Moving data from kernel space to user space and vice-versa is also a little tricky, but not too bad. 3) zero access to the standard libraries, the only calls you can make are those found in /proc/ksyms, which actually is quite a bit. Other than that, I've not had much trouble - especially since Linux actually implements POSIX threads, complete with priorities. Feel free to drop me a note if you've got questions. Michael Baumann baumann@llumc.edu From vxwexplo-errs@csg.lbl.gov Fri Jun 1 08:17:58 2001 From: "Mark Del Giorno" Date: Fri Jun 1 08:18:00 PDT 2001 Subject: Firewire PMC Module This is a multi-part message in MIME format. ------=_NextPart_000_0007_01C0EA95.5BBFD860 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable I'm looking for a firewire PMC module for which a vxWorks driver is = already written. (I can re-port to my board, which is a dual-cetia = 7400.) Anybody know of one? Thanks in advance - Mark Del Giorno Robotic Systems Technology mark@rst.com ------=_NextPart_000_0007_01C0EA95.5BBFD860 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable
I'm looking for a firewire PMC module for which = a vxWorks=20 driver is already written.  (I can re-port to my board, which is a=20 dual-cetia 7400.)
 
Anybody know of one?
 
Thanks in advance
 
- Mark Del Giorno
Robotic Systems Technology
mark@rst.com
 
------=_NextPart_000_0007_01C0EA95.5BBFD860-- From vxwexplo-errs@csg.lbl.gov Fri Jun 1 08:36:49 2001 From: David Laight Date: Fri Jun 1 08:36:51 PDT 2001 Subject: Re: float parameter > Note also that on some architectures (PowerPC, ARM, maybe others) the > shell is unable to pass floating-point (i.e. double) arguments to > functions at all, although it won't complain about it. This is because > the ABI requires FP arguments to be passed in FP registers, not the > integer registers that the shell uses. It ought to be possible to fix > this, but the WRS SPR that mentions it is ancient, so they probably never > will. This isn't true for the ARM ABI used by vxWorks. Very few ARM cpus actually contain a hardware FP system and associated registers. Passing double values is possible. Note that the ARM standard is a mixed-endian double... Anyway: -> printf( "%f\n", 0x3ff00000, 0 ) 1.000000 -> printf( "%f\n", 0x3ff80000, 0 ) 1.500000 -> printf "%x,%x,%x\n",3.0,~0 40080000,0,ffffffff -> printf "%x,%x,%x\n",sqrt(9.0),~0 40080000,ffffffff,0 It seems the target shell knows that floating point constants exist, and are 64bit. But doesn't know which functions have fp return values. Interestingly though: -> zz=0.9 new symbol "zz" added to symbol table. zz = 0x3f5881c: value = 0.9 -> printf "%x,%x,%x\n",zz,~0 3feccccc,ffffffff,0 -> printf "%x,%x,%x\n",(double)zz,~0 3feccccc,cccccccc,ffffffff however attempts to run (double)sqrt(9.0) cause the shell to crash badly. David ---------------------------------------------------------------- David Laight email: dsl@tadpole.co.uk Tadpole Technology plc phone: +44 1223 428 232 Cambridge, UK fax: +44 1223 428 201 From vxwexplo-errs@csg.lbl.gov Tue Jun 5 00:14:23 2001 From: "Michael Lawnick" Date: Tue Jun 5 00:14:25 PDT 2001 Subject: Re: pipeline problem on PowerPC Hi Chris, From: "Chris Slominski" Newsgroups: comp.os.vxworks Sent: Friday, June 01, 2001 9:10 PM Subject: pipeline problem on PowerPC > VxWorks Users, > I am porting code from a mv167 to a mv2700. I configured the board, > recompiled the application and ran. I got an "interrupt: bad VME interrupt > 0" error immediately and was told by a co-worker that this may be due to the > mv2700 instruction pipeline executing my interrupt reset instruction after > the return from the interrupt handler. I've heard of an "eieio" instruction > that can be used force sequential execution over a set of instructions. > Please let me know if this is indeed a likely candidate for my interrupt > crash, and what are good choices to correct the problem. > > Thanks, > Chris I fear this alone won't solve your problem. Standard configuration uses a 'posted write', i.e. Universe will complete write cycle independantly from CPU. Your EIEIO will only ensure, that the write to Universe will have completed before next instruction execution. The access to your VME-H/W will still happen, when the bus is aritrated. The CPU may have completed some jobs 'till then. The secure way is a read-back the IACK value. This way the CPU is forced to wait for completion. The second method (with more performance impact) is to switch off 'posted write'. HTH Mit freundlichen Grüßen, Michael Lawnick ============================================== SOFTEC GmbH Tel +49-731-96600-0 Promenade 17 Fax +49-731-96600-23 D-89073 Ulm Michael Lawnick Germany lawnick@softec.de ============================================== From vxwexplo-errs@csg.lbl.gov Tue Jun 5 02:03:52 2001 From: "Moy, Ian" Date: Tue Jun 5 02:03:55 PDT 2001 Subject: Booting problem This is a multi-part message in MIME format. --------------InterScan_NT_MIME_Boundary Content-Type: text/plain; charset="iso-8859-1" Hi, I am trying to set-up a Motorola MBX860 board for vxWorks and have a problem. I have successfully built the bootrom and downloaded that to the board but when I try and start the board booting vxWorks I get the following error: Attaching network interface cpm0... done. Attaching network interface lo0... done. Loading... Error loading file: errno = 0xd0003 The board does not make contact to the FTP server at all and I have checked the ethernet port, cables and all the boot parameters are correct. What else could be wrong? Any help would be gratefully received, Regards, Ian Moy --------------InterScan_NT_MIME_Boundary Content-Type: text/plain; name="RMRL-Disclaimer.txt" Content-Transfer-Encoding: 7bit Content-Disposition: attachment; filename="RMRL-Disclaimer.txt" The information contained in this e-mail is confidential to Roke Manor and must not be passed to any third party without permission. This communication is for information only and shall not create or change any contractual relationship. --------------InterScan_NT_MIME_Boundary-- From vxwexplo-errs@csg.lbl.gov Tue Jun 5 03:55:58 2001 From: Priya Pradhan Date: Tue Jun 5 03:56:01 PDT 2001 Subject: Using timer_create() and related routines..... > Hi! > > I am trying to create a periodic timer on VxWorks using the routines - > timer_create(), timer_connect(), timer_settime(). Everytime I try to run > the program I get the error - "kill failed" - I can't understand the > error. > > The code is included below..... > > #include > #include > #include > #include > #include > > extern int errno ; > > void my_handler(int param) ; > > timer_t timer_id ; > > int timers(void) > { > int i = 0 ; > struct timespec tp_gettime, tp_settime, tp_clkres ; > struct itimerspec time_val ; > > clock_getres(CLOCK_REALTIME, &tp_clkres) ; > > clock_gettime(CLOCK_REALTIME, &tp_gettime) ; > > printf("tp_gettime.tv_sec == %d\n",tp_gettime.tv_sec) ; > printf("tp_gettime.tv_nsec == %d\n",tp_gettime.tv_nsec) ; > > tp_settime.tv_sec = tp_clkres.tv_sec ; > tp_settime.tv_nsec = tp_clkres.tv_nsec ; > > clock_settime(CLOCK_REALTIME, &tp_settime) ; > > if(timer_create(CLOCK_REALTIME, NULL, &timer_id) == ERROR) > { > printf("timer :: Error number set is == %d\n",errno) ; > exit(1) ; > } > else > printf("created timer.... \n") ; > > if(timer_connect(timer_id, my_handler, 0) == ERROR) > { > printf("timer :: Error no set is == %d\n",errno) ; > exit(0) ; > } > else > printf("connected timer.... \n") ; > > time_val.it_interval.tv_sec = 1 ; > time_val.it_interval.tv_nsec = 0 ; > time_val.it_value.tv_sec = 1 ; > time_val.it_value.tv_nsec = 0 ; > > if(timer_settime(timer_id, CLOCK_REALTIME, &time_val, NULL) == ERROR) > { > printf("timer :: Error number set is == %d\n",errno) ; > exit(1) ; > } > else > printf("settime for timer.... \n") ; > > clock_gettime(CLOCK_REALTIME, &tp_gettime) ; > > printf("tp_gettime.tv_sec == %d\n",tp_gettime.tv_sec) ; > printf("tp_gettime.tv_nsec == %d\n",tp_gettime.tv_nsec) ; > > while(i != 1000000) > i++ ; > > return 0 ; > } > > void my_handler(int param) > { > struct itimerspec time_val ; > > printf("..........in function............ \n") ; > fflush(stdout) ; > > time_val.it_interval.tv_sec = 1 ; > time_val.it_interval.tv_nsec = 0 ; > time_val.it_value.tv_sec = 1 ; > time_val.it_value.tv_nsec = 0 ; > > if(timer_settime(timer_id, CLOCK_REALTIME, &time_val, NULL) == ERROR) > { > printf("my_handler :: Error number set is == %d\n",errno) ; > exit(0) ; > } > > return ; > } > > Hoping someone will be able to help. > > Thanks, > Priya. > Priya Pradhan > Member of Technical Staff. > Net Brahma Technologies Pvt. Ltd. > Internext Networking Software > Ph: 5521451 (Extn 212) > http://www.netbrahma.com > > > From vxwexplo-errs@csg.lbl.gov Tue Jun 5 04:03:24 2001 From: Vxworks Exploder Date: Tue Jun 5 04:03:27 PDT 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Tue Jun 5 04:03:09 PDT 2001 Subject: Re: intConnect seems to work only one time Subject: Re: intConnect seems to work only one time Subject: networking Subject: Re: vxWorks 5.3.1 and Motorola MVME 2400 rev G SBC. Subject: Green Hills MULTI 2000 Tornado 2 Kernel Build Subject: Can VxWorks Apps co-existence with Hard Real-Time Embedded Linux Subject: RE: Can VxWorks Apps co-existence with Hard Real-Time Embedded Linux Subject: ANSI + NO ANSI CODE Subject: Re: TCP Theory question puzzling me Subject: Firewire PMC Module Subject: re: float parameter Subject: Re: Weird dosFs problem? Subject: Host filesystem in simulator Subject: Re: float parameter Subject: pipeline problem on PowerPC Subject: Re: TCP Theory question puzzling me Subject: Re: zip for vxwork Subject: looking for MIPS evaluation board (re-post) Subject: Some problem with C++ Subject: Re: Why can I find snmpdLib.c? Subject: Re: ANSI + NO ANSI CODE Subject: Re: News from ANNA KOURNAKOVA !!! Subject: Re: float parameter Subject: Re: Firewire PMC Module Subject: need help Subject: connecting modem to vxworks Subject: lsil64108 with vxworks Subject: test Subject: Re: networking Subject: What is LARGE and SMALL Library? Subject: unable to send udp packet on 255.255.255.255 Subject: How to insert version info in VxWorks image? Subject: Re: How to insert version info in VxWorks image? Subject: Re: looking for MIPS evaluation board (re-post) Subject: Re: connecting modem to vxworks Subject: Re: How to insert version info in VxWorks image? Subject: Does anyone has idea about add an .obj into lib.a? Subject: Intel 82559ER END driver Subject: Re: Some problem with C++ Subject: ElfHex Tool Subject: Controling point at which socket write blocks Subject: Re: Host filesystem in simulator Subject: Re: Some problem with C++ Subject: Re: CFI driver for TrueFFS? Subject: ldarm crash... Subject: Re: Some problem with C++ Subject: Problem Pinging Other OS's Subject: Passthrough filesystem on NT. Subject: MVME2604 lockup problem Subject: Re: Some problem with C++ Subject: Use of VxWorks Exception Task - "tExcTask" Subject: informazioni Subject: Re: Some problem with C++ Subject: Re: Does anyone has idea about add an .obj into lib.a? Subject: Re: Some problem with C++ Subject: Problem of Transmit Descriptor Subject: VxWorks Network Protocol Development Engineer Subject: Re: Problem of Transmit Descriptor Subject: Re: Use of VxWorks Exception Task - "tExcTask" Subject: boot vxWorks from harddisk on mcp750 Subject: Re: connecting modem to vxworks Subject: Re: ElfHex Tool Subject: tNetTask Error Subject: re: pipeline problem on PowerPC Subject: pipeline problem on PowerPC Subject: Re: connecting modem to vxworks Subject: Booting problem Subject: C-code generated with mibcomp Subject: Re: C-code generated with mibcomp Subject: Socket creation problem Subject: problem with timeout Subject: Re: Booting problem Subject: Re: Booting problem Subject: Re: Booting problem ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: intConnect seems to work only one time Date: Fri, 1 Jun 2001 15:11:01 +0200 From: "Jonathan ILIAS" Organization: University of Angers, France. Message-ID: <9f8485$2o4n$1@news.univ-angers.fr> References: <9evh7b$mn1$1@news.univ-angers.fr> Hello all, Here is the Kelly's answer to my post : - ----- Original Message ----- From: "Kelly Hornsby" To: Sent: Tuesday, May 29, 2001 12:30 PM Subject: Re: intConnect seems to work only one time > > > you cannot unconnect an interrupt once it is connected. I fussed at WRS > about this, and my BSP instructor admitted there was a fault there, but told > me that I could create my own unconnector. Normally the intConnect just > creates a new link into its linked list for the interrupt handler. > > I suggest that you use a "fixed" interrupt handler, and supply a function > pointer to call in the "fixed" routine that does nothing but call the > function through the pointer. If the function pointer is null, then do > nothing. > > --------------------------- Newsgroups: comp.os.vxworks Subject: networking Date: 1 Jun 2001 07:59:11 MDT From: vinod m Organization: Lawrence Berkeley National Laboratory Message-ID: <9f885n$6h2$1@overload.lbl.gov> sir, i'm new to tornado. i've downloaded tornado with networking support. how can i transfer the processed data from the host system(winNT), in which tornado and VxSim are loaded, to another system in the LAN. should i install ULIP?, if so how should i configure it. please send a solution for this. vinod m. ____________________________________________________________________ Get free email and a permanent address at http://www.netaddress.com/?N=1 --------------------------- Newsgroups: comp.os.vxworks Subject: Re: vxWorks 5.3.1 and Motorola MVME 2400 rev G SBC. Date: Fri, 01 Jun 2001 10:28:31 -0400 From: Michael Reed Message-ID: <3B17A68F.B72CD299@ll.mit.edu> References: <3B153A03.8EA2AD13@appsig.com> Hi Mario, Not exactly an answer to your question, but we've also had problems with the UniverseII rev switch on mvme2400 boards. Our initial purchase involved mvme2400 boards with Rev B, and a later batch contained this rev C (CA91C142C). We're running pSOS, and we couldn't even boot the boards w/ rev C. Currently, Motorola has them, and we asked them to put the old rev back on. They don't know if they can do it. Good luck, Mike Mario Martinez wrote: > > vxWorks 5.3.1 fails sysMemProbe() on the VME bus for Motorola mvme2400 > rev G SBC. > I imagine the reason is the change of the Universe II chip (to > CA91C142C). > This problem is not present on revC boards from motorola and vxWorks > 5.3.1. > Also this problem is not present with vxWorks 5.4 and rev G boards, but > i also noticed > that vxWorks 5.4 is not backward compatible when it comes to VME > accesses in rev C boards. > > Has anyone encountered similar problems with vxWorks 5.3.1 and the > Motorola mvme2400 revG boards? > > any help would be appreciated. > thanks, > mario. --------------------------- Newsgroups: comp.os.vxworks Subject: Green Hills MULTI 2000 Tornado 2 Kernel Build Date: Fri, 1 Jun 2001 09:26:12 -0500 From: "Brad Crabtree" Organization: Raytheon Company Message-ID: Has anyone successfully linked the Green Hill's EC++ libraries into a Tornado 2 Project? We are having trouble translating what we did in Tornado 1.0.1 kernel builds to link the MULTI libraries(such as the makefile,usrConfig.c,etc. mods) into the correct project settings for Tornado 2. Thanks. p.s. Remove "-nojunk" for my real e-mail. - -------------------------------------------- Bradley D. Crabtree b-crabtree@raytheon-nojunk.com --------------------------- Newsgroups: comp.os.vxworks Subject: Can VxWorks Apps co-existence with Hard Real-Time Embedded Linux Date: Fri, 1 Jun 2001 10:20:39 EDT From: PhilP10280@aol.com Organization: Lawrence Berkeley National Laboratory Message-ID: <9f8bm7$8f0$1@overload.lbl.gov> - --part1_42.159de58c.2848feb7_boundary Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable Content-Language: en I'm beginning my search to migrateto Embedded Linux from VxWorks.=C2=A0 My g= oalis=20 to run applications with greater complexity, thus Linux.=C2=A0 I would like=20= to use=20 my existing VxWorksapplications as well.=C2=A0=C2=A0 There seems tobe few op= tions out=20 there at the moment when it comes to a hard real time Linuxsolution.=C2=A0 I= =E2=80=99ve=20 read about OnCoreSystems ( http://= www.oncoresystems.com).=C2=A0 Apparently, they=20 have what they sayis a real-time microkernel OS that is based on a Virtual=20 Memory architecture.=C2=A0 Using the processors MMU, their ad notedthat they= can=20 run both both VxWorks applications and a Hard Real-Time Embedded=20 Linuxsolution that can achieve sub 5 microsecond response times.=C2=A0=20 =20 I=E2=80=99ve also looked into using RTLinux=C2=A0 http://www.rtlinux.com/.=C2=A0 They=E2=80=99ve been=20 around a while, but its not clear if they support VxWorks and the difference= s=20 between both.=C2=A0=20 =20 Any suggestions or comments out there. =20 =20 Phil Peterson - --part1_42.159de58c.2848feb7_boundary Content-Type: text/html; charset="UTF-8" Content-Transfer-Encoding: quoted-printable Content-Language: en I'm beginning my search t= o migrateto Embedded Linux from VxWorks.=C2=A0 My goalis=20
to run applications with greater complexity, thus Linux.=C2=A0 I would l= ike to use=20
my existing VxWorksapplications as well.=C2=A0=C2=A0 There seems tobe fe= w options out=20
there at the moment when it comes to a hard real time Linuxsolution.=C2= =A0 I=E2=80=99ve=20
read about OnCoreSystems ( htt= p://www.oncoresystems.com).=C2=A0 Apparently, they=20
have what they sayis a real-time microkernel OS that is based on a Virtu= al=20
Memory architecture.=C2=A0 Using the processors MMU, their ad notedthat=20= they can=20
run both both VxWorks applications and a Hard Real-Time Embedded=20
Linuxsolution that can achieve sub 5 microsecond response times.=C2=A0=20
=20
I=E2=80=99ve also looked into using RTLinux=C2=A0 http://www.rtlinux.com/.=C2=A0 They=E2=80=99ve been=20
around a while, but its not clear if they support VxWorks and the differ= ences=20
between both.=C2=A0=20
=20
Any suggestions or comments out there.  
=20
Phil Peterson
- --part1_42.159de58c.2848feb7_boundary-- --------------------------- Newsgroups: comp.os.vxworks Subject: RE: Can VxWorks Apps co-existence with Hard Real-Time Embedded Linux Date: Fri, 1 Jun 2001 08:15:20 -0700 From: "Michael Baumann" Organization: Lawrence Berkeley National Laboratory Message-ID: <9f8bm8$8f1$1@overload.lbl.gov> Phil: Sorry I had to cut out your entire quote - but the encoding you used for your mail totally trashed the readability of your message. The gist, from what I can see, is that you want a hard real-time kernel, with linux, and able to run vxWorks tasks. In short - I think it can be done, but don't expect to be able to just copy over your vxWorks code. I am not familiar with OnCoreSystems, but the RT-Linux is an interesting solution, as it is actually a hard real-time microkernel that runs Linux as it's lowest priority task. There is also a group that is doing some work patching the scheduler to give a 'firm' real-time system - not hard, yet not as soft as the normal system. The research is interesting on that one. We are evaluating the possiblity of switching to Linux as well - in fact I've been tasked with porting one of our controllers to Linux as a proof of concept. This first attempt is on a system without hard real-time requirements, so that won't be an issue. Fortunately for us, the port has not been too difficult so far. Biggest difference comes in the device driver arena - as we have a lot of custom hardware, I've had to port our drivers from vxWorks to Linux. Some gotchas to be aware of: 1)VM and an MMU make for an interesting time doing memory-mapped IO. 2)Moving data from kernel space to user space and vice-versa is also a little tricky, but not too bad. 3) zero access to the standard libraries, the only calls you can make are those found in /proc/ksyms, which actually is quite a bit. Other than that, I've not had much trouble - especially since Linux actually implements POSIX threads, complete with priorities. Feel free to drop me a note if you've got questions. Michael Baumann baumann@llumc.edu --------------------------- Newsgroups: comp.os.vxworks Subject: ANSI + NO ANSI CODE Date: Fri, 1 Jun 2001 18:19:34 +0200 From: "Michel.Condemine" Organization: 4CE Industry Message-ID: <9f8f85$8g2$1@wanadoo.fr> Reply-To: "Michel.Condemine" Hello ALL, somebody knows how the mix ANSI and NO ANSI code in the same image ? Thanks Michel --------------------------- Newsgroups: comp.os.vxworks Subject: Re: TCP Theory question puzzling me Date: 1 Jun 2001 12:56:16 -0400 From: vanbaren@falcon.si.com (vanbaren_gerald) Organization: Smiths Industries Message-ID: <3b17c930@news.si.com> References: "Geoff" writes: >Hello >I am developing an embedded application, which requires DHCP client >functionality (Not Intel/Windows based). It was fairly easy to establish an >IP address using the DHCP call provided by the Operating System. However, >what is puzzling me is, now that I have a dynamic IP address, how do other >machines address this machine ? I would like them to access this machine by >host name, but what mechanism do I use, to tell the DNS server my name and >DHCP'd IP address? Any help or insight into this problem would be greatly >appreciated ? Thanks > Geoff You typically would use a fixed address served by your DHCP server (it should have the capability to associate a MAC address with a static IP address rather than allocating it from the pool). At that point, you can put the static IP address in your DNS server host tables. Not very elegant. Anybody have a more elegant approach? gvb - -- +---------------------------------------------------------------------------+ | Jerry Van Baren / vanbaren_gerald@si.com / Grand Rapids Mi / 616-241-7973 | | My employer is a company. Companies are artifacts of a legal system. | |________________Artifacts are incapable of having opinions.________________| --------------------------- Newsgroups: comp.os.vxworks Subject: Firewire PMC Module Date: Fri, 1 Jun 2001 12:21:17 -0400 From: "Mark Del Giorno" Organization: Lawrence Berkeley National Laboratory Message-ID: <9f8in8$bl3$1@overload.lbl.gov> This is a multi-part message in MIME format. - ------=_NextPart_000_0007_01C0EA95.5BBFD860 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable I'm looking for a firewire PMC module for which a vxWorks driver is = already written. (I can re-port to my board, which is a dual-cetia = 7400.) Anybody know of one? Thanks in advance - - Mark Del Giorno Robotic Systems Technology mark@rst.com - ------=_NextPart_000_0007_01C0EA95.5BBFD860 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable
I'm looking for a firewire PMC module for which = a vxWorks=20 driver is already written.  (I can re-port to my board, which is a=20 dual-cetia 7400.)
 
Anybody know of one?
 
Thanks in advance
 
- Mark Del Giorno
Robotic Systems Technology
mark@rst.com
 
- ------=_NextPart_000_0007_01C0EA95.5BBFD860-- --------------------------- Newsgroups: comp.os.vxworks Subject: re: float parameter Date: Fri, 1 Jun 2001 17:38:27 +0100 (BST) From: David Laight Organization: Lawrence Berkeley National Laboratory Message-ID: <200106011638.RAA08229@dsl-2.tadpole.co.uk> > Note also that on some architectures (PowerPC, ARM, maybe others) the > shell is unable to pass floating-point (i.e. double) arguments to > functions at all, although it won't complain about it. This is because > the ABI requires FP arguments to be passed in FP registers, not the > integer registers that the shell uses. It ought to be possible to fix > this, but the WRS SPR that mentions it is ancient, so they probably never > will. This isn't true for the ARM ABI used by vxWorks. Very few ARM cpus actually contain a hardware FP system and associated registers. Passing double values is possible. Note that the ARM standard is a mixed-endian double... Anyway: - -> printf( "%f\n", 0x3ff00000, 0 ) 1.000000 - -> printf( "%f\n", 0x3ff80000, 0 ) 1.500000 - -> printf "%x,%x,%x\n",3.0,~0 40080000,0,ffffffff - -> printf "%x,%x,%x\n",sqrt(9.0),~0 40080000,ffffffff,0 It seems the target shell knows that floating point constants exist, and are 64bit. But doesn't know which functions have fp return values. Interestingly though: - -> zz=0.9 new symbol "zz" added to symbol table. zz = 0x3f5881c: value = 0.9 - -> printf "%x,%x,%x\n",zz,~0 3feccccc,ffffffff,0 - -> printf "%x,%x,%x\n",(double)zz,~0 3feccccc,cccccccc,ffffffff however attempts to run (double)sqrt(9.0) cause the shell to crash badly. David - ---------------------------------------------------------------- David Laight email: dsl@tadpole.co.uk Tadpole Technology plc phone: +44 1223 428 232 Cambridge, UK fax: +44 1223 428 201 --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Weird dosFs problem? Date: 1 Jun 2001 10:40:03 -0700 From: Michael@cote.org (Michael Cote) Organization: http://groups.google.com/ Message-ID: References: <200105151415.PAA06701@dsl-2.tadpole.co.uk> I debugged that exact problem in DOSFSv1. It is fixed in V2. The scenario I found that reproduced this every time was: Create a file but don't write anything to it. Reboot the system without closing the file descriptor. After rebooting, create a new file and write to it. You'll find that the same data shows up in both file. One solution is never to leave a file empty after creating it. If you are still using DOSFS V1, stop. It has so many bugs WRS tech support generally won't even talk to you about them. Mike David Laight wrote in message news:<200105151415.PAA06701@dsl-2.tadpole.co.uk>... > Finding the 'wrong' data in a file is usually caused by an earlier corruption of > the FAT - a block is marked free when it is allocated to a file. > > This causes the same block to be allocated to two files later on. > > We have seen quite a few corrupted dosfs filesystems - but never identified a > sequence that actually corrupts the disk. > > I don't know the mechanism used by compact flash cards to support 512 byte sectors > using (I assume) standard flash parts with 64k byte erase blocks. I have seem > corrupt filesystems where it looked as though a sector was in the wrong place - > which might be a firmware bug on the CF card, but is just as likely to be a > hardware problem on our board. > > I have so far refrained from using dosfs2, partially because I'm building vxWorks > 5.3.1 and 5.4 on the same source base, and partially because I don't think it will > interact very well with the cache I use to get reasonable performance from flash. > > David > > > > > We ran into this weird dosFs problem. Checking if others had a similar > > experience. > > > > We have the following directories: > > > > /ata/cfg - for config files > > /ata/log - for log files > > > > Created many config files - startup.cfg, test.cfg, etc - under /ata/cfg. > > When startup.cfg is created, we keep backup copies of the previous 3 > > versions as startup1.cfg, startup2.cfg, startup3.cfg (rename files). > > Similarly, we rotate the log files in the directory /ata/log once a day, > > keeping the last N days' files. > > > > I displayed the content of one of the log files and found that it had what > > should have been in a config file! I'm checking our code for bugs, but one > > of the suspects is dosFs. Has anyone had problems where dosFs puts the > > content that's supposed to be in one file in another? > > > > We use Tornado 2.0.2, so I would expect it to have the latest version of > > dosFs. I've heard that the previous version was very bad. > > > > Thanks, > > Hari > > > ---------------------------------------------------------------- > David Laight email: dsl@tadpole.co.uk > Tadpole Technology plc phone: +44 1223 428 232 > Cambridge, UK fax: +44 1223 428 201 --------------------------- Newsgroups: comp.os.vxworks Subject: Host filesystem in simulator Date: Fri, 01 Jun 2001 14:08:57 -0400 From: Doug Dotson Organization: Verio Message-ID: <3B17DA39.3D836D1F@digidata.com> Greetings! Is it possible to build a customized simulator that includes host filesystem support? When using real iron, all my file I/O goes to the boot host which is fine. I would like to do s a similar thing when using the simulator where file I/O goes to the machine running Tornado. I created a new bootable project based on the simulator, enabled the Network Remote I/O Support and rebuilt. I get a bunch of errors for missing avl tree operations. Thanks, DOug --------------------------- Newsgroups: comp.os.vxworks Subject: Re: float parameter Date: Fri, 01 Jun 2001 13:32:27 -0500 From: Andrew Johnson Organization: APS, Argonne National Laboratory Message-ID: <3B17DFBB.99CC25F@aps.anl.gov> References: <200106011638.RAA08229@dsl-2.tadpole.co.uk> David Laight wrote: > > > Note also that on some architectures (PowerPC, ARM, maybe others) the > > shell is unable to pass floating-point (i.e. double) arguments to > > functions at all, although it won't complain about it. This is because > > the ABI requires FP arguments to be passed in FP registers, not the > > integer registers that the shell uses. It ought to be possible to fix > > this, but the WRS SPR that mentions it is ancient, so they probably never > > will. > > This isn't true for the ARM ABI used by vxWorks. Very few ARM cpus actually > contain a hardware FP system and associated registers. Passing double values is > possible. Note that the ARM standard is a mixed-endian double... Interesting, I stand (sit) corrected - I believed something I read in WRS's spr 6201: From the shell level, calling functions with parameters of type float does not work on architectures where there is a different parameter passing mechanism for floats and integers, namely, PowerPC, MIPS, PA-RISC and ARM. - - Andrew - -- The world is such a cheerful place when viewed from upside-down It makes a rise of every fall, a smile of every frown --------------------------- Newsgroups: comp.os.vxworks Subject: pipeline problem on PowerPC Date: Fri, 1 Jun 2001 15:10:00 -0400 From: "Chris Slominski" Organization: Jefferson Lab Message-ID: <9f8pb1$6a9$1@inn.jlab.org> VxWorks Users, I am porting code from a mv167 to a mv2700. I configured the board, recompiled the application and ran. I got an "interrupt: bad VME interrupt 0" error immediately and was told by a co-worker that this may be due to the mv2700 instruction pipeline executing my interrupt reset instruction after the return from the interrupt handler. I've heard of an "eieio" instruction that can be used force sequential execution over a set of instructions. Please let me know if this is indeed a likely candidate for my interrupt crash, and what are good choices to correct the problem. Thanks, Chris --------------------------- Newsgroups: comp.os.vxworks Subject: Re: TCP Theory question puzzling me Date: Fri, 01 Jun 2001 20:01:25 GMT From: "Geoff" Organization: blueyonder (post doesn't reflect views of blueyonder) Message-ID: References: <3b17c930@news.si.com> Hi As you say this static approach is not very elegant as it somewhat negates the purpose of DHCP in the first place. I have got a bit more info on this question now since I first asked . The solution to this problem is Dynamic DNS, this sounds like it is something that is still fairly new and standards are still evolving. In simple terms the DNS server and the DHCP Server talk to each other in order to keep the DNS's client information upto date in real time. Here is a link that explains it in more detail http://www.join.com/doc/server.4.html Regards Geoff vanbaren_gerald wrote in message news:3b17c930@news.si.com... > "Geoff" writes: > > >Hello > > >I am developing an embedded application, which requires DHCP client > >functionality (Not Intel/Windows based). It was fairly easy to establish an > >IP address using the DHCP call provided by the Operating System. However, > >what is puzzling me is, now that I have a dynamic IP address, how do other > >machines address this machine ? I would like them to access this machine by > >host name, but what mechanism do I use, to tell the DNS server my name and > >DHCP'd IP address? Any help or insight into this problem would be greatly > >appreciated ? Thanks > > > Geoff > > > You typically would use a fixed address served by your DHCP server (it > should have the capability to associate a MAC address with a static > IP address rather than allocating it from the pool). At that point, > you can put the static IP address in your DNS server host tables. > > Not very elegant. Anybody have a more elegant approach? > > gvb > > -- > +--------------------------------------------------------------------------- + > | Jerry Van Baren / vanbaren_gerald@si.com / Grand Rapids Mi / 616-241-7973 | > | My employer is a company. Companies are artifacts of a legal system. | > |________________Artifacts are incapable of having opinions.________________| --------------------------- Newsgroups: comp.os.vxworks Subject: Re: zip for vxwork Date: Fri, 01 Jun 2001 20:04:56 GMT From: Karen Organization: Bell Nexxia Message-ID: LZO data compression library might also be a good choice, and it has been reported to work under vxWorks (I haven't tried so though) http://wildsau.idv.uni-linz.ac.at/mfx/lzo.html Just also notice that if you have Tornado 2.0 with DosFs2, then you will have the tarLib UNIX tar compatible library Karen vdalal@ipunity.com (Vasu Dalal) said: > > Naveen: > > > Is there a zip/unzip utility that has been ported for vxWorks ? If not > > is there a gzip/tar utility for vxWorks ? > > Look at deflate and inflate. > > Typically under :: > > $WIND_BASE/host/$WIND_HOST_TYPE/bin/deflate > > To deflate from the command line ( also can write a Makefile rule ):: > > $) deflate < vxWorks > vxWorks.compressed > > > "inflate" is available and can be found in bootInit.c. We use the API > in our boot code.eg. > extern STATUS inflate(char *src, char *dst, int src_size); > > > From the reference manual :: > > [snip] > inflateLib > > NAME > > inflateLib - inflate code using public domain zlib functions > > ROUTINES > > inflate( ) - inflate compressed code > > DESCRIPTION > > This library is used to inflate a compressed data stream, > primarily for boot ROM decompression. Compressed boot ROMs contain a > compressed > executable in the data segment between the symbols binArrayStart > and binArrayEnd (the compressed data is generated by deflate( ) and > binToAsm). The boot ROM startup code (in > target/src/config/all/bootInit.c) calls inflate( ) to decompress the > executable and then jump to it. > > This library is based on the public domain zlib code, which has > been modified by Wind River Systems. For more information, see the zlib > home > page at http://quest.jpl.nasa.gov/zlib/. > > SEE ALSO > > inflateLib > [snip] > > > It works in-memory and not on a file pointer. If u port zlib or gzip > to vxWorks, please let me know. > > HTH, > vasu > > > -- > Posted from [12.44.179.201] > via Mailgate.ORG Server - http://www.Mailgate.ORG > --------------------------- Newsgroups: comp.os.vxworks Subject: looking for MIPS evaluation board (re-post) Date: Fri, 01 Jun 2001 20:20:01 GMT From: andreyk@home.com (Andray Kaganovsky) Organization: @Home Message-ID: I am looking for MIPS evaluation board that meets two requirements below: - it should be modern board - BSP must be available from WindRiver, and must be stable (i.e no digging in piles of WindSurf's SPRs to find out why the bloody thing hangs during the boot etc.) Could someone please provide me with some pointers in this regard. Thanks in advance, Andray --------------------------- Newsgroups: comp.os.vxworks Subject: Some problem with C++ Date: Fri, 01 Jun 2001 16:53:40 -0400 From: Beginner Organization: UUNET Canada News Reader Service Message-ID: <3B1800D4.FAE5B1DC@hotmail.com> Hi All, Could someone tell me what is wrong with the following: #include #include std::ofstream out; void initOut() { new (&out) std::ofstream(); } Gcc in VxWorks 5.4 give the follwoing error: DebugLog.cpp : In function `static void initOut()': :10: too many arguments to function `operator new(unsigned int)' DebugLog.cpp:10: at this point in file The code has no problem compiled and run under Visual C++ 6.0. I tried both the Gcc 2.7 and the 2.95.3 with the same result. TIA --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Why can I find snmpdLib.c? Date: Thu, 31 May 2001 07:05:49 -0700 From: Joe Durusau Organization: Lockheed Martin Corporation Message-ID: <3B164FBD.C6B7C1E0@lmco.com> References: <9f4jfr$jeu$1@news.seed.net.tw> Send WRS a large bag of money, and they will give you a source license. Speaking only for myself, Joe Durusau "OH!NEWS" wrote: > > Help! > When I trace code of snmp, > miss the snmpdLib.c. > Where can I find it? > > best regards > -- > Firmware Engineer > GLORY TELECOM. > Phone:(886)-3-3127832-386 > Fax:(886)-3-3127831 > http://www.glorytel.com.tw > email:kumo@glorytel.com.tw --------------------------- Newsgroups: comp.os.vxworks Subject: Re: ANSI + NO ANSI CODE Date: Sat, 2 Jun 2001 02:32:46 +0200 From: "Michel.Condemine" Organization: 4CE Industry Message-ID: <9f9c51$659$1@wanadoo.fr> References: <9f8f85$8g2$1@wanadoo.fr> Reply-To: "Michel.Condemine" One precision, I use the GNU compiler + Tornado 2 + VXDCOM. Waiting for answer. "Michel.Condemine" a écrit dans le message news: 9f8f85$8g2$1@wanadoo.fr... > Hello ALL, > > somebody knows how the mix ANSI and NO ANSI code in the same image ? > > Thanks > > Michel > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: News from ANNA KOURNAKOVA !!! Date: 01 Jun 2001 20:22:50 -0400 From: Bill Pringlemeir Organization: Bell Nexxia Message-ID: References: <9f95fl$3ku$1348@news.ukr.net> Sender: bpringlemeir@DeadDuck >>>>> "Spammer" == writes: Spammer> ANNA KOURNAKOVA NEW Anna Kournikova - Topless!!! Brand new [snip] Spammer> body!!. This one can't be missed Picture Watch Anna show her Spammer> bra and panties during this match Movie New in shot of Anna Spammer> showing just how sweaty a game of tennis can be!!!! Do you have pictures of her using vxWorks? ...perhaps in the fetish section. - -- AIRCRAFT. TORPEDOED. LEAVE BOAT. U-235. Useful vxWorks URLS, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" "http://www.xs4all.nl/~borkhuis/vxworks/vxw_pt9.html" --------------------------- Newsgroups: comp.os.vxworks Subject: Re: float parameter Date: 1 Jun 2001 18:17:51 -0700 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0106011717.7c41248f@posting.google.com> References: <200106011638.RAA08229@dsl-2.tadpole.co.uk> <3B17DFBB.99CC25F@aps.anl.gov> Andrew Johnson wrote in message news:<3B17DFBB.99CC25F@aps.anl.gov>... > David Laight wrote: > > > > This isn't true for the ARM ABI used by vxWorks. Very few ARM cpus actually > > contain a hardware FP system and associated registers. Passing double values is > > possible. Note that the ARM standard is a mixed-endian double... > > Interesting, I stand (sit) corrected - I believed something I read in > WRS's spr 6201: > > From the shell level, calling functions with parameters of type float > does > not work on architectures where there is a different parameter passing > mechanism for floats and integers, namely, PowerPC, MIPS, PA-RISC and > ARM. The SPR is accurate. It states that for floats where there are differences in the parameter passing from integers there might be problems. For ARM, this is not due to hardware floating point regs (as it is for PPC and MIPS), but is more likely due to the strange cross endian format that David mentioned. That said, if I type printf "%f\n", 4.3 at my AE shell it works as expected and displays 4.300000. HTH, John... --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Firewire PMC Module Date: Sat, 02 Jun 2001 04:43:31 GMT From: "ag" Organization: Excite@Home - The Leader in Broadband http://home.com/faster Message-ID: References: <9f8in8$bl3$1@overload.lbl.gov> You might check www.mindready.com "Mark Del Giorno" wrote in message news:9f8in8$bl3$1@overload.lbl.gov... > This is a multi-part message in MIME format. > > ------=_NextPart_000_0007_01C0EA95.5BBFD860 > Content-Type: text/plain; > charset="iso-8859-1" > Content-Transfer-Encoding: quoted-printable > > I'm looking for a firewire PMC module for which a vxWorks driver is = > already written. (I can re-port to my board, which is a dual-cetia = > 7400.) > > Anybody know of one? > > Thanks in advance > > - Mark Del Giorno > Robotic Systems Technology > mark@rst.com > > > ------=_NextPart_000_0007_01C0EA95.5BBFD860 > Content-Type: text/html; > charset="iso-8859-1" > Content-Transfer-Encoding: quoted-printable > > > > http-equiv=3DContent-Type> > > > > >
I'm looking for a firewire PMC module for which = > a vxWorks=20 > driver is already written.  (I can re-port to my board, which is a=20 > dual-cetia 7400.)
>
 
>
Anybody know of one?
>
 
>
Thanks in advance
>
 
>
- Mark Del Giorno
>
Robotic Systems Technology
>
href=3D"mailto:mark@rst.com">mark@rst.com
>
 
> > ------=_NextPart_000_0007_01C0EA95.5BBFD860-- > --------------------------- Newsgroups: comp.os.vxworks Subject: need help Date: Sat, 2 Jun 2001 12:42:05 +0200 From: PAtrixkXsat Organization: Wanadoo, l'internet avec France Telecom Message-ID: Xsat cdtv310 use vxworks interface somebody can give me some infos about thanxx PAtrickXsat http://www.xsat.com4.ws ( Unnofficial Website ) --------------------------- Newsgroups: comp.os.vxworks Subject: connecting modem to vxworks Date: 2 Jun 2001 04:28:37 -0700 From: sarveshku@lycos.com (Sarvesh) Organization: http://groups.google.com/ Message-ID: HI I want to connect my VxWorks machine to internet through modem dialing. But I am not able to write to modem. Can anyone tell me how to initialize the modem. TR LED in the Modem is not glowing. I have done pppInit() also but still it is not dialing. My modem is connected to COM port2. Please suggest me some procedure. --------------------------- Newsgroups: comp.os.vxworks Subject: lsil64108 with vxworks Date: Sat, 02 Jun 2001 12:21:06 GMT From: "Serge Haas" Organization: chello broadband Message-ID: Hello, please , where can i found object module for assign a LSIL64108 with Vxworks Thanks for the reponse serge --------------------------- Newsgroups: comp.os.vxworks Subject: test Date: Sat, 2 Jun 2001 14:33:59 +0200 From: PAtrixkXsat Organization: Wanadoo, l'internet avec France Telecom Message-ID: test --------------------------- Newsgroups: comp.os.vxworks Subject: Re: networking Date: Sat, 02 Jun 2001 08:46:31 -0400 From: Joe Durusau Organization: http://bellsouth.webusenet.com - Home of the fastest NNTP servers on the Net. Message-ID: <3B18E027.2BF5D432@bellsouth.net> References: <9f885n$6h2$1@overload.lbl.gov> I'm not sure what you are trying to do. If you are trying to run a real target, as opposed to the simulator, you have to BUY Tornado and a BSP. The version youy can download is only sutiable for playing around. If you just want to transfer data around, it depends on what the systems are. WinNT commonly lets you use the network neighborhood thing to move files around. Almost anything requires the use of FTP. Speaking only for myself, Joe Durusau vinod m wrote: > > sir, > i'm new to tornado. i've downloaded tornado with networking support. how can i > transfer the processed data from the host system(winNT), in which tornado and > VxSim are loaded, to another system in the LAN. should i install ULIP?, if so > how should i configure it. please send a solution for this. > vinod m. > > ____________________________________________________________________ > Get free email and a permanent address at http://www.netaddress.com/?N=1 --------------------------- Newsgroups: comp.os.vxworks Subject: What is LARGE and SMALL Library? Date: Sat, 2 Jun 2001 09:49:51 -0700 From: "deb" Organization: Posted via Supernews, http://www.supernews.com Message-ID: Hello, What is the diferrence between Large and Small library for MIPS library in Vxworks. Thanks Deb --------------------------- Newsgroups: comp.os.vxworks Subject: unable to send udp packet on 255.255.255.255 Date: 2 Jun 2001 09:57:57 -0700 From: pareshj@yahoo.com (Paresh Jain) Organization: http://groups.google.com/ Message-ID: <92b087a4.0106020857.7e2438e6@posting.google.com> I am writing a DHCP client on VxWorks, and facing the problem in sending a udp packet to 255.255.255.255 . As per requirement of DHCP I have to set my IP address to 0.0.0.0 and then hv to send the DHCP DISCOVER request on 255.255.255.255 . But it fails in 'sendto', returning 0x41 errno. I went thru vxWorks support site and found a workaround (http://www.windriver.com/cgi-bin/csdocs/techtips/public/viewSum.cgi?id=519) that add a routing entry directing all requests to 255.255.255.255, should be forwarded to some well identified default gateway. Well I tried this also but it doesn't work with 0.0.0.0 IP address. Function mRouteEntryAdd fails for ip address 0.0.0.0 . Catch me where I am going wrong....?? - --Paresh --------------------------- Newsgroups: comp.os.vxworks Subject: How to insert version info in VxWorks image? Date: Sun, 3 Jun 2001 21:32:47 +0200 From: "Rui Germano" Organization: Planet Internet Message-ID: <9fe3ap$5fmd0$1@reader03.wxs.nl> Well, the title says it all. If anyone has an anwer tell me. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to insert version info in VxWorks image? Date: Sun, 03 Jun 2001 19:52:38 GMT From: "Mark Fanara" Organization: Excite@Home - The Leader in Broadband http://home.com/faster Message-ID: References: <9fe3ap$5fmd0$1@reader03.wxs.nl> We have inserted version information by modifying the rominit.s file (in your BSP). There is a windriver copyright near the top of of the file, just after the program entry point. We just added some storage just prior to the copyright. We then created a tool to read the binary image and poke the version number into that storage. Rui Germano wrote in message news:9fe3ap$5fmd0$1@reader03.wxs.nl... > Well, the title says it all. > If anyone has an anwer tell me. > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: looking for MIPS evaluation board (re-post) Date: 3 Jun 2001 19:49:41 GMT From: borkhuis@agere.com (Johan Borkhuis) Organization: Agere System Message-ID: References: andreyk@home.com (Andray Kaganovsky) wrote: > > I am looking for MIPS evaluation board that meets two > requirements below: > > - it should be modern board > > - BSP must be available from WindRiver, and must be > stable (i.e no digging in piles of WindSurf's SPRs > to find out why the bloody thing hangs during the > boot etc.) I am using the IDTS334 board, and it is working pretty well. You need the IDTS134 BSP (standard WRS engineering BSP) plus the extra BSP for the IDTS334 board. But then it works OK 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: connecting modem to vxworks Date: 3 Jun 2001 19:52:35 GMT From: borkhuis@agere.com (Johan Borkhuis) Organization: Agere System Message-ID: References: sarveshku@lycos.com (Sarvesh) wrote: > HI > I want to connect my VxWorks machine to internet through modem > dialing. But I am not able to write to modem. Can anyone tell me how > to initialize the modem. TR LED in the Modem is not glowing. I have > done pppInit() also but still it is not dialing. My modem is connected > to COM port2. Please suggest me some procedure. As far as I know PPP does not do the dialing. You have to initialize the modem and send the dial-commands to it. When the connection is started you can start the PPP service. Groeten, Johan - -- o o o o o o o . . . _____________________________ o _____ || Johan Borkhuis | .][__n_n_|DD[ ====_____ | borkhuis@agere.com | >(________|__|_[_________]_|__________________________| _/oo OOOOO oo` ooo ooo 'o!o!o o!o!o` === VxWorks FAQ: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html === --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to insert version info in VxWorks image? Date: Mon, 04 Jun 2001 03:33:58 GMT From: "Dan Perlberger" Organization: Road Runner - EC Message-ID: References: <9fe3ap$5fmd0$1@reader03.wxs.nl> There's a version.h file you can edit to display a version text string when you boot. I'm not sure if that's what you're talking about or not. Dan "Rui Germano" wrote in message news:9fe3ap$5fmd0$1@reader03.wxs.nl... > Well, the title says it all. > If anyone has an anwer tell me. > > --------------------------- Newsgroups: comp.os.vxworks Subject: Does anyone has idea about add an .obj into lib.a? Date: Mon, 4 Jun 2001 12:26:37 +0800 From: "kumo Yeh" Organization: SEEDNet News Service Message-ID: <9ff2oe$gg8$1@news.seed.net.tw> I was check the GNU Linker,but it seems don't work? I am not sure the parameter is correct, using -i & -o and it gernerate default file a.out, and it's not right! Because too small. .a ~= 4Mbytes .o ~= 64kbytes a.out = 2kbytes - -- Firmware Engineer GLORY TELECOM. Phone:(886)-3-3127832-386 Fax:(886)-3-3127831 http://www.glorytel.com.tw email:kumo@glorytel.com.tw --------------------------- Newsgroups: comp.os.vxworks Subject: Intel 82559ER END driver Date: 4 Jun 2001 00:22:00 -0700 From: netproa@hanmail.net (IL HWAN, KIM) Organization: http://groups.google.com/ Message-ID: <29f27f2c.0106032322.528b9352@posting.google.com> Hello, I have two 82559ER ethernet controller on my system.(used MPC8240) When I use fei82557End driver, BSP failed in muxDevStart(). (For debugging, I use only one, not configured the other) I only changed Device ID, and don't care EEPROM checksum. I saw in this group, to use 82557End driver for 82559 is OK. Please, let me know how to modify 82557 END driver for 82559ER ethernet controller. I hope your answer. Another problem, When I changed the driver object code to rx82559end, above function go well, and FTP control connection work well(LOGIN, TYPE I, RETR ...), but 'Unable to open FTP data connection' message printed. I don't know how to fix. Best regards. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Some problem with C++ Date: Mon, 4 Jun 2001 07:23:00 -0400 From: "Chris Slominski" Organization: Jefferson Lab Message-ID: <9ffr3e$h7l$1@inn.jlab.org> References: <3B1800D4.FAE5B1DC@hotmail.com> TIA, I did not understand your example, but if you want to create a dynamic output stream on the heap I would do as follows. The "new" operator returns a pointer to an object. std::ofstream *out; ... out = new std::ofstream(); Chris --------------------------- Newsgroups: comp.os.vxworks Subject: ElfHex Tool Date: Mon, 4 Jun 2001 13:57:40 +0000 (UTC) From: ahenric@ccii.co.za (Arne Henrichsen) Organization: Mailgate.ORG Server - http://www.Mailgate.ORG Message-ID: <5.0.2.1.0.20010604155618.00a63080@alpha.ccii.co.za> Hi, A question about the elfHex tool from vxWorks which converts the vxWorks=20 object file to a Motorola s-record file. The last entry of the s-record=20 file is always S9........ The address field of S9 contains the starting=20 execution address. Mine is always 000000. I want it to be 100000. Is that=20 part of the elfHex tool or do I need to set that up in my BSP??? Thanks - ---------------------------------------------------------------- Arne Henrichsen - Software Design Engineer C=B2I=B2 Systems (Pty) Ltd Tel: (+27 21) 683 5490 Fax: (+27 21) 683 5435 Cell: (+27 82) 872 0773 Email: ahenric@ccii.co.za Url: http://www.ccii.co.za - -- Posted from alpha.ccii.co.za [66.8.24.68] via Mailgate.ORG Server - http://www.Mailgate.ORG --------------------------- Newsgroups: comp.os.vxworks Subject: Controling point at which socket write blocks Date: 4 Jun 2001 07:47:48 -0700 From: nick.gardner@europe.simoco.com (nick) Organization: http://groups.google.com/ Message-ID: Hello, We want to change the point at which at socket blocks. Our system may have in excess of 120 sockets simultaneously writing, and I would like each socket to block after 400 bytes or so has been written without a corresponding read. I have tried changing the send buffer, but this just slows things down (I think it does not work because data is copied from the send buffer to another vxworks / tcp buffer). The other methods, such as recv with MSG_PEEK, and querying the FIONREAD values only work at the other end of the socket (which is done by another company / software, and we have no control of). All this seems to point me in the direction of tweaking mbufs, mblks, clusters, etc., but to be honest I don't have a clue about all that. Any help would be greatly appreciated, Regards, Nick nick.gardner@europe.simoco.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Host filesystem in simulator Date: Mon, 04 Jun 2001 16:16:42 GMT From: andreyk@home.com (Andray Kaganovsky) Organization: @Home Message-ID: References: <3B17DA39.3D836D1F@digidata.com> Doug Dotson wrote in news:3B17DA39.3D836D1F@digidata.com: > Greetings! > > Is it possible to build a customized simulator that includes > host filesystem support? When using real iron, all my file > I/O goes to the boot host which is fine. I would like to > do s a similar thing when using the simulator where file > I/O goes to the machine running Tornado. I created a > new bootable project based on the simulator, enabled > the Network Remote I/O Support and rebuilt. I get a > bunch of errors for missing avl tree operations. > > Thanks, > DOug > > > [AndrayK]. Would Target Server File System (TSFS) be suitable for your needs ? VxWorks 5.4 Programmers Guide, Section 4.7. Andray --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Some problem with C++ Date: Mon, 04 Jun 2001 12:55:10 -0400 From: Beginner Organization: UUNET Canada News Reader Service Message-ID: <3B1BBD6E.FCE05642@hotmail.com> References: <3B1800D4.FAE5B1DC@hotmail.com> <9ffr3e$h7l$1@inn.jlab.org> Hi, Chris, Thanks. What I am trying to do is such that when I use "out", I don't need to put a * infront of it. For example, "out << test" not "*out << test", so it looks just like cout, cerr etc. I believe "new (&out) std::ofstream();" should let me do that. Under under VC++, I could. Now after doing what you suggest, I found out I could open a file but the text just didn't get stored into it, anyone use ofstrem out there under VxWorks that is working? TIA Chris Slominski wrote: > TIA, > I did not understand your example, but if you want to create a dynamic > output stream on the heap I would do as follows. The "new" operator returns > a pointer to an object. > > std::ofstream *out; > > ... > > out = new std::ofstream(); > > Chris --------------------------- Newsgroups: comp.os.vxworks Subject: Re: CFI driver for TrueFFS? Date: 4 Jun 2001 16:52:02 GMT From: aparker@vtlinx.com Organization: NewsOne.Net - Free Usenet News via the Web - http://newsone.net/ Message-ID: <9fgebi$gst$1@news.netmar.com> References: <3B0A9F9F.35CDA599@digidata.com> Followup-To: comp.os.vxworks In article <3B0A9F9F.35CDA599@digidata.com>, Doug Dotson writes: >Is there a driver for TrueFFS that supports the new CFI mechanism? > >Thanks, >Doug Dotson > >Yes there is , contact WindRiver Rep. - -Tony ----- Posted via NewsOne.Net: Free (anonymous) Usenet News via the Web ----- http://newsone.net/ -- Free reading and anonymous posting to 60,000+ groups NewsOne.Net prohibits users from posting spam. If this or other posts made through NewsOne.Net violate posting guidelines, email abuse@newsone.net --------------------------- Newsgroups: comp.os.vxworks Subject: ldarm crash... Date: 04 Jun 2001 12:59:36 -0400 From: Bill Pringlemeir Organization: Bell Nexxia Message-ID: Sender: bpringlemeir@DeadDuck I have constructed a linker script by using the `--verbose' option and then putting, TARGET(binary) INPUT(fpga.bin) TARGET(coff-arm-little) at the start of the file and including the statements after .rdata, . = ALIGN(0x10); fpga.bin This appears to crash the linker (ldarm) on my NT workstation. Do other have problems with use of the ld `INPUT' option? It appears to be this, as I can link without the input statement (and filename in .rdata). I have also used several different sets of `.o' files in my command line. I have written code to convert the binary to a `C' array, but this solution seems cleaner and it works flawlessly on a uCLinux distribution that I have used on other projects. Any hints, TSR, SPRs, etc. regards, Bill Pringlemeir. - -- Useful vxWorks URLS, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" "http://www.xs4all.nl/~borkhuis/vxworks/vxw_pt9.html" --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Some problem with C++ Date: Mon, 04 Jun 2001 18:15:26 +0100 From: Simon Farnsworth Message-ID: <3B1BC22E.7040403@snellwilcox.com> References: <3B1800D4.FAE5B1DC@hotmail.com> Beginner wrote: > Hi All, > > Could someone tell me what is wrong with the following: > > #include > #include > > std::ofstream out; > > void initOut() > { > new (&out) std::ofstream(); > } > > Gcc in VxWorks 5.4 give the follwoing error: > > > DebugLog.cpp > : In function `static void initOut()': > :10: too many arguments to function `operator new(unsigned > int)' > DebugLog.cpp:10: at this point in file > > The code has no problem compiled and run under Visual C++ 6.0. > > I tried both the Gcc 2.7 and the 2.95.3 with the same result. > > TIA > > > Your C++ is rather weird. std::ofstream out will create an object called out of type std::ofstream, which you use as in: out << "Test Output\n"; If I understand the concept of initOut(), you are trying to initialise the out object; this has already been done when you wrote std::ofstream out. --------------------------- Newsgroups: comp.os.vxworks Subject: Problem Pinging Other OS's Date: Mon, 04 Jun 2001 17:23:01 GMT From: zen_of_slack@yahoo.ca (Orest) Organization: NovAtel Inc. Message-ID: <3b1bbc7a.323839406@news.novatel.ca> I have written a device driver for a wireless LAN card, and I am having problems pinging a linux system from a VxWorks shell. The linux system is using the same wireless LAN card. I have no problems pinging another VxWorks system running my driver, and can connect two VxWorks systems using sockets and transmit data back and forth. However, when I try to ping a linux box, the VxWorks system does not appear to acknowledge the ping reply. I am running tcpdump on the linux system, and it is getting the ICMP echo request packet, and responding with an echo reply packet, but the VxWorks system doesn't seem to recognize the reply and states that there is no answer received. However, the linux system IP address does show up in the ARP table after trying to do a ping, and the table contains the correct MAC address. I have checked the raw data being received, and VxWorks system is getting the echo reply packet from the linux system, and my driver is passing the data up to the network layer. I have also tried connecting to a linux system using sockets, with no luck. Here is my setup: linux system IP address: 10.0.0.5 netmask 255.0.0.0 VxWorks System: ifAddrSet "sy0", "10.0.0.2" ifShow results: ifShow sy (unit number 0): Flags: (0x8863) UP BROADCAST MULTICAST ARP RUNNING Type: ETHERNET_CSMACD Internet address: 10.0.0.2 Broadcast address: 10.255.255.255 Netmask 0xff000000 Subnetmask 0xff000000 Ethernet address is 00:a0:f8:8d:48:4d routeShow results: ROUTE NET TABLE destination gateway flags Refcnt Use Interface - ---------------------------------------------------------------------------- 10.0.0.0 10.0.0.2 101 0 0 sy0 - ---------------------------------------------------------------------------- ROUTE HOST TABLE destination gateway flags Refcnt Use Interface - ---------------------------------------------------------------------------- - ---------------------------------------------------------------------------- After calling ping "10.0.0.5", 1, 0 from the shell, the arp table looks like: LINK LEVEL ARP TABLE destination gateway flags Refcnt Use Interface - -------------------------------------------------------------------------- 10.0.0.5 00:a0:f8:8e:ab:a2 405 0 10 sy0 - ------------------------------------------------------------------------- My knowledge of VxWorks networking is pretty limited, so I'm hoping that I'm just missing an obvious step in configuring my system. Any help would be appreciated, thanks. Orest --------------------------- Newsgroups: comp.os.vxworks Subject: Passthrough filesystem on NT. Date: Mon, 04 Jun 2001 13:41:12 -0400 From: Doug Dotson Organization: Verio Message-ID: <3B1BC838.25A229D1@digidata.com> Does anyone know how to set up a passthrough filesystem on NT. I tried to clone a standard vxSim and adding the "dosDisk" component but I get unresolved symbols when trying to build. Thanks, Doug --------------------------- Newsgroups: comp.os.vxworks Subject: MVME2604 lockup problem Date: 4 Jun 2001 10:50:26 -0700 From: keith.kingsley@lintek.aeroflex.com (Keith Kingsley) Organization: http://groups.google.com/ Message-ID: We're having a problem with the MVME2604 occasionally locking up on VME reads/writes to a certain card (specifically, a PMAC II VME Ultralite). It appears that when the lockup occurs, the MVME2604 receives the expected DTACK from the PMAC board, dutifully releases the data strobe lines (DS1 and DS2), but then fails to release the address strobe line (AS). The MVME2604 then locks up with the bottom four status lights lit. When the event happens, it also appears that the 2604 board responds unusually quickly to the DTACK-- say, 19 ns instead of the usual ~50 ns. Has anyone seen anything like this before? We suspect there is a problem with the Universe II chip. Our 2604 board has CA91C142C-33CE, although we've tried another 2604 with CA91C142-33CE installed and the same problem happens. - -Keith Kingsley Aeroflex Lintek --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Some problem with C++ Date: Mon, 04 Jun 2001 14:17:52 -0400 From: Beginner Organization: UUNET Canada News Reader Service Message-ID: <3B1BD0D0.B0BA892@hotmail.com> References: <3B1800D4.FAE5B1DC@hotmail.com> <3B1BC22E.7040403@snellwilcox.com> Hi, Simon, Thanks for your reply. Acttually it was suggested in a May 1999 C/C++ Users Journal "The Journeyman¡¯s Shop: Initialization and Cleanup, Part 3" article written by Pete Becker. (Unforfortunately the article is not on-line :( )) The reason is that there are some prolem in my out object initialization order that causes a crush (right on the out.open(filename) call) if I execute the code using dynamic loading. If I execute the code straight (i.e. not via DL) it is OK. That is when I digged out that article and try to see if that is the problem. Life is strange sometime when you just want to solve a simple problem and a Pandora (sp?) box breaks out :). Anyway, I have since using the "new" (no more crushing) to create the out object, but I still couldn't get the text output to the file (i.e. the file is empty after closing). Thanks again. Simon Farnsworth wrote: > Beginner wrote: > > > Hi All, > > > > Could someone tell me what is wrong with the following: > > > > #include > > #include > > > > std::ofstream out; > > > > void initOut() > > { > > new (&out) std::ofstream(); > > } > > > > Gcc in VxWorks 5.4 give the follwoing error: > > > > > > DebugLog.cpp > > : In function `static void initOut()': > > :10: too many arguments to function `operator new(unsigned > > int)' > > DebugLog.cpp:10: at this point in file > > > > The code has no problem compiled and run under Visual C++ 6.0. > > > > I tried both the Gcc 2.7 and the 2.95.3 with the same result. > > > > TIA > > > > > > > > Your C++ is rather weird. std::ofstream out will create an object called > out of type std::ofstream, which you use as in: out << "Test Output\n"; > If I understand the concept of initOut(), you are trying to initialise > the out object; this has already been done when you wrote std::ofstream out. --------------------------- Newsgroups: comp.os.vxworks Subject: Use of VxWorks Exception Task - "tExcTask" Date: 4 Jun 2001 11:32:19 -0700 From: jlally@drc.com (Jim Lally) Organization: http://groups.google.com/ Message-ID: <6d87d2a2.0106041032.bd6b0cd@posting.google.com> Using Tornado 2.0 - VxWorks 5.4 - Motorolla 68K platforms I'm trying to identify all uses of the vxWorks exceptions task "tExcTask". Any Help would be appreciated. Currently identified uses are: · Exception Handling Processing that must be done in a task context! Possibly when "excHookAdd" used? · Watchdog Timer support! · Debugging support! Setting up of breakpoints, etc.? · Use of Unix like signaling mechanism! Delivery of signals to multiple tasks? Signaling handler executing in tExcTask task context? · Library dependencies are: dbgLib (Debugger), sigLib (Signal Library), and intLib (hardware specific interrupt library). · Supports a task's deletion by executing taskDelete Hooks for tasks deleting themselves. Any help would be appreciated. --------------------------- Newsgroups: comp.os.vxworks Subject: informazioni Date: Mon, 4 Jun 2001 18:59:22 +0200 From: "federico torretta" Organization: Tiscali Spa Message-ID: <9fgkf2$1gn$4@lacerta.tiscalinet.it> my name is Federico and I am a major in >>Electronic Engineering >>My thesis in microelectronics is about the >>implementation in hardware of an algorhythm that >>simulates the reaction of a neuron network exposed to >>outer electomagnetic fields. >>The above mentioned algorhythm first of all models the >>behaviour of single ionic channels of the cellular >>membrane, then it joins them as to simulate the >>bahaviour of the cellular membrane (determination of >>its action potential when exposed to an >>electromagnetic field, referring above all to the >>Hadgin-Huxley approach)and finally the single cellular >>membranes are integrated in order to describe the >>behavoiur of a neuron network. >>I kindly ask you if you know any hardware device which >>already implements algorhythm that are similar to the >>above mentioned. Or more simply how I can connect DSP >>processors (which simulate the effect of the membrane >>of a single neuron) without using share memories, (foe >>example with a bus ...) >> >>Thanking you in advance I remain >> >> Sincerely Yours >> Federico --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Some problem with C++ Date: Mon, 4 Jun 2001 14:59:26 -0400 From: "Chris Slominski" Organization: Jefferson Lab Message-ID: <9fglr8$hm1$1@inn.jlab.org> References: <3B1800D4.FAE5B1DC@hotmail.com> <9ffr3e$h7l$1@inn.jlab.org> <3B1BBD6E.FCE05642@hotmail.com> Do you really want to create the ofstream object on the heap? Why not std::ofstream out; ... out.open("file.name"); ... out << "test"; --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Does anyone has idea about add an .obj into lib.a? Date: 4 Jun 2001 20:24:45 GMT From: borkhuis@agere.com (Johan Borkhuis) Organization: Agere System Message-ID: References: <9ff2oe$gg8$1@news.seed.net.tw> "kumo Yeh" wrote: > I was check the GNU Linker,but it seems > don't work? > > I am not sure the parameter is correct, > using -i & -o and it gernerate default file a.out, > and it's not right! > > Because too small. > .a ~= 4Mbytes > .o ~= 64kbytes > a.out = 2kbytes If you want to add a .o file to an archive you have to use the archive tool (ar, for example arppc for PowerPC). It is not possible to add a .out file to an archive, you can only add object-files Groeten, Johan - -- o o o o o o o . . . _____________________________ o _____ || Johan Borkhuis | .][__n_n_|DD[ ====_____ | borkhuis@agere.com | >(________|__|_[_________]_|__________________________| _/oo OOOOO oo` ooo ooo 'o!o!o o!o!o` === VxWorks FAQ: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html === --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Some problem with C++ Date: Mon, 04 Jun 2001 17:31:56 -0400 From: Beginner Organization: UUNET Canada News Reader Service Message-ID: <3B1BFE4C.EE7F347@hotmail.com> References: <3B1800D4.FAE5B1DC@hotmail.com> <9ffr3e$h7l$1@inn.jlab.org> <3B1BBD6E.FCE05642@hotmail.com> <9fglr8$hm1$1@inn.jlab.org> Hi Chris, Please see my last reply to Simon. I fact, that was what I did originally. But then when I port my application into a dynamic loadable module, it crashed at the out.open("file.name"). After much debugging, I kind of deduced that the DL application is not running at the memory space of the application that loads it or because the initialization of the ofstream is out of order, or it doesn't have enough to run it. I am not sure. After creating the obj from the heap, it no longer crashed. Like I said before, did anyone successfully use ofstream on the VxWork? because the file I created is always empty, nothing was put on the file. Chris Slominski wrote: > Do you really want to create the ofstream object on the heap? Why not > > std::ofstream out; > > ... > > out.open("file.name"); > > ... > > out << "test"; --------------------------- Newsgroups: comp.os.vxworks Subject: Problem of Transmit Descriptor Date: Mon, 04 Jun 2001 21:55:35 GMT From: Tony Rodayer Organization: Guest of ProXad - France Message-ID: <3B1C0439.9000906@free.fr> Hy, Our product crashes when we increase the Number of Transmit Descriptor up to 128. Does anyone can help me? I have got this message : data access Exception current instruction address: 0x0004397c Machine Status Register: 0x00003030 Data Access Register: 0x21724011 Condition Register: 0x44242082 Data storage interrupt Register: 0x00003030 Task: 0x1ce48e0 "t1" --------------------------- Newsgroups: comp.os.vxworks,ba.jobs Subject: VxWorks Network Protocol Development Engineer Date: Mon, 04 Jun 2001 23:16:18 GMT From: "The Vista Group" Organization: EarthLink Inc. -- http://www.EarthLink.net Keywords: VxWorks,BGP,OSPF, MPLS, Software Engineer,Career Message-ID: <6FUS6.2133$fI.216823@newsread1.prod.itd.earthlink.net> South San Francisco Bay Area Software Engineering Position Develop network protocol software with this hot optical networking company. Requirements: Experience with developing and integrating Networking Protocols: RIP, OSPF, BGP, IS-IS, MPLS,Virtual Router Must have strong programming skills in C Experience with embedded software developement, RTOS such as VxWorks is a plus B.S. (or) M.S. in computer Science with minimum 5+ years of work experience in any or all of the relevant areas. For more information contact Shawn Moonan Manager resumes@searchvista.com The Vista Group 310-937-9049 --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Problem of Transmit Descriptor Date: Mon, 04 Jun 2001 23:37:33 GMT From: "Chacha Hindustani" Organization: Road Runner Message-ID: <1ZUS6.21713$zl5.6731249@typhoon.ne.mediaone.net> References: <3B1C0439.9000906@free.fr> are you talking about socket descriptor?? try to increase it in vxWorks image by selecting networking protocols->core TCP/IP components->BSD Sockets and double click and change param "NUM_FILE" to 128. Hoping it will help . Gaurav "Tony Rodayer" wrote in message news:3B1C0439.9000906@free.fr... > Hy, > Our product crashes when we increase the Number of Transmit Descriptor > up to 128. > Does anyone can help me? > I have got this message : > data access > Exception current instruction address: 0x0004397c > Machine Status Register: 0x00003030 > Data Access Register: 0x21724011 > Condition Register: 0x44242082 > Data storage interrupt Register: 0x00003030 > Task: 0x1ce48e0 "t1" > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Use of VxWorks Exception Task - "tExcTask" Date: Tue, 05 Jun 2001 09:30:38 +0100 From: Vijay Kumar Peshkar Organization: Cisco Systems Inc. Message-ID: <3B1C98AE.6683056A@wipro.com> References: <6d87d2a2.0106041032.bd6b0cd@posting.google.com> This is a multi-part message in MIME format. - --------------484D53954B04F0F12C69F853 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Also tExcTask can be used by Interrupt Service Routines to facilitate execution of code at task level. (Code executed at ISRs is executed at the Interrupt level) _func_excJobAdd should be called at interrupt level with the correct parameters so that a request is sent to tExcTask. This communication from the ISR to the tExctask is done using message Queues. Since ISRs cannot WAIT for the message queues to be available, they just throw it into the queue (NO_WAIT) and hope it reaches the tExctask. In a system receiving large number of interrupts, there is a possibility that this might get lost. cheers, Vijay Jim Lally wrote: > Using Tornado 2.0 - VxWorks 5.4 - Motorolla 68K platforms > > I'm trying to identify all uses of the vxWorks exceptions task > "tExcTask". Any Help would be appreciated. > > Currently identified uses are: > > · Exception Handling Processing that must be done in a task context! > Possibly when "excHookAdd" used? > > · Watchdog Timer support! > > · Debugging support! Setting up of breakpoints, etc.? > > · Use of Unix like signaling mechanism! Delivery of signals to > multiple tasks? Signaling handler executing in tExcTask task context? > > · Library dependencies are: dbgLib (Debugger), sigLib (Signal > Library), and intLib (hardware specific interrupt library). > > · Supports a task's deletion by executing taskDelete Hooks for tasks > deleting themselves. > > Any help would be appreciated. - --------------484D53954B04F0F12C69F853 Content-Type: text/x-vcard; charset=us-ascii; name="vijay.peshkar.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Vijay Kumar Peshkar Content-Disposition: attachment; filename="vijay.peshkar.vcf" begin:vcard n:Peshkar;Vijay Kumar tel;work:5530034/35/35 x2237 x-mozilla-html:FALSE org:Wipro Global R & D;Wishing you blue skies, a couple of daisies, some fresh air and a smile! adr:;;;;;; version:2.1 email;internet:vpeshkar@cisco.com note:Wishing you blue skies, a couple of daisies, some fresh air and a smile! fn:Vijay Kumar Peshkar end:vcard - --------------484D53954B04F0F12C69F853-- --------------------------- Newsgroups: comp.os.vxworks Subject: boot vxWorks from harddisk on mcp750 Date: 4 Jun 2001 21:36:57 -0700 From: ellin_lin@263.net (ellin) Organization: http://groups.google.com/ Message-ID: <44a01eba.0106042036.7df252c8@posting.google.com> I wonder if someone can list the steps of boot vxWorks from harddisk on mcp750. Thx in advance. ellin --------------------------- Newsgroups: comp.os.vxworks Subject: Re: connecting modem to vxworks Date: 4 Jun 2001 22:04:13 -0700 From: sarveshku@lycos.com (Sarvesh) Organization: http://groups.google.com/ Message-ID: References: borkhuis@agere.com (Johan Borkhuis) wrote in message news:... > sarveshku@lycos.com (Sarvesh) wrote: > > > HI > > I want to connect my VxWorks machine to internet through modem > > dialing. But I am not able to write to modem. Can anyone tell me how > > to initialize the modem. TR LED in the Modem is not glowing. I have > > done pppInit() also but still it is not dialing. My modem is connected > > to COM port2. Please suggest me some procedure. > > As far as I know PPP does not do the dialing. You have to initialize the > modem and send the dial-commands to it. When the connection is started you > can start the PPP service. > > Groeten, > Johan Hi Johan Thanks for the reply. But I am not able to send the dialup commands to modem. Can you tell me the precedures to write to modem. My modem is connected to the COM port2 of the target machine. When I am writing to the serial port it is not going to the modem. Kindly tell me the procedures. With regards. Sarvesh --------------------------- Newsgroups: comp.os.vxworks Subject: Re: ElfHex Tool Date: Tue, 05 Jun 2001 10:42:55 +0100 From: Vijay Kumar Peshkar Organization: Cisco Systems Inc. Message-ID: <3B1CA99F.1D73F6E5@wipro.com> References: <5.0.2.1.0.20010604155618.00a63080@alpha.ccii.co.za> This is a multi-part message in MIME format. - --------------63B7DA9706AA3C8CFEBB1BEF Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Hi Arne, Sorry havn't used the tool. Still giving my view on this ;) S9 is a termination record for a block of S1 records. The address field may optionally contain the 2/3/4-byte address of the instruction to which control is passed. For ROM data the S9 address field is usually 000000. There is no code/data field which you need to copy, execute as in other S-records. So normally 000000 should do. Any specific reason for 10000 ?. Typically a S9 record looks like.. S9030000FC 03-size, 0000-address, FC-checksum. Hope this helps, bye, Vijay Arne Henrichsen wrote: > Hi, > > A question about the elfHex tool from vxWorks which converts the vxWorks > object file to a Motorola s-record file. The last entry of the s-record > file is always S9........ The address field of S9 contains the starting > execution address. Mine is always 000000. I want it to be 100000. Is that > part of the elfHex tool or do I need to set that up in my BSP??? > > Thanks > > ---------------------------------------------------------------- > Arne Henrichsen - Software Design Engineer > C²I² Systems (Pty) Ltd > Tel: (+27 21) 683 5490 > Fax: (+27 21) 683 5435 > Cell: (+27 82) 872 0773 > Email: ahenric@ccii.co.za > Url: http://www.ccii.co.za > > -- > Posted from alpha.ccii.co.za [66.8.24.68] > via Mailgate.ORG Server - http://www.Mailgate.ORG - --------------63B7DA9706AA3C8CFEBB1BEF Content-Type: text/x-vcard; charset=us-ascii; name="vijay.peshkar.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Vijay Kumar Peshkar Content-Disposition: attachment; filename="vijay.peshkar.vcf" begin:vcard n:Peshkar;Vijay Kumar tel;work:5530034/35/35 x2237 x-mozilla-html:FALSE org:Wipro Global R & D;Wishing you blue skies, a couple of daisies, some fresh air and a smile! adr:;;;;;; version:2.1 email;internet:vpeshkar@cisco.com note:Wishing you blue skies, a couple of daisies, some fresh air and a smile! fn:Vijay Kumar Peshkar end:vcard - --------------63B7DA9706AA3C8CFEBB1BEF-- --------------------------- Newsgroups: comp.os.vxworks Subject: tNetTask Error Date: 4 Jun 2001 23:30:57 -0700 From: m.maio@myrealbox.com (Massimo Maio) Organization: http://groups.google.com/ Message-ID: Hi VxWorkers! Our system is composed by three CPUs (a Master CPU and two slaves CPU). The tNetTask agent running on the master CPU from time to time produces the following error: "cksum: out of data" Has anyone of you already found this kind of error? Thanks in advance. Bye, Massimo. --------------------------- Newsgroups: comp.os.vxworks Subject: re: pipeline problem on PowerPC Date: Tue, 5 Jun 2001 09:08:38 +0200 From: "Michael Lawnick" Organization: Lawrence Berkeley National Laboratory Message-ID: <9fi0u7$sc1$1@overload.lbl.gov> Hi Chris, From: "Chris Slominski" Newsgroups: comp.os.vxworks Sent: Friday, June 01, 2001 9:10 PM Subject: pipeline problem on PowerPC > VxWorks Users, > I am porting code from a mv167 to a mv2700. I configured the board, > recompiled the application and ran. I got an "interrupt: bad VME interrupt > 0" error immediately and was told by a co-worker that this may be due to the > mv2700 instruction pipeline executing my interrupt reset instruction after > the return from the interrupt handler. I've heard of an "eieio" instruction > that can be used force sequential execution over a set of instructions. > Please let me know if this is indeed a likely candidate for my interrupt > crash, and what are good choices to correct the problem. > > Thanks, > Chris I fear this alone won't solve your problem. Standard configuration uses a 'posted write', i.e. Universe will complete write cycle independantly from CPU. Your EIEIO will only ensure, that the write to Universe will have completed before next instruction execution. The access to your VME-H/W will still happen, when the bus is aritrated. The CPU may have completed some jobs 'till then. The secure way is a read-back the IACK value. This way the CPU is forced to wait for completion. The second method (with more performance impact) is to switch off 'posted write'. HTH Mit freundlichen Grüßen, Michael Lawnick ============================================== SOFTEC GmbH Tel +49-731-96600-0 Promenade 17 Fax +49-731-96600-23 D-89073 Ulm Michael Lawnick Germany lawnick@softec.de ============================================== --------------------------- Newsgroups: comp.os.vxworks Subject: Re: connecting modem to vxworks Date: 5 Jun 2001 07:48:07 GMT From: borkhuis@agere.com (Johan Borkhuis) Organization: Agere System Message-ID: References: sarveshku@lycos.com (Sarvesh) wrote: > Thanks for the reply. But I am not able to send the dialup commands to > modem. Can you tell me the precedures to write to modem. My modem is > connected to the COM port2 of the target machine. When I am writing to > the serial port it is not going to the modem. Kindly tell me the > procedures. Have you checked that you can write to the serial port, using a terminal program to see the output? If that does not work you have to check your serial driver. 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: Booting problem Date: Tue, 5 Jun 2001 10:03:35 +0100 From: "Moy, Ian" Organization: Lawrence Berkeley National Laboratory Message-ID: <9fi7v7$2ut$1@overload.lbl.gov> This is a multi-part message in MIME format. - --------------InterScan_NT_MIME_Boundary Content-Type: text/plain; charset="iso-8859-1" Hi, I am trying to set-up a Motorola MBX860 board for vxWorks and have a problem. I have successfully built the bootrom and downloaded that to the board but when I try and start the board booting vxWorks I get the following error: Attaching network interface cpm0... done. Attaching network interface lo0... done. Loading... Error loading file: errno = 0xd0003 The board does not make contact to the FTP server at all and I have checked the ethernet port, cables and all the boot parameters are correct. What else could be wrong? Any help would be gratefully received, Regards, Ian Moy - --------------InterScan_NT_MIME_Boundary Content-Type: text/plain; name="RMRL-Disclaimer.txt" Content-Transfer-Encoding: 7bit Content-Disposition: attachment; filename="RMRL-Disclaimer.txt" The information contained in this e-mail is confidential to Roke Manor and must not be passed to any third party without permission. This communication is for information only and shall not create or change any contractual relationship. - --------------InterScan_NT_MIME_Boundary-- --------------------------- Newsgroups: comp.protocols.snmp,comp.lang.c,comp.os.psos,comp.os.vxworks Subject: C-code generated with mibcomp Date: Tue, 05 Jun 2001 12:21:00 +0200 From: Alex Vinokur Message-ID: <3B1CB28A.8564D436@bigfoot.com> ============================ pRISM+ for pSOSystem/PowerPC pRISM+ 1.2.3 . pSOSystem 2.2.2 Diab Data D-CC Optimizing C compiler SNMP Release 7.4 (SNMPv1) Epilogue Technology Emissary SNMP MIB Compiler, version 6.3 Host : Windows98 Target : PowerPC (MPC860) ============================ mibcomp (SNMP MIB Compiler) has been run with the -stub modes : %mibcomp -stub -o stub.c my.mib my.ctl A relevant piece of the code is shown below. Note. mibcomp's comments omitted. My comment added. // ---------- File stub.c (fragment) --------- void pna_next(OIDC_T lastmatch, int compc, OIDC_T *compl, SNMP_PKT_T *pktp, VB_T *vbp) { STRUCT_pna data; OIDC_T instance = 0; if (compc != 0) nextproc_no_next(pktp, vbp); else if (pna_lookup(&data)); // ### Suspicious line ### getproc_error(pktp, vbp, GEN_ERR); else { for (group_by_getproc_and_instance(pktp, vbp, compc, compl); vbp; vbp = vbp->vb_link) { nextproc_next_instance(pktp, vbp, 1, &instance); pna_get_value(vbp->vb_ml.ml_last_match, pktp, vbp, &data); } } } // ---------------------------------------------- A C-compiler detects error (1633): parse error near 'else' in the following line --------------------------------- else if (pna_lookup(&data)); --------------------------------- It seems that - the C-compiler is correct - mibcomp is producing illegal C-code in the line above (is semicolon needless ?). Did the mibcomp mean as following ? : - ------------------------------------------------ else if (pna_lookup(&data)) getproc_error(pktp, vbp, GEN_ERR); - ------------------------------------------------ ================================== Alex Vinokur mailto:alexvn@bigfoot.com http://up.to/alexvn http://visitweb.com/alexvn.math ================================== --------------------------- Newsgroups: comp.protocols.snmp,comp.lang.c,comp.os.psos,comp.os.vxworks Subject: Re: C-code generated with mibcomp Date: Tue, 05 Jun 2001 11:44:31 +0200 From: willem veenhoven Organization: Veenhoven Bouwinformatica Message-ID: <3B1CA9FF.6A4BC94A@veenhoven.com> References: <3B1CB28A.8564D436@bigfoot.com> Alex Vinokur wrote: > > if (compc != 0) > nextproc_no_next(pktp, vbp); > else if (pna_lookup(&data)); // ### Suspicious line ### > getproc_error(pktp, vbp, GEN_ERR); > else { ... } > } > > Did the mibcomp mean as following ? : > ------------------------------------------------ > else if (pna_lookup(&data)) > getproc_error(pktp, vbp, GEN_ERR); > ------------------------------------------------ Of course. The empty statement ';' closes the if .. else, so you have effectively written: if () do_something(); else if (); getproc_error(); else { } where the last else does not refer to any previous if. willem --------------------------- Newsgroups: comp.os.vxworks Subject: Socket creation problem Date: 5 Jun 2001 02:42:37 -0700 From: nick.gardner@europe.simoco.com (nick) Organization: http://groups.google.com/ Message-ID: Hello, I am having some very annoying problems creating sockets on a bluewave systems c6400 card. Depending on the size of my SO_SNDBUF, I am unable to create more than a few sockets. Currently my SO_SNDBUF is set to 1024bytes (confirmed with getsockopt). My test software spawns a task which creates a client socket (server running on a FORCE 680 card in same CPCI rack) which connects and then writes until blocked. When the task has blocked the parent process spawns the same function, so on and so forth. The first 3 tasks spawned create and connect their socket without a problem and block after the correct amount of data has been written (1k send buffer + 8k receive buffer on other card). The fourth task and socket blocks after about 2.5k of data has been wrote. After this no other sockets will connect. I know it is not the server because lowering the send buffer to 100bytes allows 6 sockets to be created. I have checked all the mbuf / mblk and cluster stats with netStackSysPoolShow() and netStackDataPoolShow(). I have absolutely loads of free everything. NUM_FILES is set to 150, and iosFdShow() shows that I am no where near that limit. I'm lost as to where to go from here. Any help much apreciated. Regards, Nick --------------------------- Newsgroups: comp.os.vxworks Subject: problem with timeout Date: Tue, 5 Jun 2001 11:55:03 +0200 From: "Jonathan ILIAS" Organization: University of Angers, France. Message-ID: <9fia8i$2jcd$1@news.univ-angers.fr> Hi, I need to use timeouts with semaphores. semTake returns ERROR after the timeout but errno equals S_ioLib_UNKNOWN_REQUEST (0xC0002) instead of S_objLib_OBJ_TIMEOUT (0x3D0004) as it is said in documentation. Any idea about this ? Thanks, Jonathan ILIAS --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Booting problem Date: Tue, 5 Jun 2001 13:11:39 +0100 From: "TSIIEL Lycée M Leroy Evreux" Organization: Wanadoo, l'internet avec France Telecom Message-ID: <9fic5v$cuf$1@wanadoo.fr> References: <9fi7v7$2ut$1@overload.lbl.gov> >Hi, > >I am trying to set-up a Motorola MBX860 board for vxWorks and have a problem. I have successfully built the bootrom and downloaded that to the board but when I try and start the board booting vxWorks I get the following error: > > Attaching network interface cpm0... done. > Attaching network interface lo0... done. > Loading... > Error loading file: errno = 0xd0003 > >The board does not make contact to the FTP server at all and I have checked the ethernet port, cables and all the boot parameters are correct. What else could be wrong? > >Any help would be gratefully received, > >Regards, > >Ian Moy > Have you entered the name and the password on the target and on the server? Jack --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Booting problem Date: Tue, 05 Jun 2001 16:08:54 +0100 From: Vijay Kumar Peshkar Organization: Cisco Systems Inc. Message-ID: <3B1CF606.D0AAB7FD@wipro.com> References: <9fi7v7$2ut$1@overload.lbl.gov> This is a multi-part message in MIME format. - --------------EFBD9791E27E408C6463387B Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Ian, The problem is your boot image is unable to locate the vxworks image. Check if #define DEFAULT_BOOT_LINE in config.h is having the right path, image name. good luck, Vijay "Moy, Ian" wrote: > This is a multi-part message in MIME format. > > --------------InterScan_NT_MIME_Boundary > Content-Type: text/plain; > charset="iso-8859-1" > > Hi, > > I am trying to set-up a Motorola MBX860 board for vxWorks and have a problem. I have successfully built the bootrom and downloaded that to the board but when I try and start the board booting vxWorks I get the following error: > > Attaching network interface cpm0... done. > Attaching network interface lo0... done. > Loading... > Error loading file: errno = 0xd0003 > > The board does not make contact to the FTP server at all and I have checked the ethernet port, cables and all the boot parameters are correct. What else could be wrong? > > Any help would be gratefully received, > > Regards, > > Ian Moy > > --------------InterScan_NT_MIME_Boundary > Content-Type: text/plain; > name="RMRL-Disclaimer.txt" > Content-Transfer-Encoding: 7bit > Content-Disposition: attachment; > filename="RMRL-Disclaimer.txt" > > The information contained in this e-mail is confidential to Roke Manor and must not be > passed to any third party without permission. This communication is for information only > and shall not create or change any contractual relationship. > > --------------InterScan_NT_MIME_Boundary-- - --------------EFBD9791E27E408C6463387B Content-Type: text/x-vcard; charset=us-ascii; name="vijay.peshkar.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Vijay Kumar Peshkar Content-Disposition: attachment; filename="vijay.peshkar.vcf" begin:vcard n:Peshkar;Vijay Kumar tel;work:5530034/35/35 x2237 x-mozilla-html:FALSE org:Wipro Global R & D;Wishing you blue skies, a couple of daisies, some fresh air and a smile! adr:;;;;;; version:2.1 email;internet:vpeshkar@cisco.com note:Wishing you blue skies, a couple of daisies, some fresh air and a smile! fn:Vijay Kumar Peshkar end:vcard - --------------EFBD9791E27E408C6463387B-- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Booting problem Date: 5 Jun 2001 10:45:43 GMT From: borkhuis@agere.com (Johan Borkhuis) Organization: Agere System Message-ID: References: <9fi7v7$2ut$1@overload.lbl.gov> "Moy, Ian" wrote: > > This is a multi-part message in MIME format. > > --------------InterScan_NT_MIME_Boundary > Content-Type: text/plain; > charset="iso-8859-1" > > > Hi, > > I am trying to set-up a Motorola MBX860 board for vxWorks and have a > problem. I have successfully built the bootrom and downloaded that to > the board but when I try and start the board booting vxWorks I get the > following error: > > Attaching network interface cpm0... done. > Attaching network interface lo0... done. > Loading... > Error loading file: errno = 0xd0003 Check the log of your FTP server. This usually makes it a lot clearer. Groeten, Johan - -- o o o o o o o . . . _____________________________ o _____ || Johan Borkhuis | .][__n_n_|DD[ ====_____ | borkhuis@agere.com | >(________|__|_[_________]_|__________________________| _/oo OOOOO oo` ooo ooo 'o!o!o o!o!o` === VxWorks FAQ: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html === --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Tue Jun 5 06:42:11 2001 From: "Aditi Kapoor" Date: Tue Jun 5 06:42:13 PDT 2001 Subject: netjobadd problem hi, i have been developing a wlan driver in vxworks but am stuck at a crucial location. Somehow, when i do a netjobadd on a rx intr it gives me a status of OK , but never is the function given to the netjobadd invoked. i've seen that the tNetTask is running as i can see it in the active task list, but i dont understand why isnt my function added to its queue through netjobadd ever invoked. i'd be grateful if u culd help me. thanx & regards aditi From vxwexplo-errs@csg.lbl.gov Tue Jun 5 07:42:00 2001 From: "Moy, Ian" Date: Tue Jun 5 07:42:03 PDT 2001 Subject: Launching tornado problem This is a multi-part message in MIME format. --------------InterScan_NT_MIME_Boundary Content-Type: text/plain; charset="iso-8859-1" Hi, After successfully getting my board to boot I tried to connect to it using tornado. When I try I get the following errors in the launch log: Command: tgtsvr mbx860_ism -L -C -A -V -s -n mbx860 -display jw-pc:0 -Bt 30 -Br 30 -m 8000000 tgtsvr (mbx860@hex): Tue Jun 5 15:22:13 2001 License request... authorized on host 'whel'. Attaching backend... succeeded. Connecting to target agent... succeeded. Attaching C++ interface... succeeded. Error: Couldn't open /usr/mbx860/vxWorks: status = 0x2 Error: Unable to initialize object module loader library I am quite confused as to why tornado needs to open /usr/mbx860/vxWorks, this is the path I use in the vxWorks bootloader to load the VxWorks image onto the board, but why does tornado need to know this?? Any ideas gratefully received, Regards, Ian Moy --------------InterScan_NT_MIME_Boundary Content-Type: text/plain; name="RMRL-Disclaimer.txt" Content-Transfer-Encoding: 7bit Content-Disposition: attachment; filename="RMRL-Disclaimer.txt" The information contained in this e-mail is confidential to Roke Manor and must not be passed to any third party without permission. This communication is for information only and shall not create or change any contractual relationship. --------------InterScan_NT_MIME_Boundary-- From vxwexplo-errs@csg.lbl.gov Tue Jun 5 08:11:40 2001 From: "Mike Anderson" Date: Tue Jun 5 08:11:43 PDT 2001 Subject: RE: Launching tornado problem VxWorks Greetings! > Submitted-by: "Moy, Ian" > > > > I am quite confused as to why tornado needs to open > /usr/mbx860/vxWorks, this is the path I use in the vxWorks > bootloader to load the VxWorks image onto the board, but why does > tornado need to know this?? > The target server is trying to set up the host-based symbol table. So, it needs to get to the VxWorks image that you booted to extract the symbols. There is an option in the target server configuration that specifies the location of the "core" file. If you don't specify a path to the VxWorks image here, it tries to extract that information from the target. As you've found, the target's view of the file systems and the target server's view don't always match up. HTH, Mike -- ======================================================= T h e P T R G r o u p, I n c. ======================================================= Embedded, Real-Time Solutions and Training Mike Anderson mailto:mike@ThePTRGroup.com Chief Scientist http://www.ThePTRGroup.com V: (703) 585-9384 F: (703) 430-3748 ======================================================= "Software development is like making a baby... You can't make a baby in one month by impregnating nine women. Some things just take time." From vxwexplo-errs@csg.lbl.gov Tue Jun 5 08:13:41 2001 From: sanjay krishnan Date: Tue Jun 5 08:13:44 PDT 2001 Subject: VxSim Sir, i would like to know whether we can input data from VxSim window. i'm using tornado prototyper with VxSim. i need to input data from the VxSim window in the following program: #include "vxWorks.h" #include "taskLib.h" #include "stdio.h" int taskid1; int s; void func1(void); STATUS name(void) { taskid1=taskSpawn("mytask1",100,0,1000,(FUNCPTR)func1,0,0,0,0,0,0,0,0,0,0); return(OK); } void func1(void) { printf("\ninput char:"); s=getchar(); printf("\nchar is:%c",s); } sanjay. ____________________________________________________________________ Get free email and a permanent address at http://www.netaddress.com/?N=1 From vxwexplo-errs@csg.lbl.gov Tue Jun 5 09:40:23 2001 From: David Laight Date: Tue Jun 5 09:40:26 PDT 2001 Subject: Re: Problem Pinging Other OS's > I have written a device driver for a wireless LAN card, and I am > having problems pinging a linux system from a VxWorks shell. The > linux system is using the same wireless LAN card. I would double check the length of the packets sent and received. It is also (just) possible you have a byte order problem! One of my ethernet drivers has the following: /* set to read packet size followed by data */ tad_ae_set_bap1( ð->ae_sel_off, eth->ae_rx_fid, 0x36 ); rx_length = eth->ae_data[1] + 12; if (rx_length < 17 || rx_length > 1514 || !ta->te.te_end.receiveRtn) { /* illegal packet - just skip */ mp = 0; } else { /* allocate receive buffer */ mp = netTupleGet( ta->te.te_end.pNetPool, 1516, M_DONTWAIT, MT_DATA, FALSE ); if (mp) { uint16 *p; /* 32bit align the IP header => frame on odd 16bit boundary */ p = (void *)mp->mBlkHdr.mData; p++; mp->mBlkHdr.mData = (void *)p; mp->mBlkHdr.mLen = rx_length; mp->mBlkHdr.mFlags |= M_PKTHDR; /* or arp bleats */ mp->mBlkPktHdr.len = rx_length; rx_length = (rx_length + 1) >> 1; do { *p++ = eth->ae_data[1]; } while (--rx_length); } else ta->ta_me.me_rx_lost++; } Works fine... David ---------------------------------------------------------------- David Laight email: dsl@tadpole.co.uk Tadpole Technology plc phone: +44 1223 428 232 Cambridge, UK fax: +44 1223 428 201 From vxwexplo-errs@csg.lbl.gov Tue Jun 5 11:24:04 2001 From: gokulnath kollayikal Date: Tue Jun 5 11:24:06 PDT 2001 Subject: tcl problems hi! am a newbie to both tcl/tk and vxworks. I am running vxsim and from the tcl shell am not able to run the exec command. my output when i run the following is given below a)lsearch : a tcl command , though with wrong number of parameters b)ls : using the catch statement c)main: function in a project that I have compiled and downloaded into the vxsim simulator.note : can call main directly from the shell, so i know it works. thanks in advance for all your help gokul tcl> exec lsearch couldn't execute "lsearch": no such file or directory tcl> if [catch {exec ls } output] { tcl> tcl> # Code to execute if error occurs tcl> tcl> puts "Error occurred: $output" tcl> tcl> } Error occurred: couldn't execute "ls": no such file or directory tcl> if [catch {exec main } output] { tcl> tcl> # Code to execute if error occurs tcl> tcl> puts "Error occurred: $output" tcl> tcl> } Error occurred: couldn't execute "main": no such file or directory ciao gokul __________________________________________________ Do You Yahoo!? Get personalized email addresses from Yahoo! Mail - only $35 a year! http://personal.mail.yahoo.com/ From vxwexplo-errs@csg.lbl.gov Tue Jun 5 22:11:13 2001 From: Rashmi Topno Date: Tue Jun 5 22:11:15 PDT 2001 Subject: Booting of CPCI/C6400 Hi, Could somebody please clarify the following ? In my system (cpci/c6400; MPC860) I need to set the ether net interface address during run time and later initiate socket communication through this interface. Could somebody give me some clue as to how this can be done ? I am planing to do this by invoking ifMaskSet() and ifAddrSet()... Do we require to reboot the vxworks / reset the board after setting the ether net interface address ? How do I get the interface name ? Thanks and best regards, rajendra. __________________________________________________ Do You Yahoo!? Get personalized email addresses from Yahoo! Mail - only $35 a year! http://personal.mail.yahoo.com/ From vxwexplo-errs@csg.lbl.gov Tue Jun 5 23:01:08 2001 From: chen wei seow Date: Tue Jun 5 23:01:10 PDT 2001 Subject: Debugger Hi, I'm new to Vxworks environment. Can any one help to solve my problem. Step1 : download vxworks.st image into target board Step2 : Launch target server Step3 : Launch debugger Step4 : Set the "Source Search Path" Step5 : Open the source code for example test.c file Step6 : set breakpoint in the test.c file Result : An error --> "Symbol table information for file, 'test.c' not found" instead. Is there any option I didn't aware ? thanks __________________________________________________ Do You Yahoo!? Get personalized email addresses from Yahoo! Mail - only $35 a year! http://personal.mail.yahoo.com/ From vxwexplo-errs@csg.lbl.gov Wed Jun 6 04:03:28 2001 From: Vxworks Exploder Date: Wed Jun 6 04:03:30 PDT 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Wed Jun 6 04:03:09 PDT 2001 Subject: Re: Socket creation problem Subject: Using timer_create() and related routines..... Subject: Re: Some problem with C++ Subject: Re: Some problem with C++ Subject: Re: unable to send udp packet on 255.255.255.255 Subject: interface address Subject: mibcomp : get and set function in stub.c Subject: netjobadd problem Subject: Communication Problems with MVME5100 Subject: Re: hanging in serial I/O Subject: Re: TCP Theory question puzzling me Subject: Re: problem with serial ports Subject: Launching tornado problem Subject: RE: Launching tornado problem Subject: VxSim Subject: SSH server for VxWorks Subject: Interrupt handling with compact PCI Subject: [Fwd: LoadCoreFileCheck() function call interface] Subject: Re: problem with timeout Subject: re: Problem Pinging Other OS's Subject: Re: ElfHex Tool Subject: Re: Problem of Transmit Descriptor Subject: Re: help please, very urgent Subject: Re: unable to send udp packet on 255.255.255.255 Subject: tcl problems Subject: Re: Some problem with C++ Subject: Re: tNetTask Error Subject: Re: MVME2604 lockup problem Subject: rename() does not work with dosFs2? Subject: Re: Communication Problems with MVME5100 Subject: Re: mibcomp : get and set function in stub.c Subject: Re: netjobadd problem Subject: Re: Communication Problems with MVME5100 Subject: Re: [Fwd: LoadCoreFileCheck() function call interface] Subject: Booting of CPCI/C6400 Subject: cc386 compiler crash, anybody seen this error. Subject: Debugger Subject: I can't do the "make" prcedure with snmp??? Subject: Re: rename() does not work with dosFs2? Subject: Re: Launching tornado problem Subject: Re: Debugger Subject: network interface ip address setting during run time ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Socket creation problem Date: Tue, 05 Jun 2001 16:37:29 +0100 From: Vijay Kumar Peshkar Organization: Cisco Systems Inc. Message-ID: <3B1CFCB8.FD6CBC2E@wipro.com> References: This is a multi-part message in MIME format. - --------------7CD466857472E27C0360540E Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Nick, You might also just try changing the globals that control the following default parameters to larger numbers (all the way upto 48K): tcp_sendspace (default 4K) tcp_recvspace (default 4K) udp_sendspace (default 2K) udp_recvspace (default 4K) Hope this helps, regards, Vijay nick wrote: > Hello, > > I am having some very annoying problems creating sockets on a bluewave > systems c6400 card. Depending on the size of my SO_SNDBUF, I am > unable to create more than a few sockets. > > Currently my SO_SNDBUF is set to 1024bytes (confirmed with > getsockopt). My test software spawns a task which creates a client > socket (server running on a FORCE 680 card in same CPCI rack) which > connects and then writes until blocked. When the task has blocked the > parent process spawns the same function, so on and so forth. The > first 3 tasks spawned create and connect their socket without a > problem and block after the correct amount of data has been written > (1k send buffer + 8k receive buffer on other card). The fourth task > and socket blocks after about 2.5k of data has been wrote. After this > no other sockets will connect. > > I know it is not the server because lowering the send buffer to > 100bytes allows 6 sockets to be created. > > I have checked all the mbuf / mblk and cluster stats with > netStackSysPoolShow() and netStackDataPoolShow(). I have absolutely > loads of free everything. > > NUM_FILES is set to 150, and iosFdShow() shows that I am no where near > that limit. > > I'm lost as to where to go from here. > > Any help much apreciated. > > Regards, > > Nick - --------------7CD466857472E27C0360540E Content-Type: text/x-vcard; charset=us-ascii; name="vijay.peshkar.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Vijay Kumar Peshkar Content-Disposition: attachment; filename="vijay.peshkar.vcf" begin:vcard n:Peshkar;Vijay Kumar tel;work:5530034/35/35 x2237 x-mozilla-html:FALSE org:Wipro Global R & D;Wishing you blue skies, a couple of daisies, some fresh air and a smile! adr:;;;;;; version:2.1 email;internet:vpeshkar@cisco.com note:Wishing you blue skies, a couple of daisies, some fresh air and a smile! fn:Vijay Kumar Peshkar end:vcard - --------------7CD466857472E27C0360540E-- --------------------------- Newsgroups: comp.os.vxworks Subject: Using timer_create() and related routines..... Date: Tue, 5 Jun 2001 16:24:55 +0530 From: Priya Pradhan Organization: Lawrence Berkeley National Laboratory Message-ID: <9fif08$5mq$1@overload.lbl.gov> > Hi! > > I am trying to create a periodic timer on VxWorks using the routines - > timer_create(), timer_connect(), timer_settime(). Everytime I try to run > the program I get the error - "kill failed" - I can't understand the > error. > > The code is included below..... > > #include > #include > #include > #include > #include > > extern int errno ; > > void my_handler(int param) ; > > timer_t timer_id ; > > int timers(void) > { > int i = 0 ; > struct timespec tp_gettime, tp_settime, tp_clkres ; > struct itimerspec time_val ; > > clock_getres(CLOCK_REALTIME, &tp_clkres) ; > > clock_gettime(CLOCK_REALTIME, &tp_gettime) ; > > printf("tp_gettime.tv_sec == %d\n",tp_gettime.tv_sec) ; > printf("tp_gettime.tv_nsec == %d\n",tp_gettime.tv_nsec) ; > > tp_settime.tv_sec = tp_clkres.tv_sec ; > tp_settime.tv_nsec = tp_clkres.tv_nsec ; > > clock_settime(CLOCK_REALTIME, &tp_settime) ; > > if(timer_create(CLOCK_REALTIME, NULL, &timer_id) == ERROR) > { > printf("timer :: Error number set is == %d\n",errno) ; > exit(1) ; > } > else > printf("created timer.... \n") ; > > if(timer_connect(timer_id, my_handler, 0) == ERROR) > { > printf("timer :: Error no set is == %d\n",errno) ; > exit(0) ; > } > else > printf("connected timer.... \n") ; > > time_val.it_interval.tv_sec = 1 ; > time_val.it_interval.tv_nsec = 0 ; > time_val.it_value.tv_sec = 1 ; > time_val.it_value.tv_nsec = 0 ; > > if(timer_settime(timer_id, CLOCK_REALTIME, &time_val, NULL) == ERROR) > { > printf("timer :: Error number set is == %d\n",errno) ; > exit(1) ; > } > else > printf("settime for timer.... \n") ; > > clock_gettime(CLOCK_REALTIME, &tp_gettime) ; > > printf("tp_gettime.tv_sec == %d\n",tp_gettime.tv_sec) ; > printf("tp_gettime.tv_nsec == %d\n",tp_gettime.tv_nsec) ; > > while(i != 1000000) > i++ ; > > return 0 ; > } > > void my_handler(int param) > { > struct itimerspec time_val ; > > printf("..........in function............ \n") ; > fflush(stdout) ; > > time_val.it_interval.tv_sec = 1 ; > time_val.it_interval.tv_nsec = 0 ; > time_val.it_value.tv_sec = 1 ; > time_val.it_value.tv_nsec = 0 ; > > if(timer_settime(timer_id, CLOCK_REALTIME, &time_val, NULL) == ERROR) > { > printf("my_handler :: Error number set is == %d\n",errno) ; > exit(0) ; > } > > return ; > } > > Hoping someone will be able to help. > > Thanks, > Priya. > Priya Pradhan > Member of Technical Staff. > Net Brahma Technologies Pvt. Ltd. > Internext Networking Software > Ph: 5521451 (Extn 212) > http://www.netbrahma.com > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Some problem with C++ Date: Tue, 5 Jun 2001 12:04:47 +0100 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: <3B1800D4.FAE5B1DC@hotmail.com> <3B1BC22E.7040403@snellwilcox.com> <3B1BD0D0.B0BA892@hotmail.com> Beginner wrote in message <3B1BD0D0.B0BA892@hotmail.com>... > >The reason is that there are some prolem in my out object initialization order >that causes a crush (right on the out.open(filename) call) if I execute the code >using dynamic loading. If I execute the code straight (i.e. not via DL) it is >OK. That is when I digged out that article and try to see if that is the >problem. Are you sure your makefile is executing the much stage correctly? Without munching a C++ module, any static C++ class objects in that file will not be constructed when the DL finishes loading the module. DaveK - -- They laughed at Galileo. They laughed at Copernicus. They laughed at Columbus. But remember, they also laughed at Bozo the Clown. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Some problem with C++ Date: Tue, 5 Jun 2001 12:05:20 +0100 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: <3B1800D4.FAE5B1DC@hotmail.com> <9ffr3e$h7l$1@inn.jlab.org> <3B1BBD6E.FCE05642@hotmail.com> <9fglr8$hm1$1@inn.jlab.org> Chris Slominski wrote in message <9fglr8$hm1$1@inn.jlab.org>... >Do you really want to create the ofstream object on the heap? Why not He's trying to call the constructor while supplying the address of the existing object in order to re-initialise it in place. DaveK - -- They laughed at Galileo. They laughed at Copernicus. They laughed at Columbus. But remember, they also laughed at Bozo the Clown. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: unable to send udp packet on 255.255.255.255 Date: Tue, 5 Jun 2001 12:11:43 +0100 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: <92b087a4.0106020857.7e2438e6@posting.google.com> Paresh Jain wrote in message <92b087a4.0106020857.7e2438e6@posting.google.com>... >I am writing a DHCP client on VxWorks, and facing the problem in >sending a udp packet to 255.255.255.255 . >As per requirement of DHCP I have to set my IP address to 0.0.0.0 and >then hv to send the DHCP DISCOVER request >on 255.255.255.255 . But it fails in 'sendto', returning 0x41 errno. Before you can broadcast from a socket, you need to enable broadcasting from that socket using the setsockopt function: (from VxWorks Reference Manual : Libraries : sockLib) - --8<----snip!----8<-- OPTION FOR DATAGRAM SOCKETS The following section discusses an option for datagram (UDP) sockets. SO_BROADCAST -- Sending to Multiple Destinations Specify the SO_BROADCAST option when an application needs to send data to more than one destination: setsockopt (sock, SOL_SOCKET, SO_BROADCAST, &optval, sizeof (optval)); The value at optval is an integer (type int), either 1 (on) or 0 (off). - -->8----snip!---->8-- Did you try this yet? DaveK - -- They laughed at Galileo. They laughed at Copernicus. They laughed at Columbus. But remember, they also laughed at Bozo the Clown. --------------------------- Newsgroups: comp.os.vxworks Subject: interface address Date: Tue, 05 Jun 2001 17:44:03 +0530 From: rajendra Organization: Lucent Technologies Message-ID: <3B1CCD0B.BE9F7306@lucent.com> vxworks tornado Hi, Could somebody please clarify the following ? In my system (cpci/c6400; MPC860) I need to set the ether net interface address during run time and later initiate socket communication through this interface. Could somebody give me some clue as to how this can be done ? I am planing to do this by invoking ifMaskSet() and ifAddrSet()... Do we require to reboot the vxworks / reset the board after setting the ether net interface address ? How do I get the interface name ? Thanks and best regards, rajendra. --------------------------- Newsgroups: comp.protocols.snmp,comp.os.psos,comp.os.vxworks Subject: mibcomp : get and set function in stub.c Date: Tue, 05 Jun 2001 16:52:07 +0200 From: Alex Vinokur Message-ID: <3B1CF217.E28AC965@bigfoot.com> //========================================= See http://pii3.brahms.bnl.gov/~daq/external_manuals/windriver/docs/snmp/supplement/x-mibcomp.html //========================================= File stub.c generated with mibcomp contains the following functions : - static int xxx_get_value (OIDC_T, SNMP_PKT_T*, VB_T*, STRUCT_xxx*) - void xxx_get(OIDC_T, int, OIDC_T*, VB_T*, SNMP_PKT_T*) - void xxx_set(OIDC_T, int, OIDC_T*, VB_T*, SNMP_PKT_T*) We can see that different approaches are used to generate xxx_get() and xxx_set() : a.1) xxx_get_value() calculates (retrieves) requested values : getproc_got_type (pktp, vbp, data->leaf_of_xxx); Note. data is of STRUCT_xxx* a.2) xxx_get() merely calls xxx_get_value() b) xxx_set() calculates (sets) requested values : leaf_of_xxx = VB_GET_TYPE(vbp); /* !!! */ So, to get some leaf_value mibcomp generates data->leaf_of_xxx, to set the same leaf_value mibcomp generates merely leaf_of_xxx. Is there any essential reason for that? Note. I realize that we have to change stub.c, however (I think) it is not a reason. //========================================= See http://pii3.brahms.bnl.gov/~daq/external_manuals/windriver/docs/snmp/supplement/x-mibcompOut.html //========================================= //--------- File stub.c : BEGIN --------- [omitted] #include #include #include #include #include #include #include #include #include "mibleaf.h" /* * Method routines for the ex1Objs variables: * * ex1Obj1 -- read-write * First example object. * * ex1Obj2 -- read-only * Second example object. */ /* An internal routine to retrieve the values of the variables, used * by the method routines ex1Objs_get and ex1Objs_next. * You need to replace the type STRUCT_ex1Objs with something * appropriate to your system. */ static int ex1Objs_get_value(OIDC_T lastmatch, SNMP_PKT_T *pktp, VB_T *vbp, STRUCT_ex1Objs *data) /* !!! */ { switch(lastmatch) { case LEAF_ex1Obj1: /* Values: * one(1) = VAL_ex1Obj1_one * two(2) = VAL_ex1Obj1_two */ getproc_got_int32(pktp, vbp, data->ex1Obj1); /* !!! */ break; case LEAF_ex1Obj2: /* if the data being returned is in dynamic storage and needs * to be free'd, change the 0 argument to a 1. */ getproc_got_string(pktp, vbp, string_length(data->ex1Obj2), data->ex1Obj2, 0, VT_STRING); /* !!! */ break; default: return GEN_ERR; } return NO_ERROR; } void ex1Objs_get(OIDC_T lastmatch, int compc, OIDC_T *compl, VB_T *vbp, SNMP_PKT_T *pktp) { STRUCT_ex1Objs data; /* !!! */ int error; /* find all the varbinds that share the same getproc and instance */ group_by_getproc_and_instance(pktp, vbp, compc, compl); /* check that the instance is exactly .0 */ if (!((compc == 1) && (*compl == 0))) for ( ; vbp; vbp = vbp->vb_link) getproc_nosuchins(pktp, vbp); /* grab the actual data for this variable. this lookup routine will * probably have to be changed for your system. for scalar variables * there might not even be any lookup routine. */ else if (ex1Objs_lookup(&data) != 0) /* !!! */ for ( ; vbp; vbp = vbp->vb_link) getproc_error(pktp, vbp, GEN_ERR); else { /* retrieve all the values from the same data structure */ for ( ; vbp; vbp = vbp->vb_link) { if ((error = ex1Objs_get_value(vbp->vb_ml.ml_last_match, pktp, vbp, &data)) != NO_ERROR) getproc_error(pktp, vbp, error); } } } [omitted] void ex1Objs_set(OIDC_T lastmatch, int compc, OIDC_T *compl, VB_T *vbp, SNMP_PKT_T *pktp) { for ( ; vbp; vbp = vbp->vb_link) { switch (vbp->vb_ml.ml_last_match) { case LEAF_ex1Obj1: ex1Obj1 = VB_GET_INT32(vbp); /* !!! */ setproc_good(pktp, vbp); break; default: setproc_error(pktp, vbp, COMMIT_FAILED); return; } } } [omitted] //--------- File stub.c : END ----------- Was it possible to create mibcomp generating set-functions as following ? : //--------- "updated" set : BEGIN ------------ static int ex1Objs_set_value(OIDC_T lastmatch, SNMP_PKT_T *pktp, VB_T *vbp, STRUCT_ex1Objs *data) /* !!! */ { switch(lastmatch) { case LEAF_ex1Obj1: data->ex1Obj1 = VB_GET_INT32(vbp); /* !!! */ break; default: return GEN_ERR; } return NO_ERROR; } void ex1Objs_get(OIDC_T lastmatch, int compc, OIDC_T *compl, VB_T *vbp, SNMP_PKT_T *pktp) { STRUCT_ex1Objs data; /* !!! */ // ex1Objs_set stuff using ex1Objs_set_value // (similar to ex1Objs_get using ex1Objs_get_value) //--------- "updated" set : END -------------- ================================== Alex Vinokur mailto:alexvn@bigfoot.com http://up.to/alexvn http://visitweb.com/alexvn.math ================================== --------------------------- Newsgroups: comp.os.vxworks Subject: netjobadd problem Date: Tue, 5 Jun 2001 19:22:01 +0530 From: "Aditi Kapoor" Organization: Lawrence Berkeley National Laboratory Message-ID: <9fiphn$9bj$1@overload.lbl.gov> hi, i have been developing a wlan driver in vxworks but am stuck at a crucial location. Somehow, when i do a netjobadd on a rx intr it gives me a status of OK , but never is the function given to the netjobadd invoked. i've seen that the tNetTask is running as i can see it in the active task list, but i dont understand why isnt my function added to its queue through netjobadd ever invoked. i'd be grateful if u culd help me. thanx & regards aditi --------------------------- Newsgroups: comp.os.vxworks Subject: Communication Problems with MVME5100 Date: Tue, 5 Jun 2001 09:19:38 -0500 From: "Owen Landgren" Organization: Honeywell Technology Center, Honeywell Inc. Message-ID: <9fipok$67i$1@chico.htc.honeywell.com> We're working with a Motorola MVME5100 board, which supposedly had a standard vxWorks/PPC-Bug split install already done. However, we're running into some very interesting problems. First, when we connect to the debug port, using an RJ45 to Serial cable, the output on the teletype is just garbage. Neither PPC-Bug nor vxWorks provides readable output. However, what is even odder is that the board will respond to commands such as CNFG in PPC-Bug mode, and will print more garbage to the port. The garbage does not seem to have any sort of patterned or shifted offset from standard ASCII. The computer we're connecting the MVME5100 to is a Win2K SP1 machine, and we're using the standard Hilgraeve HyperTerminal to do so. The settings all match what the documentation describes -- 9600 Baud, 8-None-1. We've also experimented with a lot of different settings within HyperTerminal, to no avail. Has anyone else had experience with this problem or found a way to solve it? Any suggestion would be greatly appreciated -- we've tried everything we can come up with. Owen Landgren --------------------------- Newsgroups: comp.os.vxworks Subject: Re: hanging in serial I/O Date: 5 Jun 2001 08:01:09 -0700 From: aparker@vtlinx.com (parreg) Organization: http://groups.google.com/ Message-ID: <79ef8b60.0106050701.1793231d@posting.google.com> References: <9e12k0$tjg$1@inn.jlab.org> "Chris Slominski" wrote in message news:<9e12k0$tjg$1@inn.jlab.org>... > VxWorks Users, > My task runs for long periods, up to two weeks, perfoming regular serial > I/O to a device every three seconds. It stops communicating because it ends > up waiting for a semaphore. The chain of ststem calls that were last made > are as follows. > > write()->iosWrite()->tyWrite()->semTake()->semBTake() > > It appears to never take the semaphore. The serial driver must be > compromised. Has anyone else encountered such a problem on an MV2700 > processor, running VxWorks 5.3.1? make sure the shell task or windshell task are preventing the semaphore from being taken. parreg --------------------------- Newsgroups: comp.os.vxworks Subject: Re: TCP Theory question puzzling me Date: Tue, 5 Jun 2001 11:05:26 -0400 From: "Kevin Livingston" Message-ID: References: <3b17c930@news.si.com> you can still use DHCP in a somewhat static way, and not defeat the purpose of DHCP you can assign IP's based on the MAC id of the NIC - this will require you to register MAC ids for each NIC to your DHCP server, but really, how often do you introduce new NICs into your system. And it would only require a little bit of time to do so (on top of the time you are already spending to put the nic in the first place). There are even automated ways of setting up your DHCP settings for this depending on your server - eg scripts to register in the MAC id etc.. This gives you two benifits - security - non registered NICs don't get IP's, and also provides added accountability. you don't need to change ANY client, it just runs DHCP - and great it gets the same IP everytime, it doesn't care. - now you get all the advantages of static IP's with almost the simplicity of pure blind DHCP. your DNS can then be set up as normal with specific IP's since you know where they go. if you want to swap move names you now have two places to do it - alter which IP address is doled out by the DHCP or change DNS as normal. the main point here is that you control where IP's go without having to configure clients. Kevin "Geoff" wrote in message news:pwSR6.22314$zb7.2860294@news1.cableinet.net... > Hi > > As you say this static approach is not very elegant as it somewhat negates > the purpose of DHCP in the first place. > I have got a bit more info on this question now since I first asked . The > solution to this problem is Dynamic DNS, this sounds like it is something > that is still fairly new and standards are still evolving. In simple terms > the DNS server and the DHCP Server talk to each other in order to keep the > DNS's client information upto date in real time. Here is a link that > explains it in more detail > > http://www.join.com/doc/server.4.html > > Regards Geoff > > > > > > vanbaren_gerald wrote in message > news:3b17c930@news.si.com... > > "Geoff" writes: > > > > >Hello > > > > >I am developing an embedded application, which requires DHCP client > > >functionality (Not Intel/Windows based). It was fairly easy to establish > an > > >IP address using the DHCP call provided by the Operating System. > However, > > >what is puzzling me is, now that I have a dynamic IP address, how do > other > > >machines address this machine ? I would like them to access this machine > by > > >host name, but what mechanism do I use, to tell the DNS server my name > and > > >DHCP'd IP address? Any help or insight into this problem would be greatly > > >appreciated ? Thanks > > > > > Geoff > > > > > > You typically would use a fixed address served by your DHCP server (it > > should have the capability to associate a MAC address with a static > > IP address rather than allocating it from the pool). At that point, > > you can put the static IP address in your DNS server host tables. > > > > Not very elegant. Anybody have a more elegant approach? > > > > gvb > > > > -- > > > +--------------------------------------------------------------------------- > + > > | Jerry Van Baren / vanbaren_gerald@si.com / Grand Rapids Mi / > 616-241-7973 | > > | My employer is a company. Companies are artifacts of a legal system. > | > > |________________Artifacts are incapable of having > opinions.________________| > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: problem with serial ports Date: 5 Jun 2001 08:04:07 -0700 From: aparker@vtlinx.com (parreg) Organization: http://groups.google.com/ Message-ID: <79ef8b60.0106050704.5206df2@posting.google.com> References: <01bd01c0df8d$af5fd550$2401880a@sammy> sami.ellonen@syspex.fi ("Sami Ellonen") wrote in message news:<01bd01c0df8d$af5fd550$2401880a@sammy>... > Hi, > I'm having a problem with the 2nd serial port in MPC860. I'm using a > Release 2.0 BSP by Embedded Planet (and Tornado2). The problem is that I > cannot get it (the 2nd port) initialized correctly. (Atleast I think > that is the problem.) fd=open("/tyCo/1",O_RDWR,0) allways returns > error. I read the addwise given here, but I'm using a different BSP so > it was not the help I needed. > > I appreciate any addwise given... :) > > ======================== > Sami Ellonen > E-mail: sami.ellonen@NOSPAM.syspex.fi > ======================== > remove NOSPAM. if e-mailing me... > > -- does the device show up in response to a "devs" command? --------------------------- Newsgroups: comp.os.vxworks Subject: Launching tornado problem Date: Tue, 5 Jun 2001 15:41:48 +0100 From: "Moy, Ian" Organization: Lawrence Berkeley National Laboratory Message-ID: <9fit29$b9k$1@overload.lbl.gov> This is a multi-part message in MIME format. - --------------InterScan_NT_MIME_Boundary Content-Type: text/plain; charset="iso-8859-1" Hi, After successfully getting my board to boot I tried to connect to it using tornado. When I try I get the following errors in the launch log: Command: tgtsvr mbx860_ism -L -C -A -V -s -n mbx860 -display jw-pc:0 -Bt 30 -Br 30 -m 8000000 tgtsvr (mbx860@hex): Tue Jun 5 15:22:13 2001 License request... authorized on host 'whel'. Attaching backend... succeeded. Connecting to target agent... succeeded. Attaching C++ interface... succeeded. Error: Couldn't open /usr/mbx860/vxWorks: status = 0x2 Error: Unable to initialize object module loader library I am quite confused as to why tornado needs to open /usr/mbx860/vxWorks, this is the path I use in the vxWorks bootloader to load the VxWorks image onto the board, but why does tornado need to know this?? Any ideas gratefully received, Regards, Ian Moy - --------------InterScan_NT_MIME_Boundary Content-Type: text/plain; name="RMRL-Disclaimer.txt" Content-Transfer-Encoding: 7bit Content-Disposition: attachment; filename="RMRL-Disclaimer.txt" The information contained in this e-mail is confidential to Roke Manor and must not be passed to any third party without permission. This communication is for information only and shall not create or change any contractual relationship. - --------------InterScan_NT_MIME_Boundary-- --------------------------- Newsgroups: comp.os.vxworks Subject: RE: Launching tornado problem Date: Tue, 5 Jun 2001 11:12:15 -0400 From: "Mike Anderson" Organization: Lawrence Berkeley National Laboratory Message-ID: <9fit2e$b9l$1@overload.lbl.gov> VxWorks Greetings! > Submitted-by: "Moy, Ian" > > > > I am quite confused as to why tornado needs to open > /usr/mbx860/vxWorks, this is the path I use in the vxWorks > bootloader to load the VxWorks image onto the board, but why does > tornado need to know this?? > The target server is trying to set up the host-based symbol table. So, it needs to get to the VxWorks image that you booted to extract the symbols. There is an option in the target server configuration that specifies the location of the "core" file. If you don't specify a path to the VxWorks image here, it tries to extract that information from the target. As you've found, the target's view of the file systems and the target server's view don't always match up. HTH, Mike - -- ======================================================= T h e P T R G r o u p, I n c. ======================================================= Embedded, Real-Time Solutions and Training Mike Anderson mailto:mike@ThePTRGroup.com Chief Scientist http://www.ThePTRGroup.com V: (703) 585-9384 F: (703) 430-3748 ======================================================= "Software development is like making a baby... You can't make a baby in one month by impregnating nine women. Some things just take time." --------------------------- Newsgroups: comp.os.vxworks Subject: VxSim Date: 5 Jun 2001 09:13:37 MDT From: sanjay krishnan Organization: Lawrence Berkeley National Laboratory Message-ID: <9fit2n$b9m$1@overload.lbl.gov> Sir, i would like to know whether we can input data from VxSim window. i'm using tornado prototyper with VxSim. i need to input data from the VxSim window in the following program: #include "vxWorks.h" #include "taskLib.h" #include "stdio.h" int taskid1; int s; void func1(void); STATUS name(void) { taskid1=taskSpawn("mytask1",100,0,1000,(FUNCPTR)func1,0,0,0,0,0,0,0,0,0,0); return(OK); } void func1(void) { printf("\ninput char:"); s=getchar(); printf("\nchar is:%c",s); } sanjay. ____________________________________________________________________ Get free email and a permanent address at http://www.netaddress.com/?N=1 --------------------------- Newsgroups: comp.os.vxworks Subject: SSH server for VxWorks Date: 5 Jun 2001 08:26:21 -0700 From: yoavg@p-cube.com (Yoav) Organization: http://groups.google.com/ Message-ID: <61809bc9.0106050726.2eb4987d@posting.google.com> hello, I'm interested in porting a SSH server to VxWorks. does anyone know of a free or commercial SSH server that is portable to VxWorks ? thanks in advance, Yoav --------------------------- Newsgroups: comp.os.vxworks Subject: Interrupt handling with compact PCI Date: Tue, 05 Jun 2001 15:33:41 GMT From: Daniel Flück Organization: Customers of Tiscali DataComm AG - http://www.tiscalinet.ch/ Message-ID: <1103_991755221@stockholm> I should implement interrupt handling with VxWorks on a mcpn765 board through compact PCI. How can I find as a novice through the jungle? Where are the important registers to trigger the interrupt? How can I set the ISR and enable it (and test it)? Sorry, if I'm at the wrong place here. But I couldn't find anything than cryptic explanations everywhere -- in my point of view. Thanks for any advise or link. Daniel --------------------------- Newsgroups: comp.os.vxworks Subject: [Fwd: LoadCoreFileCheck() function call interface] Date: Tue, 05 Jun 2001 11:33:09 -0400 From: Xiong Xue Organization: AlcaNET Canada Message-ID: <3B1CFBB5.391E9127@alcatel.com> - --------------DFC6652274C21ECDCF31EF7A Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Xiong Xue wrote: > From: Xiong Xue > company: Alcatel Canada > License No. 204957 > > ----------------- > > Would you please raise a TSR for me on the following issues. > > Background: In unix environment, a tcl-based script runs on host > machine, started target server, which connected to a target agent > which runs in lab, via wtx/wdb/tcp/ip. > > We all know that when target server launches, it establishes a > connection to target agent by wdb/tcp/ip, then it does a corefile > checksum check between the test header section on host target server > and image on target. The function called is LoadCoreFileCheck(). > > In our script, we need to run the corefile checksum check after the > connection initialization stage in a controlled, interactive manner. > i.e., I'd like to do the checksum called within the script, and I want > to capture the return value, then make program flow decision based on > the return value. > > Tornado provided 4 types of user interfaces to access target server, > via WTX protocol. they are: Tornado shell, wtx tcl API, wtx C api and > wtx java api. The one I am interested in is the wtxtcl. > > The problem: However, I checked all wtxtcl libraries, and couldn't > find a procedure which is invoking the corefile checksum check. The > closest one I get is wtxObjModuleChecksum(), which, according to its > description in online docs, should do the job, however, gave 0 return > no matter what corefiles were specified. (should return "Core File > Checksum not match" error when the corefile on host does not match the > target, as the function LoadCoreFileCheck() does). > > Would you please assign a engineer to help me on this? > > Thanks for help. > > Regards, > > -- > Xiong Xue, ATM Development Technologies, CID Alcatel > email: xiong.xue@alcatel.com > phone 720-4416 (CID internal) 613-784-4416 (external) > > - --------------DFC6652274C21ECDCF31EF7A Content-Type: text/html; charset=us-ascii Content-Transfer-Encoding: 7bit Xiong Xue wrote:
From: Xiong Xue
company: Alcatel Canada
License No. 204957 
-----------------
Would you please raise a TSR for me on the following issues.

Background: In unix environment, a tcl-based script runs on host machine, started target server, which connected to a target agent which runs in lab, via wtx/wdb/tcp/ip.

We all know that when target server launches, it establishes a connection to target agent by wdb/tcp/ip, then it does a corefile checksum check between the test header section on host target server and image on target. The function called is LoadCoreFileCheck().

In our script, we need to run the corefile checksum check after the connection initialization stage in a controlled, interactive manner. i.e., I'd like to do the checksum called within the script, and I want to capture the return value, then make program flow decision based on the return value.

Tornado provided 4 types of user interfaces to access target server, via WTX protocol.  they are: Tornado shell, wtx tcl API, wtx C api and wtx java api. The one I am interested in is the wtxtcl.

The problem: However, I checked all wtxtcl libraries, and couldn't find a procedure which is invoking the corefile checksum check. The closest one I get is wtxObjModuleChecksum(), which, according to its description in online docs, should do the job, however, gave 0 return no matter what corefiles were specified. (should return "Core File Checksum not match" error when the corefile on host does not match the target, as the function LoadCoreFileCheck() does).

Would you please assign a engineer to help me on this?

Thanks for help.

Regards, 

-- 
Xiong Xue, ATM Development Technologies, CID Alcatel
email: xiong.xue@alcatel.com
phone 720-4416 (CID internal) 613-784-4416 (external)
 
- --------------DFC6652274C21ECDCF31EF7A-- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: problem with timeout Date: Tue, 05 Jun 2001 09:56:25 -0600 From: Don Small Organization: Sandia National Laboratories Message-ID: <3B1D0129.BBAF23FF@sandia.gov> References: <9fia8i$2jcd$1@news.univ-angers.fr> Just some ideas. How did you find out what the errno was? Did you read it from the program immediately after the return from semTake or is this from the shell (with "i"). Don't forget that this is a global variable (for each task) and that it can be overwritten by another part of your task. As an example, if your task continued after the errant semTake, a different function could have failed and overwrote the errno. Jonathan ILIAS wrote: > > Hi, > > I need to use timeouts with semaphores. semTake returns ERROR after the > timeout but errno equals S_ioLib_UNKNOWN_REQUEST (0xC0002) instead of > S_objLib_OBJ_TIMEOUT (0x3D0004) as it is said in documentation. > Any idea about this ? > > Thanks, > Jonathan ILIAS --------------------------- Newsgroups: comp.os.vxworks Subject: re: Problem Pinging Other OS's Date: Tue, 5 Jun 2001 17:39:53 +0100 (BST) From: David Laight Organization: Lawrence Berkeley National Laboratory Message-ID: <200106051639.RAA15386@dsl-2.tadpole.co.uk> > I have written a device driver for a wireless LAN card, and I am > having problems pinging a linux system from a VxWorks shell. The > linux system is using the same wireless LAN card. I would double check the length of the packets sent and received. It is also (just) possible you have a byte order problem! One of my ethernet drivers has the following: /* set to read packet size followed by data */ tad_ae_set_bap1( ð->ae_sel_off, eth->ae_rx_fid, 0x36 ); rx_length = eth->ae_data[1] + 12; if (rx_length < 17 || rx_length > 1514 || !ta->te.te_end.receiveRtn) { /* illegal packet - just skip */ mp = 0; } else { /* allocate receive buffer */ mp = netTupleGet( ta->te.te_end.pNetPool, 1516, M_DONTWAIT, MT_DATA, FALSE ); if (mp) { uint16 *p; /* 32bit align the IP header => frame on odd 16bit boundary */ p = (void *)mp->mBlkHdr.mData; p++; mp->mBlkHdr.mData = (void *)p; mp->mBlkHdr.mLen = rx_length; mp->mBlkHdr.mFlags |= M_PKTHDR; /* or arp bleats */ mp->mBlkPktHdr.len = rx_length; rx_length = (rx_length + 1) >> 1; do { *p++ = eth->ae_data[1]; } while (--rx_length); } else ta->ta_me.me_rx_lost++; } Works fine... David - ---------------------------------------------------------------- David Laight email: dsl@tadpole.co.uk Tadpole Technology plc phone: +44 1223 428 232 Cambridge, UK fax: +44 1223 428 201 --------------------------- Newsgroups: comp.os.vxworks Subject: Re: ElfHex Tool Date: 5 Jun 2001 13:17:07 -0400 From: vanbaren@falcon.si.com (vanbaren_gerald) Organization: Smiths Industries Message-ID: <3b1d1413$1@news.si.com> References: <5.0.2.1.0.20010604155618.00a63080@alpha.ccii.co.za> ahenric@ccii.co.za (Arne Henrichsen) writes: >Hi, >A question about the elfHex tool from vxWorks which converts the vxWorks=20 >object file to a Motorola s-record file. The last entry of the s-record=20 >file is always S9........ The address field of S9 contains the starting=20 >execution address. Mine is always 000000. I want it to be 100000. Is that=20 >part of the elfHex tool or do I need to set that up in my BSP??? >Thanks >---------------------------------------------------------------- >Arne Henrichsen - Software Design Engineer >C=B2I=B2 Systems (Pty) Ltd >Tel: (+27 21) 683 5490 >Fax: (+27 21) 683 5435 >Cell: (+27 82) 872 0773 >Email: ahenric@ccii.co.za >Url: http://www.ccii.co.za Try the -a option. gvb elfHex NAME elfHex - convert a ELF format object file into Motorola hex records SYNOPSIS elfHex [-a adrs] [-l] [-v] [-p PC] [-s SP] file DESCRIPTION This program generates a Motorola hex format (S-record) file from a ELF format object module. Normally, the entry address in the object module is used as the starting address of the output module. However, if the - -a flag is used, then adrs is used as the starting address. Normally S1 records are generated for addresses less than 64K; S2 records are generated for addresses greater than 64K and less the 16M; and S3 records are generated for addresses greater than 16M. OPTIONS - -l generate only S2 records. - -a adrs use adrs as the entry address, rather than the address in the object module. - -v output vector information at address 0. - -p PC use PC as the PC in the vector (meaningless without -v). - -s SP use SP as the SP in the vector (meaningless without -v). - -- +---------------------------------------------------------------------------+ | Jerry Van Baren / vanbaren_gerald@si.com / Grand Rapids Mi / 616-241-7973 | | My employer is a company. Companies are artifacts of a legal system. | |________________Artifacts are incapable of having opinions.________________| --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Problem of Transmit Descriptor Date: Tue, 05 Jun 2001 18:12:44 GMT From: Tony Rodayer Organization: Guest of ProXad - France Message-ID: <3B1D217B.80602@free.fr> References: <3B1C0439.9000906@free.fr> <1ZUS6.21713$zl5.6731249@typhoon.ne.mediaone.net> No I am not talking about socket descriptor but descriptor and data buffers for host communication. Chacha Hindustani wrote: > are you talking about socket descriptor?? > > try to increase it in vxWorks image by selecting > networking protocols->core TCP/IP components->BSD Sockets and double click > and change param "NUM_FILE" to 128. > > Hoping it will help . > Gaurav > > > "Tony Rodayer" wrote in message > news:3B1C0439.9000906@free.fr... > >> Hy, >> Our product crashes when we increase the Number of Transmit Descriptor >> up to 128. >> Does anyone can help me? >> I have got this message : >> data access >> Exception current instruction address: 0x0004397c >> Machine Status Register: 0x00003030 >> Data Access Register: 0x21724011 >> Condition Register: 0x44242082 >> Data storage interrupt Register: 0x00003030 >> Task: 0x1ce48e0 "t1" >> --------------------------- Newsgroups: comp.os.vxworks Subject: Re: help please, very urgent Date: Tue, 05 Jun 2001 11:14:30 -0700 From: "Michael S. Simpson" Organization: Raytheon Company Message-ID: <3B1D2186.AF2458A1@west.raytheon.com> References: <3ACD0DF6.B75B1E14@hotmail.com> This is a multi-part message in MIME format. - --------------184FEC43E7F5C67D03FC1FF0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit The output can be seen on a workstation if you have a serial device enabled on your board. If you are using a PC as a host machine, just open a hyperterminal window and run an application on the target board. You will see your output in the terminal window. Ram wrote: > > hi > do you know if there is a way to connect a monitor to the mtx603 > target board. If yes, please tell me how to. I want to see my output on > the monitor. > i will be grateful for your information > thank you > Ram - --------------184FEC43E7F5C67D03FC1FF0 Content-Type: text/x-vcard; charset=us-ascii; name="mssimpson.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Michael S. Simpson Content-Disposition: attachment; filename="mssimpson.vcf" begin:vcard n:Simpson;Michael tel;pager:520-489-7933 tel;home:520-622-5039 tel;work:520-794-5381 x-mozilla-html:FALSE adr:;;;;;; version:2.1 email;internet:mssimpson@west.raytheon.com fn:Michael Simpson end:vcard - --------------184FEC43E7F5C67D03FC1FF0-- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: unable to send udp packet on 255.255.255.255 Date: 5 Jun 2001 11:26:09 -0700 From: pareshj@yahoo.com (Paresh Jain) Organization: http://groups.google.com/ Message-ID: <92b087a4.0106051026.453e2a98@posting.google.com> References: <92b087a4.0106020857.7e2438e6@posting.google.com> "Dave Korn" wrote in message news:... > Before you can broadcast from a socket, you need to enable broadcasting > from that socket using the setsockopt function: (from VxWorks Reference > Manual : Libraries : sockLib) > > --8<----snip!----8<-- > OPTION FOR DATAGRAM SOCKETS > The following section discusses an option for datagram (UDP) sockets. > > SO_BROADCAST -- Sending to Multiple Destinations > Specify the SO_BROADCAST option when an application needs to send data to > more than one destination: > setsockopt (sock, SOL_SOCKET, SO_BROADCAST, &optval, sizeof (optval)); > The value at optval is an integer (type int), either 1 (on) or 0 (off). > -->8----snip!---->8-- > > Did you try this yet? > > DaveK Thanks Dave, I am already doing this before sending packet, but it fails. One more thing I wanna add is I am using two interfaces on my target . And what I doubt is mRouteEntryAdd fails because it doesn't know for which interface this entry is(As we don't have any IP address assigned to any of these interfaces). - --Paresh --------------------------- Newsgroups: comp.os.vxworks Subject: tcl problems Date: Tue, 5 Jun 2001 11:24:02 -0700 (PDT) From: gokulnath kollayikal Organization: Lawrence Berkeley National Laboratory Message-ID: <9fjb4c$l56$1@overload.lbl.gov> hi! am a newbie to both tcl/tk and vxworks. I am running vxsim and from the tcl shell am not able to run the exec command. my output when i run the following is given below a)lsearch : a tcl command , though with wrong number of parameters b)ls : using the catch statement c)main: function in a project that I have compiled and downloaded into the vxsim simulator.note : can call main directly from the shell, so i know it works. thanks in advance for all your help gokul tcl> exec lsearch couldn't execute "lsearch": no such file or directory tcl> if [catch {exec ls } output] { tcl> tcl> # Code to execute if error occurs tcl> tcl> puts "Error occurred: $output" tcl> tcl> } Error occurred: couldn't execute "ls": no such file or directory tcl> if [catch {exec main } output] { tcl> tcl> # Code to execute if error occurs tcl> tcl> puts "Error occurred: $output" tcl> tcl> } Error occurred: couldn't execute "main": no such file or directory ciao gokul __________________________________________________ Do You Yahoo!? Get personalized email addresses from Yahoo! Mail - only $35 a year! http://personal.mail.yahoo.com/ --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Some problem with C++ Date: Tue, 05 Jun 2001 16:22:22 -0400 From: Beginner Organization: UUNET Canada News Reader Service Message-ID: <3B1D3F7E.7EB3F339@hotmail.com> References: <3B1800D4.FAE5B1DC@hotmail.com> <3B1BC22E.7040403@snellwilcox.com> <3B1BD0D0.B0BA892@hotmail.com> Hi, Dave, That is a point I never thought about, now how do you munch a C++ module? Any help is appreciated. TIA. Dave Korn wrote: > Beginner wrote in message <3B1BD0D0.B0BA892@hotmail.com>... > > > > >The reason is that there are some prolem in my out object initialization > order > >that causes a crush (right on the out.open(filename) call) if I execute the > code > >using dynamic loading. If I execute the code straight (i.e. not via DL) it > is > >OK. That is when I digged out that article and try to see if that is the > >problem. > > Are you sure your makefile is executing the much stage correctly? Without > munching a C++ module, any static C++ class objects in that file will not be > constructed when the DL finishes loading the module. > > DaveK > -- > They laughed at Galileo. They laughed at Copernicus. They laughed at > Columbus. But remember, they also laughed at Bozo the Clown. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: tNetTask Error Date: Tue, 05 Jun 2001 15:26:25 -0500 From: Andrew Johnson Organization: APS, Argonne National Laboratory Message-ID: <3B1D4071.1D7F71CD@aps.anl.gov> References: Massimo Maio wrote: > > Our system is composed by three CPUs (a Master CPU and two slaves > CPU). The tNetTask agent running on the master CPU from time to time > produces the following error: > > "cksum: out of data" > > Has anyone of you already found this kind of error? You're running the Proxy ARP Server on the master, which IIRC causes this when some DHCP clients on your network send out something. Check WindSurf for an SPR which mentions this (put the above error into their search box) and gives some hints on how to resolve it. - - Andrew - -- The world is such a cheerful place when viewed from upside-down It makes a rise of every fall, a smile of every frown --------------------------- Newsgroups: comp.os.vxworks Subject: Re: MVME2604 lockup problem Date: Tue, 05 Jun 2001 23:57:49 GMT From: William Dennen Organization: Motorola Computer Group Message-ID: <3B1D7289.8000706@motorola.com> References: Reply-To: william.dennen@motorola.com Keith Kingsley wrote: > We're having a problem with the MVME2604 occasionally locking up on > VME reads/writes to a certain card (specifically, a PMAC II VME > Ultralite). It appears that when the lockup occurs, the MVME2604 > receives the expected DTACK from the PMAC board, dutifully releases > the data strobe lines (DS1 and DS2), but then fails to release the > address strobe line (AS). The MVME2604 then locks up with the bottom > four status lights lit. When the event happens, it also appears that > the 2604 board responds unusually quickly to the DTACK-- say, 19 ns > instead of the usual ~50 ns. Has anyone seen anything like this > before? We suspect there is a problem with the Universe II chip. Our > 2604 board has CA91C142C-33CE, although we've tried another 2604 with > CA91C142-33CE installed and the same problem happens. > > -Keith Kingsley > Aeroflex Lintek If you have a trace of the event I'd be glad to look at it although the response time will be quicker if you present it through your distribution FAE ... Regards - -- >@< Bill Dennen william.dennen@motorola.com Motorola Computer Group --GTAC --------------------------- Newsgroups: comp.os.vxworks Subject: rename() does not work with dosFs2? Date: 5 Jun 2001 17:13:29 -0700 From: hariharan@foobox.com (Thiagarajan Hariharan) Organization: http://groups.google.com/ Message-ID: <7f10f842.0106051613.3c92083d@posting.google.com> Hello, Thanks to those who responded to the previous "weird dosFs problem" question. Got dosFs2 from wind river and installed it. So far so good, except that rename() does not work. It sets errno 0x388002, which is S_dosFsLib_FILE_NOT_FOUND. Has anyone had the same experience? My current workaround is to copy the file and delete the original. Another annoyance with dosFs2 is that in prjConfig.c, it insists on mounting hard-drive 1 also (using usrAtaConfig()) even if there isn't a second hard-drive in the system. Also, mounting the second non-existant drive does not fail when vxworks starts up - the second drive shows up as "/ata1a" shows up in the 'devs' list. You cannot have prjConfig.c call usrAtaInit() alone without calling usrAtaConfig() for drives 0 and 1. Thanks, Hari --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Communication Problems with MVME5100 Date: Wed, 06 Jun 2001 00:20:00 GMT From: "Kelly Hornsby" Organization: Excite@Home - The Leader in Broadband http://home.com/faster Message-ID: References: <9fipok$67i$1@chico.htc.honeywell.com> loop your 2-3 and 3-2 (db9) to yourself and see if you see the characters correctly. If you do, then your cable is good, and perhaps the output drivers for the maxim chip are blasted. BTW my 5100 works ok with just 2,3 and gnd --------------------------- Newsgroups: comp.protocols.snmp,comp.os.psos,comp.os.vxworks Subject: Re: mibcomp : get and set function in stub.c Date: Tue, 05 Jun 2001 19:25:55 -0500 From: Jim Jones Organization: netINS, Inc. Message-ID: <3B1D7892.B7213A55@icon-labs.com> References: <3B1CF217.E28AC965@bigfoot.com> Reply-To: jim_jones@icon-labs.com Alex Vinokur wrote: > //========================================= > See > http://pii3.brahms.bnl.gov/~daq/external_manuals/windriver/docs/snmp/supplement/x-mibcomp.html > > //========================================= > File stub.c generated with mibcomp contains > the following functions : > - static int xxx_get_value (OIDC_T, SNMP_PKT_T*, VB_T*, STRUCT_xxx*) > - void xxx_get(OIDC_T, int, OIDC_T*, VB_T*, SNMP_PKT_T*) > - void xxx_set(OIDC_T, int, OIDC_T*, VB_T*, SNMP_PKT_T*) > > We can see that different approaches are used > to generate xxx_get() and xxx_set() : > a.1) xxx_get_value() calculates (retrieves) requested values : > getproc_got_type (pktp, vbp, data->leaf_of_xxx); > Note. data is of STRUCT_xxx* > a.2) xxx_get() merely calls xxx_get_value() > > b) xxx_set() calculates (sets) requested values : > leaf_of_xxx = VB_GET_TYPE(vbp); /* !!! */ This is one of the things you have to fix. The stub.c generated code is buggy in this regard. The !!! identifies places you are likely to have to fix. Clearly you must decide if these are global variables or are fields of a data structure pointed to by "data". Once you decide, then you must make everything consistent. The conflict exists with the scalar related methods, not the table ones. > > > So, to get some leaf_value mibcomp generates data->leaf_of_xxx, > to set the same leaf_value mibcomp generates merely leaf_of_xxx. > > Is there any essential reason for that? > Note. I realize that we have to change stub.c, > however (I think) it is not a reason. Correct. It is cleary a mistake. The Epilogue team apparently felt there were other things that were more important to work on... and since we have to alter this code anyway... it would be left for us to do. JJ > > > //========================================= > See > http://pii3.brahms.bnl.gov/~daq/external_manuals/windriver/docs/snmp/supplement/x-mibcompOut.html > > //========================================= > //--------- File stub.c : BEGIN --------- > > [omitted] > > #include > #include > #include > #include > #include > #include > #include > #include > #include "mibleaf.h" > > /* > * Method routines for the ex1Objs variables: > * > * ex1Obj1 -- read-write > * First example object. > * > * ex1Obj2 -- read-only > * Second example object. > */ > > /* An internal routine to retrieve the values of the variables, used > * by the method routines ex1Objs_get and ex1Objs_next. > * You need to replace the type STRUCT_ex1Objs with something > * appropriate to your system. */ > static int > ex1Objs_get_value(OIDC_T lastmatch, > SNMP_PKT_T *pktp, > VB_T *vbp, > STRUCT_ex1Objs *data) /* !!! */ > { > switch(lastmatch) { > case LEAF_ex1Obj1: > /* Values: > * one(1) = VAL_ex1Obj1_one > * two(2) = VAL_ex1Obj1_two > */ > getproc_got_int32(pktp, vbp, > data->ex1Obj1); /* !!! */ > break; > case LEAF_ex1Obj2: > /* if the data being returned is in dynamic storage and needs > * to be free'd, change the 0 argument to a 1. */ > getproc_got_string(pktp, vbp, string_length(data->ex1Obj2), > data->ex1Obj2, 0, > VT_STRING); /* !!! */ > break; > default: > return GEN_ERR; > } > return NO_ERROR; > } > > void > ex1Objs_get(OIDC_T lastmatch, > int compc, > OIDC_T *compl, > VB_T *vbp, > SNMP_PKT_T *pktp) > { > STRUCT_ex1Objs data; /* !!! */ > int error; > > /* find all the varbinds that share the same getproc and instance */ > > group_by_getproc_and_instance(pktp, vbp, compc, compl); > > /* check that the instance is exactly .0 */ > if (!((compc == 1) && (*compl == 0))) > for ( ; vbp; vbp = vbp->vb_link) > getproc_nosuchins(pktp, vbp); > /* grab the actual data for this variable. this lookup routine will > * probably have to be changed for your system. for scalar variables > > * there might not even be any lookup routine. */ > else if (ex1Objs_lookup(&data) != > 0) /* !!! */ > for ( ; vbp; vbp = vbp->vb_link) > getproc_error(pktp, vbp, GEN_ERR); > else { > /* retrieve all the values from the same data structure */ > for ( ; vbp; vbp = vbp->vb_link) { > if ((error = ex1Objs_get_value(vbp->vb_ml.ml_last_match, > pktp, vbp, &data)) != NO_ERROR) > getproc_error(pktp, vbp, error); > } > } > } > > [omitted] > > void > ex1Objs_set(OIDC_T lastmatch, > int compc, > OIDC_T *compl, > VB_T *vbp, > SNMP_PKT_T *pktp) > { > for ( ; vbp; vbp = vbp->vb_link) { > switch (vbp->vb_ml.ml_last_match) { > case LEAF_ex1Obj1: > ex1Obj1 = VB_GET_INT32(vbp); /* !!! */ > setproc_good(pktp, vbp); > break; > default: > setproc_error(pktp, vbp, COMMIT_FAILED); > return; > } > } > } > > [omitted] > > //--------- File stub.c : END ----------- > > Was it possible to create mibcomp > generating set-functions as following ? : > //--------- "updated" set : BEGIN ------------ > static int > ex1Objs_set_value(OIDC_T lastmatch, > SNMP_PKT_T *pktp, > VB_T *vbp, > STRUCT_ex1Objs *data) /* !!! */ > { > switch(lastmatch) { > case LEAF_ex1Obj1: > data->ex1Obj1 = VB_GET_INT32(vbp); /* !!! */ > break; > default: > return GEN_ERR; > } > return NO_ERROR; > } > > void > ex1Objs_get(OIDC_T lastmatch, > int compc, > OIDC_T *compl, > VB_T *vbp, > SNMP_PKT_T *pktp) > { > STRUCT_ex1Objs data; /* !!! */ > // ex1Objs_set stuff using ex1Objs_set_value > // (similar to ex1Objs_get using ex1Objs_get_value) > > //--------- "updated" set : END -------------- > > ================================== > Alex Vinokur > mailto:alexvn@bigfoot.com > http://up.to/alexvn > http://visitweb.com/alexvn.math > ================================== --------------------------- Newsgroups: comp.os.vxworks Subject: Re: netjobadd problem Date: Wed, 6 Jun 2001 08:59:14 +0800 From: "cxj" Organization: Bentium Ltd. Message-ID: <9fjv4m$8ul$1@mail.cn99.com> References: <9fiphn$9bj$1@overload.lbl.gov> Where did you call netJobAdd? Is it in your ISR? Then determine if your ISR have been invoked. There is only 85 entries in netJob Ring. Use rngIsEmpty(netJobRing) to determine if there is entry in netJobRing. Good luck! "Aditi Kapoor" wrote in message news:9fiphn$9bj$1@overload.lbl.gov... > hi, > i have been developing a wlan driver in vxworks but am stuck at a > crucial location. Somehow, when i do a netjobadd on a rx intr it gives me a > status of OK , but never is the function given to the netjobadd invoked. > > i've seen that the tNetTask is running as i can see it in the active > task list, but i dont understand why isnt my function added to its queue > through netjobadd ever invoked. > > i'd be grateful if u culd help me. > > thanx & regards > aditi > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Communication Problems with MVME5100 Date: Tue, 05 Jun 2001 19:34:52 -0700 From: DrDiags Organization: laserlink.net Message-ID: <3B1D96CC.B98E7448@covad.net> References: <9fipok$67i$1@chico.htc.honeywell.com> Reply-To: drdiags@covad.net Owen, Don't know if what you see is similar to an experience I had with a PPC750, where the BSP was set for 50Mhz crystal and I had a 66 Mhz crystal installed. The board would boot-up but the UART was being programmed based on the 50Mhz crystal. Don't have a lot of knowledge on the MVME1500, so this is probably a WAG. HTH. Owen Landgren wrote: > > We're working with a Motorola MVME5100 board, which supposedly had a > standard vxWorks/PPC-Bug split install already done. > > However, we're running into some very interesting problems. First, when we > connect to the debug port, using an RJ45 to Serial cable, the output on the > teletype is just garbage. Neither PPC-Bug nor vxWorks provides readable > output. However, what is even odder is that the board will respond to > commands such as CNFG in PPC-Bug mode, and will print more garbage to the > port. The garbage does not seem to have any sort of patterned or shifted > offset from standard ASCII. > > The computer we're connecting the MVME5100 to is a Win2K SP1 machine, and > we're using the standard Hilgraeve HyperTerminal to do so. The settings all > match what the documentation describes -- 9600 Baud, 8-None-1. We've also > experimented with a lot of different settings within HyperTerminal, to no > avail. > > Has anyone else had experience with this problem or found a way to solve it? > Any suggestion would be greatly appreciated -- we've tried everything we can > come up with. > > Owen Landgren --------------------------- Newsgroups: comp.os.vxworks Subject: Re: [Fwd: LoadCoreFileCheck() function call interface] Date: Tue, 05 Jun 2001 19:42:59 -0700 From: DrDiags Organization: laserlink.net Message-ID: <3B1D98B3.9C6A05CC@covad.net> References: <3B1CFBB5.391E9127@alcatel.com> Reply-To: drdiags@covad.net Xiong, I think you may have wanted to send this to support@windriver.com. I don't think you wanted the world to know your License Number either, but I may be wrong. Lets see if sending this to support will work for you. To Wind River Support: Please open a TSR for the following action. Thanks! Xiong Xue wrote: > > Xiong Xue wrote: > > > From: Xiong Xue > > company: Alcatel Canada > > License No. 204957 > > > > ----------------- > > > > Would you please raise a TSR for me on the following issues. > > > > Background: In unix environment, a tcl-based script runs on host > > machine, started target server, which connected to a target agent > > which runs in lab, via wtx/wdb/tcp/ip. > > > > We all know that when target server launches, it establishes a > > connection to target agent by wdb/tcp/ip, then it does a corefile > > checksum check between the test header section on host target server > > and image on target. The function called is LoadCoreFileCheck(). > > > > In our script, we need to run the corefile checksum check after the > > connection initialization stage in a controlled, interactive manner. > > i.e., I'd like to do the checksum called within the script, and I > > want to capture the return value, then make program flow decision > > based on the return value. > > > > Tornado provided 4 types of user interfaces to access target server, > > via WTX protocol. they are: Tornado shell, wtx tcl API, wtx C api > > and wtx java api. The one I am interested in is the wtxtcl. > > > > The problem: However, I checked all wtxtcl libraries, and couldn't > > find a procedure which is invoking the corefile checksum check. The > > closest one I get is wtxObjModuleChecksum(), which, according to its > > description in online docs, should do the job, however, gave 0 > > return no matter what corefiles were specified. (should return "Core > > File Checksum not match" error when the corefile on host does not > > match the target, as the function LoadCoreFileCheck() does). > > > > Would you please assign a engineer to help me on this? > > > > Thanks for help. > > > > Regards, > > > > -- > > Xiong Xue, ATM Development Technologies, CID Alcatel > > email: xiong.xue@alcatel.com > > phone 720-4416 (CID internal) 613-784-4416 (external) > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Booting of CPCI/C6400 Date: Tue, 5 Jun 2001 22:11:11 -0700 (PDT) From: Rashmi Topno Organization: Lawrence Berkeley National Laboratory Message-ID: <9fke96$af4$1@overload.lbl.gov> Hi, Could somebody please clarify the following ? In my system (cpci/c6400; MPC860) I need to set the ether net interface address during run time and later initiate socket communication through this interface. Could somebody give me some clue as to how this can be done ? I am planing to do this by invoking ifMaskSet() and ifAddrSet()... Do we require to reboot the vxworks / reset the board after setting the ether net interface address ? How do I get the interface name ? Thanks and best regards, rajendra. __________________________________________________ Do You Yahoo!? Get personalized email addresses from Yahoo! Mail - only $35 a year! http://personal.mail.yahoo.com/ --------------------------- Newsgroups: comp.os.vxworks Subject: cc386 compiler crash, anybody seen this error. Date: Wed, 6 Jun 2001 05:20:36 +0000 (UTC) From: subodh@best.com (Subodh Nijsure) Organization: a user of Best Internet Communications, Inc. www.best.com Message-ID: <9fkej4$2vjc$1@nntp1.ba.best.com> Anybody seen VxWorks compiler crashing while compiling C++ code with -O1 or - -O2 flag? I am trying to compile some C++ code, yes it has some templates & exception, and Windriver supplied compiler crashes with following error c386: Internal compiler error: program cc1plus got fatal signal 125 Windriver compiler version is as follows gcc driver version cygnus-2.7.2-960126 egcs-971225 tornado 2.0 executing gcc version cygnus-2.7.2-960126 I have tried compiling this code with gcc version 2.95.3-4 and code itself compiles fine. Looks like an issue with Windriver's compiler, anybody seen this behavior before and what was the solution? /Subodh Nijsure --------------------------- Newsgroups: comp.os.vxworks Subject: Debugger Date: Tue, 5 Jun 2001 23:01:06 -0700 (PDT) From: chen wei seow Organization: Lawrence Berkeley National Laboratory Message-ID: <9fkhpn$bmc$1@overload.lbl.gov> Hi, I'm new to Vxworks environment. Can any one help to solve my problem. Step1 : download vxworks.st image into target board Step2 : Launch target server Step3 : Launch debugger Step4 : Set the "Source Search Path" Step5 : Open the source code for example test.c file Step6 : set breakpoint in the test.c file Result : An error --> "Symbol table information for file, 'test.c' not found" instead. Is there any option I didn't aware ? thanks __________________________________________________ Do You Yahoo!? Get personalized email addresses from Yahoo! Mail - only $35 a year! http://personal.mail.yahoo.com/ --------------------------- Newsgroups: comp.os.vxworks,comp.protocols.snmp Subject: I can't do the "make" prcedure with snmp??? Date: Wed, 6 Jun 2001 15:01:14 +0800 From: "kumo Yeh" Organization: SEEDNet News Service Message-ID: <9fkkiu$cq1$1@news.seed.net.tw> Hi I try to enable a snmp agent with vxworks. In default MIB2 , it's a''right. And I want to add a MIB&agent function. I was follow the procedure "Extending the WindNet SNMPv1/v2c Agent" in WindNet SNMPv1/v2c docu. I use the example mib,and use mibcomp get source code which will link with snmp agent. But I can't do the step5 "building the agent"???? ============================ Tornado 2.0 for PowerPC VxWorks 5.4 Host : Windows98 Target : PowerPC (MPC8240) ============================ Q1:Does I have to change anything,because the example is using MC68020? I had change the CPU=PPC603 TOOL=gnu WIND_HOST_TYPE=x86-win32. Q2:Can I add function without using mibcomp? I mean that When I write down a mib,I compile it &put in mib browser. what agent needs to add-->*.h *.c or what? Q3:I also did the dynamic mib addition,it's the error in dos mode. what's the problem??? ############################################### C:\Tornado\target\src\snmpv1\agent>ccppc -msoft-float -o -fvolatile -fno-bui ltin -Wall -ansi -nostdinc -DCPU=PPC603 -DVXWORKS -I$WIND_BASE/target/h -I$WIND_ BASE /target/h/snmp/ -I. -c demo.c demomib.c CCPPC.EXE: cannot specify -o with -c and multiple compilations C:\Tornado\target\src\snmpv1\agent> #############################################33 Firmware Engineer GLORY TELECOM. Phone:(886)-3-3127832-386 Fax:(886)-3-3127831 http://www.glorytel.com.tw email:kumo@glorytel.com.tw --------------------------- Newsgroups: comp.os.vxworks Subject: Re: rename() does not work with dosFs2? Date: 6 Jun 2001 00:15:43 -0700 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0106052315.699ae70f@posting.google.com> References: <7f10f842.0106051613.3c92083d@posting.google.com> > Another annoyance with dosFs2 is that in prjConfig.c, > it insists on mounting hard-drive 1 also (using usrAtaConfig()) > even if there isn't a second hard-drive in the system. Also, > mounting the second non-existant drive does not fail when > vxworks starts up - the second drive shows up as "/ata1a" > shows up in the 'devs' list. You cannot have prjConfig.c > call usrAtaInit() alone without calling usrAtaConfig() > for drives 0 and 1. You do have the source code for these routines though, so just change them to match your requirements. If you're using the project facility, then you can also play with the init routine settings for the DosFs2 component(s) to get whatever effect you want, but simply editing the configlette code is going to be simpler in this case. For T2 BSP builds, the code is in target/src/config, for T2 and T3 project builds it will be in target/config/comps/src. HTH, John... --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Launching tornado problem Date: 6 Jun 2001 09:27:09 GMT From: borkhuis@agere.com (Johan Borkhuis) Organization: Agere System Message-ID: References: <9fit29$b9k$1@overload.lbl.gov> "Moy, Ian" wrote: > After successfully getting my board to boot I tried to connect to it > using tornado. When I try I get the following errors in the launch log: > > Command: tgtsvr mbx860_ism -L -C -A -V -s -n mbx860 -display jw-pc:0 > -Bt 30 -Br 30 -m 8000000 > > tgtsvr (mbx860@hex): Tue Jun 5 15:22:13 2001 > License request... authorized on host 'whel'. > Attaching backend... succeeded. > Connecting to target agent... succeeded. > Attaching C++ interface... succeeded. > Error: Couldn't open /usr/mbx860/vxWorks: status = 0x2 > Error: Unable to initialize object module loader library > > I am quite confused as to why tornado needs to open > /usr/mbx860/vxWorks, this is the path I use in the vxWorks bootloader > to load the VxWorks image onto the board, but why does tornado need to > know this?? The target server needs this information because it needs the symbol information. This is needed for example to download files or to run the host shell. If you add "-c " to your command line it shoudl work. Groeten, Johan - -- o o o o o o o . . . _____________________________ o _____ || Johan Borkhuis | .][__n_n_|DD[ ====_____ | borkhuis@agere.com | >(________|__|_[_________]_|__________________________| _/oo OOOOO oo` ooo ooo 'o!o!o o!o!o` === VxWorks FAQ: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html === --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Debugger Date: 6 Jun 2001 09:35:33 GMT From: borkhuis@agere.com (Johan Borkhuis) Organization: Agere System Message-ID: References: <9fkhpn$bmc$1@overload.lbl.gov> chen wei seow wrote: > Hi, > I'm new to Vxworks environment. Can any one help to > solve my problem. > > Step1 : download vxworks.st image into target board > Step2 : Launch target server > Step3 : Launch debugger > Step4 : Set the "Source Search Path" > Step5 : Open the source code for example test.c file > Step6 : set breakpoint in the test.c file > Result : An error --> "Symbol table information for > file, 'test.c' not found" instead. Did you enable debugging when you compiled your test.c file? Groeten, Johan - -- o o o o o o o . . . _____________________________ o _____ || Johan Borkhuis | .][__n_n_|DD[ ====_____ | borkhuis@agere.com | >(________|__|_[_________]_|__________________________| _/oo OOOOO oo` ooo ooo 'o!o!o o!o!o` === VxWorks FAQ: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html === --------------------------- Newsgroups: comp.os.vxworks Subject: network interface ip address setting during run time Date: Wed, 06 Jun 2001 15:26:13 +0530 From: rajendra Organization: Lucent Technologies Message-ID: <3B1DFE3D.975D9063@lucent.com> vxworks tornado Hi, Could somebody please clarify the following ? In my system (cpci/c6400; MPC860) I need to set the ether net interface address during run time and later initiate socket communication through this interface. Could somebody give me some clue as to how this can be done ? I am planing to do this by invoking ifMaskSet() and ifAddrSet()... Do we require to reboot the vxworks / reset the board after setting the ether net interface address ? How do I get the interface name ? Thanks and best regards, rajendra. --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Wed Jun 6 05:14:33 2001 From: "Michael Lawnick" Date: Wed Jun 6 05:14:36 PDT 2001 Subject: Re: Pended Semaphore behaviour Hi S. From: "S.Rengarajan" Newsgroups: comp.os.vxworks Sent: Thursday, June 07, 2001 1:07 PM Subject: Pended Semaphore behaviour > Hello all, > Iam having a doubt.We have two tasks T1 and T2 .T2 having higher priority > then T1.Both are pended on a semaphore.T2 suspends the task T1.Some other > task (T3)resleases the semaphore.T2 takes the semaphore and after updating > the data releases the semaphore.Now will the task T1 take the semaphore as I > think it's happening because T2 is getting blocked for ever.Is the behaviour > like this? > > The code follows > > /************************************************************************/ > > /* includes */ > #include "vxworks.h" > #include "taskLib.h" > #include "semLib.h" > #include "stdio.h" > > /* defines */ > #define ITERATIONS 10 > > /* globals */ > SEM_ID semB_Id; > int global = 0; > int taskIdOne, taskIdTwo, taskIdThree; > > /* forward declarations */ > void taskOne (void); > void taskTwo (void); > void taskThree (void); > void binary (void) > { > > /* Create semaphore with semaphore available and queue tasks on FIFO > basis */ > semB_Id = semBCreate (SEM_Q_PRIORITY, SEM_FULL); > > /* Note 1: lock the semaphore for scheduling purposes */ > semTake (semB_Id, WAIT_FOREVER); > > /* spawn the the two tasks */ > taskIdOne = taskSpawn ("t1", 90, 0x100, 2000, (FUNCPTR)taskOne, 0, 0, > 0, 0, 0, 0, 0, 0, 0, 0); > taskIdTwo = taskSpawn ("t2", 10, 0x100, 2000, (FUNCPTR)taskTwo, 0, 0, > 0, 0, 0, 0, 0, 0, 0, 0); > taskIdThree = taskSpawn ("t3", 90, 0x100, 2000, (FUNCPTR)taskThree, 0, > 0, 0, 0, 0, 0, 0, 0, 0, 0); > } > void taskOne (void) > { > int i; > for (i=0; i { > semTake (semB_Id, WAIT_FOREVER); /* wait indefinitely for semaphore */ > global = 1; > printf (" I am taskOne and global = %d.............\n", global); > semGive (semB_Id); /* give up semaphore */ > } > } > > /* taskTwo */ > void taskTwo (void) > > > int i; > /* semGive (semB_Id); *//* Note 2: give up semaphore (a scheduling fix) > */ > taskSuspend(taskIdOne); > for (i=0; i { > semTake (semB_Id, WAIT_FOREVER); /* Wait indefinitely for semaphore > */ > global = 0; > printf ("I am taskTwo and global = %d-----------\n", global); > semGive (semB_Id); /* give up semaphore */ > } > taskResume(taskIdOne); > } > > /* taskThree */ > void taskThree (void) > { > semGive (semB_Id); /* give up semaphore */ > } > > /*******************************************************************/ > > We are using Tornado 2 prototyper. > > Regards > Renga > I'd expect T2 to cycle ITERATIONS times, then T1 two times. Why ? You created a binary semaphore with SEM_Q_PRIORITY. T1 is first waiting for it, then T2 should wait with higher prio. When T3 gives it first time, T2 will start and cycle ITERATIONS times. The first semGive will free T1 and transact it from PEND+SUSP to SUSP. The second semGive will be stored within the semaphore. When T2 terminates, T1 will complete its first cycle, then receive the semaphore a second time and pend again on third semTake. One big problem: The behaviour of your tasks depends on the priority of your wrapper-task, calling binary(). Do you call it from Shell I'd expect the above. Do you call it within a task with prio>90, T1 will not reach the semTake on startup. It therefore will cycle only once. (Haven't tested your code, just from looking into it) Mit freundlichen Grüßen, Michael Lawnick ============================================== SOFTEC GmbH Tel +49-731-96600-0 Promenade 17 Fax +49-731-96600-23 D-89073 Ulm Michael Lawnick Germany lawnick@softec.de ============================================== From vxwexplo-errs@csg.lbl.gov Wed Jun 6 06:35:48 2001 From: "AMASS" Date: Wed Jun 6 06:35:50 PDT 2001 Subject: File Systems This is a multi-part message in MIME format. ------=_NextPart_000_0022_01C0EE95.83F23400 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Hi All We use T2/NT on a PowerPC rs60x. Is it possible to use FTP with the raw file system? I am assuming not, = due to the lack of a 'real' file structure. Is the DOS FS the only FS = which will support the use of FTP? What performance overhead does the = use of the DOS FS impose? Cheers Andy vxworks ------=_NextPart_000_0022_01C0EE95.83F23400 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable
Hi All
 
We use T2/NT on a PowerPC = rs60x.
 
Is it possible to use FTP with the raw = file system?=20 I am assuming not, due to the lack of a 'real' file structure. Is the = DOS FS the=20 only FS which will support the use of FTP? What performance overhead = does the=20 use of the DOS FS impose?
 
Cheers
 
Andy
 
vxworks
------=_NextPart_000_0022_01C0EE95.83F23400-- From vxwexplo-errs@csg.lbl.gov Wed Jun 6 08:18:17 2001 From: "Crozier, Bruce" Date: Wed Jun 6 08:18:19 PDT 2001 Subject: RE: vxWorks POSIX timer Hello Priya, Sorry, I don't know where the 'kill failed' is coming from, however, I compiled your test program for my PowerPC target CPU, and after adding '#define INCLUDE_POSIX_TIMERS' to the vxWorks make, the test program 'timers' appears to work fine on my CPU (Force PowerCore 680). Maybe it is a problem with the particular CPU or BSP you use ? What processor architecture, vxWorks version, and CPU are you using ? -------------< my output> ------------ -> ld < /tst/clktst.o value = 33544280 = 0x1ffd858 = my_handler + 0x6a8 -> timers tp_gettime.tv_sec == 0 tp_gettime.tv_nsec == 0 created timer.... connected timer.... settime for timer.... tp_gettime.tv_sec == 0 tp_gettime.tv_nsec == 16666666 value = 0 = 0x0 -> ..........in function............ ..........in function............ < etc, ~1/sec > Regards, Bruce -----Original Message----- Subject: Using timer_create() and related routines..... Date: Tue, 5 Jun 2001 16:24:55 +0530 From: Priya Pradhan > Hi! > > I am trying to create a periodic timer on VxWorks using the routines - > timer_create(), timer_connect(), timer_settime(). Everytime I try to run > the program I get the error - "kill failed" - I can't understand the > error. > > The code is included below..... ---- > Hoping someone will be able to help. > > Thanks, > Priya. > Priya Pradhan > Member of Technical Staff. > Net Brahma Technologies Pvt. Ltd. > Internext Networking Software > Ph: 5521451 (Extn 212) > http://www.netbrahma.com From vxwexplo-errs@csg.lbl.gov Wed Jun 6 09:49:27 2001 From: David Laight Date: Wed Jun 6 09:49:30 PDT 2001 Subject: Re: File Systems > Is it possible to use FTP with the raw file system? I am assuming not, > due to the lack of a 'real' file structure. Is the DOS FS the only FS > which will support the use of FTP? What performance overhead does the > use of the DOS FS impose? The ftp server uses open/read/write/close the same as any other program, so can read (and write) raw filesystems. You will need to use a 'proper' ftp client to request the required transfer. None of the windows 'file manager' style ftp clients interact well with the vxWorks ftp server. If you create a raw file system covering the same disc space as a dosfs one, you can take an image of the current state of the filesystem and restore it at a later date. (Assuming no files are open ...) One thing I found is that the ftp server requires that the raw device driver error reads for zero bytes at the end of file - otherwise the ftp doesn't complete. David ---------------------------------------------------------------- David Laight email: dsl@tadpole.co.uk Tadpole Technology plc phone: +44 1223 428 232 Cambridge, UK fax: +44 1223 428 201 From vxwexplo-errs@csg.lbl.gov Wed Jun 6 21:25:22 2001 From: "Chandrashekar Nagaraj" Date: Wed Jun 6 21:25:24 PDT 2001 Subject: vxworks pragma pack... hi, i want to know the significance of argument to the "pragma_pack()" system call in "C" language and "VxWorks"... Is it have anything more to do other than packing... with regards, chandra. From vxwexplo-errs@csg.lbl.gov Wed Jun 6 23:05:50 2001 From: "T. Hariharan" Date: Wed Jun 6 23:05:54 PDT 2001 Subject: Re: rename() does not work with dosFs2? Thanks for the solution. It works! The following should also work for you (the diretory 'AB' cannot be in lowercase in the destination path, but can be in lower or uppercase in the source path): > rename "/I0d/AB/tmp", "/I0d/AB/fj" Hari ----- Original Message ----- From: "Dave Olson" To: Sent: Wednesday, June 06, 2001 11:12 AM Subject: re: rename() does not work with dosFs2? > [I read the vxworks newsgroup via a mail list distributor & don't know how > to respond other than via direct e-mail.] > > Here's something to check. Although all of the other file manipulation > routines in dosFs2 are case insensitive, 'rename' is not. Here's a script > that demos the issue; of course the paths and file names are specific to our > system. > > #-- rename experiment > @cd "/I0d" > mkdir "AB" > cp "/I0a/pdb/slm.dtd", "/I0d/ab/tmp" > #rename using full path, but all lower case for the directory. This fails. > rename "/I0d/ab/tmp", "/I0d/ab/fj" > #use correct name for the first argument. This fails. > rename "/I0d/AB/tmp", "/I0d/ab/fj" > #use all lower case for 1st argument, correct case for 2nd. This succeeds! > rename "/I0d/ab/tmp", "/I0d/AB/fj" > > David H. Olson > Staff Engineer > Oresis Communications, Inc. > 14670 NW Greenbrier Parkway > Beaverton, OR 97006 > 503.466.6246 Fax:503.629.8009 > dolson@oresis.com > > > From vxwexplo-errs@csg.lbl.gov Thu Jun 7 00:44:44 2001 From: "Michael Lawnick" Date: Thu Jun 7 00:44:46 PDT 2001 Subject: Re: Pended Semaphore behaviour ----- Original Message ----- From: "Kumar" Newsgroups: comp.os.vxworks Sent: Wednesday, June 06, 2001 6:51 PM Subject: Re: Pended Semaphore behaviour > I did test the code. It seems to be working perfectly. > Here is the output. > > > I am taskTwo and global = 0----------- > I am taskTwo and global = 0----------- > I am taskTwo and global = 0----------- > I am taskTwo and global = 0----------- > I am taskTwo and global = 0----------- > I am taskTwo and global = 0----------- > I am taskTwo and global = 0----------- > I am taskTwo and global = 0----------- > I am taskTwo and global = 0----------- > I am taskTwo and global = 0----------- > I am taskOne and global = 1............. > I am taskOne and global = 1............. > I am taskOne and global = 1............. > I am taskOne and global = 1............. > I am taskOne and global = 1............. > I am taskOne and global = 1............. > I am taskOne and global = 1............. > I am taskOne and global = 1............. > I am taskOne and global = 1............. > I am taskOne and global = 1............. > > > I do hope that this is what you are expecting to happen. By seeing your logic. > No, I didn't see it that way, I saw it wrong. I missed that the semTake is in the loop of t2 (not only before). When calling binary from shell, your output is right. As t2 gives the sema at the end, it gets it again at the beginning. The same works t1. The problem comes up, when calling binary with 'sp': T2 can do only one output, its semGive will be eaten by the suspended T1 and the whole mechanism gets blocked ! -> bad construction Mit freundlichen Grüßen, Michael Lawnick ============================================== SOFTEC GmbH Tel +49-731-96600-0 Promenade 17 Fax +49-731-96600-23 D-89073 Ulm Michael Lawnick Germany lawnick@softec.de ============================================== From vxwexplo-errs@csg.lbl.gov Thu Jun 7 04:00:24 2001 From: Fiederling.Ulf@dornier.dasa.de Date: Thu Jun 7 04:00:26 PDT 2001 Subject: Initializing ATA for loading shell script Hi VxWorks Gurus We are using Tornado II on a Pentium platform How can we initialize the ATA device for loading scripts from the shell? It seems the only way to do this is in usrAppIni by calling usrAtaConfigt, but when this is done, the shell is already started. Any solution? Best regards Ulf Fiederling From vxwexplo-errs@csg.lbl.gov Fri Jun 8 03:32:17 2001 From: "Harry Stox" Date: Fri Jun 8 03:32:19 PDT 2001 Subject: Intel 82559ER driver on Tornado 2.0.2? Hello, I'm interested to know if anyone managed to get the unsupported Intel Architecture driver for the Intel 82559ER on-board Ethernet chip (a559mux.exe from the wrs support site) up and running on a ServerWorks ServerSet III LE based mainboard. We are trying to build a kernel for the new FORCE Flexor IDP3S370 SBC, but when following the instructions supplied in the driver's readme file, VxWorks freezes after approximately one second after calling syseeVPciInit(), even when not automatically attaching the END device to the IP stack. We're using Tornado 2.0.2 (T2CP4), which has support for NPT drivers. I noticed that END drivers sometimes need to be slightly modified to get them to work reliably using a kernel with NPT support. I don't know if the Intel Architecture driver can be used in conjunction with a kernel having NPT support. Further details: - PC_PENTIUMPRO kernel - VxVMI included - CPU: PIII/1GHz - Memory: 256Mb - no other (external) PCI devices present on backplane Thanks in advance for your reaction. Best regards, Harry Stox - Ellips B.V. --- Harry Stox (harry@ellips.nl) -- Software Engineer -- Ellips B.V. Urkhovenseweg 11 -- 5641 KA -- Eindhoven -- The Netherlands phone: +31-40-2456540 -- fax: +31-40-2467183 -- http://www.ellips.nl From vxwexplo-errs@csg.lbl.gov Fri Jun 8 04:03:30 2001 From: Vxworks Exploder Date: Fri Jun 8 04:03:33 PDT 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Fri Jun 8 04:03:11 PDT 2001 Subject: Pended Semaphore behaviour Subject: Re: Some problem with C++ Subject: re: Pended Semaphore behaviour Subject: Pended Semaphore behaviour Subject: How to build a bootline in NvRam Subject: How to change bootline parameters Subject: File Systems Subject: Project files, Tornado 2.0 and RCS Subject: Re: rename() does not work with dosFs2? Subject: Re: Project files, Tornado 2.0 and RCS Subject: RE: vxWorks POSIX timer Subject: Using timer_create() and related routines..... Subject: Re: Project files, Tornado 2.0 and RCS Subject: Re: How to change bootline parameters Subject: Re: Pended Semaphore behaviour Subject: Re: How to change bootline parameters Subject: re: File Systems Subject: Tundra Universe IIB Weirdness Subject: DP national driver driver for VxWorks Subject: vxsim memory size Subject: Re: network interface ip address setting during run time Subject: re: File Systems Subject: Re: Project files, Tornado 2.0 and RCS Subject: Re: How to change bootline parameters Subject: Embedded S'ware (Application & Compiler) developer needed in Ottawa & Subject: reboot() problems Subject: Re: Embedded S'ware (Application & Compiler) developer needed in Ottawa & Milpitas OFF-TOPIC Subject: How to run VxWorks on an x86 machine? Subject: Network Stack Subject: Redirecting stdout problems Subject: Re: How to run VxWorks on an x86 machine? Subject: VxWorks Tools questions Subject: Re: Redirecting stdout problems Subject: vxworks pragma pack... Subject: re: rename() does not work with dosFs2? Subject: re: rename() does not work with dosFs2? Subject: Using GDB 5 (insight, cygwin) with VxWorks (WinNT) ? Subject: Re: Re: Project files, Tornado 2.0 and RCS Subject: Re: Pended Semaphore behaviour Subject: Re: vxsim memory size Subject: Re: VxSim Subject: Re: Using timer_create() and related routines..... Subject: target board behavior on reboot() of vxworks Subject: Re: VxWorks Tools questions Subject: re: Pended Semaphore behaviour Subject: Re: Pended Semaphore behaviour Subject: Initializing ATA for loading shell script Subject: Re: target board behavior on reboot() of vxworks Subject: Diab C++ fundamental access bug? Subject: Re: vxworks pragma pack... Subject: Re: target board behavior on reboot() of vxworks : BOOT_NORMAL option Subject: Re: vxsim memory size Subject: Re: Pended Semaphore behaviour Subject: =?Big5?B?pOW+zLCqr8Wk5b7Msqa3fqTlvsy+x6RoutOkaL3StXs=?= Subject: Re: VxWorks Tools questions Subject: Shell with WXworks Subject: Re: Pended Semaphore behaviour Subject: =?Big5?B?uNu4db3StXuxwLxzrfs=?= Subject: MVME5100, does the BSP support writing to Flash? Subject: Re: Redirecting stdout problems Subject: how to find the right component to link in? Subject: loadModule on from fs, not from memory Subject: Gratuitous ARP and etherAddrResolve Subject: Re: target board behavior on reboot() of vxworks : BOOT_NORMAL option Subject: Interrupts with x86 BSP Subject: Re: Interrupts with x86 BSP Subject: Re: Shell with WXworks Subject: Re: TCP Theory question puzzling me Subject: Re: Controling point at which socket write blocks Subject: Re: loadModule on from fs, not from memory Subject: equivalent of dn_expand in vxworks ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Pended Semaphore behaviour Date: Thu, 7 Jun 2001 16:37:01 +0530 From: "S.Rengarajan" Organization: Siemens Inc. Message-ID: <9fl32p$jqk$1@news.mch.sbs.de> Hello all, Iam having a doubt.We have two tasks T1 and T2 .T2 having higher priority then T1.Both are pended on a semaphore.T2 suspends the task T1.Some other task (T3)resleases the semaphore.T2 takes the semaphore and after updating the data releases the semaphore.Now will the task T1 take the semaphore as I think it's happening because T2 is getting blocked for ever.Is the behaviour like this? The code follows /************************************************************************/ /* includes */ #include "vxworks.h" #include "taskLib.h" #include "semLib.h" #include "stdio.h" /* defines */ #define ITERATIONS 10 /* globals */ SEM_ID semB_Id; int global = 0; int taskIdOne, taskIdTwo, taskIdThree; /* forward declarations */ void taskOne (void); void taskTwo (void); void taskThree (void); void binary (void) { /* Create semaphore with semaphore available and queue tasks on FIFO basis */ semB_Id = semBCreate (SEM_Q_PRIORITY, SEM_FULL); /* Note 1: lock the semaphore for scheduling purposes */ semTake (semB_Id, WAIT_FOREVER); /* spawn the the two tasks */ taskIdOne = taskSpawn ("t1", 90, 0x100, 2000, (FUNCPTR)taskOne, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); taskIdTwo = taskSpawn ("t2", 10, 0x100, 2000, (FUNCPTR)taskTwo, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); taskIdThree = taskSpawn ("t3", 90, 0x100, 2000, (FUNCPTR)taskThree, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); } void taskOne (void) { int i; for (i=0; i Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: <3B1800D4.FAE5B1DC@hotmail.com> <3B1BC22E.7040403@snellwilcox.com> <3B1BD0D0.B0BA892@hotmail.com> <3B1D3F7E.7EB3F339@hotmail.com> Beginner wrote in message <3B1D3F7E.7EB3F339@hotmail.com>... >Dave Korn wrote: >> >> Are you sure your makefile is executing the much stage correctly? >>Without munching a C++ module, any static C++ class objects in that file >> will not be constructed when the DL finishes loading the module. >> >Hi, Dave, > >That is a point I never thought about, now how do you munch a C++ module? >Any help is appreciated. From "VxWorks Programmer's Guide 5.4 Edition 1", Chapter 5 "C++ Development ", Section 5.2.5: - ---8<----snip!----8<--- 5.2.5 Munching C++ Application Modules Modules written in C++ must undergo an additional host processing step before being downloaded to a VxWorks target. This extra step (called munching, by convention) initializes support for static objects and ensures that when the module is downloaded to VxWorks, the C++ run-time support is able to call the correct constructors and destructors in the correct order for all static objects. The following commands will compile hello.cpp, then munch hello.o, resulting in the munched file hello.out suitable for loading by the Tornado module loader: cc68k -IinstallDir/target/h -DCPU=MC68020 -nostdinc -fno-builtin \ -c hello.cpp nm68k hello.o | wtxtcl installDir/host/src/hutils/munch.tcl \ -asm 68k > ctdt.c cc68k -c ctdt.c ld68k -r -o hello.out hello.o ctdt.o - ---------------------------------------------------------------------------- NOTE: You can substitute the actual name of your installDir or use $WIND_BASE (UNIX) or %WIND_BASE% (Windows). - ---------------------------------------------------------------------------- - ---------------------------------------------------------------------------- CAUTION: The GNU ToolKit User's Guide: Using ld describes a -Ur option for resolving references to C++ constructors. That option is for native development, not for cross-development. Do not use -Ur with C++ modules for VxWorks. - ---------------------------------------------------------------------------- - ---8<----snip!----8<--- hth, DaveK - -- They laughed at Galileo. They laughed at Copernicus. They laughed at Columbus. But remember, they also laughed at Bozo the Clown. --------------------------- Newsgroups: comp.os.vxworks Subject: re: Pended Semaphore behaviour Date: Wed, 6 Jun 2001 13:47:09 +0200 From: "Michael Lawnick" Organization: Lawrence Berkeley National Laboratory Message-ID: <9fl6sq$kq5$1@overload.lbl.gov> Hi S. From: "S.Rengarajan" Newsgroups: comp.os.vxworks Sent: Thursday, June 07, 2001 1:07 PM Subject: Pended Semaphore behaviour > Hello all, > Iam having a doubt.We have two tasks T1 and T2 .T2 having higher priority > then T1.Both are pended on a semaphore.T2 suspends the task T1.Some other > task (T3)resleases the semaphore.T2 takes the semaphore and after updating > the data releases the semaphore.Now will the task T1 take the semaphore as I > think it's happening because T2 is getting blocked for ever.Is the behaviour > like this? > > The code follows > > /************************************************************************/ > > /* includes */ > #include "vxworks.h" > #include "taskLib.h" > #include "semLib.h" > #include "stdio.h" > > /* defines */ > #define ITERATIONS 10 > > /* globals */ > SEM_ID semB_Id; > int global = 0; > int taskIdOne, taskIdTwo, taskIdThree; > > /* forward declarations */ > void taskOne (void); > void taskTwo (void); > void taskThree (void); > void binary (void) > { > > /* Create semaphore with semaphore available and queue tasks on FIFO > basis */ > semB_Id = semBCreate (SEM_Q_PRIORITY, SEM_FULL); > > /* Note 1: lock the semaphore for scheduling purposes */ > semTake (semB_Id, WAIT_FOREVER); > > /* spawn the the two tasks */ > taskIdOne = taskSpawn ("t1", 90, 0x100, 2000, (FUNCPTR)taskOne, 0, 0, > 0, 0, 0, 0, 0, 0, 0, 0); > taskIdTwo = taskSpawn ("t2", 10, 0x100, 2000, (FUNCPTR)taskTwo, 0, 0, > 0, 0, 0, 0, 0, 0, 0, 0); > taskIdThree = taskSpawn ("t3", 90, 0x100, 2000, (FUNCPTR)taskThree, 0, > 0, 0, 0, 0, 0, 0, 0, 0, 0); > } > void taskOne (void) > { > int i; > for (i=0; i { > semTake (semB_Id, WAIT_FOREVER); /* wait indefinitely for semaphore */ > global = 1; > printf (" I am taskOne and global = %d.............\n", global); > semGive (semB_Id); /* give up semaphore */ > } > } > > /* taskTwo */ > void taskTwo (void) > > > int i; > /* semGive (semB_Id); *//* Note 2: give up semaphore (a scheduling fix) > */ > taskSuspend(taskIdOne); > for (i=0; i { > semTake (semB_Id, WAIT_FOREVER); /* Wait indefinitely for semaphore > */ > global = 0; > printf ("I am taskTwo and global = %d-----------\n", global); > semGive (semB_Id); /* give up semaphore */ > } > taskResume(taskIdOne); > } > > /* taskThree */ > void taskThree (void) > { > semGive (semB_Id); /* give up semaphore */ > } > > /*******************************************************************/ > > We are using Tornado 2 prototyper. > > Regards > Renga > I'd expect T2 to cycle ITERATIONS times, then T1 two times. Why ? You created a binary semaphore with SEM_Q_PRIORITY. T1 is first waiting for it, then T2 should wait with higher prio. When T3 gives it first time, T2 will start and cycle ITERATIONS times. The first semGive will free T1 and transact it from PEND+SUSP to SUSP. The second semGive will be stored within the semaphore. When T2 terminates, T1 will complete its first cycle, then receive the semaphore a second time and pend again on third semTake. One big problem: The behaviour of your tasks depends on the priority of your wrapper-task, calling binary(). Do you call it from Shell I'd expect the above. Do you call it within a task with prio>90, T1 will not reach the semTake on startup. It therefore will cycle only once. (Haven't tested your code, just from looking into it) Mit freundlichen Grüßen, Michael Lawnick ============================================== SOFTEC GmbH Tel +49-731-96600-0 Promenade 17 Fax +49-731-96600-23 D-89073 Ulm Michael Lawnick Germany lawnick@softec.de ============================================== --------------------------- Newsgroups: comp.os.vxworks Subject: How to build a bootline in NvRam Date: Wed, 6 Jun 2001 20:59:54 +0800 From: "John Lo" Organization: SEEDNet News Service Message-ID: <9fl95j$3lj$1@news.seed.net.tw> Could anybody tell me how to build a bootline in NvRam Thank you --------------------------- Newsgroups: comp.os.vxworks Subject: How to change bootline parameters Date: Wed, 6 Jun 2001 21:01:20 +0800 From: "John Lo" Organization: SEEDNet News Service Message-ID: <9fl988$3m1$1@news.seed.net.tw> Could anybold tell me how could I change bootline parameters in NvRam. Thank you --------------------------- Newsgroups: comp.os.vxworks Subject: File Systems Date: Wed, 6 Jun 2001 14:32:29 +0100 From: "AMASS" Organization: Lawrence Berkeley National Laboratory Message-ID: <9fldts$o02$1@overload.lbl.gov> This is a multi-part message in MIME format. - ------=_NextPart_000_0022_01C0EE95.83F23400 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Hi All We use T2/NT on a PowerPC rs60x. Is it possible to use FTP with the raw file system? I am assuming not, = due to the lack of a 'real' file structure. Is the DOS FS the only FS = which will support the use of FTP? What performance overhead does the = use of the DOS FS impose? Cheers Andy vxworks - ------=_NextPart_000_0022_01C0EE95.83F23400 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable
Hi All
 
We use T2/NT on a PowerPC = rs60x.
 
Is it possible to use FTP with the raw = file system?=20 I am assuming not, due to the lack of a 'real' file structure. Is the = DOS FS the=20 only FS which will support the use of FTP? What performance overhead = does the=20 use of the DOS FS impose?
 
Cheers
 
Andy
 
vxworks
- ------=_NextPart_000_0022_01C0EE95.83F23400-- --------------------------- Newsgroups: comp.os.vxworks Subject: Project files, Tornado 2.0 and RCS Date: Wed, 6 Jun 2001 15:03:34 +0000 (UTC) From: igtorque@eliop.es (Ignacio González) Organization: Mailgate.ORG Server - http://www.Mailgate.ORG Message-ID: <3B1E61D2.8313.11B8B2F@localhost> This is not a big issue, but it's annoyng. I'm using RCS for version control of my Tornado / VxWorks projects. I include the *.wpj file (which is a text file) in the repository, along wit= h my c files and the like. Tornado's editor supports Unix text files (lines ending with \n) and DOS t= ext files as well (lines ending with \r\n). However, project files MUST be Uni= x text files. The RCS version I've got, when checking out a file, looks at the type of t= he host: if it is a Unix machine, extracts the file as a Unix text file. If i= t is a DOS machine (and WINNT platforms are considered DOS machines), you get a DOS text file. And here is the problem: when I try to add a project (checked out in this = way from a RCS repository) to a Tornado's workspace, the *.wpj file is not properly recognised. Present workaround: check out the project file, open it in an editor and s= ave it as a Unix file. Desired solution: convince Tornado to perform the conversion automatically= . Is it possible?. Intermediate solution: convince RCS to extract some text files as Unix fil= es over DOS machines. I've found no option for doing that, and prefer not to modify the RCS source files. Any useful comment? Thanks, Ignacio Gonz=E1lez - -- Posted from 57089.rad.tsai.es [194.224.57.89] via Mailgate.ORG Server - http://www.Mailgate.ORG --------------------------- Newsgroups: comp.os.vxworks Subject: Re: rename() does not work with dosFs2? Date: 6 Jun 2001 08:26:36 -0700 From: hariharan@foobox.com (Thiagarajan Hariharan) Organization: http://groups.google.com/ Message-ID: <7f10f842.0106060726.4bc295fb@posting.google.com> References: <7f10f842.0106051613.3c92083d@posting.google.com> <488e459a.0106052315.699ae70f@posting.google.com> john_94501@yahoo.com (John) wrote in message news:<488e459a.0106052315.699ae70f@posting.google.com>... > > Another annoyance with dosFs2 is that in prjConfig.c, > > it insists on mounting hard-drive 1 also (using usrAtaConfig()) > > even if there isn't a second hard-drive in the system. Also, > > mounting the second non-existant drive does not fail when > > vxworks starts up - the second drive shows up as "/ata1a" > > shows up in the 'devs' list. You cannot have prjConfig.c > > call usrAtaInit() alone without calling usrAtaConfig() > > for drives 0 and 1. > > You do have the source code for these routines though, so just change them > to match your requirements. If you're using the project facility, then you > can also play with the init routine settings for the DosFs2 component(s) to > get whatever effect you want, but simply editing the configlette code is > going to be simpler in this case. > > For T2 BSP builds, the code is in target/src/config, for T2 and T3 project > builds it will be in target/config/comps/src. > > HTH, > > John... Thanks for the detailed reply. Setting ATA_DEV_NAMES_1 to "" in the project settings for ATA fixed it - now it does not try to mount disk 1. Hari --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Project files, Tornado 2.0 and RCS Date: Wed, 06 Jun 2001 10:35:23 -0500 From: Francisco Pataro Organization: Verio Message-ID: <3B1E4DBB.3CEB374@dnaent.com> References: <3B1E61D2.8313.11B8B2F@localhost> Ignacio, On WINNT environment we use CS-RCS product that is built on top of RCS. They allow to specify the file type when checking out the .wpj files. Here is the link (price is very reasonably). http://www.ComponentSoftware.com Regards, Francisco. "Ignacio González" wrote: > This is not a big issue, but it's annoyng. > > I'm using RCS for version control of my Tornado / VxWorks projects. I > include the *.wpj file (which is a text file) in the repository, along with my c > files and the like. > > Tornado's editor supports Unix text files (lines ending with \n) and DOS text > files as well (lines ending with \r\n). However, project files MUST be Unix > text files. > > The RCS version I've got, when checking out a file, looks at the type of the > host: if it is a Unix machine, extracts the file as a Unix text file. If it is a DOS > machine (and WINNT platforms are considered DOS machines), you get a > DOS text file. > > And here is the problem: when I try to add a project (checked out in this way > from a RCS repository) to a Tornado's workspace, the *.wpj file is not > properly recognised. > > Present workaround: check out the project file, open it in an editor and save > it as a Unix file. > > Desired solution: convince Tornado to perform the conversion automatically. > Is it possible?. > > Intermediate solution: convince RCS to extract some text files as Unix files > over DOS machines. I've found no option for doing that, and prefer not to > modify the RCS source files. > > Any useful comment? > > Thanks, > > Ignacio González > > -- > Posted from 57089.rad.tsai.es [194.224.57.89] > via Mailgate.ORG Server - http://www.Mailgate.ORG --------------------------- Newsgroups: comp.os.vxworks Subject: RE: vxWorks POSIX timer Date: Wed, 6 Jun 2001 08:07:52 -0700 From: "Crozier, Bruce" Organization: Lawrence Berkeley National Laboratory Message-ID: <9flkur$rrb$1@overload.lbl.gov> Hello Priya, Sorry, I don't know where the 'kill failed' is coming from, however, I compiled your test program for my PowerPC target CPU, and after adding '#define INCLUDE_POSIX_TIMERS' to the vxWorks make, the test program 'timers' appears to work fine on my CPU (Force PowerCore 680). Maybe it is a problem with the particular CPU or BSP you use ? What processor architecture, vxWorks version, and CPU are you using ? - -------------< my output> ------------ -> ld < /tst/clktst.o value = 33544280 = 0x1ffd858 = my_handler + 0x6a8 -> timers tp_gettime.tv_sec == 0 tp_gettime.tv_nsec == 0 created timer.... connected timer.... settime for timer.... tp_gettime.tv_sec == 0 tp_gettime.tv_nsec == 16666666 value = 0 = 0x0 -> ..........in function............ ..........in function............ < etc, ~1/sec > Regards, Bruce - -----Original Message----- Subject: Using timer_create() and related routines..... Date: Tue, 5 Jun 2001 16:24:55 +0530 From: Priya Pradhan > Hi! > > I am trying to create a periodic timer on VxWorks using the routines - > timer_create(), timer_connect(), timer_settime(). Everytime I try to run > the program I get the error - "kill failed" - I can't understand the > error. > > The code is included below..... - ---- > Hoping someone will be able to help. > > Thanks, > Priya. > Priya Pradhan > Member of Technical Staff. > Net Brahma Technologies Pvt. Ltd. > Internext Networking Software > Ph: 5521451 (Extn 212) > http://www.netbrahma.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Project files, Tornado 2.0 and RCS Date: 6 Jun 2001 16:23:35 GMT From: borkhuis@agere.com (Johan Borkhuis) Organization: Agere System Message-ID: References: <3B1E61D2.8313.11B8B2F@localhost> igtorque@eliop.es (Ignacio González) wrote: > I'm using RCS for version control of my Tornado / VxWorks projects. I > include the *.wpj file (which is a text file) in the repository, along > with my c files and the like. > > Tornado's editor supports Unix text files (lines ending with \n) and > DOS text files as well (lines ending with \r\n). However, project files > MUST be Unix text files. > > The RCS version I've got, when checking out a file, looks at the type > of the host: if it is a Unix machine, extracts the file as a Unix text > file. If it is a DOS machine (and WINNT platforms are considered DOS > machines), you get a DOS text file. > > And here is the problem: when I try to add a project (checked out in > this way from a RCS repository) to a Tornado's workspace, the *.wpj > file is not properly recognised. Yes, take a look at this page: http://www.xs4all.nl/~borkhuis/vxworks/vxw_pt1.html#1.12-A Hope that will help. Groeten, Johan - -- o o o o o o o . . . _____________________________ o _____ || Johan Borkhuis | .][__n_n_|DD[ ====_____ | borkhuis@agere.com | >(________|__|_[_________]_|__________________________| _/oo OOOOO oo` ooo ooo 'o!o!o o!o!o` === VxWorks FAQ: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html === --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to change bootline parameters Date: Wed, 06 Jun 2001 11:48:58 -0700 From: Joe Durusau Organization: Lockheed Martin Corporation Message-ID: <3B1E7B1A.4E7747BF@lmco.com> References: <9fl988$3m1$1@news.seed.net.tw> I'm not sure I understand the question, but: Try plugging a serial terminal to the first serial port on the target, then turn the target off and back on. You should see a countdown start--- Then press any key on the term keyboard. If you see something like : type ? and press return. Then, you should be able to find the answers to your questions. Speaking only for myself, Joe Durusau John Lo wrote: > > Could anybold tell me how could I change bootline parameters > in NvRam. Thank you --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Pended Semaphore behaviour Date: 6 Jun 2001 09:51:12 -0700 From: tlksk@hotmail.com (Kumar) Organization: http://groups.google.com/ Message-ID: References: <9fl6sq$kq5$1@overload.lbl.gov> I did test the code. It seems to be working perfectly. Here is the output. I am taskTwo and global = 0----------- I am taskTwo and global = 0----------- I am taskTwo and global = 0----------- I am taskTwo and global = 0----------- I am taskTwo and global = 0----------- I am taskTwo and global = 0----------- I am taskTwo and global = 0----------- I am taskTwo and global = 0----------- I am taskTwo and global = 0----------- I am taskTwo and global = 0----------- I am taskOne and global = 1............. I am taskOne and global = 1............. I am taskOne and global = 1............. I am taskOne and global = 1............. I am taskOne and global = 1............. I am taskOne and global = 1............. I am taskOne and global = 1............. I am taskOne and global = 1............. I am taskOne and global = 1............. I am taskOne and global = 1............. I do hope that this is what you are expecting to happen. By seeing your logic. "Michael Lawnick" wrote in message news:<9fl6sq$kq5$1@overload.lbl.gov>... > Hi S. > > From: "S.Rengarajan" > Newsgroups: comp.os.vxworks > Sent: Thursday, June 07, 2001 1:07 PM > Subject: Pended Semaphore behaviour > > > > Hello all, > > Iam having a doubt.We have two tasks T1 and T2 .T2 having higher priority > > then T1.Both are pended on a semaphore.T2 suspends the task T1.Some other > > task (T3)resleases the semaphore.T2 takes the semaphore and after updating > > the data releases the semaphore.Now will the task T1 take the semaphore as > I > > think it's happening because T2 is getting blocked for ever.Is the > behaviour > > like this? > > > > The code follows > > > > /************************************************************************/ > > > > /* includes */ > > #include "vxworks.h" > > #include "taskLib.h" > > #include "semLib.h" > > #include "stdio.h" > > > > /* defines */ > > #define ITERATIONS 10 > > > > /* globals */ > > SEM_ID semB_Id; > > int global = 0; > > int taskIdOne, taskIdTwo, taskIdThree; > > > > /* forward declarations */ > > void taskOne (void); > > void taskTwo (void); > > void taskThree (void); > > void binary (void) > > { > > > > /* Create semaphore with semaphore available and queue tasks on FIFO > > basis */ > > semB_Id = semBCreate (SEM_Q_PRIORITY, SEM_FULL); > > > > /* Note 1: lock the semaphore for scheduling purposes */ > > semTake (semB_Id, WAIT_FOREVER); > > > > /* spawn the the two tasks */ > > taskIdOne = taskSpawn ("t1", 90, 0x100, 2000, (FUNCPTR)taskOne, 0, > 0, > > 0, 0, 0, 0, 0, 0, 0, 0); > > taskIdTwo = taskSpawn ("t2", 10, 0x100, 2000, (FUNCPTR)taskTwo, 0, > 0, > > 0, 0, 0, 0, 0, 0, 0, 0); > > taskIdThree = taskSpawn ("t3", 90, 0x100, 2000, (FUNCPTR)taskThree, 0, > > 0, 0, 0, 0, 0, 0, 0, 0, 0); > > } > > void taskOne (void) > > { > > int i; > > for (i=0; i > { > > semTake (semB_Id, WAIT_FOREVER); /* wait indefinitely for semaphore */ > > global = 1; > > printf (" I am taskOne and global = %d.............\n", global); > > semGive (semB_Id); /* give up semaphore */ > > } > > } > > > > /* taskTwo */ > > void taskTwo (void) > > > > > > int i; > > /* semGive (semB_Id); *//* Note 2: give up semaphore (a scheduling > fix) > > */ > > taskSuspend(taskIdOne); > > for (i=0; i > { > > semTake (semB_Id, WAIT_FOREVER); /* Wait indefinitely for > semaphore > > */ > > global = 0; > > printf ("I am taskTwo and global = %d-----------\n", global); > > semGive (semB_Id); /* give up semaphore */ > > } > > taskResume(taskIdOne); > > } > > > > /* taskThree */ > > void taskThree (void) > > { > > semGive (semB_Id); /* give up semaphore */ > > } > > > > /*******************************************************************/ > > > > We are using Tornado 2 prototyper. > > > > Regards > > Renga > > > I'd expect T2 to cycle ITERATIONS times, then T1 two times. > > Why ? You created a binary semaphore with SEM_Q_PRIORITY. > T1 is first waiting for it, then T2 should wait with higher prio. > When T3 gives it first time, T2 will start and cycle ITERATIONS times. > The first semGive will free T1 and transact it from PEND+SUSP to SUSP. > The second semGive will be stored within the semaphore. > When T2 terminates, T1 will complete its first cycle, then receive the > semaphore > a second time and pend again on third semTake. > > One big problem: > The behaviour of your tasks depends on the priority of your wrapper-task, > calling > binary(). Do you call it from Shell I'd expect the above. Do you call it > within a task > with prio>90, T1 will not reach the semTake on startup. It therefore will > cycle only once. > > (Haven't tested your code, just from looking into it) > > Mit freundlichen Grüßen, > Michael Lawnick > ============================================== > SOFTEC GmbH Tel +49-731-96600-0 > Promenade 17 Fax +49-731-96600-23 > D-89073 Ulm Michael Lawnick > Germany lawnick@softec.de > ============================================== --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to change bootline parameters Date: 6 Jun 2001 10:13:03 -0700 From: tlksk@hotmail.com (Kumar) Organization: http://groups.google.com/ Message-ID: References: <9fl988$3m1$1@news.seed.net.tw> If you are asking about how to change from vxWorks shell it is by using "bootChange" or You can also get it by doing the following 1. buf = malloc ( 300 ) 2. Initialize the buf as the bootline parameters string. 3. "sysNvRamSet buf,300,0" you should have initialized the buf with the bootline as produced in the output of version Kumar "John Lo" wrote in message news:<9fl988$3m1$1@news.seed.net.tw>... > Could anybold tell me how could I change bootline parameters > in NvRam. Thank you --------------------------- Newsgroups: comp.os.vxworks Subject: re: File Systems Date: Wed, 6 Jun 2001 17:48:53 +0100 (BST) From: David Laight Organization: Lawrence Berkeley National Laboratory Message-ID: <200106061648.RAA16892@dsl-2.tadpole.co.uk> > Is it possible to use FTP with the raw file system? I am assuming not, > due to the lack of a 'real' file structure. Is the DOS FS the only FS > which will support the use of FTP? What performance overhead does the > use of the DOS FS impose? The ftp server uses open/read/write/close the same as any other program, so can read (and write) raw filesystems. You will need to use a 'proper' ftp client to request the required transfer. None of the windows 'file manager' style ftp clients interact well with the vxWorks ftp server. If you create a raw file system covering the same disc space as a dosfs one, you can take an image of the current state of the filesystem and restore it at a later date. (Assuming no files are open ...) One thing I found is that the ftp server requires that the raw device driver error reads for zero bytes at the end of file - otherwise the ftp doesn't complete. David - ---------------------------------------------------------------- David Laight email: dsl@tadpole.co.uk Tadpole Technology plc phone: +44 1223 428 232 Cambridge, UK fax: +44 1223 428 201 --------------------------- Newsgroups: comp.os.vxworks Subject: Tundra Universe IIB Weirdness Date: Wed, 06 Jun 2001 17:25:28 GMT From: cgan76@nospam4me.com (cgan76) Organization: MTS Internet Message-ID: <3b1e6510.344858470@200.123.101.100> Howdy, My company is currently developing a VME 64x board (MPC8240) that utilizes the Universe IIB bridge. We are running VxWorks 5.4 as our operating system. Here's our current setup: MPC8240 VME 64x Board (any slot): -MPC8240 CPU running VxWorks 5.4 -64 MB SDRAM -64 bit PCI bus -Universe IIB bridge @ 0x80000000 (Local PCI) -Slave Window into SDRAM @ 0x70000000 (VME A32)and 0x00000000 (Local PCI), size is 64 MB. -All caching is turned off System Controller (slot 1): -SBS Technologies BIT3 Card -We are also performing reads/writes to the above slave window from this card. We are seeing a strange problem when performing reads/writes to the VME slave window (A32). Initially, when the system is booted and the Universe is newly initialized, reads and writes into the A32 space into MPC8240 SDRAM works properly. For example: Device name "\\.\Btw0_A32" Starting bt_read() test Attempting read #0, 0x100 bytes 00000000: f7 df d7 bd f9 7b af bd ea 7d bf d7 fb bb db e3 00000010: f3 29 9e db f7 db 5f ff 7f db fd 5d 7f ff bf dd 00000020: 94 79 fb f5 fb 7f db ff 7f 1e d3 df 7b ff ff fe data truncated.... 000000d0: f6 17 ff e9 bf ff f2 fc bf fb 7b d7 b7 df fe ff 000000e0: 57 fb e7 fb bb bf d7 b7 ff ed 3f 4f 67 76 9f fa 000000f0: b5 7f bf eb f7 fd a2 f7 ff fe b9 7f b7 fd ba bf bt_read() #0 success Exiting bt_read() test But after a certain amount of time, a read will be perfromed and the data read will have every second DWORD returned as zeros. For example: Device name "\\.\Btw0_A32" Starting bt_read() test Attempting read #0, 0x100 bytes 00000000: f7 df d7 bd 00 00 00 00 ea 7d bf d7 00 00 00 00 00000010: f3 29 9e db 00 00 00 00 7f db fd 5d 00 00 00 00 00000020: 94 79 fb f5 00 00 00 00 7f 1e d3 df 00 00 00 00 data truncated... 000000d0: f6 17 ff e9 00 00 00 00 bf fb 7b d7 00 00 00 00 000000e0: 57 fb e7 fb 00 00 00 00 ff ed 3f 4f 00 00 00 00 000000f0: b5 7f bf eb 00 00 00 00 ff fe b9 7f 00 00 00 00 bt_read() #0 success Exiting bt_read() test Once in this state, writes will also fail in the same way. i.e. Every second DWORD will not be written. We have found that we can put the Universe into the working state again by merely READING any one of it's PCI registers. For example we used the first 32 bit register at (0x80000000) which is just the device id. Again, after doing this read, the Universe will operate for a time before reverting back to reading zeros for every second DWORD. We do not know what could be causing this and your help in this matter would be greatly appreciated. Thanks! Christian Gan --------------------------- Newsgroups: comp.os.vxworks Subject: DP national driver driver for VxWorks Date: Wed, 6 Jun 2001 13:26:15 -0400 From: "Mike Kordik" Message-ID: Is anyone using this driver (DP83815)? The target server will not connect to this driver after you install the Tornado 2.02 patch. Does anyone know why? Is there a workaround? Thanks, Mike --------------------------- Newsgroups: comp.os.vxworks Subject: vxsim memory size Date: Wed, 06 Jun 2001 13:50:57 -0400 From: Doug Dotson Organization: Verio Message-ID: <3B1E6D81.8A76BB77@digidata.com> Does anybody know how to increase the memory size of vxsim? Thanks, Doug --------------------------- Newsgroups: comp.os.vxworks Subject: Re: network interface ip address setting during run time Date: 6 Jun 2001 11:33:21 -0700 From: pfliv@yahoo.com (Paul Livingston) Organization: http://groups.google.com/ Message-ID: <7b1623b6.0106061033.3d812360@posting.google.com> References: <3B1DFE3D.975D9063@lucent.com> We change our address at runtime on an MVME2400. You must bring the IF down, change the address and then bring the IF up : ifFlagChange("dc0", IFF_UP, FALSE); /* take interface down */ ifAddrSet("dc0", "nnn.nnn.nnn.nnn"); /* change IP address */ ifFlagChange("dc0", IFF_UP, TRUE); /* bring interface up */ Paul Livingston rajendra wrote in message news:<3B1DFE3D.975D9063@lucent.com>... > vxworks tornado > > Hi, > > Could somebody please clarify the following ? > > In my system (cpci/c6400; MPC860) I need to set the ether net interface address > during run time and later initiate socket communication through this interface. > > Could somebody give me some clue as to how this can be done ? I am planing to do > this by invoking ifMaskSet() and ifAddrSet()... Do we require to reboot the > vxworks / reset the board after setting the ether net interface address ? How do > I get the interface name ? > > Thanks and best regards, > rajendra. --------------------------- Newsgroups: comp.os.vxworks Subject: re: File Systems Date: Wed, 06 Jun 2001 19:50:19 +0100 From: molochai Message-ID: References: <200106061648.RAA16892@dsl-2.tadpole.co.uk> In message <200106061648.RAA16892@dsl-2.tadpole.co.uk> David Laight wrote: > > Is it possible to use FTP with the raw file system? I am assuming not, > > due to the lack of a 'real' file structure. Is the DOS FS the only FS > > which will support the use of FTP? What performance overhead does the use > > of the DOS FS impose? > > The ftp server uses open/read/write/close the same as any other program, so > can read (and write) raw filesystems. > > You will need to use a 'proper' ftp client to request the required > transfer. None of the windows 'file manager' style ftp clients interact > well with the vxWorks ftp server. The unsupported ftp server with DosFs2 is much better behaved in ny experience, and pretends to be UN*X. If you're using the built in ftp server, some ftp clients do have directory listing parsers which can cope. I've had sucess with Reflection FTP, ws_ftp and Visual SlickEdit (built in ftp client). You can incidentally FTP via vxworks to nfs mounts ;-) --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Project files, Tornado 2.0 and RCS Date: Wed, 06 Jun 2001 20:00:15 +0100 From: molochai Message-ID: <9c8882864a%molochai@vapour-trail.demon.co.uk> References: <3B1E61D2.8313.11B8B2F@localhost> <3B1E4DBB.3CEB374@dnaent.com> In message <3B1E4DBB.3CEB374@dnaent.com> Francisco Pataro wrote: > Ignacio, > > On WINNT environment we use CS-RCS product that is built on top of RCS. > They allow to specify the file type when checking out the .wpj files. > > Here is the link (price is very reasonably). > > http://www.ComponentSoftware.com > The company I work for uses CS-RCS in its infinate wisdom.. WTF anyone would pay for a wrapper around RCS when WinCVS is free, and considerably more flexible is beyond me. Anyway, the solution is to get RCS to treat the project file as binary, and for the CR/LF not to be converted. (the problem seems to be that Tornado II is very sensitive to *exact* formatting of the project file) [snip] --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to change bootline parameters Date: Wed, 06 Jun 2001 19:22:13 GMT From: kput@avtecinc.com (Kim) Organization: Info Avenue Internet Services Message-ID: <3b1e8272.28655953@192.168.1.214> References: <9fl988$3m1$1@news.seed.net.tw> On Wed, 6 Jun 2001 21:01:20 +0800, "John Lo" wrote: >Could anybold tell me how could I change bootline parameters >in NvRam. Thank you > > To change them in code something like BOOT_PARAMS params; bootStringToStruct(BOOT_LINE_ADRS,¶ms); sprintf(params.bootDev,"tffs=0,0"); bootStructToString (BOOT_LINE_ADRS, ¶ms); (void) sysNvRamSet (BOOT_LINE_ADRS, strlen (BOOT_LINE_ADRS) + 1, 0); k --------------------------- Newsgroups: comp.sys.mips,comp.os.vxworks,comp.lang.c,comp.lang.c++ Subject: Embedded S'ware (Application & Compiler) developer needed in Ottawa & Milpitas Date: Wed, 06 Jun 2001 16:04:08 -0400 From: Rabindra Guha Organization: Bell Nexxia Message-ID: <3B1E8CB8.EC80A462@empowertel.com> Empowertel Networks is seeking Embedded S'ware developers with experience in Telecom/ATM or assembly compilers/tools. If you are interested, please visit our website www.empowertel.com or contact us directly. Rabin - -- Rabindra Guha - Project Leader - guha@empowertel.com Empowertel Networks Canada 11 Holland Ave, Suite #701, Ottawa, Ontario K1Y 4S1, Canada. Tel: (613) 728-4562 --------------------------- Newsgroups: comp.os.vxworks Subject: reboot() problems Date: 6 Jun 2001 13:28:15 -0700 From: hariharan@foobox.com (Thiagarajan Hariharan) Organization: http://groups.google.com/ Message-ID: <7f10f842.0106061228.5c4a35c8@posting.google.com> Hello, Our platform is a generic PC using Pentium III. We use Tornado 2.0.2. When I use the vxworks reboot() function, it seems to work sometimes and hang other times. I was trying the warm reboot (sysWarmType = SYS_WARM_ATA). I added printf's to sysLib.c. It reads bootrom.sys successfully and transfers control to the bootrom code (the line "(*pEntry) (startType);" in sysLib.c). After that, the system just hangs (it does work on some systems, but hangs on others). Then I tried SYS_WARM_BIOS. Had better luck with it than SYS_WARM_ATA. But even with SYS_WARM_BIOS, some systems simply hang. Anyone have any similar experience? Anything I'm missing? I start reboot() in a separate task as follows (calling reboot() in a low priority thread didn't work at all): (void)taskSpawn( "reboot", 1, VX_UNBREAKABLE|VX_DEALLOC_STACK|VX_FP_TASK, 10000, (FUNCPTR)reboot, BOOT_QUICK_AUTOBOOT|BOOT_CLEAR, 0, 0, 0, 0, 0, 0, 0, 0, 0); Thanks, Hari --------------------------- Newsgroups: comp.sys.mips,comp.os.vxworks,comp.lang.c,comp.lang.c++ Subject: Re: Embedded S'ware (Application & Compiler) developer needed in Ottawa & Milpitas OFF-TOPIC Date: Wed, 06 Jun 2001 21:16:19 GMT From: David Harmon Organization: Form follows function. Message-ID: <3b3a9c3b.100521317@news.earthlink.net> References: <3B1E8CB8.EC80A462@empowertel.com> On Wed, 06 Jun 2001 16:04:08 -0400 in comp.lang.c++, Rabindra Guha wrote: >Empowertel Networks is seeking Embedded S'ware developers with >experience in Telecom/ATM or assembly compilers/tools. If you are >interested, please visit our website Job ads, off-topic in comp.lang.c++, give the impression of an employer that is either technically clueless, or abusive and irresponsible, or both. In any case, a company you do not want to work with! There are dozens of *.jobs.* newsgroups where they belong. By posting in this manner you declare to the entire world that you and Empowertel Networks just do not care enough or respect the users of comp.lang.c++ enough to follow ordinary courtesy. See the article titled "Welcome to comp.lang.c++! Read this first." posted semiweekly in this newsgroup or at: http://www.slack.net/~shiva/welcome.txt See the topic "[5.8] Which newsgroup should I post my questions?" in Marshall Cline's C++ FAQ. You can get the FAQ at: http://www.parashift.com/cpp-faq-lite/ (Posted and emailed.) --------------------------- Newsgroups: comp.os.vxworks Subject: How to run VxWorks on an x86 machine? Date: Thu, 7 Jun 2001 00:09:59 +0200 From: "Rui Germano" Organization: Planet Internet Message-ID: <9fm9l3$7oqs5$1@reader01.wxs.nl> Hello: I'm a bit of a newbie in VxWorks. I use it embedded on a board where I work but stripped of some funbctionality. I would like to be able to install it in a x86 machine with all features like TCP/IP support , fat, etc. I have Tornado and I know I can create an hex image to run on 486 machines. I have also seen a startup disk that initiated VxWorks on a x86 machine with serial port communication support and allowed to start a Target Server. Well, if anyone can give some hints I would be most grateful Regards Rui --------------------------- Newsgroups: comp.os.vxworks Subject: Network Stack Date: Thu, 7 Jun 2001 02:18:58 +0000 (UTC) From: Jen-Chieh Tang Organization: University of California, Davis Message-ID: <9fmoai$4jj$1@woodrow.ucdavis.edu> Hello All, I am rather new to vxWorks and was wondering what the community uses to debug the network stack. I am pretty familiar with the stack code as I have worked with the BSD stack before, but how does one debug in vxWorks ? Thanks for any info. - -- Tom Tang tangj@cs.ucdavis.edu --------------------------- Newsgroups: comp.os.vxworks Subject: Redirecting stdout problems Date: 6 Jun 2001 19:28:28 -0700 From: bill_harmon_mw@yahoo.com (Bill Harmon) Organization: http://groups.google.com/ Message-ID: I am having problems redirecting stdout. I am trying to provide the ability for our the CLI in our application to call certain VxWorks functions such as i(), ifShow(), and ti() so that they may be executed via our CLI through telnet (not the Tornado shell). For some of these functions, the source code is provided so I can easily direct the output to our routines, however some functions are not provided in source form. As a result, I am trying to come up with a general scheme to redirect the output of stdout so that I can capture the output of these routines, and send it to our routines for output to the user via telnet. First, is redirecting stdout the best (only?) way to accomlish this? Assuming this is the best way to proceed, I wrote some code which redirects the stdout to a pipe using ioGlobalStdSet(). While this works in most cases, it does not seem to work reliably, and sometimes crashes the system. Further, in order to capture the output of these routines I found that I needed to create a pipe such as pipeDevCreate ("/pipe/cli", 1000, 200); The effect is, I beleive, reserving a relatively huge amount of memory for the queues for this function. This does not seem right to have to allocate this amount of memory for this purpose. I have not used pipes before so part of the problem may be my misundertanding of how to use them correctly. If I use smaller amount of max_msgs or max_msg_len, I find that I miss portions of the output generated by i(), for example. The code fragment that I am using to capture the output of stdout and send it to my routine is as follows; // Open the pipe for STDOUT redirection fd = open ("/pipe/cli", O_RDWR, 0644); if (fd != ERROR) { // Setup the I/O streams for redirection taskLock(); stdin_fd = ioGlobalStdGet(0); stdout_fd = ioGlobalStdGet (1); stderr_fd = ioGlobalStdGet (2); ioGlobalStdSet(0,fd); ioGlobalStdSet(1,fd); ioGlobalStdSet(2,fd); taskUnlock(); // Clear out and residual data in the CLI pipe ioctl (fd, FIOFLUSH, 0); // Call the actual function which does the work i(); // <--- I want to capture the output of this function // Replace the original I/O streams for this task taskLock(); ioGlobalStdSet(0,stdin_fd); ioGlobalStdSet(1,stdout_fd); ioGlobalStdSet(2,stderr_fd); taskUnlock(); // Initialize the output buffer to NULL strcpy(pipeBuf,""); // Get the number of messages queued up for us... ioctl (fd, FIONMSGS, (int) &nMessages); // Get the output from the pipe while (nMessages--) { len = read (fd, dummyBuf, 200); strncat(pipeBuf, dummyBuf, len); if (strlen(pipeBuf) >= 80) { // Send pipeBuf to the output routines here... strcpy(pipeBuf,""); } } // Flush any residual data if (strlen(pipeBuf)) { // Send pipeBuf to the output routines here... } // Close the pipe for STDOUT redirection close (fd); } Is it correct that I pipeDevCreate needs to allocate a queue this large to capture this output? Is there a better was to accomplish what I am trying to do? Thanks in advance, Bill Harmon --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to run VxWorks on an x86 machine? Date: Wed, 06 Jun 2001 19:32:16 -0700 From: DrDiags Organization: laserlink.net Message-ID: <3B1EE7B0.78703DF3@covad.net> References: <9fm9l3$7oqs5$1@reader01.wxs.nl> Reply-To: drdiags@covad.net Rui, If you are saying you have T2 for x86, then from the BSP directory of your architecture type, you can make a bootrom and then use the mkboot utility like so: C:>mkboot a: bootrom This will put a image called bootrom.sys on your floppy drive which will behave exactly the same as a bootrom on a target. You can make sure the config.h file has the type of features you want to access the kernel, whether it is on an ATA drive, accross the network or also on the floppy as a vxWorks.st image. However, if you don't have the x86 flavor of Tornado, you might find the mkboot utility in the unsupported section of your architecture (PPC for example). The target/unsupported/src/dosfs2 has the mkboot.c file and vxsys.com. You might be able to get an x86 BSP and build your own x86 Toolchain from WRS GNU Toolkit CD, which should be open source and from what I know, free but painful since the makefiles are not geared to allow a straight off the CD build, but then again what else is new? I might try contacting your local Sales guy and see if you could get a boot floppy and x86 kernel for the particular x86 type you have. It may work and would be easier than trying to hack something together. But then you run into the problem of all the support libraries and tools not being there to support your x86 platform. Just my 2 cents worth... Rui Germano wrote: > > Hello: > I'm a bit of a newbie in VxWorks. I use it embedded on a board where I work > but stripped of some funbctionality. I would like to be able to install it > in a x86 machine with all features like TCP/IP support , fat, etc. I have > Tornado and I know I can create an hex image to run on 486 machines. I have > also seen a startup disk that initiated VxWorks on a x86 machine with serial > port communication support and allowed to start a Target Server. > Well, if anyone can give some hints I would be most grateful > Regards > Rui --------------------------- Newsgroups: comp.os.vxworks Subject: VxWorks Tools questions Date: Thu, 07 Jun 2001 02:37:21 GMT From: joe-woodbury@mail.com (Joseph Woodbury) Message-ID: <3b1ee72a.4738173@news.qwest.net> My company is on the verge of adopting VxWorks for some projects. However, I am totally confused by several of the Tornado tools Wind River offers. Part of this is due to Wind River buying up companies right and left, even those with competing products, which they then offer without explanation. What is the difference between CodeTEST and PerformancePAK (ProfileScope & MemScope)? Are there significant differences between the [default] GNU compiler and Diab C++? (We will be using C++ on a PowerQUICC 860) How about between the default debugger and SingleStep? (Note, for the above mentioned pSOS, we were using Diab C++ and SingleStep, though I was involved only for a few weeks of overlap when the subsidiary formerly doing the development was shut down and I took over.) Any objective (i.e. non-Wind River) analysis would be very welcome. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Redirecting stdout problems Date: Thu, 07 Jun 2001 10:42:39 +0100 From: Vijay Kumar Peshkar Organization: Cisco Systems Inc. Message-ID: <3B1F4C8F.E137BEF@wipro.com> References: This is a multi-part message in MIME format. - --------------A8EC2541722ADBAF1B8D4A7D Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Bill, You can try by sending commands to the target shell using simple ASCII (sprintf()) and get responses. You can parse the responses with sscanf(). Hope this helps, Vijay Bill Harmon wrote: > I am having problems redirecting stdout. I am trying to provide > the ability for our the CLI in our application to call certain > VxWorks functions such as i(), ifShow(), and ti() so that they > may be executed via our CLI through telnet (not the Tornado > shell). For some of these functions, the source code is provided > so I can easily direct the output to our routines, however some > functions are not provided in source form. As a result, I am > trying to come up with a general scheme to redirect the output > of stdout so that I can capture the output of these routines, > and send it to our routines for output to the user via telnet. > First, is redirecting stdout the best (only?) way to accomlish > this? Assuming this is the best way to proceed, I wrote some code > which redirects the stdout to a pipe using ioGlobalStdSet(). > While this works in most cases, it does not seem to work reliably, > and sometimes crashes the system. Further, in order to capture the > output of these routines I found that I needed to create a pipe > such as > > pipeDevCreate ("/pipe/cli", 1000, 200); > > The effect is, I beleive, reserving a relatively huge amount of > memory for the queues for this function. This does not seem right to > have to allocate this amount of memory for this purpose. I have not > used pipes before so part of the problem may be my misundertanding > of how to use them correctly. If I use smaller amount of max_msgs or > max_msg_len, I find that I miss portions of the output generated by > i(), for example. > > The code fragment that I am using to capture the output of stdout > and send it to my routine is as follows; > > // Open the pipe for STDOUT redirection > fd = open ("/pipe/cli", O_RDWR, 0644); > > if (fd != ERROR) { > > // Setup the I/O streams for redirection > taskLock(); > stdin_fd = ioGlobalStdGet(0); > stdout_fd = ioGlobalStdGet (1); > stderr_fd = ioGlobalStdGet (2); > ioGlobalStdSet(0,fd); > ioGlobalStdSet(1,fd); > ioGlobalStdSet(2,fd); > taskUnlock(); > > // Clear out and residual data in the CLI pipe > ioctl (fd, FIOFLUSH, 0); > > // Call the actual function which does the work > i(); // <--- I want to capture the output of this function > > // Replace the original I/O streams for this task > taskLock(); > ioGlobalStdSet(0,stdin_fd); > ioGlobalStdSet(1,stdout_fd); > ioGlobalStdSet(2,stderr_fd); > taskUnlock(); > > // Initialize the output buffer to NULL > strcpy(pipeBuf,""); > > // Get the number of messages queued up for us... > ioctl (fd, FIONMSGS, (int) &nMessages); > > // Get the output from the pipe > while (nMessages--) { > len = read (fd, dummyBuf, 200); > strncat(pipeBuf, dummyBuf, len); > if (strlen(pipeBuf) >= 80) { > // Send pipeBuf to the output routines here... > strcpy(pipeBuf,""); > } > } > > // Flush any residual data > if (strlen(pipeBuf)) { > // Send pipeBuf to the output routines here... > } > > // Close the pipe for STDOUT redirection > close (fd); > } > > Is it correct that I pipeDevCreate needs to allocate a queue this > large to capture this output? Is there a better was to accomplish > what I am trying to do? > > Thanks in advance, > > Bill Harmon - --------------A8EC2541722ADBAF1B8D4A7D Content-Type: text/x-vcard; charset=us-ascii; name="vijay.peshkar.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Vijay Kumar Peshkar Content-Disposition: attachment; filename="vijay.peshkar.vcf" begin:vcard n:Peshkar;Vijay Kumar tel;work:5530034/35/35 x2237 x-mozilla-html:FALSE org:Wipro Global R & D;Wishing you blue skies, a couple of daisies, some fresh air and a smile! adr:;;;;;; version:2.1 email;internet:vpeshkar@cisco.com note:Wishing you blue skies, a couple of daisies, some fresh air and a smile! fn:Vijay Kumar Peshkar end:vcard - --------------A8EC2541722ADBAF1B8D4A7D-- --------------------------- Newsgroups: comp.os.vxworks Subject: vxworks pragma pack... Date: Thu, 7 Jun 2001 09:56:41 +0530 From: "Chandrashekar Nagaraj" Organization: Lawrence Berkeley National Laboratory Message-ID: <9fn2l6$jv3$1@overload.lbl.gov> hi, i want to know the significance of argument to the "pragma_pack()" system call in "C" language and "VxWorks"... Is it have anything more to do other than packing... with regards, chandra. --------------------------- Newsgroups: comp.os.vxworks Subject: re: rename() does not work with dosFs2? Date: Wed, 6 Jun 2001 11:43:45 -0700 From: "T. Hariharan" Organization: Lawrence Berkeley National Laboratory Message-ID: <9fn65u$mh4$1@overload.lbl.gov> Thanks for the solution. It works! The following should also work for you (the diretory 'AB' cannot be in lowercase in the destination path, but can be in lower or uppercase in the source path): > rename "/I0d/AB/tmp", "/I0d/AB/fj" Hari - ----- Original Message ----- From: "Dave Olson" To: Sent: Wednesday, June 06, 2001 11:12 AM Subject: re: rename() does not work with dosFs2? > [I read the vxworks newsgroup via a mail list distributor & don't know how > to respond other than via direct e-mail.] > > Here's something to check. Although all of the other file manipulation > routines in dosFs2 are case insensitive, 'rename' is not. Here's a script > that demos the issue; of course the paths and file names are specific to our > system. > > #-- rename experiment > @cd "/I0d" > mkdir "AB" > cp "/I0a/pdb/slm.dtd", "/I0d/ab/tmp" > #rename using full path, but all lower case for the directory. This fails. > rename "/I0d/ab/tmp", "/I0d/ab/fj" > #use correct name for the first argument. This fails. > rename "/I0d/AB/tmp", "/I0d/ab/fj" > #use all lower case for 1st argument, correct case for 2nd. This succeeds! > rename "/I0d/ab/tmp", "/I0d/AB/fj" > > David H. Olson > Staff Engineer > Oresis Communications, Inc. > 14670 NW Greenbrier Parkway > Beaverton, OR 97006 > 503.466.6246 Fax:503.629.8009 > dolson@oresis.com > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Using GDB 5 (insight, cygwin) with VxWorks (WinNT) ? Date: 7 Jun 2001 00:14:11 -0700 From: heiko_elger@arburg.com (Heiko Elger) Organization: http://groups.google.com/ Message-ID: <97e6274e.0106062314.443a6d88@posting.google.com> Hello has anyone a port for using gdb 5 (insight, cygnus) with VxWorks 5.4 for x86 or simNT targets? As I know the WDB protocol has to port to gdb 5. Perhaps someone has done this x86 or SimNt? best regards and thanks a lot Heiko Elger --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Re: Project files, Tornado 2.0 and RCS Date: Thu, 7 Jun 2001 07:56:43 +0000 (UTC) From: igtorque@eliop.es (Ignacio González) Organization: Mailgate.ORG Server - http://www.Mailgate.ORG Message-ID: <3B1F4F33.23356.4BACF95@localhost> Thank you for your pointer. Useful indeed. Saludos, - --------- Ignacio Gonzalez igtorque@eliop.es - -- Posted from 57089.rad.tsai.es [194.224.57.89] via Mailgate.ORG Server - http://www.Mailgate.ORG --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Pended Semaphore behaviour Date: Thu, 07 Jun 2001 14:22:13 +0100 From: Vijay Kumar Peshkar Organization: Cisco Systems Inc. Message-ID: <3B1F8005.95C83C8@wipro.com> References: <9fl32p$jqk$1@news.mch.sbs.de> This is a multi-part message in MIME format. - --------------C563E351ABD3119D7FA9C104 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Renga, After T3 gives the semaphore, T1 is taking it. But before T1 gives, its gets suspended by T2. So both tasks T1 and T2 get blocked on semaphore. Try resuming T1 from some other task so that it releases semaphore and T2 takes it. Hope this helps, bye, Vijay "S.Rengarajan" wrote: > Hello all, > Iam having a doubt.We have two tasks T1 and T2 .T2 having higher priority > then T1.Both are pended on a semaphore.T2 suspends the task T1.Some other > task (T3)resleases the semaphore.T2 takes the semaphore and after updating > the data releases the semaphore.Now will the task T1 take the semaphore as I > think it's happening because T2 is getting blocked for ever.Is the behaviour > like this? > > The code follows > > /************************************************************************/ > > /* includes */ > #include "vxworks.h" > #include "taskLib.h" > #include "semLib.h" > #include "stdio.h" > > /* defines */ > #define ITERATIONS 10 > > /* globals */ > SEM_ID semB_Id; > int global = 0; > int taskIdOne, taskIdTwo, taskIdThree; > > /* forward declarations */ > void taskOne (void); > void taskTwo (void); > void taskThree (void); > void binary (void) > { > > /* Create semaphore with semaphore available and queue tasks on FIFO > basis */ > semB_Id = semBCreate (SEM_Q_PRIORITY, SEM_FULL); > > /* Note 1: lock the semaphore for scheduling purposes */ > semTake (semB_Id, WAIT_FOREVER); > > /* spawn the the two tasks */ > taskIdOne = taskSpawn ("t1", 90, 0x100, 2000, (FUNCPTR)taskOne, 0, 0, > 0, 0, 0, 0, 0, 0, 0, 0); > taskIdTwo = taskSpawn ("t2", 10, 0x100, 2000, (FUNCPTR)taskTwo, 0, 0, > 0, 0, 0, 0, 0, 0, 0, 0); > taskIdThree = taskSpawn ("t3", 90, 0x100, 2000, (FUNCPTR)taskThree, 0, > 0, 0, 0, 0, 0, 0, 0, 0, 0); > } > void taskOne (void) > { > int i; > for (i=0; i { > semTake (semB_Id, WAIT_FOREVER); /* wait indefinitely for semaphore */ > global = 1; > printf (" I am taskOne and global = %d.............\n", global); > semGive (semB_Id); /* give up semaphore */ > } > } > > /* taskTwo */ > void taskTwo (void) > > int i; > /* semGive (semB_Id); *//* Note 2: give up semaphore (a scheduling fix) > */ > taskSuspend(taskIdOne); > for (i=0; i { > semTake (semB_Id, WAIT_FOREVER); /* Wait indefinitely for semaphore > */ > global = 0; > printf ("I am taskTwo and global = %d-----------\n", global); > semGive (semB_Id); /* give up semaphore */ > } > taskResume(taskIdOne); > } > > /* taskThree */ > void taskThree (void) > { > semGive (semB_Id); /* give up semaphore */ > } > > /*******************************************************************/ > > We are using Tornado 2 prototyper. > > Regards > Renga - --------------C563E351ABD3119D7FA9C104 Content-Type: text/x-vcard; charset=us-ascii; name="vijay.peshkar.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Vijay Kumar Peshkar Content-Disposition: attachment; filename="vijay.peshkar.vcf" begin:vcard n:Peshkar;Vijay Kumar tel;work:5530034/35/35 x2237 x-mozilla-html:FALSE org:Wipro Global R & D;Wishing you blue skies, a couple of daisies, some fresh air and a smile! adr:;;;;;; version:2.1 email;internet:vpeshkar@cisco.com note:Wishing you blue skies, a couple of daisies, some fresh air and a smile! fn:Vijay Kumar Peshkar end:vcard - --------------C563E351ABD3119D7FA9C104-- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: vxsim memory size Date: 7 Jun 2001 06:57:23 GMT From: borkhuis@agere.com (Johan Borkhuis) Organization: Agere System Message-ID: References: <3B1E6D81.8A76BB77@digidata.com> Doug Dotson wrote: > Does anybody know how to increase the memory size of vxsim? See http://www.xs4all.nl/~borkhuis/vxworks/vxw_pt5.html#5.6-B 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: VxSim Date: 7 Jun 2001 10:30:33 GMT From: psenthilnathan@dsqsoft.com Organization: NewsOne.Net - Free Usenet News via the Web - http://newsone.net/ Message-ID: <9fnl49$7bo$1@news.netmar.com> References: <9fit2n$b9m$1@overload.lbl.gov> Followup-To: comp.os.vxworks In article <9fit2n$b9m$1@overload.lbl.gov>, sanjay krishnan writes: >Sir, >i would like to know whether we can input data from VxSim window. >i'm using tornado prototyper with VxSim. >i need to input data from the VxSim window in the following program: > Hi! You certainly can do that , whats the problem you are facing? Senthilnathan. Telecom & Networking, DSQ Software Ltd, Chennai-600035, India. Ph: 4346889 /901 /907 extn:2262/2389 email:- psenthilnathan@dsqsoft.com email:-nathan_senthil@rediffmail.com. ----- Posted via NewsOne.Net: Free (anonymous) Usenet News via the Web ----- http://newsone.net/ -- Free reading and anonymous posting to 60,000+ groups NewsOne.Net prohibits users from posting spam. If this or other posts made through NewsOne.Net violate posting guidelines, email abuse@newsone.net --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Using timer_create() and related routines..... Date: 7 Jun 2001 09:46:35 GMT From: psenthilnathan@dsqsoft.com Organization: NewsOne.Net - Free Usenet News via the Web - http://newsone.net/ Message-ID: <9fnihr$6jv$1@news.netmar.com> References: <9fif08$5mq$1@overload.lbl.gov> Followup-To: comp.os.vxworks In article <9fif08$5mq$1@overload.lbl.gov>, Priya Pradhan writes: >> Hi! >> >> I am trying to create a periodic timer on VxWorks using the routines - >> timer_create(), timer_connect(), timer_settime(). Everytime I try to run >> the program I get the error - "kill failed" - I can't understand the >> error. >> Hi! 1. Basically the error is that even before the timer expires the main routine is getting over. Means, if the while loop is made infinite , the expiry of the timer can be traced to the my_handler() routine.. 2. And the timer_connect() defines that the routine want to connect it to at the expiry of the timer should have the timer id as the first parameter..( which is missing in your case) Find here the modified code after removing those errors and unnnecesary lines.. Hope this solves your query............. Regs, Senthil nathan.P Telecom & Networking, DSQ Software Ltd, Chennai-35. Ph: 4346889/901/907 ext:2262 email: nathan_senthil@rediffmail.com, psenthilnathan@dsqsoft.com #include #include #include #include #include extern int errno ; void my_handler(timer_t tim,int param) ; timer_t timer_id ; int timers(void) { int i = 0 ; struct timespec tp_gettime, tp_settime, tp_clkres ; struct itimerspec time_val ; /* clock_getres(CLOCK_REALTIME, &tp_clkres) ; clock_gettime(CLOCK_REALTIME, &tp_gettime) ; printf("tp_gettime.tv_sec == %d\n",tp_gettime.tv_sec) ; printf("tp_gettime.tv_nsec == %d\n",tp_gettime.tv_nsec) ; */ tp_settime.tv_sec =0; /* Clock's resolution has to be set */ tp_settime.tv_nsec = 1; clock_settime(CLOCK_REALTIME, &tp_settime) ; if(timer_create(CLOCK_REALTIME, NULL, &timer_id) == ERROR) { printf("timer :: Error number set is == %d\n",errno) ; exit(1) ; } else printf("created timer.... \n") ; if(timer_connect(timer_id, my_handler, 0) == ERROR) { printf("timer :: Error no set is == %d\n",errno) ; exit(0) ; } else printf("connected timer.... \n") ; time_val.it_interval.tv_sec = 1 ; time_val.it_interval.tv_nsec = 0 ; time_val.it_value.tv_sec = 1 ; time_val.it_value.tv_nsec = 0 ; if(timer_settime(timer_id, CLOCK_REALTIME, &time_val, NULL) == ERROR) { printf("timer :: Error number set is == %d\n",errno) ; exit(1) ; } else printf("settime for timer.... \n") ; clock_gettime(CLOCK_REALTIME, &tp_gettime) ; printf("tp_gettime.tv_sec == %d\n",tp_gettime.tv_sec) ; printf("tp_gettime.tv_nsec == %d\n",tp_gettime.tv_nsec) ; /* while(i != 1000000) i++ ; */ while(1); return 0 ; } void my_handler(timer_t timer_id,int param) { struct itimerspec time_val ; printf("..........in function............ \n") ; fflush(stdout) ; /* time_val.it_interval.tv_sec = 1 ; time_val.it_interval.tv_nsec = 0 ; time_val.it_value.tv_sec = 1 ; time_val.it_value.tv_nsec = 0 ; if(timer_settime(timer_id, CLOCK_REALTIME, &time_val, NULL) == ERROR) { printf("my_handler :: Error number set is == %d\n",errno) ; exit(0) ; } */ return ; } ----- Posted via NewsOne.Net: Free (anonymous) Usenet News via the Web ----- http://newsone.net/ -- Free reading and anonymous posting to 60,000+ groups NewsOne.Net prohibits users from posting spam. If this or other posts made through NewsOne.Net violate posting guidelines, email abuse@newsone.net --------------------------- Newsgroups: comp.os.vxworks Subject: target board behavior on reboot() of vxworks Date: Thu, 07 Jun 2001 16:14:02 +0530 From: rajendra Organization: Lucent Technologies Message-ID: <3B1F5AF1.59A13C03@lucent.com> vxworks tornado Hi, I have some doubts about the vxworks reboot mechanism. In my system one of the target vxworks boards boots over the ether net and gets the vxworks image from the host board using TFTP. (host board also sits in the same rack and connected to the target through ether net). 1. If I execute the reboot command on the target, then does this target board again try to the get the vxworks image from the host using TFTP ? (Manual says control is returned to the BOOT PROMs) 2. On reboot with option BOOT_CLEAR is the current context lost ? Is this reboot equivalent to a cold start for the target board ? 3. On reboot with option BOOT_NORMAL the is current context retained ? 4. I have another board in the system which does not follow a networked boot procedure. Here vxworks image is loaded to the target and run from host through some other application that uses the backplane for this. I am not altering any boot parameters for this. There is no NVRAM in ths board. In this case how does the target behave on a reboot ? Thanks and best regards, rajendra. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks Tools questions Date: 7 Jun 2001 07:22:04 GMT From: borkhuis@agere.com (Johan Borkhuis) Organization: Agere System Message-ID: References: <3b1ee72a.4738173@news.qwest.net> joe-woodbury@mail.com (Joseph Woodbury) wrote: > What is the difference between CodeTEST and PerformancePAK > (ProfileScope & MemScope)? Same kind of tools, different implementation. > Are there significant differences between the [default] GNU compiler > and Diab C++? (We will be using C++ on a PowerQUICC 860) > How about between the default debugger and SingleStep? The GNU compiler is old, it is the 2.7.2 version of the compiler. There will be a new release soon, but it is not available. I don't know Singlestep, the default debugger is based on the GDB, the GNU debugger. Diab and Singlestep are commercial tools, so you have to purchase these seperate. > (Note, for the above mentioned pSOS, we were using Diab C++ and > SingleStep, though I was involved only for a few weeks of overlap when > the subsidiary formerly doing the development was shut down and I took > over.) In this case it might be possible to reuse the compiler and debugger. This might be an advantage for your developers. 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: Pended Semaphore behaviour Date: Thu, 7 Jun 2001 09:28:07 +0200 From: "Michael Lawnick" Organization: Lawrence Berkeley National Laboratory Message-ID: <9fnr8p$aav$1@overload.lbl.gov> - ----- Original Message ----- From: "Kumar" Newsgroups: comp.os.vxworks Sent: Wednesday, June 06, 2001 6:51 PM Subject: Re: Pended Semaphore behaviour > I did test the code. It seems to be working perfectly. > Here is the output. > > > I am taskTwo and global = 0----------- > I am taskTwo and global = 0----------- > I am taskTwo and global = 0----------- > I am taskTwo and global = 0----------- > I am taskTwo and global = 0----------- > I am taskTwo and global = 0----------- > I am taskTwo and global = 0----------- > I am taskTwo and global = 0----------- > I am taskTwo and global = 0----------- > I am taskTwo and global = 0----------- > I am taskOne and global = 1............. > I am taskOne and global = 1............. > I am taskOne and global = 1............. > I am taskOne and global = 1............. > I am taskOne and global = 1............. > I am taskOne and global = 1............. > I am taskOne and global = 1............. > I am taskOne and global = 1............. > I am taskOne and global = 1............. > I am taskOne and global = 1............. > > > I do hope that this is what you are expecting to happen. By seeing your logic. > No, I didn't see it that way, I saw it wrong. I missed that the semTake is in the loop of t2 (not only before). When calling binary from shell, your output is right. As t2 gives the sema at the end, it gets it again at the beginning. The same works t1. The problem comes up, when calling binary with 'sp': T2 can do only one output, its semGive will be eaten by the suspended T1 and the whole mechanism gets blocked ! - -> bad construction Mit freundlichen Grüßen, Michael Lawnick ============================================== SOFTEC GmbH Tel +49-731-96600-0 Promenade 17 Fax +49-731-96600-23 D-89073 Ulm Michael Lawnick Germany lawnick@softec.de ============================================== --------------------------- Newsgroups: comp.os.vxworks Subject: Initializing ATA for loading shell script Date: Thu, 7 Jun 2001 12:58:25 +0200 From: Fiederling.Ulf@dornier.dasa.de Organization: Lawrence Berkeley National Laboratory Message-ID: <9fnr9q$flg$1@overload.lbl.gov> Hi VxWorks Gurus We are using Tornado II on a Pentium platform How can we initialize the ATA device for loading scripts from the shell? It seems the only way to do this is in usrAppIni by calling usrAtaConfigt, but when this is done, the shell is already started. Any solution? Best regards Ulf Fiederling --------------------------- Newsgroups: comp.os.vxworks Subject: Re: target board behavior on reboot() of vxworks Date: Thu, 07 Jun 2001 18:03:13 +0100 From: Vijay Kumar Peshkar Organization: Cisco Systems Inc. Message-ID: <3B1FB3D1.C236A0F7@wipro.com> References: <3B1F5AF1.59A13C03@lucent.com> This is a multi-part message in MIME format. - --------------070DB223B6A86A5661F37558 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Rajendra, Please see comments inline. > In my system one of the target vxworks boards boots over the ether net and gets > the vxworks image from the host board using TFTP. (host board also sits in the > same rack and connected to the target through ether net). > > 1. If I execute the reboot command on the target, then does this target board > again try to the get the vxworks image from the host using TFTP ? (Manual says > control is returned to the BOOT PROMs) > Yes. flow is something like this.... reboot ->sysToMonitor->romInit->"tboot task"->bootcmdloop->tftpXfer > > 2. On reboot with option BOOT_CLEAR is the current context lost ? Is this reboot > equivalent to a cold start for the target board ? > Can you please explain what do you mean by "current conext". Yes. BOOT_CLEAR is equivalent to cold boot...see usrBootLineInit() in bootConfig.c > > 3. On reboot with option BOOT_NORMAL the is current context retained ? > > > 4. I have another board in the system which does not follow a networked boot > procedure. Here vxworks image is loaded to the target and run from host through > some other application that uses the backplane for this. I am not altering any > boot parameters for this. There is no NVRAM in ths board. > > In this case how does the target behave on a reboot ? > Here again the target board needs to communicate to the host. It may use a loader like program which takes it to a defined state before telling the host to send image. Hope this helps. cheers, Vijay - --------------070DB223B6A86A5661F37558 Content-Type: text/x-vcard; charset=us-ascii; name="vijay.peshkar.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Vijay Kumar Peshkar Content-Disposition: attachment; filename="vijay.peshkar.vcf" begin:vcard n:Peshkar;Vijay Kumar tel;work:5530034/35/35 x2237 x-mozilla-html:FALSE org:Wipro Global R & D;Wishing you blue skies, a couple of daisies, some fresh air and a smile! adr:;;;;;; version:2.1 email;internet:vpeshkar@cisco.com note:Wishing you blue skies, a couple of daisies, some fresh air and a smile! fn:Vijay Kumar Peshkar end:vcard - --------------070DB223B6A86A5661F37558-- --------------------------- Newsgroups: comp.os.vxworks Subject: Diab C++ fundamental access bug? Date: 7 Jun 2001 06:08:00 -0700 From: leonardmatherson@usa.net (Leonard Matherson) Organization: http://groups.google.com/ Message-ID: Hello, We are using the 4.3g Diab compiler under Tornado 2.0 Please consider the following: class X private: int XX; . . class Y X YY; foo(); . . foo() int any = YY.XX We should get a message saying that we do not have access rights to the private members of a class, but when we compile the following we get a clean compile and a good object file. Other compilers (GNU, VC6) of course give us the correct messages. Here is the source code that I compiled - no headers are needed. Please try this on your compiler, Thanks Leonard. class CTryacc { public: int ddd; private: int CanGet; }; class MiddleLevel { public: int sss; int fff(void); CTryacc TryAcc; }; int MiddleLevel::fff() { int yyy = TryAcc.CanGet; return (yyy); } void LetsSee(void) { MiddleLevel Upper; int nSome = Upper.fff(); } --------------------------- Newsgroups: comp.os.vxworks Subject: Re: vxworks pragma pack... Date: Thu, 07 Jun 2001 09:34:04 -0400 From: Joe Durusau Organization: http://bellsouth.webusenet.com - Home of the fastest NNTP servers on the Net. Message-ID: <3B1F82CC.E410FA12@bellsouth.net> References: <9fn2l6$jv3$1@overload.lbl.gov> I'm not sure if this is what you mean, but on certain types of targets you will get bus errors if you use pragma pack improperly. Speaking only for myself, Joe Durusau Chandrashekar Nagaraj wrote: > > hi, > i want to know the significance of argument to the "pragma_pack()" > system call in "C" language and "VxWorks"... > > Is it have anything more to do other than packing... > > with regards, > chandra. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: target board behavior on reboot() of vxworks : BOOT_NORMAL option Date: Thu, 07 Jun 2001 19:23:39 +0530 From: rajendra Organization: Lucent Technologies Message-ID: <3B1F8763.1E9F295E@lucent.com> References: <3B1F5AF1.59A13C03@lucent.com> <3B1FB3D1.C236A0F7@wipro.com> Vijay, Thanks for the reply. Could you please clarify the following ? 1. If BOOT_CLEAR is equivalent to cold boot then how does the system behave for a BOOT_NORMAL option ? When reboot with BOOT_NORMAL is executed, manual says that memory is not cleared. Does this mean that in a network boot environment the values stored in the RAM during the previous execution will be retained, a fresh image is obtained from the host and execution is started using this image ? 2. In this case, using reboot hooks can I store some parameters in memory and re use them after a reboot with BOOT_NORMAL option ? Thanks and best regards, rajendra. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: vxsim memory size Date: Thu, 07 Jun 2001 10:13:20 -0400 From: Doug Dotson Organization: Verio Message-ID: <3B1F8C00.C45FADB9@digidata.com> References: <3B1E6D81.8A76BB77@digidata.com> Johan, Thanks for the info. I did find out about the -r switch through WindSurf. It seems to work fine even on the default simulator if I start the simulator from the command line. I can't seem to figure out how to get the -r switch in place when launching the simulator from Tornado 2. Doug Johan Borkhuis wrote: > Doug Dotson wrote: > > > Does anybody know how to increase the memory size of vxsim? > > See http://www.xs4all.nl/~borkhuis/vxworks/vxw_pt5.html#5.6-B > > 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: Pended Semaphore behaviour Date: 7 Jun 2001 11:09:04 -0700 From: tlksk@hotmail.com (Kumar) Organization: http://groups.google.com/ Message-ID: References: <9fnr8p$aav$1@overload.lbl.gov> Hi Micheal Thanks for the clarification. Here is the scenario that you are facing. And it is and valid behaviour. 1. T1 Spawns --- Waits for semaphore and gets pended in the queue 2. T2 Spawns --- Waits for the semaphore and gets pended. 3. T3 Spawns 4. T3 gives the semaphore. 5. T2 takes the semaphore because of priority. T1 is still pended for the semaphore. 6. T2 gives the semaphore. Resulting in T1 getting control of the semaphore. 7. T1 cannot proceed because it is suspended. 8. T2 comes back to get the semaphore. Noone is giving so it waits. I hope that this time Iam precise. Kumar "Michael Lawnick" wrote in message news:<9fnr8p$aav$1@overload.lbl.gov>... > ----- Original Message ----- > From: "Kumar" > Newsgroups: comp.os.vxworks > Sent: Wednesday, June 06, 2001 6:51 PM > Subject: Re: Pended Semaphore behaviour > > > > I did test the code. It seems to be working perfectly. > > Here is the output. > > > > > > I am taskTwo and global = 0----------- > > I am taskTwo and global = 0----------- > > I am taskTwo and global = 0----------- > > I am taskTwo and global = 0----------- > > I am taskTwo and global = 0----------- > > I am taskTwo and global = 0----------- > > I am taskTwo and global = 0----------- > > I am taskTwo and global = 0----------- > > I am taskTwo and global = 0----------- > > I am taskTwo and global = 0----------- > > I am taskOne and global = 1............. > > I am taskOne and global = 1............. > > I am taskOne and global = 1............. > > I am taskOne and global = 1............. > > I am taskOne and global = 1............. > > I am taskOne and global = 1............. > > I am taskOne and global = 1............. > > I am taskOne and global = 1............. > > I am taskOne and global = 1............. > > I am taskOne and global = 1............. > > > > > > I do hope that this is what you are expecting to happen. By seeing your > logic. > > > No, I didn't see it that way, I saw it wrong. > I missed that the semTake is in the loop of t2 (not only before). > When calling binary from shell, your output is right. > As t2 gives the sema at the end, it gets it again at the beginning. > The same works t1. > The problem comes up, when calling binary with 'sp': > T2 can do only one output, its semGive will be eaten by the suspended T1 and > the whole mechanism gets blocked ! > -> bad construction > > Mit freundlichen Grüßen, > Michael Lawnick > ============================================== > SOFTEC GmbH Tel +49-731-96600-0 > Promenade 17 Fax +49-731-96600-23 > D-89073 Ulm Michael Lawnick > Germany lawnick@softec.de > ============================================== --------------------------- Newsgroups: comp.os.vxworks Subject: =?Big5?B?pOW+zLCqr8Wk5b7Msqa3fqTlvsy+x6RoutOkaL3StXs=?= Date: Thu, 7 Jun 2001 22:37:27 +0800 From: =?Big5?B?uOqwVLresnqxTbd+qPO3fA==?= Organization: Org Message-ID: <9fo3dl$4qc8496@rain.i-cable.com> IT Management Association UK University of Devonshire UK Any enquiry please contact Sunny Cheung Email: sunny@itma-uk.org URL: www.itma-uk.org Tel: 27108382 (Office hour) Add: Rm. 2103 Hollywood Plaza, 610 NAthan Rd., Mongkok, Lowloon. Open entry Anyone aged 17 or above is eligible to apply. No entry qualifications are required except for professional programs. Distance learning You will enjoy maximum flexibility through distance learning. You will learn through the printed materials, set texts and other materials that we suggest. Financial assistance Low-interest loans are available to help students with financial difficulties. Credit transfer You may transfer credits gained at local or overseas universities to our programs. It depends on your previous qualifications and their relevance to the program you wish to enroll. Recognition IT Management Association UK and University of Devonshire UK awards are widely recognized for employment¡Bprofessional bodies, and further studies both in Hong Kong and overseas. Institutional credit transfer arrangements have been signed with institutions in U.K., Canada, Australia, USA and New Zealand. Assessment Our programs are widely recognized for its high quality. Your study will be assessed by subject paper throughout a course taken and no examination is required. Tuition Master Degree HK$42,000 Bachelor Degree Final Stage HK$18,000- Diploma -> Bachelor Degree Combination HK$48,000- Diploma -> Graduate Diploma Combination HK$38,000- Diploma -> Advanced Diploma Combination HK$28,000- Diploma, Advanced Diploma, Graduate Diploma HK$18,000- Entry to any level diploma / final year of Bachelor Degree with previous education qualifications or working experience Bachelor of Business Administration Course Contents 1. Introduction to Business 2. Fundamental Accounting Principles 3. Approach to Management 4. Introduction to Business Communication 5. Principles of Economics 6. Financial Accounting 7. Marketing 8. Organizational Behavior 9. Introduction to Corporate Finance 10. Financial Markets 11. Entrepreneurship 12. Business Administration 13. Introduction to Human Resource Management 14. Introduction to Law for Business 15. Contemporary Advertising 16. International Marketing Bachelor of Science in Computer Science Course Contents 1. Introduction to Business 2. Computer Studies 3. Fundamentals of the Internet and World Wide Web 4. Learning C++ 5. Introduction to Computer Science 6. Fundamental Accounting Principles 7. Introduction to Languages and Computation 8. Software Engineering with Java 9. Database Systems Concepts 10. Introduction to Business Communication 11. Classical and Object-Oriented Software Engineering with Java and UML 12. Database Management Systems 13. Problem Solving Using C 14. Computer Organization 15. Computer Architecture 16. Machine Vision Bachelor of Science in Management Information Systems and Technology Course Contents 1. Introduction to Business 2. Communication Networks 3. Corporate Information Systems Management 4. Learning C++ 5. Fundamental Accounting Principles 6. Building the Information-Age Organization 7. Competitive Advantage (IT) 8. Software Engineering with Java 9. Introduction to Information Systems 10. Information Technology for Management 11. Information Systems 12. Accounting Information Systems 13. Principles of Management 14. Managing Information Systems in the Information Age 15. Classical and Object-Oriented Software Engineering with Java and UML 16. Problem Solving Using C Students completing all required subjects can obtain a Bachelor Degree from The University of Devonshire. The Bachelor Degree can normally be completed within 16 to 48 months. Students enrolled in the Graduate Diploma will complete 12 subjects from the list. The Graduate Diploma can normally be completed within 12 to 36 months. Students enrolled in the Advanced Diploma will complete 8 subjects from the list. The Advanced Diploma can normally be completed within 8 to 24 months. Students enrolled in the Diploma will complete 4 subjects from the list. The Diploma can normally be completed within 4 to 12 months. Master's of Business Administration Course Contents 1. Corporate Finance Essentials 2. Managerial Economics 3. Principles of Management II 4. Marketing Management 5. Management Competition 6. Integrated Business 7. Modern Business Law 8. Advanced Financial Accounting 9. Business Communication in the Professional Setting 10. Consumer Behavior and Marketing Strategy 11. Global Marketing E-Commerce MBA Course Contents 1. International Business 2. Cases in Electronic Commerce 3. E-Business 4. Management Information Systems for the Information Age 5. Investment Experience 6. Security, Risk Management, and Control 7. E-Business: Roadmap to Success 8. Internet Marketing 9. E-Business Revolution 10. Accounting Information Systems 11. Global Marketing Master's in Computer Science Course Contents 1. Advanced Studies in Programming Languages 2. Database Systems 3. Computer Systems: Organization and Operating Systems 4. Advanced Studies in Algorithm Design and Analysis 5. Software Engineering 6. Computer Graphics 7. Network Design and Analysis 2 Students completing all required subjects could obtain a Master Degree from The University of Devonshire. The Master Degree can normally be completed within 9 to 24 months. 2 Students completing 2/3 required subjects from the list could obtain a Postgraduate Diploma. 2 Students completing 1/3 required subjects from the list could obtain a Postgraduate Certificate. ±M·~¤å¾Ì (Professional Diploma) ½Òµ{¡G ±M·~¡þ±M®a¤å¾Ì½Òµ{¦®¦b´£¨Ñµ¹¨º¨Ç»Ý­n±M·~¤Î±Mªù¯S®í¦æ·~§Þ³NªÌªº¾Ç³N°V½m¡AÅý¥L­Ì¨ã³Æ±q¨Æ¸Ó±Mªù¦æ·~ªº¤èªk¤Î¨t¦Cªºª¾Ãѧޯà¡C ³øŪ¦U±M·~¤å¾Ì½Òµ{±ø¥ó¡G À³©¡­»´ä¤¤¾Ç²¦·~¥Í¶··|¦Ò­^¤å¬ì¥Ø¹F¨ì¢Ó¯Å©Î¥H¤W¡A ©Î¤¤¾Ç²¦·~µ{«×¨ã³Æ¨}¦n­^»y¾\Ū¤Î¼g§@¯à¤O¡A ©Î¤w§¹¦¨­»´ä°ªµ¥µ{«×·|¦Ò¤§¾Ç­û¡A ©Î19·³¥H¤W¦¨¦~¾Ç­û¡A¨ã³Æ¤@¦~¬ÛÃö¤u§@¸gÅç¡C 2 ¹q¸£¨t²Î¤Îºôµ¸§Þ³N¤ä´©¤uµ{®v±M·~¤å¾Ì½Òµ{ Professional Diploma in Computer System & Network Support Engineer ¦¹½Òµ{¥Øªº¦b©óµ¹¤©¦³§Ó±q¨Æ©Î²{¥¿±q¨Æ¹q¸£¨t²Î§Þ³N¤ä´©¤H¤h¡A´£°ª±M·~ª¾ÃѤιê½î¾÷·|¡A¨ä­×Ū¤§¬ì¥Ø¦p¤U¡G 1. Hardware Introduction - I¡]¹q¸£µw¥ó¤@¡^ Computer Keyboards, Data Processing, Electrical requirements of computer systems, Printer Management, Disk Management, Central Processing Units and Memory, Storage Systems, Input and Output Devices, Data Communications, Networking, Process Control. 2. Hardware Introduction - II ¡]¹q¸£µw¥ó¤G¡^ Processors, Memory and Benchmarking, Buses and Interfaces, Mass Storage Devices, Input Output Devices, Power Supplies, Testing Procedures, Configuring Peripheral Devices. 3. Operating System ¡]¹q¸£¾Þ§@¨t²Î¡^¡¯ What is an Operating System, Various parts of an Operating System, What is a Process, Operating System Load, File Management, Software Tools, Basic Features of Graphical Interfaces, Windows Objects and Components. 4. Windows 95/98 and MS-DOS Operating System ¡]µøµ¡¤E¤­¡þ¤E¤K¤ÎºÏºÐ¾Þ§@¨t²Î¡^¡¯ Why Windows 95/98, Basic Features, Starting Up, Using Applications, Managing Files and Folders, Managing the Desktop. 5. Introduction to Networking ¡]¹q¸£ºôµ¸¡^¡¯ Network Concepts Components, Adapters, Wiring Guide, ELA TIA-568 Standard, Twisted Pair, Coaxial, Fiber, OSI Model, Network Segments, Spanning Tree Algorithm, Repeaters, Bridges, Routers, Hubs, Ethernet Switches, Virtual Networking Network Topology Bus, Bus-Physical Layout, Ethernet 802.3-CSMA/CD, 10Base2-Ethernet, 10Base5-Ethernet, 10BaseT-Ethernet, Bus-Physical Cable Limits, Ring, Ring-Physical Layout, IEEE 802.5 Token Ring, Ring-Physical Cable Limits, Star, FDDI, Logical vs Physical. Connecting Computers Peer to Peer, Workgroups, Workgroup Resources, Belonging to a Workgroup, Sharing Resources, Accessing Resources, Novell File and Print Services, Domains, User Accounts, Computer Accounts, Groups, Domain Models, Single Domain Model, Trust Relationships, Single Master Domain Model, Complete Trust Domain Model. Troubleshooting Cable Testing, General Troubleshooting Principles, Hardware Diagnostics, Using a TDR, Using a multi-meter. 6. Windows 2000 Server/NT Server 4.0 ¡]µøµ¡ºô¸ô¦øªA¾¹¢±¢¯¢¯¢¯¡þ¢³¡D¢¯¡^¡¯ Operating Systems, Windows NT Services, Workgroups and Domains, Domain Name System, Active Directory, Administering Active Directory, Security and Permissions, Supporting Users, Managing S/W Applications, Managing Printers, System Recovery. 7. Introduction to Internet and Web Pages Development¡]¤¬ÁpºôÀ³¥Î¤Îºô­¶¶}µo¡^¡¯ Understand basic Internet concepts and terminology, Use a browser to explore the WWW, Locate information using search engines, Use FrontPage to create web sites, Use E-mail, Video and communication tools. 8. Linux Server Engineering ¡]Linux ºô¸ô¦øªA¾¹¡^ Hardware and Installation (x86 Architecture), Configuration and Administration, Alternate Installation Methods, Kernel Services and Configuration, TCP/IP Networking, Samba File Server Configuration, Apache Web Server Configuration, NFS Network File Service Configuration, X Window System, User and Host Security, Routers & Firewalls Configuration. 2 ºô¯¸¶}µo¤ÎºÞ²z§Þ³N¤ä´©¤uµ{®v±M·~¤å¾Ì½Òµ{ Professional Diploma in Web Site Development & Management Support Engineer ¦¹½Òµ{¥Øªº¦b©óµ¹¤©¦³§Ó±q¨Æ©Î²{¥¿±q¨Æºô¯¸ºÞ²z¤u§@¤§¤H¤h¡A´£°ª±M·~ª¾ÃѤιê½î¾÷·|¡A¨ä­×Ū¤§¬ì¥Ø¦p¤U¡G 1. Operating System ¡]¹q¸£¾Þ§@¨t²Î¡^¡¯ What is an Operating System, Various parts of an Operating System, What is a Process, Operating System Load, File Management, Software Tools, Basic Features of Graphical Interfaces, Windows Objects and Components. 2. Windows 95/98 and MS-DOS Operating System ¡]µøµ¡¤E¤­¡þ¤E¤K¤ÎºÏºÐ¾Þ§@¨t²Î¡^¡¯ Why Windows 95/98, Basic Features, Starting Up, Using Applications, Managing Files and Folders, Managing the Desktop. 3. Introduction to Networking ¡]¹q¸£ºôµ¸¡^¡¯ Network Concepts Components, Adapters, Wiring Guide, ELA TIA-568 Standard, Twisted Pair, Coaxial, Fiber, OSI Model, Network Segments, Spanning Tree Algorithm, Repeaters, Bridges, Routers, Hubs, Ethernet Switches, Virtual Networking Network Topology Bus, Bus-Physical Layout, Ethernet 802.3-CSMA/CD, 10Base2-Ethernet, 10Base5-Ethernet, 10BaseT-Ethernet, Bus-Physical Cable Limits, Ring, Ring-Physical Layout, IEEE 802.5 Token Ring, Ring-Physical Cable Limits, Star, FDDI, Logical vs Physical. Connecting Computers Peer to Peer, Workgroups, Workgroup Resources, Belonging to a Workgroup, Sharing Resources, Accessing Resources, Novell File and Print Services, Domains, User Accounts, Computer Accounts, Groups, Domain Models, Single Domain Model, Trust Relationships, Single Master Domain Model, Complete Trust Domain Model. Troubleshooting Cable Testing, General Troubleshooting Principles, Hardware Diagnostics, Using a TDR, Using a multi-meter. 4. Windows 2000 Server/NT Server 4.0 ¡]µøµ¡ºô¸ô¦øªA¾¹¢±¢¯¢¯¢¯¡þ¢³¡D¢¯¡^¡¯ Operating Systems, Windows NT Services, Workgroups and Domains, Domain Name System, Active Directory, Administering Active Directory, Security and Permissions, Supporting Users, Managing S/W Applications, Managing Printers, System Recovery. 5. Introduction to Internet and Web Pages Development¡]¤¬ÁpºôÀ³¥Î¤Îºô­¶¶}µo¡^¡¯ Understand basic Internet concepts and terminology, Use a browser to explore the WWW, Locate information using search engines, Use FrontPage to create web sites, Use E-mail, Video and communication tools. 6. Internet Management¡]¤¬ÁpºôºÞ²z¡^ Introduction to Internet Management, Networks Components, Network Topology and Cabling, Security, Management, Auditing, Intranets, Extranets, WAN's, Managing Users, Security of data, Clistering, UPS, Backup, Virus, Capacity Planning, Asset Management, Managing Web Sites, Log Files, Tracking Users. 7. Web Sites Development and Programming using JAVA Script ¡]JAVA Script ºô¯¸¶}µo¤Îµ{¦¡½s¼g¡^ What is JavaScript, HTML-Document, Frames Programming, Status bar and Timeouts Programming, Predefined Objects, Forms Programming, Image Object. _____________________________________________________________ This message posted with trial version of Express News Poster --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks Tools questions Date: Thu, 07 Jun 2001 13:12:56 -0500 From: Andrew Johnson Organization: APS, Argonne National Laboratory Message-ID: <3B1FC428.817D241B@aps.anl.gov> References: <3b1ee72a.4738173@news.qwest.net> Johan Borkhuis wrote: > > > Are there significant differences between the [default] GNU compiler > > and Diab C++? (We will be using C++ on a PowerQUICC 860) > > The GNU compiler is old, it is the 2.7.2 version of the compiler. There > will be a new release soon, but it is not available. Not quite true, the C compiler they ship is still based on gcc-2.7.2, but the C++ compiler is based on the newer but still old EGCS-1.0.2 release (see the Tornado 2.0 Release Notes, section 7.6). They just combined the two such that if you ask for the C++ compiler's version you actually get given that of the C compiler. The most annoying thing that's missing is C++ namespaces; other than that it's about as good as MS VC++ in its support of the ISO standard, but in different ways. - - Andrew - -- The world is such a cheerful place when viewed from upside-down It makes a rise of every fall, a smile of every frown --------------------------- Newsgroups: comp.os.vxworks Subject: Shell with WXworks Date: 7 Jun 2001 11:30:19 -0700 From: fburgun@club-internet.fr (Francis Burgun) Organization: http://groups.google.com/ Message-ID: <43e840db.0106071030.2ca3d802@posting.google.com> We have a strange problem with the shell on the serial link (the code is loaded in flash on a POWERPC board. On one off the serial links we have access with an hyperterminal to a menu with many choices. One off this choice is to have a shell. The prompt of the shell appears but we can not execute any shell commande like help, i etc. The error message that appears is command not found. If anybody has any idea of a solution or what this problem could be, y would be very greatful of your reply. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Pended Semaphore behaviour Date: Thu, 07 Jun 2001 19:23:32 +0100 From: Simon Farnsworth Message-ID: <3B1FC6A4.6000804@snellwilcox.com> References: <9fnr8p$aav$1@overload.lbl.gov> Kumar wrote: > 1. T1 Spawns --- Waits for semaphore and gets pended in the queue > 2. T2 Spawns --- Waits for the semaphore and gets pended. > 3. T3 Spawns > 4. T3 gives the semaphore. > 5. T2 takes the semaphore because of priority. T1 is still > pended for the semaphore. > 6. T2 gives the semaphore. Resulting in T1 getting control of > the semaphore. > 7. T1 cannot proceed because it is suspended. > 8. T2 comes back to get the semaphore. Noone is giving so it waits. > This is as it is meant to be. taskSuspend() is a debugging aid. The idea is that the programmer will suspend a task before some critical section that is misbehaving, and then resume it manually, with a debugger attached or after checking some form of state information It is *NOT* intended to be used in general code (my copy of the reference manual says: "This routine is the basis of the debugging and exception handling packages. However, as a synchronisation mechanism, this facility should be rejected...) --------------------------- Newsgroups: comp.os.vxworks Subject: =?Big5?B?uNu4db3StXuxwLxzrfs=?= Date: Fri, 8 Jun 2001 00:14:02 +0800 From: =?Big5?B?sU23fr7Ht3ytXrDqpGq+xw==?= Organization: Professional Organization/UK University Message-ID: <9fo92o$4qd8501@rain.i-cable.com> ±M·~¾Ç·|¡þ­^°ê¤j¾Ç ¸Û¸u½Òµ{±À¼s­û¡A¤¤¤­©Î¥H¤Wµ{«× À³©¡¤¤¤­²¦·~©Î¹w¬ì¥Í¥ç¥i ¦þª÷Àu«p¡A¦³·N½Ð¹q¢±¢¶¢°¢¯¢·¢²¢·¢± Sunny Cheung ¬¢ _____________________________________________________________ This message posted with trial version of Express News Poster --------------------------- Newsgroups: comp.os.vxworks Subject: MVME5100, does the BSP support writing to Flash? Date: Thu, 07 Jun 2001 21:59:14 GMT From: Martijn Kamerbeek Organization: Sierra Data Systems Message-ID: <3B1FE378.A663AC4@sierradatasys.com> Hi, I am currently working with a MVME1604 board and I see that the BSP has routines such as sysFlashSet and SysFlashWrite. Soon I'll be working with a MVME5100 and I would like to read and write to Flash memory. However, I can't find any indication anywhere that the 5100 BSP will support writing to and reading from Flash. Does anybody have some information for me? Thank you very much!! Martijn Kamerbeek --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Redirecting stdout problems Date: 7 Jun 2001 15:09:09 -0700 From: tlksk@hotmail.com (Kumar) Organization: http://groups.google.com/ Message-ID: References: <3B1F4C8F.E137BEF@wipro.com> Bill, Set the ioGlobalStdSet(1,fd); to an local file. Issue the necessary command. And then proper reading and interperting the file. This would be an better alternative than pumping to an message queue. If you cannot do on a file open a mempory file. Kumar Vijay Kumar Peshkar wrote in message news:<3B1F4C8F.E137BEF@wipro.com>... > Bill, > > You can try by sending commands to the target shell using simple ASCII (sprintf()) > and get responses. You can parse the responses with sscanf(). > > Hope this helps, > Vijay > > Bill Harmon wrote: > > > I am having problems redirecting stdout. I am trying to provide > > the ability for our the CLI in our application to call certain > > VxWorks functions such as i(), ifShow(), and ti() so that they > > may be executed via our CLI through telnet (not the Tornado > > shell). For some of these functions, the source code is provided > > so I can easily direct the output to our routines, however some > > functions are not provided in source form. As a result, I am > > trying to come up with a general scheme to redirect the output > > of stdout so that I can capture the output of these routines, > > and send it to our routines for output to the user via telnet. > > First, is redirecting stdout the best (only?) way to accomlish > > this? Assuming this is the best way to proceed, I wrote some code > > which redirects the stdout to a pipe using ioGlobalStdSet(). > > While this works in most cases, it does not seem to work reliably, > > and sometimes crashes the system. Further, in order to capture the > > output of these routines I found that I needed to create a pipe > > such as > > > > pipeDevCreate ("/pipe/cli", 1000, 200); > > > > The effect is, I beleive, reserving a relatively huge amount of > > memory for the queues for this function. This does not seem right to > > have to allocate this amount of memory for this purpose. I have not > > used pipes before so part of the problem may be my misundertanding > > of how to use them correctly. If I use smaller amount of max_msgs or > > max_msg_len, I find that I miss portions of the output generated by > > i(), for example. > > > > The code fragment that I am using to capture the output of stdout > > and send it to my routine is as follows; > > > > // Open the pipe for STDOUT redirection > > fd = open ("/pipe/cli", O_RDWR, 0644); > > > > if (fd != ERROR) { > > > > // Setup the I/O streams for redirection > > taskLock(); > > stdin_fd = ioGlobalStdGet(0); > > stdout_fd = ioGlobalStdGet (1); > > stderr_fd = ioGlobalStdGet (2); > > ioGlobalStdSet(0,fd); > > ioGlobalStdSet(1,fd); > > ioGlobalStdSet(2,fd); > > taskUnlock(); > > > > // Clear out and residual data in the CLI pipe > > ioctl (fd, FIOFLUSH, 0); > > > > // Call the actual function which does the work > > i(); // <--- I want to capture the output of this function > > > > // Replace the original I/O streams for this task > > taskLock(); > > ioGlobalStdSet(0,stdin_fd); > > ioGlobalStdSet(1,stdout_fd); > > ioGlobalStdSet(2,stderr_fd); > > taskUnlock(); > > > > // Initialize the output buffer to NULL > > strcpy(pipeBuf,""); > > > > // Get the number of messages queued up for us... > > ioctl (fd, FIONMSGS, (int) &nMessages); > > > > // Get the output from the pipe > > while (nMessages--) { > > len = read (fd, dummyBuf, 200); > > strncat(pipeBuf, dummyBuf, len); > > if (strlen(pipeBuf) >= 80) { > > // Send pipeBuf to the output routines here... > > strcpy(pipeBuf,""); > > } > > } > > > > // Flush any residual data > > if (strlen(pipeBuf)) { > > // Send pipeBuf to the output routines here... > > } > > > > // Close the pipe for STDOUT redirection > > close (fd); > > } > > > > Is it correct that I pipeDevCreate needs to allocate a queue this > > large to capture this output? Is there a better was to accomplish > > what I am trying to do? > > > > Thanks in advance, > > > > Bill Harmon > > -- --------------------------- Newsgroups: comp.os.vxworks Subject: how to find the right component to link in? Date: Thu, 7 Jun 2001 16:06:54 -0700 From: "Gang Wang" Organization: None Message-ID: Hi, The question is regarding to how to find out what vxWorks component is needed when you build your image. When this is a forward question, it is easy. For example, if I know that I will use select to do file I/O, I could just go to the component tree and link the Operating System Components/IO System components/select module into my image. However, things get hard ( at least to me ) when the question is reversed, i.e., I know a specific function I have to use then how I could find out which component(s) contains it? For example, can someone tell me which component I should use for function rmdir() and mkdir()? Or in other words, where is the usrLib? Thanks. --------------------------- Newsgroups: comp.os.vxworks Subject: loadModule on from fs, not from memory Date: Thu, 07 Jun 2001 23:50:57 GMT From: "Kelly Hornsby" Organization: Excite@Home - The Leader in Broadband http://home.com/faster Message-ID: Hello All Is it true that loadModule may only be used from a file system (TFFS, ramdisk, whatever) and not from memory? I'm looking for options to load in a module dynamically from flash memory, but without using a structured file system. Any suggestions or ideas are welcome. --------------------------- Newsgroups: comp.os.vxworks Subject: Gratuitous ARP and etherAddrResolve Date: Thu, 7 Jun 2001 22:48:24 -0400 From: "Albert So" Organization: Sprint Canada Inc. Message-ID: <73XT6.1900$F42.42217@newscontent-01.sprint.ca> Reply-To: "Albert So" I am trying to implement gratuitous ARP using the etherAddrResolve function. To recap, the function is as follows: STATUS etherAddrResolve ( struct ifnet * pIf, /* interface on which to send ARP req */ char * targetAddr, /* name or Internet address of target */ char * eHdr, /* where to return the Ethernet addr */ int numTries, /* number of times to try ARPing */ int numTicks /* number of ticks between ARPing */ ) For a gratuitous ARP, the target internet address is the same as the originator's IP address. When I tried this, I found that the ARP was not sent at all. It worked though when the target was another machine. Is there a way to force an ARP to be sent regardless whether the target address is the same as the originator or not. Thanks in advance! Albert --------------------------- Newsgroups: comp.os.vxworks Subject: Re: target board behavior on reboot() of vxworks : BOOT_NORMAL option Date: Fri, 08 Jun 2001 10:00:25 +0100 From: Vijay Kumar Peshkar Organization: Cisco Systems Inc. Message-ID: <3B209429.3A6522A0@wipro.com> References: <3B1F5AF1.59A13C03@lucent.com> <3B1FB3D1.C236A0F7@wipro.com> <3B1F8763.1E9F295E@lucent.com> This is a multi-part message in MIME format. - --------------15D86443FA42530A43C114B0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Rajendra, > 1. If BOOT_CLEAR is equivalent to cold boot then how does the system behave for > a BOOT_NORMAL option ? > When reboot with BOOT_NORMAL is executed, manual says that memory is not > cleared. Does this mean that in a network boot environment the values stored in > the RAM during the previous execution will be retained, a fresh image is > obtained from the host and execution is started using this image ? BOOT_NORMAL seems to be warm boot. So mem is not cleared and POST is skipped. The RAM should remain untouched no matter what its reload method is. Yes. We only need to get the image from host and run it just as in cold boot. > > 2. In this case, using reboot hooks can I store some parameters in memory and re > use them after a reboot with BOOT_NORMAL option ? As I understand, reboot hooks are normally used to reset some extra hardware or tell some peripheral devices some thing special. It may not be proper to use it to store parameters. A simpler way I was using was to isolate some memory for my POST results and (in your case reboot parameters) on above sysMemTop(). In kernelInit, you can say something like "sysMemTop - MY_REBOOT_STAT_SIZE". This mem will always be exempt from POST or mem clear etc. Also we need to be extra careful that the hook does not cause any exception as this may mot lead to proper reboot..... good luck. regards, Vijay - --------------15D86443FA42530A43C114B0 Content-Type: text/x-vcard; charset=us-ascii; name="vijay.peshkar.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Vijay Kumar Peshkar Content-Disposition: attachment; filename="vijay.peshkar.vcf" begin:vcard n:Peshkar;Vijay Kumar tel;work:5530034/35/35 x2237 x-mozilla-html:FALSE org:Wipro Global R & D;Wishing you blue skies, a couple of daisies, some fresh air and a smile! adr:;;;;;; version:2.1 email;internet:vpeshkar@cisco.com note:Wishing you blue skies, a couple of daisies, some fresh air and a smile! fn:Vijay Kumar Peshkar end:vcard - --------------15D86443FA42530A43C114B0-- --------------------------- Newsgroups: comp.os.vxworks Subject: Interrupts with x86 BSP Date: Fri, 8 Jun 2001 14:14:46 +0930 From: "David Hichens" Message-ID: <3b205862_3@news01.one.net.au> G'Day, I am trying to get interrupts working with a PC-104 board running a NS Geode processor (Pentium MMX equivalent). Is it necessary to include code to send an EOI command to the interrupt controller in the interrupt service routine or is this handled automatically if the service routine is linked in using the intConnect() function? Regards, David Hichens. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Interrupts with x86 BSP Date: Fri, 08 Jun 2001 11:34:52 +0100 From: Vijay Kumar Peshkar Organization: Cisco Systems Inc. Message-ID: <3B20AA4C.DD4202AE@wipro.com> References: <3b205862_3@news01.one.net.au> This is a multi-part message in MIME format. - --------------DAED1B39801BDCFB8ECA098E Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit David, Yes. You need to call EOI. The stubs added by intConnect above and below your ISR only handle the registers and stack. regards, Vijay David Hichens wrote: > G'Day, > > I am trying to get interrupts working with a PC-104 board running a NS Geode > processor (Pentium MMX equivalent). > > Is it necessary to include code to send an EOI command to the interrupt > controller in the interrupt service routine or is this handled automatically > if the service routine is linked in using the intConnect() function? > > Regards, > > David Hichens. - --------------DAED1B39801BDCFB8ECA098E Content-Type: text/x-vcard; charset=us-ascii; name="vijay.peshkar.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Vijay Kumar Peshkar Content-Disposition: attachment; filename="vijay.peshkar.vcf" begin:vcard n:Peshkar;Vijay Kumar tel;work:5530034/35/35 x2237 x-mozilla-html:FALSE org:Wipro Global R & D;Wishing you blue skies, a couple of daisies, some fresh air and a smile! adr:;;;;;; version:2.1 email;internet:vpeshkar@cisco.com note:Wishing you blue skies, a couple of daisies, some fresh air and a smile! fn:Vijay Kumar Peshkar end:vcard - --------------DAED1B39801BDCFB8ECA098E-- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Shell with WXworks Date: Thu, 07 Jun 2001 23:28:22 -0700 From: DrDiags Organization: laserlink.net Message-ID: <3B207086.5198DDE8@covad.net> References: <43e840db.0106071030.2ca3d802@posting.google.com> Reply-To: drdiags@covad.net Francis, Did you add in symbols for the Target Shell? Make sure the kernel is configured to use either a downloaded symbol table or build one in if you are not strapped for memory. These options are off by default. Francis Burgun wrote: > > We have a strange problem with the shell on the serial link (the code > is loaded in flash on a POWERPC board. > > On one off the serial links we have access with an hyperterminal to a > menu with many choices. > > One off this choice is to have a shell. > > The prompt of the shell appears but we can not execute any shell > commande like help, i etc. > > The error message that appears is command not found. > > If anybody has any idea of a solution or what this problem could be, y > would be very greatful of your reply. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: TCP Theory question puzzling me Date: Fri, 8 Jun 2001 08:51:31 +0200 From: "Michel Hendriks" Organization: XS4ALL Internet BV Message-ID: <9fpslu$7bk$1@news1.xs4all.nl> References: Sender: tcj@gateway.pijnenburg.nl Geoff wrote in message news:kfaR6.8471$zb7.1100889@news1.cableinet.net... > Hello > > I am developing an embedded application, which requires DHCP client > functionality (Not Intel/Windows based). It was fairly easy to establish an > IP address using the DHCP call provided by the Operating System. However, > what is puzzling me is, now that I have a dynamic IP address, how do other > machines address this machine ? I would like them to access this machine by > host name, but what mechanism do I use, to tell the DNS server my name and > DHCP'd IP address? Any help or insight into this problem would be greatly > appreciated ? Thanks > > Geoff > Hi, Use the zone transfer protocol of DNS. rfc1035 should get you started. Look at ftp://ftp.isi.edu/in-notes/rfc1035.txt Greetings --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Controling point at which socket write blocks Date: Fri, 8 Jun 2001 08:57:18 +0200 From: "Michel Hendriks" Organization: XS4ALL Internet BV Message-ID: <9fpt0o$87e$1@news1.xs4all.nl> References: Sender: tcj@gateway.pijnenburg.nl nick wrote in message news:d7f51a07.0106040647.6d4b776a@posting.google.com... > Hello, > > We want to change the point at which at socket blocks. Our system may > have in excess of 120 sockets simultaneously writing, and I would like > each socket to block after 400 bytes or so has been written without a > corresponding read. > > I have tried changing the send buffer, but this just slows things down > (I think it does not work because data is copied from the send buffer > to another vxworks / tcp buffer). > > The other methods, such as recv with MSG_PEEK, and querying the > FIONREAD values only work at the other end of the socket (which is > done by another company / software, and we have no control of). > > All this seems to point me in the direction of tweaking mbufs, mblks, > clusters, etc., but to be honest I don't have a clue about all that. > > Any help would be greatly appreciated, > > Regards, > > Nick > > nick.gardner@europe.simoco.com Hi, There is no such possibility in TCP/IP. You have to either solve it in your protocol (as in: wait for a response from the other end before transmitting more data). Or, if that is not possible, you could try to get it to work by suspending each task after it sends 400 bytes for a certain period of time (1 second?). But I suppose this will be a hell of lot of tweaking. With 120 sockets, that may prove to hard to do.. Greetings Michel --------------------------- Newsgroups: comp.os.vxworks Subject: Re: loadModule on from fs, not from memory Date: 8 Jun 2001 08:04:33 GMT From: borkhuis@agere.com (Johan Borkhuis) Organization: Agere System Message-ID: References: "Kelly Hornsby" wrote: > Hello All > > Is it true that loadModule may only be used from a file system (TFFS, > ramdisk, whatever) and not from memory? I'm looking for options to > load in a module dynamically from flash memory, but without using a > structured file system. Any suggestions or ideas are welcome. As you need a file descriptor for loadModule you also need a file to open. But you can use a memDrv, see the manuals for more details on this. Groeten, Johan - -- o o o o o o o . . . _____________________________ o _____ || Johan Borkhuis | .][__n_n_|DD[ ====_____ | borkhuis@agere.com | >(________|__|_[_________]_|__________________________| _/oo OOOOO oo` ooo ooo 'o!o!o o!o!o` === VxWorks FAQ: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html === --------------------------- Newsgroups: comp.os.vxworks Subject: equivalent of dn_expand in vxworks Date: 8 Jun 2001 03:04:13 -0700 From: sarveshku@lycos.com (Sarvesh) Organization: http://groups.google.com/ Message-ID: hi, can anyone tell me how to expand a compressed domain name in vxworks. is there any equivalent function call in vxworks as dn_expand in unix machines. thanx in advance, regards, sarvesh --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Fri Jun 8 07:00:46 2001 From: "Thomas Rahn" Date: Fri Jun 8 07:00:51 PDT 2001 Subject: Re: Interrupts with x86 BSP Hi David, there is a simple answer: you do not have to do take care of signaling EOI. Our int handlers typically look like void ataIntr(int ctrl) { sysInByte(0x1f7); // clears the irq condition semGive (syncSem); // continue waiting task } But there might be more (depends on your problem): Int handling on x86 is sligthly weird because of the two cascaded interrupt controllers. Ints must be acknowledged to the int controller by some I/O write. The EOI must be send to both int controllers if the int originates from the slave device (int 8..15). VxWorks wraps your interrupt handler with some code that should include a call to i8259IntEoiSlave() or similar. [Try to disassemble the address returned by intVecGet()] In VxWorks 5.3 the int controller software always ack'd both (no harm done). With VxWorks 5.4, the driver software changed. It was improved to spare the extra IO cycle if not required. However, the code got somehow mixed up and it might happen that no EOIs are sent at all. This happens if you use the combination intHandlerCreate/intVecSet instead of intConnect (never trust the docs, intConnect uses a private function intHandlerCreateX86...) This special problem can be solved, if you change the file sysLib.c back to the old style: find /* initialize the PIC (Programmable Interrupt Controller) */ sysIntInitPIC (); intEoiGet = sysIntEoiGet; /* function pointer used in intConnect () */ and replace with /* set the global function pointer to sysIntEOI() */ intEOI = sysIntEOI; Bye Thomas Rahn From vxwexplo-errs@csg.lbl.gov Fri Jun 8 07:31:06 2001 From: Wilco Rademaker Date: Fri Jun 8 07:31:09 PDT 2001 Subject: how to check whether fpp register contents is a number? Hello, We are using vxworks 5.3.1 with PowerPC750 board. with fppTaskRegsShow we get the following output: fppTaskRegsShow 0x3b89d10 fpcsr = 70 fr0 = NaN fr1 = NaN fr2 = NaN fr3 = NaN fr4 = NaN fr5 = NaN fr6 = NaN fr7 = NaN fr8 = NaN fr9 = NaN fr10 = NaN fr11 = NaN fr12 = NaN fr13 = NaN fr14 = NaN fr15 = NaN fr16 = NaN fr17 = NaN fr18 = NaN fr19 = NaN fr20 = NaN fr21 = NaN fr22 = NaN fr23 = NaN fr24 = NaN fr25 = NaN fr26 = NaN fr27 = NaN fr28 = NaN fr29 = NaN fr30 = NaN fr31 = NaN We would like to be able to perform the test on the registers for NaN (Not a Number). We were unable to trackdown the functioncall to do this. On Solaris the function isnan() is available to perform this. Who can tell us which function we should use on vxworks or with what code we perform the isnan test? Kind regards, -- Wilco Rademaker ASM Lithography phone: +31-(0)40-268.4311 PO Box 324 fax: +31-(0)40-268.4999 Room 7A2013 mailto:wilco.rademaker@asml.nl (work) 5500 AH VELDHOVEN mailto:wilco.rademaker@bigfoot.com (private) The Netherlands From vxwexplo-errs@csg.lbl.gov Sat Jun 9 04:03:18 2001 From: Vxworks Exploder Date: Sat Jun 9 04:03:20 PDT 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sat Jun 9 04:03:07 PDT 2001 Subject: Intel 82559ER driver on Tornado 2.0.2? Subject: Re: Controling point at which socket write blocks Subject: Re: How to run VxWorks on an x86 machine? Subject: Re: Network Stack Subject: Re: Redirecting stdout problems Subject: Re: Using GDB 5 (insight, cygwin) with VxWorks (WinNT) ? Subject: Re: ¤å¾Ì°ª¯Å¤å¾Ì²¦·~¤å¾Ì¾Ç¤hºÓ¤h½Òµ{ Subject: VxSim on Solaris 7/8 Subject: re: Interrupts with x86 BSP Subject: Re: BSP questions- int vector and remap Subject: sntp on vxworks 5.2 Subject: how to check whether fpp register contents is a number? Subject: Re: BSP questions- int vector and remap Subject: Where can I take a VxWorks class? Subject: Re: Some problem with C++ Subject: Where can one learn to use VxWorks Subject: Proxy ARP problems: workaround needed Subject: Re: equivalent of dn_expand in vxworks Subject: How do I set up VME address space? Subject: Re: Proxy ARP problems: workaround needed Subject: problem with inet_aton Subject: Filenames 8.3 Subject: Re: MVME5100, does the BSP support writing to Flash? Subject: Re: How do I set up VME address space? Subject: Re: reboot() problems Subject: Re: how to find the right component to link in? Subject: Re: Where can one learn to use VxWorks Subject: ARP cache Subject: Is there anybody experienced in WindNet PPP? Subject: internal compiler errors ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Intel 82559ER driver on Tornado 2.0.2? Date: Fri, 8 Jun 2001 12:28:05 +0200 From: "Harry Stox" Organization: Lawrence Berkeley National Laboratory Message-ID: <9fqc45$n6e$1@overload.lbl.gov> Hello, I'm interested to know if anyone managed to get the unsupported Intel Architecture driver for the Intel 82559ER on-board Ethernet chip (a559mux.exe from the wrs support site) up and running on a ServerWorks ServerSet III LE based mainboard. We are trying to build a kernel for the new FORCE Flexor IDP3S370 SBC, but when following the instructions supplied in the driver's readme file, VxWorks freezes after approximately one second after calling syseeVPciInit(), even when not automatically attaching the END device to the IP stack. We're using Tornado 2.0.2 (T2CP4), which has support for NPT drivers. I noticed that END drivers sometimes need to be slightly modified to get them to work reliably using a kernel with NPT support. I don't know if the Intel Architecture driver can be used in conjunction with a kernel having NPT support. Further details: - - PC_PENTIUMPRO kernel - - VxVMI included - - CPU: PIII/1GHz - - Memory: 256Mb - - no other (external) PCI devices present on backplane Thanks in advance for your reaction. Best regards, Harry Stox - Ellips B.V. - --- Harry Stox (harry@ellips.nl) -- Software Engineer -- Ellips B.V. Urkhovenseweg 11 -- 5641 KA -- Eindhoven -- The Netherlands phone: +31-40-2456540 -- fax: +31-40-2467183 -- http://www.ellips.nl --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Controling point at which socket write blocks Date: Fri, 8 Jun 2001 10:58:29 +0100 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: nick wrote in message ... >Hello, > >We want to change the point at which at socket blocks. Our system may >have in excess of 120 sockets simultaneously writing, and I would like >each socket to block after 400 bytes or so has been written without a >corresponding read. > >I have tried changing the send buffer, but this just slows things down >(I think it does not work because data is copied from the send buffer >to another vxworks / tcp buffer). It does not work because the other end still has room in its socket's read buffer. As long as there is room at the *receiver*, your transmitter will not block. If you cut down the send buffer to 400 bytes, it will fill up and then transmit those bytes, and the tcp stack at the far end will continue filling its receive buffer even if the application layer at the far end isn't reading out of that buffer. To acheive the effect you want, you would need to adjust the size of the receive buffer at the far end to 400 bytes. This is what controls the receiver's advertised window size, and it is that which controls when the transmitter blocks. >The other methods, such as recv with MSG_PEEK, and querying the >FIONREAD values only work at the other end of the socket (which is >done by another company / software, and we have no control of). You may not have control of the software, but do you have control of the OS environment that runs there? Can you set the OS default read buffer size to 400 bytes and hope that the remote software doesn't override the default with an ioctl/setsockopt ? DaveK - -- They laughed at Galileo. They laughed at Copernicus. They laughed at Columbus. But remember, they also laughed at Bozo the Clown. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to run VxWorks on an x86 machine? Date: Fri, 8 Jun 2001 11:04:20 +0100 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: <9fm9l3$7oqs5$1@reader01.wxs.nl> <3B1EE7B0.78703DF3@covad.net> DrDiags wrote in message <3B1EE7B0.78703DF3@covad.net>... > However, if you don't have the x86 flavor of Tornado, you might find >the mkboot utility in the unsupported section of your architecture (PPC >for example). The target/unsupported/src/dosfs2 has the mkboot.c file >and vxsys.com. You might be able to get an x86 BSP and build your own >x86 Toolchain from WRS GNU Toolkit CD, which should be open source and >from what I know, free but painful since the makefiles are not geared to >allow a straight off the CD build, but then again what else is new? Hiya Doc, That won't work. The toolchain is indeed free source, including WRS' modifications to it, but the BSP itself - the Vxworks libs and source - is WRS proprietary code, and they aren't about to start giving that away. DaveK - -- They laughed at Galileo. They laughed at Copernicus. They laughed at Columbus. But remember, they also laughed at Bozo the Clown. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Network Stack Date: Fri, 8 Jun 2001 11:06:47 +0100 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: <9fmoai$4jj$1@woodrow.ucdavis.edu> Jen-Chieh Tang wrote in message <9fmoai$4jj$1@woodrow.ucdavis.edu>... > I am rather new to vxWorks and was wondering what the > community uses to debug the network stack. I am pretty > familiar with the stack code as I have worked with the > BSD stack before, but how does one debug in vxWorks ? > Thanks for any info. >-- > Tom Tang Hi Tom, Generally I don't think many people do debug the network stack. It seems to just work pretty well. There's no reason you shouldn't try and step through it in the debugger, but without source code you might get lost quite easily! DaveK - -- They laughed at Galileo. They laughed at Copernicus. They laughed at Columbus. But remember, they also laughed at Bozo the Clown. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Redirecting stdout problems Date: Fri, 8 Jun 2001 11:10:16 +0100 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: <3B1F4C8F.E137BEF@wipro.com> >> Bill Harmon wrote: >> >> > I am having problems redirecting stdout. Kumar wrote in message ... >If you cannot do on a file open a mempory file. Actually I think it would be a better idea to use a memory file in any case; it's faster and less wear and tear on the hardware. Bill, see the memDrv documentation in "VxWorks Reference Manual : Libraries" for more info. DaveK - -- They laughed at Galileo. They laughed at Copernicus. They laughed at Columbus. But remember, they also laughed at Bozo the Clown. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Using GDB 5 (insight, cygwin) with VxWorks (WinNT) ? Date: Fri, 8 Jun 2001 11:14:32 +0100 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: <063U6.7$Hw6.9096@newsr2.u-net.net> References: <97e6274e.0106062314.443a6d88@posting.google.com> Heiko Elger wrote in message <97e6274e.0106062314.443a6d88@posting.google.com>... >Hello has anyone a port for using gdb 5 (insight, cygnus) with VxWorks 5.4 >for x86 or simNT targets? > >As I know the WDB protocol has to port to gdb 5. > >Perhaps someone has done this x86 or SimNt? > >best regards and thanks a lot I'd also be interested if anyone has tried this. I'm after getting hold of a source CD from WRS at the moment and if nobody else has done it, I'm going to try and port the target-comms part of the WRS gdb up to the v5 gdb release. If that isn't easy, I found some stuff at the VxHacks site that is meant to provide a remote server for GDB on a VxWorks target, and I might give that a go. DaveK - -- They laughed at Galileo. They laughed at Copernicus. They laughed at Columbus. But remember, they also laughed at Bozo the Clown. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: ¤å¾Ì°ª¯Å¤å¾Ì²¦·~¤å¾Ì¾Ç¤hºÓ¤h½Òµ{ Date: Fri, 8 Jun 2001 11:48:02 +0100 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: <763U6.8$Hw6.9059@newsr2.u-net.net> References: <9fo3dl$4qc8496@rain.i-cable.com> ¸ê°TºÞ²z±M·~¨ó·| wrote in message <9fo3dl$4qc8496@rain.i-cable.com>... >IT Management Association UK >University of Devonshire UK > >Any enquiry please contact >Sunny Cheung >Email: sunny@itma-uk.org >URL: www.itma-uk.org >Tel: 27108382 (Office hour) >Add: Rm. 2103 Hollywood Plaza, 610 NAthan Rd., Mongkok, Lowloon. I never knew there was a "Devonshire UK" in Hong Kong, what an amazing coincidence. DaveK - -- They laughed at Galileo. They laughed at Copernicus. They laughed at Columbus. But remember, they also laughed at Bozo the Clown. --------------------------- Newsgroups: comp.os.vxworks Subject: VxSim on Solaris 7/8 Date: 8 Jun 2001 07:00:37 -0700 From: ciuli_net@yahoo.com (Iulian Ciulinet) Organization: http://groups.google.com/ Message-ID: <9d61677d.0106080600.b10a87d@posting.google.com> Hi, Has anybody actually got VxSim (with the networking features) to work on Solaris 7 or 8, using either the ULIP or PPP network backend ? I have tried various things, including booting in 32-bit mode, to no success. PPP looks like it's working (at least checking with netstat/ifconfig), but the VxSim processes keep getting stuck inside sysHwInit, when doing the sysPPPOpen call. The shell keeps displaying 'sysToMonitor: restarting' in a loop. I cannot ping the virtual vxsimXXX adresses (127.0.1.XXX). The ULIP drivers don't initialize correctly. In 64-bit mode, the devices aren't even there (/dev/ulXXX). In 32-bit mode, they appear correctly, but the ulXXX interfaces don;t come up. When the ULIP service is trying to start, these errors come up on the console: starting ULIP services... ifconfig: DL_INFO_ACK timed out ifname: ul0 style: 1 ppa: 0 mod_cnt: 0 devname: /dev/ul0 ifconfig: plumb: ul0: No such file or directory ifconfig: SIOCGLIFNETMASK: ul0: no such interface Strangely enough, the VxSim process seems to attach to the ul0 interface. It comes up with Target Name: vxTarget Attaching network interface ul0... done. Attaching network interface lo0... done. NFS client support not included. VxWorks Copyright 1984-1998 Wind River Systems, Inc. CPU: SunOS 5.8 [sun4u] VxWorks: 5.4 BSP version: 1.2/0 Creation date: Jun 8 2001 WDB: Ready. But then again I cannot ping it. I'm open to any suggestions ... As you can see, I'm using VxWorks 5.4 (Tornado 2.0.1) Thanks for your time, - - Iulian --------------------------- Newsgroups: comp.os.vxworks Subject: re: Interrupts with x86 BSP Date: Fri, 8 Jun 2001 16:01:13 +0200 From: "Thomas Rahn" Organization: Lawrence Berkeley National Laboratory Message-ID: <9fqmlm$s8s$1@overload.lbl.gov> Hi David, there is a simple answer: you do not have to do take care of signaling EOI. Our int handlers typically look like void ataIntr(int ctrl) { sysInByte(0x1f7); // clears the irq condition semGive (syncSem); // continue waiting task } But there might be more (depends on your problem): Int handling on x86 is sligthly weird because of the two cascaded interrupt controllers. Ints must be acknowledged to the int controller by some I/O write. The EOI must be send to both int controllers if the int originates from the slave device (int 8..15). VxWorks wraps your interrupt handler with some code that should include a call to i8259IntEoiSlave() or similar. [Try to disassemble the address returned by intVecGet()] In VxWorks 5.3 the int controller software always ack'd both (no harm done). With VxWorks 5.4, the driver software changed. It was improved to spare the extra IO cycle if not required. However, the code got somehow mixed up and it might happen that no EOIs are sent at all. This happens if you use the combination intHandlerCreate/intVecSet instead of intConnect (never trust the docs, intConnect uses a private function intHandlerCreateX86...) This special problem can be solved, if you change the file sysLib.c back to the old style: find /* initialize the PIC (Programmable Interrupt Controller) */ sysIntInitPIC (); intEoiGet = sysIntEoiGet; /* function pointer used in intConnect () */ and replace with /* set the global function pointer to sysIntEOI() */ intEOI = sysIntEOI; Bye Thomas Rahn --------------------------- Newsgroups: comp.os.vxworks,comp.sys.arm Subject: Re: BSP questions- int vector and remap Date: 8 Jun 2001 07:30:14 -0700 From: franck.dagostini@ifrance.com (Franck D'agostini) Organization: http://groups.google.com/ Message-ID: <95f18de8.0106080630.74ca669f@posting.google.com> References: <9e9uqn$1u1$1@news.hitel.net> "wjsu" wrote in message news:... > Some fundamental questions continuing , would you please help? > > 1. Is there any sample code showing the RE-MAPPING procedure illustrating > what one should prepare for remapping? You could found a boot example with a remapping example in the document: "Implementation of an universal Boot Monitor for an ARM-based System" at this address: http://mikro.ee.tu-berlin.de/~hellmich/theses.html If you have others examples of remapping please send to me. > > 2. Also if ROM address starts from 0x0, what software approach can I do in > VxWorks to made RAM appear in 0x0 instead of ROM? > > 3. I saw that ARM allow exception vectors moved from 0x0-0x1c to > 0xffff0000-0xffff001c, since normally 0xffff0000 will not have actual RAM > there, what can I do in VxWorks to take advantage of this ARM feature? > > Thanks! > > WJ > > > Does hardware/board need to have > > "Àü¸¸¼ö" wrote in message > news:9e9uqn$1u1$1@news.hitel.net... > > 3. remap can change ram address to rom address. > > At first, Processor start at ROM. > > and Loading Progrm to RAM. > > and remap > > now RAM code Running. > > > > > > "wjsu" wrote in message > > news:tgaot4e8t4skb7@corp.supernews.com... > > > Hi, > > > > > > I have some basic and fundamental question for BSP development for the > > > following thing to ask you. > > > > > > For a clean system, clean board, with nothing ever programmed before. > > > > > > 1. Does it need to program the Flash/ROM for interrupt vectors and > > > exception pointers the first thing? > > > > > > 2. How does one know where those vectors should go at first place? > > > For example, I saw intergrator has the vector installed in address > > > 0x018 > > > 0x018 : ldr pc, 0x014; =#0x01360c > > > How did it know the address? > > > > > > 3. I learn from ARM manual that ARM has "remap" mechnism, what is > exactly > > > "remap" means and how can one achieve that? What's the benefit? Does it > mean > > > instead of programming Flash/ROM for interrupt vectors, one can do it > in > > > the RAM? Where to get more practicle implementation information? > > > > > > Thank you! > > > > > > WJ > > > > > > > > > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: sntp on vxworks 5.2 Date: Fri, 08 Jun 2001 10:48:37 -0400 From: "Beauvais, Severin [BVW:9P82:EXCH]" Organization: Nortel Networks Message-ID: <3B20E5C5.70DF65AF@americasm01.nt.com> Hello, Has anyone ported sntp to vxworks 5.2? I need a free implementation as a temporary measure (until upgrade to 5.4). All I want to do is take the system clock and broadcast the messages. Don't need to sync with a server. Thanks in advance! Severin Beauvais --------------------------- Newsgroups: comp.os.vxworks Subject: how to check whether fpp register contents is a number? Date: Fri, 08 Jun 2001 16:30:55 +0200 From: Wilco Rademaker Organization: Lawrence Berkeley National Laboratory Message-ID: <9fqq66$su$1@overload.lbl.gov> Hello, We are using vxworks 5.3.1 with PowerPC750 board. with fppTaskRegsShow we get the following output: fppTaskRegsShow 0x3b89d10 fpcsr = 70 fr0 = NaN fr1 = NaN fr2 = NaN fr3 = NaN fr4 = NaN fr5 = NaN fr6 = NaN fr7 = NaN fr8 = NaN fr9 = NaN fr10 = NaN fr11 = NaN fr12 = NaN fr13 = NaN fr14 = NaN fr15 = NaN fr16 = NaN fr17 = NaN fr18 = NaN fr19 = NaN fr20 = NaN fr21 = NaN fr22 = NaN fr23 = NaN fr24 = NaN fr25 = NaN fr26 = NaN fr27 = NaN fr28 = NaN fr29 = NaN fr30 = NaN fr31 = NaN We would like to be able to perform the test on the registers for NaN (Not a Number). We were unable to trackdown the functioncall to do this. On Solaris the function isnan() is available to perform this. Who can tell us which function we should use on vxworks or with what code we perform the isnan test? Kind regards, - -- Wilco Rademaker ASM Lithography phone: +31-(0)40-268.4311 PO Box 324 fax: +31-(0)40-268.4999 Room 7A2013 mailto:wilco.rademaker@asml.nl (work) 5500 AH VELDHOVEN mailto:wilco.rademaker@bigfoot.com (private) The Netherlands --------------------------- Newsgroups: comp.os.vxworks,comp.sys.arm Subject: Re: BSP questions- int vector and remap Date: Fri, 8 Jun 2001 18:32:11 +0200 From: "Ulf Samuelsson" Organization: A Customer of Tele2 Message-ID: References: <9e9uqn$1u1$1@news.hitel.net> <95f18de8.0106080630.74ca669f@posting.google.com> Remapping is CPU dependant. On the Atmel AT91 you would normally configure the External Bus Interface (EBI) with a string copy and then immediately do a jump to boot flash. At reset, the CS0 boot flash is mapped to address zero. At the end of the EBI is a bit which would do the remapping so the last write of the string copy will remap SRAM to address zero and flash to another location. The pipeline ensures that the following JUMP instruction is already fetched before the remapping takes effect. It is shown in the CD delivered with the Atmel AT91EBxx dev boards. - -- Best regards, ulf at atmel dot com The contents of this message is intended to be my private opinion and may or may not be shared by my employer Atmel Sweden "Franck D'agostini" skrev i meddelandet news:95f18de8.0106080630.74ca669f@posting.google.com... > "wjsu" wrote in message news:... > > Some fundamental questions continuing , would you please help? > > > > 1. Is there any sample code showing the RE-MAPPING procedure illustrating > > what one should prepare for remapping? > > > You could found a boot example with a remapping example in the document: > "Implementation of an universal Boot Monitor for an ARM-based System" > > at this address: > http://mikro.ee.tu-berlin.de/~hellmich/theses.html > > If you have others examples of remapping please send to me. > > > > > > > > > > > > > > > > > 2. Also if ROM address starts from 0x0, what software approach can I do in > > VxWorks to made RAM appear in 0x0 instead of ROM? > > > > 3. I saw that ARM allow exception vectors moved from 0x0-0x1c to > > 0xffff0000-0xffff001c, since normally 0xffff0000 will not have actual RAM > > there, what can I do in VxWorks to take advantage of this ARM feature? > > > > Thanks! > > > > WJ > > > > > > Does hardware/board need to have > > > > "Àü¸¸¼ö" wrote in message > > news:9e9uqn$1u1$1@news.hitel.net... > > > 3. remap can change ram address to rom address. > > > At first, Processor start at ROM. > > > and Loading Progrm to RAM. > > > and remap > > > now RAM code Running. > > > > > > > > > "wjsu" wrote in message > > > news:tgaot4e8t4skb7@corp.supernews.com... > > > > Hi, > > > > > > > > I have some basic and fundamental question for BSP development for the > > > > following thing to ask you. > > > > > > > > For a clean system, clean board, with nothing ever programmed before. > > > > > > > > 1. Does it need to program the Flash/ROM for interrupt vectors and > > > > exception pointers the first thing? > > > > > > > > 2. How does one know where those vectors should go at first place? > > > > For example, I saw intergrator has the vector installed in address > > > > 0x018 > > > > 0x018 : ldr pc, 0x014; =#0x01360c > > > > How did it know the address? > > > > > > > > 3. I learn from ARM manual that ARM has "remap" mechnism, what is > > exactly > > > > "remap" means and how can one achieve that? What's the benefit? Does it > > mean > > > > instead of programming Flash/ROM for interrupt vectors, one can do it > > in > > > > the RAM? Where to get more practicle implementation information? > > > > > > > > Thank you! > > > > > > > > WJ > > > > > > > > > > > > > > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Where can I take a VxWorks class? Date: Fri, 08 Jun 2001 10:39:43 -0700 From: russell cecala Organization: Sun Microsystems Message-ID: <3B210DDF.80D49B09@Sun.COM> Reply-To: Russell.Cecala@Sun.COM --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Some problem with C++ Date: Fri, 08 Jun 2001 10:38:10 -0700 From: russell cecala Organization: Sun Microsystems Message-ID: <3B210D82.B65073D3@Sun.COM> References: <3B1800D4.FAE5B1DC@hotmail.com> <9ffr3e$h7l$1@inn.jlab.org> <3B1BBD6E.FCE05642@hotmail.com> <9fglr8$hm1$1@inn.jlab.org> Reply-To: Russell.Cecala@Sun.COM Where can I take VxWorks class? --------------------------- Newsgroups: comp.os.vxworks Subject: Where can one learn to use VxWorks Date: Fri, 08 Jun 2001 10:37:20 -0700 From: russell cecala Organization: Sun Microsystems Message-ID: <3B210D50.CEA4CECD@Sun.COM> Reply-To: Russell.Cecala@Sun.COM Man there are a lot of jobs out there for VxWorks people ... where can I take a class? --------------------------- Newsgroups: comp.os.vxworks Subject: Proxy ARP problems: workaround needed Date: Fri, 08 Jun 2001 21:54:44 +0400 From: Pavel Kovalev Organization: Gamma NNTP server Moscow Russia Message-ID: <3B211164.3AA58342@auriga.ru> Hello All, We have a custom network driver for a CompactPCI backplane network and want to configure a proxy ARP server on the CompactPCI host board to make the backplane nodes visible on the Ethernet network (without additional routing). Tornado 2/VxWorks 5.4 support this configuration only for the shared memory network drivers supplied by WRS. Attempts to use proxyArpLib or configure all the necessary routing manually fail with various errors. I studied various materials available on WindSurf but haven't found a solution for this problem. There is an official SPR describing this problem (SPR# 28602), but it looks like it hasn't been processed/fixed yet. There were some discussions of this problem in the newsgroup in the past, but no good solutions have been proposed. I spent some time investigating this problem, and found that when I add a specific route to a node residing on the proxy network (route is added on the host which has to work proxy ARP server), mRouteAdd() creates the route with RTF_GATEWAY set. This prevents arpAdd() from adding a permanent published entry for the node. One solution is to develop a better version of mRouteAdd(), but I am not yet ready/willing to do so - there is no warranty that other hacks won't be needed. So, I am wondering if anybody has managed to solve this problem. Thanks, Pavel. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: equivalent of dn_expand in vxworks Date: Fri, 08 Jun 2001 19:05:47 GMT From: "Daniel Roy" Organization: bitHeads, inc. Message-ID: References: From the VxWorks resolvLib library... NAME resolvDNExpand( ) - expand a DNS compressed name from a DNS packet SYNOPSIS int resolvDNExpand ( const u_char * msg, /* ptr to the start of the DNS packet */ const u_char * eomorig, /* ptr to the last location +1 of the DNS packet */ const u_char * comp_dn, /* ptr to the compressed domain name */ u_char * exp_dn, /* ptr to where the expanded DN is output */ int length /* length of the buffer pointed by expd_dn */ ) DESCRIPTION This functions expands a compressed DNS name from a DNS packet. The msg parameter points to that start of the DNS packet. The eomorig parameter points to the last location of the DNS packet plus 1. The comp_dn parameter points to the compress domain name, and exp_dn parameter expects a pointer to a buffer. Upon function completion, this buffer contains the expanded domain name. Use the length parameter to pass in the size of the buffer referenced by the exp_dn parameter. RETURNS The length of the expanded domain name, or ERROR on failure. Daniel "Sarvesh" wrote in message news:c86bce3f.0106080204.3fd982ad@posting.google.com... > hi, > > can anyone tell me how to expand a compressed domain name in vxworks. > is there any equivalent function call in vxworks as dn_expand in unix machines. > > thanx in advance, > > regards, > sarvesh --------------------------- Newsgroups: comp.os.vxworks Subject: How do I set up VME address space? Date: Fri, 08 Jun 2001 13:58:37 -0500 From: James Toland Organization: Raytheon Company Message-ID: <3B21205D.E06762F5@raytheon.com> How do I set up the MVME2700 SBC for accessing VME bus addresses 0x8000 0000 to 0xBFFF FFFF? The BSP documentation says, "In order to modify the Extended VME mapping configuration, make the necessary changes in config.h and, possibly, sysLib.c," & " In config.h, #define the VME window variables." In the config.h file I might change #define VME_A32_MSTR_BUS 0x08000000 #define VME_A32_MSTR_SIZE 0x08000000 /* (128MB) */ to #define VME_A32_MSTR_BUS 0x08000000 #define VME_A32_MSTR_SIZE 0x40000000 /* (128MB) */ ?? What else should I change in config.h and sysLib.c? Thanks --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Proxy ARP problems: workaround needed Date: Fri, 08 Jun 2001 19:28:03 GMT From: "Michael Yeung" Organization: Magma Communications Ltd. Message-ID: <7H9U6.494$j6.27775@news> References: <3B211164.3AA58342@auriga.ru> Reply-To: "Michael Yeung" Pavel, I actually ask the WindRiver support guys about this and here is what they came back with. I haven't had the time to try it out yet, but here you go. The proxy ARP is tied to shared memory in two location. in ..\target\src\config\usrNetwork.c or in ..\target\config\comps\src\net\usrNetSmSecBoot.c if you use the Project Facility to You can see that this is where they presume that you will be using the shared memory interface, but if you call proxyArpLibInit() yourself, and proxyNetCreate() with your own set of interfaces instead of params.bad you should have no problem as far as I can see. As well, don't select proxy ARP server component when building your BSP. The linker should be capable of pulling in the right file (ArpLib.o). Let me know how it went. Michael Pavel Kovalev wrote in message news:3B211164.3AA58342@auriga.ru... > Hello All, > > We have a custom network driver for a CompactPCI backplane network > and want to configure a proxy ARP server on the CompactPCI host > board to make the backplane nodes visible on the Ethernet network > (without additional routing). > > Tornado 2/VxWorks 5.4 support this configuration only for > the shared memory network drivers supplied by WRS. Attempts > to use proxyArpLib or configure all the necessary routing > manually fail with various errors. > > I studied various materials available on WindSurf but haven't > found a solution for this problem. There is an official SPR > describing this problem (SPR# 28602), but it looks like it > hasn't been processed/fixed yet. There were some discussions > of this problem in the newsgroup in the past, but no good > solutions have been proposed. > > I spent some time investigating this problem, and found that when > I add a specific route to a node residing on the proxy network > (route is added on the host which has to work proxy ARP server), > mRouteAdd() creates the route with RTF_GATEWAY set. This prevents > arpAdd() from adding a permanent published entry for the node. > One solution is to develop a better version of mRouteAdd(), but > I am not yet ready/willing to do so - there is no warranty that > other hacks won't be needed. > > So, I am wondering if anybody has managed to solve this problem. > > Thanks, > Pavel. --------------------------- Newsgroups: comp.os.vxworks Subject: problem with inet_aton Date: Fri, 08 Jun 2001 19:38:33 GMT From: "Michael Yeung" Organization: Magma Communications Ltd. Message-ID: Reply-To: "Michael Yeung" Hi all, I'm not sure why but the inet_aton seems to always returns 0 and conversion is not done properly. struct in_addr temp; if(inet_aton("1.1.1.1", (struct in_addr *)&temp) == 0) { printf("Failed conversion\n"); } I noticed VxWorks does not support "inet_addr" or "inet_pton", as suggestion is appreciated. As well, does anyone knows the format expected for mask to function like "rtrequest"? thanx in advance, Michael --------------------------- Newsgroups: comp.os.vxworks Subject: Filenames 8.3 Date: Fri, 8 Jun 2001 20:33:24 GMT From: "Gary G." Organization: The Boeing Company Message-ID: <3B213694.307B5FC3@boeing.com> Sender: nntp@news.boeing.com (Boeing NNTP News Access) Is it true that the filename/directory paths must be 8.3 ? I am trying to create my projects and workspace on a server that has a path that is NOT 8.3 and it fails. If I use C:\<8.3>\<8.3> everything works. Using the "add files" menu will let me "drive down" to the server directory that contains the files I want to add to my project, but again, if the path is not 8.3, it fails to add the file. Does Tornado really only support 8.3 file formats, or am I doing something wrong ? (using NT 4.0, Tornado 2.0, BSP for a PowerPC) Is there a workaround / patch ? THANKS Gary --------------------------- Newsgroups: comp.os.vxworks Subject: Re: MVME5100, does the BSP support writing to Flash? Date: Fri, 08 Jun 2001 21:14:18 GMT From: William Dennen Organization: Motorola Computer Group Message-ID: <3B2140B4.4000504@motorola.com> References: <3B1FE378.A663AC4@sierradatasys.com> Reply-To: william.dennen@motorola.com Martijn Kamerbeek wrote: > Hi, > > I am currently working with a MVME1604 board and I see that the BSP has > routines such as sysFlashSet and SysFlashWrite. > > Soon I'll be working with a MVME5100 and I would like to read and write > to > Flash memory. However, I can't find any indication anywhere that the > 5100 BSP > will support writing to and reading from Flash. > > Does anybody have some information for me? > > Thank you very much!! > > Martijn Kamerbeek The Early Access BSP from MCG does not; when released by Wind River it may, but probably not ... Regards - -- >@< Bill Dennen william.dennen@motorola.com Motorola Computer Group --GTAC --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How do I set up VME address space? Date: Fri, 08 Jun 2001 16:46:17 -0500 From: Andrew Johnson Organization: APS, Argonne National Laboratory Message-ID: <3B2147A9.33C10B26@aps.anl.gov> References: <3B21205D.E06762F5@raytheon.com> James Toland wrote: > > How do I set up the MVME2700 SBC for accessing VME bus addresses 0x8000 > 0000 to 0xBFFF FFFF? > The BSP documentation says, "In order to modify the Extended VME mapping > configuration, make the necessary changes in config.h and, possibly, > sysLib.c," & " In config.h, #define the VME window variables." > > In the config.h file I might change > #define VME_A32_MSTR_BUS 0x08000000 > #define VME_A32_MSTR_SIZE 0x08000000 /* (128MB) */ > to > #define VME_A32_MSTR_BUS 0x08000000 > #define VME_A32_MSTR_SIZE 0x40000000 /* (128MB) */ ?? Which do you want, 0x80000000-0xbfffffff or 0x08000000-0x0bffffff - your #defines will give you the latter. If you want the former which is what you wrote in your question above they should be #define VME_A32_MSTR_BUS 0x80000000 /* Base VME address */ #define VME_A32_MSTR_SIZE 0x40000000 > What else should I change in config.h and sysLib.c? The D BAT entries in sysBatDesc[], found in sysLib.c. You don't want to ask the system to set up MMU page tables for 1GB of address space (that would use up 1MB of RAM just for the page tables), so delete the VME_A32_MSTR_LOCAL entry in sysPhysMemDesc[] and change the second half of your sysBatDesc[] table to look like this: /* D BAT 0 */ ((VME_A32_MSTR_BAT0 & _MMU_UBAT_BEPI_MASK) | _MMU_UBAT_BL_256M | _MMU_UBAT_VS | _MMU_UBAT_VP), ((VME_A32_MSTR_BAT0 & _MMU_LBAT_BRPN_MASK) | _MMU_LBAT_PP_RW | _MMU_LBAT_CACHE_INHIBIT), /* D BAT 1 */ ((VME_A32_MSTR_BAT1 & _MMU_UBAT_BEPI_MASK) | _MMU_UBAT_BL_256M | _MMU_UBAT_VS | _MMU_UBAT_VP), ((VME_A32_MSTR_BAT1 & _MMU_LBAT_BRPN_MASK) | _MMU_LBAT_PP_RW | _MMU_LBAT_CACHE_INHIBIT), /* D BAT 2 */ ((VME_A32_MSTR_BAT2 & _MMU_UBAT_BEPI_MASK) | _MMU_UBAT_BL_256M | _MMU_UBAT_VS | _MMU_UBAT_VP), ((VME_A32_MSTR_BAT2 & _MMU_LBAT_BRPN_MASK) | _MMU_LBAT_PP_RW | _MMU_LBAT_CACHE_INHIBIT), /* D BAT 3 */ ((VME_A32_MSTR_BAT3 & _MMU_UBAT_BEPI_MASK) | _MMU_UBAT_BL_256M | _MMU_UBAT_VS | _MMU_UBAT_VP), ((VME_A32_MSTR_BAT3 & _MMU_LBAT_BRPN_MASK) | _MMU_LBAT_PP_RW | _MMU_LBAT_CACHE_INHIBIT) Those VME_A32_MSTR_BATn constants are defined in config.h like this: #define VME_A32_MSTR_BAT0 0x80000000 /* BATs are 256MB each */ #define VME_A32_MSTR_BAT1 0x90000000 #define VME_A32_MSTR_BAT2 0xa0000000 #define VME_A32_MSTR_BAT3 0xb0000000 This assumes that you're not going to be running code from that address range. If you are, you'll need to replicate those settings for the four I BAT entries as well, and you'll probably want to enable cacheing so the flags will need to be a bit different for both sets of entries. Oh, and do make sure you keep a copy of the originals whenever you're fiddling with BSP files like this, just in case! - - Andrew - -- The world is such a cheerful place when viewed from upside-down It makes a rise of every fall, a smile of every frown --------------------------- Newsgroups: comp.os.vxworks Subject: Re: reboot() problems Date: 8 Jun 2001 15:21:03 -0700 From: hariharan@foobox.com (Thiagarajan Hariharan) Organization: http://groups.google.com/ Message-ID: <7f10f842.0106081421.5c3572a9@posting.google.com> References: <7f10f842.0106061228.5c4a35c8@posting.google.com> We were previously using a bootrom built with dosFs1. Building a bootrom with dosFs2 has made the problem go away (at least for now). Booting from the ATA disk seems to work every time now. An observation: If a task is suspended when you reboot from the disk, the reboot does not work - just hangs. May be the reboot hooks are hung somewhere. Hari --------------------------- Newsgroups: comp.os.vxworks Subject: Re: how to find the right component to link in? Date: Fri, 08 Jun 2001 18:09:16 -0700 From: "Michael R. Kesti" Organization: MK Associates Message-ID: <3B21773C.219B8236@gv.net> References: Reply-To: mkesti@gv.net Gang Wang wrote: >The question is regarding to how to find out what vxWorks component is >needed when you build your image. > >When this is a forward question, it is easy. For example, if I know that I >will use select to do file I/O, I could just go to the component tree and >link the Operating System Components/IO System components/select module into >my image. > >However, things get hard ( at least to me ) when the question is reversed, >i.e., I know a specific function I have to use then how I could find out >which component(s) contains it? For example, can someone tell me which >component I should use for function rmdir() and mkdir()? Or in other words, >where is the usrLib? Use the GNU "nm" binary utility to search the archive files in the ..Tornado\target\lib directory. RTFM: The last section of the GNU Toolkit User's Guide. - -- ======================================================================== Michael Kesti | "And like, one and one don't make | two, one and one make one." mkesti@gv.net | - The Who, Bargain --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Where can one learn to use VxWorks Date: Fri, 08 Jun 2001 18:19:04 -0700 From: "Michael R. Kesti" Organization: MK Associates Message-ID: <3B217988.EA4F128A@gv.net> References: <3B210D50.CEA4CECD@Sun.COM> Reply-To: mkesti@gv.net russell cecala wrote: >Man there are a lot of jobs out there for VxWorks >people ... where can I take a class? Wind River offers a variety of classroom training opportunities, but they're not cheap. Check it out at http://www.wrs.com - -- ======================================================================== 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: ARP cache Date: Sat, 09 Jun 2001 04:08:30 GMT From: "Michael Yeung" Organization: Magma Communications Ltd. Message-ID: <2jhU6.643$j6.34686@news> Reply-To: "Michael Yeung" Hi all, I'm looking for something in VxWorks that allows me to retrieve all entries in the ARP cache. Something similar to what "sysctl" provides. I found an API called "ip_sysctl" in netinet\ip_var.h but experience problem when loading the object file to the target. Please let me know if you know of anything that allows me to walk the ARP table. thanx, Michael --------------------------- Newsgroups: comp.os.vxworks Subject: Is there anybody experienced in WindNet PPP? Date: Sat, 9 Jun 2001 13:38:58 +0900 From: "seo park" Organization: Korea Telecom Message-ID: <9fs8n0$qq3$1@news1.kornet.net> Hello, vxWorks Experts. I want WindNet PPP to operate as PPP Server. PPP Client connects to PPP Server(My target Board) and PPP Server assigns IP Address to PPP Client. How do I set WinNet PPP IPCP Componenet or others? Thanks your answer. --------------------------- Newsgroups: comp.os.vxworks Subject: internal compiler errors Date: Sat, 09 Jun 2001 05:09:27 GMT From: jagan Organization: Road Runner Message-ID: <3B207CA7.D938FC01@mediaone.net> I am seeing an internal compiler error when compiling using Tornado 2 cross compiler for a ppc target in a cygwin environment on NT. I see cc1.exe coring with signal 5. Does anyone know why this may be happening? thanks for the help. --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Sun Jun 10 04:03:37 2001 From: Vxworks Exploder Date: Sun Jun 10 04:03:39 PDT 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sun Jun 10 04:03:16 PDT 2001 Subject: Processing of Raw TCP/IP Packets Subject: Re: Where can one learn to use VxWorks Subject: Re: Filenames 8.3 Subject: Priority inversion in semaphores Subject: Re: Proxy ARP problems: workaround needed Subject: Re: VxWorks Tools questions Subject: Re: vxsim memory size Subject: Re: VxWorks Tools questions Subject: New Altivec-enhanced VxWorks Gcc 2.95.4 compiler released! Subject: Re: Network Stack Subject: Re: Proxy ARP problems: workaround needed Subject: Re: How to run VxWorks on an x86 machine? Subject: Re: reboot() problems Subject: Proper way to unmount dosFs Subject: SM anchor error codes Subject: PPC: system tick routine stack usage Subject: getproc_got_string() and structure Subject: Re: getproc_got_string() and structure ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Processing of Raw TCP/IP Packets Date: 9 Jun 2001 07:33:38 -0700 From: ivan@blueiguana.com (Ivan Bach) Organization: http://groups.google.com/ Message-ID: <628eca30.0106090633.251aa368@posting.google.com> I was wondering if any of you knew of some code that receives, processes, and/or sends raw TCP/IP packets under VxWorks. I am aware that etherhooks are getting obsolete. There is a CAUTION note at the end of the description of etherhooks in Chapter 14 in the version of VxWorks Network Programmer's Guide that is on Windsurfer. That document says that etherhooks will not be supported in the future releases of VxWorks. Only some 5.4 Edition 2 versions of the VxWorks Network Programmer's Guide even have Chapters 13 and 14. My printed copy of that manual (also 5.4 Edition 2) no longer has those chapters. What have you found about MUX and etherhook replacements? Is the new Section 3.2.1 BPF, the BSD Packet Filter in the printed version of the VxWorks Network Programmer's Guide 5.4, Edition 2, an attempt to replace them? That sounds as replacement for only one half of the function of etherhooks, the receiving of packets. So does ipFilterLib. What happened to the generation or sending of packets? I have not yet seen vxsniff.c program, but it was apparently based on etherhooks. Does anybody actually have a copy of that program, even if it is not working? The pointers that I saw for that program don't lead anywhere. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Where can one learn to use VxWorks Date: 9 Jun 2001 16:43:10 GMT From: borkhuis@agere.com (Johan Borkhuis) Organization: Agere System Message-ID: References: <3B210D50.CEA4CECD@Sun.COM> russell cecala wrote: > Man there are a lot of jobs out there for VxWorks > people ... where can I take a class? You can try to learn it yourself, using the Tornado Prototyper, a free VxWorks simulator. Groeten, Johan - -- o o o o o o o . . . _____________________________ o _____ || Johan Borkhuis | .][__n_n_|DD[ ====_____ | borkhuis@agere.com | >(________|__|_[_________]_|__________________________| _/oo OOOOO oo` ooo ooo 'o!o!o o!o!o` === VxWorks FAQ: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html === --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Filenames 8.3 Date: 9 Jun 2001 16:48:15 GMT From: borkhuis@agere.com (Johan Borkhuis) Organization: Agere System Message-ID: References: <3B213694.307B5FC3@boeing.com> "Gary G." wrote: > Is it true that the filename/directory paths must be 8.3 ? I am trying > to create my projects and workspace on a server that has a path that is > NOT 8.3 and it fails. If I use C:\<8.3>\<8.3> everything works. I never had any problems with this. The only thing you should be carefull with is using spaces in names. Do you create it on a remote server? Do any other programs have problems with this? Groeten, Johan - -- o o o o o o o . . . _____________________________ o _____ || Johan Borkhuis | .][__n_n_|DD[ ====_____ | borkhuis@agere.com | >(________|__|_[_________]_|__________________________| _/oo OOOOO oo` ooo ooo 'o!o!o o!o!o` === VxWorks FAQ: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html === --------------------------- Newsgroups: comp.realtime,comp.os.vxworks,comp.os.psos,comp.os.qnx,comp.os.xinu Subject: Priority inversion in semaphores Date: 9 Jun 2001 10:54:29 -0700 From: amitchoudhary@my-deja.com (amit choudhary) Organization: http://groups.google.com/ Message-ID: <4a6d848e.0106090954.514a3912@posting.google.com> To handle the priority inversion problem in semaphores, can it be segregated into two types: 1. To provide mutually exclusive access to critical section or a resource, and 2. For producer-consumer kind of synchronization, Or a single abstraction can be used to handle priority inversion completely. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Proxy ARP problems: workaround needed Date: Sat, 09 Jun 2001 19:33:00 +0400 From: Pavel Kovalev Organization: Gamma NNTP server Moscow Russia Message-ID: <3B2241AC.72CB81AA@auriga.ru> References: <3B211164.3AA58342@auriga.ru> <7H9U6.494$j6.27775@news> Michael, > I actually ask the WindRiver support guys about this and here is what > they came back with. I haven't had the time to try it out yet, but here you > go. > > The proxy ARP is tied to shared memory in two location. > > in ..\target\src\config\usrNetwork.c or in > ..\target\config\comps\src\net\usrNetSmSecBoot.c if you use the Project > Facility to > > You can see that this is where they presume that you will be using the > shared memory interface, but if you call > proxyArpLibInit() > yourself, and > proxyNetCreate() > with your own set of interfaces instead of params.bad you should have no > problem as far as I can see. > > As well, don't select proxy ARP server component when building your BSP. > The linker should be capable of pulling in the right file (ArpLib.o). > > Let me know how it went. It is true that in the source code that is shipped to customers the proxy ARP is tied to the shared memory component in two location. Though, proxyNetCreate() checks whether the name of the network interface is equal to "sm" or "bp". While this problem can be worked around, this leads to nothing. proxyNetCreate() does not do very much useful by itself. The meat of proxy ARP server configuration seems to lie in proxyClientAdd(). Though, I haven't figured out how I can call this routine. Looks like shared memory clients register in the proxy ARP server using proxyReg(), and proxyClientAdd() is called from the registration packet processing function. I do not know how this functionality can be emulated in custom network drivers. I am afraid that for such drivers a new facility for registering proxy ARP clients should be created. I will continue some experiments and will post my results later. Still, I am hoping that somebody may have already passed all this way and found the solution... Regards, Pavel. > Michael > > Pavel Kovalev wrote in message > news:3B211164.3AA58342@auriga.ru... > > Hello All, > > > > We have a custom network driver for a CompactPCI backplane network > > and want to configure a proxy ARP server on the CompactPCI host > > board to make the backplane nodes visible on the Ethernet network > > (without additional routing). > > > > Tornado 2/VxWorks 5.4 support this configuration only for > > the shared memory network drivers supplied by WRS. Attempts > > to use proxyArpLib or configure all the necessary routing > > manually fail with various errors. > > > > I studied various materials available on WindSurf but haven't > > found a solution for this problem. There is an official SPR > > describing this problem (SPR# 28602), but it looks like it > > hasn't been processed/fixed yet. There were some discussions > > of this problem in the newsgroup in the past, but no good > > solutions have been proposed. > > > > I spent some time investigating this problem, and found that when > > I add a specific route to a node residing on the proxy network > > (route is added on the host which has to work proxy ARP server), > > mRouteAdd() creates the route with RTF_GATEWAY set. This prevents > > arpAdd() from adding a permanent published entry for the node. > > One solution is to develop a better version of mRouteAdd(), but > > I am not yet ready/willing to do so - there is no warranty that > > other hacks won't be needed. > > > > So, I am wondering if anybody has managed to solve this problem. > > > > Thanks, > > Pavel. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks Tools questions Date: Sat, 09 Jun 2001 13:34:49 -0700 From: Lori Fraleigh Organization: Real-Time Innovations Message-ID: <3B228869.1E23C01@rti.com> References: <3b1ee72a.4738173@news.qwest.net> Joseph Woodbury wrote: > > My company is on the verge of adopting VxWorks for some projects. > However, I am totally confused by several of the Tornado tools Wind > River offers. Part of this is due to Wind River buying up companies > right and left, even those with competing products, which they then > offer without explanation. > > What is the difference between CodeTEST and PerformancePAK > (ProfileScope & MemScope)? ProfileScope and MemScope require no code instrumentation or recompliation in order to statistically profile your system and track down memory leaks. They are able to analyze your code, the OS, and third party libraries with very little overhead allowing you to run your real-time system at full speed. The CodeTEST tools require you to instrument your code to use them and therefore do not allow you to analyze any code for which you do not have the source. They also introduce a much higher overhead. > Are there significant differences between the [default] GNU compiler > and Diab C++? (We will be using C++ on a PowerQUICC 860) > > How about between the default debugger and SingleStep? > > (Note, for the above mentioned pSOS, we were using Diab C++ and > SingleStep, though I was involved only for a few weeks of overlap when > the subsidiary formerly doing the development was shut down and I took > over.) > > Any objective (i.e. non-Wind River) analysis would be very welcome. - -- Lori Fraleigh Director of Engineering Real-Time Innovations, Inc. http://www.rti.com Phone: 408-734-4200 x121 Fax: 408-734-5009 --------------------------- Newsgroups: comp.os.vxworks Subject: Re: vxsim memory size Date: Sat, 09 Jun 2001 12:02:06 -0400 From: Andrew McGregor Organization: Bell Sympatico Message-ID: <3B22487E.16A6F31E@yahoo.ca> References: <3B1E6D81.8A76BB77@digidata.com> In NT we make a file called vxWorks.bat and in it was: start s:\np_src\proj\si\si_simHpuImage\default\vxWorks.exe %1 %2 /r10000000 exit I think that %1 and %2 were the card number arguments. We then set Tornado to start the batch file in Tools->Simulator (In Unix is is prob -r) Doug Dotson wrote: > Does anybody know how to increase the memory size of vxsim? > > Thanks, > Doug --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks Tools questions Date: Sat, 09 Jun 2001 12:05:21 -0400 From: Andrew McGregor Organization: Bell Nexxia Message-ID: <3B224941.E375FA8C@yahoo.ca> References: <3b1ee72a.4738173@news.qwest.net> We did do some testing with Diag and GNU on PPC. Diab was better for some pieces of code, but it was only 10-15%, I think. For the gain we couldn't be bothered with the cost and hassle. Personally, I've always been happy with the GNU story for 2 major products over 10 years. For the rest, I don't know... Joseph Woodbury wrote: > My company is on the verge of adopting VxWorks for some projects. > However, I am totally confused by several of the Tornado tools Wind > River offers. Part of this is due to Wind River buying up companies > right and left, even those with competing products, which they then > offer without explanation. > > What is the difference between CodeTEST and PerformancePAK > (ProfileScope & MemScope)? > > Are there significant differences between the [default] GNU compiler > and Diab C++? (We will be using C++ on a PowerQUICC 860) > > How about between the default debugger and SingleStep? > > (Note, for the above mentioned pSOS, we were using Diab C++ and > SingleStep, though I was involved only for a few weeks of overlap when > the subsidiary formerly doing the development was shut down and I took > over.) > > Any objective (i.e. non-Wind River) analysis would be very welcome. --------------------------- Newsgroups: comp.os.vxworks Subject: New Altivec-enhanced VxWorks Gcc 2.95.4 compiler released! Date: Sat, 9 Jun 2001 18:38:15 +0100 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: <5btU6.24100$YB3.4024200@news2-win.server.ntlworld.com> Hi everybody, I've just finished uploading a new compiler on my website that allows you to write Altivec programs for VxWorks. It seems WRS have one on the way as well but if you want something in the meantime or don't have a maintenance/update agreement with WRS, you might find it helpful. http://www.newgcc4vxworks4ppc.cjb.net/ DaveK --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Network Stack Date: Sun, 10 Jun 2001 00:06:36 +0400 From: Pavel Kovalev Organization: Gamma NNTP server Moscow Russia Message-ID: <3B2281CC.684D5A88@auriga.ru> References: <9fmoai$4jj$1@woodrow.ucdavis.edu> Hi Dave, Dave Korn wrote: > > Jen-Chieh Tang wrote in message <9fmoai$4jj$1@woodrow.ucdavis.edu>... > > > I am rather new to vxWorks and was wondering what the > > community uses to debug the network stack. I am pretty > > familiar with the stack code as I have worked with the > > BSD stack before, but how does one debug in vxWorks ? > > Thanks for any info. > >-- > > Tom Tang > > Hi Tom, > > Generally I don't think many people do debug the network stack. It seems > to just work pretty well. There's no reason you shouldn't try and step > through it in the debugger, but without source code you might get lost quite > easily! The point is that most people do not have the source code of the VxWorks network stack and associated utilities. Of course, there were not many changes as compared to the original BSD 4.4 code, but still there are components that are difficult to understand and debug. One example is the VxWorks proxy ARP server code. It is difficult to understand how the code does its job without having its sources. Though, one may need to do so. Now I am in the position when I need to have proxy ARP working for a non shared memory network interface, and I face several problems with the VxWorks implementation of the proxy ARP server (see my question posted on June 8 regarding this subject, as well as subsequent replies). While I am looking for some help from fellow VxWorkers who may know how to solve my specific problem, I would also be grateful for a general advice on debugging network services or something like that. Thanks, Pavel. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Proxy ARP problems: workaround needed Date: Sun, 10 Jun 2001 00:25:40 +0400 From: Pavel Kovalev Organization: Gamma NNTP server Moscow Russia Message-ID: <3B228644.622D75E6@auriga.ru> References: <3B211164.3AA58342@auriga.ru> <7H9U6.494$j6.27775@news> <3B2241AC.72CB81AA@auriga.ru> Michael et al, Here is some additional news regarding my investigation of the proxy ARP configuration: - - It is possible to configure all routes and the ARP cache properly manually from the target shell. Some ugly tricks are needed though, namely: o The main network interface should be disabled while configuring host specific routes for the proxy interface. The IP address of the main network interface should be temporarily set to 127.127.127.127 while doing so (this is described somewhere on WindSurf). o mRouteAdd should be used to configure routes to the proxy clients. Though, mRouteAdd erroneously sets the RTF_GATEWAY flag, which prevents arpAdd() from adding entries (permanent and published) referring to the proxy ARP clients. I had to set a breakpoint to rtrequest() and modify the flags using the mRegs command (cleared RTF_GATEWAY) to have arpAdd() work. - - Configuring routes and the ARP cache is not sufficient for the proxy ARP server to work. From the proxy ARP server I could ping nodes located both on the main network and on the proxy ARP network, but I could not ping nodes located on the main network from a node located on the proxy ARP network. Does anybody know how the proxy ARP server should be configured using the routing socket? I no longer hope that this can be done using the network functions exported by the VxWorks TCP/IP stack and associated libraries. Thanks, Pavel. Pavel Kovalev wrote: > > Michael, > > > I actually ask the WindRiver support guys about this and here is what > > they came back with. I haven't had the time to try it out yet, but here you > > go. > > > > The proxy ARP is tied to shared memory in two location. > > > > in ..\target\src\config\usrNetwork.c or in > > ..\target\config\comps\src\net\usrNetSmSecBoot.c if you use the Project > > Facility to > > > > You can see that this is where they presume that you will be using the > > shared memory interface, but if you call > > proxyArpLibInit() > > yourself, and > > proxyNetCreate() > > with your own set of interfaces instead of params.bad you should have no > > problem as far as I can see. > > > > As well, don't select proxy ARP server component when building your BSP. > > The linker should be capable of pulling in the right file (ArpLib.o). > > > > Let me know how it went. > > It is true that in the source code that is shipped to customers > the proxy ARP is tied to the shared memory component in two location. > > Though, proxyNetCreate() checks whether the name of the network > interface is equal to "sm" or "bp". While this problem can be > worked around, this leads to nothing. proxyNetCreate() does > not do very much useful by itself. The meat of proxy ARP server > configuration seems to lie in proxyClientAdd(). Though, I haven't > figured out how I can call this routine. Looks like shared memory > clients register in the proxy ARP server using proxyReg(), and > proxyClientAdd() is called from the registration packet processing > function. I do not know how this functionality can be emulated > in custom network drivers. I am afraid that for such drivers > a new facility for registering proxy ARP clients should be created. > > I will continue some experiments and will post my results later. > Still, I am hoping that somebody may have already passed all this > way and found the solution... > > Regards, > Pavel. > > > Michael > > > > Pavel Kovalev wrote in message > > news:3B211164.3AA58342@auriga.ru... > > > Hello All, > > > > > > We have a custom network driver for a CompactPCI backplane network > > > and want to configure a proxy ARP server on the CompactPCI host > > > board to make the backplane nodes visible on the Ethernet network > > > (without additional routing). > > > > > > Tornado 2/VxWorks 5.4 support this configuration only for > > > the shared memory network drivers supplied by WRS. Attempts > > > to use proxyArpLib or configure all the necessary routing > > > manually fail with various errors. > > > > > > I studied various materials available on WindSurf but haven't > > > found a solution for this problem. There is an official SPR > > > describing this problem (SPR# 28602), but it looks like it > > > hasn't been processed/fixed yet. There were some discussions > > > of this problem in the newsgroup in the past, but no good > > > solutions have been proposed. > > > > > > I spent some time investigating this problem, and found that when > > > I add a specific route to a node residing on the proxy network > > > (route is added on the host which has to work proxy ARP server), > > > mRouteAdd() creates the route with RTF_GATEWAY set. This prevents > > > arpAdd() from adding a permanent published entry for the node. > > > One solution is to develop a better version of mRouteAdd(), but > > > I am not yet ready/willing to do so - there is no warranty that > > > other hacks won't be needed. > > > > > > So, I am wondering if anybody has managed to solve this problem. > > > > > > Thanks, > > > Pavel. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to run VxWorks on an x86 machine? Date: Sat, 09 Jun 2001 17:16:38 -0700 From: DrDiags Organization: laserlink.net Message-ID: <3B22BC66.FB7D4AE1@covad.net> References: <9fm9l3$7oqs5$1@reader01.wxs.nl> <3B1EE7B0.78703DF3@covad.net> Reply-To: drdiags@covad.net Dave, Dave Korn wrote: > > > Hiya Doc, > > That won't work. The toolchain is indeed free source, including WRS' > modifications to it, but the BSP itself - the Vxworks libs and source - is > WRS proprietary code, and they aren't about to start giving that away. > > DaveK > -- > They laughed at Galileo. They laughed at Copernicus. They laughed at > Columbus. But remember, they also laughed at Bozo the Clown. Yeah, I was irresponsible for even considering and proposing the "if you don't have the binaries" scenario. Plus, my story had holes to boot. "Damn it Jim! ......" (:^) --------------------------- Newsgroups: comp.os.vxworks Subject: Re: reboot() problems Date: 9 Jun 2001 17:20:50 -0700 From: hariharan@foobox.com (Thiagarajan Hariharan) Organization: http://groups.google.com/ Message-ID: <7f10f842.0106091620.2f46e109@posting.google.com> References: <7f10f842.0106061228.5c4a35c8@posting.google.com> <7f10f842.0106081421.5c3572a9@posting.google.com> Turned out that one of the PCI cards in our box causes the reboot hang. If I remove that card, the warm reboot works just fine. Hari --------------------------- Newsgroups: comp.os.vxworks Subject: Proper way to unmount dosFs Date: 9 Jun 2001 17:24:53 -0700 From: hariharan@foobox.com (Thiagarajan Hariharan) Organization: http://groups.google.com/ Message-ID: <7f10f842.0106091624.70f230be@posting.google.com> Hello, To unmount dosFs that was mounted on an ATA disk at "/ata", I'm doing: char* pTail; iosDevDelete(iosDevFind("/ata", &pTail)); Is that all that's needed for the unmount? Thanks, Hari --------------------------- Newsgroups: comp.os.vxworks Subject: SM anchor error codes Date: 10 Jun 2001 00:49:34 GMT From: bte@tog.arl.psu.edu (Bruce Einfalt) Organization: Applied Research Laboratory Message-ID: <9fug6u$1pkg@r02n01.cac.psu.edu> Reply-To: bte@tog.arl.psu.edu Hi, I am trying to boot off the backplane and get SM anchor errors such as "r", "B", ">", etc.. Could someone please tell where I can find the list of error codes associated with the SM anchor? Thanks. Regard, Bruce Einfalt - -- Bruce Einfalt ARL - Penn State 814-863-4142 beinfalt@psu.edu --------------------------- Newsgroups: comp.os.vxworks Subject: PPC: system tick routine stack usage Date: Sun, 10 Jun 2001 03:48:32 GMT From: pkockritz@home.com (Pete Kockritz) Organization: Excite@Home - The Leader in Broadband http://home.com/faster Message-ID: The basic question is; does the PPC tick handler have its own stack or does it just use the stack of the task that is active when the decrementer interrupt goes off? We are having some stack corruption problems and I've traced it to our message sending routine that gets called in a watchdog timer callback function. The msg send routine has a stack frame of about 4100 bytes. The task that is having its stack corrupted has a stack allocation of 5000 bytes. I set a data write breakpoint on the location being corrupted and the stack trace showed our message callback routine and windTickAnnounce in the trace, along with the task whose stack was getting corrupted. Pete - -- Pete Kockritz --------------------------- Newsgroups: comp.protocols.snmp,comp.os.psos,comp.os.vxworks Subject: getproc_got_string() and structure Date: Sun, 10 Jun 2001 09:40:03 +0200 From: Alex Vinokur Message-ID: <3B232452.8A39582D@bigfoot.com> ============================ pRISM+ for pSOSystem/PowerPC pRISM+ 1.2.3 . pSOSystem 2.2.2 Diab Data D-CC Optimizing C compiler SNMP Release 7.4 (SNMPv1) Epilogue Technology Emissary SNMP MIB Compiler, version 6.3 Host : Windows98 Target : PowerPC (MPC860) MG-SOFT MIB Browser, Professional Edition 6.10.0.3633 ============================ Hi, I have a problem while using getproc_got_string() and initializing char[] in structure. Here are some files. //--- mib file (fragments) ----- pss OBJECT IDENTIFIER ::= { isi 1 } pssVers OBJECT-TYPE SYNTAX DisplayString ACCESS read-only STATUS mandatory DESCRIPTION "Stuff" ::= { pss 1 } pssDebugMode OBJECT-TYPE SYNTAX INTEGER { dbgsa(1), dbgxs(2), dbgxn(3), dbgap(4), storage(6) } ACCESS read-write STATUS mandatory DESCRIPTION "Stuff" ::= { pss 2 } pssSCPSOS OBJECT-TYPE SYNTAX INTEGER { no(0), yes(1) } ACCESS read-write STATUS mandatory DESCRIPTION "Stuff" ::= { pss 3 } pssSCPNA OBJECT-TYPE SYNTAX INTEGER { no(0), yes(1) } ACCESS read-write STATUS mandatory DESCRIPTION "Stuff" ::= { pss 4 } -- Here is a problem pssString1 OBJECT-TYPE SYNTAX DisplayString (SIZE (0..101)) ACCESS read-only STATUS mandatory DESCRIPTION "This is pssString1" ::= { pss 5 } -- Here is a problem pssString2 OBJECT-TYPE SYNTAX DisplayString (SIZE (0..102)) ACCESS read-write STATUS mandatory DESCRIPTION "This is pssString2" ::= { pss 6 } -- Here is OK pssString3 OBJECT-TYPE SYNTAX DisplayString (SIZE (0..103)) ACCESS read-write STATUS mandatory DESCRIPTION "This is pssString3" ::= { pss 7 } pssInteger1 OBJECT-TYPE SYNTAX INTEGER (0..121) ACCESS read-only STATUS mandatory DESCRIPTION "This is pssInteger1" ::= { pss 8 } pssInteger2 OBJECT-TYPE SYNTAX INTEGER (0..122) ACCESS read-write STATUS mandatory DESCRIPTION "This is pssInteger2" ::= { pss 9 } //------------------------------ //--- control file (fragments) - -- pss pss OBJECT-TYPE DEFAULT set-function-async pss_set_async DEFAULT get-function-async pss_get_async DEFAULT test-function-async pss_test_async DEFAULT next-function-async pss_next_async DEFAULT cookie (char *) 0 ::= { isi 1 } pssVers OBJECT-TYPE ::= { pss 1 } pssDebugMode OBJECT-TYPE ::= { pss 2 } pssSCPSOS OBJECT-TYPE ::= { pss 3 } pssSCPNA OBJECT-TYPE ::= { pss 4 } pssString1 OBJECT-TYPE DEFAULT cookie (char *) the_pss.%n ::= { pss 5 } pssString2 OBJECT-TYPE DEFAULT cookie (char *) the_pss.%n ::= { pss 6 } pssString3 OBJECT-TYPE DEFAULT cookie (char *) the_%n ::= { pss 7 } pssInteger1 OBJECT-TYPE ::= { pss 8 } pssInteger2 OBJECT-TYPE ::= { pss 9 } //------------------------------ //--- File stub.h (fragments) -- typedef struct { char *pssVers; ULONG pssDebugMode; ULONG pssSCPSOS; ULONG pssSCPNA; char pssString1[MAXSIZE_pssString1]; char pssString2[MAXSIZE_pssString2]; int pssInteger1; int pssInteger2; } STRUCT_pss; extern STRUCT_pss the_pss; extern char the_pssString3[MAXSIZE_pssString3]; #define string_length(x) strlen(x) //------------------------------ //--- File stub1.c (fragments) - static int pss_get_value ( OIDC_T lastmatch, SNMP_PKT_T *pktp, VB_T *vbp, STRUCT_pss *data) { switch(lastmatch) { case LEAF_pssVers : getproc_got_string(pktp, vbp, string_length(data->pssVers), data->pssVers, 0, VT_STRING); break; case LEAF_pssDebugMode : getproc_got_int32(pktp, vbp, data->pssDebugMode); break; case LEAF_pssSCPSOS: getproc_got_int32(pktp, vbp, data->pssSCPSOS); break; case LEAF_pssSCPNA: getproc_got_int32(pktp, vbp, data->pssSCPNA); break; // Here is a problem case LEAF_pssString1: getproc_got_string(pktp, vbp, string_length(data->pssString1), (char *) data->pssString1, 0, VT_STRING); break; // Here is a problem case LEAF_pssString2: getproc_got_string(pktp, vbp, string_length(data->pssString2), (char*) data->pssString2, 0, VT_STRING); break; // Here is OK case LEAF_pssString3: getproc_got_string(pktp, vbp, string_length(the_pssString3), the_pssString3, 0, VT_STRING); break; case LEAF_pssInteger1: getproc_got_int32(pktp, vbp, data->pssInteger1); break; case LEAF_pssInteger2: getproc_got_int32(pktp, vbp, data->pssInteger2); break; default: return GEN_ERR; } return NO_ERROR; } // pss_get_value //------------------------------ //--- File stub2.c (fragments) - const char the_pssVers[] = "Version1"; char the_pssString3[MAXSIZE_pssString3] = "This is pssString3"; // Here is OK STRUCT_pss the_pss = { the_pssVers, // pssVers SE_DEBUG_MODE, // pssDebugMode SC_PSOS, // pssSCPSOS SC_PNA, // pssSCPNA "This is pssString1", // pssString1 : Here is a problem "This is pssString2", // pssString2 : Here is a problem 201, // pssInteger1 202 // pssInteger2 }; //------------------------------ //--- Running : via MG-SOFT ---- // Walk pss ***** SNMP QUERY STARTED ***** 1: pssVers.0 (octet string) Version1 2: pssDebugMode.0 (integer) 5 3: pssSCPSOS.0 (integer) yes(1) 4: pssSCPNA.0 (integer) yes(1) 5: pssString1.0 (octet string) 00.3C.97.00.20.69.73.20.00.3C.97.08.00.3C.0D.04.00.3C (hex) 6: pssString2.0 (octet string) 00.00.00.00.3C.97.4D.00.3C.0C.FC.00.3C.98.A0.00.1B.40 (hex) 7: pssString3.0 (octet string) This is pssString3 8: pssInteger1.0 (integer) 201 9: pssInteger2.0 (integer) 202 ***** SNMP QUERY FINISHED ***** //------------------------------ We can see that there is a problem with Get-Request/Response for pssString1 and pssString2. Note. pssString1 and pssString2 are initialized in structure. There is no problem with Get-Request/Response for pssString3. Note. pssString2 are initialized as simple variable. What is wrong? ================================== Alex Vinokur mailto:alexvn@bigfoot.com http://up.to/alexvn http://visitweb.com/alexvn.math ================================== --------------------------- Newsgroups: comp.protocols.snmp,comp.os.psos,comp.os.vxworks Subject: Re: getproc_got_string() and structure Date: Sun, 10 Jun 2001 11:19:13 +0200 From: Alex Vinokur Message-ID: <3B233B8F.7E89440C@bigfoot.com> References: <3B232452.8A39582D@bigfoot.com> Alex Vinokur wrote: > ============================ > pRISM+ for pSOSystem/PowerPC > pRISM+ 1.2.3 . pSOSystem 2.2.2 > Diab Data D-CC Optimizing C compiler > SNMP Release 7.4 (SNMPv1) > Epilogue Technology Emissary SNMP MIB Compiler, version 6.3 > Host : Windows98 > Target : PowerPC (MPC860) > > MG-SOFT MIB Browser, Professional Edition 6.10.0.3633 > ============================ > > Hi, > > I have a problem while > using getproc_got_string() and > initializing char[] in structure. > > Here are some files. > > //--- mib file (fragments) ----- > > pss OBJECT IDENTIFIER ::= { isi 1 } > > pssVers OBJECT-TYPE > SYNTAX DisplayString > ACCESS read-only > STATUS mandatory > DESCRIPTION "Stuff" > ::= { pss 1 } > > pssDebugMode OBJECT-TYPE > SYNTAX INTEGER { > dbgsa(1), > dbgxs(2), > dbgxn(3), > dbgap(4), > storage(6) > } > ACCESS read-write > STATUS mandatory > DESCRIPTION "Stuff" > ::= { pss 2 } > > pssSCPSOS OBJECT-TYPE > SYNTAX INTEGER { > no(0), > yes(1) > } > ACCESS read-write > STATUS mandatory > DESCRIPTION "Stuff" > ::= { pss 3 } > > pssSCPNA OBJECT-TYPE > SYNTAX INTEGER { > no(0), > yes(1) > } > ACCESS read-write > STATUS mandatory > DESCRIPTION "Stuff" > ::= { pss 4 } > > -- Here is a problem > pssString1 OBJECT-TYPE > SYNTAX DisplayString (SIZE (0..101)) > ACCESS read-only > STATUS mandatory > DESCRIPTION "This is pssString1" > ::= { pss 5 } > > -- Here is a problem > pssString2 OBJECT-TYPE > SYNTAX DisplayString (SIZE (0..102)) > ACCESS read-write > STATUS mandatory > DESCRIPTION "This is pssString2" > ::= { pss 6 } > > -- Here is OK > pssString3 OBJECT-TYPE > SYNTAX DisplayString (SIZE (0..103)) > ACCESS read-write > STATUS mandatory > DESCRIPTION "This is pssString3" > ::= { pss 7 } > > pssInteger1 OBJECT-TYPE > SYNTAX INTEGER (0..121) > ACCESS read-only > STATUS mandatory > DESCRIPTION "This is pssInteger1" > ::= { pss 8 } > > pssInteger2 OBJECT-TYPE > SYNTAX INTEGER (0..122) > ACCESS read-write > STATUS mandatory > DESCRIPTION "This is pssInteger2" > ::= { pss 9 } > > //------------------------------ > > //--- control file (fragments) - > > -- pss > pss OBJECT-TYPE > DEFAULT set-function-async pss_set_async > DEFAULT get-function-async pss_get_async > DEFAULT test-function-async pss_test_async > DEFAULT next-function-async pss_next_async > DEFAULT cookie (char *) 0 > ::= { isi 1 } > > pssVers OBJECT-TYPE > ::= { pss 1 } > > pssDebugMode OBJECT-TYPE > ::= { pss 2 } > > pssSCPSOS OBJECT-TYPE > ::= { pss 3 } > > pssSCPNA OBJECT-TYPE > ::= { pss 4 } > > pssString1 OBJECT-TYPE > DEFAULT cookie (char *) the_pss.%n > ::= { pss 5 } > > pssString2 OBJECT-TYPE > DEFAULT cookie (char *) the_pss.%n > ::= { pss 6 } > > pssString3 OBJECT-TYPE > DEFAULT cookie (char *) the_%n > ::= { pss 7 } > > pssInteger1 OBJECT-TYPE > ::= { pss 8 } > > pssInteger2 OBJECT-TYPE > ::= { pss 9 } > > //------------------------------ > > //--- File stub.h (fragments) -- > > typedef struct > { > char *pssVers; > ULONG pssDebugMode; > ULONG pssSCPSOS; > ULONG pssSCPNA; > char pssString1[MAXSIZE_pssString1]; > char pssString2[MAXSIZE_pssString2]; > int pssInteger1; > int pssInteger2; > } STRUCT_pss; > extern STRUCT_pss the_pss; > extern char the_pssString3[MAXSIZE_pssString3]; > > #define string_length(x) strlen(x) > > //------------------------------ > > //--- File stub1.c (fragments) - > > static int pss_get_value ( > OIDC_T lastmatch, > SNMP_PKT_T *pktp, > VB_T *vbp, > STRUCT_pss *data) > { // It works OK if we are using intermediate valiable: char tmp_str[1000]; > > switch(lastmatch) > { > case LEAF_pssVers : > getproc_got_string(pktp, vbp, string_length(data->pssVers), > data->pssVers, 0, VT_STRING); > break; > > case LEAF_pssDebugMode : > getproc_got_int32(pktp, vbp, data->pssDebugMode); > break; > > case LEAF_pssSCPSOS: > getproc_got_int32(pktp, vbp, data->pssSCPSOS); > break; > > case LEAF_pssSCPNA: > getproc_got_int32(pktp, vbp, data->pssSCPNA); > break; > > // Here is a problem > case LEAF_pssString1: // It works OK : strcpy (tmp_str, data->pssString1); getproc_got_string(pktp, vbp, string_length(tmp_str), tmp_str, 0, VT_STRING); // But why do we have to use intermediate variable? //--------------------------------------------------------------------- This doesn't work OK : > > getproc_got_string(pktp, vbp, string_length(data->pssString1), > (char *) data->pssString1, 0, VT_STRING); //---------------------------------------------------------------------- > > break; > > // Here is a problem > case LEAF_pssString2: > getproc_got_string(pktp, vbp, string_length(data->pssString2), > (char*) data->pssString2, 0, VT_STRING); > break; > > // Here is OK > case LEAF_pssString3: > getproc_got_string(pktp, vbp, string_length(the_pssString3), > the_pssString3, 0, VT_STRING); > break; > > case LEAF_pssInteger1: > getproc_got_int32(pktp, vbp, data->pssInteger1); > break; > > case LEAF_pssInteger2: > getproc_got_int32(pktp, vbp, data->pssInteger2); > break; > > default: > return GEN_ERR; > } > return NO_ERROR; > > } // pss_get_value > > //------------------------------ > > //--- File stub2.c (fragments) - > > const char the_pssVers[] = "Version1"; > char the_pssString3[MAXSIZE_pssString3] = "This is pssString3"; // Here > is OK > STRUCT_pss the_pss = > { > the_pssVers, // pssVers > SE_DEBUG_MODE, // pssDebugMode > SC_PSOS, // pssSCPSOS > SC_PNA, // pssSCPNA > "This is pssString1", // pssString1 : Here is a problem > "This is pssString2", // pssString2 : Here is a problem > 201, // pssInteger1 > 202 // pssInteger2 > }; > > //------------------------------ > > //--- Running : via MG-SOFT ---- > // Walk pss > > ***** SNMP QUERY STARTED ***** > 1: pssVers.0 (octet string) Version1 > 2: pssDebugMode.0 (integer) 5 > 3: pssSCPSOS.0 (integer) yes(1) > 4: pssSCPNA.0 (integer) yes(1) > 5: pssString1.0 (octet string) > 00.3C.97.00.20.69.73.20.00.3C.97.08.00.3C.0D.04.00.3C (hex) > 6: pssString2.0 (octet string) > 00.00.00.00.3C.97.4D.00.3C.0C.FC.00.3C.98.A0.00.1B.40 (hex) > 7: pssString3.0 (octet string) This is pssString3 > 8: pssInteger1.0 (integer) 201 > 9: pssInteger2.0 (integer) 202 > ***** SNMP QUERY FINISHED ***** > > //------------------------------ > > We can see that there is a problem > with Get-Request/Response for pssString1 and pssString2. > Note. pssString1 and pssString2 are initialized in structure. > > There is no problem with Get-Request/Response for pssString3. > Note. pssString2 are initialized as simple variable. > > What is wrong? > > ================================== > Alex Vinokur > mailto:alexvn@bigfoot.com > http://up.to/alexvn > http://visitweb.com/alexvn.math > ================================== --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Mon Jun 11 02:03:19 2001 From: David Laight Date: Mon Jun 11 02:03:21 PDT 2001 Subject: Re: BSP questions- int vector and remap > > 1. Is there any sample code showing the RE-MAPPING procedure illustrating > > what one should prepare for remapping? > You could found a boot example with a remapping example in the document: > "Implementation of an universal Boot Monitor for an ARM-based System" > at this address: > http://mikro.ee.tu-berlin.de/~hellmich/theses.html > If you have others examples of remapping please send to me. > > 2. Also if ROM address starts from 0x0, what software approach can I do in > > VxWorks to made RAM appear in 0x0 instead of ROM? The vxWorks book talks of a 'walk of death' where you use an intermediate set of page tables with the ROM mapped at 0 and a higher address in order to get RAM mapped to address zero. I use a much simpler approach on our SA1100 system. The strongarm does not fully decode the low physical addresses - even if it did your board hardware probably does not. This means the ROM (at physical address 0) is aliased by the hardware to 64Mb. All of the ROM code, and any vxWorks code is built assuming the ROM starts at 64Mb. An early instruction of the ROM is: start: ldr pc,L__entry /* Jump to expected alias of this code */ entry: ... L__entry: .long ROM_TEXT_ADRS + entry - _romInit This allows me to enable the mmu with DRAM at virtual address 0 and the ROM address unchanged - much simpler. There is, of course, only 64MB of contiguous address space available for DRAM. 'Warm' boots (including the one done by the network loader) cause grief because you can end up re-initialising the MMU while running from page tables already in memory. If you are unlucky the old page tables are in memory the new build is using for something else! even initialising then in the same place is fraught with sequencing issues. All made slightly more complicated because oage table entries are cached in the TLB. David ---------------------------------------------------------------- David Laight email: dsl@tadpole.co.uk Tadpole Technology plc phone: +44 1223 428 232 Cambridge, UK fax: +44 1223 428 201 From vxwexplo-errs@csg.lbl.gov Mon Jun 11 02:19:00 2001 From: David Laight Date: Mon Jun 11 02:19:02 PDT 2001 Subject: Re: Proper way to unmount dosFs > To unmount dosFs that was mounted on an ATA disk > at "/ata", I'm doing: > > char* pTail; > iosDevDelete(iosDevFind("/ata", &pTail)); > Is that all that's needed for the unmount? No! All iosDevdelete() does is remove the entry from the list of device drivers known to the system. The device driver isn't told - or allowed to reject this request. Moreover if you have "/ata" open (as opposed to "/atafred") the file data area has a pointer to the "/ata" string in the ios device list. If you do an iosDevAdd for "/ata" before the fd is closed, then close the fd the system will free the string - causing a memory heap error! The nearest you can do to unmount a dos filesystem is to call all the volume changed rountines - however they do not free all the memory used by dosfs (eg the cache of the FAT and a tree copy of the entire directory structure). I have been told that there is no 'unmount' routine because: 1) there is no requirement 2) it is too hard to do it properly Makes supporting PCMCIA disks in vxworks a bit lively! WRS support may give you some random source which will free up the memory held by dosfs...... David ---------------------------------------------------------------- David Laight email: dsl@tadpole.co.uk Tadpole Technology plc phone: +44 1223 428 232 Cambridge, UK fax: +44 1223 428 201 From vxwexplo-errs@csg.lbl.gov Mon Jun 11 02:27:50 2001 From: David Laight Date: Mon Jun 11 02:27:52 PDT 2001 Subject: Re: how to check whether fpp register contents is a number? > > We are using vxworks 5.3.1 with PowerPC750 board. > with fppTaskRegsShow we get the following output: > fppTaskRegsShow 0x3b89d10 > > fpcsr = 70 > fr0 = NaN fr1 = NaN fr2 = NaN fr3 > = NaN > > We would like to be able to perform the test on the registers for NaN > (Not a Number). > We were unable to trackdown the functioncall to do this. > On Solaris the function isnan() is available to perform this. > Who can tell us which function we should use on vxworks or with what > code we perform the isnan test? IF your FP routines are properly conformant then any of the following should work: 1) if (x != x) return NAN 2) sprint( buf, "%f", x ); if (!strcmp( buf, "NaN" ) return NAN 3) if (((int *)&x)[0] & 0x7ff00000 == 0x7ff00000 && ( ((int *)&x)[0] & 0xfffff || ((int *)&x)[1] )) return NAN (Third example is for double, the [0] and [1] may need reversing depending on your FP number endianness) David ---------------------------------------------------------------- David Laight email: dsl@tadpole.co.uk Tadpole Technology plc phone: +44 1223 428 232 Cambridge, UK fax: +44 1223 428 201 From vxwexplo-errs@csg.lbl.gov Mon Jun 11 02:49:55 2001 From: David Laight Date: Mon Jun 11 02:49:57 PDT 2001 Subject: Re: Redirecting stdout problems > I am having problems redirecting stdout. I am trying to provide > the ability for our the CLI in our application to call certain > VxWorks functions such as i(), ifShow(), and ti() so that they > may be executed via our CLI through telnet (not the Tornado > shell). Firstly I would run these functions from a separate task whose stdout and stderr are redirected to you pipe - rather than changing the global descriptors. Actually why not make your 'telnet' server have a device driver interface - so it picks up data to transmit from write() calls. Then you have no more work to do! This is vxWorks, there is no user-kernel boundary to worry about, you can write data to a socket from within another device drivers write routine. (I would be a little careful about open(), but the data area locking in vxWorks seems (IMHO) so sloppy you will not hit a deadlock). Even a trivial pipe() device driver is easy to write. - open( "/mypipe" ) gives one side - ioctl( fd, MYP_SETNAME, "name" ) gives it a name - open( "/mypipe/name ) will open the other side - read blocks if no one writing - write blocks if no one reading - transfer data straight from user write buffer to user read buffer. - maybe optionally push read data through ttyLib David ---------------------------------------------------------------- David Laight email: dsl@tadpole.co.uk Tadpole Technology plc phone: +44 1223 428 232 Cambridge, UK fax: +44 1223 428 201 From vxwexplo-errs@csg.lbl.gov Mon Jun 11 04:03:16 2001 From: Vxworks Exploder Date: Mon Jun 11 04:03:18 PDT 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Mon Jun 11 04:03:07 PDT 2001 Subject: Help for init. 53C875 SCSI controller Subject: Re: internal compiler errors Subject: Re: Processing of Raw TCP/IP Packets Subject: Re: problem with inet_aton Subject: Please try this on Diab C++ Subject: Re: Help for init. 53C875 SCSI controller Subject: re: BSP questions- int vector and remap Subject: Command 'iosFdShow' does not work in Shell Subject: re: Proper way to unmount dosFs Subject: re: how to check whether fpp register contents is a number? Subject: re: Redirecting stdout problems Subject: Problem with stack trace and targetshell ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Help for init. 53C875 SCSI controller Date: 10 Jun 2001 04:51:06 -0700 From: lilyko2002@yahoo.com (lily) Organization: http://groups.google.com/ Message-ID: <7f208981.0106100351.2b2a5849@posting.google.com> Hi all, I need help in configuration SYM53C875 SCSI controller My Environment: Tornado 2, VxW 5.4 HW (Target) Intel x586 : board: pc104, on PCI bus connected SYM53C875 SCSI controller. I use the following code to init the SCSI controller: 1. pciFindDevice (devVendor, 0xf, index, &busNo, &devNo, &funcNo) 2. pciDevConfig(busNo, devNo, funcNo, ioBase, memBase, PCI_CMD_IO_ENABLE | PCI_CMD_MASTER_ENABLE); 3. pSysScsiCtrl = (SCSI_CTRL *) ncr810CtrlCreate ((UINT8 *)ioBase,NCR810_3750MHZ , 0xf) 4. pciIntConnect (INUM_TO_IVEC (((int)irq) + INT_NUM_IRQ0), ncr810Intr, (int) pSysScsiCtrl); 5. ncr810CtrlInit ((NCR_810_SCSI_CTRL *) pSysScsiCtrl, SCSI_DEF_CTRL_BUS_ID) 6. sysIntEnablePIC((int) Sym875.irq) ; 7. scsiCacheSnoopDisable ((SCSI_CTRL *) pSysScsiCtrl); 8. ncr810SetHwRegister ((NCR_810_SCSI_CTRL *)pSysScsiCtrl, &hwRegs); 9. scsiDebug,scsiIntsDebug,scsiErrors = TRUE; 10. scsiAutoConfig(pSysScsiCtrl); Here are the debug messages: When I'm calling scsiAutoConfig, the processor hangs. x3f08e60 (tDbgTask): scsiAutoConfig: bus ID = 0, LUN = 0 0x3f08e60 (tDbgTask): scsiReqSense: 0x3f08e60 (tDbgTask): scsi2Transact: 0x3f02e6c (tScsiTask): scsiMgrRequestProc: client request: 0 0x3f02e6c (tScsiTask): scsiMgrActivateRequest: thread 0x03ffcb8c 0x3f02e6c (tScsiTask): ncr810ThreadActivate: thread 0x03ffcb8c: activating 0x3f02e6c (tScsiTask): ncr810XferParamsCvt: converted to: offset = 0, period = 0 xferParams = 0, clockDivide = 2 0x3f02e6c (tScsiTask): ncr810ThreadStateSet: thread 0x03ffcb8c: 1 -> 6 0x3f02e6c (tScsiTask): scsiMgrTimeoutProc: thread 0x03ffcb8c (state = 6) timed o ut 0x3f02e6c (tScsiTask): ncr810ThreadAbort: thread 0x03ffcb8c (state = 6) aborting 0x3f02e6c (tScsiTask): ncr810ThreadStateSet: thread 0x03ffcb8c: 6 -> 9 Maybe somebody knows what happens? Am I configured Interrupt for SCSI controller wrong? Thanks for any suggestions, Lily --------------------------- Newsgroups: comp.os.vxworks Subject: Re: internal compiler errors Date: Mon, 11 Jun 2001 00:58:34 +0100 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: <3B207CA7.D938FC01@mediaone.net> "jagan" wrote in message news:3B207CA7.D938FC01@mediaone.net... > > > I am seeing an internal compiler error when compiling > using Tornado 2 cross compiler for a ppc target in > a cygwin environment on NT. > > I see cc1.exe coring with signal 5. > > Does anyone know why this may be happening? Not really. Could be flaky hardware or configuration - try doing the same build on a different PC. Show us the actual output that you get - copy + paste it from the screen into a reply. Davek --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Processing of Raw TCP/IP Packets Date: Mon, 11 Jun 2001 01:04:35 +0100 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: <628eca30.0106090633.251aa368@posting.google.com> "Ivan Bach" wrote in message news:628eca30.0106090633.251aa368@posting.google.com... > I have not yet seen vxsniff.c program, but it was apparently based on > etherhooks. Does anybody actually have a copy of that program, even if > it is not working? The pointers that I saw for that program don't lead > anywhere. [DK] I hope Hwa-Jin Bae won't be offended by my distributing this. /* * quick and dirty ethernet sniffer for vxworks. * Provided as an example. It did its work for me but there is no * guarantee that it will work for you. * * Written by Hwa-Jin Bae, bae@mail.com, Piedmont California */ #include "vxWorks.h" #include "sys/types.h" #include "ioLib.h" #include "iosLib.h" #include "fioLib.h" #include "stdio.h" #include "unistd.h" #include "string.h" #include "usrLib.h" #include "errnoLib.h" #include "hostLib.h" #include "sockLib.h" #include "socket.h" #include "inetLib.h" #include "in.h" #include "selectLib.h" #include "taskLib.h" #include "dosFsLib.h" #include "etherLib.h" #include "systm.h" #include "mbuf.h" #include "domain.h" #include "tickLib.h" #include "protosw.h" #include "socket.h" #include "errno.h" #include "if.h" #include "route.h" #include "in.h" #include "in_pcb.h" #include "in_systm.h" #include "in_var.h" #include "ip.h" #include "ip_var.h" #include "ip_icmp.h" #include "udp.h" #include "tcp.h" #include "if_ether.h" #include "systm.h" #include "inetLib.h" char vxSniffLog[200][200]; int vxSniffLogIdx= 0; STATUS vxSniffHook(struct ifnet *ifp, char *buf, int len) { struct ip *ip; struct udphdr *udp; struct tcphdr *tcp; unsigned long ticks; char *proto; int optlen; char srcAddr[20], dstAddr[20]; char protoxxx[10]; ip = (struct ip *)(buf + 14); inet_ntoa_b(ip->ip_src, srcAddr); inet_ntoa_b(ip->ip_dst, dstAddr); switch (ip->ip_p) { case IPPROTO_TCP: proto = "TCP"; break; case IPPROTO_UDP: proto = "UDP"; break; case IPPROTO_ICMP: proto = "ICMP"; break; default: sprintf(protoxxx, "?0x%x", ip->ip_p); proto = protoxxx; break; } ticks = tickGet (); sprintf(vxSniffLog[vxSniffLogIdx++], "0x%x:IP(p:%s,v:%d,h:%d,l:%d) %s -> %s\n", ticks, proto, (ip->ip_v_hl>>4)&0xf, (ip->ip_v_hl & 0xf)<<2, ip->ip_len, srcAddr, dstAddr); if (vxSniffLogIdx == 200) vxSniffLogIdx = 0; optlen = ((ip->ip_v_hl & 0xf)<<2) - sizeof( struct ip ); if (ip->ip_p == IPPROTO_UDP) { udp = (struct udphdr *)((char *)(ip + 1) + optlen); sprintf(vxSniffLog[vxSniffLogIdx++], " UDP(s:%d,d:%d,l:%d)\n", udp->uh_sport, udp->uh_dport, udp->uh_ulen); } else if (ip->ip_p == IPPROTO_TCP) { tcp = (struct tcphdr *)((char *)(ip + 1) + optlen); sprintf(vxSniffLog[vxSniffLogIdx++], " TCP(s:%d,d:%d,q:%d,a:%d,f:0x%x,w:%d)\n", tcp->th_sport, tcp->th_dport, tcp->th_seq, tcp->th_ack, tcp->th_flags, tcp->th_win); } if (vxSniffLogIdx == 200) vxSniffLogIdx = 0; return FALSE; } int vxSniffStartIdx = 0; void vxSniffStart() { vxSniffStartIdx = vxSniffLogIdx; } void vxSniffShow() { int end = vxSniffLogIdx; int i; if (end < vxSniffStartIdx) { for (i = vxSniffStartIdx; i < 200; i++) printf("%s", vxSniffLog[i]); for (i = 0; i < end; i++) printf("%s", vxSniffLog[i]); } else { for (i = vxSniffStartIdx; i < end; i++) printf("%s", vxSniffLog[i]); } } void vxSniffInit() { /* XXX do something here to put ethernet driver in promiscuous mode */ etherInputHookAdd(vxSniffHook); } --------------------------- Newsgroups: comp.os.vxworks Subject: Re: problem with inet_aton Date: Mon, 11 Jun 2001 11:06:20 +0100 From: Vijay Kumar Peshkar Organization: Cisco Systems Inc. Message-ID: <3B24981B.2A9CF1F5@wipro.com> References: This is a multi-part message in MIME format. - --------------5F5C6BB0224C17539821B14C Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Michael, The macro OK has value 0 and in your case as everything seems to be working fine, the func. call is returning OK ;). You should be comparing it with ERROR(-1). The inet_addr func. is supported in vxWorks. Sorry I could not understand what's "rtrequest". cheers, Vijay Michael Yeung wrote: > Hi all, > I'm not sure why but the inet_aton seems to always returns 0 and > conversion is not done properly. > > struct in_addr temp; > > if(inet_aton("1.1.1.1", (struct in_addr *)&temp) == 0) > { > printf("Failed conversion\n"); > } > > I noticed VxWorks does not support "inet_addr" or "inet_pton", as > suggestion is appreciated. > > As well, does anyone knows the format expected for mask to function like > "rtrequest"? > > thanx in advance, > Michael - --------------5F5C6BB0224C17539821B14C Content-Type: text/x-vcard; charset=us-ascii; name="vijay.peshkar.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Vijay Kumar Peshkar Content-Disposition: attachment; filename="vijay.peshkar.vcf" begin:vcard n:Peshkar;Vijay Kumar tel;work:5530034/35/35 x2237 x-mozilla-html:FALSE org:Wipro Global R & D;Wishing you blue skies, a couple of daisies, some fresh air and a smile! adr:;;;;;; version:2.1 email;internet:vpeshkar@cisco.com note:Wishing you blue skies, a couple of daisies, some fresh air and a smile! fn:Vijay Kumar Peshkar end:vcard - --------------5F5C6BB0224C17539821B14C-- --------------------------- Newsgroups: comp.os.vxworks Subject: Please try this on Diab C++ Date: 11 Jun 2001 01:26:00 -0700 From: leonardmatherson@usa.net (Leonard Matherson) Organization: http://groups.google.com/ Message-ID: Hello, We are using the 4.3g Diab compiler under Tornado 2.0 Our target is PPC860. Please consider the following: class X private: int XX; . . class Y X YY; foo(); . . foo() int any = YY.XX We should get a message saying that we do not have access rights to the private members of a class, but when we compile the following we get a clean compile and a good object file. Other compilers (GNU, VC6) of course give us the correct messages. Here is the source code that I compiled - no headers are needed. Please try this on your compiler, Thanks Leonard. class CTryacc { public: int ddd; private: int CanGet; }; class MiddleLevel { public: int sss; int fff(void); CTryacc TryAcc; }; int MiddleLevel::fff() { int yyy = TryAcc.CanGet; return (yyy); } void LetsSee(void) { MiddleLevel Upper; int nSome = Upper.fff(); } --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Help for init. 53C875 SCSI controller Date: 11 Jun 2001 01:45:50 -0700 From: shw64@hotmail.com (Julia) Organization: http://groups.google.com/ Message-ID: <3f7326d7.0106110045.39fcd8c8@posting.google.com> References: <7f208981.0106100351.2b2a5849@posting.google.com> lilyko2002@yahoo.com (lily) wrote in message news:<7f208981.0106100351.2b2a5849@posting.google.com>... > Hi all, > > I need help in configuration SYM53C875 SCSI controller > > My Environment: > Tornado 2, VxW 5.4 > HW (Target) Intel x586 : board: pc104, > on PCI bus connected SYM53C875 SCSI controller. > > I use the following code to init the SCSI controller: > > 1. pciFindDevice (devVendor, 0xf, > index, &busNo, &devNo, &funcNo) > > 2. pciDevConfig(busNo, devNo, funcNo, > ioBase, memBase, > PCI_CMD_IO_ENABLE | PCI_CMD_MASTER_ENABLE); > > 3. pSysScsiCtrl = (SCSI_CTRL *) ncr810CtrlCreate > ((UINT8 *)ioBase,NCR810_3750MHZ , 0xf) > > 4. pciIntConnect (INUM_TO_IVEC (((int)irq) + INT_NUM_IRQ0), > ncr810Intr, (int) pSysScsiCtrl); > > 5. ncr810CtrlInit ((NCR_810_SCSI_CTRL *) pSysScsiCtrl, > SCSI_DEF_CTRL_BUS_ID) > > 6. sysIntEnablePIC((int) Sym875.irq) ; > > 7. scsiCacheSnoopDisable ((SCSI_CTRL *) pSysScsiCtrl); > > 8. ncr810SetHwRegister ((NCR_810_SCSI_CTRL *)pSysScsiCtrl, > &hwRegs); > > 9. scsiDebug,scsiIntsDebug,scsiErrors = TRUE; > > 10. scsiAutoConfig(pSysScsiCtrl); > > Here are the debug messages: > > > When I'm calling scsiAutoConfig, the processor hangs. > > x3f08e60 (tDbgTask): scsiAutoConfig: bus ID = 0, LUN = 0 > 0x3f08e60 (tDbgTask): scsiReqSense: > 0x3f08e60 (tDbgTask): scsi2Transact: > 0x3f02e6c (tScsiTask): scsiMgrRequestProc: client request: 0 > 0x3f02e6c (tScsiTask): scsiMgrActivateRequest: thread 0x03ffcb8c > 0x3f02e6c (tScsiTask): ncr810ThreadActivate: thread 0x03ffcb8c: activating > 0x3f02e6c (tScsiTask): ncr810XferParamsCvt: converted to: offset = 0, > period > = 0 > xferParams = 0, clockDivide = 2 > 0x3f02e6c (tScsiTask): ncr810ThreadStateSet: thread 0x03ffcb8c: 1 -> 6 > 0x3f02e6c (tScsiTask): scsiMgrTimeoutProc: thread 0x03ffcb8c (state = 6) > timed o > ut > 0x3f02e6c (tScsiTask): ncr810ThreadAbort: thread 0x03ffcb8c (state = 6) > aborting > > 0x3f02e6c (tScsiTask): ncr810ThreadStateSet: thread 0x03ffcb8c: 6 -> 9 > > Maybe somebody knows what happens? Am I configured Interrupt for > SCSI controller wrong? > > Thanks for any suggestions, > > Lily Hi,Lily, it looks like your interrupts from the controller don't reach your CPU. Unfortunately I don't know about details of target you're using. If you're sure that your parameters in pciIntConnect are right and you enable proper level of interrupts on CPU (I use intConnect&intEnable), do one more thing: I have also option PCI_CMD_MEM_ENABLE during my PCI_CFG_COMMAND register configuration: pciConfigOutLong(pRsrc->pciBus,pRsrc->pciDevice,pRsrc->pciFunc,PCI_CFG_BASE_ADDRESS_0,pRsrc->memaddr); EIEIO_SYNC; pciConfigOutLong(pRsrc->pciBus,pRsrc->pciDevice,pRsrc->pciFunc,PCI_CFG_BASE_ADDRESS_1,pRsrc->ioaddr);EIEIO_SYNC;/*|PCI_BASE_IO*/ pciConfigOutLong(pRsrc->pciBus,pRsrc->pciDevice,pRsrc->pciFunc,PCI_CFG_BASE_ADDRESS_2,pRsrc->dramaddr);EIEIO_SYNC; pciConfigOutLong(pRsrc->pciBus,pRsrc->pciDevice,pRsrc->pciFunc,PCI_CFG_DEV_INT_LINE,pRsrc->irqvec);EIEIO_SYNC; pciConfigOutLong(pRsrc->pciBus,pRsrc->pciDevice,pRsrc->pciFunc,PCI_CFG_COMMAND,PCI_CMD_IO_ENABLE|PCI_CMD_MEM_ENABLE|PCI_CMD_MASTER_ENABLE); You'll have screen something like this if all is configured properly. 0x1c59950 (t8): scsiAutoConfig: bus ID = 0, LUN = 0 0x1c59950 (t8): scsiReqSense: 0x1c59950 (t8): scsi2Transact: 0x1ef7868 (tScsiTask): scsiMgrRequestProc: client request: 0 0x1ef7868 (tScsiTask): scsiMgrActivateRequest: thread 0x01fff990 0x1ef7868 (tScsiTask): ncr810ThreadActivate: thread 0x01fff990: activating 0x1ef7868 (tScsiTask): ncr810XferParamsCvt: converted to: offset = 0, period = 0 xferParams = 0, clockDivide = 3 interrupt: ncr810EventTypeGet:Interrupt status istat = 0x01, SCSI status0 = 0x00, SCSI status1 = 0x00DMA status = 0x84 interrupt: ncr810EventTypeGet: script completed. interrupt: ncr810Intr: DSA = 0x01ef8a3c, DSP = 0x001d2cd8, type = 0. interrupt: ncr810Intr: state 1 -> 2 0x1ef7868 (tScsiTask): ncr810ThreadStateSet: thread 0x01fff990: 1 -> 6 interrupt: ncr810EventTypeGet:Interrupt status istat = 0x02, SCSI status0 = 0x00, SCSI status1 = 0x05DMA status = 0x00 interrupt: ncr810EventTypeGet: SCSI bus timeout. interrupt: ncr810Intr: DSA = 0x01ef8a88, DSP = 0x001d2d10, type = 51. interrupt: ncr810Intr: state 2 -> 1 0x1ef7868 (tScsiTask): ncr810Event: received event 51 (thread = 0x01fff990) logTask: 4 log messages lost. 0x1ef7868 (tScsiTask): ncr810ThreadEvent: thread 0x01fff990: received event 51 0x1ef7868 (tScsiTask): ncr810InitEvent: thread 0x01fff990: select timeout. 0x1ef7868 (tScsiTask): ncr810ThreadFail: thread 0x01fff990 failed (errno = 3604490) 0x1ef7868 (tScsiTask): ncr810ThreadStateSet: thread 0x01fff990: 6 -> 0 0x1ef7868 (tScsiTask): scsiMgrThreadComplete: thread 0x01fff990 0x1ef7868 (tScsiTask): scsiMgrReplyProc: client reply: 0 0x1ef7868 (tScsiTask): scsiMgrCompleteReply: thread 0x01fff990 (status: -1, errno: 3604490) 0x1c59950 (t8): scsiCommand: thread execution failed 0x1c59950 (t8): scsiTransact: scsiCommand ERROR. 0x1c59950 (t8): scsiPhysDevCreate: REQUEST SENSE failed. 0x1c59950 (t8): scsiAutoConfig: bus ID = 1, LUN = 0 ... Regards, Julia --------------------------- Newsgroups: comp.os.vxworks Subject: re: BSP questions- int vector and remap Date: Mon, 11 Jun 2001 10:02:51 +0100 (BST) From: David Laight Organization: Lawrence Berkeley National Laboratory Message-ID: <200106110902.KAA23209@dsl-2.tadpole.co.uk> > > 1. Is there any sample code showing the RE-MAPPING procedure illustrating > > what one should prepare for remapping? > You could found a boot example with a remapping example in the document: > "Implementation of an universal Boot Monitor for an ARM-based System" > at this address: > http://mikro.ee.tu-berlin.de/~hellmich/theses.html > If you have others examples of remapping please send to me. > > 2. Also if ROM address starts from 0x0, what software approach can I do in > > VxWorks to made RAM appear in 0x0 instead of ROM? The vxWorks book talks of a 'walk of death' where you use an intermediate set of page tables with the ROM mapped at 0 and a higher address in order to get RAM mapped to address zero. I use a much simpler approach on our SA1100 system. The strongarm does not fully decode the low physical addresses - even if it did your board hardware probably does not. This means the ROM (at physical address 0) is aliased by the hardware to 64Mb. All of the ROM code, and any vxWorks code is built assuming the ROM starts at 64Mb. An early instruction of the ROM is: start: ldr pc,L__entry /* Jump to expected alias of this code */ entry: ... L__entry: .long ROM_TEXT_ADRS + entry - _romInit This allows me to enable the mmu with DRAM at virtual address 0 and the ROM address unchanged - much simpler. There is, of course, only 64MB of contiguous address space available for DRAM. 'Warm' boots (including the one done by the network loader) cause grief because you can end up re-initialising the MMU while running from page tables already in memory. If you are unlucky the old page tables are in memory the new build is using for something else! even initialising then in the same place is fraught with sequencing issues. All made slightly more complicated because oage table entries are cached in the TLB. David - ---------------------------------------------------------------- David Laight email: dsl@tadpole.co.uk Tadpole Technology plc phone: +44 1223 428 232 Cambridge, UK fax: +44 1223 428 201 --------------------------- Newsgroups: comp.os.vxworks Subject: Command 'iosFdShow' does not work in Shell Date: Mon, 11 Jun 2001 10:53:08 GMT From: "Asaf Sinai" Organization: Telrad Networks Message-ID: Reply-To: "Asaf Sinai" Sender: news@cpm.elex.co.il (NEWS SERVER CPM-SOLARIS) Hi, When I run the command 'iosFdShow' in the Shell, I get: - -> iosFdShow fd name drv value = 0 = 0x0 But, in the serial console I get all the data. Has anyone know why it happens ? --------------------------- Newsgroups: comp.os.vxworks Subject: re: Proper way to unmount dosFs Date: Mon, 11 Jun 2001 10:18:33 +0100 (BST) From: David Laight Organization: Lawrence Berkeley National Laboratory Message-ID: <200106110918.KAA23216@dsl-2.tadpole.co.uk> > To unmount dosFs that was mounted on an ATA disk > at "/ata", I'm doing: > > char* pTail; > iosDevDelete(iosDevFind("/ata", &pTail)); > Is that all that's needed for the unmount? No! All iosDevdelete() does is remove the entry from the list of device drivers known to the system. The device driver isn't told - or allowed to reject this request. Moreover if you have "/ata" open (as opposed to "/atafred") the file data area has a pointer to the "/ata" string in the ios device list. If you do an iosDevAdd for "/ata" before the fd is closed, then close the fd the system will free the string - - causing a memory heap error! The nearest you can do to unmount a dos filesystem is to call all the volume changed rountines - however they do not free all the memory used by dosfs (eg the cache of the FAT and a tree copy of the entire directory structure). I have been told that there is no 'unmount' routine because: 1) there is no requirement 2) it is too hard to do it properly Makes supporting PCMCIA disks in vxworks a bit lively! WRS support may give you some random source which will free up the memory held by dosfs...... David - ---------------------------------------------------------------- David Laight email: dsl@tadpole.co.uk Tadpole Technology plc phone: +44 1223 428 232 Cambridge, UK fax: +44 1223 428 201 --------------------------- Newsgroups: comp.os.vxworks Subject: re: how to check whether fpp register contents is a number? Date: Mon, 11 Jun 2001 10:27:23 +0100 (BST) From: David Laight Organization: Lawrence Berkeley National Laboratory Message-ID: <200106110927.KAA23220@dsl-2.tadpole.co.uk> > > We are using vxworks 5.3.1 with PowerPC750 board. > with fppTaskRegsShow we get the following output: > fppTaskRegsShow 0x3b89d10 > > fpcsr = 70 > fr0 = NaN fr1 = NaN fr2 = NaN fr3 > = NaN > > We would like to be able to perform the test on the registers for NaN > (Not a Number). > We were unable to trackdown the functioncall to do this. > On Solaris the function isnan() is available to perform this. > Who can tell us which function we should use on vxworks or with what > code we perform the isnan test? IF your FP routines are properly conformant then any of the following should work: 1) if (x != x) return NAN 2) sprint( buf, "%f", x ); if (!strcmp( buf, "NaN" ) return NAN 3) if (((int *)&x)[0] & 0x7ff00000 == 0x7ff00000 && ( ((int *)&x)[0] & 0xfffff || ((int *)&x)[1] )) return NAN (Third example is for double, the [0] and [1] may need reversing depending on your FP number endianness) David - ---------------------------------------------------------------- David Laight email: dsl@tadpole.co.uk Tadpole Technology plc phone: +44 1223 428 232 Cambridge, UK fax: +44 1223 428 201 --------------------------- Newsgroups: comp.os.vxworks Subject: re: Redirecting stdout problems Date: Mon, 11 Jun 2001 10:49:28 +0100 (BST) From: David Laight Organization: Lawrence Berkeley National Laboratory Message-ID: <200106110949.KAA23228@dsl-2.tadpole.co.uk> > I am having problems redirecting stdout. I am trying to provide > the ability for our the CLI in our application to call certain > VxWorks functions such as i(), ifShow(), and ti() so that they > may be executed via our CLI through telnet (not the Tornado > shell). Firstly I would run these functions from a separate task whose stdout and stderr are redirected to you pipe - rather than changing the global descriptors. Actually why not make your 'telnet' server have a device driver interface - so it picks up data to transmit from write() calls. Then you have no more work to do! This is vxWorks, there is no user-kernel boundary to worry about, you can write data to a socket from within another device drivers write routine. (I would be a little careful about open(), but the data area locking in vxWorks seems (IMHO) so sloppy you will not hit a deadlock). Even a trivial pipe() device driver is easy to write. - - open( "/mypipe" ) gives one side - - ioctl( fd, MYP_SETNAME, "name" ) gives it a name - - open( "/mypipe/name ) will open the other side - - read blocks if no one writing - - write blocks if no one reading - - transfer data straight from user write buffer to user read buffer. - - maybe optionally push read data through ttyLib David - ---------------------------------------------------------------- David Laight email: dsl@tadpole.co.uk Tadpole Technology plc phone: +44 1223 428 232 Cambridge, UK fax: +44 1223 428 201 --------------------------- Newsgroups: comp.os.vxworks Subject: Problem with stack trace and targetshell Date: 11 Jun 2001 03:28:02 -0700 From: anja_krause@hotmail.com (A. Krause) Organization: http://groups.google.com/ Message-ID: Hello together, I have a problem. When I include dbgLib to my vxWorks...the TargetShell will included automatically and this produces an error by starting the TargetServer later. Do anyone know how I can get the stack trace without any call to tt() ??? I have to include the dbgLib only for this function. Is there a way to use only the taskLib and get the stack trace from the task descriptor... Our problem is that we want to find out the reason when a task crashes during our product field trials. Does anyone know or have an example how you can do this ???? I would appreciate each response... Thank you in advance Anja --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Mon Jun 11 04:06:17 2001 From: "AMASS" Date: Mon Jun 11 04:06:19 PDT 2001 Subject: Calculating required stack size This is a multi-part message in MIME format. ------=_NextPart_000_003E_01C0F26E.734993A0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Hi All In addition to automatic variables, arguments, recursive algorithms and = the like, what else must be taken into consideration when calculating = the requried stack size for a task? Is it best practice to allow a = safety margin when allocating stack sizes? Does anyone have any typical = figures for this if it is recommended? Cheers Andy vxworks ------=_NextPart_000_003E_01C0F26E.734993A0 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable
Hi All
 
In addition to automatic variables, = arguments,=20 recursive algorithms and the like, what else must be taken into = consideration=20 when calculating the requried stack size for a task? Is it best practice = to=20 allow a safety margin when allocating stack sizes? Does anyone have any = typical=20 figures for this if it is recommended?
 
Cheers
 
Andy
 
vxworks
------=_NextPart_000_003E_01C0F26E.734993A0-- From vxwexplo-errs@csg.lbl.gov Mon Jun 11 07:56:31 2001 From: David Laight Date: Mon Jun 11 07:56:33 PDT 2001 Subject: Re: Calculating required stack size > In addition to automatic variables, arguments, recursive algorithms and = > the like, what else must be taken into consideration when calculating = > the requried stack size for a task? Is it best practice to allow a = > safety margin when allocating stack sizes? Does anyone have any typical = > figures for this if it is recommended? The biggest problem is working out how much space any vxWorks library routines require. For some routines this is considerable, and may depend on the local configuration. For instance, netDrv open() allocates several buffers of size BUFSIZ (for copies of the filename - without checking the filename will fit), if it has to use DNS (to resolve the host name) another buffer or two is required. This add up to several kb of stack! An fprintf() call will also eat stack! There are too many jump tables in vxWorks for simple static analysis. Unfortunately the usual method - give it loads of stack then steal some back when we think we know how much it uses - is very error prone because the limits of stack use are likely to be in error reporting and/or error recover code, which can be very difficult to enter from the required places. On some systems the interrupts also use task stack.... David ---------------------------------------------------------------- David Laight email: dsl@tadpole.co.uk Tadpole Technology plc phone: +44 1223 428 232 Cambridge, UK fax: +44 1223 428 201 From vxwexplo-errs@csg.lbl.gov Mon Jun 11 10:22:54 2001 From: "Harry Stox" Date: Mon Jun 11 10:22:56 PDT 2001 Subject: Intel 82559ER driver on Tornado 2.0.2 - solution Hello, Several people replied to my question for an Intel 82559ER driver. Many thanks to all of them. Dave Golden from Compaq was the first to suggest using the native WRS driver since the only necessary changes are a different device ID (0x1209) and a larger address EEPROM. (The size of the EEPROM can be determined by checking for the dummy zero bit on the DO-line when clocking out the address of an arbitrary read location.) Using this simple modification, I was able to get the driver up and running. A later version of the Intel-supplied driver, as can be found on their developer site (ftp://download.intel.com/design/network/drivers/82559erdrivers17.zip) did not work on our hardware. -Harry Stox. From vxwexplo-errs@csg.lbl.gov Mon Jun 11 21:59:56 2001 From: "Rajendra Prasad" Date: Mon Jun 11 21:59:58 PDT 2001 Subject: Application code entry point vxworks tornado Hi, Could somebody explain me the procedure to set the application code entry point for the vxworks image to be booted in a network and non-network boot environments ? Any relevant manual references will be greatly appreciated. Thanks and best regards, rajendra. _________________________________________________________________________ Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com. From vxwexplo-errs@csg.lbl.gov Tue Jun 12 04:03:28 2001 From: Vxworks Exploder Date: Tue Jun 12 04:03:30 PDT 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Tue Jun 12 04:03:23 PDT 2001 Subject: PPC:clockLib problems Subject: Calculating required stack size Subject: Re: New Altivec-enhanced VxWorks Gcc 2.95.4 compiler released! Subject: Watchdog Subject: Re: Calculating required stack size Subject: Re: Problem with stack trace and targetshell Subject: Re: Tundra Universe IIB Weirdness Subject: Application code entry point Subject: Re: Tundra Universe IIB Weirdness Subject: Re: Proper way to unmount dosFs Subject: Re: Help for init. 53C875 SCSI controller Subject: re: Calculating required stack size Subject: Corba over PPP problem Subject: Re: Calculating required stack size Subject: Re: problem with inet_aton Subject: Re: Filenames 8.3 Subject: Re: How to build a bootline in NvRam Subject: Intel 82559ER driver on Tornado 2.0.2 - solution Subject: VX works expert required Subject: objdump error Subject: ATA Device Driver Error Codes Subject: Re: PPC: system tick routine stack usage Subject: Re: Proper way to unmount dosFs Subject: Re: PPC: system tick routine stack usage Subject: Application code entry point Subject: Re: BSP Component Removal Subject: PLX 9054 (and vxWorks) help Subject: comparision on realtime systems Subject: Extending the WindRiver (TMS) Envoy SNMP agent Subject: Re: Application code entry point Subject: Change MAC address dynamically in PPC8260 ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: PPC:clockLib problems Date: 11 Jun 2001 04:11:27 -0700 From: mvi76@yahoo.com (Martin Ivanov) Organization: http://groups.google.com/ Message-ID: <5f86e92.0106110311.69d7b3f0@posting.google.com> Hi VxWorkers! I read few postings, describing different problems with clockLib - timer gets reset periodically, going backwards and so on. I have basicly the same problem - time() increments to some seconds, gets reset to zero, and after a that while again the same. I tried to debug, windTickAnnounce, incrementing the vxTicks, but problem seems to be complicated. Can you give me some information on possible clockLib bugs? I'm using Tornado 2 for PPC 405. Thanx. --------------------------- Newsgroups: comp.os.vxworks Subject: Calculating required stack size Date: Mon, 11 Jun 2001 12:02:56 +0100 From: "AMASS" Organization: Lawrence Berkeley National Laboratory Message-ID: <9g2985$jsm$1@overload.lbl.gov> This is a multi-part message in MIME format. - ------=_NextPart_000_003E_01C0F26E.734993A0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Hi All In addition to automatic variables, arguments, recursive algorithms and = the like, what else must be taken into consideration when calculating = the requried stack size for a task? Is it best practice to allow a = safety margin when allocating stack sizes? Does anyone have any typical = figures for this if it is recommended? Cheers Andy vxworks - ------=_NextPart_000_003E_01C0F26E.734993A0 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable
Hi All
 
In addition to automatic variables, = arguments,=20 recursive algorithms and the like, what else must be taken into = consideration=20 when calculating the requried stack size for a task? Is it best practice = to=20 allow a safety margin when allocating stack sizes? Does anyone have any = typical=20 figures for this if it is recommended?
 
Cheers
 
Andy
 
vxworks
- ------=_NextPart_000_003E_01C0F26E.734993A0-- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: New Altivec-enhanced VxWorks Gcc 2.95.4 compiler released! Date: Mon, 11 Jun 2001 13:47:41 +0200 From: Mathieu Imrazene Organization: Nortel Networks Message-ID: <9g2b4t$91q$1@qnsgh006.europe.nortel.com> References: <5btU6.24100$YB3.4024200@news2-win.server.ntlworld.com> > Hi everybody, > > I've just finished uploading a new compiler on my website that allows > you > to write Altivec programs for VxWorks. It seems WRS have one on the way > as well but if you want something in the meantime or don't have a > maintenance/update agreement with WRS, you might find it helpful. > > http://www.newgcc4vxworks4ppc.cjb.net/ > > DaveK > Great job, as always ! Does anyone has examples of performance improvements by using the Altivec compiler ? In others words, it is really useful to upgrade from the 2.95.3 version to the 2.95.4 + Altivec version ? Mathieu --------------------------- Newsgroups: comp.os.vxworks Subject: Watchdog Date: Mon, 11 Jun 2001 14:01:39 +0200 From: "Gunter Ritter" Organization: Tenovis GmbH u. Co. KG Message-ID: <9g2bpo$rik$1@proxy2.fe.internet.bosch.com> Hello, I want to realize a watchdog in a system running with VxWorks. The watchdog has to detect crashing tasks and deadlocks and has to reset the system if something is going wrong. I haven´t found much information the documentation except the watchdog timers. So I try to ask the audience (this is my last joker ;-) ). I have found out that a possible criteria for a crashed task is the "valid" value in the EXC_INFO structure of the TCB. If this value is not zero, an exception occured and there is something wrong with this task. Checking this value at every task switch (I use the taskSwitchHook functionality) does work. The problem I couldn´t solve is detecting a deadlock. One solution would be to start a watchdog timer at the beginning of a task function and stopping it before leaving the task function. But with this solution I have to modify all task functions. What I need is a solution working for all running tasks without changing the task functions because some components iin the system are only binary code. Does anybody has a hint for me. Thanks Gunter --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Calculating required stack size Date: Mon, 11 Jun 2001 18:00:30 +0100 From: Vijay Kumar Peshkar Organization: Cisco Systems Inc. Message-ID: <3B24F92D.784104CB@wipro.com> References: <9g2985$jsm$1@overload.lbl.gov> This is a multi-part message in MIME format. - --------------A2567FD6931857B2652A02CE Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hi Andy, Yes. Its a good practice to maintain a safety margin. You can also consider.... 1.Some architectures, do not permit using a separate interrupt stack. On such architectures, ISRs use the stack of the interrupted task. If you have such an architecture, you must create tasks with enough stack space to handle the worst possible combination of nested interrupts and the worst possible combination of ordinary nested calls. 2.If you are using stuff like NFS you will have to add few extra kbs(~10 ?). Generally there's no set of 'checklist' we can refer. Trial and error is the best way even after accounting for all things we know. Use the checkStack() facility during development to see how close your tasks have come to exhausting the available stack space. Hope this helps. Have fun, Vijay AMASS wrote: > This is a multi-part message in MIME format. > > ------=_NextPart_000_003E_01C0F26E.734993A0 > Content-Type: text/plain; > charset="iso-8859-1" > Content-Transfer-Encoding: quoted-printable > > Hi All > > In addition to automatic variables, arguments, recursive algorithms and = > the like, what else must be taken into consideration when calculating = > the requried stack size for a task? Is it best practice to allow a = > safety margin when allocating stack sizes? Does anyone have any typical = > figures for this if it is recommended? > > Cheers > > Andy > > vxworks > > ------=_NextPart_000_003E_01C0F26E.734993A0 > Content-Type: text/html; > charset="iso-8859-1" > Content-Transfer-Encoding: quoted-printable > > > > charset=3Diso-8859-1"> > > > > >
>
Hi All
>
 
>
In addition to automatic variables, = > arguments,=20 > recursive algorithms and the like, what else must be taken into = > consideration=20 > when calculating the requried stack size for a task? Is it best practice = > to=20 > allow a safety margin when allocating stack sizes? Does anyone have any = > typical=20 > figures for this if it is recommended?
>
 
>
Cheers
>
 
>
Andy
>
 
>
vxworks
> > ------=_NextPart_000_003E_01C0F26E.734993A0-- - --------------A2567FD6931857B2652A02CE Content-Type: text/x-vcard; charset=us-ascii; name="vijay.peshkar.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Vijay Kumar Peshkar Content-Disposition: attachment; filename="vijay.peshkar.vcf" begin:vcard n:Peshkar;Vijay Kumar tel;work:5530034/35/35 x2237 x-mozilla-html:FALSE org:Wipro Global R & D;Wishing you blue skies, a couple of daisies, some fresh air and a smile! adr:;;;;;; version:2.1 email;internet:vpeshkar@cisco.com note:Wishing you blue skies, a couple of daisies, some fresh air and a smile! fn:Vijay Kumar Peshkar end:vcard - --------------A2567FD6931857B2652A02CE-- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Problem with stack trace and targetshell Date: Mon, 11 Jun 2001 18:11:19 +0100 From: Vijay Kumar Peshkar Organization: Cisco Systems Inc. Message-ID: <3B24FBB7.B4A9A8C0@wipro.com> References: This is a multi-part message in MIME format. - --------------39D3087D87D245B72A9C427E Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Anja, A question, how do you test it on field. Do you run your application (say main() ) from target shell as ->main or you spawn a task and run as -> sp main ? In earlier case the "main" runs in target shell task's context essentially using its stack while in second case it runs as a separate task and uses a new stack. I suspect case 1.... Good luck. cheers, Vijay "A. Krause" wrote: > Hello together, > > I have a problem. When I include dbgLib to my vxWorks...the > TargetShell will included automatically and this produces an error by > starting the TargetServer later. > Do anyone know how I can get the stack trace without any call to tt() > ??? > I have to include the dbgLib only for this function. Is there a way to > use only the taskLib and get the stack trace from the task > descriptor... > > Our problem is that we want to find out the reason when a task crashes > during our product field trials. Does anyone know or have an example > how you can do this ???? > > I would appreciate each response... > Thank you in advance > > Anja - --------------39D3087D87D245B72A9C427E Content-Type: text/x-vcard; charset=us-ascii; name="vijay.peshkar.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Vijay Kumar Peshkar Content-Disposition: attachment; filename="vijay.peshkar.vcf" begin:vcard n:Peshkar;Vijay Kumar tel;work:5530034/35/35 x2237 x-mozilla-html:FALSE org:Wipro Global R & D;Wishing you blue skies, a couple of daisies, some fresh air and a smile! adr:;;;;;; version:2.1 email;internet:vpeshkar@cisco.com note:Wishing you blue skies, a couple of daisies, some fresh air and a smile! fn:Vijay Kumar Peshkar end:vcard - --------------39D3087D87D245B72A9C427E-- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tundra Universe IIB Weirdness Date: Mon, 11 Jun 2001 06:50:30 -0700 From: Melvin Gardipee <"> Organization: Raytheon Company Message-ID: <724V6.111$Mf6.2079@dfw-service2.ext.raytheon.com> References: <3b1e6510.344858470@200.123.101.100> I had a similar problem with a MVME2305 when trying to write doubles to a VME shared memory board. It turns out that my BSP configures the TundraII to be VME64 compliant. A32-D64 cycles will not work with some VME boards. I edited the include file that configures the LM/SEM registers to limit the data cycles to D32. Although the defines probably have no meaning in your BSP, the file was mv2600.h and the macro define was for VAL_LSI0_CTL through VAL_LSI3_CTL. I changed the parameter LSI_CTL_D64 to LSI_CTL_D32. Regards, Melvin Gardipee cgan76 wrote: > > Howdy, > > My company is currently developing a VME 64x board (MPC8240) that > utilizes the Universe IIB bridge. We are running VxWorks 5.4 as > our operating system. > > Here's our current setup: > > MPC8240 VME 64x Board (any slot): > -MPC8240 CPU running VxWorks 5.4 > -64 MB SDRAM > -64 bit PCI bus > -Universe IIB bridge @ 0x80000000 (Local PCI) > -Slave Window into SDRAM @ 0x70000000 (VME A32)and 0x00000000 (Local > PCI), size is 64 MB. > -All caching is turned off > > System Controller (slot 1): > -SBS Technologies BIT3 Card > -We are also performing reads/writes to the above slave window from > this card. > > We are seeing a strange problem when performing reads/writes to the > VME slave window (A32). Initially, when the system is booted and the > Universe is newly initialized, reads and writes into the A32 space > into MPC8240 SDRAM works properly. For example: > > Device name "\\.\Btw0_A32" > Starting bt_read() test > Attempting read #0, 0x100 bytes > 00000000: f7 df d7 bd f9 7b af bd ea 7d bf d7 fb bb db e3 > 00000010: f3 29 9e db f7 db 5f ff 7f db fd 5d 7f ff bf dd > 00000020: 94 79 fb f5 fb 7f db ff 7f 1e d3 df 7b ff ff fe > data truncated.... > 000000d0: f6 17 ff e9 bf ff f2 fc bf fb 7b d7 b7 df fe ff > 000000e0: 57 fb e7 fb bb bf d7 b7 ff ed 3f 4f 67 76 9f fa > 000000f0: b5 7f bf eb f7 fd a2 f7 ff fe b9 7f b7 fd ba bf > bt_read() #0 success > Exiting bt_read() test > > But after a certain amount of time, a read will be perfromed and the > data read will have every second DWORD returned as zeros. For > example: > > Device name "\\.\Btw0_A32" > Starting bt_read() test > Attempting read #0, 0x100 bytes > 00000000: f7 df d7 bd 00 00 00 00 ea 7d bf d7 00 00 00 00 > 00000010: f3 29 9e db 00 00 00 00 7f db fd 5d 00 00 00 00 > 00000020: 94 79 fb f5 00 00 00 00 7f 1e d3 df 00 00 00 00 > data truncated... > 000000d0: f6 17 ff e9 00 00 00 00 bf fb 7b d7 00 00 00 00 > 000000e0: 57 fb e7 fb 00 00 00 00 ff ed 3f 4f 00 00 00 00 > 000000f0: b5 7f bf eb 00 00 00 00 ff fe b9 7f 00 00 00 00 > bt_read() #0 success > Exiting bt_read() test > > Once in this state, writes will also fail in the same way. i.e. Every > second DWORD will not be written. > > We have found that we can put the Universe into the working state > again by merely READING any one of it's PCI registers. For example we > used the first 32 bit register at (0x80000000) which is just the > device id. > > Again, after doing this read, the Universe will operate for a time > before reverting back to reading zeros for every second DWORD. > > We do not know what could be causing this and your help in this matter > would be greatly appreciated. > > Thanks! > > Christian Gan --------------------------- Newsgroups: comp.os.vxworks Subject: Application code entry point Date: Mon, 11 Jun 2001 19:47:51 +0530 From: rajendra Organization: Lucent Technologies Message-ID: <3B24D30F.84DBDB71@lucent.com> vxworks tornado Hi, Could somebody explain me the procedure to set the application code entry point for the vxworks image to be booted in a network and non-network boot environments ? Any relevant manual references will be greatly appreciated. Thanks and best regards, rajendra. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tundra Universe IIB Weirdness Date: Mon, 11 Jun 2001 14:36:06 GMT From: cgan76@nospam4me.com (cgan76) Organization: MTS Internet Message-ID: <3b24d4cc.766660199@200.123.101.100> References: <3b1e6510.344858470@200.123.101.100> <724V6.111$Mf6.2079@dfw-service2.ext.raytheon.com> Hi Melvin, Thanks for your response. We actually found what our problem was and it does have to do with the data width of the bus that the Universe II expects. Specifically, our ACK64 line in PCI did not have a pull up resistor in place for our design. Since it floats it sometimes looks like a logic high and sometimes a logic low. When not in use, the signal should always be logic high. So in some cases, the ACK64 line looked like a logic low to the Universe and therefore it expected 64 bit wide data to come in from the PCI bus, but our MPC840 CPU was only sending 32 (that's why every second DWORD was zero). I still find it somewhat strange that the Universe chip will respond to the ACK64 line even though it did not request 64 bit data in the first place (with REQ64). It fooled us until we actually put our logic analyzers to the bus. Thanks! On Mon, 11 Jun 2001 06:50:30 -0700, Melvin Gardipee <"> wrote: > I had a similar problem with a MVME2305 when trying to write >doubles to a VME shared memory board. It turns out that my BSP >configures the TundraII to be VME64 compliant. A32-D64 cycles will not >work with some VME boards. I edited the include file that configures >the LM/SEM registers to limit the data cycles to D32. > Although the defines probably have no meaning in your BSP, the >file was mv2600.h and the macro define was for VAL_LSI0_CTL through >VAL_LSI3_CTL. I changed the parameter LSI_CTL_D64 to LSI_CTL_D32. > >Regards, >Melvin Gardipee > > >cgan76 wrote: >> >> Howdy, >> >> My company is currently developing a VME 64x board (MPC8240) that >> utilizes the Universe IIB bridge. We are running VxWorks 5.4 as >> our operating system. >> >> Here's our current setup: >> >> MPC8240 VME 64x Board (any slot): >> -MPC8240 CPU running VxWorks 5.4 >> -64 MB SDRAM >> -64 bit PCI bus >> -Universe IIB bridge @ 0x80000000 (Local PCI) >> -Slave Window into SDRAM @ 0x70000000 (VME A32)and 0x00000000 (Local >> PCI), size is 64 MB. >> -All caching is turned off >> >> System Controller (slot 1): >> -SBS Technologies BIT3 Card >> -We are also performing reads/writes to the above slave window from >> this card. >> >> We are seeing a strange problem when performing reads/writes to the >> VME slave window (A32). Initially, when the system is booted and the >> Universe is newly initialized, reads and writes into the A32 space >> into MPC8240 SDRAM works properly. For example: >> >> Device name "\\.\Btw0_A32" >> Starting bt_read() test >> Attempting read #0, 0x100 bytes >> 00000000: f7 df d7 bd f9 7b af bd ea 7d bf d7 fb bb db e3 >> 00000010: f3 29 9e db f7 db 5f ff 7f db fd 5d 7f ff bf dd >> 00000020: 94 79 fb f5 fb 7f db ff 7f 1e d3 df 7b ff ff fe >> data truncated.... >> 000000d0: f6 17 ff e9 bf ff f2 fc bf fb 7b d7 b7 df fe ff >> 000000e0: 57 fb e7 fb bb bf d7 b7 ff ed 3f 4f 67 76 9f fa >> 000000f0: b5 7f bf eb f7 fd a2 f7 ff fe b9 7f b7 fd ba bf >> bt_read() #0 success >> Exiting bt_read() test >> >> But after a certain amount of time, a read will be perfromed and the >> data read will have every second DWORD returned as zeros. For >> example: >> >> Device name "\\.\Btw0_A32" >> Starting bt_read() test >> Attempting read #0, 0x100 bytes >> 00000000: f7 df d7 bd 00 00 00 00 ea 7d bf d7 00 00 00 00 >> 00000010: f3 29 9e db 00 00 00 00 7f db fd 5d 00 00 00 00 >> 00000020: 94 79 fb f5 00 00 00 00 7f 1e d3 df 00 00 00 00 >> data truncated... >> 000000d0: f6 17 ff e9 00 00 00 00 bf fb 7b d7 00 00 00 00 >> 000000e0: 57 fb e7 fb 00 00 00 00 ff ed 3f 4f 00 00 00 00 >> 000000f0: b5 7f bf eb 00 00 00 00 ff fe b9 7f 00 00 00 00 >> bt_read() #0 success >> Exiting bt_read() test >> >> Once in this state, writes will also fail in the same way. i.e. Every >> second DWORD will not be written. >> >> We have found that we can put the Universe into the working state >> again by merely READING any one of it's PCI registers. For example we >> used the first 32 bit register at (0x80000000) which is just the >> device id. >> >> Again, after doing this read, the Universe will operate for a time >> before reverting back to reading zeros for every second DWORD. >> >> We do not know what could be causing this and your help in this matter >> would be greatly appreciated. >> >> Thanks! >> >> Christian Gan --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Proper way to unmount dosFs Date: 11 Jun 2001 07:34:36 -0700 From: trahn@macrosystem.de (Thomas Rahn) Organization: http://groups.google.com/ Message-ID: <30c8bd8b.0106110634.467b83a1@posting.google.com> References: <7f10f842.0106091624.70f230be@posting.google.com> Hi, > > To unmount dosFs that was mounted on an ATA disk > at "/ata", I'm doing: > > char* pTail; > iosDevDelete(iosDevFind("/ata", &pTail)); > > Is that all that's needed for the unmount? > > Thanks, > Hari Your code will only delete the uppermost layer of the block devices / filesystem. If you are using dosFs 2.0 with the CBIO layer, there will be an intermediate driver. That one is not deleted. In any case, the low level device created with a call to ataDevCreate() is not freed. Unfortunately, there is no way with vxwWorks to remove a block device within a running system. You will have to add some patch code to find device headers and unlink them from the device chain... AFAIK, I will be glad to hear anything else. Thomas Rahn --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Help for init. 53C875 SCSI controller Date: 11 Jun 2001 07:35:50 -0700 From: trahn@macrosystem.de (Thomas Rahn) Organization: http://groups.google.com/ Message-ID: <30c8bd8b.0106110635.b6d69dd@posting.google.com> References: <7f208981.0106100351.2b2a5849@posting.google.com> Hi, at our site, the ncr810 driver is working on a x86 target... There had been several obstacles since our first attempts: - - a NULL pointer access - - hard coded pci address range - - missing define for 875 - - offset 0x8000 0000 to pci memory range - - ... From your mail i guess WRS has fixed some of those in the current driver package. If a ncr810Show() displays some valid register entries, that part should be working. However, there might be left a bug in the ncr810 driver (and sym895 as well) regarding wide/sync negotiation. I never convinced WRS about it. You will need the driver source to fix it. Try to spawn the scsi detection as a low-priority task. If the "tScsiTask" stays ready, you might get on if you disable all non-basic scsi features for the target. [using scsiTargetOptionsSet() or similar]. Or: there might be a jumper on the drive to disable any such feature at the drive. Thomas Rahn --------------------------- Newsgroups: comp.os.vxworks Subject: re: Calculating required stack size Date: Mon, 11 Jun 2001 15:55:45 +0100 (BST) From: David Laight Organization: Lawrence Berkeley National Laboratory Message-ID: <200106111455.PAA23545@dsl-2.tadpole.co.uk> > In addition to automatic variables, arguments, recursive algorithms and = > the like, what else must be taken into consideration when calculating = > the requried stack size for a task? Is it best practice to allow a = > safety margin when allocating stack sizes? Does anyone have any typical = > figures for this if it is recommended? The biggest problem is working out how much space any vxWorks library routines require. For some routines this is considerable, and may depend on the local configuration. For instance, netDrv open() allocates several buffers of size BUFSIZ (for copies of the filename - without checking the filename will fit), if it has to use DNS (to resolve the host name) another buffer or two is required. This add up to several kb of stack! An fprintf() call will also eat stack! There are too many jump tables in vxWorks for simple static analysis. Unfortunately the usual method - give it loads of stack then steal some back when we think we know how much it uses - is very error prone because the limits of stack use are likely to be in error reporting and/or error recover code, which can be very difficult to enter from the required places. On some systems the interrupts also use task stack.... David - ---------------------------------------------------------------- David Laight email: dsl@tadpole.co.uk Tadpole Technology plc phone: +44 1223 428 232 Cambridge, UK fax: +44 1223 428 201 --------------------------- Newsgroups: comp.os.vxworks Subject: Corba over PPP problem Date: Mon, 11 Jun 2001 15:24:40 GMT From: "Carl Wolff" Message-ID: <01c0f28d$5889a4a0$3a75eed4@default> Hello Got the following strange problem: I've got an VxWorks system connected to a NT system using a PPP link over a nullmodem cable. Corba communicates to the NT system using this link. When the 'corba things' are started the VxWorks registers the object in the NT broker (messages from the VxWorks system arrive at the broker). Until sofar the link seems OK. But when the broker sends a reply the reply is send to the client with a wrong IP adress. When this is tried using a ethernet link the problem mentioned above is not present, and everything works OK. Questions 1) Can anybody give me a guideline how to solve this problem, and where to begin? 2) Where is the wronh IP adress coming from? Thanks Carl. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Calculating required stack size Date: 11 Jun 2001 12:26:04 -0400 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: References: <9g2985$jsm$1@overload.lbl.gov> Sender: bpringlemeir@DeadDuck Andy> In addition to automatic variables, arguments, recursive Andy> algorithms and = the like, what else must be taken into Andy> consideration when calculating = the requried stack size for a Andy> task? Is it best practice to allow a = safety margin when Andy> allocating stack sizes? Does anyone have any typical = figures Andy> for this if it is recommended? You can calculate the stack usage exactly. The functions must be in a `DAG' structure. Ie, there are no loops or cycles in the function call hierarchy. Gprof has similar restriction for calculating function execution times... You can not use things like alloca() and dynamic array sizes. Also, interrupt stack usage must be bound. It seems that this is too tough for most systems. There were discussions on the Linux kernel list and `comp.arch.embedded' as well. I think that David Laight gave some pragmatic advice in this thread. The major factor with vxWorks is that the source is closed and you can not calculate what amount of stack the library routines use. Most people seem to use `water marks' and then multiply by some safety factor. Watermarks or colouring are when you initialize the stack to some value, like `0x55aa55aa', and then observe what values were over-written after running the task/thread through a test suite. hth, Bill Pringlemeir. - -- Clothing is not flowers. Useful vxWorks URLS, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" "http://www.xs4all.nl/~borkhuis/vxworks/vxw_pt9.html" --------------------------- Newsgroups: comp.os.vxworks Subject: Re: problem with inet_aton Date: Mon, 11 Jun 2001 16:42:47 GMT From: "Michael Yeung" Organization: Magma Communications Ltd. Message-ID: References: <3B24981B.2A9CF1F5@wipro.com> Reply-To: "Michael Yeung" Vijay, Thanx for the info. rtrequest is a function support in BSD to add new route. Rather than using routeAdd from VxWorks. Michael Vijay Kumar Peshkar wrote in message news:3B24981B.2A9CF1F5@wipro.com... > Michael, > > The macro OK has value 0 and in your case as everything seems to be working > fine, the func. call is returning OK ;). You should be comparing it with ERROR(-1). > > The inet_addr func. is supported in vxWorks. > Sorry I could not understand what's "rtrequest". > > cheers, > Vijay > > Michael Yeung wrote: > > > Hi all, > > I'm not sure why but the inet_aton seems to always returns 0 and > > conversion is not done properly. > > > > struct in_addr temp; > > > > if(inet_aton("1.1.1.1", (struct in_addr *)&temp) == 0) > > { > > printf("Failed conversion\n"); > > } > > > > I noticed VxWorks does not support "inet_addr" or "inet_pton", as > > suggestion is appreciated. > > > > As well, does anyone knows the format expected for mask to function like > > "rtrequest"? > > > > thanx in advance, > > Michael > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Filenames 8.3 Date: Mon, 11 Jun 2001 16:40:29 GMT From: Gary Organization: The Boeing Company Message-ID: <3B24F47D.C675C14C@boeing.com> References: <3B213694.307B5FC3@boeing.com> Sender: nntp@news.boeing.com (Boeing NNTP News Access) Johan, I do create the file on a server, but if I try to create the file on the "C:" drive with a long directory name, it will also fail. Johan Borkhuis wrote: > > "Gary G." wrote: > > > Is it true that the filename/directory paths must be 8.3 ? I am trying > > to create my projects and workspace on a server that has a path that is > > NOT 8.3 and it fails. If I use C:\<8.3>\<8.3> everything works. > > I never had any problems with this. The only thing you should be carefull > with is using spaces in names. Do you create it on a remote server? Do any > other programs have problems with this? > > Groeten, > Johan > > -- > o o o o o o o . . . _____________________________ > o _____ || Johan Borkhuis | > .][__n_n_|DD[ ====_____ | borkhuis@agere.com | > >(________|__|_[_________]_|__________________________| > _/oo OOOOO oo` ooo ooo 'o!o!o o!o!o` > === VxWorks FAQ: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html === --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to build a bootline in NvRam Date: 11 Jun 2001 10:34:40 -0700 From: aparker@vtlinx.com (parreg) Organization: http://groups.google.com/ Message-ID: <79ef8b60.0106110934.24835e27@posting.google.com> References: <9fl95j$3lj$1@news.seed.net.tw> "John Lo" wrote in message news:<9fl95j$3lj$1@news.seed.net.tw>... > Could anybody tell me how to build a bootline in NvRam > Thank you Could you be a little more specific. The bootline is a character string, the same as when you type "c" at the vxWorks bootline prompt. This string gets stored in NVRAM usually EEPROM. The vxWorks commands sysNvRamGet sysNvRamSet are responsible interacting with the NVRAM if it set up corrrectly. parreg --------------------------- Newsgroups: comp.os.vxworks Subject: Intel 82559ER driver on Tornado 2.0.2 - solution Date: Mon, 11 Jun 2001 19:18:45 +0200 From: "Harry Stox" Organization: Lawrence Berkeley National Laboratory Message-ID: <9g31rl$6et$1@overload.lbl.gov> Hello, Several people replied to my question for an Intel 82559ER driver. Many thanks to all of them. Dave Golden from Compaq was the first to suggest using the native WRS driver since the only necessary changes are a different device ID (0x1209) and a larger address EEPROM. (The size of the EEPROM can be determined by checking for the dummy zero bit on the DO-line when clocking out the address of an arbitrary read location.) Using this simple modification, I was able to get the driver up and running. A later version of the Intel-supplied driver, as can be found on their developer site (ftp://download.intel.com/design/network/drivers/82559erdrivers17.zip) did not work on our hardware. - -Harry Stox. --------------------------- Newsgroups: comp.os.vxworks Subject: VX works expert required Date: Mon, 11 Jun 2001 16:35:00 -0400 From: "Brian" Organization: Bell Sympatico Message-ID: <3Q9V6.4962$nY4.912102@news20.bellglobal.com> I am looking for a Sr. VX works developer with previous and recent experience specifically with device driver development. The work would be developing device drivers and associated software for various hardware devices. You must have device driver development experience, the more the better. This would be a full time telecommute position, with full benefits. For more information on this position please contact me directly. For other information please browse our web site. Brian brian@ahiresolution.com www.ahiresolution.com --------------------------- Newsgroups: comp.os.vxworks Subject: objdump error Date: 11 Jun 2001 13:38:07 -0700 From: big_dou@yahoo.com (Amadou Diop) Organization: http://groups.google.com/ Message-ID: List file generation appl does not run properly using the mottorolla (VX_CPU_FAMILY 68k). Here is the statement: objdump -d $*.o > $*.lst and this is the error the I get "missing separator" Has anyone encountered the same problem before. --------------------------- Newsgroups: comp.os.vxworks Subject: ATA Device Driver Error Codes Date: Mon, 11 Jun 2001 21:16:40 GMT From: "Steeve Gagne" Organization: MetroNet Communications Group Inc. Message-ID: Hi ! I am working with ATA Driver of vxWorks. The device driver works fine but report an error on the set feature command : ... ataCmd: ctrl=0 drive=0 cmd=0xef ataWait: ctrl=0 request=0x40 ataWait end: ataCmd err: status=0x51 semStatus=0 err=0x4 ... Is ECC made on data really are the error returned by the ATA controller ?? in "drv/hdisk/atadrv.h" #define ATA_STAT_BUSY 0x80 /* controller busy */ #define ATA_STAT_READY 0x40 /* selected drive ready */ #define ATA_STAT_WRTFLT 0x20 /* write fault */ #define ATA_STAT_SEEKCMPLT 0x10 /* seek complete */ #define ATA_STAT_DRQ 0x08 /* data request */ #define ATA_STAT_ECCCOR 0x04 /* ECC correction made in data */ #define ATA_STAT_INDEX 0x02 /* index pulse from selected drive */ #define ATA_STAT_ERR 0x01 /* error detect */ Here is the printout of the debug information : ataInit: ctrl=0 ataInit Calling sysAtaInit (if present): ataInit sysAtaInit returned: ataInit end: ataCmd: ctrl=0 drive=0 cmd=0x90 ataWait: ctrl=0 request=0x40 ataWait end: ataCmd end: ataCmd: ctrl=0 drive=0 cmd=0x77 ataWait: ctrl=0 request=0x40 ataWait end: ataCmd end: ataPread: ctrl=0 drive=0 ataWait: ctrl=0 request=0x40 ataWait end: ataWait: ctrl=0 request=0x8 ataWait end: ataPread end: ID_DRIVE reports LBA (60-61) as 0x00007a80 ataCmd: ctrl=0 drive=0 cmd=0x91 ataWait: ctrl=0 request=0x40 ataWait end: ataCmd end: ataCmd: ctrl=0 drive=0 cmd=0x10 ataWait: ctrl=0 request=0x40 ataWait end: ataCmd end: ataCmd: ctrl=0 drive=0 cmd=0xef ataWait: ctrl=0 request=0x40 ataWait end: ataCmd err: status=0x51 semStatus=0 err=0x4 ataWait: ctrl=0 request=0x40 ataWait end: ataCmd err: status=0x51 semStatus=0 err=0x4 ataWait: ctrl=0 request=0x40 ataWait end: ataCmd err: status=0x51 semStatus=0 err=0x4 ataWait: ctrl=0 request=0x40 ataWait end: ataCmd err: status=0x51 semStatus=0 err=0x4 ataDrv Calling sysAtaInit (if present): ataDrv sysAtaInit returned: ataRW: ctrl=0 drive=0 c=0 h=0 s=0 buf=0x10b048 n=1 dir=0 ataWait: ctrl=0 request=0x40 ataWait end: ataWait: ctrl=0 request=0x80 ataWait end: ataWait: ctrl=0 request=0x8 ataWait end: ataRW: end Thanks . Steeve Gagne steeve.gagne@videotron.ca --------------------------- Newsgroups: comp.os.vxworks Subject: Re: PPC: system tick routine stack usage Date: Mon, 11 Jun 2001 21:57:18 GMT From: pkockritz@home.com (Pete Kockritz) Organization: Excite@Home - The Leader in Broadband http://home.com/faster Message-ID: References: In article , pkockritz@home.com (Pete Kockritz) wrote: Should've mentioned: T2, vxWorks 5.4, NT hosted, PPC860 target > The basic question is; does the PPC tick handler have its own stack or > does it just use the stack of the task that is active when the decrementer > interrupt goes off? I've looked at the code at 0x500 and 0x900, both call the intEnt/intExit routines. intEnt sets the stack to the interrupt stack... but only if it is not a nested interrupt. It appears to use SPRG1 as a nesting counter. But we've caught it several times using a task's stack area while in a watchdog timer callback. So I'm wondering if somehow SPRG1 is getting changed by something else, or if SPRG1 is not being reset properly - intEnt increments it and intExit decrements it. We probably would've never noticed it if the watchdog callback didn't have a large stack frame. Perhaps it happens occassionally but it is just not noticed? Pete - -- Pete Kockritz - -- Pete Kockritz --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Proper way to unmount dosFs Date: 11 Jun 2001 17:11:02 -0700 From: hariharan@foobox.com (Thiagarajan Hariharan) Organization: http://groups.google.com/ Message-ID: <7f10f842.0106111611.39022d1e@posting.google.com> References: <7f10f842.0106091624.70f230be@posting.google.com> <30c8bd8b.0106110634.467b83a1@posting.google.com> For what it's worth, I added the following before the iosDevDelete(). "/ata" is the mount point for the hard disk. fd = open("/ata", O_RDWR, 0); ioctl(fd, FIOUNMOUNT, 0); Hari --------------------------- Newsgroups: comp.os.vxworks Subject: Re: PPC: system tick routine stack usage Date: 11 Jun 2001 20:01:47 -0700 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0106111901.aba7af1@posting.google.com> References: Hello, > But we've caught it several times using a task's stack area while in a > watchdog timer callback. So I'm wondering if somehow SPRG1 is getting > changed by something else, or if SPRG1 is not being reset properly - > intEnt increments it and intExit decrements it. No, SPRG1 is probably OK. It is just that there are times when the tick announce is deferred (if it happens to go off when something else is already "in the kernel" then it is added to the work queue). In these cases, the work will happen in the scheduling code, and that will be on the bottom of a task stack. I think it will be the "old" task stack (i.e. the one running before the decrementer interrupt). Now, "in the kernel" was in quotes because it means only the brief periods where the kernel is working on critical data structures. Unlike other kernels, the Wind kernel does not lock out interrupts for most of this work; instead it uses a flag to indicate that it is in the kernel's critical region. Any interrupt making calls to kernel routines (e.g. semGive) will execute, but the operation will be queued. When the original routine exits the kernel, the work queue will be drained in order. > We probably would've never noticed it if the watchdog callback didn't have > a large stack frame. Perhaps it happens occassionally but it is just not > noticed? Indeed. As with ISRs though it is best to keep watchdog call backs simple, and not use a lot of resources (in this case stack, but also CPU time etc). Do you really need such a large stack frame? HTH, John... --------------------------- Newsgroups: comp.os.vxworks Subject: Application code entry point Date: Tue, 12 Jun 2001 04:59:49 From: "Rajendra Prasad" Organization: Lawrence Berkeley National Laboratory Message-ID: <9g48hu$ov4$1@overload.lbl.gov> vxworks tornado Hi, Could somebody explain me the procedure to set the application code entry point for the vxworks image to be booted in a network and non-network boot environments ? Any relevant manual references will be greatly appreciated. Thanks and best regards, rajendra. _________________________________________________________________________ Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: BSP Component Removal Date: Tue, 12 Jun 2001 08:37:31 +0100 From: Paul Caswell Organization: Land & Sea Systems Message-ID: <3B25C6BB.DEE1A254@baesystems.com> References: <3B0CE5D5.45629E7A@baesystems.com> Many thanks to all. Paul --------------------------- Newsgroups: comp.os.vxworks Subject: PLX 9054 (and vxWorks) help Date: Tue, 12 Jun 2001 08:41:00 +0100 From: Paul Caswell Organization: Land & Sea Systems Message-ID: <3B25C78C.E964F694@baesystems.com> Hello all, Does anyone out there have experience of this device under vxWorks ? It's a PCI I/O accelerator. I've initialised it's registers as defined in the manual but the BAR0 to BAR3 registers are not being set by the host software on boot. Any help greatly appreciated. Paul Caswell --------------------------- Newsgroups: comp.os.vxworks Subject: comparision on realtime systems Date: Tue, 12 Jun 2001 08:08:49 GMT From: rsasalm Organization: http://www.newsranger.com Message-ID: Hi, I am interested in knowing if there is any place on the internet where I can find the differences among major commercially avaiable realtime systems. Any help in this regard will be appreciated. Better if you send me answer on the following email adress: mohammad.saleem@elema.siemens.se thanks in advance regards rsasalm --------------------------- Newsgroups: comp.protocols.snmp,comp.os.vxworks Subject: Extending the WindRiver (TMS) Envoy SNMP agent Date: Tue, 12 Jun 2001 11:01:37 +0200 From: Michael Gaudlitz Message-ID: <3B25DA71.95D61C13@gaudlitz.com> hi, any guys out there have experience on enxtending the SNMP agent that comes with TMS? Actually it is an Envoy Agent but in TMS it is hidden behind some IDB+ database stuff. So instead of running mibcomp you have to run idbgen on your mib which also produces .c and .h. I also succeed in compiling and integrating it in the VxWorks image. But any query to my (enterprise) OID is not beeing answered. I suspect the error in missing code I have to add by hand in the produced .c and .h files. But there is no good documentation or example on that. Can anyone help me or point me to some docs? Thanks Michael --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Application code entry point Date: Tue, 12 Jun 2001 15:47:24 +0100 From: Vijay Kumar Peshkar Organization: Cisco Systems Inc. Message-ID: <3B262B7C.2B7D8ECE@wipro.com> References: <9g48hu$ov4$1@overload.lbl.gov> This is a multi-part message in MIME format. - --------------12649EA9972467853E158F02 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Rajendra, vxWorks goes multitasking in usrRoot(). If INCLUDE_NET_INIT is included, network is initialized with the routine usrNetInit() here. Here at the end we call the first user task(application task). Refer vxWorks Prog. guide 8.3 pg 343 "The Initial Task:usrRoot()" for details. As networking is just another component, it doesn't really effect the over all bootup. Hope this helps. regards, Vijay Rajendra Prasad wrote: > vxworks tornado > > Hi, > > Could somebody explain me the procedure to set the application code entry > point for the vxworks image to be booted in a network and non-network boot > environments ? Any relevant manual references will be greatly appreciated. > > Thanks and best regards, > rajendra. > _________________________________________________________________________ > Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com. - --------------12649EA9972467853E158F02 Content-Type: text/x-vcard; charset=us-ascii; name="vijay.peshkar.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Vijay Kumar Peshkar Content-Disposition: attachment; filename="vijay.peshkar.vcf" begin:vcard n:Peshkar;Vijay Kumar tel;work:5530034/35/35 x2237 x-mozilla-html:FALSE org:Wipro Global R & D;Wishing you blue skies, a couple of daisies, some fresh air and a smile! adr:;;;;;; version:2.1 email;internet:vpeshkar@cisco.com note:Wishing you blue skies, a couple of daisies, some fresh air and a smile! fn:Vijay Kumar Peshkar end:vcard - --------------12649EA9972467853E158F02-- --------------------------- Newsgroups: comp.os.vxworks Subject: Change MAC address dynamically in PPC8260 Date: Tue, 12 Jun 2001 10:24:18 GMT From: Gururaj patkar Organization: Optimum Online Message-ID: <3B25ED5F.E7E492A4@optonline.net> Hi, We have the need to change the Ethernet MAC address of one of the FCC channel on PPC8260 at runtime dynamically. 1) if I use the call to sysEnetAddrSet(), will these corrupt the boot parameters in the flash? 2) If (1) is not the solution how do I go about doing this and what precaution should I take for this. Thanks for the help. gururaj --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Tue Jun 12 05:14:48 2001 From: "Michael Lawnick" Date: Tue Jun 12 05:14:50 PDT 2001 Subject: Re: objdump error Hi Amadou are you using this line in a makefile ? If so, your problem is that you used spaces to ident the line, insted of using tabulator. Mit freundlichen Grüßen, Michael Lawnick ============================================== SOFTEC GmbH Tel +49-731-96600-0 Promenade 17 Fax +49-731-96600-23 D-89073 Ulm Michael Lawnick Germany lawnick@softec.de ============================================== ----- Original Message ----- From: "Amadou Diop" Newsgroups: comp.os.vxworks Sent: Monday, June 11, 2001 10:38 PM Subject: objdump error > List file generation appl does not run properly using the mottorolla (VX_CPU_FAMILY 68k). > > Here is the statement: > objdump -d $*.o > $*.lst > > and this is the error the I get > "missing separator" > > Has anyone encountered the same problem before. From vxwexplo-errs@csg.lbl.gov Tue Jun 12 07:46:30 2001 From: "B. J. Dou" Date: Tue Jun 12 07:46:32 PDT 2001 Subject: Re: objdump error It still doesn't work after I took the tab or spaces out. --- Michael Lawnick wrote: > Hi Amadou > > are you using this line in a makefile ? If so, your > problem is that you used > spaces to ident the line, insted of using tabulator. > > Mit freundlichen Grüßen, > Michael Lawnick > ============================================== > SOFTEC GmbH Tel +49-731-96600-0 > Promenade 17 Fax +49-731-96600-23 > D-89073 Ulm Michael Lawnick > Germany lawnick@softec.de > ============================================== > > ----- Original Message ----- > From: "Amadou Diop" > Newsgroups: comp.os.vxworks > Sent: Monday, June 11, 2001 10:38 PM > Subject: objdump error > > > > List file generation appl does not run properly > using the mottorolla > (VX_CPU_FAMILY 68k). > > > > Here is the statement: > > objdump -d $*.o > $*.lst > > > > and this is the error the I get > > "missing separator" > > > > Has anyone encountered the same problem before. > > __________________________________________________ Do You Yahoo!? Get personalized email addresses from Yahoo! Mail - only $35 a year! http://personal.mail.yahoo.com/ From vxwexplo-errs@csg.lbl.gov Tue Jun 12 12:05:03 2001 From: Chris Slominski Date: Tue Jun 12 12:05:06 PDT 2001 Subject: PowerPC combining consecutive VME reads VxWorks Users, I am porting code from a mv167 to a mv2700 PowerPC. My code reads consecutive 8-bit quantities from aVME board that does not support D32 acccess. I have a VME bus analyzer that I use to inspect the VME activity of the software. When the software begin, the 8-bit reads are made as D8 VME accesses, but very rapidly the hardware starts making D32 accesses of the consecutive addresses instead. The D32 accesses cause the software to fail. How do I force the hardware to not perform D32 accesses? Chris From vxwexplo-errs@csg.lbl.gov Tue Jun 12 17:08:42 2001 From: "Leung, Eugene" Date: Tue Jun 12 17:08:44 PDT 2001 Subject: midas-20 Has anyone written a vxworks driver to use this VME to PCI carrier? The board seems simple enough. Most of the work is done by the Universe II. However, I am having trouble writing to the PCI registers CF8 & CFC from the VME side to config the PCI bus. I've setup the VME slave images to access these in the io space. I can read from the register (even though I have no idea if what I read is meaningful), but write gives me bus errors. I am using Motorola MVME177 with vxworks 5.2. Hope someone can provide some pointers. Please send email to eugenel@condorsys.com as I am no longer on the exploder list. TIA. -Eugene From vxwexplo-errs@csg.lbl.gov Tue Jun 12 22:01:12 2001 From: "Sanesh Kumar K" Date: Tue Jun 12 22:01:14 PDT 2001 Subject: VxWorks : Mapping the Interrupt Levels Hi, In intLevelSet(int level), how the levels are mapped to atcual interrupt numbers ? Bye Sanesh Kumar K From vxwexplo-errs@csg.lbl.gov Tue Jun 12 23:42:04 2001 From: Brian Phelan Date: Tue Jun 12 23:42:07 PDT 2001 Subject: Re: midas-20 Eugene, I did a driver for the midas-20 (and midas-50) a few years ago. If you're trying to access the configuration registers on the PMC devices, what you want to do is to configure one of the 4 VME slave images to access PCI config space. You should then be able to access the config header of the devices in slots 1 and 2 at offsets 0x2800 and 0x3000 respectively, from the base address of the image. Hope this helps. Brian the vxWorks Users Group Exploder wrote: > > Submitted-by vxwexplo-errs@csg.lbl.gov Tue Jun 12 17:08:42 2001 > Submitted-by: "Leung, Eugene" > > Has anyone written a vxworks driver to use this VME to PCI carrier? The > board seems simple enough. Most of the work is done by the Universe II. > However, I am having trouble writing to the PCI registers CF8 & CFC from the > VME side to config the PCI bus. I've setup the VME slave images to access > these in the io space. I can read from the register (even though I have no > idea if what I read is meaningful), but write gives me bus errors. I am > using Motorola MVME177 with vxworks 5.2. Hope someone can provide some > pointers. Please send email to eugenel@condorsys.com as I am no longer on > the exploder list. TIA. > -Eugene -- ======================================================= T h e P T R G r o u p, I n c. ======================================================= Brian Phelan (703)869-9066 brian@ThePTRGroup.com www.ThePTRGroup.com Embedded, Real-Time Solutions ======================================================= From vxwexplo-errs@csg.lbl.gov Wed Jun 13 01:36:06 2001 From: David Laight Date: Wed Jun 13 01:36:08 PDT 2001 Subject: Re: clockLib problems > I read few postings, describing different problems with clockLib - > timer gets reset periodically, going backwards and so on. > I have basicly the same problem - time() increments to some seconds, > gets reset to zero, and after a that while again the same. > I tried to debug, windTickAnnounce, incrementing the vxTicks, but > problem seems to be complicated. I looked into this a few months ago. It appears that vxWorks doesn't believe in real time! The underlying problem is the way vxworks handles the tick counter wrapping. What I found was that if ANY timeout call (eg taskDelay, semTake) had an expiry time which was later than the time at which the tick counter wrapped, then the tick counter was reset to zero and all the pending timeouts adjusted. However the offset isn't transferred to the code that supports time() - so time() leaps backwards. This is easily reproducable by calling 'sp taskDelay, -2' from the shell. David ---------------------------------------------------------------- David Laight email: dsl@tadpole.co.uk Tadpole Technology plc phone: +44 1223 428 232 Cambridge, UK fax: +44 1223 428 201 From vxwexplo-errs@csg.lbl.gov Wed Jun 13 04:03:33 2001 From: Vxworks Exploder Date: Wed Jun 13 04:03:36 PDT 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Wed Jun 13 04:03:08 PDT 2001 Subject: Re: objdump error Subject: re: objdump error Subject: objdump error Subject: Re: Help for init. 53C875 SCSI controller Subject: VxWorks drivers for z85230 synchronous and DMA modes of operation Subject: Tornado Browser Problem Subject: Re: PLX 9054 (and vxWorks) help Subject: re: objdump error Subject: VxWorks routing problems Subject: WDB agent communication error Subject: Re: objdump error Subject: Re: objdump error Subject: Re: Change MAC address dynamically in PPC8260 Subject: Re: VxWorks drivers for z85230 synchronous and DMA modes of operation Subject: MAX Timers in VxWroks ! Subject: Compiling VxWorks using Tornado 2.0 from Command Line Subject: PowerPC combining consecutive VME reads Subject: Help for a newbie with vxworks Subject: Re: Where can one learn to use VxWorks Subject: Re: Proxy ARP problems: workaround needed Subject: Re: PLX 9054 (and vxWorks) help Subject: Command line syntax for compiling Tornado projects Subject: Re: Processing of Raw TCP/IP Packets Subject: Handling Multiple Memory Maps In Device Driver Subject: Re: PLX 9054 (and vxWorks) help Subject: Re: How do I send a raw Ethernet packet? Subject: Re: Interrupt handling with compact PCI Subject: Re: VxWorks drivers for z85230 synchronous and DMA modes of operation Subject: midas-20 Subject: Re: Network Stack Subject: vxworks vs. threadx ? Subject: Re: problem with inet_aton Subject: Re: ARP cache Subject: Re: Extending the WindRiver (TMS) Envoy SNMP agent Subject: Re: VxWorks routing problems Subject: VxWorks : Mapping the Interrupt Levels Subject: INT0 on x86 BSP Subject: Re: Where can one learn to use VxWorks Subject: re: midas-20 Subject: Re: Priority inversion in semaphores Subject: Re: New Altivec-enhanced VxWorks Gcc 2.95.4 compiler released! Subject: re: clockLib problems Subject: svgmd Build Boot rom Subject: Re: Extending the WindRiver (TMS) Envoy SNMP agent Subject: Re: Extending the WindRiver (TMS) Envoy SNMP agent ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: objdump error Date: Tue, 12 Jun 2001 10:36:04 +0100 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: <3qnV6.12$523.82274@newsr2.u-net.net> References: Amadou Diop wrote in message ... >List file generation appl does not run properly using the mottorolla (VX_CPU_FAMILY 68k). > >Here is the statement: > objdump -d $*.o > $*.lst > >and this is the error the I get > "missing separator" > >Has anyone encountered the same problem before. Yes, it is a very common problem in makefiles. In a makefile, the lines with commands on them must begin with a TAB character. Ordinary ASCII space chars (0x20) aren't any good; it must be a genuine 0x09 TAB. DaveK - -- They laughed at Galileo. They laughed at Copernicus. They laughed at Columbus. But remember, they also laughed at Bozo the Clown. --------------------------- Newsgroups: comp.os.vxworks Subject: re: objdump error Date: Tue, 12 Jun 2001 13:54:02 +0200 From: "Michael Lawnick" Organization: Lawrence Berkeley National Laboratory Message-ID: <9g514k$ehg$1@overload.lbl.gov> Hi Amadou are you using this line in a makefile ? If so, your problem is that you used spaces to ident the line, insted of using tabulator. Mit freundlichen Grüßen, Michael Lawnick ============================================== SOFTEC GmbH Tel +49-731-96600-0 Promenade 17 Fax +49-731-96600-23 D-89073 Ulm Michael Lawnick Germany lawnick@softec.de ============================================== - ----- Original Message ----- From: "Amadou Diop" Newsgroups: comp.os.vxworks Sent: Monday, June 11, 2001 10:38 PM Subject: objdump error > List file generation appl does not run properly using the mottorolla (VX_CPU_FAMILY 68k). > > Here is the statement: > objdump -d $*.o > $*.lst > > and this is the error the I get > "missing separator" > > Has anyone encountered the same problem before. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Help for init. 53C875 SCSI controller Date: 12 Jun 2001 06:32:05 -0700 From: lilyko2002@yahoo.com (lily) Organization: http://groups.google.com/ Message-ID: <7f208981.0106120532.1c6fd174@posting.google.com> References: <7f208981.0106100351.2b2a5849@posting.google.com> <30c8bd8b.0106110635.b6d69dd@posting.google.com> trahn@macrosystem.de (Thomas Rahn) wrote in message news:<30c8bd8b.0106110635.b6d69dd@posting.google.com>... > Hi, > > at our site, the ncr810 driver is working on a > x86 target... > > There had been several obstacles since our first > attempts: > - a NULL pointer access > - hard coded pci address range > - missing define for 875 > - offset 0x8000 0000 to pci memory range > - ... > > From your mail i guess WRS has fixed some of those > in the current driver package. If a ncr810Show() > displays some valid register entries, that part > should be working. > > However, there might be left a bug in the ncr810 driver > (and sym895 as well) regarding wide/sync negotiation. > I never convinced WRS about it. You will need the driver > source to fix it. > > Try to spawn the scsi detection as a low-priority task. > If the "tScsiTask" stays ready, you might get on if > you disable all non-basic scsi features for the target. > [using scsiTargetOptionsSet() or similar]. > Or: there might be a jumper on the drive to disable > any such feature at the drive. > > Thomas Rahn Thanks to the help of Julia and Thomas, My SCSI driver is working now. Regards, Lily --------------------------- Newsgroups: comp.os.vxworks Subject: VxWorks drivers for z85230 synchronous and DMA modes of operation Date: Tue, 12 Jun 2001 09:15:37 -0400 From: "Gerald Chang" Organization: Lockheed Martin -- Information Systems Center Message-ID: <9g54lt$73b$1@zeus.orl.lmco.com> Hello All, I'm trying to use the i8237 DMA controller to send/recv data through the z85230 in the SDLC mode. Does anyone know where I can obtain any "C" drivers that support the DMA and synchronous modes of the z85230 and i8237A that are on the Motorola MVME2700 board? VxWorks supports only the async mode. Regards, Gerald Chang Lockheed Martin Corp. --------------------------- Newsgroups: comp.os.vxworks Subject: Tornado Browser Problem Date: Tue, 12 Jun 2001 10:31:25 -0400 From: Sandy Olson Organization: Raytheon Company Message-ID: <3B2627BD.F887A9A0@fallschurch.esys.com> We're having a problem with our Browser Tool. After our applications software has been loaded, the "Tools" memory consumption bar graph, on the Browser Window, indicates that it is maxed out (the bar is totally black). This seems to be preventing us from running other tools, such as the CPU utilization tool. Is this a result of declaring global variables? Also, what does the number at the bottom end of the "Tools" bar represent? Is there a kernel parameter that can be modified to fix this? We are running VxWorks on a Radstone 750 PowerPC using BSP 1.6. Tornado (2.0) is running in a Unix environment on Sun Solaris 8.0. Thank-you in advance for any assistance. - Sandy Olson --------------------------- Newsgroups: comp.os.vxworks Subject: Re: PLX 9054 (and vxWorks) help Date: Tue, 12 Jun 2001 08:07:05 -0700 From: "Joseph Hansen" Organization: Posted via Supernews, http://www.supernews.com Message-ID: References: <3B25C78C.E964F694@baesystems.com> Paul, When you say "host software", are you talking about VxWorks/Tornado, or are you talking about the device driver? As far as I know, the VxWorks OS itself will not do any setting of any PCI BAR registers. They provide routines for initializing the PCI bus, but after that, you as a developer are responsible for writing your drivers for querying the system for PCI devices, their locations (bus, slot, and function numbers), the required sizes for memory and I/O, and then setting the BAR registers (and internally keeping track of where everything is). It get really fun when you have multiple PCI buses, and you have to go through PCI-to-PCI bridges ... Hope this helps. This has been my limited experience, so anyone is welcome to correct me if I spoke incorrectly. Joseph Hansen NAWCWPNS China Lake (760) 939-6999 HansenJR@navair.navy.mil "Paul Caswell" wrote in message news:3B25C78C.E964F694@baesystems.com... > Hello all, > Does anyone out there have experience of this device under vxWorks ? > It's a PCI I/O accelerator. I've initialised it's registers as defined > in the manual but the BAR0 to BAR3 registers are not being set by the > host software on boot. Any help greatly appreciated. > > Paul Caswell > --------------------------- Newsgroups: comp.os.vxworks Subject: re: objdump error Date: Tue, 12 Jun 2001 07:46:28 -0700 (PDT) From: "B. J. Dou" Organization: Lawrence Berkeley National Laboratory Message-ID: <9g5bm3$nph$1@overload.lbl.gov> It still doesn't work after I took the tab or spaces out. - --- Michael Lawnick wrote: > Hi Amadou > > are you using this line in a makefile ? If so, your > problem is that you used > spaces to ident the line, insted of using tabulator. > > Mit freundlichen Grüßen, > Michael Lawnick > ============================================== > SOFTEC GmbH Tel +49-731-96600-0 > Promenade 17 Fax +49-731-96600-23 > D-89073 Ulm Michael Lawnick > Germany lawnick@softec.de > ============================================== > > ----- Original Message ----- > From: "Amadou Diop" > Newsgroups: comp.os.vxworks > Sent: Monday, June 11, 2001 10:38 PM > Subject: objdump error > > > > List file generation appl does not run properly > using the mottorolla > (VX_CPU_FAMILY 68k). > > > > Here is the statement: > > objdump -d $*.o > $*.lst > > > > and this is the error the I get > > "missing separator" > > > > Has anyone encountered the same problem before. > > __________________________________________________ Do You Yahoo!? Get personalized email addresses from Yahoo! Mail - only $35 a year! http://personal.mail.yahoo.com/ --------------------------- Newsgroups: comp.os.vxworks Subject: VxWorks routing problems Date: Tue, 12 Jun 2001 15:16:39 GMT From: Ran@127.0.0.1 (Ran Shalgi) Organization: Only when absolutely necessary Message-ID: <3b262c2b.112640638@news.inter.net.il> We're developing a VxWorks-based router software, and we've come across a rather unusual problem that has so far completely eludes us. In a Cisco router, you can assign a route directly to an ethernet interface, without a default gateway. for example, the ethernet interface is IP 192.168.1.2 subnet mask 255.255.0.0 and in addition one can define a route of IP 172.16.32.0 mask 255.255.255.0 => destination ethernet 0. This means that this new subnet is considered directly connected to the ethernet port, and that port communicates with those addresses in a "normal" form, by first ARPing them directly and then sending the packets, without a gateway in between. In fact, what we need is even simpler: we want one ethernet interface to have a very narrow IP range (i.e. 192.168.1.2 mask 255.255.255.255) and in addition have ONE more machine on that LAN, e.g. to define a route 10.0.0.1 mask 255.255.255.255 => destination ethernet 0. Two disjoint IP addresses directly connected to the same ethernet port, and only these two addresses. We've not been able to get this to work. Sometimes the IP stack sends the packets to the right interface, but the interface outputs them with a destination MAC address of all 1s (broadcast address). In slightly different configuration, the intherface outputs the packets with destination MAC address equal to it's own source MAC address (?!) which of course is no good. In other cases it seems the IP stack does not forward packets at all to the i/f. Has anyone ever managed to get two disjoint subnets on the same ethernet, or managed to get any kind of routing to be considered "directly connected" to an ethernet port? WindRiver support doesn't even understand the question, and we really need this, so any help would be greatly appreciated. In case it matters, we're using x86 BSP, T2 WinNT hosts, VxWorks 5.4. - -- Ran Shalgi replace 127.0.0.1 with expand.com when replying by e-mail --------------------------- Newsgroups: comp.os.vxworks Subject: WDB agent communication error Date: 12 Jun 2001 08:22:14 -0700 From: girishvg@yahoo.com (girish) Organization: http://groups.google.com/ Message-ID: hi. i have developed an ethernet driver for a prop hw. this is a END style driver. & i suppos it basically works fine. the target board does download image from nw & boots. also trivial socket programs work. now i create target server in network mode. the vxworks image has been configured with WDB agen is nw. the target server attached when i try to attach a suspended user task in wdb it gives "WDB Exception restarting agent in 10 sec." and drops out. guess this is bascially a configuration problem. could anyone of you please guide me to solve this problem. many thanks in advance. girish. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: objdump error Date: 12 Jun 2001 11:39:05 -0400 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: References: <9g5bm3$nph$1@overload.lbl.gov> Sender: bpringlemeir@DeadDuck >>>>> "B" == B J Dou writes: >> Here is the statement: > >> objdump -d $*.o > $*.lst > > and this is the error the I get > >> "missing separator" > > Has anyone encountered the same problem >> before. B> It still doesn't work after I took the tab or spaces out. The tool name is `objdump68k' or something like that. If you have mingw or cygwin or you are on a *Unix machine, then you might have an objdump that is not compatible with your object format. regards, Bill Pringlemeir. - -- The Bible makes sense. Useful vxWorks URLS, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" "http://www.xs4all.nl/~borkhuis/vxworks/vxw_pt9.html" --------------------------- Newsgroups: comp.os.vxworks Subject: Re: objdump error Date: Tue, 12 Jun 2001 16:57:40 +0100 From: Simon Farnsworth Message-ID: <3B263BF4.4060707@snellwilcox.com> References: <9g5bm3$nph$1@overload.lbl.gov> B. J. Dou wrote: > It still doesn't work after I took the tab or spaces > out. > You don't want to take it out, you want to replace the spaces at the beginning of the line with a tab. Note that it must be a tab, not n spaces. > --- Michael Lawnick wrote: > >>Hi Amadou >> >>are you using this line in a makefile ? If so, your >>problem is that you used >>spaces to ident the line, insted of using tabulator. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Change MAC address dynamically in PPC8260 Date: Tue, 12 Jun 2001 17:25:36 GMT From: Dan Gold Organization: http://www.newsranger.com Message-ID: References: <3B25ED5F.E7E492A4@optonline.net> In article <3B25ED5F.E7E492A4@optonline.net>, Gururaj patkar says... > >Hi, > >We have the need to change the Ethernet MAC address of one of the FCC >channel on PPC8260 at runtime dynamically. Are you sure you have the need ? Really sure ? Most of the time, if you change the MAC address of an interface on the fly you're asking for trouble. Assuming you do all the required magic (and there will be magic) to change it, how do you propagate your change to all the other people on the network that have your old one ? Send out a gratuitous ARP ? You'll have to do it yourself. Wait for all the connections to drop because of no response ? :-) Not a very clean solution. Count on everyone else's ARP table timing out ? Again, not very efficient. You could (if all your stars, phase of the moon, and dow jones average all aligned), try setting the new MAC address with sysNvRamSet(), then unloading/loading/starting/attaching the FCC driver, like they do in usrNetInit(). Of course the FCC driver would need to support unloading and loading again. I have an early one and a quick perusal makes me think it would, but I'm not using that one so I can't say for sure. How many times will you need to do this ? Constantly while the board is running, or just once ? >1) if I use the call to sysEnetAddrSet(), will these corrupt the boot >parameters in the >flash? No, it shouldn't corrupt the boot parameters in FLASH. Where is your MAC address stored ? Most boards store it in serial eeprom, not FLASH so it shouldn't be affected. If your MAC address is stored in FLASH, better make sure no one cycles power or touches FLASH while you're erasing and re-progromming with the new MAC address. The MAC address is not part of the boot parameters (in the boot line, per se) so it shouldn't affect you. Your bigger problem is changing it on the fly. You'll have to write the 8260 DPRAM yourself, because the FCC driver (with ioctl(SIFADDR...) only copies it to the driver control struct). The function that writes it into the DPRAM is a local (to the driver) one, and only called from the xxxEndStart() routine. Of course if you're going to change it on the fly, you better make sure you're not currently in the middle of Tx'ing or Rx'ing a frame. I.e. you better stop the FCC (with GSMR bits) before changing the MAC address. What about data that's currently buffered up on output queues in the SENS layer ? Should that go out with the new MAC address ? Do you care ? >2) If (1) is not the solution how do I go about doing this and what >precaution should I take for this. You should take lots of precautions, the biggest one being try to find another solution other than changing the MAC address on the fly. hth, Dan Gold gold@ensemble.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks drivers for z85230 synchronous and DMA modes of operation Date: Tue, 12 Jun 2001 10:51:56 -0700 From: Christopher A Leddy Organization: Raytheon Company Message-ID: <3B2656BC.E8A58F31@west.raytheon.com> References: <9g54lt$73b$1@zeus.orl.lmco.com> Try http://kernelnewbies.org/documents/kdoc/z8530book/z85230guide.html Source code is covered under GNU General License, but it is a good as a reference. Gerald Chang wrote: > > Hello All, > > I'm trying to use the i8237 DMA controller to send/recv data through the > z85230 in the SDLC mode. > > Does anyone know where I can obtain any "C" drivers that support the DMA > and synchronous modes of the z85230 and i8237A that are on the Motorola > MVME2700 board? > > VxWorks supports only the async mode. > > Regards, > Gerald Chang > Lockheed Martin Corp. --------------------------- Newsgroups: comp.os.vxworks Subject: MAX Timers in VxWroks ! Date: Tue, 12 Jun 2001 14:17:27 -0400 From: Srinivas Pitta Organization: Marconi Message-ID: <3B265CB7.B433F58A@Marconi.com> I wonder whether there is any maximum LIMIT of timers, that we can spawn in VxWorks system. (watch dog timers, using wdStart call). - -Srini --------------------------- Newsgroups: comp.os.vxworks Subject: Compiling VxWorks using Tornado 2.0 from Command Line Date: Tue, 12 Jun 2001 12:00:08 -0700 From: "Langston" Organization: SBC Internet Services Message-ID: <3FtV6.106$bR1.60252@news.pacbell.net> Hello All, Currently we are using Tornado 2.0 to build or VxWorks applications. We generally load .wsp (Tornado Workspace Files) to the IDE and create our builds from there. I would like to be able to script this build process but if I use a 'Make' command on the Makefile then I run the risk of not getting all of the latest changes because this Makefile is generated by the Workspace and Project files. Is there a command line for building applications from the project files? If I could generate the Makefile from the command line then perform a Make on it? Thank you, Langston --------------------------- Newsgroups: comp.os.vxworks Subject: PowerPC combining consecutive VME reads Date: Tue, 12 Jun 2001 15:04:27 -0400 From: Chris Slominski Organization: Lawrence Berkeley National Laboratory Message-ID: <9g5poe$79$1@overload.lbl.gov> VxWorks Users, I am porting code from a mv167 to a mv2700 PowerPC. My code reads consecutive 8-bit quantities from aVME board that does not support D32 acccess. I have a VME bus analyzer that I use to inspect the VME activity of the software. When the software begin, the 8-bit reads are made as D8 VME accesses, but very rapidly the hardware starts making D32 accesses of the consecutive addresses instead. The D32 accesses cause the software to fail. How do I force the hardware to not perform D32 accesses? Chris --------------------------- Newsgroups: comp.os.vxworks Subject: Help for a newbie with vxworks Date: Tue, 12 Jun 2001 21:46:36 +0200 From: "Alain Bougeniere" Organization: Club-Internet (France) Message-ID: <9g5rsv$4no$1@front5.grolier.fr> Reply-To: "Alain Bougeniere" I am intersted in compilers;... and perhaps simulator ?.. Could you help me ? :) thanks a lot! --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Where can one learn to use VxWorks Date: Tue, 12 Jun 2001 22:06:57 +0200 From: "Alain Bougeniere" Organization: Club-Internet (France) Message-ID: <9g5t36$t1h$1@front7.grolier.fr> References: <3B210D50.CEA4CECD@Sun.COM> Reply-To: "Alain Bougeniere" where can you find this free tool ?... Thanks a lot! Alain --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Proxy ARP problems: workaround needed Date: Tue, 12 Jun 2001 17:11:53 -0400 From: Bill Fulton Organization: Raytheon Company Message-ID: <3B268599.3E163DD4@raytheon.com> References: <3B211164.3AA58342@auriga.ru> The solution that we came up with was to use the ATF_PUBL flag with the arpAdd() function. This seems to, in effect, have all the desired consequences - our card responds to an ARP request for the defined IP address with its own MAC address, and then routes the IP message to the appropriate place. This requires a manual setup of the ARP table ... but it seems to work well. (Of course, you have to have routing enabled.) In our case, though, we were forwarding messages between ethernet networks, not a network and a backplane. It's my understanding that the routing should still work, but .... Bill Pavel Kovalev wrote: > Hello All, > > We have a custom network driver for a CompactPCI backplane network > and want to configure a proxy ARP server on the CompactPCI host > board to make the backplane nodes visible on the Ethernet network > (without additional routing). > > Tornado 2/VxWorks 5.4 support this configuration only for > the shared memory network drivers supplied by WRS. Attempts > to use proxyArpLib or configure all the necessary routing > manually fail with various errors. > > I studied various materials available on WindSurf but haven't > found a solution for this problem. There is an official SPR > describing this problem (SPR# 28602), but it looks like it > hasn't been processed/fixed yet. There were some discussions > of this problem in the newsgroup in the past, but no good > solutions have been proposed. > > I spent some time investigating this problem, and found that when > I add a specific route to a node residing on the proxy network > (route is added on the host which has to work proxy ARP server), > mRouteAdd() creates the route with RTF_GATEWAY set. This prevents > arpAdd() from adding a permanent published entry for the node. > One solution is to develop a better version of mRouteAdd(), but > I am not yet ready/willing to do so - there is no warranty that > other hacks won't be needed. > > So, I am wondering if anybody has managed to solve this problem. > > Thanks, > Pavel. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: PLX 9054 (and vxWorks) help Date: Tue, 12 Jun 2001 17:20:53 -0400 From: "Tim Stanka" Organization: Eastman Kodak Company Message-ID: <9g613n$4ve$1@news.kodak.com> References: <3B25C78C.E964F694@baesystems.com> Paul, The PLX 9054 is a 64 bit PCI interface chip (PCI to local bus). If you have PCI autoconfig enabled in your boot ROM code, VXWORKS image or combined app it should get configured along with any otherPCI devices. I have alot of experience with PCI and PCI autoconfig software that is usually part of a board support package. I also have experience with the PLX 9080 and we are using the 9054 in our hardware. To see if pci is included you could do -> lkup "pci" and see if pciAutoconfig shows up. If it does you can use the PCI show routines (pciHeaderShow()) to see how the device is set up. How are you looking at BAR values?? are you using a bus analyzer. What you should do is write 0xFFFFFFFF to all 6 BARs (using pciConfigOutLong()) in the 9054 and then read then back. If they are implemented they should not return 0x0 otherwise they will. BAR0 and 1 should always return 0xFFFFFF00 and 0xFFFFFF01 for the CSR. Tim --------------------------- Newsgroups: comp.os.vxworks Subject: Command line syntax for compiling Tornado projects Date: Tue, 12 Jun 2001 14:26:49 -0700 From: "Langston" Organization: SBC Internet Services Message-ID: Hello All, Currently we are using Tornado 2.0 to build our VxWorks applications. We generally load .wsp (Tornado Workspace Files) to the Tornado IDE and create our builds from there. I would like to be able to script this build process but if I use a 'Make' command on the Makefile then I run the risk of not getting all of the latest changes because this Makefile is generated by the Workspace and Project files. Is there a command line for building applications from the project files? If I could generate the Makefile from the command line then perform a Make on it (so that I have the current Makefile)? Thank you, Langston --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Processing of Raw TCP/IP Packets Date: Tue, 12 Jun 2001 17:29:44 -0400 From: Bill Fulton Organization: Raytheon Company Message-ID: <3B2689C8.7B2252E3@raytheon.com> References: <628eca30.0106090633.251aa368@posting.google.com> Dave Korn wrote: > "Ivan Bach" wrote in message > news:628eca30.0106090633.251aa368@posting.google.com... > > I have not yet seen vxsniff.c program, but it was apparently based on > > etherhooks. Does anybody actually have a copy of that program, even if > > it is not working? The pointers that I saw for that program don't lead > > anywhere. > > [DK] I hope Hwa-Jin Bae won't be offended by my distributing this. > > ..... > > void > vxSniffInit() > { > /* XXX do something here to put ethernet driver in promiscuous mode Yeah - that's kinda the tricky part. :-) (All of the rest is just standard "hook in" calls and formatting of the data - pretty easy stuff.) That, plus adding the EtherHook*() hooks to your ethernet driver, if it doesn't already have it (the driver template code has it). But the promiscuous mode transition was the toughest for us - we knew what the chip wanted, but couldn't "bust in" to the chip initialization code in the ethernet driver well enough to get (dynamically) selectable promiscuity. There are many different layers of initialization, including ioctl() interactions leading to reinitialization, with logical and physical bit flags and latencies. The alternative is to have the chip run in promiscuous mode all of the time (should work ok ... just load the CPU). --------------------------- Newsgroups: comp.os.vxworks Subject: Handling Multiple Memory Maps In Device Driver Date: Tue, 12 Jun 2001 17:33:00 -0400 From: "Tim Stanka" Organization: Eastman Kodak Company Message-ID: <9g61qp$541$1@news.kodak.com> Hi, I've been put in the ugly position of having to deal with multiple memory maps for a device within one device driver. I am trying to design the driver to deal with memory map changes amung different hardware versions. I have a C structure to represent each register map and I would like to look at the hardware's version register to determine what register map I should use and then set a pointer to that map type that all my ioctls can use. The only pointer that can be set to any type is "void*" but void cannot be used because the compiler cannot figure out field offsets for the structure it points to. A pointer to a type which is a union of all register map types could be created but that would add an intermediate field entry: typedef struct { . . . .} map1; typedef struct {........)map2; union {map1; map2} combType; comnType *ptr; if (version2) ptr->map2->field; I would like to be able to do ptr->field; in my ioctls. without the if (version): map1 *ptr1; map2 *ptr2; if (version1) ptr1->field; Any thoughts??? Thanks, Tim --------------------------- Newsgroups: comp.os.vxworks Subject: Re: PLX 9054 (and vxWorks) help Date: Tue, 12 Jun 2001 22:31:31 +0100 From: molochai Message-ID: <2e66a7894a%molochai@vapour-trail.demon.co.uk> References: <3B25C78C.E964F694@baesystems.com> In message "Joseph Hansen" wrote: > Paul, > > When you say "host software", are you talking about VxWorks/Tornado, or are > you talking about the device driver? > > As far as I know, the VxWorks OS itself will not do any setting of any PCI > BAR registers. Depends. VxWorks can do automatic bus-scans in which BARs are programmed and the devices control registers automatically written to enable bus mastering [1]. Automatic determination of the interrupt line the device is connected to is also possible with BSP support. However, ISTR that there are some issues wrt PCI-PCI bridges. [1] Not on some x86 platforms with a BIOS though. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How do I send a raw Ethernet packet? Date: Tue, 12 Jun 2001 17:19:01 -0400 From: Bill Fulton Organization: Raytheon Company Message-ID: <3B268745.B9A44227@raytheon.com> References: <9avf790k94@drn.newsguy.com> You have to be careful with raw packet interfaces. Using the ATF_PUBL flag with the arpAdd() will do (proxy) ARP replies for you, you just have to "prime" the arp cache. wdm@world.std.com wrote: > I would like to reply to certain ARP requests (to implement a proxy-arp-like > scheme). > > I can access the ARP requests by binding a "SNARF" type protocol to the driver, > but how do I send the arp reply? Pointers would be greatly appreciated! > > I am using a 68360 BSP, with the QU END. > > Thanks, > Wayne --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Interrupt handling with compact PCI Date: Tue, 12 Jun 2001 17:48:15 -0400 From: "Tim Stanka" Organization: Eastman Kodak Company Message-ID: <9g62n2$58o$1@news.kodak.com> References: <1103_991755221@stockholm> Dan, Your PCI configuration software should write a value to the interrupt line field in your device's config space. I believe this is config offset 0x3C. This will tell you which pin on your interrupt controller chip will be used. With this value you can add an offset defined in your BSP for the base of external interrupts. For example, with the MCP750, a device in logical slot 2 will have it's interrupt line programmed with 0xb. The BSP define (don't remember the name) is set to 0x10. When you add the two you get 0x1B. This value should be used as an arguement to intConnect() and intEnable to register your ISR and enable the interrupt. Be aware, with CPCI, if you move your board from slot to slot, the system slot interrupt line used will change. This is accounted for when the autoconfig routine programs the interrupt line field in config space. Hope this helps --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks drivers for z85230 synchronous and DMA modes of operation Date: Tue, 12 Jun 2001 16:42:37 -0400 From: "Gerald Chang" Organization: Lockheed Martin -- Information Systems Center Message-ID: <9g5urs$av0$1@zeus.orl.lmco.com> References: <9g54lt$73b$1@zeus.orl.lmco.com> <3B2656BC.E8A58F31@west.raytheon.com> Much thanks Chris. "Christopher A Leddy" wrote in message news:3B2656BC.E8A58F31@west.raytheon.com... > Try > > http://kernelnewbies.org/documents/kdoc/z8530book/z85230guide.html > > Source code is covered under GNU General License, but it is a good > as a reference. > > > Gerald Chang wrote: > > > > Hello All, > > > > I'm trying to use the i8237 DMA controller to send/recv data through the > > z85230 in the SDLC mode. > > > > Does anyone know where I can obtain any "C" drivers that support the DMA > > and synchronous modes of the z85230 and i8237A that are on the Motorola > > MVME2700 board? > > > > VxWorks supports only the async mode. > > > > Regards, > > Gerald Chang > > Lockheed Martin Corp. --------------------------- Newsgroups: comp.os.vxworks Subject: midas-20 Date: Tue, 12 Jun 2001 17:01:14 -0700 From: "Leung, Eugene" Organization: Lawrence Berkeley National Laboratory Message-ID: <9g6bak$97b$1@overload.lbl.gov> Has anyone written a vxworks driver to use this VME to PCI carrier? The board seems simple enough. Most of the work is done by the Universe II. However, I am having trouble writing to the PCI registers CF8 & CFC from the VME side to config the PCI bus. I've setup the VME slave images to access these in the io space. I can read from the register (even though I have no idea if what I read is meaningful), but write gives me bus errors. I am using Motorola MVME177 with vxworks 5.2. Hope someone can provide some pointers. Please send email to eugenel@condorsys.com as I am no longer on the exploder list. TIA. - -Eugene --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Network Stack Date: Wed, 13 Jun 2001 03:32:28 GMT From: "Vic Sperry" Message-ID: References: <9fmoai$4jj$1@woodrow.ucdavis.edu> "Jen-Chieh Tang" wrote in message news:9fmoai$4jj$1@woodrow.ucdavis.edu... > > Hello All, > > I am rather new to vxWorks and was wondering what the > community uses to debug the network stack. I am pretty > familiar with the stack code as I have worked with the > BSD stack before, but how does one debug in vxWorks ? > Thanks for any info. It's difficult to debug the stack because the debugger uses the stack. I've had some success disassembling the code while looking at FreeBSD source code, but this is obviously painful and slow. In my opinion, the vxWorks stack is not robust enough to be shipped sans source code. When using the stack for everyday debugging (over ethernet), I've had great success. But when trying to adapt the stack to a non-ethernet medium as part of my product's functionality, I've run into difficult problems that are time-consuming and painful to solve because the source code is not available. WRS should make the source code to the stack available to all. --------------------------- Newsgroups: comp.os.vxworks Subject: vxworks vs. threadx ? Date: Wed, 13 Jun 2001 09:24:15 +0100 From: Vijay Kumar Peshkar Organization: Cisco Systems Inc. Message-ID: <3B27232F.55E4803@wipro.com> This is a multi-part message in MIME format. - --------------189C95FAA1D45942D404CFAF Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hi Friends, Any URLs/documents comparing the two OSs?. Thanks and Regards, Vijay - --------------189C95FAA1D45942D404CFAF Content-Type: text/x-vcard; charset=us-ascii; name="vijay.peshkar.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Vijay Kumar Peshkar Content-Disposition: attachment; filename="vijay.peshkar.vcf" begin:vcard n:Peshkar;Vijay Kumar tel;work:5530034/35/35 x2237 x-mozilla-html:FALSE org:Wipro Global R & D;Wishing you blue skies, a couple of daisies, some fresh air and a smile! adr:;;;;;; version:2.1 email;internet:vpeshkar@cisco.com note:Wishing you blue skies, a couple of daisies, some fresh air and a smile! fn:Vijay Kumar Peshkar end:vcard - --------------189C95FAA1D45942D404CFAF-- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: problem with inet_aton Date: Wed, 13 Jun 2001 03:57:32 GMT From: "Vic Sperry" Message-ID: References: <3B24981B.2A9CF1F5@wipro.com> vxWorks is a modified BSD stack. rtrequest is there. I have had to use rtrequest on numerous occasions when the vxWorks API (e.g. routeAdd, mRoute...) failed to work properly. If I remember correctly, netmasks are passed to rtrequest as sockaddr's. "Michael Yeung" wrote in message news:by6V6.970$j6.80351@news... > Vijay, > Thanx for the info. rtrequest is a function support in BSD to add new > route. Rather than using routeAdd from VxWorks. > > Michael > > Vijay Kumar Peshkar wrote in message > news:3B24981B.2A9CF1F5@wipro.com... > > Michael, > > > > The macro OK has value 0 and in your case as everything seems to be > working > > fine, the func. call is returning OK ;). You should be comparing it with > ERROR(-1). > > > > The inet_addr func. is supported in vxWorks. > > Sorry I could not understand what's "rtrequest". > > > > cheers, > > Vijay > > > > Michael Yeung wrote: > > > > > Hi all, > > > I'm not sure why but the inet_aton seems to always returns 0 and > > > conversion is not done properly. > > > > > > struct in_addr temp; > > > > > > if(inet_aton("1.1.1.1", (struct in_addr *)&temp) == 0) > > > { > > > printf("Failed conversion\n"); > > > } > > > > > > I noticed VxWorks does not support "inet_addr" or "inet_pton", as > > > suggestion is appreciated. > > > > > > As well, does anyone knows the format expected for mask to function like > > > "rtrequest"? > > > > > > thanx in advance, > > > Michael > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: ARP cache Date: Wed, 13 Jun 2001 04:06:41 GMT From: "Vic Sperry" Message-ID: References: <2jhU6.643$j6.34686@news> rn_walktree() can be used to walk the route table. The ARP cache entries are routes with the proper flag set. I believe the flag is LL_INFO (link-layer info). "Michael Yeung" wrote in message news:2jhU6.643$j6.34686@news... > Hi all, > I'm looking for something in VxWorks that allows me to retrieve all > entries in the ARP cache. Something similar to what "sysctl" provides. I > found an API called "ip_sysctl" in netinet\ip_var.h but experience problem > when loading the object file to the target. Please let me know if you know > of anything that allows me to walk the ARP table. > > thanx, > Michael > > --------------------------- Newsgroups: comp.protocols.snmp,comp.os.vxworks Subject: Re: Extending the WindRiver (TMS) Envoy SNMP agent Date: Tue, 12 Jun 2001 23:03:42 -0500 From: Jim Jones Organization: netINS, Inc. Message-ID: <3B26E619.1F75FCD3@icon-labs.com> References: <3B25DA71.95D61C13@gaudlitz.com> Reply-To: jim_jones@icon-labs.com Michael, I am not familiar with your particular setup. However I am familiar with Envoy. The generated .c and .h files that you mention probably have filenames like mib.c, mibleaf.h, and mibskel.h and they are generated by the mibcomp argument - -mib.c [or none, the default], -leaf, and -skel.h, respectively. You may need to find the underlying makefile to discover the exact filename if not these. You should not have to manually change these three auto-generated files. Next, start looking into the generated mib.c file to see that you indeed have leaf-nodes initialized for your extended objects. If not then your agent will be clueless about these regardless of what other things you do. If you walk the agent without any timeouts but the agent does not give you the extended objects, or if you get no-such-object[name] when you directly GET one of the new objects, then this is probably the problem. The additional mibs did not get included on the command-line when mib.c was generated or you did not regenerate mib.c. Now, if you DO see the leaf-node data for new objects in mib.c, you'll find the access method routine names there as well (get, next, test, and set methods). So... find the .c file with those new methods. That file may have been generated by the -stub argument (perhaps in the makefile). If you see comments with !!!, that is a tip that you may need to modify the code for your implementation. Stub code is buggy... so it is doubtful that you are using it without knowing about it since you would not have gotten it to compile. Another possibility is that you have empty bodies to such methods (generated by -skel, an argument I never use... I much prefer the -stub arg). If those methods got called I would expect the agent to not respond at all and your manager will timeout rather than get a no-such-object[name] response. That is because the empty methods do not properly flag things as being done so the SNMP engine doesn't send a response (it presumes that you could have another thread that is working on this that will send a response when processing is complete). Let the -stub generated code and existing source code for your working method routines be your guide for writing new ones. Regards, Jim Jones Icon Labs Michael Gaudlitz wrote: > hi, > > any guys out there have experience on enxtending the SNMP agent that > comes with TMS? > > Actually it is an Envoy Agent but in TMS it is hidden behind some IDB+ > database stuff. So instead of running mibcomp you have to run idbgen on > your mib which also produces .c and .h. I also succeed in compiling and > integrating it in the VxWorks image. But any query to my (enterprise) > OID is not beeing answered. I suspect the error in missing code I have > to add by hand in the produced .c and .h files. But there is no good > documentation or example on that. > > Can anyone help me or point me to some docs? > > Thanks > > Michael --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks routing problems Date: Wed, 13 Jun 2001 05:08:16 GMT From: "Vic Sperry" Message-ID: <4zCV6.6812$v4.565655@paloalto-snr1.gtei.net> References: <3b262c2b.112640638@news.inter.net.il> "Ran Shalgi" wrote in message news:3b262c2b.112640638@news.inter.net.il... > We're developing a VxWorks-based router software, and we've come > across a rather unusual problem that has so far completely eludes us. > > In a Cisco router, you can assign a route directly to an ethernet > interface, without a default gateway. for example, the ethernet > interface is IP 192.168.1.2 subnet mask 255.255.0.0 and in addition > one can define a route of IP 172.16.32.0 mask 255.255.255.0 => > destination ethernet 0. This means that this new subnet is considered > directly connected to the ethernet port, and that port communicates > with those addresses in a "normal" form, by first ARPing them directly > and then sending the packets, without a gateway in between. If you're talking about a single interface, you could multi-home it: ifAddrSet("foo","192.168.1.2") ifAddrAdd("foo","172.16.32.2") To route to some host (1.1.1.1) behind a router (172.16.32.254) you could do this: routeAdd("1.1.1.1","172.16.32.254") This might get you what you need. The vxWorks API's only let you use one subnet mask per interface, though. It's not robust. > > In fact, what we need is even simpler: we want one ethernet interface > to have a very narrow IP range (i.e. 192.168.1.2 mask 255.255.255.255) > and in addition have ONE more machine on that LAN, e.g. to define a > route 10.0.0.1 mask 255.255.255.255 => destination ethernet 0. Two > disjoint IP addresses directly connected to the same ethernet port, > and only these two addresses. To me, this makes no sense for ethernet. All the nodes on an ethernet segment should be -- well -- on the same segment. But it makes total sense for a point-to-point link! I have configured vxWorks to have one interface that has assigned to it a number of different IP addresses with multiple "other ends" each with an arbitrary IP address (multi-homed point- to-point for frame-relay). This required me to abandon the vxWorks APIs in favor of the Berkeley APIs, and also to poke around with the interface data structs directly. It's UGLY, but it works. And the WRS people were no help AT ALL. I feel your pain. It looks to me like you could set up your "little" ethernet interface as a p2p link, if you really need to use disjoint IP addresses. I don't remember the particulars off-hand, but I could consult my code when I get to work, if you're interested (remove "splat." from E-mail address). vsperry@splat.paradyne.com > > We've not been able to get this to work. Sometimes the IP stack sends > the packets to the right interface, but the interface outputs them > with a destination MAC address of all 1s (broadcast address). In > slightly different configuration, the intherface outputs the packets > with destination MAC address equal to it's own source MAC address (?!) > which of course is no good. In other cases it seems the IP stack does > not forward packets at all to the i/f. > > Has anyone ever managed to get two disjoint subnets on the same > ethernet, or managed to get any kind of routing to be considered > "directly connected" to an ethernet port? WindRiver support doesn't > even understand the question, and we really need this, so any help > would be greatly appreciated. > > In case it matters, we're using x86 BSP, T2 WinNT hosts, VxWorks 5.4. > > -- > Ran Shalgi > > replace 127.0.0.1 with expand.com when replying by e-mail --------------------------- Newsgroups: comp.os.vxworks Subject: VxWorks : Mapping the Interrupt Levels Date: Wed, 13 Jun 2001 10:38:28 +0530 From: "Sanesh Kumar K" Organization: Lawrence Berkeley National Laboratory Message-ID: <9g6su2$iq2$1@overload.lbl.gov> Hi, In intLevelSet(int level), how the levels are mapped to atcual interrupt numbers ? Bye Sanesh Kumar K --------------------------- Newsgroups: comp.os.vxworks Subject: INT0 on x86 BSP Date: Wed, 13 Jun 2001 15:00:42 +0930 From: "David Hichens" Message-ID: <3b26faa3_5@news01.one.net.au> G'Day, I am using Vx-Works on a Pentium MMX equivalent PC-104 board using the x86 BSP. The code is servicing an interrupt on INT5 at a 5104 Hz rate. On examining the running system using WindView, there are seemingly random INT0 interrupt (timer tick) services occuring that take 30ms to execute. This is of course unexceptable in our system. The normal INT0 services are of the order of 18us. What could be causing these extended interrupt service times? Regards, David Hichens. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Where can one learn to use VxWorks Date: Wed, 13 Jun 2001 08:27:09 +0200 From: "Martin Raabe" Organization: T-Online Message-ID: <9g714l$orq$00$1@news.t-online.com> References: <3B210D50.CEA4CECD@Sun.COM> <9g5t36$t1h$1@front7.grolier.fr> Hello Alain, go for: http://www.WindRiver.com/prototyper Ciao Martin "Alain Bougeniere" schrieb im Newsbeitrag news:9g5t36$t1h$1@front7.grolier.fr... > where can you find this free tool ?... > > Thanks a lot! > Alain > > --------------------------- Newsgroups: comp.os.vxworks Subject: re: midas-20 Date: Wed, 13 Jun 2001 02:41:05 -0400 From: Brian Phelan Organization: Lawrence Berkeley National Laboratory Message-ID: <9g73ue$lfs$1@overload.lbl.gov> Eugene, I did a driver for the midas-20 (and midas-50) a few years ago. If you're trying to access the configuration registers on the PMC devices, what you want to do is to configure one of the 4 VME slave images to access PCI config space. You should then be able to access the config header of the devices in slots 1 and 2 at offsets 0x2800 and 0x3000 respectively, from the base address of the image. Hope this helps. Brian the vxWorks Users Group Exploder wrote: > > Submitted-by vxwexplo-errs@csg.lbl.gov Tue Jun 12 17:08:42 2001 > Submitted-by: "Leung, Eugene" > > Has anyone written a vxworks driver to use this VME to PCI carrier? The > board seems simple enough. Most of the work is done by the Universe II. > However, I am having trouble writing to the PCI registers CF8 & CFC from the > VME side to config the PCI bus. I've setup the VME slave images to access > these in the io space. I can read from the register (even though I have no > idea if what I read is meaningful), but write gives me bus errors. I am > using Motorola MVME177 with vxworks 5.2. Hope someone can provide some > pointers. Please send email to eugenel@condorsys.com as I am no longer on > the exploder list. TIA. > -Eugene - -- ======================================================= T h e P T R G r o u p, I n c. ======================================================= Brian Phelan (703)869-9066 brian@ThePTRGroup.com www.ThePTRGroup.com Embedded, Real-Time Solutions ======================================================= --------------------------- Newsgroups: comp.realtime,comp.os.vxworks,comp.os.psos,comp.os.qnx,comp.os.xinu Subject: Re: Priority inversion in semaphores Date: Wed, 13 Jun 2001 08:53:28 GMT From: Gord Deinstadt Organization: @Home Network Message-ID: <3B272916.1E3978B1@home.com> References: <4a6d848e.0106090954.514a3912@posting.google.com> amit choudhary wrote: > > To handle the priority inversion problem in semaphores, can it be > segregated into two types: > 1. To provide mutually exclusive access to critical section or a > resource, and > 2. For producer-consumer kind of synchronization, > Or a single abstraction can be used to handle priority inversion > completely. Depends what you mean by completely. Most RTOSes don't prevent priority inversion for every case of producer-consumer synchronization. For example Neutrino (from QNX) uses separate synchronization mechanisms for critical sections (semaphores) and producer-consumer (synchronous messages). It prevents priority inversion only so long as applications use a many-to-one messaging model. As soon as an app uses many-to-many (via a Posix queue) there is no more prevention of priority inversion. Many-to-many is inherently difficult because the kernel has no way to know which listener might be ready next, so all it could do would be to raise the priority of all potential listeners (and the processes upon which they are waiting) which would often result in a logjam as every process was raised to the same priority. For an example of a very small single-mechanism kernel that handles priority inversion as well as Neutrino, see my nanokernel page at http://www.thersitesconsulting.com/ (The nanokernel is designed for use in embedded systems. On a COP-8, which is an 8-bit MCU, it occupies 201 bytes of ROM.) - -- Gord Deinstadt gord@thersitesconsulting.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: New Altivec-enhanced VxWorks Gcc 2.95.4 compiler released! Date: Tue, 12 Jun 2001 16:16:51 +0100 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: <3YFV6.1$g4.3022@newsr2.u-net.net> References: <5btU6.24100$YB3.4024200@news2-win.server.ntlworld.com> <9g2b4t$91q$1@qnsgh006.europe.nortel.com> Mathieu Imrazene wrote in message <9g2b4t$91q$1@qnsgh006.europe.nortel.com>... >Does anyone has examples of performance improvements by using the Altivec >compiler ? In others words, it is really useful to upgrade from the 2.95.3 >version to the 2.95.4 + Altivec version ? Hi Mathieu, The answer is that it won't make any difference unless you specifically rewrite your application to make use of the altivec unit. There are no features in GCC to automatically vectorize scalar code; what you get are a few new builtin types (vector int, float, short, char etc.) plus a whole slew of builtin functions (all beginning vec_, as specified in the Motorola Altivec Programming Interface Manual. However the speedup you can get by rewriting the inner loops of multimedia or number crunching applications are very great indeed; 4x faster is easily achievable. DaveK - -- They laughed at Galileo. They laughed at Copernicus. They laughed at Columbus. But remember, they also laughed at Bozo the Clown. --------------------------- Newsgroups: comp.os.vxworks Subject: re: clockLib problems Date: Wed, 13 Jun 2001 09:35:38 +0100 (BST) From: David Laight Organization: Lawrence Berkeley National Laboratory Message-ID: <200106130835.JAA26021@dsl-2.tadpole.co.uk> > I read few postings, describing different problems with clockLib - > timer gets reset periodically, going backwards and so on. > I have basicly the same problem - time() increments to some seconds, > gets reset to zero, and after a that while again the same. > I tried to debug, windTickAnnounce, incrementing the vxTicks, but > problem seems to be complicated. I looked into this a few months ago. It appears that vxWorks doesn't believe in real time! The underlying problem is the way vxworks handles the tick counter wrapping. What I found was that if ANY timeout call (eg taskDelay, semTake) had an expiry time which was later than the time at which the tick counter wrapped, then the tick counter was reset to zero and all the pending timeouts adjusted. However the offset isn't transferred to the code that supports time() - so time() leaps backwards. This is easily reproducable by calling 'sp taskDelay, -2' from the shell. David - ---------------------------------------------------------------- David Laight email: dsl@tadpole.co.uk Tadpole Technology plc phone: +44 1223 428 232 Cambridge, UK fax: +44 1223 428 201 --------------------------- Newsgroups: comp.os.vxworks Subject: svgmd Build Boot rom Date: 13 Jun 2001 02:52:47 -0700 From: sbhushan@asu.edu (Bhushan) Organization: http://groups.google.com/ Message-ID: <83beda69.0106130152.60ef6ba9@posting.google.com> Hi, I am trying to build a bootable application for Synergy VGMD card through tornado. But when i go to 'Build Boot ROM' under the Build menu, and select svgmd as BSP and bootrom as its image and say ok, i get the following ERROR .. but is working fine for all other BSPs. ==================================================================== In file included from ..\synergy\net\esyf885End.cc:117, from esyf885End.cc:1: D:\Tornado\target\h\m2Lib.h:1024: syntax error before `,' D:\Tornado\target\h\m2Lib.h:1025: syntax error before `,' D:\Tornado\target\h\m2Lib.h:1026: syntax error before `,' D:\Tornado\target\h\m2Lib.h:1027: syntax error before `,' ..\synergy\net\esyf885End.cc: In function `void esyfRecv(struct Esyf *)': In file included from esyf885End.cc:1: ..\synergy\net\esyf885End.cc:1295: too many arguments to function make.exe: *** [esyf885End.o] Error 0x1 ==================================================================== could any one please help me out in this . Thanks Bhushan --------------------------- Newsgroups: comp.protocols.snmp,comp.os.vxworks Subject: Re: Extending the WindRiver (TMS) Envoy SNMP agent Date: 13 Jun 2001 03:07:58 -0700 From: mathi@india.adventnet.com (V. Mathivanan) Organization: http://groups.google.com/ Message-ID: <30830a82.0106130207.7780b3e7@posting.google.com> References: <3B25DA71.95D61C13@gaudlitz.com> <3B26E619.1F75FCD3@icon-labs.com> Hi, AdventNet C Edition provides complete code generation for SNMP agent development. User need not to write any code for SNMP functionality. You can download the evaluation copy at http://www.adventnet.com Thanks Mathi Jim Jones wrote in message news:<3B26E619.1F75FCD3@icon-labs.com>... > Michael, > I am not familiar with your particular setup. However I am > familiar with Envoy. The generated .c and .h files that you > mention probably have filenames like mib.c, mibleaf.h, and > mibskel.h and they are generated by the mibcomp argument > -mib.c [or none, the default], -leaf, and -skel.h, respectively. > You may need to find the underlying makefile to discover > the exact filename if not these. You should not have to > manually change these three auto-generated files. > > Next, start looking into the generated mib.c file to see > that you indeed have leaf-nodes initialized for your extended > objects. If not then your agent will be clueless about these > regardless of what other things you do. If you walk the > agent without any timeouts but the agent does not give you > the extended objects, or if you get no-such-object[name] > when you directly GET one of the new objects, then this is > probably the problem. The additional mibs did not get > included on the command-line when mib.c was generated > or you did not regenerate mib.c. > > Now, if you DO see the leaf-node data for new objects in > mib.c, you'll find the access method routine names there as > well (get, next, test, and set methods). So... find the > .c file with those new methods. That file may > have been generated by the -stub argument (perhaps > in the makefile). If you see comments with !!!, that is a tip > that you may need to modify the code for your implementation. > Stub code is buggy... so it is doubtful that you are using it > without knowing about it since you would not have gotten > it to compile. Another possibility is that you have > empty bodies to such methods (generated by -skel, an > argument I never use... I much prefer the -stub arg). If > those methods got called I would expect the agent to not > respond at all and your manager will timeout rather than > get a no-such-object[name] response. That is because the > empty methods do not properly flag things as being done > so the SNMP engine doesn't send a response (it presumes > that you could have another thread that is working on this > that will send a response when processing is complete). > Let the -stub generated code and existing source code > for your working method routines be your guide for > writing new ones. > > Regards, > Jim Jones > Icon Labs > > Michael Gaudlitz wrote: > > > hi, > > > > any guys out there have experience on enxtending the SNMP agent that > > comes with TMS? > > > > Actually it is an Envoy Agent but in TMS it is hidden behind some IDB+ > > database stuff. So instead of running mibcomp you have to run idbgen on > > your mib which also produces .c and .h. I also succeed in compiling and > > integrating it in the VxWorks image. But any query to my (enterprise) > > OID is not beeing answered. I suspect the error in missing code I have > > to add by hand in the produced .c and .h files. But there is no good > > documentation or example on that. > > > > Can anyone help me or point me to some docs? > > > > Thanks > > > > Michael --------------------------- Newsgroups: comp.protocols.snmp,comp.os.vxworks Subject: Re: Extending the WindRiver (TMS) Envoy SNMP agent Date: 13 Jun 2001 03:07:57 -0700 From: mathi@india.adventnet.com (V. Mathivanan) Organization: http://groups.google.com/ Message-ID: <30830a82.0106130207.3cc4e2dc@posting.google.com> References: <3B25DA71.95D61C13@gaudlitz.com> <3B26E619.1F75FCD3@icon-labs.com> Hi, AdventNet C Edition provides complete code generation for SNMP agent development. User need not to write any code for SNMP functionality. You can download the evaluation copy at http://www.adventnet.com Thanks Mathi Jim Jones wrote in message news:<3B26E619.1F75FCD3@icon-labs.com>... > Michael, > I am not familiar with your particular setup. However I am > familiar with Envoy. The generated .c and .h files that you > mention probably have filenames like mib.c, mibleaf.h, and > mibskel.h and they are generated by the mibcomp argument > -mib.c [or none, the default], -leaf, and -skel.h, respectively. > You may need to find the underlying makefile to discover > the exact filename if not these. You should not have to > manually change these three auto-generated files. > > Next, start looking into the generated mib.c file to see > that you indeed have leaf-nodes initialized for your extended > objects. If not then your agent will be clueless about these > regardless of what other things you do. If you walk the > agent without any timeouts but the agent does not give you > the extended objects, or if you get no-such-object[name] > when you directly GET one of the new objects, then this is > probably the problem. The additional mibs did not get > included on the command-line when mib.c was generated > or you did not regenerate mib.c. > > Now, if you DO see the leaf-node data for new objects in > mib.c, you'll find the access method routine names there as > well (get, next, test, and set methods). So... find the > .c file with those new methods. That file may > have been generated by the -stub argument (perhaps > in the makefile). If you see comments with !!!, that is a tip > that you may need to modify the code for your implementation. > Stub code is buggy... so it is doubtful that you are using it > without knowing about it since you would not have gotten > it to compile. Another possibility is that you have > empty bodies to such methods (generated by -skel, an > argument I never use... I much prefer the -stub arg). If > those methods got called I would expect the agent to not > respond at all and your manager will timeout rather than > get a no-such-object[name] response. That is because the > empty methods do not properly flag things as being done > so the SNMP engine doesn't send a response (it presumes > that you could have another thread that is working on this > that will send a response when processing is complete). > Let the -stub generated code and existing source code > for your working method routines be your guide for > writing new ones. > > Regards, > Jim Jones > Icon Labs > > Michael Gaudlitz wrote: > > > hi, > > > > any guys out there have experience on enxtending the SNMP agent that > > comes with TMS? > > > > Actually it is an Envoy Agent but in TMS it is hidden behind some IDB+ > > database stuff. So instead of running mibcomp you have to run idbgen on > > your mib which also produces .c and .h. I also succeed in compiling and > > integrating it in the VxWorks image. But any query to my (enterprise) > > OID is not beeing answered. I suspect the error in missing code I have > > to add by hand in the produced .c and .h files. But there is no good > > documentation or example on that. > > > > Can anyone help me or point me to some docs? > > > > Thanks > > > > Michael --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Wed Jun 13 20:55:24 2001 From: "G.Gopal Krishnam Raju" Date: Wed Jun 13 20:55:26 PDT 2001 Subject: regarding TCP/IP APIs ----- Original Message ----- From: G.Gopal Krishnam Raju To: the vxWorks Users Group Exploder Sent: Wednesday, June 13, 2001 6:25 AM Subject: regarding TCP/IP APIs > hello Gurus > > Sorry for troubling u all > > After getting the packet from driver (ene/elt....). What function to be > called to pass the packet from driver to tcp/ip stack where can i get this > information. Can i get networking APIs documentations. > > My problem is i have some interface which will give ethernet packet. I want > to pass to tcp/ip stack using what API i can do this. > > Thanks in Advance > > reply ASAP > > Regards > > > GGKRaju > > > From vxwexplo-errs@csg.lbl.gov Wed Jun 13 23:20:24 2001 From: "Deepesh Aggarwal" Date: Wed Jun 13 23:20:26 PDT 2001 Subject: Regarding Sending a Packet after a fixed interval of time This is a multi-part message in MIME format. ------=_NextPart_000_00E4_01C0F4C7.E0559BF0 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="iso-8859-1" Hi All I want to "send multicast Packets after a specified fixed interval of = time repeatedly" . So to serve my purpose i write a program whose code looks like .. *************************************************************************= ****************** #include "vxWorks.h" #include "taskLib.h" #include "socket.h" #include "netinet/in.h" #include "stdio.h" #include "stdlib.h" #include "string.h" #include "sockLib.h" #include "inetLib.h" #include "ioLib.h" #include "routeLib.h" #include "logLib.h" #include "wdLib.h" #define SECONDS (15) WDOG_ID myWatchDogId; struct sockaddr_in sin; struct sockaddr_in toAddr; int toAddrLen; int sockDesc; char * sendBuf=3D"Hello"; int sendLen =3Dsizeof(sendBuf); int len; STATUS trial(void); STATUS mcastSend (char * mcastAddr, USHORT mcastPort); STATUS multi(void) { routeAdd("224.0.0.0","157.227.2.1"); taskSpawn ("maintask", 100, 0, 20000,mcastSend,"224.0.0.23" = ,"8769",0,0,0,0,0,0,0,0); return (OK); } =20 STATUS mcastSend(char * mcastAddr, USHORT mcastPort) { =20 /* create a send and recv socket */ if ((sockDesc =3D socket (AF_INET, SOCK_DGRAM, 0)) < 0 ) { printf (" cannot open send socket\n"); return (ERROR); } /* zero out the structures */ bzero ((char *)&sin, sizeof (sin)); bzero ((char *)&toAddr, sizeof (toAddr)); sin.sin_family =3D AF_INET; sin.sin_addr.s_addr =3D INADDR_ANY; sin.sin_port =3D htons(0); if (bind(sockDesc, (struct sockaddr *)&sin, sizeof(sin)) !=3D 0) { perror("bind"); if (sockDesc) close (sockDesc); return (ERROR); } toAddrLen =3D sizeof(struct sockaddr_in); toAddr.sin_family =3D AF_INET; /* initialize the address to the send */ toAddr.sin_addr.s_addr =3D inet_addr (mcastAddr); printf("addr is %d",toAddr.sin_addr.s_addr); /* initialize the port to send */ toAddr.sin_port =3D htons(mcastPort); if ((myWatchDogId =3D wdCreate( )) =3D=3D NULL) {=20 perror("Error in creating Timer\n"); return (ERROR); } if (wdStart (myWatchDogId, sysClkRateGet( ) * SECONDS,trial,0) =3D=3D = ERROR) { perror("Error in starting Timer\n"); =20 return(ERROR); } } STATUS trial(void) { /*printf ("sending \n"); printf ("Sending %s of length %d using Socket descriptor = %d\n",sendBuf,sendLen,sockDesc );*/ if ((len =3D sendto (sockDesc, sendBuf, sendLen, 0,(struct sockn = *)&toAddr, toAddrLen)) < 0 ) { printf("mcastSend sendto errno:0x%x\n", errno ); } if (wdStart (myWatchDogId, sysClkRateGet( ) * SECONDS,trial,0) =3D=3D = ERROR) { return(ERROR); } } *************************************************************************= ********************** I downloaded the application on target with the help of target server = and tried to call "multi" , it works fine . i mean it was sending the = packet after 15 seconds ..but=20 after sending two or three packets my target server gets killed and when = again i tried to launch it ....it gives an error WTX_ERROR ( Connection = busy ..). but i didn't get any error on target .and target was still sending the = packets ( That's ok .that's the behavior i want..). .until i explicitly = shut it down... I m using P-III as a target with an ethernet card FEI ( Fast Intel Ether = Express PRO100B PCI ). Can anyone please explain me the reason why Target server is behaving = like that.? and where and what can be the problem..?....or suggest me = some better way to do the same... =20 Eagerly waiting for ur precious advice Thanks n Happy Working=20 Deepesh ------=_NextPart_000_00E4_01C0F4C7.E0559BF0 Content-Transfer-Encoding: quoted-printable Content-Type: text/html; charset="iso-8859-1"
Hi All
 
I want to = "send multicast Packets=20 after a specified fixed interval of time repeatedly" = .
So to serve my purpose i write a = program whose=20 code looks like ..
 
****************************************************************= ***************************
 
#include "vxWorks.h"
#include=20 "taskLib.h"
#include "socket.h"
#include = "netinet/in.h"
#include=20 "stdio.h"
#include "stdlib.h"
#include "string.h"
#include=20 "sockLib.h"
#include "inetLib.h"
#include "ioLib.h"
#include=20 "routeLib.h"
#include "logLib.h"
#include "wdLib.h"
 
#define SECONDS (15)
 
WDOG_ID myWatchDogId;
 
struct sockaddr_in sin;
struct = sockaddr_in=20 toAddr;
int toAddrLen;
int sockDesc;
char * = sendBuf=3D"Hello";
int=20 sendLen =3Dsizeof(sendBuf);
int len;
 
STATUS trial(void);
STATUS mcastSend = (char *=20 mcastAddr, USHORT mcastPort);
STATUS=20 multi(void)
{
 routeAdd("224.0.0.0","157.227.2.1");
 t= askSpawn=20 ("maintask", 100, 0, 20000,mcastSend,"224.0.0.23"=20 ,"8769",0,0,0,0,0,0,0,0);
 return = (OK);
}
 
STATUS mcastSend(char * mcastAddr, USHORT=20 mcastPort)
{
 
/* create a send and recv socket */
if=20 ((sockDesc =3D socket (AF_INET, SOCK_DGRAM, 0)) < 0 )
{
printf = (" cannot=20 open send socket\n");
return (ERROR);
}
 
/* zero out the structures */
bzero = ((char=20 *)&sin, sizeof (sin));
bzero ((char *)&toAddr, sizeof=20 (toAddr));
 
sin.sin_family =3D = AF_INET;
sin.sin_addr.s_addr =3D=20 INADDR_ANY;
sin.sin_port =3D htons(0);
 
if (bind(sockDesc, (struct sockaddr = *)&sin,=20 sizeof(sin)) !=3D 0)
{
perror("bind");
if (sockDesc)
close=20 (sockDesc);
return (ERROR);
}
 
toAddrLen =3D sizeof(struct=20 sockaddr_in);
toAddr.sin_family =3D AF_INET;
/* initialize the = address to=20 the send */
toAddr.sin_addr.s_addr =3D inet_addr = (mcastAddr);
 
printf("addr is = %d",toAddr.sin_addr.s_addr);
/*=20 initialize the port to send */
toAddr.sin_port =3D=20 htons(mcastPort);
 
if ((myWatchDogId =3D wdCreate( )) = =3D=3D NULL)
{=20
perror("Error in creating Timer\n");
return = (ERROR);
}
 
if (wdStart (myWatchDogId, = sysClkRateGet( ) *=20 SECONDS,trial,0) =3D=3D ERROR)
{
perror("Error in starting = Timer\n"); =20
return(ERROR);
}
 
}
 

STATUS trial(void)
{
/*printf = ("sending=20 \n");
printf ("Sending %s of length %d using Socket descriptor=20 %d\n",sendBuf,sendLen,sockDesc );*/
 
if ((len =3D sendto (sockDesc, sendBuf, = sendLen,=20 0,(struct sockn *)&toAddr, toAddrLen)) < 0 = )
{
printf("mcastSend=20 sendto errno:0x%x\n", errno );
}
 
if (wdStart (myWatchDogId, = sysClkRateGet( ) *=20 SECONDS,trial,0) =3D=3D ERROR)
{
return(ERROR);
}
 
}

************************************************************= ***********************************
 
I downloaded the application on target = with the=20 help of target server and tried to call "multi" , it works fine . i mean = it was=20 sending the packet after 15 seconds ..but
after sending two or three packets my = target server=20 gets killed and when again i tried to launch it ....it gives an error = WTX_ERROR=20 ( Connection busy ..).
but i didn't get any error on target=20 .and target was still sending the packets ( That's ok .that's the = behavior=20 i want..). .until i explicitly shut it down...
I m using P-III as a target with an = ethernet card=20 FEI ( Fast Intel Ether Express PRO100B PCI ).
 
Can anyone please explain me the reason = why Target=20 server is behaving like that.? and where and what can be the = problem..?....or=20 suggest me some better way to do the same...
 
Eagerly waiting for ur precious = advice
 
Thanks n Happy=20 Working
Deepesh
 
------=_NextPart_000_00E4_01C0F4C7.E0559BF0-- From vxwexplo-errs@csg.lbl.gov Thu Jun 14 04:03:31 2001 From: Vxworks Exploder Date: Thu Jun 14 04:03:34 PDT 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Thu Jun 14 04:03:08 PDT 2001 Subject: Re: DP national driver driver for VxWorks Subject: Re: Proxy ARP problems: workaround needed Subject: loadModule and memDrv Subject: Re: loadModule and memDrv Subject: Re: loadModule and memDrv Subject: Re: loadModule and memDrv Subject: RTOS performance Subject: Re: loadModule and memDrv Subject: Re: VxWorks routing problems Subject: Re: problem with inet_aton Subject: Re: loadModule and memDrv Subject: Re: problem with inet_aton Subject: Re: DP national driver driver for VxWorks Subject: module download from command prompt Subject: Boot rom foe VX works on tornado Subject: Re: DP national driver driver for VxWorks Subject: Re: DP national driver driver for VxWorks Subject: Re: DP national driver driver for VxWorks Subject: USB driver format Subject: what is the interrupt vector associated with a VME bus error? Subject: p2v Subject: Re: MAX Timers in VxWroks ! Subject: How to create network gateway via PPP... Subject: Re: VxWorks drivers for z85230 synchronous and DMA modes of operation Subject: gdb and image tool questions Subject: AIC-7899 Subject: Re: How to create network gateway via PPP... Subject: Re: gdb and image tool questions Subject: VxSim x86 availability? Subject: Re: Priority inversion in semaphores Subject: regarding TCP/IP APIs Subject: regarding TCP/IP APIs Subject: Re: PPC: system tick routine stack usage Subject: Regarding Sending a Packet after a fixed interval of time Subject: Re: VxSim x86 availability? Subject: Problem in Performance monitoring Subject: tWdbTask stack overflow Subject: Booting from ATA disk ... crash Subject: VCOMM problem ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: DP national driver driver for VxWorks Date: Wed, 13 Jun 2001 09:34:11 -0400 From: Ian Melhuish Organization: Teklogix Inc. Message-ID: <3B276BD2.88CF4514@teklogix.com> References: I am using the driver and also cannot get a target server connected. Did your problem arise after installong the 2.02 patch? Mike Kordik wrote: > Is anyone using this driver (DP83815)? The target server will not connect to > this driver after you install the Tornado 2.02 patch. Does anyone know why? > Is there a workaround? > > Thanks, > Mike --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Proxy ARP problems: workaround needed Date: Wed, 13 Jun 2001 17:56:13 +0400 From: Pavel Kovalev Organization: Gamma NNTP server Moscow Russia Message-ID: <3B2770FD.C9BE0382@auriga.ru> References: <3B211164.3AA58342@auriga.ru> <3B268599.3E163DD4@raytheon.com> Hello Bill, > The solution that we came up with was to use the ATF_PUBL flag with the > arpAdd() function. > This seems to, in effect, have all the desired consequences - our card > responds to an ARP request for the defined IP address with its own MAC > address, and then routes the IP message to the appropriate place. > > This requires a manual setup of the ARP table ... but it seems to work > well. (Of course, you have to have routing enabled.) > > In our case, though, we were forwarding messages between ethernet > networks, not a network and a backplane. It's my understanding that the > routing should still work, but .... Our backplane network fully emulates the ethernet network, except support for multicasting. So, in theory, what works in your case should work for us. Though, when use arpAdd() I get the following: - -> ifShow dc (unit number 0): Flags: (0x8063) UP BROADCAST MULTICAST ARP RUNNING Type: ETHERNET_CSMACD Internet address: 192.168.1.154 Broadcast address: 192.168.1.255 Netmask 0xffffff00 Subnetmask 0xffffff00 Ethernet address is 08:00:3e:2d:80:00 Metric is 0 Maximum Transfer Unit size is 1500 6107 packets received; 864 packets sent 4683 multicast packets received 1 multicast packets sent 0 input errors; 0 output errors 0 collisions; 0 dropped lo (unit number 0): Flags: (0x8069) UP LOOPBACK MULTICAST ARP RUNNING Type: SOFTWARE_LOOPBACK Internet address: 127.0.0.1 Netmask 0xff000000 Subnetmask 0xff000000 Metric is 0 Maximum Transfer Unit size is 32768 0 packets received; 0 packets sent 0 multicast packets received 0 multicast packets sent 0 input errors; 0 output errors 0 collisions; 0 dropped mn (unit number 1): Flags: (0x63) UP BROADCAST ARP RUNNING Type: ETHERNET_CSMACD Internet address: 192.168.1.155 Broadcast address: 192.168.1.255 Netmask 0xffffff00 Subnetmask 0xffffff00 Ethernet address is 22:22:22:00:00:01 Metric is 0 Maximum Transfer Unit size is 4058 2 packets received; 1 packets sent 2 multicast packets received 0 multicast packets sent 0 input errors; 0 output errors 0 collisions; 0 dropped - -> arpAdd "192.168.1.151", "22:22:22:00:00:04", 8 value = 0 = 0x0 - -> arpShow LINK LEVEL ARP TABLE destination gateway flags Refcnt Use Interface - -------------------------------------------------------------------------- 192.168.1.151 22:22:22:00:00:4 4c05 0 0 dc0 192.168.1.253 00:01:02:1e:13:4 405 0 851 dc0 - -------------------------------------------------------------------------- value = 75 = 0x4b = 'K' - -> routeShow ROUTE NET TABLE destination gateway flags Refcnt Use Interface - ---------------------------------------------------------------------------- 192.168.1.0 192.168.1.154 101 0 0 dc0 - ---------------------------------------------------------------------------- ROUTE HOST TABLE destination gateway flags Refcnt Use Interface - ---------------------------------------------------------------------------- 127.0.0.1 127.0.0.1 5 0 0 lo0 - ---------------------------------------------------------------------------- value = 77 = 0x4d = 'M' - -> ping "192.168.1.151" PING 192.168.1.151: 56 data bytes no answer from 192.168.1.151 value = -1 = 0xffffffff = udbinfo + 0xffd7666b Obviously, when I call arpAdd() as shown above I do not get what I need: 192.161.1.151 resides on the proxy network (mn1 - CompactPCI), and the ARP entry is associated with the ethernet network (dc0). Therefore, to achieve the desired configuration I use the following commands: - -> ifAddrSet "dc0", "127.127.127.127" value = 0 = 0x0 - -> ifAddrSet "mn1", "192.168.1.155" value = 0 = 0x0 - -> arpAdd "192.168.1.151", "22:22:22:00:00:04", 8 value = 0 = 0x0 - -> routeDelete "192.168.1.0", "192.168.1.155" value = 0 = 0x0 - -> mRouteAdd "192.168.1.151", "192.168.1.155", 0xffffffff, 0, 0 value = 0 = 0x0 - -> ifAddrSet "dc0", "192.168.1.154" value = 0 = 0x0 - -> routeShow ROUTE NET TABLE destination gateway flags Refcnt Use Interface - ---------------------------------------------------------------------------- 192.168.1.0 192.168.1.154 101 0 0 dc0 192.168.1.151 192.168.1.155 3 0 0 mn1 - ---------------------------------------------------------------------------- ROUTE HOST TABLE destination gateway flags Refcnt Use Interface - ---------------------------------------------------------------------------- 127.0.0.1 127.0.0.1 5 0 0 lo0 - ---------------------------------------------------------------------------- value = 77 = 0x4d = 'M' - -> arpShow LINK LEVEL ARP TABLE destination gateway flags Refcnt Use Interface - -------------------------------------------------------------------------- 192.168.1.151 22:22:22:00:00:4 4c05 0 0 mn1 - -------------------------------------------------------------------------- value = 75 = 0x4b = 'K' Though, while now the configuration seems to be absolutely correct, VxWorks does not reply to ARP requests originated from one network and targeted to another network. Thus, 192.168.1.253 (residing on Ethernet) and 192.168.1.151 do not get ARP replies when they want to communicate with each other. If I setup ARP tables manually on these hosts then they can successfully ping each other; though, this is not what I need. The proxy ARP server on 192.168.1.154/155 does not seem to work at all. I am wondering what is the difference between our configurations. From your message it appears that the same setup works on your system. Am I missing something? I am using VxWorks 5.4/Tornado II. Do you have some special patches applied, or something like that? Thanks, Pavel. > Bill > > Pavel Kovalev wrote: > > > Hello All, > > > > We have a custom network driver for a CompactPCI backplane network > > and want to configure a proxy ARP server on the CompactPCI host > > board to make the backplane nodes visible on the Ethernet network > > (without additional routing). > > > > Tornado 2/VxWorks 5.4 support this configuration only for > > the shared memory network drivers supplied by WRS. Attempts > > to use proxyArpLib or configure all the necessary routing > > manually fail with various errors. > > > > I studied various materials available on WindSurf but haven't > > found a solution for this problem. There is an official SPR > > describing this problem (SPR# 28602), but it looks like it > > hasn't been processed/fixed yet. There were some discussions > > of this problem in the newsgroup in the past, but no good > > solutions have been proposed. > > > > I spent some time investigating this problem, and found that when > > I add a specific route to a node residing on the proxy network > > (route is added on the host which has to work proxy ARP server), > > mRouteAdd() creates the route with RTF_GATEWAY set. This prevents > > arpAdd() from adding a permanent published entry for the node. > > One solution is to develop a better version of mRouteAdd(), but > > I am not yet ready/willing to do so - there is no warranty that > > other hacks won't be needed. > > > > So, I am wondering if anybody has managed to solve this problem. > > > > Thanks, > > Pavel. --------------------------- Newsgroups: comp.os.vxworks Subject: loadModule and memDrv Date: Wed, 13 Jun 2001 10:12:09 -0400 From: "Daniel G Waddington" Organization: Bell Laboratories, Lucent Technologies Message-ID: <9g7sdc$c523@news.research.bell-labs.com> Hi, can anyone tell me why I can't get loadModule to load from a memDrv I/O device? I'm trying to load a module from which I have the file image in memory. Please help if you can. Thanks, DAN. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: loadModule and memDrv Date: Wed, 13 Jun 2001 07:41:15 -0700 From: "Michael R. Kesti" Organization: MK Associates Message-ID: <3B277B8B.E64E1EE7@gv.net> References: <9g7sdc$c523@news.research.bell-labs.com> Reply-To: mkesti@gv.net Daniel G Waddington wrote: > > Hi, can anyone tell me why I can't get loadModule to load from a memDrv I/O > device? I'm trying to load a module from which I have the file image in > memory. Please help if you can. I'd be willing to help if I can, but I can't without some idea of your environment, the code you have tried, and the errors you have experienced. - -- ======================================================================== Michael Kesti | "And like, one and one don't make | two, one and one make one." mkesti@gv.net | - The Who, Bargain --------------------------- Newsgroups: comp.os.vxworks Subject: Re: loadModule and memDrv Date: 13 Jun 2001 11:48:20 -0400 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: References: <9g7sdc$c523@news.research.bell-labs.com> Sender: bpringlemeir@DeadDuck Daniel> Hi, can anyone tell me why I can't get loadModule to load Daniel> from a memDrv I/O device? I'm trying to load a module from Daniel> which I have the file image in memory. Please help if you Daniel> can. Are you calling `loadModule' from code that is loaded on the target or from the shell. There are two versions of `loadModule'. One is fixed up on the host and the other is fixed up on the target. In order for you to load on the target by calling `loadModule' from a program, you must have the symbol library available and all the vxWorks symbols included. Try this on your target, - -> @lkup "SymTbl" standAloneSymTbl 0x002a90d8 bss statSymTbl 0x002a8684 bss sysSymTbl 0x002a91b0 bss value = 0 = 0x0 - -> lkup "SymTbl" _loginSymTbl 0x002a0880 bss (vxWorks.ram) _standAloneSymTbl 0x002a90d8 bss (vxWorks.ram) _statSymTbl 0x002a8684 bss (vxWorks.ram) _sysSymTbl 0x002a91b0 bss (vxWorks.ram) value = 0 = 0x0 - -> - -> standAloneSymTbl _standAloneSymTbl = 0x2a90d8: value = 0 = 0x0 - -> - -> statSymTbl _statSymTbl = 0x2a8684: value = 0 = 0x0 - -> - -> sysSymTbl _sysSymTbl = 0x2a91b0: value = 8105904 = 0x7bafb0 - -> - -> At least one of the ".*SymTbl" variables should be set to a value. If not, you don't have symbol tables in your image. Preceeding shell commands with `@' makes them execute on the target, instead of the host. hth, Bill Pringlemeir. - -- So, every young person is a peasant? You make me sick. Useful vxWorks URLS, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" "http://www.xs4all.nl/~borkhuis/vxworks/vxw_pt9.html" --------------------------- Newsgroups: comp.os.vxworks Subject: Re: loadModule and memDrv Date: Wed, 13 Jun 2001 11:48:28 -0400 From: "Daniel G Waddington" Organization: Bell Laboratories, Lucent Technologies Message-ID: <9g821u$c524@news.research.bell-labs.com> References: <9g7sdc$c523@news.research.bell-labs.com> <3B277B8B.E64E1EE7@gv.net> I'm using VxWorks 5.4 on the Intel platform. After loading the file image into memory and setting up a pseudo IO memory device (memDrv) for it, I then open the device and use loadModule on the handle. I get all sorts of errors like, "loadAoutLib error : insufficient memory strings table for 823737483839 bytes". I pretty sure the file image in memory is sound. DAN --------------------------- Newsgroups: comp.os.vxworks Subject: RTOS performance Date: Wed, 13 Jun 2001 11:28:23 -0500 From: "Ben Sun" Message-ID: Hi there! I am in the process of comparing VxWorks and a non-RTOS such as Solaris for some CPU board on which both OS is available. Could somebody give me some idea? For example, what is the OS time difference for receiving an IP packet and copy it to user space using the same CPU board? Any help, numbers, hints would be greatly appreciated. Thank you all very much. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: loadModule and memDrv Date: Wed, 13 Jun 2001 12:35:16 -0400 From: "Daniel G Waddington" Organization: Bell Laboratories, Lucent Technologies Message-ID: <9g84pn$c525@news.research.bell-labs.com> References: <9g7sdc$c523@news.research.bell-labs.com> I calling loadModule from another module loaded on the target. I've don what you said and get.. - -> @lkup "SymTbl" wtxSymTblInfoGet 0x001942c8 text (vxWorks.sym) statSymTbl 0x001c5c58 bss (vxWorks.sym) loadSymTbl 0x00111b58 text (vxWorks.sym) netLoadSymTbl 0x00111a38 text (vxWorks.sym) sysSymTbl 0x001c6254 bss (vxWorks.sym) value = 0 = 0x0 - -> lkup "SymTbl" _wtxSymTblInfoGet 0x001942c8 text (vxWorks) _statSymTbl 0x001c5c58 bss (vxWorks) _loadSymTbl 0x00111b58 text (vxWorks) _netLoadSymTbl 0x00111a38 text (vxWorks) _sysSymTbl 0x001c6254 bss (vxWorks) value = 0 = 0x0 - -> which appears to mean the symbols are there and happy. I can run loadModule on the target console. DAN --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks routing problems Date: 13 Jun 2001 12:50:56 -0400 From: vanbaren@falcon.si.com (vanbaren_gerald) Organization: Smiths Industries Message-ID: <3b2799f0$1@news.si.com> References: <3b262c2b.112640638@news.inter.net.il> Ran@127.0.0.1 (Ran Shalgi) writes: >We're developing a VxWorks-based router software, and we've come >across a rather unusual problem that has so far completely eludes us. >In a Cisco router, you can assign a route directly to an ethernet >interface, without a default gateway. for example, the ethernet >interface is IP 192.168.1.2 subnet mask 255.255.0.0 and in addition >one can define a route of IP 172.16.32.0 mask 255.255.255.0 => >destination ethernet 0. This means that this new subnet is considered >directly connected to the ethernet port, and that port communicates >with those addresses in a "normal" form, by first ARPing them directly >and then sending the packets, without a gateway in between. >In fact, what we need is even simpler: we want one ethernet interface >to have a very narrow IP range (i.e. 192.168.1.2 mask 255.255.255.255) >and in addition have ONE more machine on that LAN, e.g. to define a >route 10.0.0.1 mask 255.255.255.255 => destination ethernet 0. Two >disjoint IP addresses directly connected to the same ethernet port, >and only these two addresses. >We've not been able to get this to work. Sometimes the IP stack sends >the packets to the right interface, but the interface outputs them >with a destination MAC address of all 1s (broadcast address). In >slightly different configuration, the intherface outputs the packets >with destination MAC address equal to it's own source MAC address (?!) >which of course is no good. In other cases it seems the IP stack does >not forward packets at all to the i/f. >Has anyone ever managed to get two disjoint subnets on the same >ethernet, or managed to get any kind of routing to be considered >"directly connected" to an ethernet port? WindRiver support doesn't >even understand the question, and we really need this, so any help >would be greatly appreciated. >In case it matters, we're using x86 BSP, T2 WinNT hosts, VxWorks 5.4. >-- >Ran Shalgi What you describe can be considered a subnet with one IP address on it. This is not a normal subnet. In a normal subnet (using a class C for example) 192.168.123.x netmask 255.255.255.0, the first and last addresses are reserved: 192.168.123.0 is reserved to mean "this host" but is not generally used and 192.168.123.255 is the broadcast address for that subnet. References: RFC 950, 1219 Assuming vxWorks is doing the traditional subnetting, you need at least two "host" bits in your subnet (subnet mask 255.255.255.252) so that your real host doesn't use IP x.y.zzzzzz00 or x.y.zzzzzz11 (binary). The problem with your current setup is that your host IP address and the "broadcast" address for that subnetwork are one and the same, confusing everybody involved. gvb - -- +---------------------------------------------------------------------------+ | Jerry Van Baren / vanbaren_gerald@si.com / Grand Rapids Mi / 616-241-7973 | | My employer is a company. Companies are artifacts of a legal system. | |________________Artifacts are incapable of having opinions.________________| --------------------------- Newsgroups: comp.os.vxworks Subject: Re: problem with inet_aton Date: Wed, 13 Jun 2001 17:08:24 GMT From: "Michael Yeung" Organization: Magma Communications Ltd. Message-ID: References: <3B24981B.2A9CF1F5@wipro.com> Reply-To: "Michael Yeung" Vic and all, Actually I tried using rtrequest in VxWorks as well, but somehow I never seems to get the address conversion properly. Address "1.2.3.4" is always interpreted as "3.4.0.0". The following is a sample of what I did. Can anyone see a problem? For host route, I simply set the mask to NULL. But not sure how configure the sockaddr struct to store the mask for network route? void HostRoute (char* pDestAddrIn, char* pGwAddrIn) { int Flag; struct rtentry *pRetNrt; struct sockaddr DestAddr, GwAddr, Mask; if(pDestAddrIn != NULL) { bzero((char *) &DestAddr, sizeof(DestAddr)); DestAddr.sa_len = sizeof(struct sockaddr); DestAddr.sa_family = AF_INET; if(inet_aton(pDestAddrIn, (struct in_addr *)&DestAddr.sa_data) == ERROR) { printf("Failed conversion for DestAdd\n"); } } if(pGwAddrIn != NULL) { bzero((char *) &GwAddr, sizeof(GwAddr)); GwAddr.sa_len = sizeof(struct sockaddr); GwAddr.sa_family = AF_INET; if(inet_aton(pGwAddrIn, (struct in_addr *)&GwAddr.sa_data) == ERROR) { printf("Failed conversion for DestAdd\n"); } } Flag = RTF_HOST | RTF_UP; if(rtrequest(RTM_ADD, &DestAddr, &GwAddr, NULL, Flag, &pRetNrt) != 0) { printf("Failed to rtrequest\n"); } } Vic Sperry wrote in message news:MwBV6.6686$v4.475294@paloalto-snr1.gtei.net... > vxWorks is a modified BSD stack. rtrequest is there. I have had to use > rtrequest > on numerous occasions when the vxWorks API (e.g. routeAdd, mRoute...) failed > to work properly. If I remember correctly, netmasks are passed to rtrequest > as > sockaddr's. > > "Michael Yeung" wrote in message > news:by6V6.970$j6.80351@news... > > Vijay, > > Thanx for the info. rtrequest is a function support in BSD to add new > > route. Rather than using routeAdd from VxWorks. > > > > Michael > > > > Vijay Kumar Peshkar wrote in message > > news:3B24981B.2A9CF1F5@wipro.com... > > > Michael, > > > > > > The macro OK has value 0 and in your case as everything seems to be > > working > > > fine, the func. call is returning OK ;). You should be comparing it with > > ERROR(-1). > > > > > > The inet_addr func. is supported in vxWorks. > > > Sorry I could not understand what's "rtrequest". > > > > > > cheers, > > > Vijay > > > > > > Michael Yeung wrote: > > > > > > > Hi all, > > > > I'm not sure why but the inet_aton seems to always returns 0 and > > > > conversion is not done properly. > > > > > > > > struct in_addr temp; > > > > > > > > if(inet_aton("1.1.1.1", (struct in_addr *)&temp) == 0) > > > > { > > > > printf("Failed conversion\n"); > > > > } > > > > > > > > I noticed VxWorks does not support "inet_addr" or "inet_pton", as > > > > suggestion is appreciated. > > > > > > > > As well, does anyone knows the format expected for mask to function > like > > > > "rtrequest"? > > > > > > > > thanx in advance, > > > > Michael > > > > > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: loadModule and memDrv Date: 13 Jun 2001 13:15:17 -0400 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: References: <9g7sdc$c523@news.research.bell-labs.com> <9g84pn$c525@news.research.bell-labs.com> Sender: bpringlemeir@DeadDuck - -> lkup "SymTbl" _wtxSymTblInfoGet 0x001942c8 text (vxWorks) _statSymTbl 0x001c5c58 bss (vxWorks) _loadSymTbl 0x00111b58 text (vxWorks) _netLoadSymTbl 0x00111a38 text (vxWorks) _sysSymTbl 0x001c6254 bss (vxWorks) value = 0 = 0x0 Dan> which appears to mean the symbols are there and happy. I can Dan> run loadModule on the target console. Can you try to print the bss version to see what the values are? I think that the symbols will be in the image no matter what. You have to actually call a function to add them to the table, which consumes RAM so this is configurable. For instance, I had... - -> standAloneSymTbl _standAloneSymTbl = 0x2a90d8: value = 0 = 0x0 - -> statSymTbl _statSymTbl = 0x2a8684: value = 0 = 0x0 - -> sysSymTbl _sysSymTbl = 0x2a91b0: value = 8105904 = 0x7bafb0 Only `sysSymTbl' had memory allocated and hence symbols... but it looks like the `@lkup' succeeded, so I don't think this is your problem now, but you should check to be sure. You could also try `@ld "/memDrv/load.out"' or whatever you call your file, and see if this behaves like your program. What does `ll "/memDrv/"' show? Also try the `devs' command so we can see what it is outputting. Finally, you must use a file `int' returned from open() and not a file handle from the Ansi function like fopen()... just in case ;-) regards, Bill Pringlemeir. - -- Useful vxWorks URLS, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" "http://www.xs4all.nl/~borkhuis/vxworks/vxw_pt9.html" --------------------------- Newsgroups: comp.os.vxworks Subject: Re: problem with inet_aton Date: Wed, 13 Jun 2001 17:50:50 GMT From: "Michael Yeung" Organization: Magma Communications Ltd. Message-ID: <_JNV6.1539$j6.136827@news> References: <3B24981B.2A9CF1F5@wipro.com> Reply-To: "Michael Yeung" Hi all, I just realize I should be using sockaddr_in rather than sockaddr to set the actual content. This solves my problem with rtrequest. Michael Michael Yeung wrote in message news:c6NV6.1532$j6.135018@news... > Vic and all, > Actually I tried using rtrequest in VxWorks as well, but somehow I never > seems to get the address conversion properly. Address "1.2.3.4" is always > interpreted as "3.4.0.0". The following is a sample of what I did. Can > anyone see a problem? For host route, I simply set the mask to NULL. But > not sure how configure the sockaddr struct to store the mask for network > route? > > > void HostRoute (char* pDestAddrIn, char* pGwAddrIn) > { > int Flag; > struct rtentry *pRetNrt; > struct sockaddr DestAddr, GwAddr, Mask; > > > if(pDestAddrIn != NULL) > { > bzero((char *) &DestAddr, sizeof(DestAddr)); > DestAddr.sa_len = sizeof(struct sockaddr); > DestAddr.sa_family = AF_INET; > if(inet_aton(pDestAddrIn, (struct in_addr *)&DestAddr.sa_data) == ERROR) > { > printf("Failed conversion for DestAdd\n"); > } > } > > if(pGwAddrIn != NULL) > { > bzero((char *) &GwAddr, sizeof(GwAddr)); > GwAddr.sa_len = sizeof(struct sockaddr); > GwAddr.sa_family = AF_INET; > if(inet_aton(pGwAddrIn, (struct in_addr *)&GwAddr.sa_data) == ERROR) > { > printf("Failed conversion for DestAdd\n"); > } > } > > Flag = RTF_HOST | RTF_UP; > > if(rtrequest(RTM_ADD, &DestAddr, &GwAddr, NULL, Flag, &pRetNrt) != 0) > { > printf("Failed to rtrequest\n"); > } > } > > > Vic Sperry wrote in message > news:MwBV6.6686$v4.475294@paloalto-snr1.gtei.net... > > vxWorks is a modified BSD stack. rtrequest is there. I have had to use > > rtrequest > > on numerous occasions when the vxWorks API (e.g. routeAdd, mRoute...) > failed > > to work properly. If I remember correctly, netmasks are passed to > rtrequest > > as > > sockaddr's. > > > > "Michael Yeung" wrote in message > > news:by6V6.970$j6.80351@news... > > > Vijay, > > > Thanx for the info. rtrequest is a function support in BSD to add > new > > > route. Rather than using routeAdd from VxWorks. > > > > > > Michael > > > > > > Vijay Kumar Peshkar wrote in message > > > news:3B24981B.2A9CF1F5@wipro.com... > > > > Michael, > > > > > > > > The macro OK has value 0 and in your case as everything seems to be > > > working > > > > fine, the func. call is returning OK ;). You should be comparing it > with > > > ERROR(-1). > > > > > > > > The inet_addr func. is supported in vxWorks. > > > > Sorry I could not understand what's "rtrequest". > > > > > > > > cheers, > > > > Vijay > > > > > > > > Michael Yeung wrote: > > > > > > > > > Hi all, > > > > > I'm not sure why but the inet_aton seems to always returns 0 and > > > > > conversion is not done properly. > > > > > > > > > > struct in_addr temp; > > > > > > > > > > if(inet_aton("1.1.1.1", (struct in_addr *)&temp) == 0) > > > > > { > > > > > printf("Failed conversion\n"); > > > > > } > > > > > > > > > > I noticed VxWorks does not support "inet_addr" or "inet_pton", > as > > > > > suggestion is appreciated. > > > > > > > > > > As well, does anyone knows the format expected for mask to function > > like > > > > > "rtrequest"? > > > > > > > > > > thanx in advance, > > > > > Michael > > > > > > > > > > > > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: DP national driver driver for VxWorks Date: Wed, 13 Jun 2001 14:27:50 -0400 From: "Mike Kordik" Message-ID: References: <3B276BD2.88CF4514@teklogix.com> Yes. I do not have a fix for it yet. "Ian Melhuish" wrote in message news:3B276BD2.88CF4514@teklogix.com... > I am using the driver and also cannot get a target server connected. Did your > problem arise after installong the 2.02 patch? > > Mike Kordik wrote: > > > Is anyone using this driver (DP83815)? The target server will not connect to > > this driver after you install the Tornado 2.02 patch. Does anyone know why? > > Is there a workaround? > > > > Thanks, > > Mike > --------------------------- Newsgroups: comp.os.vxworks Subject: module download from command prompt Date: Wed, 13 Jun 2001 14:12:09 -0400 From: "Daniel G Waddington" Organization: Bell Laboratories, Lucent Technologies Message-ID: <9g8afb$c526@news.research.bell-labs.com> Hi, anyone know how to download modules from the command line, using Tornado II? Thanks DAN --------------------------- Newsgroups: comp.os.vxworks Subject: Boot rom foe VX works on tornado Date: Wed, 13 Jun 2001 11:56:09 -0700 From: "Shesha Bhushan" Organization: Arizona State University Message-ID: <9g8cvf$9m0$1@news.asu.edu> Hi, I wanted to create the boot rom for my wvworks on tornado. On the tornado GUI i go to 'Build Boot Rom' under BUILD menu. I have VGMD SBC from synergy. Now i have installed the BSP proved by synergy Inc for their board. Under the 'Build Boot Rom' menu i selected svgmd as bsp and bootrom as my image. When i click ok i get the following error... But it is working fine.. in the sense the build is successful for other BSP listed. =============================================================== ******************** E R R O R **************************** =============================================================== In file included from ..\synergy\net\esyf885End.cc:117, from esyf885End.cc:1: D:\Tornado\target\h\m2Lib.h:1024: syntax error before `,' D:\Tornado\target\h\m2Lib.h:1025: syntax error before `,' D:\Tornado\target\h\m2Lib.h:1026: syntax error before `,' D:\Tornado\target\h\m2Lib.h:1027: syntax error before `,' ..\synergy\net\esyf885End.cc: In function `void esyfRecv(struct Esyf *)': In file included from esyf885End.cc:1: ..\synergy\net\esyf885End.cc:1295: too many arguments to function make.exe: *** [esyf885End.o] Error 0x1 =============================================================== --------------------------- Newsgroups: comp.os.vxworks Subject: Re: DP national driver driver for VxWorks Date: Wed, 13 Jun 2001 15:00:06 -0400 From: Ian Melhuish Organization: Teklogix Inc. Message-ID: <3B27B836.97060194@teklogix.com> References: <3B276BD2.88CF4514@teklogix.com> I've been digging here to see what the status is re. 2.0.2. I'm developing for an ARM7 and I do not have a 2.0.2 package for ARM (just ppc and x86), but I did an upgrade for Win2k but it does not appear that I'm using 2.0.2 just 2.0. I've asked WRS about this and so far have no response. I get the message 'Unable to receive' from the target server and have traced this with a network sniffer as far as seeing that the WDB UDP packets are rejected with an ICMP response indicating destination unreachable. But everything I've looked at in usrConfig (wdbConfig) looks just fine. Mike Kordik wrote: > Yes. I do not have a fix for it yet. > "Ian Melhuish" wrote in message > news:3B276BD2.88CF4514@teklogix.com... > > I am using the driver and also cannot get a target server connected. Did > your > > problem arise after installong the 2.02 patch? > > > > Mike Kordik wrote: > > > > > Is anyone using this driver (DP83815)? The target server will not > connect to > > > this driver after you install the Tornado 2.02 patch. Does anyone know > why? > > > Is there a workaround? > > > > > > Thanks, > > > Mike > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: DP national driver driver for VxWorks Date: Wed, 13 Jun 2001 15:00:25 -0400 From: Ian Melhuish Organization: Teklogix Inc. Message-ID: <3B27B849.D7426D2B@teklogix.com> References: <3B276BD2.88CF4514@teklogix.com> I've been digging here to see what the status is re. 2.0.2. I'm developing for an ARM7 and I do not have a 2.0.2 package for ARM (just ppc and x86), but I did an upgrade for Win2k but it does not appear that I'm using 2.0.2 just 2.0. I've asked WRS about this and so far have no response. I get the message 'Unable to receive' from the target server and have traced this with a network sniffer as far as seeing that the WDB UDP packets are rejected with an ICMP response indicating destination unreachable. But everything I've looked at in usrConfig (wdbConfig) looks just fine. Mike Kordik wrote: > Yes. I do not have a fix for it yet. > "Ian Melhuish" wrote in message > news:3B276BD2.88CF4514@teklogix.com... > > I am using the driver and also cannot get a target server connected. Did > your > > problem arise after installong the 2.02 patch? > > > > Mike Kordik wrote: > > > > > Is anyone using this driver (DP83815)? The target server will not > connect to > > > this driver after you install the Tornado 2.02 patch. Does anyone know > why? > > > Is there a workaround? > > > > > > Thanks, > > > Mike > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: DP national driver driver for VxWorks Date: 13 Jun 2001 15:12:51 -0400 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: References: <3B276BD2.88CF4514@teklogix.com> <3B27B836.97060194@teklogix.com> Sender: bpringlemeir@DeadDuck >>>>> "Ian" == Ian Melhuish writes: Ian> I've been digging here to see what the status is re. 2.0.2. I'm Ian> developing for an ARM7 and I do not have a 2.0.2 package for ARM Ian> (just ppc and x86), but I did an upgrade for Win2k but it does Ian> not appear that I'm using 2.0.2 just 2.0. I've asked WRS about Ian> this and so far have no response. There is no T2.0.2 for the ARM, just T2.0.1. The ARM is *special*... I don't know why WRS has treated the ARM special. At any rate, I have T2.0.1 for the ARM architecture. I really don't know what it fixes. However, they have changed many things, such as CPU=ARM710A or CPU=ARMSA1100 to CPU=ARMARCH3 and CPU=ARMARCH4. This is nice, but cause lots of transition headaches. I am still not clear what the upgrade buys you... Ian> I get the message 'Unable to receive' from the target server and Ian> have traced this with a network sniffer as far as seeing that Ian> the WDB UDP packets are rejected with an ICMP response Ian> indicating destination unreachable. But everything I've looked Ian> at in usrConfig (wdbConfig) looks just fine. You could configure the board as a standalone image and if you have a serial port, you can manually try to configure the network to it working. Alternatively, you could try to get WDB working over a serial port (using PPP) and then try to get the network work. I have used both PPP and Ethernet on a target and you can switch interfaces to use WDB on either. If you don't have a serial port, you can throw a football at an Irishman. regards, Bill Pringlemeir. - -- I'm not just a gardener, I'm a plant manager. Useful vxWorks URLS, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" "http://www.xs4all.nl/~borkhuis/vxworks/vxw_pt9.html" --------------------------- Newsgroups: comp.os.vxworks Subject: USB driver format Date: Wed, 13 Jun 2001 15:21:58 -0400 From: "Yamin Hu" Organization: Lucent Technologies, Columbus, Ohio Message-ID: <9g8egs$n03@nntpa.cb.lucent.com> Dear Sir, I am trying to write a USB HOST device driver in vxWorks. Do I have to follow the I/O system driver routines to implement it? For example, drivers include XXcreate, XXremove, XXopen, XXwrite, XXioctl, XXinterrupt...as described in Programmer's guide. Regards, - -- Yamin --------------------------- Newsgroups: comp.os.vxworks Subject: what is the interrupt vector associated with a VME bus error? Date: Wed, 13 Jun 2001 14:29:38 -0500 From: James Toland Organization: Raytheon Company Message-ID: <3B27BF22.AD20F255@raytheon.com> I would like to detect when a BERR has occurred on the VME bus.  What is the interrupt vector and interrupt level associated with a VME bus error for a MV2700 single board computer?

The following code is what I was going to use to connect my interrupt handler:
        ...
        sysIntEnable( intLevel );
        intConnect( (VOIDFUNCPTR *) vector, (VOIDFUNCPTR)myVmeHandler, intLevel);
        ...
        void myVmeHandler(int level)
        {
            logMsg("vme interrupt level %d", arg,0,0,0,0,0);
        }

Thanks --------------------------- Newsgroups: comp.os.vxworks Subject: p2v Date: Wed, 13 Jun 2001 13:57:30 -0500 From: Surmita Chatterjee Organization: Alcatel USA Message-ID: <3B27B79A.C1E92E52@usa.alcatel.com> I would like to find out if anyone has used p2v library functions that convert pSOS function calls to vxWorks equivalent function calls. We have lot of legacy code that was written based on pSOS but now for the same product, we will be using VxWorks. For most of the functions, the transition is not that much of a problem but I am having hard time with event handling, timers and memory management/protection. Looking at the p2v library, it seems to me that it would solve my problem. But I would like to know from someone who has already got some experience with it and has any concern. thanks, Surmita --------------------------- Newsgroups: comp.os.vxworks Subject: Re: MAX Timers in VxWroks ! Date: Wed, 13 Jun 2001 14:04:13 -0500 From: Andrew Johnson Organization: APS, Argonne National Laboratory Message-ID: <3B27B92D.E72AA726@aps.anl.gov> References: <3B265CB7.B433F58A@Marconi.com> Srinivas Pitta wrote: > > I wonder whether there is any maximum LIMIT of timers, that we can spawn > in VxWorks system. > (watch dog timers, using wdStart call). There is no fixed limit, it just depends on the amount of memory and CPU time that you have available for your application. - - Andrew - -- The world is such a cheerful place when viewed from upside-down It makes a rise of every fall, a smile of every frown --------------------------- Newsgroups: comp.os.vxworks Subject: How to create network gateway via PPP... Date: Wed, 13 Jun 2001 21:33:40 +0100 From: Kees van der Bent Organization: news-service.com Message-ID: <3B27CE24.E7B57C86@support.nl> Dear people, We're having a PPP connection between our VxWorks 5.4 system and a Windows NT machine. Both the VxWorks machine and PC have Ethernet address in the same (192.168.0.0) network (which we used before we switched to PPP). We successfully created the PPP connection and can ping the PC-side PPP IP address from VxWorks. However, we're unable to ping the Ethernet IP address of the PC (being 192.168.0.2) from the VxWorks box. We leave the Ethernet interface intact, because we need this one as well. Trying get PPP working, I've removed a default route to 0.0.0.0 using gateway 192.168.0.4 (the VxWorks's Ethernet IP address), and added a new one to <192.168.0.2> using gateway . But without success. Can anyone please give me suggestions on how to ping the PC's Ethernet IP address from the VxWorks system via the PPP connection between the two. Thanks, Kees --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks drivers for z85230 synchronous and DMA modes of operation Date: Wed, 13 Jun 2001 13:09:58 -0700 From: Christopher A Leddy Organization: Raytheon Company Message-ID: <3B27C896.3ACFE8A2@west.raytheon.com> References: <9g54lt$73b$1@zeus.orl.lmco.com> <3B2656BC.E8A58F31@west.raytheon.com> <9g5urs$av0$1@zeus.orl.lmco.com> Your welcome. I have one other helpful hint (from experience with the very same chip on a Radstone board): INTERRUPT 'CONNECTS' TO THE SAME VECTOR ON A PPC UNDER VXWORKS CASCADE. Make sure your BSP does not (repeat NOT) support this chip as an additional com port with the standard vxWorks drivers. The last system I fixed had a custom driver installed at the application level to accomplish something similar to what you are doing (in my case, asych DMA com). The software engineer had inadvertantly installed this routine on top of the vendor supplied OS driver for interrupt driven serial. VxWorks for the PPC cascades the calls 'connected' to an interrupt, so the OS would call it's driver followed by the custom driver. This mess worked pretty well because no one called 'open' on the port in question, but every once and while, the unenabled OS driver would get confused and yank a single byte off the receive FIFO. This was a royal mess to find, easy to fix since they didn't need the DMA anyhow (and it was half-assed written with recv being interrupt and trans using DMA to begin with). If in doubt, look in the BSP source and see if they have code supporting the chip. Figuring out if it actually shows up in the OS can be a pain with all the ifdef's, but if you see code that tells you what device it is assigned to, you can just try an 'open("..."); and see if it actually exists in the compiled OS. Good Luck, Chris Gerald Chang wrote: > > Much thanks Chris. > > "Christopher A Leddy" wrote in message > news:3B2656BC.E8A58F31@west.raytheon.com... > > Try > > > > http://kernelnewbies.org/documents/kdoc/z8530book/z85230guide.html > > > > Source code is covered under GNU General License, but it is a good > > as a reference. > > > > > > Gerald Chang wrote: > > > > > > Hello All, > > > > > > I'm trying to use the i8237 DMA controller to send/recv data through the > > > z85230 in the SDLC mode. > > > > > > Does anyone know where I can obtain any "C" drivers that support the > DMA > > > and synchronous modes of the z85230 and i8237A that are on the Motorola > > > MVME2700 board? > > > > > > VxWorks supports only the async mode. > > > > > > Regards, > > > Gerald Chang > > > Lockheed Martin Corp. --------------------------- Newsgroups: comp.os.vxworks Subject: gdb and image tool questions Date: Wed, 13 Jun 2001 13:13:37 -0700 From: "Gang Wang" Organization: None Message-ID: Hi, Folks: Two questions: ( I am using Windows as host machine ) 1. How could I get a gdb window that I could use gdb command for debugging in Tornado 2.0 IDE? 2. Is there a tool to list all the components ( preferred in a tree structure just like in the image building window in the IDE ) for a given vxWorks image already built? This could be useful when to compare the difference between two builds. Thanks, - ---------------------------------------------------------------------------- - ----- The top three issues the world is facing: Protect the forest, Feed the hungry, and Save some malloc's. --------------------------- Newsgroups: comp.os.vxworks Subject: AIC-7899 Date: Wed, 13 Jun 2001 16:00:49 -0500 From: "Michael Murphy" Organization: Newsfeeds.com http://www.newsfeeds.com 90,000+ UNCENSORED Newsgroups. Message-ID: <3b27d7c5_6@127.0.0.1> Hi, I need VxWorks 5.4 support for an Adaptec AIC-7899 SCSI controller. Can anyone give me some pointers to what I might need for this? thanx, Michael - -----= Posted via Newsfeeds.Com, Uncensored Usenet News =----- http://www.newsfeeds.com - The #1 Newsgroup Service in the World! - -----== Over 80,000 Newsgroups - 16 Different Servers! =----- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to create network gateway via PPP... Date: Wed, 13 Jun 2001 21:56:30 GMT From: "Vic Sperry" Message-ID: References: <3B27CE24.E7B57C86@support.nl> "Kees van der Bent" wrote in message news:3B27CE24.E7B57C86@support.nl... > Dear people, > > We're having a PPP connection between our VxWorks 5.4 system and a Windows > NT machine. Both the VxWorks machine and PC have Ethernet address in the > same (192.168.0.0) network (which we used before we switched to PPP). > > We successfully created the PPP connection and can ping the PC-side PPP IP > address which is....? I'll call it ?.?.?.? from VxWorks. However, we're unable to ping the Ethernet IP address > of the PC (being 192.168.0.2) from the VxWorks box. We leave the Ethernet > interface intact, because we need this one as well. > > Trying get PPP working, I've removed a default route to 0.0.0.0 using gateway > 192.168.0.4 (the VxWorks's Ethernet IP address), and added a new one to > <192.168.0.2> using gateway . > But without success. > > Can anyone please give me suggestions on how to ping the PC's Ethernet IP > address from the VxWorks system via the PPP connection between the two. > > Thanks, > > Kees Is this accurate? 192.168.0.4 192.168.0.2 VX ------------------------------- PC | ethernet | | | |-------------------------------------| 172.19.10.102 PPP ?.?.?.? If so, you should be able to add a host route to the VX box: routeAdd("192.168.0.2","?.?.?.?") ... to route packets to the PC via the serial interface, but this seems extremely odd since you have a nice ethernet interface that is directly connected. Why, I wonder, would you want to do this? Your default route could be one of these: routeAdd("0.0.0.0","192.168.0.2") // route unknown to PC via ethernet routeAdd("0.0.0.0","?.?.?.?") // route unknown to PC via PPP --------------------------- Newsgroups: comp.os.vxworks Subject: Re: gdb and image tool questions Date: Thu, 14 Jun 2001 00:43:57 +0200 From: "Martin Raabe" Organization: T-Online Message-ID: <9g8qbe$smo$01$1@news.t-online.com> References: Hello Gang, 1. you could use the Menu Debug > Debug Windows > Debug-Commandline (Alt+1)? 2. Not as far as I know. But maybe you could go for Sourcecode comparison? This could be done by debugging from within SNiFF+, wich is a SourceCodeHandling system from WindRiver. Ciao Martin Raabe "Gang Wang" schrieb im Newsbeitrag news:DSPV6.71$P81.3775@newsfeed.avtel.net... > Hi, Folks: > > Two questions: ( I am using Windows as host machine ) > > 1. How could I get a gdb window that I could use gdb command for debugging > in Tornado 2.0 IDE? > > 2. Is there a tool to list all the components ( preferred in a tree > structure just like in the image building window in the IDE ) for a given > vxWorks image already built? This could be useful when to compare the > difference between two builds. > > Thanks, > > -------------------------------------------------------------------------- - -- > ----- > The top three issues the world is facing: > Protect the forest, Feed the hungry, and Save some malloc's. > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: VxSim x86 availability? Date: Thu, 14 Jun 2001 00:24:37 GMT From: "Paulo Sam" Organization: Excite@Home - The Leader in Broadband http://home.com/faster Message-ID: <9vTV6.223598$eK2.48460042@news4.rdc1.on.home.com> Hi all, Just wondering if it is possible to run VxSim on x86 machines using Solaris or Linux? Thanks, Paulo - -- --------------------------- Newsgroups: comp.realtime,comp.os.vxworks,comp.os.psos,comp.os.qnx,comp.os.xinu Subject: Re: Priority inversion in semaphores Date: Thu, 14 Jun 2001 12:25:38 +1000 From: Tom Evans Organization: OzEmail Ltd, Australia Message-ID: <3B2820A2.53E3@nospam.invalid> References: <4a6d848e.0106090954.514a3912@posting.google.com> <3B272916.1E3978B1@home.com> Gord Deinstadt wrote: > > amit choudhary wrote: > > > > To handle the priority inversion problem in semaphores, The following might be of general interest on this subject. For the Priority Inversion that caused problems on the Mars Pathfinder, and how they fixed it, read: http://www.ganssle.com/tem/tem14.pdf Tom Evans InitialSurnameAt tennyson.com.au --------------------------- Newsgroups: comp.os.vxworks Subject: regarding TCP/IP APIs Date: Wed, 13 Jun 2001 21:26:37 +0530 From: "G.Gopal Krishnam Raju" Organization: Lawrence Berkeley National Laboratory Message-ID: <9g9dok$mki$1@overload.lbl.gov> - ----- Original Message ----- From: G.Gopal Krishnam Raju To: the vxWorks Users Group Exploder Sent: Wednesday, June 13, 2001 6:25 AM Subject: regarding TCP/IP APIs > hello Gurus > > Sorry for troubling u all > > After getting the packet from driver (ene/elt....). What function to be > called to pass the packet from driver to tcp/ip stack where can i get this > information. Can i get networking APIs documentations. > > My problem is i have some interface which will give ethernet packet. I want > to pass to tcp/ip stack using what API i can do this. > > Thanks in Advance > > reply ASAP > > Regards > > > GGKRaju > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: PPC: system tick routine stack usage Date: Thu, 14 Jun 2001 04:32:59 GMT From: pkockritz@home.com (Pete Kockritz) Organization: Excite@Home - The Leader in Broadband http://home.com/faster Message-ID: References: <488e459a.0106111901.aba7af1@posting.google.com> In article <488e459a.0106111901.aba7af1@posting.google.com>, john_94501@yahoo.com (John) wrote: > Hello, > > > But we've caught it several times using a task's stack area while in a > > watchdog timer callback. So I'm wondering if somehow SPRG1 is getting > > changed by something else, or if SPRG1 is not being reset properly - > > intEnt increments it and intExit decrements it. > > No, SPRG1 is probably OK. It is just that there are times when the tick > announce is deferred (if it happens to go off when something else is already > "in the kernel" then it is added to the work queue). In these cases, the > work will happen in the scheduling code, and that will be on the bottom of a > task stack. I think it will be the "old" task stack (i.e. the one running > before the decrementer interrupt). We've got a test that just does a watchdog timer with a callback repeatedly. In the callback I print out the SPRG1 value. Almost always it is 0. Only once, out of hundreds of attempts, did I see SPRG1 with a value of 1, indicating that it was actually in ISR context. And when SPRG1 was 1, the stack frame was in the interrupt stack. All the other times the stack from was on some other task's stack. The vxWorks Prog Guide says that the watchdog work can be deferred to tExcTask. To me that would mean it executes with its stack frame on the stack of tExcTask. > Now, "in the kernel" was in quotes because it means only the brief periods > where the kernel is working on critical data structures. Unlike other kernels, > the Wind kernel does not lock out interrupts for most of this work; instead it > uses a flag to indicate that it is in the kernel's critical region. Any > interrupt making calls to kernel routines (e.g. semGive) will execute, but the > operation will be queued. When the original routine exits the kernel, the work > queue will be drained in order. > > We probably would've never noticed it if the watchdog callback didn't have > > a large stack frame. Perhaps it happens occassionally but it is just not > > noticed? > > Indeed. As with ISRs though it is best to keep watchdog call backs simple, and > not use a lot of resources (in this case stack, but also CPU time etc). Do you > really need such a large stack frame? No, we don't. We've modified the message sending routine to have a smaller stack frame, and call another function with the large buffer only if it needs to (and it doesn't need to when called from the callback). Thanks for your reply. Pete - -- Pete Kockritz --------------------------- Newsgroups: comp.os.vxworks Subject: Regarding Sending a Packet after a fixed interval of time Date: Thu, 14 Jun 2001 11:48:06 +0530 From: "Deepesh Aggarwal" Organization: Lawrence Berkeley National Laboratory Message-ID: <9g9oa5$r6e$1@overload.lbl.gov> This is a multi-part message in MIME format. - ------=_NextPart_000_00E4_01C0F4C7.E0559BF0 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="iso-8859-1" Hi All I want to "send multicast Packets after a specified fixed interval of = time repeatedly" . So to serve my purpose i write a program whose code looks like .. *************************************************************************= ****************** #include "vxWorks.h" #include "taskLib.h" #include "socket.h" #include "netinet/in.h" #include "stdio.h" #include "stdlib.h" #include "string.h" #include "sockLib.h" #include "inetLib.h" #include "ioLib.h" #include "routeLib.h" #include "logLib.h" #include "wdLib.h" #define SECONDS (15) WDOG_ID myWatchDogId; struct sockaddr_in sin; struct sockaddr_in toAddr; int toAddrLen; int sockDesc; char * sendBuf=3D"Hello"; int sendLen =3Dsizeof(sendBuf); int len; STATUS trial(void); STATUS mcastSend (char * mcastAddr, USHORT mcastPort); STATUS multi(void) { routeAdd("224.0.0.0","157.227.2.1"); taskSpawn ("maintask", 100, 0, 20000,mcastSend,"224.0.0.23" = ,"8769",0,0,0,0,0,0,0,0); return (OK); } =20 STATUS mcastSend(char * mcastAddr, USHORT mcastPort) { =20 /* create a send and recv socket */ if ((sockDesc =3D socket (AF_INET, SOCK_DGRAM, 0)) < 0 ) { printf (" cannot open send socket\n"); return (ERROR); } /* zero out the structures */ bzero ((char *)&sin, sizeof (sin)); bzero ((char *)&toAddr, sizeof (toAddr)); sin.sin_family =3D AF_INET; sin.sin_addr.s_addr =3D INADDR_ANY; sin.sin_port =3D htons(0); if (bind(sockDesc, (struct sockaddr *)&sin, sizeof(sin)) !=3D 0) { perror("bind"); if (sockDesc) close (sockDesc); return (ERROR); } toAddrLen =3D sizeof(struct sockaddr_in); toAddr.sin_family =3D AF_INET; /* initialize the address to the send */ toAddr.sin_addr.s_addr =3D inet_addr (mcastAddr); printf("addr is %d",toAddr.sin_addr.s_addr); /* initialize the port to send */ toAddr.sin_port =3D htons(mcastPort); if ((myWatchDogId =3D wdCreate( )) =3D=3D NULL) {=20 perror("Error in creating Timer\n"); return (ERROR); } if (wdStart (myWatchDogId, sysClkRateGet( ) * SECONDS,trial,0) =3D=3D = ERROR) { perror("Error in starting Timer\n"); =20 return(ERROR); } } STATUS trial(void) { /*printf ("sending \n"); printf ("Sending %s of length %d using Socket descriptor = %d\n",sendBuf,sendLen,sockDesc );*/ if ((len =3D sendto (sockDesc, sendBuf, sendLen, 0,(struct sockn = *)&toAddr, toAddrLen)) < 0 ) { printf("mcastSend sendto errno:0x%x\n", errno ); } if (wdStart (myWatchDogId, sysClkRateGet( ) * SECONDS,trial,0) =3D=3D = ERROR) { return(ERROR); } } *************************************************************************= ********************** I downloaded the application on target with the help of target server = and tried to call "multi" , it works fine . i mean it was sending the = packet after 15 seconds ..but=20 after sending two or three packets my target server gets killed and when = again i tried to launch it ....it gives an error WTX_ERROR ( Connection = busy ..). but i didn't get any error on target .and target was still sending the = packets ( That's ok .that's the behavior i want..). .until i explicitly = shut it down... I m using P-III as a target with an ethernet card FEI ( Fast Intel Ether = Express PRO100B PCI ). Can anyone please explain me the reason why Target server is behaving = like that.? and where and what can be the problem..?....or suggest me = some better way to do the same... =20 Eagerly waiting for ur precious advice Thanks n Happy Working=20 Deepesh - ------=_NextPart_000_00E4_01C0F4C7.E0559BF0 Content-Transfer-Encoding: quoted-printable Content-Type: text/html; charset="iso-8859-1"

Hi All
 
I want to = "send multicast Packets=20 after a specified fixed interval of time repeatedly" = .
So to serve my purpose i write a = program whose=20 code looks like ..
 
****************************************************************= ***************************
 
#include "vxWorks.h"
#include=20 "taskLib.h"
#include "socket.h"
#include = "netinet/in.h"
#include=20 "stdio.h"
#include "stdlib.h"
#include "string.h"
#include=20 "sockLib.h"
#include "inetLib.h"
#include "ioLib.h"
#include=20 "routeLib.h"
#include "logLib.h"
#include "wdLib.h"
 
#define SECONDS (15)
 
WDOG_ID myWatchDogId;
 
struct sockaddr_in sin;
struct = sockaddr_in=20 toAddr;
int toAddrLen;
int sockDesc;
char * = sendBuf=3D"Hello";
int=20 sendLen =3Dsizeof(sendBuf);
int len;
 
STATUS trial(void);
STATUS mcastSend = (char *=20 mcastAddr, USHORT mcastPort);
STATUS=20 multi(void)
{
 routeAdd("224.0.0.0","157.227.2.1");
 t= askSpawn=20 ("maintask", 100, 0, 20000,mcastSend,"224.0.0.23"=20 ,"8769",0,0,0,0,0,0,0,0);
 return = (OK);
}
 
STATUS mcastSend(char * mcastAddr, USHORT=20 mcastPort)
{
 
/* create a send and recv socket */
if=20 ((sockDesc =3D socket (AF_INET, SOCK_DGRAM, 0)) < 0 )
{
printf = (" cannot=20 open send socket\n");
return (ERROR);
}
 
/* zero out the structures */
bzero = ((char=20 *)&sin, sizeof (sin));
bzero ((char *)&toAddr, sizeof=20 (toAddr));
 
sin.sin_family =3D = AF_INET;
sin.sin_addr.s_addr =3D=20 INADDR_ANY;
sin.sin_port =3D htons(0);
 
if (bind(sockDesc, (struct sockaddr = *)&sin,=20 sizeof(sin)) !=3D 0)
{
perror("bind");
if (sockDesc)
close=20 (sockDesc);
return (ERROR);
}
 
toAddrLen =3D sizeof(struct=20 sockaddr_in);
toAddr.sin_family =3D AF_INET;
/* initialize the = address to=20 the send */
toAddr.sin_addr.s_addr =3D inet_addr = (mcastAddr);
 
printf("addr is = %d",toAddr.sin_addr.s_addr);
/*=20 initialize the port to send */
toAddr.sin_port =3D=20 htons(mcastPort);
 
if ((myWatchDogId =3D wdCreate( )) = =3D=3D NULL)
{=20
perror("Error in creating Timer\n");
return = (ERROR);
}
 
if (wdStart (myWatchDogId, = sysClkRateGet( ) *=20 SECONDS,trial,0) =3D=3D ERROR)
{
perror("Error in starting = Timer\n"); =20
return(ERROR);
}
 
}
 

STATUS trial(void)
{
/*printf = ("sending=20 \n");
printf ("Sending %s of length %d using Socket descriptor=20 %d\n",sendBuf,sendLen,sockDesc );*/
 
if ((len =3D sendto (sockDesc, sendBuf, = sendLen,=20 0,(struct sockn *)&toAddr, toAddrLen)) < 0 = )
{
printf("mcastSend=20 sendto errno:0x%x\n", errno );
}
 
if (wdStart (myWatchDogId, = sysClkRateGet( ) *=20 SECONDS,trial,0) =3D=3D ERROR)
{
return(ERROR);
}
 
}

************************************************************= ***********************************
 
I downloaded the application on target = with the=20 help of target server and tried to call "multi" , it works fine . i mean = it was=20 sending the packet after 15 seconds ..but
after sending two or three packets my = target server=20 gets killed and when again i tried to launch it ....it gives an error = WTX_ERROR=20 ( Connection busy ..).
but i didn't get any error on target=20 .and target was still sending the packets ( That's ok .that's the = behavior=20 i want..). .until i explicitly shut it down...
I m using P-III as a target with an = ethernet card=20 FEI ( Fast Intel Ether Express PRO100B PCI ).
 
Can anyone please explain me the reason = why Target=20 server is behaving like that.? and where and what can be the = problem..?....or=20 suggest me some better way to do the same...
 
Eagerly waiting for ur precious = advice
 
Thanks n Happy=20 Working
Deepesh
 
- ------=_NextPart_000_00E4_01C0F4C7.E0559BF0-- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxSim x86 availability? Date: Thu, 14 Jun 2001 10:47:48 +0200 From: "Martin Raabe" Organization: T-Online Message-ID: <9g9tnf$c5$06$1@news.t-online.com> References: <9vTV6.223598$eK2.48460042@news4.rdc1.on.home.com> Hello Paulo, this is not possible except using VMware. Speaking for my own, not for my employer. Ciao Martin "Paulo Sam" schrieb im Newsbeitrag news:9vTV6.223598$eK2.48460042@news4.rdc1.on.home.com... > Hi all, > > Just wondering if it is possible to run VxSim on x86 machines using Solaris > or Linux? > > Thanks, > Paulo > > -- > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Problem in Performance monitoring Date: Thu, 7 Jun 2001 16:23:14 +0530 From: "Amitb" Organization: VSNL Message-ID: <9foq9j$feb$1@news.vsnl.net.in> Hi, We want to find out the execution time of the tasks in VxWorks. We used spy chart in Browser for the same, but it just gives the % of cpu usage. Though the libraries (spyLib & usrLib) are available & INCLUDE_SPY macro is included in configAll.h, we are unable to use the spy routines in tornado shell. Shell doesn't recognize the command. Neither we are able to use it in the application. The error comes while downloading the application. Can anyone help us on this front? Thanks --------------------------- Newsgroups: comp.os.vxworks Subject: tWdbTask stack overflow Date: Thu, 14 Jun 2001 18:53:54 +0930 From: "David Hichens" Message-ID: <3b2882d2_3@news01.one.net.au> G'Day, I have a stack overflow problem with tWdbTask - the Vx-Works target agent task. This was observed using the browser tool. Is this a sympton of another problem in my code or is there some way to enlarge the stack allocation for this task? Thanks for any assistance. Regards, David Hichens. --------------------------- Newsgroups: comp.os.vxworks Subject: Booting from ATA disk ... crash Date: 14 Jun 2001 03:25:16 -0700 From: igtorque@eliop.es (=?ISO-8859-1?Q?Ignacio_Gonz=E1lez?=) Organization: http://groups.google.com/ Message-ID: <9ec53d11.0106140225.75f234a8@posting.google.com> Hello all. Now that we have not renewed our support (...) from our local reseller of VxWorks, I've run into an annoying problem. My new target is a Pentium PC (somewhat old). With my boot diskette, I can load VxWorks from that diskette, and even from my host PC via Ethernet. Now I'm trying to load VxWorks from the target's disk. First I boot from diskette and change the boot parameters to look like this: boot device : ata=0,0 file name : vxWorks flags : 0x0 ... Previously, I had copied file vxWorks in C:\ (the disk is FAT formatted), booting from DOS instead of from the VxWorks boot diskette. When I try load the boot file: [VxWorks Boot]: l the following message appears: Attaching to ATA disk device... Then the disk grumbles and mumbles and, after a few seconds, the screen spits: 0x7febe80(tBoot):memPartAlloc:block too big- 16448 in partition 0x5f9f4. Error during dosFsDevInit: 380001 usrAtaCofig failed. Error loading file: errno=0x380001. Surely, there is something very basic that I´m missing. Any clue? Thanks for reading. Many more thnaks if you think about it. Infinite thanks if you answer or post. --------------------------- Newsgroups: comp.os.vxworks Subject: VCOMM problem Date: Thu, 14 Jun 2001 12:50:43 +0200 From: Yolanda Barquilla Organization: =?iso-8859-1?Q?Telef=F3nica?= Sistemas Message-ID: <3B289702.7C74DCB4@ts.es> Hi all. I am developing a driver for a SmartcardReader. For communications i use the RS232 channel and the functions included in the VCOMM library. I have a problem.Although i send commands to the Smartcard it seems not to recognize those bytes.I mean, I don´t know if all bytes arrive to the card correctly prior to the card answer. How can i control the communication of this channel in order to know when all bytes have been sent and when the Smartcard starts to answer?. I am using a Protocol Analyzer and i can see that although the first time the card answers correctly, i only can see the correct answer at thirth time. Thanks a lot Yolanda --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Thu Jun 14 05:29:51 2001 From: David Laight Date: Thu Jun 14 05:29:53 PDT 2001 Subject: Re: DP national driver driver for VxWork > There is no T2.0.2 for the ARM, just T2.0.1. The ARM is *special*... > I don't know why WRS has treated the ARM special. At any rate, I have > T2.0.1 for the ARM architecture. I really don't know what it fixes. > However, they have changed many things, such as CPU=ARM710A or > CPU=ARMSA1100 to CPU=ARMARCH3 and CPU=ARMARCH4. This is nice, but cause > lots of transition headaches. I am still not clear what the upgrade > buys you... Don't know about the arm being 'special' more like 'second class'. None of the 'mandatory' patches are ever built for ARM. The floating point routines are slower than an old sinclair scientific (ok not quite that slow...) I recall some changes in the mmu setup routines when doing the arm T2 -> T201 port. Ended up throwing all the vxworks mmu setup code out with the bathwater. You do need to implement _cacheDFlushAll, _mmuTtbrSet, _mmuTLBIDFlushAll, _mmuIntLock, _mmuIntUnlock and _cache_text_update (write address to cacheLib+0x1c) Total about 60 lines of assembler. I was told that there was a new arm build due out 'soon' using a later version of gcc. However 'soon' has already passed :-) David ---------------------------------------------------------------- David Laight email: dsl@tadpole.co.uk Tadpole Technology plc phone: +44 1223 428 232 Cambridge, UK fax: +44 1223 428 201 From vxwexplo-errs@csg.lbl.gov Thu Jun 14 07:56:10 2001 From: "Ravi Malhotra" Date: Thu Jun 14 07:56:12 PDT 2001 Subject: Usage of Distributed Name Database in VxFusion hi VxWorks Gurus, does the distributed name database allow global data (such as my own defined structures) to be added as entries , such that they can be shared across all nodes in the system. any pointers to the same will be highly appreciated Thanks Ravi Malhotra From vxwexplo-errs@csg.lbl.gov Thu Jun 14 08:57:52 2001 From: "Joseph Sebastian" Date: Thu Jun 14 08:57:54 PDT 2001 Subject: Looking for VxWorks assignment Hi All, Experienced over 12 years in embedded programming. 5 years in VxWorks and pSos RTOS. Looking for new assignment, in the bay area. Joseph Sebastian. joseph_ks@hotmail.com _________________________________________________________________________ Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com. From vxwexplo-errs@csg.lbl.gov Thu Jun 14 19:18:37 2001 From: "Devatha, Pratap" Date: Thu Jun 14 19:18:39 PDT 2001 Subject: editor in vxworks Hi I wanted to know if there is any known text editor which can edit files in the vxworks file system residing in the target system. Vxworks is running on Intel Ixp1200 Network Processor. I am also planning to write a text editor for this. Does anyone know the address at which the video memory resides for the Ixp1200. Any help regarding this is appreciated. Thanking you, Pratap From vxwexplo-errs@csg.lbl.gov Fri Jun 15 00:38:11 2001 From: "Suresh Kumar Shanbhog" Date: Fri Jun 15 00:38:13 PDT 2001 Subject: a doubt on linking libraries(unix) in VxSim This is a multi-part message in MIME format. --------------InterScan_NT_MIME_Boundary Content-Type: multipart/alternative; boundary="----=_NextPart_000_0051_01C0F57C.89D59680" ------=_NextPart_000_0051_01C0F57C.89D59680 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Hello, Iam Suresh and I work for Wipro.We are working on a project where we = need to simulate a gateway.Real gateway software runs on VxWorks.Our = project involves simulating the gateway.For this we are compiling the = source code written for VxWorks on VxSim(Unix version) so that its = entire functionality is available on Unix for testing purpose. As a development we need to plug-in some code which is written in = Unix.To achieve this we need to include some libraries compiled on Unix. We tried this by including these libs after editing the makefile (under = $WIND_BASE/target/config/hpux) with=20 MACH_EXTRA =3D /usr/lib/xyzlib.a we are using loader under /bin/ld While building VxWorks image we are getting some linker errors like = unsatisfied symbols related to socket calls(gethostbyname) and some = other Unix related calls. In the library there is one such function which calls "getpwuid()". This = function is available only in Unix but not on VxWorks.If I try to = include Unix library /usr/lib/libc.a I got lots of linker errors related = to "duplicate symbols of malloc,realloc etc" I would like to know how to resolve issues such as the above where in we = have few system calls supported under Unix but not supported under = VxWorks.Do we have to set any compiler options so that it takes default = Unix libraries since we are compiling and running the image on = Simulator. Appreciate your help in this regard. Thanks, S Suresh Kumar Sr Systerms Engineer Wipro-Bangalore. ph# 5538301-2750 ESN# 6-872-8290 ------=_NextPart_000_0051_01C0F57C.89D59680 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable
Hello,
 
Iam Suresh and I work for Wipro.We are working on a project where = we need=20 to simulate a gateway.Real gateway software runs on VxWorks.Our project = involves=20 simulating the gateway.For this we are compiling the source code written = for=20 VxWorks on VxSim(Unix version) so that its = entire=20 functionality is available on Unix for testing purpose.
As a development we need to plug-in some code which is written in = Unix.To=20 achieve this we need to include some libraries compiled on Unix.
We tried this by including these libs after editing the makefile = (under=20 $WIND_BASE/target/config/hpux) with
       =20             =    =20 MACH_EXTRA =3D /usr/lib/xyzlib.a
we are using loader under=20 /bin/ld
While building VxWorks image we are getting some linker errors like = unsatisfied symbols related to socket calls(gethostbyname) and some = other Unix=20 related calls.
In the library there is one such function which calls "getpwuid()". This function is available only in = Unix but=20 not on VxWorks.If I try to include Unix library /usr/lib/libc.a I got lots of = linker=20 errors related to "duplicate symbols of malloc,realloc = etc"
I would like to know how to resolve issues such as the above where = in we=20 have few system calls supported under Unix but not supported under = VxWorks.Do we=20 have to set any compiler options so that it takes default Unix libraries = since=20 we are compiling and running the image on Simulator.
 
Appreciate your help in this regard.

Thanks,

S Suresh Kumar
Sr Systerms = Engineer
Wipro-Bangalore.
ph#=20 5538301-2750
ESN# 6-872-8290
------=_NextPart_000_0051_01C0F57C.89D59680-- --------------InterScan_NT_MIME_Boundary-- From vxwexplo-errs@csg.lbl.gov Fri Jun 15 02:10:37 2001 From: "Watson, John" Date: Fri Jun 15 02:10:39 PDT 2001 Subject: Memory Mapping This is a multi-part message in MIME format. --------------InterScan_NT_MIME_Boundary Content-Type: text/plain; charset="iso-8859-1" Hi, I'm extremely new to vxWorks and having difficulty with Memory Mapping. I'm using a mbx860 SBC and BSP, and I'm trying to add an ethernet card to the board. I need to know what values to set the memory addresses to for FEIO_DEV_SPACE FEIO_MEMBASE0 memory base for CSR FEIO_MEMBASE1 memory base for Flash FEIO_IOBASE0 IO base for CSR these defines are found in the mbx800.h file. Also when do the functions to initialise the ethernet card need to be called, is it during the boot or can it be performed after? Thanks for any help, John Watson --------------InterScan_NT_MIME_Boundary Content-Type: text/plain; name="RMRL-Disclaimer.txt" Content-Transfer-Encoding: 7bit Content-Disposition: attachment; filename="RMRL-Disclaimer.txt" The information contained in this e-mail is confidential to Roke Manor Research Ltd and must not be passed to any third party without permission. This communication is for information only and shall not create or change any contractual relationship. --------------InterScan_NT_MIME_Boundary-- From vxwexplo-errs@csg.lbl.gov Fri Jun 15 04:03:39 2001 From: Vxworks Exploder Date: Fri Jun 15 04:03:41 PDT 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Fri Jun 15 04:03:07 PDT 2001 Subject: Re: internal compiler errors Subject: Re: Booting from ATA disk ... crash Subject: Re: pciDevConfig Subject: Re: internal compiler errors Subject: Re: Booting from ATA disk ... crash Subject: re: DP national driver driver for VxWork Subject: Re: VxWorks Tools questions Subject: Usage of Distributed Name Database in VxFusion Subject: build_instance() in next_async() Subject: Looking for VxWorks assignment Subject: Re: Booting from ATA disk ... crash Subject: TrueFFS INstallation Problem Subject: Re: Regarding Sending a Packet after a fixed interval of time Subject: Maintianing Unresolved Symbols Subject: Re: DP national driver driver for VxWorks Subject: Re: Proxy ARP problems: workaround needed Subject: What is meaning of this error Subject: Prevent "Press any key to stop auto-reboot" prompt Subject: Windview wooos ? [Missing events, timing off...]... Subject: Re: DP national driver driver for VxWorks Subject: Network performance question (UDP/IP) Subject: "pcb cache lookups failed" Subject: Re: Prevent "Press any key to stop auto-reboot" prompt Subject: Re: What is meaning of this error Subject: IGMP Subject: Re: Prevent "Press any key to stop auto-reboot" prompt Subject: Re: build_instance() in next_async() Subject: editor in vxworks Subject: Re: DP national driver driver for VxWorks Subject: Re: Command 'iosFdShow' does not work in Shell Subject: Re: Windview wooos ? [Missing events, timing off...]... Subject: Re: internal compiler errors Subject: Re: Command 'iosFdShow' does not work in Shell Subject: a doubt on linking libraries(unix) in VxSim Subject: Re: Booting from ATA disk ... crash Subject: Memory Mapping Subject: Attaching network interface cpm0... 0x1fe6aa8 (tBoot): Subject: Re: PLX 9054 (and vxWorks) help Subject: Reserving Memory ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: internal compiler errors Date: Thu, 14 Jun 2001 11:07:56 GMT From: jagan Organization: Road Runner Message-ID: <3B28C4E7.6620AF6B@mediaone.net> References: <3B207CA7.D938FC01@mediaone.net> This has been a very elusive problem. I have tried it on different PCs and it strikes randomly. It has happened on almost all PCs that I have tried on at some time or the other - but usually it happens on a different pc at a different time. At this point I suspect the problem may be a combination of bash environment (cygwin), cygwin make, and the tornado compiler cc1.exe. When it happens I can get around it by compiling the file (different file at different times) individually and then calling make to compile and link all the other files. Any thoughts? We sent a message to Wind River, but they refused to look at the problem because of the mixed environment (I do not use their project tool.. and their make files). - -jagan. Dave Korn wrote: > > "jagan" wrote in message > news:3B207CA7.D938FC01@mediaone.net... > > > > > > I am seeing an internal compiler error when compiling > > using Tornado 2 cross compiler for a ppc target in > > a cygwin environment on NT. > > > > I see cc1.exe coring with signal 5. > > > > Does anyone know why this may be happening? > > Not really. Could be flaky hardware or configuration - try doing the same > build on a different PC. Show us the actual output that you get - copy + > paste it from the screen into a reply. > > Davek --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Booting from ATA disk ... crash Date: Thu, 14 Jun 2001 12:09:42 +0100 From: "timallen" Organization: ntlworld News Service Message-ID: References: <9ec53d11.0106140225.75f234a8@posting.google.com> Reply-To: "timallen" Ignacio, > Now that we have not renewed our support (...) from our local reseller > of VxWorks, I've run into an annoying problem. Hmm! ... > Previously, I had copied file vxWorks in C:\ (the disk is FAT > formatted), booting from DOS instead of from the VxWorks boot > diskette. As I understand it, VxWorks DOSFS cannot read hard drives which have been formatted from DOS, unless you have DOSFS 2.0, which is a free upgrade ... if you had support. :-( Regards, Tim. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: pciDevConfig Date: Thu, 14 Jun 2001 11:18:58 GMT From: jagan Organization: Road Runner Message-ID: <3B28C77D.EA438E6A@mediaone.net> References: <3B1407CA.8E7BA482@ll.mit.edu> That configures the PCI device to be in the respective IO and Configuration spaces. The IO for the PCI device will start at the BASE_ADDR_0_IO (4th param) - IO space can be accessed as offsets from this using specific routines or directly accesses at those memory locations. The configuration space of the device will start at the BASE_ADDR_0_MEM (5th param) - config registers will be accessed as offsets.. or if memory mapped they can be directly accessed. Michael Reed wrote: > > Hello, > I'm porting a VxWorks driver to pSOS. Can anyone shed some light on > what this function call accomplishes: > > pciDevConfig (pciBusNo, pciDevNo, pciFuncNo, > PCI_IO_PMC_ADRS + BASE_ADDR_0_IO, > PCI_MEM_PMC_ADRS + BASE_ADDR_0_MEM, > (PCI_CMD_SERR_ENABLE | PCI_CMD_MASTER_ENABLE > | PCI_CMD_MEM_ENABLE | PCI_CMD_IO_ENABLE)); > > Specifically, what do the 4th & 5th parameters do (PCI_IO_PMC.. & > PCI_MEM_PMC..). > > Thanks, > Mike --------------------------- Newsgroups: comp.os.vxworks Subject: Re: internal compiler errors Date: Thu, 14 Jun 2001 07:46:42 -0700 From: Joe Durusau Organization: Lockheed Martin Corporation Message-ID: <3B28CE52.630BE57@lmco.com> References: <3B207CA7.D938FC01@mediaone.net> <3B28C4E7.6620AF6B@mediaone.net> Well, signal 5 is usually SIGTRAP, which would typically indicate a memory problem in the PC. I have seen many, many problems caused by intermittent memory failures in X86 hardware, since many of them don't even have parity checking anymore, which in in itself a very weak error check. Your best bet (not cheapest) choice is to upgrade to better hardware, preferably something that has a fairly strong way of self-checking the hardware. Speaking only for myself, Joe Durusau jagan wrote: > > This has been a very elusive problem. I have tried > it on different PCs and it strikes randomly. It > has happened on almost all PCs that I have tried > on at some time or the other - but usually it > happens on a different pc at a different time. > > At this point I suspect the problem may be a > combination of bash environment (cygwin), cygwin make, > and the tornado compiler cc1.exe. > > When it happens I can get around it by compiling > the file (different file at different times) individually > and then calling make to compile and link all the other > files. > > Any thoughts? > > We sent a message to Wind River, but they refused > to look at the problem because of the mixed environment (I do not use > their project tool.. and their make files). > > -jagan. > > Dave Korn wrote: > > > > "jagan" wrote in message > > news:3B207CA7.D938FC01@mediaone.net... > > > > > > > > > I am seeing an internal compiler error when compiling > > > using Tornado 2 cross compiler for a ppc target in > > > a cygwin environment on NT. > > > > > > I see cc1.exe coring with signal 5. > > > > > > Does anyone know why this may be happening? > > > > Not really. Could be flaky hardware or configuration - try doing the same > > build on a different PC. Show us the actual output that you get - copy + > > paste it from the screen into a reply. > > > > Davek --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Booting from ATA disk ... crash Date: 14 Jun 2001 11:54:41 GMT From: tmb@Xisotek.co.uk (Trevor Barton) Organization: Isotek Electronics Ltd., Leeds, UK. Message-ID: References: <9ec53d11.0106140225.75f234a8@posting.google.com> Reply-To: tmb@Xisotek.co.uk On Thu, 14 Jun 2001 12:09:42 +0100, timallen wrote: >Ignacio, > >> Now that we have not renewed our support (...) from our local reseller >> of VxWorks, I've run into an annoying problem. > >Hmm! > >... >> Previously, I had copied file vxWorks in C:\ (the disk is FAT >> formatted), booting from DOS instead of from the VxWorks boot >> diskette. > >As I understand it, VxWorks DOSFS cannot read hard drives which have been >formatted from DOS, unless you have DOSFS 2.0, which is a free upgrade ... >if you had support. :-( I'm not sure that's the case, we format all our disks with Dos 6.2 and it works OK. However, we don't boot off them natively (?) we bood dos first and then use vxload.com to load vxworks. That's simply because it's useful to be able to use DOS sometimes (repair, inital loading, editing config files in the field, etc) and so we have a quick timeout config.sys meny to choose either to automatically start vxworks or stay at the dos shell. However, we're looking at FreeDos because I hate paying M$ for something that's resident only for seconds before it gets heaved out or RAM by vxworks! Trev - -- Trevor Barton Isotek Electronics Ltd, 9 Clayton Wood Bank, Leeds, LS16 6QZ, UK. Tel: +44 (113) 275 1339, Fax +44 (113) 224 9827 Remove X from tmb@Xisotek.co.uk before replying. Views expressed are my own and not necessarily those of Isotek Electronics Ltd. --------------------------- Newsgroups: comp.os.vxworks Subject: re: DP national driver driver for VxWork Date: Thu, 14 Jun 2001 13:29:24 +0100 (BST) From: David Laight Organization: Lawrence Berkeley National Laboratory Message-ID: <200106141229.NAA27914@dsl-2.tadpole.co.uk> > There is no T2.0.2 for the ARM, just T2.0.1. The ARM is *special*... > I don't know why WRS has treated the ARM special. At any rate, I have > T2.0.1 for the ARM architecture. I really don't know what it fixes. > However, they have changed many things, such as CPU=ARM710A or > CPU=ARMSA1100 to CPU=ARMARCH3 and CPU=ARMARCH4. This is nice, but cause > lots of transition headaches. I am still not clear what the upgrade > buys you... Don't know about the arm being 'special' more like 'second class'. None of the 'mandatory' patches are ever built for ARM. The floating point routines are slower than an old sinclair scientific (ok not quite that slow...) I recall some changes in the mmu setup routines when doing the arm T2 -> T201 port. Ended up throwing all the vxworks mmu setup code out with the bathwater. You do need to implement _cacheDFlushAll, _mmuTtbrSet, _mmuTLBIDFlushAll, _mmuIntLock, _mmuIntUnlock and _cache_text_update (write address to cacheLib+0x1c) Total about 60 lines of assembler. I was told that there was a new arm build due out 'soon' using a later version of gcc. However 'soon' has already passed :-) David - ---------------------------------------------------------------- David Laight email: dsl@tadpole.co.uk Tadpole Technology plc phone: +44 1223 428 232 Cambridge, UK fax: +44 1223 428 201 --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks Tools questions Date: 14 Jun 2001 06:59:03 -0700 From: prenom_nomus@yahoo.com (Mark) Organization: http://groups.google.com/ Message-ID: References: <3b1ee72a.4738173@news.qwest.net> <3B1FC428.817D241B@aps.anl.gov> Andrew Johnson wrote in message news:<3B1FC428.817D241B@aps.anl.gov>... > Johan Borkhuis wrote: > > > > > Are there significant differences between the [default] GNU compiler > > > and Diab C++? (We will be using C++ on a PowerQUICC 860) > > > > The GNU compiler is old, it is the 2.7.2 version of the compiler. There > > will be a new release soon, but it is not available. > > Not quite true, the C compiler they ship is still based on gcc-2.7.2, but > the C++ compiler is based on the newer but still old EGCS-1.0.2 release > (see the Tornado 2.0 Release Notes, section 7.6). They just combined the > two such that if you ask for the C++ compiler's version you actually get > given that of the C compiler. The most annoying thing that's missing is > C++ namespaces; other than that it's about as good as MS VC++ in its > support of the ISO standard, but in different ways. FYI: I was told by WindRiver that a newer GNU C++ compiler for Tornado (supporting namespaces) will be coming out very shortly. --------------------------- Newsgroups: comp.os.vxworks Subject: Usage of Distributed Name Database in VxFusion Date: Thu, 14 Jun 2001 18:26:44 +0530 From: "Ravi Malhotra" Organization: Lawrence Berkeley National Laboratory Message-ID: <005501c0f4d1$77ade080$1607080a@future.futsoft.com> hi VxWorks Gurus, does the distributed name database allow global data (such as my own defined structures) to be added as entries , such that they can be shared across all nodes in the system. any pointers to the same will be highly appreciated Thanks Ravi Malhotra --------------------------- Newsgroups: comp.protocols.snmp,comp.os.psos,comp.os.vxworks Subject: build_instance() in next_async() Date: Thu, 14 Jun 2001 19:02:27 +0200 From: Alex Vinokur Message-ID: <3B28EE22.94383912@bigfoot.com> ============================ pRISM+ for pSOSystem/PowerPC pRISM+ 1.2.3 . pSOSystem 2.2.2 Diab Data D-CC Optimizing C compiler SNMP Release 7.4 (SNMPv1) Host : Windows98 Target : PowerPC (MPC860) ============================ Here are fragments of files that implement a sample SNMP agent using pSOSystem //--------------------- 1 -------------------- // File ip.c /* * ipAddrEntry_next_async * * Common GET method routine for ipAddrEnt table */ void ipAddrEntry_next_async(OIDC_T lastmatch, int compc, OIDC_T *compl, SNMP_PKT_T *pktp, VB_T *vbp) { #define ipAddrEntry_INSTANCE_LEN 4 struct mib_ipaddrreq *data, *best; OIDC_T tmp_inst[ipAddrEntry_INSTANCE_LEN]; OIDC_T best_inst[ipAddrEntry_INSTANCE_LEN]; int i,no; int error; unsigned inst_len; unsigned best_inst_len; /* find all the varbinds that share the same getproc and instance */ group_by_getproc_and_instance(pktp, vbp, compc, compl); read_ipaddrent(); /* generate the instance of each table entry and find the * smallest instance that's larger than compc/compl */ best = 0; /* This loop needs to iterate over each entry in your table */ for (data = ipatab, no = ipaddrlen; no; data++,no--) { inst_len = ipAddrEntry_build_instance(data, tmp_inst); if ((oidcmp2(inst_len, tmp_inst, compc, compl) > 0) && ((!best || (oidcmp2(inst_len, tmp_inst, inst_len, best_inst) < 0)))) { best = data; for (i = 0; i < inst_len; i++) best_inst[i] = tmp_inst[i]; best_inst_len = inst_len; } } [Stuff] } /* * ipAddrEntry_build_instance * * Internal routine to build a table instance */ static ipAddrEntry_build_instance( struct mib_ipaddrreq *data, OIDC_T *tmp_inst) { int rc; return (rc = ip_to_rlist((UINT_32_T) data->ia_iaddr.s_addr, tmp_inst)); } //-------------------------------------------- //--------------------- 2 -------------------- // File if.c /* * ifEntry_next_async * * Common NEXT method routine for ifEntry Table */ void ifEntry_next_async(OIDC_T lastmatch, int compc, OIDC_T *compl, SNMP_PKT_T *pktp, VB_T *vbp) { #define ifEntry_INSTANCE_LEN 1 struct mib_ifreq *data, *best; OIDC_T tmp_inst[ifEntry_INSTANCE_LEN]; OIDC_T best_inst[ifEntry_INSTANCE_LEN]; int i; int error, no; unsigned inst_len; unsigned best_inst_len; /* find all the varbinds that share the same getproc and instance */ group_by_getproc_and_instance(pktp, vbp, compc, compl); /* generate the instance of each table entry and find the * smallest instance that's larger than compc/compl */ best = 0; /* This loop needs to iterate over each entry in your table */ for (data = iftab, no = ifnumber; no; data++,no--) { inst_len = 1; tmp_inst[0] = data->ie_iindex; if ((oidcmp2(inst_len, tmp_inst, compc, compl) > 0) && ((!best || (oidcmp2(inst_len, tmp_inst, inst_len, best_inst) < 0)))) { best = data; for (i = 0; i < inst_len; i++) best_inst[i] = tmp_inst[i]; best_inst_len = inst_len; } } [Stuff] } //-------------------------------------------- It seems that both ipAddrEntry_next_async() and have ifEntry_next_async() similar purpose. However, ipAddrEntry_next_async() is using the ipAddrEntry_build_instance() method, ifEntry_next_async() is not using a similar method. What causes the difference between two these next_async methods? May we write ipAddrEntry_next_async() without build_instance()? May we write ifEntry_next_async() with build_instance()? Why to use the build_instance() method? ================================== Alex Vinokur mailto:alexvn@bigfoot.com http://up.to/alexvn http://visitweb.com/alexvn.math ================================== --------------------------- Newsgroups: comp.os.vxworks Subject: Looking for VxWorks assignment Date: Thu, 14 Jun 2001 15:57:45 From: "Joseph Sebastian" Organization: Lawrence Berkeley National Laboratory Message-ID: <9ganvf$eqb$1@overload.lbl.gov> Hi All, Experienced over 12 years in embedded programming. 5 years in VxWorks and pSos RTOS. Looking for new assignment, in the bay area. Joseph Sebastian. joseph_ks@hotmail.com _________________________________________________________________________ Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Booting from ATA disk ... crash Date: Thu, 14 Jun 2001 17:24:23 +0100 From: "timallen" Organization: ntlworld News Service Message-ID: References: <9ec53d11.0106140225.75f234a8@posting.google.com> Reply-To: "timallen" Trevor, ... > >As I understand it, VxWorks DOSFS cannot read hard drives which have been > >formatted from DOS, unless you have DOSFS 2.0, which is a free upgrade ... > >if you had support. :-( > > I'm not sure that's the case, we format all our disks with Dos 6.2 and it works > OK. However, we don't boot off them natively (?) we bood dos first and then > use vxload.com to load vxworks. That's simply because it's useful to be > able to use DOS sometimes (repair, inital loading, editing config files in the > field, etc) and so we have a quick timeout config.sys meny to choose either > to automatically start vxworks or stay at the dos shell. Yes, that's what I usually do too, for the same reasons. But that does use DOS to access the drive. Are you then able to access the drive from VxWorks once it has booted (without dosFS2.0)? Regards, Tim. --------------------------- Newsgroups: comp.os.vxworks Subject: TrueFFS INstallation Problem Date: Thu, 14 Jun 2001 12:35:46 -0400 From: Doug Dotson Organization: Verio Message-ID: <3B28E7E2.A56DA2E@digidata.com> Greetings! I am having alot of trouble getting TFFS installed. Followed the instructions and got a successful build. After rebooting, I attempt to create a device using "tffsDevCreate" and get the following: - -> tffsDevCreate Data bus error Exception Program Counter: 0x800a60cc Status Register: 0x3000ff01 Cause Register: 0x0000001c Error address: 0xffffffff, Error ID: 0x0000 80041894 vxTaskEntry +c : shell (1, 0, 0, 0) 800bef98 shell +1c8: 800befbc (eeeeeeee, eeeeeeee, eeeeeeee, eeeeeeee) 800bf1a4 shell +3d4: execute (800fe4c8, 8010b304, 2048, &yypact) 800bf30c execute +c8 : yyparse (80c77d88, ffffffff, 1, ffffffff) 800ef040 yyparse +888: 800ed400 (eeeeeeee, 0, eeeeeeee, 0) 800ed5c4 yystart +89c: tffsDevCreate (140, 258, 80077aec, 0) 800a1a58 tffsDevCreate +4c : flCall (80c77cc0, 0, 80cfe920, 96) 800a73bc flCall +248: 800a6540 (0, 0, 8007829c, 0) 800a6588 tffscmpWords +270: flMount (80124c60, 800a6454, 0, 0) 800f2050 flMount +30 : flIdentifyFlash (0, 0, 80c77b80, 0) 800a617c flIdentifyFlash+38 : 800a609c (800ec5e8, 800ec590, 800ea708, 800e990c) My specific configuration is: 1) Board is a Galileo EV-64240 with resident Intel E28F640J3A-120 flash. 2) I modified mv177-sysTtfs.c file to suit. 3) I'm using the cfiscs.c MTD. I put in a few debug statements and have found that the rfaRegister() function is being called. I also put some debug statements in the flRegisterCFISCS() and cfiscsIdentfy() routines and they are never called. Since the exception is occurring at 0xffffffff it appears to me that the function pointers to the MTD functions are never being set up but I can't figure out how this is supposed to be done. Thanks, Doug Douglas P. Dotson Digi-Data Corporation 8580 Dorsey Run Rd. Jessup, MD 20794 301-498-0200 x248 --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Regarding Sending a Packet after a fixed interval of time Date: 14 Jun 2001 09:35:20 -0700 From: tlksk@hotmail.com (Kumar) Organization: http://groups.google.com/ Message-ID: References: <9g9oa5$r6e$1@overload.lbl.gov> Hi Deepesh, Since the routines given in the wdStart run at interrupt level. Never use the sendto or any blocking calls. Better use someother way of tackiling this issuse. Either by semGive or so. Please refer the watchdog timer and ISR manual section in the programmers guide Kumar "Deepesh Aggarwal" wrote in message news:<9g9oa5$r6e$1@overload.lbl.gov>... > This is a multi-part message in MIME format. > > ------=_NextPart_000_00E4_01C0F4C7.E0559BF0 > Content-Transfer-Encoding: quoted-printable > Content-Type: text/plain; > charset="iso-8859-1" > > Hi All > > I want to "send multicast Packets after a specified fixed interval of = > time repeatedly" . > So to serve my purpose i write a program whose code looks like .. > > *************************************************************************= > ****************** > > #include "vxWorks.h" > #include "taskLib.h" > #include "socket.h" > #include "netinet/in.h" > #include "stdio.h" > #include "stdlib.h" > #include "string.h" > #include "sockLib.h" > #include "inetLib.h" > #include "ioLib.h" > #include "routeLib.h" > #include "logLib.h" > #include "wdLib.h" > > #define SECONDS (15) > > WDOG_ID myWatchDogId; > > struct sockaddr_in sin; > struct sockaddr_in toAddr; > int toAddrLen; > int sockDesc; > char * sendBuf=3D"Hello"; > int sendLen =3Dsizeof(sendBuf); > int len; > > STATUS trial(void); > STATUS mcastSend (char * mcastAddr, USHORT mcastPort); > > STATUS multi(void) > { > routeAdd("224.0.0.0","157.227.2.1"); > taskSpawn ("maintask", 100, 0, 20000,mcastSend,"224.0.0.23" = > ,"8769",0,0,0,0,0,0,0,0); > return (OK); > } > =20 > STATUS mcastSend(char * mcastAddr, USHORT mcastPort) > { > =20 > /* create a send and recv socket */ > if ((sockDesc =3D socket (AF_INET, SOCK_DGRAM, 0)) < 0 ) > { > printf (" cannot open send socket\n"); > return (ERROR); > } > > /* zero out the structures */ > bzero ((char *)&sin, sizeof (sin)); > bzero ((char *)&toAddr, sizeof (toAddr)); > > sin.sin_family =3D AF_INET; > sin.sin_addr.s_addr =3D INADDR_ANY; > sin.sin_port =3D htons(0); > > if (bind(sockDesc, (struct sockaddr *)&sin, sizeof(sin)) !=3D 0) > { > perror("bind"); > if (sockDesc) > close (sockDesc); > return (ERROR); > } > > toAddrLen =3D sizeof(struct sockaddr_in); > toAddr.sin_family =3D AF_INET; > /* initialize the address to the send */ > toAddr.sin_addr.s_addr =3D inet_addr (mcastAddr); > > printf("addr is %d",toAddr.sin_addr.s_addr); > /* initialize the port to send */ > toAddr.sin_port =3D htons(mcastPort); > > if ((myWatchDogId =3D wdCreate( )) =3D=3D NULL) > {=20 > perror("Error in creating Timer\n"); > return (ERROR); > } > > if (wdStart (myWatchDogId, sysClkRateGet( ) * SECONDS,trial,0) =3D=3D = > ERROR) > { > perror("Error in starting Timer\n"); =20 > return(ERROR); > } > > } > > > STATUS trial(void) > { > /*printf ("sending \n"); > printf ("Sending %s of length %d using Socket descriptor = > %d\n",sendBuf,sendLen,sockDesc );*/ > > if ((len =3D sendto (sockDesc, sendBuf, sendLen, 0,(struct sockn = > *)&toAddr, toAddrLen)) < 0 ) > { > printf("mcastSend sendto errno:0x%x\n", errno ); > } > > if (wdStart (myWatchDogId, sysClkRateGet( ) * SECONDS,trial,0) =3D=3D = > ERROR) > { > return(ERROR); > } > > } > > *************************************************************************= > ********************** > > I downloaded the application on target with the help of target server = > and tried to call "multi" , it works fine . i mean it was sending the = > packet after 15 seconds ..but=20 > after sending two or three packets my target server gets killed and when = > again i tried to launch it ....it gives an error WTX_ERROR ( Connection = > busy ..). > but i didn't get any error on target .and target was still sending the = > packets ( That's ok .that's the behavior i want..). .until i explicitly = > shut it down... > I m using P-III as a target with an ethernet card FEI ( Fast Intel Ether = > Express PRO100B PCI ). > > Can anyone please explain me the reason why Target server is behaving = > like that.? and where and what can be the problem..?....or suggest me = > some better way to do the same... > =20 > Eagerly waiting for ur precious advice > > Thanks n Happy Working=20 > Deepesh > > > ------=_NextPart_000_00E4_01C0F4C7.E0559BF0 > Content-Transfer-Encoding: quoted-printable > Content-Type: text/html; > charset="iso-8859-1" > > > > http-equiv=3DContent-Type> > > > > >
Hi All
>
 
>
I want to = > "send multicast Packets=20 > after a specified fixed interval of time repeatedly" = > .
>
So to serve my purpose i write a = > program whose=20 > code looks like ..
>
 
>
size=3D2>****************************************************************= > ***************************
>
 
>
#include "vxWorks.h"
#include=20 > "taskLib.h"
#include "socket.h"
#include = > "netinet/in.h"
#include=20 > "stdio.h"
#include "stdlib.h"
#include "string.h"
#include=20 > "sockLib.h"
#include "inetLib.h"
#include "ioLib.h"
#include=20 > "routeLib.h"
#include "logLib.h"
#include "wdLib.h"
>
 
>
#define SECONDS (15)
>
 
>
WDOG_ID myWatchDogId;
>
 
>
struct sockaddr_in sin;
struct = > sockaddr_in=20 > toAddr;
int toAddrLen;
int sockDesc;
char * = > sendBuf=3D"Hello";
int=20 > sendLen =3Dsizeof(sendBuf);
int len;
>
 
>
STATUS trial(void);
STATUS mcastSend = > (char *=20 > mcastAddr, USHORT mcastPort);
>
STATUS=20 > multi(void)
{
 routeAdd("224.0.0.0","157.227.2.1");
 t= > askSpawn=20 > ("maintask", 100, 0, 20000,mcastSend,"224.0.0.23"=20 > ,"8769",0,0,0,0,0,0,0,0);
>
 return = > (OK);
}
 
 face=3DArial size=3D2>STATUS mcastSend(char * mcastAddr, USHORT=20 > mcastPort)
{
 
/* create a send and recv socket */
if=20 > ((sockDesc =3D socket (AF_INET, SOCK_DGRAM, 0)) < 0 )
{
printf = > (" cannot=20 > open send socket\n");
return (ERROR);
}
>
 
>
/* zero out the structures */
bzero = > ((char=20 > *)&sin, sizeof (sin));
bzero ((char *)&toAddr, sizeof=20 > (toAddr));
>
 
>
sin.sin_family =3D = > AF_INET;
sin.sin_addr.s_addr =3D=20 > INADDR_ANY;
sin.sin_port =3D htons(0);
>
 
>
if (bind(sockDesc, (struct sockaddr = > *)&sin,=20 > sizeof(sin)) !=3D 0)
{
perror("bind");
if (sockDesc)
close=20 > (sockDesc);
return (ERROR);
}
>
 
>
toAddrLen =3D sizeof(struct=20 > sockaddr_in);
toAddr.sin_family =3D AF_INET;
/* initialize the = > address to=20 > the send */
toAddr.sin_addr.s_addr =3D inet_addr = > (mcastAddr);
>
 
>
printf("addr is = > %d",toAddr.sin_addr.s_addr);
/*=20 > initialize the port to send */
toAddr.sin_port =3D=20 > htons(mcastPort);
>
 
>
if ((myWatchDogId =3D wdCreate( )) = > =3D=3D NULL)
{=20 >
perror("Error in creating Timer\n");
return = > (ERROR);
}
>
 
>
if (wdStart (myWatchDogId, = > sysClkRateGet( ) *=20 > SECONDS,trial,0) =3D=3D ERROR)
{
perror("Error in starting = > Timer\n"); =20 >
return(ERROR);
}
>
 
>
}
>
 
>

STATUS trial(void)
{
/*printf = > ("sending=20 > \n");
printf ("Sending %s of length %d using Socket descriptor=20 > %d\n",sendBuf,sendLen,sockDesc );*/
>
 
>
if ((len =3D sendto (sockDesc, sendBuf, = > sendLen,=20 > 0,(struct sockn *)&toAddr, toAddrLen)) < 0 = > )
{
printf("mcastSend=20 > sendto errno:0x%x\n", errno );
}
>
 
>
if (wdStart (myWatchDogId, = > sysClkRateGet( ) *=20 > SECONDS,trial,0) =3D=3D ERROR)
{
return(ERROR);
}
>
 
>
}
>
size=3D2>
************************************************************= > ***********************************
>
 
>
I downloaded the application on target = > with the=20 > help of target server and tried to call "multi" , it works fine . i mean = > it was=20 > sending the packet after 15 seconds ..but
>
after sending two or three packets my = > target server=20 > gets killed and when again i tried to launch it ....it gives an error = > WTX_ERROR=20 > ( Connection busy ..).
>
but i didn't get any error on target=20 > .and target was still sending the packets ( That's ok .that's the = > behavior=20 > i want..). .until i explicitly shut it down...
>
I m using P-III as a target with an = > ethernet card=20 > FEI ( Fast Intel Ether Express PRO100B PCI ).
>
 
>
Can anyone please explain me the reason = > why Target=20 > server is behaving like that.? and where and what can be the = > problem..?....or=20 > suggest me some better way to do the same...
>
 
>
Eagerly waiting for ur precious = > advice
>
 
>
Thanks n size=3D2>Happy=20 > Working
>
Deepesh
>
 
> > ------=_NextPart_000_00E4_01C0F4C7.E0559BF0-- --------------------------- Newsgroups: comp.os.vxworks Subject: Maintianing Unresolved Symbols Date: Thu, 14 Jun 2001 11:47:50 -0500 From: "Chris Richards" Message-ID: <3b28eab6$0$94313$39cecf19@news2.twtelecom.net> Howdy. There are several utilities that I want to use during development; they aren't used, normally. I would like them to be there when I use Windshell or telnet to the machine. Unfortunatley (in this case), ld is doing it's job and removing the .o's that are never referenced when building the VxWorks image. To give you more information, we use a combination of command-line make and project (Tornado 2.0). Individual components are built into archive libraries. The project includes the archive libraries. In this way, we are able to achieve automated builds and limit the exposure of project to the build dude. Using the .lst file for .o's doesn't work because the contents of the .lst file must have only filenames--no paths. It expects ALL .o's in one spot. The ld flag --whole-archive exists, but is for the whole line as opposed to being turned off by --no-whole-archive (which doesn't exist). I think the 'ld' options have been exhausted. There doesn't seem to be a way to do this short of specifying either a) all .o's on the ld line, b) making a .lst and moving all .o's into one directory, or c) creating a "hack" src file which references a noop function in each .o that I ant to keep. Thoughts? /cjr --------------------------- Newsgroups: comp.os.vxworks Subject: Re: DP national driver driver for VxWorks Date: Thu, 14 Jun 2001 13:38:59 -0400 From: Ian Melhuish Organization: Teklogix Inc. Message-ID: <3B28F6B3.40232609@teklogix.com> References: <3B276BD2.88CF4514@teklogix.com> I managed to get this going by setting WDB_COMM_TYPE to WDB_COMM_NETWORK (which is the default in configall.h) instead of WDB_COMM_END which I had specified in config.h. Ian Mike Kordik wrote: > Yes. I do not have a fix for it yet. > "Ian Melhuish" wrote in message > news:3B276BD2.88CF4514@teklogix.com... > > I am using the driver and also cannot get a target server connected. Did > your > > problem arise after installong the 2.02 patch? > > > > Mike Kordik wrote: > > > > > Is anyone using this driver (DP83815)? The target server will not > connect to > > > this driver after you install the Tornado 2.02 patch. Does anyone know > why? > > > Is there a workaround? > > > > > > Thanks, > > > Mike > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Proxy ARP problems: workaround needed Date: Thu, 14 Jun 2001 13:57:07 -0400 From: Bill Fulton Organization: Raytheon Company Message-ID: <3B28FAF3.F8E9A38A@raytheon.com> References: <3B211164.3AA58342@auriga.ru> <3B268599.3E163DD4@raytheon.com> <3B2770FD.C9BE0382@auriga.ru> If I'm understanding your configuration correctly ... then it's a simple fix - you just need to publish the *ethernet interface* MAC address in response to an ARP, not the backplane "MAC" address. i.e. arpAdd( "192.168.1.151", "08:00:3e:2d:80:00", 8 ) So .... after doing that, here's what should happen: > Some "peer host" on the local network ARP's for the 192.168.1.151 host. > Your card sees that it is proxying that IP address, so it responds with the local ethernet interface's MAC address (not the backplane pseudo MAC address). > The peer host, being none the wiser, passes the IP message to your local ethernet interface. > IFF your local IP stack is configured to be a router, then when it gets the IP message it sees that the destination IP address is not itself, it looks in its routing table to find another interface to forward it to, sees the mn interface, and forwards it on. Note: I have not worked with backplane interfaces (only read about them, when trying to get proxy ARP working) - I'm assuming that routing works the same for ethernet and backplane interfaces (that's my understanding of why vxWorks supplies the backplane "network interface" in the first place). BUT ... here's where my unfamiliarity with backplane interfaces becomes a roadblock. If backplane and ethernet interfaces are treated the same as far as routing goes, then I don't understand the IP addresses given to your interfaces - both interfaces have the same network address, which makes routing not work (at least according to "normal" IP routing). So .... maybe it's not that easy. Either I'm not understanding "special" backplane IP addressing and routing (can someone else help, here?), or you have the two interfaces configured incorrectly - i.e. both interfaces have the same IP network address (192.168.1.0/24). Hope it helps ... but maybe my unfamiliarity with backplane interface configuration is sending you in the wrong direction. Pavel Kovalev wrote: > Hello Bill, > > > The solution that we came up with was to use the ATF_PUBL flag with the > > arpAdd() function. > > This seems to, in effect, have all the desired consequences - our card > > responds to an ARP request for the defined IP address with its own MAC > > address, and then routes the IP message to the appropriate place. > > > > This requires a manual setup of the ARP table ... but it seems to work > > well. (Of course, you have to have routing enabled.) > > > > In our case, though, we were forwarding messages between ethernet > > networks, not a network and a backplane. It's my understanding that the > > routing should still work, but .... > > Our backplane network fully emulates the ethernet network, except support for > multicasting. So, in theory, what works in your case > should work for us. Though, when use arpAdd() I get the following: > > -> ifShow > dc (unit number 0): > Flags: (0x8063) UP BROADCAST MULTICAST ARP RUNNING > Type: ETHERNET_CSMACD > Internet address: 192.168.1.154 > Broadcast address: 192.168.1.255 > Netmask 0xffffff00 Subnetmask 0xffffff00 > Ethernet address is 08:00:3e:2d:80:00 > Metric is 0 > Maximum Transfer Unit size is 1500 > 6107 packets received; 864 packets sent > 4683 multicast packets received > 1 multicast packets sent > 0 input errors; 0 output errors > 0 collisions; 0 dropped > lo (unit number 0): > Flags: (0x8069) UP LOOPBACK MULTICAST ARP RUNNING > Type: SOFTWARE_LOOPBACK > Internet address: 127.0.0.1 > Netmask 0xff000000 Subnetmask 0xff000000 > Metric is 0 > Maximum Transfer Unit size is 32768 > 0 packets received; 0 packets sent > 0 multicast packets received > 0 multicast packets sent > 0 input errors; 0 output errors > 0 collisions; 0 dropped > mn (unit number 1): > Flags: (0x63) UP BROADCAST ARP RUNNING > Type: ETHERNET_CSMACD > Internet address: 192.168.1.155 > Broadcast address: 192.168.1.255 > Netmask 0xffffff00 Subnetmask 0xffffff00 > Ethernet address is 22:22:22:00:00:01 > Metric is 0 > Maximum Transfer Unit size is 4058 > 2 packets received; 1 packets sent > 2 multicast packets received > 0 multicast packets sent > 0 input errors; 0 output errors > 0 collisions; 0 dropped > -> arpAdd "192.168.1.151", "22:22:22:00:00:04", 8 > value = 0 = 0x0 > -> arpShow > > LINK LEVEL ARP TABLE > destination gateway flags Refcnt Use Interface > -------------------------------------------------------------------------- > 192.168.1.151 22:22:22:00:00:4 4c05 0 0 dc0 > 192.168.1.253 00:01:02:1e:13:4 405 0 851 dc0 > -------------------------------------------------------------------------- > value = 75 = 0x4b = 'K' > -> routeShow > > ROUTE NET TABLE > destination gateway flags Refcnt Use Interface > ---------------------------------------------------------------------------- > 192.168.1.0 192.168.1.154 101 0 0 dc0 > ---------------------------------------------------------------------------- > > ROUTE HOST TABLE > destination gateway flags Refcnt Use Interface > ---------------------------------------------------------------------------- > 127.0.0.1 127.0.0.1 5 0 0 lo0 > ---------------------------------------------------------------------------- > value = 77 = 0x4d = 'M' > -> ping "192.168.1.151" > PING 192.168.1.151: 56 data bytes > no answer from 192.168.1.151 > value = -1 = 0xffffffff = udbinfo + 0xffd7666b > > Obviously, when I call arpAdd() as shown above I do not get what > I need: 192.161.1.151 resides on the proxy network (mn1 - CompactPCI), > and the ARP entry is associated with the ethernet network (dc0). > Therefore, to achieve the desired configuration I use the following > commands: > > -> ifAddrSet "dc0", "127.127.127.127" > value = 0 = 0x0 > -> ifAddrSet "mn1", "192.168.1.155" > value = 0 = 0x0 > -> arpAdd "192.168.1.151", "22:22:22:00:00:04", 8 > value = 0 = 0x0 > -> routeDelete "192.168.1.0", "192.168.1.155" > value = 0 = 0x0 > -> mRouteAdd "192.168.1.151", "192.168.1.155", 0xffffffff, 0, 0 > value = 0 = 0x0 > -> ifAddrSet "dc0", "192.168.1.154" > value = 0 = 0x0 > -> routeShow > > ROUTE NET TABLE > destination gateway flags Refcnt Use Interface > ---------------------------------------------------------------------------- > 192.168.1.0 192.168.1.154 101 0 0 dc0 > 192.168.1.151 192.168.1.155 3 0 0 mn1 > ---------------------------------------------------------------------------- > > ROUTE HOST TABLE > destination gateway flags Refcnt Use Interface > ---------------------------------------------------------------------------- > 127.0.0.1 127.0.0.1 5 0 0 lo0 > ---------------------------------------------------------------------------- > value = 77 = 0x4d = 'M' > -> arpShow > > LINK LEVEL ARP TABLE > destination gateway flags Refcnt Use Interface > -------------------------------------------------------------------------- > 192.168.1.151 22:22:22:00:00:4 4c05 0 0 mn1 > -------------------------------------------------------------------------- > value = 75 = 0x4b = 'K' > > Though, while now the configuration seems to be absolutely correct, > VxWorks does not reply to ARP requests originated from one > network and targeted to another network. Thus, 192.168.1.253 > (residing on Ethernet) and 192.168.1.151 do not get ARP replies > when they want to communicate with each other. If I setup > ARP tables manually on these hosts then they can successfully > ping each other; though, this is not what I need. The proxy > ARP server on 192.168.1.154/155 does not seem to work at all. > > I am wondering what is the difference between our configurations. > From your message it appears that the same setup works on your > system. Am I missing something? I am using VxWorks 5.4/Tornado II. > Do you have some special patches applied, or something like that? > > Thanks, > Pavel. > > > Bill > > > > Pavel Kovalev wrote: > > > > > Hello All, > > > > > > We have a custom network driver for a CompactPCI backplane network > > > and want to configure a proxy ARP server on the CompactPCI host > > > board to make the backplane nodes visible on the Ethernet network > > > (without additional routing). > > > > > > Tornado 2/VxWorks 5.4 support this configuration only for > > > the shared memory network drivers supplied by WRS. Attempts > > > to use proxyArpLib or configure all the necessary routing > > > manually fail with various errors. > > > > > > I studied various materials available on WindSurf but haven't > > > found a solution for this problem. There is an official SPR > > > describing this problem (SPR# 28602), but it looks like it > > > hasn't been processed/fixed yet. There were some discussions > > > of this problem in the newsgroup in the past, but no good > > > solutions have been proposed. > > > > > > I spent some time investigating this problem, and found that when > > > I add a specific route to a node residing on the proxy network > > > (route is added on the host which has to work proxy ARP server), > > > mRouteAdd() creates the route with RTF_GATEWAY set. This prevents > > > arpAdd() from adding a permanent published entry for the node. > > > One solution is to develop a better version of mRouteAdd(), but > > > I am not yet ready/willing to do so - there is no warranty that > > > other hacks won't be needed. > > > > > > So, I am wondering if anybody has managed to solve this problem. > > > > > > Thanks, > > > Pavel. --------------------------- Newsgroups: comp.os.vxworks Subject: What is meaning of this error Date: Thu, 14 Jun 2001 11:32:24 -0700 From: "Shesha Bhushan" Organization: Arizona State University Message-ID: <9gavv2$heb$1@news.asu.edu> Hi, Gurus I'm a novice to vxWorks. when i compile usrConfig.c i get the following error. I have added usrConfig.c in to my bootable project on Tornado. ERROR = "make: *** No rule to make target `usrDepend.c', needed by `usrConfig.o'. Stop." where usrDepend.c is an include in usrConfig.c. What is the implication of this error. --------------------------- Newsgroups: comp.os.vxworks Subject: Prevent "Press any key to stop auto-reboot" prompt Date: Thu, 14 Jun 2001 11:38:12 -0700 From: Sridhar Ramanathan Organization: QUALCOMM Incorporated Message-ID: <3B290494.AC3A5FCA@qualcomm.com> Hello, Is there a way to disable the "Press any key to stop auto-reboot" prompt during the boot sequence of vxWorks? I am connecting a GPS receiver ( which sends out 15bytes of TOD data every 2 secs ) to the com port of the target running vxWorks. Doing this prevents the target from booting up because every time it tries to boot up, the data from the GPS receiver acts as the input to the "Press any key to stop auto-reboot" prompt and causes the boot up sequence to stop. Alternately, is there a way of pointing the prompt to look for data on a non-existing com port like say /tyCo/1? Thanks in advance - -Sridhar --------------------------- Newsgroups: comp.os.vxworks Subject: Windview wooos ? [Missing events, timing off...]... Date: Thu, 14 Jun 2001 20:36:53 GMT From: "Todd Sandor" Organization: Magma Communications Ltd. Message-ID: I've been trying to use Windview with Tornado 2 along with the 8260 BSP. I'm experiencing some issues that I'm trying to understand - please help me if you can ....thxs... If Windview does in fact loose events, it is of questional use as a development tool... 1. Missing Events: As a test, I start a watchdog with a 1 second internal and the routine that gets executed simply performs a msgQSend() and restarts the 1 second watchdog [to call the same routine again]. Windview sometimes records the msgQSend() and wdStart() and sometimes doesn't [I don't see these every second]. When I do see them, they get "recorded" against the "Idle" task. Recording these against the Idle task may be explained as being performed since the watchdog callback function is executed as an ISR - but I don't know why they would not alway be recorded/shown. The msgQReceive() performed by the destination task of the msgQSend() *is* always recorded - soo ...Does anyone know of any problems related to Windview missing events? 2. Event Timings - I believe this is a 8260 BSP Aux Clock issue, but the time between Windview events recorded seem to be 1/4 of what they should be. For example, the time between the msgQReceive() events of the test explained in issue #1 is ~250 mseconds rather then 1 second. Has anyone noticed similiar time between event recording problems? Cheers... --------------------------- Newsgroups: comp.os.vxworks Subject: Re: DP national driver driver for VxWorks Date: Thu, 14 Jun 2001 16:58:57 -0400 From: "Mike Kordik" Message-ID: References: <3B276BD2.88CF4514@teklogix.com> <3B28F6B3.40232609@teklogix.com> Thanks for the info. I will give that a try. BTW it did work for me with specifying WDB_COMM_END when I have Tornado 2.0 but when I went to 2.02 it quit working. My contact has told me that they think if I recompiled the source after installing 2.02 that it may fix the problem. The theory is that when WRS add the NPT (network toolkit) they modified header files and that the network driver will have to be recompiled. Unfortunately I do not have the source. Do you? Thanks, Mike "Ian Melhuish" wrote in message news:3B28F6B3.40232609@teklogix.com... > I managed to get this going by setting WDB_COMM_TYPE to WDB_COMM_NETWORK > (which is the default in configall.h) instead of WDB_COMM_END which I had > specified in config.h. > > Ian > > Mike Kordik wrote: > > > Yes. I do not have a fix for it yet. > > "Ian Melhuish" wrote in message > > news:3B276BD2.88CF4514@teklogix.com... > > > I am using the driver and also cannot get a target server connected. Did > > your > > > problem arise after installong the 2.02 patch? > > > > > > Mike Kordik wrote: > > > > > > > Is anyone using this driver (DP83815)? The target server will not > > connect to > > > > this driver after you install the Tornado 2.02 patch. Does anyone know > > why? > > > > Is there a workaround? > > > > > > > > Thanks, > > > > Mike > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Network performance question (UDP/IP) Date: 14 Jun 2001 14:09:22 -0700 From: static@ieee.org (Andrew Spiehler) Organization: http://groups.google.com/ Message-ID: <59b46a44.0106141309.6037709c@posting.google.com> Hi, very green with VxWorks and in need of much help: I'm broadcasting single frame messages over 10Base-T ethernet using UDP sockets from my host system (NT) to my target system (a 350Mhz PPC SBC in a VME cage). I've also got a packet sniffer running on a 3rd computer. I'm noticing that regardless of the size of the data, either 100 bytes or 1400 bytes, each packet is taking from 2 to 4 ms to reach the target. Once the packet is received, my task takes from 6 to 12 microseconds to read the data from the socket and takes another 400 ns or so to process the data before reading the next message from the socket. When I watch the process in Windview, I see tNetTask running to (I assume) receive the packet from the ethernet hardware and put it into the socket buffer. It then does a semGive and a semTake and blocks. The semGive releases a semaphore that causes my task to run after tNetTask blocks. I assume it did a semtake inside the recvfrom() function. From there it processes the data and does another recvfrom() at which point it blocks. Then windview shows the system IDLE task running for about 4 ms until it does a semgive, casuing tNetTask to unblock and run, starting the whole cycle over again. My questions are: 1. Why am I seeing a lag of 2 to 4 ms regardless of the packet size? 2. Why is tNetTask waiting for a semaphore from IDLE? 3. When I transmit a large number of 1400 byte packets, I occasionally lose packets. The packets are going across the network according to the packet sniffer. I don't think it's the socket receive buffer because my task finishes it's processing on the first packet long before the next packet arrives. Could it be some memory pool in the network stack? Any help would be greatly appreciated. Sincerely, Andrew Spiehler --------------------------- Newsgroups: comp.os.vxworks Subject: "pcb cache lookups failed" Date: 14 Jun 2001 14:15:23 -0700 From: static@ieee.org (Andrew Spiehler) Organization: http://groups.google.com/ Message-ID: <59b46a44.0106141315.1c1b78ab@posting.google.com> udpstatShow() returns the following information: "X pcb cache lookups failed" where X is the same as the number of incoming udp packets. Can someone tell me what that means and how I fix it? Thanks. Andrew Spiehler --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Prevent "Press any key to stop auto-reboot" prompt Date: 14 Jun 2001 15:21:30 -0700 From: tlksk@hotmail.com (Kumar) Organization: http://groups.google.com/ Message-ID: References: <3B290494.AC3A5FCA@qualcomm.com> Hi Sridhar Either edit the bootConfig.c so the autoboot takes a value of ZERO or make no autoboot in the boot parameter line. Kumar Sridhar Ramanathan wrote in message news:<3B290494.AC3A5FCA@qualcomm.com>... > Hello, > > Is there a way to disable the "Press any key to stop auto-reboot" prompt > during the boot sequence of vxWorks? > > I am connecting a GPS receiver ( which sends out 15bytes of TOD data > every 2 secs ) to the com port of the target running vxWorks. Doing this > prevents the target from booting up because every time it tries to boot > up, the data from the GPS receiver acts as the input to the "Press any > key to stop auto-reboot" prompt and causes the boot up sequence to stop. > > Alternately, is there a way of pointing the prompt to look for data on a > non-existing com port like say /tyCo/1? > > Thanks in advance > > -Sridhar --------------------------- Newsgroups: comp.os.vxworks Subject: Re: What is meaning of this error Date: 14 Jun 2001 15:25:04 -0700 From: tlksk@hotmail.com (Kumar) Organization: http://groups.google.com/ Message-ID: References: <9gavv2$heb$1@news.asu.edu> Hi Shesha Check whether the usrDepend.c file is prsent in the target/src/config. This file is requred for some configuration checks which you can go thru in the file usrConfig.c And this file is required. Kumar "Shesha Bhushan" wrote in message news:<9gavv2$heb$1@news.asu.edu>... > Hi, Gurus > > I'm a novice to vxWorks. > > when i compile usrConfig.c i get the following error. I have added > usrConfig.c in to my bootable project on Tornado. > > ERROR = "make: *** No rule to make target `usrDepend.c', needed by > `usrConfig.o'. Stop." > > where usrDepend.c is an include in usrConfig.c. > What is the implication of this error. --------------------------- Newsgroups: comp.os.vxworks Subject: IGMP Date: Thu, 14 Jun 2001 23:07:13 GMT From: Tony Rodayer Organization: Guest of ProXad - France Message-ID: <3B2943FE.1020902@free.fr> Hi, Is someone has information to implement IGMP on VxWorks? Thanks. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Prevent "Press any key to stop auto-reboot" prompt Date: Thu, 14 Jun 2001 17:52:00 -0700 From: "Michael R. Kesti" Organization: MK Associates Message-ID: <3B295C30.9CAEAD89@gv.net> References: <3B290494.AC3A5FCA@qualcomm.com> Reply-To: mkesti@gv.net Sridhar Ramanathan wrote: >Hello, > >Is there a way to disable the "Press any key to stop auto-reboot" prompt >during the boot sequence of vxWorks? > >I am connecting a GPS receiver ( which sends out 15bytes of TOD data >every 2 secs ) to the com port of the target running vxWorks. Doing this >prevents the target from booting up because every time it tries to boot >up, the data from the GPS receiver acts as the input to the "Press any >key to stop auto-reboot" prompt and causes the boot up sequence to stop. It is best to let the console be nothing but the console, and you should be using another serial port for your application. I gather, though, that your target provides only one serial port. You could, however, replace your target's BSP boot ROM with a boot ROM created by linking a downloadable application Tornado project with a bootable vxWorks Tornado project configured to not include the target shell. This has the effect of replacing the BSP boot process with a program that simply runs at target power up. The problem is that you cannot debug such a ROM, so, during developement use a BSP boot ROM but disconnect the GPS receiver and connect a terminal to the port until the target boots the debug environment, again configured to not include the target shell. >Alternately, is there a way of pointing the prompt to look for data on a >non-existing com port like say /tyCo/1? I suppose that one might hack the serial driver to include a null tyCo, but that seems a bit over the top. - -- ======================================================================== Michael Kesti | "And like, one and one don't make | two, one and one make one." mkesti@gv.net | - The Who, Bargain --------------------------- Newsgroups: comp.protocols.snmp,comp.os.psos,comp.os.vxworks Subject: Re: build_instance() in next_async() Date: Thu, 14 Jun 2001 22:07:21 -0500 From: Jim Jones Organization: netINS, Inc. Message-ID: <3B297BE4.5D015EF8@icon-labs.com> References: <3B28EE22.94383912@bigfoot.com> Reply-To: jim_jones@icon-labs.com Alex Vinokur wrote: > ============================ > pRISM+ for pSOSystem/PowerPC > pRISM+ 1.2.3 . pSOSystem 2.2.2 > Diab Data D-CC Optimizing C compiler > SNMP Release 7.4 (SNMPv1) > Host : Windows98 > Target : PowerPC (MPC860) > ============================ > > Here are fragments of files that implement > a sample SNMP agent using pSOSystem > > //--------------------- 1 -------------------- > // File ip.c > /* > * ipAddrEntry_next_async > * > * Common GET method routine for ipAddrEnt table > */ > void > ipAddrEntry_next_async(OIDC_T lastmatch, > int compc, > OIDC_T *compl, > SNMP_PKT_T *pktp, > VB_T *vbp) > { > #define ipAddrEntry_INSTANCE_LEN 4 > struct mib_ipaddrreq *data, *best; > OIDC_T tmp_inst[ipAddrEntry_INSTANCE_LEN]; > OIDC_T best_inst[ipAddrEntry_INSTANCE_LEN]; > int i,no; > int error; > unsigned inst_len; > unsigned best_inst_len; > > /* find all the varbinds that share the same getproc and instance */ > group_by_getproc_and_instance(pktp, vbp, compc, compl); > > read_ipaddrent(); > > /* generate the instance of each table entry and find the > * smallest instance that's larger than compc/compl */ > best = 0; > /* This loop needs to iterate over each entry in your table */ > for (data = ipatab, no = ipaddrlen; no; data++,no--) { > > inst_len = ipAddrEntry_build_instance(data, tmp_inst); > > if ((oidcmp2(inst_len, tmp_inst, compc, compl) > 0) && > ((!best || (oidcmp2(inst_len, tmp_inst, > inst_len, best_inst) < 0)))) { > best = data; > for (i = 0; i < inst_len; i++) > best_inst[i] = tmp_inst[i]; > best_inst_len = inst_len; > } > } > > [Stuff] > } > > /* > * ipAddrEntry_build_instance > * > * Internal routine to build a table instance > */ > static > ipAddrEntry_build_instance( > struct mib_ipaddrreq *data, > OIDC_T *tmp_inst) > { > int rc; > > return (rc = ip_to_rlist((UINT_32_T) data->ia_iaddr.s_addr, tmp_inst)); > > } > > //-------------------------------------------- > > //--------------------- 2 -------------------- > // File if.c > /* > * ifEntry_next_async > * > * Common NEXT method routine for ifEntry Table > */ > void > ifEntry_next_async(OIDC_T lastmatch, > int compc, > OIDC_T *compl, > SNMP_PKT_T *pktp, > VB_T *vbp) > { > #define ifEntry_INSTANCE_LEN 1 > struct mib_ifreq *data, *best; > OIDC_T tmp_inst[ifEntry_INSTANCE_LEN]; > OIDC_T best_inst[ifEntry_INSTANCE_LEN]; > int i; > int error, no; > unsigned inst_len; > unsigned best_inst_len; > > /* find all the varbinds that share the same getproc and instance */ > group_by_getproc_and_instance(pktp, vbp, compc, compl); > > /* generate the instance of each table entry and find the > * smallest instance that's larger than compc/compl */ > best = 0; > /* This loop needs to iterate over each entry in your table */ > for (data = iftab, no = ifnumber; no; data++,no--) { > inst_len = 1; tmp_inst[0] = data->ie_iindex; > > if ((oidcmp2(inst_len, tmp_inst, compc, compl) > 0) && > ((!best || (oidcmp2(inst_len, tmp_inst, > inst_len, best_inst) < 0)))) { > best = data; > for (i = 0; i < inst_len; i++) > best_inst[i] = tmp_inst[i]; > best_inst_len = inst_len; > } > } > > [Stuff] > } > //-------------------------------------------- > > It seems that both ipAddrEntry_next_async() and have > ifEntry_next_async() > similar purpose. > > However, ipAddrEntry_next_async() is using the > ipAddrEntry_build_instance() method, > ifEntry_next_async() is not using a similar method. > > What causes the difference between two these next_async methods? > May we write ipAddrEntry_next_async() without build_instance()? > May we write ifEntry_next_async() with build_instance()? Yes you can. The -stub generated code assumes that you will write this function. You are looking at code in which someone did that by hand. The difference is that for a simple integer it was overkill to create such a function. But for the ipAddress index, it needs to be broken up into its 4 components and put into an array of 4 integers. The writer took advantage of a utility for that and wrote the build_instance function and included it as a static function in the file. Jim Jones Icon labs > > > Why to use the build_instance() method? > > ================================== > Alex Vinokur > mailto:alexvn@bigfoot.com > http://up.to/alexvn > http://visitweb.com/alexvn.math > ================================== --------------------------- Newsgroups: comp.os.vxworks Subject: editor in vxworks Date: Thu, 14 Jun 2001 19:17:13 -0700 From: "Devatha, Pratap" Organization: Lawrence Berkeley National Laboratory Message-ID: <9gbuk2$94f$1@overload.lbl.gov> Hi I wanted to know if there is any known text editor which can edit files in the vxworks file system residing in the target system. Vxworks is running on Intel Ixp1200 Network Processor. I am also planning to write a text editor for this. Does anyone know the address at which the video memory resides for the Ixp1200. Any help regarding this is appreciated. Thanking you, Pratap --------------------------- Newsgroups: comp.os.vxworks Subject: Re: DP national driver driver for VxWorks Date: Fri, 15 Jun 2001 03:59:13 GMT From: "Martin Usher" Organization: EarthLink Inc. -- http://www.EarthLink.net Message-ID: References: <3B276BD2.88CF4514@teklogix.com> <3B28F6B3.40232609@teklogix.com> Reply-To: "Martin Usher" > "Ian Melhuish" wrote in message > news:3B28F6B3.40232609@teklogix.com... > > I managed to get this going by setting WDB_COMM_TYPE to WDB_COMM_NETWORK > > (which is the default in configall.h) instead of WDB_COMM_END which I had > > specified in config.h. > > That figures. WDB connected directly to the END hooks up the driver outside of, and before, the normal place where END drivers are initialized. They did this so that you could do 'system' level debugging. The Network hookup communicates to the normally initialized driver through the Mux like any other piece of network software so its going to work if the network works. Its only useable with 'task' debugging. There's plenty to go wrong with the END hookup and my guess is that it usually does. I've experienced this problem on several different PowerPC boards using 860, 8260 and 8240 processors interfacing with a variety of NICs (both external and internal to the PPC). I've never bothered to look into it because I have COPS (BDM) debuggers to take care of the heavy stuff leaving the WDB hosted diagnostics for just lightweight and casual use. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Command 'iosFdShow' does not work in Shell Date: Fri, 15 Jun 2001 04:32:00 GMT From: Dan Engholm Organization: AT&T Worldnet Message-ID: <3B299011.4CBC17E5@worldnet.att.net> References: Reply-To: dan.engholm@writeme.com Asaf Sinai wrote: > Hi, > > When I run the command 'iosFdShow' in the Shell, I get: > -> iosFdShow > fd name drv > value = 0 = 0x0 > > But, in the serial console I get all the data. > > Has anyone know why it happens ? I can't say why it happens, but I get around it by typing "_iosFdShow" at the shell prompt. This causes the command to be run on the target rather than on the host. - --Dan --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Windview wooos ? [Missing events, timing off...]... Date: Thu, 14 Jun 2001 22:32:21 -0700 From: DrDiags Organization: laserlink.net Message-ID: <3B299DE5.D02284ED@covad.net> References: Reply-To: drdiags@covad.net Todd, As far as item #1, I have been trying to put together a test case to reproduce this, so that I could submit it for an SPR. I cannot seem to get anything consistent, so if you come up with one, could you send it to me. I am using the SBC8260 BSP from HSI as my platform. So, I guess you are not along on this one.... Todd Sandor wrote: > > I've been trying to use Windview with Tornado 2 along with the 8260 BSP. > I'm experiencing some issues that I'm trying to understand - please help me > if you can ....thxs... If Windview does in fact loose events, it is of > questional use as a development tool... > > 1. Missing Events: As a test, I start a watchdog with a 1 second internal > and the routine that gets executed simply performs a msgQSend() and restarts > the 1 second watchdog [to call the same routine again]. Windview sometimes > records the msgQSend() and wdStart() and sometimes doesn't [I don't see > these every second]. When I do see them, they get "recorded" against the > "Idle" task. Recording these against the Idle task may be explained as > being performed since the watchdog callback function is executed as an ISR - > but I don't know why they would not alway be recorded/shown. The > msgQReceive() performed by the destination task of the msgQSend() *is* > always recorded - soo ...Does anyone know of any problems related to > Windview missing events? > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: internal compiler errors Date: Thu, 14 Jun 2001 22:42:03 -0700 From: DrDiags Organization: laserlink.net Message-ID: <3B29A02B.9B6947BD@covad.net> References: <3B207CA7.D938FC01@mediaone.net> <3B28C4E7.6620AF6B@mediaone.net> Reply-To: drdiags@covad.net Jagan, Do you have the GNAT 95 Ada compiler installed on these machines? If so, somewhere in my reading, there was a mention of problems similar to yours. If I can find the e-mail I saw that spoke to this, I will post it here. But if you don't have this configuration, never mind.......... jagan wrote: > > This has been a very elusive problem. I have tried > it on different PCs and it strikes randomly. It > has happened on almost all PCs that I have tried > on at some time or the other - but usually it > happens on a different pc at a different time. > > At this point I suspect the problem may be a > combination of bash environment (cygwin), cygwin make, > and the tornado compiler cc1.exe. > > When it happens I can get around it by compiling > the file (different file at different times) individually > and then calling make to compile and link all the other > files. > > Any thoughts? > > We sent a message to Wind River, but they refused > to look at the problem because of the mixed environment (I do not use > their project tool.. and their make files). > > -jagan. > > Dave Korn wrote: > > > > "jagan" wrote in message > > news:3B207CA7.D938FC01@mediaone.net... > > > > > > > > > I am seeing an internal compiler error when compiling > > > using Tornado 2 cross compiler for a ppc target in > > > a cygwin environment on NT. > > > > > > I see cc1.exe coring with signal 5. > > > > > > Does anyone know why this may be happening? > > > > Not really. Could be flaky hardware or configuration - try doing the same > > build on a different PC. Show us the actual output that you get - copy + > > paste it from the screen into a reply. > > > > Davek --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Command 'iosFdShow' does not work in Shell Date: Thu, 14 Jun 2001 22:54:38 -0700 From: DrDiags Organization: laserlink.net Message-ID: <3B29A31E.3068BD2E@covad.net> References: Reply-To: drdiags@covad.net Asaf, What version of Tornado? What architecture? There probably is a Tcl file which mimics the target version of the show routine. If you like hacking Tcl code, you might find out why this is so. In the $(WIND_BASE)/host/resource/tcl/shellShowCmd.tcl file, you will find the iosFdShow Tcl procedure: ############################################################################## # # iosFdShow - show all existing file descriptors # # Mutual exclusion note: This procedure is accomplished in two steps. # First, a vector of booleans is obtained from the fdTable to find # out which entries are valid. Then data is collected for each fd marked # as valid in the fdTable. # # Difference from the shell impelmentation: fds are not marked "in", "out" # and "err" if they are bound to stdio devices as the target shell version # of this command would do. # shellproc iosFdShow {args} { global __wtxTsInfo global offset global sizeof if {[lsearch -exact [lindex $__wtxTsInfo 0] RPC] != -1} { set mxGLength 1444 } elseif {[lsearch -exact [lindex $__wtxTsInfo 0] serial] != -1} { set mxGLength 950 } else { set mxGLength 444 } puts stdout [format "%3s %-20s %3s" "fd" "name" "drv"] set maxFiles [expr [wtxGopherEval "[shSymAddr maxFiles] @"]] set fdTable [wtxGopherEval "[shSymAddr fdTable] @"] set fdValid [wtxGopherEval "$fdTable ($maxFiles <+$offset(FD_ENTRY,inuse) @> +$sizeof(FD_ENTRY) )"] # # Construct a gopher string that will collect data for each valid # fd. Iterate over the vector of validity bits collected above, # and append a gopher string that will pick out the information for # that fd. # So, if you feel like tracing this down, here is a good place to start. Or, contact WRS support to see if this has been identified on your platform and resolved. I use this all the time, but I am using PowerPC or x86 targets. HTH. Asaf Sinai wrote: > > Hi, > > When I run the command 'iosFdShow' in the Shell, I get: > -> iosFdShow > fd name drv > value = 0 = 0x0 > > But, in the serial console I get all the data. > > Has anyone know why it happens ? --------------------------- Newsgroups: comp.os.vxworks Subject: a doubt on linking libraries(unix) in VxSim Date: Fri, 15 Jun 2001 09:21:20 +0530 From: "Suresh Kumar Shanbhog" Organization: Lawrence Berkeley National Laboratory Message-ID: <9gcg6i$gmr$1@overload.lbl.gov> This is a multi-part message in MIME format. - --------------InterScan_NT_MIME_Boundary Content-Type: multipart/alternative; boundary="----=_NextPart_000_0051_01C0F57C.89D59680" - ------=_NextPart_000_0051_01C0F57C.89D59680 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Hello, Iam Suresh and I work for Wipro.We are working on a project where we = need to simulate a gateway.Real gateway software runs on VxWorks.Our = project involves simulating the gateway.For this we are compiling the = source code written for VxWorks on VxSim(Unix version) so that its = entire functionality is available on Unix for testing purpose. As a development we need to plug-in some code which is written in = Unix.To achieve this we need to include some libraries compiled on Unix. We tried this by including these libs after editing the makefile (under = $WIND_BASE/target/config/hpux) with=20 MACH_EXTRA =3D /usr/lib/xyzlib.a we are using loader under /bin/ld While building VxWorks image we are getting some linker errors like = unsatisfied symbols related to socket calls(gethostbyname) and some = other Unix related calls. In the library there is one such function which calls "getpwuid()". This = function is available only in Unix but not on VxWorks.If I try to = include Unix library /usr/lib/libc.a I got lots of linker errors related = to "duplicate symbols of malloc,realloc etc" I would like to know how to resolve issues such as the above where in we = have few system calls supported under Unix but not supported under = VxWorks.Do we have to set any compiler options so that it takes default = Unix libraries since we are compiling and running the image on = Simulator. Appreciate your help in this regard. Thanks, S Suresh Kumar Sr Systerms Engineer Wipro-Bangalore. ph# 5538301-2750 ESN# 6-872-8290 - ------=_NextPart_000_0051_01C0F57C.89D59680 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable
Hello,
 
Iam Suresh and I work for Wipro.We are working on a project where = we need=20 to simulate a gateway.Real gateway software runs on VxWorks.Our project = involves=20 simulating the gateway.For this we are compiling the source code written = for=20 VxWorks on VxSim(Unix version) so that its = entire=20 functionality is available on Unix for testing purpose.
As a development we need to plug-in some code which is written in = Unix.To=20 achieve this we need to include some libraries compiled on Unix.
We tried this by including these libs after editing the makefile = (under=20 $WIND_BASE/target/config/hpux) with
       =20             =    =20 MACH_EXTRA =3D /usr/lib/xyzlib.a
we are using loader under=20 /bin/ld
While building VxWorks image we are getting some linker errors like = unsatisfied symbols related to socket calls(gethostbyname) and some = other Unix=20 related calls.
In the library there is one such function which calls "getpwuid()". This function is available only in = Unix but=20 not on VxWorks.If I try to include Unix library /usr/lib/libc.a I got lots of = linker=20 errors related to "duplicate symbols of malloc,realloc = etc"
I would like to know how to resolve issues such as the above where = in we=20 have few system calls supported under Unix but not supported under = VxWorks.Do we=20 have to set any compiler options so that it takes default Unix libraries = since=20 we are compiling and running the image on Simulator.
 
Appreciate your help in this regard.

Thanks,

S Suresh Kumar
Sr Systerms = Engineer
Wipro-Bangalore.
ph#=20 5538301-2750
ESN# 6-872-8290
- ------=_NextPart_000_0051_01C0F57C.89D59680-- - --------------InterScan_NT_MIME_Boundary-- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Booting from ATA disk ... crash Date: 15 Jun 2001 08:13:03 GMT From: tmb@Xisotek.co.uk (Trevor Barton) Organization: Isotek Electronics Ltd., Leeds, UK. Message-ID: References: <9ec53d11.0106140225.75f234a8@posting.google.com> Reply-To: tmb@Xisotek.co.uk On Thu, 14 Jun 2001 17:24:23 +0100, timallen wrote: >Trevor, > >... >> >As I understand it, VxWorks DOSFS cannot read hard drives which have been >> >formatted from DOS, unless you have DOSFS 2.0, which is a free upgrade >... >> >if you had support. :-( >> >> I'm not sure that's the case, we format all our disks with Dos 6.2 and it >works >> OK. However, we don't boot off them natively (?) we bood dos first and >then >> use vxload.com to load vxworks. That's simply because it's useful to be >> able to use DOS sometimes (repair, inital loading, editing config files in >the >> field, etc) and so we have a quick timeout config.sys meny to choose >either >> to automatically start vxworks or stay at the dos shell. > >Yes, that's what I usually do too, for the same reasons. But that does use >DOS >to access the drive. Are you then able to access the drive from VxWorks once >it has booted (without dosFS2.0)? Yes as is the boot process ... we use a 3 stage boot process which loads DOS, which uses vxload to load bootrom_uncmp which itself opens the drive to load vxWorks.st. Dunno is we have to do all of that in that way, but it's one of those things where that's what we managed to get to work in the first couple of days of having vxWorks all those many years ago, and we've just stuck with it because it works for us! But, yes, the second stage boot (bootrom_uncmp( can read the disk, and manages to find vxWorks.st in c:\vxworks, out boot line looks like #define DEFAULT_BOOT_LINE \ "ata=0,0(0,0)host:/ata0/vxworks/vxWorks.st h=192.168.2.128 ... ... e=192.168.2.129 u=target tn=WcsBots o=ene s=/startup" and all is hunky dory. After boot we access 3 primary partitions, and we have some software to mount c: d: and e: as devices /hda1/, /, and /hda3/ respectively. Application text and ini files are read from /, and logging goes to /hda3/, and that's all that goes there so that if it gets trashed it can easily be reformatted. You have to use something like partition magic to unhide D: and E: if they are primary partitions, and we used primary partitions because it's easier to decode just one partition table rather than having to code with a primary to get the extended, and then the extended's partition table. All this is with Tornado 1.0.1, no fixes, which I think is not yet dosFS2, so it realy ought to work with 2. Trev. - -- Trevor Barton Isotek Electronics Ltd, 9 Clayton Wood Bank, Leeds, LS16 6QZ, UK. Tel: +44 (113) 275 1339, Fax +44 (113) 224 9827 Remove X from tmb@Xisotek.co.uk before replying. Views expressed are my own and not necessarily those of Isotek Electronics Ltd. --------------------------- Newsgroups: comp.os.vxworks Subject: Memory Mapping Date: Fri, 15 Jun 2001 10:10:24 +0100 From: "Watson, John" Organization: Lawrence Berkeley National Laboratory Message-ID: <9gcjnb$i2t$1@overload.lbl.gov> This is a multi-part message in MIME format. - --------------InterScan_NT_MIME_Boundary Content-Type: text/plain; charset="iso-8859-1" Hi, I'm extremely new to vxWorks and having difficulty with Memory Mapping. I'm using a mbx860 SBC and BSP, and I'm trying to add an ethernet card to the board. I need to know what values to set the memory addresses to for FEIO_DEV_SPACE FEIO_MEMBASE0 memory base for CSR FEIO_MEMBASE1 memory base for Flash FEIO_IOBASE0 IO base for CSR these defines are found in the mbx800.h file. Also when do the functions to initialise the ethernet card need to be called, is it during the boot or can it be performed after? Thanks for any help, John Watson - --------------InterScan_NT_MIME_Boundary Content-Type: text/plain; name="RMRL-Disclaimer.txt" Content-Transfer-Encoding: 7bit Content-Disposition: attachment; filename="RMRL-Disclaimer.txt" The information contained in this e-mail is confidential to Roke Manor Research Ltd and must not be passed to any third party without permission. This communication is for information only and shall not create or change any contractual relationship. - --------------InterScan_NT_MIME_Boundary-- --------------------------- Newsgroups: comp.os.vxworks Subject: Attaching network interface cpm0... 0x1fe6aa8 (tBoot): Date: 15 Jun 2001 02:37:13 -0700 From: old-lian@163.net (lian) Organization: http://groups.google.com/ Message-ID: Hello all, I want to download the vxworks imag by the ethernet,but failed. Does anyone know why? (Tornado2.0/VxWorks 5.4) [VxWorks Boot]: c '.' = clear field; '-' = go to previous field; ^D = quit boot device : cpm0 processor number : 0 host name : debug file name : e:/tornado/target/config/bsc860/vxWorks inet on ethernet (e) : 169.0.0.67 169.0.0.67:ffff0000 inet on backplane (b): host inet (h) : 169.0.1.8 gateway inet (g) : user (u) : szxinwei ftp password (pw) (blank = use rsh): szxinwei flags (f) : 0x80 target name (tn) : bsc860 startup script (s) : other (o) : [VxWorks Boot]: @ boot device : cpm unit number : 0 processor number : 0 host name : debug file name : e:/tornado/target/config/bsc860/vxWorks inet on ethernet (e) : 169.0.0.67:ffff0000 host inet (h) : 169.0.1.8 user (u) : szxinwei ftp password (pw) : szxinwei flags (f) : 0x80 target name (tn) : bsc860 Attaching network interface cpm0... 0x1fe6aa8 (tBoot): memPartAlloc: block too b ig - 97280 in partition 0x1fb0334. failed: errno = 0x110001. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: PLX 9054 (and vxWorks) help Date: Fri, 15 Jun 2001 11:37:56 +0100 From: Paul Caswell Organization: Land & Sea Systems Message-ID: <3B29E584.6A528BE6@baesystems.com> References: <3B25C78C.E964F694@baesystems.com> Hello all, Thanks for the help, perhaps I should have been clearer. We have a 6U cPCI rack into which is plugged the following: Pentium Single Board Computer running NT4 Our Development Board (MPC860 board running vxWorks). This has the 9054 device on it. I expected, on system boot, for the BAR registers to be populated by the NT4 SBC host after I had initialised the 9054 registers and set its "initialisation complete bit" (LMISC[2]). To cut a long story short I was initialising the 9054 in sysHwInit(), installing interrupts in sysHwInit2(). By the time this was complete the host had given up waiting and so did not set the BAR regsiters. I moved the basic 9054 initialisation to romInit.s and now all works fine. Best Regards Paul --------------------------- Newsgroups: comp.os.vxworks Subject: Reserving Memory Date: Fri, 15 Jun 2001 11:44:33 +0100 From: Paul Caswell Organization: Land & Sea Systems Message-ID: <3B29E711.C228F8A1@baesystems.com> Hello All, I'd like to reserve a 64K window of memory at an absolute address in vxWorks. We do not have the shared memory object component. Is there an easy way to do this? Thanks Paul Caswell --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Fri Jun 15 07:17:02 2001 From: David Laight Date: Fri Jun 15 07:17:04 PDT 2001 Subject: Re: Maintianing Unresolved Symbols > There are several utilities that I want to use during development; they > aren't used, normally. I would like them to be there when I use Windshell > or telnet to the machine. Unfortunatley (in this case), ld is doing it's > job and removing the .o's that are never referenced when building the > VxWorks image. rtfm - of ld... Specify -u symbol on the ld line (before the .a containing the object you want. WRS could use this to configure vxWorks - instead of having to include calls to the relevant initialisation routines in a conditionally compiled module. The init routines could be called by scanning the linked image for names of the form (say) wxworks_init_nn and calling the routines in numeric order. Alternatively generate a assembly source file containing references to the symbols. Note all you need is: .extern _sym_name In any case you don't need a DATA reference to force the linker to pull in the symbol from a .s file. Such a file could be pulled in from a single -u to the linker. .global _sym_name .set _symname,0 is enough to make the file define a symbol - again with no code generated. David ---------------------------------------------------------------- David Laight email: dsl@tadpole.co.uk Tadpole Technology plc phone: +44 1223 428 232 Cambridge, UK fax: +44 1223 428 201 From vxwexplo-errs@csg.lbl.gov Fri Jun 15 07:18:44 2001 From: David Laight Date: Fri Jun 15 07:18:46 PDT 2001 Subject: New vxworks build for ARM My sources tell me that Tornado 2.1 for ARM (based on T2CP3) and gcc 2.96 is currently (finally) in its final stages of production. Hold your breath... David ---------------------------------------------------------------- David Laight email: dsl@tadpole.co.uk Tadpole Technology plc phone: +44 1223 428 232 Cambridge, UK fax: +44 1223 428 201 From vxwexplo-errs@csg.lbl.gov Fri Jun 15 07:46:47 2001 From: f.pertin@staubli.com Date: Fri Jun 15 07:46:49 PDT 2001 Subject: vxsys utility under WIN NT4.0 Hello VXWORKS and TORNADO gurus I am trying to run the vxsys program provided with tornado 2.0 under windows NT4.0 SP5. I get an error "the system try to access directly to the disk, this is not possible ....." Any suggestion appreciated Thanks François PERTIN STÄUBLI SCA Robotics Division Place Robert Stäubli - BP 70 F - 74210 Faverges (France) Tel.: +33 (0)4.50.65.62.59 Fax: +33 (0)4.50.65.61.30 mailto:f.pertin@staubli.com http://www.staubli.com From vxwexplo-errs@csg.lbl.gov Sat Jun 16 04:03:30 2001 From: Vxworks Exploder Date: Sat Jun 16 04:03:33 PDT 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sat Jun 16 04:03:05 PDT 2001 Subject: WDB Exception [SPR#26387] Subject: inflate.exe needed for x86-win32 Subject: Read Subject: test Subject: Re: TrueFFS INstallation Problem Subject: Re: Read Subject: Re: Read Subject: re: Maintianing Unresolved Symbols Subject: New vxworks build for ARM Subject: vxsys utility under WIN NT4.0 Subject: Tornado problem - compile on Unix, debug on Windows. Subject: Re: Proxy ARP problems: workaround needed Subject: Re: Windview wooos ? [Missing events, timing off...]... Subject: Re: New vxworks build for ARM Subject: Timing problems on VxWorks x486 Subject: Re: Timing problems on VxWorks x486 Subject: Re: DP national driver driver for VxWorks Subject: Re: Attaching network interface cpm0... 0x1fe6aa8 (tBoot): Subject: Re: vxsys utility under WIN NT4.0 Subject: routing sock or sysctl ? Subject: taskSpawn problem... Subject: Re: Maintianing Unresolved Symbols Subject: VxWorks dynamic loading Subject: Help About Posix Timer Subject: hey friendZ Subject: How common is SNiFF usage? Subject: Re: internal compiler errors Subject: GnatPro and gcc problems Subject: Re: Attaching network interface cpm0... 0x1fe6aa8 (tBoot): Subject: Re: New vxworks build for ARM Subject: Re: How common is SNiFF usage? Subject: Re: VxWorks dynamic loading ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: WDB Exception [SPR#26387] Date: 15 Jun 2001 04:21:18 -0700 From: girishvg@yahoo.com (girish) Organization: http://groups.google.com/ Message-ID: hello, goood guys. i'm developing a bsp for prop hw. currently facing a problem described below. - ------------------------------------------- [tools] tornado 2.0, vxworks 5.4 [bsp] mips r4000, custom bsp for a prop hardware. [host]sun solaris. [wdb1] agent : network debugging mode: task (system disabled, no polling at ethernet, speed 10/100 fdx) END style ethernet driver developed by self. it basically works well with applications such as http server, trivial socket applications, etc. i have checked multitasking http transfers each ranging from 100MB to 2047MB, 10 parallel downloads from netscape both from sun solaris & windows-nt hosts.in this case vxworks image also is downaloaded over the network. [wdb2] agent : serial debugging mode: task (system enabled.system mode works fine. speed 38400baud) in this configuration download & system mode debuging is okay. checked from host sun solaris. [problem] i create target server with network backend. it connects. but when i try to attach a suspended user application task the WDB gives an exception and restsrts, and thus WDB fails. this is observed in both above mentions wdb1 & wdb2 cases, essentially task mode debugging fails. [source] in the WRS distributed source file /target/config/comps/src/usrWdbTask.c the function vxTaskRegsGet() fails. vxTaskRegGet () { TCB *tcb; ... tcb = taskTcb() switch(regSetType) { case WDB_REG_SET_IU: { REG_SET dummy; taskRegsGet(context, &dummy); **** problem *** *ppRegSet = (char*)tcb->regs; break; } ... } this dummy call to taskRegsGet gives an exception & it makes WDB agent to restart. it appears to me just a vxworks configuration mismatch. under WDB agent services memory access & register access and also all services i have enabled. but still the problem remains. i checked wrs site to search this spr, looks like it is unpublished & internal spr. (fyi, i did comment out this function call, it works. the wdb moves and also tt() from host shell gives results. but this solution i am not sure whether correct or not?) i am sure someof you might have seen this problem. if you know the solution for this please reply asap. rather any help in this regards is greatly appreciated. thanks a million in advance, girish. --------------------------- Newsgroups: comp.os.vxworks Subject: inflate.exe needed for x86-win32 Date: Fri, 15 Jun 2001 14:59:38 +0200 From: "T.J.Scheffel" Organization: Internet Access Eindhoven, the Netherlands Message-ID: <9gd0sp$sms$1@news.IAEhv.nl> Hi there, Although ZLIB should contain the answer, up until now I haven't been able to properly inflate a piece of bits that went through deflate.exe in the win32 environment... I was kind of hoping inflate.exe would be where deflate.exe is in my directories. But.. Of course I am aware that instead of inflate I could just change the source to be deflated... ( That is not exactly what I need. ) It is no option to do things on my target too.. SO: If anyone has some useful hints or a ready to use inflate.exe for me: please. Thanx for your time! Teun Scheffel --------------------------- Newsgroups: comp.os.vxworks Subject: Read Date: Fri, 15 Jun 2001 15:46:36 +0100 From: "TSIIEL Lycée M Leroy Evreux" Organization: Wanadoo, l'internet avec France Telecom Message-ID: <9gd0v8$elr$1@wanadoo.fr> Hi, Our configuration : Tornado2 VxWorks5.4 PPC604 In my program, i use the read function : char mess[20]; read(fd,mess,20); fd is open and corresponds to /tyCo/1 The system connected to the serail line of our UC sends strings with 2 Organization: Lockheed Martin Corp, Valley Forge PA Message-ID: <3B2A02D3.9849BBF5@lmco.com> test --------------------------- Newsgroups: comp.os.vxworks Subject: Re: TrueFFS INstallation Problem Date: Fri, 15 Jun 2001 13:40:05 GMT From: andreyk@home.com (Andray Kaganovsky) Organization: @Home Message-ID: References: <3B28E7E2.A56DA2E@digidata.com> Doug Dotson wrote in news:3B28E7E2.A56DA2E@digidata.com: > Greetings! > > I am having alot of trouble getting TFFS installed. Followed the > instructions > and got a successful build. After rebooting, I attempt to create a > device > using "tffsDevCreate" and get the following: > > -> tffsDevCreate > > Data bus error > Exception Program Counter: 0x800a60cc > Status Register: 0x3000ff01 > Cause Register: 0x0000001c > Error address: 0xffffffff, Error ID: 0x0000 > > 80041894 vxTaskEntry +c : shell (1, 0, 0, 0) > 800bef98 shell +1c8: 800befbc (eeeeeeee, eeeeeeee, eeeeeeee, > eeeeeeee) > 800bf1a4 shell +3d4: execute (800fe4c8, 8010b304, 2048, > &yypact) > 800bf30c execute +c8 : yyparse (80c77d88, ffffffff, 1, ffffffff) > 800ef040 yyparse +888: 800ed400 (eeeeeeee, 0, eeeeeeee, 0) > 800ed5c4 yystart +89c: tffsDevCreate (140, 258, 80077aec, 0) > 800a1a58 tffsDevCreate +4c : flCall (80c77cc0, 0, 80cfe920, 96) > 800a73bc flCall +248: 800a6540 (0, 0, 8007829c, 0) > 800a6588 tffscmpWords +270: flMount (80124c60, 800a6454, 0, 0) > 800f2050 flMount +30 : flIdentifyFlash (0, 0, 80c77b80, 0) > 800a617c flIdentifyFlash+38 : 800a609c (800ec5e8, 800ec590, 800ea708, > 800e990c) > > My specific configuration is: > > 1) Board is a Galileo EV-64240 with resident Intel E28F640J3A-120 > flash. > 2) I modified mv177-sysTtfs.c file to suit. > 3) I'm using the cfiscs.c MTD. > > I put in a few debug statements and have found that the rfaRegister() > function is being > called. I also put some debug statements in the flRegisterCFISCS() and > cfiscsIdentfy() > routines and they are never called. Since the exception is occurring at > 0xffffffff it appears > to me that the function pointers to the MTD functions are never being > set up but I can't > figure out how this is supposed to be done. > > Thanks, > Doug > > > Douglas P. Dotson > Digi-Data Corporation > 8580 Dorsey Run Rd. > Jessup, MD 20794 > 301-498-0200 x248 > > > [AndrayK]. Did you #defined INCLUDE_MTD_CFISCS in sysTffs.c ? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Read Date: Fri, 15 Jun 2001 14:14:37 GMT From: "Kelly Hornsby" Organization: Excite@Home - The Leader in Broadband http://home.com/faster Message-ID: References: <9gd0v8$elr$1@wanadoo.fr> One solution: set the port to perform non-blocking reads (ioctl) and then create a task that polls the port until the newline character arrives, and then set a semaphore that your consuming process monitors so that it may get the accepted string. the maximum amount of time you need to read the string will be about 17 mSec, since at 9600 baud each character takes about 1 mSec to transfer. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Read Date: Fri, 15 Jun 2001 17:03:14 +0200 From: Emmanuel Herbreteau Organization: SEPRO Robotique (BEP) Message-ID: <3B2A23B2.DB2AED34@sepro-robotique.com> References: <9gd0v8$elr$1@wanadoo.fr> Kelly Hornsby wrote: > One solution: > set the port to perform non-blocking reads (ioctl) and then create a task > that polls the port until the newline character arrives, and then set a > semaphore Hi, A better way to perform this : attach a function to the COM interupt. ie : for the MPC860, call ppc860CallbackInstall() in the ppc860sio.c driver (for a PC target, you should look at the 16550 or 8250 SIO driver). - -- Regards Emmanuel --------------------------- Newsgroups: comp.os.vxworks Subject: re: Maintianing Unresolved Symbols Date: Fri, 15 Jun 2001 15:16:30 +0100 (BST) From: David Laight Organization: Lawrence Berkeley National Laboratory Message-ID: <200106151416.PAA29739@dsl-2.tadpole.co.uk> > There are several utilities that I want to use during development; they > aren't used, normally. I would like them to be there when I use Windshell > or telnet to the machine. Unfortunatley (in this case), ld is doing it's > job and removing the .o's that are never referenced when building the > VxWorks image. rtfm - of ld... Specify -u symbol on the ld line (before the .a containing the object you want. WRS could use this to configure vxWorks - instead of having to include calls to the relevant initialisation routines in a conditionally compiled module. The init routines could be called by scanning the linked image for names of the form (say) wxworks_init_nn and calling the routines in numeric order. Alternatively generate a assembly source file containing references to the symbols. Note all you need is: .extern _sym_name In any case you don't need a DATA reference to force the linker to pull in the symbol from a .s file. Such a file could be pulled in from a single -u to the linker. .global _sym_name .set _symname,0 is enough to make the file define a symbol - again with no code generated. David - ---------------------------------------------------------------- David Laight email: dsl@tadpole.co.uk Tadpole Technology plc phone: +44 1223 428 232 Cambridge, UK fax: +44 1223 428 201 --------------------------- Newsgroups: comp.os.vxworks Subject: New vxworks build for ARM Date: Fri, 15 Jun 2001 15:18:18 +0100 (BST) From: David Laight Organization: Lawrence Berkeley National Laboratory Message-ID: <200106151418.PAA29742@dsl-2.tadpole.co.uk> My sources tell me that Tornado 2.1 for ARM (based on T2CP3) and gcc 2.96 is currently (finally) in its final stages of production. Hold your breath... David - ---------------------------------------------------------------- David Laight email: dsl@tadpole.co.uk Tadpole Technology plc phone: +44 1223 428 232 Cambridge, UK fax: +44 1223 428 201 --------------------------- Newsgroups: comp.os.vxworks Subject: vxsys utility under WIN NT4.0 Date: Fri, 15 Jun 2001 16:46:36 +0200 From: f.pertin@staubli.com Organization: Lawrence Berkeley National Laboratory Message-ID: <9gd8u6$q7p$1@overload.lbl.gov> Hello VXWORKS and TORNADO gurus I am trying to run the vxsys program provided with tornado 2.0 under windows NT4.0 SP5. I get an error "the system try to access directly to the disk, this is not possible ....." Any suggestion appreciated Thanks François PERTIN STÄUBLI SCA Robotics Division Place Robert Stäubli - BP 70 F - 74210 Faverges (France) Tel.: +33 (0)4.50.65.62.59 Fax: +33 (0)4.50.65.61.30 mailto:f.pertin@staubli.com http://www.staubli.com --------------------------- Newsgroups: comp.os.vxworks Subject: Tornado problem - compile on Unix, debug on Windows. Date: 15 Jun 2001 08:25:07 -0700 From: matthew.becker@usa.xerox.com (Matthew Becker) Organization: http://groups.google.com/ Message-ID: <94e55b6.0106150725.35205ecc@posting.google.com> Hello all, The development group I'm in wants to use the VxWorks cross-compiler on Solaris to perform target builds, and Tornado on Windows PC's to perform debug. In the past, this strategy has worked because the IDE's they were using had some notion of associating a Unix path to a Windows driver letter (i.e. /home/jsmith -> E:\ ). Unfortunately, I have been unable to find this same functionality in Tornado. Currently the group has licences (and is using) Tornado II for Solaris and Windows. The Tornado IDE on Windows does provide a "Debugger Source Search Path" window, but it doesn't seem to be smart enough to recursively descend through a specified directory. In other words, all directories for a project must be specified individually (also, not what I'm looking for). Does anyone know of a solution to this problem for Tornado II? Is this issue resolved with Tornado 3/AE? Thanks in advance. - -Matt Becker --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Proxy ARP problems: workaround needed Date: Fri, 15 Jun 2001 15:58:02 GMT From: "Michael Yeung" Organization: Magma Communications Ltd. Message-ID: References: <3B211164.3AA58342@auriga.ru> <7H9U6.494$j6.27775@news> <3B2241AC.72CB81AA@auriga.ru> <3B228644.622D75E6@auriga.ru> Reply-To: "Michael Yeung" Hi all and Pavel, I just talked to WindRiver and they have a patch available to decouple the Proxy ARP from shared memory. So check with your WindRiver guys for the patch. It may save you some work. Michael Pavel Kovalev wrote in message news:3B228644.622D75E6@auriga.ru... > Michael et al, > > Here is some additional news regarding my investigation > of the proxy ARP configuration: > > - It is possible to configure all routes and the ARP cache > properly manually from the target shell. Some ugly tricks > are needed though, namely: > > o The main network interface should be disabled while > configuring host specific routes for the proxy interface. > The IP address of the main network interface should be > temporarily set to 127.127.127.127 while doing so (this > is described somewhere on WindSurf). > > o mRouteAdd should be used to configure routes to the > proxy clients. Though, mRouteAdd erroneously sets > the RTF_GATEWAY flag, which prevents arpAdd() from > adding entries (permanent and published) referring > to the proxy ARP clients. I had to set a breakpoint > to rtrequest() and modify the flags using the mRegs > command (cleared RTF_GATEWAY) to have arpAdd() work. > > - Configuring routes and the ARP cache is not sufficient for > the proxy ARP server to work. From the proxy ARP server > I could ping nodes located both on the main network and on > the proxy ARP network, but I could not ping nodes located > on the main network from a node located on the proxy ARP > network. > > Does anybody know how the proxy ARP server should be configured > using the routing socket? I no longer hope that this can be > done using the network functions exported by the VxWorks TCP/IP > stack and associated libraries. > > Thanks, > Pavel. > > > Pavel Kovalev wrote: > > > > Michael, > > > > > I actually ask the WindRiver support guys about this and here is what > > > they came back with. I haven't had the time to try it out yet, but here you > > > go. > > > > > > The proxy ARP is tied to shared memory in two location. > > > > > > in ..\target\src\config\usrNetwork.c or in > > > ..\target\config\comps\src\net\usrNetSmSecBoot.c if you use the Project > > > Facility to > > > > > > You can see that this is where they presume that you will be using the > > > shared memory interface, but if you call > > > proxyArpLibInit() > > > yourself, and > > > proxyNetCreate() > > > with your own set of interfaces instead of params.bad you should have no > > > problem as far as I can see. > > > > > > As well, don't select proxy ARP server component when building your BSP. > > > The linker should be capable of pulling in the right file (ArpLib.o). > > > > > > Let me know how it went. > > > > It is true that in the source code that is shipped to customers > > the proxy ARP is tied to the shared memory component in two location. > > > > Though, proxyNetCreate() checks whether the name of the network > > interface is equal to "sm" or "bp". While this problem can be > > worked around, this leads to nothing. proxyNetCreate() does > > not do very much useful by itself. The meat of proxy ARP server > > configuration seems to lie in proxyClientAdd(). Though, I haven't > > figured out how I can call this routine. Looks like shared memory > > clients register in the proxy ARP server using proxyReg(), and > > proxyClientAdd() is called from the registration packet processing > > function. I do not know how this functionality can be emulated > > in custom network drivers. I am afraid that for such drivers > > a new facility for registering proxy ARP clients should be created. > > > > I will continue some experiments and will post my results later. > > Still, I am hoping that somebody may have already passed all this > > way and found the solution... > > > > Regards, > > Pavel. > > > > > Michael > > > > > > Pavel Kovalev wrote in message > > > news:3B211164.3AA58342@auriga.ru... > > > > Hello All, > > > > > > > > We have a custom network driver for a CompactPCI backplane network > > > > and want to configure a proxy ARP server on the CompactPCI host > > > > board to make the backplane nodes visible on the Ethernet network > > > > (without additional routing). > > > > > > > > Tornado 2/VxWorks 5.4 support this configuration only for > > > > the shared memory network drivers supplied by WRS. Attempts > > > > to use proxyArpLib or configure all the necessary routing > > > > manually fail with various errors. > > > > > > > > I studied various materials available on WindSurf but haven't > > > > found a solution for this problem. There is an official SPR > > > > describing this problem (SPR# 28602), but it looks like it > > > > hasn't been processed/fixed yet. There were some discussions > > > > of this problem in the newsgroup in the past, but no good > > > > solutions have been proposed. > > > > > > > > I spent some time investigating this problem, and found that when > > > > I add a specific route to a node residing on the proxy network > > > > (route is added on the host which has to work proxy ARP server), > > > > mRouteAdd() creates the route with RTF_GATEWAY set. This prevents > > > > arpAdd() from adding a permanent published entry for the node. > > > > One solution is to develop a better version of mRouteAdd(), but > > > > I am not yet ready/willing to do so - there is no warranty that > > > > other hacks won't be needed. > > > > > > > > So, I am wondering if anybody has managed to solve this problem. > > > > > > > > Thanks, > > > > Pavel. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Windview wooos ? [Missing events, timing off...]... Date: Fri, 15 Jun 2001 16:12:46 GMT From: tangoNO_SPAM@pulp-fiction.com (Max) Message-ID: <3b2a3045.31643991@news.fgi.net> References: In my case WindView gives accurate enough timings *between* timer interrupts, used for OS ticks. However, the reported duration of the interrupts themselves is far too long and makes no sense. I tried to use for timestamping both, System and Auxiliary clocks, which in my HW are two separate timers with no visible difference. The platform is Tornado 2.0 on ARM based u-controller with 4 separate AMBA timers. Max. Cambury Technologies, Inc embedded wireless systems http://www.cambury.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: New vxworks build for ARM Date: Fri, 15 Jun 2001 16:16:50 GMT From: tangoNO_SPAM@pulp-fiction.com (Max) Message-ID: <3b2a34e0.32822977@news.fgi.net> References: <200106151418.PAA29742@dsl-2.tadpole.co.uk> On Fri, 15 Jun 2001 15:18:18 +0100 (BST), David Laight wrote: >My sources tell me that Tornado 2.1 for ARM (based on T2CP3) and gcc 2.96 is >currently (finally) in its final stages of production. Hold your breath... > > David > >---------------------------------------------------------------- >David Laight email: dsl@tadpole.co.uk >Tadpole Technology plc phone: +44 1223 428 232 >Cambridge, UK fax: +44 1223 428 201 > My sales rep quoted 2 weeks for 2.1 ARM release.... two weeks ago! (meanwhile it took them two weeks to upgrade me from 2.0 to 2.01! :) Max. Cambury Technologies, Inc embedded wireless systems http://www.cambury.com --------------------------- Newsgroups: comp.os.vxworks Subject: Timing problems on VxWorks x486 Date: 15 Jun 2001 09:49:54 -0700 From: s.starkie@ic.ac.uk (Stephen Starkie) Organization: http://groups.google.com/ Message-ID: Hi, I seem to be having problems with the timings of a piece of code I have written. It linearly searches a set of data (triangles for the closest to a point) for varying numbers of triangles. I have plotted the time per search against number of triangles, which I expected to be linear (bar caching) yet it looks like this; x x x x x x x x x x x x x which I really don't understand - Anybody any ideas???? Steve RA - Mechatronics in Medicine Imperial College, London --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Timing problems on VxWorks x486 Date: 15 Jun 2001 13:37:00 -0400 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: References: Sender: bpringlemeir@DeadDuck Stephen> Hi, I seem to be having problems with the timings of a piece Stephen> of code I have written. Does your system have cache? We need more information. regards, Bill Pringlemeir - -- Am I elected yet? Useful vxWorks URLS, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" "http://www.xs4all.nl/~borkhuis/vxworks/vxw_pt9.html" --------------------------- Newsgroups: comp.os.vxworks Subject: Re: DP national driver driver for VxWorks Date: Fri, 15 Jun 2001 14:14:50 -0400 From: "Kevin Livingston" Message-ID: References: <3B276BD2.88CF4514@teklogix.com> <3B28F6B3.40232609@teklogix.com> I have now recompilied the driver. with kernel setting to WDB_COMM_END - still no effective connecion - timout with kernel setting to WDB_COMM_NETWORK - everything appears to operate properly. I cannot yet confirm if the combination of the two changes were necessary or not. thank you for the tip Kevin "Ian Melhuish" wrote in message news:3B28F6B3.40232609@teklogix.com... > I managed to get this going by setting WDB_COMM_TYPE to WDB_COMM_NETWORK > (which is the default in configall.h) instead of WDB_COMM_END which I had > specified in config.h. > > Ian > > Mike Kordik wrote: > > > Yes. I do not have a fix for it yet. > > "Ian Melhuish" wrote in message > > news:3B276BD2.88CF4514@teklogix.com... > > > I am using the driver and also cannot get a target server connected. Did > > your > > > problem arise after installong the 2.02 patch? > > > > > > Mike Kordik wrote: > > > > > > > Is anyone using this driver (DP83815)? The target server will not > > connect to > > > > this driver after you install the Tornado 2.02 patch. Does anyone know > > why? > > > > Is there a workaround? > > > > > > > > Thanks, > > > > Mike > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Attaching network interface cpm0... 0x1fe6aa8 (tBoot): Date: Fri, 15 Jun 2001 14:20:14 -0400 From: "Kevin Livingston" Message-ID: References: I have had that problem too, although I am not sure the exact reason. I believe I was trying to use a bootrom that was not built with some of the same settings as the kernel - for the network driver. But I am not entirely certain. I have since rebuilt my bootrom and my kernel and everything appears to be operating correctly. And have since even rebuilt my kernel several times without needing to rebuild the bootrom. try rebuilding both (some recomendations I have seen also mention removing your depend. files - I really don't know if that has helped or had any effect or not) - but if you are using an old version bootrom this might be a solution. Kevin note: I am using the dp driver from national with the related hardware - possibly different from your situation. "lian" wrote in message news:ebdd146d.0106150137.2e805129@posting.google.com... > Hello all, > I want to download the vxworks imag by the ethernet,but failed. > Does anyone know why? > > > (Tornado2.0/VxWorks 5.4) > [VxWorks Boot]: c > > '.' = clear field; '-' = go to previous field; ^D = quit > > boot device : cpm0 > processor number : 0 > host name : debug > file name : e:/tornado/target/config/bsc860/vxWorks > inet on ethernet (e) : 169.0.0.67 169.0.0.67:ffff0000 > inet on backplane (b): > host inet (h) : 169.0.1.8 > gateway inet (g) : > user (u) : szxinwei > ftp password (pw) (blank = use rsh): szxinwei > flags (f) : 0x80 > target name (tn) : bsc860 > startup script (s) : > other (o) : > > [VxWorks Boot]: @ > > boot device : cpm > unit number : 0 > processor number : 0 > host name : debug > file name : e:/tornado/target/config/bsc860/vxWorks > inet on ethernet (e) : 169.0.0.67:ffff0000 > host inet (h) : 169.0.1.8 > user (u) : szxinwei > ftp password (pw) : szxinwei > flags (f) : 0x80 > target name (tn) : bsc860 > > Attaching network interface cpm0... 0x1fe6aa8 (tBoot): memPartAlloc: block too b > ig - 97280 in partition 0x1fb0334. > failed: errno = 0x110001. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: vxsys utility under WIN NT4.0 Date: Fri, 15 Jun 2001 18:44:11 GMT From: andreyk@home.com (Andray Kaganovsky) Organization: @Home Message-ID: References: <9gd8u6$q7p$1@overload.lbl.gov> f.pertin@staubli.com wrote in news:9gd8u6$q7p$1@overload.lbl.gov: > Hello VXWORKS and TORNADO gurus > > I am trying to run the vxsys program provided with tornado 2.0 under > windows NT4.0 SP5. > I get an error "the system try to access directly to the disk, this is not > possible ....." > > Any suggestion appreciated > > > Thanks > > > > François PERTIN > > > STÄUBLI SCA > Robotics Division > Place Robert Stäubli - BP 70 > F - 74210 Faverges (France) > Tel.: +33 (0)4.50.65.62.59 > Fax: +33 (0)4.50.65.61.30 > mailto:f.pertin@staubli.com > http://www.staubli.com > > [AndrayK]. I think VxSYS is DOS utility, and doesn't work under Windows (at least NT4/5). You can't run it from DOS box, you need to actually boot DOS. Andray --------------------------- Newsgroups: comp.os.vxworks Subject: routing sock or sysctl ? Date: Fri, 15 Jun 2001 12:16:34 -0700 From: Waseem Siddiqi Organization: Cisco Systems Inc. Message-ID: <3B2A5F12.21B7B07F@yahoo.com> Does anyone know if vxworks supports routing sockets or sysctl ? I am trying to programmatically get information on a devices' IP addr/netmask/GW thanks waseem --------------------------- Newsgroups: comp.os.vxworks Subject: taskSpawn problem... Date: Sat, 16 Jun 2001 06:10:23 +0900 From: "±èº´È£" Organization: Korea Telecom Message-ID: <9gdtdb$ebc$1@imsinews.kornet.net> 'taskSpawn' function doesn't work on my VxWorks. I tried it on the shell, like below. and have the exception error messages. what should I do to make 'taskSpawn' works ? - ---------------------------------------------- - -> taskSpawn() Exception number 4906: Task: 0x189720 () instruction translation miss program counter: 0x04 machine status : 0x9032 trcStack aborted: error in top frame (sp) ffffffff end +ffed846f: () value = 159704 = 0x185e88 - -> - ------------------------------------------------- and I also write the code to try 'taskSpawn' and 'etherOutputHookAdd'. and download it to vxworks and tried ... It also doesn't work.... I think it's also because of that 'taskSpawn' problem. I'll be really pleased if i can get advice to solve that problems... can anyone see if there's any problem in the code ? - --------------------------------------------------- #include #include void hook_test() { BOOL OutputHandle(struct ifnet *pIf, char *buffer, int length) { /* to see what is hooked*/ printf("hooked_packet = %c\n", *buffer); return (FALSE); } STATUS OutputHook() { if (etherOutputHookAdd(OutputHandle) == OK) { printf("etherOutputHookAdd successful\n"); return(OK); } else { printf("etherOutputHook failed\n"); return(ERROR); } } taskSpawn ("etherOutputHook", 20, 0, 2000, OutputHook, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); } - ---------------------------------------------------------------------------- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Maintianing Unresolved Symbols Date: 15 Jun 2001 14:31:33 -0700 From: hariharan@foobox.com (Thiagarajan Hariharan) Organization: http://groups.google.com/ Message-ID: <7f10f842.0106151331.798b3a60@posting.google.com> References: <3b28eab6$0$94313$39cecf19@news2.twtelecom.net> My experience with specifying all the .o's on the command line has not been good. While everthing links just fine, when the bootrom loads vxworks, instead of the usual "Starting at 0x108000...", you see some weird address instead of 0x108000. Then you drop right back into the bootrom prompt. I then switched to putting all the .o's into a .a - it worked fine. Of your options, (c) would be the simplest. You should not have to create a no-op function in each .o. In one of the .o's that you know is already getting linked in, create a dummy function that calls functions (not no-op) from each of the .o's that you want to force to be linked in. Hari --------------------------- Newsgroups: comp.os.vxworks Subject: VxWorks dynamic loading Date: Fri, 15 Jun 2001 16:12:40 -0700 From: "Alex Tselovalnikov" Organization: SBC Internet Services Message-ID: <5KwW6.709$ll3.117524@news.pacbell.net> Hi, Does anybody know if a notion of shared libraries and/or dynamic loading is available on VxWorks. I mean something similar to dlopen/dlsym (most of the unix), or LoadLibrary/GetProcedureAddress (Windows)? Thanks a lot in advance Alex --------------------------- Newsgroups: comp.os.vxworks Subject: Help About Posix Timer Date: Fri, 15 Jun 2001 16:42:29 +0800 From: "Kingwel Xie" Organization: The news server of Changzhou TeleCom Message-ID: <9gchns$2jna$1@news.cz.js.cn> Hi, all Did anyone know about the maximal count of posix timer can be started simultaneously? I made a test, and got a GP fault in the end, when the counter reached less than 2000. It seems that the signal module of kernel get in trouble eventually( Overrun ?). Thanks XJW The below shows the test code: /*------------------------------------------------------------------------*/ #include static void sss(timer_t timerid, int arg) { logMsg("***Count=%d\n", arg, 0, 0, 0, 0, 0); } void ddd() { timer_t timerId; int counter = 0; while(1) { counter++; { struct itimerspec timeToSet; struct timespec timeValue; struct timespec timeInterval; timeValue.tv_sec = 5; timeValue.tv_nsec = 0; timeInterval.tv_sec = 5; timeInterval.tv_nsec = 0; timeToSet.it_value = timeValue; timeToSet.it_interval = timeInterval; if ((timer_create (CLOCK_REALTIME, NULL, &timerId)) == ERROR) { printf("posixTimerStart: Error in allocating a timer\n"); return; } if (timer_settime (timerId, 0, &timeToSet, NULL) == ERROR) { printf("posixTimerStart: Error in setting time\n"); return; } timer_connect(timerId, sss, counter); } printf("---Count=%d\n", counter); taskDelay(2); } } --------------------------- Newsgroups: comp.os.vxworks Subject: hey friendZ Date: 15 Jun 2001 17:03:51 -0700 From: DrSec@hotmail.com (DrSec) Organization: http://groups.google.com/ Message-ID: I realy need VxWorks v5.4 & Tornado II , hope someone upload me a copy of them ... you'r welcome to get any CAD/CAM/CAE & specialy EDA tools that you need from me ,email me regardZ DrSec DrSec@hotmail.com --------------------------- Newsgroups: comp.os.vxworks Subject: How common is SNiFF usage? Date: Fri, 15 Jun 2001 18:57:27 -0600 From: "Joe Woodbury" Message-ID: My local rep is pushing SNiFF+ at us for our next project. I used SNiFF+ 2 and didn't like it much, but gave the SNiFF+ 4.x demo a whirl to see if it could change my mind. I still didn't like it. However, with future employment possibilities in mind, I'm wondering how many developers/companies use SNiFF+ (or at least insist it be used)? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: internal compiler errors Date: Fri, 15 Jun 2001 19:41:02 -0700 From: DrDiags Organization: laserlink.net Message-ID: <3B2AC73E.B29FCBD7@covad.net> References: <3B207CA7.D938FC01@mediaone.net> <3B28C4E7.6620AF6B@mediaone.net> <3B29A02B.9B6947BD@covad.net> Reply-To: drdiags@covad.net Point of clarification: ======================= The following statement is not correct. The text following my original post is in regards to GNATPro and problems found when using TornadoAE, so sorry for the misinformation. DrDiags wrote: > > Jagan, > > Do you have the GNAT 95 Ada compiler installed on these > machines? If so, somewhere in my reading, there was a mention > of problems similar to yours. If I can find the e-mail I saw > that spoke to this, I will post it here. But if you don't have > this configuration, never mind.......... ====================== First problem text ========================== Subject: GnatPro and gcc problems Folks, In an effort to save others the trouble that I went thru... I recently installed GnatPro on top of a T2-PPC tree on my PC. After that, I was having problems building in other Tornado trees. I'd get weird errors about "can't find cpp". After pulling my hair out for a day, I eventually figured out that GnatPro had added a key into my registry. Removing this key allowed me to build again in my other Tornado trees. However, w/o the key, Gnat will not work. So, leave the key in for gnat, remove it otherwise. The offending key is: [HKEY_LOCAL_MACHINE\SOFTWARE\Free Software Foundation] "GCC"="D:\\GnatPro" "MINGW32"="@GCC" "GNAT"="@GCC" "BINUTILS"="@GCC" ======================= Second problem text =========================== I can now say that there is definitely an incompatibility between Tornado for AE and GnatAda 3.14a installations on Windows hosts. Until this issue is resolved, just make a choice - you're either going to build Ada with Gnat for T2, or you are going to build with AE. When you install GnatAda 3.14a, a registry entry is created called "Free Software Foundation". It contains an entry for GCC that is "d:\GNATPro" (or whatever your gnat installation directory might be), and entries like BIN_UTILS @GCC. This messes up AE builds, causing the "can't find cpp" error. When you delete this registry setting, you can build AE again (but not Ada with gnat). .................., I did find some entries on this "cpp installation problem" that were resolved by setting GCC_EXEC_PREFIX (making sure there was a "/" at the end). I don't know if that would still work now or not. I would expect this to be a problem if other software had similar registry settings setting GCC to something else. The ACT folks are copied on this so that they know as well. To state the obvious, something will have to be done on this for their GnatAda95 with AE to work properly on Windows hosts. =================== ACT Support response ======================== >Thanks for reporting this. The 3.14a1 release of GNATPRO will no longer be >using Free Software Foundation registry subtree for its purposes, which >will hopefully address the problem you have reported. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Attaching network interface cpm0... 0x1fe6aa8 (tBoot): Date: Fri, 15 Jun 2001 19:59:13 -0700 From: DrDiags Organization: laserlink.net Message-ID: <3B2ACB81.9D3D6076@covad.net> References: Reply-To: drdiags@covad.net Lian, I see that you are using DHCP (or at least the flag is set), so I am guessing you have included the DHCPC in your bootrom. I just had the same problem with a SBC8260 w/64MB of RAM. The BSP divided the 64MB in half, as it should, so it was reserving 32MB as user defined. I found out that on an x86 platform this same problem of "block too big" was resolved by calling memAddToPool during sysHwInit2(), and adding back the memory which had been user reserved, which I didn't understand. I figured, heck I got 32MB of RAM, why should I have to add the other 32MB to my memory pool, but I did it and my error went away. I need someone smarter than I am (which is just about everybody, but that is another story), to explain why this is necessary. Since I am notorious for not providing exact detail, you should see if some else can give you the exact steps required to do this, because I don't have the modified code in front of me and I lost the e-mail that referenced the solution used for the x86 platform. lian wrote: > > Hello all, > I want to download the vxworks imag by the ethernet,but failed. > Does anyone know why? > > (Tornado2.0/VxWorks 5.4) > [VxWorks Boot]: c > > '.' = clear field; '-' = go to previous field; ^D = quit > > boot device : cpm0 > processor number : 0 > host name : debug > file name : e:/tornado/target/config/bsc860/vxWorks > inet on ethernet (e) : 169.0.0.67 169.0.0.67:ffff0000 > inet on backplane (b): > host inet (h) : 169.0.1.8 > gateway inet (g) : > user (u) : szxinwei > ftp password (pw) (blank = use rsh): szxinwei > flags (f) : 0x80 > target name (tn) : bsc860 > startup script (s) : > other (o) : > > [VxWorks Boot]: @ > > boot device : cpm > unit number : 0 > processor number : 0 > host name : debug > file name : e:/tornado/target/config/bsc860/vxWorks > inet on ethernet (e) : 169.0.0.67:ffff0000 > host inet (h) : 169.0.1.8 > user (u) : szxinwei > ftp password (pw) : szxinwei > flags (f) : 0x80 > target name (tn) : bsc860 > > Attaching network interface cpm0... 0x1fe6aa8 (tBoot): memPartAlloc: block too b > ig - 97280 in partition 0x1fb0334. > failed: errno = 0x110001. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: New vxworks build for ARM Date: Fri, 15 Jun 2001 20:08:57 -0700 From: DrDiags Organization: laserlink.net Message-ID: <3B2ACDC9.B9CA0CC4@covad.net> References: <200106151418.PAA29742@dsl-2.tadpole.co.uk> <3b2a34e0.32822977@news.fgi.net> Reply-To: drdiags@covad.net T2.1 ARM and MIPS are BETA testing, which ended today. So, maybe that pesky Sales guy may not be pulling your leg (:^), but alas, PPC and x86 are not there yet (:^( Max wrote: > > On Fri, 15 Jun 2001 15:18:18 +0100 (BST), David Laight > wrote: > > >My sources tell me that Tornado 2.1 for ARM (based on T2CP3) and gcc 2.96 is > >currently (finally) in its final stages of production. Hold your breath... > > > > David > > > >---------------------------------------------------------------- > >David Laight email: dsl@tadpole.co.uk > >Tadpole Technology plc phone: +44 1223 428 232 > >Cambridge, UK fax: +44 1223 428 201 > > > > My sales rep quoted 2 weeks for 2.1 ARM release.... two weeks ago! > (meanwhile it took them two weeks to upgrade me from 2.0 to 2.01! :) > > Max. > > Cambury Technologies, Inc > embedded wireless systems > http://www.cambury.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How common is SNiFF usage? Date: Fri, 15 Jun 2001 20:16:46 -0700 From: DrDiags Organization: laserlink.net Message-ID: <3B2ACF9E.A6709005@covad.net> References: Reply-To: drdiags@covad.net Joe, I am pretty sure that Martin R. will be more than happy to give you the details. Just wait ........ Joe Woodbury wrote: > > My local rep is pushing SNiFF+ at us for our next project. I used SNiFF+ 2 > and didn't like it much, but gave the SNiFF+ 4.x demo a whirl to see if it > could change my mind. I still didn't like it. However, with future > employment possibilities in mind, I'm wondering how many > developers/companies use SNiFF+ (or at least insist it be used)? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks dynamic loading Date: Sat, 16 Jun 2001 05:53:40 GMT From: "Merlin" Organization: [Infostrada] Message-ID: References: <5KwW6.709$ll3.117524@news.pacbell.net> You can dynamically link a file.o to your running application. vxworks I did it just a couple of month ago. So it is possible but you have to find the funcion on the help ( LoadModule or something like that .. ) Merlin "Alex Tselovalnikov" ha scritto nel messaggio news:5KwW6.709$ll3.117524@news.pacbell.net... > Hi, > > Does anybody know if a notion of shared libraries and/or dynamic loading is > available on VxWorks. I mean something similar to dlopen/dlsym (most of the > unix), or LoadLibrary/GetProcedureAddress (Windows)? > > Thanks a lot in advance > > Alex > > > --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Sun Jun 17 04:03:31 2001 From: Vxworks Exploder Date: Sun Jun 17 04:03:33 PDT 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sun Jun 17 04:03:05 PDT 2001 Subject: Re: hey friendZ Subject: hardware implementation Subject: Re: hardware implementation Subject: VxWorks web servers & server pages technologies Subject: Re: Windview wooos ? [Missing events, timing off...]... Subject: Re: TrueFFS INstallation Problem Subject: printer driver for MVME2604 Subject: how can i use 'taskSpawn' in vxworks 5.4 ? ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: hey friendZ Date: Sat, 16 Jun 2001 09:37:54 -0700 From: "Michael R. Kesti" Organization: MK Associates Message-ID: <3B2B8B62.B1C131F4@gv.net> References: Reply-To: mkesti@gv.net DrSec wrote: >I realy need VxWorks v5.4 & Tornado II , hope someone upload me a copy >of them ... My respect for vendors such as Wind River prevents me from doing such things. > you'r welcome to get any CAD/CAM/CAE & specialy EDA tools >that you need from me ,email me And my self-respect prevents me from accepting such offers. What do you respect, DrSec? - -- ======================================================================== 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: hardware implementation Date: 16 Jun 2001 10:50:38 -0700 From: john_valavi@mailcity.com (JOHN VALAVI) Organization: http://groups.google.com/ Message-ID: <1c5aef0e.0106160950.2e99fbd8@posting.google.com> i want to know about the harware control of devices using the vxworks --------------------------- Newsgroups: comp.os.vxworks Subject: Re: hardware implementation Date: Sat, 16 Jun 2001 12:14:38 -0700 From: "Michael R. Kesti" Organization: MK Associates Message-ID: <3B2BB01E.2C5CD8FC@gv.net> References: <1c5aef0e.0106160950.2e99fbd8@posting.google.com> Reply-To: mkesti@gv.net JOHN VALAVI wrote: >i want to know about the harware control of devices using the vxworks There are as many ways to do this as there are programmers and devices to be controlled! They all come down to one of two fundamental techniques, however. The first is to write a device driver for the hardware and incorporate it in the OS. This provides the ability to use the OS's open(), close(), read(), write(), and ioctl() calls from the application level, and offers a consistent and convenient programmer's interface that serves to isolate the hardware issues from the application issues. That interface may also, in some cases, be too slow or limiting to accomplish a system's mission. The second is to access the hardware devices directly from the application level. This is often faster and more flexible, at the cost of program structure and complexity. - -- ======================================================================== 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: VxWorks web servers & server pages technologies Date: Sat, 16 Jun 2001 16:36:29 -0400 From: "Lin, Ping [TOR:9T00:EXCH]" Organization: Nortel Networks Message-ID: <9ggg24$18s$1@bcarh8ab.ca.nortel.com> Hi, I'm looking at the possibility of creating dynamic web server pages (similar to JSP/ASP) on VxWorks, and was wondering if the users of this newsgroup can give me some pointers on: 1) Is Apache or any other web server available for VxWorks, besides the Wind web server? 2) Is there an implementation of JSP, Java servlets or similar server pages technology available for VxWorks? Thanks in advance, Ping Lin --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Windview wooos ? [Missing events, timing off...]... Date: Sat, 16 Jun 2001 23:16:38 GMT From: Todd Sandor Organization: @home Message-ID: <3B2BEBE4.A2BD7C98@home.com> References: <3b2a3045.31643991@news.fgi.net> I too get accurate OS/System clock ticks in Windview, but the time between other "events" in Windview is inaccurate [the real/wall time between events *is* correct]. This doesn't make sense to me, since the Aux clock would be measuring the System clock tick interrupt and recording it accurately, but it doesn't seem to record the other events with the correct time? As you say, this makes no sense.... Is anyone using Windview for performing accurate measuring? If so, what is the environment/trick? Other then starting to turn Led on/off and measuring this with a scope, what other methods/tools are people using to perform low level timing measurements? Cheers... Max wrote: > In my case WindView gives accurate enough timings *between* > timer interrupts, used for OS ticks. However, the reported duration > of the interrupts themselves is far too long and makes no sense. > I tried to use for timestamping both, System and Auxiliary clocks, > which in my HW are two separate timers with no visible difference. > > The platform is Tornado 2.0 on ARM based u-controller with 4 > separate AMBA timers. > > Max. > > Cambury Technologies, Inc > embedded wireless systems > http://www.cambury.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: TrueFFS INstallation Problem Date: Fri, 15 Jun 2001 12:48:20 -0400 From: Doug Dotson Organization: Verio Message-ID: <3B2A3C54.A6640BA5@digidata.com> References: <3B28E7E2.A56DA2E@digidata.com> Yes. And when I load it I take a look at noOfMTDs and it is 1 indicating that one entry is in the MTD table. Doug Andray Kaganovsky wrote: > Doug Dotson wrote in > news:3B28E7E2.A56DA2E@digidata.com: > > > Greetings! > > > > I am having alot of trouble getting TFFS installed. Followed the > > instructions > > and got a successful build. After rebooting, I attempt to create a > > device > > using "tffsDevCreate" and get the following: > > > > -> tffsDevCreate > > > > Data bus error > > Exception Program Counter: 0x800a60cc > > Status Register: 0x3000ff01 > > Cause Register: 0x0000001c > > Error address: 0xffffffff, Error ID: 0x0000 > > > > 80041894 vxTaskEntry +c : shell (1, 0, 0, 0) > > 800bef98 shell +1c8: 800befbc (eeeeeeee, eeeeeeee, eeeeeeee, > > eeeeeeee) > > 800bf1a4 shell +3d4: execute (800fe4c8, 8010b304, 2048, > > &yypact) > > 800bf30c execute +c8 : yyparse (80c77d88, ffffffff, 1, ffffffff) > > 800ef040 yyparse +888: 800ed400 (eeeeeeee, 0, eeeeeeee, 0) > > 800ed5c4 yystart +89c: tffsDevCreate (140, 258, 80077aec, 0) > > 800a1a58 tffsDevCreate +4c : flCall (80c77cc0, 0, 80cfe920, 96) > > 800a73bc flCall +248: 800a6540 (0, 0, 8007829c, 0) > > 800a6588 tffscmpWords +270: flMount (80124c60, 800a6454, 0, 0) > > 800f2050 flMount +30 : flIdentifyFlash (0, 0, 80c77b80, 0) > > 800a617c flIdentifyFlash+38 : 800a609c (800ec5e8, 800ec590, 800ea708, > > 800e990c) > > > > My specific configuration is: > > > > 1) Board is a Galileo EV-64240 with resident Intel E28F640J3A-120 > > flash. > > 2) I modified mv177-sysTtfs.c file to suit. > > 3) I'm using the cfiscs.c MTD. > > > > I put in a few debug statements and have found that the rfaRegister() > > function is being > > called. I also put some debug statements in the flRegisterCFISCS() and > > cfiscsIdentfy() > > routines and they are never called. Since the exception is occurring at > > 0xffffffff it appears > > to me that the function pointers to the MTD functions are never being > > set up but I can't > > figure out how this is supposed to be done. > > > > Thanks, > > Doug > > > > > > Douglas P. Dotson > > Digi-Data Corporation > > 8580 Dorsey Run Rd. > > Jessup, MD 20794 > > 301-498-0200 x248 > > > > > > > > [AndrayK]. Did you #defined INCLUDE_MTD_CFISCS in sysTffs.c ? --------------------------- Newsgroups: comp.os.vxworks Subject: printer driver for MVME2604 Date: 16 Jun 2001 23:17:47 -0700 From: chrisanc1967@yahoo.com (Christopher Sanchez) Organization: http://groups.google.com/ Message-ID: I need to print from a hp laserjet printer using the parallel port on the trasition module. Can anyone that is doing print jobs from vxWorks explain to me how its done? Thanks Christopher --------------------------- Newsgroups: comp.os.vxworks Subject: how can i use 'taskSpawn' in vxworks 5.4 ? Date: Sun, 17 Jun 2001 17:08:43 +0900 From: "SungJin Lee" Organization: Korea Telecom Message-ID: <9ghoj1$aqm$1@imsinews.kornet.net> hi vxworkers, When i try 'taskSpawn()' on the shell, it doesn't work and shows the error message following. How can i use taskSpawn function in the vxworks 5.4 ? (with tornado 2) - ----------------------------------------------------------------- Exception number 4906: Task: 0x189720 () instruction translation miss program counter: 0x04 machine status : 0x9032 trcStack aborted: error in top frame (sp) ffffffff end +ffed846f: () value = 159704 = 0x185e88 --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Sun Jun 17 05:23:49 2001 From: "David Laight" Date: Sun Jun 17 05:23:51 PDT 2001 Subject: Re: Timing problems on VxWorks x486 > I seem to be having problems with the timings of a piece of code I > have written. > It linearly searches a set of data (triangles for the closest to a > point) for varying numbers of triangles. I have plotted the time per > search against number of triangles, which I expected to be linear > (bar caching) yet it looks like this; > x > x > x x > x x > x x > x x > x x > x > which I really don't understand - Anybody any ideas???? Seems to be precisely what I would expect! 1st line - data pulled from level 1 cache 2nd line - data from level 2 cache 3rd line - data from main memory Nice to see a test showing it so well :-) This would be particularly visible if you are running through the data many times in order to get a 'accurate' benchmark. David ... vxworks... From vxwexplo-errs@csg.lbl.gov Sun Jun 17 05:36:37 2001 From: "David Laight" Date: Sun Jun 17 05:36:39 PDT 2001 Subject: Re: New vxworks build for ARM > T2.1 ARM and MIPS are BETA testing, which ended today. So, maybe > that pesky Sales guy may not be pulling your leg (:^), but > alas, PPC and x86 are not there yet (:^( I beleive that T2.1 for ARM is basically T2CP3 which has been available for PPC and x86 for a while. Even the 'mandatory patches' haven't been available for vxworks for ARM. Fortunately I don't have to put up with this any longer! We have had a 'reorganisation', further development on the project I have been working on is cancelled, and I am now a man of leasure... (for a week or two at least) David From vxwexplo-errs@csg.lbl.gov Sun Jun 17 07:34:04 2001 From: "A.RadhaKrishna" Date: Sun Jun 17 07:34:07 PDT 2001 Subject: Fw: Trigraphs encountered. This is a multi-part message in MIME format. ------=_NextPart_000_003C_01C0F768.633BD960 Content-Type: text/plain; charset="Windows-1252" Content-Transfer-Encoding: quoted-printable ----- Original Message -----=20 From: A.RadhaKrishna=20 To: vxworks_users@csg.lbl.gov ; vxwexplo-errs@csg.lbl.gov=20 Sent: Sunday, June 17, 2001 6:57 PM Subject: Trigraphs encountered. Hi VxWorkers, I am using Tornado2 for Pentium Target. I modified my harddisk driver file (ataDrv.c file) and built it using = "make CPU=3DPENTIUM" at dos prompt. It is working fine but I am getting a warning " 1 Trigraphs occurred". What is the meaning of this warning? Can you pl tell me Thanks & Regards, A.Radha Krishna ------=_NextPart_000_003C_01C0F768.633BD960 Content-Type: text/html; charset="Windows-1252" Content-Transfer-Encoding: quoted-printable
 
----- Original Message -----=20
From: A.RadhaKrishna
To: vxworks_users@csg.lbl.gov = ; vxwexplo-errs@csg.lbl.gov =
Sent: Sunday, June 17, 2001 6:57 PM
Subject: Trigraphs encountered.

Hi VxWorkers,
 
 
 
I am using Tornado2 for Pentium Target.
I modified my harddisk driver file (ataDrv.c file) = and built=20 it using "make CPU=3DPENTIUM" at dos prompt.
It is working fine but I am getting a warning " 1 = Trigraphs=20 occurred".
 
What is the meaning of this = warning?
 
 
 
Can you pl tell me
 
 
Thanks & Regards,
A.Radha Krishna
------=_NextPart_000_003C_01C0F768.633BD960-- From vxwexplo-errs@csg.lbl.gov Sun Jun 17 10:07:20 2001 From: Michael.Ben-Ari@ecitele.com Date: Sun Jun 17 10:07:22 PDT 2001 Subject: Re: Fw: Trigraphs encountered. >I am using Tornado2 for Pentium Target. >I modified my harddisk driver file (ataDrv.c file) and built it using = >"make CPU=3DPENTIUM" at dos prompt. >It is working fine but I am getting a warning " 1 Trigraphs occurred". > >What is the meaning of this warning? This has NOTHING to do with Tornado or VxWorks (per se) You most probably have in your code (EVEN in COMMENTS) a trigraph sequence -- see K&R (Kernighan & Ritchie; A12.1 - This is new to the ANSI standard -- But as the GNU manual states "You don't want to know about this brain-damage..." use the -ansi switch or -trigraphs switch upon compilation... Or better yet, get rid of any comments that comtain a trigraph sequence '??X' (see K&R for definitions of X). -Mike ----------------------------------------------- Mike Ben-Ari ECI Telecom - NGTS Ltd. Cellular Division - Celtro Mail: mbenari@ecitele.com Tel: +972-3-926-6794 Fax: +972-3-926-6460 ----------------------------------------------- From vxwexplo-errs@csg.lbl.gov Sun Jun 17 10:18:57 2001 From: Susanna Jacobson Date: Sun Jun 17 10:19:00 PDT 2001 Subject: Re: Fw: Trigraphs encountered. I got a trigraph warning once when I had put "???" in a comment, (in a comment!!) something like /* Why does this happen??? */ When I reduced the number of question marks to 1, the warning went away. Hoping this helps, Susanna > ----- Original Message -----=20 > From: A.RadhaKrishna=20 > To: vxworks_users@csg.lbl.gov ; vxwexplo-errs@csg.lbl.gov=20 > Sent: Sunday, June 17, 2001 6:57 PM > Subject: Trigraphs encountered. > > Hi VxWorkers, > > I am using Tornado2 for Pentium Target. > I modified my harddisk driver file (ataDrv.c file) and built it using = > "make CPU=3DPENTIUM" at dos prompt. > It is working fine but I am getting a warning " 1 Trigraphs occurred". > > What is the meaning of this warning? > > Can you pl tell me > > Thanks & Regards, > A.Radha Krishna -- ===================================================================== Susanna Jacobson MS 10-110 SRJacobson@LBL.Gov Lawrence Berkeley National Laboratory VOICE: (510) 486-7801 1 Cyclotron Road FAX: (510) 486-7374 Berkeley, CA 94720 ===================================================================== From vxwexplo-errs@csg.lbl.gov Mon Jun 18 04:03:28 2001 From: Vxworks Exploder Date: Mon Jun 18 04:03:30 PDT 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Mon Jun 18 04:03:10 PDT 2001 Subject: Re: how can i use 'taskSpawn' in vxworks 5.4 ? Subject: re: Timing problems on VxWorks x486 Subject: re: New vxworks build for ARM Subject: Q: Problem with sm network on mcp750 Subject: Re: how can i use 'taskSpawn' in vxworks 5.4 ? Subject: Re: how can i use 'taskSpawn' in vxworks 5.4 ? Subject: Fw: Trigraphs encountered. Subject: Trigraphs encountered. Subject: Re: hey friendZ Subject: Re: How common is SNiFF usage? Subject: Re: Fw: Trigraphs encountered. Subject: Re: Trigraphs encountered. Subject: Re: hey friendZ Subject: re: Fw: Trigraphs encountered. Subject: re: Fw: Trigraphs encountered. Subject: reaction to posix compat and sockets Subject: Re: How common is SNiFF usage? Subject: Build Error !!! Subject: Re: how can i use 'taskSpawn' in vxworks 5.4 ? Subject: Re: vxsys utility under WIN NT4.0 Subject: Re: Booting from ATA disk ... crash Subject: PowerPC VME read problem ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: how can i use 'taskSpawn' in vxworks 5.4 ? Date: 17 Jun 2001 05:50:15 -0700 From: nick.gardner@europe.simoco.com (nick) Organization: http://groups.google.com/ Message-ID: References: <9ghoj1$aqm$1@imsinews.kornet.net> Hmm, this is a dead quick answer, with not much thought at all, but I think you should use ts() to spawn tasks from the shell. Anyway, give it a try and see what happens. Regards, Nick > hi vxworkers, > > > When i try 'taskSpawn()' on the shell, it doesn't work > and shows the error message following. > > How can i use taskSpawn function in the vxworks 5.4 ? (with tornado 2) > > ----------------------------------------------------------------- > Exception number 4906: Task: 0x189720 () > > instruction translation miss > program counter: 0x04 > machine status : 0x9032 > trcStack aborted: error in top frame (sp) > ffffffff end +ffed846f: () > value = 159704 = 0x185e88 --------------------------- Newsgroups: comp.os.vxworks Subject: re: Timing problems on VxWorks x486 Date: Sun, 17 Jun 2001 13:28:00 +0100 From: "David Laight" Organization: Lawrence Berkeley National Laboratory Message-ID: <9giahq$3qe$1@overload.lbl.gov> > I seem to be having problems with the timings of a piece of code I > have written. > It linearly searches a set of data (triangles for the closest to a > point) for varying numbers of triangles. I have plotted the time per > search against number of triangles, which I expected to be linear > (bar caching) yet it looks like this; > x > x > x x > x x > x x > x x > x x > x > which I really don't understand - Anybody any ideas???? Seems to be precisely what I would expect! 1st line - data pulled from level 1 cache 2nd line - data from level 2 cache 3rd line - data from main memory Nice to see a test showing it so well :-) This would be particularly visible if you are running through the data many times in order to get a 'accurate' benchmark. David ... vxworks... --------------------------- Newsgroups: comp.os.vxworks Subject: re: New vxworks build for ARM Date: Sun, 17 Jun 2001 13:40:48 +0100 From: "David Laight" Organization: Lawrence Berkeley National Laboratory Message-ID: <9giakn$3tq$1@overload.lbl.gov> > T2.1 ARM and MIPS are BETA testing, which ended today. So, maybe > that pesky Sales guy may not be pulling your leg (:^), but > alas, PPC and x86 are not there yet (:^( I beleive that T2.1 for ARM is basically T2CP3 which has been available for PPC and x86 for a while. Even the 'mandatory patches' haven't been available for vxworks for ARM. Fortunately I don't have to put up with this any longer! We have had a 'reorganisation', further development on the project I have been working on is cancelled, and I am now a man of leasure... (for a week or two at least) David --------------------------- Newsgroups: comp.os.vxworks Subject: Q: Problem with sm network on mcp750 Date: 17 Jun 2001 08:02:26 -0700 From: dafna@septier.com (Dafna) Organization: http://groups.google.com/ Message-ID: Hi All, We are experiencing some problems using the SM network with MCP750 CPU as a master, and MCP860 CPUs (Artesyn Bajaspan Board) as "slave" processors. Here's the configuration we are using: MCP750 master CPU (#0) initializes the shared-memory area and sets the ANCHOR address to 0x4100. (A constant value set in the 750 config.h). At the same time, several MPC860 processors poll the SM area on the 750 for their boot files. When the sm is initialized and boot file is found, the mpc860 CPUs retreive it by FTP, and boot over the SM network, which is activated over a cPci backplane. During boot time, the 860 CPUs operate in poll mode, and once booted they operate in Interrupt mode. The 750 CPU operates in interrupt mode. And the problem... As we increase the number of MPC860 CPUs booting over the SM network (8 CPUs and more), we see some of the processors, sometimes all, hang when searching for the boot file, or during booting itself. Once this problem occurs, the CPUs that "got into trouble" never recover. Our research has shown that when things go wrong, the head of the "free-packets-list" at the shared memory is overridden. From what we've seen, this list is used by the 750 to receive packets from the 860. We can also see that the shared-memory area is initialized (ready value at the ANCHOR is 0x87654321), and the heartbeat advances, showing the system is alive. Has anyone, by any chance, experienced a similar problem, or have used the SM network in a similar configuration? We have contacted the Wind River support, but with no luck so far. Thanks in advance, Dafna --------------------------- Newsgroups: comp.os.vxworks Subject: Re: how can i use 'taskSpawn' in vxworks 5.4 ? Date: 17 Jun 2001 08:06:25 -0700 From: qj1020@yahoo.com (Jin Qian) Organization: http://groups.google.com/ Message-ID: References: <9ghoj1$aqm$1@imsinews.kornet.net> Hi SungJin, To try "taskSpawn", you can try taskSpawn "sungjin", 4, 0, 40000, printf, "hello world\n", 0 on the tornado shell or better yet on target shell. You will see string "hello world" as the output. After this the task terminates. Note: "sungjin" is the task name, you can change it to any string 4 is the priority of the task 0 is the option 40000 is the stack size for this task (all local variables of functions called by the task will use this space) printf is a function used as the entry point to this task. You can write some C routine and download it in target and run it. For example, suppose you have a C routine called hello(), you can put hello here instead of printf. To learn more about the VxWorks, (suppose you have NT based Tornado environment), you can go to menu "help" and choose menu index, there you will take either "Tornado user guide" or "VxWorks programmer's guide" these are excellent books. Hope it helps, Thanks. Jin "SungJin Lee" wrote in message news:<9ghoj1$aqm$1@imsinews.kornet.net>... > hi vxworkers, > > > When i try 'taskSpawn()' on the shell, it doesn't work > and shows the error message following. > > How can i use taskSpawn function in the vxworks 5.4 ? (with tornado 2) > > ----------------------------------------------------------------- > Exception number 4906: Task: 0x189720 () > > instruction translation miss > program counter: 0x04 > machine status : 0x9032 > trcStack aborted: error in top frame (sp) > ffffffff end +ffed846f: () > value = 159704 = 0x185e88 --------------------------- Newsgroups: comp.os.vxworks Subject: Re: how can i use 'taskSpawn' in vxworks 5.4 ? Date: Sun, 17 Jun 2001 16:07:53 +0100 From: molochai Message-ID: References: <9ghoj1$aqm$1@imsinews.kornet.net> In message nick.gardner@europe.simoco.com (nick) wrote: > Hmm, this is a dead quick answer, with not much thought at all, but I > think you should use ts() to spawn tasks from the shell. Anyway, give > it a try and see what happens. No. it's sp(). ts() == taskSuspend() > > Regards, > > Nick > > > hi vxworkers, > > > > > > When i try 'taskSpawn()' on the shell, it doesn't work > > and shows the error message following. > > > > How can i use taskSpawn function in the vxworks 5.4 ? (with tornado 2) > > > > ----------------------------------------------------------------- > > Exception number 4906: Task: 0x189720 () > > > > instruction translation miss > > program counter: 0x04 > > machine status : 0x9032 > > trcStack aborted: error in top frame (sp) > > ffffffff end +ffed846f: () > > value = 159704 = 0x185e88 Is taskSpawn being passed paremeters? If not it's not entirely surprising it doesn't work. taskSpawn takes about up to 14 (?) parameters, one of which is the entry point. This is not the first paremeter passed to the function either. taskSpwan(myEntryPoint) will definately not work. sp(myEntryPoint) will (parameters such as the task options, stack size and priority take default values). --------------------------- Newsgroups: comp.os.vxworks Subject: Fw: Trigraphs encountered. Date: Sun, 17 Jun 2001 20:02:07 +0530 From: "A.RadhaKrishna" Organization: Lawrence Berkeley National Laboratory Message-ID: <9gihie$685$1@overload.lbl.gov> This is a multi-part message in MIME format. - ------=_NextPart_000_003C_01C0F768.633BD960 Content-Type: text/plain; charset="Windows-1252" Content-Transfer-Encoding: quoted-printable - ----- Original Message -----=20 From: A.RadhaKrishna=20 To: vxworks_users@csg.lbl.gov ; vxwexplo-errs@csg.lbl.gov=20 Sent: Sunday, June 17, 2001 6:57 PM Subject: Trigraphs encountered. Hi VxWorkers, I am using Tornado2 for Pentium Target. I modified my harddisk driver file (ataDrv.c file) and built it using = "make CPU=3DPENTIUM" at dos prompt. It is working fine but I am getting a warning " 1 Trigraphs occurred". What is the meaning of this warning? Can you pl tell me Thanks & Regards, A.Radha Krishna - ------=_NextPart_000_003C_01C0F768.633BD960 Content-Type: text/html; charset="Windows-1252" Content-Transfer-Encoding: quoted-printable
 
----- Original Message -----=20
From: A.RadhaKrishna
To: vxworks_users@csg.lbl.gov = ; vxwexplo-errs@csg.lbl.gov =
Sent: Sunday, June 17, 2001 6:57 PM
Subject: Trigraphs encountered.

Hi VxWorkers,
 
 
 
I am using Tornado2 for Pentium Target.
I modified my harddisk driver file (ataDrv.c file) = and built=20 it using "make CPU=3DPENTIUM" at dos prompt.
It is working fine but I am getting a warning " 1 = Trigraphs=20 occurred".
 
What is the meaning of this = warning?
 
 
 
Can you pl tell me
 
 
Thanks & Regards,
A.Radha Krishna
- ------=_NextPart_000_003C_01C0F768.633BD960-- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: hey friendZ Date: 17 Jun 2001 08:33:26 -0700 From: DrSec@hotmail.com (DrSec) Organization: http://groups.google.com/ Message-ID: References: <3B2B8B62.B1C131F4@gv.net> "Michael R. Kesti" wrote in message news:<3B2B8B62.B1C131F4@gv.net>... > DrSec wrote: > > My respect for vendors such as Wind River prevents me from doing such > things. > yes , but I didn't ask about your feeling in your life , I'm a student in a small town who try to join a great university and I wanna make my resume good , and for this reason I need to work 2-3 months with thses softwares. After i joined to a good university and then a good company where they have money to buy these tools for their engineers ( like me ), i won't try to reply my feeling to request from a poor studnet in far town ... > > And my self-respect prevents me from accepting such offers. > > What do you respect, DrSec? this isn't an offer for rich company's engineers , this's only a help for ppl like me ... regards DrSec --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How common is SNiFF usage? Date: Sun, 17 Jun 2001 17:56:01 +0200 From: "Martin Raabe" Organization: T-Online Message-ID: <9gijt8$abo$03$1@news.t-online.com> References: Hello Joe, I am Martin R. as Dr. Diag referred to, and I would say, that team development is increasingly coming into the embedded world, as is is since years in the non-embedded world. I am interrested, what of SNiFF+ 4.0.1 is not to your satisfaction? Please reply to me at: Martin.Raabe at BaSystem.de "Joe Woodbury" schrieb im Newsbeitrag news:y8yW6.736$s_.560883@news.uswest.net... > My local rep is pushing SNiFF+ at us for our next project. I used SNiFF+ 2 > and didn't like it much, but gave the SNiFF+ 4.x demo a whirl to see if it > could change my mind. I still didn't like it. However, with future > employment possibilities in mind, I'm wondering how many > developers/companies use SNiFF+ (or at least insist it be used)? > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Fw: Trigraphs encountered. Date: Sun, 17 Jun 2001 17:11:37 +0100 From: molochai Message-ID: <9b4a1d8c4a%molochai@vapour-trail.demon.co.uk> References: <9gihie$685$1@overload.lbl.gov> In message <9gihie$685$1@overload.lbl.gov> "A.RadhaKrishna" wrote: > This is a multi-part message in MIME format. > > ------=_NextPart_000_003C_01C0F768.633BD960 Content-Type: text/plain; > charset="Windows-1252" Content-Transfer-Encoding: quoted-printable > > > ----- Original Message -----=20 From: A.RadhaKrishna=20 To: > vxworks_users@csg.lbl.gov ; vxwexplo-errs@csg.lbl.gov=20 Sent: Sunday, June > 17, 2001 6:57 PM Subject: Trigraphs encountered. > > > Hi VxWorkers, > > > > I am using Tornado2 for Pentium Target. I modified my harddisk driver file > (ataDrv.c file) and built it using = "make CPU=3DPENTIUM" at dos prompt. It > is working fine but I am getting a warning " 1 Trigraphs occurred". Strangely enough it menas that trigraphs were found in the code. Trigraphs are a means of generating certain various non-alphanumeric characters in character sets that use those character codes for other purposes. Examples would be ??= -> # ??( -> [ ??) -> ] and so on. There are also Digraphs which are two character combinations and various keyword equivalents for comparison operators. The one you've probably got is '??!' in a comment (which == |) This strikes me as a bad case of design-by-commitee and challange anybody to actually be able to understand code written with them. I think the authors of GCC deserve the last word on the subject: they descibe trigrpahs as brain-damage. Sadly, interpretting trigraphs is enabled with the '-ansi' option. > > What is the meaning of this warning? > > > > Can you pl tell me > > > Thanks & Regards, > A.Radha Krishna > > ------=_NextPart_000_003C_01C0F768.633BD960 > Content-Type: text/html; > charset="Windows-1252" > Content-Transfer-Encoding: quoted-printable > > > > charset=3Dwindows-1252"> > > > > >
 
>
----- Original Message -----=20 >
From: title=3Dkrishna.ar@cmcltd.com=20 > href=3D"mailto:krishna.ar@cmcltd.com">A.RadhaKrishna
>
To: href=3D"mailto:vxworks_users@csg.lbl.gov">vxworks_users@csg.lbl.gov = > ; title=3Dvxwexplo-errs@csg.lbl.gov=20 > href=3D"mailto:vxwexplo-errs@csg.lbl.gov">vxwexplo-errs@csg.lbl.gov = >
>
Sent: Sunday, June 17, 2001 6:57 PM
>
Subject: Trigraphs encountered.
>

>
Hi VxWorkers,
>
 
>
 
>
 
>
I am using Tornado2 for Pentium Target.
>
I modified my harddisk driver file (ataDrv.c file) = > and built=20 > it using "make CPU=3DPENTIUM" at dos prompt.
>
It is working fine but I am getting a warning " 1 = > Trigraphs=20 > occurred".
>
 
>
What is the meaning of this = > warning?
>
 
>
 
>
 
>
Can you pl tell me
>
 
>
 
>
Thanks & Regards,
>
A.Radha Krishna
> > ------=_NextPart_000_003C_01C0F768.633BD960-- > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Trigraphs encountered. Date: Sun, 17 Jun 2001 16:21:28 GMT From: "Gaurav Dixit" Organization: Road Runner Message-ID: References: <9gihie$685$1@overload.lbl.gov> you must be using some trigraph in your code. Trigraph sequences are three-character sequences that are replaced by a corresponding single character in Translation Phase 1, as follows: ??= # ??( [ ??< { ??/ \ ??) ] ??> } ??' ^ ??! | ??- ~ No other such sequences are recognized. The trigraph sequences provide a way to specify characters that are missing on some terminals, but that the C language uses. just search in your source file, if you have anyone of them in your file. Even within comments, inside string, or inside #if 0 ...#endif. as trigraph conversion is the first thing during compilation (or preprocessing). thanks - -Gaurav "A.RadhaKrishna" wrote in message news:9gihie$685$1@overload.lbl.gov... > This is a multi-part message in MIME format. > > ------=_NextPart_000_003C_01C0F768.633BD960 > Content-Type: text/plain; > charset="Windows-1252" > Content-Transfer-Encoding: quoted-printable > > > ----- Original Message -----=20 > From: A.RadhaKrishna=20 > To: vxworks_users@csg.lbl.gov ; vxwexplo-errs@csg.lbl.gov=20 > Sent: Sunday, June 17, 2001 6:57 PM > Subject: Trigraphs encountered. > > > Hi VxWorkers, > > > > I am using Tornado2 for Pentium Target. > I modified my harddisk driver file (ataDrv.c file) and built it using = > "make CPU=3DPENTIUM" at dos prompt. > It is working fine but I am getting a warning " 1 Trigraphs occurred". > > What is the meaning of this warning? > > > > Can you pl tell me > > > Thanks & Regards, > A.Radha Krishna > > ------=_NextPart_000_003C_01C0F768.633BD960 > Content-Type: text/html; > charset="Windows-1252" > Content-Transfer-Encoding: quoted-printable > > > > charset=3Dwindows-1252"> > > > > >
 
>
----- Original Message -----=20 >
From: title=3Dkrishna.ar@cmcltd.com=20 > href=3D"mailto:krishna.ar@cmcltd.com">A.RadhaKrishna
>
To: href=3D"mailto:vxworks_users@csg.lbl.gov">vxworks_users@csg.lbl.gov = > ; title=3Dvxwexplo-errs@csg.lbl.gov=20 > href=3D"mailto:vxwexplo-errs@csg.lbl.gov">vxwexplo-errs@csg.lbl.gov = >
>
Sent: Sunday, June 17, 2001 6:57 PM
>
Subject: Trigraphs encountered.
>

>
Hi VxWorkers,
>
 
>
 
>
 
>
I am using Tornado2 for Pentium Target.
>
I modified my harddisk driver file (ataDrv.c file) = > and built=20 > it using "make CPU=3DPENTIUM" at dos prompt.
>
It is working fine but I am getting a warning " 1 = > Trigraphs=20 > occurred".
>
 
>
What is the meaning of this = > warning?
>
 
>
 
>
 
>
Can you pl tell me
>
 
>
 
>
Thanks & Regards,
>
A.Radha Krishna
> > ------=_NextPart_000_003C_01C0F768.633BD960-- > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: hey friendZ Date: Sun, 17 Jun 2001 09:57:20 -0700 From: DrDiags Organization: laserlink.net Message-ID: <3B2CE170.9DC8BC6D@covad.net> References: <3B2B8B62.B1C131F4@gv.net> Reply-To: drdiags@covad.net DrSec, If you go to WindSurf at www.windriver.com, there is a *FREE* download of Tornado2 which will run on your host. At one time, it didn't contain full-networking support for simulated targets, but I am not sure what it contains now. Why don't you see if this will get you where you want to go? As far as your offer of an exchange of Intellectual Property, these are professionals who make their life blood on creating products that are based on some form of Non-Disclosure Acts and Intellectual Property rights. Your post seems to violate the spirit of what is expected as acceptable behavior. I don't pretend to be the moral leader here, but also think of what you are asking. DrSec wrote: > > "Michael R. Kesti" wrote in message news:<3B2B8B62.B1C131F4@gv.net>... > > DrSec wrote: > > > > My respect for vendors such as Wind River prevents me from doing such > > things. > > > yes , but I didn't ask about your feeling in your life , I'm a student > in a small town who try to join a great university and I wanna make my > resume good , and for this reason I need to work 2-3 months with thses > softwares. After i joined to a good university and then a good company > where they have money to buy these tools for their engineers ( like me > ), i won't try to reply my feeling to request from a poor studnet in > far town ... > > > > > And my self-respect prevents me from accepting such offers. > > > > What do you respect, DrSec? > > this isn't an offer for rich company's engineers , this's only a help > for ppl like me ... > > regards > DrSec --------------------------- Newsgroups: comp.os.vxworks Subject: re: Fw: Trigraphs encountered. Date: Sun, 17 Jun 2001 20:07:10 +0300 From: Michael.Ben-Ari@ecitele.com Organization: Lawrence Berkeley National Laboratory Message-ID: <9giojj$8m1$1@overload.lbl.gov> >I am using Tornado2 for Pentium Target. >I modified my harddisk driver file (ataDrv.c file) and built it using = >"make CPU=3DPENTIUM" at dos prompt. >It is working fine but I am getting a warning " 1 Trigraphs occurred". > >What is the meaning of this warning? This has NOTHING to do with Tornado or VxWorks (per se) You most probably have in your code (EVEN in COMMENTS) a trigraph sequence -- see K&R (Kernighan & Ritchie; A12.1 - This is new to the ANSI standard -- But as the GNU manual states "You don't want to know about this brain-damage..." use the -ansi switch or -trigraphs switch upon compilation... Or better yet, get rid of any comments that comtain a trigraph sequence '??X' (see K&R for definitions of X). - -Mike - ----------------------------------------------- Mike Ben-Ari ECI Telecom - NGTS Ltd. Cellular Division - Celtro Mail: mbenari@ecitele.com Tel: +972-3-926-6794 Fax: +972-3-926-6460 - ----------------------------------------------- --------------------------- Newsgroups: comp.os.vxworks Subject: re: Fw: Trigraphs encountered. Date: Sun, 17 Jun 2001 10:24:06 -0700 From: Susanna Jacobson Organization: Lawrence Berkeley National Laboratory Message-ID: <9gis3k$9t6$1@overload.lbl.gov> I got a trigraph warning once when I had put "???" in a comment, (in a comment!!) something like /* Why does this happen??? */ When I reduced the number of question marks to 1, the warning went away. Hoping this helps, Susanna > ----- Original Message -----=20 > From: A.RadhaKrishna=20 > To: vxworks_users@csg.lbl.gov ; vxwexplo-errs@csg.lbl.gov=20 > Sent: Sunday, June 17, 2001 6:57 PM > Subject: Trigraphs encountered. > > Hi VxWorkers, > > I am using Tornado2 for Pentium Target. > I modified my harddisk driver file (ataDrv.c file) and built it using = > "make CPU=3DPENTIUM" at dos prompt. > It is working fine but I am getting a warning " 1 Trigraphs occurred". > > What is the meaning of this warning? > > Can you pl tell me > > Thanks & Regards, > A.Radha Krishna - -- ===================================================================== Susanna Jacobson MS 10-110 SRJacobson@LBL.Gov Lawrence Berkeley National Laboratory VOICE: (510) 486-7801 1 Cyclotron Road FAX: (510) 486-7374 Berkeley, CA 94720 ===================================================================== --------------------------- Newsgroups: comp.os.vxworks Subject: reaction to posix compat and sockets Date: Sun, 17 Jun 2001 16:15:00 -0400 From: Andrew S Gardner Organization: Carnegie Mellon, Pittsburgh, PA Message-ID: Does anyone feel strongly one way or another about the POSIX compatibility and berkley sockets modules? I'm interested in developing a rather complex application, but would like to be able to leverage other hosts/targets for an initial test platform, and was hoping that using the POSIX and Berkely socket APIs instead of more uniquely VxWorks ones would give me that freedom. I'm wondering if anyone with experience using either of the two or both would share their experience with me ( they're to big, slower than another API, the best thing since sliced bread, etc. ). Thanks, Andrew Gardner asg@andrew.cmu.edu --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How common is SNiFF usage? Date: Sun, 17 Jun 2001 14:59:23 -0600 From: "Joe Woodbury" Message-ID: References: <9gijt8$abo$03$1@news.t-online.com> There is no direct connection between SNiFF and team development. SNiFF is an object browser, not a team collaboration tool. SNiFF may facilitate large project develpment, which is often done in a team environment, but it completely lacks the other collaboration tools more commonly associated with a team development environment. As far as particular criticisms of SNiFF go; my distaste of object browsers is that they don't really clarify anything. You end up with massive lists of functions and classes (and data and function members), none of which give you much more than a clinical antiseptic view of the project. For example, say you look at list of member functions of a class and it's parents. That view tells you nothing about how the class works except by intuition (i.e. experience in how objects work based on how they are named and how they are usually implemented.) But why should you choose one overloaded member function over another? Should you derive a new class and overload a member function or add a virtual function? Beyond that, there is the actual implementation. While it may be "nice" to show every function and every define from every file, the shear volume of information (generally without context) solves nothing for me. There's a define or typedef named something with a certain value; big deal. So now I won't create a define by that name, but I still don't know what that define is specifically for. And if I want to find out the value of a specific define, there is either the documentation or search file (grep) functionality. The advantage of the latter, which I can do very quickly, is that it tells me both the context of where it was defined, along with the associated defines which may prove even more helpful, especially if that's what I really wanted in the first place, and other places where it was used. That SNiFF can act as a giant index is not disputed, only that there is no compelling reason to use it for just that reason since otherwise, I don't find it a well written tool. It's awkward, uses screen space poorly, doesn't allow common settings to be made or changed. Like the other object browsers I've tried, it doesn't work how I intuitively want it to. Mainly, I have to do many clicks and take entirely unintuitive steps to traverse the project from an object perspective. On top of that, with the projects I'm working on, as with most projects in the real world, the objects themselves don't tell you much of anything. The browser may tell you what exists and the programatic relationship of those items, but it tells you nothing about how those objects interact. And that programatic relationship may be misleading; just because task A calls the function to send a message to task B, doesn't mean it actually does so in operation or that it may do so only sporadically. Unless you know the code, the conclusions you draw from the browser may be very misleading. If you know the code, then, as I said before, the browser in many becomes just a fancy grep engine. But that's not what's at issue. Are people REALLY using SNiFF, or are they buying it so their ignorant bosses can feel better about how they are "managing" the project. (Or so they can convince themselves they can hire inexperienced developers and let them use SNiFF to understand what's going on.) (I've yet to work with a genuine [vs. those with just the title] Sr. Developer/Engineer I respect who uses object browsers. I'm curious if my experience is atypical.) "Martin Raabe" wrote in message news:9gijt8$abo$03$1@news.t-online.com... > Hello Joe, > I am Martin R. as Dr. Diag referred to, and I would say, that team > development > is increasingly coming into the embedded world, as is is since years in the > non-embedded world. > > I am interrested, what of SNiFF+ 4.0.1 is not to your satisfaction? > Please reply to me at: > > Martin.Raabe at BaSystem.de > > > "Joe Woodbury" schrieb im Newsbeitrag > news:y8yW6.736$s_.560883@news.uswest.net... > > My local rep is pushing SNiFF+ at us for our next project. I used SNiFF+ 2 > > and didn't like it much, but gave the SNiFF+ 4.x demo a whirl to see if it > > could change my mind. I still didn't like it. However, with future > > employment possibilities in mind, I'm wondering how many > > developers/companies use SNiFF+ (or at least insist it be used)? > > > > > > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Build Error !!! Date: Sun, 17 Jun 2001 14:31:56 -0700 From: "Shesha Bhushan" Organization: Arizona State University Message-ID: <9gj7ju$7h5$1@news.asu.edu> Hi, I am a novice. I want to create a bootable image of Vx Works using Tornado. I have created a bootable project, and in the VXWorks Tab, I haden't included "hardware -) buses -) scsi -) ScsiVersion -) SCSI2" and was able to build vxworks.exe. Now i want to include this SCSI2. Then it gives the following error. How to solve this. Can anybody help me out please. ================== *** ERROR *** ================== partialImage.o: In function `_tf13bad_exception': \wpwr\host\src\gnu.simpc.cpp\gcc\cp\exception.cc(.text+0xa21): undefined reference to `sysScsiInit' make: *** [vxWorks] Error 0x1 ==================== --------------------------- Newsgroups: comp.os.vxworks Subject: Re: how can i use 'taskSpawn' in vxworks 5.4 ? Date: Mon, 18 Jun 2001 10:18:39 +0900 From: "SungJin Lee" Organization: Korea Telecom Message-ID: <9gjktr$288$1@imsinews.kornet.net> References: <9ghoj1$aqm$1@imsinews.kornet.net> I also tried taskSpawn with taskSpawn "test", 20, 0, 1000, printf("hello\n"), 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 it works ...but little strange. it works and shows the "hello" message. but it also print error message " too many arguments ... " so I decreased the arguments following printf( ) function. however it was same. o when it has 8, 9, 10 arguments, it shows " too many arguments ... " error message but printf the message " hello" on the monitor taskSpawn ( ............................. 0, 0, 0, 0, 0, 0, 0, 0, 0, 0) taskSpawn ( ............................. 0, 0, 0, 0, 0, 0, 0, 0, 0) taskSpawn ( ............................. 0, 0, 0, 0, 0, 0, 0, 0) o when it has 1 ~ 7 arguments, it shows " Exception number 4906 ... " taskSpawn ( ............................. 0, 0, 0, 0, 0, 0, 0) : : taskSpawn ( ............................. 0) I've not tried sp() yet. i'll try it more. thanks for good advises and let me know if you can find something more. =================================================================== "molochai" wrote in message news:b974178c4a%molochai@vapour-trail.demon.co.uk... : In message : nick.gardner@europe.simoco.com (nick) wrote: : : > Hmm, this is a dead quick answer, with not much thought at all, but I : > think you should use ts() to spawn tasks from the shell. Anyway, give : > it a try and see what happens. : : No. it's sp(). ts() == taskSuspend() : : > : > Regards, : > : > Nick : > : > > hi vxworkers, : > > : > > : > > When i try 'taskSpawn()' on the shell, it doesn't work : > > and shows the error message following. : > > : > > How can i use taskSpawn function in the vxworks 5.4 ? (with tornado 2) : > > : > > ----------------------------------------------------------------- : > > Exception number 4906: Task: 0x189720 () : > > : > > instruction translation miss : > > program counter: 0x04 : > > machine status : 0x9032 : > > trcStack aborted: error in top frame (sp) : > > ffffffff end +ffed846f: () : > > value = 159704 = 0x185e88 : : Is taskSpawn being passed paremeters? If not it's not entirely surprising it : doesn't work. taskSpawn takes about up to 14 (?) parameters, one of which is : the entry point. This is not the first paremeter passed to the function : either. : : taskSpwan(myEntryPoint) will definately not work. sp(myEntryPoint) will : (parameters such as the task options, stack size and priority take default : values). --------------------------- Newsgroups: comp.os.vxworks Subject: Re: vxsys utility under WIN NT4.0 Date: 17 Jun 2001 22:49:35 -0700 From: igtorque@eliop.es (=?ISO-8859-1?Q?Ignacio_Gonz=E1lez?=) Organization: http://groups.google.com/ Message-ID: <9ec53d11.0106172149.7038a19d@posting.google.com> References: <9gd8u6$q7p$1@overload.lbl.gov> > [AndrayK]. I think VxSYS is DOS utility, and doesn't work under Windows (at > least NT4/5). You can't run it from DOS box, you need to actually boot > DOS. > Right: it looks for io.sys and msdos.sys, which are not present or accesible in an MS-DOS box. However, if you have double boot (MS-DOS / WNT), just boot from MS-DOS. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Booting from ATA disk ... crash Date: 18 Jun 2001 00:31:59 -0700 From: igtorque@eliop.es (=?ISO-8859-1?Q?Ignacio_Gonz=E1lez?=) Organization: http://groups.google.com/ Message-ID: <9ec53d11.0106172331.5cf925c3@posting.google.com> References: <9ec53d11.0106140225.75f234a8@posting.google.com> Thanks for your suggestions, but I think we are missing the point. I've just taken a look at message: 15 Jun 2001 Attaching network interface cpm0... 0x1fe6aa8 (tBoot): and it reports a problem very similar to mine. In fact, I have been able to boot from MS-DOS via vxload bootrom.dat (from C:\) and, then, loading vxworks via Ethernet. What I cannot do is to load vxworks from C:\, because it still complains the same way: ...0x7febe800 (tBoot): memPartAlloc: block too big... I think I'm trying to play with memAddToPool in sysHwInir and see... I have 128MB in my motherboard. --------------------------- Newsgroups: comp.os.vxworks Subject: PowerPC VME read problem Date: Mon, 18 Jun 2001 06:58:01 -0400 From: "Chris Slominski" Organization: Jefferson Lab Message-ID: <9gkmsn$dqv$1@inn.jlab.org> VxWorks Group, The news server an my work site has been down and I never saw my first posting of this, or any responses. Sorry for the repeat. - ------------------- VxWorks Users, I am porting code from a mv167 to a mv2700 PowerPC. My code reads consecutive 8-bit quantities from aVME board that does not support D32 acccess. I have a VME bus analyzer that I use to inspect the VME activity of the software. When the software begin, the 8-bit reads are made as D8 VME accesses, but very rapidly the hardware starts making D32 accesses of the consecutive addresses instead. The D32 accesses cause the software to fail. How do I force the hardware to not perform D32 accesses? - -------------------- I am going to try to modify the VDW bit of the Universe chip's LSI2_CTL register to make 16-bit VME accesses the maximum. Does this solution sound correct to anyone? Chris --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Mon Jun 18 06:49:07 2001 From: "Watson, John" Date: Mon Jun 18 06:49:10 PDT 2001 Subject: Ethernet Initialisation This is a multi-part message in MIME format. --------------InterScan_NT_MIME_Boundary Content-Type: text/plain; charset="iso-8859-1" Hi, I'm trying to add an Ethernet card to a SBC (mbx860) which is running vxWorks. My question is about the sequence of initialisation of the card, do all init functions need to be called during boot sequence or can they be called in the user program. Also which are the involved functions in this process. I'm running a 82557 Ethernet controller. Thanks for any help, John Watson. --------------InterScan_NT_MIME_Boundary Content-Type: text/plain; name="RMRL-Disclaimer.txt" Content-Transfer-Encoding: 7bit Content-Disposition: attachment; filename="RMRL-Disclaimer.txt" The information contained in this e-mail is confidential to Roke Manor Research Ltd and must not be passed to any third party without permission. This communication is for information only and shall not create or change any contractual relationship. --------------InterScan_NT_MIME_Boundary-- From vxwexplo-errs@csg.lbl.gov Mon Jun 18 18:23:00 2001 From: "Antony Lo \(ATAN\)" Date: Mon Jun 18 18:23:02 PDT 2001 Subject: Target Hung after Telnet Client Connecting to Hi! I'm working on VxWorks V5.4 on Arm CPU. After I try to telnet to the target board, its console hang and the last info I saw is "telnetd: This system *IN USE* via telnet. shell restarted." Then I have to RESET the target. I want to provide both telnet client and console interface to users. Is there anything I miss ? Please see the info dumped to console as follows. Thanks, Antony ]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]] ]]]]]]]] ]] ]]]]]]]]]]]]] ]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]] ]]]]]]]]]]]]]]]]]]]]]]]]]]] V ] ]]]]]]]]] ]]]]]] ]]]]]]]] ]] ]]]] ]] ]]]]]]] ]]]]]]]] ]]]]]] ] ]] ]]]] ]]] ]]]]] ] ]]] ] ]]]] ]]] ]]]]]]]]] ]]]] ]] ]]]] ]] ]]]]] ]]]] ]]] ]] ] ]]] ]] ]]]]] ]]]]]] ]] ]]]]]]] ]]]] ]] ]]]] ]]]]] ] ]]]] ]]]]] ]]]]]]]] ]]]] ]] ]]]] ]]]]]]] ]]]] ]]]]]] ]]]]] ]]]]]] ] ]]]]] ]]]] ]] ]]]] ]]]]]]]] ]]]] ]]]]]]] ]]]]] ] ]]]]]] ] ]]] ]]]] ]] ]]]] ]]]] ]]]] ]]]] ]]]]]]]] ]]]]] ]]] ]]]]]]] ] ]]]]]]] ]]]] ]]]] ]]]] ]]]]] ]]]]]]]]]]]]]]]]]]]]]]]]]]]]]] ]]]]]]]]]]]]]]]]]]]]]]]]]]]]] Development System ]]]]]]]]]]]]]]]]]]]]]]]]]]]] ]]]]]]]]]]]]]]]]]]]]]]]]]]] VxWorks version 5.4 ]]]]]]]]]]]]]]]]]]]]]]]]]] KERNEL: WIND version 2.5 ]]]]]]]]]]]]]]]]]]]]]]]]] Copyright Wind River Systems, Inc., 1984-1999 CPU: KS32C50100 FOR SNDS100 Ver 1.0. Processor # 0. Memory Size: 0x400000. BSP version 1.2/0. WDB: Ready. -> telnetd: someone tried to login. telnetd: This system *IN USE* via telnet. shell restarted. From vxwexplo-errs@csg.lbl.gov Tue Jun 19 00:41:50 2001 From: "Chantreux, Guy Felix" Date: Tue Jun 19 00:41:53 PDT 2001 Subject: Gnu compiler for VxWorks powerpc version 2.95.3 instead of the wi We need to use this version of compiler for c++ design reason. I download the sources and configure it, and begin to compile it. But before doing more job, I need to know : - if I have to do some changes - there is some specific target option as targ=powerpc-eabi or targ=powerpc-wrs in configure - it's better to use the old way used by wind river in version 5.3 with the gnu toolchain - if I use the configure, do I need additional option - will I have some problem with the windriver object libraries at run-time Thanks in advance, Guy From vxwexplo-errs@csg.lbl.gov Tue Jun 19 04:03:27 2001 From: Vxworks Exploder Date: Tue Jun 19 04:03:30 PDT 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Tue Jun 19 04:03:07 PDT 2001 Subject: Where is the definition of ipFilterLib ?? Subject: Cannot start the server Subject: Re: taskSpawn problem... Subject: Re: Cannot start the server Subject: Re: Running applications at boot time... Subject: Running applications at boot time... Subject: Re: Running applications at boot time... Subject: Re: TrueFFS INstallation Problem Subject: Ethernet Initialisation Subject: Re: VxWorks web servers & server pages technologies Subject: Re: Reserving Memory Subject: Re: taskSpawn problem... Subject: Re: TrueFFS INstallation Problem Subject: spurious interrupts Subject: vxWorks.st - how to build Subject: Dual 7410 board Subject: Re: module download from command prompt Subject: Re: vxWorks.st - how to build Subject: Re: taskSpawn problem... Subject: Re: Build Error !!! Subject: Re: VxWorks web servers & server pages technologies Subject: Re: PowerPC VME read problem Subject: Wind Web Server 2.0, httpSymTabAdd function failure Subject: Re: taskSpawn problem... Subject: Triggering Subject: Re: Running applications at boot time... Subject: Exception handling? Subject: Bufferless message queues Subject: usrNetInit Subject: Re: Q: Problem with sm network on mcp750 Subject: Re: Bufferless message queues Subject: Re: how can i use 'taskSpawn' in vxworks 5.4 ? Subject: romInit in VxWorks 8260 BSP Subject: any good website to start learning VxWorks device driver development ? _EOM Subject: Re: editor in vxworks Subject: Re: Problem with sm network on mcp750 Subject: Target Hung after Telnet Client Connecting to Subject: Re: how can i use 'taskSpawn' in vxworks 5.4 ? Subject: Re: Triggering Subject: Re: module download from command prompt Subject: Re: Timing problems on VxWorks x486 Subject: the fAQ Subject: Gnu compiler for VxWorks powerpc version 2.95.3 instead of the wi Subject: vxworks porting Subject: Re: Bufferless message queues ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Where is the definition of ipFilterLib ?? Date: Mon, 18 Jun 2001 14:40:10 +0200 From: Amit Raam Organization: Verio Message-ID: <6atritgcjlj50ml9unciv1a5d702hk1tr6@4ax.com> Hi, I've done a text search of all .h files, but still couldn't fine the prototype for ipFilterhookAdd etc. when entering the prototypes manually, I'm thrown out due to missing definition for 'ifnet' structure (used in the hook function definition). Tornado 2.0 i960JX Thanks, Amit Amit Raam Email address: raama at indigo.co.il - -------------------------------------------------- ".sigless and proud of it" (R) (C) (tm) (look-and-feel) Std. disclaimer:I don't represent the opinions of anyone but myself. --------------------------- Newsgroups: comp.os.vxworks Subject: Cannot start the server Date: 18 Jun 2001 04:45:35 -0700 From: i_satya@yahoo.com (satya kumar) Organization: http://groups.google.com/ Message-ID: Hi All, I have already creted a bootable floppy for vxworks and am successfully booting the the target and able to get the message WDB :ready .But as soon as I try to launch the server this error message is flashed after a while Connecting to target agent....succeded Attaching C++ interface.....succeded Attaching a.out OMF reader I80X86 CPU Family.....succeded error : Couldn't open C:\Tornado\vxWorks :status = 0x2 Error during target server core initialization. Please suggest me what to do.I am stuck over here.Thanks in advance. Satya --------------------------- Newsgroups: comp.os.vxworks Subject: Re: taskSpawn problem... Date: 18 Jun 2001 05:04:15 -0700 From: i_satya@yahoo.com (satya kumar) Organization: http://groups.google.com/ Message-ID: References: <9gdtdb$ebc$1@imsinews.kornet.net> Hi there, Well I dont know much of VxWorks but have some codes running.What you can try out is with take an integer value suppose taskid i.e int taskid; and then try to spawn the task... like taskid=taskSpawn(..............); OR try it like this if(((taskid=taskSpawn(.......))==ERROR)) { printf("taskSpawn of taskid failed \n"); } Just try it out and please, please let me know if it works.Well I think it should.Best of Luck. Satya --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Cannot start the server Date: Mon, 18 Jun 2001 15:32:43 +0200 From: Amit Raam Organization: Verio Message-ID: <4b0sit00pab53vdo9i10erp3t65svjo3mc@4ax.com> References: i_satya@yahoo.com (satya kumar) wrote: >Hi All, > I have already creted a bootable floppy for vxworks and am >successfully booting the the target and able to get the message WDB >:ready .But as soon as I try to launch the server this error message >is flashed after a while >Connecting to target agent....succeded >Attaching C++ interface.....succeded >Attaching a.out OMF reader I80X86 CPU Family.....succeded >error : Couldn't open C:\Tornado\vxWorks :status = 0x2 >Error during target server core initialization. >Please suggest me what to do.I am stuck over here.Thanks in advance. >Satya Looks like you need to configure your server. tools -> target server -> configure -> core file and symbol make sure you're pointing to the correct file. Amit Raam Email address: raama at indigo.co.il - -------------------------------------------------- ".sigless and proud of it" (R) (C) (tm) (look-and-feel) Std. disclaimer:I don't represent the opinions of anyone but myself. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Running applications at boot time... Date: Mon, 18 Jun 2001 08:53:10 -0400 From: Jeremiah Frick Organization: Improv Systems, Inc. Message-ID: <3B2DF9B6.1A8EDE26@improvsys.com> References: <3B2DF451.A176E1A0@improvsys.com> This is a multi-part message in MIME format. - --------------7D37210FC411D12D2D60F0A7 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Nevermind, the solution was easy. All set. Jeremiah Frick wrote: > Hello all, > > I think there should be a relatively easy solution to this, but I > figured I'd ask. > > I have some code that I'd like to run automatically immediately > following VxWorks booting. Is there anywhere that I need to simply add > the call to the application, or is there something else I have to do? > Basically, I'm trying to eliminate any interaction required from the > command prompt. > > Thanks, > Jeremy - --------------7D37210FC411D12D2D60F0A7 Content-Type: text/x-vcard; charset=us-ascii; name="jeremyf.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Jeremiah Frick Content-Disposition: attachment; filename="jeremyf.vcf" begin:vcard n:Frick;Jeremiah tel;work:716-756-5411 x45 x-mozilla-html:FALSE url:www.improvsys.com org:Improv Systems Inc.;PDG version:2.1 email;internet:jeremyf@improvsys.com title:Hardware Engineer adr;quoted-printable:;;3136 Winton Road South=0D=0ASuite 100;Rochester;New York;14623;United States of America fn:Jeremiah Frick end:vcard - --------------7D37210FC411D12D2D60F0A7-- --------------------------- Newsgroups: comp.os.vxworks Subject: Running applications at boot time... Date: Mon, 18 Jun 2001 08:30:09 -0400 From: Jeremiah Frick Organization: Improv Systems, Inc. Message-ID: <3B2DF451.A176E1A0@improvsys.com> This is a multi-part message in MIME format. - --------------05F4DB45A39E8F704F931884 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hello all, I think there should be a relatively easy solution to this, but I figured I'd ask. I have some code that I'd like to run automatically immediately following VxWorks booting. Is there anywhere that I need to simply add the call to the application, or is there something else I have to do? Basically, I'm trying to eliminate any interaction required from the command prompt. Thanks, Jeremy - --------------05F4DB45A39E8F704F931884 Content-Type: text/x-vcard; charset=us-ascii; name="jeremyf.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Jeremiah Frick Content-Disposition: attachment; filename="jeremyf.vcf" begin:vcard n:Frick;Jeremiah tel;work:716-756-5411 x45 x-mozilla-html:FALSE url:www.improvsys.com org:Improv Systems Inc.;PDG version:2.1 email;internet:jeremyf@improvsys.com title:Hardware Engineer adr;quoted-printable:;;3136 Winton Road South=0D=0ASuite 100;Rochester;New York;14623;United States of America fn:Jeremiah Frick end:vcard - --------------05F4DB45A39E8F704F931884-- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Running applications at boot time... Date: Mon, 18 Jun 2001 09:22:55 -0400 From: "Kevin Livingston" Message-ID: References: <3B2DF451.A176E1A0@improvsys.com> One option is to use the start up script option. You can then specify a start up script to run from the boot line I belive. The start up script takes commands just like they were run from the shell(prompt). The other option is to modify your kernel to make the call to the code you wish to run. If you look in the manuals under booting or startup sequences you should find the approiate function to add your "hook" into. I don't think it is UsrRoot, but it is something like that. if I am not mistaken you modify usrAppInit() Kevin "Jeremiah Frick" wrote in message news:3B2DF451.A176E1A0@improvsys.com... > Hello all, > > I think there should be a relatively easy solution to this, but I > figured I'd ask. > > I have some code that I'd like to run automatically immediately > following VxWorks booting. Is there anywhere that I need to simply add > the call to the application, or is there something else I have to do? > Basically, I'm trying to eliminate any interaction required from the > command prompt. > > Thanks, > Jeremy > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: TrueFFS INstallation Problem Date: 18 Jun 2001 06:24:38 -0700 From: aparker@vtlinx.com (parreg) Organization: http://groups.google.com/ Message-ID: <79ef8b60.0106180524.58de98ca@posting.google.com> References: <3B28E7E2.A56DA2E@digidata.com> <3B2A3C54.A6640BA5@digidata.com> Doug Dotson wrote in message news:<3B2A3C54.A6640BA5@digidata.com>... > Yes. And when I load it I take a look at noOfMTDs and it is 1 > indicating that one entry is in the MTD table. > > Doug > > > Andray Kaganovsky wrote: > > > Doug Dotson wrote in > > news:3B28E7E2.A56DA2E@digidata.com: > > > > > Greetings! > > > > > > I am having alot of trouble getting TFFS installed. Followed the > > > instructions > > > and got a successful build. After rebooting, I attempt to create a > > > device > > > using "tffsDevCreate" and get the following: > > > > > > -> tffsDevCreate > > > > > > Data bus error > > > Exception Program Counter: 0x800a60cc > > > Status Register: 0x3000ff01 > > > Cause Register: 0x0000001c > > > Error address: 0xffffffff, Error ID: 0x0000 > > > > > > 80041894 vxTaskEntry +c : shell (1, 0, 0, 0) > > > 800bef98 shell +1c8: 800befbc (eeeeeeee, eeeeeeee, eeeeeeee, > > > eeeeeeee) > > > 800bf1a4 shell +3d4: execute (800fe4c8, 8010b304, 2048, > > > &yypact) > > > 800bf30c execute +c8 : yyparse (80c77d88, ffffffff, 1, ffffffff) > > > 800ef040 yyparse +888: 800ed400 (eeeeeeee, 0, eeeeeeee, 0) > > > 800ed5c4 yystart +89c: tffsDevCreate (140, 258, 80077aec, 0) > > > 800a1a58 tffsDevCreate +4c : flCall (80c77cc0, 0, 80cfe920, 96) > > > 800a73bc flCall +248: 800a6540 (0, 0, 8007829c, 0) > > > 800a6588 tffscmpWords +270: flMount (80124c60, 800a6454, 0, 0) > > > 800f2050 flMount +30 : flIdentifyFlash (0, 0, 80c77b80, 0) > > > 800a617c flIdentifyFlash+38 : 800a609c (800ec5e8, 800ec590, 800ea708, > > > 800e990c) > > > > > > My specific configuration is: > > > > > > 1) Board is a Galileo EV-64240 with resident Intel E28F640J3A-120 > > > flash. > > > 2) I modified mv177-sysTtfs.c file to suit. > > > 3) I'm using the cfiscs.c MTD. > > > > > > I put in a few debug statements and have found that the rfaRegister() > > > function is being > > > called. I also put some debug statements in the flRegisterCFISCS() and > > > cfiscsIdentfy() > > > routines and they are never called. Since the exception is occurring at > > > 0xffffffff it appears > > > to me that the function pointers to the MTD functions are never being > > > set up but I can't > > > figure out how this is supposed to be done. > > > > > > Thanks, > > > Doug > > > > > > > > > Douglas P. Dotson > > > Digi-Data Corporation > > > 8580 Dorsey Run Rd. > > > Jessup, MD 20794 > > > 301-498-0200 x248 > > > > > > > > > > > > > [AndrayK]. Did you #defined INCLUDE_MTD_CFISCS in sysTffs.c ? you should format the flash first using sysTffsFormat() then use usrTffsConfig 0,0,"/flash" to create the device. Then type "devs" confirming the device is there. Once the device is created you can check the dosFs volumne configuration using "dosFsConFigShow("/flash"). parreg --------------------------- Newsgroups: comp.os.vxworks Subject: Ethernet Initialisation Date: Mon, 18 Jun 2001 14:48:48 +0100 From: "Watson, John" Organization: Lawrence Berkeley National Laboratory Message-ID: <9gl2dj$ba3$1@overload.lbl.gov> This is a multi-part message in MIME format. - --------------InterScan_NT_MIME_Boundary Content-Type: text/plain; charset="iso-8859-1" Hi, I'm trying to add an Ethernet card to a SBC (mbx860) which is running vxWorks. My question is about the sequence of initialisation of the card, do all init functions need to be called during boot sequence or can they be called in the user program. Also which are the involved functions in this process. I'm running a 82557 Ethernet controller. Thanks for any help, John Watson. - --------------InterScan_NT_MIME_Boundary Content-Type: text/plain; name="RMRL-Disclaimer.txt" Content-Transfer-Encoding: 7bit Content-Disposition: attachment; filename="RMRL-Disclaimer.txt" The information contained in this e-mail is confidential to Roke Manor Research Ltd and must not be passed to any third party without permission. This communication is for information only and shall not create or change any contractual relationship. - --------------InterScan_NT_MIME_Boundary-- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks web servers & server pages technologies Date: 18 Jun 2001 07:19:30 -0700 From: Eran_Gluska@packetlight.com (Eran Gluska) Organization: http://groups.google.com/ Message-ID: <1c30b9ff.0106180619.7928564b@posting.google.com> References: <9ggg24$18s$1@bcarh8ab.ca.nortel.com> Try www.goahead.com - they have a free web server already ported for VxWorks. "Lin, Ping [TOR:9T00:EXCH]" wrote in message news:<9ggg24$18s$1@bcarh8ab.ca.nortel.com>... > Hi, > > I'm looking at the possibility of creating dynamic web server pages > (similar to JSP/ASP) on VxWorks, and was wondering if the users of this > newsgroup can give me some pointers on: > > 1) Is Apache or any other web server available for VxWorks, besides the Wind > web server? > > 2) Is there an implementation of JSP, Java servlets or similar server pages > technology available for VxWorks? > > > Thanks in advance, > Ping Lin --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Reserving Memory Date: Mon, 18 Jun 2001 15:50:19 +0100 From: Paul Caswell Organization: Land & Sea Systems Message-ID: <3B2E152B.DE6C056E@baesystems.com> References: <3B29E711.C228F8A1@baesystems.com> It's OK, Ive cracked it. I'm using memDrv for those interested. Regards Paul --------------------------- Newsgroups: comp.os.vxworks Subject: Re: taskSpawn problem... Date: Mon, 18 Jun 2001 08:55:56 -0600 From: Don Small Organization: Sandia National Laboratories Message-ID: <3B2E167C.A89CC163@sandia.gov> References: <9gdtdb$ebc$1@imsinews.kornet.net> "±èº´È£" wrote: > > 'taskSpawn' function doesn't work on my VxWorks. > I tried it on the shell, like below. and have the exception error messages. > what should I do to make 'taskSpawn' works ? > > ---------------------------------------------- > -> taskSpawn() > > Exception number 4906: Task: 0x189720 () > > instruction translation miss > program counter: 0x04 > machine status : 0x9032 > trcStack aborted: error in top frame (sp) > ffffffff end +ffed846f: () > value = 159704 = 0x185e88 > -> > > ------------------------------------------------- > taskSpawn require parameters, otherwise the parameters passed are all 0 and it will attempt to start a task at address 0. The simplified method of starting a task from the command line is sp > and I also write the code to try 'taskSpawn' and > 'etherOutputHookAdd'. and download it to vxworks > and tried ... It also doesn't work.... > I think it's also because of that 'taskSpawn' problem. > I'll be really pleased if i can get advice to solve that problems... > can anyone see if there's any problem in the code ? > > --------------------------------------------------- > #include > #include > > void hook_test() > { > > BOOL OutputHandle(struct ifnet *pIf, char *buffer, int length) > { > > /* to see what is hooked*/ > printf("hooked_packet = %c\n", *buffer); > > return (FALSE); > } > > STATUS OutputHook() > { > if (etherOutputHookAdd(OutputHandle) == OK) { > printf("etherOutputHookAdd successful\n"); > return(OK); > } > else { > printf("etherOutputHook failed\n"); > return(ERROR); > } > } > > taskSpawn ("etherOutputHook", 20, 0, 2000, OutputHook, 0, 0, 0, 0, 0, 0, > 0, 0, 0, 0); > > } > ---------------------------------------------------------------------------- When I compiled, loaded and tried to run this piece of code, it failed also. I did not know that gcc allowed embedded functions within functions but the compiler does not complain. But I did have a load error. ld < test1.o undefined symbol: ___trampoline which seems to be crucial for the execution of this code. I don't know if you version of gcc/bsp creates code with this same unresolved reference but any time you have an unresolved reference it needs to be solved before you run the code. The loader normally resolves all references while loading the code and unresolved references are left at address 0. In this case it appears to be a function to help gcc call functions inside of functions. When I execute hook_test, this is what I get: - -> hook_test Illegal Instruction Program Counter: 0x00000008 Status Register: 0x30004 Task: 0x3d3fd98 "etherOutputHook" - -> Now, if I move two line as below: BOOL OutputHandle(struct ifnet *pIf, char *buffer, int length) { /* to see what is hooked*/ printf("hooked_packet = %c\n", *buffer); return (FALSE); } STATUS OutputHook() { if (etherOutputHookAdd(OutputHandle) == OK) { printf("etherOutputHookAdd successful\n"); return(OK); } else { printf("etherOutputHook failed\n"); return(ERROR); } } void hook_test() { taskSpawn ("etherOutputHook", 20, 0, 2000, OutputHook, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); } compile and load this works. Don --------------------------- Newsgroups: comp.os.vxworks Subject: Re: TrueFFS INstallation Problem Date: Mon, 18 Jun 2001 11:15:41 -0400 From: Doug Dotson Organization: Verio Message-ID: <3B2E1B1D.D4D1BE1D@digidata.com> References: <3B28E7E2.A56DA2E@digidata.com> <3B2A3C54.A6640BA5@digidata.com> <79ef8b60.0106180524.58de98ca@posting.google.com> I don't have a function sysTffsFormat. Don't I have to run tffsDevCreate first? When I run usrTffsConfig I get the same error. I put some printf's in cfiscsIdentify() and it never gets called. I checked noOfMTDs and it is 1 as it should be. Doug parreg wrote: > Doug Dotson wrote in message news:<3B2A3C54.A6640BA5@digidata.com>... > > Yes. And when I load it I take a look at noOfMTDs and it is 1 > > indicating that one entry is in the MTD table. > > > > Doug > > > > > > Andray Kaganovsky wrote: > > > > > Doug Dotson wrote in > > > news:3B28E7E2.A56DA2E@digidata.com: > > > > > > > Greetings! > > > > > > > > I am having alot of trouble getting TFFS installed. Followed the > > > > instructions > > > > and got a successful build. After rebooting, I attempt to create a > > > > device > > > > using "tffsDevCreate" and get the following: > > > > > > > > -> tffsDevCreate > > > > > > > > Data bus error > > > > Exception Program Counter: 0x800a60cc > > > > Status Register: 0x3000ff01 > > > > Cause Register: 0x0000001c > > > > Error address: 0xffffffff, Error ID: 0x0000 > > > > > > > > 80041894 vxTaskEntry +c : shell (1, 0, 0, 0) > > > > 800bef98 shell +1c8: 800befbc (eeeeeeee, eeeeeeee, eeeeeeee, > > > > eeeeeeee) > > > > 800bf1a4 shell +3d4: execute (800fe4c8, 8010b304, 2048, > > > > &yypact) > > > > 800bf30c execute +c8 : yyparse (80c77d88, ffffffff, 1, ffffffff) > > > > 800ef040 yyparse +888: 800ed400 (eeeeeeee, 0, eeeeeeee, 0) > > > > 800ed5c4 yystart +89c: tffsDevCreate (140, 258, 80077aec, 0) > > > > 800a1a58 tffsDevCreate +4c : flCall (80c77cc0, 0, 80cfe920, 96) > > > > 800a73bc flCall +248: 800a6540 (0, 0, 8007829c, 0) > > > > 800a6588 tffscmpWords +270: flMount (80124c60, 800a6454, 0, 0) > > > > 800f2050 flMount +30 : flIdentifyFlash (0, 0, 80c77b80, 0) > > > > 800a617c flIdentifyFlash+38 : 800a609c (800ec5e8, 800ec590, 800ea708, > > > > 800e990c) > > > > > > > > My specific configuration is: > > > > > > > > 1) Board is a Galileo EV-64240 with resident Intel E28F640J3A-120 > > > > flash. > > > > 2) I modified mv177-sysTtfs.c file to suit. > > > > 3) I'm using the cfiscs.c MTD. > > > > > > > > I put in a few debug statements and have found that the rfaRegister() > > > > function is being > > > > called. I also put some debug statements in the flRegisterCFISCS() and > > > > cfiscsIdentfy() > > > > routines and they are never called. Since the exception is occurring at > > > > 0xffffffff it appears > > > > to me that the function pointers to the MTD functions are never being > > > > set up but I can't > > > > figure out how this is supposed to be done. > > > > > > > > Thanks, > > > > Doug > > > > > > > > > > > > Douglas P. Dotson > > > > Digi-Data Corporation > > > > 8580 Dorsey Run Rd. > > > > Jessup, MD 20794 > > > > 301-498-0200 x248 > > > > > > > > > > > > > > > > > > [AndrayK]. Did you #defined INCLUDE_MTD_CFISCS in sysTffs.c ? > > you should format the flash first using sysTffsFormat() then > use usrTffsConfig 0,0,"/flash" to create the device. Then type "devs" > confirming the device is there. Once the device is created you can > check the dosFs volumne configuration using > "dosFsConFigShow("/flash"). > > parreg --------------------------- Newsgroups: comp.os.vxworks Subject: spurious interrupts Date: 18 Jun 2001 10:15:55 -0500 From: "john cobarruvias" Message-ID: <3b2e1b2b$1_1@newsa.ev1.net> Reply-To: "john cobarruvias" Sender: "john cobarruvias" I have a PPC603 board with a serial port on the board and one on an ethernet card. When we push data out on the ethernet cars serial port and then on the cpu's port, we get "Uninitialized interrupt PIC 7" errors. The Winbond register defaults to 7. Therefore I believe it is a "spurious" interrupt and just could be ignored. Is this an acceptable course of action to take in the intHandler? John Cobarruvias --------------------------- Newsgroups: comp.os.vxworks Subject: vxWorks.st - how to build Date: Mon, 18 Jun 2001 11:44:08 -0700 From: "Sean Bose" Organization: Lucent Technologies, Columbus, Ohio Message-ID: <9gl7cc$24c@nntpa.cb.lucent.com> Hi I am a newbie to the x86 flavour of VxWorks. I am trying to compile and create a bootable floppy from which I can boot a PC with vxWorks. I understand that vxWorks.st is the standalone image from which I can create a bootable floppy using mkboot.bat and vxsys.com. I have Tornado 2.0 and I am unable to figure out how to make a working vxWorks.st. Please suggest ways to compile / make vxWorks.st. thanks to my helpful vxWorks friends Sean --------------------------- Newsgroups: comp.os.vxworks Subject: Dual 7410 board Date: Mon, 18 Jun 2001 16:02:36 GMT From: cgan76@nospam4me.com (cgan76) Organization: MTS Internet Message-ID: <3b2e2452.506384021@200.123.101.100> Greetings, Has anyone out there developed VxWorks BSPs for dual MPC boards? Does anyone have any information on how this can be done in VxWorks and where I can get started? I'm fairly sure that VxWorks does not support symettric multiprocessing but is there an alternate option that can be used? Here's the basic hardware: Dual MPC7410 128MB SDRAM Local PCI bus Your help would be greatly appreciated! --------------------------- Newsgroups: comp.os.vxworks Subject: Re: module download from command prompt Date: Mon, 18 Jun 2001 11:50:34 -0400 From: "Daniel G Waddington" Organization: Bell Laboratories, Lucent Technologies Message-ID: <9gl82c$66q2@news.research.bell-labs.com> References: <9g8afb$c526@news.research.bell-labs.com> <3B282D64.D7CE9590@support.nl> I meant from the DOS prompt. DAN "Kees van der Bent" wrote in message news:3B282D64.D7CE9590@support.nl... > Maybe you have specified the problem not clear enough... But reading > your question I would initially say using: > > -> ld < module-(full-)path-name > > This seems to simple and it's all in the manual... > > > > Daniel G Waddington wrote: > > > Hi, anyone know how to download modules from the command line, using Tornado > > II? > > > > Thanks > > > > DAN > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: vxWorks.st - how to build Date: Mon, 18 Jun 2001 18:15:11 +0200 From: "Werner Schiendl" Organization: VBS - Vienna Backbone Service Message-ID: <992881382.800716@newsmaster-04.atnet.at> References: <9gl7cc$24c@nntpa.cb.lucent.com> To make a VxWorks bootdisk, you need bootrom_uncmp and mkboot.bat. You can build the bootrom_uncmp from the command line with make bootrom_uncmp in your BSP directory. Be sure to edit config.h (in your BSP directory, too) as required. To bring it to a disk type format a: mkboot a: bootrom_uncmp (The format is optional, but saves you headaches. You can use /q to save time - quick format) You can then copy any VxWorks image onto the disk too. Using Tornado II project facility, the name of the image is usually just VxWorks, but that does not matter. Just include the target development utilities (target shell, target loader, Show Routines, etc.) as required and build the image. hth werner "Sean Bose" wrote in message news:9gl7cc$24c@nntpa.cb.lucent.com... > Hi > I am a newbie to the x86 flavour of VxWorks. I am trying to > compile and create a bootable floppy from which I can boot a > PC with vxWorks. > I understand that vxWorks.st is the standalone image from which > I can create a bootable floppy using mkboot.bat and vxsys.com. > I have Tornado 2.0 and I am unable to figure out how to make a > working vxWorks.st. Please suggest ways to compile / make > vxWorks.st. > > thanks to my helpful vxWorks friends > Sean > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: taskSpawn problem... Date: Mon, 18 Jun 2001 12:18:19 -0400 From: Charles H. Chapman Organization: NASA Message-ID: <992881271.1061574211@news.ksc.nasa.gov> References: <9gdtdb$ebc$1@imsinews.kornet.net> Followup-To: comp.os.vxworks On Fri, 15 Jun 2001, ±èº´È£ wrote: >'taskSpawn' function doesn't work on my VxWorks. >I tried it on the shell, like below. and have the exception error messages. >what should I do to make 'taskSpawn' works ? > > >---------------------------------------------- >-> taskSpawn() > >Exception number 4906: Task: 0x189720 () > >instruction translation miss >program counter: 0x04 >machine status : 0x9032 >trcStack aborted: error in top frame (sp) >ffffffff end +ffed846f: () >value = 159704 = 0x185e88 If you just typed taskSpawn() and no arguments like you showed above then it's no wonder that it crashed. You have to supply a bunch of arguments to taskSpawn to tell it things like what function to spawn, the priority of the task, the statck space to use, arguments for the spawned function, etc. Chuck --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Build Error !!! Date: Mon, 18 Jun 2001 18:39:41 +0200 From: "Werner Schiendl" Organization: VBS - Vienna Backbone Service Message-ID: <992882853.244764@newsmaster-04.atnet.at> References: <9gj7ju$7h5$1@news.asu.edu> For which platform do you try to compile this? It looks to me from the file path you are trying to build a simulator with the SCSI option? I'm not sure if this is supported, at least not with the simulator that ships for 'free' with Tornado II. regards werner "Shesha Bhushan" wrote in message news:9gj7ju$7h5$1@news.asu.edu... > Hi, > > I am a novice. I want to create a bootable image of Vx Works using Tornado. > I have created a bootable project, and in the VXWorks Tab, I haden't > included "hardware -) buses -) scsi -) ScsiVersion -) SCSI2" and was able to > build vxworks.exe. Now i want to include this SCSI2. Then it gives the > following error. How to solve this. Can anybody help me out please. > ================== > *** ERROR *** > ================== > partialImage.o: In function `_tf13bad_exception': > \wpwr\host\src\gnu.simpc.cpp\gcc\cp\exception.cc(.text+0xa21): undefined > reference to `sysScsiInit' > make: *** [vxWorks] Error 0x1 > ==================== > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks web servers & server pages technologies Date: Mon, 18 Jun 2001 11:01:20 -0600 From: Les Hammer Organization: Agilent Technologies Message-ID: <3B2E33E0.5C3408B3@agilent.com> References: <9ggg24$18s$1@bcarh8ab.ca.nortel.com> <1c30b9ff.0106180619.7928564b@posting.google.com> Eran Gluska wrote: > Try www.goahead.com - they have a free web server already ported for VxWorks. Some caveats: 1) I tried to use goahead. I ran into a problem where something wasn't working right. Since this was a free web server, there was no one to answer my questions and get the problem fixed. (They have a news group, but unless a willing and knowlegable person is there, you are on your own.) 2) Goahead requires you to display their logo on your main page. Company lawyers don't like that. 3) Goahead requires that you allow them to display your company's logo. Company lawyers *really* don't like that. (It is disliked so much, that unless you obtain special permission, you can be terminated from HP or Agilent Technologies for giving goahead permission to do it. Other companies may have similar policies.) 4) Lawyers also don't like the concept of, if goahead just happened to use somebody else's copyrighted code (knowingly or unknowingly), it is our company that can be sued. (To be more specific, our lawyers don't like that. The lawyers of the guy who's copyright you are violating will love it.) You get what you pay for. If your needs are simple and you are a small company without lawyers who worry a lot, goahead may work fine for you. If you may need help with your web page, or you want to keep your company's lawyers happy, buy a web server. The small fees and purchase price is not worth the hassles. Les --------------------------- Newsgroups: comp.os.vxworks Subject: Re: PowerPC VME read problem Date: Mon, 18 Jun 2001 11:12:28 -0600 From: Don Small Organization: Sandia National Laboratories Message-ID: <3B2E367C.D92FB2FC@sandia.gov> References: <9gkmsn$dqv$1@inn.jlab.org> I think that your problem has to do with the basic operation of the PowerPC. If you do not have a copy of the "PowerPC Microprocessor Family: The Programming Environments" or its current equivalent I would strongly suggest getting one. The following information comes from section 5.1.1 (Memory Access Ordering) and section 5.1.1.1 (Enforce In-Order Execution of I/O{I bet they had to think for quite some time to come up with a name whose acronym was EIEIO}) from this manual. In order to increase memory bandwidth, and all VME accesses are memory mapped IO, the PowerPC attempts to reorder IO in such a way as to reduce the number of accesses required to main memory. If you access byte 1, then 0, then 3, then 2 it will make this a longword (32bit) access to memory at address 0. You can override this behavior. An instruction called EIEIO can be used between each access to force the PowerPC to read/write the IO device before the next IO operation starts. In my BSP for the mvme2604 the BSP uses a function called sysInByte and sysOutByte which are located in sysALib.s. These instructions incorporate an EIEIO instruction into each memory access. If you use these functions instead of direct memory accesses it should clear up your problem. Chris Slominski wrote: > > VxWorks Group, > The news server an my work site has been down and I never saw my first > posting of this, or any responses. Sorry for the repeat. > > ------------------- > VxWorks Users, > I am porting code from a mv167 to a mv2700 PowerPC. My code reads > consecutive 8-bit quantities from aVME board that does not support D32 > acccess. I have a VME bus analyzer that I use to inspect the VME activity > of the software. When the software begin, the 8-bit reads are made as D8 > VME accesses, but very rapidly the hardware starts making D32 accesses of > the consecutive addresses instead. The D32 accesses cause the software to > fail. How do I force the hardware to not perform D32 accesses? > -------------------- > > I am going to try to modify the VDW bit of the Universe chip's LSI2_CTL > register to make 16-bit VME accesses the maximum. Does this solution sound > correct to anyone? > > Chris --------------------------- Newsgroups: comp.os.vxworks Subject: Wind Web Server 2.0, httpSymTabAdd function failure Date: 18 Jun 2001 10:16:10 -0700 From: emil.romascanu@us.xircom.com (Emil Romascanu) Organization: http://groups.google.com/ Message-ID: <2b1a2c27.0106180916.1ee5f6e3@posting.google.com> Has anybody experienced failures of httpSymTabAdd() calls ? I am running tutorial/step_07 and the server does not work properly because calls to that function return NULL. Thx, Emil Romascanu Software developer Intel Corporation --------------------------- Newsgroups: comp.os.vxworks Subject: Re: taskSpawn problem... Date: Mon, 18 Jun 2001 10:30:51 -0700 From: "Shesha Bhushan" Organization: Arizona State University Message-ID: <9gldrv$cc0$1@news.asu.edu> References: <9gdtdb$ebc$1@imsinews.kornet.net> <3B2E167C.A89CC163@sandia.gov> taskSpawn() is used to spawn the task from inside the code and u have a better control of it, since u can specify all the parameters. But sp is used to spawn a function as a task from the command line (ie, from shell). - -Shesha Bhushan "Don Small" wrote in message news:3B2E167C.A89CC163@sandia.gov... > > > "±èº´È£" wrote: > > > > 'taskSpawn' function doesn't work on my VxWorks. > > I tried it on the shell, like below. and have the exception error messages. > > what should I do to make 'taskSpawn' works ? > > > > ---------------------------------------------- > > -> taskSpawn() > > > > Exception number 4906: Task: 0x189720 () > > > > instruction translation miss > > program counter: 0x04 > > machine status : 0x9032 > > trcStack aborted: error in top frame (sp) > > ffffffff end +ffed846f: () > > value = 159704 = 0x185e88 > > -> > > > > ------------------------------------------------- > > > > taskSpawn require parameters, otherwise the parameters passed are all > 0 and it will attempt to start a task at address 0. The simplified > method > of starting a task from the command line is > > sp > > > > and I also write the code to try 'taskSpawn' and > > 'etherOutputHookAdd'. and download it to vxworks > > and tried ... It also doesn't work.... > > I think it's also because of that 'taskSpawn' problem. > > I'll be really pleased if i can get advice to solve that problems... > > can anyone see if there's any problem in the code ? > > > > --------------------------------------------------- > > #include > > #include > > > > void hook_test() > > { > > > > BOOL OutputHandle(struct ifnet *pIf, char *buffer, int length) > > { > > > > /* to see what is hooked*/ > > printf("hooked_packet = %c\n", *buffer); > > > > return (FALSE); > > } > > > > STATUS OutputHook() > > { > > if (etherOutputHookAdd(OutputHandle) == OK) { > > printf("etherOutputHookAdd successful\n"); > > return(OK); > > } > > else { > > printf("etherOutputHook failed\n"); > > return(ERROR); > > } > > } > > > > taskSpawn ("etherOutputHook", 20, 0, 2000, OutputHook, 0, 0, 0, 0, 0, 0, > > 0, 0, 0, 0); > > > > } > > -------------------------------------------------------------------------- - -- > > > When I compiled, loaded and tried to run this piece of code, it failed > also. I did not know that gcc allowed embedded functions within > functions but the compiler does not complain. But I did have a load > error. > > ld < test1.o > undefined symbol: ___trampoline > > which seems to be crucial for the execution of this code. I don't know > if you version of gcc/bsp creates code with this same unresolved > reference but any time you have an unresolved reference it needs to be > solved before you run the code. The loader normally resolves all > references while loading the code and unresolved references are left at > address 0. In this case it appears to be a function to help gcc call > functions inside of functions. When I execute hook_test, this is what I > get: > > -> hook_test > > Illegal Instruction > Program Counter: 0x00000008 > Status Register: 0x30004 > Task: 0x3d3fd98 "etherOutputHook" > > -> > > > Now, if I move two line as below: > > > BOOL OutputHandle(struct ifnet *pIf, char *buffer, int length) > { > > /* to see what is hooked*/ > printf("hooked_packet = %c\n", *buffer); > > return (FALSE); > } > > STATUS OutputHook() > { > if (etherOutputHookAdd(OutputHandle) == OK) { > printf("etherOutputHookAdd successful\n"); > return(OK); > } > else { > printf("etherOutputHook failed\n"); > return(ERROR); > } > } > > > void hook_test() > { > taskSpawn ("etherOutputHook", 20, 0, 2000, OutputHook, 0, 0, 0, 0, 0, > 0, > 0, 0, 0, 0); > > > } > > compile and load this works. > > Don --------------------------- Newsgroups: comp.os.vxworks Subject: Triggering Date: Mon, 18 Jun 2001 17:33:43 GMT From: no@way.com (Langston) Organization: SBC Internet Services Message-ID: <3b2e3b4f.262136078@news.pacbell.net> I'm trying to use WindView Triggering on my x86 target using Tornado2. I can run WindView in the normal mode but when I attempt to set up a trigger I get the error message "Target System Does not Support triggering". In my BSP I have the triggering library included (INCLUDE_TRIGGERING). Is there anything else I need enabled? Thanks In advance, Steve --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Running applications at boot time... Date: Mon, 18 Jun 2001 13:45:31 -0400 From: Jeremiah Frick Organization: Improv Systems, Inc. Message-ID: <3B2E3E3B.5E8EBE9F@improvsys.com> References: <3B2DF451.A176E1A0@improvsys.com> This is a multi-part message in MIME format. - --------------8EDED7ADD0BCEAF9DEBCBD45 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Kevin, Yep, it's usrAppInit(). Just have to add the function call. Here's another question for all you vxWorks people: I'm trying to build my bootable application, but I keep getting the following error: "undefined reference to `usrNetInit'" Now, I realize that this is included as a result of INCLUDE_NET_INIT being defined (which it is, in prjComps.h). However, I'm attempting to actually call that in my initialization routine, and it can't seem to find it. Any thoughts? The code will build fine and works great if I build it as a downloadable module. Thanks, Jeremy Kevin Livingston wrote: > One option is to use the start up script option. You can then specify a > start up script to run from the boot line I belive. The start up script > takes commands just like they were run from the shell(prompt). > > The other option is to modify your kernel to make the call to the code you > wish to run. If you look in the manuals under booting or startup sequences > you should find the approiate function to add your "hook" into. I don't > think it is UsrRoot, but it is something like that. if I am not mistaken > you modify usrAppInit() > > Kevin > > "Jeremiah Frick" wrote in message > news:3B2DF451.A176E1A0@improvsys.com... > > Hello all, > > > > I think there should be a relatively easy solution to this, but I > > figured I'd ask. > > > > I have some code that I'd like to run automatically immediately > > following VxWorks booting. Is there anywhere that I need to simply add > > the call to the application, or is there something else I have to do? > > Basically, I'm trying to eliminate any interaction required from the > > command prompt. > > > > Thanks, > > Jeremy > > - --------------8EDED7ADD0BCEAF9DEBCBD45 Content-Type: text/x-vcard; charset=us-ascii; name="jeremyf.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Jeremiah Frick Content-Disposition: attachment; filename="jeremyf.vcf" begin:vcard n:Frick;Jeremiah tel;work:716-756-5411 x45 x-mozilla-html:FALSE url:www.improvsys.com org:Improv Systems Inc.;PDG version:2.1 email;internet:jeremyf@improvsys.com title:Hardware Engineer adr;quoted-printable:;;3136 Winton Road South=0D=0ASuite 100;Rochester;New York;14623;United States of America fn:Jeremiah Frick end:vcard - --------------8EDED7ADD0BCEAF9DEBCBD45-- --------------------------- Newsgroups: comp.os.vxworks Subject: Exception handling? Date: Mon, 18 Jun 2001 18:53:15 +0100 From: vxworksintent@tao-group.com (David Given) Organization: Tao Group Ltd. Message-ID: Sender: dg@pearl.tao.co.uk We're trying to write an exception handler for VxWorks on the MIPS. What we want to do is be able to modify the processor context when an exception occurs, and then continue processing. For example: if a division by zero error occurs, we'd like to be able to adjust the register containing the value being divided by, and restart the instruction. We're having some problems. As far as we can see, there are two ways of doing this. If we use a signal with the SA_SIGINFO bit set when we register it, we get passed the signal context when the signal handler is called --- all well and geed, but some of the registers in the sigcontext are corrupted, in particular the exception status registers. We can't find out what sort of exception occurred, and where, and so on. The other way we've found is to use sigAddHook(). However, this seems to work at a very low level, and VxWorks seems to insist on suspending the task in which the exception occurred after the exception handler exits. We get garbage in the exception status registers, too, and random lockups. What's the official way of doing this? What's the difference between sigaction() and sigvec()? (The documentation is almost word-for-word identical.) Has anyone done this in the past, and have they encountered any problems? TIA... - -- David Given dg@tao-group.com --------------------------- Newsgroups: comp.os.vxworks Subject: Bufferless message queues Date: Mon, 18 Jun 2001 18:59:07 +0100 From: vxworksintent@tao-group.com (David Given) Organization: Tao Group Ltd. Message-ID: Sender: dg@pearl.tao.co.uk One of the things we need to do in my application is to send messages to a pool of tasks. If a task is ready to receive the message, it does so, if not we spawn a new task to process the message. Something like: task_main() { loop { receive message process message } } send(msg) { do { send message, nonblocking if send would block spawn task } while not sent } Normally I'd use a message queue for this. *However*, it seems that VxWorks won't let me use message queues with no buffering... it'll let me create one, but any attempt to send a message causes the send to fail. The best I can do is to create a message queue with a buffer of length one message, which is inadequate because I run the risk of the message being left in the queue with all tasks busy. I could implement my own message queue type thing with shared memory and semaphores, but it's fiddly and prone to bugs. All VxWorks pipes and message queues seem to be buffered. Is there an IPC system I've missed that's not buffered? TIA. - -- David Given dg@tao-group.com --------------------------- Newsgroups: comp.os.vxworks Subject: usrNetInit Date: Mon, 18 Jun 2001 14:45:00 -0400 From: Jeremiah Frick Organization: Improv Systems, Inc. Message-ID: <3B2E4C2C.D37891E0@improvsys.com> This is a multi-part message in MIME format. - --------------EBC5FDD36284270A0F489E16 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hello Everyone, Posted this a bit earlier under another topic, but I wanted to get it out in the open more. I'm trying to build my bootable application, but I keep getting the following error: "undefined reference to `usrNetInit'" Now, I realize that this should be included as a result of INCLUDE_NET_INIT being defined (which it is, in prjComps.h). However, I'm attempting to actually call that in my initialization routine, and it can't seem to find it. The actual body for usrNetInit is defined in usrNetwork.c. Any thoughts? The code will build fine and works great if I build it as a downloadable module. Thanks, Jeremy - --------------EBC5FDD36284270A0F489E16 Content-Type: text/x-vcard; charset=us-ascii; name="jeremyf.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Jeremiah Frick Content-Disposition: attachment; filename="jeremyf.vcf" begin:vcard n:Frick;Jeremiah tel;work:716-756-5411 x45 x-mozilla-html:FALSE url:www.improvsys.com org:Improv Systems Inc.;PDG version:2.1 email;internet:jeremyf@improvsys.com title:Hardware Engineer adr;quoted-printable:;;3136 Winton Road South=0D=0ASuite 100;Rochester;New York;14623;United States of America fn:Jeremiah Frick end:vcard - --------------EBC5FDD36284270A0F489E16-- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Q: Problem with sm network on mcp750 Date: Mon, 18 Jun 2001 15:06:35 -0500 From: Andrew Johnson Organization: APS, Argonne National Laboratory Message-ID: <3B2E5F4B.82D9F0BA@aps.anl.gov> References: Dafna wrote: > > As we increase the number of MPC860 CPUs booting over the SM > network (8 CPUs and more), we see some of the processors, sometimes > all, hang when searching for the boot file, or during booting itself. > Once this problem occurs, the CPUs that "got into trouble" never > recover. > Our research has shown that when things go wrong, the head of the > "free-packets-list" at the shared memory is overridden. From what > we've seen, this list is used by the 750 to receive packets from the > 860. > We can also see that the shared-memory area is initialized (ready > value at the ANCHOR is 0x87654321), and the heartbeat advances, > showing the system is alive. Write yourself some test code to exercise sysBusTas() and check that it really does work properly, as this seems like a classic symptom of more than one CPU believing that it has successfully claimed a shared area. I've attached my code as an example, but you'll probably have to change it before use. You might need to have the lock location in a special shared portion of memory, I can't help there as I don't know your bus architecture or the specifics of your BSP. - - Andrew - -- The world is such a cheerful place when viewed from upside-down It makes a rise of every fall, a smile of every frown /* tas.c - a soak test for sysBusTas by Andrew Johnson To use: compile and load this code on each vxWorks target (change the value of NUM_CPUS as required). Secondary CPUs don't have to use the sm network to boot, but all must have different processor numbers in their boot parameter settings. The TAS location can either be on the master (cpu 0) or some external location (currently this must be in the short VME address space). On the master, run initTas(0) from the vxWorks shell to use an onboard TAS location, or initTas(0x28cc) for example to use location A16:28CC. This will print a command line to be used subsequently to initialize the slave boards. Finally execute runTas on all CPUs simultaneously. Each cpu will print a line of counts every 65536 successful TAS operations, and all should be incrementing although the rates may vary between different CPUs. The test has failed if any CPU stops and diplays an error message, although other cpus may continue to run without error. */ #include #include #include #include #include #include #define NUM_CPUS 3 #if (CPU_FAMILY == PPC) /* This stuff may not all be necessary, but who knows? */ #define EIEIO_SYNC __asm__ volatile ("eieio;sync") void sysBusTasClear (volatile char *); #define TAS_CLEAR(loc) sysBusTasClear(loc) #else #define EIEIO_SYNC #define TAS_CLEAR(loc) *loc = 0 #endif LOCAL volatile struct tas_s { int lock; /* TAS location used when on master */ int space; /* Address space of the TAS location */ char *plock; /* VMEbus address of the TAS location */ int winner; /* Who won the TAS? */ int count[NUM_CPUS];/* Success counters for each CPU */ } master, *ptest; LOCAL char *plock; int initTas(void *pmaster) { int cpu = sysProcNumGet(); if (cpu == 0) { /* I'm the master CPU, create a struct tas_s */ int i; ptest = &master; /* ptest = cacheDmaMalloc(sizeof (struct tas_s)); if (ptest == NULL) { printf("cacheDmaMalloc returned NULL\n"); return -1; } */ ptest->lock = 0; /* 32-bit, if needed for PPC */ ptest->winner = 0; for (i=0; icount[i] = 0; if (pmaster == NULL) { /* Use ptest->lock as the TAS location */ plock = (char *) &ptest->lock; ptest->space = VME_AM_EXT_SUP_DATA; if (sysLocalToBusAdrs(ptest->space, plock, (char **)&ptest->plock)) { printf("first master sysLocalToBusAdrs(%#x, %p) failed\n", ptest->space, plock); return -1; } } else { /* TAS location is at short VMEbus address pmaster */ ptest->plock = (char *) pmaster; ptest->space = VME_AM_SUP_SHORT_IO; if (sysBusToLocalAdrs(ptest->space, ptest->plock, &plock)) { printf("master sysBusToLocalAdrs(%#x, %p) failed\n", ptest->space, ptest->plock); return -1; } } *plock = 0; /* 8-bit, off-board */ /* Convert ptest to a VMEbus address to print it */ if (sysLocalToBusAdrs(VME_AM_EXT_SUP_DATA, (char *) ptest, (char **) &pmaster)) { printf("second master sysLocalToBusAdrs(%#x, %p) failed\n", VME_AM_EXT_SUP_DATA, ptest); return -1; } printf("Master initialized, init slaves using\n"); printf("\tinitTas(%p)\nthen execute runTas() on all CPUs\n", pmaster); } else { /* Slave CPU, ptest is given by arg */ ptest = (struct tas_s *) pmaster; if (sysBusToLocalAdrs(VME_AM_EXT_SUP_DATA, (char *) pmaster, (char **) &ptest)) { printf("first slave sysBusToLocalAdrs(%#x, %p) failed\n", VME_AM_EXT_SUP_DATA, pmaster); return -1; } if (sysBusToLocalAdrs(ptest->space, ptest->plock, &plock)) { printf("second slave sysBusToLocalAdrs(%#x, %p) failed\n", ptest->space, ptest->plock); return -1; } printf("Slave initialized\n"); } EIEIO_SYNC; CACHE_DMA_FLUSH(ptest, sizeof (struct tas_s)); return 0; } LOCAL void doNothing(int i) { volatile int j=i; j++; } int runTas(void) { int cpu = sysProcNumGet(); ULONG startTick = tickGet(); ULONG startCount = ptest->count[cpu]; int tickRate = sysClkRateGet(); int i; for (;;) { while (!sysBusTas(plock)) ; /* We've won control now */ EIEIO_SYNC; CACHE_DMA_INVALIDATE(ptest, sizeof (struct tas_s)); if (ptest->winner != 0) { printf("Successful claim (cpu %d) pre-empted by cpu %d\n", cpu, ptest->winner - 1); return -1; } ptest->winner = cpu + 1; ptest->count[cpu]++; EIEIO_SYNC; CACHE_DMA_FLUSH(ptest, sizeof (struct tas_s)); if ((ptest->count[cpu] & 0xffff) == 0) { ULONG tickNow = tickGet(); ULONG countNow = ptest->count[cpu]; printf("counts = "); for (i=0; icount[i]); printf(" rate = %ld/second\n", (countNow - startCount) * tickRate / (tickNow - startTick)); startTick = tickNow; startCount = countNow; } for (i=0; i<10; i++) doNothing(i); EIEIO_SYNC; CACHE_DMA_INVALIDATE(ptest, sizeof (struct tas_s)); if (ptest->winner != cpu + 1) { printf("Successful claim (cpu %d) stolen by cpu %d\n", cpu, ptest->winner - 1); return -1; } ptest->winner = 0; EIEIO_SYNC; CACHE_DMA_FLUSH(ptest, sizeof (struct tas_s)); TAS_CLEAR(plock); EIEIO_SYNC; CACHE_DMA_FLUSH(ptest, sizeof (struct tas_s)); for (i=0; i<7; i++) doNothing(i); } return 0; } --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Bufferless message queues Date: Mon, 18 Jun 2001 16:22:29 -0400 From: Doug Dotson Organization: Verio Message-ID: <3B2E6305.1AF5A58@digidata.com> References: David, Why are you so adverse to buffered queues? Doug David Given wrote: > One of the things we need to do in my application is to send messages to a > pool of tasks. If a task is ready to receive the message, it does so, if > not we spawn a new task to process the message. Something like: > > task_main() > { > loop > { > receive message > process message > } > } > > send(msg) > { > do { > send message, nonblocking > if send would block > spawn task > } while not sent > } > > Normally I'd use a message queue for this. > > *However*, it seems that VxWorks won't let me use message queues with no > buffering... it'll let me create one, but any attempt to send a message > causes the send to fail. The best I can do is to create a message queue > with a buffer of length one message, which is inadequate because I run the > risk of the message being left in the queue with all tasks busy. > > I could implement my own message queue type thing with shared memory and > semaphores, but it's fiddly and prone to bugs. All VxWorks pipes and > message queues seem to be buffered. Is there an IPC system I've missed > that's not buffered? > > TIA. > > -- > David Given > dg@tao-group.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: how can i use 'taskSpawn' in vxworks 5.4 ? Date: Mon, 18 Jun 2001 16:26:49 -0400 From: Emory Stagmer Organization: Northrup Grumman, Litton Adv Sys, Space Sys Oper Message-ID: <3B2E6408.91EEA303@littonas.spamx.com> References: <9ghoj1$aqm$1@imsinews.kornet.net> taskSpawn from the shell prompt has the following format (example) - -> taskSpawn "TASK_NAME", 50, 0, 20000, task_code_main where 50 is the runtime priority, 20000 is the stack size and the zero is a parameter I never use and can't remember the definition of! Emory Stagmer SungJin Lee wrote: > hi vxworkers, > > When i try 'taskSpawn()' on the shell, it doesn't work > and shows the error message following. > > How can i use taskSpawn function in the vxworks 5.4 ? (with tornado 2) > > ----------------------------------------------------------------- > Exception number 4906: Task: 0x189720 () > > instruction translation miss > program counter: 0x04 > machine status : 0x9032 > trcStack aborted: error in top frame (sp) > ffffffff end +ffed846f: () > value = 159704 = 0x185e88 --------------------------- Newsgroups: comp.os.vxworks Subject: romInit in VxWorks 8260 BSP Date: 18 Jun 2001 13:42:54 -0700 From: sachin@blrmail.com (Sachin) Organization: http://groups.google.com/ Message-ID: Greetings PowerPCers, I am trying to decipher the init. code of 8260 in the Vxworks BSP & i find this in romInit function which is the initial entry point in the Flash space. ***************************************************************************** /* * When the PowerPC 8260 is powered on, the processor fetches the * instructions located at the address 0x100. We need to jump * from the address 0x100 to the Flash space. */ lis r4, HIADJ(start) /* load r4 with the address */ addi r4, r4, LO(start) /* of start */ lis r5, HIADJ(romInit) /* load r5 with the address */ addi r5, r5, LO(romInit) /* of romInit() */ lis r6, HIADJ(ROM_TEXT_ADRS) /* load r6 with the address */ addi r6, r6, LO(ROM_TEXT_ADRS) /* of ROM_TEXT_ADRS */ sub r4, r4, r5 /* */ add r4, r4, r6 mtspr LR, r4 /* save destination address*/ /* into LR register */ blr /* jump to flash mem address */ start: **************************************************************************** What do the comments signify & what is this piece of code doing?? To me, it seems like the code after initialising the IMMR to 0x4700_0000 and sundry other registers jumps from Flash space (from where this piece of code is running) BACK into Flash space, infact into the very next line (viz. the "start" label). Since the romInit function gets included only in the bootrom image & not in the downloadable image, the whole thing runs only in Flash space, does it not?? If so, isn't the whole thing useless exercise?? Thanks, Sachin. --------------------------- Newsgroups: comp.os.vxworks Subject: any good website to start learning VxWorks device driver development ? _EOM Date: Tue, 19 Jun 2001 01:49:55 GMT From: "not really" Organization: Road Runner Message-ID: <7dyX6.4392$P46.2620432@typhoon.ne.mediaone.net> --------------------------- Newsgroups: comp.os.vxworks Subject: Re: editor in vxworks Date: Tue, 19 Jun 2001 02:11:24 GMT From: Ephraim Gadsby Organization: - minimal Message-ID: <38dtit0bgqu3905nmbt1uh1ps2qnqbndm9@4ax.com> References: <9gbuk2$94f$1@overload.lbl.gov> On Thu, 14 Jun 2001 19:17:13 -0700, "Devatha, Pratap" wrote: also planning to write a text editor for this. >Does anyone know the address at which the video memory resides for the >Ixp1200. An IXP1200 does not have video memory. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Problem with sm network on mcp750 Date: Tue, 19 Jun 2001 10:14:24 +0800 From: "tanghaoyu" Organization: Bentium Ltd. Message-ID: <9gmcei$js4$1@mail.cn99.com> References: We have similar problem,our platform is CPX8216 + MCP750 + MCPN750. We contacted with Wind River support last year,but still no reply still. "Dafna" wrote in message news:ba3b0a8f.0106170702.1064b8da@posting.google.com... > Hi All, > We are experiencing some problems using the SM network with MCP750 > CPU as a master, and MCP860 CPUs (Artesyn Bajaspan Board) as "slave" > processors. > > Here's the configuration we are using: > MCP750 master CPU (#0) initializes the shared-memory area and sets > the ANCHOR address to 0x4100. (A constant value set in the 750 > config.h). At the same time, several MPC860 processors poll the SM > area on the 750 for their boot files. When the sm is initialized and > boot file is found, the mpc860 CPUs retreive it by FTP, and boot over > the SM network, which is activated over a cPci backplane. > During boot time, the 860 CPUs operate in poll mode, and once > booted they operate in Interrupt mode. The 750 CPU operates in > interrupt mode. > > > And the problem... > > As we increase the number of MPC860 CPUs booting over the SM > network (8 CPUs and more), we see some of the processors, sometimes > all, hang when searching for the boot file, or during booting itself. > Once this problem occurs, the CPUs that "got into trouble" never > recover. > Our research has shown that when things go wrong, the head of the > "free-packets-list" at the shared memory is overridden. From what > we've seen, this list is used by the 750 to receive packets from the > 860. > We can also see that the shared-memory area is initialized (ready > value at the ANCHOR is 0x87654321), and the heartbeat advances, > showing the system is alive. > > > Has anyone, by any chance, experienced a similar problem, or have used > the SM network in a similar configuration? > We have contacted the Wind River support, but with no luck so far. > > > Thanks in advance, > Dafna > --------------------------- Newsgroups: comp.os.vxworks Subject: Target Hung after Telnet Client Connecting to Date: Tue, 19 Jun 2001 09:19:43 +0800 From: "Antony Lo \(ATAN\)" Organization: Lawrence Berkeley National Laboratory Message-ID: <9gmcjj$7f6$1@overload.lbl.gov> Hi! I'm working on VxWorks V5.4 on Arm CPU. After I try to telnet to the target board, its console hang and the last info I saw is "telnetd: This system *IN USE* via telnet. shell restarted." Then I have to RESET the target. I want to provide both telnet client and console interface to users. Is there anything I miss ? Please see the info dumped to console as follows. Thanks, Antony ]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]] ]]]]]]]] ]] ]]]]]]]]]]]]] ]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]] ]]]]]]]]]]]]]]]]]]]]]]]]]]] V ] ]]]]]]]]] ]]]]]] ]]]]]]]] ]] ]]]] ]] ]]]]]]] ]]]]]]]] ]]]]]] ] ]] ]]]] ]]] ]]]]] ] ]]] ] ]]]] ]]] ]]]]]]]]] ]]]] ]] ]]]] ]] ]]]]] ]]]] ]]] ]] ] ]]] ]] ]]]]] ]]]]]] ]] ]]]]]]] ]]]] ]] ]]]] ]]]]] ] ]]]] ]]]]] ]]]]]]]] ]]]] ]] ]]]] ]]]]]]] ]]]] ]]]]]] ]]]]] ]]]]]] ] ]]]]] ]]]] ]] ]]]] ]]]]]]]] ]]]] ]]]]]]] ]]]]] ] ]]]]]] ] ]]] ]]]] ]] ]]]] ]]]] ]]]] ]]]] ]]]]]]]] ]]]]] ]]] ]]]]]]] ] ]]]]]]] ]]]] ]]]] ]]]] ]]]]] ]]]]]]]]]]]]]]]]]]]]]]]]]]]]]] ]]]]]]]]]]]]]]]]]]]]]]]]]]]]] Development System ]]]]]]]]]]]]]]]]]]]]]]]]]]]] ]]]]]]]]]]]]]]]]]]]]]]]]]]] VxWorks version 5.4 ]]]]]]]]]]]]]]]]]]]]]]]]]] KERNEL: WIND version 2.5 ]]]]]]]]]]]]]]]]]]]]]]]]] Copyright Wind River Systems, Inc., 1984-1999 CPU: KS32C50100 FOR SNDS100 Ver 1.0. Processor # 0. Memory Size: 0x400000. BSP version 1.2/0. WDB: Ready. - -> telnetd: someone tried to login. telnetd: This system *IN USE* via telnet. shell restarted. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: how can i use 'taskSpawn' in vxworks 5.4 ? Date: Mon, 18 Jun 2001 20:03:20 -0700 From: DrDiags Organization: laserlink.net Message-ID: <3B2EC0F8.6CC88E9A@covad.net> References: <9ghoj1$aqm$1@imsinews.kornet.net> <3B2E6408.91EEA303@littonas.spamx.com> Reply-To: drdiags@covad.net Emory, NAME taskSpawn( ) - spawn a task SYNOPSIS int taskSpawn ( char * name, /* name of new task (stored at pStackBase) */ int priority, /* priority of new task */ int options, /* task option word */ int stackSize, /* size (bytes) of stack needed plus name */ FUNCPTR entryPt, /* entry point of new task */ int arg1, /* 1st of 10 req'd task args to pass to func */ int arg2, int arg3, int arg4, int arg5, int arg6, int arg7, int arg8, int arg9, int arg10 ) Bits in the options argument may be set to run with the following modes: VX_FP_TASK (0x0008) execute with floating-point coprocessor support. VX_PRIVATE_ENV (0x0080) include private environment support (see envLib). VX_NO_STACK_FILL (0x0100) do not fill the stack for use by checkStack( ). VX_UNBREAKABLE (0x0002) do not allow breakpoint debugging. See the definitions in taskLib.h. Emory Stagmer wrote: > > taskSpawn from the shell prompt has the following format (example) > > -> taskSpawn "TASK_NAME", 50, 0, 20000, task_code_main > > where 50 is the runtime priority, 20000 is the stack size and > the zero is a parameter I never use and can't remember the definition of! > > Emory Stagmer > > SungJin Lee wrote: > > > hi vxworkers, > > > > When i try 'taskSpawn()' on the shell, it doesn't work > > and shows the error message following. > > > > How can i use taskSpawn function in the vxworks 5.4 ? (with tornado 2) > > > > ----------------------------------------------------------------- > > Exception number 4906: Task: 0x189720 () > > > > instruction translation miss > > program counter: 0x04 > > machine status : 0x9032 > > trcStack aborted: error in top frame (sp) > > ffffffff end +ffed846f: () > > value = 159704 = 0x185e88 --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Triggering Date: Mon, 18 Jun 2001 20:16:10 -0700 From: DrDiags Organization: laserlink.net Message-ID: <3B2EC3FA.9F9D7B70@covad.net> References: <3b2e3b4f.262136078@news.pacbell.net> Reply-To: drdiags@covad.net Langston, I not sure, but in the Project facility, the parameters the triggering library shows that the INCLUDE_TRIGGER_SHOW Macro is required. At least this if for T2/vxWorks 5.4 Langston wrote: > > I'm trying to use WindView Triggering on my x86 target using Tornado2. > > I can run WindView in the normal mode but when I attempt to set up a > trigger I > get the error message "Target System Does not Support triggering". > > In my BSP I have the triggering library included (INCLUDE_TRIGGERING). > Is there anything else I need enabled? > > Thanks In advance, Steve --------------------------- Newsgroups: comp.os.vxworks Subject: Re: module download from command prompt Date: Mon, 18 Jun 2001 20:33:47 -0700 From: DrDiags Organization: laserlink.net Message-ID: <3B2EC81B.B63AB592@covad.net> References: <9g8afb$c526@news.research.bell-labs.com> <3B282D64.D7CE9590@support.nl> <9gl82c$66q2@news.research.bell-labs.com> Reply-To: drdiags@covad.net Daniel, You could create a Tcl script that would allow you to connect to the Target server, load your modules and exit. Or you could create a DOS executable using the APIs documented in the Tornado API Programmer's Guide (though some have voiced concerns on how well these are explained, you be the judge) and Visual C++ for example. Just a couple of suggestions. I gather this is not part of your final design, just a tool to help you manage development cycles, without having to use the Tornado IDE. You can also fire up the GUI-less (widget-less?) version of the windSh executable. Daniel G Waddington wrote: > > I meant from the DOS prompt. > > DAN > > "Kees van der Bent" wrote in message > news:3B282D64.D7CE9590@support.nl... > > Maybe you have specified the problem not clear enough... But reading > > your question I would initially say using: > > > > -> ld < module-(full-)path-name > > > > This seems to simple and it's all in the manual... > > > > > > > > Daniel G Waddington wrote: > > > > > Hi, anyone know how to download modules from the command line, using > Tornado > > > II? > > > > > > Thanks > > > > > > DAN > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Timing problems on VxWorks x486 Date: Tue, 19 Jun 2001 05:38:45 GMT From: Ephraim Gadsby Organization: - minimal Message-ID: References: <9giahq$3qe$1@overload.lbl.gov> On Sun, 17 Jun 2001 13:28:00 +0100, "David Laight" wrote: >> I seem to be having problems with the timings of a piece of code I >> have written. > >> It linearly searches a set of data (triangles for the closest to a >> point) for varying numbers of triangles. I have plotted the time per >> search against number of triangles, which I expected to be linear >> (bar caching) yet it looks like this; >> x >> x >> x x >> x x >> x x >> x x >> x x >> x > >> which I really don't understand - Anybody any ideas???? > >Seems to be precisely what I would expect! > >1st line - data pulled from level 1 cache >2nd line - data from level 2 cache >3rd line - data from main memory > I don't get that. It's time taken for the search v's number of thing to be searched. Why would the time taken drop as you exaust each type of cache? --------------------------- Newsgroups: comp.os.vxworks Subject: the fAQ Date: Tue, 19 Jun 2001 07:45:38 GMT From: Xuejun Cai Organization: Lucent Technologies, Columbus, Ohio Message-ID: <20010619.7453800@blrc.china.bell-labs.com> All, Where can find the FAQ of this group? I searched the old archive and=20 found a link to http://www.pso.com/vxfaq.html. However, it seems that th= e=20 FAQ can't be found from this link. Any other places? Regards, Cai, Xuejun --------------------------- Newsgroups: comp.os.vxworks Subject: Gnu compiler for VxWorks powerpc version 2.95.3 instead of the wi Date: Tue, 19 Jun 2001 09:41:24 +0200 From: "Chantreux, Guy Felix" Organization: Lawrence Berkeley National Laboratory Message-ID: <9gn1mj$jb5$1@overload.lbl.gov> We need to use this version of compiler for c++ design reason. I download the sources and configure it, and begin to compile it. But before doing more job, I need to know : - - if I have to do some changes - - there is some specific target option as targ=powerpc-eabi or targ=powerpc-wrs in configure - - it's better to use the old way used by wind river in version 5.3 with the gnu toolchain - - if I use the configure, do I need additional option - - will I have some problem with the windriver object libraries at run-time Thanks in advance, Guy --------------------------- Newsgroups: comp.os.vxworks Subject: vxworks porting Date: 19 Jun 2001 01:20:12 -0700 From: skyheave@kornet.net (jo, Kang) Organization: http://groups.google.com/ Message-ID: <9fb6474c.0106190020.1faf577f@posting.google.com> Hi, I'm engineer of the Future Systems Corp Korea. I'm working this!!!--> hardware porting, vxworks porting and so on... I'm hope so, --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Bufferless message queues Date: Tue, 19 Jun 2001 10:17:55 +0100 From: vxworksintent@tao-group.com (David Given) Organization: I do not speak for anyone but myself, and barely that. Message-ID: <3c5ng9.j5q.ln@127.0.0.1> References: <3B2E6305.1AF5A58@digidata.com> Sender: dg@pearl.tao.co.uk In article <3B2E6305.1AF5A58@digidata.com>, Doug Dotson writes: > Why are you so adverse to buffered queues? Because then I get a situation where the message has been successfully sent, but hasn't been received yet because it's in the buffer. When I send the message, I need to know whether it's been delivered successfully, or if there's nothing ready to receive it (in which case I need to start a new handler process). I'm really looking for the same behaviour as in traditional Unix pipes. - -- David Given dg@tao-group.com --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Tue Jun 19 05:33:02 2001 From: K Sanjay Date: Tue Jun 19 05:33:04 PDT 2001 Subject: Stack overflow Friends, We are using VxWorks5.4. While testing my module, I was getting memPartAlloc() error or memPartFree error in some other tasks. Finally I found out that this was because the size of the local variables in my task is more than the allocated stack size. How do I findout the stack overflow? Another question, I am adding adding additional memory to the memory pool after loading the module. The code looks like this, if (sysPhysMemTop() > sysMemTop()) memAddToPool (sysMemTop(), (sysPhysMemTop()-sysMemTop()); After this if I try to execute sysMemTop on windshell it will crash giving exception. I could not figure out the reason for this. I would appreciate your help/suggestion regarding these problems. Thanks, Sanjay __________________________________________________ Do You Yahoo!? Spot the hottest trends in music, movies, and more. http://buzz.yahoo.com/ From vxwexplo-errs@csg.lbl.gov Tue Jun 19 05:48:12 2001 From: Ilia Bosis Date: Tue Jun 19 05:48:14 PDT 2001 Subject: RE: Target Hung after Telnet Client Connecting to Welcome to the club :) You faced the well-known feature(problem) of the VxWorks shell: There is AT MOST ONE shell console at any given time !!! So when You telnet to the target, it kills the serial console shell to give the shell to telnet. When You close the telnet session (say, by issuing the 'logout' command from telnet) the serial console shell will be alive again. Try to telnet to the same target from two machines and the second attempt will be rejected with the message that say: "target engaged by ". What to do ? First, think if You really want to give the full power of VxWorks shell to users ? If what You want is some application-specific CLI with pre-defined set of commands, employ You own CLI. Recently there has been a discussion on the subject in this mailing list and os.comp.vxworks newsgroup. Look at the archives. If You still want a number of the VxWorks shells simultaneously (and are willing to deal with the consequences) ask Your VxWorks FAE. AFAIK, they have something called "multi-user telnet" or "multi-user shell". That may help You. Or may not. I'm not familar with the beast at all. Hope, this helps Regards Ilia -------------------------------------------- Ilia Bosis Software Engineer PacketLight Networks Tel: +972-9-7644228 Fax: +972-9-7645444 E-mail: Ilia_Bosis@packetlight.com Web: www.packetlight.com > -----Original Message----- > From: vxwexplo@lbl.gov [mailto:vxwexplo@lbl.gov] > Sent: Tuesday, June 19, 2001 3:32 AM > To: vxworks_users@csg.lbl.gov > Subject: Target Hung after Telnet Client Connecting to > > > Submitted-by vxwexplo-errs@csg.lbl.gov Mon Jun 18 18:23:00 2001 > Submitted-by: "Antony Lo \(ATAN\)" > > Hi! > I'm working on VxWorks V5.4 on Arm CPU. After I try to > telnet to the target board, its console hang and > the last info I saw is "telnetd: This system *IN USE* via > telnet. shell restarted." Then I have to RESET the target. > I want to provide both telnet client and console > interface to users. Is there anything I miss ? Please see > the info dumped to console as follows. > > Thanks, > Antony > > > ]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]] ]]]]]]]] ]] > ]]]]]]]]]]]]] > ]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]] > ]]]]]]]]]]]]]]]]]]]]]]]]]]] V > ] ]]]]]]]]] ]]]]]] ]]]]]]]] ]] ]]]] > ]] ]]]]]]] ]]]]]]]] ]]]]]] ] ]] ]]]] > ]]] ]]]]] ] ]]] ] ]]]] ]]] ]]]]]]]]] ]]]] ]] > ]]]] ]] ]]]]] > ]]]] ]]] ]] ] ]]] ]] ]]]]] ]]]]]] ]] ]]]]]]] > ]]]] ]] ]]]] > ]]]]] ] ]]]] ]]]]] ]]]]]]]] ]]]] ]] ]]]] > ]]]]]]] ]]]] > ]]]]]] ]]]]] ]]]]]] ] ]]]]] ]]]] ]] ]]]] > ]]]]]]]] ]]]] > ]]]]]]] ]]]]] ] ]]]]]] ] ]]] ]]]] ]] ]]]] > ]]]] ]]]] ]]]] > ]]]]]]]] ]]]]] ]]] ]]]]]]] ] ]]]]]]] ]]]] > ]]]] ]]]] ]]]]] > ]]]]]]]]]]]]]]]]]]]]]]]]]]]]]] > ]]]]]]]]]]]]]]]]]]]]]]]]]]]]] Development System > ]]]]]]]]]]]]]]]]]]]]]]]]]]]] > ]]]]]]]]]]]]]]]]]]]]]]]]]]] VxWorks version 5.4 > ]]]]]]]]]]]]]]]]]]]]]]]]]] KERNEL: WIND version 2.5 > ]]]]]]]]]]]]]]]]]]]]]]]]] Copyright Wind River > Systems, Inc., 1984-1999 > > CPU: KS32C50100 FOR SNDS100 > Ver 1.0. Processor # 0. > Memory Size: 0x400000. BSP > version 1.2/0. > WDB: Ready. > -> telnetd: someone tried to login. > > telnetd: This system *IN USE* via telnet. > shell restarted. From vxwexplo-errs@csg.lbl.gov Tue Jun 19 10:23:05 2001 From: Joe Carissimi Date: Tue Jun 19 10:23:07 PDT 2001 Subject: EHOSTDOWN error message Does anyone know which layer generates the EHOSTDOWN error id? We are running Tornado II on a DY4 SVM179 SBC, using UDP sockets. We are "blindcasting" messages from one SBC to other SBCs. When we attempt to send a message to a non-powered target, this errno (0x43) is returned from the sendto() routine. It occurs about 5 seconds after the first message is sent. I can snoop the network and see ARP requests for the IP (with no reply of course). Is there a layer which can be excluded or a parameter which can be adjusted to remove this unnecessary error response? Can't seem to find anything in the documentation (or WindSurf) as per a solution. Note that this is not a problem in our VxWorks 5.2 system. -- Joe Carissimi BBN Technologies Middletown, RI 02842 (401) 848-3525 From vxwexplo-errs@csg.lbl.gov Tue Jun 19 17:22:08 2001 From: "Jun Wu" Date: Tue Jun 19 17:22:10 PDT 2001 Subject: strong ARM pointer align Hi all, I use intel strongARM SA1110 assabet board and Tornado 2.0. I find if a pointer is not 4-byte align, data may be corrupted. For example void foo() { unsigned int *p; /* set p point with a non-4-byte align address */ p = (int *) 0x08000002; *p = 0xffffffff; /* bytes with address 0x08000000 and 0x08000001 are corrupted!*/ } Are there any compiler options can avoid this problem ? Best regard? From vxwexplo-errs@csg.lbl.gov Wed Jun 20 04:03:16 2001 From: Vxworks Exploder Date: Wed Jun 20 04:03:19 PDT 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Wed Jun 20 04:03:06 PDT 2001 Subject: Re: Bufferless message queues Subject: Re: Target Hung after Telnet Client Connecting to Subject: Re: Gnu compiler for VxWorks powerpc version 2.95.3 instead of the wi Subject: Re: Problem with sm network on mcp750 Subject: Re: Bufferless message queues Subject: Re: Bufferless message queues Subject: Write to SCSI disk Subject: Re: TrueFFS INstallation Problem Subject: Stack overflow Subject: RE: Target Hung after Telnet Client Connecting to Subject: Re: Bufferless message queues Subject: StethoScope & TraceScope Subject: Re: any good website to start learning VxWorks device driver development ? _EOM Subject: Re: the fAQ Subject: Re: VxWorks web servers & server pages technologies Subject: Re: romInit in VxWorks 8260 BSP Subject: Memory protection and Loadmodule Subject: Re: TrueFFS INstallation Problem Subject: global objects in VxWorks Subject: Re: Memory protection and Loadmodule Subject: WindView 2.0.1 Customizing Default User Event Format Subject: priority inheritance problem Subject: EHOSTDOWN error message Subject: stl Subject: Newbie question: Including spyLib in application code (or project) Subject: Re: Memory protection and Loadmodule Subject: Re: stl Subject: Re: TrueFFS INstallation Problem Subject: How many END_TBL_ENTRY (entries) needed for 3 PHYs (each with 2 ports) Subject: Re: How common is SNiFF usage? Subject: Re: TrueFFS INstallation Problem Subject: Re: romInit in VxWorks 8260 BSP Subject: Re: romInit in VxWorks 8260 BSP Subject: Motorola 5101 Board Support package Subject: strong ARM pointer align Subject: Re: Newbie question: Including spyLib in application code (or project) Subject: Re: WindView 2.0.1 Customizing Default User Event Format Subject: Re: Newbie question: Including spyLib in application code (or project) Subject: How to get VxSim to execute commands from script file automatically after it is started with vxWorks -p ? Subject: Re: Stack overflow Subject: Re: How common is SNiFF usage? Subject: Re: Problem with sm network on mcp750 Subject: Tornado2 dependencies Subject: Re: EHOSTDOWN error message Subject: Re: Memory protection and Loadmodule ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Bufferless message queues Date: Tue, 19 Jun 2001 13:25:37 +0200 From: Urban Lindberg Organization: Ericsson Message-ID: <3B2F36B1.D6DA392E@ks.ericsson.se> References: <3B2E6305.1AF5A58@digidata.com> <3c5ng9.j5q.ln@127.0.0.1> David Given wrote: >I'm really looking for the same behaviour as in > traditional Unix pipes. > > -- > David Given > dg@tao-group.com Don't know much about unix pipes, but if you want to know if a task is waiting for a message on a message queue, have a look on msgQInfoGet ($WIND_BASE/docs/vxworks/ref/msgQShow.html#msgQInfoGet). Beware of the warnings though. HTH Urban --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Target Hung after Telnet Client Connecting to Date: Tue, 19 Jun 2001 13:00:54 +0100 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: <94HX6.25$c25.75283@newsr2.u-net.net> References: <9gmcjj$7f6$1@overload.lbl.gov> Antony Lo (ATAN) wrote in message <9gmcjj$7f6$1@overload.lbl.gov>... >Hi! > I'm working on VxWorks V5.4 on Arm CPU. After I try to telnet to the target board, its console hang and >the last info I saw is "telnetd: This system *IN USE* via telnet. shell restarted." Then I have to RESET the target. > I want to provide both telnet client and console interface to users. Is there anything I miss ? Yes, there are three things you need to know: - You can only have one shell task running at a time, so you can either have shell at the console, or shell in telnet, but not both. - The target does not hang when you telnet to it; what actually happens is that the shell inputs and outputs are redirected to the telnet connection. - There is no need to reset the target; once you close the telnet connection, the shell will return to the console. DaveK - -- They laughed at Galileo. They laughed at Copernicus. They laughed at Columbus. But remember, they also laughed at Bozo the Clown. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Gnu compiler for VxWorks powerpc version 2.95.3 instead of the wi Date: Tue, 19 Jun 2001 12:55:29 +0100 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: <94HX6.24$c25.75161@newsr2.u-net.net> References: <9gn1mj$jb5$1@overload.lbl.gov> Chantreux, Guy Felix wrote in message <9gn1mj$jb5$1@overload.lbl.gov>... >We need to use this version of compiler for c++ design reason. >I download the sources and configure it, and begin to compile it. > >But before doing more job, I need to know : >- if I have to do some changes >- there is some specific target option as targ=powerpc-eabi or >targ=powerpc-wrs in configure >- it's better to use the old way used by wind river in version 5.3 with the >gnu toolchain >- if I use the configure, do I need additional option >- will I have some problem with the windriver object libraries at run-time I've done this job already. There are some patches that need to be made to the code, there are some options that need to be used to configure it, there are complications with the WRS object libraries, target = powerpc-wrs-vxworks, and you can find out all the information you need including build procedures, patches, documentation, source code and pre-compiled binary installations at http://newgcc4vxworks4ppc.cjb.net/ hope it helps, DaveK -- They laughed at Galileo. They laughed at Copernicus. They laughed at Columbus. But remember, they also laughed at Bozo the Clown. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Problem with sm network on mcp750 Date: 19 Jun 2001 05:39:49 -0700 From: ellin_lin@263.net (ellin) Organization: http://groups.google.com/ Message-ID: <44a01eba.0106190439.75377191@posting.google.com> References: <9gmcei$js4$1@mail.cn99.com> I think you should pay attention to the limit of the count of none-system processor that mcp750 supports with sm network, and it's defined in the config.h as below. ... # define MAX_MCPN750_COUNT 7 ... You can increase the number,and try again. Speak just for myself. ellin --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Bufferless message queues Date: Tue, 19 Jun 2001 08:56:21 -0400 From: Doug Dotson Organization: Verio Message-ID: <3B2F4BF5.C0C6652F@digidata.com> References: <3B2E6305.1AF5A58@digidata.com> <3c5ng9.j5q.ln@127.0.0.1> It sounds to me like a queue is not what you are after because queuing behavior is keeping you from getting what to want. Perhaps a global variable visable to be producer and consumer whould be better with synchronization controlled by a semaphore. Doug David Given wrote: > In article <3B2E6305.1AF5A58@digidata.com>, > Doug Dotson writes: > > Why are you so adverse to buffered queues? > > Because then I get a situation where the message has been successfully > sent, but hasn't been received yet because it's in the buffer. When I send > the message, I need to know whether it's been delivered successfully, or > if there's nothing ready to receive it (in which case I need to start a > new handler process). I'm really looking for the same behaviour as in > traditional Unix pipes. > > -- > David Given > dg@tao-group.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Bufferless message queues Date: Tue, 19 Jun 2001 08:59:04 -0400 From: Doug Dotson Organization: Verio Message-ID: <3B2F4C98.87EF25AC@digidata.com> References: <3B2E6305.1AF5A58@digidata.com> <3c5ng9.j5q.ln@127.0.0.1> Unfortunately vxWorks pipes are built on top of message queues. How about using sockets through localhost. Will be a bit slower though. Doug David Given wrote: > In article <3B2E6305.1AF5A58@digidata.com>, > Doug Dotson writes: > > Why are you so adverse to buffered queues? > > Because then I get a situation where the message has been successfully > sent, but hasn't been received yet because it's in the buffer. When I send > the message, I need to know whether it's been delivered successfully, or > if there's nothing ready to receive it (in which case I need to start a > new handler process). I'm really looking for the same behaviour as in > traditional Unix pipes. > > -- > David Given > dg@tao-group.com --------------------------- Newsgroups: comp.os.vxworks Subject: Write to SCSI disk Date: 19 Jun 2001 05:57:37 -0700 From: lilyko2002@yahoo.com (lily) Organization: http://groups.google.com/ Message-ID: <7f208981.0106190457.4d20edb6@posting.google.com> I am using SCSI disk and DosFs 2.0 After call to dosFsVolFormat debug output is : " fordisk cache error: device mat failed, err3ecf7bcno= block 1 errno 0x37001a, disk write failed, data loss may have occurred." (From scsi2Lib.h errno 0x37001a = S_scsiLib_UNKNOWN_SENSE_DATA) What is error? Any ideas on what I am missing? Maybe somebody knows what happens? Thanks for any suggestions, Lily --------------------------- Newsgroups: comp.os.vxworks Subject: Re: TrueFFS INstallation Problem Date: 19 Jun 2001 06:08:52 -0700 From: aparker@vtlinx.com (parreg) Organization: http://groups.google.com/ Message-ID: <79ef8b60.0106190508.ac4a589@posting.google.com> References: <3B28E7E2.A56DA2E@digidata.com> <3B2A3C54.A6640BA5@digidata.com> <79ef8b60.0106180524.58de98ca@posting.google.com> <3B2E1B1D.D4D1BE1D@digidata.com> Doug Dotson wrote in message news:<3B2E1B1D.D4D1BE1D@digidata.com>... > I don't have a function sysTffsFormat. Don't I have to run tffsDevCreate first? > When I run usrTffsConfig I get the same error. I put some printf's in > cfiscsIdentify() and it never gets called. I checked noOfMTDs and it is 1 > as it should be. > > Doug ################# You won't get anywhere until the flash is formated. You must do sysTffsFormat or tffsDevFormat. After this can create the device usrTffsConfig ultimately calls tffsDevCreate, but this will only work after the flash has been formatted into Erasble Units. What core are you using? > > parreg wrote: > > > Doug Dotson wrote in message news:<3B2A3C54.A6640BA5@digidata.com>... > > > Yes. And when I load it I take a look at noOfMTDs and it is 1 > > > indicating that one entry is in the MTD table. > > > > > > Doug > > > > > > > > > Andray Kaganovsky wrote: > > > > > > > Doug Dotson wrote in > > > > news:3B28E7E2.A56DA2E@digidata.com: > > > > > > > > > Greetings! > > > > > > > > > > I am having alot of trouble getting TFFS installed. Followed the > > > > > instructions > > > > > and got a successful build. After rebooting, I attempt to create a > > > > > device > > > > > using "tffsDevCreate" and get the following: > > > > > > > > > > -> tffsDevCreate > > > > > > > > > > Data bus error > > > > > Exception Program Counter: 0x800a60cc > > > > > Status Register: 0x3000ff01 > > > > > Cause Register: 0x0000001c > > > > > Error address: 0xffffffff, Error ID: 0x0000 > > > > > > > > > > 80041894 vxTaskEntry +c : shell (1, 0, 0, 0) > > > > > 800bef98 shell +1c8: 800befbc (eeeeeeee, eeeeeeee, eeeeeeee, > > > > > eeeeeeee) > > > > > 800bf1a4 shell +3d4: execute (800fe4c8, 8010b304, 2048, > > > > > &yypact) > > > > > 800bf30c execute +c8 : yyparse (80c77d88, ffffffff, 1, ffffffff) > > > > > 800ef040 yyparse +888: 800ed400 (eeeeeeee, 0, eeeeeeee, 0) > > > > > 800ed5c4 yystart +89c: tffsDevCreate (140, 258, 80077aec, 0) > > > > > 800a1a58 tffsDevCreate +4c : flCall (80c77cc0, 0, 80cfe920, 96) > > > > > 800a73bc flCall +248: 800a6540 (0, 0, 8007829c, 0) > > > > > 800a6588 tffscmpWords +270: flMount (80124c60, 800a6454, 0, 0) > > > > > 800f2050 flMount +30 : flIdentifyFlash (0, 0, 80c77b80, 0) > > > > > 800a617c flIdentifyFlash+38 : 800a609c (800ec5e8, 800ec590, 800ea708, > > > > > 800e990c) > > > > > > > > > > My specific configuration is: > > > > > > > > > > 1) Board is a Galileo EV-64240 with resident Intel E28F640J3A-120 > > > > > flash. > > > > > 2) I modified mv177-sysTtfs.c file to suit. > > > > > 3) I'm using the cfiscs.c MTD. > > > > > > > > > > I put in a few debug statements and have found that the rfaRegister() > > > > > function is being > > > > > called. I also put some debug statements in the flRegisterCFISCS() and > > > > > cfiscsIdentfy() > > > > > routines and they are never called. Since the exception is occurring at > > > > > 0xffffffff it appears > > > > > to me that the function pointers to the MTD functions are never being > > > > > set up but I can't > > > > > figure out how this is supposed to be done. > > > > > > > > > > Thanks, > > > > > Doug > > > > > > > > > > > > > > > Douglas P. Dotson > > > > > Digi-Data Corporation > > > > > 8580 Dorsey Run Rd. > > > > > Jessup, MD 20794 > > > > > 301-498-0200 x248 > > > > > > > > > > > > > > > > > > > > > > > [AndrayK]. Did you #defined INCLUDE_MTD_CFISCS in sysTffs.c ? > > > > you should format the flash first using sysTffsFormat() then > > use usrTffsConfig 0,0,"/flash" to create the device. Then type "devs" > > confirming the device is there. Once the device is created you can > > check the dosFs volumne configuration using > > "dosFsConFigShow("/flash"). > > > > parreg --------------------------- Newsgroups: comp.os.vxworks Subject: Stack overflow Date: Tue, 19 Jun 2001 05:33:00 -0700 (PDT) From: K Sanjay Organization: Lawrence Berkeley National Laboratory Message-ID: <9gnj97$7li$1@overload.lbl.gov> Friends, We are using VxWorks5.4. While testing my module, I was getting memPartAlloc() error or memPartFree error in some other tasks. Finally I found out that this was because the size of the local variables in my task is more than the allocated stack size. How do I findout the stack overflow? Another question, I am adding adding additional memory to the memory pool after loading the module. The code looks like this, if (sysPhysMemTop() > sysMemTop()) memAddToPool (sysMemTop(), (sysPhysMemTop()-sysMemTop()); After this if I try to execute sysMemTop on windshell it will crash giving exception. I could not figure out the reason for this. I would appreciate your help/suggestion regarding these problems. Thanks, Sanjay __________________________________________________ Do You Yahoo!? Spot the hottest trends in music, movies, and more. http://buzz.yahoo.com/ --------------------------- Newsgroups: comp.os.vxworks Subject: RE: Target Hung after Telnet Client Connecting to Date: Tue, 19 Jun 2001 15:46:28 +0200 From: Ilia Bosis Organization: Lawrence Berkeley National Laboratory Message-ID: <9gnjd7$7ll$1@overload.lbl.gov> Welcome to the club :) You faced the well-known feature(problem) of the VxWorks shell: There is AT MOST ONE shell console at any given time !!! So when You telnet to the target, it kills the serial console shell to give the shell to telnet. When You close the telnet session (say, by issuing the 'logout' command from telnet) the serial console shell will be alive again. Try to telnet to the same target from two machines and the second attempt will be rejected with the message that say: "target engaged by ". What to do ? First, think if You really want to give the full power of VxWorks shell to users ? If what You want is some application-specific CLI with pre-defined set of commands, employ You own CLI. Recently there has been a discussion on the subject in this mailing list and os.comp.vxworks newsgroup. Look at the archives. If You still want a number of the VxWorks shells simultaneously (and are willing to deal with the consequences) ask Your VxWorks FAE. AFAIK, they have something called "multi-user telnet" or "multi-user shell". That may help You. Or may not. I'm not familar with the beast at all. Hope, this helps Regards Ilia - -------------------------------------------- Ilia Bosis Software Engineer PacketLight Networks Tel: +972-9-7644228 Fax: +972-9-7645444 E-mail: Ilia_Bosis@packetlight.com Web: www.packetlight.com > -----Original Message----- > From: vxwexplo@lbl.gov [mailto:vxwexplo@lbl.gov] > Sent: Tuesday, June 19, 2001 3:32 AM > To: vxworks_users@csg.lbl.gov > Subject: Target Hung after Telnet Client Connecting to > > > Submitted-by vxwexplo-errs@csg.lbl.gov Mon Jun 18 18:23:00 2001 > Submitted-by: "Antony Lo \(ATAN\)" > > Hi! > I'm working on VxWorks V5.4 on Arm CPU. After I try to > telnet to the target board, its console hang and > the last info I saw is "telnetd: This system *IN USE* via > telnet. shell restarted." Then I have to RESET the target. > I want to provide both telnet client and console > interface to users. Is there anything I miss ? Please see > the info dumped to console as follows. > > Thanks, > Antony > > > ]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]] ]]]]]]]] ]] > ]]]]]]]]]]]]] > ]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]] > ]]]]]]]]]]]]]]]]]]]]]]]]]]] V > ] ]]]]]]]]] ]]]]]] ]]]]]]]] ]] ]]]] > ]] ]]]]]]] ]]]]]]]] ]]]]]] ] ]] ]]]] > ]]] ]]]]] ] ]]] ] ]]]] ]]] ]]]]]]]]] ]]]] ]] > ]]]] ]] ]]]]] > ]]]] ]]] ]] ] ]]] ]] ]]]]] ]]]]]] ]] ]]]]]]] > ]]]] ]] ]]]] > ]]]]] ] ]]]] ]]]]] ]]]]]]]] ]]]] ]] ]]]] > ]]]]]]] ]]]] > ]]]]]] ]]]]] ]]]]]] ] ]]]]] ]]]] ]] ]]]] > ]]]]]]]] ]]]] > ]]]]]]] ]]]]] ] ]]]]]] ] ]]] ]]]] ]] ]]]] > ]]]] ]]]] ]]]] > ]]]]]]]] ]]]]] ]]] ]]]]]]] ] ]]]]]]] ]]]] > ]]]] ]]]] ]]]]] > ]]]]]]]]]]]]]]]]]]]]]]]]]]]]]] > ]]]]]]]]]]]]]]]]]]]]]]]]]]]]] Development System > ]]]]]]]]]]]]]]]]]]]]]]]]]]]] > ]]]]]]]]]]]]]]]]]]]]]]]]]]] VxWorks version 5.4 > ]]]]]]]]]]]]]]]]]]]]]]]]]] KERNEL: WIND version 2.5 > ]]]]]]]]]]]]]]]]]]]]]]]]] Copyright Wind River > Systems, Inc., 1984-1999 > > CPU: KS32C50100 FOR SNDS100 > Ver 1.0. Processor # 0. > Memory Size: 0x400000. BSP > version 1.2/0. > WDB: Ready. > -> telnetd: someone tried to login. > > telnetd: This system *IN USE* via telnet. > shell restarted. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Bufferless message queues Date: Tue, 19 Jun 2001 12:51:45 +0100 From: vxworksintent@tao-group.com (David Given) Organization: I do not speak for anyone but myself, and barely that. Message-ID: References: <3B2E6305.1AF5A58@digidata.com> <3c5ng9.j5q.ln@127.0.0.1> <3B2F36B1.D6DA392E@ks.ericsson.se> Sender: dg@pearl.tao.co.uk In article <3B2F36B1.D6DA392E@ks.ericsson.se>, Urban Lindberg writes: > Don't know much about unix pipes, but if you want to know if a task is > waiting for a message on a message queue, have a look on msgQInfoGet > ($WIND_BASE/docs/vxworks/ref/msgQShow.html#msgQInfoGet). Beware of the > warnings though. Thanks; that seems to have done the trick... { lock mutex no processes block on the queue? create new process send message unlock mutex } As reader processes are pretty light-weight and self-destruct if they aren't used for a certain amount of time, I don't really have to worry about inconsistent information, as the mutex means I'll only send one message at a time. Cheers. - -- David Given dg@tao-group.com --------------------------- Newsgroups: comp.os.vxworks Subject: StethoScope & TraceScope Date: 19 Jun 2001 07:39:08 -0700 From: joewoodbury@mail.com (Joe Woodbury) Organization: http://groups.google.com/ Message-ID: Has anyone on this forum used either of these tools? If so what was your impression? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: any good website to start learning VxWorks device driver development ? _EOM Date: 19 Jun 2001 07:53:31 -0700 From: aparker@vtlinx.com (parreg) Organization: http://groups.google.com/ Message-ID: <79ef8b60.0106190653.39d5a7c6@posting.google.com> References: <7dyX6.4392$P46.2620432@typhoon.ne.mediaone.net> "not really" wrote in message news:<7dyX6.4392$P46.2620432@typhoon.ne.mediaone.net>... Your best bet is WindRiver's Tornado BSP Developers Kit for VxWorks User's Guide. It's clear and there are plenty of examples. parreg --------------------------- Newsgroups: comp.os.vxworks Subject: Re: the fAQ Date: 19 Jun 2001 08:17:09 -0700 From: aparker@vtlinx.com (parreg) Organization: http://groups.google.com/ Message-ID: <79ef8b60.0106190717.1155d8f@posting.google.com> References: <20010619.7453800@blrc.china.bell-labs.com> Xuejun Cai wrote in message news:<20010619.7453800@blrc.china.bell-labs.com>... > All, > > Where can find the FAQ of this group? I searched the old archive and > found a link to http://www.pso.com/vxfaq.html. However, it seems that th > e > FAQ can't be found from this link. Any other places? > > Regards, > > Cai, Xuejun you can try: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html parreg --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks web servers & server pages technologies Date: 19 Jun 2001 08:44:51 -0700 From: joewoodbury@mail.com (Joe Woodbury) Organization: http://groups.google.com/ Message-ID: References: <9ggg24$18s$1@bcarh8ab.ca.nortel.com> <1c30b9ff.0106180619.7928564b@posting.google.com> <3B2E33E0.5C3408B3@agilent.com> What is the performance hit on the target machine using a web server such as goahead? > Eran Gluska wrote: > > > Try www.goahead.com - they have a free web server already ported for VxWorks. > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: romInit in VxWorks 8260 BSP Date: 19 Jun 2001 11:45:21 -0400 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: References: Sender: bpringlemeir@DeadDuck >>>>> "Sachin" == Sachin writes: [snip] Sachin> What do the comments signify & what is this piece of code Sachin> doing?? To me, it seems like the code after initialising the Sachin> IMMR to 0x4700_0000 and sundry other registers jumps from Sachin> Flash space (from where this piece of code is running) BACK Sachin> into Flash space, infact into the very next line (viz. the Sachin> "start" label). Since the romInit function gets included only Sachin> in the bootrom image & not in the downloadable image, the Sachin> whole thing runs only in Flash space, does it not?? If so, Sachin> isn't the whole thing useless exercise?? Thanks, Sachin. Immediate values are loaded in two instruction on PPC. Code is like, r4 = start; r5 = romInit; r6 = ROM_TEXT_ADRS; r4 -= romInit; /* r5 */ r4 += ROM_TEXT_ADRS; /* r6 */ (int(*)())r4; /* jump to address in r4 */. What is actually happening is that the `C' symbols `start' and `romInit' have a logical value that is different from where the code is actually executing from when it boots. This is the value `ROM_TEXT_ADRS'. The delta between `start' and `romInit' is calculated. This delta is then added to ROM_TEXT_ADRS which is where `romInit' is actually located at this point. This is a common idiom on all processor models that I know of. On occasion, the start address can be select between `0' or `0xffff0000' or something like that. However, the address that the processor boots from is hard coded. The code is linked to run at a different address, so care must be taken until it jumps to the link address later on. hth, Bill Pringlemeir. - -- The Kibo Bible is not emotions. Useful vxWorks URLS, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" "http://www.xs4all.nl/~borkhuis/vxworks/vxw_pt9.html" --------------------------- Newsgroups: comp.os.vxworks Subject: Memory protection and Loadmodule Date: Tue, 19 Jun 2001 11:59:49 -0400 From: Beginner Organization: UUNET Canada News Reader Service Message-ID: <3B2F76F5.5517B48F@hotmail.com> Hi, All, I have a problem that might have something to do with memory protection. I am using the new to create a class objection, such as MyObject *myobject - new MyObject; If I pass the pointer myobject to a executable that are loaded using loadModule() and VxWorks crashed each time a member in my object is referenced by this executable. Does anyone know why? Malloc seems to be ok. TIA and cheers. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: TrueFFS INstallation Problem Date: Tue, 19 Jun 2001 11:52:47 -0400 From: Doug Dotson Organization: Verio Message-ID: <3B2F754F.388FCDB9@digidata.com> References: <3B28E7E2.A56DA2E@digidata.com> <3B2A3C54.A6640BA5@digidata.com> <79ef8b60.0106180524.58de98ca@posting.google.com> <3B2E1B1D.D4D1BE1D@digidata.com> <79ef8b60.0106190508.ac4a589@posting.google.com> It is when I try to run tffsDevFormat that it blows up. It is never trying to call the cfiscsIdentify function to obtain the programming parameters for the flash device. I'm not sure what you mean by "core". Processor core? RM7000. Here is the dump. As you can see, tffsDevFormat ultimately tries to identify the flash device be calling tffsIdentifyFlash. This function is supposed to go through the list of MTDs to determine which kind of flash it is working with. In my case the only MTD registered is the CFI/SCS MTD si the cfiscsIdentify function should be called. It isn't being called. Perhaps it has a problem when only a single MTD is in the table. I'll try adding a couple more Data bus error Exception Program Counter: 0x800f115c Status Register: 0x3000ff01 Cause Register: 0x0000001c Error address: 0xffffffff, Error ID: 0x0000 800419a4 vxTaskEntry +c : shell (1, 0, 0, 0) 800bda48 shell +1c8: 800bda6c (eeeeeeee, eeeeeeee, eeeeeeee, eeeeeeee) 800bdc54 shell +3d4: execute (800fd698, 8010a4c4, 7f, 10) 800bddbc execute +c8 : yyparse (8010a520, 0, 1, ffffffff) 800edaf0 yyparse +888: 800ebeb0 (eeeeeeee, 0, eeeeeeee, 0) 800ec074 yystart +89c: tffsDevFormat (140, 258, 80077bfc, eeeeeeee) 800a2028 tffsDevFormat +b0 : flCall (800eab18, eeeeeeee, eeeeeeee, eeeeeeee) 800a5e98 flCall +274: 800a523c (80c77bf8, 0, ffffffff, 5000000) 800a52dc tffscmpWords +514: flMount (80c77cd4, 96, 80123e20, 800a4f04) 800f0b00 flMount +30 : flIdentifyFlash (800f3310, 0, 0, a0d24444) 800f120c flIdentifyFlash+38 : 800f112c (&yyval, 3, 3, 0) Thanks, Doug parreg wrote: > Doug Dotson wrote in message news:<3B2E1B1D.D4D1BE1D@digidata.com>... > > I don't have a function sysTffsFormat. Don't I have to run tffsDevCreate first? > > When I run usrTffsConfig I get the same error. I put some printf's in > > cfiscsIdentify() and it never gets called. I checked noOfMTDs and it is 1 > > as it should be. > > > > Doug > > ################# > You won't get anywhere until the flash is formated. You must do > sysTffsFormat or tffsDevFormat. After this can create the device > usrTffsConfig ultimately calls tffsDevCreate, but this will only work > after the flash has been formatted into Erasble Units. What core are > you using? > > > > parreg wrote: > > > > > Doug Dotson wrote in message news:<3B2A3C54.A6640BA5@digidata.com>... > > > > Yes. And when I load it I take a look at noOfMTDs and it is 1 > > > > indicating that one entry is in the MTD table. > > > > > > > > Doug > > > > > > > > > > > > Andray Kaganovsky wrote: > > > > > > > > > Doug Dotson wrote in > > > > > news:3B28E7E2.A56DA2E@digidata.com: > > > > > > > > > > > Greetings! > > > > > > > > > > > > I am having alot of trouble getting TFFS installed. Followed the > > > > > > instructions > > > > > > and got a successful build. After rebooting, I attempt to create a > > > > > > device > > > > > > using "tffsDevCreate" and get the following: > > > > > > > > > > > > -> tffsDevCreate > > > > > > > > > > > > Data bus error > > > > > > Exception Program Counter: 0x800a60cc > > > > > > Status Register: 0x3000ff01 > > > > > > Cause Register: 0x0000001c > > > > > > Error address: 0xffffffff, Error ID: 0x0000 > > > > > > > > > > > > 80041894 vxTaskEntry +c : shell (1, 0, 0, 0) > > > > > > 800bef98 shell +1c8: 800befbc (eeeeeeee, eeeeeeee, eeeeeeee, > > > > > > eeeeeeee) > > > > > > 800bf1a4 shell +3d4: execute (800fe4c8, 8010b304, 2048, > > > > > > &yypact) > > > > > > 800bf30c execute +c8 : yyparse (80c77d88, ffffffff, 1, ffffffff) > > > > > > 800ef040 yyparse +888: 800ed400 (eeeeeeee, 0, eeeeeeee, 0) > > > > > > 800ed5c4 yystart +89c: tffsDevCreate (140, 258, 80077aec, 0) > > > > > > 800a1a58 tffsDevCreate +4c : flCall (80c77cc0, 0, 80cfe920, 96) > > > > > > 800a73bc flCall +248: 800a6540 (0, 0, 8007829c, 0) > > > > > > 800a6588 tffscmpWords +270: flMount (80124c60, 800a6454, 0, 0) > > > > > > 800f2050 flMount +30 : flIdentifyFlash (0, 0, 80c77b80, 0) > > > > > > 800a617c flIdentifyFlash+38 : 800a609c (800ec5e8, 800ec590, 800ea708, > > > > > > 800e990c) > > > > > > > > > > > > My specific configuration is: > > > > > > > > > > > > 1) Board is a Galileo EV-64240 with resident Intel E28F640J3A-120 > > > > > > flash. > > > > > > 2) I modified mv177-sysTtfs.c file to suit. > > > > > > 3) I'm using the cfiscs.c MTD. > > > > > > > > > > > > I put in a few debug statements and have found that the rfaRegister() > > > > > > function is being > > > > > > called. I also put some debug statements in the flRegisterCFISCS() and > > > > > > cfiscsIdentfy() > > > > > > routines and they are never called. Since the exception is occurring at > > > > > > 0xffffffff it appears > > > > > > to me that the function pointers to the MTD functions are never being > > > > > > set up but I can't > > > > > > figure out how this is supposed to be done. > > > > > > > > > > > > Thanks, > > > > > > Doug > > > > > > > > > > > > > > > > > > Douglas P. Dotson > > > > > > Digi-Data Corporation > > > > > > 8580 Dorsey Run Rd. > > > > > > Jessup, MD 20794 > > > > > > 301-498-0200 x248 > > > > > > > > > > > > > > > > > > > > > > > > > > > > [AndrayK]. Did you #defined INCLUDE_MTD_CFISCS in sysTffs.c ? > > > > > > you should format the flash first using sysTffsFormat() then > > > use usrTffsConfig 0,0,"/flash" to create the device. Then type "devs" > > > confirming the device is there. Once the device is created you can > > > check the dosFs volumne configuration using > > > "dosFsConFigShow("/flash"). > > > > > > parreg --------------------------- Newsgroups: gnu.gcc.help,gnu.g++.help,comp.os.vxworks Subject: global objects in VxWorks Date: Tue, 19 Jun 2001 17:48:29 +0200 From: Max Fischer Organization: DLR Message-ID: <3B2F744D.7BF1613A@dlr.de> Hello, we are using a gcc-2.95.2 for powerpc-wrs-vxworks and would like to use global C++ objects in one relinkable object in another one, e.g. testLib.cc: class A { public: A(){} int i; }; A myGlobalA; testDrv.cc: ... extern A myGlobalA; void testit ( void ){ int i = myGlobalA.i; cout << i << endl; } When downloading those two relinkable objects to our v5.4 ppc target, we get Undefined symbols: myGlobalA nmppc output of testLib: 00000064 T _GLOBAL_$I$myGlobalA 00000000 W __1A 00000000 t __static_initialization_and_destruction_0 00000000 D _ctors 00000008 D _dtors 00000000 t gcc2_compiled. 00000000 t gcc2_compiled. 00000000 S myGlobalA This happens with all symbols of type 'S' or 'B' (nmppc output). The loader seems to ignore them. Could anybody give me a hint how we could convince our compiler/linker/loader not to ignore these symbols (except not using global objects ;-)??? Thanks Max - -- ______________________________________________________________________ | | | | Dr. rer. nat. Max Fischer | Tel: +49-8153/28-1183 | | DLR Oberpfaffenhofen | Fax: +49-8153/28-1134 | | Institute of Robotics and Mechatronics | E-mail: /| | Postfach 1116 | Max.Fischer@dlr.de --+-+--- | D-82230 Wessling | / | | / |_____________________________________________________________ ---+-+-- |/ --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Memory protection and Loadmodule Date: Tue, 19 Jun 2001 17:21:14 +0100 From: Simon Farnsworth Message-ID: <3B2F7BFA.7020509@snellwilcox.com> References: <3B2F76F5.5517B48F@hotmail.com> Beginner wrote: > Hi, All, > > I have a problem that might have something to do with memory protection. Which versions of VxWorks and Tornado are you using? Unless you are using VxWorks AE, there is no memory protection. > I am using the new to create a class objection, such as > MyObject *myobject - new MyObject; Try MyObject *myobject = new MyObject; > If I pass the pointer myobject to a executable that are loaded using > loadModule() and VxWorks crashed each time a member in my object is > referenced by this executable. Does anyone know why? Is the C++ runtime support compiled into your VxWorks OS image? Do you get a stack backtrace for the failure; if so, please post it. --------------------------- Newsgroups: comp.os.vxworks Subject: WindView 2.0.1 Customizing Default User Event Format Date: 19 Jun 2001 10:14:02 -0700 From: adufour@akara.com (Allan Dufour) Organization: http://groups.google.com/ Message-ID: <78700411.0106190914.62f78462@posting.google.com> We would like to modify the default format for User Events. There appears to be an obivous place to do this within the "WindView.win32.tcl" file in the procedure "eventFormat". Having modified this procedure, we still see no change in the format? The current default is to display the raw hex data associated with the event. Any ideas? Are we in the right place? --------------------------- Newsgroups: comp.os.vxworks Subject: priority inheritance problem Date: 19 Jun 2001 10:51:54 -0700 From: steve.ross@gvg.tv (Steve) Organization: http://groups.google.com/ Message-ID: Hello, I'm using Tornado2 on an x86 target and I'm seeing application tasks getting raised to the same priority as the network task (netTask). As I understand it VxWorks will do this to avoid priority inversion. What I don't understand is why an application task would be holding a resource that netTask would need. It seems that my application task is holding a mutex that the netTask requires. A consequence is that after a short period of time in this condition I see "Ring Buffer Overflow" messages. Has anybody seen this kind of behavior? Any suggestions on fixing this? thanks, Steve --------------------------- Newsgroups: comp.os.vxworks Subject: EHOSTDOWN error message Date: Tue, 19 Jun 2001 13:19:08 -0400 From: Joe Carissimi Organization: Lawrence Berkeley National Laboratory Message-ID: <9go4ri$bh8$1@overload.lbl.gov> Does anyone know which layer generates the EHOSTDOWN error id? We are running Tornado II on a DY4 SVM179 SBC, using UDP sockets. We are "blindcasting" messages from one SBC to other SBCs. When we attempt to send a message to a non-powered target, this errno (0x43) is returned from the sendto() routine. It occurs about 5 seconds after the first message is sent. I can snoop the network and see ARP requests for the IP (with no reply of course). Is there a layer which can be excluded or a parameter which can be adjusted to remove this unnecessary error response? Can't seem to find anything in the documentation (or WindSurf) as per a solution. Note that this is not a problem in our VxWorks 5.2 system. - -- Joe Carissimi BBN Technologies Middletown, RI 02842 (401) 848-3525 --------------------------- Newsgroups: comp.os.vxworks Subject: stl Date: Tue, 19 Jun 2001 14:36:12 -0400 From: "Kevin Livingston" Message-ID: Are there any good implementations of STL available for use with VxWorks/Tornado. I am guessing since Tornado uses the GNU compilier that the GNU STL would be good, unless it is too skewed for Linux or something. Any opinions on the matter would be greatly appreciated. Kevin --------------------------- Newsgroups: comp.os.vxworks Subject: Newbie question: Including spyLib in application code (or project) Date: 19 Jun 2001 11:44:09 -0700 From: c_campise@hotmail.com (Chris) Organization: http://groups.google.com/ Message-ID: I need to include the spyLib utility in order to profile some of my application code. Now, being new to Tornado, I was curious as to where in my project I add the configuration macro (INCLUDE_SPY)? According to the Tornado documentation, if the INCLUDE_SPY macro is defined, then the spyLib functionality will be available. I've tried both under the "C/C++ Compiler" tab (by adding: -DINCLUDE_SPY) and under the "Macros" tab for the Properties:Build Specification. And when I tried to call any of the spyLib routine (either via the shell, or from my code), I received the following: 'spy not configured into this system' Any help would be greatly appreciated. Thanks, Christopher Campise --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Memory protection and Loadmodule Date: Tue, 19 Jun 2001 15:18:29 -0400 From: Beginner Organization: UUNET Canada News Reader Service Message-ID: <3B2FA585.90907A05@hotmail.com> References: <3B2F76F5.5517B48F@hotmail.com> <3B2F7BFA.7020509@snellwilcox.com> Hi, Simon, Thanks. "MyObject *myobject - new MyObject", the - was a typo. I finally find out why, there was a bug on my object that were overwritten, but the time it got to the exec, it was already corrupted. Thanks again. And by the way, how do you do a stack backtrace? Cheers! Simon Farnsworth wrote: > Beginner wrote: > > > Hi, All, > > > > I have a problem that might have something to do with memory protection. > > Which versions of VxWorks and Tornado are you using? Unless you are > using VxWorks AE, there is no memory protection. > > > I am using the new to create a class objection, such as > > MyObject *myobject - new MyObject; > > Try MyObject *myobject = new MyObject; > > > If I pass the pointer myobject to a executable that are loaded using > > loadModule() and VxWorks crashed each time a member in my object is > > referenced by this executable. Does anyone know why? > > Is the C++ runtime support compiled into your VxWorks OS image? Do you > get a stack backtrace for the failure; if so, please post it. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: stl Date: Tue, 19 Jun 2001 15:15:12 -0400 From: "Kevin Livingston" Message-ID: References: ok, lets just forget I asked that... it's bundeled in with Tornado and all that. ALTHOUGH, I am courious if their stl stands up to the standard. Has anyone had any trouble porting it or anything to other implementations of stl etc.? Or found instances where it disagrees with the standard? (definition) it's been a long week Kevin "Kevin Livingston" wrote in message news:TVMX6.27$TQ4.1303@client... > Are there any good implementations of STL available for use with > VxWorks/Tornado. > > I am guessing since Tornado uses the GNU compilier that the GNU STL would be > good, unless it is too skewed for Linux or something. > > Any opinions on the matter would be greatly appreciated. > > Kevin > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: TrueFFS INstallation Problem Date: Tue, 19 Jun 2001 20:14:04 GMT From: andreyk@home.com (Andray Kaganovsky) Organization: @Home Message-ID: References: <3B28E7E2.A56DA2E@digidata.com> <3B2A3C54.A6640BA5@digidata.com> <79ef8b60.0106180524.58de98ca@posting.google.com> <3B2E1B1D.D4D1BE1D@digidata.com> <79ef8b60.0106190508.ac4a589@posting.google.com> <3B2F754F.388FCDB9@digidata.com> Doug Dotson wrote in news:3B2F754F.388FCDB9 @digidata.com: > It is when I try to run tffsDevFormat that it blows up. It is never trying > to call the cfiscsIdentify function to obtain the programming parameters > for the flash device. I'm not sure what you mean by "core". Processor > core? RM7000. > > Here is the dump. As you can see, tffsDevFormat ultimately tries to > identify the flash device be calling tffsIdentifyFlash. This function is > supposed to go through the list of MTDs to determine which kind > of flash it is working with. In my case the only MTD registered is the > CFI/SCS MTD si the cfiscsIdentify function should be called. It > isn't being called. Perhaps it has a problem when only a single MTD > is in the table. I'll try adding a couple more > > Data bus error > Exception Program Counter: 0x800f115c > Status Register: 0x3000ff01 > Cause Register: 0x0000001c > Error address: 0xffffffff, Error ID: 0x0000 > > 800419a4 vxTaskEntry +c : shell (1, 0, 0, 0) > 800bda48 shell +1c8: 800bda6c (eeeeeeee, eeeeeeee, eeeeeeee, eeeeeeee) > 800bdc54 shell +3d4: execute (800fd698, 8010a4c4, 7f, 10) > 800bddbc execute +c8 : yyparse (8010a520, 0, 1, ffffffff) > 800edaf0 yyparse +888: 800ebeb0 (eeeeeeee, 0, eeeeeeee, 0) > 800ec074 yystart +89c: tffsDevFormat (140, 258, 80077bfc, eeeeeeee) > 800a2028 tffsDevFormat +b0 : flCall (800eab18, eeeeeeee, eeeeeeee, eeeeeeee) > 800a5e98 flCall +274: 800a523c (80c77bf8, 0, ffffffff, 5000000) > 800a52dc tffscmpWords +514: flMount (80c77cd4, 96, 80123e20, 800a4f04) > 800f0b00 flMount +30 : flIdentifyFlash (800f3310, 0, 0, a0d24444) > 800f120c flIdentifyFlash+38 : 800f112c (&yyval, 3, 3, 0) > > Thanks, > Doug > > > parreg wrote: > >> Doug Dotson wrote in message news: <3B2E1B1D.D4D1BE1D@digidata.com>... >> > I don't have a function sysTffsFormat. Don't I have to run tffsDevCreate first? >> > When I run usrTffsConfig I get the same error. I put some printf's in >> > cfiscsIdentify() and it never gets called. I checked noOfMTDs and it is 1 >> > as it should be. >> > >> > Doug >> >> ################# >> You won't get anywhere until the flash is formated. You must do >> sysTffsFormat or tffsDevFormat. After this can create the device >> usrTffsConfig ultimately calls tffsDevCreate, but this will only work >> after the flash has been formatted into Erasble Units. What core are you using? >> > >> > parreg wrote: >> > >> > > Doug Dotson wrote in message news: <3B2A3C54.A6640BA5@digidata.com>... >> > > > Yes. And when I load it I take a look at noOfMTDs and it is 1 >> > > > indicating that one entry is in the MTD table. >> > > > >> > > > Doug >> > > > >> > > > >> > > > Andray Kaganovsky wrote: >> > > > >> > > > > Doug Dotson wrote in >> > > > > news:3B28E7E2.A56DA2E@digidata.com: >> > > > > >> > > > > > Greetings! >> > > > > > >> > > > > > I am having alot of trouble getting TFFS installed. Followed the instructions >> > > > > > and got a successful build. After rebooting, I attempt to create a device >> > > > > > using "tffsDevCreate" and get the following: >> > > > > > >> > > > > > -> tffsDevCreate >> > > > > > >> > > > > > Data bus error >> > > > > > Exception Program Counter: 0x800a60cc >> > > > > > Status Register: 0x3000ff01 >> > > > > > Cause Register: 0x0000001c >> > > > > > Error address: 0xffffffff, Error ID: 0x0000 >> > > > > > >> > > > > > 80041894 vxTaskEntry +c : shell (1, 0, 0, 0) >> > > > > > 800bef98 shell +1c8: 800befbc (eeeeeeee, eeeeeeee, eeeeeeee, eeeeeeee) >> > > > > > 800bf1a4 shell +3d4: execute (800fe4c8, 8010b304, 2048, &yypact) >> > > > > > 800bf30c execute +c8 : yyparse (80c77d88, ffffffff, 1, ffffffff) >> > > > > > 800ef040 yyparse +888: 800ed400 (eeeeeeee, 0, eeeeeeee, 0) >> > > > > > 800ed5c4 yystart +89c: tffsDevCreate (140, 258, 80077aec, 0) >> > > > > > 800a1a58 tffsDevCreate +4c : flCall (80c77cc0, 0, 80cfe920, 96) >> > > > > > 800a73bc flCall +248: 800a6540 (0, 0, 8007829c, 0) >> > > > > > 800a6588 tffscmpWords +270: flMount (80124c60, 800a6454, 0, 0) >> > > > > > 800f2050 flMount +30 : flIdentifyFlash (0, 0, 80c77b80, 0) >> > > > > > 800a617c flIdentifyFlash+38 : 800a609c (800ec5e8, 800ec590, 800ea708, 800e990c) >> > > > > > >> > > > > > My specific configuration is: >> > > > > > >> > > > > > 1) Board is a Galileo EV-64240 with resident Intel E28F640J3A- 120 flash. >> > > > > > 2) I modified mv177-sysTtfs.c file to suit. >> > > > > > 3) I'm using the cfiscs.c MTD. >> > > > > > >> > > > > > I put in a few debug statements and have found that the rfaRegister() >> > > > > > function is being >> > > > > > called. I also put some debug statements in the flRegisterCFISCS () and cfiscsIdentfy() >> > > > > > routines and they are never called. Since the exception is occurring at >> > > > > > 0xffffffff it appears >> > > > > > to me that the function pointers to the MTD functions are never being >> > > > > > set up but I can't >> > > > > > figure out how this is supposed to be done. >> > > > > > >> > > > > > Thanks, >> > > > > > Doug >> > > > > > >> > > > > > >> > > > > > Douglas P. Dotson >> > > > > > Digi-Data Corporation >> > > > > > 8580 Dorsey Run Rd. >> > > > > > Jessup, MD 20794 >> > > > > > 301-498-0200 x248 >> > > > > > >> > > > > > >> > > > > > >> > > > > >> > > > > [AndrayK]. Did you #defined INCLUDE_MTD_CFISCS in sysTffs.c ? >> > > >> > > you should format the flash first using sysTffsFormat() then >> > > use usrTffsConfig 0,0,"/flash" to create the device. Then type "devs" >> > > confirming the device is there. Once the device is created you can >> > > check the dosFs volumne configuration using >> > > "dosFsConFigShow("/flash"). >> > > >> > > parreg > > [AndrayK]. Could you please set breakpoint on flIdentifyFlash(), and run till this breakpoint is hit. At that point please set breakpoint on the following routines: flNeedVpp flMap flDelayLoop flDontNeedVpp and continue. All four routines above are called from isRAM() - the small static routine located right before the routine flIdentifyFlash(). The isRAM() is called from flIdentifyFlash() to check if memory region that socket's window has been mapped to, doesn't look like RAM. The dump above seems to indicate that isRAM() is where the CPU exception comes from. This also explains why MTD isn't called - the exception occures right before TrueFFS attempts all the MTD registered in the MTD list. I also suggest to thorougly check the socket component that you are installing in sysTffs.c. Make sure that it is installed properly, global variable 'noOfDrives' is equal to 1 at the moment you are entering flIdentifyFlash(), and first arguement to flIdentifyFlash() (which is pointer to struct FLSocket) contains meaningfull data. Please also verify that memory region that your flash is mapped to, is defined as writable in BSP's memory descriptor table. Andray --------------------------- Newsgroups: comp.os.vxworks Subject: How many END_TBL_ENTRY (entries) needed for 3 PHYs (each with 2 ports) Date: Tue, 19 Jun 2001 16:09:48 -0400 From: Babu Yama Organization: Lucent Technologies Message-ID: <3B2FB18C.9E91892C@lucent.com> I am not clear on the endDevTbl->unit field? Can some guru through some light on this, please? END_TBL_ENTRY endDevTbl[] = { 0, MOT_FCC_LOAD_FUNC, MOT_FCC_LOAD_STRING, 1, NULL, FALSE}; What does the endDevTbl->unit(the first field) refer to? Does it refer to one PHY (physical layer ethernet device) or one PORT on a PHY? How to set up the endDevTbl for multiple PHYs (with some PHYs having multiple ports)? In other words, if I have 3 PHYS (each PHY having 2 ports), how many entries in endDevTbl do I need? In this situation, how many devices are there under MUX and how many drivers? Should I have one driver instance per PHY or have one driver that handles 3 PHYs and all their ports. I guess, for the END drivers, the concept of driver and device is not clear to me. I understand that muxDevLoad() is called once for each of the entries in endDevTbl, which results in driver's xxload call. If a PHY has multiple ports, how many times xxload is called by DEFAULT by vxworks ( I am using wind river's est8260 bsp code)? Do I need to make any vxworks code changes to supports multiple ports on a PHY or multiple PHYs? Thanks a lot, Babu --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How common is SNiFF usage? Date: Tue, 19 Jun 2001 22:51:22 +0200 From: "Martin Raabe" Organization: T-Online Message-ID: <9goe18$frm$05$1@news.t-online.com> References: <9gijt8$abo$03$1@news.t-online.com> Hello Joe, thank you for expressing your impressions so detailed. I will take your input back to the factory. My short statement to this is as following: What SNiFF+ is: - SNiFF+ is a SourceCode Management system, which handles SourceCode of different languages and Platforms. - SNiFF+ is a SDE (Software Development Environment) which allows to develop independent from the target platform. - SNiFF+ supports in these rols also Team Development, which is very helpfull in small and medium sized (1-100 developers) teams. - SNiFF+ is an easy to handle Tool which behaves the same on Unix, and Windows. - SNiFF+ allows to faster get into foreign or (own) older SourceCode. - SNiFF+ helps to live project strategies, individuals would not follow without SNiFF+. What SNiFF+ is NOT: - SNiFF+ is not a debugging tool for OOP-Languages, where runtime behavior is a significant part of the debugging process. - SNiFF+ is not a Project Manager.It only adapts to VersionControl and Project Management Systems. I have a question for you. Which SNiFF+ Version are you referring to? I refer to the latest Version 4.0.1. I can understand that you do not feel that comfortable with the 2.x or 3.0.x Versions. But the feedback from customers on Verison 3.2.1 and 4.0.x is excellent! So I am afraid SNiFf+ will make it into a huge project basis accross branches and markets! Ciao Martin Raabe "Joe Woodbury" schrieb im Newsbeitrag news:HS8X6.652$ZR1.297147@news.uswest.net... > There is no direct connection between SNiFF and team development. SNiFF is > an object browser, not a team collaboration tool. SNiFF may facilitate large > project develpment, which is often done in a team environment, but it > completely lacks the other collaboration tools more commonly associated with > a team development environment. > > As far as particular criticisms of SNiFF go; my distaste of object browsers > is that they don't really clarify anything. You end up with massive lists of > functions and classes (and data and function members), none of which give > you much more than a clinical antiseptic view of the project. > > For example, say you look at list of member functions of a class and it's > parents. That view tells you nothing about how the class works except by > intuition (i.e. experience in how objects work based on how they are named > and how they are usually implemented.) But why should you choose one > overloaded member function over another? Should you derive a new class and > overload a member function or add a virtual function? > > Beyond that, there is the actual implementation. While it may be "nice" to > show every function and every define from every file, the shear volume of > information (generally without context) solves nothing for me. There's a > define or typedef named something with a certain value; big deal. So now I > won't create a define by that name, but I still don't know what that define > is specifically for. > > And if I want to find out the value of a specific define, there is either > the documentation or search file (grep) functionality. The advantage of the > latter, which I can do very quickly, is that it tells me both the context of > where it was defined, along with the associated defines which may prove even > more helpful, especially if that's what I really wanted in the first place, > and other places where it was used. > > That SNiFF can act as a giant index is not disputed, only that there is no > compelling reason to use it for just that reason since otherwise, I don't > find it a well written tool. It's awkward, uses screen space poorly, doesn't > allow common settings to be made or changed. > > Like the other object browsers I've tried, it doesn't work how I intuitively > want it to. Mainly, I have to do many clicks and take entirely unintuitive > steps to traverse the project from an object perspective. On top of that, > with the projects I'm working on, as with most projects in the real world, > the objects themselves don't tell you much of anything. The browser may tell > you what exists and the programatic relationship of those items, but it > tells you nothing about how those objects interact. > > And that programatic relationship may be misleading; just because task A > calls the function to send a message to task B, doesn't mean it actually > does so in operation or that it may do so only sporadically. Unless you know > the code, the conclusions you draw from the browser may be very misleading. > If you know the code, then, as I said before, the browser in many becomes > just a fancy grep engine. > > But that's not what's at issue. Are people REALLY using SNiFF, or are they > buying it so their ignorant bosses can feel better about how they are > "managing" the project. (Or so they can convince themselves they can hire > inexperienced developers and let them use SNiFF to understand what's going > on.) > > (I've yet to work with a genuine [vs. those with just the title] Sr. > Developer/Engineer I respect who uses object browsers. I'm curious if my > experience is atypical.) > > > > > "Martin Raabe" wrote in message > news:9gijt8$abo$03$1@news.t-online.com... > > Hello Joe, > > I am Martin R. as Dr. Diag referred to, and I would say, that team > > development > > is increasingly coming into the embedded world, as is is since years in > the > > non-embedded world. > > > > I am interrested, what of SNiFF+ 4.0.1 is not to your satisfaction? > > Please reply to me at: > > > > Martin.Raabe at BaSystem.de > > > > > > "Joe Woodbury" schrieb im Newsbeitrag > > news:y8yW6.736$s_.560883@news.uswest.net... > > > My local rep is pushing SNiFF+ at us for our next project. I used SNiFF+ > 2 > > > and didn't like it much, but gave the SNiFF+ 4.x demo a whirl to see if > it > > > could change my mind. I still didn't like it. However, with future > > > employment possibilities in mind, I'm wondering how many > > > developers/companies use SNiFF+ (or at least insist it be used)? > > > > > > > > > > > > > > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: TrueFFS INstallation Problem Date: 19 Jun 2001 14:27:06 -0700 From: aparker@vtlinx.com (parreg) Organization: http://groups.google.com/ Message-ID: <79ef8b60.0106191327.26ff9a8a@posting.google.com> References: <3B28E7E2.A56DA2E@digidata.com> <3B2A3C54.A6640BA5@digidata.com> <79ef8b60.0106180524.58de98ca@posting.google.com> <3B2E1B1D.D4D1BE1D@digidata.com> <79ef8b60.0106190508.ac4a589@posting.google.com> <3B2F754F.388FCDB9@digidata.com> Doug Dotson wrote in message news:<3B2F754F.388FCDB9@digidata.com>... > It is when I try to run tffsDevFormat that it blows up. It is never trying > to call the cfiscsIdentify function to obtain the programming parameters > for the flash device. I'm not sure what you mean by "core". Processor > core? RM7000. > > Here is the dump. As you can see, tffsDevFormat ultimately tries to > identify the flash device be calling tffsIdentifyFlash. This function is > supposed to go through the list of MTDs to determine which kind > of flash it is working with. In my case the only MTD registered is the > CFI/SCS MTD si the cfiscsIdentify function should be called. It > isn't being called. Perhaps it has a problem when only a single MTD > is in the table. I'll try adding a couple more > > Data bus error > Exception Program Counter: 0x800f115c > Status Register: 0x3000ff01 > Cause Register: 0x0000001c > Error address: 0xffffffff, Error ID: 0x0000 > > 800419a4 vxTaskEntry +c : shell (1, 0, 0, 0) > 800bda48 shell +1c8: 800bda6c (eeeeeeee, eeeeeeee, eeeeeeee, eeeeeeee) > 800bdc54 shell +3d4: execute (800fd698, 8010a4c4, 7f, 10) > 800bddbc execute +c8 : yyparse (8010a520, 0, 1, ffffffff) > 800edaf0 yyparse +888: 800ebeb0 (eeeeeeee, 0, eeeeeeee, 0) > 800ec074 yystart +89c: tffsDevFormat (140, 258, 80077bfc, eeeeeeee) > 800a2028 tffsDevFormat +b0 : flCall (800eab18, eeeeeeee, eeeeeeee, eeeeeeee) > 800a5e98 flCall +274: 800a523c (80c77bf8, 0, ffffffff, 5000000) > 800a52dc tffscmpWords +514: flMount (80c77cd4, 96, 80123e20, 800a4f04) > 800f0b00 flMount +30 : flIdentifyFlash (800f3310, 0, 0, a0d24444) > 800f120c flIdentifyFlash+38 : 800f112c (&yyval, 3, 3, 0) > > Thanks, > Doug > > > parreg wrote: > > > Doug Dotson wrote in message news:<3B2E1B1D.D4D1BE1D@digidata.com>... > > > I don't have a function sysTffsFormat. Don't I have to run tffsDevCreate first? > > > When I run usrTffsConfig I get the same error. I put some printf's in > > > cfiscsIdentify() and it never gets called. I checked noOfMTDs and it is 1 > > > as it should be. > > > > > > Doug > > > > ################# > > You won't get anywhere until the flash is formated. You must do > > sysTffsFormat or tffsDevFormat. After this can create the device > > usrTffsConfig ultimately calls tffsDevCreate, but this will only work > > after the flash has been formatted into Erasble Units. What core are > > you using? > > > > > > parreg wrote: > > > > > > > Doug Dotson wrote in message news:<3B2A3C54.A6640BA5@digidata.com>... > > > > > Yes. And when I load it I take a look at noOfMTDs and it is 1 > > > > > indicating that one entry is in the MTD table. > > > > > > > > > > Doug > > > > > > > > > > > > > > > Andray Kaganovsky wrote: > > > > > > > > > > > Doug Dotson wrote in > > > > > > news:3B28E7E2.A56DA2E@digidata.com: > > > > > > > > > > > > > Greetings! > > > > > > > > > > > > > > I am having alot of trouble getting TFFS installed. Followed the > > > > > > > instructions > > > > > > > and got a successful build. After rebooting, I attempt to create a > > > > > > > device > > > > > > > using "tffsDevCreate" and get the following: > > > > > > > > > > > > > > -> tffsDevCreate > > > > > > > > > > > > > > Data bus error > > > > > > > Exception Program Counter: 0x800a60cc > > > > > > > Status Register: 0x3000ff01 > > > > > > > Cause Register: 0x0000001c > > > > > > > Error address: 0xffffffff, Error ID: 0x0000 > > > > > > > > > > > > > > 80041894 vxTaskEntry +c : shell (1, 0, 0, 0) > > > > > > > 800bef98 shell +1c8: 800befbc (eeeeeeee, eeeeeeee, eeeeeeee, > > > > > > > eeeeeeee) > > > > > > > 800bf1a4 shell +3d4: execute (800fe4c8, 8010b304, 2048, > > > > > > > &yypact) > > > > > > > 800bf30c execute +c8 : yyparse (80c77d88, ffffffff, 1, ffffffff) > > > > > > > 800ef040 yyparse +888: 800ed400 (eeeeeeee, 0, eeeeeeee, 0) > > > > > > > 800ed5c4 yystart +89c: tffsDevCreate (140, 258, 80077aec, 0) > > > > > > > 800a1a58 tffsDevCreate +4c : flCall (80c77cc0, 0, 80cfe920, 96) > > > > > > > 800a73bc flCall +248: 800a6540 (0, 0, 8007829c, 0) > > > > > > > 800a6588 tffscmpWords +270: flMount (80124c60, 800a6454, 0, 0) > > > > > > > 800f2050 flMount +30 : flIdentifyFlash (0, 0, 80c77b80, 0) > > > > > > > 800a617c flIdentifyFlash+38 : 800a609c (800ec5e8, 800ec590, 800ea708, > > > > > > > 800e990c) > > > > > > > > > > > > > > My specific configuration is: > > > > > > > > > > > > > > 1) Board is a Galileo EV-64240 with resident Intel E28F640J3A-120 > > > > > > > flash. > > > > > > > 2) I modified mv177-sysTtfs.c file to suit. > > > > > > > 3) I'm using the cfiscs.c MTD. > > > > > > > > > > > > > > I put in a few debug statements and have found that the rfaRegister() > > > > > > > function is being > > > > > > > called. I also put some debug statements in the flRegisterCFISCS() and > > > > > > > cfiscsIdentfy() > > > > > > > routines and they are never called. Since the exception is occurring at > > > > > > > 0xffffffff it appears > > > > > > > to me that the function pointers to the MTD functions are never being > > > > > > > set up but I can't > > > > > > > figure out how this is supposed to be done. > > > > > > > > > > > > > > Thanks, > > > > > > > Doug > > > > > > > > > > > > > > > > > > > > > Douglas P. Dotson > > > > > > > Digi-Data Corporation > > > > > > > 8580 Dorsey Run Rd. > > > > > > > Jessup, MD 20794 > > > > > > > 301-498-0200 x248 > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > [AndrayK]. Did you #defined INCLUDE_MTD_CFISCS in sysTffs.c ? > > > > > > > > you should format the flash first using sysTffsFormat() then > > > > use usrTffsConfig 0,0,"/flash" to create the device. Then type "devs" > > > > confirming the device is there. Once the device is created you can > > > > check the dosFs volumne configuration using > > > > "dosFsConFigShow("/flash"). > > > > > > > > parreg I'm not sure what's going on here. You can debug the code by adding the following in the makefile in the directory wind_base/target/src/drv/tffs: ADDED_CFLAGS= "-g -DDEBUG_PRINT=logMsg". The debug option can not provide source level debugging. It gives log messages in the shell which will help you identify the problem. When you call the tffsDevformat( ) or tffsDevCreate( ), you should get log messages in the shell. You can also define INCLUDE_SHOW_ROUTINES and use tffsShow( ) and see whether the device is recognized. I hope this helps. parreg --------------------------- Newsgroups: comp.os.vxworks Subject: Re: romInit in VxWorks 8260 BSP Date: 19 Jun 2001 14:30:02 -0700 From: sachin@blrmail.com (Sachin) Organization: http://groups.google.com/ Message-ID: References: Bill Pringlemeir wrote in message news:... > Immediate values are loaded in two instruction on PPC. Code is like, > > r4 = start; > r5 = romInit; > r6 = ROM_TEXT_ADRS; > > r4 -= romInit; /* r5 */ > r4 += ROM_TEXT_ADRS; /* r6 */ > (int(*)())r4; /* jump to address in r4 */. > > What is actually happening is that the `C' symbols `start' and > `romInit' have a logical value that is different from where the code > is actually executing from when it boots. This is the value > `ROM_TEXT_ADRS'. The delta between `start' and `romInit' is > calculated. This delta is then added to ROM_TEXT_ADRS which is where > `romInit' is actually located at this point. > > This is a common idiom on all processor models that I know of. On > occasion, the start address can be select between `0' or `0xffff0000' > or something like that. However, the address that the processor boots > from is hard coded. The code is linked to run at a different address, > so care must be taken until it jumps to the link address later on. > > hth, > Bill Pringlemeir. I think i get it now, so correct me if iam wrong. So, what you are saying is that the 'bl' and 'blr' instructions use Effective Addresses and not jump offsets as i thought. If i remember rightly, in the MIPS 4000 series (which is the only ASM i have used), jumps are pure offsets and not an effective address by itself, which is what confused me. However, i still feel, that the comments (in 2 places in fact) in the file are misleading when they say "jump to Flash space", since the code aither completely executes from Flash or from RAM & there is no way that the code jumps from RAM to Flash space. Again this is my understanding, so please do correct me if iam wrong. Anyway, thanks a million, Bill, Sachin. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: romInit in VxWorks 8260 BSP Date: 19 Jun 2001 19:11:42 -0400 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: References: Sender: bpringlemeir@DeadDuck >>>>> "Sachin" == Sachin writes: [snip] Sachin> I think i get it now, so correct me if iam wrong. So, what Sachin> you are saying is that the 'bl' and 'blr' instructions use Sachin> Effective Addresses and not jump offsets as i thought. If i This is correct. Please look at the mnemonics. They are `branch and link' and `branch and link return'. The link register is special, it is the return address of the current function. When you do a BL, you put the value of PC+8 (or something like that) into the link register. Then when you want to return, you do a `blr' which put the link register into the PC. The mtspr is special in that it lets up put an arbitrary register to LR. This is not the `normal mode'. Look at the code again, /* * When the PowerPC 8260 is powered on, the processor fetches the * instructions located at the address 0x100. We need to jump * from the address 0x100 to the Flash space. */ This comment is correct. [...jump value in R4] mtspr LR, r4 /* save destination address*/ /* into LR register */ blr /* jump to flash mem address */ start: [...] Sachin> address by itself, which is what confused me. However, i Sachin> still feel, that the comments (in 2 places in fact) in the Sachin> file are misleading when they say "jump to Flash space", Sachin> since the code aither completely executes from Flash or from Sachin> RAM & there is no way that the code jumps from RAM to Flash Sachin> space. The code does jump from 0x100 (the reset vector on the PPC) to the flash address which is probably `ROM_TEXT_ADRS'... It doesn't really jump anywhere as the Flash will alias (address wraps when you try to use a bigger size than memory). It just really changes the PC from near 0x100 to near `ROM_TEXT_ADRS'; the code remains the same. <1> This is where the `C' code has been linked and the PC must be set properly or any `position dependant C code' would ~BOMB~! That isn't good. After you have changed the PC to `ROM_TEXT_ADRS', the next step is to map RAM to the zero address as this is the usual place for vectors and you would like to be able to change these dynamically. The comment is sort of wrong, put it is mostly right...at that point in time all of the address space is flash. It is difficult to put all the concepts I have just explained in a few words. I suggest you get a PDF of the PowerPC assembler syntax. I am sure that either IBM or Motorola had a PDF on the web somewhere. hth, Bill Pringlemeir. <1> when your climbing a stair way to heaven... wrong alblum. - -- Useful vxWorks URLS, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" "http://www.xs4all.nl/~borkhuis/vxworks/vxw_pt9.html" --------------------------- Newsgroups: comp.os.vxworks Subject: Motorola 5101 Board Support package Date: Tue, 19 Jun 2001 19:16:30 -0400 From: Stephen Rejto Organization: MIT Lincoln Laboratory Message-ID: <3B2FDD4E.D8805CC@ll.mit.edu> We're having problems with our BSP for a Moto 5101 board. Doe snayone have a moto5101 orking reliabaly with VxWorks ? Thanks for any info Stephen Rejto MIT Linconl Lab srejto@ll.mit.edu --------------------------- Newsgroups: comp.os.vxworks Subject: strong ARM pointer align Date: Wed, 20 Jun 2001 08:21:29 +0800 From: "Jun Wu" Organization: Lawrence Berkeley National Laboratory Message-ID: <9gotf3$odp$1@overload.lbl.gov> Hi all, I use intel strongARM SA1110 assabet board and Tornado 2.0. I find if a pointer is not 4-byte align, data may be corrupted. For example void foo() { unsigned int *p; /* set p point with a non-4-byte align address */ p = (int *) 0x08000002; *p = 0xffffffff; /* bytes with address 0x08000000 and 0x08000001 are corrupted!*/ } Are there any compiler options can avoid this problem ? Best regard? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Newbie question: Including spyLib in application code (or project) Date: Wed, 20 Jun 2001 01:34:36 GMT From: "not really" Organization: Road Runner Message-ID: References: My guess would be, the tool is controlled by license from Windriver and you don't have one. "Chris" wrote in message news:cf545555.0106191044.4e7c122c@posting.google.com... > I need to include the spyLib utility in order to profile some of my > application code. Now, being new to Tornado, I was curious as to > where in my project I add the configuration macro (INCLUDE_SPY)? > > According to the Tornado documentation, if the INCLUDE_SPY macro is > defined, then the spyLib functionality will be available. I've tried > both under the "C/C++ Compiler" tab (by adding: -DINCLUDE_SPY) and > under the "Macros" tab for the Properties:Build Specification. And > when I tried to call any of the spyLib routine (either via the shell, > or from my code), I received the following: > > 'spy not configured into this system' > > Any help would be greatly appreciated. > > Thanks, > Christopher Campise --------------------------- Newsgroups: comp.os.vxworks Subject: Re: WindView 2.0.1 Customizing Default User Event Format Date: Tue, 19 Jun 2001 19:36:43 -0700 From: DrDiags Organization: laserlink.net Message-ID: <3B300C3B.FD7112AE@covad.net> References: <78700411.0106190914.62f78462@posting.google.com> Reply-To: drdiags@covad.net Allan, Some of the text that follows may help. If you figure out how to make the GUI understand newlines in the ListBox widget, pass it on. I know in Tcl you can change how the widget deals with newlines inserted in text, but the Windows version of the GUI doesn't seem to allow you to manipulate how to place your formatted strings (i.e. your string is one long text string). See the following copy of information that was passed on to me. Allan Dufour wrote: > > We would like to modify the default format for User Events. > > There appears to be an obivous place to do this within the > "WindView.win32.tcl" file in the procedure "eventFormat". Having > modified this procedure, we still see no change in the format? > > The current default is to display the raw hex data associated with the > event. > > Any ideas? Are we in the right place? SPR# 22563 ................................................................ DATE: Oct 01 1998 TITLE: WindView 2.0: display user event data example in F.3 of User's Guide does not correctly display "Formatted Display" Problem_Description User EventBase display doesnt properly format for WindView 2.0, as shown in the User's Guide, section F.3: <...... SNIP .......> An FAE referred me to TSR# 142050. The method described in there is fairly good (i.e., the routine userFormatNNNNN actually gets called where the variable wvUsrEventFormat in eventbase.tcl never gets used) except there is still no instructions on how to take 'event' and format it into anything useful. HERE IS THE TEXT PROVIDED BY TSR 142050 ======================================= > It seems that, at the VERY last moment someone stuck in a > way of formatting user event data for display in the Show Event dialog box. > It's a tad different from the way it was done before. > I haven't tried it myself, but the theory goes as follows: > > 1) it is assumed either/or that > a) you have a HOMEDRIVE and a HOMEPATH environment variable set up, and/or > b) you have a HOME environment variable set up > >From these is constructed a search path... > > 2) in the above search path, there should be a .wind subdirectory. > > 3) in the .wind subdirectory, create a "WindView" subdirectory and a file in > that directory called "userEvents.tcl" (if they don't already exist, which they should under Unix). > e.g. > if HOME=c:\cjtc > the file creates should be > c:\cjtc\.wind\WindView\userEvents.tcl > > 4) this file needs to be edited to contain tcl code required to format the > way in which user events are displayed. > > e.g. for each user event that requires data formatting, insert something > like > > proc userFormat00200 {arg} { > return "result is $arg" > } > > NB. For each user event which need custom formatting, the name of the proc > which should be created is > > userFormatNNNNN > > where NNNNN is the decimal number of the user event to be formatted. > e.g. for user event 200 (decimal), the name of the proc to be inserted > should be > > userFormat00200 > > The tcl proc should format the arg string parameter and return a string > formatted as required. > > The end result of all of this should be that the formatted string is shown > as required in the Show Events dialog box. Note that the user data shown in > the status line (when the cursor just hovers over an event icon) is still in > straight hex. =================================================================== After staring at the WindView tcl scripts for many hours, we finally wrote the following which works: proc userFormat00014 {event} { set data [userEventFormat $event {n1 n1 n2 n4}] scan [lindex $data 0] "%2x" cpuid #messageBox "cpuid = $cpuid, data1 = [lindex $data 3]" if [cpuLittleEndian $cpuid] { set data2 [swapTargetInt [lindex $data 3]] } else { set data2 [lindex $data 3] } scan $data2 "%8x" data3 #messageBox "data2 = $data2, data3 = $data3" return "sequence number = $data3" } Please file an SPR(s) for the following: 1. The documentation on pages 246-249 (section F.3) is completely wrong. A FAQ or app note should be written with the correct information (most of it detailed in the aforementioned TSR or below). 2. The tcl procedure eventFormat in WindView.win32.tcl never gets called. It's the routine that actually pays attention to wvUsrEventFormat (the data structure the manual tells you to use). This routine expects the data in one [<]format[>] (one that you can use with usrUnpack, ie, the one described in the manual). The actual [<]format[>] used is incompatible with usrUnpack - you must call userEventFormat instead. This needs documented as well. FYI, the usr routines are mostly in %WIND_BASE%/host/resource/tcl/WindView.win32.tcl. The user routines are mostly in %WIND_BASE%/host/resource/tcl/app-config/WindView/database.tcl. 3. We can't use functions such as unpackint and unpackstring as the upvar causes a tcl error. We need to do the scan and swapTargetInt calls ourselves.The available calls need to be documented so we don't have to read the tcl source to find them. We also don't have the cpuId (since our events aren't in the format expected by usrUnpack). 4. FYI, there is a bug in unpackint (and possibly some other functions). It's goal is to format the int as a decimal representation and byte swap if necessary. However, the byte swap routine, swapTargetInt, always formats the data in hex. The scan should not be performed until after the swapTargetInt. Big endian targets will return the result as a %d, little endian targets as a %x. 5. In database.tcl, the routine wvLoadUserEventFormats does not look in WIND_BASE first (as the routine wvFindUserDefinedIcons does). I believe they should both look in WIND_BASE first. 6. Even if pages 246-249 were right, there are several typos including switching between wvUsrEventFormat and wvUserEventFormat. The routine postEvent is also mentioned but I cannot find it in a header file. <.... SNIP.....> I've found one additional bug. Bug: proc cpuLittleEndian {cpuId} { # return endiannes of target based on cpuId. This doesn't # take ambi-endian architectures into account. In this # arrangement we have no direct connection to the target # server so we can't have authoritative data. return [expr $cpuId >= 20 && $cpuId <= 29 || $cpuId >= 80 && $cpuId <= 89] } This does not work for the ARM or SIMNT platforms. ARM cpuid are >=120 and <=125 and are little endian. SIMNT has cpuID 100. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Newbie question: Including spyLib in application code (or project) Date: Tue, 19 Jun 2001 19:52:14 -0700 From: DrDiags Organization: laserlink.net Message-ID: <3B300FDE.24D5A4C3@covad.net> References: Reply-To: drdiags@covad.net Chris, If you are using Tornado 2/vxWorks 5.4, you should be able to enable the spy utility by making sure your Kernel has it selected in the vxWorks Project Facility Tab. If you are building from the command line, you should be able to #define INCLUDE_SPY to your config.h file, after the #include "configAll.h" statement. Are you using this version of Tornado? You can also review the on-line help (are you using Windows or Unix as your Host?) for spy. spyLibInit( ) NAME spyLibInit( ) - initialize task cpu utilization tool package SYNOPSIS void spyLibInit (void) DESCRIPTION This routine initializes the task cpu utilization tool package. If the configuration macro INCLUDE_SPY is defined, it is called by the root task, usrRoot( ), in usrConfig.c. RETURNS N/A SEE ALSO spyLib, usrLib For the project facility, as long as spyLib.o exists in your archive libYOURARCHgnuvx.a or whatever it is, the project will allow you to include this. I thought the High resolution clock needed to be in your BSP and I always add the AUX clock, but other than that, I am at a loss. BTW, what is your BSP? Maybe someone with the same BSP could shed some light on this one. Chris wrote: > > I need to include the spyLib utility in order to profile some of my > application code. Now, being new to Tornado, I was curious as to > where in my project I add the configuration macro (INCLUDE_SPY)? > > According to the Tornado documentation, if the INCLUDE_SPY macro is > defined, then the spyLib functionality will be available. I've tried > both under the "C/C++ Compiler" tab (by adding: -DINCLUDE_SPY) and > under the "Macros" tab for the Properties:Build Specification. And > when I tried to call any of the spyLib routine (either via the shell, > or from my code), I received the following: > > 'spy not configured into this system' > > Any help would be greatly appreciated. > > Thanks, > Christopher Campise --------------------------- Newsgroups: comp.os.vxworks Subject: How to get VxSim to execute commands from script file automatically after it is started with vxWorks -p ? Date: 19 Jun 2001 20:45:22 -0700 From: jacklv@sc.mcel.mot.com (Jack) Organization: http://groups.google.com/ Message-ID: Should I modify usrStartupScript file in config/comps/src ? How to modify this file or Is there any other way to do so? Thanks a lot! BRs, Jack --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Stack overflow Date: Wed, 20 Jun 2001 10:46:44 +0100 From: Vijay Kumar Peshkar Organization: Cisco Systems Inc. Message-ID: <3B307103.A975E24A@wipro.com> References: <9gnj97$7li$1@overload.lbl.gov> This is a multi-part message in MIME format. - --------------08055D18B87DF79215A08679 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hello Sanjay, You need to account for all the stack used by locals, lib calls, interrupt etc while you decide your stack size. Also you can use the checkStack() during development to see how close you are to overflow. Recently there were few threads of discussion on stack overflow. May be you can refer them Did you check if sysMemTop() is supported in your BSP?. Also the exception may give a clue to the error you are having. Hope this helps, Vijay K Sanjay wrote: > Friends, > We are using VxWorks5.4. > While testing my module, I was getting memPartAlloc() > error or memPartFree error in some other tasks. > Finally I found out that this was because the size of > the local variables in my task is more than the > allocated stack size. > How do I findout the stack overflow? > > Another question, > I am adding adding additional memory to the memory > pool after loading the module. The code looks like > this, > > if (sysPhysMemTop() > sysMemTop()) > memAddToPool (sysMemTop(), > (sysPhysMemTop()-sysMemTop()); > > After this if I try to execute sysMemTop on windshell > it will crash giving exception. I could not figure out > the reason for this. > I would appreciate your help/suggestion regarding > these problems. > Thanks, > Sanjay > > __________________________________________________ > Do You Yahoo!? > Spot the hottest trends in music, movies, and more. > http://buzz.yahoo.com/ - --------------08055D18B87DF79215A08679 Content-Type: text/x-vcard; charset=us-ascii; name="vijay.peshkar.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Vijay Kumar Peshkar Content-Disposition: attachment; filename="vijay.peshkar.vcf" begin:vcard n:Peshkar;Vijay Kumar tel;work:5530034/35/35 x2237 x-mozilla-html:FALSE org:Wipro Global R & D;Wishing you blue skies, a couple of daisies, some fresh air and a smile! adr:;;;;;; version:2.1 email;internet:vpeshkar@cisco.com note:Wishing you blue skies, a couple of daisies, some fresh air and a smile! fn:Vijay Kumar Peshkar end:vcard - --------------08055D18B87DF79215A08679-- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How common is SNiFF usage? Date: Tue, 19 Jun 2001 23:42:05 -0600 From: "Joe Woodbury" Message-ID: References: <9gijt8$abo$03$1@news.t-online.com> <9goe18$frm$05$1@news.t-online.com> I initially, and very briefly, used SNiFF 2.x. Most recently I was testing 4.0.1. Today, we chose to not purchase SNiFF+ for inclusion with our Tornado Pro package, so that issue is closed for now. However, I do beg to differ with you on some of your points: Unless any editor, IDE or even any tool which interfaces to a version control system has suddenly been redefined as a "Team Development" tool (in capital letters), SNiFF most certainly is not. Again, it may facilitate development in a team environment, but it lacks the collaboration tools of a true "Team Development" tool--for example, being able to tie design specifications and changes to specific objects or modules. While others may differ, the product I've been most impressed with in this regard is StarTeam from StarBase. (www.starbase.com -- no, I don't work for them). Furthermore, I find the claim that SNiFF is a "SourceCode Management system" misleading. Granted, this may be a matter of semantics, but SNiFF doesn't manage source code as much as it browses it. While the basic functionality of SNiFF may be "easy" so is the basic functionality of PageMaker. But it doesn't mean anything. The real question, in this regard, is how easy it is to fully leverage the tool. Conceding that I didn't have the incentive to become a SNiFF expert, nevertheless, as I stated before, I found it surprisingly non-intuitive. When you come out with your next version of SNiFF, I'll give it another try. Though I would advise that you hire a Windows UI specialist and do some genuine usability testing. I'd also recommend that you make the integration between SNiFF and Tornado a little less obtuse. Joe "Martin Raabe" wrote in message news:9goe18$frm$05$1@news.t-online.com... > Hello Joe, > thank you for expressing your impressions so detailed. > I will take your input back to the factory. > > My short statement to this is as following: > What SNiFF+ is: > - SNiFF+ is a SourceCode Management system, which handles SourceCode of > different languages and Platforms. > - SNiFF+ is a SDE (Software Development Environment) which allows to > develop independent from the target platform. > - SNiFF+ supports in these rols also Team Development, which is very > helpfull in small and medium sized (1-100 developers) teams. > - SNiFF+ is an easy to handle Tool which behaves the same on Unix, and > Windows. > - SNiFF+ allows to faster get into foreign or (own) older SourceCode. > - SNiFF+ helps to live project strategies, individuals would not follow > without SNiFF+. > > What SNiFF+ is NOT: > - SNiFF+ is not a debugging tool for OOP-Languages, where runtime behavior > is a significant part of the debugging process. > - SNiFF+ is not a Project Manager.It only adapts to VersionControl and > Project Management Systems. > > I have a question for you. > Which SNiFF+ Version are you referring to? > I refer to the latest Version 4.0.1. > I can understand that you do not feel that comfortable with the 2.x or 3.0.x > Versions. > But the feedback from customers on Verison 3.2.1 and 4.0.x is excellent! > > So I am afraid SNiFf+ will make it into a huge project basis accross > branches and markets! > > Ciao > > Martin Raabe > > > "Joe Woodbury" schrieb im Newsbeitrag > news:HS8X6.652$ZR1.297147@news.uswest.net... > > There is no direct connection between SNiFF and team development. SNiFF is > > an object browser, not a team collaboration tool. SNiFF may facilitate > large > > project develpment, which is often done in a team environment, but it > > completely lacks the other collaboration tools more commonly associated > with > > a team development environment. > > > > As far as particular criticisms of SNiFF go; my distaste of object > browsers > > is that they don't really clarify anything. You end up with massive lists > of > > functions and classes (and data and function members), none of which give > > you much more than a clinical antiseptic view of the project. > > > > For example, say you look at list of member functions of a class and it's > > parents. That view tells you nothing about how the class works except by > > intuition (i.e. experience in how objects work based on how they are named > > and how they are usually implemented.) But why should you choose one > > overloaded member function over another? Should you derive a new class and > > overload a member function or add a virtual function? > > > > Beyond that, there is the actual implementation. While it may be "nice" to > > show every function and every define from every file, the shear volume of > > information (generally without context) solves nothing for me. There's a > > define or typedef named something with a certain value; big deal. So now I > > won't create a define by that name, but I still don't know what that > define > > is specifically for. > > > > And if I want to find out the value of a specific define, there is either > > the documentation or search file (grep) functionality. The advantage of > the > > latter, which I can do very quickly, is that it tells me both the context > of > > where it was defined, along with the associated defines which may prove > even > > more helpful, especially if that's what I really wanted in the first > place, > > and other places where it was used. > > > > That SNiFF can act as a giant index is not disputed, only that there is no > > compelling reason to use it for just that reason since otherwise, I don't > > find it a well written tool. It's awkward, uses screen space poorly, > doesn't > > allow common settings to be made or changed. > > > > Like the other object browsers I've tried, it doesn't work how I > intuitively > > want it to. Mainly, I have to do many clicks and take entirely unintuitive > > steps to traverse the project from an object perspective. On top of that, > > with the projects I'm working on, as with most projects in the real world, > > the objects themselves don't tell you much of anything. The browser may > tell > > you what exists and the programatic relationship of those items, but it > > tells you nothing about how those objects interact. > > > > And that programatic relationship may be misleading; just because task A > > calls the function to send a message to task B, doesn't mean it actually > > does so in operation or that it may do so only sporadically. Unless you > know > > the code, the conclusions you draw from the browser may be very > misleading. > > If you know the code, then, as I said before, the browser in many becomes > > just a fancy grep engine. > > > > But that's not what's at issue. Are people REALLY using SNiFF, or are they > > buying it so their ignorant bosses can feel better about how they are > > "managing" the project. (Or so they can convince themselves they can hire > > inexperienced developers and let them use SNiFF to understand what's going > > on.) > > > > (I've yet to work with a genuine [vs. those with just the title] Sr. > > Developer/Engineer I respect who uses object browsers. I'm curious if my > > experience is atypical.) > > > > > > > > > > "Martin Raabe" wrote in message > > news:9gijt8$abo$03$1@news.t-online.com... > > > Hello Joe, > > > I am Martin R. as Dr. Diag referred to, and I would say, that team > > > development > > > is increasingly coming into the embedded world, as is is since years in > > the > > > non-embedded world. > > > > > > I am interrested, what of SNiFF+ 4.0.1 is not to your satisfaction? > > > Please reply to me at: > > > > > > Martin.Raabe at BaSystem.de > > > > > > > > > "Joe Woodbury" schrieb im Newsbeitrag > > > news:y8yW6.736$s_.560883@news.uswest.net... > > > > My local rep is pushing SNiFF+ at us for our next project. I used > SNiFF+ > > 2 > > > > and didn't like it much, but gave the SNiFF+ 4.x demo a whirl to see > if > > it > > > > could change my mind. I still didn't like it. However, with future > > > > employment possibilities in mind, I'm wondering how many > > > > developers/companies use SNiFF+ (or at least insist it be used)? > > > > > > > > > > > > > > > > > > > > > > > > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Problem with sm network on mcp750 Date: Wed, 20 Jun 2001 14:06:34 +0800 From: "tanghaoyu" Organization: Bentium Ltd. Message-ID: <9gpedl$jk3$1@mail.cn99.com> References: <9gmcei$js4$1@mail.cn99.com> <44a01eba.0106190439.75377191@posting.google.com> I had even change this macro to 14,but the problem still existed. "ellin" wrote in message news:44a01eba.0106190439.75377191@posting.google.com... > I think you should pay attention to the limit of the count of > none-system processor that mcp750 supports with sm network, and it's > defined in the config.h as below. > ... > # define MAX_MCPN750_COUNT 7 > ... > You can increase the number,and try again. > > Speak just for myself. > > ellin > --------------------------- Newsgroups: comp.os.vxworks Subject: Tornado2 dependencies Date: Wed, 20 Jun 2001 09:07:14 +0200 From: walter gamsjaeger Organization: Siemens AG Message-ID: <3B304BA2.9347F6DD@siemens.at> Hello. I'm using Tornado2 with Windows NT4.0 SP6 and I have problems when I try to generate dependencies. For some reasons Tornado2 doesn't insert the path of some include files in the dependency list. For example: caal1h_factory_suces.o: E:/SU_CES/src/target/src/aal1h/caal1h_factory_suces.cpp \ Z:/mgw_global/source/framework/include/rtbase.inl \ Z:/mgw_global/source/framework/include/cmanaged.h \ rtconfig.h \ <---- missing path "Z:/mgw_global/source/framework/project" Z:/mgw_global/source/framework/include/cstdapp.h \ The macro VPATH is set to: $(filter-out -I.,$(filter -I%,$(CFLAGS))) and CFLAGS contains: - -IZ:/mgw_global/source/framework/project - -IZ:/mgw_global/source/framework/include The only advice I got from colleagues was to move the directories up and down in the Builds->Properties->C/C++ compiler Menu of the Workspace until it works. Does somebody have a better solution? Thanks for your help. Walter Gamsjaeger --------------------------- Newsgroups: comp.os.vxworks Subject: Re: EHOSTDOWN error message Date: Wed, 20 Jun 2001 11:58:59 +0400 From: Pavel Kovalev Organization: Gamma NNTP server Moscow Russia Message-ID: <3B3057C3.A500241@auriga.ru> References: <9go4ri$bh8$1@overload.lbl.gov> Hi Joe, Joe Carissimi wrote: > > Does anyone know which layer generates the EHOSTDOWN > error id? We are running Tornado II on a DY4 SVM179 SBC, > using UDP sockets. We are "blindcasting" messages from > one SBC to other SBCs. When we attempt to send a message > to a non-powered target, this errno (0x43) is returned from the > sendto() routine. It occurs about 5 seconds after the first > message is sent. I can snoop the network and see ARP > requests for the IP (with no reply of course). Is there a > layer which can be excluded or a parameter which can be > adjusted to remove this unnecessary error response? Can't > seem to find anything in the documentation (or WindSurf) > as per a solution. Note that this is not a problem in our > VxWorks 5.2 system. When you send a packet to a host and your system does not have an ARP cache entry for the host, the following happens: - - ARP holds the mbuf chain containing the packet. - - ARP tries to resolve the destination IP address. - - Only one packet is held. If other packets are sent while ARP tries to resolve the address, the previously held packet(s) are freed, and only the last packet is held. - - On success, ARP sends out the held packet. On failure (if there is no ARP reply after 5 ARP requests, which means 5 seconds after the initial ARP request), the held packet is freed and the route to the destination is marked with the RTF_REJECT flag. - - While RTF_REJECT is set, attempts to send packets to the destination result in EHOSTDOWN or EHOSTUNREACH; ARP requests are not sent. After 20 seconds RTF_REJECT is cleared and new ARP attempts are allowed. This is a fundamental behavior of the BSD 4.4 TCP/IP stack used in VxWorks, and you cannot change this. I suggest that you add appropriate logic to you program to handle this case. Regards, Pavel. > > -- > Joe Carissimi > BBN Technologies > Middletown, RI 02842 > (401) 848-3525 --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Memory protection and Loadmodule Date: Wed, 20 Jun 2001 10:04:38 +0100 From: Simon Farnsworth Message-ID: <3B306726.1020906@snellwilcox.com> References: <3B2F76F5.5517B48F@hotmail.com> <3B2F7BFA.7020509@snellwilcox.com> <3B2FA585.90907A05@hotmail.com> Beginner wrote: > Thanks again. And by the way, how do you do a stack backtrace? If you have a shell open (either target based or host shell), a backtrace is dumped in mangled form to the shell. To get an unmangled backtrace, start the debugger and attach to the dead task. In Tornado, this will fill a backtrace window for you, and there is a GDB command to dump one as well --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Wed Jun 20 06:10:51 2001 From: s.eckardt@ixthos.com Date: Wed Jun 20 06:10:53 PDT 2001 Subject: Re: Problem with sm network on mcp750 Dafna, I just implemented a vxWorks cPCI Shared Memory network and had the same problem. Basically, the "count" field and the "head pointer" got out of sync: the "count" erroneously indicated that there was a queued packet, but the "head pointer" was NULL. In my case, the problem was due to "read aliasing" in a PCI bridge. In a nutshell: Board A requests read of location X, Board A is told to retry, Board B writes to location X, Board B requests read of location X, Board B gets the (pre-write) read response from Board A's read request. This behavior is documented in Item 6 in Section 3.10 of the Revision 2.2 PCI Local Bus Specification. This behavior can happen when two boards are trying to access the same location, or when one board is reading a cache line worth of data (e.g., 32 bytes) that overlaps the location being accessed by the other board. I had to get WRS-proprietary source code to fix my problem. I submitted TSR #241800 to document the underlying problem and the suggested fixes for future releases. Here is the text of the TSR: -------------------------------------------------------------------------------- Some Shared Memory functions do not work properly on a cPCI backplane. Specifically, functions that perform a write-readback may fail due to erroneous read responses returned by PCI bridges. Two such functions are: smUtilSoftTas() and smPktSllGet(). The recommended solution is to perform a flushing read after the write (i.e., read, but discard the response data). The smUtilSoftTas() routine can be corrected this way; refer to the workaround implemented in the sysBusTas() function of the Motorola PPMC750 BSP. The smPktSllGet() routine can also be corrected this way, but it would be better to fix the code by eliminating the unnecessary readback on "listLocalAdrs->count". -------------------------------------------------------------------------------- Suzie Eckardt Chief Software Engineer Ixthos, Inc. s.eckardt@ixthos.com 703-779-2534 x305 -------------------------------------------------------------------------------- Dafna wrote: As we increase the number of MPC860 CPUs booting over the SM network (8 CPUs and more), we see some of the processors, sometimes all, hang when searching for the boot file, or during booting itself. Once this problem occurs, the CPUs that "got into trouble" never recover. Our research has shown that when things go wrong, the head of the "free-packets-list" at the shared memory is overridden. From what we've seen, this list is used by the 750 to receive packets from the 860. We can also see that the shared-memory area is initialized (ready value at the ANCHOR is 0x87654321), and the heartbeat advances, showing the system is alive. From vxwexplo-errs@csg.lbl.gov Wed Jun 20 10:43:45 2001 From: "David Laight" Date: Wed Jun 20 10:43:47 PDT 2001 Subject: Re: strong ARM pointer align > I use intel strongARM SA1110 assabet board and Tornado 2.0. > I find if a pointer is not 4-byte align, data may be corrupted. For example > /* set p point with a non-4-byte align address */ > p = (int *) 0x08000002; > *p = 0xffffffff; > bytes with address 0x08000000 and 0x08000001 are corrupted!*/ What do you expect to happen! Support for misaligned transfers is VERY processor specific. If you read the ARM V4 architecture guide you find there are two options for the ARM: 1) trap misaligned transfers 2) access the requested word and rotate the answer (puts the addressed byte into the lsb. Unfortunately you cannot enable the trap because gcc uses misaligned 32-bit transfers for some 16-bit requests. If you MUST support misaligned pointers, you have to code the byte reads and shifts yourself. Any compiler options are not portable and will seriously affect the performance of other code. In practise it is always more efficient to 'naturally' align all data items - even on sytems (like the x86) which support misaligned tranfers in the expected way. The reason for not 'supporting' misaligned transfers is simple, the hardware to generate two bus cycles and shift the results isn't the problem. The problems arise when the transfer crosses an MMU page boundary - and the second page isn't resident. Recovering from this is quite complicated. (Ever seem the size of the fault stack frames on a 68020 - required to restart things like the CMP2 instruction correctly). David From vxwexplo-errs@csg.lbl.gov Wed Jun 20 11:51:28 2001 From: karunakaran nair Date: Wed Jun 20 11:51:31 PDT 2001 Subject: comp.os.vxworks newsdigest --0-35878476-993063087=:30902 Content-Type: text/plain; charset=us-ascii please if somebody has an understanding of stocks, shares, how companies are bought by other companies,decision factors etc, pl. write to me. or, if you know a good book that is worth reading to know all these, pl. write to me. --------------------------------- Do You Yahoo!? Get personalized email addresses from Yahoo! Mail - only $35 a year! http://personal.mail.yahoo.com/ --0-35878476-993063087=:30902 Content-Type: text/html; charset=us-ascii

please if somebody has an understanding of stocks, shares, how companies are bought by other companies,decision factors etc, pl. write to me. or, if you know a good book that is worth reading to know all these, pl. write to me.


Do You Yahoo!?
Get personalized email addresses from Yahoo! Mail - only $35 a year!
http://personal.mail.yahoo.com/ --0-35878476-993063087=:30902-- From vxwexplo-errs@csg.lbl.gov Wed Jun 20 16:43:05 2001 From: "Crozier, Bruce" Date: Wed Jun 20 16:43:07 PDT 2001 Subject: RE: DHCP problems on vxWorks Hello Serg, If you are running Tornado 2.0.2, then you have to use dhcpcOptionSet(pCookie,TAG) to request that option TAG be sent to you by the dhcp server. Then, you must create a structure to store the parameters that you want, and attach it to the bootParams structure, so that the dhcp routine will know to copy the parameter from the return packet into your bootParams structure. In bootConfig.c, dhcpGet() routine, they clear the bootParams structure using bzero(), which NULLs out all of the structure pointers. You then put back those that you want. For example, if you want to get the hostname parameter, char dhcp_hostName[BOOT_HOST_LEN]; /* create new structure to store host name */ bootParams.hostname = dhcp_hostName; dhcpcOptionSet(pCookie, _DHCP_HOSTNAME_TAG); if(bootParams.hostname[0] != EOS) printf(" Target Name: %s\n\r",bootParams.hostname); Regards, Bruce -----Original Message----- From: Serg [mailto:sergey@congruency.co.il] Sent: Wednesday, June 20, 2001 10:38 AM Hi! Please, could you help me? I wanna get not standard parameters (for example 121,122) from DHCP server. I can add non-standard options by the dhcpcBootReqSpec variable, but I havn't found way to get this options. I know that server send it. If I use low-level functions (dhcpBind, not dhcpBootBind) I can access to full offer by the dhcpcMsgIn variable way, but can't add non-standard parameters to Request becouse dhcpcBootReqSpec isn't used. Thanks so much. Serg. From vxwexplo-errs@csg.lbl.gov Thu Jun 21 02:55:10 2001 From: "Jun Wu" Date: Thu Jun 21 02:55:12 PDT 2001 Subject: system tick rate Hello vxWorks experts, I use watch dog implement a periodic event in my application. Since the event should occur every 6ms, I have to change the system tick rate to 166. Is it safe that 166 can be divide exactly by 1000? Will the remaining 4ms (1000 - 166*6 = 4) accumulate and affect the vxWorks? BR From vxwexplo-errs@csg.lbl.gov Thu Jun 21 04:03:25 2001 From: Vxworks Exploder Date: Thu Jun 21 04:03:29 PDT 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Thu Jun 21 04:03:20 PDT 2001 Subject: Updating gcc 2.7.2 for i386 Subject: Re: Updating gcc 2.7.2 for i386 Subject: Re: Motorola 5101 Board Support package Subject: re: Problem with sm network on mcp750 Subject: Re: How common is SNiFF usage? Subject: Re: vxWorks.st - how to build Subject: Re: romInit in VxWorks 8260 BSP Subject: Re: Wind Web Server 2.0, httpSymTabAdd function failure Subject: Re: strong ARM pointer align Subject: Re: global objects in VxWorks Subject: Re: How many END_TBL_ENTRY (entries) needed for 3 PHYs (each with 2 ports) Subject: re: strong ARM pointer align Subject: Re: How common is SNiFF usage? Subject: Re: romInit in VxWorks 8260 BSP Subject: Re: Newbie question: Including spyLib in application code (or project) Subject: Re: How common is SNiFF usage? Subject: mvme162 or 172 with Petra asic problem? Subject: Re: How common is SNiFF usage? Subject: Re: How common is SNiFF usage? Subject: writing a Shell for vxSim Subject: RE: DHCP problems on vxWorks Subject: Reserving Memory Subject: writing a "Shell" to run on top of "tShell" of vxSim Subject: writing a "Shell" to run on top of "tShell" of vxSim Subject: writing a "Shell" to run on top of "tShell" of vxSim Subject: SOS - how to download/upload files from/to flash memory? Subject: Re: Cannot start the server Subject: Re: Cannot start the server Subject: Re: How common is SNiFF usage? Subject: Re: romInit in VxWorks 8260 BSP Subject: TFFS limitations? Subject: system tick rate Subject: Re: Command line syntax for compiling Tornado projects Subject: Re: Cannot start the server ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Updating gcc 2.7.2 for i386 Date: 20 Jun 2001 04:06:08 -0700 From: igtorque@eliop.es (=?ISO-8859-1?Q?Ignacio_Gonz=E1lez?=) Organization: http://groups.google.com/ Message-ID: <9ec53d11.0106200306.1414178a@posting.google.com> Hello. I have VxWorks 5.4/Tornado 2.0 , PC host, i386 targets. As you may know, the compiler version supplied with this configuration is gcc 2.7.2. Now I need to use namespaces. So an updating of gcc would be fine (2.7.2 does not support namespaces). I suppose there are some tricky matters in doing this. I'm *sure* somebody has done this before. Can anybody share this knowledge with me? By the way, I heard a long time ago that version 2.9.x of gcc generates much more code than 2.7.2 when using the STL. Is that right? Thanks for your time. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Updating gcc 2.7.2 for i386 Date: Wed, 20 Jun 2001 12:37:24 +0100 From: Simon Farnsworth Message-ID: <3B308AF4.1000102@snellwilcox.com> References: <9ec53d11.0106200306.1414178a@posting.google.com> Ignacio Gonz=E1lez wrote: > Now I need to use namespaces. So an updating of gcc would be fine > (2.7.2 does not support namespaces). I suppose there are some tricky > matters in doing this. > I'm *sure* somebody has done this before. Can anybody share this > knowledge with me? DaveK has produced a gcc 2.95.x release for PPC target; his pages at http://www.newgcc4vxworks4ppc.cjb.net/ explain how he did so. Note that C++ name mangling has changed between the two versions; you=20 will need to recompile your BSP to make the new compiler work, and there = may be further issues that would require an updated C++ runtime from=20 WindRiver. Of course, when the next Tornado service pack comes out (RSN ;-), the=20 compiler update will be included. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Motorola 5101 Board Support package Date: Wed, 20 Jun 2001 12:23:54 GMT From: William Dennen Organization: Excite@Home - The Leader in Broadband http://home.com/faster Message-ID: <3B309668.F8D54FEB@home.com> References: <3B2FDD4E.D8805CC@ll.mit.edu> Reply-To: wdennen@home.com Stephen Rejto wrote: > > We're having problems with our BSP for a Moto 5101 board. > Doe snayone have a moto5101 orking reliabaly with VxWorks ? > > Thanks for any info > Stephen Rejto > MIT Linconl Lab > srejto@ll.mit.edu Direct the problem description to your local FAE for the distributor from whom you purchased the board and supplied the Early Access BSP. We'll take a look at it ... Regards - -- >@< Bill Dennen william.dennen@motorola.com Motorola Computer Group --GTAC --------------------------- Newsgroups: comp.os.vxworks Subject: re: Problem with sm network on mcp750 Date: Wed, 20 Jun 2001 09:10:35 -0400 From: s.eckardt@ixthos.com Organization: Lawrence Berkeley National Laboratory Message-ID: <9gq7l6$fsr$1@overload.lbl.gov> Dafna, I just implemented a vxWorks cPCI Shared Memory network and had the same problem. Basically, the "count" field and the "head pointer" got out of sync: the "count" erroneously indicated that there was a queued packet, but the "head pointer" was NULL. In my case, the problem was due to "read aliasing" in a PCI bridge. In a nutshell: Board A requests read of location X, Board A is told to retry, Board B writes to location X, Board B requests read of location X, Board B gets the (pre-write) read response from Board A's read request. This behavior is documented in Item 6 in Section 3.10 of the Revision 2.2 PCI Local Bus Specification. This behavior can happen when two boards are trying to access the same location, or when one board is reading a cache line worth of data (e.g., 32 bytes) that overlaps the location being accessed by the other board. I had to get WRS-proprietary source code to fix my problem. I submitted TSR #241800 to document the underlying problem and the suggested fixes for future releases. Here is the text of the TSR: - -------------------------------------------------------------------------------- Some Shared Memory functions do not work properly on a cPCI backplane. Specifically, functions that perform a write-readback may fail due to erroneous read responses returned by PCI bridges. Two such functions are: smUtilSoftTas() and smPktSllGet(). The recommended solution is to perform a flushing read after the write (i.e., read, but discard the response data). The smUtilSoftTas() routine can be corrected this way; refer to the workaround implemented in the sysBusTas() function of the Motorola PPMC750 BSP. The smPktSllGet() routine can also be corrected this way, but it would be better to fix the code by eliminating the unnecessary readback on "listLocalAdrs->count". - -------------------------------------------------------------------------------- Suzie Eckardt Chief Software Engineer Ixthos, Inc. s.eckardt@ixthos.com 703-779-2534 x305 - -------------------------------------------------------------------------------- Dafna wrote: As we increase the number of MPC860 CPUs booting over the SM network (8 CPUs and more), we see some of the processors, sometimes all, hang when searching for the boot file, or during booting itself. Once this problem occurs, the CPUs that "got into trouble" never recover. Our research has shown that when things go wrong, the head of the "free-packets-list" at the shared memory is overridden. From what we've seen, this list is used by the 750 to receive packets from the 860. We can also see that the shared-memory area is initialized (ready value at the ANCHOR is 0x87654321), and the heartbeat advances, showing the system is alive. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How common is SNiFF usage? Date: Wed, 20 Jun 2001 15:22:03 +0200 From: "Matthias Guelck" Organization: Heidelberg Digital Message-ID: <9gq81t$c2v$1@news.hdpp.de> References: <9gijt8$abo$03$1@news.t-online.com> <9goe18$frm$05$1@news.t-online.com> Hello Joe, I have been considering to evaluate the usage of SNiFF for our team. Fortunately I have read this thread before. From my point of view, WRS is able to sell high priced tools with very little benefit. With Tornado we bought WindNavigator and LOOK!, because WRS announced them as very useful to understand source code written by other people. After we paid the money, no one at WRS was able to teach us, how to take benefit from the tools. They only repeated the slogans from their colorful advertising booklets. Complaining about some debugger crashes (we have got hundreds of crash log files) and problems (default window size for watch, variables,…), WRS told us, that the whole world is very satisfied, and we are the only people in the world complaining about their famous products! So my question is: did you receive replies from other SNiFF users (non WRS employees like Martin Raabe)? Best regards Matthais "Joe Woodbury" wrote in message news:EIWX6.2104$zE1.780634@news.uswest.net... > I initially, and very briefly, used SNiFF 2.x. Most recently I was testing > 4.0.1. > > Today, we chose to not purchase SNiFF+ for inclusion with our Tornado Pro > package, so that issue is closed for now. However, I do beg to differ with > you on some of your points: > > Unless any editor, IDE or even any tool which interfaces to a version > control system has suddenly been redefined as a "Team Development" tool (in > capital letters), SNiFF most certainly is not. Again, it may facilitate > development in a team environment, but it lacks the collaboration tools of a > true "Team Development" tool--for example, being able to tie design > specifications and changes to specific objects or modules. While others may > differ, the product I've been most impressed with in this regard is StarTeam > from StarBase. (www.starbase.com -- no, I don't work for them). > > Furthermore, I find the claim that SNiFF is a "SourceCode Management system" > misleading. Granted, this may be a matter of semantics, but SNiFF doesn't > manage source code as much as it browses it. > > While the basic functionality of SNiFF may be "easy" so is the basic > functionality of PageMaker. But it doesn't mean anything. The real question, > in this regard, is how easy it is to fully leverage the tool. Conceding > that I didn't have the incentive to become a SNiFF expert, nevertheless, as > I stated before, I found it surprisingly non-intuitive. > > When you come out with your next version of SNiFF, I'll give it another try. > Though I would advise that you hire a Windows UI specialist and do some > genuine usability testing. > > I'd also recommend that you make the integration between SNiFF and Tornado a > little less obtuse. > > Joe > > > "Martin Raabe" wrote in message > news:9goe18$frm$05$1@news.t-online.com... > > Hello Joe, > > thank you for expressing your impressions so detailed. > > I will take your input back to the factory. > > > > My short statement to this is as following: > > What SNiFF+ is: > > - SNiFF+ is a SourceCode Management system, which handles SourceCode of > > different languages and Platforms. > > - SNiFF+ is a SDE (Software Development Environment) which allows to > > develop independent from the target platform. > > - SNiFF+ supports in these rols also Team Development, which is very > > helpfull in small and medium sized (1-100 developers) teams. > > - SNiFF+ is an easy to handle Tool which behaves the same on Unix, and > > Windows. > > - SNiFF+ allows to faster get into foreign or (own) older SourceCode. > > - SNiFF+ helps to live project strategies, individuals would not follow > > without SNiFF+. > > > > What SNiFF+ is NOT: > > - SNiFF+ is not a debugging tool for OOP-Languages, where runtime > behavior > > is a significant part of the debugging process. > > - SNiFF+ is not a Project Manager.It only adapts to VersionControl and > > Project Management Systems. > > > > I have a question for you. > > Which SNiFF+ Version are you referring to? > > I refer to the latest Version 4.0.1. > > I can understand that you do not feel that comfortable with the 2.x or > 3.0.x > > Versions. > > But the feedback from customers on Verison 3.2.1 and 4.0.x is excellent! > > > > So I am afraid SNiFf+ will make it into a huge project basis accross > > branches and markets! > > > > Ciao > > > > Martin Raabe > > > > > > "Joe Woodbury" schrieb im Newsbeitrag > > news:HS8X6.652$ZR1.297147@news.uswest.net... > > > There is no direct connection between SNiFF and team development. SNiFF > is > > > an object browser, not a team collaboration tool. SNiFF may facilitate > > large > > > project develpment, which is often done in a team environment, but it > > > completely lacks the other collaboration tools more commonly associated > > with > > > a team development environment. > > > > > > As far as particular criticisms of SNiFF go; my distaste of object > > browsers > > > is that they don't really clarify anything. You end up with massive > lists > > of > > > functions and classes (and data and function members), none of which > give > > > you much more than a clinical antiseptic view of the project. > > > > > > For example, say you look at list of member functions of a class and > it's > > > parents. That view tells you nothing about how the class works except by > > > intuition (i.e. experience in how objects work based on how they are > named > > > and how they are usually implemented.) But why should you choose one > > > overloaded member function over another? Should you derive a new class > and > > > overload a member function or add a virtual function? > > > > > > Beyond that, there is the actual implementation. While it may be "nice" > to > > > show every function and every define from every file, the shear volume > of > > > information (generally without context) solves nothing for me. There's a > > > define or typedef named something with a certain value; big deal. So now > I > > > won't create a define by that name, but I still don't know what that > > define > > > is specifically for. > > > > > > And if I want to find out the value of a specific define, there is > either > > > the documentation or search file (grep) functionality. The advantage of > > the > > > latter, which I can do very quickly, is that it tells me both the > context > > of > > > where it was defined, along with the associated defines which may prove > > even > > > more helpful, especially if that's what I really wanted in the first > > place, > > > and other places where it was used. > > > > > > That SNiFF can act as a giant index is not disputed, only that there is > no > > > compelling reason to use it for just that reason since otherwise, I > don't > > > find it a well written tool. It's awkward, uses screen space poorly, > > doesn't > > > allow common settings to be made or changed. > > > > > > Like the other object browsers I've tried, it doesn't work how I > > intuitively > > > want it to. Mainly, I have to do many clicks and take entirely > unintuitive > > > steps to traverse the project from an object perspective. On top of > that, > > > with the projects I'm working on, as with most projects in the real > world, > > > the objects themselves don't tell you much of anything. The browser may > > tell > > > you what exists and the programatic relationship of those items, but it > > > tells you nothing about how those objects interact. > > > > > > And that programatic relationship may be misleading; just because task A > > > calls the function to send a message to task B, doesn't mean it actually > > > does so in operation or that it may do so only sporadically. Unless you > > know > > > the code, the conclusions you draw from the browser may be very > > misleading. > > > If you know the code, then, as I said before, the browser in many > becomes > > > just a fancy grep engine. > > > > > > But that's not what's at issue. Are people REALLY using SNiFF, or are > they > > > buying it so their ignorant bosses can feel better about how they are > > > "managing" the project. (Or so they can convince themselves they can > hire > > > inexperienced developers and let them use SNiFF to understand what's > going > > > on.) > > > > > > (I've yet to work with a genuine [vs. those with just the title] Sr. > > > Developer/Engineer I respect who uses object browsers. I'm curious if my > > > experience is atypical.) > > > > > > > > > > > > > > > "Martin Raabe" wrote in message > > > news:9gijt8$abo$03$1@news.t-online.com... > > > > Hello Joe, > > > > I am Martin R. as Dr. Diag referred to, and I would say, that team > > > > development > > > > is increasingly coming into the embedded world, as is is since years > in > > > the > > > > non-embedded world. > > > > > > > > I am interrested, what of SNiFF+ 4.0.1 is not to your satisfaction? > > > > Please reply to me at: > > > > > > > > Martin.Raabe at BaSystem.de > > > > > > > > > > > > "Joe Woodbury" schrieb im Newsbeitrag > > > > news:y8yW6.736$s_.560883@news.uswest.net... > > > > > My local rep is pushing SNiFF+ at us for our next project. I used > > SNiFF+ > > > 2 > > > > > and didn't like it much, but gave the SNiFF+ 4.x demo a whirl to see > > if > > > it > > > > > could change my mind. I still didn't like it. However, with future > > > > > employment possibilities in mind, I'm wondering how many > > > > > developers/companies use SNiFF+ (or at least insist it be used)? > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: vxWorks.st - how to build Date: 20 Jun 2001 06:57:10 -0700 From: chrisanc1967@yahoo.com (Christopher Sanchez) Organization: http://groups.google.com/ Message-ID: References: <9gl7cc$24c@nntpa.cb.lucent.com> In tornado 2 window, seclect the options item from the tools menu and then under the projects tab check the "show Tornado 1.0.1 menu items" box. This will give you the "standard BSP Builds..." item in the Build menu, where you can select vxworks.st as your build. Hope this helps Christopher "Sean Bose" wrote in message news:<9gl7cc$24c@nntpa.cb.lucent.com>... > Hi > I am a newbie to the x86 flavour of VxWorks. I am trying to > compile and create a bootable floppy from which I can boot a > PC with vxWorks. > I understand that vxWorks.st is the standalone image from which > I can create a bootable floppy using mkboot.bat and vxsys.com. > I have Tornado 2.0 and I am unable to figure out how to make a > working vxWorks.st. Please suggest ways to compile / make > vxWorks.st. > > thanks to my helpful vxWorks friends > Sean --------------------------- Newsgroups: comp.os.vxworks Subject: Re: romInit in VxWorks 8260 BSP Date: 20 Jun 2001 07:45:01 -0700 From: sachin@blrmail.com (Sachin) Organization: http://groups.google.com/ Message-ID: References: Bill Pringlemeir wrote in message news:... > The code does jump from 0x100 (the reset vector on the PPC) to the > flash address which is probably `ROM_TEXT_ADRS'... It doesn't really > jump anywhere as the Flash will alias (address wraps when you try to > use a bigger size than memory). It just really changes the PC from > near 0x100 to near `ROM_TEXT_ADRS'; the code remains the same. <1> > This is where the `C' code has been linked and the PC must be set > properly or any `position dependant C code' would ~BOMB~! That isn't > good. After you have changed the PC to `ROM_TEXT_ADRS', the next step > is to map RAM to the zero address as this is the usual place for > vectors and you would like to be able to change these dynamically. > > The comment is sort of wrong, put it is mostly right...at that point I think that 0x100 is just the reset exception vector offset, and not an address by itself. This is added to 0x0000_0000/0xfff0_0000 depending on MSR[IP] bit of the 8260 and MUST map to Flash/Non-volatile memory space for the reset exception vector. But i understand what you meant by saying that the comment is partially right, and partially wrong.... > in time all of the address space is flash. It is difficult to put > all the concepts I have just explained in a few words. > > I suggest you get a PDF of the PowerPC assembler syntax. I am sure > that either IBM or Motorola had a PDF on the web somewhere. > > hth, > Bill Pringlemeir. > > <1> when your climbing a stair way to heaven... wrong alblum. Thanks, Sachin. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Wind Web Server 2.0, httpSymTabAdd function failure Date: Wed, 20 Jun 2001 08:56:57 -0700 From: Christian Kurzke Organization: Gatespace inc. Message-ID: <3B30C7C9.C2153C@gatespace.com> References: <2b1a2c27.0106180916.1ee5f6e3@posting.google.com> Emil Romascanu wrote: > > Has anybody experienced failures of httpSymTabAdd() calls ? I am > running tutorial/step_07 and the server does not work properly because > calls to that function return NULL. > > Thx, Emil Romascanu > Software developer > Intel Corporation could it be, that you exceedd the maximum (configurable) number of symbols? does it not run out of the box, or after you modified it?? (another idea: the requestpool memory could be too small?) - - chrisitan --------------------------- Newsgroups: comp.os.vxworks Subject: Re: strong ARM pointer align Date: 20 Jun 2001 16:24:55 GMT From: lied@nwsgpc.ih.lucent.com (Bob Lied) Organization: Lucent Technologies, Naperville, Illinois, USA Message-ID: <9gqion$hl9@nntpb.cb.lucent.com> References: <9gotf3$odp$1@overload.lbl.gov> Reply-To: lied@lucent.com In article <9gotf3$odp$1@overload.lbl.gov>, Jun Wu wrote: >Hi all, >I find if a pointer is not 4-byte align, data may be corrupted. For example >p = (int *) 0x08000002; >*p = 0xffffffff; If you're going to do that sort of thing, the compiler can't help you. Buried way down in an obscure description of load and store in the ARM reference manual is an explanation of how load and store works and why this doesn't do what you expect. The compiler does have alignment and packing options that force variables to long boundaries, but if you explicitly write to a non-aligned address, you have to read the fine print from ARM under LD and ST instructions. For unaligned structure members, the compiler is smart enough to take a long apart a byte at a time and write them individually, but for the sort of explicit write in your example, you got what you asked for. Perhaps your example will work if you use a short int: unsigned short *p; p = (unsigned short *)0x08000002; *p = 0xffff; *++p = 0xffff; --------------------------- Newsgroups: gnu.gcc.help,gnu.g++.help,comp.os.vxworks Subject: Re: global objects in VxWorks Date: 20 Jun 2001 09:35:46 -0700 From: igtorque@eliop.es (=?ISO-8859-1?Q?Ignacio_Gonz=E1lez?=) Organization: http://groups.google.com/ Message-ID: <9ec53d11.0106200835.731dfbc0@posting.google.com> References: <3B2F744D.7BF1613A@dlr.de> I've tested your program in my environment and it works OK. To be sure, I copy here the files I've tested: ProGloC.cpp: class A { public: A(){} int i; }; A myGlobalA; - -------- DriGloC.cpp: #include "iostream" class A { public: A(){} int i; }; extern A myGlobalA; void testit ( void ){ int i = myGlobalA.i; cout << i << endl; } I've used gcc 2.7.2, Tornado 2.0, VxWorks 5.4, an i386-based target. I've downloaded first ProGloC.o and then DriGloC.o nlm386 reports this: C:\USUARIOS\PRUEBAS\PruGloClass\PENTIUMgnu>nm386 ProGloC.o 00000000 T __GLOBAL_$I$myGlobalA 00000020 t ___1A 00000000 - 00 0000 (4) ___CTOR_LIST__ 00000000 t ___gnu_compiled_cplusplus 00000034 D _myGlobalA 00000000 t gcc2_compiled. C:\USUARIOS\PRUEBAS\PruGloClass\PENTIUMgnu>nm386 DriGloC.o 00000000 t ___gnu_compiled_cplusplus U ___ls__7ostreamPFR7ostream_R7ostream U ___ls__7ostreami U _cout U _endl__FR7ostream U _myGlobalA 00000000 T _testit__Fv 00000000 t gcc2_compiled. When launched, my host shell warns me always this: C++ Constructors/Destructors Strategy is AUTOMATIC Perhaps you downloaded the object modules in the wrong order? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How many END_TBL_ENTRY (entries) needed for 3 PHYs (each with 2 ports) Date: Wed, 20 Jun 2001 17:47:14 GMT From: Dan Gold Organization: http://www.newsranger.com Message-ID: References: <3B2FB18C.9E91892C@lucent.com> In article <3B2FB18C.9E91892C@lucent.com>, Babu Yama says... > >I am not clear on the endDevTbl->unit field? >Can some guru through some light on this, please? First, consider this a device major number (in Unix thinking). It must be unique, i.e. each device in your END table must have a uniqe unit number. Second, within your driver code, you can think of this as a minor number. But remember it must be unique. > >END_TBL_ENTRY endDevTbl[] = >{ 0, MOT_FCC_LOAD_FUNC, MOT_FCC_LOAD_STRING, 1, NULL, FALSE}; > >What does the endDevTbl->unit(the first field) refer to? This is the 'unit' number that's passed to your driver. If you have a driver (i.e. piece of code) you can use it as a minor number. >Does it refer to one PHY (physical layer >ethernet device) or one PORT on a PHY? How to set up the endDevTbl for multiple >PHYs (with some PHYs having multiple ports)? In other words, if I have 3 PHYS >(each PHY having 2 ports), how many >entries in endDevTbl do I need? In this situation, how many devices are there >under MUX and how many drivers? Should I have one driver instance per PHY or >have one driver that handles 3 PHYs and all their ports. >I guess, for the END drivers, the concept of driver and device is not clear to >me. It can be confusing. Consider your FCC ethernet driver. It is 1 piece of code, that has the xxxEnd{load, start, etc..} functions. Each routine (load, start...) will be called for each line in the END table. In the load function, it gets a unit number. You also allocate a driver control structure that stores that unit number. When any of the other functions (start, etc...) are called, either by the MUX or your driver, the driver control structure is passed, and you figure out which unit you're talking about by pDrvCtrl->unit. Remember that the MUX layer will associate an IP address with (effectively) a unit number. The MUX layer will figure out which interface a packet goes out on, get it's driver control structure, and pass it (and the mblk chain for the frame) to the send routine. >I understand that muxDevLoad() is called once for each of the entries >in endDevTbl, which results in driver's xxload call. If a PHY has multiple >ports, >how many times xxload is called by DEFAULT by vxworks ( I am using >wind river's est8260 bsp code)? Do I need to make any vxworks code changes to >supports multiple ports on a PHY or multiple PHYs? muxDevLoad() is called once for each line in the END table (except for the last one, ha ha :-)) I haven't looked at the WRS 8260 BSP in awhile, but for my own, since I have 3 FCCs, assuming unit numbers of 1, 2, 3, I'd have 3 entries in the END table that looked like: 1, FCC stuff 2, FCC stuff 3, FCC stuff. And my single xxxEndLoad() routine will be called 3 times, and each time with a different unit number. So 3 driver control structs will be allocated, each one holding a different unit number. When the MUX calls my xxxEndStart() function, the first parameter is the driver control structure, and the you look at the particular unit number (thereby treating it as a minor number) to see which FCC you're going to use. Note that you'll need different load strings for each FCC, because among other things, the load string carries the vector number, and each FCC interrupts on a different vector. hth, Dan Gold gold@ensemble.com --------------------------- Newsgroups: comp.os.vxworks Subject: re: strong ARM pointer align Date: Wed, 20 Jun 2001 18:47:22 +0100 From: "David Laight" Organization: Lawrence Berkeley National Laboratory Message-ID: <9gqp7m$sj7$1@overload.lbl.gov> > I use intel strongARM SA1110 assabet board and Tornado 2.0. > I find if a pointer is not 4-byte align, data may be corrupted. For example > /* set p point with a non-4-byte align address */ > p = (int *) 0x08000002; > *p = 0xffffffff; > bytes with address 0x08000000 and 0x08000001 are corrupted!*/ What do you expect to happen! Support for misaligned transfers is VERY processor specific. If you read the ARM V4 architecture guide you find there are two options for the ARM: 1) trap misaligned transfers 2) access the requested word and rotate the answer (puts the addressed byte into the lsb. Unfortunately you cannot enable the trap because gcc uses misaligned 32-bit transfers for some 16-bit requests. If you MUST support misaligned pointers, you have to code the byte reads and shifts yourself. Any compiler options are not portable and will seriously affect the performance of other code. In practise it is always more efficient to 'naturally' align all data items - even on sytems (like the x86) which support misaligned tranfers in the expected way. The reason for not 'supporting' misaligned transfers is simple, the hardware to generate two bus cycles and shift the results isn't the problem. The problems arise when the transfer crosses an MMU page boundary - and the second page isn't resident. Recovering from this is quite complicated. (Ever seem the size of the fault stack frames on a 68020 - required to restart things like the CMP2 instruction correctly). David --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How common is SNiFF usage? Date: 20 Jun 2001 11:14:26 -0700 From: joewoodbury@mail.com (Joe Woodbury) Organization: http://groups.google.com/ Message-ID: References: <9gijt8$abo$03$1@news.t-online.com> <9goe18$frm$05$1@news.t-online.com> <9gq81t$c2v$1@news.hdpp.de> My experience has been very similar to yours! The main responses I've gotten from WRS itself about SNiFF and many other tools are "look at the web site." Good heavens, they don't even have a comparison chart between their own Diab compiler and the GNU compiler. Neither do they have comparisons between SingleStep and GDB, Look! and SNiFF+ nor between PerformancePak and ScopePak. As for responses from other users on SNiFF+ 4.x, I've gotten none (save Martin, but that was still mostly sloganeering.) (The project I'm taking over was original prototyped in pSOS and those developers found SNiFF+ 2.x to be mostly useless.) VxWorks is the best RTOS we've evaluated and/or used but that doesn't mask a growing dislike for Wind River proper. To be fair, several employees of Wind River that I've spoken with are also incredibly frustrated. Management is running around buying stuff with little or no regard on integrating those purchases into the existing product lines. I'm thinking of buying stock in the company just so I can complain at shareholder meetings (and because the stock will go up if they ever organize the product lines properly.) Fortunately, while VxWorks has a 38% market share in the embedded market, Microsoft is making serious inroads with Windows CE. I say fortunately, because competition will force Wind River to perform better. (I don't use Macs or Linux, but I'm sure glad they're around for the same reason.) Joe "Matthias Guelck" wrote in message news:<9gq81t$c2v$1@news.hdpp.de>... > Hello Joe, > > I have been considering to evaluate the usage of SNiFF for our team. > Fortunately I have read this thread before. > From my point of view, WRS is able to sell high priced tools with very little benefit. > With Tornado we bought WindNavigator and LOOK!, because WRS announced them as very useful to understand source code written by other > people. > After we paid the money, no one at WRS was able to teach us, how to take benefit from the tools. > They only repeated the slogans from their colorful advertising booklets. > > Complaining about some debugger crashes (we have got hundreds of crash log files) and problems (default window size for watch, > variables,…), WRS told us, that the whole world is very satisfied, and we are the only people in the world complaining about their > famous products! > > So my question is: did you receive replies from other SNiFF users (non WRS employees like Martin Raabe)? > > Best regards > > Matthais > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: romInit in VxWorks 8260 BSP Date: Mon, 18 Jun 2001 19:29:27 -0400 From: "Sohacki, Timothy [NC1:GW55:EXCH]" Organization: Nortel Networks Message-ID: <3B2E8ED7.3A0BA131@americasm01.nt.com> References: Sachin, The computed branches exist in the boot ROM because the boot ROM code is linked and located to be executed at RAM_HIGH_ADRS. The initial boot ROM code (e.g., romInit.s) is linked relative to ROM_HIGH_ADRS, instead of ROM_TEXT_ADRS. The code below "fixes up" the program counter to the area where the Flash resides. Note: A typical boot ROM executes mostly from RAM, beginning at (well, near) RAM_HIGH_ADRS. Only a minimal amount of code actually executes from Flash, code enough to configure the processor and the memory controller, and then to boot strap the rest of the code to RAM. The need for a fix-up can be seen if you desire to branch to a subroutine in romInit.s. For example, if you try to add a branch to a memory test without "fixing up" the destination address, the code may jump off to somewhere near RAM_HIGH_ADRS instead of somewhere near ROM_TEXT_ADRS. In the case of the 0x100 vector, while the branch logically continues to the next line of code at the "start" label, it is necessary because the PowerQUICC may not be running with the "correct" program counter, i.e., offset from ROM_TEXT_ADRS, at this point. This fix-up ensures the program will continue at the ROM_TEXT_ADRS appropriate to your application. (I think this might have something to do with whether the PowerPC starts with the reset vector at 0x0000_0100 or 0xFFF0_0100.) Cheers, Tim. Sachin wrote: > I am trying to decipher the init. code of 8260 in the Vxworks BSP & i > find this in romInit function which is the initial entry point in the > Flash space. > > ***************************************************************************** > /* > * When the PowerPC 8260 is powered on, the processor fetches the > * instructions located at the address 0x100. We need to jump > * from the address 0x100 to the Flash space. > */ > > lis r4, HIADJ(start) /* load r4 with the address */ > addi r4, r4, LO(start) /* of start */ > lis r5, HIADJ(romInit) /* load r5 with the address */ > addi r5, r5, LO(romInit) /* of romInit() */ > lis r6, HIADJ(ROM_TEXT_ADRS) /* load r6 with the address */ > addi r6, r6, LO(ROM_TEXT_ADRS) /* of ROM_TEXT_ADRS */ > sub r4, r4, r5 /* */ > add r4, r4, r6 > mtspr LR, r4 /* save destination address*/ > /* into LR register */ > blr /* jump to flash mem address */ > start: > **************************************************************************** > What do the comments signify & what is this piece of code doing?? To > me, it seems like the code after initialising the IMMR to 0x4700_0000 > and sundry other registers jumps from Flash space (from where this > piece of code is running) BACK into Flash space, infact into the very > next line (viz. the "start" label). Since the romInit function gets > included only in the bootrom image & not in the downloadable image, > the whole thing runs only in Flash space, does it not?? If so, isn't > the whole thing useless exercise?? > Thanks, > Sachin. . . . --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Newbie question: Including spyLib in application code (or project) Date: Tue, 19 Jun 2001 19:57:20 -0400 From: "Sohacki, Timothy [NC1:GW55:EXCH]" Organization: Nortel Networks Message-ID: <3B2FE6E0.E45FE34F@americasm01.nt.com> References: Christopher, > I need to include the spyLib utility in order to profile some of my > application code. Now, being new to Tornado, I was curious as to > where in my project I add the configuration macro (INCLUDE_SPY)? The "old" style approach is to add the following line to config.h: #define INCLUDE_SPY The "new" approach is to use the project facility. In your Workspace, select the appropriate project, and click the VxWorks tab. Then, expand "development tool components" and select "spy" ... right mouse click, include 'spy,' and "OK." Cheers, Tim. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How common is SNiFF usage? Date: 20 Jun 2001 15:13:57 -0400 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: References: <9gijt8$abo$03$1@news.t-online.com> <9goe18$frm$05$1@news.t-online.com> <9gq81t$c2v$1@news.hdpp.de> Sender: bpringlemeir@DeadDuck >>>>> "Joe" == Joe Woodbury writes: Joe> My experience has been very similar to yours! The main responses Joe> I've gotten from WRS itself about SNiFF and many other tools are Joe> "look at the web site." Good heavens, they don't even have a Joe> comparison chart between their own Diab compiler and the GNU [snip] There are several things at issue here. 1) User base 2) Product maturity 3) General stupidity The vxWorks kernel gets the most use and it is pretty bullet proof in my experience. Most problems are due to improper linking and memory setup. However, if you look at the vxSim and vxVMI products, they were also fairly bug ridden upon introduction. This is similar with ugl/windml and DosFs. I do think that these products are ok at the current point in time. I think that this illustrates that the first two have an influence. That said, WRS has constantly lagged behind with its versions of gcc. I guess that this is a stability issue, but I would think that 1997 versions of tools are a little out of date. I used the Diab compiler on a vxWorks PPC target before any acquisition. It was difficult to get working due to some va_arg mess. Currently there are too many items on WRS plate. It seems like they go after "new products/markets" like PLCs. This is frustrating as it seems like they have abandoned the core functions of the OS... it has been rather stagnant and it is certainly frustrating to get bugs fixed. Which leads me to the question "Why do they care *WHO* reports a bug in their product"? Obviously many will claim bugs that are just user errors. However, how is a `selected individual' less likely to claim a user error is a bug? I hate the license verification! I just don't report bugs because the process is too cumbersome. My FAE is currently very good, but this wasn't always the case. Oh well... Joe> VxWorks is the best RTOS we've evaluated and/or used but that Joe> doesn't mask a growing dislike for Wind River proper. To be Joe> fair, several employees of Wind River that I've spoken with are Joe> also incredibly frustrated. Management is running around buying ibid... CE is garbage; Epoch is probably far better for PDA/cell phones. I think that they are more concerned with [Embedded] Linux than anything else. This takes away from the router/switch market and is fairly big volume. fwiw, Bill Pringlemeir. - -- .. I feel.. JUGULAR.. Useful vxWorks URLS, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" "http://www.xs4all.nl/~borkhuis/vxworks/vxw_pt9.html" --------------------------- Newsgroups: comp.os.vxworks Subject: mvme162 or 172 with Petra asic problem? Date: Wed, 20 Jun 2001 15:50:54 -0400 From: Michael Reed Message-ID: <3B30FE9E.76C85F42@ll.mit.edu> Hello, Is anyone having a problem with IP packs on mvme162s or mvme172s with the newer Petra asic? The IP2 chip's functions were added into a comprehensive asic, the "Petra". On the old mvme162 boards, I start this IP pack up, can write to it, and it runs fine. On the new mvme162 w/ Petra asic, the IP pack comes up ok, starts to function, but generates a bus-error on on the 8th or 9th time I write it. Thanks, Mike (IP pack is the IP-Thermistor from GreenSpring/SBS.) --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How common is SNiFF usage? Date: Wed, 20 Jun 2001 22:16:40 +0200 From: "Martin Raabe" Organization: T-Online Message-ID: <9gr0cn$sqr$03$1@news.t-online.com> References: <9gijt8$abo$03$1@news.t-online.com> <9goe18$frm$05$1@news.t-online.com> Hello Joe, thank you for you open words. I want to talk to you direcctly.but your Email Addreass is not reachable. I am heavily interessted in the meaning of the last sentence of your posting, which was: > I'd also recommend that you make the integration between > SNiFF and Tornado a little less obtuse. What do you mean especially is obtuse in this integration. I would love to talk to you on the phone. Is this possible? Are you available somehow? Please get in contact with me directly, so we can continue our conversation offline. My Email is: Martin.Raabe at WindRiver.com Ciao - -- Martin Raabe --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How common is SNiFF usage? Date: Wed, 20 Jun 2001 22:38:00 +0200 From: "Martin Raabe" Organization: T-Online Message-ID: <9gr1k7$g3k$04$1@news.t-online.com> References: <9gijt8$abo$03$1@news.t-online.com> <9goe18$frm$05$1@news.t-online.com> <9gq81t$c2v$1@news.hdpp.de> Hello Matthias, maybe you do not know, but there is a majordomo based UN-moderated email list, where SNiFF+ users meet SNiFF+ users and SNiFF+ developers. sniff-list@takefive.co.at. To subscribe look at: www.takefive.co.at or www.windriver.com/windsurf Section Embedded Development Tools -> SNiFF+ HIH Martin Raabe "Matthias Guelck" schrieb im Newsbeitrag news:9gq81t$c2v$1@news.hdpp.de... > Hello Joe, > > I have been considering to evaluate the usage of SNiFF for our team. > Fortunately I have read this thread before. > From my point of view, WRS is able to sell high priced tools with very little benefit. > With Tornado we bought WindNavigator and LOOK!, because WRS announced them as very useful to understand source code written by other > people. > After we paid the money, no one at WRS was able to teach us, how to take benefit from the tools. > They only repeated the slogans from their colorful advertising booklets. > > Complaining about some debugger crashes (we have got hundreds of crash log files) and problems (default window size for watch, > variables,…), WRS told us, that the whole world is very satisfied, and we are the only people in the world complaining about their > famous products! > > So my question is: did you receive replies from other SNiFF users (non WRS employees like Martin Raabe)? > > Best regards > > Matthais > > > > "Joe Woodbury" wrote in message news:EIWX6.2104$zE1.780634@news.uswest.net... > > I initially, and very briefly, used SNiFF 2.x. Most recently I was testing > > 4.0.1. > > > > Today, we chose to not purchase SNiFF+ for inclusion with our Tornado Pro > > package, so that issue is closed for now. However, I do beg to differ with > > you on some of your points: > > > > Unless any editor, IDE or even any tool which interfaces to a version > > control system has suddenly been redefined as a "Team Development" tool (in > > capital letters), SNiFF most certainly is not. Again, it may facilitate > > development in a team environment, but it lacks the collaboration tools of a > > true "Team Development" tool--for example, being able to tie design > > specifications and changes to specific objects or modules. While others may > > differ, the product I've been most impressed with in this regard is StarTeam > > from StarBase. (www.starbase.com -- no, I don't work for them). > > > > Furthermore, I find the claim that SNiFF is a "SourceCode Management system" > > misleading. Granted, this may be a matter of semantics, but SNiFF doesn't > > manage source code as much as it browses it. > > > > While the basic functionality of SNiFF may be "easy" so is the basic > > functionality of PageMaker. But it doesn't mean anything. The real question, > > in this regard, is how easy it is to fully leverage the tool. Conceding > > that I didn't have the incentive to become a SNiFF expert, nevertheless, as > > I stated before, I found it surprisingly non-intuitive. > > > > When you come out with your next version of SNiFF, I'll give it another try. > > Though I would advise that you hire a Windows UI specialist and do some > > genuine usability testing. > > > > I'd also recommend that you make the integration between SNiFF and Tornado a > > little less obtuse. > > > > Joe > > > > > > "Martin Raabe" wrote in message > > news:9goe18$frm$05$1@news.t-online.com... > > > Hello Joe, > > > thank you for expressing your impressions so detailed. > > > I will take your input back to the factory. > > > > > > My short statement to this is as following: > > > What SNiFF+ is: > > > - SNiFF+ is a SourceCode Management system, which handles SourceCode of > > > different languages and Platforms. > > > - SNiFF+ is a SDE (Software Development Environment) which allows to > > > develop independent from the target platform. > > > - SNiFF+ supports in these rols also Team Development, which is very > > > helpfull in small and medium sized (1-100 developers) teams. > > > - SNiFF+ is an easy to handle Tool which behaves the same on Unix, and > > > Windows. > > > - SNiFF+ allows to faster get into foreign or (own) older SourceCode. > > > - SNiFF+ helps to live project strategies, individuals would not follow > > > without SNiFF+. > > > > > > What SNiFF+ is NOT: > > > - SNiFF+ is not a debugging tool for OOP-Languages, where runtime > > behavior > > > is a significant part of the debugging process. > > > - SNiFF+ is not a Project Manager.It only adapts to VersionControl and > > > Project Management Systems. > > > > > > I have a question for you. > > > Which SNiFF+ Version are you referring to? > > > I refer to the latest Version 4.0.1. > > > I can understand that you do not feel that comfortable with the 2.x or > > 3.0.x > > > Versions. > > > But the feedback from customers on Verison 3.2.1 and 4.0.x is excellent! > > > > > > So I am afraid SNiFf+ will make it into a huge project basis accross > > > branches and markets! > > > > > > Ciao > > > > > > Martin Raabe > > > > > > > > > "Joe Woodbury" schrieb im Newsbeitrag > > > news:HS8X6.652$ZR1.297147@news.uswest.net... > > > > There is no direct connection between SNiFF and team development. SNiFF > > is > > > > an object browser, not a team collaboration tool. SNiFF may facilitate > > > large > > > > project develpment, which is often done in a team environment, but it > > > > completely lacks the other collaboration tools more commonly associated > > > with > > > > a team development environment. > > > > > > > > As far as particular criticisms of SNiFF go; my distaste of object > > > browsers > > > > is that they don't really clarify anything. You end up with massive > > lists > > > of > > > > functions and classes (and data and function members), none of which > > give > > > > you much more than a clinical antiseptic view of the project. > > > > > > > > For example, say you look at list of member functions of a class and > > it's > > > > parents. That view tells you nothing about how the class works except by > > > > intuition (i.e. experience in how objects work based on how they are > > named > > > > and how they are usually implemented.) But why should you choose one > > > > overloaded member function over another? Should you derive a new class > > and > > > > overload a member function or add a virtual function? > > > > > > > > Beyond that, there is the actual implementation. While it may be "nice" > > to > > > > show every function and every define from every file, the shear volume > > of > > > > information (generally without context) solves nothing for me. There's a > > > > define or typedef named something with a certain value; big deal. So now > > I > > > > won't create a define by that name, but I still don't know what that > > > define > > > > is specifically for. > > > > > > > > And if I want to find out the value of a specific define, there is > > either > > > > the documentation or search file (grep) functionality. The advantage of > > > the > > > > latter, which I can do very quickly, is that it tells me both the > > context > > > of > > > > where it was defined, along with the associated defines which may prove > > > even > > > > more helpful, especially if that's what I really wanted in the first > > > place, > > > > and other places where it was used. > > > > > > > > That SNiFF can act as a giant index is not disputed, only that there is > > no > > > > compelling reason to use it for just that reason since otherwise, I > > don't > > > > find it a well written tool. It's awkward, uses screen space poorly, > > > doesn't > > > > allow common settings to be made or changed. > > > > > > > > Like the other object browsers I've tried, it doesn't work how I > > > intuitively > > > > want it to. Mainly, I have to do many clicks and take entirely > > unintuitive > > > > steps to traverse the project from an object perspective. On top of > > that, > > > > with the projects I'm working on, as with most projects in the real > > world, > > > > the objects themselves don't tell you much of anything. The browser may > > > tell > > > > you what exists and the programatic relationship of those items, but it > > > > tells you nothing about how those objects interact. > > > > > > > > And that programatic relationship may be misleading; just because task A > > > > calls the function to send a message to task B, doesn't mean it actually > > > > does so in operation or that it may do so only sporadically. Unless you > > > know > > > > the code, the conclusions you draw from the browser may be very > > > misleading. > > > > If you know the code, then, as I said before, the browser in many > > becomes > > > > just a fancy grep engine. > > > > > > > > But that's not what's at issue. Are people REALLY using SNiFF, or are > > they > > > > buying it so their ignorant bosses can feel better about how they are > > > > "managing" the project. (Or so they can convince themselves they can > > hire > > > > inexperienced developers and let them use SNiFF to understand what's > > going > > > > on.) > > > > > > > > (I've yet to work with a genuine [vs. those with just the title] Sr. > > > > Developer/Engineer I respect who uses object browsers. I'm curious if my > > > > experience is atypical.) > > > > > > > > > > > > > > > > > > > > "Martin Raabe" wrote in message > > > > news:9gijt8$abo$03$1@news.t-online.com... > > > > > Hello Joe, > > > > > I am Martin R. as Dr. Diag referred to, and I would say, that team > > > > > development > > > > > is increasingly coming into the embedded world, as is is since years > > in > > > > the > > > > > non-embedded world. > > > > > > > > > > I am interrested, what of SNiFF+ 4.0.1 is not to your satisfaction? > > > > > Please reply to me at: > > > > > > > > > > Martin.Raabe at BaSystem.de > > > > > > > > > > > > > > > "Joe Woodbury" schrieb im Newsbeitrag > > > > > news:y8yW6.736$s_.560883@news.uswest.net... > > > > > > My local rep is pushing SNiFF+ at us for our next project. I used > > > SNiFF+ > > > > 2 > > > > > > and didn't like it much, but gave the SNiFF+ 4.x demo a whirl to see > > > if > > > > it > > > > > > could change my mind. I still didn't like it. However, with future > > > > > > employment possibilities in mind, I'm wondering how many > > > > > > developers/companies use SNiFF+ (or at least insist it be used)? > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: writing a Shell for vxSim Date: Wed, 20 Jun 2001 23:39:58 GMT From: "not really" Organization: Road Runner Message-ID: Friends, I am trying to write a "Shell" for vxSim that will run from "tShell". I want to run specific set of commands from the shell and if the issued command does not match any of mine it should be passed on to tShell. Any idea / sample is HIGHLY appreciated. Keeping tShell at priority 2 while myShell at 65, will it cause any problem ? - ->myShell myShell> myCmd value = 0 myShell> i *task list reported from tShell. PS: I do not have network package .. so u know ... no sockets or telnet etc.. Thanks, Bibhuti --------------------------- Newsgroups: comp.os.vxworks Subject: RE: DHCP problems on vxWorks Date: Wed, 20 Jun 2001 16:31:39 -0700 From: "Crozier, Bruce" Organization: Lawrence Berkeley National Laboratory Message-ID: <9great$8sk$1@overload.lbl.gov> Hello Serg, If you are running Tornado 2.0.2, then you have to use dhcpcOptionSet(pCookie,TAG) to request that option TAG be sent to you by the dhcp server. Then, you must create a structure to store the parameters that you want, and attach it to the bootParams structure, so that the dhcp routine will know to copy the parameter from the return packet into your bootParams structure. In bootConfig.c, dhcpGet() routine, they clear the bootParams structure using bzero(), which NULLs out all of the structure pointers. You then put back those that you want. For example, if you want to get the hostname parameter, char dhcp_hostName[BOOT_HOST_LEN]; /* create new structure to store host name */ bootParams.hostname = dhcp_hostName; dhcpcOptionSet(pCookie, _DHCP_HOSTNAME_TAG); if(bootParams.hostname[0] != EOS) printf(" Target Name: %s\n\r",bootParams.hostname); Regards, Bruce - -----Original Message----- From: Serg [mailto:sergey@congruency.co.il] Sent: Wednesday, June 20, 2001 10:38 AM Hi! Please, could you help me? I wanna get not standard parameters (for example 121,122) from DHCP server. I can add non-standard options by the dhcpcBootReqSpec variable, but I havn't found way to get this options. I know that server send it. If I use low-level functions (dhcpBind, not dhcpBootBind) I can access to full offer by the dhcpcMsgIn variable way, but can't add non-standard parameters to Request becouse dhcpcBootReqSpec isn't used. Thanks so much. Serg. --------------------------- Newsgroups: comp.os.vxworks Subject: Reserving Memory Date: Thu, 21 Jun 2001 01:31:13 GMT From: pat.dirienzo@verizon.net Message-ID: <3B314F83.121AE0AD@verizon.net> Reply-To: pat.dirienzo@verizon.net Hi all: I am trying to put data structures, arrays etc in a memory location of my choice. In the old operating system, we did this with a linker directive, but I can't seem to be able to do the same with vxWorks GNU linker. Is there another way to easily accomplish this? Again, the idea is to put an array or many arrays starting at a desired memory location. The memory location would be in 24 bit address space, that is below 2^24 as an address Thanks for reading --------------------------- Newsgroups: comp.os.vxworks Subject: writing a "Shell" to run on top of "tShell" of vxSim Date: Thu, 21 Jun 2001 02:11:31 GMT From: "not really" Organization: Road Runner Message-ID: For some reason my prev mail did not make it to the news group. Friends, I am trying to write a Shell that I can run with my own command plugins and anything unmatched will be passed to tShell. for example: - -> myShell my> myCmd Hello World value 0x0 my>lkup "badboy" badboy text 0x776399849 /*Passed to tShell for parsing*/ my>exit - -> /*back to tShell*/ I ve vxSim with NO network package. Is it possible to write such a Shell ? What FDs to be used for in/out/error? What should be the priority of tShell when myShell is running ? Thanks a lot for any hint/suggestion/sample code. - -Bibhuti --------------------------- Newsgroups: comp.os.vxworks Subject: writing a "Shell" to run on top of "tShell" of vxSim Date: Thu, 21 Jun 2001 02:16:10 GMT From: "not really" Organization: Road Runner Message-ID: For some reason my prev mail did not make it to the news group. Friends, I am trying to write a Shell that I can run with my own command plugins and anything unmatched will be passed to tShell. for example: - -> myShell my> myCmd Hello World value 0x0 my>lkup "badboy" badboy text 0x776399849 /*Passed to tShell for parsing*/ my>exit - -> /*back to tShell*/ I ve vxSim with NO network package. Is it possible to write such a Shell ? What FDs to be used for in/out/error? What should be the priority of tShell when myShell is running ? Thanks a lot for any hint/suggestion/sample code. - -Bibhuti --------------------------- Newsgroups: comp.os.vxworks Subject: writing a "Shell" to run on top of "tShell" of vxSim Date: Thu, 21 Jun 2001 02:16:24 GMT From: "not really" Organization: Road Runner Message-ID: For some reason my prev mail did not make it to the news group. Friends, I am trying to write a Shell that I can run with my own command plugins and anything unmatched will be passed to tShell. for example: - -> myShell my> myCmd Hello World value 0x0 my>lkup "badboy" badboy text 0x776399849 /*Passed to tShell for parsing*/ my>exit - -> /*back to tShell*/ I ve vxSim with NO network package. Is it possible to write such a Shell ? What FDs to be used for in/out/error? What should be the priority of tShell when myShell is running ? Thanks a lot for any hint/suggestion/sample code. - -Bibhuti --------------------------- Newsgroups: comp.os.vxworks Subject: SOS - how to download/upload files from/to flash memory? Date: Thu, 21 Jun 2001 00:06:44 -0500 From: "Lee" Message-ID: <9grurv$astd9$1@ID-92003.news.dfncis.de> Hi, This is a kind of urgent. It would be highly appreciated if someone could help me on this. I'm not very familiar with vxworks. - - Environment 1. We use vxworks on a hardware device which has a big flash memory. a big firmware file is uploaded into the flash memory. This firmware file, I assume, contains both application code and vxworks (am I right?). It also includes a few web pages (htm files). 2. We can use a web browser on the host PC to access vxworks' web server inside the flash memory, and open those web pages. 3. We can also use the NT's ftp client on the host PC to access vxworks' ftp server inside the flash memory to download / upload a file from flash memory into host PC / from host PC into flash memory without any problem. - - Qustion How to use a web browser to 1. download a file from the flash memory into the host PC. 2. upload a file from the host PC into the flash memory. We can use NT's ftp client to do it, but we do need to use a web browser to do file downloading/uploading. Thank you in advance. Lee --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Cannot start the server Date: 20 Jun 2001 23:28:30 -0700 From: i_satya@yahoo.com (satya kumar) Organization: http://groups.google.com/ Message-ID: References: <4b0sit00pab53vdo9i10erp3t65svjo3mc@4ax.com> Hi, I have already done that checked with it ..it is still giving the same problem.Please help me out. Satya --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Cannot start the server Date: Thu, 21 Jun 2001 10:19:47 +0200 From: Amit Raam Organization: Verio Message-ID: <2bb3jtk214m9rcbo45asg3m93rgv3ov15u@4ax.com> References: <4b0sit00pab53vdo9i10erp3t65svjo3mc@4ax.com> i_satya@yahoo.com (satya kumar) wrote: >Hi, > I have already done that checked with it ..it is still giving the >same problem.Please help me out. >Satya Thing is, you need to put your vxWorks at the same place where you have built it, since this information is INSIDE the vxworks image. you can't build it in C:\BUILD, move it to C:\TARGET and point the target server the C:\TARGET. this doesn't work. Hope this helps, Amit Raam Email address: raama at indigo.co.il - -------------------------------------------------- ".sigless and proud of it" (R) (C) (tm) (look-and-feel) Std. disclaimer:I don't represent the opinions of anyone but myself. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How common is SNiFF usage? Date: Thu, 21 Jun 2001 09:28:50 +0200 From: "Matthias Guelck" Organization: Heidelberg Digital Message-ID: <9gs7nq$gdr$1@news.hdpp.de> References: <9gijt8$abo$03$1@news.t-online.com> <9goe18$frm$05$1@news.t-online.com> <9gq81t$c2v$1@news.hdpp.de> Hello Joe, all WRS FAEs told me, that I am the only person in the world complaining about WRS products. Again it was very important for me to see, that I am NOT alone! (I was close to it to believe) Do you (or others) know about a forum like “complaints about WRS”? You mentioned that “VxWorks is the best RTOS we've evaluated and/or used”. Which other RTOS have you evaluated and/or used? I am asking this, because we are currently searching for an alternative RTOS like ecos, embedded Linux or something other with open source. Are you going to stay with WRS products in spite of all your dissatisfaction? Best regards Matthias "Joe Woodbury" wrote in message news:c8c13723.0106201014.22bf49fe@posting.google.com... > My experience has been very similar to yours! The main responses I've > gotten from WRS itself about SNiFF and many other tools are "look at > the web site." Good heavens, they don't even have a comparison chart > between their own Diab compiler and the GNU compiler. Neither do they > have comparisons between SingleStep and GDB, Look! and SNiFF+ nor > between PerformancePak and ScopePak. > > As for responses from other users on SNiFF+ 4.x, I've gotten none > (save Martin, but that was still mostly sloganeering.) (The project > I'm taking over was original prototyped in pSOS and those developers > found SNiFF+ 2.x to be mostly useless.) > > VxWorks is the best RTOS we've evaluated and/or used but that doesn't > mask a growing dislike for Wind River proper. To be fair, several > employees of Wind River that I've spoken with are also incredibly > frustrated. Management is running around buying stuff with little or > no regard on integrating those purchases into the existing product > lines. I'm thinking of buying stock in the company just so I can > complain at shareholder meetings (and because the stock will go up if > they ever organize the product lines properly.) > > Fortunately, while VxWorks has a 38% market share in the embedded > market, Microsoft is making serious inroads with Windows CE. I say > fortunately, because competition will force Wind River to perform > better. (I don't use Macs or Linux, but I'm sure glad they're around > for the same reason.) > > Joe > > > "Matthias Guelck" wrote in message news:<9gq81t$c2v$1@news.hdpp.de>... > > Hello Joe, > > > > I have been considering to evaluate the usage of SNiFF for our team. > > Fortunately I have read this thread before. > > From my point of view, WRS is able to sell high priced tools with very little benefit. > > With Tornado we bought WindNavigator and LOOK!, because WRS announced them as very useful to understand source code written by other > > people. > > After we paid the money, no one at WRS was able to teach us, how to take benefit from the tools. > > They only repeated the slogans from their colorful advertising booklets. > > > > Complaining about some debugger crashes (we have got hundreds of crash log files) and problems (default window size for watch, > > variables,…), WRS told us, that the whole world is very satisfied, and we are the only people in the world complaining about their > > famous products! > > > > So my question is: did you receive replies from other SNiFF users (non WRS employees like Martin Raabe)? > > > > Best regards > > > > Matthais > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: romInit in VxWorks 8260 BSP Date: Thu, 21 Jun 2001 09:54:30 +0100 From: Graham Baxter Organization: Graham Baxter (Software) Limited Message-ID: <3B31B646.2355C148@NOSPAM.bcs.org.uk> References: The trick is that when the 8260 boots up it has a single very large (128MB) chip select for the ROM boot device. Where the memory for this chip select is located is found from the hard reset configuration word located at the start of the boot device. It can either be located at: 0xFE000000 or 0x00000000 The first instruction 0x100 relative to this address. If the boot device is of size say 8MB (0x00800000) all the addresses: 0xFE000100 0xFE800100 0xFF000100 0xFF800100 0x00000100 etc. all refer to the same physical location until chip select zero is programmed with its correct size as the memory appears to wrap. If code is linked at ROM_BASE_ADRS = 0xFE000000 but we boot from vector 0x100: The trick is to jump from the running address say 0x138 to the linked next instruction address 0xFE00013C. Once we are running at the linked address, chip select zero can be programmed to its correct size so the memory wrapping will no longer occur. Regards, Graham Baxter Freelance Software Engineer gbaxter@NOSPAM.bcs.org.uk Sachin wrote: > > Bill Pringlemeir wrote in message news:... > > The code does jump from 0x100 (the reset vector on the PPC) to the > > flash address which is probably `ROM_TEXT_ADRS'... It doesn't really > > jump anywhere as the Flash will alias (address wraps when you try to > > use a bigger size than memory). It just really changes the PC from > > near 0x100 to near `ROM_TEXT_ADRS'; the code remains the same. <1> > > This is where the `C' code has been linked and the PC must be set > > properly or any `position dependant C code' would ~BOMB~! That isn't > > good. After you have changed the PC to `ROM_TEXT_ADRS', the next step > > is to map RAM to the zero address as this is the usual place for > > vectors and you would like to be able to change these dynamically. > > > > The comment is sort of wrong, put it is mostly right...at that point > > I think that 0x100 is just the reset exception vector offset, and not > an address by itself. This is added to 0x0000_0000/0xfff0_0000 > depending on MSR[IP] bit of the 8260 and MUST map to > Flash/Non-volatile memory space for the reset exception vector. But i > understand what you meant by saying that the comment is partially > right, and partially wrong.... > > > in time all of the address space is flash. It is difficult to put > > all the concepts I have just explained in a few words. > > > > I suggest you get a PDF of the PowerPC assembler syntax. I am sure > > that either IBM or Motorola had a PDF on the web somewhere. > > > > hth, > > Bill Pringlemeir. > > > > <1> when your climbing a stair way to heaven... wrong alblum. > > Thanks, > Sachin. - -- --------------------------- Newsgroups: comp.os.vxworks Subject: TFFS limitations? Date: Thu, 21 Jun 2001 10:25:54 +0100 From: Owain Phillips Organization: Siemens Inc. Message-ID: <3B31BDA2.AEEA9B89@siemenscomms.co.uk> This is a multi-part message in MIME format. - --------------D367171D689B7DCCCB966150 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hi, I have an 8MB device and the architectural limitation that the boot image has to lie at 4MB. With TFFS can I reserve arbitary regions of the prom that are not to be used in filesystems? IE. Can I build a filesystem in that 4MB below the bootimage? Or can I treat the 8MB device as though it is several devices and hence use the lover 4MB as FSys? Or must I reserve 4MB+bootimageSize and have a filesystem above the bootimage? I look forward to your responses... Regards, Owain Phillips - --------------D367171D689B7DCCCB966150 Content-Type: text/x-vcard; charset=us-ascii; name="owain.phillips.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Owain Phillips Content-Disposition: attachment; filename="owain.phillips.vcf" begin:vcard n:Phillips;Owain tel;fax:+44 115 943 4969 tel;work:+44 115 943 2167 x-mozilla-html:TRUE url:http://www.siemenscomms.co.uk org:Siemens Comunications Ltd version:2.1 email;internet:owain.phillips@siemenscomms.co.uk title:Software Engineer adr;quoted-printable:;;Technology Drive=0D=0ABeeston;Nottingham;;NG9 1LA;United Kingdom fn:Owain Phillips end:vcard - --------------D367171D689B7DCCCB966150-- --------------------------- Newsgroups: comp.os.vxworks Subject: system tick rate Date: Thu, 21 Jun 2001 17:54:14 +0800 From: "Jun Wu" Organization: Lawrence Berkeley National Laboratory Message-ID: <9gshfi$2c4$1@overload.lbl.gov> Hello vxWorks experts, I use watch dog implement a periodic event in my application. Since the event should occur every 6ms, I have to change the system tick rate to 166. Is it safe that 166 can be divide exactly by 1000? Will the remaining 4ms (1000 - 166*6 = 4) accumulate and affect the vxWorks? BR --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Command line syntax for compiling Tornado projects Date: Thu, 21 Jun 2001 18:19:06 +0800 From: "Schubert Zhang" Organization: The news server of Changzhou TeleCom Message-ID: <9gshht$25u3$1@news.cz.js.cn> References: You must write the makefile by yourself and then exe. "make". Schubert "Langston" wrote in message news:vOvV6.1426$FS6.72017@news.pacbell.net... > Hello All, > > Currently we are using Tornado 2.0 to build our VxWorks applications. > We > generally load .wsp (Tornado Workspace Files) to the Tornado IDE and create > our > builds from there. I would like to be able to script this build process but > if I use a 'Make' command on the Makefile then I run the risk of not getting > all of the latest changes because this Makefile is generated by the > Workspace and Project files. Is there a command line for building > applications from the project files? If I could generate the Makefile from > the command line then perform a Make on it (so that I have the current > Makefile)? > > Thank you, > Langston > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Cannot start the server Date: Thu, 21 Jun 2001 12:15:04 +0200 From: "Michael Lawnick" Organization: Buergernetz Dillingen Message-ID: <9gshlb$1ge$1@snoopy.bndlg.de> References: <4b0sit00pab53vdo9i10erp3t65svjo3mc@4ax.com> <2bb3jtk214m9rcbo45asg3m93rgv3ov15u@4ax.com> Reply-To: "Michael Lawnick" Sender: mlawnick@pd95249e8.dip.t-dialin.net "Amit Raam" schrieb im Newsbeitrag news:2bb3jtk214m9rcbo45asg3m93rgv3ov15u@4ax.com... > i_satya@yahoo.com (satya kumar) wrote: > > >Hi, > > I have already done that checked with it ..it is still giving the > >same problem.Please help me out. > >Satya > > Thing is, you need to put your vxWorks at the same place where you have built > it, since this information is INSIDE the vxworks image. you can't build it in > C:\BUILD, move it to C:\TARGET and point the target server the C:\TARGET. this > doesn't work. > Hope this helps, Wrong! Target boots the image given in bootline. Target Server searches either - - on the path received from target (i.e. bootline) - - on a (host side) local path When target boots from target's harddrive, target server must boot a copy on host as it has no access to targets filesystem. --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Thu Jun 21 06:58:04 2001 From: Bibhuti Kar Date: Thu Jun 21 06:58:06 PDT 2001 Subject: vxWorks - developing a Shell on vxSim w/ no network package --=====================_244998921==_.ALT Content-Type: text/plain; charset="us-ascii" Hi, I am trying to write a Shell on vxSim, that will run on top of tShell, with my command plug ins and any unmatched command will be passed to the tShell. Something like : ->mySh mySh> mySh> myCmd Hello World ! value = 0 = 0x0 mySh> i /* passed on to tShell */ mySh> exitShell -> Can it be done with no network package license ? What are the FDs to be used for in and out and err ? What should be the priority of tShell and tmyShell ? Any help, sample, suggestion is greatly appreciated. Thanks, Bibhuti Kar Cisco Systems Inc. (978) 244 5740 eFax (508)462-0368 --=====================_244998921==_.ALT Content-Type: text/html; charset="us-ascii"
Hi,
  I am trying to write a Shell on vxSim, that will run on top of tShell, with my command plug ins and any unmatched command will be passed to the tShell.

Something like :
->mySh
mySh>
mySh> myCmd
Hello World !
value = 0 = 0x0
mySh> i   /* passed on to tShell */
mySh> exitShell
->

Can it be done with no network package license ? What are the FDs to be used for in and out and err ? What should be the priority of tShell and tmyShell ?

Any help, sample, suggestion is greatly appreciated.

Thanks,


Bibhuti Kar
Cisco Systems Inc.
(978) 244 5740
eFax (508)462-0368 --=====================_244998921==_.ALT-- From vxwexplo-errs@csg.lbl.gov Thu Jun 21 20:33:16 2001 From: "Krishna Kishore Reddy Sureddy" Date: Thu Jun 21 20:33:19 PDT 2001 Subject: Paragma Pack() in vxWorks Hi vxWorks Gurus, What is sigificance of argument passed to #pragma pack() in "C" and "vxWorks". Thanks and Regards, S.K.K.Reddy. From vxwexplo-errs@csg.lbl.gov Fri Jun 22 04:03:21 2001 From: Vxworks Exploder Date: Fri Jun 22 04:03:24 PDT 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Fri Jun 22 04:03:06 PDT 2001 Subject: INF FILE Subject: VxWorks Developpment on Linux Subject: Re: mvme162 or 172 with Petra asic problem? Subject: vxWorks - developing a Shell on vxSim w/ no network package Subject: Re: loadModule and memDrv Subject: Unsubscribe Subject: Re: Unsubscribe Subject: Re: mvme162 or 172 with Petra asic problem? Subject: loadLib, dynamic link with coff file from memory Subject: MPC860 cross-compiler Subject: ATA support for RW4PPC BSP Subject: Re: Updating gcc 2.7.2 for i386 Subject: Help - Can I use Java applets with vxworks' web server? Subject: Re: Reserving Memory Subject: Building .out from a bunch of .a files?? Subject: Running rsh on VxWorks Subject: Re: Unsubscribe Subject: Update of website Subject: Re: How common is SNiFF usage? Subject: Re: Help - Can I use Java applets with vxworks' web server? Subject: How can I get a Call Stack back trace? Subject: Re: Help - Can I use Java applets with vxworks' web server? Subject: Re: Help - Can I use Java applets with vxworks' web server? Subject: Re: Running rsh on VxWorks Subject: Make Problem to use Windview Subject: Re: system tick rate Subject: Re: Running rsh on VxWorks Subject: Paragma Pack() in vxWorks Subject: All zero for Ethernet address Subject: Re: Reserving Memory Subject: Re: Unsubscribe ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: INF FILE Date: Thu, 21 Jun 2001 13:20:40 +0200 From: Yolanda Barquilla Organization: Telefonica Transmision de Datos Message-ID: <3B31D888.D3D5375C@ts.es> Dear all. I have developed a driver and now i must generate the correct inf file o the installation tool in order to install it correctly. My driver is for a not Plug and Play SmartcardReader.The driver extension is vxd and it is for Win98. Can anyone give me an example of an inffile or an exe file generated by InstallShield to install correctly a driver for this type of device?. Thanks in advance Yolanda --------------------------- Newsgroups: comp.os.vxworks Subject: VxWorks Developpment on Linux Date: Thu, 21 Jun 2001 14:41:53 +0200 From: Christophe Rabourg Organization: GACI Message-ID: <3B31EB91.56B0BAF9@gaci.fr> Hello, I am wondering whether such tools are available on Linux (ie: Cross Compiler, remode debugger, remote shell ...) ? If so, where may I download these materials ? Thanks. - -- Christophe Rabourg _/_/_/ _/_/_/ _/_/_/ _/ 6 Av. de Norvege BP46 tel: +33 1 69071686 _/ _/ _/ _/ _/ 91942 Courtaboeuf fax: +33 1 69071358 _/ _/ _/_/_/_/ _/ _/ France crabourg@gaci.fr _/_/_/ _/ _/ _/_/_/ _/ http://www.gaci.fr --------------------------- Newsgroups: comp.os.vxworks Subject: Re: mvme162 or 172 with Petra asic problem? Date: Thu, 21 Jun 2001 13:09:13 GMT From: William Dennen Organization: Excite@Home - The Leader in Broadband http://home.com/faster Message-ID: <3B31F288.4961A710@home.com> References: <3B30FE9E.76C85F42@ll.mit.edu> Reply-To: wdennen@home.com Michael Reed wrote: > > Hello, > Is anyone having a problem with IP packs on mvme162s or mvme172s with > the newer Petra asic? The IP2 chip's functions were added into a > comprehensive asic, the "Petra". On the old mvme162 boards, I start > this IP pack up, can write to it, and it runs fine. On the new mvme162 > w/ Petra asic, the IP pack comes up ok, starts to function, but > generates a bus-error on on the 8th or 9th time I write it. > > Thanks, > Mike > > (IP pack is the IP-Thermistor from GreenSpring/SBS.) Petra 1 (mvme1x2P) or Petra 2 (mvme1x2PA)? Resembles a Petra 1 errata corrected in Petra 2, but details of transfer would be needed to determine ... Regards - -- >@< Bill Dennen william.dennen@motorola.com Motorola Computer Group --GTAC --------------------------- Newsgroups: comp.os.vxworks Subject: vxWorks - developing a Shell on vxSim w/ no network package Date: Thu, 21 Jun 2001 10:02:08 -0400 From: Bibhuti Kar Organization: Lawrence Berkeley National Laboratory Message-ID: <4.1.20010621095630.00b35330@mill.cisco.com> - --=====================_244998921==_.ALT Content-Type: text/plain; charset="us-ascii" Hi, I am trying to write a Shell on vxSim, that will run on top of tShell, with my command plug ins and any unmatched command will be passed to the tShell. Something like : - ->mySh mySh> mySh> myCmd Hello World ! value = 0 = 0x0 mySh> i /* passed on to tShell */ mySh> exitShell - -> Can it be done with no network package license ? What are the FDs to be used for in and out and err ? What should be the priority of tShell and tmyShell ? Any help, sample, suggestion is greatly appreciated. Thanks, Bibhuti Kar Cisco Systems Inc. (978) 244 5740 eFax (508)462-0368 - --=====================_244998921==_.ALT Content-Type: text/html; charset="us-ascii"
Hi,
  I am trying to write a Shell on vxSim, that will run on top of tShell, with my command plug ins and any unmatched command will be passed to the tShell.

Something like :
->mySh
mySh>
mySh> myCmd
Hello World !
value = 0 = 0x0
mySh> i   /* passed on to tShell */
mySh> exitShell
->

Can it be done with no network package license ? What are the FDs to be used for in and out and err ? What should be the priority of tShell and tmyShell ?

Any help, sample, suggestion is greatly appreciated.

Thanks,


Bibhuti Kar
Cisco Systems Inc.
(978) 244 5740
eFax (508)462-0368 - --=====================_244998921==_.ALT-- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: loadModule and memDrv Date: Thu, 21 Jun 2001 10:17:11 -0400 From: "Daniel G Waddington" Organization: Bell Laboratories, Lucent Technologies Message-ID: <9gsvmp$b292@news.research.bell-labs.com> References: <9g7sdc$c523@news.research.bell-labs.com> It works, excellent!!! I'll post this on the newsgroup for anyone else who may have the same problem. Thanks again, DAN > -----Original Message----- > From: Diego Serafin [mailto:d.serafin@telegamma.it] > Sent: Thursday, June 21, 2001 9:14 AM > To: Daniel G Waddington > Subject: Re: loadModule and memDrv > > > I'm sorry I'm responding via e-mail, but my news server rejects messages > today (maybe I'm writing rubbish ?). Feel free to re-post on the group > if you want (please protect my e-mail address from spamming). > > > In article , Bill Pringlemeir says... > > > [...] > > > Just in case ... did you #define'd INCLUDE_LOADER ? I'm doing the same > as you want to do and, before #define'ing that switch I had very strange > errors... > I also #define'd INCLUDE_SYM_TBL and INCLUDE_STANDALONE_SYM_TBL > > This is an excerpt from my code, "loadModuling" fine on a MPC860: > > s = memDevCreate ( "/mem/sw", (char *)localBufferAddress, > localBufferSizeByte); > > ..... > > fd = open ("/mem/sw", O_RDONLY, 0); > > ... > > mId = loadModule (fd, LOAD_ALL_SYMBOLS); > ... > > close (fd); > memDevDelete ("/mem/sw"); > > Hope it helps, > Ciao, > Diego. > > > > -- > Ing. Diego Serafin > Design Engineer > > Telegamma S.r.l. > R&D TLC Division phone: +39 0422 297 574 > Via Noalese 63/B fax : +39 0422 297 579 > 31100 Treviso http://www.telegamma.com > ITALY mailto:d.serafin@telegamma.it > --------------------------- Newsgroups: comp.os.vxworks Subject: Unsubscribe Date: 21 Jun 2001 07:39:47 -0700 From: wizsky@my-deja.com (wizsky) Organization: http://groups.google.com/ Message-ID: <7bd35585.0106210639.37eb2ff3@posting.google.com> How can I unsubscribe from vxwexplo@lbl.gov --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Unsubscribe Date: Thu, 21 Jun 2001 10:53:47 -0400 From: me@pappajoemamma.net Organization: Posted via Supernews, http://www.supernews.com Message-ID: References: <7bd35585.0106210639.37eb2ff3@posting.google.com> wizsky@my-deja.com (wizsky) wrote: :(How can I unsubscribe from :(vxwexplo@lbl.gov Requests for subscriptions, removal from the list, address changes, etc. should be sent to "vxwexplo-request@lbl.gov". Please, use subscribe, unsubscribe or change for the subject. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: mvme162 or 172 with Petra asic problem? Date: Thu, 21 Jun 2001 11:14:54 -0400 From: Michael Reed Message-ID: <3B320F6E.5217A7B8@ll.mit.edu> References: <3B30FE9E.76C85F42@ll.mit.edu> <3B31F288.4961A710@home.com> Hi Bill, It's the mvme162PA-344SE we're using. The card is a 16-channel thermistor. What we're doing with the card is: 1) init the IP-pack 2) foreach thermistor channel select channel (write mask to 0xFFF58201) start adc (write to 0xFFF58206) wait for conversion read data (read from 0xFFF58208) Each time we call our read-therm function, it does this for each of 16 channels, and we ship the data off. On about the 9th, 10th, or 11th call, a bus-error occurs while reading from 0xFFF58208. I've tried increasing & decreasing the ip-recovery time (0, 2, 4, 8 ms) but it didn't help. I also tried adding the Bus-Turnaround-Delay, but that also didn't help (and I think that's only pertinent to DMAs, which we're not doing). Thanks, Mike William Dennen wrote: > > Michael Reed wrote: > > > > Hello, > > Is anyone having a problem with IP packs on mvme162s or mvme172s with > > the newer Petra asic? The IP2 chip's functions were added into a > > comprehensive asic, the "Petra". On the old mvme162 boards, I start > > this IP pack up, can write to it, and it runs fine. On the new mvme162 > > w/ Petra asic, the IP pack comes up ok, starts to function, but > > generates a bus-error on on the 8th or 9th time I write it. > > > > Thanks, > > Mike > > > > (IP pack is the IP-Thermistor from GreenSpring/SBS.) > > Petra 1 (mvme1x2P) or Petra 2 (mvme1x2PA)? Resembles a Petra 1 errata > corrected in Petra 2, but details of transfer would be needed to > determine ... > > Regards > -- > >@< > Bill Dennen william.dennen@motorola.com > Motorola Computer Group --GTAC --------------------------- Newsgroups: comp.os.vxworks Subject: loadLib, dynamic link with coff file from memory Date: Thu, 21 Jun 2001 17:32:50 +0200 From: "BG" Message-ID: <3b3213c0$0$17707@ams2eusosrv31.ams.ops.eu.uu.net> All, Do anyone has a suggestion on loading a coff file from memory instead from file, or do I have to go through file descriptors? Regards, Bent --------------------------- Newsgroups: comp.os.vxworks Subject: MPC860 cross-compiler Date: 21 Jun 2001 09:12:02 -0700 From: ramesh_625@yahoo.com (ramesh) Organization: http://groups.google.com/ Message-ID: Dear Sir, I am using VxWorks for debugging some applications written for MPC860. For this,( concidering I dont have the BSP for MPC860), Do I need to get a seperate cross-compiler for MPC860?. or is the GNU C compiler has a cross compiler supporting MPC860? Also please let me know how do we invoke the GNU cross-compiler. Ramesh --------------------------- Newsgroups: comp.os.vxworks Subject: ATA support for RW4PPC BSP Date: 21 Jun 2001 10:11:57 -0700 From: a1_nocrap_exh@hotmail.com (Alexander Holland) Organization: http://groups.google.com/ Message-ID: <7f38b97b.0106210911.6836ecef@posting.google.com> Hi, I am using the IBM RedwoodIV development board, which is basically a PowerPC405 with multimedia extensions (MPEG2 decoder). It has it's own custom on-chip IDE controller. We are using VxWorks, but the BSP that comes with the package doesnt include the IDE controller (although there is source code to manipulate it). How do I go about *customising* the VxWorks generic IDE driver to my hardware?? I guess you have to add the ATA support to the BSP? Anyone had any experience in doing this? And could perhaps help through the initial steps? Cheers Alex Holland Zarlink Semiconductor --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Updating gcc 2.7.2 for i386 Date: 21 Jun 2001 14:28:11 GMT From: tmb@Xisotek.co.uk (Trevor Barton) Organization: Isotek Electronics Ltd., Leeds, UK. Message-ID: References: <9ec53d11.0106200306.1414178a@posting.google.com> <3B308AF4.1000102@snellwilcox.com> Reply-To: tmb@Xisotek.co.uk On Wed, 20 Jun 2001 12:37:24 +0100, Simon Farnsworth wrote: >Ignacio González wrote: > >> Now I need to use namespaces. So an updating of gcc would be fine >> (2.7.2 does not support namespaces). I suppose there are some tricky >> matters in doing this. >> I'm *sure* somebody has done this before. Can anybody share this >> knowledge with me? > >DaveK has produced a gcc 2.95.x release for PPC target; his pages at >http://www.newgcc4vxworks4ppc.cjb.net/ explain how he did so. > >Note that C++ name mangling has changed between the two versions; you >will need to recompile your BSP to make the new compiler work, and there >may be further issues that would require an updated C++ runtime from >WindRiver. > >Of course, when the next Tornado service pack comes out (RSN ;-), the >compiler update will be included. Y'need to be careful. Exception handling changed when the compiler went through it's egcs phase, and then again after that. The mechanism for retracing the stack seems to have changed, and the functions that are supplied in the WRS supplied runtime libs obviously require the correct version of the compiler. Also, if you use a WRS supplies BSP, you will find that they have, for some reason I can't fathom, hard coded jump offsets in the early boot assembler files. The machine code produced by later assemblers is sufficiently different that the jumps no longer work as required, and you end up jumping into junk. Changing the jumps to assemble-time calculated jumps (ie jumps to labels) works AFAICT. I have not yet managed to produce a gcc version suitable for T2, but we still use 1.0.1 here and this has no exception support anyway. The reasone I rebuilt is that we now develop almost exclusively on Linux and almost never use our NT hosted development system unless we need to use WindView, which is rare. Trev - -- Trevor Barton Isotek Electronics Ltd, 9 Clayton Wood Bank, Leeds, LS16 6QZ, UK. Tel: +44 (113) 275 1339, Fax +44 (113) 224 9827 Remove X from tmb@Xisotek.co.uk before replying. Views expressed are my own and not necessarily those of Isotek Electronics Ltd. --------------------------- Newsgroups: comp.os.vxworks Subject: Help - Can I use Java applets with vxworks' web server? Date: Thu, 21 Jun 2001 12:36:03 -0500 From: "Lee" Message-ID: <9gtba5$akbke$1@ID-92003.news.dfncis.de> Hi, It would be appreciated if someone could help me on this. It's a kind of urgent. Can I use Java applets with vxworks' web server, and how? Thank you in advance. Lee --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Reserving Memory Date: Thu, 21 Jun 2001 13:10:14 -0500 From: Andrew Johnson Organization: APS, Argonne National Laboratory Message-ID: <3B323886.B2B31614@aps.anl.gov> References: <3B314F83.121AE0AD@verizon.net> pat.dirienzo@verizon.net wrote: > > I am trying to put data structures, arrays etc in a memory location of > my choice. In the old operating system, we did this with a linker > directive, but I can't seem to be able to do the same with vxWorks GNU > linker. Is there another way to easily accomplish this? The answer to your question is easy, just cast the address into the pointer of the type you want: struct mystruct { char a; } *pmystruct; long *plong; plong = (long *) 0x1234; printf ("mem[%p]=%lx\n", plong, *plong) pmydata = (struct mystruct *) 0x345678; pmydata->a = 'x'; However you need to be very sure that you know what you're pointing at, as addresses between 0 and your local RAM size usually point to vxWorks system memory. If you're trying to access a VMEbus A24 area you should use sysBusToLocalAdrs() to convert your A24 bus address into a pointer in the CPU's address space, as the VME A24 area probably won't be at local address 0. Note that neither of these will actually reserve memory for you (Subject: line) if what you actually want is to reserve some fixed location in the system RAM for your own purposes, but you didn't ask that. - - Andrew - -- The world is such a cheerful place when viewed from upside-down It makes a rise of every fall, a smile of every frown --------------------------- Newsgroups: comp.os.vxworks Subject: Building .out from a bunch of .a files?? Date: Thu, 21 Jun 2001 18:20:33 GMT From: Davis Leeman Organization: NBTel Internet Message-ID: <3B323AE4.490DF7F0@imagictv.com> I have a workspace made up of many projects. I have built each project into archive files. I have added a new project to my workspace and I would like to build one big .out file from all of my archive files from this project. How is this done? I have set up the build specification, but I am unsure how to include the archive files in the new project. I understand it has something to do with setting the PRJ_LIBS macro under the MACROS tab. How do I add my archive files? What exactly goes into the Value field? Any help would be appreciated. Davis --------------------------- Newsgroups: comp.os.vxworks Subject: Running rsh on VxWorks Date: 21 Jun 2001 11:50:29 -0700 From: mma@ece2.engr.ucf.edu (Mandar A Mathure) Organization: http://groups.google.com/ Message-ID: <639ef38a.0106211050.4e3147a4@posting.google.com> Hi I recently built a kernel for a Motorola 68K target with rlogin and telnet services included. This lets me telnet and also rsh . But there is a slight differnce between how rsh is executed on VxWorks. % rsh runs the command on a UNIX host % rsh the connection is refused. Does anyone know if and how the problem can be rectified? Thanks in advance Mandar --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Unsubscribe Date: Thu, 21 Jun 2001 15:13:30 -0400 From: Beginner Organization: UUNET Canada News Reader Service Message-ID: <3B32475A.536D10B3@hotmail.com> References: <7bd35585.0106210639.37eb2ff3@posting.google.com> Anyone knows what is the advantage of joinging this mail list? me@pappajoemamma.net wrote: > wizsky@my-deja.com (wizsky) wrote: > > :(How can I unsubscribe from > :(vxwexplo@lbl.gov > > Requests for subscriptions, removal from the list, address changes, etc. > should be sent to "vxwexplo-request@lbl.gov". Please, use subscribe, > unsubscribe or change for the subject. --------------------------- Newsgroups: comp.os.vxworks Subject: Update of website Date: 21 Jun 2001 21:12:36 GMT From: borkhuis@agere.com (Johan Borkhuis) Organization: Agere System Message-ID: Hello all VxWorks-workers outthere, I finally found the time to do a major update of the website. A lot of questions have been added. I did collect a lot of messages during the last couple of months, and there are still a lot of them in my mail-queue (together with a lot of even older messages). But I also got a lot of them out of there and into the FAQ. I hope that I can find some more time soon to move a lot more messages. For more information on the questions that have been added see the changes page (http://www.xs4all.nl/~borkhuis/vxworks/vxw_changes.html) Groeten, Johan - -- o o o o o o o . . . _____________________________ o _____ || Johan Borkhuis | .][__n_n_|DD[ ====_____ | borkhuis@agere.com | >(________|__|_[_________]_|__________________________| _/oo OOOOO oo` ooo ooo 'o!o!o o!o!o` === VxWorks FAQ: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html === --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How common is SNiFF usage? Date: 21 Jun 2001 14:16:26 -0700 From: joewoodbury@mail.com (Joe Woodbury) Organization: http://groups.google.com/ Message-ID: References: <9gijt8$abo$03$1@news.t-online.com> <9goe18$frm$05$1@news.t-online.com> <9gq81t$c2v$1@news.hdpp.de> <9gs7nq$gdr$1@news.hdpp.de> The first RTOS I used years ago was Phar Lap ETS. It was adequate, but lacked drivers. However, for what we needed (ticket vending machines) it worked just fine. The device I'm working on now was prototyped with pSOS which is showing it's age. It is missing key features and poorly implemented others. Some of this was rectified with the progressive 2.x releases, but it was clearly a case of too little too late. You save only a little over VxWorks and the support will get progressively worse as those knowledgable in pSOS go onto other things. We also evaluated Windows CE and Embedix (Lineo's embedded Linux offering) to appease management. The biggest problem with both were board support packages, which CE completely lacked and for which Lineo couldn't give us a straight answer. Lineo was also very coy about their [incident] support package costs, but we found out they were very high. Red Hat's support package for eCos "start as low as $15,000"--platinum supports starts at $40,000. Oooh,I can feel the savings now [deep sarcasm.] (eCOS claims to support our board, but there is a disclaimer on the web page that "...no proper device drivers are present in this port..." Other than that, eCOS is promising.) In the end, we chose VxWorks because a) it works with out hardware almost out-of-the-box, b) it's small and designed for embedded systems--it's not an afterthought forced into the role like CE or Linux, c) It's a genuine RTOS, which CE and Linux aren't, d) it is the industry standard, e) they are going to be around for a while, unless, of course, they totally blow it, g) it has some very nice tools (i.e. WindView), h) the OS itself is solid, i) we've gotten good development support so far from them, at least from our local office (they may be unhelpful with tools, but are very helpful in other areas.) Yes, I've been concerned about their general support and their inability to articulate the differences between products. But switching to a competitor would likely be a case of jumping from the frying pan into the fire. Face it, support in the traditional (i.e. UNIX) developer community sucks. (Microsoft may have other problems, but providing developers with nice tools and lots of resources for a very low price generally isn't one of them (except, ironically, in their embedded offerings.)) "Matthias Guelck" wrote in message news:<9gs7nq$gdr$1@news.hdpp.de>... > Hello Joe, > > all WRS FAEs told me, that I am the only person in the world complaining about WRS products. > Again it was very important for me to see, that I am NOT alone! (I was close to it to believe) > Do you (or others) know about a forum like “complaints about WRS”? > > You mentioned that “VxWorks is the best RTOS we've evaluated and/or used”. > Which other RTOS have you evaluated and/or used? > I am asking this, because we are currently searching for an alternative RTOS like ecos, embedded Linux or something other with open > source. > Are you going to stay with WRS products in spite of all your dissatisfaction? > > Best regards > > Matthias --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Help - Can I use Java applets with vxworks' web server? Date: Thu, 21 Jun 2001 17:46:23 -0400 From: "Kevin Livingston" Message-ID: <9UtY6.26$xa5.1683@client> References: <9gtba5$akbke$1@ID-92003.news.dfncis.de> What web server are you using? I know my colleagues have successfully done it with GoAhead web server. And if you just want to view them locally on your Vx computer you can use the AppletViewer (I know it comes with personal Jworks 3.1 I don't know about previous) I have done that and it works just fine. (you don't need to serve them then, you can just give it the classes directly see the example in the P JWorks book. Kevin "Lee" wrote in message news:9gtba5$akbke$1@ID-92003.news.dfncis.de... > Hi, > > It would be appreciated if someone could help me on this. It's a kind of > urgent. > > Can I use Java applets with vxworks' web server, and how? > > Thank you in advance. > > > Lee > > > --------------------------- Newsgroups: comp.os.vxworks Subject: How can I get a Call Stack back trace? Date: Thu, 21 Jun 2001 18:27:39 -0400 From: "Robert H. Fritz" Organization: GeoLogics Message-ID: Reply-To: "Robert H. Fritz" I'm generating a kernel exception which causes my vxWorks host to reboot. From what I've been able to determine, I'm blowing up in bcopy which is used by memcpy and probably a seemingly limitless number of things. I'm looking for a tool that can provide me with a back-trace of my call stack so I can determine what line of code is ultimately causing this problem. Any tool or technique suggestions appreciated. Thanks. Rob Fritz --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Help - Can I use Java applets with vxworks' web server? Date: Thu, 21 Jun 2001 17:34:37 -0500 From: "Lee" Message-ID: <9gtsq0$aq4mh$1@ID-92003.news.dfncis.de> References: <9gtba5$akbke$1@ID-92003.news.dfncis.de> <9UtY6.26$xa5.1683@client> Hi Kevin, Thank you for the info. - - Details We generate a firmware which includs vxwork, its web server, htm files and all other application files. Then upload this firmware file onto flash memory of a device. The device is connected to a host PC via ethernet. We can access web pages (the htm files inside the firmware) without any problems. This means the vxworks' web server included in the firmware is also working OK. - - Requirments Now, we need to embed an applet into one htm file, so when we hit/open that htm file (a web page) we need to see the applet (i.e. that applet is downloaded from web server onto the web browser on host PC side). Could you please give me some suggestions on how we can do this if it's doable. Thank you in advance. Lee "Kevin Livingston" wrote in message news:9UtY6.26$xa5.1683@client... > What web server are you using? > I know my colleagues have successfully done it with GoAhead web server. And > if you just want to view them locally on your Vx computer you can use the > AppletViewer (I know it comes with personal Jworks 3.1 I don't know about > previous) I have done that and it works just fine. (you don't need to serve > them then, you can just give it the classes directly see the example in the > P JWorks book. > > Kevin > > > "Lee" wrote in message > news:9gtba5$akbke$1@ID-92003.news.dfncis.de... > > Hi, > > > > It would be appreciated if someone could help me on this. It's a kind of > > urgent. > > > > Can I use Java applets with vxworks' web server, and how? > > > > Thank you in advance. > > > > > > Lee > > > > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Help - Can I use Java applets with vxworks' web server? Date: Thu, 21 Jun 2001 17:40:28 -0500 From: "Lee" Message-ID: <9gtt4t$b84pc$1@ID-92003.news.dfncis.de> References: <9gtba5$akbke$1@ID-92003.news.dfncis.de> <9UtY6.26$xa5.1683@client> <9gtsq0$aq4mh$1@ID-92003.news.dfncis.de> Hi Kevin, Thank you for the info. - Details We generate a firmware which includs vxwork, Wind Web Server (the web server comes with vxworks), htm files and all other application files. Then upload this firmware file onto flash memory of a device. The device is connected to a host PC via ethernet. We can access web pages (the htm files inside the firmware) without any problems. This means the vxworks' web server included in the firmware is also working OK. - Requirments Now, we need to embed an applet into one htm file, so when we hit/open that htm file (a web page) we need to see the applet (i.e. that applet is downloaded from web server onto the web browser on host PC side). Could you please give me some suggestions on how we can do this if it's doable. Thank you in advance. Lee > > "Kevin Livingston" wrote in message > news:9UtY6.26$xa5.1683@client... > > What web server are you using? > > I know my colleagues have successfully done it with GoAhead web server. > And > > if you just want to view them locally on your Vx computer you can use the > > AppletViewer (I know it comes with personal Jworks 3.1 I don't know about > > previous) I have done that and it works just fine. (you don't need to > serve > > them then, you can just give it the classes directly see the example in > the > > P JWorks book. > > > > Kevin > > > > > > "Lee" wrote in message > > news:9gtba5$akbke$1@ID-92003.news.dfncis.de... > > > Hi, > > > > > > It would be appreciated if someone could help me on this. It's a kind > of > > > urgent. > > > > > > Can I use Java applets with vxworks' web server, and how? > > > > > > Thank you in advance. > > > > > > > > > Lee > > > > > > > > > > > > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Running rsh on VxWorks Date: Thu, 21 Jun 2001 19:23:25 -0500 From: Andrew Johnson Organization: APS, Argonne National Laboratory Message-ID: <3B328FFD.9AA5C1CE@aps.anl.gov> References: <639ef38a.0106211050.4e3147a4@posting.google.com> Mandar A Mathure wrote: > > I recently built a kernel for a Motorola 68K target with rlogin and > telnet services included. This lets me telnet and also rsh > . You probably don't realize that the command 'rsh ' will actually do a 'rlogin ' because you haven't specified any commands for it to run (go read the manpage for rsh). rlogin and rsh are two different protocols with different TCP port numbers (513 v 514 IIRC); vxWorks comes will support for the rlogin and telnet protocols, but not the rsh one. If you want it, you'll have to write it yourself. - - Andrew - -- The world is such a cheerful place when viewed from upside-down It makes a rise of every fall, a smile of every frown --------------------------- Newsgroups: comp.os.vxworks Subject: Make Problem to use Windview Date: Fri, 22 Jun 2001 00:38:30 GMT From: Tony Rodayer Organization: Guest of ProXad - France Message-ID: <3B3293EC.5040905@free.fr> Hi I want to use windview but i have got this error when I compile for each function used by Windview. Nevertheless i have defined INCLUDE_WINDVIEW in configAll.h Thanks --------------------------- Newsgroups: comp.os.vxworks Subject: Re: system tick rate Date: Fri, 22 Jun 2001 02:01:11 GMT From: pkockritz@home.com (Pete Kockritz) Organization: Excite@Home - The Leader in Broadband http://home.com/faster Message-ID: References: <9gshfi$2c4$1@overload.lbl.gov> In article <9gshfi$2c4$1@overload.lbl.gov>, "Jun Wu" wrote: > Hello vxWorks experts, > > I use watch dog implement a periodic event in my application. > Since the event should occur every 6ms, I have to change the system tick rate to 166. > > Is it safe that 166 can be divide exactly by 1000? > > Will the remaining 4ms (1000 - 166*6 = 4) accumulate and affect the vxWorks? > > BR When you set the system tick rate to 166, that means you'll get 166 ticks per second. So the interval will be 1/166 = 6.024096 msec (+/- the accuracy of your processor input clock). You might be better off using the aux timer, if your hardware and BSP support it. Pete - -- Pete Kockritz pkockritz@home.com - -- Pete Kockritz --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Running rsh on VxWorks Date: 21 Jun 2001 19:24:42 -0700 From: Matt Wette Organization: Jet Propulsion Laboratory Message-ID: <7k8zilp7yt.fsf@mr-ed.jpl.nasa.gov> References: <639ef38a.0106211050.4e3147a4@posting.google.com> mma@ece2.engr.ucf.edu (Mandar A Mathure) writes: > Hi > I recently built a kernel for a Motorola 68K target with rlogin and > telnet services included. This lets me telnet and also rsh > . > But there is a slight differnce between how rsh is executed on > VxWorks. > % rsh runs the command on a UNIX host > % rsh the connection is refused. > > Does anyone know if and how the problem can be rectified? /* rshd.c - rshd for VxWorks * * Copyright (C) 1993-2001 Matthew R. Wette -- all rights reserved. * * $Id: rshd.c,v 1.2 2001/01/25 21:16:41 mwette Exp $ * * You may use this program if this copyright notice is kept in tact and * you don't try to sell it. * * To use: * compile to rshd * ld < rshd * rshdInit * * This program is subject to a memory leak in the VxWorks shell. Every * rsh command will eat a little memory. Also, if you're using VxWorks 5.3 * or later, you'll need the SHELL compiled in. */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include static char const rcsid[] = "$Id: rshd.c,v 1.2 2001/01/25 21:16:41 mwette Exp $"; #ifndef RSHD_PORT #define RSHD_PORT 514 #endif #ifndef RSHD_PRIO #define RSHD_PRIO 5 #endif #define SHELL_CMD 1 /* is a shell command */ #define RESET_CMD 2 /* is a reset command */ #define FEVAL_CMD 3 /* is a function eval command */ int rshdInit(void); int rshdTask(void); static int rshd(int sock, struct sockaddr_in*); static int getstr(int, char*, int, char*); static int rshdTid = 0; static int rshdStackSize = 0x8000; static char rcmd_errmsg[80]; int rshdInit(void) { rshdTid = taskSpawn("tRshd", RSHD_PRIO, VX_STDIO, rshdStackSize, rshdTask, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); if (rshdTid == 0) return ERROR; return OK; } int rshdTask(void) { int status, serv_sock, clnt_sock, clen; struct sockaddr_in servaddr, clntaddr; memset(&servaddr, '\0', sizeof(servaddr)); memset(&clntaddr, '\0', sizeof(clntaddr)); /* Open the socket. */ serv_sock = socket(AF_INET, SOCK_STREAM, 0); if (serv_sock == ERROR) return ERROR; servaddr.sin_family = AF_INET; servaddr.sin_port = htons(RSHD_PORT); status = bind(serv_sock, (struct sockaddr *)&servaddr, sizeof(servaddr)); if (status == ERROR) { close(serv_sock); logMsg("rshd: bind failed\n", 0, 0, 0, 0, 0, 0); return ERROR; } status = listen(serv_sock, 2); if (status == ERROR) { logMsg("rshd: listen failed\n", 0, 0, 0, 0, 0, 0); close(serv_sock); return ERROR; } while (1) { clen = sizeof(clntaddr); clnt_sock = accept(serv_sock, (struct sockaddr*)&clntaddr, &clen); if (clnt_sock == ERROR) { logMsg("rshd: accept failed\n", 0, 0, 0, 0, 0, 0); continue; } status = rshd(clnt_sock, &clntaddr); close(clnt_sock); } return OK; } static int rshd(int sock, struct sockaddr_in *addr) { int n, status, was_stdout, was_stderr; int secdport, secdsock, loclport; char c, *cmd, clntuser[16], servuser[16], bufr[128]; /* Get secondary port request. */ /*alarm(60);*/ secdport = 0; while ((n = read(sock, &c, 1)) == 1) { if (c == '\0') break; secdport = 10*secdport + c - '0'; } if (secdport != 0) { loclport = IPPORT_RESERVED - 1; secdsock = rresvport(&loclport); addr->sin_port = htons((unsigned short)secdport); status = connect(secdsock, (struct sockaddr*)addr, sizeof(*addr)); if (status == ERROR) { logMsg("rshd: connect() on second port %d failed\n", secdport, 0, 0, 0, 0, 0); close(sock); return status; } } /* get the command strings */ getstr(sock, clntuser, sizeof(clntuser), "clntuser"); getstr(sock, servuser, sizeof(servuser), "servuser"); getstr(sock, bufr, sizeof(bufr), "command"); # ifdef DEBUG logMsg("rshd: command=[%s]\n", bufr, 0, 0, 0, 0, 0); # endif cmd = bufr; while (*cmd == ' ' || *cmd == '\t') cmd++; /* strip leading whitespace */ if (*cmd++ != '-') { /* -- lock shell -- */ status = shellLock(TRUE); if (status == FALSE) { logMsg("rshd: failed to lock shell\n", 0, 0, 0, 0, 0, 0); write(sock, "!", 1); remCurIdGet(servuser, 0); sprintf(rcmd_errmsg, "rshd: shell locked by user \"%s\"\n", servuser); if (strlen(rcmd_errmsg) > 80) rcmd_errmsg[79] = '\n'; write(sock, rcmd_errmsg, strlen(rcmd_errmsg)); goto exit1; } status = remCurIdSet(clntuser, 0); /* -- set I/O -- */ was_stdout = ioGlobalStdGet(1); ioGlobalStdSet(1, sock); if (secdport) { was_stderr = ioGlobalStdGet(2); ioGlobalStdSet(2, secdsock); } /* -- let the client know we're alive -- */ if (write(sock, "", 1) != 1) { logMsg("rshd: null write to client failed\n", 0, 0, 0, 0, 0, 0); goto exit3; } /* -- execute -- */ status = execute(bufr); if (status == ERROR) { logMsg("rshd: execute() failed\n", 0, 0, 0, 0, 0, 0); goto exit3; } break; } else { /* insert hooks here */ logMsg("rshd: bad command\n", 0, 0, 0, 0, 0, 0); goto exit2; } exit3: /* Reestablish I/O */ ioGlobalStdSet(STD_OUT, was_stdout); if (secdport) ioGlobalStdSet(STD_ERR, was_stderr); exit2: /* unlock shell */ status = shellLock(FALSE); exit1: /* close sockets */ close(sock); if (secdport) close(secdsock); if (qCmdPending) qCmdExec(); return status; } static int getstr(int sock, char *bufr, int cnt, char* errmsg) { int i, n; char c; bufr[i=0] = '\0'; while ((n = read(sock, &c, 1)) == 1) { bufr[i++] = c; if (i == cnt) return ERROR; if (c == '\0') return OK; } if (n == -1) return ERROR; return OK; } /* SYNOPSIS int rshdInit(void) - starts rshdTask int rshdTask(void) - server task int rshd() - service a request on Unix host: rsh DESCRIPTION This is an implementation of the remote shell daemon for VxWorks hosts. It allows users on Unix hosts to run shell commands via the rsh. ERRORS If an error occurs rsh will prints a single line error message. BUGS Surely some exist. SEE ALSO: rsh(1), rcmd(1), rshd(8), in.rshd(8) AUTHOR Matt Wette , Jet Propulsion Laboratory */ /* --- last line of rshd.c --- */ - -- Matthew.R.Wette at jpl.nasa.gov -- I speak for myself, not for JPL. --------------------------- Newsgroups: comp.os.vxworks Subject: Paragma Pack() in vxWorks Date: Fri, 22 Jun 2001 09:05:58 +0530 From: "Krishna Kishore Reddy Sureddy" Organization: Lawrence Berkeley National Laboratory Message-ID: <9gugoh$9f7$1@overload.lbl.gov> Hi vxWorks Gurus, What is sigificance of argument passed to #pragma pack() in "C" and "vxWorks". Thanks and Regards, S.K.K.Reddy. --------------------------- Newsgroups: comp.os.vxworks Subject: All zero for Ethernet address Date: 21 Jun 2001 23:39:09 -0700 From: sunsplendour@my-deja.com (albertw) Organization: http://groups.google.com/ Message-ID: <6e3a7700.0106212239.ca4c187@posting.google.com> Hello! vxWorks gurus, I'm baffled by the driver of a PMC network interface card, namely Osicom 2350. The LAN controller used is DEC21140A. The driver(T2 Beta version) is not in END style. After I've done expattach, ifMaskSet and ifAddrSet, I can see this interface is up using ifShow[Windshell]. But it displays a Ethernet address with all zero, 00:00:00:00:00:00. The hardware just keeps silent. Only pinging itself from Windshell can be successfully done. These have been done upon MCP750, T2.0. Besides, I have done the same thing once on MV2604. Everything is OK. FYI: I have set this interface in a network different from the primary one. And I'm sure the SROM is correct. Any help, hints is anticipated. Thanks in advance. Best regards albertw --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Reserving Memory Date: Fri, 22 Jun 2001 10:15:06 +0100 From: Paul Caswell Organization: Land & Sea Systems Message-ID: <3B330C9A.E834C83C@baesystems.com> References: <3B314F83.121AE0AD@verizon.net> <3B323886.B2B31614@aps.anl.gov> Hi Andrew, You're suggestion is technically correct but I would be very wary of anyone using such mechanism. As you note this does not reserve any memory and as such the operating system is free to use this space as it sees fit (with malloc, new etc....). The addage 'if it can go wrong it will go wrong' is certain to apply here and in my case (and I think in Pats) I needed to reserve the memory for my applications exclusive use. Either way the memDrv driver solves the problem Regards Paul --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Unsubscribe Date: Fri, 22 Jun 2001 11:48:47 +0200 From: "Michael Lawnick" Organization: Buergernetz Dillingen Message-ID: <9gv4g3$t83$1@snoopy.bndlg.de> References: <7bd35585.0106210639.37eb2ff3@posting.google.com> <3B32475A.536D10B3@hotmail.com> Reply-To: "Michael Lawnick" Sender: mlawnick@pd9e65871.dip.t-dialin.net "Beginner" schrieb im Newsbeitrag news:3B32475A.536D10B3@hotmail.com... > Anyone knows what is the advantage of joinging this mail list? > > me@pappajoemamma.net wrote: > > > wizsky@my-deja.com (wizsky) wrote: > > > > :(How can I unsubscribe from > > :(vxwexplo@lbl.gov > > > > Requests for subscriptions, removal from the list, address changes, etc. > > should be sent to "vxwexplo-request@lbl.gov". Please, use subscribe, > > unsubscribe or change for the subject. > It's for those, who have no access to a newsserver of want an automatism of receiving the posts. But: there is a delay of about 1-2 days and some posts get lost as the server filters for the word vxworks. --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Sat Jun 23 04:03:13 2001 From: Vxworks Exploder Date: Sat Jun 23 04:03:15 PDT 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sat Jun 23 04:03:05 PDT 2001 Subject: Re: TFFS limitations? Subject: Determining RAM extent. Subject: PPC405 core+ASIC debug Subject: Re: Compiling VxWorks using Tornado 2.0 from Command Line Subject: Re: SOS - how to download/upload files from/to flash memory? Subject: Re: Updating gcc 2.7.2 for i386 Subject: Re: Paragma Pack() in vxWorks Subject: Re: MPC860 cross-compiler Subject: Re: How can I get a Call Stack back trace? Subject: Re: Make Problem to use Windview Subject: Re: Paragma Pack() in vxWorks Subject: Re: Building .out from a bunch of .a files?? Subject: Re: Running rsh on VxWorks Subject: Add Wdb on VxWorks Subject: Re: SOS - how to download/upload files from/to flash memory? Subject: nfs problems ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: TFFS limitations? Date: Fri, 22 Jun 2001 10:09:46 -0400 From: Doug Dotson Organization: Verio Message-ID: <3B3351AA.684FB899@digidata.com> References: <3B31BDA2.AEEA9B89@siemenscomms.co.uk> As far as I can tell, all you can do is specify an offset of where TFFS-managed flash begins. This means that the boot image must start low. I don;t think that you can treat a single device as multiple smaller devices because of the way that erasing happens. No read access of any portion of the flash can be going on while any erasing is in progress. It least this is my understanding. I can't even get the darn thing (TFFS) to try to identify my flash device. It throws an exception from flIdentifyFlash prior to attempting to call cfiscsIdentify. Hope this helps, Doug Owain Phillips wrote: > Hi, > > I have an 8MB device and the architectural limitation that the boot > image > has to lie at 4MB. > > With TFFS can I reserve arbitary regions of the prom that are not to be > used in filesystems? IE. Can I build a filesystem in that 4MB below the > bootimage? > > Or can I treat the 8MB device as though it is several devices and hence > use the lover 4MB as FSys? > > Or must I reserve 4MB+bootimageSize and have a filesystem > above the bootimage? > > I look forward to your responses... > > Regards, > Owain Phillips --------------------------- Newsgroups: comp.os.vxworks Subject: Determining RAM extent. Date: Fri, 22 Jun 2001 10:11:05 -0400 From: Doug Dotson Organization: Verio Message-ID: <3B3351F9.45A77BBA@digidata.com> Is there a way to determine the base address and size of the RAM in a running system. THanks, Doug --------------------------- Newsgroups: comp.os.vxworks Subject: PPC405 core+ASIC debug Date: Fri, 22 Jun 2001 22:35:02 +0800 From: "Jerry Liu" Organization: Peking University,Beijing,China Message-ID: <9gvkjh$9v0$1@sunlight.pku.edu.cn> hi,everyone, our project use a ASIC which intergated a PPC405B3 core,i'm looking for a JTAG debugger to debug this chip,it seems that VisionICE from EST or Singlestep+blackbird JTAG debugger will be a good choice,but the sales of these debugger just say VisionICE or blackbird JTAG cable will support IBM PPC405GP and not sure they will support PPC405B3 core or not,does anybody has the same experience? these two debugger is not cheap for us,i can't decide to buy which one,but our project's time is up...so please help me!!! any help or hint will be great appreciated! best regards liu --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Compiling VxWorks using Tornado 2.0 from Command Line Date: 22 Jun 2001 15:43:35 GMT From: borkhuis@agere.com (Johan Borkhuis) Organization: Agere System Message-ID: References: <3FtV6.106$bR1.60252@news.pacbell.net> "Langston" wrote: > Hello All, > > Currently we are using Tornado 2.0 to build or VxWorks > applications. We > generally load .wsp (Tornado Workspace Files) to the IDE and create our > builds from there. I would like to be able to script this build > process but if I use a 'Make' command on the Makefile then I run the > risk of not getting all of the latest changes because this Makefile is > generated by the Workspace and Project files. Is there a command line > for building applications from the project files? If I could generate > the Makefile from the command line then perform a Make on it? Take a look at http://www.xs4all.nl/~borkhuis/vxworks/vxw_pt1.html#1.7-C 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: SOS - how to download/upload files from/to flash memory? Date: Fri, 22 Jun 2001 16:52:20 +0200 From: "Ignacio González" Organization: Accesos TSAI Message-ID: <9gvojf$psu1@esiami.tsai.es> References: <9grurv$astd9$1@ID-92003.news.dfncis.de> "Lee" escribió en el mensaje news:9grurv$astd9$1@ID-92003.news.dfncis.de... > Hi, > - Qustion > How to use a web browser to > 1. download a file from the flash memory into the host PC. > 2. upload a file from the host PC into the flash memory. > Microsoft Internet Explorer 5.0 (and Netscape, I suppose) can access files via FTP protocol, instead of HTML protocol. Simply type ftp://your-target-ftp-server-name and that's it. OR do you mean you have to use HTTP protocol (the "Web protocol"). In that case, the VxWork part of your target's program must have an HTTP server (and you usually don't have it, unless the people who purchased and configured the VxWorks part had pay for it). --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Updating gcc 2.7.2 for i386 Date: Fri, 22 Jun 2001 17:22:55 +0100 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: <9ec53d11.0106200306.1414178a@posting.google.com> <3B308AF4.1000102@snellwilcox.com> Trevor Barton wrote in message ... >Y'need to be careful. Exception handling changed when the compiler >went through it's egcs phase, and then again after that. The mechanism >for retracing the stack seems to have changed, and the functions that >are supplied in the WRS supplied runtime libs obviously require the >correct version of the compiler. It is for this reason that I recommend: - remove -nostdinc and -nostdlib from your makefiles / T2 gui options. - remove C++ support and STL from the BSP/kernel, by configuring it out from the T2 gui or #undef'ing the relevant #define INCLUDE_xxx lines - removing the c++ operator new module (and a couple of others, details on my website courtesy of Yu Cao) from the vxworks library. By doing so, all the exception handling code (and in fact all the C++ abi-related code) comes from the new versions of libgcc and libstdc++. These are newer and presumably more sound / bugfree than the older ones; libstdc++ is also a good deal more complete. AFAIK there shouldn't be anything else in the kernel that throws exceptions, so that side of the code should be entirely 2.95.3 if these procedures are followed. DaveK - -- They laughed at Galileo. They laughed at Copernicus. They laughed at Columbus. But remember, they also laughed at Bozo the Clown. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Paragma Pack() in vxWorks Date: Fri, 22 Jun 2001 17:01:29 +0100 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: <9gugoh$9f7$1@overload.lbl.gov> Krishna Kishore Reddy Sureddy wrote in message <9gugoh$9f7$1@overload.lbl.gov>... >Hi vxWorks Gurus, >What is sigificance of argument passed to #pragma pack() in "C" and >"vxWorks". Alas, the only documentation on these is buried deep in the instructions for how to modify and retarget the compiler to a new cpu. Here they are: I don't know whether or not the #pragma pack (pull/push) variants are implemented or not. - ----snip--- `HANDLE_SYSV_PRAGMA' Define this macro (to a value of 1) if you want the System V style pragmas `#pragma pack()' and `#pragma weak [=]' to be supported by gcc. The pack pragma specifies the maximum alignment (in bytes) of fields within a structure, in much the same way as the `__aligned__' and `__packed__' `__attribute__'s do. A pack value of zero resets the behaviour to the default. The weak pragma only works if `SUPPORTS_WEAK' and `ASM_WEAKEN_LABEL' are defined. If enabled it allows the creation of specifically named weak labels, optionally with a value. `HANDLE_PRAGMA_PACK_PUSH_POP' Define this macro (to a value of 1) if you want to support the Win32 style pragmas `#pragma pack(push,)' and `#pragma pack(pop)'. The pack(push,) pragma specifies the maximum alignment (in bytes) of fields within a structure, in much the same way as the `__aligned__' and `__packed__' `__attribute__'s do. A pack value of zero resets the behaviour to the default. Successive invocations of this pragma cause the previous values to be stacked, so that invocations of `#pragma pack(pop)' will return to the previous value. - ----snip--- DaveK - -- They laughed at Galileo. They laughed at Copernicus. They laughed at Columbus. But remember, they also laughed at Bozo the Clown. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: MPC860 cross-compiler Date: Fri, 22 Jun 2001 17:36:09 +0100 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: ramesh wrote in message ... >Dear Sir, > I am using VxWorks for debugging some applications written for >MPC860. For this,( concidering I dont have the BSP for MPC860), Do I >need to get a seperate cross-compiler for MPC860?. or is the GNU C >compiler has a cross compiler supporting MPC860? The Gnu C compiler supports all, or most, PPc types. To specialise your code for the 860, try using the flag -mcpu=860 > Also please let me >know how do we invoke the GNU cross-compiler. From the GUI is quite a good way to do it. From the command line, ccppc DaveK - -- They laughed at Galileo. They laughed at Copernicus. They laughed at Columbus. But remember, they also laughed at Bozo the Clown. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How can I get a Call Stack back trace? Date: Fri, 22 Jun 2001 17:32:04 +0100 From: "Dave Korn" Organization: Lumber Cartel (tinlc) Members #2234-2237 (owing to browser refresh) Message-ID: References: Robert H. Fritz wrote in message ... >I'm generating a kernel exception which causes my vxWorks host to reboot. >From what I've been able to determine, I'm blowing up in bcopy which is used >by memcpy and probably a seemingly limitless number of things. I'm looking >for a tool that can provide me with a back-trace of my call stack so I can >determine what line of code is ultimately causing this problem. > >Any tool or technique suggestions appreciated. Here's what I'd do: Try and narrow down the exact region of your code where it blows up by scattering printf statements through the suspect area of code with messages along the lines of "still alive here..". Run your code and see how many of them make it out before the target reboots; you then know that the crash happens in between the last printf that you see succeeding and the next one that you never see. In fact, you have to be careful; it's possible that the very last printf might be getting lost because the target reboots before it's had time to send the bytes out, but at least you'll know the rough area where the crash happens. When you've localized it to a small part of your code, put a call to taskSuspend (0) at the start of it; now your task will suspend itself just before it blows up. At that point, use the target shell 'b' command to set a breakpoint on bcopy. Now use taskResume (tid) from the shell to restart your app. It should run for a moment and hit the bcopy breakpoint. Now get a stack trace using the 'tt' command from the shell. Then allow the task to continue by giving the 'c' command. Either it will blow up, and you know where the call came from, or it will carry on running until it gets to the next bcopy call, at which point you repeat the procedure. DaveK - -- They laughed at Galileo. They laughed at Copernicus. They laughed at Columbus. But remember, they also laughed at Bozo the Clown. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Make Problem to use Windview Date: 22 Jun 2001 10:12:54 -0700 From: tlksk@hotmail.com (Kumar) Organization: http://groups.google.com/ Message-ID: References: <3B3293EC.5040905@free.fr> Did you regenerate the dependencies once again. Check with doing dependencies once again. Tony Rodayer wrote in message news:<3B3293EC.5040905@free.fr>... > Hi > I want to use windview but i have got this error when I compile for each > function used by Windview. > Nevertheless i have defined INCLUDE_WINDVIEW in configAll.h > Thanks --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Paragma Pack() in vxWorks Date: Fri, 22 Jun 2001 09:55:20 -0700 From: Melvin Gardipee <"> Organization: Raytheon Company Message-ID: References: <9gugoh$9f7$1@overload.lbl.gov> Hi Krishna, The GNU ToolKit Users Guide only reference to #pragma gives reasons for not using them ending with the quote "It is basically a mistake to use #pragma for anything". Use of #pragma in the 1.34 version of GNU C compiler would cause the cause the compiler to stop and launch a computer game. It was their idea of implementing the ANSI standard 'arbitrary implementation-defined effect'. I would avoid using #pragma with GNU compilers. If you are communicating with a VxWorks based system from a different OS, #pragma pack your structures to match the VxWorks layout. For Win32 to VxWorks communication, you can #pragma pack on the Win32 side in a common header file using something like this: #ifdef WIN32 #pragma pack( push, some_unique_name ) #pragma pack(8) #endif #ifdef WIN32 #pragma pack( pop, some_unique_name ) #endif The above example works for PPC architectures. Regards, Melvin Gardipee Krishna Kishore Reddy Sureddy wrote: > > Hi vxWorks Gurus, > What is sigificance of argument passed to #pragma pack() in "C" and > "vxWorks". > Thanks and Regards, > S.K.K.Reddy. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Building .out from a bunch of .a files?? Date: 22 Jun 2001 17:27:12 GMT From: borkhuis@agere.com (Johan Borkhuis) Organization: Agere System Message-ID: References: <3B323AE4.490DF7F0@imagictv.com> Davis Leeman wrote: > I have set up the build specification, but I am unsure how to include > the archive files in the new project. I understand it has something to > do with setting the PRJ_LIBS macro under the MACROS tab. How do I add my > archive files? What exactly goes into the Value field? You can add the complete path to all your archives to this macro, seperated by spaces. 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: Running rsh on VxWorks Date: 22 Jun 2001 18:55:35 GMT From: stnchris@xmission.com (Steve Christensen) Organization: XMission Internet, LLC Message-ID: References: <639ef38a.0106211050.4e3147a4@posting.google.com> <3B328FFD.9AA5C1CE@aps.anl.gov> Reply-To: stnchris@xmission.com On Thu, 21 Jun 2001 19:23:25 -0500, Andrew Johnson wrote: >Mandar A Mathure wrote: >> >> I recently built a kernel for a Motorola 68K target with rlogin and >> telnet services included. This lets me telnet and also rsh >> . > >You probably don't realize that the command 'rsh ' will actually >do a 'rlogin ' because you haven't specified any commands for it >to run (go read the manpage for rsh). rlogin and rsh are two different >protocols with different TCP port numbers (513 v 514 IIRC); vxWorks comes >will support for the rlogin and telnet protocols, but not the rsh one. If >you want it, you'll have to write it yourself. > Rather than have more code running on the target system, you could also script a telnet session using python/perl/tcl. For example, in python: import telnetlib target = "targetName" tn = telnetlib.Telnet(target) print tn.read_until("->") tn.write("i\n") print tn.read_until("->") tn.close() - -Steve - -- Steve Christensen | "It's like getting a scalp massage from Lucifer" stnchris@xmission.com | -- Space Ghost, SGCC --------------------------- Newsgroups: comp.os.vxworks Subject: Add Wdb on VxWorks Date: Fri, 22 Jun 2001 16:31:42 -0500 From: Christophe Part Organization: Bell Sympatico Message-ID: <3B33B93E.F6A33A56@sympatico.ca> Hi, I try to add the Wdb on the serial port on my own board. Can somebody tell me which defines, sources files I need to add for that ? Thanks a lot, Christophe --------------------------- Newsgroups: comp.os.vxworks Subject: Re: SOS - how to download/upload files from/to flash memory? Date: Fri, 22 Jun 2001 17:48:34 -0500 From: "Lee" Message-ID: <9h0i03$bjubc$1@ID-92003.news.dfncis.de> References: <9grurv$astd9$1@ID-92003.news.dfncis.de> <9gvojf$psu1@esiami.tsai.es> Hi, Thank you for the info. It helps. Yes, it will be great if I can use IE's ftp protocal to download/upload files. I tried ftp://my-target-ftp-server-IP but IE popup an error message box showing the following error messages: An error occurred opening that folder on the FTP Server. Make sure you have permission to access that folder. Details: 200 Type set to A, ASCII mode 200 Port set okay 150 Opening ASCII mode data connection 550 No files found or invalid directory or permission problem. It seems IE tries to connect the FTP server in flash memory, and FTP server does have some response. Could you see what is wrong from the error messages? Is port important when connecting FTP server? How to find out my FTP server's port? and how to specify the port number when use IE's ftp protocol? The firmware we upload into flash memory does have both Wind Web Server and FTP Server. I can use NT's ftp client to upload/download files to/from flash memory without any problems. Thanks. Lee "Ignacio González" wrote in message news:9gvojf$psu1@esiami.tsai.es... > > "Lee" escribió en el mensaje > news:9grurv$astd9$1@ID-92003.news.dfncis.de... > > Hi, > > - Qustion > > How to use a web browser to > > 1. download a file from the flash memory into the host PC. > > 2. upload a file from the host PC into the flash memory. > > > > Microsoft Internet Explorer 5.0 (and Netscape, I suppose) can access files > via FTP protocol, instead of HTML protocol. Simply type > ftp://your-target-ftp-server-name and that's it. > > OR do you mean you have to use HTTP protocol (the "Web protocol"). In that > case, the VxWork part of your target's program must have an HTTP server (and > you usually don't have it, unless the people who purchased and configured > the VxWorks part had pay for it). > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: nfs problems Date: Fri, 22 Jun 2001 23:01:08 GMT From: "Joe Flowers" Organization: Excite@Home - The Leader in Broadband http://home.com/faster Message-ID: I'm nfs mounting a remote file system and I can Not cd to the mounted device from the shell. Originally I was doing this without a problem. Somehow my configuration has been corrupted. Does anyone have an idea? >nfsMount("remote", "/home/remotedir", "/mydir"); >cd "/mydir" response is can't find file of directory. The mount device does appear with nfsDevShow device = /mydir remote:/home/remotedir also I can opendir("/mydir"); successfully. however i cannot open or fopen /mydir. What is going on. --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Sat Jun 23 07:43:37 2001 From: "Srinivas Reddy" Date: Sat Jun 23 07:43:39 PDT 2001 Subject: about VxWorks MUX interface hello all, we want to make use of existing TCP/IP stack of VxWorks for Intel IXP1200 Netwprk processor .For this can we use the available MUX interface to send the packet to TCP/IP stack in VXWorks.we have fragmentation &reassembly modules in micro Code ,So we want to Disable the Existing Fragmentation &reassembly in the TCP/IP stack ,how can we Do this ,is there any IP LAyer API available for Doing This and we are Developing MPLS for IXP ,So how Can we Integrate This at layer2. Is there any API available for the communication between Kernel to User ( in Linux we have ntlink sockets) at layer2.we require this the "Label Distribution Protocol"which interanally MPLS will use. we want to interface the IXP with MUX ,can some body guide what are the possible API required for this (If any Document available for this please send us). any help in this regard will be appriciated. thanks in advance regards lonka _____________________________________________________ Buy Lagaan & Yaadein music for 30% less. Avail this special offer at http://shopping.rediff.com/shopping/music/offerrediffmailer.htm From vxwexplo-errs@csg.lbl.gov Sat Jun 23 07:44:42 2001 From: "Srinivas Reddy" Date: Sat Jun 23 07:44:44 PDT 2001 Subject: about VxWorks MUX interface hello all, we want to make use of existing TCP/IP stack of VxWorks for Intel IXP1200 Netwprk processor .For this can we use the available MUX interface to send the packet to TCP/IP stack in VXWorks.we have fragmentation &reassembly modules in micro Code ,So we want to Disable the Existing Fragmentation &reassembly in the TCP/IP stack ,how can we Do this ,is there any IP LAyer API available for Doing This and we are Developing MPLS for IXP ,So how Can we Integrate This at layer2. Is there any API available for the communication between Kernel to User ( in Linux we have ntlink sockets) at layer2.we require this the "Label Distribution Protocol"which interanally MPLS will use. we want to interface the IXP with MUX ,can some body guide what are the possible API required for this (If any Document available for this please send us). any help in this regard will be appriciated. thanks in advance regards lonka _____________________________________________________ Buy Lagaan & Yaadein music for 30% less. Avail this special offer at http://shopping.rediff.com/shopping/music/offerrediffmailer.htm From vxwexplo-errs@csg.lbl.gov Sun Jun 24 04:03:13 2001 From: Vxworks Exploder Date: Sun Jun 24 04:03:15 PDT 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sun Jun 24 04:03:05 PDT 2001 Subject: VXDCOM Subject: about VxWorks MUX interface Subject: about VxWorks MUX interface Subject: Re: about VxWorks MUX interface Subject: Log the output into a file automatically Subject: Re: SOS - how to download/upload files from/to flash memory? Subject: Re: about VxWorks MUX interface Subject: Re: SOS - how to download/upload files from/to flash memory? ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: VXDCOM Date: Sat, 23 Jun 2001 14:09:28 +0200 From: "Michel.Condemine" Organization: 4CE Industry Message-ID: <9h20u3$sjm$1@wanadoo.fr> Reply-To: "Michel.Condemine" Hello , I use Win 2000 Host + I80486 Target link on BNC. Win 2000 in Workgroups. When i us CocreateinstanceEx I always receive an 0x80070776 Error "Object Exporter not found".. Some VXWorks gurus have any suggestion ? thanks Michel --------------------------- Newsgroups: comp.os.vxworks Subject: about VxWorks MUX interface Date: 23 Jun 2001 14:48:57 -0000 From: "Srinivas Reddy" Organization: Lawrence Berkeley National Laboratory Message-ID: <9h2bq1$aj1$1@overload.lbl.gov> hello all, we want to make use of existing TCP/IP stack of VxWorks for Intel IXP1200 Netwprk processor .For this can we use the available MUX interface to send the packet to TCP/IP stack in VXWorks.we have fragmentation &reassembly modules in micro Code ,So we want to Disable the Existing Fragmentation &reassembly in the TCP/IP stack ,how can we Do this ,is there any IP LAyer API available for Doing This and we are Developing MPLS for IXP ,So how Can we Integrate This at layer2. Is there any API available for the co mmunication between Kernel to User ( in Linux we have ntlink sockets) at layer2.we require this the "Label Distribution Protocol"which interanally MPLS will use. we want to interface the IXP with MUX ,can some body guide what are the possible API required for this (If any Document available for this please send us). any help in this regard will be appriciated. thanks in advance regards lonka _____________________________________________________ Buy Lagaan & Yaadein music for 30% less. Avail this special offer at http://shopping.rediff.com/shopping/music/offerrediffmailer.htm --------------------------- Newsgroups: comp.os.vxworks Subject: about VxWorks MUX interface Date: 23 Jun 2001 14:50:02 -0000 From: "Srinivas Reddy" Organization: Lawrence Berkeley National Laboratory Message-ID: <9h2bq2$aj2$1@overload.lbl.gov> hello all, we want to make use of existing TCP/IP stack of VxWorks for Intel IXP1200 Netwprk processor .For this can we use the available MUX interface to send the packet to TCP/IP stack in VXWorks.we have fragmentation &reassembly modules in micro Code ,So we want to Disable the Existing Fragmentation &reassembly in the TCP/IP stack ,how can we Do this ,is there any IP LAyer API available for Doing This and we are Developing MPLS for IXP ,So how Can we Integrate This at layer2. Is there any API available for the co mmunication between Kernel to User ( in Linux we have ntlink sockets) at layer2.we require this the "Label Distribution Protocol"which interanally MPLS will use. we want to interface the IXP with MUX ,can some body guide what are the possible API required for this (If any Document available for this please send us). any help in this regard will be appriciated. thanks in advance regards lonka _____________________________________________________ Buy Lagaan & Yaadein music for 30% less. Avail this special offer at http://shopping.rediff.com/shopping/music/offerrediffmailer.htm --------------------------- Newsgroups: comp.os.vxworks Subject: Re: about VxWorks MUX interface Date: Sat, 23 Jun 2001 18:31:35 +0000 (UTC) From: Jen-Chieh Tang Organization: University of California, Davis Message-ID: <9h2na7$q07$1@woodrow.ucdavis.edu> References: <9h2bq1$aj1$1@overload.lbl.gov> - --> Srinivas, - --> - --> There is no method I am aware of which can turn off the frag/reassm in the - --> vxWorks stack. But for note, that stuff is in ip_output/ip_input. - --> - --> Hmm, I'm no expert at this, but for MPLS maybe you can try redirecting the - --> ifp struct's input/output function pointers to point to your MPLS stuff first. - --> Might be ugly but should work. - --> - --> Think you may need the stack source code bro... - --> - --> Cheers. - --> Tom Srinivas Reddy wrote: > hello all, > we want to make use of existing TCP/IP stack of VxWorks for Intel IXP1200 Netwprk processor .For this can we use the available MUX interface to send the packet to TCP/IP stack in VXWorks.we have fragmentation &reassembly modules in micro Code ,So we want to Disable the Existing Fragmentation &reassembly in the TCP/IP stack ,how can we Do this ,is there any IP LAyer API available for Doing This and we are Developing MPLS for IXP ,So how Can we Integrate This at layer2. Is there any API available for the co > mmunication between Kernel to User ( in Linux we have ntlink sockets) at layer2.we require this the "Label Distribution Protocol"which interanally MPLS will use. > we want to interface the IXP with MUX ,can some body guide what are the possible API required for this (If any Document available for this please send us). > any help in this regard will be appriciated. > thanks in advance > regards > lonka > _____________________________________________________ > Buy Lagaan & Yaadein music for 30% less. > Avail this special offer at http://shopping.rediff.com/shopping/music/offerrediffmailer.htm - -- Tom Tang tangj@cs.ucdavis.edu --------------------------- Newsgroups: comp.os.vxworks Subject: Log the output into a file automatically Date: Sat, 23 Jun 2001 17:50:48 +0800 From: "ferret" Organization: The news server of Changzhou TeleCom Message-ID: <9h1okb$1ba4$1@news.cz.js.cn> I need to examine the output, but it is so troublesome to open a log file every time running the program. How can I log the output of the target into a file instead of the screen automatically? Thank you . --------------------------- Newsgroups: comp.os.vxworks Subject: Re: SOS - how to download/upload files from/to flash memory? Date: 23 Jun 2001 16:34:05 -0700 From: hariharan@foobox.com (Thiagarajan Hariharan) Organization: http://groups.google.com/ Message-ID: <7f10f842.0106231534.1aae5447@posting.google.com> References: <9grurv$astd9$1@ID-92003.news.dfncis.de> <9gvojf$psu1@esiami.tsai.es> <9h0i03$bjubc$1@ID-92003.news.dfncis.de> To use HTTP to upload file check the RFC on forms-based file uploads (eg: if you have an email account at yahoo, this is how you would upload email attachments). Unless the web server has built-in support for forms based upload, you will have to write your own cgi scripts to process the received data and save the file. Hari --------------------------- Newsgroups: comp.os.vxworks Subject: Re: about VxWorks MUX interface Date: 23 Jun 2001 16:49:37 -0700 From: tlksk@hotmail.com (Kumar) Organization: http://groups.google.com/ Message-ID: References: <9h2bq1$aj1$1@overload.lbl.gov> <9h2na7$q07$1@woodrow.ucdavis.edu> Srinivas You can add an ipFilter Hooks at the IP layer to receive the packet do the necessary processing with respect to the mBuf data structures that you get with ( Iam not aware whether it is permitted but this is my suggestion ) and return false so the IP layer wont know that the received packet has been preprocessed by your routine. But this might lead to situations where you wont be freeing the mBufs. After the microcode pass it to the TCP stack for further processing. I dont know whether this is a good solution, but this is an suggestion that you can think of. All the best. Kumar Jen-Chieh Tang wrote in message news:<9h2na7$q07$1@woodrow.ucdavis.edu>... > --> Srinivas, > --> > --> There is no method I am aware of which can turn off the frag/reassm in the > --> vxWorks stack. But for note, that stuff is in ip_output/ip_input. > --> > --> Hmm, I'm no expert at this, but for MPLS maybe you can try redirecting the > --> ifp struct's input/output function pointers to point to your MPLS stuff first. > --> Might be ugly but should work. > --> > --> Think you may need the stack source code bro... > --> > --> Cheers. > --> Tom > > Srinivas Reddy wrote: > > hello all, > > we want to make use of existing TCP/IP stack of VxWorks for Intel IXP1200 Netwprk processor .For this can we use the available MUX interface to send the packet to TCP/IP stack in VXWorks.we have fragmentation &reassembly modules in micro Code ,So we want to Disable the Existing Fragmentation &reassembly in the TCP/IP stack ,how can we Do this ,is there any IP LAyer API available for Doing This and we are Developing MPLS for IXP ,So how Can we Integrate This at layer2. Is there any API available for the > co > > mmunication between Kernel to User ( in Linux we have ntlink sockets) at layer2.we require this the "Label Distribution Protocol"which interanally MPLS will use. > > > we want to interface the IXP with MUX ,can some body guide what are the possible API required for this (If any Document available for this please send us). > > > any help in this regard will be appriciated. > > > thanks in advance > > regards > > lonka > > > _____________________________________________________ > > Buy Lagaan & Yaadein music for 30% less. > > Avail this special offer at http://shopping.rediff.com/shopping/music/offerrediffmailer.htm --------------------------- Newsgroups: comp.os.vxworks Subject: Re: SOS - how to download/upload files from/to flash memory? Date: 23 Jun 2001 17:01:05 -0700 From: tlksk@hotmail.com (Kumar) Organization: http://groups.google.com/ Message-ID: References: <9grurv$astd9$1@ID-92003.news.dfncis.de> <9gvojf$psu1@esiami.tsai.es> <9h0i03$bjubc$1@ID-92003.news.dfncis.de> Hi Lee The better part is to debug what is happening on your target side. Do the following. 1. Point your web browser. Before make sure you have configured your file system in the target. Otherwise you wont be able to see any files over there. 2. make ftpdDebug = 1 to see what is happening on the target side. 3. Then you can upload a file by copying from PC and pasting to the target in the IE window. Similiarily downloading a file is also copying fromPC and pasting to the target on the IE window. Kumar "Lee" wrote in message news:<9h0i03$bjubc$1@ID-92003.news.dfncis.de>... > Hi, > > Thank you for the info. It helps. > > Yes, it will be great if I can use IE's ftp protocal to download/upload > files. > > I tried ftp://my-target-ftp-server-IP but IE popup an error message box > showing the following error messages: > > An error occurred opening that folder on the FTP Server. Make sure you have > permission to access that folder. > Details: > 200 Type set to A, ASCII mode > 200 Port set okay > 150 Opening ASCII mode data connection > 550 No files found or invalid directory or permission problem. > > It seems IE tries to connect the FTP server in flash memory, and FTP server > does have some response. Could you see what is wrong from the error > messages? > > Is port important when connecting FTP server? How to find out my FTP > server's port? and how to specify the port number when use IE's ftp > protocol? > > The firmware we upload into flash memory does have both Wind Web Server and > FTP Server. I can use NT's ftp client to upload/download files to/from > flash memory without any problems. > > Thanks. > > > Lee > > > > "Ignacio González" wrote in message > news:9gvojf$psu1@esiami.tsai.es... > > > > "Lee" escribió en el mensaje > > news:9grurv$astd9$1@ID-92003.news.dfncis.de... > > > Hi, > > > - Qustion > > > How to use a web browser to > > > 1. download a file from the flash memory into the host PC. > > > 2. upload a file from the host PC into the flash memory. > > > > > > > Microsoft Internet Explorer 5.0 (and Netscape, I suppose) can access files > > via FTP protocol, instead of HTML protocol. Simply type > > ftp://your-target-ftp-server-name and that's it. > > > > OR do you mean you have to use HTTP protocol (the "Web protocol"). In that > > case, the VxWork part of your target's program must have an HTTP server > (and > > you usually don't have it, unless the people who purchased and configured > > the VxWorks part had pay for it). > > > > > > > > --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Mon Jun 25 03:15:23 2001 From: Fiederling.Ulf@dornier.dasa.de Date: Mon Jun 25 03:15:25 PDT 2001 Subject: editor Hi VxWorks Gurus I am looking for a small editor for vxWorks 3.4 on x386. Can anybody help me Best regards Ulf Fiederling From vxwexplo-errs@csg.lbl.gov Mon Jun 25 04:03:13 2001 From: Vxworks Exploder Date: Mon Jun 25 04:03:15 PDT 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Mon Jun 25 04:03:04 PDT 2001 Subject: Help - got an error "Truncated class file" when opening an applet. Subject: Re: Help - got an error "Truncated class file" when opening an applet. Subject: Re: Help - got an error "Truncated class file" when opening an applet. Subject: Help - got an error "Truncated class file" when opening an applet. Subject: Re: Help - got an error "Truncated class file" when opening an applet. Subject: Re: Log the output into a file automatically Subject: Re: Change MAC address dynamically in PPC8260 Subject: MUX interface and M_L2HDR flag Subject: editor ------------------------------------------------------- Newsgroups: comp.lang.java.help,comp.lang.java.programmer,comp.os.vxworks Subject: Help - got an error "Truncated class file" when opening an applet. Date: Sun, 24 Jun 2001 13:13:30 -0500 From: "Lee" Message-ID: <9h5a2d$c52iu$1@ID-92003.news.dfncis.de> Hi, I know most people use different environment from mine. It would be appreciated if someone could help me on this, or give me your hints. - - A simple applet is used with a html file (). - - Java Plug-in 1.3.1 is installed on client side (Win2000 with IE5.5), so web browser will use JRE1.3.1 - - convert html file by using HTMLConverter1.3 - - when open html file with IE5.5, only see a gray box on IE and Java Console shows an error like "Truncated class file myApplet.class". The environment is a little bit different from most environment used by most people. The server side operating system is VxWork (a real time OS), the web server is Wind Web Server coming with VxWorks. The web server, VxWorks, converted html file (by HTMLConverter) and myApplet.class file are built into a firmware file which lives in a flash memory. The class file myApplet.class should be OK since it works OK when I use the same html file and the same class file with IIS. Why I received "Truncated class file" error. It would be appreciated if you could give me your hints or guesses. Thank you. Lee --------------------------- Newsgroups: comp.lang.java.help,comp.lang.java.programmer,comp.os.vxworks Subject: Re: Help - got an error "Truncated class file" when opening an applet. Date: Sun, 24 Jun 2001 15:41:00 -0500 From: "Lee" Message-ID: <9h5j8t$c77j2$1@ID-92003.news.dfncis.de> References: <9h5a2d$c52iu$1@ID-92003.news.dfncis.de> <3B362E73.54BDEC4F@qwest.net> Hi Danno, Thanks for the info. Q1: You mentioned "no hard drive space" in you response. The html file, class file, web server and the real time OS (i.e. VxWorks) are packed into a firmware file which is uploaded into a flash memory. So there is no hard drive in this environment at all. I think once the browser opens a html which has an imbedded applet, the browser will ask web server to sent it the applet class file. That is all, right? Why we need to worry about extra hard drive space? Q2: The html file and applet class file are in the same place (I don't use subdirectory because flash memory may not use that type of file structure). So after opening the html file, the browser should be able to find applet class file without any problem. In fact, I think it does. Otherwise why could Java Console show the error message "Truncated class file, myApplet"? Q3: By "removing the disk cache, memory cache", do you mean remove them from client side or server side? If from client side (Win2k with IE5.5), how can I do it? Server side is different because every thing (OS, web server, html file, class file, etc.) is inside a firmware file which lives in a flash memory. Thank you. Lee "Daniel Hinojosa" wrote in message news:3B362E73.54BDEC4F@qwest.net... > Two answers I have seen was no hard drive space. The other was that you > probably moved the applet to another codebase. Remove the disk cache, > memory cache, and restart your browser. Try that, see if that helps. > > Danno > > Lee wrote: > > > > Hi, > > > > I know most people use different environment from mine. It would be > > appreciated if someone could help me on this, or give me your hints. > > > > - A simple applet is used with a html file > > (). > > - Java Plug-in 1.3.1 is installed on client side (Win2000 > > with IE5.5), so web browser will use JRE1.3.1 > > - convert html file by using HTMLConverter1.3 > > - when open html file with IE5.5, only see a gray box on > > IE and Java Console shows an error like "Truncated class > > file myApplet.class". > > > > The environment is a little bit different from most environment used by most > > people. > > The server side operating system is VxWork (a real time OS), the web server > > is Wind Web Server coming with VxWorks. The web server, VxWorks, converted > > html file (by HTMLConverter) and myApplet.class file are built into a > > firmware file which lives in a flash memory. > > > > The class file myApplet.class should be OK since it works OK when I use the > > same html file and the same class file with IIS. > > > > Why I received "Truncated class file" error. It would be appreciated if you > > could give me your hints or guesses. > > > > Thank you. > > > > Lee --------------------------- Newsgroups: comp.lang.java.help,comp.lang.java.programmer,comp.os.vxworks Subject: Re: Help - got an error "Truncated class file" when opening an applet. Date: Sun, 24 Jun 2001 16:53:59 -0600 From: Daniel Hinojosa Organization: Java Programmer and Consultant Message-ID: <3B366F87.A59F1F64@qwest.net> References: <9h5a2d$c52iu$1@ID-92003.news.dfncis.de> <3B362E73.54BDEC4F@qwest.net> <9h5j8t$c77j2$1@ID-92003.news.dfncis.de> Your vxWorks makes it a bit difficult to diagnose, since you're doing this a non-standard way, and no file structure. On I.E. I beleive it is remove temporary Internet files, in the options menu, on the client side. Lee wrote: > > Hi Danno, > > Thanks for the info. > > Q1: You mentioned "no hard drive space" in you response. The html file, > class file, web server and the real time OS (i.e. VxWorks) are packed into a > firmware file which is uploaded into a flash memory. So there is no hard > drive in this environment at all. I think once the browser opens a html > which has an imbedded applet, the browser will ask web server to sent it the > applet class file. That is all, right? Why we need to worry about extra > hard drive space? > > Q2: The html file and applet class file are in the same place (I don't use > subdirectory because flash memory may not use that type of file structure). > So after opening the html file, the browser should be able to find applet > class file without any problem. In fact, I think it does. Otherwise why > could Java Console show the error message "Truncated class file, myApplet"? > > Q3: By "removing the disk cache, memory cache", do you mean remove them from > client side or server side? If from client side (Win2k with IE5.5), how can > I do it? Server side is different because every thing (OS, web server, html > file, class file, etc.) is inside a firmware file which lives in a flash > memory. > > Thank you. > > Lee > > "Daniel Hinojosa" wrote in message > news:3B362E73.54BDEC4F@qwest.net... > > Two answers I have seen was no hard drive space. The other was that you > > probably moved the applet to another codebase. Remove the disk cache, > > memory cache, and restart your browser. Try that, see if that helps. > > > > Danno > > > > Lee wrote: > > > > > > Hi, > > > > > > I know most people use different environment from mine. It would be > > > appreciated if someone could help me on this, or give me your hints. > > > > > > - A simple applet is used with a html file > > > (). > > > - Java Plug-in 1.3.1 is installed on client side (Win2000 > > > with IE5.5), so web browser will use JRE1.3.1 > > > - convert html file by using HTMLConverter1.3 > > > - when open html file with IE5.5, only see a gray box on > > > IE and Java Console shows an error like "Truncated class > > > file myApplet.class". > > > > > > The environment is a little bit different from most environment used by > most > > > people. > > > The server side operating system is VxWork (a real time OS), the web > server > > > is Wind Web Server coming with VxWorks. The web server, VxWorks, > converted > > > html file (by HTMLConverter) and myApplet.class file are built into a > > > firmware file which lives in a flash memory. > > > > > > The class file myApplet.class should be OK since it works OK when I use > the > > > same html file and the same class file with IIS. > > > > > > Why I received "Truncated class file" error. It would be appreciated if > you > > > could give me your hints or guesses. > > > > > > Thank you. > > > > > > Lee --------------------------- Newsgroups: comp.lang.java.programmer,comp.os.vxworks Subject: Help - got an error "Truncated class file" when opening an applet. Date: Sun, 24 Jun 2001 13:19:42 -0500 From: "Lee" Message-ID: <9h5ae1$c7n6v$1@ID-92003.news.dfncis.de> Hi, I know most people use different environment from mine. It would be appreciated if someone could help me on this, or give me your hints. - - A simple applet is used with a html file (). - - Java Plug-in 1.3.1 is installed on client side (Win2000 with IE5.5), so web browser will use JRE1.3.1 - - convert html file by using HTMLConverter1.3 - - when open html file with IE5.5, only see a gray box on IE and Java Console shows an error like "Truncated class file myApplet.class". The environment is a little bit different from most environment used by most people. The server side operating system is VxWork (a real time OS), the web server is Wind Web Server coming with VxWorks. The web server, VxWorks, converted html file (by HTMLConverter) and myApplet.class file are built into a firmware file which lives in a flash memory. The class file myApplet.class should be OK since it works OK when I use the same html file and the same class file with IIS. Why I received "Truncated class file" error. It would be appreciated if you could give me your hints or guesses. Thank you. Lee --------------------------- Newsgroups: comp.lang.java.help,comp.lang.java.programmer,comp.os.vxworks Subject: Re: Help - got an error "Truncated class file" when opening an applet. Date: Sun, 24 Jun 2001 12:16:19 -0600 From: Daniel Hinojosa Organization: Java Programmer and Consultant Message-ID: <3B362E73.54BDEC4F@qwest.net> References: <9h5a2d$c52iu$1@ID-92003.news.dfncis.de> Two answers I have seen was no hard drive space. The other was that you probably moved the applet to another codebase. Remove the disk cache, memory cache, and restart your browser. Try that, see if that helps. Danno Lee wrote: > > Hi, > > I know most people use different environment from mine. It would be > appreciated if someone could help me on this, or give me your hints. > > - A simple applet is used with a html file > (). > - Java Plug-in 1.3.1 is installed on client side (Win2000 > with IE5.5), so web browser will use JRE1.3.1 > - convert html file by using HTMLConverter1.3 > - when open html file with IE5.5, only see a gray box on > IE and Java Console shows an error like "Truncated class > file myApplet.class". > > The environment is a little bit different from most environment used by most > people. > The server side operating system is VxWork (a real time OS), the web server > is Wind Web Server coming with VxWorks. The web server, VxWorks, converted > html file (by HTMLConverter) and myApplet.class file are built into a > firmware file which lives in a flash memory. > > The class file myApplet.class should be OK since it works OK when I use the > same html file and the same class file with IIS. > > Why I received "Truncated class file" error. It would be appreciated if you > could give me your hints or guesses. > > Thank you. > > Lee --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Log the output into a file automatically Date: 24 Jun 2001 18:49:54 -0700 From: tlksk@hotmail.com (Kumar) Organization: http://groups.google.com/ Message-ID: References: <9h1okb$1ba4$1@news.cz.js.cn> You can redirect the STDOUT and STDERR into a file for an particular task using the ios calls. The void ioTaskStdSet ( int taskId, /* task whose std fd is to be set (0 = self) */ int stdFd, /* std input (0), output (1), or error (2) */ int newFd /* new underlying file descriptor */ ) to the file that you wanted to debug. This is in Kumar "ferret" wrote in message news:<9h1okb$1ba4$1@news.cz.js.cn>... > I need to examine the output, but it is so > troublesome to open a log file every time > running the program. > > How can I log the output of the target into a file instead of the screen > automatically? > > Thank you . --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Change MAC address dynamically in PPC8260 Date: Mon, 25 Jun 2001 03:27:47 GMT From: Gururaj patkar Organization: Optimum Online Message-ID: <3B36AF36.260B1257@optonline.net> References: <3B25ED5F.E7E492A4@optonline.net> sorry for the belated reply to some of the questions posted. Let me explain the detailed scenario - In our design, we have a master controller (PPC8260) talking to many child cards (also PPC8260) over ethernet hub. The commuincation between the master and the child card is using ethernet layer 2 only using the vxworks api to filter the ethernet packets. Hence the need for the both the master and the child card to use well known mac address bytes for the first 5 bytes and use the last byte reflecting the card location. The child card do not talk to anyone but the master card currently. The master and the child cards are in a private subnet as far as communication between them is involved. The master uses FCC3 for communicating with the child card and FCC2 for EMS needs. The child card uses FCC3 only. The manufacturing department sets the eeprom mac address which is different from the well known mac address I want to use, hence the need to change it dynamically at run-time to well-known mac address (known to both master and child card). However, before any ethernet packet is sent out of FCC3 on master card and child card I change it dynamically using sysEnetAddrSet(). All I wanted to make sure this change I am making, with the above design in mind does not have any caveats. Thanks for your help. gururaj. Dan Gold wrote: > In article <3B25ED5F.E7E492A4@optonline.net>, Gururaj patkar says... > > > >Hi, > > > >We have the need to change the Ethernet MAC address of one of the FCC > >channel on PPC8260 at runtime dynamically. > > Are you sure you have the need ? Really sure ? Most of the time, if you change > the > MAC address of an interface on the fly you're asking for trouble. Assuming you > do > all the required magic (and there will be magic) to change it, how do you > propagate > your change to all the other people on the network that have your old one ? > Send out > a gratuitous ARP ? You'll have to do it yourself. Wait for all the connections > to drop > because of no response ? :-) Not a very clean solution. Count on everyone > else's ARP > table timing out ? Again, not very efficient. > You could (if all your stars, phase of the moon, and dow jones average all > aligned), try > setting the new MAC address with sysNvRamSet(), then > unloading/loading/starting/attaching the FCC driver, > like they do in usrNetInit(). Of course the FCC driver would need to support > unloading and loading again. > I have an early one and a quick perusal makes me think it would, but I'm not > using that one so I can't say > for sure. > > How many times will you need to do this ? Constantly while the board is > running, or just once ? > > >1) if I use the call to sysEnetAddrSet(), will these corrupt the boot > >parameters in the > >flash? > No, it shouldn't corrupt the boot parameters in FLASH. Where is your MAC > address stored ? Most boards > store it in serial eeprom, not FLASH so it shouldn't be affected. If your MAC > address is stored in FLASH, > better make sure no one cycles power or touches FLASH while you're erasing and > re-progromming with > the new MAC address. The MAC address is not part of the boot parameters (in the > boot line, per se) so > it shouldn't affect you. > > Your bigger problem is changing it on the fly. You'll have to write the 8260 > DPRAM yourself, because the FCC > driver (with ioctl(SIFADDR...) only copies it to the driver control struct). > The function that writes it into the DPRAM > is a local (to the driver) one, and only called from the xxxEndStart() routine. > Of course if you're going to change it > on the fly, you better make sure you're not currently in the middle of Tx'ing or > Rx'ing a frame. I.e. you better stop > the FCC (with GSMR bits) before changing the MAC address. What about data > that's currently buffered up on output > queues in the SENS layer ? Should that go out with the new MAC address ? Do > you care ? > > >2) If (1) is not the solution how do I go about doing this and what > >precaution should I take for this. > > You should take lots of precautions, the biggest one being try to find another > solution other than changing > the MAC address on the fly. > > hth, > > Dan Gold gold@ensemble.com --------------------------- Newsgroups: comp.os.vxworks Subject: MUX interface and M_L2HDR flag Date: Mon, 25 Jun 2001 10:51:10 +0200 From: "Anders Jørgensen" Organization: Posted through some European Outpost of TDC Internet A/S Message-ID: <9h6u6p$d7t$1@news.inet.tele.dk> Hello all, I am writing a level 2 driver for VxWorks. I read the NPT 5.4 User's Guide, regarding muxTkSend, p. 90, to find out whether the outbound network packets I receive from the MUX (from IP) contain an L2 header or not. The doc says that "when sending to a device that uses an NPT driver this requires that you set the M_L2HDR flag in the mBlk header". But i cannot find the definition of the M_L2HDR flag anywhere in the header files. Any suggestions? How do I determine whether L2 headers are already there or not? best regards, Anders Jørgensen, Århus, Denmark. --------------------------- Newsgroups: comp.os.vxworks Subject: editor Date: Mon, 25 Jun 2001 12:17:42 +0200 From: Fiederling.Ulf@dornier.dasa.de Organization: Lawrence Berkeley National Laboratory Message-ID: <9h72vh$iq5$1@overload.lbl.gov> Hi VxWorks Gurus I am looking for a small editor for vxWorks 3.4 on x386. Can anybody help me Best regards Ulf Fiederling --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Mon Jun 25 04:07:08 2001 From: "Marco Raimondi" Date: Mon Jun 25 04:07:11 PDT 2001 Subject: vxworks configuration Hi. I'm working on an old project base on Tornado 1.0.1 and Motorola 68360. I have no documentation except for source code and hw schema. The debugging environment is the small one, with just two serial connections, one for the console and the other for the WDB agent. I need to reconfigure vxWorks system, but i can't get the target server connect to target agent. I get the following messages in target server window: Attaching backend... succeeded. Connecting to target agent... succeeded. Attaching C++ interface... succeeded. Error: Error parsing core filename: Error: No core file: use -f option to specify object format Error: Unable to initialize object module loader library Coul you plese help me? Thanks you all in advance. From vxwexplo-errs@csg.lbl.gov Mon Jun 25 05:58:47 2001 From: "Marco Raimondi" Date: Mon Jun 25 05:58:49 PDT 2001 Subject: gdb Hi vxworks guru! This is my problem. I'm working on an old project base on Tornado 1.0.1 and Motorola 68360. I succeeded in connecting target server with target agent, but when i try to connect gdb with the target iI get the following messages: ************ GDB is free software and you are welcome to distribute copies of it under certain conditions; type "show copying" to see the conditions. There is absolutely no warranty for GDB; type "show warranty" for details. +------------------------------------------------------------------+ | CrossWind 3.1 - Copyright (C) 1990-1996 Wind River Systems, Inc. | | This is a modified version of GDB by Wind River Systems Inc. | | Based on code from the original GDB version shown below: | +------------------------------------------------------------------+ GDB 4.12 (i386-unknown-win32 --target m68k-wrs-vxworks), Copyright 1994 Free Software Foundation, Inc. warning: wrong # args: should be "source fileName" warning: can't read "__wtxCpuType": no such variable Failed: cannot initialize debugger for target type (gdb) ************* What I have to do? Thanks in advanced. From vxwexplo-errs@csg.lbl.gov Mon Jun 25 07:51:48 2001 From: "Sinn, Larry" Date: Mon Jun 25 07:51:51 PDT 2001 Subject: Re: Log the output into a file automatically If you are running from a real OS (unix/solaris) you can have the xterm command do this for you automatically: xterm -iconic -name VxWorks -title VxWorks -geometry 80x25+560+10 -sl 700 \ -fg black -bg white -l -lf "filename" -sb -sk -e tip hardwirea & The only commands really needed are: -lf "filename" - specifies to write to a file, include the path -e tip hardwarea - tip is a routine that connects to a remote system/device, usually serial and in this case it's hardwirea that is specified in /etc/remote and has the form of: hardwirea:\ :dv=/dev/term/a:br#9600:ed=^C^S^Q^U^D:ie=%$:oe=^D: where "/dev/term/a" is the serial device. Larry. Larry Sinn KLA-Tencor MS I-1009 408 875 0247 voice 160 Rio Robles 408 875 6179 fax San Jose, Ca 95134-1809 larry.sinn@kla-tencor.com From: "ferret" I need to examine the output, but it is so troublesome to open a log file every time running the program. How can I log the output of the target into a file instead of the screen automatically? Thank you . From vxwexplo-errs@csg.lbl.gov Tue Jun 26 04:03:20 2001 From: Vxworks Exploder Date: Tue Jun 26 04:03:23 PDT 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Tue Jun 26 04:03:05 PDT 2001 Subject: vxworks configuration Subject: Re: editor Subject: gdb Subject: dosFsLib> dosFsMkFs fails with error S_memLib_NOT_ENOUGH_MEMORY Subject: Re: SOS - how to download/upload files from/to flash memory? Subject: re: Log the output into a file automatically Subject: Re: Motorola 5101 Board Support package Subject: Re: nfs problems Subject: Re: dosFsLib> dosFsMkFs fails with error S_memLib_NOT_ENOUGH_MEMORY Subject: change sector size? HELP! Subject: VxWorks email lists Subject: how to configurate DEC 21143 driver? Subject: STL, -frepo and archives Subject: Re: Cannot start the server Subject: Re: Cannot start the server ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: vxworks configuration Date: Mon, 25 Jun 2001 13:07:06 +0200 From: "Marco Raimondi" Organization: Lawrence Berkeley National Laboratory Message-ID: <9h76g0$k14$1@overload.lbl.gov> Hi. I'm working on an old project base on Tornado 1.0.1 and Motorola 68360. I have no documentation except for source code and hw schema. The debugging environment is the small one, with just two serial connections, one for the console and the other for the WDB agent. I need to reconfigure vxWorks system, but i can't get the target server connect to target agent. I get the following messages in target server window: Attaching backend... succeeded. Connecting to target agent... succeeded. Attaching C++ interface... succeeded. Error: Error parsing core filename: Error: No core file: use -f option to specify object format Error: Unable to initialize object module loader library Coul you plese help me? Thanks you all in advance. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: editor Date: Mon, 25 Jun 2001 12:56:41 +0100 From: Simon Farnsworth Message-ID: <3B3726F9.5060005@snellwilcox.com> References: <9h72vh$iq5$1@overload.lbl.gov> Fiederling.Ulf@dornier.dasa.de wrote: > Hi VxWorks Gurus > > I am looking for a small editor for vxWorks 3.4 on x386. > > Can anybody help me > > > Best regards > > Ulf Fiederling > > Try looking at this archive: ftp://ftp.atd.ucar.edu/pub/vxworks/ I believe they have a text editor available for download --------------------------- Newsgroups: comp.os.vxworks Subject: gdb Date: Mon, 25 Jun 2001 14:58:45 +0200 From: "Marco Raimondi" Organization: Lawrence Berkeley National Laboratory Message-ID: <9h7dh0$mer$1@overload.lbl.gov> Hi vxworks guru! This is my problem. I'm working on an old project base on Tornado 1.0.1 and Motorola 68360. I succeeded in connecting target server with target agent, but when i try to connect gdb with the target iI get the following messages: ************ GDB is free software and you are welcome to distribute copies of it under certain conditions; type "show copying" to see the conditions. There is absolutely no warranty for GDB; type "show warranty" for details. +------------------------------------------------------------------+ | CrossWind 3.1 - Copyright (C) 1990-1996 Wind River Systems, Inc. | | This is a modified version of GDB by Wind River Systems Inc. | | Based on code from the original GDB version shown below: | +------------------------------------------------------------------+ GDB 4.12 (i386-unknown-win32 --target m68k-wrs-vxworks), Copyright 1994 Free Software Foundation, Inc. warning: wrong # args: should be "source fileName" warning: can't read "__wtxCpuType": no such variable Failed: cannot initialize debugger for target type (gdb) ************* What I have to do? Thanks in advanced. --------------------------- Newsgroups: comp.os.vxworks Subject: dosFsLib> dosFsMkFs fails with error S_memLib_NOT_ENOUGH_MEMORY Date: 25 Jun 2001 07:05:07 -0700 From: girishvg@yahoo.com (girish) Organization: http://groups.google.com/ Message-ID: hello friends. i have seen couple of posts regarding this issue. but could not get the clear cut solution. could any of you please explain this. the problem is, i have an ide hard disk of capacity 30GB. when i try to make a file system for 3.2 GB the call dosFsMkFS() returns NULL pointer and errno says S_memLib_NOT_ENOUGH_MEMORY. while as the system has enough that is something like 246MB free!!! the previous calls to ideDrv and ideDevCreate are okay. i have searched http://www.xs4all.nl/~borkhuis/vxworks/vxw_pt3.html also. my target is mipsR4000, t2.0. many thanks in advance. girish. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: SOS - how to download/upload files from/to flash memory? Date: 25 Jun 2001 10:37:57 -0400 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: References: <9grurv$astd9$1@ID-92003.news.dfncis.de> <9gvojf$psu1@esiami.tsai.es> <9h0i03$bjubc$1@ID-92003.news.dfncis.de> Sender: bpringlemeir@DeadDuck >>>>> "Lee" == Lee writes: [snip] Lee> 550 No files found or invalid directory or permission problem. Lee> It seems IE tries to connect the FTP server in flash memory, and Lee> FTP server does have some response. Could you see what is wrong Lee> from the error messages? You must set the default directory to your flash drive. This is something like, #include ioDefPathSet("F:/"); See the vxWorks reference page on `ioLib' for more information. I believe that this was also in the FAQs listed below. hth, Bill Pringlemeir - -- Useful vxWorks URLS, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" "http://www.xs4all.nl/~borkhuis/vxworks/vxw_pt9.html" --------------------------- Newsgroups: comp.os.vxworks Subject: re: Log the output into a file automatically Date: Mon, 25 Jun 2001 07:51:41 -0700 From: "Sinn, Larry" Organization: Lawrence Berkeley National Laboratory Message-ID: <9h7ki0$os1$1@overload.lbl.gov> If you are running from a real OS (unix/solaris) you can have the xterm command do this for you automatically: xterm -iconic -name VxWorks -title VxWorks -geometry 80x25+560+10 -sl 700 \ -fg black -bg white -l -lf "filename" -sb -sk -e tip hardwirea & The only commands really needed are: -lf "filename" - specifies to write to a file, include the path -e tip hardwarea - tip is a routine that connects to a remote system/device, usually serial and in this case it's hardwirea that is specified in /etc/remote and has the form of: hardwirea:\ :dv=/dev/term/a:br#9600:ed=^C^S^Q^U^D:ie=%$:oe=^D: where "/dev/term/a" is the serial device. Larry. Larry Sinn KLA-Tencor MS I-1009 408 875 0247 voice 160 Rio Robles 408 875 6179 fax San Jose, Ca 95134-1809 larry.sinn@kla-tencor.com From: "ferret" I need to examine the output, but it is so troublesome to open a log file every time running the program. How can I log the output of the target into a file instead of the screen automatically? Thank you . --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Motorola 5101 Board Support package Date: 25 Jun 2001 09:20:41 -0700 From: jcampbell@omnigon.com (James Campbell) Organization: http://groups.google.com/ Message-ID: <46f40e0e.0106250820.5d20cead@posting.google.com> References: <3B2FDD4E.D8805CC@ll.mit.edu> <3B309668.F8D54FEB@home.com> William Dennen wrote in message news:<3B309668.F8D54FEB@home.com>... > Stephen Rejto wrote: > > > > We're having problems with our BSP for a Moto 5101 board. > > Doe snayone have a moto5101 orking reliabaly with VxWorks ? > > > > Thanks for any info > > Stephen Rejto > > MIT Linconl Lab > > srejto@ll.mit.edu > > Direct the problem description to your local FAE for the distributor > from whom you purchased the board and supplied the Early Access BSP. > We'll take a look at it ... > > Regards I assume you are talking about the Moto mv5101 VME board with the MCP7400 processor? Yes, we have had it running for a month or so now, but if your just starting to get it up, expect some pain. We've updated the install instructions and added some pretty significant things to the Motorola early access BSP, talk to your FAE for more info. If you need any help email me direct. Cheers, James Campbell jcampbell@omnigon.com Sr. Embedded Software Engineer Omnigon International --------------------------- Newsgroups: comp.os.vxworks Subject: Re: nfs problems Date: Mon, 25 Jun 2001 12:32:22 -0600 From: Don Small Organization: Sandia National Laboratories Message-ID: <3B3783B6.5C5380FE@sandia.gov> References: Have you checked file permissions? Thats one reason you might have been able to access the files before and not now. Don Joe Flowers wrote: > > I'm nfs mounting a remote file system and I can Not cd to the mounted > device from the shell. Originally I was doing this without a problem. > Somehow my configuration has been corrupted. Does anyone have an idea? > > >nfsMount("remote", "/home/remotedir", "/mydir"); > >cd "/mydir" > response is can't find file of directory. > > The mount device does appear with > nfsDevShow > device = /mydir remote:/home/remotedir > > also I can opendir("/mydir"); successfully. > however i cannot open or fopen /mydir. What is going on. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: dosFsLib> dosFsMkFs fails with error S_memLib_NOT_ENOUGH_MEMORY Date: Mon, 25 Jun 2001 22:32:22 GMT From: andreyk@home.com (Andray Kaganovsky) Organization: @Home Message-ID: References: girishvg@yahoo.com (girish) wrote in news:e3f21b2.0106250605.2df5dece@posting.google.com: > hello friends. > > i have seen couple of posts regarding this issue. but could not get > the clear cut solution. could any of you please explain this. the > problem is, i have an ide hard disk of capacity 30GB. when i try to > make a file system for 3.2 GB the call dosFsMkFS() returns NULL > pointer and errno says S_memLib_NOT_ENOUGH_MEMORY. while as the system > has enough that is something like 246MB free!!! the previous calls to > ideDrv and ideDevCreate are okay. i have searched > http://www.xs4all.nl/~borkhuis/vxworks/vxw_pt3.html also. my target is > mipsR4000, t2.0. > > many thanks in advance. > girish. > [AndrayK]. I belive I saw the dosFs patch specifically for MIPS somewhere on WindSurf. Try searching WindSurf's SPR database for dosFs and MIPS. regards, Andray --------------------------- Newsgroups: comp.os.vxworks Subject: change sector size? HELP! Date: Tue, 26 Jun 2001 02:59:29 GMT From: "George" Organization: Excite@Home - The Leader in Broadband http://home.com/faster Message-ID: Reply-To: "George" How do you change the sector size from 512 bytes to 2048? WindRiver has not been much help :-( --------------------------- Newsgroups: comp.os.vxworks Subject: VxWorks email lists Date: Tue, 26 Jun 2001 03:04:24 GMT From: "George" Organization: Excite@Home - The Leader in Broadband http://home.com/faster Message-ID: Reply-To: "George" What are the different VxWorks email lists and how are they different? What are their advantages and disadvantages? --------------------------- Newsgroups: comp.os.vxworks Subject: how to configurate DEC 21143 driver? Date: 25 Jun 2001 23:54:07 -0700 From: gcai@netease.com (cai) Organization: http://groups.google.com/ Message-ID: <31f1a7e2.0106252254.1219583b@posting.google.com> hi everyone, I am new in vxworks.now I am using a 6U CPCI Pentium board as my target.It was provided by Men, a German company(www.men.de). There is a DEC21143 NIC on it. My problem is how to config and make it run with generics Pentium BSP? There is many parameters to configurate. thanks in advance. cai --------------------------- Newsgroups: comp.os.vxworks Subject: STL, -frepo and archives Date: Tue, 26 Jun 2001 09:49:56 +0200 From: "Albert Wijnja" Message-ID: <9h9eqn$cqpbl$1@ID-67650.news.dfncis.de> Hello, For our project we use STL. We also use another library (ACE) for which we need to compile everything with the -fno-implicit-templates flag and explicitly instantiate each template. This gives problems with unresolved symbols regarding STL templates (e.g. _clear__t4list2ZiZt24__default_alloc_template2b1i0) when downloading the .out to the target. To overcome this problem I would like to use -frepo instead of -fno-implicit-templates. I have downloaded the WindRiver patch for the -frepo bug (SPR #29880). As long as all my files are in a single project, everything works fine. Before linking a "collect" is performed to determine the templates that need to be instantiated and some files are recompiled. When downloading, no unresolved symbols appear. But my problem is that not all files are in a single project. I use archives. When creating an archive with -frepo, the .rpo files are generated but not used in any "collect". And the project that produces the .out file uses the archives, but it is not aware of any rpo files and no "collect" is performed, so I still get the unresolved symbols. Has anybody struggled with the same problem and found a solution for it? Albert Wijnja --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Cannot start the server Date: 26 Jun 2001 03:16:33 -0700 From: i_satya@yahoo.com (satya kumar) Organization: http://groups.google.com/ Message-ID: References: <4b0sit00pab53vdo9i10erp3t65svjo3mc@4ax.com> <2bb3jtk214m9rcbo45asg3m93rgv3ov15u@4ax.com> <9gshlb$1ge$1@snoopy.bndlg.de> Hi All, Please help me out Guys am still stuck there without a clue.Thanks in advance. Satya --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Cannot start the server Date: 26 Jun 2001 10:32:37 GMT From: borkhuis@agere.com (Johan Borkhuis) Organization: Agere Systems Message-ID: References: i_satya@yahoo.com (satya kumar) wrote: > Connecting to target agent....succeded > Attaching C++ interface.....succeded > Attaching a.out OMF reader I80X86 CPU Family.....succeded > error : Couldn't open C:\Tornado\vxWorks :status = 0x2 > Error during target server core initialization. What is the exact commandline that starts the target server (in the bottom part of your target server config window)? Does the vxWorks file exist? Is it read/write or read-only? Groeten, Johan - -- o o o o o o o . . . _____________________________ o _____ || Johan Borkhuis | .][__n_n_|DD[ ====_____ | borkhuis@agere.com | >(________|__|_[_________]_|__________________________| _/oo OOOOO oo` ooo ooo 'o!o!o o!o!o` === VxWorks FAQ: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html === --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Tue Jun 26 06:41:26 2001 From: "Praveen Nadgir" Date: Tue Jun 26 06:41:28 PDT 2001 Subject: Dynamic linking/loading.... Hi VxWorks Gurus, I have a simple question: In vxworks, how does dynamic linking/loading work? Is it used or not? thanks in advance, Praveen Nadgir. From vxwexplo-errs@csg.lbl.gov Tue Jun 26 06:49:36 2001 From: "George" Date: Tue Jun 26 06:49:38 PDT 2001 Subject: VxWorks sector size This is a multi-part message in MIME format. ------=_NextPart_000_001E_01C0FE0C.05D4F900 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable How do I change the sector size from 512 to 2048?? George ------=_NextPart_000_001E_01C0FE0C.05D4F900 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable
How do I change the sector size from = 512 to=20 2048??
 
George
------=_NextPart_000_001E_01C0FE0C.05D4F900-- From vxwexplo-errs@csg.lbl.gov Tue Jun 26 07:23:48 2001 From: "Mike Anderson" Date: Tue Jun 26 07:23:51 PDT 2001 Subject: RE: VxWorks sector size VxWorks Greetings! > Submitted-by: "George" > > > How do I change the sector size from 512 to 2048?? > You don't say what kind of device or filesystem you're trying to work with, both of which are key pieces of information to be able to provide you an answer. However, the only device that typically uses a 2048-byte sector size is a CDROM. If that is your target, DosFsLib 2 supports the CDFS which is an ISO-9660 formatted CDROM. If you're looking to support a different device or format, you'll need to give us more info... HTH, Mike Anderson -- ======================================================= T h e P T R G r o u p, I n c. ======================================================= Embedded, Real-Time Solutions and Training Mike Anderson mailto:mike@ThePTRGroup.com Chief Scientist http://www.ThePTRGroup.com V: (703) 585-9384 F: (703) 430-3748 ======================================================= "Software development is like making a baby... You can't make a baby in one month by impregnating nine women. Some things just take time." From vxwexplo-errs@csg.lbl.gov Tue Jun 26 07:33:56 2001 From: "Mike Anderson" Date: Tue Jun 26 07:33:58 PDT 2001 Subject: RE: Dynamic linking/loading.... VxWorks Greetings! > Submitted-by: "Praveen Nadgir" > > Hi VxWorks Gurus, > > I have a simple question: > > In vxworks, how does dynamic linking/loading work? > Is it used or not? > > Are you referring to the dynamic loading of the VxWorks shell, or the C++ dynamic binding thingy? If you're asking about VxWorks dynamic loading, then that is performed using the shell (WindSh or tShell) and a symbol table. As you try to load a piece of code with unresolved references, the symbol table is interrogated and addresses are resolved automagically. Any external symbols are, by default, added to the symbol tables as a result of the loading process. This is obviously the 30,000 ft view of the process. HTH, Mike -- ======================================================= T h e P T R G r o u p, I n c. ======================================================= Embedded, Real-Time Solutions and Training Mike Anderson mailto:mike@ThePTRGroup.com Chief Scientist http://www.ThePTRGroup.com V: (703) 585-9384 F: (703) 430-3748 ======================================================= "Software development is like making a baby... You can't make a baby in one month by impregnating nine women. Some things just take time." From vxwexplo-errs@csg.lbl.gov Tue Jun 26 11:21:41 2001 From: Anuradha Pani Date: Tue Jun 26 11:21:43 PDT 2001 Subject: Threads in VxWorks hi all, I have a question, how are threads implemented in Vxworks? I found several routines for tasks in taskLib library, but none for threads. Thanks __________________________________________________ Do You Yahoo!? Get personalized email addresses from Yahoo! Mail - only $35 a year! http://personal.mail.yahoo.com/ From vxwexplo-errs@csg.lbl.gov Wed Jun 27 02:23:18 2001 From: fscheurer@manz-acs.de Date: Wed Jun 27 02:23:20 PDT 2001 Subject: Additional code & alternative tool chain This message is in MIME format. Since your mail reader does not understand this format, some or all of this message may not be legible. ------_=_NextPart_000_00014DA2.3B39C221 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 7Bit Hi VxWorkers, Additional code without TORNADO tool chain possible? we have a VxWorks based system (x86) developed by a partner where I want to add some cross-compiled object-modules. They wouldn't use VxWorks-code (no libs avail.), but only ANSI-C and the API of the partner's application code. Is it possible to load and run objects compiled e.g. with GNU gccx86 v2.95.x? Is there something special to pay attention to when cross-developing and providing standard-C based object modules, e.g. object format, options, segmentation, etc.? Would the loader accept object modules from an alternative tool-chain? MfG / Kind Regards Friedrich Scheurer, R&D -------------------------------- acs GmbH Steigaeckerstrasse 13 72768 Reutlingen Telefon (07121) 9019-23 Telefax (07121) 9019-98 Internet www.manz-acs.de e-mail: fscheurer@manz-acs.de -------------------------------- ------_=_NextPart_000_00014DA2.3B39C221 Content-Type: text/html; charset=iso-8859-1 Content-Transfer-Encoding: quoted-printable
Hi VxWorkers,
 
Additional code without TORNADO tool chain possible?
 
we have a VxWorks based system (x86) developed by a partner
where I want to add some cross-compiled object-modules.
They wouldn't use VxWorks-code (no libs avail.), but only ANSI-C and
the API of the partner's application code.
 
Is it possible to load and run objects compiled e.g.
with GNU gccx86 v2.95.x?
 
Is there something special
to pay attention to when cross-developing
and providing standard-C based object modules,
e.g. object format, options, segmentation, etc.?
 
Would the loader accept object modules from
an alternative tool-chain?
 
 
MfG / Kind Regards
 
Friedrich Scheurer, R&D
--------------------------------
acs GmbH
Steigaeckerstrasse 13
72768 Reutlingen
Telefon  (07121) 9019-23
Telefax  (07121) 9019-98
Internet www.manz-acs.de
e-mail:  fscheurer@manz-acs.de
--------------------------------
------_=_NextPart_000_00014DA2.3B39C221-- From vxwexplo-errs@csg.lbl.gov Wed Jun 27 04:03:20 2001 From: Vxworks Exploder Date: Wed Jun 27 04:03:23 PDT 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Wed Jun 27 04:03:07 PDT 2001 Subject: Re: SOS - how to download/upload files from/to flash memory? Subject: Re: dosFsLib> dosFsMkFs fails with error S_memLib_NOT_ENOUGH_MEMORY Subject: Dynamic linking/loading.... Subject: VxWorks sector size Subject: RE: VxWorks sector size Subject: RE: Dynamic linking/loading.... Subject: Re: How can I get a Call Stack back trace? Subject: struct sigcontext in signal handlers Subject: RPC problems Subject: Re: SOS - how to download/upload files from/to flash memory? Subject: packet dropping problem under UDP Subject: Re: SOS - how to download/upload files from/to flash memory? Subject: Re: change sector size? HELP! Subject: Re: packet dropping problem under UDP Subject: Threads in VxWorks Subject: Re: SOS - how to download/upload files from/to flash memory? Subject: Java and JNI Subject: Re: SOS - how to download/upload files from/to flash memory? Subject: Re: Add Wdb on VxWorks Subject: Re: Add Wdb on VxWorks Subject: Re: Tornado problem - compile on Unix, debug on Windows. Subject: Re: RPC problems Subject: Re: Threads in VxWorks Subject: Re: Threads in VxWorks Subject: ENUM# signal Subject: Re: change sector size? HELP! Subject: loadModule + memDev = loaded symbols? Subject: Re: Dynamic linking/loading.... Subject: Debug vars / system mBlk config Subject: Re: Debug vars / system mBlk config Subject: Re: Tornado problem - compile on Unix, debug on Windows. Subject: Re: loadModule + memDev = loaded symbols? Subject: Re: Tornado problem - compile on Unix, debug on Windows. Subject: Re: loadModule + memDev = loaded symbols? Subject: Re: Debug vars / system mBlk config Subject: Additional code & alternative tool chain ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: SOS - how to download/upload files from/to flash memory? Date: 26 Jun 2001 04:33:44 -0700 From: igtorque@eliop.es (=?ISO-8859-1?Q?Ignacio_Gonz=E1lez?=) Organization: http://groups.google.com/ Message-ID: <9ec53d11.0106260333.7cbbc6a6@posting.google.com> References: <9grurv$astd9$1@ID-92003.news.dfncis.de> <9gvojf$psu1@esiami.tsai.es> <9h0i03$bjubc$1@ID-92003.news.dfncis.de> tlksk@hotmail.com (Kumar) wrote in message news:... > Hi Lee > > The better part is to debug what is happening on your target side. > Do the following. > > 1. Point your web browser. Before make sure you have configured your > file system in the target. Otherwise you wont be able to see any files > over there. > > 2. make ftpdDebug = 1 to see what is happening on the target side. > > 3. Then you can upload a file by copying from PC and pasting to the > target in the IE window. Similiarily downloading a file is also > copying fromPC and pasting to the target on the IE window. > > > Kumar > I've tried this with a PC-base target, and these are my conclusions: 1) The FTP server in the target has always a "default directory". If not explicitely configured, my target uses the host (named woody:). So the ftp command ls (LIST) would try to list the root directory of my host (!), which usually has no permissions to be accesed by the target. 2) When a "standard" FTP client is run (ftp.exe from MS, WS FTP95, etc.), the client does not try to make a ls (LIST) command immediatly over the default directory: it waits for the user to specify a directory to look for: and when it is instructed to do so (ls /hdo/, for instance), it interprets it as an absolute path. In my case, /hd0/ is mounted as the target's hard disk, so CWD /hd0/ makes target's "C:\" the default drive and directory for subsequent ftp commands. So far, so good. 3) But some *intelligent* ftp clients (like that in MS Internet Explorer), after connection want to know the default path (they issue a PWD command, which gets "woody:" in my target), then try to list its contents (LIST) and then issue the message you have described (because listing "woody:" is denied.) 4) If you try to defeat those *intelligent* ftp clients, specifying ftp://192.9.200.118/hd0/, the assume that /hd0/ is RELATIVE (!) to the default directory ("woody:") and they issue something like CWD woody:/hd0/, and here we are again with the same problem. Who's more intelligent, Wind River or Micro Soft? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: dosFsLib> dosFsMkFs fails with error S_memLib_NOT_ENOUGH_MEMORY Date: 26 Jun 2001 06:53:56 -0700 From: girishvg@yahoo.com (girish) Organization: http://groups.google.com/ Message-ID: References: thanks. is this problem due to address alignment or something? andreyk@home.com (Andray Kaganovsky) wrote in message news:... > girishvg@yahoo.com (girish) wrote in > news:e3f21b2.0106250605.2df5dece@posting.google.com: > > > hello friends. > > > > i have seen couple of posts regarding this issue. but could not get > > the clear cut solution. could any of you please explain this. the > > problem is, i have an ide hard disk of capacity 30GB. when i try to > > make a file system for 3.2 GB the call dosFsMkFS() returns NULL > > pointer and errno says S_memLib_NOT_ENOUGH_MEMORY. while as the system > > has enough that is something like 246MB free!!! the previous calls to > > ideDrv and ideDevCreate are okay. i have searched > > http://www.xs4all.nl/~borkhuis/vxworks/vxw_pt3.html also. my target is > > mipsR4000, t2.0. > > > > many thanks in advance. > > girish. > > > > [AndrayK]. I belive I saw the dosFs patch specifically for MIPS somewhere > on WindSurf. Try searching WindSurf's SPR database for dosFs and MIPS. > > regards, Andray --------------------------- Newsgroups: comp.os.vxworks Subject: Dynamic linking/loading.... Date: Tue, 26 Jun 2001 19:16:06 +0530 From: "Praveen Nadgir" Organization: Lawrence Berkeley National Laboratory Message-ID: <9ha5dk$70q$1@overload.lbl.gov> Hi VxWorks Gurus, I have a simple question: In vxworks, how does dynamic linking/loading work? Is it used or not? thanks in advance, Praveen Nadgir. --------------------------- Newsgroups: comp.os.vxworks Subject: VxWorks sector size Date: Tue, 26 Jun 2001 06:48:35 -0700 From: "George" Organization: Lawrence Berkeley National Laboratory Message-ID: <9ha5e0$70r$1@overload.lbl.gov> This is a multi-part message in MIME format. - ------=_NextPart_000_001E_01C0FE0C.05D4F900 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable How do I change the sector size from 512 to 2048?? George - ------=_NextPart_000_001E_01C0FE0C.05D4F900 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable
How do I change the sector size from = 512 to=20 2048??
 
George
- ------=_NextPart_000_001E_01C0FE0C.05D4F900-- --------------------------- Newsgroups: comp.os.vxworks Subject: RE: VxWorks sector size Date: Tue, 26 Jun 2001 10:23:34 -0400 From: "Mike Anderson" Organization: Lawrence Berkeley National Laboratory Message-ID: <9ha8tv$885$1@overload.lbl.gov> VxWorks Greetings! > Submitted-by: "George" > > > How do I change the sector size from 512 to 2048?? > You don't say what kind of device or filesystem you're trying to work with, both of which are key pieces of information to be able to provide you an answer. However, the only device that typically uses a 2048-byte sector size is a CDROM. If that is your target, DosFsLib 2 supports the CDFS which is an ISO-9660 formatted CDROM. If you're looking to support a different device or format, you'll need to give us more info... HTH, Mike Anderson - -- ======================================================= T h e P T R G r o u p, I n c. ======================================================= Embedded, Real-Time Solutions and Training Mike Anderson mailto:mike@ThePTRGroup.com Chief Scientist http://www.ThePTRGroup.com V: (703) 585-9384 F: (703) 430-3748 ======================================================= "Software development is like making a baby... You can't make a baby in one month by impregnating nine women. Some things just take time." --------------------------- Newsgroups: comp.os.vxworks Subject: RE: Dynamic linking/loading.... Date: Tue, 26 Jun 2001 10:33:42 -0400 From: "Mike Anderson" Organization: Lawrence Berkeley National Laboratory Message-ID: <9ha8u0$886$1@overload.lbl.gov> VxWorks Greetings! > Submitted-by: "Praveen Nadgir" > > Hi VxWorks Gurus, > > I have a simple question: > > In vxworks, how does dynamic linking/loading work? > Is it used or not? > > Are you referring to the dynamic loading of the VxWorks shell, or the C++ dynamic binding thingy? If you're asking about VxWorks dynamic loading, then that is performed using the shell (WindSh or tShell) and a symbol table. As you try to load a piece of code with unresolved references, the symbol table is interrogated and addresses are resolved automagically. Any external symbols are, by default, added to the symbol tables as a result of the loading process. This is obviously the 30,000 ft view of the process. HTH, Mike - -- ======================================================= T h e P T R G r o u p, I n c. ======================================================= Embedded, Real-Time Solutions and Training Mike Anderson mailto:mike@ThePTRGroup.com Chief Scientist http://www.ThePTRGroup.com V: (703) 585-9384 F: (703) 430-3748 ======================================================= "Software development is like making a baby... You can't make a baby in one month by impregnating nine women. Some things just take time." --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How can I get a Call Stack back trace? Date: Tue, 26 Jun 2001 11:22:34 -0400 From: "Robert H. Fritz" Organization: GeoLogics Message-ID: References: Reply-To: "Robert H. Fritz" Dave, Thanks for the info. Judging from some similar threads I've read, printf's and gdb are about as advanced as it gets in vxWorks. Bummer. Rob --------------------------- Newsgroups: comp.os.vxworks Subject: struct sigcontext in signal handlers Date: Tue, 26 Jun 2001 16:51:44 +0100 From: vxworksintent@tao-group.com (David Given) Organization: I do not speak for anyone but myself, and barely that. Message-ID: Sender: dg@pearl.tao.co.uk In the following code: - ---snip--- void sighandler(int signum, siginfo_t* si, struct sigcontext* context) { printf("pc=%x (%x)\n", context->sc_regs.pc, main); } void main(void) { struct sigvec sv = {sighandler, -1, SA_SIGINFO}; sigvec(SIGSEGV, &sv, NULL); sigvec(SIGBUS, &sv, NULL); { int a = *(int*)0; printf("%d", a); // to prevent the deref being optimised // away } } - ---snip--- ...I find that the program counter displayed by the signal handler is somewhere inside sigqueue() and not, as I would have expected, at the address of the null reference itself. This renders the whole signal context thing rather useless for my purpose, which is a pity, because there's no other way of doing what I want... Can anyone explain this, and suggest a way in which I can fix it? On a related note, is there a way to suppress the exception notification I get on the console when this happens? - -- David Given vxworksintent@tao-group.com --------------------------- Newsgroups: comp.os.vxworks Subject: RPC problems Date: Tue, 26 Jun 2001 09:33:17 -0700 From: "Andrew Spiehler" Organization: Raytheon Company Message-ID: I have my target set up as an RPC server. When I make a large number of remote procedure calls from my RPC client, the first 15 to 20 of them return from the server in a reasonable amount of time 2-3ms. After that, they start taking from 22 to 44 ms to return from the server. I saw this happening first on Windview, but to verify that the slow down wasn't caused by Windview event logging, I also watched network traffic with a sniffer program. Does anyone know what might be causing this? I've heard second hand that WRS has said that their RPC library is buggy, but I haven't checked on that. Andrew Spiehler --------------------------- Newsgroups: comp.os.vxworks Subject: Re: SOS - how to download/upload files from/to flash memory? Date: Tue, 26 Jun 2001 12:11:06 -0500 From: "Lee" Message-ID: <9hafnb$ctlcl$1@ID-92003.news.dfncis.de> References: <9grurv$astd9$1@ID-92003.news.dfncis.de> <9gvojf$psu1@esiami.tsai.es> Thanks for the info. Suppose I use IE5.5 to access a ftp server, and that server needs UserName and Password. How do I specify UserName and Password? Something like ftp://ftp-server-ip?user:pass or what? Thanks. Lee "Ignacio González" wrote in message news:9gvojf$psu1@esiami.tsai.es... > > "Lee" escribió en el mensaje > news:9grurv$astd9$1@ID-92003.news.dfncis.de... > > Hi, > > - Qustion > > How to use a web browser to > > 1. download a file from the flash memory into the host PC. > > 2. upload a file from the host PC into the flash memory. > > > > Microsoft Internet Explorer 5.0 (and Netscape, I suppose) can access files > via FTP protocol, instead of HTML protocol. Simply type > ftp://your-target-ftp-server-name and that's it. > > OR do you mean you have to use HTTP protocol (the "Web protocol"). In that > case, the VxWork part of your target's program must have an HTTP server (and > you usually don't have it, unless the people who purchased and configured > the VxWorks part had pay for it). > > > > --------------------------- Newsgroups: comp.os.vxworks Subject: packet dropping problem under UDP Date: Tue, 26 Jun 2001 10:12:01 -0700 From: "Andrew Spiehler" Organization: Raytheon Company Message-ID: <9p3_6.160$Mf6.3315@dfw-service2.ext.raytheon.com> I'm streaming data to my VxWorks target using UDP sockets. What I have setup up is basically a loop on my station writing data to a socket (1300-1400 bytes per packet) and a loop on my target reading the data out of the socket and checking a sequence number in the data. When I transmit a large number of packets, I occasionally lose blocks of 20 to 30. Sometimes after the first 30 or 40 packets, sometimes it's after several hundred, and sometimes not at all. The packets are going across the network according to the packet sniffer I have attached to the network. I don't think it's the socket receive buffer because the task reading data out of the socket is only taking about 15 microseconds to complete a loop. Could it be some memory pool problem in the network stack? I don't really understand the mbuf/mblk intricacies... can anyone recommend a better reference for that than the VxWorks programmer's guide? Thanks, Andrew Spiehler --------------------------- Newsgroups: comp.os.vxworks Subject: Re: SOS - how to download/upload files from/to flash memory? Date: Tue, 26 Jun 2001 18:57:20 +0100 From: Simon Farnsworth Message-ID: <3B38CD00.9050202@snellwilcox.com> References: <9grurv$astd9$1@ID-92003.news.dfncis.de> <9gvojf$psu1@esiami.tsai.es> <9hafnb$ctlcl$1@ID-92003.news.dfncis.de> Lee wrote: > Suppose I use IE5.5 to access a ftp server, and that server needs UserName > and Password. How do I specify UserName and Password? Something like > ftp://ftp-server-ip?user:pass or what? I believe the syntax is ftp://user:password@server-(name|ip)[:port]/[path] --------------------------- Newsgroups: comp.os.vxworks Subject: Re: change sector size? HELP! Date: Tue, 26 Jun 2001 19:37:31 +0100 From: molochai Message-ID: <52ccd904a%molochai@vapour-trail.demon.co.uk> References: In message "George" wrote: > How do you change the sector size from 512 bytes to 2048? WindRiver has > not been much help :-( > Of what? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: packet dropping problem under UDP Date: Tue, 26 Jun 2001 20:08:44 +0100 From: molochai Message-ID: References: <9p3_6.160$Mf6.3315@dfw-service2.ext.raytheon.com> In message <9p3_6.160$Mf6.3315@dfw-service2.ext.raytheon.com> "Andrew Spiehler" wrote: > I'm streaming data to my VxWorks target using UDP sockets. What I have > setup up is basically a loop on my station writing data to a socket > (1300-1400 bytes per packet) and a loop on my target reading the data out > of the socket and checking a sequence number in the data. When I transmit > a large number of packets, I occasionally lose blocks of 20 to 30. > Sometimes after the first 30 or 40 packets, sometimes it's after several > hundred, and sometimes not at all. The packets are going across the network > according to the packet sniffer I have attached to the network. The net/3 stack discards the IP fragment queue upon an out-of-mbuf event. Quite probably the ethernet device driver has been stuffing stuff into this queue under interrupts and netTask has not been run to process them. (See TCP/IP illustrated Vol II p298 for more details). > I don't think it's the socket receive buffer because the task reading data > out of the socket is only taking about 15 microseconds to complete a loop. The socket recieve buffers are mbufs themselves so this will make no difference. > Could it be some memory pool problem in the network stack? I don't really > understand the mbuf/mblk intricacies... can anyone recommend a better > reference for that than the VxWorks programmer's guide? This could help however, you can't expect UDP to have 100% reliability. The sender, recipient, and any subsequent routers or switches inbetween are quite at liberty to discard IP datagrams if they run out of buffers. If you need 100% reliability use TCP or roll your own flow control and retry mechanism. > --------------------------- Newsgroups: comp.os.vxworks Subject: Threads in VxWorks Date: Tue, 26 Jun 2001 11:21:35 -0700 (PDT) From: Anuradha Pani Organization: Lawrence Berkeley National Laboratory Message-ID: <9hamvv$hg1$1@overload.lbl.gov> hi all, I have a question, how are threads implemented in Vxworks? I found several routines for tasks in taskLib library, but none for threads. Thanks __________________________________________________ Do You Yahoo!? Get personalized email addresses from Yahoo! Mail - only $35 a year! http://personal.mail.yahoo.com/ --------------------------- Newsgroups: comp.os.vxworks Subject: Re: SOS - how to download/upload files from/to flash memory? Date: Tue, 26 Jun 2001 14:34:37 -0500 From: "Lee" Message-ID: <9hao4f$d4r2q$1@ID-92003.news.dfncis.de> References: <9grurv$astd9$1@ID-92003.news.dfncis.de> <9gvojf$psu1@esiami.tsai.es> <9h0i03$bjubc$1@ID-92003.news.dfncis.de> <9ec53d11.0106260333.7cbbc6a6@posting.google.com> Hi, Thank you for the infor. It helps. It seems the default directory of VxWorks ftp server is "/flash". I see it's defined in application code like: #include ioDefPathSet("/flash"); Also, I can see "/flash" as the root from the connection messages under Q1 below. Q1: Why is there a difference between using Win2k's DOS ftp client and using Windows-based ftp client such as ws_ftp (ws_ftp can be downloaded from http://download.cnet.com/downloads/1,10150,0-10000-103-0-1-7,00.html?tag=&qt =wsftp&cn=&ca=10000 )? I can use NT/2000's DOS ftp client to access the ftp server of VxWorks inside flash memory, and list/download/upload files without any problem, but when I use ws_ftp, the connection to VxWorks ftp server could be established, but LIST command always failed, not even mention download/upload files. The error msg is: WINSOCK.DLL: WinSock 2.0 WS_FTP Pro 6.70T 2001.03.28, Copyright © 1992-2000 Ipswitch, Inc. - - - connecting to xxx.xxx.xxx.xxx:21 (xxx.xxx.xxx.xxx is the ftp server IP) Connected to xxx.xxx.xxx.xxx port 21 (xxx.xxx.xxx.xxx is the ftp server IP) 220 VxWorks (5.4) FTP server ready USER root 331 Password required PASS (hidden) 230 User logged in PWD 257 Current directory is "/flash" Host type (I): VxWorks PORT xxx,xxx,xxx,xxx,8,121 (xxx,xxx,xxx,xxx is the client IP) 200 Port set okay LIST 150 Opening ASCII mode data connection Received 89 bytes in 0.1 secs, (8900.00 bps), transfer succeeded 550 No files found or invalid directory or permission problem ! Retrieve of folder listing failed (5) Why ws_ftp LIST command fails while I can use "ls" (LIST command) with Win2k DOS ftp client without any problems? If I do need to use ws_ftp to work with VxWorks ftp server to list/download/upload files, is there any workaround available? Q2: - - the VxWorks ftp server's IP is "ftp-server-ip" - - the root is "/flash" (this is the only place I can see after ftp connection is established) - - ftp server also needs UserName as "usr" and Password as "ps". Now, what URL should I use on IE5.5 to connect to this ftp server? ftp://ftp-server-ip/flash?usr:ps, or something else? Thank you in advance "Ignacio González" wrote in message news:9ec53d11.0106260333.7cbbc6a6@posting.google.com... > tlksk@hotmail.com (Kumar) wrote in message news:... > > Hi Lee > > > > The better part is to debug what is happening on your target side. > > Do the following. > > > > 1. Point your web browser. Before make sure you have configured your > > file system in the target. Otherwise you wont be able to see any files > > over there. > > > > 2. make ftpdDebug = 1 to see what is happening on the target side. > > > > 3. Then you can upload a file by copying from PC and pasting to the > > target in the IE window. Similiarily downloading a file is also > > copying fromPC and pasting to the target on the IE window. > > > > > > Kumar > > > > I've tried this with a PC-base target, and these are my conclusions: > > 1) The FTP server in the target has always a "default directory". If > not explicitely configured, my target uses the host (named woody:). So > the ftp command ls (LIST) would try to list the root directory of my > host (!), which usually has no permissions to be accesed by the > target. > > 2) When a "standard" FTP client is run (ftp.exe from MS, WS FTP95, > etc.), the client does not try to make a ls (LIST) command immediatly > over the default directory: it waits for the user to specify a > directory to look for: and when it is instructed to do so (ls /hdo/, > for instance), it interprets it as an absolute path. In my case, /hd0/ > is mounted as the target's hard disk, so CWD /hd0/ makes target's > "C:\" the default drive and directory for subsequent ftp commands. So > far, so good. > > 3) But some *intelligent* ftp clients (like that in MS Internet > Explorer), after connection want to know the default path (they issue > a PWD command, which > gets "woody:" in my target), then try to list its contents (LIST) and > then issue the message you have described (because listing "woody:" is > denied.) > > 4) If you try to defeat those *intelligent* ftp clients, specifying > ftp://192.9.200.118/hd0/, the assume that /hd0/ is RELATIVE (!) to the > default directory ("woody:") and they issue something like CWD > woody:/hd0/, and here we are again with the same problem. > > Who's more intelligent, Wind River or Micro Soft? --------------------------- Newsgroups: comp.os.vxworks Subject: Java and JNI Date: Tue, 26 Jun 2001 16:49:10 -0400 From: "Edwin McKay" Message-ID: I'm trying to use the JNI portion of Java to interface to a Java database and have run into a variety of problems. It seems a C program can create and then Attach to the VM multiple times, which is good. However, a C++ program can create the VM but cannot Attach to an existing one, after running once the program core dumps the next time until a reboot. Once either a C or C++ program have created the JVM a normal 'java HelloWorld' won't work anymore, the shell just hangs. If I run 'java HelloWorld' first then my C/C++ programs won't run, they core dump. Also, it seems that only the main thread of a C program can connect with the JVM, tasks that are spawned can neither attach nor use the 'env' pointer of the main thread. This seems to imply that only 1 thread of execution can actually use the JNI environment. I can call C routines from Java, but that seems like a one-way path. I would like to have either side be able to use JNI to communicate back and forth. Our real-time code is in C++, while our user interface and persistent storage is (maybe) in Java. There is an spr (#9194) that mentions AttachCurrentThread, but I was hoping someone had some further insights on the problem. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: SOS - how to download/upload files from/to flash memory? Date: Tue, 26 Jun 2001 15:57:34 -0500 From: "Lee" Message-ID: <9hat00$d7r1t$1@ID-92003.news.dfncis.de> References: <9grurv$astd9$1@ID-92003.news.dfncis.de> <9gvojf$psu1@esiami.tsai.es> <9hafnb$ctlcl$1@ID-92003.news.dfncis.de> <3B38CD00.9050202@snellwilcox.com> Hi, Thanks for the info. Where can I find the syntax? Is there a document or web site I can check? How to specify binary or ascii when connect to the server. Which one is the default? Thanks Lee "Simon Farnsworth" wrote in message news:3B38CD00.9050202@snellwilcox.com... > Lee wrote: > > > Suppose I use IE5.5 to access a ftp server, and that server needs UserName > > and Password. How do I specify UserName and Password? Something like > > ftp://ftp-server-ip?user:pass or what? > > I believe the syntax is ftp://user:password@server-(name|ip)[:port]/[path] > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Add Wdb on VxWorks Date: Tue, 26 Jun 2001 17:19:33 -0500 From: Christophe Part Organization: Bell Sympatico Message-ID: <3B390A75.C782560A@sympatico.ca> References: <3B33B93E.F6A33A56@sympatico.ca> - -------- Original Message -------- Subject: Re: Add Wdb on VxWorks Date: Fri, 22 Jun 2001 16:56:46 -0700 From: "GERASIMOV,SERGEY (HP-Roseville,ex1)" To: "'chpart@sympatico.ca'" Hi, try to refer to pp 28 and 119 of Tornado user's guide. Also, if you are building on unix, then config.h in the target directory probably should have: #define INCLUDE_WDB #define WDB_COMM_PORT 1 (for raw serial) You will also will need to select desired baud rate and other serial connection parameters. I thin default work pretty adequate though, at least that's what we are using. Now, building the image and loading it on the target should result in the message 'WDB READY' on the console. That should be it. Hope that helps, Sergey Christophe Part wrote: > Hi, > > I try to add the Wdb on the serial port on my own board. > Can somebody tell me which defines, sources files I need to add for that > ? > > Thanks a lot, > > Christophe --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado problem - compile on Unix, debug on Windows. Date: 26 Jun 2001 14:34:01 -0700 From: matthew.becker@usa.xerox.com (Matthew Becker) Organization: http://groups.google.com/ Message-ID: <94e55b6.0106261334.2f3a53b6@posting.google.com> References: <94e55b6.0106150725.35205ecc@posting.google.com> I have still not received an adequate response to this message (yea or nay). Therefore, I'm reposting it with some additional clarification of the problem. When using the VxWorks cross-compiler (with debug flags turned on), certain information is stored in the object file that is used for debugging. Part of that information is the absolute paths to the source files used to compile that object. The problem I am having is that I want to compile on Unix (Solaris), but I want to debug on Windows. When the Tornado debugger on Windows opens that object file, it can't resolve the Unix paths that were used to compile the object. The Windows box I'm using has a mapped network drive to the Solaris box. Therefore, all of the paths to the source files are only off by the base. For example, if the file "/home/jsmith/project/USBDriver.cc" existed on Solaris, then it could be seen as "E:\project\USBDriver.cc" on Windows. I was wondering if the Tornado debugger had some internal smarts such that it could take a path-mapping (e.g. "/home/jsmith" -> "E:\"), and correctly map the Unix paths in the object file to Windows paths (so that it could then find the source files). This type of funtionality exists on other debugging solutions provided by WindRiver (e.g. SingleStep and visionClick). Does this functionality exist in Tornado II? Tornado 3/AE? Is it going to be added to AE? Thanks in advance. - -Matt matthew.becker@usa.xerox.com (Matthew Becker) wrote in message news:<94e55b6.0106150725.35205ecc@posting.google.com>... > Hello all, > > The development group I'm in wants to use the VxWorks cross-compiler > on Solaris to perform target builds, and Tornado on Windows PC's to > perform debug. In the past, this strategy has worked because the > IDE's they were using had some notion of associating a Unix path to a > Windows driver letter (i.e. /home/jsmith -> E:\ ). > > Unfortunately, I have been unable to find this same functionality in > Tornado. Currently the group has licences (and is using) Tornado II > for Solaris and Windows. The Tornado IDE on Windows does provide a > "Debugger Source Search Path" window, but it doesn't seem to be smart > enough to recursively descend through a specified directory. In other > words, all directories for a project must be specified individually > (also, not what I'm looking for). > > Does anyone know of a solution to this problem for Tornado II? Is > this issue resolved with Tornado 3/AE? > > Thanks in advance. > > -Matt Becker --------------------------- Newsgroups: comp.os.vxworks Subject: Re: RPC problems Date: Tue, 26 Jun 2001 16:12:43 -0500 From: "tazboy78" Organization: http://extra.newsguy.com Message-ID: <9hatqk02v83@enews1.newsguy.com> References: > that WRS has said that their RPC library is buggy That's kinda funny. RPC is written by Sun. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Threads in VxWorks Date: Tue, 26 Jun 2001 16:39:19 -0500 From: "tazboy78" Organization: http://extra.newsguy.com Message-ID: <9havch031b0@enews1.newsguy.com> References: <9hamvv$hg1$1@overload.lbl.gov> What is the difference between tasks and threads? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Threads in VxWorks Date: Tue, 26 Jun 2001 23:18:49 GMT From: Harvey Taylor Organization: MTS Internet Message-ID: <3B3933BE.5552@despam.autobahn.mb.ca> References: <9hamvv$hg1$1@overload.lbl.gov> In article <9hamvv$hg1$1@overload.lbl.gov>, Anuradha Pani wrote: > > I have a question, how are threads implemented in > Vxworks? I found several routines for tasks in taskLib > library, but none for threads. > There is no direct support for threads in VxWorks. When you have a flat address space, and no protection, you don't really need them. If it makes you feel better, think of tasks as fat threads. - -het - -- "So long as it's hidden in the OS, I don't care if it uses blue men from mars. The OS _can_ use any hardware primitives it likes." - -Randell Jesup (AmigaDOS) Harvey Taylor http://www.autobahn.mb.ca/~het/ --------------------------- Newsgroups: comp.os.vxworks Subject: ENUM# signal Date: Tue, 26 Jun 2001 18:25:09 +0900 From: "msjeon" Organization: Korea Telecom Hitel Message-ID: <9hb848$ipb$1@news.hitel.net> Hi,ALL I want to know about ENUM# signal . Help me --------------------------- Newsgroups: comp.os.vxworks Subject: Re: change sector size? HELP! Date: Wed, 27 Jun 2001 00:44:31 GMT From: "George" Organization: Excite@Home - The Leader in Broadband http://home.com/faster Message-ID: References: <52ccd904a%molochai@vapour-trail.demon.co.uk> Reply-To: "George" Of a disk drive. "molochai" wrote in message news:52ccd904a%molochai@vapour-trail.demon.co.uk... > In message > "George" wrote: > > > How do you change the sector size from 512 bytes to 2048? WindRiver has > > not been much help :-( > > > > Of what? > --------------------------- Newsgroups: comp.os.vxworks Subject: loadModule + memDev = loaded symbols? Date: Wed, 27 Jun 2001 01:01:15 GMT From: "Kelly Hornsby" Organization: Excite@Home - The Leader in Broadband http://home.com/faster Message-ID: Hello All I'm trying to use a memDev pseudo-file (located in flash memory) called /mem/flash to keep an ELF file that I flashed from my tornado project. It's a very simple program that prints "hello world\n", and is called hello.o. Very simple. I've verified the file is identical to the standalone .o file in my Tornado project. I find that I can load this file from the shell using ld < hello.o, and it will load properly and will execute. furthermore, the symbol will appear in the symbol table. When I do the equivalent thing from my /mem/flash memDev file that I created, the loadModule runs to completion, but nothing appears in the symbol table. I pass the flag ALL_SYMBOLS to loadModule, but I still get no symbols. The memDev is working properly because I can open the file and read it programmatically. Has anyone done this before... load a file from memory or flash into the system via a loadModule and a memDev? I'm looking for gotchas, faults or success stories. Any ideas or thoughts are welcome. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Dynamic linking/loading.... Date: Wed, 27 Jun 2001 11:15:05 +0100 From: Vijay Kumar Peshkar Organization: Cisco Systems Inc. Message-ID: <3B39B229.C8CDB9EB@wipro.com> References: <9ha8u0$886$1@overload.lbl.gov> This is a multi-part message in MIME format. - --------------609F09496F6B0CB3210DFED2 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Praveen, Have a look at "8.4.4 Downloading an Application Module" in prog. guide. cheers, Vijay Mike Anderson wrote: > VxWorks Greetings! > > > Submitted-by: "Praveen Nadgir" > > > > Hi VxWorks Gurus, > > > > I have a simple question: > > > > In vxworks, how does dynamic linking/loading work? > > Is it used or not? > > > > > > Are you referring to the dynamic loading of the VxWorks > shell, or the C++ dynamic binding thingy? If you're asking > about VxWorks dynamic loading, then that is performed using > the shell (WindSh or tShell) and a symbol table. As you try > to load a piece of code with unresolved references, the symbol > table is interrogated and addresses are resolved automagically. > Any external symbols are, by default, added to the symbol tables > as a result of the loading process. This is obviously the > 30,000 ft view of the process. > > HTH, > > Mike > -- > ======================================================= > T h e P T R G r o u p, I n c. > ======================================================= > Embedded, Real-Time Solutions and Training > > Mike Anderson mailto:mike@ThePTRGroup.com > Chief Scientist http://www.ThePTRGroup.com > V: (703) 585-9384 F: (703) 430-3748 > ======================================================= > "Software development is like making a baby... > You can't make a baby in one month by impregnating > nine women. Some things just take time." - --------------609F09496F6B0CB3210DFED2 Content-Type: text/x-vcard; charset=us-ascii; name="vijay.peshkar.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Vijay Kumar Peshkar Content-Disposition: attachment; filename="vijay.peshkar.vcf" begin:vcard n:Peshkar;Vijay Kumar tel;work:5530034/35/35 x2237 x-mozilla-html:FALSE org:Wipro Global R & D;Wishing you blue skies, a couple of daisies, some fresh air and a smile! adr:;;;;;; version:2.1 email;internet:vpeshkar@cisco.com note:Wishing you blue skies, a couple of daisies, some fresh air and a smile! fn:Vijay Kumar Peshkar end:vcard - --------------609F09496F6B0CB3210DFED2-- --------------------------- Newsgroups: comp.os.vxworks Subject: Debug vars / system mBlk config Date: Wed, 27 Jun 2001 06:42:15 +0100 From: Alan Dennis Organization: Customer of PlusNet Message-ID: Reply-To: news@lakesgc.force9.co.uk Hi VxWorks Gurus, Two apparently unrelated questions but they arose out of the same ftp server problem (ftp error 426 caused by running out of system mBlks): 1)Through WindSurf I "discovered" the ftpdDebug and ftpDebug vars that are used for enabling debug output for ftp server and client operations. I also know of scsiDebug which does similar for scsi. Is there a list of other similar "undocumented" variables anywhere? Anyone know of any others that could be useful (someday)? 2) Is there any guidance anywhere on the number of mBlks needed for various network operations? Is there any way I can calculate the number to allocate based on the maximum expected network and application usage? Thanks, Alan. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Debug vars / system mBlk config Date: Tue, 26 Jun 2001 23:16:18 -0700 From: DrDiags Organization: laserlink.net Message-ID: <3B397A32.55F5C090@covad.net> References: Reply-To: drdiags@covad.net This is a multi-part message in MIME format. - --------------7BAB0E0070D9668839B285A3 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Alan, Since you didn't have a personal e-mail address, this may get bounced by the news reader, I have tried to paste the text but I seem to lack that skill, so I tried attaching the file in hopes that it will be pasted in-line. If not, you could try Wind Surf for guidelines on how to configure your network stack. Alan Dennis wrote: > > 2) Is there any guidance anywhere on the number of mBlks needed for > various network operations? Is there any way I can calculate the > number to allocate based on the maximum expected network and > application usage? > - --------------7BAB0E0070D9668839B285A3 Content-Type: text/plain; charset=iso-8859-1; name="netPerformance.txt" Content-Transfer-Encoding: 8bit Content-Disposition: inline; filename="netPerformance.txt" SENS/TORNADO 2.0 NETWORK STACK CONFIGURATION/PERFORMANCE ISSUES -------------------------------- 1. INTRODUCTION 1.1 NETWORK STACK VERSIONS 1.2 NET POOLS 1.2.1 CLUSTERS 1.2.2 CLUSTER BLOCKS 1.2.3 MBLOCKS 1.3 RECEIVING: DATA FLOW FROM DRIVER TO APPLICATION LEVEL 1.4 SENDING: DATA FLOW FROM APPLICATION LEVEL TO DRIVER 1.5 SOCKET BUFFERS 1.5.1 UDP SOCKET BUFFERS 1.5.2 TCP SOCKET BUFFERS 2. NETWORK MEMORY RESOURCES SUMMARY 2.1 MEMORY POOLS 2.1.1 NETWORK STACK SYSTEM POOL 2.1.2 NETWORK STACK DATA POOL 2.1.3 NETWORK DRIVER POOLS 2.2 RECEIVE QUEUES 2.2.1 IP PROTOCOL RECEIVE QUEUE 2.2.2 IP FRAGMENT REASSEMBLY QUEUE 2.2.3 ARP RECEIVE QUEUE 2.2.4 TCP REASSEMBLY QUEUE 2.2.5 SOCKET RECEIVE QUEUES 2.3 SEND QUEUES 2.3.1 SOCKET SEND QUEUES 2.3.2 IP NETWORK INTERFACE SEND QUEUES 3. ESTIMATING NETWORK RESOURCES REQUIREMENTS 3.1 SOCKETS AND ROUTES 3.2 SENDING AND RECEIVING 3.2.1 LARGE SOCKET RECEIVE BUFFERS 3.2.2 TCP SMALL PACKETS 3.2.3 SENDING 3.2.4 RECEIVING 3.2.5 APPLICATION ERRORS 4. CONFIGURATION 4.1 NET POOL CONFIGURATION 4.1.1 NETWORK STACK MEMORY POOLS 4.1.2 NETWORK DRIVER MEMORY POOLS 4.2 PROTOCOL RECEIVE QUEUES 4.3 IP SEND QUEUES 5. REFERENCES FIGURE 1. MEMORY POOLS AND QUEUES DEFAULT VALUES 1. INTRODUCTION ------------ This article provides background and guidelines for configuring and performance-tuning the VxWorks network stack. This article applies to the Tornado 1.0.1 stack with SENS installed, and to the Tornado 2.0 stack with or without the Network Protocol Toolkit product installed. The assumed network programming interface is the BSD sockets interface; the zbuf sockets interface is mentioned only in passing. Tornado/setup.log file shows which version of Tornado and which components are installed. The Tornado 2.0 VxWorks Network Programmer's Guide, Section 4.6.3, has the following comments regarding network stack memory configuration: "The values shown above are reasonable defaults, but the network requirements for your system could be radically different. For example, your use of the network stack could require more clusters of a particular size. Making such a determination is a matter of experimentation and analysis.", "Change these constants after you fully understand what they do. Setting inappropriate values can make the TCP/IP stack inoperable.", and "Carefully planned changes can significantly improve performance, but reckless changes can significantly reduce performance." As many users find out, the default values for the network stack configuration parameters are adequate to download a file, open a few sockets, add a few routes to the routing table, and run simple client/server applications. Anything beyond that makes the stack hang for a while and resume a short time later. The network system requires clusters (blocks of memory) of various sizes from different pools, used to store system data structures, incoming data, and outgoing data. Clusters carrying data are stored in various holding areas (queues) along the path from application level to the network driver and vice-versa. The number of clusters of each size and the capacity of the holding areas must be configured to suit the application needs. It is also necessary to ensure that the clusters are processed regularly so that they will be available to carry new data. If clusters are held in any area for too long, the system's throughput will decrease. If the accumulation continues, data flow will eventually stop until clusters are freed up. Analyzing application needs requires basic knowledge on how the network system works. This paper will describe: 1. How data flows from application level to network interface driver and vice-versa, for TCP and UDP; 2. The role of tNetTask; 3. How the configuration of network stack and driver memory pools, application socket buffer sizes, protocol input queues, and interface output queues interact with network performance. Understanding these factors would help the reader set the network stack configuration values appropriately. Tuning configuration parameters for performance is in essence estimating the resources needed to sustain data flowing at a certain rate. As in any other inventory situation, requirements must be estimated taking into account fluctuations in demand and delays. Having enough capacity to handle the worst case is the only sure way to ensure reliability. This approach may not be cost effective in many cases, but may be the only alternative for some. Only users can make that determination based on their application needs. Section 1 presents background information on basic network stack versions and terms, followed by a description of data flow for receiving and sending, and socket buffers. Section 2 presents a summary of the memory resources used by the network system and a description of the various queues where these resources wait is presented next. Sections 3 and 4 present some guidelines for estimating requirements and configuration information. Section 5 lists resources where additional information can be found in vxWorks Documentation, reference books and articles from: WindSurf->Technical Solutions II. 1.1 NETWORK STACK VERSIONS - -------------------------- The network stack in Tornado 1.0.1 is based on BSD4.3, which is essentially the network protocol IP over the link layer protocol Ethernet. BSD4.3 Ethernet network interface drivers do not support multicasting, or polled Ethernet for system mode debugging. Under T101 the SENS component was optional. Under Tornado 2.0, SENS is the one and only network stack. SENS features: a. BSD4.4 network stack. BSD4.4 has enhancements at the network stack level such as support for CIDR (variable length subnet masks), multicasting, and various protocols such as DHCP, and RIP. b. MUX The MUX interface facilitates connecting END drivers to other user-ported network protocols besides IP. END drivers can be written for link layer protocols besides Ethernet. c. BSD4.4 and END drivers for supported Ethernet network cards. BSD4.3 drivers were modified slightly to work with the BSD4.4 network stack. Either BSD4.4 or END drivers can be used for booting. END drivers are required for multicasting and system mode debugging over Ethernet. The SENS release provided END Ethernet drivers for some BSPs. The Tornado 2.0 release provides END Ethernet drivers for all supported BSPs. The free Network Protocol Toolkit (NPT) source product contains updated documentation for porting to the END driver model and source for the MUX and supporting network libraries. In addition, NPT contains information for porting to the NPT driver model which facilitates porting non-Ethernet drivers. Network drivers developers should get the Network Protocol ToolKit product from their sales representative. The outdated BETA User's Guide is still available from WindSurf->Documentation->Tornado 2.0 Documentation. For sending, the path from IP to the physical medium can take the BSD4.4 driver path, or the END driver path. Similarly, for receiving, the BSD4.4 driver and the END driver path both converge at the IP level. The details are hidden in the initialization process. The driver’s initialization process installs the proper function pointers in the ifnet structure that IP keeps for every configured interface. For example, ipAttach will install ipOutput as the if_output function pointer of the ifnet structure for an END driver. Similarly, ether_attach will install ether_output for a BSD4.4 driver. IP selects an interface to send a packet, based solely on the routing table. IP calls the if_output function of the selected interface, so IP does not need to know what kind of driver it is. The setup.log file found in the Tornado base directory shows which network stack version is installed: SENS 1.0: T101 + WindNet SENS installed under SETUP 1.4 SENS 1.1: T101 + WindNet SENS installed under SETUP 1.5 Tornado 2.0: SETUP 2.0. The word SENS does not appear since SENS is not an optional component anymore. Note: T101 ARM BSPs have SENS 1.0 installed by default. There was no SENS 1.1 patch update for ARM BSPs. 1.2 NET POOLS --------- Tornado 1.0.1 (without the SENS component installed) had a memory pool consisting of fixed data structures, BSD style mbufs, used both to store small stack data structures such as socket addresses, and for packet data. BSD mbufs were designed to facilitate passing data between network drivers and the network stack, and contain pointers that can be adjusted as protocol headers are added or stripped. BSD mbufs contain space within them to store small amounts of data. Larger amounts of data were stored in fixed-sized clusters (typically 2048 bytes), which could be referenced and shared by more than one mbuf. In T101 + SENS and in Tornado 2.0, the BSD network buffer implementation has been changed somewhat. Slightly modified BSD mbufs are retained, but always reference external clusters rather than carrying data directly; they are now called "mblocks." Pools of clusters of multiple sizes are used, and a new structure called the "cluster block" was added, supporting the zbuf sockets interface and multiple network pools in addition to cluster sharing. The file target/h/netBufLib.h has the definitions of the new memory system. The basic structure is the NET_POOL, which contains mblocks, cluster blocks, and clusters of one or more sizes. The network stack and some network drivers create different NET_POOLs: a) The network stack DATA pool is used for actual data and headers transferred between the application, the stack, and the network interface. b) The network stack SYSTEM pool is used to allocate clusters for holding network stack control structures such as sockets, protocol control blocks, routes, and so forth. c) Network interface pools are used by (some) network drivers to hold clusters for receiving packets off the wire, and for staging packets to be transmitted by the target. Users can create pools for their own purposes as well. The header target/h/net/mbuf.h maps the Berkeley style mbuf definitions to the new style. 1.2.1 CLUSTERS -------- A NET_POOL may contain several cluster pools, each one holding clusters of a single size. Valid cluster sizes range from 64 (2 to the 6th) to 65536 (2 to the 16th). However, not all conceivable combinations of cluster pool sizes are allowed within a net pool. The NET_POOL structure contains an array of pointers to cluster pools. Each row in the array corresponds to an allowed cluster size range, measured in bytes, starting at a power of 2. The array has room for a maximum of 11 entries, with indices ranging from 0 (6 - minimum cluster size log 2 = 6 - 6) to 10 (16 - 6). When configuring a NET_POOL, choose at most one cluster size from within the range of each row, as given by this table: Row 0: 64 <= cluster size <= 127 Row 1: 128 <= cluster size <= 255 Row 2: 256 <= cluster size <= 511 Row 3: 512 <= cluster size <= 1023 Row 4: 1024 <= cluster size <= 2047 ... Row 9: 32768 <= cluster size <= 65535 Row 10: 65536 = cluster size If 2 cluster sizes fall in the same row, the pointer to the second cluster will replace the first. 1.2.2 CLUSTER BLOCKS -------------- A cluster block is a control structure pointing to a cluster. It contains a count of the mblocks referencing the cluster, as well as an optional function pointer and three function arguments. If the function pointer (pClFreeRtn field) is non-null, the function is called with the arguments when the cluster is freed; this is used to support the zbuf sockets interface. Users can provide their own functions as well. If the pClFreeRtn function pointer is null, the network stack returns the cluster to the appropriate cluster pool. The address of the cluster pool is found by subtracting 4 bytes from the clNode.pClBuf member of the cluster block. One cluster block is required for each cluster. 1.2.3 MBLOCKS ------- Mblocks are control structures which point to a section of data within clusters, and link them together into packets and chains of packets. Mblocks also contain other information about the type of data contained in the clusters they reference. There should be at least as many mblocks as there are cluster blocks. Some extra mblocks are needed for the network stack DATA pool. The stack uses DATA pool mblocks to reference data in the clusters without copying. A TCP socket's send buffer is a chain of mblocks linking together clusters containing data which has not yet been acknowledged by the socket's peer endpoint. Some of this data is also referenced by additional mblocks used to form the TCP segments which are passed to the network driver for transmission. In UDP, additional DATA mblocks are used to store source address and port information for received UDP packets. A rough estimate (sometimes erroneously treated as a hard rule) is that there should be two times as many mblocks in the stack DATA pool as cluster blocks. This "rule" certainly does not apply to the stack SYSTEM pool or to network interface pools. If the Network Protocol Toolkit product has not been installed, the IP protocol driver allocates just two mbufs from the net pool of each END interface it attaches to. Only the network driver allocates from an END or NPT driver pool if NPT has been installed. BSD 4.4 drivers may allocate mbufs from the DATA pool. When freed, mblocks, clusters, and cluster blocks are returned to the pool from which they were allocated. 1.3 RECEIVING: DATA FLOW FROM DRIVER TO APPLICATION LEVEL ----------------------------------------------------- Network drivers, and protocols like TCP and IP, schedule tNetTask to do task level processing. IP and TCP schedule tNetTask to process timers: to drop IP fragments that timed out before being reassembled, and to send TCP delayed data acknowledgements, or retransmissions. Drivers schedule tNetTask to process received packets, and sometimes to perform miscellaneous other jobs. While processing received packets, protocols may need to send output messages (ICMP errors, TCP immediate ACKs or fast retransmits, TCP data sent in response to peer's window openings, ARP replies, and so forth). Together with timer-triggered TCP retransmissions or resets, these are cases when sending is done in the context of tNetTask. tNetTask is not involved when applications call output functions like sendto() or write(). Such processing is done within the context of the application task. (However, TCP data placed into a socket send buffer by an application's write() call may be later transmitted in the context of tNetTask when the send window opens.) Below is a overview of the receive processing done by tNetTask: When a packet is received, within the driver's interrupt service routine, netJobAdd() is called to add the driver's task level receive routine to a ring buffer of jobs (fixed capacity = 85 jobs) for tNetTask. This is done to minimize the time spent in the ISR with interrupts locked. If the job is successfully added, netTaskSemId is given to awaken tNetTask. If not, a ring buffer overflow log message appears. When tNetTask (priority 50) runs, it will first process all the jobs in the ring buffer. If any IP datagrams have been queued, it then calls ipintr() to process them, and finally calls arpintr() if any ARP packets are queued. Each time tNetTask runs a job from the ring buffer, and again before calling ipintr() and arpintr(), it takes the splSemId semaphore by calling splnet() to lock out other network stack tasks. The driver's task-level receive function calls the stack receive function. The driver lends the buffer to the stack, if possible. The stack strips the packet's link layer header to determine who should process the packet next. (NPT drivers do the header "parsing" for the stack; for other drivers, the stack must do so itself.). For END drivers, the MUX hands the packet to the appropriate protocol input routine; for ARP and IP packets, this is ipReceiveRtn(). BSD 4.4 drivers call do_protocol (ether_input). ipReceiveRtn and ether_input in turn call function do_protocol_with_type(), which places the packet in the appropriate protocol queue (IP = type 0x800), (ARP = 0x806), if there is room. The default size for both the IP and ARP input queues is 50 packets. The packet is dropped if the corresponding queue is full. IP determines if the packet arrived in good condition and whether it is a fragment or a full datagram. If it is a datagram, or if it is a fragment completing a datagram, and the datagram specifies a known transport-level protocol, IP then calls the appropriate protocol receive function (e.g. tcp_input, udp_input, icmp_input ...). The receiving protocol processes the packet. If there is room in the socket receive buffer, the data is appended to the socket's receive buffer. If a task was blocked waiting to read data from the socket, it is given a semaphore; the task is then READY to run. When the priority level is such that the task runs and reads from the socket, the socket layer copies the data to the buffer provided by the user. When all the data from a cluster has been copied to the application's buffers, the cluster is returned to the driver's pool, or to the network stack DATA pool if a BSD4.4 driver had to borrow from the network stack DATA pool. (If the zbuf interface is used, the cluster is not returned to its pool until the received zbuf is deleted.) Some output may occur during processing of received packets. IP may need to send various ICMP messages. If the received packet is a UDP packet, and there is no socket bound to the destination port, UDP will call icmp_error to send an ICMP Unreachable error message. For TCP, tcp_output() may be called during reception for a number of reasons, for example, to send an immediate ACK if an out-of-sequence segment is received, to do a fast retransmit, or to send more available data, if the application is sending data, in response to a window opening. (If the application is receiving data, window updates may be sent to the peer after the application reads data from the socket buffer, but this does not happen in the context of tNetTask.) The function ip_output(), and ultimately the driver's output function are eventually called, still in the context of tNetTask. 1.4 SENDING: DATA FLOW FROM APPLICATION LEVEL TO DRIVER --------------------------------------------------- In most cases transmission of application data to the network occurs in the context of application tasks. tNetTask is involved in just a few cases: tNetTask may transmit TCP application data, both during retransmission, and when data earlier placed in the socket buffer by the application is later sent in response to the send window opening; and tNetTask may transmit any IP datagram held earlier pending an ARP reply. When send(), sendto(), or write() is called at the application level, the following happens: splSemId semaphore is taken by calling splnet to lock out other network stack tasks. If there is room in the socket send buffer, the socket layer copies data from application level buffers to clusters allocated from the network stack DATA memory pool (_pNetDpool). For example, if an application writes 1460 bytes, the socket layer will request a buffer big enough to contain the data plus some space for link layer headers. The system will return the closest size available. The data will be copied to a single buffer (2048 bytes, in the default configuration), if available, or to a chain of smaller buffers, otherwise. If there are no buffers, the system attempts to find buffers, and if it is not successful sendto, write ... will return ENOBUFS for non-blocking AND blocking sockets. The socket layer then calls the protocol user request function for output, which in turn calls the actual protocol output function (e.g., udp_output(), tcp_output()) to process the data. After the appropriate protocol processes the data and adds its header, one or more packets may be passed to IP. Based on the destination IP address, IP locates the appropriate route and calls the output function for the outgoing interface corresponding to that route. If the packet is too large for the outgoing interface's MTU, IP will fragment it before calling the interface output function. The output function will resolve the address to obtain a hardware address, if necessary. It will then add the packet to the send queue for the appropriate interface, if there is room. (When a network interface is attached to IP, the default size for the send queue is set at 50 packets.) ENOBUFS is returned if there is no room in the send queue, and the packet is dropped. The appropriate if_start() function is called to pass every packet in the interface send queue to the driver's transmit function. After all packets have been passed to the driver (or the driver has stopped accepting packets until a transmit restart) splx() is called to release splSemId. Each driver (END or BSD) is different. Some END drivers have clusters for sending, others use different data structures called various names such as descriptors or frames. Some BSD drivers have 1 transmit frame, others have many. Whether the driver uses clusters or frames, it still has to find a free one. In general for END DRIVERS, using clusters: Within the driver's output function, the driver takes its own transmit semaphore which is SEM_INVERSION_SAFE The driver attempts to allocate a buffer from its pool to copy the packet passed by the stack. (The copy is done for two reasons: the hardware may not support gather-write, and the stack buffers containing the data may not be cache-coherent.) If the driver does not have enough buffers, it returns END_ERR_BLOCK. The stack will prepend the packet to the send queue and try again the next time a packet is queued for the interface, or after the driver calls the MUX transmit restart routine. If the driver has enough buffers, the driver frees the network stack DATA buffer. It then transmits the packet. The driver gives its transmission semaphore. UDP frees the network stack DATA buffers as soon as the message is sent. TCP, on the other hand, holds the DATA buffers for possible retransmission until it receives an acknowledgement from the peer TCP. 1.5 SOCKET BUFFERS -------------- Socket buffers are used to store messages and other information required by the protocols. As compiled, the maximum size for a socket receive or send buffer is 258,111 bytes. When a new UDP socket is created, its receive and send buffer are set to the value of the global variables udp_recvspace and udp_sendspace respectively. Similarly, TCP sockets buffers are set to the value of the global variables tcp_recvspace and tcp_sendspace. The values of the global variables are: - -> tcp_sendspace tcp_sendspace = 0xf9e00: value = 8192 = 0x2000 - -> tcp_recvspace tcp_recvspace = 0xf9e04: value = 8192 = 0x2000 - -> udp_sendspace udp_sendspace = 0xfec74: value = 9216 = 0x2400 - -> udp_recvspace udp_recvspace = 0xfec78: value = 41600 = 0xa280 Another value set when a socket is created, or changed when setsockopt is called with options SO_SNDBUF/SO_RCVBUF, is the maximum number of mbufs (in bytes) that a socket may hold. mbufs is the same as clusters. The limit is set to avoid having a single socket consume too many mbufs and is calculated as follows: max number of mbufs = minimum ( 5* maximum buffer size, 256*1024) (in bytes) For both UDP and TCP, space in the send or receive buffer is calculated using the formula below prior to sending or receiving: space = min (buffer maximum size - total bytes in buffer), (max number of mbufs - total mbuf bytes in buffer) To change the default buffer sizes for all sockets, change the value of the global variables from the shell or programmatically. Use setsockopt to change the value for an individual socket. Calling setsockopt to set the maximum size for buffer space just declares what the limit will be. No memory is actually allocated until a packet is sent or received. 1.5.1 UDP SOCKET BUFFERS ------------------ For sending, if the message size is larger than the socket send buffer, EMSGSIZE error is returned. For receiving, since UDP is connectionless, UDP stores the message source IP address and port in the receive buffer. Therefore, an additional 16 bytes per message are needed for the receive buffer. If there is no space, as determined by the space calculation above, received packet is dropped and the error is reported by udpstatShow as full sockets error. For example, for a 172 byte UDP message received over Ethernet: The 172 byte message is within a buffer of size 1520. The address and port information is stored in a separate 128 byte buffer. Each buffer requires a control data structure, an mblock (size = 32 bytes): # mbuf bytes = packet mbuf bytes + address information mbuf bytes # mbuf bytes = 1520 + 32 + 128 + 32 = 1712 bytes For a 2000 byte UDP message, since the maximum UDP message size over Ethernet is 1472 bytes (1500 - 28 byte header), 2 packets are received: # mbuf bytes = (1520 + 32) * 2 + 128 + 32 = 3264 bytes 1.5.2 TCP SOCKET BUFFERS5 ------------------ For TCP sockets, the socket receive/send buffer sizes determine the maximum send and receive window sizes for the connection. setsockopt must be called before the listen call on the server, and before the connect call on the client, since maximum window sizes are among the parameters exchanged when the connection is first established. Increasing the buffers later has no effect. Periodically TCP sends window updates informing the peer how much data it can receive. The advertised window size is the available room in the receive buffer. The window size cannot exceed the maximum set when the connection was established, but can decrease to 0 if the socket is not read, and the receive buffer becomes full. To improve TCP performance, Richard Stevens (UNIX Network Programming Volume 1, Second edition, Page 192), suggests that: 1. socket buffer sizes should be at least 3 times the MSS. 2. buffer sizes should be an even multiple of the MSS for the connection and 3. Buffer sizes should be at least as large as the product of the bandwidth (bytes/sec) and the round trip time to keep the "pipe" full and get the expected performance. Round trip times (RTT) are obtained by using ping. For example, if the bandwidth is 1000 bytes per second, and the average round trip time is 2 seconds, buffer sizes would have to be at least 2000 bytes. For Ethernet, and an average RTT = 50 msecs: 10000000/8 bytes * 0.050 seconds = 62500 bytes which is approximately 42.8 * MSS. Following 1 and 2 above, socket buffer size should be 64240 (44*1460). On an Ethernet which is not overloaded, RTT is much faster than 50 msecs. The window size field in the TCP header is an unsigned short (maximum = 65535 bytes). The Tornado 2.0 stack implements RFC1323, the window scale option (See netLib.h), which is designed to get around the 65535 limit for high speed connections (ATM) or slow media (satellite) which require large window sizes. 2. NETWORK MEMORY RESOURCES SUMMARY -------------------------------- There are 3 memory pools consisting of clusters and their control structures, which are allocated at initialization time. These pools cannot be increased later: NETWORK STACK SYSTEM POOL NETWORK STACK DATA POOL NETWORK DRIVER POOLS There are 7 types of queues (linked lists) that hold data waiting to be processed. The data is contained in clusters previously allocated from the DATA or the DRIVER pools. Adding a cluster to a queue is just chaining the cluster to the queue's linked list. The stack sets default values for the maximum length of some queues, which can be changed after initialization: RECEIVE QUEUES: 1. IP PROTOCOL RECEIVE QUEUE 2. IP FRAGMENT REASSEMBLY QUEUE 3. ARP RECEIVE QUEUE 4. TCP REASSEMBLY QUEUE 5. SOCKET RECEIVE QUEUES SEND QUEUES: 1. SOCKET SEND QUEUES 2. IP NETWORK INTERFACE SEND QUEUES 2.1 MEMORY POOLS ------------ 1. NETWORK STACK SYSTEM POOL ------------------------- This pool is used to store data structures such as sockets and routes. The default values are: ------ CLUSTERS REQUIRED ----- CLUSTER SIZE TOTAL PER TCP PER UDP PER ROUTE BYTES SOCKET SOCKET ARP MAX ** NUM_SYS_64 40 0 0 2 3 NUM_SYS_128 40 1 1 0 0 NUM_SYS_256 40 1 0 1 2 NUM_SYS_512 20 1 1 0 0 **: Route requirements vary. For example, the first time a network route is added with a new gateway, (Example; routeAdd("147.11.48.0", "147.11.41.254") 3 64 and 2 256 clusters (3,0,2,0) are allocated. If another route is added using the same gateway, (Example: routeAdd("147.20.30.40", "147.11.41.254") (1,0,1,0) clusters are allocated. If an ARP entry is added with arpAdd, (2,0,1,0) clusters are allocated. routeAdd, arpAdd themselves open a raw socket (0,1,0,1) temporarily to pass information to the stack. 2. NETWORK STACK DATA POOL ----------------------- This pool is used to copy data passed by applications (sendto, write...) from application level buffers to network stack clusters. The clusters are passed to the protocols and ultimately to the network drivers. Protocols (ICMP, TCP, RIP, ...) send messages too. The stack allocates a cluster big enough to hold the message plus some extra space for protocol headers. Smaller clusters would be allocated and chained if the right size is not available. CLUSTER SIZE TOTAL BYTES NUM_64 100 NUM_128 100 NUM_256 40 NUM_512 40 NUM_1024 25 NUM_2048 25 For example, if an application calls "write" to send one 1460 byte message for a TCP socket, 1 2048 byte cluster would be allocated from the pool. TCP would hold this cluster in the socket send queue until the recipient notifies that the data was received. Clusters used by UDP are freed as soon as the driver copies the data to its own clusters. 3. NETWORK DRIVER POOLS -------------------- Tornado 2.0 offers 2 driver versions for each supported network card: BSD 4.4 and END. Either type can be used, but END drivers are required for multicasting and for system mode debugging over Ethernet. END drivers are configured by default in Tornado 2.0. muxShow displays configured END drivers. ifShow displays both END and BSD4.4 drivers. Drivers set up their own pools of non-cacheable memory and allocate buffers from it which are used for sending and receiving. Buffers can be clusters or other data structure types called frames or descriptors. Drivers copy data passed by the stack to their own buffers for transmission. Devices receive packets in the driver's buffers. END and BSD4.4 drivers (some, not all) lend buffers containing received packets to the network stack to avoid copying. Buffers have to be big enough to receive or transmit the maximum packet size allowed by the link layer. For Ethernet 10 Base T, the maximum size is 1518 bytes. Two extra bytes are required to align the IP header on a 4 byte boundary for incoming data. The default size of end driver cluster pools vary, but the number could be as few as 40 1520 clusters total. In this article, when it comes to drivers, clusters are a generic name for buffers: descriptors, frames, or actual cluster data structures. Drivers differ: END DRIVERS ----------- Some END drivers have frames for transmission and clusters for reception. Some END drivers have clusters for both. END network drivers lend all their clusters. BSD 4.4 DRIVERS --------------- BSD drivers have frames. BSD drivers borrow mblocks and cluster blocks from the network stack DATA pool when they lend their frames to the stack. BSD drivers may borrow clusters as well if they have lent the configured maximum number of LOAN frames to the stack. Some BSD drivers always borrow clusters. 2.2 RECEIVE QUEUES -------------- 1. IP PROTOCOL RECEIVE QUEUE ------------------------- This queue stores DRIVER or DATA cluster containing incoming packets passed to IP from all configured interfaces. IP processes the packets and passes them to the protocols (TCP, UDP...). Protocols process the packets and queue data in the destination socket receive buffers. If packets come in bursts and the queue is full, the excess is dropped. The default value for this queue is 50 packets. 2. IP FRAGMENT REASSEMBLY QUEUE ---------------------------- This queue stores DATA or DRIVER clusters that contain fragments of a larger IP datagram waiting to be reassembled. The maximum size for an IP packet is 65535 bytes. Packets exceeding the maximum transmission unit (MTU) for a link layer protocol must be fragmented. For example, for Ethernet 10BaseT, the maximum packet size is 1518 bytes. Subtracting the Ethernet header and checksum (18 bytes), the IP header (20 bytes), and the UDP header (8 bytes) leaves 1472 bytes. A UDP message consisting of 1473 bytes arrives in 2 separate IP packets. There is a default timeout to drop the fragments if they cannot be reassembled in time. There is no explicit queue size. IP fragments are dropped also when a pool is out of clusters and the system drains the protocols. 3. ARP RECEIVE QUEUE ----------------- This queue holds DATA or DRIVER buffers containing ARP packets which are used for hardware address resolution. If packets come in bursts and the queue is full, the excess is dropped. The default value for this queue is 50 packets. 4. TCP REASSEMBLY QUEUE -------------------- This queue contains DATA or DRIVER clusters containing TCP segments that arrive out of order. For example, the maximum TCP message size over Ethernet 10BaseT is 1460 bytes (assuming no TCP options). TCP avoids IP fragmentation, so each segment is 1460 bytes maximum. The socket receive/send buffer sizes determine the maximum send and receive window sizes for the connection. If socket receive buffers are large, the peer TCP will be able to send a large number of segments within the window. TCP has to pass data to the application in the order in which it was sent. When packets arrive out of order, or are lost, TCP stores these segments in a reassembly queue until the missing segment arrives. There is no limit on the queue size nor is there a timeout period to drop the fragments. 5. SOCKET RECEIVE QUEUES --------------------- Each socket's receive buffer stores the DATA or DRIVER clusters until the application reads the socket, at which point, the clusters would be freed. This is important: Driver or DATA clusters are freed IF AND WHEN applications read the sockets. If they don't, data flow may stop eventually. The default value is 8192 bytes for TCP sockets and 41600 bytes for UDP sockets. 2.3 SEND QUEUES ----------- 1. SOCKET SEND QUEUES ------------------ Each TCP socket's send buffer holds DATA clusters containing packets waiting to be acknowledged. The default value for TCP sockets send queues is 8192 bytes. UDP sockets send queues do not hold any data, but the stack checks that the message size does not exceed the send socket buffer size (9216 bytes). A DATA cluster is freed as soon as the driver copies the data to its own clusters prior to transmission. 2. IP NETWORK INTERFACE SEND QUEUES -------------------------------- IP has a send queue for each configured interface which stores the outgoing packets, DATA buffers, that IP needs to pass to the driver for transmission. If a driver is busy transmitting, and this queue is full, the packet is dropped and ENOBUFS error is returned to the application. The default value is 50 packets. FIGURE 1: MEMORY POOLS AND QUEUES DEFAULT VALUES -------------------------------------- RAW SOCK SEND BUFF SYSTEM POOL ****************** ****************** * DATA *-----------> * SOCKETS * * CLUSTERS * | * ROUTES * * 8192 bytes * | * PCBs, TCPCB * ****************** | ****************** IP SEND QUEUE UDP SOCK SEND BUFF FOR "FEI" END DRIVER ****************** ****************** * DATA CLUSTERS *--->* DATA CLUSTERS *----------> * 9216 bytes * * 50 PACKETS * | ****************** ****************** | | IP SEND QUEUE | TCP SOCK SEND BUFF FOR "ELT" BSD 4.4 DRIVER | ****************** ****************** | * DATA CLUSTERS *--->* DATA CLUSTERS *----------|-------- * 8192 bytes * * 50 PACKETS * | | ****************** ****************** | | | | ******************** | | * ARP RECEIVE QUEUE*<-<--------|-------|< **** 50 PACKETS *** | | || | V || UDP SOCK RCV BUFF IP FRAGMENT QUEUE | FEI END DRIVER || ****************** *************** | ************** || * DRIVER, DATA * * DRIVER, DATA* |<*END DRIVER * || * 41600 bytes *<---*CLUSTERS/FRAM* * * || ****************** *************** *RECV: * || | ^ --*CLUSTERS: 64* || <----------------- | | *CLBLKs: 64 * || | | | *MBLK: 128 * || RAW SOCK RECV BUFF IP RECEIVE QUEUE | *TX FRAMES:32* || ****************** *************** | ************** || * DRIVER, DATA * *DRIVER,DATA *<--| || *CLUSTERS/FRAMES *<---*CLUSTERS/FRAM* || * 8192 bytes * * 50 PACKETS *<-- ELT BSD4.4 DRIVER|| ****************** *************** | ************** || | | | *BSD44 DRIVER* || <--------------- | | | * * || V V | *FRAME * || TCP SOCK RECV BUFF TCP REASSEMBLY Q. |-*DESCRIPTORS:*<- | ****************** *************** * Rx: 20 * | * DRIVER, DATA *<---* DRIVER,DATA * * Tx: 1 *---> * 8192 bytes * *CLUSTERS/FRAM* * Loan: 16 * ****************** *************** ************** 3. ESTIMATING NETWORK RESOURCES REQUIREMENTS Network stack memory exhaustion problem is not unique to vxWorks. TCP and UDP do not prioritize sockets. There are no facilities to identify sockets which could be considered less important and which could be targeted for selective drops, or even closing, when resources are limited. Packets will be dropped during peak times if there are not enough driver buffers to lend to the stack to support the sum of the UDP and TCP socket receive buffers for the expected number of concurrent socket connections. This sum could be a very large number. Even if there is a single socket, there could be still be high demand. For example, applications expecting to receive bursts of 100 UDP packets even if they carry only one byte of data will still be consuming 100 driver buffers. Similarly, applications will hang and resume after a while if there are not enough SYSTEM clusters to open sockets or create routes for applications that generate a high demand for them. 3.1 SOCKETS AND ROUTES ------------------ If the stack does not have enough SYSTEM clusters, applications like FTP and Web servers can hang for a while when there is not enough memory to open sockets. iosFdShow displays open sockets. inetstatShow displays socket status. When TCP on the target closes a socket, the socket is removed from the file descriptor table, but the memory used by that socket is not de-allocated until TIME_WAIT time elapses (about 1 minute). The TCP protocol needs this time to minimize the chance that duplicate packets belonging to an earlier connection will be delivered to a new TCP connection. inetstatShow displays the socket states. If a routing protocol like RIP is configured, or if, for example, the target is not given an explicit subnet mask, when necessary, the route table can get very large since the target is really listening to ARP broadcasts for all subnets. A message may appear on the on the console: arpresolve: can't allocate llinfo netStackSysPool can be used to tune the SYSTEM pool. See 5.1.1 NETWORK STACK MEMORY POOLS for details. 3.2 SENDING AND RECEIVING --------------------- Some conditions that would generate a high demand for driver buffers: 3.2.1 LARGE SOCKET RECEIVE BUFFERS ---------------------------- A UDP or TCP message consisting of a single byte of user data will be received in a whole driver buffer. Large socket receive buffers can contain many driver buffers. UDP messages larger than the MTU (For Ethernet = 1472 bytes) are fragmented by IP. IP on the receiving side will have to reassemble the fragments. Each fragment will be received on a separate driver or DATA buffer. TCP does not send packets larger than the MTU. However, if socket receive buffers are large, the peer TCP will be able to send a large number of segments within the window. Since TCP has to pass data to the application in the order in which it was sent, when packets arrive out of order, or are lost, TCP stores these segments in a reassembly queue until the missing segment arrives. TCP has no provision to drop packets in the reassembly queue, so driver buffers can sit there for a long time or indefinitely. TCP on the receiver side sends ACKS for the missing segment every time it gets an out of sequence packet. TCP on the sender side retransmits the missing segment, or even starts over again retransmitting the first unacknowledged segment if it does not get any response. TCP on the sender side retries a number of times but eventually it gives up. This situation could happen if all of the driver buffers are held up in the TCP reassembly queue, and there are no available driver buffers for transmission or to receive the missing segment. 3.2.2 SMALL TCP PACKETS ----------------- TCP normally does not send small packets, but many users enable TCP_NODELAY to force TCP to do so. Enabling TCP_NODELAY on the sender side (See entry for setsockopt in vxWorks Reference Manual) prevents TCP from waiting to collect small packets until an MSS sized packet can be sent. For TCP over Ethernet: MSS = 1460 bytes. TCP on the sender side still is subject to the constraint that it must not send more than the peer's advertised window. It should be noted that enabling TCP_NODELAY on the receiver side will not have a direct effect on how fast the receiver side acknowledges data. When an application reads a socket, the TCP on the receiver's side calculates whether it should tell the peer that is OK to send more. It does so by sending ACKS. This ACK is not immediate. The following calculation takes place: Compare available window to amount of window known to peer. If the difference is at least two maximum size segments (For TCP = 1460 bytes) or at least 50% of the maximum possible window, then send a window update to peer. Otherwise, the ACK gets delayed until the FAST TIMER goes off about 200 msecs later. TCP works this way following RFCs to avoid contributing to network congestion. If the application message size is small (less than 1460 bytes), reducing the socket receive buffer size would make ACKS come faster. If the application message size is larger than MSS (1460 bytes), you should test various message sizes (1 MSS, 2 MSS, ...) to find a level that gives good performance. Surprisingly, large buffers are not always better. TCP ACK-every-other-segment feature improves performance since the 200 msec delay would not occur as often. It should be noted that TCP was developed to maximize reliability, not performance. TCP wants to be efficient also, so it tries not to add unnecessary packets on the network thereby increasing congestion. The THEORETICAL maximum throughput, under ideal conditions, for a single TCP socket connection on 10 BASET, is 1,183,667 bytes/second. In reality, there are delays. Achieving 90% of this maximum is considered very good. TCP can be tuned to give a reasonable performance, but the reader should understand its limitations by consulting TCP/IP reference books. 3.2.3 SENDING ------- Sending a lot of packets requires the following steps: STEP 1: INCREASE THE NETWORK STACK DATA POOL ------------------------------------ netStackDataPoolShow displays the status of the DATA pool. If there are no free buffers or if the statistics for the number of times failed to find space is greater than 0, increase the appropriate cluster pools based on the application message size. Detailed instructions can be found in Section 4.1.1 NETWORK STACK MEMORY POOLS. STEP 2: INCREASE THE NETWORK DRIVER BUFFER POOL --------------------------------------- See Section 4.1.2 NETWORK DRIVER MEMORY POOLS. STEP 3: INCREASE THE NETWORK INTERFACE SEND QUEUE ----------------------------------------- See Section 4.3 IP SEND QUEUES 3.2.4 RECEIVING --------- Receiving lots of packets for many socket connections requires the following steps: EITHER STEP 1: INCREASE THE NETWORK DRIVER BUFFER POOL --------------------------------------- If an application will be using the network intensively, performance will be enhanced if the network interface driver is configured with a sufficient number of buffers so that packets are not dropped unnecessarily. See Section 4.1.2 NETWORK DRIVER MEMORY POOLS. A utility like endPoolShow, found in attached utilities.c can help the user find the optimal number for END drivers. For BSD4.4 drivers, there is no specific function that can show how many times the driver was out of its own buffers and had to borrow from the network stack DATA pool. However, netStackDataPoolShow can be helpful. Both endPoolShow and netStackDataPoolShow would display if buffers were exhausted at any time. STEP 2: INCREASE IP INPUT QUEUE ----------------------- See Section 4.2 PROTOCOL RECEIVE QUEUES. OR STEP 1: DECREASE SOCKET RECEIVE BUFFERS STEP 2: DECREASE THE NUMBER OF CONCURRENT SOCKET CONNECTIONS STEP 3: ADJUST APPLICATION MESSAGE SIZE AS NECESSARY TO BE CLOSER TO THE MTU TO IMPROVE EFFICIENCY. 3.2.5 APPLICATION ERRORS ------------------ The network driver buffer mechanism can be disrupted if tNetTask cannot process packets, or applications errors have deadlock conditions which prevent the sockets from being read. The priority of applications that need the network should be lower (>50) than tNetTask's (= 50). User's applications usually run at priority 100. Call i from the shell to display priorities for running tasks. tNetTask takes splSemId which is SEM_INVERSION_SAFE. Do not create and then take SEM_INVERSION_SAFE semaphores before making a socket call or your task could be promoted to run at tNetTask level. See priority inversion discussion in the vxWorks Programmer's Guide (Section 2.4.3). If priority inversion occurred, the command i would show it as 0+I: - ->i tNetTask netTask 807c9750 0+I PEND Application may have deadlock conditions which prevent them from reading sockets. If inetstatShow (or equivalent in other systems) displays data backed up on the send side and on the receive side of the peer, most likely there is a deadlock situation within the client/server application code. TCP is a stream of bytes. Loops are required to read or write, since TCP can return partial reads and partial writes. Running both server and client in the target by sending to 127.0.0.l or to the target's own IP address is a good way to detect this kind of problem. 4. CONFIGURATION See WindSurf->Technical Articles II: How to build images with network support in Tornado 2.0 which contains instructions for building a kernel with network show libraries and a target shell whose output goes to serial console. The target shell can be used to debug network problems when network problems occur. The attached utilities.c file contain code for some utilities which are not part of the API: endPoolShow protocolQValuesShow changeProtocolInputQMax ifQValuesShow changeIfSendQMax 4.1 POOL SHOW FUNCTIONS ------------------- Call pool show functions to find the level where resources are adequate: netStackSysPoolShow netStackDataPoolShow endPoolShow (found in utilities file). Pool show functions display: number of times failed to find space: If not 0, pool is too small. number of times waited for space: 0 number of times drained protocols for space: IP fragments are dropped when drivers or the stack are out of clusters. Usage: the cumulative number of times clusters of each size have been allocated. Clusters: Total clusters: Free + clusters in queues. Mbufs and their use. Increase clusters, cluster blocks and mblocks as necessary until there are no shortages. 4.1.1 NETWORK STACK MEMORY POOLS -------------------------- Increase the default values as follows: 1. INCREASE NUM_FILES The socket library uses NUM_FILES to configure the size of the socket file descriptor table. The default value for NUM_FILES is 50 sockets. 2. TCP SOCKETS For every TCP socket, the system allocates 3 clusters (1 128 byte, 1 256 byte, and 1 512 byte clusters) used for the generic protocol control block (inpcb), TCP protocol control block, and the socket structure respectively. 3. UDP SOCKETS For every UDP socket, the system allocates 2 clusters (1 128 byte, 1 512 byte) used for the inpcb and socket structures respectively. 4. ROUTES For every link layer route (ARP), 3 clusters (2 64 byte, and 1 256 byte) clusters are allocated. BOOTROM IMAGE MACROS -------------------- The macros with MIN suffix (i.e. NUM_SYS_64_MIN... and NUM_64_MIN...) are used for BOOTROM images which usually need to be small. The values for the MIN macros are usually adequate to download the runtime image and do not normally need to be changed. RUNTIME IMAGE MACROS -------------------- The network stack SYSTEM POOL macros are NUM_SYS_64, NUM_SYS_128, ...etc. The network stack DATA pool MACROS are NUM_64, NUM_128...etc. TORNADO 101 BUILD METHOD ------------------------ a. Increase NUM_FILES in target/config/all/configAll.h. b. Increase the default network stack memory sizes in netBufLib.h. PROJECT FACILITY BUILD METHOD ------------------------------ a. Increase NUM_FILES Right Click on network components-> networking protocols-> BSD SOCKET to bring up the properties window. Click on Params window to change NUM_FILES. b. Change values for the runtime image in the Params window for network buffer initialization component: network components-> basic network initialization components-> network buffer initialization-> 4.1.2 NETWORK DRIVER MEMORY POOLS --------------------------- The size of a driver's pool is determined by the parameters passed to the driver, during initialization time (See the VxWorks Reference entry for the driver, since each one is different. Drivers have a ring of buffers for transmission and reception. For some drivers, the ring size may be limited by the hardware to a maximum. BSD 4.4 DRIVERS - --------------- The LOAN parameter allows having extra buffers for loaning so the ring always has enough buffers. END DRIVERS - ----------- END drivers lend all their clusters. If END driver's cluster pools are depleted, reception will stop until the clusters are freed by the network stack or the applications IF and WHEN socket buffers are read. Transmission may stop as well if the driver uses clusters for transmission. For end drivers, target/config/bsp/configNet.h is the place to change the default parameters in the driver's load string. For BSD4.4 drivers, the parameters are passed to the attach function. See basicNetConfiguration document for more details. The number of buffers is typically the number of transmit descriptors plus the number of receive descriptors plus the number of loan buffers. If the loan parameter is not part of the initialization string, then changing the LOAN parameter in the driver's header file will have no effect unless you have source for the driver and can recompile the object. Some drivers just make loan buffers equal to the number of receive or send descriptors. Example: PPC target uses dc END driver. In configNet.h /* :::::::: */ #ifdef ULTRA # define DEC_LOAD_STRING "0x81020000:0x80000000:0x09:0x09:-1:-1:-1:0:0x0" #else # define DEC_LOAD_STRING "0x81020000:0x80000000:0x0a:0x0a:-1:-1:-1:0:0x0" #endif >From dec21x40End manual entry: The default -1 means use 32 buffers So I changed the default numbers to 128 each: #define DEC_LOAD_STRING "0x81020000:0x80000000:0x0a:0x0a:128:128:-1:0:0x0" NUM_LOAN is declared in dec21x40End.h as 16. The cluster pool available to the driver will be: numRds + numTds + NUM_LOAN = 128 + 128 + 16 = 272 The receive and transmit numbers are specified separately, but the clusters in the pool are in fact used for both sending and receiving, and all buffers are loaned. - -> endPoolShow("dc",0) type number - --------- ------ FREE : 510 DATA : 2 HEADER : 0 SOCKET : 0 PCB : 0 RTABLE : 0 HTABLE : 0 ATABLE : 0 SONAME : 0 ZOMBIE : 0 SOOPTS : 0 FTABLE : 0 RIGHTS : 0 IFADDR : 0 CONTROL : 0 OOBDATA : 0 IPMOPTS : 0 IPMADDR : 0 IFMADDR : 0 MRTABLE : 0 TOTAL : 512 number of mbufs: 512 number of times failed to find space: 0 number of times waited for space: 0 number of times drained protocols for space: 0 __________________ CLUSTER POOL TABLE _________________________________________________________ size clusters free usage 1520 272 144 450 - --------------------------------------------------------- 4.2 PROTOCOL RECEIVE QUEUES ----------------------- There are 2 global variables that control the queue length for the ARP queue and the IP queue: arpintrq, and ipintrq. The queues hold the unprocessed ARP and IP packets. Packets are dropped if the queue length is at the maximum level. The default value for each is 50. No error statistics are increased. If building the Tornado 101 way, change the default value for ipintrq.ifq_maxlen by changing the value for IP_QLEN_DFLT in target/h/netLib.h. If building using the Tornado 2.0 GUI: network components networking protocols core TCP/IP components Right Click on IPv4-> Properties-> Parameters The default values and drop statistics are displayed by protocolQValuesShow. Change the default value for ARP or IP at run time calling changeProtocolInputQMax. Neither is an API function. See utilities.c. 4.3 IP SEND QUEUES -------------- When an END, BSD4.4 driver, or the loopback driver is attached to IP, one of the parameters set by IP is the size of the maximum send queue length. On sending, ipOutput for END drivers or ether_output for BSD4.4 drivers attempts to add a packet to interface's send queue, one of the fields of the ifnet structure that each driver creates (See target/h/net/ if.h). Default queue size is 50 packets. Packet is dropped if queue is full, the ifq_drops statistics increases and the socket call returns ENOBUFS. Call ifQValuesShow to display the current values and drop statistics. Use changeIfSendQMax to increase the default maximum, if an application will be sending more than 50 packets in rapid succession. ifQValuesShow and changeIfSendQMax utilities can be found in the code section. 5. REFERENCES ---------- 5.1 VXWORKS DOCUMENTATION --------------------- The Tornado 2.0 Documentation Collection has the most up-to-date manual versions. Manuals are available from the Help Menu in Tornado and also from WindSurf. In addition to manuals, WindSurf has Patches, Topic keyword search engine, and Networking Demo Code among other things. www.windriver.com -> Support -> WindSurf Access to WindSurf requires a password. Register at: http://www.windriver.com/corporate/support/private/validate.html Network documentation is distributed in the following manuals: TORNADO USER'S GUIDE Chapter 1: Overview. Description of Tornado IDE. Chapter 2: Setup and Startup. Host and target setup. Chapter 4: Projects. Project Facility Build Method. Chapter 5: Target Server. Tools/target interface. Chapter 6: Target Shell. Host-based shell: WindSh. Appendix F: FTP Server. Host system FTP server setup. TORNADO RELEASE NOTES Section 7: Host Development Environment Enhancements Section 9.1: Enhanced Network Stack Section 11: Backwards Compatibility Issues VXWORKS PROGRAMMER'S GUIDE Chapter 8: Configuration and Build. Tornado 101 Build Method. Chapter 9: Target Shell. VXWORKS NETWORK PROGRAMMER'S GUIDE Chapter 13: Booting over the Network. Chapter 14: Upgrading 4.3 BSD Network Drivers. Network drivers developers should get the Network Protocol ToolKit product from their sales representative. NPT is the most up to date manual on this subject. The BETA User's Guide is available from WindSurf which is an updated version of the SENS for Tornado manual. NPT is free and comes with some network libraries source. There is a patch also in source on WindSurf->Download->Patches for SPR# 30894. 5.2 TCP/IP REFERENCE BOOKS ---------------------- TCP/IP Illustrated Volume 1, by Richard Stevens - ----------------------------------------------- This book is an overview of TCP/IP protocols. TCP/IP Illustrated Volume 2, by Richard Stevens - ----------------------------------------------- The Tornado 2.0 network stack is based on BSD4.4 Lite. This book explains BSD4.4 code, function by function, line by line. UNIX Network Programming: Networking APIs: Sockets and XTI Volume 1, Second Edition, by Richard Stevens. - ---------------------------------------------------------- This book summarizes the TCP/IP Illustrated Volume 2 information and provides lots of client/server code examples which illustrate common programming errors. Effective TCP/IP Programming 44 Tips to Improve Your TCP/IP Programming by Jon C. Snader - ----------------------------------------------------------- High-Speed Networks TCP/IP and ATM Design Principles by William Stallings. - ---------------------------------------------------- This book discusses TCP/IP performance and memory resource issues. 5.3 WindSurf->Technical Solutions II -------------------------------- How does Tornado 2.0 (SENS) support END and BSD44 drivers at the same time How do END driver pass link layer information to the MUX Network Driver Porting Issues - --------------7BAB0E0070D9668839B285A3-- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado problem - compile on Unix, debug on Windows. Date: Wed, 27 Jun 2001 08:31:25 +0200 From: "Michael Lawnick" Organization: Buergernetz Dillingen Message-ID: <9hbuqo$eif$1@snoopy.bndlg.de> References: <94e55b6.0106150725.35205ecc@posting.google.com> <94e55b6.0106261334.2f3a53b6@posting.google.com> Reply-To: "Michael Lawnick" Sender: mlawnick@pd9e65895.dip.t-dialin.net Hi Matthew, ever tried Debug/Source Search Path ? - -- Mit freundlichen Grüßen, Michael Lawnick ============================================== SOFTEC GmbH Tel +49-731-96600-0 Promenade 17 Fax +49-731-96600-23 D-89073 Ulm Michael Lawnick Germany lawnick@softec.de ============================================== "Matthew Becker" schrieb im Newsbeitrag news:94e55b6.0106261334.2f3a53b6@posting.google.com... > I have still not received an adequate response to this message (yea or > nay). Therefore, I'm reposting it with some additional clarification > of the problem. > > When using the VxWorks cross-compiler (with debug flags turned on), > certain information is stored in the object file that is used for > debugging. Part of that information is the absolute paths to the > source files used to compile that object. > > The problem I am having is that I want to compile on Unix (Solaris), > but I want to debug on Windows. When the Tornado debugger on Windows > opens that object file, it can't resolve the Unix paths that were used > to compile the object. > > The Windows box I'm using has a mapped network drive to the Solaris > box. Therefore, all of the paths to the source files are only off by > the base. For example, if the file > "/home/jsmith/project/USBDriver.cc" existed on Solaris, then it could > be seen as "E:\project\USBDriver.cc" on Windows. I was wondering if > the Tornado debugger had some internal smarts such that it could take > a path-mapping (e.g. "/home/jsmith" -> "E:\"), and correctly map the > Unix paths in the object file to Windows paths (so that it could then > find the source files). > > This type of funtionality exists on other debugging solutions provided > by WindRiver (e.g. SingleStep and visionClick). Does this > functionality exist in Tornado II? Tornado 3/AE? Is it going to be > added to AE? > > Thanks in advance. > > -Matt > > matthew.becker@usa.xerox.com (Matthew Becker) wrote in message news:<94e55b6.0106150725.35205ecc@posting.google.com>... > > Hello all, > > > > The development group I'm in wants to use the VxWorks cross-compiler > > on Solaris to perform target builds, and Tornado on Windows PC's to > > perform debug. In the past, this strategy has worked because the > > IDE's they were using had some notion of associating a Unix path to a > > Windows driver letter (i.e. /home/jsmith -> E:\ ). > > > > Unfortunately, I have been unable to find this same functionality in > > Tornado. Currently the group has licences (and is using) Tornado II > > for Solaris and Windows. The Tornado IDE on Windows does provide a > > "Debugger Source Search Path" window, but it doesn't seem to be smart > > enough to recursively descend through a specified directory. In other > > words, all directories for a project must be specified individually > > (also, not what I'm looking for). > > > > Does anyone know of a solution to this problem for Tornado II? Is > > this issue resolved with Tornado 3/AE? > > > > Thanks in advance. > > > > -Matt Becker --------------------------- Newsgroups: comp.os.vxworks Subject: Re: loadModule + memDev = loaded symbols? Date: Wed, 27 Jun 2001 08:35:15 +0200 From: "Michael Lawnick" Organization: Buergernetz Dillingen Message-ID: <9hbv1n$ej1$1@snoopy.bndlg.de> References: Reply-To: "Michael Lawnick" Sender: mlawnick@pd9e65895.dip.t-dialin.net Hi Kelly, did you do the load from WindShell (i.e. Tornado Shell) ? As long as you haven't switched on 'target synchronisation' in your target server, the symbols will reside only on your target's symbol table. Try to find on TELNET or trget-shell (serial), should be visible. HTH - -- Mit freundlichen Grüßen, Michael Lawnick ============================================== SOFTEC GmbH Tel +49-731-96600-0 Promenade 17 Fax +49-731-96600-23 D-89073 Ulm Michael Lawnick Germany lawnick@softec.de ============================================== "Kelly Hornsby" schrieb im Newsbeitrag news:vfa_6.31836$w5.3844917@news1.rdc1.ga.home.com... > Hello All > > I'm trying to use a memDev pseudo-file (located in flash memory) called > /mem/flash to keep an ELF file that I flashed from my tornado project. It's > a very simple program that prints "hello world\n", and is called hello.o. > Very simple. I've verified the file is identical to the standalone .o file > in my Tornado project. > > I find that I can load this file from the shell using ld < hello.o, and it > will load properly and will execute. furthermore, the symbol will appear in > the symbol table. > > When I do the equivalent thing from my /mem/flash memDev file that I > created, the loadModule runs to completion, but nothing appears in the > symbol table. I pass the flag ALL_SYMBOLS to loadModule, but I still get no > symbols. The memDev is working properly because I can open the file and > read it programmatically. > > Has anyone done this before... load a file from memory or flash into the > system via a loadModule and a memDev? I'm looking for gotchas, faults or > success stories. > > Any ideas or thoughts are welcome. > > --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado problem - compile on Unix, debug on Windows. Date: 27 Jun 2001 07:22:48 GMT From: borkhuis@agere.com (Johan Borkhuis) Organization: Agere Systems Message-ID: References: <94e55b6.0106150725.35205ecc@posting.google.com> <94e55b6.0106261334.2f3a53b6@posting.google.com> matthew.becker@usa.xerox.com (Matthew Becker) wrote: > When using the VxWorks cross-compiler (with debug flags turned on), > certain information is stored in the object file that is used for > debugging. Part of that information is the absolute paths to the > source files used to compile that object. > > The problem I am having is that I want to compile on Unix (Solaris), > but I want to debug on Windows. When the Tornado debugger on Windows > opens that object file, it can't resolve the Unix paths that were used > to compile the object. Take a look at the following question in the FAQ: http://www.xs4all.nl/~borkhuis/vxworks/vxw_pt1.html#1.6-D It is about combined C and assembly files, but that was not working correctly because of the full pathnames. Try the proposed modifications on your Sun to see if this helps. 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: loadModule + memDev = loaded symbols? Date: 27 Jun 2001 07:25:37 GMT From: borkhuis@agere.com (Johan Borkhuis) Organization: Agere Systems Message-ID: References: "Kelly Hornsby" wrote: > When I do the equivalent thing from my /mem/flash memDev file that I > created, the loadModule runs to completion, but nothing appears in the > symbol table. I pass the flag ALL_SYMBOLS to loadModule, but I still > get no symbols. The memDev is working properly because I can open the > file and read it programmatically. Try @lkup "" instead of lkup (so add a '@' to the command). 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: Debug vars / system mBlk config Date: 27 Jun 2001 07:33:44 GMT From: borkhuis@agere.com (Johan Borkhuis) Organization: Agere Systems Message-ID: References: Alan Dennis wrote: > 1)Through WindSurf I "discovered" the ftpdDebug and ftpDebug vars that > are used for enabling debug output for ftp server and client > operations. I also know of scsiDebug which does similar for scsi. Is > there a list of other similar "undocumented" variables anywhere? > Anyone know of any others that could be useful (someday)? Try the following command: lkup "Debug" The variables in bss are a list of all debug variables. The meaning is unknown to me, but as far as I know they are bit-coded. So check the value of the variable, if it is 0 it could mean that it is a debugging help, because they are off by default. Then write -1 to it (so setting all bits) and wait for the data to apear on your terminal. 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: Additional code & alternative tool chain Date: 27 Jun 2001 09:23:16 UT From: fscheurer@manz-acs.de Organization: Lawrence Berkeley National Laboratory Message-ID: <9hcbnk$6tv$1@overload.lbl.gov> This message is in MIME format. Since your mail reader does not understand this format, some or all of this message may not be legible. - ------_=_NextPart_000_00014DA2.3B39C221 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 7Bit Hi VxWorkers, Additional code without TORNADO tool chain possible? we have a VxWorks based system (x86) developed by a partner where I want to add some cross-compiled object-modules. They wouldn't use VxWorks-code (no libs avail.), but only ANSI-C and the API of the partner's application code. Is it possible to load and run objects compiled e.g. with GNU gccx86 v2.95.x? Is there something special to pay attention to when cross-developing and providing standard-C based object modules, e.g. object format, options, segmentation, etc.? Would the loader accept object modules from an alternative tool-chain? MfG / Kind Regards Friedrich Scheurer, R&D - -------------------------------- acs GmbH Steigaeckerstrasse 13 72768 Reutlingen Telefon (07121) 9019-23 Telefax (07121) 9019-98 Internet www.manz-acs.de e-mail: fscheurer@manz-acs.de - -------------------------------- - ------_=_NextPart_000_00014DA2.3B39C221 Content-Type: text/html; charset=iso-8859-1 Content-Transfer-Encoding: quoted-printable
Hi VxWorkers,
 
Additional code without TORNADO tool chain possible?
 
we have a VxWorks based system (x86) developed by a partner
where I want to add some cross-compiled object-modules.
They wouldn't use VxWorks-code (no libs avail.), but only ANSI-C and
the API of the partner's application code.
 
Is it possible to load and run objects compiled e.g.
with GNU gccx86 v2.95.x?
 
Is there something special
to pay attention to when cross-developing
and providing standard-C based object modules,
e.g. object format, options, segmentation, etc.?
 
Would the loader accept object modules from
an alternative tool-chain?
 
 
MfG / Kind Regards
 
Friedrich Scheurer, R&D
--------------------------------
acs GmbH
Steigaeckerstrasse 13
72768 Reutlingen
Telefon  (07121) 9019-23
Telefax  (07121) 9019-98
Internet www.manz-acs.de
e-mail:  fscheurer@manz-acs.de
--------------------------------
- ------_=_NextPart_000_00014DA2.3B39C221-- --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Wed Jun 27 05:11:34 2001 From: fscheurer@manz-acs.de Date: Wed Jun 27 05:11:37 PDT 2001 Subject: Re: Threads in VxWorks Hi As I understand it: Due to the missing protection between tasks tasks can be interpreted as threads. In comparison to other OSes (UNIX,WinXX) VxWorks doesn't know processes. This is why you can see it the way mentioned above. Regards -------- Original Message -------- > Submitted-by vxwexplo-errs@csg.lbl.gov Tue Jun 26 11:21:41 2001 > Submitted-by: Anuradha Pani > > hi all, > > I have a question, how are threads implemented in > Vxworks? I found several routines for tasks in taskLib > library, but none for threads. > > Thanks > MfG / Kind Regards Friedrich Scheurer, R&D -------------------------------- acs GmbH Steigaeckerstrasse 13 72768 Reutlingen Telefon (07121) 9019-23 Telefax (07121) 9019-98 Internet www.manz-acs.de e-mail: fscheurer@manz-acs.de -------------------------------- From vxwexplo-errs@csg.lbl.gov Wed Jun 27 10:00:32 2001 From: Eric Martin Date: Wed Jun 27 10:00:34 PDT 2001 Subject: Disk Cache Tornado 2.0.2 - DosFs 2.0 - PPC 2603 We are using Fujitsu MO drives 230M 540M 640M and 1.3G drives. I want to be able to pull a disk out and replace it with any other and have the system handle it. Disk Initialization: I create a SCSI physical device structure (scsiPhysDevCreate), then define a logical partition on a SCSI block device (scsiBlkDevCreate), then I do a cbioDeviceVerify to create CBIO wrapper for BLK_DEV device which returns a CBIO device pointer which I use to create a disk cache (dcacheDevCreate). Then I do a dosFsDevCreate which takes the CBIO_DEV_ID device handle returned by dcacheDevCreate. And all this seems to work. NOW, when I sense that a disk has been removed. I remove the device from the I/O system device list like this if ( iosDevFind (base_dir[drive_num], &ptail)!=NULL) iosDevDelete (iosDevFind (base_dir[drive_num], &ptail)); and delete the SCSI physical-device structure scsiPhysDevDelete (pDisk[drive_num]); and set the related pointers to NULL. Then I go through the whole Disk Initialization process with the new disk and it seems to work fine. The problem is, every time I change a disk I loose a little more RAM (60K) because a new cache gets created for each new disk. How Do I get rid of the cache from the previous disk? I tried a couple ioctls, they didn't hurt anything but they didn't help either. cbioIoctl (pCbioDevId1,CBIO_RESET,1) cbioIoctl (pCbioDevId1,CBIO_CACHE_INVAL,0) Any suggestions? Eric emartin@zetec.com From vxwexplo-errs@csg.lbl.gov Wed Jun 27 11:08:24 2001 From: Sujeet Naik Date: Wed Jun 27 11:08:26 PDT 2001 Subject: help required on linker configuration on tornado Hi everyone, I have perhaps a strange requirement. I have a project with 4 files viz., f1.c, f2.c, f3.c and f4.c (all are C source files). When I make the dependencies and build, I compile all these four files and create an archive MYPROJ.a. This is later linked with other archives to create a final binary. My requirement is, I want to force the functions from f2.c to start from a fixed location, while the other files should be linked and located contiguously with other archive files. For this, I have my linker directive (.ld file) forcing f2.o to start from a fixed address. The problem is, when I am doing this, I am not supposed to have f2.o as part of MYPROJ.a. Otherwise, I will have multiple instances of all the functions in f2.c since the .ld file already has forced f2.o from a fixed location. Can someone help me on deciding how to eliminate f2.o from MYPROJ.a. I cant manually go and edit makefile of this project since it is dynamically created. Also, if this cant be done, is there any other way I can achieve my goal. Thanks in advance, Sujeet From vxwexplo-errs@csg.lbl.gov Wed Jun 27 11:14:59 2001 From: "Aravindan Jagannathan" Date: Wed Jun 27 11:15:01 PDT 2001 Subject: Question on ledOpen (vxWorks) Hello All, I'm using 'ledLib' to get input from the user (using ledOpen/ledRead). 'ledOpen' returns an Integer id (ledId). Since 'ledRead' is a blocking call, I would like to do a 'select' on the 'ledId'. But as 'ledId' is a huge number, it doesn't allow me to do that( the 'select' call just blocks ). Is there a way by which I can map the 'ledId' into a lower Integer so that I can do a 'select' on it? Any response is appreciated. Thanks & Regards, Aravind. jara_vind@hotmail.com _________________________________________________________________ Get your FREE download of MSN Explorer at http://explorer.msn.com From vxwexplo-errs@csg.lbl.gov Wed Jun 27 12:10:33 2001 From: Geoffrey Espin Date: Wed Jun 27 12:10:35 PDT 2001 Subject: Re: Question on ledOpen (vxWorks) Aravind, > I'm using 'ledLib' to get input from the user > (using ledOpen/ledRead). 'ledOpen' returns an Integer > id (ledId). > Since 'ledRead' is a blocking call, I would like > to do a 'select' on the 'ledId'. > But as 'ledId' is a huge number, it doesn't allow > me to do that( the 'select' call just blocks ). > Is there a way by which I can map the 'ledId' into > a lower Integer so that I can do a 'select' on it? ledId is a pointer to an internal structure where the first few fields are just the inFd and outFd from what ledOpen() was passed. So you could try select()'ing on inFd. But I would avoid using select() for multithreading your UI and whatever else you're doing. Create separate tasks. Geoff -- vxworks From vxwexplo-errs@csg.lbl.gov Thu Jun 28 01:59:24 2001 From: Nayan Kumar Garg Date: Thu Jun 28 01:59:26 PDT 2001 Subject: How to download vxWorks Image through serial port Hi Can anyone tell me how to download the vxWorks image through the serial port . Nayan From vxwexplo-errs@csg.lbl.gov Thu Jun 28 03:30:32 2001 From: "Kartha, Unnikrishan P. (ext)" Date: Thu Jun 28 03:30:34 PDT 2001 Subject: Re: help required on linker configuration on tornado Hi, There is no harm in editing the makefile generated. I assume that you are using Tornado2.0 There are two possibilities: You can create a makefile of your own (refer GNU Make user's guide provided by Windriver) and execute it from Windows shell after setting all the paths correctly. Easier would be to edit the makefile generated and make the changes you need (removing f2.c from the archive path) and then execute make using this make file. The problem here is that next time you build the project from Tornado2.0, the makefile again gets regenerated and your changes will not be there. When you have specific requirements, it is better to have an own makefile. You can create one for ur own referring to the makefile generated dynamically. Kartha. Submitted-by vxwexplo-errs@csg.lbl.gov Wed Jun 27 11:08:24 2001 Submitted-by: Sujeet Naik Hi everyone, I have perhaps a strange requirement. I have a project with 4 files viz., f1.c, f2.c, f3.c and f4.c (all are C source files). When I make the dependencies and build, I compile all these four files and create an archive MYPROJ.a. This is later linked with other archives to create a final binary. My requirement is, I want to force the functions from f2.c to start from a fixed location, while the other files should be linked and located contiguously with other archive files. For this, I have my linker directive (.ld file) forcing f2.o to start from a fixed address. The problem is, when I am doing this, I am not supposed to have f2.o as part of MYPROJ.a. Otherwise, I will have multiple instances of all the functions in f2.c since the .ld file already has forced f2.o from a fixed location. Can someone help me on deciding how to eliminate f2.o from MYPROJ.a. I cant manually go and edit makefile of this project since it is dynamically created. Also, if this cant be done, is there any other way I can achieve my goal. Thanks in advance, Sujeet ----------------------------------------------------------- This Mail has been checked for Viruses Attention: Encrypted Mails can NOT be checked ! *** Diese Mail wurde auf Viren ueberprueft Hinweis: Verschluesselte Mails koennen NICHT geprueft werden! ------------------------------------------------------------ From vxwexplo-errs@csg.lbl.gov Thu Jun 28 04:03:16 2001 From: Vxworks Exploder Date: Thu Jun 28 04:03:19 PDT 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Thu Jun 28 04:03:06 PDT 2001 Subject: Re: loadModule + memDev = loaded symbols? Subject: re: Threads in VxWorks Subject: Re: Tornado problem - compile on Unix, debug on Windows. Subject: Re: change sector size? HELP! Subject: ELF to BIN conversion in vxworks_rom image Subject: databases Subject: Re: change sector size? HELP! Subject: Re: TFFS limitations? Subject: any way to pass environment variables into WindShell? Subject: Re: TFFS limitations? Subject: Flash File System Alternatives Subject: Re: change sector size? HELP! Subject: Re: Change MAC address dynamically in PPC8260 Subject: Disk Cache Subject: Re: Additional code & alternative tool chain Subject: help required on linker configuration on tornado Subject: Question on ledOpen (vxWorks) Subject: Re: Question on ledOpen (vxWorks) Subject: usrNetworkInit Subject: Re: Flash File System Alternatives Subject: re: Question on ledOpen (vxWorks) Subject: Re: Disk Cache Subject: can't create bootable project in tornado2.0 IDE Subject: Re: Software Applications Architect - Telecom Subject: Re: Software Applications Architect - Telecom Subject: Re: Software Applications Architect - Telecom Subject: Re: Software Applications Architect - Telecom Subject: Re: Software Applications Architect - Telecom Subject: Re: Software Applications Architect - Telecom Subject: Re: Software Applications Architect - Telecom Subject: Re: Software Applications Architect - Telecom Subject: Re: Software Applications Architect - Telecom Subject: Re: usrNetworkInit Subject: Re: ELF to BIN conversion in vxworks_rom image Subject: Re: any way to pass environment variables into WindShell? Subject: Re: help required on linker configuration on tornado Subject: Location Monitor Function On MV2604 Subject: Demangling c++ template function Subject: How to download vxWorks Image through serial port Subject: Re: How to download vxWorks Image through serial port ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: loadModule + memDev = loaded symbols? Date: Wed, 27 Jun 2001 13:58:20 +0200 From: "Matthias Guelck" Organization: Heidelberg Digital Message-ID: <9hchou$j4b$1@news.hdpp.de> References: Hello Kelly, I am using the following commands in my code (simplified!): test.cpp: #include #include #include #include #include #include #include #include #include void testMain(void) { FILE *pFile; char *pcBuffer; struct stat SStat; FUNCPTR helloMain; SYM_TYPE symType; int iTaskID; pFile=fopen("/tgtsvr/hello.out","rb"); fstat(fileno(pFile),&SStat); pcBuffer=(char *)malloc(SStat.st_size); fread(pcBuffer,1,SStat.st_size,pFile); fclose(pFile); memDrv(); memDevCreate("/mem/diag",pcBuffer,SStat.st_size); ld(1,TRUE,"/mem/diag"); memDevDelete("/mem/diag"); free(pcBuffer); symFindByName(sysSymTbl,"helloMain__Fv",(char **)(void **)&helloMain,&symType); taskSpawn("tHello",140,0,8192,(FUNCPTR)helloMain,0,0,0,0,0,0,0,0,0,0); } hello.cpp: #include #include void helloMain(void) { printf("\nhello world \n") ; } Best regards Matthias "Kelly Hornsby" wrote in message news:vfa_6.31836$w5.3844917@news1.rdc1.ga.home.com... > Hello All > > I'm trying to use a memDev pseudo-file (located in flash memory) called > /mem/flash to keep an ELF file that I flashed from my tornado project. It's > a very simple program that prints "hello world\n", and is called hello.o. > Very simple. I've verified the file is identical to the standalone .o file > in my Tornado project. > > I find that I can load this file from the shell using ld < hello.o, and it > will load properly and will execute. furthermore, the symbol will appear in > the symbol table. > > When I do the equivalent thing from my /mem/flash memDev file that I > created, the loadModule runs to completion, but nothing appears in the > symbol table. I pass the flag ALL_SYMBOLS to loadModule, but I still get no > symbols. The memDev is working properly because I can open the file and > read it programmatically. > > Has anyone done this before... load a file from memory or flash into the > system via a loadModule and a memDev? I'm looking for gotchas, faults or > success stories. > > Any ideas or thoughts are welcome. > > --------------------------- Newsgroups: comp.os.vxworks Subject: re: Threads in VxWorks Date: 27 Jun 2001 12:11:32 UT From: fscheurer@manz-acs.de Organization: Lawrence Berkeley National Laboratory Message-ID: <9hciol$9c3$1@overload.lbl.gov> Hi As I understand it: Due to the missing protection between tasks tasks can be interpreted as threads. In comparison to other OSes (UNIX,WinXX) VxWorks doesn't know processes. This is why you can see it the way mentioned above. Regards - -------- Original Message -------- > Submitted-by vxwexplo-errs@csg.lbl.gov Tue Jun 26 11:21:41 2001 > Submitted-by: Anuradha Pani > > hi all, > > I have a question, how are threads implemented in > Vxworks? I found several routines for tasks in taskLib > library, but none for threads. > > Thanks > MfG / Kind Regards Friedrich Scheurer, R&D - -------------------------------- acs GmbH Steigaeckerstrasse 13 72768 Reutlingen Telefon (07121) 9019-23 Telefax (07121) 9019-98 Internet www.manz-acs.de e-mail: fscheurer@manz-acs.de - -------------------------------- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Tornado problem - compile on Unix, debug on Windows. Date: 27 Jun 2001 05:40:29 -0700 From: matthew.becker@usa.xerox.com (Matthew Becker) Organization: http://groups.google.com/ Message-ID: <94e55b6.0106270440.2e63aae1@posting.google.com> References: <94e55b6.0106150725.35205ecc@posting.google.com> <94e55b6.0106261334.2f3a53b6@posting.google.com> <9hbuqo$eif$1@snoopy.bndlg.de> Michael, The problem with Debug->Source Search Path is that it doesn't recurse. In other words, I can't specify the root directory to all of the source files. I have to specify each directory/sub-directory individually. When there are many files in a large directory structure this is very cumbersome. - -Matt "Michael Lawnick" wrote in message news:<9hbuqo$eif$1@snoopy.bndlg.de>... > Hi Matthew, > > ever tried Debug/Source Search Path ? > > -- > Mit freundlichen Grüßen, > Michael Lawnick > ============================================== > SOFTEC GmbH Tel +49-731-96600-0 > Promenade 17 Fax +49-731-96600-23 > D-89073 Ulm Michael Lawnick > Germany lawnick@softec.de > ============================================== > > "Matthew Becker" schrieb im Newsbeitrag > news:94e55b6.0106261334.2f3a53b6@posting.google.com... > > I have still not received an adequate response to this message (yea or > > nay). Therefore, I'm reposting it with some additional clarification > > of the problem. > > > > When using the VxWorks cross-compiler (with debug flags turned on), > > certain information is stored in the object file that is used for > > debugging. Part of that information is the absolute paths to the > > source files used to compile that object. > > > > The problem I am having is that I want to compile on Unix (Solaris), > > but I want to debug on Windows. When the Tornado debugger on Windows > > opens that object file, it can't resolve the Unix paths that were used > > to compile the object. > > > > The Windows box I'm using has a mapped network drive to the Solaris > > box. Therefore, all of the paths to the source files are only off by > > the base. For example, if the file > > "/home/jsmith/project/USBDriver.cc" existed on Solaris, then it could > > be seen as "E:\project\USBDriver.cc" on Windows. I was wondering if > > the Tornado debugger had some internal smarts such that it could take > > a path-mapping (e.g. "/home/jsmith" -> "E:\"), and correctly map the > > Unix paths in the object file to Windows paths (so that it could then > > find the source files). > > > > This type of funtionality exists on other debugging solutions provided > > by WindRiver (e.g. SingleStep and visionClick). Does this > > functionality exist in Tornado II? Tornado 3/AE? Is it going to be > > added to AE? > > > > Thanks in advance. > > > > -Matt > > > > matthew.becker@usa.xerox.com (Matthew Becker) wrote in message > news:<94e55b6.0106150725.35205ecc@posting.google.com>... > > > Hello all, > > > > > > The development group I'm in wants to use the VxWorks cross-compiler > > > on Solaris to perform target builds, and Tornado on Windows PC's to > > > perform debug. In the past, this strategy has worked because the > > > IDE's they were using had some notion of associating a Unix path to a > > > Windows driver letter (i.e. /home/jsmith -> E:\ ). > > > > > > Unfortunately, I have been unable to find this same functionality in > > > Tornado. Currently the group has licences (and is using) Tornado II > > > for Solaris and Windows. The Tornado IDE on Windows does provide a > > > "Debugger Source Search Path" window, but it doesn't seem to be smart > > > enough to recursively descend through a specified directory. In other > > > words, all directories for a project must be specified individually > > > (also, not what I'm looking for). > > > > > > Does anyone know of a solution to this problem for Tornado II? Is > > > this issue resolved with Tornado 3/AE? > > > > > > Thanks in advance. > > > > > > -Matt Becker --------------------------- Newsgroups: comp.os.vxworks Subject: Re: change sector size? HELP! Date: 27 Jun 2001 08:48:29 -0400 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: References: <52ccd904a%molochai@vapour-trail.demon.co.uk> Sender: bpringlemeir@DeadDuck >>>>> "George" == George writes: >>>>> "molochai" == writes... George> How do you change the sector size from 512 bytes to 2048? George> WindRiver has > not been much help :-( > molochai> Of what? George> Of a disk drive. Usually you don't set the sector size, you set the cluster size. With DosFs, I think the default cluster size is 1024 or 2048 bytes. I don't think there is an overhead to use a small cluster size with a solid state devices such as flash. You might gain a little disk space if you could format a drive such that it had larger sector sizes. If you really want to increase the sector size, then I would suggest doing this under DOS and then see if DosFS will read the drive. However, I have had trouble (and so have others) changing the cluster sizes. The odds that DosFs would handle larger sector sizes seems rather dubious to me. hth, Bill Pringlemeir. ps, you could be a little more specific on what it is you wish to do. Useful vxWorks URLS, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" "http://www.xs4all.nl/~borkhuis/vxworks/vxw_pt9.html" - -- Truth decays into beauty, while beauty soon becomes merely charm. Charm ends up as strangeness, and even that doesn't last, but up and down are forever. - Anonymous --------------------------- Newsgroups: comp.os.vxworks Subject: ELF to BIN conversion in vxworks_rom image Date: 27 Jun 2001 06:21:17 -0700 From: balajeep@hclt.com (Balajee Premraj) Organization: http://groups.google.com/ Message-ID: HI I have built the vxworks_rom image for the Motorola ADS860 BSP. When i tried to convert it into Binary file using the elfTobin utility provided by the Wind River it gives me an error "seg1: Expected load address 0x02800100 but file address is 0x00293100" I would like to know what mistake am i making. The same thing when done on the RAM based Image works fine and i am able to get the binary file. Thanks in advance Balajee P --------------------------- Newsgroups: comp.os.vxworks Subject: databases Date: Wed, 27 Jun 2001 10:01:03 -0400 From: "Kevin Livingston" Message-ID: Does anyone know of any good database products that run on VxWorks? preferably "free" or low cost would be good, although any suggestions you have are definately welcome. Any experience you have with the product would be nice to hear about too. thanks for your help Kevin --------------------------- Newsgroups: comp.os.vxworks Subject: Re: change sector size? HELP! Date: Wed, 27 Jun 2001 14:01:30 GMT From: "George" Organization: Excite@Home - The Leader in Broadband http://home.com/faster Message-ID: <_Gl_6.69653$%a.3391129@news1.rdc1.sdca.home.com> References: <52ccd904a%molochai@vapour-trail.demon.co.uk> Reply-To: "George" We have a number of Motorola 750 hosts hooked up to 300Gbyte RAIDs (through a Fibre Channel switch). These RAIDs are also accessed by Win32 PCs (through the switch). We need the highest possible throughput in and out of the RAIDs (looking for 40 to 80 Mbytes/second). Having a RAID (disk) sector size of 2048 bytes significantly increases the efficiency of the RAIDs allowing each drive of the RAID to have a sector size of 512. We're formatting the RAIDs with FAT32 so that both VxWorks and Win32 can access them. We have no trouble formatting the RAIDs with a sector size of 2048 (we've found it's easiest to use Linux for this), now if we can just get VxWorks to read them! Setting the sector size is such a trivial thing to do in Windows and Linux, so far this seems to be impossible in VxWorks. :-( "Bill Pringlemeir" wrote in message news:uhex22in6.fsf@yahoo.com... > >>>>> "George" == George writes: > >>>>> "molochai" == writes... > > George> How do you change the sector size from 512 bytes to 2048? > George> WindRiver has > not been much help :-( > > > molochai> Of what? > > George> Of a disk drive. > > Usually you don't set the sector size, you set the cluster size. With > DosFs, I think the default cluster size is 1024 or 2048 bytes. I > don't think there is an overhead to use a small cluster size with a > solid state devices such as flash. You might gain a little disk space > if you could format a drive such that it had larger sector sizes. > > If you really want to increase the sector size, then I would suggest > doing this under DOS and then see if DosFS will read the drive. > However, I have had trouble (and so have others) changing the cluster > sizes. The odds that DosFs would handle larger sector sizes seems > rather dubious to me. > > hth, > Bill Pringlemeir. > > ps, you could be a little more specific on what it is you wish to do. > > Useful vxWorks URLS, > "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" > "http://www.xs4all.nl/~borkhuis/vxworks/vxw_pt9.html" > > -- > Truth decays into beauty, while beauty soon becomes merely > charm. Charm ends up as strangeness, and even that doesn't last, but > up and down are forever. - Anonymous --------------------------- Newsgroups: comp.os.vxworks Subject: Re: TFFS limitations? Date: 27 Jun 2001 07:21:46 -0700 From: aparker@vtlinx.com (parreg) Organization: http://groups.google.com/ Message-ID: <79ef8b60.0106270621.33a72fc3@posting.google.com> References: <3B31BDA2.AEEA9B89@siemenscomms.co.uk> <3B3351AA.684FB899@digidata.com> Doug Dotson wrote in message news:<3B3351AA.684FB899@digidata.com>... > As far as I can tell, all you can do is specify an offset of where > TFFS-managed flash begins. This means that the boot image > must start low. I don;t think that you can treat a single device > as multiple smaller devices because of the way that erasing > happens. No read access of any portion of the flash can be > going on while any erasing is in progress. It least this is my > understanding. > > I can't even get the darn thing (TFFS) to try to identify my > flash device. It throws an exception from flIdentifyFlash prior > to attempting to call cfiscsIdentify. > > Hope this helps, > Doug > > Owain Phillips wrote: > > > Hi, > > > > I have an 8MB device and the architectural limitation that the boot > > image > > has to lie at 4MB. > > > > With TFFS can I reserve arbitary regions of the prom that are not to be > > used in filesystems? IE. Can I build a filesystem in that 4MB below the > > bootimage? > > > > Or can I treat the 8MB device as though it is several devices and hence > > use the lover 4MB as FSys? > > > > Or must I reserve 4MB+bootimageSize and have a filesystem > > above the bootimage? > > > > I look forward to your responses... > > > > Regards, > > Owain Phillips Hello, I beleive the answer to your question is no. TFFS once formatted will take up the rest of the flash, so if you have a boot image in between it will get wiped out. The safest way is to have the boot image prior to the file system. - -parreg --------------------------- Newsgroups: comp.os.vxworks Subject: any way to pass environment variables into WindShell? Date: Wed, 27 Jun 2001 08:50:42 -0600 From: "cswreg.cos.agilent.com" Organization: Agilent Technologies Message-ID: <993653367.882868@cswreg.cos.agilent.com> Hi gang, The title says it all - I have some environment settings in Windows that I need to use in commands/scripts in WindShell, which doesn't seem to recognize either the UNIX $VAR or Windows %VAR% syntax. Does anyone have any ideas how I can do this? Thanks! Rob --------------------------- Newsgroups: comp.os.vxworks Subject: Re: TFFS limitations? Date: 27 Jun 2001 08:09:48 -0700 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0106270709.302cb0d1@posting.google.com> References: <3B31BDA2.AEEA9B89@siemenscomms.co.uk> <3B3351AA.684FB899@digidata.com> Doug Dotson wrote in message news:<3B3351AA.684FB899@digidata.com>... > As far as I can tell, all you can do is specify an offset of where > TFFS-managed flash begins. This is indeed true; something to be careful of is the emulation of NVRAM that some boards (notably ARM ones) use that writes the bootline info to the last block of flash. You should be able to move the bootline block to a lower one in the flash area so that it is "contiguous" with the area reserved for the boot image - make sure it has its own block though as it erases the whole block when changes are made to the bootline. > I can't even get the darn thing (TFFS) to try to identify my > flash device. It throws an exception from flIdentifyFlash prior > to attempting to call cfiscsIdentify. There's a couple of things to be careful of with this part of the process: 1) Make sure that the flash is not marked read-only in sysPhysMemDesc (or the region tables for AE) - this will cause the MMU mappings to be set to read only, and the identify process needs to write to the flash address to get the device into command mode. 2) Make sure that you are specifying the actual start address of the flash device (even if you are reserving a block for something else). Obviously, make sure you get the offset right too if you have a boot image in the same flash ;-) HTH, John... > Hope this helps, > Doug > > Owain Phillips wrote: > > > Hi, > > > > I have an 8MB device and the architectural limitation that the boot > > image > > has to lie at 4MB. > > > > With TFFS can I reserve arbitary regions of the prom that are not to be > > used in filesystems? IE. Can I build a filesystem in that 4MB below the > > bootimage? > > > > Or can I treat the 8MB device as though it is several devices and hence > > use the lover 4MB as FSys? > > > > Or must I reserve 4MB+bootimageSize and have a filesystem > > above the bootimage? > > > > I look forward to your responses... > > > > Regards, > > Owain Phillips --------------------------- Newsgroups: comp.os.vxworks Subject: Flash File System Alternatives Date: Wed, 27 Jun 2001 17:32:38 +0200 From: "Ton C. Jaspers" Organization: XS4ALL Internet BV Message-ID: <9hcuav$1gn$1@news1.xs4all.nl> Sender: tcj@gateway.pijnenburg.nl Hi, Anyone know alternative flash file systems and their suppliers? Please don't tell me about M-systems TFFS, that is identical to the FFS supplied by Windriver. T --------------------------- Newsgroups: comp.os.vxworks Subject: Re: change sector size? HELP! Date: 27 Jun 2001 11:33:56 -0400 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: References: <52ccd904a%molochai@vapour-trail.demon.co.uk> <_Gl_6.69653$%a.3391129@news1.rdc1.sdca.home.com> Sender: bpringlemeir@DeadDuck George> We have a number of Motorola 750 hosts hooked up to 300Gbyte George> RAIDs (through a Fibre Channel switch). These RAIDs are also George> accessed by Win32 PCs (through the switch). We need the George> highest possible throughput in and out of the RAIDs (looking George> for 40 to 80 Mbytes/second). Having a RAID (disk) sector George> size of 2048 bytes significantly increases the efficiency of George> the RAIDs allowing each drive of the RAID to have a sector George> size of 512. We're formatting the RAIDs with FAT32 so that George> both VxWorks and Win32 can access them. We have no trouble George> formatting the RAIDs with a sector size of 2048 (we've found George> it's easiest to use Linux for this), now if we can just get George> VxWorks to read them! George> Setting the sector size is such a trivial thing to do in George> Windows and Linux, so far this seems to be impossible in George> VxWorks. :-( I suggest you examine the following files then, $(TORNADO)/target/src/drv/hdisk/*.c $(TORNADO)/target/src/config/usrIde.c $(TORNADO)/target/src/config/usrAta.c From the looks of the function in ataDrv.c and ideDrv.c, it is possible to configure the device so that it uses larger sector sizes. Surprise, Surprise; Sometime the source is provided with vxWorks! In theory you should be able to coerce vxWorks to use a larger sector size. The comments look like they might help a little. You might like to contact your FAE, sometimes they can help. Also, if you are really nice, you might post how you did this here. For DosFs you might try the document, "file:///C|/Tornado/docs/vxworks/guide/c-filesys2.html" There are entries for bytes/sector in the dosFsConfigShow() so I think that these maybe coded as variables. You will have to change your driver (which might be custom or ata/ide?) and DosFs may have to be configured to use a larger size. hth, Bill Pringlemeir. Useful vxWorks URLS, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" "http://www.xs4all.nl/~borkhuis/vxworks/vxw_pt9.html" - -- If Einstein's thoughts could die, would the Universe, U for Unaccompanied, be boring modern? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Change MAC address dynamically in PPC8260 Date: Wed, 27 Jun 2001 17:09:58 GMT From: Dan Gold Organization: http://www.newsranger.com Message-ID: References: <3B25ED5F.E7E492A4@optonline.net> <3B36AF36.260B1257@optonline.net> In article <3B36AF36.260B1257@optonline.net>, Gururaj patkar says... > >sorry for the belated reply to some of the questions posted. >Let me explain the detailed scenario - [most details deleted...] The manufacturing department sets >the eeprom mac address which is different from the well known mac address I want to >use, hence the need to change it dynamically at run-time to well-known mac address >(known to both master and child card). However, before any ethernet packet is sent >out of FCC3 on master card and child card I change it dynamically using >sysEnetAddrSet(). All I wanted to make sure this change I am making, with the above >design in mind does not have any caveats. Well the key statement (that was missing from the previous discussion) is that you want to do this before any ethernet packets are sent out, and that it won't happen *after* any packets have been sent out. Manufacturing burns the real mac address in, and you probably want to change the last byte to the slot in the backplane or something ? Then sysEnetAddrSet() is the correct place to do it. hth, Dan Gold gold@ensemble.com --------------------------- Newsgroups: comp.os.vxworks Subject: Disk Cache Date: Wed, 27 Jun 2001 09:54:34 -0700 From: Eric Martin Organization: Lawrence Berkeley National Laboratory Message-ID: <9hd4b0$2k0$1@overload.lbl.gov> Tornado 2.0.2 - DosFs 2.0 - PPC 2603 We are using Fujitsu MO drives 230M 540M 640M and 1.3G drives. I want to be able to pull a disk out and replace it with any other and have the system handle it. Disk Initialization: I create a SCSI physical device structure (scsiPhysDevCreate), then define a logical partition on a SCSI block device (scsiBlkDevCreate), then I do a cbioDeviceVerify to create CBIO wrapper for BLK_DEV device which returns a CBIO device pointer which I use to create a disk cache (dcacheDevCreate). Then I do a dosFsDevCreate which takes the CBIO_DEV_ID device handle returned by dcacheDevCreate. And all this seems to work. NOW, when I sense that a disk has been removed. I remove the device from the I/O system device list like this if ( iosDevFind (base_dir[drive_num], &ptail)!=NULL) iosDevDelete (iosDevFind (base_dir[drive_num], &ptail)); and delete the SCSI physical-device structure scsiPhysDevDelete (pDisk[drive_num]); and set the related pointers to NULL. Then I go through the whole Disk Initialization process with the new disk and it seems to work fine. The problem is, every time I change a disk I loose a little more RAM (60K) because a new cache gets created for each new disk. How Do I get rid of the cache from the previous disk? I tried a couple ioctls, they didn't hurt anything but they didn't help either. cbioIoctl (pCbioDevId1,CBIO_RESET,1) cbioIoctl (pCbioDevId1,CBIO_CACHE_INVAL,0) Any suggestions? Eric emartin@zetec.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Additional code & alternative tool chain Date: Wed, 27 Jun 2001 11:46:06 -0500 From: "tazboy78" Organization: http://extra.newsguy.com Message-ID: <9hd2im020ti@enews2.newsguy.com> References: <9hcbnk$6tv$1@overload.lbl.gov> > Is it possible to load and run objects compiled e.g. > with GNU gccx86 v2.95.x? Yes, but the objects will be built for dynamic linking at download. If you can handle this you should be okay. --------------------------- Newsgroups: comp.os.vxworks Subject: help required on linker configuration on tornado Date: Wed, 27 Jun 2001 23:42:28 +0530 From: Sujeet Naik Organization: Lawrence Berkeley National Laboratory Message-ID: <9hd7rm$4hs$1@overload.lbl.gov> Hi everyone, I have perhaps a strange requirement. I have a project with 4 files viz., f1.c, f2.c, f3.c and f4.c (all are C source files). When I make the dependencies and build, I compile all these four files and create an archive MYPROJ.a. This is later linked with other archives to create a final binary. My requirement is, I want to force the functions from f2.c to start from a fixed location, while the other files should be linked and located contiguously with other archive files. For this, I have my linker directive (.ld file) forcing f2.o to start from a fixed address. The problem is, when I am doing this, I am not supposed to have f2.o as part of MYPROJ.a. Otherwise, I will have multiple instances of all the functions in f2.c since the .ld file already has forced f2.o from a fixed location. Can someone help me on deciding how to eliminate f2.o from MYPROJ.a. I cant manually go and edit makefile of this project since it is dynamically created. Also, if this cant be done, is there any other way I can achieve my goal. Thanks in advance, Sujeet --------------------------- Newsgroups: comp.os.vxworks Subject: Question on ledOpen (vxWorks) Date: Wed, 27 Jun 2001 14:14:52 -0400 From: "Aravindan Jagannathan" Organization: Lawrence Berkeley National Laboratory Message-ID: <9hd7t5$4ht$1@overload.lbl.gov> Hello All, I'm using 'ledLib' to get input from the user (using ledOpen/ledRead). 'ledOpen' returns an Integer id (ledId). Since 'ledRead' is a blocking call, I would like to do a 'select' on the 'ledId'. But as 'ledId' is a huge number, it doesn't allow me to do that( the 'select' call just blocks ). Is there a way by which I can map the 'ledId' into a lower Integer so that I can do a 'select' on it? Any response is appreciated. Thanks & Regards, Aravind. jara_vind@hotmail.com _________________________________________________________________ Get your FREE download of MSN Explorer at http://explorer.msn.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Question on ledOpen (vxWorks) Date: Wed, 27 Jun 2001 14:48:45 -0400 From: Doug Dotson Organization: Verio Message-ID: <3B3A2A8D.8BA81912@digidata.com> References: <9hd7t5$4ht$1@overload.lbl.gov> Looks to me like you want the select to use the inFd or the outFd or both depending upon what you are trying to do. It doesn't look like the led itself is associated with a unique fd. Doug Aravindan Jagannathan wrote: > Hello All, > > I'm using 'ledLib' to get input from the user > (using ledOpen/ledRead). 'ledOpen' returns an Integer > id (ledId). > > Since 'ledRead' is a blocking call, I would like > to do a 'select' on the 'ledId'. > > But as 'ledId' is a huge number, it doesn't allow > me to do that( the 'select' call just blocks ). > > Is there a way by which I can map the 'ledId' into > a lower Integer so that I can do a 'select' on it? > > Any response is appreciated. > > Thanks & Regards, > Aravind. > jara_vind@hotmail.com > _________________________________________________________________ > Get your FREE download of MSN Explorer at http://explorer.msn.com --------------------------- Newsgroups: comp.os.vxworks Subject: usrNetworkInit Date: Wed, 27 Jun 2001 14:57:08 -0400 From: Jeremiah Frick Organization: Improv Systems, Inc. Message-ID: <3B3A2C84.1410D611@improvsys.com> This is a multi-part message in MIME format. - --------------28F1F69FB6015FB093FCD0F6 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hello VxWorkers :) Quick question. How do I determine what the string you use for usrNetworkInit should be? Is this even required? Right now I have two machines that will talk to each other no problem, but I think it might be that they're using DHCP and getting the correct IPs. Is there a function call to set the IP address for your vxWorks system? Thanks! Jeremy - --------------28F1F69FB6015FB093FCD0F6 Content-Type: text/x-vcard; charset=us-ascii; name="jeremyf.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Jeremiah Frick Content-Disposition: attachment; filename="jeremyf.vcf" begin:vcard n:Frick;Jeremiah tel;work:716-756-5411 x45 x-mozilla-html:FALSE url:www.improvsys.com org:Improv Systems Inc.;PDG version:2.1 email;internet:jeremyf@improvsys.com title:Hardware Engineer adr;quoted-printable:;;3136 Winton Road South=0D=0ASuite 100;Rochester;New York;14623;United States of America fn:Jeremiah Frick end:vcard - --------------28F1F69FB6015FB093FCD0F6-- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Flash File System Alternatives Date: Wed, 27 Jun 2001 20:06:26 +0100 From: molochai Message-ID: References: <9hcuav$1gn$1@news1.xs4all.nl> In message <9hcuav$1gn$1@news1.xs4all.nl> "Ton C. Jaspers" wrote: > Hi, > > Anyone know alternative flash file systems and their suppliers? > Please don't tell me about M-systems TFFS, that is identical to the FFS > supplied by Windriver. > DIY depending on the subset of features of TFFS you want and how much time you have, how many products it will be deployed on etc. --------------------------- Newsgroups: comp.os.vxworks Subject: re: Question on ledOpen (vxWorks) Date: Wed, 27 Jun 2001 12:10:29 -0700 From: Geoffrey Espin Organization: Lawrence Berkeley National Laboratory Message-ID: <9hdbbv$5q9$1@overload.lbl.gov> Aravind, > I'm using 'ledLib' to get input from the user > (using ledOpen/ledRead). 'ledOpen' returns an Integer > id (ledId). > Since 'ledRead' is a blocking call, I would like > to do a 'select' on the 'ledId'. > But as 'ledId' is a huge number, it doesn't allow > me to do that( the 'select' call just blocks ). > Is there a way by which I can map the 'ledId' into > a lower Integer so that I can do a 'select' on it? ledId is a pointer to an internal structure where the first few fields are just the inFd and outFd from what ledOpen() was passed. So you could try select()'ing on inFd. But I would avoid using select() for multithreading your UI and whatever else you're doing. Create separate tasks. Geoff - -- vxworks --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Disk Cache Date: Wed, 27 Jun 2001 20:03:45 GMT From: andreyk@home.com (Andray Kaganovsky) Organization: @Home Message-ID: References: <9hd4b0$2k0$1@overload.lbl.gov> Eric Martin wrote in news:9hd4b0$2k0$1@overload.lbl.gov: > > > Tornado 2.0.2 - DosFs 2.0 - PPC 2603 > > We are using Fujitsu MO drives 230M 540M 640M and 1.3G drives. I want to be > able to pull a disk out and replace it with any other and have the system > handle it. > > Disk Initialization: > I create a SCSI physical device structure (scsiPhysDevCreate), then define a > logical partition on a SCSI block device (scsiBlkDevCreate), then I do a > cbioDeviceVerify to create CBIO wrapper for BLK_DEV device which returns a > CBIO device pointer which I use to create a disk cache (dcacheDevCreate). > Then I do a dosFsDevCreate which takes the CBIO_DEV_ID device handle > returned by dcacheDevCreate. And all this seems to work. > > NOW, when I sense that a disk has been removed. I remove the device from the > I/O system device list like this > > if ( iosDevFind (base_dir[drive_num], &ptail)!=NULL) > iosDevDelete (iosDevFind (base_dir[drive_num], &ptail)); > > and delete the SCSI physical-device structure > > scsiPhysDevDelete (pDisk[drive_num]); > > and set the related pointers to NULL. Then I go through the whole Disk > Initialization process with the new disk and it seems to work fine. > > The problem is, every time I change a disk I loose a little more RAM (60K) > because a new cache gets created for each new disk. > How Do I get rid of the cache from the previous disk? I tried a couple > ioctls, they didn't hurt anything but they didn't help either. > > > cbioIoctl (pCbioDevId1,CBIO_RESET,1) > > cbioIoctl (pCbioDevId1,CBIO_CACHE_INVAL,0) > > Any suggestions? > > Eric > emartin@zetec.com > [AndrayK]. The second arguement to dcacheDevCreate() is the pointer to memory region where disk cache is to be placed (or NULL for malloc()). Did you try to pre-allocate this memory region, and pass it's address to all subsequent dcacheDevCreate() calls ? Andray --------------------------- Newsgroups: comp.os.vxworks Subject: can't create bootable project in tornado2.0 IDE Date: Wed, 27 Jun 2001 18:47:09 -0700 From: "Vic Zhang" Organization: The news server of Changzhou TeleCom Message-ID: <9hcdjv$2rpa$1@news.cz.js.cn> when i create bootable project in tornado2.0 IDE on windows2k, I meet the following error message, couldn't duplicate input handle: bad file number "make makeMacros BSP2PRJ=TRUE" in C:/Tornado/target/config/ziva5mk failed does anyone ever met and can solve this problem? thanks a lot vic --------------------------- Newsgroups: cern.c++,comp.os.psos,comp.os.vxworks,dfw.jobs,misc.taxes Subject: Re: Software Applications Architect - Telecom Date: Wed, 27 Jun 2001 16:39:03 -0700 From: "Fauzi Duran" Organization: SoftEdge Systems, LLC Message-ID: Reply-To: "Fauzi Duran" "Adnan Haq" wrote in message news:... > My client is a very hot Telecom company in Dallas and looking for the > following: > > Job Description: Responsible for designing the application software that > manages the policies, SLAs, routing , and external databases, as well as the > software platform. > > Requirements: BSCS or BSEE with ten years of experience > Telecom or datacom background required > Knowledge of software subsystems integration and communication > Experience in building complete systems > Self-motivation and enthusiasm a plus > > > Contact: > > Fauzi Duran > Technical Recruiter > SoftEdge Systems, LLC > 972-241-3211 X13 > duran@softedgesystems.com > > > > --------------------------- Newsgroups: cern.c++,comp.os.psos,comp.os.vxworks,dfw.jobs,misc.taxes Subject: Re: Software Applications Architect - Telecom Date: Wed, 27 Jun 2001 16:42:38 -0700 From: "Fauzi Duran" Organization: SoftEdge Systems, LLC Message-ID: Reply-To: "Fauzi Duran" "Fauzi Duran" wrote in message news:... > > "Adnan Haq" wrote in message news:... > > My client is a very hot Telecom company in Dallas and looking for the > > following: > > > > Job Description: Responsible for designing the application software that > > manages the policies, SLAs, routing , and external databases, as well as > the > > software platform. > > > > Requirements: BSCS or BSEE with ten years of experience > > Telecom or datacom background required > > Knowledge of software subsystems integration and communication > > Experience in building complete systems > > Self-motivation and enthusiasm a plus > > > > > > Contact: > > > > Fauzi Duran > > Technical Recruiter > > SoftEdge Systems, LLC > > 972-241-3211 X13 > > duran@softedgesystems.com > > > > > > > > > > --------------------------- Newsgroups: cern.c++,comp.os.psos,comp.os.vxworks,dfw.jobs,misc.taxes Subject: Re: Software Applications Architect - Telecom Date: Wed, 27 Jun 2001 16:44:43 -0700 From: "Fauzi Duran" Organization: SoftEdge Systems, LLC Message-ID: <2D4680A0F2289649.EFB5B247C981697E.342C80F478FA8BEA@lp.airnews.net> Reply-To: "Fauzi Duran" "Fauzi Duran" wrote in message news:... > > "Fauzi Duran" wrote in message news:... > > > > "Adnan Haq" wrote in message news:... > > > My client is a very hot Telecom company in Dallas and looking for the > > > following: > > > > > > Job Description: Responsible for designing the application software that > > > manages the policies, SLAs, routing , and external databases, as well as > > the > > > software platform. > > > > > > Requirements: BSCS or BSEE with ten years of experience > > > Telecom or datacom background required > > > Knowledge of software subsystems integration and communication > > > Experience in building complete systems > > > Self-motivation and enthusiasm a plus > > > > > > > > > Contact: > > > > > > Fauzi Duran > > > Technical Recruiter > > > SoftEdge Systems, LLC > > > 972-241-3211 X13 > > > duran@softedgesystems.com > > > > > > > > > > > > > > > > > > --------------------------- Newsgroups: cern.c++,comp.os.psos,comp.os.vxworks,dfw.jobs,misc.taxes Subject: Re: Software Applications Architect - Telecom Date: Wed, 27 Jun 2001 16:46:44 -0700 From: "Fauzi Duran" Organization: SoftEdge Systems, LLC Message-ID: <8718833C339AA940.E16A0A27B071C633.6FCBEF72FC416A18@lp.airnews.net> Reply-To: "Fauzi Duran" "Fauzi Duran" wrote in message news:... > > "Fauzi Duran" wrote in message news:... > > > > "Fauzi Duran" wrote in message news:... > > > > > > "Adnan Haq" wrote in message news:... > > > > My client is a very hot Telecom company in Dallas and looking for the > > > > following: > > > > > > > > Job Description: Responsible for designing the application software > that > > > > manages the policies, SLAs, routing , and external databases, as well > as > > > the > > > > software platform. > > > > > > > > Requirements: BSCS or BSEE with ten years of experience > > > > Telecom or datacom background required > > > > Knowledge of software subsystems integration and communication > > > > Experience in building complete systems > > > > Self-motivation and enthusiasm a plus > > > > > > > > > > > > Contact: > > > > > > > > Fauzi Duran > > > > Technical Recruiter > > > > SoftEdge Systems, LLC > > > > 972-241-3211 X13 > > > > duran@softedgesystems.com > > > > > > > > > > > > > > > > > > > > > > > > > > > > --------------------------- Newsgroups: cern.c++,comp.os.psos,comp.os.vxworks,dfw.jobs,misc.taxes Subject: Re: Software Applications Architect - Telecom Date: Wed, 27 Jun 2001 16:48:34 -0700 From: "Fauzi Duran" Organization: SoftEdge Systems, LLC Message-ID: <58F59D1EB3FC094E.9EFC9F5B33B89183.0A3FDCEB9470D571@lp.airnews.net> Reply-To: "Fauzi Duran" "Fauzi Duran" wrote in message news:... > > "Fauzi Duran" wrote in message news:... > > > > "Fauzi Duran" wrote in message news:... > > > > > > "Fauzi Duran" wrote in message news:... > > > > > > > > "Adnan Haq" wrote in message news:... > > > > > My client is a very hot Telecom company in Dallas and looking for > the > > > > > following: > > > > > > > > > > Job Description: Responsible for designing the application software > > that > > > > > manages the policies, SLAs, routing , and external databases, as > well > > as > > > > the > > > > > software platform. > > > > > > > > > > Requirements: BSCS or BSEE with ten years of experience > > > > > Telecom or datacom background required > > > > > Knowledge of software subsystems integration and communication > > > > > Experience in building complete systems > > > > > Self-motivation and enthusiasm a plus > > > > > > > > > > > > > > > Contact: > > > > > > > > > > Fauzi Duran > > > > > Technical Recruiter > > > > > SoftEdge Systems, LLC > > > > > 972-241-3211 X13 > > > > > duran@softedgesystems.com > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > --------------------------- Newsgroups: cern.c++,comp.os.psos,comp.os.vxworks,dfw.jobs,misc.taxes Subject: Re: Software Applications Architect - Telecom Date: Wed, 27 Jun 2001 16:51:11 -0700 From: "Fauzi Duran" Organization: SoftEdge Systems, LLC Message-ID: Reply-To: "Fauzi Duran" "Fauzi Duran" wrote in message news:... > > "Fauzi Duran" wrote in message news:... > > > > "Fauzi Duran" wrote in message news:... > > > > > > "Fauzi Duran" wrote in message news:... > > > > > > > > "Fauzi Duran" wrote in message news:... > > > > > > > > > > "Adnan Haq" wrote in message news:... > > > > > > My client is a very hot Telecom company in Dallas and looking for > > the > > > > > > following: > > > > > > > > > > > > Job Description: Responsible for designing the application > software > > > that > > > > > > manages the policies, SLAs, routing , and external databases, as > > well > > > as > > > > > the > > > > > > software platform. > > > > > > > > > > > > Requirements: BSCS or BSEE with ten years of experience > > > > > > Telecom or datacom background required > > > > > > Knowledge of software subsystems integration and communication > > > > > > Experience in building complete systems > > > > > > Self-motivation and enthusiasm a plus > > > > > > > > > > > > > > > > > > Contact: > > > > > > > > > > > > Fauzi Duran > > > > > > Technical Recruiter > > > > > > SoftEdge Systems, LLC > > > > > > 972-241-3211 X13 > > > > > > duran@softedgesystems.com > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > --------------------------- Newsgroups: cern.c++,comp.os.psos,comp.os.vxworks,dfw.jobs,misc.taxes Subject: Re: Software Applications Architect - Telecom Date: Wed, 27 Jun 2001 16:53:06 -0700 From: "Fauzi Duran" Organization: SoftEdge Systems, LLC Message-ID: <778DE477FB01C5CC.7940D3DF1EE48B23.A270E28876FA4EC3@lp.airnews.net> Reply-To: "Fauzi Duran" "Fauzi Duran" wrote in message news:... > > "Fauzi Duran" wrote in message news:... > > > > "Fauzi Duran" wrote in message news:... > > > > > > "Fauzi Duran" wrote in message news:... > > > > > > > > "Fauzi Duran" wrote in message news:... > > > > > > > > > > "Fauzi Duran" wrote in message news:... > > > > > > > > > > > > "Adnan Haq" wrote in message news:... > > > > > > > My client is a very hot Telecom company in Dallas and looking > for > > > the > > > > > > > following: > > > > > > > > > > > > > > Job Description: Responsible for designing the application > > software > > > > that > > > > > > > manages the policies, SLAs, routing , and external databases, as > > > well > > > > as > > > > > > the > > > > > > > software platform. > > > > > > > > > > > > > > Requirements: BSCS or BSEE with ten years of experience > > > > > > > Telecom or datacom background required > > > > > > > Knowledge of software subsystems integration and communication > > > > > > > Experience in building complete systems > > > > > > > Self-motivation and enthusiasm a plus > > > > > > > > > > > > > > > > > > > > > Contact: > > > > > > > > > > > > > > Fauzi Duran > > > > > > > Technical Recruiter > > > > > > > SoftEdge Systems, LLC > > > > > > > 972-241-3211 X13 > > > > > > > duran@softedgesystems.com > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > --------------------------- Newsgroups: cern.c++,comp.os.psos,comp.os.vxworks,dfw.jobs,misc.taxes Subject: Re: Software Applications Architect - Telecom Date: Wed, 27 Jun 2001 17:00:02 -0700 From: "Fauzi Duran" Organization: SoftEdge Systems, LLC Message-ID: <2B3CB2C90C612396.927009637E7C8CE3.2C6A2D4807CF633A@lp.airnews.net> Reply-To: "Fauzi Duran" "Fauzi Duran" wrote in message news:... > > "Fauzi Duran" wrote in message news:... > > > > "Fauzi Duran" wrote in message news:... > > > > > > "Fauzi Duran" wrote in message news:... > > > > > > > > "Fauzi Duran" wrote in message news:... > > > > > > > > > > "Fauzi Duran" wrote in message news:... > > > > > > > > > > > > "Fauzi Duran" wrote in message > news:... > > > > > > > > > > > > > > "Adnan Haq" wrote in message > news:... > > > > > > > > My client is a very hot Telecom company in Dallas and looking > > for > > > > the > > > > > > > > following: > > > > > > > > > > > > > > > > Job Description: Responsible for designing the application > > > software > > > > > that > > > > > > > > manages the policies, SLAs, routing , and external databases, > as > > > > well > > > > > as > > > > > > > the > > > > > > > > software platform. > > > > > > > > > > > > > > > > Requirements: BSCS or BSEE with ten years of experience > > > > > > > > Telecom or datacom background required > > > > > > > > Knowledge of software subsystems integration and > communication > > > > > > > > Experience in building complete systems > > > > > > > > Self-motivation and enthusiasm a plus > > > > > > > > > > > > > > > > > > > > > > > > Contact: > > > > > > > > > > > > > > > > Fauzi Duran > > > > > > > > Technical Recruiter > > > > > > > > SoftEdge Systems, LLC > > > > > > > > 972-241-3211 X13 > > > > > > > > duran@softedgesystems.com > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > --------------------------- Newsgroups: cern.c++,comp.os.psos,comp.os.vxworks,dfw.jobs,misc.taxes Subject: Re: Software Applications Architect - Telecom Date: Wed, 27 Jun 2001 23:47:02 +0100 From: molochai Message-ID: <1cda67914a%molochai@vapour-trail.demon.co.uk> References: <2B3CB2C90C612396.927009637E7C8CE3.2C6A2D4807CF633A@lp.airnews.net> In message <2B3CB2C90C612396.927009637E7C8CE3.2C6A2D4807CF633A@lp.airnews.net> WTF? > "Fauzi Duran" wrote: > > "Fauzi Duran" wrote in message news:... > > > > "Fauzi Duran" wrote in message news:... > > > > > > "Fauzi Duran" wrote in message news:... > > > > > > > > "Fauzi Duran" wrote in message news:... > > > > > > > > > > "Fauzi Duran" wrote in message news:... > > > > > > > > > > > > "Fauzi Duran" wrote in message > news:... > > > > > > > > > > > > > > "Fauzi Duran" wrote in message > > news:... > > > > > > > > > > > > > > > > "Adnan Haq" wrote in message > > news:... [snip] --------------------------- Newsgroups: comp.os.vxworks Subject: Re: usrNetworkInit Date: Thu, 28 Jun 2001 01:10:18 GMT From: Dan Gold Organization: http://www.newsranger.com Message-ID: <_tv_6.1499$f4.20656@www.newsranger.com> References: <3B3A2C84.1410D611@improvsys.com> In article <3B3A2C84.1410D611@improvsys.com>, Jeremiah Frick says... > >This is a multi-part message in MIME format. >--------------28F1F69FB6015FB093FCD0F6 >Content-Type: text/plain; charset=us-ascii >Content-Transfer-Encoding: 7bit > >Quick question. How do I determine what the string you use for >usrNetworkInit should be? You mean usrNetInit() ? If you look in usrConfig.c, you'll see that it passes the "boot line" address. The boot line contains all the information needed by the system to boot, including cpu, filesystem and image, IP address, and other network parameters. It is possible that the board has used DHCP to set the various network parameters in the boot line, and thus when usrNetInit() is called, it has all it needs to bring up an interface correctly. The bootline is typically stored in non-volatile ram, obtained via sysNvRamGet(), and put into RAM as a string for everyone to use. >Is this even required? Right now I have two machines that will talk to >each other no problem, but I think it might be that they're using DHCP >and getting the correct IPs. Is there a function call to set the IP >address for your vxWorks system? If you have to machines that are currently using DHCP to get an IP address, chances are it's setting your boot line with the right stuff. Usually the IP address is stored in the bootline, and you get it by "cracking" the bootline, with usrBootlineCrack(). Look carefully at usrNetInit() and watch it use params.ead, the string containing the IP address for your interface. hth, Dan Gold gold@ensemble.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: ELF to BIN conversion in vxworks_rom image Date: Wed, 27 Jun 2001 19:09:02 -0700 From: DrDiags Organization: laserlink.net Message-ID: <3B3A91BE.F479CBB5@covad.net> References: Reply-To: drdiags@covad.net Balajee, You may need to contact your Wind River Support folks. I understand there was a problem with the elfToBin utility producing this type of error for T2 when building ROM images. You should also try WindSurf and search on the error string (replacing the addresses with a wild card). There may be a patch located there also. Balajee Premraj wrote: > > HI > > I have built the vxworks_rom image for the Motorola ADS860 BSP. > When i tried to convert it into Binary file using the elfTobin utility > provided by the Wind River it gives me an error > > "seg1: Expected load address 0x02800100 but file address is > 0x00293100" > > I would like to know what mistake am i making. > The same thing when done on the RAM based Image works fine and i am > able to get the binary file. > > Thanks in advance > Balajee P --------------------------- Newsgroups: comp.os.vxworks Subject: Re: any way to pass environment variables into WindShell? Date: Wed, 27 Jun 2001 19:11:01 -0700 From: DrDiags Organization: laserlink.net Message-ID: <3B3A9235.BD1C3690@covad.net> References: <993653367.882868@cswreg.cos.agilent.com> Reply-To: drdiags@covad.net What about the global "env" variable in Tcl? Does $env(VAR) display what you want? "cswreg.cos.agilent.com" wrote: > > Hi gang, > > The title says it all - I have some environment settings in Windows that I > need to use in commands/scripts in WindShell, which doesn't seem to > recognize either the UNIX $VAR or Windows %VAR% syntax. Does anyone have any > ideas how I can do this? > > Thanks! > > Rob --------------------------- Newsgroups: comp.os.vxworks Subject: Re: help required on linker configuration on tornado Date: Thu, 28 Jun 2001 08:57:09 +0100 From: Vijay Kumar Peshkar Organization: Cisco Systems Inc. Message-ID: <3B3AE355.96503767@wipro.com> References: <9hd7rm$4hs$1@overload.lbl.gov> This is a multi-part message in MIME format. - --------------2E2AD1608C1027964C580D4D Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sujeet, Did you look at MACH_EXTRA option?. Guess it should help. regards, Vijay Sujeet Naik wrote: > Hi everyone, > I have perhaps a strange requirement. I have a project with 4 files viz., > f1.c, f2.c, f3.c and f4.c (all are C source files). When I make the > dependencies and build, I compile all these four files and create an archive > MYPROJ.a. This is later linked with other archives to create a final binary. > > My requirement is, I want to force the functions from f2.c to start from a > fixed location, while the other files should be linked and located > contiguously with other archive files. For this, I have my linker directive > (.ld file) forcing f2.o to start from a fixed address. The problem is, when > I am doing this, I am not supposed to have f2.o as part of MYPROJ.a. > Otherwise, I will have multiple instances of all the functions in f2.c since > the .ld file already has forced f2.o from a fixed location. > > Can someone help me on deciding how to eliminate f2.o from MYPROJ.a. I cant > manually go and edit makefile of this project since it is dynamically > created. > > Also, if this cant be done, is there any other way I can achieve my goal. > > Thanks in advance, > > Sujeet - --------------2E2AD1608C1027964C580D4D Content-Type: text/x-vcard; charset=us-ascii; name="vijay.peshkar.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Vijay Kumar Peshkar Content-Disposition: attachment; filename="vijay.peshkar.vcf" begin:vcard n:Peshkar;Vijay Kumar tel;work:5530034/35/35 x2237 x-mozilla-html:FALSE org:Wipro Global R & D;Wishing you blue skies, a couple of daisies, some fresh air and a smile! adr:;;;;;; version:2.1 email;internet:vpeshkar@cisco.com note:Wishing you blue skies, a couple of daisies, some fresh air and a smile! fn:Vijay Kumar Peshkar end:vcard - --------------2E2AD1608C1027964C580D4D-- --------------------------- Newsgroups: comp.os.vxworks Subject: Location Monitor Function On MV2604 Date: Thu, 28 Jun 2001 05:15:20 GMT From: "Jon Lewis" Organization: Excite@Home - The Leader in Broadband http://home.com/faster Message-ID: Does anyone have a code example of how to set one of these up on a MV2604? I have been struggling to get the location monitor function to work. If a code example isn't available could someone be so kind as to outline the steps? I figured out how to get mailboxes to work but I am lost on how to get the location monitor function to work. I know in principle it should generate an interupt when a particular VME address is read. How is all this configured? ( How do you specify which address space? Does the register have to be mapped to that address space? ) Thanks in advance. --------------------------- Newsgroups: comp.os.vxworks Subject: Demangling c++ template function Date: Thu, 28 Jun 2001 10:21:18 +0200 From: "Albert Wijnja" Message-ID: <9hepd6$dfbhv$1@ID-67650.news.dfncis.de> Hello, I use C++ and need to compile my project with -fno-implicit-templates. This requires the templates to be instantiated explicitly. I managed to do this, and also managed to instantiate some template functions. However, I still got one unresolved external regarding the insert template method for a list. The mangled symbol is: _insert__H1Zt15__list_iterator3Z12MedeaWStringZRC12MedeaWStringZPC12MedeaWSt ring_t4list2Z1 2MedeaWStringZt24__default_alloc_template2b1i0Gt15__list_iterator3Z12MedeaWS tringZR12Medea WStringZP12MedeaWStringX02T2_v I tried to demangle it using c++filt386 but that does not work. It returns the supplied mangled symbol, so I assume it is unable to demangle it. If anyone can give me any hints on demangling the symbol above, any help would be appreciated. Albert Wijnja --------------------------- Newsgroups: comp.os.vxworks Subject: How to download vxWorks Image through serial port Date: Thu, 28 Jun 2001 14:33:22 +0530 From: Nayan Kumar Garg Organization: Lawrence Berkeley National Laboratory Message-ID: <9hesiv$1m0$1@overload.lbl.gov> Hi Can anyone tell me how to download the vxWorks image through the serial port . Nayan --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to download vxWorks Image through serial port Date: 28 Jun 2001 10:25:16 GMT From: borkhuis@agere.com (Johan Borkhuis) Organization: Agere Systems Message-ID: References: <9hesiv$1m0$1@overload.lbl.gov> Nayan Kumar Garg wrote: > Can anyone tell me how to download the vxWorks image through the serial > port . One possibility is to add PPP to your boot loader en setup a PPP link between your host and your target. You can also implement something like ZModem or even a hex download program, but you have to do that yourself. Groeten, Johan - -- o o o o o o o . . . _____________________________ o _____ || Johan Borkhuis | .][__n_n_|DD[ ====_____ | borkhuis@agere.com | >(________|__|_[_________]_|__________________________| _/oo OOOOO oo` ooo ooo 'o!o!o o!o!o` === VxWorks FAQ: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html === --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Thu Jun 28 05:04:17 2001 From: "Manjunath Bhat" Date: Thu Jun 28 05:04:20 PDT 2001 Subject: Vxworks Torndo Hi, Can anybody analyse what this error message Tornado II means? This error occurs very randomly when I try to run a particular application. WTX Error 0x100cc (AGENT_RT_ERROR) 0x4b40c08 (tWdbTask): memPartAlloc: block too big - 29344 in partition 0x446298 Kindly explain what each componet of this message refers to. Thanking you, Manjunath Bhat. _________________________________________________________________________ Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com. From vxwexplo-errs@csg.lbl.gov Fri Jun 29 02:34:36 2001 From: fscheurer@manz-acs.de Date: Fri Jun 29 02:34:38 PDT 2001 Subject: Re: comp.os.vxworks newsdigest - Re: Additional code & alternative tool chain Hi VxWorkers, hi tazboy, Thank you! How do I control that the linker doesn't try to bind unreferenced symbols, that the loader is able to do the job? Another question: What is the right format? Where can I find detailed information? TIA MfG / Kind Regards Friedrich Scheurer, R&D -------------------------------- acs GmbH Steigaeckerstrasse 13 72768 Reutlingen Telefon (07121) 9019-23 Telefax (07121) 9019-98 Internet www.manz-acs.de e-mail: fscheurer@manz-acs.de -------------------------------- -------- Original Message -------- > > Newsgroups: comp.os.vxworks > Subject: Re: Additional code & alternative tool chain > Date: Wed, 27 Jun 2001 11:46:06 -0500 > From: "tazboy78" > Organization: http://extra.newsguy.com > Message-ID: <9hd2im020ti@enews2.newsguy.com> > References: <9hcbnk$6tv$1@overload.lbl.gov> > > > Is it possible to load and run objects compiled e.g. > > with GNU gccx86 v2.95.x? > > Yes, but the objects will be built for dynamic linking at download. If you > can handle this you should be okay. > INFO - I previously posted the following: Hi VxWorkers, Additional code without TORNADO tool chain possible? we have a VxWorks based system (x86) developed by a partner where I want to add some cross-compiled object-modules. They wouldn't use VxWorks-code (no libs avail.), but only ANSI-C and the API of the partner's application code. Is it possible to load and run objects compiled e.g. with GNU gccx86 v2.95.x? Is there something special to pay attention to when cross-developing and providing standard-C based object modules, e.g. object format, options, segmentation, etc.? Would the loader accept object modules from an alternative tool-chain? From vxwexplo-errs@csg.lbl.gov Fri Jun 29 04:03:30 2001 From: Vxworks Exploder Date: Fri Jun 29 04:03:32 PDT 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Fri Jun 29 04:03:16 PDT 2001 Subject: re: help required on linker configuration on tornado Subject: VGA Graphics and VxWorks (PC arch) Subject: Re: How to download vxWorks Image through serial port Subject: Vxworks Torndo Subject: Re: about VxWorks MUX interface Subject: Re: Disk Cache Subject: Re: Disk Cache Subject: Re: Disk Cache Subject: Help - why receive "Truncated class file" error when opening an applet. Subject: how to get gw address ? Subject: Assembler with VxWorks Subject: MPC8260 interrupt OSE vs vxworks Subject: Two IP stacks in one CPU Subject: how to make rpcgen(on UNIX) generated server side code work under Subject: newbie question. steps for hardisk installation Subject: Why would you re-build the boot Rom ? Subject: is there any crash dump analysis tool for VxWorks on MIPS? Subject: Re: Two IP stacks in one CPU Subject: SSH on VxWorks Subject: nest mpc 8260 interrupt? Subject: Re: Two IP stacks in one CPU Subject: Re: Why would you re-build the boot Rom ? Subject: Re: Why would you re-build the boot Rom ? Subject: about realloc memo in TornadoII Subject: function to get using cpu rate & memory (as spy) itself kill too much cpu, and another question Subject: Re: Two IP stacks in one CPU Subject: Re: nest mpc 8260 interrupt? Subject: Re: nest mpc 8260 interrupt? Subject: sample source code for VxWorks driver Subject: re: comp.os.vxworks newsdigest - Re: Additional code & alternative tool chain ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: re: help required on linker configuration on tornado Date: Thu, 28 Jun 2001 10:01:14 +0200 From: "Kartha, Unnikrishan P. (ext)" Organization: Lawrence Berkeley National Laboratory Message-ID: <9hf3jv$43v$1@overload.lbl.gov> Hi, There is no harm in editing the makefile generated. I assume that you are using Tornado2.0 There are two possibilities: You can create a makefile of your own (refer GNU Make user's guide provided by Windriver) and execute it from Windows shell after setting all the paths correctly. Easier would be to edit the makefile generated and make the changes you need (removing f2.c from the archive path) and then execute make using this make file. The problem here is that next time you build the project from Tornado2.0, the makefile again gets regenerated and your changes will not be there. When you have specific requirements, it is better to have an own makefile. You can create one for ur own referring to the makefile generated dynamically. Kartha. Submitted-by vxwexplo-errs@csg.lbl.gov Wed Jun 27 11:08:24 2001 Submitted-by: Sujeet Naik Hi everyone, I have perhaps a strange requirement. I have a project with 4 files viz., f1.c, f2.c, f3.c and f4.c (all are C source files). When I make the dependencies and build, I compile all these four files and create an archive MYPROJ.a. This is later linked with other archives to create a final binary. My requirement is, I want to force the functions from f2.c to start from a fixed location, while the other files should be linked and located contiguously with other archive files. For this, I have my linker directive (.ld file) forcing f2.o to start from a fixed address. The problem is, when I am doing this, I am not supposed to have f2.o as part of MYPROJ.a. Otherwise, I will have multiple instances of all the functions in f2.c since the .ld file already has forced f2.o from a fixed location. Can someone help me on deciding how to eliminate f2.o from MYPROJ.a. I cant manually go and edit makefile of this project since it is dynamically created. Also, if this cant be done, is there any other way I can achieve my goal. Thanks in advance, Sujeet - ----------------------------------------------------------- This Mail has been checked for Viruses Attention: Encrypted Mails can NOT be checked ! *** Diese Mail wurde auf Viren ueberprueft Hinweis: Verschluesselte Mails koennen NICHT geprueft werden! - ------------------------------------------------------------ --------------------------- Newsgroups: comp.os.vxworks Subject: VGA Graphics and VxWorks (PC arch) Date: Thu, 28 Jun 2001 14:09:48 +0200 From: Rob Uiterlinden Organization: National Aerospace Laboratory (NLR) Message-ID: <3B3B1E8C.5550A8E8@nlr.nl> Hi, I am planning to port an MSDOS application to VxWorks on an embedded PC. The application is written in Turbo C and uses the Borland Graphics Interface (BGI) to produce graphics on a VGA screen. The graphics comprise mainly X-t plots containing traces of measured data. How is graphics programming accomplished is VxWorks? Should I use WindML or Zinc or ...? Opinions, anyone? Thanks in advance. Regards, Rob Uiterlinden --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to download vxWorks Image through serial port Date: Thu, 28 Jun 2001 13:07:26 +0100 From: Paul Caswell Organization: Land & Sea Systems Message-ID: <3B3B1DFE.65DFC239@baesystems.com> References: <9hesiv$1m0$1@overload.lbl.gov> I'm assuming that you don't have a network connection and want to download your vxWorks image over the serial link. As I understand it there are a few ways to do this the simplest being to use the TSFS (Target Server File System). See Page 136 of the Tornado 2.0 User Guide. Hope this helps Paul --------------------------- Newsgroups: comp.os.vxworks Subject: Vxworks Torndo Date: Thu, 28 Jun 2001 12:04:10 -0000 From: "Manjunath Bhat" Organization: Lawrence Berkeley National Laboratory Message-ID: <9hf74f$5b2$1@overload.lbl.gov> Hi, Can anybody analyse what this error message Tornado II means? This error occurs very randomly when I try to run a particular application. WTX Error 0x100cc (AGENT_RT_ERROR) 0x4b40c08 (tWdbTask): memPartAlloc: block too big - 29344 in partition 0x446298 Kindly explain what each componet of this message refers to. Thanking you, Manjunath Bhat. _________________________________________________________________________ Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: about VxWorks MUX interface Date: Thu, 28 Jun 2001 20:26:43 +0530 From: Yusuf Motiwala Organization: Lucent Technologies, Columbus, Ohio Message-ID: <3B3B45AB.52CF4D1D@lucent.com> References: <9h2bq1$aj1$1@overload.lbl.gov> I suggest you to grab the source code and do neccessary rather than approaching for any indirect methods (yes, indirect methods are possible). Otherwise you will loose all adv of network processor and may result in performance degradation Yusuf Srinivas Reddy wrote: > > hello all, --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Disk Cache Date: 28 Jun 2001 08:40:19 -0700 From: emartin@zetec.com (Eric Martin) Organization: http://groups.google.com/ Message-ID: <158d62fa.0106280740.2aba9eb@posting.google.com> References: <9hd4b0$2k0$1@overload.lbl.gov> > > > > [AndrayK]. The second arguement to dcacheDevCreate() is the pointer to memory > region where disk cache is to be placed (or NULL for malloc()). Did you try to > pre-allocate this memory region, and pass it's address to all subsequent > dcacheDevCreate() calls ? > > Andray This keeps the ram from getting eaten up. There is also a referrence to the cbio_memBase in cbioLibP.h. But when I do a dcacheShow the old cache devices are still there and the DCACHE_MAX_DEVS is 16. So after I replace disks 16 times I get a Error: dcacheDevCreate failed: errno = 0x13 - -> printErrno 0x13 0x13 = S_errno_ENODEV value = 0 = 0x0 I need to find a way to remove the old cache devices. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Disk Cache Date: Thu, 28 Jun 2001 16:58:29 +0100 From: Simon Farnsworth Message-ID: <3B3B5425.2000503@snellwilcox.com> References: <9hd4b0$2k0$1@overload.lbl.gov> Eric Martin wrote: > Tornado 2.0.2 - DosFs 2.0 - PPC 2603 > > We are using Fujitsu MO drives 230M 540M 640M and 1.3G drives. I want to be > able to pull a disk out and replace it with any other and have the system > handle it. > > Disk Initialization: > I create a SCSI physical device structure (scsiPhysDevCreate), then define a > logical partition on a SCSI block device (scsiBlkDevCreate), then I do a > cbioDeviceVerify to create CBIO wrapper for BLK_DEV device which returns a > CBIO device pointer which I use to create a disk cache (dcacheDevCreate). > Then I do a dosFsDevCreate which takes the CBIO_DEV_ID device handle > returned by dcacheDevCreate. And all this seems to work. > > NOW, when I sense that a disk has been removed. I remove the device from the > I/O system device list like this > > if ( iosDevFind (base_dir[drive_num], &ptail)!=NULL) > iosDevDelete (iosDevFind (base_dir[drive_num], &ptail)); > > and delete the SCSI physical-device structure > > scsiPhysDevDelete (pDisk[drive_num]); > > and set the related pointers to NULL. Then I go through the whole Disk > Initialization process with the new disk and it seems to work fine. At this point, to reinitialise the CBIO layer, you should do a CBIO_RESET cbioIoctl, with the third argument the new BLK_DEV; according to the documentation, this should cause your CBIO layer and DOSFS layer to reread the new disk. Note that I haven't tested this; it is based on a bit of RTFMing. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Disk Cache Date: Thu, 28 Jun 2001 17:07:36 +0100 From: Simon Farnsworth Message-ID: <3B3B5648.9070608@snellwilcox.com> References: <9hd4b0$2k0$1@overload.lbl.gov> <3B3B5425.2000503@snellwilcox.com> Simon Farnsworth wrote: > Eric Martin wrote: > >> Tornado 2.0.2 - DosFs 2.0 - PPC 2603 >> >> We are using Fujitsu MO drives 230M 540M 640M and 1.3G drives. I want >> to be >> able to pull a disk out and replace it with any other and have the system >> handle it. >> >> Disk Initialization: >> I create a SCSI physical device structure (scsiPhysDevCreate), then >> define a >> logical partition on a SCSI block device (scsiBlkDevCreate), then I do a >> cbioDeviceVerify to create CBIO wrapper for BLK_DEV device which >> returns a >> CBIO device pointer which I use to create a disk cache (dcacheDevCreate). >> Then I do a dosFsDevCreate which takes the CBIO_DEV_ID device handle >> returned by dcacheDevCreate. And all this seems to work. >> >> NOW, when I sense that a disk has been removed. I remove the device >> from the >> I/O system device list like this >> >> if ( iosDevFind (base_dir[drive_num], &ptail)!=NULL) >> iosDevDelete (iosDevFind (base_dir[drive_num], &ptail)); >> >> and delete the SCSI physical-device structure >> >> scsiPhysDevDelete (pDisk[drive_num]); >> >> and set the related pointers to NULL. Then I go through the whole Disk >> Initialization process with the new disk and it seems to work fine. > > > > > At this point, to reinitialise the CBIO layer, you should do a > CBIO_RESET cbioIoctl, with the third argument the new BLK_DEV; according > to the documentation, this should cause your CBIO layer and DOSFS layer > to reread the new disk. Note that I haven't tested this; it is based on > a bit of RTFMing. > Ooops; forgot to snip enough. You shouldn't go through the whole initialization process. Instead of creating a new wrapper, cache and fs, you should call the ioctl. --------------------------- Newsgroups: comp.lang.java.gui,comp.lang.java.help,comp.lang.java.programmer,comp.os.vxworks Subject: Help - why receive "Truncated class file" error when opening an applet. Date: Thu, 28 Jun 2001 11:20:04 -0500 From: "Lee" Message-ID: <9hflfl$dml7n$1@ID-92003.news.dfncis.de> Hi, Please help me on this if you have a clue. I'm using WindWebServer inside VxWork real time OS. When I open a web page which contains a simple applet (myApplet.class), I can only see a gray box with text saying "Loading Java Applet". Open Java Console and find the following error message showing "Truncated class file". ... ... Applet Installation finished. Opening http://123.456.78.90/fs/dynaform/myApplet.class Cache filename: C:\Documents and Settings\lee\Local Settings\Temporary Internet Files\Content.IE5\PQGN78C5\myApplet[1].class java.lang.ClassFormatError: myApplet (Truncated class file) ... ... Questions are: Q1: why I receive "Truncated class file" while the class file itself is in good shape ( The same class file works fine if I put it under IIS)? Maybe WindWebServer doesn't send myApplet.class file as a whole piece to web browser? Q2: The cache file myApplet[1].class is only 4 bytes but myApplet.class file is about 800 bytes. What is the cached file myApplet[1].class? what is it for? Can I turn off cache and let browser get the whole class file? Thanks. Lee --------------------------- Newsgroups: comp.os.vxworks Subject: how to get gw address ? Date: Thu, 28 Jun 2001 11:47:25 -0700 From: Waseem Siddiqi Organization: Cisco Systems Inc. Message-ID: <3B3B7BBD.585E238B@yahoo.com> How can I get gw address of a VXWorks box programmatically ? I am using routing sockets as suggested by Stevens but the write() is failing thanks waseem --------------------------- Newsgroups: comp.os.vxworks Subject: Assembler with VxWorks Date: Tue, 26 Jun 2001 18:33:15 -0400 From: "Harman Bassan" Organization: The University of Western Ontario, London, Ont. Canada Message-ID: <9hb1mi$8g2$1@panther.uwo.ca> Hi everyone, Can some one give me some idea on How can i do inline assembly in vxWorks with GCC. I looked at the GCC toolkit but things didn't look that clear to me. For example , the following lines which are valid with Visual C++, How can they be conerted to run with GCC on VxWorks. _asm { mov dx,378h mov al,0ffh out dx,al } Thanks in advance, Harman. --------------------------- Newsgroups: comp.os.vxworks,comp.realtime Subject: MPC8260 interrupt OSE vs vxworks Date: Thu, 28 Jun 2001 15:51:55 GMT From: u687591552@spawnkill.ip-mobilphone.net Message-ID: Sender: u687591552@spawnkill.ip-mobilphone.net In OSE, what does it take as interrupt priority for the interrupt processes, according to the interrupt controler's priority table? Does it consider 603e's priority table? a stupid question? How is different vs vxwork in term of dealing interrupts? BT - -- Sent by guangbutun from hotmail subdomain of com This is a spam protected message. Please answer with reference header. Posted via http://www.usenet-replayer.com/cgi/content/new --------------------------- Newsgroups: comp.os.vxworks Subject: Two IP stacks in one CPU Date: 28 Jun 2001 13:31:39 -0700 From: du_bin@yahoo.com (billy du) Organization: http://groups.google.com/ Message-ID: <340587ac.0106281231.714528e0@posting.google.com> Can VxWorks support two IP stacks inside one CPU? For example, a card in the box has two interfaces. One is connected to outside world, the other is internal connection. Both has IP addresses. Unfortuately, the outside world might use the private IP addresses. If we uses private IP addresses for internal connection, there might be conflict. Therefore, we plan to use two IP stacks in one card. Can anyone help me? --------------------------- Newsgroups: comp.os.vxworks Subject: how to make rpcgen(on UNIX) generated server side code work under vxworks 5.4 Date: Thu, 28 Jun 2001 16:24:51 -0400 From: "Xu, Xiaoyi [BVW:9P81:EXCH]" Organization: Nortel Networks Message-ID: <3B3B9293.E49CE939@americasm01.nt.com> --------------------------- Newsgroups: comp.os.vxworks Subject: newbie question. steps for hardisk installation Date: Thu, 28 Jun 2001 23:38:02 +0200 From: "Rui Germano" Organization: Planet Internet Message-ID: <9hg7v7$48e8$1@reader01.wxs.nl> Hello: I would like to know where can I learn the steps to make an hardisk installation of Vxworks on a x86 system. Regards Rui --------------------------- Newsgroups: comp.os.vxworks Subject: Why would you re-build the boot Rom ? Date: Fri, 29 Jun 2001 07:15:02 +1000 From: Chris Chock Organization: Pacific Internet (Australia) Message-ID: <3B3B9E56.1040504@hunterlink.net.au> Hi All, Under what circumstances do you need to re-build the boot ROM ? I am running T2 development environment and am considering implementing dosFS2. Do I need to re-build the boot ROM that was supplied by the board manufacturer and then re-flash it ? Thanks in advance. Chris Chock. --------------------------- Newsgroups: comp.os.vxworks Subject: is there any crash dump analysis tool for VxWorks on MIPS? Date: 28 Jun 2001 15:14:34 -0700 From: ramkumar_sankar@yahoo.com (ramku) Organization: http://groups.google.com/ Message-ID: thanks in advance - -ramku --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Two IP stacks in one CPU Date: Thu, 28 Jun 2001 22:30:38 GMT From: Dan Gold Organization: http://www.newsranger.com Message-ID: References: <340587ac.0106281231.714528e0@posting.google.com> In article <340587ac.0106281231.714528e0@posting.google.com>, billy du says... > >Can VxWorks support two IP stacks inside one CPU? > >For example, a card in the box has two interfaces. One is connected to >outside >world, the other is internal connection. Both has IP addresses. >Unfortuately, the outside world might use the private IP addresses. If >we uses private IP addresses for internal connection, there might be >conflict. Therefore, we plan to use two IP stacks in one card. > >Can anyone help me? I think the first thing you might need is a basic networking course. Why would you want to use well-known (and registered) IP addresses as internal ones ? I apologize if I'm not understanding correctly, but you want to use an internal IP address such as 147.11.1.23. The problem is on the outside world, that already belongs to wrs.com ? Why don't you use the 192.168.x.x network for internal usage, since I believe that's what it's there for... hth, Dan Gold gold@ensemble.com --------------------------- Newsgroups: comp.os.vxworks Subject: SSH on VxWorks Date: 28 Jun 2001 16:50:52 -0700 From: tapankarwa@yahoo.com (Tapan Karwa) Organization: http://groups.google.com/ Message-ID: <7c4037a9.0106281550.732ff4a5@posting.google.com> Hi, Does anyone know if SSH has been ported onto VxWorks. This question has been asked before but there were very few answers. Any pointers will be really appreciated. Thanks, Tapan. --------------------------- Newsgroups: comp.os.vxworks Subject: nest mpc 8260 interrupt? Date: Thu, 28 Jun 2001 19:26:48 GMT From: u687591552@spawnkill.ip-mobilphone.net Message-ID: Sender: u687591552@spawnkill.ip-mobilphone.net I was told that for mpc 8260, an interrupt can not preempt another interrupt, if one ISR starts, it goes until finished. I am not sure if that's true at least I can not find support in the MPC8260 manual. If anyone has such experience can tell me something about it. - -- Sent by guangbutun from hotmail included in com This is a spam protected message. Please answer with reference header. Posted via http://www.usenet-replayer.com/cgi/content/new --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Two IP stacks in one CPU Date: Thu, 28 Jun 2001 19:39:10 -0700 From: "Michael R. Kesti" Organization: MK Associates Message-ID: <3B3BEA4E.85E6452E@gv.net> References: <340587ac.0106281231.714528e0@posting.google.com> Reply-To: mkesti@gv.net billy du wrote: >Can VxWorks support two IP stacks inside one CPU? > >For example, a card in the box has two interfaces. One is connected to >outside >world, the other is internal connection. Both has IP addresses. >Unfortuately, the outside world might use the private IP addresses. If >we uses private IP addresses for internal connection, there might be >conflict. Therefore, we plan to use two IP stacks in one card. Rather than run multiple IP stacks, you want to ipAttach() as many as IP_MAX_UNITS network interfaces to the IP stack and assign their net masks and addresses with ifMaskSet() and ifAddrSet(). (The interface specified on the boot line may have been attached and had its net mask and address set by the boot process.) The IP stack then resolves to which interface a request is sent based on the contents of the sockaddr structure passed to the requesting sockLib calls (sendto(), connect(), and the like) and on the interfaces' net masks and addresses. The interface whose address and'ed with its net mask is equal to the sockaddr address and'ed with the interface's net mask is the interface to which the request is sent. See sections 4.1, 4.2, and 7.2 of the VxWorks 5.4 Network Programmer's Guide for more information. - -- ======================================================================== Michael Kesti | "And like, one and one don't make | two, one and one make one." mkesti@gv.net | - The Who, Bargain --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Why would you re-build the boot Rom ? Date: Thu, 28 Jun 2001 19:43:26 -0700 From: DrDiags Organization: laserlink.net Message-ID: <3B3BEB4E.2D5AD0D5@covad.net> References: <3B3B9E56.1040504@hunterlink.net.au> Reply-To: drdiags@covad.net Chris Chock wrote: > > Hi All, > > Under what circumstances do you need to re-build the boot ROM ? - Adding code required to support a feature which by default was not enabled by the BSP developer. By this I mean such things as DHCP Client support, TFFS if I wanted to boot off an vxWorks kernel image that existed on a TFFS flash drive, PPP or SLIP, if this is how I wanted to grab my development kernel, etc. For the most part, if there were hardware changes that need to be addressed in the bootrom that could not be delayed until the kernel was initialized. This is just my two cents.. > I am running T2 development environment and am considering implementing > dosFS2. Do I need to re-build the boot ROM that was supplied by the > board manufacturer and then re-flash it ? > - I don't think you need to. Others may be able to point out why you would need this built into your bootrom and not wait for the kernel containing it, to be initialized. Had you planned on using your ATA drive as your boot device? > Thanks in advance. > > Chris Chock. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Why would you re-build the boot Rom ? Date: Thu, 28 Jun 2001 19:41:52 -0700 From: "Michael R. Kesti" Organization: MK Associates Message-ID: <3B3BEAF0.EDA40F47@gv.net> References: <3B3B9E56.1040504@hunterlink.net.au> Reply-To: mkesti@gv.net Chris Chock wrote: >Under what circumstances do you need to re-build the boot ROM ? I am >running T2 development environment and am considering implementing >dosFS2. Do I need to re-build the boot ROM that was supplied by the >board manufacturer and then re-flash it ? The boot ROM needs to be rebuilt and updated anytime you want the results of editing the BSP files to reflect the changes you have made. Adding dosFS2 is just such an occasion. - -- ======================================================================== 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: about realloc memo in TornadoII Date: Fri, 29 Jun 2001 11:18:17 +0800 From: "SimbaDa" Organization: Beijing Telegraph Administration Message-ID: <9hgs4n$mru$1@info.bta.net.cn> Reply-To: "SimbaDa" This is a multi-part message in MIME format. - ------=_NextPart_000_0046_01C1008D.31E674A0 Content-Type: text/plain; charset="gb2312" Content-Transfer-Encoding: quoted-printable Hi Vxwork guru... I have a quetion that what different between Win32 and Tonardo.. we can get buffer according to our requirement in Win32, but I do not = know how about Tornado.. can anyone tell me... for example: we can define a array with following: #define BUFSIZE 1024 char buffer[BUFSIZE]; we will not get the right result in Tornado.. can anyone help me? Best wish - ------=_NextPart_000_0046_01C1008D.31E674A0 Content-Type: text/html; charset="gb2312" Content-Transfer-Encoding: quoted-printable
Hi Vxwork guru...
 
 
  I have  a quetion that what different between Win32 and = Tonardo..
we can get buffer according to our requirement in Win32, = but I do=20 not know how about Tornado..
can anyone tell me...
 
 for example:
   we can define a array with following:
    #define    = BUFSIZE   =20 1024
       char   =20 buffer[BUFSIZE];
 
 we will not get the right result in Tornado..
 
can anyone help me?
Best wish
 
 
- ------=_NextPart_000_0046_01C1008D.31E674A0-- --------------------------- Newsgroups: comp.os.vxworks Subject: function to get using cpu rate & memory (as spy) itself kill too much cpu, and another question Date: 28 Jun 2001 21:47:18 -0700 From: akinliu@iwv.com.hk (AkinLiu) Organization: http://groups.google.com/ Message-ID: <5e64af6d.0106282047.4360efbb@posting.google.com> Hi, vxWorks guys, i need your help... now i am writing a task to moniter cpu rate for all tasks and how much every task using memory, and get cpu rate & mem of each task. but now 2 question hit me : 1st: spyLib and its functions such as spy(),spyReport() kill too much cpu rate, --even 50%+ of that, that will make result ugly if i want to know each task takes how much cpu rate when system is busying. :( and 2nd: tcb = taskTcb(tid); and GetTaskRes(tid, *tRes) to get -- Res->tid = Tid; Res->name = taskName(Tid); Res->StackSize = td.td_stackSize; Res->StackMargin = td.td_stackMargin; Res->MaxTmrs = iw_tcb->max_act_tmr; Res->TmrMbx = iw_tcb->dflt_tmr_mbx; Res->TmrPri = iw_tcb->dflt_tmr_pri; Res->NumMalloc = IWQ_LEN(&iw_tcb->mem_q); Res->MallocMem = mem_mem; But it seems these resource are allocted when task is spawned, and never changed when task is running, that means i can't get dynamic memory value. so what's the correct method i should take? please someone answer these 2 questions, thanx and appreciate your answer. Best regards! --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Two IP stacks in one CPU Date: Fri, 29 Jun 2001 08:49:11 +0200 From: Emmanuel Herbreteau Organization: SEPRO Robotique (BEP) Message-ID: <3B3C24E7.3783DE68@sepro-robotique.com> References: <340587ac.0106281231.714528e0@posting.google.com> billy du wrote: > Can VxWorks support two IP stacks inside one CPU? What ??? > Unfortuately, the outside world might use the private IP addresses. If > we uses private IP addresses for internal connection, there might be > conflict. Therefore, we plan to use two IP stacks in one card. You should use private IP address for your private network : Look at RFC 1918 on the Internet (or a good book, like "TCP/IP" from MacMillan). ie : Private address (class C) "192.168.16.44" or private address (class B) "172.16.1.27" will never conflict with an Internet address. - -- Best regards Emmanuel --------------------------- Newsgroups: comp.os.vxworks Subject: Re: nest mpc 8260 interrupt? Date: Fri, 29 Jun 2001 09:17:59 +0100 From: Graham Baxter Organization: Graham Baxter (Software) Limited Message-ID: <3B3C39B7.DA57C2F7@NOSPAM.bcs.org.uk> References: If you are using a Wind River BSP, look in m8260Int.c. You will see that you can prioritize the various 8260 interrupts that occur on the PPC external interrupt vector. For each interrupt you provide a mask of all higher priority interrupts. In the ISR: 1. The interrupt mask (SIMR) is first saved. 2. SIMR is written with a value to only allow higher priority external (8260) interrupts to get in. 3. The external interrupt is cleared in PPC MSR 4. The interrupt is serviced. 5. SIMR is restored. TIPS: If you want to use lots of nesting, ensure that your interrupt stack is large enough. If you are using caching and experience un-initialized interrupt messages being displayed, do a CACHE_PIPE_FLUSH() after writing the new value into SIMR (step 2). - -- Regards, Graham Baxter Freelance Software Engineer gbaxter@NOSPAM.bcs.org.uk u687591552@spawnkill.ip-mobilphone.net wrote: > > I was told that for mpc 8260, an interrupt can not preempt another > interrupt, if one ISR starts, it goes until finished. I am not sure > if that's true at least I can not find support in the MPC8260 --------------------------- Newsgroups: comp.os.vxworks Subject: Re: nest mpc 8260 interrupt? Date: Fri, 29 Jun 2001 09:24:47 +0100 From: Graham Baxter Organization: Graham Baxter (Software) Limited Message-ID: <3B3C3B4F.964D9DC6@NOSPAM.bcs.org.uk> References: File is m8260IntrCtl.c. u687591552@spawnkill.ip-mobilphone.net wrote: > > I was told that for mpc 8260, an interrupt can not preempt another > interrupt, if one ISR starts, it goes until finished. I am not sure > if that's true at least I can not find support in the MPC8260 > manual. If anyone has such experience can tell me something about > it. > > > -- > Sent by guangbutun from hotmail included in com > This is a spam protected message. Please answer with reference header. > Posted via http://www.usenet-replayer.com/cgi/content/new - -- Regards, Graham Baxter Freelance Software Engineer gbaxter@NOSPAM.bcs.org.uk --------------------------- Newsgroups: comp.os.vxworks Subject: sample source code for VxWorks driver Date: Fri, 29 Jun 2001 09:45:05 +0200 From: "Joerg Engel" Organization: Customer of UUNET Deutschland GmbH Message-ID: <3b3c3f6a$0$1795$4dbef881@businessnews.de.uu.net> Are there any websites whre I can find sample source code dor VxWorks dirver ? --------------------------- Newsgroups: comp.os.vxworks Subject: re: comp.os.vxworks newsdigest - Re: Additional code & alternative tool chain Date: 29 Jun 2001 09:34:29 UT From: fscheurer@manz-acs.de Organization: Lawrence Berkeley National Laboratory Message-ID: <9hhkff$2j4$1@overload.lbl.gov> Hi VxWorkers, hi tazboy, Thank you! How do I control that the linker doesn't try to bind unreferenced symbols, that the loader is able to do the job? Another question: What is the right format? Where can I find detailed information? TIA MfG / Kind Regards Friedrich Scheurer, R&D - -------------------------------- acs GmbH Steigaeckerstrasse 13 72768 Reutlingen Telefon (07121) 9019-23 Telefax (07121) 9019-98 Internet www.manz-acs.de e-mail: fscheurer@manz-acs.de - -------------------------------- - -------- Original Message -------- > > Newsgroups: comp.os.vxworks > Subject: Re: Additional code & alternative tool chain > Date: Wed, 27 Jun 2001 11:46:06 -0500 > From: "tazboy78" > Organization: http://extra.newsguy.com > Message-ID: <9hd2im020ti@enews2.newsguy.com> > References: <9hcbnk$6tv$1@overload.lbl.gov> > > > Is it possible to load and run objects compiled e.g. > > with GNU gccx86 v2.95.x? > > Yes, but the objects will be built for dynamic linking at download. If you > can handle this you should be okay. > INFO - I previously posted the following: Hi VxWorkers, Additional code without TORNADO tool chain possible? we have a VxWorks based system (x86) developed by a partner where I want to add some cross-compiled object-modules. They wouldn't use VxWorks-code (no libs avail.), but only ANSI-C and the API of the partner's application code. Is it possible to load and run objects compiled e.g. with GNU gccx86 v2.95.x? Is there something special to pay attention to when cross-developing and providing standard-C based object modules, e.g. object format, options, segmentation, etc.? Would the loader accept object modules from an alternative tool-chain? --------------------------- End of New-News digest ********************** From vxwexplo-errs@csg.lbl.gov Fri Jun 29 05:58:50 2001 From: Jeff David Date: Fri Jun 29 05:58:53 PDT 2001 Subject: How to get rid of const warning Hello vxWorks users! When the code contains lines such as the following: if (ipAttach(0, "dc") != OK) ... I get the following warning: pniend.c:666: warning: passing arg 2 of `ipAttach' discards `const' from pointer target type Does anybody know how to change the code to get rid of this warning? (Note: Suppressing the warning is not an option.) Here's the OS and compiler info: /vobs/OS/cmts/dev/pni> uname -a SunOS cc1 5.7 Generic_106541-15 sun4u sparc SUNW,Ultra-Enterprise /vobs/OS/cmts/dev/pni> ccppc -v Reading specs from /vobs/OS/wind/host/sun4-solaris2/lib/gcc-lib/powerpc-wrs-vxworks/cygnus-2.7. 2-960126/specs gcc driver version cygnus-2.7.2-960126 egcs-971225 tornado 2.0 executing gcc version cygnus-2.7.2-960126 Thanks, Jeff jdavid@riverdelta.com From vxwexplo-errs@csg.lbl.gov Fri Jun 29 10:07:59 2001 From: "Ben Kelly" Date: Fri Jun 29 10:08:01 PDT 2001 Subject: RE: How to get rid of const warning You can use the C++ casting facilities to tell the compiler that you know what you are doing and to please not spew out a ton of warnings. Do this: if (ipAttach(0, const_cast("dc") != OK) ... You should only do this when you do not have the option of fixing the underlying library to advertise const usage. That is assuming of course that the library does treat it like a const; if the library actually modifies the value I believe the compiler is free to do whatever it likes at this point (crash, reformat your hard drive, etc). Just as an FYI the other C++ casts are: static_cast - do the casting at compile time... replacement for c-style () casts dynamic_cast - do a runtime cast for casting up the hierarchy tree reinterpret_cast - the kill -9 of casts... use at your own risk... usually needed to deal with casting funtion pointers, etc. - Ben > -----Original Message----- > From: the vxWorks Users Group Exploder [mailto:vxwexplo@lbl.gov] > Sent: Friday, June 29, 2001 9:12 AM > To: vxworks_users@csg.lbl.gov > Subject: How to get rid of const warning > > > Submitted-by vxwexplo-errs@csg.lbl.gov Fri Jun 29 05:58:50 2001 > Submitted-by: Jeff David > > Hello vxWorks users! > > When the code contains lines such as the following: > > if (ipAttach(0, "dc") != OK) > ... > > I get the following warning: > pniend.c:666: warning: passing arg 2 of `ipAttach' discards `const' from > pointer target type > > Does anybody know how to change the code to get rid of this warning? (Note: > Suppressing the warning is not an option.) > From vxwexplo-errs@csg.lbl.gov Fri Jun 29 20:03:02 2001 From: "David Anderson" Date: Fri Jun 29 20:03:05 PDT 2001 Subject: FW: Network pool running out... -----Original Message----- From: David Anderson [SMTP:david.anderson@dtrack.com] Sent: 27 June 2001 16:05 To: 'Vxworks_Users (E-mail) Subject: Network pool running out... vxWorks 5.3.1 Tornado 1.0.1 Hello all, I've run into a problem which produces an apparent 'freeze up' of a PPP link, but still allows 'normal' Ethernet style TCP/IP connections to proceed. We are using the Tornado 2.0 ppp daemon. It appears that running an FTP session and a TELNET session concurrently over the PPP link causes the network pool to run out of buffers ( error 0x73000a ). However, the PPP link does not recover, the daemon tasks are running but you can no longer ping from either end to the other, or establish a new socket connection across it. The buffer error appears only to be a soft error - an Ethernet TCP/IP connection can still be set-up OK. Any ideas about why the PPP link should be 'frozen out' although all status indications suggest it is open and up. David Anderson. NEW E-MAIL ADDRESS: mailto:david.anderson@dtrack.com ------------------------------------- FF: 2B + ~2B , that is the question ? ------------------------------------- Disclaimer ---------- This email and any files transmitted with it are confidential and intended solely for the use of the individual or entity to whom they are addressed. If you have received this email in error please destroy and notify Data Track Technology Plc +44 1425 271900. ------------------------------------------------------- From vxwexplo-errs@csg.lbl.gov Fri Jun 29 20:56:38 2001 From: Nayan Kumar Garg Date: Fri Jun 29 20:56:40 PDT 2001 Subject: RE : comp.os.vxworks newsdigest Hi Paul , Thanks for the reply . I am using the same to download the vxWorks image through serial cable .Actually I am not using the network .So I want to download the vxWorks image through a serial cable .For this I am using TSFS .I have made all the changes in config.h which are required for that . But I am not able to download the image through the serial port . Whenever I boot the target machine I gets a message " Waiting for Target Server Connection " ..and on the Host Machine in the target server launching Log Console I get a message "Error: clnttty_rcv: read timed out " .And unlimately the target machine gets waiting for the VxWorks image to be downloaded .I have already checked the serial cable , it is working fine . So can you please suggest me some way to do this . Thanks & Regards nayan Newsgroups: comp.os.vxworks Subject: Re: How to download vxWorks Image through serial port Date: Thu, 28 Jun 2001 13:07:26 +0100 From: Paul Caswell Organization: Land & Sea Systems Message-ID: <3B3B1DFE.65DFC239@baesystems.com> References: <9hesiv$1m0$1@overload.lbl.gov> I'm assuming that you don't have a network connection and want to download your vxWorks image over the serial link. As I understand it there are a few ways to do this the simplest being to use the TSFS (Target Server File System). See Page 136 of the Tornado 2.0 User Guide. Hope this helps Paul From vxwexplo-errs@csg.lbl.gov Sat Jun 30 00:52:17 2001 From: "lhp" Date: Sat Jun 30 00:52:19 PDT 2001 Subject: why signal sample source code generates exception (VxWorks 5.4 for PPC) HI, I use Tornado II and VxWorks 5.4 for PPC. Why the following signal sample source code generates exception, no matter running in VxSim or running on an actual MPC860 board? HELP! my email: LHP@gpte.com.cn /* includes */ #include "vxWorks.h" #include "signal.h" /* forward declarations */ static void mysigHandle (int); void receivSig(); void sendSig(); static int rtaskid; void mysignal() { rtaskid = taskSpawn("receivSig",100,0,4000,receivSig,0,0,0,0,0,0,0,0,0,0); } void receivSig() { int n; logMsg("enter rec task.\n"); if(signal(SIGUSR1,mysigHandle)!=0) { logMsg("sigaction failed.\n"); } taskSpawn("sendSig",90,0,4000,sendSig,0,0,0,0,0,0,0,0,0,0); /*wait somewhile then exit*/ n=10; while(n-->0) { logMsg("exitting rec task...\n"); } } void sendSig() { int id; logMsg("enter send task.\n"); if(kill(rtaskid, SIGUSR1) != 0) { logMsg("send signal failed.\n"); } logMsg("exit send task.\n"); } static void mysigHandle (int signo) { logMsg("got the signal!\n"); } From vxwexplo-errs@csg.lbl.gov Sat Jun 30 04:03:28 2001 From: Vxworks Exploder Date: Sat Jun 30 04:03:31 PDT 2001 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sat Jun 30 04:03:14 PDT 2001 Subject: How to get rid of const warning Subject: Re: SSH on VxWorks Subject: Re: Why would you re-build the boot Rom ? Subject: WindView post-mortem: no luck yet Subject: Re: sample source code for VxWorks driver Subject: Re: Two IP stacks in one CPU Subject: RE: How to get rid of const warning Subject: rawIpLibInit Subject: Re: How common is SNiFF usage? Subject: Rational ClearCase with Tornado & VxWorks? Subject: Re: SSH on VxWorks Subject: blocking ip addresses Subject: Re: How to get rid of const warning Subject: FW: Network pool running out... Subject: Network pool running out... Subject: RE : comp.os.vxworks newsdigest Subject: Re: How to download vxWorks Image through serial port Subject: vxworks device, tcp port 513 open Subject: How can I block VxWorks default shell (->)? Subject: why signal sample source code generates exception (VxWorks 5.4 for PPC) ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: How to get rid of const warning Date: Fri, 29 Jun 2001 08:54:43 -0400 From: Jeff David Organization: Lawrence Berkeley National Laboratory Message-ID: <9hhv0v$68c$1@overload.lbl.gov> Hello vxWorks users! When the code contains lines such as the following: if (ipAttach(0, "dc") != OK) ... I get the following warning: pniend.c:666: warning: passing arg 2 of `ipAttach' discards `const' from pointer target type Does anybody know how to change the code to get rid of this warning? (Note: Suppressing the warning is not an option.) Here's the OS and compiler info: /vobs/OS/cmts/dev/pni> uname -a SunOS cc1 5.7 Generic_106541-15 sun4u sparc SUNW,Ultra-Enterprise /vobs/OS/cmts/dev/pni> ccppc -v Reading specs from /vobs/OS/wind/host/sun4-solaris2/lib/gcc-lib/powerpc-wrs-vxworks/cygnus-2.7. 2-960126/specs gcc driver version cygnus-2.7.2-960126 egcs-971225 tornado 2.0 executing gcc version cygnus-2.7.2-960126 Thanks, Jeff jdavid@riverdelta.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: SSH on VxWorks Date: 29 Jun 2001 09:24:07 -0400 From: Bill Pringlemeir Organization: Bell Sympatico Message-ID: References: <7c4037a9.0106281550.732ff4a5@posting.google.com> Sender: bpringlemeir@DeadDuck Tapan> Hi, Does anyone know if SSH has been ported onto VxWorks. This Tapan> question has been asked before but there were very few Tapan> answers. Any pointers will be really appreciated. "http://www.xs4all.nl/~borkhuis/vxworks/vxw_pt6.html#6.9" OpenSSH will work on top of OpenSSL. "http://www.openssh.org/" hth, Bill Pringlemeir. - -- Useful vxWorks URLS, "http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html" "http://www.xs4all.nl/~borkhuis/vxworks/vxw_pt9.html" --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Why would you re-build the boot Rom ? Date: Fri, 29 Jun 2001 13:53:44 GMT From: andreyk@home.com (Andray Kaganovsky) Organization: @Home Message-ID: References: <3B3B9E56.1040504@hunterlink.net.au> <3B3BEB4E.2D5AD0D5@covad.net> DrDiags wrote in news:3B3BEB4E.2D5AD0D5@covad.net: > > > Chris Chock wrote: >> >> Hi All, >> >> Under what circumstances do you need to re-build the boot ROM ? > > - Adding code required to support a feature which by default > was not enabled by the BSP developer. By this I mean such > things as DHCP Client support, TFFS if I wanted to boot > off an vxWorks kernel image that existed on a TFFS flash > drive, PPP or SLIP, if this is how I wanted to grab my > development kernel, etc. For the most part, if there were > hardware changes that need to be addressed in the bootrom > that could not be delayed until the kernel was initialized. > This is just my two cents.. > >> I am running T2 development environment and am considering implementing >> dosFS2. Do I need to re-build the boot ROM that was supplied by the >> board manufacturer and then re-flash it ? >> > > - I don't think you need to. Others may be able to point out > why you would need this built into your bootrom and not > wait for the kernel containing it, to be initialized. Had > you planned on using your ATA drive as your boot device? > >> Thanks in advance. >> >> Chris Chock. [AndrayK]. If you are using bootrom exclusively for loading vxWorks application from some mass storage device (SCSI disk, ATA disk, TrueFFS device etc.), than I belive you don't have to re-build bootrom, as far as migration from dosFs to dosFs-2 concerned. Andray --------------------------- Newsgroups: comp.os.vxworks Subject: WindView post-mortem: no luck yet Date: Fri, 29 Jun 2001 16:55:23 +0200 From: Ton Janssen Organization: Oce Technologies B.V. Message-ID: <3B3C96DB.D012CF9E@oce.nl> Hi there, Anyone experience with post-mortem mode Windview logging: I've got the following problem: I'm trying to debug a situation were the last thing that happens in a session is a hangup in (probably) interrupt context. So I tried to use post-mortem logging in WindView, together with the addition of a whole bunch of user events (wvevent(nr,0,0)) in the suspected code area. The problem is that when I hit the stop button, the target application hangs at interrupt level. After a reboot (the target has disconnected and connected again) I try to upload the windView buffer via NFS. However, the tool now complaints that the target buffer has not been created; The buffer is created in an area which is not cleared at boot-time, so it should work... I tried to type wvEvtLogStop from the shell after reboot but the target upload still fails. Anyone had success with post-mortem NFS uploads in comparable situations. If yes I might check the buffer settings over and over again... Kind Regards, - -- ############################################################ # Oce Technologies B.V. name: Ton Janssen (3N-38) # # P.O. Box 101 email: djan@oce.nl # # 5900 MA Venlo Private: dpc.janssen@hccnet.nl # # Tel: +31 (0)77-359 40 97 # # The Netherlands Fax: +31 (0)77-359 54 50 # ############################################################ This note does not necessarily represent the position of Oce Technoligies B.V. Therefore no liability or responsibility for whatever will be accepted. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: sample source code for VxWorks driver Date: 29 Jun 2001 08:08:46 -0700 From: john_94501@yahoo.com (John) Organization: http://groups.google.com/ Message-ID: <488e459a.0106290708.32a234ed@posting.google.com> References: <3b3c3f6a$0$1795$4dbef881@businessnews.de.uu.net> Hello Joerg, You can find some example driver source code in your VxWorks tree... look under target/src/drv. HTH, John... "Joerg Engel" wrote in message news:<3b3c3f6a$0$1795$4dbef881@businessnews.de.uu.net>... > Are there any websites whre I can find sample source code dor VxWorks dirver > ? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Two IP stacks in one CPU Date: 29 Jun 2001 09:32:19 -0700 From: du_bin@yahoo.com (billy du) Organization: http://groups.google.com/ Message-ID: <340587ac.0106290832.1d50b660@posting.google.com> References: <340587ac.0106281231.714528e0@posting.google.com> <3B3C24E7.3783DE68@sepro-robotique.com> Emmanuel Herbreteau wrote in message news:<3B3C24E7.3783DE68@sepro-robotique.com>... > billy du wrote: > > Can VxWorks support two IP stacks inside one CPU? > > What ??? > > > Unfortuately, the outside world might use the private IP addresses. If > > we uses private IP addresses for internal connection, there might be > > conflict. Therefore, we plan to use two IP stacks in one card. > > You should use private IP address for your private network : Look at > RFC 1918 on the Internet (or a good book, like "TCP/IP" from MacMillan). > > ie : Private address (class C) "192.168.16.44" or private address (class > B) > "172.16.1.27" will never conflict with an Internet address. Sorry, my question is not clear. We want to support private IP addresses in the outside world as well because our equipment might connect to a private network using IP address like 10.x.x.x 192.168.x.x. Then we have a problem to assign IP address inside the equipment. That's why we need two IP stack or any other approach. --------------------------- Newsgroups: comp.os.vxworks Subject: RE: How to get rid of const warning Date: Fri, 29 Jun 2001 13:07:01 -0400 From: "Ben Kelly" Organization: Lawrence Berkeley National Laboratory Message-ID: <9hid31$br0$1@overload.lbl.gov> You can use the C++ casting facilities to tell the compiler that you know what you are doing and to please not spew out a ton of warnings. Do this: if (ipAttach(0, const_cast("dc") != OK) ... You should only do this when you do not have the option of fixing the underlying library to advertise const usage. That is assuming of course that the library does treat it like a const; if the library actually modifies the value I believe the compiler is free to do whatever it likes at this point (crash, reformat your hard drive, etc). Just as an FYI the other C++ casts are: static_cast - do the casting at compile time... replacement for c-style () casts dynamic_cast - do a runtime cast for casting up the hierarchy tree reinterpret_cast - the kill -9 of casts... use at your own risk... usually needed to deal with casting funtion pointers, etc. - - Ben > -----Original Message----- > From: the vxWorks Users Group Exploder [mailto:vxwexplo@lbl.gov] > Sent: Friday, June 29, 2001 9:12 AM > To: vxworks_users@csg.lbl.gov > Subject: How to get rid of const warning > > > Submitted-by vxwexplo-errs@csg.lbl.gov Fri Jun 29 05:58:50 2001 > Submitted-by: Jeff David > > Hello vxWorks users! > > When the code contains lines such as the following: > > if (ipAttach(0, "dc") != OK) > ... > > I get the following warning: > pniend.c:666: warning: passing arg 2 of `ipAttach' discards `const' from > pointer target type > > Does anybody know how to change the code to get rid of this warning? (Note: > Suppressing the warning is not an option.) > --------------------------- Newsgroups: comp.os.vxworks Subject: rawIpLibInit Date: Fri, 29 Jun 2001 14:12:09 -0400 From: Jeremiah Frick Organization: Improv Systems, Inc. Message-ID: <3B3CC4F9.27D6655D@improvsys.com> This is a multi-part message in MIME format. - --------------538D54EAA5807D51432F11CB Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Anyone know what it does? :) I'm getting an error during my initialization when I call usrNetInit due to rawIpLibInit... Thanks! Jeremy - --------------538D54EAA5807D51432F11CB Content-Type: text/x-vcard; charset=us-ascii; name="jeremyf.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Jeremiah Frick Content-Disposition: attachment; filename="jeremyf.vcf" begin:vcard n:Frick;Jeremiah tel;work:716-756-5411 x45 x-mozilla-html:FALSE url:www.improvsys.com org:Improv Systems Inc.;PDG version:2.1 email;internet:jeremyf@improvsys.com title:Hardware Engineer adr;quoted-printable:;;3136 Winton Road South=0D=0ASuite 100;Rochester;New York;14623;United States of America fn:Jeremiah Frick end:vcard - --------------538D54EAA5807D51432F11CB-- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How common is SNiFF usage? Date: Fri, 29 Jun 2001 19:18:55 +0000 (UTC) From: Denis Organization: Your Company Message-ID: <9hikav$8g9$1@siamese.noc.ucla.edu> References: <9gijt8$abo$03$1@news.t-online.com> <9goe18$frm$05$1@news.t-online.com> <9gq81t$c2v$1@news.hdpp.de> <9gs7nq$gdr$1@news.hdpp.de> Joe, while i get a feeling you are not really into source-code browsing utils, let me suggest a couple which i use (which i havent seen mentioned in the thread, but which you may know about anyway) there is the source-navigator (now owned by redhat, i believe). runs on linux and windows platforms. it is a bit slow, but if you use it occasionally for call-tree analysis type of stuff, its fine. it handles macros much better then sniff does. http://sources.redhat.com/sourcenav/ it is not a full-fledged IDE however, so you would need to manually update the project if you were to change some files in it. again, i am not sure if this is what you are looking for. i think sniff takes care of that, and that's one of the nice things about it. further, i am not sure if sourcenav easily interfaces with SCM tools, but i am sure you could use some macros or whatnot to achieve that. i've had good experience evaluating "understand for c++" from www.scitools.com. this is a commercial product, which is comparable (but i think faster and slightly better supported then sourcenav). then, there is the plain-ol combinations of ctags and emacs/vim/textpad whatever else that people use. ctags allows for easy cross-referencing. in any case, in my own, fairly inexperienced :) opinion, i found that a collection of good tools are generally more convenient then one big app (like sniff). hth, denis ps. sorry, martin. i am not against sniff+ or anything ... just suggesting some options :) joewoodbury@mail.com (Joe Woodbury) wrote in news:c8c13723.0106211316.585460b5@posting.google.com: > The first RTOS I used years ago was Phar Lap ETS. It was adequate, but > lacked drivers. However, for what we needed (ticket vending machines) > it worked just fine. > > The device I'm working on now was prototyped with pSOS which is > showing it's age. It is missing key features and poorly implemented > others. Some of this was rectified with the progressive 2.x releases, > but it was clearly a case of too little too late. You save only a > little over VxWorks and the support will get progressively worse as > those knowledgable in pSOS go onto other things. > > We also evaluated Windows CE and Embedix (Lineo's embedded Linux > offering) to appease management. The biggest problem with both were > board support packages, which CE completely lacked and for which Lineo > couldn't give us a straight answer. Lineo was also very coy about > their [incident] support package costs, but we found out they were > very high. Red Hat's support package for eCos "start as low as > $15,000"--platinum supports starts at $40,000. Oooh,I can feel the > savings now [deep sarcasm.] (eCOS claims to support our board, but > there is a disclaimer on the web page that "...no proper device > drivers are present in this port..." Other than that, eCOS is > promising.) > > In the end, we chose VxWorks because a) it works with out hardware > almost out-of-the-box, b) it's small and designed for embedded > systems--it's not an afterthought forced into the role like CE or > Linux, c) It's a genuine RTOS, which CE and Linux aren't, d) it is the > industry standard, e) they are going to be around for a while, unless, > of course, they totally blow it, g) it has some very nice tools (i.e. > WindView), h) the OS itself is solid, i) we've gotten good development > support so far from them, at least from our local office (they may be > unhelpful with tools, but are very helpful in other areas.) > > Yes, I've been concerned about their general support and their > inability to articulate the differences between products. But > switching to a competitor would likely be a case of jumping from the > frying pan into the fire. Face it, support in the traditional (i.e. > UNIX) developer community sucks. (Microsoft may have other problems, > but providing developers with nice tools and lots of resources for a > very low price generally isn't one of them (except, ironically, in > their embedded offerings.)) > > > > > "Matthias Guelck" wrote in message > news:<9gs7nq$gdr$1@news.hdpp.de>... >> Hello Joe, >> >> all WRS FAEs told me, that I am the only person in the world >> complaining about WRS products. Again it was very important for me to >> see, that I am NOT alone! (I was close to it to believe) Do you (or >> others) know about a forum like “complaints about WRS”? >> >> You mentioned that “VxWorks is the best RTOS we've evaluated and/or >> used”. Which other RTOS have you evaluated and/or used? >> I am asking this, because we are currently searching for an >> alternative RTOS like ecos, embedded Linux or something other with >> open source. Are you going to stay with WRS products in spite of all >> your dissatisfaction? >> >> Best regards >> >> Matthias > --------------------------- Newsgroups: comp.os.vxworks Subject: Rational ClearCase with Tornado & VxWorks? Date: Fri, 29 Jun 2001 22:14:29 +0100 From: molochai Message-ID: Anybody using clearcase for maanging VxWorks projects? Amy experiences you'd like yo share? --------------------------- Newsgroups: comp.os.vxworks Subject: Re: SSH on VxWorks Date: Fri, 29 Jun 2001 22:18:17 +0100 From: molochai Message-ID: <196667924a%molochai@vapour-trail.demon.co.uk> References: <7c4037a9.0106281550.732ff4a5@posting.google.com> In message <7c4037a9.0106281550.732ff4a5@posting.google.com> tapankarwa@yahoo.com (Tapan Karwa) wrote: OpenSSL ought to port quite easily. Not done so myself though. > Hi, > > Does anyone know if SSH has been ported onto VxWorks. This question > has been asked before but there were very few answers. Any pointers > will be really appreciated. > > Thanks, > Tapan. --------------------------- Newsgroups: comp.os.vxworks Subject: blocking ip addresses Date: Fri, 29 Jun 2001 17:29:29 -0600 From: Les Hammer Organization: Agilent Technologies Message-ID: <3B3D0F59.62AE7E49@agilent.com> Hi, I'm creating a vxWorks device that will be on the LAN. Actually, there will be several of them on the LAN. What I want to do is to set up some sort of security system to prevent George in the next cubicle from mistakenly accessing my device. (He doesn't type in IP addresses very well and types mine by mistake.) What's the best way to set my device to not accept anything from his computer's IP address? I've found the ipFilter functions. It looks like that is a first level pass at IP traffic. Can I use it to examine incoming traffic, and if the ip address isn't on my list of users just return TRUE (claiming I've handled the packet, don't do normal processing)? If it's on my list, return FALSE and let the normal process take place. Is this the best way to handle this, or is there another way? Thanks, Les Hammer Agilent Technologies --------------------------- Newsgroups: comp.os.vxworks Subject: Re: How to get rid of const warning Date: Fri, 29 Jun 2001 21:08:32 -0400 From: "Kevin Livingston" Message-ID: References: <9hid31$br0$1@overload.lbl.gov> "Ben Kelly" wrote in message news:9hid31$br0$1@overload.lbl.gov... > You should only do this when you do not have the option of fixing the underlying library to advertise const usage. That > is assuming of course that the library does treat it like a const; if the library actually modifies the value I believe > the compiler is free to do whatever it likes at this point (crash, reformat your hard drive, etc). Just to agree and reaffirm... The C++ Programming Language 3rd Ed. Bjarne Stroustrup Chapter 10 p232 in reference to const_cast: "This is hardly elegant, and is not guaranteed to work when applied to an object that was originally declared as a const." "...is not guaranteed to give a single predictable result on all implementations." his [Stroustrup] comments were mostly directed to something like const my_class my_variable_name; I don't know how more or less applicable to something like the return value from const return_type* function(parameters); but I imagine the same holds. I had to look that up for myself the other day. I imagine you could test it and if it seems to work, then you are probably ok for that compilier... but it doesn't sound like the best thing to count on. (certainly if you want your code to port well... ok for some reason nothing ports well but at least better) Kevin --------------------------- Newsgroups: comp.os.vxworks Subject: FW: Network pool running out... Date: Fri, 29 Jun 2001 16:53:20 +0100 From: "David Anderson" Organization: Lawrence Berkeley National Laboratory Message-ID: <9hjg80$oqs$1@overload.lbl.gov> - -----Original Message----- From: David Anderson [SMTP:david.anderson@dtrack.com] Sent: 27 June 2001 16:05 To: 'Vxworks_Users (E-mail) Subject: Network pool running out... vxWorks 5.3.1 Tornado 1.0.1 Hello all, I've run into a problem which produces an apparent 'freeze up' of a PPP link, but still allows 'normal' Ethernet style TCP/IP connections to proceed. We are using the Tornado 2.0 ppp daemon. It appears that running an FTP session and a TELNET session concurrently over the PPP link causes the network pool to run out of buffers ( error 0x73000a ). However, the PPP link does not recover, the daemon tasks are running but you can no longer ping from either end to the other, or establish a new socket connection across it. The buffer error appears only to be a soft error - an Ethernet TCP/IP connection can still be set-up OK. Any ideas about why the PPP link should be 'frozen out' although all status indications suggest it is open and up. David Anderson. NEW E-MAIL ADDRESS: mailto:david.anderson@dtrack.com - ------------------------------------- FF: 2B + ~2B , that is the question ? - ------------------------------------- Disclaimer - ---------- This email and any files transmitted with it are confidential and intended solely for the use of the individual or entity to whom they are addressed. If you have received this email in error please destroy and notify Data Track Technology Plc +44 1425 271900. - ------------------------------------------------------- --------------------------- Newsgroups: comp.os.vxworks Subject: RE : comp.os.vxworks newsdigest Date: Sat, 30 Jun 2001 09:29:54 +0530 From: Nayan Kumar Garg Organization: Lawrence Berkeley National Laboratory Message-ID: <9hjjof$q1v$1@overload.lbl.gov> Hi Paul , Thanks for the reply . I am using the same to download the vxWorks image through serial cable .Actually I am not using the network .So I want to download the vxWorks image through a serial cable .For this I am using TSFS .I have made all the changes in config.h which are required for that . But I am not able to download the image through the serial port . Whenever I boot the target machine I gets a message " Waiting for Target Server Connection " ..and on the Host Machine in the target server launching Log Console I get a message "Error: clnttty_rcv: read timed out " .And unlimately the target machine gets waiting for the VxWorks image to be downloaded .I have already checked the serial cable , it is working fine . So can you please suggest me some way to do this . Thanks & Regards nayan Newsgroups: comp.os.vxworks Subject: Re: How to download vxWorks Image through serial port Date: Thu, 28 Jun 2001 13:07:26 +0100 From: Paul Caswell Organization: Land & Sea Systems Message-ID: <3B3B1DFE.65DFC239@baesystems.com> References: <9hesiv$1m0$1@overload.lbl.gov> I'm assuming that you don't have a network connection and want to download your vxWorks image over the serial link. As I understand it there are a few ways to do this the simplest being to use the TSFS (Target Server File System). See Page 136 of the Tornado 2.0 User Guide. Hope this helps Paul --------------------------- Newsgroups: comp.os.vxworks Subject: vxworks device, tcp port 513 open Date: Sat, 30 Jun 2001 06:19:57 GMT From: "fejed" Organization: @Home Network Message-ID: Hi, I have some hardware running VxWorks Version: 5.3, it has port 513 open, (rlogind) but on connect I get nothing with a rlogin client or if I telnet to it, advice anyone? fejed --------------------------- Newsgroups: comp.os.vxworks Subject: How can I block VxWorks default shell (->)? Date: Sat, 30 Jun 2001 15:54:10 +0900 From: "±èÀÀ½Ä" Organization: Korea Telecom Message-ID: <9hjt5a$ee2$1@news1.kornet.net> As far as I know windriver don't show the source of ``telnet'' and ``shell''. However I am looking forward to enter my own shell command interpreter when I connect target throught telnet. I've run the vxworks shell in background but my all application didn't work well. Because they run properly under tShell task only. plz, someone help me... ====================================== Eung-Shik Kim (Ed) ed@ebridgecom.com PowerVoIP Gateway Hardware Devl. Dept. 467-12 Dogok-Dong Kangnam-Gu MMAA Bldg 29th Fl. Seoul (zip 135-270) KOREA, Republic of Tel+82 2 575 6476 ext 517 Fax+82 2 575 6478 C.P. +82 11 9704 0840 ICQ#61658172 Send me plain text please... ======================================= --------------------------- Newsgroups: comp.os.vxworks Subject: why signal sample source code generates exception (VxWorks 5.4 for PPC) Date: Sat, 30 Jun 2001 15:57:34 +0800 From: "lhp" Organization: Lawrence Berkeley National Laboratory Message-ID: <9hk1qe$36e$1@overload.lbl.gov> HI, I use Tornado II and VxWorks 5.4 for PPC. Why the following signal sample source code generates exception, no matter running in VxSim or running on an actual MPC860 board? HELP! my email: LHP@gpte.com.cn /* includes */ #include "vxWorks.h" #include "signal.h" /* forward declarations */ static void mysigHandle (int); void receivSig(); void sendSig(); static int rtaskid; void mysignal() { rtaskid = taskSpawn("receivSig",100,0,4000,receivSig,0,0,0,0,0,0,0,0,0,0); } void receivSig() { int n; logMsg("enter rec task.\n"); if(signal(SIGUSR1,mysigHandle)!=0) { logMsg("sigaction failed.\n"); } taskSpawn("sendSig",90,0,4000,sendSig,0,0,0,0,0,0,0,0,0,0); /*wait somewhile then exit*/ n=10; while(n-->0) { logMsg("exitting rec task...\n"); } } void sendSig() { int id; logMsg("enter send task.\n"); if(kill(rtaskid, SIGUSR1) != 0) { logMsg("send signal failed.\n"); } logMsg("exit send task.\n"); } static void mysigHandle (int signo) { logMsg("got the signal!\n"); } --------------------------- End of New-News digest **********************