From daemon@vxw.ee.lbl.gov Sat Apr 1 04:00:30 1995 From: daemon@vxw.ee.lbl.gov Date: Sat Apr 1 04:00:38 PST 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sat Apr 1 04:00:20 PST 1995 Subject: HELP for novice Subject: Mailbox Interrupts on force 3ce? Subject: yield the cpu? Subject: US-HI Real-time & GUI programmer/analyst positions Subject: Re: yield the cpu? ------------------------------------------------------- Newsgroups: comp.sys.m68k,comp.os.vxworks Subject: HELP for novice Date: 31 Mar 1995 07:01:00 GMT From: lfini@lfiniarcetri.astro.it (Luca Fini) Organization: Osservatorio Astrofisico di Arcetri Message-ID: <3lg9bc$dof@sisifo.arcetri.astro.it> Sender: lfini@lfini (Luca Fini) I've got a Motorola MVME167 single board computer for the VME bus running VxWorks. I've also got a CAEN V 260 16 channel scaler which is stetd to be addressable as an A24 module (but the same problem arises with all the A24 modules). My question is: how can I address the board without conflicting with the upper part of the installed 16 Mb of memory? The VME manual specifies that I can set an "address modifier" to switch between short I/O (16 bit), standard (24 bit), extended (32 bit) and even long (64 bit) addresses, but I don't know how either the MVME167 architecture or a C program running under VxWorks can do that. BTW I understand from the MVME167 manual that the Short I/O adresses are mapped in a slice of the full 32 bit addresses (i.e. at $FFF00000-$FFFEFFFF) so I think that when addressing something within that range I actually do a Short I/O addressing mode, but there isn't anything similar for the 24 bit standard mode. I'd appreciate very much any hints on the subject (maybe by mail because I'm not very sure I always get all the messages from the News system) Cheers, \\|// +----------------------------------------+ |O-O| | Luca Fini +----------oOO-(_)-OOo----------+ | Osservatorio Astrofisico di Arcetri | | L.go E.Fermi, 5 Tel. +39 55 2752 307 | | 50125 Firenze Fax.: +39 55 2752 292 | | Italia e-mail: lfini@arcetri.astro.it | +------------------------------------------------------------------------+ --------------------------- Newsgroups: comp.os.vxworks Subject: Mailbox Interrupts on force 3ce? Date: 31 Mar 1995 23:00:12 GMT From: najam@biostat.wisc.edu (Haroon Najam) Organization: Department of Biostatistics, University of Wisconsin-Madison Keywords: 3ce, Mailbox Message-ID: <3li1hs$eae@news.doit.wisc.edu> Hello Everybody, I am trying to use Mailbox Interrupts available on the SPARC based cpu-3ce boards from Force. I would really appreciate if anybody can share some suggestions/advice/code fragments with me on this topic. An outline of how to go about this might get get me started. Thanks a lot. - -Haroon Najam najam@biostat.wisc.edu --------------------------- Newsgroups: comp.os.vxworks Subject: yield the cpu? Date: 31 Mar 1995 17:02:21 GMT From: rtm@cabernet.harvard.edu (Robert Morris) Organization: Division of Applied Sciences, Harvard University Message-ID: Followup-To: comp.os.vxworks Is there any way to yield the CPU in VxWorks without giving it up for a tick? I want to give up the CPU to any other processes that want to run, but get the CPU back immediately when they are done. taskDelay(0) seems to do nothing at all. taskDelay(1) always makes my task pause a tick. Thanks, Robert Morris rtm@das.harvard.edu --------------------------- Newsgroups: misc.jobs.contract,misc.jobs.offered,comp.os.vxworks Subject: US-HI Real-time & GUI programmer/analyst positions Date: Fri, 31 Mar 1995 19:27:06 GMT From: cruise@orion.cfht.hawaii.edu (Bill Cruise) Organization: Canada-France-Hawaii Telescope Keywords: vxworks EPICS real-time DVXDesigner GUI Message-ID: Sender: news@news.hawaii.edu - -- Real-time control system programmer/analyst Graphical user interface programmer/analyst The Canada-France-Hawaii Telescope, which operates a 3.6 metre optical and infrared telescope on Mauna Kea, Hawaii, has immediate openings for programmer/analysts to work in a project to replace the existing telescope control system. These are temporary positions with a maximum duration of two years. Duties include developing both real time device level control code, and high level graphical user interfaces. The control system will be divided into a real-time part and a user interface part. The real-time part will utilize a VME based system running vxWorks and EPICS. The user interface system will be on a UNIX workstation and utilize MOTIF, DVXDesigner, DataViews, and a CFHT proprietary GUI. Minimum requirements: BS in computer science, or in physics or electrical engineering with an emphasis on computer science work; 4 years experience in software design and implementation; strong C or C++ language skills; 3 years UNIX and X-Windows programming experience for GUI pos; 2 years experience in a real-time, high performance software system directly controlling electromechanical hardware for RT pos; good written and verbal communication skills. Desirable skills: vxWorks, EPICS, DVXDesigner, DataViews, servo systems, astronomical observatory operations. Depending on the mix of programming skills available, project work may not be divided on the RT/GUI lines. A combination of real time and GUI skills would be a plus. Most of the project work will be done in our offices, but occasional trips to the observatory, which is at an altitude of 4200 metres, will be required. Salary will be commensurate with qualifications. To apply please send a detailed resume and three professional references to tcs_programmer@cfht.hawaii.edu, or to the U.S. Mail address or FAX number below. Direct any inquiries to cruise@cfht.hawaii.edu. Please, no telephone inquiries. Applications will be accepted until 31 May 1995. Mail applications to: TCS Programmer Position Canada-France-Hawaii Telescope P.O. Box 1597 Kamuela, HI 96743 FAX (808) 885-7288 --------------------------- Newsgroups: comp.os.vxworks Subject: Re: yield the cpu? Date: 31 Mar 1995 20:34:12 GMT From: friedl@cher.heurikon.com (Ted Friedl) Organization: Heurikon Corporation Message-ID: <3lhp04$gl@zebu.heurikon.com> References: In article , Robert Morris wrote: >Is there any way to yield the CPU in VxWorks without giving it up for >a tick? I want to give up the CPU to any other processes that want to >run, but get the CPU back immediately when they are done. taskDelay(0) >seems to do nothing at all. taskDelay(1) always makes my task pause a >tick. Use taskPrioritySet() and set your task's priority below that of the tasks you want to run. Bump it back to your regular priority when your task resumes. Ted Friedl Software Engineer Heurion Corporation --------------------------- End of New-News digest ********************** From AS_Bridge@msmail3.hac.com Sat Apr 1 09:18:54 1995 From: "AS Bridge" Date: Sat Apr 1 09:19:00 PST 1995 Subject: Undeliverable Mail Unknown Microsoft mail form. Approximate representation follows. Message: SCSI tape drive Sent: Fri, Mar 31, 1995 9:02 AM To: Claudio, Michael A On Server: 631 Segment4 Date: Sat, Apr 1, 1995 9:18 AM Reason: Could not be delivered because the destination Microsoft Mail server could not be found. From AS_Bridge@msmail3.hac.com Sat Apr 1 10:51:53 1995 From: "AS Bridge" Date: Sat Apr 1 10:52:05 PST 1995 Subject: Undeliverable Mail Unknown Microsoft mail form. Approximate representation follows. Message: Re: Industry Pack Drivers (IP-Serial) Sent: Fri, Mar 31, 1995 10:26 AM To: Claudio, Michael A On Server: 631 Segment4 Date: Sat, Apr 1, 1995 10:51 AM Reason: Could not be delivered because the destination Microsoft Mail server could not be found. From AS_Bridge@msmail3.hac.com Sat Apr 1 10:51:54 1995 From: "AS Bridge" Date: Sat Apr 1 10:52:06 PST 1995 Subject: Undeliverable Mail Unknown Microsoft mail form. Approximate representation follows. Message: Re: SCSI Tape Driver Sent: Fri, Mar 31, 1995 10:26 AM To: Claudio, Michael A On Server: 631 Segment4 Date: Sat, Apr 1, 1995 10:51 AM Reason: Could not be delivered because the destination Microsoft Mail server could not be found. From AS_Bridge@msmail3.hac.com Sat Apr 1 11:11:57 1995 From: "AS Bridge" Date: Sat Apr 1 11:12:03 PST 1995 Subject: Undeliverable Mail Unknown Microsoft mail form. Approximate representation follows. Message: Re: PCMCIA drivers for VxWorks. Sent: Fri, Mar 31, 1995 11:03 AM To: Claudio, Michael A On Server: 631 Segment4 Date: Sat, Apr 1, 1995 11:11 AM Reason: Could not be delivered because the destination Microsoft Mail server could not be found. From AS_Bridge@msmail3.hac.com Sat Apr 1 11:43:07 1995 From: "AS Bridge" Date: Sat Apr 1 11:43:15 PST 1995 Subject: Undeliverable Mail Unknown Microsoft mail form. Approximate representation follows. Message: Little vs. Big Endian Processors Sent: Fri, Mar 31, 1995 11:35 AM To: Claudio, Michael A On Server: 631 Segment4 Date: Sat, Apr 1, 1995 11:42 AM Reason: Could not be delivered because the destination Microsoft Mail server could not be found. From AS_Bridge@msmail3.hac.com Sat Apr 1 12:13:08 1995 From: "AS Bridge" Date: Sat Apr 1 12:13:22 PST 1995 Subject: Undeliverable Mail Unknown Microsoft mail form. Approximate representation follows. Message: Re: Industry Pack Drivers and Mot MVIP301 in particular Sent: Fri, Mar 31, 1995 11:59 AM To: Claudio, Michael A On Server: 631 Segment4 Date: Sat, Apr 1, 1995 12:13 PM Reason: Could not be delivered because the destination Microsoft Mail server could not be found. From AS_Bridge@msmail3.hac.com Sat Apr 1 12:13:08 1995 From: "AS Bridge" Date: Sat Apr 1 12:13:22 PST 1995 Subject: Undeliverable Mail Unknown Microsoft mail form. Approximate representation follows. Message: Re: Industry Pack Drivers (IP-Serial) Sent: Fri, Mar 31, 1995 12:01 PM To: Claudio, Michael A On Server: 631 Segment4 Date: Sat, Apr 1, 1995 12:13 PM Reason: Could not be delivered because the destination Microsoft Mail server could not be found. From AS_Bridge@msmail3.hac.com Sat Apr 1 14:26:14 1995 From: "AS Bridge" Date: Sat Apr 1 14:26:21 PST 1995 Subject: Undeliverable Mail Unknown Microsoft mail form. Approximate representation follows. Message: Looking for Flash File System for VxWorks Sent: Fri, Mar 31, 1995 2:16 PM To: Claudio, Michael A On Server: 631 Segment4 Date: Sat, Apr 1, 1995 2:26 PM Reason: Could not be delivered because the destination Microsoft Mail server could not be found. From AS_Bridge@msmail3.hac.com Sat Apr 1 16:07:15 1995 From: "AS Bridge" Date: Sat Apr 1 16:07:22 PST 1995 Subject: Undeliverable Mail Unknown Microsoft mail form. Approximate representation follows. Message: RE: Industry Pack Drivers and Mot MVIP301 in particular Sent: Fri, Mar 31, 1995 4:01 PM To: Claudio, Michael A On Server: 631 Segment4 Date: Sat, Apr 1, 1995 4:06 PM Reason: Could not be delivered because the destination Microsoft Mail server could not be found. From AS_Bridge@msmail3.hac.com Sat Apr 1 16:27:20 1995 From: "AS Bridge" Date: Sat Apr 1 16:27:29 PST 1995 Subject: Undeliverable Mail Unknown Microsoft mail form. Approximate representation follows. Message: Re: Looking for Flash File System for VxWorks Sent: Fri, Mar 31, 1995 4:24 PM To: Claudio, Michael A On Server: 631 Segment4 Date: Sat, Apr 1, 1995 4:26 PM Reason: Could not be delivered because the destination Microsoft Mail server could not be found. From leonid@rst.co.il Sat Apr 1 23:47:57 1995 From: leonid@rst.co.il (Leonid Rosenboim) Date: Sat Apr 1 23:48:05 PST 1995 Subject: Re: SCSI tape drive McHale wrote on Mar 31, 1995: ... My problem is that I need to get updated vxWorks images from the SUN host to the SCSI drive (without the ethernet to the host). >From what I understand, vxWorks does not provide utilities for tapes. What I would really like is something similar to a UNIX "tar" so that I can make my tapes on the SUN using tar, and then copy the files off the tape onto the SCSI hard disk. ... Our "SCSI Tape Driver for VxWorks" package includes a set of functions that read and write tapes in UNIX-tar format, and it has been tested to do just that, among other things. We also now include an application note woth this package that explains how to make boot-roms that can actually boot and initialize a virgin system from a SCSI tape. Our software was tested with mv167 and Archive Viper, among other hardware platforms. ----------------------------------------------------------------------- Leonid Rosenboim Phone: +972-3-559-8144 R S T Software Industries Ltd. Mobile: +972-50-307-142 3 Hashikma st. Industrial Zone, Fax: +972-3-559-8244 P.O.Box 11502, AZUR 58017, Israel E-Mail: leonid@rst.co.il From AS_Bridge@msmail3.hac.com Sun Apr 2 01:25:04 1995 From: "AS Bridge" Date: Sun Apr 2 01:25:12 PST 1995 Subject: Undeliverable Mail Unknown Microsoft mail form. Approximate representation follows. Message: Monthly VxWorks archive posting Sent: Sat, Apr 1, 1995 12:59 AM To: Claudio, Michael A On Server: 631 Segment4 Date: Sun, Apr 2, 1995 1:24 AM Reason: Could not be delivered because the destination Microsoft Mail server could not be found. From daemon@vxw.ee.lbl.gov Sun Apr 2 04:00:32 1995 From: daemon@vxw.ee.lbl.gov Date: Sun Apr 2 04:00:40 PDT 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sun Apr 2 04:00:18 PDT 1995 Subject: Re: Looking for Flash File System for VxWorks ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Looking for Flash File System for VxWorks Date: 1 Apr 1995 23:30:09 GMT From: D'Anne Thompson Organization: National Optical Astronomy Observatories, Tucson, AZ, USA Message-ID: <3lknm2$bvt@noao.edu> References: <2F7C8104@gicpo.gic.gi.com> "Moore, Rich (HT-MS)" wrote: > > > Hello all, > > Does anybody know if I can run the VxWorks RamDisk system over a flash > device? > My target is Motorola MVME162, which has 1 MBytes of Flash Memory. > > Thanks in advance, > Rich Moore > General Instrument, Communications Division > Hatboro, PA, USA Hi Rich, You can use the 'ramDrv' to read data from a flash device, but you can't use it to 'write' to the flash. What you can do easily is this. 0. I assume you have routines to write data to flash memory. 1. Use 'ramDrv' to create a new dosFs system in ram. Write your data to this disk image. 2. When satisfied with the disk image created above, use your flash writing routines to copy the ram data to flash memory. (Flush the ram device first). 3. Now you can create a new 'ramDrv' device that points to the new updated flash memory image. At Synergy we had to create a complete flash memory driver to support our new DEFL/DELF options that can provide from 2 to 16 MB of flash memory on board. You can't use the scheme above when you've got 16MB of flash and only 4MB of ram, in some configurations. This scheme only works when you've got enough ram to hold a complete image of the flash memory in ram. D'Anne Thompson Synergy Microsystems dee@ioinc.tucson.az.us dat85715@indirect.com --------------------------- End of New-News digest ********************** From AS_Bridge@msmail3.hac.com Sun Apr 2 09:16:11 1995 From: "AS Bridge" Date: Sun Apr 2 09:16:18 PDT 1995 Subject: Undeliverable Mail Unknown Microsoft mail form. Approximate representation follows. Message: comp.os.vxworks newsdigest Sent: Sat, Apr 1, 1995 5:50 AM To: Claudio, Michael A On Server: 631 Segment4 Date: Sun, Apr 2, 1995 9:15 AM Reason: Could not be delivered because the destination Microsoft Mail server could not be found. From AS_Bridge@msmail3.hac.com Sun Apr 2 11:17:32 1995 From: "AS Bridge" Date: Sun Apr 2 11:17:39 PDT 1995 Subject: Undeliverable Mail Unknown Microsoft mail form. Approximate representation follows. Message: Undeliverable Mail Sent: Sat, Apr 1, 1995 11:10 AM To: Claudio, Michael A On Server: 631 Segment4 Date: Sun, Apr 2, 1995 11:16 AM Reason: Could not be delivered because the destination Microsoft Mail server could not be found. From AS_Bridge@msmail3.hac.com Sun Apr 2 13:00:31 1995 From: "AS Bridge" Date: Sun Apr 2 13:00:43 PDT 1995 Subject: Undeliverable Mail Unknown Microsoft mail form. Approximate representation follows. Message: Undeliverable Mail Sent: Sat, Apr 1, 1995 12:38 PM To: Claudio, Michael A On Server: 631 Segment4 Date: Sun, Apr 2, 1995 12:59 PM Reason: Could not be delivered because the destination Microsoft Mail server could not be found. From AS_Bridge@msmail3.hac.com Sun Apr 2 13:00:33 1995 From: "AS Bridge" Date: Sun Apr 2 13:00:46 PDT 1995 Subject: Undeliverable Mail Unknown Microsoft mail form. Approximate representation follows. Message: Undeliverable Mail Sent: Sat, Apr 1, 1995 12:44 PM To: Claudio, Michael A On Server: 631 Segment4 Date: Sun, Apr 2, 1995 12:59 PM Reason: Could not be delivered because the destination Microsoft Mail server could not be found. From AS_Bridge@msmail3.hac.com Sun Apr 2 13:29:34 1995 From: "AS Bridge" Date: Sun Apr 2 13:29:44 PDT 1995 Subject: Undeliverable Mail Unknown Microsoft mail form. Approximate representation follows. Message: Undeliverable Mail Sent: Sat, Apr 1, 1995 1:24 PM To: Claudio, Michael A On Server: 631 Segment4 Date: Sun, Apr 2, 1995 1:28 PM Reason: Could not be delivered because the destination Microsoft Mail server could not be found. From AS_Bridge@msmail3.hac.com Sun Apr 2 13:29:34 1995 From: "AS Bridge" Date: Sun Apr 2 13:29:52 PDT 1995 Subject: Undeliverable Mail Unknown Microsoft mail form. Approximate representation follows. Message: Undeliverable Mail Sent: Sat, Apr 1, 1995 1:04 PM To: Claudio, Michael A On Server: 631 Segment4 Date: Sun, Apr 2, 1995 1:28 PM Reason: Could not be delivered because the destination Microsoft Mail server could not be found. From AS_Bridge@msmail3.hac.com Sun Apr 2 13:59:34 1995 From: "AS Bridge" Date: Sun Apr 2 13:59:47 PDT 1995 Subject: Undeliverable Mail Unknown Microsoft mail form. Approximate representation follows. Message: Undeliverable Mail Sent: Sat, Apr 1, 1995 1:54 PM To: Claudio, Michael A On Server: 631 Segment4 Date: Sun, Apr 2, 1995 1:59 PM Reason: Could not be delivered because the destination Microsoft Mail server could not be found. From daemon@vxw.ee.lbl.gov Mon Apr 3 04:00:28 1995 From: daemon@vxw.ee.lbl.gov Date: Mon Apr 3 04:00:37 PDT 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Mon Apr 3 04:00:17 PDT 1995 Subject: taskLock()/taskUnlock() question Subject: VxWorks UDP performance Subject: Is there a group/mailing list covering Vertex? Subject: A short review of tgdb for vxworks ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: taskLock()/taskUnlock() question Date: 2 Apr 1995 12:29:05 GMT From: michaelv@qi.qualcomm.com (Michael Vakulenko) Organization: Qualcomm Israel Message-ID: Hi, VxWorks maintains task switch locking on per task basis. When task, which has issued taskLock(), blocks, task switching is enabled again. My question is following: let's say, that task, in it's forever loop, calls read() from socket function in each circle, and the socket has always data for task to read. Before entering forever loop, task has issued taskLock() command. Will other tasks, running at the same priority, be able to get their time slice ? In other words, could task, locked task switching, be preempted inside read() or write() functions, inspite of that system will have enough resource to perform read() or write() without blocking? Thanx in advance, Michael Vakulenko, Software Engineer. Qualcomm Israel Ltd., Haifa, Israel. E-mail: michaelv@qualcomm.com --------------------------- Newsgroups: comp.os.vxworks Subject: VxWorks UDP performance Date: 2 Apr 1995 12:32:26 GMT From: michaelv@qi.qualcomm.com (Michael Vakulenko) Organization: Qualcomm Israel Message-ID: Hi, I was tried to configure our VxWorks system to max UDP perfromace with short packets (test description follows). I hope, that some more experienced VxWorks people could advice me, how to improve VxWorks UDP performance. I've tested VxWorks performance in terms of max UDP packet rate, that VxWorks application is able process without packet loss. Special "Echo" program was written. The Echo receives data at particular UDP port ( recvfrom()) and echoes the data packet back to the sender( sendto()). The program introduces minimum overhead and runs as a VxWorks task. It does nothing with data, just sends it back. In the following experiments, Echo was the only task running in addition to standard VxWorks tasks. NCC LAN analyzer was used to generate UDP packets at given rate and send them to VxWorks. After playing around with several VxWorks parameters, I've got following results: Echo was able to return 100% of UDP packets at rates: UDP packet size 70 bytes - 1450 packets/second, this corresponds to throughput 99.1 K/second; UDP packet size 162 bytes - 1100 packets/second, this correspondes to throughput 174 K/second. When two concurrent Echo tasks were listening on different UDP ports, both of them were able to return 100% of UDP packets 162 bytes long at rate 970 packets/second (throughput 153.4 K/second). In this tests I was running VxWorks 5.1.1 on MOTOROLA MVME162LX VME card with M68040 CPU and 16M RAM and Intel 82596 Ethernet chip. The best results I've got with following VxWorks parameters: o Ethernet interface driver was configured to 300 transmit and 300 receive buffers. o Initially allocated mbufs number was set to 200. VxWorks never needed to allocate more buffers. I've checked this with "mbufShow" command. o Echo task priority greater than tNetTask priority. o tShell task run at priority lower (80) than tNetTask & Echo tasks and was idle. Has anybody managed to get better results ? Thanx in advance, Michael Vakulenko, Software Engineer. Qualcomm Israel Ltd., Haifa, Israel. E-mail: michaelv@qualcomm.com --------------------------- Newsgroups: comp.os.vxworks Subject: Is there a group/mailing list covering Vertex? Date: Wed, 29 Mar 1995 22:53:27 GMT From: news@demon.co.uk (Usenet Administration) Organization: Quantrill Computer Limited Message-ID: Subject line says it all really. Product is by Ready Systems (or as I prefer to call them, Ethelred Systems ;^) Replies/flames by email please.... | Adam Quantrill | ,-| | E-mail: adam@maxwells.demon.co.uk | | Cheap modems | | | | S-nail: QCL, 61 Akeman St, Cambridge, | | for sale in UK | `-| | U-hail: +44 223 323409 | CB4 3HE, UK | | V34 150 quid | `- | Modem sales: modems@maxwells.demon.co.uk| --------------------------- Newsgroups: comp.os.vxworks Subject: A short review of tgdb for vxworks Date: 02 Apr 1995 20:18:43 GMT From: duhrenho@casey.sanders.com (David Uhrenholdt) Organization: Lockheed Sanders Message-ID: When I posted a request about using tgdb with VxWorks I got two responses. Neither responder had used tgdb - one thought it sounded like a good idea to try it out, the other requested a list of features that might recommend it over xxgdb. I have since tried to use and I will provide a short review. 1) tgdb is built from: tcl7.3 tk3.6 (or tk3.6pl1) tclX7.3a (or tclX7.3b) BLT1.3 (or newer) TkSteal3.6c (or newer) expect5.3 (or newer) The following tgdb_wish (the interpreter) binaries are available: sunsite.unc.edu:/pub/Linux/devel/debuggers/tgdb_wish-linux.* ftp.aud.alcatel.com:/tcl/code/tgdb_wish-sunos4.1.3.* ftp.aud.alcatel.com:/tcl/code/tgdb-HPUX-bin.* ftp.aud.alcatel.com:/tcl/code/tgdb_wish-AIX32RS6k.* The tcl source is available from: sunsite.unc.edu:/pub/Linux/devel/debuggers/ tgdb is shareware and there is a small fee per user to use it. I tried it out with sunos4.1.4 as a host and VxWorks 5.1.? 68040 as target. I was able to build gdb 4.14 without much trouble. (tgdb requires 4.10 or higher) The gdb in my release of VxWorks is 4.6 2) List of features. 1) Like xxgdb it has a source frame (scrollable, etc), a button frame, and a command frame all on the main window. It also has a separate menu bar. Unlike xxgdb I wasn't able to figure out how to change the relative size of source and command windows. It also has a program output window - useless as far as I can see for remote VxWorks debugging. 2) It has an assembly language which you can single step in. 3) It has a tgdb help window and a gdb help window (compiled the first time you use gdb for a target). Each of these can be more helpful to a new user than the traditional gdb help. 4) It has a few other scrollable window such as mem dump, display window, and register window. 5) The way to select the a C structure is different than in xxgdb: A single click selects a simple expression, a second click selects the next component of the structure, etc. Also holding down the mouse works as expected. 6) There are many key bindings available for short cuts. 7) It has an easier to use command recall than xxgdb and also has command line completion. 8) There are many other small features that I won't go over, but a lot of thought went in to marking it more user friendly than gdb. 3) My impressions 1) tgdb give much the same functionality as you get with xxgdb, plus some important new features (not for debuggers in general, but in the gdb world they are new). 2) It seems to be a beta version, I keep getting pop-ups saying this or that internal error occurred and asking me if I would like an internal stack dump. Also the tgdb help window will frequently just start resizing its self making getting help difficult. The register display doesn't work. - etc. 3) I couldn't figure out how to get a custom menu up without added the capability to get a new init file that is loaded in a different stage of initialization. I like the ability to write my own tcl routines, I wish it where easier to incorporate then. 4) It is much slower than xxgdb. I am not an experienced tcl user so I am may be missing something, but this is a serious drawback. 5) It seems to have inherited a weird input focus policy from Tk. Once I display the register window I am prevented from using any other window until I dismiss the register window. There may and should be a way around this but it isn't easy to figure out. Dave - -- Dave Uhrenholdt duhrenho@vette.sanders.lockheed.com Phone (603) 885-7398 --------------------------- End of New-News digest ********************** From schultzr@source.asset.com Mon Apr 3 06:02:30 1995 From: schultzr@source.asset.com (Richard Schultz) Date: Mon Apr 3 06:02:37 PDT 1995 Subject: Re: PCMCIA drivers for VxWorks. If you find any info on VxWorks type PCMCIA device drivers, please let me know. Thanks, Richard. From AS_Bridge@msmail3.hac.com Mon Apr 3 06:42:31 1995 From: "AS Bridge" Date: Mon Apr 3 06:42:41 PDT 1995 Subject: Undeliverable Mail Unknown Microsoft mail form. Approximate representation follows. Message: Undeliverable Mail Sent: Sat, Apr 1, 1995 4:25 PM To: Claudio, Michael A On Server: 631 Segment4 Date: Sun, Apr 2, 1995 8:25 PM Reason: Could not be delivered because the destination Microsoft Mail server could not be found. From AS_Bridge@msmail3.hac.com Mon Apr 3 06:42:36 1995 From: "AS Bridge" Date: Mon Apr 3 06:42:58 PDT 1995 Subject: Undeliverable Mail Unknown Microsoft mail form. Approximate representation follows. Message: Undeliverable Mail Sent: Sat, Apr 1, 1995 6:05 PM To: Claudio, Michael A On Server: 631 Segment4 Date: Sun, Apr 2, 1995 8:25 PM Reason: Could not be delivered because the destination Microsoft Mail server could not be found. From AS_Bridge@msmail3.hac.com Mon Apr 3 06:42:38 1995 From: "AS Bridge" Date: Mon Apr 3 06:43:00 PDT 1995 Subject: Undeliverable Mail Unknown Microsoft mail form. Approximate representation follows. Message: Undeliverable Mail Sent: Sat, Apr 1, 1995 6:08 PM To: Claudio, Michael A On Server: 631 Segment4 Date: Sun, Apr 2, 1995 8:25 PM Reason: Could not be delivered because the destination Microsoft Mail server could not be found. From AS_Bridge@msmail3.hac.com Mon Apr 3 06:44:50 1995 From: "AS Bridge" Date: Mon Apr 3 06:44:59 PDT 1995 Subject: Undeliverable Mail Unknown Microsoft mail form. Approximate representation follows. Message: Re: SCSI tape drive Sent: Sun, Apr 2, 1995 1:46 AM To: Claudio, Michael A On Server: 631 Segment4 Date: Mon, Apr 3, 1995 2:08 AM Reason: Could not be delivered because the destination Microsoft Mail server could not be found. From AS_Bridge@msmail3.hac.com Mon Apr 3 06:44:57 1995 From: "AS Bridge" Date: Mon Apr 3 06:45:14 PDT 1995 Subject: Undeliverable Mail Unknown Microsoft mail form. Approximate representation follows. Message: Undeliverable Mail Sent: Sun, Apr 2, 1995 3:34 AM To: Claudio, Michael A On Server: 631 Segment4 Date: Mon, Apr 3, 1995 3:40 AM Reason: Could not be delivered because the destination Microsoft Mail server could not be found. From AS_Bridge@msmail3.hac.com Mon Apr 3 06:45:24 1995 From: "AS Bridge" Date: Mon Apr 3 06:45:34 PDT 1995 Subject: Undeliverable Mail Unknown Microsoft mail form. Approximate representation follows. Message: comp.os.vxworks newsdigest Sent: Sun, Apr 2, 1995 6:22 AM To: Claudio, Michael A On Server: 631 Segment4 Date: Mon, Apr 3, 1995 6:39 AM Reason: Could not be delivered because the destination Microsoft Mail server could not be found. From michaud@alaxzo.cern.ch Mon Apr 3 09:05:05 1995 From: michaud@alaxzo.cern.ch (Dominique Michaud div. PPE/ALD CERN-Geneve) Date: Mon Apr 3 09:05:14 PDT 1995 Subject: VXWORKS and TCP/IP Hi , I'm running VxWorks 4.2 and have it communicate using TCP/ip through a single ethernet cable with a SunOs 4.1.1b station. Sometimes, like every few thousand Megabytes transmitted from the vxworks card to the Sun, it looks like the sender does not take into account the acknoledgement sent by the unix card, though the unix card odes send acknoledgement (Has anybody heard about this problem ? Is it solved in a next version of Vxworks, or might it be due to a misuse of sockets ? Thanks for any information on the subject, D. Michaud Cern, Geneva Switzerland PS: of course after retrying a few times the vxworks station close the connection... From jpc@ssd.ray.com Mon Apr 3 10:35:36 1995 From: Carissimi Date: Mon Apr 3 10:35:42 PDT 1995 Subject: TCP Sockets and select I have an MV167 with V5.1.1 running as a TCP server to a number of TCP clients. The server is using the select() facility to determine when data is available on any of the sockets. Most of the time, all works well. But in some instances, a client sends a message to the server and the server receives the message in parts (segment). The case that I see often is that a client send a message of 0xe30 bytes, but the server wakes up and read 0xc00 bytes and then the next time read 0x230 bytes. On the other side (client), the send returns the indication that 0xe30 bytes were sent. My question is...when is select() activated? When a segment of data (1K in this instance) is available on the socket, or when the entire message, as sent by the client, is reassembled? Is there any way to not have the server wake up until the entire message is receive?? Or do I, in my TCP input driver have to accommodate for segmentation? Thanks in advance... Joe Carissimi Raytheon Electronic Systems Division 1847 West Main Road, Mail Stop 177 Portsmouth, RI 02871-1087 Email: jpc@ssd.ray.com From aisie.aisinc.com!aisinc.com!pew@msen.com Mon Apr 3 11:20:26 1995 From: pew@aisinc.com (Paul E. Wilt) Date: Mon Apr 3 11:20:34 PDT 1995 Subject: Re: Little vs. Big Endian Processors Mike Anderson (mea@mclean.sparta.com) writes: > > Greetings! > > I've been trying to fill in some matrices for the FAQ and I need some information. > What is the endian-ness of the following processors as used by VxWorks? > ... Stuff deleted ... > > Any others? > Mike: Has anyone mentioned the Intel i960 to you? The i960 by default is a little- endian machine. However, it contains 16 memory configuration registers to allow you to specify the endianess of each 256-Mb chunk of the 4 Gb address space. At least two of the boards I know of support byte swapping data coming out of the VME address space. Hope this is of use to you! I hope that we will see the new FAQ soon! (;-) Thanks, Paul =========================================================================== Paul Wilt (pew@aisinc.com) Applied Intelligent Systems, Inc. 110 Parkland Plaza Ann Arbor, MI 48103 (313) 995-2035 (x276) From AS_Bridge@msmail3.hac.com Mon Apr 3 11:33:22 1995 From: "AS Bridge" Date: Mon Apr 3 11:33:28 PDT 1995 Subject: Undeliverable Mail Unknown Microsoft mail form. Approximate representation follows. Message: Undeliverable Mail Sent: Sun, Apr 2, 1995 11:26 AM To: Claudio, Michael A On Server: 631 Segment4 Date: Mon, Apr 3, 1995 11:32 AM Reason: Could not be delivered because the destination Microsoft Mail server could not be found. From AS_Bridge@msmail3.hac.com Mon Apr 3 13:37:47 1995 From: "AS Bridge" Date: Mon Apr 3 13:37:54 PDT 1995 Subject: Undeliverable Mail Unknown Microsoft mail form. Approximate representation follows. Message: Undeliverable Mail Sent: Sun, Apr 2, 1995 1:27 PM To: Claudio, Michael A On Server: 631 Segment4 Date: Mon, Apr 3, 1995 1:37 PM Reason: Could not be delivered because the destination Microsoft Mail server could not be found. From hebo@mbari.org Mon Apr 3 13:53:47 1995 From: Bob Herlien Date: Mon Apr 3 13:53:55 PDT 1995 Subject: Re: TCP Sockets and select > My question is...when is select() activated? When a segment of data > (1K in this instance) is available on the socket, or when the entire > message, as sent by the client, is reassembled? Is there any way to > not have the server wake up until the entire message is receive?? Or > do I, in my TCP input driver have to accommodate for segmentation? The latter. The TCP socket interface is a reliable *stream oriented* interface. The receiving side is guaranteed to get exactly the same byte stream, in order, as was sent. There is no guarantee about *when* those bytes arrive (within timeout constraints). There are no message boundaries inherent in sockets. If you need a message oriented interface using sockets, you need to build it. This is independent of whether you're using select(). Fortunately, building such a message interface is almost trivial. You simply put a length field in the header of your application message, and call read() until you have the whole message. Network gurus -- please let me know (gently) if I'm wrong. -------------------------------------------------------------- Bob Herlien MBARI (Monterey Bay Aquarium Research Institute) hebo@mbari.org From isys!bt@uunet.uu.net Mon Apr 3 15:13:18 1995 From: isys!bt@uunet.uu.net (Bret Tilson) Date: Mon Apr 3 15:13:24 PDT 1995 Subject: Re: Undeliverable Mail HELP!! From guthrie@power.amasd.anatcp.rockwell.com Mon Apr 3 15:45:28 1995 From: guthrie@power.amasd.anatcp.rockwell.com (Bob Guthrie) Date: Mon Apr 3 15:45:38 PDT 1995 Subject: D64 Block mode on the MVME166 ---------- X-Sun-Data-Type: text X-Sun-Data-Description: text X-Sun-Data-Name: text X-Sun-Content-Lines: 15 To all: With the program include below I was experiencing similar problems as reported earlier by Jo Schambach (JSchamba@physics.utexas.edu) where the VME address was being broadcast much more often then expected and slowing down the transfer rates. I noticed this when the MVME166 as the VME master and doing writes to a micro memory board. I did not notice this when the MVME166 was the VME master reading the micro memory board. I got the following transfer rates: MVME166 master,write to MicroMemory board 37 Mbytes/sec MVME166 master,read from MicroMemory board 76 Mbytes/sec Actually, the D32 block mode writes are as fast as D64? ---------- X-Sun-Data-Type: c-file X-Sun-Data-Description: c-file X-Sun-Data-Name: master_block_test.c X-Sun-Content-Lines: 209 #include "vxWorks.h" #include "string.h" #include "stdlib.h" #include "stdio.h" #include "vme.h" #include "mv167.h" #include "vmechip2.h" #include "taskLib.h" #include "vmLib.h" #define NBYTES_TEST 0x100000 #define DONE_MASK (ICR_MLBE| \ ICR_MLPE| \ ICR_MLOB| \ ICR_MLTO| \ ICR_DLBE| \ ICR_DLPE| \ ICR_DLOB| \ ICR_DLTO| \ ICR_TBL| \ ICR_VME| \ ICR_DONE) #define TYPE_OF_BLOCK DMACR2_D64_BLOCK #define TEST_ADDRESS_MODIFIER 0xD #define VME_ADDRESS 0x18000000 void master_block_test(option) int option; { int i; u_long *input_buffer; u_long *output_buffer; u_long input_checksum; u_long output_checksum; u_long temp_status; u_long temp_control_word; u_short *short_ptr; u_long mmu_page_size; u_long nbytes; mmu_page_size = vmBasePageSizeGet(); nbytes = (NBYTES_TEST+mmu_page_size-1)/mmu_page_size; nbytes *= mmu_page_size; input_buffer = (u_long *)valloc(NBYTES_TEST); vmBaseStateSet(NULL, input_buffer, nbytes, VM_STATE_MASK_CACHEABLE, VM_STATE_CACHEABLE_NOT); printf("address of input_buffer %08X (hex)\n",(u_long)input_buffer); output_buffer = (u_long *)valloc(NBYTES_TEST); vmBaseStateSet(NULL, output_buffer, nbytes, VM_STATE_MASK_CACHEABLE, VM_STATE_CACHEABLE_NOT); printf("address of output_buffer %08X (hex)\n",(u_long)output_buffer); bzero((char *)input_buffer,NBYTES_TEST); bzero((char *)output_buffer,NBYTES_TEST); switch(option) { case 0: bfill((char *)output_buffer,NBYTES_TEST,0); break; case 1: bfill((char *)output_buffer,NBYTES_TEST,0xff); break; case 2: bfill((char *)output_buffer,NBYTES_TEST,0xaa); break; case 3: bfill((char *)output_buffer,NBYTES_TEST,0x55); break; case 4: default: short_ptr = (u_short *)output_buffer; for (i=0;i<(NBYTES_TEST/2);i++) short_ptr[i] = rand(); break; } output_checksum = 0; for (i=0;i<(NBYTES_TEST/4);i++) output_checksum += output_buffer[i]; temp_control_word = (DMACR2_SNOOP_INHIBIT| DMACR2_VINC| DMACR2_LINC| DMACR2_TO_VME| DMACR2_D32| TYPE_OF_BLOCK| TEST_ADDRESS_MODIFIER); printf("DMACR2 set to %08X\n",temp_control_word); *VMECHIP2_DMACR2 = temp_control_word; temp_control_word = (TIMEOUTCR_OFF_0US| TIMEOUTCR_ON_DONE| TIMEOUTCR_VGTO_DISABLE| TIMEOUTCR_VATO_DISABLE| TIMEOUTCR_LBTO_DISABLE| TIMEOUTCR_WDTO_64S| 0xDF); printf("TIMEOUTCR set to %08X\n",temp_control_word); *VMECHIP2_TIMEOUTCR = temp_control_word; *VMECHIP2_DMACLBAC = (u_long)output_buffer; *VMECHIP2_DMACVAC = (u_long)VME_ADDRESS; *VMECHIP2_DMACBC = NBYTES_TEST; printf("ICR set to %08X\n",ICR_MCLR); *VMECHIP2_ICR = ICR_MCLR; temp_control_word = (DMACR1_ROM0_OFF| DMACR1_SNOOP_INHIBIT| DMACR1_165NS| DMACR1_PRIORITY| DMACR1_LVRWD| DMACR1_LVREQ_L3| DMACR1_DEN| DMACR1_BRX| DMACR1_DREQ_L3); printf("DMACR1 set to %08X\n",temp_control_word); *VMECHIP2_DMACR1 = temp_control_word; temp_status = 0; while (1) { temp_status = *VMECHIP2_ICR & DONE_MASK; if (temp_status) break; } printf("*VMECHIP2_ICR after 166 -> VME %08X (hex)\n", temp_status); taskDelay(60); temp_control_word = (DMACR2_SNOOP_INHIBIT| DMACR2_VINC| DMACR2_LINC| DMACR2_TO_LOCAL| DMACR2_D32| TYPE_OF_BLOCK| TEST_ADDRESS_MODIFIER); printf("DMACR2 set to %08X\n",temp_control_word); *VMECHIP2_DMACR2 = temp_control_word; temp_control_word = (TIMEOUTCR_OFF_0US| TIMEOUTCR_ON_DONE| TIMEOUTCR_VGTO_DISABLE| TIMEOUTCR_VATO_DISABLE| TIMEOUTCR_LBTO_DISABLE| TIMEOUTCR_WDTO_64S| 0xDF); printf("TIMEOUTCR set to %08X\n",temp_control_word); *VMECHIP2_TIMEOUTCR = temp_control_word; *VMECHIP2_DMACLBAC = (u_long)input_buffer; *VMECHIP2_DMACVAC = (u_long)VME_ADDRESS; *VMECHIP2_DMACBC = NBYTES_TEST; printf("ICR set to %08X\n",ICR_MCLR); *VMECHIP2_ICR = ICR_MCLR; temp_control_word = (DMACR1_ROM0_OFF| DMACR1_SNOOP_INHIBIT| DMACR1_165NS| DMACR1_PRIORITY| DMACR1_LVRWD| DMACR1_LVREQ_L3| DMACR1_DEN| DMACR1_BRX| DMACR1_DREQ_L3); printf("DMACR1 set to %08X\n",temp_control_word); *VMECHIP2_DMACR1 = temp_control_word; temp_status = 0; while (1) { temp_status = *VMECHIP2_ICR & DONE_MASK; if (temp_status) break; } printf("*VMECHIP2_ICR after VME -> 166 %08X (hex)\n", temp_status); input_checksum = 0; for (i=0;i<(NBYTES_TEST/4);i++) input_checksum += input_buffer[i]; printf("output buffer checksum %08X (hex)\n",output_checksum); printf("input buffer checksum %08X (hex)\n",input_checksum); for (i=0;i<(NBYTES_TEST/4);i++) { if (input_buffer[i] != output_buffer[i]) { printf("mismatch long word # %08X (hex)\n", i); printf("input_buffer %08X\n", input_buffer[i]); printf("output_buffer %08X\n", output_buffer[i]); break; } } free(input_buffer); free(output_buffer); } From daemon@vxw.ee.lbl.gov Tue Apr 4 04:00:28 1995 From: daemon@vxw.ee.lbl.gov Date: Tue Apr 4 04:00:37 PDT 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Tue Apr 4 04:00:19 PDT 1995 Subject: VxWorks dying in semQPut... Subject: Re: Little vs. Big Endian Processors Subject: Re: VxWorks dying in semQPut... Subject: Update to closing date ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: VxWorks dying in semQPut... Date: 3 Apr 1995 21:31:23 GMT From: Charles Shannon Hendrix Organization: InfiNet Message-ID: <3lppfb$q3q@lucy.infi.net> I have a simple network program that is basically a network data reflector. It reads data from a host, writes it to a RAM location on an mv162 board running VxWorks, reads some variables from the same block, and sends that back to a UNIX host computer. I'm using UDP. The program runs for awhile... usually 30 minuts to 2 hours... and then just stops with no error messages or anything. Tracing the task from the shell shows this: 29778 _vxTaskEntry +10 : _main_init (311554, 311540, 31152c, 7e4, 5, 0, 0, 0, 0, 0) 312294 _main_init +ec : _main_loop (5, 6, 30b5ac, 3057ec, 5, 5, 5) 3136ea _main_loop +196: _va_read_poll (5, 30b5ac, 30193c, 5) 312a70 _va_read_poll +78 : _send_recv_data (1, 3018ba, 6) 3127ea _send_recv_data+a6 : _recv_data (10, 301870) 312ce4 _recv_data +54 : _semQPut (20, 301808, 0, 0, 301810) value = 0 = 0x0 ..so it looks to me like it's stuck in a kernel/lib call. I need ideas on why this might happen because my code looks OK and it runs fine until it gets hung in semQPut. Here is a quick pseudocode so you'll have some idea what this program's main loop does. main_loop(...) { set up interval timer... usually 5-10hz while(1) { get specified list of variables from a UNIX host via UDP write that data to 0x300000 plus some offset read specified variables from 0x300000 send that data to UNIX host via UDP pause(); } } The interval timer kicks me out of the pause() at the end of the loop. I have no idea why this program would run for awhile and die. I think the time it runs is variable but I am still really testing that (have to share this board with another programmer). Now, I do have another problem which may or may not be related so I will mention it here. I cannot get the interval timers to work properly. When I set up the timers so I'll get 10hz (100ms interval) I actually get 15hz (around 67ms). An mv162 at another site runs correctly at 10hz. I don't know if that is related to failing semaphores in the network call or not. I sincerely hope nothing is wrong with our hardware. This could be a kernel problem or an improperly configured kernel. Any ideas greatly appreciated. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Little vs. Big Endian Processors Date: Sun, 2 Apr 1995 23:56:55 GMT From: ronk@PROBLEM_WITH_INEWS_GATEWAY_FILE (Ron Kellam) Organization: Hewlett-Packard Message-ID: References: <9503311711.AA07438@locutus> Sender: news@hparc12.aus.hp.com (News Daemon ID) Mike Anderson (mea@mclean.sparta.com) wrote: : Greetings! : I've been trying to fill in some matrices for the FAQ and I need some information. : What is the endian-ness of the following processors as used by VxWorks? : DEC Alpha : Little endian? DEC is typically Little for their stuff : R4600 : Big endian? SGI uses them as little endian, I think. : Intel 386/486/586 : Little? : PowerPC : Little or Big? I've heard both. : Mot 88K : Big? : RS 6000 : Little? (Related to PowerPC) : HP PA-Risc : Big? : Mot 68HC16 : Big? : Any others? Intel i960 : Little endian - although the i960 can support a mixture of either in different memory regions. Regards, Ron.K. (who is looking forward to ESC East) - ---------------------------------------------------------------------------- Ron Kellam | Phone: +613 272 4016 Hewlett Packard | Fax: +613 898 9257 .-_!\ Australian Telecom Operation | Email: ronk@hpato.aus.hp.com / ATO \ 31-41 Joseph St; Blackburn | HPDesk: KELLAM_RON/HP9061_RG@hpausa1 \_.-._/ Victoria, Australia, 3130 | "My opinions are my own, not HP's" v - ---------------------------------------------------------------------------- "Life is like the weather - what comes, comes. Even with a sunny forecast you can still get wet." - ---------------------------------------------------------------------------- --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks dying in semQPut... Date: Mon, 3 Apr 1995 23:40:43 GMT From: jfinley@netcom.com (John Finley) Organization: NETCOM On-line Communication Services (408 261-4700 guest) Message-ID: References: <3lppfb$q3q@lucy.infi.net> Sender: jfinley@netcom22.netcom.com Charles Shannon Hendrix (shendrix@taurustech.com) wrote: : I have a simple network program that is basically a network data : reflector. It reads data from a host, writes it to a RAM location : on an mv162 board running VxWorks, reads some variables from the : same block, and sends that back to a UNIX host computer. : I'm using UDP. : The program runs for awhile... usually 30 minuts to 2 hours... and : then just stops with no error messages or anything. Tracing the : task from the shell shows this: (edited) Things that go bang when you're not looking include: * Stack overflows: Perhaps you could do a checkStack every 100th call or something, to see if you're getting close. * Memory corruption: Are you using data from the network to compute the offset where you poke? If so, be sure to validate the offset. UDP is probably not checksumming your packets, and especially if you are going across a serial link, you're likely to eventually get an error. If it's all local Ethernet, then you shouldn't get an error since the Enet hardware does a checksum. * The infamous 68k NMI: If you have any devices on interrupt level 7, and their drivers may invoke kernel services, then you will eventually crash since the kernel can't protect itself. From your task trace, your task looks OK, like it's patiently waiting for something from the network. Are any other tasks looking sick? Good luck, John Finley --------------------------- Newsgroups: comp.os.vxworks Subject: Update to closing date Date: Tue, 4 Apr 1995 02:06:40 GMT From: cruise@orion.cfht.hawaii.edu (Bill Cruise) Organization: Canada-France-Hawaii Telescope Corp Keywords: real-time GUI programmer EPICS Message-ID: Followup-To: US-HI Real-time & GUI programmer/analyst positions Sender: news@news.hawaii.edu - -- NOTE: this is a repost In the previous post the closing date was wrong. It is actually 30 Apr, not 31 May. Real-time control system programmer/analyst Graphical user interface programmer/analyst The Canada-France-Hawaii Telescope, which operates a 3.6 metre optical and infrared telescope on Mauna Kea, Hawaii, has immediate openings for programmer/analysts to work in a project to replace the existing telescope control system. These are temporary positions with a maximum duration of two years. Duties include developing both real time device level control code, and high level graphical user interfaces. The control system will be divided into a real-time part and a user interface part. The real-time part will utilize a VME based system running vxWorks and EPICS. The user interface system will be on a UNIX workstation and utilize MOTIF, DVXDesigner, DataViews, and a CFHT proprietary GUI. Minimum requirements: BS in computer science, or in physics or electrical engineering with an emphasis on computer science work; 4 years experience in software design and implementation; strong C or C++ language skills; 3 years UNIX and X-Windows programming experience for GUI pos; 2 years experience in a real-time, high performance software system directly controlling electromechanical hardware for RT pos; good written and verbal communication skills. Desirable skills: vxWorks, EPICS, DVXDesigner, DataViews, servo systems, astronomical observatory operations. Depending on the mix of programming skills available, project work may not be divided on the RT/GUI lines. A combination of real time and GUI skills would be a plus. Most of the project work will be done in our offices, but occasional trips to the observatory, which is at an altitude of 4200 metres, will be required. Salary will be commensurate with qualifications. To apply please send a detailed resume and three professional references to tcs_programmer@cfht.hawaii.edu, or to the U.S. Mail address or FAX number below. Direct any inquiries to cruise@cfht.hawaii.edu. Please, no telephone inquiries. Applications will be accepted until 30 Apr 1995. Mail applications to: TCS Programmer Position Canada-France-Hawaii Telescope P.O. Box 1597 Kamuela, HI 96743 FAX (808) 885-7288 --------------------------- End of New-News digest ********************** From @mail.uunet.ca:corcoran@array Tue Apr 4 09:14:57 1995 From: corcoran@array.ca (Rob Corcoran) Date: Tue Apr 4 09:15:05 PDT 1995 Subject: MVME167 boards Can anyone suggest a location where I can find used or short-term leased MVME167 boards (1-2)? We're undertaking a VadsWorks prototyping effort, and the intended target is a PowerPC card, but the VxWorks port is not slated to be available until mid-summer. The '167 is our choice for interim (2-3 months) efforts. Thanks, Rob Corcoran corcoran@array.ca Array Systems Computing Inc. From jtosey@glenayre.com Wed Apr 5 00:41:00 1995 From: jtosey@glenayre.com (Joseph Tosey [4572]) Date: Wed Apr 5 00:41:07 PDT 1995 Subject: IP/360 drivers Does anyone know of an archive for VxWorks or pSOS+ IP/68360 drivers? We're looking for any of the following: ethernet, frame relay, ISDN 2B+D (with PPP/ML), etc. I didn't see anything relevent at ncar, and Motorola doesn't seem to provide os-specific drivers in their European archive (why not, I wonder?). From daemon@vxw.ee.lbl.gov Wed Apr 5 04:00:26 1995 From: daemon@vxw.ee.lbl.gov Date: Wed Apr 5 04:00:36 PDT 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Wed Apr 5 04:00:18 PDT 1995 Subject: Re: VxWorks dying in semQPut... Subject: Re: VxWorks dying in semQPut... Subject: Re: VxWorks dying in semQPut... Subject: Re: VxWorks dying in semQPut... Subject: Re: VxWorks dying in semQPut... ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks dying in semQPut... Date: Tue, 04 Apr 1995 04:35:55 GMT From: jfhughes@jtechcons.win.net (John F. Hughes) Message-ID: <195@jtechcons.win.net> References: <3lppfb$q3q@lucy.infi.net> Reply-To: jfhughes@jtechcons.win.net (John F. Hughes) In article <3lppfb$q3q@lucy.infi.net>, Charles Shannon Hendrix (shendrix@taurustech.com) writes: > >I have a simple network program that is basically a network data >reflector. It reads data from a host, writes it to a RAM location >on an mv162 board running VxWorks, reads some variables from the >same block, and sends that back to a UNIX host computer. > >I'm using UDP. > >The program runs for awhile... usually 30 minuts to 2 hours... and >then just stops with no error messages or anything. Tracing the >task from the shell shows this: > > 29778 _vxTaskEntry +10 : _main_init (311554, 311540, 31152c, 7e4, 5, 0, 0, 0, > 0, 0) >312294 _main_init +ec : _main_loop (5, 6, 30b5ac, 3057ec, 5, 5, 5) >3136ea _main_loop +196: _va_read_poll (5, 30b5ac, 30193c, 5) >312a70 _va_read_poll +78 : _send_recv_data (1, 3018ba, 6) >3127ea _send_recv_data+a6 : _recv_data (10, 301870) >312ce4 _recv_data +54 : _semQPut (20, 301808, 0, 0, 301810) >value = 0 = 0x0 > >..so it looks to me like it's stuck in a kernel/lib call. > I have been having a similar problem, using TCP/IP streams. Everything runs great for a variable amount of time (we have run anywhere from a few seconds to 5 hours), then one or more of the tasks that are doing the communication stop in semQPut. I have been tearing my hair out for months trying to figure out what is wrong with our code. We just started pointing the finger at the board vendor, but maybe it is a vxWorks problem? In our case, we frequently start seeing input errors on the Ethernet interface (using ifShow "ln") when the communications stops working. I sent email to WRS (TSR 21870) asking what the errors mean and what causes the counter to increment. If anyone out there has a clue what can be causing this, we could really use the help. regards, jfh --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks dying in semQPut... Date: 4 Apr 1995 17:40:30 GMT From: Charles Shannon Hendrix Organization: InfiNet Message-ID: <3ls0ae$eeg@lucy.infi.net> References: <3lppfb$q3q@lucy.infi.net> jfinley@netcom.com (John Finley) wrote: > > Charles Shannon Hendrix (shendrix@taurustech.com) wrote: > Things that go bang when you're not looking include: > > * Stack overflows: Perhaps you could do a checkStack every 100th call > or something, to see if you're getting close. This is in a while loop and I don't believe anything is going on the stack. checkStack looks OK to me when it's hung and while it's running. > * Memory corruption: Are you using data from the network to compute > the offset where you poke? If so, be sure to validate the offset. > UDP is probably not checksumming your packets, and especially if you > are going across a serial link, you're likely to eventually get an > error. If it's all local Ethernet, then you shouldn't get an error > since the Enet hardware does a checksum. This is ethernet but I guess I could do a little sanity check on the offsets... however, it's hanging during a read before the data is written to an offset. > * The infamous 68k NMI: If you have any devices on interrupt level 7, > and their drivers may invoke kernel services, then you will > eventually crash since the kernel can't protect itself. Nasty... hope it's nothing like that. This system is nothing but a 162 board, two VMIC cards (1110 and a 2112), and a custom designed carrier board. None of the cards are driver by interrupt level that I am aware of... I don't know if the MV162 uses that or not. I would hope Wind River systems would have taken care of that. > From your task trace, your task looks OK, like it's patiently waiting > for something from the network. Are any other tasks looking sick? Nope... this is the only task running outside of the normal vxworks tasks. Would any of them interfere in some way? I have the remote debugger, NFS, and the timer lib compiled in. > Good luck, > John Finley --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks dying in semQPut... Date: 4 Apr 1995 08:35:25 -0700 From: mwette@mr-ed.jpl.nasa.gov (Matt Wette) Organization: Jet Propulsion Laboratory Message-ID: <3lrovt$e44@mr-ed.jpl.nasa.gov> References: <3lppfb$q3q@lucy.infi.net> <195@jtechcons.win.net> In article <195@jtechcons.win.net>, John F. Hughes wrote: > >In article <3lppfb$q3q@lucy.infi.net>, Charles Shannon Hendrix (shendrix@taurustech.com) writes: >> >>I have a simple network program that is basically a network data >>reflector. It reads data from a host, writes it to a RAM location >>on an mv162 board running VxWorks, reads some variables from the >>same block, and sends that back to a UNIX host computer. >> >>I'm using UDP. >> >>The program runs for awhile... usually 30 minuts to 2 hours... and >>then just stops with no error messages or anything. Tracing the >>task from the shell shows this: >> >> 29778 _vxTaskEntry +10 : _main_init (311554, 311540, 31152c, 7e4, 5, 0, 0, 0, >> 0, 0) >>312294 _main_init +ec : _main_loop (5, 6, 30b5ac, 3057ec, 5, 5, 5) >>3136ea _main_loop +196: _va_read_poll (5, 30b5ac, 30193c, 5) >>312a70 _va_read_poll +78 : _send_recv_data (1, 3018ba, 6) >>3127ea _send_recv_data+a6 : _recv_data (10, 301870) >>312ce4 _recv_data +54 : _semQPut (20, 301808, 0, 0, 301810) >>value = 0 = 0x0 >> >>..so it looks to me like it's stuck in a kernel/lib call. >> >I have been having a similar problem, using TCP/IP streams. >Everything runs great for a variable amount of time (we have run >anywhere from a few seconds to 5 hours), then one or more of the >tasks that are doing the communication stop in semQPut. I have >been tearing my hair out for months trying to figure out what is >wrong with our code. We just started pointing the finger at the >board vendor, but maybe it is a vxWorks problem? > >In our case, we frequently start seeing input errors on the >Ethernet interface (using ifShow "ln") when the communications >stops working. I sent email to WRS (TSR 21870) asking what the >errors mean and what causes the counter to increment. > >If anyone out there has a clue what can be causing this, we could >really use the help. > >regards, > jfh > > - -- matthew.r.wette@jpl.nasa.gov --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks dying in semQPut... Date: 4 Apr 1995 18:01:59 GMT From: Charles Shannon Hendrix Organization: InfiNet Message-ID: <3ls1in$eeg@lucy.infi.net> References: <3lppfb$q3q@lucy.infi.net> <195@jtechcons.win.net> jfhughes@jtechcons.win.net (John F. Hughes) wrote: > > > I have been having a similar problem, using TCP/IP streams. > Everything runs great for a variable amount of time (we have run > anywhere from a few seconds to 5 hours), then one or more of the > tasks that are doing the communication stop in semQPut. I have > been tearing my hair out for months trying to figure out what is > wrong with our code. We just started pointing the finger at the > board vendor, but maybe it is a vxWorks problem? I thought about it being the board but now I'm not so sure. I know our 162 has timer problems that another local company's 162 does not but even that could be a problem with vxworks. > In our case, we frequently start seeing input errors on the > Ethernet interface (using ifShow "ln") when the communications > stops working. I sent email to WRS (TSR 21870) asking what the > errors mean and what causes the counter to increment. I will have to start checking with ifShow() to see how things are during the run and after it fails. > If anyone out there has a clue what can be causing this, we could > really use the help. Please let us know what happens when tech support gets back to you on this. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks dying in semQPut... Date: 4 Apr 1995 09:09:30 -0700 From: mwette@mr-ed.jpl.nasa.gov (Matt Wette) Organization: Jet Propulsion Laboratory Message-ID: <3lrqvq$e61@mr-ed.jpl.nasa.gov> References: <3lppfb$q3q@lucy.infi.net> <195@jtechcons.win.net> <3lrovt$e44@mr-ed.jpl.nasa.gov> Let me try again ... We have seen similar problems. We are running TCP/IP between two HKV4Fs over the SM backplane (5.1.1). Our simulation will run for usually about 5 to 20 minutes and then just freezes. If I rememer right the guy last working on this problem said that he was able to get things rolling again by starting up spy. This has been a major headache for our project. Matt In article <195@jtechcons.win.net>, John F. Hughes wrote: > >In article <3lppfb$q3q@lucy.infi.net>, >.. Charles Shannon Hendrix (shendrix@taurustech.com) writes: >> >>I have a simple network program that is basically a network data >>reflector. It reads data from a host, writes it to a RAM location >>on an mv162 board running VxWorks, reads some variables from the >>same block, and sends that back to a UNIX host computer. >> >>I'm using UDP. >> >>The program runs for awhile... usually 30 minuts to 2 hours... and >>then just stops with no error messages or anything. Tracing the >>task from the shell shows this: >> >> 29778 _vxTaskEntry +10 : _main_init (311554, 311540, 31152c, 7e4, 5, 0, 0, 0, >> 0, 0) >>312294 _main_init +ec : _main_loop (5, 6, 30b5ac, 3057ec, 5, 5, 5) >>3136ea _main_loop +196: _va_read_poll (5, 30b5ac, 30193c, 5) >>312a70 _va_read_poll +78 : _send_recv_data (1, 3018ba, 6) >>3127ea _send_recv_data+a6 : _recv_data (10, 301870) >>312ce4 _recv_data +54 : _semQPut (20, 301808, 0, 0, 301810) >>value = 0 = 0x0 >> >>..so it looks to me like it's stuck in a kernel/lib call. >> >I have been having a similar problem, using TCP/IP streams. >Everything runs great for a variable amount of time (we have run >anywhere from a few seconds to 5 hours), then one or more of the >tasks that are doing the communication stop in semQPut. I have >been tearing my hair out for months trying to figure out what is >wrong with our code. We just started pointing the finger at the >board vendor, but maybe it is a vxWorks problem? > >In our case, we frequently start seeing input errors on the >Ethernet interface (using ifShow "ln") when the communications >stops working. I sent email to WRS (TSR 21870) asking what the >errors mean and what causes the counter to increment. > >If anyone out there has a clue what can be causing this, we could >really use the help. - -- matthew.r.wette@jpl.nasa.gov --------------------------- End of New-News digest ********************** From dillow@ataway.aptec.com Wed Apr 5 09:40:32 1995 From: dillow@ataway.aptec.com (Mike Dillow (staff)) Date: Wed Apr 5 09:40:38 PDT 1995 Subject: Job Opportunity Applied Technology Associates, a small Albuquerque, NM based R&D company, is looking for a Realtime Programmer who is experienced with VxWorks and VME. Experience with X11/Motif and TCP/IP is a plus. The new employee would be involved in a diverse range of interesting research projects. If you are interested, please E-mail or FAX your resume. Sorry, US Citizens Only. Send resumes to Mike Dillow Applied Technology Associates 1900 Randolph, SE Albuquerque, NM 87106 or E-mail to dillow@ataway.aptec.com or Fax to (505) 768-1379 Any questions can be sent via E-mail. ATA is an equal opportunity employer. From pardoe@ncp.gpt.co.uk Thu Apr 6 02:21:49 1995 From: "Matthew Pardoe" Date: Thu Apr 6 02:21:57 PDT 1995 Subject: SCSI disk configuration We have a vxWorks configuration where we boot up form SCSI disk. For production we need a way of quickly programming lots of disks with vxWorks images. Our production people have facilities that will copy PC SCSI disks, but when they try to copy a disk that has been configured via VxWorks it does not recognise the disk format correctly. My question is are there any constraints in configuring SCSI disks eg. block cluster size, No. of sectors etc that need to be conformed to to be able to take a disk out of a vxWorks system and put in a PC. =============================================================================== Matthew Pardoe _/_/ _/_/ _/ _/_/_/_/ email : pardoe@ncp.gpt.co.uk GPT Ltd. _/ _/ _/ _/ _/ _/ tel : +44 1202 782653 Sopers Lane _/ _/ _/ _/ _/ _/_/_/_/ fax : +44 1202 782163 Poole BH17 8BH _/ _/ _/_/_/_/ _/ England =============================================================================== From daemon@vxw.ee.lbl.gov Thu Apr 6 04:00:27 1995 From: daemon@vxw.ee.lbl.gov Date: Thu Apr 6 04:00:36 PDT 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Thu Apr 6 04:00:18 PDT 1995 Subject: VxGdb Subject: Re: MVME167 boards ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: VxGdb Date: 5 Apr 95 17:07:22 GMT From: Hans Peter Eberhard Message-ID: <1995Apr5.170722.14095@zippy.dct.ac.uk> Hi, is it possible to start the tRdbTask for remote debugging with VxGdb without rebuilding the kernel (inlude #define INLUDE_RDB in configAll.h and rebuilt it)? Thanks for any help Hans-Peter Eberhard email: hp.eberhard@unn.ac.uk --------------------------- Newsgroups: comp.os.vxworks Subject: Re: MVME167 boards Date: Wed, 5 Apr 1995 13:02:06 GMT From: clegg@ssd.fsi.com Organization: Flightsafety International, Broken Arrow, OK Message-ID: <1995Apr5.130206.72038@dev1.ssd.fsi.com> References: <9504032110.AA02320@array.ca> Reply-To: clegg@ssd.fsi.com Sender: news@dev1.ssd.fsi.com In <9504032110.AA02320@array.ca>, corcoran@array.ca (Rob Corcoran) writes: >Can anyone suggest a location where I can find used or short-term >leased MVME167 boards (1-2)? We're undertaking a VadsWorks >prototyping effort, and the intended target is a PowerPC card, >but the VxWorks port is not slated to be available until mid-summer. >The '167 is our choice for interim (2-3 months) efforts. >Rob Corcoran corcoran@array.ca >Array Systems Computing Inc. I recently asked WRS about the PowerPC port and they said that you could buy a "Developers Kit" but that you would have to buy a third party (non-GNU) compiler to develope your application. FYI R. Clegg clegg@ssd.fsi.com --------------------------- End of New-News digest ********************** From rpitts@scires.com Thu Apr 6 06:31:05 1995 From: rpitts@scires.com Date: Thu Apr 6 06:31:17 PDT 1995 Subject: Re: MVME167 boards >Can anyone suggest a location where I can find used or short-term >leased MVME167 boards (1-2)? We're undertaking a VadsWorks >prototyping effort, and the intended target is a PowerPC card, >but the VxWorks port is not slated to be available until mid-summer. >The '167 is our choice for interim (2-3 months) efforts. We get our demo boards from our distributor (Arrow Electronics) in the Atlanta area. Distributors often have demo boards available and if you are planning on buying a PowerPC and your distributor handles both boards, you might be able to get a MVME167 for 2-3 months. You may initially get a 1 month demo, but you can often extend the loaner period. Richard Pitts Senior Software Engineer Scientific Research Corporation 2300 Windy Ridge Parkway Suite 400 South Atlanta, Georgia 30339 (404)859-9161 (404)953-6241 FAX rpitts@scires.com From mea@mclean.sparta.com Thu Apr 6 07:05:37 1995 From: mea@mclean.sparta.com (Mike Anderson) Date: Thu Apr 6 07:05:44 PDT 1995 Subject: Re: SCSI disk configuration Greetings! > Submitted-by: "Matthew Pardoe" > > > We have a vxWorks configuration where we boot up form SCSI disk. For production > we need a way of quickly programming lots of disks with vxWorks images. > Our production people have facilities that will copy PC SCSI disks, but when > they try to copy a disk that has been configured via VxWorks it does not > recognise the disk format correctly. > > My question is are there any constraints in configuring SCSI disks eg. block > cluster size, No. of sectors etc that need to be conformed to to be able to > take a disk out of a vxWorks system and put in a PC. > > Well... this is problem that I have been investigating for the past few weeks. What I have found is that the MSDOS file system under VxWorks is a typical "floppy-disk" type of format. That is, under the "real" MSDOS filesystem, a floppy has a single master partition table/master boot record in sector 0 of the floppy. Sector 1 starts with data. That is not how MSDOS hard disks are formatted. In a hard disk, Sector 0 contains the master partition table which has pointers for up to 4 disk partitions. The disk partitions then start with master boot records, 2 FATs then data. The trick on the first sector (sector 0) is that typical MSDOS "wastes" the rest of the first *track* (for future bookkeeping purposes so sayeth MicroSoft) and the first partition actually starts at the first sector on track 1. This is definately *not* how the VxWorks version of the MSDOS filesystem works for hard disks. Therefore, you can't take a hard disk from a VxWorks system and simply use it in MSDOS (although you can take floppy disks and use them interchangeably). What I'm experimenting with now is a technique wherein the disk is formatted under MSDOS and I lay the partition form the VxWorks system down on the MSDOS disk where it's expected to be. This is obviously an off-line operation. I believe this incompatibility is a *major* annoyance that I hope will be remedied in the next release after VxWorks 5.2 (my guess is that it's too late in the development cycle to make it into 5.2). BTW, watch out for the definition of a megabyte. As computer types, we tend to define a megabyte as 1,048,576 bytes. But, disk vendors rate their drives using 1MB = 1,000,000 bytes. Therefore, a 2Gb disk has 2,000,000,000 bytes on it instead of 2,097,152,000 bytes as you might expect. 97 MBs here, 97 MBs there and sooner or later it starts to add up ;-). However, to simply duplicate disks for VxWorks, it is a fairly straightforward thing to do. First, format the target disk under VxWorks (e.g., use the FIOFORMAT ioctl call). This can be on the same or different SCSI daisychain as the source disk. Next, physically connect the target disk to the *same* SCSI daisychain as the original disk, but at a different SCSI target ID. Do the scsiPhysDevCreates, scsiBlkDevCreates on both drives. Do the dosFsDevInit on the source and a rawFsDevInit on the target. Open both drives in raw mode (e.g., fd0 = open("/sd0/", 2) ). Simply read from the source and write to the target block at a time until you hit the end of the source drive. Use buffer sizes in the 256K-1MB range for maximum efficiency. Close both fds and start again for the next target. For example: -> copyDisk("/sd0/", "/sd1/", 64, 0x2000000, 0, 0) would copy from "/sd0/" to "/sd1/" 64*COPY_WINDOW_SIZE bytes at a time (1MB) for 32 MBs starting at the beginning of each disk (source and destination starting locations are both 0). And Here's the Code: #define COPY_WINDOW_SIZE 16384 /* we'll copy in multiples of 16KB */ #define DISK_VENDOR_MEGABYTE 1000000 int copyDisk(source, destination, blockSize, totalSize, srcStartingLocation, destStartingLocation) char *source; char *destination; int blockSize; int totalSize; int srcStartingLocation; int destStartingLocation; { char *buffer; int srcFd, destFd, nBytes, nRdBytes, nextTarget, status; UINT32 srcFileOffset = 0; UINT32 destFileOffset = 0; UINT32 numBytesToWrite = 0; UINT32 numBytesWritten = 0; UINT32 tickStart, tickStop, elapsedTicks; float msecs; if ((srcFd=open(source, O_RDWR)) == ERROR) { printf("Could not open source file for reading!\n"); return(ERROR); } if ((destFd=open(destination, O_RDWR)) == ERROR) { printf("Could not open destination file for writing!\n"); return(ERROR); } /* set up buffer and fill it with something */ buffer = (char *) malloc(blockSize * COPY_WINDOW_SIZE); bfill(buffer, blockSize * COPY_WINDOW_SIZE, 0); numBytesToWrite = (UINT32) (blockSize * COPY_WINDOW_SIZE); /* get the time */ tickStart = tickGet(); nextTarget = 10; /* print a msg after 10 MBytes have been copied */ printf("\n"); /* get us to a new line */ if (totalSize <= 0) return(OK); destFileOffset = destStartingLocation; srcFileOffset = srcStartingLocation; status = ioctl(srcFd, FIOSEEK, srcFileOffset ); status = ioctl(destFd, FIOSEEK, destFileOffset ); while (numBytesWritten < totalSize ) { if ((totalSize - numBytesWritten) < numBytesToWrite) { numBytesToWrite = (totalSize - numBytesWritten); printf("\rLast read is %ld bytes ", numBytesToWrite); } /* read it */ if((nRdBytes = read(srcFd, buffer, numBytesToWrite)) < 1) { logMsg("Error reading from disk file!\n"); printErrno(errnoGet()); close(srcFd); close(destFd); taskDelay(5 * sysClkRateGet()); return(ERROR); } /* and write it */ if((nBytes = write(destFd, buffer, nRdBytes)) != nRdBytes) { logMsg("Error writing to disk file!\n"); printErrno(errnoGet()); close(srcFd); close(destFd); taskDelay(5 * sysClkRateGet()); return(ERROR); } /* adjust file pointer and number of bytes written */ srcFileOffset += numBytesToWrite; destFileOffset += numBytesToWrite; numBytesWritten += numBytesToWrite; if ((numBytesWritten / DISK_VENDOR_MEGABYTE) > nextTarget ) { printf("\r "); printf("\r%ld Mbytes Copied ", nextTarget); nextTarget += 10; } else { printf("."); } } /* get the time */ tickStop = tickGet(); elapsedTicks = tickStop - tickStart; /* express in millisecs */ msecs = (float) (elapsedTicks * (1.0 / sysClkRateGet()) * 1000.0); /* print results */ printf("\nTotal time : %10.2f secs\n", msecs / 1000.0); printf("Transfer rate: %10.2f Mbytes/sec\n", (float) ((numBytesWritten/msecs) / 1000.0)); close(destFd); close(srcFd); free(buffer); return(OK); } =========================== end sample code ================================== HTH, =============================================================================== __ Real-Time System Development, Integration, Training and Services //\\ // \\ Mike Anderson // /\ \\ Chief Engineer // / \ \\ SPARTA, Inc. Voice : (703) 448-0210 ext. 235 // \ \\ 7926 Jones Branch Drive FAX : (703) 734-3323 \\ \ // Suite 900 EMAIL : mea@mclean.sparta.com \\ \ / // McLean, VA 22102 \\ \/ // "Software development is like making \\ // a baby... You can't make a baby in one \\// month by impregnating nine women. -- "Pride in Performance" Some things just take time." =============================================================================== From carl@themis.com Thu Apr 6 08:49:51 1995 From: "Carl C. Chesbrough" Date: Thu Apr 6 08:49:59 PDT 1995 Subject: Re: SCSI disk configuration > We have a vxWorks configuration where we boot up form SCSI disk. For production > we need a way of quickly programming lots of disks with vxWorks images. > Our production people have facilities that will copy PC SCSI disks, but when > they try to copy a disk that has been configured via VxWorks it does not > recognise the disk format correctly. > > My question is are there any constraints in configuring SCSI disks eg. block > cluster size, No. of sectors etc that need to be conformed to to be able to > take a disk out of a vxWorks system and put in a PC. I have had problems with taking a hard disk formatted on a VxWorks system, and moving it to a PC. If you format a disk on a PC, then put it on the VxWorks system and copy your files, then you should not have a problem copying the disks. ******************************************************************** * * * Carl C. Chesbrough Telephone: (510) 252-0870 x 126 * * Themis Computer Facsimile: (510) 490-5529 * * 3185 Laurelview Court E-Mail : carl@themis.com * * Fremont, California 94538 * * * ******************************************************************** From km@spacenet.lanl.gov Thu Apr 6 14:17:21 1995 From: km@spacenet.lanl.gov (Kim Mcculla) Date: Thu Apr 6 14:17:28 PDT 1995 Subject: PCI meeting in Atlanta? Does anyone have information on a PCI meeting in Atlanta on April 17-18? I seem to have misplaced this information. Thanks! -Kim ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Kimberly McCulla E-mail: km@spacenet.lanl.gov Space Data Systems, NIS-3 PO Box 1663, MS-D440 Phone: (505) 667-3295 Los Alamos National Laboratory FAX: (505) 665-4197 Los Alamos, NM 87545 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ From daemon@vxw.ee.lbl.gov Fri Apr 7 04:00:29 1995 From: daemon@vxw.ee.lbl.gov Date: Fri Apr 7 04:00:40 PDT 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Fri Apr 7 04:00:19 PDT 1995 Subject: Re: VxWorks dying in semQPut... Subject: Digital and Wind River Systems Announcement Subject: Mupltiple writes to a socket Subject: A24 Addressing for MVME-167 Board with 16MBytes local RAM ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks dying in semQPut... Date: 6 Apr 1995 14:09:39 GMT From: billag@b4pphff (Bill Gutknecht) Organization: Bell-Northern Research Message-ID: <3m0sn3$6is@nrtphba6.bnr.ca> References: <3lppfb$q3q@lucy.infi.net> I've seen the same thing happen using TCP/IP. Task is hung on a semQPut() ... at first, I thought it was just the TCP queue filling up, but that wasn't the case. I'd be interested in hearing about the fix for this! Bill - ------------------------------------------------------------------------------ Bill Gutknecht "If I die, I will go before Crom and he will BNR/Northern Telecom ask me 'What is the Riddle of Steel?' If I Research Triangle Park, NC do not know it, he will cast me out of billag@bnr.ca Valhalla and laugh at me ... " - ------------------------------------------------------------------------------ --------------------------- Newsgroups: comp.os.vxworks Subject: Digital and Wind River Systems Announcement Date: Thu, 6 Apr 1995 15:29:57 GMT From: segrest@bobseg.enet.dec.com () Organization: Me and only Me Message-ID: <1995Apr6.152957.1394@datum.nyo.dec.com> Reply-To: segrest@bobseg.enet.dec.com () Sender: usenet@datum.nyo.dec.com (USENET News System) DIGITAL EQUIPMENT CORPORATION AND WIND RIVER SYSTEMS FORM NEW ALLIANCE TO EXTEND ALPHA OPTIONS FOR OEMS AND EMBEDDED DEVELOPERS -- VxWorks 5.2, The Industry's Premier Embedded Development Environment, Is Now Available On Alpha And Non-Alpha Workstation Hosts And Single Board Computer Targets -- -- New Solutions Are Distributed By Both Wind River Systems And Digital Equipment Corporation -- MAYNARD, Mass. -- April 5, 1995 -- Digital Equipment Corporation and Wind River Systems, Inc. today announced a new joint development and marketing agreement that strengthens their existing alliance and substantially increases the portfolio of Alpha-based embedded applications, one year after introducing VxWorks for Alpha -- the industry's first embedded, realtime development environment optimized for next-generation 64-bit RISC architecture. Realizing greater opportunities to service large industrial OEMs with high-end VMEbus platform requirements and high-volume custom embedded applications, Alameda, CA-based Wind River Systems -- the developer of the VxWorks real-time operating system and WindPower development tools -- is increasing its support for Alpha. Wind River and Digital will make VxWorks 5.2 -- the latest version of the industry's leading embedded software -- available as a standard offering for both Alpha and non-Alpha workstation hosts, thereby expanding the VxWorks advantage to a wider range of developers. In addition, VxWorks 5.2 will support Alpha targets -- in the form of Alpha AXPvme VME single board computers -- from several non-Alpha workstations, including Sun Microsystems, Hewlett-Packard Co. and Windows platforms. Digital Equipment, which entered the embedded market 38 years ago, and expanded its embedded offerings in March 1994 with the industry's most comprehensive line of VME and PCI-based hardware and realtime software, can now move users with non-Alpha host systems to higher performance platforms quickly and cost-effectively. Under the terms of the agreement, Wind River is merging VxWorks 5.2 with the current VxWorks for Alpha. Through collaborative engineering efforts, the WindPower tools, including the WindView system visualizer and the VxSim simulation tool, are being expanding to support Alpha hosts and targets. A new marketing and sales component has been added to the Digital-Wind River alliance. VxWorks 5.2 will be sold by both Wind River and Digital. Digital has become an authorized distributor of all Wind River products, for both Alpha and non-Alpha hosts and targets. In the interim, Alpha-based VxWorks products are available from both Digital and Wind River. David Wilner, Wind River's co-founder and vice president of marketing, said that the new alliance will offer embedded systems developers seeking the high performance of an Alpha-based target system -- the most powerful development environment available on the market. "Digital's Alpha architecture is quite popular among OEMs in high-end VME markets including process control, SCADA, military and simulation, telecommunications and videographics. The robust VxWorks real-time operating system and productivity-enhancing WindPower tools are ideal for developing applications in these areas." William Armitage, vice president of Digital's Embedded and Realtime Products Group, stated, "Our new agreement with Wind River improves the breadth and depth of our realtime software offerings. Digital's large OEM customers, which include Raytheon Corp. and Bailey Controls, are eager to add greater performance and new software options to their embedded environments. VxWorks will continue to play a pivotal role in our next-generation realtime strategy as it moves forward." "The barriers that have prevented VME and embedded developers from rapidly migrating to Alpha have been removed now that non-Alpha workstation sites can access VxWorks 5.2 on Alpha targets," said Warren Andrews, industry analyst and editor of Embedded Computer Trends. "Digital's and Wind River's sales forces can service embedded OEMs -- specifically those with a substantial investment in non-Alpha workstations and PCs -- that are ready to port to Alpha VME single board computer targets. This represents a substantial portion of the market," he said. Paul Zorfass, an industry analyst covering embedded and realtime software for Framingham, MA-based market research firm International Data Corp., agrees. "Embedded OEMs in cost-conscious industries can now access Alpha 64-bit RISC performance and the latest version of VxWorks without sacrificing their existing hardware and software investments or retraining personnel. Digital and Wind River have created a welcome alternative," he said. VxWorks 5.2 will be available on Alpha host workstations and can be used by OEMs developing applications for non-Alpha targets, according to Steve Hargrave, manager of software alliances for Digital's Embedded and Realtime Products Group. "Digital's Alpha workstations have emerged as the most efficient development platform for the embedded market. Some OEM customers have requested VxWorks on Alpha host workstations and Motorola 68000 targets as they prefer the richer Alpha environment backed by UNIX software tools and 64-bit RISC performance." The New VxWorks Will Be The Most Robust, Standard Offering As part of the new agreement, Digital will have continuing access to new Wind River products as soon as they become available. Beginning with version 5.2, the Alpha platform will be available as a standard VxWorks offering, thereby remaining consistent with future kernel extensions and new or modified layered products that will be developed, according to Hargrave. "All forthcoming VxWorks applications integrated with Alpha hosts and targets will be the latest and most robust solutions available." Pricing And Availability VxWorks 5.2 for Alpha targets and Alpha host workstations will begin shipping in the fall of 1995. Pricing will be announced in summer 1995. Wind River Systems, Inc. develops, markets and supports software for developing real-time embedded applications. Its extensive WindPower product line enables customers to standardize their designs across products and enjoy rapid time-to-market. Wind River is a major supplier to the medical instrumentation, automotive, telecommunications, networking, industrial automation, electronics, and aerospace industries. Wind River is headquartered in Alameda, California with offices throughout the United States, and overseas operations in France, Germany, the United Kingdom, Scandinavia, Japan, and the Asia-Pacific region. Digital Equipment Corporation is the world's leader in open client/server solutions from personal computing to integrated worldwide information systems. Digital's scalable Alpha platforms, storage, networking, software and services, together with industry- focused solutions from business partners, help organizations compete and win in today's global marketplace. --------------------------- Newsgroups: comp.os.vxworks Subject: Mupltiple writes to a socket Date: 6 Apr 1995 15:27:08 GMT From: billag@b4pphff (Bill Gutknecht) Organization: Bell-Northern Research Message-ID: <3m118c$6is@nrtphba6.bnr.ca> If two tasks of equal priority both peform a write() operation on a socket simultaniously, what happens? Bill - ------------------------------------------------------------------------------ Bill Gutknecht "If I die, I will go before Crom and he will BNR/Northern Telecom ask me 'What is the Riddle of Steel?' If I Research Triangle Park, NC do not know it, he will cast me out of billag@bnr.ca Valhalla and laugh at me ... " - ------------------------------------------------------------------------------ --------------------------- Newsgroups: comp.os.vxworks Subject: A24 Addressing for MVME-167 Board with 16MBytes local RAM Date: 7 Apr 1995 01:18:42 GMT From: angelo@mrad.com.au (Angelo Capozza) Organization: MRad Pty Ltd, Australia Message-ID: <3m23ti$ld1@zeppo.mrad.com.au> Reply-To: angelo@mrad.com.au - --- I asked Wind River Systems (WRS) support@wrs.com if there is a specific Board Support Package (BSP) for the MVME-167 with 16 MBytes of RAM that allows as a minimum A24 addressing for the total RAM space with the MVME-167 control registers set correctly? I received the following response: "Unfortunately, we dont have a version of the mv167 that uses A24 addressing. It looks to me like you have modified the proper places. With the possible exception that I think you may have to change something in sysHwInit() in sysLib.c for the VMEchip2 operation. A good place to also ask is the customer base, which has far more experience with such things than we do here at Wind River. To submit articles to the vxWorks exploder, send to: vxwexplo@lbl.gov to subscribe send to: vxwexplo-request@lbl.gov Thanks, and sorry I dont have a bsp for this and only have limited information." Consequently , I am here asking for help. The background information (that was also sent to WRS) follows: We are using the following Development Environment: Processor: MVME167-33, 33MHz 68040 processor with 16 Mbytes RAM Operating System: Wind River VxWorks Version 5.1.1 The following peripheral (MATRIX MD-SCSIFLP 6U high Floppy disk drive) is also used: Part number: DSM-FL-FFF-NNN Serial Number: 941261002 Board type: D0520-9105-01J Revision: J Serial Number: 940180803 Address: 0x2000 (in short address space and unchanged by us) Factory Configurations: Unchanged by us Mass Storage: Floppy Disk only Part Number: 5501-0316 Drive ID: 1 Size and Type: 3.5' floppy Manufacturer: TEAC Model Number: FD-235HF (Can not read last two characters properly) Serial number: L349142 Factory Configuration: Unchanged by us. VxWorks drivers and VxWorks Version 5.0.x (Received from MATRIX Corporation) Part number: VXSCSIFLP0554 Description: MATRIX MD/DSM-SCIFLP Source level Driver for VxWorks Format: QIC 150 TAR Version: V2.3 Serial number: 941260501 Date: july 7 1994 To date we have been having some trouble integrating the MATRIX supplied VxWorks drivers into our development system. Some progress has been made but we require some assistance. During our investigation we found that the MVME167 card with 16 Mbytes of memory can work with the DMA driver on the DSM Floppy disk drive. The VxWorks kernel needed modification to allow A24 addressing from an external VME bus device. The Board Support Package (BSP) for the MVME-167 assumes a default memory size of 4 MByte and allows A24 addressing from 0x00800000 to 0x00bfffff. We need A24 addressing from the external VME bus to be from 0x00000000 to 0x00ffffff (i.e. the full 16 MByte RAM range) since DMA memory is dynamically allocated by malloc() and the DMA driver on the DSM Floppy disk drive sets up the Address Modifier for A24 addressing only. We had to modify the procedure sysProcNumSet() and sysLocalToBusAdrs() in sysLib.c as follows: void sysProcNumSet ( int procNum /* processor number */ ) { int memsize; sysProcNum = procNum; /* Set memory base address, as seen from VME bus */ if (procNum == 0) { memsize = (4 << (*MEMC_MCR & 0x07)); /* Test was this: if (memsize == 4) */ if ((*MEMC_MCR & 0x07) <= 2) /* below 16 MByte RAM */ { /* Was *VMECHIP2_VSAR2 = 0x003f0000 | */ *VMECHIP2_VSAR2 = 0x00ff0000 | LOCAL_MEM_BUS_A24 | (LOCAL_MEM_BUS_A24 >> 16); /* map 4Mb of DRAM */ *VMECHIP2_VSATR2 = LOCAL_MEM_BUS_A24 >> 16; /* to 0x00800000..0x00bfffff */ *VMECHIP2_VSAMSR |= (VSAMSR2_SNP_INHIBIT | VSAMSR2_WP | VSAMSR2_SUP | VSAMSR2_USR | VSAMSR2_A24 | VSAMSR2_A32 | VSAMSR2_D64 | VSAMSR2_BLK | VSAMSR2_PGM | VSAMSR2_DAT); /* all but A32.. Not now! */ } *VMECHIP2_VSAR1 = (((memsize << 4) - 1) << 16) | (memsize << 21) | (memsize << 5); /* map all DRAM to VMEbus access */ *VMECHIP2_VSATR1 = (memsize << 5); /* Following was this: *VMECHIP2_VSAMSR |= (VSAMSR1_SNP_INHIBIT | VSAMSR1_WP | VSAMSR1_SUP | VSAMSR1_USR | VSAMSR1_A32 | VSAMSR1_D64 | VSAMSR1_BLK | VSAMSR1_PGM | VSAMSR1_DAT); /* all but A24 */ /* It now becomes A24 and A32 addressing */ *VMECHIP2_VSAMSR |= (VSAMSR1_SNP_INHIBIT | VSAMSR1_WP | VSAMSR1_SUP | VSAMSR1_USR | VSAMSR1_A24 | VSAMSR1_A32 | VSAMSR1_D64 | VSAMSR1_BLK | VSAMSR1_PGM | VSAMSR1_DAT); /* all */ } else *VMECHIP2_VSAMSR = 0x00; /* turn off all bits */ /* set global control register VME address, based on processor number */ if (sysProcNum < 15) { /* VME address of GCSR is function of processor number */ memsize = *VMECHIP2_LBTVCR; memsize &= ~(0xfff00000); /* clear bits */ *VMECHIP2_LBTVCR = (memsize | (sysProcNum << 20) | (GCSR_GROUP_ADDR << 24)); #ifdef INCLUDE_VSB /* configure mapping of local RAM to/from VSB */ sysVsbBusConfig (sysProcNum); #endif /* INCLUDE_VSB */ } } - ----------------------------------------------------------------------------- STATUS sysLocalToBusAdrs ( int adrsSpace, /* bus address space in which busAdrs resides */ char *localAdrs, /* local address to convert */ char **pBusAdrs /* where to return bus address */ ) { if ((int)localAdrs < LOCAL_MEM_LOCAL_ADRS || localAdrs >= sysMemTop ()) { /* this is off-board memory - just return local address */ *pBusAdrs = localAdrs; return (OK); } /* this is on-board memory - map to bus address space; * the following memory mapping is established in sysProcNumSet(): * - only processor 0 has memory on bus, * - the memory is placed in STD space at * address LOCAL_MEM_BUS_24 if < 8 Meg of local memory * and at twice the LOCAL_MEM_SIZE for EXT space. * For 8MB memory 0x01000000 * 16MB memory 0x02000000 * 32MB memory 0x04000000 * - short I/O is not mapped locally. */ switch (adrsSpace) { case VME_AM_SUP_SHORT_IO: case VME_AM_USR_SHORT_IO: /* no local map */ return (ERROR); case VME_AM_STD_SUP_PGM: case VME_AM_STD_SUP_DATA: case VME_AM_STD_USR_PGM: case VME_AM_STD_USR_DATA: /* Original test was the following which did not work for a 167 card with 16Mb RAM. (Assumed a 4 MByte Board ). It was change to pass if board memory is <= 16MBytes. if ((4 << (*MEMC_MCR & 0x07)) == 4) */ /* New test follows */ if ((*MEMC_MCR & 0x07)<=2) { *pBusAdrs = localAdrs + LOCAL_MEM_BUS_A24 - LOCAL_MEM_LOCAL_ADRS; return (OK); } else return (ERROR); case VME_AM_EXT_SUP_PGM: case VME_AM_EXT_SUP_DATA: case VME_AM_EXT_USR_PGM: case VME_AM_EXT_USR_DATA: *pBusAdrs = localAdrs + (((4 << (*MEMC_MCR & 0x07)) << 1) << 20) - LOCAL_MEM_LOCAL_ADRS; return (OK); default: return (ERROR); } } - ------------------------------------------------------------------------------- We had to modify the #defines in config.h as follows: #define LOCAL_MEM_LOCAL_ADRS 0x00000000 /* fixed at zero */ #define LOCAL_MEM_SIZE 0x01000000 /* Min memory: 16 Mbytes */ /*#define LOCAL_MEM_SIZE 0x00400000 /* Min memory: 4 Mbytes */ /*#define LOCAL_MEM_BUS_A24 0x00800000 /* Bus address in A24 range */ /* for boards with 4 Mbytes of RAM */ #define LOCAL_MEM_BUS_A24 0x0000000 /* A24 range offset (for 16MByte) */ - ------------------------------------------------------------------------------ The above changes are not optimal but they seem to work most of the time. With that background information, can anybody with similar problems (and hopefully solutions or helpful hints) respond to this newsgroup. Thanking all in anticipation of your advice. Regards - ------- Angelo Capozza Email : angelo@mrad.com.au Project Engineer Phone : (08) 260 8942 (work) MRad Pty Ltd Fax : (08) 260 8980 (work) Adelaide, AUSTRALIA --------------------------- End of New-News digest ********************** From darren@wrs.com Fri Apr 7 08:38:16 1995 From: Darren Cathey Date: Fri Apr 7 08:38:22 PDT 1995 Hello: I have a customer who is looking for a Unix filesystem port to VxWorks. If anyone knows of such a beast for sale or free via the archive, could you please provide me with the appropriate pointers. Also, I'm looking for a list of available PCI chipsets on the market other than the PCI Ethernet and SCSI chips available from Mot and AMD. If anyone knows of others, please let me know. Thanks, Darren Cathey Field Applications Engineer Wind River Systems 703-771-7934 ph darren@wrs.com From Harald.Grundner@erno.de Fri Apr 7 08:40:20 1995 From: Harald.Grundner@erno.de (Harald Grundner) Date: Fri Apr 7 08:40:29 PDT 1995 Subject: VxWorks Driver for SCC 82C52 ? Does anyone know where to get a driver for the serial comm. controller 82C52 or, even better, HAVE one? Thanks for any responses, -Harald ________________________________________________________________ Harald Grundner e-mail : harald.grundner@erno.de RIT 23 Phone : +49 421 539-4841 Daimler-Benz Aerospace AG Fax : +49 421 539-5577 Huenefeldstrasse 1-5 ___ D-28199 Bremen C ..\__ Germany _ | 00\ _ _________________________________/ \_|_______/___/ \____________ \_/ \_/ From tweadon@sadira.gb.nrao.edu Fri Apr 7 08:58:24 1995 From: tweadon@sadira.gb.nrao.edu (TIM WEADON) Date: Fri Apr 7 08:58:31 PDT 1995 Subject: Re: Mupltiple writes to a socket > If two tasks of equal priority both peform a write() operation on > a socket simultaniously, what happens? This can't happen! Although both tasks are equal priority only one task can be in control of the CPU at a time. Whichever one gets to the write routine first should win out. The semaphore mechanism should then hold off the second task until the first is complete. This actually is no different then asking what happens if a lower priority task is doing a write and a higher priority task attempts to write on the same socket. The first task to the semaphore wins. The only thing which gets harry is trying to prove who will actually get it first. Since tasks of equal priority are treated as round-robin tasks, relative to each other, the task you might have thought was going to get to do the write first may not because it was pre-empted by a higher priority task first then the round-robin mechanism caused the other task to get control. Bottom line,,, the write semaphore mechanism should take care of insuring the two tasks won't write over each others data. Tim Weadon National Radio Astronomy Observatory Green Bank, WV. 24944 From mmilde@scires.com Fri Apr 7 08:58:54 1995 From: mmilde@scires.com Date: Fri Apr 7 08:59:05 PDT 1995 Subject: C++ and taskSpawn I have two questions that I can't seem to find an answer to. Any help from a VxWorks/C++ guru is greatly appreciated! I need to have a function in an instance of a class spawn a function in that same class that runs with the same visibility as any other function in the instance. Doing such a thing seems perfectly reasonable but I don't know if VxWorks/WindC++ support it. #1: How do you spawn a task that sits inside a C++ class? For example, sp(class_name::function()) doesn't work. I've also tried several different possibilities, but none seem to work. The only way I can do this is to add an external "wrapper" function, spawn it, and have that function call the function inside the class. #2: If a particular instance of a class spawns a function that is in the same class, does that task have visibility as if it were just another function in that instance? Thanks in advance for the help! Mike Milde Scientific Research Corp. (404) 859-9161 mmilde@scires.com From brodrigu@polaris.cv.nrao.edu Fri Apr 7 09:12:09 1995 From: "Billie Rodriguez,205,312,7039436481" Date: Fri Apr 7 09:12:15 PDT 1995 Subject: Job Opening at NRAO The National Radio Astronomy Observatory (NRAO) has an opening for an Electronics Engineer (Software and Digital Systems) to work in Green Bank, West Virginia on the Green Bank Telescope Project . The GBT will be the world's most fully steerable radio telescope. Construction is in progress, with a completion date of early 1996. DESCRIPTION: Will assist in the design, construction, and maintenance of software and hardware systems for the Green Bank Telescope. Responsibilities will include coding of real-time control software, hardware/software integration, and maintenance of hardware/software systems. Some hardware design and maintenance of existing systems will be required. SPECIAL QUALIFICATIONS: At least a Bachelors degree in Electrical or Computer Engineering, with experience in programming real-time systems, digital and analog electronics, writing driver software, and integration of hardware/software systems. Familiarity with modern engineering and programming practices and state-of-the-art digital devices is necessary. The applicant should be an experienced user of oscilloscopes, logic analyzers, and various other hardware/software integration tools. Familiarity with some or all of the following is required: Object Oriented design, C/C++, UNIX, VxWorks, and sockets/RPCs. APPLICATION DEADLINE: May 1, 1995 For immediate consideration please send resume, via e-mail, to the NRAO Headquarters Personnel Office: brodrigu@nrao.edu From tweadon@sadira.gb.nrao.edu Fri Apr 7 09:59:26 1995 From: tweadon@sadira.gb.nrao.edu (TIM WEADON) Date: Fri Apr 7 09:59:35 PDT 1995 Subject: Re: C++ and taskSpawn > > I have two questions that I can't seem to find an answer to. Any help > from a VxWorks/C++ guru is greatly appreciated! > > I need to have a function in an instance of a class spawn a function > in that same class that runs with the same visibility as any other > function in the instance. Doing such a thing seems perfectly > reasonable but I don't know if VxWorks/WindC++ support it. > > #1: > How do you spawn a task that sits inside a C++ class? For example, > sp(class_name::function()) doesn't work. I've also tried several > different possibilities, but none seem to work. The only way I can do > this is to add an external "wrapper" function, spawn it, and have that > function call the function inside the class. > You can't spawn a C++ method (function) unless it is static. The way we do it is to make a static method and spawn it. Pass as one of your arguments to that method *this. Then either reference everything via the *this pointer (ie: thisPtr->var1 etc) or have that method invoke a non-static member function using the this pointer. See code below. #2: If a particular instance of a class spawns a function that is in the same class, does that task have visibility as if it were just another function in that instance? This is obviously yes now that you see that the way we keep track of the instance is by using *this for the instance we are referring to. *********************************************************************** The following code spawns the static method passing the *this ptr. monitorTaskId = taskSpawn (monitorTaskName, // task name monitorPriority, // task priority (VX_DEALLOC_STACK + // task options VX_FP_TASK + VX_STDIO), 7000, // task stack byte size (FUNCPTR) ClassName::monitorTask, // task (int)this,2,3,4,5,6,7,8,9,10); // input args. if (monitorTaskId == ERROR) { cout <<"error creating Receiver monitor task." << endl; } *********************************************************************** This is the static method which in turn calls a non-static method using the passed *this ptr. void ClassName::monitorTask(ClassName* obj) { obj->nonStaticMethod(); } *********************************************************************** Tim Weadon National Radio Astronomy Observatory Green Bank, WV. 24944 From glwicks@sandia.gov Fri Apr 7 12:45:54 1995 From: Greg Wickstrom Date: Fri Apr 7 12:46:01 PDT 1995 Subject: qsort problems I am running qsort under VxWorks 5.1 on a Force CPU30 board. I recently ported over a bunch of code developed under Visual C++, and everything is working fine except the qsort routine. Both Visual C++ and VxWorks claim to be ANSI compliant, and the prototype for the VxWorks version is the same as that for Visual C++. A closer invesigation in the way the compare function is called by the VxWorks qsort looks like the implementation actually required arguments in a slightly different order. Even with that problem solved, it doesn't seem to terminate properly. Any suggestions? Thanks in Advance, Greg Wickstrom glwicks@sandia.gov From daemon@vxw.ee.lbl.gov Sat Apr 8 04:00:29 1995 From: daemon@vxw.ee.lbl.gov Date: Sat Apr 8 04:00:36 PDT 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sat Apr 8 04:00:20 PDT 1995 Subject: Re: Mupltiple writes to a socket ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Mupltiple writes to a socket Date: Fri, 7 Apr 1995 13:46:18 GMT From: magna@crosfield.co.uk (John Hartridge) Organization: Crosfield, Hemel Hempstead, UK Message-ID: <1995Apr7.134618.16884@crosfield.co.uk> References: <3m118c$6is@nrtphba6.bnr.ca> In article <3m118c$6is@nrtphba6.bnr.ca> billag@b4pphff (Bill Gutknecht) writes: > If two tasks of equal priority both peform a write() operation on > a socket simultaniously, what happens? > > Bill It is impossible for two tasks on one processor to do anything simultaneously. If they have equal priority, it just depends which task is schduled first, but one of then must get there first. John - -- John Hartridge ext 3282 | ____ Crosfield Electronics Ltd | / 0,0 /__/ /| / /__/ magna@crosfield.co.uk | /__/ \_/ / / / |/ / / o --------------------------- End of New-News digest ********************** From daemon@vxw.ee.lbl.gov Sun Apr 9 04:00:26 1995 From: daemon@vxw.ee.lbl.gov Date: Sun Apr 9 04:00:34 PDT 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sun Apr 9 04:00:18 PDT 1995 Subject: Re: MVME167 boards Subject: Re: C++ and taskSpawn Subject: Backplane driver ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: MVME167 boards Date: 5 Apr 1995 23:13:43 -0400 From: kab1lm@aol.com (KAB1LM) Organization: America Online, Inc. (1-800-827-6364) Message-ID: <3lvm97$eh7@newsbf02.news.aol.com> References: <9504032110.AA02320@array.ca> Reply-To: kab1lm@aol.com (KAB1LM) Sender: root@newsbf02.news.aol.com Have you aske Motorola? If you need an e-mail address for Motorola support, let me know, I do some work with them, and they usually are pretty helpful. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: C++ and taskSpawn Date: 8 Apr 1995 12:28:53 +1000 From: rjl@f111.iassf.easams.com.au (Rohan LENARD) Organization: EASAMS (Australia) Pty Ltd Message-ID: <3m4sd5$3dp@f111.iassf.easams.com.au> References: <9503077972.AA797281065@mailhost.scires.com> In article <9503077972.AA797281065@mailhost.scires.com>, wrote: > > #1: > How do you spawn a task that sits inside a C++ class? For example, > sp(class_name::function()) doesn't work. I've also tried several > different possibilities, but none seem to work. The only way I can do > this is to add an external "wrapper" function, spawn it, and have that > function call the function inside the class. Okay, the reason it doesn't work is because an instance of a class (an object) has a this pointer. This is used in all object method calls. There are basically two options - 1. Use an external wrapper function. (This is the "correct" way, and you can infact use a static member function to achieve this).) 2. Rely on some knowledge about the compiler. If you use the gnu c++ compiler, then on a m68k (it's the only architecture I've tried), the first argument passed is the "this" pointer. eg. (something like this) class foo { /* ...... */ DoSomething(int Param); }; main() { foo AnObject; int Param = 5; sp((FUNCPTR)(&foo::DoSomething()),&AnObject,Param,0,0,0,...); } > > #2: > If a particular instance of a class spawns a function that is in > the same class, does that task have visibility as if it were just > another function in that instance? > If you provide the "this" pointer yes. If you somehow use another mechanism then maybe not. The way to think of objects in c++ is that the code for the functions is "shared", the data is what is different for each instance of the class. Regards, Rohan - -- - ---------------------------------------------------------------------------- rjl@iassf.easams.com.au | All quotes can be attributed to my automated quote Rohan Lenard | writing tool. Yours for just $19.95; and if you +61-2-367-4555 | call now you'll get a free set of steak knives ... --------------------------- Newsgroups: comp.os.vxworks Subject: Backplane driver Date: 7 Apr 1995 23:44:38 -0500 From: jford@telerama.lm.com Organization: Telerama Public Access Internet, Pittsburgh, PA Message-ID: <3m54bm$56k@africa.lm.com> Hi, I'm about to start developing a mixed VxWorks (MVME-16X) and VME based PC system. I would like to connect the two boards with a backplane TCP/IP connection. I know that VxWorks includes a backplane driver that works between VxWorks boards. What sort of software is needed on the other end? Has anyone connected a VxWorks system to a non-VxWorks system using the backplane driver? Thanks. John Ford Contraves, Inc. --------------------------- End of New-News digest ********************** From dowling@cub.kscorp.com Mon Apr 10 13:37:56 1995 From: Bob Dowling Date: Mon Apr 10 13:38:06 PDT 1995 Subject: PCI Chipsets > I'm looking for a list of available PCI chipsets > on the market other than the PCI Ethernet and SCSI > chips available from Mot and AMD. The Feb. 2, 95 issue of EDN has an article on PCI bridge chipsets listing the following manufacturers: Chips & Technologies, DEC, IBM Microelectronics, Intel, Mentor Arc Inc., Motorola, Newbridge Microsystems, Oki Semiconductor, Opti Inc., PLX Technology Inc., SIS Microelectronics, TI, and Toshiba. The March 30, 95 issue of EDN is a special issue devoted to PCI but a quick scan did not turn up any other vendor lists, but Applied Micro Circuits Corp. has a featured article & there are a few ads for PCI chips. HTH! Bob Dowling (dowling@kscorp.com) KineticSystems Corp. From YFILLION@spar.ca Mon Apr 10 14:08:23 1995 From: YFILLION@spar.ca Date: Mon Apr 10 14:08:29 PDT 1995 Subject: Problem loading vxWorks from dosFs on top of a RAM device This has probably been done before but I am still having difficulties loading vxWorks from a local dosFs file system defined on top of a RAM disk. begin 666 loadflas.net M_U=00Y($```!"@`!`````/O_!0`R`$@$```'`!8```!"````#P!6````6``` M```"0`,``*X````,`%H```#N`P``5&EM97,@3F5W(%)O;6%N("`H5%0I`/`! M__^%`#(`2@`]`"X`+`$!`````6=I)@)*`/0:7!(:"0```!`@4(X`'#91$0,0 M)@)0_O[^_O[^_O___________O__________________________``$B`((` M_____S4!__^%`?__^P'_______]``O_______UXN/4M<7)F/(3T]7&@N/2XS M7%Q<7%Q<7%Q<7#,S:&AH4JF$>GJ$<&:$A#U'A'"CA(1FA'IF<(2$K82$<#TS M/59GI2<%)P4G!2<%(],STS/3,] M,X1E)Z4@````!P4G!2<%*$7(1<``"$`(1DQF1V9'3D```"X`4UP]4EQ< M7%P`7#,YBHI<-P"+BUR*-P`]/0!24ERW/3T``%QE)24E)2,S,S M,UQ<7%Q<7%QD7%Q<7%Q<7$A0($QA28"4/O_!0`R```````&`!````!Z!```__\&````B@0` M``@``@```)`$````````````````""-\`$H````"```!#P```$-%"``1TP`` M"C$M($D@9&5F:6YE9"!A(&1O2!&3$%32"!M M96UO2!D;W-&$!T/`!```=0T+2!)(&UO9&EF:65D(&)O;W1#;VYF:671E(&)L;V-K3!T/`!```=0W+2!! M2!T;PW4`1```0H2`+`$>!Z7'0\`$``!U&%T=&%C:"!T;R!T:&4@ M;F5T=V]R:R!L;C`@*&)O;W0@<&%R86UE=&5R("=O=&AE Reply-To: skv@pcdoh.pgh.wec.com Sender: usenet@tron.bwi.wec.com (Usenet_news poster) We need to port NTP to VxWorks. I was wondering if anybody did this before. Any pointers would be helpful and appreciated. Suresh (skv@pcdoh.pgh.wec.com) --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks dying in semQPut... Date: Sat, 08 Apr 95 10:14:44 PDT From: "Mark A. Menge" Organization: ICNET... Your Link To The Internet... +1.313.998.0090 Message-ID: References: <3lppfb$q3q@lucy.infi.net> <195@jtechcons.win.net> In article <195@jtechcons.win.net>, writes: > Path: condor.ic.net!news1.oakland.edu!wsu-cs!newsxfer.itd.umich.edu!zip.eecs.umich.ed u!newshost.marcam.com!news.mathworks.com!usenet.eel.ufl.edu!spool.mu.edu!howlan d.reston.ans.net!news.sprintlink.net!ns1.win.net!witch!jtechcons!jfhughes > Newsgroups: comp.os.vxworks > Message-ID: <195@jtechcons.win.net> > References: <3lppfb$q3q@lucy.infi.net> > Reply-To: jfhughes@jtechcons.win.net (John F. Hughes) > From: jfhughes@jtechcons.win.net (John F. Hughes) > Date: Tue, 04 Apr 1995 04:35:55 GMT > Subject: Re: VxWorks dying in semQPut... > Lines: 44 > > > In article <3lppfb$q3q@lucy.infi.net>, Charles Shannon Hendrix (shendrix@taurustech.com) writes: > > > >I have a simple network program that is basically a network data > >reflector. It reads data from a host, writes it to a RAM location > >on an mv162 board running VxWorks, reads some variables from the > >same block, and sends that back to a UNIX host computer. > > > >I'm using UDP. > > > >The program runs for awhile... usually 30 minuts to 2 hours... and > >then just stops with no error messages or anything. Tracing the > >task from the shell shows this: > > > > 29778 _vxTaskEntry +10 : _main_init (311554, 311540, 31152c, 7e4, 5, 0, 0, 0, > > 0, 0) > >312294 _main_init +ec : _main_loop (5, 6, 30b5ac, 3057ec, 5, 5, 5) > >3136ea _main_loop +196: _va_read_poll (5, 30b5ac, 30193c, 5) > >312a70 _va_read_poll +78 : _send_recv_data (1, 3018ba, 6) > >3127ea _send_recv_data+a6 : _recv_data (10, 301870) > >312ce4 _recv_data +54 : _semQPut (20, 301808, 0, 0, 301810) > >value = 0 = 0x0 > > > >..so it looks to me like it's stuck in a kernel/lib call. > > > I have been having a similar problem, using TCP/IP streams. > Everything runs great for a variable amount of time (we have run > anywhere from a few seconds to 5 hours), then one or more of the > tasks that are doing the communication stop in semQPut. I have > been tearing my hair out for months trying to figure out what is > wrong with our code. We just started pointing the finger at the > board vendor, but maybe it is a vxWorks problem? > > In our case, we frequently start seeing input errors on the > Ethernet interface (using ifShow "ln") when the communications > stops working. I sent email to WRS (TSR 21870) asking what the > errors mean and what causes the counter to increment. > > If anyone out there has a clue what can be causing this, we could > really use the help. > > regards, > jfh > > > From Mark Menge using a GMSV46 68040. Hello there fellow pullers of hair! I have ran into a very similar situation where it dies in semQPut and after about 5 hours. However, my ifShow "ln" shows all input packets have errors. The source of the problem appears to be related to a jumper with watchdog timer (Hardware not software) that is supposed to reboot the board in certain instances, and appears to introduce a very brief transients on the reset line (we are investigating further). I remember we ran into a similar problem on the MVME 68030 board when we incorrectly had our ADC boards jumpered to halt CPU activity during conversions (delayed DTACK for 30 - 130 uSecs) which caused the ethernet chip to lockup. My guess is the ethernet chip is more sensitive to brief assertions on the reset line than the CPU is. Is that possible? What is going on? Mark Menge DSP Technologies --------------------------- Newsgroups: comp.os.vxworks Subject: Anybody used VxWorks on MVME177 Date: Mon, 10 Apr 1995 16:04:05 +0000 From: Paul@pgt1.demon.co.uk (Paul G Tindall) Organization: Myorganisation Message-ID: <797529845snz@pgt1.demon.co.uk> Reply-To: Paul@pgt1.demon.co.uk Sender: usenet@demon.co.uk Does anyone have any experience of trying to run vxWorks on a Motorola MVME177 card. We are currently trying a 32MB card with VADSworks, the first stage boot is fine but the system crashes soon after, either whilst loading the symbol table from the host or during subsequent network activity. I've set the addresses and memory size in all the appropriate places and the 177 is a 167 with a 68060 at 50MHz. The local bus is running at 25MHz and all device addresses and types are as for the MVME167. Anyone else tried this and managed to get it to work OK? - -- Paul G Tindall --------------------------- Newsgroups: comp.os.vxworks Subject: RPC over SLIP with VxWorks for Windows? Date: Mon, 10 Apr 1995 16:22:12 GMT From: lee@dradeb.demon.co.uk (Lee) Organization: DRA Bedford Message-ID: Sender: news@demon.co.uk (Usenet Administration) RPC over SLIP with VxWorks for Windows? I'm having problems with RPC over SLIP with VXW for Windows. I can boot over the SLIP link using FTP but after booting I get an RPC error then an RDB error from the target after about 1 min. The errors are :- 'Cannot register service : RPC Timed Out' 'rdbTask : couldn't register RDB service' This means that the rpc and rbd tasks fail. I can boot over ethernet and the RPC and RDB tasks come up OK. The only difference is I boot from 'ne' and not sl=/tyCo/1. On examination of the SLIP traffic with my winsock I have have IP packets of length 85 with UDP protocol 17 (UDP). These are from ports 1005/1007 to 111 BUT the IP addreses are identical (146.80.46.164) my target IP address. host = 146.80.46.163 target = 146.80.46.164 Having looked at the routetable I get the following :- 127.0.0.0 127.0.0.0 5 0 0 lo0 146.80.46.163 146.80.46.164 5 0 131 sl0 146.80.46.164 146.80.46.164 5 2 42 sl0 This does not seem right. ?? Can anyone tell me if the table is correct and how it got like this and if not how to fix it ?. Has anyone done remote debugging over SLIP with VxWorks for Windows. martin@drabed.demon.co.uk --------------------------- Newsgroups: comp.os.vxworks Subject: Checksum on object load Date: Mon, 10 Apr 1995 16:19:28 GMT From: dfox@netcom.com (Don Fox) Organization: NETCOM On-line Communication Services (408 261-4700 guest) Message-ID: Sender: dfox@netcom10.netcom.com I have been perusing through documentation on loadLib and on various object formats (COFF, a.out, ECOFF,...) and have not seen any indication that there is a way to get loadModule to perform a checksum during the loading of an object file from disk. I would like to compare what is read off of disk to what I think the checksum of the file should be. If there are routines that already exist in VxWorks to handle this, then I would like to use them, otherwise, I will have to write something myself. Any pointers to info would be appreciated. Thanks in advance. Don Fox dfox@netcom.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Unix filesystem port to VxWorks wanted Date: Mon, 10 Apr 1995 19:07:46 GMT From: joshisx@ntmtv.com (Suhas Joshi) Organization: Northern Telecom, Mountain View Message-ID: References: <199504071537.AA24253@darya.wrs.com> Sender: news@ntmtv.com In article <199504071537.AA24253@darya.wrs.com>, Darren Cathey wrote: >Hello: > >I have a customer who is looking for a Unix filesystem >port to VxWorks. If anyone knows of such a beast for >sale or free via the archive, could you please provide >me with the appropriate pointers. I have a doubt whether Unix FS is suitable enough for real time applications. The file system can get easily currupted if machine is switched on and off a few times. Can any one clarify this. SJ --------------------------- Newsgroups: comp.os.vxworks Subject: Re: C++ and taskSpawn Date: 10 Apr 1995 23:40:19 GMT From: billr@saab.CNA.TEK.COM (Bill Randle) Organization: Tektronix, Inc., Redmond, OR Message-ID: <3mcfl3$7js@marlyn.cna.tek.com> References: <9503077972.AA797281065@mailhost.scires.com> <3m4sd5$3dp@f111.iassf.easams.com.au> Reply-To: billr@saab.CNA.TEK.COM In article <3m4sd5$3dp@f111.iassf.easams.com.au>, rjl@f111.iassf.easams.com.au (Rohan LENARD) writes: |> In article <9503077972.AA797281065@mailhost.scires.com>, |> wrote: |> > |> > #1: |> > How do you spawn a task that sits inside a C++ class? For example, |> > sp(class_name::function()) doesn't work. I've also tried several |> > different possibilities, but none seem to work. The only way I can do |> > this is to add an external "wrapper" function, spawn it, and have that |> > function call the function inside the class. |> |> Okay, the reason it doesn't work is because an instance of a class (an object) |> has a this pointer. This is used in all object method calls. There |> are basically two options - |> 1. Use an external wrapper function. |> (This is the "correct" way, and you can infact use a static member |> function to achieve this).) |> |> 2. Rely on some knowledge about the compiler. |> If you use the gnu c++ compiler, then on a m68k (it's the only |> architecture I've tried), the first argument passed is the "this" |> pointer. There is yet another way, depending on your system requirements. What we did was to create a Process class that encapsulates (and isolates) the rest of the system from the details of the OS [this becomes really handy if you need to port an application to another OS!]. Our Process class has a run() member function (which is the main processing loop) and includes member functions to set/get task priority and preemption and member functions to send messages to the process. The task then is actually spawn in the constructor for the class. Since you are inside the class, you know the address of the run() function - it doesn't have to be static. In many cases, the spawned task then just goes to sleep, waiting on a message queue until told to do something. Finally, we use one of the unused fields in the VxWorks tcb to save the 'this' pointer in. This way you can always get a pointer to a particular process (or the current process), given its taskId. For example, if you define "_curProc" to be a pointer to the currently running Process class in a header file: #define _curProc ((Process *)taskIdCurrent->spare1) then you caan do things like: Process myproc(initialPriority, stackSize); // create and spawn a new task _curProc->sendMsg(); // send a message to current process or _curProc->nopreempt(ON); // don't let anyone else in The best method for you (or anyone else) will depend on the actual application requirements. - -- -Bill Randle Tektronix, Inc. billr@saab.CNA.TEK.COM --------------------------- End of New-News digest ********************** From HARRIS.KDREWRY@ic1d.harris.com Tue Apr 11 06:37:46 1995 From: "KDrewry" Date: Tue Apr 11 06:37:53 PDT 1995 Subject: Flash-Writing Routines for Intel 28F008SA I am trying to set up a filesystem on an IP-FLASH-8M module on an MVME162LX. I have instructions on how to accomplish this, but I don't have any way to write my data to the Flash. Does anybody have any routines to write to the Intel 28F008SA? This IP-FLASH uses 2-chip banks: the high byte is written to one chip and the low byte is written to the other. Any help would be greatly appreciated. Kyle Drewry Harris Corporation Melbourne, FL 407-727-4043 kdrewry@harris.com From pf1122@tcville.es.hac.com Tue Apr 11 07:48:25 1995 From: pf1122@tcville.es.hac.com (Kim Chacon) Date: Tue Apr 11 07:48:39 PDT 1995 Subject: Re: Flash-Writing Routines for Intel 28F008SA Kyle Drewry wrote: ------------------ >Submitted-by HARRIS.KDREWRY@ic1d.harris.com Tue Apr 11 06:37:46 1995 >Submitted-by: "KDrewry" > >I am trying to set up a filesystem on an IP-FLASH-8M module on an MVME162LX. I >have instructions on how to accomplish this, but I don't have any way to write >my data to the Flash. Does anybody have any routines to write to the Intel >28F008SA? This IP-FLASH uses 2-chip banks: the high byte is written to one >chip and the low byte is written to the other. Any help would be greatly >appreciated. We had to solve the same problem. We purchased some software from a company called CompWare Software. Contact Paul Anderson at (602) 998-8650 for more information. The software we purchased included a flashLib, which was downloaded to VxWorks and provided routines such as writeFlash(), eraseFlash() and more. I hope this helps. Kim M. Chacon Hughes Aerospace and Electronics Company El Segundo, CA email: chacon@tcville.es.hac.com From hebo@mbari.org Tue Apr 11 11:12:41 1995 From: Bob Herlien Date: Tue Apr 11 11:12:47 PDT 1995 Subject: Re: NTP for VxWorks > > We need to port NTP to VxWorks. I was wondering if anybody did this before. > Any pointers would be helpful and appreciated. > > Suresh > (skv@pcdoh.pgh.wec.com) Suresh, Get ntpv3.1.tar.gz from the VxWorks archive. This port of NTP 3.1 may or may not still require the usrTime library that I wrote when I ported NTP 1.0 several years ago. I haven't looked at it for quite a while, and someone else did the 3.1 port. If it does require usrTime you can also pick up usrTime.shar from the archive. P.S. Although I was a beta site for VxWorks 5.2, I wasn't able to actually test the beta due to illness. VxWorks 5.2 includes many of the time functions necessary for NTP, but as I said, I wasn't able to check if it's fully compatible with NTP. I suspect we'll have more work to do when 5.2 is released. -------------------------------------------------------------- Bob Herlien MBARI (Monterey Bay Aquarium Research Institute) hebo@mbari.org From mwbrown@somnet.sandia.gov Tue Apr 11 14:50:47 1995 From: "Matthew W. Brown" Date: Tue Apr 11 14:50:55 PDT 1995 Subject: Motorola S-records Hopefully this is a basic question for someone out there ... WindRiver ships many of the gnu utilities with their VxWorks distribution; however, they don't ship "objcopy". I suppose that most prople never need objcopy. I only need it because I am running as68k for another (non-VxWorks) board in my system. Unfortunately, this board needs Motorola S-records. With objcopy, one can easily convert a.out to S-records objcopy -O srec inputFile outputFile Is there another way to generate S-records? Maybe generating them in the first place instead of converting them w/ objcopy??? I really was hoping to not have to build a gnu cross-compiler for this board as I really already have one. Any easy ideas? Thanks, Matthew Brown mwbrown@sandia.gov From susanna@cse.lbl.gov Tue Apr 11 16:10:18 1995 From: susanna@cse.lbl.gov (Susanna Jacobson) Date: Tue Apr 11 16:10:24 PDT 1995 Subject: Re: Motorola S-records Matthew W. Brown (mwbrown@sandia.gov) writes: > WindRiver ships many of the gnu utilities with their VxWorks > distribution; however, they don't ship "objcopy". I suppose > that most prople never need objcopy. I only need it because > I am running as68k for another (non-VxWorks) board in my > system. Unfortunately, this board needs Motorola S-records. > With objcopy, one can easily convert a.out to S-records > > objcopy -O srec inputFile outputFile > > Is there another way to generate S-records? Maybe generating > them in the first place instead of converting them w/ objcopy??? Here is how I generate Motorola S-records for my non-VxWorks code: I use a linker script that specifies S-records as the output format. With a command line such as this I link object files into a single Motorola S-record file: ld68k -M -Tlink.cmd -Ttext 00004000 -Tdata 00006000 -Tbss 00008000 \ -o app.x app.o syslib.o > app.map (Note that I specify segment addresses and also generate a link map.) Here is the script I use (your address map may vary): --------------------------------------------------------- OUTPUT_FORMAT(srec) MEMORY { user : ORIGIN = 0x4000, LENGTH = 0x3fc000 } SECTIONS { .text : { *(.text) } > user .data : { *(.data) } > user .bss : { *(.bss) } > user } --------------------------------------------------------- There is also a Wind River utility, hex, that converts a.out files to S-records. It places the data segment immediately at the end of the text segment, which was unacceptable for my purposes. Hope this helps! ======================================================================== Susanna Jacobson MS 70A-3363 SRJacobson@LBL.Gov Lawrence Berkeley Laboratory VOICE: (510) 486-7801 1 Cyclotron Road FAX: (510) 486-7698 Berkeley, CA 94720 ======================================================================== From greg@eng.nmr.varian.com Tue Apr 11 16:44:26 1995 From: greg@eng.nmr.varian.com (Greg Brissey x6951) Date: Tue Apr 11 16:44:33 PDT 1995 Subject: Re: Motorola S-records Matthew Brown write: > > > Hopefully this is a basic question for someone out there ... > > WindRiver ships many of the gnu utilities with their VxWorks > distribution; however, they don't ship "objcopy". I suppose > that most prople never need objcopy. I only need it because > I am running as68k for another (non-VxWorks) board in my > system. Unfortunately, this board needs Motorola S-records. > With objcopy, one can easily convert a.out to S-records > > objcopy -O srec inputFile outputFile > > Is there another way to generate S-records? Maybe generating > them in the first place instead of converting them w/ objcopy??? > > I really was hoping to not have to build a gnu cross-compiler > for this board as I really already have one. Any easy ideas? > > Thanks, > > Matthew Brown > mwbrown@sandia.gov > Take a look at the makefile for the vxWorks kernel, when given the target vxWorks_rom.hex it creates an S-record file. Just use the utilities they use.. Greg Brissey Varian greg.brissey@nmr.varian.com From morriss@smtplink.indigo.co.il Tue Apr 11 23:23:10 1995 From: Steve Morris Date: Tue Apr 11 23:23:17 PDT 1995 Subject: VxWorks on PowerPC and Intel 80x86 - help needed Text item: Text_1 We are hotly debating whether to develop a new fairly heavy, interrupt driven, system on a PowerPC platform or on a Intel 80x86 platform. Either way we will be using VxWorks. I urgently need to talk anyone who has experience using VxWorks on Intel 80x86 or has already started to use VxWorks on PowerPC. If you can help, please let me have your phone number and I will contact you. ---------------------------------------------------------------------- Steve Morris voice: (972)-8-381055 Software Team Manager (972)-8-381818 Indigo Fax: (972)-8-381341 P.O. Box 150, Rehovot, Israel e-mail: morriss@indigo.co.il ---------------------------------------------------------------------- From daemon@vxw.ee.lbl.gov Wed Apr 12 04:00:31 1995 From: daemon@vxw.ee.lbl.gov Date: Wed Apr 12 04:00:40 PDT 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Wed Apr 12 04:00:23 PDT 1995 Subject: Re: VxWorks dying in semQPut... Subject: VxWorks 1553 drivers Subject: How to free memory/interrupt vector used by intConnect() Subject: Help! Memory artifacts after task death! Subject: Re: C++ and taskSpawn Subject: Re: VxWorks dying in semQPut... ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks dying in semQPut... Date: Tue, 11 Apr 1995 09:08:15 GMT From: jfhughes@jtechcons.win.net (John F. Hughes) Message-ID: <207@jtechcons.win.net> Reply-To: jfhughes@jtechcons.win.net (John F. Hughes) Here is what I have found to date; 1. The chip we are using in our Ethernet board is an AMD 7990 (lance). The lance chip/driver combination in vxWorks has several SPRs against it. Wind Rivers attitude basically seems to be that the chip has problems, it's not their fault. Documentation for the 79C90 (CMOS version of the 7990 with numerous improvements) indicates that it has corrections for a transmit lockout and receive lockup problem. These problems could cause the symptoms we are seeing, but resetting the lance chip does not unstick the task waiting in semQPut. 2. Using the shell to find the semaphore that semQPut is waiting for and "giving" this semaphore using semGive, does not unstick the network. 3. Not only is the task that is doing the socket communication affected, but any "network" task gets stuck the same way. Our application can have multiple streams open at the same time. Once any one of the tasks suspends and stops receiving data, all the tasks that are handling the streams stop. No network task will work from this point on, including telnet, rsh, rlogin, etc. I have not tested NFS but I suspect it would get stuck as well. 4. The problem seems to occur less frequently on faster boards. 5. The problem is more frequent if the lance chip's access to memory is slowed down by CPU and/or VME access to the memory. 6. Nothing we have tried so far allows us to continue network operations once the problem has occured. Any one have any more ideas? --------------------------- Newsgroups: comp.os.vxworks Subject: VxWorks 1553 drivers Date: Tue, 11 Apr 1995 16:19:33 +0000 From: currie@asd1.nasa.jsc.gov (Alan Currie) Organization: NASA EV5 Message-ID: Followup-To: comp.os.vxworks Does anyone know of an archive of 1553 drivers that will work with VxWorks? I'm interested in RS422 also. Please respond to Alan Currie currie@asd1.jsc.nasa.gov Thanks. --------------------------- Newsgroups: comp.os.vxworks Subject: How to free memory/interrupt vector used by intConnect() Date: 11 Apr 1995 18:29:10 +1000 From: dmr@f111.iassf.easams.com.au (Dewan RASHID) Organization: EASAMS (Australia) Pty Ltd Message-ID: <3mdekm$6ej@f111.iassf.easams.com.au> Dear VxPerts: It appears to me that whenever intConnect(VOIDFUNCPTR.....) is executed it takes 44 bytes of memory and never released. I can free this allocated memory by executing the following function calls: FUNCPTR intVecGet(FUNCPTR* vector) /* returns pointerToISR */ free(FUNCPTR pointerToISR) intVecSet(FUNCPTR* vector, FUNCPTR excIntStub) Now is there any other way to do the same? Anything like intFree(..)? Any ideas will be appreciated. Thanks, Dewan RASHID. --------------------------- Newsgroups: comp.os.vxworks Subject: Help! Memory artifacts after task death! Date: 12 Apr 1995 02:16:39 GMT From: jjwang@clark.net (Jason J Wang) Organization: Clark Internet Services, Inc., Ellicott City, MD USA Message-ID: <3mfd67$589@clarknet.clark.net> I've got tasks that hang on to some one or two megabytes of allocated memory after program exit. Is this strictly a matter of not freeing dynamically allocated memory before program exit? Or is it perhaps due to a possible occurrence of a task getting moduleDeleted before the task actually dies? Any info would be greatly appreciated. Is there an exit hook that I may implement to completely free any memory used by the dying task? E-mail or post. I can be reached at jjwang@clark.net um, thank you very very very much. - -jason "uncertain career future" wang --------------------------- Newsgroups: comp.os.vxworks Subject: Re: C++ and taskSpawn Date: 11 Apr 1995 15:42:21 GMT From: billr@saab.CNA.TEK.COM (Bill Randle) Organization: Tektronix, Inc., Redmond, OR Message-ID: <3me80t$lvn@marlyn.cna.tek.com> References: <9503077972.AA797281065@mailhost.scires.com> <3m4sd5$3dp@f111.iassf.easams.com.au> <3mcfl3$7js@marlyn.cna.tek.com> Reply-To: billr@saab.CNA.TEK.COM I have to apologize for a slip of the eyeballs in my previous post - it's been awhile since I looked at the Process class code and I missed the fact that even with the class encapsulation, there is still a friendly static function involved in the task creation that gets passed the 'this' pointer of the process to spawn. For those interested in an example of process encapsulation, this is a snippet of code that deals with the process creation and initialization. The other stuff that's not shown is that we actually keep a list of all processes that are created then call a single static function activateAll() to activate all of the processes at the same time - after all the constructors and startup code has run. Again my apologies for any confusion my previous post may have caused. -Bill Randle Tektronix, Inc. billr@saab.CNA.TEK.COM - ----------------------- class Process { private: int tid; WIND_TCB tcb; Boolean wasactivated; char *name; public: Process(char *name, int priority, int userStackSize, Boolean fpucode=FALSE, Boolean stdiocode=TRUE); ~Process(); // Delete the process and remove it from the process list. protected: friend void Process_spawn(int); // Calls run() originally virtual void run(); // This member is called after the process is activated, and should // be an infinite loop representing the function of the process. public: STATUS activate(); // Start execution of this process (i.e. start // execution of the code defined by the member function named 'run') static void activateAll(); // Activate all the processes in the list which have been // constructed, but which haven't been activated so far. /** other stuff deleted **/ }; Process::Process(char* procname, int priority, int stacksize, Boolean fpucode, Boolean stdiocode) { STATUS status; int options = VX_DEALLOC_STACK; if (fpucode) options |= VX_FP_TASK; if (stdiocode) options |= VX_STDIO; name = (char*) malloc(strlen(procname)+1); strcpy(name, procname); wasactivated = !spawn; char* stk_base = (char*) malloc(stacksize); stk_base = (char*) ((int) stk_base + stacksize); if (spawn) { status = taskInit(&tcb, name, priority, options, stk_base, stacksize, (FUNCPTR) Process_spawn, (int) this, 0, 0, 0, 0, 0, 0, 0, 0, 0); tid = (int) &tcb; // Risky. WRS does not guarantee tid is & of tcb. tcb.spare1 = (int) this; // Used by _curProc } else { status = taskPrioritySet(tid, priority); } } void Process_spawn(int procid) { ((Process*) procid)->run(); } STATUS Process::activate() { STATUS err = 0; if (!wasactivated) { err = taskActivate(tid); } wasactivated = TRUE; return err; } /*virtual*/ void Process::run() { fprintf(stderr, "Process::run() : no run defined for process %s.\n", name); for (;;) taskDelay(10); } /** other stuff deleted **/ - -- -Bill Randle Tektronix, Inc. billr@saab.CNA.TEK.COM --------------------------- Newsgroups: comp.os.vxworks Subject: Re: VxWorks dying in semQPut... Date: Tue, 11 Apr 1995 15:59:40 GMT From: jfhughes@jtechcons.win.net (John F. Hughes) Message-ID: <212@jtechcons.win.net> References: <3lppfb$q3q@lucy.infi.net> <3m0sn3$6is@nrtphba6.bnr.ca> Reply-To: jfhughes@jtechcons.win.net (John F. Hughes) In article , Ed Goforth (egoforth@netcom.com) writes: : original post deleted : >I believe that this is mentioned in the 5.1.1 release notes among "Known >problems". The select() call messes up the semaphore information tables (or >something like that). I don't have the book in front of me for the exact >explanation. As I recall, there is no known workaround mentioned in the >documentation. > >Ed... > >======================================================================== >Edward R. Goforth | egoforth@netcom.com | GO >Software Engineer | | TEAM >Digital System Resources, Inc. | egoforth@dsrnet.com | Ada >-- >Edward R. Goforth | egoforth@netcom.com | GO >Software Engineer | | TEAM >Digital System Resources, Inc. | egoforth@dsrnet.com | Ada > This is good to know, however our application is not using select. (At least I don't think it is). We still get stuck under as yet unknown conditions. thanks for the input. jfh --------------------------- End of New-News digest ********************** From daveg@datasrv.co.il Thu Apr 13 02:52:57 1995 From: David Gasul Date: Thu Apr 13 02:53:06 PDT 1995 Subject: hashLib, sllLib and other undocumented libraries Hello VxWorkers, Does anybody know where information re. undocumented libraries of VxWorks (i.e. not in manuals and vxman) can be found? The interesting information is obviously the APIs and semantics, but also the status of these libraries. Specifically, we're currently interested in the hashLib and sllLib libraries. Many thanks in advance, David Gasul Email: daveg@zeus.datasrv.co.il Telegate Ltd. Voice: +972-3-5388634 Aryt Building FAX: +972-3-5335877 7 Haplada St. 60218 Or-Yehuda, ISRAEL From daemon@vxw.ee.lbl.gov Thu Apr 13 04:00:34 1995 From: daemon@vxw.ee.lbl.gov Date: Thu Apr 13 04:00:52 PDT 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Thu Apr 13 04:00:20 PDT 1995 Subject: ACM Workshop on Languages, Compilers & Tools for Real-Time Systems Subject: Re: Motorola S-records Subject: Re: Unix filesystem port to VxWorks wanted Subject: Re: Flash-Writing Routines for Intel 28F008SA Subject: Re: console ? Subject: Proxy ARP w/o if_sm? ------------------------------------------------------- Newsgroups: comp.realtime,comp.lang.ada,comp.specification,comp.os.lynx,comp.os.qnx,comp.os.vxworks,comp.os.os9,comp.os.chorus,comp.os.mach Subject: ACM Workshop on Languages, Compilers & Tools for Real-Time Systems Date: 11 Apr 1995 23:31:15 -0400 From: rich@cs.umd.edu (Richard Gerber) Organization: U of Maryland, Dept. of Computer Science, Coll. Pk., MD 20742 Message-ID: <3mfhi3$5hh@coltrane.cs.umd.edu> ============================================================================ ACM SIGPLAN Workshop on Languages, Compilers and Tools for Real-Time Systems (LCT-RTS '95) (In Conjunction with ACM SIGPLAN PLDI/PEPM) La Jolla, California June 21-22, 1995 Program and Registration Information ------------------------------------- ACM SIGPLAN LCT-RTS '95 is an interface between two dynamic fields of computer science and engineering: programming languages and real-time systems. LCT-RTS provides a forum where top researchers in these areas can share their results and directions, and where they can form new collaborations based on common interests. LCT-RTS '95 directly follows the conclusion of PLDI '95. The workshop program includes 16 presentations on new research results, and one panel session on the topic of Software Interfaces. WEDNESDAY, June 22 - ------------------- 12:00 Conclusion of PLDI '95 1:30 - 2:00 Introduction to LCT-RTS '95 2:00 - 3:30 Session I: Scheduling & Timing Analysis o T.-Y. Huang, J. Liu, (Illinois, USA), "Predicting the worst-case execution time of the concurrent execution of instructions and cycle-stealing DMA I/O operations" o J. Lee, S. Lee, H. Kim, (Korea Telecomm & Kyunghee U, Korea), "Scheduling of hard-aperiodic tasks in hybrid static/dynamic priority systems" o K. D. Nielsen, B. Rygg, (Iowa State, USA), "Worst-case execution time analysis on modern processors" 3:30 - 4:00 Break 4:00 - 6:00 Session II: Language Design & Development o F. Bonfatti, G. Gadda, P. D. Monari, (U Modena & CMA Spa, Italy), "Re-usable software design for programmable logic controllers" o T. Chung, H. Dietz, (Purdue, USA), "Language constructs and transformation for hard real-time systems" o S. Ren, G. Agha, (Illinois, USA), "RTsynchronizer: Language support for real-time specifications in distributed systems" o F. Thoen, M. Cornero, G. Goossens, H. De Man, (IMEC, Belgium), "Software synthesis for real-time information systems" 6:00 - 8:00 Dinner (on your own) 8:00 -10:00 Discussions/Birds-of-a-feather THURSDAY, June 22 - ----------------- 7:30 - 8:30 Continental Breakfast 8:30 - 10:00 Session III: Analysis Methods o S. Campos, E. Clarke, W. Marrero, M. Minea, (Carnegie Mellon, USA), "Verus: A tool for quantitative analysis of finite-state real-time systems" o M. Jourdan, F. Maraninchi, (VERIMAG, France), "Static timing analysis of real-time systems" o Y. Li, S. Malik, (Princeton, USA), "Performance analysis of embedded software using implicit path enumeration" 10:00 - 10:30 Break 10:30 - 12:00 Session IV: Tools o L. Ko, D. Whalley, (Florida State, USA), "Supporting user-friendly analysis of timing constraints" o H. Kopetz, Nossal, (Tech U Wien, Austria), "The Cluster compiler -- a tool for the design of time-triggered real-time systems" o D. Wilner, (Wind River Systems, USA), "WindView: A tool for understanding real-time embedded software through system visualization" 12:00 - 1:30 Lunch 1:30 - 3:00 Panel: Appropriate Interfaces Between Design Tools, Languages Compilers and Runtimes in Real-Time Systems Development - Where are the connections and separations? - How can a separation of concerns be maintained? 3:00 - 3:30 Break 3:30 - 5:00 Session V: Compilers o F. Mueller, (Humboldt-Universitaet zu Berlin, Germany), "Compiler support for software-based cache partitioning" o A. Bakkers, J. Sunter, E. Ploeg, (U. Twente, Netherlands), "Automatic generation of scheduling and inter-process communication code in parallel real-time programs for transputers" o S. Schneider, V. Chen, J. Steele, G. Pardo-Castellote, (Real-Time Innovations Inc, USA), "The ControlShell component-based real-time programming system, and its applications to the Marsokhod Martian rover" 5:00 - 5:30 Wrap-up =================================================================== PROGRAM COMMITTEE Alan Burns (University of York) Richard Gerber, Co-Chair (University of Maryland) Rajiv Gupta (Univ of Pittsburgh) Mary Hall (Caltech) Connie Heitmeyer (Naval Research Lab) Insup Lee (University of Pennsylvania) Al Mok (University of Texas at Austin) Thomas Marlowe, Co-Chair (Seton Hall University, NJIT RTCL) Steve Tjiang (Synopsys Inc.) PROGRAM CO-CHAIRS Richard Gerber Thomas Marlowe Department of Computer Science Department of Mathematics University of Maryland Seton Hall University College Park, MD 20742, USA South Orange NJ 07079, USA telephone: +1 301 405 2710 telephone: +1 201 761 9784 fax: +1 301 405 6707 fax: +1 201 761 9596 rich@cs.umd.edu marlowe@cs.rutgers.edu =================================================================== ACM Programming Languages Summer Extravaganza --- ----------- --------- ------ ------------ June 18-28, 1995, Hyatt Regency Hotel, La Jolla, California Registration information ACM SIGPLAN Conference on Programming Language Design and Implementation (PLDI '95) 18-21 June 1995 ACM SIGPLAN Workshop on Language, Compiler, and Tools for Real-Time Systems (LCT-RTS '95) 21-22 June 1995 ACM SIGPLAN Symposium on Partial Evaluation and Semantics-Based Program Manipulation (PEPM '95) 21-23 June 1995 Haskell Workshop 25 June 1995 SIGPLAN/SIGARCH/WG2.8 Conference on Functional Programming Languages and Computer Architecture (FPCA '95) 25-28 June 1995 All information is available via the World-Wide Web at http://www.research.digital.com/wrl/projects/sigplan95/ or via ftp in the directory gatekeeper.dec.com:/pub/DEC/sigplan95/* - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - PLDI/PEPM/FPCA '95 Hotel Reservations Mention "Association for Computing Machinery" to receive the conference rates, valid if you register by 28 May 1995. By Mail By Phone Hyatt Regency La Jolla Voice: (800) 233--1234 3777 La Jolla Village Drive +1 (619) 552--1234 San Diego, CA 92122 Fax: +1 (619) 552--6066 Single or Double rate: $107 (+ 10.5% state tax) Name(s):_________________________________________________________________ Affiliation:_____________________________________________________________ Address:_________________________________________________________________ _________________________________________________________________________ Phone:____________________________ Fax:__________________________________ arrival date________________________________ number of nights____________ number of rooms________________________ number of people_________________ Room type (King bed or Two double beds)__________________________________ Smoking or Non-smoking?__________________________________________________ Special Needs:___________________________________________________________ Guarantee room by credit card? (VISA, MC, AMEX, Diners, Discover) ___Visa ___Mastercard ___American Express ___Diners Club ___Discover Card number:_______________________________________ Expires______________ Signature:_______________________________________________________________ Registration Information Submit one photocopy of the form for each attendee. Please print or type all information to avoid errors on your badge. To qualify for the lower rates, registration forms with full payment must be postmarked by May 25, 1995. * Please make checks and money orders payable in U. S. dollars to ACM SIGPLAN '95. * Phone, e-mail, and faxed registrations must be paid by Visa, MasterCard, or American Express. American Express users, please provide your billing address if it is different from the address on the registration form. * Government vouchers and purchase orders are not accepted. Requests for refunds must be postmarked by June 7, 1995. After this date, cancellations and no-shows are liable for the full fees. PLDI '95 registration includes a copy of the PLDI proceedings, Tuesday night's catered Gab-Fest, 3 continental breakfasts, coffee breaks, and lunch Monday and Tuesday of the conference. PEPM '95 registration includes a copy of the PEPM proceedings, the banquet, 2 continental breakfasts, coffee breaks, and lunch Thursday and Friday of the conference. FPCA '95 registration includes a copy of the FPCA proceedings, admission to the FPCA Tutorial, the Sunday evening reception, 3 continental breakfasts, coffee breaks, and lunch Monday and Tuesday of the conference. PLDI Tutorial registration includes a copy of the notes, lunch Sunday, and coffee breaks. Real-Time Workshop registration includes a copy of the proceedings preprints, lunch Thursday, and coffee breaks. Haskell Workshop registration includes a copy of the notes and coffee breaks. Send registration form and full payment to: ACM SIGPLAN '95 P. O. Box 8304 Maitland, FL 32794-8304 For overnight mail: or Fax to: (407) 628-3186 Carole Mann Phone: (407) 628-3602 ACM SIGPLAN '95 E-mail: mann@cs.ucf.edu Registration Systems Lab 2060 Goldwater Court Maitland, FL 32751 Name: ___________________________________________________________________ Name tag should read:____________________________________________________ Affiliation: ____________________________________________________________ Address: ________________________________________________________________ _________________________________________________________________________ Phone: ______________________________ Fax: _____________________________ Electronic mail: ________________________________________________________ ACM Membership Number: __________________________________________________ The list of attendees will be sent electronically, only to attendees. May we include you on this list? ___yes ___no Dietary preference? ___Vegetarian (may contain dairy, eggs) ___Vegan (no animal products) ___Kosher Special needs or accommodations: ________________________________________ Fee Schedule (in U. S. dollars) please circle ALL applicable fees Tutorial PLDI'95 Real-Time Early Late Early Late Workshop ACM & SIGPLAN member 150 17 275 300 90 ACM or SIGPLAN member 175 200 300 325 90 Non-member 175 200 325 350 100 full-time student 60 60 125 125 PEPM '95 FPCA '95 Haskell Early Late Early Late Workshop ACM & SIGPLAN member 310 385 310 310 30 ACM or SIGPLAN member 310 385 325 325 30 Non-member 385 460 350 450 30 full-time student 150 150 150 150 Students are welcome at the workshops, but there is no reduced fee. Total (Conferences, Tutorial, Workshops):________________________________ ___check (US$, payable to SIGPLAN '95) ___Visa ___MasterCard ___American Express Card #: _________________________________________ Expires _______________ Signature: ______________________________________________________________ Signature of e-mail registrants will be required at the conference. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Motorola S-records Date: Wed, 12 Apr 1995 13:25:51 GMT From: clegg@ssd.fsi.com Organization: Flightsafety International, Broken Arrow, OK Message-ID: <1995Apr12.132551.17710@dev1.ssd.fsi.com> References: <199504112156.PAA21250@somnet.sandia.gov> Reply-To: clegg@ssd.fsi.com Sender: news@dev1.ssd.fsi.com In <199504112156.PAA21250@somnet.sandia.gov>, "Matthew W. Brown" writes: > >Hopefully this is a basic question for someone out there ... > >WindRiver ships many of the gnu utilities with their VxWorks >distribution; however, they don't ship "objcopy". >Is there another way to generate S-records? Maybe generating >them in the first place instead of converting them w/ objcopy??? You can use the "hex" function which shipped with your WRS software. Its sole purpose is to convert a.out to S-Record, however, it screws up when resolving links to initialized data. I solved this problem by sed'ing the assembly source generated by gcc, replacing .data with text. I don't find this a particularly appealing solution, but it works and for my application the exact location of initialized variables doesn't matter. >I really was hoping to not have to build a gnu cross-compiler >for this board as I really already have one. Any easy ideas? The latest version of GNU ld68k supports direct S-record generation using something like - -oformat srec. Unfortunately, the version we obtained from Hundred Acres Consulting (HAC) had problems with its object module format which rendered it useless for a.out generation :( (when it generated jump tables for switch() statements it created links to them which WRS loader couldn't resolve, among other things). HAC agreed that a fix was needed, but has since ceased answering our calls or mail. FYI clegg@ssd.fsi.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Unix filesystem port to VxWorks wanted Date: Wed, 12 Apr 1995 09:23:24 GMT From: prayson@eltec.win-uk.net (Eltec International PLC) Message-ID: <65@eltec.win-uk.net> References: <199504071537.AA24253@darya.wrs.com> Reply-To: prayson@eltec.win-uk.net (Eltec International PLC) In article , Suhas Joshi (joshisx@ntmtv.com) writes: >In article <199504071537.AA24253@darya.wrs.com>, >Darren Cathey wrote: >>Hello: >> >>I have a customer who is looking for a Unix filesystem >>port to VxWorks. If anyone knows of such a beast for >>sale or free via the archive, could you please provide >>me with the appropriate pointers. > >I have a doubt whether Unix FS is suitable enough for real >time applications. The file system can get easily currupted >if machine is switched on and off a few times. > >Can any one clarify this. > >SJ > > > > Unix FS is robust enough if write caching is disabled, but v. slow. Darren, why do you need a Unix FS ? Surely how the data is stored on the local HDD makes no difference. Unless you are planning to share devices on the SCSI bus of course, but I would strongly advise against using a Unix FS for this. (From past experience I would advise against this anyway :-)) Peter. - --- Eltec International PLC Tel: +44 1908 366499 (a subsidiary of Eltec Elektronik GmbH) Fax: +44 1908 274600 Milton Keynes, MK7 8LF, UK email: support@eltec.win-uk.net --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Flash-Writing Routines for Intel 28F008SA Date: Wed, 12 Apr 1995 14:39:50 GMT From: gerlach@netcom.com (Matthew H. Gerlach) Organization: Netcom Online Communications Services (408-241-9760 login: guest) Message-ID: References: <9504111456.AA26431@tcville.ES.HAC.COM> Sender: gerlach@netcom16.netcom.com Hi, I am currently doing a lot of work with i960 boards. One of the nice things about this processor is that Intel was nice enough to port GCC and tools to the i960. Furthermore they freely distribute a ROM monitor that has the same functionality as VxMon. Part of this ROM monitor has code for writing flash. Included below is flash.c from Intel. If you are interested in getting at all of their stuff, try ftp'ng to ftp.intel.com. Good Luck, Matthew H. Gerlach ******************************************************* * Gerlach Computer Consulting * * * * Real-Time Embedded Systems * * Networking Protocols * * Software Development Environments * * Network and Sun System Administration * * * * gerlach@netcom.com * * * ******************************************************* /*(cb*/ /******************************************************************************* * * Copyright (c) 1993, 1994 Intel Corporation * * Intel hereby grants you permission to copy, modify, and distribute this * software and its documentation. Intel grants this permission provided * that the above copyright notice appears in all copies and that both the * copyright notice and this permission notice appear in supporting * documentation. In addition, Intel grants this permission provided that * you prominently mark as "not part of the original" any modifications * made to this software or documentation, and that the name of Intel * Corporation not be used in advertising or publicity pertaining to * distribution of the software or the documentation without specific, * written prior permission. * * Intel Corporation provides this AS IS, WITHOUT ANY WARRANTY, EXPRESS OR * IMPLIED, INCLUDING, WITHOUT LIMITATION, ANY WARRANTY OF MERCHANTABILITY * OR FITNESS FOR A PARTICULAR PURPOSE. Intel makes no guarantee or * representations regarding the use of, or the results of the use of, * the software and documentation in terms of correctness, accuracy, * reliability, currentness, or otherwise; and you rely on the software, * documentation and results solely at your own risk. * * IN NO EVENT SHALL INTEL BE LIABLE FOR ANY LOSS OF USE, LOSS OF BUSINESS, * LOSS OF PROFITS, INDIRECT, INCIDENTAL, SPECIAL OR CONSEQUENTIAL DAMAGES * OF ANY KIND. IN NO EVENT SHALL INTEL'S TOTAL LIABILITY EXCEED THE SUM * PAID TO INTEL FOR THE PRODUCT LICENSED HEREUNDER. * ******************************************************************************/ /*)ce*/ /* This driver can be used for any board which has a single bank of * Flash EPROM. The memory may be 1, 2, or 4 bytes wide. The driver * will erase and program all devices in parallel. * The include file 'this_hw.h' must define the following symbols: * FLASH_ADDR - the base address of the Flash memory * FLASH_WIDTH - the number of devices which are accessed in parallel * * PROC_FREQ - the processor clock frequency in MHz * or * TIMER_CNTL - the address of the 8254 timer control register * TIMER_0 - the address of the 8254 count register for timer 0 * TIMER_XTAL - the frequency in Hz of the timer crystal * * If PROC_FREQ is not defined, timer 0 of the 8254 is used to calibrate * the timing delays. The timer is used only during initialization; it is * available to the application after that. * To use this driver with a board which has more than one bank of * flash, you must modify the code in init_eeprom() which determines the * flash type and the code in several other places which validates the * address and length of the region to be programmed or erased. */ #include "common.h" #include "i960.h" #include "hdi_errs.h" #include "hdi_arch.h" #include "this_hw.h" #if FLASH_WIDTH == 4 typedef unsigned long flash_type; #define MASK 0xffffffff #elif FLASH_WIDTH == 2 typedef unsigned short flash_type; #define MASK 0xffff #elif FLASH_WIDTH == 1 typedef unsigned char flash_type; #define MASK 0xff #endif /* FLASH_WIDTH */ #define READ_CMD (0x00000000 & MASK) #define WRITE_CMD (0x40404040 & MASK) #define STOP_CMD (0xc0c0c0c0 & MASK) #define ERASE_CMD (0x20202020 & MASK) #define VERIFY_CMD (0xa0a0a0a0 & MASK) #define RESET_CMD (0xffffffff & MASK) #define READ_ID_CMD (0x90909090 & MASK) #define ERASED_VALUE (0xffffffff & MASK) /* ASSEMBLY INLINE */ #if defined (__GNUC__) || defined(GNU_ASM) __inline static void delay(int loops ) { int cnt; asm volatile ( "mov %1,%0" : "=d"(cnt) : "d"(loops) ); asm volatile ( "0: cmpdeco 0,%1,%0; bl 0b" : "=d"(cnt) : "0"(cnt) ); } #elif defined(_MCC960) static void delay(loops) { asm("loop: mov cmpdeco 0,`loops`,`loops`"); asm(" bl loop"); } #else /*intel*/ asm void delay(int loops) { %void return; tmpreg loops; label loop; loop: cmpdeco 0, loops, loops bl loop %error; } #endif /* __GNUC__ */ unsigned long eeprom_size; ADDR eeprom_prog_first, eeprom_prog_last; static int fltype; static long timeout_6; /* number of loops needed for 6uS timeout */ static long timeout_10; /* number of loops needed for 10uS timeout */ static long timeout_100; /* number of loops needed for 100uS timeout */ static long timeout_10000; /* number of loops needed for 10mS timeout */ static int init_loopcnt(); static long loopcnt(int t); static int program_zero(ADDR addr, unsigned long length); static int program_word(ADDR addr, flash_type data, flash_type mask); /********************************************************/ /* INIT FLASH */ /* */ /* This routine initializes the variables for timing */ /* with any board configuration. This is used to get */ /* exact timing every time. */ /********************************************************/ void init_eeprom() { volatile flash_type *check; /* Set defaults */ fltype = 0; eeprom_size = 0; /* set up timing loop numbers */ /* If init_loopcnt returns -1, the processor architecture * is unknown and timing cannot be done. */ if (init_loopcnt() < 0) return; timeout_6 = loopcnt(6); timeout_10 = loopcnt(10); timeout_100 = loopcnt(100); timeout_10000 = loopcnt(10000); check = (flash_type *)FLASH_ADDR; *check = RESET_CMD; /* Reset device, in case it was being */ *check = RESET_CMD; /* programmed when the board was reset. */ *check = READ_ID_CMD; /* Read-Intelligent-Identifier command */ /* find out if devices are really Flash */ if (check[0] == (0x89898989 & MASK)) { /* find out which type of device is present */ switch ( check[1] ){ case 0xB1B1B1B1 & MASK: case 0xB2B2B2B2 & MASK: fltype = 256; /* 256 Kbit Flash */ eeprom_size = 0x8000 * FLASH_WIDTH; break; case 0xB9B9B9B9 & MASK: fltype = 257; /* 256 Kbit Flash new process */ eeprom_size = 0x8000 * FLASH_WIDTH; break; case 0xB8B8B8B8 & MASK: fltype = 512; /* 512 Kbit Flash */ eeprom_size = 0x10000 * FLASH_WIDTH; break; case 0xB4B4B4B4 & MASK: fltype = 1024; /* 1 Mbit Flash */ eeprom_size = 0x20000 * FLASH_WIDTH; break; case 0xBDBDBDBD & MASK: fltype = 2048; /* 2 Mbit Flash */ eeprom_size = 0x40000 * FLASH_WIDTH; break; default: /* fltype was set to 0 above */ break; } } *check = READ_CMD; } /********************************************************/ /* IS EEPROM */ /* Check if memory is Flash */ /* */ /* returns TRUE if it is; FALSE if not eeprom ; */ /* returns ERROR bad addr or partial eeprom */ /* */ /********************************************************/ int is_eeprom(ADDR addr, unsigned long length) { ADDR eeprom_end = FLASH_ADDR + eeprom_size - 1; ADDR block_end = addr + length - 1; /* Check for wrap: if the address and length given wrap past * the end of memory, it is an error. */ if (block_end < addr) return ERR; if (addr >= FLASH_ADDR && block_end <= eeprom_end) return TRUE; if (addr > eeprom_end || block_end < FLASH_ADDR) return FALSE; /* If the block was partly within the Flash, it is an error. */ return ERR; } /********************************************************/ /* CHECK EEPROM */ /* Check if Flash is Blank */ /* */ /* returns OK if it is; returns ERROR and sets cmd_stat */ /* to an error code if memory region is not Flash or if */ /* it is not blank. */ /* */ /********************************************************/ int check_eeprom(ADDR addr, unsigned long length) { unsigned char *p, *end; if (fltype == 0) { cmd_stat = E_NO_FLASH; return ERR; } if (addr == NO_ADDR) { addr = FLASH_ADDR; length = eeprom_size; } else if (length == 0) length = 1; if (is_eeprom(addr, length) != 1) { cmd_stat = E_EEPROM_ADDR; return ERR; } p = (unsigned char *)addr; end = p + length; /* find first non_blank address */ while (p != end) { if (*p != 0xff) { cmd_stat = E_EEPROM_PROG; eeprom_prog_first = (ADDR)p; /* find last non_blank address */ for (p = end; *--p == 0xff; ) ; eeprom_prog_last = (ADDR)p; return ERR; } p++; } return OK; } /********************************************************/ /* ERASE EEPROM */ /* */ /* returns OK if successful; otherwise returns ERROR */ /* and sets cmd_stat to an error code */ /* */ /* Flash Programming Algorithm */ /* Copyright (c) 1988, Intel Corporation */ /* All Rights Reserved */ /********************************************************/ int erase_eeprom(ADDR addr, unsigned long length) { flash_type *end_addr; flash_type value = 0; volatile flash_type *erase_data; flash_type mask; long cumtime, erase_time; long erase_timeout; int error, ERRNUM; if (fltype == 0) { cmd_stat = E_NO_FLASH; return ERR; } if (addr != NO_ADDR && (addr != FLASH_ADDR || (length != 0 && length != eeprom_size))) { cmd_stat = E_EEPROM_ADDR; return ERR; } /* If it's already erased, don't bother. */ if (check_eeprom(FLASH_ADDR, eeprom_size) == OK) return OK; /* set up number of errors permissible for Flash type */ if (fltype == 256) ERRNUM = 64; else ERRNUM = 1000; erase_time = 10000; /* 10 mS */ erase_timeout = timeout_10000; cumtime = 0; /* cumulative erase time */ error = 0; /* no cumulative errors */ /* program device to all zeros */ if (program_zero(FLASH_ADDR, eeprom_size) != OK) { cmd_stat = E_EEPROM_FAIL; return (ERR); } erase_data = (flash_type *)FLASH_ADDR; end_addr = (flash_type *)(FLASH_ADDR + eeprom_size); mask = MASK; while (1) { *erase_data = ERASE_CMD & mask; /* set up erase */ *erase_data = ERASE_CMD & mask; /* begin erase */ delay(erase_timeout); /* wait XX mS */ /* check for devices being erased */ while (erase_data != end_addr) { *erase_data = VERIFY_CMD; /* end and verify erase */ delay(timeout_6); /* wait 6 uS */ value = *erase_data; if (value != ERASED_VALUE) break; erase_data++; } /* if completely erased */ if (erase_data == end_addr) break; /* check for max errors */ if (++error == ERRNUM) { *erase_data = READ_CMD; cmd_stat = E_EEPROM_FAIL; return (ERR); } /* erase only the invalid banks */ mask = 0; if ((value & 0x000000ff) != 0x000000ff) mask |= 0x000000ff; #if FLASH_WIDTH > 1 if ((value & 0x0000ff00) != 0x0000ff00) mask |= 0x0000ff00; #if FLASH_WIDTH > 2 if ((value & 0x00ff0000) != 0x00ff0000) mask |= 0x00ff0000; if ((value & 0xff000000) != 0xff000000) mask |= 0xff000000; #endif /* FLASH_WIDTH > 2 */ #endif /* FLASH_WIDTH > 1 */ /* For the 28F256, the erase time changes each time. */ if (fltype == 256) { /* count cumulative erase time */ cumtime += erase_time; /* adjust erase time */ erase_time = cumtime / 8; erase_timeout = loopcnt(erase_time); } } erase_data = (flash_type *)FLASH_ADDR; *erase_data = READ_CMD; return OK; } /********************************************************/ /* PROGRAM ZEROS */ /* */ /* Returns OK if successful; otherwise returns ERROR. */ /* */ /* Flash Programming Algorithm */ /* Copyright (c) 1988, Intel Corporation */ /* All Rights Reserved */ /********************************************************/ static int program_zero(ADDR addr, unsigned long length) { unsigned long i; /* begin writing */ for (i = 0; i < length; i += FLASH_WIDTH, addr += FLASH_WIDTH) if (program_word(addr, 0, MASK) != OK) return ERR; return OK; } /********************************************************/ /* WRITE EEPROM */ /* */ /* returns OK if successful; otherwise returns ERROR */ /* and sets cmd_stat to an error code */ /* */ /* Flash Programming Algorithm */ /* Copyright (c) 1988, Intel Corporation */ /* All Rights Reserved */ /********************************************************/ int write_eeprom(ADDR start_addr, const void *data_arg, int data_size) { const flash_type *dataptr = data_arg; flash_type data, mask; int i, shift, leftover; if (fltype == 0) { cmd_stat = E_NO_FLASH; return ERR; } /* need to make sure the address begins on a word boundary */ /* program bytes until it is there */ #if FLASH_WIDTH > 1 /* find number of bytes done */ shift = start_addr % FLASH_WIDTH; if (shift != 0) { /* get starting address to write extra bytes */ start_addr -= shift; data = *dataptr << 8*shift; mask = MASK << 8*shift; if (data_size < FLASH_WIDTH-shift) mask &= MASK >> 8*(FLASH_WIDTH-(shift+data_size)); if (program_word(start_addr, data, mask) != OK) return ERR; /* fix start address for rest of bytes */ start_addr += FLASH_WIDTH; dataptr = (const flash_type *) ((const char *)dataptr + FLASH_WIDTH-shift); data_size -= FLASH_WIDTH - shift; if (data_size <= 0) return OK; } /* write all full width words possible */ /* find out how many leftover bytes must be written at end */ leftover = data_size % FLASH_WIDTH; data_size -= leftover; #endif /* FLASH_WIDTH > 1 */ for (i=0; i 1 /* write out leftover bytes */ if (leftover) { mask = MASK >> 8*(FLASH_WIDTH-leftover); data = *dataptr & mask; if (program_word(start_addr, data, mask) != OK) return ERR; } #endif /* FLASH_WIDTH > 1 */ return OK; } static int program_word(ADDR addr, flash_type data, flash_type mask) { volatile flash_type *write_addr = (flash_type *)addr; long standby; flash_type verify_data; int error; if (fltype == 256) standby = timeout_100; else standby = timeout_10; error = 0; while (1) { *write_addr = WRITE_CMD & mask; /* send write cmd */ *write_addr = data & mask; /* write data */ delay(standby); /* standby for programming */ *write_addr = STOP_CMD & mask; /* send stop cmd */ delay(timeout_6); /* wait 6 uS */ verify_data = *write_addr & mask; if (verify_data == (data & mask)) /* verify data */ break; if (++error == 25) { *write_addr = READ_CMD; cmd_stat = E_EEPROM_FAIL; return (ERR); } /* mask off valid data */ if ((verify_data & 0x000000ff) == (data & 0x000000ff)) mask &= 0xffffff00; #if FLASH_WIDTH > 1 if ((verify_data & 0x0000ff00) == (data & 0x0000ff00)) mask &= 0xffff00ff; #if FLASH_WIDTH > 2 if ((verify_data & 0x00ff0000) == (data & 0x00ff0000)) mask &= 0xff00ffff; if ((verify_data & 0xff000000) == (data & 0xff000000)) mask &= 0x00ffffff; #endif /* FLASH_WIDTH > 2 */ #endif /* FLASH_WIDTH > 1 */ } *write_addr = READ_CMD; return OK; } #ifdef PROC_FREQ /* These definitions give the number of clocks per loop in the delay routine. */ #define CA_CLOCKS 2 #define KX_CLOCKS 4 /* PROC_FREQ is the processor clock frequency, in MHz. It is defined in the */ /* board-specific include file. (If the processor is running at 25 MHz, */ /* PROC_FREQ is 25, not 25000000.) */ /* The variable proc_freq is global so it can be changed in init_hardware if */ /* desired. */ int proc_freq = PROC_FREQ; static int clocks; static int init_loopcnt() { switch (arch) { case ARCH_CA: clocks = CA_CLOCKS; break; case ARCH_KA: case ARCH_KB: case ARCH_SA: case ARCH_SB: clocks = KX_CLOCKS; break; default: /* This can only happen if a new processor type is defined. */ /* At this point in the code, we have no way to issue a message. */ return -1; } return 0; } /* Return the delay constant required to delay t us. */ static long loopcnt(int t) { return (t * proc_freq) / clocks; } #else /* PROC_FREQ */ static long time(int loops); static long loops_ratio; /* Initialize the loop ratio, which is used by loopcnt to calculate the * delay constant for any length of time. */ static int init_loopcnt() { /* Since this is all integral arithmetic, the units must be small * enough to keep rounding error within reason. That is why all * calculations are done it ns. */ /* time() returns the time in ns for the given loop count */ long time1000 = time(1000); /* loops_ration is ns per loop */ loops_ratio = time1000 / 1000; return 0; } /* Return the delay constant required to delay t us. */ static long loopcnt(int t) { long t_ns = (long)t * 1000; return 1 + (t_ns / loops_ratio); } /* * return mul1 * mul2 / divisor, using a 64-bit multiply to avoid overflow */ #if defined (__GNUC__) || defined(GNU_ASM) __inline static unsigned emuldiv(unsigned mul1, unsigned mul2, unsigned divisor) { struct { unsigned low; unsigned high; } tmp; struct { unsigned remainder; unsigned quotient; } result; asm("emul %1, %2, %0" : "=d"(tmp) : "d"(mul1), "d"(mul2)); asm("ediv %1, %2, %0" : "=d"(result) : "d"(divisor), "d"(tmp)); return result.quotient; } #elif defined(_MCC960) static unsigned emuldiv(unsigned mul1, unsigned mul2, unsigned divisor) { struct{unsigned remainder; unsigned quotient;} result; asm(" mov `mul2`,r4"); asm(" emul `mul1`,r4,r6"); asm(" ediv `divisor`,r6,r6"); asm(" mov r6,`result`"); return result.quotient; } #else /*intel*/ asm unsigned emuldiv(unsigned mul1, unsigned mul2, unsigned divisor) { %tmpreg return; reglit mul1; reglit mul2; reglit divisor; tmpreg(2) tmp; emul mul1, mul2, tmp ediv divisor, tmp, tmp mov tmp(1), return # return quotient; ignore remainder %error } #endif /* __GNUC__ */ /* This routine uses timer 0 of the 8254 timer to determine how * fast the timing loop is. It is called only during initialization. After * that all timing is done with delay loops. */ static long time(int loops) { unsigned char lsb, msb; int ticks; volatile unsigned char dummy; /* init 8254 timer */ *(volatile unsigned char *)TIMER_CNTL = 0x30; /* set up timer 0 */ dummy = 0; /* Avoid two consecutive writes to hw registers. */ *(volatile unsigned char *)TIMER_0 = 0xff; /* counter lsb */ dummy = 0; *(volatile unsigned char *)TIMER_0 = 0xff; /* counter msb */ delay(loops); /* latch and read timer */ *(volatile unsigned char *)TIMER_CNTL = 0x00; /* latch timer */ dummy = 0; /* Avoid two consecutive writes to hw registers. */ lsb = *(volatile unsigned char *)TIMER_0; msb = *(volatile unsigned char *)TIMER_0; /* figure out total number of ticks */ ticks = 0xffff - ((msb << 8) + lsb); /* Return the time in ns */ return (long)emuldiv(ticks, 1000000000, TIMER_XTAL); } #endif /* PROC_FREQ */ - -- ******************************************************* * Gerlach Computer Consulting * * * * Real-Time Embedded Systems * * Networking Protocols * * Software Development Environments * * Network and Sun System Administration * * * * gerlach@netcom.com * * * ******************************************************* --------------------------- Newsgroups: comp.os.vxworks Subject: Re: console ? Date: Wed, 12 Apr 95 11:26:20 -0500 From: Kelly Cochran Organization: Delphi (info@delphi.com email, 800-695-4005 voice) Message-ID: References: <3l5mdq$jo6@overload.lbl.gov> Kuang-chun Cheng writes: > We use a SGI indy to be a host computer of our DAQ system (mv167 on >VME crate). Usually, I use telnet to connect mv167. But when reset the >mv167 form SGI. The telnet secssion will be disconnect. I must wait for >mv167 re-boot and then telnet again. > Try using "cu". I was in a similar position as you about two months ago when I had to switch from a SUN to an SGI onyx. John --------------------------- Newsgroups: comp.os.vxworks Subject: Proxy ARP w/o if_sm? Date: 12 Apr 1995 22:13:16 GMT From: bonaker@amasd.anatcp.rockwell.com (Joe Bonaker) Organization: Rockwell Message-ID: <3mhj9s$gu5@huron.eel.ufl.edu> Reply-To: bonaker@amasd.anatcp.rockwell.com In certain configurations, I need to run Proxy ARP across two network interfaces, neither of which is the shared memory (backplane) interface supported by if_sm. Apparently the VxWorks proxy ARP server verifies that the "proxy" network is a shared memory network. As such, proxyNetCreate(2) against one of the Ethernet interfaces fails with an errno of S_proxyArpLib_INVALID_INTERFACE. Is this correct, i.e., does VxWorks' proxy ARP server only support operation over the shared memory network? If so, how can proxy ARP be implemented across two non-shared memory networks? Does anyone have static {arp,route}{Add,Delete}(2) workaround examples to share? Thanks, Joe B. --------------------------- End of New-News digest ********************** From nps@radstone.co.uk Thu Apr 13 07:18:34 1995 From: nps@radstone.co.uk (Nigel P Street) Date: Thu Apr 13 07:18:41 PDT 1995 Subject: dosFsLib and Floppies Hello, There's been talk recently of DOS format hard disks and possible incompatibilities. What is the score with floppy drives? Is it reasonable to assume that floppies formatted from DOS or VxWorks are fully interchangeable between PCs and VxWorks? TIA, Nige St ------------------------------------------------------------------------- | Nigel P Street C Eng Radstone Technology PLC | | Principal Software Engineer Water Lane | | Towcester voice : +44 (0)1327 359444 | | Northants fax : +44 (0)1327 358112 | | NN12 6JN e-mail : nps@radstone.co.uk | | England | ------------------------------------------------------------------------- From mea@mclean.sparta.com Thu Apr 13 07:53:19 1995 From: mea@mclean.sparta.com (Mike Anderson) Date: Thu Apr 13 07:53:26 PDT 1995 Subject: Re: dosFsLib and Floppies Greetings! > Submitted-by: nps@radstone.co.uk (Nigel P Street) > > Hello, > > There's been talk recently of DOS format hard disks and possible > incompatibilities. What is the score with floppy drives? Is it > reasonable to assume that floppies formatted from DOS or VxWorks > are fully interchangeable between PCs and VxWorks? > I've experienced no difficulties exchanging floppies made under VxWorks 5.1.X and MSDOS when using the Teac FD235 3.5" floppy drives. Regards, =============================================================================== __ Real-Time System Development, Integration, Training and Services //\\ // \\ Mike Anderson // /\ \\ Chief Engineer // / \ \\ SPARTA, Inc. Voice : (703) 448-0210 ext. 235 // \ \\ 7926 Jones Branch Drive FAX : (703) 734-3323 \\ \ // Suite 900 EMAIL : mea@mclean.sparta.com \\ \ / // McLean, VA 22102 \\ \/ // "Software development is like making \\ // a baby... You can't make a baby in one \\// month by impregnating nine women. -- "Pride in Performance" Some things just take time." =============================================================================== From sblachma@aoc.nrao.edu Thu Apr 13 10:47:40 1995 From: Steve Blachman Date: Thu Apr 13 10:47:46 PDT 1995 Subject: Bug in taskNameToId ()? I hate to blame the OS, but I am getting a Bus Error inside a call to taskNameToId (). It happens in a call to strcmp () in which the first pointer is out of the address space, and the second argument is the pointer to the name passed to taskNameToId (). This happens at a time of mass task dieoffs and I suspect that the tcb dissapears while taskNameToId () is using it. The task name passed does not exist (the code is trying to be sure of that) and the error is only occasional, < 1% of the calls. What I would like to know is if anyone has seen this before or if it is a known problem. I can work around it, but I do not want to cover up any problems in the application which this might be showing. Thanks, Steven Blachman From daemon@vxw.ee.lbl.gov Fri Apr 14 04:00:29 1995 From: daemon@vxw.ee.lbl.gov Date: Fri Apr 14 04:00:37 PDT 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Fri Apr 14 04:00:19 PDT 1995 Subject: Re: Proxy ARP w/o if_sm? Subject: VxWORKS PPP Subject: Getting back memory when closing pipes Subject: Renaming symbols ? Subject: Motorola M68360QUADS-040 eval board Subject: WANTED: Utility to flag non-reentrant functions. ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Proxy ARP w/o if_sm? Date: Thu, 13 Apr 1995 08:43:36 GMT From: rsgawera@ss11.wg.icl.co.uk (Raj Gawera) Organization: International Computers Limited Message-ID: References: <3mhj9s$gu5@huron.eel.ufl.edu> Reply-To: rsgawera@ss11.wg.icl.co.uk Sender: news@oasis.icl.co.uk Hi Joe, I had a similar desire to run Proxy ARP but without using shared memory networks. This is the reply I got from Wind River :- !!!Start of insert!!! > Dear VxWorks Support, > I have been looking at using VxWorks PROXY ARP ? I have looked in the > manual but I am a bit confused. The manual section on configuring Proxy ARP > keeps on referring to the shared memory network. I have two networks on > my VxWorks card, one ethernet and one proprietry network (for which I have > written my own driver). I wish to use Proxy ARP to route packets between both > interfaces. Is this possible or does Proxy ARP on VxWorks only apply when > one interface is the shared memory network ? The proxyNetCreate function does check for backplane interface and returns an error if the interface is not either "bp" or "sm". I guess the proxy arp could run with other interfaces, but this need modifications in the source code. > > If you have any other tips/hints/notes on Proxy ARP on VxWorks I would be > most grateful. > You could try, but with NO GARANTY at all, to rename you second interface as "sm" or "bp". That way you will pass the following test (within proxyNetCreate): /* proxy only for backplane interfaces */ if (proxyArpParanoia && (strcmp (pIfa->ifa_ifp->if_name, "bp") != 0) && (strcmp (pIfa->ifa_ifp->if_name, "sm") != 0)) { logMsg ("proxying for non backplane device\n", 0, 0, 0, 0, 0, 0); return (ERROR); } But you may (or may not) run into serious network problem latter on. Hope this help, Regards, Jean-Marc De Maraumont. !!!End of Inser!!! So you can see that vxWorks checks the name of the interface. You could rename your interface but I wouldn't advise it. We got round this problem by implementing our own ARP server which did accessed a lookup table it gets from another server. This involved playing around at the device driver level although you could use EtherHooks to achieve the same affect. cheers, Raj email : rsgawera@wg.icl.co.uk --------------------------- Newsgroups: comp.os.vxworks Subject: VxWORKS PPP Date: 13 Apr 1995 15:03:16 -0500 From: sureshr@bnr.ca (Marco Marchetti) Organization: Bell-Northern Research, Richardson, TX Message-ID: <3mk024$ob5@crchha82.bnr.ca> Hello Folks, We are trying to establish a communication between the VxWORKS PPP and the SUN Solaris 2.4 PPP. Is there anyone who was successful in establishing a communication between these two. Many thanks in advance. Balaji Holur --------------------------- Newsgroups: comp.os.vxworks Subject: Getting back memory when closing pipes Date: 13 Apr 1995 21:27:19 GMT From: Ma.Marino@agora.stm.it Organization: I.Net S.p.A. Message-ID: Getting back memory when closing pipes I currently use VxWorks v.5.0.2b on some Force SYS68k CPU-40 B/16 boards. I have noticed a behaviour not acceptable in the application I have to develop. I did a test as follows. First I create and open some pipes by: pipeDevCreate (pipename, pipesize, msgsize); fd = open (pipename, O_RDWR); so that 2 Mbytes of memory are allocated for those pipes. Afterwards, I close the pipes by: close(fd); and I see that all the close's succeded but the memory is still allocated and never freed. I use memShow to see the amount of free and allocated memory. If I repeat the process I get the same result with less memory. And so on. I also tried a FIOFLUSH with ioctl before the close's, but nothing changed. My questions are: is this behaviour correct, and how can I free the memory allocated to the pipes by pipeDevCreate when I close them ? Please, if possible answer directly to me through email. Thanks a lot, Massimo Marino --------------------------- Newsgroups: comp.os.vxworks Subject: Renaming symbols ? Date: Thu, 13 Apr 1995 14:59:38 GMT From: rvdouder@ms.philips.nl (Rob van der Ouderaa) Organization: Philips Medical Systems, Best Message-ID: <1995Apr13.145938.3705@ms.philips.nl> Sender: news@ms.philips.nl My problem is the following. A computer that uses vxworks has a global symbol table. All vxworks processes share this symbol table. So if I have an application that I want to port to vxworks, potential symbol table clashes can occure between processes already running under vxworks and the newly ported application. Is there some tool to make the symbols of the ported application unique under vxworks ? Can you add a unique prefix to the symbols of the ported application for example ? (Not for all symbols, for example printf should stay the same but all symbols defined by the ported application should be unique) All input will be highly appreciated, Rob van der Ouderaa (rvdouder@ms.philips.nl) --------------------------- Newsgroups: comp.os.vxworks Subject: Motorola M68360QUADS-040 eval board Date: 13 Apr 1995 17:26:24 GMT From: Eric Oosterhof Organization: PSI Public Usenet Link Message-ID: <3mjms0$dvj@pubxfer2.news.psi.net> We at RadiSys are desperately attempting to purchase or rent a QUADS-040 board. While we have an order in with a vendor, the time lag that is being imposed on us by off the shelf items not being on the shelves is rapidly reducing the usefulness of the purchase. Consequently, I am hoping that someone out there has one of these boards that you are no longer using or are willing to part with for a period of about 2 months. This board has a MC68040 with a MC68360 running in 68040 companion mode. It is not the dual MC68360 with one QUICC in master and the other in slave mode. If you have one of these boards, and can part with it, either temporarily or as a sale, please E-mail me at Eric.Oosterhof@RadiSys.com or call at the numbers below. Our news feed is as yet not very robust, so email is preferable. Eric Oosterhof Eric.Oosterhof@RadiSys.com (503) 646-1800 x7495 (503) 646-1850 (Fax) --------------------------- Newsgroups: comp.realtime,comp.os.vxworks Subject: WANTED: Utility to flag non-reentrant functions. Date: Fri, 14 Apr 1995 03:01:33 GMT From: mlfdyn@world.std.com (Mark Fisher) Organization: The World Public Access UNIX, Brookline, MA Message-ID: Hi, Does anyone know of a utility that will parse C code and flag functions that may be non-reentrant. Thanks for any suggestions. - ----------------------------------------------------------------------------- Mark Fisher mfisher@dyn.com --------------------------- End of New-News digest ********************** From greg@eng.nmr.varian.com Fri Apr 14 09:09:23 1995 From: greg@eng.nmr.varian.com (Greg Brissey x6951) Date: Fri Apr 14 09:09:30 PDT 1995 Subject: Re: comp.os.vxworks newsdigest Massimo Marino writes: > ..... >.....and how can I free the memory allocated to the pipes by pipeDevCreate > when I close them ? >.... I had the same problem, and the following solved it for me. I am Using 5.1.1 so I hope you have these functions. Use iosDevFind to obtain the DEV_HDR of the device. Then use iosDevDelete(DEV_HDR *mypipe) to free the resources. E.G. DEV_HDR *mypipeDevHdr char *pNameTail mypipeDevHdr = iosDevFind("/pipe/mypipe",&pNameTail); iosDevDelete(mypipeDevHdr); pNameTail = '' if an exact match is found I hope this helps. Greg Brissey Varian greg.brissey@nmr.varian.com From burt@xylan.com Fri Apr 14 11:44:22 1995 From: burt@xylan.com (Burt Cyr) Date: Fri Apr 14 11:44:28 PDT 1995 Subject: Re: Renaming symbols ? Hi Rob - you wrote: > My problem is the following. A computer that uses vxworks has a > global symbol table. All vxworks processes share this symbol table. > So if I have an application that I want to port to vxworks, > potential symbol table clashes can occure between processes already > running under vxworks and the newly ported application. Actually, only the tshell (which is basically a symbol interpreter) uses this table. Programs that are loaded via the tshell have their unresolved externals patched using this table. If you define your own printf, for example, in your downloadable image, your code will use it, but other code which was linked into the kernel (had printf satisfied at link time) will not use your new one. However, code that is loaded after your module, will use your printf. > Is there some tool to make the symbols of the ported application unique > under vxworks ? Can you add a unique prefix to the symbols of the ported > application for example ? (Not for all symbols, for example printf should > stay the same but all symbols defined by the ported application should > be unique) Unfortunately, the lack of coding conventions in vxWorks has bit us too. This problem is known as the "weak externals" problem. To solve this, we adopted a coding convention early on whereby each public symbol has a unique prefix - usually two letters - to avoid clashes of this type. I don't know of any tool that your using but one thing you could do is post process the link map output from your loader into a sed script and pass sed over your modules. Hope this helps. Burt ================================================================== v _ ____(_)___ Burt Cyr Xylan, Inc. _-_-__-_-_-_ burt@xylan.com 26679 Agoura Rd. Suite 100 --_-_--_-_- (818) 880-3500 x3533 Calabasas, CA 91302 ================================================================== From cook@positron.med.ge.com Fri Apr 14 14:44:54 1995 From: cook@positron.med.ge.com (Mike Cook) Date: Fri Apr 14 14:45:00 PDT 1995 Subject: anyone using mv177 BSP with INCLUDE_SCSI? Is anyone out there successfully using the WRS supplied BSP for the mv177 and defining INCLUDE_SCSI? I recently got the BSP for eval, and am trying to run it on a 32 meg mv177 under VxWorks 5.1.1. If I define INCLUDE_SCSI & SCSI_AUTO_CONFIG, I get the following when scsiAutoConfig(pSysScsiCtrl) is invoked by tRootTask: Attaching network interface ei0... done. Attaching network interface lo0... done. Loading... 513260 + 33040 + 43634 Starting at 0x1000... Auto-configuring SCSI bus... Trap to uninitialized vector number 329 (0-255). Invalid ESF type 0x9 Task: 0x1fffe80 "tRootTask" Can anyone shed any light? Mike Cook cookm@med.ge.com From daemon@vxw.ee.lbl.gov Sat Apr 15 04:00:26 1995 From: daemon@vxw.ee.lbl.gov Date: Sat Apr 15 04:00:35 PDT 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sat Apr 15 04:00:17 PDT 1995 Subject: Press Release: NobleNet Releases New EZ-RPC for VxWorks Subject: Re: Dos file system memory leak and Memory fragmentation Subject: Wanted: driver for 85c30 in SDLC mode Subject: OS for intel based realtime audio server? Subject: X.25 / Ethernet ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Press Release: NobleNet Releases New EZ-RPC for VxWorks Date: Fri, 14 Apr 1995 15:49:14 GMT From: artc@world.std.com (Art Campbell) Organization: The World Public Access UNIX, Brookline, MA Message-ID: PRESS RELEASE NobleNet Ships EZ-RPC Development Tools For Wind River Systems' VxWorks Facilitates Distribution of Embedded Client/Server Applications Atlanta, GA. - April 18, 1995 - NobleNet, Inc. today announced availability of the latest version of its award-winning EZ-RPC client/server development tool for Wind River Systems' embedded real-time operating system, VxWorks. EZ-RPC lets developers distribute applications across VxWorks clients and servers as well as mix VxWorks environments in heterogeneous networks of UNIX platforms and workstations; Windows, Windows/NT, OS/2, and Macintosh clients and servers; and NetWare local area networks. EZ-RPC uses proven, cost-effective remote procedure call (RPC) technology. "We see the market for distributed application development and deployment in the embedded systems market expanding rapidly," said Dennis Ford, President and CEO of NobleNet. "NobleNet is pleased to be working with the industry leader. NobleNet's tools will help Wind River grow its customer base by providing the tools to integrate VxWorks applications across the many platforms typically found in enterprise environments." VxWorks customers develop applications for a variety of embedded products including telephone switch systems, factory-floor automation and robotics, and medical imaging database engines. With EZ-RPC, VxWorks applications can be distributed across heterogeneous environments without changing the existing network hardware and software infrastructure. The EZ-RPC family of compilers automatically generates both the client and server source codes, representing interactions between clients and servers as procedure calls across a network. EZ-RPC also provides a solution for distributing existing applications without re-writing or modifying source code. "The use of RPCs is proven technology for distributing applications across a variety of desktop platforms," said Stephen Li, vice president of business development for Wind River Systems. "However, programming with RPCs can be time-consuming. With NobleNet's EZ-RPC, embedded system developers can quickly migrate VxWorks-based applications to a client-server architecture, without any re-programming. Once again, VxWorks users are reaffirming Wind River's leadership position by pioneering client-server applications on real- time platforms." Wind River Systems, Inc. develops, markets, and supports software for developing real-time embedded applications. Its extensive WindPower product line enables customers to standardize their designs across products and enjoy a brief time-to-market. Wind River is a major supplier to the medical instrumentation, automotive, telecommunications, networking, industrial automation, electronics, and aerospace industries. Wind River is headquartered in Alameda, Calif. with offices throughout the United States, and overseas operations in France, Germany, the UK, Scandinavia, Japan, and the Asia-Pacific region. NobleNet's EZ-RPC tools have been recognized by industry press and analysts for their excellence; they are emerging standards in client/server application development. Based in Southboro, MA, NobleNet offers total one-stop shopping for multi-platform middleware products and services. In addition to its successful commercialization of RPC technology, NobleNet sells and services IONA Technologies' Orbix CORBA-compliant multi-platform Object Request Broker and provides technical support for its customers' complete middleware needs in both the procedural and object paradigms. NobleNet products are installed in more than 200 development sites worldwide, with resulting NobleNet run-time code supporting more than 300,000 clients and 27,000 servers. For more information, media representatives can contact: Dennis Byron at NobleNet Phone: 508-460-8222 E-mail: byron@noblenet.com Stephanie Schwarz at Wind River Systems Phone 510-814-2573 Email:steph@wrs.com NobleNet and EZ-RPC are trademarks of NobleNet Inc. All other product and company names mentioned herein may be registered trademarks and /or trademarks of their respective companies. - -- Art Campbell artc@world.std.com 72227.1375@compuserve.com "... In my opinion, there's nothing in this world beats a '52 Vincent and a redheaded girl." -- Richard Thompson No disclaimers apply. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Dos file system memory leak and Memory fragmentation Date: Fri, 14 Apr 1995 18:30:55 GMT From: kent@wrs.com (Kent Long) Organization: Wind River Systems, Inc. Message-ID: References: <3kii36$s2t@asgard.mlb.dmt.csiro.au> Sender: news@wrs.com (News Manager) Hi, Folks - Although it's been a few weeks since the original post, I thought I'd provide some comments on the memory issues with dosFs. Hao Nguyen (hao@mlb.dmt.csiro.au) writes: >We use VxWorks 5.1 on a 68040 board. We used HeapTrace to and found >that the DOS file system leaks 80 bytes for every open() call. After a >running for a while, the system memory get very fragmented, memShow() >shows the available memory around say 1.7MB and the maximum contiguous >space is only sa y 60K. > > This is bad because our DOS cluster size is around 63K and an open() >on a DOS file requires a contiguous space of the cluster size, 63k! >Consequently our system failed when the max contiguous space gets below >the DOS cluster size. The memory in question is the file system's internal version of the directory entry for a file. These individual entries are kept in linked lists defining directories, and these lists are in turn maintained in a tree representing the directory hierarchy. The problem at hand stems from the persistence of these directory entries. When a file is deleted under regular MS-DOS, its entry in the directory remains present on the disk, it is merely marked as deleted (by overwriting the first character in the name). The VxWorks dosFsLib does the same with its internal entries; so, the entry for the deleted file remains allocated. This is what was described as a "leak," but in fact the memory is not really lost. It will be re-used when the next file is created in that directory. The memory will only be freed when the directory is removed. However, the complaint that this leads to fragmentation is entirely valid. Because these small (80 bytes in 5.1.1) allocations are persistent, and because there are other allocations by dosFs that *are* in fact freed once the file is closed and/or deleted, the system memory partition becomes fragmented. Unfortunately, there are not any desirable workarounds. One which was described in a followup posting is to insert a new version of malloc()/calloc() which checks the size and allocates from a dedicated partition if the allocation is for 80 bytes. However, this is not an easy proposition. What dosFs *should* do is to use a dedicated partition itself for these allocations. Unfortunately, this is still not the case in the new 5.2 release, either. The more hopeful news is that Wind River is currently working on a set of enhancements that will not only address this fragmentation issue, but will also greatly reduce the overall memory consumption by the file system. This new support will allow pre-allocation of memory pools for both directory and FAT buffering, which will no longer be fully represented in memory. (This will be an option; maintaining a full memory image will still be possible for implementations where determinism outweighs ram usage.) The release vehicle for this dosFs work has not been finalized, but look for it to be available later this year. Anyway, I hope this helps clarify things a bit. Bye for now, - kent Kent Long Wind River Systems kent@wrs.com --------------------------- Newsgroups: comp.os.vxworks Subject: Wanted: driver for 85c30 in SDLC mode Date: Fri, 14 Apr 1995 18:39:25 GMT From: gerlach@netcom.com (Matthew H. Gerlach) Organization: NETCOM On-line Communication Services (408 261-4700 guest) Message-ID: Sender: gerlach@netcom14.netcom.com Hello fellow embedded hackers, I am looking for some driver code for a Zilog 85c30 SCC in SDLC mode. This code does not have to be for VxWorks. Any ideas as to where I can get some? Many Thanks, Matthew H. Gerlach P.S. I will happily mail a 32 oz. bottle of my finest home made beer to the first individual who gives me C-code or a pointer to code that pans out! - -- ******************************************************* * Gerlach Computer Consulting * * * * Real-Time Embedded Systems * * Networking Protocols * * Software Development Environments * * Network and Sun System Administration * * * * gerlach@netcom.com * * * ******************************************************* --------------------------- Newsgroups: comp.os.misc,comp.realtime,comp.os.vxworks,comp.os.qnx,fj.comp.dev.scsi Subject: OS for intel based realtime audio server? Date: 14 Apr 1995 18:58:42 GMT From: elvis@well.sf.ca.us (Michael Taylor) Organization: The Whole Earth 'Lectronic Link, Sausalito, CA Message-ID: <3mmgl2$5fn@nkosi.well.com> We're looking for an operating system suitable for a realtime audio server system which has the following qualities: * standard IBM PC (486 or greater) compatible * able to handle alot of realtime ISA traffic (wavefiles/control) to the multiple soundcards (via shared memory and I/O ports) * able to use SCSI card as a comm device to a host computer. The PC will virtually be a SCSI peripheal to the host - this implies that the PC/OS/SCSI card must be able to to handle 'SCSI target mode.' We expect this will be a very complex com system whereby the host will be able to send multiple wavefiles and control to/from the PC concurrently, and in realtime. * Easy to develop for and debug. Gnu based dev tools would be nice. * Simple TCP/IP networking would be a plus, too, but not necessary. * oh, and a cheap/free runtime license. any ideas? thanks for any response, - -andrew --------------------------- Newsgroups: comp.os.vxworks Subject: X.25 / Ethernet Date: Fri, 14 Apr 1995 19:22:43 GMT From: arlen_baker@email.mot.com (Arlen Baker) Organization: Motorola GSTG Message-ID: <1995Apr14.192243.5783@schbbs.mot.com> Sender: news@schbbs.mot.com (SCHBBS News Account) I have a requirement to execute the X.25 protocol over an Ethernet bus on a VME 68040 board running VxWorks. The requirement stems from a protocol requirement coupled with a speed and throughput requirement (real-time environment). Does anyone know: 1) If this can be accomplished and if so, how? 2) If anyone has done this and iff so, what's his/here phone number???!!! Any and all help would be greatly appreciated. --------------------------- End of New-News digest ********************** From daemon@vxw.ee.lbl.gov Mon Apr 17 04:00:25 1995 From: daemon@vxw.ee.lbl.gov Date: Mon Apr 17 04:00:32 PDT 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Mon Apr 17 04:00:17 PDT 1995 Subject: GNU Compiler ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: GNU Compiler Date: Sun, 16 Apr 95 22:28:29 PDT From: "Mark A. Menge" Organization: ICNET... Your Link To The Internet... +1.313.998.0090 Message-ID: * Does anyone know if the GNU compiler is capable of compiling for 680x0 code in DOS ? Mark Menge --------------------------- End of New-News digest ********************** From stan@lassen.rti.com Mon Apr 17 08:15:36 1995 From: stan@lassen.rti.com (Stan Schneider) Date: Mon Apr 17 08:15:42 PDT 1995 Subject: Press Release: NDDS =============================================================================== A N N O U N C E M E N T =============================================================================== Real-Time Innovations Announces NDDS ------------------------------------ April 18, 1995 --- Real-Time Innovations, Inc. (Sunnyvale, CA) today advanced real-time network communications with a novel system called the "Network Data Delivery Service". The Network Data Delivery Service (NDDS) makes it easy to share data and event information among distributed systems without additional programming. Its subscription communications are a revolutionary new way to share data on a network. Designed especially for complex connectivity applications, subscription architectures are much more modular and robust than client-server grids. They naturally handle communications failures and network changes. Moreover, subscribers do not have to re-request data for each transaction, making NDDS much more efficient than client-server designs. With a subscription architecture, nodes on the network independently "subscribe to" and "produce" data items by name. Each time an item is produced, the NDDS agent distributes it to all subscribers. NDDS supports multiple consumers and producers. Its real-time communications model lets NDDS support real-time many-to-many communications transparently. NDDS provides update guarantees, deadlines, "hot standby" and override capability. No request traffic and no central server make subscription updates fast. Anonymous subscriptions make the system modular. Transparent multiple nodes and on-line reconfiguration make the system robust. And, automatic code generation makes NDDS easy to use. A typical application that would have taken thousands of lines of network or RPC code can be done with a few NDDS library calls. In addition, NDDS will support replicated nodes (e.g. many GUIs), be robust to network changes and failures, run faster, allow moving programs from processor to processor transparently, run immediately on all flavors of Unix and VxWorks, and be much more modular and understandable. To help blast off NDDS and VxWorks 5.2, center stage at the Wind River booth at the ESC features a big screen demonstration of VxWorks release 5.2 and NDDS running on eight different target boards. The demonstration underscores Wind River's committment to timely release of new OS features on all architectures simultaneously, and RTI's committment to offer the best distributed multi-target and multi-Unix communications package on the market. Drop by and see it! NDDS will change the way you think about distributed applications. --- For further information, contact "info@rti.com", or call (408) 720-8312. Extremely aggressive academic discount program available. =============================================================================== = = = = Real-Time Innovations, Inc. = Phone: (408) 720-8312 = = 954 Aster, Sunnyvale, CA 94086 = email: info@rti.com = = = = =============================================================================== From cliao@me.umn.edu Mon Apr 17 11:02:31 1995 From: "Chen-Fu Liao (MD)" Date: Mon Apr 17 11:02:38 PDT 1995 Subject: Load booting parameters to EPROM, How? Dear Sir, Does any one know how to load the booting parameters to EPROM for mv133a board? I have mv167 as the master and want 2 mv133a to be the slaves for multi-processor by using VxMp. We have mv133XT, too. Can I find some booting file in the mv133XT and download the file to EPROM? How? Thanks for your help! ChenFu Liao ------------------------------------------------------------------ _/_/_/ _/_/ _/_/ _/_/ _/_/_/ _/_/ _/_/_/ _/ _/ _/ _/ _/ _/ _/ _/ _/ _/ _/ _/ _/ _/ _/ _/ _/_/ _/ _/ _/ _/ _/ _/ _/ _/_/_/_/ _/ _/ _/ _/ _/ _/_/_/_/ _/ _/_/_/ _/ _/ _/ _/ _/_/_/ _/ _/ _/_/_/ Robotics Laboratory ME55 TEL: 612-625-5561 Department of Mechanical Engineering FAX: 612-625-8884 University of Minnesota E-mail: cliao@me.umn.edu ------------------------------------------------------------------ From daemon@vxw.ee.lbl.gov Tue Apr 18 04:03:59 1995 From: daemon@vxw.ee.lbl.gov Date: Tue Apr 18 04:04:08 PDT 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Tue Apr 18 04:02:23 PDT 1995 Subject: Re: Mupltiple writes to a socket Subject: Problems with fprintf ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Mupltiple writes to a socket Date: Mon, 17 Apr 1995 08:43:25 -0700 From: cobarruvias@asd1.jsc.nasa.gov (John R. Cobarruvias) Organization: NASA/JSC Message-ID: References: <3m118c$6is@nrtphba6.bnr.ca> In article <3m118c$6is@nrtphba6.bnr.ca>, billag@b4pphff (Bill Gutknecht) wrote: > If two tasks of equal priority both peform a write() operation on > a socket simultaniously, what happens? Is this a quiz? If so, both task will never simultaniously perform a write, and since the tasks have the same priority they execute in FIFO order. > > Bill > > ------------------------------------------------------------------------------ > Bill Gutknecht "If I die, I will go before Crom and he will > BNR/Northern Telecom ask me 'What is the Riddle of Steel?' If I > Research Triangle Park, NC do not know it, he will cast me out of > billag@bnr.ca Valhalla and laugh at me ... " > ------------------------------------------------------------------------------ _ _ _ / _ /_ _ /_/ / ` _ /_ _ _ _ . _ _ (_/ /_// // / / \ /_, /_//_//_|/ / /_/|// /_|_\ John R. Cobarruvias cobarruvias@asd1.jsc.nasa.gov --------------------------- Newsgroups: comp.os.vxworks Subject: Problems with fprintf Date: 17 Apr 1995 21:23:35 GMT From: nils@eecs.nwu.edu (Robert Nils Itschner) Organization: EECS Department, Northwestern University Message-ID: <3mum8n$n6v@news.eecs.nwu.edu> Hello, I am having some random errors while using fprintf to write to a file over the Ethernet (from a MVME167). Sometimes characters turn up in the file instead of doubles as supposed. This happens only if I am using the application I am developing. If I am using a simple little programm to write numbers to a file these errors never occur. Only one task is ever writing to this file, so I am simply clueless what is causing these errors. Did anybody have similar problems ? Can anybody give me a hint what could be a reason for random errors like that ? As this is my first project working with VxWorks it could be a quite basic error as well.. Robert --------------------------- End of New-News digest ********************** From kmartin@astro.ge.com Tue Apr 18 11:49:46 1995 From: Kevin Martin Date: Tue Apr 18 11:49:53 PDT 1995 Subject: Quad Serial I/O card for VME Does anyone have knowledge of vendors which supply Quad Serial I/O Modules for VME architectures which have device drivers that are integrated with VxWorks version 5.1.1? The only vendor I've been able to find is APLabs (Xycom serial board). From kevinr@photon.ceco.com Tue Apr 18 13:24:54 1995 From: kevinr@photon.ceco.com (Kevin R. Rumbaugh) Date: Tue Apr 18 13:25:00 PDT 1995 Subject: Re: Quad Serial I/O card for VME > > >Does anyone have knowledge of vendors which supply Quad Serial I/O Modules for VME >architectures which have device drivers that are integrated with VxWorks >version 5.1.1? The only vendor I've been able to find is APLabs (Xycom serial board). > The Xycom 400 board uses the Zilog 8530 chip which is the same chip that is used on the MVME147 board from Motorola so the driver code is easily extensible to include the extra serial ports. I have done that in the past and it is pretty straigtforward. If you are not using the MVME147 in your particular application, but have the BSP for the 147. I would think that the serial driver could easily be installed as an extra driver on your board but only talk to the Xycom 400 card. -- Kevin Rumbaugh 'The Marauder' --- Commonwealth Edison --- A Unicom division E-MAIL: kevinr@photon.ceco.com AT&T : (312) 394-8894 FAX : (312) 394-4405 USMAIL: Process Computer Technologies;Room 1139;125 S. Clark;Chicago IL 60603 From 100067.355@compuserve.com Wed Apr 19 01:36:38 1995 From: Philippe Weber <100067.355@compuserve.com> Date: Wed Apr 19 01:36:46 PDT 1995 Subject: Re: Quad Serial I/O card for VME Kevin Martin wrote : > > >Does anyone have knowledge of vendors which supply Quad Serial I/O Modules for VME >architectures which have device drivers that are integrated with VxWorks >version 5.1.1? The only vendor I've been able to find is APLabs (Xycom serial board). > Themis has the TSVME500, a Quad Serial card based on 68681 DUARTs, with a VxWorks 5.1.1 driver associated. For info, please mail sales@themis.com. ************************************* * Philippe WEBER * * Product Marketing Manager * * THEMIS COMPUTER * * 1, rue des Essarts * * 38610 GIERES, FRANCE * * Voice : +33 76596046 * * Fax : +33 76630030 * * E-mail: 100067.355@CompuServe.com * ************************************* From sergi@hpbpq11.bpo.hp.com Wed Apr 19 02:10:44 1995 From: Sergi Casas(QA Software testing) Date: Wed Apr 19 02:10:51 PDT 1995 Subject: Re: problems with fprintf Robert Nils Itschner writes: > Hello, > I am having some random errors while using fprintf to write to a file over > the Ethernet (from a MVME167). Sometimes characters turn up in the file > instead of doubles as supposed. This happens only if I am using the application > I am developing. If I am using a simple little programm to write numbers to > a file these errors never occur. I'd try to check the stack: as fprintf needs memory to convert doubles into strings for writing them into the file, and it is also on top of a quite large hierarchy of routine calls (variable args management, string formating, C standard buffered files, OS I/O layers, file system, disk or NFS driver ...), it might take more stack than you expected and allowed for that task. Use checkStack and look at the "Margin" field. Otherwise, you might be simply overwriting the formatted string of fprintf from another task. Maybe you could "taskLock()" before the fprintf() or even modify the relative priorities of application tasks trying to control which tasks may be executing whenever fprintf blocks its calling task (for NFS file access, disk driver or anything else). Hope this helps. -- |============================================================================| | Sergi Casas ##### / ##### Firmware Engineer | | Hewlett-Packard Company ### /_ _ ### | | Barcelona Division (BCD) ## / / / / ## e-mail: sergi@hp-bpo.bpo.hp.com | | Avda. Graells, 501 ## / / /_/ ## Phone : +34 3 582 1475 | | 08190 S. Cugat (Barcelona) ### / ### Fax : +34 3 582 2515 | | SPAIN ##### / ##### HP-Telnet: 712 1475 | |============================================================================| From daemon@vxw.ee.lbl.gov Wed Apr 19 04:00:30 1995 From: daemon@vxw.ee.lbl.gov Date: Wed Apr 19 04:00:38 PDT 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Wed Apr 19 04:00:23 PDT 1995 Subject: Re: Quad Serial I/O card for VME ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Quad Serial I/O card for VME Date: Wed, 19 Apr 1995 03:03:31 GMT From: afraser@hookup.net (Andrew Fraser) Organization: HookUp Communication Corporation, Oakville, Ontario, CANADA Message-ID: References: <199504181851.OAA14002@a9972.ASTRO.GE.COM> In article <199504181851.OAA14002@a9972.ASTRO.GE.COM> Kevin Martin writes: >From: Kevin Martin >Subject: Quad Serial I/O card for VME >Date: Tue, 18 Apr 1995 14:51:06 -0400 >Does anyone have knowledge of vendors which supply Quad Serial I/O Modules for >VME >architectures which have device drivers that are integrated with VxWorks >version 5.1.1? The only vendor I've been able to find is APLabs (Xycom serial >board). Our SVME-739 would fit the bill. It's based a 68020, plus two 68302s, each providing three serial I/O channels. The '302s have internal DMAs and one of the 302s operates as a master to oversee interrupts and the DMA activity. The 68020 is the overall controller, running either your s/w or our Data Link Communications Suite (establish port, open session, session management, etc). We also have the VxWorks drivers to support it. Give one of our offices a call to get more information: DY 4 Systems Inc. DY 4 Systems Ltd. 21 Fitzgerald Road 1475 S. Bascom Ave., Suite 202 Nepean, Ontario Campbell, California 95008 Canada K2H 9J4 phone (408) 377-9822 phone (613) 596-9911 fax (408) 377-4725 fax (613) 596-0574 Regards - -Andy --------------------------------------------------------------------- ANDREW FRASER DY 4 Systems Inc **** *** Senior Hardware Designer 21 Fitzgerald Rd *** * *** *** *** Product Support Group Nepean, ON ** ** ** *** *** * Canada * *** * *** *** ** Email: afraser@dy4.com K2H 9J4 **** *** *** Tel: (613) 596-9922 ext 251 *** Fax: (613) 596-0574 CUSTOMER FIRST, QUALITY ALWAYS --------------------------- End of New-News digest ********************** From sidney@opstech.ca.boeing.com Wed Apr 19 06:58:41 1995 From: sidney@opstech.ca.boeing.com (Sidney Ly) Date: Wed Apr 19 06:59:34 PDT 1995 Subject: Re: Quad Serial I/O card for VME Matrix Corporation has this kind of board with driver for vxWorks. From leonid@rst.co.il Wed Apr 19 09:08:13 1995 From: leonid@rst.co.il (Leonid Rosenboim) Date: Wed Apr 19 09:08:25 PDT 1995 Subject: Re: Quad Serial I/O card for VME We recommend checking the C407 board from AItech, its got a 68030 as the main CPU, VIC/VAC VME interface, and up to four 302s, providing three serial channels each. The board runs VxWorks by means of an appropriate Board Support Package, which also includes a high-speed asynchronous driver for the 68302 on-board. Feel free to contact me for more details, or AItech at ----------------------------------------------------------------------- Leonid Rosenboim Phone: +972-3-559-8144 R S T Software Industries Ltd. Mobile: +972-50-307-142 3 Hashikma st. Industrial Zone, Fax: +972-3-559-8244 P.O.Box 11502, AZUR 58017, Israel E-Mail: leonid@rst.co.il From mpd@ssd.ray.com Wed Apr 19 12:31:04 1995 From: Dragicevich Date: Wed Apr 19 12:31:11 PDT 1995 Subject: SCSI tape question I am currently using VxWorks 5.1.1 on a MVME167 with an HP35470A DAT tape drive. I am getting a sustained write throughput of only 140 Kbytes/s, for block writes of 10,000 to 50,000 bytes. Running 'spy' while a task is continuously writing to the tape drive indicates that the task is only utilizing 5% of the CPU. The tape drive is being written while in variable mode. The tape drive uses SCSI-II and is spec'd at 5 Mbytes/s synchronous and 1.5 Mbytes/s asynchrounous, and we need at least a 250 Kbytes/s sustained rate. The SCSI bus also has a disk drive but no accesses are being performed to it during tape accesses. The Buffered Mode field of the Mode Select Parameter is set to 1 as suggested in the tape drive interface manual for better performance. Anyone using this tape drive, or any others, have experience they would like to share to increase my throughput to something reasonable?? Any suggestions are appreciated. =========================================================================== = Mike Dragicevich = email : mpd@ssd.ray.com = = Raytheon = phone : (401)-842-3476 = = Electronic Systems Division = fax : (401)-842-5200 = = 1847 West Main Rd M/S 177 = = = Portsmouth, R.I. 02871 = = =========================================================================== From mea@mclean.sparta.com Wed Apr 19 13:30:31 1995 From: mea@mclean.sparta.com (Mike Anderson) Date: Wed Apr 19 13:30:37 PDT 1995 Subject: Re: Quad Serial I/O card for VME Greetings! I've got a driver that I wrote for the Xycom X400 series 4 port board. I'll send it to you gratis if you want to use this board. It's a hack from the Heurikon HKV3E tyCo driver ported to 5.1 and probably worth what you pay for it ;-). It did work the last time I used it. Regards, =============================================================================== __ Real-Time System Development, Integration, Training and Services //\\ // \\ Mike Anderson // /\ \\ Chief Engineer // / \ \\ SPARTA, Inc. Voice : (703) 448-0210 ext. 235 // \ \\ 7926 Jones Branch Drive FAX : (703) 734-3323 \\ \ // Suite 900 EMAIL : mea@mclean.sparta.com \\ \ / // McLean, VA 22102 \\ \/ // "Software development is like making \\ // a baby... You can't make a baby in one \\// month by impregnating nine women. -- "Pride in Performance" Some things just take time." =============================================================================== From browen@aoc.nrao.edu Wed Apr 19 14:13:02 1995 From: Bruce Rowen Date: Wed Apr 19 14:13:10 PDT 1995 Subject: Re: SCSI tape question > > > I am currently using VxWorks 5.1.1 on a MVME167 with an HP35470A > DAT tape drive. I am getting a sustained write throughput of only > 140 Kbytes/s, for block writes of 10,000 to 50,000 bytes. > Running 'spy' while a task is continuously writing to the tape drive > indicates that the task is only utilizing 5% of the CPU. > The tape drive is being written while in variable mode. > > The tape drive uses SCSI-II and is spec'd at 5 Mbytes/s > synchronous and 1.5 Mbytes/s asynchrounous, and we need at least a > 250 Kbytes/s sustained rate. The SCSI bus also has a disk drive but no > accesses are being performed to it during tape accesses. The Buffered > Mode field of the Mode Select Parameter is set to 1 as suggested > in the tape drive interface manual for better performance. > Your quoted SCSI rates are for tape drive buffer transfer rates. Once the buffer is full, your throughput drops down to the tape buffer to media transfer rate which is about 180 KB/sec for the drive you indicate. > > Anyone using this tape drive, or any others, have experience they > would like to share to increase my throughput to something reasonable?? > Any suggestions are appreciated. The only way to get higher rates is to use a compression drive. The HP 35480 does 2:1 so *if* your data can be compressed 2:1, your sustained data transfer rate *should* approach >300 KB/sec. Newer drives can store/compress even better and should yield proportional gains. YMMV > > > =========================================================================== > = Mike Dragicevich = email : mpd@ssd.ray.com = > = Raytheon = phone : (401)-842-3476 = > = Electronic Systems Division = fax : (401)-842-5200 = > = 1847 West Main Rd M/S 177 = = > = Portsmouth, R.I. 02871 = = > =========================================================================== ------------------------------------------------------------------------------- Bruce Rowen National Radio Astronomy Observatory Scientific Programmer Array Operations Center browen@aoc.nrao.edu P.O. Box O Socorro, NM 87801 (505)835-7329 (505)835-7000 ------------------------------------------------------------------------------- From enorgren@ehsl.mitre.org Wed Apr 19 16:04:27 1995 From: enorgren@ehsl.mitre.org (Eric J Norgren) Date: Wed Apr 19 16:04:34 PDT 1995 Subject: ATM net i/f driver Hello, Our group is searching for ANY vxWorks drivers that currently exist or will exist by 5/15/95 for ANY 3U or 6U VMEbus-based ATM network adapter boards. The ATM hardware and driver need to satisfy the following minimal requirements: 1. User Network Interface Specification (Version 3.0) compliance 2. OC-3 SONET (fiber) interface 3. Permanent Virtual Circuit (PVC) support 4. Interoperability with FORE Systems ASX-200 ATM switch 5. TCP/IP encapsulation over the ATM datalink layer 6. ATM Adaptation Layers (AAL's) 3,4, and 5 support We have looked at the following 3 VMEbus-based ATM network adapters: 1. FORE VMA-200 2. Interphase 5215 3. Dynatem DATM13x (3U) and DATM16x (6U) FORE currently has no vxWorks driver the VMA-200 but hopes to have one within 3 to 6 months. Sparta, Inc. plans to develop a vxWorks driver for the Interphase 5215 but this effort has not yet begun. Trillium Digital Systems, Inc. offers ATM Forum compliant ATM communications software that can be integrated with the Dynatem ATM boards, but there seems to be no full-up vxWorks drivers available for adapters. LSI Logic's ATMizer R/T System Development Platform provides an excellent integrated ATM protocol/vxWorks development environment which bodes well for the future availability of vxWorks drivers - are there any ATMizer-based VME ATM network adapter boards with vxWorks drivers out there today or which are planned to be released by 5/15/95 ? Has anyone out there come across a vxWorks driver for the ATM network adapters listed above? Has anyone out there come across a vxWorks driver for a VMEbus-based ATM network interface adapter that is not listed above? Information regarding any such vxWorks drivers which would be available by 5/15/95 would be most helpful. Thank you in advance for any insights. Eric Norgren MITRE Corporation Embedded Hardware Standards Specialty Group email: enorgren@ehsl.mitre.org voice: 703-883-5632 fax: 703-883-3337 From pardoe@ncp.gpt.co.uk Thu Apr 20 01:31:04 1995 From: "Matthew Pardoe" Date: Thu Apr 20 01:31:12 PDT 1995 Subject: SLIP for Sparc Can anyone tell me where I can get hold of a version of SLIP that will run on a SUN 4 sparc station. =============================================================================== Matthew Pardoe _/_/ _/_/ _/ _/_/_/_/ email : pardoe@ncp.gpt.co.uk GPT Ltd. _/ _/ _/ _/ _/ _/ tel : +44 1202 782653 Sopers Lane _/ _/ _/ _/ _/ _/_/_/_/ fax : +44 1202 782163 Poole BH17 8BH _/ _/ _/_/_/_/ _/ England =============================================================================== From daemon@vxw.ee.lbl.gov Thu Apr 20 04:00:35 1995 From: daemon@vxw.ee.lbl.gov Date: Thu Apr 20 04:00:44 PDT 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Thu Apr 20 04:00:26 PDT 1995 Subject: Re: problems with fprintf Subject: Help! Why does puts work, while printf doesn't? Subject: iRMX EMB OS Subject: Re: Quad Serial I/O card for VME Subject: Re: Help! Why does puts work, while printf doesn't? Subject: problems with 167 and 147 board on same bus Subject: Re: SCSI tape question Subject: Socket processing with VxWorks Subject: Loading an object file assembled with AS68k ? ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: problems with fprintf Date: Wed, 19 Apr 1995 16:48:38 GMT From: jfinley@netcom.com (John Finley) Organization: NETCOM On-line Communication Services (408 261-4700 guest) Message-ID: References: <199504190910.AA039242635@hpbpq11.bpo.hp.com> Sender: jfinley@netcom22.netcom.com Robert Nils Itschner writes: > Hello, > I am having some random errors while using fprintf to write to a file over > the Ethernet (from a MVME167). Sometimes characters turn up in the file > instead of doubles as supposed. This happens only if I am using the application > I am developing. If I am using a simple little programm to write numbers to > a file these errors never occur. Make sure the task calling fprintf has floating point enabled. If not, then it will "usually" work, but will do the unpredictable when an interrupt comes in during FP operations. Enabling FP for a task just makes it so the FP context is saved on task switches. If it's not enabled, and you use FP anyway, it works for awhile if your task is never interrupted when the FP context is important. Eventually, it will probably get interrupted at the wrong time, and the FP context will get messed up. When your task comes back, it continues on with the messed up FP context. John Finley --------------------------- Newsgroups: comp.os.vxworks Subject: Help! Why does puts work, while printf doesn't? Date: 19 Apr 1995 20:23:50 GMT From: chip@b4pph1a6.rus.uni-stuttgart.de (Chip Salzenberg) Organization: BNR, RTP, NC Message-ID: <3n3rgm$8tq@crchh327.bnr.ca> Here's a weird one -- too weird for me, a VxWorks newbie. Can you help? I'm using tshell, and this works: -> puts "hello" But this _doesn't_ work: -> printf "hello\n" For some reason, printf is returning -1 and doing nothing. What's wrong? --------------------------- Newsgroups: comp.os,comp.os.vxworks Subject: iRMX EMB OS Date: 19 Apr 1995 20:36:33 GMT From: Bill Schell Organization: AT&T Bell Labs, Holmdel, NJ, USA Message-ID: <3n3s8i$27g@big.info.att.com> Has anyone used the iRMX EMB operating system from Intel for developing embedded applications? Any comments are appreciated, horror stories, great praise, etc. We're about to begin writing a fairly complex application that will live on an embedded 386EX and don't want to jump into an OS that is known to be problematic. Please reply via email and I'll post replies if people are interested. I apologize if this is appearing in the wrong newsgroup. Thanks very much, Bill Schell AT&T Bell Labs, Murray Hill, NJ bill@allegra.att.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Quad Serial I/O card for VME Date: Wed, 19 Apr 95 09:15:02 EDT From: kai@ade.no (Kai Atle Myrvang) Message-ID: <020fjee5f@ade.no> References: <199504181851.OAA14002@a9972.ASTRO.GE.COM> Reply-To: kai@ade.no >Does anyone have knowledge of vendors which supply Quad Serial I/O Modules for VME >architectures which have device drivers that are integrated with VxWorks >version 5.1.1? The only vendor I've been able to find is APLabs (Xycom serial board). PEP Modular Computers provide VxWorks drivers for virtually all their modules. In the product range, you find solutions with up to eight serial channels on one VME board. As you are in US, here is their US address: PEP Modular Computers Inc., 750 Holiday Drive, Building 9 Pittsburg, PA 15220 Phone: 412 921-3322, 800 228-1737 Fax: 412 921-3356 - -- * Kai Atle Myrvang (kai@ade.no) * * AD Elektronikk AS, P.O.Box 104, N-1405 Langhus, Norway * * Tel: int+47 64 86 99 70 | Fax: int+47 64 86 99 20 * ********************************************************************* --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Help! Why does puts work, while printf doesn't? Date: 19 Apr 1995 17:47:06 -0700 From: kla@leland.Stanford.EDU (Earl Mitchell) Organization: Stanford University, CA 94305, USA Message-ID: <3n4aua$me0@elaine30.Stanford.EDU> References: <3n3rgm$8tq@crchh327.bnr.ca> In article <3n3rgm$8tq@crchh327.bnr.ca>, Chip Salzenberg wrote: >Here's a weird one -- too weird for me, a VxWorks newbie. Can you help? > >I'm using tshell, and this works: > > -> puts "hello" > >But this _doesn't_ work: > > -> printf "hello\n" > >For some reason, printf is returning -1 and doing nothing. > >What's wrong? This seems to work fine on my system. --------------------------- Newsgroups: comp.os.vxworks Subject: problems with 167 and 147 board on same bus Date: 20 Apr 1995 07:27:43 GMT From: s#slucas@eduserv.its.unimelb.EDU.AU (Stuart Ronald Lucas) Organization: University of Melbourne Message-ID: <3n52df$106k@news.unimelb.EDU.AU> Hi all, I have a mvme167-01B (4M) configured as the system controller in slot 1 and a mvme147A-1 (8M) in slot 2. The system boots ok, and I can rlogin to both. From the 147 I can read all of the 167 memory using 'd' this is mapped to vme address space 0x00800000-0x00bfffff. However I cannot do the same from the 167 the 147 memory is mapped to vme addresses 0x01000000-0x017fffff I get a bus error if i try any addresses in this space eg. - -> d 0x01000000 01000000: Bus Error Program Counter: 0x0005acd4 Status Register: 0x3004 Access Address : 0x01000000 Special Status : 0x0145 27600 _vxTaskEntry +10 : _shell (1, 0, 0, 0, 0, 0, 0, 0, 0, 0) 4e1e2 _shell +12e: 4e200 ([1, 0, 0, 4e010, 0]) 4e378 _shell +2c4: _execute (3a6802) 4e49c _execute +ac : _yyparse ([0, 1, 0, 3a6802, 0]) 52660 _yyparse +514: 5102a (3ab9b0, 3ab9d0) 51136 _yystart +758: _d (1000000, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0) shell restarted. If any one has any ideas I would like to hear them. regards Stuart Lucas - -- Stuart Lucas Robotics & Design Group Department of Mechanical & Manufacturing Engineering Ph.: (03) 344 6680 University of Melbourne, Victoria Melbourne. Fax.: (03) 347 8784 Email: stuart@spike.mame.mu.OZ.AU Ah: (03) 568 2559 --------------------------- Newsgroups: comp.os.vxworks Subject: Re: SCSI tape question Date: Thu, 20 Apr 1995 09:15:40 GMT From: prayson@eltec.win-uk.net (Eltec International PLC) Message-ID: <69@eltec.win-uk.net> References: <199504191927.PAA25457@guppy.ssd.ray.com> Reply-To: prayson@eltec.win-uk.net (Eltec International PLC) In article <199504191927.PAA25457@guppy.ssd.ray.com>, Dragicevich (mpd@ssd.ray.com) writes: > > >I am currently using VxWorks 5.1.1 on a MVME167 with an HP35470A >DAT tape drive. I am getting a sustained write throughput of only >140 Kbytes/s, for block writes of 10,000 to 50,000 bytes. >Running 'spy' while a task is continuously writing to the tape drive >indicates that the task is only utilizing 5% of the CPU. >The tape drive is being written while in variable mode. > > According to HPs Product Brief, the 35470A has a maximum sustained rate of 183 KBytes/sec, so your 140 is reasonable. The 35480A is spec'ed at 732KBytes/sec, but I havn't tried it. Peter Rayson. - --- Eltec International PLC Tel: +44 1908 366499 (a subsidiary of Eltec Elektronik GmbH) Fax: +44 1908 274600 Milton Keynes, MK7 8LF, UK email: support@eltec.win-uk.net --------------------------- Newsgroups: comp.os.vxworks Subject: Socket processing with VxWorks Date: Wed, 19 Apr 95 15:07:49 GMT From: gilley@america.net (Charles Gilley) Organization: AEL/Cross Systems Message-ID: <3n4imt$48v@peach.america.net> I'd like to make sure our messaging system using sockets is kosher. I have a concern that I need a mutual exclusion semephore but perhaps not. The system under development is based on the VME chassis with one master processor and the rest slaves. The master processor publishes the networkservice and then stalls on the connect request (not really the processor, just the task). Now when a slave processor issues a connect request, the network connection is made, socket commnications are established, and the master processor creates two additional tasks for send and receive message processing. Each message processing task is at the same priority and queue driven. We're not time slicing, instead, we operate in the pre-emptive multitasking mode. Will the message tasks conflict? Do I need an exclusion semephore, or can I dpend on the kernal to get things sorted out? Sig area under construction..... --------------------------- Newsgroups: comp.os.vxworks Subject: Loading an object file assembled with AS68k ? Date: 20 Apr 1995 10:30:17 GMT From: fedou@obspm.fr (Pierre Fedou) Organization: Observatoire de Meudon, France Message-ID: <3n5d3p$r17@upsn2.u-psud.fr> Reply-To: fedou@megasx.obspm.fr We have a test program written in 68040 assembler, but we can't load it in VxWORKS kernel's. Do we need any assembler directives in the source program, in order to use it with VxWORKS or do we need anything else? Any and all help would be greatly appreciated. --------------------------- End of New-News digest ********************** From leif@tsl.uu.se Thu Apr 20 04:13:20 1995 From: leif@tsl.uu.se (Leif Thuresson) Date: Thu Apr 20 04:13:26 PDT 1995 Subject: ansiTime and TIMEZONE Anyone managed to get the TIMEZONE and ansiTime library working ? When I run the test program below it seam to ignore the daylight saving parameters. -Leif #include #include #include #include void timezone_test() { time_t tp; putenv("TIMEZONE=CET::-60:040102:100102"); tp = time(0); printf(ctime(&tp)); putenv("TIMEZONE=CET::-60:060102:100102"); tp = time(0); printf(ctime(&tp)); return(OK); } Output from program on Apr 20: -> timezone_test THU APR 20 12:07:18 1995 THU APR 20 12:07:18 1995 value = 0 = 0x0 -> ------------------------------------------------------------------------------- Leif Thuresson, The Svedberg Laboratory, Uppsala University, Sweden leif@tsl.uu.se From leonid@rst.co.il Thu Apr 20 05:55:25 1995 From: leonid@rst.co.il (Leonid Rosenboim) Date: Thu Apr 20 05:55:32 PDT 1995 Subject: Re: SCSI tape question As you well know we sell an off-the-shelf driver and utilities package for SCSI Tapes on VxWorks, and here is what I can offer you from our (and our customers) collective experience: The sustained rate of 140 KB/s is truly what a primitive DAT drive provides. To go higher you can either opt for a drive with compression, but then if your data comes from a CCD or something, the noise will make your compression ration quite low. You can actually "test" the typical compression ration of your data if you pump a couple of megs to your work station, and run "compress -v" on it. The algorythms are similar, so are the results. You can get a real sustained rate of 250 KB/s from an 8mm Exabyte 8200, without any compression being used, but that is an old drive. The common 8mm drive now is the EXB-8505 which provides 500KB/s sustained without compression, and compression there does help sometimes. ANother point - I have encountered drives that in some cases REDUCED the actual transfer rate when compression was turned on, and the data was not compressible. The usual tape vendor figures are based on a typical backup application, with lots of ASCII and executable files. ----------------------------------------------------------------------- Leonid Rosenboim Phone: +972-3-559-8144 R S T Software Industries Ltd. Mobile: +972-50-307-142 3 Hashikma st. Industrial Zone, Fax: +972-3-559-8244 P.O.Box 11502, AZUR 58017, Israel E-Mail: leonid@rst.co.il From burt@xylan.com Thu Apr 20 10:18:12 1995 From: burt@xylan.com (Burt Cyr) Date: Thu Apr 20 10:18:18 PDT 1995 Subject: Re: SLIP for Sparc Hi Matthew, Regarding: > Can anyone tell me where I can get hold of a version of SLIP that will run > on a SUN 4 sparc station. We're using the cslip package and it works just fine. Try ftp.uu.net:/networking/ip/slip/cslip/cslip-2.7.tar.Z Burt ================================================================== v _ ____(_)___ Burt Cyr Xylan, Inc. _-_-__-_-_-_ burt@xylan.com 26679 Agoura Rd. Suite 100 --_-_--_-_- (818) 880-3500 x3533 Calabasas, CA 91302 ================================================================== From mpd@ssd.ray.com Thu Apr 20 11:08:57 1995 From: Dragicevich Date: Thu Apr 20 11:09:03 PDT 1995 Subject: SCSI tape question resolution Thanks to all who responded to my SCSI tape drive question. The key was that the HP35470A is only spec'd at 180 Kbytes/sec, and that information was not contained in the pages of my tape drive interface manual. I've also just learned that the project's tape drive is soon to be upgraded to an HP C1533A, which is spec'd at better than 510 Kbytes/s sustained. This should eliminate out current problem. Anyone have any real world experience with this tape drive ? =========================================================================== = Mike Dragicevich = email : mpd@ssd.ray.com = = Raytheon = phone : (401)-842-3476 = = Equipment Division Portsmouth Facility = fax : (401)-842-5200 = = 1847 West Main Rd M/S 177 = = = Portsmouth, R.I. 02871 = = =========================================================================== From jpc@ssd.ray.com Thu Apr 20 11:50:29 1995 From: Carissimi Date: Thu Apr 20 11:50:36 PDT 1995 Subject: Re: SCSI tape question resolution > Anyone have any real world experience with this tape drive ? > No. Joe Carissimi Raytheon Electronic Systems Division 1847 West Main Road, Mail Stop 177 Portsmouth, RI 02871-1087 Email: jpc@ssd.ray.com From sergi@hpbpq11.bpo.hp.com Fri Apr 21 01:13:11 1995 From: Sergi Casas(QA Software testing) Date: Fri Apr 21 01:13:19 PDT 1995 Subject: Unreachable user TO: Shane Arnold, from E-Systems Inc. You contacted me through this users group, but I cannot reach your e-mail address (slarnold@unicomp.net) to respond. Please, e-mail me again after checking the cause of the problem. -------------------------------------------------------------- While connected to ns1.unicomp.net. [199.1.42.2] (tcp): >>> RCPT To: <<< 550 ... User unknown 550 ... User unknown -------------------------------------------------------------- Thanks. -- |============================================================================| | Sergi Casas ##### / ##### Firmware Engineer | | Hewlett-Packard Company ### /_ _ ### | | Barcelona Division (BCD) ## / / / / ## e-mail: sergi@hp-bpo.bpo.hp.com | | Avda. Graells, 501 ## / / /_/ ## Phone : +34 3 582 1475 | | 08190 S. Cugat (Barcelona) ### / ### Fax : +34 3 582 2515 | | SPAIN ##### / ##### HP-Telnet: 712 1475 | |============================================================================| From daemon@vxw.ee.lbl.gov Fri Apr 21 04:00:27 1995 From: daemon@vxw.ee.lbl.gov Date: Fri Apr 21 04:00:35 PDT 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Fri Apr 21 04:00:19 PDT 1995 Subject: Re: Socket processing with VxWorks Subject: help with hkv3500 board Subject: Re: C++ and taskSpawn Subject: Motorola 162-FX ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Socket processing with VxWorks Date: Thu, 20 Apr 1995 17:23:42 GMT From: jfinley@netcom.com (John Finley) Organization: NETCOM On-line Communication Services (408 261-4700 guest) Message-ID: References: <3n4imt$48v@peach.america.net> Sender: jfinley@netcom22.netcom.com Charles Gilley (gilley@america.net) wrote: : I'd like to make sure our messaging system using sockets is kosher. I have a concern that I : need a mutual exclusion semephore but perhaps not. The system under development is based : on the VME chassis with one master processor and the rest slaves. The master processor publishes : the networkservice and then stalls on the connect request (not really the processor, just the task). : Now when a slave processor issues a connect request, the network connection is made, socket : commnications are established, and the master processor creates two additional tasks for send and : receive message processing. : Each message processing task is at the same priority and queue driven. We're not time slicing, instead, : we operate in the pre-emptive multitasking mode. Will the message tasks conflict? Do I need an : exclusion semephore, or can I dpend on the kernal to get things sorted out? : Sig area under construction..... My response would be that if you're considering a mutex (in this case, and probably most cases), then it's probably not worth thinking about, and just putting one in. Even if you don't *really* need one right now, someone's going to modify and break your code some day if you don't protect yourself. Since your application is already dependent on the network for some of its timing, you probably don't mind the extra microseconds the mutex will take. Just my opinion, John Finley --------------------------- Newsgroups: comp.os.vxworks Subject: help with hkv3500 board Date: Thu, 20 Apr 1995 18:30:13 GMT From: mew@ece.scarolina.edu (Howard A. Mew, Jr. "Al") Organization: Univ. of S. Carolina - ECE Dept. Message-ID: <4147cb$d1ed.186@aragorn.ece.scarolina.edu> We are currently using a hkv3500 board running VxWorks 5.1.1 We seem to be having problems with any tasks that access the network while a seperate task uses the I/O ports. We generally can not access the network -- currently write to a network disk. I was wondering if anyone out there has experience problems like this or any other similiar problems with this board or VxWorks in general Thanks Al Mew mew@ece.scarolina.edu --------------------------- Newsgroups: comp.os.vxworks Subject: Re: C++ and taskSpawn Date: 21 Apr 1995 07:55:50 +1000 From: rjl@f111.iassf.easams.com.au (Rohan LENARD) Organization: EASAMS (Australia) Pty Ltd Message-ID: <3n6l96$atl@f111.iassf.easams.com.au> References: <9503077972.AA797281065@mailhost.scires.com> <3m4sd5$3dp@f111.iassf.easams.com.au> <3mcfl3$7js@marlyn.cna.tek.com> Hi there, In article <3mcfl3$7js@marlyn.cna.tek.com>, Bill Randle wrote: >There is yet another way, depending on your system requirements. What we did was >to create a Process class that encapsulates (and isolates) the rest of the >system from the details of the OS [this becomes really handy if you need >to port an application to another OS!]. Our Process class has a run() member >function (which is the main processing loop) and includes member functions >to set/get task priority and preemption and member functions to send messages >to the process. > >The task then is actually spawn in the constructor for the class. Since you >are inside the class, you know the address of the run() function - it doesn't >have to be static. In many cases, the spawned task then just goes to sleep, >waiting on a message queue until told to do something. > But you still have to provide the "this" pointer, because a member function needs it. The other problem with this method is that it constrains people to inherit from process, or you could develop this further to use a helper which contains a this and member fn pointer to call, and attach a generic callback to process. eg. class Callback { public: virtual void operator()(void) = 0; } class Process { public: CreateTask(Callback& cb) { // spawn the task cb.(); }; }; class Application { public: void FunctionToRun(); //... }; template AppCB : public Callback { public: AppCB(T* pObj, void (T::*pmf)(void)) : Object(pObj), PMF(pmf) {}; virtual void operator()(void) { (pObj->*PMF)(); };; private: T* Object; void (T::*PMF)(void); }; main() { Application* pTask1Prog = new Application; Process.Run(AppCB(pTask1Prog,&Application::FunctionToRun)); } Regards, rohan - -- - ---------------------------------------------------------------------------- rjl@iassf.easams.com.au | All quotes can be attributed to my automated quote Rohan Lenard | writing tool. Yours for just $19.95; and if you +61-2-367-4555 | call now you'll get a free set of steak knives ... --------------------------- Newsgroups: comp.os.vxworks Subject: Motorola 162-FX Date: 21 Apr 1995 07:06:16 GMT From: matthieu@elwood.laas.fr (Matthieu Herrb) Organization: LAAS-CNRS Toulouse, France Message-ID: Reply-To: matthieu@laas.fr Hi, Has someone experiences with the new Motorola 162 FX boards ? are they supported by the 162 BSP ? Thank you in advance. - -- Matthieu Herrb CNRS - LAAS | War, Matthieu.Herrb@laas.fr 7, avenue du Colonel Roche | What is it good for ? 31077 Toulouse Cedex - France | Absolutly Nothing ! --------------------------- End of New-News digest ********************** From phillip.shaffer@ae.ge.com Fri Apr 21 04:56:04 1995 From: phillip.shaffer@ae.ge.com (Phil Shaffer) Date: Fri Apr 21 04:56:11 PDT 1995 Subject: FTP address for Gnu software / 68060 Two questions: 1) Can someone provide the FTP address and file names for obtaining the latest (or at least recent) versions of GCC and GAS from FSF or other FTP site(s)? 2) Has anyone adapted GCC / GAS for 68060? If so, can you share the mods? Many thanks. Phillip L. Shaffer phillip.shaffer@ae.ge.com GE Aircraft Engines, MS G57 1 Neumann Way Cincinnati, OH 45215 From noacco@maloja.egt.fr Fri Apr 21 08:42:54 1995 From: Alain Noacco Date: Fri Apr 21 08:43:01 PDT 1995 Subject: Re: FTP address for Gnu software / 68060 > > 1) Can someone provide the FTP address and file names for obtaining > the latest (or at least recent) versions of GCC and GAS from FSF or > other FTP site(s)? > > Phillip L. Shaffer phillip.shaffer@ae.ge.com > GE Aircraft Engines, MS G57 > 1 Neumann Way > Cincinnati, OH 45215 You need these files : gcc-2.6.3.tar.gz gas-2.5.1.tar.gz The files are available via anonymous ftp under directory ~ftp/gnu. Emacs and other GNU programs may be available via anonymous ftp from these US sites: ftp.kpc.com:/pub/mirror/gnu (Silicon Valley, CA) ftp.hawaii.edu:/mirrors/gnu, f.ms.uky.edu:/pub3/gnu, ftp.digex.net:/pub/gnu (Internet address 164.109.10.23, nightly full mirror, ran by mcguire@digex.net), wuarchive.wustl.edu:/systems/gnu, col.hp.com:/mirrors/gnu, ftp.cs.columbia.edu:/archives/gnu/prep, vixen.cso.uiuc.edu:/gnu, mrcnext.cso.uiuc.edu:/pub/gnu, jaguar.utah.edu:/gnustuff, gatekeeper.dec.com:/pub/GNU, labrea.stanford.edu, archive.cis.ohio-state.edu, and ftp.uu.net:/archive/systems/gnu. And these foreign sites: ftp.cs.ubc.ca:/mirror2/gnu (Western Canada, daily full mirror, ran by ftp-admin@cs.ubc.ca), ftp.inf.utfsm.cl:/pub/gnu (Chile 146.83.198.3 nightly full mirror, ran by ftp@inf.utfsm.cl), ftp.unicamp.br:/pub/gnu (Brazil manual mirror, ran by oliva@dcc.unicamp.br), archie.au:/gnu (Australia (archie.oz or archie.oz.au for ACSnet)), ftp.technion.ac.il:/pub/unsupported/gnu (Israel, daily full mirror, ran by ftp-admin), ftp.sun.ac.za:/pub/gnu (South Africa), ftp.mcc.ac.uk:/pub/gnu (130.88.203.12 daily full mirror, ran by root@ftp.mcc.ac.uk), unix.hensa.ac.uk:/pub/uunet/systems/gnu, ftp.warwick.ac.uk (137.205.192.14 daily full mirror, ran by unixhelp@warwick.ac.uk), ftp.informatik.tu-muenchen.de, ftp.informatik.rwth-aachen.de or germany.eu.net (mirror ran by archive-admin@germany.eu.net) (Germany), isy.liu.se (Sweden), ftp.stacken.kth.se or ftp.luth.se:/pub/unix/gnu (Sweden), ftp.sunet.se:/pub/gnu (Sweden 130.238.127.3 daily mirror, ran by archive@ftp.sunet.se (also mirrors the Mailing List Archives) hp4nl.nluug.nl (Netherlands), ftp.win.tue.nl:/pub/gnu (Netherlands 131.155.70.100 daily mirror, ran by ftp@win.tue.nl), ftp.funet.fi:/pub/gnu (Finland 128.214.6.100, ran by gnu-adm), ftp.denet.dk (Denmark), ugle.unit.no (Norway 129.241.1.97), ftp.eunet.ch or nic.switch.ch:/mirror/gnu (Switzerland), irisa.irisa.fr:/pub/gnu or ftp.univ-lyon1.fr:pub/gnu (ran by ftpmaint@ftp.univ-lyon1.fr) (France), ftp.ieunet.ie:pub/gnu (Ireland 192.111.39.1 weekly mirror, ran by archive@ieunet.ie), archive.eu.net (Europe 192.16.202.1), cair-archive.kaist.ac.kr:/pub/gnu (Korea 143.248.11.171, ran by ftpkeeper@cair-archive.kaist.ac.kr), ftp.nectec.or.th:/pub/mirrors/gnu (Thailand 192.150.251.32 daily mirror, ran by ftp@nwg.nectec.or.th), utsun.s.u-tokyo.ac.jp:/ftpsync/prep or ftp.cs.titech.ac.jp (Japan, nemacs, the japanese port of GNU Emacs, is under ~ftp/JAPAN). Un*x), which will run Un*x user ============================================================================= | Alain NOACCO | European Gas Turbines | | EGT PEDL | 3, Avenue des trois chenes | | tel : (33) 84.55.49.57 | 90018 BELFORT Cedex | | fax : (33) 84.55.35.21 | FRANCE | | email : anoacco@egt.fr | | ============================================================================= From djb@rtp.co.uk Fri Apr 21 09:44:45 1995 From: djb@rtp.co.uk (David Bullimore) Date: Fri Apr 21 09:44:51 PDT 1995 Subject: Re: comp.os.vxworks newsdigest [matthieu@laas.fr wrote:] > > Has someone experiences with the new Motorola 162 FX boards ? are they > supported by the 162 BSP ? The standard mv162 BSP does not work with the FX. There is a new BSP for this board but it is only due for release with VxWorks 5.2. David *********************************************** * David Bullimore * * Real Time Products Ltd. (RTP Ltd.) * * Unit 8 * * Holt Court * * Aston Science Park * * Birmingham * * B7 4EJ * * England * * * * Phone: +44 121 333 6955 * * Fax: +44 121 333 5433 * * * * Email djb@rtp.co.uk * *********************************************** From johnk@photon.ceco.com Fri Apr 21 13:34:00 1995 From: johnk@photon.ceco.com (John M. Katrenak) Date: Fri Apr 21 13:34:08 PDT 1995 Subject: Solaris 2.3 / SO_KEEPALIVE / VxWorks 5.11 Troubles Has anyone experienced the following problem and found a solution? Problem: -------- VxWorks "server" reboots and the Sun "client" connected to the server using a socket fails to detect the failed socket while polling the file descriptor for in-bound data. All attempts to use the socket option SO_KEEPALIVE have been fruitless. I am aware of some problems with Solaris 2.3 and the socket interface, but I have not been able to locate any information in the Sun bug reports relating to SO_KEEPALIVE not working. Environment: ------------ Sun SPARCstation 10 running Solaris 2.3 MVME167 running VxWorks 5.11 Thanks... -- John M. Katrenak (johnk@ceco.ceco.com) Process Systems Department Commonwealth Edison Company Chicago, Illinois From daemon@vxw.ee.lbl.gov Sun Apr 23 04:00:29 1995 From: daemon@vxw.ee.lbl.gov Date: Sun Apr 23 04:00:38 PDT 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sun Apr 23 04:00:20 PDT 1995 Subject: problem with qsort Subject: some questions about VxWorks Subject: VME FDDI boards. Subject: Re: Help! Why does puts work, while printf doesn't? Subject: Any PEP/vxWorks users out there? Subject: Support for Industry Packs Subject: Advanced vxWorks/Realtime class available! Subject: Re: anyone using mv177 BSP with INCLUDE_SCSI? Subject: Re: Help! Why does puts work, while printf doesn't? ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: problem with qsort Date: 22 Apr 1995 06:16:38 -0500 From: kevin@htilbom.ernet.in (Kevin Menezes) Organization: UTexas Mail-to-News Gateway Message-ID: <9504220740.AA06309@htilbom.ernet.in> Sender: nobody@cs.utexas.edu Hello, I have read the FAQ, but I do not know from where to get a copy of qsort() from the BSD UNIX distribution.Can someone please mail me the address of a site which carries this ? Please reply to kevin@htilbom.ernet.in. Thanks, Kevin --------------------------- Newsgroups: comp.os.vxworks Subject: some questions about VxWorks Date: Sat, 22 Apr 1995 02:25:23 GMT From: James Glass Organization: Polytechnic University, New York Message-ID: Sender: news@poly.edu (netnews) Just a few questions.... I am currently designing an autonomous helicopter to compete at Gergia tech... I am in the process of either designing my own kernal, or choosing one to run the on board processes... Obviously, I have been looking at realtime Os', and someone recommended this group to me... Is there a FAQ for VX? Do the binaries have to be purchased, or can it be downloaded from somewhere? The helicoptor has all of the traditional needs of a realtime systesm.. will VX work? I would appreciate any help... --------------------------- Newsgroups: comp.os.vxworks Subject: VME FDDI boards. Date: 21 Apr 1995 13:31:41 GMT From: David Kim Organization: NASA Goddard Space Flight Center -- Greenbelt, Maryland USA Message-ID: <3n8c3t$5jt@post.gsfc.nasa.gov> I am in search of VME based FDDI interface cards that will be able to sustain 34 Mbps with TCP/IP. If any of you know of such commercial FDDI board, please let me know. Even possible lead will be appreciated. Thanks. David Kim email: calypso@vlsi9.gsfc.nasa.gov --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Help! Why does puts work, while printf doesn't? Date: Thu, 20 Apr 1995 12:56:58 GMT From: clegg@ssd.fsi.com Organization: Flightsafety International, Broken Arrow, OK Message-ID: <1995Apr20.125658.63835@dev1.ssd.fsi.com> References: <3n3rgm$8tq@crchh327.bnr.ca> Reply-To: clegg@ssd.fsi.com Sender: news@dev1.ssd.fsi.com In <3n3rgm$8tq@crchh327.bnr.ca>, chip@b4pph1a6.rus.uni-stuttgart.de (Chip Salzenberg) writes: > -> printf "hello\n" >For some reason, printf is returning -1 and doing nothing. >What's wrong? Are you getting an unknown command response also? If so you may wish to make certain that your symbol table loaded successfully. --------------------------- Newsgroups: comp.os.vxworks Subject: Any PEP/vxWorks users out there? Date: 21 Apr 1995 15:13:44 GMT From: johncoby@blkbox.com (John Cobarruvias) Organization: The Black Box, Houston, Tx (713) 480-2686 Message-ID: <3n8i38$6b7@news.blkbox.com> We recently have purchased a PEP VM30 board from PEP cpmputers and their BSP. I am curious if any other organization is using PEP boards with vxWorks? - -- _____ __ ______ _| |.-----.| |--.-----. | __ \ | || _ || | | | < John R. Cobarruvias |_______||_____||__|__|__|__| |___|__| johncoby@blkbox.com ______ __ __ | |.-----.| |--.---.-.----.----.--.--.--.--.|__|.---.-.-----. | ---|| _ || _ | _ | _| _| | | | || || _ |__ --| |______||_____||_____|___._|__| |__| |_____|\___/ |__||___._|_____| --------------------------- Newsgroups: comp.os.vxworks Subject: Support for Industry Packs Date: 22 Apr 1995 15:50:49 GMT From: alanc@mistral.co.uk (Alan Campbell) Organization: ObjecTime. Message-ID: <3nb8kp$9a@news.mistral.co.uk> Does anyone know of a list of which Industry Packs have vxWorks drivers or less complete interfaces available? I'm thinking of using IPs on a Motorola MVME162. Ta, Alan C --------------------------- Newsgroups: comp.os.vxworks Subject: Advanced vxWorks/Realtime class available! Date: 21 Apr 1995 19:33:00 GMT From: johncoby@blkbox.com (John Cobarruvias) Organization: The Black Box, Houston, Tx (713) 480-2686 Message-ID: <3n919c$9ok@news.blkbox.com> There is a class titled Advanced Realtime programming featuring vxWorks in mid June, McLEan Virginia (Wash DC). I am looking for people to fill the class. Anyone interested, please contact me @ 713-483-9357 or cobarruvias@asd1.jsc.nasa.gov I will send you the details. - -- __ __ _ _ _ _ ___ ( )/ \( )( )( \( ) ( ,) John R. Cobarruvias __)(( () ))__( ) ( ) \ johncoby@blkbox.com (___/ \__/(_)(_)(_)\_) (_)\_) __ __ ___ __ ___ ___ _ _ _ _ __ __ ___ / _)/ \( ,) ( ) ( ,) ( ,)( )( )( )( )( ) ( ) / __) ( (_( () )) ,\ /__\ ) \ ) \ )()( \\// )( /__\ \__ \ \__)\__/(___/(_)(_)(_)\_)(_)\_)\__/ (__) (__)(_)(_)(___/ --------------------------- Newsgroups: comp.os.vxworks Subject: Re: anyone using mv177 BSP with INCLUDE_SCSI? Date: Fri, 21 Apr 1995 18:11:59 GMT From: georg@sgl.ists.ca (Georg Feil) Organization: Institute for Space and Terrestrial Science Message-ID: References: <9504142141.AA27593@positron.med.ge.com> Sender: news@newshub.ists.ca cook@positron.med.ge.com (Mike Cook) writes: >Is anyone out there successfully using the WRS supplied BSP for the mv177 >and defining INCLUDE_SCSI? . . . >Trap to uninitialized vector number 329 (0-255). >Invalid ESF type 0x9 >Task: 0x1fffe80 "tRootTask" We got exactly the same thing when evaluating a 177. I tried to find out whether SCSI was *supposed* to work on a 177 under 5.1.1, but Wind River never came out with a direct answer. I did find out that there is an improved 177 BSP but it is for VxWorks 5.2 only. Therefore I concluded that any further 177 work has to wait until 5.2 comes out. Georg. - -- Georg Feil Space Geodynamics Laboratory | Internet: georg@sgl.ists.ca Institute for Space and Terrestrial Science | Phone: (416) 665-5458 4850 Keele St./North York/Ont/Canada/M3J 3K1 | Fax: (416) 665-1815 --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Help! Why does puts work, while printf doesn't? Date: 21 Apr 1995 19:59:16 GMT From: chip@cybernetics.net Organization: BNR, RTP, NC Message-ID: <3n92qk$4n4@crchh327.bnr.ca> References: <3n3rgm$8tq@crchh327.bnr.ca> In article <3n3rgm$8tq@crchh327.bnr.ca>, Chip Salzenberg wrote: >I'm using tshell, and this works: > -> puts "hello" >But this _doesn't_ work: > -> printf "hello\n" Well, it turns out that I linked my program with the libvx.a intended for linking the KERNEL! Aargh! :-( All is well now that I am no longer loading a duplicate printf... :-) --------------------------- End of New-News digest ********************** From leonid@rst.co.il Sun Apr 23 04:38:39 1995 From: leonid@rst.co.il (Leonid Rosenboim) Date: Sun Apr 23 04:38:45 PDT 1995 Subject: Re: SCSI tape question resolution Here is some more advise based on our experience with SCSI Tapes on VxWorks: Please beware the fast DAT drives. The "standard" sustained rate of most DAT drives of 180 KBytes/sec is based on the "audio" transfer rate that the mechanics of DAT was designed to deliver. In order to acheive a higher transfer rate some DAT drives have increased the machanical tape speed, whereas the tape itself is not thick enougth, and will heat up during movement beyond the temperature it was designed to sustain. Hence, the life of the cartrige is significantly shortened. Not all drives use this technique, check your specifications before you buy a drive. Other drives that use DDS-2 recording format are incerasing the transfer rate by means of increasing the recording bit densitities. The problem however with DDS-2 drives is that the technology is not mature enougth yet. Our recommendation to transfer-rate-hungry users is 8mm. ----------------------------------------------------------------------- Leonid Rosenboim Phone: +972-3-559-8144 R S T Software Industries Ltd. Mobile: +972-50-307-142 3 Hashikma st. Industrial Zone, Fax: +972-3-559-8244 P.O.Box 11502, AZUR 58017, Israel E-Mail: leonid@rst.co.il From leonid@rst.co.il Sun Apr 23 08:07:52 1995 From: leonid@rst.co.il (Leonid Rosenboim) Date: Sun Apr 23 08:07:58 PDT 1995 Subject: BSP for IDT boards Has anyone got any information as to where one can get a VxWorks Board Support Package for IDT evaluation boards, specifically with R3081 micro-controller ? Any pointers will be greatly appreciated. Also, does IDT have an email contact ? ----------------------------------------------------------------------- Leonid Rosenboim Phone: +972-3-559-8144 R S T Software Industries Ltd. Fax: +972-3-559-8244 3 Hashikma st. Industrial Zone, E-Mail: leonid@rst.co.il P.O.Box 11502, AZUR 58017, Israel WindRiver Distributor From wilhite@email.ncsc.navy.mil Sun Apr 23 13:51:46 1995 From: wilhite@email.ncsc.navy.mil (Wilhite) Date: Sun Apr 23 13:51:53 PDT 1995 Subject: X-WINDOWS CLIENT UNDER VX-5.02B? Want to develop an x-windows client under VxWorks 5.02b to drive display on Sun 4 running Open Windows release 3. Has anyone done this? Having problems with include file configuration... Thanks in Advance, David Wilhite From vdbent@hea.sron.ruu.nl Mon Apr 24 04:20:47 1995 From: vdbent@hea.sron.ruu.nl (Kees van der Bent) Date: Mon Apr 24 04:20:53 PDT 1995 Subject: The use of source code control Hi, Can some of you please comment on what source code control system they like the most (for use with small/medium sized VxWorks applications). Currently I can choose between SCCS and RCS; which one is best and/or are there others? Thanks in advance, | ### ## | #### ### ## # | Kees van der Bent | #### ### | ## ## # ### ##### | Space Research Organisation Netherlands | #### ## | #### ## ## ## ## | Sorbonnelaan 2 3584 CA Utrecht | #### | ## ## ### ## ## | The Netherlands | ## #### | | | ### #### | | +31-30-538595 | ## ### | | vdbent@hea.sron.ruu.nl From vdbent@hea.sron.ruu.nl Mon Apr 24 04:56:59 1995 From: vdbent@hea.sron.ruu.nl (Kees van der Bent) Date: Mon Apr 24 04:57:05 PDT 1995 Subject: Re: Solaris 2.3 / SO_KEEPALIVE / VxWorks 5.11 Troubles John M. Katrenak write: > Has anyone experienced the following problem and found a solution? > > Problem: > -------- > VxWorks "server" reboots and the Sun "client" connected to the server > using a socket fails to detect the failed socket while polling the file > descriptor for in-bound data. All attempts to use the socket option > SO_KEEPALIVE have been fruitless. I am aware of some problems with > Solaris 2.3 and the socket interface, but I have not been able to > locate any information in the Sun bug reports relating to SO_KEEPALIVE > not working. Unless SO_KEEPALIVE has been finally implemented for both Solaris and for VxWorks, the cause of your problems is clear, As far as I know it was not imlemented for Solaris 1.x nor for VxWorks 5.0.2. To solve the problem I would suggest to temporarily write() (dummy) data from the "client". The write() will surely fail the second time you write after the socket broke. (The "server" can simply throw away the 'dummy' data.) In fact now you're implementing your own keep-alive socket option. Good luck, | ### ## | #### ### ## # | Kees van der Bent | #### ### | ## ## # ### ##### | Space Research Organisation Netherlands | #### ## | #### ## ## ## ## | Sorbonnelaan 2 3584 CA Utrecht | #### | ## ## ### ## ## | The Netherlands | ## #### | | | ### #### | | +31-30-538595 | ## ### | | vdbent@hea.sron.ruu.nl From QMGATE@qmgate.trw.com Mon Apr 24 10:22:51 1995 From: "QMGATE" Date: Mon Apr 24 10:23:19 PDT 1995 Subject: Error in delivery-resend Mail*Link( SMTP comp.os.vxworks newsdigest Submitted-by daemon@vxw.ee.lbl.gov Sun Apr 23 04:00:29 1995 Submitted-by: daemon@vxw.ee.lbl.gov Comp.Os.Vxworks Daily Digest Sun Apr 23 04:00:20 PDT 1995 Subject: problem with qsort Subject: some questions about VxWorks Subject: VME FDDI boards. Subject: Re: Help! Why does puts work, while printf doesn't? Subject: Any PEP/vxWorks users out there? Subject: Support for Industry Packs Subject: Advanced vxWorks/Realtime class available! Subject: Re: anyone using mv177 BSP with INCLUDE_SCSI? Subject: Re: Help! Why does puts work, while printf doesn't? ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: problem with qsort Date: 22 Apr 1995 06:16:38 -0500 From: kevin@htilbom.ernet.in (Kevin Menezes) Organization: UTexas Mail-to-News Gateway Message-ID: <9504220740.AA06309@htilbom.ernet.in> Sender: nobody@cs.utexas.edu Hello, I have read the FAQ, but I do not know from where to get a copy of qsort() from the BSD UNIX distribution.Can someone please mail me the address of a site which carries this ? Please reply to kevin@htilbom.ernet.in. Thanks, Kevin --------------------------- Newsgroups: comp.os.vxworks Subject: some questions about VxWorks Date: Sat, 22 Apr 1995 02:25:23 GMT From: James Glass Organization: Polytechnic University, New York Message-ID: Sender: news@poly.edu (netnews) Just a few questions.... I am currently designing an autonomous helicopter to compete at Gergia tech... I am in the process of either designing my own kernal, or choosing one to run the on board processes... Obviously, I have been looking at realtime Os', and someone recommended this group to me... Is there a FAQ for VX? Do the binaries have to be purchased, or can it be downloaded from somewhere? The helicoptor has all of the traditional needs of a realtime systesm.. will VX work? I would appreciate any help... --------------------------- Newsgroups: comp.os.vxworks Subject: VME FDDI boards. Date: 21 Apr 1995 13:31:41 GMT From: David Kim Organization: NASA Goddard Space Flight Center -- Greenbelt, Maryland USA Message-ID: <3n8c3t$5jt@post.gsfc.nasa.gov> I am in search of VME based FDDI interface cards that will be able to sustain 34 Mbps with TCP/IP. If any of you know of such commercial FDDI board, please let me know. Even possible lead will be appreciated. Thanks. David Kim email: calypso@vlsi9.gsfc.nasa.gov --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Help! Why does puts work, while printf doesn't? Date: Thu, 20 Apr 1995 12:56:58 GMT From: clegg@ssd.fsi.com Organization: Flightsafety International, Broken Arrow, OK Message-ID: <1995Apr20.125658.63835@dev1.ssd.fsi.com> References: <3n3rgm$8tq@crchh327.bnr.ca> Reply-To: clegg@ssd.fsi.com Sender: news@dev1.ssd.fsi.com In <3n3rgm$8tq@crchh327.bnr.ca>, chip@b4pph1a6.rus.uni-stuttgart.de (Chip Salzenberg) writes: > -> printf "hello\n" >For some reason, printf is returning -1 and doing nothing. >What's wrong? Are you getting an unknown command response also? If so you may wish to make certain that your symbol table loaded successfully. --------------------------- Newsgroups: comp.os.vxworks Subject: Any PEP/vxWorks users out there? Date: 21 Apr 1995 15:13:44 GMT From: johncoby@blkbox.com (John Cobarruvias) Organization: The Black Box, Houston, Tx (713) 480-2686 Message-ID: <3n8i38$6b7@news.blkbox.com> We recently have purchased a PEP VM30 board from PEP cpmputers and their BSP. I am curious if any other organization is using PEP boards with vxWorks? - -- _____ __ ______ _| |.-----.| |--.-----. | __ \ | || _ || | | | < John R. Cobarruvias |_______||_____||__|__|__|__| |___|__| johncoby@blkbox.com ______ __ __ | |.-----.| |--.---.-.----.----.--.--.--.--.|__|.---.-.-----. | ---|| _ || _ | _ | _| _| | | | || || _ |__ --| |______||_____||_____|___._|__| |__| |_____|\___/ |__||___._|_____| --------------------------- Newsgroups: comp.os.vxworks Subject: Support for Industry Packs Date: 22 Apr 1995 15:50:49 GMT From: alanc@mistral.co.uk (Alan Campbell) Organization: ObjecTime. Message-ID: <3nb8kp$9a@news.mistral.co.uk> Does anyone know of a list of which Industry Packs have vxWorks drivers or less complete interfaces available? I'm thinking of using IPs on a Motorola MVME162. Ta, Alan C --------------------------- Newsgroups: comp.os.vxworks Subject: Advanced vxWorks/Realtime class available! Date: 21 Apr 1995 19:33:00 GMT From: johncoby@blkbox.com (John Cobarruvias) Organization: The Black Box, Houston, Tx (713) 480-2686 Message-ID: <3n919c$9ok@news.blkbox.com> There is a class titled Advanced Realtime programming featuring vxWorks in mid June, McLEan Virginia (Wash DC). I am looking for people to fill the class. Anyone interested, please contact me @ 713-483-9357 or cobarruvias@asd1.jsc.nasa.gov I will send you the details. - -- __ __ _ _ _ _ ___ ( )/ \( )( )( \( ) ( ,) John R. Cobarruvias __)(( () ))__( ) ( ) \ johncoby@blkbox.com (___/ \__/(_)(_)(_)\_) (_)\_) __ __ ___ __ ___ ___ _ _ _ _ __ __ ___ / _)/ \( ,) ( ) ( ,) ( ,)( )( )( )( )( ) ( ) / __) ( (_( () )) ,\ /__\ ) \ ) \ )()( \\// )( /__\ \__ \ \__)\__/(___/(_)(_)(_)\_)(_)\_)\__/ (__) (__)(_)(_)(___/ --------------------------- Newsgroups: comp.os.vxworks Subject: Re: anyone using mv177 BSP with INCLUDE_SCSI? Date: Fri, 21 Apr 1995 18:11:59 GMT From: georg@sgl.ists.ca (Georg Feil) Organization: Institute for Space and Terrestrial Science Message-ID: References: <9504142141.AA27593@positron.med.ge.com> Sender: news@newshub.ists.ca cook@positron.med.ge.com (Mike Cook) writes: >Is anyone out there successfully using the WRS supplied BSP for the mv177 >and defining INCLUDE_SCSI? From aisie.aisinc.com!aisinc.com!tam@msen.com Mon Apr 24 13:08:51 1995 From: tam@aisinc.com (Tod More) Date: Mon Apr 24 13:09:02 PDT 1995 Subject: Re: The use of source code control > > Submitted-by vdbent@hea.sron.ruu.nl Mon Apr 24 04:20:47 1995 > Submitted-by: vdbent@hea.sron.ruu.nl (Kees van der Bent) > > Hi, > > Can some of you please comment on what source code control system they like > the most (for use with small/medium sized VxWorks applications). Currently > I can choose between SCCS and RCS; which one is best and/or are there others? > > Thanks in advance, > If you have a few programmers RCS, if you have many (Greater than 4) use CVS which uses RCS. I have worked with three companies that have converted from SCCS to RCS, RCS handels branching, and version control much better than SCCS. Good luck Tod From cgonter@inept.stanford.edu Mon Apr 24 16:20:49 1995 From: cgonter@inept.stanford.edu Date: Mon Apr 24 16:20:56 PDT 1995 Subject: tclvx7.3 In the release of tclvx7.3, the README.vxworks says: 4) When using the "tcp" command: Security is non-existent - you may need to establish your own in "tcpTrustedHost()", or implement your own "login" command. You can demo tclTCP by sourcing tests/test.tcp from tclMain, and then unixshell% telnet 1230 info commands I could not find the test.tcp file. Does someone have this file? craig cgonter@inept.stanford.edu From daemon@vxw.ee.lbl.gov Tue Apr 25 04:00:28 1995 From: daemon@vxw.ee.lbl.gov Date: Tue Apr 25 04:00:38 PDT 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Tue Apr 25 04:00:18 PDT 1995 Subject: Re: The use of source code control Subject: Using Tcl with VxWorks Subject: Using Tcl with VxWorks Subject: Re: SCSI tape question resolution Subject: NFS Server Subject: Re: The use of source code control ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: The use of source code control Date: 24 Apr 1995 13:32:19 GMT From: dyer@alx.alx.sticomet.com (Doug Dyer) Organization: Naval Research Laboratory Message-ID: <3ng993$cti@ra.nrl.navy.mil> References: <9504241120.AA04899@hea.sron.ruu.nl> vdbent@hea.sron.ruu.nl (Kees van der Bent) writes: >Hi, >Can some of you please comment on what source code control system they like >the most (for use with small/medium sized VxWorks applications). Currently >I can choose between SCCS and RCS; which one is best and/or are there others? >Thanks in advance, RCS is much faster than SCCS. RCS will use "reverse-diffs" meaning that it is very fast when extracting revisions close to the current release. SCCS will get slower the deeper your revision history is (unless you extract the original entries). release RCS stores 2.0 diff 2.0-1.9 1.9 diff 1.9-1.5 1.5 diff 1.5-1.0 1.0 1.0 The GNU RCS package contains a paper on this (SCCS vs RCS). With some handy CSH scripts I built a reasonably nice collection of tools (fairly project specific, btw) for RCS. SCCS is "over the hill" so to speak. pitfalls and good stuff on RCS: - ------------------------------- It is difficult at best to control revisions when your application involves clusters of source organized by subdirectories. RCS will store the file reverse diff, but not file location or organization information, therefore if you want to do that, you need to integrate Make and shell scripts with RCS. It can be done, but not very well. You can store a symbolic release name to your file as well. Ie: a particular library is up to v19.2, but the package the library contains is "v2.0". This and the usual abilities to merge and branch revisions (and an "rcsdiff" utility) work out fairly well. Locks! RCS will lock files preventing access by others (if desired). CVS... RCS "solution"? - --------------------------------- While RCS is good for small projects that do not span directory trees or "multi-developer", "multi-group" environments, CVS builds on RCS (it requires RCS) and again comes with a good paper on itself. The problem with CVS (and why I did not use it) was its inability to lock source files. Locking is apparently some controversial issue in revision control circles, but dammit I want it :) To me, merging is a scary issue and Id rather loose the parallel development in sharing files for this reason. Its your call, though. Get both, read the two papers, and play! - --------------------- A very simple sample script using RCS to extract a file with locking. It is a good idea to shelter developers from the revision software as you can build an extra layer of security between the user and the utilities. #! /bin/csh -f # STI reserve script clone from vax CMS to unix RCS... # Doug Dyer -STI # $Id: reserve,v 1.1 1993/08/03 18:24:06 dyer Exp $ # #obtain filename if ($#argv >= 1) then set fname = $argv[1] else echo "Enter FILENAME:" set fname = $< endif #get a path? if ($#argv == 2) then set fpath = $argv[2] if ( !(-d $fpath)) then echo path invalid: $fpath exit endif goto PERFORM else set fpath = "." endif #print information for user verification echo DATA: echo FILE...$fname echo # proceed with operation? echo "Proceed (y/n)?";set proceed = $< if ($proceed != y) then echo "Reserve will checkout a file for you" exit endif PERFORM: #perform operation pushd $fpath > /dev/null co -l $RCSROOT/$fname,v popd > /dev/null - ------------------------- good luck, Doug - -- Doug Dyer - dyer@alx.sticomet.com | Real-time systems . * //[]// Software Technology, Inc. (STI) | Space flight/ground software ^^ . DC office: (703) 329-9707 | vxWorks BSPs, etc. . --------------------------- Newsgroups: comp.os.vxworks Subject: Using Tcl with VxWorks Date: 24 Apr 1995 13:31:06 GMT From: petrovci@tif145.ed.ray.com (Steve Petrovcin) Organization: Equipment Division, Raytheon Company Message-ID: <3ng96q$c1n@news.ed.ray.com> - -- Opinions expressed here are mine, and not my employer's. =================================================================== Steve Petrovcin petrovci@prt.ed.ray.com Raytheon Electronic Systems Division (508) 490-2042 1001 Boston Post Road, Marlborough, MA 01752 m/s 3-2-3925 --------------------------- Newsgroups: comp.os.vxworks Subject: Using Tcl with VxWorks Date: 24 Apr 1995 13:54:52 GMT From: petrovci@tif145.ed.ray.com (Steve Petrovcin) Organization: Equipment Division, Raytheon Company Message-ID: <3ngajc$cf4@news.ed.ray.com> I am trying to use the Tcl scripting language to allow an interpreter running on a Sun to communicate with an interpreter running under VxWorks. We have found a version of Tcl-TCP that allows this (the VxWorks hooks to Tcl-TCP were in place). We were hoping to use Tcl-dp so that remote procedure calls are available. Has anyone ported this for VxWorks? Does anyone know if it needs to be ported or what the scope of such a port would be? Any help would be greatly appreciated. - -- =================================================================== Steve Petrovcin petrovci@prt.ed.ray.com Raytheon Electronic Systems Division (508) 490-2042 1001 Boston Post Road, Marlborough, MA 01752 m/s 3-2-3925 --------------------------- Newsgroups: comp.os.vxworks Subject: Re: SCSI tape question resolution Date: Mon, 24 Apr 1995 17:35:23 GMT From: jfinley@netcom.com (John Finley) Organization: NETCOM On-line Communication Services (408 261-4700 guest) Message-ID: References: <9504231131.AA00930@rst.co.il> Sender: jfinley@netcom22.netcom.com Leonid Rosenboim (leonid@rst.co.il) wrote: : Here is some more advise based on our experience with SCSI Tapes : on VxWorks: [warnings re: "turbo-dats" deleted] : Our recommendation to transfer-rate-hungry users is 8mm. Has anyone used a "tape-RAID" system? These put (e.g.) five drives in a box and stripe across four + parity, making the group look like a single SCSI target (I think). Seems like a good solution technically, though I don't know how much they cost. John Finley --------------------------- Newsgroups: comp.os.vxworks Subject: NFS Server Date: Mon, 24 Apr 1995 20:07:35 GMT From: danc@lds.loral.com (Dan Cunningham) Organization: Loral Data Systems, Sarasota, FL Message-ID: Sender: news@lds.loral.com I am looking for an NFS server that runs with vxWorks. The FAQ pointed me to ftp.netcom.com:/pub/hjb/vxnfsd.tar.gz but the vxnfsd.tar.gz file is no longer there. Does anyone know where this file is now maintained?? Does anyone know of other NFS servers that run on vxWorks (free or not)?? Thanks, Dan - -- ******************************************************************************* Dan Cunningham e-mail to: danc@lds.loral.com __ ______ ______ ______ __ / / / __ / / __ / / __ / / / / / / / / / / /_/ / / /_/ / / / / /___ / /_/ / / __ / / __ / / /___ /_____/ /_____/ /_/ \_\ /_/ /_/ /_____/ All standard disclaimers, etc., etc.... ******************************************************************************* --------------------------- Newsgroups: comp.os.vxworks Subject: Re: The use of source code control Date: 25 Apr 1995 02:52:15 GMT From: D'Anne Thompson Organization: National Optical Astronomy Observatories, Tucson, AZ, USA Message-ID: <3nho4v$jq@noao.edu> References: <9504241120.AA04899@hea.sron.ruu.nl> <3ng993$cti@ra.nrl.navy.mil> dyer@alx.alx.sticomet.com (Doug Dyer) wrote: > > vdbent@hea.sron.ruu.nl (Kees van der Bent) writes: > > >Hi, > > >Can some of you please comment on what source code control system they like > >the most (for use with small/medium sized VxWorks applications). Currently > >I can choose between SCCS and RCS; which one is best and/or are there others? > > >Thanks in advance, > > RCS is much faster than SCCS. RCS will use "reverse-diffs" meaning that > it is very fast when extracting revisions close to the current > release. SCCS will get slower the deeper your revision history is > (unless you extract the original entries). > ..snip.. > > The GNU RCS package contains a paper on this (SCCS vs RCS). With some > handy CSH scripts I built a reasonably nice collection of tools (fairly project > specific, btw) for RCS. > > SCCS is "over the hill" so to speak. > ..snip... Let's make the case for SCCS. Since in my normal environment, working (read-only) copies of each file are normally checked out at all times, the issue of 10 msec versus 5 msec to retrieve a specific version of a file is totally moot. The one real gotcha to RCS would be the fact that it is not fully compatible with standard 'make' packages (sun os or solaris - mileage may vary). SCCS is fully integrated with all conventional make packages to the extent that make will automatically look for SCCS directories and check out SCCS files as needed to perform a make unit. You can get around this by hacking up all your makefiles to make it RCS aware, or you can switch to the GNU make utility. Do you have to distribute your code and Makefiles? If you do then realize that SCCS is the least common denominator. A lot more sites have SCCS than RCS. I still use SCCS because I don't want to be dragged into the whole GNU scene. RCS is part of the GNU package. It is only fully compatible with the GNU make program. Then you need GNU bin utils, GNU compiler, and everything else GNU, even the dreaded emacs. (I'm flame immune - - do your worst!) I like to keep my life simple. I like simple tools. SCCS performs the basic source code control function very nicely. RCS does provide some new features, but I haven't need them yet. Note too, that there are commercial packages that provide very advanced features - there are alternatives to both RCS and SCCS. D'Anne Thompson dat85715@indirect.com --------------------------- End of New-News digest ********************** From mukherje@kns.com Tue Apr 25 08:02:45 1995 From: mukherje@kns.com (Arup Mukherjee) Date: Tue Apr 25 08:02:51 PDT 1995 Subject: Tcl for VxWorks Hi all, Can anyone point me to a FTP site or otherwise where I can get the source or binaries for using Tcl/Tk on VxWorks 5.1.x Thanks in advance -Arup Mukherjee (arup@kns.com) Kulicke and Soffa Industries Willow Grove, Pa 19090 From djb@rtp.co.uk Tue Apr 25 08:39:15 1995 From: djb@rtp.co.uk (David Bullimore) Date: Tue Apr 25 08:39:21 PDT 1995 Subject: ObjecTime Dear All, Re: ObjecTime for VxWorks Seminar. ---------------------------------- As a part of Wind River Systems' WindNet Partners program, ObjecTime Ltd of Canada are now offering ObjecTime with complete code generation for VxWorks. ObjecTime is the first object orientated CASE tool specifically designed for implementing real time systems. Using ROOM (Real Time Object Orientated Modelling methodology), developers can construct models which may run under simulation on teh host or directly on a VxWorks target. RTP are running a two days hands-on workshop on ObjecTime for VxWorks users on 25-26 May 1995 in Birmingham, Emgland. If anyone is interested in attending (or just having information on the product), please contact me by email or phone. David *********************************************** * David Bullimore * * Real Time Products Ltd. (RTP Ltd.) * * Unit 8 * * Holt Court * * Aston Science Park * * Birmingham * * B7 4EJ * * England * * * * Phone: +44 121 333 6955 * * Fax: +44 121 333 5433 * * * * Email djb@rtp.co.uk * *********************************************** From hill@luke.atdiv.lanl.gov Tue Apr 25 09:53:08 1995 From: hill@luke.atdiv.lanl.gov (Jeff Hill) Date: Tue Apr 25 09:53:14 PDT 1995 Subject: efficient max free block test Hello all, I have implemented a TCP/IP server. This server needs to allocate memory on behalf of clients. However, the server should refuse to allocate memory on behalf of clients if some measure of the amount of free dynamic memory remaining has dropped below a predetermined threshold. Maintaining this dynamic memory pad will guarantee that the rest of our system has a reasonable dynamic memory pool to support reliable operation. Under vxWorks I have found the routine "memFindMax()" which returns the maximum free block in the system memory partition. However, after further investigation, I discovered that this routine performs a linear search checking every free block to see which one is the largest. While this will produce the correct result I feel uncomfortable calling this routine just prior to each call to malloc() because of concerns about the overhead involved. My current solution is to poll "memFindMax()" infrequently and then set a flag which determines if the server may allocate memory on behalf of clients. This appears (to me) to be a less than satisfactory solution because the server runs very quickly and might consume all available memory including a reasonable pad which I wish to maintain for proper operation before the next poll to "memFindMax()" gets a chance to run. So my question is: Is there a routine that will provide information about the amount of free memory remaining that does not incur the linear search overhead of memFindMax()? Could the dynamic memory system be rewritten to maintain the max free block size with the partition each time that malloc() and free() are called? If so then memFindMax() could return the value of this variable immediately without incurring any significant overhead. I suppose that I could create a independent memory partition for the sole use of this server however this would require prediction of the actual resource loading mix that will actually occur in the system at run time. My preference is to keep all allocations coming from the same pool and to have access to a fast test that will give me a reasonable measure of the remaining free memory. Jeff ______________________________________________________________________ Jeffrey O. Hill Internet johill@lanl.gov LANL MS H820 Voice 505 665 1831 Los Alamos, NM 87545 USA FAX 505 665 5107 From rpitts@scires.com Tue Apr 25 15:21:23 1995 From: rpitts@scires.com Date: Tue Apr 25 15:21:29 PDT 1995 Subject: Re: SCSI tape question resolution >Has anyone used a "tape-RAID" system? These put (e.g.) five drives >in a box and stripe across four + parity, making the group look like >a single SCSI target (I think). Seems like a good solution >technically, though I don't know how much they cost. The 8mm 'tape-RAID' drives I've seen are by Contemporary Cybernetics and HI-PAR. Both vendor's prices, I believe, are easily over $10K and push $18-21K depending on whether there are 2 or four drives. A four drive unit's storage capacity is ~20 Gigabytes and the transfer rates are 2 Megabytes/second uncompressed, which is essentially 5GB at 500 KB/s each for four 8 mm drives. These are SCSI II devices. HI-PAR tape format is ANSI and they support 4mm and 8mm. Other tape drives that are of interest are: Quantum's DLT4000 SCSI II OEM drive - 20 GB/tape (uncompressed) at 1.5 MB/s. (currently available.) I've had quotes of $6K, with list at ~$8K. Proprietary tape format. Exabyte's MAMMOTH 8 mm drive to be released 2Q95(beta?): 20 GB/tape at 3 MB/s. Target price is $6K. IRDC (Improved Data Recording Capability) tape format and is SCSI II. [All of the above drives quote double the capacity and storage rate with 2:1 compression.] We are looking seriously at the Quantum drive, which was a Digital Equipment product, and if our schedule slips, we will move toward the MAMMOTH. Richard Pitts Senior Software Engineer Scientific Research Corporation (404)859-9161 2300 Windy Ridge Parkway Suite 400 South (404)953-6241 FAX Atlanta, Georgia 30339 rpitts@scires.com From daemon@vxw.ee.lbl.gov Wed Apr 26 04:00:30 1995 From: daemon@vxw.ee.lbl.gov Date: Wed Apr 26 04:00:39 PDT 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Wed Apr 26 04:00:20 PDT 1995 Subject: Re: The use of source code control Subject: Any Token Ring Drivers out there? Subject: Re: NFS Server Subject: WWW/URL to access Wind River Systems ? Subject: VxWorks 5.2 advantages?? Subject: repost VxWorks 5.2 advantages?? ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: The use of source code control Date: 25 Apr 1995 11:47:10 GMT From: dyer@alx.alx.sticomet.com (Doug Dyer) Organization: Naval Research Laboratory Message-ID: <3ninfu$9h5@ra.nrl.navy.mil> References: <9504241120.AA04899@hea.sron.ruu.nl> <3ng993$cti@ra.nrl.navy.mil> <3nho4v$jq@noao.edu> D'Anne Thompson writes: >dyer@alx.alx.sticomet.com (Doug Dyer) wrote: >> >> vdbent@hea.sron.ruu.nl (Kees van der Bent) writes: >> >> >Hi, >> >> >Can some of you please comment on what source code control system they like >> >the most (for use with small/medium sized VxWorks applications). Currently >> >I can choose between SCCS and RCS; which one is best and/or are there others? >> >> >Thanks in advance, >> >> RCS is much faster than SCCS. RCS will use "reverse-diffs" meaning that >> it is very fast when extracting revisions close to the current >> release. SCCS will get slower the deeper your revision history is >> (unless you extract the original entries). >> >..snip.. >> >> The GNU RCS package contains a paper on this (SCCS vs RCS). With some >> handy CSH scripts I built a reasonably nice collection of tools (fairly project >> specific, btw) for RCS. >> >> SCCS is "over the hill" so to speak. >> >..snip... >Let's make the case for SCCS. OK, but I get a rebuttal :) >Since in my normal environment, working (read-only) copies of each file >are normally checked out at all times, the issue of 10 msec versus >5 msec to retrieve a specific version of a file is totally moot. Your mileage will vary on this one. In SCCS the deeper the revision list, the more significant the changes, the longer you wait. I have experienced 10-15 seconds for older source code. >The one real gotcha to RCS would be the fact that it is not fully >compatible with standard 'make' packages (sun os or solaris - mileage >may vary). SCCS is fully integrated with all conventional make packages >to the extent that make will automatically look for SCCS directories >and check out SCCS files as needed to perform a make unit. You can >get around this by hacking up all your makefiles to make it RCS aware, >or you can switch to the GNU make utility. Good point. But you do not need RCS support integrated into Make since it modifies the end of file names not the beginning. Here is an addition you can include in your Makefile to solve this: - -- snip -- .SUFFIXES : .c,v # "refresh" our rules, a pain but ... CO = co # define the "Check Out" RCS program to use #RCS Rules for you folks #Rule one: The object is dependent on the RCS history file .c,v.o : ${CO} $< #check out the .c file from .c,v ${CC} ${CFLAGS} -c $*.c #compile the .c file to a .o - rm -f $*.c #delete the source.. tada! #Rule 2: The source file is dependent on the RCS history file .c,v.c : ${CO} $? #just need source, check it out - -- snip -- Now the above assumes your RCS provides the ",v" extension. SCCS *requires* that make understands it since rather than taking joe.c -> joe.c,v as RCS will do, it modifies a file as joe.c -> ,joe.c Make naturally supports extensions and we can take advantage of it :) >Do you have to distribute your code and Makefiles? If you do then >realize that SCCS is the least common denominator. A lot more sites >have SCCS than RCS. But there are those without SCCS or Make, making GNU very attractive (as we see in their GCC compiler). In my projects, I always establish gnu-make early on on all the systems. The reason is that Makes themselves differ in many strange ways. One nice thing about GNU make is its support of VPATH (but Solaris has some attractive features too). But I digress :) Back to revision control: >I still use SCCS because I don't want to be dragged into the whole >GNU scene. RCS is part of the GNU package. It is only fully compatible >with the GNU make program. Then you need GNU bin utils, GNU compiler, >and everything else GNU, even the dreaded emacs. (I'm flame immune >- do your worst!) I hear you! You make valid points, and I hate emacs too :) >D'Anne Thompson >dat85715@indirect.com - -- Doug Dyer - dyer@alx.sticomet.com | Real-time systems . * //[]// Software Technology, Inc. (STI) | Space flight/ground software ^^ . DC office: (703) 329-9707 | vxWorks BSPs, etc. . --------------------------- Newsgroups: comp.os.vxworks Subject: Any Token Ring Drivers out there? Date: 25 Apr 1995 16:38:35 GMT From: Alon Segal Organization: Electronics for Imaging, San Mateo, CA Message-ID: <3nj8ib$rjr@rmstar.efi.com> Hello, I was wondering if anyone knows about VxWorks drivers for Token Ring Devices (TI or any other). WRS response was try the news group. Help!!! Alon Segal Electronics For Imaging alon@efi.com --------------------------- Newsgroups: comp.os.vxworks Subject: Re: NFS Server Date: Tue, 25 Apr 1995 16:46:31 GMT From: jfinley@netcom.com (John Finley) Organization: NETCOM On-line Communication Services (408 261-4700 guest) Message-ID: References: Sender: jfinley@netcom22.netcom.com Dan Cunningham (danc@lds.loral.com) wrote: : I am looking for an NFS server that runs with vxWorks. The FAQ pointed : me to ftp.netcom.com:/pub/hjb/vxnfsd.tar.gz but the vxnfsd.tar.gz file : is no longer there. Does anyone know where this file is now maintained?? : Does anyone know of other NFS servers that run on vxWorks (free or not)?? There will be one in 5.2, I think. Perhaps you could talk to Wind River and see what your options are there (for an unbundled and/or early version). Otherwise, there's {Free,Net,386}BSD, and Linux, for free sources. Unfortunately, the hard part of NFS is the file handle and directory stuff, which will be very different for the UFS versions (from your DOSFS goal). Maybe not so much hard, as tricky. John Finley --------------------------- Newsgroups: comp.os.vxworks Subject: WWW/URL to access Wind River Systems ? Date: Tue, 25 Apr 1995 14:25:05 GMT From: Marc BRANDT Organization: Hewlett-Packard France Message-ID: <3nj0o1$hti@hpgndlab.grenoble.hp.com> Any address (ftp) or URL for accessing info about Wind River Systems ' products ? tia marc --------------------------- Newsgroups: comp.os.vxworks Subject: VxWorks 5.2 advantages?? Date: Tue, 25 Apr 1995 07:47:39 From: hwest@cheetah.sed.csc.com (Harold West) Organization: Computer Sciences Corporation Message-ID: Hello --------------------------- Newsgroups: comp.os.vxworks Subject: repost VxWorks 5.2 advantages?? Date: Tue, 25 Apr 1995 07:50:54 From: hwest@cheetah.sed.csc.com (Harold West) Organization: Computer Sciences Corporation Message-ID: Hello Has anyone working with the VxWorks 5.2 betas seen any increase in the performance of the TCP/IP network drivers? We currently use 5.1 and are wondering if there is any advantage to upgrading to 5.2? Thanks Harold West --------------------------- End of New-News digest ********************** From daemon@vxw.ee.lbl.gov Thu Apr 27 04:59:37 1995 From: daemon@vxw.ee.lbl.gov Date: Thu Apr 27 04:59:46 PDT 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Thu Apr 27 04:00:28 PDT 1995 Subject: Re: Using Tcl with VxWorks Subject: Re: The use of source code control Subject: Re: Any PEP/vxWorks users out there? Subject: Looking for Code Coverage/RMA tools Subject: Re: WWW/URL to access Wind River Systems ? Subject: Metrum VLDS SCSI drivers Subject: Exabyte SCSI Drivers Subject: write data to specified address, help needed Subject: Re: NFS Server ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Using Tcl with VxWorks Date: 26 Apr 1995 16:33:38 GMT From: nigels2@synoptics.com (Nigel Standing) Organization: Bay Networks, Inc. Message-ID: <3nlsl2$79@inn.synoptics.com> References: <3ngajc$cf4@news.ed.ray.com> petrovci@tif145.ed.ray.com (Steve Petrovcin) writes: > I am trying to use the Tcl scripting language to allow an interpreter > running on a Sun to communicate with an interpreter running under > VxWorks. We have found a version of Tcl-TCP that allows this (the > VxWorks hooks to Tcl-TCP were in place). We were hoping to use > Tcl-dp so that remote procedure calls are available. Instead of using Tcl-TCP (sorry no experience with that) you might want to consider using expect. In the words of John Ousterhout, "expect is one of the oldest Tcl applications and also one of the most popular. It is a program that 'talks' to interactive programs." What this means is that an expect script being interpreted on your Sun can 'talk' to the VxWorks shell interpreter via tip, telnet, or rlogin (whatever you want). Once your "remote procedures" are loaded on the VxWorks target, then from within the expect script, you would just execute the remote procedures as you would from the VxWorks shell (of course the expect script could check to see whether the remote procedures were already loaded, and if not automatically load them for you). BTW The VxWorks BSP Porting Kit uses expect as the basis for the validation test suite. Nigel - --------------------------- Newsgroups: comp.os.vxworks Subject: Re: The use of source code control Date: 26 Apr 1995 16:35:24 GMT From: nigels2@synoptics.com (Nigel Standing) Organization: Bay Networks, Inc. Message-ID: <3nlsoc$7d@inn.synoptics.com> References: <9504241120.AA04899@hea.sron.ruu.nl> vdbent@hea.sron.ruu.nl (Kees van der Bent) writes: > Can some of you please comment on what source code control system they > like the most (for use with small/medium sized VxWorks applications). > Currently I can choose between SCCS and RCS; which one is best and/or > are there others? There are many commercial and non-commercial offerings out there; check out "comp.software.config-mgmt", also the derived FAQ available on the Web at: http://www.iac.honeywell.com/Pub/Tech/CM/index.html another interesting page is: http://www.cs.colorado.edu:80/users/andre/configuration_management.html Since you only mention SCCS and RCS, perhaps you're looking for a non-commercial (read: free) solution, in which case you might want to take a look at CVS; CVS is built on top of RCS and provides the ability to manage groups of configuration items. More info at: http://www.winternet.com/~zoo/cvs/ happy hunting. Nigel - --------------------------- Newsgroups: comp.os.vxworks Subject: Re: Any PEP/vxWorks users out there? Date: Wed, 26 Apr 1995 17:00:28 GMT From: mikewd@leica.co.uk (Mike Wilmot-Dear) Organization: Leica Cambridge Ltd. Message-ID: References: <3n8i38$6b7@news.blkbox.com> John Cobarruvias (johncoby@blkbox.com) wrote: : We recently have purchased a PEP VM30 board from PEP cpmputers and : their BSP. I am curious if any other organization is using PEP boards : with vxWorks? We are using PEP VSBC4 boards with VxWorks we use VLAN boards to provide the ethernet i/f for development. I'm not directly involved with that project group any more but I know they did originally have some problems in building a working standalone PROM based system (without ethernet). But this was with the 5.02 BSP and I think it's now working o.k. with the 5.1 BSP. (I think part of the difficulty originally is the same as occurs with all the BSP's that are done by board manufacturers and not directly supported by WRS, in that problems get bounced between them as to whether it's in the BSP or the generic VxWorks). Otherwise I think we've been happy enough with the boards and they're useful if you just want a small embedded controller and are used to using VxWorks. A lot of the other boards with VxWorks BSP's available for them tend to be rather high end and rather too expensive to use for simple controller type aplications. Mike. - -- - ------------------------------------------------------------------------------ +*I'm not paid to make official comments for Leica, so the above isn't one.*+ Mike Wilmot-Dear (MW342) e-mail: mikewd@leica.co.uk Leica Cambridge Ltd. Tel: +44 1223-411411 (x347) Clifton Road Fax: +44 1223-210692 Cambridge, CB1 3QH, UK --------------------------- Newsgroups: comp.os.vxworks Subject: Looking for Code Coverage/RMA tools Date: 26 Apr 1995 16:22:27 GMT From: dwybo01@kocrsv01.delcoelect.com (David Wybo) Organization: Delco Electronics Corp. Message-ID: <3nls03$sro@mloeff01.elec.mid.gmeds.com> I'm looking for info on what other VxWorks customers might be using in the way of Code Coverage tools. Rate Monotonic Analysis tools are also of interest. Anyone out there using such things in their projects? - -- Dave Wybo Phone: (810) 684-3547 Wind River Systems Fax: (810) 684-3755 483-303-301 General Motors Proving Grounds --------------------------- Newsgroups: comp.os.vxworks Subject: Re: WWW/URL to access Wind River Systems ? Date: 26 Apr 1995 19:30:55 GMT From: pwang@bnr.ca (Peter-Zhi Wang) Organization: Bell-Northern Research Ltd. Message-ID: <3nm71f$scn@bcarh8ab.bnr.ca> References: <3nj0o1$hti@hpgndlab.grenoble.hp.com> In article <3nj0o1$hti@hpgndlab.grenoble.hp.com>, Marc BRANDT wrote: >Any address (ftp) or URL for accessing info about Wind River Systems ' products >? http://www.wrs.com Peter --------------------------- Newsgroups: comp.os.vxworks Subject: Metrum VLDS SCSI drivers Date: 26 Apr 1995 21:30:59 GMT From: kressig@ti.com (Bob Kressig) Organization: Texas Instruments Message-ID: <3nme2j$1fr@ganesh.mc.ti.com> I'm looking for VLDS SCSI Drivers for vxWorks. Presently we are running vxWorks 5.1.1 on a Force CPU-44. Thanks, - -- Bob Kressig Msgid: BOBK Email: kressig@ti.com Texas Instruments Voice: 214-995-3404 Fax: 214-995-0456 --------------------------- Newsgroups: comp.os.vxworks Subject: Exabyte SCSI Drivers Date: 26 Apr 1995 21:32:28 GMT From: kressig@ti.com (Bob Kressig) Organization: Texas Instruments Message-ID: <3nme5c$1fr@ganesh.mc.ti.com> I'm looking for exabyte scsi drivers for an exabyte 8200 drive and vxworks version 5.1.1. Thanks, - -- Bob Kressig Email: kressig@ti.com Texas Instruments Voice: 214-995-3404 Fax: 214-995-0456 --------------------------- Newsgroups: comp.os.vxworks Subject: write data to specified address, help needed Date: 27 Apr 1995 07:18:34 GMT From: d797302@Oz.nthu.edu.tw (Shen-Kuang Chin) Organization: National Tsing Hua University, HsinChu, Taiwan Message-ID: <3nngga$po@Emerald.Oz.nthu.edu.tw> Dear Netter, I'm using VxWorks 5.1.1 with SUN SPARC 10 as our host. Our target is v4f (CPU: 68040). We have two other boards installed on VME bus. We need to write data (usually, a Word or a Long Word) to a specified absolute address. Is there any function available for doing these ? I also tried to use pointer but with some problems. It would be appreciated if someone could give me some instructions. I'm new in both VxWorks and GNU C compiler (for cross compile). Thanks, S.K.King 4/27/95 /------------------------------------------------------------------\ | Graduate in Physics | Sun-Kun King | | National Tsing Hua University | | | Hsinchu, Taiwan, R.O.C. | g797302@phys.nthu.edu.tw | \------------------------------------------------------------------/ --------------------------- Newsgroups: comp.os.vxworks Subject: Re: NFS Server Date: Thu, 27 Apr 95 07:57:07 GMT From: squeeby@hooked.net (squeeby) Organization: peaceful star Message-ID: <3nnip5$g09@its.hooked.net> References: In article , danc@lds.loral.com (Dan Cunningham) wrote: >I am looking for an NFS server that runs with vxWorks. The FAQ pointed >me to ftp.netcom.com:/pub/hjb/vxnfsd.tar.gz but the vxnfsd.tar.gz file >is no longer there. Does anyone know where this file is now maintained?? > >Does anyone know of other NFS servers that run on vxWorks (free or not)?? > >Thanks, >Dan i don't post to this group anymore, but since this inquiry deserved an answer... vxnfsd stuff is deleted because WRS reportedly has nfs server already. it was a quick hack anyway. ftp.netcom.com does not have vxworks stuff i used to keep. the FAQ is no longer there for example. in fact, i'm no longer a subscriber of netcom. you should check teacher.sparta.mclean.com instead. and, mike a., please, for the last time, update the FAQ via news.answers so that the old archived version disappears. thanks. love, hwajin Squeeby http://www.hooked.net/users/squeeby/ --------------------------- End of New-News digest ********************** From brightca@ncp.gpt.co.uk Thu Apr 27 06:18:22 1995 From: "Chris Brightman" Date: Thu Apr 27 06:18:29 PDT 1995 Subject: How to unsubscribe? How do I unsubscribe from the vxworks_users mailing list? I have tried sending mail to vxwexplo-request, but have had no success. +--------------------------------------------------------+ | Chris Brightman GPT Business Systems | | Software Design Engineer Technology Drive | | brightca@ncp.gpt.co.uk Beeston | | Telephone: +44 115 943 2254 Nottingham, UK | | Fax: +44 115 943 3805 NG9 1LA | +--------------------------------------------------------+ From luke@wrs.com Thu Apr 27 11:24:56 1995 From: Luke Teyssier Date: Thu Apr 27 11:25:02 PDT 1995 Subject: RMA Tools Tri-Pacific offers a tool called PERTS for RMA analysis. They also offer a whole set of services relating to Rate Monotonic Analysis training. Their phone number is (510) 814-1770 or CompuServe at 76434,117 From swtillman@ccgate.hac.com Thu Apr 27 17:59:45 1995 From: swtillman@ccgate.hac.com Date: Thu Apr 27 17:59:51 PDT 1995 Subject: Development Tools We are considering using Centerline's ObjectCenter or Green Hills' MULTI to develop C++ software for VxWorks. I would like to hear about anyone's experiences with these tools. Thank you Scott Tillman Software Engineer Hughes Missile Systems Co. From vilardebo@nrlvax.nrl.navy.mil Thu Apr 27 20:07:57 1995 From: vilardebo@nrlvax.nrl.navy.mil Date: Thu Apr 27 20:08:05 PDT 1995 Subject: Using VXGDB with multiple tasks How do I use VxGDB to debug tasks that are spawned from a main task? I have tried using the file and attach commands but I get a strange message that I am reading floating point registers--not what I expected at all. I'm using VxWorks 5.1, VxGDB 2.0 on Force CPU 40. Any hits would be appreciated. Ken Vilardebo Mnemonics, Inc. From clundhol@sw.seisy.abb.se Thu Apr 27 23:35:26 1995 From: 'Claes Lundholm' Date: Thu Apr 27 23:35:32 PDT 1995 Subject: Profibus Dear all, I'm trying to find out if there is any VxWorks support for Profibus. Any information regarding VxWorks drivers/boards or whatever would be useful. Thanks. Claes -- +---------------------------.----------------------------------+ ! Claes Lundholm ! Tel: +46 21 342665 ! ! ABB Industrial Systems ! Fax: +46 21 342555 ! ! Dept. LKSK ! e-mail: clundhol@sw.seisy.abb.se ! ! S-721 67 Vasteras, Sweden ! ! +---------------------------'----------------------------------+ From simon.trew@def.bae.co.uk Fri Apr 28 00:33:05 1995 From: simon.trew@def.bae.co.uk (Simon Trew) Date: Fri Apr 28 00:33:15 PDT 1995 Subject: Re: Development Tools swtillman@ccgate.hac.com (Scott Tillman) wrote: We are considering using Centerline's ObjectCenter or Green Hills' MULTI to develop C++ software for VxWorks. I would like to hear about anyone's experiences with these tools. We are using MULTI in a Windows environment. I assume it is also available under X-Window as it is based on one of these Windows ports rather than a native Windows look & feel. Therefore one has all X Window controls which, while similar to Windows, take a little getting used to, e.g. the focussing and position of scroll bars is different. It also has an internal copy/paste buffer so things aren't copied/pasted between other Windows apps. It saves files with different line-end format from most DOS applications so they can't be easily read in Notepad etc. On the good side, it does seem to work reasonably well, although for ourselves we cannot use any of the interactive debugging tools. There is a built-in source code control system which is OKish, but we use MKS RCS as this is standard on all projects. The compiler seems reasonable although it does not have many of the more modern messages abt variables being unused, assignments inside if's etc., which makes one wonder how much optimisation it does. The worst thing abt the compiler/makefile is that we have never been able to get it to work across multiple directories, i.e. it always seems to put the outfile in the current directory, which is a pain on any nontrivial project where files are installed in several directories as one cannot do an all-encompassing build, one must build each directory separately. This possibly paints MULTI in a worse light than it actually is, as it does work well enough and for day-to-day use, once one has got used to the X Window windows, it is more bug-free than usual and one can have confidence in it. But there are many "niggles" which stops it being a truly great package. If you are considering Multi in a non-Windows environment, some of these comments may not apply. Si. -------------------------------------------------------------------------- --- A fool and his money are soon elected. -------------------------------------------------------------------------- --- Work: 01438 753663 | Simon Trew, Software Engineer | email:trew@def.bae.co.uk Home: 01438 455931 | British Aerospace Defence Ltd | Fax: 0438 756293 From js@rtp.co.uk Fri Apr 28 03:06:11 1995 From: js@rtp.co.uk (Jatinder Sangha) Date: Fri Apr 28 03:06:18 PDT 1995 Subject: Benelux Countries Hi VxWorkers, Sorry to waste bandwidth like this but can organisations/individuals working with VxWorks in the Benelux countries - Belgium, the Netherlands, Luxembourg, Holland etc please email me at js@rtp.co.uk TIA --Jatinder From daemon@vxw.ee.lbl.gov Fri Apr 28 04:00:33 1995 From: daemon@vxw.ee.lbl.gov Date: Fri Apr 28 04:00:41 PDT 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Fri Apr 28 04:00:22 PDT 1995 Subject: FTP Login - Possible to establish userid and passwd in 486 Arch Subject: Task locking in VxWorks... Subject: Re: FTP address for Gnu software / 68060 Subject: VME mfr. hiring ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: FTP Login - Possible to establish userid and passwd in 486 Arch Date: 27 Apr 1995 08:40:16 -0400 From: et1@baretta.ideas.com (Everett Tom) Organization: Ideas, Inc. Message-ID: Does VxWorks Version 5.11 support FTP logins which require a userid and password? If it means anything to you, I am running in a 486 architecture. My full purpose is to create a list of valid userid with passwords who can ftp login to my target. Any helpful input would be GREATLY apprciated. If possible, please send a response directly to me. If not, a post to the newsgroup would do just fine. Thank you in advance. Everett --------------------------- Newsgroups: comp.os.vxworks Subject: Task locking in VxWorks... Date: Thu, 27 Apr 1995 05:22:39 GMT From: joshisx@ntmtv.com (Suhas Joshi) Organization: Northern Telecom, Mountain View Message-ID: Sender: news@ntmtv.com Hi, I have a doubt using the taskLock. I want to protect certain sections of my code. So what will be the effect of taskLock(); .... taskUnLock(); against .. simTake(); ... simGive(); What are the advantages of each approach and dis advantages. Will it affect me in debugging the code. As of my understanding I won't be able put a break point in the code in the critical section. Is it true. Regards Suhas Joshi. --------------------------- Newsgroups: comp.os.vxworks Subject: Re: FTP address for Gnu software / 68060 Date: 27 Apr 1995 18:08:34 GMT From: synergy@nic.cerf.net (John E. Clark) Organization: CERFnet Dial n' CERF Customer Message-ID: <3nomj2$ijf@news.cerf.net> References: <9504211155.AA22199@ada3.ge.com> In article <9504211155.AA22199@ada3.ge.com> phillip.shaffer@ae.ge.com (Phil Shaffer) writes: + +2) Has anyone adapted GCC / GAS for 68060? If so, can you share the +mods? As of gcc-2.6.3 the compiler code gneration templates emit 'mulul dx,dx:dx' which have been elimnated in the '60 instruction set, as well as divul dx,dx:dy and a few other instructions which I've never seen a compiler emit. The solution here is to get the ISP(and FPSP) package mentioned in the MC68060 Manual, Appendix C, via their BBS. (one can telnet to it via pirs.aus.sps.mot.com) It's a bit clumsy, and in the download operation one must select the 'attachment' option to get the actual code. In the case of 'gas' and new '60 control register defintions, obviously it's back to '.word', or '.long' defintions for a while. --------------------------- Newsgroups: comp.os.vxworks Subject: VME mfr. hiring Date: Thu, 27 Apr 1995 23:55:06 GMT From: sarge314@indirect.com (Tom sargent) Organization: Internet Direct, indirect.com Message-ID: Sender: usenet@indirect.com (Internet Direct Admin) Synergy Microsystems, a major VMEbus CPU manufacturer, is looking for a hardware engineer and a diagnostics programmer to work in its Tucson office. The hardware position will work on all phases of VMEbus Single Board Computer development/debug, including a wide variety of I/O functions (Ethernet, SCSI, sync/async serial, parallel, etc.). Experience in complex digital design & debug essential, VMEbus experience very desirable, Motorola 680x0 family experience desirable. The programmer position will be principally concerned with writing and maintaining hardware diagnostics (in C and assembler) used in manufacturing and testing of VMEbus products. Experience with Motorola 680x0 family processors and Unix very desirable. Experience with VxWorks, pSOS, OS9, a plus. Contact Tom Sargent at sarge@ioinc.tucson.az.us or call 520-690-1709 FAX: 520-690-1796. - -- "Experience is what you get when you don't get what you want." --------------------------- End of New-News digest ********************** From ld0304!uchenick@uu5.psi.com Fri Apr 28 06:26:53 1995 From: ld0304!uchenick@uu5.psi.com (Gordon Uchenick) Date: Fri Apr 28 06:27:01 PDT 1995 Subject: Re: Using VXGDB with multiple tasks With vxgdb, you can attach to any task without a problem. If I had your symptoms, I would first make sure that the tasks had been compiled with the -g option. Another interesting problem is how you can use vxgdb to debug a task that crashes the processor very early in its execution. What I did was use the console to put a breakpoint on the task entry point and then let it start. Bring up vxgdb and target the machine. On the console, delete the breakpoint via bdall command. This will leave the task suspended. Then, attach to the task with vxgdb. You can step through it until you find the reason for the crash. Good luck, Gordon ,,, (o o) --------------------------oOO--(_)--OOo---------------------- _/_/_/_/_/_/_/ _/_/_/ _/_/_/_/_/_/_/ _/_/_/_/_/_/ _/_/ _/_/_/_/ _/_/ _/_/ _/_/ _/_/ _/_/ _/_/ _/_/ _/_/ _/_/_/_/_/_/ _/_/ _/_/_/_/_/_/ _/_/ _/_/ _/_/ _/_/ _/_/ _/_/ _/_/ _/_/ _/_/ _/_/ _/_/ _/_/ _/_/ _/_/ _/_/_/_/_/_/ INTEGRATED SYSTEMS, L.P. ------------------------------------------------------------- Gordon M. Uchenick Phone: 410-581-0422 11431 Cronhill Drive, Suite J FAX: 410-581-5738 Owings Mills, MD 21117 EMAIL: uchenick@tate.com "When you lose your momentum, all you have left is inertia." From interactive.com!jimc@indev.portal.com Fri Apr 28 10:55:51 1995 From: "ENG Jim Chee" Date: Fri Apr 28 10:56:01 PDT 1995 Subject: fopen fails when 256 fds in use While running my application, fopen always fails due to S_rpcLib_RPC_TIMEDOUT or S_errno_EACCES when 256 or more fds are in use. I've confirmed that the file I'm trying to open does exist & the user does have the correct privileges. Also, only 7 of the 256 fds in use are for NFS files. I've managed to capture the message traffic between the SUN and my Tadpole TP34 68030 board when this problem occurs, and it shows the following: - NFS GETPORT request from TP34 being retransmitted 11 times. The SUN promptly responds to each GETPORT request, even the 1st request. - NFS LOOKUP request from TP34 being retransmitted 4 times. Again, the SUN promptly responds to each NFS LOOKUP request. I've tried increasing nfsTimeoutSec from the shell & this only delays the amount of time it takes fopen to return an error. I've also tried increasing the number of mbufs, mbuf clusters, & ethernet transmit & receive frames, all to no avail. Does anyone have any clues on what I should try next? JimC. From dowling@cub.kscorp.com Fri Apr 28 12:16:00 1995 From: Bob Dowling Date: Fri Apr 28 12:16:07 PDT 1995 Subject: Device Driver Positions At KineticSystems, we design and manufacture high-performance Data Acquisition equipment for industrial and scientific markets. As our product lines grow, so have our needs for experienced SW professionals to develop device drivers for our many computer bus adapters. Qualifications are a BSCS or equivalent with 3+ years experience with device drivers for any of the following Operating Systems: - VAX/VMS or OpenVMS - Windows 3.X or NT - UNIX: DEC ULTRIX, OSF/1; HP-UX; SunOS, Solaris; SGI Irix; etc. - VxWorks Our focus is on VMS or Windows driver experience, but we will train motivated candidates who have a proven track record on platforms like VxWorks. We offer an excellent compensation package, flexible working hours, a generous relocation allowance, and a full range of benefits. For immediate consideration, please send, fax, or e-mail your resume to: KineticSystems Corporation (ATTN: Personnel/RCD) 900 North State Street FAX: (815) 838-0095 Lockport IL 60441 E-mail: dowling@kscorp.com From daemon@vxw.ee.lbl.gov Sat Apr 29 04:00:30 1995 From: daemon@vxw.ee.lbl.gov Date: Sat Apr 29 04:00:38 PDT 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sat Apr 29 04:00:21 PDT 1995 Subject: Re: Any PEP/vxWorks users out there? Subject: interrupt question ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: Any PEP/vxWorks users out there? Date: 28 Apr 1995 16:31:46 -0400 From: contrav@telerama.lm.com (Contraves) Organization: Telerama Public Access Internet, Pittsburgh, PA Message-ID: <3nrjbi$hik@india.lm.com> References: <3n8i38$6b7@news.blkbox.com> Mike Wilmot-Dear (mikewd@leica.co.uk) wrote: > John Cobarruvias (johncoby@blkbox.com) wrote: > : We recently have purchased a PEP VM30 board from PEP cpmputers and > : their BSP. I am curious if any other organization is using PEP boards > : with vxWorks? > We are using PEP VSBC4 boards with VxWorks we use VLAN boards to > provide the ethernet i/f for development. > I'm not directly involved with that project group any more but I know > they did originally have some problems in building a working standalone > PROM based system (without ethernet). But this was with the 5.02 BSP and > I think it's now working o.k. with the 5.1 BSP. > (I think part of the difficulty originally is the same as occurs with all > the BSP's that are done by board manufacturers and not directly supported > by WRS, in that problems get bounced between them as to whether it's in the > BSP or the generic VxWorks). > Otherwise I think we've been happy enough with the boards and > they're useful if you just want a small embedded controller > and are used to using VxWorks. A lot of the other boards with > VxWorks BSP's available for them tend to be rather high end and > rather too expensive to use for simple controller type aplications. > Mike. > -- > ------------------------------------------------------------------------------ > +*I'm not paid to make official comments for Leica, so the above isn't one.*+ > Mike Wilmot-Dear (MW342) e-mail: mikewd@leica.co.uk > Leica Cambridge Ltd. Tel: +44 1223-411411 (x347) > Clifton Road Fax: +44 1223-210692 > Cambridge, CB1 3QH, UK We are using PEP boards with VLAN for ethernet succesfully here at Contraves. If you have specific questions, I can direct you to the people to contact here. Ivan Dzombak Contraves, Inc. Pittsburgh, PA ivan.dzombak@industry.net --------------------------- Newsgroups: comp.os.vxworks Subject: interrupt question Date: 28 Apr 1995 23:18:35 GMT From: kccheng@beauty2.lbl.gov (Kuang-chun Cheng) Organization: Lawrence Berkeley Laboratory, Berkeley CA Message-ID: <3nrt4b$m2e@overload.lbl.gov> Hi, In "Wind Technical Notes" No.27 Mutltiprocessing and the Backplane, p.27-27, There are two programs (intGen.c and intSync.c) to demonstrate synchronizing among tasks running on 2 CPU. intGen() will generate a VME bus interrupt, and intSync() will waiting for a semaphore which will be given by interrupt handler. The basic structure of these two program is intGen() #define LEVEL 1 #define VECTOR 255 sysBusIntGen(LEVEL, VECTOR); intSync() intInit() --------------------------> sysIntEnable(LEVEL) semTake(semID, WAIT_FOREVER) intConnect(INUM_TO_IVEC(VECTOR), semGive, semID) I think this example is simple enough and I did run it without any problem. However, if I run these two programs at the same CPU by the following command, -> sp inSync -> sp inGen I got the following messages which I don't understand: interrupt: Uninitialized Interrupt! Vector number 24 (0-255). Spurious Interrupt Program Counter: 0x0003ce96 Status Register: 0x3108 interrupt generated I did use any interrupt with vector=24, why it appear here ??? Why they work very well in 2 CPU case, but can't work together at the same CPU ? Thanks for your help. Sincerely, Kuang-chun Cheng kccheng@phys.sinica.edu.tw kccheng@beauty1.lbl.gov --------------------------- End of New-News digest ********************** From daemon@vxw.ee.lbl.gov Sun Apr 30 04:00:28 1995 From: daemon@vxw.ee.lbl.gov Date: Sun Apr 30 04:00:36 PDT 1995 Subject: comp.os.vxworks newsdigest Comp.Os.Vxworks Daily Digest Sun Apr 30 04:00:20 PDT 1995 Subject: Re: interrupt question ------------------------------------------------------- Newsgroups: comp.os.vxworks Subject: Re: interrupt question Date: Sat, 29 Apr 1995 07:39:57 From: sarnold@unicomp.net (Shane Arnold) Organization: CCS Integra Software Message-ID: References: <3nrt4b$m2e@overload.lbl.gov> In article <3nrt4b$m2e@overload.lbl.gov> kccheng@beauty2.lbl.gov (Kuang-chun Cheng) writes: >I got the following messages which I don't understand: > interrupt: > Uninitialized Interrupt! > Vector number 24 (0-255). Spurious Interrupt > Program Counter: 0x0003ce96 > Status Register: 0x3108 > interrupt generated >I did use any interrupt with vector=24, why it appear here ??? ^didn't I think what has happened here is your interrupter is waiting for a VME interrupt acknowlege so it can respond with a vector. Since your only using one CPU it would have to simultaneously act as an interrupter and an interrupt handler. If the CPU you are using does not support this, it will wait a while for the interrupt acknowledge and timeout generating a spurious interrupt exception (24). I've never attempted to use a VME interrupter to interrupt itself so I don't know if it's possible. I would suggest using semaphores directly. Good Luck! Shane Arnold E-Systems, Inc. - Greenville Division sla1@esygvl.com P.O. Box 6056 (903) 457-6846 Greenville, Texas 75401 --------------------------- End of New-News digest ********************** From thor@thor.atd.ucar.edu Sun Apr 30 23:00:15 1995 From: thor@thor.atd.ucar.edu (Rich Neitzel) Date: Sun Apr 30 23:00:23 PDT 1995 Subject: Monthly VxWorks archive posting This is the monthly posting showing the current holdings in the VxWorks Software Archive. To get more detailed infomation send email to: vxworks_archive@ncar.ucar.edu The message body must read: send index send index from vx ------------------------------------------------ VxWorks sources: total 14594 lrwxrwxrwx 1 root rsf 5 Feb 10 19:32 README -> index -rw-r--r-- 1 thor rsf 22132 Sep 20 1994 ansi.p1 -rw-r--r-- 1 thor rsf 22717 Sep 20 1994 ansi.p2 -rw-r--r-- 1 thor rsf 24174 Sep 20 1994 ansi.p3 -rw-r--r-- 1 thor rsf 8108 Sep 20 1994 ansi.patch1 -rw-r--r-- 1 thor rsf 37126 Sep 20 1994 ansilib01 -rw-r--r-- 1 thor rsf 18913 Sep 20 1994 ansilib02 -rw-r--r-- 1 thor rsf 2671 Sep 20 1994 benchmarks -rw-r--r-- 1 thor rsf 7168 Sep 20 1994 bitcnt -rw-r--r-- 1 thor rsf 11437 Sep 20 1994 c++builtin.shar -rw-r--r-- 1 thor rsf 22330 Sep 20 1994 c++headers.p1 -rw-r--r-- 1 thor rsf 22775 Sep 20 1994 c++headers.p2 -rw-r--r-- 1 thor rsf 29052 Sep 20 1994 camaclib1 -rw-r--r-- 1 thor rsf 25095 Sep 20 1994 camaclib2 -rw-r--r-- 1 thor rsf 31005 Sep 20 1994 camaclib3 -rw-r--r-- 1 thor rsf 37770 Sep 20 1994 cbench.shar -rw-r--r-- 1 thor rsf 7371 Sep 20 1994 cntsem_class.shar -rw-r--r-- 1 thor rsf 5853 Sep 20 1994 crc.shar -rw-r--r-- 1 thor rsf 8917 Sep 20 1994 deadman.shar -rw-r--r-- 1 thor rsf 41669 Sep 20 1994 dhrystones01 -rw-r--r-- 1 thor rsf 19170 Sep 20 1994 dirlib01 -rw-r--r-- 1 thor rsf 25681 Sep 20 1994 dt1451 -rw-r--r-- 1 thor rsf 4905 Sep 20 1994 dup.shar -rw-r--r-- 1 thor rsf 5944 Sep 20 1994 fcompress.shar -rw-r--r-- 1 thor rsf 11561 Sep 20 1994 flags_class.shar -rw-r--r-- 1 thor rsf 88109 Sep 20 1994 flash.zip -rw-r--r-- 1 thor rsf 44762 Sep 20 1994 force.p1 -rw-r--r-- 1 thor rsf 40154 Sep 20 1994 force.p2 -rw-r--r-- 1 thor rsf 80491 Sep 20 1994 force.shar -rw-r--r-- 1 thor rsf 2453 Sep 20 1994 gcc+68040 -rw-r--r-- 1 thor rsf 6106 Sep 20 1994 getdate -rw-r--r-- 1 thor rsf 9774 Sep 20 1994 hkv30extintutil.shar -rw-r--r-- 1 thor rsf 24394 Jan 31 15:04 index -rw-r--r-- 1 thor rsf 2694 Sep 20 1994 ivecalloc.shar -rw-r--r-- 1 thor rsf 35245 Sep 20 1994 joblib2.p1 -rw-r--r-- 1 thor rsf 18110 Sep 20 1994 joblib2.p2 -rw-r--r-- 1 thor rsf 9079 Sep 20 1994 lclflag.shar -rw-r--r-- 1 thor rsf 174829 Sep 20 1994 lexyacc.tar.gz drwxr-xr-x 2 thor rsf 512 Sep 20 1994 libX11 -rw-r--r-- 1 thor rsf 14059 Nov 4 16:03 libg++-2.6.1.patch -rw-r--r-- 1 thor rsf 1302 Nov 23 10:34 libgcc2-2.6.2.patch -rw-r--r-- 1 thor rsf 1862 Dec 29 10:50 libgcc2-2.6.3.patch -rw-r--r-- 1 thor rsf 1222 Nov 4 16:03 libio-2.6.1.patch -rw-r--r-- 1 thor rsf 277 Nov 4 16:03 libobjc.patch drwxr-xr-x 2 thor rsf 512 Sep 20 1994 libx11 -rw-rw-r-- 1 ftp rsf 3515 Nov 2 15:29 loadmeter.shar -rw-r--r-- 1 thor rsf 10399 Sep 20 1994 math.shar -rw-r--r-- 1 thor rsf 11950 Sep 20 1994 math2 -rw-r--r-- 1 thor rsf 26655 Sep 20 1994 monitor.shar -rw-r--r-- 1 thor rsf 18733 Sep 20 1994 msgque_class.shar -rw-r--r-- 1 thor rsf 829713 Sep 20 1994 ntpv3.1.tar.gz -rw-r--r-- 1 thor rsf 1010176 Sep 20 1994 ntpv3.tar.gz -rw-r--r-- 1 thor rsf 1082 Sep 20 1994 objc.patch -rw-r--r-- 1 thor rsf 19422 Sep 20 1994 ping01 -rw-r--r-- 1 thor rsf 20494 Sep 20 1994 pipe.shar -rw-r--r-- 1 thor rsf 15418 Sep 20 1994 poolLib -rw-r--r-- 1 thor rsf 15418 Sep 20 1994 poollib -rw-r--r-- 1 thor rsf 13204 Sep 20 1994 ring.shar -rw-r--r-- 1 thor rsf 6614 Sep 20 1994 semCnt -rw-r--r-- 1 thor rsf 6614 Sep 20 1994 semcnt -rw-r--r-- 1 thor rsf 48659 Dec 2 10:25 snmp2.01 -rw-r--r-- 1 thor rsf 48658 Dec 2 10:25 snmp2.02 -rw-r--r-- 1 thor rsf 48666 Dec 2 10:25 snmp2.03 -rw-r--r-- 1 thor rsf 48723 Dec 2 10:25 snmp2.04 -rw-r--r-- 1 thor rsf 48869 Dec 2 10:25 snmp2.05 -rw-r--r-- 1 thor rsf 48632 Dec 2 10:25 snmp2.06 -rw-r--r-- 1 thor rsf 48658 Dec 2 10:25 snmp2.07 -rw-r--r-- 1 thor rsf 48652 Dec 2 10:25 snmp2.08 -rw-r--r-- 1 thor rsf 48714 Dec 2 10:25 snmp2.09 -rw-r--r-- 1 thor rsf 48615 Dec 2 10:25 snmp2.10 -rw-r--r-- 1 thor rsf 49185 Dec 2 10:25 snmp2.11 -rw-r--r-- 1 thor rsf 49214 Dec 2 10:25 snmp2.12 -rw-r--r-- 1 thor rsf 49203 Dec 2 10:25 snmp2.13 -rw-r--r-- 1 thor rsf 49239 Dec 2 10:25 snmp2.14 -rw-r--r-- 1 thor rsf 49176 Dec 2 10:25 snmp2.15 -rw-r--r-- 1 thor rsf 49216 Dec 2 10:25 snmp2.16 -rw-r--r-- 1 thor rsf 49182 Dec 2 10:25 snmp2.17 -rw-r--r-- 1 thor rsf 49188 Dec 2 10:25 snmp2.18 -rw-r--r-- 1 thor rsf 49189 Dec 2 10:25 snmp2.19 -rw-r--r-- 1 thor rsf 49200 Dec 2 10:25 snmp2.20 -rw-r--r-- 1 thor rsf 48646 Dec 2 10:25 snmp2.21 -rw-r--r-- 1 thor rsf 48671 Dec 2 10:25 snmp2.22 -rw-r--r-- 1 thor rsf 49229 Dec 2 10:25 snmp2.23 -rw-r--r-- 1 thor rsf 49235 Dec 2 10:25 snmp2.24 -rw-r--r-- 1 thor rsf 49268 Dec 2 10:25 snmp2.25 -rw-r--r-- 1 thor rsf 48682 Dec 2 10:25 snmp2.26 -rw-r--r-- 1 thor rsf 49223 Dec 2 10:25 snmp2.27 -rw-r--r-- 1 thor rsf 48676 Dec 2 10:25 snmp2.28 -rw-r--r-- 1 thor rsf 48818 Dec 2 10:25 snmp2.29 -rw-r--r-- 1 thor rsf 49308 Dec 2 10:25 snmp2.30 -rw-r--r-- 1 thor rsf 49294 Dec 2 10:25 snmp2.31 -rw-r--r-- 1 thor rsf 48634 Dec 2 10:26 snmp2.32 -rw-r--r-- 1 thor rsf 48708 Dec 2 10:26 snmp2.33 -rw-r--r-- 1 thor rsf 49055 Dec 2 10:26 snmp2.34 -rw-r--r-- 1 thor rsf 48724 Dec 2 10:26 snmp2.35 -rw-r--r-- 1 thor rsf 48736 Dec 2 10:26 snmp2.36 -rw-r--r-- 1 thor rsf 48667 Dec 2 10:26 snmp2.37 -rw-r--r-- 1 thor rsf 48626 Dec 2 10:26 snmp2.38 -rw-r--r-- 1 thor rsf 48653 Dec 2 10:26 snmp2.39 -rw-r--r-- 1 thor rsf 14122 Dec 2 10:26 snmp2.40 -rw-r--r-- 1 thor rsf 2308 Sep 20 1994 ss1.bnch -rw-r--r-- 1 thor rsf 41196 Sep 20 1994 stevie01 -rw-r--r-- 1 thor rsf 35279 Sep 20 1994 stevie02 -rw-r--r-- 1 thor rsf 35278 Sep 20 1994 stevie03 -rw-r--r-- 1 thor rsf 35012 Sep 20 1994 stevie04 -rw-r--r-- 1 thor rsf 34502 Sep 20 1994 stevie05 -rw-r--r-- 1 thor rsf 37476 Sep 20 1994 stevie06 -rw-r--r-- 1 thor rsf 30073 Sep 20 1994 stevie07 -rw-r--r-- 1 thor rsf 31562 Sep 20 1994 stevie08 -rw-r--r-- 1 thor rsf 37360 Sep 20 1994 stevie09 -rw-r--r-- 1 thor rsf 20662 Sep 20 1994 stevie10 -rw-r--r-- 1 thor rsf 25717 Sep 20 1994 stevie11 -rw-r--r-- 1 thor rsf 28075 Sep 20 1994 stevie12 -rw-r--r-- 1 thor rsf 31852 Sep 20 1994 stevie13 -rw-r--r-- 1 thor rsf 16225 Sep 20 1994 string.shar -rw-r--r-- 1 thor rsf 8424 Sep 20 1994 syslog.shar -rw-r--r-- 1 thor rsf 15096 Sep 20 1994 task_class.shar -rw-r--r-- 1 thor rsf 16171 Sep 20 1994 taskmon.shar -rw-rw-r-- 1 thor rsf 417608 Jan 31 14:43 tclvx7.0v5.tar.gz -rw-rw-r-- 1 vanandel rsf 473525 Apr 25 08:42 tclvx7.3.tar.gz -rw-r--r-- 1 thor rsf 10523 Sep 20 1994 tod.shar -rw-r--r-- 1 thor rsf 19912 Sep 20 1994 tp41.shar -rw-r--r-- 1 thor rsf 25790 Sep 20 1994 ty335.shar -rw-r--r-- 1 thor rsf 44504 Sep 20 1994 usrTime.shar -rw-r--r-- 1 thor rsf 44504 Sep 20 1994 usrtime.shar -rw-r--r-- 1 thor rsf 4945 Sep 20 1994 veclist01 -rw-r--r-- 1 thor rsf 25814 Sep 20 1994 vtape.shar -rw-r--r-- 1 thor rsf 43671 Sep 20 1994 vwcurses01 -rw-r--r-- 1 thor rsf 40180 Sep 20 1994 vwcurses02 -rw-r--r-- 1 thor rsf 38308 Sep 20 1994 vwcurses03 -rw-r--r-- 1 thor rsf 31181 Sep 20 1994 vwcurses04 -rw-r--r-- 1 thor rsf 31798 Sep 20 1994 vwcurses05 -rw-r--r-- 1 thor rsf 31459 Sep 20 1994 vwcurses06 -rw-r--r-- 1 thor rsf 24279 Sep 20 1994 vwcurses07 -rw-r--r-- 1 thor rsf 4973 Nov 4 16:01 vx_cplusplus -rw-r--r-- 1 thor rsf 40485 Sep 20 1994 vxrouted01 -rw-r--r-- 1 thor rsf 40512 Sep 20 1994 vxrouted02 -rw-r--r-- 1 thor rsf 40482 Sep 20 1994 vxrouted03 -rw-r--r-- 1 thor rsf 40472 Sep 20 1994 vxrouted04 -rw-r--r-- 1 thor rsf 40526 Sep 20 1994 vxrouted05 -rw-r--r-- 1 thor rsf 40540 Sep 20 1994 vxrouted06 -rw-r--r-- 1 thor rsf 40552 Sep 20 1994 vxrouted07 -rw-r--r-- 1 thor rsf 22675 Sep 20 1994 vxrouted08 -rw-r--r-- 1 thor rsf 29720 Sep 20 1994 vxrsh.p1 -rw-r--r-- 1 thor rsf 26002 Sep 20 1994 vxrsh.p2 -rw-r--r-- 1 thor rsf 13713 Sep 20 1994 vxrsh.p3 -rw-r--r-- 1 thor rsf 4702 Sep 20 1994 wdog_class -rw-r--r-- 1 thor rsf 40860 Sep 20 1994 xmodem.01 -rw-r--r-- 1 thor rsf 40463 Sep 20 1994 xmodem.02 -rw-r--r-- 1 thor rsf 40403 Sep 20 1994 xmodem.03 -rw-r--r-- 1 thor rsf 40451 Sep 20 1994 xmodem.04 -rw-r--r-- 1 thor rsf 40721 Sep 20 1994 xmodem.05 -rw-r--r-- 1 thor rsf 9333 Sep 20 1994 xmodem.06