Organization: Mascot Systems Limited
Message-ID: <3A2A9E9D.6E815673@us.mascotsystems.com>
Reply-To: gkota@eu.mascotsystems.com
- --------------7A0D567B6A45D8095A68C692
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Mascot systems is a part of Igate capital (NASDAQ : IGTE)(formerly
called Mastech Corp) is a a worldwide supplier of IT consulting
services (http://www.mascotsystems.com). Services include the delivery
of specialized IT skills in teams as required for specific projects by
Mascot's clients. Projects are typically completed onsite However,
Mascot also has extensive offshore software factories, which may be
utilized for the delivery of certain aspects ofservices. Igate has over
5000 employees, and 1998 revenues over US$ 400 million. If you are
looking to make a career move, and interested in a fast growing
multi-national corporation with unlimited opportunities, then look no
further.
We are looking for Dynamic information technology professionals with
good communication skills in written and spoken English for Mascot
Netherlands operations . We have an interesting relocation plan for
people immigrating from other coutries
The Open vacancies are as follows
* Embedded systems and Real-time Operating systems(RTOS)
programmers with experience in the following areas C/C++ ,
Assembler , Motorola ,Siemens, VXWorks, CAN, program Control
Switching, H323 protocol Filters , UMTS layer 2 (Medium Access
Control) and layer 3 (Network and Architectures) techniques
and protocols, Hatley and Phirbai , , Motorola MPC860, Intel
SA110, Tundra Qspan2, Intel 21554,Trimedia, RF, OSE, ASIC design
and so on
Interested candidates can send their CV to gkota@eu.mascotsystems.com
- --------------7A0D567B6A45D8095A68C692
Content-Type: text/html; charset=us-ascii
Content-Transfer-Encoding: 7bit
Mascot systems is a part of Igate capital (NASDAQ : IGTE)(formerly
called Mastech Corp) is a a worldwide supplier of IT consulting
services (http://www.mascotsystems.com).
Services include the delivery of specialized IT skills in teams as required
for specific projects by Mascot's clients. Projects are typically
completed onsite However, Mascot also has extensive offshore software
factories, which may be utilized for the delivery of certain aspects ofservices.
Igate has over 5000 employees, and 1998 revenues over US$ 400 million.
If you are looking to make a career move, and interested in a fast growing
multi-national corporation with unlimited opportunities, then look no further.
We are looking for Dynamic information technology professionals with
good communication skills in written and spoken English for Mascot Netherlands
operations . We have an interesting relocation plan for people immigrating
from other coutries
The Open vacancies are as follows
-
Embedded systems and Real-time Operating
systems(RTOS) programmers with experience in the
following areas C/C++ , Assembler , Motorola ,Siemens, VXWorks, CAN,
program Control Switching, H323 protocol Filters , UMTS
layer 2 (Medium Access Control) and layer 3 (Network
and Architectures) techniques and protocols, Hatley and
Phirbai , , Motorola MPC860, Intel SA110, Tundra Qspan2, Intel 21554,Trimedia,
RF, OSE, ASIC design and so on
Interested candidates can send their CV to gkota@eu.mascotsystems.com
- --------------7A0D567B6A45D8095A68C692--
---------------------------
Newsgroups: comp.os.vxworks
Subject: Re: Enabling interrupt using intSRSet on MIPS??
Date: Sun, 03 Dec 2000 23:12:45 GMT
From: "nazneen shaikh"
Organization: Global Online Japan // Exodus Communications KK.
Message-ID:
References: <3A29A982.35AEE24@wellknit.com>
call intSRSet() for MIPS.
ajay wrote in message
news:3A29A982.35AEE24@wellknit.com...
> Hi All,
> Is there a special way to set SR register. When I set the SR register
> it gets updated only in the task's context. But is does not enable the
> interrupts that I want to enable. I am working on QED's MIPS R7000 and
> using vxworks 5.4 with bsp from windriver. The vxworks programmer's
> guide indicates that SR must be set before spawning task's. In my case I
> need to do this after all the task's have been spawned.
>
> Any help will be appreciated.
>
> Thanks,
> vanshi
>
---------------------------
Newsgroups: comp.os.vxworks
Subject: [Question] Difference VxWorks Boot Sequence??
Date: Mon, 04 Dec 2000 00:00:07 GMT
From: kingseft@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <90emq6$pp0$1@nnrp1.deja.com>
Hi. Vxworks Developers...
I'm a novice with VxWorks.
I have a basic question about VxWorks Boot Sequence ...
when I read some materials about boot sequence.
I found the diferrences between building from the traditional command
line method and Vxworks Project Facility methdo..
Could someone describe some differences about that??
Why some developers use command line and whey some developers use
project facility method??
They said that there is some differences ...
But I can't find that..
Please answer this basic questions..... Really thanks......
kingseft
Sent via Deja.com http://www.deja.com/
Before you buy.
---------------------------
Newsgroups: comp.os.vxworks
Subject: integrating network drivers
Date: 4 Dec 2000 00:13:18 GMT
From: skenda@start.com.au
Organization: NewsOne.Net - Free Usenet News via the Web - http://newsone.net/
Message-ID: <90eniu$cms$1@news.netmar.com>
Followup-To: comp.os.vxworks
I have been using vxWorks\Tornado 2 for a short time.
My target processor board is a pentium
I am trying to integrate the crystal CS8900 network driver
Up to now the majority of my configuration changes have been done so through
the project facility. On this occaision however it seems that this device is
only supported through the device \target\src\drv\netif\if_cs
ie it has not been integrated into the project facility\config.h
I'm still using an external floppy to boot during this intial development
phase so that changes can be incorporated rapidly.
My question is that since there is no mention of this driver in either
a) config.h
b) project facility
how do I make the required changes to:
a) make a bootrom supporting this driver
b) make a bootable kernal
any tips are most welcome
Sineon
----- Posted via NewsOne.Net: Free (anonymous) Usenet News via the Web -----
http://newsone.net/ -- Free reading and anonymous posting to 60,000+ groups
NewsOne.Net prohibits users from posting spam. If this or other posts
made through NewsOne.Net violate posting guidelines, email abuse@newsone.net
---------------------------
Newsgroups: comp.os.vxworks
Subject: debugging in the Windows simulator
Date: Mon, 04 Dec 2000 00:38:23 GMT
From: wv12@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <90ep1v$rdm$1@nnrp1.deja.com>
Hi
Sometimes vxworks.exe crashes before windsh gets a chance to start.
Windsh then gets hopelessly confused. I am wondering if it's possible
to set breakpoint(s) even before starting vxworks.exe. Also, what is
the syntax to start the target server at a DOS prompt.
Thank you
Will
Sent via Deja.com http://www.deja.com/
Before you buy.
---------------------------
Newsgroups: comp.os.vxworks
Subject: Re: Enabling interrupt using intSRSet on MIPS??
Date: Mon, 04 Dec 2000 00:55:12 GMT
From: ajay
Organization: EarthLink Inc. -- http://www.EarthLink.net
Message-ID: <3A2AEB5D.659D7BF3@wellknit.com>
References: <3A29A982.35AEE24@wellknit.com>
Reply-To: ajay@wellknit.com
Sorry, for not mentioning that I was using intSRSet for modifying the SR
register. This API seems to modify SR in context of the task (confirmed using
taskRegsShow) when I try to execute from shell. The interrupts are not
disabled using this API.
The assembly code seems to be fine. Does it mean that intSRSet cannot be
called from shell? or may be I am confused here...
Assembly code:
mfc0 v0,sr
nop
jr ra
mtc0 a0,sr
nop
- -vanshi
nazneen shaikh wrote:
> call intSRSet() for MIPS.
>
> ajay wrote in message
> news:3A29A982.35AEE24@wellknit.com...
> > Hi All,
> > Is there a special way to set SR register. When I set the SR register
> > it gets updated only in the task's context. But is does not enable the
> > interrupts that I want to enable. I am working on QED's MIPS R7000 and
> > using vxworks 5.4 with bsp from windriver. The vxworks programmer's
> > guide indicates that SR must be set before spawning task's. In my case I
> > need to do this after all the task's have been spawned.
> >
> > Any help will be appreciated.
> >
> > Thanks,
> > vanshi
> >
---------------------------
Newsgroups: comp.os.vxworks
Subject: re: comp.os.vxworks newsdigest
Date: Mon, 04 Dec 2000 09:20:44 +0900
From: "Girish V. Gulawani"
Organization: Lawrence Berkeley National Laboratory
Message-ID: <90er54$a2l$1@overload.lbl.gov>
> Newsgroups: comp.os.vxworks
> Subject: Enabling interrupt using intSRSet on MIPS??
> Date: Sun, 03 Dec 2000 02:01:57 GMT
> From: ajay
> Organization: EarthLink Inc. -- http://www.EarthLink.net
> Message-ID: <3A29A982.35AEE24@wellknit.com>
> Reply-To: ajay@wellknit.com
>
> Hi All,
> Is there a special way to set SR register. When I set the SR register
> it gets updated only in the task's context. But is does not enable the
> interrupts that I want to enable. I am working on QED's MIPS R7000 and
> using vxworks 5.4 with bsp from windriver. The vxworks programmer's
> guide indicates that SR must be set before spawning task's. In my case I
> need to do this after all the task's have been spawned.
pls see there are two separate calls intSRSet & taskSRInit.
hth.
---------------------------
Newsgroups: comp.os.vxworks
Subject: Re: [Question] Difference VxWorks Boot Sequence??
Date: Sun, 03 Dec 2000 17:25:53 -0800
From: DrDiags
Organization: Posted via Supernews, http://www.supernews.com
Message-ID: <3A2AF2A1.71AE88A3@flashcom.net>
References: <90emq6$pp0$1@nnrp1.deja.com>
Reply-To: drdiags@flashcom.net
Kingseft,
I will give you my best.
Command line build allows you to make changes to the BSP and build the
various images you may want. You could take a project directory and modify
it to also build from the command line. When building from the IDE, you
can base it off of a particular BSP, but any changes made to the BSP
source after you have created the project will not be picked up unless you
go to the command line in the project and do a make clean. The "Rebuild
All" button does not remove the prjConfig.c file and others which were
inherited from your BSP and modified by changes you make in the IDE. Also,
the directory structure of the files used by the two techniques differ.
Building and modifying the BSP uses the target/config/all files and your
target/config/BSP header, source and dependency files. For the IDE, the
target/config/BSP directory plus the target/config/comps/src. Of course
the libraries are the same no matter which of the two you choose. You
might try looking in your BSP's Makefile, config.h and depend.BSP file.
Also review the target.nr file if you haven't already.
From my experience, you have a little more flexiblity building from
the command line, but the IDE is really helpful if you don't want to
figure all the interdependencies of some of the Kernel features. I find
it a good exercise to build a kernel and bootrom from the command line.
But when I am not familiar with some new feature such as dosFS 2.0 or
VxFusion, then the IDE can help me understand what the parameters are and
how I might set them. I hope this helps.
kingseft@my-deja.com wrote:
> Hi. Vxworks Developers...
>
> I'm a novice with VxWorks.
> I have a basic question about VxWorks Boot Sequence ...
>
> when I read some materials about boot sequence.
> I found the diferrences between building from the traditional command
> line method and Vxworks Project Facility methdo..
>
> Could someone describe some differences about that??
>
> Why some developers use command line and whey some developers use
> project facility method??
> They said that there is some differences ...
> But I can't find that..
>
> Please answer this basic questions..... Really thanks......
>
> kingseft
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.
---------------------------
Newsgroups: comp.os.vxworks
Subject: Re: integrating network drivers
Date: Sun, 03 Dec 2000 17:44:43 -0800
From: DrDiags
Organization: Posted via Supernews, http://www.supernews.com
Message-ID: <3A2AF70B.A74CA02@flashcom.net>
References: <90eniu$cms$1@news.netmar.com>
Reply-To: drdiags@flashcom.net
Sineon,
I don't know if you tried using the MACH_EXTRA line in your makefile. You could
put the source locally in your BSP directory and create your own #define
INCLUDE_MY_NETIF. You could then copy from the target/config/all the bootConfig.c
file or the usrConfig.c file and modify the Makefile so that it will use your local
files. Start reviewing the bootConfig (which is used for bootrom system startup) or
usrConfig.c (which is the path the kernel starts up after being expanded into local
RAM or begins running out of ROM if it is resident). If you can get your hands on a
BSP developer manual, there are some really good flow diagrams that show where you
need to do what (as suggested by WRS). Back to the MACH_EXTRA line, it could look
like so in the BSP makefile:
MACH_EXTRA = myNetIf.o
skenda@start.com.au wrote:
> I have been using vxWorks\Tornado 2 for a short time.
>
> My target processor board is a pentium
>
> I am trying to integrate the crystal CS8900 network driver
>
> Up to now the majority of my configuration changes have been done so through
> the project facility. On this occaision however it seems that this device is
> only supported through the device \target\src\drv\netif\if_cs
> ie it has not been integrated into the project facility\config.h
>
> I'm still using an external floppy to boot during this intial development
> phase so that changes can be incorporated rapidly.
>
> My question is that since there is no mention of this driver in either
> a) config.h
> b) project facility
>
> how do I make the required changes to:
> a) make a bootrom supporting this driver
> b) make a bootable kernal
>
> any tips are most welcome
> Sineon
>
> ----- Posted via NewsOne.Net: Free (anonymous) Usenet News via the Web -----
> http://newsone.net/ -- Free reading and anonymous posting to 60,000+ groups
> NewsOne.Net prohibits users from posting spam. If this or other posts
> made through NewsOne.Net violate posting guidelines, email abuse@newsone.net
---------------------------
Newsgroups: comp.os.vxworks
Subject: Re: Can the target write to the host's file system?
Date: Mon, 04 Dec 2000 01:40:05 GMT
From: Mike McNaughton
Organization: Sympatico
Message-ID: <3A2B1222.94D6B002@sympatico.ca>
References: <9093m2$c3t$1@murdoch.harvard.net>
Andy
I've had some succes using using the NFS file sytem on the targets adn the Omni
nfs server on the PC. I'm using tornado 1.02. I had to futz around with it a
bit but it works.
Mike
Andrew Carlisle wrote:
> Hello,
>
> Is it possible to have the vxWorks target write to the host's file system?
> For instance, I have a target running vxWorks that has no hard drive. I
> would like to be able to open and close files on the NT host's file system.
> For instance, I would like to write to c:/sysdump/dump.txt from the target.
> Is this possible?
>
> Andy
>
> acarlisle@envoynetworks.com
---------------------------
Newsgroups: comp.os.vxworks
Subject: Re: debugging in the Windows simulator
Date: Sun, 03 Dec 2000 17:50:20 -0800
From: DrDiags
Organization: Posted via Supernews, http://www.supernews.com
Message-ID: <3A2AF85C.7458D7E0@flashcom.net>
References: <90ep1v$rdm$1@nnrp1.deja.com>
Reply-To: drdiags@flashcom.net
Will,
Here is a dump of the tgtsvr.exe command with a -help switch.
Probably more than you really wanted to know. As far as the debugging of
vxworks.exe, I leave that to my more experienced colleagues.
==========================================================
This command creates a NT/Win95 host based target server. Use the
command
tgtsvr [options] targetName -c corefileName
Options available are:
- -Bt[imeout] value WDB RPC Backend request timeout (default 1sec.).
- -Br[esend] number WDB RPC Backend request re-send number(default 3).
- -p[ort] number UDP port number to connect (default 0x4321).
- -A[llsyms] Include global and local core symbols.
- -B[ackend] bkendName Type of backend to use. (default wdbrpc).
- -Bd[ebug] fileName Log backend debug trace to named file.
- -Bm[ax] wdbLogMaxSize Size max for the WDB log file.
- -b[ps] bit/s Host tty device speed. (default 9600).
- -C[onsole] Start a console window.
- -c[ore] fileName Filename of target core.
- -d[evice] device Host tty device to use. (default com2).
- -display hostName:0 Where to display console window.
- -f[ormat] formatName Object module format (a.out, coff).
- -h[elp] Print help.
- -hfc Include hardware flow control.
- -L[ock] Restrict access to targer server creator.
- -m[emory] nbytes Size of target server memory cache.
- -n[ame] server-name Name of target server.
- -N[osyms] Include no core symbols.
- -redirectIO Redirect the target global IO.
- -redirectShell Start a target console with target shell IO redirected.
- -R[oot] directory Enable Target Server file system with given root.
- -RW Make Target Server file system read/write.
- -s[ynchro] Synchronize target/host symbol tables.
- -use_portmapper Use the local portmapper to register rpc services.
- -u[sers] fileName Filename of list of authorized server users.
- -V[erbose] Turn on verbose mode.
- -v[ersion] Display Tornado version.
- -Wd[ebug] fileName Log wtx debug trace to named file.
- -Wf[ilter] request Remove request from log (default: WTX_EVENT_GET)
- -Wm[ax] wtxLogMaxSize Size max for the WTX log file.
Target Server will exit
===========================================================
wv12@my-deja.com wrote:
> Hi
> Sometimes vxworks.exe crashes before windsh gets a chance to start.
> Windsh then gets hopelessly confused. I am wondering if it's possible
> to set breakpoint(s) even before starting vxworks.exe. Also, what is
> the syntax to start the target server at a DOS prompt.
> Thank you
> Will
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.
---------------------------
Newsgroups: comp.os.vxworks
Subject: re: comp.os.vxworks newsdigest
Date: Mon, 04 Dec 2000 10:26:48 +0900
From: "Girish V. Gulawani"
Organization: Lawrence Berkeley National Laboratory
Message-ID: <90eulk$alk$1@overload.lbl.gov>
reply is below -
> Newsgroups: comp.os.vxworks
> Subject: DosFS 2.0 and File Size jumps >4GB
> Date: Sat, 02 Dec 2000 12:08:38 -0800
> From: DrDiags
> Organization: Posted via Supernews, http://www.supernews.com
> Message-ID: <3A2956C6.55E9C0AE@flashcom.net>
> Reply-To: drdiags@flashcom.net
>
> Hello,
>
> I am trying to track down a problem I am experiencing with DosFS
> 2.0. Using a SanDisk 440MB ATA Flash Disk on a MPC750 based on a
> YellowKnife 750 BSP. The disk has several 1-4MB zip files. Using zLib
> 1.1.3 ported to VxWorks, I have a routine which will unzip the file to
> local RAM, read a location and close the file. During this time, I can
> also have up to 18 processors across a message queue on a shared memory
> backplane reading this file into their local RAM. During this time, I
> have found by doing fstat() on the file, that it will grow to >4GB. Here
> are some observations made:
>
> 1 - The corruption occurs during a lseek on the file.
> 2 - The 18 readers do not seem to cause this.
> 3 - Putting MUTEX around the routines used by DosFS 2.0 have no affect.
> 4 - Found the taskSpawn used by the DosFS 2.0 service routines do not
> have PRIORITY_INVERSION_SAFE set, doing this slightly changes the
> behavior (lseek now will fail as it detects attempts to seek past the
> EOF when a write is done to the file, but does help the scenario of read
> only corruption. This point may be side noise, due to throwing the
> kitchen sink at this problem and trying anything to isolate it. It just
> smells of timing dependency)
> 5 - Found that during debugging, if the unzip task is held at certain
> breakpoints, the corruption is clear by the other reader tasks doing
> lseek().
>
> I guess my main question is has anyone seen something similar as far as
> the file size corruption, I have found that there are at least 2 cases
> which have not been resolved. Any ideas? Sorry about the long winded and
> maybe not properly documented question.
.1> if seek after end of file, new clusters get allocated.
.2> i guess any read after such seek fails.
.3> however if write is performed, file size increases with addition of
newly allocated clusters. note step #1, those new cluster(s) come in
middle.
.4> note however that in this case there is no activity on hw but the
file pointer.
.5> new clusters allocated in step#1, shd be zeroed out, by file system.
if it isnt happening, this exactly is the case you've come across.
am i getting the question?
please let me also know if its something else.
hth.
---------------------------
Newsgroups: comp.os.vxworks
Subject: [Question] What functions called in Vxworks ??
Date: Mon, 04 Dec 2000 02:21:26 GMT
From: kingseft@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <90ev33$vo8$1@nnrp1.deja.com>
Hi. Vxworks Developers.
Really thanks to my previous question..
I finally got some guideline from replying to my question.
Here. I have another question on ROM-based VxWorks Startup sequence.
I quoted from Tornado Training WorksShop Book. Chater 13.
Reconfingurring Vxworks.
In this chapters. they presented ROM-based VxWorks Start Up sequence.
1. romInit
2. romStart in romStart.c or bootInit.c
3. usrInit in prjConfig.c or bootConfig.c
4. usrRoot int prjConfig.c or bootConfig.c
5. boot program or user application
Here. I really curious above procedures...
in step 2. I knew romInit called romStart() function, but when I found
romStart(), I founnd 2 version of romStart(). i.e.
one is romStart.c and the other is bootInit.c
I really want to know what romInit really called
and whey romStart.c and bootInit.c have same romStart() routine..
and step 3, step 4 are the same question.
I really eager to know...
your comments will help us.. thanks for reading......
kingseft.
Sent via Deja.com http://www.deja.com/
Before you buy.
---------------------------
Newsgroups: comp.os.vxworks
Subject: Re: [Question] What functions called in Vxworks ??
Date: Sun, 03 Dec 2000 19:33:17 -0800
From: DrDiags
Organization: Posted via Supernews, http://www.supernews.com
Message-ID: <3A2B107C.629DE424@flashcom.net>
References: <90ev33$vo8$1@nnrp1.deja.com>
Reply-To: drdiags@flashcom.net
Kingseft,
What you are seeing here is the dual path of how a kernel startup
flows. The file prjConfig.c (for IDE build) replaces the usrConfig.c file
in the BSP are for the startup of the full-blown Kernel and the
bootConfig.c and bootInit.c are for the bootstrap startup. So what happens
on power-up can be followed through the boot*.c files. Now, depending on
whether you are ROM resident, compressed, uncompressed determines how you
kernel image is treated. The short story is that, lets say you are
downloading your modified kernel across ethernet with all the components in
the world enabled (too big to put in your ROM), your boot time kernel has
initialized your H/W and started up minimal services which will allow you
to download this kernel. After it has been successfully loaded and is ready
to run, the bootstrap services hand off control of the Processor to this
new Kernel, who doesn't trust anyone, so he is going to go through some
type of sanity check and initialization to make sure interrupts and memory
is in some stable condition. He begins to walk down a initialization path
similar to what was taken by the boot time kernel, that is why a somewhat
duplication of functions, but one only exists in the boot prom and the
other only exists in the downloaded/uncompressed/rom resident kernel. There
are some things not done in the full kernel image, which have to be done in
the boot image.
That is why if you need to use DHCP or serial or PPP for download, it
has to exist in the bootprom, since what is in the bootprom is all the boot
code can execute. I hope this helps. I try to keep things at a high level,
because that is as far as I can go.
kingseft@my-deja.com wrote:
> Hi. Vxworks Developers.
>
> Really thanks to my previous question..
> I finally got some guideline from replying to my question.
>
> Here. I have another question on ROM-based VxWorks Startup sequence.
>
> I quoted from Tornado Training WorksShop Book. Chater 13.
> Reconfingurring Vxworks.
>
> In this chapters. they presented ROM-based VxWorks Start Up sequence.
>
> 1. romInit
> 2. romStart in romStart.c or bootInit.c
> 3. usrInit in prjConfig.c or bootConfig.c
> 4. usrRoot int prjConfig.c or bootConfig.c
> 5. boot program or user application
>
> Here. I really curious above procedures...
> in step 2. I knew romInit called romStart() function, but when I found
> romStart(), I founnd 2 version of romStart(). i.e.
> one is romStart.c and the other is bootInit.c
>
> I really want to know what romInit really called
> and whey romStart.c and bootInit.c have same romStart() routine..
> and step 3, step 4 are the same question.
>
> I really eager to know...
> your comments will help us.. thanks for reading......
>
> kingseft.
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.
---------------------------
Newsgroups: comp.os.vxworks
Subject: ataDrv and a DOS disk
Date: Sun, 3 Dec 2000 20:26:55 -0800
From:
Organization: SBC Internet Services
Message-ID: <7%EW5.1067$R95.203918@news.pacbell.net>
Well, I am trying to get a DOS disk working on a StrongArm BSP originally
based on the ebsa285. I can configure my IDE interface chip (an Intel 82371)
and I can read/write the ATA drive registers, I can run ataInit successfully
if I disable the interrupt in ataDrv.c by polling for the busy bit on the
drive to be not busy and not wait forever for an interrupt to read the
status register. Using that trick, I can complete ataInit and then
ataDevCreate followed by dosFsMkfs. If I then do a 'cd' to the new device, I
can copy files to it and even list those files to the console. But if I turn
the target off and back on, I cannot see my files anymore. I tried copying a
rather large file (several megabytes) and I can hear the drive make noise,
and I can even list that file to the console. I am suspecting that there is
another command other then ataDevCreate or dosFsMkfs for a drive that has
allready been initialized, or I am fooling myself and the drive really is
nto initialized at all.
On a seperate but related question, it is not apparent to me how interrupts
need to be enabled through the 82371/87309/StrongArm triage for this
somewhat obscure IRQ14 (from legacy PC land), perhaps someone with more
experience might enlighten me. And as a third question, I find that I can do
a 'cd' to a device from a target shell but not from a host shell. Is there
some slightly additional incantation necessary to do a 'cd' to a drive on a
target from a host shell so I can use the Tornado debugger and not feel like
I have to fall back on the terminal from the target's shell
Thanks in advance.
---------------------------
Newsgroups: comp.os.vxworks
Subject: Re: Message queues in VxWorks
Date: Sun, 3 Dec 2000 21:10:58 -0600
From: "Wilson Lin"
Organization: SBC Internet Services
Message-ID:
References: <907fjm$qaj$1@overload.lbl.gov>
"R R, Rajendra Prabhala (Rajendra)** CTR **" wrote in
message news:907fjm$qaj$1@overload.lbl.gov...
> Hi vxworkers,
>
> Could anybody please clarify the following two points.
>
> 1. Is there any routine available in vxworks to flush all the messages at
> once in a message queue ?
> i.e. to discard all the messages waiting in a queue to be read (Posix
> or Wind queue)
I don't think so.
>
> 2. Is there any method to pick up nth message from the queue selectively ?
NO.
The best way to do that is to implement your own queue with sempaphores.
---------------------------
Newsgroups: comp.os.vxworks
Subject: Re: ataDrv and a DOS disk
Date: Sun, 03 Dec 2000 21:54:33 -0800
From: DrDiags
Organization: Posted via Supernews, http://www.supernews.com
Message-ID: <3A2B3199.8DA1A8E9@flashcom.net>
References: <7%EW5.1067$R95.203918@news.pacbell.net>
Reply-To: drdiags@flashcom.net
cfk@pacbell.net wrote:
> I am suspecting that there is
> another command other then ataDevCreate or dosFsMkfs for a drive that has
> allready been initialized, or I am fooling myself and the drive really is
> nto initialized at all.
>
dosFsDevInit( )
NAME
dosFsDevInit( ) - associate a block device with dosFs file system functions
SYNOPSIS
DOS_VOL_DESC *dosFsDevInit
(
char * devName, /* device name */
BLK_DEV * pBlkDev, /* pointer to block device struct */
DOS_VOL_CONFIG * pConfig /* pointer to volume config data */
)
DESCRIPTION
<.......................SNIP............................>
If the device being initialized already has a valid dosFs (MS-DOS) file
system on it, the pConfig parameter may be NULL. In this case, the volume will
be mounted and the configuration data will be read from the boot sector of the
disk. (If pConfig is NULL, both change-no-warn and auto-sync options are
initially disabled. These can be enabled using the dosFsVolOptionsSet( )
routine.)
>
> On a seperate but related question, it is not apparent to me how interrupts
> need to be enabled through the 82371/87309/StrongArm triage for this
> somewhat obscure IRQ14 (from legacy PC land), perhaps someone with more
> experience might enlighten me.
Don't know.
> And as a third question, I find that I can do
> a 'cd' to a device from a target shell but not from a host shell. Is there
> some slightly additional incantation necessary to do a 'cd' to a drive on a
> target from a host shell so I can use the Tornado debugger and not feel like
> I have to fall back on the terminal from the target's shell
Use the @ in front of your commands. This will route them to the target. Such
as:
- ->@cd "diag:D:"
- ->@ls
---------------------------
Newsgroups: comp.os.vxworks
Subject: unsupported source code..
Date: Mon, 04 Dec 2000 07:54:56 GMT
From: "Jyjung"
Organization: Dacom
Message-ID:
Hi,, VxWorkers..
I seek the RIP version1/2 free source code ported to VxWorks OS.
If you don't have these, please Tell me the site I can get the source code.
thank you
---------------------------
Newsgroups: comp.os.vxworks
Subject: upnp
Date: Mon, 04 Dec 2000 07:54:38 GMT
From: gandalf107@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <90fijt$dnb$1@nnrp1.deja.com>
hi,
is there any solution for including the upnp feature into vxworks?
thanks
Sent via Deja.com http://www.deja.com/
Before you buy.
---------------------------
Newsgroups: comp.os.vxworks
Subject: [Question] usrAppInit() function in Vxworks Project
Date: Mon, 04 Dec 2000 08:05:05 GMT
From: kingseft@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <90fj7h$e4o$1@nnrp1.deja.com>
Hi, VxWorks Developers.
I have one question about usrAppInit() in VxWorks Project file.
I added following lines to my BSP directory's config.h
#define INCLUDE_USER_APPL
#define USER_APPL_INIT taskSpawn ("sample_app", 20, 0,
2000, (FUNCPTR) sample_app, 0,0,0,0,0,0,0,0,0,0)
of course, I defined sample_app function also....
and I create new bootable project based on my BSP directory.
So, I got some project files and I found that usrAppInit()
void usrAppInit (void)
{
#ifdef USER_APPL_INIT
USER_APPL_INIT; /* for backwards compatibility */
#endif
/* add application specific code here */
}
When I made ROM based Vxworks (ie. Vxworks_rom.hex) and succeeded in
booting. but usrAppInit() function never run....
so, I checked whether checked USER_APPL_INIT is defined.
unfortunately, in my usrAppInit() fucntion .
this function said that ......
void usrAppInit (void)
{
#ifdef USER_APPL_INIT
USER_APPL_INIT; /* for backwards compatibility */
printf("defined.......");
#else
printf("undedined.......");
#endif
/* add application specific code here */
}
the result is "undefined......."
Is something that I forgot ??
I defined
#define INCLUDE_USER_APPL
#define USER_APPL_INIT
in my config.h file in my BSP directory...
I don't know what else should I change or add....
please post your comments......
Thank you..........
kingseft.
Sent via Deja.com http://www.deja.com/
Before you buy.
---------------------------
Newsgroups: comp.os.vxworks
Subject: Error during using JavaCodeCompact
Date: Mon, 04 Dec 2000 09:41:43 GMT
From: tobyhage@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <90fosm$ip5$1@nnrp1.deja.com>
We want to use JavaCode compact. When we configure the Java classes
with "complete ROMmized configuration" and try to start java -version
or a java application we receive the following error messages in the
shell:
- -> javaConfigShow()
- ------------------------
Native libraries file zip net awt jpeg math
- ------------------------
java.class.path /java/lib/classes.zip:/java/lib/wrs/jar
java.home /java
java.user.name target
java.user.home /java
java.user.dir /java
path.separator :
file.separator /
- ------------------------
Native stack size 32768 bytes
Max Java heap size 33554432 bytes
Stack overflow checks FAST (minimal)
Java thread semaphores priority inversion safe
Native library loading DISABLED
- ------------------------
Idle thread frequency 10 Hz
Async GC period 1 seconds
Async GC threshold 99 %
- ------------------------
On VxWorks exception ABORT VM
Java VM restart ENABLED
JNI arg init hook NULL
Instrumentation hook _javaInstrumentHook (DISABLED)
- ------------------------
awt.toolkit
windriver.awt.chameleon.ChameleonToolkit
sun.graphicssystem sun.awt.gfU.GraphicsSystem
sun.graphicssystem.width 800
sun.graphicssystem.height 600
sun.awt.gfU.pixelType color:8
value = 0 = 0x0
- -> java -version
Exception number 16: Task: 0x3f85bb0 (t27)
Data abort
Program Counter: 0x17399c
Status Register: 0xa0000013
d67bc _vxTaskEntry +28 : 325b60 ()
325c04 _wdbFuncCallLibInit+128: _java ()
173bbc _java +14 : _javaWait ()
173b18 _javaWait +1c : 173950 ()
value = 0 = 0x0
- -> java -version
Exception number 16: Task: 0x3f80a04 (t28)
Data abort
Program Counter: 0x17399c
Status Register: 0xa0000013
d67bc _vxTaskEntry +28 : 325b60 ()
325c04 _wdbFuncCallLibInit+128: _java ()
173bbc _java +14 : _javaWait ()
173b18 _javaWait +1c : 173950 ()
value = 0 = 0x0
- -> java("-verbose HelloWorld")
value = 0 = 0x0
- ->
Exception number 16: Task: 0x3ee7fd4 (tJmain)
Data abort
Program Counter: 0xf3ccc
Status Register: 0x60000013
0 : ???? ()
0 : _ROMPureMethods ()
243044 _ROMPureMethods+310e4: .text2 ()
302e94 .text2 +893c: _ExecuteJava ()
101190 _ExecuteJava +37f0: _excEnterUndef ()
352f14 _excEnterUndef +9c : _excExcContinue ()
323214 _excExcContinue+94 : _excExcHandle ()
322f94 _excExcHandle +174: 32faa8 ()
32fad4 _sigPendDestroy+2a8: 32fa30 ()
32fa88 _sigPendDestroy+25c: _sigPendKill ()
32f7e4 _sigPendKill +10c: 32f380 ()
32f6a0 _sigqueue +938: 32efc8 ()
32f0f4 _sigqueue +38c: 177994 ()
177a08 __jhpiThreadUnlock+3a8: _jhpiAbort ()
1794c8 _jhpiAbort +9c : _javaDumpThreads ()
fd47c _javaDumpThreads+64 : _jhpiThreadEnumerateOver ()
178c54 _jhpiThreadEnumerateOver+60 : _DumpThreadsHelper ()
fd34c _DumpThreadsHelper+20 : _threadDumpInfo ()
fd1d4 _threadDumpInfo+38 : _jio_fprintf ()
120dfc _jio_fprintf +18 : _jio_vfprintf ()
120e20 _jio_vfprintf +1c : _vfprintf ()
33e580 _vfprintf +148: _fflush ()
335c64 _fflush +70 : ___sflush ()
335d34 ___sflush +c4 : ___swrite ()
35b070 ___swrite +4c : _write ()
e9e74 _write +c : _iosWrite ()
3307a4 _iosWrite +b0 : _tyWrite ()
350d18 _tyWrite +a4 : 351670 ()
3516b0 _tyIRd +668: 3260c8 ()
3260d8 _wdbVioDrv +200: _wdbEventPost ()
323ff0 _wdbEventPost +80 : _wdbNotifyHost ()
31cc20 _wdbNotifyHost +34 : 322a60 ()
322a80 _udpRcv +26c: 466c8 ()
and the following on the output screen:
Usage: (null) [-options] class
where options include:
-help print out this message
-version print out the build version
-v -verbose turn on verbose mode
-debug enable remote JAVA debugging
-asyncgc allow asynchronous garbage collection
-verbosegc print a message when garbage collection occurs
-noclassgc disable class garbage collection
-ss set the maximum native stack size for any thread
-oss set the maximum Java stack size for any thread
-ms set the initial Java heap size
-mx set the maximum Java heap size
-mr set the red heap reserve size
-my set the yellow heap reserve size
-D= set a system property
-classpath
list directories in which to look for classes
-verify verify all classes when read in
-verifyremote verify classes read in over the network [default]
-noverify do not verify any class
Data abort
Exception address: 0x0017399c
Current Processor Status Register: 0xa0000013
Task: 0x3f85bb0 "t27"
Data abort
Exception address: 0x0017399c
Current Processor Status Register: 0xa0000013
Task: 0x3f80a04 "t28"
SIGSEGV occured.
Full thread dump:
"main" (TID:0x1edffe0, sys_thread_t:0x3ef4288, taskId:0x3ee7fd4
state:runnable name:tJmain, stack_base:0x3ee7fd4, stack_high:0xee0
stack_size:32092) prio=5 *current thread*
java.lang.System.initializeSystemClass()
"Finalizer thread" (TID:0x1ee0010, sys_thread_t:0x3eeb898,
taskId:0x3ef38f0 state:runnable name:Finalizer threa,
stack_base:0x3ef38e8, stack_high:0x138 stack_size:32784) prio=1
Monitor Cache Dump:
Registered Monitor Dump:
Thread queue lock: 0x3ef42f0 0x0 n 00 m 0x3f7ad10 b
0x3f7acec #e 0
Name and type hash table lock: 0x3ef434c 0x0 n 00 m
0x3f7ad58 b 0x3f7ad34 #e 0
String intern lock: 0x3ef439c 0x0 n 00 m 0x3f7ada0 b
0x3f7ad7c #e 0
JNI pinning lock: 0x3ef4bd8 0x0 n 00 m 0x3f7ade8 b
0x3f7adc4 #e 0
JNI global reference lock: 0x3ef4c30 0x0 n 00 m 0x3f7ae30
b 0x3f7ae0c #e 0
BinClass lock: 0x3ef4c7c 0x0 n 00 m 0x3f7ae78 b 0x3f7ae54
#e 0
Class loading lock: 0x3ef4ccc 0x0 n 00 m 0x3f7aec0 b
0x3f7ae9c #e 0
Java stack lock: 0x3ef4d18 0x0 n 00 m 0x3f7af08 b
0x3f7aee4 #e 0
Code rewrite lock: 0x3ef4d68 0x0 n 00 m 0x3f7af50 b
0x3f7af2c #e 0
Heap lock: 0x3ef4db0 0x0 n 00 m 0x3f7af98 b 0x3f7af74 #e 0
Has finalization queue lock: 0x3ef4e08 0x0 n 00 m
0x3f7afe0 b 0x3f7afbc #e 0
Finalize me queue lock: 0x3ef4e5c 0x0 n 00 m 0x3f7b028 b
0x3f7b004 #e 0
Monitor registry: 0x3ef8100 0x3ee7fd4 n 01 m 0x3f7b998 b
0x3f7b974 #e 0
Data abort
Exception address: 0x000f3ccc
Current Processor Status Register: 0x60000013
Task: 0x3ee7fd4 "tJmain"
what could this be??
with kind regards,
Toby Hage
Sent via Deja.com http://www.deja.com/
Before you buy.
---------------------------
Newsgroups: comp.os.vxworks
Subject: Error during using JavaCodeCompact
Date: Mon, 04 Dec 2000 09:41:55 GMT
From: tobyhage@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <90fot2$ip8$1@nnrp1.deja.com>
We want to use JavaCode compact. When we configure the Java classes
with "complete ROMmized configuration" and try to start java -version
or a java application we receive the following error messages in the
shell:
- -> javaConfigShow()
- ------------------------
Native libraries file zip net awt jpeg math
- ------------------------
java.class.path /java/lib/classes.zip:/java/lib/wrs/jar
java.home /java
java.user.name target
java.user.home /java
java.user.dir /java
path.separator :
file.separator /
- ------------------------
Native stack size 32768 bytes
Max Java heap size 33554432 bytes
Stack overflow checks FAST (minimal)
Java thread semaphores priority inversion safe
Native library loading DISABLED
- ------------------------
Idle thread frequency 10 Hz
Async GC period 1 seconds
Async GC threshold 99 %
- ------------------------
On VxWorks exception ABORT VM
Java VM restart ENABLED
JNI arg init hook NULL
Instrumentation hook _javaInstrumentHook (DISABLED)
- ------------------------
awt.toolkit
windriver.awt.chameleon.ChameleonToolkit
sun.graphicssystem sun.awt.gfU.GraphicsSystem
sun.graphicssystem.width 800
sun.graphicssystem.height 600
sun.awt.gfU.pixelType color:8
value = 0 = 0x0
- -> java -version
Exception number 16: Task: 0x3f85bb0 (t27)
Data abort
Program Counter: 0x17399c
Status Register: 0xa0000013
d67bc _vxTaskEntry +28 : 325b60 ()
325c04 _wdbFuncCallLibInit+128: _java ()
173bbc _java +14 : _javaWait ()
173b18 _javaWait +1c : 173950 ()
value = 0 = 0x0
- -> java -version
Exception number 16: Task: 0x3f80a04 (t28)
Data abort
Program Counter: 0x17399c
Status Register: 0xa0000013
d67bc _vxTaskEntry +28 : 325b60 ()
325c04 _wdbFuncCallLibInit+128: _java ()
173bbc _java +14 : _javaWait ()
173b18 _javaWait +1c : 173950 ()
value = 0 = 0x0
- -> java("-verbose HelloWorld")
value = 0 = 0x0
- ->
Exception number 16: Task: 0x3ee7fd4 (tJmain)
Data abort
Program Counter: 0xf3ccc
Status Register: 0x60000013
0 : ???? ()
0 : _ROMPureMethods ()
243044 _ROMPureMethods+310e4: .text2 ()
302e94 .text2 +893c: _ExecuteJava ()
101190 _ExecuteJava +37f0: _excEnterUndef ()
352f14 _excEnterUndef +9c : _excExcContinue ()
323214 _excExcContinue+94 : _excExcHandle ()
322f94 _excExcHandle +174: 32faa8 ()
32fad4 _sigPendDestroy+2a8: 32fa30 ()
32fa88 _sigPendDestroy+25c: _sigPendKill ()
32f7e4 _sigPendKill +10c: 32f380 ()
32f6a0 _sigqueue +938: 32efc8 ()
32f0f4 _sigqueue +38c: 177994 ()
177a08 __jhpiThreadUnlock+3a8: _jhpiAbort ()
1794c8 _jhpiAbort +9c : _javaDumpThreads ()
fd47c _javaDumpThreads+64 : _jhpiThreadEnumerateOver ()
178c54 _jhpiThreadEnumerateOver+60 : _DumpThreadsHelper ()
fd34c _DumpThreadsHelper+20 : _threadDumpInfo ()
fd1d4 _threadDumpInfo+38 : _jio_fprintf ()
120dfc _jio_fprintf +18 : _jio_vfprintf ()
120e20 _jio_vfprintf +1c : _vfprintf ()
33e580 _vfprintf +148: _fflush ()
335c64 _fflush +70 : ___sflush ()
335d34 ___sflush +c4 : ___swrite ()
35b070 ___swrite +4c : _write ()
e9e74 _write +c : _iosWrite ()
3307a4 _iosWrite +b0 : _tyWrite ()
350d18 _tyWrite +a4 : 351670 ()
3516b0 _tyIRd +668: 3260c8 ()
3260d8 _wdbVioDrv +200: _wdbEventPost ()
323ff0 _wdbEventPost +80 : _wdbNotifyHost ()
31cc20 _wdbNotifyHost +34 : 322a60 ()
322a80 _udpRcv +26c: 466c8 ()
and the following on the output screen:
Usage: (null) [-options] class
where options include:
-help print out this message
-version print out the build version
-v -verbose turn on verbose mode
-debug enable remote JAVA debugging
-asyncgc allow asynchronous garbage collection
-verbosegc print a message when garbage collection occurs
-noclassgc disable class garbage collection
-ss set the maximum native stack size for any thread
-oss set the maximum Java stack size for any thread
-ms set the initial Java heap size
-mx set the maximum Java heap size
-mr set the red heap reserve size
-my set the yellow heap reserve size
-D= set a system property
-classpath
list directories in which to look for classes
-verify verify all classes when read in
-verifyremote verify classes read in over the network [default]
-noverify do not verify any class
Data abort
Exception address: 0x0017399c
Current Processor Status Register: 0xa0000013
Task: 0x3f85bb0 "t27"
Data abort
Exception address: 0x0017399c
Current Processor Status Register: 0xa0000013
Task: 0x3f80a04 "t28"
SIGSEGV occured.
Full thread dump:
"main" (TID:0x1edffe0, sys_thread_t:0x3ef4288, taskId:0x3ee7fd4
state:runnable name:tJmain, stack_base:0x3ee7fd4, stack_high:0xee0
stack_size:32092) prio=5 *current thread*
java.lang.System.initializeSystemClass()
"Finalizer thread" (TID:0x1ee0010, sys_thread_t:0x3eeb898,
taskId:0x3ef38f0 state:runnable name:Finalizer threa,
stack_base:0x3ef38e8, stack_high:0x138 stack_size:32784) prio=1
Monitor Cache Dump:
Registered Monitor Dump:
Thread queue lock: 0x3ef42f0 0x0 n 00 m 0x3f7ad10 b
0x3f7acec #e 0
Name and type hash table lock: 0x3ef434c 0x0 n 00 m
0x3f7ad58 b 0x3f7ad34 #e 0
String intern lock: 0x3ef439c 0x0 n 00 m 0x3f7ada0 b
0x3f7ad7c #e 0
JNI pinning lock: 0x3ef4bd8 0x0 n 00 m 0x3f7ade8 b
0x3f7adc4 #e 0
JNI global reference lock: 0x3ef4c30 0x0 n 00 m 0x3f7ae30
b 0x3f7ae0c #e 0
BinClass lock: 0x3ef4c7c 0x0 n 00 m 0x3f7ae78 b 0x3f7ae54
#e 0
Class loading lock: 0x3ef4ccc 0x0 n 00 m 0x3f7aec0 b
0x3f7ae9c #e 0
Java stack lock: 0x3ef4d18 0x0 n 00 m 0x3f7af08 b
0x3f7aee4 #e 0
Code rewrite lock: 0x3ef4d68 0x0 n 00 m 0x3f7af50 b
0x3f7af2c #e 0
Heap lock: 0x3ef4db0 0x0 n 00 m 0x3f7af98 b 0x3f7af74 #e 0
Has finalization queue lock: 0x3ef4e08 0x0 n 00 m
0x3f7afe0 b 0x3f7afbc #e 0
Finalize me queue lock: 0x3ef4e5c 0x0 n 00 m 0x3f7b028 b
0x3f7b004 #e 0
Monitor registry: 0x3ef8100 0x3ee7fd4 n 01 m 0x3f7b998 b
0x3f7b974 #e 0
Data abort
Exception address: 0x000f3ccc
Current Processor Status Register: 0x60000013
Task: 0x3ee7fd4 "tJmain"
what could this be??
with kind regards,
Toby Hage
Sent via Deja.com http://www.deja.com/
Before you buy.
---------------------------
Newsgroups: comp.os.vxworks
Subject: Re: Can the target write to the host's file system?
Date: Mon, 04 Dec 2000 12:30:06 +0200
From: Amit Raam
Organization: Verio
Message-ID: <14sm2tsag2kpgv1ct6ft978tkn5vofmrfg@4ax.com>
References: <9093m2$c3t$1@murdoch.harvard.net>
Yes: just do an fopen; we use an FTP server on NT to accomplish this. vxWorks
does a nice job of virtualizing a remote file system via ftp.
To write to c:\sysdump, you'll need to define the FTP home at c:\ .
Make sure to handle security on your NT host (is it a closed network?)
FILE *f;
if (f = fopen(perfDataCollFileName, "w") == NULL)
handleError();
if (fprintf(f, "text") < 0)
handleError();
if (fclose(f) == EOF)
handleError();
"Andrew Carlisle" wrote:
>Hello,
>
>Is it possible to have the vxWorks target write to the host's file system?
>For instance, I have a target running vxWorks that has no hard drive. I
>would like to be able to open and close files on the NT host's file system.
>For instance, I would like to write to c:/sysdump/dump.txt from the target.
>Is this possible?
>
>Andy
>
>acarlisle@envoynetworks.com
>
Amit Raam
Email address: raama at indigo.co.il
- --------------------------------------------------
".sigless and proud of it" (R) (C) (tm) (look-and-feel)
Std. disclaimer:I don't represent the opinions of anyone but myself.
---------------------------
End of New-News digest
**********************
From vxwexplo-errs@csg.lbl.gov Mon Dec 4 06:09:00 2000
From: David Laight
Date: Mon Dec 4 06:09:02 PST 2000
Subject: Re: Touch Screen Calibration
You will need to do the calibration in the touch screen driver. It (typically)
has to convert an A-D output to pixels.
I recall the vxWorks UGL driver template using some horrid sequence of linear
transformations for each of X and Y. I don't do it that way....
What I found with our 4-wire touch screen is that the calibrated X output has to
depend on both raw input values. However the transformation is linear in each
term - but does need a product term.
X = a + b.x + c.y + d.x.y (b is dominant)
Y = e + f.x + g.y + h.x.y (g is dominant)
This requires test points in all four corners.
(The x.y term is required because the resistance of the 'bus bars' of the touch
panel isn't completely insignificant.)
You really do need the raw data to work out these numbers - not the converted
values using the existing values. Also you probably need to average a
considerable number of readings.
One day I even thought that the calibration was temperature sensitive!
David
> Subject: Re: Touch Screen Calibration
> X-UIDL: (jC!!a5N"!]J
> >My problem is, how do I al of this. I have created a window using Zinc with
> >the two dot positions as
> >described. How do I capture and save the touch position when the user
> >touches the screen at the dot positions? Is there some way of getting this
> >information from the UGL?
>
> You may find the information you require in the UGL Developer's Kit manual.
> The Input Driver API section describes some of the detail about the absolute
> pointer event information. There's also discussion of calibration in the
> Writing A New Input Driver Section.
>
> I don't know if this event information is available at Zinc, I've done touch
> screen recalibration at the UGL level in the past.
>
----------------------------------------------------------------
David Laight email: dsl@tadpole.co.uk
Tadpole Technology plc phone: +44 1223 428 232
Cambridge, UK fax: +44 1223 428 201
From vxwexplo-errs@csg.lbl.gov Mon Dec 4 10:05:09 2000
From: "Komuraiah Poodari"
Date: Mon Dec 4 10:05:12 PST 2000
Subject: RouteLookup function
Hi,
Is there any function in VxWorks SENS stack to perform routeLookup, i.e. to find out the interface corresponding to given destination address ?
The function routeShow() displays the entire route table.
thanks for the help.
- Komuraiah
From vxwexplo-errs@csg.lbl.gov Mon Dec 4 14:57:10 2000
From: s.eckardt@ixthos.com
Date: Mon Dec 4 14:57:12 PST 2000
Subject: Re: usrAppInit() function in Vxworks Project
I believe that in Tornado 2, the recommended way to spawn user tasks is to
code them inline after the comment: /* add application specific code here
*/.
If you prefer to spawn your tasks in the Tornado 1 style, by using the
USER_APPL_INIT macro, you can #include config.h at the start of
usrAppInit.c. Then your definition of USER_APPL_INIT will be available to
the source file.
Suzie Eckardt
Chief Software Engineer
Ixthos, Inc.
s.eckardt@ixthos.com
From vxwexplo-errs@csg.lbl.gov Mon Dec 4 21:22:17 2000
From: karunakaran nair
Date: Mon Dec 4 21:22:20 PST 2000
Subject: comp.os.vxworks newsdigest
--0-608413784-975993736=:25484
Content-Type: text/plain; charset=us-ascii
hello,
please somebody tell me how to make scripts at the shell and execute. for example instead of typing commands at the shell i want to capture them all in a file and execute them at the shell. thanks,KN
---------------------------------
Do You Yahoo!?
Yahoo! Shopping - Thousands of Stores. Millions of Products.
--0-608413784-975993736=:25484
Content-Type: text/html; charset=us-ascii
hello,
please somebody tell me how to make scripts at the shell and execute. for example instead of typing commands at the shell i want to capture them all in a file and execute them at the shell. thanks,KN
Do You Yahoo!?
Yahoo! Shopping -
Thousands of Stores. Millions of Products.
--0-608413784-975993736=:25484--
From vxwexplo-errs@csg.lbl.gov Tue Dec 5 04:03:08 2000
From: Vxworks Exploder
Date: Tue Dec 5 04:03:10 PST 2000
Subject: comp.os.vxworks newsdigest
Comp.Os.Vxworks Daily Digest Tue Dec 5 04:03:04 PST 2000
Subject: NaN - how can rid of it?
Subject: re: Touch Screen Calibration
Subject: Re: NaN - how can rid of it?
Subject: Re: integrating network drivers
Subject: Re: upnp
Subject: mcp750ha
Subject: Post-load module verification
Subject: Re: NaN - how can rid of it?
Subject: RouteLookup function
Subject: Re: Maximum number of semaphores in VxWorks
Subject: Re: How to calculate Task stack space?
Subject: Re: Post-load module verification
Subject: Re: [Question] Difference VxWorks Boot Sequence??
Subject: Sio vs. Serial
Subject: wdb agent restart
Subject: Assembler constants.
Subject: tffs and ads8260
Subject: Shared memory network problems
Subject: Re: integrating network drivers
Subject: please help me
Subject: re: usrAppInit() function in Vxworks Project
Subject: re: usrAppInit() function in Vxworks Project
Subject: Re: How to calculate Task stack space?
Subject: VxWorks for Linux emulation
Subject: Re: Tornado Compiler
Subject: snmp...snmpdLib.c??
Subject: Re: Maximum number of semaphores in VxWorks
Subject: RTC for PPC mcpn765
-------------------------------------------------------
Newsgroups: comp.os.vxworks
Subject: NaN - how can rid of it?
Date: Mon, 4 Dec 2000 16:59:12 +0200
From: "Elbit user"
Organization: NetVision Israel
Message-ID: <90dna1$7um$1@news.netvision.net.il>
In my system (vxworks, tornado, pentium-target) I have several tasks.
some of them make - floating point calc.
in some of the cases I get NaN as a result of those calcs?
I discovers that when I put taskLock() and taskUnlock() before and after the
calulation the NaN frequancy is much lower (but still exists).
do u have any idea how can I get rid of it?
thanx - Yonatan
---------------------------
Newsgroups: comp.os.vxworks
Subject: re: Touch Screen Calibration
Date: Mon, 4 Dec 2000 14:09:38 +0000 (GMT)
From: David Laight
Organization: Lawrence Berkeley National Laboratory
Message-ID: <200012041409.OAA13714@dsl-2.tadpole.co.uk>
You will need to do the calibration in the touch screen driver. It (typically)
has to convert an A-D output to pixels.
I recall the vxWorks UGL driver template using some horrid sequence of linear
transformations for each of X and Y. I don't do it that way....
What I found with our 4-wire touch screen is that the calibrated X output has to
depend on both raw input values. However the transformation is linear in each
term - but does need a product term.
X = a + b.x + c.y + d.x.y (b is dominant)
Y = e + f.x + g.y + h.x.y (g is dominant)
This requires test points in all four corners.
(The x.y term is required because the resistance of the 'bus bars' of the touch
panel isn't completely insignificant.)
You really do need the raw data to work out these numbers - not the converted
values using the existing values. Also you probably need to average a
considerable number of readings.
One day I even thought that the calibration was temperature sensitive!
David
> Subject: Re: Touch Screen Calibration
> X-UIDL: (jC!!a5N"!]J
> >My problem is, how do I al of this. I have created a window using Zinc with
> >the two dot positions as
> >described. How do I capture and save the touch position when the user
> >touches the screen at the dot positions? Is there some way of getting this
> >information from the UGL?
>
> You may find the information you require in the UGL Developer's Kit manual.
> The Input Driver API section describes some of the detail about the absolute
> pointer event information. There's also discussion of calibration in the
> Writing A New Input Driver Section.
>
> I don't know if this event information is available at Zinc, I've done touch
> screen recalibration at the UGL level in the past.
>
- ----------------------------------------------------------------
David Laight email: dsl@tadpole.co.uk
Tadpole Technology plc phone: +44 1223 428 232
Cambridge, UK fax: +44 1223 428 201
---------------------------
Newsgroups: comp.os.vxworks
Subject: Re: NaN - how can rid of it?
Date: Mon, 4 Dec 2000 13:46:32 GMT
From: Lee DeRaud
Organization: None to speak of
Message-ID:
References: <90dna1$7um$1@news.netvision.net.il>
Sender: nntp@news.boeing.com (Boeing NNTP News Access)
On Mon, 4 Dec 2000 16:59:12 +0200, "Elbit user"
wrote:
>In my system (vxworks, tornado, pentium-target) I have several tasks.
>some of them make - floating point calc.
>in some of the cases I get NaN as a result of those calcs?
>I discovers that when I put taskLock() and taskUnlock() before and after the
>calulation the NaN frequancy is much lower (but still exists).
>
>do u have any idea how can I get rid of it?
Two things necessary for floating-point:
1. Make sure you're setting the VX_FP_TASK flag in the 'options'
parameter to taskSpawn for *all* tasks that use floating-point. That
forces the kernel to save/restore the FPP context during task
switches.
2. Use 'fppSave' and 'fppRestore' to do the same thing in any ISRs
that perform floating-point computations. (See the man pages for
fppLib and fppArchLib.)
Since the taskLock/taskUnlock helped but didn't completely solve your
problem, I'd suspect #2 is your problem.
Lee
---------------------------
Newsgroups: comp.os.vxworks
Subject: Re: integrating network drivers
Date: Mon, 04 Dec 2000 14:25:58 GMT
From: Bill Pringlemeir
Organization: Factory of the mind
Message-ID:
References: <90eniu$cms$1@news.netmar.com> <3A2AF70B.A74CA02@flashcom.net>
Sender: bpringlemeir@DeadDuck
>>>>> "DrDiags" == DrDiags writes:
DrDiags> Sineon, I don't know if you tried using the MACH_EXTRA line
DrDiags> in your makefile. You could put the source locally in your
[snip]
If you are using an `END' system, then you have to also add an entry
to the `End Table'. See configNet.h for more. This is the alternative
for the #INCLUDE_CS8900 etc macros that where used in older versions
of vxWorks.
hth,
Bill
- --
Ft. Bragg Khaddafi NSA CIA Ortega North Korea security South Africa
Saddam Hussein Legion of Doom World Trade Center Cocaine Nazi Honduras
arrangements
---------------------------
Newsgroups: comp.os.vxworks
Subject: Re: upnp
Date: Mon, 04 Dec 2000 14:34:57 GMT
From: Bill Pringlemeir
Organization: Factory of the mind
Message-ID:
References: <90fijt$dnb$1@nnrp1.deja.com>
Sender: bpringlemeir@DeadDuck
>>>>> "gandalf107" == gandalf107 writes:
gandalf107> hi, is there any solution for including the upnp feature
gandalf107> into vxworks? thanks
This question has been /debated/ on the uPnP mailing list. I think it
is hosted by SourceForge. You may be able to get mail archives there.
The basic suggestion is that the Linux uPnP source should be a fairly
easy port to vxWorks.
I don't know of any other solutions.
hth,
Bill
- --
Have you ever tried to talk with someone you saw at a party while
being operated on for club foot? Or thought it was quite worthwhile
to have four Ann Arbor Ambassador terminals to defend against the
bobbie? You will. And the company that will bring it to you: AT&T.
---------------------------
Newsgroups: comp.os.vxworks
Subject: mcp750ha
Date: Mon, 04 Dec 2000 16:00:29 GMT
From: jayantadas@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <90gf2n$3hc$1@nnrp1.deja.com>
Hi,
I am using mcp750ha board. The BSP we have with the vxWorks is for
mcp750. Does anyone know if Windriver has the BSP for MCP750HA? Or we
have to get it from the hardware vendor?
Any help will be appreciated.
Thanks.
Sent via Deja.com http://www.deja.com/
Before you buy.
---------------------------
Newsgroups: comp.os.vxworks
Subject: Post-load module verification
Date: Mon, 04 Dec 2000 16:51:11 GMT
From: jgiurleo@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <90gi1r$69c$1@nnrp1.deja.com>
Hello,
I need a way to verify that my module has loaded correctly, perhaps by
using a checksum. How do I determine the starting address and length for the
executable portion of the module? I don't want to add any local or global
symbols to the calculated checksum so that I can continue to re-verify the
integrity of the executable even after the module has been spawned.
Any help at all is greatly appreciated. Thanks!
Sent via Deja.com http://www.deja.com/
Before you buy.
---------------------------
Newsgroups: comp.os.vxworks
Subject: Re: NaN - how can rid of it?
Date: Mon, 04 Dec 2000 10:00:24 -0800
From: Lori Fraleigh
Organization: Real-Time Innovations
Message-ID: <3A2BDBB8.7F420D9@rti.com>
References: <90dna1$7um$1@news.netvision.net.il>
If you're printing floating point numbers, you need to make
sure you have INCLUDE_FLOATING_POINT defined.
Regards,
Lori
Elbit user wrote:
>
> In my system (vxworks, tornado, pentium-target) I have several tasks.
> some of them make - floating point calc.
> in some of the cases I get NaN as a result of those calcs?
> I discovers that when I put taskLock() and taskUnlock() before and after the
> calulation the NaN frequancy is much lower (but still exists).
>
> do u have any idea how can I get rid of it?
>
> thanx - Yonatan
- --
Lori Fraleigh Real-Time Innovations
Senior Software Engineer ScopeTools Manager
Phone: 408-734-4200 x121 Fax: 408-734-5009 http://www.rti.com
Please send/cc all support related email to support@rti.com
---------------------------
Newsgroups: comp.os.vxworks
Subject: RouteLookup function
Date: Mon, 4 Dec 2000 09:49:21 -0800
From: "Komuraiah Poodari"
Organization: Lawrence Berkeley National Laboratory
Message-ID: <90gmtj$p26$1@overload.lbl.gov>
Hi,
Is there any function in VxWorks SENS stack to perform routeLookup, i.e. to find out the interface corresponding to given destination address ?
The function routeShow() displays the entire route table.
thanks for the help.
- - Komuraiah
---------------------------
Newsgroups: comp.os.vxworks
Subject: Re: Maximum number of semaphores in VxWorks
Date: Mon, 04 Dec 2000 13:48:38 -0600
From: Andrew Johnson
Organization: APS, Argonne National Laboratory
Message-ID: <3A2BF516.FD99356D@aps.anl.gov>
References: <3A285508.DDC57211@c-cube.com>
Venki Mohan wrote:
>
> Is there any limit on the number of semaphores that I can
> use in VxWorks?
>
> I have an application in which i need to instantiate plenty
> of semaphores( > 500). I was wondering if I will have any performance hit.
No, semaphores take up a certain amount of memory but that's usually all.
I'm sure we have systems here that have more than 500 semaphores.
- - Andrew
- --
Every great idea appears crazy to start with.
---------------------------
Newsgroups: comp.os.vxworks
Subject: Re: How to calculate Task stack space?
Date: Mon, 4 Dec 2000 20:09:24 -0000
From: "David Laight"
Organization: BT Internet
Message-ID: <90gtk7$nfm$1@neptunium.btinternet.com>
References: <200011301401.GAA03645@mail22.bigmailbox.com>
> 1.How to calculate accurately the Task stack space?.
> 2.Do we have any "formula"? or technique to calculate?
The 'usual' technique seems to be 'give it some and if it odd things happen
try giving it some more'. As you say this isn't ideal!
Years ago I wrote a program that parsed the object file listing (from a
PASCAL compiler) which tracked the stack depth in each routine at which it
called other routines. This let me calculate a maximum stack usage for the
high level routines. Maximum stack use was often in obscure error
paths......
Doing the same for vxWorks would be a lot harder due to the number of
procedure variables and call tables. Parsing a running system - with a
little manual help for some of the call tables would be enlightening! (and
probably possible.)
The last big one that exploded on me was open() - the number of on-stack
buffers into which the filename is put has to be seen to believed (and not a
size check amongst them). The deep part was netdrv using DNS to find the IP
address of the host.
Certainly it would be useful for windriver to give a hint - even though it
would be very architecture (and compiler) specific.
David
---------------------------
Newsgroups: comp.os.vxworks
Subject: Re: Post-load module verification
Date: Mon, 04 Dec 2000 14:12:34 -0600
From: Andrew Johnson
Organization: APS, Argonne National Laboratory
Message-ID: <3A2BFAB2.35D1E102@aps.anl.gov>
References: <90gi1r$69c$1@nnrp1.deja.com>
jgiurleo@my-deja.com wrote:
>
> I need a way to verify that my module has loaded correctly, perhaps by
> using a checksum. How do I determine the starting address and length for the
> executable portion of the module? I don't want to add any local or global
> symbols to the calculated checksum so that I can continue to re-verify the
> integrity of the executable even after the module has been spawned.
Look up moduleLib in the vxWorks reference manual. moduleCheck() may be
of some use to you for example.
- - Andrew
- --
Every great idea appears crazy to start with.
---------------------------
Newsgroups: comp.os.vxworks
Subject: Re: [Question] Difference VxWorks Boot Sequence??
Date: 4 Dec 2000 20:26:50 GMT
From: Johan Borkhuis
Organization: Lucent Technologies, Nieuwegein NL
Message-ID: <90guma$746$1@news1.xs4all.nl>
References: <90emq6$pp0$1@nnrp1.deja.com>
kingseft@my-deja.com wrote:
: Hi. Vxworks Developers...
: I'm a novice with VxWorks.
: I have a basic question about VxWorks Boot Sequence ...
: when I read some materials about boot sequence.
: I found the diferrences between building from the traditional command
: line method and Vxworks Project Facility methdo..
: Could someone describe some differences about that??
: Why some developers use command line and whey some developers use
: project facility method??
: They said that there is some differences ...
: But I can't find that..
The main difference is that you can only build a bootrom from the
commandline. The bootrom is based on the items as specified in the file
config.h in your BSP directory.
If you want to build a VxWorks image you can use the commandline or the
GUI. In my opinion the best way to build a VxWorks image is using the GUI.
It is the easiest way to configure your image and selecting the right
components.
When you want to do this by hand it is also possible, but you have to
build something like the prjConfig.c and the other files.
Groeten,
Johan
- --
o o o o o o o . . . ______________________________
o _____ || Johan Borkhuis |
.][__n_n_|DD[ ====_____ | borkhuis@lucent.com |
>(________|__|_[_________]_|___________________________|
_/oo OOOOO oo` ooo ooo 'o!o!o o!o!o`
=== VxWorks FAQ: http://www.xs4all.nl/~borkhuis/vxworks/vxworks.html ===
---------------------------
Newsgroups: comp.os.vxworks
Subject: Sio vs. Serial
Date: Mon, 4 Dec 2000 13:53:51 -0700
From: Stephan Esterhuizen
Organization: University of Colorado, Boulder
Message-ID:
Can anybody tell me what the difference is between the src/drv/sio and
src/drv/serial directories are?
I am having trouble figuring out which one to use.
It seems as if the /sio dir has more up to date drivers... is that the
only difference?
Stephan Esterhuizen
Boulder,CO,USA
Icq : 2785232
Homepage : http://www.pythion.com
An optimist sees a glass that is half full.
A pessimist sees a glass that is half empty.
An engineer sees a glass that is twice as big as it needs to be.
---------------------------
Newsgroups: comp.os.vxworks
Subject: wdb agent restart
Date: Mon, 04 Dec 2000 16:22:55 -0500
From: Geoffrey Brown
Message-ID: <3A2C0B2F.C424FF04@ibnets.com>
I've got a 2M binary on an ppc 860T which runs just fine until
I try to attach the tornado tools at which point the wdb agent
restarts. I had no problems when my binary was smaller. The problem
occurs even if usrAppInit simply calls a task which goes to sleep (and
hence doesn't execute the added code).
Are there resource limits in wdb agent that are likely to be
exposed as binaries grow ?
Geoffrey
---------------------------
Newsgroups: comp.os.vxworks
Subject: Assembler constants.
Date: Mon, 04 Dec 2000 21:29:27 GMT
From: Bill Pringlemeir
Organization: Factory of the mind
Message-ID:
Sender: bpringlemeir@DeadDuck
I am trying to compose a major mode in Emacs for editing VxWorks
assembler. Does anyone know if one exists already? I have the
comments and pseudo-ops, and pre-processor directives highlighting
with font-lock. As well, strings are quite fine. I can not find any
reference as to what the `#' and `=' characters do. I think this is
consistent between assembler languages on different architectures.
For example,
ldr r0,=0x55555555
str r0,[sp,#-4]!
What is the meaning of "=0x55555555" and "#-4". Actually, the # is
a value; I just wonder if the `=' has an identical meaning. Is this
built into gas? I would like to set these in font-lock-constant-face,
but I am not sure if that is correct for an equals `='.
Thanks,
Bill
- --
How can you say that if you've seen one source license, you've seen
them all? The oceans are full of dirty fish, right?
---------------------------
Newsgroups: comp.os.vxworks
Subject: tffs and ads8260
Date: Mon, 04 Dec 2000 21:33:44 GMT
From: "Jim"
Organization: Magma Communications Ltd.
Message-ID:
Hello all,
has anyone got the tffs file system to work on the ads8260 ? If anyone has
the file system running on this board it would be very reassuring to know it
is possible ...
It looks like I am writting/formatting the file system to flash correctly
... this can be seen by the below memory dump from the tornado shell ...
however I am still encountering errors.
Cheers and Thanks,
Jim
-> sysTffsFormat
....
....
WRITING TO BLOCK: 0xff800140 < CardAddress = 0x140, length = 4 >
value = -1 = 0xffffffff = udbinfo + 0xfff0abd7
<~~~~~~~~~~~~~~~~~~~~~NOTE THE ERROR
-> tffsShow
0: socket=SIMM: type=0x89aa, unitSize=0x40000, mediaSize=0x400000
value = 50 = 0x32 = '2'
-> d 0xff800000
ff800000: 1303 4349 5346 ff00 4654 4c31 3030 0001 *..CISF..FTL100..*
ff800010: 0100 0000 0000 0912 0000 1000 0058 3a00 *.............X:.*
ff800020: 0000 0400 3b00 00ff 0000 0000 0000 0000 *....;...........*
ff800030: 4400 0000 ffff ffff ffff ffff ffff ffff *D...............*
ff800040: ffff ffff 3000 0000 3000 0000 3000 0000 *....0...0...0...*
ff800050: 3000 0000 3000 0000 408a ffff 408c ffff *0...0...@...@...*
ff800060: 408e ffff 4090 ffff 4092 ffff 4094 ffff *@...@...@...@...*
ff800070: 4096 ffff 4098 ffff 409a ffff 409c ffff *@...@...@...@...*
ff800080: 409e ffff 40a0 ffff 40a2 ffff 40a4 ffff *@...@...@...@...*
ff800090: 40a6 ffff 40a8 ffff 40aa ffff 40ac ffff *@...@...@...@...*
ff8000a0: 40ae ffff 40b0 ffff 40b2 ffff 40b4 ffff *@...@...@...@...*
ff8000b0: 40b6 ffff 40b8 ffff 40ba ffff 40bc ffff *@...@...@...@...*
ff8000c0: 40be ffff 40c0 ffff 40c2 ffff 40c4 ffff *@...@...@...@...*
ff8000d0: 40c6 ffff 40c8 ffff 40ca ffff 40cc ffff *@...@...@...@...*
ff8000e0: 40ce ffff 40d0 ffff 40d2 ffff 40d4 ffff *@...@...@...@...*
ff8000f0: 40d6 ffff 40d8 ffff 40da ffff 40dc ffff *@...@...@...@...*
value = 21 = 0x15
->
---------------------------
Newsgroups: comp.os.vxworks
Subject: Shared memory network problems
Date: Mon, 04 Dec 2000 21:33:51 GMT
From: mark@SPAMYENOTthelambs.freeserve.co.uk.SPAMYENOT (Mark Lamb)
Organization: www.ic24.net
Message-ID: <3a2c0d56.3444883@news.ic24.net>
Hi,
I'm currently having problems using TCP sockets over the shared memory
network. I get panics from (usually) tNetTask and eventually mbufShow
shows corruption of its mbuf statistics (the totals don't agree). This
is often accompanied by a fail light coming on on my slot 0
controller, a National Instruments VXIcpu-030 (68030-based). I'm
currently running vxWorks 5.1.1 (can't change at the moment).
I have a couple of questions:
1. Has anyone else had problems using the shared memory network?
2. Has anyone developed or know of an alternative implementation of
TCP sockets which doesn't use the shared memory?
Thanks,
Mark Lamb
---------------------------
Newsgroups: comp.os.vxworks
Subject: Re: integrating network drivers
Date: Mon, 4 Dec 2000 22:57:13 +0100
From: "JJS"
Organization: LT
Message-ID: <90h3q3$1o07$1@news5.isdnet.net>
References: <90eniu$cms$1@news.netmar.com>
skenda@start.com.au a écrit dans le message
<90eniu$cms$1@news.netmar.com>...
>
>I have been using vxWorks\Tornado 2 for a short time.
>
>My target processor board is a pentium
>
>I am trying to integrate the crystal CS8900 network driver
>
>Up to now the majority of my configuration changes have been done so
through
>the project facility. On this occaision however it seems that this device
is
>only supported through the device \target\src\drv\netif\if_cs
>ie it has not been integrated into the project facility\config.h
>
>I'm still using an external floppy to boot during this intial development
>phase so that changes can be incorporated rapidly.
>
>My question is that since there is no mention of this driver in either
>a) config.h
>b) project facility
>
>how do I make the required changes to:
>b) make a bootable kernal
With the project facilities :
Build Tab -> expand VxWorks build -> properties -> Select macros ->
EXTRA_MODULES -> add the .o driver file
Jack
>
---------------------------
Newsgroups: comp.os.vxworks
Subject: please help me
Date: Mon, 04 Dec 2000 15:21:15 -0700
From: Ben
Message-ID: <3A2C18DB.A32EAA47@hotmail.com>
Reply-To: kanigeri@hotmail.com
hi
i am trying to install vxworks image on my target board(ppc), but i
am getting the following error. can you please tell me where am i going
wrong. i am also attaching the information which i am getting in ftp.
i will be grateful to you for any assistance in this regard.
thank you
following is the information i am getting in terminal emulator
[VxWorks Boot]: @
boot device : dc
processor number : 0
host name : U.S.U..usu.edu
file name : c:\tornado\target\config\mtx603\vxWorks
inet on ethernet (e) : 129.123.4.127:ffffff00
host inet (h) : 129.123.4.237
user (u) : sl149
ftp password (pw) : goud
flags (f) : 0x0
target name (tn) : tornado
Attaching network interface dc0... done.
Attaching network interface lo0... done.
Loading... 670192
......and it gets stuck here...
the follwoing information in ftp is
[L 0010] 11/30/00 15:46:51 Connection accepted from 129.123.4.127
[C 0010] 11/30/00 15:46:51 Command "USER sl149" received
[C 0010] 11/30/00 15:46:51 PASSword accepted
[L 0010] 11/30/00 15:46:51 User sl149 logged in.
[C 0010] 11/30/00 15:46:51 Command "TYPE I" received
[C 0010] 11/30/00 15:46:51 TYPE set to I N
[C 0010] 11/30/00 15:46:51 Command "PORT 129,123,4,127,4,1" received
[C 0010] 11/30/00 15:46:51 PORT set to 129.123.4.127 - 1025 (4,1)
[C 0010] 11/30/00 15:46:51 Command "RETR
c:\tornado\target\config\mtx603\vxworks" received
[C 0010] 11/30/00 15:46:51 RETRieve started on file
c:\tornado\target\config\mtx603\vxworks
[C 0010] 11/30/00 15:46:58 Transfer aborted
[G 0010] 11/30/00 15:46:58 Got file
c:\tornado\target\config\mtx603\vxworks unsuccessfully
[! 0010] 11/30/00 15:46:58 Timeout while sending data
[C 0010] 11/30/00 15:46:58 QUIT or close - user sl149 logged ou
---------------------------
Newsgroups: comp.os.vxworks
Subject: re: usrAppInit() function in Vxworks Project
Date: Mon, 4 Dec 2000 17:43:35 -0500
From: s.eckardt@ixthos.com
Organization: Lawrence Berkeley National Laboratory
Message-ID: <90h8g4$4q9$1@overload.lbl.gov>
I believe that in Tornado 2, the recommended way to spawn user tasks is to
code them inline after the comment: /* add application specific code here
*/.
If you prefer to spawn your tasks in the Tornado 1 style, by using the
USER_APPL_INIT macro, you can #include config.h at the start of
usrAppInit.c. Then your definition of USER_APPL_INIT will be available to
the source file.
Suzie Eckardt
Chief Software Engineer
Ixthos, Inc.
s.eckardt@ixthos.com
---------------------------
Newsgroups: comp.os.vxworks
Subject: re: usrAppInit() function in Vxworks Project
Date: Tue, 05 Dec 2000 01:00:32 GMT
From: kingseft@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <90hend$vls$1@nnrp1.deja.com>
References: <90h8g4$4q9$1@overload.lbl.gov>
Really Thank you for your help..
kingseft.
In article <90h8g4$4q9$1@overload.lbl.gov>,
s.eckardt@ixthos.com wrote:
> I believe that in Tornado 2, the recommended way to spawn user tasks
is to
> code them inline after the comment: /* add application specific code
here
> */.
>
> If you prefer to spawn your tasks in the Tornado 1 style, by using the
> USER_APPL_INIT macro, you can #include config.h at the start of
> usrAppInit.c. Then your definition of USER_APPL_INIT will be
available to
> the source file.
>
> Suzie Eckardt
> Chief Software Engineer
> Ixthos, Inc.
> s.eckardt@ixthos.com
>
>
Sent via Deja.com http://www.deja.com/
Before you buy.
---------------------------
Newsgroups: comp.os.vxworks
Subject: Re: How to calculate Task stack space?
Date: Tue, 05 Dec 2000 01:47:02 GMT
From: snaphook@southwesternbell.net (Bruce)
Organization: SBC Internet Services
Message-ID: <3a2c48d7.6857835@news.swbell.net>
References: <200011301401.GAA03645@mail22.bigmailbox.com> <90gtk7$nfm$1@neptunium.btinternet.com>
In comp.os.vxworks
"David Laight" wrote:
>> 1.How to calculate accurately the Task stack space?.
>> 2.Do we have any "formula"? or technique to calculate?
>
>The 'usual' technique seems to be 'give it some and if it odd things happen
>try giving it some more'. As you say this isn't ideal!
>
>Years ago I wrote a program that parsed the object file listing (from a
>PASCAL compiler) which tracked the stack depth in each routine at which it
>called other routines. This let me calculate a maximum stack usage for the
>high level routines. Maximum stack use was often in obscure error
>paths......
>
>Doing the same for vxWorks would be a lot harder due to the number of
>procedure variables and call tables. Parsing a running system - with a
>little manual help for some of the call tables would be enlightening! (and
>probably possible.)
VxWorks is actually easier. Not only does the browser show task usage, you
can call checkStack() or is it stackCheck() from the command line and see
the usage and high water marks.
Bruce
---------------------------
Newsgroups: comp.os.vxworks
Subject: VxWorks for Linux emulation
Date: Tue, 05 Dec 2000 05:52:52 GMT
From: bsjung8472@my-deja.com
Organization: Deja.com - Before you buy.
Message-ID: <90hvri$cs6$1@nnrp1.deja.com>
I'm new to Vxworks... but familar to the embedded
system such as embedded linux...
I've designed the embedded linux for SA-1110...
Anyway, I found the pSOS2Linux project
for pSOS emulation in Linux environment...
and I've known the FreeDOS and ReactOS (NT
clone)...
Do anybody know similar project for VxWorks?
If it does not exist, I would like to start
new project for hobby...
I hope this is regal and help anybody who wants
to know VxWorks internals.
From Seoul...
Byungsoo Jung
Sent via Deja.com http://www.deja.com/
Before you buy.
---------------------------
Newsgroups: comp.os.vxworks
Subject: Re: Tornado Compiler
Date: Mon, 4 Dec 2000 22:09:11 -0800
From: "Jianfei XU"
Organization: SBC Internet Services
Message-ID: <6D%W5.251$4T1.189523@news.pacbell.net>
References: <8vv0t6$f7h$1@overload.lbl.gov>
with gcc, you can use -x to specify the language.
"Thach Vu" wrote in message
news:8vv0t6$f7h$1@overload.lbl.gov...
> This message is in MIME format. Since your mail reader does not understand
> this format, some or all of this message may not be legible.
>
> ------_=_NextPart_001_01C058C5.050C2EE0
> Content-Type: text/plain;
> charset="iso-8859-1"
>
> Hi,
>
> I have a question regarding to Tornado 2.0 Compiler.
> I tried to compile a header file that contains C++ comment (//) but it
> didn't accept this comment style.
> I had to change it to C comment style (/* */) in order to make the
compiler
> work.
> I found it strange that if the file is cpp files then it will accept the
C++
> comment style.
> Is there any way that the compiler will read the C++ comment style inside
my
> header file.
> The reason I ask is that the header file contains an awful lot of C++
> comments and I don't want to convert them into C comment.
> Thank you for any of your help.
>
> Thach Vu
>
>
> ------_=_NextPart_001_01C058C5.050C2EE0
> Content-Type: text/html;
> charset="iso-8859-1"
> Content-Transfer-Encoding: quoted-printable
>
>
>
>
> charset=3Diso-8859-1">
> 5.5.2652.35">
> Tornado Compiler
>
>
>
> Hi,
>
>
> I have a question regarding to Tornado =
> 2.0 Compiler.
>
I tried to compile a header file that =
> contains C++ comment (//) but it didn't accept this comment =
> style.
>
I had to change it to C comment style =
> (/* */) in order to make the compiler work.
>
I found it strange that if the file =
> is cpp files then it will accept the C++ comment style.
>
Is there any way that the compiler =
> will read the C++ comment style inside my header file.
>
The reason I ask is that the header =
> file contains an awful lot of C++ comments and I don't want to convert =
> them into C comment.
>
> Thank you for any of your help.
>
>
> Thach Vu
>
>
>
>
> ------_=_NextPart_001_01C058C5.050C2EE0--
>
---------------------------
Newsgroups: comp.os.vxworks
Subject: snmp...snmpdLib.c??
Date: Tue, 5 Dec 2000 15:25:19 +0800
From: "Bibo"
Organization: SEEDNet News Service
Message-ID: <90i52e$4mq$1@news.seed.net.tw>
hi all:
I had try to use WindNet SNMPv1/v2c porting to my target(8240).
But I has some question....
1.When I make my object code integrated to the library, it show the error
message
" out of envirnment"....what does it mean??
2.I find that \target\src\snmpv1\agent\snmpdLib.c is not exist. When the
SNMP agent initialize(from snmpdInit() to spawn the task "tSnmpd"),it always
went to snmpdLib.o find the function.How can I modify this"snmpdLib.o"??Is
the snmpIoLib.c has no effect??
thanks!
Best Regards!
Roy
---------------------------
Newsgroups: comp.os.vxworks
Subject: Re: Maximum number of semaphores in VxWorks
Date: Tue, 05 Dec 2000 07:34:16 GMT
From: pwhicker@home.com (Paul Whicker)
Organization: Excite@Home - The Leader in Broadband http://home.com/faster
Message-ID:
References: <3A285508.DDC57211@c-cube.com> <3A2BF516.FD99356D@aps.anl.gov>
A semaphore is just a malloced data structure with a bunch of suitable access
routines to control its use, which is why you can get invalid object errors by
trying to do semaphore things to a random pointer. So there is no
computational overhead, if you don't use them they don't consumer anything
except a little memory.
Paul
In article <3A2BF516.FD99356D@aps.anl.gov>, Andrew Johnson
wrote:
>Venki Mohan wrote:
>>
>> Is there any limit on the number of semaphores that I can
>> use in VxWorks?
>>
>> I have an application in which i need to instantiate plenty
>> of semaphores( > 500). I was wondering if I will have any performance hit.
>
>No, semaphores take up a certain amount of memory but that's usually all.
>I'm sure we have systems here that have more than 500 semaphores.
>
>- Andrew
---------------------------
Newsgroups: comp.os.vxworks
Subject: RTC for PPC mcpn765
Date: Tue, 05 Dec 2000 10:46:47 +0100
From: Luca Garbo
Organization: Swisscom IP+ (post doesn't reflect views of Swisscom)
Message-ID: <3A2CB987.6FCEF3E9@stest.ch>
- --------------BF31490FA2C9B3931E487429
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Hi guys , does anybody know anything about real time clock functions on
mcpn765 board as well as RTC struct on cpv5350?
Thanks in advance
- --
Luca Garbo
Phone: +41 763 55 6795, http://www.stest.com
Wavetek Wandel & Goltermann, Foerrlibuckstrasse 62, CH-8037 Zurich
- --------------BF31490FA2C9B3931E487429
Content-Type: text/html; charset=us-ascii
Content-Transfer-Encoding: 7bit
Hi guys , does anybody know anything about real time clock functions on
mcpn765 board as well as RTC struct on cpv5350?
Thanks in advance
--
Luca Garbo
Phone: +41 763 55 6795, http://www.stest.com
Wavetek Wandel & Goltermann, Foerrlibuckstrasse 62, CH-8037 Zurich
- --------------BF31490FA2C9B3931E487429--
---------------------------
End of New-News digest
**********************
From vxwexplo-errs@csg.lbl.gov Tue Dec 5 07:07:40 2000
From: David Laight
Date: Tue Dec 5 07:07:43 PST 2000
Subject: Re: How to calculate Task stack space?
> > Years ago I wrote a program that parsed the object file listing (from a
> > PASCAL compiler) which tracked the stack depth in each routine at which it
> > called other routines. This let me calculate a maximum stack usage for the
> > high level routines. Maximum stack use was often in obscure error
> > paths......
> >
> > Doing the same for vxWorks would be a lot harder due to the number of
> > procedure variables and call tables. Parsing a running system - with a
> > little manual help for some of the call tables would be enlightening! (and
> > probably possible.)
> VxWorks is actually easier. Not only does the browser show task usage, you
> can call checkStack() or is it stackCheck() from the command line and see
> the usage and high water marks.
This is the answer to the wrong question!
The browser and stackCheck() only show the amount of stack the task HAS USED.
What you need to determine is the amount a task MIGHT USE.
Unless you can guarantee your task has executed ALL its code paths (and all
those in the vxWorks kernel it MIGHT go through) then the numbers returned by
stackCheck() are almost useless.
IF you have memory to throw at stacks you can give every process several kb
spare and hope that is enough. If you are tight on memory - or trying to run a
lot of tasks - you will typically leave tasks that CAN blow their allocated
stack space.
David
----------------------------------------------------------------
David Laight email: dsl@tadpole.co.uk
Tadpole Technology plc phone: +44 1223 428 232
Cambridge, UK fax: +44 1223 428 201
From vxwexplo-errs@csg.lbl.gov Tue Dec 5 07:21:43 2000
From: David Laight
Date: Tue Dec 5 07:21:45 PST 2000
Subject: Re: Assembler constants.
> I think this is consistent between assembler languages on different
> architectures. For example,
> ldr r0,=0x55555555
> str r0,[sp,#-4]!
> What is the meaning of "=0x55555555" and "#-4".
This is obviously ARM assembler!
As you guessed, # just identifies the value following as a constant.
The '=' is an ARM special. The ARM processor can only load 8 bit constants
(rotated by an even number of bits). Generating a 32 bit constant that way
would require 4 instructions - and be a real pain for a linker. The '=' is a
short form for a PC relative load of a memory location (typically not far below
the instruction) into which the compiler will place the constant.
ie
ldr r0,=0x55555555
is equivalent to:
ldr r0,[pc,#lbl-.-8]
...
lbl: .long 0x55555555
David
PS: isn't EMACS for those who don't understand TECO? - maybe I shouldn't start
editor flame wars in the vxWorks newsgroup.
----------------------------------------------------------------
David Laight email: dsl@tadpole.co.uk
Tadpole Technology plc phone: +44 1223 428 232
Cambridge, UK fax: +44 1223 428 201
From vxwexplo-errs@csg.lbl.gov Tue Dec 5 13:33:30 2000
From: "timallen"
Date: Tue Dec 5 13:33:33 PST 2000
Subject: Re:stephan@pythion.com
>Can anybody tell me what the difference is between the src/drv/sio and
>src/drv/serial directories are?
The sio drivers include polled mode functions which allow system mode
debugging if the target server is connected via the serial ports. They
effectively replaced the older serial drivers, which seem to be kept for
backwards compatibility.
Regards,
Tim.
--------------------
Tim Allen Ltd
Real Time Embedded Software Consultancy
VxWorks
From vxwexplo-errs@csg.lbl.gov Tue Dec 5 18:14:41 2000
From: "rtos E"
Date: Tue Dec 5 18:14:43 PST 2000
Subject: Re: mcp750ha
Hi jayantadas!
We do the thing similar to yours. Windriver has the BSP for MCP750HA for
sure. We've got it from wrs' local distributer this September. We use it
well. Of course some note from the hardware vendor may be necessary also.
HTH.
Regards
wwen
------------------------------------------------------------------------------------------------------------------------
Date: Mon, 04 Dec 2000 16:00:29 GMT
jayantadas wrote
>Hi,
>
>I am using mcp750ha board. The BSP we have with the vxWorks is for
>mcp750. Does anyone know if Windriver has the BSP for MCP750HA? Or we
>have to get it from the hardware vendor?
>
>Any help will be appreciated.
>
>Thanks.
_____________________________________________________________________________________
Get more from the Web. FREE MSN Explorer download : http://explorer.msn.com
From vxwexplo-errs@csg.lbl.gov Tue Dec 5 21:33:31 2000
From: "Girish V. Gulawani"
Date: Tue Dec 5 21:33:33 PST 2000
Subject: Re: How to calculate Task stack space?
reply is below -
> Newsgroups: comp.os.vxworks
> Subject: How to calculate Task stack space?
> Date: Thu, 30 Nov 2000 06:01:40 -0800
> From: "Deja User"