PDA

View Full Version : Old computers running machine tools and ISA slots..



Milacron of PM
10-17-2005, 11:11 AM
My latest "antique" computer problem...1996 vintage Starrett DCC CMM joined at the hip to 1996 Dell PC via ($3,000) interface card in ISA slot of same. Starrett Apogee software was upgraded to the point that the old computer just can't handle it, but I can't upgrade the computer because they no longer have ISA slots !

Any (cheap) solution to this ?

3 Phase Lightbulb
10-17-2005, 11:22 AM
You should still be able to buy a PCI to ISA bridge card. The PCI to ISA bridge is a PCI card that plugs into a PCI slot and provides one or more ISA slots. There might be some mechanical issues depending on what motherboard/case you use.

-Adrian

Evan
10-17-2005, 11:25 AM
What are your minimum requirements for the software? There are tons of motherboards out there for the high end PIII coppermine cpus using a slocket adapter that will run XP just fine and have ISA slots. The ASUS P2B and the SOYO 6BA series will work and have 2 or 3 ISA slots.

Milacron of PM
10-17-2005, 11:27 AM
I'll check the software for min requirements and get back...thanks.

ngray
10-17-2005, 02:21 PM
You're probably restricted to pentium 3 or Thunderbird class systems, topping out at 1.4mhz:

http://www.newegg.com/Product/Product.asp?Item=N82E16813138015
http://www.shopontheweb.net/biostar-m7mia-amd-761-socket-a-atx-mb-wsound-p-9907.html
http://froogle.google.com/froogle?num=100&hl=en&hs=rAt&lr=lang_en&c2coff=1&safe=off&client=firefox-a&rls=org.mozilla%3Aen-US%3Aofficial&tab=ff&oi=froogler&scoring=p&q=motherboard+i sa+usb+%28%22pentium+4%22+OR+athlon%29&btnG=Search+Froogle

Honestly though, with what you're going to pay in surplus prices, you're probably better to buy an old gateway or dell at 1.4mhz, put in fresh RAM, HD and Win2000, and blow it out with compresssed air.

Cheapest and better off.

Milacron of PM
10-17-2005, 04:23 PM
<font face="Verdana, Arial" size="2">put in fresh RAM, HD and Win2000,</font> That's kinda what I was hoping, as the OS doesn't matter much..just need to get it to handle the newer software. So adding RAM to the current computer might do the trick ? (I need to check the processor speed on this one first of course..wonder what was typical back in 1996 on a mid range Dell PC ?)

Leigh
10-17-2005, 05:21 PM
Take a look at the offerings in the PC-104 world. This is a huge assortment of Intel-based industrial computers which conform to a standard form factor (about 4 x 4.5 inches). Google on PC-104 and you'll get over 1.5 Million hits http://bbs.homeshopmachinist.net//eek.gif

They have adapters which will accept ISA cards. You could use a single-board PC-104 computer and an interface for the ISA card.

And you can get any processor you want, from an 8085 up http://bbs.homeshopmachinist.net//eek.gif

------------------
Leigh W3NLB

[This message has been edited by Leigh (edited 10-17-2005).]

Leigh
10-17-2005, 05:40 PM
<font face="Verdana, Arial" size="2">Originally posted by D. Thomas:
...wonder what was typical back in 1996 on a mid range Dell PC ?)</font>

Windows '95, of course. Could have been Win3.11, but I doubt it.

------------------
Leigh W3NLB

3 Phase Lightbulb
10-17-2005, 05:46 PM
Windows NT 3.51 was my weapon of choice back then. I tri-booted into NT 3.51, OS/2 Warp, or MS-DOS 6. All the software I made worked on all three.

-Adrian

amcfall
10-17-2005, 06:13 PM
<font face="Verdana, Arial" size="2">Originally posted by 3 Phase Lightbulb:
Windows NT 3.51 was my weapon of choice back then. I tri-booted into NT 3.51, OS/2 Warp, or MS-DOS 6. All the software I made worked on all three.

-Adrian</font>


I was an OS/2 guy myself :-)

There are actually plenty of current motherboards with ISA, you just need to know where to look. I work for an industrial PC company, we sell Pentium 4 based systems with ISA (passive or active backplane). Of course they're pricey, but sometimes it's worth it (like the many customers with those Allen Bradley $1500 ISA cards). If you want a new PC just look us up :-)

Other wise look used, I think I (and many others) have a 1Ghz'ish AMD board with ISA. Depends if you want new or used I guess.

Avery

Milacron of PM
10-17-2005, 06:18 PM
Possible bad news...I'm not at the point where I can hook it up yet, but took a better look and the original (1996) computer was replaced back in 2000 with a Dell Optiplex GX1. Sticker on the front of the case says Pentium III "configured for Widows NT and 98"

So, I say "bad news" because the CPU may in fact be "too new" and already near the limits of ISA slot technology motherboards ??

Looked inside the case and what I assume is main processor chip..on the board where the ISA slots are... is Intel 21152-AB, 1997 date. On the board this board plugs into is a "Parallel Tasking II, Performance" chip..apparently by 3COM, and a Intel PCIset chip...if any of that means anything for this dilemma.

Hard drive is Western Digital 102BA, 10.2 GB "enhanced IDE" type. Year 2000 on that one.

As to RAM..dunno..have to hook it up and see I guess...

Software CD's and documentation have no "requirements" printed on them anywhere. Since this is propriatary Starrett stuff I presume they felt no need to reveal this sort of "secret" as they would handle all that themselves in the CMM maintenance contracts.

Starrett themselves are little help with this sort of thing, as their goal is to claim the whole deal "obsolete" and sell a whole 'nuther "package" for $6,000 and up (even though the latest edition is dated 2002) http://bbs.homeshopmachinist.net//eek.gif

[This message has been edited by D. Thomas (edited 10-17-2005).]

Leigh
10-17-2005, 06:40 PM
<font face="Verdana, Arial" size="2">Originally posted by 3 Phase Lightbulb:
Windows NT 3.51 was my weapon of choice back then. I tri-booted into NT 3.51, OS/2 Warp, or MS-DOS 6. All the software I made worked on all three.</font>

That's because NT was OS/2, and OS/2 was the best multi-tasking RTOS ever designed.

------------------
Leigh W3NLB

Milacron of PM
10-17-2005, 06:42 PM
What about this puppy ?

www.soyousa.com/products/proddesc.php?id=194 (http://www.soyousa.com/products/proddesc.php?id=194)

$175 from Tiger Direct

Bruce Griffing
10-18-2005, 01:08 AM
Don-
The thing to look out for is driver problems. If the software you are running needs a special driver to run the ISA board, you may be hosed. Everything in the new system may work except for that board.

Milacron of PM
10-18-2005, 01:26 AM
Bruce...it has a "dongle" as well...could that further complicate matters ? I'm not even sure where the dongle plugs in yet...

3 Phase Lightbulb
10-18-2005, 01:55 AM
<font face="Verdana, Arial" size="2">Originally posted by Leigh:
That's because NT was OS/2, and OS/2 was the best multi-tasking RTOS ever designed.

</font>

Windows NT is based on the VMS kernel, not OS/2. OS/2 is _NOT_ even remotely an RTOS. OS/2 is just a classic round robin high priority scheduler with non-deterministic interrupt latency. OS/2 had way too many high-level flaws, like the notorious 1 GUI thread shared by all processes/applications. At least NT (Win32) had a seperate message pump thread per process. In 1994/1995 IBM was paying us a lot of money to help bringup OS/2 on IBM's new PowerPC 603 platform right up until they cancled the OS/2 for PowerPC project. IBM was completely lost without Microsoft's help.

-Adrian

ibewgypsie
10-18-2005, 06:40 AM
What are you running? I bought a Pci-BIO board from Access with 48 bits DMA. Has two card connectors on board that tie direct to a opto edgecard cable. And the DCC CMM card does what? if it is 5v ttl signals you are back in business. This replaced the antique Opto ISA card.

You just find the windows assigned address and modify the software. Or reassign the address manually.



------------------
David Cofer, Of:
Tunnel Hill, North Georgia

Milacron of PM
10-18-2005, 10:33 AM
For those that don't already know, a DCC CMM is a CNC Coordinate Measuring Machine (I assume DCC stands for Direct Computer Control...not sure why they don't just say CNC CMM !)

Anyhoo, the ISA card is probably a miniature CNC control/driver/interface for the small motors that move X, Y and Z axis on the CMM. The motors can be small, even though the CMM is quite heavy, as the axis ride on air bearings and motion is relatively slow. I haven't looked, so I'm not sure if the motors are steppers, servo, or "something else" but each axis has Heidenhain glass scales so there must be feedback as well. Travel is excuted via bands rather than screws.

Bruce Griffing
10-18-2005, 01:24 PM
Don-
When you say in your original message that the software upgrades made the machine unusable, what is the specific problem? Speed? Memory? You may be better off upgrading your existing machine.

Wirecutter
10-18-2005, 01:40 PM
<font face="Verdana, Arial" size="2">Originally posted by 3 Phase Lightbulb:
Windows NT is based on the VMS kernel, not OS/2.
-Adrian</font>

Then you have doubtless heard the old "legend" about the "WNT", or Win NT moniker. In the movie 2001, "HAL" was rumored to have come from a one-letter bump of the letters "IBM". Later, the rumor was that "WNT" was a similar one-letter bump of "VMS".

3 Phase Lightbulb
10-18-2005, 02:16 PM
<font face="Verdana, Arial" size="2">Originally posted by Wirecutter:
Then you have doubtless heard the old "legend" about the "WNT", or Win NT moniker. In the movie 2001, "HAL" was rumored to have come from a one-letter bump of the letters "IBM". Later, the rumor was that "WNT" was a similar one-letter bump of "VMS".</font>

Microsoft was going to call the next version of OS/2 "NT OS/2" long before they hired Dave Cutler to design the new NT kernel so "NT" didn't come from Dave, but the association did http://bbs.homeshopmachinist.net//smile.gif

-Adrian

Leigh
10-18-2005, 02:42 PM
<font face="Verdana, Arial" size="2">Originally posted by 3 Phase Lightbulb:
...OS/2 is _NOT_ even remotely an RTOS.</font>

Well, I guess the Washington Post will have to call back a lot of newspapers that were printed under control of OS/2. If you've ever seen a printing plant that puts out a million papers a day, with a 20-foot wide web of newsprint moving at 60 miles per hour, you learn what real-time means.

And I have about 50 other RT projects in operation with customers quite happy with OS/2 in an RT environment. I guess none of them read the book. Amazing what you can do if you don't know it's impossible.

------------------
Leigh W3NLB

[This message has been edited by Leigh (edited 10-18-2005).]

3 Phase Lightbulb
10-18-2005, 03:01 PM
<font face="Verdana, Arial" size="2">Originally posted by Leigh:
Well, I guess the Washington Post will have to call back a lot of newspapers that were printed under control of OS/2. If you've ever seen a printing plant that puts out a million papers a day, with a 20-foot wide web of newsprint moving at 60 miles per hour, you learn what real-time means.</font>

What is real time about that?


<font face="Verdana, Arial" size="2">Originally posted by Leigh:
And I have about 50 other RT projects in operation with customers quite happy with OS/2 in an RT environment. I guess none of them read the book. Amazing what you can do if you don't know it's impossible.
</font>

I'm just curious. What is your definition of real time?

-Adrian

Evan
10-18-2005, 03:56 PM
<font face="Verdana, Arial" size="2">
OS/2 can be used for the softer varieties of real-time tasks. It is not strictly deterministic. According to the bible on the subject, "The Design of OS/2" by Deitel and Kogan (ISBN 0-201-54889-5), "the system guarantees that there is a maximum interrupt disable time of 400 microseconds, and that time-critical threads are dispatched within 6
milliseconds of becoming ready to run." On the other hand, another section in the same book claims the 6ms figure is actually 4ms.... Typical latency times are much less and depending on the configuration the dispatch delay can be higher or lower.
</font>

http://www.ifi.unizh.ch/ailab/embedded/os2rt.txt

Incidentally, I have a copy of OS/2 WARP V3. New in the box.

3 Phase Lightbulb
10-18-2005, 04:42 PM
<font face="Verdana, Arial" size="2">OS/2 can be used for the softer varieties of real-time tasks. It is not strictly deterministic.</font>

Almost all virtual memory based operating systems (OS/2, Windows NT, Linux, Unix, Solaris, AIX, BSD, OpenVMS) can be configured and used in a limited way for soft real-time tasks. However, none of them are fully deterministic and can never replace a hard real time operating system when there are actually real time requirements.

The question is, what is your definition of real time?

-Adrian

Evan
10-18-2005, 04:59 PM
I'm not going to try and define it. It depends on your application. Even without an OS a microprocessor based system is going to exhibit variable latency unless extraordinary measures are used. There is always interrupt jitter to contend with.

ibewgypsie
10-18-2005, 05:15 PM
Let's take it apart to see how it works.

Burr brown made a analog isa board back in the 80s.. I got the job of replacing them without replacing the external cards. Now that was exciting. I ended up going with a Pamux system and specialty software drivers. DMA with analog outputs yeah, there is some challenges, i/o buffers memory and shift gates.

Do you have a picture of the cabling and input output rack?

David

3 Phase Lightbulb
10-18-2005, 06:37 PM
<font face="Verdana, Arial" size="2">Originally posted by Evan:
I'm not going to try and define it. It depends on your application. </font>

Real Time systems are deterministic which means OS/2 can only be used for applications that don't really require a deterministic real time response (I.E: Soft Real Time).


<font face="Verdana, Arial" size="2">Originally posted by Evan:
Even without an OS a microprocessor based system is going to exhibit variable latency unless extraordinary measures are used. There is always interrupt jitter to contend with.</font>

External hardware interrupt latency is always fixed and not variable.

Interrupt jitter has nothing to do with the microprocessor, it has to do with the variable latency seen on a fixed frequency interrupt due to interrupts being masked/disabled by the OS. If you have a clock interrupt firing every 10ms that runs your scheduler, you're not always going to get an interrupt exactly every 10 ms unless interrupts are always enabled/unmasked. In a real system, interrupts are getting enabled/disabled for short periods of time causing jitter (latency) on fixed frequency interrupts.

Interrupt Latency is defined by how long the operating system takes to store the current thread's context and state information before transfering control to an interrupt handler.

Software interrupt latency is defined by all kinds of factors and is where a real time system differs from a non-soft real time system.

-Adrian

Evan
10-18-2005, 06:49 PM
Even without an OS there is interrupt jitter. It is entirely dependent on the CPU. It depends on the intruction that is currently executing and will vary according to the number of clock cycles required to finish execution of the instruction.

Evan
10-18-2005, 07:04 PM
<font face="Verdana, Arial" size="2">
The interrupt latency for a given interrupt is the maximum time from the point the interrupt is asserted to the time the processor begins executing the interrupt's ISR. Many systems have hard real-time constraints on the response to one or more interrupts.</font>


<font face="Verdana, Arial" size="2">
So, you must calculate the latency, adding all the following:

* The longest interrupt execution time of any other single interrupt at that interrupt priority level,
* plus the longest interrupt execution time of every interrupt at any higher-priority levels
* The longest time interrupts at that priority level are disabled
* The longest instruction time (usually a "divide" instruction)
* The time needed to switch context and start the ISR (save the stack pointer and processor state and jump to the starting address of the ISR)
</font>

http://www.embedded.com/showArticle.jhtml?articleID=12803540

3 Phase Lightbulb
10-18-2005, 07:16 PM
<font face="Verdana, Arial" size="2">Originally posted by Evan:
Even without an OS there is interrupt jitter. It is entirely dependent on the CPU. It depends on the intruction that is currently executing and will vary according to the number of clock cycles required to finish execution of the instruction. </font>

This is defined as Interrupt latency, not interrupt jitter. Interrupts are by default disabled while instructions execute atomically. There is nothing that can be done to remove that latency other than use the Wait for Interrupt instruction if one exists.

-Adrian

3 Phase Lightbulb
10-18-2005, 07:42 PM
<font face="Verdana, Arial" size="2">Originally posted by Evan:
http://www.embedded.com/showArticle.jhtml?articleID=12803540

</font>

I don't have to google search to answer questions. I'm trying to imagine what having a technical conversation with you in person would be like. It must be like talking to a mime.

In fact, you might get a better understanding of what Jitter is by reading that article yourself.

-Adrian

Evan
10-18-2005, 07:52 PM
I provide you with links Adrian since you don't seem to be inclined to believe me. Speaking just of the CPU, interrupt latency is the time it takes for the processor to respond to an interrupt. Jitter is the variation in response time.

Incidentally, I wasn't answering a question here.

[This message has been edited by Evan (edited 10-18-2005).]

Leigh
10-18-2005, 08:05 PM
Adrian,

In a real-time system, the important criterion is how long it takes from the time an interrupt is generated until its service routine is COMPLETED. The time before execution is initiated is irrelevant.

This is the point on which many novices trip themselves up. For a system to run reliably and continuously, it must handle all ISRs such that none are lost, nor delayed to the point where their execution is meaningless or even detrimental to the system as a whole.

If you don't understand the real-time implications of running a large printing press, then it's obvious why you don't understand real-time concepts at their basic level.

------------------
Leigh W3NLB

[This message has been edited by Leigh (edited 10-18-2005).]

Leigh
10-18-2005, 08:12 PM
<font face="Verdana, Arial" size="2">Originally posted by 3 Phase Lightbulb:
...However, none of them are fully deterministic and can never replace a hard real time operating system when there are actually real time requirements.</font>

Well, Adrian, it sounds like you're real good at quoting gibberish from the Computer Book of the Month Club.

Have you ever actually done anything, or just read about it? My money goes with the obvious answer.

------------------
Leigh W3NLB

Evan
10-18-2005, 08:25 PM
"The time before execution is initiated is irrelevant."

No it isn't. I once wrote a ML routine that depended critically on the response time to an interrupt as it was doing video display timing for scan lines. Since the time the interrupt would occur was well known I put the cpu into a chain of instructions with the shortest execution time possible just before the expected interrupt and after finishing scan line processing. The interrupt itself was generated based on a count in a register that was clocked by the system clock. The interrupt was generated synchronously with the system clock as the counter register overflowed. This eliminated jitter as the cpu was currently executing a predictable sequence of instructions of all the same length based on the same clock. Latency was held to exactly the execution time of the instruction.

This insured that every scan line started at exactly the same amount of elaspsed time since the last one began.

mochinist
10-18-2005, 08:28 PM
I love it when you computer nerds(&lt;--not a bad thing to be nowaday's, so take no offense to that label) argue, 34 responses and I have no idea who is right or wrong. lol

3 Phase Lightbulb
10-18-2005, 08:44 PM
<font face="Verdana, Arial" size="2">Originally posted by Leigh:
Adrian,

In a real-time system, the important criterion is how long it takes from the time an interrupt is generated until its service routine is COMPLETED. The time before execution is initiated is irrelevant.</font>

In a real-time system, the interrupt service routine does nothing except signal an _event_ and return back to the interrupted task (IF) and only (IF) no other higher priority tasks are ready to run otherwise the interrupt doesn't return to the interrupted task, it returns to the highest runnable task. The _event_ that is signaled by the ISR is usualy blocking a high-priority device driver task so unless a higher priority task was interrupted, the ISR returns to the task that was waiting for that _event_. Interrupts are only disabled long enough to trigger the event. The actuall interrupt handing is being done in a driver task that has interrupts enabled and also has a priority relevant to the hardware it's servicing.


<font face="Verdana, Arial" size="2">Originally posted by Leigh:
This is the point on which many novices trip themselves up. For a system to run reliably and continuously, it must handle all ISRs such that none are lost, nor delayed to the point where their execution is meaningless or even detrimental to the system as a whole.</font>

In a real time system, Interrupts only signal events. Each interrupt usually has a different event that it signals and there are different tasks with relevant task priorities that are waiting(blocked) for those event(s) to get signaled. In a real time system, a device driver's ISR handling code is run in a task with interrupts ENABLED and never runs within the context of an ISR.



<font face="Verdana, Arial" size="2">Originally posted by Leigh:
If you don't understand the real-time implications of running a large printing press, then it's obvious why you don't understand real-time concepts at their basic level.</font>

I don't understand the real-time implications of running a large printing press so that's why I asked "What's real time about it?".

The fact that you think OS/2 is the best RTOS ever tells me that you don't understand real-time concepts.

-Adrian

3 Phase Lightbulb
10-18-2005, 08:54 PM
<font face="Verdana, Arial" size="2">Originally posted by Leigh:
Well, Adrian, it sounds like you're real good at quoting gibberish from the Computer Book of the Month Club.

Have you ever actually done anything, or just read about it? My money goes with the obvious answer.

</font>

Yes, I design and develop hard real time operating systems, device drivers, and cross development tools for a living. What part of a hard real time operating system's __implementation__ would you like to see?

May I see a real time scheduler and interrupt/exception handling code that work you implemented?

You can see any of mine, just ask. Actually, I have an entire real time operating system that runs on X86, PowerPC, MIPS, and SuperH that I wrote from scratch so you're welcome to see any of that as well.

-Adrian

3 Phase Lightbulb
10-18-2005, 09:45 PM
<font face="Verdana, Arial" size="2">Originally posted by mochinist:
I love it when you computer nerds(&lt;--not a bad thing to be nowaday's, so take no offense to that label) argue, 34 responses and I have no idea who is right or wrong. lol</font>

It doesn't look like this "argument" is going to last much longer. We've moved from "I know this and that" to "Show me that you know this and that" and it got quiet all of a sudden http://bbs.homeshopmachinist.net//smile.gif

mochinist
10-18-2005, 10:53 PM
<font face="Verdana, Arial" size="2">Originally posted by 3 Phase Lightbulb:
and it got quiet all of a sudden http://bbs.homeshopmachinist.net//smile.gif

</font>

http://muextension.missouri.edu/explore/wildthing/images/cricket.jpg

Leigh
10-18-2005, 11:25 PM
<font face="Verdana, Arial" size="2">Originally posted by 3 Phase Lightbulb:
...May I see a real time scheduler and interrupt/exception handling code that work you implemented?</font>

The guidance and control systems and launch control systems for the three-stage Pegasus rocket. 100% my code.

I've been writing interrupt handlers, device drivers, communications systems, and such since the 8080 was introduced in 1974. And I've written code for every CPU that's been introduced since then.

------------------
Leigh W3NLB

[This message has been edited by Leigh (edited 10-18-2005).]

cuemaker
10-18-2005, 11:52 PM
Oh yeah?? I built my own computer!! And I have fixed it 3 times!! (someday it will run right)

I could never understand what my dad did, (computer geek since the 50's) even though he tried to explain it to me. I know that has done programing most of his life for JPL, Honeywell, CSC, CSC sold his unit to ?GSM?, he was working in Chantille VA on the new CIA building that was being built. And then congress found out how much it was costing and started investigating. Since ?GSM? was majority owned by a Japanese company and it was a no-no to do govt contract work on the new CIA building and be owned by the Japanese, he was sold/transfered to Infonet rather quickly.

He will retire in Feb after 41yrs or 42yrs of service to the same company (even though he switched a lot) Hasnt changed his office since I was 10yrs old (25yrs ago)

What I really wanted to say was that he once told me that he worked for a year writing a program to figure out the size of the gas tank on the Phoenix missle during the 70's. I sure that he spent most of his time waiting on others to give him info, plus this was like 20+ yrs ago also. So I maybe off.

Ok forget the post. It doesnt compare to a freakin Pegasus rocket, but my dad is the best damn program writer there is, so there.

ibewgypsie
10-19-2005, 12:06 AM
Cuemaker:

can your dad come over and play? I got this allen bradley plc on the bench and this cdrom of all the ascii protocols.. (your volunteered him didn't cha?)

David

cuemaker
10-19-2005, 12:24 AM
I will send him right over. He would have fun helping you out.

He will be moving to Hawaii in Feb to retire. So get his help while you can.

3 Phase Lightbulb
10-19-2005, 03:49 PM
<font face="Verdana, Arial" size="2">Originally posted by Leigh:
The guidance and control systems and launch control systems for the three-stage Pegasus rocket. 100% my code.</font>

Was that a day or two of work, or a full weeks worth of work?


<font face="Verdana, Arial" size="2">Originally posted by Leigh:
I've been writing interrupt handlers, device drivers, communications systems, and such since the 8080 was introduced in 1974. </font>

If that's the case, then I would expect you to at least know what an RTOS is. Your previous statements are quite inept:

"OS/2 was the best multi-tasking RTOS ever designed."

and

"In a real-time system, the important criterion is how long it takes from the time an interrupt is generated until its service routine is COMPLETED. The time before execution is initiated is irrelevant."


<font face="Verdana, Arial" size="2">Originally posted by Leigh:
And I've written code for every CPU that's been introduced since then.</font>

What kind of code have you written for "every CPU that's been introduced"? I've only written code for 32 CPU's:

http://www.gnuxtools.com/targets.html

I've designed/developed an entire hard real time operating system (With fully protected independent process address space, hard real-time task scheduling, multiple filesystem support (FAT12/16/32/ISO9660/UFS), full networking support, etc. My real time operating system currently runs on x86 based platforms, PowerPC based platforms (Sandpoint, Yellowknife, Blackhawk, Artis A3000), Mips (ddb54xx eval boards), 68000 (Arnewsh-sbc68k), SuperH (Bigsur SH4 CPCI), etc. I'm currently working on adding ARM7/9 support for an ARM Evaluator board I have.

Here is a header file that shows all of my kernel level APIs that live in my kernel. User level APIs are implemented on a per-process basis and can implement different trap handling per process making it possible to create a Linux GLIB compatible process and run native Linux binaries at the same time while running a native Solaris binary in another process as well as a native FreeBSD binary in another process.

http://www.gnuxtools.com/prokernel/prokernel.h

-Adrian

Leigh
10-19-2005, 04:02 PM
That's nice, Adrian. I'm sure the natives are very happy.

As to your statement "Was that a day or two of work, or a full weeks worth of work?"

That's obviously nonsense, and only demonstrates how little you know about real programming.

And if you don't understand the concept of ISR completion, it's obvious why your stuff breaks. Mine doesn't.

------------------
Leigh W3NLB

[This message has been edited by Leigh (edited 10-19-2005).]

Leigh
10-19-2005, 07:58 PM
Adrian,

I'm sure you're quite proud of your RTOS, and justifiably so. Such an endeavor is quite an achievement, and you're to be complimented.

We might even have used it on the rocket project, except it wasn't available then, so I had to write my own.

------------------
Leigh W3NLB

[This message has been edited by Leigh (edited 10-20-2005).]