I have received permission from STMicroelectronics to make the VM Labs MPE processor core available as open source under the MIT license. This is a core that was used in the VM Labs NUON DVD chip and has shipped in numerous products. It was designed for implementation on a SoC though and as far as I know has never been ported to an FPGA. I'd like to post the Verilog for this core to OpenCores but I don't have an FPGA example showing how to use it. Is it acceptable to post just the core without a corresponding testbench? As I said, I know that this core works because it has been used in commercial products. I just don't have an example of it working on an FPGA. Alternatively, would someone here be interested in helping to port this core to an FPGA?

By the way, one of the interesting features of the MPE is that there is a port of GCC available for it. I've attached a document describing the MPE processor core as well as the VM Labs SoC that used it.

Thanks,
David Betz

Go ahead and create a project under processors and post what you have. You can always add to it later.
Be sure that all the licensing and boilerplate is in order. You say that you "have permission" then I would also include a copy of that as well.

If you have a software tool chain then you should also include that. A working fpga example is nice but it is more important to include a basic simulation self checking test suite that exercises as muuch as possile.

Documention is always welcome.

John Eaton

Not meaning to be a stick in the mud here, but the reference document (.pdf file) that you attached has "Confidential and Proprietary" markings stamped all over it... This seems a risky business to be involving OpenCores in... If you have permission to post the Verilog sources, then you should probably also get 'hold of original format documentation so that it can be legitimately redacted prior to posting here... The last thing OpenCores needs is a lawsuit... Apologies for perhaps being overly careful...

Go ahead, sometimes a little caution is in order. The document he posted covers a complete SOC using the STM processor and perhaps should not have been shared. I was some what surprised that STM would give permission to an outside party to opensource one of their products but given that it looks at least 20 years old then it is possible that they are retiring the product and did ok the transfer.

But I would like to see the letter from STM showing what they think that this means and how high up the signing manager is.


John Eaton

Hi David,

This is really cool and I hope that we can see more of these kind of news in the future.

I know personally that it can be really hard to get companies to open source their products even if they have been obsolete for a decade, so good job getting the permissions to do this

Regards,
Olof Kindgren

I actually have permission to release the source for the entire SoC. Also, I'm not a third party. I work for STMicroelectronics. I have an email exchange giving me permission to release all of this. Is that sufficient or do I need to get ST to write a letter saying this? Also, I'm pretty sure that the original source for the document that I posted is no longer available so there is no way to create a version of it without the proprietary markings. How is something like that typically handled?

I have the source code for the toolchain but I haven't tried building it in a long time. Also, I have nothing that can be used to verify the RTL, no test suite, nothing. I believe that some of that may be included in the source tree but it probably uses expensive software to which I have no access. Essentially, this code is availabe as-is. If you want to use it you'll have to create your own test environment. This is what I was worried about. Does this make it essentially useless? I'd like to find someone who has an interest in this and is proficient at FPGAs who could help create a test environment. While I have an interest in FPGA programming, I have no professional experience in that area and very little personal experience either.

I should point out that I am not representing OpenCores here... I'm nothing more than a concerned user... I'm also quite excited that you're posting this... My suggestion... Pull down the .pdf as it's posted now, and pre-pend a few pages to the front of the pdf that include the permission documentation and an explanation that the proprietary markings are no longer valid... (along with licensing info)

I think posting is far from worthless... Others may disagree... Please post it... It'll either find it's way to maturity or not... If you don't post it, I'll guarantee "not"... ;-)

My fear is that STM is like most large companies and sometimes it is hard to find all the right people to make a decision like this. Or if you do find them then they will either forget or be replaced by some one else who has no idea of the history.

If somebody in STM legal is told in two years that their IP is on opencores.org will they look at your email and say thats OK or will they say that email sender did not have the authority to make that decision and start trouble?


John Eaton

If the software is released under an opensource license then you should also include it. Releasing a cpu with no assembler or c compiler is pretty worthless. I used to hand assemble code for a M6502 but I would never consider that for a VLIW processor.

John Eaton

Someone here suggested that it would be safer if I requested a letter from ST releasing this core under an open source license. I asked for such a letter and have now been told that it will not be possible to get one. I guess the person who gave me permission through email was just hoping no one else would notice that the source had been distributed.

I tried to figure out how to delete this entire thread or at least to delete the attachment to the first message but that doesn't seem to be possible with this forum software. Could the moderator please delete the attachment?

Sorry for taking everyone on a wild goose chase. It looks like the MPE processor core will remain proprietary and probably never be seen again since ST has no plans to use it any longer.

ST does not understand Open Source :
http://www.eetimes.com/document.asp?doc_id=1153778

:-( :-(

Could one of the moderators please delete the attachment to the first message in this thread?

Thanks!

That's sad but at least you gave it a good try and you won't have to worry about being walked out the door in a few years if the S%$t hits the fan.

When dealing with a corporation you should always remember that if any human being were to act with the same ethics and morales as any corp then they would be classified as a psychopath. If they do someting nice then you have to ask why.

It is really frustrating for all the designers who put a major chunk of their life into creating something only to watch it abandoned and left to die. Actually thats not right. If you leave something to die then at least someone else can find it and save it. Nope, they are putting the kittens in a sack and throwing it off the bridge. If they don't want it then nobody gets it.

Maybe wikileaks will take it.

John Eaton