XTAG2 firmware update?

Technical discussions related to any XMOS development kit or reference design. Eg XK-1A, sliceKIT, etc.
User avatar
trousers
Active Member
Posts: 44
Joined: Fri Dec 11, 2009 10:20 am
Contact:

Post by trousers »

Folknology wrote:The point here is that we don't just program them and ship, they are actually custom programmed in situ. All of the designs in the family are like this, it is a key feature and emerging market.
If all you need to do is program them from a host then a simple OTP or SPI flash based boot loader would surely be simpler, easier and cheaper. Why do you need an XTAG2 on board at deployment time?


Best friends with the code fairy.
User avatar
Folknology
XCore Legend
Posts: 1274
Joined: Thu Dec 10, 2009 10:20 pm
Contact:

Post by Folknology »

Some of it is compiled in situ based on code we don't have.

(Actually tested and debugged insitu in many cases)
User avatar
trousers
Active Member
Posts: 44
Joined: Fri Dec 11, 2009 10:20 am
Contact:

Post by trousers »

Folknology wrote:Some of it is compiled in situ based on code we don't have.
Then surely you need either the compilation suite on your host (which is rather beyond what you were originally asking for), or you need your own compiler. In either case you can still use a simple boot custom boot loader. The above comment doesn't explain, at least not in language I understand, why you need an XTAG-2 in the deployed system.
Best friends with the code fairy.
User avatar
Folknology
XCore Legend
Posts: 1274
Joined: Thu Dec 10, 2009 10:20 pm
Contact:

Post by Folknology »

How can it be debugged without xtag2?
User avatar
trousers
Active Member
Posts: 44
Joined: Fri Dec 11, 2009 10:20 am
Contact:

Post by trousers »

Folknology wrote:Some of it is compiled in situ based on code we don't have.

(Actually tested and debugged insitu in many cases)
You added that after I originally replied!

An application which includes testing and debugging hard/software sounds like an unusual billion-chip opportunity. This doesn't involve mobile software agents does it?
Best friends with the code fairy.
User avatar
trousers
Active Member
Posts: 44
Joined: Fri Dec 11, 2009 10:20 am
Contact:

Post by trousers »

I think we're getting messages out of sync here.
Folknology wrote:How can it be debugged without xtag2?
That depends on exactly what boundaries you place on what programs can be downloaded and run, which is to say, it depends on your application. As I don't know what you're trying to do I can only take your word that you need an xtag2 and therefore, tautalogically, you do need an xtag2.

There's been quite alot of research done on checking binaries for safeness. If you have a safe program then you can implement co-opertive debugging in the same you one might implement co-operative context-switching.
Best friends with the code fairy.
User avatar
Folknology
XCore Legend
Posts: 1274
Joined: Thu Dec 10, 2009 10:20 pm
Contact:

Post by Folknology »

Actually its not difficult to imagine

There is one important principle that makes this whole thing work.

We cannot know in advance the exact usage scenario in which the shipped products will operate.

This is a market based on that very opportunity, it fulfills the space of something that can't be fulfilled by pre-programmed means. All we can do is make the best product we can for those markets. here is what has been said about it:
"We are selling a creative product, not a finished product. We are selling possibilities. The participant decides what runs on it, not us. We just give them as much as we can to help them through that process."
Its not a big mystery, I have no idea how big it is, its to diverse to contemplate.


Actually for an analogy I feel like Apple or Microsoft must have felt when talking to IBM about PCs, IBM probably kept trying to re-frame the idea in terms of Mainframes.

This isn't just re-programming either, it can actually be self programming at a higher level also.

Its exciting, I don't see why Xmos aren't all over it, they and this community are in an ideal position to take the high ground with XS1..
User avatar
Folknology
XCore Legend
Posts: 1274
Joined: Thu Dec 10, 2009 10:20 pm
Contact:

Post by Folknology »

Anyhow I actually have to do some work today :cry:

Already spent enough time on this thread, the requirements are clear.

I look forward to Xtag2 code being open sourced like its design. :D

If that doesn't come then I will have to go with the alternative FPGA+MCU design

regards
Al
User avatar
lilltroll
XCore Expert
Posts: 956
Joined: Fri Dec 11, 2009 3:53 am
Location: Sweden, Eskilstuna

Post by lilltroll »

It's (at least) something I do not understand here. As I understand it, the XTAG2 or any other solution is limited to USB2.0 wince it has a USB2.0 hardware transciever. So the day USB2.0 is an too old standard the solutions we are talking about here is out of date anyway.

The XTAG2 uses a bootloader written in OTP.
What is the difference to "include XTAG2" and to include "USB bootloading" in a consumer product or any product - isn't it the same thing on the XMOS side of the USB cable ?

By the way - how do you burn the OTP in the lab, and how do you burn it in a large production ?
(Microchip offers to pre-program your chip, for a fee of course)
Probably not the most confused programmer anymore on the XCORE forum.
User avatar
paul
XCore Addict
Posts: 169
Joined: Fri Jan 08, 2010 12:13 am
Contact:

Post by paul »

lilltroll wrote: By the way - how do you burn the OTP in the lab, and how do you burn it in a large production ?
(Microchip offers to pre-program your chip, for a fee of course)
There is a tool called xburn (use it very carefully) to do it in the lab.

For large production XMOS have the ability to provide pre-programmed devices.
Paul

On two occasions I have been asked, 'Pray, Mr. Babbage, if you put into the machine wrong figures, will the right answers come out?' I am not able rightly to apprehend the kind of confusion of ideas that could provoke such a question.
Post Reply