wiznet - USB device not Recognised

[espruino-discuss3]

Posted at 2014-11-26 by DaveNI

I've received my wiz550io so decided to install the latest wiznet espruino binary:

http://www.espruino.com/binaries/espruino_1v71_espruino_1r3_wiznet.bin

After installing my board is no-longer recognised as a USB device!

I've gone back to the latest standard binary and all is fine but I would like to get going with working on an http server!

Any ideas?

[espruino-discuss3]

Posted at 2014-11-26 by @gfwilliams

I'll try and have a look at this tomorrow - it could just be a build issue. I know I switched to a space-saving compile and the builds have been unreliable in the past.

Maybe try the 1v70 build?

---

Posted at 2014-11-26 by DrAzzy

Hm.

I've tried recent github wiznet builds, and they work without issue, as do post v71 wiznet bigram builds. So if there is a problem with this, it's a narrow window of effected builds...

---

Posted at 2014-11-27 by @gfwilliams

@drazzy - yes, it would be. I think there's some issue with -Os on the compiles - and it's a nightmare to track down, because the second you change the build in any way the problem goes away.

---

Posted at 2014-11-28 by DaveNI

Just downloaded the latest Wiznet build and it worked perfectly!

It was maybe just a build issue but if you made changes to fix - thanks.

---

Posted at 2014-11-28 by DrAzzy

Actually, maybe this is why some of my v71 wiznet bigram builds weren't working.

I thought it was due to my not having incorporated the changes to EspruinoBoard.py* in commit 91164f259c40d64b93b764ae44b488a57a277afe (which I don't understand the purpose of), since after taking those changes in, the builds started working.

*the way my script works is it clones the repo, then copies in the modified EspruinoBoard.py, and then builds. So I need to manually update it if anything is changed in EspruinoBoard.py

---

Posted at 2014-12-01 by @gfwilliams

I did make some changes in Git, so they should be in the latest builds - glad it's working.

It appears I might have hit a bug in GCC. Sometimes it was randomly choosing to re-order accesses to the USB hardware registers (even though they were volatile). I posted on StackOverflow about it, but nobody seems to have a clue :)