Sat Feb 12 10:31:20 EST 2011

cell123 on AT91

I'm trying to run the cel123 test program on the AT91.  It seems to
compile fine, but there is something wrong with the IO.

Q: When using the main() start point in eCos as opposed to
   cyg_user_start(), does the scheduler get started to enable blocking

It seems main is a thread[1].  In retrospect it's apparent from the
stack trace as there is no caller of main() in the backtrace.

Anyway, somewhere in this i->o data gets lost:

    while(1) {

Ok, it is using the debug input port, but it doesn't get any data, so
I suspect something else is also reading from this port.

Ok, changing main -> cyg_user_start gives me a working console.  That
might be the trick.

[1] http://sourceware.org/ml/ecos-discuss/2008-11/msg00034.html