Richard A. Smith
richard at laptop.org
Tue Apr 3 22:11:44 CEST 2007
>> Is the EEPROM file correct?
> The linux based eeprom flashing process is somehow flawed.
> The windows based method as recommended by FTDI using their MProg works,
> though :(
> We once had somebody volunteering to look at this problem, they even
> hada USB protocol analyzer. Unfortuantely no result so far.
That was me..
Looking at bugzilla I see that you fixed the string writes. Do I still
need to look at this?
Getting the windows tool to work is a real pita. It only works if the
ID's are default. So once you program it you can't re-program it. But
it appears to fail silently.
At least thats how it looks from the USB traces I grabbed. I programmed
the device and then re-programmed it with the analyzer enabled. The
logs are filled with single byte setup IN transactions where I see the
characters one at a time aka 'M' 'o' 'k' 'o'. So I suspect thats the
reads I'm seeing and not the writing method.
At the time I was doing the grabs the owner of the windows machine was
trying to get useful work done so I was not able to re-do the tests over
and over. Its a bit time consuming since I have to put the unit back in
a linuxbox and reprogram it back to the default ids.
I did not mess with it more since it appeared un-necessary. The linux
tool write the ID's fine (for me at least) and openOCD and ftdi_sio only
care about the IDs. Not the strings.
But If you still want more info I'll certainly get everything all hooked
up and do another grab.
Richard Smith <richard at laptop.org>
One Laptop Per Child
More information about the neo1973-hardware