Amaury Pouly wrote:
>
>... seems to indicate that the problem lies in the
> initialization code but we have been unable to fix it.

Just a quick check - in reset_endpoints:

| /* Setup next chain for IN eps */
| FOR_EACH_IN_EP_AND_EP0(i, ep)
| {
| int next_ep = in_ep_list[(i + 2) % (NUM_IN_EP + 1)];
| DIEPCTL(ep) = (DIEPCTL(ep) & ~bitm(DEPCTL, nextep)) | (next_ep << DEPCTL_nextep_bitp);
| }

Is the "i + 2" correct?
Received on 2010-10-10