On Fri, Mar 07, 2003 at 03:12:17PM -0500, Dan Malek wrote:
> Jun Sun wrote:
>
> > More than likely this is due to the interrupt routing for USB controller
> > not being setup correctly.
>
> How did you come to this conclusion? Is this a PCI USB controller or the
> on-chip peripheral? I have Au1xxx boards were on-chip usb is
> required and is working fine. There aren't any options to configuring
> on-chip USB interrupts on Au1xxx. It could just as likely be a Linux kernel
> configuration problem. We know there is something amiss with Au1xxx USB
> in big endian mode, but all LE boards should work fine.
>
I have seen USB working under BE mode in many instances. (Acutally
that I have two that work.). I don't think endianess is the issue.
I have seen "number assign failures" a couple of times before, and it
is all because interrupt not routed correctedly. That is where "more than
likely" comes from.
Of course, I have only dealt with PCI USB controllers. On-chip ones may
have another set of additional issues that I am not aware of.
Jun
|