I've posted some fixes to address the endianness problem in  OHCI (when 
used with an OHCI controller whose native endianness is not little, such 
as found on the Alchemy parts).  I'd be very grateful if someone with 
OHCI knoweldge could review my patch.

The patch has been posted to port-evbmips/31912, because it also 
includes changes to the Alchemy OHCI driver, but by far most of the 
change is in the machine-independent OHCI code.  (Perhaps this should 
have been filed as two PRs, one for the MI portion and one for the MD 
portion?  I'm not sure about the preferred process here.)

Thank-you very much.

-- 
Garrett D'Amore                          http://www.tadpolecomputer.com/
Sr. Staff Engineer          Extending the Power of 64-bit UNIX Computing
Tadpole Computer, Inc.                             Phone: (951) 325-2134