Hmm, the overwhelming number of responses seem to indicate noone is realy
sure - or in other words: there is no well defined endianess guarantee.

Ok, I can live with that. If we ever change this, we currently need to
fix two drivers in the whole source tree ;-)

For now I'll leave it as is, that's: machfb assumes aperture endianess to be
identical to host endianess. In the mach case all other scenarios are easy to
cope with, since the card can work with both endianesses - it's all a matter
of deciding which way to use. If we define a bus_space_vaddr_endianess macro
to test bus_space_tag_t or bus_space_handle_t's for their endianess, or
define that to always be bus-endianess - fine, we can adjust the driver then.

Martin