> I never got down to facts about whether it was an issue of the PPC
> supporting it in HW or the PPC developers just mis-designed their
> framework and now require big endian friendly devices just to work.
PPC has byte-swapped load and store instructions, and the
in*/out*/read*/write* family of macros use them. It doesn't have the
endianness selectable on a per-page basis like the ultra does, where
you can do a normal load/store to a little-endian page and the
hardware byte-swaps it for you (if I've understood it correctly).
So if you have a frame buffer with multi-byte pixels, you can either
use a big-endian aperture if there is one, or have versions of all
your framebuffer access routines which use the byte-swapped loads and
stores instead of normal loads and stores. That would be possible
(and would even be just as fast) but it would be a little ugly.
Paul.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/