Sure. That is in fact my situation So the macros that I use to wrap these accesses both byte swap and fetch data byte at a time. I was trying to save the original writer some effort since his Alpha processor does support unaligned accesses. Ales Fiala Miles Bader wrote:
bugzilla-daemon@freedesktop.org writes:
Most, if not all, of these unaligned accesses are in the AtomBIOS. This is just initialization code that is not performance critical. So if you can put up with the polluted kernel logs, I don't think it is much of a problem.
What about processors that don't support unaligned accesses? If it's not performance critical then it should be fine to just do byte-by-byte r/w...
-Miles
-- To unsubscribe, e-mail: radeonhd+unsubscribe@opensuse.org For additional commands, e-mail: radeonhd+help@opensuse.org