Atomic operations updates: add cmpxchg_local

Hi Andrew,

I have been asked for a cmpxchg_local implementation by Christoph Lameter so he
can use it to accelerate the slub allocator.

On some architectures, where cmpxchg is nothing more that a simple save/restore
interrupts surrounding the cmpxchg non atomic ops, I remove the code and make
them use the new "generic" version.

cmpxchg_local and cmpxchg64_local are made available on all architectures in
these patches. They use the best primitive known, either an atomic operation or
save/restore interrupts.

Mathieu

-- 
Mathieu Desnoyers
Computer Engineering Ph.D. Student, Ecole Polytechnique de Montreal
OpenPGP key fingerprint: 8CD5 52C3 8E3C 4140 715F  BA06 3F25 A8FE 3BAE 9A68