This patch reorganizes and cleans up the FPU code. Main changes include:

1. Related functions are re-arranged together.
2. FPU save/restore code are extracted out as independent functions 
(fsave/frstor, fxsave/fxrstor, xsave/xrstor).
3. Two FPU entry functions, fpu_save() and fpu_restore(), are defined. 
They utilize xsave/xrstor to save/restore FPU state if CPU supports 
extended state. Otherwise they fall back to fxsave/fxrstor or fsave/frstor.

Signed-off-by: Wei Huang <wei.huang2@amd.com>