Swap optimization pass

[Michael P. Hayes]

Jeffrey A Law writes:
 > 
 > Yup -- this is the basic idea behind my idea of using reorg to combine
 > unrelated insns after reload.

I would prefer this sort of pass to run before register allocation
otherwise the chances of combining these insns gets slim since the
register constraints are usually tighter on multi-pack insns.


 > Note that introducing "foreign" insns into a libcall block tends to
 > do bad things in the register allocator, and I guess this might happen
 > in that case.

Well we could restrict these "foreign" insns to be moves.  I'm finding
that the incoming register arguments get poorly copied to the
corresponding arguments of a nested call.

Michael.