Swap optimization pass

[Jeffrey A Law]

  In message < l03130300b0df4e557e96@[206.119.242.170] >you write:
  > At 10:19 PM -0500 1/11/98, John Wehle wrote:
  > 
  >  * Hello,
  >  *
  >  * Here's an initial implementation of a swap optimization pass.  The
  > 
  > Well, I'm sure it finds a lot of things this misses, (and possibly
  > avoids register pressure problems), but I'll bet that putting
  > something like this into your MD file would get 90% there and is
  > way faster.....
Or even better, make it a real pattern instead of a define_peephole.

That will allow combine to create it, and combine will handle making
sure the right operands are dead (note, it's not really safe to call
dead_or_set_p in a define_peephole, regardless of what the docs might
say -- I can explain in more detail in another thread if you're interested).

jeff