nightmares with biased stack pointers

[Jeffrey A Law]

  In message < 341B044D.D30D9594@execpc.com >you write:
  > Richard Henderson wrote:
  > > 
  > > > 3) Hey look, taking into account the known alignment of the stack
  > > >    frame and the "add sp, const, reg; srlx 4; sllx 4;" sequence,
  > > >    I can determine that this is really an alignment of the 'constant'
  > > >    so I can merge this into a constant operation and blow the two
  > > >    shifts away.
  > 
  > 	I'm new to this list and this was the first message I got, so I'm not
  > quite sure exactly what it is you're talking about... 
Sounds to me like he's describing an optimization in combine,
which is screwing this particular port.

  > But if you want to
  > get rid of the two shifts, why don't you just use a logical and?
Well, internally the "and" will get turned into shifts during
the combine phase so that combine can optimize them better.

The problem is in this case combine over-optimizes stuff.

Jeff