nightmares with biased stack pointers

[Richard Henderson]

> 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 presume you tried subtracting out the bias before applying the 
const alignment logic?  I don't know where you found the code that
was doing this so I don't know what else it would impact.

Failing that, I suppose you could try to use elimination from a 
"soft" stack pointer to the hard stack pointer.  I don't know
how many places would then have to be taught to treat manipulations
of HARD_STACK_POINTER_REGNUM specially like they sometimes do
for the current stack pointer.


r~