Code gen error on PowerPC

[Franz Sirl]

At 00:35 11.03.99 , Richard Henderson wrote:
>On Wed, Mar 10, 1999 at 10:54:56PM +0100, Franz Sirl wrote:
>> Insn 71: dest address src reg 81 benefit 7 used 1 lifetime 1 replaceable 
>> mult
>> 8 add (plus:SI (plus:SI (reg:SI 88)         (const_int -8))
>>     (const_int 4))
>>
>> Is this valid at this point? Shouldn't there be a single (plus:SI 
>> (reg:SI 88)
>> (const_int -4))?
>
>It's certainly not good, but it isn't really evil either, since that
>non-canonical expression should only inhibit proper giv reduction.

yeah, it seems that this is harmless and not related to this problem.

This damned bug keeps slipping thru my fingers during debugging. I _think_ 
that it's the changed cost analysis in Richard's loop optimization that 
keeps the bug from happening, but currently I'm lost in debugging...

A few more hints please... :-)

Franz.