Re: [cond-optab] Convert cris

[Hans-Peter Nilsson <hp at bitrange dot com>]

On Thu, 26 Mar 2009, Paolo Bonzini wrote:
> No, I would have had to get away from "@ in *cmpsi because of
>
>   if (which_alternative == 0 && TARGET_V32)
>     return "cmpq 0,%0";
>   return "test.d %0";
>
> and didn't really like to.

But that just counts an assembly code difference. :)
Oh wait, I think I'm using the non-carry-modification.

(I'll look into the folding.  Don't want multiple patterns where
only match_operand operands differ.  The pre-v32 targets may
take a hit for maintenance.)

I forgot to say: after fixing the things I mentioned, it's ok.
Again thanks.

brgds, H-P