patch that caused regression PR middle-end/8808

[Jan Hubicka]

> On Thu, Dec 19, 2002 at 10:37:53AM +0100, Jan Hubicka wrote:
> > Yes, however I am not at all sure how this can be done.  Fixed register,
> > for instance can be used in the asm("xxx") in rare situations, so I
> > really don't see how one can validate such things.
> 
> We could have another array, along side FIXED and CALL_USED,
> that was UNAVAILABLE.  Errors would be issued for using asm
> registers in this set.
You mean defining new register array based on new target macro and do
testing only when such macro exists?  That sounds like a plan, I can
implement it over the christmas.

Honza
> 
> 
> r~