Why is targetm.fixed_condition_code_regs a function?

[Ian Lance Taylor]

Richard Earnshaw <rearnsha@arm.com> writes:

> More importantly, why is it a function that returns one or two registers.  
> If it has to mess around with pointers, why doesn't it return an array 
> (terminated by INVALID_REGNUM) of condition registers.  In fact, the macro 
> could point directly at the array without the need for a function call (a 
> -m switch would simply change the array pointed to).

I returned two registers merely because the ix86 has two, but every
other target has one, and I thought it was quite unlikely that any
other target would ever have more than one.  The code is more
efficient when there is a known limitation on the number of condition
code registers, since otherwise there have to be some loops.  Since
those loops would be to a variable bound which would always in
practice be 1 or 2, I thought it was more efficient to just impose the
requirement.

Note that targets like MIPS4 with an array of condition code registers
wouldn't use this code.  This code is for cases where a fixed
condition code register is used directly in RTL.  The MIPS4 uses a
pseudo-register which is assigned to a hard condition code register in
the usual manner for register allocation.

Ian