ARM: allow factorization of constants into addressing insns when optimizing for space

[Richard Earnshaw]

> > This is hardly significant enough to get upset about, but I think I'm 
> > going to revert the code to using a limit of 1 for two reasons:
> > 
> > 1) We now have two testcases that show an increase overall, versus one 
> > that shows a decrease.
> > 2) When using a limit of 2 code degradation for pathological cases seems 
> > to be more severe than the degradation when the limit is 1.
> 
> Agreed.
> 
> What about a -mconstant_limit switch?  This would allow for overriding the
> default when it is known to be beneficial.
> 

My inclination would be no.  It adds yet another flag that we have to 
manage and document, and it's not clear what would have to be done with it 
if we changed the heuristics in the compiler (for example, if the constant 
limit were to change depending on the phase that the compiler was in).

That being said, if you want to submit a patch, and it is correctly 
documented and implemented, I'm not going to turn it down outright.

R.