mn10300 relaxing bug

[DJ Delorie]

> For something like this, I would be more happy with it not introducing 
> any regressions into the G++ testsuite for the mn10300...

Tested, none.

> Hmm - is this going to make the mn10300 port very slow at linking ?

It's O(n) so it shouldn't be too bad.  If we wanted to hash it, we'd
have to sprintf the address to make a key.  That would also be O(n)
and likely as complex as what I've already got.

> Anyway the patch looks OK to me, so if you are happy with it please feel 
> free to apply it.

Done.