[patch] Fix the failures created by fix for PR 27144

[Andrew Pinski]

> 
> Hello,
> 
> > Is there a strong reason to have unsigned_type_for and signed_type_for,
> > when lang_hooks.types.unsigned_type and lang_hooks.types.signed_type
> > would appear to be sufficient?
> 
> They do not handle pointers correctly.  Also, I am too lazy to type
> ugly long names like these :-)

How are they ugly, they tell what is going on instead of how
unsigned_type_for and signed_type_for where you have to
look into a different file to figure out what is going on.

Also the comment on those two functions could be exanded to be
like:

Returns the unsigned variant of TYPE that the language requires.

-- Pinski