[M32C] : No error is reported when relocation gets overflowed.

[DJ Delorie]

> I guess, the above behavior is because of the patch provided in link
> above.	
> 
> It would be better to complain for overflowed relocations instead of 	
> generating wrong code, which finally does not work.	
> 
> Also, I would like to know, if this is done in order to fix or get some
> workaround for 24 bit pointer issue in GCC for M16C targets.		

The problem is that the m32c chips don't have signed offsets against
the address registers, so to simulate a signed offset, gcc uses a huge
offset and lets it wrap.

I guess the solution is that we can't use any of the built-in overflow
detectors, and we'll have to do it ourselves.