[PATH] fix PR 13087, documentation does not follow IMA/IMI (all at once compiling)

[Andrew Pinski]

> I didn't try understanding all of the logic in the code.  I got hung 
> up on the items above.  Mostly it seems pretty reasonable, but I'm 
> concerned about how many options need so much special handling.  Maybe 
> there is a better way to do this?  Maybe we shouldn't be risking 
> gcc-3.4 breakage to get this working?  I'll have to take another look 
> at this at some point.

I think we should defer this until after 3.4 branches, I will submit a 
doc fix for 3.4 then.
This is really needed for the tree-ssa as we will then able to do IPA 
for functions that are
in different translation units (really for SPEC compiling) without 
getting into trouble with
RMS.

Thanks,
Andrew Pinski
apinski@apple.com