[llvm-commits] [patch][RFC] Correctly produce R_X86_64_32 or R_X86_64_32S

[Chris Lattner]

On Sep 23, 2010, at 10:43 AM, Rafael Espindola wrote:

>> fwiw - I havent observed any ill effects with this patch + it fixes
>> all signed/unsigned reloc mismatches I was having when linking
>> FreeBSD kenrel.
> 
> Nice!
> 
> An update patch against current trunk is attached. I also expanded the
> included test.

Hi Rafael,

How is reloc_signed_4byte different than FK_Data_4?  Assuming that there is a good reason, please improve the documentation in X86FixupKinds.h for it, and commit.  Thanks!

-Chris