Home Original page

GOLD handling of weak symbols (including x86 vs. ARM)

Cary Coutant ccoutant@google.com
Fri Nov 5 18:47:00 GMT 2010 
> In other words, FUNCTION_CALL isn't used to mean a relocation which is a
> function call.  It means a symbol which is a function.  It should
> probably be renamed to IS_FUNCTION.  Or just omitted, and
> needs_dynamic_reloc can check the symbol type itself.

It actually was intended to mean that it's a function call, but in the
case of i386 and x86_64, it's only a best guess. We don't always set
FUNCTION_CALL when the symbol type is STT_FUNC -- only when the
relocation type also suggests that it might be a function call.

> It seems to me that the distinguishing characteristic of "foo@PLT" is
> not that it is a function call, it is that the relocation specifically
> requires the PLT.  I suspect that the right thing to do here is to
> always use the use the PLT offset, if there is one, for R_386_PLT32.

Agreed. If the relocation type forces the use of a PLT, we shouldn't
even call use_plt_offset(). Same for R_X86_64_PLT32 and (I think)
R_X86_64_PLTOFF64.

-cary
More information about the Binutils mailing list