https://gcc.gnu.org/bugzilla/show_bug.cgi?id=119386
--- Comment #14 from Michael Matz <matz at gcc dot gnu.org> --- (In reply to H.J. Lu from comment #13) > (In reply to Michael Matz from comment #11) > > > access to the respective GOT slot). Upstream binutils now silently do emit > > a > > route via PLT, our binutils complain. I'm not sure that upstream behaviour > > is > > indeed better. > > PLT32 relocation is used by ld to distinguish call/jump vs other PC relative > accesses. Linker allocates a PLT slot only if it is needed. If it isn't the > case, please open a binutils bug report. I'm aware of that, and that's fine. I don't think there's a binutils bug per se here. Older binutils simply always used a PC32 reloc for naked symbol refs, and newer binutils use a PLT32 reloc when the context of the naked reference is a call instruction. (Older binutils then sometimes would have to allocate a PLT slot even when not strictly necessary). If anything, then binutils could become even _more_ strict, and simply not invent any specific PIC-related relocation at all for undecorated symbols, but that discussion is for another time and a different bug tracker :-) In this specific case the bug lies within GCC in emitting a naked symbol ref under -fPIC.
