https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114065
--- Comment #69 from Eric Botcazou <ebotcazou at gcc dot gnu.org> --- So the current implementation is in fact incorrect for all 32-bit architectures since the size of timespec is wrong, although this is (apparently) harmless on little-endian architectures? And the kernel implementation of the problematic tv_nsec is
