https://gcc.gnu.org/bugzilla/show_bug.cgi?id=119408
--- Comment #18 from Jakub Jelinek <jakub at gcc dot gnu.org> --- _Float128 is definitely not for backward compatibility, it is a standard C23 type and in C++ used for the standard std::float128_t type. I think aarch64 doesn't build libquadmath either, but maybe it only supports long double and _Float128/f128 and not __float128/Q.
