https://gcc.gnu.org/bugzilla/show_bug.cgi?id=119408
--- Comment #22 from Jakub Jelinek <jakub at gcc dot gnu.org> --- I think for libgfortran the cleanest would be in the configure check whether long double is IEEE quad and if so, have libgfor_cv_have_float128 no. That can be e.g. #if __LDBL_MANT_DIG__ == __FLT128_MANT_DIG__ && __LDBL_MIN_EXP__ == __FLT128_MIN_EXP__ #error long double is IEEE quad, no need for _Float128/__float128 #endif
