Applying: x86: Do not use __builtin_fpclassify for _Float64x/long double
Applying: x86: Do not use __builtin_isinf_sign for _Float64x/long double
Applying: stdio: Fix tst-vfprintf-user-type on clang
Applying: math: Use math_opt_barrier on ldbl-128 powl underflow/overflow handling
Applying: misc: Enable tst-atomic for clang