https://gcc.gnu.org/bugzilla/show_bug.cgi?id=117766
--- Comment #8 from Andrew Pinski <pinskia at gcc dot gnu.org> --- Also the ARM ARM says the following here about fp/simd on A32/T32: E1.3.10 Implications of not including Advanced SIMD and floating-point support "Exceptionally, for implementation targeting specialized markets, Arm might produce or license an Armv8-A implementation that does not provide any support for Advanced SIMD and floating-point instructions. In such an implementation, in AArch32 state:" ....
