branches: 1.1.4; 1.1.8;Add __floatunsi{sd}f for IEEE754.Unlike the quad versions, these don't use floating point.