float64_to_int64_round_to_zero
#define float64_to_int64_round_to_zero __fixdfdi
int64 float64_to_int64_round_to_zero( float64 a )