math: use 0x1p-120f and 0x1p120f for tiny and huge values

[musl] / src / stdlib / llabs.c

long long llabs(long long a)
{
        return a>0 ? a : -a;
}