Convert the double-precision floating point value x to a signed integer value in round-down (to negative infinity) mode.
x