BigInteger中destructiveMulAdd方法疑问

[_冬木]

其中第15行和第17行分别取低32位和高32位是什么意思？

private static void destructiveMulAdd(int[] x, int y, int z) {
        // Perform the multiplication word by word
    	//将y与z转换为long类型
        long ylong = y & LONG_MASK;
        long zlong = z & LONG_MASK;
        int len = x.length;

        long product = 0;
        long carry = 0;
        //从低位到高位分别与y相乘,每次都加上之前的进位,和传统乘法一模一样.
        for (int i = len-1; i >= 0; i--) {
        	//每次相乘时将x[i]转换为long,这样其32位数就可转变为其真正代表的数
            product = ylong * (x[i] & LONG_MASK) + carry;
            //x[i]取乘积的低32位.
            x[i] = (int)product;
            //高32位为进位数,留到下次循环相加
            carry = product >>> 32;
        }

        // Perform the addition
        //执行加z
        //mag最低位转换为long后与z相加
        long sum = (x[len-1] & LONG_MASK) + zlong;
        //mag最低位保留相加结果的低32位.
        x[len-1] = (int)sum;
        //高32位当成进位数
        carry = sum >>> 32;
        //和传统加法一样进位数不断向高位加
        for (int i = len-2; i >= 0; i--) {
            sum = (x[i] & LONG_MASK) + carry;
            x[i] = (int)sum;
            carry = sum >>> 32;
        }
    }