This makes insertion to block list O(log n) instead of O(n) when blocking with wake time. This is more heavy and currently slower than the previous implementation, but that is just because we don't really have too many threads running at any given time. This will be more scalable in the future and even allows implementing priorities if I ever wish so