You can’t get to 0 by taking a positive number and dividing by two. The number will get smaller and smaller and never reach 0. The best you can hope for is to divide until you have a number smaller than your floating point library can handle which still may or may not get to 0. The answer is definitely NOT O(log n)
What a crap question, can you get rid of it please?