Given a square 2D list holding only 0's and 1's, determine the greatest number of 1's that can be covered by a square of smaller size. More formally, given an n x n 2D list in [row][col] order holding only 0's and 1's, and an integer k <= n, determine the greatest number of 1's that can be covered by a square of size k x k. Return a tuple with the number of 1's this optimal k x k square covers as the first element, the row of the top left corner of this optimal k x k square as the second element, and the col of the top left corner as the third element.

If several different k x k squares cover the same (maximum) number of 1's, return the row, col pair of the top left corner of the top left optimal placement, i. e., the placement for which row + col is smallest. There will not be a case where multiple optimal placements with the same row + col sum exist. If given k=0 or an empty 2D list, return 0 for the number of 1's covered, 0 for the row and 0 for the col.