Best single buy-then-sell profit from a price list

An array gives a stock's price on each successive day. You may buy once and later sell once (the sale must be on a strictly later day). Describe an algorithm returning the largest achievable profit, or $0$ if no trade makes money. For example, prices $[7,1,5,3,6,4]$ give $5$ and $[7,6,4,3,1]$ give $0$.

One pass tracking the running minimum: $O(n)$ time, $O(1)$ space. Scan left to right, keeping the lowest price seen so far; at each day the best sale today is price minus that minimum, and you keep the largest such value (starting at $0$). This replaces the naive $O(n^2)$ check of all buy/sell pairs.