Hmm. That's a reasonable ruling. It's not exactly what's happening though; after all Orcus does have enough movement, it's just that it ends when he's hit - and that ending occurs when the OA occurs, by which time (necessarily) orcus must have already moved several squares into a position in which he can't end his movement.
Ending movement means he has no movement left. It doesn't end the action, it ends the movement, and that's an important distinction.
The distinction is clearer is you imagine a third creature on this hypothetical battlefield. Orcus is walking over that third creature, and at some point during that movement he provokes an OA from the fighter - which hits and ends the move. Obviously, you can't retrospectively pretend Orcus never moved over the creature and that his move ended in the last legal position since from that position, the fighter may well have not been able to hit orcus - or for that matter, he may have a feat or an ability that allows him to shift or move, so that even the original location is impossible!
Orcus attempts to step over the creature. The fighter (who, by the way, has power equivalent to a demigod, a world-saving hero of legend, and is at the absolute maximum a demigod hero of legend can be. This isn't even Heracles, this is Heracles at his absolute best.) does an epic heroic assault, knocking Orcus back where he ways, using all the enemy-stopping techniques he's developed over his entire career.
This idea you can't 'retroactively' pretend anything in a game with immediate interrupts irks me. There's entire action types devoted to that exact concept. The ability in question, Combat Superiority, is essentially 'retroactively pretend that




never got to happen.'
In short, no interpretation of orcus's movement rules can prevent a situation in which an OA occurs in a position where he may not end and which could not have been made from the last safe position, nor can it prevent situations whereby there are no legal positions for orcus to end in (if a reaction or interrupt to the movement changes the battlefield).
The former is a non-issue, and while the latter is a corner case, it can be resolved in a chain of interrupts.
Obviously the movement interrupt that put a guy in Orcus's spot is resolved at that time. But the Fighter's interrupt occurs before Orcus's movement, moving him back to his original square. The thing is, that original movement interrupt has -that- guy in an illegal square, so he then has to move back to the first square he had moved into where he could legally occupy.
Is this RAW? Seriously, this comes up so utterly rarely that there doesn't need to be RAW for this. I've said it before, and I've said it again.
Roleplaying games do not need rules to cover every single rules corner case or loophole. That job belongs to the guy at the head of your table with the DM's screen.
So, a DM may need to adjudicate ending in an occupied square regardless.
In a very rare corner case, yes, a DM would have to actually adjudicate something in this game.
As is his job.