Much of the psychiatric literature refers to psychopathy as a "mental disorder".
I guess it's partly a matter of what we mean. We could decide that if there's an adaptive explanation for some trait, we won't call it a disorder. That said, the fact that there could be an adaptive explanation for a trait doesn't mean that there is and in particular, the fact that there could be an adaptive explanation for psychopathy doesn't mean there is, and we may well never know
All that said, perhaps the real point here is that what we count as a "disorder" isn't just a matter of whether it makes the person who has it more likely to survive or reproduce. We might be able to make the case that on average, psychopathy doesn't contribute to reproductive fitness or other related matters. But even if it turned out that it actually does, we would most likely still classify psychopathy as a disorder.
All of which tends to confirm the suspicion that diagnoses sometimes (often? always?) rest partly on value judgments. That raises some red flags, as the case of homosexuality demonstrates. Homosexuality was once diagnosed as a mental disorder. Looking back on the matter now, it seems clear that this was the result of a debatable (I'd say mistaken) value judgment. That's just one among many such cases and so there's lots of room to worry.
Worry or not, however, it's not clear that we can expect to have a purely biological, non-evaluative notion of health and illness -- physical or psychological. The fact that something is adaptive in the biologist's sense doesn't necessarily mean that we should simply accept it and not try to find ways of preventing it or mitigating it. And the fact that something isn't adaptive in the biologist's sense doesn't mean that we should treat it as pathological. How exactly all this should be sorted out, however, is clearly not easy to say.