If you don't want it to be expensive as hell - any part.
Also with space resources you can make the whole thing quite comfortable - big ship, lots of fuel->deltav->short travel time and lots of shielding for when the Sun decides to reach out and touch you.
Compared to the total cost of the mission I highly doubt that that is going to be the thing it hinges on, besides, all you need to be able to do is get back to orbit from Mars, you don't need to lift the fuel to get back to Earth from the Mars surface, you just leave that parked in orbit.
Anyway, I don't see it happening at all so debating the execution details of things that are in the realm of the solvable already or at best an optimization isn't going to move the needle, the things that need to be solved that we have no clue about are life support, mental health, exposure to radiation for a prolonged period, waste management. Keep in mind that there is no way to re-supply a mission like that en-route and a crew of say four (which would seem to be a minimum for such a mission, and probably is too low from a redundancy perspective) will eat and drink their way through a small mountain over the course of 4 years+ total mission duration.
Believe it or not it's around a 60% reduction in launch mass, so it should cut costs significantly. Tsiolkovsky equation do be like that. https://ntrs.nasa.gov/citations/20140009943
As for all the human related needs, the ISS has been a decades long study on practically all of those. That's kind of why we have it.
Supplying the mission isn't really a fundamental problem since you can send multiple times redundant amounts of all the supplies there on the previous cycle. But yes if something goes wrong they're on their own to solve it for (2 years - time spent there), much like any other mission not in close proximity to Earth. That's why you don't send idiots.
And you're right that there's no point in debating it further, because no amount of arguing will stop humans landing on Mars sooner or later.
