I think you missed the biggest two: overdoses and neglected healthcare. My county is relatively untouched by covid, we've had like 15 deaths total. At the same time, we jumped from 10 overdose deaths a month to about 30 since we locked down. And then between the hospitals cancelling a lot of "elective" procedures and people generally avoiding the hospitals at all costs, a lot of people's health is being neglected. That's the insidiousness of the "deaths year over year" suggestion people make. Imagine an older man gets stressed over the lockdown and has chest pains that might be a heart attack, and then decides not to go to the hospital to avoid covid, and then dies of the heart attack. People would say covid caused that death, and I agree in a sense, but that doesn't help if we're trying to figure out how much caution is justified around the virus itself.
Additionally, the comparison isn't covid deaths vs additional non-covid deaths, because nobody dying of covid-19 isn't an option. The comparison would be to the covid deaths actually prevented. Which, I understand, is a practically impossible analysis. But that was the whole point of those flatten the curve graphs. In theory, by avoiding overburdening of the hospitals, you prevent the majority of the preventable deaths. We might not have succeeded at that in New York City and a few other early outbreaks, but at this point it seems the hospitals everywhere are doing just fine, so it's unclear what deaths are being prevented rather than just being a delayed eventuality. I'll put it this way: it doesn't matter to the analysis if 100,000 have died and 1,000,000 deaths have been prevented if further lockdown is going to save 10 covid deaths and kill 50 more elsewhere. That the scales of the numbers are so far apart is irrelevant to the cost benefit analysis of further action.