First lets constrain this conversation a bit- let's only look at the US and the likelihood of famine occurring there (as we have already been talking about it). Secondly I think we can both agree that the result this is heavily dependent on the type of strike, how may cities are actually lost and the surrounding context of the exchange (is the US and Russia/China at war), but for the sake of the conversation lets say that the US immediately won the prevailing conflict and doesn't have to concern itself with preforming or dealing additional military operations. So which cities are mostly likely targeted? One needn't look far to see that targeting logistic hubs is fairly high on the list of likely enemy action (1, 2) following close behind domestic nuclear facilities and military targets. It is hard to say but perhaps we can go with a few dozen of Americas top cities were targeted along with a few Nebraskan wheat fields and military bases (it seems to me this is the scenario that most in this thread have been talking about).

Okay, with that context lets revisit the initial problem area- what does the logistics chain of bulk commodities following a nuclear exchange look like? Firstly, I invite you to take a look at this map of US rail lines (note that this does not include all rail connections- I'm assuming spurs and private lines are not shown). Notice many of the large intersection points are in large cities which are very likely to be hit (also most of these rail yards are actually very close to the city center where fireball damage would occur). As you say most of the rail lines outside cities would likely avoid significant damage, however most of the important logistics infrastructure lies within the rail yards this includes the rolling stock, offloading equipment, signaling/coordinating structures, and arguably the most important the human capital- would all be destroyed. Yes, other less important or used track could be cannibalized for use and various surviving heavy machine shops somehow forced into helping with the reconstruction efforts, I am left a bit skeptical how they get a bunch of workers (who maybe just lost everything in a blast) to go clear, sort debris, re-lay track and signaling, and learn how to operate the whole system all within inside a highly radioactive zone.

I digress though, thank you for bringing up the inland water ways- I think this is one of the best reasons the US could weather intentional logistics damage the best. However I will again point out that the limited barge stock available, and particularly where the largest inland parts are located- again in large cities that would likely be targeted.

From a meta point of view, the key thing to remember here is not what is theoretically possible but what is likely using previous experience, to happen. As an example - theoretically yes, someone with access to crude oil, some metal stock, and a small amount of equipment could make a very rudimentary oil distillation system but why would they do that? I get that you originally used it just to illustrate how basic some industrial processes are (which I don't know if I really agree), but missing in your thought process is the 'why?' or 'who?' would do all of this. Logistics is Coordination and coordination is hard. Sidenote: Martial Law or at minimum a nationwide state of emergency would likely be declared in such an event likely simplifying decision making and coordination. However, how quickly this could be ramped up to nationwide scale following significant damage to communication systems I will leave unassessed.

I would recommend reading about the supply chain issues that have been caused by COVID and the ongoing Ukrainian War, and just extrapolate out. I realize that because we are discussing famine it narrows the scope of what we are concerned about (ie only moving food from A to B) and with a short time horizon, the issue is still surprisingly complex.

My apologies if my line of reasoning may have been hard to follow at the end let me know if there is anything you find unclear.