Alright, so I posted this a few months ago and finally got around to trying this out for myself rather than simply reading stuff on the internet on how to mitigate this issue.
So, this is what my brass looked like before. It is very apparent that there are donuts.
I bought a 7mm inside neck reamer and tested the two most common methods on a few sacrificial pieces of 12+ times fired lapua brass (this is a 284 Shehane chambering). Bear in mind I use a bushing die as well. This brass will be tossed soon, FYI.
Method 1: I took a fired case and used a 7mm neck die to get the whole neck sized. I then used an expander mandrel and sized the brass back up to push all the material back to the outside. I then turned the neck as usual. That is the case on the left in the bottom first picture. My observation with this method are that it definitely got rid of the donut and actually took the rounded section off my shoulder/neck junction and now I have what looks like a normal looking case. I could not tell how much material was left in the shoulder/neck region and I will admit I have done this before and never had any of those cases split after 5+ firings.
Method 2: I took a fired case and simply used a Wilson 7mm inside neck reaming tool. This case is on the right. It is obvious where the material was taken out as indicated by the shiny sections in the neck and the majority of it was at the top of the neck and then at the base of the neck where the donut would reside. I also took a bullet and it dropped right into the case with no resistance (I have a .317 neck and my fired case neck OD measures at .316 on these cases). Perfect! I don't know if I necessarily like that it faced off a bit of the neck at the top.
Method #1 yields cases with nice sharp shoulder/neck junctions, but it is a much longer process. Method #2 is the easiest, but I will still be living with rounded off necks that will never be as sharp as a FL non-bushing die. So, I think I am going to take the best of both methods.
I bought a non-bushing FL sizing die from Whidden with my desired neck OD of .310. This will give me a couple options. I can turn my necks and use the die as is and will get my desired neck tension, or no-turn and then use an expander mandrel as a final sizing step to get the desired neck tension (some argue is a more consistent way of getting consistent neck ID, and I believe that). When/if a donut forms (and I wonder if I will see donuts this bad with a FL non-bushing die), I can then use an inside neck reamer, which will be faster than turning the OD of the necks. Donut solved, and rounded off shoulder junctions solved.