I am not sure about the answer to your question on viral load, but regarding linearization of plasmids, I can direct you to this post (you've probably already seen it by now):
Apparently it is important to linearize plasmids for the lowest Ct value. However, you can always do a side by side comparison yourself and see if for your gene, you see a difference.
To maintain sensitive detection in samples, many people use a carrier RNA in the stored cDNA to help prevent loss of detection due to stickiness of the cDNA to the sides of the tube. We use 10ng/ul and some people use 30 ng/ul:
This seems to help allow for detection down to low copy numbers. You can use tRNA or homopolymer A.
Regarding the amount of RNA to use, I think this really depends on how much you have to start with. Many people doing viral load testing on blood, serum, and plasma will have ng-pg quantities to work with and still be able to sensitively detect virus. Sometimes people will have to do a nested PCR to get the product to a detectable level. Since it sounds like you have a lot of RNA to work with, I think you have an advantage. Why not try priming RT with a gene specific primer for the virus instead of RTing the entire mix with random primers or polydT?
Hopefully someone with direct viral load testing experience can give you more specific insight.