hi and good question.
however, really you need to seperate out the requirements.
what is needed for SQL and what is needed for HyperV are not the same thing.
your last question is the relevent one, do i create LUNS exclusively for SQL and that is the case, looking at the netapp best practice is the way to go...
you then present these into your hyperv machines, either as ISCSI or pass-thru disks...
so the questions about CSV are really hyperv specfic (SQL 2014 now adding CSV support into SQL clusters) and are to do with your hyperv environment, the SQL disk layout is about presenting disks back to your SQL envronmnet, and about these disks been dedicated to that task - if you use VHD files to house them, then that is a seperate set of considerations.
Hope that makes some sense...