On Thu, Dec 21, 2017 at 02:29:14PM +0100, Maciej Purski wrote: > Now I can understand your point, but I still have doubts what is the > advantage of that solution. For non-coupled regulators we end up with > useless data structure - coupling_desc. That also might cause some > confusion. We expect coupled regulators to be a very rare case, so in most > of the cases we will have a pointless structure in reg_dev with a pointer to > itself. Maybe you suggest that coupling_desc should contain something > different? It's precisely because it's such an unusual case that I'm looking to see it as part of the common path - if it's not then it seems very liklely that it'll get broken down the line and nobody will notice because the code is never run. Yes, it's unusual but having the unusual thing be a bit more visible than it is used seems better than having support there that gets broken.
Attachment:
signature.asc
Description: PGP signature
