Hi Matthias, On Mon, Sep 7, 2020 at 4:40 AM Matthias Schiffer <matthias.schiffer@xxxxxxxxxxxxxxx> wrote: > My rationale here is the following: As broken as the native CS of these > controllers is, it isn't an unreasonable assumption that it is working > fine with *some* devices or for some usecases - after all the support > was implemented at some point, and has existed for a long time now. If > we really want to remove this feature, a deprecation period with a > warning message seems like the proper way to deal with this. > > Hypothetically, existing out-of-tree DTS could have used the native CS > with num-cs set to 4. Always setting num_chipselect to 4 ensures that > we don't break such DTS with the num-cs removal. I still think this is more of a theoretical issue rather than a real use case one. Anyway, I have a proposal that I think will make both of us happy :-) I will submit a patch shortly.