On Mon, Oct 20, 2014 at 03:10:24PM +0900, Gyungoh Yoo wrote: > On Fri, Oct 17, 2014 at 04:26:05PM +0200, Mark Brown wrote: > > The thing I'm seeing is that the binding for your device with the > > subnode looks very much like the device trees of devices with multiple > > regulators. The fact that you only have one regulator is a bit > > difference but not that much. It seems like drivers should fit into one > > of two patterns: either the regulator is described in the root node for > > the device for single purpose devices or there should be a collection of > > regulators like is supported with this helper API. Having a collection > > with only one node doesn't seem to be a problem in any way. > Thank you for your kind comments. > My understanding is getting better. > For my clear understanding: > I think the original designed which I wanted to design is similar > with arizona-ldo1.c > It seems that this is 1st pattern your explained above. > Can I ask what is different between arizona-ldo1.c and > this sky81452-regulator.c? > I think both are designed under root node. Were that code being written today I'd probably be asking for it to use a regulators subnode to be consistent with everything else, it's a slightly older driver and sometimes we've not spotted the patterns when they're added, once we get to the point of adding the helpers as we have now things are definitely getting baked in.
Attachment:
signature.asc
Description: Digital signature
