On Fri, May 06, 2016 at 02:33:04PM +0000, Opensource [Adam Thomson] wrote:
On May 06, 2016, 13:27, Mark Brown wrote:
Is this *really* sensible? DT idioms don't always match up with ACPI idioms well and this isn't a trivial DT binding.
For what we're doing here, both DT and ACPI match up well, and so what I've implemented works on both sides. There are other examples of this already in the Linux kernel, so I don't think it's anything particularly new.
No, not really - your DT is fairly unusual in how it's done and the lack of ACPI helpers is not a good sign on that side. The _DSD things are really only supposed to work for simple properties on devices.
There's nothing device specific about this, it should go in generic code.
The intention was to just match against DT or ACPI and nothing else, so that didn't feel generic enough to be pushed into the fwnode framework. However I will take another look.
That's currently the entire set of things that fwnode supports so...