On Tue, Jun 13, 2023 at 02:54:36PM +0100, Mark Brown wrote:
On Tue, Jun 13, 2023 at 01:46:41PM +0100, Conor Dooley wrote:
So it is entirely a software construct? Why does it need a dt-binding then? Your commit message says the controller is present on the device!
A typical embedded (or power efficient laptop) audio design will consist of multiple devices connected together (frequently via non-control buses) together with system level passive components and plastics which are also important to the audio configuration. A card binding describes the interconections between the devices in the system and provides identification information for the audio subsystem. This system level audio integration is a physical thing that can be pointed at that requires real software control.
The bit you were responding to with that was a disingenuous question. Probably not fair of me to ask one of a non-native speaker like that, when all I wanted to know was whether it was appropriate not to add a more specific compatible, or whether this was something invariant.
Like I said before please look at the existing audio card bindings.
Yah, ofc I did that to see if there were other similar bindings that used specific compatibles...