On 27/01/2023 18:00, Chris Morgan wrote: > On Fri, Jan 27, 2023 at 05:53:49PM +0100, Krzysztof Kozlowski wrote: >> On 27/01/2023 17:46, Chris Morgan wrote: >>> On Fri, Jan 27, 2023 at 11:24:03AM +0100, Krzysztof Kozlowski wrote: >>>> On 26/01/2023 17:55, Chris Morgan wrote: >>>>> From: Chris Morgan <macromorgan@xxxxxxxxxxx> >>>>> >>>>> Add compatible string for RTL8821CS for existing Realtek Bluetooth >>>>> driver. >>>>> >>>>> Signed-off-by: Chris Morgan <macromorgan@xxxxxxxxxxx> >>>>> --- >>>>> Documentation/devicetree/bindings/net/realtek-bluetooth.yaml | 1 + >>>>> 1 file changed, 1 insertion(+) >>>>> >>>> >>>> Based on your driver it is compatible with rtl8822cs, so you can >>>> indicate it in the binding and use just one of_device_id entry. >>> >>> It's very similar, and uses some of the same constants, but has a >>> different firmware. I assumed the firmware difference is enough to >>> require it to be differentiated, but I wasn't sure. You are saying >>> it does not? I just want to be clear. >> >> If by "differentiated" you mean "incompatible", then depends: >> 1. You have firmware-name property to indicate the firmware to load and >> then you can use one compatible to bind and just load different firmware, >> 2. If device variant is autodectable, it's compatible. >> >> realtek,rtl8723ds-bt also was added to the driver uselessly... > > Gotcha. > > They are incompatible, but the driver should be able to autodetect each > device as best I can tell. In fact all of the devices listed in the > bindings should be autodetectable by the driver (again, as best I can > tell). Honestly though that's assuming I'm using the correct firmware, > which is why I have this tagged as an RFC. I'm really hoping to get > Realtek's attention for them to chime in to confirm I'm using the > most recent firmware and under what license/terms the firmware can > be redistributed so that the firmware may be added to linux-firmware. > > Assuming everything is good though, I can resubmit V2 and instead of > adding a new compatible just noting that the 8822CS and 8821CS use > the same bindings. They cannot use the same bindings. You always need specific compatible and this was not discussed here. What's discussed is to drop the driver change in of_device_id. Best regards, Krzysztof
