On 4.03.2023 18:53, Bryan O'Donoghue wrote:
> On 04/03/2023 17:45, Bryan O'Donoghue wrote:
>> On 04/03/2023 17:35, Konrad Dybcio wrote:
>>>> you'll see no error. However if you just do this
>>>>
>>>> diff --git a/arch/arm64/boot/dts/qcom/msm8916.dtsi b/arch/arm64/boot/dts/qcom/msm8916.dtsi
>>>> index 0733c2f4f3798..829fbe05b5713 100644
>>>> --- a/arch/arm64/boot/dts/qcom/msm8916.dtsi
>>>> +++ b/arch/arm64/boot/dts/qcom/msm8916.dtsi
>>>> @@ -1094,8 +1094,7 @@ mdp5_intf1_out: endpoint {
>>>> };
>>>>
>>>> dsi0: dsi@1a98000 {
>>>> - compatible = "qcom,msm8916-dsi-ctrl",
>>>> - "qcom,mdss-dsi-ctrl";
>>>> + compatible = "qcom,mdss-dsi-ctrl";
>>>> reg = <0x01a98000 0x25c>;
>>>> reg-names = "dsi_ctrl";
>>>>
>>>>
>>>> and run the same test you get
>>> Yes, correct. It's valid but it's deprecated, so the bindings are
>>> sane. Keep in mind there's an ABI-like aspect to this.
>>>
>>> Konrad
>>
>> The _driver_ will still accept "qcom,mdss-dsi-ctrl" which is ABI compliant but, I don't see why the yaml should.
>>
>> If you declare a new .dts with only "qcom,mdss-dsi-ctrl", that should throw a yaml check error.
>>
>> ---
>> bod
>
> Actually. I agree with you, I just dislike it.
If I understand correctly, you are dissatisfied with dt_binding_check
not even throwing a warning when a deprecated binding is present.. I
agree, that could be improved..
Konrad
>
> - "qcom,mdss-dsi-ctrl" <- the driver will accept this
> - "qcom,dsi-ctrl-6g-qcm2290" <- the driver will not accept this
>
> bah
>
> Reviewed-by: Bryan O'Donoghue <bryan.odonoghue@xxxxxxxxxx>
