Re: [PATCH RESEND v4 3/3] arm64: dts: qcom: oneplus-kebab: new device

["Frieder Hannenheim" <mail@fhannenheim.net>]

On Mon Nov 9, 2026 at 3:57 PM CET, Konrad Dybcio wrote:
> On 11/9/26 3:53 PM, Frieder Hannenheim wrote:
>> On Mon Nov 9, 2026 at 12:08 PM CET, Konrad Dybcio wrote:
>>> On 11/5/26 9:20 PM, Frieder Hannenheim wrote:
>>>> Add initial device-tree for the OnePlus 8T. Downstream reference is [1].
>>>>
>>>> [1] https://github.com/LineageOS/android_kernel_oneplus_sm8250/blob/lineage-23.2/arch/arm64/boot/dts/vendor/oplus/kona-kebab-overlay.dts
>>>>
>>>> Co-developed-by: Robert Oscilowski <drgitx@gmail.com>
>>>> Signed-off-by: Robert Oscilowski <drgitx@gmail.com>
>>>> Signed-off-by: Frieder Hannenheim <git@fhannenheim.net>
>>>> ---
>>>
>>> [...]
>>>
>>>> +&i2c2 {
>>>> +	status = "okay";
>>>> +
>>>> +	pm8008: pmic@8 {
>>>> +		compatible = "qcom,pm8008";
>>>
>>> No PM8008 on the other phone?
>> 
>> There's 4 phones in total: 8 (instantnoodle), 8 Pro (instantnoodlep), 8T (kebab)
>> and 9R (lemonades). The OnePlus 8 doesn't have a PM8008. Should I define
>> this in -common? It would need to be disabled when a dts for the OnePlus 8 is
>> submitted.
>
> Okay, I thought there were only two. Do the remaining two feature
> this PMIC? Generally we try to do a lowest common denominator in
> -common, whatever that ends up being
>

I all except the 8 have it according to the downstream dts.

>
>>>> +		reg = <0x8>;
>>>> +
>>>> +		interrupt-parent = <&tlmm>;
>>>> +		interrupts = <112 IRQ_TYPE_EDGE_RISING>;
>>>
>>> interrupts-extended = <&tlmm 112 IRQ..>
>> 
>> May I ask why? According to the docs interrupts-extended is only useful with
>> multiple interrupt parents which is not the case here. Or just for brevity?
>
> Yep just to be concise
>
>>>
>>> [...]
>>>
>>>> +&vreg_f {
>>>> +	status = "disabled";
>>>> +};
>>>
>>> I think it may be cleaner to move the definition to the other
>>> phone's DT (are there just 2, or more?)
>>>
>>> Konrad
>> 
>> The f block is used on the 8 and 8 Pro I believe. So I think it's best to keep
>> it this way.
>
> Thinking about it again, I'm a little surprised there would be variance
> between them. Can you do
>
> `cat /sys/kernel/debug/qcom_socinfo/pmic_model_array`
>
> ?
>
> Konrad

fried@oneplus-kebab ~> sudo cat /sys/kernel/debug/qcom_socinfo/pmic_model_array
PM8150 3.0
PM8150B 2.0
unknown (39)
PMK8002 2.0
unknown (42)

It's not used on kebab and errors out when enabled so it's disabled.

Frieder