Re: [PATCH v6 2/5] dt-bindings: media: Add qcom,x1e80100-camss

[Konrad Dybcio]

On 4/24/25 11:10 PM, Bryan O'Donoghue wrote:
> On 24/04/2025 21:08, Konrad Dybcio wrote:
>>> The Test Pattern Generators - TPGs would then also fit into this new model for the PHYs.
>> Maybe we could consider modeling various camss subdevices as separate DT nodes,
>> perhaps on some sort of a `camss` simple-pm-bus or something alike
> 
> I hadn't though of that specifically, call it option 0.
> 
> I had been thinking of
> 
> 1. Doing like venus with a subdevice based on a compat
> 2. Doing it like DPU PHYs and just moving everything into drivers/phy
> 
> The fact that the CSIPHYs are technically inside of the CAMSS collapsible power-domain seem to mitigate against option 2.

Option 1 sounds just a little horrid, I'd be interested in something that
combines 0 and 2..

We may either repeat the power domain everywhere (meh), or model camss
as a bus, (see e.g. agnoc@0 in msm8996.dtsi) and focus on other resources
required by specific blocks (e.g. CSIPHY clocks for the CSIPHY)

Konrad