Re: [PATCH v2 01/10] dt-bindings: ptp: Document the TSCSS on Qualcomm SA8775P and QCS8300 SoCs

[Imran Shaik]

On 09-11-2026 06:18 pm, Krzysztof Kozlowski wrote:
> On Sun, Nov 08, 2026 at 06:48:22PM +0530, Imran Shaik wrote:
> > +properties:
> > +  compatible:
> > +    oneOf:
> > +      - const: qcom,sa8775p-tscss
> > +      - items:
> > +          - const: qcom,qcs8300-tscss
> > +          - const: qcom,sa8775p-tscss
> > +
> > +  reg:
> > +    minItems: 1
> > +    maxItems: 2
> > +
> > +  reg-names:
> > +    minItems: 1
> > +    items:
> > +      - const: tsc
> > +      - const: etu
>
> ...
>
> > +required:
> > +  - compatible
> > +  - reg
> > +  - reg-names
> > +  - clocks
> > +  - clock-names
> > +
> > +allOf:
> > +  - if:
> > +      properties:
> > +        reg-names:
> > +          contains:
> > +            const: etu
>
> Reason for same devices with different address space details should be
> provided somewhere, preferrably in description. Assuming in the first
> place, that this stays because how two devices can be basically
> identical (same compatible) if they have different programming
> interface?

TSCSS is a single hardware device, split into two sub-blocks - TSC and
ETU, each with its own register address. Every SoC with this compatible
has both blocks in hardware.

What changes is whether the vendor/board actually wires up the ETU's
external event. The ETU block captures a timestamp from the TSC block
when an external event fires one of the ETU slices, which are driven
through board-level GPIO/IRQs. If a board doesn't wire any slice,
there's nothing in the ETU region for Linux to use.

So interrupts/pinctrl/qcom,etu-* only matter once a slice is actually
connected. Because of this, only the tsc reg space is described in the
DTS patches today.

Instead, I'll use the "qcom,etu-slice" property to drive the required
constraints, and add reg/reg-names support in DTS by default for both
TSC and ETU, and update the description with the above details in the
next series. I hope that works.

Thanks,
Imran