Re: [BUG] OV02C10 on Dell 16 Premium DA16250 (ARL): INT3472 handshake-derived "dvdd" regulator registered but never linked to sensor, sensor probe fails with -EREMOTEIO

[Angioli Samuele (gmail)]

Marco,

Agreed on all counts, and thanks for closing the loop on it.

I won't push the standalone dvdd_second_sensor quirk: a field with no 
enabled board behind it is dead code, and the second_sensor mechanism is 
already there for whenever a board that actually needs it shows up. The 
validation stands on its own as a data point if that day comes.

I'll take the NVS topology issue to Dell through what channel I have -- 
the RGB-side PMIC (DSC1/LTH1/CLP1) being _STA=0 while only the IR-side 
DSC0 is live looks like a factory-config decision on the Linux SKU, not 
a hardware limit, since the Windows image drives the sensor fine. If 
they flip the gate (or expose the rails on a live instance), the 
existing int3472 paths should pick it up with no kernel change.

Thanks again to you, Hans and Sakari for the time on this one.

Samuele

Il 18/06/26 09:22, Marco Nenciarini ha scritto:
> Samuele,
>
> The _STA dump settles it, thanks for running it down. With DSC1/LTH1/CLP1 all
> _STA=0 and only DSC0 (INT3472:0c) live, the OV02C10 has no live PMIC instance on
> the DA16250, and the gate is a BIOS/NVS config (CL01/C1TP read but never
> assigned, no _OSI branch), so acpi_osi spoofing can't reach it. avdd/dovdd/reset/
> clock have no live provider to key to.
>
> That closes the kernel angle: there is nothing for int3472 to fix here, and the
> dvdd re-key, though correct, can't bring the sensor up. This is a Dell firmware
> matter (the RGB-side PMIC topology left disabled in NVS), worth reporting to them
> if you have a channel.
>
> On the standalone dvdd second_sensor quirk: I'd hold off. The second_sensor
> mechanism already exists; a dvdd_second_sensor field with no enabled board behind
> it would be unused code, and the kernel doesn't carry a quirk that has no
> consumer. If a board turns up that actually benefits, the field and its DMI entry
> can land together in that series, with the consumer that justifies them. Until
> then there is nothing to upstream.
>
> Thanks again for the thorough work on this one.
>
> Marco