Re: [PATCH 1/2] dt-bindings: iio: temperature: Add ADT7604 support to adi,ltc2983

[Conor Dooley <conor@kernel.org>]

[-- Attachment #1: Type: text/plain, Size: 2467 bytes --]

On Wed, May 06, 2026 at 01:06:35PM +0000, Stan, Liviu wrote:
> Thank you for the comments, and I apologize for the delayed answer. 
> 
> On Mon, Apr 27, 2026, Conor Dooley wrote:
> > I have to wonder if this is the right approach, if it's the same device
> > just with a different label and advertised purpose.
> 
> Could you expand on this? Are you suggesting the copper-trace@ and
> leak-detector@ node types aren't needed, or something about the overall
> approach of adding ADT7604 to the existing ltc2983 binding? I want to
> make sure I understand the concern.

Yeah, if it is the same device, just with different uses for the same
pins, I was wondering whether the correct approach is to reuse the
existing child nodes, with some way of indicating what they are
measuring (e.g. use the compatible to decide). I think Jonathan
expressed a similar sentiment.

> 
> > Pedantry perhaps, but isn't this an "ohmmeter"?
> 
> For the naming I followed the convention used in this binding: rtd@, 
> thermocouple@, diode@ are named after what they are, not what
> they measure. copper-trace@ names the component being sensed. Jonathan

An "ohmmeter" is what it is, not what it measures. What it measures
would be "ohms".

> has a related question below about whether this should just reuse rtd@
> with type 18, so the name might change depending on where that lands.
> 
> > I want to look into this property, where is the datasheet?
> 
> The ADT7604 datasheet is not yet publicly available. Will add the URL
> once it is.

Right. It'll be hard to provide a full review without being able to look
at the datasheet.

>  
> > Are these ever linked in a different way?
> > Ditto here and in the rtd node. Are these fixed linkages or actually
> > dynamic?

Please don't trim this aggressively, you've removed the context for what
this is about, and it was 2 weeks ago so there's no way I remember what
it was.

> The linkage is board-dependent. Any rsense on channels 2-20 can be
> paired with any sensor channel in that range. The example uses two
> separate sense resistors because the copper trace and leak detector
> channels need different values (100Ω vs 10kΩ, these are the recommended
> values in the datasheet).

This question of mine is fairly moot anyway, given the existing child
nodes all have have it, especially if there ends up being reuse of them,
so you can ignore this comment.

Cheers,
Conor.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 228 bytes --]