Re: [PATCH v5] i2c: mux: reg: use device property accessors

[Abdurrahman Hussain]

On Tue May 19, 2026 at 1:39 PM PDT, Peter Rosin wrote:
>
> Hi!
>
> Thanks for your patience in seeing this through!
>
> Acked-by: Peter Rosin <peda@xxxxxxxxxxxxxx>
>
> Wolfram, can you please pick this one? Perhaps after waiting a bit
> for extra tags, see below.
>

Thank you for taking the time, and all the feedback!

>> ---
>> - Discriminate ACPI children with is_acpi_device_node() instead of
>>    is_acpi_node().  v4 routed every is_acpi_node() child to
>>    acpi_get_local_address(), which is wrong for ACPI data nodes
>>    (the _DSD hierarchical-property children PRP0001-style firmware
>>    uses): data nodes have no ACPI handle, so ACPI_HANDLE_FWNODE()
>>    returns NULL, acpi_get_local_address() falls back to evaluating
>>    _ADR against the root namespace, and the probe fails with
>>    -ENODATA.  is_acpi_device_node() narrows the match to ACPI
>>    device nodes (which do carry an _ADR), and the else-branch now
>>    resolves "reg" via fwnode_property_read_u32() for OF, swnode,
>>    and ACPI data nodes alike.  Flagged by Sashiko on v4.
>
> According to Documentation/process/coding-assistants.rst, you should
> perhaps credit the tool with an Assisted-by tag? Given your mail
> address I suspect you might have used other AI tools as well? Just
> add tags for whatever is appropriate.
>
> Cheers,
> Peter

I used claude-opus-4-7, mostly for helping with the wording of commit
messages, and sashiko (gemini-3.1-pro-preview) helped catch the
is_acpi_node()/is_acpi_device_node() distinction.

Will add the tags and send v6 shortly.

Best regards,
Abdurrahman