Re: [PATCH v2 2/2] iio: adc: ti-ads1015: Add support for label

[Flaviu Nistor <flaviu.nistor@gmail.com>]

On Sat, Sep 5, 2026 at 5:01 AM, Jonathan Cameron wrote:
>On Thu, 3 Sep 2026 22:03:55 +0300
>Andy Shevchenko <andriy.shevchenko@intel.com> wrote:
>
>> On Thu, Sep 03, 2026 at 06:16:27PM +0300, Flaviu Nistor wrote:
>> > On Thu, Sep 3, 2026 at 9:18 AM, Andy Shevchenko wrote: =20
>>=20
>> ...
>>=20
>> > Since the iio core exposes in sysfs the in_volatagex_label
>> > for all channels, there are 3 situations: =20
>>=20
>> > 3. The channel is not defined in the dts (in_voltagex_label is still vi=
sible
>> > in the sysfs), and reading the label will cause "(null)\n" to be printe=
>d in
>> > the console (as indicated also by Sashiko boot), since the default "" v=
>alue
>> > is not assigned during device_for_each_child_node_scoped(). I think thi=
>s is
>> > not an issue since is just the way sysfs_emit() handles an NULL string.=
> =20
>>=20
>> Is this current state of affairs?
>>=20
>> If so, the patch must not change that =E2=80=94 it's an ABI now and (null=
>) is valid
>> response from the user space point of view.
>>=20
>> > What would you prefer? Any further suggestions? =20
>>=20
>Unfortunately we never defined what should happen this way explicitly.
>In practice assumption is that this is a corner case we probably won't
>see. People don't tend to want to label just one random pin, they tend
>to label the lot. =20
>
>So go with whatever is most common in other drivers,
>

Hello, I took some time and went through all of the drivers that use read_label()
to get the most common solution. Most of the drivers only use datasheet_name and only
a few actually get the label (if present) from the device tree during probe.
I now have the best solution (from my point of view), which covers all (3) cases. If
the label string is present in the device tree it will be use and printed in the console
in user space. If label not present or channel is not defined at all in the dts, the
datasheet_name will be used similar to most drivers. I will send a v3.

PS: many thanks for everybody for pushing for a better solution.

Best regards,
Flaviu

>Jonathan