Re: [PATCH v2 02/11] dt-bindings: hwmon: lm75: use common hwmon schema

["Théo Lebrun" <theo.lebrun@bootlin.com>]

Hello,

On Fri Mar 1, 2024 at 4:35 PM CET, Rob Herring wrote:
> On Fri, Mar 01, 2024 at 11:44:37AM +0100, Théo Lebrun wrote:
> > Hello,
> > 
> > On Fri Mar 1, 2024 at 11:13 AM CET, Krzysztof Kozlowski wrote:
> > > On 01/03/2024 10:41, Théo Lebrun wrote:
> > > > Hello,
> > > > 
> > > > On Fri Mar 1, 2024 at 7:53 AM CET, Guenter Roeck wrote:
> > > >> On 2/29/24 22:37, Krzysztof Kozlowski wrote:
> > > >>> On 29/02/2024 19:10, Théo Lebrun wrote:
> > > >>>> Reference common hwmon schema which has the generic "label" property,
> > > >>>> parsed by Linux hwmon subsystem.
> > > >>>>
> > > >>>
> > > >>> Please do not mix independent patchsets. You create unneeded
> > > >>> dependencies blocking this patch. This patch depends on hwmon work, so
> > > >>> it cannot go through different tree.
> > > > 
> > > > I had to pick between this or dtbs_check failing on my DTS that uses a
> > > > label on temperature-sensor@48.
> > >
> > > I don't see how is that relevant. You can organize your branches as you
> > > wish, e.g. base one b4 branch on another and you will not have any warnings.
> > 
> > That is what I do, I however do not want mips-next to have errors when
> > running dtbs_check. Having dtbs_check return errors is not an issue?
>
> That's a good goal, but difficult to achieve as you can see. Given 
> dtbs_check in general has tons of warnings already, we currently don't 
> worry about more warnings in specific branches. We just look at mainline 
> and linux-next. And for that it's still so many, I'm just looking at 
> general trends. It runs daily here[1].

Here's my opportunity to ask a question I've had for a while: do you
have a way to filter out dtbs that are known to be bad actors (ie have
many many warnings)? Maybe a list of platforms you talk about below
that aim at zero warnings?

Another way to ask this: what would be a good default DT_SCHEMA_FILES
value? Not filtering leads to way too many errors.

Regards,

--
Théo Lebrun, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com