Re: [PATCH v1 1/1] hwmon: (isl28022) Don't check for specific errors when parsing properties

[Andy Shevchenko <andriy.shevchenko@linux.intel.com>]

On Thu, Feb 19, 2026 at 03:21:29PM +0100, Geert Uytterhoeven wrote:
> On Thu, 19 Feb 2026 at 15:06, Andy Shevchenko
> <andriy.shevchenko@linux.intel.com> wrote:
> > Instead of checking for the specific error codes (that can be considered
> > a layering violation to some extent) check for the property existence first
> > and then either parse it, or apply a default value.

> IIRC, we have removed superfluous presence checks all over the tree
> during the past few years? E.g. of_property_read_*() is documented to
> return -EINVAL if a property does not exist.

Even though, it's still fragile. When we have a check for explicit device
presence, we wouldn't care of the error code we get in case of unsuccessful
parsing.

> So this patch looks like a step back to me...

Obviously I have a disagreement here, this is step forward to weaken
the dependency on the certain error code in the cases when we can avoid
that. Motivation is mentioned in the commit message.

Also note, -EINVAL can sneak in tons of mysterious ways as it's one of
the most overloaded error code in the kernel, its semantic is basically
equals to "an error happened".

Having the code above, we make it robust against some subtle nuances which
may not be discovered in time.

-- 
With Best Regards,
Andy Shevchenko