On Thu, Nov 12, 2020 at 05:36:17PM +0100, Alexandre Belloni wrote:
> On 12/11/2020 17:57:51+0200, Andy Shevchenko wrote:
> > The commit 9c19b8930d2c ("rtc: ds1307: Add ACPI support") added non-valid
> > ACPI IDs (all of them abusing ACPI specification). Moreover there is
> > no even a single evidence that vendor registered any of such device.
> >
> > Remove broken ACPI IDs from the driver. For prototyping one may use PRP0001
> > with device tree defined bindings. The following patches will add support
> > of that to the driver.
>
> I'm intrigued, how does PRP0001 work? Where would the device tree come
> from?
>From nowhere :-)
There is no device tree. The properties are coming from _DSD ACPI Method with
compatible string provided. ACPI glue layer provides a mechanism to parse that
and match against OF ID table (even when CONFIG_OF=n!).
More about PRP0001 is in documentation [1].
> You probably want to have a look at:
> https://lore.kernel.org/linux-rtc/20201112130734.331094-3-ch@xxxxxxx/T/#u
Thanks for heads up!
I'll check that thread and answer there.
[1]: Documentation/firmware-guide/acpi/enumeration.rst
--
With Best Regards,
Andy Shevchenko
