RE: [PATCH v5 1/2] dt-bindings: timer: Add bindings for Intel Keem Bay SoC Timer

-----Original Message-----
From: Andy Shevchenko <redacted>
Sent: Wednesday, August 4, 2021 1:08 PM
To: Sanil, Shruthi <redacted>
Cc: Rob Herring <robh@kernel.org>; daniel.lezcano@linaro.org;
tglx@linutronix.de; linux-kernel@vger.kernel.org;
devicetree@vger.kernel.org; andriy.shevchenko@linux.intel.com;
kris.pan@linux.intel.com; mgross@linux.intel.com; Thokala, Srikanth
[off-list ref]; Raja Subramanian, Lakshmi Bai
[off-list ref]; Sangannavar, Mallikarjunappa
[off-list ref]
Subject: Re: [PATCH v5 1/2] dt-bindings: timer: Add bindings for Intel Keem
Bay SoC Timer

On Wed, Aug 4, 2021 at 8:35 AM Sanil, Shruthi [off-list ref]
wrote:

quoted

quoted

From: Rob Herring <robh@kernel.org>
Sent: Tuesday, August 3, 2021 4:14 AM

...

quoted

quoted

quoted

+properties:

You need a 'compatible' here. Otherwise, how does one know what 'reg'
contains. Also, without it, this schema will never be applied.

This is a parent block that has the common configuration register address

defined which we would need during the initialization of the child nodes. This
block in itself is not doing anything. We have this because, we have a
common register that is required to be accessed during all the timers and
counter initialization.

quoted

The child nodes have the compatible string, which is used in the driver. I

have validated this on the Keem Bay HW and see that the timer probes are
being called and the timers are functional as expected.

I think I understand now. The problem is that the current state of affairs with
this block is incorrect software representation. What you need is to create an
MFD device driver (for which the compatible will exactly the one Rob is telling
about) and from it you register the rest of your drivers. The existing drivers
for this block should be converted to MFD schema.

Sure Andy, I shall check on this and get back.
Thank You!

--
With Best Regards,
Andy Shevchenko