Re: [PATCH v3 00/10] Remove support for platform data from samsung keypad

On Sun, Jul 12, 2026 at 02:30:50PM +0200, Krzysztof Kozlowski wrote:

On Sat, Jul 11, 2026 at 07:12:41PM -0700, Dmitry Torokhov wrote:

quoted

Hi,

This series converts the Samsung keypad driver and the Cragganmore 6410
board to generic device properties, allowing the removal of legacy platform
data and GPIO lookup tables.

We rework the Samsung keypad driver to stop using platform data, utilizing
a new compact matrix keypad binding. To support this on the Cragganmore
board, we add infrastructure to associate software nodes with Samsung
gpio_chips, and add software node propagation support to the wm831x MFD
core to route DVS GPIO to the PMIC regulator.

Finally, we convert all peripherals on the Cragganmore board (including
keypad, GPIO keys, PMIC, WM1250, SPI0, basic-mmio-gpio, and LEDs) to
software properties. This allows us to eliminate all legacy GPIO lookup
tables and fixes incorrect GPIO chip names from previous conversions.

Mark, it would be great if you could give this a spin on Cragganmore.

The majority of changes are in arch/board code, with a couple of input
matches and one for MFD. If everyone agrees maybe we can merge it all
through the soc tree?

Yes, I could take it via samsung-soc. I assume I have your acks for
unput parts.

Yes, of course.

Regarding Sashiko review:

- IRQ concerns are pre-existing and will be addressed in a separate
  series

- TOCTOU with device tree overlay changing DT properties in the middle
  of the probe I will not attempt to address. If this is a real concern
  it should be addressed globally 

- Introduction of 5ms default debounce on proximity and suspend seem
  harmless

- Copying of MFD cells is necessary because we need to "mix in" software
  nodes from platform data. I originally tried to handle this is the
  regulator, but the result was worse as it needed to add software node
  to the device after it was created. Doing it in the core is much
  cleaner overall.

Thanks.

-- 
Dmitry