On Fri, Jan 27, 2023, at 11:32, Nicolas Ferre wrote: > Hi Arnd, > > On 26/01/2023 at 14:50, Arnd Bergmann wrote: >> From: Arnd Bergmann <arnd@xxxxxxxx> >> >> All Atmel (now Microchip) machines boot using DT, so the >> old platform_data for this driver is no longer used by any >> boards. >> >> Removing the pdata probe lets us simplify the GPIO handling >> with the use of the descriptor API. > > Thanks for your patch. I would like to know if it's related to the > initiative started by Bala in this series: > > https://lore.kernel.org/all/20221226073908.17317-1-balamanikandan.gunasundar@xxxxxxxxxxxxx/ > > It's true that it didn't come to a conclusion yet... I hadn't seen that one, my patch was a somewhat older series to convert a couple of subsystems over from the old of_get_named_gpio(), with no special interest in this platform. It looks like the patches are fairly similar, with two differences I can see immediately: - I use the normal devm_gpiod_get_optional(), while Bala uses the fwnode variant that should not be needed here since the fwnode is the one for the platform device itself. - I use the normal gpiod_get_value(), which will handle polarity as specified in DT, while Bala's patch uses gpiod_get_raw_value(). The difference is what happens when the GPIO_ACTIVE_LOW flag is set in the DT, and the possible combinations with the "cd-inverted" flag. I did not test my version, so I assume Bala is correct here, but it would be good to review this carefully either way. Arnd
