On Mon, Nov 13, 2023 at 02:12:48PM +0200, Andy Shevchenko wrote: > On Fri, Nov 03, 2023 at 07:57:38AM +0200, Mika Westerberg wrote: > > On Mon, Oct 30, 2023 at 04:10:33PM +0200, Andy Shevchenko wrote: > > ... > > > > +config PINCTRL_INTEL_PLATFORM > > > + tristate "Intel pinctrl and GPIO platform driver" > > > + depends on ACPI > > > + select PINCTRL_INTEL > > > + help > > > + This pinctrl driver provides an interface that allows configuring > > > + of Intel PCH pins and using them as GPIOs. > > > > Add here some description that explains why this needs to be enabled, > > for example for Lunar Lake. Now it is all too generic for distro folks > > to understand if this is needed or not. > > OK! > > ... > > > > + * Copyright (C) 2021-2023, Intel Corporation > > > > That's 2023 > > As-is it is still valid and reflects the history. > > ... > > > > + ngpps = device_get_child_node_count(dev); > > > + if (ngpps == 0) > > > > if (!nggps) > > 0 is a plain number here (as count) and explicit comparison makes sense. > But I'm okay with another form. > > > > > + return -ENODEV; > > ... > > > > + ncommunities = 1, > > > > Why this is 1? Can't we have more communities? > > As for now (version 1.0 of the specification) it's assumed that it's one > community per device node in the ACPI, so I would leave this as is (we have > also drivers with single community per device node, hence this is kinda > pattern. Should I add a comment? > Yes, I think it warrants a comment. > ... > > > > + struct device *dev = &pdev->dev; > > > + struct intel_pinctrl_soc_data *data; > > > > > > Change the ordering of the above: > > > > struct intel_pinctrl_soc_data *data; > > struct device *dev = &pdev->dev; > > Sure. > > ... > > > > +static const struct acpi_device_id intel_platform_pinctrl_acpi_match[] = { > > > + { } > > > > And add the _CID here in this patch as I commented in the last patch. > > OK! I'll squash the next patch into this one. > > > > +}; > > -- > With Best Regards, > Andy Shevchenko >