Re: [PATCH v4 01/20] platform/x86: x86-android-tablets: convert Goodix devices to GPIO references

[Andy Shevchenko <andriy.shevchenko@intel.com>]

On Wed, Feb 11, 2026 at 07:44:05PM -0800, Dmitry Torokhov wrote:
> On Tue, Feb 10, 2026 at 01:53:11AM -0800, Bartosz Golaszewski wrote:
> > On Mon, 9 Feb 2026 21:41:49 +0100, Dmitry Torokhov
> > <dmitry.torokhov@gmail.com> said:

...

> > Agreed, I'm not advocating for using lookup tables. I'm advocating for using
> > swnode correctly.
> 
> I am curious how would you approach to fixing
> drivers/platform/x86/meraki-mx100.c ? The GPIOs in question belong to
> "gpio_ich" which is a cell of a MFD PCI device (lpc_ich).
> 
> Maybe it has a standard ACPI HID, Andy, do you know?

For PCI devices the ACPI companion may or may not be present using _ADR()
method. And it's an optional, so some BIOSes might have it, but my gut
feelings that the majority don't provide that. It means there is no associated
firmware node.

To solve this one may think of adding the swnode to the GPIO device by default
at registration, the problem is that fwnode is a single linked list of maximum
two elements, and we have already GPIO drivers that use firmware node _and_
software node (Intel Galileo boards with Synopsys DesignWare GPIO IP).

To me it looks like we again bump into the very old design problem of fwnode.
The roadmap to fix that is:
- make sure we never ever dereference fwnode/of_node of struct device anywhere
except the driver core
- convert fwnode to be a (double linked) list of more than two elements
- decouple of_node/fwnode from struct device, so it will only have a list head
- introduce a simple priority for nodes, the head for the firmware, the tail
for the software nodes
- kill primary/secondary fwnode approach for good


-- 
With Best Regards,
Andy Shevchenko