[PATCH] gpio: fix up CONFIG_OF dependencies

[Bartosz Golaszewski]

On Wed, 25 Mar 2026 11:01:14 +0100, Arnd Bergmann <arnd at kernel.org> said:
> From: Arnd Bergmann <arnd at arndb.de>
>
> A number of GPIO drivers that used to have a CONFIG_OF_GPIO dependency now fail
> to build on targets without CONFIG_OF:
>
> WARNING: unmet direct dependencies detected for GPIO_SYSCON
>   Depends on [n]: GPIOLIB [=y] && HAS_IOMEM [=y] && MFD_SYSCON [=y] && OF [=n]
>   Selected by [y]:
>   - GPIO_SAMA5D2_PIOBU [=y] && GPIOLIB [=y] && HAS_IOMEM [=y] && MFD_SYSCON [=y] && (ARCH_AT91 || COMPILE_TEST [=y])
>
> drivers/gpio/gpio-mt7621.c: In function 'mediatek_gpio_bank_probe':
> drivers/gpio/gpio-mt7621.c:254:20: error: 'struct gpio_chip' has no member named 'of_gpio_n_cells'
>   254 |         rg->chip.gc.of_gpio_n_cells = 2;
>       |                    ^
> drivers/gpio/gpio-tegra186.c: In function 'tegra186_gpio_of_xlate':
> drivers/gpio/gpio-tegra186.c:502:25: error: 'struct gpio_chip' has no member named 'of_gpio_n_cells'
>   502 |         if (WARN_ON(chip->of_gpio_n_cells < 2))
>       |                         ^~
> drivers/gpio/gpio-lpc32xx.c: In function 'lpc32xx_gpio_probe':
> drivers/gpio/gpio-lpc32xx.c:523:49: error: 'struct gpio_chip' has no member named 'of_xlate'
>   523 |                         lpc32xx_gpiochip[i].chip.of_xlate = lpc32xx_of_xlate;
>       |                                                 ^
> drivers/gpio/gpio-spacemit-k1.c: In function 'spacemit_gpio_add_bank':
> drivers/gpio/gpio-spacemit-k1.c:234:11: error: 'struct gpio_chip' has no member named 'of_gpio_n_cells'
>   234 |         gc->of_gpio_n_cells     = 3;
>       |           ^~
>
> Bring that back as a dependency.
>
> Fixes: 7803501e5754 ("gpio: drop unneeded Kconfig dependencies on OF_GPIO")
> Signed-off-by: Arnd Bergmann <arnd at arndb.de>
> ---
> There may be more fallout from the change, this is just what I found immediately,
> but you could already fold the changes into the broken patch while I look
> for additional ones.
> ---

Thanks and sorry for the breakage. However, I'm wondering if it wouldn't make
sense to do the following:

diff --git a/include/linux/gpio/driver.h b/include/linux/gpio/driver.h
index 17511434ed077..ce0e0df68dbaa 100644
--- a/include/linux/gpio/driver.h
+++ b/include/linux/gpio/driver.h
@@ -473,7 +473,6 @@ struct gpio_chip {
 	struct gpio_irq_chip irq;
 #endif /* CONFIG_GPIOLIB_IRQCHIP */

-#if defined(CONFIG_OF_GPIO)
 	/*
 	 * If CONFIG_OF_GPIO is enabled, then all GPIO controllers described in
 	 * the device tree automatically may have an OF translation
@@ -516,7 +515,6 @@ struct gpio_chip {
 	 */
 	int (*of_xlate)(struct gpio_chip *gc,
 			const struct of_phandle_args *gpiospec, u32 *flags);
-#endif /* CONFIG_OF_GPIO */
 };

 char *gpiochip_dup_line_label(struct gpio_chip *gc, unsigned int offset);

Symbols from linux/of.h are stubbed out and these drivers can build just fine
with !CONFIG_OF. This would naturally increase the build coverage.

Bartosz