gpio: of: Build fails if CONFIG_OF_DYNAMIC enabled without CONFIG_OF_GPIO

The symbol 'gpio_of_notifier' doesn't exist without both CONFIG_OF_GPIO and CONFIG_OF_DYNAMIC enabled, but is referenced when only CONFIG_OF_DYNAMIC is enabled.

This broke building with 'make ARCH=um allyesconfig':
---------------
/usr/bin/ld: drivers/gpio/gpiolib.o: in function `gpiolib_dev_init':
./drivers/gpio/gpiolib.c:5293: undefined reference to `gpio_of_notifier'
collect2: error: ld returned 1 exit status
---------------

---
 drivers/gpio/gpiolib.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

index 40f2d7f69be2..5c292fb3355c 100644
@@ -5289,8 +5289,9 @@ static int __init gpiolib_dev_init(void)
 	gpiolib_initialized = true;
 	gpiochip_setup_devs();
 
-	if (IS_ENABLED(CONFIG_OF_DYNAMIC))
-		WARN_ON(of_reconfig_notifier_register(&gpio_of_notifier));
+#if IS_ENABLED(CONFIG_OF_DYNAMIC) && IS_ENABLED(CONFIG_OF_GPIO)
+	WARN_ON(of_reconfig_notifier_register(&gpio_of_notifier));
+#endif /* CONFIG_OF_DYNAMIC && CONFIG_OF_GPIO */
 
 	return ret;
 }

2.26.2.303.gf8c07b1a785-goog

Hi David,

Right, so you have CONFIG_OF=y, but CONFIG_OF_GPIO=n, as the latter depends on HAS_IOMEM, which is not set for UML.

Interestingly, the latter dependency claims to have been added because gpiolib-of.c uses ioremap()/iounmap(). However, it seems to have never called ioremap(), only of_iomap() and iounmap().
of_iomap() itself is available, as it depends on OF_ADDRESS, which depends on HAS_IOMEM || UML. Interestingly, of_iomap() calls ioremap(), without a dependency on HAS_IOMEM?

So perhaps the dependency of CONFIG_OF_GPIO on HAS_IOMEM can be dropped these days, solving this issue as well?

Anyway:

Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Patch applied for fixes since it fixes a regression, I recommend looking a bit down the dependency chain as indicated by Geert if you have time!

Yours,
Linus Walleij