pinctrl: make the CONFIG_PINCTRL=n gpio config stubs return -ENOTSUPP

With CONFIG_PINCTRL=n, pinctrl_gpio_get_config() and
pinctrl_gpio_set_config() are stubs that return 0. The getter reports
success without touching the config it was asked to fill, so a caller
reads the packed parameter it passed in back as if it were the pin's
state; the setter reports success for a configuration nobody applied.

With CONFIG_PINCTRL=y the same callers already get -ENOTSUPP from
gpiochip_generic_config() and gpiochip_generic_get_config() for a chip
without pin ranges, and gpiolib treats that value as "the backend cannot
do this" and carries on. Return it from the stubs too, so a kernel
without pinctrl behaves like a chip without pin ranges instead of
claiming success.

Suggested-by: Haibo Chen <haibo.chen@nxp.com>
Signed-off-by: Mehmet Fide <mehmet.fide@screeningeagle.com>
Reviewed-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>
Signed-off-by: Linus Walleij <linusw@kernel.org>

1 files changed, 2 insertions, 2 deletions

diff --git a/include/linux/pinctrl/consumer.h b/include/linux/pinctrl/consumer.h
index 11b8f0b8da0c..7be49d447426 100644
--- a/include/linux/pinctrl/consumer.h
+++ b/include/linux/pinctrl/consumer.h
@@ -107,14 +107,14 @@ static inline int
 pinctrl_gpio_get_config(struct gpio_chip *gc, unsigned int offset,
 			unsigned long *config)
 {
-	return 0;
+	return -ENOTSUPP;
 }
 
 static inline int
 pinctrl_gpio_set_config(struct gpio_chip *gc, unsigned int offset,
 			    unsigned long config)
 {
-	return 0;
+	return -ENOTSUPP;
 }
 
 static inline struct pinctrl * __must_check pinctrl_get(struct device *dev)