gpio: add API to be strict about GPIO IRQ usage

It is currently often possible in many GPIO drivers to request
a GPIO line to be used as IRQ after calling gpio_to_irq() and,
as the gpiolib is not aware of this, set the same line to
output and start driving it, with undesired side effects.

As it is a bogus usage scenario to request a line flagged as
output to used as IRQ, we introduce APIs to let gpiolib track
the use of a line as IRQ, and also set this flag from the
userspace ABI.

The API is symmetric so that lines can also be flagged from
.irq_enable() and unflagged from IRQ by .irq_disable().
The debugfs file is altered so that we see if a line is
reserved for IRQ.

Cc: Enric Balletbo i Serra <eballetbo@gmail.com>
Cc: Grant Likely <grant.likely@linaro.org>
Cc: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
Cc: Santosh Shilimkar <santosh.shilimkar@ti.com>
Acked-by: Alexandre Courbot <acourbot@nvidia.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
Reviewed-by: Javier Martinez Canillas <javier.martinez@collabora.co.uk>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>

1 files changed, 12 insertions, 0 deletions

diff --git a/include/linux/gpio.h b/include/linux/gpio.h
index 552e3f4..a06ec3e 100644
--- a/include/linux/gpio.h
+++ b/include/linux/gpio.h
@@ -204,6 +204,18 @@ static inline int gpio_to_irq(unsigned gpio)
 	return -EINVAL;
 }
 
+static inline int gpio_lock_as_irq(struct gpio_chip *chip, unsigned int offset)
+{
+	WARN_ON(1);
+	return -EINVAL;
+}
+
+static inline void gpio_unlock_as_irq(struct gpio_chip *chip,
+				      unsigned int offset)
+{
+	WARN_ON(1);
+}
+
 static inline int irq_to_gpio(unsigned irq)
 {
 	/* irq can never have been returned from gpio_to_irq() */