[PATCH 1/1] gpio/omap: use gpiolib API to mark a GPIO used as an IRQ

[PATCH 1/1] gpio/omap: use gpiolib API to mark a GPIO used as an IRQ

[Javier Martinez Canillas]

The OMAP GPIO driver keeps track about GPIO pins that are
used as IRQ lines for two reasons:

1) To prevent GPIO banks to be disabled while one of their
   GPIO pins are only used as an interrupt line.

2) To not allow another caller to set the GPIO pin as output.

Now gpiolib has an API to mark GPIO pins as used as IRQ lines
so the GPIO core only allows to set as output GPIO pins not
tied to an IRQ. So there is no need to have custom code for 2).

The IRQ usage still has to be maintained locally for 1) though.

Signed-off-by: Javier Martinez Canillas <javier.martinez@collabora.co.uk>
---
 drivers/gpio/gpio-omap.c | 20 ++++++++++----------
 1 file changed, 10 insertions(+), 10 deletions(-)

diff --git a/drivers/gpio/gpio-omap.c b/drivers/gpio/gpio-omap.c
index 89675f8..f319c9f 100644
--- a/drivers/gpio/gpio-omap.c
+++ b/drivers/gpio/gpio-omap.c
@@ -514,6 +514,14 @@ static int gpio_irq_type(struct irq_data *d, unsigned type)
 		return -EINVAL;
 	}
 
+	retval = gpio_lock_as_irq(&bank->chip, offset);
+	if (retval) {
+		dev_err(bank->dev, "unable to lock offset %d for IRQ\n",
+			offset);
+		spin_unlock_irqrestore(&bank->lock, flags);
+		return retval;
+	}
+
 	bank->irq_usage |= 1 << GPIO_INDEX(bank, gpio);
 	spin_unlock_irqrestore(&bank->lock, flags);
 
@@ -797,6 +805,7 @@ static void gpio_irq_shutdown(struct irq_data *d)
 	unsigned offset = GPIO_INDEX(bank, gpio);
 
 	spin_lock_irqsave(&bank->lock, flags);
+	gpio_unlock_as_irq(&bank->chip, offset);
 	bank->irq_usage &= ~(1 << offset);
 	_disable_gpio_module(bank, offset);
 	_reset_gpio(bank, gpio);
@@ -957,22 +966,13 @@ static int gpio_output(struct gpio_chip *chip, unsigned offset, int value)
 {
 	struct gpio_bank *bank;
 	unsigned long flags;
-	int retval = 0;
 
 	bank = container_of(chip, struct gpio_bank, chip);
 	spin_lock_irqsave(&bank->lock, flags);
-
-	if (LINE_USED(bank->irq_usage, offset)) {
-			retval = -EINVAL;
-			goto exit;
-	}
-
 	bank->set_dataout(bank, offset, value);
 	_set_gpio_direction(bank, offset, 0);
-
-exit:
 	spin_unlock_irqrestore(&bank->lock, flags);
-	return retval;
+	return 0;
 }
 
 static int gpio_debounce(struct gpio_chip *chip, unsigned offset,
-- 
1.8.4.rc3

Re: [PATCH 1/1] gpio/omap: use gpiolib API to mark a GPIO used as an IRQ

[Linus Walleij]

On Wed, Oct 16, 2013 at 2:47 AM, Javier Martinez Canillas
<javier.martinez@collabora.co.uk> wrote:

> The OMAP GPIO driver keeps track about GPIO pins that are
> used as IRQ lines for two reasons:
>
> 1) To prevent GPIO banks to be disabled while one of their
>    GPIO pins are only used as an interrupt line.
>
> 2) To not allow another caller to set the GPIO pin as output.
>
> Now gpiolib has an API to mark GPIO pins as used as IRQ lines
> so the GPIO core only allows to set as output GPIO pins not
> tied to an IRQ. So there is no need to have custom code for 2).
>
> The IRQ usage still has to be maintained locally for 1) though.
>
> Signed-off-by: Javier Martinez Canillas <javier.martinez@collabora.co.uk>

Patch applied. I had to merge in v3.12-rc4 to get the dependency
fixes, so I guess you did something similar when developing
this...

Yours,
Linus Walleij

Re: [PATCH 1/1] gpio/omap: use gpiolib API to mark a GPIO used as an IRQ

[Javier Martinez Canillas]

On 10/16/2013 10:06 AM, Linus Walleij wrote:
> On Wed, Oct 16, 2013 at 2:47 AM, Javier Martinez Canillas
> <javier.martinez@collabora.co.uk> wrote:
> 
>> The OMAP GPIO driver keeps track about GPIO pins that are
>> used as IRQ lines for two reasons:
>>
>> 1) To prevent GPIO banks to be disabled while one of their
>>    GPIO pins are only used as an interrupt line.
>>
>> 2) To not allow another caller to set the GPIO pin as output.
>>
>> Now gpiolib has an API to mark GPIO pins as used as IRQ lines
>> so the GPIO core only allows to set as output GPIO pins not
>> tied to an IRQ. So there is no need to have custom code for 2).
>>
>> The IRQ usage still has to be maintained locally for 1) though.
>>
>> Signed-off-by: Javier Martinez Canillas <javier.martinez@collabora.co.uk>
> 
> Patch applied. I had to merge in v3.12-rc4 to get the dependency
> fixes, so I guess you did something similar when developing
> this...
>

Thanks and yes since this is not a critical bug it was meant to be for v3.13 on
top of fac7fa162 ("gpio/omap: auto-setup a GPIO when used as an IRQ") and after
you have merged your flag-irqs branch.

Sorry for not explicitly mentioning this dependency.

> Yours,
> Linus Walleij
> 

Best regards,
Javier