[v2] gpio: acpi: Call enable_irq_wake for _IAE GpioInts with Wake set

Submitted by Hans de Goede on March 20, 2017, 5:32 p.m.

Details

Message ID 20170320173221.3397-1-hdegoede@redhat.com
State New
Headers show

Commit Message

Hans de Goede March 20, 2017, 5:32 p.m.
On Bay Trail / Cherry Trail systems with a LID switch, the LID switch is
often connect to a gpioint handled by an _IAE event handler.
Before this commit such systems would not wake up when opening the lid,
requiring the powerbutton to be pressed after opening the lid to wakeup.

Note that Bay Trail / Cherry Trail systems use suspend-to-idle, so
the interrupts are generated anyway on those lines on lid switch changes,
but they are treated by the IRQ subsystem as spurious while suspended if
not marked as wakeup IRQs.

This commit calls enable_irq_wake() for _IAE GpioInts with a valid
event handler which have their Wake flag set. This fixes such systems
not waking up when opening the lid.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Mika Westerberg <mika.westerberg@linux.intel.com>
---
Changes in v2:
-Improve commit msg
-Add Mika's Acked-by
---
 drivers/gpio/gpiolib-acpi.c | 9 +++++++++
 1 file changed, 9 insertions(+)

Comments

Rafael J. Wysocki March 20, 2017, 9:59 p.m.
On Monday, March 20, 2017 06:32:21 PM Hans de Goede wrote:
> On Bay Trail / Cherry Trail systems with a LID switch, the LID switch is
> often connect to a gpioint handled by an _IAE event handler.
> Before this commit such systems would not wake up when opening the lid,
> requiring the powerbutton to be pressed after opening the lid to wakeup.
> 
> Note that Bay Trail / Cherry Trail systems use suspend-to-idle, so
> the interrupts are generated anyway on those lines on lid switch changes,
> but they are treated by the IRQ subsystem as spurious while suspended if
> not marked as wakeup IRQs.
> 
> This commit calls enable_irq_wake() for _IAE GpioInts with a valid
> event handler which have their Wake flag set. This fixes such systems
> not waking up when opening the lid.
> 
> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
> Acked-by: Mika Westerberg <mika.westerberg@linux.intel.com>
> ---
> Changes in v2:
> -Improve commit msg
> -Add Mika's Acked-by
> ---
>  drivers/gpio/gpiolib-acpi.c | 9 +++++++++
>  1 file changed, 9 insertions(+)
> 
> diff --git a/drivers/gpio/gpiolib-acpi.c b/drivers/gpio/gpiolib-acpi.c
> index 8cd3f66..18207b2 100644
> --- a/drivers/gpio/gpiolib-acpi.c
> +++ b/drivers/gpio/gpiolib-acpi.c
> @@ -28,6 +28,7 @@ struct acpi_gpio_event {
>  	acpi_handle handle;
>  	unsigned int pin;
>  	unsigned int irq;
> +	bool irq_wake_enabled;
>  	struct gpio_desc *desc;
>  };
>  
> @@ -266,6 +267,11 @@ static acpi_status acpi_gpiochip_request_interrupt(struct acpi_resource *ares,
>  		goto fail_free_event;
>  	}
>  
> +	if (agpio->wake_capable == ACPI_WAKE_CAPABLE) {
> +		enable_irq_wake(irq);
> +		event->irq_wake_enabled = true;
> +	}
> +
>  	list_add_tail(&event->node, &acpi_gpio->events);
>  	return AE_OK;
>  
> @@ -339,6 +345,9 @@ void acpi_gpiochip_free_interrupts(struct gpio_chip *chip)
>  	list_for_each_entry_safe_reverse(event, ep, &acpi_gpio->events, node) {
>  		struct gpio_desc *desc;
>  
> +		if (event->irq_wake_enabled)

It has just occurred to me that if the event is in the list, the IRQ will be
enabled to wake up as long as agpio->wake_capable == ACPI_WAKE_CAPABLE, so it
looks like it should be sufficient to check

	if (agpio->wake_capable == ACPI_WAKE_CAPABLE)

here and the new flag is not necessary.  Or is it?

> +			disable_irq_wake(event->irq);
> +
>  		free_irq(event->irq, event);
>  		desc = event->desc;
>  		if (WARN_ON(IS_ERR(desc)))
> 

Thanks,
Rafael

--
To unsubscribe from this list: send the line "unsubscribe linux-gpio" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Andy Shevchenko March 21, 2017, 12:30 p.m.
On Mon, 2017-03-20 at 18:32 +0100, Hans de Goede wrote:
> On Bay Trail / Cherry Trail systems with a LID switch, the LID switch
> is
> often connect to a gpioint handled by an _IAE event handler.
> Before this commit such systems would not wake up when opening the
> lid,
> requiring the powerbutton to be pressed after opening the lid to
> wakeup.
> 
> Note that Bay Trail / Cherry Trail systems use suspend-to-idle, so
> the interrupts are generated anyway on those lines on lid switch
> changes,
> but they are treated by the IRQ subsystem as spurious while suspended
> if
> not marked as wakeup IRQs.
> 
> This commit calls enable_irq_wake() for _IAE GpioInts with a valid
> event handler which have their Wake flag set. This fixes such systems
> not waking up when opening the lid.
> 
> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
> Acked-by: Mika Westerberg <mika.westerberg@linux.intel.com>
> 

> +	bool irq_wake_enabled;
> 

> +		if (event->irq_wake_enabled)
> 

Same (new) comment as in v1.
Rafael J. Wysocki March 21, 2017, 12:36 p.m.
On Tue, Mar 21, 2017 at 1:30 PM, Andy Shevchenko
<andriy.shevchenko@linux.intel.com> wrote:
> On Mon, 2017-03-20 at 18:32 +0100, Hans de Goede wrote:
>> On Bay Trail / Cherry Trail systems with a LID switch, the LID switch
>> is
>> often connect to a gpioint handled by an _IAE event handler.
>> Before this commit such systems would not wake up when opening the
>> lid,
>> requiring the powerbutton to be pressed after opening the lid to
>> wakeup.
>>
>> Note that Bay Trail / Cherry Trail systems use suspend-to-idle, so
>> the interrupts are generated anyway on those lines on lid switch
>> changes,
>> but they are treated by the IRQ subsystem as spurious while suspended
>> if
>> not marked as wakeup IRQs.
>>
>> This commit calls enable_irq_wake() for _IAE GpioInts with a valid
>> event handler which have their Wake flag set. This fixes such systems
>> not waking up when opening the lid.
>>
>> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
>> Acked-by: Mika Westerberg <mika.westerberg@linux.intel.com>
>>
>
>> +     bool irq_wake_enabled;
>>
>
>> +             if (event->irq_wake_enabled)
>>
>
> Same (new) comment as in v1.

Why not just check for agpio->wake_capable == ACPI_WAKE_CAPABLE
instead as I said?

Thanks,
Rafael
--
To unsubscribe from this list: send the line "unsubscribe linux-gpio" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Andy Shevchenko March 21, 2017, 12:40 p.m.
On Tue, 2017-03-21 at 13:36 +0100, Rafael J. Wysocki wrote:
> On Tue, Mar 21, 2017 at 1:30 PM, Andy Shevchenko
> <andriy.shevchenko@linux.intel.com> wrote:
> > On Mon, 2017-03-20 at 18:32 +0100, Hans de Goede wrote:

> > > +     bool irq_wake_enabled;
> > > 
> > > +             if (event->irq_wake_enabled)
> > > 
> > 
> > Same (new) comment as in v1.
> 
> Why not just check for agpio->wake_capable == ACPI_WAKE_CAPABLE
> instead as I said?

Ah, yes, it could work as well as far as I can see from the code point
of view.
Hans de Goede March 24, 2017, 9:51 a.m.
Hi,

On 20-03-17 22:59, Rafael J. Wysocki wrote:
> On Monday, March 20, 2017 06:32:21 PM Hans de Goede wrote:
>> On Bay Trail / Cherry Trail systems with a LID switch, the LID switch is
>> often connect to a gpioint handled by an _IAE event handler.
>> Before this commit such systems would not wake up when opening the lid,
>> requiring the powerbutton to be pressed after opening the lid to wakeup.
>>
>> Note that Bay Trail / Cherry Trail systems use suspend-to-idle, so
>> the interrupts are generated anyway on those lines on lid switch changes,
>> but they are treated by the IRQ subsystem as spurious while suspended if
>> not marked as wakeup IRQs.
>>
>> This commit calls enable_irq_wake() for _IAE GpioInts with a valid
>> event handler which have their Wake flag set. This fixes such systems
>> not waking up when opening the lid.
>>
>> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
>> Acked-by: Mika Westerberg <mika.westerberg@linux.intel.com>
>> ---
>> Changes in v2:
>> -Improve commit msg
>> -Add Mika's Acked-by
>> ---
>>  drivers/gpio/gpiolib-acpi.c | 9 +++++++++
>>  1 file changed, 9 insertions(+)
>>
>> diff --git a/drivers/gpio/gpiolib-acpi.c b/drivers/gpio/gpiolib-acpi.c
>> index 8cd3f66..18207b2 100644
>> --- a/drivers/gpio/gpiolib-acpi.c
>> +++ b/drivers/gpio/gpiolib-acpi.c
>> @@ -28,6 +28,7 @@ struct acpi_gpio_event {
>>  	acpi_handle handle;
>>  	unsigned int pin;
>>  	unsigned int irq;
>> +	bool irq_wake_enabled;
>>  	struct gpio_desc *desc;
>>  };
>>
>> @@ -266,6 +267,11 @@ static acpi_status acpi_gpiochip_request_interrupt(struct acpi_resource *ares,
>>  		goto fail_free_event;
>>  	}
>>
>> +	if (agpio->wake_capable == ACPI_WAKE_CAPABLE) {
>> +		enable_irq_wake(irq);
>> +		event->irq_wake_enabled = true;
>> +	}
>> +
>>  	list_add_tail(&event->node, &acpi_gpio->events);
>>  	return AE_OK;
>>
>> @@ -339,6 +345,9 @@ void acpi_gpiochip_free_interrupts(struct gpio_chip *chip)
>>  	list_for_each_entry_safe_reverse(event, ep, &acpi_gpio->events, node) {
>>  		struct gpio_desc *desc;
>>
>> +		if (event->irq_wake_enabled)
>
> It has just occurred to me that if the event is in the list, the IRQ will be
> enabled to wake up as long as agpio->wake_capable == ACPI_WAKE_CAPABLE, so it
> looks like it should be sufficient to check
>
> 	if (agpio->wake_capable == ACPI_WAKE_CAPABLE)
>
> here and the new flag is not necessary.  Or is it?

We don't have (readily available) access to the acpi_resource_gpio struct
in acpi_gpiochip_free_interrupts, so I'm going to go with Andy's suggestion
instead and change the if to:

                 if (irqd_is_wakeup_set(irq_get_irq_data(event->irq)))
                         disable_irq_wake(event->irq);

Still need to run some quick tests and then I will send v3 with this
change.

Regards,

Hans
--
To unsubscribe from this list: send the line "unsubscribe linux-gpio" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Rafael J. Wysocki March 24, 2017, 3:16 p.m.
On Friday, March 24, 2017 10:51:53 AM Hans de Goede wrote:
> Hi,
> 
> On 20-03-17 22:59, Rafael J. Wysocki wrote:
> > On Monday, March 20, 2017 06:32:21 PM Hans de Goede wrote:
> >> On Bay Trail / Cherry Trail systems with a LID switch, the LID switch is
> >> often connect to a gpioint handled by an _IAE event handler.
> >> Before this commit such systems would not wake up when opening the lid,
> >> requiring the powerbutton to be pressed after opening the lid to wakeup.
> >>
> >> Note that Bay Trail / Cherry Trail systems use suspend-to-idle, so
> >> the interrupts are generated anyway on those lines on lid switch changes,
> >> but they are treated by the IRQ subsystem as spurious while suspended if
> >> not marked as wakeup IRQs.
> >>
> >> This commit calls enable_irq_wake() for _IAE GpioInts with a valid
> >> event handler which have their Wake flag set. This fixes such systems
> >> not waking up when opening the lid.
> >>
> >> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
> >> Acked-by: Mika Westerberg <mika.westerberg@linux.intel.com>
> >> ---
> >> Changes in v2:
> >> -Improve commit msg
> >> -Add Mika's Acked-by
> >> ---
> >>  drivers/gpio/gpiolib-acpi.c | 9 +++++++++
> >>  1 file changed, 9 insertions(+)
> >>
> >> diff --git a/drivers/gpio/gpiolib-acpi.c b/drivers/gpio/gpiolib-acpi.c
> >> index 8cd3f66..18207b2 100644
> >> --- a/drivers/gpio/gpiolib-acpi.c
> >> +++ b/drivers/gpio/gpiolib-acpi.c
> >> @@ -28,6 +28,7 @@ struct acpi_gpio_event {
> >>  	acpi_handle handle;
> >>  	unsigned int pin;
> >>  	unsigned int irq;
> >> +	bool irq_wake_enabled;
> >>  	struct gpio_desc *desc;
> >>  };
> >>
> >> @@ -266,6 +267,11 @@ static acpi_status acpi_gpiochip_request_interrupt(struct acpi_resource *ares,
> >>  		goto fail_free_event;
> >>  	}
> >>
> >> +	if (agpio->wake_capable == ACPI_WAKE_CAPABLE) {
> >> +		enable_irq_wake(irq);
> >> +		event->irq_wake_enabled = true;
> >> +	}
> >> +
> >>  	list_add_tail(&event->node, &acpi_gpio->events);
> >>  	return AE_OK;
> >>
> >> @@ -339,6 +345,9 @@ void acpi_gpiochip_free_interrupts(struct gpio_chip *chip)
> >>  	list_for_each_entry_safe_reverse(event, ep, &acpi_gpio->events, node) {
> >>  		struct gpio_desc *desc;
> >>
> >> +		if (event->irq_wake_enabled)
> >
> > It has just occurred to me that if the event is in the list, the IRQ will be
> > enabled to wake up as long as agpio->wake_capable == ACPI_WAKE_CAPABLE, so it
> > looks like it should be sufficient to check
> >
> > 	if (agpio->wake_capable == ACPI_WAKE_CAPABLE)
> >
> > here and the new flag is not necessary.  Or is it?
> 
> We don't have (readily available) access to the acpi_resource_gpio struct
> in acpi_gpiochip_free_interrupts, so I'm going to go with Andy's suggestion
> instead and change the if to:
> 
>                  if (irqd_is_wakeup_set(irq_get_irq_data(event->irq)))
>                          disable_irq_wake(event->irq);
> 
> Still need to run some quick tests and then I will send v3 with this
> change.

OK, fair enough.

Thanks,
Rafael

--
To unsubscribe from this list: send the line "unsubscribe linux-gpio" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Patch hide | download patch | download mbox

diff --git a/drivers/gpio/gpiolib-acpi.c b/drivers/gpio/gpiolib-acpi.c
index 8cd3f66..18207b2 100644
--- a/drivers/gpio/gpiolib-acpi.c
+++ b/drivers/gpio/gpiolib-acpi.c
@@ -28,6 +28,7 @@  struct acpi_gpio_event {
 	acpi_handle handle;
 	unsigned int pin;
 	unsigned int irq;
+	bool irq_wake_enabled;
 	struct gpio_desc *desc;
 };
 
@@ -266,6 +267,11 @@  static acpi_status acpi_gpiochip_request_interrupt(struct acpi_resource *ares,
 		goto fail_free_event;
 	}
 
+	if (agpio->wake_capable == ACPI_WAKE_CAPABLE) {
+		enable_irq_wake(irq);
+		event->irq_wake_enabled = true;
+	}
+
 	list_add_tail(&event->node, &acpi_gpio->events);
 	return AE_OK;
 
@@ -339,6 +345,9 @@  void acpi_gpiochip_free_interrupts(struct gpio_chip *chip)
 	list_for_each_entry_safe_reverse(event, ep, &acpi_gpio->events, node) {
 		struct gpio_desc *desc;
 
+		if (event->irq_wake_enabled)
+			disable_irq_wake(event->irq);
+
 		free_irq(event->irq, event);
 		desc = event->desc;
 		if (WARN_ON(IS_ERR(desc)))