Re: [PATCH V2 10/12] gpio: tegra: implement wake event support for Tegra210 and prior GPIO

[Sowjanya Komatineni]

On 5/29/19 7:03 AM, Thierry Reding wrote:
> On Tue, May 28, 2019 at 04:08:54PM -0700, Sowjanya Komatineni wrote:
> > The GPIO controller doesn't have any controls to enable the system to
> > wake up from low power states based on activity on GPIO pins. An extra
> > hardware block that is part of the power management controller (PMC)
> > contains these controls. In order for the GPIO controller to be able
> > to cooperate with the PMC, obtain a reference to the PMC's IRQ domain
> > and make it a parent to the GPIO controller's IRQ domain. This way the
> > PMC gets an opportunity to program the additional registers required
> > to enable wakeup sources on suspend.
> >
> > Signed-off-by: Sowjanya Komatineni <skomatineni@xxxxxxxxxx>
> > ---
> >   drivers/gpio/gpio-tegra.c | 116 +++++++++++++++++++++++++++++++++++++++++++---
> >   1 file changed, 110 insertions(+), 6 deletions(-)
> >
> > diff --git a/drivers/gpio/gpio-tegra.c b/drivers/gpio/gpio-tegra.c
> > index 6d9b6906b9d0..5190129668d3 100644
> > --- a/drivers/gpio/gpio-tegra.c
> > +++ b/drivers/gpio/gpio-tegra.c
> > @@ -32,6 +32,8 @@
> >   #include <linux/pinctrl/consumer.h>
> >   #include <linux/pm.h>
> >   
> > +#include <dt-bindings/interrupt-controller/arm-gic.h>
> > +
> >   #define GPIO_BANK(x)		((x) >> 5)
> >   #define GPIO_PORT(x)		(((x) >> 3) & 0x3)
> >   #define GPIO_BIT(x)		((x) & 0x7)
> > @@ -275,8 +277,22 @@ static int tegra_gpio_set_config(struct gpio_chip *chip, unsigned int offset,
> >   static int tegra_gpio_to_irq(struct gpio_chip *chip, unsigned int offset)
> >   {
> >   	struct tegra_gpio_info *tgi = gpiochip_get_data(chip);
> > +	struct irq_domain *domain = tgi->irq_domain;
> > +
> > +	if (!gpiochip_irqchip_irq_valid(chip, offset))
> > +		return -ENXIO;
> > +
> > +	if (irq_domain_is_hierarchy(domain)) {
> > +		struct irq_fwspec spec;
> > +
> > +		spec.fwnode = domain->fwnode;
> > +		spec.param_count = 2;
> > +		spec.param[0] = offset;
> > +		spec.param[1] = IRQ_TYPE_NONE;
> > +		return irq_domain_alloc_irqs(domain, 1, NUMA_NO_NODE, &spec);
> This looks like it was copied from the equivalent Tegra186 patch. I have
> since then changed the implementation, based on feedback by Linus, to
> not call irq_domain_alloc_irqs() here and instead call
> irq_create_fwspec_mapping(). This has the advantage of not requiring the
> irq_domain_alloc_irqs() function to be exported. It ends up calling that
> function internally, but as discussed with Linus it's also a nicer way
> to create these mappings.
existing gpio-tegra driver maps hwirq to virtual interrupt number using 
irq_create_mapping during gpio probe

and irq_create_fwspec_mapping() will always return virq number as virq 
number exists already and irq_domain_alloc_irqs doesn't happen.

So I was using irq_domain_alloc_irqs().


> > +	}
> >   
> > -	return irq_find_mapping(tgi->irq_domain, offset);
> > +	return irq_find_mapping(domain, offset);
> >   }
> >   
> >   static void tegra_gpio_irq_ack(struct irq_data *d)
> > @@ -365,7 +381,10 @@ static int tegra_gpio_irq_set_type(struct irq_data *d, unsigned int type)
> >   	else if (type & (IRQ_TYPE_EDGE_FALLING | IRQ_TYPE_EDGE_RISING))
> >   		irq_set_handler_locked(d, handle_edge_irq);
> >   
> > -	return 0;
> > +	if (d->parent_data)
> > +		return irq_chip_set_type_parent(d, type);
> > +	else
> > +		return 0;
> There's no need for this final else. Just make it a regular "return 0;"
> at the end of the function, without the extra else branch.
>
> >   }
> >   
> >   static void tegra_gpio_irq_shutdown(struct irq_data *d)
> > @@ -503,6 +522,7 @@ static int tegra_gpio_irq_set_wake(struct irq_data *d, unsigned int enable)
> >   	struct tegra_gpio_bank *bank = irq_data_get_irq_chip_data(d);
> >   	unsigned int gpio = d->hwirq;
> >   	u32 port, bit, mask;
> > +	int ret;
> >   
> >   	port = GPIO_PORT(gpio);
> >   	bit = GPIO_BIT(gpio);
> > @@ -513,7 +533,14 @@ static int tegra_gpio_irq_set_wake(struct irq_data *d, unsigned int enable)
> >   	else
> >   		bank->wake_enb[port] &= ~mask;
> >   
> > -	return irq_set_irq_wake(bank->irq, enable);
> > +	ret = irq_set_irq_wake(bank->irq, enable);
> > +	if (ret < 0)
> > +		return ret;
> > +
> > +	if (d->parent_data)
> > +		return irq_chip_set_wake_parent(d, enable);
> > +	else
> > +		return 0;
> Same here.
>
> Thierry