Message ID | 20230216-gpio-at91-immutable-v1-1-44f52f148ab9@kernel.org |
---|---|
State | New |
Headers | show |
Series | pinctrl: at91: Minor cleanups | expand |
On Thu, Feb 16, 2023 at 4:49 PM Mark Brown <broonie@kernel.org> wrote: > To help gpiolib not fiddle around with the internals of the irqchip > flag the chip as immutable, adding the calls into the gpiolib core > required to do so. > > Signed-off-by: Mark Brown <broonie@kernel.org> 1) I'm impressed that you're using AT91 hardware 2) Can you respin this on top of my pinctrl devel branch: https://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-pinctrl.git/log/?h=devel There are some Andy-cleanups already queued for AT91 so I am a bit worried of collisions. (If you feel confident they are orthogonal just use v6.3-rc1) Yours, Linus Walleij
On Mon, Mar 06, 2023 at 02:20:56PM +0100, Linus Walleij wrote: > 2) Can you respin this on top of my pinctrl devel branch: > https://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-pinctrl.git/log/?h=devel > There are some Andy-cleanups already queued for AT91 so I am a bit > worried of collisions. (If you feel confident they are orthogonal just > use v6.3-rc1) Argh, I'd already started testing with a rebase onto 6.3-rc1. I've restarted for your devel branch (neither rebase caused any conflicts).
On 06/03/2023 at 18:50, Linus Walleij wrote: > EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe > > On Thu, Feb 16, 2023 at 4:49 PM Mark Brown <broonie@kernel.org> wrote: > >> To help gpiolib not fiddle around with the internals of the irqchip >> flag the chip as immutable, adding the calls into the gpiolib core >> required to do so. >> >> Signed-off-by: Mark Brown <broonie@kernel.org> > > 1) I'm impressed that you're using AT91 hardware I'm delighted that you're using AT91 hardware ;-) Thanks for your help Mark! Regards, Nicolas > 2) Can you respin this on top of my pinctrl devel branch: > https://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-pinctrl.git/log/?h=devel > There are some Andy-cleanups already queued for AT91 so I am a bit > worried of collisions. (If you feel confident they are orthogonal just > use v6.3-rc1) > > Yours, > Linus Walleij
diff --git a/drivers/pinctrl/pinctrl-at91.c b/drivers/pinctrl/pinctrl-at91.c index 1e1813d7c550..8ecf52ec9b9b 100644 --- a/drivers/pinctrl/pinctrl-at91.c +++ b/drivers/pinctrl/pinctrl-at91.c @@ -1536,6 +1536,20 @@ static void at91_gpio_dbg_show(struct seq_file *s, struct gpio_chip *chip) #define at91_gpio_dbg_show NULL #endif +static int gpio_irq_request_resources(struct irq_data *d) +{ + struct at91_gpio_chip *at91_gpio = irq_data_get_irq_chip_data(d); + + return gpiochip_lock_as_irq(&at91_gpio->chip, irqd_to_hwirq(d)); +} + +static void gpio_irq_release_resources(struct irq_data *d) +{ + struct at91_gpio_chip *at91_gpio = irq_data_get_irq_chip_data(d); + + gpiochip_unlock_as_irq(&at91_gpio->chip, irqd_to_hwirq(d)); +} + /* Several AIC controller irqs are dispatched through this GPIO handler. * To use any AT91_PIN_* as an externally triggered IRQ, first call * at91_set_gpio_input() then maybe enable its glitch filter. @@ -1555,6 +1569,9 @@ static void gpio_irq_mask(struct irq_data *d) struct at91_gpio_chip *at91_gpio = irq_data_get_irq_chip_data(d); void __iomem *pio = at91_gpio->regbase; unsigned mask = 1 << d->hwirq; + unsigned gpio = irqd_to_hwirq(d); + + gpiochip_disable_irq(&at91_gpio->chip, gpio); if (pio) writel_relaxed(mask, pio + PIO_IDR); @@ -1565,6 +1582,9 @@ static void gpio_irq_unmask(struct irq_data *d) struct at91_gpio_chip *at91_gpio = irq_data_get_irq_chip_data(d); void __iomem *pio = at91_gpio->regbase; unsigned mask = 1 << d->hwirq; + unsigned gpio = irqd_to_hwirq(d); + + gpiochip_enable_irq(&at91_gpio->chip, gpio); if (pio) writel_relaxed(mask, pio + PIO_IER); @@ -1731,12 +1751,15 @@ static int at91_gpio_of_irq_setup(struct platform_device *pdev, at91_gpio->pioc_hwirq = irqd_to_hwirq(d); gpio_irqchip->name = "GPIO"; + gpio_irqchip->irq_request_resources = gpio_irq_request_resources; + gpio_irqchip->irq_release_resources = gpio_irq_release_resources; gpio_irqchip->irq_ack = gpio_irq_ack; gpio_irqchip->irq_disable = gpio_irq_mask; gpio_irqchip->irq_mask = gpio_irq_mask; gpio_irqchip->irq_unmask = gpio_irq_unmask; gpio_irqchip->irq_set_wake = pm_ptr(gpio_irq_set_wake); gpio_irqchip->irq_set_type = at91_gpio->ops->irq_type; + gpio_irqchip->flags = IRQCHIP_IMMUTABLE; /* Disable irqs of this PIO controller */ writel_relaxed(~0, at91_gpio->regbase + PIO_IDR); @@ -1747,7 +1770,7 @@ static int at91_gpio_of_irq_setup(struct platform_device *pdev, * interrupt. */ girq = &at91_gpio->chip.irq; - girq->chip = gpio_irqchip; + gpio_irq_chip_set_chip(girq, gpio_irqchip); girq->default_type = IRQ_TYPE_NONE; girq->handler = handle_edge_irq;
To help gpiolib not fiddle around with the internals of the irqchip flag the chip as immutable, adding the calls into the gpiolib core required to do so. Signed-off-by: Mark Brown <broonie@kernel.org> --- drivers/pinctrl/pinctrl-at91.c | 25 ++++++++++++++++++++++++- 1 file changed, 24 insertions(+), 1 deletion(-)