Re: [PATCH v3 4/9] drivers: irqchip: add PDC irqdomain for wakeup capable GPIOs

[Marc Zyngier <marc.zyngier@arm.com>]

On 22/02/2019 22:18, Lina Iyer wrote:
> Introduce a new domain for wakeup capable GPIOs. The domain can be
> requested using the bus token DOMAIN_BUS_WAKEUP. In the following
> patches, we will specify PDC as the wakeup-parent for the TLMM GPIO
> irqchip. Requesting a wakeup GPIO will setup the GPIO and the
> corresponding PDC interrupt as its parent.
> 
> Co-developed-by: Stephen Boyd <swboyd@chromium.org>
> Signed-off-by: Lina Iyer <ilina@codeaurora.org>
> ---
> Changes in v3:
> 	- Remove PDC GPIO map data (moved to DT)
> 	- hwirq passed in .alloc() is a PDC pin now
> Changes in v2:
> 	- Remove separate file for PDC GPIO map data
> 	- Error checks and return
> 	- Whitespace fixes
> ---
>  drivers/irqchip/qcom-pdc.c   | 106 ++++++++++++++++++++++++++++++++---
>  include/linux/soc/qcom/irq.h |  23 ++++++++
>  2 files changed, 120 insertions(+), 9 deletions(-)
>  create mode 100644 include/linux/soc/qcom/irq.h
> 
> diff --git a/drivers/irqchip/qcom-pdc.c b/drivers/irqchip/qcom-pdc.c
> index faa7d61b9d6c..5a64cbeec410 100644
> --- a/drivers/irqchip/qcom-pdc.c
> +++ b/drivers/irqchip/qcom-pdc.c
> @@ -13,12 +13,13 @@
>  #include <linux/of.h>
>  #include <linux/of_address.h>
>  #include <linux/of_device.h>
> +#include <linux/soc/qcom/irq.h>
>  #include <linux/spinlock.h>
> -#include <linux/platform_device.h>
>  #include <linux/slab.h>
>  #include <linux/types.h>
>  
>  #define PDC_MAX_IRQS		126
> +#define PDC_MAX_GPIO_IRQS	256
>  
>  #define CLEAR_INTR(reg, intr)	(reg & ~(1 << intr))
>  #define ENABLE_INTR(reg, intr)	(reg | (1 << intr))
> @@ -32,6 +33,16 @@ struct pdc_pin_region {
>  	u32 cnt;
>  };
>  
> +struct pdc_gpio_pin_map {
> +	unsigned int gpio;
> +	u32 pdc_pin;
> +};
> +
> +struct pdc_gpio_pin_data {
> +	size_t size;
> +	const struct pdc_gpio_pin_map *map;
> +};

I can't see this being used anywhere.

> +
>  static DEFINE_RAW_SPINLOCK(pdc_lock);
>  static void __iomem *pdc_base;
>  static struct pdc_pin_region *pdc_region;
> @@ -47,9 +58,8 @@ static u32 pdc_reg_read(int reg, u32 i)
>  	return readl_relaxed(pdc_base + reg + i * sizeof(u32));
>  }
>  
> -static void pdc_enable_intr(struct irq_data *d, bool on)
> +static void pdc_enable_intr(irq_hw_number_t pin_out, bool on)
>  {
> -	int pin_out = d->hwirq;

Why do you need this change? As far as I can tell, you can always use
the irq_data here, and it would reduce the size of the patch.

>  	u32 index, mask;
>  	u32 enable;
>  
> @@ -65,13 +75,23 @@ static void pdc_enable_intr(struct irq_data *d, bool on)
>  
>  static void qcom_pdc_gic_mask(struct irq_data *d)
>  {
> -	pdc_enable_intr(d, false);
> +	irq_hw_number_t hwirq = d->hwirq;
> +
> +	if (hwirq == ULONG_MAX)

Can you define a specific constant here? PDC_WAKEUP_IRQ? Or something?
Also, I can't work out where d->hwirq gets set to such value.

> +		return;
> +
> +	pdc_enable_intr(hwirq, false);
>  	irq_chip_mask_parent(d);
>  }
>  
>  static void qcom_pdc_gic_unmask(struct irq_data *d)
>  {
> -	pdc_enable_intr(d, true);
> +	irq_hw_number_t hwirq = d->hwirq;
> +
> +	if (hwirq == ULONG_MAX)
> +		return;
> +
> +	pdc_enable_intr(hwirq, true);
>  	irq_chip_unmask_parent(d);
>  }
>  
> @@ -111,9 +131,12 @@ enum pdc_irq_config_bits {
>   */
>  static int qcom_pdc_gic_set_type(struct irq_data *d, unsigned int type)
>  {
> -	int pin_out = d->hwirq;
> +	irq_hw_number_t pin_out = d->hwirq;
>  	enum pdc_irq_config_bits pdc_type;
>  
> +	if (pin_out == ULONG_MAX)
> +		return 0;
> +

It is quite odd that you're accepting any odd trigger, even if they are
illegal. The API should be consistent, even if you're not applying any
change.

>  	switch (type) {
>  	case IRQ_TYPE_EDGE_RISING:
>  		pdc_type = PDC_EDGE_RISING;
> @@ -157,7 +180,7 @@ static struct irq_chip qcom_pdc_gic_chip = {
>  	.irq_set_affinity	= irq_chip_set_affinity_parent,
>  };
>  
> -static irq_hw_number_t get_parent_hwirq(int pin)
> +static irq_hw_number_t get_parent_hwirq(irq_hw_number_t pin)
>  {
>  	int i;
>  	struct pdc_pin_region *region;
> @@ -169,7 +192,6 @@ static irq_hw_number_t get_parent_hwirq(int pin)
>  			return (region->parent_base + pin - region->pin_base);
>  	}
>  
> -	WARN_ON(1);
>  	return ~0UL;

Above, you're using ULONG_MAX, and here ~0UL. You need to make up your
mind and use one single identifier. Or maybe these are not the same
thing? In any case, you need to lift the ambiguity.

>  }
>  
> @@ -232,6 +254,60 @@ static const struct irq_domain_ops qcom_pdc_ops = {
>  	.free		= irq_domain_free_irqs_common,
>  };
>  
> +static int qcom_pdc_gpio_alloc(struct irq_domain *domain, unsigned int virq,
> +			       unsigned int nr_irqs, void *data)
> +{
> +	struct qcom_irq_fwspec *qcom_fwspec = data;
> +	struct irq_fwspec *fwspec = &qcom_fwspec->fwspec;
> +	struct irq_fwspec parent_fwspec;
> +	irq_hw_number_t hwirq, parent_hwirq;
> +	unsigned int type;
> +	int ret;
> +
> +	ret = qcom_pdc_translate(domain, fwspec, &hwirq, &type);
> +	if (ret)
> +		return -EINVAL;
> +
> +	parent_hwirq = get_parent_hwirq(hwirq);
> +	if (parent_hwirq == ~0UL)
> +		return -EINVAL;
> +
> +	ret = irq_domain_set_hwirq_and_chip(domain, virq, hwirq,
> +					    &qcom_pdc_gic_chip, NULL);
> +	if (ret)
> +		return ret;
> +
> +	qcom_fwspec->mask = true;
> +
> +	if (type & IRQ_TYPE_EDGE_BOTH)
> +		type = IRQ_TYPE_EDGE_RISING;
> +
> +	if (type & IRQ_TYPE_LEVEL_MASK)
> +		type = IRQ_TYPE_LEVEL_HIGH;
> +
> +	parent_fwspec.fwnode      = domain->parent->fwnode;
> +	parent_fwspec.param_count = 3;
> +	parent_fwspec.param[0]    = 0;
> +	parent_fwspec.param[1]    = parent_hwirq;
> +	parent_fwspec.param[2]    = type;
> +
> +	return irq_domain_alloc_irqs_parent(domain, virq, nr_irqs,
> +					    &parent_fwspec);
> +}
> +
> +static int qcom_pdc_gpio_domain_select(struct irq_domain *d,
> +				       struct irq_fwspec *fwspec,
> +				       enum irq_domain_bus_token bus_token)
> +{
> +	return (bus_token == DOMAIN_BUS_WAKEUP);
> +}
> +
> +static const struct irq_domain_ops qcom_pdc_gpio_ops = {
> +	.select		= qcom_pdc_gpio_domain_select,
> +	.alloc		= qcom_pdc_gpio_alloc,
> +	.free		= irq_domain_free_irqs_common,
> +};
> +
>  static int pdc_setup_pin_mapping(struct device_node *np)
>  {
>  	int ret, n;
> @@ -270,7 +346,7 @@ static int pdc_setup_pin_mapping(struct device_node *np)
>  
>  static int qcom_pdc_init(struct device_node *node, struct device_node *parent)
>  {
> -	struct irq_domain *parent_domain, *pdc_domain;
> +	struct irq_domain *parent_domain, *pdc_domain, *pdc_gpio_domain;
>  	int ret;
>  
>  	pdc_base = of_iomap(node, 0);
> @@ -301,6 +377,18 @@ static int qcom_pdc_init(struct device_node *node, struct device_node *parent)
>  		goto fail;
>  	}
>  
> +	pdc_gpio_domain = irq_domain_create_hierarchy(parent_domain, 0,
> +						      PDC_MAX_GPIO_IRQS,
> +						      of_fwnode_handle(node),
> +						      &qcom_pdc_gpio_ops, NULL);
> +	if (!pdc_gpio_domain) {
> +		pr_err("%pOF: GIC domain add failed for GPIO domain\n", node);
> +		ret = -ENOMEM;
> +		goto fail;
> +	}
> +
> +	irq_domain_update_bus_token(pdc_gpio_domain, DOMAIN_BUS_WAKEUP);
> +
>  	return 0;
>  
>  fail:
> diff --git a/include/linux/soc/qcom/irq.h b/include/linux/soc/qcom/irq.h
> new file mode 100644
> index 000000000000..bacc9edbce0d
> --- /dev/null
> +++ b/include/linux/soc/qcom/irq.h
> @@ -0,0 +1,23 @@
> +/* SPDX-License-Identifier: GPL-2.0 */
> +
> +#ifndef __QCOM_IRQ_H
> +#define __QCOM_IRQ_H
> +
> +#include <linux/irqdomain.h>
> +
> +/**
> + * struct qcom_irq_fwspec - qcom specific irq fwspec wrapper
> + * @fwspec: irq fwspec
> + * @mask: if true, keep the irq masked in the gpio controller
> + *
> + * Use this structure to communicate between the parent irq chip, MPM or PDC,
> + * to the gpio chip, TLMM, about the gpio being allocated in the parent
> + * and if the gpio chip should keep the line masked because the parent irq
> + * chip is handling everything about the irq line.
> + */
> +struct qcom_irq_fwspec {
> +	struct irq_fwspec fwspec;
> +	bool mask;
> +};
> +
> +#endif
> 

Thanks,

	M.
-- 
Jazz is not dead. It just smells funny...