Re: [PATCH v9 2/4] pinctrl: cygnus: add gpio/pinconf driver

From: Linus Walleij
Date: Wed Mar 04 2015 - 04:48:14 EST


On Tue, Feb 10, 2015 at 11:16 PM, Ray Jui <rjui@xxxxxxxxxxxx> wrote:

> This adds the initial support of the Broadcom Cygnus GPIO/PINCONF driver
> that supports all 3 GPIO controllers on Cygnus including the ASIU GPIO
> controller, the chipCommonG GPIO controller, and the always-on GPIO
> controller. Basic PINCONF configurations such as bias pull up/down, and
> drive strength are also supported in this driver.
>
> Pins from the ASIU GPIO controller can be individually muxed to GPIO
> function, through interaction with the Cygnus IOMUX controller
>
> Signed-off-by: Ray Jui <rjui@xxxxxxxxxxxx>
> Reviewed-by: Scott Branden <sbranden@xxxxxxxxxxxx>

OK!

> diff --git a/drivers/pinctrl/bcm/pinctrl-cygnus-gpio.c b/drivers/pinctrl/bcm/pinctrl-cygnus-gpio.c
(...)

> +#include <linux/kernel.h>
> +#include <linux/slab.h>
> +#include <linux/module.h>
> +#include <linux/interrupt.h>
> +#include <linux/io.h>
> +#include <linux/gpio.h>

This should be:
#include <linux/gpio/driver.h>

> +static u32 cygnus_readl(struct cygnus_gpio *chip, unsigned int offset)
> +{
> + return readl(chip->base + offset);
> +}
> +
> +static void cygnus_writel(struct cygnus_gpio *chip, unsigned int offset,
> + u32 val)
> +{
> + writel(val, chip->base + offset);
> +}

If these look like so, just use readl(val, chip->base +
offset)/writel() at all sites.
These functions just adds a pointless layer of abstraction.

> +static void cygnus_set_bit(struct cygnus_gpio *chip, unsigned int reg,
> + unsigned gpio, int set)

"set" should be bool, right?

> +{
> + unsigned int offset = CYGNUS_GPIO_REG(gpio, reg);
> + unsigned int shift = CYGNUS_GPIO_SHIFT(gpio);
> + u32 val;
> +
> + val = cygnus_readl(chip, offset);
> + if (set)
> + val |= BIT(shift);
> + else
> + val &= ~BIT(shift);
> + cygnus_writel(chip, offset, val);
> +}
> +
> +static int cygnus_get_bit(struct cygnus_gpio *chip, unsigned int reg,
> + unsigned gpio)

This should be bool right, not int?

> +{
> + unsigned int offset = CYGNUS_GPIO_REG(gpio, reg);
> + unsigned int shift = CYGNUS_GPIO_SHIFT(gpio);
> + u32 val;
> +
> + val =
> + if (val)
> + return 1;
> + else
> + return 0;

Just:
return !!(cygnus_readl(chip, offset) & BIT(shift));

Both of these are a bit overzealous like the readl/writel functions above,
consider just inlining them.

> +static int cygnus_gpio_irq_set_type(struct irq_data *d, unsigned int type)
> +{
> + struct gpio_chip *gc = irq_data_get_irq_chip_data(d);
> + struct cygnus_gpio *chip = to_cygnus_gpio(gc);
> + unsigned gpio = d->hwirq;
> + int int_type = 0, dual_edge = 0, edge_lvl = 0;

It looks like some of these should be bool.

> + unsigned long flags;
> +
> + switch (type & IRQ_TYPE_SENSE_MASK) {
> + case IRQ_TYPE_EDGE_RISING:
> + edge_lvl = 1;

= true;

etc.

> + break;
> +
> + case IRQ_TYPE_EDGE_FALLING:
> + break;
> +
> + case IRQ_TYPE_EDGE_BOTH:
> + dual_edge = 1;
> + break;
> +
> + case IRQ_TYPE_LEVEL_HIGH:
> + int_type = 1;
> + edge_lvl = 1;
> + break;
> +
> + case IRQ_TYPE_LEVEL_LOW:
> + int_type = 1;
> + break;
> +
> + default:
> + dev_err(chip->dev, "invalid GPIO IRQ type 0x%x\n",
> + type);
> + return -EINVAL;
> + }
> +
> + spin_lock_irqsave(&chip->lock, flags);
> + cygnus_set_bit(chip, CYGNUS_GPIO_IN_TYPE_OFFSET, gpio, int_type);
> + cygnus_set_bit(chip, CYGNUS_GPIO_INT_DE_OFFSET, gpio, dual_edge);
> + cygnus_set_bit(chip, CYGNUS_GPIO_INT_EDGE_OFFSET, gpio,
> + edge_lvl);
> + spin_unlock_irqrestore(&chip->lock, flags);
> +
> + dev_dbg(chip->dev,
> + "gpio:%u set int_type:%d dual_edge:%d edge_lvl:%d\n",
> + gpio, int_type, dual_edge, edge_lvl);
> +
> + return 0;
> +}

(...)
> +/*
> + * Request the Cygnus IOMUX pinmux controller to mux individual pins to GPIO
> + */
> +static int cygnus_gpio_request(struct gpio_chip *gc, unsigned offset)
> +{
> + struct cygnus_gpio *chip = to_cygnus_gpio(gc);
> + unsigned gpio = gc->base + offset;
> +
> + /* not all Cygnus GPIO pins can be muxed individually */
> + if (!chip->pinmux_is_supported)
> + return 0;
> +
> + return pinctrl_request_gpio(gpio);
> +}
> +
> +static void cygnus_gpio_free(struct gpio_chip *gc, unsigned offset)
> +{
> + struct cygnus_gpio *chip = to_cygnus_gpio(gc);
> + unsigned gpio = gc->base + offset;
> +
> + if (!chip->pinmux_is_supported)
> + return;
> +
> + pinctrl_free_gpio(gpio);
> +}

Nice!

> +static const struct pinconf_ops cygnus_pconf_ops = {
> + .is_generic = true,
> + .pin_config_get = cygnus_pin_config_get,
> + .pin_config_set = cygnus_pin_config_set,
> +};

This pin config thing looks really nice.

> +/*
> + * Map a GPIO in the local gpio_chip pin space to a pin in the Cygnus IOMUX
> + * pinctrl pin space
> + */
> +struct cygnus_gpio_pin_range {
> + unsigned offset;
> + unsigned pin_base;
> + unsigned num_pins;
> +};
> +
> +#define CYGNUS_PINRANGE(o, p, n) { .offset = o, .pin_base = p, .num_pins = n }

Aha so this mapping is really very sparse...

> +/*
> + * Pin mapping table for mapping local GPIO pins to Cygnus IOMUX pinctrl pins
> + */
> +static const struct cygnus_gpio_pin_range cygnus_gpio_pintable[] = {
> + CYGNUS_PINRANGE(0, 42, 1),
> + CYGNUS_PINRANGE(1, 44, 3),
> + CYGNUS_PINRANGE(4, 48, 1),
> + CYGNUS_PINRANGE(5, 50, 3),
> + CYGNUS_PINRANGE(8, 126, 1),
> + CYGNUS_PINRANGE(9, 155, 1),
> + CYGNUS_PINRANGE(10, 152, 1),
> + CYGNUS_PINRANGE(11, 154, 1),
> + CYGNUS_PINRANGE(12, 153, 1),
> + CYGNUS_PINRANGE(13, 127, 3),
> + CYGNUS_PINRANGE(16, 140, 1),
> + CYGNUS_PINRANGE(17, 145, 7),
> + CYGNUS_PINRANGE(24, 130, 10),
> + CYGNUS_PINRANGE(34, 141, 4),
> + CYGNUS_PINRANGE(38, 54, 1),
> + CYGNUS_PINRANGE(39, 56, 3),
> + CYGNUS_PINRANGE(42, 60, 3),
> + CYGNUS_PINRANGE(45, 64, 3),
> + CYGNUS_PINRANGE(48, 68, 2),
> + CYGNUS_PINRANGE(50, 84, 6),
> + CYGNUS_PINRANGE(56, 94, 6),
> + CYGNUS_PINRANGE(62, 72, 1),
> + CYGNUS_PINRANGE(63, 70, 1),
> + CYGNUS_PINRANGE(64, 80, 1),
> + CYGNUS_PINRANGE(65, 74, 3),
> + CYGNUS_PINRANGE(68, 78, 1),
> + CYGNUS_PINRANGE(69, 82, 1),
> + CYGNUS_PINRANGE(70, 156, 17),
> + CYGNUS_PINRANGE(87, 104, 12),
> + CYGNUS_PINRANGE(99, 102, 2),
> + CYGNUS_PINRANGE(101, 90, 4),
> + CYGNUS_PINRANGE(105, 116, 10),
> + CYGNUS_PINRANGE(123, 11, 1),
> + CYGNUS_PINRANGE(124, 38, 4),
> + CYGNUS_PINRANGE(128, 43, 1),
> + CYGNUS_PINRANGE(129, 47, 1),
> + CYGNUS_PINRANGE(130, 49, 1),
> + CYGNUS_PINRANGE(131, 53, 1),
> + CYGNUS_PINRANGE(132, 55, 1),
> + CYGNUS_PINRANGE(133, 59, 1),
> + CYGNUS_PINRANGE(134, 63, 1),
> + CYGNUS_PINRANGE(135, 67, 1),
> + CYGNUS_PINRANGE(136, 71, 1),
> + CYGNUS_PINRANGE(137, 73, 1),
> + CYGNUS_PINRANGE(138, 77, 1),
> + CYGNUS_PINRANGE(139, 79, 1),
> + CYGNUS_PINRANGE(140, 81, 1),
> + CYGNUS_PINRANGE(141, 83, 1),
> + CYGNUS_PINRANGE(142, 10, 1)
> +};
> +
> +/*
> + * The Cygnus IOMUX controller mainly supports group based mux configuration,
> + * but certain pins can be muxed to GPIO individually. Only the ASIU GPIO
> + * controller can support this, so it's an optional configuration
> + *
> + * Return -ENODEV means no support and that's fine
> + */
> +static int cygnus_gpio_pinmux_add_range(struct cygnus_gpio *chip)
> +{
> + struct device_node *node = chip->dev->of_node;
> + struct device_node *pinmux_node;
> + struct platform_device *pinmux_pdev;
> + struct gpio_chip *gc = &chip->gc;
> + int i, ret = 0;
> +
> + /* parse DT to find the phandle to the pinmux controller */
> + pinmux_node = of_parse_phandle(node, "pinmux", 0);
> + if (!pinmux_node)
> + return -ENODEV;
> +
> + pinmux_pdev = of_find_device_by_node(pinmux_node);
> + /* no longer need the pinmux node */
> + of_node_put(pinmux_node);
> + if (!pinmux_pdev) {
> + dev_err(chip->dev, "failed to get pinmux device\n");
> + return -EINVAL;
> + }
> +
> + /* now need to create the mapping between local GPIO and PINMUX pins */
> + for (i = 0; i < ARRAY_SIZE(cygnus_gpio_pintable); i++) {
> + ret = gpiochip_add_pin_range(gc, dev_name(&pinmux_pdev->dev),
> + cygnus_gpio_pintable[i].offset,
> + cygnus_gpio_pintable[i].pin_base,
> + cygnus_gpio_pintable[i].num_pins);
> + if (ret) {
> + dev_err(chip->dev, "unable to add GPIO pin range\n");
> + goto err_put_device;
> + }
> + }

This is really nice. Awesome job, exactly how it should be done,
even if it's a bit complex.

> + chip->pinmux_is_supported = (ret == 0);

Just = true;

You cannot get here with ret != 0.

> +static void cygnus_gpio_pinmux_remove_range(struct cygnus_gpio *chip)
> +{
> + struct gpio_chip *gc = &chip->gc;
> +
> + if (chip->pinmux_is_supported)
> + gpiochip_remove_pin_ranges(gc);
> +}

You don't need to do this. Look in gpiochip_remove() and you see it
will remove the range for you.

+
> +err_unregister_pinconf:
> + cygnus_gpio_unregister_pinconf(chip);
> +
> +err_rm_range:
> + cygnus_gpio_pinmux_remove_range(chip);

Not needed I think.

> +
> +err_rm_gpiochip:
> + gpiochip_remove(gc);

Because this will do it.

> +
> + return ret;
> +}

Apart from that this looks really good!

If you resend with the above nitpicks fixed I will merge this.

Yours,
Linus Walleij
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/