blob: 5ae30de3490ac84118da2da0b169f8f24b9da4b9 [file] [log] [blame]
/*
* GPIOs on MPC512x/8349/8572/8610/QorIQ and compatible
*
* Copyright (C) 2008 Peter Korsgaard <jacmet@sunsite.dk>
* Copyright (C) 2016 Freescale Semiconductor Inc.
*
* This file is licensed under the terms of the GNU General Public License
* version 2. This program is licensed "as is" without any warranty of any
* kind, whether express or implied.
*/
#include <linux/kernel.h>
#include <linux/init.h>
#include <linux/spinlock.h>
#include <linux/io.h>
#include <linux/of.h>
#include <linux/of_gpio.h>
#include <linux/of_address.h>
#include <linux/of_irq.h>
#include <linux/of_platform.h>
#include <linux/slab.h>
#include <linux/irq.h>
#include <linux/gpio/driver.h>
#include <linux/bitops.h>
#include <linux/interrupt.h>
#define MPC8XXX_GPIO_PINS 32
#define GPIO_DIR 0x00
#define GPIO_ODR 0x04
#define GPIO_DAT 0x08
#define GPIO_IER 0x0c
#define GPIO_IMR 0x10
#define GPIO_ICR 0x14
#define GPIO_ICR2 0x18
#define GPIO_IBE 0x18
struct mpc8xxx_gpio_chip {
struct gpio_chip gc;
void __iomem *regs;
raw_spinlock_t lock;
int (*direction_output)(struct gpio_chip *chip,
unsigned offset, int value);
struct irq_domain *irq;
unsigned int irqn;
};
/* The GPIO Input Buffer Enable register(GPIO_IBE) is used to
* control the input enable of each individual GPIO port.
* When an individual GPIO port’s direction is set to
* input (GPIO_GPDIR[DRn=0]), the associated input enable must be
* set (GPIOxGPIE[IEn]=1) to propagate the port value to the GPIO
* Data Register.
*/
static int ls1028a_gpio_dir_in_init(struct gpio_chip *gc)
{
unsigned long flags;
struct mpc8xxx_gpio_chip *mpc8xxx_gc = gpiochip_get_data(gc);
spin_lock_irqsave(&gc->bgpio_lock, flags);
gc->write_reg(mpc8xxx_gc->regs + GPIO_IBE, 0xffffffff);
spin_unlock_irqrestore(&gc->bgpio_lock, flags);
return 0;
}
/*
* This hardware has a big endian bit assignment such that GPIO line 0 is
* connected to bit 31, line 1 to bit 30 ... line 31 to bit 0.
* This inline helper give the right bitmask for a certain line.
*/
static inline u32 mpc_pin2mask(unsigned int offset)
{
return BIT(31 - offset);
}
/* Workaround GPIO 1 errata on MPC8572/MPC8536. The status of GPIOs
* defined as output cannot be determined by reading GPDAT register,
* so we use shadow data register instead. The status of input pins
* is determined by reading GPDAT register.
*/
static int mpc8572_gpio_get(struct gpio_chip *gc, unsigned int gpio)
{
u32 val;
struct mpc8xxx_gpio_chip *mpc8xxx_gc = gpiochip_get_data(gc);
u32 out_mask, out_shadow;
out_mask = gc->read_reg(mpc8xxx_gc->regs + GPIO_DIR);
val = gc->read_reg(mpc8xxx_gc->regs + GPIO_DAT) & ~out_mask;
out_shadow = gc->bgpio_data & out_mask;
return !!((val | out_shadow) & mpc_pin2mask(gpio));
}
static int mpc5121_gpio_dir_out(struct gpio_chip *gc,
unsigned int gpio, int val)
{
struct mpc8xxx_gpio_chip *mpc8xxx_gc = gpiochip_get_data(gc);
/* GPIO 28..31 are input only on MPC5121 */
if (gpio >= 28)
return -EINVAL;
return mpc8xxx_gc->direction_output(gc, gpio, val);
}
static int mpc5125_gpio_dir_out(struct gpio_chip *gc,
unsigned int gpio, int val)
{
struct mpc8xxx_gpio_chip *mpc8xxx_gc = gpiochip_get_data(gc);
/* GPIO 0..3 are input only on MPC5125 */
if (gpio <= 3)
return -EINVAL;
return mpc8xxx_gc->direction_output(gc, gpio, val);
}
static int mpc8xxx_gpio_to_irq(struct gpio_chip *gc, unsigned offset)
{
struct mpc8xxx_gpio_chip *mpc8xxx_gc = gpiochip_get_data(gc);
if (mpc8xxx_gc->irq && offset < MPC8XXX_GPIO_PINS)
return irq_create_mapping(mpc8xxx_gc->irq, offset);
else
return -ENXIO;
}
static irqreturn_t mpc8xxx_gpio_irq_cascade(int irq, void *data)
{
struct mpc8xxx_gpio_chip *mpc8xxx_gc = data;
struct gpio_chip *gc = &mpc8xxx_gc->gc;
unsigned long mask;
int i;
mask = gc->read_reg(mpc8xxx_gc->regs + GPIO_IER)
& gc->read_reg(mpc8xxx_gc->regs + GPIO_IMR);
for_each_set_bit(i, &mask, 32)
generic_handle_irq(irq_linear_revmap(mpc8xxx_gc->irq, 31 - i));
return IRQ_HANDLED;
}
static void mpc8xxx_irq_unmask(struct irq_data *d)
{
struct mpc8xxx_gpio_chip *mpc8xxx_gc = irq_data_get_irq_chip_data(d);
struct gpio_chip *gc = &mpc8xxx_gc->gc;
unsigned long flags;
raw_spin_lock_irqsave(&mpc8xxx_gc->lock, flags);
gc->write_reg(mpc8xxx_gc->regs + GPIO_IMR,
gc->read_reg(mpc8xxx_gc->regs + GPIO_IMR)
| mpc_pin2mask(irqd_to_hwirq(d)));
raw_spin_unlock_irqrestore(&mpc8xxx_gc->lock, flags);
}
static void mpc8xxx_irq_mask(struct irq_data *d)
{
struct mpc8xxx_gpio_chip *mpc8xxx_gc = irq_data_get_irq_chip_data(d);
struct gpio_chip *gc = &mpc8xxx_gc->gc;
unsigned long flags;
raw_spin_lock_irqsave(&mpc8xxx_gc->lock, flags);
gc->write_reg(mpc8xxx_gc->regs + GPIO_IMR,
gc->read_reg(mpc8xxx_gc->regs + GPIO_IMR)
& ~mpc_pin2mask(irqd_to_hwirq(d)));
raw_spin_unlock_irqrestore(&mpc8xxx_gc->lock, flags);
}
static void mpc8xxx_irq_ack(struct irq_data *d)
{
struct mpc8xxx_gpio_chip *mpc8xxx_gc = irq_data_get_irq_chip_data(d);
struct gpio_chip *gc = &mpc8xxx_gc->gc;
gc->write_reg(mpc8xxx_gc->regs + GPIO_IER,
mpc_pin2mask(irqd_to_hwirq(d)));
}
static int mpc8xxx_irq_set_type(struct irq_data *d, unsigned int flow_type)
{
struct mpc8xxx_gpio_chip *mpc8xxx_gc = irq_data_get_irq_chip_data(d);
struct gpio_chip *gc = &mpc8xxx_gc->gc;
unsigned long flags;
switch (flow_type) {
case IRQ_TYPE_EDGE_FALLING:
raw_spin_lock_irqsave(&mpc8xxx_gc->lock, flags);
gc->write_reg(mpc8xxx_gc->regs + GPIO_ICR,
gc->read_reg(mpc8xxx_gc->regs + GPIO_ICR)
| mpc_pin2mask(irqd_to_hwirq(d)));
raw_spin_unlock_irqrestore(&mpc8xxx_gc->lock, flags);
break;
case IRQ_TYPE_EDGE_BOTH:
raw_spin_lock_irqsave(&mpc8xxx_gc->lock, flags);
gc->write_reg(mpc8xxx_gc->regs + GPIO_ICR,
gc->read_reg(mpc8xxx_gc->regs + GPIO_ICR)
& ~mpc_pin2mask(irqd_to_hwirq(d)));
raw_spin_unlock_irqrestore(&mpc8xxx_gc->lock, flags);
break;
default:
return -EINVAL;
}
return 0;
}
static int mpc512x_irq_set_type(struct irq_data *d, unsigned int flow_type)
{
struct mpc8xxx_gpio_chip *mpc8xxx_gc = irq_data_get_irq_chip_data(d);
struct gpio_chip *gc = &mpc8xxx_gc->gc;
unsigned long gpio = irqd_to_hwirq(d);
void __iomem *reg;
unsigned int shift;
unsigned long flags;
if (gpio < 16) {
reg = mpc8xxx_gc->regs + GPIO_ICR;
shift = (15 - gpio) * 2;
} else {
reg = mpc8xxx_gc->regs + GPIO_ICR2;
shift = (15 - (gpio % 16)) * 2;
}
switch (flow_type) {
case IRQ_TYPE_EDGE_FALLING:
case IRQ_TYPE_LEVEL_LOW:
raw_spin_lock_irqsave(&mpc8xxx_gc->lock, flags);
gc->write_reg(reg, (gc->read_reg(reg) & ~(3 << shift))
| (2 << shift));
raw_spin_unlock_irqrestore(&mpc8xxx_gc->lock, flags);
break;
case IRQ_TYPE_EDGE_RISING:
case IRQ_TYPE_LEVEL_HIGH:
raw_spin_lock_irqsave(&mpc8xxx_gc->lock, flags);
gc->write_reg(reg, (gc->read_reg(reg) & ~(3 << shift))
| (1 << shift));
raw_spin_unlock_irqrestore(&mpc8xxx_gc->lock, flags);
break;
case IRQ_TYPE_EDGE_BOTH:
raw_spin_lock_irqsave(&mpc8xxx_gc->lock, flags);
gc->write_reg(reg, (gc->read_reg(reg) & ~(3 << shift)));
raw_spin_unlock_irqrestore(&mpc8xxx_gc->lock, flags);
break;
default:
return -EINVAL;
}
return 0;
}
static struct irq_chip mpc8xxx_irq_chip = {
.name = "mpc8xxx-gpio",
.irq_unmask = mpc8xxx_irq_unmask,
.irq_mask = mpc8xxx_irq_mask,
.irq_ack = mpc8xxx_irq_ack,
/* this might get overwritten in mpc8xxx_probe() */
.irq_set_type = mpc8xxx_irq_set_type,
};
static int mpc8xxx_gpio_irq_map(struct irq_domain *h, unsigned int irq,
irq_hw_number_t hwirq)
{
irq_set_chip_data(irq, h->host_data);
irq_set_chip_and_handler(irq, &mpc8xxx_irq_chip, handle_edge_irq);
return 0;
}
static const struct irq_domain_ops mpc8xxx_gpio_irq_ops = {
.map = mpc8xxx_gpio_irq_map,
.xlate = irq_domain_xlate_twocell,
};
struct mpc8xxx_gpio_devtype {
int (*gpio_dir_in_init)(struct gpio_chip *chip);
int (*gpio_dir_out)(struct gpio_chip *, unsigned int, int);
int (*gpio_get)(struct gpio_chip *, unsigned int);
int (*irq_set_type)(struct irq_data *, unsigned int);
};
static const struct mpc8xxx_gpio_devtype mpc512x_gpio_devtype = {
.gpio_dir_out = mpc5121_gpio_dir_out,
.irq_set_type = mpc512x_irq_set_type,
};
static const struct mpc8xxx_gpio_devtype ls1028a_gpio_devtype = {
.gpio_dir_in_init = ls1028a_gpio_dir_in_init,
};
static const struct mpc8xxx_gpio_devtype mpc5125_gpio_devtype = {
.gpio_dir_out = mpc5125_gpio_dir_out,
.irq_set_type = mpc512x_irq_set_type,
};
static const struct mpc8xxx_gpio_devtype mpc8572_gpio_devtype = {
.gpio_get = mpc8572_gpio_get,
};
static const struct mpc8xxx_gpio_devtype mpc8xxx_gpio_devtype_default = {
.irq_set_type = mpc8xxx_irq_set_type,
};
static const struct of_device_id mpc8xxx_gpio_ids[] = {
{ .compatible = "fsl,mpc8349-gpio", },
{ .compatible = "fsl,mpc8572-gpio", .data = &mpc8572_gpio_devtype, },
{ .compatible = "fsl,mpc8610-gpio", },
{ .compatible = "fsl,mpc5121-gpio", .data = &mpc512x_gpio_devtype, },
{ .compatible = "fsl,mpc5125-gpio", .data = &mpc5125_gpio_devtype, },
{ .compatible = "fsl,pq3-gpio", },
{ .compatible = "fsl,ls1028a-gpio", .data = &ls1028a_gpio_devtype, },
{ .compatible = "fsl,ls1088a-gpio", .data = &ls1028a_gpio_devtype, },
{ .compatible = "fsl,qoriq-gpio", },
{}
};
static int mpc8xxx_probe(struct platform_device *pdev)
{
struct device_node *np = pdev->dev.of_node;
struct mpc8xxx_gpio_chip *mpc8xxx_gc;
struct gpio_chip *gc;
const struct mpc8xxx_gpio_devtype *devtype =
of_device_get_match_data(&pdev->dev);
int ret;
mpc8xxx_gc = devm_kzalloc(&pdev->dev, sizeof(*mpc8xxx_gc), GFP_KERNEL);
if (!mpc8xxx_gc)
return -ENOMEM;
platform_set_drvdata(pdev, mpc8xxx_gc);
raw_spin_lock_init(&mpc8xxx_gc->lock);
mpc8xxx_gc->regs = of_iomap(np, 0);
if (!mpc8xxx_gc->regs)
return -ENOMEM;
gc = &mpc8xxx_gc->gc;
gc->parent = &pdev->dev;
if (of_property_read_bool(np, "little-endian")) {
ret = bgpio_init(gc, &pdev->dev, 4,
mpc8xxx_gc->regs + GPIO_DAT,
NULL, NULL,
mpc8xxx_gc->regs + GPIO_DIR, NULL,
BGPIOF_BIG_ENDIAN);
if (ret)
goto err;
dev_dbg(&pdev->dev, "GPIO registers are LITTLE endian\n");
} else {
ret = bgpio_init(gc, &pdev->dev, 4,
mpc8xxx_gc->regs + GPIO_DAT,
NULL, NULL,
mpc8xxx_gc->regs + GPIO_DIR, NULL,
BGPIOF_BIG_ENDIAN
| BGPIOF_BIG_ENDIAN_BYTE_ORDER);
if (ret)
goto err;
dev_dbg(&pdev->dev, "GPIO registers are BIG endian\n");
}
mpc8xxx_gc->direction_output = gc->direction_output;
if (!devtype)
devtype = &mpc8xxx_gpio_devtype_default;
/*
* It's assumed that only a single type of gpio controller is available
* on the current machine, so overwriting global data is fine.
*/
if (devtype->irq_set_type)
mpc8xxx_irq_chip.irq_set_type = devtype->irq_set_type;
if (devtype->gpio_dir_out)
gc->direction_output = devtype->gpio_dir_out;
if (devtype->gpio_get)
gc->get = devtype->gpio_get;
gc->to_irq = mpc8xxx_gpio_to_irq;
if (of_device_is_compatible(np, "fsl,qoriq-gpio"))
gc->write_reg(mpc8xxx_gc->regs + GPIO_IBE, 0xffffffff);
ret = gpiochip_add_data(gc, mpc8xxx_gc);
if (ret) {
pr_err("%pOF: GPIO chip registration failed with status %d\n",
np, ret);
goto err;
}
mpc8xxx_gc->irqn = irq_of_parse_and_map(np, 0);
if (!mpc8xxx_gc->irqn)
return 0;
mpc8xxx_gc->irq = irq_domain_add_linear(np, MPC8XXX_GPIO_PINS,
&mpc8xxx_gpio_irq_ops, mpc8xxx_gc);
if (!mpc8xxx_gc->irq)
return 0;
/* ack and mask all irqs */
gc->write_reg(mpc8xxx_gc->regs + GPIO_IER, 0xffffffff);
gc->write_reg(mpc8xxx_gc->regs + GPIO_IMR, 0);
/* enable input buffer */
if (devtype->gpio_dir_in_init)
devtype->gpio_dir_in_init(gc);
ret = devm_request_irq(&pdev->dev, mpc8xxx_gc->irqn,
mpc8xxx_gpio_irq_cascade,
IRQF_NO_THREAD | IRQF_SHARED, "gpio-cascade",
mpc8xxx_gc);
if (ret) {
dev_err(&pdev->dev, "%s: failed to devm_request_irq(%d), ret = %d\n",
np->full_name, mpc8xxx_gc->irqn, ret);
goto err;
}
return 0;
err:
iounmap(mpc8xxx_gc->regs);
return ret;
}
static int mpc8xxx_remove(struct platform_device *pdev)
{
struct mpc8xxx_gpio_chip *mpc8xxx_gc = platform_get_drvdata(pdev);
if (mpc8xxx_gc->irq) {
irq_set_chained_handler_and_data(mpc8xxx_gc->irqn, NULL, NULL);
irq_domain_remove(mpc8xxx_gc->irq);
}
gpiochip_remove(&mpc8xxx_gc->gc);
iounmap(mpc8xxx_gc->regs);
return 0;
}
static struct platform_driver mpc8xxx_plat_driver = {
.probe = mpc8xxx_probe,
.remove = mpc8xxx_remove,
.driver = {
.name = "gpio-mpc8xxx",
.of_match_table = mpc8xxx_gpio_ids,
},
};
static int __init mpc8xxx_init(void)
{
return platform_driver_register(&mpc8xxx_plat_driver);
}
arch_initcall(mpc8xxx_init);