Re: [PATCH v4] gpio: driver for PrimeCell PL061 GPIO controller

From: Baruch Siach
Date: Wed Jun 10 2009 - 03:23:29 EST


Hi Andrew,

On Tue, Jun 09, 2009 at 02:02:39PM -0700, Andrew Morton wrote:
> On Sun, 7 Jun 2009 21:38:55 +0300
> Baruch Siach <baruch@xxxxxxxxxx> wrote:
>
> > This is a driver for the ARM PrimeCell PL061 GPIO AMBA peripheral. The driver
> > is implemented using the gpiolib framework.
> >
> > This driver also includes support for the use of the PL061 as an interrupt
> > controller (secondary).
> >
> > Signed-off-by: Baruch Siach <baruch@xxxxxxxxxx>
> > ---
> > Changes in v4:
> > - Depend on CONFIG_ARM_AMBA
> > - Do not request the gpio for IRQ automatically, just warn if it's not
> > requested
> > - Remove SZ_4K
> > - Iterate through a linked list find the source of interrupt when
> > multiple PL061s are connected to the same IRQ trigger
> > - Move hardware register defines into the .c file
>
> I generated an incremental diff so that we can see what you did.
>
>
> > drivers/gpio/Kconfig | 1
> > drivers/gpio/pl061.c | 92 ++++++++++++++++-------------------
> > include/linux/amba/pl061.h | 19 +++----
> > 3 files changed, 54 insertions(+), 58 deletions(-)
> >
> > diff -puN drivers/gpio/Kconfig~gpio-driver-for-primecell-pl061-gpio-controller-v4 drivers/gpio/Kconfig
> > --- a/drivers/gpio/Kconfig~gpio-driver-for-primecell-pl061-gpio-controller-v4
> > +++ a/drivers/gpio/Kconfig
> > @@ -69,6 +69,7 @@ comment "Memory mapped GPIO expanders:"
> >
> > config GPIO_PL061
> > bool "PrimeCell PL061 GPIO support"
> > + depends on ARM_AMBA
> > help
> > Say yes here to support the PrimeCell PL061 GPIO device
> >
> > diff -puN drivers/gpio/pl061.c~gpio-driver-for-primecell-pl061-gpio-controller-v4 drivers/gpio/pl061.c
> > --- a/drivers/gpio/pl061.c~gpio-driver-for-primecell-pl061-gpio-controller-v4
> > +++ a/drivers/gpio/pl061.c
> > @@ -14,6 +14,7 @@
> > #include <linux/spinlock.h>
> > #include <linux/errno.h>
> > #include <linux/module.h>
> > +#include <linux/list.h>
> > #include <linux/io.h>
> > #include <linux/ioport.h>
> > #include <linux/irq.h>
> > @@ -24,9 +25,22 @@
> > #include <linux/amba/bus.h>
> > #include <linux/amba/pl061.h>
> >
> > +#define GPIODIR 0x400
> > +#define GPIOIS 0x404
> > +#define GPIOIBE 0x408
> > +#define GPIOIEV 0x40C
> > +#define GPIOIE 0x410
> > +#define GPIORIS 0x414
> > +#define GPIOMIS 0x418
> > +#define GPIOIC 0x41C
> > +
> > #define PL061_GPIO_NR 8
> >
> > +#define PL061_REG_SIZE (4*1024)
> > +
> > struct pl061_gpio {
> > + struct list_head list;
>
> Some commentary which describes this field would be nice. What lock
> protects it, what other list_head this list is anchored to, etc

I'll add a comment.

Regarding the lock, the only list_add() is done in the probe method, and this
method is serialized, as far as I know.

> > /* Each of the two spinlocks protects a different set of hardware
> > * regiters and data structurs. This decouples the code of the IRQ from
> > * the GPIO code. This also makes the case of a GPIO routine call from
> > @@ -37,12 +51,8 @@ struct pl061_gpio {
> >
> > void __iomem *base;
> > struct gpio_chip gc;
> > - struct work_struct gpio_free_work;
> > - DECLARE_BITMAP(gpios_to_free, PL061_GPIO_NR);
> > };
> >
> > -static u32 (*pl061_pending_irq)(int irq);
> > -
> > static int pl061_direction_input(struct gpio_chip *gc, unsigned offset)
> > {
> > struct pl061_gpio *chip = container_of(gc, struct pl061_gpio, gc);
> > @@ -112,16 +122,6 @@ static void pl061_irq_disable(unsigned i
> > spin_unlock_irqrestore(&chip->irq_lock, flags);
> > }
> >
> > -static void pl061_irq_shutdown(unsigned irq)
> > -{
> > - struct pl061_gpio *chip = get_irq_chip_data(irq);
> > - int offset = irq_to_gpio(irq) - chip->gc.base;
> > -
> > - pl061_irq_disable(irq);
> > - set_bit(offset, chip->gpios_to_free);
> > - schedule_work(&chip->gpio_free_work);
> > -}
> > -
> > static void pl061_irq_enable(unsigned irq)
> > {
> > struct pl061_gpio *chip = get_irq_chip_data(irq);
> > @@ -138,16 +138,10 @@ static void pl061_irq_enable(unsigned ir
> >
> > static unsigned int pl061_irq_startup(unsigned irq)
> > {
> > - int ret;
> > -
> > - ret = gpio_request(irq_to_gpio(irq), "IRQ");
> > - if (ret < 0) {
> > - pr_warning("%s: warning: gpio_request(%d) returned %d\n",
> > - __func__, irq_to_gpio(irq), ret);
> > - return 0;
> > - }
> > + if (gpio_request(irq_to_gpio(irq), "IRQ") == 0)
> > + pr_warning("%s: warning: GPIO%d has not been requested\n",
> > + __func__, irq_to_gpio(irq));
>
> This is wrong, isn't it? gpio_request() returns 0 on success.

Russell said that gpio configuration is the responsibility of the platform
code. Here I just warn when the gpio has not been requested, and thus
gpio_request() succeeds. I'll add a comment.

> > - gpio_direction_input(irq_to_gpio(irq));
> > pl061_irq_enable(irq);
> >
> > return 0;
> > @@ -202,45 +196,38 @@ static struct irq_chip pl061_irqchip = {
> > .startup = pl061_irq_startup,
> > .enable = pl061_irq_enable,
> > .disable = pl061_irq_disable,
> > - .shutdown = pl061_irq_shutdown,
> > .set_type = pl061_irq_type,
> > };
> >
> > static void pl061_irq_handler(unsigned irq, struct irq_desc *desc)
> > {
> > + struct list_head *chip_list = get_irq_chip_data(irq);
> > + struct list_head *ptr;
> > + struct pl061_gpio *chip;
> > +
> > desc->chip->ack(irq);
> > - while (1) {
> > + list_for_each(ptr, chip_list) {
>
> What locking protects the newly-added list?

Do we need locking even though we list_add() only at probe time? (Compiling as
a module is not supported, so this only happens at boot time).

> > unsigned long pending;
> > int gpio;
> >
> > - pending = pl061_pending_irq(irq);
> > + chip = list_entry(ptr, struct pl061_gpio, list);
> > + pending = readb(chip->base + GPIOMIS);
> > + writeb(pending, chip->base + GPIOIC);
> > +
> > if (pending == 0)
> > - break;
> > + continue;
> >
> > - for_each_bit(gpio, &pending, BITS_PER_LONG)
> > + for_each_bit(gpio, &pending, PL061_GPIO_NR)
> > generic_handle_irq(gpio_to_irq(gpio));
> > }
> > desc->chip->unmask(irq);
> > }
> >
> > -static void pl061_gpio_free(struct work_struct *work)
> > -{
> > - struct pl061_gpio *chip = container_of(work, struct pl061_gpio,
> > - gpio_free_work);
> > - int offset;
> > -
> > - for_each_bit(offset, chip->gpios_to_free, PL061_GPIO_NR) {
> > - int gpio = offset + chip->gc.base;
> > -
> > - if (test_and_clear_bit(offset, chip->gpios_to_free))
> > - gpio_free(gpio);
> > - }
> > -}
> > -
> > static int __init pl061_probe(struct amba_device *dev, struct amba_id *id)
> > {
> > struct pl061_platform_data *pdata;
> > struct pl061_gpio *chip;
> > + struct list_head *chip_list;
> > int ret, irq, i;
> >
> > pdata = dev->dev.platform_data;
> > @@ -251,12 +238,12 @@ static int __init pl061_probe(struct amb
> > if (chip == NULL)
> > return -ENOMEM;
> >
> > - if (request_mem_region(dev->res.start, SZ_4K, "pl061") == NULL) {
> > + if (!request_mem_region(dev->res.start, PL061_REG_SIZE, "pl061")) {
> > ret = -EBUSY;
> > goto free_mem;
> > }
> >
> > - chip->base = ioremap(dev->res.start, SZ_4K);
> > + chip->base = ioremap(dev->res.start, PL061_REG_SIZE);
> > if (chip->base == NULL) {
> > ret = -ENOMEM;
> > goto release_region;
> > @@ -264,8 +251,7 @@ static int __init pl061_probe(struct amb
> >
> > spin_lock_init(&chip->lock);
> > spin_lock_init(&chip->irq_lock);
> > -
> > - INIT_WORK(&chip->gpio_free_work, pl061_gpio_free);
> > + INIT_LIST_HEAD(&chip->list);
> >
> > chip->gc.direction_input = pl061_direction_input;
> > chip->gc.direction_output = pl061_direction_output;
> > @@ -291,7 +277,17 @@ static int __init pl061_probe(struct amb
> > goto iounmap;
> > }
> > set_irq_chained_handler(irq, pl061_irq_handler);
> > - pl061_pending_irq = pdata->pending_irqs;
> > + if (!pdata->is_irq_initialized || !pdata->is_irq_initialized(irq)) {
> > + chip_list = kmalloc(sizeof(*chip_list), GFP_KERNEL);
> > + if (chip_list == NULL) {
> > + ret = -ENOMEM;
> > + goto iounmap;
> > + }
> > + INIT_LIST_HEAD(chip_list);
> > + set_irq_chip_data(irq, chip_list);
> > + } else
> > + chip_list = get_irq_chip_data(irq);
> > + list_add(&chip->list, chip_list);
> >
> > for (i = 0; i < PL061_GPIO_NR; i++) {
> > if (pdata->directions & (1 << i))
> > @@ -312,7 +308,7 @@ static int __init pl061_probe(struct amb
> > iounmap:
> > iounmap(chip->base);
> > release_region:
> > - release_mem_region(dev->res.start, SZ_4K);
> > + release_mem_region(dev->res.start, PL061_REG_SIZE);
> > free_mem:
> > kfree(chip);
> >
> > diff -puN include/linux/amba/pl061.h~gpio-driver-for-primecell-pl061-gpio-controller-v4 include/linux/amba/pl061.h
> > --- a/include/linux/amba/pl061.h~gpio-driver-for-primecell-pl061-gpio-controller-v4
> > +++ a/include/linux/amba/pl061.h
> > @@ -1,18 +1,17 @@
> > /* platform data for the PL061 GPIO driver */
> >
> > -#define GPIODIR 0x400
> > -#define GPIOIS 0x404
> > -#define GPIOIBE 0x408
> > -#define GPIOIEV 0x40C
> > -#define GPIOIE 0x410
> > -#define GPIORIS 0x414
> > -#define GPIOMIS 0x418
> > -#define GPIOIC 0x41C
> > -
> > struct pl061_platform_data {
> > /* number of the first GPIO */
> > unsigned gpio_base;
> > - u32 (*pending_irqs)(int irq);
> > +
> > u8 directions; /* startup directions, 1: out, 0: in */
> > u8 values; /* startup values */
> > +
> > + /* This callback may be used when you have more than one PL061
> > + * connected to the same IRQ trigger. This callback must return 0 on
> > + * the first time it is called for each irq, and 1 on any other call.
> > + * In case you are not unfortunate enough to have this setup, this
> > + * pointer must be set to NULL.
> > + */
> > + int (*is_irq_initialized)(int irq);
> > };
> > _
> >
> >

baruch

--
~. .~ Tk Open Systems
=}------------------------------------------------ooO--U--Ooo------------{=
- baruch@xxxxxxxxxx - tel: +972.2.679.5364, http://www.tkos.co.il -
--
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/