linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH v4 1/2] gpio: add GPO driver for PCA9570
@ 2020-07-02 12:17 Sungbo Eo
  2020-07-02 12:36 ` Andy Shevchenko
  0 siblings, 1 reply; 5+ messages in thread
From: Sungbo Eo @ 2020-07-02 12:17 UTC (permalink / raw)
  To: Linus Walleij, Bartosz Golaszewski, Andy Shevchenko,
	Michael Walle, linux-kernel, linux-gpio
  Cc: Sungbo Eo

NXP PCA9570 is 4-bit I2C GPO expander without interrupt functionality.
Its ports are controlled only by a data byte without register address.

As there is no other driver similar enough to be adapted for it, a new
driver is introduced here.

Signed-off-by: Sungbo Eo <mans0n@gorani.run>
---
v4:
* removed ->direction_input() and ->direction_output()
  (Seems unnecessary to me)
* removed ->set_multiple()
  (I'm not sure this implementation is really correct)
* added ->get()
  (DS says we can read the status from the device)
* read current status during probe

v3:
* remove mutex
* rename buffer to out
* simplify return statements
* replace ->probe() to ->probe_new()
* move ngpio to driver_data
  (PCA9571 is 8-bit so I thought making ngpio configurable is a good idea)

v2:
* move the direction functions below the set functions
* use devm_gpiochip_add_data() and remove the remove callback

v1:
Tested in kernel 5.4 on an ipq40xx platform.

This is my first time submitting a whole driver patch, and I'm not really
familiar with this PCA expander series.
Please let me know how I can improve this patch further.

FYI there's an unmerged patch for this chip.
http://driverdev.linuxdriverproject.org/pipermail/driverdev-devel/2017-May/105602.html
I don't have PCA9571 either so I didn't add support for it.
---
 drivers/gpio/Kconfig        |   8 +++
 drivers/gpio/Makefile       |   1 +
 drivers/gpio/gpio-pca9570.c | 138 ++++++++++++++++++++++++++++++++++++
 3 files changed, 147 insertions(+)
 create mode 100644 drivers/gpio/gpio-pca9570.c

diff --git a/drivers/gpio/Kconfig b/drivers/gpio/Kconfig
index c6b5c65c8405..d10dcb81b841 100644
--- a/drivers/gpio/Kconfig
+++ b/drivers/gpio/Kconfig
@@ -962,6 +962,14 @@ config GPIO_PCA953X_IRQ
 	  Say yes here to enable the pca953x to be used as an interrupt
 	  controller. It requires the driver to be built in the kernel.
 
+config GPIO_PCA9570
+	tristate "PCA9570 4-Bit I2C GPO expander"
+	help
+	  Say yes here to enable the GPO driver for the NXP PCA9570 chip.
+
+	  To compile this driver as a module, choose M here: the module will
+	  be called gpio-pca9570.
+
 config GPIO_PCF857X
 	tristate "PCF857x, PCA{85,96}7x, and MAX732[89] I2C GPIO expanders"
 	select GPIOLIB_IRQCHIP
diff --git a/drivers/gpio/Makefile b/drivers/gpio/Makefile
index 1e4894e0bf0f..33cb40c28a61 100644
--- a/drivers/gpio/Makefile
+++ b/drivers/gpio/Makefile
@@ -110,6 +110,7 @@ obj-$(CONFIG_GPIO_OCTEON)		+= gpio-octeon.o
 obj-$(CONFIG_GPIO_OMAP)			+= gpio-omap.o
 obj-$(CONFIG_GPIO_PALMAS)		+= gpio-palmas.o
 obj-$(CONFIG_GPIO_PCA953X)		+= gpio-pca953x.o
+obj-$(CONFIG_GPIO_PCA9570)		+= gpio-pca9570.o
 obj-$(CONFIG_GPIO_PCF857X)		+= gpio-pcf857x.o
 obj-$(CONFIG_GPIO_PCH)			+= gpio-pch.o
 obj-$(CONFIG_GPIO_PCIE_IDIO_24)		+= gpio-pcie-idio-24.o
diff --git a/drivers/gpio/gpio-pca9570.c b/drivers/gpio/gpio-pca9570.c
new file mode 100644
index 000000000000..fc031f2f03c8
--- /dev/null
+++ b/drivers/gpio/gpio-pca9570.c
@@ -0,0 +1,138 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Driver for PCA9570 I2C GPO expander
+ *
+ * Copyright (C) 2020 Sungbo Eo <mans0n@gorani.run>
+ *
+ * Based on gpio-tpic2810.c
+ * Copyright (C) 2015 Texas Instruments Incorporated - http://www.ti.com/
+ *	Andrew F. Davis <afd@ti.com>
+ */
+
+#include <linux/gpio/driver.h>
+#include <linux/i2c.h>
+#include <linux/module.h>
+
+/**
+ * struct pca9570 - GPIO driver data
+ * @chip: GPIO controller chip
+ * @client: I2C device pointer
+ * @out: Buffer for device register
+ */
+struct pca9570 {
+	struct gpio_chip chip;
+	struct i2c_client *client;
+	u8 out;
+};
+
+static int pca9570_read(struct pca9570 *gpio, u8 *value)
+{
+	s32 ret;
+
+	ret = i2c_smbus_read_byte(gpio->client);
+	if (ret < 0)
+		return ret;
+
+	*value = (u8)ret;
+	return 0;
+}
+
+static int pca9570_write(struct pca9570 *gpio, u8 value)
+{
+	return i2c_smbus_write_byte(gpio->client, value);
+}
+
+static int pca9570_get_direction(struct gpio_chip *chip,
+				 unsigned offset)
+{
+	/* This device always output */
+	return GPIO_LINE_DIRECTION_OUT;
+}
+
+static int pca9570_get(struct gpio_chip *chip, unsigned offset)
+{
+	struct pca9570 *gpio = gpiochip_get_data(chip);
+	u8 buffer;
+	int ret;
+
+	ret = pca9570_read(gpio, &buffer);
+	if (ret)
+		return ret;
+
+	return !!(buffer & BIT(offset));
+}
+
+static void pca9570_set(struct gpio_chip *chip, unsigned offset, int value)
+{
+	struct pca9570 *gpio = gpiochip_get_data(chip);
+	u8 buffer = gpio->out;
+	int ret;
+
+	if (value)
+		buffer |= BIT(offset);
+	else
+		buffer &= ~BIT(offset);
+
+	ret = pca9570_write(gpio, buffer);
+	if (ret)
+		return;
+
+	gpio->out = buffer;
+}
+
+static const struct gpio_chip template_chip = {
+	.label			= "pca9570",
+	.owner			= THIS_MODULE,
+	.get_direction		= pca9570_get_direction,
+	.get			= pca9570_get,
+	.set			= pca9570_set,
+	.base			= -1,
+	.can_sleep		= true,
+};
+
+static const struct i2c_device_id pca9570_id_table[] = {
+	{ "pca9570", 4 },
+	{ /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(i2c, pca9570_id_table);
+
+static const struct of_device_id pca9570_of_match_table[] = {
+	{ .compatible = "nxp,pca9570" },
+	{ /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(of, pca9570_of_match_table);
+
+static int pca9570_probe(struct i2c_client *client)
+{
+	struct pca9570 *gpio;
+
+	gpio = devm_kzalloc(&client->dev, sizeof(*gpio), GFP_KERNEL);
+	if (!gpio)
+		return -ENOMEM;
+
+	gpio->chip = template_chip;
+	gpio->chip.parent = &client->dev;
+	gpio->chip.ngpio = i2c_match_id(pca9570_id_table, client)->driver_data;
+	gpio->client = client;
+
+	/* Read the current output level */
+	(void) pca9570_read(gpio, &gpio->out);
+
+	i2c_set_clientdata(client, gpio);
+
+	return devm_gpiochip_add_data(&client->dev, &gpio->chip, gpio);
+}
+
+static struct i2c_driver pca9570_driver = {
+	.driver = {
+		.name = "pca9570",
+		.of_match_table = pca9570_of_match_table,
+	},
+	.probe_new = pca9570_probe,
+	.id_table = pca9570_id_table,
+};
+module_i2c_driver(pca9570_driver);
+
+MODULE_AUTHOR("Sungbo Eo <mans0n@gorani.run>");
+MODULE_DESCRIPTION("GPIO expander driver for PCA9570");
+MODULE_LICENSE("GPL v2");
-- 
2.27.0


^ permalink raw reply related	[flat|nested] 5+ messages in thread

* Re: [PATCH v4 1/2] gpio: add GPO driver for PCA9570
  2020-07-02 12:17 [PATCH v4 1/2] gpio: add GPO driver for PCA9570 Sungbo Eo
@ 2020-07-02 12:36 ` Andy Shevchenko
  2020-07-02 12:38   ` Andy Shevchenko
  2020-07-03  0:12   ` Sungbo Eo
  0 siblings, 2 replies; 5+ messages in thread
From: Andy Shevchenko @ 2020-07-02 12:36 UTC (permalink / raw)
  To: Sungbo Eo
  Cc: Linus Walleij, Bartosz Golaszewski, Michael Walle,
	Linux Kernel Mailing List, open list:GPIO SUBSYSTEM

On Thu, Jul 2, 2020 at 3:18 PM Sungbo Eo <mans0n@gorani.run> wrote:
>
> NXP PCA9570 is 4-bit I2C GPO expander without interrupt functionality.

is a 4-bit

> Its ports are controlled only by a data byte without register address.
>
> As there is no other driver similar enough to be adapted for it, a new
> driver is introduced here.

Can we have a Datasheet: tag with URL attached to it?

...

> +static int pca9570_read(struct pca9570 *gpio, u8 *value)
> +{
> +       s32 ret;
> +
> +       ret = i2c_smbus_read_byte(gpio->client);
> +       if (ret < 0)
> +               return ret;
> +

> +       *value = (u8)ret;

I doubt casting is needed.

> +       return 0;
> +}

...

> +static const struct i2c_device_id pca9570_id_table[] = {
> +       { "pca9570", 4 },
> +       { /* sentinel */ }
> +};
> +MODULE_DEVICE_TABLE(i2c, pca9570_id_table);
> +
> +static const struct of_device_id pca9570_of_match_table[] = {

> +       { .compatible = "nxp,pca9570" },

This driver data should also have 4.

> +       { /* sentinel */ }
> +};

...

> +       gpio->chip.ngpio = i2c_match_id(pca9570_id_table, client)->driver_data;

Oh, avoid direct access to the table like this. And you may simply use
device_get_match_data().
...

> +       /* Read the current output level */
> +       (void) pca9570_read(gpio, &gpio->out);

(void) casting is not needed. And I'm not sure hiding an error is a
good idea. But the latter is up to you.

-- 
With Best Regards,
Andy Shevchenko

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH v4 1/2] gpio: add GPO driver for PCA9570
  2020-07-02 12:36 ` Andy Shevchenko
@ 2020-07-02 12:38   ` Andy Shevchenko
  2020-07-03  0:12   ` Sungbo Eo
  1 sibling, 0 replies; 5+ messages in thread
From: Andy Shevchenko @ 2020-07-02 12:38 UTC (permalink / raw)
  To: Sungbo Eo
  Cc: Linus Walleij, Bartosz Golaszewski, Michael Walle,
	Linux Kernel Mailing List, open list:GPIO SUBSYSTEM

On Thu, Jul 2, 2020 at 3:36 PM Andy Shevchenko
<andy.shevchenko@gmail.com> wrote:
> On Thu, Jul 2, 2020 at 3:18 PM Sungbo Eo <mans0n@gorani.run> wrote:

> > +       s32 ret;

And why not simple int ret; ?

-- 
With Best Regards,
Andy Shevchenko

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH v4 1/2] gpio: add GPO driver for PCA9570
  2020-07-02 12:36 ` Andy Shevchenko
  2020-07-02 12:38   ` Andy Shevchenko
@ 2020-07-03  0:12   ` Sungbo Eo
  2020-07-04 10:05     ` Andy Shevchenko
  1 sibling, 1 reply; 5+ messages in thread
From: Sungbo Eo @ 2020-07-03  0:12 UTC (permalink / raw)
  To: Andy Shevchenko
  Cc: Linus Walleij, Bartosz Golaszewski, Michael Walle,
	Linux Kernel Mailing List, open list:GPIO SUBSYSTEM

On 2020-07-02 21:36, Andy Shevchenko wrote:
>> +       gpio->chip.ngpio = i2c_match_id(pca9570_id_table, client)->driver_data;
> 
> Oh, avoid direct access to the table like this. And you may simply use
> device_get_match_data().

I'm not sure if it really does the same thing, but I'll try following 
your suggestion.

> ...
> 
>> +       /* Read the current output level */
>> +       (void) pca9570_read(gpio, &gpio->out);
> 
> (void) casting is not needed. And I'm not sure hiding an error is a
> good idea. But the latter is up to you.
> 

If it returns an error then it might be because the chip could not be 
detected on the bus at that time. But I think aborting probe for that is 
too much.
(void) casting was to indicate that I want to ignore the error, but I'll 
remove it as you said.

Thanks.

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH v4 1/2] gpio: add GPO driver for PCA9570
  2020-07-03  0:12   ` Sungbo Eo
@ 2020-07-04 10:05     ` Andy Shevchenko
  0 siblings, 0 replies; 5+ messages in thread
From: Andy Shevchenko @ 2020-07-04 10:05 UTC (permalink / raw)
  To: Sungbo Eo
  Cc: Linus Walleij, Bartosz Golaszewski, Michael Walle,
	Linux Kernel Mailing List, open list:GPIO SUBSYSTEM

On Fri, Jul 3, 2020 at 3:13 AM Sungbo Eo <mans0n@gorani.run> wrote:
> On 2020-07-02 21:36, Andy Shevchenko wrote:
> >> +       gpio->chip.ngpio = i2c_match_id(pca9570_id_table, client)->driver_data;
> >
> > Oh, avoid direct access to the table like this. And you may simply use
> > device_get_match_data().
>
> I'm not sure if it really does the same thing, but I'll try following
> your suggestion.

It's an agnostic way to get it from fwnode enumerated devices. Since
you are using ->probe_new() the driver should rely on fwnode, that's
why the OF table has to have driver data be defined.

-- 
With Best Regards,
Andy Shevchenko

^ permalink raw reply	[flat|nested] 5+ messages in thread

end of thread, other threads:[~2020-07-04 10:06 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-07-02 12:17 [PATCH v4 1/2] gpio: add GPO driver for PCA9570 Sungbo Eo
2020-07-02 12:36 ` Andy Shevchenko
2020-07-02 12:38   ` Andy Shevchenko
2020-07-03  0:12   ` Sungbo Eo
2020-07-04 10:05     ` Andy Shevchenko

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).