[PATCH v2 0/5] i2c: pca-platform: updates

[PATCH v2 0/5] i2c: pca-platform: updates

[Chris Packham]

We're using a PCA9564 on an embedded platform to provide an i2c
controller. This adds devicetree support and I've included some
cleanups for the driver while I was at it.

Chris Packham (5):
  dt-bindings: add bindings for i2c-pca-platform
  i2c: pca-platform: add devicetree awareness
  i2c: pca-platform: use gpio_is_valid
  i2c: pca-platform: use device managed allocations
  i2c: pca-platform: use dev_warn/dev_info instead of printk

 .../devicetree/bindings/i2c/i2c-pca-platform.txt   | 18 ++++
 drivers/i2c/busses/i2c-pca-platform.c              | 97 +++++++++-------------
 2 files changed, 57 insertions(+), 58 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/i2c/i2c-pca-platform.txt

-- 
2.13.0

[PATCH v2 1/5] dt-bindings: add bindings for i2c-pca-platform

[Chris Packham]

Signed-off-by: Chris Packham <chris.packham@alliedtelesis.co.nz>
---
Changes in v2:
- Split dt-binding to separate patch, use "reset-gpios" instead of "gpios".

 .../devicetree/bindings/i2c/i2c-pca-platform.txt       | 18 ++++++++++++++++++
 1 file changed, 18 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/i2c/i2c-pca-platform.txt

diff --git a/Documentation/devicetree/bindings/i2c/i2c-pca-platform.txt b/Documentation/devicetree/bindings/i2c/i2c-pca-platform.txt
new file mode 100644
index 000000000000..86abf249b194
--- /dev/null
+++ b/Documentation/devicetree/bindings/i2c/i2c-pca-platform.txt
@@ -0,0 +1,18 @@
+* NXP PCA PCA9564/PCA9665 I2C controller
+
+The PCA9564/PCA9665 serves as an interface between most standard
+parallel-bus microcontrollers/microprocessors and the serial I2C-bus
+and allows the parallel bus system to communicate bi-directionally
+with the I2C-bus.
+
+Required properties :
+
+ - reg : Offset and length of the register set for the device
+ - compatible : one of "nxp,pca9564" or "nxp,pca9665"
+
+Optional properties
+ - interrupts : the interrupt number
+ - interrupt-parent : the phandle for the interrupt controller.
+   If an interrupt is not specified polling will be used.
+ - reset-gpios : gpio specifier for gpio connected to RESET_N pin.
+ - clock-frequency : I2C bus frequency.
-- 
2.13.0

Re: [PATCH v2 1/5] dt-bindings: add bindings for i2c-pca-platform

[Wolfram Sang]

[-- Attachment #1: Type: text/plain, Size: 142 bytes --]


> + - reset-gpios : gpio specifier for gpio connected to RESET_N pin.

What about the 'active' state that Rob mentioned in his last review?


[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: [PATCH v2 1/5] dt-bindings: add bindings for i2c-pca-platform

[Wolfram Sang]

[-- Attachment #1: Type: text/plain, Size: 142 bytes --]


> + - reset-gpios : gpio specifier for gpio connected to RESET_N pin.

What about the 'active' state that Rob mentioned in his last review?


[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: [PATCH v2 1/5] dt-bindings: add bindings for i2c-pca-platform

[Chris Packham]

On 22/06/17 20:23, Wolfram Sang wrote:
> 
>> + - reset-gpios : gpio specifier for gpio connected to RESET_N pin.
> 
> What about the 'active' state that Rob mentioned in his last review?
> 

My intention was that by saying it is connected to the RESET_N bin the 
active state is covered.

I personally always get a little confused when talking about reset 
lines. The _N denotes that the line is active low but because it is a 
reset line writing 1 releases the reset so from a what-the-user-wants 
perspective it's active high.

dt-binding-wise what we want here is GPIO_ACTIVE_HIGH, as in the driver 
writes 1 to release reset so don't do any polarity inversion. I'd be 
happy do add something to that effect in a v3 but I actually felt saying 
the gpio is connected to RESET_N was less ambiguous.

Re: [PATCH v2 1/5] dt-bindings: add bindings for i2c-pca-platform

[Wolfram Sang]

[-- Attachment #1: Type: text/plain, Size: 1021 bytes --]

On Thu, Jun 22, 2017 at 09:20:59PM +0000, Chris Packham wrote:
> On 22/06/17 20:23, Wolfram Sang wrote:
> > 
> >> + - reset-gpios : gpio specifier for gpio connected to RESET_N pin.
> > 
> > What about the 'active' state that Rob mentioned in his last review?
> > 
> 
> My intention was that by saying it is connected to the RESET_N bin the 
> active state is covered.
> 
> I personally always get a little confused when talking about reset 
> lines. The _N denotes that the line is active low but because it is a 
> reset line writing 1 releases the reset so from a what-the-user-wants 
> perspective it's active high.
> 
> dt-binding-wise what we want here is GPIO_ACTIVE_HIGH, as in the driver 
> writes 1 to release reset so don't do any polarity inversion. I'd be 
> happy do add something to that effect in a v3 but I actually felt saying 
> the gpio is connected to RESET_N was less ambiguous.

Okay. Thanks for the explanation. I am fine with it but I'll leave the
final decision to Rob.


[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: [PATCH v2 1/5] dt-bindings: add bindings for i2c-pca-platform

[Wolfram Sang]

[-- Attachment #1: Type: text/plain, Size: 1021 bytes --]

On Thu, Jun 22, 2017 at 09:20:59PM +0000, Chris Packham wrote:
> On 22/06/17 20:23, Wolfram Sang wrote:
> > 
> >> + - reset-gpios : gpio specifier for gpio connected to RESET_N pin.
> > 
> > What about the 'active' state that Rob mentioned in his last review?
> > 
> 
> My intention was that by saying it is connected to the RESET_N bin the 
> active state is covered.
> 
> I personally always get a little confused when talking about reset 
> lines. The _N denotes that the line is active low but because it is a 
> reset line writing 1 releases the reset so from a what-the-user-wants 
> perspective it's active high.
> 
> dt-binding-wise what we want here is GPIO_ACTIVE_HIGH, as in the driver 
> writes 1 to release reset so don't do any polarity inversion. I'd be 
> happy do add something to that effect in a v3 but I actually felt saying 
> the gpio is connected to RESET_N was less ambiguous.

Okay. Thanks for the explanation. I am fine with it but I'll leave the
final decision to Rob.


[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: [PATCH v2 1/5] dt-bindings: add bindings for i2c-pca-platform

[Rob Herring]

On Thu, Jun 22, 2017 at 09:20:59PM +0000, Chris Packham wrote:
> On 22/06/17 20:23, Wolfram Sang wrote:
> > 
> >> + - reset-gpios : gpio specifier for gpio connected to RESET_N pin.
> > 
> > What about the 'active' state that Rob mentioned in his last review?
> > 
> 
> My intention was that by saying it is connected to the RESET_N bin the 
> active state is covered.

I would say okay, but based on below it's not.

> I personally always get a little confused when talking about reset 
> lines. The _N denotes that the line is active low but because it is a 
> reset line writing 1 releases the reset so from a what-the-user-wants 
> perspective it's active high.
> 
> dt-binding-wise what we want here is GPIO_ACTIVE_HIGH, as in the driver 
> writes 1 to release reset so don't do any polarity inversion. I'd be 
> happy do add something to that effect in a v3 but I actually felt saying 
> the gpio is connected to RESET_N was less ambiguous.

No, the driver is wrong. The binding should say GPIO_ACTIVE_LOW because 
that is how the pin is defined. The driver needs to set it to inactive.

Rob

[PATCH v2 2/5] i2c: pca-platform: add devicetree awareness

[Chris Packham]

Allow devices that use this driver to be registered via a
devicetree.

Signed-off-by: Chris Packham <chris.packham@alliedtelesis.co.nz>
---
Changes in v2:
- Set i2c->adap.dev.of_node so that child nodes are automatically probed
- Split dt-binding to separate patch, use "reset-gpios" instead of "gpios".

 drivers/i2c/busses/i2c-pca-platform.c | 22 ++++++++++++++++++++++
 1 file changed, 22 insertions(+)

diff --git a/drivers/i2c/busses/i2c-pca-platform.c b/drivers/i2c/busses/i2c-pca-platform.c
index 3bd2e7d06e4b..87ca2f02f607 100644
--- a/drivers/i2c/busses/i2c-pca-platform.c
+++ b/drivers/i2c/busses/i2c-pca-platform.c
@@ -23,6 +23,9 @@
 #include <linux/i2c-pca-platform.h>
 #include <linux/gpio.h>
 #include <linux/io.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+#include <linux/of_gpio.h>
 
 #include <asm/irq.h>
 
@@ -136,12 +139,15 @@ static int i2c_pca_pf_probe(struct platform_device *pdev)
 	struct resource *res;
 	struct i2c_pca9564_pf_platform_data *platform_data =
 				dev_get_platdata(&pdev->dev);
+	struct device_node *np = pdev->dev.of_node;
 	int ret = 0;
 	int irq;
 
 	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
 	irq = platform_get_irq(pdev, 0);
 	/* If irq is 0, we do polling. */
+	if (irq < 0)
+		irq = 0;
 
 	if (res == NULL) {
 		ret = -ENODEV;
@@ -177,11 +183,17 @@ static int i2c_pca_pf_probe(struct platform_device *pdev)
 		 (unsigned long) res->start);
 	i2c->adap.algo_data = &i2c->algo_data;
 	i2c->adap.dev.parent = &pdev->dev;
+	i2c->adap.dev.of_node = np;
 
 	if (platform_data) {
 		i2c->adap.timeout = platform_data->timeout;
 		i2c->algo_data.i2c_clock = platform_data->i2c_clock_speed;
 		i2c->gpio = platform_data->gpio;
+	} else if (np) {
+		i2c->adap.timeout = HZ;
+		i2c->gpio = of_get_named_gpio(np, "reset-gpios", 0);
+		of_property_read_u32_index(np, "clock-frequency", 0,
+					   &i2c->algo_data.i2c_clock);
 	} else {
 		i2c->adap.timeout = HZ;
 		i2c->algo_data.i2c_clock = 59000;
@@ -275,11 +287,21 @@ static int i2c_pca_pf_remove(struct platform_device *pdev)
 	return 0;
 }
 
+#ifdef CONFIG_OF
+static const struct of_device_id i2c_pca_of_match_table[] = {
+	{ .compatible = "nxp,pca9564" },
+	{ .compatible = "nxp,pca9665" },
+	{},
+};
+MODULE_DEVICE_TABLE(of, i2c_pca_of_match_table);
+#endif
+
 static struct platform_driver i2c_pca_pf_driver = {
 	.probe = i2c_pca_pf_probe,
 	.remove = i2c_pca_pf_remove,
 	.driver = {
 		.name = "i2c-pca-platform",
+		.of_match_table = of_match_ptr(i2c_pca_of_match_table),
 	},
 };
 
-- 
2.13.0

[PATCH v2 3/5] i2c: pca-platform: use gpio_is_valid

[Chris Packham]

Use gpio_is_valid() instead of gpio > -1.

Signed-off-by: Chris Packham <chris.packham@alliedtelesis.co.nz>
---
 drivers/i2c/busses/i2c-pca-platform.c | 9 ++++-----
 1 file changed, 4 insertions(+), 5 deletions(-)

diff --git a/drivers/i2c/busses/i2c-pca-platform.c b/drivers/i2c/busses/i2c-pca-platform.c
index 87ca2f02f607..3a9845bae9a4 100644
--- a/drivers/i2c/busses/i2c-pca-platform.c
+++ b/drivers/i2c/busses/i2c-pca-platform.c
@@ -197,7 +197,7 @@ static int i2c_pca_pf_probe(struct platform_device *pdev)
 	} else {
 		i2c->adap.timeout = HZ;
 		i2c->algo_data.i2c_clock = 59000;
-		i2c->gpio = -1;
+		i2c->gpio = -ENODEV;
 	}
 
 	i2c->algo_data.data = i2c;
@@ -220,8 +220,7 @@ static int i2c_pca_pf_probe(struct platform_device *pdev)
 		break;
 	}
 
-	/* Use gpio_is_valid() when in mainline */
-	if (i2c->gpio > -1) {
+	if (gpio_is_valid(i2c->gpio)) {
 		ret = gpio_request(i2c->gpio, i2c->adap.name);
 		if (ret == 0) {
 			gpio_direction_output(i2c->gpio, 1);
@@ -255,7 +254,7 @@ static int i2c_pca_pf_probe(struct platform_device *pdev)
 	if (irq)
 		free_irq(irq, i2c);
 e_reqirq:
-	if (i2c->gpio > -1)
+	if (gpio_is_valid(i2c->gpio))
 		gpio_free(i2c->gpio);
 
 	iounmap(i2c->reg_base);
@@ -277,7 +276,7 @@ static int i2c_pca_pf_remove(struct platform_device *pdev)
 	if (i2c->irq)
 		free_irq(i2c->irq, i2c);
 
-	if (i2c->gpio > -1)
+	if (gpio_is_valid(i2c->gpio))
 		gpio_free(i2c->gpio);
 
 	iounmap(i2c->reg_base);
-- 
2.13.0

[PATCH v2 3/5] i2c: pca-platform: use gpio_is_valid

[Chris Packham]

Use gpio_is_valid() instead of gpio > -1.

Signed-off-by: Chris Packham <chris.packham-6g8wRflRTwXFdCa3tKVlE6U/zSkkHjvu@public.gmane.org>
---
 drivers/i2c/busses/i2c-pca-platform.c | 9 ++++-----
 1 file changed, 4 insertions(+), 5 deletions(-)

diff --git a/drivers/i2c/busses/i2c-pca-platform.c b/drivers/i2c/busses/i2c-pca-platform.c
index 87ca2f02f607..3a9845bae9a4 100644
--- a/drivers/i2c/busses/i2c-pca-platform.c
+++ b/drivers/i2c/busses/i2c-pca-platform.c
@@ -197,7 +197,7 @@ static int i2c_pca_pf_probe(struct platform_device *pdev)
 	} else {
 		i2c->adap.timeout = HZ;
 		i2c->algo_data.i2c_clock = 59000;
-		i2c->gpio = -1;
+		i2c->gpio = -ENODEV;
 	}
 
 	i2c->algo_data.data = i2c;
@@ -220,8 +220,7 @@ static int i2c_pca_pf_probe(struct platform_device *pdev)
 		break;
 	}
 
-	/* Use gpio_is_valid() when in mainline */
-	if (i2c->gpio > -1) {
+	if (gpio_is_valid(i2c->gpio)) {
 		ret = gpio_request(i2c->gpio, i2c->adap.name);
 		if (ret == 0) {
 			gpio_direction_output(i2c->gpio, 1);
@@ -255,7 +254,7 @@ static int i2c_pca_pf_probe(struct platform_device *pdev)
 	if (irq)
 		free_irq(irq, i2c);
 e_reqirq:
-	if (i2c->gpio > -1)
+	if (gpio_is_valid(i2c->gpio))
 		gpio_free(i2c->gpio);
 
 	iounmap(i2c->reg_base);
@@ -277,7 +276,7 @@ static int i2c_pca_pf_remove(struct platform_device *pdev)
 	if (i2c->irq)
 		free_irq(i2c->irq, i2c);
 
-	if (i2c->gpio > -1)
+	if (gpio_is_valid(i2c->gpio))
 		gpio_free(i2c->gpio);
 
 	iounmap(i2c->reg_base);
-- 
2.13.0

--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH v2 4/5] i2c: pca-platform: use device managed allocations

[Chris Packham]

Switch to using the devm_ APIs and remove the now unnecessary error
handling and most of the device removal code.

Signed-off-by: Chris Packham <chris.packham@alliedtelesis.co.nz>
---
 drivers/i2c/busses/i2c-pca-platform.c | 61 +++++++----------------------------
 1 file changed, 11 insertions(+), 50 deletions(-)

diff --git a/drivers/i2c/busses/i2c-pca-platform.c b/drivers/i2c/busses/i2c-pca-platform.c
index 3a9845bae9a4..12ab0cafa7c9 100644
--- a/drivers/i2c/busses/i2c-pca-platform.c
+++ b/drivers/i2c/busses/i2c-pca-platform.c
@@ -143,35 +143,23 @@ static int i2c_pca_pf_probe(struct platform_device *pdev)
 	int ret = 0;
 	int irq;
 
-	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
 	irq = platform_get_irq(pdev, 0);
 	/* If irq is 0, we do polling. */
 	if (irq < 0)
 		irq = 0;
 
-	if (res == NULL) {
-		ret = -ENODEV;
-		goto e_print;
-	}
+	i2c = devm_kzalloc(&pdev->dev, sizeof(*i2c), GFP_KERNEL);
+	if (!i2c)
+		return -ENOMEM;
 
-	if (!request_mem_region(res->start, resource_size(res), res->name)) {
-		ret = -ENOMEM;
-		goto e_print;
-	}
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	i2c->reg_base = devm_ioremap_resource(&pdev->dev, res);
+	if (IS_ERR(i2c->reg_base))
+		return PTR_ERR(i2c->reg_base);
 
-	i2c = kzalloc(sizeof(struct i2c_pca_pf_data), GFP_KERNEL);
-	if (!i2c) {
-		ret = -ENOMEM;
-		goto e_alloc;
-	}
 
 	init_waitqueue_head(&i2c->wait);
 
-	i2c->reg_base = ioremap(res->start, resource_size(res));
-	if (!i2c->reg_base) {
-		ret = -ENOMEM;
-		goto e_remap;
-	}
 	i2c->io_base = res->start;
 	i2c->io_size = resource_size(res);
 	i2c->irq = irq;
@@ -221,7 +209,7 @@ static int i2c_pca_pf_probe(struct platform_device *pdev)
 	}
 
 	if (gpio_is_valid(i2c->gpio)) {
-		ret = gpio_request(i2c->gpio, i2c->adap.name);
+		ret = devm_gpio_request(&pdev->dev, i2c->gpio, i2c->adap.name);
 		if (ret == 0) {
 			gpio_direction_output(i2c->gpio, 1);
 			i2c->algo_data.reset_chip = i2c_pca_pf_resetchip;
@@ -233,15 +221,14 @@ static int i2c_pca_pf_probe(struct platform_device *pdev)
 	}
 
 	if (irq) {
-		ret = request_irq(irq, i2c_pca_pf_handler,
+		ret = devm_request_irq(&pdev->dev, irq, i2c_pca_pf_handler,
 			IRQF_TRIGGER_FALLING, pdev->name, i2c);
 		if (ret)
-			goto e_reqirq;
+			return ret;
 	}
 
 	if (i2c_pca_add_numbered_bus(&i2c->adap) < 0) {
-		ret = -ENODEV;
-		goto e_adapt;
+		return -ENODEV;
 	}
 
 	platform_set_drvdata(pdev, i2c);
@@ -249,22 +236,6 @@ static int i2c_pca_pf_probe(struct platform_device *pdev)
 	printk(KERN_INFO "%s registered.\n", i2c->adap.name);
 
 	return 0;
-
-e_adapt:
-	if (irq)
-		free_irq(irq, i2c);
-e_reqirq:
-	if (gpio_is_valid(i2c->gpio))
-		gpio_free(i2c->gpio);
-
-	iounmap(i2c->reg_base);
-e_remap:
-	kfree(i2c);
-e_alloc:
-	release_mem_region(res->start, resource_size(res));
-e_print:
-	printk(KERN_ERR "Registering PCA9564/PCA9665 FAILED! (%d)\n", ret);
-	return ret;
 }
 
 static int i2c_pca_pf_remove(struct platform_device *pdev)
@@ -273,16 +244,6 @@ static int i2c_pca_pf_remove(struct platform_device *pdev)
 
 	i2c_del_adapter(&i2c->adap);
 
-	if (i2c->irq)
-		free_irq(i2c->irq, i2c);
-
-	if (gpio_is_valid(i2c->gpio))
-		gpio_free(i2c->gpio);
-
-	iounmap(i2c->reg_base);
-	release_mem_region(i2c->io_base, i2c->io_size);
-	kfree(i2c);
-
 	return 0;
 }
 
-- 
2.13.0

[PATCH v2 4/5] i2c: pca-platform: use device managed allocations

[Chris Packham]

Switch to using the devm_ APIs and remove the now unnecessary error
handling and most of the device removal code.

Signed-off-by: Chris Packham <chris.packham-6g8wRflRTwXFdCa3tKVlE6U/zSkkHjvu@public.gmane.org>
---
 drivers/i2c/busses/i2c-pca-platform.c | 61 +++++++----------------------------
 1 file changed, 11 insertions(+), 50 deletions(-)

diff --git a/drivers/i2c/busses/i2c-pca-platform.c b/drivers/i2c/busses/i2c-pca-platform.c
index 3a9845bae9a4..12ab0cafa7c9 100644
--- a/drivers/i2c/busses/i2c-pca-platform.c
+++ b/drivers/i2c/busses/i2c-pca-platform.c
@@ -143,35 +143,23 @@ static int i2c_pca_pf_probe(struct platform_device *pdev)
 	int ret = 0;
 	int irq;
 
-	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
 	irq = platform_get_irq(pdev, 0);
 	/* If irq is 0, we do polling. */
 	if (irq < 0)
 		irq = 0;
 
-	if (res == NULL) {
-		ret = -ENODEV;
-		goto e_print;
-	}
+	i2c = devm_kzalloc(&pdev->dev, sizeof(*i2c), GFP_KERNEL);
+	if (!i2c)
+		return -ENOMEM;
 
-	if (!request_mem_region(res->start, resource_size(res), res->name)) {
-		ret = -ENOMEM;
-		goto e_print;
-	}
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	i2c->reg_base = devm_ioremap_resource(&pdev->dev, res);
+	if (IS_ERR(i2c->reg_base))
+		return PTR_ERR(i2c->reg_base);
 
-	i2c = kzalloc(sizeof(struct i2c_pca_pf_data), GFP_KERNEL);
-	if (!i2c) {
-		ret = -ENOMEM;
-		goto e_alloc;
-	}
 
 	init_waitqueue_head(&i2c->wait);
 
-	i2c->reg_base = ioremap(res->start, resource_size(res));
-	if (!i2c->reg_base) {
-		ret = -ENOMEM;
-		goto e_remap;
-	}
 	i2c->io_base = res->start;
 	i2c->io_size = resource_size(res);
 	i2c->irq = irq;
@@ -221,7 +209,7 @@ static int i2c_pca_pf_probe(struct platform_device *pdev)
 	}
 
 	if (gpio_is_valid(i2c->gpio)) {
-		ret = gpio_request(i2c->gpio, i2c->adap.name);
+		ret = devm_gpio_request(&pdev->dev, i2c->gpio, i2c->adap.name);
 		if (ret == 0) {
 			gpio_direction_output(i2c->gpio, 1);
 			i2c->algo_data.reset_chip = i2c_pca_pf_resetchip;
@@ -233,15 +221,14 @@ static int i2c_pca_pf_probe(struct platform_device *pdev)
 	}
 
 	if (irq) {
-		ret = request_irq(irq, i2c_pca_pf_handler,
+		ret = devm_request_irq(&pdev->dev, irq, i2c_pca_pf_handler,
 			IRQF_TRIGGER_FALLING, pdev->name, i2c);
 		if (ret)
-			goto e_reqirq;
+			return ret;
 	}
 
 	if (i2c_pca_add_numbered_bus(&i2c->adap) < 0) {
-		ret = -ENODEV;
-		goto e_adapt;
+		return -ENODEV;
 	}
 
 	platform_set_drvdata(pdev, i2c);
@@ -249,22 +236,6 @@ static int i2c_pca_pf_probe(struct platform_device *pdev)
 	printk(KERN_INFO "%s registered.\n", i2c->adap.name);
 
 	return 0;
-
-e_adapt:
-	if (irq)
-		free_irq(irq, i2c);
-e_reqirq:
-	if (gpio_is_valid(i2c->gpio))
-		gpio_free(i2c->gpio);
-
-	iounmap(i2c->reg_base);
-e_remap:
-	kfree(i2c);
-e_alloc:
-	release_mem_region(res->start, resource_size(res));
-e_print:
-	printk(KERN_ERR "Registering PCA9564/PCA9665 FAILED! (%d)\n", ret);
-	return ret;
 }
 
 static int i2c_pca_pf_remove(struct platform_device *pdev)
@@ -273,16 +244,6 @@ static int i2c_pca_pf_remove(struct platform_device *pdev)
 
 	i2c_del_adapter(&i2c->adap);
 
-	if (i2c->irq)
-		free_irq(i2c->irq, i2c);
-
-	if (gpio_is_valid(i2c->gpio))
-		gpio_free(i2c->gpio);
-
-	iounmap(i2c->reg_base);
-	release_mem_region(i2c->io_base, i2c->io_size);
-	kfree(i2c);
-
 	return 0;
 }
 
-- 
2.13.0

--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH v2 5/5] i2c: pca-platform: use dev_warn/dev_info instead of printk

[Chris Packham]

Signed-off-by: Chris Packham <chris.packham@alliedtelesis.co.nz>
---
 drivers/i2c/busses/i2c-pca-platform.c | 9 ++++-----
 1 file changed, 4 insertions(+), 5 deletions(-)

diff --git a/drivers/i2c/busses/i2c-pca-platform.c b/drivers/i2c/busses/i2c-pca-platform.c
index 12ab0cafa7c9..27175d681548 100644
--- a/drivers/i2c/busses/i2c-pca-platform.c
+++ b/drivers/i2c/busses/i2c-pca-platform.c
@@ -107,8 +107,8 @@ static int i2c_pca_pf_waitforcompletion(void *pd)
 static void i2c_pca_pf_dummyreset(void *pd)
 {
 	struct i2c_pca_pf_data *i2c = pd;
-	printk(KERN_WARNING "%s: No reset-pin found. Chip may get stuck!\n",
-		i2c->adap.name);
+
+	dev_warn(&i2c->adap.dev, "No reset-pin found. Chip may get stuck!\n");
 }
 
 static void i2c_pca_pf_resetchip(void *pd)
@@ -214,8 +214,7 @@ static int i2c_pca_pf_probe(struct platform_device *pdev)
 			gpio_direction_output(i2c->gpio, 1);
 			i2c->algo_data.reset_chip = i2c_pca_pf_resetchip;
 		} else {
-			printk(KERN_WARNING "%s: Registering gpio failed!\n",
-				i2c->adap.name);
+			dev_warn(&pdev->dev, "Registering gpio failed!\n");
 			i2c->gpio = ret;
 		}
 	}
@@ -233,7 +232,7 @@ static int i2c_pca_pf_probe(struct platform_device *pdev)
 
 	platform_set_drvdata(pdev, i2c);
 
-	printk(KERN_INFO "%s registered.\n", i2c->adap.name);
+	dev_info(&pdev->dev, "registered.\n");
 
 	return 0;
 }
-- 
2.13.0

[PATCH v2 5/5] i2c: pca-platform: use dev_warn/dev_info instead of printk

[Chris Packham]

Signed-off-by: Chris Packham <chris.packham-6g8wRflRTwXFdCa3tKVlE6U/zSkkHjvu@public.gmane.org>
---
 drivers/i2c/busses/i2c-pca-platform.c | 9 ++++-----
 1 file changed, 4 insertions(+), 5 deletions(-)

diff --git a/drivers/i2c/busses/i2c-pca-platform.c b/drivers/i2c/busses/i2c-pca-platform.c
index 12ab0cafa7c9..27175d681548 100644
--- a/drivers/i2c/busses/i2c-pca-platform.c
+++ b/drivers/i2c/busses/i2c-pca-platform.c
@@ -107,8 +107,8 @@ static int i2c_pca_pf_waitforcompletion(void *pd)
 static void i2c_pca_pf_dummyreset(void *pd)
 {
 	struct i2c_pca_pf_data *i2c = pd;
-	printk(KERN_WARNING "%s: No reset-pin found. Chip may get stuck!\n",
-		i2c->adap.name);
+
+	dev_warn(&i2c->adap.dev, "No reset-pin found. Chip may get stuck!\n");
 }
 
 static void i2c_pca_pf_resetchip(void *pd)
@@ -214,8 +214,7 @@ static int i2c_pca_pf_probe(struct platform_device *pdev)
 			gpio_direction_output(i2c->gpio, 1);
 			i2c->algo_data.reset_chip = i2c_pca_pf_resetchip;
 		} else {
-			printk(KERN_WARNING "%s: Registering gpio failed!\n",
-				i2c->adap.name);
+			dev_warn(&pdev->dev, "Registering gpio failed!\n");
 			i2c->gpio = ret;
 		}
 	}
@@ -233,7 +232,7 @@ static int i2c_pca_pf_probe(struct platform_device *pdev)
 
 	platform_set_drvdata(pdev, i2c);
 
-	printk(KERN_INFO "%s registered.\n", i2c->adap.name);
+	dev_info(&pdev->dev, "registered.\n");
 
 	return 0;
 }
-- 
2.13.0

--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH v3 0/5] i2c: pca-platform: updates

[Chris Packham]

We're using a PCA9564 on an embedded platform to provide an i2c
controller. This adds devicetree support and I've included some
cleanups for the driver while I was at it.

Chris Packham (5):
  dt-bindings: add bindings for i2c-pca-platform
  i2c: pca-platform: switch to struct gpio_desc
  i2c: pca-platform: add devicetree awareness
  i2c: pca-platform: use device managed allocations
  i2c: pca-platform: use dev_warn/dev_info instead of printk

 .../devicetree/bindings/i2c/i2c-pca-platform.txt   |  29 +++++
 drivers/i2c/busses/i2c-pca-platform.c              | 125 +++++++++------------
 2 files changed, 84 insertions(+), 70 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/i2c/i2c-pca-platform.txt

-- 
2.13.0

[PATCH v3 1/5] dt-bindings: add bindings for i2c-pca-platform

[Chris Packham]

Signed-off-by: Chris Packham <chris.packham@alliedtelesis.co.nz>
---
Changes in v2:
- Split dt-binding to separate patch, use "reset-gpios" instead of "gpios".
Changes in v3:
- State the gpio is active low
- Add example

 .../devicetree/bindings/i2c/i2c-pca-platform.txt   | 29 ++++++++++++++++++++++
 1 file changed, 29 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/i2c/i2c-pca-platform.txt

diff --git a/Documentation/devicetree/bindings/i2c/i2c-pca-platform.txt b/Documentation/devicetree/bindings/i2c/i2c-pca-platform.txt
new file mode 100644
index 000000000000..f1f3876bb8e8
--- /dev/null
+++ b/Documentation/devicetree/bindings/i2c/i2c-pca-platform.txt
@@ -0,0 +1,29 @@
+* NXP PCA PCA9564/PCA9665 I2C controller
+
+The PCA9564/PCA9665 serves as an interface between most standard
+parallel-bus microcontrollers/microprocessors and the serial I2C-bus
+and allows the parallel bus system to communicate bi-directionally
+with the I2C-bus.
+
+Required properties :
+
+ - reg : Offset and length of the register set for the device
+ - compatible : one of "nxp,pca9564" or "nxp,pca9665"
+
+Optional properties
+ - interrupts : the interrupt number
+ - interrupt-parent : the phandle for the interrupt controller.
+   If an interrupt is not specified polling will be used.
+ - reset-gpios : gpio specifier for gpio connected to RESET_N pin. As the line
+   is active low, it should be marked GPIO_ACTIVE_LOW.
+ - clock-frequency : I2C bus frequency.
+
+Example:
+	i2c0: i2c@80000 {
+		compatible = "nxp,pca9564";
+		#address-cells = <1>;
+		#size-cells = <0>;
+		reg = <0x80000 0x4>;
+		reset-gpios = <&gpio1 0 GPIO_ACTIVE_LOW>;
+		clock-frequency = <100000>;
+	};
-- 
2.13.0

Re: [PATCH v3 1/5] dt-bindings: add bindings for i2c-pca-platform

[Rob Herring]

On Sun, Jun 25, 2017 at 7:44 PM, Chris Packham
<chris.packham@alliedtelesis.co.nz> wrote:
> Signed-off-by: Chris Packham <chris.packham@alliedtelesis.co.nz>
> ---
> Changes in v2:
> - Split dt-binding to separate patch, use "reset-gpios" instead of "gpios".
> Changes in v3:
> - State the gpio is active low
> - Add example
>
>  .../devicetree/bindings/i2c/i2c-pca-platform.txt   | 29 ++++++++++++++++++++++
>  1 file changed, 29 insertions(+)
>  create mode 100644 Documentation/devicetree/bindings/i2c/i2c-pca-platform.txt

Acked-by: Rob Herring <robh@kernel.org>

Re: [PATCH v3 1/5] dt-bindings: add bindings for i2c-pca-platform

[Rob Herring]

On Sun, Jun 25, 2017 at 7:44 PM, Chris Packham
<chris.packham-6g8wRflRTwXFdCa3tKVlE6U/zSkkHjvu@public.gmane.org> wrote:
> Signed-off-by: Chris Packham <chris.packham-6g8wRflRTwXFdCa3tKVlE6U/zSkkHjvu@public.gmane.org>
> ---
> Changes in v2:
> - Split dt-binding to separate patch, use "reset-gpios" instead of "gpios".
> Changes in v3:
> - State the gpio is active low
> - Add example
>
>  .../devicetree/bindings/i2c/i2c-pca-platform.txt   | 29 ++++++++++++++++++++++
>  1 file changed, 29 insertions(+)
>  create mode 100644 Documentation/devicetree/bindings/i2c/i2c-pca-platform.txt

Acked-by: Rob Herring <robh-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH v3 1/5] dt-bindings: add bindings for i2c-pca-platform

[Wolfram Sang]

[-- Attachment #1: Type: text/plain, Size: 161 bytes --]

On Mon, Jun 26, 2017 at 12:44:30PM +1200, Chris Packham wrote:
> Signed-off-by: Chris Packham <chris.packham@alliedtelesis.co.nz>

Applied to for-next, thanks!


[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

[PATCH v3 2/5] i2c: pca-platform: switch to struct gpio_desc

[Chris Packham]

Make use of struct gpio_desc which allows us to specify the active state
of the reset pin.

Signed-off-by: Chris Packham <chris.packham@alliedtelesis.co.nz>
---
Changes in v3:
- New, replaces "i2c: pca-platform: use gpio_is_valid"
- Using struct gpio_desc is required to make the flags stick. Because those
  flags now correctly indicate that RESET_N is active low, the logic for
  resetting the chip is inverted.

 drivers/i2c/busses/i2c-pca-platform.c | 43 ++++++++++++++++-------------------
 1 file changed, 19 insertions(+), 24 deletions(-)

diff --git a/drivers/i2c/busses/i2c-pca-platform.c b/drivers/i2c/busses/i2c-pca-platform.c
index 3bd2e7d06e4b..9f995b8ed587 100644
--- a/drivers/i2c/busses/i2c-pca-platform.c
+++ b/drivers/i2c/busses/i2c-pca-platform.c
@@ -22,6 +22,7 @@
 #include <linux/i2c-algo-pca.h>
 #include <linux/i2c-pca-platform.h>
 #include <linux/gpio.h>
+#include <linux/gpio/consumer.h>
 #include <linux/io.h>
 
 #include <asm/irq.h>
@@ -29,7 +30,7 @@
 struct i2c_pca_pf_data {
 	void __iomem			*reg_base;
 	int				irq;	/* if 0, use polling */
-	int				gpio;
+	struct gpio_desc		*gpio;
 	wait_queue_head_t		wait;
 	struct i2c_adapter		adap;
 	struct i2c_algo_pca_data	algo_data;
@@ -112,9 +113,9 @@ static void i2c_pca_pf_resetchip(void *pd)
 {
 	struct i2c_pca_pf_data *i2c = pd;
 
-	gpio_set_value(i2c->gpio, 0);
+	gpiod_set_value(i2c->gpio, 1);
 	ndelay(100);
-	gpio_set_value(i2c->gpio, 1);
+	gpiod_set_value(i2c->gpio, 0);
 }
 
 static irqreturn_t i2c_pca_pf_handler(int this_irq, void *dev_id)
@@ -181,11 +182,24 @@ static int i2c_pca_pf_probe(struct platform_device *pdev)
 	if (platform_data) {
 		i2c->adap.timeout = platform_data->timeout;
 		i2c->algo_data.i2c_clock = platform_data->i2c_clock_speed;
-		i2c->gpio = platform_data->gpio;
+		if (gpio_is_valid(platform_data->gpio)) {
+			ret = devm_gpio_request_one(&pdev->dev,
+						    platform_data->gpio,
+						    GPIOF_ACTIVE_LOW,
+						    i2c->adap.name);
+			if (ret == 0) {
+				i2c->gpio = gpio_to_desc(platform_data->gpio);
+				gpiod_direction_output(i2c->gpio, 0);
+				i2c->algo_data.reset_chip = i2c_pca_pf_resetchip;
+			} else {
+				dev_warn(&pdev->dev, "Registering gpio failed!\n");
+				i2c->gpio = NULL;
+			}
+		}
 	} else {
 		i2c->adap.timeout = HZ;
 		i2c->algo_data.i2c_clock = 59000;
-		i2c->gpio = -1;
+		i2c->gpio = NULL;
 	}
 
 	i2c->algo_data.data = i2c;
@@ -208,19 +222,6 @@ static int i2c_pca_pf_probe(struct platform_device *pdev)
 		break;
 	}
 
-	/* Use gpio_is_valid() when in mainline */
-	if (i2c->gpio > -1) {
-		ret = gpio_request(i2c->gpio, i2c->adap.name);
-		if (ret == 0) {
-			gpio_direction_output(i2c->gpio, 1);
-			i2c->algo_data.reset_chip = i2c_pca_pf_resetchip;
-		} else {
-			printk(KERN_WARNING "%s: Registering gpio failed!\n",
-				i2c->adap.name);
-			i2c->gpio = ret;
-		}
-	}
-
 	if (irq) {
 		ret = request_irq(irq, i2c_pca_pf_handler,
 			IRQF_TRIGGER_FALLING, pdev->name, i2c);
@@ -243,9 +244,6 @@ static int i2c_pca_pf_probe(struct platform_device *pdev)
 	if (irq)
 		free_irq(irq, i2c);
 e_reqirq:
-	if (i2c->gpio > -1)
-		gpio_free(i2c->gpio);
-
 	iounmap(i2c->reg_base);
 e_remap:
 	kfree(i2c);
@@ -265,9 +263,6 @@ static int i2c_pca_pf_remove(struct platform_device *pdev)
 	if (i2c->irq)
 		free_irq(i2c->irq, i2c);
 
-	if (i2c->gpio > -1)
-		gpio_free(i2c->gpio);
-
 	iounmap(i2c->reg_base);
 	release_mem_region(i2c->io_base, i2c->io_size);
 	kfree(i2c);
-- 
2.13.0

Re: [PATCH v3 2/5] i2c: pca-platform: switch to struct gpio_desc

[Wolfram Sang]

[-- Attachment #1: Type: text/plain, Size: 267 bytes --]

On Mon, Jun 26, 2017 at 12:44:31PM +1200, Chris Packham wrote:
> Make use of struct gpio_desc which allows us to specify the active state
> of the reset pin.
> 
> Signed-off-by: Chris Packham <chris.packham@alliedtelesis.co.nz>

Applied to for-next, thanks!


[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: [PATCH v3 2/5] i2c: pca-platform: switch to struct gpio_desc

[Andy Shevchenko]

On Mon, Jun 26, 2017 at 3:44 AM, Chris Packham
<chris.packham@alliedtelesis.co.nz> wrote:
> Make use of struct gpio_desc which allows us to specify the active state
> of the reset pin.

> +               if (gpio_is_valid(platform_data->gpio)) {

No way, it should be provided by GPIO lookup table.

> +                       ret = devm_gpio_request_one(&pdev->dev,
> +                                                   platform_data->gpio,
> +                                                   GPIOF_ACTIVE_LOW,
> +                                                   i2c->adap.name);
> +                       if (ret == 0) {
> +                               i2c->gpio = gpio_to_desc(platform_data->gpio);

This has to be devm_gpiod_get() instead.

Wolfram, I suggested to drop the series for now and wait for improvements.

-- 
With Best Regards,
Andy Shevchenko

Re: [PATCH v3 2/5] i2c: pca-platform: switch to struct gpio_desc

[Chris Packham]

On 28/06/17 21:25, Andy Shevchenko wrote:
> On Mon, Jun 26, 2017 at 3:44 AM, Chris Packham
> <chris.packham@alliedtelesis.co.nz> wrote:
>> Make use of struct gpio_desc which allows us to specify the active state
>> of the reset pin.
> 
>> +               if (gpio_is_valid(platform_data->gpio)) {
> 
> No way, it should be provided by GPIO lookup table.
> 

Can do. The only platform that would be affected is 
arch/sh/boards/board-sh7785lcr.c and the fact that it sets gpio = 0 
makes me think that this might be an error.

>> +                       ret = devm_gpio_request_one(&pdev->dev,
>> +                                                   platform_data->gpio,
>> +                                                   GPIOF_ACTIVE_LOW,
>> +                                                   i2c->adap.name);
>> +                       if (ret == 0) {
>> +                               i2c->gpio = gpio_to_desc(platform_data->gpio);
> 
> This has to be devm_gpiod_get() instead.

I can just make the devm_gpiod_get_optional() call I add in 3/5 
unconditional.

> 
> Wolfram, I suggested to drop the series for now and wait for improvements.
> 

Wolframs call. I don't mind either way. I can send a v4 or send a new 
series on top of i2c.git#for-next.

Re: [PATCH v3 2/5] i2c: pca-platform: switch to struct gpio_desc

[Wolfram Sang]

[-- Attachment #1: Type: text/plain, Size: 194 bytes --]


> Wolframs call. I don't mind either way. I can send a v4 or send a new 
> series on top of i2c.git#for-next.

I think we can work incrementally on what I pushed out already.

Thanks!


[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

[PATCH v3 3/5] i2c: pca-platform: add devicetree awareness

[Chris Packham]

Allow devices that use this driver to be registered via a
devicetree.

Signed-off-by: Chris Packham <chris.packham@alliedtelesis.co.nz>
---
Changes in v2:
- Set i2c->adap.dev.of_node so that child nodes are automatically probed
- Split dt-binding to separate patch, use "reset-gpios" instead of "gpios".
Changes in v3:
- None

 drivers/i2c/busses/i2c-pca-platform.c | 23 +++++++++++++++++++++++
 1 file changed, 23 insertions(+)

diff --git a/drivers/i2c/busses/i2c-pca-platform.c b/drivers/i2c/busses/i2c-pca-platform.c
index 9f995b8ed587..a6df6b8d2289 100644
--- a/drivers/i2c/busses/i2c-pca-platform.c
+++ b/drivers/i2c/busses/i2c-pca-platform.c
@@ -24,6 +24,8 @@
 #include <linux/gpio.h>
 #include <linux/gpio/consumer.h>
 #include <linux/io.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
 
 #include <asm/irq.h>
 
@@ -137,12 +139,15 @@ static int i2c_pca_pf_probe(struct platform_device *pdev)
 	struct resource *res;
 	struct i2c_pca9564_pf_platform_data *platform_data =
 				dev_get_platdata(&pdev->dev);
+	struct device_node *np = pdev->dev.of_node;
 	int ret = 0;
 	int irq;
 
 	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
 	irq = platform_get_irq(pdev, 0);
 	/* If irq is 0, we do polling. */
+	if (irq < 0)
+		irq = 0;
 
 	if (res == NULL) {
 		ret = -ENODEV;
@@ -178,6 +183,7 @@ static int i2c_pca_pf_probe(struct platform_device *pdev)
 		 (unsigned long) res->start);
 	i2c->adap.algo_data = &i2c->algo_data;
 	i2c->adap.dev.parent = &pdev->dev;
+	i2c->adap.dev.of_node = np;
 
 	if (platform_data) {
 		i2c->adap.timeout = platform_data->timeout;
@@ -196,6 +202,13 @@ static int i2c_pca_pf_probe(struct platform_device *pdev)
 				i2c->gpio = NULL;
 			}
 		}
+	} else if (np) {
+		i2c->adap.timeout = HZ;
+		i2c->gpio = devm_gpiod_get_optional(&pdev->dev, "reset-gpios", GPIOD_OUT_LOW);
+		if (IS_ERR(i2c->gpio))
+			return PTR_ERR(i2c->gpio);
+		of_property_read_u32_index(np, "clock-frequency", 0,
+					   &i2c->algo_data.i2c_clock);
 	} else {
 		i2c->adap.timeout = HZ;
 		i2c->algo_data.i2c_clock = 59000;
@@ -270,11 +283,21 @@ static int i2c_pca_pf_remove(struct platform_device *pdev)
 	return 0;
 }
 
+#ifdef CONFIG_OF
+static const struct of_device_id i2c_pca_of_match_table[] = {
+	{ .compatible = "nxp,pca9564" },
+	{ .compatible = "nxp,pca9665" },
+	{},
+};
+MODULE_DEVICE_TABLE(of, i2c_pca_of_match_table);
+#endif
+
 static struct platform_driver i2c_pca_pf_driver = {
 	.probe = i2c_pca_pf_probe,
 	.remove = i2c_pca_pf_remove,
 	.driver = {
 		.name = "i2c-pca-platform",
+		.of_match_table = of_match_ptr(i2c_pca_of_match_table),
 	},
 };
 
-- 
2.13.0

Re: [PATCH v3 3/5] i2c: pca-platform: add devicetree awareness

[Wolfram Sang]

[-- Attachment #1: Type: text/plain, Size: 673 bytes --]

On Mon, Jun 26, 2017 at 12:44:32PM +1200, Chris Packham wrote:
> Allow devices that use this driver to be registered via a
> devicetree.
> 
> Signed-off-by: Chris Packham <chris.packham@alliedtelesis.co.nz>

...

> +		if (IS_ERR(i2c->gpio))
> +			return PTR_ERR(i2c->gpio);

coccicheck reported:

drivers/i2c/busses/i2c-pca-platform.c:209:3-9: ERROR: missing iounmap; ioremap on line 170 and execution via conditional on line 208

I fixed it like this (incremental diff):

-		if (IS_ERR(i2c->gpio))
-			return PTR_ERR(i2c->gpio);
+		if (IS_ERR(i2c->gpio)) {
+			ret = PTR_ERR(i2c->gpio);
+			goto e_reqirq;
+		}

and applied to for-next, thanks!


[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: [PATCH v3 3/5] i2c: pca-platform: add devicetree awareness

[Andy Shevchenko]

On Mon, Jun 26, 2017 at 3:44 AM, Chris Packham
<chris.packham@alliedtelesis.co.nz> wrote:
> Allow devices that use this driver to be registered via a
> devicetree.

Device Tree

Wolfram was being a bit faster to apply it...
> +#include <linux/of.h>
> +#include <linux/of_device.h>

> +       struct device_node *np = pdev->dev.of_node;

>         /* If irq is 0, we do polling. */
> +       if (irq < 0)
> +               irq = 0;

This does not belong to what is written in commit message. Looks like
a separate fix.

> +       i2c->adap.dev.of_node = np;

> +       } else if (np) {

> +               i2c->adap.timeout = HZ;
> +               i2c->gpio = devm_gpiod_get_optional(&pdev->dev, "reset-gpios", GPIOD_OUT_LOW);
> +               if (IS_ERR(i2c->gpio))
> +                       return PTR_ERR(i2c->gpio);
> +               of_property_read_u32_index(np, "clock-frequency", 0,
> +                                          &i2c->algo_data.i2c_clock);

And what prevents you to use device_property_read_*() here and get rid
of OFstuff?

> +#ifdef CONFIG_OF

Better not to put it...

> +               .of_match_table = of_match_ptr(i2c_pca_of_match_table),

...and get rid of of_match_ptr().

-- 
With Best Regards,
Andy Shevchenko

Re: [PATCH v3 3/5] i2c: pca-platform: add devicetree awareness

[Wolfram Sang]

[-- Attachment #1: Type: text/plain, Size: 687 bytes --]


> >         /* If irq is 0, we do polling. */
> > +       if (irq < 0)
> > +               irq = 0;
> 
> This does not belong to what is written in commit message. Looks like
> a separate fix.

True, this needs explanation.

> > +               of_property_read_u32_index(np, "clock-frequency", 0,
> > +                                          &i2c->algo_data.i2c_clock);
> 
> And what prevents you to use device_property_read_*() here and get rid
> of OFstuff?

device_* functions are not super much known yet. Even I tend to forget
them occasionally like here. We can fix it incrementally, too. But I
agree, it would be nice to have it right from the beginning.


[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: [PATCH v3 3/5] i2c: pca-platform: add devicetree awareness

[Chris Packham]

On 28/06/17 21:20, Andy Shevchenko wrote:
>> +       i2c->adap.dev.of_node = np;
>> +       } else if (np) {
>> +               i2c->adap.timeout = HZ;
>> +               i2c->gpio = devm_gpiod_get_optional(&pdev->dev, "reset-gpios", GPIOD_OUT_LOW);
>> +               if (IS_ERR(i2c->gpio))
>> +                       return PTR_ERR(i2c->gpio);
>> +               of_property_read_u32_index(np, "clock-frequency", 0,
>> +                                          &i2c->algo_data.i2c_clock);
> And what prevents you to use device_property_read_*() here and get rid
> of OFstuff?
> 

Nothing particular. I think I just found more instances of parsing a 
"clock-frequency" this way. Happy to change this.

>> +#ifdef CONFIG_OF
> Better not to put it...
> 
>> +               .of_match_table = of_match_ptr(i2c_pca_of_match_table),
> ...and get rid of of_match_ptr().

So what's the current best practice with this? I gather the intent is to 
keep the kernel size down by only including the of_match tables on 
platforms that actually use a device tree. There are just shy of 1000 
instances of of_match_ptr in the current tree (a handful of which are in 
drivers/i2c). Have we now reached a point where there are more dt-aware 
platforms than unaware ones?

[PATCH v3 4/5] i2c: pca-platform: use device managed allocations

[Chris Packham]

Switch to using the devm_ APIs and remove the now unnecessary error
handling and most of the device removal code.

Signed-off-by: Chris Packham <chris.packham@alliedtelesis.co.nz>
---
 drivers/i2c/busses/i2c-pca-platform.c | 53 +++++++----------------------------
 1 file changed, 10 insertions(+), 43 deletions(-)

diff --git a/drivers/i2c/busses/i2c-pca-platform.c b/drivers/i2c/busses/i2c-pca-platform.c
index a6df6b8d2289..7db481cbf402 100644
--- a/drivers/i2c/busses/i2c-pca-platform.c
+++ b/drivers/i2c/busses/i2c-pca-platform.c
@@ -143,35 +143,23 @@ static int i2c_pca_pf_probe(struct platform_device *pdev)
 	int ret = 0;
 	int irq;
 
-	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
 	irq = platform_get_irq(pdev, 0);
 	/* If irq is 0, we do polling. */
 	if (irq < 0)
 		irq = 0;
 
-	if (res == NULL) {
-		ret = -ENODEV;
-		goto e_print;
-	}
+	i2c = devm_kzalloc(&pdev->dev, sizeof(*i2c), GFP_KERNEL);
+	if (!i2c)
+		return -ENOMEM;
 
-	if (!request_mem_region(res->start, resource_size(res), res->name)) {
-		ret = -ENOMEM;
-		goto e_print;
-	}
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	i2c->reg_base = devm_ioremap_resource(&pdev->dev, res);
+	if (IS_ERR(i2c->reg_base))
+		return PTR_ERR(i2c->reg_base);
 
-	i2c = kzalloc(sizeof(struct i2c_pca_pf_data), GFP_KERNEL);
-	if (!i2c) {
-		ret = -ENOMEM;
-		goto e_alloc;
-	}
 
 	init_waitqueue_head(&i2c->wait);
 
-	i2c->reg_base = ioremap(res->start, resource_size(res));
-	if (!i2c->reg_base) {
-		ret = -ENOMEM;
-		goto e_remap;
-	}
 	i2c->io_base = res->start;
 	i2c->io_size = resource_size(res);
 	i2c->irq = irq;
@@ -236,15 +224,14 @@ static int i2c_pca_pf_probe(struct platform_device *pdev)
 	}
 
 	if (irq) {
-		ret = request_irq(irq, i2c_pca_pf_handler,
+		ret = devm_request_irq(&pdev->dev, irq, i2c_pca_pf_handler,
 			IRQF_TRIGGER_FALLING, pdev->name, i2c);
 		if (ret)
-			goto e_reqirq;
+			return ret;
 	}
 
 	if (i2c_pca_add_numbered_bus(&i2c->adap) < 0) {
-		ret = -ENODEV;
-		goto e_adapt;
+		return -ENODEV;
 	}
 
 	platform_set_drvdata(pdev, i2c);
@@ -252,19 +239,6 @@ static int i2c_pca_pf_probe(struct platform_device *pdev)
 	printk(KERN_INFO "%s registered.\n", i2c->adap.name);
 
 	return 0;
-
-e_adapt:
-	if (irq)
-		free_irq(irq, i2c);
-e_reqirq:
-	iounmap(i2c->reg_base);
-e_remap:
-	kfree(i2c);
-e_alloc:
-	release_mem_region(res->start, resource_size(res));
-e_print:
-	printk(KERN_ERR "Registering PCA9564/PCA9665 FAILED! (%d)\n", ret);
-	return ret;
 }
 
 static int i2c_pca_pf_remove(struct platform_device *pdev)
@@ -273,13 +247,6 @@ static int i2c_pca_pf_remove(struct platform_device *pdev)
 
 	i2c_del_adapter(&i2c->adap);
 
-	if (i2c->irq)
-		free_irq(i2c->irq, i2c);
-
-	iounmap(i2c->reg_base);
-	release_mem_region(i2c->io_base, i2c->io_size);
-	kfree(i2c);
-
 	return 0;
 }
 
-- 
2.13.0

Re: [PATCH v3 4/5] i2c: pca-platform: use device managed allocations

[Wolfram Sang]

[-- Attachment #1: Type: text/plain, Size: 577 bytes --]

On Mon, Jun 26, 2017 at 12:44:33PM +1200, Chris Packham wrote:
> Switch to using the devm_ APIs and remove the now unnecessary error
> handling and most of the device removal code.
> 
> Signed-off-by: Chris Packham <chris.packham@alliedtelesis.co.nz>

And "reverted" the previous fix here:

-               if (IS_ERR(i2c->gpio)) {
-                       ret = PTR_ERR(i2c->gpio);
-                       goto e_reqirq;
-               }
+               if (IS_ERR(i2c->gpio))
+                       return PTR_ERR(i2c->gpio);

Applied to for-next, thanks!


[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: [PATCH v3 4/5] i2c: pca-platform: use device managed allocations

[Andy Shevchenko]

On Mon, Jun 26, 2017 at 3:44 AM, Chris Packham
<chris.packham@alliedtelesis.co.nz> wrote:
> Switch to using the devm_ APIs and remove the now unnecessary error
> handling and most of the device removal code.


>         if (i2c_pca_add_numbered_bus(&i2c->adap) < 0) {
> -               ret = -ENODEV;
> -               goto e_adapt;
> +               return -ENODEV;

This is still shadows the actual error code.

-- 
With Best Regards,
Andy Shevchenko

Re: [PATCH v3 4/5] i2c: pca-platform: use device managed allocations

[Wolfram Sang]

[-- Attachment #1: Type: text/plain, Size: 317 bytes --]


> >         if (i2c_pca_add_numbered_bus(&i2c->adap) < 0) {
> > -               ret = -ENODEV;
> > -               goto e_adapt;
> > +               return -ENODEV;
> 
> This is still shadows the actual error code.

Nice catch. But since it fixes a seperate issue, I'd prefer an
incremental change here.


[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: [PATCH v3 4/5] i2c: pca-platform: use device managed allocations

[Andy Shevchenko]

On Wed, Jun 28, 2017 at 12:38 PM, Wolfram Sang <wsa@the-dreams.de> wrote:
>
>> >         if (i2c_pca_add_numbered_bus(&i2c->adap) < 0) {
>> > -               ret = -ENODEV;
>> > -               goto e_adapt;
>> > +               return -ENODEV;
>>
>> This is still shadows the actual error code.
>
> Nice catch. But since it fixes a seperate issue, I'd prefer an
> incremental change here.

Agreed.

-- 
With Best Regards,
Andy Shevchenko

Re: [PATCH v3 4/5] i2c: pca-platform: use device managed allocations

[Chris Packham]

On 28/06/17 22:54, Andy Shevchenko wrote:
> On Wed, Jun 28, 2017 at 12:38 PM, Wolfram Sang <wsa@the-dreams.de> wrote:
>>
>>>>          if (i2c_pca_add_numbered_bus(&i2c->adap) < 0) {
>>>> -               ret = -ENODEV;
>>>> -               goto e_adapt;
>>>> +               return -ENODEV;
>>>
>>> This is still shadows the actual error code.
>>
>> Nice catch. But since it fixes a seperate issue, I'd prefer an
>> incremental change here.
> 
> Agreed.
> 

OK I can send another patch for that.

[PATCH v3 5/5] i2c: pca-platform: use dev_warn/dev_info instead of printk

[Chris Packham]

Signed-off-by: Chris Packham <chris.packham@alliedtelesis.co.nz>
---
 drivers/i2c/busses/i2c-pca-platform.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/i2c/busses/i2c-pca-platform.c b/drivers/i2c/busses/i2c-pca-platform.c
index 7db481cbf402..395eca0cbb1f 100644
--- a/drivers/i2c/busses/i2c-pca-platform.c
+++ b/drivers/i2c/busses/i2c-pca-platform.c
@@ -107,8 +107,8 @@ static int i2c_pca_pf_waitforcompletion(void *pd)
 static void i2c_pca_pf_dummyreset(void *pd)
 {
 	struct i2c_pca_pf_data *i2c = pd;
-	printk(KERN_WARNING "%s: No reset-pin found. Chip may get stuck!\n",
-		i2c->adap.name);
+
+	dev_warn(&i2c->adap.dev, "No reset-pin found. Chip may get stuck!\n");
 }
 
 static void i2c_pca_pf_resetchip(void *pd)
@@ -236,7 +236,7 @@ static int i2c_pca_pf_probe(struct platform_device *pdev)
 
 	platform_set_drvdata(pdev, i2c);
 
-	printk(KERN_INFO "%s registered.\n", i2c->adap.name);
+	dev_info(&pdev->dev, "registered.\n");
 
 	return 0;
 }
-- 
2.13.0

Re: [PATCH v3 5/5] i2c: pca-platform: use dev_warn/dev_info instead of printk

[Wolfram Sang]

[-- Attachment #1: Type: text/plain, Size: 161 bytes --]

On Mon, Jun 26, 2017 at 12:44:34PM +1200, Chris Packham wrote:
> Signed-off-by: Chris Packham <chris.packham@alliedtelesis.co.nz>

Applied to for-next, thanks!


[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: [PATCH v3 5/5] i2c: pca-platform: use dev_warn/dev_info instead of printk

[Andy Shevchenko]

On Mon, Jun 26, 2017 at 3:44 AM, Chris Packham
<chris.packham@alliedtelesis.co.nz> wrote:

On the first glance I have thought this is _the_ only patch that
doesn't make my eyes hurt.

Not a chance! Where is the commit message?

Code below though looks fine.

> @@ -107,8 +107,8 @@ static int i2c_pca_pf_waitforcompletion(void *pd)
>  static void i2c_pca_pf_dummyreset(void *pd)
>  {
>         struct i2c_pca_pf_data *i2c = pd;
> -       printk(KERN_WARNING "%s: No reset-pin found. Chip may get stuck!\n",
> -               i2c->adap.name);
> +
> +       dev_warn(&i2c->adap.dev, "No reset-pin found. Chip may get stuck!\n");
>  }
>
>  static void i2c_pca_pf_resetchip(void *pd)
> @@ -236,7 +236,7 @@ static int i2c_pca_pf_probe(struct platform_device *pdev)
>
>         platform_set_drvdata(pdev, i2c);
>
> -       printk(KERN_INFO "%s registered.\n", i2c->adap.name);
> +       dev_info(&pdev->dev, "registered.\n");
>
>         return 0;
>  }
> --
> 2.13.0
>



-- 
With Best Regards,
Andy Shevchenko

Re: [PATCH v3 5/5] i2c: pca-platform: use dev_warn/dev_info instead of printk

[Wolfram Sang]

[-- Attachment #1: Type: text/plain, Size: 90 bytes --]


> Not a chance! Where is the commit message?

Don't you think $subject is clear enough?


[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: [PATCH v3 5/5] i2c: pca-platform: use dev_warn/dev_info instead of printk

[Andy Shevchenko]

On Wed, Jun 28, 2017 at 12:37 PM, Wolfram Sang <wsa@the-dreams.de> wrote:
>
>> Not a chance! Where is the commit message?
>
> Don't you think $subject is clear enough?

I would add the comment at least on how the output is changed since
pr_*("%s:"..., i2c->adap.name,...)  to dev_*(&pdev->dev, ...)
replacement.

-- 
With Best Regards,
Andy Shevchenko

Re: [PATCH v3 5/5] i2c: pca-platform: use dev_warn/dev_info instead of printk

[Wolfram Sang]

[-- Attachment #1: Type: text/plain, Size: 442 bytes --]

On Wed, Jun 28, 2017 at 01:57:01PM +0300, Andy Shevchenko wrote:
> On Wed, Jun 28, 2017 at 12:37 PM, Wolfram Sang <wsa@the-dreams.de> wrote:
> >
> >> Not a chance! Where is the commit message?
> >
> > Don't you think $subject is clear enough?
> 
> I would add the comment at least on how the output is changed since
> pr_*("%s:"..., i2c->adap.name,...)  to dev_*(&pdev->dev, ...)
> replacement.

Yes, that could have been done.


[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]