[PATCH v2 0/3] ieee802154: ca8210: Sparse fix and GPIOd conversion

[PATCH v2 0/3] ieee802154: ca8210: Sparse fix and GPIOd conversion

[Andy Shevchenko]

The main part is the patch 3 that converts the driver to GPIO descriptor APIs,
the first one is just an ad-hoc fix WRT sparse complains on the bitwise
types misuse. The second one is a small cleanup that helps patch 3 to be nicer.

In v2:
- split and extended cleanup pieces into patch 2 (Miquel)
- updated kernel doc for changed members (Miquel)
- unfolded PTR_ERR_OR_ZERO() to the preferred pattern (Miquel)
- collected tags (Miquel)

Andy Shevchenko (3):
  ieee802154: ca8210: Use proper setter and getters for bitwise types
  ieee802154: ca8210: Get platform data via dev_get_platdata()
  ieee802154: ca8210: Switch to using gpiod API

 drivers/net/ieee802154/ca8210.c | 78 +++++++++++++--------------------
 1 file changed, 31 insertions(+), 47 deletions(-)

-- 
2.47.2

[PATCH net-next v2 1/3] ieee802154: ca8210: Use proper setter and getters for bitwise types

[Andy Shevchenko]

Sparse complains that the driver doesn't respect the bitwise types:

drivers/net/ieee802154/ca8210.c:1796:27: warning: incorrect type in assignment (different base types)
drivers/net/ieee802154/ca8210.c:1796:27:    expected restricted __le16 [addressable] [assigned] [usertype] pan_id
drivers/net/ieee802154/ca8210.c:1796:27:    got unsigned short [usertype]
drivers/net/ieee802154/ca8210.c:1801:25: warning: incorrect type in assignment (different base types)
drivers/net/ieee802154/ca8210.c:1801:25:    expected restricted __le16 [addressable] [assigned] [usertype] pan_id
drivers/net/ieee802154/ca8210.c:1801:25:    got unsigned short [usertype]
drivers/net/ieee802154/ca8210.c:1928:28: warning: incorrect type in argument 3 (different base types)
drivers/net/ieee802154/ca8210.c:1928:28:    expected unsigned short [usertype] dst_pan_id
drivers/net/ieee802154/ca8210.c:1928:28:    got restricted __le16 [addressable] [usertype] pan_id

Use proper setter and getters for bitwise types.

Note, in accordance with [1] the protocol is little endian.

Link: https://www.cascoda.com/wp-content/uploads/2018/11/CA-8210_datasheet_0418.pdf [1]
Reviewed-by: Miquel Raynal <miquel.raynal@bootlin.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
---
 drivers/net/ieee802154/ca8210.c | 9 ++++-----
 1 file changed, 4 insertions(+), 5 deletions(-)

diff --git a/drivers/net/ieee802154/ca8210.c b/drivers/net/ieee802154/ca8210.c
index 753215ebc67c..a036910f6082 100644
--- a/drivers/net/ieee802154/ca8210.c
+++ b/drivers/net/ieee802154/ca8210.c
@@ -1446,8 +1446,7 @@ static u8 mcps_data_request(
 	command.pdata.data_req.src_addr_mode = src_addr_mode;
 	command.pdata.data_req.dst.mode = dst_address_mode;
 	if (dst_address_mode != MAC_MODE_NO_ADDR) {
-		command.pdata.data_req.dst.pan_id[0] = LS_BYTE(dst_pan_id);
-		command.pdata.data_req.dst.pan_id[1] = MS_BYTE(dst_pan_id);
+		put_unaligned_le16(dst_pan_id, command.pdata.data_req.dst.pan_id);
 		if (dst_address_mode == MAC_MODE_SHORT_ADDR) {
 			command.pdata.data_req.dst.address[0] = LS_BYTE(
 				dst_addr->short_address
@@ -1795,12 +1794,12 @@ static int ca8210_skb_rx(
 	}
 	hdr.source.mode = data_ind[0];
 	dev_dbg(&priv->spi->dev, "srcAddrMode: %#03x\n", hdr.source.mode);
-	hdr.source.pan_id = *(u16 *)&data_ind[1];
+	hdr.source.pan_id = cpu_to_le16(get_unaligned_le16(&data_ind[1]));
 	dev_dbg(&priv->spi->dev, "srcPanId: %#06x\n", hdr.source.pan_id);
 	memcpy(&hdr.source.extended_addr, &data_ind[3], 8);
 	hdr.dest.mode = data_ind[11];
 	dev_dbg(&priv->spi->dev, "dstAddrMode: %#03x\n", hdr.dest.mode);
-	hdr.dest.pan_id = *(u16 *)&data_ind[12];
+	hdr.dest.pan_id = cpu_to_le16(get_unaligned_le16(&data_ind[12]));
 	dev_dbg(&priv->spi->dev, "dstPanId: %#06x\n", hdr.dest.pan_id);
 	memcpy(&hdr.dest.extended_addr, &data_ind[14], 8);
 
@@ -1927,7 +1926,7 @@ static int ca8210_skb_tx(
 	status =  mcps_data_request(
 		header.source.mode,
 		header.dest.mode,
-		header.dest.pan_id,
+		le16_to_cpu(header.dest.pan_id),
 		(union macaddr *)&header.dest.extended_addr,
 		skb->len - mac_len,
 		&skb->data[mac_len],
-- 
2.47.2

Re: [PATCH net-next v2 1/3] ieee802154: ca8210: Use proper setter and getters for bitwise types

[Linus Walleij]

On Mon, Mar 3, 2025 at 5:49 PM Andy Shevchenko
<andriy.shevchenko@linux.intel.com> wrote:

> Sparse complains that the driver doesn't respect the bitwise types:
>
> drivers/net/ieee802154/ca8210.c:1796:27: warning: incorrect type in assignment (different base types)
> drivers/net/ieee802154/ca8210.c:1796:27:    expected restricted __le16 [addressable] [assigned] [usertype] pan_id
> drivers/net/ieee802154/ca8210.c:1796:27:    got unsigned short [usertype]
> drivers/net/ieee802154/ca8210.c:1801:25: warning: incorrect type in assignment (different base types)
> drivers/net/ieee802154/ca8210.c:1801:25:    expected restricted __le16 [addressable] [assigned] [usertype] pan_id
> drivers/net/ieee802154/ca8210.c:1801:25:    got unsigned short [usertype]
> drivers/net/ieee802154/ca8210.c:1928:28: warning: incorrect type in argument 3 (different base types)
> drivers/net/ieee802154/ca8210.c:1928:28:    expected unsigned short [usertype] dst_pan_id
> drivers/net/ieee802154/ca8210.c:1928:28:    got restricted __le16 [addressable] [usertype] pan_id
>
> Use proper setter and getters for bitwise types.
>
> Note, in accordance with [1] the protocol is little endian.
>
> Link: https://www.cascoda.com/wp-content/uploads/2018/11/CA-8210_datasheet_0418.pdf [1]
> Reviewed-by: Miquel Raynal <miquel.raynal@bootlin.com>
> Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

Reviewed-by: Linus Walleij <linus.walleij@linaro.org>

Yours,
Linus Walleij

[PATCH net-next v2 2/3] ieee802154: ca8210: Get platform data via dev_get_platdata()

[Andy Shevchenko]

Access to platform data via dev_get_platdata() getter to make code cleaner.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
---
 drivers/net/ieee802154/ca8210.c | 12 ++++++++----
 1 file changed, 8 insertions(+), 4 deletions(-)

diff --git a/drivers/net/ieee802154/ca8210.c b/drivers/net/ieee802154/ca8210.c
index a036910f6082..65f042c8734b 100644
--- a/drivers/net/ieee802154/ca8210.c
+++ b/drivers/net/ieee802154/ca8210.c
@@ -627,7 +627,8 @@ static int ca8210_spi_transfer(
  */
 static void ca8210_reset_send(struct spi_device *spi, unsigned int ms)
 {
-	struct ca8210_platform_data *pdata = spi->dev.platform_data;
+	struct device *dev = &spi->dev;
+	struct ca8210_platform_data *pdata = dev_get_platdata(dev);
 	struct ca8210_priv *priv = spi_get_drvdata(spi);
 	long status;
 
@@ -2736,9 +2737,10 @@ static int ca8210_config_extern_clk(
  */
 static int ca8210_register_ext_clock(struct spi_device *spi)
 {
+	struct device *dev = &spi->dev;
+	struct ca8210_platform_data *pdata = dev_get_platdata(dev);
 	struct device_node *np = spi->dev.of_node;
 	struct ca8210_priv *priv = spi_get_drvdata(spi);
-	struct ca8210_platform_data *pdata = spi->dev.platform_data;
 
 	if (!np)
 		return -EFAULT;
@@ -2784,8 +2786,9 @@ static void ca8210_unregister_ext_clock(struct spi_device *spi)
  */
 static int ca8210_reset_init(struct spi_device *spi)
 {
+	struct device *dev = &spi->dev;
+	struct ca8210_platform_data *pdata = dev_get_platdata(dev);
 	int ret;
-	struct ca8210_platform_data *pdata = spi->dev.platform_data;
 
 	pdata->gpio_reset = of_get_named_gpio(
 		spi->dev.of_node,
@@ -2813,8 +2816,9 @@ static int ca8210_reset_init(struct spi_device *spi)
  */
 static int ca8210_interrupt_init(struct spi_device *spi)
 {
+	struct device *dev = &spi->dev;
+	struct ca8210_platform_data *pdata = dev_get_platdata(dev);
 	int ret;
-	struct ca8210_platform_data *pdata = spi->dev.platform_data;
 
 	pdata->gpio_irq = of_get_named_gpio(
 		spi->dev.of_node,
-- 
2.47.2

Re: [PATCH net-next v2 2/3] ieee802154: ca8210: Get platform data via dev_get_platdata()

[Linus Walleij]

On Mon, Mar 3, 2025 at 5:49 PM Andy Shevchenko
<andriy.shevchenko@linux.intel.com> wrote:

> Access to platform data via dev_get_platdata() getter to make code cleaner.
>
> Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

Reviewed-by: Linus Walleij <linus.walleij@linaro.org>

Yours,
Linus Walleij

Re: [PATCH net-next v2 2/3] ieee802154: ca8210: Get platform data via dev_get_platdata()

[Miquel Raynal]

On 03/03/2025 at 18:49:09 +02, Andy Shevchenko <andriy.shevchenko@linux.intel.com> wrote:

> Access to platform data via dev_get_platdata() getter to make code cleaner.

Reviewed-by: Miquel Raynal <miquel.raynal@bootlin.com>

Thanks,
Miquèl

[PATCH net-next v2 3/3] ieee802154: ca8210: Switch to using gpiod API

[Andy Shevchenko]

This updates the driver to gpiod API, and removes yet another use of
of_get_named_gpio().

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
---
 drivers/net/ieee802154/ca8210.c | 57 +++++++++++----------------------
 1 file changed, 19 insertions(+), 38 deletions(-)

diff --git a/drivers/net/ieee802154/ca8210.c b/drivers/net/ieee802154/ca8210.c
index 65f042c8734b..d4c1f00d29fe 100644
--- a/drivers/net/ieee802154/ca8210.c
+++ b/drivers/net/ieee802154/ca8210.c
@@ -52,12 +52,10 @@
 #include <linux/debugfs.h>
 #include <linux/delay.h>
 #include <linux/gpio/consumer.h>
-#include <linux/gpio.h>
 #include <linux/ieee802154.h>
 #include <linux/io.h>
 #include <linux/kfifo.h>
 #include <linux/of.h>
-#include <linux/of_gpio.h>
 #include <linux/module.h>
 #include <linux/mutex.h>
 #include <linux/poll.h>
@@ -350,8 +348,8 @@ struct work_priv_container {
  * @extclockenable: true if the external clock is to be enabled
  * @extclockfreq:   frequency of the external clock
  * @extclockgpio:   ca8210 output gpio of the external clock
- * @gpio_reset:     gpio number of ca8210 reset line
- * @gpio_irq:       gpio number of ca8210 interrupt line
+ * @reset_gpio:     ca8210 reset GPIO descriptor
+ * @irq_gpio:       ca8210 interrupt GPIO descriptor
  * @irq_id:         identifier for the ca8210 irq
  *
  */
@@ -359,8 +357,8 @@ struct ca8210_platform_data {
 	bool extclockenable;
 	unsigned int extclockfreq;
 	unsigned int extclockgpio;
-	int gpio_reset;
-	int gpio_irq;
+	struct gpio_desc *reset_gpio;
+	struct gpio_desc *irq_gpio;
 	int irq_id;
 };
 
@@ -632,10 +630,10 @@ static void ca8210_reset_send(struct spi_device *spi, unsigned int ms)
 	struct ca8210_priv *priv = spi_get_drvdata(spi);
 	long status;
 
-	gpio_set_value(pdata->gpio_reset, 0);
+	gpiod_set_value(pdata->reset_gpio, 0);
 	reinit_completion(&priv->ca8210_is_awake);
 	msleep(ms);
-	gpio_set_value(pdata->gpio_reset, 1);
+	gpiod_set_value(pdata->reset_gpio, 1);
 	priv->promiscuous = false;
 
 	/* Wait until wakeup indication seen */
@@ -2788,24 +2786,14 @@ static int ca8210_reset_init(struct spi_device *spi)
 {
 	struct device *dev = &spi->dev;
 	struct ca8210_platform_data *pdata = dev_get_platdata(dev);
-	int ret;
 
-	pdata->gpio_reset = of_get_named_gpio(
-		spi->dev.of_node,
-		"reset-gpio",
-		0
-	);
-
-	ret = gpio_direction_output(pdata->gpio_reset, 1);
-	if (ret < 0) {
-		dev_crit(
-			&spi->dev,
-			"Reset GPIO %d did not set to output mode\n",
-			pdata->gpio_reset
-		);
+	pdata->reset_gpio = devm_gpiod_get(dev, "reset", GPIOD_OUT_HIGH);
+	if (IS_ERR(pdata->reset_gpio)) {
+		dev_crit(dev, "Reset GPIO did not set to output mode\n");
+		return PTR_ERR(pdata->reset_gpio);
 	}
 
-	return ret;
+	return 0;
 }
 
 /**
@@ -2820,20 +2808,15 @@ static int ca8210_interrupt_init(struct spi_device *spi)
 	struct ca8210_platform_data *pdata = dev_get_platdata(dev);
 	int ret;
 
-	pdata->gpio_irq = of_get_named_gpio(
-		spi->dev.of_node,
-		"irq-gpio",
-		0
-	);
+	pdata->irq_gpio = devm_gpiod_get(dev, "irq", GPIOD_IN);
+	if (IS_ERR(pdata->irq_gpio)) {
+		dev_crit(dev, "Could not retrieve IRQ GPIO\n");
+		return PTR_ERR(pdata->irq_gpio);
+	}
 
-	pdata->irq_id = gpio_to_irq(pdata->gpio_irq);
+	pdata->irq_id = gpiod_to_irq(pdata->irq_gpio);
 	if (pdata->irq_id < 0) {
-		dev_crit(
-			&spi->dev,
-			"Could not get irq for gpio pin %d\n",
-			pdata->gpio_irq
-		);
-		gpio_free(pdata->gpio_irq);
+		dev_crit(dev, "Could not get irq for IRQ GPIO\n");
 		return pdata->irq_id;
 	}
 
@@ -2844,10 +2827,8 @@ static int ca8210_interrupt_init(struct spi_device *spi)
 		"ca8210-irq",
 		spi_get_drvdata(spi)
 	);
-	if (ret) {
+	if (ret)
 		dev_crit(&spi->dev, "request_irq %d failed\n", pdata->irq_id);
-		gpio_free(pdata->gpio_irq);
-	}
 
 	return ret;
 }
-- 
2.47.2

Re: [PATCH net-next v2 3/3] ieee802154: ca8210: Switch to using gpiod API

[Linus Walleij]

On Mon, Mar 3, 2025 at 5:49 PM Andy Shevchenko
<andriy.shevchenko@linux.intel.com> wrote:

> This updates the driver to gpiod API, and removes yet another use of
> of_get_named_gpio().
>
> Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

Reviewed-by: Linus Walleij <linus.walleij@linaro.org>

But note:

> @@ -632,10 +630,10 @@ static void ca8210_reset_send(struct spi_device *spi, unsigned int ms)
>         struct ca8210_priv *priv = spi_get_drvdata(spi);
>         long status;
>
> -       gpio_set_value(pdata->gpio_reset, 0);
> +       gpiod_set_value(pdata->reset_gpio, 0);
>         reinit_completion(&priv->ca8210_is_awake);
>         msleep(ms);
> -       gpio_set_value(pdata->gpio_reset, 1);
> +       gpiod_set_value(pdata->reset_gpio, 1);

This drives the GPIO low to assert reset, meaning it is something
that should have GPIO_ACTIVE_LOW set in the device tree,
and it might even have, so let's check what we can check:

git grep cascoda,ca8210
Documentation/devicetree/bindings/net/ieee802154/ca8210.txt:    -
compatible:           Should be "cascoda,ca8210"
Documentation/devicetree/bindings/net/ieee802154/ca8210.txt:
 compatible = "cascoda,ca8210";
drivers/net/ieee802154/ca8210.c:        {.compatible = "cascoda,ca8210", },

well ain't that typical, all users are out of tree. The example
in the bindings file is wrong, setting ACTIVE_HIGH. Sigh, let's
assume all those DTS files somewhere are wrong and they
didn't set GPIO_ACTIVE_LOW in them...

Maybe add a comment in the code that this is wrong and the
driver and DTS files should be fixed.

Alternatively patch Documentation/devicetree/bindings/net/ieee802154/ca8210.txt
to set GPIO_ACTIVE_LOW and fix the code to invert it both
here and when getting the GPIO, but it could cause problems
for outoftree users.

Either way, this is good progress:
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>

Yours,
Linus Walleij

Re: [PATCH net-next v2 3/3] ieee802154: ca8210: Switch to using gpiod API

[Andy Shevchenko]

Mon, Mar 03, 2025 at 09:00:39PM +0100, Linus Walleij kirjoitti:
> On Mon, Mar 3, 2025 at 5:49 PM Andy Shevchenko
> <andriy.shevchenko@linux.intel.com> wrote:

...

> >         reinit_completion(&priv->ca8210_is_awake);
> >         msleep(ms);
> > -       gpio_set_value(pdata->gpio_reset, 1);
> > +       gpiod_set_value(pdata->reset_gpio, 1);
> 
> This drives the GPIO low to assert reset, meaning it is something
> that should have GPIO_ACTIVE_LOW set in the device tree,

Yeah, the pin 27 is marked as NRESET and description is pointing out to it as
active low.

> and it might even have, so let's check what we can check:
> 
> git grep cascoda,ca8210
> Documentation/devicetree/bindings/net/ieee802154/ca8210.txt:    -
> compatible:           Should be "cascoda,ca8210"
> Documentation/devicetree/bindings/net/ieee802154/ca8210.txt:
>  compatible = "cascoda,ca8210";
> drivers/net/ieee802154/ca8210.c:        {.compatible = "cascoda,ca8210", },
> 
> well ain't that typical, all users are out of tree. The example
> in the bindings file is wrong, setting ACTIVE_HIGH. Sigh, let's
> assume all those DTS files somewhere are wrong and they
> didn't set GPIO_ACTIVE_LOW in them...

> Maybe add a comment in the code that this is wrong and the
> driver and DTS files should be fixed.

Or maybe fix in the driver and schema and add a quirk to gpiolib-of.c?

> Alternatively patch Documentation/devicetree/bindings/net/ieee802154/ca8210.txt
> to set GPIO_ACTIVE_LOW and fix the code to invert it both
> here and when getting the GPIO, but it could cause problems
> for outoftree users.

Would it? We have such quirks to fix a polarity for other drivers/devices.

> Either way, this is good progress:
> Reviewed-by: Linus Walleij <linus.walleij@linaro.org>

Thank you!

-- 
With Best Regards,
Andy Shevchenko

Re: [PATCH net-next v2 3/3] ieee802154: ca8210: Switch to using gpiod API

[Linus Walleij]

On Mon, Mar 3, 2025 at 9:39 PM Andy Shevchenko
<andy.shevchenko@gmail.com> wrote:

> > Maybe add a comment in the code that this is wrong and the
> > driver and DTS files should be fixed.
>
> Or maybe fix in the driver and schema and add a quirk to gpiolib-of.c?

Even better!

Yours,
Linus Walleij

Re: [PATCH net-next v2 3/3] ieee802154: ca8210: Switch to using gpiod API

[Andy Shevchenko]

On Tue, Mar 04, 2025 at 01:03:41AM +0100, Linus Walleij wrote:
> On Mon, Mar 3, 2025 at 9:39 PM Andy Shevchenko
> <andy.shevchenko@gmail.com> wrote:
> 
> > > Maybe add a comment in the code that this is wrong and the
> > > driver and DTS files should be fixed.
> >
> > Or maybe fix in the driver and schema and add a quirk to gpiolib-of.c?
> 
> Even better!

I am about to send a v3, I'm going to leave your tag despite a few changes as
discussed. I hope this is okay with you.

-- 
With Best Regards,
Andy Shevchenko

Re: [PATCH net-next v2 3/3] ieee802154: ca8210: Switch to using gpiod API

[Linus Walleij]

On Tue, Mar 4, 2025 at 12:13 PM Andy Shevchenko
<andy.shevchenko@gmail.com> wrote:
> On Tue, Mar 04, 2025 at 01:03:41AM +0100, Linus Walleij wrote:
> > On Mon, Mar 3, 2025 at 9:39 PM Andy Shevchenko
> > <andy.shevchenko@gmail.com> wrote:
> >
> > > > Maybe add a comment in the code that this is wrong and the
> > > > driver and DTS files should be fixed.
> > >
> > > Or maybe fix in the driver and schema and add a quirk to gpiolib-of.c?
> >
> > Even better!
>
> I am about to send a v3, I'm going to leave your tag despite a few changes as
> discussed. I hope this is okay with you.

Of course!

Linus Walleij