Re: [PATCH 1/2] hwmon: shtc1: add support for device tree bindings

From: Chris Ruehl
Date: Sat Jul 04 2020 - 20:42:22 EST


Hi Guenter,

On 3/7/2020 1:48 pm, Guenter Roeck wrote:
On 7/2/20 8:48 PM, Chris Ruehl wrote:
Add support for DTS bindings to the shtc driver, use CONFIG_OF
to compile in the code if needed.


Ah, here it is. The introducing patch should say something like "[PATCH 0/2]".

Signed-off-by: Chris Ruehl <chris.ruehl@xxxxxxxxxxxx>
---
drivers/hwmon/shtc1.c | 30 ++++++++++++++++++++++++++++++
1 file changed, 30 insertions(+)

diff --git a/drivers/hwmon/shtc1.c b/drivers/hwmon/shtc1.c
index a0078ccede03..3bcabc1cbce8 100644
--- a/drivers/hwmon/shtc1.c
+++ b/drivers/hwmon/shtc1.c
@@ -14,6 +14,9 @@
#include <linux/err.h>
#include <linux/delay.h>
#include <linux/platform_data/shtc1.h>
+#ifdef CONFIG_OF

No. Please no conditional includes.

OK, will be fixed and same for below.


+#include <linux/of.h>
+#endif
/* commands (high precision mode) */
static const unsigned char shtc1_cmd_measure_blocking_hpm[] = { 0x7C, 0xA2 };
@@ -196,6 +199,10 @@ static int shtc1_probe(struct i2c_client *client,
enum shtcx_chips chip = id->driver_data;
struct i2c_adapter *adap = client->adapter;
struct device *dev = &client->dev;
+#ifdef CONFIG_OF
+ struct device_node *np = dev->of_node;
+ u8 value;
+#endif
if (!i2c_check_functionality(adap, I2C_FUNC_I2C)) {
dev_err(dev, "plain i2c transactions not supported\n");
@@ -235,6 +242,20 @@ static int shtc1_probe(struct i2c_client *client,
if (client->dev.platform_data)
data->setup = *(struct shtc1_platform_data *)dev->platform_data;
+
+#ifdef CONFIG_OF

Unnecessary ifdef. Selection of devicetree data or not can be made
based on np != NULL. Also, if devictree data is available, platform
data should be ignored to avoid confusion.

Ok, I wasn't aware this rule. Will change it.


+ if (np) {
+ if (of_property_read_bool(np, "sensirion,blocking_io")) {
+ of_property_read_u8(np, "sensirion,blocking_io", &value);
+ data->setup.blocking_io = (value > 0) ? true : false;
+ }
Why this complexity, and why not just make the property a boolean ?

+ if (of_property_read_bool(np, "sensicon,high_precision")) {
+ of_property_read_u8(np, "sensirion,high_precision", &value);
+ data->setup.high_precision = (value > 0) ? true : false;

"sensicon,high_precision" should also be a boolean.

+ }
+ }
+#endif
+
shtc1_select_command(data);
mutex_init(&data->update_lock);
@@ -257,6 +278,15 @@ static const struct i2c_device_id shtc1_id[] = {
};
MODULE_DEVICE_TABLE(i2c, shtc1_id);
+#ifdef CONFIG_OF
+static const struct of_device_id shtc1_of_match[] = {
+ { .compatible = "sensirion,shtc1" },
+ { .compatible = "sensirion,shtw1" },
+ { .compatible = "sensirion,shtc3" },
+ { }
+};
+MODULE_DEVICE_TABLE(of, shtc1_of_match);
+#endif
static struct i2c_driver shtc1_i2c_driver = {
.driver.name = "shtc1",
.probe = shtc1_probe,

Not sure how this works without setting of_match_table. I guess it just works
accidentally because .id_table also provides a devicetree match. Still,
of_match_table should be set.

Thanks, I will take care of this in the v2 version.


Guenter