This compliments the existing scheme that lets it be set via
platform_data.
Signed-off-by: Jason Gunthorpe <[email protected]>
---
Documentation/devicetree/bindings/hwmon/lm87.txt | 27 ++++++++++++++++++++++++
drivers/hwmon/lm87.c | 8 ++++++-
2 files changed, 34 insertions(+), 1 deletion(-)
create mode 100644 Documentation/devicetree/bindings/hwmon/lm87.txt
diff --git a/Documentation/devicetree/bindings/hwmon/lm87.txt b/Documentation/devicetree/bindings/hwmon/lm87.txt
new file mode 100644
index 00000000000000..4ea2646ce8d2f0
--- /dev/null
+++ b/Documentation/devicetree/bindings/hwmon/lm87.txt
@@ -0,0 +1,27 @@
+* LM87 I2C Temperature and Fan sensor
+
+Required properties:
+- compatible: should contain one of
+ * "national,lm87"
+ * "ti,lm87"
+
+Optional properties:
+- channel-mode: Write the specified value to register 0x16. If not specified
+ the driver will read the value from the chip at startup.
+
+ The value controls how the driver operates certain multi-use
+ pins, eg bit 2 will select between temp_input3 or
+ in0&in5 operation.
+
+See Documentation/devicetree/bindings/i2c/i2c.txt for more required and
+optional properties.
+
+Example:
+
+i2c_master {
+ temperature-sensor@0 {
+ compatible = "ti,lm87";
+ channel-mode = <0x4>;
+ reg = <0>;
+ };
+};
diff --git a/drivers/hwmon/lm87.c b/drivers/hwmon/lm87.c
index 81cb898245a107..ccc0c913b4534e 100644
--- a/drivers/hwmon/lm87.c
+++ b/drivers/hwmon/lm87.c
@@ -853,9 +853,15 @@ static void lm87_restore_config(void *arg)
static int lm87_init_client(struct i2c_client *client)
{
struct lm87_data *data = i2c_get_clientdata(client);
+ u32 channel;
int rc;
- if (dev_get_platdata(&client->dev)) {
+ if (client->dev.of_node &&
+ !of_property_read_u32(client->dev.of_node, "channel-mode",
+ &channel)) {
+ data->channel = channel;
+ lm87_write_value(client, LM87_REG_CHANNEL_MODE, data->channel);
+ } else if (dev_get_platdata(&client->dev)) {
data->channel = *(u8 *)dev_get_platdata(&client->dev);
lm87_write_value(client,
LM87_REG_CHANNEL_MODE, data->channel);
--
2.7.4
On Thu, Nov 10, 2016 at 10:19:36AM -0700, Jason Gunthorpe wrote:
> This compliments the existing scheme that lets it be set via
> platform_data.
>
> Signed-off-by: Jason Gunthorpe <[email protected]>
I would prefer https://www.spinics.net/lists/devicetree/msg148819.html
which appears to address the same problem and is less cryptic.
NACK unless someone convinces me otherwise.
Guenter
> ---
> Documentation/devicetree/bindings/hwmon/lm87.txt | 27 ++++++++++++++++++++++++
> drivers/hwmon/lm87.c | 8 ++++++-
> 2 files changed, 34 insertions(+), 1 deletion(-)
> create mode 100644 Documentation/devicetree/bindings/hwmon/lm87.txt
>
> diff --git a/Documentation/devicetree/bindings/hwmon/lm87.txt b/Documentation/devicetree/bindings/hwmon/lm87.txt
> new file mode 100644
> index 00000000000000..4ea2646ce8d2f0
> --- /dev/null
> +++ b/Documentation/devicetree/bindings/hwmon/lm87.txt
> @@ -0,0 +1,27 @@
> +* LM87 I2C Temperature and Fan sensor
> +
> +Required properties:
> +- compatible: should contain one of
> + * "national,lm87"
> + * "ti,lm87"
> +
> +Optional properties:
> +- channel-mode: Write the specified value to register 0x16. If not specified
> + the driver will read the value from the chip at startup.
> +
> + The value controls how the driver operates certain multi-use
> + pins, eg bit 2 will select between temp_input3 or
> + in0&in5 operation.
> +
> +See Documentation/devicetree/bindings/i2c/i2c.txt for more required and
> +optional properties.
> +
> +Example:
> +
> +i2c_master {
> + temperature-sensor@0 {
> + compatible = "ti,lm87";
> + channel-mode = <0x4>;
> + reg = <0>;
> + };
> +};
> diff --git a/drivers/hwmon/lm87.c b/drivers/hwmon/lm87.c
> index 81cb898245a107..ccc0c913b4534e 100644
> --- a/drivers/hwmon/lm87.c
> +++ b/drivers/hwmon/lm87.c
> @@ -853,9 +853,15 @@ static void lm87_restore_config(void *arg)
> static int lm87_init_client(struct i2c_client *client)
> {
> struct lm87_data *data = i2c_get_clientdata(client);
> + u32 channel;
> int rc;
>
> - if (dev_get_platdata(&client->dev)) {
> + if (client->dev.of_node &&
> + !of_property_read_u32(client->dev.of_node, "channel-mode",
> + &channel)) {
> + data->channel = channel;
> + lm87_write_value(client, LM87_REG_CHANNEL_MODE, data->channel);
> + } else if (dev_get_platdata(&client->dev)) {
> data->channel = *(u8 *)dev_get_platdata(&client->dev);
> lm87_write_value(client,
> LM87_REG_CHANNEL_MODE, data->channel);
> --
> 2.7.4
On Thu, Nov 10, 2016 at 03:47:07PM -0800, Guenter Roeck wrote:
> On Thu, Nov 10, 2016 at 10:19:36AM -0700, Jason Gunthorpe wrote:
> > This compliments the existing scheme that lets it be set via
> > platform_data.
> >
> > Signed-off-by: Jason Gunthorpe <[email protected]>
>
> I would prefer https://www.spinics.net/lists/devicetree/msg148819.html
> which appears to address the same problem and is less cryptic.
Is there a patch someplace that implements that proposed DT binding? I
couldn't find anything.
> NACK unless someone convinces me otherwise.
*shrug* both styles of definition seem to be used in various device
tree bindings.
The downside with Mahoda's proposal is that it only covers 3 of the 8
configuration bits, and will require alot of code for parsing. It
seems really clunky, and why is a regulator involved to set the A/D
sensitivity? Weird.
But I only need has-temp3, so if a patch appears please cc me and I'll
review it here..
Regards,
Jason