| Message ID | 20160920052240.10242-1-mahoda.ratnayaka@alliedtelesis.co.nz (mailing list archive) |
|---|---|
| State | Changes Requested |
| Headers | show |
On 09/19/2016 10:22 PM, Mahoda Ratnayaka wrote: > Currently there is no method for setting the channel > value from the DTS file. When, the driver uses a dts > file to initialize the driver platform_data is not set. > As a the result channel variable may not be set correctly. > > Without the channel variable set correctly, some of the > sensors will not be initialized correctly. For example > temp3 sensor sysfs entries. > > This adds the required functionality to set the channel > variable from the DTS file. This is done via reading the > reading a property named "channel" from the lm87 driver. > > Signed-off-by: Mahoda Ratnayaka <mahoda.ratnayaka@alliedtelesis.co.nz> > --- > > Notes: > changes since v1: > Removed unncessary variables channel and np. > Update the code as per review comments. > > Documentation/devicetree/bindings/hwmon/lm87.txt | 9 +++++++++ > drivers/hwmon/lm87.c | 7 ++++++- > 2 files changed, 15 insertions(+), 1 deletion(-) > > diff --git a/Documentation/devicetree/bindings/hwmon/lm87.txt b/Documentation/devicetree/bindings/hwmon/lm87.txt > index fefcb48..1906e08 100644 > --- a/Documentation/devicetree/bindings/hwmon/lm87.txt > +++ b/Documentation/devicetree/bindings/hwmon/lm87.txt > @@ -6,9 +6,18 @@ Required properties: > > - reg: I2C address > > +optional properties: > +- channels: Value for the channel mode register. > + This allows configuration of pins with > + selectable uses on the LM87 device (e.g. > + choosing between voltage and temperature > + inputs). > + See hwmon/lm87 for further information Rob is the ultimate person to comment here, but I think the property description by itself should be complete and technically independent of the implementation. Documentation/hwmon/lm87 describes the implementation, so an (incomplete) reference to it seems inappropriate. > + > Example: > > lm87@2e { > compatible = "ti,lm87"; > reg = <0x2e>; > + channels=<0x4>; > }; Please move those property descriptions into the first patch. It is confusing to have two patches touching the properties, and the first patch as it is right now is really useless. > diff --git a/drivers/hwmon/lm87.c b/drivers/hwmon/lm87.c > index a5e2958..358c1d4 100644 > --- a/drivers/hwmon/lm87.c > +++ b/drivers/hwmon/lm87.c > @@ -858,8 +858,13 @@ static void lm87_remove_files(struct i2c_client *client) > static void lm87_init_client(struct i2c_client *client) > { > struct lm87_data *data = i2c_get_clientdata(client); > + u8 val = 0; > > - if (dev_get_platdata(&client->dev)) { > + if (!of_property_read_u8(client->dev.of_node, "channels", &val)) { This ignores property value errors (-ENODATA, -EOVERFLOW). Is this what you want ? > + data->channel = val; > + 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); > -- To unsubscribe from this list: send the line "unsubscribe linux-hwmon" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On 09/19/16 22:22, Mahoda Ratnayaka wrote: > Currently there is no method for setting the channel > value from the DTS file. When, the driver uses a dts > file to initialize the driver platform_data is not set. > As a the result channel variable may not be set correctly. > > Without the channel variable set correctly, some of the > sensors will not be initialized correctly. For example > temp3 sensor sysfs entries. > > This adds the required functionality to set the channel > variable from the DTS file. This is done via reading the > reading a property named "channel" from the lm87 driver. > > Signed-off-by: Mahoda Ratnayaka <mahoda.ratnayaka@alliedtelesis.co.nz> > --- > > Notes: > changes since v1: > Removed unncessary variables channel and np. > Update the code as per review comments. > > Documentation/devicetree/bindings/hwmon/lm87.txt | 9 +++++++++ > drivers/hwmon/lm87.c | 7 ++++++- > 2 files changed, 15 insertions(+), 1 deletion(-) > > diff --git a/Documentation/devicetree/bindings/hwmon/lm87.txt b/Documentation/devicetree/bindings/hwmon/lm87.txt > index fefcb48..1906e08 100644 > --- a/Documentation/devicetree/bindings/hwmon/lm87.txt > +++ b/Documentation/devicetree/bindings/hwmon/lm87.txt I do not see the file Documentation/devicetree/bindings/hwmon/lm87.txt in Linux 4.8-rc7. I am guessing that it might already be in a tree headed upstream. Can you point me to it? > @@ -6,9 +6,18 @@ Required properties: > > - reg: I2C address > > +optional properties: > +- channels: Value for the channel mode register. > + This allows configuration of pins with > + selectable uses on the LM87 device (e.g. > + choosing between voltage and temperature > + inputs). > + See hwmon/lm87 for further information > + That is configuration information, not hardware description. Devicetree contains hardware description. Hardware description would instead be a property that said what is hooked up to the shared pins, eg temp3, fan1, fan2, VID, or IRQ. The driver should know what value to place in channel for each of the different cases. > Example: > > lm87@2e { > compatible = "ti,lm87"; > reg = <0x2e>; > + channels=<0x4>; > }; > diff --git a/drivers/hwmon/lm87.c b/drivers/hwmon/lm87.c > index a5e2958..358c1d4 100644 > --- a/drivers/hwmon/lm87.c > +++ b/drivers/hwmon/lm87.c > @@ -858,8 +858,13 @@ static void lm87_remove_files(struct i2c_client *client) > static void lm87_init_client(struct i2c_client *client) > { > struct lm87_data *data = i2c_get_clientdata(client); > + u8 val = 0; > > - if (dev_get_platdata(&client->dev)) { > + if (!of_property_read_u8(client->dev.of_node, "channels", &val)) { > + data->channel = val; > + 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); > -- To unsubscribe from this list: send the line "unsubscribe linux-hwmon" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Thanks Guenter for the reply, I'll update things as you suggested and send the patches back again. I'll try to get the documentation approved first and then send back the lm87 driver update patch. Thanks, Mahoda
On Tue, Sep 20, 2016 at 02:55:42AM -0700, Guenter Roeck wrote: > On 09/19/2016 10:22 PM, Mahoda Ratnayaka wrote: > >Currently there is no method for setting the channel > >value from the DTS file. When, the driver uses a dts > >file to initialize the driver platform_data is not set. > >As a the result channel variable may not be set correctly. > > > >Without the channel variable set correctly, some of the > >sensors will not be initialized correctly. For example > >temp3 sensor sysfs entries. > > > >This adds the required functionality to set the channel > >variable from the DTS file. This is done via reading the > >reading a property named "channel" from the lm87 driver. > > > >Signed-off-by: Mahoda Ratnayaka <mahoda.ratnayaka@alliedtelesis.co.nz> > >--- > > > >Notes: > > changes since v1: > > Removed unncessary variables channel and np. > > Update the code as per review comments. > > > > Documentation/devicetree/bindings/hwmon/lm87.txt | 9 +++++++++ > > drivers/hwmon/lm87.c | 7 ++++++- > > 2 files changed, 15 insertions(+), 1 deletion(-) > > > >diff --git a/Documentation/devicetree/bindings/hwmon/lm87.txt b/Documentation/devicetree/bindings/hwmon/lm87.txt > >index fefcb48..1906e08 100644 > >--- a/Documentation/devicetree/bindings/hwmon/lm87.txt > >+++ b/Documentation/devicetree/bindings/hwmon/lm87.txt > >@@ -6,9 +6,18 @@ Required properties: > > > > - reg: I2C address > > > >+optional properties: > >+- channels: Value for the channel mode register. > >+ This allows configuration of pins with > >+ selectable uses on the LM87 device (e.g. > >+ choosing between voltage and temperature > >+ inputs). > >+ See hwmon/lm87 for further information > > Rob is the ultimate person to comment here, but I think the property description > by itself should be complete and technically independent of the implementation. > Documentation/hwmon/lm87 describes the implementation, so an (incomplete) > reference to it seems inappropriate. Yes, please make the binding stand on its own. And please put all of the binding in one patch. You're describing h/w, not adding s/w features. As far as whether this is configuration or h/w description, it depends on what defines the configuration. If it is fixed by the h/w design then this is okay. If it is a user choice, then no it should not be in DT. Seems like this would be the former case as probably temperature measurement requires a thermistor or something. Rob -- To unsubscribe from this list: send the line "unsubscribe linux-hwmon" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/Documentation/devicetree/bindings/hwmon/lm87.txt b/Documentation/devicetree/bindings/hwmon/lm87.txt index fefcb48..1906e08 100644 --- a/Documentation/devicetree/bindings/hwmon/lm87.txt +++ b/Documentation/devicetree/bindings/hwmon/lm87.txt @@ -6,9 +6,18 @@ Required properties: - reg: I2C address +optional properties: +- channels: Value for the channel mode register. + This allows configuration of pins with + selectable uses on the LM87 device (e.g. + choosing between voltage and temperature + inputs). + See hwmon/lm87 for further information + Example: lm87@2e { compatible = "ti,lm87"; reg = <0x2e>; + channels=<0x4>; }; diff --git a/drivers/hwmon/lm87.c b/drivers/hwmon/lm87.c index a5e2958..358c1d4 100644 --- a/drivers/hwmon/lm87.c +++ b/drivers/hwmon/lm87.c @@ -858,8 +858,13 @@ static void lm87_remove_files(struct i2c_client *client) static void lm87_init_client(struct i2c_client *client) { struct lm87_data *data = i2c_get_clientdata(client); + u8 val = 0; - if (dev_get_platdata(&client->dev)) { + if (!of_property_read_u8(client->dev.of_node, "channels", &val)) { + data->channel = val; + 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);
Currently there is no method for setting the channel value from the DTS file. When, the driver uses a dts file to initialize the driver platform_data is not set. As a the result channel variable may not be set correctly. Without the channel variable set correctly, some of the sensors will not be initialized correctly. For example temp3 sensor sysfs entries. This adds the required functionality to set the channel variable from the DTS file. This is done via reading the reading a property named "channel" from the lm87 driver. Signed-off-by: Mahoda Ratnayaka <mahoda.ratnayaka@alliedtelesis.co.nz> --- Notes: changes since v1: Removed unncessary variables channel and np. Update the code as per review comments. Documentation/devicetree/bindings/hwmon/lm87.txt | 9 +++++++++ drivers/hwmon/lm87.c | 7 ++++++- 2 files changed, 15 insertions(+), 1 deletion(-)