[PATCH] iio: chemical: ams-iaq-core: update driver name and DT match table

[PATCH] iio: chemical: ams-iaq-core: update driver name and DT match table

[Matt Ranostay]

DT match name should be similar to the i2c device table entry, and also
make the iio name match the driver name versus the i2c device location.

Signed-off-by: Matt Ranostay <mranostay@gmail.com>
---
 drivers/iio/chemical/ams-iaq-core.c | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

diff --git a/drivers/iio/chemical/ams-iaq-core.c b/drivers/iio/chemical/ams-iaq-core.c
index 41a8e6f2e31d..4c7d1f918d34 100644
--- a/drivers/iio/chemical/ams-iaq-core.c
+++ b/drivers/iio/chemical/ams-iaq-core.c
@@ -27,6 +27,8 @@
 #define AMS_IAQCORE_VOC_RESISTANCE_IDX	1
 #define AMS_IAQCORE_VOC_TVOC_IDX	2
 
+#define AMS_IAQCORE_DRVNAME		"ams-iaq-core"
+
 struct ams_iaqcore_reading {
 	__be16 co2_ppm;
 	u8 status;
@@ -164,7 +166,7 @@ static int ams_iaqcore_probe(struct i2c_client *client,
 
 	indio_dev->dev.parent = &client->dev;
 	indio_dev->info = &ams_iaqcore_info,
-	indio_dev->name = dev_name(&client->dev);
+	indio_dev->name = AMS_IAQCORE_DRVNAME;
 	indio_dev->modes = INDIO_DIRECT_MODE;
 
 	indio_dev->channels = ams_iaqcore_channels;
@@ -180,14 +182,14 @@ static const struct i2c_device_id ams_iaqcore_id[] = {
 MODULE_DEVICE_TABLE(i2c, ams_iaqcore_id);
 
 static const struct of_device_id ams_iaqcore_dt_ids[] = {
-	{ .compatible = "ams,iaq-core" },
+	{ .compatible = "ams,ams-iaq-core" },
 	{ }
 };
 MODULE_DEVICE_TABLE(of, ams_iaqcore_dt_ids);
 
 static struct i2c_driver ams_iaqcore_driver = {
 	.driver = {
-		.name	= "ams-iaq-core",
+		.name = AMS_IAQCORE_DRVNAME,
 		.of_match_table = of_match_ptr(ams_iaqcore_dt_ids),
 	},
 	.probe = ams_iaqcore_probe,
-- 
2.7.4

Re: [PATCH] iio: chemical: ams-iaq-core: update driver name and DT match table

[Lars-Peter Clausen]

On 08/02/2016 06:25 AM, Matt Ranostay wrote:
> DT match name should be similar to the i2c device table entry.

Why?

Re: [PATCH] iio: chemical: ams-iaq-core: update driver name and DT match table

[Matt Ranostay]

On Wed, Aug 3, 2016 at 5:13 AM, Lars-Peter Clausen <lars@metafoo.de> wrote:
> On 08/02/2016 06:25 AM, Matt Ranostay wrote:
>> DT match name should be similar to the i2c device table entry.
>
> Why?

To be consistent mostly.   But I have noticed  compatible =
"ams,iaq-core" doesn't actually match...
Does anyone here know if i2c-core just checks the id_table, and not
of_match_table as well?

Anyone have thoughts on this but I dig into this too deep?

>

Re: [PATCH] iio: chemical: ams-iaq-core: update driver name and DT match table

[Lars-Peter Clausen]

On 08/03/2016 10:02 PM, Matt Ranostay wrote:
> On Wed, Aug 3, 2016 at 5:13 AM, Lars-Peter Clausen <lars@metafoo.de> wrote:
>> On 08/02/2016 06:25 AM, Matt Ranostay wrote:
>>> DT match name should be similar to the i2c device table entry.
>>
>> Why?
> 
> To be consistent mostly.   But I have noticed  compatible =
> "ams,iaq-core" doesn't actually match...

Well, consistent with what? The normal pattern for devicetree compatible
strings is vendor,partname. So that would suggest that ams,iaq-core is the
preferred solution if it is supposed to be consistent. Btw. there seem to be
no binding documents for this part.

> Does anyone here know if i2c-core just checks the id_table, and not
> of_match_table as well?
> 
> Anyone have thoughts on this but I dig into this too deep?

It uses of_match_table when it is set. See
http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/tree/drivers/i2c/i2c-core.c#n609

Re: [PATCH] iio: chemical: ams-iaq-core: update driver name and DT match table

[Matt Ranostay]

On Wed, Aug 3, 2016 at 1:08 PM, Lars-Peter Clausen <lars@metafoo.de> wrote:
> On 08/03/2016 10:02 PM, Matt Ranostay wrote:
>> On Wed, Aug 3, 2016 at 5:13 AM, Lars-Peter Clausen <lars@metafoo.de> wrote:
>>> On 08/02/2016 06:25 AM, Matt Ranostay wrote:
>>>> DT match name should be similar to the i2c device table entry.
>>>
>>> Why?
>>
>> To be consistent mostly.   But I have noticed  compatible =
>> "ams,iaq-core" doesn't actually match...
>
> Well, consistent with what? The normal pattern for devicetree compatible
> strings is vendor,partname. So that would suggest that ams,iaq-core is the
> preferred solution if it is supposed to be consistent. Btw. there seem to be
> no binding documents for this part.
>
>> Does anyone here know if i2c-core just checks the id_table, and not
>> of_match_table as well?
>>
>> Anyone have thoughts on this but I dig into this too deep?
>
> It uses of_match_table when it is set. See
> http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/tree/drivers/i2c/i2c-core.c#n609

Yeah thought would enumerate it as well.. but could be that someone
broke something in i2c controller or core. Will test again :)

>