On 17.09.2015 05:11, Andreas Dannenberg wrote:

On Wed, Sep 16, 2015 at 05:31:58PM +0900, Krzysztof Kozlowski wrote:

quoted

quoted

@@ -1087,7 +1176,13 @@ static int bq24257_probe(struct i2c_client *client,
 		return ret;
 	}
 
-	if (client->irq <= 0)
+	/*
+	 * When using device firmware we either take the IRQ specified directly
+	 * or probe for a pin named BQ24257_STAT_IRQ. In case of using platform
+	 * data always run the IRQ pin probe function to finish IRQ and GPIO
+	 * setup based on the stat pin specified in the platform data.
+	 */

Why?

The original driver would allow configuring the IRQ either through
client->irq or through a separete DT entry called "stat_gpios" (which
seems redundant... might it be required for ACPI? Or...?). I modeled the
configuration via pdata the same way, this way aligning the fields in
the pdata structure with the DT entries. Let me know if you have a
preference to simplify this.

Right, it makes sense now.

Best regards,
Krzysztof