smb347-charger: Move IRQ enabling to the end of probe

There is a potential problem if we call smb347_irq_enable() from
smb347_irq_init() because smb347_irq_enable() makes the device registers
read-only once it returns and smb347_irq_init() expects them to still be
read-write. Currently no harm happens because it is the last call we make
in smb347_irq_init().

Anyway a better place for enabling IRQs is at the end of probe function
and this is also symmetric to call smb347_irq_disable() which is done at
the beginning of remove function.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Signed-off-by: Anton Vorontsov <anton.vorontsov@linaro.org>

drivers/power/smb347-charger.c

@@ -925,10 +925,6 @@ static int smb347_irq_init(struct smb347_charger *smb)
 	if (ret < 0)
 		goto fail_readonly;
 
-	ret = smb347_irq_enable(smb);
-	if (ret < 0)
-		goto fail_readonly;
-
 	smb347_set_writable(smb, false);
 	smb->client->irq = irq;
 	return 0;
@@ -1241,6 +1237,8 @@ static int smb347_probe(struct i2c_client *client,
 		if (ret < 0) {
 			dev_warn(dev, "failed to initialize IRQ: %d\n", ret);
 			dev_warn(dev, "disabling IRQ support\n");
+		} else {
+			smb347_irq_enable(smb);
 		}
 	}