On 5/3/2018 11:29 AM, Sergei Shtylyov wrote:

quoted

From: Peter Chen <redacted>

With that we can clear any pending events and the port is registered
so driver can be ready to handle typec events once we request irq.

Signed-off-by: Peter Chen <redacted>
Signed-off-by: Li Jun <redacted>
---
  drivers/staging/typec/tcpci.c | 15 +++++++++------
  1 file changed, 9 insertions(+), 6 deletions(-)

diff --git a/drivers/staging/typec/tcpci.c b/drivers/staging/typec/tcpci.c
index a21aaf5..139c2ab 100644
--- a/drivers/staging/typec/tcpci.c
+++ b/drivers/staging/typec/tcpci.c

@@ -537,24 +537,27 @@ static int tcpci_probe(struct i2c_client *client,

      if (IS_ERR(chip->data.regmap))
          return PTR_ERR(chip->data.regmap);
+    i2c_set_clientdata(client, chip);
+
      /* Disable chip interrupts before requesting irq */
      err = regmap_raw_write(chip->data.regmap, TCPC_ALERT_MASK, &val,
                     sizeof(u16));
      if (err < 0)
          return err;
+    chip->tcpci = tcpci_register_port(&client->dev, &chip->data);
+    if (PTR_ERR_OR_ZERO(chip->tcpci))

    Is NULL an error indication too?

quoted

+        return PTR_ERR(chip->tcpci);

    If so, you'll return a success indication here.

    ... when chip->tcpci is NULL, I mean. Seems wrong...

quoted

+
      err = devm_request_threaded_irq(&client->dev, client->irq, NULL,
                      _tcpci_irq,
                      IRQF_ONESHOT | IRQF_TRIGGER_LOW,
                      dev_name(&client->dev), chip);
-    if (err < 0)
+    if (err < 0) {
+        tcpci_unregister_port(chip->tcpci);
          return err;
+    }
-    chip->tcpci = tcpci_register_port(&client->dev, &chip->data);
-    if (PTR_ERR_OR_ZERO(chip->tcpci))
-        return PTR_ERR(chip->tcpci);
-

    Ah, you're only moving the questionable code...

quoted

-    i2c_set_clientdata(client, chip);
      return 0;
  }

MBR, Sergei