On Wed, Oct 30, 2024 at 12:53:57PM +0300, Dan Carpenter wrote:
> Hello Andy Shevchenko,
>
> Commit 14686836fb69 ("iio: light: isl29018: Replace a variant of
> iio_get_acpi_device_name_and_data()") from Oct 24, 2024 (linux-next),
> leads to the following Smatch static checker warning:
>
> drivers/iio/light/isl29018.c:724 isl29018_probe() error: uninitialized symbol 'ddata'.
> drivers/iio/light/ltr501.c:1514 ltr501_probe() error: uninitialized symbol 'ddata'.
>
> drivers/iio/light/isl29018.c
> 701 static int isl29018_probe(struct i2c_client *client)
> 702 {
> 703 const struct i2c_device_id *id = i2c_client_get_device_id(client);
> 704 struct isl29018_chip *chip;
> 705 struct iio_dev *indio_dev;
> 706 const void *ddata;
> 707 const char *name;
> 708 int dev_id;
> 709 int err;
> 710
> 711 indio_dev = devm_iio_device_alloc(&client->dev, sizeof(*chip));
> 712 if (!indio_dev)
> 713 return -ENOMEM;
> 714
> 715 chip = iio_priv(indio_dev);
> 716
> 717 i2c_set_clientdata(client, indio_dev);
> 718
> 719 if (id) {
> 720 name = id->name;
> 721 dev_id = id->driver_data;
> 722 } else {
> 723 name = iio_get_acpi_device_name_and_data(&client->dev, &ddata);
> --> 724 dev_id = (intptr_t)ddata;
>
> How do we know that iio_get_acpi_device_name_and_data() will succeed?
Ideally we need to file &ddata with NULL in such case, but it will be
equal to 0, so it only works with the chip_info in place.
Let me look into this once more, thanks for the good catch!
> 725 }
> 726
> 727 mutex_init(&chip->lock);
> 728
> 729 chip->type = dev_id;
--
With Best Regards,
Andy Shevchenko
