On Sat, 3 Nov 2018 19:49:46 -0300
Matheus Tavares <matheus.bernardino@xxxxxx> wrote:
> Previously, devm_iio_device_register was being called before the
> spi_setup call and the spi_device's max_speed_hz and mode assignments.
> This could lead to a race condition since the driver was still being
> set up after it was already made ready to use. To fix it, this patch
> moves the device registration to the end of ad2s90_probe.
>
> Signed-off-by: Matheus Tavares <matheus.bernardino@xxxxxx>
Applied. This is definitely looking more like a 'modern' IIO driver.
All sorts of less than sensible things like this used to be done
when all of us were new to drivers.
Thanks,
Jonathan
> ---
> drivers/staging/iio/resolver/ad2s90.c | 6 +-----
> 1 file changed, 1 insertion(+), 5 deletions(-)
>
> diff --git a/drivers/staging/iio/resolver/ad2s90.c b/drivers/staging/iio/resolver/ad2s90.c
> index 54ad85bd9dc6..8f79cccf4814 100644
> --- a/drivers/staging/iio/resolver/ad2s90.c
> +++ b/drivers/staging/iio/resolver/ad2s90.c
> @@ -79,10 +79,6 @@ static int ad2s90_probe(struct spi_device *spi)
> indio_dev->num_channels = 1;
> indio_dev->name = spi_get_device_id(spi)->name;
>
> - ret = devm_iio_device_register(indio_dev->dev.parent, indio_dev);
> - if (ret)
> - return ret;
> -
> /* need 600ns between CS and the first falling edge of SCLK */
> spi->max_speed_hz = 830000;
> spi->mode = SPI_MODE_3;
> @@ -93,7 +89,7 @@ static int ad2s90_probe(struct spi_device *spi)
> return ret;
> }
>
> - return 0;
> + return devm_iio_device_register(indio_dev->dev.parent, indio_dev);
> }
>
> static const struct spi_device_id ad2s90_id[] = {
