vl53l0x_read_raw() starts a single-shot ranging measurement and reads
back the result. Once the triggered buffer is enabled the sensor runs in
continuous mode and its data-ready interrupt is routed to the trigger,
so a concurrent in_distance_raw read disturbs the streaming setup and
never gets its completion, returning -ETIMEDOUT.

The original submission claimed direct mode here, but it was dropped
during review because the driver had no buffer support at the time [1].
Continuous (buffered) mode was later added without restoring the
claim [2], reintroducing the conflict.

Reject direct reads while buffered capture is active by claiming direct
mode around the measurement, as the vl53l1x sibling already does.

Fixes: 762186c6e7b1 ("iio: proximity: vl53l0x-i2c: Added continuous mode support")
Link: https://lore.kernel.org/linux-iio/20180911160300.GA9212@himanshu-Vostro-3559/ [1]
Link: https://lore.kernel.org/linux-iio/20240909101508.263085-3-abhashkumarjha123@gmail.com/ [2]
Signed-off-by: Donggeun Yoo <donggeunyoo.kernel@gmail.com>
---
Only compile-tested; I don't have VL53L0X hardware to exercise the race
yet. The change mirrors the existing vl53l1x sibling.

 drivers/iio/proximity/vl53l0x-i2c.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/drivers/iio/proximity/vl53l0x-i2c.c b/drivers/iio/proximity/vl53l0x-i2c.c
index 9fe14ceb8be7..173e69753df7 100644
--- a/drivers/iio/proximity/vl53l0x-i2c.c
+++ b/drivers/iio/proximity/vl53l0x-i2c.c
@@ -219,7 +219,10 @@ static int vl53l0x_read_raw(struct iio_dev *indio_dev,
 
 	switch (mask) {
 	case IIO_CHAN_INFO_RAW:
+		if (!iio_device_claim_direct(indio_dev))
+			return -EBUSY;
 		ret = vl53l0x_read_proximity(data, chan, val);
+		iio_device_release_direct(indio_dev);
 		if (ret < 0)
 			return ret;
 
-- 
2.53.0

On Wed,  2 Sep 2026 00:21:26 +0900
Donggeun Yoo <donggeunyoo.kernel@gmail.com> wrote:

> vl53l0x_read_raw() starts a single-shot ranging measurement and reads
> back the result. Once the triggered buffer is enabled the sensor runs in
> continuous mode and its data-ready interrupt is routed to the trigger,
> so a concurrent in_distance_raw read disturbs the streaming setup and
> never gets its completion, returning -ETIMEDOUT.
> 
> The original submission claimed direct mode here, but it was dropped
> during review because the driver had no buffer support at the time [1].
> Continuous (buffered) mode was later added without restoring the
> claim [2], reintroducing the conflict.
> 
> Reject direct reads while buffered capture is active by claiming direct
> mode around the measurement, as the vl53l1x sibling already does.
> 
> Fixes: 762186c6e7b1 ("iio: proximity: vl53l0x-i2c: Added continuous mode support")
> Link: https://lore.kernel.org/linux-iio/20180911160300.GA9212@himanshu-Vostro-3559/ [1]
> Link: https://lore.kernel.org/linux-iio/20240909101508.263085-3-abhashkumarjha123@gmail.com/ [2]
> Signed-off-by: Donggeun Yoo <donggeunyoo.kernel@gmail.com>
Applied to the fixes-togreg branch of iio.git.

Thanks,

Jonathan

> ---
> Only compile-tested; I don't have VL53L0X hardware to exercise the race
> yet. The change mirrors the existing vl53l1x sibling.
> 
>  drivers/iio/proximity/vl53l0x-i2c.c | 3 +++
>  1 file changed, 3 insertions(+)
> 
> diff --git a/drivers/iio/proximity/vl53l0x-i2c.c b/drivers/iio/proximity/vl53l0x-i2c.c
> index 9fe14ceb8be7..173e69753df7 100644
> --- a/drivers/iio/proximity/vl53l0x-i2c.c
> +++ b/drivers/iio/proximity/vl53l0x-i2c.c
> @@ -219,7 +219,10 @@ static int vl53l0x_read_raw(struct iio_dev *indio_dev,
>  
>  	switch (mask) {
>  	case IIO_CHAN_INFO_RAW:
> +		if (!iio_device_claim_direct(indio_dev))
> +			return -EBUSY;
>  		ret = vl53l0x_read_proximity(data, chan, val);
> +		iio_device_release_direct(indio_dev);
>  		if (ret < 0)
>  			return ret;
>