This series deals with cleaning up the TI OPT3001 sensor driver,
moving it to more modern kernel practices and improving the code style.

While reviewing, Jonathan Cameron (and eventually Sashiko) found a race
condition where userspace could start interacting with the device
before the hardware IRQ was set up.

Changes include:
- moving the driver to use devm_* functions
- IWYU cleanups
- removal of unnecessary macros and comments
- using dev_err_probe() in probe and probe path functions
- checkpatch.pl warning cleanups
- fixing a race condition found in opt3001_probe() function

Signed-off-by: Joshua Crofts <joshua.crofts1@gmail.com>
---
Changes in v4:
- Fix bad merge
- Edit commit messages
- PATCH 1: add free_irq on iio_device_register error
- PATCH 8: change comments
- Link to v3: https://lore.kernel.org/r/20260521-opt3001-cleanup-v3-0-820169dec8c3@gmail.com

Changes in v3:
- PATCH 1: fix build error
- PATCH 2: remove struct device member from struct opt3001
- PATCH 6: edit return statements
- Pull additional trailers from previous version's reviews
- Edit commit messages
- Link to v2: https://lore.kernel.org/r/20260512-opt3001-cleanup-v2-0-8018cf3a8a0a@gmail.com

Changes in v2:
- PATCH 1: added patch that fixes race condition
- PATCH 3: remove wrong usage of GENMASK()
- PATCH 4: add patch that moves driver to use local structs
- PATCH 5: add patch that ensures correct parenthesis alignment
- PATCH 6: change int to unsigned int
- PATCH 7: moved opt3001_read_id() function to use dev_err_probe()
- PATCH 9: removed unnecessary dev_err_probe() calls, reordering
- PATCH 10: added patch that adds a comment to mutex declaration
- Link to v1: https://lore.kernel.org/r/20260511-opt3001-cleanup-v1-0-f7879dc3455c@gmail.com

---
Joshua Crofts (8):
      iio: light: opt3001: move device registration to end of probe()
      iio: light: opt3001: use local struct device and i2c_client variables
      iio: light: opt3001: ensure correct parenthesis alignment
      iio: light: opt3001: localize for loop iterator
      iio: light: opt3001: prefer dev_err_probe()
      iio: light: opt3001: move driver to guard(mutex)() use
      iio: light: opt3001: switch driver to managed resources
      iio: light: opt3001: add comment to mutex

 drivers/iio/light/opt3001.c | 382 +++++++++++++++++++++-----------------------
 1 file changed, 183 insertions(+), 199 deletions(-)
---
base-commit: 0e7dbde323808f28c5220295bfc1c5bc6f08c3f4
change-id: 20260511-opt3001-cleanup-c1411de14392

Best regards,
-- 
Joshua Crofts <joshua.crofts1@gmail.com>

On Mon, 25 May 2026 at 10:20, Joshua Crofts via B4 Relay
<devnull+joshua.crofts1.gmail.com@kernel.org> wrote:
>
> This series deals with cleaning up the TI OPT3001 sensor driver,
> moving it to more modern kernel practices and improving the code style.
>
> While reviewing, Jonathan Cameron (and eventually Sashiko) found a race
> condition where userspace could start interacting with the device
> before the hardware IRQ was set up.
>
> Changes include:
> - moving the driver to use devm_* functions
> - IWYU cleanups
> - removal of unnecessary macros and comments
> - using dev_err_probe() in probe and probe path functions
> - checkpatch.pl warning cleanups
> - fixing a race condition found in opt3001_probe() function
>
> Signed-off-by: Joshua Crofts <joshua.crofts1@gmail.com>
> ---

Sashiko came back with a dead code one liner and then found
a new issue where a function returns instead of calling cleanup
functions. This is definitely a bug that should be marked for stable.

-- 
Kind regards

CJD