There are a lot of devm_*() APIs defined in the device.h and sometimes either
the rest of the device.h is in use or only stuff from (new) devres.h.
Especially helpful to have it separated for the headers that want to use
devm_add_action_or_reset() or similar APIs in inline functions.

I don't mind any subsystem in question to take it (IIO or driver core), but
I would need an immutable branch/tag since there is a pending series that
adds yet another devm_*() into the device.h and supposed to go via one of
mine trees. Of course it would be possible to route that series via another
tree.

Andy Shevchenko (2):
  driver core: Split devres APIs to device/devres.h
  iio: imu: st_lsm9ds0: Replace device.h with what is needed

 drivers/iio/imu/st_lsm9ds0/st_lsm9ds0_i2c.c |   2 +-
 drivers/iio/imu/st_lsm9ds0/st_lsm9ds0_spi.c |   2 +-
 include/linux/device.h                      | 122 +------------------
 include/linux/device/devres.h               | 124 ++++++++++++++++++++
 4 files changed, 129 insertions(+), 121 deletions(-)
 create mode 100644 include/linux/device/devres.h

-- 
2.43.0.rc1.1336.g36b5255a03ac

On Tue, Dec 03, 2024 at 09:48:50PM +0200, Andy Shevchenko wrote:
> There are a lot of devm_*() APIs defined in the device.h and sometimes either
> the rest of the device.h is in use or only stuff from (new) devres.h.
> Especially helpful to have it separated for the headers that want to use
> devm_add_action_or_reset() or similar APIs in inline functions.
> 
> I don't mind any subsystem in question to take it (IIO or driver core), but
> I would need an immutable branch/tag since there is a pending series that
> adds yet another devm_*() into the device.h and supposed to go via one of
> mine trees. Of course it would be possible to route that series via another
> tree.

FWIW, the mentioned change that may influence this series (and vice versa)
is https://lore.kernel.org/r/20241126172240.6044-1-raag.jadav@intel.com.

-- 
With Best Regards,
Andy Shevchenko