It's been another year of discussing the object life-time problems at
conferences. I2C is one of the offenders and its problems are more
complex than those of some other subsystems. It seems the revocable[1]
API may make its way into the kernel this year but even with it in
place, I2C won't be able to use it as there's currently nothing to
*revoke*. The struct device is embedded within the i2c_adapter struct
whose lifetime is tied to the provider device being bound to its driver.

Fixing this won't be fast and easy but nothing's going to happen if we
don't start chipping away at it. The ultimate goal in order to be able
to use an SRCU-based solution (revocable or otherwise) is to convert the
embedded struct device in struct i2c_adapter into an __rcu pointer that
can be *revoked*. To that end we need to hide all dereferences of
adap->dev in drivers.

I2C drivers use device_set_node() to assign a firmware node directly to
the struct device embedded in i2c_adapter. In order to hide the direct
dereferencing: provide a dedicated interface that wraps the call to
device_set_node() but takes the adapter as argument instead.

[1] https://lore.kernel.org/all/20251106152330.11733-1-tzungbi@kernel.org/

Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>
---
Bartosz Golaszewski (7):
      i2c: provide i2c_adapter_set_node()
      i2c: designware-common: use i2c_adapter_set_node()
      i2c: gpio: use i2c_adapter_set_node()
      i2c: ljca: use i2c_adapter_set_node()
      i2c: ls2x: use i2c_adapter_set_node()
      i2c: rzv2m: use i2c_adapter_set_node()
      i2c: usbio: use i2c_adapter_set_node()

 drivers/i2c/busses/i2c-designware-common.c | 2 +-
 drivers/i2c/busses/i2c-gpio.c              | 2 +-
 drivers/i2c/busses/i2c-ljca.c              | 2 +-
 drivers/i2c/busses/i2c-ls2x.c              | 2 +-
 drivers/i2c/busses/i2c-rzv2m.c             | 2 +-
 drivers/i2c/busses/i2c-usbio.c             | 2 +-
 include/linux/i2c.h                        | 7 +++++++
 7 files changed, 13 insertions(+), 6 deletions(-)
---
base-commit: 9448598b22c50c8a5bb77a9103e2d49f134c9578
change-id: 20251223-i2c-ada-dev-set-node-0636a8d3666c

Best regards,
-- 
Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>

On Tue, Dec 23, 2025 at 11:06:47AM +0100, Bartosz Golaszewski wrote:
> It's been another year of discussing the object life-time problems at
> conferences. I2C is one of the offenders and its problems are more
> complex than those of some other subsystems. It seems the revocable[1]
> API may make its way into the kernel this year but even with it in
> place, I2C won't be able to use it as there's currently nothing to
> *revoke*. The struct device is embedded within the i2c_adapter struct
> whose lifetime is tied to the provider device being bound to its driver.
> 
> Fixing this won't be fast and easy but nothing's going to happen if we
> don't start chipping away at it. The ultimate goal in order to be able
> to use an SRCU-based solution (revocable or otherwise) is to convert the
> embedded struct device in struct i2c_adapter into an __rcu pointer that
> can be *revoked*. To that end we need to hide all dereferences of
> adap->dev in drivers.
> 
> I2C drivers use device_set_node() to assign a firmware node directly to
> the struct device embedded in i2c_adapter. In order to hide the direct
> dereferencing: provide a dedicated interface that wraps the call to
> device_set_node() but takes the adapter as argument instead.

> [1] https://lore.kernel.org/all/20251106152330.11733-1-tzungbi@kernel.org/

Can be also made as a Link tag and some people use `b4 shazam`, that uses the
above test as a merge commit message where other tags are also being
recognized.

Have no objection and code looks good,
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

-- 
With Best Regards,
Andy Shevchenko