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.

This series addresses the usage of adap->dev in probe() callbacks where
drivers assign the parent device address and the associated OF-node
directly to the struct device embedded in i2c_adapter. We extend the
latter struct to accept the parent struct device and of_node directly
and make it assign it to its internal struct device inside
i2c_register_adapter(). For now just 12 patches but I'll keep on doing it
if these get accepted. Once these get upstream for v6.20/7.0, we'll be
able to also start converting i2c drivers outside of drivers/i2c/.

Link: https://lore.kernel.org/all/20251106152330.11733-1-tzungbi@kernel.org/
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>
---
Changes in v2:
- fix a NULL-pointer dereference in gpio-rcar (Geert)
- rebase on top of v7.0-rc1
- collect tags
- Link to v1: https://lore.kernel.org/r/20251223-i2c-adap-dev-config-v1-0-4829b1cf0834@oss.qualcomm.com

---
Bartosz Golaszewski (12):
      i2c: allow setting the parent device and OF node through the adapter struct
      i2c: qcom-geni: set device parent and of_node through the adapter struct
      i2c: bcm-kona: set device parent and of_node through the adapter struct
      i2c: keba: set device parent and of_node through the adapter struct
      i2c: omap: set device parent and of_node through the adapter struct
      i2c: rcar: set device parent and of_node through the adapter struct
      i2c: st: set device parent and of_node through the adapter struct
      i2c: mxs: set device parent and of_node through the adapter struct
      i2c: highlander: set device parent and of_node through the adapter struct
      i2c: gpio: set device parent and of_node through the adapter struct
      i2c: nomadik: set device parent and of_node through the adapter struct
      i2c: bcm2835: set device parent and of_node through the adapter struct

 drivers/i2c/busses/i2c-bcm-kona.c   | 4 ++--
 drivers/i2c/busses/i2c-bcm2835.c    | 4 ++--
 drivers/i2c/busses/i2c-gpio.c       | 2 +-
 drivers/i2c/busses/i2c-highlander.c | 2 +-
 drivers/i2c/busses/i2c-keba.c       | 2 +-
 drivers/i2c/busses/i2c-mxs.c        | 4 ++--
 drivers/i2c/busses/i2c-nomadik.c    | 4 ++--
 drivers/i2c/busses/i2c-omap.c       | 4 ++--
 drivers/i2c/busses/i2c-qcom-geni.c  | 4 ++--
 drivers/i2c/busses/i2c-rcar.c       | 6 +++---
 drivers/i2c/busses/i2c-st.c         | 4 ++--
 drivers/i2c/i2c-core-base.c         | 5 +++++
 include/linux/i2c.h                 | 4 ++++
 13 files changed, 29 insertions(+), 20 deletions(-)
---
base-commit: 6de23f81a5e08be8fbf5e8d7e9febc72a5b5f27f
change-id: 20251223-i2c-adap-dev-config-9403ab811795

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

On Mon, Feb 23, 2026 at 10:05:46AM +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.
> 
> This series addresses the usage of adap->dev in probe() callbacks where
> drivers assign the parent device address and the associated OF-node
> directly to the struct device embedded in i2c_adapter. We extend the
> latter struct to accept the parent struct device and of_node directly
> and make it assign it to its internal struct device inside
> i2c_register_adapter(). For now just 12 patches but I'll keep on doing it
> if these get accepted. Once these get upstream for v6.20/7.0, we'll be
> able to also start converting i2c drivers outside of drivers/i2c/.
> 
> Link: https://lore.kernel.org/all/20251106152330.11733-1-tzungbi@kernel.org/
> Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>

Also applied (and squashed) to for-current. Tags are retained BTW.

Hi Bartosz,

On Mon, 23 Feb 2026 at 10:06, Bartosz Golaszewski
<bartosz.golaszewski@oss.qualcomm.com> wrote:
> Changes in v2:
> - fix a NULL-pointer dereference in gpio-rcar (Geert)
                                      ^^^^
I guess that is bound to happen when a GPIO maintainer touches
the I2C subsystem? ;-)

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

On Mon, Feb 23, 2026 at 3:15 PM Geert Uytterhoeven <geert@linux-m68k.org> wrote:
>
> Hi Bartosz,
>
> On Mon, 23 Feb 2026 at 10:06, Bartosz Golaszewski
> <bartosz.golaszewski@oss.qualcomm.com> wrote:
> > Changes in v2:
> > - fix a NULL-pointer dereference in gpio-rcar (Geert)
>                                       ^^^^
> I guess that is bound to happen when a GPIO maintainer touches
> the I2C subsystem? ;-)
>

That or caffeine had not hit the system yet. :)

Bartosz