At the moment, the RT5033 MFD and battery driver suggest that the battery driver should probe as a sub-device of the MFD driver. However, this does not make any sense since the fuel gauge part of RT5033 has its own I2C device and interrupt line. It was also documented as separate I2C device in the original device tree bindings [1] (that were never finished up and merged) but for some reason the code does not match the documentation (and reality). :/ Given other fairly critical mistakes like setting the wrong bits in the regulator driver (see [2]), unfortunately I get the feeling that none of the RT5033 drivers were ever tested properly. :( This patch sets adds a proper of_match_table to rt5033-battery and removes the rt5033-battery sub-device from the MFD driver. There is no compile/runtime dependency of the power supply / MFD patch so they can just be applied separately through the power supply / MFD tree. With these changes, rt5033-battery seems to work fine on the Samsung Galaxy A5 (2015) at least (it reports a reasonable battery percentage). [1]: https://lore.kernel.org/linux-pm/1425864191-4121-3-git-send-email-beomho.seo@xxxxxxxxxxx/ [2]: https://lore.kernel.org/lkml/20201110130047.8097-1-michael.srba@xxxxxxxxx/ Changes in v2: Fix stupid typo in second patch :( v1: Honestly, not worth looking at :) Stephan Gerhold (3): dt-bindings: power: supply: Add DT schema for richtek,rt5033-battery power: supply: rt5033_battery: Fix device tree enumeration mfd: rt5033: Drop rt5033-battery sub-device .../power/supply/richtek,rt5033-battery.yaml | 54 +++++++++++++++++++ drivers/mfd/rt5033.c | 3 -- drivers/power/supply/Kconfig | 3 +- drivers/power/supply/rt5033_battery.c | 7 +++ 4 files changed, 63 insertions(+), 4 deletions(-) create mode 100644 Documentation/devicetree/bindings/power/supply/richtek,rt5033-battery.yaml -- 2.31.1
