Re: [RFC PATCH 2/2] platform: arm64: qcom-hamoa-ec: test sensor and variant handling

[Birk Skyum]

Thanks for checking this.

The test client is bound to the I2C core's dummy driver by
i2c_new_dummy_device(). Unregistering it unbinds that driver and releases
the managed hwmon child before the adapter is deleted. I reran the arm64
KUnit image on the stated 7.3-rc2 base; all seven cases completed and the
guest rebooted normally, including cleanup after the sensor-probe test.

The dummy-driver binding is also documented here:
https://docs.kernel.org/driver-api/i2c.html#c.i2c_new_dummy_device

For the hwmon placement, I kept the small read-only sensor path in the
existing EC driver, as huawei-gaokun-ec does. Guenter and linux-hwmon are
already copied. Would an auxiliary driver under drivers/hwmon be preferred
here? I'm happy to split it if that is the preferred direction.