Hi Greg, On 10/18/2017 03:33 PM, Greg Kroah-Hartman wrote:
The SPI core should have some internal housekeeping to do there as well, right? Is this the only bus that gets this "wrong"?
hmm, i checked i2c code, it has the same issue: void i2c_unregister_device(struct i2c_client *client) { if (client->dev.of_node) of_node_clear_flag(client->dev.of_node, OF_POPULATED); if (ACPI_COMPANION(&client->dev)) acpi_device_clear_enumerated(ACPI_COMPANION(&client->dev)); device_unregister(&client->dev); } not sure about others, maybe mfd-core too...
thanks, greg k-h