The case where we reboot with flight mode being enable is still valid. I did some further analysis on this as I stumbled upon this again during my current BlueZ 5 work and the underlaying problems goes down to the kernel.
The bluetooth driver we have on krillin isn't implemented as a module but is initialized by a subsystem it uses. It now turns out that this is bound to the WiFi driver being initialized too. If WiFi doesn't get enabled at boot time the Bluetooth driver can proceed providing a HCI device and we end up without any Bluetooth functionality.
Needs further investigation on the kernel side what we can do to work around this.
The case where we reboot with flight mode being enable is still valid. I did some further analysis on this as I stumbled upon this again during my current BlueZ 5 work and the underlaying problems goes down to the kernel.
The bluetooth driver we have on krillin isn't implemented as a module but is initialized by a subsystem it uses. It now turns out that this is bound to the WiFi driver being initialized too. If WiFi doesn't get enabled at boot time the Bluetooth driver can proceed providing a HCI device and we end up without any Bluetooth functionality.
Needs further investigation on the kernel side what we can do to work around this.