Stefan x Nilsson <stefan.x.nilsson@xxxxxxxx> writes: > Would a quirk on the affected modem be an better/acceptable solution? More acceptable but still needs justification. I'm probably missing something, but I don't see the described issue even if I assume a device crashing if it sees an ethernet header. The netdev and associated chardev are created when a USB device is connected. The device knows nothing about this. It answers a few USB descriptor requests and that's it. The host is free to switch the netdev to raw-ip framing before bringing it up. There is no need to send any frames with eithernet header. There is no relationship between framing and the QMI control channel. But the host is in complete control of any communication there too, so it wouldn't be a problam if there were. Bjørn
