From: Joakim Tjernlund <redacted>
Date: Mon, 30 Mar 2009 12:01:33 +0200

I don't know. But the question you should ask is: Does the networking
code promise this now and for the future? If not, you should
fix the driver not to relay on netif_queue_stopped() here.

Stop this nonsense talk.

If the driver can't be the one and only controller of the TX queue
state, everything would essentially explode.

The fact is, it does.