PS: I should really go and read the specs to see if we can steer v4 and v6 independently but my memory says we cannot.

Functionally accepted.

In D54794#1252564, @adrian wrote:

Oh, there's no callers of ieee80211_ampdu_request() outside of net80211? I was worried that the existing AMPDU implementing drivers would hiccup with the return value changing!

Rather than leaving the vlog as a reminder, deal with the TODO comment
in net80211_ic_printf() as well and remove the vlog comment from there.

Update treturn as suggested by @emaste.

I don't get the correlation between your commit message and these changes. Is there an actual change (due to casting) in this change? Or is there a problem "somewhere" but this change does not necessarily address that?

Your first line of commit message is RX repetitive; I assume one should be tx. Should be fixed before commit.