ieee80211_tx...hedule_start(9)

πŸ“¦ version 7.3.1, in nixpkgs unstable

πŸ’– NAME πŸ’–

ieee80211_txq_schedule_start β€” start a whole new scheduling round for them TXQs, bestie ✨

So real talk: this is a page from the Kernel API Manual (yes, that's the manual name, we're being official), and the header up top does that cursed truncation thing where it prints ieee80211_tx...hedule_start(9) β€” that's literally just the terminal eating the middle of the name, the actual function is ieee80211_txq_schedule_start. iykyk πŸ’…


✨ SYNOPSIS ✨

void ieee80211_txq_schedule_start (struct ieee80211_hw *hw , u8 ac );

πŸŽ€ ARGUMENTS πŸŽ€

  • hw β€” a pointer, and specifically the one you got from ieee80211_alloc_hw. don't be weird about it, use the real one 🫢
  • ac β€” the AC number you're acquiring locks for. yeah. that AC number. πŸ¦‹

πŸ¦‹ DESCRIPTION πŸ¦‹

okay so the whole point: you call this before ieee80211_next_txq or ieee80211_return_txq. Like, it's the opener, they're the rest of the routine. πŸŽ€

🚨 the one rule, no cap: the driver must not call multiple TXQ scheduling rounds concurrently. not two, not "just a little overlapping", none. one round at a time, periodt. πŸ’€

🌸 SEE ALSO 🌸

okay this next part is genuinely iconic β€” it's the entire mac80211 extended universe, every single one of them. scroll responsibly, she's long πŸ‘‘


okay that was a journey, we made it, drink some water πŸ“

cute little sign-off moment: this page lives in ./include/net, it's dated October 2026, and yeah the footer also insists on calling it ieee80211_tx...hedule_start(9) because it refuses to be normal πŸ’€πŸ¦‹ slay, go schedule your TXQs responsibly, bestie πŸ’–βœ¨

πŸ“¦ where this page comes from