On 2018-10-01 23:58, Rajkumar Manoharan wrote:
>> Great! I'll fold in the rest, test it with ath9k and submit as a 
>> proper patch :)
>> 
> Toke,
> 
> I noticed a race condition b/w sta cleanup and kick_airtime tasklet. 
> How do you
> plan to exit kick_airtime gracefully during sta_cleanup?
> 
If kick_airtime tasklet is only used for adjusting deficit for all 
throttled txq,
then below rcu lock issue is not observed. I am testing with 50 clients 
and the
crash happens only during sta cleanup. Releasing active_txq_lock from 
tasklet is
yielding handle to txq_purge(). I am thinking of get rid of tasklet and 
handle
adjustment directly in API.


diff --git a/net/mac80211/util.c b/net/mac80211/util.c
index 0bb590928dd0..277dbf8e0a4b 100644
--- a/net/mac80211/util.c
+++ b/net/mac80211/util.c
@@ -261,14 +261,7 @@ static void __ieee80211_kick_airtime(struct 
ieee80211_local *local, int ac)

                 if (sta->airtime[ac].deficit >= 0) {
                         seen_eligible = true;
-
-                       if 
(!test_and_clear_bit(IEEE80211_TXQ_AIRTIME_THROTTLE,
-                                               &txqi->flags))
-                               continue;
-
-                       spin_unlock_bh(&local->active_txq_lock[ac]);
-                       drv_wake_tx_queue(local, txqi);
-                       spin_lock_bh(&local->active_txq_lock[ac]);
+                       clear_bit(IEEE80211_TXQ_AIRTIME_THROTTLE, 
&txqi->flags);

-Rajkumar