mt76: fix race condition in station removal

If there are still pending packets in the tx queue when removing a station,
it could possibly lead to a call to further attempts to pull packets from
the mac80211 tx queue after it has already been removed from the scheduling
list.
Prevent this from happening by calling synchronize_rcu after deleting the
wcid pointer before further cleaning up the tx queues.
To be extra careful, ensure that mtxq->list is always initialized properly.

Also drop the useless call to mt76x02_mac_wcid_setup, which only re-assigns
the bss index of the wcid entry, but does not help with the cleanup in any
way.

Signed-off-by: Felix Fietkau <[email protected]>

mt76x02_util.c

@@ -214,14 +214,15 @@ int mt76x02_sta_remove(struct ieee80211_hw *hw, struct ieee80211_vif *vif,
 	int idx = msta->wcid.idx;
 	int i;
 
+	rcu_assign_pointer(dev->mt76.wcid[idx], NULL);
+	synchronize_rcu();
+
 	mutex_lock(&dev->mt76.mutex);
 	mt76_tx_status_check(&dev->mt76, &msta->wcid, true);
-	rcu_assign_pointer(dev->mt76.wcid[idx], NULL);
 	for (i = 0; i < ARRAY_SIZE(sta->txq); i++)
 		mt76_txq_remove(&dev->mt76, sta->txq[i]);
 	mt76x02_mac_wcid_set_drop(dev, idx, true);
 	mt76_wcid_free(dev->mt76.wcid_mask, idx);
-	mt76x02_mac_wcid_setup(dev, idx, 0, NULL);
 	mutex_unlock(&dev->mt76.mutex);
 
 	return 0;

tx.c

@@ -590,7 +590,7 @@ void mt76_txq_remove(struct mt76_dev *dev, struct ieee80211_txq *txq)
 
 	spin_lock_bh(&hwq->lock);
 	if (!list_empty(&mtxq->list))
-		list_del(&mtxq->list);
+		list_del_init(&mtxq->list);
 	spin_unlock_bh(&hwq->lock);
 
 	while ((skb = skb_dequeue(&mtxq->retry_q)) != NULL)