Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next

Pull networking updates from David Miller:

 1) BPF debugger and asm tool by Daniel Borkmann.

 2) Speed up create/bind in AF_PACKET, also from Daniel Borkmann.

 3) Correct reciprocal_divide and update users, from Hannes Frederic
    Sowa and Daniel Borkmann.

 4) Currently we only have a "set" operation for the hw timestamp socket
    ioctl, add a "get" operation to match.  From Ben Hutchings.

 5) Add better trace events for debugging driver datapath problems, also
    from Ben Hutchings.

 6) Implement auto corking in TCP, from Eric Dumazet.  Basically, if we
    have a small send and a previous packet is already in the qdisc or
    device queue, defer until TX completion or we get more data.

 7) Allow userspace to manage ipv6 temporary addresses, from Jiri Pirko.

 8) Add a qdisc bypass option for AF_PACKET sockets, from Daniel
    Borkmann.

 9) Share IP header compression code between Bluetooth and IEEE802154
    layers, from Jukka Rissanen.

10) Fix ipv6 router reachability probing, from Jiri Benc.

11) Allow packets to be captured on macvtap devices, from Vlad Yasevich.

12) Support tunneling in GRO layer, from Jerry Chu.

13) Allow bonding to be configured fully using netlink, from Scott
    Feldman.

14) Allow AF_PACKET users to obtain the VLAN TPID, just like they can
    already get the TCI.  From Atzm Watanabe.

15) New "Heavy Hitter" qdisc, from Terry Lam.

16) Significantly improve the IPSEC support in pktgen, from Fan Du.

17) Allow ipv4 tunnels to cache routes, just like sockets.  From Tom
    Herbert.

18) Add Proportional Integral Enhanced packet scheduler, from Vijay
    Subramanian.

19) Allow openvswitch to mmap'd netlink, from Thomas Graf.

20) Key TCP metrics blobs also by source address, not just destination
    address.  From Christoph Paasch.

21) Support 10G in generic phylib.  From Andy Fleming.

22) Try to short-circuit GRO flow compares using device provided RX
    hash, if provided.  From Tom Herbert.

The wireless and netfilter folks have been busy little bees too.

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next: (2064 commits)
  net/cxgb4: Fix referencing freed adapter
  ipv6: reallocate addrconf router for ipv6 address when lo device up
  fib_frontend: fix possible NULL pointer dereference
  rtnetlink: remove IFLA_BOND_SLAVE definition
  rtnetlink: remove check for fill_slave_info in rtnl_have_link_slave_info
  qlcnic: update version to 5.3.55
  qlcnic: Enhance logic to calculate msix vectors.
  qlcnic: Refactor interrupt coalescing code for all adapters.
  qlcnic: Update poll controller code path
  qlcnic: Interrupt code cleanup
  qlcnic: Enhance Tx timeout debugging.
  qlcnic: Use bool for rx_mac_learn.
  bonding: fix u64 division
  rtnetlink: add missing IFLA_BOND_AD_INFO_UNSPEC
  sfc: Use the correct maximum TX DMA ring size for SFC9100
  Add Shradha Shah as the sfc driver maintainer.
  net/vxlan: Share RX skb de-marking and checksum checks with ovs
  tulip: cleanup by using ARRAY_SIZE()
  ip_tunnel: clear IPCB in ip_tunnel_xmit() in case dst_link_failure() is called
  net/cxgb4: Don't retrieve stats during recovery
  ...

1 files changed, 219 insertions, 127 deletions

diff --git a/net/mac80211/sta_info.c b/net/mac80211/sta_info.c
index 1eb66e26e49d..decd30c1e290 100644
--- a/net/mac80211/sta_info.c
+++ b/net/mac80211/sta_info.c
@@ -99,23 +99,6 @@ static void cleanup_single_sta(struct sta_info *sta)
 	struct ieee80211_local *local = sdata->local;
 	struct ps_data *ps;
 
-	/*
-	 * At this point, when being called as call_rcu callback,
-	 * neither mac80211 nor the driver can reference this
-	 * sta struct any more except by still existing timers
-	 * associated with this station that we clean up below.
-	 *
-	 * Note though that this still uses the sdata and even
-	 * calls the driver in AP and mesh mode, so interfaces
-	 * of those types mush use call sta_info_flush_cleanup()
-	 * (typically via sta_info_flush()) before deconfiguring
-	 * the driver.
-	 *
-	 * In station mode, nothing happens here so it doesn't
-	 * have to (and doesn't) do that, this is intentional to
-	 * speed up roaming.
-	 */
-
 	if (test_sta_flag(sta, WLAN_STA_PS_STA)) {
 		if (sta->sdata->vif.type == NL80211_IFTYPE_AP ||
 		    sta->sdata->vif.type == NL80211_IFTYPE_AP_VLAN)
@@ -160,37 +143,6 @@ static void cleanup_single_sta(struct sta_info *sta)
 	sta_info_free(local, sta);
 }
 
-void ieee80211_cleanup_sdata_stas(struct ieee80211_sub_if_data *sdata)
-{
-	struct sta_info *sta;
-
-	spin_lock_bh(&sdata->cleanup_stations_lock);
-	while (!list_empty(&sdata->cleanup_stations)) {
-		sta = list_first_entry(&sdata->cleanup_stations,
-				       struct sta_info, list);
-		list_del(&sta->list);
-		spin_unlock_bh(&sdata->cleanup_stations_lock);
-
-		cleanup_single_sta(sta);
-
-		spin_lock_bh(&sdata->cleanup_stations_lock);
-	}
-
-	spin_unlock_bh(&sdata->cleanup_stations_lock);
-}
-
-static void free_sta_rcu(struct rcu_head *h)
-{
-	struct sta_info *sta = container_of(h, struct sta_info, rcu_head);
-	struct ieee80211_sub_if_data *sdata = sta->sdata;
-
-	spin_lock(&sdata->cleanup_stations_lock);
-	list_add_tail(&sta->list, &sdata->cleanup_stations);
-	spin_unlock(&sdata->cleanup_stations_lock);
-
-	ieee80211_queue_work(&sdata->local->hw, &sdata->cleanup_stations_wk);
-}
-
 /* protected by RCU */
 struct sta_info *sta_info_get(struct ieee80211_sub_if_data *sdata,
 			      const u8 *addr)
@@ -266,9 +218,17 @@ struct sta_info *sta_info_get_by_idx(struct ieee80211_sub_if_data *sdata,
  */
 void sta_info_free(struct ieee80211_local *local, struct sta_info *sta)
 {
+	int i;
+
 	if (sta->rate_ctrl)
 		rate_control_free_sta(sta);
 
+	if (sta->tx_lat) {
+		for (i = 0; i < IEEE80211_NUM_TIDS; i++)
+			kfree(sta->tx_lat[i].bins);
+		kfree(sta->tx_lat);
+	}
+
 	sta_dbg(sta->sdata, "Destroyed STA %pM\n", sta->sta.addr);
 
 	kfree(sta);
@@ -333,12 +293,42 @@ struct sta_info *sta_info_alloc(struct ieee80211_sub_if_data *sdata,
 	struct ieee80211_local *local = sdata->local;
 	struct sta_info *sta;
 	struct timespec uptime;
+	struct ieee80211_tx_latency_bin_ranges *tx_latency;
 	int i;
 
 	sta = kzalloc(sizeof(*sta) + local->hw.sta_data_size, gfp);
 	if (!sta)
 		return NULL;
 
+	rcu_read_lock();
+	tx_latency = rcu_dereference(local->tx_latency);
+	/* init stations Tx latency statistics && TID bins */
+	if (tx_latency) {
+		sta->tx_lat = kzalloc(IEEE80211_NUM_TIDS *
+				      sizeof(struct ieee80211_tx_latency_stat),
+				      GFP_ATOMIC);
+		if (!sta->tx_lat) {
+			rcu_read_unlock();
+			goto free;
+		}
+
+		if (tx_latency->n_ranges) {
+			for (i = 0; i < IEEE80211_NUM_TIDS; i++) {
+				/* size of bins is size of the ranges +1 */
+				sta->tx_lat[i].bin_count =
+					tx_latency->n_ranges + 1;
+				sta->tx_lat[i].bins =
+					kcalloc(sta->tx_lat[i].bin_count,
+						sizeof(u32), GFP_ATOMIC);
+				if (!sta->tx_lat[i].bins) {
+					rcu_read_unlock();
+					goto free;
+				}
+			}
+		}
+	}
+	rcu_read_unlock();
+
 	spin_lock_init(&sta->lock);
 	INIT_WORK(&sta->drv_unblock_wk, sta_unblock);
 	INIT_WORK(&sta->ampdu_mlme.work, ieee80211_ba_session_work);
@@ -363,10 +353,8 @@ struct sta_info *sta_info_alloc(struct ieee80211_sub_if_data *sdata,
 	for (i = 0; i < ARRAY_SIZE(sta->chain_signal_avg); i++)
 		ewma_init(&sta->chain_signal_avg[i], 1024, 8);
 
-	if (sta_prepare_rate_control(local, sta, gfp)) {
-		kfree(sta);
-		return NULL;
-	}
+	if (sta_prepare_rate_control(local, sta, gfp))
+		goto free;
 
 	for (i = 0; i < IEEE80211_NUM_TIDS; i++) {
 		/*
@@ -411,8 +399,16 @@ struct sta_info *sta_info_alloc(struct ieee80211_sub_if_data *sdata,
 	}
 
 	sta_dbg(sdata, "Allocated STA %pM\n", sta->sta.addr);
-
 	return sta;
+
+free:
+	if (sta->tx_lat) {
+		for (i = 0; i < IEEE80211_NUM_TIDS; i++)
+			kfree(sta->tx_lat[i].bins);
+		kfree(sta->tx_lat);
+	}
+	kfree(sta);
+	return NULL;
 }
 
 static int sta_info_insert_check(struct sta_info *sta)
@@ -507,6 +503,7 @@ static int sta_info_insert_finish(struct sta_info *sta) __acquires(RCU)
 
 	set_sta_flag(sta, WLAN_STA_INSERTED);
 
+	ieee80211_recalc_min_chandef(sdata);
 	ieee80211_sta_debugfs_add(sta);
 	rate_control_add_sta_debugfs(sta);
 
@@ -630,8 +627,8 @@ void sta_info_recalc_tim(struct sta_info *sta)
 #ifdef CONFIG_MAC80211_MESH
 	} else if (ieee80211_vif_is_mesh(&sta->sdata->vif)) {
 		ps = &sta->sdata->u.mesh.ps;
-		/* TIM map only for PLID <= IEEE80211_MAX_AID */
-		id = le16_to_cpu(sta->plid) % IEEE80211_MAX_AID;
+		/* TIM map only for 1 <= PLID <= IEEE80211_MAX_AID */
+		id = sta->plid % (IEEE80211_MAX_AID + 1);
 #endif
 	} else {
 		return;
@@ -807,7 +804,7 @@ static bool sta_info_cleanup_expire_buffered(struct ieee80211_local *local,
 	return have_buffered;
 }
 
-int __must_check __sta_info_destroy(struct sta_info *sta)
+static int __must_check __sta_info_destroy_part1(struct sta_info *sta)
 {
 	struct ieee80211_local *local;
 	struct ieee80211_sub_if_data *sdata;
@@ -833,12 +830,35 @@ int __must_check __sta_info_destroy(struct sta_info *sta)
 	ieee80211_sta_tear_down_BA_sessions(sta, AGG_STOP_DESTROY_STA);
 
 	ret = sta_info_hash_del(local, sta);
-	if (ret)
+	if (WARN_ON(ret))
 		return ret;
 
 	list_del_rcu(&sta->list);
 
-	/* this always calls synchronize_net() */
+	drv_sta_pre_rcu_remove(local, sta->sdata, sta);
+
+	if (sdata->vif.type == NL80211_IFTYPE_AP_VLAN &&
+	    rcu_access_pointer(sdata->u.vlan.sta) == sta)
+		RCU_INIT_POINTER(sdata->u.vlan.sta, NULL);
+
+	return 0;
+}
+
+static void __sta_info_destroy_part2(struct sta_info *sta)
+{
+	struct ieee80211_local *local = sta->local;
+	struct ieee80211_sub_if_data *sdata = sta->sdata;
+	int ret;
+
+	/*
+	 * NOTE: This assumes at least synchronize_net() was done
+	 *	 after _part1 and before _part2!
+	 */
+
+	might_sleep();
+	lockdep_assert_held(&local->sta_mtx);
+
+	/* now keys can no longer be reached */
 	ieee80211_free_sta_keys(local, sta);
 
 	sta->dead = true;
@@ -846,9 +866,6 @@ int __must_check __sta_info_destroy(struct sta_info *sta)
 	local->num_sta--;
 	local->sta_generation++;
 
-	if (sdata->vif.type == NL80211_IFTYPE_AP_VLAN)
-		RCU_INIT_POINTER(sdata->u.vlan.sta, NULL);
-
 	while (sta->sta_state > IEEE80211_STA_NONE) {
 		ret = sta_info_move_state(sta, sta->sta_state - 1);
 		if (ret) {
@@ -869,8 +886,21 @@ int __must_check __sta_info_destroy(struct sta_info *sta)
 
 	rate_control_remove_sta_debugfs(sta);
 	ieee80211_sta_debugfs_remove(sta);
+	ieee80211_recalc_min_chandef(sdata);
+
+	cleanup_single_sta(sta);
+}
 
-	call_rcu(&sta->rcu_head, free_sta_rcu);
+int __must_check __sta_info_destroy(struct sta_info *sta)
+{
+	int err = __sta_info_destroy_part1(sta);
+
+	if (err)
+		return err;
+
+	synchronize_net();
+
+	__sta_info_destroy_part2(sta);
 
 	return 0;
 }
@@ -940,32 +970,38 @@ void sta_info_stop(struct ieee80211_local *local)
 }
 
 
-int sta_info_flush_defer(struct ieee80211_sub_if_data *sdata)
+int __sta_info_flush(struct ieee80211_sub_if_data *sdata, bool vlans)
 {
 	struct ieee80211_local *local = sdata->local;
 	struct sta_info *sta, *tmp;
+	LIST_HEAD(free_list);
 	int ret = 0;
 
 	might_sleep();
 
+	WARN_ON(vlans && sdata->vif.type != NL80211_IFTYPE_AP);
+	WARN_ON(vlans && !sdata->bss);
+
 	mutex_lock(&local->sta_mtx);
 	list_for_each_entry_safe(sta, tmp, &local->sta_list, list) {
-		if (sdata == sta->sdata) {
-			WARN_ON(__sta_info_destroy(sta));
+		if (sdata == sta->sdata ||
+		    (vlans && sdata->bss == sta->sdata->bss)) {
+			if (!WARN_ON(__sta_info_destroy_part1(sta)))
+				list_add(&sta->free_list, &free_list);
 			ret++;
 		}
 	}
+
+	if (!list_empty(&free_list)) {
+		synchronize_net();
+		list_for_each_entry_safe(sta, tmp, &free_list, free_list)
+			__sta_info_destroy_part2(sta);
+	}
 	mutex_unlock(&local->sta_mtx);
 
 	return ret;
 }
 
-void sta_info_flush_cleanup(struct ieee80211_sub_if_data *sdata)
-{
-	ieee80211_cleanup_sdata_stas(sdata);
-	cancel_work_sync(&sdata->cleanup_stations_wk);
-}
-
 void ieee80211_sta_expire(struct ieee80211_sub_if_data *sdata,
 			  unsigned long exp_time)
 {
@@ -1117,7 +1153,8 @@ void ieee80211_sta_ps_deliver_wakeup(struct sta_info *sta)
 
 static void ieee80211_send_null_response(struct ieee80211_sub_if_data *sdata,
 					 struct sta_info *sta, int tid,
-					 enum ieee80211_frame_release_type reason)
+					 enum ieee80211_frame_release_type reason,
+					 bool call_driver)
 {
 	struct ieee80211_local *local = sdata->local;
 	struct ieee80211_qos_hdr *nullfunc;
@@ -1175,7 +1212,9 @@ static void ieee80211_send_null_response(struct ieee80211_sub_if_data *sdata,
 		       IEEE80211_TX_STATUS_EOSP |
 		       IEEE80211_TX_CTL_REQ_TX_STATUS;
 
-	drv_allow_buffered_frames(local, sta, BIT(tid), 1, reason, false);
+	if (call_driver)
+		drv_allow_buffered_frames(local, sta, BIT(tid), 1,
+					  reason, false);
 
 	skb->dev = sdata->dev;
 
@@ -1191,6 +1230,17 @@ static void ieee80211_send_null_response(struct ieee80211_sub_if_data *sdata,
 	rcu_read_unlock();
 }
 
+static int find_highest_prio_tid(unsigned long tids)
+{
+	/* lower 3 TIDs aren't ordered perfectly */
+	if (tids & 0xF8)
+		return fls(tids) - 1;
+	/* TID 0 is BE just like TID 3 */
+	if (tids & BIT(0))
+		return 0;
+	return fls(tids) - 1;
+}
+
 static void
 ieee80211_sta_ps_deliver_response(struct sta_info *sta,
 				  int n_frames, u8 ignored_acs,
@@ -1198,7 +1248,6 @@ ieee80211_sta_ps_deliver_response(struct sta_info *sta,
 {
 	struct ieee80211_sub_if_data *sdata = sta->sdata;
 	struct ieee80211_local *local = sdata->local;
-	bool found = false;
 	bool more_data = false;
 	int ac;
 	unsigned long driver_release_tids = 0;
@@ -1209,9 +1258,7 @@ ieee80211_sta_ps_deliver_response(struct sta_info *sta,
 
 	__skb_queue_head_init(&frames);
 
-	/*
-	 * Get response frame(s) and more data bit for it.
-	 */
+	/* Get response frame(s) and more data bit for the last one. */
 	for (ac = 0; ac < IEEE80211_NUM_ACS; ac++) {
 		unsigned long tids;
 
@@ -1220,43 +1267,48 @@ ieee80211_sta_ps_deliver_response(struct sta_info *sta,
 
 		tids = ieee80211_tids_for_ac(ac);
 
-		if (!found) {
-			driver_release_tids = sta->driver_buffered_tids & tids;
-			if (driver_release_tids) {
-				found = true;
-			} else {
-				struct sk_buff *skb;
-
-				while (n_frames > 0) {
-					skb = skb_dequeue(&sta->tx_filtered[ac]);
-					if (!skb) {
-						skb = skb_dequeue(
-							&sta->ps_tx_buf[ac]);
-						if (skb)
-							local->total_ps_buffered--;
-					}
-					if (!skb)
-						break;
-					n_frames--;
-					found = true;
-					__skb_queue_tail(&frames, skb);
-				}
-			}
+		/* if we already have frames from software, then we can't also
+		 * release from hardware queues
+		 */
+		if (skb_queue_empty(&frames))
+			driver_release_tids |= sta->driver_buffered_tids & tids;
 
-			/*
-			 * If the driver has data on more than one TID then
+		if (driver_release_tids) {
+			/* If the driver has data on more than one TID then
 			 * certainly there's more data if we release just a
-			 * single frame now (from a single TID).
+			 * single frame now (from a single TID). This will
+			 * only happen for PS-Poll.
 			 */
 			if (reason == IEEE80211_FRAME_RELEASE_PSPOLL &&
 			    hweight16(driver_release_tids) > 1) {
 				more_data = true;
 				driver_release_tids =
-					BIT(ffs(driver_release_tids) - 1);
+					BIT(find_highest_prio_tid(
+						driver_release_tids));
 				break;
 			}
+		} else {
+			struct sk_buff *skb;
+
+			while (n_frames > 0) {
+				skb = skb_dequeue(&sta->tx_filtered[ac]);
+				if (!skb) {
+					skb = skb_dequeue(
+						&sta->ps_tx_buf[ac]);
+					if (skb)
+						local->total_ps_buffered--;
+				}
+				if (!skb)
+					break;
+				n_frames--;
+				__skb_queue_tail(&frames, skb);
+			}
 		}
 
+		/* If we have more frames buffered on this AC, then set the
+		 * more-data bit and abort the loop since we can't send more
+		 * data from other ACs before the buffered frames from this.
+		 */
 		if (!skb_queue_empty(&sta->tx_filtered[ac]) ||
 		    !skb_queue_empty(&sta->ps_tx_buf[ac])) {
 			more_data = true;
@@ -1264,7 +1316,7 @@ ieee80211_sta_ps_deliver_response(struct sta_info *sta,
 		}
 	}
 
-	if (!found) {
+	if (skb_queue_empty(&frames) && !driver_release_tids) {
 		int tid;
 
 		/*
@@ -1285,15 +1337,13 @@ ieee80211_sta_ps_deliver_response(struct sta_info *sta,
 		/* This will evaluate to 1, 3, 5 or 7. */
 		tid = 7 - ((ffs(~ignored_acs) - 1) << 1);
 
-		ieee80211_send_null_response(sdata, sta, tid, reason);
-		return;
-	}
-
-	if (!driver_release_tids) {
+		ieee80211_send_null_response(sdata, sta, tid, reason, true);
+	} else if (!driver_release_tids) {
 		struct sk_buff_head pending;
 		struct sk_buff *skb;
 		int num = 0;
 		u16 tids = 0;
+		bool need_null = false;
 
 		skb_queue_head_init(&pending);
 
@@ -1327,22 +1377,57 @@ ieee80211_sta_ps_deliver_response(struct sta_info *sta,
 			    ieee80211_is_qos_nullfunc(hdr->frame_control))
 				qoshdr = ieee80211_get_qos_ctl(hdr);
 
-			/* end service period after last frame */
-			if (skb_queue_empty(&frames)) {
-				if (reason == IEEE80211_FRAME_RELEASE_UAPSD &&
-				    qoshdr)
-					*qoshdr |= IEEE80211_QOS_CTL_EOSP;
+			tids |= BIT(skb->priority);
+
+			__skb_queue_tail(&pending, skb);
+
+			/* end service period after last frame or add one */
+			if (!skb_queue_empty(&frames))
+				continue;
 
+			if (reason != IEEE80211_FRAME_RELEASE_UAPSD) {
+				/* for PS-Poll, there's only one frame */
 				info->flags |= IEEE80211_TX_STATUS_EOSP |
 					       IEEE80211_TX_CTL_REQ_TX_STATUS;
+				break;
 			}
 
-			if (qoshdr)
-				tids |= BIT(*qoshdr & IEEE80211_QOS_CTL_TID_MASK);
-			else
-				tids |= BIT(0);
+			/* For uAPSD, things are a bit more complicated. If the
+			 * last frame has a QoS header (i.e. is a QoS-data or
+			 * QoS-nulldata frame) then just set the EOSP bit there
+			 * and be done.
+			 * If the frame doesn't have a QoS header (which means
+			 * it should be a bufferable MMPDU) then we can't set
+			 * the EOSP bit in the QoS header; add a QoS-nulldata
+			 * frame to the list to send it after the MMPDU.
+			 *
+			 * Note that this code is only in the mac80211-release
+			 * code path, we assume that the driver will not buffer
+			 * anything but QoS-data frames, or if it does, will
+			 * create the QoS-nulldata frame by itself if needed.
+			 *
+			 * Cf. 802.11-2012 10.2.1.10 (c).
+			 */
+			if (qoshdr) {
+				*qoshdr |= IEEE80211_QOS_CTL_EOSP;
 
-			__skb_queue_tail(&pending, skb);
+				info->flags |= IEEE80211_TX_STATUS_EOSP |
+					       IEEE80211_TX_CTL_REQ_TX_STATUS;
+			} else {
+				/* The standard isn't completely clear on this
+				 * as it says the more-data bit should be set
+				 * if there are more BUs. The QoS-Null frame
+				 * we're about to send isn't buffered yet, we
+				 * only create it below, but let's pretend it
+				 * was buffered just in case some clients only
+				 * expect more-data=0 when eosp=1.
+				 */
+				hdr->frame_control |=
+					cpu_to_le16(IEEE80211_FCTL_MOREDATA);
+				need_null = true;
+				num++;
+			}
+			break;
 		}
 
 		drv_allow_buffered_frames(local, sta, tids, num,
@@ -1350,17 +1435,22 @@ ieee80211_sta_ps_deliver_response(struct sta_info *sta,
 
 		ieee80211_add_pending_skbs(local, &pending);
 
+		if (need_null)
+			ieee80211_send_null_response(
+				sdata, sta, find_highest_prio_tid(tids),
+				reason, false);
+
 		sta_info_recalc_tim(sta);
 	} else {
 		/*
 		 * We need to release a frame that is buffered somewhere in the
 		 * driver ... it'll have to handle that.
-		 * Note that, as per the comment above, it'll also have to see
-		 * if there is more than just one frame on the specific TID that
-		 * we're releasing from, and it needs to set the more-data bit
-		 * accordingly if we tell it that there's no more data. If we do
-		 * tell it there's more data, then of course the more-data bit
-		 * needs to be set anyway.
+		 * Note that the driver also has to check the number of frames
+		 * on the TIDs we're releasing from - if there are more than
+		 * n_frames it has to set the more-data bit (if we didn't ask
+		 * it to set it anyway due to other buffered frames); if there
+		 * are fewer than n_frames it has to make sure to adjust that
+		 * to allow the service period to end properly.
 		 */
 		drv_release_buffered_frames(local, sta, driver_release_tids,
 					    n_frames, reason, more_data);
@@ -1368,9 +1458,9 @@ ieee80211_sta_ps_deliver_response(struct sta_info *sta,
 		/*
 		 * Note that we don't recalculate the TIM bit here as it would
 		 * most likely have no effect at all unless the driver told us
-		 * that the TID became empty before returning here from the
+		 * that the TID(s) became empty before returning here from the
 		 * release function.
-		 * Either way, however, when the driver tells us that the TID
+		 * Either way, however, when the driver tells us that the TID(s)
 		 * became empty we'll do the TIM recalculation.
 		 */
 	}
@@ -1459,6 +1549,8 @@ void ieee80211_sta_set_buffered(struct ieee80211_sta *pubsta,
 	if (WARN_ON(tid >= IEEE80211_NUM_TIDS))
 		return;
 
+	trace_api_sta_set_buffered(sta->local, pubsta, tid, buffered);
+
 	if (buffered)
 		set_bit(tid, &sta->driver_buffered_tids);
 	else