From 0bb7ed426373cebf406ff3ae7b6d2344f2f4364c Mon Sep 17 00:00:00 2001
From: Ben Greear <greearb@candelatech.com>
Date: Fri, 13 May 2016 11:29:33 -0700
Subject: mac80211: add vht cap decode to debugfs

This makes it a lot easier to understand the capabilities used
by the station:

VHT supported
cap: 0x300819b2
		MAX-MPDU-11454
		80Mhz
		RXLDPC
		SHORT-GI-80
		TXSTBC
		RXSTBC_1
		SU-BEAMFORMER-CAPABLE
		SU-BEAMFORMEE-CAPABLE
		BEAMFORMEE-STS: 0x0
		SOUNDING-DIMENSIONS: 0x0
		MU-BEAMFORMER-CAPABLE
		MPDU-LENGTH-EXPONENT: 0x0
		LINK-ADAPTATION-VHT-MRQ-MFB: 0x0
		RX-ANTENNA-PATTERN
		TX-ANTENNA-PATTERN
RX MCS: fffe
TX MCS: fffe

Signed-off-by: Ben Greear <greearb@candelatech.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
---
 net/mac80211/debugfs_sta.c | 78 ++++++++++++++++++++++++++++++++++++++++++++--
 1 file changed, 76 insertions(+), 2 deletions(-)

(limited to 'net/mac80211')

diff --git a/net/mac80211/debugfs_sta.c b/net/mac80211/debugfs_sta.c
index 33dfcbc2bf9c..fd334133ff45 100644
--- a/net/mac80211/debugfs_sta.c
+++ b/net/mac80211/debugfs_sta.c
@@ -328,14 +328,88 @@ STA_OPS(ht_capa);
 static ssize_t sta_vht_capa_read(struct file *file, char __user *userbuf,
 				 size_t count, loff_t *ppos)
 {
-	char buf[128], *p = buf;
+	char buf[512], *p = buf;
 	struct sta_info *sta = file->private_data;
 	struct ieee80211_sta_vht_cap *vhtc = &sta->sta.vht_cap;
 
 	p += scnprintf(p, sizeof(buf) + buf - p, "VHT %ssupported\n",
 			vhtc->vht_supported ? "" : "not ");
 	if (vhtc->vht_supported) {
-		p += scnprintf(p, sizeof(buf)+buf-p, "cap: %#.8x\n", vhtc->cap);
+		p += scnprintf(p, sizeof(buf) + buf - p, "cap: %#.8x\n",
+			       vhtc->cap);
+#define PFLAG(a, b)							\
+		do {							\
+			if (vhtc->cap & IEEE80211_VHT_CAP_ ## a)	\
+				p += scnprintf(p, sizeof(buf) + buf - p, \
+					       "\t\t%s\n", b);		\
+		} while (0)
+
+		switch (vhtc->cap & 0x3) {
+		case IEEE80211_VHT_CAP_MAX_MPDU_LENGTH_3895:
+			p += scnprintf(p, sizeof(buf) + buf - p,
+				       "\t\tMAX-MPDU-3895\n");
+			break;
+		case IEEE80211_VHT_CAP_MAX_MPDU_LENGTH_7991:
+			p += scnprintf(p, sizeof(buf) + buf - p,
+				       "\t\tMAX-MPDU-7991\n");
+			break;
+		case IEEE80211_VHT_CAP_MAX_MPDU_LENGTH_11454:
+			p += scnprintf(p, sizeof(buf) + buf - p,
+				       "\t\tMAX-MPDU-11454\n");
+			break;
+		default:
+			p += scnprintf(p, sizeof(buf) + buf - p,
+				       "\t\tMAX-MPDU-UNKNOWN\n");
+		};
+		switch (vhtc->cap & IEEE80211_VHT_CAP_SUPP_CHAN_WIDTH_MASK) {
+		case 0:
+			p += scnprintf(p, sizeof(buf) + buf - p,
+				       "\t\t80Mhz\n");
+			break;
+		case IEEE80211_VHT_CAP_SUPP_CHAN_WIDTH_160MHZ:
+			p += scnprintf(p, sizeof(buf) + buf - p,
+				       "\t\t160Mhz\n");
+			break;
+		case IEEE80211_VHT_CAP_SUPP_CHAN_WIDTH_160_80PLUS80MHZ:
+			p += scnprintf(p, sizeof(buf) + buf - p,
+				       "\t\t80+80Mhz\n");
+			break;
+		default:
+			p += scnprintf(p, sizeof(buf) + buf - p,
+				       "\t\tUNKNOWN-MHZ: 0x%x\n",
+				       (vhtc->cap >> 2) & 0x3);
+		};
+		PFLAG(RXLDPC, "RXLDPC");
+		PFLAG(SHORT_GI_80, "SHORT-GI-80");
+		PFLAG(SHORT_GI_160, "SHORT-GI-160");
+		PFLAG(TXSTBC, "TXSTBC");
+		p += scnprintf(p, sizeof(buf) + buf - p,
+			       "\t\tRXSTBC_%d\n", (vhtc->cap >> 8) & 0x7);
+		PFLAG(SU_BEAMFORMER_CAPABLE, "SU-BEAMFORMER-CAPABLE");
+		PFLAG(SU_BEAMFORMEE_CAPABLE, "SU-BEAMFORMEE-CAPABLE");
+		p += scnprintf(p, sizeof(buf) + buf - p,
+			"\t\tBEAMFORMEE-STS: 0x%x\n",
+			(vhtc->cap & IEEE80211_VHT_CAP_BEAMFORMEE_STS_MASK) >>
+			IEEE80211_VHT_CAP_BEAMFORMEE_STS_SHIFT);
+		p += scnprintf(p, sizeof(buf) + buf - p,
+			"\t\tSOUNDING-DIMENSIONS: 0x%x\n",
+			(vhtc->cap & IEEE80211_VHT_CAP_SOUNDING_DIMENSIONS_MASK)
+			>> IEEE80211_VHT_CAP_SOUNDING_DIMENSIONS_SHIFT);
+		PFLAG(MU_BEAMFORMER_CAPABLE, "MU-BEAMFORMER-CAPABLE");
+		PFLAG(MU_BEAMFORMEE_CAPABLE, "MU-BEAMFORMEE-CAPABLE");
+		PFLAG(VHT_TXOP_PS, "TXOP-PS");
+		PFLAG(HTC_VHT, "HTC-VHT");
+		p += scnprintf(p, sizeof(buf) + buf - p,
+			"\t\tMPDU-LENGTH-EXPONENT: 0x%x\n",
+			(vhtc->cap & IEEE80211_VHT_CAP_MAX_A_MPDU_LENGTH_EXPONENT_MASK) >>
+			IEEE80211_VHT_CAP_MAX_A_MPDU_LENGTH_EXPONENT_SHIFT);
+		PFLAG(VHT_LINK_ADAPTATION_VHT_UNSOL_MFB,
+		      "LINK-ADAPTATION-VHT-UNSOL-MFB");
+		p += scnprintf(p, sizeof(buf) + buf - p,
+			"\t\tLINK-ADAPTATION-VHT-MRQ-MFB: 0x%x\n",
+			(vhtc->cap & IEEE80211_VHT_CAP_VHT_LINK_ADAPTATION_VHT_MRQ_MFB) >> 26);
+		PFLAG(RX_ANTENNA_PATTERN, "RX-ANTENNA-PATTERN");
+		PFLAG(TX_ANTENNA_PATTERN, "TX-ANTENNA-PATTERN");
 
 		p += scnprintf(p, sizeof(buf)+buf-p, "RX MCS: %.4x\n",
 			       le16_to_cpu(vhtc->vht_mcs.rx_mcs_map));
-- 
cgit v1.2.3


From 80a83cfc434b1e3afe38974570b460db4898bec6 Mon Sep 17 00:00:00 2001
From: Michal Kazior <michal.kazior@tieto.com>
Date: Thu, 19 May 2016 10:37:48 +0200
Subject: mac80211: skip netdev queue control with software queuing

Qdiscs are designed with no regard to 802.11
aggregation requirements and hand out
packet-by-packet with no guarantee they are
destined to the same tid. This does more bad than
good no matter how fairly a given qdisc may behave
on an ethernet interface.

Software queuing used per-AC netdev subqueue
congestion control whenever a global AC limit was
hit. This meant in practice a single station or
tid queue could starve others rather easily. This
could resonate with qdiscs in a bad way or could
just end up with poor aggregation performance.
Increasing the AC limit would increase induced
latency which is also bad.

Disabling qdiscs by default and performing
taildrop instead of netdev subqueue congestion
control on the other hand makes it possible for
tid queues to fill up "in the meantime" while
preventing stations starving each other.

This increases aggregation opportunities and
should allow software queuing based drivers
achieve better performance by utilizing airtime
more efficiently with big aggregates.

Signed-off-by: Michal Kazior <michal.kazior@tieto.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
---
 net/mac80211/ieee80211_i.h |  2 +-
 net/mac80211/iface.c       | 18 +++++++++--
 net/mac80211/main.c        |  3 --
 net/mac80211/sta_info.c    |  2 +-
 net/mac80211/tx.c          | 75 ++++++++++++++++++++++++++--------------------
 net/mac80211/util.c        | 11 +++----
 6 files changed, 66 insertions(+), 45 deletions(-)

(limited to 'net/mac80211')

diff --git a/net/mac80211/ieee80211_i.h b/net/mac80211/ieee80211_i.h
index 9438c9406687..634603320374 100644
--- a/net/mac80211/ieee80211_i.h
+++ b/net/mac80211/ieee80211_i.h
@@ -856,7 +856,7 @@ struct ieee80211_sub_if_data {
 	bool control_port_no_encrypt;
 	int encrypt_headroom;
 
-	atomic_t txqs_len[IEEE80211_NUM_ACS];
+	atomic_t num_tx_queued;
 	struct ieee80211_tx_queue_params tx_conf[IEEE80211_NUM_ACS];
 	struct mac80211_qos_map __rcu *qos_map;
 
diff --git a/net/mac80211/iface.c b/net/mac80211/iface.c
index c59af3eb9fa4..609c5174d798 100644
--- a/net/mac80211/iface.c
+++ b/net/mac80211/iface.c
@@ -976,13 +976,13 @@ static void ieee80211_do_stop(struct ieee80211_sub_if_data *sdata,
 
 	if (sdata->vif.txq) {
 		struct txq_info *txqi = to_txq_info(sdata->vif.txq);
+		int n = skb_queue_len(&txqi->queue);
 
 		spin_lock_bh(&txqi->queue.lock);
 		ieee80211_purge_tx_queue(&local->hw, &txqi->queue);
+		atomic_sub(n, &sdata->num_tx_queued);
 		txqi->byte_cnt = 0;
 		spin_unlock_bh(&txqi->queue.lock);
-
-		atomic_set(&sdata->txqs_len[txqi->txq.ac], 0);
 	}
 
 	if (local->open_count == 0)
@@ -1198,6 +1198,12 @@ static void ieee80211_if_setup(struct net_device *dev)
 	dev->destructor = ieee80211_if_free;
 }
 
+static void ieee80211_if_setup_no_queue(struct net_device *dev)
+{
+	ieee80211_if_setup(dev);
+	dev->priv_flags |= IFF_NO_QUEUE;
+}
+
 static void ieee80211_iface_work(struct work_struct *work)
 {
 	struct ieee80211_sub_if_data *sdata =
@@ -1707,6 +1713,7 @@ int ieee80211_if_add(struct ieee80211_local *local, const char *name,
 	struct net_device *ndev = NULL;
 	struct ieee80211_sub_if_data *sdata = NULL;
 	struct txq_info *txqi;
+	void (*if_setup)(struct net_device *dev);
 	int ret, i;
 	int txqs = 1;
 
@@ -1734,12 +1741,17 @@ int ieee80211_if_add(struct ieee80211_local *local, const char *name,
 			txq_size += sizeof(struct txq_info) +
 				    local->hw.txq_data_size;
 
+		if (local->ops->wake_tx_queue)
+			if_setup = ieee80211_if_setup_no_queue;
+		else
+			if_setup = ieee80211_if_setup;
+
 		if (local->hw.queues >= IEEE80211_NUM_ACS)
 			txqs = IEEE80211_NUM_ACS;
 
 		ndev = alloc_netdev_mqs(size + txq_size,
 					name, name_assign_type,
-					ieee80211_if_setup, txqs, 1);
+					if_setup, txqs, 1);
 		if (!ndev)
 			return -ENOMEM;
 		dev_net_set(ndev, wiphy_net(local->hw.wiphy));
diff --git a/net/mac80211/main.c b/net/mac80211/main.c
index 7ee91d6151d1..160ac6b8b9a1 100644
--- a/net/mac80211/main.c
+++ b/net/mac80211/main.c
@@ -1055,9 +1055,6 @@ int ieee80211_register_hw(struct ieee80211_hw *hw)
 
 	local->dynamic_ps_forced_timeout = -1;
 
-	if (!local->hw.txq_ac_max_pending)
-		local->hw.txq_ac_max_pending = 64;
-
 	result = ieee80211_wep_init(local);
 	if (result < 0)
 		wiphy_debug(local->hw.wiphy, "Failed to initialize wep: %d\n",
diff --git a/net/mac80211/sta_info.c b/net/mac80211/sta_info.c
index 5ccfdbd406bd..177cc6cd6416 100644
--- a/net/mac80211/sta_info.c
+++ b/net/mac80211/sta_info.c
@@ -116,7 +116,7 @@ static void __cleanup_single_sta(struct sta_info *sta)
 			int n = skb_queue_len(&txqi->queue);
 
 			ieee80211_purge_tx_queue(&local->hw, &txqi->queue);
-			atomic_sub(n, &sdata->txqs_len[txqi->txq.ac]);
+			atomic_sub(n, &sdata->num_tx_queued);
 			txqi->byte_cnt = 0;
 		}
 	}
diff --git a/net/mac80211/tx.c b/net/mac80211/tx.c
index 203044379ce0..3e77da195ce8 100644
--- a/net/mac80211/tx.c
+++ b/net/mac80211/tx.c
@@ -1236,27 +1236,21 @@ ieee80211_tx_prepare(struct ieee80211_sub_if_data *sdata,
 	return TX_CONTINUE;
 }
 
-static void ieee80211_drv_tx(struct ieee80211_local *local,
-			     struct ieee80211_vif *vif,
-			     struct ieee80211_sta *pubsta,
-			     struct sk_buff *skb)
+static struct txq_info *ieee80211_get_txq(struct ieee80211_local *local,
+					  struct ieee80211_vif *vif,
+					  struct ieee80211_sta *pubsta,
+					  struct sk_buff *skb)
 {
 	struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) skb->data;
-	struct ieee80211_sub_if_data *sdata = vif_to_sdata(vif);
 	struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb);
-	struct ieee80211_tx_control control = {
-		.sta = pubsta,
-	};
 	struct ieee80211_txq *txq = NULL;
-	struct txq_info *txqi;
-	u8 ac;
 
 	if ((info->flags & IEEE80211_TX_CTL_SEND_AFTER_DTIM) ||
 	    (info->control.flags & IEEE80211_TX_CTRL_PS_RESPONSE))
-		goto tx_normal;
+		return NULL;
 
 	if (!ieee80211_is_data(hdr->frame_control))
-		goto tx_normal;
+		return NULL;
 
 	if (pubsta) {
 		u8 tid = skb->priority & IEEE80211_QOS_CTL_TID_MASK;
@@ -1267,25 +1261,28 @@ static void ieee80211_drv_tx(struct ieee80211_local *local,
 	}
 
 	if (!txq)
-		goto tx_normal;
+		return NULL;
 
-	ac = txq->ac;
-	txqi = to_txq_info(txq);
-	atomic_inc(&sdata->txqs_len[ac]);
-	if (atomic_read(&sdata->txqs_len[ac]) >= local->hw.txq_ac_max_pending)
-		netif_stop_subqueue(sdata->dev, ac);
+	return to_txq_info(txq);
+}
 
-	spin_lock_bh(&txqi->queue.lock);
-	txqi->byte_cnt += skb->len;
-	__skb_queue_tail(&txqi->queue, skb);
-	spin_unlock_bh(&txqi->queue.lock);
+static void ieee80211_txq_enqueue(struct ieee80211_local *local,
+				  struct txq_info *txqi,
+				  struct sk_buff *skb)
+{
+	struct ieee80211_sub_if_data *sdata = vif_to_sdata(txqi->txq.vif);
 
-	drv_wake_tx_queue(local, txqi);
+	lockdep_assert_held(&txqi->queue.lock);
 
-	return;
+	if (atomic_read(&sdata->num_tx_queued) >= TOTAL_MAX_TX_BUFFER ||
+	    txqi->queue.qlen >= STA_MAX_TX_BUFFER) {
+		ieee80211_free_txskb(&local->hw, skb);
+		return;
+	}
 
-tx_normal:
-	drv_tx(local, &control, skb);
+	atomic_inc(&sdata->num_tx_queued);
+	txqi->byte_cnt += skb->len;
+	__skb_queue_tail(&txqi->queue, skb);
 }
 
 struct sk_buff *ieee80211_tx_dequeue(struct ieee80211_hw *hw,
@@ -1296,7 +1293,6 @@ struct sk_buff *ieee80211_tx_dequeue(struct ieee80211_hw *hw,
 	struct txq_info *txqi = container_of(txq, struct txq_info, txq);
 	struct ieee80211_hdr *hdr;
 	struct sk_buff *skb = NULL;
-	u8 ac = txq->ac;
 
 	spin_lock_bh(&txqi->queue.lock);
 
@@ -1307,12 +1303,9 @@ struct sk_buff *ieee80211_tx_dequeue(struct ieee80211_hw *hw,
 	if (!skb)
 		goto out;
 
+	atomic_dec(&sdata->num_tx_queued);
 	txqi->byte_cnt -= skb->len;
 
-	atomic_dec(&sdata->txqs_len[ac]);
-	if (__netif_subqueue_stopped(sdata->dev, ac))
-		ieee80211_propagate_queue_wake(local, sdata->vif.hw_queue[ac]);
-
 	hdr = (struct ieee80211_hdr *)skb->data;
 	if (txq->sta && ieee80211_is_data_qos(hdr->frame_control)) {
 		struct sta_info *sta = container_of(txq->sta, struct sta_info,
@@ -1343,7 +1336,9 @@ static bool ieee80211_tx_frags(struct ieee80211_local *local,
 			       struct sk_buff_head *skbs,
 			       bool txpending)
 {
+	struct ieee80211_tx_control control = {};
 	struct sk_buff *skb, *tmp;
+	struct txq_info *txqi;
 	unsigned long flags;
 
 	skb_queue_walk_safe(skbs, skb, tmp) {
@@ -1358,6 +1353,21 @@ static bool ieee80211_tx_frags(struct ieee80211_local *local,
 		}
 #endif
 
+		txqi = ieee80211_get_txq(local, vif, sta, skb);
+		if (txqi) {
+			info->control.vif = vif;
+
+			__skb_unlink(skb, skbs);
+
+			spin_lock_bh(&txqi->queue.lock);
+			ieee80211_txq_enqueue(local, txqi, skb);
+			spin_unlock_bh(&txqi->queue.lock);
+
+			drv_wake_tx_queue(local, txqi);
+
+			continue;
+		}
+
 		spin_lock_irqsave(&local->queue_stop_reason_lock, flags);
 		if (local->queue_stop_reasons[q] ||
 		    (!txpending && !skb_queue_empty(&local->pending[q]))) {
@@ -1400,9 +1410,10 @@ static bool ieee80211_tx_frags(struct ieee80211_local *local,
 		spin_unlock_irqrestore(&local->queue_stop_reason_lock, flags);
 
 		info->control.vif = vif;
+		control.sta = sta;
 
 		__skb_unlink(skb, skbs);
-		ieee80211_drv_tx(local, vif, sta, skb);
+		drv_tx(local, &control, skb);
 	}
 
 	return true;
diff --git a/net/mac80211/util.c b/net/mac80211/util.c
index 905003f75c4d..0db46442bdcf 100644
--- a/net/mac80211/util.c
+++ b/net/mac80211/util.c
@@ -244,6 +244,9 @@ void ieee80211_propagate_queue_wake(struct ieee80211_local *local, int queue)
 	struct ieee80211_sub_if_data *sdata;
 	int n_acs = IEEE80211_NUM_ACS;
 
+	if (local->ops->wake_tx_queue)
+		return;
+
 	if (local->hw.queues < IEEE80211_NUM_ACS)
 		n_acs = 1;
 
@@ -260,11 +263,6 @@ void ieee80211_propagate_queue_wake(struct ieee80211_local *local, int queue)
 		for (ac = 0; ac < n_acs; ac++) {
 			int ac_queue = sdata->vif.hw_queue[ac];
 
-			if (local->ops->wake_tx_queue &&
-			    (atomic_read(&sdata->txqs_len[ac]) >
-			     local->hw.txq_ac_max_pending))
-				continue;
-
 			if (ac_queue == queue ||
 			    (sdata->vif.cab_queue == queue &&
 			     local->queue_stop_reasons[ac_queue] == 0 &&
@@ -352,6 +350,9 @@ static void __ieee80211_stop_queue(struct ieee80211_hw *hw, int queue,
 	if (__test_and_set_bit(reason, &local->queue_stop_reasons[queue]))
 		return;
 
+	if (local->ops->wake_tx_queue)
+		return;
+
 	if (local->hw.queues < IEEE80211_NUM_ACS)
 		n_acs = 1;
 
-- 
cgit v1.2.3


From fa962b92120bb70693a4db545f89067eb3373294 Mon Sep 17 00:00:00 2001
From: Michal Kazior <michal.kazior@tieto.com>
Date: Thu, 19 May 2016 10:37:49 +0200
Subject: mac80211: implement fair queueing per txq

mac80211's software queues were designed to work
very closely with device tx queues. They are
required to make use of 802.11 packet aggregation
easily and efficiently.

Due to the way 802.11 aggregation is designed it
only makes sense to keep fair queuing as close to
hardware as possible to reduce induced latency and
inertia and provide the best flow responsiveness.

This change doesn't translate directly to
immediate and significant gains. End result
depends on driver's induced latency. Best results
can be achieved if driver keeps its own tx
queue/fifo fill level to a minimum.

Signed-off-by: Michal Kazior <michal.kazior@tieto.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
---
 net/mac80211/agg-tx.c      |   8 ++-
 net/mac80211/ieee80211_i.h |  24 ++++++--
 net/mac80211/iface.c       |  12 ++--
 net/mac80211/main.c        |   7 +++
 net/mac80211/rx.c          |   2 +-
 net/mac80211/sta_info.c    |  14 ++---
 net/mac80211/tx.c          | 136 ++++++++++++++++++++++++++++++++++++++-------
 net/mac80211/util.c        |  23 +-------
 8 files changed, 162 insertions(+), 64 deletions(-)

(limited to 'net/mac80211')

diff --git a/net/mac80211/agg-tx.c b/net/mac80211/agg-tx.c
index 42fa81031dfa..5650c46bf91a 100644
--- a/net/mac80211/agg-tx.c
+++ b/net/mac80211/agg-tx.c
@@ -194,17 +194,21 @@ static void
 ieee80211_agg_stop_txq(struct sta_info *sta, int tid)
 {
 	struct ieee80211_txq *txq = sta->sta.txq[tid];
+	struct ieee80211_sub_if_data *sdata;
+	struct fq *fq;
 	struct txq_info *txqi;
 
 	if (!txq)
 		return;
 
 	txqi = to_txq_info(txq);
+	sdata = vif_to_sdata(txq->vif);
+	fq = &sdata->local->fq;
 
 	/* Lock here to protect against further seqno updates on dequeue */
-	spin_lock_bh(&txqi->queue.lock);
+	spin_lock_bh(&fq->lock);
 	set_bit(IEEE80211_TXQ_STOP, &txqi->flags);
-	spin_unlock_bh(&txqi->queue.lock);
+	spin_unlock_bh(&fq->lock);
 }
 
 static void
diff --git a/net/mac80211/ieee80211_i.h b/net/mac80211/ieee80211_i.h
index 634603320374..6f8375f1df88 100644
--- a/net/mac80211/ieee80211_i.h
+++ b/net/mac80211/ieee80211_i.h
@@ -30,6 +30,7 @@
 #include <net/ieee80211_radiotap.h>
 #include <net/cfg80211.h>
 #include <net/mac80211.h>
+#include <net/fq.h>
 #include "key.h"
 #include "sta_info.h"
 #include "debug.h"
@@ -805,10 +806,17 @@ enum txq_info_flags {
 	IEEE80211_TXQ_NO_AMSDU,
 };
 
+/**
+ * struct txq_info - per tid queue
+ *
+ * @tin: contains packets split into multiple flows
+ * @def_flow: used as a fallback flow when a packet destined to @tin hashes to
+ *	a fq_flow which is already owned by a different tin
+ */
 struct txq_info {
-	struct sk_buff_head queue;
+	struct fq_tin tin;
+	struct fq_flow def_flow;
 	unsigned long flags;
-	unsigned long byte_cnt;
 
 	/* keep last! */
 	struct ieee80211_txq txq;
@@ -1099,6 +1107,8 @@ struct ieee80211_local {
 	 * it first anyway so they become a no-op */
 	struct ieee80211_hw hw;
 
+	struct fq fq;
+
 	const struct ieee80211_ops *ops;
 
 	/*
@@ -1931,9 +1941,13 @@ static inline bool ieee80211_can_run_worker(struct ieee80211_local *local)
 	return true;
 }
 
-void ieee80211_init_tx_queue(struct ieee80211_sub_if_data *sdata,
-			     struct sta_info *sta,
-			     struct txq_info *txq, int tid);
+int ieee80211_txq_setup_flows(struct ieee80211_local *local);
+void ieee80211_txq_teardown_flows(struct ieee80211_local *local);
+void ieee80211_txq_init(struct ieee80211_sub_if_data *sdata,
+			struct sta_info *sta,
+			struct txq_info *txq, int tid);
+void ieee80211_txq_purge(struct ieee80211_local *local,
+			 struct txq_info *txqi);
 void ieee80211_send_auth(struct ieee80211_sub_if_data *sdata,
 			 u16 transaction, u16 auth_alg, u16 status,
 			 const u8 *extra, size_t extra_len, const u8 *bssid,
diff --git a/net/mac80211/iface.c b/net/mac80211/iface.c
index 609c5174d798..b123a9e325b3 100644
--- a/net/mac80211/iface.c
+++ b/net/mac80211/iface.c
@@ -779,6 +779,7 @@ static void ieee80211_do_stop(struct ieee80211_sub_if_data *sdata,
 			      bool going_down)
 {
 	struct ieee80211_local *local = sdata->local;
+	struct fq *fq = &local->fq;
 	unsigned long flags;
 	struct sk_buff *skb, *tmp;
 	u32 hw_reconf_flags = 0;
@@ -976,13 +977,10 @@ static void ieee80211_do_stop(struct ieee80211_sub_if_data *sdata,
 
 	if (sdata->vif.txq) {
 		struct txq_info *txqi = to_txq_info(sdata->vif.txq);
-		int n = skb_queue_len(&txqi->queue);
 
-		spin_lock_bh(&txqi->queue.lock);
-		ieee80211_purge_tx_queue(&local->hw, &txqi->queue);
-		atomic_sub(n, &sdata->num_tx_queued);
-		txqi->byte_cnt = 0;
-		spin_unlock_bh(&txqi->queue.lock);
+		spin_lock_bh(&fq->lock);
+		ieee80211_txq_purge(local, txqi);
+		spin_unlock_bh(&fq->lock);
 	}
 
 	if (local->open_count == 0)
@@ -1792,7 +1790,7 @@ int ieee80211_if_add(struct ieee80211_local *local, const char *name,
 
 		if (txq_size) {
 			txqi = netdev_priv(ndev) + size;
-			ieee80211_init_tx_queue(sdata, NULL, txqi, 0);
+			ieee80211_txq_init(sdata, NULL, txqi, 0);
 		}
 
 		sdata->dev = ndev;
diff --git a/net/mac80211/main.c b/net/mac80211/main.c
index 160ac6b8b9a1..d00ea9b13f49 100644
--- a/net/mac80211/main.c
+++ b/net/mac80211/main.c
@@ -1086,6 +1086,10 @@ int ieee80211_register_hw(struct ieee80211_hw *hw)
 
 	rtnl_unlock();
 
+	result = ieee80211_txq_setup_flows(local);
+	if (result)
+		goto fail_flows;
+
 #ifdef CONFIG_INET
 	local->ifa_notifier.notifier_call = ieee80211_ifa_changed;
 	result = register_inetaddr_notifier(&local->ifa_notifier);
@@ -1111,6 +1115,8 @@ int ieee80211_register_hw(struct ieee80211_hw *hw)
 #if defined(CONFIG_INET) || defined(CONFIG_IPV6)
  fail_ifa:
 #endif
+	ieee80211_txq_teardown_flows(local);
+ fail_flows:
 	rtnl_lock();
 	rate_control_deinitialize(local);
 	ieee80211_remove_interfaces(local);
@@ -1169,6 +1175,7 @@ void ieee80211_unregister_hw(struct ieee80211_hw *hw)
 	skb_queue_purge(&local->skb_queue);
 	skb_queue_purge(&local->skb_queue_unreliable);
 	skb_queue_purge(&local->skb_queue_tdls_chsw);
+	ieee80211_txq_teardown_flows(local);
 
 	destroy_workqueue(local->workqueue);
 	wiphy_unregister(local->hw.wiphy);
diff --git a/net/mac80211/rx.c b/net/mac80211/rx.c
index 5e65e838992a..9a1eb70cb120 100644
--- a/net/mac80211/rx.c
+++ b/net/mac80211/rx.c
@@ -1268,7 +1268,7 @@ static void sta_ps_start(struct sta_info *sta)
 	for (tid = 0; tid < ARRAY_SIZE(sta->sta.txq); tid++) {
 		struct txq_info *txqi = to_txq_info(sta->sta.txq[tid]);
 
-		if (!skb_queue_len(&txqi->queue))
+		if (!txqi->tin.backlog_packets)
 			set_bit(tid, &sta->txq_buffered_tids);
 		else
 			clear_bit(tid, &sta->txq_buffered_tids);
diff --git a/net/mac80211/sta_info.c b/net/mac80211/sta_info.c
index 177cc6cd6416..76b737dcc36f 100644
--- a/net/mac80211/sta_info.c
+++ b/net/mac80211/sta_info.c
@@ -90,6 +90,7 @@ static void __cleanup_single_sta(struct sta_info *sta)
 	struct tid_ampdu_tx *tid_tx;
 	struct ieee80211_sub_if_data *sdata = sta->sdata;
 	struct ieee80211_local *local = sdata->local;
+	struct fq *fq = &local->fq;
 	struct ps_data *ps;
 
 	if (test_sta_flag(sta, WLAN_STA_PS_STA) ||
@@ -113,11 +114,10 @@ static void __cleanup_single_sta(struct sta_info *sta)
 	if (sta->sta.txq[0]) {
 		for (i = 0; i < ARRAY_SIZE(sta->sta.txq); i++) {
 			struct txq_info *txqi = to_txq_info(sta->sta.txq[i]);
-			int n = skb_queue_len(&txqi->queue);
 
-			ieee80211_purge_tx_queue(&local->hw, &txqi->queue);
-			atomic_sub(n, &sdata->num_tx_queued);
-			txqi->byte_cnt = 0;
+			spin_lock_bh(&fq->lock);
+			ieee80211_txq_purge(local, txqi);
+			spin_unlock_bh(&fq->lock);
 		}
 	}
 
@@ -368,7 +368,7 @@ struct sta_info *sta_info_alloc(struct ieee80211_sub_if_data *sdata,
 		for (i = 0; i < ARRAY_SIZE(sta->sta.txq); i++) {
 			struct txq_info *txq = txq_data + i * size;
 
-			ieee80211_init_tx_queue(sdata, sta, txq, i);
+			ieee80211_txq_init(sdata, sta, txq, i);
 		}
 	}
 
@@ -1211,7 +1211,7 @@ void ieee80211_sta_ps_deliver_wakeup(struct sta_info *sta)
 		for (i = 0; i < ARRAY_SIZE(sta->sta.txq); i++) {
 			struct txq_info *txqi = to_txq_info(sta->sta.txq[i]);
 
-			if (!skb_queue_len(&txqi->queue))
+			if (!txqi->tin.backlog_packets)
 				continue;
 
 			drv_wake_tx_queue(local, txqi);
@@ -1648,7 +1648,7 @@ ieee80211_sta_ps_deliver_response(struct sta_info *sta,
 		for (tid = 0; tid < ARRAY_SIZE(sta->sta.txq); tid++) {
 			struct txq_info *txqi = to_txq_info(sta->sta.txq[tid]);
 
-			if (!(tids & BIT(tid)) || skb_queue_len(&txqi->queue))
+			if (!(tids & BIT(tid)) || txqi->tin.backlog_packets)
 				continue;
 
 			sta_info_recalc_tim(sta);
diff --git a/net/mac80211/tx.c b/net/mac80211/tx.c
index 3e77da195ce8..1d8343fca6d4 100644
--- a/net/mac80211/tx.c
+++ b/net/mac80211/tx.c
@@ -25,6 +25,7 @@
 #include <net/cfg80211.h>
 #include <net/mac80211.h>
 #include <asm/unaligned.h>
+#include <net/fq_impl.h>
 
 #include "ieee80211_i.h"
 #include "driver-ops.h"
@@ -1266,46 +1267,121 @@ static struct txq_info *ieee80211_get_txq(struct ieee80211_local *local,
 	return to_txq_info(txq);
 }
 
+static struct sk_buff *fq_tin_dequeue_func(struct fq *fq,
+					   struct fq_tin *tin,
+					   struct fq_flow *flow)
+{
+	return fq_flow_dequeue(fq, flow);
+}
+
+static void fq_skb_free_func(struct fq *fq,
+			     struct fq_tin *tin,
+			     struct fq_flow *flow,
+			     struct sk_buff *skb)
+{
+	struct ieee80211_local *local;
+
+	local = container_of(fq, struct ieee80211_local, fq);
+	ieee80211_free_txskb(&local->hw, skb);
+}
+
+static struct fq_flow *fq_flow_get_default_func(struct fq *fq,
+						struct fq_tin *tin,
+						int idx,
+						struct sk_buff *skb)
+{
+	struct txq_info *txqi;
+
+	txqi = container_of(tin, struct txq_info, tin);
+	return &txqi->def_flow;
+}
+
 static void ieee80211_txq_enqueue(struct ieee80211_local *local,
 				  struct txq_info *txqi,
 				  struct sk_buff *skb)
 {
-	struct ieee80211_sub_if_data *sdata = vif_to_sdata(txqi->txq.vif);
+	struct fq *fq = &local->fq;
+	struct fq_tin *tin = &txqi->tin;
 
-	lockdep_assert_held(&txqi->queue.lock);
+	fq_tin_enqueue(fq, tin, skb,
+		       fq_skb_free_func,
+		       fq_flow_get_default_func);
+}
 
-	if (atomic_read(&sdata->num_tx_queued) >= TOTAL_MAX_TX_BUFFER ||
-	    txqi->queue.qlen >= STA_MAX_TX_BUFFER) {
-		ieee80211_free_txskb(&local->hw, skb);
-		return;
+void ieee80211_txq_init(struct ieee80211_sub_if_data *sdata,
+			struct sta_info *sta,
+			struct txq_info *txqi, int tid)
+{
+	fq_tin_init(&txqi->tin);
+	fq_flow_init(&txqi->def_flow);
+
+	txqi->txq.vif = &sdata->vif;
+
+	if (sta) {
+		txqi->txq.sta = &sta->sta;
+		sta->sta.txq[tid] = &txqi->txq;
+		txqi->txq.tid = tid;
+		txqi->txq.ac = ieee802_1d_to_ac[tid & 7];
+	} else {
+		sdata->vif.txq = &txqi->txq;
+		txqi->txq.tid = 0;
+		txqi->txq.ac = IEEE80211_AC_BE;
 	}
+}
 
-	atomic_inc(&sdata->num_tx_queued);
-	txqi->byte_cnt += skb->len;
-	__skb_queue_tail(&txqi->queue, skb);
+void ieee80211_txq_purge(struct ieee80211_local *local,
+			 struct txq_info *txqi)
+{
+	struct fq *fq = &local->fq;
+	struct fq_tin *tin = &txqi->tin;
+
+	fq_tin_reset(fq, tin, fq_skb_free_func);
+}
+
+int ieee80211_txq_setup_flows(struct ieee80211_local *local)
+{
+	struct fq *fq = &local->fq;
+	int ret;
+
+	if (!local->ops->wake_tx_queue)
+		return 0;
+
+	ret = fq_init(fq, 4096);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+void ieee80211_txq_teardown_flows(struct ieee80211_local *local)
+{
+	struct fq *fq = &local->fq;
+
+	if (!local->ops->wake_tx_queue)
+		return;
+
+	fq_reset(fq, fq_skb_free_func);
 }
 
 struct sk_buff *ieee80211_tx_dequeue(struct ieee80211_hw *hw,
 				     struct ieee80211_txq *txq)
 {
 	struct ieee80211_local *local = hw_to_local(hw);
-	struct ieee80211_sub_if_data *sdata = vif_to_sdata(txq->vif);
 	struct txq_info *txqi = container_of(txq, struct txq_info, txq);
 	struct ieee80211_hdr *hdr;
 	struct sk_buff *skb = NULL;
+	struct fq *fq = &local->fq;
+	struct fq_tin *tin = &txqi->tin;
 
-	spin_lock_bh(&txqi->queue.lock);
+	spin_lock_bh(&fq->lock);
 
 	if (test_bit(IEEE80211_TXQ_STOP, &txqi->flags))
 		goto out;
 
-	skb = __skb_dequeue(&txqi->queue);
+	skb = fq_tin_dequeue(fq, tin, fq_tin_dequeue_func);
 	if (!skb)
 		goto out;
 
-	atomic_dec(&sdata->num_tx_queued);
-	txqi->byte_cnt -= skb->len;
-
 	hdr = (struct ieee80211_hdr *)skb->data;
 	if (txq->sta && ieee80211_is_data_qos(hdr->frame_control)) {
 		struct sta_info *sta = container_of(txq->sta, struct sta_info,
@@ -1320,7 +1396,7 @@ struct sk_buff *ieee80211_tx_dequeue(struct ieee80211_hw *hw,
 	}
 
 out:
-	spin_unlock_bh(&txqi->queue.lock);
+	spin_unlock_bh(&fq->lock);
 
 	if (skb && skb_has_frag_list(skb) &&
 	    !ieee80211_hw_check(&local->hw, TX_FRAG_LIST))
@@ -1337,6 +1413,7 @@ static bool ieee80211_tx_frags(struct ieee80211_local *local,
 			       bool txpending)
 {
 	struct ieee80211_tx_control control = {};
+	struct fq *fq = &local->fq;
 	struct sk_buff *skb, *tmp;
 	struct txq_info *txqi;
 	unsigned long flags;
@@ -1359,9 +1436,9 @@ static bool ieee80211_tx_frags(struct ieee80211_local *local,
 
 			__skb_unlink(skb, skbs);
 
-			spin_lock_bh(&txqi->queue.lock);
+			spin_lock_bh(&fq->lock);
 			ieee80211_txq_enqueue(local, txqi, skb);
-			spin_unlock_bh(&txqi->queue.lock);
+			spin_unlock_bh(&fq->lock);
 
 			drv_wake_tx_queue(local, txqi);
 
@@ -2893,6 +2970,9 @@ static bool ieee80211_amsdu_aggregate(struct ieee80211_sub_if_data *sdata,
 				      struct sk_buff *skb)
 {
 	struct ieee80211_local *local = sdata->local;
+	struct fq *fq = &local->fq;
+	struct fq_tin *tin;
+	struct fq_flow *flow;
 	u8 tid = skb->priority & IEEE80211_QOS_CTL_TAG1D_MASK;
 	struct ieee80211_txq *txq = sta->sta.txq[tid];
 	struct txq_info *txqi;
@@ -2904,6 +2984,7 @@ static bool ieee80211_amsdu_aggregate(struct ieee80211_sub_if_data *sdata,
 	__be16 len;
 	void *data;
 	bool ret = false;
+	unsigned int orig_len;
 	int n = 1, nfrags;
 
 	if (!ieee80211_hw_check(&local->hw, TX_AMSDU))
@@ -2920,12 +3001,20 @@ static bool ieee80211_amsdu_aggregate(struct ieee80211_sub_if_data *sdata,
 		max_amsdu_len = min_t(int, max_amsdu_len,
 				      sta->sta.max_rc_amsdu_len);
 
-	spin_lock_bh(&txqi->queue.lock);
+	spin_lock_bh(&fq->lock);
 
-	head = skb_peek_tail(&txqi->queue);
+	/* TODO: Ideally aggregation should be done on dequeue to remain
+	 * responsive to environment changes.
+	 */
+
+	tin = &txqi->tin;
+	flow = fq_flow_classify(fq, tin, skb, fq_flow_get_default_func);
+	head = skb_peek_tail(&flow->queue);
 	if (!head)
 		goto out;
 
+	orig_len = head->len;
+
 	if (skb->len + head->len > max_amsdu_len)
 		goto out;
 
@@ -2964,8 +3053,13 @@ static bool ieee80211_amsdu_aggregate(struct ieee80211_sub_if_data *sdata,
 	head->data_len += skb->len;
 	*frag_tail = skb;
 
+	flow->backlog += head->len - orig_len;
+	tin->backlog_bytes += head->len - orig_len;
+
+	fq_recalc_backlog(fq, tin, flow);
+
 out:
-	spin_unlock_bh(&txqi->queue.lock);
+	spin_unlock_bh(&fq->lock);
 
 	return ret;
 }
diff --git a/net/mac80211/util.c b/net/mac80211/util.c
index 0db46442bdcf..42bf0b6685e8 100644
--- a/net/mac80211/util.c
+++ b/net/mac80211/util.c
@@ -3389,25 +3389,6 @@ u8 *ieee80211_add_wmm_info_ie(u8 *buf, u8 qosinfo)
 	return buf;
 }
 
-void ieee80211_init_tx_queue(struct ieee80211_sub_if_data *sdata,
-			     struct sta_info *sta,
-			     struct txq_info *txqi, int tid)
-{
-	skb_queue_head_init(&txqi->queue);
-	txqi->txq.vif = &sdata->vif;
-
-	if (sta) {
-		txqi->txq.sta = &sta->sta;
-		sta->sta.txq[tid] = &txqi->txq;
-		txqi->txq.tid = tid;
-		txqi->txq.ac = ieee802_1d_to_ac[tid & 7];
-	} else {
-		sdata->vif.txq = &txqi->txq;
-		txqi->txq.tid = 0;
-		txqi->txq.ac = IEEE80211_AC_BE;
-	}
-}
-
 void ieee80211_txq_get_depth(struct ieee80211_txq *txq,
 			     unsigned long *frame_cnt,
 			     unsigned long *byte_cnt)
@@ -3415,9 +3396,9 @@ void ieee80211_txq_get_depth(struct ieee80211_txq *txq,
 	struct txq_info *txqi = to_txq_info(txq);
 
 	if (frame_cnt)
-		*frame_cnt = txqi->queue.qlen;
+		*frame_cnt = txqi->tin.backlog_packets;
 
 	if (byte_cnt)
-		*byte_cnt = txqi->byte_cnt;
+		*byte_cnt = txqi->tin.backlog_bytes;
 }
 EXPORT_SYMBOL(ieee80211_txq_get_depth);
-- 
cgit v1.2.3


From 9399b86c0e9a058928e8b5f1e695056714814873 Mon Sep 17 00:00:00 2001
From: Michal Kazior <michal.kazior@tieto.com>
Date: Thu, 19 May 2016 10:37:50 +0200
Subject: mac80211: add debug knobs for fair queuing

This adds a debugfs entry to read and modify some fq parameters.

This makes it easy to debug, test and experiment.

Signed-off-by: Michal Kazior <michal.kazior@tieto.com>
[remove module parameter for now]
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
---
 net/mac80211/debugfs.c | 173 +++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 173 insertions(+)

(limited to 'net/mac80211')

diff --git a/net/mac80211/debugfs.c b/net/mac80211/debugfs.c
index b251b2f7f8dd..2906c1004e1a 100644
--- a/net/mac80211/debugfs.c
+++ b/net/mac80211/debugfs.c
@@ -10,6 +10,7 @@
 
 #include <linux/debugfs.h>
 #include <linux/rtnetlink.h>
+#include <linux/vmalloc.h>
 #include "ieee80211_i.h"
 #include "driver-ops.h"
 #include "rate.h"
@@ -70,6 +71,177 @@ DEBUGFS_READONLY_FILE(wep_iv, "%#08x",
 DEBUGFS_READONLY_FILE(rate_ctrl_alg, "%s",
 	local->rate_ctrl ? local->rate_ctrl->ops->name : "hw/driver");
 
+struct aqm_info {
+	struct ieee80211_local *local;
+	size_t size;
+	size_t len;
+	unsigned char buf[0];
+};
+
+#define AQM_HDR_LEN 200
+#define AQM_HW_ENTRY_LEN 40
+#define AQM_TXQ_ENTRY_LEN 110
+
+static int aqm_open(struct inode *inode, struct file *file)
+{
+	struct ieee80211_local *local = inode->i_private;
+	struct ieee80211_sub_if_data *sdata;
+	struct sta_info *sta;
+	struct txq_info *txqi;
+	struct fq *fq = &local->fq;
+	struct aqm_info *info = NULL;
+	int len = 0;
+	int i;
+
+	if (!local->ops->wake_tx_queue)
+		return -EOPNOTSUPP;
+
+	len += AQM_HDR_LEN;
+	len += 6 * AQM_HW_ENTRY_LEN;
+
+	rcu_read_lock();
+	list_for_each_entry_rcu(sdata, &local->interfaces, list)
+		len += AQM_TXQ_ENTRY_LEN;
+	list_for_each_entry_rcu(sta, &local->sta_list, list)
+		len += AQM_TXQ_ENTRY_LEN * ARRAY_SIZE(sta->sta.txq);
+	rcu_read_unlock();
+
+	info = vmalloc(len);
+	if (!info)
+		return -ENOMEM;
+
+	spin_lock_bh(&local->fq.lock);
+	rcu_read_lock();
+
+	file->private_data = info;
+	info->local = local;
+	info->size = len;
+	len = 0;
+
+	len += scnprintf(info->buf + len, info->size - len,
+			 "* hw\n"
+			 "access name value\n"
+			 "R fq_flows_cnt %u\n"
+			 "R fq_backlog %u\n"
+			 "R fq_overlimit %u\n"
+			 "R fq_collisions %u\n"
+			 "RW fq_limit %u\n"
+			 "RW fq_quantum %u\n",
+			 fq->flows_cnt,
+			 fq->backlog,
+			 fq->overlimit,
+			 fq->collisions,
+			 fq->limit,
+			 fq->quantum);
+
+	len += scnprintf(info->buf + len,
+			 info->size - len,
+			 "* vif\n"
+			 "ifname addr ac backlog-bytes backlog-packets flows overlimit collisions tx-bytes tx-packets\n");
+
+	list_for_each_entry_rcu(sdata, &local->interfaces, list) {
+		txqi = to_txq_info(sdata->vif.txq);
+		len += scnprintf(info->buf + len, info->size - len,
+				 "%s %pM %u %u %u %u %u %u %u %u\n",
+				 sdata->name,
+				 sdata->vif.addr,
+				 txqi->txq.ac,
+				 txqi->tin.backlog_bytes,
+				 txqi->tin.backlog_packets,
+				 txqi->tin.flows,
+				 txqi->tin.overlimit,
+				 txqi->tin.collisions,
+				 txqi->tin.tx_bytes,
+				 txqi->tin.tx_packets);
+	}
+
+	len += scnprintf(info->buf + len,
+			 info->size - len,
+			 "* sta\n"
+			 "ifname addr tid ac backlog-bytes backlog-packets flows overlimit collisions tx-bytes tx-packets\n");
+
+	list_for_each_entry_rcu(sta, &local->sta_list, list) {
+		sdata = sta->sdata;
+		for (i = 0; i < ARRAY_SIZE(sta->sta.txq); i++) {
+			txqi = to_txq_info(sta->sta.txq[i]);
+			len += scnprintf(info->buf + len, info->size - len,
+					 "%s %pM %d %d %u %u %u %u %u %u %u\n",
+					 sdata->name,
+					 sta->sta.addr,
+					 txqi->txq.tid,
+					 txqi->txq.ac,
+					 txqi->tin.backlog_bytes,
+					 txqi->tin.backlog_packets,
+					 txqi->tin.flows,
+					 txqi->tin.overlimit,
+					 txqi->tin.collisions,
+					 txqi->tin.tx_bytes,
+					 txqi->tin.tx_packets);
+		}
+	}
+
+	info->len = len;
+
+	rcu_read_unlock();
+	spin_unlock_bh(&local->fq.lock);
+
+	return 0;
+}
+
+static int aqm_release(struct inode *inode, struct file *file)
+{
+	vfree(file->private_data);
+	return 0;
+}
+
+static ssize_t aqm_read(struct file *file,
+			char __user *user_buf,
+			size_t count,
+			loff_t *ppos)
+{
+	struct aqm_info *info = file->private_data;
+
+	return simple_read_from_buffer(user_buf, count, ppos,
+				       info->buf, info->len);
+}
+
+static ssize_t aqm_write(struct file *file,
+			 const char __user *user_buf,
+			 size_t count,
+			 loff_t *ppos)
+{
+	struct aqm_info *info = file->private_data;
+	struct ieee80211_local *local = info->local;
+	char buf[100];
+	size_t len;
+
+	if (count > sizeof(buf))
+		return -EINVAL;
+
+	if (copy_from_user(buf, user_buf, count))
+		return -EFAULT;
+
+	buf[sizeof(buf) - 1] = '\0';
+	len = strlen(buf);
+	if (len > 0 && buf[len-1] == '\n')
+		buf[len-1] = 0;
+
+	if (sscanf(buf, "fq_limit %u", &local->fq.limit) == 1)
+		return count;
+	else if (sscanf(buf, "fq_quantum %u", &local->fq.quantum) == 1)
+		return count;
+
+	return -EINVAL;
+}
+
+static const struct file_operations aqm_ops = {
+	.write = aqm_write,
+	.read = aqm_read,
+	.open = aqm_open,
+	.release = aqm_release,
+	.llseek = default_llseek,
+};
+
 #ifdef CONFIG_PM
 static ssize_t reset_write(struct file *file, const char __user *user_buf,
 			   size_t count, loff_t *ppos)
@@ -256,6 +428,7 @@ void debugfs_hw_add(struct ieee80211_local *local)
 	DEBUGFS_ADD(hwflags);
 	DEBUGFS_ADD(user_power);
 	DEBUGFS_ADD(power);
+	DEBUGFS_ADD_MODE(aqm, 0600);
 
 	statsd = debugfs_create_dir("statistics", phyd);
 
-- 
cgit v1.2.3


From 5caa328e3811b7cfa33fd02c93280ffa622deb0e Mon Sep 17 00:00:00 2001
From: Michal Kazior <michal.kazior@tieto.com>
Date: Thu, 19 May 2016 10:37:51 +0200
Subject: mac80211: implement codel on fair queuing flows

There is no other limit other than a global
packet count limit when using software queuing.
This means a single flow queue can grow insanely
long. This is particularly bad for TCP congestion
algorithms which requires a little more
sophisticated frame dropping scheme than a mere
headdrop on limit overflow.

Hence apply (a slighly modified, to fit the knobs)
CoDel5 on flow queues. This improves TCP
convergence and stability when combined with
wireless driver which keeps its own tx queue/fifo
at a minimum fill level for given link conditions.

Signed-off-by: Michal Kazior <michal.kazior@tieto.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
---
 net/mac80211/ieee80211_i.h |   5 +++
 net/mac80211/tx.c          | 109 ++++++++++++++++++++++++++++++++++++++++++++-
 2 files changed, 113 insertions(+), 1 deletion(-)

(limited to 'net/mac80211')

diff --git a/net/mac80211/ieee80211_i.h b/net/mac80211/ieee80211_i.h
index 6f8375f1df88..54edfb6fc1d1 100644
--- a/net/mac80211/ieee80211_i.h
+++ b/net/mac80211/ieee80211_i.h
@@ -812,10 +812,12 @@ enum txq_info_flags {
  * @tin: contains packets split into multiple flows
  * @def_flow: used as a fallback flow when a packet destined to @tin hashes to
  *	a fq_flow which is already owned by a different tin
+ * @def_cvars: codel vars for @def_flow
  */
 struct txq_info {
 	struct fq_tin tin;
 	struct fq_flow def_flow;
+	struct codel_vars def_cvars;
 	unsigned long flags;
 
 	/* keep last! */
@@ -1108,6 +1110,9 @@ struct ieee80211_local {
 	struct ieee80211_hw hw;
 
 	struct fq fq;
+	struct codel_vars *cvars;
+	struct codel_params cparams;
+	struct codel_stats cstats;
 
 	const struct ieee80211_ops *ops;
 
diff --git a/net/mac80211/tx.c b/net/mac80211/tx.c
index 1d8343fca6d4..44ec605a5682 100644
--- a/net/mac80211/tx.c
+++ b/net/mac80211/tx.c
@@ -24,6 +24,8 @@
 #include <net/ieee80211_radiotap.h>
 #include <net/cfg80211.h>
 #include <net/mac80211.h>
+#include <net/codel.h>
+#include <net/codel_impl.h>
 #include <asm/unaligned.h>
 #include <net/fq_impl.h>
 
@@ -1267,11 +1269,92 @@ static struct txq_info *ieee80211_get_txq(struct ieee80211_local *local,
 	return to_txq_info(txq);
 }
 
+static void ieee80211_set_skb_enqueue_time(struct sk_buff *skb)
+{
+	IEEE80211_SKB_CB(skb)->control.enqueue_time = codel_get_time();
+}
+
+static void ieee80211_set_skb_vif(struct sk_buff *skb, struct txq_info *txqi)
+{
+	IEEE80211_SKB_CB(skb)->control.vif = txqi->txq.vif;
+}
+
+static u32 codel_skb_len_func(const struct sk_buff *skb)
+{
+	return skb->len;
+}
+
+static codel_time_t codel_skb_time_func(const struct sk_buff *skb)
+{
+	const struct ieee80211_tx_info *info;
+
+	info = (const struct ieee80211_tx_info *)skb->cb;
+	return info->control.enqueue_time;
+}
+
+static struct sk_buff *codel_dequeue_func(struct codel_vars *cvars,
+					  void *ctx)
+{
+	struct ieee80211_local *local;
+	struct txq_info *txqi;
+	struct fq *fq;
+	struct fq_flow *flow;
+
+	txqi = ctx;
+	local = vif_to_sdata(txqi->txq.vif)->local;
+	fq = &local->fq;
+
+	if (cvars == &txqi->def_cvars)
+		flow = &txqi->def_flow;
+	else
+		flow = &fq->flows[cvars - local->cvars];
+
+	return fq_flow_dequeue(fq, flow);
+}
+
+static void codel_drop_func(struct sk_buff *skb,
+			    void *ctx)
+{
+	struct ieee80211_local *local;
+	struct ieee80211_hw *hw;
+	struct txq_info *txqi;
+
+	txqi = ctx;
+	local = vif_to_sdata(txqi->txq.vif)->local;
+	hw = &local->hw;
+
+	ieee80211_free_txskb(hw, skb);
+}
+
 static struct sk_buff *fq_tin_dequeue_func(struct fq *fq,
 					   struct fq_tin *tin,
 					   struct fq_flow *flow)
 {
-	return fq_flow_dequeue(fq, flow);
+	struct ieee80211_local *local;
+	struct txq_info *txqi;
+	struct codel_vars *cvars;
+	struct codel_params *cparams;
+	struct codel_stats *cstats;
+
+	local = container_of(fq, struct ieee80211_local, fq);
+	txqi = container_of(tin, struct txq_info, tin);
+	cparams = &local->cparams;
+	cstats = &local->cstats;
+
+	if (flow == &txqi->def_flow)
+		cvars = &txqi->def_cvars;
+	else
+		cvars = &local->cvars[flow - fq->flows];
+
+	return codel_dequeue(txqi,
+			     &flow->backlog,
+			     cparams,
+			     cvars,
+			     cstats,
+			     codel_skb_len_func,
+			     codel_skb_time_func,
+			     codel_drop_func,
+			     codel_dequeue_func);
 }
 
 static void fq_skb_free_func(struct fq *fq,
@@ -1303,6 +1386,7 @@ static void ieee80211_txq_enqueue(struct ieee80211_local *local,
 	struct fq *fq = &local->fq;
 	struct fq_tin *tin = &txqi->tin;
 
+	ieee80211_set_skb_enqueue_time(skb);
 	fq_tin_enqueue(fq, tin, skb,
 		       fq_skb_free_func,
 		       fq_flow_get_default_func);
@@ -1314,6 +1398,7 @@ void ieee80211_txq_init(struct ieee80211_sub_if_data *sdata,
 {
 	fq_tin_init(&txqi->tin);
 	fq_flow_init(&txqi->def_flow);
+	codel_vars_init(&txqi->def_cvars);
 
 	txqi->txq.vif = &sdata->vif;
 
@@ -1342,6 +1427,7 @@ int ieee80211_txq_setup_flows(struct ieee80211_local *local)
 {
 	struct fq *fq = &local->fq;
 	int ret;
+	int i;
 
 	if (!local->ops->wake_tx_queue)
 		return 0;
@@ -1350,6 +1436,22 @@ int ieee80211_txq_setup_flows(struct ieee80211_local *local)
 	if (ret)
 		return ret;
 
+	codel_params_init(&local->cparams);
+	codel_stats_init(&local->cstats);
+	local->cparams.interval = MS2TIME(100);
+	local->cparams.target = MS2TIME(20);
+	local->cparams.ecn = true;
+
+	local->cvars = kcalloc(fq->flows_cnt, sizeof(local->cvars[0]),
+			       GFP_KERNEL);
+	if (!local->cvars) {
+		fq_reset(fq, fq_skb_free_func);
+		return -ENOMEM;
+	}
+
+	for (i = 0; i < fq->flows_cnt; i++)
+		codel_vars_init(&local->cvars[i]);
+
 	return 0;
 }
 
@@ -1360,6 +1462,9 @@ void ieee80211_txq_teardown_flows(struct ieee80211_local *local)
 	if (!local->ops->wake_tx_queue)
 		return;
 
+	kfree(local->cvars);
+	local->cvars = NULL;
+
 	fq_reset(fq, fq_skb_free_func);
 }
 
@@ -1382,6 +1487,8 @@ struct sk_buff *ieee80211_tx_dequeue(struct ieee80211_hw *hw,
 	if (!skb)
 		goto out;
 
+	ieee80211_set_skb_vif(skb, txqi);
+
 	hdr = (struct ieee80211_hdr *)skb->data;
 	if (txq->sta && ieee80211_is_data_qos(hdr->frame_control)) {
 		struct sta_info *sta = container_of(txq->sta, struct sta_info,
-- 
cgit v1.2.3