be2net: support multiple TX queues

This patch provides support for multiple TX queues.

Signed-off-by: Sathya Perla <[e-mail]>
Signed-off-by: David S. Miller <[e-mail]>

drivers/net/benet/be.h

···
 
 #define MAX_RSS_QS		4	/* BE limit is 4 queues/port */
 #define MAX_RX_QS		(MAX_RSS_QS + 1) /* RSS qs + 1 def Rx */
+#define MAX_TX_QS		8
 #define BE_MAX_MSIX_VECTORS	(MAX_RX_QS + 1)/* RX + TX */
 #define BE_NAPI_WEIGHT		64
 #define MAX_RX_POST 		BE_NAPI_WEIGHT /* Frags posted at a time */
···
 	u32 be_tx_reqs;		/* number of TX requests initiated */
 	u32 be_tx_stops;	/* number of times TX Q was stopped */
 	u32 be_tx_wrbs;		/* number of tx WRBs used */
-	u32 be_tx_events;	/* number of tx completion events  */
 	u32 be_tx_compl;	/* number of tx completion entries processed */
 	ulong be_tx_jiffies;
 	u64 be_tx_bytes;
···
 	struct be_queue_info cq;
 	/* Remember the skbs that were transmitted */
 	struct sk_buff *sent_skb_list[TX_Q_LEN];
+	struct be_tx_stats stats;
 };
 
 /* Struct to remember the pages posted for rx frags */
···
 
 	/* TX Rings */
 	struct be_eq_obj tx_eq;
-	struct be_tx_obj tx_obj;
-	struct be_tx_stats tx_stats;
+	struct be_tx_obj tx_obj[MAX_TX_QS];
+	u8 num_tx_qs;
 
 	u32 cache_line_break[8];
 
···
 extern const struct ethtool_ops be_ethtool_ops;
 
 #define msix_enabled(adapter)		(adapter->num_msix_vec > 0)
-#define tx_stats(adapter)		(&adapter->tx_stats)
+#define tx_stats(txo)			(&txo->stats)
 #define rx_stats(rxo)			(&rxo->stats)
 
 #define BE_SET_NETDEV_OPS(netdev, ops)	(netdev->netdev_ops = ops)
···
 #define for_all_rss_queues(adapter, rxo, i)				\
 	for (i = 0, rxo = &adapter->rx_obj[i+1]; i < (adapter->num_rx_qs - 1);\
 		i++, rxo++)
+
+#define for_all_tx_queues(adapter, txo, i)				\
+	for (i = 0, txo = &adapter->tx_obj[i]; i < adapter->num_tx_qs;	\
+		i++, txo++)
 
 #define PAGE_SHIFT_4K		12
 #define PAGE_SIZE_4K		(1 << PAGE_SHIFT_4K)

drivers/net/benet/be_ethtool.c

···
 	{NETSTAT_INFO(tx_errors)},
 	{NETSTAT_INFO(rx_dropped)},
 	{NETSTAT_INFO(tx_dropped)},
-	{DRVSTAT_TX_INFO(be_tx_rate)},
-	{DRVSTAT_TX_INFO(be_tx_reqs)},
-	{DRVSTAT_TX_INFO(be_tx_wrbs)},
-	{DRVSTAT_TX_INFO(be_tx_stops)},
-	{DRVSTAT_TX_INFO(be_tx_events)},
-	{DRVSTAT_TX_INFO(be_tx_compl)},
+	{DRVSTAT_INFO(be_tx_events)},
 	{DRVSTAT_INFO(rx_crc_errors)},
 	{DRVSTAT_INFO(rx_alignment_symbol_errors)},
 	{DRVSTAT_INFO(rx_pause_frames)},
···
 	{ERXSTAT_INFO(rx_drops_no_fragments)}
 };
 #define ETHTOOL_RXSTATS_NUM (ARRAY_SIZE(et_rx_stats))
+
+/* Stats related to multi TX queues */
+static const struct be_ethtool_stat et_tx_stats[] = {
+	{DRVSTAT_TX_INFO(be_tx_rate)},
+	{DRVSTAT_TX_INFO(be_tx_reqs)},
+	{DRVSTAT_TX_INFO(be_tx_wrbs)},
+	{DRVSTAT_TX_INFO(be_tx_stops)},
+	{DRVSTAT_TX_INFO(be_tx_compl)}
+};
+#define ETHTOOL_TXSTATS_NUM (ARRAY_SIZE(et_tx_stats))
 
 static const char et_self_tests[][ETH_GSTRING_LEN] = {
 	"MAC Loopback test",
···
 {
 	struct be_adapter *adapter = netdev_priv(netdev);
 	struct be_rx_obj *rxo;
+	struct be_tx_obj *txo;
 	void *p = NULL;
-	int i, j;
+	int i, j, base;
 
 	for (i = 0; i < ETHTOOL_STATS_NUM; i++) {
 		switch (et_stats[i].type) {
 		case NETSTAT:
 			p = &netdev->stats;
-			break;
-		case DRVSTAT_TX:
-			p = &adapter->tx_stats;
 			break;
 		case DRVSTAT:
 			p = &adapter->drv_stats;
···
 				*(u64 *)p: *(u32 *)p;
 	}
 
+	base = ETHTOOL_STATS_NUM;
 	for_all_rx_queues(adapter, rxo, j) {
 		for (i = 0; i < ETHTOOL_RXSTATS_NUM; i++) {
 			switch (et_rx_stats[i].type) {
···
 								rxo->q.id;
 				break;
 			}
-			data[ETHTOOL_STATS_NUM + j * ETHTOOL_RXSTATS_NUM + i] =
+			data[base + j * ETHTOOL_RXSTATS_NUM + i] =
 				(et_rx_stats[i].size == sizeof(u64)) ?
+					*(u64 *)p: *(u32 *)p;
+		}
+	}
+
+	base = ETHTOOL_STATS_NUM + adapter->num_rx_qs * ETHTOOL_RXSTATS_NUM;
+	for_all_tx_queues(adapter, txo, j) {
+		for (i = 0; i < ETHTOOL_TXSTATS_NUM; i++) {
+			p = (u8 *)&txo->stats + et_tx_stats[i].offset;
+			data[base + j * ETHTOOL_TXSTATS_NUM + i] =
+				(et_tx_stats[i].size == sizeof(u64)) ?
 					*(u64 *)p: *(u32 *)p;
 		}
 	}
···
 				data += ETH_GSTRING_LEN;
 			}
 		}
+		for (i = 0; i < adapter->num_tx_qs; i++) {
+			for (j = 0; j < ETHTOOL_TXSTATS_NUM; j++) {
+				sprintf(data, "txq%d: %s", i,
+					et_tx_stats[j].desc);
+				data += ETH_GSTRING_LEN;
+			}
+		}
 		break;
 	case ETH_SS_TEST:
 		for (i = 0; i < ETHTOOL_TESTS_NUM; i++) {
···
 		return ETHTOOL_TESTS_NUM;
 	case ETH_SS_STATS:
 		return ETHTOOL_STATS_NUM +
-			adapter->num_rx_qs * ETHTOOL_RXSTATS_NUM;
+			adapter->num_rx_qs * ETHTOOL_RXSTATS_NUM +
+			adapter->num_tx_qs * ETHTOOL_TXSTATS_NUM;
 	default:
 		return -EINVAL;
 	}
···
 	struct be_adapter *adapter = netdev_priv(netdev);
 
 	ring->rx_max_pending = adapter->rx_obj[0].q.len;
-	ring->tx_max_pending = adapter->tx_obj.q.len;
+	ring->tx_max_pending = adapter->tx_obj[0].q.len;
 
 	ring->rx_pending = atomic_read(&adapter->rx_obj[0].q.used);
-	ring->tx_pending = atomic_read(&adapter->tx_obj.q.used);
+	ring->tx_pending = atomic_read(&adapter->tx_obj[0].q.used);
 }
 
 static void

drivers/net/benet/be_main.c

···
 	struct be_drv_stats *drvs = &adapter->drv_stats;
 	struct net_device_stats *dev_stats = &adapter->netdev->stats;
 	struct be_rx_obj *rxo;
+	struct be_tx_obj *txo;
 	int i;
 
 	memset(dev_stats, 0, sizeof(*dev_stats));
···
 		}
 	}
 
-	dev_stats->tx_packets = tx_stats(adapter)->be_tx_pkts;
-	dev_stats->tx_bytes = tx_stats(adapter)->be_tx_bytes;
+	for_all_tx_queues(adapter, txo, i) {
+		dev_stats->tx_packets += tx_stats(txo)->be_tx_pkts;
+		dev_stats->tx_bytes += tx_stats(txo)->be_tx_bytes;
+	}
 
 	/* bad pkts received */
 	dev_stats->rx_errors = drvs->rx_crc_errors +
···
 	return rate;
 }
 
-static void be_tx_rate_update(struct be_adapter *adapter)
+static void be_tx_rate_update(struct be_tx_obj *txo)
 {
-	struct be_tx_stats *stats = tx_stats(adapter);
+	struct be_tx_stats *stats = tx_stats(txo);
 	ulong now = jiffies;
 
 	/* Wrapped around? */
···
 	}
 }
 
-static void be_tx_stats_update(struct be_adapter *adapter,
+static void be_tx_stats_update(struct be_tx_obj *txo,
 			u32 wrb_cnt, u32 copied, u32 gso_segs, bool stopped)
 {
-	struct be_tx_stats *stats = tx_stats(adapter);
+	struct be_tx_stats *stats = tx_stats(txo);
+
 	stats->be_tx_reqs++;
 	stats->be_tx_wrbs += wrb_cnt;
 	stats->be_tx_bytes += copied;
···
 	}
 }
 
-static int make_tx_wrbs(struct be_adapter *adapter,
+static int make_tx_wrbs(struct be_adapter *adapter, struct be_queue_info *txq,
 		struct sk_buff *skb, u32 wrb_cnt, bool dummy_wrb)
 {
 	dma_addr_t busaddr;
 	int i, copied = 0;
 	struct device *dev = &adapter->pdev->dev;
 	struct sk_buff *first_skb = skb;
-	struct be_queue_info *txq = &adapter->tx_obj.q;
 	struct be_eth_wrb *wrb;
 	struct be_eth_hdr_wrb *hdr;
 	bool map_single = false;
···
 			struct net_device *netdev)
 {
 	struct be_adapter *adapter = netdev_priv(netdev);
-	struct be_tx_obj *tx_obj = &adapter->tx_obj;
-	struct be_queue_info *txq = &tx_obj->q;
+	struct be_tx_obj *txo = &adapter->tx_obj[skb_get_queue_mapping(skb)];
+	struct be_queue_info *txq = &txo->q;
 	u32 wrb_cnt = 0, copied = 0;
 	u32 start = txq->head;
 	bool dummy_wrb, stopped = false;
 
 	wrb_cnt = wrb_cnt_for_skb(adapter, skb, &dummy_wrb);
 
-	copied = make_tx_wrbs(adapter, skb, wrb_cnt, dummy_wrb);
+	copied = make_tx_wrbs(adapter, txq, skb, wrb_cnt, dummy_wrb);
 	if (copied) {
 		/* record the sent skb in the sent_skb table */
-		BUG_ON(tx_obj->sent_skb_list[start]);
-		tx_obj->sent_skb_list[start] = skb;
+		BUG_ON(txo->sent_skb_list[start]);
+		txo->sent_skb_list[start] = skb;
 
 		/* Ensure txq has space for the next skb; Else stop the queue
 		 * *BEFORE* ringing the tx doorbell, so that we serialze the
···
 		atomic_add(wrb_cnt, &txq->used);
 		if ((BE_MAX_TX_FRAG_COUNT + atomic_read(&txq->used)) >=
 								txq->len) {
-			netif_stop_queue(netdev);
+			netif_stop_subqueue(netdev, skb_get_queue_mapping(skb));
 			stopped = true;
 		}
 
 		be_txq_notify(adapter, txq->id, wrb_cnt);
 
-		be_tx_stats_update(adapter, wrb_cnt, copied,
+		be_tx_stats_update(txo, wrb_cnt, copied,
 				skb_shinfo(skb)->gso_segs, stopped);
 	} else {
 		txq->head = start;
···
 	return txcp;
 }
 
-static u16 be_tx_compl_process(struct be_adapter *adapter, u16 last_index)
+static u16 be_tx_compl_process(struct be_adapter *adapter,
+		struct be_tx_obj *txo, u16 last_index)
 {
-	struct be_queue_info *txq = &adapter->tx_obj.q;
+	struct be_queue_info *txq = &txo->q;
 	struct be_eth_wrb *wrb;
-	struct sk_buff **sent_skbs = adapter->tx_obj.sent_skb_list;
+	struct sk_buff **sent_skbs = txo->sent_skb_list;
 	struct sk_buff *sent_skb;
 	u16 cur_index, num_wrbs = 1; /* account for hdr wrb */
 	bool unmap_skb_hdr = true;
···
 }
 
 static int event_handle(struct be_adapter *adapter,
-			struct be_eq_obj *eq_obj)
+			struct be_eq_obj *eq_obj,
+			bool rearm)
 {
 	struct be_eq_entry *eqe;
 	u16 num = 0;
···
 	/* Deal with any spurious interrupts that come
 	 * without events
 	 */
-	be_eq_notify(adapter, eq_obj->q.id, true, true, num);
+	if (!num)
+		rearm = true;
+
+	be_eq_notify(adapter, eq_obj->q.id, rearm, true, num);
 	if (num)
 		napi_schedule(&eq_obj->napi);
 
···
 	BUG_ON(atomic_read(&rxq->used));
 }
 
-static void be_tx_compl_clean(struct be_adapter *adapter)
+static void be_tx_compl_clean(struct be_adapter *adapter,
+				struct be_tx_obj *txo)
 {
-	struct be_queue_info *tx_cq = &adapter->tx_obj.cq;
-	struct be_queue_info *txq = &adapter->tx_obj.q;
+	struct be_queue_info *tx_cq = &txo->cq;
+	struct be_queue_info *txq = &txo->q;
 	struct be_eth_tx_compl *txcp;
 	u16 end_idx, cmpl = 0, timeo = 0, num_wrbs = 0;
-	struct sk_buff **sent_skbs = adapter->tx_obj.sent_skb_list;
+	struct sk_buff **sent_skbs = txo->sent_skb_list;
 	struct sk_buff *sent_skb;
 	bool dummy_wrb;
 
···
 		while ((txcp = be_tx_compl_get(tx_cq))) {
 			end_idx = AMAP_GET_BITS(struct amap_eth_tx_compl,
 					wrb_index, txcp);
-			num_wrbs += be_tx_compl_process(adapter, end_idx);
+			num_wrbs += be_tx_compl_process(adapter, txo, end_idx);
 			cmpl++;
 		}
 		if (cmpl) {
···
 		index_adv(&end_idx,
 			wrb_cnt_for_skb(adapter, sent_skb, &dummy_wrb) - 1,
 			txq->len);
-		num_wrbs = be_tx_compl_process(adapter, end_idx);
+		num_wrbs = be_tx_compl_process(adapter, txo, end_idx);
 		atomic_sub(num_wrbs, &txq->used);
 	}
 }
···
 static void be_tx_queues_destroy(struct be_adapter *adapter)
 {
 	struct be_queue_info *q;
+	struct be_tx_obj *txo;
+	u8 i;
 
-	q = &adapter->tx_obj.q;
-	if (q->created)
-		be_cmd_q_destroy(adapter, q, QTYPE_TXQ);
-	be_queue_free(adapter, q);
+	for_all_tx_queues(adapter, txo, i) {
+		q = &txo->q;
+		if (q->created)
+			be_cmd_q_destroy(adapter, q, QTYPE_TXQ);
+		be_queue_free(adapter, q);
 
-	q = &adapter->tx_obj.cq;
-	if (q->created)
-		be_cmd_q_destroy(adapter, q, QTYPE_CQ);
-	be_queue_free(adapter, q);
+		q = &txo->cq;
+		if (q->created)
+			be_cmd_q_destroy(adapter, q, QTYPE_CQ);
+		be_queue_free(adapter, q);
+	}
 
 	/* Clear any residual events */
 	be_eq_clean(adapter, &adapter->tx_eq);
···
 	be_queue_free(adapter, q);
 }
 
+/* One TX event queue is shared by all TX compl qs */
 static int be_tx_queues_create(struct be_adapter *adapter)
 {
 	struct be_queue_info *eq, *q, *cq;
+	struct be_tx_obj *txo;
+	u8 i;
 
 	adapter->tx_eq.max_eqd = 0;
 	adapter->tx_eq.min_eqd = 0;
 	adapter->tx_eq.cur_eqd = 96;
 	adapter->tx_eq.enable_aic = false;
-	/* Alloc Tx Event queue */
+
 	eq = &adapter->tx_eq.q;
-	if (be_queue_alloc(adapter, eq, EVNT_Q_LEN, sizeof(struct be_eq_entry)))
+	if (be_queue_alloc(adapter, eq, EVNT_Q_LEN,
+		sizeof(struct be_eq_entry)))
 		return -1;
 
-	/* Ask BE to create Tx Event queue */
 	if (be_cmd_eq_create(adapter, eq, adapter->tx_eq.cur_eqd))
-		goto tx_eq_free;
-
+		goto err;
 	adapter->tx_eq.eq_idx = adapter->eq_next_idx++;
 
-
-	/* Alloc TX eth compl queue */
-	cq = &adapter->tx_obj.cq;
-	if (be_queue_alloc(adapter, cq, TX_CQ_LEN,
+	for_all_tx_queues(adapter, txo, i) {
+		cq = &txo->cq;
+		if (be_queue_alloc(adapter, cq, TX_CQ_LEN,
 			sizeof(struct be_eth_tx_compl)))
-		goto tx_eq_destroy;
+			goto err;
 
-	/* Ask BE to create Tx eth compl queue */
-	if (be_cmd_cq_create(adapter, cq, eq, false, false, 3))
-		goto tx_cq_free;
+		if (be_cmd_cq_create(adapter, cq, eq, false, false, 3))
+			goto err;
 
-	/* Alloc TX eth queue */
-	q = &adapter->tx_obj.q;
-	if (be_queue_alloc(adapter, q, TX_Q_LEN, sizeof(struct be_eth_wrb)))
-		goto tx_cq_destroy;
+		q = &txo->q;
+		if (be_queue_alloc(adapter, q, TX_Q_LEN,
+			sizeof(struct be_eth_wrb)))
+			goto err;
 
-	/* Ask BE to create Tx eth queue */
-	if (be_cmd_txq_create(adapter, q, cq))
-		goto tx_q_free;
+		if (be_cmd_txq_create(adapter, q, cq))
+			goto err;
+	}
 	return 0;
 
-tx_q_free:
-	be_queue_free(adapter, q);
-tx_cq_destroy:
-	be_cmd_q_destroy(adapter, cq, QTYPE_CQ);
-tx_cq_free:
-	be_queue_free(adapter, cq);
-tx_eq_destroy:
-	be_cmd_q_destroy(adapter, eq, QTYPE_EQ);
-tx_eq_free:
-	be_queue_free(adapter, eq);
+err:
+	be_tx_queues_destroy(adapter);
 	return -1;
 }
 
···
 
 	if (lancer_chip(adapter)) {
 		if (event_peek(&adapter->tx_eq))
-			tx = event_handle(adapter, &adapter->tx_eq);
+			tx = event_handle(adapter, &adapter->tx_eq, false);
 		for_all_rx_queues(adapter, rxo, i) {
 			if (event_peek(&rxo->rx_eq))
-				rx |= event_handle(adapter, &rxo->rx_eq);
+				rx |= event_handle(adapter, &rxo->rx_eq, true);
 		}
 
 		if (!(tx || rx))
···
 			return IRQ_NONE;
 
 		if ((1 << adapter->tx_eq.eq_idx & isr))
-			event_handle(adapter, &adapter->tx_eq);
+			event_handle(adapter, &adapter->tx_eq, false);
 
 		for_all_rx_queues(adapter, rxo, i) {
 			if ((1 << rxo->rx_eq.eq_idx & isr))
-				event_handle(adapter, &rxo->rx_eq);
+				event_handle(adapter, &rxo->rx_eq, true);
 		}
 	}
 
···
 	struct be_rx_obj *rxo = dev;
 	struct be_adapter *adapter = rxo->adapter;
 
-	event_handle(adapter, &rxo->rx_eq);
+	event_handle(adapter, &rxo->rx_eq, true);
 
 	return IRQ_HANDLED;
 }
···
 {
 	struct be_adapter *adapter = dev;
 
-	event_handle(adapter, &adapter->tx_eq);
+	event_handle(adapter, &adapter->tx_eq, false);
 
 	return IRQ_HANDLED;
 }
···
 	struct be_eq_obj *tx_eq = container_of(napi, struct be_eq_obj, napi);
 	struct be_adapter *adapter =
 		container_of(tx_eq, struct be_adapter, tx_eq);
-	struct be_queue_info *txq = &adapter->tx_obj.q;
-	struct be_queue_info *tx_cq = &adapter->tx_obj.cq;
+	struct be_tx_obj *txo;
 	struct be_eth_tx_compl *txcp;
-	int tx_compl = 0, mcc_compl, status = 0;
-	u16 end_idx, num_wrbs = 0;
+	int tx_compl, mcc_compl, status = 0;
+	u8 i;
+	u16 num_wrbs;
 
-	while ((txcp = be_tx_compl_get(tx_cq))) {
-		end_idx = AMAP_GET_BITS(struct amap_eth_tx_compl,
-				wrb_index, txcp);
-		num_wrbs += be_tx_compl_process(adapter, end_idx);
-		tx_compl++;
+	for_all_tx_queues(adapter, txo, i) {
+		tx_compl = 0;
+		num_wrbs = 0;
+		while ((txcp = be_tx_compl_get(&txo->cq))) {
+			num_wrbs += be_tx_compl_process(adapter, txo,
+				AMAP_GET_BITS(struct amap_eth_tx_compl,
+					wrb_index, txcp));
+			tx_compl++;
+		}
+		if (tx_compl) {
+			be_cq_notify(adapter, txo->cq.id, true, tx_compl);
+
+			atomic_sub(num_wrbs, &txo->q.used);
+
+			/* As Tx wrbs have been freed up, wake up netdev queue
+			 * if it was stopped due to lack of tx wrbs.  */
+			if (__netif_subqueue_stopped(adapter->netdev, i) &&
+				atomic_read(&txo->q.used) < txo->q.len / 2) {
+				netif_wake_subqueue(adapter->netdev, i);
+			}
+
+			adapter->drv_stats.be_tx_events++;
+			txo->stats.be_tx_compl += tx_compl;
+		}
 	}
 
 	mcc_compl = be_process_mcc(adapter, &status);
-
-	napi_complete(napi);
 
 	if (mcc_compl) {
 		struct be_mcc_obj *mcc_obj = &adapter->mcc_obj;
 		be_cq_notify(adapter, mcc_obj->cq.id, true, mcc_compl);
 	}
 
-	if (tx_compl) {
-		be_cq_notify(adapter, adapter->tx_obj.cq.id, true, tx_compl);
+	napi_complete(napi);
 
-		atomic_sub(num_wrbs, &txq->used);
-
-		/* As Tx wrbs have been freed up, wake up netdev queue if
-		 * it was stopped due to lack of tx wrbs.
-		 */
-		if (netif_queue_stopped(adapter->netdev) &&
-			atomic_read(&txq->used) < txq->len / 2) {
-			netif_wake_queue(adapter->netdev);
-		}
-
-		tx_stats(adapter)->be_tx_events++;
-		tx_stats(adapter)->be_tx_compl += tx_compl;
-	}
-
+	be_eq_notify(adapter, tx_eq->q.id, true, false, 0);
 	return 1;
 }
 
···
 	struct be_adapter *adapter =
 		container_of(work, struct be_adapter, work.work);
 	struct be_rx_obj *rxo;
+	struct be_tx_obj *txo;
 	int i;
 
 	if (!adapter->ue_detected && !lancer_chip(adapter))
···
 		else
 			be_cmd_get_stats(adapter, &adapter->stats_cmd);
 	}
-	be_tx_rate_update(adapter);
+
+	for_all_tx_queues(adapter, txo, i)
+		be_tx_rate_update(txo);
 
 	for_all_rx_queues(adapter, rxo, i) {
 		be_rx_rate_update(rxo);
···
 {
 	struct be_adapter *adapter = netdev_priv(netdev);
 	struct be_rx_obj *rxo;
+	struct be_tx_obj *txo;
 	struct be_eq_obj *tx_eq = &adapter->tx_eq;
 	int vec, i;
 
···
 	napi_disable(&tx_eq->napi);
 
 	if (lancer_chip(adapter)) {
-		be_cq_notify(adapter, adapter->tx_obj.cq.id, false, 0);
 		be_cq_notify(adapter, adapter->mcc_obj.cq.id, false, 0);
 		for_all_rx_queues(adapter, rxo, i)
 			 be_cq_notify(adapter, rxo->cq.id, false, 0);
+		for_all_tx_queues(adapter, txo, i)
+			 be_cq_notify(adapter, txo->cq.id, false, 0);
 	}
 
 	if (msix_enabled(adapter)) {
···
 	/* Wait for all pending tx completions to arrive so that
 	 * all tx skbs are freed.
 	 */
-	be_tx_compl_clean(adapter);
+	for_all_tx_queues(adapter, txo, i)
+		be_tx_compl_clean(adapter, txo);
 
 	return 0;
 }
···
 		return status;
 
 	be_cmd_check_native_mode(adapter);
+
+	if ((num_vfs && adapter->sriov_enabled) ||
+		(adapter->function_mode & 0x400) ||
+		lancer_chip(adapter) || !be_physfn(adapter)) {
+		adapter->num_tx_qs = 1;
+		netif_set_real_num_tx_queues(adapter->netdev,
+			adapter->num_tx_qs);
+	} else {
+		adapter->num_tx_qs = MAX_TX_QS;
+	}
+
 	return 0;
 }
 
···
 		goto disable_dev;
 	pci_set_master(pdev);
 
-	netdev = alloc_etherdev(sizeof(struct be_adapter));
+	netdev = alloc_etherdev_mq(sizeof(struct be_adapter), MAX_TX_QS);
 	if (netdev == NULL) {
 		status = -ENOMEM;
 		goto rel_reg;