Merge commit 'v2.6.30-rc5' into next

11 files changed, 792 insertions, 349 deletions

diff --git a/drivers/net/cxgb3/adapter.h b/drivers/net/cxgb3/adapter.h
index a89d8cc51205..714df2b675e6 100644
--- a/drivers/net/cxgb3/adapter.h
+++ b/drivers/net/cxgb3/adapter.h
@@ -42,7 +42,6 @@
 #include <linux/cache.h>
 #include <linux/mutex.h>
 #include <linux/bitops.h>
-#include <linux/inet_lro.h>
 #include "t3cdev.h"
 #include <asm/io.h>
 
@@ -69,6 +68,8 @@ struct port_info {
 	struct net_device_stats netstats;
 	int activity;
 	__be32 iscsi_ipv4addr;
+
+	int link_fault; /* link fault was detected */
 };
 
 enum {				/* adapter flags */
@@ -84,6 +85,8 @@ struct fl_pg_chunk {
 	struct page *page;
 	void *va;
 	unsigned int offset;
+	u64 *p_cnt;
+	DECLARE_PCI_UNMAP_ADDR(mapping);
 };
 
 struct rx_desc;
@@ -92,6 +95,7 @@ struct rx_sw_desc;
 struct sge_fl {                     /* SGE per free-buffer list state */
 	unsigned int buf_size;      /* size of each Rx buffer */
 	unsigned int credits;       /* # of available Rx buffers */
+	unsigned int pend_cred;     /* new buffers since last FL DB ring */
 	unsigned int size;          /* capacity of free list */
 	unsigned int cidx;          /* consumer index */
 	unsigned int pidx;          /* producer index */
@@ -99,6 +103,7 @@ struct sge_fl {                     /* SGE per free-buffer list state */
 	struct fl_pg_chunk pg_chunk;/* page chunk cache */
 	unsigned int use_pages;     /* whether FL uses pages or sk_buffs */
 	unsigned int order;	    /* order of page allocations */
+	unsigned int alloc_size;    /* size of allocated buffer */
 	struct rx_desc *desc;       /* address of HW Rx descriptor ring */
 	struct rx_sw_desc *sdesc;   /* address of SW Rx descriptor ring */
 	dma_addr_t   phys_addr;     /* physical address of HW ring start */
@@ -178,15 +183,11 @@ enum {				/* per port SGE statistics */
 	SGE_PSTAT_TX_CSUM,	/* # of TX checksum offloads */
 	SGE_PSTAT_VLANEX,	/* # of VLAN tag extractions */
 	SGE_PSTAT_VLANINS,	/* # of VLAN tag insertions */
-	SGE_PSTAT_LRO_AGGR,	/* # of page chunks added to LRO sessions */
-	SGE_PSTAT_LRO_FLUSHED,	/* # of flushed LRO sessions */
-	SGE_PSTAT_LRO_NO_DESC,	/* # of overflown LRO sessions */
 
 	SGE_PSTAT_MAX		/* must be last */
 };
 
-#define T3_MAX_LRO_SES 8
-#define T3_MAX_LRO_MAX_PKTS 64
+struct napi_gro_fraginfo;
 
 struct sge_qset {		/* an SGE queue set */
 	struct adapter *adap;
@@ -194,17 +195,14 @@ struct sge_qset {		/* an SGE queue set */
 	struct sge_rspq rspq;
 	struct sge_fl fl[SGE_RXQ_PER_SET];
 	struct sge_txq txq[SGE_TXQ_PER_SET];
-	struct net_lro_mgr lro_mgr;
-	struct net_lro_desc lro_desc[T3_MAX_LRO_SES];
-	struct skb_frag_struct *lro_frag_tbl;
-	int lro_nfrags;
+	struct napi_gro_fraginfo lro_frag_tbl;
 	int lro_enabled;
-	int lro_frag_len;
 	void *lro_va;
 	struct net_device *netdev;
 	struct netdev_queue *tx_q;	/* associated netdev TX queue */
 	unsigned long txq_stopped;	/* which Tx queues are stopped */
 	struct timer_list tx_reclaim_timer;	/* reclaims TX buffers */
+	struct timer_list rx_reclaim_timer;	/* reclaims RX buffers */
 	unsigned long port_stats[SGE_PSTAT_MAX];
 } ____cacheline_aligned;
 
@@ -230,6 +228,7 @@ struct adapter {
 	unsigned int slow_intr_mask;
 	unsigned long irq_stats[IRQ_NUM_STATS];
 
+	int msix_nvectors;
 	struct {
 		unsigned short vec;
 		char desc[22];
@@ -247,6 +246,7 @@ struct adapter {
 	struct delayed_work adap_check_task;
 	struct work_struct ext_intr_handler_task;
 	struct work_struct fatal_error_handler_task;
+	struct work_struct link_fault_handler_task;
 
 	struct dentry *debugfs_root;
 
@@ -289,9 +289,12 @@ void t3_os_ext_intr_handler(struct adapter *adapter);
 void t3_os_link_changed(struct adapter *adapter, int port_id, int link_status,
 			int speed, int duplex, int fc);
 void t3_os_phymod_changed(struct adapter *adap, int port_id);
+void t3_os_link_fault(struct adapter *adapter, int port_id, int state);
+void t3_os_link_fault_handler(struct adapter *adapter, int port_id);
 
 void t3_sge_start(struct adapter *adap);
 void t3_sge_stop(struct adapter *adap);
+void t3_start_sge_timers(struct adapter *adap);
 void t3_stop_sge_timers(struct adapter *adap);
 void t3_free_sge_resources(struct adapter *adap);
 void t3_sge_err_intr_handler(struct adapter *adapter);
diff --git a/drivers/net/cxgb3/ael1002.c b/drivers/net/cxgb3/ael1002.c
index 5c3c05da4d96..e1b22490ff59 100644
--- a/drivers/net/cxgb3/ael1002.c
+++ b/drivers/net/cxgb3/ael1002.c
@@ -1005,7 +1005,8 @@ static int ael2005_reset(struct cphy *phy, int wait)
 		{ 0, 0, 0, 0 }
 	};
 
-	int err, lasi_ctrl;
+	int err;
+	unsigned int lasi_ctrl;
 
 	err = mdio_read(phy, MDIO_DEV_PMA_PMD, LASI_CTRL, &lasi_ctrl);
 	if (err)
diff --git a/drivers/net/cxgb3/common.h b/drivers/net/cxgb3/common.h
index db4f4f575b6a..e508dc32f3ec 100644
--- a/drivers/net/cxgb3/common.h
+++ b/drivers/net/cxgb3/common.h
@@ -191,7 +191,8 @@ struct mdio_ops {
 };
 
 struct adapter_info {
-	unsigned char nports;	/* # of ports */
+	unsigned char nports0;        /* # of ports on channel 0 */
+	unsigned char nports1;        /* # of ports on channel 1 */
 	unsigned char phy_base_addr;	/* MDIO PHY base address */
 	unsigned int gpio_out;	/* GPIO output settings */
 	unsigned char gpio_intr[MAX_NPORTS]; /* GPIO PHY IRQ pins */
@@ -280,6 +281,7 @@ struct mac_stats {
 	unsigned long num_toggled; /* # times toggled TxEn due to stuck TX */
 	unsigned long num_resets;  /* # times reset due to stuck TX */
 
+	unsigned long link_faults;  /* # detected link faults */
 };
 
 struct tp_mib_stats {
@@ -421,6 +423,7 @@ struct adapter_params {
 	unsigned short b_wnd[NCCTRL_WIN];
 
 	unsigned int nports;	/* # of ethernet ports */
+	unsigned int chan_map;  /* bitmap of in-use Tx channels */
 	unsigned int stats_update_period;	/* MAC stats accumulation period */
 	unsigned int linkpoll_period;	/* link poll period in 0.1s */
 	unsigned int rev;	/* chip revision */
@@ -701,6 +704,8 @@ int t3_phy_lasi_intr_handler(struct cphy *phy);
 void t3_intr_enable(struct adapter *adapter);
 void t3_intr_disable(struct adapter *adapter);
 void t3_intr_clear(struct adapter *adapter);
+void t3_xgm_intr_enable(struct adapter *adapter, int idx);
+void t3_xgm_intr_disable(struct adapter *adapter, int idx);
 void t3_port_intr_enable(struct adapter *adapter, int idx);
 void t3_port_intr_disable(struct adapter *adapter, int idx);
 void t3_port_intr_clear(struct adapter *adapter, int idx);
@@ -708,6 +713,7 @@ int t3_slow_intr_handler(struct adapter *adapter);
 int t3_phy_intr_handler(struct adapter *adapter);
 
 void t3_link_changed(struct adapter *adapter, int port_id);
+void t3_link_fault(struct adapter *adapter, int port_id);
 int t3_link_start(struct cphy *phy, struct cmac *mac, struct link_config *lc);
 const struct adapter_info *t3_get_adapter_info(unsigned int board_id);
 int t3_seeprom_read(struct adapter *adapter, u32 addr, __le32 *data);
@@ -744,6 +750,8 @@ int t3_mc7_bd_read(struct mc7 *mc7, unsigned int start, unsigned int n,
 
 int t3_mac_reset(struct cmac *mac);
 void t3b_pcs_reset(struct cmac *mac);
+void t3_mac_disable_exact_filters(struct cmac *mac);
+void t3_mac_enable_exact_filters(struct cmac *mac);
 int t3_mac_enable(struct cmac *mac, int which);
 int t3_mac_disable(struct cmac *mac, int which);
 int t3_mac_set_mtu(struct cmac *mac, unsigned int mtu);
diff --git a/drivers/net/cxgb3/cxgb3_main.c b/drivers/net/cxgb3/cxgb3_main.c
index bab8a934c33d..7ea48414c6cb 100644
--- a/drivers/net/cxgb3/cxgb3_main.c
+++ b/drivers/net/cxgb3/cxgb3_main.c
@@ -170,6 +170,40 @@ static void link_report(struct net_device *dev)
 	}
 }
 
+void t3_os_link_fault(struct adapter *adap, int port_id, int state)
+{
+	struct net_device *dev = adap->port[port_id];
+	struct port_info *pi = netdev_priv(dev);
+
+	if (state == netif_carrier_ok(dev))
+		return;
+
+	if (state) {
+		struct cmac *mac = &pi->mac;
+
+		netif_carrier_on(dev);
+
+		/* Clear local faults */
+		t3_xgm_intr_disable(adap, pi->port_id);
+		t3_read_reg(adap, A_XGM_INT_STATUS +
+				    pi->mac.offset);
+		t3_write_reg(adap,
+			     A_XGM_INT_CAUSE + pi->mac.offset,
+			     F_XGM_INT);
+
+		t3_set_reg_field(adap,
+				 A_XGM_INT_ENABLE +
+				 pi->mac.offset,
+				 F_XGM_INT, F_XGM_INT);
+		t3_xgm_intr_enable(adap, pi->port_id);
+
+		t3_mac_enable(mac, MAC_DIRECTION_TX);
+	} else
+		netif_carrier_off(dev);
+
+	link_report(dev);
+}
+
 /**
  *	t3_os_link_changed - handle link status changes
  *	@adapter: the adapter associated with the link change
@@ -197,10 +231,34 @@ void t3_os_link_changed(struct adapter *adapter, int port_id, int link_stat,
 	if (link_stat != netif_carrier_ok(dev)) {
 		if (link_stat) {
 			t3_mac_enable(mac, MAC_DIRECTION_RX);
+
+			/* Clear local faults */
+			t3_xgm_intr_disable(adapter, pi->port_id);
+			t3_read_reg(adapter, A_XGM_INT_STATUS +
+				    pi->mac.offset);
+			t3_write_reg(adapter,
+				     A_XGM_INT_CAUSE + pi->mac.offset,
+				     F_XGM_INT);
+
+			t3_set_reg_field(adapter,
+					 A_XGM_INT_ENABLE + pi->mac.offset,
+					 F_XGM_INT, F_XGM_INT);
+			t3_xgm_intr_enable(adapter, pi->port_id);
+
 			netif_carrier_on(dev);
 		} else {
 			netif_carrier_off(dev);
-			pi->phy.ops->power_down(&pi->phy, 1);
+
+			t3_xgm_intr_disable(adapter, pi->port_id);
+			t3_read_reg(adapter, A_XGM_INT_STATUS + pi->mac.offset);
+			t3_set_reg_field(adapter,
+					 A_XGM_INT_ENABLE + pi->mac.offset,
+					 F_XGM_INT, 0);
+
+			if (is_10G(adapter))
+				pi->phy.ops->power_down(&pi->phy, 1);
+
+			t3_read_reg(adapter, A_XGM_INT_STATUS + pi->mac.offset);
 			t3_mac_disable(mac, MAC_DIRECTION_RX);
 			t3_link_start(&pi->phy, mac, &pi->link_config);
 		}
@@ -339,7 +397,7 @@ static void free_irq_resources(struct adapter *adapter)
 
 		free_irq(adapter->msix_info[0].vec, adapter);
 		for_each_port(adapter, i)
-		    n += adap2pinfo(adapter, i)->nqsets;
+			n += adap2pinfo(adapter, i)->nqsets;
 
 		for (i = 0; i < n; ++i)
 			free_irq(adapter->msix_info[i + 1].vec,
@@ -509,19 +567,9 @@ static void set_qset_lro(struct net_device *dev, int qset_idx, int val)
 {
 	struct port_info *pi = netdev_priv(dev);
 	struct adapter *adapter = pi->adapter;
-	int i, lro_on = 1;
 
 	adapter->params.sge.qset[qset_idx].lro = !!val;
 	adapter->sge.qs[qset_idx].lro_enabled = !!val;
-
-	/* let ethtool report LRO on only if all queues are LRO enabled */
-	for (i = pi->first_qset; i < pi->first_qset + pi->nqsets; ++i)
-		lro_on &= adapter->params.sge.qset[i].lro;
-
-	if (lro_on)
-		dev->features |= NETIF_F_LRO;
-	else
-		dev->features &= ~NETIF_F_LRO;
 }
 
 /**
@@ -554,7 +602,6 @@ static int setup_sge_qsets(struct adapter *adap)
 				&adap->params.sge.qset[qset_idx], ntxq, dev,
 				netdev_get_tx_queue(dev, j));
 			if (err) {
-				t3_stop_sge_timers(adap);
 				t3_free_sge_resources(adap);
 				return err;
 			}
@@ -998,6 +1045,8 @@ static int cxgb_up(struct adapter *adap)
 		setup_rss(adap);
 		if (!(adap->flags & NAPI_INIT))
 			init_napi(adap);
+
+		t3_start_sge_timers(adap);
 		adap->flags |= FULL_INIT_DONE;
 	}
 
@@ -1068,8 +1117,8 @@ static void cxgb_down(struct adapter *adapter)
 	spin_unlock_irq(&adapter->work_lock);
 
 	free_irq_resources(adapter);
-	flush_workqueue(cxgb3_wq);	/* wait for external IRQ handler */
 	quiesce_rx(adapter);
+	flush_workqueue(cxgb3_wq);	/* wait for external IRQ handler */
 }
 
 static void schedule_chk_task(struct adapter *adap)
@@ -1138,6 +1187,9 @@ static int offload_close(struct t3cdev *tdev)
 
 	sysfs_remove_group(&tdev->lldev->dev.kobj, &offload_attr_group);
 
+	/* Flush work scheduled while releasing TIDs */
+	flush_scheduled_work();
+
 	tdev->lldev = NULL;
 	cxgb3_set_dummy_ops(tdev);
 	t3_tp_set_offload_mode(adapter, 0);
@@ -1183,6 +1235,14 @@ static int cxgb_close(struct net_device *dev)
 	struct port_info *pi = netdev_priv(dev);
 	struct adapter *adapter = pi->adapter;
 
+	
+	if (!adapter->open_device_map)
+		return 0;
+
+	/* Stop link fault interrupts */
+	t3_xgm_intr_disable(adapter, pi->port_id);
+	t3_read_reg(adapter, A_XGM_INT_STATUS + pi->mac.offset);
+
 	t3_port_intr_disable(adapter, pi->port_id);
 	netif_tx_stop_all_queues(dev);
 	pi->phy.ops->power_down(&pi->phy, 1);
@@ -1194,8 +1254,7 @@ static int cxgb_close(struct net_device *dev)
 	spin_unlock_irq(&adapter->work_lock);
 
 	if (!(adapter->open_device_map & PORT_MASK))
-		cancel_rearming_delayed_workqueue(cxgb3_wq,
-						  &adapter->adap_check_task);
+		cancel_delayed_work_sync(&adapter->adap_check_task);
 
 	if (!adapter->open_device_map)
 		cxgb_down(adapter);
@@ -1309,6 +1368,7 @@ static char stats_strings[][ETH_GSTRING_LEN] = {
 	"CheckTXEnToggled   ",
 	"CheckResets        ",
 
+	"LinkFaults         ",
 };
 
 static int get_sset_count(struct net_device *dev, int sset)
@@ -1434,13 +1494,15 @@ static void get_stats(struct net_device *dev, struct ethtool_stats *stats,
 	*data++ = collect_sge_port_stats(adapter, pi, SGE_PSTAT_VLANINS);
 	*data++ = collect_sge_port_stats(adapter, pi, SGE_PSTAT_TX_CSUM);
 	*data++ = collect_sge_port_stats(adapter, pi, SGE_PSTAT_RX_CSUM_GOOD);
-	*data++ = collect_sge_port_stats(adapter, pi, SGE_PSTAT_LRO_AGGR);
-	*data++ = collect_sge_port_stats(adapter, pi, SGE_PSTAT_LRO_FLUSHED);
-	*data++ = collect_sge_port_stats(adapter, pi, SGE_PSTAT_LRO_NO_DESC);
+	*data++ = 0;
+	*data++ = 0;
+	*data++ = 0;
 	*data++ = s->rx_cong_drops;
 
 	*data++ = s->num_toggled;
 	*data++ = s->num_resets;
+
+	*data++ = s->link_faults;
 }
 
 static inline void reg_block_dump(struct adapter *ap, void *buf,
@@ -1576,7 +1638,6 @@ static int speed_duplex_to_caps(int speed, int duplex)
 
 static int set_settings(struct net_device *dev, struct ethtool_cmd *cmd)
 {
-	int cap;
 	struct port_info *p = netdev_priv(dev);
 	struct link_config *lc = &p->link_config;
 
@@ -1586,7 +1647,7 @@ static int set_settings(struct net_device *dev, struct ethtool_cmd *cmd)
 		 * being requested.
 		 */
 		if (cmd->autoneg == AUTONEG_DISABLE) {
-			cap = speed_duplex_to_caps(cmd->speed, cmd->duplex);
+			int cap = speed_duplex_to_caps(cmd->speed, cmd->duplex);
 			if (lc->supported & cap)
 				return 0;
 		}
@@ -1827,28 +1888,6 @@ static void get_wol(struct net_device *dev, struct ethtool_wolinfo *wol)
 	memset(&wol->sopass, 0, sizeof(wol->sopass));
 }
 
-static int cxgb3_set_flags(struct net_device *dev, u32 data)
-{
-	struct port_info *pi = netdev_priv(dev);
-	int i;
-
-	if (data & ETH_FLAG_LRO) {
-		if (!(pi->rx_offload & T3_RX_CSUM))
-			return -EINVAL;
-
-		pi->rx_offload |= T3_LRO;
-		for (i = pi->first_qset; i < pi->first_qset + pi->nqsets; i++)
-			set_qset_lro(dev, i, 1);
-
-	} else {
-		pi->rx_offload &= ~T3_LRO;
-		for (i = pi->first_qset; i < pi->first_qset + pi->nqsets; i++)
-			set_qset_lro(dev, i, 0);
-	}
-
-	return 0;
-}
-
 static const struct ethtool_ops cxgb_ethtool_ops = {
 	.get_settings = get_settings,
 	.set_settings = set_settings,
@@ -1878,8 +1917,6 @@ static const struct ethtool_ops cxgb_ethtool_ops = {
 	.get_regs = get_regs,
 	.get_wol = get_wol,
 	.set_tso = ethtool_op_set_tso,
-	.get_flags = ethtool_op_get_flags,
-	.set_flags = cxgb3_set_flags,
 };
 
 static int in_range(int val, int lo, int hi)
@@ -2460,8 +2497,21 @@ static void check_link_status(struct adapter *adapter)
 		struct net_device *dev = adapter->port[i];
 		struct port_info *p = netdev_priv(dev);
 
-		if (!(p->phy.caps & SUPPORTED_IRQ) && netif_running(dev))
+		spin_lock_irq(&adapter->work_lock);
+		if (p->link_fault) {
+			t3_link_fault(adapter, i);
+			spin_unlock_irq(&adapter->work_lock);
+			continue;
+		}
+		spin_unlock_irq(&adapter->work_lock);
+
+		if (!(p->phy.caps & SUPPORTED_IRQ) && netif_running(dev)) {
+			t3_xgm_intr_disable(adapter, i);
+			t3_read_reg(adapter, A_XGM_INT_STATUS + p->mac.offset);
+
 			t3_link_changed(adapter, i);
+			t3_xgm_intr_enable(adapter, i);
+		}
 	}
 }
 
@@ -2506,12 +2556,12 @@ static void t3_adap_check_task(struct work_struct *work)
 	struct adapter *adapter = container_of(work, struct adapter,
 					       adap_check_task.work);
 	const struct adapter_params *p = &adapter->params;
+	int port;
+	unsigned int v, status, reset;
 
 	adapter->check_task_cnt++;
 
-	/* Check link status for PHYs without interrupts */
-	if (p->linkpoll_period)
-		check_link_status(adapter);
+	check_link_status(adapter);
 
 	/* Accumulate MAC stats if needed */
 	if (!p->linkpoll_period ||
@@ -2524,6 +2574,54 @@ static void t3_adap_check_task(struct work_struct *work)
 	if (p->rev == T3_REV_B2)
 		check_t3b2_mac(adapter);
 
+	/*
+	 * Scan the XGMAC's to check for various conditions which we want to
+	 * monitor in a periodic polling manner rather than via an interrupt
+	 * condition.  This is used for conditions which would otherwise flood
+	 * the system with interrupts and we only really need to know that the
+	 * conditions are "happening" ...  For each condition we count the
+	 * detection of the condition and reset it for the next polling loop.
+	 */
+	for_each_port(adapter, port) {
+		struct cmac *mac =  &adap2pinfo(adapter, port)->mac;
+		u32 cause;
+
+		cause = t3_read_reg(adapter, A_XGM_INT_CAUSE + mac->offset);
+		reset = 0;
+		if (cause & F_RXFIFO_OVERFLOW) {
+			mac->stats.rx_fifo_ovfl++;
+			reset |= F_RXFIFO_OVERFLOW;
+		}
+
+		t3_write_reg(adapter, A_XGM_INT_CAUSE + mac->offset, reset);
+	}
+
+	/*
+	 * We do the same as above for FL_EMPTY interrupts.
+	 */
+	status = t3_read_reg(adapter, A_SG_INT_CAUSE);
+	reset = 0;
+
+	if (status & F_FLEMPTY) {
+		struct sge_qset *qs = &adapter->sge.qs[0];
+		int i = 0;
+
+		reset |= F_FLEMPTY;
+
+		v = (t3_read_reg(adapter, A_SG_RSPQ_FL_STATUS) >> S_FL0EMPTY) &
+		    0xffff;
+
+		while (v) {
+			qs->fl[i].empty += (v & 1);
+			if (i)
+				qs++;
+			i ^= 1;
+			v >>= 1;
+		}
+	}
+
+	t3_write_reg(adapter, A_SG_INT_CAUSE, reset);
+
 	/* Schedule the next check update if any port is active. */
 	spin_lock_irq(&adapter->work_lock);
 	if (adapter->open_device_map & PORT_MASK)
@@ -2538,9 +2636,23 @@ static void ext_intr_task(struct work_struct *work)
 {
 	struct adapter *adapter = container_of(work, struct adapter,
 					       ext_intr_handler_task);
+	int i;
+
+	/* Disable link fault interrupts */
+	for_each_port(adapter, i) {
+		struct net_device *dev = adapter->port[i];
+		struct port_info *p = netdev_priv(dev);
+
+		t3_xgm_intr_disable(adapter, i);
+		t3_read_reg(adapter, A_XGM_INT_STATUS + p->mac.offset);
+	}
 
+	/* Re-enable link fault interrupts */
 	t3_phy_intr_handler(adapter);
 
+	for_each_port(adapter, i)
+		t3_xgm_intr_enable(adapter, i);
+
 	/* Now reenable external interrupts */
 	spin_lock_irq(&adapter->work_lock);
 	if (adapter->slow_intr_mask) {
@@ -2573,10 +2685,26 @@ void t3_os_ext_intr_handler(struct adapter *adapter)
 	spin_unlock(&adapter->work_lock);
 }
 
+void t3_os_link_fault_handler(struct adapter *adapter, int port_id)
+{
+	struct net_device *netdev = adapter->port[port_id];
+	struct port_info *pi = netdev_priv(netdev);
+
+	spin_lock(&adapter->work_lock);
+	pi->link_fault = 1;
+	spin_unlock(&adapter->work_lock);
+}
+
 static int t3_adapter_error(struct adapter *adapter, int reset)
 {
 	int i, ret = 0;
 
+	if (is_offload(adapter) &&
+	    test_bit(OFFLOAD_DEVMAP_BIT, &adapter->open_device_map)) {
+		cxgb3_err_notify(&adapter->tdev, OFFLOAD_STATUS_DOWN, 0);
+		offload_close(&adapter->tdev);
+	}
+
 	/* Stop all ports */
 	for_each_port(adapter, i) {
 		struct net_device *netdev = adapter->port[i];
@@ -2585,10 +2713,6 @@ static int t3_adapter_error(struct adapter *adapter, int reset)
 			cxgb_close(netdev);
 	}
 
-	if (is_offload(adapter) &&
-	    test_bit(OFFLOAD_DEVMAP_BIT, &adapter->open_device_map))
-		offload_close(&adapter->tdev);
-
 	/* Stop SGE timers */
 	t3_stop_sge_timers(adapter);
 
@@ -2640,6 +2764,9 @@ static void t3_resume_ports(struct adapter *adapter)
 			}
 		}
 	}
+
+	if (is_offload(adapter) && !ofld_disable)
+		cxgb3_err_notify(&adapter->tdev, OFFLOAD_STATUS_UP, 0);
 }
 
 /*
@@ -2684,7 +2811,6 @@ void t3_fatal_err(struct adapter *adapter)
 		CH_ALERT(adapter, "FW status: 0x%x, 0x%x, 0x%x, 0x%x\n",
 			 fw_status[0], fw_status[1],
 			 fw_status[2], fw_status[3]);
-
 }
 
 /**
@@ -2701,6 +2827,9 @@ static pci_ers_result_t t3_io_error_detected(struct pci_dev *pdev,
 	struct adapter *adapter = pci_get_drvdata(pdev);
 	int ret;
 
+	if (state == pci_channel_io_perm_failure)
+		return PCI_ERS_RESULT_DISCONNECT;
+
 	ret = t3_adapter_error(adapter, 0);
 
 	/* Request a slot reset. */
@@ -2734,6 +2863,9 @@ static void t3_io_resume(struct pci_dev *pdev)
 {
 	struct adapter *adapter = pci_get_drvdata(pdev);
 
+	CH_ALERT(adapter, "adapter recovering, PEX ERR 0x%x\n",
+		 t3_read_reg(adapter, A_PCIE_PEX_ERR));
+
 	t3_resume_ports(adapter);
 }
 
@@ -2753,7 +2885,7 @@ static void set_nqsets(struct adapter *adap)
 	int i, j = 0;
 	int num_cpus = num_online_cpus();
 	int hwports = adap->params.nports;
-	int nqsets = SGE_QSETS;
+	int nqsets = adap->msix_nvectors - 1;
 
 	if (adap->params.rev > 0 && adap->flags & USING_MSIX) {
 		if (hwports == 2 &&
@@ -2782,18 +2914,30 @@ static void set_nqsets(struct adapter *adap)
 static int __devinit cxgb_enable_msix(struct adapter *adap)
 {
 	struct msix_entry entries[SGE_QSETS + 1];
+	int vectors;
 	int i, err;
 
-	for (i = 0; i < ARRAY_SIZE(entries); ++i)
+	vectors = ARRAY_SIZE(entries);
+	for (i = 0; i < vectors; ++i)
 		entries[i].entry = i;
 
-	err = pci_enable_msix(adap->pdev, entries, ARRAY_SIZE(entries));
+	while ((err = pci_enable_msix(adap->pdev, entries, vectors)) > 0)
+		vectors = err;
+
+	if (err < 0)
+		pci_disable_msix(adap->pdev);
+
+	if (!err && vectors < (adap->params.nports + 1)) {
+		pci_disable_msix(adap->pdev);
+		err = -1;
+	}
+
 	if (!err) {
-		for (i = 0; i < ARRAY_SIZE(entries); ++i)
+		for (i = 0; i < vectors; ++i)
 			adap->msix_info[i].vec = entries[i].vector;
-	} else if (err > 0)
-		dev_info(&adap->pdev->dev,
-		       "only %d MSI-X vectors left, not using MSI-X\n", err);
+		adap->msix_nvectors = vectors;
+	}
+
 	return err;
 }
 
@@ -2859,7 +3003,7 @@ static int __devinit init_one(struct pci_dev *pdev,
 	static int version_printed;
 
 	int i, err, pci_using_dac = 0;
-	unsigned long mmio_start, mmio_len;
+	resource_size_t mmio_start, mmio_len;
 	const struct adapter_info *ai;
 	struct adapter *adapter = NULL;
 	struct port_info *pi;
@@ -2891,15 +3035,15 @@ static int __devinit init_one(struct pci_dev *pdev,
 		goto out_release_regions;
 	}
 
-	if (!pci_set_dma_mask(pdev, DMA_64BIT_MASK)) {
+	if (!pci_set_dma_mask(pdev, DMA_BIT_MASK(64))) {
 		pci_using_dac = 1;
-		err = pci_set_consistent_dma_mask(pdev, DMA_64BIT_MASK);
+		err = pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(64));
 		if (err) {
 			dev_err(&pdev->dev, "unable to obtain 64-bit DMA for "
 			       "coherent allocations\n");
 			goto out_disable_device;
 		}
-	} else if ((err = pci_set_dma_mask(pdev, DMA_32BIT_MASK)) != 0) {
+	} else if ((err = pci_set_dma_mask(pdev, DMA_BIT_MASK(32))) != 0) {
 		dev_err(&pdev->dev, "no usable DMA configuration\n");
 		goto out_disable_device;
 	}
@@ -2938,7 +3082,7 @@ static int __devinit init_one(struct pci_dev *pdev,
 	INIT_WORK(&adapter->fatal_error_handler_task, fatal_error_task);
 	INIT_DELAYED_WORK(&adapter->adap_check_task, t3_adap_check_task);
 
-	for (i = 0; i < ai->nports; ++i) {
+	for (i = 0; i < ai->nports0 + ai->nports1; ++i) {
 		struct net_device *netdev;
 
 		netdev = alloc_etherdev_mq(sizeof(struct port_info), SGE_QSETS);
@@ -2961,7 +3105,7 @@ static int __devinit init_one(struct pci_dev *pdev,
 		netdev->mem_end = mmio_start + mmio_len - 1;
 		netdev->features |= NETIF_F_SG | NETIF_F_IP_CSUM | NETIF_F_TSO;
 		netdev->features |= NETIF_F_LLTX;
-		netdev->features |= NETIF_F_LRO;
+		netdev->features |= NETIF_F_GRO;
 		if (pci_using_dac)
 			netdev->features |= NETIF_F_HIGHDMA;
 
@@ -3028,7 +3172,7 @@ static int __devinit init_one(struct pci_dev *pdev,
 
 out_free_dev:
 	iounmap(adapter->regs);
-	for (i = ai->nports - 1; i >= 0; --i)
+	for (i = ai->nports0 + ai->nports1 - 1; i >= 0; --i)
 		if (adapter->port[i])
 			free_netdev(adapter->port[i]);
 
diff --git a/drivers/net/cxgb3/cxgb3_offload.c b/drivers/net/cxgb3/cxgb3_offload.c
index 2d7f69aff1d9..620d80be6aac 100644
--- a/drivers/net/cxgb3/cxgb3_offload.c
+++ b/drivers/net/cxgb3/cxgb3_offload.c
@@ -153,6 +153,18 @@ void cxgb3_remove_clients(struct t3cdev *tdev)
 	mutex_unlock(&cxgb3_db_lock);
 }
 
+void cxgb3_err_notify(struct t3cdev *tdev, u32 status, u32 error)
+{
+	struct cxgb3_client *client;
+
+	mutex_lock(&cxgb3_db_lock);
+	list_for_each_entry(client, &client_list, client_list) {
+		if (client->err_handler)
+			client->err_handler(tdev, status, error);
+	}
+	mutex_unlock(&cxgb3_db_lock);
+}
+
 static struct net_device *get_iff_from_mac(struct adapter *adapter,
 					   const unsigned char *mac,
 					   unsigned int vlan)
diff --git a/drivers/net/cxgb3/cxgb3_offload.h b/drivers/net/cxgb3/cxgb3_offload.h
index d514e5019dfc..a8e8e5fcdf84 100644
--- a/drivers/net/cxgb3/cxgb3_offload.h
+++ b/drivers/net/cxgb3/cxgb3_offload.h
@@ -64,10 +64,16 @@ void cxgb3_register_client(struct cxgb3_client *client);
 void cxgb3_unregister_client(struct cxgb3_client *client);
 void cxgb3_add_clients(struct t3cdev *tdev);
 void cxgb3_remove_clients(struct t3cdev *tdev);
+void cxgb3_err_notify(struct t3cdev *tdev, u32 status, u32 error);
 
 typedef int (*cxgb3_cpl_handler_func)(struct t3cdev *dev,
 				      struct sk_buff *skb, void *ctx);
 
+enum {
+	OFFLOAD_STATUS_UP,
+	OFFLOAD_STATUS_DOWN
+};
+
 struct cxgb3_client {
 	char *name;
 	void (*add) (struct t3cdev *);
@@ -76,6 +82,7 @@ struct cxgb3_client {
 	int (*redirect)(void *ctx, struct dst_entry *old,
 			struct dst_entry *new, struct l2t_entry *l2t);
 	struct list_head client_list;
+	void (*err_handler)(struct t3cdev *tdev, u32 status, u32 error);
 };
 
 /*
diff --git a/drivers/net/cxgb3/regs.h b/drivers/net/cxgb3/regs.h
index a035d5c24442..1b5327b5a965 100644
--- a/drivers/net/cxgb3/regs.h
+++ b/drivers/net/cxgb3/regs.h
@@ -170,6 +170,10 @@
 
 #define S_RSPQ0DISABLED    8
 
+#define S_FL0EMPTY    16
+#define V_FL0EMPTY(x) ((x) << S_FL0EMPTY)
+#define F_FL0EMPTY    V_FL0EMPTY(1U)
+
 #define A_SG_EGR_RCQ_DRB_THRSH 0x54
 
 #define S_HIRCQDRBTHRSH    16
@@ -258,6 +262,10 @@
 #define V_RSPQCREDITOVERFOW(x) ((x) << S_RSPQCREDITOVERFOW)
 #define F_RSPQCREDITOVERFOW    V_RSPQCREDITOVERFOW(1U)
 
+#define S_FLEMPTY    1
+#define V_FLEMPTY(x) ((x) << S_FLEMPTY)
+#define F_FLEMPTY    V_FLEMPTY(1U)
+
 #define A_SG_INT_ENABLE 0x60
 
 #define A_SG_CMDQ_CREDIT_TH 0x64
@@ -2207,6 +2215,15 @@
 
 #define A_XGM_RX_EXACT_MATCH_LOW_8 0x854
 
+#define A_XGM_INT_STATUS 0x86c
+
+#define S_LINKFAULTCHANGE    9
+#define V_LINKFAULTCHANGE(x) ((x) << S_LINKFAULTCHANGE)
+#define F_LINKFAULTCHANGE    V_LINKFAULTCHANGE(1U)
+
+#define A_XGM_XGM_INT_ENABLE 0x874
+#define A_XGM_XGM_INT_DISABLE 0x878
+
 #define A_XGM_STAT_CTRL 0x880
 
 #define S_CLRSTATS    2
@@ -2405,6 +2422,10 @@
 #define V_XAUIPCSALIGNCHANGE(x) ((x) << S_XAUIPCSALIGNCHANGE)
 #define F_XAUIPCSALIGNCHANGE    V_XAUIPCSALIGNCHANGE(1U)
 
+#define S_XGM_INT    0
+#define V_XGM_INT(x) ((x) << S_XGM_INT)
+#define F_XGM_INT    V_XGM_INT(1U)
+
 #define A_XGM_INT_CAUSE 0x8d8
 
 #define A_XGM_XAUI_ACT_CTRL 0x8dc
diff --git a/drivers/net/cxgb3/sge.c b/drivers/net/cxgb3/sge.c
index d31791f60292..26d3587f3399 100644
--- a/drivers/net/cxgb3/sge.c
+++ b/drivers/net/cxgb3/sge.c
@@ -50,6 +50,7 @@
 #define SGE_RX_COPY_THRES  256
 #define SGE_RX_PULL_LEN    128
 
+#define SGE_PG_RSVD SMP_CACHE_BYTES
 /*
  * Page chunk size for FL0 buffers if FL0 is to be populated with page chunks.
  * It must be a divisor of PAGE_SIZE.  If set to 0 FL0 will use sk_buffs
@@ -57,16 +58,25 @@
  */
 #define FL0_PG_CHUNK_SIZE  2048
 #define FL0_PG_ORDER 0
+#define FL0_PG_ALLOC_SIZE (PAGE_SIZE << FL0_PG_ORDER)
 #define FL1_PG_CHUNK_SIZE (PAGE_SIZE > 8192 ? 16384 : 8192)
 #define FL1_PG_ORDER (PAGE_SIZE > 8192 ? 0 : 1)
+#define FL1_PG_ALLOC_SIZE (PAGE_SIZE << FL1_PG_ORDER)
 
 #define SGE_RX_DROP_THRES 16
+#define RX_RECLAIM_PERIOD (HZ/4)
 
 /*
+ * Max number of Rx buffers we replenish at a time.
+ */
+#define MAX_RX_REFILL 16U
+/*
  * Period of the Tx buffer reclaim timer.  This timer does not need to run
  * frequently as Tx buffers are usually reclaimed by new Tx packets.
  */
 #define TX_RECLAIM_PERIOD (HZ / 4)
+#define TX_RECLAIM_TIMER_CHUNK 64U
+#define TX_RECLAIM_CHUNK 16U
 
 /* WR size in bytes */
 #define WR_LEN (WR_FLITS * 8)
@@ -304,21 +314,25 @@ static void free_tx_desc(struct adapter *adapter, struct sge_txq *q,
  *	reclaim_completed_tx - reclaims completed Tx descriptors
  *	@adapter: the adapter
  *	@q: the Tx queue to reclaim completed descriptors from
+ *	@chunk: maximum number of descriptors to reclaim
  *
  *	Reclaims Tx descriptors that the SGE has indicated it has processed,
  *	and frees the associated buffers if possible.  Called with the Tx
  *	queue's lock held.
  */
-static inline void reclaim_completed_tx(struct adapter *adapter,
-					struct sge_txq *q)
+static inline unsigned int reclaim_completed_tx(struct adapter *adapter,
+						struct sge_txq *q,
+						unsigned int chunk)
 {
 	unsigned int reclaim = q->processed - q->cleaned;
 
+	reclaim = min(chunk, reclaim);
 	if (reclaim) {
 		free_tx_desc(adapter, q, reclaim);
 		q->cleaned += reclaim;
 		q->in_use -= reclaim;
 	}
+	return q->processed - q->cleaned;
 }
 
 /**
@@ -334,6 +348,26 @@ static inline int should_restart_tx(const struct sge_txq *q)
 	return q->in_use - r < (q->size >> 1);
 }
 
+static void clear_rx_desc(struct pci_dev *pdev, const struct sge_fl *q,
+			  struct rx_sw_desc *d)
+{
+	if (q->use_pages && d->pg_chunk.page) {
+		(*d->pg_chunk.p_cnt)--;
+		if (!*d->pg_chunk.p_cnt)
+			pci_unmap_page(pdev,
+				       pci_unmap_addr(&d->pg_chunk, mapping),
+				       q->alloc_size, PCI_DMA_FROMDEVICE);
+
+		put_page(d->pg_chunk.page);
+		d->pg_chunk.page = NULL;
+	} else {
+		pci_unmap_single(pdev, pci_unmap_addr(d, dma_addr),
+				 q->buf_size, PCI_DMA_FROMDEVICE);
+		kfree_skb(d->skb);
+		d->skb = NULL;
+	}
+}
+
 /**
  *	free_rx_bufs - free the Rx buffers on an SGE free list
  *	@pdev: the PCI device associated with the adapter
@@ -349,16 +383,8 @@ static void free_rx_bufs(struct pci_dev *pdev, struct sge_fl *q)
 	while (q->credits--) {
 		struct rx_sw_desc *d = &q->sdesc[cidx];
 
-		pci_unmap_single(pdev, pci_unmap_addr(d, dma_addr),
-				 q->buf_size, PCI_DMA_FROMDEVICE);
-		if (q->use_pages) {
-			if (d->pg_chunk.page)
-				put_page(d->pg_chunk.page);
-			d->pg_chunk.page = NULL;
-		} else {
-			kfree_skb(d->skb);
-			d->skb = NULL;
-		}
+
+		clear_rx_desc(pdev, q, d);
 		if (++cidx == q->size)
 			cidx = 0;
 	}
@@ -401,18 +427,39 @@ static inline int add_one_rx_buf(void *va, unsigned int len,
 	return 0;
 }
 
-static int alloc_pg_chunk(struct sge_fl *q, struct rx_sw_desc *sd, gfp_t gfp,
+static inline int add_one_rx_chunk(dma_addr_t mapping, struct rx_desc *d,
+				   unsigned int gen)
+{
+	d->addr_lo = cpu_to_be32(mapping);
+	d->addr_hi = cpu_to_be32((u64) mapping >> 32);
+	wmb();
+	d->len_gen = cpu_to_be32(V_FLD_GEN1(gen));
+	d->gen2 = cpu_to_be32(V_FLD_GEN2(gen));
+	return 0;
+}
+
+static int alloc_pg_chunk(struct adapter *adapter, struct sge_fl *q,
+			  struct rx_sw_desc *sd, gfp_t gfp,
 			  unsigned int order)
 {
 	if (!q->pg_chunk.page) {
+		dma_addr_t mapping;
+
 		q->pg_chunk.page = alloc_pages(gfp, order);
 		if (unlikely(!q->pg_chunk.page))
 			return -ENOMEM;
 		q->pg_chunk.va = page_address(q->pg_chunk.page);
+		q->pg_chunk.p_cnt = q->pg_chunk.va + (PAGE_SIZE << order) -
+				    SGE_PG_RSVD;
 		q->pg_chunk.offset = 0;
+		mapping = pci_map_page(adapter->pdev, q->pg_chunk.page,
+				       0, q->alloc_size, PCI_DMA_FROMDEVICE);
+		pci_unmap_addr_set(&q->pg_chunk, mapping, mapping);
 	}
 	sd->pg_chunk = q->pg_chunk;
 
+	prefetch(sd->pg_chunk.p_cnt);
+
 	q->pg_chunk.offset += q->buf_size;
 	if (q->pg_chunk.offset == (PAGE_SIZE << order))
 		q->pg_chunk.page = NULL;
@@ -420,9 +467,23 @@ static int alloc_pg_chunk(struct sge_fl *q, struct rx_sw_desc *sd, gfp_t gfp,
 		q->pg_chunk.va += q->buf_size;
 		get_page(q->pg_chunk.page);
 	}
+
+	if (sd->pg_chunk.offset == 0)
+		*sd->pg_chunk.p_cnt = 1;
+	else
+		*sd->pg_chunk.p_cnt += 1;
+
 	return 0;
 }
 
+static inline void ring_fl_db(struct adapter *adap, struct sge_fl *q)
+{
+	if (q->pend_cred >= q->credits / 4) {
+		q->pend_cred = 0;
+		t3_write_reg(adap, A_SG_KDOORBELL, V_EGRCNTX(q->cntxt_id));
+	}
+}
+
 /**
  *	refill_fl - refill an SGE free-buffer list
  *	@adapter: the adapter
@@ -436,38 +497,43 @@ static int alloc_pg_chunk(struct sge_fl *q, struct rx_sw_desc *sd, gfp_t gfp,
  */
 static int refill_fl(struct adapter *adap, struct sge_fl *q, int n, gfp_t gfp)
 {
-	void *buf_start;
 	struct rx_sw_desc *sd = &q->sdesc[q->pidx];
 	struct rx_desc *d = &q->desc[q->pidx];
 	unsigned int count = 0;
 
 	while (n--) {
+		dma_addr_t mapping;
 		int err;
 
 		if (q->use_pages) {
-			if (unlikely(alloc_pg_chunk(q, sd, gfp, q->order))) {
+			if (unlikely(alloc_pg_chunk(adap, q, sd, gfp,
+						    q->order))) {
 nomem:				q->alloc_failed++;
 				break;
 			}
-			buf_start = sd->pg_chunk.va;
+			mapping = pci_unmap_addr(&sd->pg_chunk, mapping) +
+						 sd->pg_chunk.offset;
+			pci_unmap_addr_set(sd, dma_addr, mapping);
+
+			add_one_rx_chunk(mapping, d, q->gen);
+			pci_dma_sync_single_for_device(adap->pdev, mapping,
+						q->buf_size - SGE_PG_RSVD,
+						PCI_DMA_FROMDEVICE);
 		} else {
-			struct sk_buff *skb = alloc_skb(q->buf_size, gfp);
+			void *buf_start;
 
+			struct sk_buff *skb = alloc_skb(q->buf_size, gfp);
 			if (!skb)
 				goto nomem;
 
 			sd->skb = skb;
 			buf_start = skb->data;
-		}
-
-		err = add_one_rx_buf(buf_start, q->buf_size, d, sd, q->gen,
-				     adap->pdev);
-		if (unlikely(err)) {
-			if (!q->use_pages) {
-				kfree_skb(sd->skb);
-				sd->skb = NULL;
+			err = add_one_rx_buf(buf_start, q->buf_size, d, sd,
+					     q->gen, adap->pdev);
+			if (unlikely(err)) {
+				clear_rx_desc(adap->pdev, q, sd);
+				break;
 			}
-			break;
 		}
 
 		d++;
@@ -478,19 +544,19 @@ nomem:				q->alloc_failed++;
 			sd = q->sdesc;
 			d = q->desc;
 		}
-		q->credits++;
 		count++;
 	}
-	wmb();
-	if (likely(count))
-		t3_write_reg(adap, A_SG_KDOORBELL, V_EGRCNTX(q->cntxt_id));
+
+	q->credits += count;
+	q->pend_cred += count;
+	ring_fl_db(adap, q);
 
 	return count;
 }
 
 static inline void __refill_fl(struct adapter *adap, struct sge_fl *fl)
 {
-	refill_fl(adap, fl, min(16U, fl->size - fl->credits),
+	refill_fl(adap, fl, min(MAX_RX_REFILL, fl->size - fl->credits),
 		  GFP_ATOMIC | __GFP_COMP);
 }
 
@@ -515,13 +581,15 @@ static void recycle_rx_buf(struct adapter *adap, struct sge_fl *q,
 	wmb();
 	to->len_gen = cpu_to_be32(V_FLD_GEN1(q->gen));
 	to->gen2 = cpu_to_be32(V_FLD_GEN2(q->gen));
-	q->credits++;
 
 	if (++q->pidx == q->size) {
 		q->pidx = 0;
 		q->gen ^= 1;
 	}
-	t3_write_reg(adap, A_SG_KDOORBELL, V_EGRCNTX(q->cntxt_id));
+
+	q->credits++;
+	q->pend_cred++;
+	ring_fl_db(adap, q);
 }
 
 /**
@@ -585,8 +653,8 @@ static void t3_reset_qset(struct sge_qset *q)
 	memset(q->txq, 0, sizeof(struct sge_txq) * SGE_TXQ_PER_SET);
 	q->txq_stopped = 0;
 	q->tx_reclaim_timer.function = NULL; /* for t3_stop_sge_timers() */
-	kfree(q->lro_frag_tbl);
-	q->lro_nfrags = q->lro_frag_len = 0;
+	q->rx_reclaim_timer.function = NULL;
+	q->lro_frag_tbl.nr_frags = q->lro_frag_tbl.len = 0;
 }
 
 
@@ -733,7 +801,9 @@ recycle:
 		return skb;
 	}
 
-	if (unlikely(fl->credits < drop_thres))
+	if (unlikely(fl->credits < drop_thres) &&
+	    refill_fl(adap, fl, min(MAX_RX_REFILL, fl->size - fl->credits - 1),
+		      GFP_ATOMIC | __GFP_COMP) == 0)
 		goto recycle;
 
 use_orig_buf:
@@ -770,19 +840,19 @@ static struct sk_buff *get_packet_pg(struct adapter *adap, struct sge_fl *fl,
 	struct sk_buff *newskb, *skb;
 	struct rx_sw_desc *sd = &fl->sdesc[fl->cidx];
 
-	newskb = skb = q->pg_skb;
+	dma_addr_t dma_addr = pci_unmap_addr(sd, dma_addr);
 
+	newskb = skb = q->pg_skb;
 	if (!skb && (len <= SGE_RX_COPY_THRES)) {
 		newskb = alloc_skb(len, GFP_ATOMIC);
 		if (likely(newskb != NULL)) {
 			__skb_put(newskb, len);
-			pci_dma_sync_single_for_cpu(adap->pdev,
-					    pci_unmap_addr(sd, dma_addr), len,
+			pci_dma_sync_single_for_cpu(adap->pdev, dma_addr, len,
 					    PCI_DMA_FROMDEVICE);
 			memcpy(newskb->data, sd->pg_chunk.va, len);
-			pci_dma_sync_single_for_device(adap->pdev,
-					    pci_unmap_addr(sd, dma_addr), len,
-					    PCI_DMA_FROMDEVICE);
+			pci_dma_sync_single_for_device(adap->pdev, dma_addr,
+						       len,
+						       PCI_DMA_FROMDEVICE);
 		} else if (!drop_thres)
 			return NULL;
 recycle:
@@ -795,16 +865,25 @@ recycle:
 	if (unlikely(q->rx_recycle_buf || (!skb && fl->credits <= drop_thres)))
 		goto recycle;
 
+	prefetch(sd->pg_chunk.p_cnt);
+
 	if (!skb)
 		newskb = alloc_skb(SGE_RX_PULL_LEN, GFP_ATOMIC);
+
 	if (unlikely(!newskb)) {
 		if (!drop_thres)
 			return NULL;
 		goto recycle;
 	}
 
-	pci_unmap_single(adap->pdev, pci_unmap_addr(sd, dma_addr),
-			 fl->buf_size, PCI_DMA_FROMDEVICE);
+	pci_dma_sync_single_for_cpu(adap->pdev, dma_addr, len,
+				    PCI_DMA_FROMDEVICE);
+	(*sd->pg_chunk.p_cnt)--;
+	if (!*sd->pg_chunk.p_cnt)
+		pci_unmap_page(adap->pdev,
+			       pci_unmap_addr(&sd->pg_chunk, mapping),
+			       fl->alloc_size,
+			       PCI_DMA_FROMDEVICE);
 	if (!skb) {
 		__skb_put(newskb, SGE_RX_PULL_LEN);
 		memcpy(newskb->data, sd->pg_chunk.va, SGE_RX_PULL_LEN);
@@ -813,14 +892,15 @@ recycle:
 				   len - SGE_RX_PULL_LEN);
 		newskb->len = len;
 		newskb->data_len = len - SGE_RX_PULL_LEN;
+		newskb->truesize += newskb->data_len;
 	} else {
 		skb_fill_page_desc(newskb, skb_shinfo(newskb)->nr_frags,
 				   sd->pg_chunk.page,
 				   sd->pg_chunk.offset, len);
 		newskb->len += len;
 		newskb->data_len += len;
+		newskb->truesize += len;
 	}
-	newskb->truesize += newskb->data_len;
 
 	fl->credits--;
 	/*
@@ -1063,7 +1143,7 @@ static void write_tx_pkt_wr(struct adapter *adap, struct sk_buff *skb,
 	struct tx_desc *d = &q->desc[pidx];
 	struct cpl_tx_pkt *cpl = (struct cpl_tx_pkt *)d;
 
-	cpl->len = htonl(skb->len | 0x80000000);
+	cpl->len = htonl(skb->len);
 	cntrl = V_TXPKT_INTF(pi->port_id);
 
 	if (vlan_tx_tag_present(skb) && pi->vlan_grp)
@@ -1161,7 +1241,7 @@ int t3_eth_xmit(struct sk_buff *skb, struct net_device *dev)
 	txq = netdev_get_tx_queue(dev, qidx);
 
 	spin_lock(&q->lock);
-	reclaim_completed_tx(adap, q);
+	reclaim_completed_tx(adap, q, TX_RECLAIM_CHUNK);
 
 	credits = q->size - q->in_use;
 	ndesc = calc_tx_descs(skb);
@@ -1570,7 +1650,7 @@ static int ofld_xmit(struct adapter *adap, struct sge_txq *q,
 	unsigned int ndesc = calc_tx_descs_ofld(skb), pidx, gen;
 
 	spin_lock(&q->lock);
-      again:reclaim_completed_tx(adap, q);
+again:	reclaim_completed_tx(adap, q, TX_RECLAIM_CHUNK);
 
 	ret = check_desc_avail(adap, q, skb, ndesc, TXQ_OFLD);
 	if (unlikely(ret)) {
@@ -1612,7 +1692,7 @@ static void restart_offloadq(unsigned long data)
 	struct adapter *adap = pi->adapter;
 
 	spin_lock(&q->lock);
-      again:reclaim_completed_tx(adap, q);
+again:	reclaim_completed_tx(adap, q, TX_RECLAIM_CHUNK);
 
 	while ((skb = skb_peek(&q->sendq)) != NULL) {
 		unsigned int gen, pidx;
@@ -1932,12 +2012,13 @@ static void rx_eth(struct adapter *adap, struct sge_rspq *rq,
 	skb_pull(skb, sizeof(*p) + pad);
 	skb->protocol = eth_type_trans(skb, adap->port[p->iff]);
 	pi = netdev_priv(skb->dev);
-	if ((pi->rx_offload & T3_RX_CSUM) && p->csum_valid && p->csum == htons(0xffff) &&
-	    !p->fragment) {
+	if ((pi->rx_offload & T3_RX_CSUM) && p->csum_valid &&
+	    p->csum == htons(0xffff) && !p->fragment) {
 		qs->port_stats[SGE_PSTAT_RX_CSUM_GOOD]++;
 		skb->ip_summed = CHECKSUM_UNNECESSARY;
 	} else
 		skb->ip_summed = CHECKSUM_NONE;
+	skb_record_rx_queue(skb, qs - &adap->sge.qs[0]);
 
 	if (unlikely(p->vlan_valid)) {
 		struct vlan_group *grp = pi->vlan_grp;
@@ -1945,10 +2026,8 @@ static void rx_eth(struct adapter *adap, struct sge_rspq *rq,
 		qs->port_stats[SGE_PSTAT_VLANEX]++;
 		if (likely(grp))
 			if (lro)
-				lro_vlan_hwaccel_receive_skb(&qs->lro_mgr, skb,
-							     grp,
-							     ntohs(p->vlan),
-							     p);
+				vlan_gro_receive(&qs->napi, grp,
+						 ntohs(p->vlan), skb);
 			else {
 				if (unlikely(pi->iscsi_ipv4addr &&
 				    is_arp(skb))) {
@@ -1965,7 +2044,7 @@ static void rx_eth(struct adapter *adap, struct sge_rspq *rq,
 			dev_kfree_skb_any(skb);
 	} else if (rq->polling) {
 		if (lro)
-			lro_receive_skb(&qs->lro_mgr, skb, p);
+			napi_gro_receive(&qs->napi, skb);
 		else {
 			if (unlikely(pi->iscsi_ipv4addr && is_arp(skb)))
 				cxgb3_arp_process(adap, skb);
@@ -1981,59 +2060,6 @@ static inline int is_eth_tcp(u32 rss)
 }
 
 /**
- *	lro_frame_ok - check if an ingress packet is eligible for LRO
- *	@p: the CPL header of the packet
- *
- *	Returns true if a received packet is eligible for LRO.
- *	The following conditions must be true:
- *	- packet is TCP/IP Ethernet II (checked elsewhere)
- *	- not an IP fragment
- *	- no IP options
- *	- TCP/IP checksums are correct
- *	- the packet is for this host
- */
-static inline int lro_frame_ok(const struct cpl_rx_pkt *p)
-{
-	const struct ethhdr *eh = (struct ethhdr *)(p + 1);
-	const struct iphdr *ih = (struct iphdr *)(eh + 1);
-
-	return (*((u8 *)p + 1) & 0x90) == 0x10 && p->csum == htons(0xffff) &&
-		eh->h_proto == htons(ETH_P_IP) && ih->ihl == (sizeof(*ih) >> 2);
-}
-
-static int t3_get_lro_header(void **eh,  void **iph, void **tcph,
-			     u64 *hdr_flags, void *priv)
-{
-	const struct cpl_rx_pkt *cpl = priv;
-
-	if (!lro_frame_ok(cpl))
-		return -1;
-
-	*eh = (struct ethhdr *)(cpl + 1);
-	*iph = (struct iphdr *)((struct ethhdr *)*eh + 1);
-	*tcph = (struct tcphdr *)((struct iphdr *)*iph + 1);
-
-	*hdr_flags = LRO_IPV4 | LRO_TCP;
-	return 0;
-}
-
-static int t3_get_skb_header(struct sk_buff *skb,
-			      void **iph, void **tcph, u64 *hdr_flags,
-			      void *priv)
-{
-	void *eh;
-
-	return t3_get_lro_header(&eh, iph, tcph, hdr_flags, priv);
-}
-
-static int t3_get_frag_header(struct skb_frag_struct *frag, void **eh,
-			      void **iph, void **tcph, u64 *hdr_flags,
-			      void *priv)
-{
-	return t3_get_lro_header(eh, iph, tcph, hdr_flags, priv);
-}
-
-/**
  *	lro_add_page - add a page chunk to an LRO session
  *	@adap: the adapter
  *	@qs: the associated queue set
@@ -2049,8 +2075,9 @@ static void lro_add_page(struct adapter *adap, struct sge_qset *qs,
 {
 	struct rx_sw_desc *sd = &fl->sdesc[fl->cidx];
 	struct cpl_rx_pkt *cpl;
-	struct skb_frag_struct *rx_frag = qs->lro_frag_tbl;
-	int nr_frags = qs->lro_nfrags, frag_len = qs->lro_frag_len;
+	struct skb_frag_struct *rx_frag = qs->lro_frag_tbl.frags;
+	int nr_frags = qs->lro_frag_tbl.nr_frags;
+	int frag_len = qs->lro_frag_tbl.len;
 	int offset = 0;
 
 	if (!nr_frags) {
@@ -2061,21 +2088,33 @@ static void lro_add_page(struct adapter *adap, struct sge_qset *qs,
 	fl->credits--;
 
 	len -= offset;
-	pci_unmap_single(adap->pdev, pci_unmap_addr(sd, dma_addr),
-			 fl->buf_size, PCI_DMA_FROMDEVICE);
+	pci_dma_sync_single_for_cpu(adap->pdev,
+				    pci_unmap_addr(sd, dma_addr),
+				    fl->buf_size - SGE_PG_RSVD,
+				    PCI_DMA_FROMDEVICE);
+
+	(*sd->pg_chunk.p_cnt)--;
+	if (!*sd->pg_chunk.p_cnt)
+		pci_unmap_page(adap->pdev,
+			       pci_unmap_addr(&sd->pg_chunk, mapping),
+			       fl->alloc_size,
+			       PCI_DMA_FROMDEVICE);
+
+	prefetch(qs->lro_va);
 
 	rx_frag += nr_frags;
 	rx_frag->page = sd->pg_chunk.page;
 	rx_frag->page_offset = sd->pg_chunk.offset + offset;
 	rx_frag->size = len;
 	frag_len += len;
-	qs->lro_nfrags++;
-	qs->lro_frag_len = frag_len;
+	qs->lro_frag_tbl.nr_frags++;
+	qs->lro_frag_tbl.len = frag_len;
+
 
 	if (!complete)
 		return;
 
-	qs->lro_nfrags = qs->lro_frag_len = 0;
+	qs->lro_frag_tbl.ip_summed = CHECKSUM_UNNECESSARY;
 	cpl = qs->lro_va;
 
 	if (unlikely(cpl->vlan_valid)) {
@@ -2084,36 +2123,15 @@ static void lro_add_page(struct adapter *adap, struct sge_qset *qs,
 		struct vlan_group *grp = pi->vlan_grp;
 
 		if (likely(grp != NULL)) {
-			lro_vlan_hwaccel_receive_frags(&qs->lro_mgr,
-						       qs->lro_frag_tbl,
-						       frag_len, frag_len,
-						       grp, ntohs(cpl->vlan),
-						       cpl, 0);
-			return;
+			vlan_gro_frags(&qs->napi, grp, ntohs(cpl->vlan),
+				       &qs->lro_frag_tbl);
+			goto out;
 		}
 	}
-	lro_receive_frags(&qs->lro_mgr, qs->lro_frag_tbl,
-			  frag_len, frag_len, cpl, 0);
-}
+	napi_gro_frags(&qs->napi, &qs->lro_frag_tbl);
 
-/**
- *	init_lro_mgr - initialize a LRO manager object
- *	@lro_mgr: the LRO manager object
- */
-static void init_lro_mgr(struct sge_qset *qs, struct net_lro_mgr *lro_mgr)
-{
-	lro_mgr->dev = qs->netdev;
-	lro_mgr->features = LRO_F_NAPI;
-	lro_mgr->frag_align_pad = NET_IP_ALIGN;
-	lro_mgr->ip_summed = CHECKSUM_UNNECESSARY;
-	lro_mgr->ip_summed_aggr = CHECKSUM_UNNECESSARY;
-	lro_mgr->max_desc = T3_MAX_LRO_SES;
-	lro_mgr->lro_arr = qs->lro_desc;
-	lro_mgr->get_frag_header = t3_get_frag_header;
-	lro_mgr->get_skb_header = t3_get_skb_header;
-	lro_mgr->max_aggr = T3_MAX_LRO_MAX_PKTS;
-	if (lro_mgr->max_aggr > MAX_SKB_FRAGS)
-		lro_mgr->max_aggr = MAX_SKB_FRAGS;
+out:
+	qs->lro_frag_tbl.nr_frags = qs->lro_frag_tbl.len = 0;
 }
 
 /**
@@ -2282,6 +2300,8 @@ no_mem:
 			if (fl->use_pages) {
 				void *addr = fl->sdesc[fl->cidx].pg_chunk.va;
 
+				prefetch(&qs->lro_frag_tbl);
+
 				prefetch(addr);
 #if L1_CACHE_BYTES < 128
 				prefetch(addr + L1_CACHE_BYTES);
@@ -2356,10 +2376,6 @@ next_fl:
 	}
 
 	deliver_partial_bundle(&adap->tdev, q, offload_skbs, ngathered);
-	lro_flush_all(&qs->lro_mgr);
-	qs->port_stats[SGE_PSTAT_LRO_AGGR] = qs->lro_mgr.stats.aggregated;
-	qs->port_stats[SGE_PSTAT_LRO_FLUSHED] = qs->lro_mgr.stats.flushed;
-	qs->port_stats[SGE_PSTAT_LRO_NO_DESC] = qs->lro_mgr.stats.no_desc;
 
 	if (sleeping)
 		check_ring_db(adap, qs, sleeping);
@@ -2775,7 +2791,8 @@ irq_handler_t t3_intr_handler(struct adapter *adap, int polling)
  */
 void t3_sge_err_intr_handler(struct adapter *adapter)
 {
-	unsigned int v, status = t3_read_reg(adapter, A_SG_INT_CAUSE);
+	unsigned int v, status = t3_read_reg(adapter, A_SG_INT_CAUSE) &
+				 ~F_FLEMPTY;
 
 	if (status & SGE_PARERR)
 		CH_ALERT(adapter, "SGE parity error (0x%x)\n",
@@ -2805,13 +2822,13 @@ void t3_sge_err_intr_handler(struct adapter *adapter)
 }
 
 /**
- *	sge_timer_cb - perform periodic maintenance of an SGE qset
+ *	sge_timer_tx - perform periodic maintenance of an SGE qset
  *	@data: the SGE queue set to maintain
  *
  *	Runs periodically from a timer to perform maintenance of an SGE queue
  *	set.  It performs two tasks:
  *
- *	a) Cleans up any completed Tx descriptors that may still be pending.
+ *	Cleans up any completed Tx descriptors that may still be pending.
  *	Normal descriptor cleanup happens when new packets are added to a Tx
  *	queue so this timer is relatively infrequent and does any cleanup only
  *	if the Tx queue has not seen any new packets in a while.  We make a
@@ -2821,51 +2838,87 @@ void t3_sge_err_intr_handler(struct adapter *adapter)
  *	up).  Since control queues use immediate data exclusively we don't
  *	bother cleaning them up here.
  *
- *	b) Replenishes Rx queues that have run out due to memory shortage.
- *	Normally new Rx buffers are added when existing ones are consumed but
- *	when out of memory a queue can become empty.  We try to add only a few
- *	buffers here, the queue will be replenished fully as these new buffers
- *	are used up if memory shortage has subsided.
  */
-static void sge_timer_cb(unsigned long data)
+static void sge_timer_tx(unsigned long data)
 {
-	spinlock_t *lock;
 	struct sge_qset *qs = (struct sge_qset *)data;
-	struct adapter *adap = qs->adap;
+	struct port_info *pi = netdev_priv(qs->netdev);
+	struct adapter *adap = pi->adapter;
+	unsigned int tbd[SGE_TXQ_PER_SET] = {0, 0};
+	unsigned long next_period;
 
 	if (spin_trylock(&qs->txq[TXQ_ETH].lock)) {
-		reclaim_completed_tx(adap, &qs->txq[TXQ_ETH]);
+		tbd[TXQ_ETH] = reclaim_completed_tx(adap, &qs->txq[TXQ_ETH],
+						    TX_RECLAIM_TIMER_CHUNK);
 		spin_unlock(&qs->txq[TXQ_ETH].lock);
 	}
 	if (spin_trylock(&qs->txq[TXQ_OFLD].lock)) {
-		reclaim_completed_tx(adap, &qs->txq[TXQ_OFLD]);
+		tbd[TXQ_OFLD] = reclaim_completed_tx(adap, &qs->txq[TXQ_OFLD],
+						     TX_RECLAIM_TIMER_CHUNK);
 		spin_unlock(&qs->txq[TXQ_OFLD].lock);
 	}
-	lock = (adap->flags & USING_MSIX) ? &qs->rspq.lock :
-					    &adap->sge.qs[0].rspq.lock;
-	if (spin_trylock_irq(lock)) {
-		if (!napi_is_scheduled(&qs->napi)) {
-			u32 status = t3_read_reg(adap, A_SG_RSPQ_FL_STATUS);
-
-			if (qs->fl[0].credits < qs->fl[0].size)
-				__refill_fl(adap, &qs->fl[0]);
-			if (qs->fl[1].credits < qs->fl[1].size)
-				__refill_fl(adap, &qs->fl[1]);
-
-			if (status & (1 << qs->rspq.cntxt_id)) {
-				qs->rspq.starved++;
-				if (qs->rspq.credits) {
-					refill_rspq(adap, &qs->rspq, 1);
-					qs->rspq.credits--;
-					qs->rspq.restarted++;
-					t3_write_reg(adap, A_SG_RSPQ_FL_STATUS,
-						     1 << qs->rspq.cntxt_id);
-				}
+
+	next_period = TX_RECLAIM_PERIOD >>
+		      (max(tbd[TXQ_ETH], tbd[TXQ_OFLD]) /
+		       TX_RECLAIM_TIMER_CHUNK);
+	mod_timer(&qs->tx_reclaim_timer, jiffies + next_period);
+}
+
+/*
+ *	sge_timer_rx - perform periodic maintenance of an SGE qset
+ *	@data: the SGE queue set to maintain
+ *
+ *	a) Replenishes Rx queues that have run out due to memory shortage.
+ *	Normally new Rx buffers are added when existing ones are consumed but
+ *	when out of memory a queue can become empty.  We try to add only a few
+ *	buffers here, the queue will be replenished fully as these new buffers
+ *	are used up if memory shortage has subsided.
+ *
+ *	b) Return coalesced response queue credits in case a response queue is
+ *	starved.
+ *
+ */
+static void sge_timer_rx(unsigned long data)
+{
+	spinlock_t *lock;
+	struct sge_qset *qs = (struct sge_qset *)data;
+	struct port_info *pi = netdev_priv(qs->netdev);
+	struct adapter *adap = pi->adapter;
+	u32 status;
+
+	lock = adap->params.rev > 0 ?
+	       &qs->rspq.lock : &adap->sge.qs[0].rspq.lock;
+
+	if (!spin_trylock_irq(lock))
+		goto out;
+
+	if (napi_is_scheduled(&qs->napi))
+		goto unlock;
+
+	if (adap->params.rev < 4) {
+		status = t3_read_reg(adap, A_SG_RSPQ_FL_STATUS);
+
+		if (status & (1 << qs->rspq.cntxt_id)) {
+			qs->rspq.starved++;
+			if (qs->rspq.credits) {
+				qs->rspq.credits--;
+				refill_rspq(adap, &qs->rspq, 1);
+				qs->rspq.restarted++;
+				t3_write_reg(adap, A_SG_RSPQ_FL_STATUS,
+					     1 << qs->rspq.cntxt_id);
 			}
 		}
-		spin_unlock_irq(lock);
 	}
-	mod_timer(&qs->tx_reclaim_timer, jiffies + TX_RECLAIM_PERIOD);
+
+	if (qs->fl[0].credits < qs->fl[0].size)
+		__refill_fl(adap, &qs->fl[0]);
+	if (qs->fl[1].credits < qs->fl[1].size)
+		__refill_fl(adap, &qs->fl[1]);
+
+unlock:
+	spin_unlock_irq(lock);
+out:
+	mod_timer(&qs->rx_reclaim_timer, jiffies + RX_RECLAIM_PERIOD);
 }
 
 /**
@@ -2906,10 +2959,10 @@ int t3_sge_alloc_qset(struct adapter *adapter, unsigned int id, int nports,
 {
 	int i, avail, ret = -ENOMEM;
 	struct sge_qset *q = &adapter->sge.qs[id];
-	struct net_lro_mgr *lro_mgr = &q->lro_mgr;
 
 	init_qset_cntxt(q, id);
-	setup_timer(&q->tx_reclaim_timer, sge_timer_cb, (unsigned long)q);
+	setup_timer(&q->tx_reclaim_timer, sge_timer_tx, (unsigned long)q);
+	setup_timer(&q->rx_reclaim_timer, sge_timer_rx, (unsigned long)q);
 
 	q->fl[0].desc = alloc_ring(adapter->pdev, p->fl_size,
 				   sizeof(struct rx_desc),
@@ -2985,25 +3038,23 @@ int t3_sge_alloc_qset(struct adapter *adapter, unsigned int id, int nports,
 	q->fl[1].use_pages = FL1_PG_CHUNK_SIZE > 0;
 	q->fl[0].order = FL0_PG_ORDER;
 	q->fl[1].order = FL1_PG_ORDER;
+	q->fl[0].alloc_size = FL0_PG_ALLOC_SIZE;
+	q->fl[1].alloc_size = FL1_PG_ALLOC_SIZE;
 
-	q->lro_frag_tbl = kcalloc(MAX_FRAME_SIZE / FL1_PG_CHUNK_SIZE + 1,
-				  sizeof(struct skb_frag_struct),
-				  GFP_KERNEL);
-	q->lro_nfrags = q->lro_frag_len = 0;
 	spin_lock_irq(&adapter->sge.reg_lock);
 
 	/* FL threshold comparison uses < */
 	ret = t3_sge_init_rspcntxt(adapter, q->rspq.cntxt_id, irq_vec_idx,
 				   q->rspq.phys_addr, q->rspq.size,
-				   q->fl[0].buf_size, 1, 0);
+				   q->fl[0].buf_size - SGE_PG_RSVD, 1, 0);
 	if (ret)
 		goto err_unlock;
 
 	for (i = 0; i < SGE_RXQ_PER_SET; ++i) {
 		ret = t3_sge_init_flcntxt(adapter, q->fl[i].cntxt_id, 0,
 					  q->fl[i].phys_addr, q->fl[i].size,
-					  q->fl[i].buf_size, p->cong_thres, 1,
-					  0);
+					  q->fl[i].buf_size - SGE_PG_RSVD,
+					  p->cong_thres, 1, 0);
 		if (ret)
 			goto err_unlock;
 	}
@@ -3041,8 +3092,6 @@ int t3_sge_alloc_qset(struct adapter *adapter, unsigned int id, int nports,
 	q->tx_q = netdevq;
 	t3_update_qset_coalesce(q, p);
 
-	init_lro_mgr(q, lro_mgr);
-
 	avail = refill_fl(adapter, &q->fl[0], q->fl[0].size,
 			  GFP_KERNEL | __GFP_COMP);
 	if (!avail) {
@@ -3063,7 +3112,6 @@ int t3_sge_alloc_qset(struct adapter *adapter, unsigned int id, int nports,
 	t3_write_reg(adapter, A_SG_GTS, V_RSPQ(q->rspq.cntxt_id) |
 		     V_NEWTIMER(q->rspq.holdoff_tmr));
 
-	mod_timer(&q->tx_reclaim_timer, jiffies + TX_RECLAIM_PERIOD);
 	return 0;
 
 err_unlock:
@@ -3074,6 +3122,27 @@ err:
 }
 
 /**
+ *      t3_start_sge_timers - start SGE timer call backs
+ *      @adap: the adapter
+ *
+ *      Starts each SGE queue set's timer call back
+ */
+void t3_start_sge_timers(struct adapter *adap)
+{
+	int i;
+
+	for (i = 0; i < SGE_QSETS; ++i) {
+		struct sge_qset *q = &adap->sge.qs[i];
+
+	if (q->tx_reclaim_timer.function)
+		mod_timer(&q->tx_reclaim_timer, jiffies + TX_RECLAIM_PERIOD);
+
+	if (q->rx_reclaim_timer.function)
+		mod_timer(&q->rx_reclaim_timer, jiffies + RX_RECLAIM_PERIOD);
+	}
+}
+
+/**
  *	t3_stop_sge_timers - stop SGE timer call backs
  *	@adap: the adapter
  *
@@ -3088,6 +3157,8 @@ void t3_stop_sge_timers(struct adapter *adap)
 
 		if (q->tx_reclaim_timer.function)
 			del_timer_sync(&q->tx_reclaim_timer);
+		if (q->rx_reclaim_timer.function)
+			del_timer_sync(&q->rx_reclaim_timer);
 	}
 }
 
diff --git a/drivers/net/cxgb3/t3_hw.c b/drivers/net/cxgb3/t3_hw.c
index ac2a974dfe37..4f68aeb2679a 100644
--- a/drivers/net/cxgb3/t3_hw.c
+++ b/drivers/net/cxgb3/t3_hw.c
@@ -493,20 +493,20 @@ int t3_phy_lasi_intr_handler(struct cphy *phy)
 }
 
 static const struct adapter_info t3_adap_info[] = {
-	{2, 0,
+	{1, 1, 0,
 	 F_GPIO2_OEN | F_GPIO4_OEN |
 	 F_GPIO2_OUT_VAL | F_GPIO4_OUT_VAL, { S_GPIO3, S_GPIO5 }, 0,
 	 &mi1_mdio_ops, "Chelsio PE9000"},
-	{2, 0,
+	{1, 1, 0,
 	 F_GPIO2_OEN | F_GPIO4_OEN |
 	 F_GPIO2_OUT_VAL | F_GPIO4_OUT_VAL, { S_GPIO3, S_GPIO5 }, 0,
 	 &mi1_mdio_ops, "Chelsio T302"},
-	{1, 0,
+	{1, 0, 0,
 	 F_GPIO1_OEN | F_GPIO6_OEN | F_GPIO7_OEN | F_GPIO10_OEN |
 	 F_GPIO11_OEN | F_GPIO1_OUT_VAL | F_GPIO6_OUT_VAL | F_GPIO10_OUT_VAL,
 	 { 0 }, SUPPORTED_10000baseT_Full | SUPPORTED_AUI,
 	 &mi1_mdio_ext_ops, "Chelsio T310"},
-	{2, 0,
+	{1, 1, 0,
 	 F_GPIO1_OEN | F_GPIO2_OEN | F_GPIO4_OEN | F_GPIO5_OEN | F_GPIO6_OEN |
 	 F_GPIO7_OEN | F_GPIO10_OEN | F_GPIO11_OEN | F_GPIO1_OUT_VAL |
 	 F_GPIO5_OUT_VAL | F_GPIO6_OUT_VAL | F_GPIO10_OUT_VAL,
@@ -514,7 +514,7 @@ static const struct adapter_info t3_adap_info[] = {
 	 &mi1_mdio_ext_ops, "Chelsio T320"},
 	{},
 	{},
-	{1, 0,
+	{1, 0, 0,
 	 F_GPIO1_OEN | F_GPIO2_OEN | F_GPIO4_OEN | F_GPIO6_OEN | F_GPIO7_OEN |
 	 F_GPIO10_OEN | F_GPIO1_OUT_VAL | F_GPIO6_OUT_VAL | F_GPIO10_OUT_VAL,
 	 { S_GPIO9 }, SUPPORTED_10000baseT_Full | SUPPORTED_AUI,
@@ -1153,6 +1153,38 @@ int t3_cim_ctl_blk_read(struct adapter *adap, unsigned int addr,
 	return ret;
 }
 
+static void t3_gate_rx_traffic(struct cmac *mac, u32 *rx_cfg,
+			       u32 *rx_hash_high, u32 *rx_hash_low)
+{
+	/* stop Rx unicast traffic */
+	t3_mac_disable_exact_filters(mac);
+
+	/* stop broadcast, multicast, promiscuous mode traffic */
+	*rx_cfg = t3_read_reg(mac->adapter, A_XGM_RX_CFG);
+	t3_set_reg_field(mac->adapter, A_XGM_RX_CFG,
+			 F_ENHASHMCAST | F_DISBCAST | F_COPYALLFRAMES,
+			 F_DISBCAST);
+
+	*rx_hash_high = t3_read_reg(mac->adapter, A_XGM_RX_HASH_HIGH);
+	t3_write_reg(mac->adapter, A_XGM_RX_HASH_HIGH, 0);
+
+	*rx_hash_low = t3_read_reg(mac->adapter, A_XGM_RX_HASH_LOW);
+	t3_write_reg(mac->adapter, A_XGM_RX_HASH_LOW, 0);
+
+	/* Leave time to drain max RX fifo */
+	msleep(1);
+}
+
+static void t3_open_rx_traffic(struct cmac *mac, u32 rx_cfg,
+			       u32 rx_hash_high, u32 rx_hash_low)
+{
+	t3_mac_enable_exact_filters(mac);
+	t3_set_reg_field(mac->adapter, A_XGM_RX_CFG,
+			 F_ENHASHMCAST | F_DISBCAST | F_COPYALLFRAMES,
+			 rx_cfg);
+	t3_write_reg(mac->adapter, A_XGM_RX_HASH_HIGH, rx_hash_high);
+	t3_write_reg(mac->adapter, A_XGM_RX_HASH_LOW, rx_hash_low);
+}
 
 /**
  *	t3_link_changed - handle interface link changes
@@ -1173,6 +1205,23 @@ void t3_link_changed(struct adapter *adapter, int port_id)
 
 	phy->ops->get_link_status(phy, &link_ok, &speed, &duplex, &fc);
 
+	if (!lc->link_ok && link_ok) {
+		u32 rx_cfg, rx_hash_high, rx_hash_low;
+		u32 status;
+
+		t3_xgm_intr_enable(adapter, port_id);
+		t3_gate_rx_traffic(mac, &rx_cfg, &rx_hash_high, &rx_hash_low);
+		t3_write_reg(adapter, A_XGM_RX_CTRL + mac->offset, 0);
+		t3_mac_enable(mac, MAC_DIRECTION_RX);
+
+		status = t3_read_reg(adapter, A_XGM_INT_STATUS + mac->offset);
+		if (status & F_LINKFAULTCHANGE) {
+			mac->stats.link_faults++;
+			pi->link_fault = 1;
+		}
+		t3_open_rx_traffic(mac, rx_cfg, rx_hash_high, rx_hash_low);
+	}
+
 	if (lc->requested_fc & PAUSE_AUTONEG)
 		fc &= lc->requested_fc;
 	else
@@ -1202,6 +1251,54 @@ void t3_link_changed(struct adapter *adapter, int port_id)
 	t3_os_link_changed(adapter, port_id, link_ok, speed, duplex, fc);
 }
 
+void t3_link_fault(struct adapter *adapter, int port_id)
+{
+	struct port_info *pi = adap2pinfo(adapter, port_id);
+	struct cmac *mac = &pi->mac;
+	struct cphy *phy = &pi->phy;
+	struct link_config *lc = &pi->link_config;
+	int link_ok, speed, duplex, fc, link_fault;
+	u32 rx_cfg, rx_hash_high, rx_hash_low;
+
+	t3_gate_rx_traffic(mac, &rx_cfg, &rx_hash_high, &rx_hash_low);
+
+	if (adapter->params.rev > 0 && uses_xaui(adapter))
+		t3_write_reg(adapter, A_XGM_XAUI_ACT_CTRL + mac->offset, 0);
+
+	t3_write_reg(adapter, A_XGM_RX_CTRL + mac->offset, 0);
+	t3_mac_enable(mac, MAC_DIRECTION_RX);
+
+	t3_open_rx_traffic(mac, rx_cfg, rx_hash_high, rx_hash_low);
+
+	link_fault = t3_read_reg(adapter,
+				 A_XGM_INT_STATUS + mac->offset);
+	link_fault &= F_LINKFAULTCHANGE;
+
+	phy->ops->get_link_status(phy, &link_ok, &speed, &duplex, &fc);
+
+	if (link_fault) {
+		lc->link_ok = 0;
+		lc->speed = SPEED_INVALID;
+		lc->duplex = DUPLEX_INVALID;
+
+		t3_os_link_fault(adapter, port_id, 0);
+
+		/* Account link faults only when the phy reports a link up */
+		if (link_ok)
+			mac->stats.link_faults++;
+	} else {
+		if (link_ok)
+			t3_write_reg(adapter, A_XGM_XAUI_ACT_CTRL + mac->offset,
+				     F_TXACTENABLE | F_RXEN);
+
+		pi->link_fault = 0;
+		lc->link_ok = (unsigned char)link_ok;
+		lc->speed = speed < 0 ? SPEED_INVALID : speed;
+		lc->duplex = duplex < 0 ? DUPLEX_INVALID : duplex;
+		t3_os_link_fault(adapter, port_id, link_ok);
+	}
+}
+
 /**
  *	t3_link_start - apply link configuration to MAC/PHY
  *	@phy: the PHY to setup
@@ -1323,7 +1420,7 @@ static int t3_handle_intr_status(struct adapter *adapter, unsigned int reg,
 #define MC7_INTR_MASK (F_AE | F_UE | F_CE | V_PE(M_PE))
 #define XGM_INTR_MASK (V_TXFIFO_PRTY_ERR(M_TXFIFO_PRTY_ERR) | \
 		       V_RXFIFO_PRTY_ERR(M_RXFIFO_PRTY_ERR) | \
-		       F_TXFIFO_UNDERRUN | F_RXFIFO_OVERFLOW)
+		       F_TXFIFO_UNDERRUN)
 #define PCIX_INTR_MASK (F_MSTDETPARERR | F_SIGTARABT | F_RCVTARABT | \
 			F_RCVMSTABT | F_SIGSYSERR | F_DETPARERR | \
 			F_SPLCMPDIS | F_UNXSPLCMP | F_RCVSPLCMPERR | \
@@ -1360,11 +1457,11 @@ static int t3_handle_intr_status(struct adapter *adapter, unsigned int reg,
 		       V_TX1TPPARERRENB(M_TX1TPPARERRENB) | \
 		       V_RXTPPARERRENB(M_RXTPPARERRENB) | \
 		       V_MCAPARERRENB(M_MCAPARERRENB))
+#define XGM_EXTRA_INTR_MASK (F_LINKFAULTCHANGE)
 #define PL_INTR_MASK (F_T3DBG | F_XGMAC0_0 | F_XGMAC0_1 | F_MC5A | F_PM1_TX | \
 		      F_PM1_RX | F_ULP2_TX | F_ULP2_RX | F_TP1 | F_CIM | \
 		      F_MC7_CM | F_MC7_PMTX | F_MC7_PMRX | F_SGE3 | F_PCIM0 | \
 		      F_MPS0 | F_CPL_SWITCH)
-
 /*
  * Interrupt handler for the PCIX1 module.
  */
@@ -1695,7 +1792,14 @@ static void mc7_intr_handler(struct mc7 *mc7)
 static int mac_intr_handler(struct adapter *adap, unsigned int idx)
 {
 	struct cmac *mac = &adap2pinfo(adap, idx)->mac;
-	u32 cause = t3_read_reg(adap, A_XGM_INT_CAUSE + mac->offset);
+	/*
+	 * We mask out interrupt causes for which we're not taking interrupts.
+	 * This allows us to use polling logic to monitor some of the other
+	 * conditions when taking interrupts would impose too much load on the
+	 * system.
+	 */
+	u32 cause = t3_read_reg(adap, A_XGM_INT_CAUSE + mac->offset) &
+		    ~F_RXFIFO_OVERFLOW;
 
 	if (cause & V_TXFIFO_PRTY_ERR(M_TXFIFO_PRTY_ERR)) {
 		mac->stats.tx_fifo_parity_err++;
@@ -1715,10 +1819,20 @@ static int mac_intr_handler(struct adapter *adap, unsigned int idx)
 		mac->stats.xaui_pcs_ctc_err++;
 	if (cause & F_XAUIPCSALIGNCHANGE)
 		mac->stats.xaui_pcs_align_change++;
+	if (cause & F_XGM_INT) {
+		t3_set_reg_field(adap,
+				 A_XGM_INT_ENABLE + mac->offset,
+				 F_XGM_INT, 0);
+		mac->stats.link_faults++;
+
+		t3_os_link_fault_handler(adap, idx);
+	}
 
 	t3_write_reg(adap, A_XGM_INT_CAUSE + mac->offset, cause);
+
 	if (cause & XGM_INTR_FATAL)
 		t3_fatal_err(adap);
+
 	return cause != 0;
 }
 
@@ -1924,6 +2038,22 @@ void t3_intr_clear(struct adapter *adapter)
 	t3_read_reg(adapter, A_PL_INT_CAUSE0);	/* flush */
 }
 
+void t3_xgm_intr_enable(struct adapter *adapter, int idx)
+{
+	struct port_info *pi = adap2pinfo(adapter, idx);
+
+	t3_write_reg(adapter, A_XGM_XGM_INT_ENABLE + pi->mac.offset,
+		     XGM_EXTRA_INTR_MASK);
+}
+
+void t3_xgm_intr_disable(struct adapter *adapter, int idx)
+{
+	struct port_info *pi = adap2pinfo(adapter, idx);
+
+	t3_write_reg(adapter, A_XGM_XGM_INT_DISABLE + pi->mac.offset,
+		     0x7ff);
+}
+
 /**
  *	t3_port_intr_enable - enable port-specific interrupts
  *	@adapter: associated adapter
@@ -1989,16 +2119,40 @@ void t3_port_intr_clear(struct adapter *adapter, int idx)
 static int t3_sge_write_context(struct adapter *adapter, unsigned int id,
 				unsigned int type)
 {
-	t3_write_reg(adapter, A_SG_CONTEXT_MASK0, 0xffffffff);
-	t3_write_reg(adapter, A_SG_CONTEXT_MASK1, 0xffffffff);
-	t3_write_reg(adapter, A_SG_CONTEXT_MASK2, 0xffffffff);
-	t3_write_reg(adapter, A_SG_CONTEXT_MASK3, 0xffffffff);
+	if (type == F_RESPONSEQ) {
+		/*
+		 * Can't write the Response Queue Context bits for
+		 * Interrupt Armed or the Reserve bits after the chip
+		 * has been initialized out of reset.  Writing to these
+		 * bits can confuse the hardware.
+		 */
+		t3_write_reg(adapter, A_SG_CONTEXT_MASK0, 0xffffffff);
+		t3_write_reg(adapter, A_SG_CONTEXT_MASK1, 0xffffffff);
+		t3_write_reg(adapter, A_SG_CONTEXT_MASK2, 0x17ffffff);
+		t3_write_reg(adapter, A_SG_CONTEXT_MASK3, 0xffffffff);
+	} else {
+		t3_write_reg(adapter, A_SG_CONTEXT_MASK0, 0xffffffff);
+		t3_write_reg(adapter, A_SG_CONTEXT_MASK1, 0xffffffff);
+		t3_write_reg(adapter, A_SG_CONTEXT_MASK2, 0xffffffff);
+		t3_write_reg(adapter, A_SG_CONTEXT_MASK3, 0xffffffff);
+	}
 	t3_write_reg(adapter, A_SG_CONTEXT_CMD,
 		     V_CONTEXT_CMD_OPCODE(1) | type | V_CONTEXT(id));
 	return t3_wait_op_done(adapter, A_SG_CONTEXT_CMD, F_CONTEXT_CMD_BUSY,
 			       0, SG_CONTEXT_CMD_ATTEMPTS, 1);
 }
 
+/**
+ *	clear_sge_ctxt - completely clear an SGE context
+ *	@adapter: the adapter
+ *	@id: the context id
+ *	@type: the context type
+ *
+ *	Completely clear an SGE context.  Used predominantly at post-reset
+ *	initialization.  Note in particular that we don't skip writing to any
+ *	"sensitive bits" in the contexts the way that t3_sge_write_context()
+ *	does ...
+ */
 static int clear_sge_ctxt(struct adapter *adap, unsigned int id,
 			  unsigned int type)
 {
@@ -2006,7 +2160,14 @@ static int clear_sge_ctxt(struct adapter *adap, unsigned int id,
 	t3_write_reg(adap, A_SG_CONTEXT_DATA1, 0);
 	t3_write_reg(adap, A_SG_CONTEXT_DATA2, 0);
 	t3_write_reg(adap, A_SG_CONTEXT_DATA3, 0);
-	return t3_sge_write_context(adap, id, type);
+	t3_write_reg(adap, A_SG_CONTEXT_MASK0, 0xffffffff);
+	t3_write_reg(adap, A_SG_CONTEXT_MASK1, 0xffffffff);
+	t3_write_reg(adap, A_SG_CONTEXT_MASK2, 0xffffffff);
+	t3_write_reg(adap, A_SG_CONTEXT_MASK3, 0xffffffff);
+	t3_write_reg(adap, A_SG_CONTEXT_CMD,
+		     V_CONTEXT_CMD_OPCODE(1) | type | V_CONTEXT(id));
+	return t3_wait_op_done(adap, A_SG_CONTEXT_CMD, F_CONTEXT_CMD_BUSY,
+			       0, SG_CONTEXT_CMD_ATTEMPTS, 1);
 }
 
 /**
@@ -2590,10 +2751,10 @@ static void tp_config(struct adapter *adap, const struct tp_params *p)
 		     F_TCPCHECKSUMOFFLOAD | V_IPTTL(64));
 	t3_write_reg(adap, A_TP_TCP_OPTIONS, V_MTUDEFAULT(576) |
 		     F_MTUENABLE | V_WINDOWSCALEMODE(1) |
-		     V_TIMESTAMPSMODE(0) | V_SACKMODE(1) | V_SACKRX(1));
+		     V_TIMESTAMPSMODE(1) | V_SACKMODE(1) | V_SACKRX(1));
 	t3_write_reg(adap, A_TP_DACK_CONFIG, V_AUTOSTATE3(1) |
 		     V_AUTOSTATE2(1) | V_AUTOSTATE1(0) |
-		     V_BYTETHRESHOLD(16384) | V_MSSTHRESHOLD(2) |
+		     V_BYTETHRESHOLD(26880) | V_MSSTHRESHOLD(2) |
 		     F_AUTOCAREFUL | F_AUTOENABLE | V_DACK_MODE(1));
 	t3_set_reg_field(adap, A_TP_IN_CONFIG, F_RXFBARBPRIO | F_TXFBARBPRIO,
 			 F_IPV6ENABLE | F_NICMODE);
@@ -3057,20 +3218,22 @@ int t3_mps_set_active_ports(struct adapter *adap, unsigned int port_mask)
 }
 
 /*
- * Perform the bits of HW initialization that are dependent on the number
- * of available ports.
+ * Perform the bits of HW initialization that are dependent on the Tx
+ * channels being used.
  */
-static void init_hw_for_avail_ports(struct adapter *adap, int nports)
+static void chan_init_hw(struct adapter *adap, unsigned int chan_map)
 {
 	int i;
 
-	if (nports == 1) {
+	if (chan_map != 3) {                                 /* one channel */
 		t3_set_reg_field(adap, A_ULPRX_CTL, F_ROUND_ROBIN, 0);
 		t3_set_reg_field(adap, A_ULPTX_CONFIG, F_CFG_RR_ARB, 0);
-		t3_write_reg(adap, A_MPS_CFG, F_TPRXPORTEN | F_TPTXPORT0EN |
-			     F_PORT0ACTIVE | F_ENFORCEPKT);
-		t3_write_reg(adap, A_PM1_TX_CFG, 0xffffffff);
-	} else {
+		t3_write_reg(adap, A_MPS_CFG, F_TPRXPORTEN | F_ENFORCEPKT |
+			     (chan_map == 1 ? F_TPTXPORT0EN | F_PORT0ACTIVE :
+					      F_TPTXPORT1EN | F_PORT1ACTIVE));
+		t3_write_reg(adap, A_PM1_TX_CFG,
+			     chan_map == 1 ? 0xffffffff : 0);
+	} else {                                             /* two channels */
 		t3_set_reg_field(adap, A_ULPRX_CTL, 0, F_ROUND_ROBIN);
 		t3_set_reg_field(adap, A_ULPTX_CONFIG, 0, F_CFG_RR_ARB);
 		t3_write_reg(adap, A_ULPTX_DMA_WEIGHT,
@@ -3378,7 +3541,7 @@ int t3_init_hw(struct adapter *adapter, u32 fw_params)
 	t3_write_reg(adapter, A_PM1_RX_CFG, 0xffffffff);
 	t3_write_reg(adapter, A_PM1_RX_MODE, 0);
 	t3_write_reg(adapter, A_PM1_TX_MODE, 0);
-	init_hw_for_avail_ports(adapter, adapter->params.nports);
+	chan_init_hw(adapter, adapter->params.chan_map);
 	t3_sge_init(adapter, &adapter->params.sge);
 
 	t3_write_reg(adapter, A_T3DBG_GPIO_ACT_LOW, calc_gpio_intr(adapter));
@@ -3615,9 +3778,18 @@ int t3_prep_adapter(struct adapter *adapter, const struct adapter_info *ai,
 	get_pci_mode(adapter, &adapter->params.pci);
 
 	adapter->params.info = ai;
-	adapter->params.nports = ai->nports;
+	adapter->params.nports = ai->nports0 + ai->nports1;
+	adapter->params.chan_map = (!!ai->nports0) | (!!ai->nports1 << 1);
 	adapter->params.rev = t3_read_reg(adapter, A_PL_REV);
-	adapter->params.linkpoll_period = 0;
+	/*
+	 * We used to only run the "adapter check task" once a second if
+	 * we had PHYs which didn't support interrupts (we would check
+	 * their link status once a second).  Now we check other conditions
+	 * in that routine which could potentially impose a very high
+	 * interrupt load on the system.  As such, we now always scan the
+	 * adapter state once a second ...
+	 */
+	adapter->params.linkpoll_period = 10;
 	adapter->params.stats_update_period = is_10G(adapter) ?
 	    MAC_STATS_ACCUM_SECS : (MAC_STATS_ACCUM_SECS * 10);
 	adapter->params.pci.vpd_cap_addr =
@@ -3638,7 +3810,7 @@ int t3_prep_adapter(struct adapter *adapter, const struct adapter_info *ai,
 		mc7_prep(adapter, &adapter->pmtx, MC7_PMTX_BASE_ADDR, "PMTX");
 		mc7_prep(adapter, &adapter->cm, MC7_CM_BASE_ADDR, "CM");
 
-		p->nchan = ai->nports;
+		p->nchan = adapter->params.chan_map == 3 ? 2 : 1;
 		p->pmrx_size = t3_mc7_size(&adapter->pmrx);
 		p->pmtx_size = t3_mc7_size(&adapter->pmtx);
 		p->cm_size = t3_mc7_size(&adapter->cm);
@@ -3707,7 +3879,14 @@ int t3_prep_adapter(struct adapter *adapter, const struct adapter_info *ai,
 		       ETH_ALEN);
 		init_link_config(&p->link_config, p->phy.caps);
 		p->phy.ops->power_down(&p->phy, 1);
-		if (!(p->phy.caps & SUPPORTED_IRQ))
+
+		/*
+		 * If the PHY doesn't support interrupts for link status
+		 * changes, schedule a scan of the adapter links at least
+		 * once a second.
+		 */
+		if (!(p->phy.caps & SUPPORTED_IRQ) &&
+		    adapter->params.linkpoll_period > 10)
 			adapter->params.linkpoll_period = 10;
 	}
 
diff --git a/drivers/net/cxgb3/version.h b/drivers/net/cxgb3/version.h
index b1b25c37aa10..7bf963ec5548 100644
--- a/drivers/net/cxgb3/version.h
+++ b/drivers/net/cxgb3/version.h
@@ -35,10 +35,10 @@
 #define DRV_DESC "Chelsio T3 Network Driver"
 #define DRV_NAME "cxgb3"
 /* Driver version */
-#define DRV_VERSION "1.1.1-ko"
+#define DRV_VERSION "1.1.2-ko"
 
 /* Firmware version */
 #define FW_VERSION_MAJOR 7
-#define FW_VERSION_MINOR 0
+#define FW_VERSION_MINOR 1
 #define FW_VERSION_MICRO 0
 #endif				/* __CHELSIO_VERSION_H */
diff --git a/drivers/net/cxgb3/xgmac.c b/drivers/net/cxgb3/xgmac.c
index 9d7786937aad..f87f9435049f 100644
--- a/drivers/net/cxgb3/xgmac.c
+++ b/drivers/net/cxgb3/xgmac.c
@@ -150,7 +150,8 @@ int t3_mac_reset(struct cmac *mac)
 static int t3b2_mac_reset(struct cmac *mac)
 {
 	struct adapter *adap = mac->adapter;
-	unsigned int oft = mac->offset;
+	unsigned int oft = mac->offset, store;
+	int idx = macidx(mac);
 	u32 val;
 
 	if (!macidx(mac))
@@ -158,14 +159,28 @@ static int t3b2_mac_reset(struct cmac *mac)
 	else
 		t3_set_reg_field(adap, A_MPS_CFG, F_PORT1ACTIVE, 0);
 
+	/* Stop NIC traffic to reduce the number of TXTOGGLES */
+	t3_set_reg_field(adap, A_MPS_CFG, F_ENFORCEPKT, 0);
+	/* Ensure TX drains */
+	t3_set_reg_field(adap, A_XGM_TX_CFG + oft, F_TXPAUSEEN, 0);
+
 	t3_write_reg(adap, A_XGM_RESET_CTRL + oft, F_MAC_RESET_);
 	t3_read_reg(adap, A_XGM_RESET_CTRL + oft);    /* flush */
 
+	/* Store A_TP_TX_DROP_CFG_CH0 */
+	t3_write_reg(adap, A_TP_PIO_ADDR, A_TP_TX_DROP_CFG_CH0 + idx);
+	store = t3_read_reg(adap, A_TP_TX_DROP_CFG_CH0 + idx);
+
 	msleep(10);
 
+	/* Change DROP_CFG to 0xc0000011 */
+	t3_write_reg(adap, A_TP_PIO_ADDR, A_TP_TX_DROP_CFG_CH0 + idx);
+	t3_write_reg(adap, A_TP_PIO_DATA, 0xc0000011);
+
 	/* Check for xgm Rx fifo empty */
+	/* Increased loop count to 1000 from 5 cover 1G and 100Mbps case */
 	if (t3_wait_op_done(adap, A_XGM_RX_MAX_PKT_SIZE_ERR_CNT + oft,
-			    0x80000000, 1, 5, 2)) {
+			    0x80000000, 1, 1000, 2)) {
 		CH_ERR(adap, "MAC %d Rx fifo drain failed\n",
 		       macidx(mac));
 		return -1;
@@ -191,11 +206,21 @@ static int t3b2_mac_reset(struct cmac *mac)
 		     F_DISPAUSEFRAMES | F_EN1536BFRAMES |
 		     F_RMFCS | F_ENJUMBO | F_ENHASHMCAST);
 
-	if (!macidx(mac))
+	/* Restore the DROP_CFG */
+	t3_write_reg(adap, A_TP_PIO_ADDR, A_TP_TX_DROP_CFG_CH0 + idx);
+	t3_write_reg(adap, A_TP_PIO_DATA, store);
+
+	if (!idx)
 		t3_set_reg_field(adap, A_MPS_CFG, 0, F_PORT0ACTIVE);
 	else
 		t3_set_reg_field(adap, A_MPS_CFG, 0, F_PORT1ACTIVE);
 
+	/* re-enable nic traffic */
+	t3_set_reg_field(adap, A_MPS_CFG, F_ENFORCEPKT, 1);
+
+	/*  Set: re-enable NIC traffic */
+	t3_set_reg_field(adap, A_MPS_CFG, F_ENFORCEPKT, 1);
+
 	return 0;
 }
 
@@ -236,7 +261,7 @@ int t3_mac_set_num_ucast(struct cmac *mac, int n)
 	return 0;
 }
 
-static void disable_exact_filters(struct cmac *mac)
+void t3_mac_disable_exact_filters(struct cmac *mac)
 {
 	unsigned int i, reg = mac->offset + A_XGM_RX_EXACT_MATCH_LOW_1;
 
@@ -247,7 +272,7 @@ static void disable_exact_filters(struct cmac *mac)
 	t3_read_reg(mac->adapter, A_XGM_RX_EXACT_MATCH_LOW_1);	/* flush */
 }
 
-static void enable_exact_filters(struct cmac *mac)
+void t3_mac_enable_exact_filters(struct cmac *mac)
 {
 	unsigned int i, reg = mac->offset + A_XGM_RX_EXACT_MATCH_HIGH_1;
 
@@ -332,18 +357,9 @@ int t3_mac_set_mtu(struct cmac *mac, unsigned int mtu)
 		return -EINVAL;
 	t3_write_reg(adap, A_XGM_RX_MAX_PKT_SIZE + mac->offset, mtu);
 
-	/*
-	 * Adjust the PAUSE frame watermarks.  We always set the LWM, and the
-	 * HWM only if flow-control is enabled.
-	 */
-	hwm = max_t(unsigned int, MAC_RXFIFO_SIZE - 3 * mtu,
-		    MAC_RXFIFO_SIZE * 38 / 100);
-	hwm = min(hwm, MAC_RXFIFO_SIZE - 8192);
-	lwm = min(3 * (int)mtu, MAC_RXFIFO_SIZE / 4);
-
 	if (adap->params.rev >= T3_REV_B2 &&
 	    (t3_read_reg(adap, A_XGM_RX_CTRL + mac->offset) & F_RXEN)) {
-		disable_exact_filters(mac);
+		t3_mac_disable_exact_filters(mac);
 		v = t3_read_reg(adap, A_XGM_RX_CFG + mac->offset);
 		t3_set_reg_field(adap, A_XGM_RX_CFG + mac->offset,
 				 F_ENHASHMCAST | F_COPYALLFRAMES, F_DISBCAST);
@@ -355,14 +371,14 @@ int t3_mac_set_mtu(struct cmac *mac, unsigned int mtu)
 		if (t3_wait_op_done(adap, reg + mac->offset,
 				    F_RXFIFO_EMPTY, 1, 20, 5)) {
 			t3_write_reg(adap, A_XGM_RX_CFG + mac->offset, v);
-			enable_exact_filters(mac);
+			t3_mac_enable_exact_filters(mac);
 			return -EIO;
 		}
 		t3_set_reg_field(adap, A_XGM_RX_MAX_PKT_SIZE + mac->offset,
 				 V_RXMAXPKTSIZE(M_RXMAXPKTSIZE),
 				 V_RXMAXPKTSIZE(mtu));
 		t3_write_reg(adap, A_XGM_RX_CFG + mac->offset, v);
-		enable_exact_filters(mac);
+		t3_mac_enable_exact_filters(mac);
 	} else
 		t3_set_reg_field(adap, A_XGM_RX_MAX_PKT_SIZE + mac->offset,
 				 V_RXMAXPKTSIZE(M_RXMAXPKTSIZE),
@@ -452,9 +468,12 @@ int t3_mac_enable(struct cmac *mac, int which)
 
 	if (which & MAC_DIRECTION_TX) {
 		t3_write_reg(adap, A_TP_PIO_ADDR, A_TP_TX_DROP_CFG_CH0 + idx);
-		t3_write_reg(adap, A_TP_PIO_DATA, 0xc0ede401);
+		t3_write_reg(adap, A_TP_PIO_DATA,
+			     adap->params.rev == T3_REV_C ?
+			     0xc4ffff01 : 0xc0ede401);
 		t3_write_reg(adap, A_TP_PIO_ADDR, A_TP_TX_DROP_MODE);
-		t3_set_reg_field(adap, A_TP_PIO_DATA, 1 << idx, 1 << idx);
+		t3_set_reg_field(adap, A_TP_PIO_DATA, 1 << idx,
+				 adap->params.rev == T3_REV_C ? 0 : 1 << idx);
 
 		t3_write_reg(adap, A_XGM_TX_CTRL + oft, F_TXEN);
 
@@ -510,15 +529,12 @@ int t3b2_mac_watchdog_task(struct cmac *mac)
 	struct adapter *adap = mac->adapter;
 	struct mac_stats *s = &mac->stats;
 	unsigned int tx_tcnt, tx_xcnt;
-	unsigned int tx_mcnt = s->tx_frames;
-	unsigned int rx_mcnt = s->rx_frames;
-	unsigned int rx_xcnt;
+	u64 tx_mcnt = s->tx_frames;
 	int status;
 
 	status = 0;
 	tx_xcnt = 1;		/* By default tx_xcnt is making progress */
 	tx_tcnt = mac->tx_tcnt;	/* If tx_mcnt is progressing ignore tx_tcnt */
-	rx_xcnt = 1;		/* By default rx_xcnt is making progress */
 	if (tx_mcnt == mac->tx_mcnt && mac->rx_pause == s->rx_pause) {
 		tx_xcnt = (G_TXSPI4SOPCNT(t3_read_reg(adap,
 						A_XGM_TX_SPI4_SOP_EOP_CNT +
@@ -529,11 +545,11 @@ int t3b2_mac_watchdog_task(struct cmac *mac)
 			tx_tcnt = (G_TXDROPCNTCH0RCVD(t3_read_reg(adap,
 						      A_TP_PIO_DATA)));
 		} else {
-			goto rxcheck;
+			goto out;
 		}
 	} else {
 		mac->toggle_cnt = 0;
-		goto rxcheck;
+		goto out;
 	}
 
 	if ((tx_tcnt != mac->tx_tcnt) && (mac->tx_xcnt == 0)) {
@@ -546,23 +562,6 @@ int t3b2_mac_watchdog_task(struct cmac *mac)
 		}
 	} else {
 		mac->toggle_cnt = 0;
-		goto rxcheck;
-	}
-
-rxcheck:
-	if (rx_mcnt != mac->rx_mcnt) {
-		rx_xcnt = (G_TXSPI4SOPCNT(t3_read_reg(adap,
-						A_XGM_RX_SPI4_SOP_EOP_CNT +
-						mac->offset))) +
-						(s->rx_fifo_ovfl -
-						 mac->rx_ocnt);
-		mac->rx_ocnt = s->rx_fifo_ovfl;
-	} else
-		goto out;
-
-	if (mac->rx_mcnt != s->rx_frames && rx_xcnt == 0 &&
-	    mac->rx_xcnt == 0) {
-		status = 2;
 		goto out;
 	}
 
@@ -570,8 +569,6 @@ out:
 	mac->tx_tcnt = tx_tcnt;
 	mac->tx_xcnt = tx_xcnt;
 	mac->tx_mcnt = s->tx_frames;
-	mac->rx_xcnt = rx_xcnt;
-	mac->rx_mcnt = s->rx_frames;
 	mac->rx_pause = s->rx_pause;
 	if (status == 1) {
 		t3_write_reg(adap, A_XGM_TX_CTRL + mac->offset, 0);