Merge commit 'origin/master' into next

Manual merge of:
	arch/powerpc/include/asm/elf.h
	drivers/i2c/busses/i2c-mpc.c

11 files changed, 796 insertions, 340 deletions

diff --git a/drivers/net/cxgb3/adapter.h b/drivers/net/cxgb3/adapter.h
index a89d8cc5120..714df2b675e 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 5c3c05da4d9..e1b22490ff5 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 db4f4f575b6..e508dc32f3e 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 bab8a934c33..2c2aaa74145 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;
 	}
 
@@ -1183,6 +1232,10 @@ static int cxgb_close(struct net_device *dev)
 	struct port_info *pi = netdev_priv(dev);
 	struct adapter *adapter = pi->adapter;
 
+	/* 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);
@@ -1309,6 +1362,7 @@ static char stats_strings[][ETH_GSTRING_LEN] = {
 	"CheckTXEnToggled   ",
 	"CheckResets        ",
 
+	"LinkFaults         ",
 };
 
 static int get_sset_count(struct net_device *dev, int sset)
@@ -1434,13 +1488,15 @@ 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 +1632,6 @@ 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 +1641,7 @@ 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 +1882,6 @@ 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 +1911,6 @@ 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 +2491,20 @@ 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) {
+			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);
+		}
 	}
 }
 
@@ -2506,6 +2549,8 @@ 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++;
 
@@ -2524,6 +2569,54 @@ 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 +2631,23 @@ 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 +2680,42 @@ void t3_os_ext_intr_handler(struct adapter *adapter)
 	spin_unlock(&adapter->work_lock);
 }
 
+static void link_fault_task(struct work_struct *work)
+{
+	struct adapter *adapter = container_of(work, struct adapter,
+					       link_fault_handler_task);
+	int i;
+
+	for_each_port(adapter, i) {
+		struct net_device *netdev = adapter->port[i];
+		struct port_info *pi = netdev_priv(netdev);
+
+		if (pi->link_fault)
+			t3_link_fault(adapter, i);
+	}
+}
+
+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;
+	queue_work(cxgb3_wq, &adapter->link_fault_handler_task);
+	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 +2724,6 @@ 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 +2775,9 @@ 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 +2822,6 @@ 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]);
-
 }
 
 /**
@@ -2734,6 +2871,9 @@ 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 +2893,7 @@ 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 +2922,25 @@ 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 && vectors < (adap->params.nports + 1))
+		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 +3006,7 @@ 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;
@@ -2935,10 +3082,11 @@ static int __devinit init_one(struct pci_dev *pdev,
 
 	INIT_LIST_HEAD(&adapter->adapter_list);
 	INIT_WORK(&adapter->ext_intr_handler_task, ext_intr_task);
+	INIT_WORK(&adapter->link_fault_handler_task, link_fault_task);
 	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 +3109,7 @@ 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 +3176,7 @@ 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 2d7f69aff1d..620d80be6aa 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 d514e5019df..a8e8e5fcdf8 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 a035d5c2444..1b5327b5a96 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 d31791f6029..26d3587f339 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, s