1 files changed, 138 insertions, 101 deletions

diff --git a/drivers/net/igb/igb_main.c b/drivers/net/igb/igb_main.c
index e25343588fc..adb09d32625 100644
--- a/drivers/net/igb/igb_main.c
+++ b/drivers/net/igb/igb_main.c
@@ -127,14 +127,48 @@ static void igb_restore_vlan(struct igb_adapter *);
 static void igb_ping_all_vfs(struct igb_adapter *);
 static void igb_msg_task(struct igb_adapter *);
 static int igb_rcv_msg_from_vf(struct igb_adapter *, u32);
-static inline void igb_set_rah_pool(struct e1000_hw *, int , int);
 static void igb_set_mc_list_pools(struct igb_adapter *, int, u16);
 static void igb_vmm_control(struct igb_adapter *);
-static inline void igb_set_vmolr(struct e1000_hw *, int);
-static inline int igb_set_vf_rlpml(struct igb_adapter *, int, int);
 static int igb_set_vf_mac(struct igb_adapter *adapter, int, unsigned char *);
 static void igb_restore_vf_multicasts(struct igb_adapter *adapter);
 
+static inline void igb_set_vmolr(struct e1000_hw *hw, int vfn)
+{
+	u32 reg_data;
+
+	reg_data = rd32(E1000_VMOLR(vfn));
+	reg_data |= E1000_VMOLR_BAM |	 /* Accept broadcast */
+	            E1000_VMOLR_ROPE |   /* Accept packets matched in UTA */
+	            E1000_VMOLR_ROMPE |  /* Accept packets matched in MTA */
+	            E1000_VMOLR_AUPE |   /* Accept untagged packets */
+	            E1000_VMOLR_STRVLAN; /* Strip vlan tags */
+	wr32(E1000_VMOLR(vfn), reg_data);
+}
+
+static inline int igb_set_vf_rlpml(struct igb_adapter *adapter, int size,
+                                 int vfn)
+{
+	struct e1000_hw *hw = &adapter->hw;
+	u32 vmolr;
+
+	vmolr = rd32(E1000_VMOLR(vfn));
+	vmolr &= ~E1000_VMOLR_RLPML_MASK;
+	vmolr |= size | E1000_VMOLR_LPE;
+	wr32(E1000_VMOLR(vfn), vmolr);
+
+	return 0;
+}
+
+static inline void igb_set_rah_pool(struct e1000_hw *hw, int pool, int entry)
+{
+	u32 reg_data;
+
+	reg_data = rd32(E1000_RAH(entry));
+	reg_data &= ~E1000_RAH_POOL_MASK;
+	reg_data |= E1000_RAH_POOL_1 << pool;;
+	wr32(E1000_RAH(entry), reg_data);
+}
+
 #ifdef CONFIG_PM
 static int igb_suspend(struct pci_dev *, pm_message_t);
 static int igb_resume(struct pci_dev *);
@@ -942,6 +976,8 @@ int igb_up(struct igb_adapter *adapter)
 	rd32(E1000_ICR);
 	igb_irq_enable(adapter);
 
+	netif_tx_start_all_queues(adapter->netdev);
+
 	/* Fire a link change interrupt to start the watchdog. */
 	wr32(E1000_ICS, E1000_ICS_LSC);
 	return 0;
@@ -994,6 +1030,11 @@ void igb_down(struct igb_adapter *adapter)
 		igb_reset(adapter);
 	igb_clean_all_tx_rings(adapter);
 	igb_clean_all_rx_rings(adapter);
+#ifdef CONFIG_IGB_DCA
+
+	/* since we reset the hardware DCA settings were cleared */
+	igb_setup_dca(adapter);
+#endif
 }
 
 void igb_reinit_locked(struct igb_adapter *adapter)
@@ -1343,6 +1384,9 @@ static int __devinit igb_probe(struct pci_dev *pdev,
 	if (pci_using_dac)
 		netdev->features |= NETIF_F_HIGHDMA;
 
+	if (adapter->hw.mac.type == e1000_82576)
+		netdev->features |= NETIF_F_SCTP_CSUM;
+
 	adapter->en_mng_pt = igb_enable_mng_pass_thru(&adapter->hw);
 
 	/* before reading the NVM, reset the controller to put the device in a
@@ -1390,8 +1434,6 @@ static int __devinit igb_probe(struct pci_dev *pdev,
 
 	igb_validate_mdi_setting(hw);
 
-	adapter->rx_csum = 1;
-
 	/* Initial Wake on LAN setting If APM wake is enabled in the EEPROM,
 	 * enable the ACPI Magic Packet filter
 	 */
@@ -1442,22 +1484,18 @@ static int __devinit igb_probe(struct pci_dev *pdev,
 	 * driver. */
 	igb_get_hw_control(adapter);
 
-	/* tell the stack to leave us alone until igb_open() is called */
-	netif_carrier_off(netdev);
-	netif_tx_stop_all_queues(netdev);
-
 	strcpy(netdev->name, "eth%d");
 	err = register_netdev(netdev);
 	if (err)
 		goto err_register;
 
+	/* carrier off reporting is important to ethtool even BEFORE open */
+	netif_carrier_off(netdev);
+
 #ifdef CONFIG_IGB_DCA
 	if (dca_add_requester(&pdev->dev) == 0) {
 		adapter->flags |= IGB_FLAG_DCA_ENABLED;
 		dev_info(&pdev->dev, "DCA enabled\n");
-		/* Always use CB2 mode, difference is masked
-		 * in the CB driver. */
-		wr32(E1000_DCA_CTRL, E1000_DCA_CTRL_DCA_MODE_CB2);
 		igb_setup_dca(adapter);
 	}
 #endif
@@ -1699,6 +1737,8 @@ static int igb_open(struct net_device *netdev)
 	if (test_bit(__IGB_TESTING, &adapter->state))
 		return -EBUSY;
 
+	netif_carrier_off(netdev);
+
 	/* allocate transmit descriptors */
 	err = igb_setup_all_tx_resources(adapter);
 	if (err)
@@ -2231,29 +2271,24 @@ static void igb_configure_rx(struct igb_adapter *adapter)
 		mrqc |= (E1000_MRQC_RSS_FIELD_IPV6_UDP_EX |
 			 E1000_MRQC_RSS_FIELD_IPV6_TCP_EX);
 
-
 		wr32(E1000_MRQC, mrqc);
-
-		/* Multiqueue and raw packet checksumming are mutually
-		 * exclusive.  Note that this not the same as TCP/IP
-		 * checksumming, which works fine. */
-		rxcsum = rd32(E1000_RXCSUM);
-		rxcsum |= E1000_RXCSUM_PCSD;
-		wr32(E1000_RXCSUM, rxcsum);
-	} else {
+	} else if (adapter->vfs_allocated_count) {
 		/* Enable multi-queue for sr-iov */
-		if (adapter->vfs_allocated_count)
-			wr32(E1000_MRQC, E1000_MRQC_ENABLE_VMDQ);
-		/* Enable Receive Checksum Offload for TCP and UDP */
-		rxcsum = rd32(E1000_RXCSUM);
-		if (adapter->rx_csum)
-			rxcsum |= E1000_RXCSUM_TUOFL | E1000_RXCSUM_IPPCSE;
-		else
-			rxcsum &= ~(E1000_RXCSUM_TUOFL | E1000_RXCSUM_IPPCSE);
-
-		wr32(E1000_RXCSUM, rxcsum);
+		wr32(E1000_MRQC, E1000_MRQC_ENABLE_VMDQ);
 	}
 
+	/* Enable Receive Checksum Offload for TCP and UDP */
+	rxcsum = rd32(E1000_RXCSUM);
+	/* Disable raw packet checksumming */
+	rxcsum |= E1000_RXCSUM_PCSD;
+
+	if (adapter->hw.mac.type == e1000_82576)
+		/* Enable Receive Checksum Offload for SCTP */
+		rxcsum |= E1000_RXCSUM_CRCOFL;
+
+	/* Don't need to set TUOFL or IPOFL, they default to 1 */
+	wr32(E1000_RXCSUM, rxcsum);
+
 	/* Set the default pool for the PF's first queue */
 	igb_configure_vt_default_pool(adapter);
 
@@ -2661,7 +2696,6 @@ static void igb_watchdog_task(struct work_struct *work)
 			}
 
 			netif_carrier_on(netdev);
-			netif_tx_wake_all_queues(netdev);
 
 			igb_ping_all_vfs(adapter);
 
@@ -2678,7 +2712,6 @@ static void igb_watchdog_task(struct work_struct *work)
 			printk(KERN_INFO "igb: %s NIC Link is Down\n",
 			       netdev->name);
 			netif_carrier_off(netdev);
-			netif_tx_stop_all_queues(netdev);
 
 			igb_ping_all_vfs(adapter);
 
@@ -2712,6 +2745,8 @@ link_up:
 			 * (Do the reset outside of interrupt context). */
 			adapter->tx_timeout_count++;
 			schedule_work(&adapter->reset_task);
+			/* return immediately since reset is imminent */
+			return;
 		}
 	}
 
@@ -2895,13 +2930,13 @@ static void igb_set_itr(struct igb_adapter *adapter)
 	switch (current_itr) {
 	/* counts and packets in update_itr are dependent on these numbers */
 	case lowest_latency:
-		new_itr = 70000;
+		new_itr = 56;  /* aka 70,000 ints/sec */
 		break;
 	case low_latency:
-		new_itr = 20000; /* aka hwitr = ~200 */
+		new_itr = 196; /* aka 20,000 ints/sec */
 		break;
 	case bulk_latency:
-		new_itr = 4000;
+		new_itr = 980; /* aka 4,000 ints/sec */
 		break;
 	default:
 		break;
@@ -2920,7 +2955,8 @@ set_itr_now:
 		 * by adding intermediate steps when interrupt rate is
 		 * increasing */
 		new_itr = new_itr > adapter->itr ?
-			     min(adapter->itr + (new_itr >> 2), new_itr) :
+			     max((new_itr * adapter->itr) /
+			         (new_itr + (adapter->itr >> 2)), new_itr) :
 			     new_itr;
 		/* Don't write the value here; it resets the adapter's
 		 * internal timer, and causes us to delay far longer than
@@ -2929,7 +2965,7 @@ set_itr_now:
 		 * ends up being correct.
 		 */
 		adapter->itr = new_itr;
-		adapter->rx_ring->itr_val = 1000000000 / (new_itr * 256);
+		adapter->rx_ring->itr_val = new_itr;
 		adapter->rx_ring->set_itr = 1;
 	}
 
@@ -3068,11 +3104,15 @@ static inline bool igb_tx_csum_adv(struct igb_adapter *adapter,
 				tu_cmd |= E1000_ADVTXD_TUCMD_IPV4;
 				if (ip_hdr(skb)->protocol == IPPROTO_TCP)
 					tu_cmd |= E1000_ADVTXD_TUCMD_L4T_TCP;
+				else if (ip_hdr(skb)->protocol == IPPROTO_SCTP)
+					tu_cmd |= E1000_ADVTXD_TUCMD_L4T_SCTP;
 				break;
 			case cpu_to_be16(ETH_P_IPV6):
 				/* XXX what about other V6 headers?? */
 				if (ipv6_hdr(skb)->nexthdr == IPPROTO_TCP)
 					tu_cmd |= E1000_ADVTXD_TUCMD_L4T_TCP;
+				else if (ipv6_hdr(skb)->nexthdr == IPPROTO_SCTP)
+					tu_cmd |= E1000_ADVTXD_TUCMD_L4T_SCTP;
 				break;
 			default:
 				if (unlikely(net_ratelimit()))
@@ -3133,8 +3173,7 @@ static inline int igb_tx_map_adv(struct igb_adapter *adapter,
 	/* set time_stamp *before* dma to help avoid a possible race */
 	buffer_info->time_stamp = jiffies;
 	buffer_info->next_to_watch = i;
-	buffer_info->dma = map[count];
-	count++;
+	buffer_info->dma = skb_shinfo(skb)->dma_head;
 
 	for (f = 0; f < skb_shinfo(skb)->nr_frags; f++) {
 		struct skb_frag_struct *frag;
@@ -3158,7 +3197,7 @@ static inline int igb_tx_map_adv(struct igb_adapter *adapter,
 	tx_ring->buffer_info[i].skb = skb;
 	tx_ring->buffer_info[first].next_to_watch = i;
 
-	return count;
+	return count + 1;
 }
 
 static inline void igb_tx_queue_adv(struct igb_adapter *adapter,
@@ -3338,7 +3377,6 @@ static int igb_xmit_frame_ring_adv(struct sk_buff *skb,
 	if (count) {
 		igb_tx_queue_adv(adapter, tx_ring, tx_flags, count,
 			         skb->len, hdr_len);
-		netdev->trans_start = jiffies;
 		/* Make sure there is space in the ring for the next send. */
 		igb_maybe_stop_tx(netdev, tx_ring, MAX_SKB_FRAGS + 4);
 	} else {
@@ -3582,8 +3620,35 @@ void igb_update_stats(struct igb_adapter *adapter)
 
 	/* Rx Errors */
 
+	if (hw->mac.type != e1000_82575) {
+		u32 rqdpc_tmp;
+		u64 rqdpc_total = 0;
+		int i;
+		/* Read out drops stats per RX queue.  Notice RQDPC (Receive
+		 * Queue Drop Packet Count) stats only gets incremented, if
+		 * the DROP_EN but it set (in the SRRCTL register for that
+		 * queue).  If DROP_EN bit is NOT set, then the some what
+		 * equivalent count is stored in RNBC (not per queue basis).
+		 * Also note the drop count is due to lack of available
+		 * descriptors.
+		 */
+		for (i = 0; i < adapter->num_rx_queues; i++) {
+			rqdpc_tmp = rd32(E1000_RQDPC(i)) & 0xFFF;
+			adapter->rx_ring[i].rx_stats.drops += rqdpc_tmp;
+			rqdpc_total += adapter->rx_ring[i].rx_stats.drops;
+		}
+		adapter->net_stats.rx_fifo_errors = rqdpc_total;
+	}
+
+	/* Note RNBC (Receive No Buffers Count) is an not an exact
+	 * drop count as the hardware FIFO might save the day.  Thats
+	 * one of the reason for saving it in rx_fifo_errors, as its
+	 * potentially not a true drop.
+	 */
+	adapter->net_stats.rx_fifo_errors += adapter->stats.rnbc;
+
 	/* RLEC on some newer hardware can be incorrect so build
-	* our own version based on RUC and ROC */
+	 * our own version based on RUC and ROC */
 	adapter->net_stats.rx_errors = adapter->stats.rxerrc +
 		adapter->stats.crcerrs + adapter->stats.algnerrc +
 		adapter->stats.ruc + adapter->stats.roc +
@@ -3767,11 +3832,15 @@ static void igb_update_tx_dca(struct igb_ring *tx_ring)
 
 static void igb_setup_dca(struct igb_adapter *adapter)
 {
+	struct e1000_hw *hw = &adapter->hw;
 	int i;
 
 	if (!(adapter->flags & IGB_FLAG_DCA_ENABLED))
 		return;
 
+	/* Always use CB2 mode, difference is masked in the CB driver. */
+	wr32(E1000_DCA_CTRL, E1000_DCA_CTRL_DCA_MODE_CB2);
+
 	for (i = 0; i < adapter->num_tx_queues; i++) {
 		adapter->tx_ring[i].cpu = -1;
 		igb_update_tx_dca(&adapter->tx_ring[i]);
@@ -4434,20 +4503,12 @@ static void igb_receive_skb(struct igb_ring *ring, u8 status,
 	bool vlan_extracted = (adapter->vlgrp && (status & E1000_RXD_STAT_VP));
 
 	skb_record_rx_queue(skb, ring->queue_index);
-	if (skb->ip_summed == CHECKSUM_UNNECESSARY) {
-		if (vlan_extracted)
-			vlan_gro_receive(&ring->napi, adapter->vlgrp,
-			                 le16_to_cpu(rx_desc->wb.upper.vlan),
-			                 skb);
-		else
-			napi_gro_receive(&ring->napi, skb);
-	} else {
-		if (vlan_extracted)
-			vlan_hwaccel_receive_skb(skb, adapter->vlgrp,
-			                  le16_to_cpu(rx_desc->wb.upper.vlan));
-		else
-			netif_receive_skb(skb);
-	}
+	if (vlan_extracted)
+		vlan_gro_receive(&ring->napi, adapter->vlgrp,
+		                 le16_to_cpu(rx_desc->wb.upper.vlan),
+		                 skb);
+	else
+		napi_gro_receive(&ring->napi, skb);
 }
 
 static inline void igb_rx_checksum_adv(struct igb_adapter *adapter,
@@ -4456,19 +4517,28 @@ static inline void igb_rx_checksum_adv(struct igb_adapter *adapter,
 	skb->ip_summed = CHECKSUM_NONE;
 
 	/* Ignore Checksum bit is set or checksum is disabled through ethtool */
-	if ((status_err & E1000_RXD_STAT_IXSM) || !adapter->rx_csum)
+	if ((status_err & E1000_RXD_STAT_IXSM) ||
+	    (adapter->flags & IGB_FLAG_RX_CSUM_DISABLED))
 		return;
 	/* TCP/UDP checksum error bit is set */
 	if (status_err &
 	    (E1000_RXDEXT_STATERR_TCPE | E1000_RXDEXT_STATERR_IPE)) {
+		/*
+		 * work around errata with sctp packets where the TCPE aka
+		 * L4E bit is set incorrectly on 64 byte (60 byte w/o crc)
+		 * packets, (aka let the stack check the crc32c)
+		 */
+		if (!((adapter->hw.mac.type == e1000_82576) &&
+		      (skb->len == 60)))
+			adapter->hw_csum_err++;
 		/* let the stack verify checksum errors */
-		adapter->hw_csum_err++;
 		return;
 	}
 	/* It must be a TCP or UDP packet with a valid checksum */
 	if (status_err & (E1000_RXD_STAT_TCPCS | E1000_RXD_STAT_UDPCS))
 		skb->ip_summed = CHECKSUM_UNNECESSARY;
 
+	dev_dbg(&adapter->pdev->dev, "cksum success: bits %08X\n", status_err);
 	adapter->hw_csum_good++;
 }
 
@@ -4513,11 +4583,12 @@ static bool igb_clean_rx_irq_adv(struct igb_ring *rx_ring,
 		cleaned = true;
 		cleaned_count++;
 
+		/* this is the fast path for the non-packet split case */
 		if (!adapter->rx_ps_hdr_size) {
 			pci_unmap_single(pdev, buffer_info->dma,
-					 adapter->rx_buffer_len +
-					   NET_IP_ALIGN,
+					 adapter->rx_buffer_len,
 					 PCI_DMA_FROMDEVICE);
+			buffer_info->dma = 0;
 			skb_put(skb, length);
 			goto send_up;
 		}
@@ -4534,8 +4605,9 @@ static bool igb_clean_rx_irq_adv(struct igb_ring *rx_ring,
 
 		if (!skb_shinfo(skb)->nr_frags) {
 			pci_unmap_single(pdev, buffer_info->dma,
-					 adapter->rx_ps_hdr_size + NET_IP_ALIGN,
+					 adapter->rx_ps_hdr_size,
 					 PCI_DMA_FROMDEVICE);
+			buffer_info->dma = 0;
 			skb_put(skb, hlen);
 		}
 
@@ -4677,7 +4749,6 @@ static void igb_alloc_rx_buffers_adv(struct igb_ring *rx_ring,
 		bufsz = adapter->rx_ps_hdr_size;
 	else
 		bufsz = adapter->rx_buffer_len;
-	bufsz += NET_IP_ALIGN;
 
 	while (cleaned_count--) {
 		rx_desc = E1000_RX_DESC_ADV(*rx_ring, i);
@@ -4701,7 +4772,7 @@ static void igb_alloc_rx_buffers_adv(struct igb_ring *rx_ring,
 		}
 
 		if (!buffer_info->skb) {
-			skb = netdev_alloc_skb(netdev, bufsz);
+			skb = netdev_alloc_skb(netdev, bufsz + NET_IP_ALIGN);
 			if (!skb) {
 				adapter->alloc_rx_buff_failed++;
 				goto no_buffers;
@@ -5302,6 +5373,9 @@ static pci_ers_result_t igb_io_error_detected(struct pci_dev *pdev,
 
 	netif_device_detach(netdev);
 
+	if (state == pci_channel_io_perm_failure)
+		return PCI_ERS_RESULT_DISCONNECT;
+
 	if (netif_running(netdev))
 		igb_down(adapter);
 	pci_disable_device(pdev);
@@ -5378,43 +5452,6 @@ static void igb_io_resume(struct pci_dev *pdev)
 	igb_get_hw_control(adapter);
 }
 
-static inline void igb_set_vmolr(struct e1000_hw *hw, int vfn)
-{
-	u32 reg_data;
-
-	reg_data = rd32(E1000_VMOLR(vfn));
-	reg_data |= E1000_VMOLR_BAM |	 /* Accept broadcast */
-	            E1000_VMOLR_ROPE |   /* Accept packets matched in UTA */
-	            E1000_VMOLR_ROMPE |  /* Accept packets matched in MTA */
-	            E1000_VMOLR_AUPE |   /* Accept untagged packets */
-	            E1000_VMOLR_STRVLAN; /* Strip vlan tags */
-	wr32(E1000_VMOLR(vfn), reg_data);
-}
-
-static inline int igb_set_vf_rlpml(struct igb_adapter *adapter, int size,
-                                 int vfn)
-{
-	struct e1000_hw *hw = &adapter->hw;
-	u32 vmolr;
-
-	vmolr = rd32(E1000_VMOLR(vfn));
-	vmolr &= ~E1000_VMOLR_RLPML_MASK;
-	vmolr |= size | E1000_VMOLR_LPE;
-	wr32(E1000_VMOLR(vfn), vmolr);
-
-	return 0;
-}
-
-static inline void igb_set_rah_pool(struct e1000_hw *hw, int pool, int entry)
-{
-	u32 reg_data;
-
-	reg_data = rd32(E1000_RAH(entry));
-	reg_data &= ~E1000_RAH_POOL_MASK;
-	reg_data |= E1000_RAH_POOL_1 << pool;;
-	wr32(E1000_RAH(entry), reg_data);
-}
-
 static void igb_set_mc_list_pools(struct igb_adapter *adapter,
 				  int entry_count, u16 total_rar_filters)
 {