Merge davem@nuts.davemloft.net:/disk1/BK/net-2.6

into kernel.bkbits.net:/home/davem/net-2.6

Documentation/networking/netdevices.txt

@@ -43,8 +43,21 @@ dev->get_stats:
 
 dev->hard_start_xmit:
 	Synchronization: dev->xmit_lock spinlock.
+	When the driver sets NETIF_F_LLTX in dev->features this will be
+	called without holding xmit_lock. In this case the driver 
+	has to lock by itself when needed. It is recommended to use a try lock
+	for this and return -1 when the spin lock fails. 
+	The locking there should also properly protect against 
+	set_multicast_list
 	Context: BHs disabled
 	Notes: netif_queue_stopped() is guaranteed false
+	Return codes: 
+	o NETDEV_TX_OK everything ok. 
+	o NETDEV_TX_BUSY Cannot transmit packet, try later 
+	  Usually a bug, means queue start/stop flow control is broken in
+	  the driver. Note: the driver must NOT put the skb in its DMA ring.
+	o NETDEV_TX_LOCKED Locking failed, please retry quickly.
+	  Only valid when NETIF_F_LLTX is set.
 
 dev->tx_timeout:
 	Synchronization: dev->xmit_lock spinlock.

drivers/net/e1000/e1000_main.c

@@ -1778,7 +1778,7 @@ e1000_xmit_frame(struct sk_buff *skb, struct net_device *netdev)
 
 	if(unlikely(skb->len <= 0)) {
 		dev_kfree_skb_any(skb);
-		return 0;
+		return NETDEV_TX_OK;
 	}
 
 #ifdef NETIF_F_TSO
@@ -1817,7 +1817,7 @@ e1000_xmit_frame(struct sk_buff *skb, struct net_device *netdev)
  	if (!spin_trylock(&adapter->tx_lock)) { 
  		/* Collision - tell upper layer to requeue */ 
  		local_irq_restore(flags); 
- 		return -1; 
+ 		return NETDEV_TX_LOCKED; 
  	} 
 
 	/* need: count + 2 desc gap to keep tail from touching
@@ -1825,7 +1825,7 @@ e1000_xmit_frame(struct sk_buff *skb, struct net_device *netdev)
 	if(E1000_DESC_UNUSED(&adapter->tx_ring) < count + 2) {
 		netif_stop_queue(netdev);
 		spin_unlock_irqrestore(&adapter->tx_lock, flags);
-		return 1;
+		return NETDEV_TX_BUSY;
 	}
 
 	if(unlikely(adapter->hw.mac_type == e1000_82547)) {
@@ -1833,7 +1833,7 @@ e1000_xmit_frame(struct sk_buff *skb, struct net_device *netdev)
 			netif_stop_queue(netdev);
 			mod_timer(&adapter->tx_fifo_stall_timer, jiffies);
 			spin_unlock_irqrestore(&adapter->tx_lock, flags);
-			return 1;
+			return NETDEV_TX_BUSY;
 		}
 	}
 
@@ -1856,7 +1856,7 @@ e1000_xmit_frame(struct sk_buff *skb, struct net_device *netdev)
 	netdev->trans_start = jiffies;
 
 	spin_unlock_irqrestore(&adapter->tx_lock, flags);
-	return 0;
+	return NETDEV_TX_OK;
 }
 
 /**

drivers/net/sk98lin/skge.c

@@ -4956,7 +4956,6 @@ static int __devinit skge_probe_one(struct pci_dev *pdev,
 	dev->stop =		&SkGeClose;
 	dev->hard_start_xmit =	&SkGeXmit;
 	dev->get_stats =	&SkGeStats;
-	dev->last_stats =	&SkGeStats;
 	dev->set_multicast_list = &SkGeSetRxMode;
 	dev->set_mac_address =	&SkGeSetMacAddr;
 	dev->do_ioctl =		&SkGeIoctl;
@@ -5037,7 +5036,6 @@ static int __devinit skge_probe_one(struct pci_dev *pdev,
 		dev->stop               = &SkGeClose;
 		dev->hard_start_xmit    = &SkGeXmit;
 		dev->get_stats          = &SkGeStats;
-		dev->last_stats         = &SkGeStats;
 		dev->set_multicast_list = &SkGeSetRxMode;
 		dev->set_mac_address    = &SkGeSetMacAddr;
 		dev->do_ioctl           = &SkGeIoctl;

drivers/net/tg3.c

@@ -2956,6 +2956,7 @@ static void tg3_set_txd(struct tg3 *tp, int entry,
 			dma_addr_t mapping, int len, u32 flags,
 			u32 mss_and_is_end)
 {
+	struct tg3_tx_buffer_desc *txd = &tp->tx_ring[entry];
 	int is_end = (mss_and_is_end & 0x1);
 	u32 mss = (mss_and_is_end >> 1);
 	u32 vlan_tag = 0;
@@ -2967,35 +2968,11 @@ static void tg3_set_txd(struct tg3 *tp, int entry,
 		flags &= 0xffff;
 	}
 	vlan_tag |= (mss << TXD_MSS_SHIFT);
-	if (tp->tg3_flags & TG3_FLAG_HOST_TXDS) {
-		struct tg3_tx_buffer_desc *txd = &tp->tx_ring[entry];
 
-		txd->addr_hi = ((u64) mapping >> 32);
-		txd->addr_lo = ((u64) mapping & 0xffffffff);
-		txd->len_flags = (len << TXD_LEN_SHIFT) | flags;
-		txd->vlan_tag = vlan_tag << TXD_VLAN_TAG_SHIFT;
-	} else {
-		struct tx_ring_info *txr = &tp->tx_buffers[entry];
-		void __iomem *txd;
-
-		txd = (tp->regs +
-		       NIC_SRAM_WIN_BASE +
-		       NIC_SRAM_TX_BUFFER_DESC);
-		txd += (entry * TXD_SIZE);
-
-		/* Save some PIOs */
-		if (sizeof(dma_addr_t) != sizeof(u32))
-			writel(((u64) mapping >> 32),
-			       txd + TXD_ADDR + TG3_64BIT_REG_HIGH);
-
-		writel(((u64) mapping & 0xffffffff),
-		       txd + TXD_ADDR + TG3_64BIT_REG_LOW);
-		writel(len << TXD_LEN_SHIFT | flags, txd + TXD_LEN_FLAGS);
-		if (txr->prev_vlan_tag != vlan_tag) {
-			writel(vlan_tag << TXD_VLAN_TAG_SHIFT, txd + TXD_VLAN_TAG);
-			txr->prev_vlan_tag = vlan_tag;
-		}
-	}
+	txd->addr_hi = ((u64) mapping >> 32);
+	txd->addr_lo = ((u64) mapping & 0xffffffff);
+	txd->len_flags = (len << TXD_LEN_SHIFT) | flags;
+	txd->vlan_tag = vlan_tag << TXD_VLAN_TAG_SHIFT;
 }
 
 static inline int tg3_4g_overflow_test(dma_addr_t mapping, int len)
@@ -3039,7 +3016,7 @@ static int tg3_start_xmit(struct sk_buff *skb, struct net_device *dev)
 	local_irq_save(flags);
 	if (!spin_trylock(&tp->tx_lock)) { 
 		local_irq_restore(flags);
-		return -1; 
+		return NETDEV_TX_LOCKED; 
 	} 
 
 	/* This is a hard error, log it. */
@@ -3048,7 +3025,7 @@ static int tg3_start_xmit(struct sk_buff *skb, struct net_device *dev)
 		spin_unlock_irqrestore(&tp->tx_lock, flags);
 		printk(KERN_ERR PFX "%s: BUG! Tx Ring full when queue awake!\n",
 		       dev->name);
-		return 1;
+		return NETDEV_TX_BUSY;
 	}
 
 	entry = tp->tx_prod;
@@ -3188,19 +3165,7 @@ static int tg3_start_xmit(struct sk_buff *skb, struct net_device *dev)
 	}
 
 	/* Packets are ready, update Tx producer idx local and on card. */
-	if (tp->tg3_flags & TG3_FLAG_HOST_TXDS) {
-		tw32_tx_mbox((MAILBOX_SNDHOST_PROD_IDX_0 +
-			      TG3_64BIT_REG_LOW), entry);
-	} else {
-		/* First, make sure tg3 sees last descriptor fully
-		 * in SRAM.
-		 */
-		if (tp->tg3_flags & TG3_FLAG_MBOX_WRITE_REORDER)
-			tr32(MAILBOX_SNDNIC_PROD_IDX_0 + TG3_64BIT_REG_LOW);
-
-		tw32_tx_mbox((MAILBOX_SNDNIC_PROD_IDX_0 +
-			      TG3_64BIT_REG_LOW), entry);
-	}
+	tw32_tx_mbox((MAILBOX_SNDHOST_PROD_IDX_0 + TG3_64BIT_REG_LOW), entry);
 
 	tp->tx_prod = entry;
 	if (TX_BUFFS_AVAIL(tp) <= (MAX_SKB_FRAGS + 1))
@@ -3211,7 +3176,7 @@ static int tg3_start_xmit(struct sk_buff *skb, struct net_device *dev)
 
 	dev->trans_start = jiffies;
 
-	return 0;
+	return NETDEV_TX_OK;
 }
 
 static inline void tg3_set_mtu(struct net_device *dev, struct tg3 *tp,
@@ -3339,7 +3304,6 @@ static void tg3_free_rings(struct tg3 *tp)
  */
 static void tg3_init_rings(struct tg3 *tp)
 {
-	void __iomem *start, *end;
 	u32 i;
 
 	/* Free up all the SKBs. */
@@ -3349,21 +3313,7 @@ static void tg3_init_rings(struct tg3 *tp)
 	memset(tp->rx_std, 0, TG3_RX_RING_BYTES);
 	memset(tp->rx_jumbo, 0, TG3_RX_JUMBO_RING_BYTES);
 	memset(tp->rx_rcb, 0, TG3_RX_RCB_RING_BYTES(tp));
-
-	if (tp->tg3_flags & TG3_FLAG_HOST_TXDS) {
-		memset(tp->tx_ring, 0, TG3_TX_RING_BYTES);
-	} else {
-		start = (tp->regs +
-			 NIC_SRAM_WIN_BASE +
-			 NIC_SRAM_TX_BUFFER_DESC);
-		end = start + TG3_TX_RING_BYTES;
-		while (start < end) {
-			writel(0, start);
-			start += 4;
-		}
-		for (i = 0; i < TG3_TX_RING_SIZE; i++)
-			tp->tx_buffers[i].prev_vlan_tag = 0;
-	}
+	memset(tp->tx_ring, 0, TG3_TX_RING_BYTES);
 
 	/* Initialize invariants of the rings, we only set this
 	 * stuff once.  This works because the card does not
@@ -3494,15 +3444,10 @@ static int tg3_alloc_consistent(struct tg3 *tp)
 	if (!tp->rx_rcb)
 		goto err_out;
 
-	if (tp->tg3_flags & TG3_FLAG_HOST_TXDS) {
-		tp->tx_ring = pci_alloc_consistent(tp->pdev, TG3_TX_RING_BYTES,
-						   &tp->tx_desc_mapping);
-		if (!tp->tx_ring)
-			goto err_out;
-	} else {
-		tp->tx_ring = NULL;
-		tp->tx_desc_mapping = 0;
-	}
+	tp->tx_ring = pci_alloc_consistent(tp->pdev, TG3_TX_RING_BYTES,
+					   &tp->tx_desc_mapping);
+	if (!tp->tx_ring)
+		goto err_out;
 
 	tp->hw_status = pci_alloc_consistent(tp->pdev,
 					     TG3_HW_STATUS_SIZE,
@@ -4966,10 +4911,7 @@ static int tg3_reset_hw(struct tg3 *tp)
 			  GRC_MODE_4X_NIC_SEND_RINGS |
 			  GRC_MODE_NO_TX_PHDR_CSUM |
 			  GRC_MODE_NO_RX_PHDR_CSUM);
-	if (tp->tg3_flags & TG3_FLAG_HOST_TXDS)
-		tp->grc_mode |= GRC_MODE_HOST_SENDBDS;
-	else
-		tp->grc_mode |= GRC_MODE_4X_NIC_SEND_RINGS;
+	tp->grc_mode |= GRC_MODE_HOST_SENDBDS;
 	if (tp->tg3_flags & TG3_FLAG_NO_TX_PSEUDO_CSUM)
 		tp->grc_mode |= GRC_MODE_NO_TX_PHDR_CSUM;
 	if (tp->tg3_flags & TG3_FLAG_NO_RX_PSEUDO_CSUM)
@@ -5122,18 +5064,11 @@ static int tg3_reset_hw(struct tg3 *tp)
 	tw32_mailbox(MAILBOX_SNDHOST_PROD_IDX_0 + TG3_64BIT_REG_LOW, 0);
 	tw32_tx_mbox(MAILBOX_SNDNIC_PROD_IDX_0 + TG3_64BIT_REG_LOW, 0);
 
-	if (tp->tg3_flags & TG3_FLAG_HOST_TXDS) {
-		tg3_set_bdinfo(tp, NIC_SRAM_SEND_RCB,
-			       tp->tx_desc_mapping,
-			       (TG3_TX_RING_SIZE <<
-				BDINFO_FLAGS_MAXLEN_SHIFT),
-			       NIC_SRAM_TX_BUFFER_DESC);
-	} else {
-		tg3_set_bdinfo(tp, NIC_SRAM_SEND_RCB,
-			       0,
-			       BDINFO_FLAGS_DISABLED,
-			       NIC_SRAM_TX_BUFFER_DESC);
-	}
+	tg3_set_bdinfo(tp, NIC_SRAM_SEND_RCB,
+		       tp->tx_desc_mapping,
+		       (TG3_TX_RING_SIZE <<
+			BDINFO_FLAGS_MAXLEN_SHIFT),
+		       NIC_SRAM_TX_BUFFER_DESC);
 
 	/* There is only one receive return ring on 5705/5750, no need
 	 * to explicitly disable the others.
@@ -5679,8 +5614,8 @@ static int tg3_open(struct net_device *dev)
 	spin_unlock(&tp->tx_lock);
 	spin_unlock_irq(&tp->lock);
 
-	/* If you move this call, make sure TG3_FLAG_HOST_TXDS in
-	 * tp->tg3_flags is accurate at that new place.
+	/* The placement of this call is tied
+	 * to the setup and use of Host TX descriptors.
 	 */
 	err = tg3_alloc_consistent(tp);
 	if (err)
@@ -6295,7 +6230,9 @@ static void tg3_set_rx_mode(struct net_device *dev)
 	struct tg3 *tp = netdev_priv(dev);
 
 	spin_lock_irq(&tp->lock);
+	spin_lock(&tp->tx_lock);
 	__tg3_set_rx_mode(dev);
+	spin_unlock(&tp->tx_lock);
 	spin_unlock_irq(&tp->lock);
 }
 
@@ -7633,23 +7570,6 @@ static int __devinit tg3_get_invariants(struct tg3 *tp)
 	udelay(50);
 	tg3_nvram_init(tp);
 
-	/* Always use host TXDs, it performs better in particular
-	 * with multi-frag packets.  The tests below are kept here
-	 * as documentation should we change this decision again
-	 * in the future.
-	 */
-	tp->tg3_flags |= TG3_FLAG_HOST_TXDS;
-
-#if 0
-	/* Determine if TX descriptors will reside in
-	 * main memory or in the chip SRAM.
-	 */
-	if ((tp->tg3_flags & TG3_FLAG_PCIX_TARGET_HWBUG) != 0 ||
-	    GET_ASIC_REV(tp->pci_chip_rev_id) == ASIC_REV_5705 ||
-	    GET_ASIC_REV(tp->pci_chip_rev_id) == ASIC_REV_5750)
-		tp->tg3_flags |= TG3_FLAG_HOST_TXDS;
-#endif
-
 	grc_misc_cfg = tr32(GRC_MISC_CFG);
 	grc_misc_cfg &= GRC_MISC_CFG_BOARD_ID_MASK;
 
@@ -8452,11 +8372,10 @@ static int __devinit tg3_init_one(struct pci_dev *pdev,
 		printk("%2.2x%c", dev->dev_addr[i],
 		       i == 5 ? '\n' : ':');
 
-	printk(KERN_INFO "%s: HostTXDS[%d] RXcsums[%d] LinkChgREG[%d] "
+	printk(KERN_INFO "%s: RXcsums[%d] LinkChgREG[%d] "
 	       "MIirq[%d] ASF[%d] Split[%d] WireSpeed[%d] "
 	       "TSOcap[%d] \n",
 	       dev->name,
-	       (tp->tg3_flags & TG3_FLAG_HOST_TXDS) != 0,
 	       (tp->tg3_flags & TG3_FLAG_RX_CHECKSUMS) != 0,
 	       (tp->tg3_flags & TG3_FLAG_USE_LINKCHG_REG) != 0,
 	       (tp->tg3_flags & TG3_FLAG_USE_MI_INTERRUPT) != 0,

drivers/net/tg3.h

@@ -1549,7 +1549,7 @@
  * exist only in the cards on-chip SRAM.  All 16 send bds are under
  * the same mode, they may not be configured individually.
  *
- * The mode we use is controlled by TG3_FLAG_HOST_TXDS in tp->tg3_flags.
+ * This driver always uses host memory TX descriptors.
  *
  * To use host memory TX descriptors:
  *	1) Set GRC_MODE_HOST_SENDBDS in GRC_MODE register.
@@ -2005,7 +2005,6 @@ struct tg3 {
 
 	spinlock_t			tx_lock;
 
-	/* TX descs are only used if TG3_FLAG_HOST_TXDS is set. */
 	struct tg3_tx_buffer_desc	*tx_ring;
 	struct tx_ring_info		*tx_buffers;
 	dma_addr_t			tx_desc_mapping;
@@ -2041,7 +2040,6 @@ struct tg3 {
 
 	u32				rx_offset;
 	u32				tg3_flags;
-#define TG3_FLAG_HOST_TXDS		0x00000001
 #define TG3_FLAG_TXD_MBOX_HWBUG		0x00000002
 #define TG3_FLAG_RX_CHECKSUMS		0x00000004
 #define TG3_FLAG_USE_LINKCHG_REG	0x00000008

include/linux/netdevice.h

@@ -73,6 +73,11 @@ struct ethtool_ops;
 
 #define MAX_ADDR_LEN	32		/* Largest hardware address length */
 
+/* Driver transmit return codes */
+#define NETDEV_TX_OK 0		/* driver took care of packet */
+#define NETDEV_TX_BUSY 1	/* driver tx path was busy*/
+#define NETDEV_TX_LOCKED -1	/* driver tx lock was already taken */
+
 /*
  *	Compute the worst case header length according to the protocols
  *	used.
@@ -479,7 +484,6 @@ struct net_device
 
 	/* class/net/name entry */
 	struct class_device	class_dev;
-	struct net_device_stats* (*last_stats)(struct net_device *);
 	/* how much padding had been added by alloc_netdev() */
 	int padded;
 };

include/linux/netfilter_ipv4/ip_conntrack_protocol.h

@@ -3,10 +3,7 @@
 #define _IP_CONNTRACK_PROTOCOL_H
 #include <linux/netfilter_ipv4/ip_conntrack.h>
 
-/* length of buffer to which print_tuple/print_conntrack members are
- * writing */
-
-#define IP_CT_PRINT_BUFLEN 100
+struct seq_file;
 
 struct ip_conntrack_protocol
 {
@@ -31,13 +28,12 @@ struct ip_conntrack_protocol
 	int (*invert_tuple)(struct ip_conntrack_tuple *inverse,
 			    const struct ip_conntrack_tuple *orig);
 
-	/* Print out the per-protocol part of the tuple. */
-	unsigned int (*print_tuple)(char *buffer,
-				    const struct ip_conntrack_tuple *);
+	/* Print out the per-protocol part of the tuple. Return like seq_* */
+	int (*print_tuple)(struct seq_file *,
+			   const struct ip_conntrack_tuple *);
 
 	/* Print out the private part of the conntrack. */
-	unsigned int (*print_conntrack)(char *buffer,
-					const struct ip_conntrack *);
+	int (*print_conntrack)(struct seq_file *, const struct ip_conntrack *);
 
 	/* Returns verdict for packet, or -1 for invalid. */
 	int (*packet)(struct ip_conntrack *conntrack,

include/linux/skbuff.h

@@ -1159,6 +1159,12 @@ static inline void nf_reset(struct sk_buff *skb)
 	skb->nf_debug = 0;
 #endif
 }
+static inline void nf_reset_debug(struct sk_buff *skb)
+{
+#ifdef CONFIG_NETFILTER_DEBUG
+	skb->nf_debug = 0;
+#endif
+}
 
 #ifdef CONFIG_BRIDGE_NETFILTER
 static inline void nf_bridge_put(struct nf_bridge_info *nf_bridge)

net/core/dev.c

@@ -861,18 +861,6 @@ static int default_rebuild_header(struct sk_buff *skb)
 }
 
 
-/*
- * Some old buggy device drivers change get_stats after registering
- * the device.  Try and trap them here.
- * This can be elimnated when all devices are known fixed.
- */
-static inline int get_stats_changed(struct net_device *dev)
-{
-	int changed = dev->last_stats != dev->get_stats;
-	dev->last_stats = dev->get_stats;
-	return changed;
-}
-
 /**
  *	dev_open	- prepare an interface for use.
  *	@dev:	device to open
@@ -896,14 +884,6 @@ int dev_open(struct net_device *dev)
 	if (dev->flags & IFF_UP)
 		return 0;
 
-	/*
-	 *	 Check for broken device drivers.
-	 */
-	if (get_stats_changed(dev) && net_ratelimit()) {
-		printk(KERN_ERR "%s: driver changed get_stats after register\n",
-		       dev->name);
-	}
-
 	/*
 	 *	Is it even present?
 	 */
@@ -920,14 +900,6 @@ int dev_open(struct net_device *dev)
 			clear_bit(__LINK_STATE_START, &dev->state);
 	}
 
-	/*
-	 * 	Check for more broken device drivers.
-	 */
-	if (get_stats_changed(dev) && net_ratelimit()) {
-		printk(KERN_ERR "%s: driver changed get_stats in open\n",
-		       dev->name);
-	}
-
  	/*
 	 *	If it went open OK then:
 	 */

net/core/net-sysfs.c

@@ -408,7 +408,6 @@ int netdev_register_sysfs(struct net_device *net)
 
 	class_dev->class = &net_class;
 	class_dev->class_data = net;
-	net->last_stats = net->get_stats;
 
 	strlcpy(class_dev->class_id, net->name, BUS_ID_SIZE);
 	if ((ret = class_device_register(class_dev)))

net/core/rtnetlink.c

@@ -265,7 +265,7 @@ static int do_setlink(struct sk_buff *skb, struct nlmsghdr *nlh, void *arg)
 	struct ifinfomsg  *ifm = NLMSG_DATA(nlh);
 	struct rtattr    **ida = arg;
 	struct net_device *dev;
-	int err;
+	int err, send_addr_notify = 0;
 
 	dev = dev_get_by_index(ifm->ifi_index);
 	if (!dev)
@@ -312,6 +312,7 @@ static int do_setlink(struct sk_buff *skb, struct nlmsghdr *nlh, void *arg)
 		err = dev->set_mac_address(dev, RTA_DATA(ida[IFLA_ADDRESS - 1]));
 		if (err)
 			goto out;
+		send_addr_notify = 1;
 	}
 
 	if (ida[IFLA_BROADCAST - 1]) {
@@ -319,6 +320,7 @@ static int do_setlink(struct sk_buff *skb, struct nlmsghdr *nlh, void *arg)
 			goto out;
 		memcpy(dev->broadcast, RTA_DATA(ida[IFLA_BROADCAST - 1]),
 		       dev->addr_len);
+		send_addr_notify = 1;
 	}
 
 	if (ida[IFLA_MTU - 1]) {
@@ -365,7 +367,7 @@ static int do_setlink(struct sk_buff *skb, struct nlmsghdr *nlh, void *arg)
 	err = 0;
 
 out:
-	if (!err)
+	if (send_addr_notify)
 		call_netdevice_notifiers(NETDEV_CHANGEADDR, dev);
 
 	dev_put(dev);

net/ipv4/ipvs/ip_vs_core.c

@@ -743,13 +743,6 @@ ip_vs_out(unsigned int hooknum, struct sk_buff **pskb,
 	if (skb->nfcache & NFC_IPVS_PROPERTY)
 		return NF_ACCEPT;
 
-	if (skb->ip_summed == CHECKSUM_HW) {
-		if (skb_checksum_help(pskb, (out == NULL)))
-			return NF_DROP;
-		if (skb != *pskb)
-			skb = *pskb;
-	}
-
 	iph = skb->nh.iph;
 	if (unlikely(iph->protocol == IPPROTO_ICMP)) {
 		int related, verdict = ip_vs_out_icmp(pskb, &related);
@@ -993,13 +986,6 @@ ip_vs_in(unsigned int hooknum, struct sk_buff **pskb,
 		return NF_ACCEPT;
 	}
 
-	if (skb->ip_summed == CHECKSUM_HW) {
-		if (skb_checksum_help(pskb, (out == NULL)))
-			return NF_DROP;
-		if (skb != *pskb)
-			skb = *pskb;
-	}
-
 	iph = skb->nh.iph;
 	if (unlikely(iph->protocol == IPPROTO_ICMP)) {
 		int related, verdict = ip_vs_in_icmp(pskb, &related);

net/ipv4/ipvs/ip_vs_xmit.c

@@ -124,11 +124,11 @@ ip_vs_dst_reset(struct ip_vs_dest *dest)
 	dst_release(old_dst);
 }
 
-
 #define IP_VS_XMIT(skb, rt)				\
 do {							\
-	nf_reset(skb);					\
+	nf_reset_debug(skb);				\
 	(skb)->nfcache |= NFC_IPVS_PROPERTY;		\
+	(skb)->ip_summed = CHECKSUM_NONE;		\
 	NF_HOOK(PF_INET, NF_IP_LOCAL_OUT, (skb), NULL,	\
 		(rt)->u.dst.dev, dst_output);		\
 } while (0)
@@ -408,8 +408,6 @@ ip_vs_tunnel_xmit(struct sk_buff *skb, struct ip_vs_conn *cp,
 	ip_select_ident(iph, &rt->u.dst, NULL);
 	ip_send_check(iph);
 
-	skb->ip_summed = CHECKSUM_NONE;
-
 	/* Another hack: avoid icmp_send in ip_fragment */
 	skb->local_df = 1;
 

net/ipv4/netfilter/ip_conntrack_proto_generic.c

@@ -34,15 +34,15 @@ static int generic_invert_tuple(struct ip_conntrack_tuple *tuple,
 }
 
 /* Print out the per-protocol part of the tuple. */
-static unsigned int generic_print_tuple(char *buffer,
-					const struct ip_conntrack_tuple *tuple)
+static int generic_print_tuple(struct seq_file *s,
+			       const struct ip_conntrack_tuple *tuple)
 {
 	return 0;
 }
 
 /* Print out the private part of the conntrack. */
-static unsigned int generic_print_conntrack(char *buffer,
-					    const struct ip_conntrack *state)
+static int generic_print_conntrack(struct seq_file *s,
+				   const struct ip_conntrack *state)
 {
 	return 0;
 }

net/ipv4/netfilter/ip_conntrack_proto_icmp.c

@@ -12,6 +12,7 @@
 #include <linux/netfilter.h>
 #include <linux/in.h>
 #include <linux/icmp.h>
+#include <linux/seq_file.h>
 #include <net/ip.h>
 #include <net/checksum.h>
 #include <linux/netfilter.h>
@@ -70,18 +71,18 @@ static int icmp_invert_tuple(struct ip_conntrack_tuple *tuple,
 }
 
 /* Print out the per-protocol part of the tuple. */
-static unsigned int icmp_print_tuple(char *buffer,
-				     const struct ip_conntrack_tuple *tuple)
+static int icmp_print_tuple(struct seq_file *s,
+			    const struct ip_conntrack_tuple *tuple)
 {
-	return sprintf(buffer, "type=%u code=%u id=%u ",
-		       tuple->dst.u.icmp.type,
-		       tuple->dst.u.icmp.code,
-		       ntohs(tuple->src.u.icmp.id));
+	return seq_printf(s, "type=%u code=%u id=%u ",
+			  tuple->dst.u.icmp.type,
+			  tuple->dst.u.icmp.code,
+			  ntohs(tuple->src.u.icmp.id));
 }
 
 /* Print out the private part of the conntrack. */
-static unsigned int icmp_print_conntrack(char *buffer,
-				     const struct ip_conntrack *conntrack)
+static int icmp_print_conntrack(struct seq_file *s,
+				const struct ip_conntrack *conntrack)
 {
 	return 0;
 }

net/ipv4/netfilter/ip_conntrack_proto_sctp.c

@@ -22,6 +22,7 @@
 #include <linux/ip.h>
 #include <linux/sctp.h>
 #include <linux/string.h>
+#include <linux/seq_file.h>
 
 #include <linux/netfilter_ipv4/ip_conntrack.h>
 #include <linux/netfilter_ipv4/ip_conntrack_protocol.h>
@@ -178,20 +179,20 @@ static int sctp_invert_tuple(struct ip_conntrack_tuple *tuple,
 }
 
 /* Print out the per-protocol part of the tuple. */
-static unsigned int sctp_print_tuple(char *buffer,
-				     const struct ip_conntrack_tuple *tuple)
+static int sctp_print_tuple(struct seq_file *s,
+			    const struct ip_conntrack_tuple *tuple)
 {
 	DEBUGP(__FUNCTION__);
 	DEBUGP("\n");
 
-	return sprintf(buffer, "sport=%hu dport=%hu ",
-		       ntohs(tuple->src.u.sctp.port),
-		       ntohs(tuple->dst.u.sctp.port));
+	return seq_printf(s, "sport=%hu dport=%hu ",
+			  ntohs(tuple->src.u.sctp.port),
+			  ntohs(tuple->dst.u.sctp.port));
 }
 
 /* Print out the private part of the conntrack. */
-static unsigned int sctp_print_conntrack(char *buffer,
-					 const struct ip_conntrack *conntrack)
+static int sctp_print_conntrack(struct seq_file *s,
+				const struct ip_conntrack *conntrack)
 {
 	enum sctp_conntrack state;
 
@@ -202,7 +203,7 @@ static unsigned int sctp_print_conntrack(char *buffer,
 	state = conntrack->proto.sctp.state;
 	READ_UNLOCK(&sctp_lock);
 
-	return sprintf(buffer, "%s ", sctp_conntrack_names[state]);
+	return seq_printf(s, "%s ", sctp_conntrack_names[state]);
 }
 
 #define for_each_sctp_chunk(skb, sch, offset, count)	\

net/ipv4/netfilter/ip_conntrack_proto_tcp.c

@@ -316,17 +316,17 @@ static int tcp_invert_tuple(struct ip_conntrack_tuple *tuple,
 }
 
 /* Print out the per-protocol part of the tuple. */
-static unsigned int tcp_print_tuple(char *buffer,
-				    const struct ip_conntrack_tuple *tuple)
+static int tcp_print_tuple(struct seq_file *s,
+			   const struct ip_conntrack_tuple *tuple)
 {
-	return sprintf(buffer, "sport=%hu dport=%hu ",
-		       ntohs(tuple->src.u.tcp.port),
-		       ntohs(tuple->dst.u.tcp.port));
+	return seq_printf(s, "sport=%hu dport=%hu ",
+			  ntohs(tuple->src.u.tcp.port),
+			  ntohs(tuple->dst.u.tcp.port));
 }
 
 /* Print out the private part of the conntrack. */
-static unsigned int tcp_print_conntrack(char *buffer,
-					const struct ip_conntrack *conntrack)
+static int tcp_print_conntrack(struct seq_file *s,
+			       const struct ip_conntrack *conntrack)
 {
 	enum tcp_conntrack state;
 
@@ -334,7 +334,7 @@ static unsigned int tcp_print_conntrack(char *buffer,
 	state = conntrack->proto.tcp.state;
 	READ_UNLOCK(&tcp_lock);
 
-	return sprintf(buffer, "%s ", tcp_conntrack_names[state]);
+	return seq_printf(s, "%s ", tcp_conntrack_names[state]);
 }
 
 static unsigned int get_conntrack_index(const struct tcphdr *tcph)

net/ipv4/netfilter/ip_conntrack_proto_udp.c

@@ -12,6 +12,7 @@
 #include <linux/netfilter.h>
 #include <linux/in.h>
 #include <linux/udp.h>
+#include <linux/seq_file.h>
 #include <net/checksum.h>
 #include <linux/netfilter.h>
 #include <linux/netfilter_ipv4.h>
@@ -46,17 +47,17 @@ static int udp_invert_tuple(struct ip_conntrack_tuple *tuple,
 }
 
 /* Print out the per-protocol part of the tuple. */
-static unsigned int udp_print_tuple(char *buffer,
-				    const struct ip_conntrack_tuple *tuple)
+static int udp_print_tuple(struct seq_file *s,
+			   const struct ip_conntrack_tuple *tuple)
 {
-	return sprintf(buffer, "sport=%hu dport=%hu ",
-		       ntohs(tuple->src.u.udp.port),
-		       ntohs(tuple->dst.u.udp.port));
+	return seq_printf(s, "sport=%hu dport=%hu ",
+			  ntohs(tuple->src.u.udp.port),
+			  ntohs(tuple->dst.u.udp.port));
 }
 
 /* Print out the private part of the conntrack. */
-static unsigned int udp_print_conntrack(char *buffer,
-					const struct ip_conntrack *conntrack)
+static int udp_print_conntrack(struct seq_file *s,
+			       const struct ip_conntrack *conntrack)
 {
 	return 0;
 }

net/ipv4/netfilter/ip_conntrack_standalone.c

@@ -57,18 +57,13 @@ static int kill_proto(const struct ip_conntrack *i, void *data)
 }
 
 #ifdef CONFIG_PROC_FS
-static unsigned int
-print_tuple(char *buffer, const struct ip_conntrack_tuple *tuple,
+static int
+print_tuple(struct seq_file *s, const struct ip_conntrack_tuple *tuple,
 	    struct ip_conntrack_protocol *proto)
 {
-	int len;
-
-	len = sprintf(buffer, "src=%u.%u.%u.%u dst=%u.%u.%u.%u ",
-		      NIPQUAD(tuple->src.ip), NIPQUAD(tuple->dst.ip));
-
-	len += proto->print_tuple(buffer + len, tuple);
-
-	return len;
+	seq_printf(s, "src=%u.%u.%u.%u dst=%u.%u.%u.%u ",
+		   NIPQUAD(tuple->src.ip), NIPQUAD(tuple->dst.ip));
+	return proto->print_tuple(s, tuple);
 }
 
 #ifdef CONFIG_IP_NF_CT_ACCT
@@ -85,48 +80,29 @@ seq_print_counters(struct seq_file *s, struct ip_conntrack_counter *counter)
 
 static void *ct_seq_start(struct seq_file *s, loff_t *pos)
 {
-	unsigned int *bucket;
-
-	/* strange seq_file api calls stop even if we fail,
-	 * thus we need to grab lock since stop unlocks */
-	READ_LOCK(&ip_conntrack_lock);
-  
 	if (*pos >= ip_conntrack_htable_size)
 		return NULL;
-
-	bucket = kmalloc(sizeof(unsigned int), GFP_KERNEL);
-	if (!bucket) {
-		return ERR_PTR(-ENOMEM);
-	}
-  
-	*bucket = *pos;
-	return bucket;
+	return &ip_conntrack_hash[*pos];
 }
   
-static void *ct_seq_next(struct seq_file *s, void *v, loff_t *pos)
+static void ct_seq_stop(struct seq_file *s, void *v)
 {
-	unsigned int *bucket = (unsigned int *) v;
+}
 
-	*pos = ++(*bucket);
-	if (*pos >= ip_conntrack_htable_size) {
-		kfree(v);
+static void *ct_seq_next(struct seq_file *s, void *v, loff_t *pos)
+{
+	(*pos)++;
+	if (*pos >= ip_conntrack_htable_size)
 		return NULL;
-	}
-	return bucket;
+	return &ip_conntrack_hash[*pos];
 }
   
-static void ct_seq_stop(struct seq_file *s, void *v)
-{
-	READ_UNLOCK(&ip_conntrack_lock);
-}
-
 /* return 0 on success, 1 in case of error */
 static int ct_seq_real_show(const struct ip_conntrack_tuple_hash *hash,
 			    struct seq_file *s)
 {
 	struct ip_conntrack *conntrack = hash->ctrack;
 	struct ip_conntrack_protocol *proto;
-	char buffer[IP_CT_PRINT_BUFLEN];
 
 	MUST_BE_READ_LOCKED(&ip_conntrack_lock);
 
@@ -147,12 +123,12 @@ static int ct_seq_real_show(const struct ip_conntrack_tuple_hash *hash,
 		      ? (conntrack->timeout.expires - jiffies)/HZ : 0) != 0)
 		return 1;
 
-	proto->print_conntrack(buffer, conntrack);
-	if (seq_puts(s, buffer))
+	if (proto->print_conntrack(s, conntrack))
 		return 1;
   
-	print_tuple(buffer, &conntrack->tuplehash[IP_CT_DIR_ORIGINAL].tuple,
-		    proto);
+	if (print_tuple(s, &conntrack->tuplehash[IP_CT_DIR_ORIGINAL].tuple,
+			proto))
+		return 1;
 
  	if (seq_print_counters(s, &conntrack->counters[IP_CT_DIR_ORIGINAL]))
 		return 1;
@@ -161,9 +137,8 @@ static int ct_seq_real_show(const struct ip_conntrack_tuple_hash *hash,
 		if (seq_printf(s, "[UNREPLIED] "))
 			return 1;
 
-	print_tuple(buffer, &conntrack->tuplehash[IP_CT_DIR_REPLY].tuple,
-		    proto);
-	if (seq_puts(s, buffer))
+	if (print_tuple(s, &conntrack->tuplehash[IP_CT_DIR_REPLY].tuple,
+			proto))
 		return 1;
 
  	if (seq_print_counters(s, &conntrack->counters[IP_CT_DIR_REPLY]))
@@ -179,17 +154,18 @@ static int ct_seq_real_show(const struct ip_conntrack_tuple_hash *hash,
 	return 0;
 }
 
-
 static int ct_seq_show(struct seq_file *s, void *v)
 {
-	unsigned int *bucket = (unsigned int *) v;
+	struct list_head *list = v;
+	int ret = 0;
 
-	if (LIST_FIND(&ip_conntrack_hash[*bucket], ct_seq_real_show,
-		      struct ip_conntrack_tuple_hash *, s)) {
-		/* buffer was filled and unable to print that tuple */
-		return 1;
-	}
-	return 0;
+	/* FIXME: Simply truncates if hash chain too long. */
+	READ_LOCK(&ip_conntrack_lock);
+	if (LIST_FIND(list, ct_seq_real_show,
+		      struct ip_conntrack_tuple_hash *, s))
+		ret = -ENOSPC;
+	READ_UNLOCK(&ip_conntrack_lock);
+	return ret;
 }
 	
 static struct seq_operations ct_seq_ops = {
@@ -255,7 +231,6 @@ static void exp_seq_stop(struct seq_file *s, void *v)
 static int exp_seq_show(struct seq_file *s, void *v)
 {
 	struct ip_conntrack_expect *expect = v;
-	char buffer[IP_CT_PRINT_BUFLEN];
 
 	if (expect->expectant->helper->timeout)
 		seq_printf(s, "%lu ", timer_pending(&expect->timeout)
@@ -266,9 +241,8 @@ static int exp_seq_show(struct seq_file *s, void *v)
 	seq_printf(s, "use=%u proto=%u ", atomic_read(&expect->use),
 		   expect->tuple.dst.protonum);
 
-	print_tuple(buffer, &expect->tuple,
-		    __ip_ct_find_proto(expect->tuple.dst.protonum));
-	return seq_printf(s, "%s\n", buffer);
+	return print_tuple(s, &expect->tuple,
+			   __ip_ct_find_proto(expect->tuple.dst.protonum));
 }
 
 static struct seq_operations exp_seq_ops = {

net/ipv4/tcp_input.c

@@ -1032,7 +1032,7 @@ tcp_sacktag_write_queue(struct sock *sk, struct sk_buff *ack_skb, u32 prior_snd_
 			if(!before(TCP_SKB_CB(skb)->seq, end_seq))
 				break;
 
-			fack_count++;
+			fack_count += TCP_SKB_CB(skb)->tso_factor;
 
 			in_sack = !after(start_seq, TCP_SKB_CB(skb)->seq) &&
 				!before(end_seq, TCP_SKB_CB(skb)->end_seq);

net/ipv4/tcp_output.c

@@ -495,10 +495,20 @@ static int tcp_fragment(struct sock *sk, struct sk_buff *skb, u32 len)
 	 */
 	TCP_SKB_CB(buff)->when = TCP_SKB_CB(skb)->when;
 
+	if (TCP_SKB_CB(skb)->sacked & TCPCB_LOST) {
+		tcp_dec_pcount(&tp->lost_out, skb);
+		tcp_dec_pcount(&tp->left_out, skb);
+	}
+
 	/* Fix up tso_factor for both original and new SKB.  */
 	tcp_set_skb_tso_factor(skb, tp->mss_cache_std);
 	tcp_set_skb_tso_factor(buff, tp->mss_cache_std);
 
+	if (TCP_SKB_CB(skb)->sacked & TCPCB_LOST) {
+		tcp_inc_pcount(&tp->lost_out, skb);
+		tcp_inc_pcount(&tp->left_out, skb);
+	}
+
 	if (TCP_SKB_CB(buff)->sacked&TCPCB_LOST) {
 		tcp_inc_pcount(&tp->lost_out, buff);
 		tcp_inc_pcount(&tp->left_out, buff);

net/ipv6/route.c

@@ -184,7 +184,8 @@ static __inline__ struct rt6_info *rt6_device_match(struct rt6_info *rt,
 			if (dev->ifindex == oif)
 				return sprt;
 			if (dev->flags & IFF_LOOPBACK) {
-				if (sprt->rt6i_idev->dev->ifindex != oif) {
+				if (sprt->rt6i_idev == NULL ||
+				    sprt->rt6i_idev->dev->ifindex != oif) {
 					if (strict && oif)
 						continue;
 					if (local && (!oif || 

net/sched/sch_generic.c

@@ -139,13 +139,8 @@ int qdisc_restart(struct net_device *dev)
 				if (netdev_nit)
 					dev_queue_xmit_nit(skb, dev);
 
-				/* hard_start_xmit returns: 
-				   0  device not ready
-				   1  everything ok
-				   -1 didn't get device lock (for LLTX)
-				*/ 
 				ret = dev->hard_start_xmit(skb, dev);
-				if (ret == 0) { 
+				if (ret == NETDEV_TX_OK) { 
 					if (!nolock) {
 						dev->xmit_lock_owner = -1;
 						spin_unlock(&dev->xmit_lock);
@@ -153,10 +148,11 @@ int qdisc_restart(struct net_device *dev)
 					spin_lock(&dev->queue_lock);
 					return -1;
 				}
-				if (ret == -1 && nolock)
+				if (ret == NETDEV_TX_LOCKED && nolock)
 					goto collision; 
 			}
 
+			/* NETDEV_TX_BUSY - we need to requeue */
 			/* Release the driver */
 			if (!nolock) { 
 				dev->xmit_lock_owner = -1;
@@ -176,7 +172,7 @@ int qdisc_restart(struct net_device *dev)
 		   3. device is buggy (ppp)
 		 */
 
-	requeue:
+requeue:
 		q->ops->requeue(skb, q);
 		netif_schedule(dev);
 		return 1;