1 files changed, 349 insertions, 41 deletions

diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c
index 269a5e407349..2c930da90a85 100644
--- a/drivers/net/bonding/bond_main.c
+++ b/drivers/net/bonding/bond_main.c
@@ -475,7 +475,18 @@
  *        Solution is to move call to dev_remove_pack outside of the
  *        spinlock.
  *        Set version to 2.6.1.
- *
+ * 2005/06/05 - Jay Vosburgh <fubar@us.ibm.com>
+ * 	- Support for generating gratuitous ARPs in active-backup mode.
+ * 	  Includes support for VLAN tagging all bonding-generated ARPs
+ * 	  as needed.  Set version to 2.6.2.
+ * 2005/06/08 - Jason Gabler <jygabler at lbl dot gov>
+ *	- alternate hashing policy support for mode 2
+ *	  * Added kernel parameter "xmit_hash_policy" to allow the selection
+ *	    of different hashing policies for mode 2.  The original mode 2
+ *	    policy is the default, now found in xmit_hash_policy_layer2().
+ *	  * Added xmit_hash_policy_layer34()
+ *	- Modified by Jay Vosburgh <fubar@us.ibm.com> to also support mode 4.
+ *	  Set version to 2.6.3.
  */
 
 //#define BONDING_DEBUG 1
@@ -490,7 +501,10 @@
 #include <linux/ptrace.h>
 #include <linux/ioport.h>
 #include <linux/in.h>
+#include <net/ip.h>
 #include <linux/ip.h>
+#include <linux/tcp.h>
+#include <linux/udp.h>
 #include <linux/slab.h>
 #include <linux/string.h>
 #include <linux/init.h>
@@ -519,6 +533,7 @@
 #include <linux/ethtool.h>
 #include <linux/if_vlan.h>
 #include <linux/if_bonding.h>
+#include <net/route.h>
 #include "bonding.h"
 #include "bond_3ad.h"
 #include "bond_alb.h"
@@ -537,6 +552,7 @@ static int use_carrier	= 1;
 static char *mode	= NULL;
 static char *primary	= NULL;
 static char *lacp_rate	= NULL;
+static char *xmit_hash_policy = NULL;
 static int arp_interval = BOND_LINK_ARP_INTERV;
 static char *arp_ip_target[BOND_MAX_ARP_TARGETS] = { NULL, };
 
@@ -556,6 +572,8 @@ module_param(primary, charp, 0);
 MODULE_PARM_DESC(primary, "Primary network device to use");
 module_param(lacp_rate, charp, 0);
 MODULE_PARM_DESC(lacp_rate, "LACPDU tx rate to request from 802.3ad partner (slow/fast)");
+module_param(xmit_hash_policy, charp, 0);
+MODULE_PARM_DESC(xmit_hash_policy, "XOR hashing method : 0 for layer 2 (default), 1 for layer 3+4");
 module_param(arp_interval, int, 0);
 MODULE_PARM_DESC(arp_interval, "arp interval in milliseconds");
 module_param_array(arp_ip_target, charp, NULL, 0);
@@ -574,8 +592,8 @@ static struct proc_dir_entry *bond_proc_dir = NULL;
 
 static u32 arp_target[BOND_MAX_ARP_TARGETS] = { 0, } ;
 static int arp_ip_count	= 0;
-static u32 my_ip	= 0;
 static int bond_mode	= BOND_MODE_ROUNDROBIN;
+static int xmit_hashtype= BOND_XMIT_POLICY_LAYER2;
 static int lacp_fast	= 0;
 static int app_abi_ver	= 0;
 static int orig_app_abi_ver = -1; /* This is used to save the first ABI version
@@ -585,7 +603,6 @@ static int orig_app_abi_ver = -1; /* This is used to save the first ABI version
 				   * command comes from an application using
 				   * another ABI version.
 				   */
-
 struct bond_parm_tbl {
 	char *modename;
 	int mode;
@@ -608,9 +625,16 @@ static struct bond_parm_tbl bond_mode_tbl[] = {
 {	NULL,			-1},
 };
 
+static struct bond_parm_tbl xmit_hashtype_tbl[] = {
+{	"layer2",		BOND_XMIT_POLICY_LAYER2},
+{	"layer3+4",		BOND_XMIT_POLICY_LAYER34},
+{	NULL,			-1},
+};
+
 /*-------------------------- Forward declarations ---------------------------*/
 
-static inline void bond_set_mode_ops(struct net_device *bond_dev, int mode);
+static inline void bond_set_mode_ops(struct bonding *bond, int mode);
+static void bond_send_gratuitous_arp(struct bonding *bond);
 
 /*---------------------------- General routines -----------------------------*/
 
@@ -659,6 +683,7 @@ static int bond_add_vlan(struct bonding *bond, unsigned short vlan_id)
 
 	INIT_LIST_HEAD(&vlan->vlan_list);
 	vlan->vlan_id = vlan_id;
+	vlan->vlan_ip = 0;
 
 	write_lock_bh(&bond->lock);
 
@@ -1468,16 +1493,6 @@ static void bond_change_active_slave(struct bonding *bond, struct slave *new_act
 		}
 	}
 
-	if (bond->params.mode == BOND_MODE_ACTIVEBACKUP) {
-		if (old_active) {
-			bond_set_slave_inactive_flags(old_active);
-		}
-
-		if (new_active) {
-			bond_set_slave_active_flags(new_active);
-		}
-	}
-
 	if (USES_PRIMARY(bond->params.mode)) {
 		bond_mc_swap(bond, new_active, old_active);
 	}
@@ -1488,6 +1503,17 @@ static void bond_change_active_slave(struct bonding *bond, struct slave *new_act
 	} else {
 		bond->curr_active_slave = new_active;
 	}
+
+	if (bond->params.mode == BOND_MODE_ACTIVEBACKUP) {
+		if (old_active) {
+			bond_set_slave_inactive_flags(old_active);
+		}
+
+		if (new_active) {
+			bond_set_slave_active_flags(new_active);
+		}
+		bond_send_gratuitous_arp(bond);
+	}
 }
 
 /**
@@ -2694,15 +2720,180 @@ out:
 	read_unlock(&bond->lock);
 }
 
+
+static u32 bond_glean_dev_ip(struct net_device *dev)
+{
+	struct in_device *idev;
+	struct in_ifaddr *ifa;
+	u32 addr = 0;
+
+	if (!dev)
+		return 0;
+
+	rcu_read_lock();
+	idev = __in_dev_get(dev);
+	if (!idev)
+		goto out;
+
+	ifa = idev->ifa_list;
+	if (!ifa)
+		goto out;
+
+	addr = ifa->ifa_local;
+out:
+	rcu_read_unlock();
+	return addr;
+}
+
+static int bond_has_ip(struct bonding *bond)
+{
+	struct vlan_entry *vlan, *vlan_next;
+
+	if (bond->master_ip)
+		return 1;
+
+	if (list_empty(&bond->vlan_list))
+		return 0;
+
+	list_for_each_entry_safe(vlan, vlan_next, &bond->vlan_list,
+				 vlan_list) {
+		if (vlan->vlan_ip)
+			return 1;
+	}
+
+	return 0;
+}
+
+/*
+ * We go to the (large) trouble of VLAN tagging ARP frames because
+ * switches in VLAN mode (especially if ports are configured as
+ * "native" to a VLAN) might not pass non-tagged frames.
+ */
+static void bond_arp_send(struct net_device *slave_dev, int arp_op, u32 dest_ip, u32 src_ip, unsigned short vlan_id)
+{
+	struct sk_buff *skb;
+
+	dprintk("arp %d on slave %s: dst %x src %x vid %d\n", arp_op,
+	       slave_dev->name, dest_ip, src_ip, vlan_id);
+	       
+	skb = arp_create(arp_op, ETH_P_ARP, dest_ip, slave_dev, src_ip,
+			 NULL, slave_dev->dev_addr, NULL);
+
+	if (!skb) {
+		printk(KERN_ERR DRV_NAME ": ARP packet allocation failed\n");
+		return;
+	}
+	if (vlan_id) {
+		skb = vlan_put_tag(skb, vlan_id);
+		if (!skb) {
+			printk(KERN_ERR DRV_NAME ": failed to insert VLAN tag\n");
+			return;
+		}
+	}
+	arp_xmit(skb);
+}
+
+
 static void bond_arp_send_all(struct bonding *bond, struct slave *slave)
 {
-	int i;
+	int i, vlan_id, rv;
 	u32 *targets = bond->params.arp_targets;
+	struct vlan_entry *vlan, *vlan_next;
+	struct net_device *vlan_dev;
+	struct flowi fl;
+	struct rtable *rt;
 
 	for (i = 0; (i < BOND_MAX_ARP_TARGETS) && targets[i]; i++) {
-		arp_send(ARPOP_REQUEST, ETH_P_ARP, targets[i], slave->dev,
-			 my_ip, NULL, slave->dev->dev_addr,
-			 NULL);
+		dprintk("basa: target %x\n", targets[i]);
+		if (list_empty(&bond->vlan_list)) {
+			dprintk("basa: empty vlan: arp_send\n");
+			bond_arp_send(slave->dev, ARPOP_REQUEST, targets[i],
+				      bond->master_ip, 0);
+			continue;
+		}
+
+		/*
+		 * If VLANs are configured, we do a route lookup to
+		 * determine which VLAN interface would be used, so we
+		 * can tag the ARP with the proper VLAN tag.
+		 */
+		memset(&fl, 0, sizeof(fl));
+		fl.fl4_dst = targets[i];
+		fl.fl4_tos = RTO_ONLINK;
+
+		rv = ip_route_output_key(&rt, &fl);
+		if (rv) {
+			if (net_ratelimit()) {
+				printk(KERN_WARNING DRV_NAME
+			     ": %s: no route to arp_ip_target %u.%u.%u.%u\n",
+				       bond->dev->name, NIPQUAD(fl.fl4_dst));
+			}
+			continue;
+		}
+
+		/*
+		 * This target is not on a VLAN
+		 */
+		if (rt->u.dst.dev == bond->dev) {
+			dprintk("basa: rtdev == bond->dev: arp_send\n");
+			bond_arp_send(slave->dev, ARPOP_REQUEST, targets[i],
+				      bond->master_ip, 0);
+			continue;
+		}
+
+		vlan_id = 0;
+		list_for_each_entry_safe(vlan, vlan_next, &bond->vlan_list,
+					 vlan_list) {
+			vlan_dev = bond->vlgrp->vlan_devices[vlan->vlan_id];
+			if (vlan_dev == rt->u.dst.dev) {
+				vlan_id = vlan->vlan_id;
+				dprintk("basa: vlan match on %s %d\n",
+				       vlan_dev->name, vlan_id);
+				break;
+			}
+		}
+
+		if (vlan_id) {
+			bond_arp_send(slave->dev, ARPOP_REQUEST, targets[i],
+				      vlan->vlan_ip, vlan_id);
+			continue;
+		}
+
+		if (net_ratelimit()) {
+			printk(KERN_WARNING DRV_NAME
+	       ": %s: no path to arp_ip_target %u.%u.%u.%u via rt.dev %s\n",
+			       bond->dev->name, NIPQUAD(fl.fl4_dst),
+			       rt->u.dst.dev ? rt->u.dst.dev->name : "NULL");
+		}
+	}
+}
+
+/*
+ * Kick out a gratuitous ARP for an IP on the bonding master plus one
+ * for each VLAN above us.
+ */
+static void bond_send_gratuitous_arp(struct bonding *bond)
+{
+	struct slave *slave = bond->curr_active_slave;
+	struct vlan_entry *vlan;
+	struct net_device *vlan_dev;
+
+	dprintk("bond_send_grat_arp: bond %s slave %s\n", bond->dev->name,
+				slave ? slave->dev->name : "NULL");
+	if (!slave)
+		return;
+
+	if (bond->master_ip) {
+		bond_arp_send(slave->dev, ARPOP_REPLY, bond->master_ip,
+				  bond->master_ip, 0);
+	}
+
+	list_for_each_entry(vlan, &bond->vlan_list, vlan_list) {
+		vlan_dev = bond->vlgrp->vlan_devices[vlan->vlan_id];
+		if (vlan->vlan_ip) {
+			bond_arp_send(slave->dev, ARPOP_REPLY, vlan->vlan_ip,
+				      vlan->vlan_ip, vlan->vlan_id);
+		}
 	}
 }
 
@@ -2781,7 +2972,7 @@ static void bond_loadbalance_arp_mon(struct net_device *bond_dev)
 			 */
 			if (((jiffies - slave->dev->trans_start) >= (2*delta_in_ticks)) ||
 			    (((jiffies - slave->dev->last_rx) >= (2*delta_in_ticks)) &&
-			     my_ip)) {
+			     bond_has_ip(bond))) {
 
 				slave->link  = BOND_LINK_DOWN;
 				slave->state = BOND_STATE_BACKUP;
@@ -2920,7 +3111,7 @@ static void bond_activebackup_arp_mon(struct net_device *bond_dev)
 			if ((slave != bond->curr_active_slave) &&
 			    (!bond->current_arp_slave) &&
 			    (((jiffies - slave->dev->last_rx) >= 3*delta_in_ticks) &&
-			     my_ip)) {
+			     bond_has_ip(bond))) {
 				/* a backup slave has gone down; three times
 				 * the delta allows the current slave to be
 				 * taken out before the backup slave.
@@ -2966,8 +3157,8 @@ static void bond_activebackup_arp_mon(struct net_device *bond_dev)
 		 * if it is up and needs to take over as the curr_active_slave
 		 */
 		if ((((jiffies - slave->dev->trans_start) >= (2*delta_in_ticks)) ||
-		     (((jiffies - slave->dev->last_rx) >= (2*delta_in_ticks)) &&
-		      my_ip)) &&
+	    (((jiffies - slave->dev->last_rx) >= (2*delta_in_ticks)) &&
+	     bond_has_ip(bond))) &&
 		    ((jiffies - slave->jiffies) >= 2*delta_in_ticks)) {
 
 			slave->link  = BOND_LINK_DOWN;
@@ -3019,7 +3210,7 @@ static void bond_activebackup_arp_mon(struct net_device *bond_dev)
 		/* the current slave must tx an arp to ensure backup slaves
 		 * rx traffic
 		 */
-		if (slave && my_ip) {
+		if (slave && bond_has_ip(bond)) {
 			bond_arp_send_all(bond, slave);
 		}
 	}
@@ -3471,10 +3662,67 @@ static int bond_netdev_event(struct notifier_block *this, unsigned long event, v
 	return NOTIFY_DONE;
 }
 
+/*
+ * bond_inetaddr_event: handle inetaddr notifier chain events.
+ *
+ * We keep track of device IPs primarily to use as source addresses in
+ * ARP monitor probes (rather than spewing out broadcasts all the time).
+ *
+ * We track one IP for the main device (if it has one), plus one per VLAN.
+ */
+static int bond_inetaddr_event(struct notifier_block *this, unsigned long event, void *ptr)
+{
+	struct in_ifaddr *ifa = ptr;
+	struct net_device *vlan_dev, *event_dev = ifa->ifa_dev->dev;
+	struct bonding *bond, *bond_next;
+	struct vlan_entry *vlan, *vlan_next;
+
+	list_for_each_entry_safe(bond, bond_next, &bond_dev_list, bond_list) {
+		if (bond->dev == event_dev) {
+			switch (event) {
+			case NETDEV_UP:
+				bond->master_ip = ifa->ifa_local;
+				return NOTIFY_OK;
+			case NETDEV_DOWN:
+				bond->master_ip = bond_glean_dev_ip(bond->dev);
+				return NOTIFY_OK;
+			default:
+				return NOTIFY_DONE;
+			}
+		}
+
+		if (list_empty(&bond->vlan_list))
+			continue;
+
+		list_for_each_entry_safe(vlan, vlan_next, &bond->vlan_list,
+					 vlan_list) {
+			vlan_dev = bond->vlgrp->vlan_devices[vlan->vlan_id];
+			if (vlan_dev == event_dev) {
+				switch (event) {
+				case NETDEV_UP:
+					vlan->vlan_ip = ifa->ifa_local;
+					return NOTIFY_OK;
+				case NETDEV_DOWN:
+					vlan->vlan_ip =
+						bond_glean_dev_ip(vlan_dev);
+					return NOTIFY_OK;
+				default:
+					return NOTIFY_DONE;
+				}
+			}
+		}
+	}
+	return NOTIFY_DONE;
+}
+
 static struct notifier_block bond_netdev_notifier = {
 	.notifier_call = bond_netdev_event,
 };
 
+static struct notifier_block bond_inetaddr_notifier = {
+	.notifier_call = bond_inetaddr_event,
+};
+
 /*-------------------------- Packet type handling ---------------------------*/
 
 /* register to receive lacpdus on a bond */
@@ -3496,6 +3744,46 @@ static void bond_unregister_lacpdu(struct bonding *bond)
 	dev_remove_pack(&(BOND_AD_INFO(bond).ad_pkt_type));
 }
 
+/*---------------------------- Hashing Policies -----------------------------*/
+
+/*
+ * Hash for the the output device based upon layer 3 and layer 4 data. If
+ * the packet is a frag or not TCP or UDP, just use layer 3 data.  If it is
+ * altogether not IP, mimic bond_xmit_hash_policy_l2()
+ */
+static int bond_xmit_hash_policy_l34(struct sk_buff *skb,
+				    struct net_device *bond_dev, int count)
+{
+	struct ethhdr *data = (struct ethhdr *)skb->data;
+	struct iphdr *iph = skb->nh.iph;
+	u16 *layer4hdr = (u16 *)((u32 *)iph + iph->ihl);
+	int layer4_xor = 0;
+
+	if (skb->protocol == __constant_htons(ETH_P_IP)) {
+		if (!(iph->frag_off & __constant_htons(IP_MF|IP_OFFSET)) &&
+		    (iph->protocol == IPPROTO_TCP ||
+		     iph->protocol == IPPROTO_UDP)) {
+			layer4_xor = htons((*layer4hdr ^ *(layer4hdr + 1)));
+		}
+		return (layer4_xor ^
+			((ntohl(iph->saddr ^ iph->daddr)) & 0xffff)) % count;
+
+	}
+
+	return (data->h_dest[5] ^ bond_dev->dev_addr[5]) % count;
+}
+
+/*
+ * Hash for the output device based upon layer 2 data
+ */
+static int bond_xmit_hash_policy_l2(struct sk_buff *skb,
+				   struct net_device *bond_dev, int count)
+{
+	struct ethhdr *data = (struct ethhdr *)skb->data;
+
+	return (data->h_dest[5] ^ bond_dev->dev_addr[5]) % count;
+}
+
 /*-------------------------- Device entry points ----------------------------*/
 
 static int bond_open(struct net_device *bond_dev)
@@ -4060,17 +4348,6 @@ static int bond_xmit_activebackup(struct sk_buff *skb, struct net_device *bond_d
 	struct bonding *bond = bond_dev->priv;
 	int res = 1;
 
-	/* if we are sending arp packets, try to at least
-	   identify our own ip address */
-	if (bond->params.arp_interval && !my_ip &&
-		(skb->protocol == __constant_htons(ETH_P_ARP))) {
-		char *the_ip = (char *)skb->data +
-				sizeof(struct ethhdr) +
-				sizeof(struct arphdr) +
-				ETH_ALEN;
-		memcpy(&my_ip, the_ip, 4);
-	}
-
 	read_lock(&bond->lock);
 	read_lock(&bond->curr_slave_lock);
 
@@ -4093,14 +4370,13 @@ out:
 }
 
 /*
- * in XOR mode, we determine the output device by performing xor on
- * the source and destination hw adresses.  If this device is not
- * enabled, find the next slave following this xor slave.
+ * In bond_xmit_xor() , we determine the output device by using a pre-
+ * determined xmit_hash_policy(), If the selected device is not enabled,
+ * find the next active slave.
  */
 static int bond_xmit_xor(struct sk_buff *skb, struct net_device *bond_dev)
 {
 	struct bonding *bond = bond_dev->priv;
-	struct ethhdr *data = (struct ethhdr *)skb->data;
 	struct slave *slave, *start_at;
 	int slave_no;
 	int i;
@@ -4112,7 +4388,7 @@ static int bond_xmit_xor(struct sk_buff *skb, struct net_device *bond_dev)
 		goto out;
 	}
 
-	slave_no = (data->h_dest[5]^bond_dev->dev_addr[5]) % bond->slave_cnt;
+	slave_no = bond->xmit_hash_policy(skb, bond_dev, bond->slave_cnt);
 
 	bond_for_each_slave(bond, slave, i) {
 		slave_no--;
@@ -4208,8 +4484,10 @@ out:
 /*
  * set bond mode specific net device operations
  */
-static inline void bond_set_mode_ops(struct net_device *bond_dev, int mode)
+static inline void bond_set_mode_ops(struct bonding *bond, int mode)
 {
+	struct net_device *bond_dev = bond->dev;
+
 	switch (mode) {
 	case BOND_MODE_ROUNDROBIN:
 		bond_dev->hard_start_xmit = bond_xmit_roundrobin;
@@ -4219,12 +4497,20 @@ static inline void bond_set_mode_ops(struct net_device *bond_dev, int mode)
 		break;
 	case BOND_MODE_XOR:
 		bond_dev->hard_start_xmit = bond_xmit_xor;
+		if (bond->params.xmit_policy == BOND_XMIT_POLICY_LAYER34)
+			bond->xmit_hash_policy = bond_xmit_hash_policy_l34;
+		else
+			bond->xmit_hash_policy = bond_xmit_hash_policy_l2;
 		break;
 	case BOND_MODE_BROADCAST:
 		bond_dev->hard_start_xmit = bond_xmit_broadcast;
 		break;
 	case BOND_MODE_8023AD:
 		bond_dev->hard_start_xmit = bond_3ad_xmit_xor;
+		if (bond->params.xmit_policy == BOND_XMIT_POLICY_LAYER34)
+			bond->xmit_hash_policy = bond_xmit_hash_policy_l34;
+		else
+			bond->xmit_hash_policy = bond_xmit_hash_policy_l2;
 		break;
 	case BOND_MODE_TLB:
 	case BOND_MODE_ALB:
@@ -4273,7 +4559,7 @@ static int __init bond_init(struct net_device *bond_dev, struct bond_params *par
 	bond_dev->change_mtu = bond_change_mtu;
 	bond_dev->set_mac_address = bond_set_mac_address;
 
-	bond_set_mode_ops(bond_dev, bond->params.mode);
+	bond_set_mode_ops(bond, bond->params.mode);
 
 	bond_dev->destructor = free_netdev;
 
@@ -4384,6 +4670,25 @@ static int bond_check_params(struct bond_params *params)
 		}
 	}
 
+	if (xmit_hash_policy) {
+		if ((bond_mode != BOND_MODE_XOR) &&
+		    (bond_mode != BOND_MODE_8023AD)) {
+			printk(KERN_INFO DRV_NAME
+			       ": xor_mode param is irrelevant in mode %s\n",
+			       bond_mode_name(bond_mode));
+		} else {
+			xmit_hashtype = bond_parse_parm(xmit_hash_policy,
+							xmit_hashtype_tbl);
+			if (xmit_hashtype == -1) {
+				printk(KERN_ERR DRV_NAME
+			       	": Error: Invalid xmit_hash_policy \"%s\"\n",
+			       	xmit_hash_policy == NULL ? "NULL" :
+				       xmit_hash_policy);
+				return -EINVAL;
+			}
+		}
+	}
+
 	if (lacp_rate) {
 		if (bond_mode != BOND_MODE_8023AD) {
 			printk(KERN_INFO DRV_NAME
@@ -4595,6 +4900,7 @@ static int bond_check_params(struct bond_params *params)
 
 	/* fill params struct with the proper values */
 	params->mode = bond_mode;
+	params->xmit_policy = xmit_hashtype;
 	params->miimon = miimon;
 	params->arp_interval = arp_interval;
 	params->updelay = updelay;
@@ -4669,6 +4975,7 @@ static int __init bonding_init(void)
 
 	rtnl_unlock();
 	register_netdevice_notifier(&bond_netdev_notifier);
+	register_inetaddr_notifier(&bond_inetaddr_notifier);
 
 	return 0;
 
@@ -4684,6 +4991,7 @@ out_err:
 static void __exit bonding_exit(void)
 {
 	unregister_netdevice_notifier(&bond_netdev_notifier);
+	unregister_inetaddr_notifier(&bond_inetaddr_notifier);
 
 	rtnl_lock();
 	bond_free_all();