bond_alb.c

linux-2.6.15.6

{
	struct alb_bond_info *bond_info = &(BOND_ALB_INFO(bond));
	struct rlb_client_info *client_info;
	int ntt = 0;
	u32 hash_index;

	_lock_rx_hashtbl(bond);

	hash_index = bond_info->rx_hashtbl_head;
	for (; hash_index != RLB_NULL_INDEX; hash_index = client_info->next) {
		client_info = &(bond_info->rx_hashtbl[hash_index]);

		if ((client_info->slave == slave) &&
		    memcmp(client_info->mac_dst, mac_bcast, ETH_ALEN)) {
			client_info->ntt = 1;
			ntt = 1;
		}
	}

	// update the team's flag only after the whole iteration
	if (ntt) {
		bond_info->rx_ntt = 1;
		//fasten the change
		bond_info->rlb_update_retry_counter = RLB_UPDATE_RETRY;
	}

	_unlock_rx_hashtbl(bond);
}

/* mark all clients using src_ip to be updated */
static void rlb_req_update_subnet_clients(struct bonding *bond, u32 src_ip)
{
	struct alb_bond_info *bond_info = &(BOND_ALB_INFO(bond));
	struct rlb_client_info *client_info;
	u32 hash_index;

	_lock_rx_hashtbl(bond);

	hash_index = bond_info->rx_hashtbl_head;
	for (; hash_index != RLB_NULL_INDEX; hash_index = client_info->next) {
		client_info = &(bond_info->rx_hashtbl[hash_index]);

		if (!client_info->slave) {
			printk(KERN_ERR DRV_NAME
			       ": Error: found a client with no channel in "
			       "the client's hash table\n");
			continue;
		}
		/*update all clients using this src_ip, that are not assigned
		 * to the team's address (curr_active_slave) and have a known
		 * unicast mac address.
		 */
		if ((client_info->ip_src == src_ip) &&
		    memcmp(client_info->slave->dev->dev_addr,
			   bond->dev->dev_addr, ETH_ALEN) &&
		    memcmp(client_info->mac_dst, mac_bcast, ETH_ALEN)) {
			client_info->ntt = 1;
			bond_info->rx_ntt = 1;
		}
	}

	_unlock_rx_hashtbl(bond);
}

/* Caller must hold both bond and ptr locks for read */
static struct slave *rlb_choose_channel(struct sk_buff *skb, struct bonding *bond)
{
	struct alb_bond_info *bond_info = &(BOND_ALB_INFO(bond));
	struct arp_pkt *arp = (struct arp_pkt *)skb->nh.raw;
	struct slave *assigned_slave;
	struct rlb_client_info *client_info;
	u32 hash_index = 0;

	_lock_rx_hashtbl(bond);

	hash_index = _simple_hash((u8 *)&arp->ip_dst, sizeof(arp->ip_src));
	client_info = &(bond_info->rx_hashtbl[hash_index]);

	if (client_info->assigned) {
		if ((client_info->ip_src == arp->ip_src) &&
		    (client_info->ip_dst == arp->ip_dst)) {
			/* the entry is already assigned to this client */
			if (memcmp(arp->mac_dst, mac_bcast, ETH_ALEN)) {
				/* update mac address from arp */
				memcpy(client_info->mac_dst, arp->mac_dst, ETH_ALEN);
			}

			assigned_slave = client_info->slave;
			if (assigned_slave) {
				_unlock_rx_hashtbl(bond);
				return assigned_slave;
			}
		} else {
			/* the entry is already assigned to some other client,
			 * move the old client to primary (curr_active_slave) so
			 * that the new client can be assigned to this entry.
			 */
			if (bond->curr_active_slave &&
			    client_info->slave != bond->curr_active_slave) {
				client_info->slave = bond->curr_active_slave;
				rlb_update_client(client_info);
			}
		}
	}
	/* assign a new slave */
	assigned_slave = rlb_next_rx_slave(bond);

	if (assigned_slave) {
		client_info->ip_src = arp->ip_src;
		client_info->ip_dst = arp->ip_dst;
		/* arp->mac_dst is broadcast for arp reqeusts.
		 * will be updated with clients actual unicast mac address
		 * upon receiving an arp reply.
		 */
		memcpy(client_info->mac_dst, arp->mac_dst, ETH_ALEN);
		client_info->slave = assigned_slave;

		if (memcmp(client_info->mac_dst, mac_bcast, ETH_ALEN)) {
			client_info->ntt = 1;
			bond->alb_info.rx_ntt = 1;
		} else {
			client_info->ntt = 0;
		}

		if (!list_empty(&bond->vlan_list)) {
			unsigned short vlan_id;
			int res = vlan_get_tag(skb, &vlan_id);
			if (!res) {
				client_info->tag = 1;
				client_info->vlan_id = vlan_id;
			}
		}

		if (!client_info->assigned) {
			u32 prev_tbl_head = bond_info->rx_hashtbl_head;
			bond_info->rx_hashtbl_head = hash_index;
			client_info->next = prev_tbl_head;
			if (prev_tbl_head != RLB_NULL_INDEX) {
				bond_info->rx_hashtbl[prev_tbl_head].prev =
					hash_index;
			}
			client_info->assigned = 1;
		}
	}

	_unlock_rx_hashtbl(bond);

	return assigned_slave;
}

/* chooses (and returns) transmit channel for arp reply
 * does not choose channel for other arp types since they are
 * sent on the curr_active_slave
 */
static struct slave *rlb_arp_xmit(struct sk_buff *skb, struct bonding *bond)
{
	struct arp_pkt *arp = (struct arp_pkt *)skb->nh.raw;
	struct slave *tx_slave = NULL;

	if (arp->op_code == __constant_htons(ARPOP_REPLY)) {
		/* the arp must be sent on the selected
		* rx channel
		*/
		tx_slave = rlb_choose_channel(skb, bond);
		if (tx_slave) {
			memcpy(arp->mac_src,tx_slave->dev->dev_addr, ETH_ALEN);
		}
		dprintk("Server sent ARP Reply packet\n");
	} else if (arp->op_code == __constant_htons(ARPOP_REQUEST)) {
		/* Create an entry in the rx_hashtbl for this client as a
		 * place holder.
		 * When the arp reply is received the entry will be updated
		 * with the correct unicast address of the client.
		 */
		rlb_choose_channel(skb, bond);

		/* The ARP relpy packets must be delayed so that
		 * they can cancel out the influence of the ARP request.
		 */
		bond->alb_info.rlb_update_delay_counter = RLB_UPDATE_DELAY;

		/* arp requests are broadcast and are sent on the primary
		 * the arp request will collapse all clients on the subnet to
		 * the primary slave. We must register these clients to be
		 * updated with their assigned mac.
		 */
		rlb_req_update_subnet_clients(bond, arp->ip_src);
		dprintk("Server sent ARP Request packet\n");
	}

	return tx_slave;
}

/* Caller must hold bond lock for read */
static void rlb_rebalance(struct bonding *bond)
{
	struct alb_bond_info *bond_info = &(BOND_ALB_INFO(bond));
	struct slave *assigned_slave;
	struct rlb_client_info *client_info;
	int ntt;
	u32 hash_index;

	_lock_rx_hashtbl(bond);

	ntt = 0;
	hash_index = bond_info->rx_hashtbl_head;
	for (; hash_index != RLB_NULL_INDEX; hash_index = client_info->next) {
		client_info = &(bond_info->rx_hashtbl[hash_index]);
		assigned_slave = rlb_next_rx_slave(bond);
		if (assigned_slave && (client_info->slave != assigned_slave)) {
			client_info->slave = assigned_slave;
			client_info->ntt = 1;
			ntt = 1;
		}
	}

	/* update the team's flag only after the whole iteration */
	if (ntt) {
		bond_info->rx_ntt = 1;
	}
	_unlock_rx_hashtbl(bond);
}

/* Caller must hold rx_hashtbl lock */
static void rlb_init_table_entry(struct rlb_client_info *entry)
{
	memset(entry, 0, sizeof(struct rlb_client_info));
	entry->next = RLB_NULL_INDEX;
	entry->prev = RLB_NULL_INDEX;
}

static int rlb_initialize(struct bonding *bond)
{
	struct alb_bond_info *bond_info = &(BOND_ALB_INFO(bond));
	struct packet_type *pk_type = &(BOND_ALB_INFO(bond).rlb_pkt_type);
	int size = RLB_HASH_TABLE_SIZE * sizeof(struct rlb_client_info);
	int i;

	spin_lock_init(&(bond_info->rx_hashtbl_lock));

	_lock_rx_hashtbl(bond);

	bond_info->rx_hashtbl = kmalloc(size, GFP_KERNEL);
	if (!bond_info->rx_hashtbl) {
		printk(KERN_ERR DRV_NAME
		       ": Error: %s: Failed to allocate RLB hash table\n",
		       bond->dev->name);
		_unlock_rx_hashtbl(bond);
		return -1;
	}

	bond_info->rx_hashtbl_head = RLB_NULL_INDEX;

	for (i = 0; i < RLB_HASH_TABLE_SIZE; i++) {
		rlb_init_table_entry(bond_info->rx_hashtbl + i);
	}

	_unlock_rx_hashtbl(bond);

	/*initialize packet type*/
	pk_type->type = __constant_htons(ETH_P_ARP);
	pk_type->dev = bond->dev;
	pk_type->func = rlb_arp_recv;

	/* register to receive ARPs */
	dev_add_pack(pk_type);

	return 0;
}

static void rlb_deinitialize(struct bonding *bond)
{
	struct alb_bond_info *bond_info = &(BOND_ALB_INFO(bond));

	dev_remove_pack(&(bond_info->rlb_pkt_type));

	_lock_rx_hashtbl(bond);

	kfree(bond_info->rx_hashtbl);
	bond_info->rx_hashtbl = NULL;
	bond_info->rx_hashtbl_head = RLB_NULL_INDEX;

	_unlock_rx_hashtbl(bond);
}

static void rlb_clear_vlan(struct bonding *bond, unsigned short vlan_id)
{
	struct alb_bond_info *bond_info = &(BOND_ALB_INFO(bond));
	u32 curr_index;

	_lock_rx_hashtbl(bond);

	curr_index = bond_info->rx_hashtbl_head;
	while (curr_index != RLB_NULL_INDEX) {
		struct rlb_client_info *curr = &(bond_info->rx_hashtbl[curr_index]);
		u32 next_index = bond_info->rx_hashtbl[curr_index].next;
		u32 prev_index = bond_info->rx_hashtbl[curr_index].prev;

		if (curr->tag && (curr->vlan_id == vlan_id)) {
			if (curr_index == bond_info->rx_hashtbl_head) {
				bond_info->rx_hashtbl_head = next_index;
			}
			if (prev_index != RLB_NULL_INDEX) {
				bond_info->rx_hashtbl[prev_index].next = next_index;
			}
			if (next_index != RLB_NULL_INDEX) {
				bond_info->rx_hashtbl[next_index].prev = prev_index;
			}

			rlb_init_table_entry(curr);
		}

		curr_index = next_index;
	}

	_unlock_rx_hashtbl(bond);
}

/*********************** tlb/rlb shared functions *********************/

static void alb_send_learning_packets(struct slave *slave, u8 mac_addr[])
{
	struct bonding *bond = bond_get_bond_by_slave(slave);
	struct learning_pkt pkt;
	int size = sizeof(struct learning_pkt);
	int i;

	memset(&pkt, 0, size);
	memcpy(pkt.mac_dst, mac_addr, ETH_ALEN);
	memcpy(pkt.mac_src, mac_addr, ETH_ALEN);
	pkt.type = __constant_htons(ETH_P_LOOP);

	for (i = 0; i < MAX_LP_BURST; i++) {
		struct sk_buff *skb;
		char *data;

		skb = dev_alloc_skb(size);
		if (!skb) {
			return;
		}

		data = skb_put(skb, size);
		memcpy(data, &pkt, size);

		skb->mac.raw = data;
		skb->nh.raw = data + ETH_HLEN;
		skb->protocol = pkt.type;
		skb->priority = TC_PRIO_CONTROL;
		skb->dev = slave->dev;

		if (!list_empty(&bond->vlan_list)) {
			struct vlan_entry *vlan;

			vlan = bond_next_vlan(bond,
					      bond->alb_info.current_alb_vlan);

			bond->alb_info.current_alb_vlan = vlan;
			if (!vlan) {
				kfree_skb(skb);
				continue;
			}

			skb = vlan_put_tag(skb, vlan->vlan_id);
			if (!skb) {
				printk(KERN_ERR DRV_NAME
				       ": Error: failed to insert VLAN tag\n");
				continue;
			}
		}

		dev_queue_xmit(skb);
	}
}

/* hw is a boolean parameter that determines whether we should try and
 * set the hw address of the device as well as the hw address of the
 * net_device
 */
static int alb_set_slave_mac_addr(struct slave *slave, u8 addr[], int hw)
{
	struct net_device *dev = slave->dev;
	struct sockaddr s_addr;

	if (!hw) {
		memcpy(dev->dev_addr, addr, dev->addr_len);
		return 0;
	}

	/* for rlb each slave must have a unique hw mac addresses so that */
	/* each slave will receive packets destined to a different mac */
	memcpy(s_addr.sa_data, addr, dev->addr_len);
	s_addr.sa_family = dev->type;
	if (dev_set_mac_address(dev, &s_addr)) {
		printk(KERN_ERR DRV_NAME
		       ": Error: dev_set_mac_address of dev %s failed! ALB "
		       "mode requires that the base driver support setting "
		       "the hw address also when the network device's "
		       "interface is open\n",
		       dev->name);
		return -EOPNOTSUPP;
	}
	return 0;
}

/* Caller must hold bond lock for write or curr_slave_lock for write*/
static void alb_swap_mac_addr(struct bonding *bond, struct slave *slave1, struct slave *slave2)
{
	struct slave *disabled_slave = NULL;
	u8 tmp_mac_addr[ETH_ALEN];
	int slaves_state_differ;

	slaves_state_differ = (SLAVE_IS_OK(slave1) != SLAVE_IS_OK(slave2));

	memcpy(tmp_mac_addr, slave1->dev->dev_addr, ETH_ALEN);
	alb_set_slave_mac_addr(slave1, slave2->dev->dev_addr, bond->alb_info.rlb_enabled);
	alb_set_slave_mac_addr(slave2, tmp_mac_addr, bond->alb_info.rlb_enabled);

	/* fasten the change in the switch */
	if (SLAVE_IS_OK(slave1)) {
		alb_send_learning_packets(slave1, slave1->dev->dev_addr);
		if (bond->alb_info.rlb_enabled) {
			/* inform the clients that the mac address
			 * has changed
			 */
			rlb_req_update_slave_clients(bond, slave1);
		}
	} else {
		disabled_slave = slave1;
	}

	if (SLAVE_IS_OK(slave2)) {
		alb_send_learning_packets(slave2, slave2->dev->dev_addr);
		if (bond->alb_info.rlb_enabled) {
			/* inform the clients that the mac address
			 * has changed
			 */
			rlb_req_update_slave_clients(bond, slave2);
		}
	} else {
		disabled_slave = slave2;
	}

	if (bond->alb_info.rlb_enabled && slaves_state_differ) {
		/* A disabled slave was assigned an active mac addr */
		rlb_teach_disabled_mac_on_primary(bond,
						  disabled_slave->dev->dev_addr);
	}
}

/**
 * alb_change_hw_addr_on_detach
 * @bond: bonding we're working on
 * @slave: the slave that was just detached
 *
 * We assume that @slave was already detached from the slave list.
 *
 * If @slave's permanent hw address is different both from its current
 * address and from @bond's address, then somewhere in the bond there's
 * a slave that has @slave's permanet address as its current address.
 * We'll make sure that that slave no longer uses @slave's permanent address.
 *
 * Caller must hold bond lock
 */
static void alb_change_hw_addr_on_detach(struct bonding *bond, struct slave *slave)
{
	int perm_curr_diff;
	int perm_bond_diff;

	perm_curr_diff = memcmp(slave->perm_hwaddr,
				slave->dev->dev_addr,
				ETH_ALEN);
	perm_bond_diff = memcmp(slave->perm_hwaddr,
				bond->dev->dev_addr,
				ETH_ALEN);

	if (perm_curr_diff && perm_bond_diff) {
		struct slave *tmp_slave;
		int i, found = 0;

		bond_for_each_slave(bond, tmp_slave, i) {
			if (!memcmp(slave->perm_hwaddr,
				    tmp_slave->dev->dev_addr,
				    ETH_ALEN)) {
				found = 1;
				break;
			}
		}

		if (found) {
			alb_swap_mac_addr(bond, slave, tmp_slave);
		}
	}
}

/**
 * alb_handle_addr_collision_on_attach
 * @bond: bonding we're working on
 * @slave: the slave that was just attached
 *
 * checks uniqueness of slave's mac address and handles the case the
 * new slave uses the bonds mac address.
 *
 * If the permanent hw address of @slave is @bond's hw address, we need to
 * find a different hw address to give @slave, that isn't in use by any other
 * slave in the bond. This address must be, of course, one of the premanent
 * addresses of the other slaves.
 *
 * We go over the slave list, and for each slave there we compare its
 * permanent hw address with the current address of all the other slaves.
 * If no match was found, then we've found a slave with a permanent address
 * that isn't used by any other slave in the bond, so we can assign it to
 * @slave.
 *
 * assumption: this function is called before @slave is attached to the
 * 	       bond slave list.
 *
 * caller must hold the bond lock for write since the mac addresses are compared
 * and may be swapped.
 */
static int alb_handle_addr_collision_on_attach(struct bonding *bond, struct slave *slave)
{
	struct slave *tmp_slave1, *tmp_slave2, *free_mac_slave;
	struct slave *has_bond_addr = bond->curr_active_slave;
	int i, j, found = 0;

	if (bond->slave_cnt == 0) {
		/* this is the first slave */
		return 0;
	}

	/* if slave's mac address differs from bond's mac address
	 * check uniqueness of slave's mac address against the other
	 * slaves in the bond.
	 */
	if (memcmp(slave->perm_hwaddr, bond->dev->dev_addr, ETH_ALEN)) {
		bond_for_each_slave(bond, tmp_slave1, i) {
			if (!memcmp(tmp_slave1->dev->dev_addr, slave->dev->dev_addr,
				    ETH_ALEN)) {
				found = 1;
				break;
			}
		}

		if (!found)
			return 0;

		/* Try setting slave mac to bond address and fall-through
		   to code handling that situation below... */
		alb_set_slave_mac_addr(slave, bond->dev->dev_addr,
				       bond->alb_info.rlb_enabled);
	}

	/* The slave's address is equal to the address of the bond.
	 * Search for a spare address in the bond for this slave.
	 */
	free_mac_slave = NULL;

	bond_for_each_slave(bond, tmp_slave1, i) {
		found = 0;
		bond_for_each_slave(bond, tmp_slave2, j) {
			if (!memcmp(tmp_slave1->perm_hwaddr,
				    tmp_slave2->dev->dev_addr,
				    ETH_ALEN)) {
				found = 1;