bond_alb.c

linux-2.6.15.6

/*
 * Copyright(c) 1999 - 2004 Intel Corporation. All rights reserved.
 *
 * This program is free software; you can redistribute it and/or modify it
 * under the terms of the GNU General Public License as published by the
 * Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 *
 * This program is distributed in the hope that it will be useful, but
 * WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
 * or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
 * for more details.
 *
 * You should have received a copy of the GNU General Public License along
 * with this program; if not, write to the Free Software Foundation, Inc.,
 * 59 Temple Place - Suite 330, Boston, MA  02111-1307, USA.
 *
 * The full GNU General Public License is included in this distribution in the
 * file called LICENSE.
 *
 *
 * Changes:
 *
 * 2003/06/25 - Shmulik Hen <shmulik.hen at intel dot com>
 *	- Fixed signed/unsigned calculation errors that caused load sharing
 *	  to collapse to one slave under very heavy UDP Tx stress.
 *
 * 2003/08/06 - Amir Noam <amir.noam at intel dot com>
 *	- Add support for setting bond's MAC address with special
 *	  handling required for ALB/TLB.
 *
 * 2003/12/01 - Shmulik Hen <shmulik.hen at intel dot com>
 *	- Code cleanup and style changes
 *
 * 2003/12/30 - Amir Noam <amir.noam at intel dot com>
 *	- Fixed: Cannot remove and re-enslave the original active slave.
 *
 * 2004/01/14 - Shmulik Hen <shmulik.hen at intel dot com>
 *	- Add capability to tag self generated packets in ALB/TLB modes.
 */

//#define BONDING_DEBUG 1

#include <linux/skbuff.h>
#include <linux/netdevice.h>
#include <linux/etherdevice.h>
#include <linux/pkt_sched.h>
#include <linux/spinlock.h>
#include <linux/slab.h>
#include <linux/timer.h>
#include <linux/ip.h>
#include <linux/ipv6.h>
#include <linux/if_arp.h>
#include <linux/if_ether.h>
#include <linux/if_bonding.h>
#include <linux/if_vlan.h>
#include <linux/in.h>
#include <net/ipx.h>
#include <net/arp.h>
#include <asm/byteorder.h>
#include "bonding.h"
#include "bond_alb.h"


#define ALB_TIMER_TICKS_PER_SEC	    10	/* should be a divisor of HZ */
#define BOND_TLB_REBALANCE_INTERVAL 10	/* In seconds, periodic re-balancing.
					 * Used for division - never set
					 * to zero !!!
					 */
#define BOND_ALB_LP_INTERVAL	    1	/* In seconds, periodic send of
					 * learning packets to the switch
					 */

#define BOND_TLB_REBALANCE_TICKS (BOND_TLB_REBALANCE_INTERVAL \
				  * ALB_TIMER_TICKS_PER_SEC)

#define BOND_ALB_LP_TICKS (BOND_ALB_LP_INTERVAL \
			   * ALB_TIMER_TICKS_PER_SEC)

#define TLB_HASH_TABLE_SIZE 256	/* The size of the clients hash table.
				 * Note that this value MUST NOT be smaller
				 * because the key hash table is BYTE wide !
				 */


#define TLB_NULL_INDEX		0xffffffff
#define MAX_LP_BURST		3

/* rlb defs */
#define RLB_HASH_TABLE_SIZE	256
#define RLB_NULL_INDEX		0xffffffff
#define RLB_UPDATE_DELAY	2*ALB_TIMER_TICKS_PER_SEC /* 2 seconds */
#define RLB_ARP_BURST_SIZE	2
#define RLB_UPDATE_RETRY	3	/* 3-ticks - must be smaller than the rlb
					 * rebalance interval (5 min).
					 */
/* RLB_PROMISC_TIMEOUT = 10 sec equals the time that the current slave is
 * promiscuous after failover
 */
#define RLB_PROMISC_TIMEOUT	10*ALB_TIMER_TICKS_PER_SEC

static const u8 mac_bcast[ETH_ALEN] = {0xff,0xff,0xff,0xff,0xff,0xff};
static const int alb_delta_in_ticks = HZ / ALB_TIMER_TICKS_PER_SEC;

#pragma pack(1)
struct learning_pkt {
	u8 mac_dst[ETH_ALEN];
	u8 mac_src[ETH_ALEN];
	u16 type;
	u8 padding[ETH_ZLEN - ETH_HLEN];
};

struct arp_pkt {
	u16     hw_addr_space;
	u16     prot_addr_space;
	u8      hw_addr_len;
	u8      prot_addr_len;
	u16     op_code;
	u8      mac_src[ETH_ALEN];	/* sender hardware address */
	u32     ip_src;			/* sender IP address */
	u8      mac_dst[ETH_ALEN];	/* target hardware address */
	u32     ip_dst;			/* target IP address */
};
#pragma pack()

/* Forward declaration */
static void alb_send_learning_packets(struct slave *slave, u8 mac_addr[]);

static inline u8 _simple_hash(u8 *hash_start, int hash_size)
{
	int i;
	u8 hash = 0;

	for (i = 0; i < hash_size; i++) {
		hash ^= hash_start[i];
	}

	return hash;
}

/*********************** tlb specific functions ***************************/

static inline void _lock_tx_hashtbl(struct bonding *bond)
{
	spin_lock(&(BOND_ALB_INFO(bond).tx_hashtbl_lock));
}

static inline void _unlock_tx_hashtbl(struct bonding *bond)
{
	spin_unlock(&(BOND_ALB_INFO(bond).tx_hashtbl_lock));
}

/* Caller must hold tx_hashtbl lock */
static inline void tlb_init_table_entry(struct tlb_client_info *entry, int save_load)
{
	if (save_load) {
		entry->load_history = 1 + entry->tx_bytes /
				      BOND_TLB_REBALANCE_INTERVAL;
		entry->tx_bytes = 0;
	}

	entry->tx_slave = NULL;
	entry->next = TLB_NULL_INDEX;
	entry->prev = TLB_NULL_INDEX;
}

static inline void tlb_init_slave(struct slave *slave)
{
	SLAVE_TLB_INFO(slave).load = 0;
	SLAVE_TLB_INFO(slave).head = TLB_NULL_INDEX;
}

/* Caller must hold bond lock for read */
static void tlb_clear_slave(struct bonding *bond, struct slave *slave, int save_load)
{
	struct tlb_client_info *tx_hash_table;
	u32 index;

	_lock_tx_hashtbl(bond);

	/* clear slave from tx_hashtbl */
	tx_hash_table = BOND_ALB_INFO(bond).tx_hashtbl;

	index = SLAVE_TLB_INFO(slave).head;
	while (index != TLB_NULL_INDEX) {
		u32 next_index = tx_hash_table[index].next;
		tlb_init_table_entry(&tx_hash_table[index], save_load);
		index = next_index;
	}

	_unlock_tx_hashtbl(bond);

	tlb_init_slave(slave);
}

/* Must be called before starting the monitor timer */
static int tlb_initialize(struct bonding *bond)
{
	struct alb_bond_info *bond_info = &(BOND_ALB_INFO(bond));
	int size = TLB_HASH_TABLE_SIZE * sizeof(struct tlb_client_info);
	int i;

	spin_lock_init(&(bond_info->tx_hashtbl_lock));

	_lock_tx_hashtbl(bond);

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

	memset(bond_info->tx_hashtbl, 0, size);

	for (i = 0; i < TLB_HASH_TABLE_SIZE; i++) {
		tlb_init_table_entry(&bond_info->tx_hashtbl[i], 1);
	}

	_unlock_tx_hashtbl(bond);

	return 0;
}

/* Must be called only after all slaves have been released */
static void tlb_deinitialize(struct bonding *bond)
{
	struct alb_bond_info *bond_info = &(BOND_ALB_INFO(bond));

	_lock_tx_hashtbl(bond);

	kfree(bond_info->tx_hashtbl);
	bond_info->tx_hashtbl = NULL;

	_unlock_tx_hashtbl(bond);
}

/* Caller must hold bond lock for read */
static struct slave *tlb_get_least_loaded_slave(struct bonding *bond)
{
	struct slave *slave, *least_loaded;
	s64 max_gap;
	int i, found = 0;

	/* Find the first enabled slave */
	bond_for_each_slave(bond, slave, i) {
		if (SLAVE_IS_OK(slave)) {
			found = 1;
			break;
		}
	}

	if (!found) {
		return NULL;
	}

	least_loaded = slave;
	max_gap = (s64)(slave->speed << 20) - /* Convert to Megabit per sec */
			(s64)(SLAVE_TLB_INFO(slave).load << 3); /* Bytes to bits */

	/* Find the slave with the largest gap */
	bond_for_each_slave_from(bond, slave, i, least_loaded) {
		if (SLAVE_IS_OK(slave)) {
			s64 gap = (s64)(slave->speed << 20) -
					(s64)(SLAVE_TLB_INFO(slave).load << 3);
			if (max_gap < gap) {
				least_loaded = slave;
				max_gap = gap;
			}
		}
	}

	return least_loaded;
}

/* Caller must hold bond lock for read */
static struct slave *tlb_choose_channel(struct bonding *bond, u32 hash_index, u32 skb_len)
{
	struct alb_bond_info *bond_info = &(BOND_ALB_INFO(bond));
	struct tlb_client_info *hash_table;
	struct slave *assigned_slave;

	_lock_tx_hashtbl(bond);

	hash_table = bond_info->tx_hashtbl;
	assigned_slave = hash_table[hash_index].tx_slave;
	if (!assigned_slave) {
		assigned_slave = tlb_get_least_loaded_slave(bond);

		if (assigned_slave) {
			struct tlb_slave_info *slave_info =
				&(SLAVE_TLB_INFO(assigned_slave));
			u32 next_index = slave_info->head;

			hash_table[hash_index].tx_slave = assigned_slave;
			hash_table[hash_index].next = next_index;
			hash_table[hash_index].prev = TLB_NULL_INDEX;

			if (next_index != TLB_NULL_INDEX) {
				hash_table[next_index].prev = hash_index;
			}

			slave_info->head = hash_index;
			slave_info->load +=
				hash_table[hash_index].load_history;
		}
	}

	if (assigned_slave) {
		hash_table[hash_index].tx_bytes += skb_len;
	}

	_unlock_tx_hashtbl(bond);

	return assigned_slave;
}

/*********************** rlb specific functions ***************************/
static inline void _lock_rx_hashtbl(struct bonding *bond)
{
	spin_lock(&(BOND_ALB_INFO(bond).rx_hashtbl_lock));
}

static inline void _unlock_rx_hashtbl(struct bonding *bond)
{
	spin_unlock(&(BOND_ALB_INFO(bond).rx_hashtbl_lock));
}

/* when an ARP REPLY is received from a client update its info
 * in the rx_hashtbl
 */
static void rlb_update_entry_from_arp(struct bonding *bond, struct arp_pkt *arp)
{
	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 = _simple_hash((u8*)&(arp->ip_src), sizeof(arp->ip_src));
	client_info = &(bond_info->rx_hashtbl[hash_index]);

	if ((client_info->assigned) &&
	    (client_info->ip_src == arp->ip_dst) &&
	    (client_info->ip_dst == arp->ip_src)) {
		/* update the clients MAC address */
		memcpy(client_info->mac_dst, arp->mac_src, ETH_ALEN);
		client_info->ntt = 1;
		bond_info->rx_ntt = 1;
	}

	_unlock_rx_hashtbl(bond);
}

static int rlb_arp_recv(struct sk_buff *skb, struct net_device *bond_dev, struct packet_type *ptype, struct net_device *orig_dev)
{
	struct bonding *bond = bond_dev->priv;
	struct arp_pkt *arp = (struct arp_pkt *)skb->data;
	int res = NET_RX_DROP;

	if (!(bond_dev->flags & IFF_MASTER))
		goto out;

	if (!arp) {
		dprintk("Packet has no ARP data\n");
		goto out;
	}

	if (skb->len < sizeof(struct arp_pkt)) {
		dprintk("Packet is too small to be an ARP\n");
		goto out;
	}

	if (arp->op_code == htons(ARPOP_REPLY)) {
		/* update rx hash table for this ARP */
		rlb_update_entry_from_arp(bond, arp);
		dprintk("Server received an ARP Reply from client\n");
	}

	res = NET_RX_SUCCESS;

out:
	dev_kfree_skb(skb);

	return res;
}

/* Caller must hold bond lock for read */
static struct slave *rlb_next_rx_slave(struct bonding *bond)
{
	struct alb_bond_info *bond_info = &(BOND_ALB_INFO(bond));
	struct slave *rx_slave, *slave, *start_at;
	int i = 0;

	if (bond_info->next_rx_slave) {
		start_at = bond_info->next_rx_slave;
	} else {
		start_at = bond->first_slave;
	}

	rx_slave = NULL;

	bond_for_each_slave_from(bond, slave, i, start_at) {
		if (SLAVE_IS_OK(slave)) {
			if (!rx_slave) {
				rx_slave = slave;
			} else if (slave->speed > rx_slave->speed) {
				rx_slave = slave;
			}
		}
	}

	if (rx_slave) {
		bond_info->next_rx_slave = rx_slave->next;
	}

	return rx_slave;
}

/* teach the switch the mac of a disabled slave
 * on the primary for fault tolerance
 *
 * Caller must hold bond->curr_slave_lock for write or bond lock for write
 */
static void rlb_teach_disabled_mac_on_primary(struct bonding *bond, u8 addr[])
{
	if (!bond->curr_active_slave) {
		return;
	}

	if (!bond->alb_info.primary_is_promisc) {
		bond->alb_info.primary_is_promisc = 1;
		dev_set_promiscuity(bond->curr_active_slave->dev, 1);
	}

	bond->alb_info.rlb_promisc_timeout_counter = 0;

	alb_send_learning_packets(bond->curr_active_slave, addr);
}

/* slave being removed should not be active at this point
 *
 * Caller must hold bond lock for read
 */
static void rlb_clear_slave(struct bonding *bond, struct slave *slave)
{
	struct alb_bond_info *bond_info = &(BOND_ALB_INFO(bond));
	struct rlb_client_info *rx_hash_table;
	u32 index, next_index;

	/* clear slave from rx_hashtbl */
	_lock_rx_hashtbl(bond);

	rx_hash_table = bond_info->rx_hashtbl;
	index = bond_info->rx_hashtbl_head;
	for (; index != RLB_NULL_INDEX; index = next_index) {
		next_index = rx_hash_table[index].next;
		if (rx_hash_table[index].slave == slave) {
			struct slave *assigned_slave = rlb_next_rx_slave(bond);

			if (assigned_slave) {
				rx_hash_table[index].slave = assigned_slave;
				if (memcmp(rx_hash_table[index].mac_dst,
					   mac_bcast, ETH_ALEN)) {
					bond_info->rx_hashtbl[index].ntt = 1;
					bond_info->rx_ntt = 1;
					/* A slave has been removed from the
					 * table because it is either disabled
					 * or being released. We must retry the
					 * update to avoid clients from not
					 * being updated & disconnecting when
					 * there is stress
					 */
					bond_info->rlb_update_retry_counter =
						RLB_UPDATE_RETRY;
				}
			} else {  /* there is no active slave */
				rx_hash_table[index].slave = NULL;
			}
		}
	}

	_unlock_rx_hashtbl(bond);

	write_lock(&bond->curr_slave_lock);

	if (slave != bond->curr_active_slave) {
		rlb_teach_disabled_mac_on_primary(bond, slave->dev->dev_addr);
	}

	write_unlock(&bond->curr_slave_lock);
}

static void rlb_update_client(struct rlb_client_info *client_info)
{
	int i;

	if (!client_info->slave) {
		return;
	}

	for (i = 0; i < RLB_ARP_BURST_SIZE; i++) {
		struct sk_buff *skb;

		skb = arp_create(ARPOP_REPLY, ETH_P_ARP,
				 client_info->ip_dst,
				 client_info->slave->dev,
				 client_info->ip_src,
				 client_info->mac_dst,
				 client_info->slave->dev->dev_addr,
				 client_info->mac_dst);
		if (!skb) {
			printk(KERN_ERR DRV_NAME
			       ": Error: failed to create an ARP packet\n");
			continue;
		}

		skb->dev = client_info->slave->dev;

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

		arp_xmit(skb);
	}
}

/* sends ARP REPLIES that update the clients that need updating */
static void rlb_update_rx_clients(struct bonding *bond)
{
	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->ntt) {
			rlb_update_client(client_info);
			if (bond_info->rlb_update_retry_counter == 0) {
				client_info->ntt = 0;
			}
		}
	}

	/* do not update the entries again untill this counter is zero so that
	 * not to confuse the clients.
	 */
	bond_info->rlb_update_delay_counter = RLB_UPDATE_DELAY;

	_unlock_rx_hashtbl(bond);
}

/* The slave was assigned a new mac address - update the clients */
static void rlb_req_update_slave_clients(struct bonding *bond, struct slave *slave)