ip_vs_lblc.c

linux-2.4.29操作系统的源码

/*
 * IPVS:        Locality-Based Least-Connection scheduling module
 *
 * Version:     $Id: ip_vs_lblc.c,v 1.9 2002/03/25 12:44:35 wensong Exp $
 *
 * Authors:     Wensong Zhang <[e-mail]>
 *
 *              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.
 *
 * Changes:
 *     Martin Hamilton         :    fixed the terrible locking bugs
 *                                   *lock(tbl->lock) ==> *lock(&tbl->lock)
 *     Wensong Zhang           :    fixed the uninitilized tbl->lock bug
 *     Wensong Zhang           :    added doing full expiration check to
 *                                   collect stale entries of 24+ hours when
 *                                   no partial expire check in a half hour
 *     Julian Anastasov        :    replaced del_timer call with del_timer_sync
 *                                   to avoid the possible race between timer
 *                                   handler and del_timer thread in SMP
 *
 */

/*
 * The lblc algorithm is as follows (pseudo code):
 *
 *       if cachenode[dest_ip] is null then
 *               n, cachenode[dest_ip] <- {weighted least-conn node};
 *       else
 *               n <- cachenode[dest_ip];
 *               if (n is dead) OR
 *                  (n.conns>n.weight AND
 *                   there is a node m with m.conns<m.weight/2) then
 *                 n, cachenode[dest_ip] <- {weighted least-conn node};
 *
 *       return n;
 *
 * Thanks must go to Wenzhuo Zhang for talking WCCP to me and pushing
 * me to write this module.
 */

#include <linux/module.h>
#include <linux/kernel.h>

/* for sysctl */
#include <linux/fs.h>
#include <linux/sysctl.h>

#include <net/ip_vs.h>


/*
 *    It is for garbage collection of stale IPVS lblc entries,
 *    when the table is full.
 */
#define CHECK_EXPIRE_INTERVAL   (60*HZ)
#define ENTRY_TIMEOUT           (6*60*HZ)

/*
 *    It is for full expiration check.
 *    When there is no partial expiration check (garbage collection)
 *    in a half hour, do a full expiration check to collect stale
 *    entries that haven't been touched for a day.
 */
#define COUNT_FOR_FULL_EXPIRATION   30
static int sysctl_ip_vs_lblc_expiration = 24*60*60*HZ;


/*
 *     for IPVS lblc entry hash table
 */
#ifndef CONFIG_IP_VS_LBLC_TAB_BITS
#define CONFIG_IP_VS_LBLC_TAB_BITS      10
#endif
#define IP_VS_LBLC_TAB_BITS     CONFIG_IP_VS_LBLC_TAB_BITS
#define IP_VS_LBLC_TAB_SIZE     (1 << IP_VS_LBLC_TAB_BITS)
#define IP_VS_LBLC_TAB_MASK     (IP_VS_LBLC_TAB_SIZE - 1)


/*
 *      IPVS lblc entry represents an association between destination
 *      IP address and its destination server
 */
struct ip_vs_lblc_entry {
	struct list_head        list;
	__u32                   addr;           /* destination IP address */
	struct ip_vs_dest       *dest;          /* real server (cache) */
	unsigned long           lastuse;        /* last used time */
};


/*
 *      IPVS lblc hash table
 */
struct ip_vs_lblc_table {
	rwlock_t	        lock;           /* lock for this table */
	struct list_head        bucket[IP_VS_LBLC_TAB_SIZE];  /* hash bucket */
	atomic_t                entries;        /* number of entries */
	int                     max_size;       /* maximum size of entries */
	struct timer_list       periodic_timer; /* collect stale entries */
	int                     rover;          /* rover for expire check */
	int                     counter;        /* counter for no expire */
};


/*
 *      IPVS LBLC sysctl table
 */
struct ip_vs_lblc_sysctl_table {
	struct ctl_table_header *sysctl_header;
	ctl_table vs_vars[2];
	ctl_table vs_dir[2];
	ctl_table ipv4_dir[2];
	ctl_table root_dir[2];
};


static struct ip_vs_lblc_sysctl_table lblc_sysctl_table = {
	NULL,
	{{NET_IPV4_VS_LBLC_EXPIRE, "lblc_expiration",
	  &sysctl_ip_vs_lblc_expiration,
	  sizeof(int), 0644, NULL, &proc_dointvec_jiffies},
	 {0}},
	{{NET_IPV4_VS, "vs", NULL, 0, 0555, lblc_sysctl_table.vs_vars},
	 {0}},
	{{NET_IPV4, "ipv4", NULL, 0, 0555, lblc_sysctl_table.vs_dir},
	 {0}},
	{{CTL_NET, "net", NULL, 0, 0555, lblc_sysctl_table.ipv4_dir},
	 {0}}
};


/*
 *      new/free a ip_vs_lblc_entry, which is a mapping of a destionation
 *      IP address to a server.
 */
static inline struct ip_vs_lblc_entry *
ip_vs_lblc_new(__u32 daddr, struct ip_vs_dest *dest)
{
	struct ip_vs_lblc_entry *en;

	en = kmalloc(sizeof(struct ip_vs_lblc_entry), GFP_ATOMIC);
	if (en == NULL) {
		IP_VS_ERR("ip_vs_lblc_new(): no memory\n");
		return NULL;
	}

	INIT_LIST_HEAD(&en->list);
	en->addr = daddr;

	atomic_inc(&dest->refcnt);
	en->dest = dest;

	return en;
}


static inline void ip_vs_lblc_free(struct ip_vs_lblc_entry *en)
{
	list_del(&en->list);
	/*
	 * We don't kfree dest because it is refered either by its service
	 * or the trash dest list.
	 */
	atomic_dec(&en->dest->refcnt);
	kfree(en);
}


/*
 *	Returns hash value for IPVS LBLC entry
 */
static inline unsigned ip_vs_lblc_hashkey(__u32 addr)
{
	return (ntohl(addr)*2654435761UL) & IP_VS_LBLC_TAB_MASK;
}


/*
 *	Hash an entry in the ip_vs_lblc_table.
 *	returns bool success.
 */
static int
ip_vs_lblc_hash(struct ip_vs_lblc_table *tbl, struct ip_vs_lblc_entry *en)
{
	unsigned hash;

	if (!list_empty(&en->list)) {
		IP_VS_ERR("ip_vs_lblc_hash(): request for already hashed, "
			  "called from %p\n", __builtin_return_address(0));
		return 0;
	}

	/*
	 *	Hash by destination IP address
	 */
	hash = ip_vs_lblc_hashkey(en->addr);

	write_lock(&tbl->lock);
	list_add(&en->list, &tbl->bucket[hash]);
	atomic_inc(&tbl->entries);
	write_unlock(&tbl->lock);

	return 1;
}


#if 0000
/*
 *	Unhash ip_vs_lblc_entry from ip_vs_lblc_table.
 *	returns bool success.
 */
static int ip_vs_lblc_unhash(struct ip_vs_lblc_table *tbl,
			     struct ip_vs_lblc_entry *en)
{
	if (list_empty(&en->list)) {
		IP_VS_ERR("ip_vs_lblc_unhash(): request for not hashed entry, "
			  "called from %p\n", __builtin_return_address(0));
		return 0;
	}

	/*
	 * Remove it from the table
	 */
	write_lock(&tbl->lock);
	list_del(&en->list);
	INIT_LIST_HEAD(&en->list);
	write_unlock(&tbl->lock);

	return 1;
}
#endif


/*
 *  Get ip_vs_lblc_entry associated with supplied parameters.
 */
static inline struct ip_vs_lblc_entry *
ip_vs_lblc_get(struct ip_vs_lblc_table *tbl, __u32 addr)
{
	unsigned hash;
	struct ip_vs_lblc_entry *en;
	struct list_head *l,*e;

	hash = ip_vs_lblc_hashkey(addr);
	l = &tbl->bucket[hash];

	read_lock(&tbl->lock);

	for (e=l->next; e!=l; e=e->next) {
		en = list_entry(e, struct ip_vs_lblc_entry, list);
		if (en->addr == addr) {
			/* HIT */
			read_unlock(&tbl->lock);
			return en;
		}
	}

	read_unlock(&tbl->lock);

	return NULL;
}


/*
 *      Flush all the entries of the specified table.
 */
static void ip_vs_lblc_flush(struct ip_vs_lblc_table *tbl)
{
	int i;
	struct list_head *l;
	struct ip_vs_lblc_entry *en;

	for (i=0; i<IP_VS_LBLC_TAB_SIZE; i++) {
		write_lock(&tbl->lock);
		for (l=&tbl->bucket[i]; l->next!=l; ) {
			en = list_entry(l->next,
					struct ip_vs_lblc_entry, list);
			ip_vs_lblc_free(en);
			atomic_dec(&tbl->entries);
		}
		write_unlock(&tbl->lock);
	}
}


static inline void ip_vs_lblc_full_check(struct ip_vs_lblc_table *tbl)
{
	unsigned long now = jiffies;
	int i, j;
	struct list_head *l, *e;
	struct ip_vs_lblc_entry *en;

	for (i=0, j=tbl->rover; i<IP_VS_LBLC_TAB_SIZE; i++) {
		j = (j + 1) & IP_VS_LBLC_TAB_MASK;
		e = l = &tbl->bucket[j];
		write_lock(&tbl->lock);
		while (e->next != l) {
			en = list_entry(e->next,
					struct ip_vs_lblc_entry, list);
			if ((now - en->lastuse) <
			    sysctl_ip_vs_lblc_expiration) {
				e = e->next;
				continue;
			}
			ip_vs_lblc_free(en);
			atomic_dec(&tbl->entries);
		}
		write_unlock(&tbl->lock);