ip_vs_core.c

linux-2.6.15.6

/*
 * IPVS         An implementation of the IP virtual server support for the
 *              LINUX operating system.  IPVS is now implemented as a module
 *              over the Netfilter framework. IPVS can be used to build a
 *              high-performance and highly available server based on a
 *              cluster of servers.
 *
 * Version:     $Id: ip_vs_core.c,v 1.34 2003/05/10 03:05:23 wensong Exp $
 *
 * Authors:     Wensong Zhang <[e-mail]>
 *              Peter Kese <[e-mail]>
 *              Julian Anastasov <[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.
 *
 * The IPVS code for kernel 2.2 was done by Wensong Zhang and Peter Kese,
 * with changes/fixes from Julian Anastasov, Lars Marowsky-Bree, Horms
 * and others.
 *
 * Changes:
 *	Paul `Rusty' Russell		properly handle non-linear skbs
 *	Harald Welte			don't use nfcache
 *
 */

#include <linux/module.h>
#include <linux/kernel.h>
#include <linux/ip.h>
#include <linux/tcp.h>
#include <linux/icmp.h>

#include <net/ip.h>
#include <net/tcp.h>
#include <net/udp.h>
#include <net/icmp.h>                   /* for icmp_send */
#include <net/route.h>

#include <linux/netfilter.h>
#include <linux/netfilter_ipv4.h>

#include <net/ip_vs.h>


EXPORT_SYMBOL(register_ip_vs_scheduler);
EXPORT_SYMBOL(unregister_ip_vs_scheduler);
EXPORT_SYMBOL(ip_vs_skb_replace);
EXPORT_SYMBOL(ip_vs_proto_name);
EXPORT_SYMBOL(ip_vs_conn_new);
EXPORT_SYMBOL(ip_vs_conn_in_get);
EXPORT_SYMBOL(ip_vs_conn_out_get);
#ifdef CONFIG_IP_VS_PROTO_TCP
EXPORT_SYMBOL(ip_vs_tcp_conn_listen);
#endif
EXPORT_SYMBOL(ip_vs_conn_put);
#ifdef CONFIG_IP_VS_DEBUG
EXPORT_SYMBOL(ip_vs_get_debug_level);
#endif
EXPORT_SYMBOL(ip_vs_make_skb_writable);


/* ID used in ICMP lookups */
#define icmp_id(icmph)          (((icmph)->un).echo.id)

const char *ip_vs_proto_name(unsigned proto)
{
	static char buf[20];

	switch (proto) {
	case IPPROTO_IP:
		return "IP";
	case IPPROTO_UDP:
		return "UDP";
	case IPPROTO_TCP:
		return "TCP";
	case IPPROTO_ICMP:
		return "ICMP";
	default:
		sprintf(buf, "IP_%d", proto);
		return buf;
	}
}

void ip_vs_init_hash_table(struct list_head *table, int rows)
{
	while (--rows >= 0)
		INIT_LIST_HEAD(&table[rows]);
}

static inline void
ip_vs_in_stats(struct ip_vs_conn *cp, struct sk_buff *skb)
{
	struct ip_vs_dest *dest = cp->dest;
	if (dest && (dest->flags & IP_VS_DEST_F_AVAILABLE)) {
		spin_lock(&dest->stats.lock);
		dest->stats.inpkts++;
		dest->stats.inbytes += skb->len;
		spin_unlock(&dest->stats.lock);

		spin_lock(&dest->svc->stats.lock);
		dest->svc->stats.inpkts++;
		dest->svc->stats.inbytes += skb->len;
		spin_unlock(&dest->svc->stats.lock);

		spin_lock(&ip_vs_stats.lock);
		ip_vs_stats.inpkts++;
		ip_vs_stats.inbytes += skb->len;
		spin_unlock(&ip_vs_stats.lock);
	}
}


static inline void
ip_vs_out_stats(struct ip_vs_conn *cp, struct sk_buff *skb)
{
	struct ip_vs_dest *dest = cp->dest;
	if (dest && (dest->flags & IP_VS_DEST_F_AVAILABLE)) {
		spin_lock(&dest->stats.lock);
		dest->stats.outpkts++;
		dest->stats.outbytes += skb->len;
		spin_unlock(&dest->stats.lock);

		spin_lock(&dest->svc->stats.lock);
		dest->svc->stats.outpkts++;
		dest->svc->stats.outbytes += skb->len;
		spin_unlock(&dest->svc->stats.lock);

		spin_lock(&ip_vs_stats.lock);
		ip_vs_stats.outpkts++;
		ip_vs_stats.outbytes += skb->len;
		spin_unlock(&ip_vs_stats.lock);
	}
}


static inline void
ip_vs_conn_stats(struct ip_vs_conn *cp, struct ip_vs_service *svc)
{
	spin_lock(&cp->dest->stats.lock);
	cp->dest->stats.conns++;
	spin_unlock(&cp->dest->stats.lock);

	spin_lock(&svc->stats.lock);
	svc->stats.conns++;
	spin_unlock(&svc->stats.lock);

	spin_lock(&ip_vs_stats.lock);
	ip_vs_stats.conns++;
	spin_unlock(&ip_vs_stats.lock);
}


static inline int
ip_vs_set_state(struct ip_vs_conn *cp, int direction,
		const struct sk_buff *skb,
		struct ip_vs_protocol *pp)
{
	if (unlikely(!pp->state_transition))
		return 0;
	return pp->state_transition(cp, direction, skb, pp);
}


int ip_vs_make_skb_writable(struct sk_buff **pskb, int writable_len)
{
	struct sk_buff *skb = *pskb;

	/* skb is already used, better copy skb and its payload */
	if (unlikely(skb_shared(skb) || skb->sk))
		goto copy_skb;

	/* skb data is already used, copy it */
	if (unlikely(skb_cloned(skb)))
		goto copy_data;

	return pskb_may_pull(skb, writable_len);

  copy_data:
	if (unlikely(writable_len > skb->len))
		return 0;
	return !pskb_expand_head(skb, 0, 0, GFP_ATOMIC);

  copy_skb:
	if (unlikely(writable_len > skb->len))
		return 0;
	skb = skb_copy(skb, GFP_ATOMIC);
	if (!skb)
		return 0;
	BUG_ON(skb_is_nonlinear(skb));

	/* Rest of kernel will get very unhappy if we pass it a
	   suddenly-orphaned skbuff */
	if ((*pskb)->sk)
		skb_set_owner_w(skb, (*pskb)->sk);
	kfree_skb(*pskb);
	*pskb = skb;
	return 1;
}

/*
 *  IPVS persistent scheduling function
 *  It creates a connection entry according to its template if exists,
 *  or selects a server and creates a connection entry plus a template.
 *  Locking: we are svc user (svc->refcnt), so we hold all dests too
 *  Protocols supported: TCP, UDP
 */
static struct ip_vs_conn *
ip_vs_sched_persist(struct ip_vs_service *svc,
		    const struct sk_buff *skb,
		    __u16 ports[2])
{
	struct ip_vs_conn *cp = NULL;
	struct iphdr *iph = skb->nh.iph;
	struct ip_vs_dest *dest;
	struct ip_vs_conn *ct;
	__u16  dport;	 /* destination port to forward */
	__u32  snet;	 /* source network of the client, after masking */

	/* Mask saddr with the netmask to adjust template granularity */
	snet = iph->saddr & svc->netmask;

	IP_VS_DBG(6, "p-schedule: src %u.%u.%u.%u:%u dest %u.%u.%u.%u:%u "
		  "mnet %u.%u.%u.%u\n",
		  NIPQUAD(iph->saddr), ntohs(ports[0]),
		  NIPQUAD(iph->daddr), ntohs(ports[1]),
		  NIPQUAD(snet));

	/*
	 * As far as we know, FTP is a very complicated network protocol, and
	 * it uses control connection and data connections. For active FTP,
	 * FTP server initialize data connection to the client, its source port
	 * is often 20. For passive FTP, FTP server tells the clients the port
	 * that it passively listens to,  and the client issues the data
	 * connection. In the tunneling or direct routing mode, the load
	 * balancer is on the client-to-server half of connection, the port
	 * number is unknown to the load balancer. So, a conn template like
	 * <caddr, 0, vaddr, 0, daddr, 0> is created for persistent FTP
	 * service, and a template like <caddr, 0, vaddr, vport, daddr, dport>
	 * is created for other persistent services.
	 */
	if (ports[1] == svc->port) {
		/* Check if a template already exists */
		if (svc->port != FTPPORT)
			ct = ip_vs_ct_in_get(iph->protocol, snet, 0,
					       iph->daddr, ports[1]);
		else
			ct = ip_vs_ct_in_get(iph->protocol, snet, 0,
					       iph->daddr, 0);

		if (!ct || !ip_vs_check_template(ct)) {
			/*
			 * No template found or the dest of the connection
			 * template is not available.
			 */
			dest = svc->scheduler->schedule(svc, skb);
			if (dest == NULL) {
				IP_VS_DBG(1, "p-schedule: no dest found.\n");
				return NULL;
			}

			/*
			 * Create a template like <protocol,caddr,0,
			 * vaddr,vport,daddr,dport> for non-ftp service,
			 * and <protocol,caddr,0,vaddr,0,daddr,0>
			 * for ftp service.
			 */
			if (svc->port != FTPPORT)
				ct = ip_vs_conn_new(iph->protocol,
						    snet, 0,
						    iph->daddr,
						    ports[1],
						    dest->addr, dest->port,
						    IP_VS_CONN_F_TEMPLATE,
						    dest);
			else
				ct = ip_vs_conn_new(iph->protocol,
						    snet, 0,
						    iph->daddr, 0,
						    dest->addr, 0,
						    IP_VS_CONN_F_TEMPLATE,
						    dest);
			if (ct == NULL)
				return NULL;

			ct->timeout = svc->timeout;
		} else {
			/* set destination with the found template */
			dest = ct->dest;
		}
		dport = dest->port;
	} else {
		/*
		 * Note: persistent fwmark-based services and persistent
		 * port zero service are handled here.
		 * fwmark template: <IPPROTO_IP,caddr,0,fwmark,0,daddr,0>
		 * port zero template: <protocol,caddr,0,vaddr,0,daddr,0>
		 */
		if (svc->fwmark)
			ct = ip_vs_ct_in_get(IPPROTO_IP, snet, 0,
					       htonl(svc->fwmark), 0);
		else
			ct = ip_vs_ct_in_get(iph->protocol, snet, 0,
					       iph->daddr, 0);

		if (!ct || !ip_vs_check_template(ct)) {
			/*
			 * If it is not persistent port zero, return NULL,
			 * otherwise create a connection template.
			 */
			if (svc->port)
				return NULL;

			dest = svc->scheduler->schedule(svc, skb);
			if (dest == NULL) {
				IP_VS_DBG(1, "p-schedule: no dest found.\n");
				return NULL;
			}

			/*
			 * Create a template according to the service
			 */
			if (svc->fwmark)
				ct = ip_vs_conn_new(IPPROTO_IP,
						    snet, 0,
						    htonl(svc->fwmark), 0,
						    dest->addr, 0,
						    IP_VS_CONN_F_TEMPLATE,
						    dest);
			else
				ct = ip_vs_conn_new(iph->protocol,
						    snet, 0,
						    iph->daddr, 0,
						    dest->addr, 0,
						    IP_VS_CONN_F_TEMPLATE,
						    dest);
			if (ct == NULL)
				return NULL;

			ct->timeout = svc->timeout;
		} else {
			/* set destination with the found template */
			dest = ct->dest;
		}
		dport = ports[1];
	}

	/*
	 *    Create a new connection according to the template
	 */
	cp = ip_vs_conn_new(iph->protocol,
			    iph->saddr, ports[0],
			    iph->daddr, ports[1],
			    dest->addr, dport,
			    0,
			    dest);
	if (cp == NULL) {
		ip_vs_conn_put(ct);
		return NULL;
	}

	/*
	 *    Add its control
	 */
	ip_vs_control_add(cp, ct);
	ip_vs_conn_put(ct);

	ip_vs_conn_stats(cp, svc);
	return cp;
}


/*
 *  IPVS main scheduling function
 *  It selects a server according to the virtual service, and
 *  creates a connection entry.
 *  Protocols supported: TCP, UDP
 */
struct ip_vs_conn *
ip_vs_schedule(struct ip_vs_service *svc, const struct sk_buff *skb)
{
	struct ip_vs_conn *cp = NULL;
	struct iphdr *iph = skb->nh.iph;
	struct ip_vs_dest *dest;
	__u16 _ports[2], *pptr;

	pptr = skb_header_pointer(skb, iph->ihl*4,
				  sizeof(_ports), _ports);
	if (pptr == NULL)
		return NULL;

	/*
	 *    Persistent service
	 */
	if (svc->flags & IP_VS_SVC_F_PERSISTENT)
		return ip_vs_sched_persist(svc, skb, pptr);