[net-next-2.6.git] / net / netfilter / ipvs / ip_vs_proto.c

/*
 * ip_vs_proto.c: transport protocol load balancing support for IPVS
 *
 * Authors:     Wensong Zhang <wensong@linuxvirtualserver.org>
 *              Julian Anastasov <ja@ssi.bg>
 *
 *              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:
 *
 */

#define KMSG_COMPONENT "IPVS"
#define pr_fmt(fmt) KMSG_COMPONENT ": " fmt

#include <linux/module.h>
#include <linux/kernel.h>
#include <linux/skbuff.h>
#include <linux/gfp.h>
#include <linux/in.h>
#include <linux/ip.h>
#include <net/protocol.h>
#include <net/tcp.h>
#include <net/udp.h>
#include <asm/system.h>
#include <linux/stat.h>
#include <linux/proc_fs.h>

#include <net/ip_vs.h>


/*
 * IPVS protocols can only be registered/unregistered when the ipvs
 * module is loaded/unloaded, so no lock is needed in accessing the
 * ipvs protocol table.
 */

#define IP_VS_PROTO_TAB_SIZE		32	/* must be power of 2 */
#define IP_VS_PROTO_HASH(proto)		((proto) & (IP_VS_PROTO_TAB_SIZE-1))

static struct ip_vs_protocol *ip_vs_proto_table[IP_VS_PROTO_TAB_SIZE];


/*
 *	register an ipvs protocol
 */
static int __used __init register_ip_vs_protocol(struct ip_vs_protocol *pp)
{
	unsigned hash = IP_VS_PROTO_HASH(pp->protocol);

	pp->next = ip_vs_proto_table[hash];
	ip_vs_proto_table[hash] = pp;

	if (pp->init != NULL)
		pp->init(pp);

	return 0;
}


/*
 *	unregister an ipvs protocol
 */
static int unregister_ip_vs_protocol(struct ip_vs_protocol *pp)
{
	struct ip_vs_protocol **pp_p;
	unsigned hash = IP_VS_PROTO_HASH(pp->protocol);

	pp_p = &ip_vs_proto_table[hash];
	for (; *pp_p; pp_p = &(*pp_p)->next) {
		if (*pp_p == pp) {
			*pp_p = pp->next;
			if (pp->exit != NULL)
				pp->exit(pp);
			return 0;
		}
	}

	return -ESRCH;
}


/*
 *	get ip_vs_protocol object by its proto.
 */
struct ip_vs_protocol * ip_vs_proto_get(unsigned short proto)
{
	struct ip_vs_protocol *pp;
	unsigned hash = IP_VS_PROTO_HASH(proto);

	for (pp = ip_vs_proto_table[hash]; pp; pp = pp->next) {
		if (pp->protocol == proto)
			return pp;
	}

	return NULL;
}
EXPORT_SYMBOL(ip_vs_proto_get);


/*
 *	Propagate event for state change to all protocols
 */
void ip_vs_protocol_timeout_change(int flags)
{
	struct ip_vs_protocol *pp;
	int i;

	for (i = 0; i < IP_VS_PROTO_TAB_SIZE; i++) {
		for (pp = ip_vs_proto_table[i]; pp; pp = pp->next) {
			if (pp->timeout_change)
				pp->timeout_change(pp, flags);
		}
	}
}


int *
ip_vs_create_timeout_table(int *table, int size)
{
	return kmemdup(table, size, GFP_ATOMIC);
}


/*
 *	Set timeout value for state specified by name
 */
int
ip_vs_set_state_timeout(int *table, int num, const char *const *names,
			const char *name, int to)
{
	int i;

	if (!table || !name || !to)
		return -EINVAL;

	for (i = 0; i < num; i++) {
		if (strcmp(names[i], name))
			continue;
		table[i] = to * HZ;
		return 0;
	}
	return -ENOENT;
}


const char * ip_vs_state_name(__u16 proto, int state)
{
	struct ip_vs_protocol *pp = ip_vs_proto_get(proto);

	if (pp == NULL || pp->state_name == NULL)
		return (IPPROTO_IP == proto) ? "NONE" : "ERR!";
	return pp->state_name(state);
}


static void
ip_vs_tcpudp_debug_packet_v4(struct ip_vs_protocol *pp,
			     const struct sk_buff *skb,
			     int offset,
			     const char *msg)
{
	char buf[128];
	struct iphdr _iph, *ih;

	ih = skb_header_pointer(skb, offset, sizeof(_iph), &_iph);
	if (ih == NULL)
		sprintf(buf, "TRUNCATED");
	else if (ih->frag_off & htons(IP_OFFSET))
		sprintf(buf, "%pI4->%pI4 frag", &ih->saddr, &ih->daddr);
	else {
		__be16 _ports[2], *pptr;

		pptr = skb_header_pointer(skb, offset + ih->ihl*4,
					  sizeof(_ports), _ports);
		if (pptr == NULL)
			sprintf(buf, "TRUNCATED %pI4->%pI4",
				&ih->saddr, &ih->daddr);
		else
			sprintf(buf, "%pI4:%u->%pI4:%u",
				&ih->saddr, ntohs(pptr[0]),
				&ih->daddr, ntohs(pptr[1]));
	}

	pr_debug("%s: %s %s\n", msg, pp->name, buf);
}

#ifdef CONFIG_IP_VS_IPV6
static void
ip_vs_tcpudp_debug_packet_v6(struct ip_vs_protocol *pp,
			     const struct sk_buff *skb,
			     int offset,
			     const char *msg)
{
	char buf[192];
	struct ipv6hdr _iph, *ih;

	ih = skb_header_pointer(skb, offset, sizeof(_iph), &_iph);
	if (ih == NULL)
		sprintf(buf, "TRUNCATED");
	else if (ih->nexthdr == IPPROTO_FRAGMENT)
		sprintf(buf, "%pI6->%pI6 frag",	&ih->saddr, &ih->daddr);
	else {
		__be16 _ports[2], *pptr;

		pptr = skb_header_pointer(skb, offset + sizeof(struct ipv6hdr),
					  sizeof(_ports), _ports);
		if (pptr == NULL)
			sprintf(buf, "TRUNCATED %pI6->%pI6",
				&ih->saddr, &ih->daddr);
		else
			sprintf(buf, "%pI6:%u->%pI6:%u",
				&ih->saddr, ntohs(pptr[0]),
				&ih->daddr, ntohs(pptr[1]));
	}

	pr_debug("%s: %s %s\n", msg, pp->name, buf);
}
#endif


void
ip_vs_tcpudp_debug_packet(int af, struct ip_vs_protocol *pp,
			  const struct sk_buff *skb,
			  int offset,
			  const char *msg)
{
#ifdef CONFIG_IP_VS_IPV6
	if (af == AF_INET6)
		ip_vs_tcpudp_debug_packet_v6(pp, skb, offset, msg);
	else
#endif
		ip_vs_tcpudp_debug_packet_v4(pp, skb, offset, msg);
}


int __init ip_vs_protocol_init(void)
{
	char protocols[64];
#define REGISTER_PROTOCOL(p)			\
	do {					\
		register_ip_vs_protocol(p);	\
		strcat(protocols, ", ");	\
		strcat(protocols, (p)->name);	\
	} while (0)

	protocols[0] = '\0';
	protocols[2] = '\0';
#ifdef CONFIG_IP_VS_PROTO_TCP
	REGISTER_PROTOCOL(&ip_vs_protocol_tcp);
#endif
#ifdef CONFIG_IP_VS_PROTO_UDP
	REGISTER_PROTOCOL(&ip_vs_protocol_udp);
#endif
#ifdef CONFIG_IP_VS_PROTO_SCTP
	REGISTER_PROTOCOL(&ip_vs_protocol_sctp);
#endif
#ifdef CONFIG_IP_VS_PROTO_AH
	REGISTER_PROTOCOL(&ip_vs_protocol_ah);
#endif
#ifdef CONFIG_IP_VS_PROTO_ESP
	REGISTER_PROTOCOL(&ip_vs_protocol_esp);
#endif
	pr_info("Registered protocols (%s)\n", &protocols[2]);

	return 0;
}


void ip_vs_protocol_cleanup(void)
{
	struct ip_vs_protocol *pp;
	int i;

	/* unregister all the ipvs protocols */
	for (i = 0; i < IP_VS_PROTO_TAB_SIZE; i++) {
		while ((pp = ip_vs_proto_table[i]) != NULL)
			unregister_ip_vs_protocol(pp);
	}
}
Commit	Line	Data
1da177e4 LT	1	/*
	2	* ip_vs_proto.c: transport protocol load balancing support for IPVS
	3	*
1da177e4 LT	4	* Authors: Wensong Zhang <wensong@linuxvirtualserver.org>
	5	* Julian Anastasov <ja@ssi.bg>
	6	*
	7	* This program is free software; you can redistribute it and/or
	8	* modify it under the terms of the GNU General Public License
	9	* as published by the Free Software Foundation; either version
	10	* 2 of the License, or (at your option) any later version.
	11	*
	12	* Changes:
	13	*
	14	*/
	15
9aada7ac HE	16	#define KMSG_COMPONENT "IPVS"
	17	#define pr_fmt(fmt) KMSG_COMPONENT ": " fmt
	18
1da177e4 LT	19	#include <linux/module.h>
	20	#include <linux/kernel.h>
	21	#include <linux/skbuff.h>
5a0e3ad6	22	#include <linux/gfp.h>
1da177e4 LT	23	#include <linux/in.h>
	24	#include <linux/ip.h>
	25	#include <net/protocol.h>
	26	#include <net/tcp.h>
	27	#include <net/udp.h>
	28	#include <asm/system.h>
	29	#include <linux/stat.h>
	30	#include <linux/proc_fs.h>
	31
	32	#include <net/ip_vs.h>
	33
	34
	35	/*
	36	* IPVS protocols can only be registered/unregistered when the ipvs
	37	* module is loaded/unloaded, so no lock is needed in accessing the
	38	* ipvs protocol table.
	39	*/
	40
	41	#define IP_VS_PROTO_TAB_SIZE 32 /* must be power of 2 */
	42	#define IP_VS_PROTO_HASH(proto) ((proto) & (IP_VS_PROTO_TAB_SIZE-1))
	43
	44	static struct ip_vs_protocol *ip_vs_proto_table[IP_VS_PROTO_TAB_SIZE];
	45
	46
	47	/*
	48	* register an ipvs protocol
	49	*/
048cf48b	50	static int __used __init register_ip_vs_protocol(struct ip_vs_protocol *pp)
1da177e4 LT	51	{
	52	unsigned hash = IP_VS_PROTO_HASH(pp->protocol);
	53
	54	pp->next = ip_vs_proto_table[hash];
	55	ip_vs_proto_table[hash] = pp;
	56
	57	if (pp->init != NULL)
	58	pp->init(pp);
	59
	60	return 0;
	61	}
	62
	63
	64	/*
	65	* unregister an ipvs protocol
	66	*/
	67	static int unregister_ip_vs_protocol(struct ip_vs_protocol *pp)
	68	{
	69	struct ip_vs_protocol **pp_p;
	70	unsigned hash = IP_VS_PROTO_HASH(pp->protocol);
	71
	72	pp_p = &ip_vs_proto_table[hash];
	73	for (; pp_p; pp_p = &(pp_p)->next) {
	74	if (*pp_p == pp) {
	75	*pp_p = pp->next;
	76	if (pp->exit != NULL)
	77	pp->exit(pp);
	78	return 0;
	79	}
	80	}
	81
	82	return -ESRCH;
	83	}
	84
	85
	86	/*
	87	* get ip_vs_protocol object by its proto.
	88	*/
	89	struct ip_vs_protocol * ip_vs_proto_get(unsigned short proto)
	90	{
	91	struct ip_vs_protocol *pp;
	92	unsigned hash = IP_VS_PROTO_HASH(proto);
	93
	94	for (pp = ip_vs_proto_table[hash]; pp; pp = pp->next) {
	95	if (pp->protocol == proto)
	96	return pp;
	97	}
	98
	99	return NULL;
	100	}
9c3e1c39	101	EXPORT_SYMBOL(ip_vs_proto_get);
1da177e4 LT	102
	103
	104	/*
	105	* Propagate event for state change to all protocols
	106	*/
	107	void ip_vs_protocol_timeout_change(int flags)
	108	{
	109	struct ip_vs_protocol *pp;
	110	int i;
	111
	112	for (i = 0; i < IP_VS_PROTO_TAB_SIZE; i++) {
	113	for (pp = ip_vs_proto_table[i]; pp; pp = pp->next) {
	114	if (pp->timeout_change)
	115	pp->timeout_change(pp, flags);
	116	}
	117	}
	118	}
	119
	120
	121	int *
	122	ip_vs_create_timeout_table(int *table, int size)
	123	{
8b2ed4bb	124	return kmemdup(table, size, GFP_ATOMIC);
1da177e4 LT	125	}
	126
	127
	128	/*
	129	* Set timeout value for state specified by name
	130	*/
	131	int
36cbd3dc JE	132	ip_vs_set_state_timeout(int table, int num, const char const *names,
36cbd3dc JE	133	const char *name, int to)
1da177e4 LT	134	{
	135	int i;
	136
	137	if (!table \|\| !name \|\| !to)
	138	return -EINVAL;
	139
	140	for (i = 0; i < num; i++) {
	141	if (strcmp(names[i], name))
	142	continue;
	143	table[i] = to * HZ;
	144	return 0;
	145	}
	146	return -ENOENT;
	147	}
	148
	149
	150	const char * ip_vs_state_name(__u16 proto, int state)
	151	{
	152	struct ip_vs_protocol *pp = ip_vs_proto_get(proto);
	153
	154	if (pp == NULL \|\| pp->state_name == NULL)
2ad17def	155	return (IPPROTO_IP == proto) ? "NONE" : "ERR!";
1da177e4 LT	156	return pp->state_name(state);
	157	}
	158
	159
77eb8516	160	static void
3b047d9d JV	161	ip_vs_tcpudp_debug_packet_v4(struct ip_vs_protocol *pp,
	162	const struct sk_buff *skb,
	163	int offset,
	164	const char *msg)
1da177e4 LT	165	{
	166	char buf[128];
	167	struct iphdr _iph, *ih;
	168
	169	ih = skb_header_pointer(skb, offset, sizeof(_iph), &_iph);
	170	if (ih == NULL)
3d91c1a8	171	sprintf(buf, "TRUNCATED");
5661df7b	172	else if (ih->frag_off & htons(IP_OFFSET))
3d91c1a8	173	sprintf(buf, "%pI4->%pI4 frag", &ih->saddr, &ih->daddr);
1da177e4	174	else {
0d79641a JA	175	__be16 _ports[2], *pptr;
0d79641a JA	176
1da177e4 LT	177	pptr = skb_header_pointer(skb, offset + ih->ihl*4,
	178	sizeof(_ports), _ports);
	179	if (pptr == NULL)
3d91c1a8 PM	180	sprintf(buf, "TRUNCATED %pI4->%pI4",
3d91c1a8 PM	181	&ih->saddr, &ih->daddr);
1da177e4	182	else
3d91c1a8	183	sprintf(buf, "%pI4:%u->%pI4:%u",
14d5e834 HH	184	&ih->saddr, ntohs(pptr[0]),
14d5e834 HH	185	&ih->daddr, ntohs(pptr[1]));
1da177e4 LT	186	}
1da177e4 LT	187
3d91c1a8	188	pr_debug("%s: %s %s\n", msg, pp->name, buf);
1da177e4 LT	189	}
1da177e4 LT	190
3b047d9d	191	#ifdef CONFIG_IP_VS_IPV6
77eb8516	192	static void
3b047d9d JV	193	ip_vs_tcpudp_debug_packet_v6(struct ip_vs_protocol *pp,
	194	const struct sk_buff *skb,
	195	int offset,
	196	const char *msg)
	197	{
	198	char buf[192];
	199	struct ipv6hdr _iph, *ih;
	200
	201	ih = skb_header_pointer(skb, offset, sizeof(_iph), &_iph);
	202	if (ih == NULL)
3d91c1a8	203	sprintf(buf, "TRUNCATED");
3b047d9d	204	else if (ih->nexthdr == IPPROTO_FRAGMENT)
3d91c1a8	205	sprintf(buf, "%pI6->%pI6 frag", &ih->saddr, &ih->daddr);
3b047d9d JV	206	else {
	207	__be16 _ports[2], *pptr;
	208
	209	pptr = skb_header_pointer(skb, offset + sizeof(struct ipv6hdr),
	210	sizeof(_ports), _ports);
	211	if (pptr == NULL)
3d91c1a8 PM	212	sprintf(buf, "TRUNCATED %pI6->%pI6",
3d91c1a8 PM	213	&ih->saddr, &ih->daddr);
3b047d9d	214	else
3d91c1a8	215	sprintf(buf, "%pI6:%u->%pI6:%u",
38ff4fa4 HH	216	&ih->saddr, ntohs(pptr[0]),
38ff4fa4 HH	217	&ih->daddr, ntohs(pptr[1]));
3b047d9d JV	218	}
3b047d9d JV	219
3d91c1a8	220	pr_debug("%s: %s %s\n", msg, pp->name, buf);
3b047d9d JV	221	}
	222	#endif
	223
	224
	225	void
0d79641a	226	ip_vs_tcpudp_debug_packet(int af, struct ip_vs_protocol *pp,
3b047d9d JV	227	const struct sk_buff *skb,
	228	int offset,
	229	const char *msg)
	230	{
	231	#ifdef CONFIG_IP_VS_IPV6
0d79641a	232	if (af == AF_INET6)
3b047d9d JV	233	ip_vs_tcpudp_debug_packet_v6(pp, skb, offset, msg);
	234	else
	235	#endif
	236	ip_vs_tcpudp_debug_packet_v4(pp, skb, offset, msg);
	237	}
	238
1da177e4	239
048cf48b	240	int __init ip_vs_protocol_init(void)
1da177e4 LT	241	{
	242	char protocols[64];
	243	#define REGISTER_PROTOCOL(p) \
	244	do { \
	245	register_ip_vs_protocol(p); \
	246	strcat(protocols, ", "); \
	247	strcat(protocols, (p)->name); \
	248	} while (0)
	249
	250	protocols[0] = '\0';
	251	protocols[2] = '\0';
	252	#ifdef CONFIG_IP_VS_PROTO_TCP
	253	REGISTER_PROTOCOL(&ip_vs_protocol_tcp);
	254	#endif
	255	#ifdef CONFIG_IP_VS_PROTO_UDP
	256	REGISTER_PROTOCOL(&ip_vs_protocol_udp);
	257	#endif
2906f66a VMR	258	#ifdef CONFIG_IP_VS_PROTO_SCTP
	259	REGISTER_PROTOCOL(&ip_vs_protocol_sctp);
	260	#endif
1da177e4 LT	261	#ifdef CONFIG_IP_VS_PROTO_AH
	262	REGISTER_PROTOCOL(&ip_vs_protocol_ah);
	263	#endif
	264	#ifdef CONFIG_IP_VS_PROTO_ESP
	265	REGISTER_PROTOCOL(&ip_vs_protocol_esp);
	266	#endif
1e3e238e	267	pr_info("Registered protocols (%s)\n", &protocols[2]);
1da177e4 LT	268
	269	return 0;
	270	}
	271
	272
	273	void ip_vs_protocol_cleanup(void)
	274	{
	275	struct ip_vs_protocol *pp;
	276	int i;
	277
	278	/* unregister all the ipvs protocols */
	279	for (i = 0; i < IP_VS_PROTO_TAB_SIZE; i++) {
	280	while ((pp = ip_vs_proto_table[i]) != NULL)
	281	unregister_ip_vs_protocol(pp);
	282	}
	283	}