Fix common misspellings
[GitHub/mt8127/android_kernel_alcatel_ttab.git] / include / net / ip_vs.h
CommitLineData
1da177e4
LT
1/*
2 * IP Virtual Server
3 * data structure and functionality definitions
4 */
5
bc4768eb
JV
6#ifndef _NET_IP_VS_H
7#define _NET_IP_VS_H
1da177e4 8
bc4768eb 9#include <linux/ip_vs.h> /* definitions shared with userland */
1da177e4 10
bc4768eb 11/* old ipvsadm versions still include this file directly */
1da177e4
LT
12#ifdef __KERNEL__
13
bc4768eb
JV
14#include <asm/types.h> /* for __uXX types */
15
16#include <linux/sysctl.h> /* for ctl_path */
1da177e4
LT
17#include <linux/list.h> /* for struct list_head */
18#include <linux/spinlock.h> /* for struct rwlock_t */
1da177e4 19#include <asm/atomic.h> /* for struct atomic_t */
1da177e4 20#include <linux/compiler.h>
14c85021 21#include <linux/timer.h>
1da177e4 22
14c85021 23#include <net/checksum.h>
e7ade46a 24#include <linux/netfilter.h> /* for union nf_inet_addr */
1668e010 25#include <linux/ip.h>
e7ade46a
JV
26#include <linux/ipv6.h> /* for struct ipv6hdr */
27#include <net/ipv6.h> /* for ipv6_addr_copy */
cf356d69 28#if defined(CONFIG_NF_CONNTRACK) || defined(CONFIG_NF_CONNTRACK_MODULE)
f4bc17cd
JA
29#include <net/netfilter/nf_conntrack.h>
30#endif
61b1ab45
HS
31#include <net/net_namespace.h> /* Netw namespace */
32
33/*
34 * Generic access of ipvs struct
35 */
36static inline struct netns_ipvs *net_ipvs(struct net* net)
37{
38 return net->ipvs;
39}
fc723250
HS
40/*
41 * Get net ptr from skb in traffic cases
42 * use skb_sknet when call is from userland (ioctl or netlink)
43 */
a0840e2e 44static inline struct net *skb_net(const struct sk_buff *skb)
fc723250
HS
45{
46#ifdef CONFIG_NET_NS
47#ifdef CONFIG_IP_VS_DEBUG
48 /*
49 * This is used for debug only.
50 * Start with the most likely hit
51 * End with BUG
52 */
53 if (likely(skb->dev && skb->dev->nd_net))
54 return dev_net(skb->dev);
55 if (skb_dst(skb)->dev)
56 return dev_net(skb_dst(skb)->dev);
57 WARN(skb->sk, "Maybe skb_sknet should be used in %s() at line:%d\n",
58 __func__, __LINE__);
59 if (likely(skb->sk && skb->sk->sk_net))
60 return sock_net(skb->sk);
61 pr_err("There is no net ptr to find in the skb in %s() line:%d\n",
62 __func__, __LINE__);
63 BUG();
64#else
65 return dev_net(skb->dev ? : skb_dst(skb)->dev);
66#endif
67#else
68 return &init_net;
69#endif
70}
71
a0840e2e 72static inline struct net *skb_sknet(const struct sk_buff *skb)
fc723250
HS
73{
74#ifdef CONFIG_NET_NS
75#ifdef CONFIG_IP_VS_DEBUG
76 /* Start with the most likely hit */
77 if (likely(skb->sk && skb->sk->sk_net))
78 return sock_net(skb->sk);
79 WARN(skb->dev, "Maybe skb_net should be used instead in %s() line:%d\n",
80 __func__, __LINE__);
81 if (likely(skb->dev && skb->dev->nd_net))
82 return dev_net(skb->dev);
83 pr_err("There is no net ptr to find in the skb in %s() line:%d\n",
84 __func__, __LINE__);
85 BUG();
86#else
87 return sock_net(skb->sk);
88#endif
89#else
90 return &init_net;
91#endif
92}
b17fc996
HS
93/*
94 * This one needed for single_open_net since net is stored directly in
25985edc 95 * private not as a struct i.e. seq_file_net can't be used.
b17fc996
HS
96 */
97static inline struct net *seq_file_single_net(struct seq_file *seq)
98{
99#ifdef CONFIG_NET_NS
100 return (struct net *)seq->private;
101#else
102 return &init_net;
103#endif
104}
6f7edb48
CB
105
106/* Connections' size value needed by ip_vs_ctl.c */
107extern int ip_vs_conn_tab_size;
108
109
64aae3cb
JV
110struct ip_vs_iphdr {
111 int len;
112 __u8 protocol;
113 union nf_inet_addr saddr;
114 union nf_inet_addr daddr;
115};
116
117static inline void
118ip_vs_fill_iphdr(int af, const void *nh, struct ip_vs_iphdr *iphdr)
119{
120#ifdef CONFIG_IP_VS_IPV6
121 if (af == AF_INET6) {
122 const struct ipv6hdr *iph = nh;
123 iphdr->len = sizeof(struct ipv6hdr);
124 iphdr->protocol = iph->nexthdr;
125 ipv6_addr_copy(&iphdr->saddr.in6, &iph->saddr);
126 ipv6_addr_copy(&iphdr->daddr.in6, &iph->daddr);
127 } else
128#endif
129 {
130 const struct iphdr *iph = nh;
131 iphdr->len = iph->ihl * 4;
132 iphdr->protocol = iph->protocol;
133 iphdr->saddr.ip = iph->saddr;
134 iphdr->daddr.ip = iph->daddr;
135 }
136}
137
138static inline void ip_vs_addr_copy(int af, union nf_inet_addr *dst,
139 const union nf_inet_addr *src)
140{
141#ifdef CONFIG_IP_VS_IPV6
142 if (af == AF_INET6)
143 ipv6_addr_copy(&dst->in6, &src->in6);
144 else
145#endif
146 dst->ip = src->ip;
147}
148
149static inline int ip_vs_addr_equal(int af, const union nf_inet_addr *a,
150 const union nf_inet_addr *b)
151{
152#ifdef CONFIG_IP_VS_IPV6
153 if (af == AF_INET6)
154 return ipv6_addr_equal(&a->in6, &b->in6);
155#endif
156 return a->ip == b->ip;
157}
158
1da177e4 159#ifdef CONFIG_IP_VS_DEBUG
14c85021
ACM
160#include <linux/net.h>
161
1da177e4 162extern int ip_vs_get_debug_level(void);
c842a3ad
JV
163
164static inline const char *ip_vs_dbg_addr(int af, char *buf, size_t buf_len,
165 const union nf_inet_addr *addr,
166 int *idx)
167{
168 int len;
169#ifdef CONFIG_IP_VS_IPV6
170 if (af == AF_INET6)
5b095d98 171 len = snprintf(&buf[*idx], buf_len - *idx, "[%pI6]",
0c6ce78a 172 &addr->in6) + 1;
c842a3ad
JV
173 else
174#endif
3685f25d
HH
175 len = snprintf(&buf[*idx], buf_len - *idx, "%pI4",
176 &addr->ip) + 1;
c842a3ad
JV
177
178 *idx += len;
179 BUG_ON(*idx > buf_len + 1);
180 return &buf[*idx - len];
181}
182
9aada7ac
HE
183#define IP_VS_DBG_BUF(level, msg, ...) \
184 do { \
185 char ip_vs_dbg_buf[160]; \
186 int ip_vs_dbg_idx = 0; \
187 if (level <= ip_vs_get_debug_level()) \
188 printk(KERN_DEBUG pr_fmt(msg), ##__VA_ARGS__); \
189 } while (0)
190#define IP_VS_ERR_BUF(msg...) \
191 do { \
192 char ip_vs_dbg_buf[160]; \
193 int ip_vs_dbg_idx = 0; \
194 pr_err(msg); \
195 } while (0)
c842a3ad
JV
196
197/* Only use from within IP_VS_DBG_BUF() or IP_VS_ERR_BUF macros */
9aada7ac
HE
198#define IP_VS_DBG_ADDR(af, addr) \
199 ip_vs_dbg_addr(af, ip_vs_dbg_buf, \
200 sizeof(ip_vs_dbg_buf), addr, \
201 &ip_vs_dbg_idx)
202
203#define IP_VS_DBG(level, msg, ...) \
204 do { \
205 if (level <= ip_vs_get_debug_level()) \
206 printk(KERN_DEBUG pr_fmt(msg), ##__VA_ARGS__); \
207 } while (0)
208#define IP_VS_DBG_RL(msg, ...) \
209 do { \
210 if (net_ratelimit()) \
211 printk(KERN_DEBUG pr_fmt(msg), ##__VA_ARGS__); \
212 } while (0)
0d79641a 213#define IP_VS_DBG_PKT(level, af, pp, skb, ofs, msg) \
9aada7ac
HE
214 do { \
215 if (level <= ip_vs_get_debug_level()) \
0d79641a 216 pp->debug_packet(af, pp, skb, ofs, msg); \
9aada7ac 217 } while (0)
0d79641a 218#define IP_VS_DBG_RL_PKT(level, af, pp, skb, ofs, msg) \
9aada7ac
HE
219 do { \
220 if (level <= ip_vs_get_debug_level() && \
221 net_ratelimit()) \
0d79641a 222 pp->debug_packet(af, pp, skb, ofs, msg); \
9aada7ac 223 } while (0)
1da177e4 224#else /* NO DEBUGGING at ALL */
c842a3ad
JV
225#define IP_VS_DBG_BUF(level, msg...) do {} while (0)
226#define IP_VS_ERR_BUF(msg...) do {} while (0)
1da177e4
LT
227#define IP_VS_DBG(level, msg...) do {} while (0)
228#define IP_VS_DBG_RL(msg...) do {} while (0)
0d79641a
JA
229#define IP_VS_DBG_PKT(level, af, pp, skb, ofs, msg) do {} while (0)
230#define IP_VS_DBG_RL_PKT(level, af, pp, skb, ofs, msg) do {} while (0)
1da177e4
LT
231#endif
232
233#define IP_VS_BUG() BUG()
1e3e238e 234#define IP_VS_ERR_RL(msg, ...) \
9aada7ac
HE
235 do { \
236 if (net_ratelimit()) \
1e3e238e 237 pr_err(msg, ##__VA_ARGS__); \
9aada7ac 238 } while (0)
1da177e4
LT
239
240#ifdef CONFIG_IP_VS_DEBUG
241#define EnterFunction(level) \
9aada7ac
HE
242 do { \
243 if (level <= ip_vs_get_debug_level()) \
244 printk(KERN_DEBUG \
245 pr_fmt("Enter: %s, %s line %i\n"), \
246 __func__, __FILE__, __LINE__); \
247 } while (0)
248#define LeaveFunction(level) \
249 do { \
250 if (level <= ip_vs_get_debug_level()) \
251 printk(KERN_DEBUG \
252 pr_fmt("Leave: %s, %s line %i\n"), \
253 __func__, __FILE__, __LINE__); \
254 } while (0)
1da177e4
LT
255#else
256#define EnterFunction(level) do {} while (0)
257#define LeaveFunction(level) do {} while (0)
258#endif
259
260#define IP_VS_WAIT_WHILE(expr) while (expr) { cpu_relax(); }
261
262
263/*
264 * The port number of FTP service (in network order).
265 */
f3a7c66b
HH
266#define FTPPORT cpu_to_be16(21)
267#define FTPDATA cpu_to_be16(20)
1da177e4 268
1da177e4
LT
269/*
270 * TCP State Values
271 */
272enum {
273 IP_VS_TCP_S_NONE = 0,
274 IP_VS_TCP_S_ESTABLISHED,
275 IP_VS_TCP_S_SYN_SENT,
276 IP_VS_TCP_S_SYN_RECV,
277 IP_VS_TCP_S_FIN_WAIT,
278 IP_VS_TCP_S_TIME_WAIT,
279 IP_VS_TCP_S_CLOSE,
280 IP_VS_TCP_S_CLOSE_WAIT,
281 IP_VS_TCP_S_LAST_ACK,
282 IP_VS_TCP_S_LISTEN,
283 IP_VS_TCP_S_SYNACK,
284 IP_VS_TCP_S_LAST
285};
286
287/*
288 * UDP State Values
289 */
290enum {
291 IP_VS_UDP_S_NORMAL,
292 IP_VS_UDP_S_LAST,
293};
294
295/*
296 * ICMP State Values
297 */
298enum {
299 IP_VS_ICMP_S_NORMAL,
300 IP_VS_ICMP_S_LAST,
301};
302
2906f66a
VMR
303/*
304 * SCTP State Values
305 */
306enum ip_vs_sctp_states {
307 IP_VS_SCTP_S_NONE,
308 IP_VS_SCTP_S_INIT_CLI,
309 IP_VS_SCTP_S_INIT_SER,
310 IP_VS_SCTP_S_INIT_ACK_CLI,
311 IP_VS_SCTP_S_INIT_ACK_SER,
312 IP_VS_SCTP_S_ECHO_CLI,
313 IP_VS_SCTP_S_ECHO_SER,
314 IP_VS_SCTP_S_ESTABLISHED,
315 IP_VS_SCTP_S_SHUT_CLI,
316 IP_VS_SCTP_S_SHUT_SER,
317 IP_VS_SCTP_S_SHUT_ACK_CLI,
318 IP_VS_SCTP_S_SHUT_ACK_SER,
319 IP_VS_SCTP_S_CLOSED,
320 IP_VS_SCTP_S_LAST
321};
322
1da177e4
LT
323/*
324 * Delta sequence info structure
325 * Each ip_vs_conn has 2 (output AND input seq. changes).
326 * Only used in the VS/NAT.
327 */
328struct ip_vs_seq {
329 __u32 init_seq; /* Add delta from this seq */
330 __u32 delta; /* Delta in sequence numbers */
331 __u32 previous_delta; /* Delta in sequence numbers
332 before last resized pkt */
333};
334
b17fc996
HS
335/*
336 * counters per cpu
337 */
338struct ip_vs_counters {
339 __u32 conns; /* connections scheduled */
340 __u32 inpkts; /* incoming packets */
341 __u32 outpkts; /* outgoing packets */
342 __u64 inbytes; /* incoming bytes */
343 __u64 outbytes; /* outgoing bytes */
344};
345/*
346 * Stats per cpu
347 */
348struct ip_vs_cpu_stats {
349 struct ip_vs_counters ustats;
350 struct u64_stats_sync syncp;
351};
1da177e4
LT
352
353/*
3a14a313 354 * IPVS statistics objects
1da177e4 355 */
3a14a313
SW
356struct ip_vs_estimator {
357 struct list_head list;
358
359 u64 last_inbytes;
360 u64 last_outbytes;
361 u32 last_conns;
362 u32 last_inpkts;
363 u32 last_outpkts;
364
365 u32 cps;
366 u32 inpps;
367 u32 outpps;
368 u32 inbps;
369 u32 outbps;
370};
371
fd2c3ef7 372struct ip_vs_stats {
b17fc996 373 struct ip_vs_stats_user ustats; /* statistics */
e9c0ce23 374 struct ip_vs_estimator est; /* estimator */
b17fc996
HS
375 struct ip_vs_cpu_stats *cpustats; /* per cpu counters */
376 spinlock_t lock; /* spin lock */
55a3d4e1 377 struct ip_vs_stats_user ustats0; /* reset values */
1da177e4
LT
378};
379
14c85021
ACM
380struct dst_entry;
381struct iphdr;
1da177e4
LT
382struct ip_vs_conn;
383struct ip_vs_app;
14c85021 384struct sk_buff;
252c6410 385struct ip_vs_proto_data;
1da177e4
LT
386
387struct ip_vs_protocol {
388 struct ip_vs_protocol *next;
389 char *name;
2ad17def
JA
390 u16 protocol;
391 u16 num_states;
1da177e4 392 int dont_defrag;
1da177e4
LT
393
394 void (*init)(struct ip_vs_protocol *pp);
395
396 void (*exit)(struct ip_vs_protocol *pp);
397
252c6410
HS
398 void (*init_netns)(struct net *net, struct ip_vs_proto_data *pd);
399
400 void (*exit_netns)(struct net *net, struct ip_vs_proto_data *pd);
401
51ef348b 402 int (*conn_schedule)(int af, struct sk_buff *skb,
9330419d 403 struct ip_vs_proto_data *pd,
1da177e4
LT
404 int *verdict, struct ip_vs_conn **cpp);
405
406 struct ip_vs_conn *
51ef348b
JV
407 (*conn_in_get)(int af,
408 const struct sk_buff *skb,
51ef348b 409 const struct ip_vs_iphdr *iph,
1da177e4
LT
410 unsigned int proto_off,
411 int inverse);
412
413 struct ip_vs_conn *
51ef348b
JV
414 (*conn_out_get)(int af,
415 const struct sk_buff *skb,
51ef348b 416 const struct ip_vs_iphdr *iph,
1da177e4
LT
417 unsigned int proto_off,
418 int inverse);
419
3db05fea 420 int (*snat_handler)(struct sk_buff *skb,
1da177e4
LT
421 struct ip_vs_protocol *pp, struct ip_vs_conn *cp);
422
3db05fea 423 int (*dnat_handler)(struct sk_buff *skb,
1da177e4
LT
424 struct ip_vs_protocol *pp, struct ip_vs_conn *cp);
425
51ef348b
JV
426 int (*csum_check)(int af, struct sk_buff *skb,
427 struct ip_vs_protocol *pp);
1da177e4
LT
428
429 const char *(*state_name)(int state);
430
431 int (*state_transition)(struct ip_vs_conn *cp, int direction,
432 const struct sk_buff *skb,
9330419d 433 struct ip_vs_proto_data *pd);
1da177e4 434
ab8a5e84 435 int (*register_app)(struct net *net, struct ip_vs_app *inc);
1da177e4 436
ab8a5e84 437 void (*unregister_app)(struct net *net, struct ip_vs_app *inc);
1da177e4
LT
438
439 int (*app_conn_bind)(struct ip_vs_conn *cp);
440
0d79641a 441 void (*debug_packet)(int af, struct ip_vs_protocol *pp,
1da177e4
LT
442 const struct sk_buff *skb,
443 int offset,
444 const char *msg);
445
9330419d 446 void (*timeout_change)(struct ip_vs_proto_data *pd, int flags);
1da177e4
LT
447};
448
252c6410
HS
449/*
450 * protocol data per netns
451 */
452struct ip_vs_proto_data {
453 struct ip_vs_proto_data *next;
454 struct ip_vs_protocol *pp;
455 int *timeout_table; /* protocol timeout table */
456 atomic_t appcnt; /* counter of proto app incs. */
457 struct tcp_states_t *tcp_state_table;
458};
459
460extern struct ip_vs_protocol *ip_vs_proto_get(unsigned short proto);
461extern struct ip_vs_proto_data *ip_vs_proto_data_get(struct net *net,
462 unsigned short proto);
1da177e4 463
f11017ec 464struct ip_vs_conn_param {
6e67e586 465 struct net *net;
f11017ec
SH
466 const union nf_inet_addr *caddr;
467 const union nf_inet_addr *vaddr;
468 __be16 cport;
469 __be16 vport;
470 __u16 protocol;
471 u16 af;
85999283
SH
472
473 const struct ip_vs_pe *pe;
474 char *pe_data;
475 __u8 pe_data_len;
f11017ec
SH
476};
477
1da177e4
LT
478/*
479 * IP_VS structure allocated for each dynamically scheduled connection
480 */
481struct ip_vs_conn {
731109e7 482 struct hlist_node c_list; /* hashed list heads */
6e67e586
HS
483#ifdef CONFIG_NET_NS
484 struct net *net; /* Name space */
485#endif
1da177e4 486 /* Protocol, addresses and port numbers */
6e67e586
HS
487 u16 af; /* address family */
488 __be16 cport;
489 __be16 vport;
490 __be16 dport;
491 __u32 fwmark; /* Fire wall mark from skb */
492 union nf_inet_addr caddr; /* client address */
493 union nf_inet_addr vaddr; /* virtual address */
494 union nf_inet_addr daddr; /* destination address */
495 volatile __u32 flags; /* status flags */
1da177e4
LT
496 __u16 protocol; /* Which protocol (TCP/UDP) */
497
498 /* counter and timer */
499 atomic_t refcnt; /* reference count */
500 struct timer_list timer; /* Expiration timer */
501 volatile unsigned long timeout; /* timeout */
502
503 /* Flags and state transition */
504 spinlock_t lock; /* lock for state transition */
1da177e4 505 volatile __u16 state; /* state info */
efac5276
RB
506 volatile __u16 old_state; /* old state, to be used for
507 * state transition triggerd
508 * synchronization
509 */
1da177e4
LT
510
511 /* Control members */
512 struct ip_vs_conn *control; /* Master control connection */
513 atomic_t n_control; /* Number of controlled ones */
514 struct ip_vs_dest *dest; /* real server */
515 atomic_t in_pkts; /* incoming packet counter */
516
517 /* packet transmitter for different forwarding methods. If it
518 mangles the packet, it must return NF_DROP or better NF_STOLEN,
519 otherwise this must be changed to a sk_buff **.
fc604767 520 NF_ACCEPT can be returned when destination is local.
1da177e4
LT
521 */
522 int (*packet_xmit)(struct sk_buff *skb, struct ip_vs_conn *cp,
523 struct ip_vs_protocol *pp);
524
525 /* Note: we can group the following members into a structure,
526 in order to save more space, and the following members are
527 only used in VS/NAT anyway */
528 struct ip_vs_app *app; /* bound ip_vs_app object */
529 void *app_data; /* Application private data */
530 struct ip_vs_seq in_seq; /* incoming seq. struct */
531 struct ip_vs_seq out_seq; /* outgoing seq. struct */
85999283 532
e9e5eee8 533 const struct ip_vs_pe *pe;
85999283
SH
534 char *pe_data;
535 __u8 pe_data_len;
1da177e4
LT
536};
537
6e67e586
HS
538/*
539 * To save some memory in conn table when name space is disabled.
540 */
541static inline struct net *ip_vs_conn_net(const struct ip_vs_conn *cp)
542{
543#ifdef CONFIG_NET_NS
544 return cp->net;
545#else
546 return &init_net;
547#endif
548}
549static inline void ip_vs_conn_net_set(struct ip_vs_conn *cp, struct net *net)
550{
551#ifdef CONFIG_NET_NS
552 cp->net = net;
553#endif
554}
555
556static inline int ip_vs_conn_net_eq(const struct ip_vs_conn *cp,
557 struct net *net)
558{
559#ifdef CONFIG_NET_NS
560 return cp->net == net;
561#else
562 return 1;
563#endif
564}
1da177e4 565
c860c6b1
JV
566/*
567 * Extended internal versions of struct ip_vs_service_user and
568 * ip_vs_dest_user for IPv6 support.
569 *
570 * We need these to conveniently pass around service and destination
571 * options, but unfortunately, we also need to keep the old definitions to
572 * maintain userspace backwards compatibility for the setsockopt interface.
573 */
574struct ip_vs_service_user_kern {
575 /* virtual service addresses */
576 u16 af;
577 u16 protocol;
578 union nf_inet_addr addr; /* virtual ip address */
579 u16 port;
580 u32 fwmark; /* firwall mark of service */
581
582 /* virtual service options */
583 char *sched_name;
0d1e71b0 584 char *pe_name;
c860c6b1
JV
585 unsigned flags; /* virtual service flags */
586 unsigned timeout; /* persistent timeout in sec */
587 u32 netmask; /* persistent netmask */
588};
589
590
591struct ip_vs_dest_user_kern {
592 /* destination server address */
593 union nf_inet_addr addr;
594 u16 port;
595
596 /* real server options */
597 unsigned conn_flags; /* connection flags */
598 int weight; /* destination weight */
599
600 /* thresholds for active connections */
601 u32 u_threshold; /* upper threshold */
602 u32 l_threshold; /* lower threshold */
603};
604
605
1da177e4
LT
606/*
607 * The information about the virtual service offered to the net
608 * and the forwarding entries
609 */
610struct ip_vs_service {
611 struct list_head s_list; /* for normal service table */
612 struct list_head f_list; /* for fwmark-based service table */
613 atomic_t refcnt; /* reference counter */
614 atomic_t usecnt; /* use counter */
615
e7ade46a 616 u16 af; /* address family */
1da177e4 617 __u16 protocol; /* which protocol (TCP/UDP) */
e7ade46a 618 union nf_inet_addr addr; /* IP address for virtual service */
014d730d 619 __be16 port; /* port number for the service */
1da177e4
LT
620 __u32 fwmark; /* firewall mark of the service */
621 unsigned flags; /* service status flags */
622 unsigned timeout; /* persistent timeout in ticks */
014d730d 623 __be32 netmask; /* grouping granularity */
fc723250 624 struct net *net;
1da177e4
LT
625
626 struct list_head destinations; /* real server d-linked list */
627 __u32 num_dests; /* number of servers */
628 struct ip_vs_stats stats; /* statistics for the service */
629 struct ip_vs_app *inc; /* bind conns to this app inc */
630
631 /* for scheduling */
632 struct ip_vs_scheduler *scheduler; /* bound scheduler object */
633 rwlock_t sched_lock; /* lock sched_data */
634 void *sched_data; /* scheduler application data */
85999283
SH
635
636 /* alternate persistence engine */
637 struct ip_vs_pe *pe;
1da177e4
LT
638};
639
640
641/*
642 * The real server destination forwarding entry
643 * with ip address, port number, and so on.
644 */
645struct ip_vs_dest {
646 struct list_head n_list; /* for the dests in the service */
647 struct list_head d_list; /* for table with all the dests */
648
e7ade46a 649 u16 af; /* address family */
014d730d 650 __be16 port; /* port number of the server */
f2431e6e 651 union nf_inet_addr addr; /* IP address of the server */
1da177e4
LT
652 volatile unsigned flags; /* dest status flags */
653 atomic_t conn_flags; /* flags to copy to conn */
654 atomic_t weight; /* server weight */
655
656 atomic_t refcnt; /* reference counter */
657 struct ip_vs_stats stats; /* statistics */
658
659 /* connection counters and thresholds */
660 atomic_t activeconns; /* active connections */
661 atomic_t inactconns; /* inactive connections */
662 atomic_t persistconns; /* persistent connections */
663 __u32 u_threshold; /* upper threshold */
664 __u32 l_threshold; /* lower threshold */
665
666 /* for destination cache */
667 spinlock_t dst_lock; /* lock of dst_cache */
668 struct dst_entry *dst_cache; /* destination cache entry */
669 u32 dst_rtos; /* RT_TOS(tos) for dst */
714f095f
HS
670 u32 dst_cookie;
671#ifdef CONFIG_IP_VS_IPV6
672 struct in6_addr dst_saddr;
673#endif
1da177e4
LT
674
675 /* for virtual service */
676 struct ip_vs_service *svc; /* service it belongs to */
677 __u16 protocol; /* which protocol (TCP/UDP) */
014d730d 678 __be16 vport; /* virtual port number */
f2431e6e 679 union nf_inet_addr vaddr; /* virtual IP address */
1da177e4
LT
680 __u32 vfwmark; /* firewall mark of service */
681};
682
683
684/*
685 * The scheduler object
686 */
687struct ip_vs_scheduler {
688 struct list_head n_list; /* d-linked list head */
689 char *name; /* scheduler name */
690 atomic_t refcnt; /* reference counter */
691 struct module *module; /* THIS_MODULE/NULL */
692
693 /* scheduler initializing service */
694 int (*init_service)(struct ip_vs_service *svc);
695 /* scheduling service finish */
696 int (*done_service)(struct ip_vs_service *svc);
697 /* scheduler updating service */
698 int (*update_service)(struct ip_vs_service *svc);
699
700 /* selecting a server from the given service */
701 struct ip_vs_dest* (*schedule)(struct ip_vs_service *svc,
702 const struct sk_buff *skb);
703};
704
85999283
SH
705/* The persistence engine object */
706struct ip_vs_pe {
707 struct list_head n_list; /* d-linked list head */
708 char *name; /* scheduler name */
709 atomic_t refcnt; /* reference counter */
710 struct module *module; /* THIS_MODULE/NULL */
711
712 /* get the connection template, if any */
713 int (*fill_param)(struct ip_vs_conn_param *p, struct sk_buff *skb);
714 bool (*ct_match)(const struct ip_vs_conn_param *p,
715 struct ip_vs_conn *ct);
716 u32 (*hashkey_raw)(const struct ip_vs_conn_param *p, u32 initval,
717 bool inverse);
a3c918ac 718 int (*show_pe_data)(const struct ip_vs_conn *cp, char *buf);
85999283 719};
1da177e4
LT
720
721/*
722 * The application module object (a.k.a. app incarnation)
723 */
fd2c3ef7 724struct ip_vs_app {
1da177e4
LT
725 struct list_head a_list; /* member in app list */
726 int type; /* IP_VS_APP_TYPE_xxx */
727 char *name; /* application module name */
728 __u16 protocol;
729 struct module *module; /* THIS_MODULE/NULL */
730 struct list_head incs_list; /* list of incarnations */
731
732 /* members for application incarnations */
733 struct list_head p_list; /* member in proto app list */
734 struct ip_vs_app *app; /* its real application */
014d730d 735 __be16 port; /* port number in net order */
1da177e4
LT
736 atomic_t usecnt; /* usage counter */
737
8b27b10f
JA
738 /*
739 * output hook: Process packet in inout direction, diff set for TCP.
740 * Return: 0=Error, 1=Payload Not Mangled/Mangled but checksum is ok,
741 * 2=Mangled but checksum was not updated
742 */
1da177e4 743 int (*pkt_out)(struct ip_vs_app *, struct ip_vs_conn *,
3db05fea 744 struct sk_buff *, int *diff);
1da177e4 745
8b27b10f
JA
746 /*
747 * input hook: Process packet in outin direction, diff set for TCP.
748 * Return: 0=Error, 1=Payload Not Mangled/Mangled but checksum is ok,
749 * 2=Mangled but checksum was not updated
750 */
1da177e4 751 int (*pkt_in)(struct ip_vs_app *, struct ip_vs_conn *,
3db05fea 752 struct sk_buff *, int *diff);
1da177e4
LT
753
754 /* ip_vs_app initializer */
755 int (*init_conn)(struct ip_vs_app *, struct ip_vs_conn *);
756
757 /* ip_vs_app finish */
758 int (*done_conn)(struct ip_vs_app *, struct ip_vs_conn *);
759
760
761 /* not used now */
762 int (*bind_conn)(struct ip_vs_app *, struct ip_vs_conn *,
763 struct ip_vs_protocol *);
764
765 void (*unbind_conn)(struct ip_vs_app *, struct ip_vs_conn *);
766
767 int * timeout_table;
768 int * timeouts;
769 int timeouts_size;
770
771 int (*conn_schedule)(struct sk_buff *skb, struct ip_vs_app *app,
772 int *verdict, struct ip_vs_conn **cpp);
773
774 struct ip_vs_conn *
775 (*conn_in_get)(const struct sk_buff *skb, struct ip_vs_app *app,
776 const struct iphdr *iph, unsigned int proto_off,
777 int inverse);
778
779 struct ip_vs_conn *
780 (*conn_out_get)(const struct sk_buff *skb, struct ip_vs_app *app,
781 const struct iphdr *iph, unsigned int proto_off,
782 int inverse);
783
784 int (*state_transition)(struct ip_vs_conn *cp, int direction,
785 const struct sk_buff *skb,
786 struct ip_vs_app *app);
787
788 void (*timeout_change)(struct ip_vs_app *app, int flags);
789};
790
2553d064
JA
791/* IPVS in network namespace */
792struct netns_ipvs {
793 int gen; /* Generation */
794 /*
795 * Hash table: for real service lookups
796 */
797 #define IP_VS_RTAB_BITS 4
798 #define IP_VS_RTAB_SIZE (1 << IP_VS_RTAB_BITS)
799 #define IP_VS_RTAB_MASK (IP_VS_RTAB_SIZE - 1)
800
801 struct list_head rs_table[IP_VS_RTAB_SIZE];
802 /* ip_vs_app */
803 struct list_head app_list;
2553d064
JA
804
805 /* ip_vs_proto */
806 #define IP_VS_PROTO_TAB_SIZE 32 /* must be power of 2 */
807 struct ip_vs_proto_data *proto_data_table[IP_VS_PROTO_TAB_SIZE];
808 /* ip_vs_proto_tcp */
809#ifdef CONFIG_IP_VS_PROTO_TCP
810 #define TCP_APP_TAB_BITS 4
811 #define TCP_APP_TAB_SIZE (1 << TCP_APP_TAB_BITS)
812 #define TCP_APP_TAB_MASK (TCP_APP_TAB_SIZE - 1)
813 struct list_head tcp_apps[TCP_APP_TAB_SIZE];
814 spinlock_t tcp_app_lock;
815#endif
816 /* ip_vs_proto_udp */
817#ifdef CONFIG_IP_VS_PROTO_UDP
818 #define UDP_APP_TAB_BITS 4
819 #define UDP_APP_TAB_SIZE (1 << UDP_APP_TAB_BITS)
820 #define UDP_APP_TAB_MASK (UDP_APP_TAB_SIZE - 1)
821 struct list_head udp_apps[UDP_APP_TAB_SIZE];
822 spinlock_t udp_app_lock;
823#endif
824 /* ip_vs_proto_sctp */
825#ifdef CONFIG_IP_VS_PROTO_SCTP
826 #define SCTP_APP_TAB_BITS 4
827 #define SCTP_APP_TAB_SIZE (1 << SCTP_APP_TAB_BITS)
828 #define SCTP_APP_TAB_MASK (SCTP_APP_TAB_SIZE - 1)
829 /* Hash table for SCTP application incarnations */
830 struct list_head sctp_apps[SCTP_APP_TAB_SIZE];
831 spinlock_t sctp_app_lock;
832#endif
833 /* ip_vs_conn */
834 atomic_t conn_count; /* connection counter */
835
836 /* ip_vs_ctl */
2a0751af 837 struct ip_vs_stats tot_stats; /* Statistics & est. */
2553d064
JA
838
839 int num_services; /* no of virtual services */
f2247fbd
SH
840
841 rwlock_t rs_lock; /* real services table */
842 /* semaphore for IPVS sockopts. And, [gs]etsockopt may sleep. */
843 struct lock_class_key ctl_key; /* ctl_mutex debuging */
844 /* Trash for destinations */
845 struct list_head dest_trash;
846 /* Service counters */
847 atomic_t ftpsvc_counter;
848 atomic_t nullsvc_counter;
849
850#ifdef CONFIG_SYSCTL
2553d064
JA
851 /* 1/rate drop and drop-entry variables */
852 struct delayed_work defense_work; /* Work handler */
853 int drop_rate;
854 int drop_counter;
855 atomic_t dropentry;
856 /* locks in ctl.c */
857 spinlock_t dropentry_lock; /* drop entry handling */
858 spinlock_t droppacket_lock; /* drop packet handling */
859 spinlock_t securetcp_lock; /* state and timeout tables */
2553d064
JA
860
861 /* sys-ctl struct */
862 struct ctl_table_header *sysctl_hdr;
863 struct ctl_table *sysctl_tbl;
f2247fbd
SH
864#endif
865
2553d064
JA
866 /* sysctl variables */
867 int sysctl_amemthresh;
868 int sysctl_am_droprate;
869 int sysctl_drop_entry;
870 int sysctl_drop_packet;
871 int sysctl_secure_tcp;
872#ifdef CONFIG_IP_VS_NFCT
873 int sysctl_conntrack;
874#endif
875 int sysctl_snat_reroute;
876 int sysctl_sync_ver;
877 int sysctl_cache_bypass;
878 int sysctl_expire_nodest_conn;
879 int sysctl_expire_quiescent_template;
880 int sysctl_sync_threshold[2];
881 int sysctl_nat_icmp_send;
882
883 /* ip_vs_lblc */
884 int sysctl_lblc_expiration;
885 struct ctl_table_header *lblc_ctl_header;
886 struct ctl_table *lblc_ctl_table;
887 /* ip_vs_lblcr */
888 int sysctl_lblcr_expiration;
889 struct ctl_table_header *lblcr_ctl_header;
890 struct ctl_table *lblcr_ctl_table;
891 /* ip_vs_est */
892 struct list_head est_list; /* estimator list */
893 spinlock_t est_lock;
894 struct timer_list est_timer; /* Estimation timer */
895 /* ip_vs_sync */
896 struct list_head sync_queue;
897 spinlock_t sync_lock;
898 struct ip_vs_sync_buff *sync_buff;
899 spinlock_t sync_buff_lock;
900 struct sockaddr_in sync_mcast_addr;
901 struct task_struct *master_thread;
902 struct task_struct *backup_thread;
903 int send_mesg_maxlen;
904 int recv_mesg_maxlen;
905 volatile int sync_state;
906 volatile int master_syncid;
907 volatile int backup_syncid;
908 /* multicast interface name */
909 char master_mcast_ifn[IP_VS_IFNAME_MAXLEN];
910 char backup_mcast_ifn[IP_VS_IFNAME_MAXLEN];
911 /* net name space ptr */
912 struct net *net; /* Needed by timer routines */
913};
1da177e4 914
59e0350e
SH
915#define DEFAULT_SYNC_THRESHOLD 3
916#define DEFAULT_SYNC_PERIOD 50
7532e8d4 917#define DEFAULT_SYNC_VER 1
59e0350e
SH
918
919#ifdef CONFIG_SYSCTL
920
921static inline int sysctl_sync_threshold(struct netns_ipvs *ipvs)
922{
923 return ipvs->sysctl_sync_threshold[0];
924}
925
926static inline int sysctl_sync_period(struct netns_ipvs *ipvs)
927{
928 return ipvs->sysctl_sync_threshold[1];
929}
930
7532e8d4
SH
931static inline int sysctl_sync_ver(struct netns_ipvs *ipvs)
932{
933 return ipvs->sysctl_sync_ver;
934}
935
59e0350e
SH
936#else
937
938static inline int sysctl_sync_threshold(struct netns_ipvs *ipvs)
939{
940 return DEFAULT_SYNC_THRESHOLD;
941}
942
943static inline int sysctl_sync_period(struct netns_ipvs *ipvs)
944{
945 return DEFAULT_SYNC_PERIOD;
946}
947
7532e8d4
SH
948static inline int sysctl_sync_ver(struct netns_ipvs *ipvs)
949{
950 return DEFAULT_SYNC_VER;
951}
952
59e0350e
SH
953#endif
954
1da177e4
LT
955/*
956 * IPVS core functions
957 * (from ip_vs_core.c)
958 */
959extern const char *ip_vs_proto_name(unsigned proto);
960extern void ip_vs_init_hash_table(struct list_head *table, int rows);
afdd6140 961#define IP_VS_INIT_HASH_TABLE(t) ip_vs_init_hash_table((t), ARRAY_SIZE((t)))
1da177e4 962
1da177e4
LT
963#define IP_VS_APP_TYPE_FTP 1
964
965/*
966 * ip_vs_conn handling functions
967 * (from ip_vs_conn.c)
968 */
969
1da177e4
LT
970enum {
971 IP_VS_DIR_INPUT = 0,
972 IP_VS_DIR_OUTPUT,
973 IP_VS_DIR_INPUT_ONLY,
974 IP_VS_DIR_LAST,
975};
976
6e67e586 977static inline void ip_vs_conn_fill_param(struct net *net, int af, int protocol,
f11017ec
SH
978 const union nf_inet_addr *caddr,
979 __be16 cport,
980 const union nf_inet_addr *vaddr,
981 __be16 vport,
982 struct ip_vs_conn_param *p)
983{
6e67e586 984 p->net = net;
f11017ec
SH
985 p->af = af;
986 p->protocol = protocol;
987 p->caddr = caddr;
988 p->cport = cport;
989 p->vaddr = vaddr;
990 p->vport = vport;
85999283
SH
991 p->pe = NULL;
992 p->pe_data = NULL;
f11017ec 993}
28364a59 994
f11017ec
SH
995struct ip_vs_conn *ip_vs_conn_in_get(const struct ip_vs_conn_param *p);
996struct ip_vs_conn *ip_vs_ct_in_get(const struct ip_vs_conn_param *p);
28364a59 997
5c0d2374 998struct ip_vs_conn * ip_vs_conn_in_get_proto(int af, const struct sk_buff *skb,
5c0d2374
SH
999 const struct ip_vs_iphdr *iph,
1000 unsigned int proto_off,
1001 int inverse);
1002
f11017ec 1003struct ip_vs_conn *ip_vs_conn_out_get(const struct ip_vs_conn_param *p);
1da177e4 1004
5c0d2374 1005struct ip_vs_conn * ip_vs_conn_out_get_proto(int af, const struct sk_buff *skb,
5c0d2374
SH
1006 const struct ip_vs_iphdr *iph,
1007 unsigned int proto_off,
1008 int inverse);
1009
1da177e4
LT
1010/* put back the conn without restarting its timer */
1011static inline void __ip_vs_conn_put(struct ip_vs_conn *cp)
1012{
1013 atomic_dec(&cp->refcnt);
1014}
1015extern void ip_vs_conn_put(struct ip_vs_conn *cp);
014d730d 1016extern void ip_vs_conn_fill_cport(struct ip_vs_conn *cp, __be16 cport);
1da177e4 1017
f11017ec
SH
1018struct ip_vs_conn *ip_vs_conn_new(const struct ip_vs_conn_param *p,
1019 const union nf_inet_addr *daddr,
1020 __be16 dport, unsigned flags,
0e051e68 1021 struct ip_vs_dest *dest, __u32 fwmark);
1da177e4
LT
1022extern void ip_vs_conn_expire_now(struct ip_vs_conn *cp);
1023
1024extern const char * ip_vs_state_name(__u16 proto, int state);
1025
4a85b96c 1026extern void ip_vs_tcp_conn_listen(struct net *net, struct ip_vs_conn *cp);
1da177e4 1027extern int ip_vs_check_template(struct ip_vs_conn *ct);
f6340ee0 1028extern void ip_vs_random_dropentry(struct net *net);
1da177e4
LT
1029extern int ip_vs_conn_init(void);
1030extern void ip_vs_conn_cleanup(void);
1031
1032static inline void ip_vs_control_del(struct ip_vs_conn *cp)
1033{
1034 struct ip_vs_conn *ctl_cp = cp->control;
1035 if (!ctl_cp) {
cfc78c5a
JV
1036 IP_VS_ERR_BUF("request control DEL for uncontrolled: "
1037 "%s:%d to %s:%d\n",
1038 IP_VS_DBG_ADDR(cp->af, &cp->caddr),
1039 ntohs(cp->cport),
1040 IP_VS_DBG_ADDR(cp->af, &cp->vaddr),
1041 ntohs(cp->vport));
1042
1da177e4
LT
1043 return;
1044 }
1045
cfc78c5a
JV
1046 IP_VS_DBG_BUF(7, "DELeting control for: "
1047 "cp.dst=%s:%d ctl_cp.dst=%s:%d\n",
1048 IP_VS_DBG_ADDR(cp->af, &cp->caddr),
1049 ntohs(cp->cport),
1050 IP_VS_DBG_ADDR(cp->af, &ctl_cp->caddr),
1051 ntohs(ctl_cp->cport));
1da177e4
LT
1052
1053 cp->control = NULL;
1054 if (atomic_read(&ctl_cp->n_control) == 0) {
cfc78c5a
JV
1055 IP_VS_ERR_BUF("BUG control DEL with n=0 : "
1056 "%s:%d to %s:%d\n",
1057 IP_VS_DBG_ADDR(cp->af, &cp->caddr),
1058 ntohs(cp->cport),
1059 IP_VS_DBG_ADDR(cp->af, &cp->vaddr),
1060 ntohs(cp->vport));
1061
1da177e4
LT
1062 return;
1063 }
1064 atomic_dec(&ctl_cp->n_control);
1065}
1066
1067static inline void
1068ip_vs_control_add(struct ip_vs_conn *cp, struct ip_vs_conn *ctl_cp)
1069{
1070 if (cp->control) {
cfc78c5a
JV
1071 IP_VS_ERR_BUF("request control ADD for already controlled: "
1072 "%s:%d to %s:%d\n",
1073 IP_VS_DBG_ADDR(cp->af, &cp->caddr),
1074 ntohs(cp->cport),
1075 IP_VS_DBG_ADDR(cp->af, &cp->vaddr),
1076 ntohs(cp->vport));
1077
1da177e4
LT
1078 ip_vs_control_del(cp);
1079 }
1080
cfc78c5a
JV
1081 IP_VS_DBG_BUF(7, "ADDing control for: "
1082 "cp.dst=%s:%d ctl_cp.dst=%s:%d\n",
1083 IP_VS_DBG_ADDR(cp->af, &cp->caddr),
1084 ntohs(cp->cport),
1085 IP_VS_DBG_ADDR(cp->af, &ctl_cp->caddr),
1086 ntohs(ctl_cp->cport));
1da177e4
LT
1087
1088 cp->control = ctl_cp;
1089 atomic_inc(&ctl_cp->n_control);
1090}
1091
1092
1093/*
1094 * IPVS application functions
1095 * (from ip_vs_app.c)
1096 */
1097#define IP_VS_APP_MAX_PORTS 8
ab8a5e84
HS
1098extern int register_ip_vs_app(struct net *net, struct ip_vs_app *app);
1099extern void unregister_ip_vs_app(struct net *net, struct ip_vs_app *app);
1da177e4
LT
1100extern int ip_vs_bind_app(struct ip_vs_conn *cp, struct ip_vs_protocol *pp);
1101extern void ip_vs_unbind_app(struct ip_vs_conn *cp);
ab8a5e84
HS
1102extern int register_ip_vs_app_inc(struct net *net, struct ip_vs_app *app,
1103 __u16 proto, __u16 port);
1da177e4
LT
1104extern int ip_vs_app_inc_get(struct ip_vs_app *inc);
1105extern void ip_vs_app_inc_put(struct ip_vs_app *inc);
1106
3db05fea
HX
1107extern int ip_vs_app_pkt_out(struct ip_vs_conn *, struct sk_buff *skb);
1108extern int ip_vs_app_pkt_in(struct ip_vs_conn *, struct sk_buff *skb);
1da177e4
LT
1109extern int ip_vs_app_init(void);
1110extern void ip_vs_app_cleanup(void);
1111
8be67a66
SH
1112void ip_vs_bind_pe(struct ip_vs_service *svc, struct ip_vs_pe *pe);
1113void ip_vs_unbind_pe(struct ip_vs_service *svc);
1114int register_ip_vs_pe(struct ip_vs_pe *pe);
1115int unregister_ip_vs_pe(struct ip_vs_pe *pe);
e9e5eee8 1116struct ip_vs_pe *ip_vs_pe_getbyname(const char *name);
fe5e7a1e 1117struct ip_vs_pe *__ip_vs_pe_getbyname(const char *pe_name);
e9e5eee8
SH
1118
1119static inline void ip_vs_pe_get(const struct ip_vs_pe *pe)
1120{
1121 if (pe && pe->module)
1122 __module_get(pe->module);
1123}
1124
1125static inline void ip_vs_pe_put(const struct ip_vs_pe *pe)
1126{
1127 if (pe && pe->module)
1128 module_put(pe->module);
1129}
1da177e4
LT
1130
1131/*
1132 * IPVS protocol functions (from ip_vs_proto.c)
1133 */
1134extern int ip_vs_protocol_init(void);
1135extern void ip_vs_protocol_cleanup(void);
9330419d 1136extern void ip_vs_protocol_timeout_change(struct netns_ipvs *ipvs, int flags);
1da177e4
LT
1137extern int *ip_vs_create_timeout_table(int *table, int size);
1138extern int
36cbd3dc
JE
1139ip_vs_set_state_timeout(int *table, int num, const char *const *names,
1140 const char *name, int to);
1da177e4 1141extern void
0d79641a
JA
1142ip_vs_tcpudp_debug_packet(int af, struct ip_vs_protocol *pp,
1143 const struct sk_buff *skb,
1da177e4
LT
1144 int offset, const char *msg);
1145
1146extern struct ip_vs_protocol ip_vs_protocol_tcp;
1147extern struct ip_vs_protocol ip_vs_protocol_udp;
1148extern struct ip_vs_protocol ip_vs_protocol_icmp;
1149extern struct ip_vs_protocol ip_vs_protocol_esp;
1150extern struct ip_vs_protocol ip_vs_protocol_ah;
2906f66a 1151extern struct ip_vs_protocol ip_vs_protocol_sctp;
1da177e4
LT
1152
1153/*
1154 * Registering/unregistering scheduler functions
1155 * (from ip_vs_sched.c)
1156 */
1157extern int register_ip_vs_scheduler(struct ip_vs_scheduler *scheduler);
1158extern int unregister_ip_vs_scheduler(struct ip_vs_scheduler *scheduler);
1159extern int ip_vs_bind_scheduler(struct ip_vs_service *svc,
1160 struct ip_vs_scheduler *scheduler);
1161extern int ip_vs_unbind_scheduler(struct ip_vs_service *svc);
1162extern struct ip_vs_scheduler *ip_vs_scheduler_get(const char *sched_name);
1163extern void ip_vs_scheduler_put(struct ip_vs_scheduler *scheduler);
1164extern struct ip_vs_conn *
190ecd27 1165ip_vs_schedule(struct ip_vs_service *svc, struct sk_buff *skb,
9330419d 1166 struct ip_vs_proto_data *pd, int *ignored);
1da177e4 1167extern int ip_vs_leave(struct ip_vs_service *svc, struct sk_buff *skb,
9330419d 1168 struct ip_vs_proto_data *pd);
1da177e4 1169
41ac51ee
PS
1170extern void ip_vs_scheduler_err(struct ip_vs_service *svc, const char *msg);
1171
1da177e4
LT
1172
1173/*
1174 * IPVS control data and functions (from ip_vs_ctl.c)
1175 */
1da177e4 1176extern struct ip_vs_stats ip_vs_stats;
5587da55 1177extern const struct ctl_path net_vs_ctl_path[];
b880c1f0 1178extern int sysctl_ip_vs_sync_ver;
1da177e4 1179
f131315f 1180extern void ip_vs_sync_switch_mode(struct net *net, int mode);
1da177e4 1181extern struct ip_vs_service *
fc723250 1182ip_vs_service_get(struct net *net, int af, __u32 fwmark, __u16 protocol,
3c2e0505 1183 const union nf_inet_addr *vaddr, __be16 vport);
1da177e4
LT
1184
1185static inline void ip_vs_service_put(struct ip_vs_service *svc)
1186{
1187 atomic_dec(&svc->usecnt);
1188}
1189
1190extern struct ip_vs_dest *
fc723250 1191ip_vs_lookup_real_service(struct net *net, int af, __u16 protocol,
7937df15
JV
1192 const union nf_inet_addr *daddr, __be16 dport);
1193
1da177e4
LT
1194extern int ip_vs_use_count_inc(void);
1195extern void ip_vs_use_count_dec(void);
1196extern int ip_vs_control_init(void);
1197extern void ip_vs_control_cleanup(void);
1e356f9c 1198extern struct ip_vs_dest *
fc723250
HS
1199ip_vs_find_dest(struct net *net, int af, const union nf_inet_addr *daddr,
1200 __be16 dport, const union nf_inet_addr *vaddr, __be16 vport,
1201 __u16 protocol, __u32 fwmark);
1e356f9c 1202extern struct ip_vs_dest *ip_vs_try_bind_dest(struct ip_vs_conn *cp);
1da177e4
LT
1203
1204
1205/*
1206 * IPVS sync daemon data and function prototypes
1207 * (from ip_vs_sync.c)
1208 */
f131315f
HS
1209extern int start_sync_thread(struct net *net, int state, char *mcast_ifn,
1210 __u8 syncid);
1211extern int stop_sync_thread(struct net *net, int state);
1212extern void ip_vs_sync_conn(struct net *net, struct ip_vs_conn *cp);
61b1ab45
HS
1213extern int ip_vs_sync_init(void);
1214extern void ip_vs_sync_cleanup(void);
1da177e4
LT
1215
1216
1217/*
1218 * IPVS rate estimator prototypes (from ip_vs_est.c)
1219 */
a919cf4b
SW
1220extern int ip_vs_estimator_init(void);
1221extern void ip_vs_estimator_cleanup(void);
6ef757f9
JA
1222extern void ip_vs_start_estimator(struct net *net, struct ip_vs_stats *stats);
1223extern void ip_vs_stop_estimator(struct net *net, struct ip_vs_stats *stats);
1da177e4 1224extern void ip_vs_zero_estimator(struct ip_vs_stats *stats);
ea9f22cc
JA
1225extern void ip_vs_read_estimator(struct ip_vs_stats_user *dst,
1226 struct ip_vs_stats *stats);
1da177e4
LT
1227
1228/*
1229 * Various IPVS packet transmitters (from ip_vs_xmit.c)
1230 */
1231extern int ip_vs_null_xmit
1232(struct sk_buff *skb, struct ip_vs_conn *cp, struct ip_vs_protocol *pp);
1233extern int ip_vs_bypass_xmit
1234(struct sk_buff *skb, struct ip_vs_conn *cp, struct ip_vs_protocol *pp);
1235extern int ip_vs_nat_xmit
1236(struct sk_buff *skb, struct ip_vs_conn *cp, struct ip_vs_protocol *pp);
1237extern int ip_vs_tunnel_xmit
1238(struct sk_buff *skb, struct ip_vs_conn *cp, struct ip_vs_protocol *pp);
1239extern int ip_vs_dr_xmit
1240(struct sk_buff *skb, struct ip_vs_conn *cp, struct ip_vs_protocol *pp);
1241extern int ip_vs_icmp_xmit
1242(struct sk_buff *skb, struct ip_vs_conn *cp, struct ip_vs_protocol *pp, int offset);
1243extern void ip_vs_dst_reset(struct ip_vs_dest *dest);
1244
b3cdd2a7
JV
1245#ifdef CONFIG_IP_VS_IPV6
1246extern int ip_vs_bypass_xmit_v6
1247(struct sk_buff *skb, struct ip_vs_conn *cp, struct ip_vs_protocol *pp);
1248extern int ip_vs_nat_xmit_v6
1249(struct sk_buff *skb, struct ip_vs_conn *cp, struct ip_vs_protocol *pp);
1250extern int ip_vs_tunnel_xmit_v6
1251(struct sk_buff *skb, struct ip_vs_conn *cp, struct ip_vs_protocol *pp);
1252extern int ip_vs_dr_xmit_v6
1253(struct sk_buff *skb, struct ip_vs_conn *cp, struct ip_vs_protocol *pp);
1254extern int ip_vs_icmp_xmit_v6
1255(struct sk_buff *skb, struct ip_vs_conn *cp, struct ip_vs_protocol *pp,
1256 int offset);
1257#endif
1da177e4 1258
3a1bbf18 1259#ifdef CONFIG_SYSCTL
1da177e4
LT
1260/*
1261 * This is a simple mechanism to ignore packets when
1262 * we are loaded. Just set ip_vs_drop_rate to 'n' and
1263 * we start to drop 1/rate of the packets
1264 */
1da177e4 1265
a0840e2e 1266static inline int ip_vs_todrop(struct netns_ipvs *ipvs)
1da177e4 1267{
a0840e2e
HS
1268 if (!ipvs->drop_rate)
1269 return 0;
1270 if (--ipvs->drop_counter > 0)
1271 return 0;
1272 ipvs->drop_counter = ipvs->drop_rate;
1da177e4
LT
1273 return 1;
1274}
3a1bbf18
SH
1275#else
1276static inline int ip_vs_todrop(struct netns_ipvs *ipvs) { return 0; }
1277#endif
1da177e4
LT
1278
1279/*
1280 * ip_vs_fwd_tag returns the forwarding tag of the connection
1281 */
1282#define IP_VS_FWD_METHOD(cp) (cp->flags & IP_VS_CONN_F_FWD_MASK)
1283
732db659 1284static inline char ip_vs_fwd_tag(struct ip_vs_conn *cp)
1da177e4
LT
1285{
1286 char fwd;
1287
1288 switch (IP_VS_FWD_METHOD(cp)) {
1289 case IP_VS_CONN_F_MASQ:
1290 fwd = 'M'; break;
1291 case IP_VS_CONN_F_LOCALNODE:
1292 fwd = 'L'; break;
1293 case IP_VS_CONN_F_TUNNEL:
1294 fwd = 'T'; break;
1295 case IP_VS_CONN_F_DROUTE:
1296 fwd = 'R'; break;
1297 case IP_VS_CONN_F_BYPASS:
1298 fwd = 'B'; break;
1299 default:
1300 fwd = '?'; break;
1301 }
1302 return fwd;
1303}
1304
1da177e4 1305extern void ip_vs_nat_icmp(struct sk_buff *skb, struct ip_vs_protocol *pp,
b3cdd2a7
JV
1306 struct ip_vs_conn *cp, int dir);
1307
1308#ifdef CONFIG_IP_VS_IPV6
1309extern void ip_vs_nat_icmp_v6(struct sk_buff *skb, struct ip_vs_protocol *pp,
1310 struct ip_vs_conn *cp, int dir);
1311#endif
1da177e4 1312
b1550f22 1313extern __sum16 ip_vs_checksum_complete(struct sk_buff *skb, int offset);
1da177e4 1314
f9214b26 1315static inline __wsum ip_vs_check_diff4(__be32 old, __be32 new, __wsum oldsum)
1da177e4 1316{
f9214b26 1317 __be32 diff[2] = { ~old, new };
1da177e4 1318
07f0757a 1319 return csum_partial(diff, sizeof(diff), oldsum);
f9214b26
AV
1320}
1321
0bbdd42b
JV
1322#ifdef CONFIG_IP_VS_IPV6
1323static inline __wsum ip_vs_check_diff16(const __be32 *old, const __be32 *new,
1324 __wsum oldsum)
1325{
1326 __be32 diff[8] = { ~old[3], ~old[2], ~old[1], ~old[0],
1327 new[3], new[2], new[1], new[0] };
1328
07f0757a 1329 return csum_partial(diff, sizeof(diff), oldsum);
0bbdd42b
JV
1330}
1331#endif
1332
f9214b26
AV
1333static inline __wsum ip_vs_check_diff2(__be16 old, __be16 new, __wsum oldsum)
1334{
1335 __be16 diff[2] = { ~old, new };
1336
07f0757a 1337 return csum_partial(diff, sizeof(diff), oldsum);
1da177e4
LT
1338}
1339
cf356d69
JA
1340/*
1341 * Forget current conntrack (unconfirmed) and attach notrack entry
1342 */
1343static inline void ip_vs_notrack(struct sk_buff *skb)
1344{
1345#if defined(CONFIG_NF_CONNTRACK) || defined(CONFIG_NF_CONNTRACK_MODULE)
1346 enum ip_conntrack_info ctinfo;
06b69390 1347 struct nf_conn *ct = nf_ct_get(skb, &ctinfo);
cf356d69
JA
1348
1349 if (!ct || !nf_ct_is_untracked(ct)) {
1350 nf_reset(skb);
1351 skb->nfct = &nf_ct_untracked_get()->ct_general;
1352 skb->nfctinfo = IP_CT_NEW;
1353 nf_conntrack_get(skb->nfct);
1354 }
1355#endif
1356}
1357
f4bc17cd
JA
1358#ifdef CONFIG_IP_VS_NFCT
1359/*
1360 * Netfilter connection tracking
1361 * (from ip_vs_nfct.c)
1362 */
a0840e2e 1363static inline int ip_vs_conntrack_enabled(struct netns_ipvs *ipvs)
f4bc17cd 1364{
a4e2f5a7 1365#ifdef CONFIG_SYSCTL
a0840e2e 1366 return ipvs->sysctl_conntrack;
a4e2f5a7
SH
1367#else
1368 return 0;
1369#endif
f4bc17cd
JA
1370}
1371
6523ce15
JA
1372extern void ip_vs_update_conntrack(struct sk_buff *skb, struct ip_vs_conn *cp,
1373 int outin);
f4bc17cd
JA
1374extern int ip_vs_confirm_conntrack(struct sk_buff *skb, struct ip_vs_conn *cp);
1375extern void ip_vs_nfct_expect_related(struct sk_buff *skb, struct nf_conn *ct,
1376 struct ip_vs_conn *cp, u_int8_t proto,
1377 const __be16 port, int from_rs);
1378extern void ip_vs_conn_drop_conntrack(struct ip_vs_conn *cp);
1379
1380#else
1381
a0840e2e 1382static inline int ip_vs_conntrack_enabled(struct netns_ipvs *ipvs)
f4bc17cd
JA
1383{
1384 return 0;
1385}
1386
1387static inline void ip_vs_update_conntrack(struct sk_buff *skb,
1388 struct ip_vs_conn *cp, int outin)
1389{
1390}
1391
1392static inline int ip_vs_confirm_conntrack(struct sk_buff *skb,
1393 struct ip_vs_conn *cp)
1394{
1395 return NF_ACCEPT;
1396}
1397
1398static inline void ip_vs_conn_drop_conntrack(struct ip_vs_conn *cp)
1399{
1400}
1401/* CONFIG_IP_VS_NFCT */
1402#endif
6523ce15 1403
b552f7e3
CG
1404static inline unsigned int
1405ip_vs_dest_conn_overhead(struct ip_vs_dest *dest)
1406{
1407 /*
1408 * We think the overhead of processing active connections is 256
1409 * times higher than that of inactive connections in average. (This
1410 * 256 times might not be accurate, we will change it later) We
1411 * use the following formula to estimate the overhead now:
1412 * dest->activeconns*256 + dest->inactconns
1413 */
1414 return (atomic_read(&dest->activeconns) << 8) +
1415 atomic_read(&dest->inactconns);
1416}
1417
1da177e4
LT
1418#endif /* __KERNEL__ */
1419
bc4768eb 1420#endif /* _NET_IP_VS_H */