Commit | Line | Data |
---|---|---|
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 | */ | |
36 | static 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 | 44 | static 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 | 72 | static 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 | |
95 | * private not as a struct i.e. seq_file_net cant be used. | |
96 | */ | |
97 | static 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 */ | |
107 | extern int ip_vs_conn_tab_size; | |
108 | ||
109 | ||
64aae3cb JV |
110 | struct ip_vs_iphdr { |
111 | int len; | |
112 | __u8 protocol; | |
113 | union nf_inet_addr saddr; | |
114 | union nf_inet_addr daddr; | |
115 | }; | |
116 | ||
117 | static inline void | |
118 | ip_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 | ||
138 | static 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 | ||
149 | static 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 | 162 | extern int ip_vs_get_debug_level(void); |
c842a3ad JV |
163 | |
164 | static 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 | */ | |
272 | enum { | |
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 | */ | |
290 | enum { | |
291 | IP_VS_UDP_S_NORMAL, | |
292 | IP_VS_UDP_S_LAST, | |
293 | }; | |
294 | ||
295 | /* | |
296 | * ICMP State Values | |
297 | */ | |
298 | enum { | |
299 | IP_VS_ICMP_S_NORMAL, | |
300 | IP_VS_ICMP_S_LAST, | |
301 | }; | |
302 | ||
2906f66a VMR |
303 | /* |
304 | * SCTP State Values | |
305 | */ | |
306 | enum 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 | */ | |
328 | struct 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 | */ | |
338 | struct 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 | */ | |
348 | struct 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 |
356 | struct 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 | 372 | struct 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 */ | |
1da177e4 LT |
377 | }; |
378 | ||
b17fc996 HS |
379 | /* |
380 | * Helper Macros for per cpu | |
381 | * ipvs->tot_stats->ustats.count | |
382 | */ | |
383 | #define IPVS_STAT_INC(ipvs, count) \ | |
384 | __this_cpu_inc((ipvs)->ustats->count) | |
385 | ||
386 | #define IPVS_STAT_ADD(ipvs, count, value) \ | |
387 | do {\ | |
388 | write_seqcount_begin(per_cpu_ptr((ipvs)->ustats_seq, \ | |
389 | raw_smp_processor_id())); \ | |
390 | __this_cpu_add((ipvs)->ustats->count, value); \ | |
391 | write_seqcount_end(per_cpu_ptr((ipvs)->ustats_seq, \ | |
392 | raw_smp_processor_id())); \ | |
393 | } while (0) | |
394 | ||
14c85021 ACM |
395 | struct dst_entry; |
396 | struct iphdr; | |
1da177e4 LT |
397 | struct ip_vs_conn; |
398 | struct ip_vs_app; | |
14c85021 | 399 | struct sk_buff; |
252c6410 | 400 | struct ip_vs_proto_data; |
1da177e4 LT |
401 | |
402 | struct ip_vs_protocol { | |
403 | struct ip_vs_protocol *next; | |
404 | char *name; | |
2ad17def JA |
405 | u16 protocol; |
406 | u16 num_states; | |
1da177e4 | 407 | int dont_defrag; |
1da177e4 LT |
408 | |
409 | void (*init)(struct ip_vs_protocol *pp); | |
410 | ||
411 | void (*exit)(struct ip_vs_protocol *pp); | |
412 | ||
252c6410 HS |
413 | void (*init_netns)(struct net *net, struct ip_vs_proto_data *pd); |
414 | ||
415 | void (*exit_netns)(struct net *net, struct ip_vs_proto_data *pd); | |
416 | ||
51ef348b | 417 | int (*conn_schedule)(int af, struct sk_buff *skb, |
9330419d | 418 | struct ip_vs_proto_data *pd, |
1da177e4 LT |
419 | int *verdict, struct ip_vs_conn **cpp); |
420 | ||
421 | struct ip_vs_conn * | |
51ef348b JV |
422 | (*conn_in_get)(int af, |
423 | const struct sk_buff *skb, | |
51ef348b | 424 | const struct ip_vs_iphdr *iph, |
1da177e4 LT |
425 | unsigned int proto_off, |
426 | int inverse); | |
427 | ||
428 | struct ip_vs_conn * | |
51ef348b JV |
429 | (*conn_out_get)(int af, |
430 | const struct sk_buff *skb, | |
51ef348b | 431 | const struct ip_vs_iphdr *iph, |
1da177e4 LT |
432 | unsigned int proto_off, |
433 | int inverse); | |
434 | ||
3db05fea | 435 | int (*snat_handler)(struct sk_buff *skb, |
1da177e4 LT |
436 | struct ip_vs_protocol *pp, struct ip_vs_conn *cp); |
437 | ||
3db05fea | 438 | int (*dnat_handler)(struct sk_buff *skb, |
1da177e4 LT |
439 | struct ip_vs_protocol *pp, struct ip_vs_conn *cp); |
440 | ||
51ef348b JV |
441 | int (*csum_check)(int af, struct sk_buff *skb, |
442 | struct ip_vs_protocol *pp); | |
1da177e4 LT |
443 | |
444 | const char *(*state_name)(int state); | |
445 | ||
446 | int (*state_transition)(struct ip_vs_conn *cp, int direction, | |
447 | const struct sk_buff *skb, | |
9330419d | 448 | struct ip_vs_proto_data *pd); |
1da177e4 | 449 | |
ab8a5e84 | 450 | int (*register_app)(struct net *net, struct ip_vs_app *inc); |
1da177e4 | 451 | |
ab8a5e84 | 452 | void (*unregister_app)(struct net *net, struct ip_vs_app *inc); |
1da177e4 LT |
453 | |
454 | int (*app_conn_bind)(struct ip_vs_conn *cp); | |
455 | ||
0d79641a | 456 | void (*debug_packet)(int af, struct ip_vs_protocol *pp, |
1da177e4 LT |
457 | const struct sk_buff *skb, |
458 | int offset, | |
459 | const char *msg); | |
460 | ||
9330419d | 461 | void (*timeout_change)(struct ip_vs_proto_data *pd, int flags); |
1da177e4 LT |
462 | }; |
463 | ||
252c6410 HS |
464 | /* |
465 | * protocol data per netns | |
466 | */ | |
467 | struct ip_vs_proto_data { | |
468 | struct ip_vs_proto_data *next; | |
469 | struct ip_vs_protocol *pp; | |
470 | int *timeout_table; /* protocol timeout table */ | |
471 | atomic_t appcnt; /* counter of proto app incs. */ | |
472 | struct tcp_states_t *tcp_state_table; | |
473 | }; | |
474 | ||
475 | extern struct ip_vs_protocol *ip_vs_proto_get(unsigned short proto); | |
476 | extern struct ip_vs_proto_data *ip_vs_proto_data_get(struct net *net, | |
477 | unsigned short proto); | |
1da177e4 | 478 | |
f11017ec | 479 | struct ip_vs_conn_param { |
6e67e586 | 480 | struct net *net; |
f11017ec SH |
481 | const union nf_inet_addr *caddr; |
482 | const union nf_inet_addr *vaddr; | |
483 | __be16 cport; | |
484 | __be16 vport; | |
485 | __u16 protocol; | |
486 | u16 af; | |
85999283 SH |
487 | |
488 | const struct ip_vs_pe *pe; | |
489 | char *pe_data; | |
490 | __u8 pe_data_len; | |
f11017ec SH |
491 | }; |
492 | ||
1da177e4 LT |
493 | /* |
494 | * IP_VS structure allocated for each dynamically scheduled connection | |
495 | */ | |
496 | struct ip_vs_conn { | |
731109e7 | 497 | struct hlist_node c_list; /* hashed list heads */ |
6e67e586 HS |
498 | #ifdef CONFIG_NET_NS |
499 | struct net *net; /* Name space */ | |
500 | #endif | |
1da177e4 | 501 | /* Protocol, addresses and port numbers */ |
6e67e586 HS |
502 | u16 af; /* address family */ |
503 | __be16 cport; | |
504 | __be16 vport; | |
505 | __be16 dport; | |
506 | __u32 fwmark; /* Fire wall mark from skb */ | |
507 | union nf_inet_addr caddr; /* client address */ | |
508 | union nf_inet_addr vaddr; /* virtual address */ | |
509 | union nf_inet_addr daddr; /* destination address */ | |
510 | volatile __u32 flags; /* status flags */ | |
1da177e4 LT |
511 | __u16 protocol; /* Which protocol (TCP/UDP) */ |
512 | ||
513 | /* counter and timer */ | |
514 | atomic_t refcnt; /* reference count */ | |
515 | struct timer_list timer; /* Expiration timer */ | |
516 | volatile unsigned long timeout; /* timeout */ | |
517 | ||
518 | /* Flags and state transition */ | |
519 | spinlock_t lock; /* lock for state transition */ | |
1da177e4 | 520 | volatile __u16 state; /* state info */ |
efac5276 RB |
521 | volatile __u16 old_state; /* old state, to be used for |
522 | * state transition triggerd | |
523 | * synchronization | |
524 | */ | |
1da177e4 LT |
525 | |
526 | /* Control members */ | |
527 | struct ip_vs_conn *control; /* Master control connection */ | |
528 | atomic_t n_control; /* Number of controlled ones */ | |
529 | struct ip_vs_dest *dest; /* real server */ | |
530 | atomic_t in_pkts; /* incoming packet counter */ | |
531 | ||
532 | /* packet transmitter for different forwarding methods. If it | |
533 | mangles the packet, it must return NF_DROP or better NF_STOLEN, | |
534 | otherwise this must be changed to a sk_buff **. | |
fc604767 | 535 | NF_ACCEPT can be returned when destination is local. |
1da177e4 LT |
536 | */ |
537 | int (*packet_xmit)(struct sk_buff *skb, struct ip_vs_conn *cp, | |
538 | struct ip_vs_protocol *pp); | |
539 | ||
540 | /* Note: we can group the following members into a structure, | |
541 | in order to save more space, and the following members are | |
542 | only used in VS/NAT anyway */ | |
543 | struct ip_vs_app *app; /* bound ip_vs_app object */ | |
544 | void *app_data; /* Application private data */ | |
545 | struct ip_vs_seq in_seq; /* incoming seq. struct */ | |
546 | struct ip_vs_seq out_seq; /* outgoing seq. struct */ | |
85999283 | 547 | |
e9e5eee8 | 548 | const struct ip_vs_pe *pe; |
85999283 SH |
549 | char *pe_data; |
550 | __u8 pe_data_len; | |
1da177e4 LT |
551 | }; |
552 | ||
6e67e586 HS |
553 | /* |
554 | * To save some memory in conn table when name space is disabled. | |
555 | */ | |
556 | static inline struct net *ip_vs_conn_net(const struct ip_vs_conn *cp) | |
557 | { | |
558 | #ifdef CONFIG_NET_NS | |
559 | return cp->net; | |
560 | #else | |
561 | return &init_net; | |
562 | #endif | |
563 | } | |
564 | static inline void ip_vs_conn_net_set(struct ip_vs_conn *cp, struct net *net) | |
565 | { | |
566 | #ifdef CONFIG_NET_NS | |
567 | cp->net = net; | |
568 | #endif | |
569 | } | |
570 | ||
571 | static inline int ip_vs_conn_net_eq(const struct ip_vs_conn *cp, | |
572 | struct net *net) | |
573 | { | |
574 | #ifdef CONFIG_NET_NS | |
575 | return cp->net == net; | |
576 | #else | |
577 | return 1; | |
578 | #endif | |
579 | } | |
1da177e4 | 580 | |
c860c6b1 JV |
581 | /* |
582 | * Extended internal versions of struct ip_vs_service_user and | |
583 | * ip_vs_dest_user for IPv6 support. | |
584 | * | |
585 | * We need these to conveniently pass around service and destination | |
586 | * options, but unfortunately, we also need to keep the old definitions to | |
587 | * maintain userspace backwards compatibility for the setsockopt interface. | |
588 | */ | |
589 | struct ip_vs_service_user_kern { | |
590 | /* virtual service addresses */ | |
591 | u16 af; | |
592 | u16 protocol; | |
593 | union nf_inet_addr addr; /* virtual ip address */ | |
594 | u16 port; | |
595 | u32 fwmark; /* firwall mark of service */ | |
596 | ||
597 | /* virtual service options */ | |
598 | char *sched_name; | |
0d1e71b0 | 599 | char *pe_name; |
c860c6b1 JV |
600 | unsigned flags; /* virtual service flags */ |
601 | unsigned timeout; /* persistent timeout in sec */ | |
602 | u32 netmask; /* persistent netmask */ | |
603 | }; | |
604 | ||
605 | ||
606 | struct ip_vs_dest_user_kern { | |
607 | /* destination server address */ | |
608 | union nf_inet_addr addr; | |
609 | u16 port; | |
610 | ||
611 | /* real server options */ | |
612 | unsigned conn_flags; /* connection flags */ | |
613 | int weight; /* destination weight */ | |
614 | ||
615 | /* thresholds for active connections */ | |
616 | u32 u_threshold; /* upper threshold */ | |
617 | u32 l_threshold; /* lower threshold */ | |
618 | }; | |
619 | ||
620 | ||
1da177e4 LT |
621 | /* |
622 | * The information about the virtual service offered to the net | |
623 | * and the forwarding entries | |
624 | */ | |
625 | struct ip_vs_service { | |
626 | struct list_head s_list; /* for normal service table */ | |
627 | struct list_head f_list; /* for fwmark-based service table */ | |
628 | atomic_t refcnt; /* reference counter */ | |
629 | atomic_t usecnt; /* use counter */ | |
630 | ||
e7ade46a | 631 | u16 af; /* address family */ |
1da177e4 | 632 | __u16 protocol; /* which protocol (TCP/UDP) */ |
e7ade46a | 633 | union nf_inet_addr addr; /* IP address for virtual service */ |
014d730d | 634 | __be16 port; /* port number for the service */ |
1da177e4 LT |
635 | __u32 fwmark; /* firewall mark of the service */ |
636 | unsigned flags; /* service status flags */ | |
637 | unsigned timeout; /* persistent timeout in ticks */ | |
014d730d | 638 | __be32 netmask; /* grouping granularity */ |
fc723250 | 639 | struct net *net; |
1da177e4 LT |
640 | |
641 | struct list_head destinations; /* real server d-linked list */ | |
642 | __u32 num_dests; /* number of servers */ | |
643 | struct ip_vs_stats stats; /* statistics for the service */ | |
644 | struct ip_vs_app *inc; /* bind conns to this app inc */ | |
645 | ||
646 | /* for scheduling */ | |
647 | struct ip_vs_scheduler *scheduler; /* bound scheduler object */ | |
648 | rwlock_t sched_lock; /* lock sched_data */ | |
649 | void *sched_data; /* scheduler application data */ | |
85999283 SH |
650 | |
651 | /* alternate persistence engine */ | |
652 | struct ip_vs_pe *pe; | |
1da177e4 LT |
653 | }; |
654 | ||
655 | ||
656 | /* | |
657 | * The real server destination forwarding entry | |
658 | * with ip address, port number, and so on. | |
659 | */ | |
660 | struct ip_vs_dest { | |
661 | struct list_head n_list; /* for the dests in the service */ | |
662 | struct list_head d_list; /* for table with all the dests */ | |
663 | ||
e7ade46a | 664 | u16 af; /* address family */ |
014d730d | 665 | __be16 port; /* port number of the server */ |
f2431e6e | 666 | union nf_inet_addr addr; /* IP address of the server */ |
1da177e4 LT |
667 | volatile unsigned flags; /* dest status flags */ |
668 | atomic_t conn_flags; /* flags to copy to conn */ | |
669 | atomic_t weight; /* server weight */ | |
670 | ||
671 | atomic_t refcnt; /* reference counter */ | |
672 | struct ip_vs_stats stats; /* statistics */ | |
673 | ||
674 | /* connection counters and thresholds */ | |
675 | atomic_t activeconns; /* active connections */ | |
676 | atomic_t inactconns; /* inactive connections */ | |
677 | atomic_t persistconns; /* persistent connections */ | |
678 | __u32 u_threshold; /* upper threshold */ | |
679 | __u32 l_threshold; /* lower threshold */ | |
680 | ||
681 | /* for destination cache */ | |
682 | spinlock_t dst_lock; /* lock of dst_cache */ | |
683 | struct dst_entry *dst_cache; /* destination cache entry */ | |
684 | u32 dst_rtos; /* RT_TOS(tos) for dst */ | |
714f095f HS |
685 | u32 dst_cookie; |
686 | #ifdef CONFIG_IP_VS_IPV6 | |
687 | struct in6_addr dst_saddr; | |
688 | #endif | |
1da177e4 LT |
689 | |
690 | /* for virtual service */ | |
691 | struct ip_vs_service *svc; /* service it belongs to */ | |
692 | __u16 protocol; /* which protocol (TCP/UDP) */ | |
014d730d | 693 | __be16 vport; /* virtual port number */ |
f2431e6e | 694 | union nf_inet_addr vaddr; /* virtual IP address */ |
1da177e4 LT |
695 | __u32 vfwmark; /* firewall mark of service */ |
696 | }; | |
697 | ||
698 | ||
699 | /* | |
700 | * The scheduler object | |
701 | */ | |
702 | struct ip_vs_scheduler { | |
703 | struct list_head n_list; /* d-linked list head */ | |
704 | char *name; /* scheduler name */ | |
705 | atomic_t refcnt; /* reference counter */ | |
706 | struct module *module; /* THIS_MODULE/NULL */ | |
707 | ||
708 | /* scheduler initializing service */ | |
709 | int (*init_service)(struct ip_vs_service *svc); | |
710 | /* scheduling service finish */ | |
711 | int (*done_service)(struct ip_vs_service *svc); | |
712 | /* scheduler updating service */ | |
713 | int (*update_service)(struct ip_vs_service *svc); | |
714 | ||
715 | /* selecting a server from the given service */ | |
716 | struct ip_vs_dest* (*schedule)(struct ip_vs_service *svc, | |
717 | const struct sk_buff *skb); | |
718 | }; | |
719 | ||
85999283 SH |
720 | /* The persistence engine object */ |
721 | struct ip_vs_pe { | |
722 | struct list_head n_list; /* d-linked list head */ | |
723 | char *name; /* scheduler name */ | |
724 | atomic_t refcnt; /* reference counter */ | |
725 | struct module *module; /* THIS_MODULE/NULL */ | |
726 | ||
727 | /* get the connection template, if any */ | |
728 | int (*fill_param)(struct ip_vs_conn_param *p, struct sk_buff *skb); | |
729 | bool (*ct_match)(const struct ip_vs_conn_param *p, | |
730 | struct ip_vs_conn *ct); | |
731 | u32 (*hashkey_raw)(const struct ip_vs_conn_param *p, u32 initval, | |
732 | bool inverse); | |
a3c918ac | 733 | int (*show_pe_data)(const struct ip_vs_conn *cp, char *buf); |
85999283 | 734 | }; |
1da177e4 LT |
735 | |
736 | /* | |
737 | * The application module object (a.k.a. app incarnation) | |
738 | */ | |
fd2c3ef7 | 739 | struct ip_vs_app { |
1da177e4 LT |
740 | struct list_head a_list; /* member in app list */ |
741 | int type; /* IP_VS_APP_TYPE_xxx */ | |
742 | char *name; /* application module name */ | |
743 | __u16 protocol; | |
744 | struct module *module; /* THIS_MODULE/NULL */ | |
745 | struct list_head incs_list; /* list of incarnations */ | |
746 | ||
747 | /* members for application incarnations */ | |
748 | struct list_head p_list; /* member in proto app list */ | |
749 | struct ip_vs_app *app; /* its real application */ | |
014d730d | 750 | __be16 port; /* port number in net order */ |
1da177e4 LT |
751 | atomic_t usecnt; /* usage counter */ |
752 | ||
8b27b10f JA |
753 | /* |
754 | * output hook: Process packet in inout direction, diff set for TCP. | |
755 | * Return: 0=Error, 1=Payload Not Mangled/Mangled but checksum is ok, | |
756 | * 2=Mangled but checksum was not updated | |
757 | */ | |
1da177e4 | 758 | int (*pkt_out)(struct ip_vs_app *, struct ip_vs_conn *, |
3db05fea | 759 | struct sk_buff *, int *diff); |
1da177e4 | 760 | |
8b27b10f JA |
761 | /* |
762 | * input hook: Process packet in outin direction, diff set for TCP. | |
763 | * Return: 0=Error, 1=Payload Not Mangled/Mangled but checksum is ok, | |
764 | * 2=Mangled but checksum was not updated | |
765 | */ | |
1da177e4 | 766 | int (*pkt_in)(struct ip_vs_app *, struct ip_vs_conn *, |
3db05fea | 767 | struct sk_buff *, int *diff); |
1da177e4 LT |
768 | |
769 | /* ip_vs_app initializer */ | |
770 | int (*init_conn)(struct ip_vs_app *, struct ip_vs_conn *); | |
771 | ||
772 | /* ip_vs_app finish */ | |
773 | int (*done_conn)(struct ip_vs_app *, struct ip_vs_conn *); | |
774 | ||
775 | ||
776 | /* not used now */ | |
777 | int (*bind_conn)(struct ip_vs_app *, struct ip_vs_conn *, | |
778 | struct ip_vs_protocol *); | |
779 | ||
780 | void (*unbind_conn)(struct ip_vs_app *, struct ip_vs_conn *); | |
781 | ||
782 | int * timeout_table; | |
783 | int * timeouts; | |
784 | int timeouts_size; | |
785 | ||
786 | int (*conn_schedule)(struct sk_buff *skb, struct ip_vs_app *app, | |
787 | int *verdict, struct ip_vs_conn **cpp); | |
788 | ||
789 | struct ip_vs_conn * | |
790 | (*conn_in_get)(const struct sk_buff *skb, struct ip_vs_app *app, | |
791 | const struct iphdr *iph, unsigned int proto_off, | |
792 | int inverse); | |
793 | ||
794 | struct ip_vs_conn * | |
795 | (*conn_out_get)(const struct sk_buff *skb, struct ip_vs_app *app, | |
796 | const struct iphdr *iph, unsigned int proto_off, | |
797 | int inverse); | |
798 | ||
799 | int (*state_transition)(struct ip_vs_conn *cp, int direction, | |
800 | const struct sk_buff *skb, | |
801 | struct ip_vs_app *app); | |
802 | ||
803 | void (*timeout_change)(struct ip_vs_app *app, int flags); | |
804 | }; | |
805 | ||
806 | ||
807 | /* | |
808 | * IPVS core functions | |
809 | * (from ip_vs_core.c) | |
810 | */ | |
811 | extern const char *ip_vs_proto_name(unsigned proto); | |
812 | extern void ip_vs_init_hash_table(struct list_head *table, int rows); | |
afdd6140 | 813 | #define IP_VS_INIT_HASH_TABLE(t) ip_vs_init_hash_table((t), ARRAY_SIZE((t))) |
1da177e4 | 814 | |
1da177e4 LT |
815 | #define IP_VS_APP_TYPE_FTP 1 |
816 | ||
817 | /* | |
818 | * ip_vs_conn handling functions | |
819 | * (from ip_vs_conn.c) | |
820 | */ | |
821 | ||
1da177e4 LT |
822 | enum { |
823 | IP_VS_DIR_INPUT = 0, | |
824 | IP_VS_DIR_OUTPUT, | |
825 | IP_VS_DIR_INPUT_ONLY, | |
826 | IP_VS_DIR_LAST, | |
827 | }; | |
828 | ||
6e67e586 | 829 | static inline void ip_vs_conn_fill_param(struct net *net, int af, int protocol, |
f11017ec SH |
830 | const union nf_inet_addr *caddr, |
831 | __be16 cport, | |
832 | const union nf_inet_addr *vaddr, | |
833 | __be16 vport, | |
834 | struct ip_vs_conn_param *p) | |
835 | { | |
6e67e586 | 836 | p->net = net; |
f11017ec SH |
837 | p->af = af; |
838 | p->protocol = protocol; | |
839 | p->caddr = caddr; | |
840 | p->cport = cport; | |
841 | p->vaddr = vaddr; | |
842 | p->vport = vport; | |
85999283 SH |
843 | p->pe = NULL; |
844 | p->pe_data = NULL; | |
f11017ec | 845 | } |
28364a59 | 846 | |
f11017ec SH |
847 | struct ip_vs_conn *ip_vs_conn_in_get(const struct ip_vs_conn_param *p); |
848 | struct ip_vs_conn *ip_vs_ct_in_get(const struct ip_vs_conn_param *p); | |
28364a59 | 849 | |
5c0d2374 | 850 | struct ip_vs_conn * ip_vs_conn_in_get_proto(int af, const struct sk_buff *skb, |
5c0d2374 SH |
851 | const struct ip_vs_iphdr *iph, |
852 | unsigned int proto_off, | |
853 | int inverse); | |
854 | ||
f11017ec | 855 | struct ip_vs_conn *ip_vs_conn_out_get(const struct ip_vs_conn_param *p); |
1da177e4 | 856 | |
5c0d2374 | 857 | struct ip_vs_conn * ip_vs_conn_out_get_proto(int af, const struct sk_buff *skb, |
5c0d2374 SH |
858 | const struct ip_vs_iphdr *iph, |
859 | unsigned int proto_off, | |
860 | int inverse); | |
861 | ||
1da177e4 LT |
862 | /* put back the conn without restarting its timer */ |
863 | static inline void __ip_vs_conn_put(struct ip_vs_conn *cp) | |
864 | { | |
865 | atomic_dec(&cp->refcnt); | |
866 | } | |
867 | extern void ip_vs_conn_put(struct ip_vs_conn *cp); | |
014d730d | 868 | extern void ip_vs_conn_fill_cport(struct ip_vs_conn *cp, __be16 cport); |
1da177e4 | 869 | |
f11017ec SH |
870 | struct ip_vs_conn *ip_vs_conn_new(const struct ip_vs_conn_param *p, |
871 | const union nf_inet_addr *daddr, | |
872 | __be16 dport, unsigned flags, | |
0e051e68 | 873 | struct ip_vs_dest *dest, __u32 fwmark); |
1da177e4 LT |
874 | extern void ip_vs_conn_expire_now(struct ip_vs_conn *cp); |
875 | ||
876 | extern const char * ip_vs_state_name(__u16 proto, int state); | |
877 | ||
4a85b96c | 878 | extern void ip_vs_tcp_conn_listen(struct net *net, struct ip_vs_conn *cp); |
1da177e4 | 879 | extern int ip_vs_check_template(struct ip_vs_conn *ct); |
f6340ee0 | 880 | extern void ip_vs_random_dropentry(struct net *net); |
1da177e4 LT |
881 | extern int ip_vs_conn_init(void); |
882 | extern void ip_vs_conn_cleanup(void); | |
883 | ||
884 | static inline void ip_vs_control_del(struct ip_vs_conn *cp) | |
885 | { | |
886 | struct ip_vs_conn *ctl_cp = cp->control; | |
887 | if (!ctl_cp) { | |
cfc78c5a JV |
888 | IP_VS_ERR_BUF("request control DEL for uncontrolled: " |
889 | "%s:%d to %s:%d\n", | |
890 | IP_VS_DBG_ADDR(cp->af, &cp->caddr), | |
891 | ntohs(cp->cport), | |
892 | IP_VS_DBG_ADDR(cp->af, &cp->vaddr), | |
893 | ntohs(cp->vport)); | |
894 | ||
1da177e4 LT |
895 | return; |
896 | } | |
897 | ||
cfc78c5a JV |
898 | IP_VS_DBG_BUF(7, "DELeting control for: " |
899 | "cp.dst=%s:%d ctl_cp.dst=%s:%d\n", | |
900 | IP_VS_DBG_ADDR(cp->af, &cp->caddr), | |
901 | ntohs(cp->cport), | |
902 | IP_VS_DBG_ADDR(cp->af, &ctl_cp->caddr), | |
903 | ntohs(ctl_cp->cport)); | |
1da177e4 LT |
904 | |
905 | cp->control = NULL; | |
906 | if (atomic_read(&ctl_cp->n_control) == 0) { | |
cfc78c5a JV |
907 | IP_VS_ERR_BUF("BUG control DEL with n=0 : " |
908 | "%s:%d to %s:%d\n", | |
909 | IP_VS_DBG_ADDR(cp->af, &cp->caddr), | |
910 | ntohs(cp->cport), | |
911 | IP_VS_DBG_ADDR(cp->af, &cp->vaddr), | |
912 | ntohs(cp->vport)); | |
913 | ||
1da177e4 LT |
914 | return; |
915 | } | |
916 | atomic_dec(&ctl_cp->n_control); | |
917 | } | |
918 | ||
919 | static inline void | |
920 | ip_vs_control_add(struct ip_vs_conn *cp, struct ip_vs_conn *ctl_cp) | |
921 | { | |
922 | if (cp->control) { | |
cfc78c5a JV |
923 | IP_VS_ERR_BUF("request control ADD for already controlled: " |
924 | "%s:%d to %s:%d\n", | |
925 | IP_VS_DBG_ADDR(cp->af, &cp->caddr), | |
926 | ntohs(cp->cport), | |
927 | IP_VS_DBG_ADDR(cp->af, &cp->vaddr), | |
928 | ntohs(cp->vport)); | |
929 | ||
1da177e4 LT |
930 | ip_vs_control_del(cp); |
931 | } | |
932 | ||
cfc78c5a JV |
933 | IP_VS_DBG_BUF(7, "ADDing control for: " |
934 | "cp.dst=%s:%d ctl_cp.dst=%s:%d\n", | |
935 | IP_VS_DBG_ADDR(cp->af, &cp->caddr), | |
936 | ntohs(cp->cport), | |
937 | IP_VS_DBG_ADDR(cp->af, &ctl_cp->caddr), | |
938 | ntohs(ctl_cp->cport)); | |
1da177e4 LT |
939 | |
940 | cp->control = ctl_cp; | |
941 | atomic_inc(&ctl_cp->n_control); | |
942 | } | |
943 | ||
944 | ||
945 | /* | |
946 | * IPVS application functions | |
947 | * (from ip_vs_app.c) | |
948 | */ | |
949 | #define IP_VS_APP_MAX_PORTS 8 | |
ab8a5e84 HS |
950 | extern int register_ip_vs_app(struct net *net, struct ip_vs_app *app); |
951 | extern void unregister_ip_vs_app(struct net *net, struct ip_vs_app *app); | |
1da177e4 LT |
952 | extern int ip_vs_bind_app(struct ip_vs_conn *cp, struct ip_vs_protocol *pp); |
953 | extern void ip_vs_unbind_app(struct ip_vs_conn *cp); | |
ab8a5e84 HS |
954 | extern int register_ip_vs_app_inc(struct net *net, struct ip_vs_app *app, |
955 | __u16 proto, __u16 port); | |
1da177e4 LT |
956 | extern int ip_vs_app_inc_get(struct ip_vs_app *inc); |
957 | extern void ip_vs_app_inc_put(struct ip_vs_app *inc); | |
958 | ||
3db05fea HX |
959 | extern int ip_vs_app_pkt_out(struct ip_vs_conn *, struct sk_buff *skb); |
960 | extern int ip_vs_app_pkt_in(struct ip_vs_conn *, struct sk_buff *skb); | |
1da177e4 LT |
961 | extern int ip_vs_app_init(void); |
962 | extern void ip_vs_app_cleanup(void); | |
963 | ||
8be67a66 SH |
964 | void ip_vs_bind_pe(struct ip_vs_service *svc, struct ip_vs_pe *pe); |
965 | void ip_vs_unbind_pe(struct ip_vs_service *svc); | |
966 | int register_ip_vs_pe(struct ip_vs_pe *pe); | |
967 | int unregister_ip_vs_pe(struct ip_vs_pe *pe); | |
e9e5eee8 | 968 | struct ip_vs_pe *ip_vs_pe_getbyname(const char *name); |
fe5e7a1e | 969 | struct ip_vs_pe *__ip_vs_pe_getbyname(const char *pe_name); |
e9e5eee8 SH |
970 | |
971 | static inline void ip_vs_pe_get(const struct ip_vs_pe *pe) | |
972 | { | |
973 | if (pe && pe->module) | |
974 | __module_get(pe->module); | |
975 | } | |
976 | ||
977 | static inline void ip_vs_pe_put(const struct ip_vs_pe *pe) | |
978 | { | |
979 | if (pe && pe->module) | |
980 | module_put(pe->module); | |
981 | } | |
1da177e4 LT |
982 | |
983 | /* | |
984 | * IPVS protocol functions (from ip_vs_proto.c) | |
985 | */ | |
986 | extern int ip_vs_protocol_init(void); | |
987 | extern void ip_vs_protocol_cleanup(void); | |
9330419d | 988 | extern void ip_vs_protocol_timeout_change(struct netns_ipvs *ipvs, int flags); |
1da177e4 LT |
989 | extern int *ip_vs_create_timeout_table(int *table, int size); |
990 | extern int | |
36cbd3dc JE |
991 | ip_vs_set_state_timeout(int *table, int num, const char *const *names, |
992 | const char *name, int to); | |
1da177e4 | 993 | extern void |
0d79641a JA |
994 | ip_vs_tcpudp_debug_packet(int af, struct ip_vs_protocol *pp, |
995 | const struct sk_buff *skb, | |
1da177e4 LT |
996 | int offset, const char *msg); |
997 | ||
998 | extern struct ip_vs_protocol ip_vs_protocol_tcp; | |
999 | extern struct ip_vs_protocol ip_vs_protocol_udp; | |
1000 | extern struct ip_vs_protocol ip_vs_protocol_icmp; | |
1001 | extern struct ip_vs_protocol ip_vs_protocol_esp; | |
1002 | extern struct ip_vs_protocol ip_vs_protocol_ah; | |
2906f66a | 1003 | extern struct ip_vs_protocol ip_vs_protocol_sctp; |
1da177e4 LT |
1004 | |
1005 | /* | |
1006 | * Registering/unregistering scheduler functions | |
1007 | * (from ip_vs_sched.c) | |
1008 | */ | |
1009 | extern int register_ip_vs_scheduler(struct ip_vs_scheduler *scheduler); | |
1010 | extern int unregister_ip_vs_scheduler(struct ip_vs_scheduler *scheduler); | |
1011 | extern int ip_vs_bind_scheduler(struct ip_vs_service *svc, | |
1012 | struct ip_vs_scheduler *scheduler); | |
1013 | extern int ip_vs_unbind_scheduler(struct ip_vs_service *svc); | |
1014 | extern struct ip_vs_scheduler *ip_vs_scheduler_get(const char *sched_name); | |
1015 | extern void ip_vs_scheduler_put(struct ip_vs_scheduler *scheduler); | |
1016 | extern struct ip_vs_conn * | |
190ecd27 | 1017 | ip_vs_schedule(struct ip_vs_service *svc, struct sk_buff *skb, |
9330419d | 1018 | struct ip_vs_proto_data *pd, int *ignored); |
1da177e4 | 1019 | extern int ip_vs_leave(struct ip_vs_service *svc, struct sk_buff *skb, |
9330419d | 1020 | struct ip_vs_proto_data *pd); |
1da177e4 | 1021 | |
41ac51ee PS |
1022 | extern void ip_vs_scheduler_err(struct ip_vs_service *svc, const char *msg); |
1023 | ||
1da177e4 LT |
1024 | |
1025 | /* | |
1026 | * IPVS control data and functions (from ip_vs_ctl.c) | |
1027 | */ | |
1da177e4 | 1028 | extern struct ip_vs_stats ip_vs_stats; |
5587da55 | 1029 | extern const struct ctl_path net_vs_ctl_path[]; |
b880c1f0 | 1030 | extern int sysctl_ip_vs_sync_ver; |
1da177e4 | 1031 | |
f131315f | 1032 | extern void ip_vs_sync_switch_mode(struct net *net, int mode); |
1da177e4 | 1033 | extern struct ip_vs_service * |
fc723250 | 1034 | ip_vs_service_get(struct net *net, int af, __u32 fwmark, __u16 protocol, |
3c2e0505 | 1035 | const union nf_inet_addr *vaddr, __be16 vport); |
1da177e4 LT |
1036 | |
1037 | static inline void ip_vs_service_put(struct ip_vs_service *svc) | |
1038 | { | |
1039 | atomic_dec(&svc->usecnt); | |
1040 | } | |
1041 | ||
1042 | extern struct ip_vs_dest * | |
fc723250 | 1043 | ip_vs_lookup_real_service(struct net *net, int af, __u16 protocol, |
7937df15 JV |
1044 | const union nf_inet_addr *daddr, __be16 dport); |
1045 | ||
1da177e4 LT |
1046 | extern int ip_vs_use_count_inc(void); |
1047 | extern void ip_vs_use_count_dec(void); | |
1048 | extern int ip_vs_control_init(void); | |
1049 | extern void ip_vs_control_cleanup(void); | |
1e356f9c | 1050 | extern struct ip_vs_dest * |
fc723250 HS |
1051 | ip_vs_find_dest(struct net *net, int af, const union nf_inet_addr *daddr, |
1052 | __be16 dport, const union nf_inet_addr *vaddr, __be16 vport, | |
1053 | __u16 protocol, __u32 fwmark); | |
1e356f9c | 1054 | extern struct ip_vs_dest *ip_vs_try_bind_dest(struct ip_vs_conn *cp); |
1da177e4 LT |
1055 | |
1056 | ||
1057 | /* | |
1058 | * IPVS sync daemon data and function prototypes | |
1059 | * (from ip_vs_sync.c) | |
1060 | */ | |
f131315f HS |
1061 | extern int start_sync_thread(struct net *net, int state, char *mcast_ifn, |
1062 | __u8 syncid); | |
1063 | extern int stop_sync_thread(struct net *net, int state); | |
1064 | extern void ip_vs_sync_conn(struct net *net, struct ip_vs_conn *cp); | |
61b1ab45 HS |
1065 | extern int ip_vs_sync_init(void); |
1066 | extern void ip_vs_sync_cleanup(void); | |
1da177e4 LT |
1067 | |
1068 | ||
1069 | /* | |
1070 | * IPVS rate estimator prototypes (from ip_vs_est.c) | |
1071 | */ | |
a919cf4b SW |
1072 | extern int ip_vs_estimator_init(void); |
1073 | extern void ip_vs_estimator_cleanup(void); | |
29c2026f HS |
1074 | extern void ip_vs_new_estimator(struct net *net, struct ip_vs_stats *stats); |
1075 | extern void ip_vs_kill_estimator(struct net *net, struct ip_vs_stats *stats); | |
1da177e4 LT |
1076 | extern void ip_vs_zero_estimator(struct ip_vs_stats *stats); |
1077 | ||
1078 | /* | |
1079 | * Various IPVS packet transmitters (from ip_vs_xmit.c) | |
1080 | */ | |
1081 | extern int ip_vs_null_xmit | |
1082 | (struct sk_buff *skb, struct ip_vs_conn *cp, struct ip_vs_protocol *pp); | |
1083 | extern int ip_vs_bypass_xmit | |
1084 | (struct sk_buff *skb, struct ip_vs_conn *cp, struct ip_vs_protocol *pp); | |
1085 | extern int ip_vs_nat_xmit | |
1086 | (struct sk_buff *skb, struct ip_vs_conn *cp, struct ip_vs_protocol *pp); | |
1087 | extern int ip_vs_tunnel_xmit | |
1088 | (struct sk_buff *skb, struct ip_vs_conn *cp, struct ip_vs_protocol *pp); | |
1089 | extern int ip_vs_dr_xmit | |
1090 | (struct sk_buff *skb, struct ip_vs_conn *cp, struct ip_vs_protocol *pp); | |
1091 | extern int ip_vs_icmp_xmit | |
1092 | (struct sk_buff *skb, struct ip_vs_conn *cp, struct ip_vs_protocol *pp, int offset); | |
1093 | extern void ip_vs_dst_reset(struct ip_vs_dest *dest); | |
1094 | ||
b3cdd2a7 JV |
1095 | #ifdef CONFIG_IP_VS_IPV6 |
1096 | extern int ip_vs_bypass_xmit_v6 | |
1097 | (struct sk_buff *skb, struct ip_vs_conn *cp, struct ip_vs_protocol *pp); | |
1098 | extern int ip_vs_nat_xmit_v6 | |
1099 | (struct sk_buff *skb, struct ip_vs_conn *cp, struct ip_vs_protocol *pp); | |
1100 | extern int ip_vs_tunnel_xmit_v6 | |
1101 | (struct sk_buff *skb, struct ip_vs_conn *cp, struct ip_vs_protocol *pp); | |
1102 | extern int ip_vs_dr_xmit_v6 | |
1103 | (struct sk_buff *skb, struct ip_vs_conn *cp, struct ip_vs_protocol *pp); | |
1104 | extern int ip_vs_icmp_xmit_v6 | |
1105 | (struct sk_buff *skb, struct ip_vs_conn *cp, struct ip_vs_protocol *pp, | |
1106 | int offset); | |
1107 | #endif | |
1da177e4 LT |
1108 | |
1109 | /* | |
1110 | * This is a simple mechanism to ignore packets when | |
1111 | * we are loaded. Just set ip_vs_drop_rate to 'n' and | |
1112 | * we start to drop 1/rate of the packets | |
1113 | */ | |
1da177e4 | 1114 | |
a0840e2e | 1115 | static inline int ip_vs_todrop(struct netns_ipvs *ipvs) |
1da177e4 | 1116 | { |
a0840e2e HS |
1117 | if (!ipvs->drop_rate) |
1118 | return 0; | |
1119 | if (--ipvs->drop_counter > 0) | |
1120 | return 0; | |
1121 | ipvs->drop_counter = ipvs->drop_rate; | |
1da177e4 LT |
1122 | return 1; |
1123 | } | |
1124 | ||
1125 | /* | |
1126 | * ip_vs_fwd_tag returns the forwarding tag of the connection | |
1127 | */ | |
1128 | #define IP_VS_FWD_METHOD(cp) (cp->flags & IP_VS_CONN_F_FWD_MASK) | |
1129 | ||
732db659 | 1130 | static inline char ip_vs_fwd_tag(struct ip_vs_conn *cp) |
1da177e4 LT |
1131 | { |
1132 | char fwd; | |
1133 | ||
1134 | switch (IP_VS_FWD_METHOD(cp)) { | |
1135 | case IP_VS_CONN_F_MASQ: | |
1136 | fwd = 'M'; break; | |
1137 | case IP_VS_CONN_F_LOCALNODE: | |
1138 | fwd = 'L'; break; | |
1139 | case IP_VS_CONN_F_TUNNEL: | |
1140 | fwd = 'T'; break; | |
1141 | case IP_VS_CONN_F_DROUTE: | |
1142 | fwd = 'R'; break; | |
1143 | case IP_VS_CONN_F_BYPASS: | |
1144 | fwd = 'B'; break; | |
1145 | default: | |
1146 | fwd = '?'; break; | |
1147 | } | |
1148 | return fwd; | |
1149 | } | |
1150 | ||
1da177e4 | 1151 | extern void ip_vs_nat_icmp(struct sk_buff *skb, struct ip_vs_protocol *pp, |
b3cdd2a7 JV |
1152 | struct ip_vs_conn *cp, int dir); |
1153 | ||
1154 | #ifdef CONFIG_IP_VS_IPV6 | |
1155 | extern void ip_vs_nat_icmp_v6(struct sk_buff *skb, struct ip_vs_protocol *pp, | |
1156 | struct ip_vs_conn *cp, int dir); | |
1157 | #endif | |
1da177e4 | 1158 | |
b1550f22 | 1159 | extern __sum16 ip_vs_checksum_complete(struct sk_buff *skb, int offset); |
1da177e4 | 1160 | |
f9214b26 | 1161 | static inline __wsum ip_vs_check_diff4(__be32 old, __be32 new, __wsum oldsum) |
1da177e4 | 1162 | { |
f9214b26 | 1163 | __be32 diff[2] = { ~old, new }; |
1da177e4 | 1164 | |
07f0757a | 1165 | return csum_partial(diff, sizeof(diff), oldsum); |
f9214b26 AV |
1166 | } |
1167 | ||
0bbdd42b JV |
1168 | #ifdef CONFIG_IP_VS_IPV6 |
1169 | static inline __wsum ip_vs_check_diff16(const __be32 *old, const __be32 *new, | |
1170 | __wsum oldsum) | |
1171 | { | |
1172 | __be32 diff[8] = { ~old[3], ~old[2], ~old[1], ~old[0], | |
1173 | new[3], new[2], new[1], new[0] }; | |
1174 | ||
07f0757a | 1175 | return csum_partial(diff, sizeof(diff), oldsum); |
0bbdd42b JV |
1176 | } |
1177 | #endif | |
1178 | ||
f9214b26 AV |
1179 | static inline __wsum ip_vs_check_diff2(__be16 old, __be16 new, __wsum oldsum) |
1180 | { | |
1181 | __be16 diff[2] = { ~old, new }; | |
1182 | ||
07f0757a | 1183 | return csum_partial(diff, sizeof(diff), oldsum); |
1da177e4 LT |
1184 | } |
1185 | ||
cf356d69 JA |
1186 | /* |
1187 | * Forget current conntrack (unconfirmed) and attach notrack entry | |
1188 | */ | |
1189 | static inline void ip_vs_notrack(struct sk_buff *skb) | |
1190 | { | |
1191 | #if defined(CONFIG_NF_CONNTRACK) || defined(CONFIG_NF_CONNTRACK_MODULE) | |
1192 | enum ip_conntrack_info ctinfo; | |
06b69390 | 1193 | struct nf_conn *ct = nf_ct_get(skb, &ctinfo); |
cf356d69 JA |
1194 | |
1195 | if (!ct || !nf_ct_is_untracked(ct)) { | |
1196 | nf_reset(skb); | |
1197 | skb->nfct = &nf_ct_untracked_get()->ct_general; | |
1198 | skb->nfctinfo = IP_CT_NEW; | |
1199 | nf_conntrack_get(skb->nfct); | |
1200 | } | |
1201 | #endif | |
1202 | } | |
1203 | ||
f4bc17cd JA |
1204 | #ifdef CONFIG_IP_VS_NFCT |
1205 | /* | |
1206 | * Netfilter connection tracking | |
1207 | * (from ip_vs_nfct.c) | |
1208 | */ | |
a0840e2e | 1209 | static inline int ip_vs_conntrack_enabled(struct netns_ipvs *ipvs) |
f4bc17cd | 1210 | { |
a0840e2e | 1211 | return ipvs->sysctl_conntrack; |
f4bc17cd JA |
1212 | } |
1213 | ||
6523ce15 JA |
1214 | extern void ip_vs_update_conntrack(struct sk_buff *skb, struct ip_vs_conn *cp, |
1215 | int outin); | |
f4bc17cd JA |
1216 | extern int ip_vs_confirm_conntrack(struct sk_buff *skb, struct ip_vs_conn *cp); |
1217 | extern void ip_vs_nfct_expect_related(struct sk_buff *skb, struct nf_conn *ct, | |
1218 | struct ip_vs_conn *cp, u_int8_t proto, | |
1219 | const __be16 port, int from_rs); | |
1220 | extern void ip_vs_conn_drop_conntrack(struct ip_vs_conn *cp); | |
1221 | ||
1222 | #else | |
1223 | ||
a0840e2e | 1224 | static inline int ip_vs_conntrack_enabled(struct netns_ipvs *ipvs) |
f4bc17cd JA |
1225 | { |
1226 | return 0; | |
1227 | } | |
1228 | ||
1229 | static inline void ip_vs_update_conntrack(struct sk_buff *skb, | |
1230 | struct ip_vs_conn *cp, int outin) | |
1231 | { | |
1232 | } | |
1233 | ||
1234 | static inline int ip_vs_confirm_conntrack(struct sk_buff *skb, | |
1235 | struct ip_vs_conn *cp) | |
1236 | { | |
1237 | return NF_ACCEPT; | |
1238 | } | |
1239 | ||
1240 | static inline void ip_vs_conn_drop_conntrack(struct ip_vs_conn *cp) | |
1241 | { | |
1242 | } | |
1243 | /* CONFIG_IP_VS_NFCT */ | |
1244 | #endif | |
6523ce15 | 1245 | |
b552f7e3 CG |
1246 | static inline unsigned int |
1247 | ip_vs_dest_conn_overhead(struct ip_vs_dest *dest) | |
1248 | { | |
1249 | /* | |
1250 | * We think the overhead of processing active connections is 256 | |
1251 | * times higher than that of inactive connections in average. (This | |
1252 | * 256 times might not be accurate, we will change it later) We | |
1253 | * use the following formula to estimate the overhead now: | |
1254 | * dest->activeconns*256 + dest->inactconns | |
1255 | */ | |
1256 | return (atomic_read(&dest->activeconns) << 8) + | |
1257 | atomic_read(&dest->inactconns); | |
1258 | } | |
1259 | ||
1da177e4 LT |
1260 | #endif /* __KERNEL__ */ |
1261 | ||
bc4768eb | 1262 | #endif /* _NET_IP_VS_H */ |