FreeBSD/Linux Kernel Cross Reference
sys/netinet/ip_fw.h
1 /*-
2 * Copyright (c) 2002-2009 Luigi Rizzo, Universita` di Pisa
3 *
4 * Redistribution and use in source and binary forms, with or without
5 * modification, are permitted provided that the following conditions
6 * are met:
7 * 1. Redistributions of source code must retain the above copyright
8 * notice, this list of conditions and the following disclaimer.
9 * 2. Redistributions in binary form must reproduce the above copyright
10 * notice, this list of conditions and the following disclaimer in the
11 * documentation and/or other materials provided with the distribution.
12 *
13 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
14 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
15 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
16 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
17 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
18 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
19 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
20 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
21 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
22 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
23 * SUCH DAMAGE.
24 *
25 * $FreeBSD: releng/9.0/sys/netinet/ip_fw.h 225044 2011-08-20 17:05:11Z bz $
26 */
27
28 #ifndef _IPFW2_H
29 #define _IPFW2_H
30
31 /*
32 * The default rule number. By the design of ip_fw, the default rule
33 * is the last one, so its number can also serve as the highest number
34 * allowed for a rule. The ip_fw code relies on both meanings of this
35 * constant.
36 */
37 #define IPFW_DEFAULT_RULE 65535
38
39 /*
40 * The number of ipfw tables. The maximum allowed table number is the
41 * (IPFW_TABLES_MAX - 1).
42 */
43 #define IPFW_TABLES_MAX 128
44
45 /*
46 * Most commands (queue, pipe, tag, untag, limit...) can have a 16-bit
47 * argument between 1 and 65534. The value 0 is unused, the value
48 * 65535 (IP_FW_TABLEARG) is used to represent 'tablearg', i.e. the
49 * can be 1..65534, or 65535 to indicate the use of a 'tablearg'
50 * result of the most recent table() lookup.
51 * Note that 16bit is only a historical limit, resulting from
52 * the use of a 16-bit fields for that value. In reality, we can have
53 * 2^32 pipes, queues, tag values and so on, and use 0 as a tablearg.
54 */
55 #define IPFW_ARG_MIN 1
56 #define IPFW_ARG_MAX 65534
57 #define IP_FW_TABLEARG 65535 /* XXX should use 0 */
58
59 /*
60 * Number of entries in the call stack of the call/return commands.
61 * Call stack currently is an uint16_t array with rule numbers.
62 */
63 #define IPFW_CALLSTACK_SIZE 16
64
65 /*
66 * The kernel representation of ipfw rules is made of a list of
67 * 'instructions' (for all practical purposes equivalent to BPF
68 * instructions), which specify which fields of the packet
69 * (or its metadata) should be analysed.
70 *
71 * Each instruction is stored in a structure which begins with
72 * "ipfw_insn", and can contain extra fields depending on the
73 * instruction type (listed below).
74 * Note that the code is written so that individual instructions
75 * have a size which is a multiple of 32 bits. This means that, if
76 * such structures contain pointers or other 64-bit entities,
77 * (there is just one instance now) they may end up unaligned on
78 * 64-bit architectures, so the must be handled with care.
79 *
80 * "enum ipfw_opcodes" are the opcodes supported. We can have up
81 * to 256 different opcodes. When adding new opcodes, they should
82 * be appended to the end of the opcode list before O_LAST_OPCODE,
83 * this will prevent the ABI from being broken, otherwise users
84 * will have to recompile ipfw(8) when they update the kernel.
85 */
86
87 enum ipfw_opcodes { /* arguments (4 byte each) */
88 O_NOP,
89
90 O_IP_SRC, /* u32 = IP */
91 O_IP_SRC_MASK, /* ip = IP/mask */
92 O_IP_SRC_ME, /* none */
93 O_IP_SRC_SET, /* u32=base, arg1=len, bitmap */
94
95 O_IP_DST, /* u32 = IP */
96 O_IP_DST_MASK, /* ip = IP/mask */
97 O_IP_DST_ME, /* none */
98 O_IP_DST_SET, /* u32=base, arg1=len, bitmap */
99
100 O_IP_SRCPORT, /* (n)port list:mask 4 byte ea */
101 O_IP_DSTPORT, /* (n)port list:mask 4 byte ea */
102 O_PROTO, /* arg1=protocol */
103
104 O_MACADDR2, /* 2 mac addr:mask */
105 O_MAC_TYPE, /* same as srcport */
106
107 O_LAYER2, /* none */
108 O_IN, /* none */
109 O_FRAG, /* none */
110
111 O_RECV, /* none */
112 O_XMIT, /* none */
113 O_VIA, /* none */
114
115 O_IPOPT, /* arg1 = 2*u8 bitmap */
116 O_IPLEN, /* arg1 = len */
117 O_IPID, /* arg1 = id */
118
119 O_IPTOS, /* arg1 = id */
120 O_IPPRECEDENCE, /* arg1 = precedence << 5 */
121 O_IPTTL, /* arg1 = TTL */
122
123 O_IPVER, /* arg1 = version */
124 O_UID, /* u32 = id */
125 O_GID, /* u32 = id */
126 O_ESTAB, /* none (tcp established) */
127 O_TCPFLAGS, /* arg1 = 2*u8 bitmap */
128 O_TCPWIN, /* arg1 = desired win */
129 O_TCPSEQ, /* u32 = desired seq. */
130 O_TCPACK, /* u32 = desired seq. */
131 O_ICMPTYPE, /* u32 = icmp bitmap */
132 O_TCPOPTS, /* arg1 = 2*u8 bitmap */
133
134 O_VERREVPATH, /* none */
135 O_VERSRCREACH, /* none */
136
137 O_PROBE_STATE, /* none */
138 O_KEEP_STATE, /* none */
139 O_LIMIT, /* ipfw_insn_limit */
140 O_LIMIT_PARENT, /* dyn_type, not an opcode. */
141
142 /*
143 * These are really 'actions'.
144 */
145
146 O_LOG, /* ipfw_insn_log */
147 O_PROB, /* u32 = match probability */
148
149 O_CHECK_STATE, /* none */
150 O_ACCEPT, /* none */
151 O_DENY, /* none */
152 O_REJECT, /* arg1=icmp arg (same as deny) */
153 O_COUNT, /* none */
154 O_SKIPTO, /* arg1=next rule number */
155 O_PIPE, /* arg1=pipe number */
156 O_QUEUE, /* arg1=queue number */
157 O_DIVERT, /* arg1=port number */
158 O_TEE, /* arg1=port number */
159 O_FORWARD_IP, /* fwd sockaddr */
160 O_FORWARD_MAC, /* fwd mac */
161 O_NAT, /* nope */
162 O_REASS, /* none */
163
164 /*
165 * More opcodes.
166 */
167 O_IPSEC, /* has ipsec history */
168 O_IP_SRC_LOOKUP, /* arg1=table number, u32=value */
169 O_IP_DST_LOOKUP, /* arg1=table number, u32=value */
170 O_ANTISPOOF, /* none */
171 O_JAIL, /* u32 = id */
172 O_ALTQ, /* u32 = altq classif. qid */
173 O_DIVERTED, /* arg1=bitmap (1:loop, 2:out) */
174 O_TCPDATALEN, /* arg1 = tcp data len */
175 O_IP6_SRC, /* address without mask */
176 O_IP6_SRC_ME, /* my addresses */
177 O_IP6_SRC_MASK, /* address with the mask */
178 O_IP6_DST,
179 O_IP6_DST_ME,
180 O_IP6_DST_MASK,
181 O_FLOW6ID, /* for flow id tag in the ipv6 pkt */
182 O_ICMP6TYPE, /* icmp6 packet type filtering */
183 O_EXT_HDR, /* filtering for ipv6 extension header */
184 O_IP6,
185
186 /*
187 * actions for ng_ipfw
188 */
189 O_NETGRAPH, /* send to ng_ipfw */
190 O_NGTEE, /* copy to ng_ipfw */
191
192 O_IP4,
193
194 O_UNREACH6, /* arg1=icmpv6 code arg (deny) */
195
196 O_TAG, /* arg1=tag number */
197 O_TAGGED, /* arg1=tag number */
198
199 O_SETFIB, /* arg1=FIB number */
200 O_FIB, /* arg1=FIB desired fib number */
201
202 O_SOCKARG, /* socket argument */
203
204 O_CALLRETURN, /* arg1=called rule number */
205
206 O_FORWARD_IP6, /* fwd sockaddr_in6 */
207
208 O_LAST_OPCODE /* not an opcode! */
209 };
210
211
212 /*
213 * The extension header are filtered only for presence using a bit
214 * vector with a flag for each header.
215 */
216 #define EXT_FRAGMENT 0x1
217 #define EXT_HOPOPTS 0x2
218 #define EXT_ROUTING 0x4
219 #define EXT_AH 0x8
220 #define EXT_ESP 0x10
221 #define EXT_DSTOPTS 0x20
222 #define EXT_RTHDR0 0x40
223 #define EXT_RTHDR2 0x80
224
225 /*
226 * Template for instructions.
227 *
228 * ipfw_insn is used for all instructions which require no operands,
229 * a single 16-bit value (arg1), or a couple of 8-bit values.
230 *
231 * For other instructions which require different/larger arguments
232 * we have derived structures, ipfw_insn_*.
233 *
234 * The size of the instruction (in 32-bit words) is in the low
235 * 6 bits of "len". The 2 remaining bits are used to implement
236 * NOT and OR on individual instructions. Given a type, you can
237 * compute the length to be put in "len" using F_INSN_SIZE(t)
238 *
239 * F_NOT negates the match result of the instruction.
240 *
241 * F_OR is used to build or blocks. By default, instructions
242 * are evaluated as part of a logical AND. An "or" block
243 * { X or Y or Z } contains F_OR set in all but the last
244 * instruction of the block. A match will cause the code
245 * to skip past the last instruction of the block.
246 *
247 * NOTA BENE: in a couple of places we assume that
248 * sizeof(ipfw_insn) == sizeof(u_int32_t)
249 * this needs to be fixed.
250 *
251 */
252 typedef struct _ipfw_insn { /* template for instructions */
253 u_int8_t opcode;
254 u_int8_t len; /* number of 32-bit words */
255 #define F_NOT 0x80
256 #define F_OR 0x40
257 #define F_LEN_MASK 0x3f
258 #define F_LEN(cmd) ((cmd)->len & F_LEN_MASK)
259
260 u_int16_t arg1;
261 } ipfw_insn;
262
263 /*
264 * The F_INSN_SIZE(type) computes the size, in 4-byte words, of
265 * a given type.
266 */
267 #define F_INSN_SIZE(t) ((sizeof (t))/sizeof(u_int32_t))
268
269 /*
270 * This is used to store an array of 16-bit entries (ports etc.)
271 */
272 typedef struct _ipfw_insn_u16 {
273 ipfw_insn o;
274 u_int16_t ports[2]; /* there may be more */
275 } ipfw_insn_u16;
276
277 /*
278 * This is used to store an array of 32-bit entries
279 * (uid, single IPv4 addresses etc.)
280 */
281 typedef struct _ipfw_insn_u32 {
282 ipfw_insn o;
283 u_int32_t d[1]; /* one or more */
284 } ipfw_insn_u32;
285
286 /*
287 * This is used to store IP addr-mask pairs.
288 */
289 typedef struct _ipfw_insn_ip {
290 ipfw_insn o;
291 struct in_addr addr;
292 struct in_addr mask;
293 } ipfw_insn_ip;
294
295 /*
296 * This is used to forward to a given address (ip).
297 */
298 typedef struct _ipfw_insn_sa {
299 ipfw_insn o;
300 struct sockaddr_in sa;
301 } ipfw_insn_sa;
302
303 /*
304 * This is used to forward to a given address (ipv6).
305 */
306 typedef struct _ipfw_insn_sa6 {
307 ipfw_insn o;
308 struct sockaddr_in6 sa;
309 } ipfw_insn_sa6;
310
311 /*
312 * This is used for MAC addr-mask pairs.
313 */
314 typedef struct _ipfw_insn_mac {
315 ipfw_insn o;
316 u_char addr[12]; /* dst[6] + src[6] */
317 u_char mask[12]; /* dst[6] + src[6] */
318 } ipfw_insn_mac;
319
320 /*
321 * This is used for interface match rules (recv xx, xmit xx).
322 */
323 typedef struct _ipfw_insn_if {
324 ipfw_insn o;
325 union {
326 struct in_addr ip;
327 int glob;
328 } p;
329 char name[IFNAMSIZ];
330 } ipfw_insn_if;
331
332 /*
333 * This is used for storing an altq queue id number.
334 */
335 typedef struct _ipfw_insn_altq {
336 ipfw_insn o;
337 u_int32_t qid;
338 } ipfw_insn_altq;
339
340 /*
341 * This is used for limit rules.
342 */
343 typedef struct _ipfw_insn_limit {
344 ipfw_insn o;
345 u_int8_t _pad;
346 u_int8_t limit_mask; /* combination of DYN_* below */
347 #define DYN_SRC_ADDR 0x1
348 #define DYN_SRC_PORT 0x2
349 #define DYN_DST_ADDR 0x4
350 #define DYN_DST_PORT 0x8
351
352 u_int16_t conn_limit;
353 } ipfw_insn_limit;
354
355 /*
356 * This is used for log instructions.
357 */
358 typedef struct _ipfw_insn_log {
359 ipfw_insn o;
360 u_int32_t max_log; /* how many do we log -- 0 = all */
361 u_int32_t log_left; /* how many left to log */
362 } ipfw_insn_log;
363
364 /*
365 * Data structures required by both ipfw(8) and ipfw(4) but not part of the
366 * management API are protected by IPFW_INTERNAL.
367 */
368 #ifdef IPFW_INTERNAL
369 /* Server pool support (LSNAT). */
370 struct cfg_spool {
371 LIST_ENTRY(cfg_spool) _next; /* chain of spool instances */
372 struct in_addr addr;
373 u_short port;
374 };
375 #endif
376
377 /* Redirect modes id. */
378 #define REDIR_ADDR 0x01
379 #define REDIR_PORT 0x02
380 #define REDIR_PROTO 0x04
381
382 #ifdef IPFW_INTERNAL
383 /* Nat redirect configuration. */
384 struct cfg_redir {
385 LIST_ENTRY(cfg_redir) _next; /* chain of redir instances */
386 u_int16_t mode; /* type of redirect mode */
387 struct in_addr laddr; /* local ip address */
388 struct in_addr paddr; /* public ip address */
389 struct in_addr raddr; /* remote ip address */
390 u_short lport; /* local port */
391 u_short pport; /* public port */
392 u_short rport; /* remote port */
393 u_short pport_cnt; /* number of public ports */
394 u_short rport_cnt; /* number of remote ports */
395 int proto; /* protocol: tcp/udp */
396 struct alias_link **alink;
397 /* num of entry in spool chain */
398 u_int16_t spool_cnt;
399 /* chain of spool instances */
400 LIST_HEAD(spool_chain, cfg_spool) spool_chain;
401 };
402 #endif
403
404 #ifdef IPFW_INTERNAL
405 /* Nat configuration data struct. */
406 struct cfg_nat {
407 /* chain of nat instances */
408 LIST_ENTRY(cfg_nat) _next;
409 int id; /* nat id */
410 struct in_addr ip; /* nat ip address */
411 char if_name[IF_NAMESIZE]; /* interface name */
412 int mode; /* aliasing mode */
413 struct libalias *lib; /* libalias instance */
414 /* number of entry in spool chain */
415 int redir_cnt;
416 /* chain of redir instances */
417 LIST_HEAD(redir_chain, cfg_redir) redir_chain;
418 };
419 #endif
420
421 #define SOF_NAT sizeof(struct cfg_nat)
422 #define SOF_REDIR sizeof(struct cfg_redir)
423 #define SOF_SPOOL sizeof(struct cfg_spool)
424
425 /* Nat command. */
426 typedef struct _ipfw_insn_nat {
427 ipfw_insn o;
428 struct cfg_nat *nat;
429 } ipfw_insn_nat;
430
431 /* Apply ipv6 mask on ipv6 addr */
432 #define APPLY_MASK(addr,mask) \
433 (addr)->__u6_addr.__u6_addr32[0] &= (mask)->__u6_addr.__u6_addr32[0]; \
434 (addr)->__u6_addr.__u6_addr32[1] &= (mask)->__u6_addr.__u6_addr32[1]; \
435 (addr)->__u6_addr.__u6_addr32[2] &= (mask)->__u6_addr.__u6_addr32[2]; \
436 (addr)->__u6_addr.__u6_addr32[3] &= (mask)->__u6_addr.__u6_addr32[3];
437
438 /* Structure for ipv6 */
439 typedef struct _ipfw_insn_ip6 {
440 ipfw_insn o;
441 struct in6_addr addr6;
442 struct in6_addr mask6;
443 } ipfw_insn_ip6;
444
445 /* Used to support icmp6 types */
446 typedef struct _ipfw_insn_icmp6 {
447 ipfw_insn o;
448 uint32_t d[7]; /* XXX This number si related to the netinet/icmp6.h
449 * define ICMP6_MAXTYPE
450 * as follows: n = ICMP6_MAXTYPE/32 + 1
451 * Actually is 203
452 */
453 } ipfw_insn_icmp6;
454
455 /*
456 * Here we have the structure representing an ipfw rule.
457 *
458 * It starts with a general area (with link fields and counters)
459 * followed by an array of one or more instructions, which the code
460 * accesses as an array of 32-bit values.
461 *
462 * Given a rule pointer r:
463 *
464 * r->cmd is the start of the first instruction.
465 * ACTION_PTR(r) is the start of the first action (things to do
466 * once a rule matched).
467 *
468 * When assembling instruction, remember the following:
469 *
470 * + if a rule has a "keep-state" (or "limit") option, then the
471 * first instruction (at r->cmd) MUST BE an O_PROBE_STATE
472 * + if a rule has a "log" option, then the first action
473 * (at ACTION_PTR(r)) MUST be O_LOG
474 * + if a rule has an "altq" option, it comes after "log"
475 * + if a rule has an O_TAG option, it comes after "log" and "altq"
476 *
477 * NOTE: we use a simple linked list of rules because we never need
478 * to delete a rule without scanning the list. We do not use
479 * queue(3) macros for portability and readability.
480 */
481
482 struct ip_fw {
483 struct ip_fw *x_next; /* linked list of rules */
484 struct ip_fw *next_rule; /* ptr to next [skipto] rule */
485 /* 'next_rule' is used to pass up 'set_disable' status */
486
487 uint16_t act_ofs; /* offset of action in 32-bit units */
488 uint16_t cmd_len; /* # of 32-bit words in cmd */
489 uint16_t rulenum; /* rule number */
490 uint8_t set; /* rule set (0..31) */
491 #define RESVD_SET 31 /* set for default and persistent rules */
492 uint8_t _pad; /* padding */
493 uint32_t id; /* rule id */
494
495 /* These fields are present in all rules. */
496 uint64_t pcnt; /* Packet counter */
497 uint64_t bcnt; /* Byte counter */
498 uint32_t timestamp; /* tv_sec of last match */
499
500 ipfw_insn cmd[1]; /* storage for commands */
501 };
502
503 #define ACTION_PTR(rule) \
504 (ipfw_insn *)( (u_int32_t *)((rule)->cmd) + ((rule)->act_ofs) )
505
506 #define RULESIZE(rule) (sizeof(struct ip_fw) + \
507 ((struct ip_fw *)(rule))->cmd_len * 4 - 4)
508
509 #if 1 // should be moved to in.h
510 /*
511 * This structure is used as a flow mask and a flow id for various
512 * parts of the code.
513 * addr_type is used in userland and kernel to mark the address type.
514 * fib is used in the kernel to record the fib in use.
515 * _flags is used in the kernel to store tcp flags for dynamic rules.
516 */
517 struct ipfw_flow_id {
518 uint32_t dst_ip;
519 uint32_t src_ip;
520 uint16_t dst_port;
521 uint16_t src_port;
522 uint8_t fib;
523 uint8_t proto;
524 uint8_t _flags; /* protocol-specific flags */
525 uint8_t addr_type; /* 4=ip4, 6=ip6, 1=ether ? */
526 struct in6_addr dst_ip6;
527 struct in6_addr src_ip6;
528 uint32_t flow_id6;
529 uint32_t extra; /* queue/pipe or frag_id */
530 };
531 #endif
532
533 #define IS_IP6_FLOW_ID(id) ((id)->addr_type == 6)
534
535 /*
536 * Dynamic ipfw rule.
537 */
538 typedef struct _ipfw_dyn_rule ipfw_dyn_rule;
539
540 struct _ipfw_dyn_rule {
541 ipfw_dyn_rule *next; /* linked list of rules. */
542 struct ip_fw *rule; /* pointer to rule */
543 /* 'rule' is used to pass up the rule number (from the parent) */
544
545 ipfw_dyn_rule *parent; /* pointer to parent rule */
546 u_int64_t pcnt; /* packet match counter */
547 u_int64_t bcnt; /* byte match counter */
548 struct ipfw_flow_id id; /* (masked) flow id */
549 u_int32_t expire; /* expire time */
550 u_int32_t bucket; /* which bucket in hash table */
551 u_int32_t state; /* state of this rule (typically a
552 * combination of TCP flags)
553 */
554 u_int32_t ack_fwd; /* most recent ACKs in forward */
555 u_int32_t ack_rev; /* and reverse directions (used */
556 /* to generate keepalives) */
557 u_int16_t dyn_type; /* rule type */
558 u_int16_t count; /* refcount */
559 };
560
561 /*
562 * Definitions for IP option names.
563 */
564 #define IP_FW_IPOPT_LSRR 0x01
565 #define IP_FW_IPOPT_SSRR 0x02
566 #define IP_FW_IPOPT_RR 0x04
567 #define IP_FW_IPOPT_TS 0x08
568
569 /*
570 * Definitions for TCP option names.
571 */
572 #define IP_FW_TCPOPT_MSS 0x01
573 #define IP_FW_TCPOPT_WINDOW 0x02
574 #define IP_FW_TCPOPT_SACK 0x04
575 #define IP_FW_TCPOPT_TS 0x08
576 #define IP_FW_TCPOPT_CC 0x10
577
578 #define ICMP_REJECT_RST 0x100 /* fake ICMP code (send a TCP RST) */
579 #define ICMP6_UNREACH_RST 0x100 /* fake ICMPv6 code (send a TCP RST) */
580
581 /*
582 * These are used for lookup tables.
583 */
584 typedef struct _ipfw_table_entry {
585 in_addr_t addr; /* network address */
586 u_int32_t value; /* value */
587 u_int16_t tbl; /* table number */
588 u_int8_t masklen; /* mask length */
589 } ipfw_table_entry;
590
591 typedef struct _ipfw_table {
592 u_int32_t size; /* size of entries in bytes */
593 u_int32_t cnt; /* # of entries */
594 u_int16_t tbl; /* table number */
595 ipfw_table_entry ent[0]; /* entries */
596 } ipfw_table;
597
598 #endif /* _IPFW2_H */
Cache object: 18b92453850925b63af104071c4b4ca9
|