2 * Copyright (c) 2002-2009 Luigi Rizzo, Universita` di Pisa
4 * Redistribution and use in source and binary forms, with or without
5 * modification, are permitted provided that the following conditions
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.
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
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
37 #define IPFW_DEFAULT_RULE 65535
40 * The number of ipfw tables. The maximum allowed table number is the
41 * (IPFW_TABLES_MAX - 1).
43 #define IPFW_TABLES_MAX 128
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.
55 #define IPFW_ARG_MIN 1
56 #define IPFW_ARG_MAX 65534
57 #define IP_FW_TABLEARG 65535 /* XXX should use 0 */
60 * The kernel representation of ipfw rules is made of a list of
61 * 'instructions' (for all practical purposes equivalent to BPF
62 * instructions), which specify which fields of the packet
63 * (or its metadata) should be analysed.
65 * Each instruction is stored in a structure which begins with
66 * "ipfw_insn", and can contain extra fields depending on the
67 * instruction type (listed below).
68 * Note that the code is written so that individual instructions
69 * have a size which is a multiple of 32 bits. This means that, if
70 * such structures contain pointers or other 64-bit entities,
71 * (there is just one instance now) they may end up unaligned on
72 * 64-bit architectures, so the must be handled with care.
74 * "enum ipfw_opcodes" are the opcodes supported. We can have up
75 * to 256 different opcodes. When adding new opcodes, they should
76 * be appended to the end of the opcode list before O_LAST_OPCODE,
77 * this will prevent the ABI from being broken, otherwise users
78 * will have to recompile ipfw(8) when they update the kernel.
81 enum ipfw_opcodes { /* arguments (4 byte each) */
84 O_IP_SRC, /* u32 = IP */
85 O_IP_SRC_MASK, /* ip = IP/mask */
86 O_IP_SRC_ME, /* none */
87 O_IP_SRC_SET, /* u32=base, arg1=len, bitmap */
89 O_IP_DST, /* u32 = IP */
90 O_IP_DST_MASK, /* ip = IP/mask */
91 O_IP_DST_ME, /* none */
92 O_IP_DST_SET, /* u32=base, arg1=len, bitmap */
94 O_IP_SRCPORT, /* (n)port list:mask 4 byte ea */
95 O_IP_DSTPORT, /* (n)port list:mask 4 byte ea */
96 O_PROTO, /* arg1=protocol */
98 O_MACADDR2, /* 2 mac addr:mask */
99 O_MAC_TYPE, /* same as srcport */
109 O_IPOPT, /* arg1 = 2*u8 bitmap */
110 O_IPLEN, /* arg1 = len */
111 O_IPID, /* arg1 = id */
113 O_IPTOS, /* arg1 = id */
114 O_IPPRECEDENCE, /* arg1 = precedence << 5 */
115 O_IPTTL, /* arg1 = TTL */
117 O_IPVER, /* arg1 = version */
118 O_UID, /* u32 = id */
119 O_GID, /* u32 = id */
120 O_ESTAB, /* none (tcp established) */
121 O_TCPFLAGS, /* arg1 = 2*u8 bitmap */
122 O_TCPWIN, /* arg1 = desired win */
123 O_TCPSEQ, /* u32 = desired seq. */
124 O_TCPACK, /* u32 = desired seq. */
125 O_ICMPTYPE, /* u32 = icmp bitmap */
126 O_TCPOPTS, /* arg1 = 2*u8 bitmap */
128 O_VERREVPATH, /* none */
129 O_VERSRCREACH, /* none */
131 O_PROBE_STATE, /* none */
132 O_KEEP_STATE, /* none */
133 O_LIMIT, /* ipfw_insn_limit */
134 O_LIMIT_PARENT, /* dyn_type, not an opcode. */
137 * These are really 'actions'.
140 O_LOG, /* ipfw_insn_log */
141 O_PROB, /* u32 = match probability */
143 O_CHECK_STATE, /* none */
146 O_REJECT, /* arg1=icmp arg (same as deny) */
148 O_SKIPTO, /* arg1=next rule number */
149 O_PIPE, /* arg1=pipe number */
150 O_QUEUE, /* arg1=queue number */
151 O_DIVERT, /* arg1=port number */
152 O_TEE, /* arg1=port number */
153 O_FORWARD_IP, /* fwd sockaddr */
154 O_FORWARD_MAC, /* fwd mac */
161 O_IPSEC, /* has ipsec history */
162 O_IP_SRC_LOOKUP, /* arg1=table number, u32=value */
163 O_IP_DST_LOOKUP, /* arg1=table number, u32=value */
164 O_ANTISPOOF, /* none */
165 O_JAIL, /* u32 = id */
166 O_ALTQ, /* u32 = altq classif. qid */
167 O_DIVERTED, /* arg1=bitmap (1:loop, 2:out) */
168 O_TCPDATALEN, /* arg1 = tcp data len */
169 O_IP6_SRC, /* address without mask */
170 O_IP6_SRC_ME, /* my addresses */
171 O_IP6_SRC_MASK, /* address with the mask */
175 O_FLOW6ID, /* for flow id tag in the ipv6 pkt */
176 O_ICMP6TYPE, /* icmp6 packet type filtering */
177 O_EXT_HDR, /* filtering for ipv6 extension header */
181 * actions for ng_ipfw
183 O_NETGRAPH, /* send to ng_ipfw */
184 O_NGTEE, /* copy to ng_ipfw */
188 O_UNREACH6, /* arg1=icmpv6 code arg (deny) */
190 O_TAG, /* arg1=tag number */
191 O_TAGGED, /* arg1=tag number */
193 O_SETFIB, /* arg1=FIB number */
194 O_FIB, /* arg1=FIB desired fib number */
196 O_LAST_OPCODE /* not an opcode! */
200 * The extension header are filtered only for presence using a bit
201 * vector with a flag for each header.
203 #define EXT_FRAGMENT 0x1
204 #define EXT_HOPOPTS 0x2
205 #define EXT_ROUTING 0x4
208 #define EXT_DSTOPTS 0x20
209 #define EXT_RTHDR0 0x40
210 #define EXT_RTHDR2 0x80
213 * Template for instructions.
215 * ipfw_insn is used for all instructions which require no operands,
216 * a single 16-bit value (arg1), or a couple of 8-bit values.
218 * For other instructions which require different/larger arguments
219 * we have derived structures, ipfw_insn_*.
221 * The size of the instruction (in 32-bit words) is in the low
222 * 6 bits of "len". The 2 remaining bits are used to implement
223 * NOT and OR on individual instructions. Given a type, you can
224 * compute the length to be put in "len" using F_INSN_SIZE(t)
226 * F_NOT negates the match result of the instruction.
228 * F_OR is used to build or blocks. By default, instructions
229 * are evaluated as part of a logical AND. An "or" block
230 * { X or Y or Z } contains F_OR set in all but the last
231 * instruction of the block. A match will cause the code
232 * to skip past the last instruction of the block.
234 * NOTA BENE: in a couple of places we assume that
235 * sizeof(ipfw_insn) == sizeof(u_int32_t)
236 * this needs to be fixed.
239 typedef struct _ipfw_insn { /* template for instructions */
241 u_int8_t len; /* number of 32-bit words */
244 #define F_LEN_MASK 0x3f
245 #define F_LEN(cmd) ((cmd)->len & F_LEN_MASK)
251 * The F_INSN_SIZE(type) computes the size, in 4-byte words, of
254 #define F_INSN_SIZE(t) ((sizeof (t))/sizeof(u_int32_t))
257 * This is used to store an array of 16-bit entries (ports etc.)
259 typedef struct _ipfw_insn_u16 {
261 u_int16_t ports[2]; /* there may be more */
265 * This is used to store an array of 32-bit entries
266 * (uid, single IPv4 addresses etc.)
268 typedef struct _ipfw_insn_u32 {
270 u_int32_t d[1]; /* one or more */
274 * This is used to store IP addr-mask pairs.
276 typedef struct _ipfw_insn_ip {
283 * This is used to forward to a given address (ip).
285 typedef struct _ipfw_insn_sa {
287 struct sockaddr_in sa;
291 * This is used for MAC addr-mask pairs.
293 typedef struct _ipfw_insn_mac {
295 u_char addr[12]; /* dst[6] + src[6] */
296 u_char mask[12]; /* dst[6] + src[6] */
300 * This is used for interface match rules (recv xx, xmit xx).
302 typedef struct _ipfw_insn_if {
312 * This is used for storing an altq queue id number.
314 typedef struct _ipfw_insn_altq {
320 * This is used for limit rules.
322 typedef struct _ipfw_insn_limit {
325 u_int8_t limit_mask; /* combination of DYN_* below */
326 #define DYN_SRC_ADDR 0x1
327 #define DYN_SRC_PORT 0x2
328 #define DYN_DST_ADDR 0x4
329 #define DYN_DST_PORT 0x8
331 u_int16_t conn_limit;
335 * This is used for log instructions.
337 typedef struct _ipfw_insn_log {
339 u_int32_t max_log; /* how many do we log -- 0 = all */
340 u_int32_t log_left; /* how many left to log */
344 * Data structures required by both ipfw(8) and ipfw(4) but not part of the
345 * management API are protected by IPFW_INTERNAL.
348 /* Server pool support (LSNAT). */
350 LIST_ENTRY(cfg_spool) _next; /* chain of spool instances */
356 /* Redirect modes id. */
357 #define REDIR_ADDR 0x01
358 #define REDIR_PORT 0x02
359 #define REDIR_PROTO 0x04
362 /* Nat redirect configuration. */
364 LIST_ENTRY(cfg_redir) _next; /* chain of redir instances */
365 u_int16_t mode; /* type of redirect mode */
366 struct in_addr laddr; /* local ip address */
367 struct in_addr paddr; /* public ip address */
368 struct in_addr raddr; /* remote ip address */
369 u_short lport; /* local port */
370 u_short pport; /* public port */
371 u_short rport; /* remote port */
372 u_short pport_cnt; /* number of public ports */
373 u_short rport_cnt; /* number of remote ports */
374 int proto; /* protocol: tcp/udp */
375 struct alias_link **alink;
376 /* num of entry in spool chain */
378 /* chain of spool instances */
379 LIST_HEAD(spool_chain, cfg_spool) spool_chain;
383 #define NAT_BUF_LEN 1024
386 /* Nat configuration data struct. */
388 /* chain of nat instances */
389 LIST_ENTRY(cfg_nat) _next;
391 struct in_addr ip; /* nat ip address */
392 char if_name[IF_NAMESIZE]; /* interface name */
393 int mode; /* aliasing mode */
394 struct libalias *lib; /* libalias instance */
395 /* number of entry in spool chain */
397 /* chain of redir instances */
398 LIST_HEAD(redir_chain, cfg_redir) redir_chain;
402 #define SOF_NAT sizeof(struct cfg_nat)
403 #define SOF_REDIR sizeof(struct cfg_redir)
404 #define SOF_SPOOL sizeof(struct cfg_spool)
407 typedef struct _ipfw_insn_nat {
412 /* Apply ipv6 mask on ipv6 addr */
413 #define APPLY_MASK(addr,mask) \
414 (addr)->__u6_addr.__u6_addr32[0] &= (mask)->__u6_addr.__u6_addr32[0]; \
415 (addr)->__u6_addr.__u6_addr32[1] &= (mask)->__u6_addr.__u6_addr32[1]; \
416 (addr)->__u6_addr.__u6_addr32[2] &= (mask)->__u6_addr.__u6_addr32[2]; \
417 (addr)->__u6_addr.__u6_addr32[3] &= (mask)->__u6_addr.__u6_addr32[3];
419 /* Structure for ipv6 */
420 typedef struct _ipfw_insn_ip6 {
422 struct in6_addr addr6;
423 struct in6_addr mask6;
426 /* Used to support icmp6 types */
427 typedef struct _ipfw_insn_icmp6 {
429 uint32_t d[7]; /* XXX This number si related to the netinet/icmp6.h
430 * define ICMP6_MAXTYPE
431 * as follows: n = ICMP6_MAXTYPE/32 + 1
437 * Here we have the structure representing an ipfw rule.
439 * It starts with a general area (with link fields and counters)
440 * followed by an array of one or more instructions, which the code
441 * accesses as an array of 32-bit values.
443 * Given a rule pointer r:
445 * r->cmd is the start of the first instruction.
446 * ACTION_PTR(r) is the start of the first action (things to do
447 * once a rule matched).
449 * When assembling instruction, remember the following:
451 * + if a rule has a "keep-state" (or "limit") option, then the
452 * first instruction (at r->cmd) MUST BE an O_PROBE_STATE
453 * + if a rule has a "log" option, then the first action
454 * (at ACTION_PTR(r)) MUST be O_LOG
455 * + if a rule has an "altq" option, it comes after "log"
456 * + if a rule has an O_TAG option, it comes after "log" and "altq"
458 * NOTE: we use a simple linked list of rules because we never need
459 * to delete a rule without scanning the list. We do not use
460 * queue(3) macros for portability and readability.
464 struct ip_fw *x_next; /* linked list of rules */
465 struct ip_fw *next_rule; /* ptr to next [skipto] rule */
466 /* 'next_rule' is used to pass up 'set_disable' status */
468 uint16_t act_ofs; /* offset of action in 32-bit units */
469 uint16_t cmd_len; /* # of 32-bit words in cmd */
470 uint16_t rulenum; /* rule number */
471 uint8_t set; /* rule set (0..31) */
472 #define RESVD_SET 31 /* set for default and persistent rules */
473 uint8_t _pad; /* padding */
474 uint32_t id; /* rule id */
476 /* These fields are present in all rules. */
477 uint64_t pcnt; /* Packet counter */
478 uint64_t bcnt; /* Byte counter */
479 uint32_t timestamp; /* tv_sec of last match */
481 ipfw_insn cmd[1]; /* storage for commands */
484 #define ACTION_PTR(rule) \
485 (ipfw_insn *)( (u_int32_t *)((rule)->cmd) + ((rule)->act_ofs) )
487 #define RULESIZE(rule) (sizeof(struct ip_fw) + \
488 ((struct ip_fw *)(rule))->cmd_len * 4 - 4)
490 #if 1 // should be moved to in.h
492 * This structure is used as a flow mask and a flow id for various
494 * addr_type is used in userland and kernel to mark the address type.
495 * fib is used in the kernel to record the fib in use.
496 * _flags is used in the kernel to store tcp flags for dynamic rules.
498 struct ipfw_flow_id {
505 uint8_t _flags; /* protocol-specific flags */
506 uint8_t addr_type; /* 4=ip4, 6=ip6, 1=ether ? */
507 struct in6_addr dst_ip6;
508 struct in6_addr src_ip6;
510 uint32_t extra; /* queue/pipe or frag_id */
514 #define IS_IP6_FLOW_ID(id) ((id)->addr_type == 6)
519 typedef struct _ipfw_dyn_rule ipfw_dyn_rule;
521 struct _ipfw_dyn_rule {
522 ipfw_dyn_rule *next; /* linked list of rules. */
523 struct ip_fw *rule; /* pointer to rule */
524 /* 'rule' is used to pass up the rule number (from the parent) */
526 ipfw_dyn_rule *parent; /* pointer to parent rule */
527 u_int64_t pcnt; /* packet match counter */
528 u_int64_t bcnt; /* byte match counter */
529 struct ipfw_flow_id id; /* (masked) flow id */
530 u_int32_t expire; /* expire time */
531 u_int32_t bucket; /* which bucket in hash table */
532 u_int32_t state; /* state of this rule (typically a
533 * combination of TCP flags)
535 u_int32_t ack_fwd; /* most recent ACKs in forward */
536 u_int32_t ack_rev; /* and reverse directions (used */
537 /* to generate keepalives) */
538 u_int16_t dyn_type; /* rule type */
539 u_int16_t count; /* refcount */
543 * Definitions for IP option names.
545 #define IP_FW_IPOPT_LSRR 0x01
546 #define IP_FW_IPOPT_SSRR 0x02
547 #define IP_FW_IPOPT_RR 0x04
548 #define IP_FW_IPOPT_TS 0x08
551 * Definitions for TCP option names.
553 #define IP_FW_TCPOPT_MSS 0x01
554 #define IP_FW_TCPOPT_WINDOW 0x02
555 #define IP_FW_TCPOPT_SACK 0x04
556 #define IP_FW_TCPOPT_TS 0x08
557 #define IP_FW_TCPOPT_CC 0x10
559 #define ICMP_REJECT_RST 0x100 /* fake ICMP code (send a TCP RST) */
560 #define ICMP6_UNREACH_RST 0x100 /* fake ICMPv6 code (send a TCP RST) */
563 * These are used for lookup tables.
565 typedef struct _ipfw_table_entry {
566 in_addr_t addr; /* network address */
567 u_int32_t value; /* value */
568 u_int16_t tbl; /* table number */
569 u_int8_t masklen; /* mask length */
572 typedef struct _ipfw_table {
573 u_int32_t size; /* size of entries in bytes */
574 u_int32_t cnt; /* # of entries */
575 u_int16_t tbl; /* table number */
576 ipfw_table_entry ent[0]; /* entries */
579 #endif /* _IPFW2_H */