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 * Default number of ipfw tables.
42 #define IPFW_TABLES_MAX 65535
43 #define IPFW_TABLES_DEFAULT 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 * Number of entries in the call stack of the call/return commands.
61 * Call stack currently is an uint16_t array with rule numbers.
63 #define IPFW_CALLSTACK_SIZE 16
65 /* IP_FW3 header/opcodes */
66 typedef struct _ip_fw3_opheader {
67 uint16_t opcode; /* Operation opcode */
68 uint16_t reserved[3]; /* Align to 64-bit boundary */
72 /* IPFW extented tables support */
73 #define IP_FW_TABLE_XADD 86 /* add entry */
74 #define IP_FW_TABLE_XDEL 87 /* delete entry */
75 #define IP_FW_TABLE_XGETSIZE 88 /* get table size */
76 #define IP_FW_TABLE_XLIST 89 /* list table contents */
79 * The kernel representation of ipfw rules is made of a list of
80 * 'instructions' (for all practical purposes equivalent to BPF
81 * instructions), which specify which fields of the packet
82 * (or its metadata) should be analysed.
84 * Each instruction is stored in a structure which begins with
85 * "ipfw_insn", and can contain extra fields depending on the
86 * instruction type (listed below).
87 * Note that the code is written so that individual instructions
88 * have a size which is a multiple of 32 bits. This means that, if
89 * such structures contain pointers or other 64-bit entities,
90 * (there is just one instance now) they may end up unaligned on
91 * 64-bit architectures, so the must be handled with care.
93 * "enum ipfw_opcodes" are the opcodes supported. We can have up
94 * to 256 different opcodes. When adding new opcodes, they should
95 * be appended to the end of the opcode list before O_LAST_OPCODE,
96 * this will prevent the ABI from being broken, otherwise users
97 * will have to recompile ipfw(8) when they update the kernel.
100 enum ipfw_opcodes { /* arguments (4 byte each) */
103 O_IP_SRC, /* u32 = IP */
104 O_IP_SRC_MASK, /* ip = IP/mask */
105 O_IP_SRC_ME, /* none */
106 O_IP_SRC_SET, /* u32=base, arg1=len, bitmap */
108 O_IP_DST, /* u32 = IP */
109 O_IP_DST_MASK, /* ip = IP/mask */
110 O_IP_DST_ME, /* none */
111 O_IP_DST_SET, /* u32=base, arg1=len, bitmap */
113 O_IP_SRCPORT, /* (n)port list:mask 4 byte ea */
114 O_IP_DSTPORT, /* (n)port list:mask 4 byte ea */
115 O_PROTO, /* arg1=protocol */
117 O_MACADDR2, /* 2 mac addr:mask */
118 O_MAC_TYPE, /* same as srcport */
128 O_IPOPT, /* arg1 = 2*u8 bitmap */
129 O_IPLEN, /* arg1 = len */
130 O_IPID, /* arg1 = id */
132 O_IPTOS, /* arg1 = id */
133 O_IPPRECEDENCE, /* arg1 = precedence << 5 */
134 O_IPTTL, /* arg1 = TTL */
136 O_IPVER, /* arg1 = version */
137 O_UID, /* u32 = id */
138 O_GID, /* u32 = id */
139 O_ESTAB, /* none (tcp established) */
140 O_TCPFLAGS, /* arg1 = 2*u8 bitmap */
141 O_TCPWIN, /* arg1 = desired win */
142 O_TCPSEQ, /* u32 = desired seq. */
143 O_TCPACK, /* u32 = desired seq. */
144 O_ICMPTYPE, /* u32 = icmp bitmap */
145 O_TCPOPTS, /* arg1 = 2*u8 bitmap */
147 O_VERREVPATH, /* none */
148 O_VERSRCREACH, /* none */
150 O_PROBE_STATE, /* none */
151 O_KEEP_STATE, /* none */
152 O_LIMIT, /* ipfw_insn_limit */
153 O_LIMIT_PARENT, /* dyn_type, not an opcode. */
156 * These are really 'actions'.
159 O_LOG, /* ipfw_insn_log */
160 O_PROB, /* u32 = match probability */
162 O_CHECK_STATE, /* none */
165 O_REJECT, /* arg1=icmp arg (same as deny) */
167 O_SKIPTO, /* arg1=next rule number */
168 O_PIPE, /* arg1=pipe number */
169 O_QUEUE, /* arg1=queue number */
170 O_DIVERT, /* arg1=port number */
171 O_TEE, /* arg1=port number */
172 O_FORWARD_IP, /* fwd sockaddr */
173 O_FORWARD_MAC, /* fwd mac */
180 O_IPSEC, /* has ipsec history */
181 O_IP_SRC_LOOKUP, /* arg1=table number, u32=value */
182 O_IP_DST_LOOKUP, /* arg1=table number, u32=value */
183 O_ANTISPOOF, /* none */
184 O_JAIL, /* u32 = id */
185 O_ALTQ, /* u32 = altq classif. qid */
186 O_DIVERTED, /* arg1=bitmap (1:loop, 2:out) */
187 O_TCPDATALEN, /* arg1 = tcp data len */
188 O_IP6_SRC, /* address without mask */
189 O_IP6_SRC_ME, /* my addresses */
190 O_IP6_SRC_MASK, /* address with the mask */
194 O_FLOW6ID, /* for flow id tag in the ipv6 pkt */
195 O_ICMP6TYPE, /* icmp6 packet type filtering */
196 O_EXT_HDR, /* filtering for ipv6 extension header */
200 * actions for ng_ipfw
202 O_NETGRAPH, /* send to ng_ipfw */
203 O_NGTEE, /* copy to ng_ipfw */
207 O_UNREACH6, /* arg1=icmpv6 code arg (deny) */
209 O_TAG, /* arg1=tag number */
210 O_TAGGED, /* arg1=tag number */
212 O_SETFIB, /* arg1=FIB number */
213 O_FIB, /* arg1=FIB desired fib number */
215 O_SOCKARG, /* socket argument */
217 O_CALLRETURN, /* arg1=called rule number */
219 O_FORWARD_IP6, /* fwd sockaddr_in6 */
221 O_DSCP, /* 2 u32 = DSCP mask */
222 O_SETDSCP, /* arg1=DSCP value */
224 O_LAST_OPCODE /* not an opcode! */
229 * The extension header are filtered only for presence using a bit
230 * vector with a flag for each header.
232 #define EXT_FRAGMENT 0x1
233 #define EXT_HOPOPTS 0x2
234 #define EXT_ROUTING 0x4
237 #define EXT_DSTOPTS 0x20
238 #define EXT_RTHDR0 0x40
239 #define EXT_RTHDR2 0x80
242 * Template for instructions.
244 * ipfw_insn is used for all instructions which require no operands,
245 * a single 16-bit value (arg1), or a couple of 8-bit values.
247 * For other instructions which require different/larger arguments
248 * we have derived structures, ipfw_insn_*.
250 * The size of the instruction (in 32-bit words) is in the low
251 * 6 bits of "len". The 2 remaining bits are used to implement
252 * NOT and OR on individual instructions. Given a type, you can
253 * compute the length to be put in "len" using F_INSN_SIZE(t)
255 * F_NOT negates the match result of the instruction.
257 * F_OR is used to build or blocks. By default, instructions
258 * are evaluated as part of a logical AND. An "or" block
259 * { X or Y or Z } contains F_OR set in all but the last
260 * instruction of the block. A match will cause the code
261 * to skip past the last instruction of the block.
263 * NOTA BENE: in a couple of places we assume that
264 * sizeof(ipfw_insn) == sizeof(u_int32_t)
265 * this needs to be fixed.
268 typedef struct _ipfw_insn { /* template for instructions */
270 u_int8_t len; /* number of 32-bit words */
273 #define F_LEN_MASK 0x3f
274 #define F_LEN(cmd) ((cmd)->len & F_LEN_MASK)
280 * The F_INSN_SIZE(type) computes the size, in 4-byte words, of
283 #define F_INSN_SIZE(t) ((sizeof (t))/sizeof(u_int32_t))
286 * This is used to store an array of 16-bit entries (ports etc.)
288 typedef struct _ipfw_insn_u16 {
290 u_int16_t ports[2]; /* there may be more */
294 * This is used to store an array of 32-bit entries
295 * (uid, single IPv4 addresses etc.)
297 typedef struct _ipfw_insn_u32 {
299 u_int32_t d[1]; /* one or more */
303 * This is used to store IP addr-mask pairs.
305 typedef struct _ipfw_insn_ip {
312 * This is used to forward to a given address (ip).
314 typedef struct _ipfw_insn_sa {
316 struct sockaddr_in sa;
320 * This is used to forward to a given address (ipv6).
322 typedef struct _ipfw_insn_sa6 {
324 struct sockaddr_in6 sa;
328 * This is used for MAC addr-mask pairs.
330 typedef struct _ipfw_insn_mac {
332 u_char addr[12]; /* dst[6] + src[6] */
333 u_char mask[12]; /* dst[6] + src[6] */
337 * This is used for interface match rules (recv xx, xmit xx).
339 typedef struct _ipfw_insn_if {
349 * This is used for storing an altq queue id number.
351 typedef struct _ipfw_insn_altq {
357 * This is used for limit rules.
359 typedef struct _ipfw_insn_limit {
362 u_int8_t limit_mask; /* combination of DYN_* below */
363 #define DYN_SRC_ADDR 0x1
364 #define DYN_SRC_PORT 0x2
365 #define DYN_DST_ADDR 0x4
366 #define DYN_DST_PORT 0x8
368 u_int16_t conn_limit;
372 * This is used for log instructions.
374 typedef struct _ipfw_insn_log {
376 u_int32_t max_log; /* how many do we log -- 0 = all */
377 u_int32_t log_left; /* how many left to log */
381 * Data structures required by both ipfw(8) and ipfw(4) but not part of the
382 * management API are protected by IPFW_INTERNAL.
385 /* Server pool support (LSNAT). */
387 LIST_ENTRY(cfg_spool) _next; /* chain of spool instances */
393 /* Redirect modes id. */
394 #define REDIR_ADDR 0x01
395 #define REDIR_PORT 0x02
396 #define REDIR_PROTO 0x04
399 /* Nat redirect configuration. */
401 LIST_ENTRY(cfg_redir) _next; /* chain of redir instances */
402 u_int16_t mode; /* type of redirect mode */
403 struct in_addr laddr; /* local ip address */
404 struct in_addr paddr; /* public ip address */
405 struct in_addr raddr; /* remote ip address */
406 u_short lport; /* local port */
407 u_short pport; /* public port */
408 u_short rport; /* remote port */
409 u_short pport_cnt; /* number of public ports */
410 u_short rport_cnt; /* number of remote ports */
411 int proto; /* protocol: tcp/udp */
412 struct alias_link **alink;
413 /* num of entry in spool chain */
415 /* chain of spool instances */
416 LIST_HEAD(spool_chain, cfg_spool) spool_chain;
421 /* Nat configuration data struct. */
423 /* chain of nat instances */
424 LIST_ENTRY(cfg_nat) _next;
426 struct in_addr ip; /* nat ip address */
427 char if_name[IF_NAMESIZE]; /* interface name */
428 int mode; /* aliasing mode */
429 struct libalias *lib; /* libalias instance */
430 /* number of entry in spool chain */
432 /* chain of redir instances */
433 LIST_HEAD(redir_chain, cfg_redir) redir_chain;
437 #define SOF_NAT sizeof(struct cfg_nat)
438 #define SOF_REDIR sizeof(struct cfg_redir)
439 #define SOF_SPOOL sizeof(struct cfg_spool)
442 typedef struct _ipfw_insn_nat {
447 /* Apply ipv6 mask on ipv6 addr */
448 #define APPLY_MASK(addr,mask) \
449 (addr)->__u6_addr.__u6_addr32[0] &= (mask)->__u6_addr.__u6_addr32[0]; \
450 (addr)->__u6_addr.__u6_addr32[1] &= (mask)->__u6_addr.__u6_addr32[1]; \
451 (addr)->__u6_addr.__u6_addr32[2] &= (mask)->__u6_addr.__u6_addr32[2]; \
452 (addr)->__u6_addr.__u6_addr32[3] &= (mask)->__u6_addr.__u6_addr32[3];
454 /* Structure for ipv6 */
455 typedef struct _ipfw_insn_ip6 {
457 struct in6_addr addr6;
458 struct in6_addr mask6;
461 /* Used to support icmp6 types */
462 typedef struct _ipfw_insn_icmp6 {
464 uint32_t d[7]; /* XXX This number si related to the netinet/icmp6.h
465 * define ICMP6_MAXTYPE
466 * as follows: n = ICMP6_MAXTYPE/32 + 1
472 * Here we have the structure representing an ipfw rule.
474 * It starts with a general area (with link fields and counters)
475 * followed by an array of one or more instructions, which the code
476 * accesses as an array of 32-bit values.
478 * Given a rule pointer r:
480 * r->cmd is the start of the first instruction.
481 * ACTION_PTR(r) is the start of the first action (things to do
482 * once a rule matched).
484 * When assembling instruction, remember the following:
486 * + if a rule has a "keep-state" (or "limit") option, then the
487 * first instruction (at r->cmd) MUST BE an O_PROBE_STATE
488 * + if a rule has a "log" option, then the first action
489 * (at ACTION_PTR(r)) MUST be O_LOG
490 * + if a rule has an "altq" option, it comes after "log"
491 * + if a rule has an O_TAG option, it comes after "log" and "altq"
493 * NOTE: we use a simple linked list of rules because we never need
494 * to delete a rule without scanning the list. We do not use
495 * queue(3) macros for portability and readability.
499 struct ip_fw *x_next; /* linked list of rules */
500 struct ip_fw *next_rule; /* ptr to next [skipto] rule */
501 /* 'next_rule' is used to pass up 'set_disable' status */
503 uint16_t act_ofs; /* offset of action in 32-bit units */
504 uint16_t cmd_len; /* # of 32-bit words in cmd */
505 uint16_t rulenum; /* rule number */
506 uint8_t set; /* rule set (0..31) */
507 #define RESVD_SET 31 /* set for default and persistent rules */
508 uint8_t _pad; /* padding */
509 uint32_t id; /* rule id */
511 /* These fields are present in all rules. */
512 uint64_t pcnt; /* Packet counter */
513 uint64_t bcnt; /* Byte counter */
514 uint32_t timestamp; /* tv_sec of last match */
516 ipfw_insn cmd[1]; /* storage for commands */
519 #define ACTION_PTR(rule) \
520 (ipfw_insn *)( (u_int32_t *)((rule)->cmd) + ((rule)->act_ofs) )
522 #define RULESIZE(rule) (sizeof(struct ip_fw) + \
523 ((struct ip_fw *)(rule))->cmd_len * 4 - 4)
525 #if 1 // should be moved to in.h
527 * This structure is used as a flow mask and a flow id for various
529 * addr_type is used in userland and kernel to mark the address type.
530 * fib is used in the kernel to record the fib in use.
531 * _flags is used in the kernel to store tcp flags for dynamic rules.
533 struct ipfw_flow_id {
540 uint8_t _flags; /* protocol-specific flags */
541 uint8_t addr_type; /* 4=ip4, 6=ip6, 1=ether ? */
542 struct in6_addr dst_ip6;
543 struct in6_addr src_ip6;
545 uint32_t extra; /* queue/pipe or frag_id */
549 #define IS_IP6_FLOW_ID(id) ((id)->addr_type == 6)
554 typedef struct _ipfw_dyn_rule ipfw_dyn_rule;
556 struct _ipfw_dyn_rule {
557 ipfw_dyn_rule *next; /* linked list of rules. */
558 struct ip_fw *rule; /* pointer to rule */
559 /* 'rule' is used to pass up the rule number (from the parent) */
561 ipfw_dyn_rule *parent; /* pointer to parent rule */
562 u_int64_t pcnt; /* packet match counter */
563 u_int64_t bcnt; /* byte match counter */
564 struct ipfw_flow_id id; /* (masked) flow id */
565 u_int32_t expire; /* expire time */
566 u_int32_t bucket; /* which bucket in hash table */
567 u_int32_t state; /* state of this rule (typically a
568 * combination of TCP flags)
570 u_int32_t ack_fwd; /* most recent ACKs in forward */
571 u_int32_t ack_rev; /* and reverse directions (used */
572 /* to generate keepalives) */
573 u_int16_t dyn_type; /* rule type */
574 u_int16_t count; /* refcount */
578 * Definitions for IP option names.
580 #define IP_FW_IPOPT_LSRR 0x01
581 #define IP_FW_IPOPT_SSRR 0x02
582 #define IP_FW_IPOPT_RR 0x04
583 #define IP_FW_IPOPT_TS 0x08
586 * Definitions for TCP option names.
588 #define IP_FW_TCPOPT_MSS 0x01
589 #define IP_FW_TCPOPT_WINDOW 0x02
590 #define IP_FW_TCPOPT_SACK 0x04
591 #define IP_FW_TCPOPT_TS 0x08
592 #define IP_FW_TCPOPT_CC 0x10
594 #define ICMP_REJECT_RST 0x100 /* fake ICMP code (send a TCP RST) */
595 #define ICMP6_UNREACH_RST 0x100 /* fake ICMPv6 code (send a TCP RST) */
598 * These are used for lookup tables.
601 #define IPFW_TABLE_CIDR 1 /* Table for holding IPv4/IPv6 prefixes */
602 #define IPFW_TABLE_INTERFACE 2 /* Table for holding interface names */
603 #define IPFW_TABLE_MAXTYPE 2 /* Maximum valid number */
605 typedef struct _ipfw_table_entry {
606 in_addr_t addr; /* network address */
607 u_int32_t value; /* value */
608 u_int16_t tbl; /* table number */
609 u_int8_t masklen; /* mask length */
612 typedef struct _ipfw_table_xentry {
613 uint16_t len; /* Total entry length */
614 uint8_t type; /* entry type */
615 uint8_t masklen; /* mask length */
616 uint16_t tbl; /* table number */
617 uint16_t flags; /* record flags */
618 uint32_t value; /* value */
620 /* Longest field needs to be aligned by 4-byte boundary */
621 struct in6_addr addr6; /* IPv6 address */
622 char iface[IF_NAMESIZE]; /* interface name */
625 #define IPFW_TCF_INET 0x01 /* CIDR flags: IPv4 record */
627 typedef struct _ipfw_table {
628 u_int32_t size; /* size of entries in bytes */
629 u_int32_t cnt; /* # of entries */
630 u_int16_t tbl; /* table number */
631 ipfw_table_entry ent[0]; /* entries */
634 typedef struct _ipfw_xtable {
635 ip_fw3_opheader opheader; /* eXtended tables are controlled via IP_FW3 */
636 uint32_t size; /* size of entries in bytes */
637 uint32_t cnt; /* # of entries */
638 uint16_t tbl; /* table number */
639 uint8_t type; /* table type */
640 ipfw_table_xentry xent[0]; /* entries */
643 #endif /* _IPFW2_H */