bsd/netinet/ip_fw2.h

   1 /*
   2  * Copyright (c) 2002 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: src/sys/netinet/ip_fw2.h,v 1.1.2.4 2003/07/17 06:03:39 luigi Exp $
  26  */
  27
  28 #ifndef _IPFW2_H
  29 #define _IPFW2_H
  30
  31 /*
  32  * Define IP Firewall event subclass, and associated events.
  33  */
  34
  35 /*!
  36         @defined KEV_IPFW_SUBCLASS
  37         @discussion The kernel event subclass for IP Firewall.
  38 */
  39 #define KEV_IPFW_SUBCLASS       1
  40
  41 /*!
  42         @defined KEV_IPFW_ADD
  43         @discussion The event code indicating a rule has been added.
  44 */
  45 #define KEV_IPFW_ADD            1
  46
  47 /*!
  48         @defined KEV_IPFW_DEL
  49         @discussion The event code indicating a rule has been removed.
  50 */
  51 #define KEV_IPFW_DEL            2
  52
  53 /*!
  54         @defined KEV_IPFW_FLUSH
  55         @discussion The event code indicating the rule set has been flushed.
  56 */
  57 #define KEV_IPFW_FLUSH          3
  58
  59 /*!
  60         @defined KEV_IPFW_ENABLE
  61         @discussion The event code indicating the enable flag has been changed
  62 */
  63 #define KEV_IPFW_ENABLE         4
  64
  65
  66
  67 #if !__LP64__
  68
  69 /*
  70  * The kernel representation of ipfw rules is made of a list of
  71  * 'instructions' (for all practical purposes equivalent to BPF
  72  * instructions), which specify which fields of the packet
  73  * (or its metadata) should be analysed.
  74  *
  75  * Each instruction is stored in a structure which begins with
  76  * "ipfw_insn", and can contain extra fields depending on the
  77  * instruction type (listed below).
  78  * Note that the code is written so that individual instructions
  79  * have a size which is a multiple of 32 bits. This means that, if
  80  * such structures contain pointers or other 64-bit entities,
  81  * (there is just one instance now) they may end up unaligned on
  82  * 64-bit architectures, so the must be handled with care.
  83  *
  84  * "enum ipfw_opcodes" are the opcodes supported. We can have up
  85  * to 256 different opcodes.
  86  */
  87
  88 enum ipfw_opcodes {             /* arguments (4 byte each)      */
  89         O_NOP,
  90
  91         O_IP_SRC,               /* u32 = IP                     */
  92         O_IP_SRC_MASK,          /* ip = IP/mask                 */
  93         O_IP_SRC_ME,            /* none                         */
  94         O_IP_SRC_SET,           /* u32=base, arg1=len, bitmap   */
  95
  96         O_IP_DST,               /* u32 = IP                     */
  97         O_IP_DST_MASK,          /* ip = IP/mask                 */
  98         O_IP_DST_ME,            /* none                         */
  99         O_IP_DST_SET,           /* u32=base, arg1=len, bitmap   */
 100
 101         O_IP_SRCPORT,           /* (n)port list:mask 4 byte ea  */
 102         O_IP_DSTPORT,           /* (n)port list:mask 4 byte ea  */
 103         O_PROTO,                /* arg1=protocol                */
 104
 105         O_MACADDR2,             /* 2 mac addr:mask              */
 106         O_MAC_TYPE,             /* same as srcport              */
 107
 108         O_LAYER2,               /* none                         */
 109         O_IN,                   /* none                         */
 110         O_FRAG,                 /* none                         */
 111
 112         O_RECV,                 /* none                         */
 113         O_XMIT,                 /* none                         */
 114         O_VIA,                  /* none                         */
 115
 116         O_IPOPT,                /* arg1 = 2*u8 bitmap           */
 117         O_IPLEN,                /* arg1 = len                   */
 118         O_IPID,                 /* arg1 = id                    */
 119
 120         O_IPTOS,                /* arg1 = id                    */
 121         O_IPPRECEDENCE,         /* arg1 = precedence << 5       */
 122         O_IPTTL,                /* arg1 = TTL                   */
 123
 124         O_IPVER,                /* arg1 = version               */
 125         O_UID,                  /* u32 = id                     */
 126         O_GID,                  /* u32 = id                     */
 127         O_ESTAB,                /* none (tcp established)       */
 128         O_TCPFLAGS,             /* arg1 = 2*u8 bitmap           */
 129         O_TCPWIN,               /* arg1 = desired win           */
 130         O_TCPSEQ,               /* u32 = desired seq.           */
 131         O_TCPACK,               /* u32 = desired seq.           */
 132         O_ICMPTYPE,             /* u32 = icmp bitmap            */
 133         O_TCPOPTS,              /* arg1 = 2*u8 bitmap           */
 134
 135         O_VERREVPATH,           /* 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
 162         /*
 163          * More opcodes.
 164          */
 165         O_IPSEC,                /* has ipsec history            */
 166
 167         O_LAST_OPCODE           /* not an opcode!               */
 168 };
 169
 170 /*
 171  * Template for instructions.
 172  *
 173  * ipfw_insn is used for all instructions which require no operands,
 174  * a single 16-bit value (arg1), or a couple of 8-bit values.
 175  *
 176  * For other instructions which require different/larger arguments
 177  * we have derived structures, ipfw_insn_*.
 178  *
 179  * The size of the instruction (in 32-bit words) is in the low
 180  * 6 bits of "len". The 2 remaining bits are used to implement
 181  * NOT and OR on individual instructions. Given a type, you can
 182  * compute the length to be put in "len" using F_INSN_SIZE(t)
 183  *
 184  * F_NOT        negates the match result of the instruction.
 185  *
 186  * F_OR         is used to build or blocks. By default, instructions
 187  *              are evaluated as part of a logical AND. An "or" block
 188  *              { X or Y or Z } contains F_OR set in all but the last
 189  *              instruction of the block. A match will cause the code
 190  *              to skip past the last instruction of the block.
 191  *
 192  * NOTA BENE: in a couple of places we assume that
 193  *      sizeof(ipfw_insn) == sizeof(u_int32_t)
 194  * this needs to be fixed.
 195  *
 196  */
 197 typedef struct  _ipfw_insn {    /* template for instructions */
 198         enum ipfw_opcodes       opcode:8;
 199         u_int8_t        len;    /* numer of 32-byte words */
 200 #define F_NOT           0x80
 201 #define F_OR            0x40
 202 #define F_LEN_MASK      0x3f
 203 #define F_LEN(cmd)      ((cmd)->len & F_LEN_MASK)
 204
 205         u_int16_t       arg1;
 206 } ipfw_insn;
 207
 208 /*
 209  * The F_INSN_SIZE(type) computes the size, in 4-byte words, of
 210  * a given type.
 211  */
 212 #define F_INSN_SIZE(t)  ((sizeof (t))/sizeof(u_int32_t))
 213
 214 /*
 215  * This is used to store an array of 16-bit entries (ports etc.)
 216  */
 217 typedef struct  _ipfw_insn_u16 {
 218         ipfw_insn o;
 219         u_int16_t ports[2];     /* there may be more */
 220 } ipfw_insn_u16;
 221
 222 /*
 223  * This is used to store an array of 32-bit entries
 224  * (uid, single IPv4 addresses etc.)
 225  */
 226 typedef struct  _ipfw_insn_u32 {
 227         ipfw_insn o;
 228         u_int32_t d[1]; /* one or more */
 229 } ipfw_insn_u32;
 230
 231 /*
 232  * This is used to store IP addr-mask pairs.
 233  */
 234 typedef struct  _ipfw_insn_ip {
 235         ipfw_insn o;
 236         struct in_addr  addr;
 237         struct in_addr  mask;
 238 } ipfw_insn_ip;
 239
 240 /*
 241  * This is used to forward to a given address (ip).
 242  */
 243 typedef struct  _ipfw_insn_sa {
 244         ipfw_insn o;
 245         struct sockaddr_in sa;
 246 } ipfw_insn_sa;
 247
 248 /*
 249  * This is used for MAC addr-mask pairs.
 250  */
 251 typedef struct  _ipfw_insn_mac {
 252         ipfw_insn o;
 253         u_char addr[12];        /* dst[6] + src[6] */
 254         u_char mask[12];        /* dst[6] + src[6] */
 255 } ipfw_insn_mac;
 256
 257 /*
 258  * This is used for interface match rules (recv xx, xmit xx).
 259  */
 260 typedef struct  _ipfw_insn_if {
 261         ipfw_insn o;
 262         union {
 263                 struct in_addr ip;
 264                 int32_t unit;
 265         } p;
 266         char name[IFNAMSIZ];
 267 } ipfw_insn_if;
 268
 269 /*
 270  * This is used for pipe and queue actions, which need to store
 271  * a single pointer (which can have different size on different
 272  * architectures.
 273  * Note that, because of previous instructions, pipe_ptr might
 274  * be unaligned in the overall structure, so it needs to be
 275  * manipulated with care.
 276  */
 277 typedef struct  _ipfw_insn_pipe {
 278         ipfw_insn       o;
 279         void            *pipe_ptr;      /* XXX */
 280 } ipfw_insn_pipe;
 281
 282 /*
 283  * This is used for limit rules.
 284  */
 285 typedef struct  _ipfw_insn_limit {
 286         ipfw_insn o;
 287         u_int8_t _pad;
 288         u_int8_t limit_mask;    /* combination of DYN_* below   */
 289 #define DYN_SRC_ADDR    0x1
 290 #define DYN_SRC_PORT    0x2
 291 #define DYN_DST_ADDR    0x4
 292 #define DYN_DST_PORT    0x8
 293
 294         u_int16_t conn_limit;
 295 } ipfw_insn_limit;
 296
 297 /*
 298  * This is used for log instructions.
 299  */
 300 typedef struct  _ipfw_insn_log {
 301         ipfw_insn o;
 302         u_int32_t max_log;      /* how many do we log -- 0 = all */
 303         u_int32_t log_left;     /* how many left to log         */
 304 } ipfw_insn_log;
 305
 306 /* Version of this API */
 307 #define IP_FW_VERSION_NONE 0
 308 #define IP_FW_VERSION_0 10      /* old ipfw */
 309 #define IP_FW_VERSION_1 20      /* ipfw in Jaguar/Panther */
 310 #define IP_FW_VERSION_2 30      /* ipfw2 */
 311 #define IP_FW_CURRENT_API_VERSION IP_FW_VERSION_2
 312
 313 /*
 314  * Here we have the structure representing an ipfw rule.
 315  *
 316  * It starts with a general area (with link fields and counters)
 317  * followed by an array of one or more instructions, which the code
 318  * accesses as an array of 32-bit values.
 319  *
 320  * Given a rule pointer  r:
 321  *
 322  *  r->cmd              is the start of the first instruction.
 323  *  ACTION_PTR(r)       is the start of the first action (things to do
 324  *                      once a rule matched).
 325  *
 326  * When assembling instruction, remember the following:
 327  *
 328  *  + if a rule has a "keep-state" (or "limit") option, then the
 329  *      first instruction (at r->cmd) MUST BE an O_PROBE_STATE
 330  *  + if a rule has a "log" option, then the first action
 331  *      (at ACTION_PTR(r)) MUST be O_LOG
 332  *
 333  * NOTE: we use a simple linked list of rules because we never need
 334  *      to delete a rule without scanning the list. We do not use
 335  *      queue(3) macros for portability and readability.
 336  */
 337
 338 struct ip_fw {
 339         u_int32_t version;              /* Version of this structure. MUST be set */
 340                                                         /* by clients. Should always be */
 341                                                         /* set to IP_FW_CURRENT_API_VERSION. */
 342         void *context;                  /* Context that is usable by user processes to */
 343                                                         /* identify this rule. */
 344         struct ip_fw    *next;          /* linked list of rules         */
 345         struct ip_fw    *next_rule;     /* ptr to next [skipto] rule    */
 346         /* 'next_rule' is used to pass up 'set_disable' status          */
 347
 348         u_int16_t       act_ofs;        /* offset of action in 32-bit units */
 349         u_int16_t       cmd_len;        /* # of 32-bit words in cmd     */
 350         u_int16_t       rulenum;        /* rule number                  */
 351         u_int8_t        set;            /* rule set (0..31)             */
 352         u_int32_t       set_masks[2];   /* masks for manipulating sets atomically */
 353 #define RESVD_SET       31      /* set for default and persistent rules */
 354         u_int8_t        _pad;           /* padding                      */
 355
 356         /* These fields are present in all rules.                       */
 357         u_int64_t       pcnt;           /* Packet counter               */
 358         u_int64_t       bcnt;           /* Byte counter                 */
 359         u_int32_t       timestamp;      /* tv_sec of last match         */
 360
 361         u_int32_t       reserved_1;     /* reserved - set to 0 */
 362         u_int32_t       reserved_2;     /* reserved - set to 0 */
 363
 364         ipfw_insn       cmd[1];         /* storage for commands         */
 365 };
 366
 367 #define ACTION_PTR(rule)                                \
 368         (ipfw_insn *)( (u_int32_t *)((rule)->cmd) + ((rule)->act_ofs) )
 369
 370 #define RULESIZE(rule)  (sizeof(struct ip_fw) + \
 371         ((struct ip_fw *)(rule))->cmd_len * 4 - 4)
 372
 373 /*
 374  * This structure is used as a flow mask and a flow id for various
 375  * parts of the code.
 376  */
 377 struct ipfw_flow_id {
 378         u_int32_t       dst_ip;
 379         u_int32_t       src_ip;
 380         u_int16_t       dst_port;
 381         u_int16_t       src_port;
 382         u_int8_t        proto;
 383         u_int8_t        flags;  /* protocol-specific flags */
 384 };
 385
 386 /*
 387  * Dynamic ipfw rule.
 388  */
 389 typedef struct _ipfw_dyn_rule ipfw_dyn_rule;
 390
 391 struct _ipfw_dyn_rule {
 392         ipfw_dyn_rule   *next;          /* linked list of rules.        */
 393         struct ip_fw *rule;             /* pointer to rule              */
 394         /* 'rule' is used to pass up the rule number (from the parent)  */
 395
 396         ipfw_dyn_rule *parent;          /* pointer to parent rule       */
 397         u_int64_t       pcnt;           /* packet match counter         */
 398         u_int64_t       bcnt;           /* byte match counter           */
 399         struct ipfw_flow_id id;         /* (masked) flow id             */
 400         u_int32_t       expire;         /* expire time                  */
 401         u_int32_t       bucket;         /* which bucket in hash table   */
 402         u_int32_t       state;          /* state of this rule (typically a
 403                                          * combination of TCP flags)
 404                                          */
 405         u_int32_t       ack_fwd;        /* most recent ACKs in forward  */
 406         u_int32_t       ack_rev;        /* and reverse directions (used */
 407                                         /* to generate keepalives)      */
 408         u_int16_t       dyn_type;       /* rule type                    */
 409         u_int16_t       count;          /* refcount                     */
 410 };
 411
 412 /*
 413  * Definitions for IP option names.
 414  */
 415 #define IP_FW_IPOPT_LSRR        0x01
 416 #define IP_FW_IPOPT_SSRR        0x02
 417 #define IP_FW_IPOPT_RR          0x04
 418 #define IP_FW_IPOPT_TS          0x08
 419
 420 /*
 421  * Definitions for TCP option names.
 422  */
 423 #define IP_FW_TCPOPT_MSS        0x01
 424 #define IP_FW_TCPOPT_WINDOW     0x02
 425 #define IP_FW_TCPOPT_SACK       0x04
 426 #define IP_FW_TCPOPT_TS         0x08
 427 #define IP_FW_TCPOPT_CC         0x10
 428
 429 #define ICMP_REJECT_RST         0x100   /* fake ICMP code (send a TCP RST) */
 430
 431 /*
 432  * Main firewall chains definitions and global var's definitions.
 433  */
 434 #ifdef KERNEL
 435
 436 #define IP_FW_PORT_DYNT_FLAG    0x10000
 437 #define IP_FW_PORT_TEE_FLAG     0x20000
 438 #define IP_FW_PORT_DENY_FLAG    0x40000
 439
 440 /*
 441  * Arguments for calling ipfw_chk() and dummynet_io(). We put them
 442  * all into a structure because this way it is easier and more
 443  * efficient to pass variables around and extend the interface.
 444  */
 445 struct ip_fw_args {
 446         struct mbuf     *m;             /* the mbuf chain               */
 447         struct ifnet    *oif;           /* output interface             */
 448         struct sockaddr_in *next_hop;   /* forward address              */
 449         struct ip_fw    *rule;          /* matching rule                */
 450         struct ether_header *eh;        /* for bridged packets          */
 451
 452         struct route    *ro;            /* for dummynet                 */
 453         struct sockaddr_in *dst;        /* for dummynet                 */
 454         int flags;                      /* for dummynet                 */
 455
 456         struct ipfw_flow_id f_id;       /* grabbed from IP header       */
 457         u_int16_t       divert_rule;    /* divert cookie                */
 458         u_int32_t       retval;
 459 };
 460
 461 /*
 462  * Function definitions.
 463  */
 464
 465 /* Firewall hooks */
 466 struct sockopt;
 467 struct dn_flow_set;
 468
 469 void flush_pipe_ptrs(struct dn_flow_set *match); /* used by dummynet */
 470 void ipfw_init(void);   /* called from raw_ip.c: load_ipfw() */
 471
 472 typedef int ip_fw_chk_t (struct ip_fw_args *args);
 473 typedef int ip_fw_ctl_t (struct sockopt *);
 474 extern ip_fw_chk_t *ip_fw_chk_ptr;
 475 extern ip_fw_ctl_t *ip_fw_ctl_ptr;
 476 extern int fw_one_pass;
 477 extern int fw_enable;
 478 #define IPFW_LOADED     (ip_fw_chk_ptr != NULL)
 479 #endif /* KERNEL */
 480
 481 #endif /* !__LP64__ */
 482 #endif /* _IPFW2_H */