net/netfilter/nf_nat_core.c

   1 /*
   2  * (C) 1999-2001 Paul `Rusty' Russell
   3  * (C) 2002-2006 Netfilter Core Team <coreteam@netfilter.org>
   4  * (C) 2011 Patrick McHardy <kaber@trash.net>
   5  *
   6  * This program is free software; you can redistribute it and/or modify
   7  * it under the terms of the GNU General Public License version 2 as
   8  * published by the Free Software Foundation.
   9  */
  10
  11 #include <linux/module.h>
  12 #include <linux/types.h>
  13 #include <linux/timer.h>
  14 #include <linux/skbuff.h>
  15 #include <linux/gfp.h>
  16 #include <net/xfrm.h>
  17 #include <linux/jhash.h>
  18 #include <linux/rtnetlink.h>
  19
  20 #include <net/netfilter/nf_conntrack.h>
  21 #include <net/netfilter/nf_conntrack_core.h>
  22 #include <net/netfilter/nf_nat.h>
  23 #include <net/netfilter/nf_nat_l3proto.h>
  24 #include <net/netfilter/nf_nat_l4proto.h>
  25 #include <net/netfilter/nf_nat_core.h>
  26 #include <net/netfilter/nf_nat_helper.h>
  27 #include <net/netfilter/nf_conntrack_helper.h>
  28 #include <net/netfilter/nf_conntrack_seqadj.h>
  29 #include <net/netfilter/nf_conntrack_l3proto.h>
  30 #include <net/netfilter/nf_conntrack_zones.h>
  31 #include <linux/netfilter/nf_nat.h>
  32
  33 static DEFINE_SPINLOCK(nf_nat_lock);
  34
  35 static DEFINE_MUTEX(nf_nat_proto_mutex);
  36 static const struct nf_nat_l3proto __rcu *nf_nat_l3protos[NFPROTO_NUMPROTO]
  37                                                 __read_mostly;
  38 static const struct nf_nat_l4proto __rcu **nf_nat_l4protos[NFPROTO_NUMPROTO]
  39                                                 __read_mostly;
  40
  41 static struct hlist_head *nf_nat_bysource __read_mostly;
  42 static unsigned int nf_nat_htable_size __read_mostly;
  43 static unsigned int nf_nat_hash_rnd __read_mostly;
  44
  45 inline const struct nf_nat_l3proto *
  46 __nf_nat_l3proto_find(u8 family)
  47 {
  48         return rcu_dereference(nf_nat_l3protos[family]);
  49 }
  50
  51 inline const struct nf_nat_l4proto *
  52 __nf_nat_l4proto_find(u8 family, u8 protonum)
  53 {
  54         return rcu_dereference(nf_nat_l4protos[family][protonum]);
  55 }
  56 EXPORT_SYMBOL_GPL(__nf_nat_l4proto_find);
  57
  58 #ifdef CONFIG_XFRM
  59 static void __nf_nat_decode_session(struct sk_buff *skb, struct flowi *fl)
  60 {
  61         const struct nf_nat_l3proto *l3proto;
  62         const struct nf_conn *ct;
  63         enum ip_conntrack_info ctinfo;
  64         enum ip_conntrack_dir dir;
  65         unsigned  long statusbit;
  66         u8 family;
  67
  68         ct = nf_ct_get(skb, &ctinfo);
  69         if (ct == NULL)
  70                 return;
  71
  72         family = ct->tuplehash[IP_CT_DIR_ORIGINAL].tuple.src.l3num;
  73         rcu_read_lock();
  74         l3proto = __nf_nat_l3proto_find(family);
  75         if (l3proto == NULL)
  76                 goto out;
  77
  78         dir = CTINFO2DIR(ctinfo);
  79         if (dir == IP_CT_DIR_ORIGINAL)
  80                 statusbit = IPS_DST_NAT;
  81         else
  82                 statusbit = IPS_SRC_NAT;
  83
  84         l3proto->decode_session(skb, ct, dir, statusbit, fl);
  85 out:
  86         rcu_read_unlock();
  87 }
  88
  89 int nf_xfrm_me_harder(struct net *net, struct sk_buff *skb, unsigned int family)
  90 {
  91         struct flowi fl;
  92         unsigned int hh_len;
  93         struct dst_entry *dst;
  94         int err;
  95
  96         err = xfrm_decode_session(skb, &fl, family);
  97         if (err < 0)
  98                 return err;
  99
 100         dst = skb_dst(skb);
 101         if (dst->xfrm)
 102                 dst = ((struct xfrm_dst *)dst)->route;
 103         dst_hold(dst);
 104
 105         dst = xfrm_lookup(net, dst, &fl, skb->sk, 0);
 106         if (IS_ERR(dst))
 107                 return PTR_ERR(dst);
 108
 109         skb_dst_drop(skb);
 110         skb_dst_set(skb, dst);
 111
 112         /* Change in oif may mean change in hh_len. */
 113         hh_len = skb_dst(skb)->dev->hard_header_len;
 114         if (skb_headroom(skb) < hh_len &&
 115             pskb_expand_head(skb, hh_len - skb_headroom(skb), 0, GFP_ATOMIC))
 116                 return -ENOMEM;
 117         return 0;
 118 }
 119 EXPORT_SYMBOL(nf_xfrm_me_harder);
 120 #endif /* CONFIG_XFRM */
 121
 122 /* We keep an extra hash for each conntrack, for fast searching. */
 123 static inline unsigned int
 124 hash_by_src(const struct net *n, const struct nf_conntrack_tuple *tuple)
 125 {
 126         unsigned int hash;
 127
 128         get_random_once(&nf_nat_hash_rnd, sizeof(nf_nat_hash_rnd));
 129
 130         /* Original src, to ensure we map it consistently if poss. */
 131         hash = jhash2((u32 *)&tuple->src, sizeof(tuple->src) / sizeof(u32),
 132                       tuple->dst.protonum ^ nf_nat_hash_rnd ^ net_hash_mix(n));
 133
 134         return reciprocal_scale(hash, nf_nat_htable_size);
 135 }
 136
 137 /* Is this tuple already taken? (not by us) */
 138 int
 139 nf_nat_used_tuple(const struct nf_conntrack_tuple *tuple,
 140                   const struct nf_conn *ignored_conntrack)
 141 {
 142         /* Conntrack tracking doesn't keep track of outgoing tuples; only
 143          * incoming ones.  NAT means they don't have a fixed mapping,
 144          * so we invert the tuple and look for the incoming reply.
 145          *
 146          * We could keep a separate hash if this proves too slow.
 147          */
 148         struct nf_conntrack_tuple reply;
 149
 150         nf_ct_invert_tuplepr(&reply, tuple);
 151         return nf_conntrack_tuple_taken(&reply, ignored_conntrack);
 152 }
 153 EXPORT_SYMBOL(nf_nat_used_tuple);
 154
 155 /* If we source map this tuple so reply looks like reply_tuple, will
 156  * that meet the constraints of range.
 157  */
 158 static int in_range(const struct nf_nat_l3proto *l3proto,
 159                     const struct nf_nat_l4proto *l4proto,
 160                     const struct nf_conntrack_tuple *tuple,
 161                     const struct nf_nat_range *range)
 162 {
 163         /* If we are supposed to map IPs, then we must be in the
 164          * range specified, otherwise let this drag us onto a new src IP.
 165          */
 166         if (range->flags & NF_NAT_RANGE_MAP_IPS &&
 167             !l3proto->in_range(tuple, range))
 168                 return 0;
 169
 170         if (!(range->flags & NF_NAT_RANGE_PROTO_SPECIFIED) ||
 171             l4proto->in_range(tuple, NF_NAT_MANIP_SRC,
 172                               &range->min_proto, &range->max_proto))
 173                 return 1;
 174
 175         return 0;
 176 }
 177
 178 static inline int
 179 same_src(const struct nf_conn *ct,
 180          const struct nf_conntrack_tuple *tuple)
 181 {
 182         const struct nf_conntrack_tuple *t;
 183
 184         t = &ct->tuplehash[IP_CT_DIR_ORIGINAL].tuple;
 185         return (t->dst.protonum == tuple->dst.protonum &&
 186                 nf_inet_addr_cmp(&t->src.u3, &tuple->src.u3) &&
 187                 t->src.u.all == tuple->src.u.all);
 188 }
 189
 190 /* Only called for SRC manip */
 191 static int
 192 find_appropriate_src(struct net *net,
 193                      const struct nf_conntrack_zone *zone,
 194                      const struct nf_nat_l3proto *l3proto,
 195                      const struct nf_nat_l4proto *l4proto,
 196                      const struct nf_conntrack_tuple *tuple,
 197                      struct nf_conntrack_tuple *result,
 198                      const struct nf_nat_range *range)
 199 {
 200         unsigned int h = hash_by_src(net, tuple);
 201         const struct nf_conn_nat *nat;
 202         const struct nf_conn *ct;
 203
 204         hlist_for_each_entry_rcu(nat, &nf_nat_bysource[h], bysource) {
 205                 ct = nat->ct;
 206                 if (same_src(ct, tuple) &&
 207                     net_eq(net, nf_ct_net(ct)) &&
 208                     nf_ct_zone_equal(ct, zone, IP_CT_DIR_ORIGINAL)) {
 209                         /* Copy source part from reply tuple. */
 210                         nf_ct_invert_tuplepr(result,
 211                                        &ct->tuplehash[IP_CT_DIR_REPLY].tuple);
 212                         result->dst = tuple->dst;
 213
 214                         if (in_range(l3proto, l4proto, result, range))
 215                                 return 1;
 216                 }
 217         }
 218         return 0;
 219 }
 220
 221 /* For [FUTURE] fragmentation handling, we want the least-used
 222  * src-ip/dst-ip/proto triple.  Fairness doesn't come into it.  Thus
 223  * if the range specifies 1.2.3.4 ports 10000-10005 and 1.2.3.5 ports
 224  * 1-65535, we don't do pro-rata allocation based on ports; we choose
 225  * the ip with the lowest src-ip/dst-ip/proto usage.
 226  */
 227 static void
 228 find_best_ips_proto(const struct nf_conntrack_zone *zone,
 229                     struct nf_conntrack_tuple *tuple,
 230                     const struct nf_nat_range *range,
 231                     const struct nf_conn *ct,
 232                     enum nf_nat_manip_type maniptype)
 233 {
 234         union nf_inet_addr *var_ipp;
 235         unsigned int i, max;
 236         /* Host order */
 237         u32 minip, maxip, j, dist;
 238         bool full_range;
 239
 240         /* No IP mapping?  Do nothing. */
 241         if (!(range->flags & NF_NAT_RANGE_MAP_IPS))
 242                 return;
 243
 244         if (maniptype == NF_NAT_MANIP_SRC)
 245                 var_ipp = &tuple->src.u3;
 246         else
 247                 var_ipp = &tuple->dst.u3;
 248
 249         /* Fast path: only one choice. */
 250         if (nf_inet_addr_cmp(&range->min_addr, &range->max_addr)) {
 251                 *var_ipp = range->min_addr;
 252                 return;
 253         }
 254
 255         if (nf_ct_l3num(ct) == NFPROTO_IPV4)
 256                 max = sizeof(var_ipp->ip) / sizeof(u32) - 1;
 257         else
 258                 max = sizeof(var_ipp->ip6) / sizeof(u32) - 1;
 259
 260         /* Hashing source and destination IPs gives a fairly even
 261          * spread in practice (if there are a small number of IPs
 262          * involved, there usually aren't that many connections
 263          * anyway).  The consistency means that servers see the same
 264          * client coming from the same IP (some Internet Banking sites
 265          * like this), even across reboots.
 266          */
 267         j = jhash2((u32 *)&tuple->src.u3, sizeof(tuple->src.u3) / sizeof(u32),
 268                    range->flags & NF_NAT_RANGE_PERSISTENT ?
 269                         0 : (__force u32)tuple->dst.u3.all[max] ^ zone->id);
 270
 271         full_range = false;
 272         for (i = 0; i <= max; i++) {
 273                 /* If first bytes of the address are at the maximum, use the
 274                  * distance. Otherwise use the full range.
 275                  */
 276                 if (!full_range) {
 277                         minip = ntohl((__force __be32)range->min_addr.all[i]);
 278                         maxip = ntohl((__force __be32)range->max_addr.all[i]);
 279                         dist  = maxip - minip + 1;
 280                 } else {
 281                         minip = 0;
 282                         dist  = ~0;
 283                 }
 284
 285                 var_ipp->all[i] = (__force __u32)
 286                         htonl(minip + reciprocal_scale(j, dist));
 287                 if (var_ipp->all[i] != range->max_addr.all[i])
 288                         full_range = true;
 289
 290                 if (!(range->flags & NF_NAT_RANGE_PERSISTENT))
 291                         j ^= (__force u32)tuple->dst.u3.all[i];
 292         }
 293 }
 294
 295 /* Manipulate the tuple into the range given. For NF_INET_POST_ROUTING,
 296  * we change the source to map into the range. For NF_INET_PRE_ROUTING
 297  * and NF_INET_LOCAL_OUT, we change the destination to map into the
 298  * range. It might not be possible to get a unique tuple, but we try.
 299  * At worst (or if we race), we will end up with a final duplicate in
 300  * __ip_conntrack_confirm and drop the packet. */
 301 static void
 302 get_unique_tuple(struct nf_conntrack_tuple *tuple,
 303                  const struct nf_conntrack_tuple *orig_tuple,
 304                  const struct nf_nat_range *range,
 305                  struct nf_conn *ct,
 306                  enum nf_nat_manip_type maniptype)
 307 {
 308         const struct nf_conntrack_zone *zone;
 309         const struct nf_nat_l3proto *l3proto;
 310         const struct nf_nat_l4proto *l4proto;
 311         struct net *net = nf_ct_net(ct);
 312
 313         zone = nf_ct_zone(ct);
 314
 315         rcu_read_lock();
 316         l3proto = __nf_nat_l3proto_find(orig_tuple->src.l3num);
 317         l4proto = __nf_nat_l4proto_find(orig_tuple->src.l3num,
 318                                         orig_tuple->dst.protonum);
 319
 320         /* 1) If this srcip/proto/src-proto-part is currently mapped,
 321          * and that same mapping gives a unique tuple within the given
 322          * range, use that.
 323          *
 324          * This is only required for source (ie. NAT/masq) mappings.
 325          * So far, we don't do local source mappings, so multiple
 326          * manips not an issue.
 327          */
 328         if (maniptype == NF_NAT_MANIP_SRC &&
 329             !(range->flags & NF_NAT_RANGE_PROTO_RANDOM_ALL)) {
 330                 /* try the original tuple first */
 331                 if (in_range(l3proto, l4proto, orig_tuple, range)) {
 332                         if (!nf_nat_used_tuple(orig_tuple, ct)) {
 333                                 *tuple = *orig_tuple;
 334                                 goto out;
 335                         }
 336                 } else if (find_appropriate_src(net, zone, l3proto, l4proto,
 337                                                 orig_tuple, tuple, range)) {
 338                         pr_debug("get_unique_tuple: Found current src map\n");
 339                         if (!nf_nat_used_tuple(tuple, ct))
 340                                 goto out;
 341                 }
 342         }
 343
 344         /* 2) Select the least-used IP/proto combination in the given range */
 345         *tuple = *orig_tuple;
 346         find_best_ips_proto(zone, tuple, range, ct, maniptype);
 347
 348         /* 3) The per-protocol part of the manip is made to map into
 349          * the range to make a unique tuple.
 350          */
 351
 352         /* Only bother mapping if it's not already in range and unique */
 353         if (!(range->flags & NF_NAT_RANGE_PROTO_RANDOM_ALL)) {
 354                 if (range->flags & NF_NAT_RANGE_PROTO_SPECIFIED) {
 355                         if (l4proto->in_range(tuple, maniptype,
 356                                               &range->min_proto,
 357                                               &range->max_proto) &&
 358                             (range->min_proto.all == range->max_proto.all ||
 359                              !nf_nat_used_tuple(tuple, ct)))
 360                                 goto out;
 361                 } else if (!nf_nat_used_tuple(tuple, ct)) {
 362                         goto out;
 363                 }
 364         }
 365
 366         /* Last change: get protocol to try to obtain unique tuple. */
 367         l4proto->unique_tuple(l3proto, tuple, range, maniptype, ct);
 368 out:
 369         rcu_read_unlock();
 370 }
 371
 372 struct nf_conn_nat *nf_ct_nat_ext_add(struct nf_conn *ct)
 373 {
 374         struct nf_conn_nat *nat = nfct_nat(ct);
 375         if (nat)
 376                 return nat;
 377
 378         if (!nf_ct_is_confirmed(ct))
 379                 nat = nf_ct_ext_add(ct, NF_CT_EXT_NAT, GFP_ATOMIC);
 380
 381         return nat;
 382 }
 383 EXPORT_SYMBOL_GPL(nf_ct_nat_ext_add);
 384
 385 unsigned int
 386 nf_nat_setup_info(struct nf_conn *ct,
 387                   const struct nf_nat_range *range,
 388                   enum nf_nat_manip_type maniptype)
 389 {
 390         struct net *net = nf_ct_net(ct);
 391         struct nf_conntrack_tuple curr_tuple, new_tuple;
 392         struct nf_conn_nat *nat;
 393
 394         /* nat helper or nfctnetlink also setup binding */
 395         nat = nf_ct_nat_ext_add(ct);
 396         if (nat == NULL)
 397                 return NF_ACCEPT;
 398
 399         NF_CT_ASSERT(maniptype == NF_NAT_MANIP_SRC ||
 400                      maniptype == NF_NAT_MANIP_DST);
 401         BUG_ON(nf_nat_initialized(ct, maniptype));
 402
 403         /* What we've got will look like inverse of reply. Normally
 404          * this is what is in the conntrack, except for prior
 405          * manipulations (future optimization: if num_manips == 0,
 406          * orig_tp = ct->tuplehash[IP_CT_DIR_ORIGINAL].tuple)
 407          */
 408         nf_ct_invert_tuplepr(&curr_tuple,
 409                              &ct->tuplehash[IP_CT_DIR_REPLY].tuple);
 410
 411         get_unique_tuple(&new_tuple, &curr_tuple, range, ct, maniptype);
 412
 413         if (!nf_ct_tuple_equal(&new_tuple, &curr_tuple)) {
 414                 struct nf_conntrack_tuple reply;
 415
 416                 /* Alter conntrack table so will recognize replies. */
 417                 nf_ct_invert_tuplepr(&reply, &new_tuple);
 418                 nf_conntrack_alter_reply(ct, &reply);
 419
 420                 /* Non-atomic: we own this at the moment. */
 421                 if (maniptype == NF_NAT_MANIP_SRC)
 422                         ct->status |= IPS_SRC_NAT;
 423                 else
 424                         ct->status |= IPS_DST_NAT;
 425
 426                 if (nfct_help(ct))
 427                         nfct_seqadj_ext_add(ct);
 428         }
 429
 430         if (maniptype == NF_NAT_MANIP_SRC) {
 431                 unsigned int srchash;
 432
 433                 srchash = hash_by_src(net,
 434                                       &ct->tuplehash[IP_CT_DIR_ORIGINAL].tuple);
 435                 spin_lock_bh(&nf_nat_lock);
 436                 /* nf_conntrack_alter_reply might re-allocate extension aera */
 437                 nat = nfct_nat(ct);
 438                 nat->ct = ct;
 439                 hlist_add_head_rcu(&nat->bysource,
 440                                    &nf_nat_bysource[srchash]);
 441                 spin_unlock_bh(&nf_nat_lock);
 442         }
 443
 444         /* It's done. */
 445         if (maniptype == NF_NAT_MANIP_DST)
 446                 ct->status |= IPS_DST_NAT_DONE;
 447         else
 448                 ct->status |= IPS_SRC_NAT_DONE;
 449
 450         return NF_ACCEPT;
 451 }
 452 EXPORT_SYMBOL(nf_nat_setup_info);
 453
 454 static unsigned int
 455 __nf_nat_alloc_null_binding(struct nf_conn *ct, enum nf_nat_manip_type manip)
 456 {
 457         /* Force range to this IP; let proto decide mapping for
 458          * per-proto parts (hence not IP_NAT_RANGE_PROTO_SPECIFIED).
 459          * Use reply in case it's already been mangled (eg local packet).
 460          */
 461         union nf_inet_addr ip =
 462                 (manip == NF_NAT_MANIP_SRC ?
 463                 ct->tuplehash[IP_CT_DIR_REPLY].tuple.dst.u3 :
 464                 ct->tuplehash[IP_CT_DIR_REPLY].tuple.src.u3);
 465         struct nf_nat_range range = {
 466                 .flags          = NF_NAT_RANGE_MAP_IPS,
 467                 .min_addr       = ip,
 468                 .max_addr       = ip,
 469         };
 470         return nf_nat_setup_info(ct, &range, manip);
 471 }
 472
 473 unsigned int
 474 nf_nat_alloc_null_binding(struct nf_conn *ct, unsigned int hooknum)
 475 {
 476         return __nf_nat_alloc_null_binding(ct, HOOK2MANIP(hooknum));
 477 }
 478 EXPORT_SYMBOL_GPL(nf_nat_alloc_null_binding);
 479
 480 /* Do packet manipulations according to nf_nat_setup_info. */
 481 unsigned int nf_nat_packet(struct nf_conn *ct,
 482                            enum ip_conntrack_info ctinfo,
 483                            unsigned int hooknum,
 484                            struct sk_buff *skb)
 485 {
 486         const struct nf_nat_l3proto *l3proto;
 487         const struct nf_nat_l4proto *l4proto;
 488         enum ip_conntrack_dir dir = CTINFO2DIR(ctinfo);
 489         unsigned long statusbit;
 490         enum nf_nat_manip_type mtype = HOOK2MANIP(hooknum);
 491
 492         if (mtype == NF_NAT_MANIP_SRC)
 493                 statusbit = IPS_SRC_NAT;
 494         else
 495                 statusbit = IPS_DST_NAT;
 496
 497         /* Invert if this is reply dir. */
 498         if (dir == IP_CT_DIR_REPLY)
 499                 statusbit ^= IPS_NAT_MASK;
 500
 501         /* Non-atomic: these bits don't change. */
 502         if (ct->status & statusbit) {
 503                 struct nf_conntrack_tuple target;
 504
 505                 /* We are aiming to look like inverse of other direction. */
 506                 nf_ct_invert_tuplepr(&target, &ct->tuplehash[!dir].tuple);
 507
 508                 l3proto = __nf_nat_l3proto_find(target.src.l3num);
 509                 l4proto = __nf_nat_l4proto_find(target.src.l3num,
 510                                                 target.dst.protonum);
 511                 if (!l3proto->manip_pkt(skb, 0, l4proto, &target, mtype))
 512                         return NF_DROP;
 513         }
 514         return NF_ACCEPT;
 515 }
 516 EXPORT_SYMBOL_GPL(nf_nat_packet);
 517
 518 struct nf_nat_proto_clean {
 519         u8      l3proto;
 520         u8      l4proto;
 521 };
 522
 523 /* kill conntracks with affected NAT section */
 524 static int nf_nat_proto_remove(struct nf_conn *i, void *data)
 525 {
 526         const struct nf_nat_proto_clean *clean = data;
 527         struct nf_conn_nat *nat = nfct_nat(i);
 528
 529         if (!nat)
 530                 return 0;
 531
 532         if ((clean->l3proto && nf_ct_l3num(i) != clean->l3proto) ||
 533             (clean->l4proto && nf_ct_protonum(i) != clean->l4proto))
 534                 return 0;
 535
 536         return i->status & IPS_NAT_MASK ? 1 : 0;
 537 }
 538
 539 static int nf_nat_proto_clean(struct nf_conn *ct, void *data)
 540 {
 541         struct nf_conn_nat *nat = nfct_nat(ct);
 542
 543         if (nf_nat_proto_remove(ct, data))
 544                 return 1;
 545
 546         if (!nat || !nat->ct)
 547                 return 0;
 548
 549         /* This netns is being destroyed, and conntrack has nat null binding.
 550          * Remove it from bysource hash, as the table will be freed soon.
 551          *
 552          * Else, when the conntrack is destoyed, nf_nat_cleanup_conntrack()
 553          * will delete entry from already-freed table.
 554          */
 555         if (!del_timer(&ct->timeout))
 556                 return 1;
 557
 558         spin_lock_bh(&nf_nat_lock);
 559         hlist_del_rcu(&nat->bysource);
 560         ct->status &= ~IPS_NAT_DONE_MASK;
 561         nat->ct = NULL;
 562         spin_unlock_bh(&nf_nat_lock);
 563
 564         add_timer(&ct->timeout);
 565
 566         /* don't delete conntrack.  Although that would make things a lot
 567          * simpler, we'd end up flushing all conntracks on nat rmmod.
 568          */
 569         return 0;
 570 }
 571
 572 static void nf_nat_l4proto_clean(u8 l3proto, u8 l4proto)
 573 {
 574         struct nf_nat_proto_clean clean = {
 575                 .l3proto = l3proto,
 576                 .l4proto = l4proto,
 577         };
 578         struct net *net;
 579
 580         rtnl_lock();
 581         for_each_net(net)
 582                 nf_ct_iterate_cleanup(net, nf_nat_proto_remove, &clean, 0, 0);
 583         rtnl_unlock();
 584 }
 585
 586 static void nf_nat_l3proto_clean(u8 l3proto)
 587 {
 588         struct nf_nat_proto_clean clean = {
 589                 .l3proto = l3proto,
 590         };
 591         struct net *net;
 592
 593         rtnl_lock();
 594
 595         for_each_net(net)
 596                 nf_ct_iterate_cleanup(net, nf_nat_proto_remove, &clean, 0, 0);
 597         rtnl_unlock();
 598 }
 599
 600 /* Protocol registration. */
 601 int nf_nat_l4proto_register(u8 l3proto, const struct nf_nat_l4proto *l4proto)
 602 {
 603         const struct nf_nat_l4proto **l4protos;
 604         unsigned int i;
 605         int ret = 0;
 606
 607         mutex_lock(&nf_nat_proto_mutex);
 608         if (nf_nat_l4protos[l3proto] == NULL) {
 609                 l4protos = kmalloc(IPPROTO_MAX * sizeof(struct nf_nat_l4proto *),
 610                                    GFP_KERNEL);
 611                 if (l4protos == NULL) {
 612                         ret = -ENOMEM;
 613                         goto out;
 614                 }
 615
 616                 for (i = 0; i < IPPROTO_MAX; i++)
 617                         RCU_INIT_POINTER(l4protos[i], &nf_nat_l4proto_unknown);
 618
 619                 /* Before making proto_array visible to lockless readers,
 620                  * we must make sure its content is committed to memory.
 621                  */
 622                 smp_wmb();
 623
 624                 nf_nat_l4protos[l3proto] = l4protos;
 625         }
 626
 627         if (rcu_dereference_protected(
 628                         nf_nat_l4protos[l3proto][l4proto->l4proto],
 629                         lockdep_is_held(&nf_nat_proto_mutex)
 630                         ) != &nf_nat_l4proto_unknown) {
 631                 ret = -EBUSY;
 632                 goto out;
 633         }
 634         RCU_INIT_POINTER(nf_nat_l4protos[l3proto][l4proto->l4proto], l4proto);
 635  out:
 636         mutex_unlock(&nf_nat_proto_mutex);
 637         return ret;
 638 }
 639 EXPORT_SYMBOL_GPL(nf_nat_l4proto_register);
 640
 641 /* No one stores the protocol anywhere; simply delete it. */
 642 void nf_nat_l4proto_unregister(u8 l3proto, const struct nf_nat_l4proto *l4proto)
 643 {
 644         mutex_lock(&nf_nat_proto_mutex);
 645         RCU_INIT_POINTER(nf_nat_l4protos[l3proto][l4proto->l4proto],
 646                          &nf_nat_l4proto_unknown);
 647         mutex_unlock(&nf_nat_proto_mutex);
 648         synchronize_rcu();
 649
 650         nf_nat_l4proto_clean(l3proto, l4proto->l4proto);
 651 }
 652 EXPORT_SYMBOL_GPL(nf_nat_l4proto_unregister);
 653
 654 int nf_nat_l3proto_register(const struct nf_nat_l3proto *l3proto)
 655 {
 656         int err;
 657
 658         err = nf_ct_l3proto_try_module_get(l3proto->l3proto);
 659         if (err < 0)
 660                 return err;
 661
 662         mutex_lock(&nf_nat_proto_mutex);
 663         RCU_INIT_POINTER(nf_nat_l4protos[l3proto->l3proto][IPPROTO_TCP],
 664                          &nf_nat_l4proto_tcp);
 665         RCU_INIT_POINTER(nf_nat_l4protos[l3proto->l3proto][IPPROTO_UDP],
 666                          &nf_nat_l4proto_udp);
 667         mutex_unlock(&nf_nat_proto_mutex);
 668
 669         RCU_INIT_POINTER(nf_nat_l3protos[l3proto->l3proto], l3proto);
 670         return 0;
 671 }
 672 EXPORT_SYMBOL_GPL(nf_nat_l3proto_register);
 673
 674 void nf_nat_l3proto_unregister(const struct nf_nat_l3proto *l3proto)
 675 {
 676         mutex_lock(&nf_nat_proto_mutex);
 677         RCU_INIT_POINTER(nf_nat_l3protos[l3proto->l3proto], NULL);
 678         mutex_unlock(&nf_nat_proto_mutex);
 679         synchronize_rcu();
 680
 681         nf_nat_l3proto_clean(l3proto->l3proto);
 682         nf_ct_l3proto_module_put(l3proto->l3proto);
 683 }
 684 EXPORT_SYMBOL_GPL(nf_nat_l3proto_unregister);
 685
 686 /* No one using conntrack by the time this called. */
 687 static void nf_nat_cleanup_conntrack(struct nf_conn *ct)
 688 {
 689         struct nf_conn_nat *nat = nf_ct_ext_find(ct, NF_CT_EXT_NAT);
 690
 691         if (nat == NULL || nat->ct == NULL)
 692                 return;
 693
 694         NF_CT_ASSERT(nat->ct->status & IPS_SRC_NAT_DONE);
 695
 696         spin_lock_bh(&nf_nat_lock);
 697         hlist_del_rcu(&nat->bysource);
 698         spin_unlock_bh(&nf_nat_lock);
 699 }
 700
 701 static void nf_nat_move_storage(void *new, void *old)
 702 {
 703         struct nf_conn_nat *new_nat = new;
 704         struct nf_conn_nat *old_nat = old;
 705         struct nf_conn *ct = old_nat->ct;
 706
 707         if (!ct || !(ct->status & IPS_SRC_NAT_DONE))
 708                 return;
 709
 710         spin_lock_bh(&nf_nat_lock);
 711         hlist_replace_rcu(&old_nat->bysource, &new_nat->bysource);
 712         spin_unlock_bh(&nf_nat_lock);
 713 }
 714
 715 static struct nf_ct_ext_type nat_extend __read_mostly = {
 716         .len            = sizeof(struct nf_conn_nat),
 717         .align          = __alignof__(struct nf_conn_nat),
 718         .destroy        = nf_nat_cleanup_conntrack,
 719         .move           = nf_nat_move_storage,
 720         .id             = NF_CT_EXT_NAT,
 721         .flags          = NF_CT_EXT_F_PREALLOC,
 722 };
 723
 724 #if IS_ENABLED(CONFIG_NF_CT_NETLINK)
 725
 726 #include <linux/netfilter/nfnetlink.h>
 727 #include <linux/netfilter/nfnetlink_conntrack.h>
 728
 729 static const struct nla_policy protonat_nla_policy[CTA_PROTONAT_MAX+1] = {
 730         [CTA_PROTONAT_PORT_MIN] = { .type = NLA_U16 },
 731         [CTA_PROTONAT_PORT_MAX] = { .type = NLA_U16 },
 732 };
 733
 734 static int nfnetlink_parse_nat_proto(struct nlattr *attr,
 735                                      const struct nf_conn *ct,
 736                                      struct nf_nat_range *range)
 737 {
 738         struct nlattr *tb[CTA_PROTONAT_MAX+1];
 739         const struct nf_nat_l4proto *l4proto;
 740         int err;
 741
 742         err = nla_parse_nested(tb, CTA_PROTONAT_MAX, attr, protonat_nla_policy);
 743         if (err < 0)
 744                 return err;
 745
 746         l4proto = __nf_nat_l4proto_find(nf_ct_l3num(ct), nf_ct_protonum(ct));
 747         if (l4proto->nlattr_to_range)
 748                 err = l4proto->nlattr_to_range(tb, range);
 749
 750         return err;
 751 }
 752
 753 static const struct nla_policy nat_nla_policy[CTA_NAT_MAX+1] = {
 754         [CTA_NAT_V4_MINIP]      = { .type = NLA_U32 },
 755         [CTA_NAT_V4_MAXIP]      = { .type = NLA_U32 },
 756         [CTA_NAT_V6_MINIP]      = { .len = sizeof(struct in6_addr) },
 757         [CTA_NAT_V6_MAXIP]      = { .len = sizeof(struct in6_addr) },
 758         [CTA_NAT_PROTO]         = { .type = NLA_NESTED },
 759 };
 760
 761 static int
 762 nfnetlink_parse_nat(const struct nlattr *nat,
 763                     const struct nf_conn *ct, struct nf_nat_range *range,
 764                     const struct nf_nat_l3proto *l3proto)
 765 {
 766         struct nlattr *tb[CTA_NAT_MAX+1];
 767         int err;
 768
 769         memset(range, 0, sizeof(*range));
 770
 771         err = nla_parse_nested(tb, CTA_NAT_MAX, nat, nat_nla_policy);
 772         if (err < 0)
 773                 return err;
 774
 775         err = l3proto->nlattr_to_range(tb, range);
 776         if (err < 0)
 777                 return err;
 778
 779         if (!tb[CTA_NAT_PROTO])
 780                 return 0;
 781
 782         return nfnetlink_parse_nat_proto(tb[CTA_NAT_PROTO], ct, range);
 783 }
 784
 785 /* This function is called under rcu_read_lock() */
 786 static int
 787 nfnetlink_parse_nat_setup(struct nf_conn *ct,
 788                           enum nf_nat_manip_type manip,
 789                           const struct nlattr *attr)
 790 {
 791         struct nf_nat_range range;
 792         const struct nf_nat_l3proto *l3proto;
 793         int err;
 794
 795         /* Should not happen, restricted to creating new conntracks
 796          * via ctnetlink.
 797          */
 798         if (WARN_ON_ONCE(nf_nat_initialized(ct, manip)))
 799                 return -EEXIST;
 800
 801         /* Make sure that L3 NAT is there by when we call nf_nat_setup_info to
 802          * attach the null binding, otherwise this may oops.
 803          */
 804         l3proto = __nf_nat_l3proto_find(nf_ct_l3num(ct));
 805         if (l3proto == NULL)
 806                 return -EAGAIN;
 807
 808         /* No NAT information has been passed, allocate the null-binding */
 809         if (attr == NULL)
 810                 return __nf_nat_alloc_null_binding(ct, manip);
 811
 812         err = nfnetlink_parse_nat(attr, ct, &range, l3proto);
 813         if (err < 0)
 814                 return err;
 815
 816         return nf_nat_setup_info(ct, &range, manip);
 817 }
 818 #else
 819 static int
 820 nfnetlink_parse_nat_setup(struct nf_conn *ct,
 821                           enum nf_nat_manip_type manip,
 822                           const struct nlattr *attr)
 823 {
 824         return -EOPNOTSUPP;
 825 }
 826 #endif
 827
 828 static void __net_exit nf_nat_net_exit(struct net *net)
 829 {
 830         struct nf_nat_proto_clean clean = {};
 831
 832         nf_ct_iterate_cleanup(net, nf_nat_proto_clean, &clean, 0, 0);
 833 }
 834
 835 static struct pernet_operations nf_nat_net_ops = {
 836         .exit = nf_nat_net_exit,
 837 };
 838
 839 static struct nf_ct_helper_expectfn follow_master_nat = {
 840         .name           = "nat-follow-master",
 841         .expectfn       = nf_nat_follow_master,
 842 };
 843
 844 static int __init nf_nat_init(void)
 845 {
 846         int ret;
 847
 848         /* Leave them the same for the moment. */
 849         nf_nat_htable_size = nf_conntrack_htable_size;
 850
 851         nf_nat_bysource = nf_ct_alloc_hashtable(&nf_nat_htable_size, 0);
 852         if (!nf_nat_bysource)
 853                 return -ENOMEM;
 854
 855         ret = nf_ct_extend_register(&nat_extend);
 856         if (ret < 0) {
 857                 nf_ct_free_hashtable(nf_nat_bysource, nf_nat_htable_size);
 858                 printk(KERN_ERR "nf_nat_core: Unable to register extension\n");
 859                 return ret;
 860         }
 861
 862         ret = register_pernet_subsys(&nf_nat_net_ops);
 863         if (ret < 0)
 864                 goto cleanup_extend;
 865
 866         nf_ct_helper_expectfn_register(&follow_master_nat);
 867
 868         /* Initialize fake conntrack so that NAT will skip it */
 869         nf_ct_untracked_status_or(IPS_NAT_DONE_MASK);
 870
 871         BUG_ON(nfnetlink_parse_nat_setup_hook != NULL);
 872         RCU_INIT_POINTER(nfnetlink_parse_nat_setup_hook,
 873                            nfnetlink_parse_nat_setup);
 874 #ifdef CONFIG_XFRM
 875         BUG_ON(nf_nat_decode_session_hook != NULL);
 876         RCU_INIT_POINTER(nf_nat_decode_session_hook, __nf_nat_decode_session);
 877 #endif
 878         return 0;
 879
 880  cleanup_extend:
 881         nf_ct_free_hashtable(nf_nat_bysource, nf_nat_htable_size);
 882         nf_ct_extend_unregister(&nat_extend);
 883         return ret;
 884 }
 885
 886 static void __exit nf_nat_cleanup(void)
 887 {
 888         unsigned int i;
 889
 890         unregister_pernet_subsys(&nf_nat_net_ops);
 891         nf_ct_extend_unregister(&nat_extend);
 892         nf_ct_helper_expectfn_unregister(&follow_master_nat);
 893         RCU_INIT_POINTER(nfnetlink_parse_nat_setup_hook, NULL);
 894 #ifdef CONFIG_XFRM
 895         RCU_INIT_POINTER(nf_nat_decode_session_hook, NULL);
 896 #endif
 897         for (i = 0; i < NFPROTO_NUMPROTO; i++)
 898                 kfree(nf_nat_l4protos[i]);
 899         synchronize_net();
 900         nf_ct_free_hashtable(nf_nat_bysource, nf_nat_htable_size);
 901 }
 902
 903 MODULE_LICENSE("GPL");
 904
 905 module_init(nf_nat_init);
 906 module_exit(nf_nat_cleanup);