net/core/filter.c

   1 /*
   2  * Linux Socket Filter - Kernel level socket filtering
   3  *
   4  * Author:
   5  *     Jay Schulist <jschlst@samba.org>
   6  *
   7  * Based on the design of:
   8  *     - The Berkeley Packet Filter
   9  *
  10  * This program is free software; you can redistribute it and/or
  11  * modify it under the terms of the GNU General Public License
  12  * as published by the Free Software Foundation; either version
  13  * 2 of the License, or (at your option) any later version.
  14  *
  15  * Andi Kleen - Fix a few bad bugs and races.
  16  * Kris Katterjohn - Added many additional checks in sk_chk_filter()
  17  */
  18
  19 #include <linux/module.h>
  20 #include <linux/types.h>
  21 #include <linux/mm.h>
  22 #include <linux/fcntl.h>
  23 #include <linux/socket.h>
  24 #include <linux/in.h>
  25 #include <linux/inet.h>
  26 #include <linux/netdevice.h>
  27 #include <linux/if_packet.h>
  28 #include <linux/gfp.h>
  29 #include <net/ip.h>
  30 #include <net/protocol.h>
  31 #include <net/netlink.h>
  32 #include <linux/skbuff.h>
  33 #include <net/sock.h>
  34 #include <linux/errno.h>
  35 #include <linux/timer.h>
  36 #include <asm/uaccess.h>
  37 #include <asm/unaligned.h>
  38 #include <linux/filter.h>
  39 #include <linux/reciprocal_div.h>
  40 #include <linux/ratelimit.h>
  41 #include <linux/seccomp.h>
  42
  43 /* No hurry in this branch
  44  *
  45  * Exported for the bpf jit load helper.
  46  */
  47 void *bpf_internal_load_pointer_neg_helper(const struct sk_buff *skb, int k, unsigned int size)
  48 {
  49         u8 *ptr = NULL;
  50
  51         if (k >= SKF_NET_OFF)
  52                 ptr = skb_network_header(skb) + k - SKF_NET_OFF;
  53         else if (k >= SKF_LL_OFF)
  54                 ptr = skb_mac_header(skb) + k - SKF_LL_OFF;
  55
  56         if (ptr >= skb->head && ptr + size <= skb_tail_pointer(skb))
  57                 return ptr;
  58         return NULL;
  59 }
  60
  61 static inline void *load_pointer(const struct sk_buff *skb, int k,
  62                                  unsigned int size, void *buffer)
  63 {
  64         if (k >= 0)
  65                 return skb_header_pointer(skb, k, size, buffer);
  66         return bpf_internal_load_pointer_neg_helper(skb, k, size);
  67 }
  68
  69 /**
  70  *      sk_filter - run a packet through a socket filter
  71  *      @sk: sock associated with &sk_buff
  72  *      @skb: buffer to filter
  73  *
  74  * Run the filter code and then cut skb->data to correct size returned by
  75  * sk_run_filter. If pkt_len is 0 we toss packet. If skb->len is smaller
  76  * than pkt_len we keep whole skb->data. This is the socket level
  77  * wrapper to sk_run_filter. It returns 0 if the packet should
  78  * be accepted or -EPERM if the packet should be tossed.
  79  *
  80  */
  81 int sk_filter(struct sock *sk, struct sk_buff *skb)
  82 {
  83         int err;
  84         struct sk_filter *filter;
  85
  86         /*
  87          * If the skb was allocated from pfmemalloc reserves, only
  88          * allow SOCK_MEMALLOC sockets to use it as this socket is
  89          * helping free memory
  90          */
  91         if (skb_pfmemalloc(skb) && !sock_flag(sk, SOCK_MEMALLOC))
  92                 return -ENOMEM;
  93
  94         err = security_sock_rcv_skb(sk, skb);
  95         if (err)
  96                 return err;
  97
  98         rcu_read_lock();
  99         filter = rcu_dereference(sk->sk_filter);
 100         if (filter) {
 101                 unsigned int pkt_len = SK_RUN_FILTER(filter, skb);
 102
 103                 err = pkt_len ? pskb_trim(skb, pkt_len) : -EPERM;
 104         }
 105         rcu_read_unlock();
 106
 107         return err;
 108 }
 109 EXPORT_SYMBOL(sk_filter);
 110
 111 /**
 112  *      sk_run_filter - run a filter on a socket
 113  *      @skb: buffer to run the filter on
 114  *      @fentry: filter to apply
 115  *
 116  * Decode and apply filter instructions to the skb->data.
 117  * Return length to keep, 0 for none. @skb is the data we are
 118  * filtering, @filter is the array of filter instructions.
 119  * Because all jumps are guaranteed to be before last instruction,
 120  * and last instruction guaranteed to be a RET, we dont need to check
 121  * flen. (We used to pass to this function the length of filter)
 122  */
 123 unsigned int sk_run_filter(const struct sk_buff *skb,
 124                            const struct sock_filter *fentry)
 125 {
 126         void *ptr;
 127         u32 A = 0;                      /* Accumulator */
 128         u32 X = 0;                      /* Index Register */
 129         u32 mem[BPF_MEMWORDS];          /* Scratch Memory Store */
 130         u32 tmp;
 131         int k;
 132
 133         /*
 134          * Process array of filter instructions.
 135          */
 136         for (;; fentry++) {
 137 #if defined(CONFIG_X86_32)
 138 #define K (fentry->k)
 139 #else
 140                 const u32 K = fentry->k;
 141 #endif
 142
 143                 switch (fentry->code) {
 144                 case BPF_S_ALU_ADD_X:
 145                         A += X;
 146                         continue;
 147                 case BPF_S_ALU_ADD_K:
 148                         A += K;
 149                         continue;
 150                 case BPF_S_ALU_SUB_X:
 151                         A -= X;
 152                         continue;
 153                 case BPF_S_ALU_SUB_K:
 154                         A -= K;
 155                         continue;
 156                 case BPF_S_ALU_MUL_X:
 157                         A *= X;
 158                         continue;
 159                 case BPF_S_ALU_MUL_K:
 160                         A *= K;
 161                         continue;
 162                 case BPF_S_ALU_DIV_X:
 163                         if (X == 0)
 164                                 return 0;
 165                         A /= X;
 166                         continue;
 167                 case BPF_S_ALU_DIV_K:
 168                         A = reciprocal_divide(A, K);
 169                         continue;
 170                 case BPF_S_ALU_MOD_X:
 171                         if (X == 0)
 172                                 return 0;
 173                         A %= X;
 174                         continue;
 175                 case BPF_S_ALU_MOD_K:
 176                         A %= K;
 177                         continue;
 178                 case BPF_S_ALU_AND_X:
 179                         A &= X;
 180                         continue;
 181                 case BPF_S_ALU_AND_K:
 182                         A &= K;
 183                         continue;
 184                 case BPF_S_ALU_OR_X:
 185                         A |= X;
 186                         continue;
 187                 case BPF_S_ALU_OR_K:
 188                         A |= K;
 189                         continue;
 190                 case BPF_S_ANC_ALU_XOR_X:
 191                 case BPF_S_ALU_XOR_X:
 192                         A ^= X;
 193                         continue;
 194                 case BPF_S_ALU_XOR_K:
 195                         A ^= K;
 196                         continue;
 197                 case BPF_S_ALU_LSH_X:
 198                         A <<= X;
 199                         continue;
 200                 case BPF_S_ALU_LSH_K:
 201                         A <<= K;
 202                         continue;
 203                 case BPF_S_ALU_RSH_X:
 204                         A >>= X;
 205                         continue;
 206                 case BPF_S_ALU_RSH_K:
 207                         A >>= K;
 208                         continue;
 209                 case BPF_S_ALU_NEG:
 210                         A = -A;
 211                         continue;
 212                 case BPF_S_JMP_JA:
 213                         fentry += K;
 214                         continue;
 215                 case BPF_S_JMP_JGT_K:
 216                         fentry += (A > K) ? fentry->jt : fentry->jf;
 217                         continue;
 218                 case BPF_S_JMP_JGE_K:
 219                         fentry += (A >= K) ? fentry->jt : fentry->jf;
 220                         continue;
 221                 case BPF_S_JMP_JEQ_K:
 222                         fentry += (A == K) ? fentry->jt : fentry->jf;
 223                         continue;
 224                 case BPF_S_JMP_JSET_K:
 225                         fentry += (A & K) ? fentry->jt : fentry->jf;
 226                         continue;
 227                 case BPF_S_JMP_JGT_X:
 228                         fentry += (A > X) ? fentry->jt : fentry->jf;
 229                         continue;
 230                 case BPF_S_JMP_JGE_X:
 231                         fentry += (A >= X) ? fentry->jt : fentry->jf;
 232                         continue;
 233                 case BPF_S_JMP_JEQ_X:
 234                         fentry += (A == X) ? fentry->jt : fentry->jf;
 235                         continue;
 236                 case BPF_S_JMP_JSET_X:
 237                         fentry += (A & X) ? fentry->jt : fentry->jf;
 238                         continue;
 239                 case BPF_S_LD_W_ABS:
 240                         k = K;
 241 load_w:
 242                         ptr = load_pointer(skb, k, 4, &tmp);
 243                         if (ptr != NULL) {
 244                                 A = get_unaligned_be32(ptr);
 245                                 continue;
 246                         }
 247                         return 0;
 248                 case BPF_S_LD_H_ABS:
 249                         k = K;
 250 load_h:
 251                         ptr = load_pointer(skb, k, 2, &tmp);
 252                         if (ptr != NULL) {
 253                                 A = get_unaligned_be16(ptr);
 254                                 continue;
 255                         }
 256                         return 0;
 257                 case BPF_S_LD_B_ABS:
 258                         k = K;
 259 load_b:
 260                         ptr = load_pointer(skb, k, 1, &tmp);
 261                         if (ptr != NULL) {
 262                                 A = *(u8 *)ptr;
 263                                 continue;
 264                         }
 265                         return 0;
 266                 case BPF_S_LD_W_LEN:
 267                         A = skb->len;
 268                         continue;
 269                 case BPF_S_LDX_W_LEN:
 270                         X = skb->len;
 271                         continue;
 272                 case BPF_S_LD_W_IND:
 273                         k = X + K;
 274                         goto load_w;
 275                 case BPF_S_LD_H_IND:
 276                         k = X + K;
 277                         goto load_h;
 278                 case BPF_S_LD_B_IND:
 279                         k = X + K;
 280                         goto load_b;
 281                 case BPF_S_LDX_B_MSH:
 282                         ptr = load_pointer(skb, K, 1, &tmp);
 283                         if (ptr != NULL) {
 284                                 X = (*(u8 *)ptr & 0xf) << 2;
 285                                 continue;
 286                         }
 287                         return 0;
 288                 case BPF_S_LD_IMM:
 289                         A = K;
 290                         continue;
 291                 case BPF_S_LDX_IMM:
 292                         X = K;
 293                         continue;
 294                 case BPF_S_LD_MEM:
 295                         A = mem[K];
 296                         continue;
 297                 case BPF_S_LDX_MEM:
 298                         X = mem[K];
 299                         continue;
 300                 case BPF_S_MISC_TAX:
 301                         X = A;
 302                         continue;
 303                 case BPF_S_MISC_TXA:
 304                         A = X;
 305                         continue;
 306                 case BPF_S_RET_K:
 307                         return K;
 308                 case BPF_S_RET_A:
 309                         return A;
 310                 case BPF_S_ST:
 311                         mem[K] = A;
 312                         continue;
 313                 case BPF_S_STX:
 314                         mem[K] = X;
 315                         continue;
 316                 case BPF_S_ANC_PROTOCOL:
 317                         A = ntohs(skb->protocol);
 318                         continue;
 319                 case BPF_S_ANC_PKTTYPE:
 320                         A = skb->pkt_type;
 321                         continue;
 322                 case BPF_S_ANC_IFINDEX:
 323                         if (!skb->dev)
 324                                 return 0;
 325                         A = skb->dev->ifindex;
 326                         continue;
 327                 case BPF_S_ANC_MARK:
 328                         A = skb->mark;
 329                         continue;
 330                 case BPF_S_ANC_QUEUE:
 331                         A = skb->queue_mapping;
 332                         continue;
 333                 case BPF_S_ANC_HATYPE:
 334                         if (!skb->dev)
 335                                 return 0;
 336                         A = skb->dev->type;
 337                         continue;
 338                 case BPF_S_ANC_RXHASH:
 339                         A = skb->rxhash;
 340                         continue;
 341                 case BPF_S_ANC_CPU:
 342                         A = raw_smp_processor_id();
 343                         continue;
 344                 case BPF_S_ANC_NLATTR: {
 345                         struct nlattr *nla;
 346
 347                         if (skb_is_nonlinear(skb))
 348                                 return 0;
 349                         if (A > skb->len - sizeof(struct nlattr))
 350                                 return 0;
 351
 352                         nla = nla_find((struct nlattr *)&skb->data[A],
 353                                        skb->len - A, X);
 354                         if (nla)
 355                                 A = (void *)nla - (void *)skb->data;
 356                         else
 357                                 A = 0;
 358                         continue;
 359                 }
 360                 case BPF_S_ANC_NLATTR_NEST: {
 361                         struct nlattr *nla;
 362
 363                         if (skb_is_nonlinear(skb))
 364                                 return 0;
 365                         if (A > skb->len - sizeof(struct nlattr))
 366                                 return 0;
 367
 368                         nla = (struct nlattr *)&skb->data[A];
 369                         if (nla->nla_len > A - skb->len)
 370                                 return 0;
 371
 372                         nla = nla_find_nested(nla, X);
 373                         if (nla)
 374                                 A = (void *)nla - (void *)skb->data;
 375                         else
 376                                 A = 0;
 377                         continue;
 378                 }
 379 #ifdef CONFIG_SECCOMP_FILTER
 380                 case BPF_S_ANC_SECCOMP_LD_W:
 381                         A = seccomp_bpf_load(fentry->k);
 382                         continue;
 383 #endif
 384                 default:
 385                         WARN_RATELIMIT(1, "Unknown code:%u jt:%u tf:%u k:%u\n",
 386                                        fentry->code, fentry->jt,
 387                                        fentry->jf, fentry->k);
 388                         return 0;
 389                 }
 390         }
 391
 392         return 0;
 393 }
 394 EXPORT_SYMBOL(sk_run_filter);
 395
 396 /*
 397  * Security :
 398  * A BPF program is able to use 16 cells of memory to store intermediate
 399  * values (check u32 mem[BPF_MEMWORDS] in sk_run_filter())
 400  * As we dont want to clear mem[] array for each packet going through
 401  * sk_run_filter(), we check that filter loaded by user never try to read
 402  * a cell if not previously written, and we check all branches to be sure
 403  * a malicious user doesn't try to abuse us.
 404  */
 405 static int check_load_and_stores(struct sock_filter *filter, int flen)
 406 {
 407         u16 *masks, memvalid = 0; /* one bit per cell, 16 cells */
 408         int pc, ret = 0;
 409
 410         BUILD_BUG_ON(BPF_MEMWORDS > 16);
 411         masks = kmalloc(flen * sizeof(*masks), GFP_KERNEL);
 412         if (!masks)
 413                 return -ENOMEM;
 414         memset(masks, 0xff, flen * sizeof(*masks));
 415
 416         for (pc = 0; pc < flen; pc++) {
 417                 memvalid &= masks[pc];
 418
 419                 switch (filter[pc].code) {
 420                 case BPF_S_ST:
 421                 case BPF_S_STX:
 422                         memvalid |= (1 << filter[pc].k);
 423                         break;
 424                 case BPF_S_LD_MEM:
 425                 case BPF_S_LDX_MEM:
 426                         if (!(memvalid & (1 << filter[pc].k))) {
 427                                 ret = -EINVAL;
 428                                 goto error;
 429                         }
 430                         break;
 431                 case BPF_S_JMP_JA:
 432                         /* a jump must set masks on target */
 433                         masks[pc + 1 + filter[pc].k] &= memvalid;
 434                         memvalid = ~0;
 435                         break;
 436                 case BPF_S_JMP_JEQ_K:
 437                 case BPF_S_JMP_JEQ_X:
 438                 case BPF_S_JMP_JGE_K:
 439                 case BPF_S_JMP_JGE_X:
 440                 case BPF_S_JMP_JGT_K:
 441                 case BPF_S_JMP_JGT_X:
 442                 case BPF_S_JMP_JSET_X:
 443                 case BPF_S_JMP_JSET_K:
 444                         /* a jump must set masks on targets */
 445                         masks[pc + 1 + filter[pc].jt] &= memvalid;
 446                         masks[pc + 1 + filter[pc].jf] &= memvalid;
 447                         memvalid = ~0;
 448                         break;
 449                 }
 450         }
 451 error:
 452         kfree(masks);
 453         return ret;
 454 }
 455
 456 /**
 457  *      sk_chk_filter - verify socket filter code
 458  *      @filter: filter to verify
 459  *      @flen: length of filter
 460  *
 461  * Check the user's filter code. If we let some ugly
 462  * filter code slip through kaboom! The filter must contain
 463  * no references or jumps that are out of range, no illegal
 464  * instructions, and must end with a RET instruction.
 465  *
 466  * All jumps are forward as they are not signed.
 467  *
 468  * Returns 0 if the rule set is legal or -EINVAL if not.
 469  */
 470 int sk_chk_filter(struct sock_filter *filter, unsigned int flen)
 471 {
 472         /*
 473          * Valid instructions are initialized to non-0.
 474          * Invalid instructions are initialized to 0.
 475          */
 476         static const u8 codes[] = {
 477                 [BPF_ALU|BPF_ADD|BPF_K]  = BPF_S_ALU_ADD_K,
 478                 [BPF_ALU|BPF_ADD|BPF_X]  = BPF_S_ALU_ADD_X,
 479                 [BPF_ALU|BPF_SUB|BPF_K]  = BPF_S_ALU_SUB_K,
 480                 [BPF_ALU|BPF_SUB|BPF_X]  = BPF_S_ALU_SUB_X,
 481                 [BPF_ALU|BPF_MUL|BPF_K]  = BPF_S_ALU_MUL_K,
 482                 [BPF_ALU|BPF_MUL|BPF_X]  = BPF_S_ALU_MUL_X,
 483                 [BPF_ALU|BPF_DIV|BPF_X]  = BPF_S_ALU_DIV_X,
 484                 [BPF_ALU|BPF_MOD|BPF_K]  = BPF_S_ALU_MOD_K,
 485                 [BPF_ALU|BPF_MOD|BPF_X]  = BPF_S_ALU_MOD_X,
 486                 [BPF_ALU|BPF_AND|BPF_K]  = BPF_S_ALU_AND_K,
 487                 [BPF_ALU|BPF_AND|BPF_X]  = BPF_S_ALU_AND_X,
 488                 [BPF_ALU|BPF_OR|BPF_K]   = BPF_S_ALU_OR_K,
 489                 [BPF_ALU|BPF_OR|BPF_X]   = BPF_S_ALU_OR_X,
 490                 [BPF_ALU|BPF_XOR|BPF_K]  = BPF_S_ALU_XOR_K,
 491                 [BPF_ALU|BPF_XOR|BPF_X]  = BPF_S_ALU_XOR_X,
 492                 [BPF_ALU|BPF_LSH|BPF_K]  = BPF_S_ALU_LSH_K,
 493                 [BPF_ALU|BPF_LSH|BPF_X]  = BPF_S_ALU_LSH_X,
 494                 [BPF_ALU|BPF_RSH|BPF_K]  = BPF_S_ALU_RSH_K,
 495                 [BPF_ALU|BPF_RSH|BPF_X]  = BPF_S_ALU_RSH_X,
 496                 [BPF_ALU|BPF_NEG]        = BPF_S_ALU_NEG,
 497                 [BPF_LD|BPF_W|BPF_ABS]   = BPF_S_LD_W_ABS,
 498                 [BPF_LD|BPF_H|BPF_ABS]   = BPF_S_LD_H_ABS,
 499                 [BPF_LD|BPF_B|BPF_ABS]   = BPF_S_LD_B_ABS,
 500                 [BPF_LD|BPF_W|BPF_LEN]   = BPF_S_LD_W_LEN,
 501                 [BPF_LD|BPF_W|BPF_IND]   = BPF_S_LD_W_IND,
 502                 [BPF_LD|BPF_H|BPF_IND]   = BPF_S_LD_H_IND,
 503                 [BPF_LD|BPF_B|BPF_IND]   = BPF_S_LD_B_IND,
 504                 [BPF_LD|BPF_IMM]         = BPF_S_LD_IMM,
 505                 [BPF_LDX|BPF_W|BPF_LEN]  = BPF_S_LDX_W_LEN,
 506                 [BPF_LDX|BPF_B|BPF_MSH]  = BPF_S_LDX_B_MSH,
 507                 [BPF_LDX|BPF_IMM]        = BPF_S_LDX_IMM,
 508                 [BPF_MISC|BPF_TAX]       = BPF_S_MISC_TAX,
 509                 [BPF_MISC|BPF_TXA]       = BPF_S_MISC_TXA,
 510                 [BPF_RET|BPF_K]          = BPF_S_RET_K,
 511                 [BPF_RET|BPF_A]          = BPF_S_RET_A,
 512                 [BPF_ALU|BPF_DIV|BPF_K]  = BPF_S_ALU_DIV_K,
 513                 [BPF_LD|BPF_MEM]         = BPF_S_LD_MEM,
 514                 [BPF_LDX|BPF_MEM]        = BPF_S_LDX_MEM,
 515                 [BPF_ST]                 = BPF_S_ST,
 516                 [BPF_STX]                = BPF_S_STX,
 517                 [BPF_JMP|BPF_JA]         = BPF_S_JMP_JA,
 518                 [BPF_JMP|BPF_JEQ|BPF_K]  = BPF_S_JMP_JEQ_K,
 519                 [BPF_JMP|BPF_JEQ|BPF_X]  = BPF_S_JMP_JEQ_X,
 520                 [BPF_JMP|BPF_JGE|BPF_K]  = BPF_S_JMP_JGE_K,
 521                 [BPF_JMP|BPF_JGE|BPF_X]  = BPF_S_JMP_JGE_X,
 522                 [BPF_JMP|BPF_JGT|BPF_K]  = BPF_S_JMP_JGT_K,
 523                 [BPF_JMP|BPF_JGT|BPF_X]  = BPF_S_JMP_JGT_X,
 524                 [BPF_JMP|BPF_JSET|BPF_K] = BPF_S_JMP_JSET_K,
 525                 [BPF_JMP|BPF_JSET|BPF_X] = BPF_S_JMP_JSET_X,
 526         };
 527         int pc;
 528
 529         if (flen == 0 || flen > BPF_MAXINSNS)
 530                 return -EINVAL;
 531
 532         /* check the filter code now */
 533         for (pc = 0; pc < flen; pc++) {
 534                 struct sock_filter *ftest = &filter[pc];
 535                 u16 code = ftest->code;
 536
 537                 if (code >= ARRAY_SIZE(codes))
 538                         return -EINVAL;
 539                 code = codes[code];
 540                 if (!code)
 541                         return -EINVAL;
 542                 /* Some instructions need special checks */
 543                 switch (code) {
 544                 case BPF_S_ALU_DIV_K:
 545                         /* check for division by zero */
 546                         if (ftest->k == 0)
 547                                 return -EINVAL;
 548                         ftest->k = reciprocal_value(ftest->k);
 549                         break;
 550                 case BPF_S_ALU_MOD_K:
 551                         /* check for division by zero */
 552                         if (ftest->k == 0)
 553                                 return -EINVAL;
 554                         break;
 555                 case BPF_S_LD_MEM:
 556                 case BPF_S_LDX_MEM:
 557                 case BPF_S_ST:
 558                 case BPF_S_STX:
 559                         /* check for invalid memory addresses */
 560                         if (ftest->k >= BPF_MEMWORDS)
 561                                 return -EINVAL;
 562                         break;
 563                 case BPF_S_JMP_JA:
 564                         /*
 565                          * Note, the large ftest->k might cause loops.
 566                          * Compare this with conditional jumps below,
 567                          * where offsets are limited. --ANK (981016)
 568                          */
 569                         if (ftest->k >= (unsigned int)(flen-pc-1))
 570                                 return -EINVAL;
 571                         break;
 572                 case BPF_S_JMP_JEQ_K:
 573                 case BPF_S_JMP_JEQ_X:
 574                 case BPF_S_JMP_JGE_K:
 575                 case BPF_S_JMP_JGE_X:
 576                 case BPF_S_JMP_JGT_K:
 577                 case BPF_S_JMP_JGT_X:
 578                 case BPF_S_JMP_JSET_X:
 579                 case BPF_S_JMP_JSET_K:
 580                         /* for conditionals both must be safe */
 581                         if (pc + ftest->jt + 1 >= flen ||
 582                             pc + ftest->jf + 1 >= flen)
 583                                 return -EINVAL;
 584                         break;
 585                 case BPF_S_LD_W_ABS:
 586                 case BPF_S_LD_H_ABS:
 587                 case BPF_S_LD_B_ABS:
 588 #define ANCILLARY(CODE) case SKF_AD_OFF + SKF_AD_##CODE:        \
 589                                 code = BPF_S_ANC_##CODE;        \
 590                                 break
 591                         switch (ftest->k) {
 592                         ANCILLARY(PROTOCOL);
 593                         ANCILLARY(PKTTYPE);
 594                         ANCILLARY(IFINDEX);
 595                         ANCILLARY(NLATTR);
 596                         ANCILLARY(NLATTR_NEST);
 597                         ANCILLARY(MARK);
 598                         ANCILLARY(QUEUE);
 599                         ANCILLARY(HATYPE);
 600                         ANCILLARY(RXHASH);
 601                         ANCILLARY(CPU);
 602                         ANCILLARY(ALU_XOR_X);
 603                         }
 604                 }
 605                 ftest->code = code;
 606         }
 607
 608         /* last instruction must be a RET code */
 609         switch (filter[flen - 1].code) {
 610         case BPF_S_RET_K:
 611         case BPF_S_RET_A:
 612                 return check_load_and_stores(filter, flen);
 613         }
 614         return -EINVAL;
 615 }
 616 EXPORT_SYMBOL(sk_chk_filter);
 617
 618 /**
 619  *      sk_filter_release_rcu - Release a socket filter by rcu_head
 620  *      @rcu: rcu_head that contains the sk_filter to free
 621  */
 622 void sk_filter_release_rcu(struct rcu_head *rcu)
 623 {
 624         struct sk_filter *fp = container_of(rcu, struct sk_filter, rcu);
 625
 626         bpf_jit_free(fp);
 627         kfree(fp);
 628 }
 629 EXPORT_SYMBOL(sk_filter_release_rcu);
 630
 631 static int __sk_prepare_filter(struct sk_filter *fp)
 632 {
 633         int err;
 634
 635         fp->bpf_func = sk_run_filter;
 636
 637         err = sk_chk_filter(fp->insns, fp->len);
 638         if (err)
 639                 return err;
 640
 641         bpf_jit_compile(fp);
 642         return 0;
 643 }
 644
 645 /**
 646  *      sk_unattached_filter_create - create an unattached filter
 647  *      @fprog: the filter program
 648  *      @pfp: the unattached filter that is created
 649  *
 650  * Create a filter independent of any socket. We first run some
 651  * sanity checks on it to make sure it does not explode on us later.
 652  * If an error occurs or there is insufficient memory for the filter
 653  * a negative errno code is returned. On success the return is zero.
 654  */
 655 int sk_unattached_filter_create(struct sk_filter **pfp,
 656                                 struct sock_fprog *fprog)
 657 {
 658         struct sk_filter *fp;
 659         unsigned int fsize = sizeof(struct sock_filter) * fprog->len;
 660         int err;
 661
 662         /* Make sure new filter is there and in the right amounts. */
 663         if (fprog->filter == NULL)
 664                 return -EINVAL;
 665
 666         fp = kmalloc(fsize + sizeof(*fp), GFP_KERNEL);
 667         if (!fp)
 668                 return -ENOMEM;
 669         memcpy(fp->insns, fprog->filter, fsize);
 670
 671         atomic_set(&fp->refcnt, 1);
 672         fp->len = fprog->len;
 673
 674         err = __sk_prepare_filter(fp);
 675         if (err)
 676                 goto free_mem;
 677
 678         *pfp = fp;
 679         return 0;
 680 free_mem:
 681         kfree(fp);
 682         return err;
 683 }
 684 EXPORT_SYMBOL_GPL(sk_unattached_filter_create);
 685
 686 void sk_unattached_filter_destroy(struct sk_filter *fp)
 687 {
 688         sk_filter_release(fp);
 689 }
 690 EXPORT_SYMBOL_GPL(sk_unattached_filter_destroy);
 691
 692 /**
 693  *      sk_attach_filter - attach a socket filter
 694  *      @fprog: the filter program
 695  *      @sk: the socket to use
 696  *
 697  * Attach the user's filter code. We first run some sanity checks on
 698  * it to make sure it does not explode on us later. If an error
 699  * occurs or there is insufficient memory for the filter a negative
 700  * errno code is returned. On success the return is zero.
 701  */
 702 int sk_attach_filter(struct sock_fprog *fprog, struct sock *sk)
 703 {
 704         struct sk_filter *fp, *old_fp;
 705         unsigned int fsize = sizeof(struct sock_filter) * fprog->len;
 706         int err;
 707
 708         /* Make sure new filter is there and in the right amounts. */
 709         if (fprog->filter == NULL)
 710                 return -EINVAL;
 711
 712         fp = sock_kmalloc(sk, fsize+sizeof(*fp), GFP_KERNEL);
 713         if (!fp)
 714                 return -ENOMEM;
 715         if (copy_from_user(fp->insns, fprog->filter, fsize)) {
 716                 sock_kfree_s(sk, fp, fsize+sizeof(*fp));
 717                 return -EFAULT;
 718         }
 719
 720         atomic_set(&fp->refcnt, 1);
 721         fp->len = fprog->len;
 722
 723         err = __sk_prepare_filter(fp);
 724         if (err) {
 725                 sk_filter_uncharge(sk, fp);
 726                 return err;
 727         }
 728
 729         old_fp = rcu_dereference_protected(sk->sk_filter,
 730                                            sock_owned_by_user(sk));
 731         rcu_assign_pointer(sk->sk_filter, fp);
 732
 733         if (old_fp)
 734                 sk_filter_uncharge(sk, old_fp);
 735         return 0;
 736 }
 737 EXPORT_SYMBOL_GPL(sk_attach_filter);
 738
 739 int sk_detach_filter(struct sock *sk)
 740 {
 741         int ret = -ENOENT;
 742         struct sk_filter *filter;
 743
 744         filter = rcu_dereference_protected(sk->sk_filter,
 745                                            sock_owned_by_user(sk));
 746         if (filter) {
 747                 RCU_INIT_POINTER(sk->sk_filter, NULL);
 748                 sk_filter_uncharge(sk, filter);
 749                 ret = 0;
 750         }
 751         return ret;
 752 }
 753 EXPORT_SYMBOL_GPL(sk_detach_filter);