1 // SPDX-License-Identifier: GPL-2.0-only 2 #define pr_fmt(fmt) "IPsec: " fmt 3 4 #include <crypto/aead.h> 5 #include <crypto/authenc.h> 6 #include <linux/err.h> 7 #include <linux/module.h> 8 #include <net/ip.h> 9 #include <net/xfrm.h> 10 #include <net/esp.h> 11 #include <linux/scatterlist.h> 12 #include <linux/kernel.h> 13 #include <linux/pfkeyv2.h> 14 #include <linux/rtnetlink.h> 15 #include <linux/slab.h> 16 #include <linux/spinlock.h> 17 #include <linux/in6.h> 18 #include <net/icmp.h> 19 #include <net/protocol.h> 20 #include <net/udp.h> 21 #include <net/tcp.h> 22 #include <net/espintcp.h> 23 24 #include <linux/highmem.h> 25 26 struct esp_skb_cb { 27 struct xfrm_skb_cb xfrm; 28 void *tmp; 29 }; 30 31 struct esp_output_extra { 32 __be32 seqhi; 33 u32 esphoff; 34 }; 35 36 #define ESP_SKB_CB(__skb) ((struct esp_skb_cb *)&((__skb)->cb[0])) 37 38 /* 39 * Allocate an AEAD request structure with extra space for SG and IV. 40 * 41 * For alignment considerations the IV is placed at the front, followed 42 * by the request and finally the SG list. 43 * 44 * TODO: Use spare space in skb for this where possible. 45 */ 46 static void *esp_alloc_tmp(struct crypto_aead *aead, int nfrags, int extralen) 47 { 48 unsigned int len; 49 50 len = extralen; 51 52 len += crypto_aead_ivsize(aead); 53 54 if (len) { 55 len += crypto_aead_alignmask(aead) & 56 ~(crypto_tfm_ctx_alignment() - 1); 57 len = ALIGN(len, crypto_tfm_ctx_alignment()); 58 } 59 60 len += sizeof(struct aead_request) + crypto_aead_reqsize(aead); 61 len = ALIGN(len, __alignof__(struct scatterlist)); 62 63 len += sizeof(struct scatterlist) * nfrags; 64 65 return kmalloc(len, GFP_ATOMIC); 66 } 67 68 static inline void *esp_tmp_extra(void *tmp) 69 { 70 return PTR_ALIGN(tmp, __alignof__(struct esp_output_extra)); 71 } 72 73 static inline u8 *esp_tmp_iv(struct crypto_aead *aead, void *tmp, int extralen) 74 { 75 return crypto_aead_ivsize(aead) ? 76 PTR_ALIGN((u8 *)tmp + extralen, 77 crypto_aead_alignmask(aead) + 1) : tmp + extralen; 78 } 79 80 static inline struct aead_request *esp_tmp_req(struct crypto_aead *aead, u8 *iv) 81 { 82 struct aead_request *req; 83 84 req = (void *)PTR_ALIGN(iv + crypto_aead_ivsize(aead), 85 crypto_tfm_ctx_alignment()); 86 aead_request_set_tfm(req, aead); 87 return req; 88 } 89 90 static inline struct scatterlist *esp_req_sg(struct crypto_aead *aead, 91 struct aead_request *req) 92 { 93 return (void *)ALIGN((unsigned long)(req + 1) + 94 crypto_aead_reqsize(aead), 95 __alignof__(struct scatterlist)); 96 } 97 98 static void esp_ssg_unref(struct xfrm_state *x, void *tmp) 99 { 100 struct crypto_aead *aead = x->data; 101 int extralen = 0; 102 u8 *iv; 103 struct aead_request *req; 104 struct scatterlist *sg; 105 106 if (x->props.flags & XFRM_STATE_ESN) 107 extralen += sizeof(struct esp_output_extra); 108 109 iv = esp_tmp_iv(aead, tmp, extralen); 110 req = esp_tmp_req(aead, iv); 111 112 /* Unref skb_frag_pages in the src scatterlist if necessary. 113 * Skip the first sg which comes from skb->data. 114 */ 115 if (req->src != req->dst) 116 for (sg = sg_next(req->src); sg; sg = sg_next(sg)) 117 put_page(sg_page(sg)); 118 } 119 120 #ifdef CONFIG_INET_ESPINTCP 121 struct esp_tcp_sk { 122 struct sock *sk; 123 struct rcu_head rcu; 124 }; 125 126 static void esp_free_tcp_sk(struct rcu_head *head) 127 { 128 struct esp_tcp_sk *esk = container_of(head, struct esp_tcp_sk, rcu); 129 130 sock_put(esk->sk); 131 kfree(esk); 132 } 133 134 static struct sock *esp_find_tcp_sk(struct xfrm_state *x) 135 { 136 struct xfrm_encap_tmpl *encap = x->encap; 137 struct esp_tcp_sk *esk; 138 __be16 sport, dport; 139 struct sock *nsk; 140 struct sock *sk; 141 142 sk = rcu_dereference(x->encap_sk); 143 if (sk && sk->sk_state == TCP_ESTABLISHED) 144 return sk; 145 146 spin_lock_bh(&x->lock); 147 sport = encap->encap_sport; 148 dport = encap->encap_dport; 149 nsk = rcu_dereference_protected(x->encap_sk, 150 lockdep_is_held(&x->lock)); 151 if (sk && sk == nsk) { 152 esk = kmalloc(sizeof(*esk), GFP_ATOMIC); 153 if (!esk) { 154 spin_unlock_bh(&x->lock); 155 return ERR_PTR(-ENOMEM); 156 } 157 RCU_INIT_POINTER(x->encap_sk, NULL); 158 esk->sk = sk; 159 call_rcu(&esk->rcu, esp_free_tcp_sk); 160 } 161 spin_unlock_bh(&x->lock); 162 163 sk = inet_lookup_established(xs_net(x), &tcp_hashinfo, x->id.daddr.a4, 164 dport, x->props.saddr.a4, sport, 0); 165 if (!sk) 166 return ERR_PTR(-ENOENT); 167 168 if (!tcp_is_ulp_esp(sk)) { 169 sock_put(sk); 170 return ERR_PTR(-EINVAL); 171 } 172 173 spin_lock_bh(&x->lock); 174 nsk = rcu_dereference_protected(x->encap_sk, 175 lockdep_is_held(&x->lock)); 176 if (encap->encap_sport != sport || 177 encap->encap_dport != dport) { 178 sock_put(sk); 179 sk = nsk ?: ERR_PTR(-EREMCHG); 180 } else if (sk == nsk) { 181 sock_put(sk); 182 } else { 183 rcu_assign_pointer(x->encap_sk, sk); 184 } 185 spin_unlock_bh(&x->lock); 186 187 return sk; 188 } 189 190 static int esp_output_tcp_finish(struct xfrm_state *x, struct sk_buff *skb) 191 { 192 struct sock *sk; 193 int err; 194 195 rcu_read_lock(); 196 197 sk = esp_find_tcp_sk(x); 198 err = PTR_ERR_OR_ZERO(sk); 199 if (err) 200 goto out; 201 202 bh_lock_sock(sk); 203 if (sock_owned_by_user(sk)) 204 err = espintcp_queue_out(sk, skb); 205 else 206 err = espintcp_push_skb(sk, skb); 207 bh_unlock_sock(sk); 208 209 out: 210 rcu_read_unlock(); 211 return err; 212 } 213 214 static int esp_output_tcp_encap_cb(struct net *net, struct sock *sk, 215 struct sk_buff *skb) 216 { 217 struct dst_entry *dst = skb_dst(skb); 218 struct xfrm_state *x = dst->xfrm; 219 220 return esp_output_tcp_finish(x, skb); 221 } 222 223 static int esp_output_tail_tcp(struct xfrm_state *x, struct sk_buff *skb) 224 { 225 int err; 226 227 local_bh_disable(); 228 err = xfrm_trans_queue_net(xs_net(x), skb, esp_output_tcp_encap_cb); 229 local_bh_enable(); 230 231 /* EINPROGRESS just happens to do the right thing. It 232 * actually means that the skb has been consumed and 233 * isn't coming back. 234 */ 235 return err ?: -EINPROGRESS; 236 } 237 #else 238 static int esp_output_tail_tcp(struct xfrm_state *x, struct sk_buff *skb) 239 { 240 kfree_skb(skb); 241 242 return -EOPNOTSUPP; 243 } 244 #endif 245 246 static void esp_output_done(struct crypto_async_request *base, int err) 247 { 248 struct sk_buff *skb = base->data; 249 struct xfrm_offload *xo = xfrm_offload(skb); 250 void *tmp; 251 struct xfrm_state *x; 252 253 if (xo && (xo->flags & XFRM_DEV_RESUME)) { 254 struct sec_path *sp = skb_sec_path(skb); 255 256 x = sp->xvec[sp->len - 1]; 257 } else { 258 x = skb_dst(skb)->xfrm; 259 } 260 261 tmp = ESP_SKB_CB(skb)->tmp; 262 esp_ssg_unref(x, tmp); 263 kfree(tmp); 264 265 if (xo && (xo->flags & XFRM_DEV_RESUME)) { 266 if (err) { 267 XFRM_INC_STATS(xs_net(x), LINUX_MIB_XFRMOUTSTATEPROTOERROR); 268 kfree_skb(skb); 269 return; 270 } 271 272 skb_push(skb, skb->data - skb_mac_header(skb)); 273 secpath_reset(skb); 274 xfrm_dev_resume(skb); 275 } else { 276 if (!err && 277 x->encap && x->encap->encap_type == TCP_ENCAP_ESPINTCP) 278 esp_output_tail_tcp(x, skb); 279 else 280 xfrm_output_resume(skb->sk, skb, err); 281 } 282 } 283 284 /* Move ESP header back into place. */ 285 static void esp_restore_header(struct sk_buff *skb, unsigned int offset) 286 { 287 struct ip_esp_hdr *esph = (void *)(skb->data + offset); 288 void *tmp = ESP_SKB_CB(skb)->tmp; 289 __be32 *seqhi = esp_tmp_extra(tmp); 290 291 esph->seq_no = esph->spi; 292 esph->spi = *seqhi; 293 } 294 295 static void esp_output_restore_header(struct sk_buff *skb) 296 { 297 void *tmp = ESP_SKB_CB(skb)->tmp; 298 struct esp_output_extra *extra = esp_tmp_extra(tmp); 299 300 esp_restore_header(skb, skb_transport_offset(skb) + extra->esphoff - 301 sizeof(__be32)); 302 } 303 304 static struct ip_esp_hdr *esp_output_set_extra(struct sk_buff *skb, 305 struct xfrm_state *x, 306 struct ip_esp_hdr *esph, 307 struct esp_output_extra *extra) 308 { 309 /* For ESN we move the header forward by 4 bytes to 310 * accommodate the high bits. We will move it back after 311 * encryption. 312 */ 313 if ((x->props.flags & XFRM_STATE_ESN)) { 314 __u32 seqhi; 315 struct xfrm_offload *xo = xfrm_offload(skb); 316 317 if (xo) 318 seqhi = xo->seq.hi; 319 else 320 seqhi = XFRM_SKB_CB(skb)->seq.output.hi; 321 322 extra->esphoff = (unsigned char *)esph - 323 skb_transport_header(skb); 324 esph = (struct ip_esp_hdr *)((unsigned char *)esph - 4); 325 extra->seqhi = esph->spi; 326 esph->seq_no = htonl(seqhi); 327 } 328 329 esph->spi = x->id.spi; 330 331 return esph; 332 } 333 334 static void esp_output_done_esn(struct crypto_async_request *base, int err) 335 { 336 struct sk_buff *skb = base->data; 337 338 esp_output_restore_header(skb); 339 esp_output_done(base, err); 340 } 341 342 static struct ip_esp_hdr *esp_output_udp_encap(struct sk_buff *skb, 343 int encap_type, 344 struct esp_info *esp, 345 __be16 sport, 346 __be16 dport) 347 { 348 struct udphdr *uh; 349 __be32 *udpdata32; 350 unsigned int len; 351 352 len = skb->len + esp->tailen - skb_transport_offset(skb); 353 if (len + sizeof(struct iphdr) > IP_MAX_MTU) 354 return ERR_PTR(-EMSGSIZE); 355 356 uh = (struct udphdr *)esp->esph; 357 uh->source = sport; 358 uh->dest = dport; 359 uh->len = htons(len); 360 uh->check = 0; 361 362 *skb_mac_header(skb) = IPPROTO_UDP; 363 364 if (encap_type == UDP_ENCAP_ESPINUDP_NON_IKE) { 365 udpdata32 = (__be32 *)(uh + 1); 366 udpdata32[0] = udpdata32[1] = 0; 367 return (struct ip_esp_hdr *)(udpdata32 + 2); 368 } 369 370 return (struct ip_esp_hdr *)(uh + 1); 371 } 372 373 #ifdef CONFIG_INET_ESPINTCP 374 static struct ip_esp_hdr *esp_output_tcp_encap(struct xfrm_state *x, 375 struct sk_buff *skb, 376 struct esp_info *esp) 377 { 378 __be16 *lenp = (void *)esp->esph; 379 struct ip_esp_hdr *esph; 380 unsigned int len; 381 struct sock *sk; 382 383 len = skb->len + esp->tailen - skb_transport_offset(skb); 384 if (len > IP_MAX_MTU) 385 return ERR_PTR(-EMSGSIZE); 386 387 rcu_read_lock(); 388 sk = esp_find_tcp_sk(x); 389 rcu_read_unlock(); 390 391 if (IS_ERR(sk)) 392 return ERR_CAST(sk); 393 394 *lenp = htons(len); 395 esph = (struct ip_esp_hdr *)(lenp + 1); 396 397 return esph; 398 } 399 #else 400 static struct ip_esp_hdr *esp_output_tcp_encap(struct xfrm_state *x, 401 struct sk_buff *skb, 402 struct esp_info *esp) 403 { 404 return ERR_PTR(-EOPNOTSUPP); 405 } 406 #endif 407 408 static int esp_output_encap(struct xfrm_state *x, struct sk_buff *skb, 409 struct esp_info *esp) 410 { 411 struct xfrm_encap_tmpl *encap = x->encap; 412 struct ip_esp_hdr *esph; 413 __be16 sport, dport; 414 int encap_type; 415 416 spin_lock_bh(&x->lock); 417 sport = encap->encap_sport; 418 dport = encap->encap_dport; 419 encap_type = encap->encap_type; 420 spin_unlock_bh(&x->lock); 421 422 switch (encap_type) { 423 default: 424 case UDP_ENCAP_ESPINUDP: 425 case UDP_ENCAP_ESPINUDP_NON_IKE: 426 esph = esp_output_udp_encap(skb, encap_type, esp, sport, dport); 427 break; 428 case TCP_ENCAP_ESPINTCP: 429 esph = esp_output_tcp_encap(x, skb, esp); 430 break; 431 } 432 433 if (IS_ERR(esph)) 434 return PTR_ERR(esph); 435 436 esp->esph = esph; 437 438 return 0; 439 } 440 441 int esp_output_head(struct xfrm_state *x, struct sk_buff *skb, struct esp_info *esp) 442 { 443 u8 *tail; 444 int nfrags; 445 int esph_offset; 446 struct page *page; 447 struct sk_buff *trailer; 448 int tailen = esp->tailen; 449 450 /* this is non-NULL only with TCP/UDP Encapsulation */ 451 if (x->encap) { 452 int err = esp_output_encap(x, skb, esp); 453 454 if (err < 0) 455 return err; 456 } 457 458 if (ALIGN(tailen, L1_CACHE_BYTES) > PAGE_SIZE || 459 ALIGN(skb->data_len, L1_CACHE_BYTES) > PAGE_SIZE) 460 goto cow; 461 462 if (!skb_cloned(skb)) { 463 if (tailen <= skb_tailroom(skb)) { 464 nfrags = 1; 465 trailer = skb; 466 tail = skb_tail_pointer(trailer); 467 468 goto skip_cow; 469 } else if ((skb_shinfo(skb)->nr_frags < MAX_SKB_FRAGS) 470 && !skb_has_frag_list(skb)) { 471 int allocsize; 472 struct sock *sk = skb->sk; 473 struct page_frag *pfrag = &x->xfrag; 474 475 esp->inplace = false; 476 477 allocsize = ALIGN(tailen, L1_CACHE_BYTES); 478 479 spin_lock_bh(&x->lock); 480 481 if (unlikely(!skb_page_frag_refill(allocsize, pfrag, GFP_ATOMIC))) { 482 spin_unlock_bh(&x->lock); 483 goto cow; 484 } 485 486 page = pfrag->page; 487 get_page(page); 488 489 tail = page_address(page) + pfrag->offset; 490 491 esp_output_fill_trailer(tail, esp->tfclen, esp->plen, esp->proto); 492 493 nfrags = skb_shinfo(skb)->nr_frags; 494 495 __skb_fill_page_desc(skb, nfrags, page, pfrag->offset, 496 tailen); 497 skb_shinfo(skb)->nr_frags = ++nfrags; 498 499 pfrag->offset = pfrag->offset + allocsize; 500 501 spin_unlock_bh(&x->lock); 502 503 nfrags++; 504 505 skb->len += tailen; 506 skb->data_len += tailen; 507 skb->truesize += tailen; 508 if (sk && sk_fullsock(sk)) 509 refcount_add(tailen, &sk->sk_wmem_alloc); 510 511 goto out; 512 } 513 } 514 515 cow: 516 esph_offset = (unsigned char *)esp->esph - skb_transport_header(skb); 517 518 nfrags = skb_cow_data(skb, tailen, &trailer); 519 if (nfrags < 0) 520 goto out; 521 tail = skb_tail_pointer(trailer); 522 esp->esph = (struct ip_esp_hdr *)(skb_transport_header(skb) + esph_offset); 523 524 skip_cow: 525 esp_output_fill_trailer(tail, esp->tfclen, esp->plen, esp->proto); 526 pskb_put(skb, trailer, tailen); 527 528 out: 529 return nfrags; 530 } 531 EXPORT_SYMBOL_GPL(esp_output_head); 532 533 int esp_output_tail(struct xfrm_state *x, struct sk_buff *skb, struct esp_info *esp) 534 { 535 u8 *iv; 536 int alen; 537 void *tmp; 538 int ivlen; 539 int assoclen; 540 int extralen; 541 struct page *page; 542 struct ip_esp_hdr *esph; 543 struct crypto_aead *aead; 544 struct aead_request *req; 545 struct scatterlist *sg, *dsg; 546 struct esp_output_extra *extra; 547 int err = -ENOMEM; 548 549 assoclen = sizeof(struct ip_esp_hdr); 550 extralen = 0; 551 552 if (x->props.flags & XFRM_STATE_ESN) { 553 extralen += sizeof(*extra); 554 assoclen += sizeof(__be32); 555 } 556 557 aead = x->data; 558 alen = crypto_aead_authsize(aead); 559 ivlen = crypto_aead_ivsize(aead); 560 561 tmp = esp_alloc_tmp(aead, esp->nfrags + 2, extralen); 562 if (!tmp) 563 goto error; 564 565 extra = esp_tmp_extra(tmp); 566 iv = esp_tmp_iv(aead, tmp, extralen); 567 req = esp_tmp_req(aead, iv); 568 sg = esp_req_sg(aead, req); 569 570 if (esp->inplace) 571 dsg = sg; 572 else 573 dsg = &sg[esp->nfrags]; 574 575 esph = esp_output_set_extra(skb, x, esp->esph, extra); 576 esp->esph = esph; 577 578 sg_init_table(sg, esp->nfrags); 579 err = skb_to_sgvec(skb, sg, 580 (unsigned char *)esph - skb->data, 581 assoclen + ivlen + esp->clen + alen); 582 if (unlikely(err < 0)) 583 goto error_free; 584 585 if (!esp->inplace) { 586 int allocsize; 587 struct page_frag *pfrag = &x->xfrag; 588 589 allocsize = ALIGN(skb->data_len, L1_CACHE_BYTES); 590 591 spin_lock_bh(&x->lock); 592 if (unlikely(!skb_page_frag_refill(allocsize, pfrag, GFP_ATOMIC))) { 593 spin_unlock_bh(&x->lock); 594 goto error_free; 595 } 596 597 skb_shinfo(skb)->nr_frags = 1; 598 599 page = pfrag->page; 600 get_page(page); 601 /* replace page frags in skb with new page */ 602 __skb_fill_page_desc(skb, 0, page, pfrag->offset, skb->data_len); 603 pfrag->offset = pfrag->offset + allocsize; 604 spin_unlock_bh(&x->lock); 605 606 sg_init_table(dsg, skb_shinfo(skb)->nr_frags + 1); 607 err = skb_to_sgvec(skb, dsg, 608 (unsigned char *)esph - skb->data, 609 assoclen + ivlen + esp->clen + alen); 610 if (unlikely(err < 0)) 611 goto error_free; 612 } 613 614 if ((x->props.flags & XFRM_STATE_ESN)) 615 aead_request_set_callback(req, 0, esp_output_done_esn, skb); 616 else 617 aead_request_set_callback(req, 0, esp_output_done, skb); 618 619 aead_request_set_crypt(req, sg, dsg, ivlen + esp->clen, iv); 620 aead_request_set_ad(req, assoclen); 621 622 memset(iv, 0, ivlen); 623 memcpy(iv + ivlen - min(ivlen, 8), (u8 *)&esp->seqno + 8 - min(ivlen, 8), 624 min(ivlen, 8)); 625 626 ESP_SKB_CB(skb)->tmp = tmp; 627 err = crypto_aead_encrypt(req); 628 629 switch (err) { 630 case -EINPROGRESS: 631 goto error; 632 633 case -ENOSPC: 634 err = NET_XMIT_DROP; 635 break; 636 637 case 0: 638 if ((x->props.flags & XFRM_STATE_ESN)) 639 esp_output_restore_header(skb); 640 } 641 642 if (sg != dsg) 643 esp_ssg_unref(x, tmp); 644 645 if (!err && x->encap && x->encap->encap_type == TCP_ENCAP_ESPINTCP) 646 err = esp_output_tail_tcp(x, skb); 647 648 error_free: 649 kfree(tmp); 650 error: 651 return err; 652 } 653 EXPORT_SYMBOL_GPL(esp_output_tail); 654 655 static int esp_output(struct xfrm_state *x, struct sk_buff *skb) 656 { 657 int alen; 658 int blksize; 659 struct ip_esp_hdr *esph; 660 struct crypto_aead *aead; 661 struct esp_info esp; 662 663 esp.inplace = true; 664 665 esp.proto = *skb_mac_header(skb); 666 *skb_mac_header(skb) = IPPROTO_ESP; 667 668 /* skb is pure payload to encrypt */ 669 670 aead = x->data; 671 alen = crypto_aead_authsize(aead); 672 673 esp.tfclen = 0; 674 if (x->tfcpad) { 675 struct xfrm_dst *dst = (struct xfrm_dst *)skb_dst(skb); 676 u32 padto; 677 678 padto = min(x->tfcpad, xfrm_state_mtu(x, dst->child_mtu_cached)); 679 if (skb->len < padto) 680 esp.tfclen = padto - skb->len; 681 } 682 blksize = ALIGN(crypto_aead_blocksize(aead), 4); 683 esp.clen = ALIGN(skb->len + 2 + esp.tfclen, blksize); 684 esp.plen = esp.clen - skb->len - esp.tfclen; 685 esp.tailen = esp.tfclen + esp.plen + alen; 686 687 esp.esph = ip_esp_hdr(skb); 688 689 esp.nfrags = esp_output_head(x, skb, &esp); 690 if (esp.nfrags < 0) 691 return esp.nfrags; 692 693 esph = esp.esph; 694 esph->spi = x->id.spi; 695 696 esph->seq_no = htonl(XFRM_SKB_CB(skb)->seq.output.low); 697 esp.seqno = cpu_to_be64(XFRM_SKB_CB(skb)->seq.output.low + 698 ((u64)XFRM_SKB_CB(skb)->seq.output.hi << 32)); 699 700 skb_push(skb, -skb_network_offset(skb)); 701 702 return esp_output_tail(x, skb, &esp); 703 } 704 705 static inline int esp_remove_trailer(struct sk_buff *skb) 706 { 707 struct xfrm_state *x = xfrm_input_state(skb); 708 struct crypto_aead *aead = x->data; 709 int alen, hlen, elen; 710 int padlen, trimlen; 711 __wsum csumdiff; 712 u8 nexthdr[2]; 713 int ret; 714 715 alen = crypto_aead_authsize(aead); 716 hlen = sizeof(struct ip_esp_hdr) + crypto_aead_ivsize(aead); 717 elen = skb->len - hlen; 718 719 if (skb_copy_bits(skb, skb->len - alen - 2, nexthdr, 2)) 720 BUG(); 721 722 ret = -EINVAL; 723 padlen = nexthdr[0]; 724 if (padlen + 2 + alen >= elen) { 725 net_dbg_ratelimited("ipsec esp packet is garbage padlen=%d, elen=%d\n", 726 padlen + 2, elen - alen); 727 goto out; 728 } 729 730 trimlen = alen + padlen + 2; 731 if (skb->ip_summed == CHECKSUM_COMPLETE) { 732 csumdiff = skb_checksum(skb, skb->len - trimlen, trimlen, 0); 733 skb->csum = csum_block_sub(skb->csum, csumdiff, 734 skb->len - trimlen); 735 } 736 pskb_trim(skb, skb->len - trimlen); 737 738 ret = nexthdr[1]; 739 740 out: 741 return ret; 742 } 743 744 int esp_input_done2(struct sk_buff *skb, int err) 745 { 746 const struct iphdr *iph; 747 struct xfrm_state *x = xfrm_input_state(skb); 748 struct xfrm_offload *xo = xfrm_offload(skb); 749 struct crypto_aead *aead = x->data; 750 int hlen = sizeof(struct ip_esp_hdr) + crypto_aead_ivsize(aead); 751 int ihl; 752 753 if (!xo || !(xo->flags & CRYPTO_DONE)) 754 kfree(ESP_SKB_CB(skb)->tmp); 755 756 if (unlikely(err)) 757 goto out; 758 759 err = esp_remove_trailer(skb); 760 if (unlikely(err < 0)) 761 goto out; 762 763 iph = ip_hdr(skb); 764 ihl = iph->ihl * 4; 765 766 if (x->encap) { 767 struct xfrm_encap_tmpl *encap = x->encap; 768 struct tcphdr *th = (void *)(skb_network_header(skb) + ihl); 769 struct udphdr *uh = (void *)(skb_network_header(skb) + ihl); 770 __be16 source; 771 772 switch (x->encap->encap_type) { 773 case TCP_ENCAP_ESPINTCP: 774 source = th->source; 775 break; 776 case UDP_ENCAP_ESPINUDP: 777 case UDP_ENCAP_ESPINUDP_NON_IKE: 778 source = uh->source; 779 break; 780 default: 781 WARN_ON_ONCE(1); 782 err = -EINVAL; 783 goto out; 784 } 785 786 /* 787 * 1) if the NAT-T peer's IP or port changed then 788 * advertize the change to the keying daemon. 789 * This is an inbound SA, so just compare 790 * SRC ports. 791 */ 792 if (iph->saddr != x->props.saddr.a4 || 793 source != encap->encap_sport) { 794 xfrm_address_t ipaddr; 795 796 ipaddr.a4 = iph->saddr; 797 km_new_mapping(x, &ipaddr, source); 798 799 /* XXX: perhaps add an extra 800 * policy check here, to see 801 * if we should allow or 802 * reject a packet from a 803 * different source 804 * address/port. 805 */ 806 } 807 808 /* 809 * 2) ignore UDP/TCP checksums in case 810 * of NAT-T in Transport Mode, or 811 * perform other post-processing fixes 812 * as per draft-ietf-ipsec-udp-encaps-06, 813 * section 3.1.2 814 */ 815 if (x->props.mode == XFRM_MODE_TRANSPORT) 816 skb->ip_summed = CHECKSUM_UNNECESSARY; 817 } 818 819 skb_pull_rcsum(skb, hlen); 820 if (x->props.mode == XFRM_MODE_TUNNEL) 821 skb_reset_transport_header(skb); 822 else 823 skb_set_transport_header(skb, -ihl); 824 825 /* RFC4303: Drop dummy packets without any error */ 826 if (err == IPPROTO_NONE) 827 err = -EINVAL; 828 829 out: 830 return err; 831 } 832 EXPORT_SYMBOL_GPL(esp_input_done2); 833 834 static void esp_input_done(struct crypto_async_request *base, int err) 835 { 836 struct sk_buff *skb = base->data; 837 838 xfrm_input_resume(skb, esp_input_done2(skb, err)); 839 } 840 841 static void esp_input_restore_header(struct sk_buff *skb) 842 { 843 esp_restore_header(skb, 0); 844 __skb_pull(skb, 4); 845 } 846 847 static void esp_input_set_header(struct sk_buff *skb, __be32 *seqhi) 848 { 849 struct xfrm_state *x = xfrm_input_state(skb); 850 struct ip_esp_hdr *esph; 851 852 /* For ESN we move the header forward by 4 bytes to 853 * accommodate the high bits. We will move it back after 854 * decryption. 855 */ 856 if ((x->props.flags & XFRM_STATE_ESN)) { 857 esph = skb_push(skb, 4); 858 *seqhi = esph->spi; 859 esph->spi = esph->seq_no; 860 esph->seq_no = XFRM_SKB_CB(skb)->seq.input.hi; 861 } 862 } 863 864 static void esp_input_done_esn(struct crypto_async_request *base, int err) 865 { 866 struct sk_buff *skb = base->data; 867 868 esp_input_restore_header(skb); 869 esp_input_done(base, err); 870 } 871 872 /* 873 * Note: detecting truncated vs. non-truncated authentication data is very 874 * expensive, so we only support truncated data, which is the recommended 875 * and common case. 876 */ 877 static int esp_input(struct xfrm_state *x, struct sk_buff *skb) 878 { 879 struct crypto_aead *aead = x->data; 880 struct aead_request *req; 881 struct sk_buff *trailer; 882 int ivlen = crypto_aead_ivsize(aead); 883 int elen = skb->len - sizeof(struct ip_esp_hdr) - ivlen; 884 int nfrags; 885 int assoclen; 886 int seqhilen; 887 __be32 *seqhi; 888 void *tmp; 889 u8 *iv; 890 struct scatterlist *sg; 891 int err = -EINVAL; 892 893 if (!pskb_may_pull(skb, sizeof(struct ip_esp_hdr) + ivlen)) 894 goto out; 895 896 if (elen <= 0) 897 goto out; 898 899 assoclen = sizeof(struct ip_esp_hdr); 900 seqhilen = 0; 901 902 if (x->props.flags & XFRM_STATE_ESN) { 903 seqhilen += sizeof(__be32); 904 assoclen += seqhilen; 905 } 906 907 if (!skb_cloned(skb)) { 908 if (!skb_is_nonlinear(skb)) { 909 nfrags = 1; 910 911 goto skip_cow; 912 } else if (!skb_has_frag_list(skb)) { 913 nfrags = skb_shinfo(skb)->nr_frags; 914 nfrags++; 915 916 goto skip_cow; 917 } 918 } 919 920 err = skb_cow_data(skb, 0, &trailer); 921 if (err < 0) 922 goto out; 923 924 nfrags = err; 925 926 skip_cow: 927 err = -ENOMEM; 928 tmp = esp_alloc_tmp(aead, nfrags, seqhilen); 929 if (!tmp) 930 goto out; 931 932 ESP_SKB_CB(skb)->tmp = tmp; 933 seqhi = esp_tmp_extra(tmp); 934 iv = esp_tmp_iv(aead, tmp, seqhilen); 935 req = esp_tmp_req(aead, iv); 936 sg = esp_req_sg(aead, req); 937 938 esp_input_set_header(skb, seqhi); 939 940 sg_init_table(sg, nfrags); 941 err = skb_to_sgvec(skb, sg, 0, skb->len); 942 if (unlikely(err < 0)) { 943 kfree(tmp); 944 goto out; 945 } 946 947 skb->ip_summed = CHECKSUM_NONE; 948 949 if ((x->props.flags & XFRM_STATE_ESN)) 950 aead_request_set_callback(req, 0, esp_input_done_esn, skb); 951 else 952 aead_request_set_callback(req, 0, esp_input_done, skb); 953 954 aead_request_set_crypt(req, sg, sg, elen + ivlen, iv); 955 aead_request_set_ad(req, assoclen); 956 957 err = crypto_aead_decrypt(req); 958 if (err == -EINPROGRESS) 959 goto out; 960 961 if ((x->props.flags & XFRM_STATE_ESN)) 962 esp_input_restore_header(skb); 963 964 err = esp_input_done2(skb, err); 965 966 out: 967 return err; 968 } 969 970 static int esp4_err(struct sk_buff *skb, u32 info) 971 { 972 struct net *net = dev_net(skb->dev); 973 const struct iphdr *iph = (const struct iphdr *)skb->data; 974 struct ip_esp_hdr *esph = (struct ip_esp_hdr *)(skb->data+(iph->ihl<<2)); 975 struct xfrm_state *x; 976 977 switch (icmp_hdr(skb)->type) { 978 case ICMP_DEST_UNREACH: 979 if (icmp_hdr(skb)->code != ICMP_FRAG_NEEDED) 980 return 0; 981 break; 982 case ICMP_REDIRECT: 983 break; 984 default: 985 return 0; 986 } 987 988 x = xfrm_state_lookup(net, skb->mark, (const xfrm_address_t *)&iph->daddr, 989 esph->spi, IPPROTO_ESP, AF_INET); 990 if (!x) 991 return 0; 992 993 if (icmp_hdr(skb)->type == ICMP_DEST_UNREACH) 994 ipv4_update_pmtu(skb, net, info, 0, IPPROTO_ESP); 995 else 996 ipv4_redirect(skb, net, 0, IPPROTO_ESP); 997 xfrm_state_put(x); 998 999 return 0; 1000 } 1001 1002 static void esp_destroy(struct xfrm_state *x) 1003 { 1004 struct crypto_aead *aead = x->data; 1005 1006 if (!aead) 1007 return; 1008 1009 crypto_free_aead(aead); 1010 } 1011 1012 static int esp_init_aead(struct xfrm_state *x) 1013 { 1014 char aead_name[CRYPTO_MAX_ALG_NAME]; 1015 struct crypto_aead *aead; 1016 int err; 1017 1018 err = -ENAMETOOLONG; 1019 if (snprintf(aead_name, CRYPTO_MAX_ALG_NAME, "%s(%s)", 1020 x->geniv, x->aead->alg_name) >= CRYPTO_MAX_ALG_NAME) 1021 goto error; 1022 1023 aead = crypto_alloc_aead(aead_name, 0, 0); 1024 err = PTR_ERR(aead); 1025 if (IS_ERR(aead)) 1026 goto error; 1027 1028 x->data = aead; 1029 1030 err = crypto_aead_setkey(aead, x->aead->alg_key, 1031 (x->aead->alg_key_len + 7) / 8); 1032 if (err) 1033 goto error; 1034 1035 err = crypto_aead_setauthsize(aead, x->aead->alg_icv_len / 8); 1036 if (err) 1037 goto error; 1038 1039 error: 1040 return err; 1041 } 1042 1043 static int esp_init_authenc(struct xfrm_state *x) 1044 { 1045 struct crypto_aead *aead; 1046 struct crypto_authenc_key_param *param; 1047 struct rtattr *rta; 1048 char *key; 1049 char *p; 1050 char authenc_name[CRYPTO_MAX_ALG_NAME]; 1051 unsigned int keylen; 1052 int err; 1053 1054 err = -EINVAL; 1055 if (!x->ealg) 1056 goto error; 1057 1058 err = -ENAMETOOLONG; 1059 1060 if ((x->props.flags & XFRM_STATE_ESN)) { 1061 if (snprintf(authenc_name, CRYPTO_MAX_ALG_NAME, 1062 "%s%sauthencesn(%s,%s)%s", 1063 x->geniv ?: "", x->geniv ? "(" : "", 1064 x->aalg ? x->aalg->alg_name : "digest_null", 1065 x->ealg->alg_name, 1066 x->geniv ? ")" : "") >= CRYPTO_MAX_ALG_NAME) 1067 goto error; 1068 } else { 1069 if (snprintf(authenc_name, CRYPTO_MAX_ALG_NAME, 1070 "%s%sauthenc(%s,%s)%s", 1071 x->geniv ?: "", x->geniv ? "(" : "", 1072 x->aalg ? x->aalg->alg_name : "digest_null", 1073 x->ealg->alg_name, 1074 x->geniv ? ")" : "") >= CRYPTO_MAX_ALG_NAME) 1075 goto error; 1076 } 1077 1078 aead = crypto_alloc_aead(authenc_name, 0, 0); 1079 err = PTR_ERR(aead); 1080 if (IS_ERR(aead)) 1081 goto error; 1082 1083 x->data = aead; 1084 1085 keylen = (x->aalg ? (x->aalg->alg_key_len + 7) / 8 : 0) + 1086 (x->ealg->alg_key_len + 7) / 8 + RTA_SPACE(sizeof(*param)); 1087 err = -ENOMEM; 1088 key = kmalloc(keylen, GFP_KERNEL); 1089 if (!key) 1090 goto error; 1091 1092 p = key; 1093 rta = (void *)p; 1094 rta->rta_type = CRYPTO_AUTHENC_KEYA_PARAM; 1095 rta->rta_len = RTA_LENGTH(sizeof(*param)); 1096 param = RTA_DATA(rta); 1097 p += RTA_SPACE(sizeof(*param)); 1098 1099 if (x->aalg) { 1100 struct xfrm_algo_desc *aalg_desc; 1101 1102 memcpy(p, x->aalg->alg_key, (x->aalg->alg_key_len + 7) / 8); 1103 p += (x->aalg->alg_key_len + 7) / 8; 1104 1105 aalg_desc = xfrm_aalg_get_byname(x->aalg->alg_name, 0); 1106 BUG_ON(!aalg_desc); 1107 1108 err = -EINVAL; 1109 if (aalg_desc->uinfo.auth.icv_fullbits / 8 != 1110 crypto_aead_authsize(aead)) { 1111 pr_info("ESP: %s digestsize %u != %hu\n", 1112 x->aalg->alg_name, 1113 crypto_aead_authsize(aead), 1114 aalg_desc->uinfo.auth.icv_fullbits / 8); 1115 goto free_key; 1116 } 1117 1118 err = crypto_aead_setauthsize( 1119 aead, x->aalg->alg_trunc_len / 8); 1120 if (err) 1121 goto free_key; 1122 } 1123 1124 param->enckeylen = cpu_to_be32((x->ealg->alg_key_len + 7) / 8); 1125 memcpy(p, x->ealg->alg_key, (x->ealg->alg_key_len + 7) / 8); 1126 1127 err = crypto_aead_setkey(aead, key, keylen); 1128 1129 free_key: 1130 kfree(key); 1131 1132 error: 1133 return err; 1134 } 1135 1136 static int esp_init_state(struct xfrm_state *x) 1137 { 1138 struct crypto_aead *aead; 1139 u32 align; 1140 int err; 1141 1142 x->data = NULL; 1143 1144 if (x->aead) 1145 err = esp_init_aead(x); 1146 else 1147 err = esp_init_authenc(x); 1148 1149 if (err) 1150 goto error; 1151 1152 aead = x->data; 1153 1154 x->props.header_len = sizeof(struct ip_esp_hdr) + 1155 crypto_aead_ivsize(aead); 1156 if (x->props.mode == XFRM_MODE_TUNNEL) 1157 x->props.header_len += sizeof(struct iphdr); 1158 else if (x->props.mode == XFRM_MODE_BEET && x->sel.family != AF_INET6) 1159 x->props.header_len += IPV4_BEET_PHMAXLEN; 1160 if (x->encap) { 1161 struct xfrm_encap_tmpl *encap = x->encap; 1162 1163 switch (encap->encap_type) { 1164 default: 1165 err = -EINVAL; 1166 goto error; 1167 case UDP_ENCAP_ESPINUDP: 1168 x->props.header_len += sizeof(struct udphdr); 1169 break; 1170 case UDP_ENCAP_ESPINUDP_NON_IKE: 1171 x->props.header_len += sizeof(struct udphdr) + 2 * sizeof(u32); 1172 break; 1173 #ifdef CONFIG_INET_ESPINTCP 1174 case TCP_ENCAP_ESPINTCP: 1175 /* only the length field, TCP encap is done by 1176 * the socket 1177 */ 1178 x->props.header_len += 2; 1179 break; 1180 #endif 1181 } 1182 } 1183 1184 align = ALIGN(crypto_aead_blocksize(aead), 4); 1185 x->props.trailer_len = align + 1 + crypto_aead_authsize(aead); 1186 1187 error: 1188 return err; 1189 } 1190 1191 static int esp4_rcv_cb(struct sk_buff *skb, int err) 1192 { 1193 return 0; 1194 } 1195 1196 static const struct xfrm_type esp_type = 1197 { 1198 .owner = THIS_MODULE, 1199 .proto = IPPROTO_ESP, 1200 .flags = XFRM_TYPE_REPLAY_PROT, 1201 .init_state = esp_init_state, 1202 .destructor = esp_destroy, 1203 .input = esp_input, 1204 .output = esp_output, 1205 }; 1206 1207 static struct xfrm4_protocol esp4_protocol = { 1208 .handler = xfrm4_rcv, 1209 .input_handler = xfrm_input, 1210 .cb_handler = esp4_rcv_cb, 1211 .err_handler = esp4_err, 1212 .priority = 0, 1213 }; 1214 1215 static int __init esp4_init(void) 1216 { 1217 if (xfrm_register_type(&esp_type, AF_INET) < 0) { 1218 pr_info("%s: can't add xfrm type\n", __func__); 1219 return -EAGAIN; 1220 } 1221 if (xfrm4_protocol_register(&esp4_protocol, IPPROTO_ESP) < 0) { 1222 pr_info("%s: can't add protocol\n", __func__); 1223 xfrm_unregister_type(&esp_type, AF_INET); 1224 return -EAGAIN; 1225 } 1226 return 0; 1227 } 1228 1229 static void __exit esp4_fini(void) 1230 { 1231 if (xfrm4_protocol_deregister(&esp4_protocol, IPPROTO_ESP) < 0) 1232 pr_info("%s: can't remove protocol\n", __func__); 1233 xfrm_unregister_type(&esp_type, AF_INET); 1234 } 1235 1236 module_init(esp4_init); 1237 module_exit(esp4_fini); 1238 MODULE_LICENSE("GPL"); 1239 MODULE_ALIAS_XFRM_TYPE(AF_INET, XFRM_PROTO_ESP); 1240