net/tls/tls_device.c

   1 /* Copyright (c) 2018, Mellanox Technologies All rights reserved.
   2  *
   3  * This software is available to you under a choice of one of two
   4  * licenses.  You may choose to be licensed under the terms of the GNU
   5  * General Public License (GPL) Version 2, available from the file
   6  * COPYING in the main directory of this source tree, or the
   7  * OpenIB.org BSD license below:
   8  *
   9  *     Redistribution and use in source and binary forms, with or
  10  *     without modification, are permitted provided that the following
  11  *     conditions are met:
  12  *
  13  *      - Redistributions of source code must retain the above
  14  *        copyright notice, this list of conditions and the following
  15  *        disclaimer.
  16  *
  17  *      - Redistributions in binary form must reproduce the above
  18  *        copyright notice, this list of conditions and the following
  19  *        disclaimer in the documentation and/or other materials
  20  *        provided with the distribution.
  21  *
  22  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
  23  * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
  24  * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
  25  * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
  26  * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
  27  * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  28  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
  29  * SOFTWARE.
  30  */
  31
  32 #include <crypto/aead.h>
  33 #include <linux/highmem.h>
  34 #include <linux/module.h>
  35 #include <linux/netdevice.h>
  36 #include <net/dst.h>
  37 #include <net/inet_connection_sock.h>
  38 #include <net/tcp.h>
  39 #include <net/tls.h>
  40
  41 /* device_offload_lock is used to synchronize tls_dev_add
  42  * against NETDEV_DOWN notifications.
  43  */
  44 static DECLARE_RWSEM(device_offload_lock);
  45
  46 static void tls_device_gc_task(struct work_struct *work);
  47
  48 static DECLARE_WORK(tls_device_gc_work, tls_device_gc_task);
  49 static LIST_HEAD(tls_device_gc_list);
  50 static LIST_HEAD(tls_device_list);
  51 static DEFINE_SPINLOCK(tls_device_lock);
  52
  53 static void tls_device_free_ctx(struct tls_context *ctx)
  54 {
  55         if (ctx->tx_conf == TLS_HW) {
  56                 kfree(tls_offload_ctx_tx(ctx));
  57                 kfree(ctx->tx.rec_seq);
  58                 kfree(ctx->tx.iv);
  59         }
  60
  61         if (ctx->rx_conf == TLS_HW)
  62                 kfree(tls_offload_ctx_rx(ctx));
  63
  64         kfree(ctx);
  65 }
  66
  67 static void tls_device_gc_task(struct work_struct *work)
  68 {
  69         struct tls_context *ctx, *tmp;
  70         unsigned long flags;
  71         LIST_HEAD(gc_list);
  72
  73         spin_lock_irqsave(&tls_device_lock, flags);
  74         list_splice_init(&tls_device_gc_list, &gc_list);
  75         spin_unlock_irqrestore(&tls_device_lock, flags);
  76
  77         list_for_each_entry_safe(ctx, tmp, &gc_list, list) {
  78                 struct net_device *netdev = ctx->netdev;
  79
  80                 if (netdev && ctx->tx_conf == TLS_HW) {
  81                         netdev->tlsdev_ops->tls_dev_del(netdev, ctx,
  82                                                         TLS_OFFLOAD_CTX_DIR_TX);
  83                         dev_put(netdev);
  84                         ctx->netdev = NULL;
  85                 }
  86
  87                 list_del(&ctx->list);
  88                 tls_device_free_ctx(ctx);
  89         }
  90 }
  91
  92 static void tls_device_attach(struct tls_context *ctx, struct sock *sk,
  93                               struct net_device *netdev)
  94 {
  95         if (sk->sk_destruct != tls_device_sk_destruct) {
  96                 refcount_set(&ctx->refcount, 1);
  97                 dev_hold(netdev);
  98                 ctx->netdev = netdev;
  99                 spin_lock_irq(&tls_device_lock);
 100                 list_add_tail(&ctx->list, &tls_device_list);
 101                 spin_unlock_irq(&tls_device_lock);
 102
 103                 ctx->sk_destruct = sk->sk_destruct;
 104                 sk->sk_destruct = tls_device_sk_destruct;
 105         }
 106 }
 107
 108 static void tls_device_queue_ctx_destruction(struct tls_context *ctx)
 109 {
 110         unsigned long flags;
 111
 112         spin_lock_irqsave(&tls_device_lock, flags);
 113         list_move_tail(&ctx->list, &tls_device_gc_list);
 114
 115         /* schedule_work inside the spinlock
 116          * to make sure tls_device_down waits for that work.
 117          */
 118         schedule_work(&tls_device_gc_work);
 119
 120         spin_unlock_irqrestore(&tls_device_lock, flags);
 121 }
 122
 123 /* We assume that the socket is already connected */
 124 static struct net_device *get_netdev_for_sock(struct sock *sk)
 125 {
 126         struct dst_entry *dst = sk_dst_get(sk);
 127         struct net_device *netdev = NULL;
 128
 129         if (likely(dst)) {
 130                 netdev = dst->dev;
 131                 dev_hold(netdev);
 132         }
 133
 134         dst_release(dst);
 135
 136         return netdev;
 137 }
 138
 139 static void destroy_record(struct tls_record_info *record)
 140 {
 141         int nr_frags = record->num_frags;
 142         skb_frag_t *frag;
 143
 144         while (nr_frags-- > 0) {
 145                 frag = &record->frags[nr_frags];
 146                 __skb_frag_unref(frag);
 147         }
 148         kfree(record);
 149 }
 150
 151 static void delete_all_records(struct tls_offload_context_tx *offload_ctx)
 152 {
 153         struct tls_record_info *info, *temp;
 154
 155         list_for_each_entry_safe(info, temp, &offload_ctx->records_list, list) {
 156                 list_del(&info->list);
 157                 destroy_record(info);
 158         }
 159
 160         offload_ctx->retransmit_hint = NULL;
 161 }
 162
 163 static void tls_icsk_clean_acked(struct sock *sk, u32 acked_seq)
 164 {
 165         struct tls_context *tls_ctx = tls_get_ctx(sk);
 166         struct tls_record_info *info, *temp;
 167         struct tls_offload_context_tx *ctx;
 168         u64 deleted_records = 0;
 169         unsigned long flags;
 170
 171         if (!tls_ctx)
 172                 return;
 173
 174         ctx = tls_offload_ctx_tx(tls_ctx);
 175
 176         spin_lock_irqsave(&ctx->lock, flags);
 177         info = ctx->retransmit_hint;
 178         if (info && !before(acked_seq, info->end_seq)) {
 179                 ctx->retransmit_hint = NULL;
 180                 list_del(&info->list);
 181                 destroy_record(info);
 182                 deleted_records++;
 183         }
 184
 185         list_for_each_entry_safe(info, temp, &ctx->records_list, list) {
 186                 if (before(acked_seq, info->end_seq))
 187                         break;
 188                 list_del(&info->list);
 189
 190                 destroy_record(info);
 191                 deleted_records++;
 192         }
 193
 194         ctx->unacked_record_sn += deleted_records;
 195         spin_unlock_irqrestore(&ctx->lock, flags);
 196 }
 197
 198 /* At this point, there should be no references on this
 199  * socket and no in-flight SKBs associated with this
 200  * socket, so it is safe to free all the resources.
 201  */
 202 void tls_device_sk_destruct(struct sock *sk)
 203 {
 204         struct tls_context *tls_ctx = tls_get_ctx(sk);
 205         struct tls_offload_context_tx *ctx = tls_offload_ctx_tx(tls_ctx);
 206
 207         tls_ctx->sk_destruct(sk);
 208
 209         if (tls_ctx->tx_conf == TLS_HW) {
 210                 if (ctx->open_record)
 211                         destroy_record(ctx->open_record);
 212                 delete_all_records(ctx);
 213                 crypto_free_aead(ctx->aead_send);
 214                 clean_acked_data_disable(inet_csk(sk));
 215         }
 216
 217         if (refcount_dec_and_test(&tls_ctx->refcount))
 218                 tls_device_queue_ctx_destruction(tls_ctx);
 219 }
 220 EXPORT_SYMBOL(tls_device_sk_destruct);
 221
 222 static void tls_append_frag(struct tls_record_info *record,
 223                             struct page_frag *pfrag,
 224                             int size)
 225 {
 226         skb_frag_t *frag;
 227
 228         frag = &record->frags[record->num_frags - 1];
 229         if (frag->page.p == pfrag->page &&
 230             frag->page_offset + frag->size == pfrag->offset) {
 231                 frag->size += size;
 232         } else {
 233                 ++frag;
 234                 frag->page.p = pfrag->page;
 235                 frag->page_offset = pfrag->offset;
 236                 frag->size = size;
 237                 ++record->num_frags;
 238                 get_page(pfrag->page);
 239         }
 240
 241         pfrag->offset += size;
 242         record->len += size;
 243 }
 244
 245 static int tls_push_record(struct sock *sk,
 246                            struct tls_context *ctx,
 247                            struct tls_offload_context_tx *offload_ctx,
 248                            struct tls_record_info *record,
 249                            struct page_frag *pfrag,
 250                            int flags,
 251                            unsigned char record_type)
 252 {
 253         struct tls_prot_info *prot = &ctx->prot_info;
 254         struct tcp_sock *tp = tcp_sk(sk);
 255         struct page_frag dummy_tag_frag;
 256         skb_frag_t *frag;
 257         int i;
 258
 259         /* fill prepend */
 260         frag = &record->frags[0];
 261         tls_fill_prepend(ctx,
 262                          skb_frag_address(frag),
 263                          record->len - prot->prepend_size,
 264                          record_type,
 265                          ctx->crypto_send.info.version);
 266
 267         /* HW doesn't care about the data in the tag, because it fills it. */
 268         dummy_tag_frag.page = skb_frag_page(frag);
 269         dummy_tag_frag.offset = 0;
 270
 271         tls_append_frag(record, &dummy_tag_frag, prot->tag_size);
 272         record->end_seq = tp->write_seq + record->len;
 273         spin_lock_irq(&offload_ctx->lock);
 274         list_add_tail(&record->list, &offload_ctx->records_list);
 275         spin_unlock_irq(&offload_ctx->lock);
 276         offload_ctx->open_record = NULL;
 277         tls_advance_record_sn(sk, &ctx->tx, ctx->crypto_send.info.version);
 278
 279         for (i = 0; i < record->num_frags; i++) {
 280                 frag = &record->frags[i];
 281                 sg_unmark_end(&offload_ctx->sg_tx_data[i]);
 282                 sg_set_page(&offload_ctx->sg_tx_data[i], skb_frag_page(frag),
 283                             frag->size, frag->page_offset);
 284                 sk_mem_charge(sk, frag->size);
 285                 get_page(skb_frag_page(frag));
 286         }
 287         sg_mark_end(&offload_ctx->sg_tx_data[record->num_frags - 1]);
 288
 289         /* all ready, send */
 290         return tls_push_sg(sk, ctx, offload_ctx->sg_tx_data, 0, flags);
 291 }
 292
 293 static int tls_create_new_record(struct tls_offload_context_tx *offload_ctx,
 294                                  struct page_frag *pfrag,
 295                                  size_t prepend_size)
 296 {
 297         struct tls_record_info *record;
 298         skb_frag_t *frag;
 299
 300         record = kmalloc(sizeof(*record), GFP_KERNEL);
 301         if (!record)
 302                 return -ENOMEM;
 303
 304         frag = &record->frags[0];
 305         __skb_frag_set_page(frag, pfrag->page);
 306         frag->page_offset = pfrag->offset;
 307         skb_frag_size_set(frag, prepend_size);
 308
 309         get_page(pfrag->page);
 310         pfrag->offset += prepend_size;
 311
 312         record->num_frags = 1;
 313         record->len = prepend_size;
 314         offload_ctx->open_record = record;
 315         return 0;
 316 }
 317
 318 static int tls_do_allocation(struct sock *sk,
 319                              struct tls_offload_context_tx *offload_ctx,
 320                              struct page_frag *pfrag,
 321                              size_t prepend_size)
 322 {
 323         int ret;
 324
 325         if (!offload_ctx->open_record) {
 326                 if (unlikely(!skb_page_frag_refill(prepend_size, pfrag,
 327                                                    sk->sk_allocation))) {
 328                         sk->sk_prot->enter_memory_pressure(sk);
 329                         sk_stream_moderate_sndbuf(sk);
 330                         return -ENOMEM;
 331                 }
 332
 333                 ret = tls_create_new_record(offload_ctx, pfrag, prepend_size);
 334                 if (ret)
 335                         return ret;
 336
 337                 if (pfrag->size > pfrag->offset)
 338                         return 0;
 339         }
 340
 341         if (!sk_page_frag_refill(sk, pfrag))
 342                 return -ENOMEM;
 343
 344         return 0;
 345 }
 346
 347 static int tls_push_data(struct sock *sk,
 348                          struct iov_iter *msg_iter,
 349                          size_t size, int flags,
 350                          unsigned char record_type)
 351 {
 352         struct tls_context *tls_ctx = tls_get_ctx(sk);
 353         struct tls_prot_info *prot = &tls_ctx->prot_info;
 354         struct tls_offload_context_tx *ctx = tls_offload_ctx_tx(tls_ctx);
 355         int tls_push_record_flags = flags | MSG_SENDPAGE_NOTLAST;
 356         int more = flags & (MSG_SENDPAGE_NOTLAST | MSG_MORE);
 357         struct tls_record_info *record = ctx->open_record;
 358         struct page_frag *pfrag;
 359         size_t orig_size = size;
 360         u32 max_open_record_len;
 361         int copy, rc = 0;
 362         bool done = false;
 363         long timeo;
 364
 365         if (flags &
 366             ~(MSG_MORE | MSG_DONTWAIT | MSG_NOSIGNAL | MSG_SENDPAGE_NOTLAST))
 367                 return -ENOTSUPP;
 368
 369         if (sk->sk_err)
 370                 return -sk->sk_err;
 371
 372         timeo = sock_sndtimeo(sk, flags & MSG_DONTWAIT);
 373         if (tls_is_partially_sent_record(tls_ctx)) {
 374                 rc = tls_push_partial_record(sk, tls_ctx, flags);
 375                 if (rc < 0)
 376                         return rc;
 377         }
 378
 379         pfrag = sk_page_frag(sk);
 380
 381         /* TLS_HEADER_SIZE is not counted as part of the TLS record, and
 382          * we need to leave room for an authentication tag.
 383          */
 384         max_open_record_len = TLS_MAX_PAYLOAD_SIZE +
 385                               prot->prepend_size;
 386         do {
 387                 rc = tls_do_allocation(sk, ctx, pfrag,
 388                                        prot->prepend_size);
 389                 if (rc) {
 390                         rc = sk_stream_wait_memory(sk, &timeo);
 391                         if (!rc)
 392                                 continue;
 393
 394                         record = ctx->open_record;
 395                         if (!record)
 396                                 break;
 397 handle_error:
 398                         if (record_type != TLS_RECORD_TYPE_DATA) {
 399                                 /* avoid sending partial
 400                                  * record with type !=
 401                                  * application_data
 402                                  */
 403                                 size = orig_size;
 404                                 destroy_record(record);
 405                                 ctx->open_record = NULL;
 406                         } else if (record->len > prot->prepend_size) {
 407                                 goto last_record;
 408                         }
 409
 410                         break;
 411                 }
 412
 413                 record = ctx->open_record;
 414                 copy = min_t(size_t, size, (pfrag->size - pfrag->offset));
 415                 copy = min_t(size_t, copy, (max_open_record_len - record->len));
 416
 417                 if (copy_from_iter_nocache(page_address(pfrag->page) +
 418                                                pfrag->offset,
 419                                            copy, msg_iter) != copy) {
 420                         rc = -EFAULT;
 421                         goto handle_error;
 422                 }
 423                 tls_append_frag(record, pfrag, copy);
 424
 425                 size -= copy;
 426                 if (!size) {
 427 last_record:
 428                         tls_push_record_flags = flags;
 429                         if (more) {
 430                                 tls_ctx->pending_open_record_frags =
 431                                                 !!record->num_frags;
 432                                 break;
 433                         }
 434
 435                         done = true;
 436                 }
 437
 438                 if (done || record->len >= max_open_record_len ||
 439                     (record->num_frags >= MAX_SKB_FRAGS - 1)) {
 440                         rc = tls_push_record(sk,
 441                                              tls_ctx,
 442                                              ctx,
 443                                              record,
 444                                              pfrag,
 445                                              tls_push_record_flags,
 446                                              record_type);
 447                         if (rc < 0)
 448                                 break;
 449                 }
 450         } while (!done);
 451
 452         if (orig_size - size > 0)
 453                 rc = orig_size - size;
 454
 455         return rc;
 456 }
 457
 458 int tls_device_sendmsg(struct sock *sk, struct msghdr *msg, size_t size)
 459 {
 460         unsigned char record_type = TLS_RECORD_TYPE_DATA;
 461         int rc;
 462
 463         lock_sock(sk);
 464
 465         if (unlikely(msg->msg_controllen)) {
 466                 rc = tls_proccess_cmsg(sk, msg, &record_type);
 467                 if (rc)
 468                         goto out;
 469         }
 470
 471         rc = tls_push_data(sk, &msg->msg_iter, size,
 472                            msg->msg_flags, record_type);
 473
 474 out:
 475         release_sock(sk);
 476         return rc;
 477 }
 478
 479 int tls_device_sendpage(struct sock *sk, struct page *page,
 480                         int offset, size_t size, int flags)
 481 {
 482         struct iov_iter msg_iter;
 483         char *kaddr = kmap(page);
 484         struct kvec iov;
 485         int rc;
 486
 487         if (flags & MSG_SENDPAGE_NOTLAST)
 488                 flags |= MSG_MORE;
 489
 490         lock_sock(sk);
 491
 492         if (flags & MSG_OOB) {
 493                 rc = -ENOTSUPP;
 494                 goto out;
 495         }
 496
 497         iov.iov_base = kaddr + offset;
 498         iov.iov_len = size;
 499         iov_iter_kvec(&msg_iter, WRITE, &iov, 1, size);
 500         rc = tls_push_data(sk, &msg_iter, size,
 501                            flags, TLS_RECORD_TYPE_DATA);
 502         kunmap(page);
 503
 504 out:
 505         release_sock(sk);
 506         return rc;
 507 }
 508
 509 struct tls_record_info *tls_get_record(struct tls_offload_context_tx *context,
 510                                        u32 seq, u64 *p_record_sn)
 511 {
 512         u64 record_sn = context->hint_record_sn;
 513         struct tls_record_info *info;
 514
 515         info = context->retransmit_hint;
 516         if (!info ||
 517             before(seq, info->end_seq - info->len)) {
 518                 /* if retransmit_hint is irrelevant start
 519                  * from the beggining of the list
 520                  */
 521                 info = list_first_entry(&context->records_list,
 522                                         struct tls_record_info, list);
 523                 record_sn = context->unacked_record_sn;
 524         }
 525
 526         list_for_each_entry_from(info, &context->records_list, list) {
 527                 if (before(seq, info->end_seq)) {
 528                         if (!context->retransmit_hint ||
 529                             after(info->end_seq,
 530                                   context->retransmit_hint->end_seq)) {
 531                                 context->hint_record_sn = record_sn;
 532                                 context->retransmit_hint = info;
 533                         }
 534                         *p_record_sn = record_sn;
 535                         return info;
 536                 }
 537                 record_sn++;
 538         }
 539
 540         return NULL;
 541 }
 542 EXPORT_SYMBOL(tls_get_record);
 543
 544 static int tls_device_push_pending_record(struct sock *sk, int flags)
 545 {
 546         struct iov_iter msg_iter;
 547
 548         iov_iter_kvec(&msg_iter, WRITE, NULL, 0, 0);
 549         return tls_push_data(sk, &msg_iter, 0, flags, TLS_RECORD_TYPE_DATA);
 550 }
 551
 552 void tls_device_write_space(struct sock *sk, struct tls_context *ctx)
 553 {
 554         int rc = 0;
 555
 556         if (!sk->sk_write_pending && tls_is_partially_sent_record(ctx)) {
 557                 gfp_t sk_allocation = sk->sk_allocation;
 558
 559                 sk->sk_allocation = GFP_ATOMIC;
 560                 rc = tls_push_partial_record(sk, ctx,
 561                                              MSG_DONTWAIT | MSG_NOSIGNAL);
 562                 sk->sk_allocation = sk_allocation;
 563         }
 564 }
 565
 566 void handle_device_resync(struct sock *sk, u32 seq, u64 rcd_sn)
 567 {
 568         struct tls_context *tls_ctx = tls_get_ctx(sk);
 569         struct net_device *netdev = tls_ctx->netdev;
 570         struct tls_offload_context_rx *rx_ctx;
 571         u32 is_req_pending;
 572         s64 resync_req;
 573         u32 req_seq;
 574
 575         if (tls_ctx->rx_conf != TLS_HW)
 576                 return;
 577
 578         rx_ctx = tls_offload_ctx_rx(tls_ctx);
 579         resync_req = atomic64_read(&rx_ctx->resync_req);
 580         req_seq = ntohl(resync_req >> 32) - ((u32)TLS_HEADER_SIZE - 1);
 581         is_req_pending = resync_req;
 582
 583         if (unlikely(is_req_pending) && req_seq == seq &&
 584             atomic64_try_cmpxchg(&rx_ctx->resync_req, &resync_req, 0))
 585                 netdev->tlsdev_ops->tls_dev_resync_rx(netdev, sk,
 586                                                       seq + TLS_HEADER_SIZE - 1,
 587                                                       rcd_sn);
 588 }
 589
 590 static int tls_device_reencrypt(struct sock *sk, struct sk_buff *skb)
 591 {
 592         struct strp_msg *rxm = strp_msg(skb);
 593         int err = 0, offset = rxm->offset, copy, nsg;
 594         struct sk_buff *skb_iter, *unused;
 595         struct scatterlist sg[1];
 596         char *orig_buf, *buf;
 597
 598         orig_buf = kmalloc(rxm->full_len + TLS_HEADER_SIZE +
 599                            TLS_CIPHER_AES_GCM_128_IV_SIZE, sk->sk_allocation);
 600         if (!orig_buf)
 601                 return -ENOMEM;
 602         buf = orig_buf;
 603
 604         nsg = skb_cow_data(skb, 0, &unused);
 605         if (unlikely(nsg < 0)) {
 606                 err = nsg;
 607                 goto free_buf;
 608         }
 609
 610         sg_init_table(sg, 1);
 611         sg_set_buf(&sg[0], buf,
 612                    rxm->full_len + TLS_HEADER_SIZE +
 613                    TLS_CIPHER_AES_GCM_128_IV_SIZE);
 614         skb_copy_bits(skb, offset, buf,
 615                       TLS_HEADER_SIZE + TLS_CIPHER_AES_GCM_128_IV_SIZE);
 616
 617         /* We are interested only in the decrypted data not the auth */
 618         err = decrypt_skb(sk, skb, sg);
 619         if (err != -EBADMSG)
 620                 goto free_buf;
 621         else
 622                 err = 0;
 623
 624         copy = min_t(int, skb_pagelen(skb) - offset,
 625                      rxm->full_len - TLS_CIPHER_AES_GCM_128_TAG_SIZE);
 626
 627         if (skb->decrypted)
 628                 skb_store_bits(skb, offset, buf, copy);
 629
 630         offset += copy;
 631         buf += copy;
 632
 633         skb_walk_frags(skb, skb_iter) {
 634                 copy = min_t(int, skb_iter->len,
 635                              rxm->full_len - offset + rxm->offset -
 636                              TLS_CIPHER_AES_GCM_128_TAG_SIZE);
 637
 638                 if (skb_iter->decrypted)
 639                         skb_store_bits(skb_iter, offset, buf, copy);
 640
 641                 offset += copy;
 642                 buf += copy;
 643         }
 644
 645 free_buf:
 646         kfree(orig_buf);
 647         return err;
 648 }
 649
 650 int tls_device_decrypted(struct sock *sk, struct sk_buff *skb)
 651 {
 652         struct tls_context *tls_ctx = tls_get_ctx(sk);
 653         struct tls_offload_context_rx *ctx = tls_offload_ctx_rx(tls_ctx);
 654         int is_decrypted = skb->decrypted;
 655         int is_encrypted = !is_decrypted;
 656         struct sk_buff *skb_iter;
 657
 658         /* Skip if it is already decrypted */
 659         if (ctx->sw.decrypted)
 660                 return 0;
 661
 662         /* Check if all the data is decrypted already */
 663         skb_walk_frags(skb, skb_iter) {
 664                 is_decrypted &= skb_iter->decrypted;
 665                 is_encrypted &= !skb_iter->decrypted;
 666         }
 667
 668         ctx->sw.decrypted |= is_decrypted;
 669
 670         /* Return immedeatly if the record is either entirely plaintext or
 671          * entirely ciphertext. Otherwise handle reencrypt partially decrypted
 672          * record.
 673          */
 674         return (is_encrypted || is_decrypted) ? 0 :
 675                 tls_device_reencrypt(sk, skb);
 676 }
 677
 678 int tls_set_device_offload(struct sock *sk, struct tls_context *ctx)
 679 {
 680         u16 nonce_size, tag_size, iv_size, rec_seq_size;
 681         struct tls_context *tls_ctx = tls_get_ctx(sk);
 682         struct tls_prot_info *prot = &tls_ctx->prot_info;
 683         struct tls_record_info *start_marker_record;
 684         struct tls_offload_context_tx *offload_ctx;
 685         struct tls_crypto_info *crypto_info;
 686         struct net_device *netdev;
 687         char *iv, *rec_seq;
 688         struct sk_buff *skb;
 689         int rc = -EINVAL;
 690         __be64 rcd_sn;
 691
 692         if (!ctx)
 693                 goto out;
 694
 695         if (ctx->priv_ctx_tx) {
 696                 rc = -EEXIST;
 697                 goto out;
 698         }
 699
 700         start_marker_record = kmalloc(sizeof(*start_marker_record), GFP_KERNEL);
 701         if (!start_marker_record) {
 702                 rc = -ENOMEM;
 703                 goto out;
 704         }
 705
 706         offload_ctx = kzalloc(TLS_OFFLOAD_CONTEXT_SIZE_TX, GFP_KERNEL);
 707         if (!offload_ctx) {
 708                 rc = -ENOMEM;
 709                 goto free_marker_record;
 710         }
 711
 712         crypto_info = &ctx->crypto_send.info;
 713         switch (crypto_info->cipher_type) {
 714         case TLS_CIPHER_AES_GCM_128:
 715                 nonce_size = TLS_CIPHER_AES_GCM_128_IV_SIZE;
 716                 tag_size = TLS_CIPHER_AES_GCM_128_TAG_SIZE;
 717                 iv_size = TLS_CIPHER_AES_GCM_128_IV_SIZE;
 718                 iv = ((struct tls12_crypto_info_aes_gcm_128 *)crypto_info)->iv;
 719                 rec_seq_size = TLS_CIPHER_AES_GCM_128_REC_SEQ_SIZE;
 720                 rec_seq =
 721                  ((struct tls12_crypto_info_aes_gcm_128 *)crypto_info)->rec_seq;
 722                 break;
 723         default:
 724                 rc = -EINVAL;
 725                 goto free_offload_ctx;
 726         }
 727
 728         prot->prepend_size = TLS_HEADER_SIZE + nonce_size;
 729         prot->tag_size = tag_size;
 730         prot->overhead_size = prot->prepend_size + prot->tag_size;
 731         prot->iv_size = iv_size;
 732         ctx->tx.iv = kmalloc(iv_size + TLS_CIPHER_AES_GCM_128_SALT_SIZE,
 733                              GFP_KERNEL);
 734         if (!ctx->tx.iv) {
 735                 rc = -ENOMEM;
 736                 goto free_offload_ctx;
 737         }
 738
 739         memcpy(ctx->tx.iv + TLS_CIPHER_AES_GCM_128_SALT_SIZE, iv, iv_size);
 740
 741         prot->rec_seq_size = rec_seq_size;
 742         ctx->tx.rec_seq = kmemdup(rec_seq, rec_seq_size, GFP_KERNEL);
 743         if (!ctx->tx.rec_seq) {
 744                 rc = -ENOMEM;
 745                 goto free_iv;
 746         }
 747
 748         rc = tls_sw_fallback_init(sk, offload_ctx, crypto_info);
 749         if (rc)
 750                 goto free_rec_seq;
 751
 752         /* start at rec_seq - 1 to account for the start marker record */
 753         memcpy(&rcd_sn, ctx->tx.rec_seq, sizeof(rcd_sn));
 754         offload_ctx->unacked_record_sn = be64_to_cpu(rcd_sn) - 1;
 755
 756         start_marker_record->end_seq = tcp_sk(sk)->write_seq;
 757         start_marker_record->len = 0;
 758         start_marker_record->num_frags = 0;
 759
 760         INIT_LIST_HEAD(&offload_ctx->records_list);
 761         list_add_tail(&start_marker_record->list, &offload_ctx->records_list);
 762         spin_lock_init(&offload_ctx->lock);
 763         sg_init_table(offload_ctx->sg_tx_data,
 764                       ARRAY_SIZE(offload_ctx->sg_tx_data));
 765
 766         clean_acked_data_enable(inet_csk(sk), &tls_icsk_clean_acked);
 767         ctx->push_pending_record = tls_device_push_pending_record;
 768
 769         /* TLS offload is greatly simplified if we don't send
 770          * SKBs where only part of the payload needs to be encrypted.
 771          * So mark the last skb in the write queue as end of record.
 772          */
 773         skb = tcp_write_queue_tail(sk);
 774         if (skb)
 775                 TCP_SKB_CB(skb)->eor = 1;
 776
 777         /* We support starting offload on multiple sockets
 778          * concurrently, so we only need a read lock here.
 779          * This lock must precede get_netdev_for_sock to prevent races between
 780          * NETDEV_DOWN and setsockopt.
 781          */
 782         down_read(&device_offload_lock);
 783         netdev = get_netdev_for_sock(sk);
 784         if (!netdev) {
 785                 pr_err_ratelimited("%s: netdev not found\n", __func__);
 786                 rc = -EINVAL;
 787                 goto release_lock;
 788         }
 789
 790         if (!(netdev->features & NETIF_F_HW_TLS_TX)) {
 791                 rc = -ENOTSUPP;
 792                 goto release_netdev;
 793         }
 794
 795         /* Avoid offloading if the device is down
 796          * We don't want to offload new flows after
 797          * the NETDEV_DOWN event
 798          */
 799         if (!(netdev->flags & IFF_UP)) {
 800                 rc = -EINVAL;
 801                 goto release_netdev;
 802         }
 803
 804         ctx->priv_ctx_tx = offload_ctx;
 805         rc = netdev->tlsdev_ops->tls_dev_add(netdev, sk, TLS_OFFLOAD_CTX_DIR_TX,
 806                                              &ctx->crypto_send.info,
 807                                              tcp_sk(sk)->write_seq);
 808         if (rc)
 809                 goto release_netdev;
 810
 811         tls_device_attach(ctx, sk, netdev);
 812
 813         /* following this assignment tls_is_sk_tx_device_offloaded
 814          * will return true and the context might be accessed
 815          * by the netdev's xmit function.
 816          */
 817         smp_store_release(&sk->sk_validate_xmit_skb, tls_validate_xmit_skb);
 818         dev_put(netdev);
 819         up_read(&device_offload_lock);
 820         goto out;
 821
 822 release_netdev:
 823         dev_put(netdev);
 824 release_lock:
 825         up_read(&device_offload_lock);
 826         clean_acked_data_disable(inet_csk(sk));
 827         crypto_free_aead(offload_ctx->aead_send);
 828 free_rec_seq:
 829         kfree(ctx->tx.rec_seq);
 830 free_iv:
 831         kfree(ctx->tx.iv);
 832 free_offload_ctx:
 833         kfree(offload_ctx);
 834         ctx->priv_ctx_tx = NULL;
 835 free_marker_record:
 836         kfree(start_marker_record);
 837 out:
 838         return rc;
 839 }
 840
 841 int tls_set_device_offload_rx(struct sock *sk, struct tls_context *ctx)
 842 {
 843         struct tls_offload_context_rx *context;
 844         struct net_device *netdev;
 845         int rc = 0;
 846
 847         /* We support starting offload on multiple sockets
 848          * concurrently, so we only need a read lock here.
 849          * This lock must precede get_netdev_for_sock to prevent races between
 850          * NETDEV_DOWN and setsockopt.
 851          */
 852         down_read(&device_offload_lock);
 853         netdev = get_netdev_for_sock(sk);
 854         if (!netdev) {
 855                 pr_err_ratelimited("%s: netdev not found\n", __func__);
 856                 rc = -EINVAL;
 857                 goto release_lock;
 858         }
 859
 860         if (!(netdev->features & NETIF_F_HW_TLS_RX)) {
 861                 pr_err_ratelimited("%s: netdev %s with no TLS offload\n",
 862                                    __func__, netdev->name);
 863                 rc = -ENOTSUPP;
 864                 goto release_netdev;
 865         }
 866
 867         /* Avoid offloading if the device is down
 868          * We don't want to offload new flows after
 869          * the NETDEV_DOWN event
 870          */
 871         if (!(netdev->flags & IFF_UP)) {
 872                 rc = -EINVAL;
 873                 goto release_netdev;
 874         }
 875
 876         context = kzalloc(TLS_OFFLOAD_CONTEXT_SIZE_RX, GFP_KERNEL);
 877         if (!context) {
 878                 rc = -ENOMEM;
 879                 goto release_netdev;
 880         }
 881
 882         ctx->priv_ctx_rx = context;
 883         rc = tls_set_sw_offload(sk, ctx, 0);
 884         if (rc)
 885                 goto release_ctx;
 886
 887         rc = netdev->tlsdev_ops->tls_dev_add(netdev, sk, TLS_OFFLOAD_CTX_DIR_RX,
 888                                              &ctx->crypto_recv.info,
 889                                              tcp_sk(sk)->copied_seq);
 890         if (rc) {
 891                 pr_err_ratelimited("%s: The netdev has refused to offload this socket\n",
 892                                    __func__);
 893                 goto free_sw_resources;
 894         }
 895
 896         tls_device_attach(ctx, sk, netdev);
 897         goto release_netdev;
 898
 899 free_sw_resources:
 900         tls_sw_free_resources_rx(sk);
 901 release_ctx:
 902         ctx->priv_ctx_rx = NULL;
 903 release_netdev:
 904         dev_put(netdev);
 905 release_lock:
 906         up_read(&device_offload_lock);
 907         return rc;
 908 }
 909
 910 void tls_device_offload_cleanup_rx(struct sock *sk)
 911 {
 912         struct tls_context *tls_ctx = tls_get_ctx(sk);
 913         struct net_device *netdev;
 914
 915         down_read(&device_offload_lock);
 916         netdev = tls_ctx->netdev;
 917         if (!netdev)
 918                 goto out;
 919
 920         if (!(netdev->features & NETIF_F_HW_TLS_RX)) {
 921                 pr_err_ratelimited("%s: device is missing NETIF_F_HW_TLS_RX cap\n",
 922                                    __func__);
 923                 goto out;
 924         }
 925
 926         netdev->tlsdev_ops->tls_dev_del(netdev, tls_ctx,
 927                                         TLS_OFFLOAD_CTX_DIR_RX);
 928
 929         if (tls_ctx->tx_conf != TLS_HW) {
 930                 dev_put(netdev);
 931                 tls_ctx->netdev = NULL;
 932         }
 933 out:
 934         up_read(&device_offload_lock);
 935         kfree(tls_ctx->rx.rec_seq);
 936         kfree(tls_ctx->rx.iv);
 937         tls_sw_release_resources_rx(sk);
 938 }
 939
 940 static int tls_device_down(struct net_device *netdev)
 941 {
 942         struct tls_context *ctx, *tmp;
 943         unsigned long flags;
 944         LIST_HEAD(list);
 945
 946         /* Request a write lock to block new offload attempts */
 947         down_write(&device_offload_lock);
 948
 949         spin_lock_irqsave(&tls_device_lock, flags);
 950         list_for_each_entry_safe(ctx, tmp, &tls_device_list, list) {
 951                 if (ctx->netdev != netdev ||
 952                     !refcount_inc_not_zero(&ctx->refcount))
 953                         continue;
 954
 955                 list_move(&ctx->list, &list);
 956         }
 957         spin_unlock_irqrestore(&tls_device_lock, flags);
 958
 959         list_for_each_entry_safe(ctx, tmp, &list, list) {
 960                 if (ctx->tx_conf == TLS_HW)
 961                         netdev->tlsdev_ops->tls_dev_del(netdev, ctx,
 962                                                         TLS_OFFLOAD_CTX_DIR_TX);
 963                 if (ctx->rx_conf == TLS_HW)
 964                         netdev->tlsdev_ops->tls_dev_del(netdev, ctx,
 965                                                         TLS_OFFLOAD_CTX_DIR_RX);
 966                 ctx->netdev = NULL;
 967                 dev_put(netdev);
 968                 list_del_init(&ctx->list);
 969
 970                 if (refcount_dec_and_test(&ctx->refcount))
 971                         tls_device_free_ctx(ctx);
 972         }
 973
 974         up_write(&device_offload_lock);
 975
 976         flush_work(&tls_device_gc_work);
 977
 978         return NOTIFY_DONE;
 979 }
 980
 981 static int tls_dev_event(struct notifier_block *this, unsigned long event,
 982                          void *ptr)
 983 {
 984         struct net_device *dev = netdev_notifier_info_to_dev(ptr);
 985
 986         if (!(dev->features & (NETIF_F_HW_TLS_RX | NETIF_F_HW_TLS_TX)))
 987                 return NOTIFY_DONE;
 988
 989         switch (event) {
 990         case NETDEV_REGISTER:
 991         case NETDEV_FEAT_CHANGE:
 992                 if ((dev->features & NETIF_F_HW_TLS_RX) &&
 993                     !dev->tlsdev_ops->tls_dev_resync_rx)
 994                         return NOTIFY_BAD;
 995
 996                 if  (dev->tlsdev_ops &&
 997                      dev->tlsdev_ops->tls_dev_add &&
 998                      dev->tlsdev_ops->tls_dev_del)
 999                         return NOTIFY_DONE;
1000                 else
1001                         return NOTIFY_BAD;
1002         case NETDEV_DOWN:
1003                 return tls_device_down(dev);
1004         }
1005         return NOTIFY_DONE;
1006 }
1007
1008 static struct notifier_block tls_dev_notifier = {
1009         .notifier_call  = tls_dev_event,
1010 };
1011
1012 void __init tls_device_init(void)
1013 {
1014         register_netdevice_notifier(&tls_dev_notifier);
1015 }
1016
1017 void __exit tls_device_cleanup(void)
1018 {
1019         unregister_netdevice_notifier(&tls_dev_notifier);
1020         flush_work(&tls_device_gc_work);
1021 }