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 #include "trace.h"
42
43 /* device_offload_lock is used to synchronize tls_dev_add
44 * against NETDEV_DOWN notifications.
45 */
46 static DECLARE_RWSEM(device_offload_lock);
47
48 static struct workqueue_struct *destruct_wq __read_mostly;
49
50 static LIST_HEAD(tls_device_list);
51 static LIST_HEAD(tls_device_down_list);
52 static DEFINE_SPINLOCK(tls_device_lock);
53
54 static struct page *dummy_page;
55
tls_device_free_ctx(struct tls_context * ctx)56 static void tls_device_free_ctx(struct tls_context *ctx)
57 {
58 if (ctx->tx_conf == TLS_HW) {
59 kfree(tls_offload_ctx_tx(ctx));
60 kfree(ctx->tx.rec_seq);
61 kfree(ctx->tx.iv);
62 }
63
64 if (ctx->rx_conf == TLS_HW)
65 kfree(tls_offload_ctx_rx(ctx));
66
67 tls_ctx_free(NULL, ctx);
68 }
69
tls_device_tx_del_task(struct work_struct * work)70 static void tls_device_tx_del_task(struct work_struct *work)
71 {
72 struct tls_offload_context_tx *offload_ctx =
73 container_of(work, struct tls_offload_context_tx, destruct_work);
74 struct tls_context *ctx = offload_ctx->ctx;
75 struct net_device *netdev = ctx->netdev;
76
77 netdev->tlsdev_ops->tls_dev_del(netdev, ctx, TLS_OFFLOAD_CTX_DIR_TX);
78 dev_put(netdev);
79 ctx->netdev = NULL;
80 tls_device_free_ctx(ctx);
81 }
82
tls_device_queue_ctx_destruction(struct tls_context * ctx)83 static void tls_device_queue_ctx_destruction(struct tls_context *ctx)
84 {
85 unsigned long flags;
86 bool async_cleanup;
87
88 spin_lock_irqsave(&tls_device_lock, flags);
89 if (unlikely(!refcount_dec_and_test(&ctx->refcount))) {
90 spin_unlock_irqrestore(&tls_device_lock, flags);
91 return;
92 }
93
94 list_del(&ctx->list); /* Remove from tls_device_list / tls_device_down_list */
95 async_cleanup = ctx->netdev && ctx->tx_conf == TLS_HW;
96 if (async_cleanup) {
97 struct tls_offload_context_tx *offload_ctx = tls_offload_ctx_tx(ctx);
98
99 /* queue_work inside the spinlock
100 * to make sure tls_device_down waits for that work.
101 */
102 queue_work(destruct_wq, &offload_ctx->destruct_work);
103 }
104 spin_unlock_irqrestore(&tls_device_lock, flags);
105
106 if (!async_cleanup)
107 tls_device_free_ctx(ctx);
108 }
109
110 /* We assume that the socket is already connected */
get_netdev_for_sock(struct sock * sk)111 static struct net_device *get_netdev_for_sock(struct sock *sk)
112 {
113 struct dst_entry *dst = sk_dst_get(sk);
114 struct net_device *netdev = NULL;
115
116 if (likely(dst)) {
117 netdev = netdev_sk_get_lowest_dev(dst->dev, sk);
118 dev_hold(netdev);
119 }
120
121 dst_release(dst);
122
123 return netdev;
124 }
125
destroy_record(struct tls_record_info * record)126 static void destroy_record(struct tls_record_info *record)
127 {
128 int i;
129
130 for (i = 0; i < record->num_frags; i++)
131 __skb_frag_unref(&record->frags[i], false);
132 kfree(record);
133 }
134
delete_all_records(struct tls_offload_context_tx * offload_ctx)135 static void delete_all_records(struct tls_offload_context_tx *offload_ctx)
136 {
137 struct tls_record_info *info, *temp;
138
139 list_for_each_entry_safe(info, temp, &offload_ctx->records_list, list) {
140 list_del(&info->list);
141 destroy_record(info);
142 }
143
144 offload_ctx->retransmit_hint = NULL;
145 }
146
tls_icsk_clean_acked(struct sock * sk,u32 acked_seq)147 static void tls_icsk_clean_acked(struct sock *sk, u32 acked_seq)
148 {
149 struct tls_context *tls_ctx = tls_get_ctx(sk);
150 struct tls_record_info *info, *temp;
151 struct tls_offload_context_tx *ctx;
152 u64 deleted_records = 0;
153 unsigned long flags;
154
155 if (!tls_ctx)
156 return;
157
158 ctx = tls_offload_ctx_tx(tls_ctx);
159
160 spin_lock_irqsave(&ctx->lock, flags);
161 info = ctx->retransmit_hint;
162 if (info && !before(acked_seq, info->end_seq))
163 ctx->retransmit_hint = NULL;
164
165 list_for_each_entry_safe(info, temp, &ctx->records_list, list) {
166 if (before(acked_seq, info->end_seq))
167 break;
168 list_del(&info->list);
169
170 destroy_record(info);
171 deleted_records++;
172 }
173
174 ctx->unacked_record_sn += deleted_records;
175 spin_unlock_irqrestore(&ctx->lock, flags);
176 }
177
178 /* At this point, there should be no references on this
179 * socket and no in-flight SKBs associated with this
180 * socket, so it is safe to free all the resources.
181 */
tls_device_sk_destruct(struct sock * sk)182 void tls_device_sk_destruct(struct sock *sk)
183 {
184 struct tls_context *tls_ctx = tls_get_ctx(sk);
185 struct tls_offload_context_tx *ctx = tls_offload_ctx_tx(tls_ctx);
186
187 tls_ctx->sk_destruct(sk);
188
189 if (tls_ctx->tx_conf == TLS_HW) {
190 if (ctx->open_record)
191 destroy_record(ctx->open_record);
192 delete_all_records(ctx);
193 crypto_free_aead(ctx->aead_send);
194 clean_acked_data_disable(inet_csk(sk));
195 }
196
197 tls_device_queue_ctx_destruction(tls_ctx);
198 }
199 EXPORT_SYMBOL_GPL(tls_device_sk_destruct);
200
tls_device_free_resources_tx(struct sock * sk)201 void tls_device_free_resources_tx(struct sock *sk)
202 {
203 struct tls_context *tls_ctx = tls_get_ctx(sk);
204
205 tls_free_partial_record(sk, tls_ctx);
206 }
207
tls_offload_tx_resync_request(struct sock * sk,u32 got_seq,u32 exp_seq)208 void tls_offload_tx_resync_request(struct sock *sk, u32 got_seq, u32 exp_seq)
209 {
210 struct tls_context *tls_ctx = tls_get_ctx(sk);
211
212 trace_tls_device_tx_resync_req(sk, got_seq, exp_seq);
213 WARN_ON(test_and_set_bit(TLS_TX_SYNC_SCHED, &tls_ctx->flags));
214 }
215 EXPORT_SYMBOL_GPL(tls_offload_tx_resync_request);
216
tls_device_resync_tx(struct sock * sk,struct tls_context * tls_ctx,u32 seq)217 static void tls_device_resync_tx(struct sock *sk, struct tls_context *tls_ctx,
218 u32 seq)
219 {
220 struct net_device *netdev;
221 struct sk_buff *skb;
222 int err = 0;
223 u8 *rcd_sn;
224
225 skb = tcp_write_queue_tail(sk);
226 if (skb)
227 TCP_SKB_CB(skb)->eor = 1;
228
229 rcd_sn = tls_ctx->tx.rec_seq;
230
231 trace_tls_device_tx_resync_send(sk, seq, rcd_sn);
232 down_read(&device_offload_lock);
233 netdev = tls_ctx->netdev;
234 if (netdev)
235 err = netdev->tlsdev_ops->tls_dev_resync(netdev, sk, seq,
236 rcd_sn,
237 TLS_OFFLOAD_CTX_DIR_TX);
238 up_read(&device_offload_lock);
239 if (err)
240 return;
241
242 clear_bit_unlock(TLS_TX_SYNC_SCHED, &tls_ctx->flags);
243 }
244
tls_append_frag(struct tls_record_info * record,struct page_frag * pfrag,int size)245 static void tls_append_frag(struct tls_record_info *record,
246 struct page_frag *pfrag,
247 int size)
248 {
249 skb_frag_t *frag;
250
251 frag = &record->frags[record->num_frags - 1];
252 if (skb_frag_page(frag) == pfrag->page &&
253 skb_frag_off(frag) + skb_frag_size(frag) == pfrag->offset) {
254 skb_frag_size_add(frag, size);
255 } else {
256 ++frag;
257 __skb_frag_set_page(frag, pfrag->page);
258 skb_frag_off_set(frag, pfrag->offset);
259 skb_frag_size_set(frag, size);
260 ++record->num_frags;
261 get_page(pfrag->page);
262 }
263
264 pfrag->offset += size;
265 record->len += size;
266 }
267
tls_push_record(struct sock * sk,struct tls_context * ctx,struct tls_offload_context_tx * offload_ctx,struct tls_record_info * record,int flags)268 static int tls_push_record(struct sock *sk,
269 struct tls_context *ctx,
270 struct tls_offload_context_tx *offload_ctx,
271 struct tls_record_info *record,
272 int flags)
273 {
274 struct tls_prot_info *prot = &ctx->prot_info;
275 struct tcp_sock *tp = tcp_sk(sk);
276 skb_frag_t *frag;
277 int i;
278
279 record->end_seq = tp->write_seq + record->len;
280 list_add_tail_rcu(&record->list, &offload_ctx->records_list);
281 offload_ctx->open_record = NULL;
282
283 if (test_bit(TLS_TX_SYNC_SCHED, &ctx->flags))
284 tls_device_resync_tx(sk, ctx, tp->write_seq);
285
286 tls_advance_record_sn(sk, prot, &ctx->tx);
287
288 for (i = 0; i < record->num_frags; i++) {
289 frag = &record->frags[i];
290 sg_unmark_end(&offload_ctx->sg_tx_data[i]);
291 sg_set_page(&offload_ctx->sg_tx_data[i], skb_frag_page(frag),
292 skb_frag_size(frag), skb_frag_off(frag));
293 sk_mem_charge(sk, skb_frag_size(frag));
294 get_page(skb_frag_page(frag));
295 }
296 sg_mark_end(&offload_ctx->sg_tx_data[record->num_frags - 1]);
297
298 /* all ready, send */
299 return tls_push_sg(sk, ctx, offload_ctx->sg_tx_data, 0, flags);
300 }
301
tls_device_record_close(struct sock * sk,struct tls_context * ctx,struct tls_record_info * record,struct page_frag * pfrag,unsigned char record_type)302 static void tls_device_record_close(struct sock *sk,
303 struct tls_context *ctx,
304 struct tls_record_info *record,
305 struct page_frag *pfrag,
306 unsigned char record_type)
307 {
308 struct tls_prot_info *prot = &ctx->prot_info;
309 struct page_frag dummy_tag_frag;
310
311 /* append tag
312 * device will fill in the tag, we just need to append a placeholder
313 * use socket memory to improve coalescing (re-using a single buffer
314 * increases frag count)
315 * if we can't allocate memory now use the dummy page
316 */
317 if (unlikely(pfrag->size - pfrag->offset < prot->tag_size) &&
318 !skb_page_frag_refill(prot->tag_size, pfrag, sk->sk_allocation)) {
319 dummy_tag_frag.page = dummy_page;
320 dummy_tag_frag.offset = 0;
321 pfrag = &dummy_tag_frag;
322 }
323 tls_append_frag(record, pfrag, prot->tag_size);
324
325 /* fill prepend */
326 tls_fill_prepend(ctx, skb_frag_address(&record->frags[0]),
327 record->len - prot->overhead_size,
328 record_type);
329 }
330
tls_create_new_record(struct tls_offload_context_tx * offload_ctx,struct page_frag * pfrag,size_t prepend_size)331 static int tls_create_new_record(struct tls_offload_context_tx *offload_ctx,
332 struct page_frag *pfrag,
333 size_t prepend_size)
334 {
335 struct tls_record_info *record;
336 skb_frag_t *frag;
337
338 record = kmalloc(sizeof(*record), GFP_KERNEL);
339 if (!record)
340 return -ENOMEM;
341
342 frag = &record->frags[0];
343 __skb_frag_set_page(frag, pfrag->page);
344 skb_frag_off_set(frag, pfrag->offset);
345 skb_frag_size_set(frag, prepend_size);
346
347 get_page(pfrag->page);
348 pfrag->offset += prepend_size;
349
350 record->num_frags = 1;
351 record->len = prepend_size;
352 offload_ctx->open_record = record;
353 return 0;
354 }
355
tls_do_allocation(struct sock * sk,struct tls_offload_context_tx * offload_ctx,struct page_frag * pfrag,size_t prepend_size)356 static int tls_do_allocation(struct sock *sk,
357 struct tls_offload_context_tx *offload_ctx,
358 struct page_frag *pfrag,
359 size_t prepend_size)
360 {
361 int ret;
362
363 if (!offload_ctx->open_record) {
364 if (unlikely(!skb_page_frag_refill(prepend_size, pfrag,
365 sk->sk_allocation))) {
366 READ_ONCE(sk->sk_prot)->enter_memory_pressure(sk);
367 sk_stream_moderate_sndbuf(sk);
368 return -ENOMEM;
369 }
370
371 ret = tls_create_new_record(offload_ctx, pfrag, prepend_size);
372 if (ret)
373 return ret;
374
375 if (pfrag->size > pfrag->offset)
376 return 0;
377 }
378
379 if (!sk_page_frag_refill(sk, pfrag))
380 return -ENOMEM;
381
382 return 0;
383 }
384
tls_device_copy_data(void * addr,size_t bytes,struct iov_iter * i)385 static int tls_device_copy_data(void *addr, size_t bytes, struct iov_iter *i)
386 {
387 size_t pre_copy, nocache;
388
389 pre_copy = ~((unsigned long)addr - 1) & (SMP_CACHE_BYTES - 1);
390 if (pre_copy) {
391 pre_copy = min(pre_copy, bytes);
392 if (copy_from_iter(addr, pre_copy, i) != pre_copy)
393 return -EFAULT;
394 bytes -= pre_copy;
395 addr += pre_copy;
396 }
397
398 nocache = round_down(bytes, SMP_CACHE_BYTES);
399 if (copy_from_iter_nocache(addr, nocache, i) != nocache)
400 return -EFAULT;
401 bytes -= nocache;
402 addr += nocache;
403
404 if (bytes && copy_from_iter(addr, bytes, i) != bytes)
405 return -EFAULT;
406
407 return 0;
408 }
409
tls_push_data(struct sock * sk,struct iov_iter * msg_iter,size_t size,int flags,unsigned char record_type)410 static int tls_push_data(struct sock *sk,
411 struct iov_iter *msg_iter,
412 size_t size, int flags,
413 unsigned char record_type)
414 {
415 struct tls_context *tls_ctx = tls_get_ctx(sk);
416 struct tls_prot_info *prot = &tls_ctx->prot_info;
417 struct tls_offload_context_tx *ctx = tls_offload_ctx_tx(tls_ctx);
418 struct tls_record_info *record;
419 int tls_push_record_flags;
420 struct page_frag *pfrag;
421 size_t orig_size = size;
422 u32 max_open_record_len;
423 bool more = false;
424 bool done = false;
425 int copy, rc = 0;
426 long timeo;
427
428 if (flags &
429 ~(MSG_MORE | MSG_DONTWAIT | MSG_NOSIGNAL | MSG_SENDPAGE_NOTLAST))
430 return -EOPNOTSUPP;
431
432 if (unlikely(sk->sk_err))
433 return -sk->sk_err;
434
435 flags |= MSG_SENDPAGE_DECRYPTED;
436 tls_push_record_flags = flags | MSG_SENDPAGE_NOTLAST;
437
438 timeo = sock_sndtimeo(sk, flags & MSG_DONTWAIT);
439 if (tls_is_partially_sent_record(tls_ctx)) {
440 rc = tls_push_partial_record(sk, tls_ctx, flags);
441 if (rc < 0)
442 return rc;
443 }
444
445 pfrag = sk_page_frag(sk);
446
447 /* TLS_HEADER_SIZE is not counted as part of the TLS record, and
448 * we need to leave room for an authentication tag.
449 */
450 max_open_record_len = TLS_MAX_PAYLOAD_SIZE +
451 prot->prepend_size;
452 do {
453 rc = tls_do_allocation(sk, ctx, pfrag, prot->prepend_size);
454 if (unlikely(rc)) {
455 rc = sk_stream_wait_memory(sk, &timeo);
456 if (!rc)
457 continue;
458
459 record = ctx->open_record;
460 if (!record)
461 break;
462 handle_error:
463 if (record_type != TLS_RECORD_TYPE_DATA) {
464 /* avoid sending partial
465 * record with type !=
466 * application_data
467 */
468 size = orig_size;
469 destroy_record(record);
470 ctx->open_record = NULL;
471 } else if (record->len > prot->prepend_size) {
472 goto last_record;
473 }
474
475 break;
476 }
477
478 record = ctx->open_record;
479 copy = min_t(size_t, size, (pfrag->size - pfrag->offset));
480 copy = min_t(size_t, copy, (max_open_record_len - record->len));
481
482 if (copy) {
483 rc = tls_device_copy_data(page_address(pfrag->page) +
484 pfrag->offset, copy, msg_iter);
485 if (rc)
486 goto handle_error;
487 tls_append_frag(record, pfrag, copy);
488 }
489
490 size -= copy;
491 if (!size) {
492 last_record:
493 tls_push_record_flags = flags;
494 if (flags & (MSG_SENDPAGE_NOTLAST | MSG_MORE)) {
495 more = true;
496 break;
497 }
498
499 done = true;
500 }
501
502 if (done || record->len >= max_open_record_len ||
503 (record->num_frags >= MAX_SKB_FRAGS - 1)) {
504 tls_device_record_close(sk, tls_ctx, record,
505 pfrag, record_type);
506
507 rc = tls_push_record(sk,
508 tls_ctx,
509 ctx,
510 record,
511 tls_push_record_flags);
512 if (rc < 0)
513 break;
514 }
515 } while (!done);
516
517 tls_ctx->pending_open_record_frags = more;
518
519 if (orig_size - size > 0)
520 rc = orig_size - size;
521
522 return rc;
523 }
524
tls_device_sendmsg(struct sock * sk,struct msghdr * msg,size_t size)525 int tls_device_sendmsg(struct sock *sk, struct msghdr *msg, size_t size)
526 {
527 unsigned char record_type = TLS_RECORD_TYPE_DATA;
528 struct tls_context *tls_ctx = tls_get_ctx(sk);
529 int rc;
530
531 mutex_lock(&tls_ctx->tx_lock);
532 lock_sock(sk);
533
534 if (unlikely(msg->msg_controllen)) {
535 rc = tls_proccess_cmsg(sk, msg, &record_type);
536 if (rc)
537 goto out;
538 }
539
540 rc = tls_push_data(sk, &msg->msg_iter, size,
541 msg->msg_flags, record_type);
542
543 out:
544 release_sock(sk);
545 mutex_unlock(&tls_ctx->tx_lock);
546 return rc;
547 }
548
tls_device_sendpage(struct sock * sk,struct page * page,int offset,size_t size,int flags)549 int tls_device_sendpage(struct sock *sk, struct page *page,
550 int offset, size_t size, int flags)
551 {
552 struct tls_context *tls_ctx = tls_get_ctx(sk);
553 struct iov_iter msg_iter;
554 char *kaddr;
555 struct kvec iov;
556 int rc;
557
558 if (flags & MSG_SENDPAGE_NOTLAST)
559 flags |= MSG_MORE;
560
561 mutex_lock(&tls_ctx->tx_lock);
562 lock_sock(sk);
563
564 if (flags & MSG_OOB) {
565 rc = -EOPNOTSUPP;
566 goto out;
567 }
568
569 kaddr = kmap(page);
570 iov.iov_base = kaddr + offset;
571 iov.iov_len = size;
572 iov_iter_kvec(&msg_iter, WRITE, &iov, 1, size);
573 rc = tls_push_data(sk, &msg_iter, size,
574 flags, TLS_RECORD_TYPE_DATA);
575 kunmap(page);
576
577 out:
578 release_sock(sk);
579 mutex_unlock(&tls_ctx->tx_lock);
580 return rc;
581 }
582
tls_get_record(struct tls_offload_context_tx * context,u32 seq,u64 * p_record_sn)583 struct tls_record_info *tls_get_record(struct tls_offload_context_tx *context,
584 u32 seq, u64 *p_record_sn)
585 {
586 u64 record_sn = context->hint_record_sn;
587 struct tls_record_info *info, *last;
588
589 info = context->retransmit_hint;
590 if (!info ||
591 before(seq, info->end_seq - info->len)) {
592 /* if retransmit_hint is irrelevant start
593 * from the beginning of the list
594 */
595 info = list_first_entry_or_null(&context->records_list,
596 struct tls_record_info, list);
597 if (!info)
598 return NULL;
599 /* send the start_marker record if seq number is before the
600 * tls offload start marker sequence number. This record is
601 * required to handle TCP packets which are before TLS offload
602 * started.
603 * And if it's not start marker, look if this seq number
604 * belongs to the list.
605 */
606 if (likely(!tls_record_is_start_marker(info))) {
607 /* we have the first record, get the last record to see
608 * if this seq number belongs to the list.
609 */
610 last = list_last_entry(&context->records_list,
611 struct tls_record_info, list);
612
613 if (!between(seq, tls_record_start_seq(info),
614 last->end_seq))
615 return NULL;
616 }
617 record_sn = context->unacked_record_sn;
618 }
619
620 /* We just need the _rcu for the READ_ONCE() */
621 rcu_read_lock();
622 list_for_each_entry_from_rcu(info, &context->records_list, list) {
623 if (before(seq, info->end_seq)) {
624 if (!context->retransmit_hint ||
625 after(info->end_seq,
626 context->retransmit_hint->end_seq)) {
627 context->hint_record_sn = record_sn;
628 context->retransmit_hint = info;
629 }
630 *p_record_sn = record_sn;
631 goto exit_rcu_unlock;
632 }
633 record_sn++;
634 }
635 info = NULL;
636
637 exit_rcu_unlock:
638 rcu_read_unlock();
639 return info;
640 }
641 EXPORT_SYMBOL(tls_get_record);
642
tls_device_push_pending_record(struct sock * sk,int flags)643 static int tls_device_push_pending_record(struct sock *sk, int flags)
644 {
645 struct iov_iter msg_iter;
646
647 iov_iter_kvec(&msg_iter, WRITE, NULL, 0, 0);
648 return tls_push_data(sk, &msg_iter, 0, flags, TLS_RECORD_TYPE_DATA);
649 }
650
tls_device_write_space(struct sock * sk,struct tls_context * ctx)651 void tls_device_write_space(struct sock *sk, struct tls_context *ctx)
652 {
653 if (tls_is_partially_sent_record(ctx)) {
654 gfp_t sk_allocation = sk->sk_allocation;
655
656 WARN_ON_ONCE(sk->sk_write_pending);
657
658 sk->sk_allocation = GFP_ATOMIC;
659 tls_push_partial_record(sk, ctx,
660 MSG_DONTWAIT | MSG_NOSIGNAL |
661 MSG_SENDPAGE_DECRYPTED);
662 sk->sk_allocation = sk_allocation;
663 }
664 }
665
tls_device_resync_rx(struct tls_context * tls_ctx,struct sock * sk,u32 seq,u8 * rcd_sn)666 static void tls_device_resync_rx(struct tls_context *tls_ctx,
667 struct sock *sk, u32 seq, u8 *rcd_sn)
668 {
669 struct tls_offload_context_rx *rx_ctx = tls_offload_ctx_rx(tls_ctx);
670 struct net_device *netdev;
671
672 trace_tls_device_rx_resync_send(sk, seq, rcd_sn, rx_ctx->resync_type);
673 rcu_read_lock();
674 netdev = READ_ONCE(tls_ctx->netdev);
675 if (netdev)
676 netdev->tlsdev_ops->tls_dev_resync(netdev, sk, seq, rcd_sn,
677 TLS_OFFLOAD_CTX_DIR_RX);
678 rcu_read_unlock();
679 TLS_INC_STATS(sock_net(sk), LINUX_MIB_TLSRXDEVICERESYNC);
680 }
681
682 static bool
tls_device_rx_resync_async(struct tls_offload_resync_async * resync_async,s64 resync_req,u32 * seq,u16 * rcd_delta)683 tls_device_rx_resync_async(struct tls_offload_resync_async *resync_async,
684 s64 resync_req, u32 *seq, u16 *rcd_delta)
685 {
686 u32 is_async = resync_req & RESYNC_REQ_ASYNC;
687 u32 req_seq = resync_req >> 32;
688 u32 req_end = req_seq + ((resync_req >> 16) & 0xffff);
689 u16 i;
690
691 *rcd_delta = 0;
692
693 if (is_async) {
694 /* shouldn't get to wraparound:
695 * too long in async stage, something bad happened
696 */
697 if (WARN_ON_ONCE(resync_async->rcd_delta == USHRT_MAX))
698 return false;
699
700 /* asynchronous stage: log all headers seq such that
701 * req_seq <= seq <= end_seq, and wait for real resync request
702 */
703 if (before(*seq, req_seq))
704 return false;
705 if (!after(*seq, req_end) &&
706 resync_async->loglen < TLS_DEVICE_RESYNC_ASYNC_LOGMAX)
707 resync_async->log[resync_async->loglen++] = *seq;
708
709 resync_async->rcd_delta++;
710
711 return false;
712 }
713
714 /* synchronous stage: check against the logged entries and
715 * proceed to check the next entries if no match was found
716 */
717 for (i = 0; i < resync_async->loglen; i++)
718 if (req_seq == resync_async->log[i] &&
719 atomic64_try_cmpxchg(&resync_async->req, &resync_req, 0)) {
720 *rcd_delta = resync_async->rcd_delta - i;
721 *seq = req_seq;
722 resync_async->loglen = 0;
723 resync_async->rcd_delta = 0;
724 return true;
725 }
726
727 resync_async->loglen = 0;
728 resync_async->rcd_delta = 0;
729
730 if (req_seq == *seq &&
731 atomic64_try_cmpxchg(&resync_async->req,
732 &resync_req, 0))
733 return true;
734
735 return false;
736 }
737
tls_device_rx_resync_new_rec(struct sock * sk,u32 rcd_len,u32 seq)738 void tls_device_rx_resync_new_rec(struct sock *sk, u32 rcd_len, u32 seq)
739 {
740 struct tls_context *tls_ctx = tls_get_ctx(sk);
741 struct tls_offload_context_rx *rx_ctx;
742 u8 rcd_sn[TLS_MAX_REC_SEQ_SIZE];
743 u32 sock_data, is_req_pending;
744 struct tls_prot_info *prot;
745 s64 resync_req;
746 u16 rcd_delta;
747 u32 req_seq;
748
749 if (tls_ctx->rx_conf != TLS_HW)
750 return;
751 if (unlikely(test_bit(TLS_RX_DEV_DEGRADED, &tls_ctx->flags)))
752 return;
753
754 prot = &tls_ctx->prot_info;
755 rx_ctx = tls_offload_ctx_rx(tls_ctx);
756 memcpy(rcd_sn, tls_ctx->rx.rec_seq, prot->rec_seq_size);
757
758 switch (rx_ctx->resync_type) {
759 case TLS_OFFLOAD_SYNC_TYPE_DRIVER_REQ:
760 resync_req = atomic64_read(&rx_ctx->resync_req);
761 req_seq = resync_req >> 32;
762 seq += TLS_HEADER_SIZE - 1;
763 is_req_pending = resync_req;
764
765 if (likely(!is_req_pending) || req_seq != seq ||
766 !atomic64_try_cmpxchg(&rx_ctx->resync_req, &resync_req, 0))
767 return;
768 break;
769 case TLS_OFFLOAD_SYNC_TYPE_CORE_NEXT_HINT:
770 if (likely(!rx_ctx->resync_nh_do_now))
771 return;
772
773 /* head of next rec is already in, note that the sock_inq will
774 * include the currently parsed message when called from parser
775 */
776 sock_data = tcp_inq(sk);
777 if (sock_data > rcd_len) {
778 trace_tls_device_rx_resync_nh_delay(sk, sock_data,
779 rcd_len);
780 return;
781 }
782
783 rx_ctx->resync_nh_do_now = 0;
784 seq += rcd_len;
785 tls_bigint_increment(rcd_sn, prot->rec_seq_size);
786 break;
787 case TLS_OFFLOAD_SYNC_TYPE_DRIVER_REQ_ASYNC:
788 resync_req = atomic64_read(&rx_ctx->resync_async->req);
789 is_req_pending = resync_req;
790 if (likely(!is_req_pending))
791 return;
792
793 if (!tls_device_rx_resync_async(rx_ctx->resync_async,
794 resync_req, &seq, &rcd_delta))
795 return;
796 tls_bigint_subtract(rcd_sn, rcd_delta);
797 break;
798 }
799
800 tls_device_resync_rx(tls_ctx, sk, seq, rcd_sn);
801 }
802
tls_device_core_ctrl_rx_resync(struct tls_context * tls_ctx,struct tls_offload_context_rx * ctx,struct sock * sk,struct sk_buff * skb)803 static void tls_device_core_ctrl_rx_resync(struct tls_context *tls_ctx,
804 struct tls_offload_context_rx *ctx,
805 struct sock *sk, struct sk_buff *skb)
806 {
807 struct strp_msg *rxm;
808
809 /* device will request resyncs by itself based on stream scan */
810 if (ctx->resync_type != TLS_OFFLOAD_SYNC_TYPE_CORE_NEXT_HINT)
811 return;
812 /* already scheduled */
813 if (ctx->resync_nh_do_now)
814 return;
815 /* seen decrypted fragments since last fully-failed record */
816 if (ctx->resync_nh_reset) {
817 ctx->resync_nh_reset = 0;
818 ctx->resync_nh.decrypted_failed = 1;
819 ctx->resync_nh.decrypted_tgt = TLS_DEVICE_RESYNC_NH_START_IVAL;
820 return;
821 }
822
823 if (++ctx->resync_nh.decrypted_failed <= ctx->resync_nh.decrypted_tgt)
824 return;
825
826 /* doing resync, bump the next target in case it fails */
827 if (ctx->resync_nh.decrypted_tgt < TLS_DEVICE_RESYNC_NH_MAX_IVAL)
828 ctx->resync_nh.decrypted_tgt *= 2;
829 else
830 ctx->resync_nh.decrypted_tgt += TLS_DEVICE_RESYNC_NH_MAX_IVAL;
831
832 rxm = strp_msg(skb);
833
834 /* head of next rec is already in, parser will sync for us */
835 if (tcp_inq(sk) > rxm->full_len) {
836 trace_tls_device_rx_resync_nh_schedule(sk);
837 ctx->resync_nh_do_now = 1;
838 } else {
839 struct tls_prot_info *prot = &tls_ctx->prot_info;
840 u8 rcd_sn[TLS_MAX_REC_SEQ_SIZE];
841
842 memcpy(rcd_sn, tls_ctx->rx.rec_seq, prot->rec_seq_size);
843 tls_bigint_increment(rcd_sn, prot->rec_seq_size);
844
845 tls_device_resync_rx(tls_ctx, sk, tcp_sk(sk)->copied_seq,
846 rcd_sn);
847 }
848 }
849
tls_device_reencrypt(struct sock * sk,struct sk_buff * skb)850 static int tls_device_reencrypt(struct sock *sk, struct sk_buff *skb)
851 {
852 struct strp_msg *rxm = strp_msg(skb);
853 int err = 0, offset = rxm->offset, copy, nsg, data_len, pos;
854 struct sk_buff *skb_iter, *unused;
855 struct scatterlist sg[1];
856 char *orig_buf, *buf;
857
858 orig_buf = kmalloc(rxm->full_len + TLS_HEADER_SIZE +
859 TLS_CIPHER_AES_GCM_128_IV_SIZE, sk->sk_allocation);
860 if (!orig_buf)
861 return -ENOMEM;
862 buf = orig_buf;
863
864 nsg = skb_cow_data(skb, 0, &unused);
865 if (unlikely(nsg < 0)) {
866 err = nsg;
867 goto free_buf;
868 }
869
870 sg_init_table(sg, 1);
871 sg_set_buf(&sg[0], buf,
872 rxm->full_len + TLS_HEADER_SIZE +
873 TLS_CIPHER_AES_GCM_128_IV_SIZE);
874 err = skb_copy_bits(skb, offset, buf,
875 TLS_HEADER_SIZE + TLS_CIPHER_AES_GCM_128_IV_SIZE);
876 if (err)
877 goto free_buf;
878
879 /* We are interested only in the decrypted data not the auth */
880 err = decrypt_skb(sk, skb, sg);
881 if (err != -EBADMSG)
882 goto free_buf;
883 else
884 err = 0;
885
886 data_len = rxm->full_len - TLS_CIPHER_AES_GCM_128_TAG_SIZE;
887
888 if (skb_pagelen(skb) > offset) {
889 copy = min_t(int, skb_pagelen(skb) - offset, data_len);
890
891 if (skb->decrypted) {
892 err = skb_store_bits(skb, offset, buf, copy);
893 if (err)
894 goto free_buf;
895 }
896
897 offset += copy;
898 buf += copy;
899 }
900
901 pos = skb_pagelen(skb);
902 skb_walk_frags(skb, skb_iter) {
903 int frag_pos;
904
905 /* Practically all frags must belong to msg if reencrypt
906 * is needed with current strparser and coalescing logic,
907 * but strparser may "get optimized", so let's be safe.
908 */
909 if (pos + skb_iter->len <= offset)
910 goto done_with_frag;
911 if (pos >= data_len + rxm->offset)
912 break;
913
914 frag_pos = offset - pos;
915 copy = min_t(int, skb_iter->len - frag_pos,
916 data_len + rxm->offset - offset);
917
918 if (skb_iter->decrypted) {
919 err = skb_store_bits(skb_iter, frag_pos, buf, copy);
920 if (err)
921 goto free_buf;
922 }
923
924 offset += copy;
925 buf += copy;
926 done_with_frag:
927 pos += skb_iter->len;
928 }
929
930 free_buf:
931 kfree(orig_buf);
932 return err;
933 }
934
tls_device_decrypted(struct sock * sk,struct tls_context * tls_ctx,struct sk_buff * skb,struct strp_msg * rxm)935 int tls_device_decrypted(struct sock *sk, struct tls_context *tls_ctx,
936 struct sk_buff *skb, struct strp_msg *rxm)
937 {
938 struct tls_offload_context_rx *ctx = tls_offload_ctx_rx(tls_ctx);
939 int is_decrypted = skb->decrypted;
940 int is_encrypted = !is_decrypted;
941 struct sk_buff *skb_iter;
942
943 /* Check if all the data is decrypted already */
944 skb_walk_frags(skb, skb_iter) {
945 is_decrypted &= skb_iter->decrypted;
946 is_encrypted &= !skb_iter->decrypted;
947 }
948
949 trace_tls_device_decrypted(sk, tcp_sk(sk)->copied_seq - rxm->full_len,
950 tls_ctx->rx.rec_seq, rxm->full_len,
951 is_encrypted, is_decrypted);
952
953 if (unlikely(test_bit(TLS_RX_DEV_DEGRADED, &tls_ctx->flags))) {
954 if (likely(is_encrypted || is_decrypted))
955 return is_decrypted;
956
957 /* After tls_device_down disables the offload, the next SKB will
958 * likely have initial fragments decrypted, and final ones not
959 * decrypted. We need to reencrypt that single SKB.
960 */
961 return tls_device_reencrypt(sk, skb);
962 }
963
964 /* Return immediately if the record is either entirely plaintext or
965 * entirely ciphertext. Otherwise handle reencrypt partially decrypted
966 * record.
967 */
968 if (is_decrypted) {
969 ctx->resync_nh_reset = 1;
970 return is_decrypted;
971 }
972 if (is_encrypted) {
973 tls_device_core_ctrl_rx_resync(tls_ctx, ctx, sk, skb);
974 return 0;
975 }
976
977 ctx->resync_nh_reset = 1;
978 return tls_device_reencrypt(sk, skb);
979 }
980
tls_device_attach(struct tls_context * ctx,struct sock * sk,struct net_device * netdev)981 static void tls_device_attach(struct tls_context *ctx, struct sock *sk,
982 struct net_device *netdev)
983 {
984 if (sk->sk_destruct != tls_device_sk_destruct) {
985 refcount_set(&ctx->refcount, 1);
986 dev_hold(netdev);
987 ctx->netdev = netdev;
988 spin_lock_irq(&tls_device_lock);
989 list_add_tail(&ctx->list, &tls_device_list);
990 spin_unlock_irq(&tls_device_lock);
991
992 ctx->sk_destruct = sk->sk_destruct;
993 smp_store_release(&sk->sk_destruct, tls_device_sk_destruct);
994 }
995 }
996
tls_set_device_offload(struct sock * sk,struct tls_context * ctx)997 int tls_set_device_offload(struct sock *sk, struct tls_context *ctx)
998 {
999 u16 nonce_size, tag_size, iv_size, rec_seq_size, salt_size;
1000 struct tls_context *tls_ctx = tls_get_ctx(sk);
1001 struct tls_prot_info *prot = &tls_ctx->prot_info;
1002 struct tls_record_info *start_marker_record;
1003 struct tls_offload_context_tx *offload_ctx;
1004 struct tls_crypto_info *crypto_info;
1005 struct net_device *netdev;
1006 char *iv, *rec_seq;
1007 struct sk_buff *skb;
1008 __be64 rcd_sn;
1009 int rc;
1010
1011 if (!ctx)
1012 return -EINVAL;
1013
1014 if (ctx->priv_ctx_tx)
1015 return -EEXIST;
1016
1017 start_marker_record = kmalloc(sizeof(*start_marker_record), GFP_KERNEL);
1018 if (!start_marker_record)
1019 return -ENOMEM;
1020
1021 offload_ctx = kzalloc(TLS_OFFLOAD_CONTEXT_SIZE_TX, GFP_KERNEL);
1022 if (!offload_ctx) {
1023 rc = -ENOMEM;
1024 goto free_marker_record;
1025 }
1026
1027 crypto_info = &ctx->crypto_send.info;
1028 if (crypto_info->version != TLS_1_2_VERSION) {
1029 rc = -EOPNOTSUPP;
1030 goto free_offload_ctx;
1031 }
1032
1033 switch (crypto_info->cipher_type) {
1034 case TLS_CIPHER_AES_GCM_128:
1035 nonce_size = TLS_CIPHER_AES_GCM_128_IV_SIZE;
1036 tag_size = TLS_CIPHER_AES_GCM_128_TAG_SIZE;
1037 iv_size = TLS_CIPHER_AES_GCM_128_IV_SIZE;
1038 iv = ((struct tls12_crypto_info_aes_gcm_128 *)crypto_info)->iv;
1039 rec_seq_size = TLS_CIPHER_AES_GCM_128_REC_SEQ_SIZE;
1040 salt_size = TLS_CIPHER_AES_GCM_128_SALT_SIZE;
1041 rec_seq =
1042 ((struct tls12_crypto_info_aes_gcm_128 *)crypto_info)->rec_seq;
1043 break;
1044 default:
1045 rc = -EINVAL;
1046 goto free_offload_ctx;
1047 }
1048
1049 /* Sanity-check the rec_seq_size for stack allocations */
1050 if (rec_seq_size > TLS_MAX_REC_SEQ_SIZE) {
1051 rc = -EINVAL;
1052 goto free_offload_ctx;
1053 }
1054
1055 prot->version = crypto_info->version;
1056 prot->cipher_type = crypto_info->cipher_type;
1057 prot->prepend_size = TLS_HEADER_SIZE + nonce_size;
1058 prot->tag_size = tag_size;
1059 prot->overhead_size = prot->prepend_size + prot->tag_size;
1060 prot->iv_size = iv_size;
1061 prot->salt_size = salt_size;
1062 ctx->tx.iv = kmalloc(iv_size + TLS_CIPHER_AES_GCM_128_SALT_SIZE,
1063 GFP_KERNEL);
1064 if (!ctx->tx.iv) {
1065 rc = -ENOMEM;
1066 goto free_offload_ctx;
1067 }
1068
1069 memcpy(ctx->tx.iv + TLS_CIPHER_AES_GCM_128_SALT_SIZE, iv, iv_size);
1070
1071 prot->rec_seq_size = rec_seq_size;
1072 ctx->tx.rec_seq = kmemdup(rec_seq, rec_seq_size, GFP_KERNEL);
1073 if (!ctx->tx.rec_seq) {
1074 rc = -ENOMEM;
1075 goto free_iv;
1076 }
1077
1078 rc = tls_sw_fallback_init(sk, offload_ctx, crypto_info);
1079 if (rc)
1080 goto free_rec_seq;
1081
1082 /* start at rec_seq - 1 to account for the start marker record */
1083 memcpy(&rcd_sn, ctx->tx.rec_seq, sizeof(rcd_sn));
1084 offload_ctx->unacked_record_sn = be64_to_cpu(rcd_sn) - 1;
1085
1086 start_marker_record->end_seq = tcp_sk(sk)->write_seq;
1087 start_marker_record->len = 0;
1088 start_marker_record->num_frags = 0;
1089
1090 INIT_WORK(&offload_ctx->destruct_work, tls_device_tx_del_task);
1091 offload_ctx->ctx = ctx;
1092
1093 INIT_LIST_HEAD(&offload_ctx->records_list);
1094 list_add_tail(&start_marker_record->list, &offload_ctx->records_list);
1095 spin_lock_init(&offload_ctx->lock);
1096 sg_init_table(offload_ctx->sg_tx_data,
1097 ARRAY_SIZE(offload_ctx->sg_tx_data));
1098
1099 clean_acked_data_enable(inet_csk(sk), &tls_icsk_clean_acked);
1100 ctx->push_pending_record = tls_device_push_pending_record;
1101
1102 /* TLS offload is greatly simplified if we don't send
1103 * SKBs where only part of the payload needs to be encrypted.
1104 * So mark the last skb in the write queue as end of record.
1105 */
1106 skb = tcp_write_queue_tail(sk);
1107 if (skb)
1108 TCP_SKB_CB(skb)->eor = 1;
1109
1110 netdev = get_netdev_for_sock(sk);
1111 if (!netdev) {
1112 pr_err_ratelimited("%s: netdev not found\n", __func__);
1113 rc = -EINVAL;
1114 goto disable_cad;
1115 }
1116
1117 if (!(netdev->features & NETIF_F_HW_TLS_TX)) {
1118 rc = -EOPNOTSUPP;
1119 goto release_netdev;
1120 }
1121
1122 /* Avoid offloading if the device is down
1123 * We don't want to offload new flows after
1124 * the NETDEV_DOWN event
1125 *
1126 * device_offload_lock is taken in tls_devices's NETDEV_DOWN
1127 * handler thus protecting from the device going down before
1128 * ctx was added to tls_device_list.
1129 */
1130 down_read(&device_offload_lock);
1131 if (!(netdev->flags & IFF_UP)) {
1132 rc = -EINVAL;
1133 goto release_lock;
1134 }
1135
1136 ctx->priv_ctx_tx = offload_ctx;
1137 rc = netdev->tlsdev_ops->tls_dev_add(netdev, sk, TLS_OFFLOAD_CTX_DIR_TX,
1138 &ctx->crypto_send.info,
1139 tcp_sk(sk)->write_seq);
1140 trace_tls_device_offload_set(sk, TLS_OFFLOAD_CTX_DIR_TX,
1141 tcp_sk(sk)->write_seq, rec_seq, rc);
1142 if (rc)
1143 goto release_lock;
1144
1145 tls_device_attach(ctx, sk, netdev);
1146 up_read(&device_offload_lock);
1147
1148 /* following this assignment tls_is_sk_tx_device_offloaded
1149 * will return true and the context might be accessed
1150 * by the netdev's xmit function.
1151 */
1152 smp_store_release(&sk->sk_validate_xmit_skb, tls_validate_xmit_skb);
1153 dev_put(netdev);
1154
1155 return 0;
1156
1157 release_lock:
1158 up_read(&device_offload_lock);
1159 release_netdev:
1160 dev_put(netdev);
1161 disable_cad:
1162 clean_acked_data_disable(inet_csk(sk));
1163 crypto_free_aead(offload_ctx->aead_send);
1164 free_rec_seq:
1165 kfree(ctx->tx.rec_seq);
1166 free_iv:
1167 kfree(ctx->tx.iv);
1168 free_offload_ctx:
1169 kfree(offload_ctx);
1170 ctx->priv_ctx_tx = NULL;
1171 free_marker_record:
1172 kfree(start_marker_record);
1173 return rc;
1174 }
1175
tls_set_device_offload_rx(struct sock * sk,struct tls_context * ctx)1176 int tls_set_device_offload_rx(struct sock *sk, struct tls_context *ctx)
1177 {
1178 struct tls12_crypto_info_aes_gcm_128 *info;
1179 struct tls_offload_context_rx *context;
1180 struct net_device *netdev;
1181 int rc = 0;
1182
1183 if (ctx->crypto_recv.info.version != TLS_1_2_VERSION)
1184 return -EOPNOTSUPP;
1185
1186 netdev = get_netdev_for_sock(sk);
1187 if (!netdev) {
1188 pr_err_ratelimited("%s: netdev not found\n", __func__);
1189 return -EINVAL;
1190 }
1191
1192 if (!(netdev->features & NETIF_F_HW_TLS_RX)) {
1193 rc = -EOPNOTSUPP;
1194 goto release_netdev;
1195 }
1196
1197 /* Avoid offloading if the device is down
1198 * We don't want to offload new flows after
1199 * the NETDEV_DOWN event
1200 *
1201 * device_offload_lock is taken in tls_devices's NETDEV_DOWN
1202 * handler thus protecting from the device going down before
1203 * ctx was added to tls_device_list.
1204 */
1205 down_read(&device_offload_lock);
1206 if (!(netdev->flags & IFF_UP)) {
1207 rc = -EINVAL;
1208 goto release_lock;
1209 }
1210
1211 context = kzalloc(TLS_OFFLOAD_CONTEXT_SIZE_RX, GFP_KERNEL);
1212 if (!context) {
1213 rc = -ENOMEM;
1214 goto release_lock;
1215 }
1216 context->resync_nh_reset = 1;
1217
1218 ctx->priv_ctx_rx = context;
1219 rc = tls_set_sw_offload(sk, ctx, 0);
1220 if (rc)
1221 goto release_ctx;
1222
1223 rc = netdev->tlsdev_ops->tls_dev_add(netdev, sk, TLS_OFFLOAD_CTX_DIR_RX,
1224 &ctx->crypto_recv.info,
1225 tcp_sk(sk)->copied_seq);
1226 info = (void *)&ctx->crypto_recv.info;
1227 trace_tls_device_offload_set(sk, TLS_OFFLOAD_CTX_DIR_RX,
1228 tcp_sk(sk)->copied_seq, info->rec_seq, rc);
1229 if (rc)
1230 goto free_sw_resources;
1231
1232 tls_device_attach(ctx, sk, netdev);
1233 up_read(&device_offload_lock);
1234
1235 dev_put(netdev);
1236
1237 return 0;
1238
1239 free_sw_resources:
1240 up_read(&device_offload_lock);
1241 tls_sw_free_resources_rx(sk);
1242 down_read(&device_offload_lock);
1243 release_ctx:
1244 ctx->priv_ctx_rx = NULL;
1245 release_lock:
1246 up_read(&device_offload_lock);
1247 release_netdev:
1248 dev_put(netdev);
1249 return rc;
1250 }
1251
tls_device_offload_cleanup_rx(struct sock * sk)1252 void tls_device_offload_cleanup_rx(struct sock *sk)
1253 {
1254 struct tls_context *tls_ctx = tls_get_ctx(sk);
1255 struct net_device *netdev;
1256
1257 down_read(&device_offload_lock);
1258 netdev = tls_ctx->netdev;
1259 if (!netdev)
1260 goto out;
1261
1262 netdev->tlsdev_ops->tls_dev_del(netdev, tls_ctx,
1263 TLS_OFFLOAD_CTX_DIR_RX);
1264
1265 if (tls_ctx->tx_conf != TLS_HW) {
1266 dev_put(netdev);
1267 tls_ctx->netdev = NULL;
1268 } else {
1269 set_bit(TLS_RX_DEV_CLOSED, &tls_ctx->flags);
1270 }
1271 out:
1272 up_read(&device_offload_lock);
1273 tls_sw_release_resources_rx(sk);
1274 }
1275
tls_device_down(struct net_device * netdev)1276 static int tls_device_down(struct net_device *netdev)
1277 {
1278 struct tls_context *ctx, *tmp;
1279 unsigned long flags;
1280 LIST_HEAD(list);
1281
1282 /* Request a write lock to block new offload attempts */
1283 down_write(&device_offload_lock);
1284
1285 spin_lock_irqsave(&tls_device_lock, flags);
1286 list_for_each_entry_safe(ctx, tmp, &tls_device_list, list) {
1287 if (ctx->netdev != netdev ||
1288 !refcount_inc_not_zero(&ctx->refcount))
1289 continue;
1290
1291 list_move(&ctx->list, &list);
1292 }
1293 spin_unlock_irqrestore(&tls_device_lock, flags);
1294
1295 list_for_each_entry_safe(ctx, tmp, &list, list) {
1296 /* Stop offloaded TX and switch to the fallback.
1297 * tls_is_sk_tx_device_offloaded will return false.
1298 */
1299 WRITE_ONCE(ctx->sk->sk_validate_xmit_skb, tls_validate_xmit_skb_sw);
1300
1301 /* Stop the RX and TX resync.
1302 * tls_dev_resync must not be called after tls_dev_del.
1303 */
1304 WRITE_ONCE(ctx->netdev, NULL);
1305
1306 /* Start skipping the RX resync logic completely. */
1307 set_bit(TLS_RX_DEV_DEGRADED, &ctx->flags);
1308
1309 /* Sync with inflight packets. After this point:
1310 * TX: no non-encrypted packets will be passed to the driver.
1311 * RX: resync requests from the driver will be ignored.
1312 */
1313 synchronize_net();
1314
1315 /* Release the offload context on the driver side. */
1316 if (ctx->tx_conf == TLS_HW)
1317 netdev->tlsdev_ops->tls_dev_del(netdev, ctx,
1318 TLS_OFFLOAD_CTX_DIR_TX);
1319 if (ctx->rx_conf == TLS_HW &&
1320 !test_bit(TLS_RX_DEV_CLOSED, &ctx->flags))
1321 netdev->tlsdev_ops->tls_dev_del(netdev, ctx,
1322 TLS_OFFLOAD_CTX_DIR_RX);
1323
1324 dev_put(netdev);
1325
1326 /* Move the context to a separate list for two reasons:
1327 * 1. When the context is deallocated, list_del is called.
1328 * 2. It's no longer an offloaded context, so we don't want to
1329 * run offload-specific code on this context.
1330 */
1331 spin_lock_irqsave(&tls_device_lock, flags);
1332 list_move_tail(&ctx->list, &tls_device_down_list);
1333 spin_unlock_irqrestore(&tls_device_lock, flags);
1334
1335 /* Device contexts for RX and TX will be freed in on sk_destruct
1336 * by tls_device_free_ctx. rx_conf and tx_conf stay in TLS_HW.
1337 * Now release the ref taken above.
1338 */
1339 if (refcount_dec_and_test(&ctx->refcount)) {
1340 /* sk_destruct ran after tls_device_down took a ref, and
1341 * it returned early. Complete the destruction here.
1342 */
1343 list_del(&ctx->list);
1344 tls_device_free_ctx(ctx);
1345 }
1346 }
1347
1348 up_write(&device_offload_lock);
1349
1350 flush_workqueue(destruct_wq);
1351
1352 return NOTIFY_DONE;
1353 }
1354
tls_dev_event(struct notifier_block * this,unsigned long event,void * ptr)1355 static int tls_dev_event(struct notifier_block *this, unsigned long event,
1356 void *ptr)
1357 {
1358 struct net_device *dev = netdev_notifier_info_to_dev(ptr);
1359
1360 if (!dev->tlsdev_ops &&
1361 !(dev->features & (NETIF_F_HW_TLS_RX | NETIF_F_HW_TLS_TX)))
1362 return NOTIFY_DONE;
1363
1364 switch (event) {
1365 case NETDEV_REGISTER:
1366 case NETDEV_FEAT_CHANGE:
1367 if (netif_is_bond_master(dev))
1368 return NOTIFY_DONE;
1369 if ((dev->features & NETIF_F_HW_TLS_RX) &&
1370 !dev->tlsdev_ops->tls_dev_resync)
1371 return NOTIFY_BAD;
1372
1373 if (dev->tlsdev_ops &&
1374 dev->tlsdev_ops->tls_dev_add &&
1375 dev->tlsdev_ops->tls_dev_del)
1376 return NOTIFY_DONE;
1377 else
1378 return NOTIFY_BAD;
1379 case NETDEV_DOWN:
1380 return tls_device_down(dev);
1381 }
1382 return NOTIFY_DONE;
1383 }
1384
1385 static struct notifier_block tls_dev_notifier = {
1386 .notifier_call = tls_dev_event,
1387 };
1388
tls_device_init(void)1389 int __init tls_device_init(void)
1390 {
1391 int err;
1392
1393 dummy_page = alloc_page(GFP_KERNEL);
1394 if (!dummy_page)
1395 return -ENOMEM;
1396
1397 destruct_wq = alloc_workqueue("ktls_device_destruct", 0, 0);
1398 if (!destruct_wq) {
1399 err = -ENOMEM;
1400 goto err_free_dummy;
1401 }
1402
1403 err = register_netdevice_notifier(&tls_dev_notifier);
1404 if (err)
1405 goto err_destroy_wq;
1406
1407 return 0;
1408
1409 err_destroy_wq:
1410 destroy_workqueue(destruct_wq);
1411 err_free_dummy:
1412 put_page(dummy_page);
1413 return err;
1414 }
1415
tls_device_cleanup(void)1416 void __exit tls_device_cleanup(void)
1417 {
1418 unregister_netdevice_notifier(&tls_dev_notifier);
1419 flush_workqueue(destruct_wq);
1420 destroy_workqueue(destruct_wq);
1421 clean_acked_data_flush();
1422 put_page(dummy_page);
1423 }
1424