1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
|
/* Copyright (C) 2015-2017 Jason A. Donenfeld <Jason@zx2c4.com>. All Rights Reserved. */
#include "queueing.h"
#include "device.h"
#include "peer.h"
#include "timers.h"
#include "messages.h"
#include "cookie.h"
#include "socket.h"
#include <linux/ip.h>
#include <linux/ipv6.h>
#include <linux/udp.h>
#include <net/ip_tunnels.h>
static inline void rx_stats(struct wireguard_peer *peer, size_t len)
{
struct pcpu_sw_netstats *tstats = get_cpu_ptr(peer->device->dev->tstats);
u64_stats_update_begin(&tstats->syncp);
tstats->rx_bytes += len;
++tstats->rx_packets;
u64_stats_update_end(&tstats->syncp);
put_cpu_ptr(tstats);
peer->rx_bytes += len;
}
static inline void update_latest_addr(struct wireguard_peer *peer, struct sk_buff *skb)
{
struct endpoint endpoint;
if (!socket_endpoint_from_skb(&endpoint, skb))
socket_set_peer_endpoint(peer, &endpoint);
}
#define SKB_TYPE_LE32(skb) ((struct message_header *)(skb)->data)->type
static inline size_t validate_header_len(struct sk_buff *skb)
{
if (unlikely(skb->len < sizeof(struct message_header)))
return 0;
if (SKB_TYPE_LE32(skb) == cpu_to_le32(MESSAGE_DATA) && skb->len >= MESSAGE_MINIMUM_LENGTH)
return sizeof(struct message_data);
if (SKB_TYPE_LE32(skb) == cpu_to_le32(MESSAGE_HANDSHAKE_INITIATION) && skb->len == sizeof(struct message_handshake_initiation))
return sizeof(struct message_handshake_initiation);
if (SKB_TYPE_LE32(skb) == cpu_to_le32(MESSAGE_HANDSHAKE_RESPONSE) && skb->len == sizeof(struct message_handshake_response))
return sizeof(struct message_handshake_response);
if (SKB_TYPE_LE32(skb) == cpu_to_le32(MESSAGE_HANDSHAKE_COOKIE) && skb->len == sizeof(struct message_handshake_cookie))
return sizeof(struct message_handshake_cookie);
return 0;
}
static inline int skb_prepare_header(struct sk_buff *skb, struct wireguard_device *wg)
{
struct udphdr *udp;
size_t data_offset, data_len, header_len;
if (unlikely(skb_examine_untrusted_ip_hdr(skb) != skb->protocol || skb_transport_header(skb) < skb->head || (skb_transport_header(skb) + sizeof(struct udphdr)) > skb_tail_pointer(skb)))
return -EINVAL; /* Bogus IP header */
udp = udp_hdr(skb);
data_offset = (u8 *)udp - skb->data;
if (unlikely(data_offset > U16_MAX || data_offset + sizeof(struct udphdr) > skb->len))
return -EINVAL; /* Packet has offset at impossible location or isn't big enough to have UDP fields*/
data_len = ntohs(udp->len);
if (unlikely(data_len < sizeof(struct udphdr) || data_len > skb->len - data_offset))
return -EINVAL; /* UDP packet is reporting too small of a size or lying about its size */
data_len -= sizeof(struct udphdr);
data_offset = (u8 *)udp + sizeof(struct udphdr) - skb->data;
if (unlikely(!pskb_may_pull(skb, data_offset + sizeof(struct message_header)) || pskb_trim(skb, data_len + data_offset) < 0))
return -EINVAL;
skb_pull(skb, data_offset);
if (unlikely(skb->len != data_len))
return -EINVAL; /* Final len does not agree with calculated len */
header_len = validate_header_len(skb);
if (unlikely(!header_len))
return -EINVAL;
__skb_push(skb, data_offset);
if (unlikely(!pskb_may_pull(skb, data_offset + header_len)))
return -EINVAL;
__skb_pull(skb, data_offset);
return 0;
}
static void receive_handshake_packet(struct wireguard_device *wg, struct sk_buff *skb)
{
static unsigned long last_under_load = 0; /* Yes this is global, so that our load calculation applies to the whole system. */
struct wireguard_peer *peer = NULL;
bool under_load;
enum cookie_mac_state mac_state;
bool packet_needs_cookie;
if (SKB_TYPE_LE32(skb) == cpu_to_le32(MESSAGE_HANDSHAKE_COOKIE)) {
net_dbg_skb_ratelimited("%s: Receiving cookie response from %pISpfsc\n", wg->dev->name, skb);
cookie_message_consume((struct message_handshake_cookie *)skb->data, wg);
return;
}
under_load = skb_queue_len(&wg->incoming_handshakes) >= MAX_QUEUED_INCOMING_HANDSHAKES / 8;
if (under_load)
last_under_load = jiffies;
else
under_load = time_is_after_jiffies(last_under_load + HZ);
mac_state = cookie_validate_packet(&wg->cookie_checker, skb, under_load);
if ((under_load && mac_state == VALID_MAC_WITH_COOKIE) || (!under_load && mac_state == VALID_MAC_BUT_NO_COOKIE))
packet_needs_cookie = false;
else if (under_load && mac_state == VALID_MAC_BUT_NO_COOKIE)
packet_needs_cookie = true;
else {
net_dbg_skb_ratelimited("%s: Invalid MAC of handshake, dropping packet from %pISpfsc\n", wg->dev->name, skb);
return;
}
switch (SKB_TYPE_LE32(skb)) {
case cpu_to_le32(MESSAGE_HANDSHAKE_INITIATION): {
struct message_handshake_initiation *message = (struct message_handshake_initiation *)skb->data;
if (packet_needs_cookie) {
packet_send_handshake_cookie(wg, skb, message->sender_index);
return;
}
peer = noise_handshake_consume_initiation(message, wg);
if (unlikely(!peer)) {
net_dbg_skb_ratelimited("%s: Invalid handshake initiation from %pISpfsc\n", wg->dev->name, skb);
return;
}
update_latest_addr(peer, skb);
net_dbg_ratelimited("%s: Receiving handshake initiation from peer %Lu (%pISpfsc)\n", wg->dev->name, peer->internal_id, &peer->endpoint.addr);
packet_send_handshake_response(peer);
break;
}
case cpu_to_le32(MESSAGE_HANDSHAKE_RESPONSE): {
struct message_handshake_response *message = (struct message_handshake_response *)skb->data;
if (packet_needs_cookie) {
packet_send_handshake_cookie(wg, skb, message->sender_index);
return;
}
peer = noise_handshake_consume_response(message, wg);
if (unlikely(!peer)) {
net_dbg_skb_ratelimited("%s: Invalid handshake response from %pISpfsc\n", wg->dev->name, skb);
return;
}
update_latest_addr(peer, skb);
net_dbg_ratelimited("%s: Receiving handshake response from peer %Lu (%pISpfsc)\n", wg->dev->name, peer->internal_id, &peer->endpoint.addr);
if (noise_handshake_begin_session(&peer->handshake, &peer->keypairs)) {
timers_session_derived(peer);
timers_handshake_complete(peer);
/* Calling this function will either send any existing packets in the queue
* and not send a keepalive, which is the best case, Or, if there's nothing
* in the queue, it will send a keepalive, in order to give immediate
* confirmation of the session. */
packet_send_keepalive(peer);
}
break;
}
default:
WARN(1, "Somehow a wrong type of packet wound up in the handshake queue!\n");
return;
}
BUG_ON(!peer);
rx_stats(peer, skb->len);
timers_any_authenticated_packet_received(peer);
timers_any_authenticated_packet_traversal(peer);
peer_put(peer);
}
void packet_handshake_receive_worker(struct work_struct *work)
{
struct wireguard_device *wg = container_of(work, struct multicore_worker, work)->ptr;
struct sk_buff *skb;
while ((skb = skb_dequeue(&wg->incoming_handshakes)) != NULL) {
receive_handshake_packet(wg, skb);
dev_kfree_skb(skb);
cond_resched();
}
}
static inline void keep_key_fresh(struct wireguard_peer *peer)
{
struct noise_keypair *keypair;
bool send = false;
if (peer->sent_lastminute_handshake)
return;
rcu_read_lock_bh();
keypair = rcu_dereference_bh(peer->keypairs.current_keypair);
if (likely(keypair && keypair->sending.is_valid) && keypair->i_am_the_initiator &&
unlikely(time_is_before_eq_jiffies64(keypair->sending.birthdate + REJECT_AFTER_TIME - KEEPALIVE_TIMEOUT - REKEY_TIMEOUT)))
send = true;
rcu_read_unlock_bh();
if (send) {
peer->sent_lastminute_handshake = true;
packet_send_queued_handshake_initiation(peer, false);
}
}
static inline bool skb_decrypt(struct sk_buff *skb, struct noise_symmetric_key *key)
{
struct scatterlist sg[MAX_SKB_FRAGS * 2 + 1];
struct sk_buff *trailer;
int num_frags;
if (unlikely(!key))
return false;
if (unlikely(!key->is_valid || time_is_before_eq_jiffies64(key->birthdate + REJECT_AFTER_TIME) || key->counter.receive.counter >= REJECT_AFTER_MESSAGES)) {
key->is_valid = false;
return false;
}
PACKET_CB(skb)->nonce = le64_to_cpu(((struct message_data *)skb->data)->counter);
skb_pull(skb, sizeof(struct message_data));
num_frags = skb_cow_data(skb, 0, &trailer);
if (unlikely(num_frags < 0 || num_frags > ARRAY_SIZE(sg)))
return false;
sg_init_table(sg, num_frags);
if (skb_to_sgvec(skb, sg, 0, skb->len) <= 0)
return false;
if (!chacha20poly1305_decrypt_sg(sg, sg, skb->len, NULL, 0, PACKET_CB(skb)->nonce, key->key))
return false;
return !pskb_trim(skb, skb->len - noise_encrypted_len(0));
}
/* This is RFC6479, a replay detection bitmap algorithm that avoids bitshifts */
static inline bool counter_validate(union noise_counter *counter, u64 their_counter)
{
bool ret = false;
unsigned long index, index_current, top, i;
spin_lock_bh(&counter->receive.lock);
if (unlikely(counter->receive.counter >= REJECT_AFTER_MESSAGES + 1 || their_counter >= REJECT_AFTER_MESSAGES))
goto out;
++their_counter;
if (unlikely((COUNTER_WINDOW_SIZE + their_counter) < counter->receive.counter))
goto out;
index = their_counter >> ilog2(BITS_PER_LONG);
if (likely(their_counter > counter->receive.counter)) {
index_current = counter->receive.counter >> ilog2(BITS_PER_LONG);
top = min_t(unsigned long, index - index_current, COUNTER_BITS_TOTAL / BITS_PER_LONG);
for (i = 1; i <= top; ++i)
counter->receive.backtrack[(i + index_current) & ((COUNTER_BITS_TOTAL / BITS_PER_LONG) - 1)] = 0;
counter->receive.counter = their_counter;
}
index &= (COUNTER_BITS_TOTAL / BITS_PER_LONG) - 1;
ret = !test_and_set_bit(their_counter & (BITS_PER_LONG - 1), &counter->receive.backtrack[index]);
out:
spin_unlock_bh(&counter->receive.lock);
return ret;
}
#include "selftest/counter.h"
static void packet_consume_data_done(struct sk_buff *skb, struct wireguard_peer *peer, struct endpoint *endpoint, bool used_new_key)
{
struct net_device *dev = peer->device->dev;
struct wireguard_peer *routed_peer;
unsigned int len;
socket_set_peer_endpoint(peer, endpoint);
if (unlikely(used_new_key)) {
timers_handshake_complete(peer);
packet_send_staged_packets(peer);
}
keep_key_fresh(peer);
/* A packet with length 0 is a keepalive packet */
if (unlikely(!skb->len)) {
net_dbg_ratelimited("%s: Receiving keepalive packet from peer %Lu (%pISpfsc)\n", dev->name, peer->internal_id, &peer->endpoint.addr);
goto packet_processed;
}
if (unlikely(skb_network_header(skb) < skb->head))
goto dishonest_packet_size;
if (unlikely(!(pskb_network_may_pull(skb, sizeof(struct iphdr)) && (ip_hdr(skb)->version == 4 || (ip_hdr(skb)->version == 6 && pskb_network_may_pull(skb, sizeof(struct ipv6hdr)))))))
goto dishonest_packet_type;
skb->dev = dev;
skb->ip_summed = CHECKSUM_UNNECESSARY;
skb->protocol = skb_examine_untrusted_ip_hdr(skb);
if (skb->protocol == htons(ETH_P_IP)) {
len = ntohs(ip_hdr(skb)->tot_len);
if (unlikely(len < sizeof(struct iphdr)))
goto dishonest_packet_size;
if (INET_ECN_is_ce(PACKET_CB(skb)->ds))
IP_ECN_set_ce(ip_hdr(skb));
} else if (skb->protocol == htons(ETH_P_IPV6)) {
len = ntohs(ipv6_hdr(skb)->payload_len) + sizeof(struct ipv6hdr);
if (INET_ECN_is_ce(PACKET_CB(skb)->ds))
IP6_ECN_set_ce(skb, ipv6_hdr(skb));
} else
goto dishonest_packet_type;
if (unlikely(len > skb->len))
goto dishonest_packet_size;
if (unlikely(pskb_trim(skb, len)))
goto packet_processed;
timers_data_received(peer);
routed_peer = routing_table_lookup_src(&peer->device->peer_routing_table, skb);
peer_put(routed_peer); /* We don't need the extra reference. */
if (unlikely(routed_peer != peer))
goto dishonest_packet_peer;
len = skb->len;
if (unlikely(netif_receive_skb(skb) == NET_RX_DROP)) {
++dev->stats.rx_dropped;
net_dbg_ratelimited("%s: Failed to give packet to userspace from peer %Lu (%pISpfsc)\n", dev->name, peer->internal_id, &peer->endpoint.addr);
} else
rx_stats(peer, len);
goto continue_processing;
dishonest_packet_peer:
net_dbg_skb_ratelimited("%s: Packet has unallowed src IP (%pISc) from peer %Lu (%pISpfsc)\n", dev->name, skb, peer->internal_id, &peer->endpoint.addr);
++dev->stats.rx_errors;
++dev->stats.rx_frame_errors;
goto packet_processed;
dishonest_packet_type:
net_dbg_ratelimited("%s: Packet is neither ipv4 nor ipv6 from peer %Lu (%pISpfsc)\n", dev->name, peer->internal_id, &peer->endpoint.addr);
++dev->stats.rx_errors;
++dev->stats.rx_frame_errors;
goto packet_processed;
dishonest_packet_size:
net_dbg_ratelimited("%s: Packet has incorrect size from peer %Lu (%pISpfsc)\n", dev->name, peer->internal_id, &peer->endpoint.addr);
++dev->stats.rx_errors;
++dev->stats.rx_length_errors;
goto packet_processed;
packet_processed:
dev_kfree_skb(skb);
continue_processing:
timers_any_authenticated_packet_received(peer);
timers_any_authenticated_packet_traversal(peer);
}
void packet_rx_worker(struct work_struct *work)
{
struct crypt_ctx *ctx;
struct crypt_queue *queue = container_of(work, struct crypt_queue, work);
local_bh_disable();
spin_lock_bh(&queue->ring.consumer_lock);
while ((ctx = __ptr_ring_peek(&queue->ring)) != NULL && atomic_read(&ctx->is_finished)) {
__ptr_ring_discard_one(&queue->ring);
if (likely(ctx->skb)) {
if (likely(counter_validate(&ctx->keypair->receiving.counter, PACKET_CB(ctx->skb)->nonce))) {
skb_reset(ctx->skb);
packet_consume_data_done(ctx->skb, ctx->peer, &ctx->endpoint, noise_received_with_keypair(&ctx->peer->keypairs, ctx->keypair));
}
else {
net_dbg_ratelimited("%s: Packet has invalid nonce %Lu (max %Lu)\n", ctx->peer->device->dev->name, PACKET_CB(ctx->skb)->nonce, ctx->keypair->receiving.counter.receive.counter);
dev_kfree_skb(ctx->skb);
}
}
noise_keypair_put(ctx->keypair);
peer_put(ctx->peer);
kmem_cache_free(crypt_ctx_cache, ctx);
}
spin_unlock_bh(&queue->ring.consumer_lock);
local_bh_enable();
}
void packet_decrypt_worker(struct work_struct *work)
{
struct crypt_ctx *ctx;
struct crypt_queue *queue = container_of(work, struct multicore_worker, work)->ptr;
struct wireguard_peer *peer;
while ((ctx = ptr_ring_consume_bh(&queue->ring)) != NULL) {
if (unlikely(socket_endpoint_from_skb(&ctx->endpoint, ctx->skb) < 0 || !skb_decrypt(ctx->skb, &ctx->keypair->receiving))) {
dev_kfree_skb(ctx->skb);
ctx->skb = NULL;
}
/* Dereferencing ctx is unsafe once ctx->is_finished == true, so
* we take a reference here first. */
peer = peer_rcu_get(ctx->peer);
atomic_set(&ctx->is_finished, true);
queue_work_on(cpumask_choose_online(&peer->serial_work_cpu, peer->internal_id), peer->device->packet_crypt_wq, &peer->rx_queue.work);
peer_put(peer);
}
}
static void packet_consume_data(struct wireguard_device *wg, struct sk_buff *skb)
{
struct crypt_ctx *ctx;
struct noise_keypair *keypair;
__le32 idx = ((struct message_data *)skb->data)->key_idx;
int ret;
rcu_read_lock_bh();
keypair = noise_keypair_get((struct noise_keypair *)index_hashtable_lookup(&wg->index_hashtable, INDEX_HASHTABLE_KEYPAIR, idx));
rcu_read_unlock_bh();
if (unlikely(!keypair)) {
dev_kfree_skb(skb);
return;
}
ctx = kmem_cache_alloc(crypt_ctx_cache, GFP_ATOMIC);
if (unlikely(!ctx)) {
dev_kfree_skb(skb);
peer_put(keypair->entry.peer);
noise_keypair_put(keypair);
return;
}
atomic_set(&ctx->is_finished, false);
ctx->keypair = keypair;
ctx->skb = skb;
/* We already have a reference to peer from index_hashtable_lookup. */
ctx->peer = ctx->keypair->entry.peer;
ret = queue_enqueue_per_device_and_peer(&wg->decrypt_queue, &ctx->peer->rx_queue, ctx, wg->packet_crypt_wq, &wg->decrypt_queue.last_cpu);
if (likely(!ret))
return; /* Successful. No need to drop references below. */
dev_kfree_skb(ctx->skb);
if (ret == -EPIPE) {
ctx->skb = NULL;
atomic_set(&ctx->is_finished, true);
queue_work_on(cpumask_choose_online(&ctx->peer->serial_work_cpu, ctx->peer->internal_id), ctx->peer->device->packet_crypt_wq, &ctx->peer->rx_queue.work);
} else {
noise_keypair_put(ctx->keypair);
peer_put(ctx->peer);
kmem_cache_free(crypt_ctx_cache, ctx);
}
}
void packet_receive(struct wireguard_device *wg, struct sk_buff *skb)
{
if (unlikely(skb_prepare_header(skb, wg) < 0))
goto err;
switch (SKB_TYPE_LE32(skb)) {
case cpu_to_le32(MESSAGE_HANDSHAKE_INITIATION):
case cpu_to_le32(MESSAGE_HANDSHAKE_RESPONSE):
case cpu_to_le32(MESSAGE_HANDSHAKE_COOKIE): {
int cpu;
if (skb_queue_len(&wg->incoming_handshakes) > MAX_QUEUED_INCOMING_HANDSHAKES) {
net_dbg_skb_ratelimited("%s: Too many handshakes queued, dropping packet from %pISpfsc\n", wg->dev->name, skb);
goto err;
}
skb_queue_tail(&wg->incoming_handshakes, skb);
/* Queues up a call to packet_process_queued_handshake_packets(skb): */
cpu = cpumask_next_online(&wg->incoming_handshake_cpu);
queue_work_on(cpu, wg->handshake_receive_wq, &per_cpu_ptr(wg->incoming_handshakes_worker, cpu)->work);
break;
}
case cpu_to_le32(MESSAGE_DATA):
PACKET_CB(skb)->ds = ip_tunnel_get_dsfield(ip_hdr(skb), skb);
packet_consume_data(wg, skb);
break;
default:
net_dbg_skb_ratelimited("%s: Invalid packet from %pISpfsc\n", wg->dev->name, skb);
goto err;
}
return;
err:
dev_kfree_skb(skb);
}
|