-
Notifications
You must be signed in to change notification settings - Fork 4
/
tcp.c
1500 lines (1303 loc) · 39.4 KB
/
tcp.c
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
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
#include <pthread.h>
#include <stdint.h>
#include <stdlib.h>
#include <string.h>
#include <sys/time.h>
#include <time.h>
#include <unistd.h>
#include "ip.h"
#include "util.h"
// TODO: user timeout should set by user
#define USER_TIMEOUT (20) /* user timeout (seconds) */
#define TIME_WAIT_TIMEOUT (2 * 10) /* TIME_WAIT timeout (seconds) */
#define TCP_SND_BUF_SIZE (10 * 1024)
#define TCP_CB_TABLE_SIZE 128
#define TCP_SOURCE_PORT_MIN 49152
#define TCP_SOURCE_PORT_MAX 65535
#define TCP_CB_STATE_CLOSED 0
#define TCP_CB_STATE_LISTEN 1
#define TCP_CB_STATE_SYN_SENT 2
#define TCP_CB_STATE_SYN_RCVD 3
#define TCP_CB_STATE_ESTABLISHED 4
#define TCP_CB_STATE_FIN_WAIT1 5
#define TCP_CB_STATE_FIN_WAIT2 6
#define TCP_CB_STATE_CLOSING 7
#define TCP_CB_STATE_TIME_WAIT 8
#define TCP_CB_STATE_CLOSE_WAIT 9
#define TCP_CB_STATE_LAST_ACK 10
#define TCP_FLG_FIN 0x01
#define TCP_FLG_SYN 0x02
#define TCP_FLG_RST 0x04
#define TCP_FLG_PSH 0x08
#define TCP_FLG_ACK 0x10
#define TCP_FLG_URG 0x20
#define TCP_FLG_IS(x, y) (((x)&0x3f) == (y))
#define TCP_FLG_ISSET(x, y) (((x)&0x3f) & (y))
#define TCP_HDR_LEN(hdr) (((hdr)->off >> 4) << 2)
#define TCP_DATA_LEN(hdr, len) ((len)-TCP_HDR_LEN(hdr))
#define IS_FREE_CB(cb) (!(cb)->used && (cb)->state == TCP_CB_STATE_CLOSED)
#define TCP_SOCKET_INVALID(x) ((x) < 0 || (x) >= TCP_CB_TABLE_SIZE)
#ifndef TCP_DEBUG
#ifdef DEBUG
#define TCP_DEBUG 1
#endif
#endif
struct tcp_hdr {
uint16_t src;
uint16_t dst;
uint32_t seq;
uint32_t ack;
uint8_t off;
uint8_t flg;
uint16_t win;
uint16_t sum;
uint16_t urg;
};
struct tcp_txq_entry {
struct tcp_hdr *segment;
uint16_t len;
struct timeval timestamp;
struct tcp_txq_entry *next;
};
struct tcp_txq_head {
struct tcp_txq_entry *head;
struct tcp_txq_entry *tail;
uint16_t snt;
};
struct tcp_cb {
uint8_t used;
uint8_t state;
struct netif *iface;
uint16_t port; // network byte order
struct {
ip_addr_t addr;
uint16_t port; // network byte order
} peer;
struct {
uint32_t nxt;
uint32_t una;
uint16_t up;
uint32_t wl1;
uint32_t wl2;
uint16_t wnd;
} snd;
uint32_t iss;
struct {
uint32_t nxt;
uint16_t up;
uint16_t wnd;
} rcv;
uint32_t irs;
struct tcp_txq_head txq;
uint8_t window[65535];
struct tcp_cb *parent;
struct queue_head backlog;
pthread_cond_t cond;
long timeout;
};
static struct tcp_cb cb_table[TCP_CB_TABLE_SIZE];
static pthread_mutex_t mutex = PTHREAD_MUTEX_INITIALIZER;
static pthread_t timer_thread;
static pthread_cond_t timer_cond;
static ssize_t tcp_tx(struct tcp_cb *cb, uint32_t seq, uint32_t ack,
uint8_t flg, struct timeval *now, uint8_t *buf,
size_t len);
static char *tcp_flg_ntop(uint8_t flg, char *buf, int len) {
int i = 0;
if (TCP_FLG_ISSET(flg, TCP_FLG_FIN)) {
buf[i++] = 'F';
}
if (TCP_FLG_ISSET(flg, TCP_FLG_SYN)) {
buf[i++] = 'S';
}
if (TCP_FLG_ISSET(flg, TCP_FLG_RST)) {
buf[i++] = 'R';
}
if (TCP_FLG_ISSET(flg, TCP_FLG_PSH)) {
buf[i++] = 'P';
}
if (TCP_FLG_ISSET(flg, TCP_FLG_ACK)) {
buf[i++] = 'A';
}
if (TCP_FLG_ISSET(flg, TCP_FLG_URG)) {
buf[i++] = 'U';
}
buf[i] = 0;
return buf;
}
static char *tcp_state_ntop(uint8_t state) {
switch (state) {
case TCP_CB_STATE_CLOSED:
return "CLOSED";
case TCP_CB_STATE_LISTEN:
return "LISTEN";
case TCP_CB_STATE_SYN_SENT:
return "SYN_SENT";
case TCP_CB_STATE_SYN_RCVD:
return "SYN_RCVD";
case TCP_CB_STATE_ESTABLISHED:
return "ESTABLISHED";
case TCP_CB_STATE_FIN_WAIT1:
return "FIN_WAIT1";
case TCP_CB_STATE_FIN_WAIT2:
return "FIN_WAIT2";
case TCP_CB_STATE_CLOSING:
return "CLOSING";
case TCP_CB_STATE_TIME_WAIT:
return "TIME_WAIT";
case TCP_CB_STATE_CLOSE_WAIT:
return "CLOSE_WAIT";
case TCP_CB_STATE_LAST_ACK:
return "LAST_ACK";
default:
return "UNKNOWN";
}
}
static void tcp_state_dump(struct tcp_cb *cb) {
char buf[64];
fprintf(stderr, " used: %d\n", cb->used);
fprintf(stderr, " state: %s\n", tcp_state_ntop(cb->state));
fprintf(stderr, " self.port: %u\n", ntoh16(cb->port));
fprintf(stderr, " peer.addr: %s\n", ip_addr_ntop(&cb->peer.addr, buf, 64));
fprintf(stderr, " peer.port: %u\n", ntoh16(cb->peer.port));
fprintf(stderr, " snd.nxt: %u\n", cb->snd.nxt);
fprintf(stderr, " snd.una: %u\n", cb->snd.una);
fprintf(stderr, " snd.wnd: %u\n", cb->snd.wnd);
fprintf(stderr, " txq.snt: %u\n", cb->txq.snt);
fprintf(stderr, " rcv.nxt: %u\n", cb->rcv.nxt);
fprintf(stderr, " rcv.wnd: %u\n", cb->rcv.wnd);
fprintf(stderr, " n_backlog: %u\n", cb->backlog.num);
fprintf(stderr, " timeout: %ld\n", cb->timeout);
}
static void tcp_dump(struct tcp_cb *cb, struct tcp_hdr *hdr, size_t plen) {
char buf[64];
tcp_state_dump(cb);
fprintf(stderr, " len: %lu\n", plen);
fprintf(stderr, " src: %u\n", ntoh16(hdr->src));
fprintf(stderr, " dst: %u\n", ntoh16(hdr->dst));
fprintf(stderr, " seq: %u\n", ntoh32(hdr->seq));
fprintf(stderr, " ack: %u\n", ntoh32(hdr->ack));
fprintf(stderr, " off: %u\n", hdr->off);
fprintf(stderr, " flg: [%s]\n", tcp_flg_ntop(hdr->flg, buf, 64));
fprintf(stderr, " win: %u\n", ntoh16(hdr->win));
fprintf(stderr, " sum: %u\n", ntoh16(hdr->sum));
fprintf(stderr, " urg: %u\n", ntoh16(hdr->urg));
}
static uint16_t tcp_checksum(ip_addr_t self, ip_addr_t peer, uint8_t *segment,
size_t len) {
uint32_t pseudo = 0;
pseudo += (self >> 16) & 0xffff;
pseudo += self & 0xffff;
pseudo += (peer >> 16) & 0xffff;
pseudo += self & 0xffff;
pseudo += hton16((uint16_t)IP_PROTOCOL_TCP);
pseudo += hton16(len);
return cksum16((uint16_t *)segment, len, pseudo);
}
/*
* Segment Queue
*/
static struct tcp_txq_entry *tcp_txq_add(struct tcp_cb *cb, struct tcp_hdr *hdr,
size_t len) {
struct tcp_txq_entry *txq;
txq = malloc(sizeof(struct tcp_txq_entry));
if (!txq) {
return NULL;
}
txq->segment = hdr;
txq->len = len;
// clear timestamp
memset(&txq->timestamp, 0, sizeof(txq->timestamp));
txq->next = NULL;
// set txq to next of tail entry
if (cb->txq.head == NULL) {
cb->txq.head = txq;
} else {
cb->txq.tail->next = txq;
}
// update tail entry
cb->txq.tail = txq;
return txq;
}
static void tcp_txq_clear_all(struct tcp_cb *cb) {
struct tcp_txq_entry *txq = cb->txq.head, *next;
while (txq) {
next = txq->next;
free(txq->segment);
free(txq);
txq = next;
}
cb->txq.head = cb->txq.tail = NULL;
}
/*
* EVENT PROCESSING
* https://tools.ietf.org/html/rfc793#section-3.9
*/
static void tcp_close_cb(struct tcp_cb *cb) {
cb->state = TCP_CB_STATE_CLOSED;
memset(&cb->snd, 0, sizeof(cb->snd));
cb->iss = 0;
memset(&cb->rcv, 0, sizeof(cb->rcv));
cb->irs = 0;
tcp_txq_clear_all(cb);
return;
}
// SEGMENT ARRIVES
// https://tools.ietf.org/html/rfc793#page-65
static void tcp_event_segment_arrives(struct tcp_cb *cb, struct tcp_hdr *hdr,
size_t len) {
size_t plen;
int acceptable = 0;
struct timeval now;
plen = TCP_DATA_LEN(hdr, len);
if (gettimeofday(&now, NULL) == -1) {
perror("gettimeofday");
return;
}
switch (cb->state) {
case TCP_CB_STATE_CLOSED:
if (!TCP_FLG_ISSET(hdr->flg, TCP_FLG_RST)) {
if (TCP_FLG_ISSET(hdr->flg, TCP_FLG_ACK)) {
tcp_tx(cb, ntoh32(hdr->ack), 0, TCP_FLG_RST, &now, NULL, 0);
} else {
tcp_tx(cb, 0, ntoh32(hdr->seq) + plen, TCP_FLG_RST | TCP_FLG_ACK,
&now, NULL, 0);
}
}
return;
case TCP_CB_STATE_LISTEN:
// first check for an RST
if (TCP_FLG_ISSET(hdr->flg, TCP_FLG_RST)) {
// incoming RST is ignored
goto ERROR_RX_LISTEN;
}
// second check for an ACK
if (TCP_FLG_ISSET(hdr->flg, TCP_FLG_ACK)) {
tcp_tx(cb, ntoh32(hdr->ack), 0, TCP_FLG_RST, &now, NULL, 0);
goto ERROR_RX_LISTEN;
}
// third check for a SYN
if (TCP_FLG_ISSET(hdr->flg, TCP_FLG_SYN)) {
// TODO: check the security
// TODO: If the SEG.PRC is greater than the TCB.PRC
// else
cb->rcv.wnd = sizeof(cb->window);
cb->rcv.nxt = ntoh32(hdr->seq) + 1;
cb->irs = ntoh32(hdr->seq);
cb->iss = (uint32_t)random();
tcp_tx(cb, cb->iss, cb->rcv.nxt, TCP_FLG_SYN | TCP_FLG_ACK, &now, NULL,
0);
cb->snd.nxt = cb->iss + 1;
cb->snd.una = cb->iss;
cb->timeout = now.tv_sec + USER_TIMEOUT;
cb->state = TCP_CB_STATE_SYN_RCVD;
// TODO: ? queue to backlog ?
// TODO: increment hdr->seq for save text
goto CHECK_URG;
}
// no packet should come here. drop segment
ERROR_RX_LISTEN:
// return state to CLOSED
tcp_close_cb(cb);
pthread_cond_broadcast(&cb->cond);
cb->parent = NULL;
return;
case TCP_CB_STATE_SYN_SENT:
// first check the ACK bit
if (TCP_FLG_ISSET(hdr->flg, TCP_FLG_ACK)) {
if (ntoh32(hdr->ack) <= cb->iss || ntoh32(hdr->ack) > cb->snd.nxt) {
tcp_tx(cb, ntoh32(hdr->ack), 0, TCP_FLG_RST, &now, NULL, 0);
return;
}
if (cb->snd.una <= ntoh32(hdr->ack) &&
ntoh32(hdr->ack) <= cb->snd.nxt) {
acceptable = 1;
} else {
// drop invalid ack
return;
}
}
// second check the RST bit
if (TCP_FLG_ISSET(hdr->flg, TCP_FLG_RST)) {
if (!acceptable) {
// drop segment
return;
}
fprintf(stderr, "error: connection reset\n");
tcp_close_cb(cb);
pthread_cond_signal(&cb->cond);
return;
}
// TODO: third check the security and precedence
// fourth check the SYN bit
if (TCP_FLG_ISSET(hdr->flg, TCP_FLG_SYN)) {
cb->rcv.nxt = ntoh32(hdr->seq) + 1;
cb->irs = ntoh32(hdr->seq);
// TODO: ? if there is an ACK ?
if (cb->snd.una < ntoh32(hdr->ack)) {
// update snd.una and user timeout
cb->snd.una = ntoh32(hdr->ack);
cb->timeout = now.tv_sec + USER_TIMEOUT;
pthread_cond_signal(&timer_cond);
}
// TODO: clear all retransmission queue
if (cb->snd.una > cb->iss) {
// our SYN has been ACKed
cb->state = TCP_CB_STATE_ESTABLISHED;
tcp_tx(cb, cb->snd.nxt, cb->rcv.nxt, TCP_FLG_ACK, &now, NULL, 0);
pthread_cond_signal(&cb->cond);
if (plen > 0 || TCP_FLG_ISSET(hdr->flg, TCP_FLG_URG)) {
goto CHECK_URG;
}
return;
} else {
cb->state = TCP_CB_STATE_SYN_RCVD;
tcp_tx(cb, cb->iss, cb->rcv.nxt, TCP_FLG_SYN | TCP_FLG_ACK, &now,
NULL, 0);
pthread_cond_signal(&cb->cond);
// TODO: If there are other controls or text in the segment, queue
// them for processing after the ESTABLISHED state has been reached,
return;
}
}
// fifth, if neither of the SYN or RST bits is set
if (!TCP_FLG_ISSET(hdr->flg, TCP_FLG_ACK | TCP_FLG_RST)) {
// drop segment
return;
}
return;
case TCP_CB_STATE_SYN_RCVD:
case TCP_CB_STATE_ESTABLISHED:
case TCP_CB_STATE_FIN_WAIT1:
case TCP_CB_STATE_FIN_WAIT2:
case TCP_CB_STATE_CLOSING:
case TCP_CB_STATE_TIME_WAIT:
case TCP_CB_STATE_CLOSE_WAIT:
case TCP_CB_STATE_LAST_ACK:
break;
default:
fprintf(stderr, ">>> segment recv : not implement tcp state (%d) <<<\n",
cb->state);
return;
}
// first check sequence number
if (plen > 0) {
if (cb->rcv.wnd > 0) {
acceptable = (cb->rcv.nxt <= ntoh32(hdr->seq) &&
ntoh32(hdr->seq) < cb->rcv.nxt + cb->rcv.wnd) ||
(cb->rcv.nxt <= ntoh32(hdr->seq) &&
ntoh32(hdr->seq) + plen - 1 < cb->rcv.nxt + cb->rcv.wnd);
} else {
acceptable = 0;
}
} else {
if (cb->rcv.wnd > 0) {
acceptable = (cb->rcv.nxt <= ntoh32(hdr->seq) &&
ntoh32(hdr->seq) < cb->rcv.nxt + cb->rcv.wnd);
} else {
acceptable = ntoh32(hdr->seq) == cb->rcv.nxt;
}
}
if (!acceptable) {
if (!TCP_FLG_ISSET(hdr->flg, TCP_FLG_RST)) {
fprintf(stderr, "is not acceptable !!!\n");
tcp_tx(cb, cb->snd.nxt, cb->rcv.nxt, TCP_FLG_ACK, &now, NULL, 0);
}
// drop segment
return;
}
// second check the RST bit
switch (cb->state) {
case TCP_CB_STATE_SYN_RCVD:
if (TCP_FLG_ISSET(hdr->flg, TCP_FLG_RST)) {
// close connection
tcp_close_cb(cb);
pthread_cond_signal(&cb->cond);
return;
}
break;
case TCP_CB_STATE_ESTABLISHED:
case TCP_CB_STATE_FIN_WAIT1:
case TCP_CB_STATE_FIN_WAIT2:
case TCP_CB_STATE_CLOSE_WAIT:
if (TCP_FLG_ISSET(hdr->flg, TCP_FLG_RST)) {
// TODO: signal to SEND, RECEIVE waiting thread.
tcp_close_cb(cb);
pthread_cond_broadcast(&cb->cond);
return;
}
break;
case TCP_CB_STATE_CLOSING:
case TCP_CB_STATE_TIME_WAIT:
case TCP_CB_STATE_LAST_ACK:
if (TCP_FLG_ISSET(hdr->flg, TCP_FLG_RST)) {
// close connection
tcp_close_cb(cb);
pthread_cond_broadcast(&cb->cond);
return;
}
break;
default:
fprintf(stderr, ">>> seg recv (2) : not implement tcp state (%d) <<<\n",
cb->state);
break;
}
// TODO: third check security and precedence
// fourth, check the SYN bit
if (TCP_FLG_ISSET(hdr->flg, TCP_FLG_SYN)) {
tcp_tx(cb, 0, cb->rcv.nxt, TCP_FLG_RST, &now, NULL, 0);
tcp_close_cb(cb);
pthread_cond_broadcast(&cb->cond);
return;
}
// TODO: ? If the SYN is not in the window this step would not be reached and
// an ack would have been sent in the first step (sequence number check). ?
// fifth check the ACK field
if (TCP_FLG_ISSET(hdr->flg, TCP_FLG_ACK)) {
switch (cb->state) {
case TCP_CB_STATE_SYN_RCVD:
if (!(cb->snd.una <= ntoh32(hdr->ack) &&
ntoh32(hdr->ack) <= cb->snd.nxt)) {
// hdr->ack is not acceptable
tcp_tx(cb, ntoh32(hdr->ack), cb->rcv.nxt, TCP_FLG_RST, &now, NULL, 0);
// The connection remains in the same state after send RST
break;
}
cb->state = TCP_CB_STATE_ESTABLISHED;
if (cb->parent) {
// add cb to backlog
queue_push(&cb->parent->backlog, cb, sizeof(*cb));
pthread_cond_signal(&cb->parent->cond);
} else {
// parent == NULL means cb is created by user and first state was
// SYN_SENT
pthread_cond_signal(&cb->cond);
}
// enter ESTABLISHED state and continue processing
case TCP_CB_STATE_ESTABLISHED:
case TCP_CB_STATE_FIN_WAIT1:
case TCP_CB_STATE_FIN_WAIT2:
case TCP_CB_STATE_CLOSE_WAIT:
case TCP_CB_STATE_CLOSING:
if (cb->snd.una <= ntoh32(hdr->ack) &&
ntoh32(hdr->ack) <= cb->snd.nxt) {
if (cb->snd.una < ntoh32(hdr->ack)) {
// update snd.una and user timeout
cb->snd.una = ntoh32(hdr->ack);
cb->timeout = now.tv_sec + USER_TIMEOUT;
pthread_cond_signal(&timer_cond);
}
// TODO: retransmission queue send
pthread_cond_broadcast(&cb->cond);
if ((cb->snd.wl1 < ntoh32(hdr->seq)) ||
(cb->snd.wl1 == ntoh32(hdr->seq) &&
cb->snd.wl2 <= ntoh32(hdr->ack))) {
cb->snd.wnd = ntoh16(hdr->win);
cb->snd.wl1 = ntoh32(hdr->seq);
cb->snd.wl2 = ntoh32(hdr->ack);
}
} else if (ntoh32(hdr->ack) > cb->snd.nxt) {
fprintf(stderr, "recv ack but ack is advanced to snd.nxt\n");
tcp_tx(cb, cb->snd.nxt, cb->rcv.nxt, TCP_FLG_ACK, &now, NULL, 0);
// drop the segment
return;
}
// else if SEG.ACK < SND.UNA then it can be ignored
if (cb->state == TCP_CB_STATE_FIN_WAIT1) {
// if this ACK is for sent FIN
if (ntoh32(hdr->ack) + 1 == cb->snd.nxt) {
cb->state = TCP_CB_STATE_FIN_WAIT2;
}
} else if (cb->state == TCP_CB_STATE_FIN_WAIT2) {
// TODO: if the retransmission queue is empty, the user's CLOSE can be
// acknowledged ("ok")
} else if (cb->state == TCP_CB_STATE_CLOSING) {
// if this ACK is for sent FIN
if (ntoh32(hdr->ack) + 1 == cb->snd.nxt) {
cb->state = TCP_CB_STATE_TIME_WAIT;
}
}
break;
case TCP_CB_STATE_LAST_ACK:
// if this ACK is for sent FIN
if (ntoh32(hdr->ack) == cb->snd.nxt) {
tcp_close_cb(cb);
pthread_cond_broadcast(&cb->cond);
return;
}
break;
case TCP_CB_STATE_TIME_WAIT:
// TODO: restart the 2 MSL timeout
break;
default:
fprintf(stderr, ">>> seg recv (5) : not implement tcp state (%d) <<<\n",
cb->state);
break;
}
}
CHECK_URG:
// sixth, check the URG bit
if (TCP_FLG_ISSET(hdr->flg, TCP_FLG_URG)) {
switch (cb->state) {
case TCP_CB_STATE_ESTABLISHED:
case TCP_CB_STATE_FIN_WAIT1:
case TCP_CB_STATE_FIN_WAIT2:
cb->rcv.up = MAX(cb->rcv.up, ntoh16(hdr->urg));
// TODO: signal
break;
case TCP_CB_STATE_CLOSING:
case TCP_CB_STATE_TIME_WAIT:
case TCP_CB_STATE_CLOSE_WAIT:
case TCP_CB_STATE_LAST_ACK:
// this should not occur. ignore the urg
break;
case TCP_CB_STATE_SYN_RCVD:
// do nothing
break;
default:
fprintf(stderr, ">>> seg recv (6) : not implement tcp state (%d) <<<\n",
cb->state);
break;
}
}
// seventh, process the segment text
switch (cb->state) {
case TCP_CB_STATE_ESTABLISHED:
case TCP_CB_STATE_FIN_WAIT1:
case TCP_CB_STATE_FIN_WAIT2:
// TODO: accept not ordered packet
if (plen > 0 && cb->rcv.nxt == ntoh32(hdr->seq)) {
// copy segment to receive buffer
memcpy(cb->window + (sizeof(cb->window) - cb->rcv.wnd),
(uint8_t *)hdr + TCP_HDR_LEN(hdr), plen);
cb->rcv.nxt = ntoh32(hdr->seq) + plen;
cb->rcv.wnd -= plen;
tcp_tx(cb, cb->snd.nxt, cb->rcv.nxt, TCP_FLG_ACK, &now, NULL, 0);
pthread_cond_broadcast(&cb->cond);
} else if (TCP_FLG_ISSET(hdr->flg, TCP_FLG_PSH)) {
tcp_tx(cb, cb->snd.nxt, cb->rcv.nxt, TCP_FLG_ACK, &now, NULL, 0);
pthread_cond_broadcast(&cb->cond);
}
break;
case TCP_CB_STATE_CLOSING:
case TCP_CB_STATE_TIME_WAIT:
case TCP_CB_STATE_CLOSE_WAIT:
case TCP_CB_STATE_LAST_ACK:
// this should not occur. ignore the text
break;
case TCP_CB_STATE_SYN_RCVD:
// do nothing
break;
default:
fprintf(stderr, ">>> seg recv (7) : not implement tcp state (%d) <<<\n",
cb->state);
break;
}
// eighth, check the FIN bit
if (TCP_FLG_ISSET(hdr->flg, TCP_FLG_FIN)) {
cb->rcv.nxt = ntoh32(hdr->seq) + 1;
tcp_tx(cb, cb->snd.nxt, cb->rcv.nxt, TCP_FLG_ACK, &now, NULL, 0);
switch (cb->state) {
case TCP_CB_STATE_SYN_RCVD:
case TCP_CB_STATE_ESTABLISHED:
cb->state = TCP_CB_STATE_CLOSE_WAIT;
break;
case TCP_CB_STATE_FIN_WAIT1:
cb->state = TCP_CB_STATE_CLOSING;
break;
case TCP_CB_STATE_FIN_WAIT2:
cb->state = TCP_CB_STATE_TIME_WAIT;
// start time-wait timer
cb->timeout = now.tv_sec + TIME_WAIT_TIMEOUT;
// TODO: turn off other timers
break;
case TCP_CB_STATE_CLOSING:
case TCP_CB_STATE_CLOSE_WAIT:
case TCP_CB_STATE_LAST_ACK:
// remain state
break;
case TCP_CB_STATE_TIME_WAIT:
// remain state
// restart the 2MSL timeout
cb->timeout = now.tv_sec + TIME_WAIT_TIMEOUT;
break;
default:
fprintf(stderr, ">>> seg recv (8) : not implement tcp state (%d) <<<\n",
cb->state);
break;
}
// signal "connection closing"
pthread_cond_broadcast(&cb->cond);
}
return;
}
/*
* TCP APPLICATION CONTROLLER
*/
// TODO: where to get current time.
static ssize_t tcp_tx(struct tcp_cb *cb, uint32_t seq, uint32_t ack,
uint8_t flg, struct timeval *now, uint8_t *buf,
size_t len) {
uint8_t *segment;
struct tcp_hdr *hdr;
ip_addr_t self, peer;
struct tcp_txq_entry *txq = NULL;
int have_unsent;
// allocate segment
segment = malloc(sizeof(struct tcp_hdr) + len);
if (!segment) {
return -1;
}
// set header params
hdr = (struct tcp_hdr *)segment;
memset(hdr, 0, sizeof(struct tcp_hdr));
hdr->src = cb->port;
hdr->dst = cb->peer.port;
hdr->seq = hton32(seq);
hdr->ack = hton32(ack);
hdr->off = (sizeof(struct tcp_hdr) >> 2) << 4;
hdr->flg = flg;
hdr->win = hton16(cb->rcv.wnd);
hdr->sum = 0;
hdr->urg = 0;
// copy data
memcpy(hdr + 1, buf, len);
if (len > 0 || flg & (TCP_FLG_SYN | TCP_FLG_FIN)) {
// add txq list only packets which have ack reply
// check unsent segment exists in cb->txq
have_unsent = cb->txq.tail && cb->txq.tail->timestamp.tv_sec == 0;
// add txq
txq = tcp_txq_add(cb, hdr, sizeof(struct tcp_hdr) + len);
if (!txq) {
free(segment);
return -1;
}
// flow control
// TODO: cb->txq.snt is not exactly the sliding window size
if (have_unsent || (!TCP_FLG_ISSET(hdr->flg, TCP_FLG_SYN) &&
cb->txq.snt + len > cb->snd.wnd)) {
// before established snd.wnd is 0. so SYN does not care snd.wnd
// segment is stored in txq and send later in timer_thread
fprintf(stderr, ">>> not send havesent %d %d<<<\n", have_unsent,
cb->txq.tail);
tcp_dump(cb, hdr, len);
return 0;
}
}
// calculate checksum
self = ((struct netif_ip *)cb->iface)->unicast;
peer = cb->peer.addr;
hdr->sum =
tcp_checksum(self, peer, (uint8_t *)hdr, sizeof(struct tcp_hdr) + len);
#ifdef TCP_DEBUG
fprintf(stderr, ">>> tcp_tx <<<\n");
tcp_dump(cb, hdr, len);
#endif
// send packet
if (ip_tx(cb->iface, IP_PROTOCOL_TCP, (uint8_t *)hdr,
sizeof(struct tcp_hdr) + len, &peer) == -1) {
// failed to send ip packet
if (!txq) {
// this packet does not expect reply so not queued into txq list
free(segment);
}
return -1;
}
if (txq) {
// set timestamp
txq->timestamp = *now;
cb->txq.snt += len;
} else {
// this packet does not expect reply so not queued into txq list
free(segment);
}
return len;
}
static void tcp_rx(uint8_t *segment, size_t len, ip_addr_t *src, ip_addr_t *dst,
struct netif *iface) {
struct tcp_hdr *hdr;
int i;
struct tcp_cb *cb, *fcb = NULL, *lcb = NULL;
// validate tcp packet
if (*dst != ((struct netif_ip *)iface)->unicast) {
return;
}
if (len < sizeof(struct tcp_hdr)) {
return;
}
// validate checksum
hdr = (struct tcp_hdr *)segment;
if (tcp_checksum(*src, *dst, segment, len) != 0) {
fprintf(stderr, "tcp checksum error\n");
return;
}
pthread_mutex_lock(&mutex);
// find connection cb or listener cb
for (i = 0; i < TCP_CB_TABLE_SIZE; i++) {
cb = &cb_table[i];
if (IS_FREE_CB(cb)) {
// cache cb for the case SYN message and no cb is connected to this peer
if (fcb == NULL) {
fcb = cb;
}
} else if ((!cb->iface || cb->iface == iface) && cb->port == hdr->dst) {
// if ip address and dst port number matches
if (cb->peer.addr == *src && cb->peer.port == hdr->src) {
// this cb is connection for this tcp packet
break;
} else if (cb->state == TCP_CB_STATE_LISTEN && !lcb) {
// listener socket is found
lcb = cb;
}
}
}
// cb that matches this tcp packet is not found.
// create socket if listener socket exists and packet is SYN packet.
if (i == TCP_CB_TABLE_SIZE) {
if (!fcb) {
// cb resource is run out
// TODO: send RST
pthread_mutex_unlock(&mutex);
return;
}
// create accept socket
cb = fcb;
cb->iface = iface;
if (lcb) {
// TODO: ? if SYN is not set ?
cb->state = lcb->state;
cb->port = lcb->port;
cb->parent = lcb;
} else {
// this port is not listened.
// this packet is invalid. no connection is found.
cb->port = 0;
}
cb->peer.addr = *src;
cb->peer.port = hdr->src;
}
// else cb that matches this tcp packet is found.
#ifdef TCP_DEBUG
fprintf(stderr, ">>> tcp_rx <<<\n");
tcp_dump(cb, hdr, TCP_DATA_LEN(hdr, len));
#endif
// handle message
tcp_event_segment_arrives(cb, hdr, len);
pthread_mutex_unlock(&mutex);
return;
}
static void *tcp_timer_thread(void *arg) {
struct timeval timestamp, diff;
struct timespec timeout;
struct tcp_cb *cb;
struct tcp_txq_entry *txq, *prev, *tmp;
ip_addr_t self, peer;
size_t sum = 0;
int i;
diff.tv_sec = 0;
diff.tv_usec = 100 * 1000;
pthread_mutex_lock(&mutex);
while (1) {
gettimeofday(×tamp, NULL);
for (i = 0; i < TCP_CB_TABLE_SIZE; i++) {
cb = &cb_table[i];
if (cb->state == TCP_CB_STATE_CLOSED) {
// skip check timeout
continue;
}
// check user timeout
if ((cb->snd.una != cb->snd.nxt || cb->state == TCP_CB_STATE_TIME_WAIT) &&
cb->timeout < timestamp.tv_sec) {
// force close connection because of ack timeout or TIME_WAIT timeout
#ifdef TCP_DEBUG
fprintf(stderr, ">>> find user timeout (%ld - %ld) <<<\n",
timestamp.tv_sec, cb->timeout);
tcp_state_dump(cb);
#endif
tcp_close_cb(cb);
pthread_cond_broadcast(&cb->cond);
continue;
}
// check retransmission timeout and vacuum acked segment entry
prev = NULL;
txq = cb->txq.head;
sum = 0;
if (cb->txq.head) {
self = ((struct netif_ip *)cb->iface)->unicast;
peer = cb->peer.addr;
}
while (txq) {
if (ntoh32(txq->segment->seq) >= cb->snd.una) {
// TODO: check sum + datalen should compare with snd.wnd. but
// txq->segment does not support splitting. so sum add later
/* sum += TCP_DATA_LEN(txq->segment, txq->len); */
if (sum < cb->snd.wnd) {
// this txq is in sliding send window
if (txq->timestamp.tv_sec == 0) {
// this txq is not sent
// re-calc checksum
txq->segment->ack = hton32(cb->rcv.nxt);
txq->segment->sum = 0;
txq->segment->sum =
tcp_checksum(self, peer, txq->segment, txq->len);
#ifdef TCP_DEBUG
fprintf(stderr, ">>> tcp_tx in timer_thread <<<\n");
tcp_dump(cb, txq->segment, TCP_DATA_LEN(txq->segment, txq->len));
#endif
// send packet
ip_tx(cb->iface, IP_PROTOCOL_TCP, (uint8_t *)txq->segment,
txq->len, &cb->peer.addr);
txq->timestamp = timestamp;
} else if (timestamp.tv_sec - txq->timestamp.tv_sec > 3) {
// if retransmission timeout (3 seconds) then resend
// re-calc checksum
txq->segment->ack = hton32(cb->rcv.nxt);
txq->segment->sum = 0;
txq->segment->sum =
tcp_checksum(self, peer, txq->segment, txq->len);
#ifdef TCP_DEBUG
fprintf(stderr,
">>> find retransmission timeout (%ld - %ld) <<<\n",
timestamp.tv_sec, txq->timestamp.tv_sec);
tcp_dump(cb, txq->segment, TCP_DATA_LEN(txq->segment, txq->len));
#endif
// resend packet
ip_tx(cb->iface, IP_PROTOCOL_TCP, (uint8_t *)txq->segment,
txq->len, &cb->peer.addr);
txq->timestamp = timestamp;
}
}
// FIXME: temporary sum add here. no support splitting
sum += TCP_DATA_LEN(txq->segment, txq->len);
// update previous tcp_txq_entry
prev = txq;
txq = txq->next;
} else {
// remove tcp_txq_entry from list
cb->txq.snt -= TCP_DATA_LEN(txq->segment, txq->len);