summaryrefslogtreecommitdiffstats
path: root/src/net/tcp.c
blob: 7f2f21663a855b86604e9eaf09a11880c26e3d8d (plain) (blame)
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
#include <stdio.h>
#include <string.h>
#include <stdlib.h>
#include <assert.h>
#include <byteswap.h>
#include <latch.h>
#include <errno.h>
#include <gpxe/process.h>
#include <gpxe/init.h>
#include <gpxe/netdevice.h>
#include <gpxe/pkbuff.h>
#include <gpxe/ip.h>
#include <gpxe/tcp.h>
#include <gpxe/tcpip.h>
#include <gpxe/retry.h>
#include "uip/uip.h"

/** @file
 *
 * TCP protocol
 *
 * The gPXE TCP stack is currently implemented on top of the uIP
 * protocol stack.  This file provides wrappers around uIP so that
 * higher-level protocol implementations do not need to talk directly
 * to uIP (which has a somewhat baroque API).
 *
 * Basic operation is to create a #tcp_connection structure, call
 * tcp_connect() and then call run_tcpip() in a loop until the
 * operation has completed.  The TCP stack will call the various
 * methods defined in the #tcp_operations structure in order to send
 * and receive data.
 *
 * See hello.c for a trivial example of a TCP protocol using this
 * API.
 *
 */

#if USE_UIP

/**
 * TCP transmit buffer
 *
 * When a tcp_operations::senddata() method is called, it is
 * guaranteed to be able to use this buffer as temporary space for
 * constructing the data to be sent.  For example, code such as
 *
 * @code
 *
 *     static void my_senddata ( struct tcp_connection *conn, void *buf,
 *				 size_t len ) {
 *         len = snprintf ( buf, len, "FETCH %s\r\n", filename );
 *         tcp_send ( conn, buf + already_sent, len - already_sent );
 *     }
 *
 * @endcode
 *
 * is allowed, and is probably the best way to deal with
 * variably-sized data.
 *
 * Note that you cannot use this simple mechanism if you want to be
 * able to construct single data blocks of more than #len bytes.
 */
static void *tcp_buffer = uip_buf + ( 40 + UIP_LLH_LEN );

/** Size of #tcp_buffer */
static size_t tcp_buflen = UIP_BUFSIZE - ( 40 + UIP_LLH_LEN );

/**
 * Open a TCP connection
 *
 * @v conn	TCP connection
 * 
 * This sets up a new TCP connection to the remote host specified in
 * tcp_connection::sin.
 */
void tcp_connect ( struct tcp_connection *conn ) {
	struct uip_conn *uip_conn;
	u16_t ipaddr[2];

	assert ( conn->sin.sin_addr.s_addr != 0 );
	assert ( conn->sin.sin_port != 0 );
	assert ( conn->tcp_op != NULL );
	assert ( sizeof ( uip_conn->appstate ) == sizeof ( conn ) );

	* ( ( uint32_t * ) ipaddr ) = conn->sin.sin_addr.s_addr;
	uip_conn = uip_connect ( ipaddr, conn->sin.sin_port );
#warning "Use linked lists so that uip_connect() cannot fail"
	assert ( uip_conn != NULL );
	*( ( void ** ) uip_conn->appstate ) = conn;
}

/**
 * Send data via a TCP connection
 *
 * @v conn	TCP connection
 * @v data	Data to send
 * @v len	Length of data
 *
 * Data will be automatically limited to the current TCP window size.
 *
 * If retransmission is required, the connection's
 * tcp_operations::senddata() method will be called again in order to
 * regenerate the data.
 */
void tcp_send ( struct tcp_connection *conn __unused,
		const void *data, size_t len ) {

	assert ( conn = *( ( void ** ) uip_conn->appstate ) );

	if ( len > tcp_buflen )
		len = tcp_buflen;
	memmove ( tcp_buffer, data, len );

	uip_send ( tcp_buffer, len );
}

/**
 * Close a TCP connection
 *
 * @v conn	TCP connection
 */
void tcp_close ( struct tcp_connection *conn __unused ) {
	assert ( conn = *( ( void ** ) uip_conn->appstate ) );
	uip_close();
}

/**
 * uIP TCP application call interface
 *
 * This is the entry point of gPXE from the point of view of the uIP
 * protocol stack.  This function calls the appropriate methods from
 * the connection's @tcp_operations table in order to process received
 * data, transmit new data etc.
 */
void uip_tcp_appcall ( void ) {
	struct tcp_connection *conn = *( ( void ** ) uip_conn->appstate );
	struct tcp_operations *op = conn->tcp_op;

	if ( op->closed ) {
		if ( uip_aborted() )
			op->closed ( conn, -ECONNABORTED );
		if ( uip_timedout() )
			op->closed ( conn, -ETIMEDOUT );
		if ( uip_closed() )
			op->closed ( conn, 0 );
	}
	if ( uip_connected() && op->connected )
		op->connected ( conn );
	if ( uip_acked() && op->acked )
		op->acked ( conn, uip_conn->len );
	if ( uip_newdata() && op->newdata )
		op->newdata ( conn, ( void * ) uip_appdata, uip_len );
	if ( ( uip_rexmit() || uip_newdata() || uip_acked() ||
	       uip_connected() || uip_poll() ) && op->senddata )
		op->senddata ( conn, tcp_buffer, tcp_buflen );
}

/* Present here to allow everything to link.  Will go into separate
 * udp.c file
 */
void uip_udp_appcall ( void ) {
}

/**
 * Perform periodic processing of all TCP connections
 *
 * This allows TCP connections to retransmit data if necessary.
 */
static void tcp_periodic ( void ) {
	struct pk_buff *pkb;
	int i;

	for ( i = 0 ; i < UIP_CONNS ; i++ ) {
		uip_periodic ( i );
		if ( uip_len > 0 ) {
			pkb = alloc_pkb ( uip_len + MAX_LL_HEADER_LEN);
			if ( ! pkb )
				continue;
				
			pkb_reserve ( pkb, MAX_LL_HEADER_LEN );
			pkb_put ( pkb, uip_len );
			memcpy ( pkb->data, uip_buf, uip_len );

			ipv4_uip_tx ( pkb );
		}
	}
}

/**
 * Kick a connection into life
 *
 * @v conn	TCP connection
 *
 * Call this function when you have new data to send and are not
 * already being called as part of TCP processing.
 */
void tcp_kick ( struct tcp_connection *conn __unused ) {
	/* Just kick all the connections; this will work for now */
	tcp_periodic();
}

/**
 * Single-step the TCP stack
 *
 * @v process	TCP process
 *
 * This calls tcp_periodic() at regular intervals.
 */
static void tcp_step ( struct process *process ) {
	static unsigned long timeout = 0;

	if ( currticks() > timeout ) {
		timeout = currticks() + ( TICKS_PER_SEC / 10 );
		tcp_periodic ();
	}

	schedule ( process );
}

/** TCP stack process */
static struct process tcp_process = {
	.step = tcp_step,
};

/** Initialise the TCP stack */
static void init_tcp ( void ) {
	schedule ( &tcp_process );
}

INIT_FN ( INIT_PROCESS, init_tcp, NULL, NULL );

#else

/**
 * List of registered TCP connections
 */
static LIST_HEAD ( tcp_conns );

/**
 * List of TCP states
 */
static const char *tcp_states[] = {
	"CLOSED",
	"LISTEN",
	"SYN_SENT",
	"SYN_RCVD",
	"ESTABLISHED",
	"FIN_WAIT_1",
	"FIN_WAIT_2",
	"CLOSING",
	"TIME_WAIT",
	"CLOSE_WAIT",
	"LAST_ACK",
	"INVALID" };

/**
 * TCP state transition function
 *
 * @v conn	TCP connection
 * @v nxt_state Next TCP state
 */
void tcp_set_flags ( struct tcp_connection *conn ) {

	/* Set the TCP flags */
	switch ( conn->tcp_state ) {
	case TCP_CLOSED:
		if ( conn->tcp_lstate == TCP_SYN_RCVD ) {
			conn->tcp_flags |= TCP_RST;
		}
		break;
	case TCP_LISTEN:
		break;
	case TCP_SYN_SENT:
		if ( conn->tcp_lstate == TCP_LISTEN ||
		     conn->tcp_lstate == TCP_CLOSED ) {
			conn->tcp_flags |= TCP_SYN;
		}
		break;
	case TCP_SYN_RCVD:
		if ( conn->tcp_lstate == TCP_LISTEN ||
		     conn->tcp_lstate == TCP_SYN_SENT ) {
			conn->tcp_flags |= ( TCP_SYN | TCP_ACK );
		}
		break;
	case TCP_ESTABLISHED:
		if ( conn->tcp_lstate == TCP_SYN_SENT ) {
			conn->tcp_flags |= TCP_ACK;
		}
		break;
	case TCP_FIN_WAIT_1:
		if ( conn->tcp_lstate == TCP_SYN_RCVD ||
		     conn->tcp_lstate == TCP_ESTABLISHED ) {
			conn->tcp_flags |= TCP_FIN;
		}
		break;
	case TCP_FIN_WAIT_2:
		break;
	case TCP_CLOSING:
		if ( conn->tcp_lstate == TCP_FIN_WAIT_1 ) {
			conn->tcp_flags |= TCP_ACK;
		}
		break;
	case TCP_TIME_WAIT:
		if ( conn->tcp_lstate == TCP_FIN_WAIT_1 ||
		     conn->tcp_lstate == TCP_FIN_WAIT_2 ) {
			conn->tcp_flags |= TCP_ACK;
		}
		break;
	case TCP_CLOSE_WAIT:
		if ( conn->tcp_lstate == TCP_ESTABLISHED ) {
			conn->tcp_flags |= TCP_ACK;
		}
		break;
	case TCP_LAST_ACK:
		if ( conn->tcp_lstate == TCP_CLOSE_WAIT ) {
			conn->tcp_flags |= TCP_FIN;
		}
		if ( conn->tcp_lstate == TCP_ESTABLISHED ) {
			conn->tcp_flags |= ( TCP_FIN | TCP_ACK );
		}
		break;
	default:
		DBG ( "TCP_INVALID state %d\n", conn->tcp_state );
		return;
	}
}

void tcp_trans ( struct tcp_connection *conn, int nxt_state ) {
	/* Remember the last state */
	conn->tcp_lstate = conn->tcp_state;
	conn->tcp_state = nxt_state;

	DBG ( "Transition from %s to %s\n", tcp_states[conn->tcp_lstate], tcp_states[conn->tcp_state] );

	/* TODO: Check if this check is required */
	if ( conn->tcp_lstate == conn->tcp_state || 
	     conn->tcp_state == TCP_INVALID ) {
		conn->tcp_flags = 0;
		return;
	}
	tcp_set_flags ( conn );
}

/**
 * Dump TCP header
 *
 * @v tcphdr	TCP header
 */
void tcp_dump ( struct tcp_header *tcphdr ) {
	DBG ( "TCP %p src:%d dest:%d seq:%lx ack:%lx hlen:%hd flags:%#hx\n",
		tcphdr, ntohs ( tcphdr->src ), ntohs ( tcphdr->dest ), ntohl ( tcphdr->seq ),
		ntohl ( tcphdr->ack ), ( ( tcphdr->hlen & TCP_MASK_HLEN ) / 16 ), ( tcphdr->flags & TCP_MASK_FLAGS ) );
}

/**
 * Initialize a TCP connection
 *
 * @v conn	TCP connection
 *
 * This function assigns initial values to some fields in the connection
 * structure. The application should call tcp_init_conn after creating a new
 * connection before calling any other "tcp_*" function.
 *
 * struct tcp_connection my_conn;
 * tcp_init_conn ( &my_conn );
 * ... 
 */
void tcp_init_conn ( struct tcp_connection *conn ) {
	conn->local_port = 0;
	conn->tcp_state = TCP_CLOSED;
	conn->tcp_lstate = TCP_INVALID;
	conn->tx_pkb = NULL;
	conn->tcp_op = NULL;
}

/** Retry timer
 *
 * @v timer	Retry timer
 * @v over	Failure indicator
 */
void tcp_expired ( struct retry_timer *timer, int over ) {
	struct tcp_connection *conn;
	conn = ( struct tcp_connection * ) container_of ( timer, 
					struct tcp_connection, timer );
	DBG ( "Timer expired in %s\n", tcp_states[conn->tcp_state] );
	switch ( conn->tcp_state ) {
	case TCP_SYN_SENT:
		if ( over ) {
			tcp_trans ( conn, TCP_CLOSED );
			DBG ( "Timeout! Connection closed\n" );
			return;
		}
		goto send_tcp_nomsg;
	case TCP_SYN_RCVD:
		if ( over ) {
			tcp_trans ( conn, TCP_CLOSED );
			goto send_tcp_nomsg;
		}
		goto send_tcp_nomsg;
	case TCP_ESTABLISHED:
		if ( conn->tcp_lstate == TCP_SYN_SENT ) {
			goto send_tcp_nomsg;
		}
		break;
	case TCP_CLOSE_WAIT:
		if ( conn->tcp_lstate == TCP_ESTABLISHED ) {
			goto send_tcp_nomsg;
		}
		break;
	case TCP_FIN_WAIT_1:
	case TCP_FIN_WAIT_2:
		goto send_tcp_nomsg;
	case TCP_CLOSING:
	case TCP_LAST_ACK:
		if ( conn->tcp_lstate == TCP_CLOSE_WAIT ) {
			goto send_tcp_nomsg;
		}
		return;
	case TCP_TIME_WAIT:
		tcp_trans ( conn, TCP_CLOSED );
		return;
	}
	/* Retransmit the data */
	tcp_set_flags ( conn );
	tcp_senddata ( conn );
	return;

  send_tcp_nomsg:
	free_pkb ( conn->tx_pkb );
	conn->tx_pkb = alloc_pkb ( MIN_PKB_LEN );
	pkb_reserve ( conn->tx_pkb, MAX_HDR_LEN );
	tcp_set_flags ( conn );
	int rc;
	if ( ( rc = tcp_send ( conn, TCP_NOMSG, TCP_NOMSG_LEN ) ) != 0 ) {
		DBG ( "Error sending TCP message (rc = %d)\n", rc );
	}
	return;
}

/**
 * Connect to a remote server
 *
 * @v conn	TCP connection
 * @v peer	Remote socket address
 *
 * This function initiates a TCP connection to the socket address specified in
 * peer. It sends a SYN packet to peer. When the connection is established, the
 * TCP stack calls the connected() callback function.
 */
int tcp_connectto ( struct tcp_connection *conn,
		    struct sockaddr_tcpip *peer ) {
	int rc;

	/* A connection can only be established from the CLOSED state */
	if ( conn->tcp_state != TCP_CLOSED ) {
		DBG ( "Error opening connection: Invalid state %s\n",
				tcp_states[conn->tcp_state] );
		return -EISCONN;
	}

	/* Add the connection to the set of listening connections */
	if ( ( rc = tcp_listen ( conn, conn->local_port ) ) != 0 ) {
		return rc;
	}
	memcpy ( &conn->peer, peer, sizeof ( conn->peer ) );

	/* Initialize the TCP timer */
	conn->timer.expired = tcp_expired;

	/* Send a SYN packet and transition to TCP_SYN_SENT */
	conn->snd_una = random();
	tcp_trans ( conn, TCP_SYN_SENT );
	/* Allocate space for the packet */
	free_pkb ( conn->tx_pkb );
	conn->tx_pkb = alloc_pkb ( MIN_PKB_LEN );
	pkb_reserve ( conn->tx_pkb, MAX_HDR_LEN );
	conn->rcv_win = MAX_PKB_LEN - MAX_HDR_LEN; /* TODO: Is this OK? */
	return tcp_send ( conn, TCP_NOMSG, TCP_NOMSG_LEN );
}

int tcp_connect ( struct tcp_connection *conn ) {
	return tcp_connectto ( conn, &conn->peer );
}

/**
 * Close the connection
 *
 * @v conn
 *
 * This function sends a FIN packet to the remote end of the connection. When
 * the remote end of the connection ACKs the FIN (FIN consumes one byte on the
 * snd stream), the stack invokes the closed() callback function.
 */
int tcp_close ( struct tcp_connection *conn ) {
	/* A connection can only be closed if it is a connected state */
	switch ( conn->tcp_state ) {
	case TCP_SYN_RCVD:
	case TCP_ESTABLISHED:
		tcp_trans ( conn, TCP_FIN_WAIT_1 );
		conn->tcp_op->closed ( conn, CONN_SNDCLOSE ); /* TODO: Check! */
		/* FIN consumes one byte on the snd stream */
//		conn->snd_una++;
		goto send_tcp_nomsg;
	case TCP_SYN_SENT:
	case TCP_LISTEN:
		/**
		 * Since the connection does not expect any packets from the
		 * remote end, it can be removed from the set of listening
		 * connections.
		 */
		list_del ( &conn->list );
		tcp_trans ( conn, TCP_CLOSED );
		conn->tcp_op->closed ( conn, CONN_SNDCLOSE );
		return 0;
	case TCP_CLOSE_WAIT:
		tcp_trans ( conn, TCP_LAST_ACK );
		conn->tcp_op->closed ( conn, CONN_SNDCLOSE ); /* TODO: Check! */
		/* FIN consumes one byte on the snd stream */
//		conn->snd_una++;
		goto send_tcp_nomsg;
	default:
		DBG ( "tcp_close(): Invalid state %s\n",
					tcp_states[conn->tcp_state] );
		return -EPROTO;
	}

  send_tcp_nomsg:
	free_pkb ( conn->tx_pkb );
	conn->tx_pkb = alloc_pkb ( MIN_PKB_LEN );
	conn->tcp_flags = TCP_FIN;
	pkb_reserve ( conn->tx_pkb, MAX_HDR_LEN );
	return tcp_send ( conn, TCP_NOMSG, TCP_NOMSG_LEN );
}

/**
 * Bind TCP connection to local port
 *
 * @v conn		TCP connection
 * @v local_port	Local port, in network byte order
 * @ret rc		Return status code
 */
int tcp_bind ( struct tcp_connection *conn, uint16_t local_port ) {
	struct tcp_connection *existing;

	list_for_each_entry ( existing, &tcp_conns, list ) {
		if ( existing->local_port == local_port )
			return -EADDRINUSE;
	}
	conn->local_port = local_port;
	return 0;
}


/**
 * Listen for a packet
 *
 * @v conn		TCP connection
 * @v local_port	Local port, in network byte order
 *
 * This function adds the connection to a list of registered tcp
 * connections. If the local port is 0, the connection is assigned an
 * available port between MIN_TCP_PORT and 65535.
 */
int tcp_listen ( struct tcp_connection *conn, uint16_t local_port ) {
	static uint16_t try_port = 1024;
	int rc;

#warning "Fix the port re-use bug"
	/* If we re-use the same port, the connection should be reset
	 * and a new connection set up.  This doesn't happen yet, so
	 * randomise the port to avoid hitting the problem.
	 */
	try_port = random();

	/* If no port specified, find the first available port */
	if ( ! local_port ) {
		for ( ; try_port ; try_port++ ) {
			if ( try_port < 1024 )
				continue;
			if ( tcp_listen ( conn, htons ( try_port ) ) == 0 )
				return 0;
		}
		return -EADDRINUSE;
	}

	/* Attempt bind to local port */
	if ( ( rc = tcp_bind ( conn, local_port ) ) != 0 )
		return rc;

	/* Add to TCP connection list */
	list_add ( &conn->list, &tcp_conns );
	DBG ( "TCP opened %p on port %d\n", conn, ntohs ( local_port ) );

	return 0;
}

/**
 * Send data
 *
 * @v conn	TCP connection
 * 
 * This function allocates space to the transmit buffer and invokes the
 * senddata() callback function. It passes the allocated buffer to senddata().
 * The applicaion may use this space to write it's data.
 */
int tcp_senddata ( struct tcp_connection *conn ) {
	/* The connection must be in a state in which the user can send data */
	switch ( conn->tcp_state ) {
	case TCP_LISTEN:
		tcp_trans ( conn, TCP_SYN_SENT );
		conn->snd_una = random();
		break;
	case TCP_ESTABLISHED:
	case TCP_CLOSE_WAIT:
		break;
	default:
		DBG ( "tcp_senddata: Invalid state %s\n",
				tcp_states[conn->tcp_state] );
		return -EPROTO;
	}

	/* Allocate space to the TX buffer */
	free_pkb ( conn->tx_pkb );
	conn->tx_pkb = alloc_pkb ( MAX_PKB_LEN );
	if ( !conn->tx_pkb ) {
		DBG ( "Insufficient memory\n" );
		return -ENOMEM;
	}
	pkb_reserve ( conn->tx_pkb, MAX_HDR_LEN );
	/* Set the advertised window */
	conn->rcv_win = pkb_available ( conn->tx_pkb );
	/* Call the senddata() call back function */
	conn->tcp_op->senddata ( conn, conn->tx_pkb->data, 
					pkb_available ( conn->tx_pkb ) );
	/* Send pure ACK if senddata() didn't call tcp_send() */
	if ( conn->tx_pkb ) {
		tcp_send ( conn, TCP_NOMSG, TCP_NOMSG_LEN );
	}
	return 0;
}

/**
 * Transmit data
 *
 * @v conn	TCP connection
 * @v data	Data to be sent
 * @v len	Length of the data
 *
 * This function sends data to the peer socket address
 */
int tcp_send ( struct tcp_connection *conn, const void *data, size_t len ) {
	struct sockaddr_tcpip *peer = &conn->peer;
	struct pk_buff *pkb;
	int slen;

	/* Take ownership of the TX buffer from the connection */
	pkb = conn->tx_pkb;
	conn->tx_pkb = NULL;

	/* Determine the amount of data to be sent */
	slen = len < conn->snd_win ? len : conn->snd_win;
	/* Copy payload */
	memmove ( pkb_put ( pkb, slen ), data, slen );

	/* Fill up the TCP header */
	struct tcp_header *tcphdr = pkb_push ( pkb, sizeof ( *tcphdr ) );

	/* Source port, assumed to be in network byte order in conn */
	tcphdr->src = conn->local_port;
	/* Destination port, assumed to be in network byte order in peer */
	tcphdr->dest = peer->st_port;
	tcphdr->seq = htonl ( conn->snd_una );
	tcphdr->ack = htonl ( conn->rcv_nxt );
	/* Header length, = 0x50 (without TCP options) */
	tcphdr->hlen = ( uint8_t ) ( ( sizeof ( *tcphdr ) / 4 ) << 4 );
	/* Copy TCP flags, and then reset the variable */
	tcphdr->flags = conn->tcp_flags;
	conn->tcp_flags = 0;
	/* Advertised window, in network byte order */
	tcphdr->win = htons ( conn->rcv_win );
	/* Set urgent pointer to 0 */
	tcphdr->urg = 0;
	/* Calculate and store partial checksum, in host byte order */
	tcphdr->csum = 0;
	tcphdr->csum = tcpip_chksum ( pkb->data, pkb_len ( pkb ) );
	
	/* Dump the TCP header */
	tcp_dump ( tcphdr );

	/* Start the timer */
	if ( ( conn->tcp_state == TCP_ESTABLISHED && conn->tcp_lstate == TCP_SYN_SENT ) ||
	     ( conn->tcp_state == TCP_LISTEN && conn->tcp_lstate == TCP_SYN_RCVD ) ||
	     ( conn->tcp_state == TCP_CLOSED && conn->tcp_lstate == TCP_SYN_RCVD ) ||
	     ( conn->tcp_state == TCP_ESTABLISHED && ( len == 0 ) ) ) {
		// Don't start the timer
	} else {
		start_timer ( &conn->timer );
	}

	/* Transmit packet */
	return tcpip_tx ( pkb, &tcp_protocol, peer );
}

/**
 * Process received packet
 *
 * @v pkb	Packet buffer
 * @v partial	Partial checksum
 */
static int tcp_rx ( struct pk_buff *pkb,
		    struct sockaddr_tcpip *st_src __unused,
		    struct sockaddr_tcpip *st_dest __unused ) {
	struct tcp_connection *conn;
	struct tcp_header *tcphdr;
	uint32_t acked, toack;
	int hlen;
	int rc;

	/* Sanity check */
	if ( pkb_len ( pkb ) < sizeof ( *tcphdr ) ) {
		DBG ( "Packet too short (%d bytes)\n", pkb_len ( pkb ) );
		rc = -EINVAL;
		goto done;
	}

	/* Process TCP header */
	tcphdr = pkb->data;
	tcp_dump ( tcphdr );

	/* Verify header length */
	hlen = ( ( tcphdr->hlen & TCP_MASK_HLEN ) / 16 ) * 4;
	if ( hlen < sizeof ( *tcphdr ) ) {
		DBG ( "Bad header length (%d bytes)\n", hlen );
		rc = -EINVAL;
		goto done;
	}
	/* TODO: Parse TCP options */
	if ( hlen != sizeof ( *tcphdr ) ) {
		DBG ( "Ignoring TCP options\n" );
	}

	/* TODO: Verify checksum */
	
	/* Demux TCP connection */
	list_for_each_entry ( conn, &tcp_conns, list ) {
		if ( tcphdr->dest == conn->local_port ) {
			goto found_conn;
		}
	}
	
	DBG ( "No connection found on port %d\n", ntohs ( tcphdr->dest ) );
	rc = 0;
	goto done;

  found_conn:
	/* Stop the timer */
	stop_timer ( &conn->timer );

	/* Set the advertised window */
	conn->snd_win = tcphdr->win;

	/* TCP State Machine */
	conn->tcp_lstate = conn->tcp_state;
	switch ( conn->tcp_state ) {
	case TCP_CLOSED:
		DBG ( "tcp_rx(): Invalid state %s\n",
				tcp_states[conn->tcp_state] );
		rc = -EINVAL;
		goto done;
	case TCP_LISTEN:
		if ( tcphdr->flags & TCP_SYN ) {
			tcp_trans ( conn, TCP_SYN_RCVD );
			/* Synchronize the sequence numbers */
			conn->rcv_nxt = ntohl ( tcphdr->seq ) + 1;
			conn->tcp_flags |= TCP_ACK;

			/* Set the sequence number for the snd stream */
			conn->snd_una = random();
			conn->tcp_flags |= TCP_SYN;

			/* Send a SYN,ACK packet */
			goto send_tcp_nomsg;
		}
		/* Unexpected packet */
		goto unexpected;
	case TCP_SYN_SENT:
		if ( tcphdr->flags & TCP_SYN ) {
			/* Synchronize the sequence number in rcv stream */
			conn->rcv_nxt = ntohl ( tcphdr->seq ) + 1;
			conn->tcp_flags |= TCP_ACK;

			if ( tcphdr->flags & TCP_ACK ) {
				tcp_trans ( conn, TCP_ESTABLISHED );
				/**
				 * Process ACK of SYN. This does not invoke the
				 * acked() callback function.
				 */
				conn->snd_una = ntohl ( tcphdr->ack );
				conn->tcp_op->connected ( conn );
				conn->tcp_flags |= TCP_ACK;
				tcp_senddata ( conn );
				rc = 0;
				goto done;
			} else {
				tcp_trans ( conn, TCP_SYN_RCVD );
				conn->tcp_flags |= TCP_SYN;
				goto send_tcp_nomsg;
			}
		}
		/* Unexpected packet */
		goto unexpected;
	case TCP_SYN_RCVD:
		if ( tcphdr->flags & TCP_RST ) {
			tcp_trans ( conn, TCP_LISTEN );
			conn->tcp_op->closed ( conn, CONN_RESTART );
			rc = 0;
			goto done;
		}
		if ( tcphdr->flags & TCP_ACK ) {
			tcp_trans ( conn, TCP_ESTABLISHED );
			/**
			 * Process ACK of SYN. It neither invokes the callback
			 * function nor does it send an ACK.
			 */
			conn->snd_una = tcphdr->ack - 1;
			conn->tcp_op->connected ( conn );
			rc = 0;
			goto done;
		}
		/* Unexpected packet */
		goto unexpected;
	case TCP_ESTABLISHED:
		if ( tcphdr->flags & TCP_FIN ) {
			if ( tcphdr->flags & TCP_ACK ) {
				tcp_trans ( conn, TCP_LAST_ACK );
				conn->tcp_flags |= TCP_FIN;
			} else {
				tcp_trans ( conn, TCP_CLOSE_WAIT );
			}
			/* FIN consumes one byte */
			conn->rcv_nxt++;
			conn->tcp_flags |= TCP_ACK;
			/* Send the packet */
			goto send_tcp_nomsg;
		}
		/* Packet might contain data */
		break;
	case TCP_FIN_WAIT_1:
		if ( tcphdr->flags & TCP_FIN ) {
			conn->rcv_nxt++;
			conn->tcp_flags |= TCP_ACK;
			conn->tcp_op->closed ( conn, CONN_SNDCLOSE );

			if ( tcphdr->flags & TCP_ACK ) {
				tcp_trans ( conn, TCP_TIME_WAIT );
			} else {
				tcp_trans ( conn, TCP_CLOSING );
			}
			/* Send an acknowledgement */
			goto send_tcp_nomsg;
		}
		if ( tcphdr->flags & TCP_ACK ) {
			tcp_trans ( conn, TCP_FIN_WAIT_2 );
		}
		/* Packet might contain data */
		break;
	case TCP_FIN_WAIT_2:
		if ( tcphdr->flags & TCP_FIN ) {
			tcp_trans ( conn, TCP_TIME_WAIT );
			/* FIN consumes one byte */
			conn->rcv_nxt++;
			conn->tcp_flags |= TCP_ACK;
			goto send_tcp_nomsg;
		}
		/* Packet might contain data */
		break;
	case TCP_CLOSING:
		if ( tcphdr->flags & TCP_ACK ) {
			tcp_trans ( conn, TCP_TIME_WAIT );
			start_timer ( &conn->timer );
			rc = 0;
			goto done;
		}
		/* Unexpected packet */
		goto unexpected;
	case TCP_TIME_WAIT:
		/* Unexpected packet */
		goto unexpected;
	case TCP_CLOSE_WAIT:
		/* Packet could acknowledge data */
		break;
	case TCP_LAST_ACK:
		if ( tcphdr->flags & TCP_ACK ) {
			tcp_trans ( conn, TCP_CLOSED );
			rc = 0;
			goto done;
		}
		/* Unexpected packet */
		goto unexpected;
	}

	/**
	 * Any packet reaching this point either contains new data or
	 * acknowledges previously transmitted data.
	 */
	assert ( ( tcphdr->flags & TCP_ACK ) ||
		 pkb_len ( pkb ) > sizeof ( *tcphdr ) );

	/**
	 * Check if the received packet ACKs sent data
	 */
	if ( tcphdr->flags & TCP_ACK ) {
		acked = ntohl ( tcphdr->ack ) - conn->snd_una;
		if ( acked < 0 ) {
			/* Packet ACKs previously ACKed data */
			DBG ( "Previously ACKed data %lx\n", 
						ntohl ( tcphdr->ack ) );
			rc = 0;
			goto done;
		}
		/* Invoke the acked() callback */
		conn->snd_una += acked;
		conn->tcp_op->acked ( conn, acked );
	}
	
	/**
	 * Check if packet contains new data
	 */
	toack = pkb_len ( pkb ) - hlen;
	if ( toack >= 0 ) {
		/* Check the sequence number */
		if ( conn->rcv_nxt == ntohl ( tcphdr->seq ) ) {
			conn->rcv_nxt += toack;
			conn->tcp_op->newdata ( conn,
						pkb->data + hlen, toack );
		} else {
			DBG ( "Unexpected sequence number %lx (wanted %lx)\n",
				ntohl ( tcphdr->ack ), conn->rcv_nxt );
		}
		conn->tcp_flags |= TCP_ACK;
	}
	
	/**
	 * Send data
	 */
	tcp_senddata ( conn );
	return 0;

  send_tcp_nomsg:
	free_pkb ( conn->tx_pkb );
	conn->tx_pkb = alloc_pkb ( MIN_PKB_LEN );
	pkb_reserve ( conn->tx_pkb, MAX_HDR_LEN );
	if ( ( rc = tcp_send ( conn, TCP_NOMSG, TCP_NOMSG_LEN ) ) != 0 ) {
		DBG ( "Error sending TCP message (rc = %d)\n", rc );
	}
	goto done;

  unexpected:
	DBG ( "Unexpected packet received in %s with flags = %#hx\n",
			tcp_states[conn->tcp_state], tcphdr->flags & TCP_MASK_FLAGS );
	tcp_close ( conn );
	free_pkb ( conn->tx_pkb );
	conn->tx_pkb = NULL;
	rc = -EINVAL;
	goto done;

 done:
	free_pkb ( pkb );
	return rc;
}

/** TCP protocol */
struct tcpip_protocol tcp_protocol __tcpip_protocol = {
	.name = "TCP",
	.rx = tcp_rx,
	.tcpip_proto = IP_TCP,
	.csum_offset = 16,
};

#endif /* USE_UIP */