-
Notifications
You must be signed in to change notification settings - Fork 24
/
tcp_output.c
1716 lines (1605 loc) · 47.4 KB
/
tcp_output.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
/* tcp_output.c,v 1.161.4.2 2008/01/09 01:57:30 matt Exp */
/*
* Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project.
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. Neither the name of the project nor the names of its contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*/
/*
* @(#)COPYRIGHT 1.1 (NRL) 17 January 1995
*
* NRL grants permission for redistribution and use in source and binary
* forms, with or without modification, of the software and documentation
* created at NRL provided that the following conditions are met:
*
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. All advertising materials mentioning features or use of this software
* must display the following acknowledgements:
* This product includes software developed by the University of
* California, Berkeley and its contributors.
* This product includes software developed at the Information
* Technology Division, US Naval Research Laboratory.
* 4. Neither the name of the NRL nor the names of its contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* THE SOFTWARE PROVIDED BY NRL IS PROVIDED BY NRL AND CONTRIBUTORS ``AS
* IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
* TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A
* PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL NRL OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
* LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
* NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
* SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*
* The views and conclusions contained in the software and documentation
* are those of the authors and should not be interpreted as representing
* official policies, either expressed or implied, of the US Naval
* Research Laboratory (NRL).
*/
/*-
* Copyright (c) 1997, 1998, 2001, 2005, 2006 The NetBSD Foundation, Inc.
* All rights reserved.
*
* This code is derived from software contributed to The NetBSD Foundation
* by Jason R. Thorpe and Kevin M. Lahey of the Numerical Aerospace Simulation
* Facility, NASA Ames Research Center.
* This code is derived from software contributed to The NetBSD Foundation
* by Charles M. Hannum.
* This code is derived from software contributed to The NetBSD Foundation
* by Rui Paulo.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. All advertising materials mentioning features or use of this software
* must display the following acknowledgement:
* This product includes software developed by the NetBSD
* Foundation, Inc. and its contributors.
* 4. Neither the name of The NetBSD Foundation nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS
* ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
* TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS
* BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
* CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
* SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
* INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
* CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
* ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
* POSSIBILITY OF SUCH DAMAGE.
*/
/*
* Copyright (c) 1982, 1986, 1988, 1990, 1993, 1995
* The Regents of the University of California. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. Neither the name of the University nor the names of its contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
* @(#)tcp_output.c 8.4 (Berkeley) 5/24/95
*/
#include <sys/cdefs.h>
__KERNEL_RCSID(0, "tcp_output.c,v 1.161.4.2 2008/01/09 01:57:30 matt Exp");
#include "opt_inet.h"
#include "opt_ipsec.h"
#include "opt_tcp_debug.h"
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/malloc.h>
#include <sys/mbuf.h>
#include <sys/protosw.h>
#include <sys/socket.h>
#include <sys/socketvar.h>
#include <sys/errno.h>
#include <sys/domain.h>
#include <sys/kernel.h>
#ifdef TCP_SIGNATURE
#include <sys/md5.h>
#endif
#include <net/if.h>
#include <net/route.h>
#include <netinet/in.h>
#include <netinet/in_systm.h>
#include <netinet/ip.h>
#include <netinet/in_pcb.h>
#include <netinet/ip_var.h>
#ifdef INET6
#ifndef INET
#include <netinet/in.h>
#endif
#include <netinet/ip6.h>
#include <netinet6/in6_var.h>
#include <netinet6/ip6_var.h>
#include <netinet6/in6_pcb.h>
#include <netinet6/nd6.h>
#endif
#ifdef FAST_IPSEC
#include <netipsec/ipsec.h>
#include <netipsec/key.h>
#ifdef INET6
#include <netipsec/ipsec6.h>
#endif
#endif /* FAST_IPSEC*/
#ifdef IPSEC
#include <netinet6/ipsec.h>
#endif
#include <netinet/tcp.h>
#define TCPOUTFLAGS
#include <netinet/tcp_fsm.h>
#include <netinet/tcp_seq.h>
#include <netinet/tcp_timer.h>
#include <netinet/tcp_var.h>
#include <netinet/tcp_congctl.h>
#include <netinet/tcpip.h>
#include <netinet/tcp_debug.h>
#include <netinet/in_offload.h>
#include <netinet6/in6_offload.h>
#ifdef IPSEC
#include <netkey/key.h>
#endif
#ifdef notyet
extern struct mbuf *m_copypack();
#endif
/*
* Knob to enable Congestion Window Monitoring, and control
* the burst size it allows. Default burst is 4 packets, per
* the Internet draft.
*/
int tcp_cwm = 0;
int tcp_cwm_burstsize = 4;
int tcp_do_autosndbuf = 0;
int tcp_autosndbuf_inc = 8 * 1024;
int tcp_autosndbuf_max = 256 * 1024;
#ifdef TCP_OUTPUT_COUNTERS
#include <sys/device.h>
extern struct evcnt tcp_output_bigheader;
extern struct evcnt tcp_output_predict_hit;
extern struct evcnt tcp_output_predict_miss;
extern struct evcnt tcp_output_copysmall;
extern struct evcnt tcp_output_copybig;
extern struct evcnt tcp_output_refbig;
#define TCP_OUTPUT_COUNTER_INCR(ev) (ev)->ev_count++
#else
#define TCP_OUTPUT_COUNTER_INCR(ev) /* nothing */
#endif /* TCP_OUTPUT_COUNTERS */
static
#ifndef GPROF
inline
#endif
int
tcp_segsize(struct tcpcb *tp, int *txsegsizep, int *rxsegsizep,
bool *alwaysfragp)
{
#ifdef INET
struct inpcb *inp = tp->t_inpcb;
#endif
#ifdef INET6
struct in6pcb *in6p = tp->t_in6pcb;
#endif
struct socket *so = NULL;
struct rtentry *rt;
struct ifnet *ifp;
int size;
int hdrlen;
int optlen;
*alwaysfragp = false;
#ifdef DIAGNOSTIC
if (tp->t_inpcb && tp->t_in6pcb)
panic("tcp_segsize: both t_inpcb and t_in6pcb are set");
#endif
switch (tp->t_family) {
#ifdef INET
case AF_INET:
hdrlen = sizeof(struct ip) + sizeof(struct tcphdr);
break;
#endif
#ifdef INET6
case AF_INET6:
hdrlen = sizeof(struct ip6_hdr) + sizeof(struct tcphdr);
break;
#endif
default:
size = tcp_mssdflt;
goto out;
}
rt = NULL;
#ifdef INET
if (inp) {
rt = in_pcbrtentry(inp);
so = inp->inp_socket;
}
#endif
#ifdef INET6
if (in6p) {
rt = in6_pcbrtentry(in6p);
so = in6p->in6p_socket;
}
#endif
if (rt == NULL) {
size = tcp_mssdflt;
goto out;
}
ifp = rt->rt_ifp;
size = tcp_mssdflt;
if (tp->t_mtudisc && rt->rt_rmx.rmx_mtu != 0) {
#ifdef INET6
if (in6p && rt->rt_rmx.rmx_mtu < IPV6_MMTU) {
/*
* RFC2460 section 5, last paragraph: if path MTU is
* smaller than 1280, use 1280 as packet size and
* attach fragment header.
*/
size = IPV6_MMTU - hdrlen - sizeof(struct ip6_frag);
*alwaysfragp = true;
} else
size = rt->rt_rmx.rmx_mtu - hdrlen;
#else
size = rt->rt_rmx.rmx_mtu - hdrlen;
#endif
} else if (ifp->if_flags & IFF_LOOPBACK)
size = ifp->if_mtu - hdrlen;
#ifdef INET
else if (inp && tp->t_mtudisc)
size = ifp->if_mtu - hdrlen;
else if (inp && in_localaddr(inp->inp_faddr))
size = ifp->if_mtu - hdrlen;
#endif
#ifdef INET6
else if (in6p) {
#ifdef INET
if (IN6_IS_ADDR_V4MAPPED(&in6p->in6p_faddr)) {
/* mapped addr case */
struct in_addr d;
bcopy(&in6p->in6p_faddr.s6_addr32[3], &d, sizeof(d));
if (tp->t_mtudisc || in_localaddr(d))
size = ifp->if_mtu - hdrlen;
} else
#endif
{
/*
* for IPv6, path MTU discovery is always turned on,
* or the node must use packet size <= 1280.
*/
size = tp->t_mtudisc ? IN6_LINKMTU(ifp) : IPV6_MMTU;
size -= hdrlen;
}
}
#endif
out:
/*
* Now we must make room for whatever extra TCP/IP options are in
* the packet.
*/
optlen = tcp_optlen(tp);
/*
* XXX tp->t_ourmss should have the right size, but without this code
* fragmentation will occur... need more investigation
*/
#ifdef INET
if (inp) {
#if defined(IPSEC) || defined(FAST_IPSEC)
if (! IPSEC_PCB_SKIP_IPSEC(inp->inp_sp, IPSEC_DIR_OUTBOUND))
optlen += ipsec4_hdrsiz_tcp(tp);
#endif
optlen += ip_optlen(inp);
}
#endif
#ifdef INET6
#ifdef INET
if (in6p && tp->t_family == AF_INET) {
#if defined(IPSEC) || defined(FAST_IPSEC)
if (! IPSEC_PCB_SKIP_IPSEC(in6p->in6p_sp, IPSEC_DIR_OUTBOUND))
optlen += ipsec4_hdrsiz_tcp(tp);
#endif
/* XXX size -= ip_optlen(in6p); */
} else
#endif
if (in6p && tp->t_family == AF_INET6) {
#if defined(IPSEC) || defined(FAST_IPSEC)
if (! IPSEC_PCB_SKIP_IPSEC(in6p->in6p_sp, IPSEC_DIR_OUTBOUND))
optlen += ipsec6_hdrsiz_tcp(tp);
#endif
optlen += ip6_optlen(in6p);
}
#endif
size -= optlen;
/* there may not be any room for data if mtu is too small */
if (size < 0)
return (EMSGSIZE);
/*
* *rxsegsizep holds *estimated* inbound segment size (estimation
* assumes that path MTU is the same for both ways). this is only
* for silly window avoidance, do not use the value for other purposes.
*
* ipseclen is subtracted from both sides, this may not be right.
* I'm not quite sure about this (could someone comment).
*/
*txsegsizep = min(tp->t_peermss - optlen, size);
/*
* Never send more than half a buffer full. This insures that we can
* always keep 2 packets on the wire, no matter what SO_SNDBUF is, and
* therefore acks will never be delayed unless we run out of data to
* transmit.
*/
if (so)
*txsegsizep = min(so->so_snd.sb_hiwat >> 1, *txsegsizep);
*rxsegsizep = min(tp->t_ourmss - optlen, size);
if (*txsegsizep != tp->t_segsz) {
/*
* If the new segment size is larger, we don't want to
* mess up the congestion window, but if it is smaller
* we'll have to reduce the congestion window to ensure
* that we don't get into trouble with initial windows
* and the rest. In any case, if the segment size
* has changed, chances are the path has, too, and
* our congestion window will be different.
*/
if (*txsegsizep < tp->t_segsz) {
tp->snd_cwnd = max((tp->snd_cwnd / tp->t_segsz)
* *txsegsizep, *txsegsizep);
tp->snd_ssthresh = max((tp->snd_ssthresh / tp->t_segsz)
* *txsegsizep, *txsegsizep);
}
tp->t_segsz = *txsegsizep;
}
return (0);
}
static
#ifndef GPROF
inline
#endif
int
tcp_build_datapkt(struct tcpcb *tp, struct socket *so, int off,
long len, int hdrlen, struct mbuf **mp)
{
struct mbuf *m, *m0;
if (tp->t_force && len == 1)
tcpstat.tcps_sndprobe++;
else if (SEQ_LT(tp->snd_nxt, tp->snd_max)) {
tcpstat.tcps_sndrexmitpack++;
tcpstat.tcps_sndrexmitbyte += len;
} else {
tcpstat.tcps_sndpack++;
tcpstat.tcps_sndbyte += len;
}
#ifdef notyet
if ((m = m_copypack(so->so_snd.sb_mb, off,
(int)len, max_linkhdr + hdrlen)) == 0)
return (ENOBUFS);
/*
* m_copypack left space for our hdr; use it.
*/
m->m_len += hdrlen;
m->m_data -= hdrlen;
#else
MGETHDR(m, M_DONTWAIT, MT_HEADER);
if (__predict_false(m == NULL))
return (ENOBUFS);
MCLAIM(m, &tcp_tx_mowner);
/*
* XXX Because other code assumes headers will fit in
* XXX one header mbuf.
*
* (This code should almost *never* be run.)
*/
if (__predict_false((max_linkhdr + hdrlen) > MHLEN)) {
TCP_OUTPUT_COUNTER_INCR(&tcp_output_bigheader);
MCLGET(m, M_DONTWAIT);
if ((m->m_flags & M_EXT) == 0) {
m_freem(m);
return (ENOBUFS);
}
}
m->m_data += max_linkhdr;
m->m_len = hdrlen;
/*
* To avoid traversing the whole sb_mb chain for correct
* data to send, remember last sent mbuf, its offset and
* the sent size. When called the next time, see if the
* data to send is directly following the previous transfer.
* This is important for large TCP windows.
*/
if (off == 0 || tp->t_lastm == NULL ||
(tp->t_lastoff + tp->t_lastlen) != off) {
TCP_OUTPUT_COUNTER_INCR(&tcp_output_predict_miss);
/*
* Either a new packet or a retransmit.
* Start from the beginning.
*/
tp->t_lastm = so->so_snd.sb_mb;
tp->t_inoff = off;
} else {
TCP_OUTPUT_COUNTER_INCR(&tcp_output_predict_hit);
tp->t_inoff += tp->t_lastlen;
}
/* Traverse forward to next packet */
while (tp->t_inoff > 0) {
if (tp->t_lastm == NULL)
panic("tp->t_lastm == NULL");
if (tp->t_inoff < tp->t_lastm->m_len)
break;
tp->t_inoff -= tp->t_lastm->m_len;
tp->t_lastm = tp->t_lastm->m_next;
}
tp->t_lastoff = off;
tp->t_lastlen = len;
m0 = tp->t_lastm;
off = tp->t_inoff;
if (len <= M_TRAILINGSPACE(m)) {
m_copydata(m0, off, (int) len, mtod(m, char *) + hdrlen);
m->m_len += len;
TCP_OUTPUT_COUNTER_INCR(&tcp_output_copysmall);
} else {
m->m_next = m_copym(m0, off, (int) len, M_DONTWAIT);
if (m->m_next == NULL) {
m_freem(m);
return (ENOBUFS);
}
#ifdef TCP_OUTPUT_COUNTERS
if (m->m_next->m_flags & M_EXT)
TCP_OUTPUT_COUNTER_INCR(&tcp_output_refbig);
else
TCP_OUTPUT_COUNTER_INCR(&tcp_output_copybig);
#endif /* TCP_OUTPUT_COUNTERS */
}
#endif
*mp = m;
return (0);
}
/*
* Tcp output routine: figure out what should be sent and send it.
*/
int
tcp_output(struct tcpcb *tp)
{
struct rtentry *rt;
struct socket *so;
struct route *ro;
long len, win;
int off, flags, error;
struct mbuf *m;
struct ip *ip;
#ifdef INET6
struct ip6_hdr *ip6;
#endif
struct tcphdr *th;
u_char opt[MAX_TCPOPTLEN];
unsigned optlen, hdrlen, packetlen;
unsigned int sack_numblks;
int idle, sendalot, txsegsize, rxsegsize;
int txsegsize_nosack;
int maxburst = TCP_MAXBURST;
int af; /* address family on the wire */
int iphdrlen;
int has_tso4, has_tso6;
int has_tso, use_tso;
bool alwaysfrag;
int sack_rxmit;
int sack_bytes_rxmt;
struct sackhole *p;
#ifdef TCP_SIGNATURE
int sigoff = 0;
#endif
#ifdef DIAGNOSTIC
if (tp->t_inpcb && tp->t_in6pcb)
panic("tcp_output: both t_inpcb and t_in6pcb are set");
#endif
so = NULL;
ro = NULL;
if (tp->t_inpcb) {
so = tp->t_inpcb->inp_socket;
ro = &tp->t_inpcb->inp_route;
}
#ifdef INET6
else if (tp->t_in6pcb) {
so = tp->t_in6pcb->in6p_socket;
ro = &tp->t_in6pcb->in6p_route;
}
#endif
switch (af = tp->t_family) {
#ifdef INET
case AF_INET:
if (tp->t_inpcb)
break;
#ifdef INET6
/* mapped addr case */
if (tp->t_in6pcb)
break;
#endif
return (EINVAL);
#endif
#ifdef INET6
case AF_INET6:
if (tp->t_in6pcb)
break;
return (EINVAL);
#endif
default:
return (EAFNOSUPPORT);
}
if (tcp_segsize(tp, &txsegsize, &rxsegsize, &alwaysfrag))
return (EMSGSIZE);
idle = (tp->snd_max == tp->snd_una);
/*
* Determine if we can use TCP segmentation offload:
* - If we're using IPv4
* - If there is not an IPsec policy that prevents it
* - If the interface can do it
*/
has_tso4 = has_tso6 = false;
#if defined(INET)
has_tso4 = tp->t_inpcb != NULL &&
#if defined(IPSEC) || defined(FAST_IPSEC)
IPSEC_PCB_SKIP_IPSEC(tp->t_inpcb->inp_sp,
IPSEC_DIR_OUTBOUND) &&
#endif
(rt = rtcache_validate(&tp->t_inpcb->inp_route)) != NULL &&
(rt->rt_ifp->if_capenable & IFCAP_TSOv4) != 0;
#endif /* defined(INET) */
#if defined(INET6)
has_tso6 = tp->t_in6pcb != NULL &&
#if defined(IPSEC) || defined(FAST_IPSEC)
IPSEC_PCB_SKIP_IPSEC(tp->t_in6pcb->in6p_sp,
IPSEC_DIR_OUTBOUND) &&
#endif
(rt = rtcache_validate(&tp->t_in6pcb->in6p_route)) != NULL &&
(rt->rt_ifp->if_capenable & IFCAP_TSOv6) != 0;
#endif /* defined(INET6) */
has_tso = (has_tso4 || has_tso6) && !alwaysfrag;
/*
* Restart Window computation. From draft-floyd-incr-init-win-03:
*
* Optionally, a TCP MAY set the restart window to the
* minimum of the value used for the initial window and
* the current value of cwnd (in other words, using a
* larger value for the restart window should never increase
* the size of cwnd).
*/
if (tcp_cwm) {
/*
* Hughes/Touch/Heidemann Congestion Window Monitoring.
* Count the number of packets currently pending
* acknowledgement, and limit our congestion window
* to a pre-determined allowed burst size plus that count.
* This prevents bursting once all pending packets have
* been acknowledged (i.e. transmission is idle).
*
* XXX Link this to Initial Window?
*/
tp->snd_cwnd = min(tp->snd_cwnd,
(tcp_cwm_burstsize * txsegsize) +
(tp->snd_nxt - tp->snd_una));
} else {
if (idle && (tcp_now - tp->t_rcvtime) >= tp->t_rxtcur) {
/*
* We have been idle for "a while" and no acks are
* expected to clock out any data we send --
* slow start to get ack "clock" running again.
*/
int ss = tcp_init_win;
#ifdef INET
if (tp->t_inpcb &&
in_localaddr(tp->t_inpcb->inp_faddr))
ss = tcp_init_win_local;
#endif
#ifdef INET6
if (tp->t_in6pcb &&
in6_localaddr(&tp->t_in6pcb->in6p_faddr))
ss = tcp_init_win_local;
#endif
tp->snd_cwnd = min(tp->snd_cwnd,
TCP_INITIAL_WINDOW(ss, txsegsize));
}
}
txsegsize_nosack = txsegsize;
again:
use_tso = has_tso;
if ((tp->t_flags & (TF_ECN_SND_CWR|TF_ECN_SND_ECE)) != 0) {
/* don't duplicate CWR/ECE. */
use_tso = 0;
}
TCP_REASS_LOCK(tp);
sack_numblks = tcp_sack_numblks(tp);
if (sack_numblks) {
int sackoptlen;
sackoptlen = TCP_SACK_OPTLEN(sack_numblks);
if (sackoptlen > txsegsize_nosack) {
sack_numblks = 0; /* give up SACK */
txsegsize = txsegsize_nosack;
} else {
if ((tp->rcv_sack_flags & TCPSACK_HAVED) != 0) {
/* don't duplicate D-SACK. */
use_tso = 0;
}
txsegsize = txsegsize_nosack - sackoptlen;
}
} else {
txsegsize = txsegsize_nosack;
}
/*
* Determine length of data that should be transmitted, and
* flags that should be used. If there is some data or critical
* controls (SYN, RST) to send, then transmit; otherwise,
* investigate further.
*
* Readjust SACK information to avoid resending duplicate data.
*/
if (TCP_SACK_ENABLED(tp) && SEQ_LT(tp->snd_nxt, tp->snd_max))
tcp_sack_adjust(tp);
sendalot = 0;
off = tp->snd_nxt - tp->snd_una;
win = min(tp->snd_wnd, tp->snd_cwnd);
flags = tcp_outflags[tp->t_state];
/*
* Send any SACK-generated retransmissions. If we're explicitly trying
* to send out new data (when sendalot is 1), bypass this function.
* If we retransmit in fast recovery mode, decrement snd_cwnd, since
* we're replacing a (future) new transmission with a retransmission
* now, and we previously incremented snd_cwnd in tcp_input().
*/
/*
* Still in sack recovery , reset rxmit flag to zero.
*/
sack_rxmit = 0;
sack_bytes_rxmt = 0;
len = 0;
p = NULL;
do {
long cwin;
if (!TCP_SACK_ENABLED(tp))
break;
if (tp->t_partialacks < 0)
break;
p = tcp_sack_output(tp, &sack_bytes_rxmt);
if (p == NULL)
break;
cwin = min(tp->snd_wnd, tp->snd_cwnd) - sack_bytes_rxmt;
if (cwin < 0)
cwin = 0;
/* Do not retransmit SACK segments beyond snd_recover */
if (SEQ_GT(p->end, tp->snd_recover)) {
/*
* (At least) part of sack hole extends beyond
* snd_recover. Check to see if we can rexmit data
* for this hole.
*/
if (SEQ_GEQ(p->rxmit, tp->snd_recover)) {
/*
* Can't rexmit any more data for this hole.
* That data will be rexmitted in the next
* sack recovery episode, when snd_recover
* moves past p->rxmit.
*/
p = NULL;
break;
}
/* Can rexmit part of the current hole */
len = ((long)ulmin(cwin, tp->snd_recover - p->rxmit));
} else
len = ((long)ulmin(cwin, p->end - p->rxmit));
off = p->rxmit - tp->snd_una;
if (off + len > so->so_snd.sb_cc) {
/* 1 for TH_FIN */
KASSERT(off + len == so->so_snd.sb_cc + 1);
KASSERT(p->rxmit + len == tp->snd_max);
len = so->so_snd.sb_cc - off;
}
if (len > 0) {
sack_rxmit = 1;
sendalot = 1;
}
} while (/*CONSTCOND*/0);
/*
* If in persist timeout with window of 0, send 1 byte.
* Otherwise, if window is small but nonzero
* and timer expired, we will send what we can
* and go to transmit state.
*/
if (tp->t_force) {
if (win == 0) {
/*
* If we still have some data to send, then
* clear the FIN bit. Usually this would
* happen below when it realizes that we
* aren't sending all the data. However,
* if we have exactly 1 byte of unset data,
* then it won't clear the FIN bit below,
* and if we are in persist state, we wind
* up sending the packet without recording
* that we sent the FIN bit.
*
* We can't just blindly clear the FIN bit,
* because if we don't have any more data
* to send then the probe will be the FIN
* itself.
*/
if (off < so->so_snd.sb_cc)
flags &= ~TH_FIN;
win = 1;
} else {
TCP_TIMER_DISARM(tp, TCPT_PERSIST);
tp->t_rxtshift = 0;
}
}
if (sack_rxmit == 0) {
if (TCP_SACK_ENABLED(tp) && tp->t_partialacks >= 0) {
long cwin;
/*
* We are inside of a SACK recovery episode and are
* sending new data, having retransmitted all the
* data possible in the scoreboard.
*/
if (tp->snd_wnd < so->so_snd.sb_cc) {
len = tp->snd_wnd - off;
flags &= ~TH_FIN;
} else {
len = so->so_snd.sb_cc - off;
}
/*
* From FreeBSD:
* Don't remove this (len > 0) check !
* We explicitly check for len > 0 here (although it
* isn't really necessary), to work around a gcc
* optimization issue - to force gcc to compute
* len above. Without this check, the computation
* of len is bungled by the optimizer.
*/
if (len > 0) {
cwin = tp->snd_cwnd -
(tp->snd_nxt - tp->sack_newdata) -
sack_bytes_rxmt;
if (cwin < 0)
cwin = 0;
if (cwin < len) {
len = cwin;
flags &= ~TH_FIN;
}
}
} else if (win < so->so_snd.sb_cc) {
len = win - off;
flags &= ~TH_FIN;
} else {
len = so->so_snd.sb_cc - off;
}
}
if (len < 0) {
/*
* If FIN has been sent but not acked,
* but we haven't been called to retransmit,
* len will be -1. Otherwise, window shrank
* after we sent into it. If window shrank to 0,
* cancel pending retransmit, pull snd_nxt back
* to (closed) window, and set the persist timer
* if it isn't already going. If the window didn't
* close completely, just wait for an ACK.
*
* If we have a pending FIN, either it has already been
* transmitted or it is outside the window, so drop it.
* If the FIN has been transmitted, but this is not a
* retransmission, then len must be -1. Therefore we also
* prevent here the sending of `gratuitous FINs'. This
* eliminates the need to check for that case below (e.g.
* to back up snd_nxt before the FIN so that the sequence
* number is correct).
*/
len = 0;
flags &= ~TH_FIN;
if (win == 0) {
TCP_TIMER_DISARM(tp, TCPT_REXMT);
tp->t_rxtshift = 0;
tp->snd_nxt = tp->snd_una;
if (TCP_TIMER_ISARMED(tp, TCPT_PERSIST) == 0)
tcp_setpersist(tp);
}
}
/*
* Automatic sizing enables the performance of large buffers
* and most of the efficiency of small ones by only allocating
* space when it is needed.
*
* The criteria to step up the send buffer one notch are:
* 1. receive window of remote host is larger than send buffer
* (with a fudge factor of 5/4th);
* 2. send buffer is filled to 7/8th with data (so we actually
* have data to make use of it);
* 3. send buffer fill has not hit maximal automatic size;
* 4. our send window (slow start and cogestion controlled) is
* larger than sent but unacknowledged data in send buffer.
*
* The remote host receive window scaling factor may limit the
* growing of the send buffer before it reaches its allowed
* maximum.
*
* It scales directly with slow start or congestion window
* and does at most one step per received ACK. This fast
* scaling has the drawback of growing the send buffer beyond
* what is strictly necessary to make full use of a given
* delay*bandwith product. However testing has shown this not
* to be much of an problem. At worst we are trading wasting
* of available bandwith (the non-use of it) for wasting some
* socket buffer memory.
*
* TODO: Shrink send buffer during idle periods together
* with congestion window. Requires another timer.
*/
if (tcp_do_autosndbuf && so->so_snd.sb_flags & SB_AUTOSIZE) {
if ((tp->snd_wnd / 4 * 5) >= so->so_snd.sb_hiwat &&
so->so_snd.sb_cc >= (so->so_snd.sb_hiwat / 8 * 7) &&
so->so_snd.sb_cc < tcp_autosndbuf_max &&
win >= (so->so_snd.sb_cc - (tp->snd_nxt - tp->snd_una))) {
if (!sbreserve(&so->so_snd,
min(so->so_snd.sb_hiwat + tcp_autosndbuf_inc,
tcp_autosndbuf_max), so))
so->so_snd.sb_flags &= ~SB_AUTOSIZE;
}
}
if (len > txsegsize) {
if (use_tso) {
/*
* Truncate TSO transfers to IP_MAXPACKET, and make
* sure that we send equal size transfers down the
* stack (rather than big-small-big-small-...).
*/
#ifdef INET6
#if IPV6_MAXPACKET != IP_MAXPACKET
#error IPV6_MAXPACKET != IP_MAXPACKET
#endif
#endif
len = (min(len, IP_MAXPACKET) / txsegsize) * txsegsize;
if (len <= txsegsize) {
use_tso = 0;
}
} else
len = txsegsize;
flags &= ~TH_FIN;
sendalot = 1;
} else
use_tso = 0;
if (sack_rxmit) {
if (SEQ_LT(p->rxmit + len, tp->snd_una + so->so_snd.sb_cc))
flags &= ~TH_FIN;
}
win = sbspace(&so->so_rcv);
/*
* Sender silly window avoidance. If connection is idle
* and can send all data, a maximum segment,
* at least a maximum default-size segment do it,
* or are forced, do it; otherwise don't bother.
* If peer's buffer is tiny, then send
* when window is at least half open.
* If retransmitting (possibly after persist timer forced us
* to send into a small window), then must resend.
*/
if (len) {
if (len >= txsegsize)
goto send;
if ((so->so_state & SS_MORETOCOME) == 0 &&
((idle || tp->t_flags & TF_NODELAY) &&
len + off >= so->so_snd.sb_cc))
goto send;