forked from chronolaw/annotated_nginx
-
Notifications
You must be signed in to change notification settings - Fork 0
/
ngx_connection.c
1709 lines (1339 loc) · 48.3 KB
/
ngx_connection.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
// annotated by chrono since 2016
//
// * ngx_create_listening
// * ngx_open_listening_sockets
// * ngx_configure_listening_sockets
// * ngx_get_connection
// * ngx_close_connection
// * ngx_close_listening_sockets
//
// 1.10增加对reuseport的支持
/*
* Copyright (C) Igor Sysoev
* Copyright (C) Nginx, Inc.
*/
#include <ngx_config.h>
#include <ngx_core.h>
#include <ngx_event.h>
// os/unix/ngx_os.h
// 操作系统提供的底层数据收发接口
// ngx_posix_init.c里初始化为linux的底层接口
// 在epoll模块的ngx_epoll_init里设置
//
// typedef struct {
// ngx_recv_pt recv;
// ngx_recv_chain_pt recv_chain;
// ngx_recv_pt udp_recv;
// ngx_send_pt send;
// ngx_send_chain_pt send_chain;
// ngx_uint_t flags;
// } ngx_os_io_t;
ngx_os_io_t ngx_io;
// 检查最多32个在可复用连接队列里的元素
// 设置为连接关闭c->close = 1;
// 调用事件的处理函数，里面会检查c->close
// 这样就会调用ngx_http_close_connection
// 释放连接，加入空闲链表，可以再次使用
// 早期函数参数是void，1.12改成cycle
static void ngx_drain_connections(ngx_cycle_t *cycle);
// http/ngx_http.c:ngx_http_add_listening()里调用
// 添加到cycle的监听端口数组，只是添加，没有其他动作
// 添加后需要用ngx_open_listening_sockets()才能打开端口监听
ngx_listening_t *
ngx_create_listening(ngx_conf_t *cf, struct sockaddr *sockaddr,
socklen_t socklen)
{
size_t len;
ngx_listening_t *ls;
struct sockaddr *sa;
u_char text[NGX_SOCKADDR_STRLEN];
// 添加到cycle的监听端口数组
ls = ngx_array_push(&cf->cycle->listening);
if (ls == NULL) {
return NULL;
}
// 内存清零，这样ls->worker就是0
ngx_memzero(ls, sizeof(ngx_listening_t));
// 根据socklen分配sockaddr的内存空间
sa = ngx_palloc(cf->pool, socklen);
if (sa == NULL) {
return NULL;
}
ngx_memcpy(sa, sockaddr, socklen);
ls->sockaddr = sa;
ls->socklen = socklen;
// 准备addr的字符串形式
len = ngx_sock_ntop(sa, socklen, text, NGX_SOCKADDR_STRLEN, 1);
ls->addr_text.len = len;
switch (ls->sockaddr->sa_family) {
#if (NGX_HAVE_INET6)
case AF_INET6:
ls->addr_text_max_len = NGX_INET6_ADDRSTRLEN;
break;
#endif
#if (NGX_HAVE_UNIX_DOMAIN)
case AF_UNIX:
ls->addr_text_max_len = NGX_UNIX_ADDRSTRLEN;
len++;
break;
#endif
case AF_INET:
ls->addr_text_max_len = NGX_INET_ADDRSTRLEN;
break;
default:
ls->addr_text_max_len = NGX_SOCKADDR_STRLEN;
break;
}
ls->addr_text.data = ngx_pnalloc(cf->pool, len);
if (ls->addr_text.data == NULL) {
return NULL;
}
// 拷贝addr的字符串形式
ngx_memcpy(ls->addr_text.data, text, len);
#if !(NGX_WIN32)
ngx_rbtree_init(&ls->rbtree, &ls->sentinel, ngx_udp_rbtree_insert_value);
#endif
ls->fd = (ngx_socket_t) -1;
ls->type = SOCK_STREAM;
// os/unix/ngx_linux_config.h:#define NGX_LISTEN_BACKLOG 511
ls->backlog = NGX_LISTEN_BACKLOG;
ls->rcvbuf = -1;
ls->sndbuf = -1;
#if (NGX_HAVE_SETFIB)
ls->setfib = -1;
#endif
#if (NGX_HAVE_TCP_FASTOPEN)
ls->fastopen = -1;
#endif
return ls;
}
// 1.10新函数，专为reuseport使用
// 拷贝了worker数量个的监听结构体
// 在ngx_stream_optimize_servers等函数创建端口时调用
ngx_int_t
ngx_clone_listening(ngx_cycle_t *cycle, ngx_listening_t *ls)
{
// configure脚本可以检测系统是否支持reuseport
// 使用宏来控制条件编译
#if (NGX_HAVE_REUSEPORT)
ngx_int_t n;
ngx_core_conf_t *ccf;
ngx_listening_t ols;
// 监听指令需要配置了reuseport
if (!ls->reuseport || ls->worker != 0) {
return NGX_OK;
}
// ls在之前已经正确设置了若干参数
// 例如type/handler/backlog等等
ols = *ls;
ccf = (ngx_core_conf_t *) ngx_get_conf(cycle->conf_ctx, ngx_core_module);
// ccf->worker_processes是nginx的worker进程数
// 拷贝了worker数量个的监听结构体
//
// 注意从1开始，这样拷贝后总数就是worker数量个
// 最开始的一个就是被克隆的ls监听结构体
for (n = 1; n < ccf->worker_processes; n++) {
/* create a socket for each worker process */
ls = ngx_array_push(&cycle->listening);
if (ls == NULL) {
return NGX_ERROR;
}
// 完全拷贝结构体
*ls = ols;
// 设置worker的序号
// 被克隆的对象的worker是0
//
// worker的使用是在ngx_event.c:ngx_event_process_init
// 只有worker id是本worker的listen才会enable
ls->worker = n;
}
#endif
return NGX_OK;
}
// 根据传递过来的socket描述符，使用系统调用获取之前设置的参数
// 填入ngx_listeing_t结构体
ngx_int_t
ngx_set_inherited_sockets(ngx_cycle_t *cycle)
{
size_t len;
ngx_uint_t i;
ngx_listening_t *ls;
socklen_t olen;
#if (NGX_HAVE_DEFERRED_ACCEPT || NGX_HAVE_TCP_FASTOPEN)
ngx_err_t err;
#endif
#if (NGX_HAVE_DEFERRED_ACCEPT && defined SO_ACCEPTFILTER)
struct accept_filter_arg af;
#endif
#if (NGX_HAVE_DEFERRED_ACCEPT && defined TCP_DEFER_ACCEPT)
int timeout;
#endif
#if (NGX_HAVE_REUSEPORT)
int reuseport;
#endif
// 遍历监听端口数组，里面应该有之前传递过来的socket
ls = cycle->listening.elts;
for (i = 0; i < cycle->listening.nelts; i++) {
// 监听端口的实际地址，先分配内存
ls[i].sockaddr = ngx_palloc(cycle->pool, sizeof(ngx_sockaddr_t));
if (ls[i].sockaddr == NULL) {
return NGX_ERROR;
}
// 地址内存的长度
ls[i].socklen = sizeof(ngx_sockaddr_t);
// 系统调用，获取socket的地址
if (getsockname(ls[i].fd, ls[i].sockaddr, &ls[i].socklen) == -1) {
ngx_log_error(NGX_LOG_CRIT, cycle->log, ngx_socket_errno,
"getsockname() of the inherited "
"socket #%d failed", ls[i].fd);
ls[i].ignore = 1;
continue;
}
// 调整为正确的长度
if (ls[i].socklen > (socklen_t) sizeof(ngx_sockaddr_t)) {
ls[i].socklen = sizeof(ngx_sockaddr_t);
}
// 接下来计算地址的文本形式
switch (ls[i].sockaddr->sa_family) {
#if (NGX_HAVE_INET6)
case AF_INET6:
ls[i].addr_text_max_len = NGX_INET6_ADDRSTRLEN;
len = NGX_INET6_ADDRSTRLEN + sizeof("[]:65535") - 1;
break;
#endif
#if (NGX_HAVE_UNIX_DOMAIN)
case AF_UNIX:
ls[i].addr_text_max_len = NGX_UNIX_ADDRSTRLEN;
len = NGX_UNIX_ADDRSTRLEN;
break;
#endif
case AF_INET:
ls[i].addr_text_max_len = NGX_INET_ADDRSTRLEN;
len = NGX_INET_ADDRSTRLEN + sizeof(":65535") - 1;
break;
default:
ngx_log_error(NGX_LOG_CRIT, cycle->log, ngx_socket_errno,
"the inherited socket #%d has "
"an unsupported protocol family", ls[i].fd);
ls[i].ignore = 1;
continue;
}
// 分配文本的内存
ls[i].addr_text.data = ngx_pnalloc(cycle->pool, len);
if (ls[i].addr_text.data == NULL) {
return NGX_ERROR;
}
// socket地址转换为字符串
len = ngx_sock_ntop(ls[i].sockaddr, ls[i].socklen,
ls[i].addr_text.data, len, 1);
if (len == 0) {
return NGX_ERROR;
}
ls[i].addr_text.len = len;
ls[i].backlog = NGX_LISTEN_BACKLOG;
// 逐个获取socket的各个参数
olen = sizeof(int);
// type，即tcp或udp
if (getsockopt(ls[i].fd, SOL_SOCKET, SO_TYPE, (void *) &ls[i].type,
&olen)
== -1)
{
ngx_log_error(NGX_LOG_CRIT, cycle->log, ngx_socket_errno,
"getsockopt(SO_TYPE) %V failed", &ls[i].addr_text);
ls[i].ignore = 1;
continue;
}
olen = sizeof(int);
// rcvbuf
if (getsockopt(ls[i].fd, SOL_SOCKET, SO_RCVBUF, (void *) &ls[i].rcvbuf,
&olen)
== -1)
{
ngx_log_error(NGX_LOG_ALERT, cycle->log, ngx_socket_errno,
"getsockopt(SO_RCVBUF) %V failed, ignored",
&ls[i].addr_text);
ls[i].rcvbuf = -1;
}
olen = sizeof(int);
// sndbuf
if (getsockopt(ls[i].fd, SOL_SOCKET, SO_SNDBUF, (void *) &ls[i].sndbuf,
&olen)
== -1)
{
ngx_log_error(NGX_LOG_ALERT, cycle->log, ngx_socket_errno,
"getsockopt(SO_SNDBUF) %V failed, ignored",
&ls[i].addr_text);
ls[i].sndbuf = -1;
}
#if 0
/* SO_SETFIB is currently a set only option */
#if (NGX_HAVE_SETFIB)
olen = sizeof(int);
if (getsockopt(ls[i].fd, SOL_SOCKET, SO_SETFIB,
(void *) &ls[i].setfib, &olen)
== -1)
{
ngx_log_error(NGX_LOG_ALERT, cycle->log, ngx_socket_errno,
"getsockopt(SO_SETFIB) %V failed, ignored",
&ls[i].addr_text);
ls[i].setfib = -1;
}
#endif
#endif
#if (NGX_HAVE_REUSEPORT)
reuseport = 0;
olen = sizeof(int);
#ifdef SO_REUSEPORT_LB
if (getsockopt(ls[i].fd, SOL_SOCKET, SO_REUSEPORT_LB,
(void *) &reuseport, &olen)
== -1)
{
ngx_log_error(NGX_LOG_ALERT, cycle->log, ngx_socket_errno,
"getsockopt(SO_REUSEPORT_LB) %V failed, ignored",
&ls[i].addr_text);
} else {
ls[i].reuseport = reuseport ? 1 : 0;
}
#else
if (getsockopt(ls[i].fd, SOL_SOCKET, SO_REUSEPORT,
(void *) &reuseport, &olen)
== -1)
{
ngx_log_error(NGX_LOG_ALERT, cycle->log, ngx_socket_errno,
"getsockopt(SO_REUSEPORT) %V failed, ignored",
&ls[i].addr_text);
} else {
ls[i].reuseport = reuseport ? 1 : 0;
}
#endif
#endif
// udp不检查下面的fastopen等参数
if (ls[i].type != SOCK_STREAM) {
continue;
}
#if (NGX_HAVE_TCP_FASTOPEN)
olen = sizeof(int);
// tcp专用的fastopen
if (getsockopt(ls[i].fd, IPPROTO_TCP, TCP_FASTOPEN,
(void *) &ls[i].fastopen, &olen)
== -1)
{
err = ngx_socket_errno;
if (err != NGX_EOPNOTSUPP && err != NGX_ENOPROTOOPT
&& err != NGX_EINVAL)
{
ngx_log_error(NGX_LOG_NOTICE, cycle->log, err,
"getsockopt(TCP_FASTOPEN) %V failed, ignored",
&ls[i].addr_text);
}
ls[i].fastopen = -1;
}
#endif
#if (NGX_HAVE_DEFERRED_ACCEPT && defined SO_ACCEPTFILTER)
// tcp专用的defered accept
ngx_memzero(&af, sizeof(struct accept_filter_arg));
olen = sizeof(struct accept_filter_arg);
if (getsockopt(ls[i].fd, SOL_SOCKET, SO_ACCEPTFILTER, &af, &olen)
== -1)
{
err = ngx_socket_errno;
if (err == NGX_EINVAL) {
continue;
}
ngx_log_error(NGX_LOG_NOTICE, cycle->log, err,
"getsockopt(SO_ACCEPTFILTER) for %V failed, ignored",
&ls[i].addr_text);
continue;
}
if (olen < sizeof(struct accept_filter_arg) || af.af_name[0] == '\0') {
continue;
}
ls[i].accept_filter = ngx_palloc(cycle->pool, 16);
if (ls[i].accept_filter == NULL) {
return NGX_ERROR;
}
(void) ngx_cpystrn((u_char *) ls[i].accept_filter,
(u_char *) af.af_name, 16);
#endif
#if (NGX_HAVE_DEFERRED_ACCEPT && defined TCP_DEFER_ACCEPT)
timeout = 0;
olen = sizeof(int);
if (getsockopt(ls[i].fd, IPPROTO_TCP, TCP_DEFER_ACCEPT, &timeout, &olen)
== -1)
{
err = ngx_socket_errno;
if (err == NGX_EOPNOTSUPP) {
continue;
}
ngx_log_error(NGX_LOG_NOTICE, cycle->log, err,
"getsockopt(TCP_DEFER_ACCEPT) for %V failed, ignored",
&ls[i].addr_text);
continue;
}
if (olen < sizeof(int) || timeout == 0) {
continue;
}
ls[i].deferred_accept = 1;
#endif
}
return NGX_OK;
}
// ngx_cycle.c : init_cycle()里被调用
// 创建socket, bind/listen
// 之后会调用ngx_configure_listening_sockets()
// 配置监听端口的rcvbuf/sndbuf等参数，调用setsockopt()
ngx_int_t
ngx_open_listening_sockets(ngx_cycle_t *cycle)
{
int reuseaddr;
ngx_uint_t i, tries, failed;
ngx_err_t err;
ngx_log_t *log;
ngx_socket_t s;
ngx_listening_t *ls;
// 默认使用reuseaddr选项
reuseaddr = 1;
#if (NGX_SUPPRESS_WARN)
failed = 0;
#endif
log = cycle->log;
/* TODO: configurable try number */
for (tries = 5; tries; tries--) {
failed = 0;
/* for each listening socket */
// 遍历监听端口链表
// 如果设置了reuseport，那么一个端口会有worker个克隆的结构体
ls = cycle->listening.elts;
for (i = 0; i < cycle->listening.nelts; i++) {
if (ls[i].ignore) {
continue;
}
#if (NGX_HAVE_REUSEPORT)
// 检查是否已经设置的标志位，只是用在inherited的时候，见注释
// 在init_cycle里设置
if (ls[i].add_reuseport) {
/*
* to allow transition from a socket without SO_REUSEPORT
* to multiple sockets with SO_REUSEPORT, we have to set
* SO_REUSEPORT on the old socket before opening new ones
*/
int reuseport = 1;
#ifdef SO_REUSEPORT_LB
if (setsockopt(ls[i].fd, SOL_SOCKET, SO_REUSEPORT_LB,
(const void *) &reuseport, sizeof(int))
== -1)
{
ngx_log_error(NGX_LOG_ALERT, cycle->log, ngx_socket_errno,
"setsockopt(SO_REUSEPORT_LB) %V failed, "
"ignored",
&ls[i].addr_text);
}
#else
if (setsockopt(ls[i].fd, SOL_SOCKET, SO_REUSEPORT,
(const void *) &reuseport, sizeof(int))
== -1)
{
ngx_log_error(NGX_LOG_ALERT, cycle->log, ngx_socket_errno,
"setsockopt(SO_REUSEPORT) %V failed, ignored",
&ls[i].addr_text);
}
#endif
// 标志位清零
ls[i].add_reuseport = 0;
}
#endif
// 已经打开的端口不再处理
if (ls[i].fd != (ngx_socket_t) -1) {
continue;
}
// 从前一个nginx进程继承过来的
// 已经打开，所以也不需要再处理
// 在ngx_set_inherited_sockets里操作
if (ls[i].inherited) {
/* TODO: close on exit */
/* TODO: nonblocking */
/* TODO: deferred accept */
continue;
}
// 创建socket
s = ngx_socket(ls[i].sockaddr->sa_family, ls[i].type, 0);
if (s == (ngx_socket_t) -1) {
ngx_log_error(NGX_LOG_EMERG, log, ngx_socket_errno,
ngx_socket_n " %V failed", &ls[i].addr_text);
return NGX_ERROR;
}
// 设置SO_REUSEADDR选项
if (setsockopt(s, SOL_SOCKET, SO_REUSEADDR,
(const void *) &reuseaddr, sizeof(int))
== -1)
{
ngx_log_error(NGX_LOG_EMERG, log, ngx_socket_errno,
"setsockopt(SO_REUSEADDR) %V failed",
&ls[i].addr_text);
if (ngx_close_socket(s) == -1) {
ngx_log_error(NGX_LOG_EMERG, log, ngx_socket_errno,
ngx_close_socket_n " %V failed",
&ls[i].addr_text);
}
return NGX_ERROR;
}
// 设置SO_REUSEPORT选项
// 这样多个进程可以打开相同的端口，由内核负责负载均衡
#if (NGX_HAVE_REUSEPORT)
if (ls[i].reuseport && !ngx_test_config) {
int reuseport;
reuseport = 1;
#ifdef SO_REUSEPORT_LB
if (setsockopt(s, SOL_SOCKET, SO_REUSEPORT_LB,
(const void *) &reuseport, sizeof(int))
== -1)
{
ngx_log_error(NGX_LOG_EMERG, log, ngx_socket_errno,
"setsockopt(SO_REUSEPORT_LB) %V failed",
&ls[i].addr_text);
if (ngx_close_socket(s) == -1) {
ngx_log_error(NGX_LOG_EMERG, log, ngx_socket_errno,
ngx_close_socket_n " %V failed",
&ls[i].addr_text);
}
return NGX_ERROR;
}
#else
if (setsockopt(s, SOL_SOCKET, SO_REUSEPORT,
(const void *) &reuseport, sizeof(int))
== -1)
{
ngx_log_error(NGX_LOG_EMERG, log, ngx_socket_errno,
"setsockopt(SO_REUSEPORT) %V failed",
&ls[i].addr_text);
if (ngx_close_socket(s) == -1) {
ngx_log_error(NGX_LOG_EMERG, log, ngx_socket_errno,
ngx_close_socket_n " %V failed",
&ls[i].addr_text);
}
return NGX_ERROR;
}
#endif
}
#endif
#if (NGX_HAVE_INET6 && defined IPV6_V6ONLY)
if (ls[i].sockaddr->sa_family == AF_INET6) {
int ipv6only;
ipv6only = ls[i].ipv6only;
if (setsockopt(s, IPPROTO_IPV6, IPV6_V6ONLY,
(const void *) &ipv6only, sizeof(int))
== -1)
{
ngx_log_error(NGX_LOG_EMERG, log, ngx_socket_errno,
"setsockopt(IPV6_V6ONLY) %V failed, ignored",
&ls[i].addr_text);
}
}
#endif
/* TODO: close on exit */
if (!(ngx_event_flags & NGX_USE_IOCP_EVENT)) {
// 设置为nonblocking，使用FIONBIO
if (ngx_nonblocking(s) == -1) {
ngx_log_error(NGX_LOG_EMERG, log, ngx_socket_errno,
ngx_nonblocking_n " %V failed",
&ls[i].addr_text);
if (ngx_close_socket(s) == -1) {
ngx_log_error(NGX_LOG_EMERG, log, ngx_socket_errno,
ngx_close_socket_n " %V failed",
&ls[i].addr_text);
}
return NGX_ERROR;
}
}
ngx_log_debug2(NGX_LOG_DEBUG_CORE, log, 0,
"bind() %V #%d ", &ls[i].addr_text, s);
//绑定地址
if (bind(s, ls[i].sockaddr, ls[i].socklen) == -1) {
err = ngx_socket_errno;
if (err != NGX_EADDRINUSE || !ngx_test_config) {
ngx_log_error(NGX_LOG_EMERG, log, err,
"bind() to %V failed", &ls[i].addr_text);
}
if (ngx_close_socket(s) == -1) {
ngx_log_error(NGX_LOG_EMERG, log, ngx_socket_errno,
ngx_close_socket_n " %V failed",
&ls[i].addr_text);
}
if (err != NGX_EADDRINUSE) {
return NGX_ERROR;
}
if (!ngx_test_config) {
failed = 1;
}
continue;
}
#if (NGX_HAVE_UNIX_DOMAIN)
if (ls[i].sockaddr->sa_family == AF_UNIX) {
mode_t mode;
u_char *name;
name = ls[i].addr_text.data + sizeof("unix:") - 1;
mode = (S_IRUSR|S_IWUSR|S_IRGRP|S_IWGRP|S_IROTH|S_IWOTH);
if (chmod((char *) name, mode) == -1) {
ngx_log_error(NGX_LOG_EMERG, cycle->log, ngx_errno,
"chmod() \"%s\" failed", name);
}
if (ngx_test_config) {
if (ngx_delete_file(name) == NGX_FILE_ERROR) {
ngx_log_error(NGX_LOG_EMERG, cycle->log, ngx_errno,
ngx_delete_file_n " %s failed", name);
}
}
}
#endif
// 检查是否是tcp，即SOCK_STREAM
if (ls[i].type != SOCK_STREAM) {
// 如果不是tcp，即udp，那么无需监听
ls[i].fd = s;
// 继续处理下一个监听结构体
continue;
}
// 开始监听，设置backlog
if (listen(s, ls[i].backlog) == -1) {
err = ngx_socket_errno;
/*
* on OpenVZ after suspend/resume EADDRINUSE
* may be returned by listen() instead of bind(), see
* https://bugzilla.openvz.org/show_bug.cgi?id=2470
*/
if (err != NGX_EADDRINUSE || !ngx_test_config) {
ngx_log_error(NGX_LOG_EMERG, log, err,
"listen() to %V, backlog %d failed",
&ls[i].addr_text, ls[i].backlog);
}
if (ngx_close_socket(s) == -1) {
ngx_log_error(NGX_LOG_EMERG, log, ngx_socket_errno,
ngx_close_socket_n " %V failed",
&ls[i].addr_text);
}
if (err != NGX_EADDRINUSE) {
return NGX_ERROR;
}
if (!ngx_test_config) {
failed = 1;
}
continue;
}
// 设置已经监听标志
ls[i].listen = 1;
// 设置socket描述符
ls[i].fd = s;
}
if (!failed) {
break;
}
/* TODO: delay configurable */
ngx_log_error(NGX_LOG_NOTICE, log, 0,
"try again to bind() after 500ms");
ngx_msleep(500);
}
if (failed) {
ngx_log_error(NGX_LOG_EMERG, log, 0, "still could not bind()");
return NGX_ERROR;
}
return NGX_OK;
}
// ngx_init_cycle()里调用，在ngx_open_listening_sockets()之后
// 配置监听端口的rcvbuf/sndbuf等参数，调用setsockopt()
void
ngx_configure_listening_sockets(ngx_cycle_t *cycle)
{
int value;
ngx_uint_t i;
ngx_listening_t *ls;
#if (NGX_HAVE_DEFERRED_ACCEPT && defined SO_ACCEPTFILTER)
struct accept_filter_arg af;
#endif
// 遍历监听端口链表
ls = cycle->listening.elts;
for (i = 0; i < cycle->listening.nelts; i++) {
ls[i].log = *ls[i].logp;
// rcvbuf
if (ls[i].rcvbuf != -1) {
if (setsockopt(ls[i].fd, SOL_SOCKET, SO_RCVBUF,
(const void *) &ls[i].rcvbuf, sizeof(int))
== -1)
{
ngx_log_error(NGX_LOG_ALERT, cycle->log, ngx_socket_errno,
"setsockopt(SO_RCVBUF, %d) %V failed, ignored",
ls[i].rcvbuf, &ls[i].addr_text);
}
}
// sndbuf
if (ls[i].sndbuf != -1) {
if (setsockopt(ls[i].fd, SOL_SOCKET, SO_SNDBUF,
(const void *) &ls[i].sndbuf, sizeof(int))
== -1)
{
ngx_log_error(NGX_LOG_ALERT, cycle->log, ngx_socket_errno,
"setsockopt(SO_SNDBUF, %d) %V failed, ignored",
ls[i].sndbuf, &ls[i].addr_text);
}
}
// keepalive
if (ls[i].keepalive) {
value = (ls[i].keepalive == 1) ? 1 : 0;
if (setsockopt(ls[i].fd, SOL_SOCKET, SO_KEEPALIVE,
(const void *) &value, sizeof(int))
== -1)
{
ngx_log_error(NGX_LOG_ALERT, cycle->log, ngx_socket_errno,
"setsockopt(SO_KEEPALIVE, %d) %V failed, ignored",
value, &ls[i].addr_text);
}
}
#if (NGX_HAVE_KEEPALIVE_TUNABLE)
if (ls[i].keepidle) {
value = ls[i].keepidle;
#if (NGX_KEEPALIVE_FACTOR)
value *= NGX_KEEPALIVE_FACTOR;
#endif
if (setsockopt(ls[i].fd, IPPROTO_TCP, TCP_KEEPIDLE,
(const void *) &value, sizeof(int))
== -1)
{
ngx_log_error(NGX_LOG_ALERT, cycle->log, ngx_socket_errno,
"setsockopt(TCP_KEEPIDLE, %d) %V failed, ignored",
value, &ls[i].addr_text);
}
}
if (ls[i].keepintvl) {
value = ls[i].keepintvl;
#if (NGX_KEEPALIVE_FACTOR)
value *= NGX_KEEPALIVE_FACTOR;
#endif
if (setsockopt(ls[i].fd, IPPROTO_TCP, TCP_KEEPINTVL,
(const void *) &value, sizeof(int))
== -1)
{
ngx_log_error(NGX_LOG_ALERT, cycle->log, ngx_socket_errno,
"setsockopt(TCP_KEEPINTVL, %d) %V failed, ignored",
value, &ls[i].addr_text);
}
}
if (ls[i].keepcnt) {
if (setsockopt(ls[i].fd, IPPROTO_TCP, TCP_KEEPCNT,
(const void *) &ls[i].keepcnt, sizeof(int))
== -1)
{
ngx_log_error(NGX_LOG_ALERT, cycle->log, ngx_socket_errno,
"setsockopt(TCP_KEEPCNT, %d) %V failed, ignored",
ls[i].keepcnt, &ls[i].addr_text);
}
}
#endif
#if (NGX_HAVE_SETFIB)
if (ls[i].setfib != -1) {
if (setsockopt(ls[i].fd, SOL_SOCKET, SO_SETFIB,
(const void *) &ls[i].setfib, sizeof(int))
== -1)
{
ngx_log_error(NGX_LOG_ALERT, cycle->log, ngx_socket_errno,
"setsockopt(SO_SETFIB, %d) %V failed, ignored",
ls[i].setfib, &ls[i].addr_text);
}
}
#endif
// tcp fast open， 可以优化tcp三次握手的延迟，提高响应速度
#if (NGX_HAVE_TCP_FASTOPEN)
if (ls[i].fastopen != -1) {
if (setsockopt(ls[i].fd, IPPROTO_TCP, TCP_FASTOPEN,
(const void *) &ls[i].fastopen, sizeof(int))
== -1)
{
ngx_log_error(NGX_LOG_ALERT, cycle->log, ngx_socket_errno,
"setsockopt(TCP_FASTOPEN, %d) %V failed, ignored",
ls[i].fastopen, &ls[i].addr_text);
}
}
#endif
#if 0
if (1) {
int tcp_nodelay = 1;
if (setsockopt(ls[i].fd, IPPROTO_TCP, TCP_NODELAY,
(const void *) &tcp_nodelay, sizeof(int))
== -1)
{
ngx_log_error(NGX_LOG_ALERT, cycle->log, ngx_socket_errno,
"setsockopt(TCP_NODELAY) %V failed, ignored",
&ls[i].addr_text);
}
}
#endif
// 已经监听标志, backlog
if (ls[i].listen) {
/* change backlog via listen() */
// 修改了tcp选项，重新监听
if (listen(ls[i].fd, ls[i].backlog) == -1) {
ngx_log_error(NGX_LOG_ALERT, cycle->log, ngx_socket_errno,
"listen() to %V, backlog %d failed, ignored",
&ls[i].addr_text, ls[i].backlog);
}
}
/*
* setting deferred mode should be last operation on socket,
* because code may prematurely continue cycle on failure
*/
#if (NGX_HAVE_DEFERRED_ACCEPT)
// 这个是freebsd的设置
#ifdef SO_ACCEPTFILTER
if (ls[i].delete_deferred) {
if (setsockopt(ls[i].fd, SOL_SOCKET, SO_ACCEPTFILTER, NULL, 0)
== -1)
{
ngx_log_error(NGX_LOG_ALERT, cycle->log, ngx_socket_errno,
"setsockopt(SO_ACCEPTFILTER, NULL) "
"for %V failed, ignored",
&ls[i].addr_text);
if (ls[i].accept_filter) {
ngx_log_error(NGX_LOG_ALERT, cycle->log, 0,
"could not change the accept filter "
"to \"%s\" for %V, ignored",