mirror/nginx
1
0
Fork 0
mirror of https://github.com/nginx/nginx.git synced 2024-12-16 14:19:00 +08:00
Code Issues Actions Packages Projects Releases Wiki Activity
2ce791f2cd
nginx/src/event/ngx_event_accept.c

835 lines
21 KiB
C
Raw Normal View History

nginx-0.0.1-2002-08-06-20:39:45 import The first code that uses "ngx_" prefix, the previous one used "gx_" prefix. At that point the code is not yet usable. The first draft ideas are dated back to 23.10.2001.
2002-08-07 00:39:45 +08:00
nginx-0.1.0-2004-09-28-12:34:51 import; set copyright and remove unused files
2004-09-28 16:34:51 +08:00
/*
nginx-0.1.0-2004-09-29-20:00:49 import; remove years from copyright
2004-09-30 00:00:49 +08:00
* Copyright (C) Igor Sysoev
Copyright updated.
2012-01-18 23:07:43 +08:00
* Copyright (C) Nginx, Inc.
nginx-0.1.0-2004-09-28-12:34:51 import; set copyright and remove unused files
2004-09-28 16:34:51 +08:00
*/
nginx-0.0.1-2002-08-06-20:39:45 import The first code that uses "ngx_" prefix, the previous one used "gx_" prefix. At that point the code is not yet usable. The first draft ideas are dated back to 23.10.2001.
2002-08-07 00:39:45 +08:00
#include <ngx_config.h>
nginx-0.0.1-2002-08-26-19:18:19 import
2002-08-26 23:18:19 +08:00
#include <ngx_core.h>
nginx-0.0.1-2002-08-06-20:39:45 import The first code that uses "ngx_" prefix, the previous one used "gx_" prefix. At that point the code is not yet usable. The first draft ideas are dated back to 23.10.2001.
2002-08-07 00:39:45 +08:00
#include <ngx_event.h>
nginx-0.3.8-RELEASE import *) Security: nginx now checks URI got from a backend in "X-Accel-Redirect" header line or in SSI file for the "/../" paths and zeroes. *) Change: nginx now does not treat the empty user name in the "Authorization" header line as valid one. *) Feature: the "ssl_session_timeout" directives of the ngx_http_ssl_module and ngx_imap_ssl_module. *) Feature: the "auth_http_header" directive of the ngx_imap_auth_http_module. *) Feature: the "add_header" directive. *) Feature: the ngx_http_realip_module. *) Feature: the new variables to use in the "log_format" directive: $bytes_sent, $apache_bytes_sent, $status, $time_gmt, $uri, $request_time, $request_length, $upstream_status, $upstream_response_time, $gzip_ratio, $uid_got, $uid_set, $connection, $pipe, and $msec. The parameters in the "%name" form will be canceled soon. *) Change: now the false variable values in the "if" directive are the empty string "" and string starting with "0". *) Bugfix: while using proxied or FastCGI-server nginx may leave connections and temporary files with client requests in open state. *) Bugfix: the worker processes did not flush the buffered logs on graceful exit. *) Bugfix: if the request URI was changes by the "rewrite" directive and the request was proxied in location given by regular expression, then the incorrect request was transferred to backend; the bug had appeared in 0.2.6. *) Bugfix: the "expires" directive did not remove the previous "Expires" header. *) Bugfix: nginx may stop to accept requests if the "rtsig" method and several worker processes were used. *) Bugfix: the "\"" and "\'" escape symbols were incorrectly handled in SSI commands. *) Bugfix: if the response was ended just after the SSI command and gzipping was used, then the response did not transferred complete or did not transferred at all.
2005-11-10 01:25:55 +08:00
static ngx_int_t ngx_enable_accept_events(ngx_cycle_t *cycle);
The "reuseport" option of the "listen" directive. When configured, an individual listen socket on a given address is created for each worker process. This allows to reduce in-kernel lock contention on configurations with high accept rates, resulting in better performance. As of now it works on Linux and DragonFly BSD. Note that on Linux incoming connection requests are currently tied up to a specific listen socket, and if some sockets are closed, connection requests will be reset, see https://lwn.net/Articles/542629/. With nginx, this may happen if the number of worker processes is reduced. There is no such problem on DragonFly BSD. Based on previous work by Sepherosa Ziehau and Yingqi Lu.
2015-05-20 20:51:56 +08:00
static ngx_int_t ngx_disable_accept_events(ngx_cycle_t *cycle, ngx_uint_t all);
nginx-0.2.0-RELEASE import *) The pid-file names used during online upgrade was changed and now is not required a manual rename operation. The old master process adds the ".oldbin" suffix to its pid-file and executes a new binary file. The new master process creates usual pid-file without the ".newbin" suffix. If the master process exits, then old master process renames back its pid-file with the ".oldbin" suffix to the pid-file without suffix. *) Change: the "worker_connections" directive, new name of the "connections" directive; now the directive specifies maximum number of connections, but not maximum socket descriptor number. *) Feature: SSL supports the session cache inside one worker process. *) Feature: the "satisfy_any" directive. *) Change: the ngx_http_access_module and ngx_http_auth_basic_module do not run for subrequests. *) Feature: the "worker_rlimit_nofile" and "worker_rlimit_sigpending" directives. *) Bugfix: if all backend using in load-balancing failed after one error, then nginx did not try do connect to them during 60 seconds. *) Bugfix: in IMAP/POP3 command argument parsing. Thanks to Rob Mueller. *) Bugfix: errors while using SSL in IMAP/POP3 proxy. *) Bugfix: errors while using SSI and gzipping. *) Bugfix: the "Expires" and "Cache-Control" header lines were omitted from the 304 responses. Thanks to Alexandr Kukushkin.
2005-09-23 19:02:22 +08:00
static void ngx_close_accepted_connection(ngx_connection_t *c);
Stream: UDP proxy.
2016-01-21 00:52:12 +08:00
#if (NGX_DEBUG)
static void ngx_debug_accepted_connection(ngx_event_conf_t *ecf,
ngx_connection_t *c);
#endif
nginx-0.0.1-2003-11-19-19:26:41 import
2003-11-20 00:26:41 +08:00
nginx-0.1.21-RELEASE import *) Bugfix: the ngx_http_stub_status_module showed incorrect statistics if "rtsig" method was used or if several worker process ran on SMP. *) Bugfix: nginx could not be built by the icc compiler on Linux or if the zlib-1.2.x library was building from sources. *) Bugfix: nginx could not be built on NetBSD 2.0.
2005-02-22 22:40:13 +08:00
void
ngx_event_accept(ngx_event_t *ev)
nginx-0.0.1-2002-08-06-20:39:45 import The first code that uses "ngx_" prefix, the previous one used "gx_" prefix. At that point the code is not yet usable. The first draft ideas are dated back to 23.10.2001.
2002-08-07 00:39:45 +08:00
{
rename variable
2006-09-25 22:34:29 +08:00
socklen_t socklen;
nginx-0.1.16-RELEASE import *) Bugfix: if the response were transferred by chunks, then on the HEAD request the final chunk was issued. *) Bugfix: the "Connection: keep-alive" header were issued, even if the keepalive_timeout directive forbade the keep-alive use. *) Bugfix: the errors in the ngx_http_fastcgi_module caused the segmentation faults. *) Bugfix: the compressed response encrypted by SSL may not transferred complete. *) Bugfix: the TCP-specific TCP_NODELAY, TCP_NOPSUH, and TCP_CORK options, are not used for the unix domain sockets. *) Feature: the rewrite directive supports the arguments rewriting. *) Bugfix: the response code 400 was returned for the POST request with the "Content-Length: 0" header; the bug had appeared in 0.1.14.
2005-01-25 20:27:35 +08:00
ngx_err_t err;
ngx_log_t *log;
Accept moderation in case of EMFILE/ENFILE. In case of EMFILE/ENFILE returned from accept() we disable accept events, and (in case of no accept mutex used) arm timer to re-enable them later. With accept mutex we just drop it, and rely on normal accept mutex handling to re-enable accept events once it's acquired again. As we now handle errors in question, logging level was changed to "crit" (instead of "alert" used for unknown errors). Note: the code might call ngx_enable_accept_events() multiple times if there are many listen sockets. The ngx_enable_accept_events() function was modified to check if connection is already active (via c->read->active) and skip it then, thus making multiple calls safe.
2012-05-11 21:33:06 +08:00
ngx_uint_t level;
nginx-0.1.16-RELEASE import *) Bugfix: if the response were transferred by chunks, then on the HEAD request the final chunk was issued. *) Bugfix: the "Connection: keep-alive" header were issued, even if the keepalive_timeout directive forbade the keep-alive use. *) Bugfix: the errors in the ngx_http_fastcgi_module caused the segmentation faults. *) Bugfix: the compressed response encrypted by SSL may not transferred complete. *) Bugfix: the TCP-specific TCP_NODELAY, TCP_NOPSUH, and TCP_CORK options, are not used for the unix domain sockets. *) Feature: the rewrite directive supports the arguments rewriting. *) Bugfix: the response code 400 was returned for the POST request with the "Content-Length: 0" header; the bug had appeared in 0.1.14.
2005-01-25 20:27:35 +08:00
ngx_socket_t s;
ngx_event_t *rev, *wev;
nginx-0.2.0-RELEASE import *) The pid-file names used during online upgrade was changed and now is not required a manual rename operation. The old master process adds the ".oldbin" suffix to its pid-file and executes a new binary file. The new master process creates usual pid-file without the ".newbin" suffix. If the master process exits, then old master process renames back its pid-file with the ".oldbin" suffix to the pid-file without suffix. *) Change: the "worker_connections" directive, new name of the "connections" directive; now the directive specifies maximum number of connections, but not maximum socket descriptor number. *) Feature: SSL supports the session cache inside one worker process. *) Feature: the "satisfy_any" directive. *) Change: the ngx_http_access_module and ngx_http_auth_basic_module do not run for subrequests. *) Feature: the "worker_rlimit_nofile" and "worker_rlimit_sigpending" directives. *) Bugfix: if all backend using in load-balancing failed after one error, then nginx did not try do connect to them during 60 seconds. *) Bugfix: in IMAP/POP3 command argument parsing. Thanks to Rob Mueller. *) Bugfix: errors while using SSL in IMAP/POP3 proxy. *) Bugfix: errors while using SSI and gzipping. *) Bugfix: the "Expires" and "Cache-Control" header lines were omitted from the 304 responses. Thanks to Alexandr Kukushkin.
2005-09-23 19:02:22 +08:00
ngx_listening_t *ls;
ngx_connection_t *c, *lc;
nginx-0.1.16-RELEASE import *) Bugfix: if the response were transferred by chunks, then on the HEAD request the final chunk was issued. *) Bugfix: the "Connection: keep-alive" header were issued, even if the keepalive_timeout directive forbade the keep-alive use. *) Bugfix: the errors in the ngx_http_fastcgi_module caused the segmentation faults. *) Bugfix: the compressed response encrypted by SSL may not transferred complete. *) Bugfix: the TCP-specific TCP_NODELAY, TCP_NOPSUH, and TCP_CORK options, are not used for the unix domain sockets. *) Feature: the rewrite directive supports the arguments rewriting. *) Bugfix: the response code 400 was returned for the POST request with the "Content-Length: 0" header; the bug had appeared in 0.1.14.
2005-01-25 20:27:35 +08:00
ngx_event_conf_t *ecf;
a prelimiary IPv6 support, HTTP listen
2009-02-21 15:02:02 +08:00
u_char sa[NGX_SOCKADDRLEN];
fallback to accept() if accept4() is not implemented, the issue has been introduced in r3787
2010-12-02 18:05:41 +08:00
#if (NGX_HAVE_ACCEPT4)
static ngx_uint_t use_accept4 = 1;
#endif
nginx-0.0.1-2003-05-29-17:02:09 import
2003-05-29 21:02:09 +08:00
Accept moderation in case of EMFILE/ENFILE. In case of EMFILE/ENFILE returned from accept() we disable accept events, and (in case of no accept mutex used) arm timer to re-enable them later. With accept mutex we just drop it, and rely on normal accept mutex handling to re-enable accept events once it's acquired again. As we now handle errors in question, logging level was changed to "crit" (instead of "alert" used for unknown errors). Note: the code might call ngx_enable_accept_events() multiple times if there are many listen sockets. The ngx_enable_accept_events() function was modified to check if connection is already active (via c->read->active) and skip it then, thus making multiple calls safe.
2012-05-11 21:33:06 +08:00
if (ev->timedout) {
if (ngx_enable_accept_events((ngx_cycle_t *) ngx_cycle) != NGX_OK) {
return;
}
ev->timedout = 0;
}
nginx-0.0.1-2003-07-07-10:11:50 import
2003-07-07 14:11:50 +08:00
ecf = ngx_event_get_conf(ngx_cycle->conf_ctx, ngx_event_core_module);
nginx-0.0.1-2003-01-30-10:28:09 import
2003-01-30 15:28:09 +08:00
Removed the obsolete rtsig module.
2015-04-23 19:17:40 +08:00
if (!(ngx_event_flags & NGX_USE_KQUEUE_EVENT)) {
nginx-0.0.2-2004-02-03-00:19:52 import
2004-02-03 05:19:52 +08:00
ev->available = ecf->multi_accept;
}
nginx-0.2.0-RELEASE import *) The pid-file names used during online upgrade was changed and now is not required a manual rename operation. The old master process adds the ".oldbin" suffix to its pid-file and executes a new binary file. The new master process creates usual pid-file without the ".newbin" suffix. If the master process exits, then old master process renames back its pid-file with the ".oldbin" suffix to the pid-file without suffix. *) Change: the "worker_connections" directive, new name of the "connections" directive; now the directive specifies maximum number of connections, but not maximum socket descriptor number. *) Feature: SSL supports the session cache inside one worker process. *) Feature: the "satisfy_any" directive. *) Change: the ngx_http_access_module and ngx_http_auth_basic_module do not run for subrequests. *) Feature: the "worker_rlimit_nofile" and "worker_rlimit_sigpending" directives. *) Bugfix: if all backend using in load-balancing failed after one error, then nginx did not try do connect to them during 60 seconds. *) Bugfix: in IMAP/POP3 command argument parsing. Thanks to Rob Mueller. *) Bugfix: errors while using SSL in IMAP/POP3 proxy. *) Bugfix: errors while using SSI and gzipping. *) Bugfix: the "Expires" and "Cache-Control" header lines were omitted from the 304 responses. Thanks to Alexandr Kukushkin.
2005-09-23 19:02:22 +08:00
lc = ev->data;
ls = lc->listening;
ev->ready = 0;
nginx-0.0.1-2002-12-15-09:25:09 import
2002-12-15 14:25:09 +08:00
nginx-0.0.2-2004-02-11-20:08:49 import
2004-02-12 01:08:49 +08:00
ngx_log_debug2(NGX_LOG_DEBUG_EVENT, ev->log, 0,
nginx-0.2.0-RELEASE import *) The pid-file names used during online upgrade was changed and now is not required a manual rename operation. The old master process adds the ".oldbin" suffix to its pid-file and executes a new binary file. The new master process creates usual pid-file without the ".newbin" suffix. If the master process exits, then old master process renames back its pid-file with the ".oldbin" suffix to the pid-file without suffix. *) Change: the "worker_connections" directive, new name of the "connections" directive; now the directive specifies maximum number of connections, but not maximum socket descriptor number. *) Feature: SSL supports the session cache inside one worker process. *) Feature: the "satisfy_any" directive. *) Change: the ngx_http_access_module and ngx_http_auth_basic_module do not run for subrequests. *) Feature: the "worker_rlimit_nofile" and "worker_rlimit_sigpending" directives. *) Bugfix: if all backend using in load-balancing failed after one error, then nginx did not try do connect to them during 60 seconds. *) Bugfix: in IMAP/POP3 command argument parsing. Thanks to Rob Mueller. *) Bugfix: errors while using SSL in IMAP/POP3 proxy. *) Bugfix: errors while using SSI and gzipping. *) Bugfix: the "Expires" and "Cache-Control" header lines were omitted from the 304 responses. Thanks to Alexandr Kukushkin.
2005-09-23 19:02:22 +08:00
"accept on %V, ready: %d", &ls->addr_text, ev->available);
nginx-0.0.1-2003-01-30-10:28:09 import
2003-01-30 15:28:09 +08:00
nginx-0.0.1-2002-08-06-20:39:45 import The first code that uses "ngx_" prefix, the previous one used "gx_" prefix. At that point the code is not yet usable. The first draft ideas are dated back to 23.10.2001.
2002-08-07 00:39:45 +08:00
do {
a prelimiary IPv6 support, HTTP listen
2009-02-21 15:02:02 +08:00
socklen = NGX_SOCKADDRLEN;
nginx-0.0.1-2003-05-20-19:37:55 import
2003-05-20 23:37:55 +08:00
Linux accept4() support
2010-11-25 18:15:04 +08:00
#if (NGX_HAVE_ACCEPT4)
fallback to accept() if accept4() is not implemented, the issue has been introduced in r3787
2010-12-02 18:05:41 +08:00
if (use_accept4) {
s = accept4(lc->fd, (struct sockaddr *) sa, &socklen,
SOCK_NONBLOCK);
} else {
s = accept(lc->fd, (struct sockaddr *) sa, &socklen);
}
Linux accept4() support
2010-11-25 18:15:04 +08:00
#else
rename variable
2006-09-25 22:34:29 +08:00
s = accept(lc->fd, (struct sockaddr *) sa, &socklen);
Linux accept4() support
2010-11-25 18:15:04 +08:00
#endif
nginx-0.1.13-RELEASE import *) Feature: the server_names_hash and server_names_hash_threshold directives. *) Bugfix: the *.domain.tld names in the "server_name" directive did not work. *) Bugfix: the %request_length log parameter logged the incorrect length.
2004-12-21 20:30:30 +08:00
Win32: MinGW GCC compatibility. Several warnings silenced, notably (ngx_socket_t) -1 is now checked on socket operations instead of -1, as ngx_socket_t is unsigned on win32 and gcc complains on comparison. With this patch, it's now possible to compile nginx using mingw gcc, with options we normally compile on win32.
2013-09-05 00:48:28 +08:00
if (s == (ngx_socket_t) -1) {
nginx-0.0.1-2002-08-06-20:39:45 import The first code that uses "ngx_" prefix, the previous one used "gx_" prefix. At that point the code is not yet usable. The first draft ideas are dated back to 23.10.2001.
2002-08-07 00:39:45 +08:00
err = ngx_socket_errno;
nginx-0.0.1-2003-01-30-10:28:09 import
2003-01-30 15:28:09 +08:00
nginx-0.0.1-2002-08-06-20:39:45 import The first code that uses "ngx_" prefix, the previous one used "gx_" prefix. At that point the code is not yet usable. The first draft ideas are dated back to 23.10.2001.
2002-08-07 00:39:45 +08:00
if (err == NGX_EAGAIN) {
Solaris 10 event ports support
2006-09-26 20:20:12 +08:00
ngx_log_debug0(NGX_LOG_DEBUG_EVENT, ev->log, err,
"accept() not ready");
nginx-0.0.1-2003-05-13-20:02:32 import
2003-05-14 00:02:32 +08:00
return;
nginx-0.0.1-2002-08-06-20:39:45 import The first code that uses "ngx_" prefix, the previous one used "gx_" prefix. At that point the code is not yet usable. The first draft ideas are dated back to 23.10.2001.
2002-08-07 00:39:45 +08:00
}
nginx-0.0.1-2002-08-26-19:18:19 import
2002-08-26 23:18:19 +08:00
Accept moderation in case of EMFILE/ENFILE. In case of EMFILE/ENFILE returned from accept() we disable accept events, and (in case of no accept mutex used) arm timer to re-enable them later. With accept mutex we just drop it, and rely on normal accept mutex handling to re-enable accept events once it's acquired again. As we now handle errors in question, logging level was changed to "crit" (instead of "alert" used for unknown errors). Note: the code might call ngx_enable_accept_events() multiple times if there are many listen sockets. The ngx_enable_accept_events() function was modified to check if connection is already active (via c->read->active) and skip it then, thus making multiple calls safe.
2012-05-11 21:33:06 +08:00
level = NGX_LOG_ALERT;
if (err == NGX_ECONNABORTED) {
level = NGX_LOG_ERR;
} else if (err == NGX_EMFILE || err == NGX_ENFILE) {
level = NGX_LOG_CRIT;
}
fix building by msvc8 introduced by the previous commit
2010-12-02 22:20:51 +08:00
#if (NGX_HAVE_ACCEPT4)
Accept moderation in case of EMFILE/ENFILE. In case of EMFILE/ENFILE returned from accept() we disable accept events, and (in case of no accept mutex used) arm timer to re-enable them later. With accept mutex we just drop it, and rely on normal accept mutex handling to re-enable accept events once it's acquired again. As we now handle errors in question, logging level was changed to "crit" (instead of "alert" used for unknown errors). Note: the code might call ngx_enable_accept_events() multiple times if there are many listen sockets. The ngx_enable_accept_events() function was modified to check if connection is already active (via c->read->active) and skip it then, thus making multiple calls safe.
2012-05-11 21:33:06 +08:00
ngx_log_error(level, ev->log, err,
fallback to accept() if accept4() is not implemented, the issue has been introduced in r3787
2010-12-02 18:05:41 +08:00
use_accept4 ? "accept4() failed" : "accept() failed");
if (use_accept4 && err == NGX_ENOSYS) {
use_accept4 = 0;
ngx_inherited_nonblocking = 0;
continue;
}
fix building by msvc8 introduced by the previous commit
2010-12-02 22:20:51 +08:00
#else
Accept moderation in case of EMFILE/ENFILE. In case of EMFILE/ENFILE returned from accept() we disable accept events, and (in case of no accept mutex used) arm timer to re-enable them later. With accept mutex we just drop it, and rely on normal accept mutex handling to re-enable accept events once it's acquired again. As we now handle errors in question, logging level was changed to "crit" (instead of "alert" used for unknown errors). Note: the code might call ngx_enable_accept_events() multiple times if there are many listen sockets. The ngx_enable_accept_events() function was modified to check if connection is already active (via c->read->active) and skip it then, thus making multiple calls safe.
2012-05-11 21:33:06 +08:00
ngx_log_error(level, ev->log, err, "accept() failed");
fallback to accept() if accept4() is not implemented, the issue has been introduced in r3787
2010-12-02 18:05:41 +08:00
#endif
nginx-0.0.1-2003-05-29-17:02:09 import
2003-05-29 21:02:09 +08:00
nginx-0.0.2-2004-02-03-00:19:52 import
2004-02-03 05:19:52 +08:00
if (err == NGX_ECONNABORTED) {
nginx-0.1.2-RELEASE import *) Feature: the --user=USER, --group=GROUP, and --with-ld-opt=OPTIONS options in configure. *) Feature: the server_name directive supports *.domain.tld. *) Bugfix: the portability improvements. *) Bugfix: if configuration file was set in command line, the reconfiguration was impossible; the bug had appeared in 0.1.1. *) Bugfix: proxy module may get caught in an endless loop when sendfile is not used. *) Bugfix: with sendfile the response was not recoded according to the charset module directives; the bug had appeared in 0.1.1. *) Bugfix: very seldom bug in the kqueue processing. *) Bugfix: the gzip module compressed the proxied responses that was already compressed.
2004-10-21 23:34:38 +08:00
if (ngx_event_flags & NGX_USE_KQUEUE_EVENT) {
nginx-0.0.2-2004-02-03-00:19:52 import
2004-02-03 05:19:52 +08:00
ev->available--;
}
if (ev->available) {
continue;
}
}
Accept moderation in case of EMFILE/ENFILE. In case of EMFILE/ENFILE returned from accept() we disable accept events, and (in case of no accept mutex used) arm timer to re-enable them later. With accept mutex we just drop it, and rely on normal accept mutex handling to re-enable accept events once it's acquired again. As we now handle errors in question, logging level was changed to "crit" (instead of "alert" used for unknown errors). Note: the code might call ngx_enable_accept_events() multiple times if there are many listen sockets. The ngx_enable_accept_events() function was modified to check if connection is already active (via c->read->active) and skip it then, thus making multiple calls safe.
2012-05-11 21:33:06 +08:00
if (err == NGX_EMFILE || err == NGX_ENFILE) {
The "reuseport" option of the "listen" directive. When configured, an individual listen socket on a given address is created for each worker process. This allows to reduce in-kernel lock contention on configurations with high accept rates, resulting in better performance. As of now it works on Linux and DragonFly BSD. Note that on Linux incoming connection requests are currently tied up to a specific listen socket, and if some sockets are closed, connection requests will be reset, see https://lwn.net/Articles/542629/. With nginx, this may happen if the number of worker processes is reduced. There is no such problem on DragonFly BSD. Based on previous work by Sepherosa Ziehau and Yingqi Lu.
2015-05-20 20:51:56 +08:00
if (ngx_disable_accept_events((ngx_cycle_t *) ngx_cycle, 1)
Accept moderation in case of EMFILE/ENFILE. In case of EMFILE/ENFILE returned from accept() we disable accept events, and (in case of no accept mutex used) arm timer to re-enable them later. With accept mutex we just drop it, and rely on normal accept mutex handling to re-enable accept events once it's acquired again. As we now handle errors in question, logging level was changed to "crit" (instead of "alert" used for unknown errors). Note: the code might call ngx_enable_accept_events() multiple times if there are many listen sockets. The ngx_enable_accept_events() function was modified to check if connection is already active (via c->read->active) and skip it then, thus making multiple calls safe.
2012-05-11 21:33:06 +08:00
!= NGX_OK)
{
return;
}
if (ngx_use_accept_mutex) {
if (ngx_accept_mutex_held) {
ngx_shmtx_unlock(&ngx_accept_mutex);
ngx_accept_mutex_held = 0;
}
ngx_accept_disabled = 1;
} else {
ngx_add_timer(ev, ecf->accept_mutex_delay);
}
}
nginx-0.0.1-2003-05-29-17:02:09 import
2003-05-29 21:02:09 +08:00
return;
}
nginx-0.0.10-2004-09-14-23:39:54 import
2004-09-15 03:39:54 +08:00
#if (NGX_STAT_STUB)
ignore ngx_atomic_fetch_add() result this fixes building at least by gcc 4.2.1 on Mac OS X 10.6
2009-06-18 21:14:51 +08:00
(void) ngx_atomic_fetch_add(ngx_stat_accepted, 1);
nginx-0.0.10-2004-09-14-23:39:54 import
2004-09-15 03:39:54 +08:00
#endif
nginx-0.0.3-2004-04-21-22:54:33 import
2004-04-22 02:54:33 +08:00
dynamic accept threshold
2008-09-19 20:47:13 +08:00
ngx_accept_disabled = ngx_cycle->connection_n / 8
nginx-0.2.0-RELEASE import *) The pid-file names used during online upgrade was changed and now is not required a manual rename operation. The old master process adds the ".oldbin" suffix to its pid-file and executes a new binary file. The new master process creates usual pid-file without the ".newbin" suffix. If the master process exits, then old master process renames back its pid-file with the ".oldbin" suffix to the pid-file without suffix. *) Change: the "worker_connections" directive, new name of the "connections" directive; now the directive specifies maximum number of connections, but not maximum socket descriptor number. *) Feature: SSL supports the session cache inside one worker process. *) Feature: the "satisfy_any" directive. *) Change: the ngx_http_access_module and ngx_http_auth_basic_module do not run for subrequests. *) Feature: the "worker_rlimit_nofile" and "worker_rlimit_sigpending" directives. *) Bugfix: if all backend using in load-balancing failed after one error, then nginx did not try do connect to them during 60 seconds. *) Bugfix: in IMAP/POP3 command argument parsing. Thanks to Rob Mueller. *) Bugfix: errors while using SSL in IMAP/POP3 proxy. *) Bugfix: errors while using SSI and gzipping. *) Bugfix: the "Expires" and "Cache-Control" header lines were omitted from the 304 responses. Thanks to Alexandr Kukushkin.
2005-09-23 19:02:22 +08:00
- ngx_cycle->free_connection_n;
c = ngx_get_connection(s, ev->log);
if (c == NULL) {
if (ngx_close_socket(s) == -1) {
ngx_log_error(NGX_LOG_ALERT, ev->log, ngx_socket_errno,
ngx_close_socket_n " failed");
}
return;
}
Stream: UDP proxy.
2016-01-21 00:52:12 +08:00
c->type = SOCK_STREAM;
nginx-0.3.39-RELEASE import *) Feature: the "uninitialized_variable_warn" directive; the logging level of the "uninitialized variable" message was lowered from "alert" to "warn". *) Feature: the "override_charset" directive. *) Change: now if the unknown variable is used in the "echo" and "if expr='$name'" SSI-commands, then the "unknown variable" message is not logged. *) Bugfix: the active connection counter increased on the exceeding of the connection limit specified by the "worker_connections" directive; the bug had appeared in 0.2.0. *) Bugfix: the limit rate might not work on some condition; the bug had appeared in 0.3.38.
2006-04-18 03:55:41 +08:00
#if (NGX_STAT_STUB)
ignore ngx_atomic_fetch_add() result this fixes building at least by gcc 4.2.1 on Mac OS X 10.6
2009-06-18 21:14:51 +08:00
(void) ngx_atomic_fetch_add(ngx_stat_active, 1);
nginx-0.3.39-RELEASE import *) Feature: the "uninitialized_variable_warn" directive; the logging level of the "uninitialized variable" message was lowered from "alert" to "warn". *) Feature: the "override_charset" directive. *) Change: now if the unknown variable is used in the "echo" and "if expr='$name'" SSI-commands, then the "unknown variable" message is not logged. *) Bugfix: the active connection counter increased on the exceeding of the connection limit specified by the "worker_connections" directive; the bug had appeared in 0.2.0. *) Bugfix: the limit rate might not work on some condition; the bug had appeared in 0.3.38.
2006-04-18 03:55:41 +08:00
#endif
nginx-0.2.0-RELEASE import *) The pid-file names used during online upgrade was changed and now is not required a manual rename operation. The old master process adds the ".oldbin" suffix to its pid-file and executes a new binary file. The new master process creates usual pid-file without the ".newbin" suffix. If the master process exits, then old master process renames back its pid-file with the ".oldbin" suffix to the pid-file without suffix. *) Change: the "worker_connections" directive, new name of the "connections" directive; now the directive specifies maximum number of connections, but not maximum socket descriptor number. *) Feature: SSL supports the session cache inside one worker process. *) Feature: the "satisfy_any" directive. *) Change: the ngx_http_access_module and ngx_http_auth_basic_module do not run for subrequests. *) Feature: the "worker_rlimit_nofile" and "worker_rlimit_sigpending" directives. *) Bugfix: if all backend using in load-balancing failed after one error, then nginx did not try do connect to them during 60 seconds. *) Bugfix: in IMAP/POP3 command argument parsing. Thanks to Rob Mueller. *) Bugfix: errors while using SSL in IMAP/POP3 proxy. *) Bugfix: errors while using SSI and gzipping. *) Bugfix: the "Expires" and "Cache-Control" header lines were omitted from the 304 responses. Thanks to Alexandr Kukushkin.
2005-09-23 19:02:22 +08:00
c->pool = ngx_create_pool(ls->pool_size, ev->log);
if (c->pool == NULL) {
ngx_close_accepted_connection(c);
return;
}
nginx-0.0.1-2003-07-21-01:15:59 import
2003-07-21 05:15:59 +08:00
rename variable
2006-09-25 22:34:29 +08:00
c->sockaddr = ngx_palloc(c->pool, socklen);
nginx-0.2.0-RELEASE import *) The pid-file names used during online upgrade was changed and now is not required a manual rename operation. The old master process adds the ".oldbin" suffix to its pid-file and executes a new binary file. The new master process creates usual pid-file without the ".newbin" suffix. If the master process exits, then old master process renames back its pid-file with the ".oldbin" suffix to the pid-file without suffix. *) Change: the "worker_connections" directive, new name of the "connections" directive; now the directive specifies maximum number of connections, but not maximum socket descriptor number. *) Feature: SSL supports the session cache inside one worker process. *) Feature: the "satisfy_any" directive. *) Change: the ngx_http_access_module and ngx_http_auth_basic_module do not run for subrequests. *) Feature: the "worker_rlimit_nofile" and "worker_rlimit_sigpending" directives. *) Bugfix: if all backend using in load-balancing failed after one error, then nginx did not try do connect to them during 60 seconds. *) Bugfix: in IMAP/POP3 command argument parsing. Thanks to Rob Mueller. *) Bugfix: errors while using SSL in IMAP/POP3 proxy. *) Bugfix: errors while using SSI and gzipping. *) Bugfix: the "Expires" and "Cache-Control" header lines were omitted from the 304 responses. Thanks to Alexandr Kukushkin.
2005-09-23 19:02:22 +08:00
if (c->sockaddr == NULL) {
ngx_close_accepted_connection(c);
return;
}
nginx-0.0.1-2003-05-29-17:02:09 import
2003-05-29 21:02:09 +08:00
rename variable
2006-09-25 22:34:29 +08:00
ngx_memcpy(c->sockaddr, sa, socklen);
nginx-0.0.1-2003-05-29-17:02:09 import
2003-05-29 21:02:09 +08:00
nginx-0.2.0-RELEASE import *) The pid-file names used during online upgrade was changed and now is not required a manual rename operation. The old master process adds the ".oldbin" suffix to its pid-file and executes a new binary file. The new master process creates usual pid-file without the ".newbin" suffix. If the master process exits, then old master process renames back its pid-file with the ".oldbin" suffix to the pid-file without suffix. *) Change: the "worker_connections" directive, new name of the "connections" directive; now the directive specifies maximum number of connections, but not maximum socket descriptor number. *) Feature: SSL supports the session cache inside one worker process. *) Feature: the "satisfy_any" directive. *) Change: the ngx_http_access_module and ngx_http_auth_basic_module do not run for subrequests. *) Feature: the "worker_rlimit_nofile" and "worker_rlimit_sigpending" directives. *) Bugfix: if all backend using in load-balancing failed after one error, then nginx did not try do connect to them during 60 seconds. *) Bugfix: in IMAP/POP3 command argument parsing. Thanks to Rob Mueller. *) Bugfix: errors while using SSL in IMAP/POP3 proxy. *) Bugfix: errors while using SSI and gzipping. *) Bugfix: the "Expires" and "Cache-Control" header lines were omitted from the 304 responses. Thanks to Alexandr Kukushkin.
2005-09-23 19:02:22 +08:00
log = ngx_palloc(c->pool, sizeof(ngx_log_t));
if (log == NULL) {
ngx_close_accepted_connection(c);
nginx-0.0.1-2003-05-13-20:02:32 import
2003-05-14 00:02:32 +08:00
return;
nginx-0.0.1-2002-08-06-20:39:45 import The first code that uses "ngx_" prefix, the previous one used "gx_" prefix. At that point the code is not yet usable. The first draft ideas are dated back to 23.10.2001.
2002-08-07 00:39:45 +08:00
}
nginx-0.0.1-2002-08-26-19:18:19 import
2002-08-26 23:18:19 +08:00
Removed the obsolete aio module.
2015-04-22 23:57:32 +08:00
/* set a blocking mode for iocp and non-blocking mode for others */
nginx-0.0.1-2003-05-21-17:28:21 import
2003-05-21 21:28:21 +08:00
if (ngx_inherited_nonblocking) {
Removed the obsolete aio module.
2015-04-22 23:57:32 +08:00
if (ngx_event_flags & NGX_USE_IOCP_EVENT) {
nginx-0.0.1-2003-05-21-17:28:21 import
2003-05-21 21:28:21 +08:00
if (ngx_blocking(s) == -1) {
nginx-0.2.0-RELEASE import *) The pid-file names used during online upgrade was changed and now is not required a manual rename operation. The old master process adds the ".oldbin" suffix to its pid-file and executes a new binary file. The new master process creates usual pid-file without the ".newbin" suffix. If the master process exits, then old master process renames back its pid-file with the ".oldbin" suffix to the pid-file without suffix. *) Change: the "worker_connections" directive, new name of the "connections" directive; now the directive specifies maximum number of connections, but not maximum socket descriptor number. *) Feature: SSL supports the session cache inside one worker process. *) Feature: the "satisfy_any" directive. *) Change: the ngx_http_access_module and ngx_http_auth_basic_module do not run for subrequests. *) Feature: the "worker_rlimit_nofile" and "worker_rlimit_sigpending" directives. *) Bugfix: if all backend using in load-balancing failed after one error, then nginx did not try do connect to them during 60 seconds. *) Bugfix: in IMAP/POP3 command argument parsing. Thanks to Rob Mueller. *) Bugfix: errors while using SSL in IMAP/POP3 proxy. *) Bugfix: errors while using SSI and gzipping. *) Bugfix: the "Expires" and "Cache-Control" header lines were omitted from the 304 responses. Thanks to Alexandr Kukushkin.
2005-09-23 19:02:22 +08:00
ngx_log_error(NGX_LOG_ALERT, ev->log, ngx_socket_errno,
nginx-0.0.1-2003-11-19-19:26:41 import
2003-11-20 00:26:41 +08:00
ngx_blocking_n " failed");
nginx-0.2.0-RELEASE import *) The pid-file names used during online upgrade was changed and now is not required a manual rename operation. The old master process adds the ".oldbin" suffix to its pid-file and executes a new binary file. The new master process creates usual pid-file without the ".newbin" suffix. If the master process exits, then old master process renames back its pid-file with the ".oldbin" suffix to the pid-file without suffix. *) Change: the "worker_connections" directive, new name of the "connections" directive; now the directive specifies maximum number of connections, but not maximum socket descriptor number. *) Feature: SSL supports the session cache inside one worker process. *) Feature: the "satisfy_any" directive. *) Change: the ngx_http_access_module and ngx_http_auth_basic_module do not run for subrequests. *) Feature: the "worker_rlimit_nofile" and "worker_rlimit_sigpending" directives. *) Bugfix: if all backend using in load-balancing failed after one error, then nginx did not try do connect to them during 60 seconds. *) Bugfix: in IMAP/POP3 command argument parsing. Thanks to Rob Mueller. *) Bugfix: errors while using SSL in IMAP/POP3 proxy. *) Bugfix: errors while using SSI and gzipping. *) Bugfix: the "Expires" and "Cache-Control" header lines were omitted from the 304 responses. Thanks to Alexandr Kukushkin.
2005-09-23 19:02:22 +08:00
ngx_close_accepted_connection(c);
nginx-0.0.1-2003-05-21-17:28:21 import
2003-05-21 21:28:21 +08:00
return;
}
nginx-0.0.1-2003-02-06-20:21:13 import
2003-02-07 01:21:13 +08:00
}
nginx-0.0.1-2003-05-21-17:28:21 import
2003-05-21 21:28:21 +08:00
} else {
Removed the obsolete rtsig module.
2015-04-23 19:17:40 +08:00
if (!(ngx_event_flags & NGX_USE_IOCP_EVENT)) {
nginx-0.0.1-2003-05-21-17:28:21 import
2003-05-21 21:28:21 +08:00
if (ngx_nonblocking(s) == -1) {
nginx-0.2.0-RELEASE import *) The pid-file names used during online upgrade was changed and now is not required a manual rename operation. The old master process adds the ".oldbin" suffix to its pid-file and executes a new binary file. The new master process creates usual pid-file without the ".newbin" suffix. If the master process exits, then old master process renames back its pid-file with the ".oldbin" suffix to the pid-file without suffix. *) Change: the "worker_connections" directive, new name of the "connections" directive; now the directive specifies maximum number of connections, but not maximum socket descriptor number. *) Feature: SSL supports the session cache inside one worker process. *) Feature: the "satisfy_any" directive. *) Change: the ngx_http_access_module and ngx_http_auth_basic_module do not run for subrequests. *) Feature: the "worker_rlimit_nofile" and "worker_rlimit_sigpending" directives. *) Bugfix: if all backend using in load-balancing failed after one error, then nginx did not try do connect to them during 60 seconds. *) Bugfix: in IMAP/POP3 command argument parsing. Thanks to Rob Mueller. *) Bugfix: errors while using SSL in IMAP/POP3 proxy. *) Bugfix: errors while using SSI and gzipping. *) Bugfix: the "Expires" and "Cache-Control" header lines were omitted from the 304 responses. Thanks to Alexandr Kukushkin.
2005-09-23 19:02:22 +08:00
ngx_log_error(NGX_LOG_ALERT, ev->log, ngx_socket_errno,
nginx-0.0.1-2003-11-19-19:26:41 import
2003-11-20 00:26:41 +08:00
ngx_nonblocking_n " failed");
nginx-0.2.0-RELEASE import *) The pid-file names used during online upgrade was changed and now is not required a manual rename operation. The old master process adds the ".oldbin" suffix to its pid-file and executes a new binary file. The new master process creates usual pid-file without the ".newbin" suffix. If the master process exits, then old master process renames back its pid-file with the ".oldbin" suffix to the pid-file without suffix. *) Change: the "worker_connections" directive, new name of the "connections" directive; now the directive specifies maximum number of connections, but not maximum socket descriptor number. *) Feature: SSL supports the session cache inside one worker process. *) Feature: the "satisfy_any" directive. *) Change: the ngx_http_access_module and ngx_http_auth_basic_module do not run for subrequests. *) Feature: the "worker_rlimit_nofile" and "worker_rlimit_sigpending" directives. *) Bugfix: if all backend using in load-balancing failed after one error, then nginx did not try do connect to them during 60 seconds. *) Bugfix: in IMAP/POP3 command argument parsing. Thanks to Rob Mueller. *) Bugfix: errors while using SSL in IMAP/POP3 proxy. *) Bugfix: errors while using SSI and gzipping. *) Bugfix: the "Expires" and "Cache-Control" header lines were omitted from the 304 responses. Thanks to Alexandr Kukushkin.
2005-09-23 19:02:22 +08:00
ngx_close_accepted_connection(c);
nginx-0.0.1-2003-05-21-17:28:21 import
2003-05-21 21:28:21 +08:00
return;
}
nginx-0.0.1-2003-02-06-20:21:13 import
2003-02-07 01:21:13 +08:00
}
}
nginx-0.2.0-RELEASE import *) The pid-file names used during online upgrade was changed and now is not required a manual rename operation. The old master process adds the ".oldbin" suffix to its pid-file and executes a new binary file. The new master process creates usual pid-file without the ".newbin" suffix. If the master process exits, then old master process renames back its pid-file with the ".oldbin" suffix to the pid-file without suffix. *) Change: the "worker_connections" directive, new name of the "connections" directive; now the directive specifies maximum number of connections, but not maximum socket descriptor number. *) Feature: SSL supports the session cache inside one worker process. *) Feature: the "satisfy_any" directive. *) Change: the ngx_http_access_module and ngx_http_auth_basic_module do not run for subrequests. *) Feature: the "worker_rlimit_nofile" and "worker_rlimit_sigpending" directives. *) Bugfix: if all backend using in load-balancing failed after one error, then nginx did not try do connect to them during 60 seconds. *) Bugfix: in IMAP/POP3 command argument parsing. Thanks to Rob Mueller. *) Bugfix: errors while using SSL in IMAP/POP3 proxy. *) Bugfix: errors while using SSI and gzipping. *) Bugfix: the "Expires" and "Cache-Control" header lines were omitted from the 304 responses. Thanks to Alexandr Kukushkin.
2005-09-23 19:02:22 +08:00
*log = ls->log;
nginx-0.0.1-2003-06-11-19:28:34 import
2003-06-11 23:28:34 +08:00
nginx-0.2.0-RELEASE import *) The pid-file names used during online upgrade was changed and now is not required a manual rename operation. The old master process adds the ".oldbin" suffix to its pid-file and executes a new binary file. The new master process creates usual pid-file without the ".newbin" suffix. If the master process exits, then old master process renames back its pid-file with the ".oldbin" suffix to the pid-file without suffix. *) Change: the "worker_connections" directive, new name of the "connections" directive; now the directive specifies maximum number of connections, but not maximum socket descriptor number. *) Feature: SSL supports the session cache inside one worker process. *) Feature: the "satisfy_any" directive. *) Change: the ngx_http_access_module and ngx_http_auth_basic_module do not run for subrequests. *) Feature: the "worker_rlimit_nofile" and "worker_rlimit_sigpending" directives. *) Bugfix: if all backend using in load-balancing failed after one error, then nginx did not try do connect to them during 60 seconds. *) Bugfix: in IMAP/POP3 command argument parsing. Thanks to Rob Mueller. *) Bugfix: errors while using SSL in IMAP/POP3 proxy. *) Bugfix: errors while using SSI and gzipping. *) Bugfix: the "Expires" and "Cache-Control" header lines were omitted from the 304 responses. Thanks to Alexandr Kukushkin.
2005-09-23 19:02:22 +08:00
c->recv = ngx_recv;
c->send = ngx_send;
nginx-0.3.10-RELEASE import *) Change: the "valid_referers" directive and the "$invalid_referer" variable were moved to the new ngx_http_referer_module from the ngx_http_rewrite_module. *) Change: the "$apache_bytes_sent" variable name was changed to "$body_bytes_sent". *) Feature: the "$sent_http_..." variables. *) Feature: the "if" directive supports the "=" and "!=" operations. *) Feature: the "proxy_pass" directive supports the HTTPS protocol. *) Feature: the "proxy_set_body" directive. *) Feature: the "post_action" directive. *) Feature: the ngx_http_empty_gif_module. *) Feature: the "worker_cpu_affinity" directive for Linux. *) Bugfix: the "rewrite" directive did not unescape URI part in redirect, now it is unescaped except the %00-%25 and %7F-%FF characters. *) Bugfix: nginx could not be built by the icc 9.0 compiler. *) Bugfix: if the SSI was enabled for zero size static file, then the chunked response was encoded incorrectly.
2005-11-15 21:30:52 +08:00
c->recv_chain = ngx_recv_chain;
nginx-0.2.0-RELEASE import *) The pid-file names used during online upgrade was changed and now is not required a manual rename operation. The old master process adds the ".oldbin" suffix to its pid-file and executes a new binary file. The new master process creates usual pid-file without the ".newbin" suffix. If the master process exits, then old master process renames back its pid-file with the ".oldbin" suffix to the pid-file without suffix. *) Change: the "worker_connections" directive, new name of the "connections" directive; now the directive specifies maximum number of connections, but not maximum socket descriptor number. *) Feature: SSL supports the session cache inside one worker process. *) Feature: the "satisfy_any" directive. *) Change: the ngx_http_access_module and ngx_http_auth_basic_module do not run for subrequests. *) Feature: the "worker_rlimit_nofile" and "worker_rlimit_sigpending" directives. *) Bugfix: if all backend using in load-balancing failed after one error, then nginx did not try do connect to them during 60 seconds. *) Bugfix: in IMAP/POP3 command argument parsing. Thanks to Rob Mueller. *) Bugfix: errors while using SSL in IMAP/POP3 proxy. *) Bugfix: errors while using SSI and gzipping. *) Bugfix: the "Expires" and "Cache-Control" header lines were omitted from the 304 responses. Thanks to Alexandr Kukushkin.
2005-09-23 19:02:22 +08:00
c->send_chain = ngx_send_chain;
nginx-0.0.1-2003-06-11-19:28:34 import
2003-06-11 23:28:34 +08:00
nginx-0.2.0-RELEASE import *) The pid-file names used during online upgrade was changed and now is not required a manual rename operation. The old master process adds the ".oldbin" suffix to its pid-file and executes a new binary file. The new master process creates usual pid-file without the ".newbin" suffix. If the master process exits, then old master process renames back its pid-file with the ".oldbin" suffix to the pid-file without suffix. *) Change: the "worker_connections" directive, new name of the "connections" directive; now the directive specifies maximum number of connections, but not maximum socket descriptor number. *) Feature: SSL supports the session cache inside one worker process. *) Feature: the "satisfy_any" directive. *) Change: the ngx_http_access_module and ngx_http_auth_basic_module do not run for subrequests. *) Feature: the "worker_rlimit_nofile" and "worker_rlimit_sigpending" directives. *) Bugfix: if all backend using in load-balancing failed after one error, then nginx did not try do connect to them during 60 seconds. *) Bugfix: in IMAP/POP3 command argument parsing. Thanks to Rob Mueller. *) Bugfix: errors while using SSL in IMAP/POP3 proxy. *) Bugfix: errors while using SSI and gzipping. *) Bugfix: the "Expires" and "Cache-Control" header lines were omitted from the 304 responses. Thanks to Alexandr Kukushkin.
2005-09-23 19:02:22 +08:00
c->log = log;
c->pool->log = log;
nginx-0.0.1-2002-12-15-09:25:09 import
2002-12-15 14:25:09 +08:00
rename variable
2006-09-25 22:34:29 +08:00
c->socklen = socklen;
a prelimiary IPv6 support, HTTP listen
2009-02-21 15:02:02 +08:00
c->listening = ls;
c->local_sockaddr = ls->sockaddr;
Core: keep the length of the local sockaddr.
2013-12-09 14:14:51 +08:00
c->local_socklen = ls->socklen;
nginx-0.0.1-2003-05-12-19:52:24 import
2003-05-12 23:52:24 +08:00
nginx-0.2.0-RELEASE import *) The pid-file names used during online upgrade was changed and now is not required a manual rename operation. The old master process adds the ".oldbin" suffix to its pid-file and executes a new binary file. The new master process creates usual pid-file without the ".newbin" suffix. If the master process exits, then old master process renames back its pid-file with the ".oldbin" suffix to the pid-file without suffix. *) Change: the "worker_connections" directive, new name of the "connections" directive; now the directive specifies maximum number of connections, but not maximum socket descriptor number. *) Feature: SSL supports the session cache inside one worker process. *) Feature: the "satisfy_any" directive. *) Change: the ngx_http_access_module and ngx_http_auth_basic_module do not run for subrequests. *) Feature: the "worker_rlimit_nofile" and "worker_rlimit_sigpending" directives. *) Bugfix: if all backend using in load-balancing failed after one error, then nginx did not try do connect to them during 60 seconds. *) Bugfix: in IMAP/POP3 command argument parsing. Thanks to Rob Mueller. *) Bugfix: errors while using SSL in IMAP/POP3 proxy. *) Bugfix: errors while using SSI and gzipping. *) Bugfix: the "Expires" and "Cache-Control" header lines were omitted from the 304 responses. Thanks to Alexandr Kukushkin.
2005-09-23 19:02:22 +08:00
c->unexpected_eof = 1;
nginx-0.0.7-2004-07-07-10:15:04 import
2004-07-07 14:15:04 +08:00
http listen unix domain sockets
2009-10-26 19:43:32 +08:00
#if (NGX_HAVE_UNIX_DOMAIN)
if (c->sockaddr->sa_family == AF_UNIX) {
c->tcp_nopush = NGX_TCP_NOPUSH_DISABLED;
c->tcp_nodelay = NGX_TCP_NODELAY_DISABLED;
#if (NGX_SOLARIS)
/* Solaris's sendfilev() supports AF_NCA, AF_INET, and AF_INET6 */
c->sendfile = 0;
#endif
}
#endif
nginx-0.3.2-RELEASE import *) Feature: the Sun Studio 10 C compiler support. *) Feature: the "proxy_upstream_max_fails", "proxy_upstream_fail_timeout", "fastcgi_upstream_max_fails", and "fastcgi_upstream_fail_timeout" directives.
2005-10-12 21:50:36 +08:00
rev = c->read;
wev = c->write;
nginx-0.0.1-2002-08-26-19:18:19 import
2002-08-26 23:18:19 +08:00
nginx-0.0.1-2003-10-29-11:30:44 import
2003-10-29 16:30:44 +08:00
wev->ready = 1;
nginx-0.0.1-2002-08-06-20:39:45 import The first code that uses "ngx_" prefix, the previous one used "gx_" prefix. At that point the code is not yet usable. The first draft ideas are dated back to 23.10.2001.
2002-08-07 00:39:45 +08:00
Removed the obsolete rtsig module.
2015-04-23 19:17:40 +08:00
if (ngx_event_flags & NGX_USE_IOCP_EVENT) {
nginx-0.0.1-2003-10-29-11:30:44 import
2003-10-29 16:30:44 +08:00
rev->ready = 1;
nginx-0.0.1-2003-02-06-20:21:13 import
2003-02-07 01:21:13 +08:00
}
nginx-0.0.1-2004-01-30-00:45:01 import
2004-01-30 05:45:01 +08:00
if (ev->deferred_accept) {
rev->ready = 1;
nginx-0.1.36-RELEASE import *) Change: if the request header has duplicate the "Host", "Connection", "Content-Length", or "Authorization" lines, then nginx now returns the 400 error. *) Change: the "post_accept_timeout" directive was canceled. *) Feature: the "default", "af=", "bl=", "deferred", and "bind" parameters of the "listen" directive. *) Feature: the FreeBSD accept filters support. *) Feature: the Linux TCP_DEFER_ACCEPT support. *) Bugfix: the ngx_http_autoindex_module did not support the file names in UTF-8. *) Bugfix: the new log file can be rotated by the -USR1 signal only if the reconfiguration by the -HUP signal was made twice.
2005-06-16 02:33:41 +08:00
#if (NGX_HAVE_KQUEUE)
rev->available = 1;
#endif
nginx-0.0.1-2004-01-30-00:45:01 import
2004-01-30 05:45:01 +08:00
}
nginx-0.0.1-2003-11-19-19:26:41 import
2003-11-20 00:26:41 +08:00
rev->log = log;
wev->log = log;
nginx-0.0.1-2002-08-06-20:39:45 import The first code that uses "ngx_" prefix, the previous one used "gx_" prefix. At that point the code is not yet usable. The first draft ideas are dated back to 23.10.2001.
2002-08-07 00:39:45 +08:00
nginx-0.0.1-2003-12-08-23:48:12 import
2003-12-09 04:48:12 +08:00
/*
nginx-0.3.2-RELEASE import *) Feature: the Sun Studio 10 C compiler support. *) Feature: the "proxy_upstream_max_fails", "proxy_upstream_fail_timeout", "fastcgi_upstream_max_fails", and "fastcgi_upstream_fail_timeout" directives.
2005-10-12 21:50:36 +08:00
* TODO: MT: - ngx_atomic_fetch_add()
nginx-0.0.2-2004-02-23-23:57:12 import
2004-02-24 04:57:12 +08:00
* or protection by critical section or light mutex
nginx-0.0.1-2003-12-08-23:48:12 import
2003-12-09 04:48:12 +08:00
*
* TODO: MP: - allocated in a shared memory
nginx-0.3.2-RELEASE import *) Feature: the Sun Studio 10 C compiler support. *) Feature: the "proxy_upstream_max_fails", "proxy_upstream_fail_timeout", "fastcgi_upstream_max_fails", and "fastcgi_upstream_fail_timeout" directives.
2005-10-12 21:50:36 +08:00
* - ngx_atomic_fetch_add()
nginx-0.0.2-2004-02-23-23:57:12 import
2004-02-24 04:57:12 +08:00
* or protection by critical section or light mutex
nginx-0.0.1-2003-12-08-23:48:12 import
2003-12-09 04:48:12 +08:00
*/
nginx-0.0.1-2003-12-19-11:15:11 import
2003-12-19 16:15:11 +08:00
nginx-0.3.2-RELEASE import *) Feature: the Sun Studio 10 C compiler support. *) Feature: the "proxy_upstream_max_fails", "proxy_upstream_fail_timeout", "fastcgi_upstream_max_fails", and "fastcgi_upstream_fail_timeout" directives.
2005-10-12 21:50:36 +08:00
c->number = ngx_atomic_fetch_add(ngx_connection_counter, 1);
nginx-0.0.1-2002-08-06-20:39:45 import The first code that uses "ngx_" prefix, the previous one used "gx_" prefix. At that point the code is not yet usable. The first draft ideas are dated back to 23.10.2001.
2002-08-07 00:39:45 +08:00
nginx-0.1.22-RELEASE import *) Bugfix: the ngx_http_stub_status_module showed incorrect handled connections statistics if the proxying or FastCGI server were used. *) Bugfix: the installation paths were incorrectly quoted on Linux and Solaris; the bug had appeared in 0.1.21.
2005-02-24 20:29:09 +08:00
#if (NGX_STAT_STUB)
ignore ngx_atomic_fetch_add() result this fixes building at least by gcc 4.2.1 on Mac OS X 10.6
2009-06-18 21:14:51 +08:00
(void) ngx_atomic_fetch_add(ngx_stat_handled, 1);
nginx-0.1.22-RELEASE import *) Bugfix: the ngx_http_stub_status_module showed incorrect handled connections statistics if the proxying or FastCGI server were used. *) Bugfix: the installation paths were incorrectly quoted on Linux and Solaris; the bug had appeared in 0.1.21.
2005-02-24 20:29:09 +08:00
#endif
nginx-0.2.0-RELEASE import *) The pid-file names used during online upgrade was changed and now is not required a manual rename operation. The old master process adds the ".oldbin" suffix to its pid-file and executes a new binary file. The new master process creates usual pid-file without the ".newbin" suffix. If the master process exits, then old master process renames back its pid-file with the ".oldbin" suffix to the pid-file without suffix. *) Change: the "worker_connections" directive, new name of the "connections" directive; now the directive specifies maximum number of connections, but not maximum socket descriptor number. *) Feature: SSL supports the session cache inside one worker process. *) Feature: the "satisfy_any" directive. *) Change: the ngx_http_access_module and ngx_http_auth_basic_module do not run for subrequests. *) Feature: the "worker_rlimit_nofile" and "worker_rlimit_sigpending" directives. *) Bugfix: if all backend using in load-balancing failed after one error, then nginx did not try do connect to them during 60 seconds. *) Bugfix: in IMAP/POP3 command argument parsing. Thanks to Rob Mueller. *) Bugfix: errors while using SSL in IMAP/POP3 proxy. *) Bugfix: errors while using SSI and gzipping. *) Bugfix: the "Expires" and "Cache-Control" header lines were omitted from the 304 responses. Thanks to Alexandr Kukushkin.
2005-09-23 19:02:22 +08:00
if (ls->addr_ntop) {
*) back out r2040 *) refactor ngx_palloc() *) introduce ngx_pnalloc() *) additional pool blocks have smaller header
2008-06-17 23:00:30 +08:00
c->addr_text.data = ngx_pnalloc(c->pool, ls->addr_text_max_len);
nginx-0.0.10-2004-09-09-19:40:48 import
2004-09-09 23:40:48 +08:00
if (c->addr_text.data == NULL) {
nginx-0.2.0-RELEASE import *) The pid-file names used during online upgrade was changed and now is not required a manual rename operation. The old master process adds the ".oldbin" suffix to its pid-file and executes a new binary file. The new master process creates usual pid-file without the ".newbin" suffix. If the master process exits, then old master process renames back its pid-file with the ".oldbin" suffix to the pid-file without suffix. *) Change: the "worker_connections" directive, new name of the "connections" directive; now the directive specifies maximum number of connections, but not maximum socket descriptor number. *) Feature: SSL supports the session cache inside one worker process. *) Feature: the "satisfy_any" directive. *) Change: the ngx_http_access_module and ngx_http_auth_basic_module do not run for subrequests. *) Feature: the "worker_rlimit_nofile" and "worker_rlimit_sigpending" directives. *) Bugfix: if all backend using in load-balancing failed after one error, then nginx did not try do connect to them during 60 seconds. *) Bugfix: in IMAP/POP3 command argument parsing. Thanks to Rob Mueller. *) Bugfix: errors while using SSL in IMAP/POP3 proxy. *) Bugfix: errors while using SSI and gzipping. *) Bugfix: the "Expires" and "Cache-Control" header lines were omitted from the 304 responses. Thanks to Alexandr Kukushkin.
2005-09-23 19:02:22 +08:00
ngx_close_accepted_connection(c);
nginx-0.0.10-2004-09-09-19:40:48 import
2004-09-09 23:40:48 +08:00
return;
}
nginx-0.3.10-RELEASE import *) Change: the "valid_referers" directive and the "$invalid_referer" variable were moved to the new ngx_http_referer_module from the ngx_http_rewrite_module. *) Change: the "$apache_bytes_sent" variable name was changed to "$body_bytes_sent". *) Feature: the "$sent_http_..." variables. *) Feature: the "if" directive supports the "=" and "!=" operations. *) Feature: the "proxy_pass" directive supports the HTTPS protocol. *) Feature: the "proxy_set_body" directive. *) Feature: the "post_action" directive. *) Feature: the ngx_http_empty_gif_module. *) Feature: the "worker_cpu_affinity" directive for Linux. *) Bugfix: the "rewrite" directive did not unescape URI part in redirect, now it is unescaped except the %00-%25 and %7F-%FF characters. *) Bugfix: nginx could not be built by the icc 9.0 compiler. *) Bugfix: if the SSI was enabled for zero size static file, then the chunked response was encoded incorrectly.
2005-11-15 21:30:52 +08:00
Core: extended ngx_sock_ntop() with socklen parameter. On Linux, sockaddr length is required to process unix socket addresses properly due to unnamed sockets (which don't have sun_path set at all) and abstract namespace sockets.
2013-07-11 20:07:25 +08:00
c->addr_text.len = ngx_sock_ntop(c->sockaddr, c->socklen,
c->addr_text.data,
a prelimiary IPv6 support, HTTP listen
2009-02-21 15:02:02 +08:00
ls->addr_text_max_len, 0);
nginx-0.0.10-2004-09-09-19:40:48 import
2004-09-09 23:40:48 +08:00
if (c->addr_text.len == 0) {
nginx-0.2.0-RELEASE import *) The pid-file names used during online upgrade was changed and now is not required a manual rename operation. The old master process adds the ".oldbin" suffix to its pid-file and executes a new binary file. The new master process creates usual pid-file without the ".newbin" suffix. If the master process exits, then old master process renames back its pid-file with the ".oldbin" suffix to the pid-file without suffix. *) Change: the "worker_connections" directive, new name of the "connections" directive; now the directive specifies maximum number of connections, but not maximum socket descriptor number. *) Feature: SSL supports the session cache inside one worker process. *) Feature: the "satisfy_any" directive. *) Change: the ngx_http_access_module and ngx_http_auth_basic_module do not run for subrequests. *) Feature: the "worker_rlimit_nofile" and "worker_rlimit_sigpending" directives. *) Bugfix: if all backend using in load-balancing failed after one error, then nginx did not try do connect to them during 60 seconds. *) Bugfix: in IMAP/POP3 command argument parsing. Thanks to Rob Mueller. *) Bugfix: errors while using SSL in IMAP/POP3 proxy. *) Bugfix: errors while using SSI and gzipping. *) Bugfix: the "Expires" and "Cache-Control" header lines were omitted from the 304 responses. Thanks to Alexandr Kukushkin.
2005-09-23 19:02:22 +08:00
ngx_close_accepted_connection(c);
nginx-0.0.10-2004-09-09-19:40:48 import
2004-09-09 23:40:48 +08:00
return;
}
}
nginx-0.0.3-2004-04-08-19:58:25 import
2004-04-08 23:58:25 +08:00
#if (NGX_DEBUG)
{
Stream: UDP proxy.
2016-01-21 00:52:12 +08:00
ngx_str_t addr;
u_char text[NGX_SOCKADDR_STRLEN];
nginx-0.0.3-2004-04-08-19:58:25 import
2004-04-08 23:58:25 +08:00
Stream: UDP proxy.
2016-01-21 00:52:12 +08:00
ngx_debug_accepted_connection(ecf, c);
if (log->log_level & NGX_LOG_DEBUG_EVENT) {
addr.data = text;
addr.len = ngx_sock_ntop(c->sockaddr, c->socklen, text,
NGX_SOCKADDR_STRLEN, 1);
ngx_log_debug3(NGX_LOG_DEBUG_EVENT, log, 0,
"*%uA accept: %V fd:%d", c->number, &addr, s);
}
}
debug_connection: added the IPv6 and UNIX-domain socket support.
2012-04-30 06:02:18 +08:00
#endif
nginx-0.0.3-2004-04-08-19:58:25 import
2004-04-08 23:58:25 +08:00
Stream: UDP proxy.
2016-01-21 00:52:12 +08:00
if (ngx_add_conn && (ngx_event_flags & NGX_USE_EPOLL_EVENT) == 0) {
if (ngx_add_conn(c) == NGX_ERROR) {
ngx_close_accepted_connection(c);
return;
debug_connection: added the IPv6 and UNIX-domain socket support.
2012-04-30 06:02:18 +08:00
}
Stream: UDP proxy.
2016-01-21 00:52:12 +08:00
}
debug_connection: added the IPv6 and UNIX-domain socket support.
2012-04-30 06:02:18 +08:00
Stream: UDP proxy.
2016-01-21 00:52:12 +08:00
log->data = NULL;
log->handler = NULL;
debug_connection: added the IPv6 and UNIX-domain socket support.
2012-04-30 06:02:18 +08:00
Stream: UDP proxy.
2016-01-21 00:52:12 +08:00
ls->handler(c);
if (ngx_event_flags & NGX_USE_KQUEUE_EVENT) {
ev->available--;
}
} while (ev->available);
}
#if !(NGX_WIN32)
void
ngx_event_recvmsg(ngx_event_t *ev)
{
ssize_t n;
ngx_log_t *log;
ngx_err_t err;
ngx_event_t *rev, *wev;
struct iovec iov[1];
struct msghdr msg;
ngx_listening_t *ls;
ngx_event_conf_t *ecf;
ngx_connection_t *c, *lc;
u_char sa[NGX_SOCKADDRLEN];
static u_char buffer[65535];
#if (NGX_HAVE_MSGHDR_MSG_CONTROL)
#if (NGX_HAVE_IP_RECVDSTADDR)
u_char msg_control[CMSG_SPACE(sizeof(struct in_addr))];
#elif (NGX_HAVE_IP_PKTINFO)
u_char msg_control[CMSG_SPACE(sizeof(struct in_pktinfo))];
debug_connection: added the IPv6 and UNIX-domain socket support.
2012-04-30 06:02:18 +08:00
#endif
Stream: UDP proxy.
2016-01-21 00:52:12 +08:00
#if (NGX_HAVE_INET6 && NGX_HAVE_IPV6_RECVPKTINFO)
u_char msg_control6[CMSG_SPACE(sizeof(struct in6_pktinfo))];
debug_connection: added the IPv6 and UNIX-domain socket support.
2012-04-30 06:02:18 +08:00
#endif
Stream: UDP proxy.
2016-01-21 00:52:12 +08:00
#endif
if (ev->timedout) {
if (ngx_enable_accept_events((ngx_cycle_t *) ngx_cycle) != NGX_OK) {
return;
}
ev->timedout = 0;
}
ecf = ngx_event_get_conf(ngx_cycle->conf_ctx, ngx_event_core_module);
if (!(ngx_event_flags & NGX_USE_KQUEUE_EVENT)) {
ev->available = ecf->multi_accept;
}
lc = ev->data;
ls = lc->listening;
ev->ready = 0;
ngx_log_debug2(NGX_LOG_DEBUG_EVENT, ev->log, 0,
"recvmsg on %V, ready: %d", &ls->addr_text, ev->available);
do {
ngx_memzero(&msg, sizeof(struct msghdr));
iov[0].iov_base = (void *) buffer;
iov[0].iov_len = sizeof(buffer);
msg.msg_name = &sa;
msg.msg_namelen = sizeof(sa);
msg.msg_iov = iov;
msg.msg_iovlen = 1;
#if (NGX_HAVE_MSGHDR_MSG_CONTROL)
if (ls->wildcard) {
#if (NGX_HAVE_IP_RECVDSTADDR || NGX_HAVE_IP_PKTINFO)
if (ls->sockaddr->sa_family == AF_INET) {
msg.msg_control = &msg_control;
msg.msg_controllen = sizeof(msg_control);
nginx-0.0.3-2004-04-08-19:58:25 import
2004-04-08 23:58:25 +08:00
}
Stream: UDP proxy.
2016-01-21 00:52:12 +08:00
#endif
debug_connection: added the IPv6 and UNIX-domain socket support.
2012-04-30 06:02:18 +08:00
Stream: UDP proxy.
2016-01-21 00:52:12 +08:00
#if (NGX_HAVE_INET6 && NGX_HAVE_IPV6_RECVPKTINFO)
if (ls->sockaddr->sa_family == AF_INET6) {
msg.msg_control = &msg_control6;
msg.msg_controllen = sizeof(msg_control6);
}
#endif
}
#endif
n = recvmsg(lc->fd, &msg, 0);
if (n == -1) {
err = ngx_socket_errno;
if (err == NGX_EAGAIN) {
ngx_log_debug0(NGX_LOG_DEBUG_EVENT, ev->log, err,
"recvmsg() not ready");
return;
}
ngx_log_error(NGX_LOG_ALERT, ev->log, err, "recvmsg() failed");
return;
}
#if (NGX_STAT_STUB)
(void) ngx_atomic_fetch_add(ngx_stat_accepted, 1);
#endif
debug_connection: added the IPv6 and UNIX-domain socket support.
2012-04-30 06:02:18 +08:00
Stream: UDP proxy.
2016-01-21 00:52:12 +08:00
#if (NGX_HAVE_MSGHDR_MSG_CONTROL)
if (msg.msg_flags & (MSG_TRUNC|MSG_CTRUNC)) {
ngx_log_error(NGX_LOG_ALERT, ev->log, 0,
"recvmsg() truncated data");
debug_connection: added the IPv6 and UNIX-domain socket support.
2012-04-30 06:02:18 +08:00
continue;
nginx-0.0.3-2004-04-08-19:58:25 import
2004-04-08 23:58:25 +08:00
}
Stream: UDP proxy.
2016-01-21 00:52:12 +08:00
#endif
nginx-0.0.3-2004-04-08-19:58:25 import
2004-04-08 23:58:25 +08:00
Stream: UDP proxy.
2016-01-21 00:52:12 +08:00
ngx_accept_disabled = ngx_cycle->connection_n / 8
- ngx_cycle->free_connection_n;
Core: output client port number when logging accept event.
2014-05-22 04:16:09 +08:00
Stream: UDP proxy.
2016-01-21 00:52:12 +08:00
c = ngx_get_connection(lc->fd, ev->log);
if (c == NULL) {
return;
}
c->shared = 1;
c->type = SOCK_DGRAM;
c->socklen = msg.msg_namelen;
#if (NGX_STAT_STUB)
(void) ngx_atomic_fetch_add(ngx_stat_active, 1);
#endif
c->pool = ngx_create_pool(ls->pool_size, ev->log);
if (c->pool == NULL) {
ngx_close_accepted_connection(c);
return;
nginx-0.0.3-2004-04-08-19:58:25 import
2004-04-08 23:58:25 +08:00
}
Stream: UDP proxy.
2016-01-21 00:52:12 +08:00
c->sockaddr = ngx_palloc(c->pool, c->socklen);
if (c->sockaddr == NULL) {
ngx_close_accepted_connection(c);
return;
Core: output client port number when logging accept event.
2014-05-22 04:16:09 +08:00
}
Stream: UDP proxy.
2016-01-21 00:52:12 +08:00
ngx_memcpy(c->sockaddr, msg.msg_name, c->socklen);
log = ngx_palloc(c->pool, sizeof(ngx_log_t));
if (log == NULL) {
ngx_close_accepted_connection(c);
return;
}
*log = ls->log;
c->send = ngx_udp_send;
c->log = log;
c->pool->log = log;
c->listening = ls;
c->local_sockaddr = ls->sockaddr;
c->local_socklen = ls->socklen;
#if (NGX_HAVE_MSGHDR_MSG_CONTROL)
if (ls->wildcard) {
struct cmsghdr *cmsg;
struct sockaddr *sockaddr;
sockaddr = ngx_palloc(c->pool, c->local_socklen);
if (sockaddr == NULL) {
ngx_close_accepted_connection(c);
return;
}
ngx_memcpy(sockaddr, c->local_sockaddr, c->local_socklen);
c->local_sockaddr = sockaddr;
for (cmsg = CMSG_FIRSTHDR(&msg);
cmsg != NULL;
cmsg = CMSG_NXTHDR(&msg, cmsg))
{
#if (NGX_HAVE_IP_RECVDSTADDR)
if (cmsg->cmsg_level == IPPROTO_IP
&& cmsg->cmsg_type == IP_RECVDSTADDR
&& sockaddr->sa_family == AF_INET)
{
struct in_addr *addr;
struct sockaddr_in *sin;
addr = (struct in_addr *) CMSG_DATA(cmsg);
sin = (struct sockaddr_in *) sockaddr;
sin->sin_addr = *addr;
break;
}
#elif (NGX_HAVE_IP_PKTINFO)
if (cmsg->cmsg_level == IPPROTO_IP
&& cmsg->cmsg_type == IP_PKTINFO
&& sockaddr->sa_family == AF_INET)
{
struct in_pktinfo *pkt;
struct sockaddr_in *sin;
pkt = (struct in_pktinfo *) CMSG_DATA(cmsg);
sin = (struct sockaddr_in *) sockaddr;
sin->sin_addr = pkt->ipi_addr;
break;
}
Core: output client port number when logging accept event.
2014-05-22 04:16:09 +08:00
#endif
nginx-0.3.54-RELEASE import *) Feature: nginx now logs the subrequest information to the error log. *) Feature: the "proxy_next_upstream", "fastcgi_next_upstream", and "memcached_next_upstream" directives support the "off" parameter. *) Feature: the "debug_connection" directive supports the CIDR address form. *) Bugfix: if a response of proxied server or FastCGI server was converted from UTF-8 or back, then it may be transferred incomplete. *) Bugfix: the $upstream_response_time variable had the time of the first request to a backend only. *) Bugfix: nginx could not be built on amd64 platform; the bug had appeared in 0.3.53.
2006-07-11 21:20:19 +08:00
Stream: UDP proxy.
2016-01-21 00:52:12 +08:00
#if (NGX_HAVE_INET6 && NGX_HAVE_IPV6_RECVPKTINFO)
if (cmsg->cmsg_level == IPPROTO_IPV6
&& cmsg->cmsg_type == IPV6_PKTINFO
&& sockaddr->sa_family == AF_INET6)
{
struct in6_pktinfo *pkt6;
struct sockaddr_in6 *sin6;
pkt6 = (struct in6_pktinfo *) CMSG_DATA(cmsg);
sin6 = (struct sockaddr_in6 *) sockaddr;
sin6->sin6_addr = pkt6->ipi6_addr;
break;
}
#endif
}
}
#endif
c->buffer = ngx_create_temp_buf(c->pool, n);
if (c->buffer == NULL) {
ngx_close_accepted_connection(c);
return;
}
c->buffer->last = ngx_cpymem(c->buffer->last, buffer, n);
rev = c->read;
wev = c->write;
wev->ready = 1;
rev->log = log;
wev->log = log;
/*
* TODO: MT: - ngx_atomic_fetch_add()
* or protection by critical section or light mutex
*
* TODO: MP: - allocated in a shared memory
* - ngx_atomic_fetch_add()
* or protection by critical section or light mutex
*/
c->number = ngx_atomic_fetch_add(ngx_connection_counter, 1);
#if (NGX_STAT_STUB)
(void) ngx_atomic_fetch_add(ngx_stat_handled, 1);
#endif
if (ls->addr_ntop) {
c->addr_text.data = ngx_pnalloc(c->pool, ls->addr_text_max_len);
if (c->addr_text.data == NULL) {
nginx-0.2.0-RELEASE import *) The pid-file names used during online upgrade was changed and now is not required a manual rename operation. The old master process adds the ".oldbin" suffix to its pid-file and executes a new binary file. The new master process creates usual pid-file without the ".newbin" suffix. If the master process exits, then old master process renames back its pid-file with the ".oldbin" suffix to the pid-file without suffix. *) Change: the "worker_connections" directive, new name of the "connections" directive; now the directive specifies maximum number of connections, but not maximum socket descriptor number. *) Feature: SSL supports the session cache inside one worker process. *) Feature: the "satisfy_any" directive. *) Change: the ngx_http_access_module and ngx_http_auth_basic_module do not run for subrequests. *) Feature: the "worker_rlimit_nofile" and "worker_rlimit_sigpending" directives. *) Bugfix: if all backend using in load-balancing failed after one error, then nginx did not try do connect to them during 60 seconds. *) Bugfix: in IMAP/POP3 command argument parsing. Thanks to Rob Mueller. *) Bugfix: errors while using SSL in IMAP/POP3 proxy. *) Bugfix: errors while using SSI and gzipping. *) Bugfix: the "Expires" and "Cache-Control" header lines were omitted from the 304 responses. Thanks to Alexandr Kukushkin.
2005-09-23 19:02:22 +08:00
ngx_close_accepted_connection(c);
nginx-0.0.1-2003-05-13-20:02:32 import
2003-05-14 00:02:32 +08:00
return;
nginx-0.0.1-2003-03-04-09:33:48 import
2003-03-04 14:33:48 +08:00
}
Stream: UDP proxy.
2016-01-21 00:52:12 +08:00
c->addr_text.len = ngx_sock_ntop(c->sockaddr, c->socklen,
c->addr_text.data,
ls->addr_text_max_len, 0);
if (c->addr_text.len == 0) {
ngx_close_accepted_connection(c);
return;
}
}
#if (NGX_DEBUG)
{
ngx_str_t addr;
u_char text[NGX_SOCKADDR_STRLEN];
ngx_debug_accepted_connection(ecf, c);
if (log->log_level & NGX_LOG_DEBUG_EVENT) {
addr.data = text;
addr.len = ngx_sock_ntop(c->sockaddr, c->socklen, text,
NGX_SOCKADDR_STRLEN, 1);
ngx_log_debug4(NGX_LOG_DEBUG_EVENT, log, 0,
"*%uA recvmsg: %V fd:%d n:%z",
c->number, &addr, c->fd, n);
}
nginx-0.0.1-2003-03-04-09:33:48 import
2003-03-04 14:33:48 +08:00
}
Stream: UDP proxy.
2016-01-21 00:52:12 +08:00
#endif
nginx-0.0.1-2003-03-04-09:33:48 import
2003-03-04 14:33:48 +08:00
nginx-0.0.1-2003-11-19-19:26:41 import
2003-11-20 00:26:41 +08:00
log->data = NULL;
log->handler = NULL;
nginx-0.2.0-RELEASE import *) The pid-file names used during online upgrade was changed and now is not required a manual rename operation. The old master process adds the ".oldbin" suffix to its pid-file and executes a new binary file. The new master process creates usual pid-file without the ".newbin" suffix. If the master process exits, then old master process renames back its pid-file with the ".oldbin" suffix to the pid-file without suffix. *) Change: the "worker_connections" directive, new name of the "connections" directive; now the directive specifies maximum number of connections, but not maximum socket descriptor number. *) Feature: SSL supports the session cache inside one worker process. *) Feature: the "satisfy_any" directive. *) Change: the ngx_http_access_module and ngx_http_auth_basic_module do not run for subrequests. *) Feature: the "worker_rlimit_nofile" and "worker_rlimit_sigpending" directives. *) Bugfix: if all backend using in load-balancing failed after one error, then nginx did not try do connect to them during 60 seconds. *) Bugfix: in IMAP/POP3 command argument parsing. Thanks to Rob Mueller. *) Bugfix: errors while using SSL in IMAP/POP3 proxy. *) Bugfix: errors while using SSI and gzipping. *) Bugfix: the "Expires" and "Cache-Control" header lines were omitted from the 304 responses. Thanks to Alexandr Kukushkin.
2005-09-23 19:02:22 +08:00
ls->handler(c);
nginx-0.0.1-2002-08-06-20:39:45 import The first code that uses "ngx_" prefix, the previous one used "gx_" prefix. At that point the code is not yet usable. The first draft ideas are dated back to 23.10.2001.
2002-08-07 00:39:45 +08:00
nginx-0.1.2-RELEASE import *) Feature: the --user=USER, --group=GROUP, and --with-ld-opt=OPTIONS options in configure. *) Feature: the server_name directive supports *.domain.tld. *) Bugfix: the portability improvements. *) Bugfix: if configuration file was set in command line, the reconfiguration was impossible; the bug had appeared in 0.1.1. *) Bugfix: proxy module may get caught in an endless loop when sendfile is not used. *) Bugfix: with sendfile the response was not recoded according to the charset module directives; the bug had appeared in 0.1.1. *) Bugfix: very seldom bug in the kqueue processing. *) Bugfix: the gzip module compressed the proxied responses that was already compressed.
2004-10-21 23:34:38 +08:00
if (ngx_event_flags & NGX_USE_KQUEUE_EVENT) {
Stream: UDP proxy.
2016-01-21 00:52:12 +08:00
ev->available -= n;
nginx-0.0.1-2003-01-30-10:28:09 import
2003-01-30 15:28:09 +08:00
}
nginx-0.0.1-2003-11-19-19:26:41 import
2003-11-20 00:26:41 +08:00
nginx-0.0.1-2002-08-06-20:39:45 import The first code that uses "ngx_" prefix, the previous one used "gx_" prefix. At that point the code is not yet usable. The first draft ideas are dated back to 23.10.2001.
2002-08-07 00:39:45 +08:00
} while (ev->available);
}
nginx-0.0.1-2003-11-19-19:26:41 import
2003-11-20 00:26:41 +08:00
Stream: UDP proxy.
2016-01-21 00:52:12 +08:00
#endif
nginx-0.0.1-2003-11-19-19:26:41 import
2003-11-20 00:26:41 +08:00
nginx-0.1.21-RELEASE import *) Bugfix: the ngx_http_stub_status_module showed incorrect statistics if "rtsig" method was used or if several worker process ran on SMP. *) Bugfix: nginx could not be built by the icc compiler on Linux or if the zlib-1.2.x library was building from sources. *) Bugfix: nginx could not be built on NetBSD 2.0.
2005-02-22 22:40:13 +08:00
ngx_int_t
ngx_trylock_accept_mutex(ngx_cycle_t *cycle)
nginx-0.0.3-2004-03-31-19:26:46 import
2004-03-31 23:26:46 +08:00
{
nginx-0.3.27-RELEASE import *) Change: the "variables_hash_max_size" and "variables_hash_bucket_size" directives. *) Feature: the $body_bytes_sent variable can be used not only in the "log_format" directive. *) Feature: the $ssl_protocol and $ssl_cipher variables. *) Feature: the cache line size detection for widespread CPUs at start time. *) Feature: now the "accept_mutex" directive is supported using fcntl(2) on platforms different from i386, amd64, sparc64, and ppc. *) Feature: the "lock_file" directive and the --with-lock-path=PATH autoconfiguration directive. *) Bugfix: if the HTTPS protocol was used in the "proxy_pass" directive then the requests with the body was not transferred.
2006-02-08 23:33:12 +08:00
if (ngx_shmtx_trylock(&ngx_accept_mutex)) {
nginx-0.0.3-2004-03-31-19:26:46 import
2004-03-31 23:26:46 +08:00
ngx_log_debug0(NGX_LOG_DEBUG_EVENT, cycle->log, 0,
"accept mutex locked");
Removed the obsolete rtsig module.
2015-04-23 19:17:40 +08:00
if (ngx_accept_mutex_held && ngx_accept_events == 0) {
nginx-0.3.8-RELEASE import *) Security: nginx now checks URI got from a backend in "X-Accel-Redirect" header line or in SSI file for the "/../" paths and zeroes. *) Change: nginx now does not treat the empty user name in the "Authorization" header line as valid one. *) Feature: the "ssl_session_timeout" directives of the ngx_http_ssl_module and ngx_imap_ssl_module. *) Feature: the "auth_http_header" directive of the ngx_imap_auth_http_module. *) Feature: the "add_header" directive. *) Feature: the ngx_http_realip_module. *) Feature: the new variables to use in the "log_format" directive: $bytes_sent, $apache_bytes_sent, $status, $time_gmt, $uri, $request_time, $request_length, $upstream_status, $upstream_response_time, $gzip_ratio, $uid_got, $uid_set, $connection, $pipe, and $msec. The parameters in the "%name" form will be canceled soon. *) Change: now the false variable values in the "if" directive are the empty string "" and string starting with "0". *) Bugfix: while using proxied or FastCGI-server nginx may leave connections and temporary files with client requests in open state. *) Bugfix: the worker processes did not flush the buffered logs on graceful exit. *) Bugfix: if the request URI was changes by the "rewrite" directive and the request was proxied in location given by regular expression, then the incorrect request was transferred to backend; the bug had appeared in 0.2.6. *) Bugfix: the "expires" directive did not remove the previous "Expires" header. *) Bugfix: nginx may stop to accept requests if the "rtsig" method and several worker processes were used. *) Bugfix: the "\"" and "\'" escape symbols were incorrectly handled in SSI commands. *) Bugfix: if the response was ended just after the SSI command and gzipping was used, then the response did not transferred complete or did not transferred at all.
2005-11-10 01:25:55 +08:00
return NGX_OK;
}
nginx-0.0.3-2004-03-31-19:26:46 import
2004-03-31 23:26:46 +08:00
nginx-0.3.8-RELEASE import *) Security: nginx now checks URI got from a backend in "X-Accel-Redirect" header line or in SSI file for the "/../" paths and zeroes. *) Change: nginx now does not treat the empty user name in the "Authorization" header line as valid one. *) Feature: the "ssl_session_timeout" directives of the ngx_http_ssl_module and ngx_imap_ssl_module. *) Feature: the "auth_http_header" directive of the ngx_imap_auth_http_module. *) Feature: the "add_header" directive. *) Feature: the ngx_http_realip_module. *) Feature: the new variables to use in the "log_format" directive: $bytes_sent, $apache_bytes_sent, $status, $time_gmt, $uri, $request_time, $request_length, $upstream_status, $upstream_response_time, $gzip_ratio, $uid_got, $uid_set, $connection, $pipe, and $msec. The parameters in the "%name" form will be canceled soon. *) Change: now the false variable values in the "if" directive are the empty string "" and string starting with "0". *) Bugfix: while using proxied or FastCGI-server nginx may leave connections and temporary files with client requests in open state. *) Bugfix: the worker processes did not flush the buffered logs on graceful exit. *) Bugfix: if the request URI was changes by the "rewrite" directive and the request was proxied in location given by regular expression, then the incorrect request was transferred to backend; the bug had appeared in 0.2.6. *) Bugfix: the "expires" directive did not remove the previous "Expires" header. *) Bugfix: nginx may stop to accept requests if the "rtsig" method and several worker processes were used. *) Bugfix: the "\"" and "\'" escape symbols were incorrectly handled in SSI commands. *) Bugfix: if the response was ended just after the SSI command and gzipping was used, then the response did not transferred complete or did not transferred at all.
2005-11-10 01:25:55 +08:00
if (ngx_enable_accept_events(cycle) == NGX_ERROR) {
nginx-0.3.27-RELEASE import *) Change: the "variables_hash_max_size" and "variables_hash_bucket_size" directives. *) Feature: the $body_bytes_sent variable can be used not only in the "log_format" directive. *) Feature: the $ssl_protocol and $ssl_cipher variables. *) Feature: the cache line size detection for widespread CPUs at start time. *) Feature: now the "accept_mutex" directive is supported using fcntl(2) on platforms different from i386, amd64, sparc64, and ppc. *) Feature: the "lock_file" directive and the --with-lock-path=PATH autoconfiguration directive. *) Bugfix: if the HTTPS protocol was used in the "proxy_pass" directive then the requests with the body was not transferred.
2006-02-08 23:33:12 +08:00
ngx_shmtx_unlock(&ngx_accept_mutex);
nginx-0.3.8-RELEASE import *) Security: nginx now checks URI got from a backend in "X-Accel-Redirect" header line or in SSI file for the "/../" paths and zeroes. *) Change: nginx now does not treat the empty user name in the "Authorization" header line as valid one. *) Feature: the "ssl_session_timeout" directives of the ngx_http_ssl_module and ngx_imap_ssl_module. *) Feature: the "auth_http_header" directive of the ngx_imap_auth_http_module. *) Feature: the "add_header" directive. *) Feature: the ngx_http_realip_module. *) Feature: the new variables to use in the "log_format" directive: $bytes_sent, $apache_bytes_sent, $status, $time_gmt, $uri, $request_time, $request_length, $upstream_status, $upstream_response_time, $gzip_ratio, $uid_got, $uid_set, $connection, $pipe, and $msec. The parameters in the "%name" form will be canceled soon. *) Change: now the false variable values in the "if" directive are the empty string "" and string starting with "0". *) Bugfix: while using proxied or FastCGI-server nginx may leave connections and temporary files with client requests in open state. *) Bugfix: the worker processes did not flush the buffered logs on graceful exit. *) Bugfix: if the request URI was changes by the "rewrite" directive and the request was proxied in location given by regular expression, then the incorrect request was transferred to backend; the bug had appeared in 0.2.6. *) Bugfix: the "expires" directive did not remove the previous "Expires" header. *) Bugfix: nginx may stop to accept requests if the "rtsig" method and several worker processes were used. *) Bugfix: the "\"" and "\'" escape symbols were incorrectly handled in SSI commands. *) Bugfix: if the response was ended just after the SSI command and gzipping was used, then the response did not transferred complete or did not transferred at all.
2005-11-10 01:25:55 +08:00
return NGX_ERROR;
nginx-0.0.3-2004-03-31-19:26:46 import
2004-03-31 23:26:46 +08:00
}
Solaris 10 event ports support
2006-09-26 20:20:12 +08:00
ngx_accept_events = 0;
nginx-0.3.8-RELEASE import *) Security: nginx now checks URI got from a backend in "X-Accel-Redirect" header line or in SSI file for the "/../" paths and zeroes. *) Change: nginx now does not treat the empty user name in the "Authorization" header line as valid one. *) Feature: the "ssl_session_timeout" directives of the ngx_http_ssl_module and ngx_imap_ssl_module. *) Feature: the "auth_http_header" directive of the ngx_imap_auth_http_module. *) Feature: the "add_header" directive. *) Feature: the ngx_http_realip_module. *) Feature: the new variables to use in the "log_format" directive: $bytes_sent, $apache_bytes_sent, $status, $time_gmt, $uri, $request_time, $request_length, $upstream_status, $upstream_response_time, $gzip_ratio, $uid_got, $uid_set, $connection, $pipe, and $msec. The parameters in the "%name" form will be canceled soon. *) Change: now the false variable values in the "if" directive are the empty string "" and string starting with "0". *) Bugfix: while using proxied or FastCGI-server nginx may leave connections and temporary files with client requests in open state. *) Bugfix: the worker processes did not flush the buffered logs on graceful exit. *) Bugfix: if the request URI was changes by the "rewrite" directive and the request was proxied in location given by regular expression, then the incorrect request was transferred to backend; the bug had appeared in 0.2.6. *) Bugfix: the "expires" directive did not remove the previous "Expires" header. *) Bugfix: nginx may stop to accept requests if the "rtsig" method and several worker processes were used. *) Bugfix: the "\"" and "\'" escape symbols were incorrectly handled in SSI commands. *) Bugfix: if the response was ended just after the SSI command and gzipping was used, then the response did not transferred complete or did not transferred at all.
2005-11-10 01:25:55 +08:00
ngx_accept_mutex_held = 1;
nginx-0.0.3-2004-03-31-19:26:46 import
2004-03-31 23:26:46 +08:00
return NGX_OK;
}
Solaris 10 event ports support
2006-09-26 20:20:12 +08:00
ngx_log_debug1(NGX_LOG_DEBUG_EVENT, cycle->log, 0,
"accept mutex lock failed: %ui", ngx_accept_mutex_held);
nginx-0.0.3-2004-04-01-10:21:13 import
2004-04-01 14:21:13 +08:00
if (ngx_accept_mutex_held) {
The "reuseport" option of the "listen" directive. When configured, an individual listen socket on a given address is created for each worker process. This allows to reduce in-kernel lock contention on configurations with high accept rates, resulting in better performance. As of now it works on Linux and DragonFly BSD. Note that on Linux incoming connection requests are currently tied up to a specific listen socket, and if some sockets are closed, connection requests will be reset, see https://lwn.net/Articles/542629/. With nginx, this may happen if the number of worker processes is reduced. There is no such problem on DragonFly BSD. Based on previous work by Sepherosa Ziehau and Yingqi Lu.
2015-05-20 20:51:56 +08:00
if (ngx_disable_accept_events(cycle, 0) == NGX_ERROR) {
nginx-0.0.3-2004-03-31-19:26:46 import
2004-03-31 23:26:46 +08:00
return NGX_ERROR;
}
nginx-0.0.3-2004-04-01-10:21:13 import
2004-04-01 14:21:13 +08:00
ngx_accept_mutex_held = 0;
nginx-0.0.3-2004-03-31-19:26:46 import
2004-03-31 23:26:46 +08:00
}
return NGX_OK;
}
nginx-0.3.8-RELEASE import *) Security: nginx now checks URI got from a backend in "X-Accel-Redirect" header line or in SSI file for the "/../" paths and zeroes. *) Change: nginx now does not treat the empty user name in the "Authorization" header line as valid one. *) Feature: the "ssl_session_timeout" directives of the ngx_http_ssl_module and ngx_imap_ssl_module. *) Feature: the "auth_http_header" directive of the ngx_imap_auth_http_module. *) Feature: the "add_header" directive. *) Feature: the ngx_http_realip_module. *) Feature: the new variables to use in the "log_format" directive: $bytes_sent, $apache_bytes_sent, $status, $time_gmt, $uri, $request_time, $request_length, $upstream_status, $upstream_response_time, $gzip_ratio, $uid_got, $uid_set, $connection, $pipe, and $msec. The parameters in the "%name" form will be canceled soon. *) Change: now the false variable values in the "if" directive are the empty string "" and string starting with "0". *) Bugfix: while using proxied or FastCGI-server nginx may leave connections and temporary files with client requests in open state. *) Bugfix: the worker processes did not flush the buffered logs on graceful exit. *) Bugfix: if the request URI was changes by the "rewrite" directive and the request was proxied in location given by regular expression, then the incorrect request was transferred to backend; the bug had appeared in 0.2.6. *) Bugfix: the "expires" directive did not remove the previous "Expires" header. *) Bugfix: nginx may stop to accept requests if the "rtsig" method and several worker processes were used. *) Bugfix: the "\"" and "\'" escape symbols were incorrectly handled in SSI commands. *) Bugfix: if the response was ended just after the SSI command and gzipping was used, then the response did not transferred complete or did not transferred at all.
2005-11-10 01:25:55 +08:00
static ngx_int_t
nginx-0.1.21-RELEASE import *) Bugfix: the ngx_http_stub_status_module showed incorrect statistics if "rtsig" method was used or if several worker process ran on SMP. *) Bugfix: nginx could not be built by the icc compiler on Linux or if the zlib-1.2.x library was building from sources. *) Bugfix: nginx could not be built on NetBSD 2.0.
2005-02-22 22:40:13 +08:00
ngx_enable_accept_events(ngx_cycle_t *cycle)
nginx-0.0.3-2004-03-31-19:26:46 import
2004-03-31 23:26:46 +08:00
{
nginx-0.2.0-RELEASE import *) The pid-file names used during online upgrade was changed and now is not required a manual rename operation. The old master process adds the ".oldbin" suffix to its pid-file and executes a new binary file. The new master process creates usual pid-file without the ".newbin" suffix. If the master process exits, then old master process renames back its pid-file with the ".oldbin" suffix to the pid-file without suffix. *) Change: the "worker_connections" directive, new name of the "connections" directive; now the directive specifies maximum number of connections, but not maximum socket descriptor number. *) Feature: SSL supports the session cache inside one worker process. *) Feature: the "satisfy_any" directive. *) Change: the ngx_http_access_module and ngx_http_auth_basic_module do not run for subrequests. *) Feature: the "worker_rlimit_nofile" and "worker_rlimit_sigpending" directives. *) Bugfix: if all backend using in load-balancing failed after one error, then nginx did not try do connect to them during 60 seconds. *) Bugfix: in IMAP/POP3 command argument parsing. Thanks to Rob Mueller. *) Bugfix: errors while using SSL in IMAP/POP3 proxy. *) Bugfix: errors while using SSI and gzipping. *) Bugfix: the "Expires" and "Cache-Control" header lines were omitted from the 304 responses. Thanks to Alexandr Kukushkin.
2005-09-23 19:02:22 +08:00
ngx_uint_t i;
ngx_listening_t *ls;
ngx_connection_t *c;
nginx-0.0.3-2004-03-31-19:26:46 import
2004-03-31 23:26:46 +08:00
nginx-0.2.0-RELEASE import *) The pid-file names used during online upgrade was changed and now is not required a manual rename operation. The old master process adds the ".oldbin" suffix to its pid-file and executes a new binary file. The new master process creates usual pid-file without the ".newbin" suffix. If the master process exits, then old master process renames back its pid-file with the ".oldbin" suffix to the pid-file without suffix. *) Change: the "worker_connections" directive, new name of the "connections" directive; now the directive specifies maximum number of connections, but not maximum socket descriptor number. *) Feature: SSL supports the session cache inside one worker process. *) Feature: the "satisfy_any" directive. *) Change: the ngx_http_access_module and ngx_http_auth_basic_module do not run for subrequests. *) Feature: the "worker_rlimit_nofile" and "worker_rlimit_sigpending" directives. *) Bugfix: if all backend using in load-balancing failed after one error, then nginx did not try do connect to them during 60 seconds. *) Bugfix: in IMAP/POP3 command argument parsing. Thanks to Rob Mueller. *) Bugfix: errors while using SSL in IMAP/POP3 proxy. *) Bugfix: errors while using SSI and gzipping. *) Bugfix: the "Expires" and "Cache-Control" header lines were omitted from the 304 responses. Thanks to Alexandr Kukushkin.
2005-09-23 19:02:22 +08:00
ls = cycle->listening.elts;
nginx-0.0.3-2004-03-31-19:26:46 import
2004-03-31 23:26:46 +08:00
for (i = 0; i < cycle->listening.nelts; i++) {
nginx-0.2.0-RELEASE import *) The pid-file names used during online upgrade was changed and now is not required a manual rename operation. The old master process adds the ".oldbin" suffix to its pid-file and executes a new binary file. The new master process creates usual pid-file without the ".newbin" suffix. If the master process exits, then old master process renames back its pid-file with the ".oldbin" suffix to the pid-file without suffix. *) Change: the "worker_connections" directive, new name of the "connections" directive; now the directive specifies maximum number of connections, but not maximum socket descriptor number. *) Feature: SSL supports the session cache inside one worker process. *) Feature: the "satisfy_any" directive. *) Change: the ngx_http_access_module and ngx_http_auth_basic_module do not run for subrequests. *) Feature: the "worker_rlimit_nofile" and "worker_rlimit_sigpending" directives. *) Bugfix: if all backend using in load-balancing failed after one error, then nginx did not try do connect to them during 60 seconds. *) Bugfix: in IMAP/POP3 command argument parsing. Thanks to Rob Mueller. *) Bugfix: errors while using SSL in IMAP/POP3 proxy. *) Bugfix: errors while using SSI and gzipping. *) Bugfix: the "Expires" and "Cache-Control" header lines were omitted from the 304 responses. Thanks to Alexandr Kukushkin.
2005-09-23 19:02:22 +08:00
c = ls[i].connection;
nginx-0.0.3-2004-03-31-19:26:46 import
2004-03-31 23:26:46 +08:00
The "reuseport" option of the "listen" directive. When configured, an individual listen socket on a given address is created for each worker process. This allows to reduce in-kernel lock contention on configurations with high accept rates, resulting in better performance. As of now it works on Linux and DragonFly BSD. Note that on Linux incoming connection requests are currently tied up to a specific listen socket, and if some sockets are closed, connection requests will be reset, see https://lwn.net/Articles/542629/. With nginx, this may happen if the number of worker processes is reduced. There is no such problem on DragonFly BSD. Based on previous work by Sepherosa Ziehau and Yingqi Lu.
2015-05-20 20:51:56 +08:00
if (c == NULL || c->read->active) {
Accept moderation in case of EMFILE/ENFILE. In case of EMFILE/ENFILE returned from accept() we disable accept events, and (in case of no accept mutex used) arm timer to re-enable them later. With accept mutex we just drop it, and rely on normal accept mutex handling to re-enable accept events once it's acquired again. As we now handle errors in question, logging level was changed to "crit" (instead of "alert" used for unknown errors). Note: the code might call ngx_enable_accept_events() multiple times if there are many listen sockets. The ngx_enable_accept_events() function was modified to check if connection is already active (via c->read->active) and skip it then, thus making multiple calls safe.
2012-05-11 21:33:06 +08:00
continue;
}
Removed the obsolete rtsig module.
2015-04-23 19:17:40 +08:00
if (ngx_add_event(c->read, NGX_READ_EVENT, 0) == NGX_ERROR) {
return NGX_ERROR;
nginx-0.0.3-2004-03-31-19:26:46 import
2004-03-31 23:26:46 +08:00
}
}
return NGX_OK;
}
nginx-0.3.8-RELEASE import *) Security: nginx now checks URI got from a backend in "X-Accel-Redirect" header line or in SSI file for the "/../" paths and zeroes. *) Change: nginx now does not treat the empty user name in the "Authorization" header line as valid one. *) Feature: the "ssl_session_timeout" directives of the ngx_http_ssl_module and ngx_imap_ssl_module. *) Feature: the "auth_http_header" directive of the ngx_imap_auth_http_module. *) Feature: the "add_header" directive. *) Feature: the ngx_http_realip_module. *) Feature: the new variables to use in the "log_format" directive: $bytes_sent, $apache_bytes_sent, $status, $time_gmt, $uri, $request_time, $request_length, $upstream_status, $upstream_response_time, $gzip_ratio, $uid_got, $uid_set, $connection, $pipe, and $msec. The parameters in the "%name" form will be canceled soon. *) Change: now the false variable values in the "if" directive are the empty string "" and string starting with "0". *) Bugfix: while using proxied or FastCGI-server nginx may leave connections and temporary files with client requests in open state. *) Bugfix: the worker processes did not flush the buffered logs on graceful exit. *) Bugfix: if the request URI was changes by the "rewrite" directive and the request was proxied in location given by regular expression, then the incorrect request was transferred to backend; the bug had appeared in 0.2.6. *) Bugfix: the "expires" directive did not remove the previous "Expires" header. *) Bugfix: nginx may stop to accept requests if the "rtsig" method and several worker processes were used. *) Bugfix: the "\"" and "\'" escape symbols were incorrectly handled in SSI commands. *) Bugfix: if the response was ended just after the SSI command and gzipping was used, then the response did not transferred complete or did not transferred at all.
2005-11-10 01:25:55 +08:00
static ngx_int_t
The "reuseport" option of the "listen" directive. When configured, an individual listen socket on a given address is created for each worker process. This allows to reduce in-kernel lock contention on configurations with high accept rates, resulting in better performance. As of now it works on Linux and DragonFly BSD. Note that on Linux incoming connection requests are currently tied up to a specific listen socket, and if some sockets are closed, connection requests will be reset, see https://lwn.net/Articles/542629/. With nginx, this may happen if the number of worker processes is reduced. There is no such problem on DragonFly BSD. Based on previous work by Sepherosa Ziehau and Yingqi Lu.
2015-05-20 20:51:56 +08:00
ngx_disable_accept_events(ngx_cycle_t *cycle, ngx_uint_t all)
nginx-0.0.3-2004-03-31-19:26:46 import
2004-03-31 23:26:46 +08:00
{
nginx-0.2.0-RELEASE import *) The pid-file names used during online upgrade was changed and now is not required a manual rename operation. The old master process adds the ".oldbin" suffix to its pid-file and executes a new binary file. The new master process creates usual pid-file without the ".newbin" suffix. If the master process exits, then old master process renames back its pid-file with the ".oldbin" suffix to the pid-file without suffix. *) Change: the "worker_connections" directive, new name of the "connections" directive; now the directive specifies maximum number of connections, but not maximum socket descriptor number. *) Feature: SSL supports the session cache inside one worker process. *) Feature: the "satisfy_any" directive. *) Change: the ngx_http_access_module and ngx_http_auth_basic_module do not run for subrequests. *) Feature: the "worker_rlimit_nofile" and "worker_rlimit_sigpending" directives. *) Bugfix: if all backend using in load-balancing failed after one error, then nginx did not try do connect to them during 60 seconds. *) Bugfix: in IMAP/POP3 command argument parsing. Thanks to Rob Mueller. *) Bugfix: errors while using SSL in IMAP/POP3 proxy. *) Bugfix: errors while using SSI and gzipping. *) Bugfix: the "Expires" and "Cache-Control" header lines were omitted from the 304 responses. Thanks to Alexandr Kukushkin.
2005-09-23 19:02:22 +08:00
ngx_uint_t i;
ngx_listening_t *ls;
ngx_connection_t *c;
nginx-0.0.3-2004-03-31-19:26:46 import
2004-03-31 23:26:46 +08:00
nginx-0.2.0-RELEASE import *) The pid-file names used during online upgrade was changed and now is not required a manual rename operation. The old master process adds the ".oldbin" suffix to its pid-file and executes a new binary file. The new master process creates usual pid-file without the ".newbin" suffix. If the master process exits, then old master process renames back its pid-file with the ".oldbin" suffix to the pid-file without suffix. *) Change: the "worker_connections" directive, new name of the "connections" directive; now the directive specifies maximum number of connections, but not maximum socket descriptor number. *) Feature: SSL supports the session cache inside one worker process. *) Feature: the "satisfy_any" directive. *) Change: the ngx_http_access_module and ngx_http_auth_basic_module do not run for subrequests. *) Feature: the "worker_rlimit_nofile" and "worker_rlimit_sigpending" directives. *) Bugfix: if all backend using in load-balancing failed after one error, then nginx did not try do connect to them during 60 seconds. *) Bugfix: in IMAP/POP3 command argument parsing. Thanks to Rob Mueller. *) Bugfix: errors while using SSL in IMAP/POP3 proxy. *) Bugfix: errors while using SSI and gzipping. *) Bugfix: the "Expires" and "Cache-Control" header lines were omitted from the 304 responses. Thanks to Alexandr Kukushkin.
2005-09-23 19:02:22 +08:00
ls = cycle->listening.elts;
nginx-0.0.3-2004-03-31-19:26:46 import
2004-03-31 23:26:46 +08:00
for (i = 0; i < cycle->listening.nelts; i++) {
nginx-0.2.0-RELEASE import *) The pid-file names used during online upgrade was changed and now is not required a manual rename operation. The old master process adds the ".oldbin" suffix to its pid-file and executes a new binary file. The new master process creates usual pid-file without the ".newbin" suffix. If the master process exits, then old master process renames back its pid-file with the ".oldbin" suffix to the pid-file without suffix. *) Change: the "worker_connections" directive, new name of the "connections" directive; now the directive specifies maximum number of connections, but not maximum socket descriptor number. *) Feature: SSL supports the session cache inside one worker process. *) Feature: the "satisfy_any" directive. *) Change: the ngx_http_access_module and ngx_http_auth_basic_module do not run for subrequests. *) Feature: the "worker_rlimit_nofile" and "worker_rlimit_sigpending" directives. *) Bugfix: if all backend using in load-balancing failed after one error, then nginx did not try do connect to them during 60 seconds. *) Bugfix: in IMAP/POP3 command argument parsing. Thanks to Rob Mueller. *) Bugfix: errors while using SSL in IMAP/POP3 proxy. *) Bugfix: errors while using SSI and gzipping. *) Bugfix: the "Expires" and "Cache-Control" header lines were omitted from the 304 responses. Thanks to Alexandr Kukushkin.
2005-09-23 19:02:22 +08:00
c = ls[i].connection;
nginx-0.0.3-2004-03-31-19:26:46 import
2004-03-31 23:26:46 +08:00
The "reuseport" option of the "listen" directive. When configured, an individual listen socket on a given address is created for each worker process. This allows to reduce in-kernel lock contention on configurations with high accept rates, resulting in better performance. As of now it works on Linux and DragonFly BSD. Note that on Linux incoming connection requests are currently tied up to a specific listen socket, and if some sockets are closed, connection requests will be reset, see https://lwn.net/Articles/542629/. With nginx, this may happen if the number of worker processes is reduced. There is no such problem on DragonFly BSD. Based on previous work by Sepherosa Ziehau and Yingqi Lu.
2015-05-20 20:51:56 +08:00
if (c == NULL || !c->read->active) {
nginx-0.2.0-RELEASE import *) The pid-file names used during online upgrade was changed and now is not required a manual rename operation. The old master process adds the ".oldbin" suffix to its pid-file and executes a new binary file. The new master process creates usual pid-file without the ".newbin" suffix. If the master process exits, then old master process renames back its pid-file with the ".oldbin" suffix to the pid-file without suffix. *) Change: the "worker_connections" directive, new name of the "connections" directive; now the directive specifies maximum number of connections, but not maximum socket descriptor number. *) Feature: SSL supports the session cache inside one worker process. *) Feature: the "satisfy_any" directive. *) Change: the ngx_http_access_module and ngx_http_auth_basic_module do not run for subrequests. *) Feature: the "worker_rlimit_nofile" and "worker_rlimit_sigpending" directives. *) Bugfix: if all backend using in load-balancing failed after one error, then nginx did not try do connect to them during 60 seconds. *) Bugfix: in IMAP/POP3 command argument parsing. Thanks to Rob Mueller. *) Bugfix: errors while using SSL in IMAP/POP3 proxy. *) Bugfix: errors while using SSI and gzipping. *) Bugfix: the "Expires" and "Cache-Control" header lines were omitted from the 304 responses. Thanks to Alexandr Kukushkin.
2005-09-23 19:02:22 +08:00
continue;
}
nginx-0.0.3-2004-04-14-21:44:28 import
2004-04-15 01:44:28 +08:00
The "reuseport" option of the "listen" directive. When configured, an individual listen socket on a given address is created for each worker process. This allows to reduce in-kernel lock contention on configurations with high accept rates, resulting in better performance. As of now it works on Linux and DragonFly BSD. Note that on Linux incoming connection requests are currently tied up to a specific listen socket, and if some sockets are closed, connection requests will be reset, see https://lwn.net/Articles/542629/. With nginx, this may happen if the number of worker processes is reduced. There is no such problem on DragonFly BSD. Based on previous work by Sepherosa Ziehau and Yingqi Lu.
2015-05-20 20:51:56 +08:00
#if (NGX_HAVE_REUSEPORT)
/*
* do not disable accept on worker's own sockets
* when disabling accept events due to accept mutex
*/
if (ls[i].reuseport && !all) {
continue;
}
#endif
Removed the obsolete rtsig module.
2015-04-23 19:17:40 +08:00
if (ngx_del_event(c->read, NGX_READ_EVENT, NGX_DISABLE_EVENT)
== NGX_ERROR)
{
return NGX_ERROR;
nginx-0.0.3-2004-03-31-19:26:46 import
2004-03-31 23:26:46 +08:00
}
}
return NGX_OK;
}
nginx-0.1.21-RELEASE import *) Bugfix: the ngx_http_stub_status_module showed incorrect statistics if "rtsig" method was used or if several worker process ran on SMP. *) Bugfix: nginx could not be built by the icc compiler on Linux or if the zlib-1.2.x library was building from sources. *) Bugfix: nginx could not be built on NetBSD 2.0.
2005-02-22 22:40:13 +08:00
static void
nginx-0.2.0-RELEASE import *) The pid-file names used during online upgrade was changed and now is not required a manual rename operation. The old master process adds the ".oldbin" suffix to its pid-file and executes a new binary file. The new master process creates usual pid-file without the ".newbin" suffix. If the master process exits, then old master process renames back its pid-file with the ".oldbin" suffix to the pid-file without suffix. *) Change: the "worker_connections" directive, new name of the "connections" directive; now the directive specifies maximum number of connections, but not maximum socket descriptor number. *) Feature: SSL supports the session cache inside one worker process. *) Feature: the "satisfy_any" directive. *) Change: the ngx_http_access_module and ngx_http_auth_basic_module do not run for subrequests. *) Feature: the "worker_rlimit_nofile" and "worker_rlimit_sigpending" directives. *) Bugfix: if all backend using in load-balancing failed after one error, then nginx did not try do connect to them during 60 seconds. *) Bugfix: in IMAP/POP3 command argument parsing. Thanks to Rob Mueller. *) Bugfix: errors while using SSL in IMAP/POP3 proxy. *) Bugfix: errors while using SSI and gzipping. *) Bugfix: the "Expires" and "Cache-Control" header lines were omitted from the 304 responses. Thanks to Alexandr Kukushkin.
2005-09-23 19:02:22 +08:00
ngx_close_accepted_connection(ngx_connection_t *c)
nginx-0.0.10-2004-09-09-19:40:48 import
2004-09-09 23:40:48 +08:00
{
nginx-0.2.0-RELEASE import *) The pid-file names used during online upgrade was changed and now is not required a manual rename operation. The old master process adds the ".oldbin" suffix to its pid-file and executes a new binary file. The new master process creates usual pid-file without the ".newbin" suffix. If the master process exits, then old master process renames back its pid-file with the ".oldbin" suffix to the pid-file without suffix. *) Change: the "worker_connections" directive, new name of the "connections" directive; now the directive specifies maximum number of connections, but not maximum socket descriptor number. *) Feature: SSL supports the session cache inside one worker process. *) Feature: the "satisfy_any" directive. *) Change: the ngx_http_access_module and ngx_http_auth_basic_module do not run for subrequests. *) Feature: the "worker_rlimit_nofile" and "worker_rlimit_sigpending" directives. *) Bugfix: if all backend using in load-balancing failed after one error, then nginx did not try do connect to them during 60 seconds. *) Bugfix: in IMAP/POP3 command argument parsing. Thanks to Rob Mueller. *) Bugfix: errors while using SSL in IMAP/POP3 proxy. *) Bugfix: errors while using SSI and gzipping. *) Bugfix: the "Expires" and "Cache-Control" header lines were omitted from the 304 responses. Thanks to Alexandr Kukushkin.
2005-09-23 19:02:22 +08:00
ngx_socket_t fd;
ngx_free_connection(c);
fd = c->fd;
c->fd = (ngx_socket_t) -1;
Stream: UDP proxy.
2016-01-21 00:52:12 +08:00
if (!c->shared && ngx_close_socket(fd) == -1) {
nginx-0.2.0-RELEASE import *) The pid-file names used during online upgrade was changed and now is not required a manual rename operation. The old master process adds the ".oldbin" suffix to its pid-file and executes a new binary file. The new master process creates usual pid-file without the ".newbin" suffix. If the master process exits, then old master process renames back its pid-file with the ".oldbin" suffix to the pid-file without suffix. *) Change: the "worker_connections" directive, new name of the "connections" directive; now the directive specifies maximum number of connections, but not maximum socket descriptor number. *) Feature: SSL supports the session cache inside one worker process. *) Feature: the "satisfy_any" directive. *) Change: the ngx_http_access_module and ngx_http_auth_basic_module do not run for subrequests. *) Feature: the "worker_rlimit_nofile" and "worker_rlimit_sigpending" directives. *) Bugfix: if all backend using in load-balancing failed after one error, then nginx did not try do connect to them during 60 seconds. *) Bugfix: in IMAP/POP3 command argument parsing. Thanks to Rob Mueller. *) Bugfix: errors while using SSL in IMAP/POP3 proxy. *) Bugfix: errors while using SSI and gzipping. *) Bugfix: the "Expires" and "Cache-Control" header lines were omitted from the 304 responses. Thanks to Alexandr Kukushkin.
2005-09-23 19:02:22 +08:00
ngx_log_error(NGX_LOG_ALERT, c->log, ngx_socket_errno,
nginx-0.0.10-2004-09-09-19:40:48 import
2004-09-09 23:40:48 +08:00
ngx_close_socket_n " failed");
}
nginx-0.1.22-RELEASE import *) Bugfix: the ngx_http_stub_status_module showed incorrect handled connections statistics if the proxying or FastCGI server were used. *) Bugfix: the installation paths were incorrectly quoted on Linux and Solaris; the bug had appeared in 0.1.21.
2005-02-24 20:29:09 +08:00
nginx-0.2.0-RELEASE import *) The pid-file names used during online upgrade was changed and now is not required a manual rename operation. The old master process adds the ".oldbin" suffix to its pid-file and executes a new binary file. The new master process creates usual pid-file without the ".newbin" suffix. If the master process exits, then old master process renames back its pid-file with the ".oldbin" suffix to the pid-file without suffix. *) Change: the "worker_connections" directive, new name of the "connections" directive; now the directive specifies maximum number of connections, but not maximum socket descriptor number. *) Feature: SSL supports the session cache inside one worker process. *) Feature: the "satisfy_any" directive. *) Change: the ngx_http_access_module and ngx_http_auth_basic_module do not run for subrequests. *) Feature: the "worker_rlimit_nofile" and "worker_rlimit_sigpending" directives. *) Bugfix: if all backend using in load-balancing failed after one error, then nginx did not try do connect to them during 60 seconds. *) Bugfix: in IMAP/POP3 command argument parsing. Thanks to Rob Mueller. *) Bugfix: errors while using SSL in IMAP/POP3 proxy. *) Bugfix: errors while using SSI and gzipping. *) Bugfix: the "Expires" and "Cache-Control" header lines were omitted from the 304 responses. Thanks to Alexandr Kukushkin.
2005-09-23 19:02:22 +08:00
if (c->pool) {
ngx_destroy_pool(c->pool);
}
nginx-0.1.22-RELEASE import *) Bugfix: the ngx_http_stub_status_module showed incorrect handled connections statistics if the proxying or FastCGI server were used. *) Bugfix: the installation paths were incorrectly quoted on Linux and Solaris; the bug had appeared in 0.1.21.
2005-02-24 20:29:09 +08:00
#if (NGX_STAT_STUB)
ignore ngx_atomic_fetch_add() result this fixes building at least by gcc 4.2.1 on Mac OS X 10.6
2009-06-18 21:14:51 +08:00
(void) ngx_atomic_fetch_add(ngx_stat_active, -1);
nginx-0.1.22-RELEASE import *) Bugfix: the ngx_http_stub_status_module showed incorrect handled connections statistics if the proxying or FastCGI server were used. *) Bugfix: the installation paths were incorrectly quoted on Linux and Solaris; the bug had appeared in 0.1.21.
2005-02-24 20:29:09 +08:00
#endif
nginx-0.0.10-2004-09-09-19:40:48 import
2004-09-09 23:40:48 +08:00
}
nginx-0.2.0-RELEASE import *) The pid-file names used during online upgrade was changed and now is not required a manual rename operation. The old master process adds the ".oldbin" suffix to its pid-file and executes a new binary file. The new master process creates usual pid-file without the ".newbin" suffix. If the master process exits, then old master process renames back its pid-file with the ".oldbin" suffix to the pid-file without suffix. *) Change: the "worker_connections" directive, new name of the "connections" directive; now the directive specifies maximum number of connections, but not maximum socket descriptor number. *) Feature: SSL supports the session cache inside one worker process. *) Feature: the "satisfy_any" directive. *) Change: the ngx_http_access_module and ngx_http_auth_basic_module do not run for subrequests. *) Feature: the "worker_rlimit_nofile" and "worker_rlimit_sigpending" directives. *) Bugfix: if all backend using in load-balancing failed after one error, then nginx did not try do connect to them during 60 seconds. *) Bugfix: in IMAP/POP3 command argument parsing. Thanks to Rob Mueller. *) Bugfix: errors while using SSL in IMAP/POP3 proxy. *) Bugfix: errors while using SSI and gzipping. *) Bugfix: the "Expires" and "Cache-Control" header lines were omitted from the 304 responses. Thanks to Alexandr Kukushkin.
2005-09-23 19:02:22 +08:00
u_char *
nginx-0.1.21-RELEASE import *) Bugfix: the ngx_http_stub_status_module showed incorrect statistics if "rtsig" method was used or if several worker process ran on SMP. *) Bugfix: nginx could not be built by the icc compiler on Linux or if the zlib-1.2.x library was building from sources. *) Bugfix: nginx could not be built on NetBSD 2.0.
2005-02-22 22:40:13 +08:00
ngx_accept_log_error(ngx_log_t *log, u_char *buf, size_t len)
nginx-0.0.1-2003-11-19-19:26:41 import
2003-11-20 00:26:41 +08:00
{
nginx-0.2.0-RELEASE import *) The pid-file names used during online upgrade was changed and now is not required a manual rename operation. The old master process adds the ".oldbin" suffix to its pid-file and executes a new binary file. The new master process creates usual pid-file without the ".newbin" suffix. If the master process exits, then old master process renames back its pid-file with the ".oldbin" suffix to the pid-file without suffix. *) Change: the "worker_connections" directive, new name of the "connections" directive; now the directive specifies maximum number of connections, but not maximum socket descriptor number. *) Feature: SSL supports the session cache inside one worker process. *) Feature: the "satisfy_any" directive. *) Change: the ngx_http_access_module and ngx_http_auth_basic_module do not run for subrequests. *) Feature: the "worker_rlimit_nofile" and "worker_rlimit_sigpending" directives. *) Bugfix: if all backend using in load-balancing failed after one error, then nginx did not try do connect to them during 60 seconds. *) Bugfix: in IMAP/POP3 command argument parsing. Thanks to Rob Mueller. *) Bugfix: errors while using SSL in IMAP/POP3 proxy. *) Bugfix: errors while using SSI and gzipping. *) Bugfix: the "Expires" and "Cache-Control" header lines were omitted from the 304 responses. Thanks to Alexandr Kukushkin.
2005-09-23 19:02:22 +08:00
return ngx_snprintf(buf, len, " while accepting new connection on %V",
log->data);
nginx-0.0.1-2003-11-19-19:26:41 import
2003-11-20 00:26:41 +08:00
}
Stream: UDP proxy.
2016-01-21 00:52:12 +08:00
#if (NGX_DEBUG)
static void
ngx_debug_accepted_connection(ngx_event_conf_t *ecf, ngx_connection_t *c)
{
struct sockaddr_in *sin;
ngx_cidr_t *cidr;
ngx_uint_t i;
#if (NGX_HAVE_INET6)
struct sockaddr_in6 *sin6;
ngx_uint_t n;
#endif
cidr = ecf->debug_connection.elts;
for (i = 0; i < ecf->debug_connection.nelts; i++) {
if (cidr[i].family != (ngx_uint_t) c->sockaddr->sa_family) {
goto next;
}
switch (cidr[i].family) {
#if (NGX_HAVE_INET6)
case AF_INET6:
sin6 = (struct sockaddr_in6 *) c->sockaddr;
for (n = 0; n < 16; n++) {
if ((sin6->sin6_addr.s6_addr[n]
& cidr[i].u.in6.mask.s6_addr[n])
!= cidr[i].u.in6.addr.s6_addr[n])
{
goto next;
}
}
break;
#endif
#if (NGX_HAVE_UNIX_DOMAIN)
case AF_UNIX:
break;
#endif
default: /* AF_INET */
sin = (struct sockaddr_in *) c->sockaddr;
if ((sin->sin_addr.s_addr & cidr[i].u.in.mask)
!= cidr[i].u.in.addr)
{
goto next;
}
break;
}
c->log->log_level = NGX_LOG_DEBUG_CONNECTION|NGX_LOG_DEBUG_ALL;
break;
next:
continue;
}
}
#endif
Reference in New Issue Copy Permalink