Mirrors/libwebsockets
1
0
Fork 0
mirror of https://github.com/warmcat/libwebsockets.git synced 2025-03-16 00:00:07 +01:00
Code Issues Releases Wiki Activity
libwebsockets/lib/libwebsockets.c

3049 lines
71 KiB
C
Raw Normal View History

remove bogus bytes on first line of websockets.c Reported-by: Pierre-Paul Lavoie <ppl@idios.org> Signed-off-by: Andy Green <andy@warmcat.com>
2011-03-07 17:54:06 +00:00
/*
add-COPYING-lgpl-2.1.patch Signed-off-by: Andy Green <andy@warmcat.com>
2010-11-08 17:12:19 +00:00
* libwebsockets - small server side websockets and web server implementation
clean-whitespace.patch Signed-off-by: Andy Green <andy@warmcat.com>
2010-12-19 22:13:26 +00:00
*
clean
2017-09-23 12:55:21 +08:00
* Copyright (C) 2010-2017 Andy Green <andy@warmcat.com>
add-COPYING-lgpl-2.1.patch Signed-off-by: Andy Green <andy@warmcat.com>
2010-11-08 17:12:19 +00:00
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation:
* version 2.1 of the License.
*
* This library is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with this library; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston,
* MA 02110-1301 USA
add-kernel-doc-script-and-generated-api.txt Signed-off-by: Andy Green <andy@warmcat.com>
2010-10-31 17:51:39 +00:00
*/
break-out-lib-sources.patch Signed-off-by: Andy Green <andy@warmcat.com>
2010-11-08 20:20:42 +00:00
#include "private-libwebsockets.h"
wince minor adaptations https://github.com/warmcat/libwebsockets/issues/444
2016-02-29 18:48:55 +08:00
#ifdef LWS_HAVE_SYS_TYPES_H
cgi Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-21 21:25:48 +08:00
#include <sys/types.h>
wince minor adaptations https://github.com/warmcat/libwebsockets/issues/444
2016-02-29 18:48:55 +08:00
#endif
CMake: convert all LWS_USE_... to LWS_WITH_... Almost all the CMake options begin with LWS_WITH_..., but many of the symbols passed to lws are LWS_USE_... , this causes neededless confusion, compounded by the fact that a few CMake options also begin with LWS_USE_. This patch globally converts all LWS_USE_... to LWS_WITH_..., so there is only one prefix to remember in both CMake and the code. The affected public CMake options are LWS_USE_BORINGSSL -> LWS_WITH_BORINGSSL LWS_USE_CYASSL -> LWS_WITH_CYASSL LWS_USE_WOLFSSL -> LWS_WITH_WOLFSSL LWS_USE_MBEDTLS -> LWS_WITH_MBEDTLS LWS_USE_BUNDLED_ZLIB -> LWS_WITH_BUNDLED_ZLIB
2017-09-28 11:29:03 +08:00
#ifdef LWS_WITH_IPV6
ipv6-allow-binding-to-ipv6-address-in-iface ipv4 and ipv6 binding to a named interface works OK. ipv4 binding to an IP also works, but we need some extra ipv6 magic to identify the ipv6 interface from an ipv6 address. This patch based on code from "user3546716" at http://stackoverflow.com/questions/13504934/binding-sockets-to-ipv6-addresses adds the necessary magic. https://github.com/warmcat/libwebsockets/issues/717
2016-12-15 09:58:20 +08:00
#if defined(WIN32) || defined(_WIN32)
#include <Iphlpapi.h>
#else
#include <net/if.h>
#endif
#endif
refactor context related functions into context.c Signed-off-by: Andy Green <andy.green@linaro.org>
2014-04-03 07:29:50 +08:00
int log_level = LLL_ERR | LLL_WARN | LLL_NOTICE;
plat-optee and boringssl adaptations
2017-01-17 07:01:02 +08:00
static void (*lwsl_emit)(int level, const char *line)
#ifndef LWS_PLAT_OPTEE
= lwsl_emit_stderr
#endif
;
#ifndef LWS_PLAT_OPTEE
style cleanup This brings the library sources into compliance with checkpatch style except for three or four exceptions like WIN32 related stuff and one long string constant I don't want to break into multiple sprintf calls. There should be no functional or compilability change from all this (hopefully). Signed-off-by: Andy Green <andy.green@linaro.org>
2013-02-11 17:13:32 +08:00
static const char * const log_level_names[] = {
introduce logging api and convert all library output to use it - multiple debug context calls lwsl_ err, warn, debug, parser, ext, client - api added to set which contexts output to stderr using a bitfield log_level - --disable-debug on configure removes all code that is not err or warn severity - err and warn contexts always output to stderr unless disabled by log_level - err and warn enabled by default in log_level Signed-off-by: Andy Green <andy@warmcat.com>
2013-01-10 19:50:35 +08:00
"ERR",
"WARN",
add lwsl_notice Signed-off-by: Andy Green <andy.green@linaro.org>
2013-01-19 12:18:07 +08:00
"NOTICE",
introduce logging api and convert all library output to use it - multiple debug context calls lwsl_ err, warn, debug, parser, ext, client - api added to set which contexts output to stderr using a bitfield log_level - --disable-debug on configure removes all code that is not err or warn severity - err and warn contexts always output to stderr unless disabled by log_level - err and warn enabled by default in log_level Signed-off-by: Andy Green <andy@warmcat.com>
2013-01-10 19:50:35 +08:00
"INFO",
"DEBUG",
"PARSER",
"HEADER",
"EXTENSION",
"CLIENT",
introduce lws_latency Signed-off-by: Andy Green <andy.green@linaro.org>
2013-01-29 12:36:17 +08:00
"LATENCY",
coverity 175437: forgot to add name to array for new USER log level
2017-02-06 10:16:45 +08:00
"USER",
"?",
"?"
introduce logging api and convert all library output to use it - multiple debug context calls lwsl_ err, warn, debug, parser, ext, client - api added to set which contexts output to stderr using a bitfield log_level - --disable-debug on configure removes all code that is not err or warn severity - err and warn contexts always output to stderr unless disabled by log_level - err and warn enabled by default in log_level Signed-off-by: Andy Green <andy@warmcat.com>
2013-01-10 19:50:35 +08:00
};
plat-optee and boringssl adaptations
2017-01-17 07:01:02 +08:00
#endif
introduce logging api and convert all library output to use it - multiple debug context calls lwsl_ err, warn, debug, parser, ext, client - api added to set which contexts output to stderr using a bitfield log_level - --disable-debug on configure removes all code that is not err or warn severity - err and warn contexts always output to stderr unless disabled by log_level - err and warn enabled by default in log_level Signed-off-by: Andy Green <andy@warmcat.com>
2013-01-10 19:50:35 +08:00
Fix potential memory leak - Got rid of ifdef _WIN32 stuff adn moved to plat_ files instead. - Also, check all calls to lws_zalloc, was potential failure on WIN32 - Made context destory enable to destroy a half inited context as well. This way I got get rid of some of the error handling complexity in libwebsocket_create_context - Added TODOs for some potential problems I see where things might be leaking and such
2015-06-25 17:51:07 +02:00
void
api rationalization: eliminate all libwebsocket[s]_ prefixes This nukes all the oldstyle prefixes except in the compatibility code. struct libwebsockets becomes struct lws too. The api docs are updated accordingly as are the READMEs that mention those apis. Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-04 11:08:32 +08:00
lws_free_wsi(struct lws *wsi)
Fix potential memory leak - Got rid of ifdef _WIN32 stuff adn moved to plat_ files instead. - Also, check all calls to lws_zalloc, was potential failure on WIN32 - Made context destory enable to destroy a half inited context as well. This way I got get rid of some of the error handling complexity in libwebsocket_create_context - Added TODOs for some potential problems I see where things might be leaking and such
2015-06-25 17:51:07 +02:00
{
pmd: autobahn fixes
2017-07-28 13:12:03 +08:00
struct lws_context_per_thread *pt;
ah pool: change to dynamic linked list For some targets like ESP32, the ah pool is mainly sitting idle wasting memory. For HTTP/2, if the client sends a series of pipelined headers on different SIDs that exist simultaneously, there is no way to stall the headers to wait for an ah, because we must read the stream for stuff like WINDOW_UPDATE on the other streams. In both these cases having the ability to free unused ah completely and allocate more dynamically if there is memory is useful, so this patch makes the ah pool an initially-empty linked list that allocates on demand up to the "max pool size" limit from the context info. When nobody wants an ah, it is freed (if someone was waiting for it, it is directly reused). For ESP32 it means no large, permanent alloc when lws starts and dynamic alloc according to how many streams the client opens, which can be controlled by SETTINGS.
2017-10-06 16:07:57 +08:00
struct allocated_headers *ah;
pmd: autobahn fixes
2017-07-28 13:12:03 +08:00
Fix potential memory leak - Got rid of ifdef _WIN32 stuff adn moved to plat_ files instead. - Also, check all calls to lws_zalloc, was potential failure on WIN32 - Made context destory enable to destroy a half inited context as well. This way I got get rid of some of the error handling complexity in libwebsocket_create_context - Added TODOs for some potential problems I see where things might be leaking and such
2015-06-25 17:51:07 +02:00
if (!wsi)
return;
esp8266 initial support
2016-07-23 14:18:25 +08:00
pmd: autobahn fixes
2017-07-28 13:12:03 +08:00
pt = &wsi->context->pt[(int)wsi->tsi];
clean
2017-09-23 12:55:21 +08:00
/*
* Protocol user data may be allocated either internally by lws
* or by specified the user. We should only free what we allocated.
*/
cleanups after api changes and mbed update Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-04 16:54:12 +08:00
if (wsi->protocol && wsi->protocol->per_session_data_size &&
wsi->user_space && !wsi->user_space_externally_allocated)
Fix potential memory leak - Got rid of ifdef _WIN32 stuff adn moved to plat_ files instead. - Also, check all calls to lws_zalloc, was potential failure on WIN32 - Made context destory enable to destroy a half inited context as well. This way I got get rid of some of the error handling complexity in libwebsocket_create_context - Added TODOs for some potential problems I see where things might be leaking and such
2015-06-25 17:51:07 +02:00
lws_free(wsi->user_space);
clean internal refactor - Mainly symbol length reduction - Whitespace clean - Code refactor for linear flow - Audit @Context for API docs vs changes Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-17 17:03:59 +08:00
lws_free_set_NULL(wsi->rxflow_buffer);
lws_free_set_NULL(wsi->trunc_alloc);
ah owns rxbuf This is intended to solve a longstanding problem with the relationship between http/1.1 keep-alive and the service loop. Ah now contain an rx buffer which is used during header processing, and the ah may not be detached from the wsi until the rx buffer is exhausted. Having the rx buffer in the ah means we can delay using the rx until a later service loop. Ah which have pending rx force POLLIN service on the wsi they are attached to automatically, so we can interleave general service / connections with draining each ah rx buffer. The possible http/1.1 situations and their dispositions are: 1) exactly one set of http headers come. After processing, the ah is detached since no pending rx left. If more headers come later, a fresh ah is aqcuired when available and the rx flow control blocks the read until then. 2) more that one whole set of headers come and we remain in http mode (no upgrade). The ah is left attached and returns to the service loop after the first set of headers. We will get forced service due to the ah having pending content (respecting flowcontrol) and process the pending rx in the ah. If we use it all up, we will detach the ah. 3) one set of http headers come with ws traffic appended. We service the headers, do the upgrade, and keep the ah until the remaining ws content is used. When we exhausted the ws traffix in the ah rx buffer, we detach the ah. Since there can be any amount of http/1.1 pipelining on a connection, and each may be expensive to service, it's now enforced there is a return to the service loop after each header set is serviced on a connection. When I added the forced service for ah with pending buffering, I added support for it to the windows plat code. However this is untested. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-15 12:37:04 +08:00
close wsi must do detatch ah flow even if no ah Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-25 21:50:49 +08:00
/* we may not have an ah, but may be on the waiting list... */
esp8266 initial support
2016-07-23 14:18:25 +08:00
lwsl_info("ah det due to close\n");
pmd: autobahn fixes
2017-07-28 13:12:03 +08:00
/* we're closing, losing some rx is OK */
lws_header_table_force_to_detachable_state(wsi);
adopt readbuf do service Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-27 11:42:22 +08:00
lws_header_table_detach(wsi, 0);
multithread stability Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-26 20:56:56 +08:00
Plugins: add ssh-base ssh server plugin
2017-10-16 16:59:57 +08:00
if (wsi->vhost->lserv_wsi == wsi)
wsi->vhost->lserv_wsi = NULL;
pmd: autobahn fixes
2017-07-28 13:12:03 +08:00
lws_pt_lock(pt);
ah pool: change to dynamic linked list For some targets like ESP32, the ah pool is mainly sitting idle wasting memory. For HTTP/2, if the client sends a series of pipelined headers on different SIDs that exist simultaneously, there is no way to stall the headers to wait for an ah, because we must read the stream for stuff like WINDOW_UPDATE on the other streams. In both these cases having the ability to free unused ah completely and allocate more dynamically if there is memory is useful, so this patch makes the ah pool an initially-empty linked list that allocates on demand up to the "max pool size" limit from the context info. When nobody wants an ah, it is freed (if someone was waiting for it, it is directly reused). For ESP32 it means no large, permanent alloc when lws starts and dynamic alloc according to how many streams the client opens, which can be controlled by SETTINGS.
2017-10-06 16:07:57 +08:00
ah = pt->ah_list;
while (ah) {
if (ah->in_use && ah->wsi == wsi) {
pmd: autobahn fixes
2017-07-28 13:12:03 +08:00
lwsl_err("%s: ah leak: wsi %p\n", __func__, wsi);
ah pool: change to dynamic linked list For some targets like ESP32, the ah pool is mainly sitting idle wasting memory. For HTTP/2, if the client sends a series of pipelined headers on different SIDs that exist simultaneously, there is no way to stall the headers to wait for an ah, because we must read the stream for stuff like WINDOW_UPDATE on the other streams. In both these cases having the ability to free unused ah completely and allocate more dynamically if there is memory is useful, so this patch makes the ah pool an initially-empty linked list that allocates on demand up to the "max pool size" limit from the context info. When nobody wants an ah, it is freed (if someone was waiting for it, it is directly reused). For ESP32 it means no large, permanent alloc when lws starts and dynamic alloc according to how many streams the client opens, which can be controlled by SETTINGS.
2017-10-06 16:07:57 +08:00
ah->in_use = 0;
ah->wsi = NULL;
pmd: autobahn fixes
2017-07-28 13:12:03 +08:00
pt->ah_count_in_use--;
ah pool: change to dynamic linked list For some targets like ESP32, the ah pool is mainly sitting idle wasting memory. For HTTP/2, if the client sends a series of pipelined headers on different SIDs that exist simultaneously, there is no way to stall the headers to wait for an ah, because we must read the stream for stuff like WINDOW_UPDATE on the other streams. In both these cases having the ability to free unused ah completely and allocate more dynamically if there is memory is useful, so this patch makes the ah pool an initially-empty linked list that allocates on demand up to the "max pool size" limit from the context info. When nobody wants an ah, it is freed (if someone was waiting for it, it is directly reused). For ESP32 it means no large, permanent alloc when lws starts and dynamic alloc according to how many streams the client opens, which can be controlled by SETTINGS.
2017-10-06 16:07:57 +08:00
break;
pmd: autobahn fixes
2017-07-28 13:12:03 +08:00
}
ah pool: change to dynamic linked list For some targets like ESP32, the ah pool is mainly sitting idle wasting memory. For HTTP/2, if the client sends a series of pipelined headers on different SIDs that exist simultaneously, there is no way to stall the headers to wait for an ah, because we must read the stream for stuff like WINDOW_UPDATE on the other streams. In both these cases having the ability to free unused ah completely and allocate more dynamically if there is memory is useful, so this patch makes the ah pool an initially-empty linked list that allocates on demand up to the "max pool size" limit from the context info. When nobody wants an ah, it is freed (if someone was waiting for it, it is directly reused). For ESP32 it means no large, permanent alloc when lws starts and dynamic alloc according to how many streams the client opens, which can be controlled by SETTINGS.
2017-10-06 16:07:57 +08:00
ah = ah->next;
pmd: autobahn fixes
2017-07-28 13:12:03 +08:00
}
LWS_WITH_PEER_LIMITS Introduces an optional, settable restriction on the number of simultaneous wsi connections and ah that a single peer IP can obtain. The default is disabled for build, and if enabled, unlimited. However when enabled at CMake, setting info.ip_limit_ah and / or info.ip_limit_wsi at context creation time will enforce the limits. Connections past the info.ip_limit_wsi limit are dropped, and connections needing an ah are forced to stay in the ah waiting list even when ah are available, while the peer is at the number of ah in info.ip_limit_ah.
2017-09-14 13:14:11 +08:00
#if defined(LWS_WITH_PEER_LIMITS)
lws_peer_track_wsi_close(wsi->context, wsi->peer);
wsi->peer = NULL;
#endif
http2: make usable HTTP/2 support is now able to serve the test server, complete with websockets, from a single vhost. - This works the same with both OpenSSL and mbedTLS. - POST is now wired up and works (also for file upload). - CGI is wired up and works. - Redirect is adapted and works - lwsws works. - URI urldecode, sanitation and argument parsing wired up for :path valgrind clean (aside from openssl-style false uninit data usage in mbedtls send occasionally) h2spec reports: $ h2spec -h 127.0.0.1 -p 7681 -t -k -o 1 ... 145 tests, 145 passed, 0 skipped, 0 failed" Incorporates: - "https://github.com/warmcat/libwebsockets/pull/1039 Fixes issue with -Werror=unused-variable flag - https://github.com/sartura/libwebsockets/commit/2c843a1395bdca5c23b40d989f508fc96aa5ba3b ssl: fix infinite loop on client cert verification failure Signed-off-by: Petar Paradzik <petar.paradzik@sartura.hr>" Caused and fixes Coverity 184887 - 184892
2017-10-13 10:33:02 +08:00
#if defined(LWS_WITH_HTTP2)
if (wsi->upgraded_to_http2 || wsi->http2_substream) {
lws_hpack_destroy_dynamic_header(wsi);
if (wsi->u.h2.h2n)
lws_free_set_NULL(wsi->u.h2.h2n);
}
#endif
pmd: autobahn fixes
2017-07-28 13:12:03 +08:00
lws_pt_unlock(pt);
close path: make sure a second time timeout and ssl buffered lists are scrubbed lws_meta children can have a different close path
2017-08-04 13:27:34 +08:00
/* since we will destroy the wsi, make absolutely sure now */
lws_ssl_remove_wsi_from_buffered_list(wsi);
lws_remove_from_timeout_list(wsi);
libuv integration This gets the libuv stuff plumbed in and working. Currently it's only workable for some service thread, and there is an isolated valgrind problem left ==28425== 128 bytes in 1 blocks are definitely lost in loss record 3 of 3 ==28425== at 0x4C28C50: malloc (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so) ==28425== by 0x4C2AB1E: realloc (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so) ==28425== by 0x58BBB27: maybe_resize (core.c:748) ==28425== by 0x58BBB27: uv__io_start (core.c:787) ==28425== by 0x58C1B80: uv__signal_loop_once_init (signal.c:225) ==28425== by 0x58C1B80: uv_signal_init (signal.c:260) ==28425== by 0x58BF7A6: uv_loop_init (loop.c:66) ==28425== by 0x4157F5: lws_uv_initloop (libuv.c:89) ==28425== by 0x405536: main (test-server-libuv.c:284) libuv wants to sign off on all libuv 'handles' that will close, and callback to do the close confirmation asynchronously. The wsi close function is adapted when libuv is in use to work with libuv accordingly and exit the uv loop the number of remaining wsi is zero. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-14 09:27:41 +08:00
wsi->context->count_wsi_allocated--;
lwsl_debug("%s: %p, remaining wsi %d\n", __func__, wsi,
wsi->context->count_wsi_allocated);
Fix potential memory leak - Got rid of ifdef _WIN32 stuff adn moved to plat_ files instead. - Also, check all calls to lws_zalloc, was potential failure on WIN32 - Made context destory enable to destroy a half inited context as well. This way I got get rid of some of the error handling complexity in libwebsocket_create_context - Added TODOs for some potential problems I see where things might be leaking and such
2015-06-25 17:51:07 +02:00
lws_free(wsi);
}
client connect: oom4 clean up timeout list https://github.com/warmcat/libwebsockets/issues/810
2017-02-21 23:38:40 +08:00
void
timeout move to doubly linked list In the case we have a lot of connections, checking them all for timeout state once a second becomes burdensome. At the moment if you have 100K connections, once a second they all get checked for timeout in a loop. This patch adds a doubly-linked list based in the context to each wsi, and only wsi with pending timeouts appear on it. At checking time, we traverse the list, which costs nothing if empty because nobody has a pending timeout. Similarly adding and removing from the list costs almost nothing since no iteration is required no matter how big the list. The extra 8 or 16 bytes in the wsi are offset a little bit by demoting .pps from int to char (save 3 bytes). And trim max act exts to 2, since we only provide one, saving 8 /16 bytes by itself if exts enabled. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-19 04:32:14 +08:00
lws_remove_from_timeout_list(struct lws *wsi)
{
multithread stability Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-26 20:56:56 +08:00
struct lws_context_per_thread *pt = &wsi->context->pt[(int)wsi->tsi];
improve timeout and ah list comments Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-25 21:39:01 +08:00
if (!wsi->timeout_list_prev) /* ie, not part of the list */
timeout move to doubly linked list In the case we have a lot of connections, checking them all for timeout state once a second becomes burdensome. At the moment if you have 100K connections, once a second they all get checked for timeout in a loop. This patch adds a doubly-linked list based in the context to each wsi, and only wsi with pending timeouts appear on it. At checking time, we traverse the list, which costs nothing if empty because nobody has a pending timeout. Similarly adding and removing from the list costs almost nothing since no iteration is required no matter how big the list. The extra 8 or 16 bytes in the wsi are offset a little bit by demoting .pps from int to char (save 3 bytes). And trim max act exts to 2, since we only provide one, saving 8 /16 bytes by itself if exts enabled. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-19 04:32:14 +08:00
return;
multithread stability Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-26 20:56:56 +08:00
lws_pt_lock(pt);
improve timeout and ah list comments Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-25 21:39:01 +08:00
/* if we have a next guy, set his prev to our prev */
timeout also fixup for0middle guy deletion case Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-19 23:11:39 +08:00
if (wsi->timeout_list)
wsi->timeout_list->timeout_list_prev = wsi->timeout_list_prev;
improve timeout and ah list comments Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-25 21:39:01 +08:00
/* set our prev guy to our next guy instead of us */
timeout move to doubly linked list In the case we have a lot of connections, checking them all for timeout state once a second becomes burdensome. At the moment if you have 100K connections, once a second they all get checked for timeout in a loop. This patch adds a doubly-linked list based in the context to each wsi, and only wsi with pending timeouts appear on it. At checking time, we traverse the list, which costs nothing if empty because nobody has a pending timeout. Similarly adding and removing from the list costs almost nothing since no iteration is required no matter how big the list. The extra 8 or 16 bytes in the wsi are offset a little bit by demoting .pps from int to char (save 3 bytes). And trim max act exts to 2, since we only provide one, saving 8 /16 bytes by itself if exts enabled. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-19 04:32:14 +08:00
*wsi->timeout_list_prev = wsi->timeout_list;
timeout also fixup for0middle guy deletion case Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-19 23:11:39 +08:00
improve timeout and ah list comments Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-25 21:39:01 +08:00
/* we're out of the list, we should not point anywhere any more */
timeout move to doubly linked list In the case we have a lot of connections, checking them all for timeout state once a second becomes burdensome. At the moment if you have 100K connections, once a second they all get checked for timeout in a loop. This patch adds a doubly-linked list based in the context to each wsi, and only wsi with pending timeouts appear on it. At checking time, we traverse the list, which costs nothing if empty because nobody has a pending timeout. Similarly adding and removing from the list costs almost nothing since no iteration is required no matter how big the list. The extra 8 or 16 bytes in the wsi are offset a little bit by demoting .pps from int to char (save 3 bytes). And trim max act exts to 2, since we only provide one, saving 8 /16 bytes by itself if exts enabled. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-19 04:32:14 +08:00
wsi->timeout_list_prev = NULL;
wsi->timeout_list = NULL;
multithread stability Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-26 20:56:56 +08:00
lws_pt_unlock(pt);
timeout move to doubly linked list In the case we have a lot of connections, checking them all for timeout state once a second becomes burdensome. At the moment if you have 100K connections, once a second they all get checked for timeout in a loop. This patch adds a doubly-linked list based in the context to each wsi, and only wsi with pending timeouts appear on it. At checking time, we traverse the list, which costs nothing if empty because nobody has a pending timeout. Similarly adding and removing from the list costs almost nothing since no iteration is required no matter how big the list. The extra 8 or 16 bytes in the wsi are offset a little bit by demoting .pps from int to char (save 3 bytes). And trim max act exts to 2, since we only provide one, saving 8 /16 bytes by itself if exts enabled. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-19 04:32:14 +08:00
}
multithread stability Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-26 20:56:56 +08:00
LWS_VISIBLE void
lws_set_timeout(struct lws *wsi, enum pending_timeout reason, int secs)
{
struct lws_context_per_thread *pt = &wsi->context->pt[(int)wsi->tsi];
time_t now;
lws-meta
2017-07-19 04:39:14 +08:00
if (secs == LWS_TO_KILL_SYNC) {
lws_remove_from_timeout_list(wsi);
lwsl_debug("synchronously killing %p\n", wsi);
lws_close_free_wsi(wsi, LWS_CLOSE_STATUS_NOSTATUS);
return;
}
multithread stability Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-26 20:56:56 +08:00
lws_pt_lock(pt);
time(&now);
improve timeout and ah list comments Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-25 21:39:01 +08:00
if (reason && !wsi->timeout_list_prev) {
/* our next guy is current first guy */
multithread stability Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-26 20:56:56 +08:00
wsi->timeout_list = pt->timeout_list;
improve timeout and ah list comments Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-25 21:39:01 +08:00
/* if there is a next guy, set his prev ptr to our next ptr */
multithread stability Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-26 20:56:56 +08:00
if (wsi->timeout_list)
wsi->timeout_list->timeout_list_prev = &wsi->timeout_list;
improve timeout and ah list comments Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-25 21:39:01 +08:00
/* our prev ptr is first ptr */
multithread stability Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-26 20:56:56 +08:00
wsi->timeout_list_prev = &pt->timeout_list;
improve timeout and ah list comments Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-25 21:39:01 +08:00
/* set the first guy to be us */
multithread stability Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-26 20:56:56 +08:00
*wsi->timeout_list_prev = wsi;
}
cgi Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-21 21:25:48 +08:00
lwsl_debug("%s: %p: %d secs\n", __func__, wsi, secs);
multithread stability Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-26 20:56:56 +08:00
wsi->pending_timeout_limit = now + secs;
wsi->pending_timeout = reason;
lws_pt_unlock(pt);
if (!reason)
lws_remove_from_timeout_list(wsi);
}
timeout move to doubly linked list In the case we have a lot of connections, checking them all for timeout state once a second becomes burdensome. At the moment if you have 100K connections, once a second they all get checked for timeout in a loop. This patch adds a doubly-linked list based in the context to each wsi, and only wsi with pending timeouts appear on it. At checking time, we traverse the list, which costs nothing if empty because nobody has a pending timeout. Similarly adding and removing from the list costs almost nothing since no iteration is required no matter how big the list. The extra 8 or 16 bytes in the wsi are offset a little bit by demoting .pps from int to char (save 3 bytes). And trim max act exts to 2, since we only provide one, saving 8 /16 bytes by itself if exts enabled. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-19 04:32:14 +08:00
adopt: allow binding to parent at same time - if protocol set, allocate own user_space If the child wsi wants the parent wsi user_space, it can use lws_wsi_user(lws_get_parent(child_wsi)) - raw file close processing handles parent-child relationship
2017-03-03 07:36:08 +08:00
static void
lws_remove_child_from_any_parent(struct lws *wsi)
{
struct lws **pwsi;
lws_remove_child_from_any_parent: clear parent pointer and fix failure message
2017-03-30 08:31:36 +08:00
int seen = 0;
adopt: allow binding to parent at same time - if protocol set, allocate own user_space If the child wsi wants the parent wsi user_space, it can use lws_wsi_user(lws_get_parent(child_wsi)) - raw file close processing handles parent-child relationship
2017-03-03 07:36:08 +08:00
lws_remove_child_from_any_parent: clear parent pointer and fix failure message
2017-03-30 08:31:36 +08:00
if (!wsi->parent)
return;
/* detach ourselves from parent's child list */
pwsi = &wsi->parent->child_list;
while (*pwsi) {
if (*pwsi == wsi) {
clean
2017-09-23 12:55:21 +08:00
lwsl_info("%s: detach %p from parent %p\n", __func__,
wsi, wsi->parent);
lws-meta
2017-07-19 04:39:14 +08:00
if (wsi->parent->protocol)
wsi->parent->protocol->callback(wsi,
LWS_CALLBACK_CHILD_CLOSING,
wsi->parent->user_space, wsi, 0);
lws_remove_child_from_any_parent: clear parent pointer and fix failure message
2017-03-30 08:31:36 +08:00
*pwsi = wsi->sibling_list;
seen = 1;
break;
adopt: allow binding to parent at same time - if protocol set, allocate own user_space If the child wsi wants the parent wsi user_space, it can use lws_wsi_user(lws_get_parent(child_wsi)) - raw file close processing handles parent-child relationship
2017-03-03 07:36:08 +08:00
}
lws_remove_child_from_any_parent: clear parent pointer and fix failure message
2017-03-30 08:31:36 +08:00
pwsi = &(*pwsi)->sibling_list;
adopt: allow binding to parent at same time - if protocol set, allocate own user_space If the child wsi wants the parent wsi user_space, it can use lws_wsi_user(lws_get_parent(child_wsi)) - raw file close processing handles parent-child relationship
2017-03-03 07:36:08 +08:00
}
lws_remove_child_from_any_parent: clear parent pointer and fix failure message
2017-03-30 08:31:36 +08:00
if (!seen)
lwsl_err("%s: failed to detach from parent\n", __func__);
wsi->parent = NULL;
adopt: allow binding to parent at same time - if protocol set, allocate own user_space If the child wsi wants the parent wsi user_space, it can use lws_wsi_user(lws_get_parent(child_wsi)) - raw file close processing handles parent-child relationship
2017-03-03 07:36:08 +08:00
}
raw: enable server and client raw sockets
2017-03-07 16:06:05 +08:00
int
lws_bind_protocol(struct lws *wsi, const struct lws_protocols *p)
{
// if (wsi->protocol == p)
// return 0;
raw: take care about same vh protocol linked list
2017-04-06 13:49:17 +08:00
const struct lws_protocols *vp = wsi->vhost->protocols, *vpo;
raw: enable server and client raw sockets
2017-03-07 16:06:05 +08:00
if (wsi->protocol)
wsi->protocol->callback(wsi, LWS_CALLBACK_HTTP_DROP_PROTOCOL,
wsi->user_space, NULL, 0);
if (!wsi->user_space_externally_allocated)
lws_free_set_NULL(wsi->user_space);
raw: take care about same vh protocol linked list
2017-04-06 13:49:17 +08:00
lws_same_vh_protocol_remove(wsi);
raw: enable server and client raw sockets
2017-03-07 16:06:05 +08:00
wsi->protocol = p;
if (!p)
return 0;
if (lws_ensure_user_space(wsi))
return 1;
raw: take care about same vh protocol linked list
2017-04-06 13:49:17 +08:00
if (p > vp && p < &vp[wsi->vhost->count_protocols])
lws_same_vh_protocol_insert(wsi, p - vp);
else {
int n = wsi->vhost->count_protocols;
int hit = 0;
vpo = vp;
while (n--) {
bind protocol: fix for NULL names
2017-04-07 18:22:54 +08:00
if (p->name && vp->name && !strcmp(p->name, vp->name)) {
raw: take care about same vh protocol linked list
2017-04-06 13:49:17 +08:00
hit = 1;
lws_same_vh_protocol_insert(wsi, vp - vpo);
break;
}
vp++;
}
if (!hit)
clean
2017-09-23 12:55:21 +08:00
lwsl_err("%s: %p is not in vhost '%s' protocols list\n",
__func__, p, wsi->vhost->name);
raw: take care about same vh protocol linked list
2017-04-06 13:49:17 +08:00
}
raw: enable server and client raw sockets
2017-03-07 16:06:05 +08:00
if (wsi->protocol->callback(wsi, LWS_CALLBACK_HTTP_BIND_PROTOCOL,
wsi->user_space, NULL, 0))
return 1;
return 0;
}
clean-whitespace.patch Signed-off-by: Andy Green <andy@warmcat.com>
2010-12-19 22:13:26 +00:00
void
refactor needless context with wsi paramater passing Now we bit the bullet and gave each wsi an lws_context *, many internal apis that take both a context and wsi parameter only need the wsi. Also simplify parser code by making a temp var for allocated_headers * instead of the longwinded dereference chain everywhere. Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-15 21:15:58 +08:00
lws_close_free_wsi(struct lws *wsi, enum lws_close_status reason)
add-per-session-user-data.patch Signed-off-by: Andy Green <andy@warmcat.com>
2010-11-03 11:13:06 +00:00
{
coverity 156860 wsi dereference before NULL check ... bit in reality, wsi can never be NULL Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-19 21:32:08 +08:00
struct lws_context_per_thread *pt;
adopt: allow binding to parent at same time - if protocol set, allocate own user_space If the child wsi wants the parent wsi user_space, it can use lws_wsi_user(lws_get_parent(child_wsi)) - raw file close processing handles parent-child relationship
2017-03-03 07:36:08 +08:00
struct lws *wsi1, *wsi2;
add explicit parent child wsi relationships wsi can have a full tree relationship with each other using linked lists. closing the parent ensures the children are closed first. Convert cgi to use this instead of his cgi-specific sub-wsi management. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-03-02 09:17:22 +08:00
struct lws_context *context;
flush extensions when closing Extensions might be caching stuff that we should spill before a controlled close. It's not allowed to send anything on the wire after the close request, so we need to make the extensions spill just before. Signed-off-by: Andy Green <andy@warmcat.com>
2011-03-07 07:08:18 +00:00
struct lws_tokens eff_buf;
add explicit parent child wsi relationships wsi can have a full tree relationship with each other using linked lists. closing the parent ensures the children are closed first. Convert cgi to use this instead of his cgi-specific sub-wsi management. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-03-02 09:17:22 +08:00
int n, m, ret;
fork-sever-process-and-introduce-broadcast-api.patch Signed-off-by: Andy Green <andy@warmcat.com>
2010-12-18 15:13:50 +00:00
lws-meta
2017-07-19 04:39:14 +08:00
lwsl_debug("%s: %p\n", __func__, wsi);
unify server and client close Signed-off-by: Andy Green <andy@warmcat.com>
2011-02-14 08:03:48 +00:00
if (!wsi)
fork-sever-process-and-introduce-broadcast-api.patch Signed-off-by: Andy Green <andy@warmcat.com>
2010-12-18 15:13:50 +00:00
return;
lws access log option and lwsws conf This adds the ability to store apache-compatible logs to a file given at vhost-creation time. lwsws conf can set it per-vhost using "access-log": "<filepath>" The feature defaults to disabled at cmake, it can be set independently but LWS_WITH_LWSWS set it on. Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-15 12:00:23 +08:00
lws_access_log(wsi);
esp8266 initial support
2016-07-23 14:18:25 +08:00
#if defined(LWS_WITH_ESP8266)
if (wsi->premature_rx)
lws_free(wsi->premature_rx);
if (wsi->pending_send_completion && !wsi->close_is_pending_send_completion) {
lwsl_notice("delaying close\n");
wsi->close_is_pending_send_completion = 1;
return;
}
#endif
lws access log option and lwsws conf This adds the ability to store apache-compatible logs to a file given at vhost-creation time. lwsws conf can set it per-vhost using "access-log": "<filepath>" The feature defaults to disabled at cmake, it can be set independently but LWS_WITH_LWSWS set it on. Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-15 12:00:23 +08:00
ah: require parsing complete before detach Introduce helpers to force to detachable state and to test the ah is in a detachable state. Require not only the ah rx buffer is all used, but that the wsi has completed a full set of headers.
2017-06-28 12:13:13 +08:00
/* we're closing, losing some rx is OK */
lws_header_table_force_to_detachable_state(wsi);
closing drops any pending ah rx immediately
2016-10-04 18:05:10 +08:00
coverity 156860 wsi dereference before NULL check ... bit in reality, wsi can never be NULL Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-19 21:32:08 +08:00
context = wsi->context;
pt = &context->pt[(int)wsi->tsi];
LWS_WITH_STATS
2017-05-07 10:02:03 +08:00
lws_stats_atomic_bump(wsi->context, pt, LWSSTATS_C_API_CLOSE, 1);
add-per-session-user-data.patch Signed-off-by: Andy Green <andy@warmcat.com>
2010-11-03 11:13:06 +00:00
add explicit parent child wsi relationships wsi can have a full tree relationship with each other using linked lists. closing the parent ensures the children are closed first. Convert cgi to use this instead of his cgi-specific sub-wsi management. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-03-02 09:17:22 +08:00
/* if we have children, close them first */
if (wsi->child_list) {
wsi2 = wsi->child_list;
while (wsi2) {
wsi1 = wsi2->sibling_list;
proxy rewrite If you enable -DLWS_WITH_HTTP_PROXY=1 at cmake, the test server has a new URI path http://localhost:7681/proxytest If you visit here, a client connection to http://example.com:80 is spawned, and the results piped on to your original connection. Also with LWS_WITH_HTTP_PROXY enabled at cmake, lws wants to link to an additional library, "libhubbub". This allows lws to do html rewriting on the fly, adjusting proxied urls in a lightweight and fast way.
2016-03-20 11:59:53 +08:00
wsi2->parent = NULL;
/* stop it doing shutdown processing */
wsi2->socket_is_permanently_unusable = 1;
add explicit parent child wsi relationships wsi can have a full tree relationship with each other using linked lists. closing the parent ensures the children are closed first. Convert cgi to use this instead of his cgi-specific sub-wsi management. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-03-02 09:17:22 +08:00
lws_close_free_wsi(wsi2, reason);
wsi2 = wsi1;
}
proxy rewrite If you enable -DLWS_WITH_HTTP_PROXY=1 at cmake, the test server has a new URI path http://localhost:7681/proxytest If you visit here, a client connection to http://example.com:80 is spawned, and the results piped on to your original connection. Also with LWS_WITH_HTTP_PROXY enabled at cmake, lws wants to link to an additional library, "libhubbub". This allows lws to do html rewriting on the fly, adjusting proxied urls in a lightweight and fast way.
2016-03-20 11:59:53 +08:00
wsi->child_list = NULL;
add explicit parent child wsi relationships wsi can have a full tree relationship with each other using linked lists. closing the parent ensures the children are closed first. Convert cgi to use this instead of his cgi-specific sub-wsi management. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-03-02 09:17:22 +08:00
}
http2: make usable HTTP/2 support is now able to serve the test server, complete with websockets, from a single vhost. - This works the same with both OpenSSL and mbedTLS. - POST is now wired up and works (also for file upload). - CGI is wired up and works. - Redirect is adapted and works - lwsws works. - URI urldecode, sanitation and argument parsing wired up for :path valgrind clean (aside from openssl-style false uninit data usage in mbedtls send occasionally) h2spec reports: $ h2spec -h 127.0.0.1 -p 7681 -t -k -o 1 ... 145 tests, 145 passed, 0 skipped, 0 failed" Incorporates: - "https://github.com/warmcat/libwebsockets/pull/1039 Fixes issue with -Werror=unused-variable flag - https://github.com/sartura/libwebsockets/commit/2c843a1395bdca5c23b40d989f508fc96aa5ba3b ssl: fix infinite loop on client cert verification failure Signed-off-by: Petar Paradzik <petar.paradzik@sartura.hr>" Caused and fixes Coverity 184887 - 184892
2017-10-13 10:33:02 +08:00
#if defined(LWS_WITH_HTTP2)
if (wsi->u.h2.parent_wsi) {
lwsl_info(" wsi: %p, his parent %p: siblings:\n", wsi, wsi->u.h2.parent_wsi);
lws_start_foreach_llp(struct lws **, w, wsi->u.h2.parent_wsi->u.h2.child_list) {
lwsl_info(" \\---- child %p\n", *w);
} lws_end_foreach_llp(w, u.h2.sibling_list);
}
if (wsi->upgraded_to_http2 || wsi->http2_substream) {
lwsl_info("closing %p: parent %p\n", wsi, wsi->u.h2.parent_wsi);
if (wsi->u.h2.child_list) {
lwsl_info(" parent %p: closing children: list:\n", wsi);
lws_start_foreach_llp(struct lws **, w, wsi->u.h2.child_list) {
lwsl_info(" \\---- child %p\n", *w);
} lws_end_foreach_llp(w, u.h2.sibling_list);
/* trigger closing of all of our http2 children first */
lws_start_foreach_llp(struct lws **, w, wsi->u.h2.child_list) {
lwsl_info(" closing child %p\n", *w);
/* disconnect from siblings */
wsi2 = (*w)->u.h2.sibling_list;
(*w)->u.h2.sibling_list = NULL;
(*w)->socket_is_permanently_unusable = 1;
lws_close_free_wsi(*w, reason);
*w = wsi2;
continue;
} lws_end_foreach_llp(w, u.h2.sibling_list);
}
}
if (wsi->upgraded_to_http2) {
/* remove pps */
struct lws_h2_protocol_send *w = wsi->u.h2.h2n->pps, *w1;
while (w) {
w1 = wsi->u.h2.h2n->pps->next;
free(w);
w = w1;
}
wsi->u.h2.h2n->pps = NULL;
}
if (wsi->http2_substream && wsi->u.h2.parent_wsi) {
lwsl_info(" %p: disentangling from siblings\n", wsi);
lws_start_foreach_llp(struct lws **, w,
wsi->u.h2.parent_wsi->u.h2.child_list) {
/* disconnect from siblings */
if (*w == wsi) {
wsi2 = (*w)->u.h2.sibling_list;
(*w)->u.h2.sibling_list = NULL;
*w = wsi2;
lwsl_info(" %p disentangled from sibling %p\n", wsi, wsi2);
break;
}
} lws_end_foreach_llp(w, u.h2.sibling_list);
wsi->u.h2.parent_wsi->u.h2.child_count--;
wsi->u.h2.parent_wsi = NULL;
if (wsi->u.h2.pending_status_body)
lws_free_set_NULL(wsi->u.h2.pending_status_body);
}
if (wsi->upgraded_to_http2 && wsi->u.h2.h2n &&
wsi->u.h2.h2n->rx_scratch)
lws_free_set_NULL(wsi->u.h2.h2n->rx_scratch);
#endif
adopt: allow binding to parent at same time - if protocol set, allocate own user_space If the child wsi wants the parent wsi user_space, it can use lws_wsi_user(lws_get_parent(child_wsi)) - raw file close processing handles parent-child relationship
2017-03-03 07:36:08 +08:00
if (wsi->mode == LWSCM_RAW_FILEDESC) {
http2: make usable HTTP/2 support is now able to serve the test server, complete with websockets, from a single vhost. - This works the same with both OpenSSL and mbedTLS. - POST is now wired up and works (also for file upload). - CGI is wired up and works. - Redirect is adapted and works - lwsws works. - URI urldecode, sanitation and argument parsing wired up for :path valgrind clean (aside from openssl-style false uninit data usage in mbedtls send occasionally) h2spec reports: $ h2spec -h 127.0.0.1 -p 7681 -t -k -o 1 ... 145 tests, 145 passed, 0 skipped, 0 failed" Incorporates: - "https://github.com/warmcat/libwebsockets/pull/1039 Fixes issue with -Werror=unused-variable flag - https://github.com/sartura/libwebsockets/commit/2c843a1395bdca5c23b40d989f508fc96aa5ba3b ssl: fix infinite loop on client cert verification failure Signed-off-by: Petar Paradzik <petar.paradzik@sartura.hr>" Caused and fixes Coverity 184887 - 184892
2017-10-13 10:33:02 +08:00
lws_remove_child_from_any_parent(wsi);
remove_wsi_socket_from_fds(wsi);
wsi->protocol->callback(wsi,
LWS_CALLBACK_RAW_CLOSE_FILE,
wsi->user_space, NULL, 0);
goto async_close;
adopt: allow binding to parent at same time - if protocol set, allocate own user_space If the child wsi wants the parent wsi user_space, it can use lws_wsi_user(lws_get_parent(child_wsi)) - raw file close processing handles parent-child relationship
2017-03-03 07:36:08 +08:00
}
cgi Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-21 21:25:48 +08:00
#ifdef LWS_WITH_CGI
if (wsi->mode == LWSCM_CGI) {
/* we are not a network connection, but a handler for CGI io */
cgi: remove and kill CGI after closing its STDOUT handler Signed-off-by: Petar Paradzik <petar.paradzik@sartura.hr>
2017-08-26 12:00:16 +08:00
if (wsi->parent && wsi->parent->cgi) {
if (wsi->cgi_channel == LWS_STDOUT)
lws_cgi_remove_and_kill(wsi->parent);
lwsws cgi integration Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-13 11:49:07 +08:00
/* end the binding between us and master */
wsi->parent->cgi->stdwsi[(int)wsi->cgi_channel] = NULL;
cgi: remove and kill CGI after closing its STDOUT handler Signed-off-by: Petar Paradzik <petar.paradzik@sartura.hr>
2017-08-26 12:00:16 +08:00
}
cgi Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-21 21:25:48 +08:00
wsi->socket_is_permanently_unusable = 1;
goto just_kill_connection;
}
cgi: remove and kill CGI after closing its STDOUT handler Signed-off-by: Petar Paradzik <petar.paradzik@sartura.hr>
2017-08-26 12:00:16 +08:00
if (wsi->cgi)
lws_cgi_remove_and_kill(wsi);
cgi Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-21 21:25:48 +08:00
#endif
client-fix-header-stash-leak-on-close-before-success
2017-07-21 21:49:24 +08:00
#if !defined(LWS_NO_CLIENT)
if (wsi->mode == LWSCM_HTTP_CLIENT ||
wsi->mode == LWSCM_WSCL_WAITING_CONNECT ||
wsi->mode == LWSCM_WSCL_WAITING_PROXY_REPLY ||
wsi->mode == LWSCM_WSCL_ISSUE_HANDSHAKE ||
wsi->mode == LWSCM_WSCL_ISSUE_HANDSHAKE2 ||
wsi->mode == LWSCM_WSCL_WAITING_SSL ||
wsi->mode == LWSCM_WSCL_WAITING_SERVER_REPLY ||
wsi->mode == LWSCM_WSCL_WAITING_EXTENSION_CONNECT ||
wsi->mode == LWSCM_WSCL_WAITING_SOCKS_GREETING_REPLY ||
wsi->mode == LWSCM_WSCL_WAITING_SOCKS_CONNECT_REPLY ||
wsi->mode == LWSCM_WSCL_WAITING_SOCKS_AUTH_REPLY)
if (wsi->u.hdr.stash)
lws_free_set_NULL(wsi->u.hdr.stash);
#endif
raw: adoption and processing
2017-02-12 20:32:49 +08:00
if (wsi->mode == LWSCM_RAW) {
adoption: make union for socket and file fds This lets lws support adopting raw file FDs and raw socket fds. A test plugin creates a FIFO and prints data sent on it, using the lws event loop.
2017-02-27 12:55:56 +08:00
wsi->protocol->callback(wsi,
raw: adoption and processing
2017-02-12 20:32:49 +08:00
LWS_CALLBACK_RAW_CLOSE, wsi->user_space, NULL, 0);
wsi->socket_is_permanently_unusable = 1;
goto just_kill_connection;
}
http2: make usable HTTP/2 support is now able to serve the test server, complete with websockets, from a single vhost. - This works the same with both OpenSSL and mbedTLS. - POST is now wired up and works (also for file upload). - CGI is wired up and works. - Redirect is adapted and works - lwsws works. - URI urldecode, sanitation and argument parsing wired up for :path valgrind clean (aside from openssl-style false uninit data usage in mbedtls send occasionally) h2spec reports: $ h2spec -h 127.0.0.1 -p 7681 -t -k -o 1 ... 145 tests, 145 passed, 0 skipped, 0 failed" Incorporates: - "https://github.com/warmcat/libwebsockets/pull/1039 Fixes issue with -Werror=unused-variable flag - https://github.com/sartura/libwebsockets/commit/2c843a1395bdca5c23b40d989f508fc96aa5ba3b ssl: fix infinite loop on client cert verification failure Signed-off-by: Petar Paradzik <petar.paradzik@sartura.hr>" Caused and fixes Coverity 184887 - 184892
2017-10-13 10:33:02 +08:00
if ((wsi->mode == LWSCM_HTTP_SERVING_ACCEPTED ||
wsi->mode == LWSCM_HTTP2_SERVING) &&
fops: refactor around lws_fops_fd_t
2017-02-25 12:42:45 +08:00
wsi->u.http.fop_fd != NULL) {
fops-zip
2017-03-03 12:38:10 +08:00
lws_vfs_file_close(&wsi->u.http.fop_fd);
introduce urlencode decode and sql escape public apis This adds - simple lws_urlencode() - simple lws_urldecode() - simple lws_sql_purify Those expect the data to all be there and process it up until the first '\0'. There is also a larger opaque apis for handling POST_BODY urldecode. To enable these, you need to give cmake -DLWS_WITH_STATEFUL_URLDECODE=1 (or arrange any larger feature that relies on it sets that in CMakeLists.txt) - stateful urldecode with parameter array These have create / process / destroy semantics on a struct that maintains decode state. Stateful urldecode is capable of dealing with large POST data in multiple POST_BODY callbacks cleanly, eg, file transfer by POST. Stateful urldecode with parameter array wraps the above with a canned callback that stores the urldecoded data and indexes them in a pointer array matching an array of parameter names. You may also pass it an optional callback when creating it, that will recieve uploaded file content. The test html is updated to support both urlencoded and multipart forms, with some javascript to do clientside validation of an arbitrary 100KB file size limit (there is no file size limit in the apis). Signed-off-by: Andy Green <andy@warmcat.com>
2016-06-08 10:07:02 +08:00
wsi->vhost->protocols->callback(wsi,
introduce vhosts This patch splits out some lws_context members into a new lws_vhost struct. - ssl state and options per vhost - SSL_CTX for serving and client per vhost - protocols[] per vhost - extensions[] per vhost lws_context maintains a linked list of lws_vhosts. The same lws_context_creation_info struct is used to regulate both the context creation and to create vhosts: for backward compatibility if you didn't provide the new LWS_SERVER_OPTION_EXPLICIT_VHOSTS option, then a default vhost is created at context creation time using the same info data as the context itself. If you will have multiple vhosts though, you should give the LWS_SERVER_OPTION_EXPLICIT_VHOSTS option at context creation time, create the context first and then the vhosts afterwards using lws_create_vhost(contest, &info); Although there is a lot of housekeeping to implement this change, there is almost no additional overhead if you don't use multiple vhosts and very little api impact (no changes to test apps). Signed-off-by: Andy Green <andy@warmcat.com>
2016-03-28 10:10:43 +08:00
LWS_CALLBACK_CLOSED_HTTP, wsi->user_space, NULL, 0);
client http: extra read notification after close https://github.com/warmcat/libwebsockets/issues/620
2016-09-06 15:36:51 +08:00
wsi->told_user_closed = 1;
cleanups after api changes and mbed update Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-04 16:54:12 +08:00
}
timeout counts as pernmanently unusable close Signed-off-by: Andy Green <andy.green@linaro.org>
2015-04-12 08:17:26 +08:00
if (wsi->socket_is_permanently_unusable ||
multithread stability Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-26 20:56:56 +08:00
reason == LWS_CLOSE_STATUS_NOSTATUS_CONTEXT_DESTROY ||
wsi->state == LWSS_SHUTDOWN)
trac 89 add flag on fatal send fail to use simple close Signed-off-by: Andy Green <andy.green@linaro.org>
2014-10-16 08:23:46 +08:00
goto just_kill_connection;
multithread stability Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-26 20:56:56 +08:00
wsi->state_pre_close = wsi->state;
switch (wsi->state_pre_close) {
clean internal refactor - Mainly symbol length reduction - Whitespace clean - Code refactor for linear flow - Audit @Context for API docs vs changes Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-17 17:03:59 +08:00
case LWSS_DEAD_SOCKET:
introduce-new-05-socket-closure-rules.patch Signed-off-by: Andy Green <andy.green@linaro.org>
2011-02-10 09:07:05 +00:00
return;
close if we tried to close politely just close next time Signed-off-by: Andy Green <andy.green@linaro.org>
2013-02-15 22:48:58 +08:00
/* we tried the polite way... */
close: make close notification go through writable Until now we took the approach if just writing the close notification broke something, we didn't care because we were closing the connection anyway. But with lws_meta, breaking stuff in the parent connection would be a sticky problem outliving the closing child connection. So this adds a new wsi state LWSS_WAITING_TO_SEND_CLOSE_NOTIFICATION and makes the send go via the writable callback mechanism.
2017-07-17 10:11:17 +08:00
case LWSS_WAITING_TO_SEND_CLOSE_NOTIFICATION:
clean internal refactor - Mainly symbol length reduction - Whitespace clean - Code refactor for linear flow - Audit @Context for API docs vs changes Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-17 17:03:59 +08:00
case LWSS_AWAITING_CLOSE_ACK:
close if we tried to close politely just close next time Signed-off-by: Andy Green <andy.green@linaro.org>
2013-02-15 22:48:58 +08:00
goto just_kill_connection;
clean internal refactor - Mainly symbol length reduction - Whitespace clean - Code refactor for linear flow - Audit @Context for API docs vs changes Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-17 17:03:59 +08:00
case LWSS_FLUSHING_STORED_SEND_BEFORE_CLOSE:
if (wsi->trunc_len) {
public api remove superfluous context params API BREAK Extend the cleanout caused by wsi having a context pointer into the public api. There's no point keeping the 1.5 compatibility work, we have changed the api in several places and rebuilt wasn't going to be enough a while ago. Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-16 18:19:08 +08:00
lws_callback_on_writable(wsi);
close should flush any truncated send before really closing Signed-off-by: Andy Green <andy.green@linaro.org>
2014-04-10 14:25:24 +08:00
return;
}
clean
2017-09-23 12:55:21 +08:00
lwsl_info("%p: end FLUSHING_STORED_SEND_BEFORE_CLOSE\n", wsi);
close should flush any truncated send before really closing Signed-off-by: Andy Green <andy.green@linaro.org>
2014-04-10 14:25:24 +08:00
goto just_kill_connection;
default:
clean internal refactor - Mainly symbol length reduction - Whitespace clean - Code refactor for linear flow - Audit @Context for API docs vs changes Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-17 17:03:59 +08:00
if (wsi->trunc_len) {
clean
2017-09-23 12:55:21 +08:00
lwsl_info("%p: start FLUSHING_STORED_SEND_BEFORE_CLOSE\n", wsi);
clean internal refactor - Mainly symbol length reduction - Whitespace clean - Code refactor for linear flow - Audit @Context for API docs vs changes Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-17 17:03:59 +08:00
wsi->state = LWSS_FLUSHING_STORED_SEND_BEFORE_CLOSE;
api rationalization use new names internally Change all internal uses of rationalized public apis to reflect the new names. Theer are a few things that got changed as side effect of search/replace matches, but these are almost all internal. I added a compatibility define for the public enum that got renamed. Theoretically existing code should not notice the difference from these two patches. And new code will find the new names. https://github.com/warmcat/libwebsockets/issues/357 Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-04 08:43:54 +08:00
lws_set_timeout(wsi, PENDING_FLUSH_STORED_SEND_BEFORE_CLOSE, 5);
close should flush any truncated send before really closing Signed-off-by: Andy Green <andy.green@linaro.org>
2014-04-10 14:25:24 +08:00
return;
}
break;
}
clean internal refactor - Mainly symbol length reduction - Whitespace clean - Code refactor for linear flow - Audit @Context for API docs vs changes Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-17 17:03:59 +08:00
if (wsi->mode == LWSCM_WSCL_WAITING_CONNECT ||
wsi->mode == LWSCM_WSCL_ISSUE_HANDSHAKE)
trac 40 client connection properly nonblocking As spotted by JM on Trac#40 http://libwebsockets.org/trac/libwebsockets/ticket/40 client connect didn't do anything about being truly nonblocking. This patch should hopefully solve that. Signed-off-by: Andy Green <andy.green@linaro.org>
2013-09-20 20:26:12 +08:00
goto just_kill_connection;
http2: make usable HTTP/2 support is now able to serve the test server, complete with websockets, from a single vhost. - This works the same with both OpenSSL and mbedTLS. - POST is now wired up and works (also for file upload). - CGI is wired up and works. - Redirect is adapted and works - lwsws works. - URI urldecode, sanitation and argument parsing wired up for :path valgrind clean (aside from openssl-style false uninit data usage in mbedtls send occasionally) h2spec reports: $ h2spec -h 127.0.0.1 -p 7681 -t -k -o 1 ... 145 tests, 145 passed, 0 skipped, 0 failed" Incorporates: - "https://github.com/warmcat/libwebsockets/pull/1039 Fixes issue with -Werror=unused-variable flag - https://github.com/sartura/libwebsockets/commit/2c843a1395bdca5c23b40d989f508fc96aa5ba3b ssl: fix infinite loop on client cert verification failure Signed-off-by: Petar Paradzik <petar.paradzik@sartura.hr>" Caused and fixes Coverity 184887 - 184892
2017-10-13 10:33:02 +08:00
if (wsi->mode == LWSCM_HTTP_SERVING ||
wsi->mode == LWSCM_HTTP2_SERVING) {
introduce urlencode decode and sql escape public apis This adds - simple lws_urlencode() - simple lws_urldecode() - simple lws_sql_purify Those expect the data to all be there and process it up until the first '\0'. There is also a larger opaque apis for handling POST_BODY urldecode. To enable these, you need to give cmake -DLWS_WITH_STATEFUL_URLDECODE=1 (or arrange any larger feature that relies on it sets that in CMakeLists.txt) - stateful urldecode with parameter array These have create / process / destroy semantics on a struct that maintains decode state. Stateful urldecode is capable of dealing with large POST data in multiple POST_BODY callbacks cleanly, eg, file transfer by POST. Stateful urldecode with parameter array wraps the above with a canned callback that stores the urldecoded data and indexes them in a pointer array matching an array of parameter names. You may also pass it an optional callback when creating it, that will recieve uploaded file content. The test html is updated to support both urlencoded and multipart forms, with some javascript to do clientside validation of an arbitrary 100KB file size limit (there is no file size limit in the apis). Signed-off-by: Andy Green <andy@warmcat.com>
2016-06-08 10:07:02 +08:00
if (wsi->user_space)
wsi->vhost->protocols->callback(wsi,
LWS_CALLBACK_HTTP_DROP_PROTOCOL,
public api remove context from user callback API BREAK Since struct lws (wsi) now has his own context pointer, we were able to remove the need for passing context almost everywhere in the apis. In turn, that means there's no real use for context being passed to every callback; in the rare cases context is needed user code can get it with lws_get_ctx(wsi) Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-17 07:54:44 +08:00
wsi->user_space, NULL, 0);
introduce urlencode decode and sql escape public apis This adds - simple lws_urlencode() - simple lws_urldecode() - simple lws_sql_purify Those expect the data to all be there and process it up until the first '\0'. There is also a larger opaque apis for handling POST_BODY urldecode. To enable these, you need to give cmake -DLWS_WITH_STATEFUL_URLDECODE=1 (or arrange any larger feature that relies on it sets that in CMakeLists.txt) - stateful urldecode with parameter array These have create / process / destroy semantics on a struct that maintains decode state. Stateful urldecode is capable of dealing with large POST data in multiple POST_BODY callbacks cleanly, eg, file transfer by POST. Stateful urldecode with parameter array wraps the above with a canned callback that stores the urldecoded data and indexes them in a pointer array matching an array of parameter names. You may also pass it an optional callback when creating it, that will recieve uploaded file content. The test html is updated to support both urlencoded and multipart forms, with some javascript to do clientside validation of an arbitrary 100KB file size limit (there is no file size limit in the apis). Signed-off-by: Andy Green <andy@warmcat.com>
2016-06-08 10:07:02 +08:00
wsi->vhost->protocols->callback(wsi, LWS_CALLBACK_CLOSED_HTTP,
wsi->user_space, NULL, 0);
client http: extra read notification after close https://github.com/warmcat/libwebsockets/issues/620
2016-09-06 15:36:51 +08:00
wsi->told_user_closed = 1;
introduce urlencode decode and sql escape public apis This adds - simple lws_urlencode() - simple lws_urldecode() - simple lws_sql_purify Those expect the data to all be there and process it up until the first '\0'. There is also a larger opaque apis for handling POST_BODY urldecode. To enable these, you need to give cmake -DLWS_WITH_STATEFUL_URLDECODE=1 (or arrange any larger feature that relies on it sets that in CMakeLists.txt) - stateful urldecode with parameter array These have create / process / destroy semantics on a struct that maintains decode state. Stateful urldecode is capable of dealing with large POST data in multiple POST_BODY callbacks cleanly, eg, file transfer by POST. Stateful urldecode with parameter array wraps the above with a canned callback that stores the urldecoded data and indexes them in a pointer array matching an array of parameter names. You may also pass it an optional callback when creating it, that will recieve uploaded file content. The test html is updated to support both urlencoded and multipart forms, with some javascript to do clientside validation of an arbitrary 100KB file size limit (there is no file size limit in the apis). Signed-off-by: Andy Green <andy@warmcat.com>
2016-06-08 10:07:02 +08:00
}
Callback with reason LWS_CALLBACK_CLOSED_HTTP when HTTP closed before serve. If the remote HTTP client closes the connection before serving commences, the only notification client code receives is LWS_CALLBACK_WSI_DESTROY. This commit gives LWS_CALLBACK_CLOSED_HTTP as well, as would happen if HTTP serving had already commenced.
2015-01-28 04:15:13 +08:00
extensions introduce pre close veto This establishes a pre-close extension check to allow an extension to veto a close. x-google-mux then uses this to stop ch1 going down (subchannel 1 is the original socket connection that turns into a mux parent) if it has active mux children; it just marks ch1 as closed in its conn struct in that case and returns 1 from the callback to veto. Code is also added to take care of the case ch1 is 'closed', and the last child is subsequently closed, it actively calls close on the mux parent then. Signed-off-by: Andy Green <andy@warmcat.com>
2011-05-25 21:41:57 +01:00
/*
* are his extensions okay with him closing? Eg he might be a mux
* parent and just his ch1 aspect is closing?
*/
whitespace trailing mass cleanout Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-14 08:52:03 +08:00
clean
2017-09-23 12:55:21 +08:00
if (lws_ext_cb_active(wsi, LWS_EXT_CB_CHECK_OK_TO_REALLY_CLOSE, NULL, 0) > 0) {
plat change to build selection by cmake Signed-off-by: Andy Green <andy.green@linaro.org>
2014-04-02 21:02:54 +08:00
lwsl_ext("extension vetoed close\n");
return;
extensions introduce pre close veto This establishes a pre-close extension check to allow an extension to veto a close. x-google-mux then uses this to stop ch1 going down (subchannel 1 is the original socket connection that turns into a mux parent) if it has active mux children; it just marks ch1 as closed in its conn struct in that case and returns 1 from the callback to veto. Code is also added to take care of the case ch1 is 'closed', and the last child is subsequently closed, it actively calls close on the mux parent then. Signed-off-by: Andy Green <andy@warmcat.com>
2011-05-25 21:41:57 +01:00
}
flush extensions when closing Extensions might be caching stuff that we should spill before a controlled close. It's not allowed to send anything on the wire after the close request, so we need to make the extensions spill just before. Signed-off-by: Andy Green <andy@warmcat.com>
2011-03-07 07:08:18 +00:00
/*
* flush any tx pending from extensions, since we may send close packet
* if there are problems with send, just nuke the connection
*/
plat change to build selection by cmake Signed-off-by: Andy Green <andy.green@linaro.org>
2014-04-02 21:02:54 +08:00
do {
flush extensions when closing Extensions might be caching stuff that we should spill before a controlled close. It's not allowed to send anything on the wire after the close request, so we need to make the extensions spill just before. Signed-off-by: Andy Green <andy@warmcat.com>
2011-03-07 07:08:18 +00:00
ret = 0;
eff_buf.token = NULL;
eff_buf.token_len = 0;
/* show every extension the new incoming data */
extension permessage deflate Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-11 11:34:01 +08:00
m = lws_ext_cb_active(wsi,
LWS_EXT_CB_FLUSH_PENDING_TX, &eff_buf, 0);
plat change to build selection by cmake Signed-off-by: Andy Green <andy.green@linaro.org>
2014-04-02 21:02:54 +08:00
if (m < 0) {
lwsl_ext("Extension reports fatal error\n");
goto just_kill_connection;
flush extensions when closing Extensions might be caching stuff that we should spill before a controlled close. It's not allowed to send anything on the wire after the close request, so we need to make the extensions spill just before. Signed-off-by: Andy Green <andy@warmcat.com>
2011-03-07 07:08:18 +00:00
}
plat change to build selection by cmake Signed-off-by: Andy Green <andy.green@linaro.org>
2014-04-02 21:02:54 +08:00
if (m)
/*
* at least one extension told us he has more
* to spill, so we will go around again after
*/
ret = 1;
flush extensions when closing Extensions might be caching stuff that we should spill before a controlled close. It's not allowed to send anything on the wire after the close request, so we need to make the extensions spill just before. Signed-off-by: Andy Green <andy@warmcat.com>
2011-03-07 07:08:18 +00:00
/* assuming they left us something to send, send it */
if (eff_buf.token_len)
if (lws_issue_raw(wsi, (unsigned char *)eff_buf.token,
cleanups after api changes and mbed update Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-04 16:54:12 +08:00
eff_buf.token_len) !=
eff_buf.token_len) {
style cleanup This brings the library sources into compliance with checkpatch style except for three or four exceptions like WIN32 related stuff and one long string constant I don't want to break into multiple sprintf calls. There should be no functional or compilability change from all this (hopefully). Signed-off-by: Andy Green <andy.green@linaro.org>
2013-02-11 17:13:32 +08:00
lwsl_debug("close: ext spill failed\n");
flush extensions when closing Extensions might be caching stuff that we should spill before a controlled close. It's not allowed to send anything on the wire after the close request, so we need to make the extensions spill just before. Signed-off-by: Andy Green <andy@warmcat.com>
2011-03-07 07:08:18 +00:00
goto just_kill_connection;
check errors on shutdown close Also make sure CLOSE doesn't go through extension munging Reduce wait for close ack to 1s Signed-off-by: Andy Green <andy.green@linaro.org>
2013-01-17 14:46:43 +08:00
}
plat change to build selection by cmake Signed-off-by: Andy Green <andy.green@linaro.org>
2014-04-02 21:02:54 +08:00
} while (ret);
flush extensions when closing Extensions might be caching stuff that we should spill before a controlled close. It's not allowed to send anything on the wire after the close request, so we need to make the extensions spill just before. Signed-off-by: Andy Green <andy@warmcat.com>
2011-03-07 07:08:18 +00:00
add extra state for waiting on close ack with timeout Signed-off-by: Andy Green <andy@warmcat.com>
2011-03-07 07:08:12 +00:00
/*
api rationalization use new names internally Change all internal uses of rationalized public apis to reflect the new names. Theer are a few things that got changed as side effect of search/replace matches, but these are almost all internal. I added a compatibility define for the public enum that got renamed. Theoretically existing code should not notice the difference from these two patches. And new code will find the new names. https://github.com/warmcat/libwebsockets/issues/357 Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-04 08:43:54 +08:00
* signal we are closing, lws_write will
add extra state for waiting on close ack with timeout Signed-off-by: Andy Green <andy@warmcat.com>
2011-03-07 07:08:12 +00:00
* add any necessary version-specific stuff. If the write fails,
* no worries we are closing anyway. If we didn't initiate this
* close, then our state has been changed to
clean internal refactor - Mainly symbol length reduction - Whitespace clean - Code refactor for linear flow - Audit @Context for API docs vs changes Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-17 17:03:59 +08:00
* LWSS_RETURNED_CLOSE_ALREADY and we will skip this.
add extra state for waiting on close ack with timeout Signed-off-by: Andy Green <andy@warmcat.com>
2011-03-07 07:08:12 +00:00
*
* Likewise if it's a second call to close this connection after we
* sent the close indication to the peer already, we are in state
clean internal refactor - Mainly symbol length reduction - Whitespace clean - Code refactor for linear flow - Audit @Context for API docs vs changes Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-17 17:03:59 +08:00
* LWSS_AWAITING_CLOSE_ACK and will skip doing this a second time.
add extra state for waiting on close ack with timeout Signed-off-by: Andy Green <andy@warmcat.com>
2011-03-07 07:08:12 +00:00
*/
multithread stability Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-26 20:56:56 +08:00
if (wsi->state_pre_close == LWSS_ESTABLISHED &&
close add api to control sent close frame contents This adds an api lws_close_reason() which lets you control what will be sent in the close frame when the connection is closed by returning nonzero from the user callback. The test server demo is extended to prove it works in both directions. With this, we should have nice close support. https://github.com/warmcat/libwebsockets/issues/196 Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-26 17:20:34 +08:00
(wsi->u.ws.close_in_ping_buffer_len || /* already a reason */
(reason != LWS_CLOSE_STATUS_NOSTATUS &&
(reason != LWS_CLOSE_STATUS_NOSTATUS_CONTEXT_DESTROY)))) {
introduce logging api and convert all library output to use it - multiple debug context calls lwsl_ err, warn, debug, parser, ext, client - api added to set which contexts output to stderr using a bitfield log_level - --disable-debug on configure removes all code that is not err or warn severity - err and warn contexts always output to stderr unless disabled by log_level - err and warn enabled by default in log_level Signed-off-by: Andy Green <andy@warmcat.com>
2013-01-10 19:50:35 +08:00
lwsl_debug("sending close indication...\n");
close add api to control sent close frame contents This adds an api lws_close_reason() which lets you control what will be sent in the close frame when the connection is closed by returning nonzero from the user callback. The test server demo is extended to prove it works in both directions. With this, we should have nice close support. https://github.com/warmcat/libwebsockets/issues/196 Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-26 17:20:34 +08:00
/* if no prepared close reason, use 1000 and no aux data */
if (!wsi->u.ws.close_in_ping_buffer_len) {
wsi->u.ws.close_in_ping_buffer_len = 2;
extension permessage deflate Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-11 11:34:01 +08:00
wsi->u.ws.ping_payload_buf[LWS_PRE] =
fix close packet index coding https://github.com/warmcat/libwebsockets/issues/792
2017-02-14 23:14:09 +08:00
(reason >> 8) & 0xff;
extension permessage deflate Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-11 11:34:01 +08:00
wsi->u.ws.ping_payload_buf[LWS_PRE + 1] =
close add api to control sent close frame contents This adds an api lws_close_reason() which lets you control what will be sent in the close frame when the connection is closed by returning nonzero from the user callback. The test server demo is extended to prove it works in both directions. With this, we should have nice close support. https://github.com/warmcat/libwebsockets/issues/196 Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-26 17:20:34 +08:00
reason & 0xff;
}
esp8266 initial support
2016-07-23 14:18:25 +08:00
#if defined (LWS_WITH_ESP8266)
wsi->close_is_pending_send_completion = 1;
#endif
add extra state for waiting on close ack with timeout Signed-off-by: Andy Green <andy@warmcat.com>
2011-03-07 07:08:12 +00:00
close: make close notification go through writable Until now we took the approach if just writing the close notification broke something, we didn't care because we were closing the connection anyway. But with lws_meta, breaking stuff in the parent connection would be a sticky problem outliving the closing child connection. So this adds a new wsi state LWSS_WAITING_TO_SEND_CLOSE_NOTIFICATION and makes the send go via the writable callback mechanism.
2017-07-17 10:11:17 +08:00
lwsl_debug("waiting for chance to send close\n");
wsi->waiting_to_send_close_frame = 1;
wsi->state = LWSS_WAITING_TO_SEND_CLOSE_NOTIFICATION;
lws_set_timeout(wsi, PENDING_TIMEOUT_CLOSE_SEND, 2);
lws_callback_on_writable(wsi);
add extra state for waiting on close ack with timeout Signed-off-by: Andy Green <andy@warmcat.com>
2011-03-07 07:08:12 +00:00
close: make close notification go through writable Until now we took the approach if just writing the close notification broke something, we didn't care because we were closing the connection anyway. But with lws_meta, breaking stuff in the parent connection would be a sticky problem outliving the closing child connection. So this adds a new wsi state LWSS_WAITING_TO_SEND_CLOSE_NOTIFICATION and makes the send go via the writable callback mechanism.
2017-07-17 10:11:17 +08:00
return;
add extra state for waiting on close ack with timeout Signed-off-by: Andy Green <andy@warmcat.com>
2011-03-07 07:08:12 +00:00
}
flush extensions when closing Extensions might be caching stuff that we should spill before a controlled close. It's not allowed to send anything on the wire after the close request, so we need to make the extensions spill just before. Signed-off-by: Andy Green <andy@warmcat.com>
2011-03-07 07:08:18 +00:00
just_kill_connection:
esp8266 initial support
2016-07-23 14:18:25 +08:00
adopt: allow binding to parent at same time - if protocol set, allocate own user_space If the child wsi wants the parent wsi user_space, it can use lws_wsi_user(lws_get_parent(child_wsi)) - raw file close processing handles parent-child relationship
2017-03-03 07:36:08 +08:00
lws_remove_child_from_any_parent(wsi);
http2: make usable HTTP/2 support is now able to serve the test server, complete with websockets, from a single vhost. - This works the same with both OpenSSL and mbedTLS. - POST is now wired up and works (also for file upload). - CGI is wired up and works. - Redirect is adapted and works - lwsws works. - URI urldecode, sanitation and argument parsing wired up for :path valgrind clean (aside from openssl-style false uninit data usage in mbedtls send occasionally) h2spec reports: $ h2spec -h 127.0.0.1 -p 7681 -t -k -o 1 ... 145 tests, 145 passed, 0 skipped, 0 failed" Incorporates: - "https://github.com/warmcat/libwebsockets/pull/1039 Fixes issue with -Werror=unused-variable flag - https://github.com/sartura/libwebsockets/commit/2c843a1395bdca5c23b40d989f508fc96aa5ba3b ssl: fix infinite loop on client cert verification failure Signed-off-by: Petar Paradzik <petar.paradzik@sartura.hr>" Caused and fixes Coverity 184887 - 184892
2017-10-13 10:33:02 +08:00
n = 0;
adopt: allow binding to parent at same time - if protocol set, allocate own user_space If the child wsi wants the parent wsi user_space, it can use lws_wsi_user(lws_get_parent(child_wsi)) - raw file close processing handles parent-child relationship
2017-03-03 07:36:08 +08:00
client: reorder close callbacks This will make sure you get - DROP_PROTOCOL - CONNECTION_ERROR - CLOSED_CLIENT_HTTP as expected for failed client connections.
2017-10-15 12:15:28 +08:00
if (wsi->user_space) {
lwsl_debug("%s: %p: DROP_PROTOCOL %s\n", __func__, wsi,
wsi->protocol->name);
wsi->protocol->callback(wsi,
LWS_CALLBACK_HTTP_DROP_PROTOCOL,
wsi->user_space, NULL, 0);
}
if ((wsi->mode == LWSCM_WSCL_WAITING_SERVER_REPLY ||
wsi->mode == LWSCM_WSCL_WAITING_CONNECT) &&
!wsi->already_did_cce) {
wsi->vhost->protocols[0].callback(wsi,
LWS_CALLBACK_CLIENT_CONNECTION_ERROR,
wsi->user_space, NULL, 0);
}
if (wsi->mode & LWSCM_FLAG_IMPLIES_CALLBACK_CLOSED_CLIENT_HTTP) {
wsi->vhost->protocols[0].callback(wsi,
LWS_CALLBACK_CLOSED_CLIENT_HTTP,
wsi->user_space, NULL, 0);
wsi->told_user_closed = 1;
}
multithread stability Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-26 20:56:56 +08:00
#if LWS_POSIX
/*
* Testing with ab shows that we have to stage the socket close when
* the system is under stress... shutdown any further TX, change the
* state to one that won't emit anything more, and wait with a timeout
* for the POLLIN to show a zero-size rx before coming back and doing
* the actual close.
*/
esp32: separate factory setup
2017-03-16 10:46:31 +08:00
if (wsi->mode != LWSCM_RAW &&
!(wsi->mode & LWSCM_FLAG_IMPLIES_CALLBACK_CLOSED_CLIENT_HTTP) &&
wsi->state != LWSS_SHUTDOWN &&
client fix reaction to tls failure https://github.com/warmcat/libwebsockets/issues/508 Signed-off-by: Andy Green <andy@warmcat.com>
2016-05-03 07:26:10 +08:00
wsi->state != LWSS_CLIENT_UNCONNECTED &&
libuv integration This gets the libuv stuff plumbed in and working. Currently it's only workable for some service thread, and there is an isolated valgrind problem left ==28425== 128 bytes in 1 blocks are definitely lost in loss record 3 of 3 ==28425== at 0x4C28C50: malloc (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so) ==28425== by 0x4C2AB1E: realloc (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so) ==28425== by 0x58BBB27: maybe_resize (core.c:748) ==28425== by 0x58BBB27: uv__io_start (core.c:787) ==28425== by 0x58C1B80: uv__signal_loop_once_init (signal.c:225) ==28425== by 0x58C1B80: uv_signal_init (signal.c:260) ==28425== by 0x58BF7A6: uv_loop_init (loop.c:66) ==28425== by 0x4157F5: lws_uv_initloop (libuv.c:89) ==28425== by 0x405536: main (test-server-libuv.c:284) libuv wants to sign off on all libuv 'handles' that will close, and callback to do the close confirmation asynchronously. The wsi close function is adapted when libuv is in use to work with libuv accordingly and exit the uv loop the number of remaining wsi is zero. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-14 09:27:41 +08:00
reason != LWS_CLOSE_STATUS_NOSTATUS_CONTEXT_DESTROY &&
!wsi->socket_is_permanently_unusable) {
ssl close: do explicit ssl shutdown instead of socket shutdown if ssl mode
2017-02-20 05:44:56 +08:00
#ifdef LWS_OPENSSL_SUPPORT
clean
2017-09-23 12:55:21 +08:00
if (lws_is_ssl(wsi) && wsi->ssl) {
ssl close: do explicit ssl shutdown instead of socket shutdown if ssl mode
2017-02-20 05:44:56 +08:00
n = SSL_shutdown(wsi->ssl);
clean
2017-09-23 12:55:21 +08:00
/*
* If finished the SSL shutdown, then do socket
* shutdown, else need to retry SSL shutdown
*/
switch (n) {
case 0:
Subject: ssl: stop spinning on close
2017-04-06 19:37:14 +02:00
lws_change_pollfd(wsi, LWS_POLLOUT, LWS_POLLIN);
clean
2017-09-23 12:55:21 +08:00
break;
case 1:
n = shutdown(wsi->desc.sockfd, SHUT_WR);
break;
default:
http2: make usable HTTP/2 support is now able to serve the test server, complete with websockets, from a single vhost. - This works the same with both OpenSSL and mbedTLS. - POST is now wired up and works (also for file upload). - CGI is wired up and works. - Redirect is adapted and works - lwsws works. - URI urldecode, sanitation and argument parsing wired up for :path valgrind clean (aside from openssl-style false uninit data usage in mbedtls send occasionally) h2spec reports: $ h2spec -h 127.0.0.1 -p 7681 -t -k -o 1 ... 145 tests, 145 passed, 0 skipped, 0 failed" Incorporates: - "https://github.com/warmcat/libwebsockets/pull/1039 Fixes issue with -Werror=unused-variable flag - https://github.com/sartura/libwebsockets/commit/2c843a1395bdca5c23b40d989f508fc96aa5ba3b ssl: fix infinite loop on client cert verification failure Signed-off-by: Petar Paradzik <petar.paradzik@sartura.hr>" Caused and fixes Coverity 184887 - 184892
2017-10-13 10:33:02 +08:00
if (SSL_want_read(wsi->ssl)) {
lws_change_pollfd(wsi, 0, LWS_POLLIN);
Subject: ssl: stop spinning on close
2017-04-06 19:37:14 +02:00
n = 0;
clean
2017-09-23 12:55:21 +08:00
break;
http2: make usable HTTP/2 support is now able to serve the test server, complete with websockets, from a single vhost. - This works the same with both OpenSSL and mbedTLS. - POST is now wired up and works (also for file upload). - CGI is wired up and works. - Redirect is adapted and works - lwsws works. - URI urldecode, sanitation and argument parsing wired up for :path valgrind clean (aside from openssl-style false uninit data usage in mbedtls send occasionally) h2spec reports: $ h2spec -h 127.0.0.1 -p 7681 -t -k -o 1 ... 145 tests, 145 passed, 0 skipped, 0 failed" Incorporates: - "https://github.com/warmcat/libwebsockets/pull/1039 Fixes issue with -Werror=unused-variable flag - https://github.com/sartura/libwebsockets/commit/2c843a1395bdca5c23b40d989f508fc96aa5ba3b ssl: fix infinite loop on client cert verification failure Signed-off-by: Petar Paradzik <petar.paradzik@sartura.hr>" Caused and fixes Coverity 184887 - 184892
2017-10-13 10:33:02 +08:00
}
if (SSL_want_write(wsi->ssl)) {
lws_change_pollfd(wsi, 0, LWS_POLLOUT);
Subject: ssl: stop spinning on close
2017-04-06 19:37:14 +02:00
n = 0;
clean
2017-09-23 12:55:21 +08:00
break;
Subject: ssl: stop spinning on close
2017-04-06 19:37:14 +02:00
}
http2: make usable HTTP/2 support is now able to serve the test server, complete with websockets, from a single vhost. - This works the same with both OpenSSL and mbedTLS. - POST is now wired up and works (also for file upload). - CGI is wired up and works. - Redirect is adapted and works - lwsws works. - URI urldecode, sanitation and argument parsing wired up for :path valgrind clean (aside from openssl-style false uninit data usage in mbedtls send occasionally) h2spec reports: $ h2spec -h 127.0.0.1 -p 7681 -t -k -o 1 ... 145 tests, 145 passed, 0 skipped, 0 failed" Incorporates: - "https://github.com/warmcat/libwebsockets/pull/1039 Fixes issue with -Werror=unused-variable flag - https://github.com/sartura/libwebsockets/commit/2c843a1395bdca5c23b40d989f508fc96aa5ba3b ssl: fix infinite loop on client cert verification failure Signed-off-by: Petar Paradzik <petar.paradzik@sartura.hr>" Caused and fixes Coverity 184887 - 184892
2017-10-13 10:33:02 +08:00
n = shutdown(wsi->desc.sockfd, SHUT_WR);
break;
Subject: ssl: stop spinning on close
2017-04-06 19:37:14 +02:00
}
clean
2017-09-23 12:55:21 +08:00
} else
ssl close: do explicit ssl shutdown instead of socket shutdown if ssl mode
2017-02-20 05:44:56 +08:00
#endif
{
http2: make usable HTTP/2 support is now able to serve the test server, complete with websockets, from a single vhost. - This works the same with both OpenSSL and mbedTLS. - POST is now wired up and works (also for file upload). - CGI is wired up and works. - Redirect is adapted and works - lwsws works. - URI urldecode, sanitation and argument parsing wired up for :path valgrind clean (aside from openssl-style false uninit data usage in mbedtls send occasionally) h2spec reports: $ h2spec -h 127.0.0.1 -p 7681 -t -k -o 1 ... 145 tests, 145 passed, 0 skipped, 0 failed" Incorporates: - "https://github.com/warmcat/libwebsockets/pull/1039 Fixes issue with -Werror=unused-variable flag - https://github.com/sartura/libwebsockets/commit/2c843a1395bdca5c23b40d989f508fc96aa5ba3b ssl: fix infinite loop on client cert verification failure Signed-off-by: Petar Paradzik <petar.paradzik@sartura.hr>" Caused and fixes Coverity 184887 - 184892
2017-10-13 10:33:02 +08:00
lwsl_info("%s: shutdown conn: %p (sock %d, state %d)\n",
clean
2017-09-23 12:55:21 +08:00
__func__, wsi, (int)(long)wsi->desc.sockfd,
wsi->state);
http2: make usable HTTP/2 support is now able to serve the test server, complete with websockets, from a single vhost. - This works the same with both OpenSSL and mbedTLS. - POST is now wired up and works (also for file upload). - CGI is wired up and works. - Redirect is adapted and works - lwsws works. - URI urldecode, sanitation and argument parsing wired up for :path valgrind clean (aside from openssl-style false uninit data usage in mbedtls send occasionally) h2spec reports: $ h2spec -h 127.0.0.1 -p 7681 -t -k -o 1 ... 145 tests, 145 passed, 0 skipped, 0 failed" Incorporates: - "https://github.com/warmcat/libwebsockets/pull/1039 Fixes issue with -Werror=unused-variable flag - https://github.com/sartura/libwebsockets/commit/2c843a1395bdca5c23b40d989f508fc96aa5ba3b ssl: fix infinite loop on client cert verification failure Signed-off-by: Petar Paradzik <petar.paradzik@sartura.hr>" Caused and fixes Coverity 184887 - 184892
2017-10-13 10:33:02 +08:00
if (!wsi->socket_is_permanently_unusable &&
lws_sockfd_valid(wsi->desc.sockfd))
n = shutdown(wsi->desc.sockfd, SHUT_WR);
ssl close: do explicit ssl shutdown instead of socket shutdown if ssl mode
2017-02-20 05:44:56 +08:00
}
multithread stability Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-26 20:56:56 +08:00
if (n)
clean
2017-09-23 12:55:21 +08:00
lwsl_debug("closing: shutdown (state %d) ret %d\n",
wsi->state, LWS_ERRNO);
defeat POLLOUT if socket in shutdown wait After andrejs.hanins@ubnt.com Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-24 21:27:46 +08:00
clean
2017-09-23 12:55:21 +08:00
/*
* This causes problems on WINCE / ESP32 with disconnection
* when the events are half closing connection
*/
esp32: separate factory setup
2017-03-16 10:46:31 +08:00
#if !defined(_WIN32_WCE) && !defined(LWS_WITH_ESP32)
libuv when in use skip shutdown close phase Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-29 10:34:29 +08:00
/* libuv: no event available to guarantee completion */
http2: make usable HTTP/2 support is now able to serve the test server, complete with websockets, from a single vhost. - This works the same with both OpenSSL and mbedTLS. - POST is now wired up and works (also for file upload). - CGI is wired up and works. - Redirect is adapted and works - lwsws works. - URI urldecode, sanitation and argument parsing wired up for :path valgrind clean (aside from openssl-style false uninit data usage in mbedtls send occasionally) h2spec reports: $ h2spec -h 127.0.0.1 -p 7681 -t -k -o 1 ... 145 tests, 145 passed, 0 skipped, 0 failed" Incorporates: - "https://github.com/warmcat/libwebsockets/pull/1039 Fixes issue with -Werror=unused-variable flag - https://github.com/sartura/libwebsockets/commit/2c843a1395bdca5c23b40d989f508fc96aa5ba3b ssl: fix infinite loop on client cert verification failure Signed-off-by: Petar Paradzik <petar.paradzik@sartura.hr>" Caused and fixes Coverity 184887 - 184892
2017-10-13 10:33:02 +08:00
if (!wsi->socket_is_permanently_unusable &&
lws_sockfd_valid(wsi->desc.sockfd) &&
!LWS_LIBUV_ENABLED(context)) {
libuv when in use skip shutdown close phase Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-29 10:34:29 +08:00
lws_change_pollfd(wsi, LWS_POLLOUT, LWS_POLLIN);
wsi->state = LWSS_SHUTDOWN;
lws_set_timeout(wsi, PENDING_TIMEOUT_SHUTDOWN_FLUSH,
context->timeout_secs);
esp32: separate factory setup
2017-03-16 10:46:31 +08:00
libuv when in use skip shutdown close phase Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-29 10:34:29 +08:00
return;
}
wince minor adaptations https://github.com/warmcat/libwebsockets/issues/444
2016-02-29 18:48:55 +08:00
#endif
multithread stability Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-26 20:56:56 +08:00
}
#endif
clean
2017-09-23 12:55:21 +08:00
lwsl_debug("%s: real just_kill_connection: %p (sockfd %d)\n", __func__,
wsi, wsi->desc.sockfd);
esp8266 initial support
2016-07-23 14:18:25 +08:00
cgi env Improve cgi support so it's capable of running cgit
2016-03-20 11:55:25 +08:00
#ifdef LWS_WITH_HTTP_PROXY
proxy rewrite If you enable -DLWS_WITH_HTTP_PROXY=1 at cmake, the test server has a new URI path http://localhost:7681/proxytest If you visit here, a client connection to http://example.com:80 is spawned, and the results piped on to your original connection. Also with LWS_WITH_HTTP_PROXY enabled at cmake, lws wants to link to an additional library, "libhubbub". This allows lws to do html rewriting on the fly, adjusting proxied urls in a lightweight and fast way.
2016-03-20 11:59:53 +08:00
if (wsi->rw) {
lws_rewrite_destroy(wsi->rw);
wsi->rw = NULL;
}
cgi env Improve cgi support so it's capable of running cgit
2016-03-20 11:55:25 +08:00
#endif
add extra state for waiting on close ack with timeout Signed-off-by: Andy Green <andy@warmcat.com>
2011-03-07 07:08:12 +00:00
/*
* we won't be servicing or receiving anything further from this guy
replace hashtable polltable management This rips out the connection hashtable implementation along with MAX_CLIENTS and replaces it with a dynamically allocated fds array and lookup table along the same lines as the new extpoll implementation from Edwin van den Oetelaar. It detects the max number of file descriptors possible at context init time and allocates accordingly; this can be externally controlled by ulimit and the server run as a specific user to facilitate targeting specific ulimit rules at it. Many operations that translated between socket descriptors and struct websocket or pollfd objects have had iteration removed by this patch and under load will be a lot faster. Signed-off-by: Andy Green <andy.green@linaro.org>
2013-01-17 12:26:48 +08:00
* delete socket from the internal poll list if still present
add extra state for waiting on close ack with timeout Signed-off-by: Andy Green <andy@warmcat.com>
2011-03-07 07:08:12 +00:00
*/
refactor needless context with wsi paramater passing Now we bit the bullet and gave each wsi an lws_context *, many internal apis that take both a context and wsi parameter only need the wsi. Also simplify parser code by making a temp var for allocated_headers * instead of the longwinded dereference chain everywhere. Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-15 21:15:58 +08:00
lws_ssl_remove_wsi_from_buffered_list(wsi);
timeout move to doubly linked list In the case we have a lot of connections, checking them all for timeout state once a second becomes burdensome. At the moment if you have 100K connections, once a second they all get checked for timeout in a loop. This patch adds a doubly-linked list based in the context to each wsi, and only wsi with pending timeouts appear on it. At checking time, we traverse the list, which costs nothing if empty because nobody has a pending timeout. Similarly adding and removing from the list costs almost nothing since no iteration is required no matter how big the list. The extra 8 or 16 bytes in the wsi are offset a little bit by demoting .pps from int to char (save 3 bytes). And trim max act exts to 2, since we only provide one, saving 8 /16 bytes by itself if exts enabled. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-19 04:32:14 +08:00
lws_remove_from_timeout_list(wsi);
ssl pending buffered reads use linked list Signed-off-by: Andy Green <andy.green@linaro.org>
2015-01-29 08:36:18 +08:00
cleanups after api changes and mbed update Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-04 16:54:12 +08:00
/* checking return redundant since we anyway close */
adoption: make union for socket and file fds This lets lws support adopting raw file FDs and raw socket fds. A test plugin creates a FIFO and prints data sent on it, using the lws event loop.
2017-02-27 12:55:56 +08:00
if (wsi->desc.sockfd != LWS_SOCK_INVALID)
coverity 169271 - take care about sockfd of -1 on close
2016-08-28 09:21:56 +08:00
remove_wsi_socket_from_fds(wsi);
lws-meta
2017-07-19 04:39:14 +08:00
else
lws_same_vh_protocol_remove(wsi);
unify server and client close Signed-off-by: Andy Green <andy@warmcat.com>
2011-02-14 08:03:48 +00:00
esp8266 initial support
2016-07-23 14:18:25 +08:00
#if defined(LWS_WITH_ESP8266)
adoption: make union for socket and file fds This lets lws support adopting raw file FDs and raw socket fds. A test plugin creates a FIFO and prints data sent on it, using the lws event loop.
2017-02-27 12:55:56 +08:00
espconn_disconnect(wsi->desc.sockfd);
esp8266 initial support
2016-07-23 14:18:25 +08:00
#endif
clean internal refactor - Mainly symbol length reduction - Whitespace clean - Code refactor for linear flow - Audit @Context for API docs vs changes Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-17 17:03:59 +08:00
wsi->state = LWSS_DEAD_SOCKET;
Subject: [PATCH] Introduce lws_free2() helper to free and re-NULL pointers Signed-off-by: Alejandro Mery <amery@geeks.cl>
2014-12-05 00:09:20 +01:00
clean internal refactor - Mainly symbol length reduction - Whitespace clean - Code refactor for linear flow - Audit @Context for API docs vs changes Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-17 17:03:59 +08:00
lws_free_set_NULL(wsi->rxflow_buffer);
multithread stability Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-26 20:56:56 +08:00
if (wsi->state_pre_close == LWSS_ESTABLISHED ||
clean internal refactor - Mainly symbol length reduction - Whitespace clean - Code refactor for linear flow - Audit @Context for API docs vs changes Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-17 17:03:59 +08:00
wsi->mode == LWSCM_WS_SERVING ||
wsi->mode == LWSCM_WS_CLIENT) {
security harden http parser a bit Drop the connection during parsing for a few more cases that can't be legit. Take care about trying to free rxflow_buffer only if we reached a connmode where it exists Change behaviour on setting unknown HTTP method to kill connection Signed-off-by: Andy Green <andy.green@linaro.org>
2013-02-12 12:52:39 +08:00
extension permessage deflate Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-11 11:34:01 +08:00
if (wsi->u.ws.rx_draining_ext) {
multithreaded service This adds support for multithreaded service to lws without adding any threading or locking code in the library. At context creation time you can request split the service part of the context into n service domains, which are load-balanced so that the most idle one gets the next listen socket accept. There's a single listen socket on one port still. User code may then spawn n threads doing n service loops / poll()s simultaneously. Locking is only required (I think) in the existing FD lock callbacks already handled by the pthreads server example, and that locking takes place in user code. So the library remains completely agnostic about the threading / locking scheme. And by default, it's completely compatible with one service thread so no changes are required by people uninterested in multithreaded service. However for people interested in extremely lightweight mass http[s]/ ws[s] service with minimum provisioning, the library can now do everything out of the box. To test it, just try $ libwebsockets-test-server-pthreads -j 8 where -j controls the number of service threads Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-19 03:34:24 +08:00
struct lws **w = &pt->rx_draining_ext_list;
extension permessage deflate Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-11 11:34:01 +08:00
wsi->u.ws.rx_draining_ext = 0;
/* remove us from context draining ext list */
while (*w) {
if (*w == wsi) {
*w = wsi->u.ws.rx_draining_ext_list;
break;
}
w = &((*w)->u.ws.rx_draining_ext_list);
}
wsi->u.ws.rx_draining_ext_list = NULL;
}
if (wsi->u.ws.tx_draining_ext) {
multithreaded service This adds support for multithreaded service to lws without adding any threading or locking code in the library. At context creation time you can request split the service part of the context into n service domains, which are load-balanced so that the most idle one gets the next listen socket accept. There's a single listen socket on one port still. User code may then spawn n threads doing n service loops / poll()s simultaneously. Locking is only required (I think) in the existing FD lock callbacks already handled by the pthreads server example, and that locking takes place in user code. So the library remains completely agnostic about the threading / locking scheme. And by default, it's completely compatible with one service thread so no changes are required by people uninterested in multithreaded service. However for people interested in extremely lightweight mass http[s]/ ws[s] service with minimum provisioning, the library can now do everything out of the box. To test it, just try $ libwebsockets-test-server-pthreads -j 8 where -j controls the number of service threads Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-19 03:34:24 +08:00
struct lws **w = &pt->tx_draining_ext_list;
extension permessage deflate Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-11 11:34:01 +08:00
wsi->u.ws.tx_draining_ext = 0;
/* remove us from context draining ext list */
while (*w) {
if (*w == wsi) {
*w = wsi->u.ws.tx_draining_ext_list;
break;
}
w = &((*w)->u.ws.tx_draining_ext_list);
}
wsi->u.ws.tx_draining_ext_list = NULL;
}
lws_free_set_NULL(wsi->u.ws.rx_ubuf);
http2 can keep upgraded connection up Signed-off-by: Andy Green <andy.green@linaro.org>
2014-10-08 12:00:53 +08:00
clean internal refactor - Mainly symbol length reduction - Whitespace clean - Code refactor for linear flow - Audit @Context for API docs vs changes Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-17 17:03:59 +08:00
if (wsi->trunc_alloc)
add explicit error for partial send This patch adds code to handle the situation that a prepared user buffer could not all be sent on the socket at once. There are two kinds of situation to handle 1) User code handles it: The connection only has extensions active that do not rewrite the buffer. In this case, the patch caused libwebsocket_write() to simply return the amount of user buffer that was consumed (this is specifically the amount of user buffer used in sending what was accepted, nothing else). So user code can just advance its buffer that much and resume sending when the socket is writable again. This continues the frame rather than starting a new one or new fragment. 2) The connections has extensions active which actually send something quite different than what the user buffer contains, for example a compression extension. In this case, libwebsockets will dynamically malloc a buffer to contain a copy of the remaining unsent data, request notifiction when writeable again, and automatically spill and free this buffer with the highest priority before passing on the writable notification to anything else. For this situation, the call to write will return that it used the whole user buffer, even though part is still rebuffered. This patch should enable libwebsockets to detect the two cases and take the appropriate action. There are also two choices for user code to deal with partial sends. 1) Leave the no_buffer_all_partial_tx member in the protocol struct at zero. The library will dyamically buffer anything you send that did not get completely written to the socket, and automatically spill it next time the socket is writable. You can use this method if your sent frames are relatvely small and unlikely to get truncated anyway. 2) Set the no_buffer_all_partial_tx member in the protocol struct. User code now needs to take care of the return value from libwebsocket_write() and deal with resending the remainder if not all of the requested amount got sent. You should use this method if you are sending large messages and want to maximize throughput and efficiency. Since the new member no_buffer_all_partial_tx will be zero by default, this patch will auto-rebuffer any partial sends by default. That's good for most cases but if you attempt to send large blocks, make sure you follow option 2) above. Signed-off-by: Andy Green <andy.green@linaro.org>
2013-10-17 08:09:19 +08:00
/* not going to be completed... nuke it */
clean internal refactor - Mainly symbol length reduction - Whitespace clean - Code refactor for linear flow - Audit @Context for API docs vs changes Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-17 17:03:59 +08:00
lws_free_set_NULL(wsi->trunc_alloc);
cleanups after api changes and mbed update Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-04 16:54:12 +08:00
ws ping buffer allocate in wsi ws union member Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-25 13:14:09 +08:00
wsi->u.ws.ping_payload_len = 0;
wsi->u.ws.ping_pending_flag = 0;
remove fixed rx buffer allow definition per protocol A new protocol member is defined that controls the size of rx buffer allocation per connection. For compatibility 0 size allocates 4096, but you should adapt your protocol definition array in the user code to declare an appropriate value. See the changelog for more detail. The advantage is the rx frame buffer size is now tailored to what is expected from the protocol, rather than being fixed to a default of 4096. If your protocol only sends frames of a dozen bytes this allows you to only allocate an rx frame buffer of the same size. For example the per-connection allocation (excluding headers) for the test server fell from ~4500 to < 750 bytes with this. Signed-off-by: Andy Green <andy.green@linaro.org>
2013-02-06 21:10:16 +09:00
}
unify server and client close Signed-off-by: Andy Green <andy@warmcat.com>
2011-02-14 08:03:48 +00:00
/* tell the user it's all over for this guy */
client: reorder close callbacks This will make sure you get - DROP_PROTOCOL - CONNECTION_ERROR - CLOSED_CLIENT_HTTP as expected for failed client connections.
2017-10-15 12:15:28 +08:00
if (!wsi->told_user_closed &&
wsi->mode != LWSCM_RAW && wsi->protocol &&
clean
2017-09-23 12:55:21 +08:00
wsi->protocol->callback &&
http2: make usable HTTP/2 support is now able to serve the test server, complete with websockets, from a single vhost. - This works the same with both OpenSSL and mbedTLS. - POST is now wired up and works (also for file upload). - CGI is wired up and works. - Redirect is adapted and works - lwsws works. - URI urldecode, sanitation and argument parsing wired up for :path valgrind clean (aside from openssl-style false uninit data usage in mbedtls send occasionally) h2spec reports: $ h2spec -h 127.0.0.1 -p 7681 -t -k -o 1 ... 145 tests, 145 passed, 0 skipped, 0 failed" Incorporates: - "https://github.com/warmcat/libwebsockets/pull/1039 Fixes issue with -Werror=unused-variable flag - https://github.com/sartura/libwebsockets/commit/2c843a1395bdca5c23b40d989f508fc96aa5ba3b ssl: fix infinite loop on client cert verification failure Signed-off-by: Petar Paradzik <petar.paradzik@sartura.hr>" Caused and fixes Coverity 184887 - 184892
2017-10-13 10:33:02 +08:00
(wsi->state_pre_close == LWSS_ESTABLISHED ||
wsi->state_pre_close == LWSS_HTTP2_ESTABLISHED ||
wsi->state_pre_close == LWSS_HTTP_BODY ||
wsi->state_pre_close == LWSS_HTTP ||
wsi->state_pre_close == LWSS_RETURNED_CLOSE_ALREADY ||
wsi->state_pre_close == LWSS_AWAITING_CLOSE_ACK ||
wsi->state_pre_close == LWSS_WAITING_TO_SEND_CLOSE_NOTIFICATION ||
wsi->state_pre_close == LWSS_FLUSHING_STORED_SEND_BEFORE_CLOSE ||
client callback closed if not upgraded also for server connection https://github.com/warmcat/libwebsockets/issues/437 https://github.com/warmcat/libwebsockets/pull/440
2016-02-24 12:54:37 +02:00
(wsi->mode == LWSCM_WS_CLIENT && wsi->state_pre_close == LWSS_HTTP) ||
(wsi->mode == LWSCM_WS_SERVING && wsi->state_pre_close == LWSS_HTTP))) {
client: reorder close callbacks This will make sure you get - DROP_PROTOCOL - CONNECTION_ERROR - CLOSED_CLIENT_HTTP as expected for failed client connections.
2017-10-15 12:15:28 +08:00
lwsl_debug("calling back CLOSED %d %d\n", wsi->mode, wsi->state);
public api remove context from user callback API BREAK Since struct lws (wsi) now has his own context pointer, we were able to remove the need for passing context almost everywhere in the apis. In turn, that means there's no real use for context being passed to every callback; in the rare cases context is needed user code can get it with lws_get_ctx(wsi) Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-17 07:54:44 +08:00
wsi->protocol->callback(wsi, LWS_CALLBACK_CLOSED,
cleanups after api changes and mbed update Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-04 16:54:12 +08:00
wsi->user_space, NULL, 0);
clean internal refactor - Mainly symbol length reduction - Whitespace clean - Code refactor for linear flow - Audit @Context for API docs vs changes Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-17 17:03:59 +08:00
} else if (wsi->mode == LWSCM_HTTP_SERVING_ACCEPTED) {
Added additional LWS_CALLBACK_CLOSED_HTTP callback to just_kill_connection branch of close_and_free_session
2013-06-29 10:16:18 +08:00
lwsl_debug("calling back CLOSED_HTTP\n");
introduce urlencode decode and sql escape public apis This adds - simple lws_urlencode() - simple lws_urldecode() - simple lws_sql_purify Those expect the data to all be there and process it up until the first '\0'. There is also a larger opaque apis for handling POST_BODY urldecode. To enable these, you need to give cmake -DLWS_WITH_STATEFUL_URLDECODE=1 (or arrange any larger feature that relies on it sets that in CMakeLists.txt) - stateful urldecode with parameter array These have create / process / destroy semantics on a struct that maintains decode state. Stateful urldecode is capable of dealing with large POST data in multiple POST_BODY callbacks cleanly, eg, file transfer by POST. Stateful urldecode with parameter array wraps the above with a canned callback that stores the urldecoded data and indexes them in a pointer array matching an array of parameter names. You may also pass it an optional callback when creating it, that will recieve uploaded file content. The test html is updated to support both urlencoded and multipart forms, with some javascript to do clientside validation of an arbitrary 100KB file size limit (there is no file size limit in the apis). Signed-off-by: Andy Green <andy@warmcat.com>
2016-06-08 10:07:02 +08:00
wsi->vhost->protocols->callback(wsi, LWS_CALLBACK_CLOSED_HTTP,
public api remove context from user callback API BREAK Since struct lws (wsi) now has his own context pointer, we were able to remove the need for passing context almost everywhere in the apis. In turn, that means there's no real use for context being passed to every callback; in the rare cases context is needed user code can get it with lws_get_ctx(wsi) Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-17 07:54:44 +08:00
wsi->user_space, NULL, 0 );
debug move most spew into debug macro This quietens the spew so all typical debug info now is coming from the user code (mirror protocol in the sample server / client case). Signed-off-by: Andy Green <andy@warmcat.com>
2011-11-07 19:53:23 +08:00
} else
client error on timeout before connection Signed-off-by: Andy Green <andy.green@linaro.org>
2015-04-16 19:55:42 +08:00
lwsl_debug("not calling back closed mode=%d state=%d\n",
multithread stability Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-26 20:56:56 +08:00
wsi->mode, wsi->state_pre_close);
add-per-session-user-data.patch Signed-off-by: Andy Green <andy@warmcat.com>
2010-11-03 11:13:06 +00:00
deallocate extension contexts on connection close Signed-off-by: Andy Green <andy@warmcat.com>
2011-03-06 10:29:38 +00:00
/* deallocate any active extension contexts */
whitespace trailing mass cleanout Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-14 08:52:03 +08:00
extension permessage deflate Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-11 11:34:01 +08:00
if (lws_ext_cb_active(wsi, LWS_EXT_CB_DESTROY, NULL, 0) < 0)
plat change to build selection by cmake Signed-off-by: Andy Green <andy.green@linaro.org>
2014-04-02 21:02:54 +08:00
lwsl_warn("extension destruction failed\n");
introduce x google mux very draft indeed This is initial x-google-mux support. It's disabled by default since it's very pre-alpha. 1) To enable it, reconfigure with --enable-x-google-mux 2) It conflicts with deflate-stream, use the -u switch on the test client to disable deflate-stream 3) It deviates from the google standard by sending full headers in the addchannel subcommand rather than just changed ones from original connect 4) Quota is not implemented yet 5) Close of subchannel is not really implemented yet 6) Google opcode 0xf is changed to 0x7 to account for v7 protocol changes to opcode layout However despite those caveats, in fact it can run the test client reliably over one socket (both dumb-increment and lws-mirror-protocol), you can open a browser on the same test server too and see the circles, etc. Signed-off-by: Andy Green <andy@warmcat.com>
2011-05-23 10:00:03 +01:00
/*
* inform all extensions in case they tracked this guy out of band
* even though not active on him specifically
*/
clean internal refactor - Mainly symbol length reduction - Whitespace clean - Code refactor for linear flow - Audit @Context for API docs vs changes Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-17 17:03:59 +08:00
if (lws_ext_cb_all_exts(context, wsi,
extension permessage deflate Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-11 11:34:01 +08:00
LWS_EXT_CB_DESTROY_ANY_WSI_CLOSING, NULL, 0) < 0)
plat change to build selection by cmake Signed-off-by: Andy Green <andy.green@linaro.org>
2014-04-02 21:02:54 +08:00
lwsl_warn("ext destroy wsi failed\n");
introduce x google mux very draft indeed This is initial x-google-mux support. It's disabled by default since it's very pre-alpha. 1) To enable it, reconfigure with --enable-x-google-mux 2) It conflicts with deflate-stream, use the -u switch on the test client to disable deflate-stream 3) It deviates from the google standard by sending full headers in the addchannel subcommand rather than just changed ones from original connect 4) Quota is not implemented yet 5) Close of subchannel is not really implemented yet 6) Google opcode 0xf is changed to 0x7 to account for v7 protocol changes to opcode layout However despite those caveats, in fact it can run the test client reliably over one socket (both dumb-increment and lws-mirror-protocol), you can open a browser on the same test server too and see the circles, etc. Signed-off-by: Andy Green <andy@warmcat.com>
2011-05-23 10:00:03 +01:00
fops-zip
2017-03-03 12:38:10 +08:00
async_close:
multithread stability Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-26 20:56:56 +08:00
wsi->socket_is_permanently_unusable = 1;
CMake: convert all LWS_USE_... to LWS_WITH_... Almost all the CMake options begin with LWS_WITH_..., but many of the symbols passed to lws are LWS_USE_... , this causes neededless confusion, compounded by the fact that a few CMake options also begin with LWS_USE_. This patch globally converts all LWS_USE_... to LWS_WITH_..., so there is only one prefix to remember in both CMake and the code. The affected public CMake options are LWS_USE_BORINGSSL -> LWS_WITH_BORINGSSL LWS_USE_CYASSL -> LWS_WITH_CYASSL LWS_USE_WOLFSSL -> LWS_WITH_WOLFSSL LWS_USE_MBEDTLS -> LWS_WITH_MBEDTLS LWS_USE_BUNDLED_ZLIB -> LWS_WITH_BUNDLED_ZLIB
2017-09-28 11:29:03 +08:00
#ifdef LWS_WITH_LIBUV
http2: make usable HTTP/2 support is now able to serve the test server, complete with websockets, from a single vhost. - This works the same with both OpenSSL and mbedTLS. - POST is now wired up and works (also for file upload). - CGI is wired up and works. - Redirect is adapted and works - lwsws works. - URI urldecode, sanitation and argument parsing wired up for :path valgrind clean (aside from openssl-style false uninit data usage in mbedtls send occasionally) h2spec reports: $ h2spec -h 127.0.0.1 -p 7681 -t -k -o 1 ... 145 tests, 145 passed, 0 skipped, 0 failed" Incorporates: - "https://github.com/warmcat/libwebsockets/pull/1039 Fixes issue with -Werror=unused-variable flag - https://github.com/sartura/libwebsockets/commit/2c843a1395bdca5c23b40d989f508fc96aa5ba3b ssl: fix infinite loop on client cert verification failure Signed-off-by: Petar Paradzik <petar.paradzik@sartura.hr>" Caused and fixes Coverity 184887 - 184892
2017-10-13 10:33:02 +08:00
if (!wsi->parent_carries_io &&
lws_sockfd_valid(wsi->desc.sockfd))
lws-meta
2017-07-19 04:39:14 +08:00
if (LWS_LIBUV_ENABLED(context)) {
if (wsi->listener) {
clean
2017-09-23 12:55:21 +08:00
lwsl_debug("%s: stop listener poll\n", __func__);
lws-meta
2017-07-19 04:39:14 +08:00
uv_poll_stop(&wsi->w_read.uv_watcher);
}
clean
2017-09-23 12:55:21 +08:00
lwsl_debug("%s: lws_libuv_closehandle: wsi %p\n",
__func__, wsi);
/*
* libuv has to do his own close handle processing
* asynchronously
*/
lws-meta
2017-07-19 04:39:14 +08:00
lws_libuv_closehandle(wsi);
libuv integration This gets the libuv stuff plumbed in and working. Currently it's only workable for some service thread, and there is an isolated valgrind problem left ==28425== 128 bytes in 1 blocks are definitely lost in loss record 3 of 3 ==28425== at 0x4C28C50: malloc (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so) ==28425== by 0x4C2AB1E: realloc (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so) ==28425== by 0x58BBB27: maybe_resize (core.c:748) ==28425== by 0x58BBB27: uv__io_start (core.c:787) ==28425== by 0x58C1B80: uv__signal_loop_once_init (signal.c:225) ==28425== by 0x58C1B80: uv_signal_init (signal.c:260) ==28425== by 0x58BF7A6: uv_loop_init (loop.c:66) ==28425== by 0x4157F5: lws_uv_initloop (libuv.c:89) ==28425== by 0x405536: main (test-server-libuv.c:284) libuv wants to sign off on all libuv 'handles' that will close, and callback to do the close confirmation asynchronously. The wsi close function is adapted when libuv is in use to work with libuv accordingly and exit the uv loop the number of remaining wsi is zero. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-14 09:27:41 +08:00
lws-meta
2017-07-19 04:39:14 +08:00
return;
}
libuv integration This gets the libuv stuff plumbed in and working. Currently it's only workable for some service thread, and there is an isolated valgrind problem left ==28425== 128 bytes in 1 blocks are definitely lost in loss record 3 of 3 ==28425== at 0x4C28C50: malloc (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so) ==28425== by 0x4C2AB1E: realloc (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so) ==28425== by 0x58BBB27: maybe_resize (core.c:748) ==28425== by 0x58BBB27: uv__io_start (core.c:787) ==28425== by 0x58C1B80: uv__signal_loop_once_init (signal.c:225) ==28425== by 0x58C1B80: uv_signal_init (signal.c:260) ==28425== by 0x58BF7A6: uv_loop_init (loop.c:66) ==28425== by 0x4157F5: lws_uv_initloop (libuv.c:89) ==28425== by 0x405536: main (test-server-libuv.c:284) libuv wants to sign off on all libuv 'handles' that will close, and callback to do the close confirmation asynchronously. The wsi close function is adapted when libuv is in use to work with libuv accordingly and exit the uv loop the number of remaining wsi is zero. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-14 09:27:41 +08:00
#endif
lws_close_free_wsi_final(wsi);
}
void
lws_close_free_wsi_final(struct lws *wsi)
{
int n;
http2: make usable HTTP/2 support is now able to serve the test server, complete with websockets, from a single vhost. - This works the same with both OpenSSL and mbedTLS. - POST is now wired up and works (also for file upload). - CGI is wired up and works. - Redirect is adapted and works - lwsws works. - URI urldecode, sanitation and argument parsing wired up for :path valgrind clean (aside from openssl-style false uninit data usage in mbedtls send occasionally) h2spec reports: $ h2spec -h 127.0.0.1 -p 7681 -t -k -o 1 ... 145 tests, 145 passed, 0 skipped, 0 failed" Incorporates: - "https://github.com/warmcat/libwebsockets/pull/1039 Fixes issue with -Werror=unused-variable flag - https://github.com/sartura/libwebsockets/commit/2c843a1395bdca5c23b40d989f508fc96aa5ba3b ssl: fix infinite loop on client cert verification failure Signed-off-by: Petar Paradzik <petar.paradzik@sartura.hr>" Caused and fixes Coverity 184887 - 184892
2017-10-13 10:33:02 +08:00
if (lws_socket_is_valid(wsi->desc.sockfd) && !lws_ssl_close(wsi)) {
mbed3 plat Signed-off-by: Andy Green <andy.green@linaro.org>
2015-11-02 20:34:12 +08:00
#if LWS_POSIX
adoption: make union for socket and file fds This lets lws support adopting raw file FDs and raw socket fds. A test plugin creates a FIFO and prints data sent on it, using the lws event loop.
2017-02-27 12:55:56 +08:00
n = compatible_close(wsi->desc.sockfd);
refactor move ssl server service to ssl.c Signed-off-by: Andy Green <andy.green@linaro.org>
2014-04-12 10:07:02 +08:00
if (n)
lwsl_debug("closing: close ret %d\n", LWS_ERRNO);
mbed3 plat Signed-off-by: Andy Green <andy.green@linaro.org>
2015-11-02 20:34:12 +08:00
mbed3 workable plus or minus mbed3 net stack bug https://github.com/ARMmbed/sockets/issues/35 Signed-off-by: Andy Green <andy.green@linaro.org>
2015-11-08 12:10:26 +08:00
#else
adoption: make union for socket and file fds This lets lws support adopting raw file FDs and raw socket fds. A test plugin creates a FIFO and prints data sent on it, using the lws event loop.
2017-02-27 12:55:56 +08:00
compatible_close(wsi->desc.sockfd);
esp8266 initial support
2016-07-23 14:18:25 +08:00
(void)n;
mbed3 plat Signed-off-by: Andy Green <andy.green@linaro.org>
2015-11-02 20:34:12 +08:00
#endif
adoption: make union for socket and file fds This lets lws support adopting raw file FDs and raw socket fds. A test plugin creates a FIFO and prints data sent on it, using the lws event loop.
2017-02-27 12:55:56 +08:00
wsi->desc.sockfd = LWS_SOCK_INVALID;
add-wss-ssl-openssl-support.patch Signed-off-by: Andy Green <andy@warmcat.com>
2010-11-08 17:03:03 +00:00
}
add outermost wsi lifetime callbacks Signed-off-by: Andy Green <andy.green@linaro.org>
2014-02-15 19:25:50 +08:00
/* outermost destroy notification for wsi (user_space still intact) */
introduce vhosts This patch splits out some lws_context members into a new lws_vhost struct. - ssl state and options per vhost - SSL_CTX for serving and client per vhost - protocols[] per vhost - extensions[] per vhost lws_context maintains a linked list of lws_vhosts. The same lws_context_creation_info struct is used to regulate both the context creation and to create vhosts: for backward compatibility if you didn't provide the new LWS_SERVER_OPTION_EXPLICIT_VHOSTS option, then a default vhost is created at context creation time using the same info data as the context itself. If you will have multiple vhosts though, you should give the LWS_SERVER_OPTION_EXPLICIT_VHOSTS option at context creation time, create the context first and then the vhosts afterwards using lws_create_vhost(contest, &info); Although there is a lot of housekeeping to implement this change, there is almost no additional overhead if you don't use multiple vhosts and very little api impact (no changes to test apps). Signed-off-by: Andy Green <andy@warmcat.com>
2016-03-28 10:10:43 +08:00
wsi->vhost->protocols[0].callback(wsi, LWS_CALLBACK_WSI_DESTROY,
clean
2017-09-23 12:55:21 +08:00
wsi->user_space, NULL, 0);
add outermost wsi lifetime callbacks Signed-off-by: Andy Green <andy.green@linaro.org>
2014-02-15 19:25:50 +08:00
cgi post Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-13 11:42:53 +08:00
#ifdef LWS_WITH_CGI
if (wsi->cgi) {
cgi: fix not getting POLLHUP on fd associated to CGI STDOUT Signed-off-by: Petar Paradzik <petar.paradzik@sartura.hr>
2017-07-28 18:01:14 +02:00
for (n = 0; n < 3; n++) {
if (wsi->cgi->pipe_fds[n][!!(n == 0)] == 0)
context deprecation 1) This makes lwsws run a parent process with the original permissions. But this process is only able to respond to SIGHUP, it doesn't do anything else. 2) You can send this parent process a SIGHUP now to cause it to - close listening sockets in existing lwsws processes - mark those processes as to exit when the number of active connections on the falls to zero - spawn a fresh child process from scratch, using latest configuration file content, latest plugins, etc. It can now reopen listening sockets if it chooses to, or open different listen ports or whatever. Notes: 1) lws_context_destroy() has been split into two pieces... the reason for the split is the first part closes the per-vhost protocols, but since they may have created libuv objects in the per-vhost protocol storage, these cannot be freed until after the loop has been run. That's the purpose of the second part of the context destruction, lws_context_destroy2(). For compatibility, if you are not using libuv, the first part calls the second part. However if you are using libuv, you must now call the second part from your own main.c after the first part.
2016-12-16 07:37:43 +08:00
lwsl_err("ZERO FD IN CGI CLOSE");
cgi: fix not getting POLLHUP on fd associated to CGI STDOUT Signed-off-by: Petar Paradzik <petar.paradzik@sartura.hr>
2017-07-28 18:01:14 +02:00
if (wsi->cgi->pipe_fds[n][!!(n == 0)] >= 0)
close(wsi->cgi->pipe_fds[n][!!(n == 0)]);
context deprecation 1) This makes lwsws run a parent process with the original permissions. But this process is only able to respond to SIGHUP, it doesn't do anything else. 2) You can send this parent process a SIGHUP now to cause it to - close listening sockets in existing lwsws processes - mark those processes as to exit when the number of active connections on the falls to zero - spawn a fresh child process from scratch, using latest configuration file content, latest plugins, etc. It can now reopen listening sockets if it chooses to, or open different listen ports or whatever. Notes: 1) lws_context_destroy() has been split into two pieces... the reason for the split is the first part closes the per-vhost protocols, but since they may have created libuv objects in the per-vhost protocol storage, these cannot be freed until after the loop has been run. That's the purpose of the second part of the context destruction, lws_context_destroy2(). For compatibility, if you are not using libuv, the first part calls the second part. However if you are using libuv, you must now call the second part from your own main.c after the first part.
2016-12-16 07:37:43 +08:00
}
cgi lifecycle logging Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-20 06:10:56 +08:00
cgi post Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-13 11:42:53 +08:00
lws_free(wsi->cgi);
}
#endif
Fix potential memory leak - Got rid of ifdef _WIN32 stuff adn moved to plat_ files instead. - Also, check all calls to lws_zalloc, was potential failure on WIN32 - Made context destory enable to destroy a half inited context as well. This way I got get rid of some of the error handling complexity in libwebsocket_create_context - Added TODOs for some potential problems I see where things might be leaking and such
2015-06-25 17:51:07 +02:00
lws_free_wsi(wsi);
add-per-session-user-data.patch Signed-off-by: Andy Green <andy@warmcat.com>
2010-11-03 11:13:06 +00:00
}
lws_get_urlarg_by_name Adds a convenient way to directly get the value of a URL argument like ...?x=y&v=1, regardless of position in the parameter list. Signed-off-by: Andy Green <andy@warmcat.com>
2016-06-03 09:04:15 +08:00
LWS_VISIBLE LWS_EXTERN const char *
lws_get_urlarg_by_name(struct lws *wsi, const char *name, char *buf, int len)
{
int n = 0, sl = strlen(name);
while (lws_hdr_copy_fragment(wsi, buf, len,
WSI_TOKEN_HTTP_URI_ARGS, n) >= 0) {
if (!strncmp(buf, name, sl))
return buf + sl;
n++;
}
return NULL;
}
ESP32 platform This is enough for all the test app features to work on ESP32 without SSL.
2017-02-18 17:26:40 +08:00
#if LWS_POSIX && !defined(LWS_WITH_ESP32)
lws_interface_to_sa Expose interface handling code as an api
2016-01-16 12:09:38 +08:00
LWS_VISIBLE int
clean
2017-09-23 12:55:21 +08:00
interface_to_sa(struct lws_vhost *vh, const char *ifname,
struct sockaddr_in *addr, size_t addrlen)
lws_interface_to_sa Expose interface handling code as an api
2016-01-16 12:09:38 +08:00
{
int ipv6 = 0;
CMake: convert all LWS_USE_... to LWS_WITH_... Almost all the CMake options begin with LWS_WITH_..., but many of the symbols passed to lws are LWS_USE_... , this causes neededless confusion, compounded by the fact that a few CMake options also begin with LWS_USE_. This patch globally converts all LWS_USE_... to LWS_WITH_..., so there is only one prefix to remember in both CMake and the code. The affected public CMake options are LWS_USE_BORINGSSL -> LWS_WITH_BORINGSSL LWS_USE_CYASSL -> LWS_WITH_CYASSL LWS_USE_WOLFSSL -> LWS_WITH_WOLFSSL LWS_USE_MBEDTLS -> LWS_WITH_MBEDTLS LWS_USE_BUNDLED_ZLIB -> LWS_WITH_BUNDLED_ZLIB
2017-09-28 11:29:03 +08:00
#ifdef LWS_WITH_IPV6
ipv6 move disable to vhost option Server ipv6 support disable is now controlled by vhost->options rather than context->options, allowing it to be set per-vhost. Signed-off-by: Andy Green <andy@warmcat.com>
2016-06-03 21:19:40 +08:00
ipv6 = LWS_IPV6_ENABLED(vh);
lws_interface_to_sa Expose interface handling code as an api
2016-01-16 12:09:38 +08:00
#endif
ipv6 move disable to vhost option Server ipv6 support disable is now controlled by vhost->options rather than context->options, allowing it to be set per-vhost. Signed-off-by: Andy Green <andy@warmcat.com>
2016-06-03 21:19:40 +08:00
(void)vh;
lws_interface_to_sa Expose interface handling code as an api
2016-01-16 12:09:38 +08:00
return lws_interface_to_sa(ipv6, ifname, addr, addrlen);
}
#endif
plat-optee and boringssl adaptations
2017-01-17 07:01:02 +08:00
#ifndef LWS_PLAT_OPTEE
esp8266 initial support
2016-07-23 14:18:25 +08:00
#if LWS_POSIX
ipv6 move disable to vhost option Server ipv6 support disable is now controlled by vhost->options rather than context->options, allowing it to be set per-vhost. Signed-off-by: Andy Green <andy@warmcat.com>
2016-06-03 21:19:40 +08:00
static int
lws_get_addresses(struct lws_vhost *vh, void *ads, char *name,
clean tidy the worst whitespace alignment probs due to mass token name length changes Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-04 11:30:53 +08:00
int name_len, char *rip, int rip_len)
gethostbyname removal from lws get peer addresses This replaces gethostbyname in libwebsockets_get_peer_addresses and in the client handshake path. There's one left in lws-plat-win but that can be done another time. Let's see how much damage that did to the cross-platform and option builds... Signed-off-by: Andy Green <andy.green@linaro.org>
2015-01-28 21:03:49 +08:00
{
mbed3 plat Signed-off-by: Andy Green <andy.green@linaro.org>
2015-11-02 20:34:12 +08:00
#if LWS_POSIX
gethostbyname removal from lws get peer addresses This replaces gethostbyname in libwebsockets_get_peer_addresses and in the client handshake path. There's one left in lws-plat-win but that can be done another time. Let's see how much damage that did to the cross-platform and option builds... Signed-off-by: Andy Green <andy.green@linaro.org>
2015-01-28 21:03:49 +08:00
struct addrinfo ai, *res;
Fix getaddrinfo memory leak
2015-10-15 13:02:03 +03:00
struct sockaddr_in addr4;
gethostbyname removal from lws get peer addresses This replaces gethostbyname in libwebsockets_get_peer_addresses and in the client handshake path. There's one left in lws-plat-win but that can be done another time. Let's see how much damage that did to the cross-platform and option builds... Signed-off-by: Andy Green <andy.green@linaro.org>
2015-01-28 21:03:49 +08:00
coverity 182068: 155650: unnecessary check against illegal NULL
2017-07-28 14:19:24 +08:00
rip[0] = '\0';
gethostbyname removal from lws get peer addresses This replaces gethostbyname in libwebsockets_get_peer_addresses and in the client handshake path. There's one left in lws-plat-win but that can be done another time. Let's see how much damage that did to the cross-platform and option builds... Signed-off-by: Andy Green <andy.green@linaro.org>
2015-01-28 21:03:49 +08:00
name[0] = '\0';
Fix getaddrinfo memory leak
2015-10-15 13:02:03 +03:00
addr4.sin_family = AF_UNSPEC;
gethostbyname removal from lws get peer addresses This replaces gethostbyname in libwebsockets_get_peer_addresses and in the client handshake path. There's one left in lws-plat-win but that can be done another time. Let's see how much damage that did to the cross-platform and option builds... Signed-off-by: Andy Green <andy.green@linaro.org>
2015-01-28 21:03:49 +08:00
CMake: convert all LWS_USE_... to LWS_WITH_... Almost all the CMake options begin with LWS_WITH_..., but many of the symbols passed to lws are LWS_USE_... , this causes neededless confusion, compounded by the fact that a few CMake options also begin with LWS_USE_. This patch globally converts all LWS_USE_... to LWS_WITH_..., so there is only one prefix to remember in both CMake and the code. The affected public CMake options are LWS_USE_BORINGSSL -> LWS_WITH_BORINGSSL LWS_USE_CYASSL -> LWS_WITH_CYASSL LWS_USE_WOLFSSL -> LWS_WITH_WOLFSSL LWS_USE_MBEDTLS -> LWS_WITH_MBEDTLS LWS_USE_BUNDLED_ZLIB -> LWS_WITH_BUNDLED_ZLIB
2017-09-28 11:29:03 +08:00
#ifdef LWS_WITH_IPV6
ipv6 move disable to vhost option Server ipv6 support disable is now controlled by vhost->options rather than context->options, allowing it to be set per-vhost. Signed-off-by: Andy Green <andy@warmcat.com>
2016-06-03 21:19:40 +08:00
if (LWS_IPV6_ENABLED(vh)) {
clean
2017-09-23 12:55:21 +08:00
if (!lws_plat_inet_ntop(AF_INET6,
&((struct sockaddr_in6 *)ads)->sin6_addr,
rip, rip_len)) {
gcc-format-strings: ipv6
2017-02-14 17:55:13 +08:00
lwsl_err("inet_ntop: %s", strerror(LWS_ERRNO));
gethostbyname removal from lws get peer addresses This replaces gethostbyname in libwebsockets_get_peer_addresses and in the client handshake path. There's one left in lws-plat-win but that can be done another time. Let's see how much damage that did to the cross-platform and option builds... Signed-off-by: Andy Green <andy.green@linaro.org>
2015-01-28 21:03:49 +08:00
return -1;
}
// Strip off the IPv4 to IPv6 header if one exists
if (strncmp(rip, "::ffff:", 7) == 0)
memmove(rip, rip + 7, strlen(rip) - 6);
getnameinfo((struct sockaddr *)ads,
clean
2017-09-23 12:55:21 +08:00
sizeof(struct sockaddr_in6), name, name_len, NULL, 0, 0);
whitespace trailing mass cleanout Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-14 08:52:03 +08:00
gethostbyname removal from lws get peer addresses This replaces gethostbyname in libwebsockets_get_peer_addresses and in the client handshake path. There's one left in lws-plat-win but that can be done another time. Let's see how much damage that did to the cross-platform and option builds... Signed-off-by: Andy Green <andy.green@linaro.org>
2015-01-28 21:03:49 +08:00
return 0;
} else
#endif
{
Fix getaddrinfo memory leak
2015-10-15 13:02:03 +03:00
struct addrinfo *result;
gethostbyname removal from lws get peer addresses This replaces gethostbyname in libwebsockets_get_peer_addresses and in the client handshake path. There's one left in lws-plat-win but that can be done another time. Let's see how much damage that did to the cross-platform and option builds... Signed-off-by: Andy Green <andy.green@linaro.org>
2015-01-28 21:03:49 +08:00
memset(&ai, 0, sizeof ai);
ai.ai_family = PF_UNSPEC;
ai.ai_socktype = SOCK_STREAM;
ai.ai_flags = AI_CANONNAME;
ESP32 platform This is enough for all the test app features to work on ESP32 without SSL.
2017-02-18 17:26:40 +08:00
#if !defined(LWS_WITH_ESP32)
gethostbyname removal from lws get peer addresses This replaces gethostbyname in libwebsockets_get_peer_addresses and in the client handshake path. There's one left in lws-plat-win but that can be done another time. Let's see how much damage that did to the cross-platform and option builds... Signed-off-by: Andy Green <andy.green@linaro.org>
2015-01-28 21:03:49 +08:00
if (getnameinfo((struct sockaddr *)ads,
sizeof(struct sockaddr_in),
name, name_len, NULL, 0, 0))
return -1;
ESP32 platform This is enough for all the test app features to work on ESP32 without SSL.
2017-02-18 17:26:40 +08:00
#endif
gethostbyname removal from lws get peer addresses This replaces gethostbyname in libwebsockets_get_peer_addresses and in the client handshake path. There's one left in lws-plat-win but that can be done another time. Let's see how much damage that did to the cross-platform and option builds... Signed-off-by: Andy Green <andy.green@linaro.org>
2015-01-28 21:03:49 +08:00
Fix getaddrinfo memory leak
2015-10-15 13:02:03 +03:00
if (getaddrinfo(name, NULL, &ai, &result))
gethostbyname removal from lws get peer addresses This replaces gethostbyname in libwebsockets_get_peer_addresses and in the client handshake path. There's one left in lws-plat-win but that can be done another time. Let's see how much damage that did to the cross-platform and option builds... Signed-off-by: Andy Green <andy.green@linaro.org>
2015-01-28 21:03:49 +08:00
return -1;
Fix getaddrinfo memory leak
2015-10-15 13:02:03 +03:00
res = result;
while (addr4.sin_family == AF_UNSPEC && res) {
gethostbyname removal from lws get peer addresses This replaces gethostbyname in libwebsockets_get_peer_addresses and in the client handshake path. There's one left in lws-plat-win but that can be done another time. Let's see how much damage that did to the cross-platform and option builds... Signed-off-by: Andy Green <andy.green@linaro.org>
2015-01-28 21:03:49 +08:00
switch (res->ai_family) {
case AF_INET:
clean
2017-09-23 12:55:21 +08:00
addr4.sin_addr =
((struct sockaddr_in *)res->ai_addr)->sin_addr;
Fix getaddrinfo memory leak
2015-10-15 13:02:03 +03:00
addr4.sin_family = AF_INET;
gethostbyname removal from lws get peer addresses This replaces gethostbyname in libwebsockets_get_peer_addresses and in the client handshake path. There's one left in lws-plat-win but that can be done another time. Let's see how much damage that did to the cross-platform and option builds... Signed-off-by: Andy Green <andy.green@linaro.org>
2015-01-28 21:03:49 +08:00
break;
}
res = res->ai_next;
}
Fix getaddrinfo memory leak
2015-10-15 13:02:03 +03:00
freeaddrinfo(result);
gethostbyname removal from lws get peer addresses This replaces gethostbyname in libwebsockets_get_peer_addresses and in the client handshake path. There's one left in lws-plat-win but that can be done another time. Let's see how much damage that did to the cross-platform and option builds... Signed-off-by: Andy Green <andy.green@linaro.org>
2015-01-28 21:03:49 +08:00
}
Fix getaddrinfo memory leak
2015-10-15 13:02:03 +03:00
if (addr4.sin_family == AF_UNSPEC)
gethostbyname removal from lws get peer addresses This replaces gethostbyname in libwebsockets_get_peer_addresses and in the client handshake path. There's one left in lws-plat-win but that can be done another time. Let's see how much damage that did to the cross-platform and option builds... Signed-off-by: Andy Green <andy.green@linaro.org>
2015-01-28 21:03:49 +08:00
return -1;
introduce LWS_WARN_UNUSED_RESULT Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-20 16:56:06 +08:00
if (lws_plat_inet_ntop(AF_INET, &addr4.sin_addr, rip, rip_len) == NULL)
return -1;
gethostbyname removal from lws get peer addresses This replaces gethostbyname in libwebsockets_get_peer_addresses and in the client handshake path. There's one left in lws-plat-win but that can be done another time. Let's see how much damage that did to the cross-platform and option builds... Signed-off-by: Andy Green <andy.green@linaro.org>
2015-01-28 21:03:49 +08:00
return 0;
mbed3 plat Signed-off-by: Andy Green <andy.green@linaro.org>
2015-11-02 20:34:12 +08:00
#else
ipv6 move disable to vhost option Server ipv6 support disable is now controlled by vhost->options rather than context->options, allowing it to be set per-vhost. Signed-off-by: Andy Green <andy@warmcat.com>
2016-06-03 21:19:40 +08:00
(void)vh;
mbed3 plat Signed-off-by: Andy Green <andy.green@linaro.org>
2015-11-02 20:34:12 +08:00
(void)ads;
(void)name;
(void)name_len;
(void)rip;
(void)rip_len;
return -1;
#endif
gethostbyname removal from lws get peer addresses This replaces gethostbyname in libwebsockets_get_peer_addresses and in the client handshake path. There's one left in lws-plat-win but that can be done another time. Let's see how much damage that did to the cross-platform and option builds... Signed-off-by: Andy Green <andy.green@linaro.org>
2015-01-28 21:03:49 +08:00
}
esp8266 initial support
2016-07-23 14:18:25 +08:00
#endif
gethostbyname removal from lws get peer addresses This replaces gethostbyname in libwebsockets_get_peer_addresses and in the client handshake path. There's one left in lws-plat-win but that can be done another time. Let's see how much damage that did to the cross-platform and option builds... Signed-off-by: Andy Green <andy.green@linaro.org>
2015-01-28 21:03:49 +08:00
plat-optee and boringssl adaptations
2017-01-17 07:01:02 +08:00
lws_get_peer_simple move to user api https://github.com/warmcat/libwebsockets/issues/537 Signed-off-by: Andy Green <andy@warmcat.com>
2016-05-25 08:04:52 +08:00
LWS_VISIBLE const char *
lws access log option and lwsws conf This adds the ability to store apache-compatible logs to a file given at vhost-creation time. lwsws conf can set it per-vhost using "access-log": "<filepath>" The feature defaults to disabled at cmake, it can be set independently but LWS_WITH_LWSWS set it on. Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-15 12:00:23 +08:00
lws_get_peer_simple(struct lws *wsi, char *name, int namelen)
{
#if LWS_POSIX
socklen_t len, olen;
CMake: convert all LWS_USE_... to LWS_WITH_... Almost all the CMake options begin with LWS_WITH_..., but many of the symbols passed to lws are LWS_USE_... , this causes neededless confusion, compounded by the fact that a few CMake options also begin with LWS_USE_. This patch globally converts all LWS_USE_... to LWS_WITH_..., so there is only one prefix to remember in both CMake and the code. The affected public CMake options are LWS_USE_BORINGSSL -> LWS_WITH_BORINGSSL LWS_USE_CYASSL -> LWS_WITH_CYASSL LWS_USE_WOLFSSL -> LWS_WITH_WOLFSSL LWS_USE_MBEDTLS -> LWS_WITH_MBEDTLS LWS_USE_BUNDLED_ZLIB -> LWS_WITH_BUNDLED_ZLIB
2017-09-28 11:29:03 +08:00
#ifdef LWS_WITH_IPV6
lws access log option and lwsws conf This adds the ability to store apache-compatible logs to a file given at vhost-creation time. lwsws conf can set it per-vhost using "access-log": "<filepath>" The feature defaults to disabled at cmake, it can be set independently but LWS_WITH_LWSWS set it on. Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-15 12:00:23 +08:00
struct sockaddr_in6 sin6;
#endif
struct sockaddr_in sin4;
int af = AF_INET;
void *p, *q;
http2: make usable HTTP/2 support is now able to serve the test server, complete with websockets, from a single vhost. - This works the same with both OpenSSL and mbedTLS. - POST is now wired up and works (also for file upload). - CGI is wired up and works. - Redirect is adapted and works - lwsws works. - URI urldecode, sanitation and argument parsing wired up for :path valgrind clean (aside from openssl-style false uninit data usage in mbedtls send occasionally) h2spec reports: $ h2spec -h 127.0.0.1 -p 7681 -t -k -o 1 ... 145 tests, 145 passed, 0 skipped, 0 failed" Incorporates: - "https://github.com/warmcat/libwebsockets/pull/1039 Fixes issue with -Werror=unused-variable flag - https://github.com/sartura/libwebsockets/commit/2c843a1395bdca5c23b40d989f508fc96aa5ba3b ssl: fix infinite loop on client cert verification failure Signed-off-by: Petar Paradzik <petar.paradzik@sartura.hr>" Caused and fixes Coverity 184887 - 184892
2017-10-13 10:33:02 +08:00
#if defined(LWS_WITH_HTTP2)
if (wsi->http2_substream)
wsi = wsi->u.h2.parent_wsi;
#endif
lws-meta
2017-07-19 04:39:14 +08:00
if (wsi->parent_carries_io)
wsi = wsi->parent;
CMake: convert all LWS_USE_... to LWS_WITH_... Almost all the CMake options begin with LWS_WITH_..., but many of the symbols passed to lws are LWS_USE_... , this causes neededless confusion, compounded by the fact that a few CMake options also begin with LWS_USE_. This patch globally converts all LWS_USE_... to LWS_WITH_..., so there is only one prefix to remember in both CMake and the code. The affected public CMake options are LWS_USE_BORINGSSL -> LWS_WITH_BORINGSSL LWS_USE_CYASSL -> LWS_WITH_CYASSL LWS_USE_WOLFSSL -> LWS_WITH_WOLFSSL LWS_USE_MBEDTLS -> LWS_WITH_MBEDTLS LWS_USE_BUNDLED_ZLIB -> LWS_WITH_BUNDLED_ZLIB
2017-09-28 11:29:03 +08:00
#ifdef LWS_WITH_IPV6
ipv6 move disable to vhost option Server ipv6 support disable is now controlled by vhost->options rather than context->options, allowing it to be set per-vhost. Signed-off-by: Andy Green <andy@warmcat.com>
2016-06-03 21:19:40 +08:00
if (LWS_IPV6_ENABLED(wsi->vhost)) {
lws access log option and lwsws conf This adds the ability to store apache-compatible logs to a file given at vhost-creation time. lwsws conf can set it per-vhost using "access-log": "<filepath>" The feature defaults to disabled at cmake, it can be set independently but LWS_WITH_LWSWS set it on. Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-15 12:00:23 +08:00
len = sizeof(sin6);
p = &sin6;
af = AF_INET6;
q = &sin6.sin6_addr;
} else
#endif
{
len = sizeof(sin4);
p = &sin4;
q = &sin4.sin_addr;
}
olen = len;
adoption: make union for socket and file fds This lets lws support adopting raw file FDs and raw socket fds. A test plugin creates a FIFO and prints data sent on it, using the lws event loop.
2017-02-27 12:55:56 +08:00
if (getpeername(wsi->desc.sockfd, p, &len) < 0 || len > olen) {
lws access log option and lwsws conf This adds the ability to store apache-compatible logs to a file given at vhost-creation time. lwsws conf can set it per-vhost using "access-log": "<filepath>" The feature defaults to disabled at cmake, it can be set independently but LWS_WITH_LWSWS set it on. Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-15 12:00:23 +08:00
lwsl_warn("getpeername: %s\n", strerror(LWS_ERRNO));
return NULL;
}
Fix inet_ntop linking error on Windows
2016-04-17 19:26:04 -03:00
return lws_plat_inet_ntop(af, q, name, namelen);
esp8266 initial support
2016-07-23 14:18:25 +08:00
#else
#if defined(LWS_WITH_ESP8266)
return lws_plat_get_peer_simple(wsi, name, namelen);
lws access log option and lwsws conf This adds the ability to store apache-compatible logs to a file given at vhost-creation time. lwsws conf can set it per-vhost using "access-log": "<filepath>" The feature defaults to disabled at cmake, it can be set independently but LWS_WITH_LWSWS set it on. Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-15 12:00:23 +08:00
#else
return NULL;
#endif
esp8266 initial support
2016-07-23 14:18:25 +08:00
#endif
lws access log option and lwsws conf This adds the ability to store apache-compatible logs to a file given at vhost-creation time. lwsws conf can set it per-vhost using "access-log": "<filepath>" The feature defaults to disabled at cmake, it can be set independently but LWS_WITH_LWSWS set it on. Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-15 12:00:23 +08:00
}
plat-optee and boringssl adaptations
2017-01-17 07:01:02 +08:00
#endif
lws access log option and lwsws conf This adds the ability to store apache-compatible logs to a file given at vhost-creation time. lwsws conf can set it per-vhost using "access-log": "<filepath>" The feature defaults to disabled at cmake, it can be set independently but LWS_WITH_LWSWS set it on. Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-15 12:00:23 +08:00
Hide private library symbols Signed-off-by: Peter Pentchev <roam@ringlet.net>
2013-03-30 09:52:21 +08:00
LWS_VISIBLE void
public api remove superfluous context params API BREAK Extend the cleanout caused by wsi having a context pointer into the public api. There's no point keeping the 1.5 compatibility work, we have changed the api in several places and rebuilt wasn't going to be enough a while ago. Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-16 18:19:08 +08:00
lws_get_peer_addresses(struct lws *wsi, lws_sockfd_type fd, char *name,
int name_len, char *rip, int rip_len)
introduce-network-connect-filter-callback.patch Signed-off-by: Andy Green <andy.green@linaro.org>
2011-02-13 08:37:12 +00:00
{
plat-optee and boringssl adaptations
2017-01-17 07:01:02 +08:00
#ifndef LWS_PLAT_OPTEE
mbed3 plat Signed-off-by: Andy Green <andy.green@linaro.org>
2015-11-02 20:34:12 +08:00
#if LWS_POSIX
cygwin compatibility Signed-off-by: Bob Roberts <bobroberts177@gmail.com>
2013-04-25 09:16:30 +08:00
socklen_t len;
CMake: convert all LWS_USE_... to LWS_WITH_... Almost all the CMake options begin with LWS_WITH_..., but many of the symbols passed to lws are LWS_USE_... , this causes neededless confusion, compounded by the fact that a few CMake options also begin with LWS_USE_. This patch globally converts all LWS_USE_... to LWS_WITH_..., so there is only one prefix to remember in both CMake and the code. The affected public CMake options are LWS_USE_BORINGSSL -> LWS_WITH_BORINGSSL LWS_USE_CYASSL -> LWS_WITH_CYASSL LWS_USE_WOLFSSL -> LWS_WITH_WOLFSSL LWS_USE_MBEDTLS -> LWS_WITH_MBEDTLS LWS_USE_BUNDLED_ZLIB -> LWS_WITH_BUNDLED_ZLIB
2017-09-28 11:29:03 +08:00
#ifdef LWS_WITH_IPV6
add ipv6 support (changed to support runtime disable + integration by andy@warmcat.com)
2014-03-24 16:09:25 +08:00
struct sockaddr_in6 sin6;
#endif
struct sockaddr_in sin4;
public api remove superfluous context params API BREAK Extend the cleanout caused by wsi having a context pointer into the public api. There's no point keeping the 1.5 compatibility work, we have changed the api in several places and rebuilt wasn't going to be enough a while ago. Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-16 18:19:08 +08:00
struct lws_context *context = wsi->context;
add ipv6 support (changed to support runtime disable + integration by andy@warmcat.com)
2014-03-24 16:09:25 +08:00
int ret = -1;
gethostbyname removal from lws get peer addresses This replaces gethostbyname in libwebsockets_get_peer_addresses and in the client handshake path. There's one left in lws-plat-win but that can be done another time. Let's see how much damage that did to the cross-platform and option builds... Signed-off-by: Andy Green <andy.green@linaro.org>
2015-01-28 21:03:49 +08:00
void *p;
introduce-network-connect-filter-callback.patch Signed-off-by: Andy Green <andy.green@linaro.org>
2011-02-13 08:37:12 +00:00
rip[0] = '\0';
name[0] = '\0';
change get_peer_addresses to use context wsi latency Signed-off-by: Andy Green <andy.green@linaro.org>
2013-01-30 08:12:20 +08:00
lws_latency_pre(context, wsi);
CMake: convert all LWS_USE_... to LWS_WITH_... Almost all the CMake options begin with LWS_WITH_..., but many of the symbols passed to lws are LWS_USE_... , this causes neededless confusion, compounded by the fact that a few CMake options also begin with LWS_USE_. This patch globally converts all LWS_USE_... to LWS_WITH_..., so there is only one prefix to remember in both CMake and the code. The affected public CMake options are LWS_USE_BORINGSSL -> LWS_WITH_BORINGSSL LWS_USE_CYASSL -> LWS_WITH_CYASSL LWS_USE_WOLFSSL -> LWS_WITH_WOLFSSL LWS_USE_MBEDTLS -> LWS_WITH_MBEDTLS LWS_USE_BUNDLED_ZLIB -> LWS_WITH_BUNDLED_ZLIB
2017-09-28 11:29:03 +08:00
#ifdef LWS_WITH_IPV6
ipv6 move disable to vhost option Server ipv6 support disable is now controlled by vhost->options rather than context->options, allowing it to be set per-vhost. Signed-off-by: Andy Green <andy@warmcat.com>
2016-06-03 21:19:40 +08:00
if (LWS_IPV6_ENABLED(wsi->vhost)) {
add ipv6 support (changed to support runtime disable + integration by andy@warmcat.com)
2014-03-24 16:09:25 +08:00
len = sizeof(sin6);
gethostbyname removal from lws get peer addresses This replaces gethostbyname in libwebsockets_get_peer_addresses and in the client handshake path. There's one left in lws-plat-win but that can be done another time. Let's see how much damage that did to the cross-platform and option builds... Signed-off-by: Andy Green <andy.green@linaro.org>
2015-01-28 21:03:49 +08:00
p = &sin6;
add ipv6 support (changed to support runtime disable + integration by andy@warmcat.com)
2014-03-24 16:09:25 +08:00
} else
Re: Windows port of your libwebsocket The latest code changed need some win32 fixes: Signed-off-by: Peter Hinz <cerebusrc@gmail.com>
2011-03-10 18:14:01 +00:00
#endif
add ipv6 support (changed to support runtime disable + integration by andy@warmcat.com)
2014-03-24 16:09:25 +08:00
{
len = sizeof(sin4);
gethostbyname removal from lws get peer addresses This replaces gethostbyname in libwebsockets_get_peer_addresses and in the client handshake path. There's one left in lws-plat-win but that can be done another time. Let's see how much damage that did to the cross-platform and option builds... Signed-off-by: Andy Green <andy.green@linaro.org>
2015-01-28 21:03:49 +08:00
p = &sin4;
}
introduce-network-connect-filter-callback.patch Signed-off-by: Andy Green <andy.green@linaro.org>
2011-02-13 08:37:12 +00:00
gethostbyname removal from lws get peer addresses This replaces gethostbyname in libwebsockets_get_peer_addresses and in the client handshake path. There's one left in lws-plat-win but that can be done another time. Let's see how much damage that did to the cross-platform and option builds... Signed-off-by: Andy Green <andy.green@linaro.org>
2015-01-28 21:03:49 +08:00
if (getpeername(fd, p, &len) < 0) {
lwsl_warn("getpeername: %s\n", strerror(LWS_ERRNO));
goto bail;
introduce-network-connect-filter-callback.patch Signed-off-by: Andy Green <andy.green@linaro.org>
2011-02-13 08:37:12 +00:00
}
whitespace trailing mass cleanout Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-14 08:52:03 +08:00
ipv6 move disable to vhost option Server ipv6 support disable is now controlled by vhost->options rather than context->options, allowing it to be set per-vhost. Signed-off-by: Andy Green <andy@warmcat.com>
2016-06-03 21:19:40 +08:00
ret = lws_get_addresses(wsi->vhost, p, name, name_len, rip, rip_len);
change get_peer_addresses to use context wsi latency Signed-off-by: Andy Green <andy.green@linaro.org>
2013-01-30 08:12:20 +08:00
bail:
api rationalization use new names internally Change all internal uses of rationalized public apis to reflect the new names. Theer are a few things that got changed as side effect of search/replace matches, but these are almost all internal. I added a compatibility define for the public enum that got renamed. Theoretically existing code should not notice the difference from these two patches. And new code will find the new names. https://github.com/warmcat/libwebsockets/issues/357 Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-04 08:43:54 +08:00
lws_latency(context, wsi, "lws_get_peer_addresses", ret, 1);
plat-optee and boringssl adaptations
2017-01-17 07:01:02 +08:00
#endif
#endif
mbed3 plat Signed-off-by: Andy Green <andy.green@linaro.org>
2015-11-02 20:34:12 +08:00
(void)wsi;
(void)fd;
(void)name;
(void)name_len;
(void)rip;
(void)rip_len;
plat-optee and boringssl adaptations
2017-01-17 07:01:02 +08:00
introduce-network-connect-filter-callback.patch Signed-off-by: Andy Green <andy.green@linaro.org>
2011-02-13 08:37:12 +00:00
}
export-libwebsocket_service_fd-api.patch Signed-off-by: Andy Green <andy.green@linaro.org>
2011-02-12 11:57:45 +00:00
vhost: allow user pointer to be attached at creation time https://github.com/warmcat/libwebsockets/issues/1008
2017-08-29 15:37:16 +08:00
LWS_EXTERN void *
lws_vhost_user(struct lws_vhost *vhost)
{
return vhost->user;
}
libwebsocket_context: add userspace pointer for use before wsi creation Signed-off-by: Alon Levy <alevy@redhat.com>
2012-10-19 11:21:56 +02:00
LWS_EXTERN void *
api rationalization: eliminate all libwebsocket[s]_ prefixes This nukes all the oldstyle prefixes except in the compatibility code. struct libwebsockets becomes struct lws too. The api docs are updated accordingly as are the READMEs that mention those apis. Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-04 11:08:32 +08:00
lws_context_user(struct lws_context *context)
libwebsocket_context: add userspace pointer for use before wsi creation Signed-off-by: Alon Levy <alevy@redhat.com>
2012-10-19 11:21:56 +02:00
{
style cleanup This brings the library sources into compliance with checkpatch style except for three or four exceptions like WIN32 related stuff and one long string constant I don't want to break into multiple sprintf calls. There should be no functional or compilability change from all this (hopefully). Signed-off-by: Andy Green <andy.green@linaro.org>
2013-02-11 17:13:32 +08:00
return context->user_space;
libwebsocket_context: add userspace pointer for use before wsi creation Signed-off-by: Alon Levy <alevy@redhat.com>
2012-10-19 11:21:56 +02:00
}
plugins This adds support for dynamically loaded plugins at runtime, which can expose their own protocols or extensions transparently. With these changes lwsws defaults to OFF in cmake, and if enabled it automatically enables plugins and libuv support. Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-06 16:15:40 +08:00
LWS_VISIBLE struct lws_vhost *
lws_vhost_get(struct lws *wsi)
{
return wsi->vhost;
}
deprecate misnamed lws_vhost_get add lws_get_vhost All the other simple accessors begin lws_get_... Signed-off-by: Andy Green <andy@warmcat.com>
2016-05-17 13:47:44 +08:00
LWS_VISIBLE struct lws_vhost *
lws_get_vhost(struct lws *wsi)
{
return wsi->vhost;
}
plugins This adds support for dynamically loaded plugins at runtime, which can expose their own protocols or extensions transparently. With these changes lwsws defaults to OFF in cmake, and if enabled it automatically enables plugins and libuv support. Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-06 16:15:40 +08:00
LWS_VISIBLE const struct lws_protocols *
lws_protocol_get(struct lws *wsi)
{
return wsi->protocol;
}
http2: make usable HTTP/2 support is now able to serve the test server, complete with websockets, from a single vhost. - This works the same with both OpenSSL and mbedTLS. - POST is now wired up and works (also for file upload). - CGI is wired up and works. - Redirect is adapted and works - lwsws works. - URI urldecode, sanitation and argument parsing wired up for :path valgrind clean (aside from openssl-style false uninit data usage in mbedtls send occasionally) h2spec reports: $ h2spec -h 127.0.0.1 -p 7681 -t -k -o 1 ... 145 tests, 145 passed, 0 skipped, 0 failed" Incorporates: - "https://github.com/warmcat/libwebsockets/pull/1039 Fixes issue with -Werror=unused-variable flag - https://github.com/sartura/libwebsockets/commit/2c843a1395bdca5c23b40d989f508fc96aa5ba3b ssl: fix infinite loop on client cert verification failure Signed-off-by: Petar Paradzik <petar.paradzik@sartura.hr>" Caused and fixes Coverity 184887 - 184892
2017-10-13 10:33:02 +08:00
LWS_VISIBLE struct lws *
lws_get_network_wsi(struct lws *wsi)
{
if (!wsi)
return NULL;
#if defined(LWS_WITH_HTTP2)
if (!wsi->http2_substream)
return wsi;
while (wsi->u.h2.parent_wsi)
wsi = wsi->u.h2.parent_wsi;
#endif
return wsi;
}
client: allow http[s] to select targeted protocol name in vhost
2017-02-15 09:12:39 +08:00
LWS_VISIBLE LWS_EXTERN const struct lws_protocols *
lws_vhost_name_to_protocol(struct lws_vhost *vh, const char *name)
{
int n;
for (n = 0; n < vh->count_protocols; n++)
if (!strcmp(name, vh->protocols[n].name))
return &vh->protocols[n];
return NULL;
}
libwebsocket_callback_all_protocol Signed-off-by: Andy Green <andy.green@linaro.org>
2014-02-15 16:36:38 +08:00
LWS_VISIBLE int
make protocols const require explicit context API BREAK The user protocols struct has not been const until now. This has been painful for a while because the semantics of the protocols struct look like it's going to be treated as const. At context creation, the protocols struct has been getting marked with the context, and three apis exploited that to only need to be passed a pointer to a protocol to get access to the context. This patch removes the two writeable members in the context (these were never directly used by user code), changes all pointers to protocols to be const, and adds an explicit first argument to the three affected apis so they can have access to context. The three affected apis are these LWS_VISIBLE LWS_EXTERN int -lws_callback_on_writable_all_protocol(const struct lws_protocols *protocol); +lws_callback_on_writable_all_protocol(const struct lws_context *context, + const struct lws_protocols *protocol); LWS_VISIBLE LWS_EXTERN int -lws_callback_all_protocol(const struct lws_protocols *protocol, int reason); +lws_callback_all_protocol(struct lws_context *context, + const struct lws_protocols *protocol, int reason); LWS_VISIBLE LWS_EXTERN void -lws_rx_flow_allow_all_protocol(const struct lws_protocols *protocol); +lws_rx_flow_allow_all_protocol(const struct lws_context *context, + const struct lws_protocols *protocol); unfortunately the original apis can no longer be emulated and users of them must update. Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-11 10:45:35 +08:00
lws_callback_all_protocol(struct lws_context *context,
const struct lws_protocols *protocol, int reason)
libwebsocket_callback_all_protocol Signed-off-by: Andy Green <andy.green@linaro.org>
2014-02-15 16:36:38 +08:00
{
multithreaded service This adds support for multithreaded service to lws without adding any threading or locking code in the library. At context creation time you can request split the service part of the context into n service domains, which are load-balanced so that the most idle one gets the next listen socket accept. There's a single listen socket on one port still. User code may then spawn n threads doing n service loops / poll()s simultaneously. Locking is only required (I think) in the existing FD lock callbacks already handled by the pthreads server example, and that locking takes place in user code. So the library remains completely agnostic about the threading / locking scheme. And by default, it's completely compatible with one service thread so no changes are required by people uninterested in multithreaded service. However for people interested in extremely lightweight mass http[s]/ ws[s] service with minimum provisioning, the library can now do everything out of the box. To test it, just try $ libwebsockets-test-server-pthreads -j 8 where -j controls the number of service threads Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-19 03:34:24 +08:00
struct lws_context_per_thread *pt = &context->pt[0];
multithread stability Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-26 20:56:56 +08:00
unsigned int n, m = context->count_threads;
clean pre 1.7 Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-29 21:18:54 +08:00
struct lws *wsi;
multithreaded service This adds support for multithreaded service to lws without adding any threading or locking code in the library. At context creation time you can request split the service part of the context into n service domains, which are load-balanced so that the most idle one gets the next listen socket accept. There's a single listen socket on one port still. User code may then spawn n threads doing n service loops / poll()s simultaneously. Locking is only required (I think) in the existing FD lock callbacks already handled by the pthreads server example, and that locking takes place in user code. So the library remains completely agnostic about the threading / locking scheme. And by default, it's completely compatible with one service thread so no changes are required by people uninterested in multithreaded service. However for people interested in extremely lightweight mass http[s]/ ws[s] service with minimum provisioning, the library can now do everything out of the box. To test it, just try $ libwebsockets-test-server-pthreads -j 8 where -j controls the number of service threads Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-19 03:34:24 +08:00
while (m--) {
for (n = 0; n < pt->fds_count; n++) {
wsi = wsi_from_fd(context, pt->fds[n].fd);
if (!wsi)
continue;
if (wsi->protocol == protocol)
clean pre 1.7 Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-29 21:18:54 +08:00
protocol->callback(wsi, reason, wsi->user_space,
NULL, 0);
multithreaded service This adds support for multithreaded service to lws without adding any threading or locking code in the library. At context creation time you can request split the service part of the context into n service domains, which are load-balanced so that the most idle one gets the next listen socket accept. There's a single listen socket on one port still. User code may then spawn n threads doing n service loops / poll()s simultaneously. Locking is only required (I think) in the existing FD lock callbacks already handled by the pthreads server example, and that locking takes place in user code. So the library remains completely agnostic about the threading / locking scheme. And by default, it's completely compatible with one service thread so no changes are required by people uninterested in multithreaded service. However for people interested in extremely lightweight mass http[s]/ ws[s] service with minimum provisioning, the library can now do everything out of the box. To test it, just try $ libwebsockets-test-server-pthreads -j 8 where -j controls the number of service threads Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-19 03:34:24 +08:00
}
pt++;
libwebsocket_callback_all_protocol Signed-off-by: Andy Green <andy.green@linaro.org>
2014-02-15 16:36:38 +08:00
}
return 0;
}
plugins This adds support for dynamically loaded plugins at runtime, which can expose their own protocols or extensions transparently. With these changes lwsws defaults to OFF in cmake, and if enabled it automatically enables plugins and libuv support. Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-06 16:15:40 +08:00
LWS_VISIBLE int
lws_callback_all_protocol_vhost_args
2017-08-05 10:38:59 +08:00
lws_callback_all_protocol_vhost_args(struct lws_vhost *vh,
const struct lws_protocols *protocol, int reason,
void *argp, size_t len)
plugins This adds support for dynamically loaded plugins at runtime, which can expose their own protocols or extensions transparently. With these changes lwsws defaults to OFF in cmake, and if enabled it automatically enables plugins and libuv support. Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-06 16:15:40 +08:00
{
struct lws_context *context = vh->context;
struct lws_context_per_thread *pt = &context->pt[0];
unsigned int n, m = context->count_threads;
struct lws *wsi;
while (m--) {
for (n = 0; n < pt->fds_count; n++) {
wsi = wsi_from_fd(context, pt->fds[n].fd);
if (!wsi)
continue;
lws_callback_all_protocol_vhost_args
2017-08-05 10:38:59 +08:00
if (wsi->vhost == vh && (wsi->protocol == protocol ||
!protocol))
wsi->protocol->callback(wsi, reason,
wsi->user_space, argp, len);
plugins This adds support for dynamically loaded plugins at runtime, which can expose their own protocols or extensions transparently. With these changes lwsws defaults to OFF in cmake, and if enabled it automatically enables plugins and libuv support. Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-06 16:15:40 +08:00
}
pt++;
}
return 0;
}
lws_callback_all_protocol_vhost_args
2017-08-05 10:38:59 +08:00
LWS_VISIBLE int
lws_callback_all_protocol_vhost(struct lws_vhost *vh,
const struct lws_protocols *protocol, int reason)
{
return lws_callback_all_protocol_vhost_args(vh, protocol, reason, NULL, 0);
}
lws_callback_vhost_protocols This gives protocols a way to talk to each other via per-vhost callbacks, one per protocol (including the sender). Signed-off-by: Andy Green <andy@warmcat.com>
2016-06-15 10:46:58 +08:00
LWS_VISIBLE LWS_EXTERN int
lws_callback_vhost_protocols(struct lws *wsi, int reason, void *in, int len)
{
int n;
for (n = 0; n < wsi->vhost->count_protocols; n++)
if (wsi->vhost->protocols[n].callback(wsi, reason, NULL, in, len))
return 1;
return 0;
}
ESP32 platform This is enough for all the test app features to work on ESP32 without SSL.
2017-02-18 17:26:40 +08:00
LWS_VISIBLE LWS_EXTERN void
esp32: align fops member name defines
2017-03-08 11:11:41 +08:00
lws_set_fops(struct lws_context *context, const struct lws_plat_file_ops *fops)
ESP32 platform This is enough for all the test app features to work on ESP32 without SSL.
2017-02-18 17:26:40 +08:00
{
esp32: align fops member name defines
2017-03-08 11:11:41 +08:00
context->fops = fops;
fops-zip
2017-03-03 12:38:10 +08:00
}
LWS_VISIBLE LWS_EXTERN lws_filepos_t
lws_vfs_tell(lws_fop_fd_t fop_fd)
{
return fop_fd->pos;
}
LWS_VISIBLE LWS_EXTERN lws_filepos_t
lws_vfs_get_length(lws_fop_fd_t fop_fd)
{
return fop_fd->len;
}
LWS_VISIBLE LWS_EXTERN uint32_t
lws_vfs_get_mod_time(lws_fop_fd_t fop_fd)
{
return fop_fd->mod_time;
}
LWS_VISIBLE lws_fileofs_t
lws_vfs_file_seek_set(lws_fop_fd_t fop_fd, lws_fileofs_t offset)
{
lws_fileofs_t ofs;
clean
2017-09-23 12:55:21 +08:00
fops-zip
2017-03-03 12:38:10 +08:00
ofs = fop_fd->fops->LWS_FOP_SEEK_CUR(fop_fd, offset - fop_fd->pos);
clean
2017-09-23 12:55:21 +08:00
fops-zip
2017-03-03 12:38:10 +08:00
return ofs;
}
LWS_VISIBLE lws_fileofs_t
lws_vfs_file_seek_end(lws_fop_fd_t fop_fd, lws_fileofs_t offset)
{
clean
2017-09-23 12:55:21 +08:00
return fop_fd->fops->LWS_FOP_SEEK_CUR(fop_fd, fop_fd->len +
fop_fd->pos + offset);
fops: allow setting from context creation and introduce lws_select_fops_by_vfs_path 1) There's now a .fops pointer that can be set in the context creation info. If set, the array of fops it points to (terminated by an entry with .open = NULL) is walked to find out the best vfs filesystem path match (comparing the vfs path to fops.path_prefix) for which fops to use. If none given (.fops is NULL in info) then behaviour is as before, fops are the platform-provided one only. 2) The built in fileserving now walks any array of fops looking for the best fops match automatically. 3) lws_plat_file_... apis are renamed to lws_vfs_file_...
2017-03-01 14:28:56 +08:00
}
fops-zip
2017-03-03 12:38:10 +08:00
const struct lws_plat_file_ops *
lws_vfs_select_fops(const struct lws_plat_file_ops *fops, const char *vfs_path,
const char **vpath)
fops: allow setting from context creation and introduce lws_select_fops_by_vfs_path 1) There's now a .fops pointer that can be set in the context creation info. If set, the array of fops it points to (terminated by an entry with .open = NULL) is walked to find out the best vfs filesystem path match (comparing the vfs path to fops.path_prefix) for which fops to use. If none given (.fops is NULL in info) then behaviour is as before, fops are the platform-provided one only. 2) The built in fileserving now walks any array of fops looking for the best fops match automatically. 3) lws_plat_file_... apis are renamed to lws_vfs_file_...
2017-03-01 14:28:56 +08:00
{
fops-zip
2017-03-03 12:38:10 +08:00
const struct lws_plat_file_ops *pf;
const char *p = vfs_path;
int n;
fops: allow setting from context creation and introduce lws_select_fops_by_vfs_path 1) There's now a .fops pointer that can be set in the context creation info. If set, the array of fops it points to (terminated by an entry with .open = NULL) is walked to find out the best vfs filesystem path match (comparing the vfs path to fops.path_prefix) for which fops to use. If none given (.fops is NULL in info) then behaviour is as before, fops are the platform-provided one only. 2) The built in fileserving now walks any array of fops looking for the best fops match automatically. 3) lws_plat_file_... apis are renamed to lws_vfs_file_...
2017-03-01 14:28:56 +08:00
fops-zip
2017-03-03 12:38:10 +08:00
*vpath = NULL;
/* no non-platform fops, just use that */
if (!fops->next)
return fops;
/*
* scan the vfs path looking for indications we are to be
* handled by a specific fops
*/
fops: allow setting from context creation and introduce lws_select_fops_by_vfs_path 1) There's now a .fops pointer that can be set in the context creation info. If set, the array of fops it points to (terminated by an entry with .open = NULL) is walked to find out the best vfs filesystem path match (comparing the vfs path to fops.path_prefix) for which fops to use. If none given (.fops is NULL in info) then behaviour is as before, fops are the platform-provided one only. 2) The built in fileserving now walks any array of fops looking for the best fops match automatically. 3) lws_plat_file_... apis are renamed to lws_vfs_file_...
2017-03-01 14:28:56 +08:00
esp32: align fops member name defines
2017-03-08 11:11:41 +08:00
while (p && *p) {
fops-zip
2017-03-03 12:38:10 +08:00
if (*p != '/') {
p++;
continue;
}
/* the first one is always platform fops, so skip */
pf = fops->next;
while (pf) {
n = 0;
while (n < ARRAY_SIZE(pf->fi) && pf->fi[n].sig) {
if (p >= vfs_path + pf->fi[n].len)
if (!strncmp(p - (pf->fi[n].len - 1),
pf->fi[n].sig,
pf->fi[n].len - 1)) {
*vpath = p + 1;
return pf;
}
n++;
fops: allow setting from context creation and introduce lws_select_fops_by_vfs_path 1) There's now a .fops pointer that can be set in the context creation info. If set, the array of fops it points to (terminated by an entry with .open = NULL) is walked to find out the best vfs filesystem path match (comparing the vfs path to fops.path_prefix) for which fops to use. If none given (.fops is NULL in info) then behaviour is as before, fops are the platform-provided one only. 2) The built in fileserving now walks any array of fops looking for the best fops match automatically. 3) lws_plat_file_... apis are renamed to lws_vfs_file_...
2017-03-01 14:28:56 +08:00
}
fops-zip
2017-03-03 12:38:10 +08:00
pf = pf->next;
}
p++;
fops: allow setting from context creation and introduce lws_select_fops_by_vfs_path 1) There's now a .fops pointer that can be set in the context creation info. If set, the array of fops it points to (terminated by an entry with .open = NULL) is walked to find out the best vfs filesystem path match (comparing the vfs path to fops.path_prefix) for which fops to use. If none given (.fops is NULL in info) then behaviour is as before, fops are the platform-provided one only. 2) The built in fileserving now walks any array of fops looking for the best fops match automatically. 3) lws_plat_file_... apis are renamed to lws_vfs_file_...
2017-03-01 14:28:56 +08:00
}
fops-zip
2017-03-03 12:38:10 +08:00
return fops;
}
LWS_VISIBLE LWS_EXTERN lws_fop_fd_t LWS_WARN_UNUSED_RESULT
lws_vfs_file_open(const struct lws_plat_file_ops *fops, const char *vfs_path,
lws_fop_flags_t *flags)
{
esp32: align fops member name defines
2017-03-08 11:11:41 +08:00
const char *vpath = "";
clean
2017-09-23 12:55:21 +08:00
const struct lws_plat_file_ops *selected;
selected = lws_vfs_select_fops(fops, vfs_path, &vpath);
fops: allow setting from context creation and introduce lws_select_fops_by_vfs_path 1) There's now a .fops pointer that can be set in the context creation info. If set, the array of fops it points to (terminated by an entry with .open = NULL) is walked to find out the best vfs filesystem path match (comparing the vfs path to fops.path_prefix) for which fops to use. If none given (.fops is NULL in info) then behaviour is as before, fops are the platform-provided one only. 2) The built in fileserving now walks any array of fops looking for the best fops match automatically. 3) lws_plat_file_... apis are renamed to lws_vfs_file_...
2017-03-01 14:28:56 +08:00
fops-zip
2017-03-03 12:38:10 +08:00
return selected->LWS_FOP_OPEN(fops, vfs_path, vpath, flags);
ESP32 platform This is enough for all the test app features to work on ESP32 without SSL.
2017-02-18 17:26:40 +08:00
}
fops-zip
2017-03-03 12:38:10 +08:00
lws_now_secs Signed-off-by: Andy Green <andy@warmcat.com>
2016-06-17 09:41:22 +08:00
/**
* lws_now_secs() - seconds since 1970-1-1
*
*/
LWS_VISIBLE LWS_EXTERN unsigned long
lws_now_secs(void)
{
struct timeval tv;
gettimeofday(&tv, NULL);
return tv.tv_sec;
}
mbed3 plat Signed-off-by: Andy Green <andy.green@linaro.org>
2015-11-02 20:34:12 +08:00
#if LWS_POSIX
Hide private library symbols Signed-off-by: Peter Pentchev <roam@ringlet.net>
2013-03-30 09:52:21 +08:00
LWS_VISIBLE int
api rationalization: eliminate all libwebsocket[s]_ prefixes This nukes all the oldstyle prefixes except in the compatibility code. struct libwebsockets becomes struct lws too. The api docs are updated accordingly as are the READMEs that mention those apis. Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-04 11:08:32 +08:00
lws_get_socket_fd(struct lws *wsi)
introduce-test-ping-app.patch Signed-off-by: Andy Green <andy@warmcat.com>
2011-01-27 20:06:03 +00:00
{
cgi: also allow for generic spawn
2017-09-19 07:58:04 +08:00
if (!wsi)
return -1;
adoption: make union for socket and file fds This lets lws support adopting raw file FDs and raw socket fds. A test plugin creates a FIFO and prints data sent on it, using the lws event loop.
2017-02-27 12:55:56 +08:00
return wsi->desc.sockfd;
introduce-test-ping-app.patch Signed-off-by: Andy Green <andy@warmcat.com>
2011-01-27 20:06:03 +00:00
}
mbed3 plat Signed-off-by: Andy Green <andy.green@linaro.org>
2015-11-02 20:34:12 +08:00
#endif
introduce lws_latency Signed-off-by: Andy Green <andy.green@linaro.org>
2013-01-29 12:36:17 +08:00
#ifdef LWS_LATENCY
void
clean tidy the worst whitespace alignment probs due to mass token name length changes Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-04 11:30:53 +08:00
lws_latency(struct lws_context *context, struct lws *wsi, const char *action,
int ret, int completed)
introduce lws_latency Signed-off-by: Andy Green <andy.green@linaro.org>
2013-01-29 12:36:17 +08:00
{
Avoid calls to gettimeofday() Add a new function to get the current time in microseconds, since gettimeofday() does not exist on Windows. Keep the current implementation for the test applications.
2014-02-26 21:37:31 +01:00
unsigned long long u;
introduce lws_latency Signed-off-by: Andy Green <andy.green@linaro.org>
2013-01-29 12:36:17 +08:00
char buf[256];
Avoid calls to gettimeofday() Add a new function to get the current time in microseconds, since gettimeofday() does not exist on Windows. Keep the current implementation for the test applications.
2014-02-26 21:37:31 +01:00
u = time_in_microseconds();
introduce lws_latency Signed-off-by: Andy Green <andy.green@linaro.org>
2013-01-29 12:36:17 +08:00
clean libwebsockets.c Signed-off-by: None <None>
2014-03-31 11:01:32 +08:00
if (!action) {
introduce lws_latency Signed-off-by: Andy Green <andy.green@linaro.org>
2013-01-29 12:36:17 +08:00
wsi->latency_start = u;
if (!wsi->action_start)
wsi->action_start = u;
clean libwebsockets.c Signed-off-by: None <None>
2014-03-31 11:01:32 +08:00
return;
introduce lws_latency Signed-off-by: Andy Green <andy.green@linaro.org>
2013-01-29 12:36:17 +08:00
}
clean libwebsockets.c Signed-off-by: None <None>
2014-03-31 11:01:32 +08:00
if (completed) {
if (wsi->action_start == wsi->latency_start)
sprintf(buf,
fix confusion ull in latency reporting
2014-07-05 10:50:47 +08:00
"Completion first try lat %lluus: %p: ret %d: %s\n",
clean libwebsockets.c Signed-off-by: None <None>
2014-03-31 11:01:32 +08:00
u - wsi->latency_start,
(void *)wsi, ret, action);
else
sprintf(buf,
fix confusion ull in latency reporting
2014-07-05 10:50:47 +08:00
"Completion %lluus: lat %lluus: %p: ret %d: %s\n",
clean libwebsockets.c Signed-off-by: None <None>
2014-03-31 11:01:32 +08:00
u - wsi->action_start,
u - wsi->latency_start,
(void *)wsi, ret, action);
wsi->action_start = 0;
} else
fix confusion ull in latency reporting
2014-07-05 10:50:47 +08:00
sprintf(buf, "lat %lluus: %p: ret %d: %s\n",
clean libwebsockets.c Signed-off-by: None <None>
2014-03-31 11:01:32 +08:00
u - wsi->latency_start, (void *)wsi, ret, action);
if (u - wsi->latency_start > context->worst_latency) {
context->worst_latency = u - wsi->latency_start;
strcpy(context->worst_latency_info, buf);
}
lwsl_latency("%s", buf);
introduce lws_latency Signed-off-by: Andy Green <andy.green@linaro.org>
2013-01-29 12:36:17 +08:00
}
#endif
Hide private library symbols Signed-off-by: Peter Pentchev <roam@ringlet.net>
2013-03-30 09:52:21 +08:00
LWS_VISIBLE int
http2: make usable HTTP/2 support is now able to serve the test server, complete with websockets, from a single vhost. - This works the same with both OpenSSL and mbedTLS. - POST is now wired up and works (also for file upload). - CGI is wired up and works. - Redirect is adapted and works - lwsws works. - URI urldecode, sanitation and argument parsing wired up for :path valgrind clean (aside from openssl-style false uninit data usage in mbedtls send occasionally) h2spec reports: $ h2spec -h 127.0.0.1 -p 7681 -t -k -o 1 ... 145 tests, 145 passed, 0 skipped, 0 failed" Incorporates: - "https://github.com/warmcat/libwebsockets/pull/1039 Fixes issue with -Werror=unused-variable flag - https://github.com/sartura/libwebsockets/commit/2c843a1395bdca5c23b40d989f508fc96aa5ba3b ssl: fix infinite loop on client cert verification failure Signed-off-by: Petar Paradzik <petar.paradzik@sartura.hr>" Caused and fixes Coverity 184887 - 184892
2017-10-13 10:33:02 +08:00
lws_rx_flow_control(struct lws *wsi, int _enable)
solve flowcontrol problems Problems with rx flow control implementation were the underlying cause of the connection stalling issue that was covered up with the udelay() patch that was removed recently. This get rx flow control working properly and corrects problems with fifo management in the test server mirror protocol code too. The rxfow control api has been changed to just set a flag, so it's very cheap to call from user code. After the callbacks that might use the rxflow control api the flag is checked and any pending actions done. rx flow control now stops any rx packet coming immediately, with compessed connections "just what was left in the pipe" might be hundreds of KBytes. To implement that the current packet being decoded is copied into a malloc'd buffer by the rx processing code now. When rxflow is allows to come again, the buffer is drained and freed before any new packet content is accepted. Signed-off-by: Andy Green <andy.green@linaro.org>
2013-01-17 16:50:35 +08:00
{
http2: make usable HTTP/2 support is now able to serve the test server, complete with websockets, from a single vhost. - This works the same with both OpenSSL and mbedTLS. - POST is now wired up and works (also for file upload). - CGI is wired up and works. - Redirect is adapted and works - lwsws works. - URI urldecode, sanitation and argument parsing wired up for :path valgrind clean (aside from openssl-style false uninit data usage in mbedtls send occasionally) h2spec reports: $ h2spec -h 127.0.0.1 -p 7681 -t -k -o 1 ... 145 tests, 145 passed, 0 skipped, 0 failed" Incorporates: - "https://github.com/warmcat/libwebsockets/pull/1039 Fixes issue with -Werror=unused-variable flag - https://github.com/sartura/libwebsockets/commit/2c843a1395bdca5c23b40d989f508fc96aa5ba3b ssl: fix infinite loop on client cert verification failure Signed-off-by: Petar Paradzik <petar.paradzik@sartura.hr>" Caused and fixes Coverity 184887 - 184892
2017-10-13 10:33:02 +08:00
int en = _enable;
lwsl_info("%s: %p 0x%x\n", __func__, wsi, _enable);
if (!(_enable & LWS_RXFLOW_REASON_APPLIES)) {
/*
* convert user bool style to bitmap style... in user simple
* bool style _enable = 0 = flow control it, = 1 = allow rx
*/
en = LWS_RXFLOW_REASON_APPLIES | LWS_RXFLOW_REASON_USER_BOOL;
if (_enable & 1)
en |= LWS_RXFLOW_REASON_APPLIES_ENABLE_BIT;
}
/* any bit set in rxflow_bitmap DISABLEs rxflow control */
if (en & LWS_RXFLOW_REASON_APPLIES_ENABLE_BIT)
wsi->rxflow_bitmap &= ~(en & 0xff);
else
wsi->rxflow_bitmap |= en & 0xff;
if ((LWS_RXFLOW_PENDING_CHANGE | (!wsi->rxflow_bitmap)) ==
wsi->rxflow_change_to)
rxflow remove recursion and simplify Signed-off-by: Andy Green <andy.green@linaro.org>
2013-03-16 11:24:23 +08:00
return 0;
http2: make usable HTTP/2 support is now able to serve the test server, complete with websockets, from a single vhost. - This works the same with both OpenSSL and mbedTLS. - POST is now wired up and works (also for file upload). - CGI is wired up and works. - Redirect is adapted and works - lwsws works. - URI urldecode, sanitation and argument parsing wired up for :path valgrind clean (aside from openssl-style false uninit data usage in mbedtls send occasionally) h2spec reports: $ h2spec -h 127.0.0.1 -p 7681 -t -k -o 1 ... 145 tests, 145 passed, 0 skipped, 0 failed" Incorporates: - "https://github.com/warmcat/libwebsockets/pull/1039 Fixes issue with -Werror=unused-variable flag - https://github.com/sartura/libwebsockets/commit/2c843a1395bdca5c23b40d989f508fc96aa5ba3b ssl: fix infinite loop on client cert verification failure Signed-off-by: Petar Paradzik <petar.paradzik@sartura.hr>" Caused and fixes Coverity 184887 - 184892
2017-10-13 10:33:02 +08:00
wsi->rxflow_change_to = LWS_RXFLOW_PENDING_CHANGE | !wsi->rxflow_bitmap;
lwsl_info("%s: 0x%p: bitmap 0x%x: en 0x%x, ch 0x%x\n", __func__, wsi,
wsi->rxflow_bitmap, en, wsi->rxflow_change_to);
if (_enable & LWS_RXFLOW_REASON_FLAG_PROCESS_NOW ||
!wsi->rxflow_will_be_applied)
return _lws_rx_flow_control(wsi);
solve flowcontrol problems Problems with rx flow control implementation were the underlying cause of the connection stalling issue that was covered up with the udelay() patch that was removed recently. This get rx flow control working properly and corrects problems with fifo management in the test server mirror protocol code too. The rxfow control api has been changed to just set a flag, so it's very cheap to call from user code. After the callbacks that might use the rxflow control api the flag is checked and any pending actions done. rx flow control now stops any rx packet coming immediately, with compessed connections "just what was left in the pipe" might be hundreds of KBytes. To implement that the current packet being decoded is copied into a malloc'd buffer by the rx processing code now. When rxflow is allows to come again, the buffer is drained and freed before any new packet content is accepted. Signed-off-by: Andy Green <andy.green@linaro.org>
2013-01-17 16:50:35 +08:00
return 0;
}
Hide private library symbols Signed-off-by: Peter Pentchev <roam@ringlet.net>
2013-03-30 09:52:21 +08:00
LWS_VISIBLE void
make protocols const require explicit context API BREAK The user protocols struct has not been const until now. This has been painful for a while because the semantics of the protocols struct look like it's going to be treated as const. At context creation, the protocols struct has been getting marked with the context, and three apis exploited that to only need to be passed a pointer to a protocol to get access to the context. This patch removes the two writeable members in the context (these were never directly used by user code), changes all pointers to protocols to be const, and adds an explicit first argument to the three affected apis so they can have access to context. The three affected apis are these LWS_VISIBLE LWS_EXTERN int -lws_callback_on_writable_all_protocol(const struct lws_protocols *protocol); +lws_callback_on_writable_all_protocol(const struct lws_context *context, + const struct lws_protocols *protocol); LWS_VISIBLE LWS_EXTERN int -lws_callback_all_protocol(const struct lws_protocols *protocol, int reason); +lws_callback_all_protocol(struct lws_context *context, + const struct lws_protocols *protocol, int reason); LWS_VISIBLE LWS_EXTERN void -lws_rx_flow_allow_all_protocol(const struct lws_protocols *protocol); +lws_rx_flow_allow_all_protocol(const struct lws_context *context, + const struct lws_protocols *protocol); unfortunately the original apis can no longer be emulated and users of them must update. Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-11 10:45:35 +08:00
lws_rx_flow_allow_all_protocol(const struct lws_context *context,
const struct lws_protocols *protocol)
introduce api for unthrottle all connections of protocol Signed-off-by: Andy Green <andy.green@linaro.org>
2013-03-16 12:32:27 +08:00
{
multithreaded service This adds support for multithreaded service to lws without adding any threading or locking code in the library. At context creation time you can request split the service part of the context into n service domains, which are load-balanced so that the most idle one gets the next listen socket accept. There's a single listen socket on one port still. User code may then spawn n threads doing n service loops / poll()s simultaneously. Locking is only required (I think) in the existing FD lock callbacks already handled by the pthreads server example, and that locking takes place in user code. So the library remains completely agnostic about the threading / locking scheme. And by default, it's completely compatible with one service thread so no changes are required by people uninterested in multithreaded service. However for people interested in extremely lightweight mass http[s]/ ws[s] service with minimum provisioning, the library can now do everything out of the box. To test it, just try $ libwebsockets-test-server-pthreads -j 8 where -j controls the number of service threads Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-19 03:34:24 +08:00
const struct lws_context_per_thread *pt = &context->pt[0];
api rationalization: eliminate all libwebsocket[s]_ prefixes This nukes all the oldstyle prefixes except in the compatibility code. struct libwebsockets becomes struct lws too. The api docs are updated accordingly as are the READMEs that mention those apis. Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-04 11:08:32 +08:00
struct lws *wsi;
multithread stability Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-26 20:56:56 +08:00
unsigned int n, m = context->count_threads;
multithreaded service This adds support for multithreaded service to lws without adding any threading or locking code in the library. At context creation time you can request split the service part of the context into n service domains, which are load-balanced so that the most idle one gets the next listen socket accept. There's a single listen socket on one port still. User code may then spawn n threads doing n service loops / poll()s simultaneously. Locking is only required (I think) in the existing FD lock callbacks already handled by the pthreads server example, and that locking takes place in user code. So the library remains completely agnostic about the threading / locking scheme. And by default, it's completely compatible with one service thread so no changes are required by people uninterested in multithreaded service. However for people interested in extremely lightweight mass http[s]/ ws[s] service with minimum provisioning, the library can now do everything out of the box. To test it, just try $ libwebsockets-test-server-pthreads -j 8 where -j controls the number of service threads Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-19 03:34:24 +08:00
while (m--) {
for (n = 0; n < pt->fds_count; n++) {
wsi = wsi_from_fd(context, pt->fds[n].fd);
if (!wsi)
continue;
if (wsi->protocol == protocol)
lws_rx_flow_control(wsi, LWS_RXFLOW_ALLOW);
}
pt++;
introduce api for unthrottle all connections of protocol Signed-off-by: Andy Green <andy.green@linaro.org>
2013-03-16 12:32:27 +08:00
}
}
Hide private library symbols Signed-off-by: Peter Pentchev <roam@ringlet.net>
2013-03-30 09:52:21 +08:00
LWS_VISIBLE extern const char *
api rationalization: eliminate all libwebsocket[s]_ prefixes This nukes all the oldstyle prefixes except in the compatibility code. struct libwebsockets becomes struct lws too. The api docs are updated accordingly as are the READMEs that mention those apis. Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-04 11:08:32 +08:00
lws_canonical_hostname(struct lws_context *context)
fix-host-header-contents--introduce-canonical-hostname-api.patch I?aki pointed out the dummy host field used in client test and ping is not valid http. This patch changes it to use the actual host name and adds an api to collect that from the context cheaply. Reported-by: I?aki Baz Castillo <ibc@aliax.net> Signed-off-by: Andy Green <andy@warmcat.com>
2011-01-28 10:00:18 +00:00
{
introduce win32 build capability This adds win32 build compatability to libwebsockets. The patch is from Peter Hinz, Andy Green has cleaned it up a bit and possibly broken win32 compatability since I can't test it, so there may be followup patches. It compiles fine under Linux after this patch anyway. Much of the patch is changing a reserved keyword for Visual C compiler "this" to "context", but there is no real C99 support in the MSFT compiler even though it is 2011 so C99 style array declarations have been mangled back into "ancient C" style. Some windows-isms are also added like closesocket() but these are quite localized. Win32 random is just using C library random() call at the moment vs Linux /dev/urandom. canonical hostname detection is broken in win32 at the moment. Signed-off-by: Peter Hinz <cerebusrc@gmail.com> Signed-off-by: Andy Green <andy@warmcat.com>
2011-03-02 22:03:47 +00:00
return (const char *)context->canonical_hostname;
fix-host-header-contents--introduce-canonical-hostname-api.patch I?aki pointed out the dummy host field used in client test and ping is not valid http. This patch changes it to use the actual host name and adds an api to collect that from the context cheaply. Reported-by: I?aki Baz Castillo <ibc@aliax.net> Signed-off-by: Andy Green <andy@warmcat.com>
2011-01-28 10:00:18 +00:00
}
fix callback typedefs and declarations Remove declarations of callback and extension_callback as these are functions declared in header but not defined anywhere. Also rename typedefs callback_function and extension_callback_function to lws_callback_function and lws_extension_callback_function so all symbolx exported by header have lws prefix; Signed-off-by: Denis Osvald <denis.osvald@sartura.hr>
2015-12-21 18:06:38 +01:00
int user_callback_handle_rxflow(lws_callback_function callback_function,
public api remove context from user callback API BREAK Since struct lws (wsi) now has his own context pointer, we were able to remove the need for passing context almost everywhere in the apis. In turn, that means there's no real use for context being passed to every callback; in the rare cases context is needed user code can get it with lws_get_ctx(wsi) Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-17 07:54:44 +08:00
struct lws *wsi,
clean tidy the worst whitespace alignment probs due to mass token name length changes Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-04 11:30:53 +08:00
enum lws_callback_reasons reason, void *user,
void *in, size_t len)
solve flowcontrol problems Problems with rx flow control implementation were the underlying cause of the connection stalling issue that was covered up with the udelay() patch that was removed recently. This get rx flow control working properly and corrects problems with fifo management in the test server mirror protocol code too. The rxfow control api has been changed to just set a flag, so it's very cheap to call from user code. After the callbacks that might use the rxflow control api the flag is checked and any pending actions done. rx flow control now stops any rx packet coming immediately, with compessed connections "just what was left in the pipe" might be hundreds of KBytes. To implement that the current packet being decoded is copied into a malloc'd buffer by the rx processing code now. When rxflow is allows to come again, the buffer is drained and freed before any new packet content is accepted. Signed-off-by: Andy Green <andy.green@linaro.org>
2013-01-17 16:50:35 +08:00
{
int n;
http2: make usable HTTP/2 support is now able to serve the test server, complete with websockets, from a single vhost. - This works the same with both OpenSSL and mbedTLS. - POST is now wired up and works (also for file upload). - CGI is wired up and works. - Redirect is adapted and works - lwsws works. - URI urldecode, sanitation and argument parsing wired up for :path valgrind clean (aside from openssl-style false uninit data usage in mbedtls send occasionally) h2spec reports: $ h2spec -h 127.0.0.1 -p 7681 -t -k -o 1 ... 145 tests, 145 passed, 0 skipped, 0 failed" Incorporates: - "https://github.com/warmcat/libwebsockets/pull/1039 Fixes issue with -Werror=unused-variable flag - https://github.com/sartura/libwebsockets/commit/2c843a1395bdca5c23b40d989f508fc96aa5ba3b ssl: fix infinite loop on client cert verification failure Signed-off-by: Petar Paradzik <petar.paradzik@sartura.hr>" Caused and fixes Coverity 184887 - 184892
2017-10-13 10:33:02 +08:00
wsi->rxflow_will_be_applied = 1;
public api remove context from user callback API BREAK Since struct lws (wsi) now has his own context pointer, we were able to remove the need for passing context almost everywhere in the apis. In turn, that means there's no real use for context being passed to every callback; in the rare cases context is needed user code can get it with lws_get_ctx(wsi) Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-17 07:54:44 +08:00
n = callback_function(wsi, reason, user, in, len);
http2: make usable HTTP/2 support is now able to serve the test server, complete with websockets, from a single vhost. - This works the same with both OpenSSL and mbedTLS. - POST is now wired up and works (also for file upload). - CGI is wired up and works. - Redirect is adapted and works - lwsws works. - URI urldecode, sanitation and argument parsing wired up for :path valgrind clean (aside from openssl-style false uninit data usage in mbedtls send occasionally) h2spec reports: $ h2spec -h 127.0.0.1 -p 7681 -t -k -o 1 ... 145 tests, 145 passed, 0 skipped, 0 failed" Incorporates: - "https://github.com/warmcat/libwebsockets/pull/1039 Fixes issue with -Werror=unused-variable flag - https://github.com/sartura/libwebsockets/commit/2c843a1395bdca5c23b40d989f508fc96aa5ba3b ssl: fix infinite loop on client cert verification failure Signed-off-by: Petar Paradzik <petar.paradzik@sartura.hr>" Caused and fixes Coverity 184887 - 184892
2017-10-13 10:33:02 +08:00
wsi->rxflow_will_be_applied = 0;
dont close in user callback wrapper let ancestor do it Signed-off-by: Andy Green <andy.green@linaro.org>
2013-02-10 21:21:24 +08:00
if (!n)
api rationalization use new names internally Change all internal uses of rationalized public apis to reflect the new names. Theer are a few things that got changed as side effect of search/replace matches, but these are almost all internal. I added a compatibility define for the public enum that got renamed. Theoretically existing code should not notice the difference from these two patches. And new code will find the new names. https://github.com/warmcat/libwebsockets/issues/357 Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-04 08:43:54 +08:00
n = _lws_rx_flow_control(wsi);
solve flowcontrol problems Problems with rx flow control implementation were the underlying cause of the connection stalling issue that was covered up with the udelay() patch that was removed recently. This get rx flow control working properly and corrects problems with fifo management in the test server mirror protocol code too. The rxfow control api has been changed to just set a flag, so it's very cheap to call from user code. After the callbacks that might use the rxflow control api the flag is checked and any pending actions done. rx flow control now stops any rx packet coming immediately, with compessed connections "just what was left in the pipe" might be hundreds of KBytes. To implement that the current packet being decoded is copied into a malloc'd buffer by the rx processing code now. When rxflow is allows to come again, the buffer is drained and freed before any new packet content is accepted. Signed-off-by: Andy Green <andy.green@linaro.org>
2013-01-17 16:50:35 +08:00
dont close in user callback wrapper let ancestor do it Signed-off-by: Andy Green <andy.green@linaro.org>
2013-02-10 21:21:24 +08:00
return n;
solve flowcontrol problems Problems with rx flow control implementation were the underlying cause of the connection stalling issue that was covered up with the udelay() patch that was removed recently. This get rx flow control working properly and corrects problems with fifo management in the test server mirror protocol code too. The rxfow control api has been changed to just set a flag, so it's very cheap to call from user code. After the callbacks that might use the rxflow control api the flag is checked and any pending actions done. rx flow control now stops any rx packet coming immediately, with compessed connections "just what was left in the pipe" might be hundreds of KBytes. To implement that the current packet being decoded is copied into a malloc'd buffer by the rx processing code now. When rxflow is allows to come again, the buffer is drained and freed before any new packet content is accepted. Signed-off-by: Andy Green <andy.green@linaro.org>
2013-01-17 16:50:35 +08:00
}
esp8266 initial support
2016-07-23 14:18:25 +08:00
#if defined(LWS_WITH_ESP8266)
#undef strchr
#define strchr ets_strchr
#endif
manual proxy forcing api add function to manually setup proxy. Useful on iOS where getenv doesn't return proxy settings Simplified by AG Signed-off-by: shys <shyswork@zoho.com> Signed-off-by: Andy Green <andy@warmcat.com>
2013-10-24 22:12:03 +08:00
LWS_VISIBLE int
introduce vhosts This patch splits out some lws_context members into a new lws_vhost struct. - ssl state and options per vhost - SSL_CTX for serving and client per vhost - protocols[] per vhost - extensions[] per vhost lws_context maintains a linked list of lws_vhosts. The same lws_context_creation_info struct is used to regulate both the context creation and to create vhosts: for backward compatibility if you didn't provide the new LWS_SERVER_OPTION_EXPLICIT_VHOSTS option, then a default vhost is created at context creation time using the same info data as the context itself. If you will have multiple vhosts though, you should give the LWS_SERVER_OPTION_EXPLICIT_VHOSTS option at context creation time, create the context first and then the vhosts afterwards using lws_create_vhost(contest, &info); Although there is a lot of housekeeping to implement this change, there is almost no additional overhead if you don't use multiple vhosts and very little api impact (no changes to test apps). Signed-off-by: Andy Green <andy@warmcat.com>
2016-03-28 10:10:43 +08:00
lws_set_proxy(struct lws_vhost *vhost, const char *proxy)
manual proxy forcing api add function to manually setup proxy. Useful on iOS where getenv doesn't return proxy settings Simplified by AG Signed-off-by: shys <shyswork@zoho.com> Signed-off-by: Andy Green <andy@warmcat.com>
2013-10-24 22:12:03 +08:00
{
esp8266 initial support
2016-07-23 14:18:25 +08:00
#if !defined(LWS_WITH_ESP8266)
manual proxy forcing api add function to manually setup proxy. Useful on iOS where getenv doesn't return proxy settings Simplified by AG Signed-off-by: shys <shyswork@zoho.com> Signed-off-by: Andy Green <andy@warmcat.com>
2013-10-24 22:12:03 +08:00
char *p;
proxy auth Simplifies proxy code to use the existing libwebsocket_set_proxy. Enables libwebsocket_set_proxy() to parse username:password@ at front of servername in both http_proxy and info->http_proxy_address. If given the base64 version of the credentials are sent in the CONNECT header to the proxy. Port is now taken from info->http_proxy_address server:port syntax, but if a port is given in the now deprecated info->http_proxy_port (ie, is nonzero) then it is allowed to be missed out and the info port used instead for backwards compatibility. Signed-off-by: Andy Green <andy.green@linaro.org>
2015-11-08 10:15:01 +08:00
char authstring[96];
whitespace trailing mass cleanout Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-14 08:52:03 +08:00
manual proxy forcing api add function to manually setup proxy. Useful on iOS where getenv doesn't return proxy settings Simplified by AG Signed-off-by: shys <shyswork@zoho.com> Signed-off-by: Andy Green <andy@warmcat.com>
2013-10-24 22:12:03 +08:00
if (!proxy)
return -1;
http_proxy: deal with redundant protocol leader https://github.com/warmcat/libwebsockets/issues/764
2017-01-26 07:27:11 +08:00
/* we have to deal with a possible redundant leading http:// */
if (!strncmp(proxy, "http://", 7))
proxy += 7;
proxy auth Simplifies proxy code to use the existing libwebsocket_set_proxy. Enables libwebsocket_set_proxy() to parse username:password@ at front of servername in both http_proxy and info->http_proxy_address. If given the base64 version of the credentials are sent in the CONNECT header to the proxy. Port is now taken from info->http_proxy_address server:port syntax, but if a port is given in the now deprecated info->http_proxy_port (ie, is nonzero) then it is allowed to be missed out and the info port used instead for backwards compatibility. Signed-off-by: Andy Green <andy.green@linaro.org>
2015-11-08 10:15:01 +08:00
p = strchr(proxy, '@');
if (p) { /* auth is around */
mbed3 workable plus or minus mbed3 net stack bug https://github.com/ARMmbed/sockets/issues/35 Signed-off-by: Andy Green <andy.green@linaro.org>
2015-11-08 12:10:26 +08:00
if ((unsigned int)(p - proxy) > sizeof(authstring) - 1)
proxy auth Simplifies proxy code to use the existing libwebsocket_set_proxy. Enables libwebsocket_set_proxy() to parse username:password@ at front of servername in both http_proxy and info->http_proxy_address. If given the base64 version of the credentials are sent in the CONNECT header to the proxy. Port is now taken from info->http_proxy_address server:port syntax, but if a port is given in the now deprecated info->http_proxy_port (ie, is nonzero) then it is allowed to be missed out and the info port used instead for backwards compatibility. Signed-off-by: Andy Green <andy.green@linaro.org>
2015-11-08 10:15:01 +08:00
goto auth_too_long;
proxy auth fix Signed-off-by: Andy Green <andy.green@linaro.org>
2015-11-18 19:32:01 +08:00
strncpy(authstring, proxy, p - proxy);
proxy auth Simplifies proxy code to use the existing libwebsocket_set_proxy. Enables libwebsocket_set_proxy() to parse username:password@ at front of servername in both http_proxy and info->http_proxy_address. If given the base64 version of the credentials are sent in the CONNECT header to the proxy. Port is now taken from info->http_proxy_address server:port syntax, but if a port is given in the now deprecated info->http_proxy_port (ie, is nonzero) then it is allowed to be missed out and the info port used instead for backwards compatibility. Signed-off-by: Andy Green <andy.green@linaro.org>
2015-11-08 10:15:01 +08:00
// null termination not needed on input
if (lws_b64_encode_string(authstring, (p - proxy),
introduce vhosts This patch splits out some lws_context members into a new lws_vhost struct. - ssl state and options per vhost - SSL_CTX for serving and client per vhost - protocols[] per vhost - extensions[] per vhost lws_context maintains a linked list of lws_vhosts. The same lws_context_creation_info struct is used to regulate both the context creation and to create vhosts: for backward compatibility if you didn't provide the new LWS_SERVER_OPTION_EXPLICIT_VHOSTS option, then a default vhost is created at context creation time using the same info data as the context itself. If you will have multiple vhosts though, you should give the LWS_SERVER_OPTION_EXPLICIT_VHOSTS option at context creation time, create the context first and then the vhosts afterwards using lws_create_vhost(contest, &info); Although there is a lot of housekeeping to implement this change, there is almost no additional overhead if you don't use multiple vhosts and very little api impact (no changes to test apps). Signed-off-by: Andy Green <andy@warmcat.com>
2016-03-28 10:10:43 +08:00
vhost->proxy_basic_auth_token,
sizeof vhost->proxy_basic_auth_token) < 0)
proxy auth Simplifies proxy code to use the existing libwebsocket_set_proxy. Enables libwebsocket_set_proxy() to parse username:password@ at front of servername in both http_proxy and info->http_proxy_address. If given the base64 version of the credentials are sent in the CONNECT header to the proxy. Port is now taken from info->http_proxy_address server:port syntax, but if a port is given in the now deprecated info->http_proxy_port (ie, is nonzero) then it is allowed to be missed out and the info port used instead for backwards compatibility. Signed-off-by: Andy Green <andy.green@linaro.org>
2015-11-08 10:15:01 +08:00
goto auth_too_long;
esp8266 initial support
2016-07-23 14:18:25 +08:00
lwsl_info(" Proxy auth in use\n");
whitespace trailing mass cleanout Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-14 08:52:03 +08:00
proxy auth Simplifies proxy code to use the existing libwebsocket_set_proxy. Enables libwebsocket_set_proxy() to parse username:password@ at front of servername in both http_proxy and info->http_proxy_address. If given the base64 version of the credentials are sent in the CONNECT header to the proxy. Port is now taken from info->http_proxy_address server:port syntax, but if a port is given in the now deprecated info->http_proxy_port (ie, is nonzero) then it is allowed to be missed out and the info port used instead for backwards compatibility. Signed-off-by: Andy Green <andy.green@linaro.org>
2015-11-08 10:15:01 +08:00
proxy = p + 1;
} else
introduce vhosts This patch splits out some lws_context members into a new lws_vhost struct. - ssl state and options per vhost - SSL_CTX for serving and client per vhost - protocols[] per vhost - extensions[] per vhost lws_context maintains a linked list of lws_vhosts. The same lws_context_creation_info struct is used to regulate both the context creation and to create vhosts: for backward compatibility if you didn't provide the new LWS_SERVER_OPTION_EXPLICIT_VHOSTS option, then a default vhost is created at context creation time using the same info data as the context itself. If you will have multiple vhosts though, you should give the LWS_SERVER_OPTION_EXPLICIT_VHOSTS option at context creation time, create the context first and then the vhosts afterwards using lws_create_vhost(contest, &info); Although there is a lot of housekeeping to implement this change, there is almost no additional overhead if you don't use multiple vhosts and very little api impact (no changes to test apps). Signed-off-by: Andy Green <andy@warmcat.com>
2016-03-28 10:10:43 +08:00
vhost->proxy_basic_auth_token[0] = '\0';
proxy auth Simplifies proxy code to use the existing libwebsocket_set_proxy. Enables libwebsocket_set_proxy() to parse username:password@ at front of servername in both http_proxy and info->http_proxy_address. If given the base64 version of the credentials are sent in the CONNECT header to the proxy. Port is now taken from info->http_proxy_address server:port syntax, but if a port is given in the now deprecated info->http_proxy_port (ie, is nonzero) then it is allowed to be missed out and the info port used instead for backwards compatibility. Signed-off-by: Andy Green <andy.green@linaro.org>
2015-11-08 10:15:01 +08:00
introduce vhosts This patch splits out some lws_context members into a new lws_vhost struct. - ssl state and options per vhost - SSL_CTX for serving and client per vhost - protocols[] per vhost - extensions[] per vhost lws_context maintains a linked list of lws_vhosts. The same lws_context_creation_info struct is used to regulate both the context creation and to create vhosts: for backward compatibility if you didn't provide the new LWS_SERVER_OPTION_EXPLICIT_VHOSTS option, then a default vhost is created at context creation time using the same info data as the context itself. If you will have multiple vhosts though, you should give the LWS_SERVER_OPTION_EXPLICIT_VHOSTS option at context creation time, create the context first and then the vhosts afterwards using lws_create_vhost(contest, &info); Although there is a lot of housekeeping to implement this change, there is almost no additional overhead if you don't use multiple vhosts and very little api impact (no changes to test apps). Signed-off-by: Andy Green <andy@warmcat.com>
2016-03-28 10:10:43 +08:00
strncpy(vhost->http_proxy_address, proxy,
sizeof(vhost->http_proxy_address) - 1);
vhost->http_proxy_address[
sizeof(vhost->http_proxy_address) - 1] = '\0';
proxy auth Simplifies proxy code to use the existing libwebsocket_set_proxy. Enables libwebsocket_set_proxy() to parse username:password@ at front of servername in both http_proxy and info->http_proxy_address. If given the base64 version of the credentials are sent in the CONNECT header to the proxy. Port is now taken from info->http_proxy_address server:port syntax, but if a port is given in the now deprecated info->http_proxy_port (ie, is nonzero) then it is allowed to be missed out and the info port used instead for backwards compatibility. Signed-off-by: Andy Green <andy.green@linaro.org>
2015-11-08 10:15:01 +08:00
introduce vhosts This patch splits out some lws_context members into a new lws_vhost struct. - ssl state and options per vhost - SSL_CTX for serving and client per vhost - protocols[] per vhost - extensions[] per vhost lws_context maintains a linked list of lws_vhosts. The same lws_context_creation_info struct is used to regulate both the context creation and to create vhosts: for backward compatibility if you didn't provide the new LWS_SERVER_OPTION_EXPLICIT_VHOSTS option, then a default vhost is created at context creation time using the same info data as the context itself. If you will have multiple vhosts though, you should give the LWS_SERVER_OPTION_EXPLICIT_VHOSTS option at context creation time, create the context first and then the vhosts afterwards using lws_create_vhost(contest, &info); Although there is a lot of housekeeping to implement this change, there is almost no additional overhead if you don't use multiple vhosts and very little api impact (no changes to test apps). Signed-off-by: Andy Green <andy@warmcat.com>
2016-03-28 10:10:43 +08:00
p = strchr(vhost->http_proxy_address, ':');
if (!p && !vhost->http_proxy_port) {
manual proxy forcing api add function to manually setup proxy. Useful on iOS where getenv doesn't return proxy settings Simplified by AG Signed-off-by: shys <shyswork@zoho.com> Signed-off-by: Andy Green <andy@warmcat.com>
2013-10-24 22:12:03 +08:00
lwsl_err("http_proxy needs to be ads:port\n");
return -1;
proxy auth Simplifies proxy code to use the existing libwebsocket_set_proxy. Enables libwebsocket_set_proxy() to parse username:password@ at front of servername in both http_proxy and info->http_proxy_address. If given the base64 version of the credentials are sent in the CONNECT header to the proxy. Port is now taken from info->http_proxy_address server:port syntax, but if a port is given in the now deprecated info->http_proxy_port (ie, is nonzero) then it is allowed to be missed out and the info port used instead for backwards compatibility. Signed-off-by: Andy Green <andy.green@linaro.org>
2015-11-08 10:15:01 +08:00
} else {
coverity 155649 medium possible write to null pointer Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-06 11:04:05 +08:00
if (p) {
*p = '\0';
introduce vhosts This patch splits out some lws_context members into a new lws_vhost struct. - ssl state and options per vhost - SSL_CTX for serving and client per vhost - protocols[] per vhost - extensions[] per vhost lws_context maintains a linked list of lws_vhosts. The same lws_context_creation_info struct is used to regulate both the context creation and to create vhosts: for backward compatibility if you didn't provide the new LWS_SERVER_OPTION_EXPLICIT_VHOSTS option, then a default vhost is created at context creation time using the same info data as the context itself. If you will have multiple vhosts though, you should give the LWS_SERVER_OPTION_EXPLICIT_VHOSTS option at context creation time, create the context first and then the vhosts afterwards using lws_create_vhost(contest, &info); Although there is a lot of housekeeping to implement this change, there is almost no additional overhead if you don't use multiple vhosts and very little api impact (no changes to test apps). Signed-off-by: Andy Green <andy@warmcat.com>
2016-03-28 10:10:43 +08:00
vhost->http_proxy_port = atoi(p + 1);
coverity 155649 medium possible write to null pointer Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-06 11:04:05 +08:00
}
manual proxy forcing api add function to manually setup proxy. Useful on iOS where getenv doesn't return proxy settings Simplified by AG Signed-off-by: shys <shyswork@zoho.com> Signed-off-by: Andy Green <andy@warmcat.com>
2013-10-24 22:12:03 +08:00
}
proxy auth Simplifies proxy code to use the existing libwebsocket_set_proxy. Enables libwebsocket_set_proxy() to parse username:password@ at front of servername in both http_proxy and info->http_proxy_address. If given the base64 version of the credentials are sent in the CONNECT header to the proxy. Port is now taken from info->http_proxy_address server:port syntax, but if a port is given in the now deprecated info->http_proxy_port (ie, is nonzero) then it is allowed to be missed out and the info port used instead for backwards compatibility. Signed-off-by: Andy Green <andy.green@linaro.org>
2015-11-08 10:15:01 +08:00
esp8266 initial support
2016-07-23 14:18:25 +08:00
lwsl_info(" Proxy %s:%u\n", vhost->http_proxy_address,
introduce vhosts This patch splits out some lws_context members into a new lws_vhost struct. - ssl state and options per vhost - SSL_CTX for serving and client per vhost - protocols[] per vhost - extensions[] per vhost lws_context maintains a linked list of lws_vhosts. The same lws_context_creation_info struct is used to regulate both the context creation and to create vhosts: for backward compatibility if you didn't provide the new LWS_SERVER_OPTION_EXPLICIT_VHOSTS option, then a default vhost is created at context creation time using the same info data as the context itself. If you will have multiple vhosts though, you should give the LWS_SERVER_OPTION_EXPLICIT_VHOSTS option at context creation time, create the context first and then the vhosts afterwards using lws_create_vhost(contest, &info); Although there is a lot of housekeeping to implement this change, there is almost no additional overhead if you don't use multiple vhosts and very little api impact (no changes to test apps). Signed-off-by: Andy Green <andy@warmcat.com>
2016-03-28 10:10:43 +08:00
vhost->http_proxy_port);
manual proxy forcing api add function to manually setup proxy. Useful on iOS where getenv doesn't return proxy settings Simplified by AG Signed-off-by: shys <shyswork@zoho.com> Signed-off-by: Andy Green <andy@warmcat.com>
2013-10-24 22:12:03 +08:00
return 0;
proxy auth Simplifies proxy code to use the existing libwebsocket_set_proxy. Enables libwebsocket_set_proxy() to parse username:password@ at front of servername in both http_proxy and info->http_proxy_address. If given the base64 version of the credentials are sent in the CONNECT header to the proxy. Port is now taken from info->http_proxy_address server:port syntax, but if a port is given in the now deprecated info->http_proxy_port (ie, is nonzero) then it is allowed to be missed out and the info port used instead for backwards compatibility. Signed-off-by: Andy Green <andy.green@linaro.org>
2015-11-08 10:15:01 +08:00
auth_too_long:
lwsl_err("proxy auth too long\n");
esp8266 initial support
2016-07-23 14:18:25 +08:00
#endif
proxy auth Simplifies proxy code to use the existing libwebsocket_set_proxy. Enables libwebsocket_set_proxy() to parse username:password@ at front of servername in both http_proxy and info->http_proxy_address. If given the base64 version of the credentials are sent in the CONNECT header to the proxy. Port is now taken from info->http_proxy_address server:port syntax, but if a port is given in the now deprecated info->http_proxy_port (ie, is nonzero) then it is allowed to be missed out and the info port used instead for backwards compatibility. Signed-off-by: Andy Green <andy.green@linaro.org>
2015-11-08 10:15:01 +08:00
return -1;
manual proxy forcing api add function to manually setup proxy. Useful on iOS where getenv doesn't return proxy settings Simplified by AG Signed-off-by: shys <shyswork@zoho.com> Signed-off-by: Andy Green <andy@warmcat.com>
2013-10-24 22:12:03 +08:00
}
client: added socks5 proxy support AG: - move creation info members to end of struct - add LWS_WITH_SOCKS5 CMake var, defaults to OFF - cast away some warnings about signed / unsigned in strncpy Signed-off-by: Andy Ning <andy.ning@windriver.com>
2017-05-05 11:38:34 -04:00
#if defined(LWS_WITH_SOCKS5)
LWS_VISIBLE int
lws_set_socks(struct lws_vhost *vhost, const char *socks)
{
#if !defined(LWS_WITH_ESP8266)
char *p_at, *p_colon;
char user[96];
char password[96];
if (!socks)
return -1;
vhost->socks_user[0] = '\0';
vhost->socks_password[0] = '\0';
p_at = strchr(socks, '@');
if (p_at) { /* auth is around */
if ((unsigned int)(p_at - socks) > (sizeof(user)
+ sizeof(password) - 2)) {
lwsl_err("Socks auth too long\n");
goto bail;
}
p_colon = strchr(socks, ':');
if (p_colon) {
if ((unsigned int)(p_colon - socks) > (sizeof(user)
- 1) ) {
lwsl_err("Socks user too long\n");
goto bail;
}
if ((unsigned int)(p_at - p_colon) > (sizeof(password)
- 1) ) {
lwsl_err("Socks password too long\n");
goto bail;
}
coverity fixes
2017-09-23 13:27:11 +08:00
strncpy(vhost->socks_user, socks, p_colon - socks);
strncpy(vhost->socks_password, p_colon + 1,
p_at - (p_colon + 1));
client: added socks5 proxy support AG: - move creation info members to end of struct - add LWS_WITH_SOCKS5 CMake var, defaults to OFF - cast away some warnings about signed / unsigned in strncpy Signed-off-by: Andy Ning <andy.ning@windriver.com>
2017-05-05 11:38:34 -04:00
}
lwsl_info(" Socks auth, user: %s, password: %s\n",
vhost->socks_user, vhost->socks_password );
socks = p_at + 1;
}
strncpy(vhost->socks_proxy_address, socks,
sizeof(vhost->socks_proxy_address) - 1);
vhost->socks_proxy_address[sizeof(vhost->socks_proxy_address) - 1]
= '\0';
p_colon = strchr(vhost->socks_proxy_address, ':');
if (!p_colon && !vhost->socks_proxy_port) {
lwsl_err("socks_proxy needs to be address:port\n");
return -1;
} else {
if (p_colon) {
*p_colon = '\0';
vhost->socks_proxy_port = atoi(p_colon + 1);
}
}
lwsl_info(" Socks %s:%u\n", vhost->socks_proxy_address,
vhost->socks_proxy_port);
return 0;
bail:
#endif
return -1;
}
#endif
api rationalization: eliminate all libwebsocket[s]_ prefixes This nukes all the oldstyle prefixes except in the compatibility code. struct libwebsockets becomes struct lws too. The api docs are updated accordingly as are the READMEs that mention those apis. Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-04 11:08:32 +08:00
LWS_VISIBLE const struct lws_protocols *
lws_get_protocol(struct lws *wsi)
fork-sever-process-and-introduce-broadcast-api.patch Signed-off-by: Andy Green <andy@warmcat.com>
2010-12-18 15:13:50 +00:00
{
return wsi->protocol;
}
Hide private library symbols Signed-off-by: Peter Pentchev <roam@ringlet.net>
2013-03-30 09:52:21 +08:00
LWS_VISIBLE int
api rationalization: eliminate all libwebsocket[s]_ prefixes This nukes all the oldstyle prefixes except in the compatibility code. struct libwebsockets becomes struct lws too. The api docs are updated accordingly as are the READMEs that mention those apis. Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-04 11:08:32 +08:00
lws_is_final_fragment(struct lws *wsi)
introduce libwebsocket_is_final_fragment Signed-off-by: Andy Green <andy@warmcat.com>
2011-03-07 21:16:31 +00:00
{
cosmetic fix trailing newline on log
2017-02-28 18:29:57 +08:00
lwsl_info("%s: final %d, rx pk length %ld, draining %ld\n", __func__,
gcc- format strings: debug and extra plugins
2017-02-05 22:07:34 +08:00
wsi->u.ws.final, (long)wsi->u.ws.rx_packet_length,
(long)wsi->u.ws.rx_draining_ext);
clean
2017-09-23 12:55:21 +08:00
return wsi->u.ws.final && !wsi->u.ws.rx_packet_length &&
!wsi->u.ws.rx_draining_ext;
introduce libwebsocket_is_final_fragment Signed-off-by: Andy Green <andy@warmcat.com>
2011-03-07 21:16:31 +00:00
}
Add libwebsockets_ensure_user_space This patch unifies the code for per-connection user allocation, and allows it to be allocated earlier, duiring HTTP service time. It's also OK to attempt allocation multiple times, it will just return any existing allocation. Signed-off-by: Alex Bligh <alex@alex.org.uk>
2011-11-07 17:19:25 +08:00
lws-meta
2017-07-19 04:39:14 +08:00
LWS_VISIBLE int
lws_is_first_fragment(struct lws *wsi)
{
return wsi->u.ws.first_fragment;
}
Hide private library symbols Signed-off-by: Peter Pentchev <roam@ringlet.net>
2013-03-30 09:52:21 +08:00
LWS_VISIBLE unsigned char
api rationalization: eliminate all libwebsocket[s]_ prefixes This nukes all the oldstyle prefixes except in the compatibility code. struct libwebsockets becomes struct lws too. The api docs are updated accordingly as are the READMEs that mention those apis. Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-04 11:08:32 +08:00
lws_get_reserved_bits(struct lws *wsi)
Added support for extensions that only manipulate application data.
2013-01-09 18:06:55 +08:00
{
unionize mutually exclusive wsi members Large chunks of struct libwebsocket members actually have a mutually exclusive lifecycle, eg, once the http headers are finished they sit there unused until the instance is destroyed. This makes a big improvement in memory efficiency by making four categories of member: always needed, needed for header processing, needed for http processing, and needed for ws processing. The last three are mutually exclusive and bound into a union inside the wsi. Care needs taking now at "union transitions", although we zeroed down the struct at init, the other union siblings have been writing the same memory by the time later member siblings start to use it. So it must be cleared down appropriately when we cross from one mutually-exclusive use to another. Signed-off-by: Andy Green <andy.green@linaro.org>
2013-01-21 11:04:23 +08:00
return wsi->u.ws.rsv;
Added support for extensions that only manipulate application data.
2013-01-09 18:06:55 +08:00
}
remove lws_ensure_user_space from public api change return The function has a logical problem when the size of the requested allocation is 0, it will return NULL which is overloaded as failure. Actually the whole function is evil as an api, this patch moves it out of the public API space and fixes it to return 0 for success or 1 for fail. Private code does not need to to return wsi->user_space and public code should only get that from the callback as discussed on trac recently. Thanks to Edwin for debugging the problem. Reported-by: Edwin van den Oetelaar <oetelaar.automatisering@gmail.com> Signed-off-by: Andy Green <andy.green@linaro.org>
2013-02-18 16:30:10 +08:00
int
api rationalization: eliminate all libwebsocket[s]_ prefixes This nukes all the oldstyle prefixes except in the compatibility code. struct libwebsockets becomes struct lws too. The api docs are updated accordingly as are the READMEs that mention those apis. Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-04 11:08:32 +08:00
lws_ensure_user_space(struct lws *wsi)
Add libwebsockets_ensure_user_space This patch unifies the code for per-connection user allocation, and allows it to be allocated earlier, duiring HTTP service time. It's also OK to attempt allocation multiple times, it will just return any existing allocation. Signed-off-by: Alex Bligh <alex@alex.org.uk>
2011-11-07 17:19:25 +08:00
{
robustness protect and document ensure_user_space Signed-off-by: Andy Green <andy.green@linaro.org>
2013-02-15 22:31:55 +08:00
if (!wsi->protocol)
http2: make usable HTTP/2 support is now able to serve the test server, complete with websockets, from a single vhost. - This works the same with both OpenSSL and mbedTLS. - POST is now wired up and works (also for file upload). - CGI is wired up and works. - Redirect is adapted and works - lwsws works. - URI urldecode, sanitation and argument parsing wired up for :path valgrind clean (aside from openssl-style false uninit data usage in mbedtls send occasionally) h2spec reports: $ h2spec -h 127.0.0.1 -p 7681 -t -k -o 1 ... 145 tests, 145 passed, 0 skipped, 0 failed" Incorporates: - "https://github.com/warmcat/libwebsockets/pull/1039 Fixes issue with -Werror=unused-variable flag - https://github.com/sartura/libwebsockets/commit/2c843a1395bdca5c23b40d989f508fc96aa5ba3b ssl: fix infinite loop on client cert verification failure Signed-off-by: Petar Paradzik <petar.paradzik@sartura.hr>" Caused and fixes Coverity 184887 - 184892
2017-10-13 10:33:02 +08:00
return 0;
robustness protect and document ensure_user_space Signed-off-by: Andy Green <andy.green@linaro.org>
2013-02-15 22:31:55 +08:00
Add libwebsockets_ensure_user_space This patch unifies the code for per-connection user allocation, and allows it to be allocated earlier, duiring HTTP service time. It's also OK to attempt allocation multiple times, it will just return any existing allocation. Signed-off-by: Alex Bligh <alex@alex.org.uk>
2011-11-07 17:19:25 +08:00
/* allocate the per-connection user memory (if any) */
if (wsi->protocol->per_session_data_size && !wsi->user_space) {
heap instrumentation
2017-10-04 07:10:39 +08:00
wsi->user_space = lws_zalloc(wsi->protocol->per_session_data_size, "user space");
http2: make usable HTTP/2 support is now able to serve the test server, complete with websockets, from a single vhost. - This works the same with both OpenSSL and mbedTLS. - POST is now wired up and works (also for file upload). - CGI is wired up and works. - Redirect is adapted and works - lwsws works. - URI urldecode, sanitation and argument parsing wired up for :path valgrind clean (aside from openssl-style false uninit data usage in mbedtls send occasionally) h2spec reports: $ h2spec -h 127.0.0.1 -p 7681 -t -k -o 1 ... 145 tests, 145 passed, 0 skipped, 0 failed" Incorporates: - "https://github.com/warmcat/libwebsockets/pull/1039 Fixes issue with -Werror=unused-variable flag - https://github.com/sartura/libwebsockets/commit/2c843a1395bdca5c23b40d989f508fc96aa5ba3b ssl: fix infinite loop on client cert verification failure Signed-off-by: Petar Paradzik <petar.paradzik@sartura.hr>" Caused and fixes Coverity 184887 - 184892
2017-10-13 10:33:02 +08:00
if (wsi->user_space == NULL) {
clean
2017-09-23 12:55:21 +08:00
lwsl_err("%s: OOM\n", __func__);
remove lws_ensure_user_space from public api change return The function has a logical problem when the size of the requested allocation is 0, it will return NULL which is overloaded as failure. Actually the whole function is evil as an api, this patch moves it out of the public API space and fixes it to return 0 for success or 1 for fail. Private code does not need to to return wsi->user_space and public code should only get that from the callback as discussed on trac recently. Thanks to Edwin for debugging the problem. Reported-by: Edwin van den Oetelaar <oetelaar.automatisering@gmail.com> Signed-off-by: Andy Green <andy.green@linaro.org>
2013-02-18 16:30:10 +08:00
return 1;
Add libwebsockets_ensure_user_space This patch unifies the code for per-connection user allocation, and allows it to be allocated earlier, duiring HTTP service time. It's also OK to attempt allocation multiple times, it will just return any existing allocation. Signed-off-by: Alex Bligh <alex@alex.org.uk>
2011-11-07 17:19:25 +08:00
}
http2 alpn npn pollout This adds npn / alpn support if your openssl can handle it. Then, browsers that understand alpn will by default negotiate http/1.1 and work as normal. Clients that understand http2.0 can negotiate h2-14 and use the basic but working http2.0 support automatically Signed-off-by: Andy Green <andy.green@linaro.org>
2014-10-22 15:37:28 +08:00
} else
clean
2017-09-23 12:55:21 +08:00
lwsl_debug("%s: %p protocol pss %lu, user_space=%p\n", __func__,
wsi, (long)wsi->protocol->per_session_data_size,
wsi->user_space);
remove lws_ensure_user_space from public api change return The function has a logical problem when the size of the requested allocation is 0, it will return NULL which is overloaded as failure. Actually the whole function is evil as an api, this patch moves it out of the public API space and fixes it to return 0 for success or 1 for fail. Private code does not need to to return wsi->user_space and public code should only get that from the callback as discussed on trac recently. Thanks to Edwin for debugging the problem. Reported-by: Edwin van den Oetelaar <oetelaar.automatisering@gmail.com> Signed-off-by: Andy Green <andy.green@linaro.org>
2013-02-18 16:30:10 +08:00
return 0;
Add libwebsockets_ensure_user_space This patch unifies the code for per-connection user allocation, and allows it to be allocated earlier, duiring HTTP service time. It's also OK to attempt allocation multiple times, it will just return any existing allocation. Signed-off-by: Alex Bligh <alex@alex.org.uk>
2011-11-07 17:19:25 +08:00
}
introduce logging api and convert all library output to use it - multiple debug context calls lwsl_ err, warn, debug, parser, ext, client - api added to set which contexts output to stderr using a bitfield log_level - --disable-debug on configure removes all code that is not err or warn severity - err and warn contexts always output to stderr unless disabled by log_level - err and warn enabled by default in log_level Signed-off-by: Andy Green <andy@warmcat.com>
2013-01-10 19:50:35 +08:00
Plugins: add ssh-base ssh server plugin
2017-10-16 16:59:57 +08:00
LWS_VISIBLE void *
lws_adjust_protocol_psds(struct lws *wsi, size_t new_size)
{
((struct lws_protocols *)lws_get_protocol(wsi))->per_session_data_size =
new_size;
if (lws_ensure_user_space(wsi))
return NULL;
return wsi->user_space;
}
logging timestamp creation expose as api Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-25 15:01:55 +08:00
LWS_VISIBLE int
lwsl_timestamp(int level, char *p, int len)
logging extend level set api to allow setting emission function Signed-off-by: Andy Green <andy.green@linaro.org>
2013-01-12 09:17:42 +08:00
{
plat-optee and boringssl adaptations
2017-01-17 07:01:02 +08:00
#ifndef LWS_PLAT_OPTEE
lwsl stderr sink use formatted time AG change to use localtime_r for threadsafety and leave old method as fallback
2016-02-21 07:42:49 +08:00
time_t o_now = time(NULL);
Avoid calls to gettimeofday() Add a new function to get the current time in microseconds, since gettimeofday() does not exist on Windows. Keep the current implementation for the test applications.
2014-02-26 21:37:31 +01:00
unsigned long long now;
windows doesnt have localtime_r Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-21 21:36:48 +08:00
struct tm *ptm = NULL;
#ifndef WIN32
struct tm tm;
#endif
int n;
introduce logging api and convert all library output to use it - multiple debug context calls lwsl_ err, warn, debug, parser, ext, client - api added to set which contexts output to stderr using a bitfield log_level - --disable-debug on configure removes all code that is not err or warn severity - err and warn contexts always output to stderr unless disabled by log_level - err and warn enabled by default in log_level Signed-off-by: Andy Green <andy@warmcat.com>
2013-01-10 19:50:35 +08:00
wince minor adaptations https://github.com/warmcat/libwebsockets/issues/444
2016-02-29 18:48:55 +08:00
#ifndef _WIN32_WCE
windows doesnt have localtime_r Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-21 21:36:48 +08:00
#ifdef WIN32
ptm = localtime(&o_now);
#else
if (localtime_r(&o_now, &tm))
ptm = &tm;
wince minor adaptations https://github.com/warmcat/libwebsockets/issues/444
2016-02-29 18:48:55 +08:00
#endif
windows doesnt have localtime_r Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-21 21:36:48 +08:00
#endif
logging timestamp creation expose as api Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-25 15:01:55 +08:00
p[0] = '\0';
cleanups after api changes and mbed update Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-04 16:54:12 +08:00
for (n = 0; n < LLL_COUNT; n++) {
if (level != (1 << n))
continue;
now = time_in_microseconds() / 100;
windows doesnt have localtime_r Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-21 21:36:48 +08:00
if (ptm)
lws_snprintf Thanks to Fabrice Gilot for reporting the problem that led to uncovering this. Due to a misunderstanding of the return value of snprintf (it is not truncated according to the max size passed in) in several places relying on snprintf to truncate the length overflows are possible. This patch wraps snprintf with a new lws_snprintf() which does truncate its length to allow the buffer limiting scheme to work properly. All users should update with these fixes.
2016-09-15 02:22:57 +08:00
n = lws_snprintf(p, len,
logging timestamp creation expose as api Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-25 15:01:55 +08:00
"[%04d/%02d/%02d %02d:%02d:%02d:%04d] %s: ",
windows doesnt have localtime_r Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-21 21:36:48 +08:00
ptm->tm_year + 1900,
lwsl_timestamp month off by one
2016-06-01 08:32:18 +08:00
ptm->tm_mon + 1,
windows doesnt have localtime_r Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-21 21:36:48 +08:00
ptm->tm_mday,
ptm->tm_hour,
ptm->tm_min,
ptm->tm_sec,
lwsl stderr sink use formatted time AG change to use localtime_r for threadsafety and leave old method as fallback
2016-02-21 07:42:49 +08:00
(int)(now % 10000), log_level_names[n]);
else
lws_snprintf Thanks to Fabrice Gilot for reporting the problem that led to uncovering this. Due to a misunderstanding of the return value of snprintf (it is not truncated according to the max size passed in) in several places relying on snprintf to truncate the length overflows are possible. This patch wraps snprintf with a new lws_snprintf() which does truncate its length to allow the buffer limiting scheme to work properly. All users should update with these fixes.
2016-09-15 02:22:57 +08:00
n = lws_snprintf(p, len, "[%llu:%04d] %s: ",
lwsl stderr sink use formatted time AG change to use localtime_r for threadsafety and leave old method as fallback
2016-02-21 07:42:49 +08:00
(unsigned long long) now / 10000,
(int)(now % 10000), log_level_names[n]);
logging timestamp creation expose as api Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-25 15:01:55 +08:00
return n;
cleanups after api changes and mbed update Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-04 16:54:12 +08:00
}
plat-optee and boringssl adaptations
2017-01-17 07:01:02 +08:00
#endif
logging timestamp creation expose as api Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-25 15:01:55 +08:00
return 0;
}
http2: make usable HTTP/2 support is now able to serve the test server, complete with websockets, from a single vhost. - This works the same with both OpenSSL and mbedTLS. - POST is now wired up and works (also for file upload). - CGI is wired up and works. - Redirect is adapted and works - lwsws works. - URI urldecode, sanitation and argument parsing wired up for :path valgrind clean (aside from openssl-style false uninit data usage in mbedtls send occasionally) h2spec reports: $ h2spec -h 127.0.0.1 -p 7681 -t -k -o 1 ... 145 tests, 145 passed, 0 skipped, 0 failed" Incorporates: - "https://github.com/warmcat/libwebsockets/pull/1039 Fixes issue with -Werror=unused-variable flag - https://github.com/sartura/libwebsockets/commit/2c843a1395bdca5c23b40d989f508fc96aa5ba3b ssl: fix infinite loop on client cert verification failure Signed-off-by: Petar Paradzik <petar.paradzik@sartura.hr>" Caused and fixes Coverity 184887 - 184892
2017-10-13 10:33:02 +08:00
static const char * const colours[] = {
"[31;1m", /* LLL_ERR */
"[36;1m", /* LLL_WARN */
"[35;1m", /* LLL_NOTICE */
"[32;1m", /* LLL_INFO */
"[34;1m", /* LLL_DEBUG */
"[33;1m", /* LLL_PARSER */
"[33;1m", /* LLL_HEADER */
"[33;1m", /* LLL_EXT */
"[33;1m", /* LLL_CLIENT */
"[33;1m", /* LLL_LATENCY */
"[30;1m", /* LLL_USER */
};
plat-optee and boringssl adaptations
2017-01-17 07:01:02 +08:00
#ifndef LWS_PLAT_OPTEE
logging timestamp creation expose as api Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-25 15:01:55 +08:00
LWS_VISIBLE void lwsl_emit_stderr(int level, const char *line)
{
esp8266 initial support
2016-07-23 14:18:25 +08:00
#if !defined(LWS_WITH_ESP8266)
logging timestamp creation expose as api Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-25 15:01:55 +08:00
char buf[50];
http2: make usable HTTP/2 support is now able to serve the test server, complete with websockets, from a single vhost. - This works the same with both OpenSSL and mbedTLS. - POST is now wired up and works (also for file upload). - CGI is wired up and works. - Redirect is adapted and works - lwsws works. - URI urldecode, sanitation and argument parsing wired up for :path valgrind clean (aside from openssl-style false uninit data usage in mbedtls send occasionally) h2spec reports: $ h2spec -h 127.0.0.1 -p 7681 -t -k -o 1 ... 145 tests, 145 passed, 0 skipped, 0 failed" Incorporates: - "https://github.com/warmcat/libwebsockets/pull/1039 Fixes issue with -Werror=unused-variable flag - https://github.com/sartura/libwebsockets/commit/2c843a1395bdca5c23b40d989f508fc96aa5ba3b ssl: fix infinite loop on client cert verification failure Signed-off-by: Petar Paradzik <petar.paradzik@sartura.hr>" Caused and fixes Coverity 184887 - 184892
2017-10-13 10:33:02 +08:00
static char tty;
int n, m = ARRAY_SIZE(colours) - 1;
if (!tty)
tty = isatty(2) | 2;
logging timestamp creation expose as api Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-25 15:01:55 +08:00
lwsl_timestamp(level, buf, sizeof(buf));
http2: make usable HTTP/2 support is now able to serve the test server, complete with websockets, from a single vhost. - This works the same with both OpenSSL and mbedTLS. - POST is now wired up and works (also for file upload). - CGI is wired up and works. - Redirect is adapted and works - lwsws works. - URI urldecode, sanitation and argument parsing wired up for :path valgrind clean (aside from openssl-style false uninit data usage in mbedtls send occasionally) h2spec reports: $ h2spec -h 127.0.0.1 -p 7681 -t -k -o 1 ... 145 tests, 145 passed, 0 skipped, 0 failed" Incorporates: - "https://github.com/warmcat/libwebsockets/pull/1039 Fixes issue with -Werror=unused-variable flag - https://github.com/sartura/libwebsockets/commit/2c843a1395bdca5c23b40d989f508fc96aa5ba3b ssl: fix infinite loop on client cert verification failure Signed-off-by: Petar Paradzik <petar.paradzik@sartura.hr>" Caused and fixes Coverity 184887 - 184892
2017-10-13 10:33:02 +08:00
if (tty == 3) {
n = 1 << (ARRAY_SIZE(colours) - 1);
while (n) {
if (level & n)
break;
m--;
n >>= 1;
}
fprintf(stderr, "%c%s%s%s%c[0m", 27, colours[m], buf, line, 27);
} else
fprintf(stderr, "%s%s", buf, line);
esp8266 initial support
2016-07-23 14:18:25 +08:00
#endif
expose log level in emit Signed-off-by: Andy Green <andy.green@linaro.org>
2013-01-19 11:17:56 +08:00
}
plat-optee and boringssl adaptations
2017-01-17 07:01:02 +08:00
#endif
helper api: log through syslog Signed-off-by: Andy Green <andy.green@linaro.org>
2013-01-19 11:12:16 +08:00
Subject: [PATCH] add _lws_logv to make loging accessible to other wrappers. use _lws_logv in _lws_log
2014-12-10 18:50:28 -06:00
LWS_VISIBLE void _lws_logv(int filter, const char *format, va_list vl)
expose log level in emit Signed-off-by: Andy Green <andy.green@linaro.org>
2013-01-19 11:17:56 +08:00
{
esp8266 initial support
2016-07-23 14:18:25 +08:00
#if defined(LWS_WITH_ESP8266)
char buf[128];
#else
expose log level in emit Signed-off-by: Andy Green <andy.green@linaro.org>
2013-01-19 11:17:56 +08:00
char buf[256];
esp8266 initial support
2016-07-23 14:18:25 +08:00
#endif
int n;
expose log level in emit Signed-off-by: Andy Green <andy.green@linaro.org>
2013-01-19 11:17:56 +08:00
if (!(log_level & filter))
return;
introduce logging api and convert all library output to use it - multiple debug context calls lwsl_ err, warn, debug, parser, ext, client - api added to set which contexts output to stderr using a bitfield log_level - --disable-debug on configure removes all code that is not err or warn severity - err and warn contexts always output to stderr unless disabled by log_level - err and warn enabled by default in log_level Signed-off-by: Andy Green <andy@warmcat.com>
2013-01-10 19:50:35 +08:00
esp8266 initial support
2016-07-23 14:18:25 +08:00
n = vsnprintf(buf, sizeof(buf) - 1, format, vl);
(void)n;
#if defined(LWS_WITH_ESP8266)
style cleanup This brings the library sources into compliance with checkpatch style except for three or four exceptions like WIN32 related stuff and one long string constant I don't want to break into multiple sprintf calls. There should be no functional or compilability change from all this (hopefully). Signed-off-by: Andy Green <andy.green@linaro.org>
2013-02-11 17:13:32 +08:00
buf[sizeof(buf) - 1] = '\0';
esp8266 initial support
2016-07-23 14:18:25 +08:00
#else
/* vnsprintf returns what it would have written, even if truncated */
if (n > sizeof(buf) - 1)
n = sizeof(buf) - 1;
if (n > 0)
buf[n] = '\0';
#endif
logging extend level set api to allow setting emission function Signed-off-by: Andy Green <andy.green@linaro.org>
2013-01-12 09:17:42 +08:00
expose log level in emit Signed-off-by: Andy Green <andy.green@linaro.org>
2013-01-19 11:17:56 +08:00
lwsl_emit(filter, buf);
introduce logging api and convert all library output to use it - multiple debug context calls lwsl_ err, warn, debug, parser, ext, client - api added to set which contexts output to stderr using a bitfield log_level - --disable-debug on configure removes all code that is not err or warn severity - err and warn contexts always output to stderr unless disabled by log_level - err and warn enabled by default in log_level Signed-off-by: Andy Green <andy@warmcat.com>
2013-01-10 19:50:35 +08:00
}
Subject: [PATCH] add _lws_logv to make loging accessible to other wrappers. use _lws_logv in _lws_log
2014-12-10 18:50:28 -06:00
LWS_VISIBLE void _lws_log(int filter, const char *format, ...)
{
va_list ap;
va_start(ap, format);
_lws_logv(filter, format, ap);
va_end(ap);
}
clean tidy the worst whitespace alignment probs due to mass token name length changes Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-04 11:30:53 +08:00
LWS_VISIBLE void lws_set_log_level(int level,
cleanups after api changes and mbed update Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-04 16:54:12 +08:00
void (*func)(int level, const char *line))
introduce logging api and convert all library output to use it - multiple debug context calls lwsl_ err, warn, debug, parser, ext, client - api added to set which contexts output to stderr using a bitfield log_level - --disable-debug on configure removes all code that is not err or warn severity - err and warn contexts always output to stderr unless disabled by log_level - err and warn enabled by default in log_level Signed-off-by: Andy Green <andy@warmcat.com>
2013-01-10 19:50:35 +08:00
{
log_level = level;
cleanups after api changes and mbed update Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-04 16:54:12 +08:00
if (func)
lwsl_emit = func;
Disable compile warning: no newline at end of file
2014-04-27 13:28:22 +02:00
}
add lws_is_ssl api Suggested by Michael Harbler on the mailing list Signed-off-by: Andy Green <andy.green@linaro.org>
2014-08-16 09:54:27 +08:00
lwsl_visible
2017-02-09 09:11:17 +08:00
LWS_VISIBLE int lwsl_visible(int level)
{
return log_level & level;
}
refactor: subdirs for source in lib Split out some optional code into own sources to shrink down libwebsockets.c and server.c a bit
2017-10-16 12:52:32 +08:00
LWS_VISIBLE void
lwsl_hexdump_level(int hexdump_level, const void *vbuf, size_t len)
{
unsigned char *buf = (unsigned char *)vbuf;
unsigned int n, m, start;
char line[80];
char *p;
if (!lwsl_visible(hexdump_level))
return;
_lws_log(hexdump_level, "\n");
for (n = 0; n < len;) {
start = n;
p = line;
p += sprintf(p, "%04X: ", start);
for (m = 0; m < 16 && n < len; m++)
p += sprintf(p, "%02X ", buf[n++]);
while (m++ < 16)
p += sprintf(p, " ");
p += sprintf(p, " ");
for (m = 0; m < 16 && (start + m) < len; m++) {
if (buf[start + m] >= ' ' && buf[start + m] < 127)
*p++ = buf[start + m];
else
*p++ = '.';
}
while (m++ < 16)
*p++ = ' ';
*p++ = '\n';
*p = '\0';
_lws_log(hexdump_level, "%s", line);
(void)line;
}
_lws_log(hexdump_level, "\n");
}
LWS_VISIBLE void
lwsl_hexdump(const void *vbuf, size_t len)
{
lwsl_hexdump_level(LLL_DEBUG, vbuf, len);
}
add lws_is_ssl api Suggested by Michael Harbler on the mailing list Signed-off-by: Andy Green <andy.green@linaro.org>
2014-08-16 09:54:27 +08:00
LWS_VISIBLE int
api rationalization: eliminate all libwebsocket[s]_ prefixes This nukes all the oldstyle prefixes except in the compatibility code. struct libwebsockets becomes struct lws too. The api docs are updated accordingly as are the READMEs that mention those apis. Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-04 11:08:32 +08:00
lws_is_ssl(struct lws *wsi)
add lws_is_ssl api Suggested by Michael Harbler on the mailing list Signed-off-by: Andy Green <andy.green@linaro.org>
2014-08-16 09:54:27 +08:00
{
fix build without ssl support Signed-off-by: Andy Green <andy.green@linaro.org>
2014-08-19 08:41:26 +08:00
#ifdef LWS_OPENSSL_SUPPORT
add lws_is_ssl api Suggested by Michael Harbler on the mailing list Signed-off-by: Andy Green <andy.green@linaro.org>
2014-08-16 09:54:27 +08:00
return wsi->use_ssl;
fix build without ssl support Signed-off-by: Andy Green <andy.green@linaro.org>
2014-08-19 08:41:26 +08:00
#else
mbed3 warning cleaning Signed-off-by: Andy Green <andy.green@linaro.org>
2015-11-02 13:10:33 +08:00
(void)wsi;
fix build without ssl support Signed-off-by: Andy Green <andy.green@linaro.org>
2014-08-19 08:41:26 +08:00
return 0;
#endif
add lws_is_ssl api Suggested by Michael Harbler on the mailing list Signed-off-by: Andy Green <andy.green@linaro.org>
2014-08-16 09:54:27 +08:00
}
introduce lws_partial_buffered Signed-off-by: Andy Green <andy.green@linaro.org>
2014-08-18 22:49:39 +08:00
ssl expose public wsi->ssl getter Signed-off-by: Denis Osvald <denis.osvald@sartura.hr>
2017-01-23 19:34:46 +08:00
#ifdef LWS_OPENSSL_SUPPORT
LWS_VISIBLE SSL*
lws_get_ssl(struct lws *wsi)
{
return wsi->ssl;
}
#endif
introduce lws_partial_buffered Signed-off-by: Andy Green <andy.green@linaro.org>
2014-08-18 22:49:39 +08:00
LWS_VISIBLE int
api rationalization: eliminate all libwebsocket[s]_ prefixes This nukes all the oldstyle prefixes except in the compatibility code. struct libwebsockets becomes struct lws too. The api docs are updated accordingly as are the READMEs that mention those apis. Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-04 11:08:32 +08:00
lws_partial_buffered(struct lws *wsi)
introduce lws_partial_buffered Signed-off-by: Andy Green <andy.green@linaro.org>
2014-08-18 22:49:39 +08:00
{
clean internal refactor - Mainly symbol length reduction - Whitespace clean - Code refactor for linear flow - Audit @Context for API docs vs changes Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-17 17:03:59 +08:00
return !!wsi->trunc_len;
introduce lws_partial_buffered Signed-off-by: Andy Green <andy.green@linaro.org>
2014-08-18 22:49:39 +08:00
}
http2 can keep upgraded connection up Signed-off-by: Andy Green <andy.green@linaro.org>
2014-10-08 12:00:53 +08:00
http2 window update Signed-off-by: Andy Green <andy.green@linaro.org>
2014-10-29 09:39:08 +08:00
LWS_VISIBLE size_t
api rationalization: eliminate all libwebsocket[s]_ prefixes This nukes all the oldstyle prefixes except in the compatibility code. struct libwebsockets becomes struct lws too. The api docs are updated accordingly as are the READMEs that mention those apis. Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-04 11:08:32 +08:00
lws_get_peer_write_allowance(struct lws *wsi)
http2 window update Signed-off-by: Andy Green <andy.green@linaro.org>
2014-10-29 09:39:08 +08:00
{
CMake: convert all LWS_USE_... to LWS_WITH_... Almost all the CMake options begin with LWS_WITH_..., but many of the symbols passed to lws are LWS_USE_... , this causes neededless confusion, compounded by the fact that a few CMake options also begin with LWS_USE_. This patch globally converts all LWS_USE_... to LWS_WITH_..., so there is only one prefix to remember in both CMake and the code. The affected public CMake options are LWS_USE_BORINGSSL -> LWS_WITH_BORINGSSL LWS_USE_CYASSL -> LWS_WITH_CYASSL LWS_USE_WOLFSSL -> LWS_WITH_WOLFSSL LWS_USE_MBEDTLS -> LWS_WITH_MBEDTLS LWS_USE_BUNDLED_ZLIB -> LWS_WITH_BUNDLED_ZLIB
2017-09-28 11:29:03 +08:00
#ifdef LWS_WITH_HTTP2
http2 window update Signed-off-by: Andy Green <andy.green@linaro.org>
2014-10-29 09:39:08 +08:00
/* only if we are using HTTP2 on this connection */
clean internal refactor - Mainly symbol length reduction - Whitespace clean - Code refactor for linear flow - Audit @Context for API docs vs changes Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-17 17:03:59 +08:00
if (wsi->mode != LWSCM_HTTP2_SERVING)
http2 window update Signed-off-by: Andy Green <andy.green@linaro.org>
2014-10-29 09:39:08 +08:00
return -1;
whitespace trailing mass cleanout Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-14 08:52:03 +08:00
http2: make usable HTTP/2 support is now able to serve the test server, complete with websockets, from a single vhost. - This works the same with both OpenSSL and mbedTLS. - POST is now wired up and works (also for file upload). - CGI is wired up and works. - Redirect is adapted and works - lwsws works. - URI urldecode, sanitation and argument parsing wired up for :path valgrind clean (aside from openssl-style false uninit data usage in mbedtls send occasionally) h2spec reports: $ h2spec -h 127.0.0.1 -p 7681 -t -k -o 1 ... 145 tests, 145 passed, 0 skipped, 0 failed" Incorporates: - "https://github.com/warmcat/libwebsockets/pull/1039 Fixes issue with -Werror=unused-variable flag - https://github.com/sartura/libwebsockets/commit/2c843a1395bdca5c23b40d989f508fc96aa5ba3b ssl: fix infinite loop on client cert verification failure Signed-off-by: Petar Paradzik <petar.paradzik@sartura.hr>" Caused and fixes Coverity 184887 - 184892
2017-10-13 10:33:02 +08:00
return lws_h2_tx_cr_get(wsi);
http2 window update Signed-off-by: Andy Green <andy.green@linaro.org>
2014-10-29 09:39:08 +08:00
#else
mbed3 warning cleaning Signed-off-by: Andy Green <andy.green@linaro.org>
2015-11-02 13:10:33 +08:00
(void)wsi;
http2 window update Signed-off-by: Andy Green <andy.green@linaro.org>
2014-10-29 09:39:08 +08:00
return -1;
#endif
}
lws_union_transition Signed-off-by: Andy Green <andy.green@linaro.org>
2014-11-08 11:18:47 +08:00
LWS_VISIBLE void
api rationalization: eliminate all libwebsocket[s]_ prefixes This nukes all the oldstyle prefixes except in the compatibility code. struct libwebsockets becomes struct lws too. The api docs are updated accordingly as are the READMEs that mention those apis. Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-04 11:08:32 +08:00
lws_union_transition(struct lws *wsi, enum connection_mode mode)
lws_union_transition Signed-off-by: Andy Green <andy.green@linaro.org>
2014-11-08 11:18:47 +08:00
{
http header malloc pool implement pool Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-25 12:44:12 +08:00
lwsl_debug("%s: %p: mode %d\n", __func__, wsi, mode);
lws_union_transition Signed-off-by: Andy Green <andy.green@linaro.org>
2014-11-08 11:18:47 +08:00
memset(&wsi->u, 0, sizeof(wsi->u));
wsi->mode = mode;
}
api rationalization add cmake switch to export old api wrappers This is off by default, use -D LWS_WITH_OLD_API_WRAPPERS=1 on cmake to get the old api names exported from the library as wrappers around the new api names. This allows the library to continue to be compatible with apps that are not rebuilt with the new libwebsockets.h api compatibility defines. Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-04 10:39:23 +08:00
lws_plat_fd implement platform default handlers This is a rewrite of the patch from Soapyman here https://github.com/warmcat/libwebsockets/pull/363 The main changes compared to Soapyman's original patch are - There's no new stuff in the info struct user code does any overrides it may want to do explicitly after lws_context_create returns - User overrides for file ops can call through (subclass) to the original platform implementation using lws_get_fops_plat() - A typedef is provided for plat-specific fd type - Public helpers are provided to allow user code to be platform-independent about file access, using the lws platform file operations underneath: static inline lws_filefd_type lws_plat_file_open(struct lws_plat_file_ops *fops, const char *filename, unsigned long *filelen, int flags) static inline int lws_plat_file_close(struct lws_plat_file_ops *fops, lws_filefd_type fd) static inline unsigned long lws_plat_file_seek_cur(struct lws_plat_file_ops *fops, lws_filefd_type fd, long offset_from_cur_pos) static inline int lws_plat_file_read(struct lws_plat_file_ops *fops, lws_filefd_type fd, unsigned long *amount, unsigned char *buf, unsigned long len) static inline int lws_plat_file_write(struct lws_plat_file_ops *fops, lws_filefd_type fd, unsigned long *amount, unsigned char *buf, unsigned long len) There's example documentation and implementation in the test server. Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-10 07:58:58 +08:00
LWS_VISIBLE struct lws_plat_file_ops *
lws_get_fops(struct lws_context *context)
{
fops-zip
2017-03-03 12:38:10 +08:00
return (struct lws_plat_file_ops *)context->fops;
lws_plat_fd implement platform default handlers This is a rewrite of the patch from Soapyman here https://github.com/warmcat/libwebsockets/pull/363 The main changes compared to Soapyman's original patch are - There's no new stuff in the info struct user code does any overrides it may want to do explicitly after lws_context_create returns - User overrides for file ops can call through (subclass) to the original platform implementation using lws_get_fops_plat() - A typedef is provided for plat-specific fd type - Public helpers are provided to allow user code to be platform-independent about file access, using the lws platform file operations underneath: static inline lws_filefd_type lws_plat_file_open(struct lws_plat_file_ops *fops, const char *filename, unsigned long *filelen, int flags) static inline int lws_plat_file_close(struct lws_plat_file_ops *fops, lws_filefd_type fd) static inline unsigned long lws_plat_file_seek_cur(struct lws_plat_file_ops *fops, lws_filefd_type fd, long offset_from_cur_pos) static inline int lws_plat_file_read(struct lws_plat_file_ops *fops, lws_filefd_type fd, unsigned long *amount, unsigned char *buf, unsigned long len) static inline int lws_plat_file_write(struct lws_plat_file_ops *fops, lws_filefd_type fd, unsigned long *amount, unsigned char *buf, unsigned long len) There's example documentation and implementation in the test server. Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-10 07:58:58 +08:00
}
api rationalization add cmake switch to export old api wrappers This is off by default, use -D LWS_WITH_OLD_API_WRAPPERS=1 on cmake to get the old api names exported from the library as wrappers around the new api names. This allows the library to continue to be compatible with apps that are not rebuilt with the new libwebsockets.h api compatibility defines. Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-04 10:39:23 +08:00
lws_get_ctx conversion Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-11 09:36:14 +08:00
LWS_VISIBLE LWS_EXTERN struct lws_context *
lws_get_context not _ctx Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-17 18:25:25 +08:00
lws_get_context(const struct lws *wsi)
lws_get_ctx conversion Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-11 09:36:14 +08:00
{
return wsi->context;
}
multithreaded service This adds support for multithreaded service to lws without adding any threading or locking code in the library. At context creation time you can request split the service part of the context into n service domains, which are load-balanced so that the most idle one gets the next listen socket accept. There's a single listen socket on one port still. User code may then spawn n threads doing n service loops / poll()s simultaneously. Locking is only required (I think) in the existing FD lock callbacks already handled by the pthreads server example, and that locking takes place in user code. So the library remains completely agnostic about the threading / locking scheme. And by default, it's completely compatible with one service thread so no changes are required by people uninterested in multithreaded service. However for people interested in extremely lightweight mass http[s]/ ws[s] service with minimum provisioning, the library can now do everything out of the box. To test it, just try $ libwebsockets-test-server-pthreads -j 8 where -j controls the number of service threads Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-19 03:34:24 +08:00
LWS_VISIBLE LWS_EXTERN int
lws_get_count_threads(struct lws_context *context)
{
return context->count_threads;
}
introduce lws_wsi_user Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-14 07:16:32 +08:00
LWS_VISIBLE LWS_EXTERN void *
lws_wsi_user(struct lws *wsi)
{
return wsi->user_space;
}
close add api to control sent close frame contents This adds an api lws_close_reason() which lets you control what will be sent in the close frame when the connection is closed by returning nonzero from the user callback. The test server demo is extended to prove it works in both directions. With this, we should have nice close support. https://github.com/warmcat/libwebsockets/issues/196 Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-26 17:20:34 +08:00
client: allow change externally-defined wsi user_data https://github.com/warmcat/libwebsockets/issues/819 AG: add explanation in doxygen comment and check user_space was externally set
2017-02-28 21:17:25 +08:00
LWS_VISIBLE LWS_EXTERN void
lws_set_wsi_user(struct lws *wsi, void *data)
{
if (wsi->user_space_externally_allocated)
wsi->user_space = data;
else
lwsl_err("%s: Cannot set internally-allocated user_space\n",
__func__);
}
add explicit parent child wsi relationships wsi can have a full tree relationship with each other using linked lists. closing the parent ensures the children are closed first. Convert cgi to use this instead of his cgi-specific sub-wsi management. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-03-02 09:17:22 +08:00
LWS_VISIBLE LWS_EXTERN struct lws *
lws_get_parent(const struct lws *wsi)
{
return wsi->parent;
}
LWS_VISIBLE LWS_EXTERN struct lws *
lws_get_child(const struct lws *wsi)
{
return wsi->child_list;
}
lws-meta
2017-07-19 04:39:14 +08:00
LWS_VISIBLE LWS_EXTERN void
lws_set_parent_carries_io(struct lws *wsi)
{
wsi->parent_carries_io = 1;
}
LWS_VISIBLE LWS_EXTERN void *
lws_get_opaque_parent_data(const struct lws *wsi)
{
return wsi->opaque_parent_data;
}
LWS_VISIBLE LWS_EXTERN void
lws_set_opaque_parent_data(struct lws *wsi, void *data)
{
wsi->opaque_parent_data = data;
}
LWS_VISIBLE LWS_EXTERN int
lws_get_child_pending_on_writable(const struct lws *wsi)
{
return wsi->parent_pending_cb_on_writable;
}
LWS_VISIBLE LWS_EXTERN void
lws_clear_child_pending_on_writable(struct lws *wsi)
{
wsi->parent_pending_cb_on_writable = 0;
}
LWS_VISIBLE LWS_EXTERN int
lws_get_close_length(struct lws *wsi)
{
return wsi->u.ws.close_in_ping_buffer_len;
}
LWS_VISIBLE LWS_EXTERN unsigned char *
lws_get_close_payload(struct lws *wsi)
{
return &wsi->u.ws.ping_payload_buf[LWS_PRE];
}
close add api to control sent close frame contents This adds an api lws_close_reason() which lets you control what will be sent in the close frame when the connection is closed by returning nonzero from the user callback. The test server demo is extended to prove it works in both directions. With this, we should have nice close support. https://github.com/warmcat/libwebsockets/issues/196 Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-26 17:20:34 +08:00
LWS_VISIBLE LWS_EXTERN void
lws_close_reason(struct lws *wsi, enum lws_close_status status,
unsigned char *buf, size_t len)
{
unsigned char *p, *start;
multithreaded service This adds support for multithreaded service to lws without adding any threading or locking code in the library. At context creation time you can request split the service part of the context into n service domains, which are load-balanced so that the most idle one gets the next listen socket accept. There's a single listen socket on one port still. User code may then spawn n threads doing n service loops / poll()s simultaneously. Locking is only required (I think) in the existing FD lock callbacks already handled by the pthreads server example, and that locking takes place in user code. So the library remains completely agnostic about the threading / locking scheme. And by default, it's completely compatible with one service thread so no changes are required by people uninterested in multithreaded service. However for people interested in extremely lightweight mass http[s]/ ws[s] service with minimum provisioning, the library can now do everything out of the box. To test it, just try $ libwebsockets-test-server-pthreads -j 8 where -j controls the number of service threads Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-19 03:34:24 +08:00
int budget = sizeof(wsi->u.ws.ping_payload_buf) - LWS_PRE;
close add api to control sent close frame contents This adds an api lws_close_reason() which lets you control what will be sent in the close frame when the connection is closed by returning nonzero from the user callback. The test server demo is extended to prove it works in both directions. With this, we should have nice close support. https://github.com/warmcat/libwebsockets/issues/196 Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-26 17:20:34 +08:00
assert(wsi->mode == LWSCM_WS_SERVING || wsi->mode == LWSCM_WS_CLIENT);
extension permessage deflate Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-11 11:34:01 +08:00
start = p = &wsi->u.ws.ping_payload_buf[LWS_PRE];
close add api to control sent close frame contents This adds an api lws_close_reason() which lets you control what will be sent in the close frame when the connection is closed by returning nonzero from the user callback. The test server demo is extended to prove it works in both directions. With this, we should have nice close support. https://github.com/warmcat/libwebsockets/issues/196 Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-26 17:20:34 +08:00
*p++ = (((int)status) >> 8) & 0xff;
*p++ = ((int)status) & 0xff;
if (buf)
while (len-- && p < start + budget)
*p++ = *buf++;
wsi->u.ws.close_in_ping_buffer_len = p - start;
}
autobahn add same serverside rxflow cache to client Server side has had immediate RX flow control for quite a while. But client side made do with RX continuing until what had been received was exhausted. For what Autobahn tests, that's not enough. This patch gives clientside RX flow control the same immediate effect as the server side enjoys, re-using the same code. Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-28 14:24:49 +08:00
LWS_EXTERN int
_lws_rx_flow_control(struct lws *wsi)
{
rx flow: handle child state change during parent cb
2017-03-05 15:32:47 +08:00
struct lws *wsic = wsi->child_list;
/* if he has children, do those if they were changed */
while (wsic) {
if (wsic->rxflow_change_to & LWS_RXFLOW_PENDING_CHANGE)
_lws_rx_flow_control(wsic);
wsic = wsic->sibling_list;
}
autobahn add same serverside rxflow cache to client Server side has had immediate RX flow control for quite a while. But client side made do with RX continuing until what had been received was exhausted. For what Autobahn tests, that's not enough. This patch gives clientside RX flow control the same immediate effect as the server side enjoys, re-using the same code. Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-28 14:24:49 +08:00
/* there is no pending change */
clean
2017-09-23 12:55:21 +08:00
if (!(wsi->rxflow_change_to & LWS_RXFLOW_PENDING_CHANGE))
autobahn add same serverside rxflow cache to client Server side has had immediate RX flow control for quite a while. But client side made do with RX continuing until what had been received was exhausted. For what Autobahn tests, that's not enough. This patch gives clientside RX flow control the same immediate effect as the server side enjoys, re-using the same code. Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-28 14:24:49 +08:00
return 0;
/* stuff is still buffered, not ready to really accept new input */
if (wsi->rxflow_buffer) {
/* get ourselves called back to deal with stashed buffer */
lws_callback_on_writable(wsi);
return 0;
}
/* pending is cleared, we can change rxflow state */
wsi->rxflow_change_to &= ~LWS_RXFLOW_PENDING_CHANGE;
lwsl_info("rxflow: wsi %p change_to %d\n", wsi,
wsi->rxflow_change_to & LWS_RXFLOW_ALLOW);
/* adjust the pollfd for this wsi */
if (wsi->rxflow_change_to & LWS_RXFLOW_ALLOW) {
if (lws_change_pollfd(wsi, 0, LWS_POLLIN)) {
lwsl_info("%s: fail\n", __func__);
return -1;
}
} else
if (lws_change_pollfd(wsi, LWS_POLLIN, 0))
return -1;
return 0;
}
utf8 check compatible with extensions Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-30 11:43:36 +08:00
LWS_EXTERN int
lws_check_utf8(unsigned char *state, unsigned char *buf, size_t len)
{
static const unsigned char e0f4[] = {
0xa0 | ((2 - 1) << 2) | 1, /* e0 */
0x80 | ((4 - 1) << 2) | 1, /* e1 */
0x80 | ((4 - 1) << 2) | 1, /* e2 */
0x80 | ((4 - 1) << 2) | 1, /* e3 */
0x80 | ((4 - 1) << 2) | 1, /* e4 */
0x80 | ((4 - 1) << 2) | 1, /* e5 */
0x80 | ((4 - 1) << 2) | 1, /* e6 */
0x80 | ((4 - 1) << 2) | 1, /* e7 */
0x80 | ((4 - 1) << 2) | 1, /* e8 */
0x80 | ((4 - 1) << 2) | 1, /* e9 */
0x80 | ((4 - 1) << 2) | 1, /* ea */
0x80 | ((4 - 1) << 2) | 1, /* eb */
0x80 | ((4 - 1) << 2) | 1, /* ec */
0x80 | ((2 - 1) << 2) | 1, /* ed */
0x80 | ((4 - 1) << 2) | 1, /* ee */
0x80 | ((4 - 1) << 2) | 1, /* ef */
0x90 | ((3 - 1) << 2) | 2, /* f0 */
0x80 | ((4 - 1) << 2) | 2, /* f1 */
0x80 | ((4 - 1) << 2) | 2, /* f2 */
0x80 | ((4 - 1) << 2) | 2, /* f3 */
0x80 | ((1 - 1) << 2) | 2, /* f4 */
0, /* s0 */
0x80 | ((4 - 1) << 2) | 0, /* s2 */
0x80 | ((4 - 1) << 2) | 1, /* s3 */
};
unsigned char s = *state;
while (len--) {
unsigned char c = *buf++;
if (!s) {
if (c >= 0x80) {
if (c < 0xc2 || c > 0xf4)
return 1;
if (c < 0xe0)
s = 0x80 | ((4 - 1) << 2);
else
s = e0f4[c - 0xe0];
}
} else {
if (c < (s & 0xf0) ||
c >= (s & 0xf0) + 0x10 + ((s << 2) & 0x30))
return 1;
s = e0f4[21 + (s & 3)];
}
}
*state = s;
return 0;
}
extension permessage deflate Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-11 11:34:01 +08:00
parse_uri Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-14 13:39:02 +08:00
LWS_VISIBLE LWS_EXTERN int
multithreaded service This adds support for multithreaded service to lws without adding any threading or locking code in the library. At context creation time you can request split the service part of the context into n service domains, which are load-balanced so that the most idle one gets the next listen socket accept. There's a single listen socket on one port still. User code may then spawn n threads doing n service loops / poll()s simultaneously. Locking is only required (I think) in the existing FD lock callbacks already handled by the pthreads server example, and that locking takes place in user code. So the library remains completely agnostic about the threading / locking scheme. And by default, it's completely compatible with one service thread so no changes are required by people uninterested in multithreaded service. However for people interested in extremely lightweight mass http[s]/ ws[s] service with minimum provisioning, the library can now do everything out of the box. To test it, just try $ libwebsockets-test-server-pthreads -j 8 where -j controls the number of service threads Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-19 03:34:24 +08:00
lws_parse_uri(char *p, const char **prot, const char **ads, int *port,
const char **path)
parse_uri Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-14 13:39:02 +08:00
{
const char *end;
static const char *slash = "/";
/* cut up the location into address, port and path */
*prot = p;
while (*p && (*p != ':' || p[1] != '/' || p[2] != '/'))
p++;
if (!*p) {
end = p;
p = (char *)*prot;
*prot = end;
} else {
*p = '\0';
p += 3;
}
*ads = p;
if (!strcmp(*prot, "http") || !strcmp(*prot, "ws"))
*port = 80;
else if (!strcmp(*prot, "https") || !strcmp(*prot, "wss"))
*port = 443;
timeout move to doubly linked list In the case we have a lot of connections, checking them all for timeout state once a second becomes burdensome. At the moment if you have 100K connections, once a second they all get checked for timeout in a loop. This patch adds a doubly-linked list based in the context to each wsi, and only wsi with pending timeouts appear on it. At checking time, we traverse the list, which costs nothing if empty because nobody has a pending timeout. Similarly adding and removing from the list costs almost nothing since no iteration is required no matter how big the list. The extra 8 or 16 bytes in the wsi are offset a little bit by demoting .pps from int to char (save 3 bytes). And trim max act exts to 2, since we only provide one, saving 8 /16 bytes by itself if exts enabled. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-19 04:32:14 +08:00
url-parser: handle ipv6 [] addresses
2017-02-21 07:27:07 +08:00
if (*p == '[')
{
++(*ads);
while (*p && *p != ']')
p++;
if (*p)
*p++ = '\0';
}
else
{
while (*p && *p != ':' && *p != '/')
p++;
}
parse_uri Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-14 13:39:02 +08:00
if (*p == ':') {
*p++ = '\0';
*port = atoi(p);
while (*p && *p != '/')
p++;
}
*path = slash;
if (*p) {
*p++ = '\0';
if (*p)
*path = p;
}
return 0;
}
extension permessage deflate Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-11 11:34:01 +08:00
#ifdef LWS_NO_EXTENSIONS
/* we need to provide dummy callbacks for internal exts
* so user code runs when faced with a lib compiled with
* extensions disabled.
*/
int
lws_extension_callback_pm_deflate(struct lws_context *context,
const struct lws_extension *ext,
struct lws *wsi,
enum lws_extension_callback_reasons reason,
void *user, void *in, size_t len)
{
mbed align with pt changes Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-20 17:35:18 +08:00
(void)context;
(void)ext;
(void)wsi;
(void)reason;
(void)user;
(void)in;
(void)len;
extension permessage deflate Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-11 11:34:01 +08:00
return 0;
}
#endif
socket interface bind generalize to lws_socket_bind Move the socket bind to interface code out of server into libwebsockets.c and make a private api for it. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-03-12 08:18:58 +08:00
LWS_EXTERN int
lws_socket_bind: use lws_sockfd_type
2016-11-16 08:59:47 +08:00
lws_socket_bind(struct lws_vhost *vhost, lws_sockfd_type sockfd, int port,
socket interface bind generalize to lws_socket_bind Move the socket bind to interface code out of server into libwebsockets.c and make a private api for it. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-03-12 08:18:58 +08:00
const char *iface)
{
#if LWS_POSIX
CMake: convert all LWS_USE_... to LWS_WITH_... Almost all the CMake options begin with LWS_WITH_..., but many of the symbols passed to lws are LWS_USE_... , this causes neededless confusion, compounded by the fact that a few CMake options also begin with LWS_USE_. This patch globally converts all LWS_USE_... to LWS_WITH_..., so there is only one prefix to remember in both CMake and the code. The affected public CMake options are LWS_USE_BORINGSSL -> LWS_WITH_BORINGSSL LWS_USE_CYASSL -> LWS_WITH_CYASSL LWS_USE_WOLFSSL -> LWS_WITH_WOLFSSL LWS_USE_MBEDTLS -> LWS_WITH_MBEDTLS LWS_USE_BUNDLED_ZLIB -> LWS_WITH_BUNDLED_ZLIB
2017-09-28 11:29:03 +08:00
#ifdef LWS_WITH_UNIX_SOCK
Add unix domain socket
2016-03-30 22:47:02 -07:00
struct sockaddr_un serv_unix;
#endif
CMake: convert all LWS_USE_... to LWS_WITH_... Almost all the CMake options begin with LWS_WITH_..., but many of the symbols passed to lws are LWS_USE_... , this causes neededless confusion, compounded by the fact that a few CMake options also begin with LWS_USE_. This patch globally converts all LWS_USE_... to LWS_WITH_..., so there is only one prefix to remember in both CMake and the code. The affected public CMake options are LWS_USE_BORINGSSL -> LWS_WITH_BORINGSSL LWS_USE_CYASSL -> LWS_WITH_CYASSL LWS_USE_WOLFSSL -> LWS_WITH_WOLFSSL LWS_USE_MBEDTLS -> LWS_WITH_MBEDTLS LWS_USE_BUNDLED_ZLIB -> LWS_WITH_BUNDLED_ZLIB
2017-09-28 11:29:03 +08:00
#ifdef LWS_WITH_IPV6
socket interface bind generalize to lws_socket_bind Move the socket bind to interface code out of server into libwebsockets.c and make a private api for it. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-03-12 08:18:58 +08:00
struct sockaddr_in6 serv_addr6;
#endif
struct sockaddr_in serv_addr4;
plat-optee and boringssl adaptations
2017-01-17 07:01:02 +08:00
#ifndef LWS_PLAT_OPTEE
Subject: Support to IPv6 on Windows
2017-05-30 18:31:06 -03:00
socklen_t len = sizeof(struct sockaddr_storage);
plat-optee and boringssl adaptations
2017-01-17 07:01:02 +08:00
#endif
socket interface bind generalize to lws_socket_bind Move the socket bind to interface code out of server into libwebsockets.c and make a private api for it. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-03-12 08:18:58 +08:00
int n;
Subject: Support to IPv6 on Windows
2017-05-30 18:31:06 -03:00
struct sockaddr_storage sin;
socket interface bind generalize to lws_socket_bind Move the socket bind to interface code out of server into libwebsockets.c and make a private api for it. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-03-12 08:18:58 +08:00
struct sockaddr *v;
CMake: convert all LWS_USE_... to LWS_WITH_... Almost all the CMake options begin with LWS_WITH_..., but many of the symbols passed to lws are LWS_USE_... , this causes neededless confusion, compounded by the fact that a few CMake options also begin with LWS_USE_. This patch globally converts all LWS_USE_... to LWS_WITH_..., so there is only one prefix to remember in both CMake and the code. The affected public CMake options are LWS_USE_BORINGSSL -> LWS_WITH_BORINGSSL LWS_USE_CYASSL -> LWS_WITH_CYASSL LWS_USE_WOLFSSL -> LWS_WITH_WOLFSSL LWS_USE_MBEDTLS -> LWS_WITH_MBEDTLS LWS_USE_BUNDLED_ZLIB -> LWS_WITH_BUNDLED_ZLIB
2017-09-28 11:29:03 +08:00
#ifdef LWS_WITH_UNIX_SOCK
unix sockets make vhost option Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-14 12:11:51 +08:00
if (LWS_UNIX_SOCK_ENABLED(vhost)) {
Add unix domain socket
2016-03-30 22:47:02 -07:00
v = (struct sockaddr *)&serv_unix;
n = sizeof(struct sockaddr_un);
bzero((char *) &serv_unix, sizeof(serv_unix));
serv_unix.sun_family = AF_UNIX;
if (sizeof(serv_unix.sun_path) <= strlen(iface)) {
lwsl_err("\"%s\" too long for UNIX domain socket\n",
iface);
return -1;
}
strcpy(serv_unix.sun_path, iface);
Support for abstract socket
2016-04-10 13:19:16 +09:00
if (serv_unix.sun_path[0] == '@')
serv_unix.sun_path[0] = '\0';
Add unix domain socket
2016-03-30 22:47:02 -07:00
} else
#endif
CMake: convert all LWS_USE_... to LWS_WITH_... Almost all the CMake options begin with LWS_WITH_..., but many of the symbols passed to lws are LWS_USE_... , this causes neededless confusion, compounded by the fact that a few CMake options also begin with LWS_USE_. This patch globally converts all LWS_USE_... to LWS_WITH_..., so there is only one prefix to remember in both CMake and the code. The affected public CMake options are LWS_USE_BORINGSSL -> LWS_WITH_BORINGSSL LWS_USE_CYASSL -> LWS_WITH_CYASSL LWS_USE_WOLFSSL -> LWS_WITH_WOLFSSL LWS_USE_MBEDTLS -> LWS_WITH_MBEDTLS LWS_USE_BUNDLED_ZLIB -> LWS_WITH_BUNDLED_ZLIB
2017-09-28 11:29:03 +08:00
#if defined(LWS_WITH_IPV6) && !defined(LWS_WITH_ESP32)
ipv6 move disable to vhost option Server ipv6 support disable is now controlled by vhost->options rather than context->options, allowing it to be set per-vhost. Signed-off-by: Andy Green <andy@warmcat.com>
2016-06-03 21:19:40 +08:00
if (LWS_IPV6_ENABLED(vhost)) {
socket interface bind generalize to lws_socket_bind Move the socket bind to interface code out of server into libwebsockets.c and make a private api for it. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-03-12 08:18:58 +08:00
v = (struct sockaddr *)&serv_addr6;
n = sizeof(struct sockaddr_in6);
bzero((char *) &serv_addr6, sizeof(serv_addr6));
ipv6-allow-binding-to-ipv6-address-in-iface ipv4 and ipv6 binding to a named interface works OK. ipv4 binding to an IP also works, but we need some extra ipv6 magic to identify the ipv6 interface from an ipv6 address. This patch based on code from "user3546716" at http://stackoverflow.com/questions/13504934/binding-sockets-to-ipv6-addresses adds the necessary magic. https://github.com/warmcat/libwebsockets/issues/717
2016-12-15 09:58:20 +08:00
if (iface) {
Subject: windows: support to bind to a specific IPv6 address
2017-06-05 13:59:22 -03:00
if (interface_to_sa(vhost, iface,
(struct sockaddr_in *)v, n) < 0) {
clean
2017-09-23 12:55:21 +08:00
lwsl_err("Unable to find if %s\n", iface);
Subject: windows: support to bind to a specific IPv6 address
2017-06-05 13:59:22 -03:00
return -1;
ipv6-allow-binding-to-ipv6-address-in-iface ipv4 and ipv6 binding to a named interface works OK. ipv4 binding to an IP also works, but we need some extra ipv6 magic to identify the ipv6 interface from an ipv6 address. This patch based on code from "user3546716" at http://stackoverflow.com/questions/13504934/binding-sockets-to-ipv6-addresses adds the necessary magic. https://github.com/warmcat/libwebsockets/issues/717
2016-12-15 09:58:20 +08:00
}
Subject: windows: support to bind to a specific IPv6 address
2017-06-05 13:59:22 -03:00
serv_addr6.sin6_scope_id = lws_get_addr_scope(iface);
ipv6-allow-binding-to-ipv6-address-in-iface ipv4 and ipv6 binding to a named interface works OK. ipv4 binding to an IP also works, but we need some extra ipv6 magic to identify the ipv6 interface from an ipv6 address. This patch based on code from "user3546716" at http://stackoverflow.com/questions/13504934/binding-sockets-to-ipv6-addresses adds the necessary magic. https://github.com/warmcat/libwebsockets/issues/717
2016-12-15 09:58:20 +08:00
}
socket interface bind generalize to lws_socket_bind Move the socket bind to interface code out of server into libwebsockets.c and make a private api for it. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-03-12 08:18:58 +08:00
serv_addr6.sin6_family = AF_INET6;
serv_addr6.sin6_port = htons(port);
} else
#endif
{
v = (struct sockaddr *)&serv_addr4;
n = sizeof(serv_addr4);
bzero((char *) &serv_addr4, sizeof(serv_addr4));
serv_addr4.sin_addr.s_addr = INADDR_ANY;
serv_addr4.sin_family = AF_INET;
ESP32 platform This is enough for all the test app features to work on ESP32 without SSL.
2017-02-18 17:26:40 +08:00
#if !defined(LWS_WITH_ESP32)
socket interface bind generalize to lws_socket_bind Move the socket bind to interface code out of server into libwebsockets.c and make a private api for it. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-03-12 08:18:58 +08:00
if (iface &&
ipv6 move disable to vhost option Server ipv6 support disable is now controlled by vhost->options rather than context->options, allowing it to be set per-vhost. Signed-off-by: Andy Green <andy@warmcat.com>
2016-06-03 21:19:40 +08:00
interface_to_sa(vhost, iface,
socket interface bind generalize to lws_socket_bind Move the socket bind to interface code out of server into libwebsockets.c and make a private api for it. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-03-12 08:18:58 +08:00
(struct sockaddr_in *)v, n) < 0) {
lwsl_err("Unable to find interface %s\n", iface);
return -1;
}
ESP32 platform This is enough for all the test app features to work on ESP32 without SSL.
2017-02-18 17:26:40 +08:00
#endif
socket interface bind generalize to lws_socket_bind Move the socket bind to interface code out of server into libwebsockets.c and make a private api for it. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-03-12 08:18:58 +08:00
serv_addr4.sin_port = htons(port);
} /* ipv4 */
n = bind(sockfd, v, n);
CMake: convert all LWS_USE_... to LWS_WITH_... Almost all the CMake options begin with LWS_WITH_..., but many of the symbols passed to lws are LWS_USE_... , this causes neededless confusion, compounded by the fact that a few CMake options also begin with LWS_USE_. This patch globally converts all LWS_USE_... to LWS_WITH_..., so there is only one prefix to remember in both CMake and the code. The affected public CMake options are LWS_USE_BORINGSSL -> LWS_WITH_BORINGSSL LWS_USE_CYASSL -> LWS_WITH_CYASSL LWS_USE_WOLFSSL -> LWS_WITH_WOLFSSL LWS_USE_MBEDTLS -> LWS_WITH_MBEDTLS LWS_USE_BUNDLED_ZLIB -> LWS_WITH_BUNDLED_ZLIB
2017-09-28 11:29:03 +08:00
#ifdef LWS_WITH_UNIX_SOCK
unix sockets make vhost option Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-14 12:11:51 +08:00
if (n < 0 && LWS_UNIX_SOCK_ENABLED(vhost)) {
Add unix domain socket
2016-03-30 22:47:02 -07:00
lwsl_err("ERROR on binding fd %d to \"%s\" (%d %d)\n",
sockfd, iface, n, LWS_ERRNO);
return -1;
} else
#endif
socket interface bind generalize to lws_socket_bind Move the socket bind to interface code out of server into libwebsockets.c and make a private api for it. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-03-12 08:18:58 +08:00
if (n < 0) {
introduce vhosts This patch splits out some lws_context members into a new lws_vhost struct. - ssl state and options per vhost - SSL_CTX for serving and client per vhost - protocols[] per vhost - extensions[] per vhost lws_context maintains a linked list of lws_vhosts. The same lws_context_creation_info struct is used to regulate both the context creation and to create vhosts: for backward compatibility if you didn't provide the new LWS_SERVER_OPTION_EXPLICIT_VHOSTS option, then a default vhost is created at context creation time using the same info data as the context itself. If you will have multiple vhosts though, you should give the LWS_SERVER_OPTION_EXPLICIT_VHOSTS option at context creation time, create the context first and then the vhosts afterwards using lws_create_vhost(contest, &info); Although there is a lot of housekeeping to implement this change, there is almost no additional overhead if you don't use multiple vhosts and very little api impact (no changes to test apps). Signed-off-by: Andy Green <andy@warmcat.com>
2016-03-28 10:10:43 +08:00
lwsl_err("ERROR on binding fd %d to port %d (%d %d)\n",
sockfd, port, n, LWS_ERRNO);
socket interface bind generalize to lws_socket_bind Move the socket bind to interface code out of server into libwebsockets.c and make a private api for it. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-03-12 08:18:58 +08:00
return -1;
}
plat-optee and boringssl adaptations
2017-01-17 07:01:02 +08:00
#ifndef LWS_PLAT_OPTEE
socket interface bind generalize to lws_socket_bind Move the socket bind to interface code out of server into libwebsockets.c and make a private api for it. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-03-12 08:18:58 +08:00
if (getsockname(sockfd, (struct sockaddr *)&sin, &len) == -1)
lwsl_warn("getsockname: %s\n", strerror(LWS_ERRNO));
else
plat-optee and boringssl adaptations
2017-01-17 07:01:02 +08:00
#endif
CMake: convert all LWS_USE_... to LWS_WITH_... Almost all the CMake options begin with LWS_WITH_..., but many of the symbols passed to lws are LWS_USE_... , this causes neededless confusion, compounded by the fact that a few CMake options also begin with LWS_USE_. This patch globally converts all LWS_USE_... to LWS_WITH_..., so there is only one prefix to remember in both CMake and the code. The affected public CMake options are LWS_USE_BORINGSSL -> LWS_WITH_BORINGSSL LWS_USE_CYASSL -> LWS_WITH_CYASSL LWS_USE_WOLFSSL -> LWS_WITH_WOLFSSL LWS_USE_MBEDTLS -> LWS_WITH_MBEDTLS LWS_USE_BUNDLED_ZLIB -> LWS_WITH_BUNDLED_ZLIB
2017-09-28 11:29:03 +08:00
#if defined(LWS_WITH_IPV6)
Subject: Support to IPv6 on Windows
2017-05-30 18:31:06 -03:00
port = (sin.ss_family == AF_INET6) ?
ntohs(((struct sockaddr_in6 *) &sin)->sin6_port) :
ntohs(((struct sockaddr_in *) &sin)->sin_port);
#else
build: strict-aliasing on old compiler workaround https://github.com/warmcat/libwebsockets/issues/1067
2017-10-24 19:27:38 +08:00
{
struct sockaddr_in sain;
memcpy(&sain, &sin, sizeof(sain));
port = ntohs(sain.sin_port);
}
Subject: Support to IPv6 on Windows
2017-05-30 18:31:06 -03:00
#endif
socket interface bind generalize to lws_socket_bind Move the socket bind to interface code out of server into libwebsockets.c and make a private api for it. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-03-12 08:18:58 +08:00
#endif
return port;
}
CMake: convert all LWS_USE_... to LWS_WITH_... Almost all the CMake options begin with LWS_WITH_..., but many of the symbols passed to lws are LWS_USE_... , this causes neededless confusion, compounded by the fact that a few CMake options also begin with LWS_USE_. This patch globally converts all LWS_USE_... to LWS_WITH_..., so there is only one prefix to remember in both CMake and the code. The affected public CMake options are LWS_USE_BORINGSSL -> LWS_WITH_BORINGSSL LWS_USE_CYASSL -> LWS_WITH_CYASSL LWS_USE_WOLFSSL -> LWS_WITH_WOLFSSL LWS_USE_MBEDTLS -> LWS_WITH_MBEDTLS LWS_USE_BUNDLED_ZLIB -> LWS_WITH_BUNDLED_ZLIB
2017-09-28 11:29:03 +08:00
#if defined(LWS_WITH_IPV6)
Subject: windows: support to bind to a specific IPv6 address
2017-06-05 13:59:22 -03:00
LWS_EXTERN unsigned long
lws_get_addr_scope(const char *ipaddr)
{
unsigned long scope = 0;
#ifndef WIN32
struct ifaddrs *addrs, *addr;
char ip[NI_MAXHOST];
unsigned int i;
getifaddrs(&addrs);
for (addr = addrs; addr; addr = addr->ifa_next) {
if (!addr->ifa_addr ||
addr->ifa_addr->sa_family != AF_INET6)
continue;
getnameinfo(addr->ifa_addr,
sizeof(struct sockaddr_in6),
ip, sizeof(ip),
NULL, 0, NI_NUMERICHOST);
i = 0;
while (ip[i])
if (ip[i++] == '%') {
ip[i - 1] = '\0';
break;
}
if (!strcmp(ip, ipaddr)) {
scope = if_nametoindex(addr->ifa_name);
break;
}
}
freeifaddrs(addrs);
#else
PIP_ADAPTER_ADDRESSES adapter, addrs = NULL;
PIP_ADAPTER_UNICAST_ADDRESS addr;
ULONG size = 0;
DWORD ret;
struct sockaddr_in6 *sockaddr;
char ip[NI_MAXHOST];
unsigned int i;
int found = 0;
for (i = 0; i < 5; i++)
{
ret = GetAdaptersAddresses(AF_INET6, GAA_FLAG_INCLUDE_PREFIX,
NULL, addrs, &size);
if ((ret == NO_ERROR) || (ret == ERROR_NO_DATA)) {
break;
} else if (ret == ERROR_BUFFER_OVERFLOW)
{
if (addrs)
free(addrs);
heap instrumentation
2017-10-04 07:10:39 +08:00
addrs = (IP_ADAPTER_ADDRESSES *)malloc(size);
Subject: windows: support to bind to a specific IPv6 address
2017-06-05 13:59:22 -03:00
} else
{
if (addrs)
{
free(addrs);
addrs = NULL;
}
lwsl_err("Failed to get IPv6 address table (%d)", ret);
break;
}
}
clean
2017-09-23 12:55:21 +08:00
if ((ret == NO_ERROR) && (addrs)) {
Subject: windows: support to bind to a specific IPv6 address
2017-06-05 13:59:22 -03:00
adapter = addrs;
clean
2017-09-23 12:55:21 +08:00
while (adapter && !found) {
Subject: windows: support to bind to a specific IPv6 address
2017-06-05 13:59:22 -03:00
addr = adapter->FirstUnicastAddress;
clean
2017-09-23 12:55:21 +08:00
while (addr && !found) {
if (addr->Address.lpSockaddr->sa_family == AF_INET6) {
sockaddr = (struct sockaddr_in6 *)
(addr->Address.lpSockaddr);
Subject: windows: support to bind to a specific IPv6 address
2017-06-05 13:59:22 -03:00
clean
2017-09-23 12:55:21 +08:00
lws_plat_inet_ntop(sockaddr->sin6_family,
&sockaddr->sin6_addr,
Subject: windows: support to bind to a specific IPv6 address
2017-06-05 13:59:22 -03:00
ip, sizeof(ip));
if (!strcmp(ip, ipaddr)) {
scope = sockaddr->sin6_scope_id;
found = 1;
break;
}
}
addr = addr->Next;
}
adapter = adapter->Next;
}
}
if (addrs)
free(addrs);
#endif
return scope;
}
#endif
ws ping pong on idle connections This adds a new member to the context creation info struct "ws_ping_pong_interval". If nonzero, it sets the number of seconds that established ws connections are allowed to be idle before a PING is forced to be sent. If zero (the default) then tracking of idle connection is disabled for backwards compatibility. Timeouts cover both the period between decision to send the ping and it being sent (because it needs the socket to become writeable), and the period between the ping being sent and the PONG coming back. INFO debug logs are issues when the timeout stuff is operating. You can test the server side by running the test server hacked to set ws_ping_pong_interval and debug log mask of 15. Both the mirror protocol and the server-status protocol are idle if nothing is happening and will trigger the PING / PONG testing. (You can also test using lwsws and /etc/lwsws/conf with "ws-pingpong-secs": "20" in the global section) For client, run the test client with -n -P 20 for 20s interval. -n stops the test client writing using the mirror protocol, so it will be idle and trigger the PING / PONGs. The timeout interval may be up to +10s late, as lws checks for affected connections every 10s.
2016-07-15 13:41:38 +08:00
LWS_EXTERN void
lws_restart_ws_ping_pong_timer(struct lws *wsi)
{
if (!wsi->context->ws_ping_pong_interval)
return;
if (wsi->state != LWSS_ESTABLISHED)
return;
wsi->u.ws.time_next_ping_check = (time_t)lws_now_secs() +
wsi->context->ws_ping_pong_interval;
}
introduce urlencode decode and sql escape public apis This adds - simple lws_urlencode() - simple lws_urldecode() - simple lws_sql_purify Those expect the data to all be there and process it up until the first '\0'. There is also a larger opaque apis for handling POST_BODY urldecode. To enable these, you need to give cmake -DLWS_WITH_STATEFUL_URLDECODE=1 (or arrange any larger feature that relies on it sets that in CMakeLists.txt) - stateful urldecode with parameter array These have create / process / destroy semantics on a struct that maintains decode state. Stateful urldecode is capable of dealing with large POST data in multiple POST_BODY callbacks cleanly, eg, file transfer by POST. Stateful urldecode with parameter array wraps the above with a canned callback that stores the urldecoded data and indexes them in a pointer array matching an array of parameter names. You may also pass it an optional callback when creating it, that will recieve uploaded file content. The test html is updated to support both urlencoded and multipart forms, with some javascript to do clientside validation of an arbitrary 100KB file size limit (there is no file size limit in the apis). Signed-off-by: Andy Green <andy@warmcat.com>
2016-06-08 10:07:02 +08:00
static const char *hex = "0123456789ABCDEF";
LWS_VISIBLE LWS_EXTERN const char *
lws_sql_purify(char *escaped, const char *string, int len)
{
const char *p = string;
char *q = escaped;
while (*p && len-- > 2) {
if (*p == '\'') {
protocol generic sessions Signed-off-by: Andy Green <andy@warmcat.com>
2016-05-19 15:28:31 +08:00
*q++ = '\'';
introduce urlencode decode and sql escape public apis This adds - simple lws_urlencode() - simple lws_urldecode() - simple lws_sql_purify Those expect the data to all be there and process it up until the first '\0'. There is also a larger opaque apis for handling POST_BODY urldecode. To enable these, you need to give cmake -DLWS_WITH_STATEFUL_URLDECODE=1 (or arrange any larger feature that relies on it sets that in CMakeLists.txt) - stateful urldecode with parameter array These have create / process / destroy semantics on a struct that maintains decode state. Stateful urldecode is capable of dealing with large POST data in multiple POST_BODY callbacks cleanly, eg, file transfer by POST. Stateful urldecode with parameter array wraps the above with a canned callback that stores the urldecoded data and indexes them in a pointer array matching an array of parameter names. You may also pass it an optional callback when creating it, that will recieve uploaded file content. The test html is updated to support both urlencoded and multipart forms, with some javascript to do clientside validation of an arbitrary 100KB file size limit (there is no file size limit in the apis). Signed-off-by: Andy Green <andy@warmcat.com>
2016-06-08 10:07:02 +08:00
*q++ = '\'';
len --;
p++;
} else
*q++ = *p++;
}
*q = '\0';
return escaped;
}
lws_json_escape Signed-off-by: Andy Green <andy@warmcat.com>
2016-06-17 10:05:23 +08:00
LWS_VISIBLE LWS_EXTERN const char *
lws_json_purify(char *escaped, const char *string, int len)
{
const char *p = string;
char *q = escaped;
plugin table dirlisting
2016-06-26 06:29:20 +08:00
if (!p) {
escaped[0] = '\0';
return escaped;
}
lws_json_escape Signed-off-by: Andy Green <andy@warmcat.com>
2016-06-17 10:05:23 +08:00
while (*p && len-- > 6) {
if (*p == '\"' || *p == '\\' || *p < 0x20) {
*q++ = '\\';
*q++ = 'u';
*q++ = '0';
*q++ = '0';
*q++ = hex[((*p) >> 4) & 15];
*q++ = hex[(*p) & 15];
len -= 5;
p++;
} else
*q++ = *p++;
}
*q = '\0';
return escaped;
}
introduce urlencode decode and sql escape public apis This adds - simple lws_urlencode() - simple lws_urldecode() - simple lws_sql_purify Those expect the data to all be there and process it up until the first '\0'. There is also a larger opaque apis for handling POST_BODY urldecode. To enable these, you need to give cmake -DLWS_WITH_STATEFUL_URLDECODE=1 (or arrange any larger feature that relies on it sets that in CMakeLists.txt) - stateful urldecode with parameter array These have create / process / destroy semantics on a struct that maintains decode state. Stateful urldecode is capable of dealing with large POST data in multiple POST_BODY callbacks cleanly, eg, file transfer by POST. Stateful urldecode with parameter array wraps the above with a canned callback that stores the urldecoded data and indexes them in a pointer array matching an array of parameter names. You may also pass it an optional callback when creating it, that will recieve uploaded file content. The test html is updated to support both urlencoded and multipart forms, with some javascript to do clientside validation of an arbitrary 100KB file size limit (there is no file size limit in the apis). Signed-off-by: Andy Green <andy@warmcat.com>
2016-06-08 10:07:02 +08:00
LWS_VISIBLE LWS_EXTERN const char *
lws_urlencode(char *escaped, const char *string, int len)
{
const char *p = string;
char *q = escaped;
while (*p && len-- > 3) {
if (*p == ' ') {
*q++ = '+';
p++;
continue;
}
if ((*p >= '0' && *p <= '9') ||
(*p >= 'A' && *p <= 'Z') ||
(*p >= 'a' && *p <= 'z')) {
*q++ = *p++;
continue;
}
*q++ = '%';
*q++ = hex[(*p >> 4) & 0xf];
*q++ = hex[*p & 0xf];
len -= 2;
p++;
}
*q = '\0';
return escaped;
}
LWS_VISIBLE LWS_EXTERN int
lws_urldecode(char *string, const char *escaped, int len)
{
int state = 0, n;
char sum = 0;
while (*escaped && len) {
switch (state) {
case 0:
if (*escaped == '%') {
state++;
escaped++;
continue;
}
if (*escaped == '+') {
escaped++;
*string++ = ' ';
len--;
continue;
}
*string++ = *escaped++;
len--;
break;
case 1:
n = char_to_hex(*escaped);
if (n < 0)
return -1;
escaped++;
sum = n << 4;
state++;
break;
case 2:
n = char_to_hex(*escaped);
if (n < 0)
return -1;
escaped++;
*string++ = sum | n;
len--;
state = 0;
break;
}
}
*string = '\0';
return 0;
}
introduce vhosts This patch splits out some lws_context members into a new lws_vhost struct. - ssl state and options per vhost - SSL_CTX for serving and client per vhost - protocols[] per vhost - extensions[] per vhost lws_context maintains a linked list of lws_vhosts. The same lws_context_creation_info struct is used to regulate both the context creation and to create vhosts: for backward compatibility if you didn't provide the new LWS_SERVER_OPTION_EXPLICIT_VHOSTS option, then a default vhost is created at context creation time using the same info data as the context itself. If you will have multiple vhosts though, you should give the LWS_SERVER_OPTION_EXPLICIT_VHOSTS option at context creation time, create the context first and then the vhosts afterwards using lws_create_vhost(contest, &info); Although there is a lot of housekeeping to implement this change, there is almost no additional overhead if you don't use multiple vhosts and very little api impact (no changes to test apps). Signed-off-by: Andy Green <andy@warmcat.com>
2016-03-28 10:10:43 +08:00
LWS_VISIBLE LWS_EXTERN int
lws_finalize_startup(struct lws_context *context)
{
struct lws_context_creation_info info;
info.uid = context->uid;
info.gid = context->gid;
capabilities support
2017-06-07 06:10:02 +08:00
#if defined(LWS_HAVE_SYS_CAPABILITY_H) && defined(LWS_HAVE_LIBCAP)
memcpy(info.caps, context->caps, sizeof(info.caps));
info.count_caps = context->count_caps;
#endif
introduce vhosts This patch splits out some lws_context members into a new lws_vhost struct. - ssl state and options per vhost - SSL_CTX for serving and client per vhost - protocols[] per vhost - extensions[] per vhost lws_context maintains a linked list of lws_vhosts. The same lws_context_creation_info struct is used to regulate both the context creation and to create vhosts: for backward compatibility if you didn't provide the new LWS_SERVER_OPTION_EXPLICIT_VHOSTS option, then a default vhost is created at context creation time using the same info data as the context itself. If you will have multiple vhosts though, you should give the LWS_SERVER_OPTION_EXPLICIT_VHOSTS option at context creation time, create the context first and then the vhosts afterwards using lws_create_vhost(contest, &info); Although there is a lot of housekeeping to implement this change, there is almost no additional overhead if you don't use multiple vhosts and very little api impact (no changes to test apps). Signed-off-by: Andy Green <andy@warmcat.com>
2016-03-28 10:10:43 +08:00
if (lws_check_opt(context->options, LWS_SERVER_OPTION_EXPLICIT_VHOSTS))
lws_plat_drop_app_privileges(&info);
return 0;
}
lws_snprintf Thanks to Fabrice Gilot for reporting the problem that led to uncovering this. Due to a misunderstanding of the return value of snprintf (it is not truncated according to the max size passed in) in several places relying on snprintf to truncate the length overflows are possible. This patch wraps snprintf with a new lws_snprintf() which does truncate its length to allow the buffer limiting scheme to work properly. All users should update with these fixes.
2016-09-15 02:22:57 +08:00
int
lws_snprintf(char *str, size_t size, const char *format, ...)
{
va_list ap;
int n;
if (!size)
return 0;
va_start(ap, format);
n = vsnprintf(str, size, format, ap);
va_end(ap);
win32: enable 64-bit file lengths https://github.com/warmcat/libwebsockets/issues/916 AG added more appveyor warning cleaning and stitched in cmake tests
2017-06-09 20:20:42 +08:00
if (n >= (int)size)
lws_snprintf Thanks to Fabrice Gilot for reporting the problem that led to uncovering this. Due to a misunderstanding of the return value of snprintf (it is not truncated according to the max size passed in) in several places relying on snprintf to truncate the length overflows are possible. This patch wraps snprintf with a new lws_snprintf() which does truncate its length to allow the buffer limiting scheme to work properly. All users should update with these fixes.
2016-09-15 02:22:57 +08:00
return size;
return n;
}
cgi Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-21 21:25:48 +08:00
LWS_VISIBLE LWS_EXTERN int
lws_is_cgi(struct lws *wsi) {
#ifdef LWS_WITH_CGI
return !!wsi->cgi;
#else
return 0;
#endif
}
cgi: also allow for generic spawn
2017-09-19 07:58:04 +08:00
extensions add api for user code option manipulation Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-08 09:45:49 +08:00
#ifdef LWS_NO_EXTENSIONS
LWS_EXTERN int
lws_set_extension_option(struct lws *wsi, const char *ext_name,
const char *opt_name, const char *opt_val)
{
return -1;
}
#endif
json dump vhost Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-14 15:07:44 +08:00
context deprecation 1) This makes lwsws run a parent process with the original permissions. But this process is only able to respond to SIGHUP, it doesn't do anything else. 2) You can send this parent process a SIGHUP now to cause it to - close listening sockets in existing lwsws processes - mark those processes as to exit when the number of active connections on the falls to zero - spawn a fresh child process from scratch, using latest configuration file content, latest plugins, etc. It can now reopen listening sockets if it chooses to, or open different listen ports or whatever. Notes: 1) lws_context_destroy() has been split into two pieces... the reason for the split is the first part closes the per-vhost protocols, but since they may have created libuv objects in the per-vhost protocol storage, these cannot be freed until after the loop has been run. That's the purpose of the second part of the context destruction, lws_context_destroy2(). For compatibility, if you are not using libuv, the first part calls the second part. However if you are using libuv, you must now call the second part from your own main.c after the first part.
2016-12-16 07:37:43 +08:00
void
lws_sum_stats(const struct lws_context *ctx, struct lws_conn_stats *cs)
{
const struct lws_vhost *vh = ctx->vhost_list;
while (vh) {
cs->rx += vh->conn_stats.rx;
cs->tx += vh->conn_stats.tx;
http2: make usable HTTP/2 support is now able to serve the test server, complete with websockets, from a single vhost. - This works the same with both OpenSSL and mbedTLS. - POST is now wired up and works (also for file upload). - CGI is wired up and works. - Redirect is adapted and works - lwsws works. - URI urldecode, sanitation and argument parsing wired up for :path valgrind clean (aside from openssl-style false uninit data usage in mbedtls send occasionally) h2spec reports: $ h2spec -h 127.0.0.1 -p 7681 -t -k -o 1 ... 145 tests, 145 passed, 0 skipped, 0 failed" Incorporates: - "https://github.com/warmcat/libwebsockets/pull/1039 Fixes issue with -Werror=unused-variable flag - https://github.com/sartura/libwebsockets/commit/2c843a1395bdca5c23b40d989f508fc96aa5ba3b ssl: fix infinite loop on client cert verification failure Signed-off-by: Petar Paradzik <petar.paradzik@sartura.hr>" Caused and fixes Coverity 184887 - 184892
2017-10-13 10:33:02 +08:00
cs->h1_conn += vh->conn_stats.h1_conn;
cs->h1_trans += vh->conn_stats.h1_trans;
cs->h2_trans += vh->conn_stats.h2_trans;
context deprecation 1) This makes lwsws run a parent process with the original permissions. But this process is only able to respond to SIGHUP, it doesn't do anything else. 2) You can send this parent process a SIGHUP now to cause it to - close listening sockets in existing lwsws processes - mark those processes as to exit when the number of active connections on the falls to zero - spawn a fresh child process from scratch, using latest configuration file content, latest plugins, etc. It can now reopen listening sockets if it chooses to, or open different listen ports or whatever. Notes: 1) lws_context_destroy() has been split into two pieces... the reason for the split is the first part closes the per-vhost protocols, but since they may have created libuv objects in the per-vhost protocol storage, these cannot be freed until after the loop has been run. That's the purpose of the second part of the context destruction, lws_context_destroy2(). For compatibility, if you are not using libuv, the first part calls the second part. However if you are using libuv, you must now call the second part from your own main.c after the first part.
2016-12-16 07:37:43 +08:00
cs->ws_upg += vh->conn_stats.ws_upg;
http2: make usable HTTP/2 support is now able to serve the test server, complete with websockets, from a single vhost. - This works the same with both OpenSSL and mbedTLS. - POST is now wired up and works (also for file upload). - CGI is wired up and works. - Redirect is adapted and works - lwsws works. - URI urldecode, sanitation and argument parsing wired up for :path valgrind clean (aside from openssl-style false uninit data usage in mbedtls send occasionally) h2spec reports: $ h2spec -h 127.0.0.1 -p 7681 -t -k -o 1 ... 145 tests, 145 passed, 0 skipped, 0 failed" Incorporates: - "https://github.com/warmcat/libwebsockets/pull/1039 Fixes issue with -Werror=unused-variable flag - https://github.com/sartura/libwebsockets/commit/2c843a1395bdca5c23b40d989f508fc96aa5ba3b ssl: fix infinite loop on client cert verification failure Signed-off-by: Petar Paradzik <petar.paradzik@sartura.hr>" Caused and fixes Coverity 184887 - 184892
2017-10-13 10:33:02 +08:00
cs->h2_upg += vh->conn_stats.h2_upg;
cs->h2_alpn += vh->conn_stats.h2_alpn;
cs->h2_subs += vh->conn_stats.h2_subs;
context deprecation 1) This makes lwsws run a parent process with the original permissions. But this process is only able to respond to SIGHUP, it doesn't do anything else. 2) You can send this parent process a SIGHUP now to cause it to - close listening sockets in existing lwsws processes - mark those processes as to exit when the number of active connections on the falls to zero - spawn a fresh child process from scratch, using latest configuration file content, latest plugins, etc. It can now reopen listening sockets if it chooses to, or open different listen ports or whatever. Notes: 1) lws_context_destroy() has been split into two pieces... the reason for the split is the first part closes the per-vhost protocols, but since they may have created libuv objects in the per-vhost protocol storage, these cannot be freed until after the loop has been run. That's the purpose of the second part of the context destruction, lws_context_destroy2(). For compatibility, if you are not using libuv, the first part calls the second part. However if you are using libuv, you must now call the second part from your own main.c after the first part.
2016-12-16 07:37:43 +08:00
cs->rejected += vh->conn_stats.rejected;
vh = vh->vhost_next;
}
}
plugin lws server status Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-15 14:01:29 +08:00
#ifdef LWS_WITH_SERVER_STATUS
json dump vhost Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-14 15:07:44 +08:00
LWS_EXTERN int
lws_json_dump_vhost(const struct lws_vhost *vh, char *buf, int len)
{
static const char * const prots[] = {
"http://",
"https://",
"file://",
"cgi://",
">http://",
">https://",
add protocol plugin for post demo Signed-off-by: Andy Green <andy@warmcat.com>
2016-05-09 09:37:01 +08:00
"callback://"
json dump vhost Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-14 15:07:44 +08:00
};
char *orig = buf, *end = buf + len - 1, first = 1;
int n = 0;
if (len < 100)
return 0;
lws_snprintf Thanks to Fabrice Gilot for reporting the problem that led to uncovering this. Due to a misunderstanding of the return value of snprintf (it is not truncated according to the max size passed in) in several places relying on snprintf to truncate the length overflows are possible. This patch wraps snprintf with a new lws_snprintf() which does truncate its length to allow the buffer limiting scheme to work properly. All users should update with these fixes.
2016-09-15 02:22:57 +08:00
buf += lws_snprintf(buf, end - buf,
json dump vhost Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-14 15:07:44 +08:00
"{\n \"name\":\"%s\",\n"
" \"port\":\"%d\",\n"
plugin lws server status Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-15 14:01:29 +08:00
" \"use_ssl\":\"%d\",\n"
json dump vhost Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-14 15:07:44 +08:00
" \"sts\":\"%d\",\n"
stats upgrade rx tx to long long Signed-off-by: Andy Green <andy@warmcat.com>
2016-05-04 15:59:27 +08:00
" \"rx\":\"%llu\",\n"
" \"tx\":\"%llu\",\n"
http2: make usable HTTP/2 support is now able to serve the test server, complete with websockets, from a single vhost. - This works the same with both OpenSSL and mbedTLS. - POST is now wired up and works (also for file upload). - CGI is wired up and works. - Redirect is adapted and works - lwsws works. - URI urldecode, sanitation and argument parsing wired up for :path valgrind clean (aside from openssl-style false uninit data usage in mbedtls send occasionally) h2spec reports: $ h2spec -h 127.0.0.1 -p 7681 -t -k -o 1 ... 145 tests, 145 passed, 0 skipped, 0 failed" Incorporates: - "https://github.com/warmcat/libwebsockets/pull/1039 Fixes issue with -Werror=unused-variable flag - https://github.com/sartura/libwebsockets/commit/2c843a1395bdca5c23b40d989f508fc96aa5ba3b ssl: fix infinite loop on client cert verification failure Signed-off-by: Petar Paradzik <petar.paradzik@sartura.hr>" Caused and fixes Coverity 184887 - 184892
2017-10-13 10:33:02 +08:00
" \"h1_conn\":\"%lu\",\n"
" \"h1_trans\":\"%lu\",\n"
" \"h2_trans\":\"%lu\",\n"
plugin lws server status Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-15 14:01:29 +08:00
" \"ws_upg\":\"%lu\",\n"
context deprecation 1) This makes lwsws run a parent process with the original permissions. But this process is only able to respond to SIGHUP, it doesn't do anything else. 2) You can send this parent process a SIGHUP now to cause it to - close listening sockets in existing lwsws processes - mark those processes as to exit when the number of active connections on the falls to zero - spawn a fresh child process from scratch, using latest configuration file content, latest plugins, etc. It can now reopen listening sockets if it chooses to, or open different listen ports or whatever. Notes: 1) lws_context_destroy() has been split into two pieces... the reason for the split is the first part closes the per-vhost protocols, but since they may have created libuv objects in the per-vhost protocol storage, these cannot be freed until after the loop has been run. That's the purpose of the second part of the context destruction, lws_context_destroy2(). For compatibility, if you are not using libuv, the first part calls the second part. However if you are using libuv, you must now call the second part from your own main.c after the first part.
2016-12-16 07:37:43 +08:00
" \"rejected\":\"%lu\",\n"
http2: make usable HTTP/2 support is now able to serve the test server, complete with websockets, from a single vhost. - This works the same with both OpenSSL and mbedTLS. - POST is now wired up and works (also for file upload). - CGI is wired up and works. - Redirect is adapted and works - lwsws works. - URI urldecode, sanitation and argument parsing wired up for :path valgrind clean (aside from openssl-style false uninit data usage in mbedtls send occasionally) h2spec reports: $ h2spec -h 127.0.0.1 -p 7681 -t -k -o 1 ... 145 tests, 145 passed, 0 skipped, 0 failed" Incorporates: - "https://github.com/warmcat/libwebsockets/pull/1039 Fixes issue with -Werror=unused-variable flag - https://github.com/sartura/libwebsockets/commit/2c843a1395bdca5c23b40d989f508fc96aa5ba3b ssl: fix infinite loop on client cert verification failure Signed-off-by: Petar Paradzik <petar.paradzik@sartura.hr>" Caused and fixes Coverity 184887 - 184892
2017-10-13 10:33:02 +08:00
" \"h2_upg\":\"%lu\",\n"
" \"h2_alpn\":\"%lu\",\n"
" \"h2_subs\":\"%lu\""
plugin lws server status Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-15 14:01:29 +08:00
,
json dump vhost Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-14 15:07:44 +08:00
vh->name, vh->listen_port,
#ifdef LWS_OPENSSL_SUPPORT
vh->use_ssl,
#else
0,
#endif
!!(vh->options & LWS_SERVER_OPTION_STS),
context deprecation 1) This makes lwsws run a parent process with the original permissions. But this process is only able to respond to SIGHUP, it doesn't do anything else. 2) You can send this parent process a SIGHUP now to cause it to - close listening sockets in existing lwsws processes - mark those processes as to exit when the number of active connections on the falls to zero - spawn a fresh child process from scratch, using latest configuration file content, latest plugins, etc. It can now reopen listening sockets if it chooses to, or open different listen ports or whatever. Notes: 1) lws_context_destroy() has been split into two pieces... the reason for the split is the first part closes the per-vhost protocols, but since they may have created libuv objects in the per-vhost protocol storage, these cannot be freed until after the loop has been run. That's the purpose of the second part of the context destruction, lws_context_destroy2(). For compatibility, if you are not using libuv, the first part calls the second part. However if you are using libuv, you must now call the second part from your own main.c after the first part.
2016-12-16 07:37:43 +08:00
vh->conn_stats.rx, vh->conn_stats.tx,
http2: make usable HTTP/2 support is now able to serve the test server, complete with websockets, from a single vhost. - This works the same with both OpenSSL and mbedTLS. - POST is now wired up and works (also for file upload). - CGI is wired up and works. - Redirect is adapted and works - lwsws works. - URI urldecode, sanitation and argument parsing wired up for :path valgrind clean (aside from openssl-style false uninit data usage in mbedtls send occasionally) h2spec reports: $ h2spec -h 127.0.0.1 -p 7681 -t -k -o 1 ... 145 tests, 145 passed, 0 skipped, 0 failed" Incorporates: - "https://github.com/warmcat/libwebsockets/pull/1039 Fixes issue with -Werror=unused-variable flag - https://github.com/sartura/libwebsockets/commit/2c843a1395bdca5c23b40d989f508fc96aa5ba3b ssl: fix infinite loop on client cert verification failure Signed-off-by: Petar Paradzik <petar.paradzik@sartura.hr>" Caused and fixes Coverity 184887 - 184892
2017-10-13 10:33:02 +08:00
vh->conn_stats.h1_conn,
vh->conn_stats.h1_trans,
vh->conn_stats.h2_trans,
context deprecation 1) This makes lwsws run a parent process with the original permissions. But this process is only able to respond to SIGHUP, it doesn't do anything else. 2) You can send this parent process a SIGHUP now to cause it to - close listening sockets in existing lwsws processes - mark those processes as to exit when the number of active connections on the falls to zero - spawn a fresh child process from scratch, using latest configuration file content, latest plugins, etc. It can now reopen listening sockets if it chooses to, or open different listen ports or whatever. Notes: 1) lws_context_destroy() has been split into two pieces... the reason for the split is the first part closes the per-vhost protocols, but since they may have created libuv objects in the per-vhost protocol storage, these cannot be freed until after the loop has been run. That's the purpose of the second part of the context destruction, lws_context_destroy2(). For compatibility, if you are not using libuv, the first part calls the second part. However if you are using libuv, you must now call the second part from your own main.c after the first part.
2016-12-16 07:37:43 +08:00
vh->conn_stats.ws_upg,
vh->conn_stats.rejected,
http2: make usable HTTP/2 support is now able to serve the test server, complete with websockets, from a single vhost. - This works the same with both OpenSSL and mbedTLS. - POST is now wired up and works (also for file upload). - CGI is wired up and works. - Redirect is adapted and works - lwsws works. - URI urldecode, sanitation and argument parsing wired up for :path valgrind clean (aside from openssl-style false uninit data usage in mbedtls send occasionally) h2spec reports: $ h2spec -h 127.0.0.1 -p 7681 -t -k -o 1 ... 145 tests, 145 passed, 0 skipped, 0 failed" Incorporates: - "https://github.com/warmcat/libwebsockets/pull/1039 Fixes issue with -Werror=unused-variable flag - https://github.com/sartura/libwebsockets/commit/2c843a1395bdca5c23b40d989f508fc96aa5ba3b ssl: fix infinite loop on client cert verification failure Signed-off-by: Petar Paradzik <petar.paradzik@sartura.hr>" Caused and fixes Coverity 184887 - 184892
2017-10-13 10:33:02 +08:00
vh->conn_stats.h2_upg,
vh->conn_stats.h2_alpn,
vh->conn_stats.h2_subs
json dump vhost Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-14 15:07:44 +08:00
);
if (vh->mount_list) {
create_vhost move mounts list to info There's no reason to not have the mounts linked list init also in the info struct, rather than provide as a paramater to lws_create_vhost(). Now is a good time to normalize that since this api only exists in master. This also allows oldstyle "do everything at context creation time in one vhost" guys to leverage mounts. Also there's no reason the mounts linked-list pointer and all uses in lws are non-const, so make them all explicitly const *. Update the info struct docs to clarify which members are used when creating a vhost and which for context creation. Signed-off-by: Andy Green <andy@warmcat.com>
2016-05-02 04:59:54 +08:00
const struct lws_http_mount *m = vh->mount_list;
json dump vhost Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-14 15:07:44 +08:00
lws_snprintf Thanks to Fabrice Gilot for reporting the problem that led to uncovering this. Due to a misunderstanding of the return value of snprintf (it is not truncated according to the max size passed in) in several places relying on snprintf to truncate the length overflows are possible. This patch wraps snprintf with a new lws_snprintf() which does truncate its length to allow the buffer limiting scheme to work properly. All users should update with these fixes.
2016-09-15 02:22:57 +08:00
buf += lws_snprintf(buf, end - buf, ",\n \"mounts\":[");
json dump vhost Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-14 15:07:44 +08:00
while (m) {
if (!first)
lws_snprintf Thanks to Fabrice Gilot for reporting the problem that led to uncovering this. Due to a misunderstanding of the return value of snprintf (it is not truncated according to the max size passed in) in several places relying on snprintf to truncate the length overflows are possible. This patch wraps snprintf with a new lws_snprintf() which does truncate its length to allow the buffer limiting scheme to work properly. All users should update with these fixes.
2016-09-15 02:22:57 +08:00
buf += lws_snprintf(buf, end - buf, ",");
buf += lws_snprintf(buf, end - buf,
json dump vhost Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-14 15:07:44 +08:00
"\n {\n \"mountpoint\":\"%s\",\n"
http cache policy This allows mounts to define the caching policy of the files inside them. Support is added in lwsws for controlling it from the config files. The api for serializing a mount struct opaquely is removed and lws_http_mount struct made public... it was getting out of control trying to hide the options. Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-22 08:53:49 +08:00
" \"origin\":\"%s%s\",\n"
" \"cache_max_age\":\"%d\",\n"
" \"cache_reuse\":\"%d\",\n"
" \"cache_revalidate\":\"%d\",\n"
" \"cache_intermediaries\":\"%d\"\n"
json dump vhost Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-14 15:07:44 +08:00
,
m->mountpoint,
prots[m->origin_protocol],
http cache policy This allows mounts to define the caching policy of the files inside them. Support is added in lwsws for controlling it from the config files. The api for serializing a mount struct opaquely is removed and lws_http_mount struct made public... it was getting out of control trying to hide the options. Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-22 08:53:49 +08:00
m->origin,
m->cache_max_age,
m->cache_reusable,
m->cache_revalidate,
m->cache_intermediaries);
json dump vhost Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-14 15:07:44 +08:00
if (m->def)
lws_snprintf Thanks to Fabrice Gilot for reporting the problem that led to uncovering this. Due to a misunderstanding of the return value of snprintf (it is not truncated according to the max size passed in) in several places relying on snprintf to truncate the length overflows are possible. This patch wraps snprintf with a new lws_snprintf() which does truncate its length to allow the buffer limiting scheme to work properly. All users should update with these fixes.
2016-09-15 02:22:57 +08:00
buf += lws_snprintf(buf, end - buf,
json dump vhost Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-14 15:07:44 +08:00
",\n \"default\":\"%s\"",
m->def);
lws_snprintf Thanks to Fabrice Gilot for reporting the problem that led to uncovering this. Due to a misunderstanding of the return value of snprintf (it is not truncated according to the max size passed in) in several places relying on snprintf to truncate the length overflows are possible. This patch wraps snprintf with a new lws_snprintf() which does truncate its length to allow the buffer limiting scheme to work properly. All users should update with these fixes.
2016-09-15 02:22:57 +08:00
buf += lws_snprintf(buf, end - buf, "\n }");
json dump vhost Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-14 15:07:44 +08:00
first = 0;
m = m->mount_next;
}
lws_snprintf Thanks to Fabrice Gilot for reporting the problem that led to uncovering this. Due to a misunderstanding of the return value of snprintf (it is not truncated according to the max size passed in) in several places relying on snprintf to truncate the length overflows are possible. This patch wraps snprintf with a new lws_snprintf() which does truncate its length to allow the buffer limiting scheme to work properly. All users should update with these fixes.
2016-09-15 02:22:57 +08:00
buf += lws_snprintf(buf, end - buf, "\n ]");
json dump vhost Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-14 15:07:44 +08:00
}
if (vh->protocols) {
n = 0;
first = 1;
lws_snprintf Thanks to Fabrice Gilot for reporting the problem that led to uncovering this. Due to a misunderstanding of the return value of snprintf (it is not truncated according to the max size passed in) in several places relying on snprintf to truncate the length overflows are possible. This patch wraps snprintf with a new lws_snprintf() which does truncate its length to allow the buffer limiting scheme to work properly. All users should update with these fixes.
2016-09-15 02:22:57 +08:00
buf += lws_snprintf(buf, end - buf, ",\n \"ws-protocols\":[");
json dump vhost Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-14 15:07:44 +08:00
while (n < vh->count_protocols) {
if (!first)
lws_snprintf Thanks to Fabrice Gilot for reporting the problem that led to uncovering this. Due to a misunderstanding of the return value of snprintf (it is not truncated according to the max size passed in) in several places relying on snprintf to truncate the length overflows are possible. This patch wraps snprintf with a new lws_snprintf() which does truncate its length to allow the buffer limiting scheme to work properly. All users should update with these fixes.
2016-09-15 02:22:57 +08:00
buf += lws_snprintf(buf, end - buf, ",");
buf += lws_snprintf(buf, end - buf,
win fix warnings from appveyor Signed-off-by: Andy Green <andy@warmcat.com>
2016-05-15 08:29:37 +08:00
"\n {\n \"%s\":{\n"
json dump vhost Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-14 15:07:44 +08:00
" \"status\":\"ok\"\n }\n }"
,
vh->protocols[n].name);
first = 0;
n++;
}
lws_snprintf Thanks to Fabrice Gilot for reporting the problem that led to uncovering this. Due to a misunderstanding of the return value of snprintf (it is not truncated according to the max size passed in) in several places relying on snprintf to truncate the length overflows are possible. This patch wraps snprintf with a new lws_snprintf() which does truncate its length to allow the buffer limiting scheme to work properly. All users should update with these fixes.
2016-09-15 02:22:57 +08:00
buf += lws_snprintf(buf, end - buf, "\n ]");
json dump vhost Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-14 15:07:44 +08:00
}
lws_snprintf Thanks to Fabrice Gilot for reporting the problem that led to uncovering this. Due to a misunderstanding of the return value of snprintf (it is not truncated according to the max size passed in) in several places relying on snprintf to truncate the length overflows are possible. This patch wraps snprintf with a new lws_snprintf() which does truncate its length to allow the buffer limiting scheme to work properly. All users should update with these fixes.
2016-09-15 02:22:57 +08:00
buf += lws_snprintf(buf, end - buf, "\n}");
json dump vhost Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-14 15:07:44 +08:00
return buf - orig;
}
plugin lws server status Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-15 14:01:29 +08:00
LWS_EXTERN LWS_VISIBLE int
context deprecation 1) This makes lwsws run a parent process with the original permissions. But this process is only able to respond to SIGHUP, it doesn't do anything else. 2) You can send this parent process a SIGHUP now to cause it to - close listening sockets in existing lwsws processes - mark those processes as to exit when the number of active connections on the falls to zero - spawn a fresh child process from scratch, using latest configuration file content, latest plugins, etc. It can now reopen listening sockets if it chooses to, or open different listen ports or whatever. Notes: 1) lws_context_destroy() has been split into two pieces... the reason for the split is the first part closes the per-vhost protocols, but since they may have created libuv objects in the per-vhost protocol storage, these cannot be freed until after the loop has been run. That's the purpose of the second part of the context destruction, lws_context_destroy2(). For compatibility, if you are not using libuv, the first part calls the second part. However if you are using libuv, you must now call the second part from your own main.c after the first part.
2016-12-16 07:37:43 +08:00
lws_json_dump_context(const struct lws_context *context, char *buf, int len,
int hide_vhosts)
plugin lws server status Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-15 14:01:29 +08:00
{
char *orig = buf, *end = buf + len - 1, first = 1;
cgi lifecycle logging Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-20 06:10:56 +08:00
const struct lws_vhost *vh = context->vhost_list;
context deprecation 1) This makes lwsws run a parent process with the original permissions. But this process is only able to respond to SIGHUP, it doesn't do anything else. 2) You can send this parent process a SIGHUP now to cause it to - close listening sockets in existing lwsws processes - mark those processes as to exit when the number of active connections on the falls to zero - spawn a fresh child process from scratch, using latest configuration file content, latest plugins, etc. It can now reopen listening sockets if it chooses to, or open different listen ports or whatever. Notes: 1) lws_context_destroy() has been split into two pieces... the reason for the split is the first part closes the per-vhost protocols, but since they may have created libuv objects in the per-vhost protocol storage, these cannot be freed until after the loop has been run. That's the purpose of the second part of the context destruction, lws_context_destroy2(). For compatibility, if you are not using libuv, the first part calls the second part. However if you are using libuv, you must now call the second part from your own main.c after the first part.
2016-12-16 07:37:43 +08:00
const struct lws_context_per_thread *pt;
time_t t = time(NULL);
coverity 175436: dead cruft
2017-02-06 10:04:04 +08:00
int n, listening = 0, cgi_count = 0;
context deprecation 1) This makes lwsws run a parent process with the original permissions. But this process is only able to respond to SIGHUP, it doesn't do anything else. 2) You can send this parent process a SIGHUP now to cause it to - close listening sockets in existing lwsws processes - mark those processes as to exit when the number of active connections on the falls to zero - spawn a fresh child process from scratch, using latest configuration file content, latest plugins, etc. It can now reopen listening sockets if it chooses to, or open different listen ports or whatever. Notes: 1) lws_context_destroy() has been split into two pieces... the reason for the split is the first part closes the per-vhost protocols, but since they may have created libuv objects in the per-vhost protocol storage, these cannot be freed until after the loop has been run. That's the purpose of the second part of the context destruction, lws_context_destroy2(). For compatibility, if you are not using libuv, the first part calls the second part. However if you are using libuv, you must now call the second part from your own main.c after the first part.
2016-12-16 07:37:43 +08:00
struct lws_conn_stats cs;
double d = 0;
cgi lifecycle logging Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-20 06:10:56 +08:00
#ifdef LWS_WITH_CGI
struct lws_cgi * const *pcgi;
#endif
context deprecation 1) This makes lwsws run a parent process with the original permissions. But this process is only able to respond to SIGHUP, it doesn't do anything else. 2) You can send this parent process a SIGHUP now to cause it to - close listening sockets in existing lwsws processes - mark those processes as to exit when the number of active connections on the falls to zero - spawn a fresh child process from scratch, using latest configuration file content, latest plugins, etc. It can now reopen listening sockets if it chooses to, or open different listen ports or whatever. Notes: 1) lws_context_destroy() has been split into two pieces... the reason for the split is the first part closes the per-vhost protocols, but since they may have created libuv objects in the per-vhost protocol storage, these cannot be freed until after the loop has been run. That's the purpose of the second part of the context destruction, lws_context_destroy2(). For compatibility, if you are not using libuv, the first part calls the second part. However if you are using libuv, you must now call the second part from your own main.c after the first part.
2016-12-16 07:37:43 +08:00
CMake: convert all LWS_USE_... to LWS_WITH_... Almost all the CMake options begin with LWS_WITH_..., but many of the symbols passed to lws are LWS_USE_... , this causes neededless confusion, compounded by the fact that a few CMake options also begin with LWS_USE_. This patch globally converts all LWS_USE_... to LWS_WITH_..., so there is only one prefix to remember in both CMake and the code. The affected public CMake options are LWS_USE_BORINGSSL -> LWS_WITH_BORINGSSL LWS_USE_CYASSL -> LWS_WITH_CYASSL LWS_USE_WOLFSSL -> LWS_WITH_WOLFSSL LWS_USE_MBEDTLS -> LWS_WITH_MBEDTLS LWS_USE_BUNDLED_ZLIB -> LWS_WITH_BUNDLED_ZLIB
2017-09-28 11:29:03 +08:00
#ifdef LWS_WITH_LIBUV
context deprecation 1) This makes lwsws run a parent process with the original permissions. But this process is only able to respond to SIGHUP, it doesn't do anything else. 2) You can send this parent process a SIGHUP now to cause it to - close listening sockets in existing lwsws processes - mark those processes as to exit when the number of active connections on the falls to zero - spawn a fresh child process from scratch, using latest configuration file content, latest plugins, etc. It can now reopen listening sockets if it chooses to, or open different listen ports or whatever. Notes: 1) lws_context_destroy() has been split into two pieces... the reason for the split is the first part closes the per-vhost protocols, but since they may have created libuv objects in the per-vhost protocol storage, these cannot be freed until after the loop has been run. That's the purpose of the second part of the context destruction, lws_context_destroy2(). For compatibility, if you are not using libuv, the first part calls the second part. However if you are using libuv, you must now call the second part from your own main.c after the first part.
2016-12-16 07:37:43 +08:00
uv_uptime(&d);
#endif
plugin lws server status Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-15 14:01:29 +08:00
lws_snprintf Thanks to Fabrice Gilot for reporting the problem that led to uncovering this. Due to a misunderstanding of the return value of snprintf (it is not truncated according to the max size passed in) in several places relying on snprintf to truncate the length overflows are possible. This patch wraps snprintf with a new lws_snprintf() which does truncate its length to allow the buffer limiting scheme to work properly. All users should update with these fixes.
2016-09-15 02:22:57 +08:00
buf += lws_snprintf(buf, end - buf, "{ "
context deprecation 1) This makes lwsws run a parent process with the original permissions. But this process is only able to respond to SIGHUP, it doesn't do anything else. 2) You can send this parent process a SIGHUP now to cause it to - close listening sockets in existing lwsws processes - mark those processes as to exit when the number of active connections on the falls to zero - spawn a fresh child process from scratch, using latest configuration file content, latest plugins, etc. It can now reopen listening sockets if it chooses to, or open different listen ports or whatever. Notes: 1) lws_context_destroy() has been split into two pieces... the reason for the split is the first part closes the per-vhost protocols, but since they may have created libuv objects in the per-vhost protocol storage, these cannot be freed until after the loop has been run. That's the purpose of the second part of the context destruction, lws_context_destroy2(). For compatibility, if you are not using libuv, the first part calls the second part. However if you are using libuv, you must now call the second part from your own main.c after the first part.
2016-12-16 07:37:43 +08:00
"\"version\":\"%s\",\n"
"\"uptime\":\"%ld\",\n",
lws_get_library_version(),
(long)d);
server status add getloadavg and lib version and more css Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-26 07:45:45 +08:00
#ifdef LWS_HAVE_GETLOADAVG
{
double d[3];
int m;
m = getloadavg(d, 3);
for (n = 0; n < m; n++) {
lws_snprintf Thanks to Fabrice Gilot for reporting the problem that led to uncovering this. Due to a misunderstanding of the return value of snprintf (it is not truncated according to the max size passed in) in several places relying on snprintf to truncate the length overflows are possible. This patch wraps snprintf with a new lws_snprintf() which does truncate its length to allow the buffer limiting scheme to work properly. All users should update with these fixes.
2016-09-15 02:22:57 +08:00
buf += lws_snprintf(buf, end - buf,
server status add getloadavg and lib version and more css Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-26 07:45:45 +08:00
"\"l%d\":\"%.2f\",\n",
n + 1, d[n]);
}
}
#endif
plugin lws server status Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-15 14:01:29 +08:00
context deprecation 1) This makes lwsws run a parent process with the original permissions. But this process is only able to respond to SIGHUP, it doesn't do anything else. 2) You can send this parent process a SIGHUP now to cause it to - close listening sockets in existing lwsws processes - mark those processes as to exit when the number of active connections on the falls to zero - spawn a fresh child process from scratch, using latest configuration file content, latest plugins, etc. It can now reopen listening sockets if it chooses to, or open different listen ports or whatever. Notes: 1) lws_context_destroy() has been split into two pieces... the reason for the split is the first part closes the per-vhost protocols, but since they may have created libuv objects in the per-vhost protocol storage, these cannot be freed until after the loop has been run. That's the purpose of the second part of the context destruction, lws_context_destroy2(). For compatibility, if you are not using libuv, the first part calls the second part. However if you are using libuv, you must now call the second part from your own main.c after the first part.
2016-12-16 07:37:43 +08:00
buf += lws_snprintf(buf, end - buf, "\"contexts\":[\n");
buf += lws_snprintf(buf, end - buf, "{ "
"\"context_uptime\":\"%ld\",\n"
"\"cgi_spawned\":\"%d\",\n"
"\"pt_fd_max\":\"%d\",\n"
"\"ah_pool_max\":\"%d\",\n"
"\"deprecated\":\"%d\",\n"
"\"wsi_alive\":\"%d\",\n",
(unsigned long)(t - context->time_up),
context->count_cgi_spawned,
context->fd_limit_per_thread,
context->max_http_header_pool,
context->deprecated,
context->count_wsi_allocated);
lws_snprintf Thanks to Fabrice Gilot for reporting the problem that led to uncovering this. Due to a misunderstanding of the return value of snprintf (it is not truncated according to the max size passed in) in several places relying on snprintf to truncate the length overflows are possible. This patch wraps snprintf with a new lws_snprintf() which does truncate its length to allow the buffer limiting scheme to work properly. All users should update with these fixes.
2016-09-15 02:22:57 +08:00
buf += lws_snprintf(buf, end - buf, "\"pt\":[\n ");
cgi lifecycle logging Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-20 06:10:56 +08:00
for (n = 0; n < context->count_threads; n++) {
pt = &context->pt[n];
if (n)
lws_snprintf Thanks to Fabrice Gilot for reporting the problem that led to uncovering this. Due to a misunderstanding of the return value of snprintf (it is not truncated according to the max size passed in) in several places relying on snprintf to truncate the length overflows are possible. This patch wraps snprintf with a new lws_snprintf() which does truncate its length to allow the buffer limiting scheme to work properly. All users should update with these fixes.
2016-09-15 02:22:57 +08:00
buf += lws_snprintf(buf, end - buf, ",");
buf += lws_snprintf(buf, end - buf,
cgi lifecycle logging Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-20 06:10:56 +08:00
"\n {\n"
" \"fds_count\":\"%d\",\n"
" \"ah_pool_inuse\":\"%d\",\n"
" \"ah_wait_list\":\"%d\"\n"
" }",
pt->fds_count,
pt->ah_count_in_use,
pt->ah_wait_list_length);
}
context deprecation 1) This makes lwsws run a parent process with the original permissions. But this process is only able to respond to SIGHUP, it doesn't do anything else. 2) You can send this parent process a SIGHUP now to cause it to - close listening sockets in existing lwsws processes - mark those processes as to exit when the number of active connections on the falls to zero - spawn a fresh child process from scratch, using latest configuration file content, latest plugins, etc. It can now reopen listening sockets if it chooses to, or open different listen ports or whatever. Notes: 1) lws_context_destroy() has been split into two pieces... the reason for the split is the first part closes the per-vhost protocols, but since they may have created libuv objects in the per-vhost protocol storage, these cannot be freed until after the loop has been run. That's the purpose of the second part of the context destruction, lws_context_destroy2(). For compatibility, if you are not using libuv, the first part calls the second part. However if you are using libuv, you must now call the second part from your own main.c after the first part.
2016-12-16 07:37:43 +08:00
buf += lws_snprintf(buf, end - buf, "]");
cgi lifecycle logging Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-20 06:10:56 +08:00
context deprecation 1) This makes lwsws run a parent process with the original permissions. But this process is only able to respond to SIGHUP, it doesn't do anything else. 2) You can send this parent process a SIGHUP now to cause it to - close listening sockets in existing lwsws processes - mark those processes as to exit when the number of active connections on the falls to zero - spawn a fresh child process from scratch, using latest configuration file content, latest plugins, etc. It can now reopen listening sockets if it chooses to, or open different listen ports or whatever. Notes: 1) lws_context_destroy() has been split into two pieces... the reason for the split is the first part closes the per-vhost protocols, but since they may have created libuv objects in the per-vhost protocol storage, these cannot be freed until after the loop has been run. That's the purpose of the second part of the context destruction, lws_context_destroy2(). For compatibility, if you are not using libuv, the first part calls the second part. However if you are using libuv, you must now call the second part from your own main.c after the first part.
2016-12-16 07:37:43 +08:00
buf += lws_snprintf(buf, end - buf, ", \"vhosts\":[\n ");
first = 1;
vh = context->vhost_list;
listening = 0;
cs = context->conn_stats;
lws_sum_stats(context, &cs);
plugin lws server status Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-15 14:01:29 +08:00
while (vh) {
context deprecation 1) This makes lwsws run a parent process with the original permissions. But this process is only able to respond to SIGHUP, it doesn't do anything else. 2) You can send this parent process a SIGHUP now to cause it to - close listening sockets in existing lwsws processes - mark those processes as to exit when the number of active connections on the falls to zero - spawn a fresh child process from scratch, using latest configuration file content, latest plugins, etc. It can now reopen listening sockets if it chooses to, or open different listen ports or whatever. Notes: 1) lws_context_destroy() has been split into two pieces... the reason for the split is the first part closes the per-vhost protocols, but since they may have created libuv objects in the per-vhost protocol storage, these cannot be freed until after the loop has been run. That's the purpose of the second part of the context destruction, lws_context_destroy2(). For compatibility, if you are not using libuv, the first part calls the second part. However if you are using libuv, you must now call the second part from your own main.c after the first part.
2016-12-16 07:37:43 +08:00
if (!hide_vhosts) {
if (!first)
if(buf != end)
*buf++ = ',';
buf += lws_json_dump_vhost(vh, buf, end - buf);
first = 0;
}
cgi lifecycle logging Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-20 06:10:56 +08:00
if (vh->lserv_wsi)
listening++;
plugin lws server status Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-15 14:01:29 +08:00
vh = vh->vhost_next;
}
context deprecation 1) This makes lwsws run a parent process with the original permissions. But this process is only able to respond to SIGHUP, it doesn't do anything else. 2) You can send this parent process a SIGHUP now to cause it to - close listening sockets in existing lwsws processes - mark those processes as to exit when the number of active connections on the falls to zero - spawn a fresh child process from scratch, using latest configuration file content, latest plugins, etc. It can now reopen listening sockets if it chooses to, or open different listen ports or whatever. Notes: 1) lws_context_destroy() has been split into two pieces... the reason for the split is the first part closes the per-vhost protocols, but since they may have created libuv objects in the per-vhost protocol storage, these cannot be freed until after the loop has been run. That's the purpose of the second part of the context destruction, lws_context_destroy2(). For compatibility, if you are not using libuv, the first part calls the second part. However if you are using libuv, you must now call the second part from your own main.c after the first part.
2016-12-16 07:37:43 +08:00
buf += lws_snprintf(buf, end - buf,
"],\n\"listen_wsi\":\"%d\",\n"
" \"rx\":\"%llu\",\n"
" \"tx\":\"%llu\",\n"
http2: make usable HTTP/2 support is now able to serve the test server, complete with websockets, from a single vhost. - This works the same with both OpenSSL and mbedTLS. - POST is now wired up and works (also for file upload). - CGI is wired up and works. - Redirect is adapted and works - lwsws works. - URI urldecode, sanitation and argument parsing wired up for :path valgrind clean (aside from openssl-style false uninit data usage in mbedtls send occasionally) h2spec reports: $ h2spec -h 127.0.0.1 -p 7681 -t -k -o 1 ... 145 tests, 145 passed, 0 skipped, 0 failed" Incorporates: - "https://github.com/warmcat/libwebsockets/pull/1039 Fixes issue with -Werror=unused-variable flag - https://github.com/sartura/libwebsockets/commit/2c843a1395bdca5c23b40d989f508fc96aa5ba3b ssl: fix infinite loop on client cert verification failure Signed-off-by: Petar Paradzik <petar.paradzik@sartura.hr>" Caused and fixes Coverity 184887 - 184892
2017-10-13 10:33:02 +08:00
" \"h1_conn\":\"%lu\",\n"
" \"h1_trans\":\"%lu\",\n"
" \"h2_trans\":\"%lu\",\n"
context deprecation 1) This makes lwsws run a parent process with the original permissions. But this process is only able to respond to SIGHUP, it doesn't do anything else. 2) You can send this parent process a SIGHUP now to cause it to - close listening sockets in existing lwsws processes - mark those processes as to exit when the number of active connections on the falls to zero - spawn a fresh child process from scratch, using latest configuration file content, latest plugins, etc. It can now reopen listening sockets if it chooses to, or open different listen ports or whatever. Notes: 1) lws_context_destroy() has been split into two pieces... the reason for the split is the first part closes the per-vhost protocols, but since they may have created libuv objects in the per-vhost protocol storage, these cannot be freed until after the loop has been run. That's the purpose of the second part of the context destruction, lws_context_destroy2(). For compatibility, if you are not using libuv, the first part calls the second part. However if you are using libuv, you must now call the second part from your own main.c after the first part.
2016-12-16 07:37:43 +08:00
" \"ws_upg\":\"%lu\",\n"
" \"rejected\":\"%lu\",\n"
http2: make usable HTTP/2 support is now able to serve the test server, complete with websockets, from a single vhost. - This works the same with both OpenSSL and mbedTLS. - POST is now wired up and works (also for file upload). - CGI is wired up and works. - Redirect is adapted and works - lwsws works. - URI urldecode, sanitation and argument parsing wired up for :path valgrind clean (aside from openssl-style false uninit data usage in mbedtls send occasionally) h2spec reports: $ h2spec -h 127.0.0.1 -p 7681 -t -k -o 1 ... 145 tests, 145 passed, 0 skipped, 0 failed" Incorporates: - "https://github.com/warmcat/libwebsockets/pull/1039 Fixes issue with -Werror=unused-variable flag - https://github.com/sartura/libwebsockets/commit/2c843a1395bdca5c23b40d989f508fc96aa5ba3b ssl: fix infinite loop on client cert verification failure Signed-off-by: Petar Paradzik <petar.paradzik@sartura.hr>" Caused and fixes Coverity 184887 - 184892
2017-10-13 10:33:02 +08:00
" \"h2_alpn\":\"%lu\",\n"
" \"h2_subs\":\"%lu\",\n"
" \"h2_upg\":\"%lu\"",
listening, cs.rx, cs.tx,
cs.h1_conn,
cs.h1_trans,
cs.h2_trans,
cs.ws_upg,
cs.rejected,
cs.h2_alpn,
cs.h2_subs,
cs.h2_upg);
cgi lifecycle logging Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-20 06:10:56 +08:00
#ifdef LWS_WITH_CGI
for (n = 0; n < context->count_threads; n++) {
pt = &context->pt[n];
pcgi = &pt->cgi_list;
while (*pcgi) {
pcgi = &(*pcgi)->cgi_list;
cgi_count++;
}
}
#endif
lws_snprintf Thanks to Fabrice Gilot for reporting the problem that led to uncovering this. Due to a misunderstanding of the return value of snprintf (it is not truncated according to the max size passed in) in several places relying on snprintf to truncate the length overflows are possible. This patch wraps snprintf with a new lws_snprintf() which does truncate its length to allow the buffer limiting scheme to work properly. All users should update with these fixes.
2016-09-15 02:22:57 +08:00
buf += lws_snprintf(buf, end - buf, ",\n \"cgi_alive\":\"%d\"\n ",
cgi lifecycle logging Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-20 06:10:56 +08:00
cgi_count);
context deprecation 1) This makes lwsws run a parent process with the original permissions. But this process is only able to respond to SIGHUP, it doesn't do anything else. 2) You can send this parent process a SIGHUP now to cause it to - close listening sockets in existing lwsws processes - mark those processes as to exit when the number of active connections on the falls to zero - spawn a fresh child process from scratch, using latest configuration file content, latest plugins, etc. It can now reopen listening sockets if it chooses to, or open different listen ports or whatever. Notes: 1) lws_context_destroy() has been split into two pieces... the reason for the split is the first part closes the per-vhost protocols, but since they may have created libuv objects in the per-vhost protocol storage, these cannot be freed until after the loop has been run. That's the purpose of the second part of the context destruction, lws_context_destroy2(). For compatibility, if you are not using libuv, the first part calls the second part. However if you are using libuv, you must now call the second part from your own main.c after the first part.
2016-12-16 07:37:43 +08:00
buf += lws_snprintf(buf, end - buf, "}");
buf += lws_snprintf(buf, end - buf, "]}\n ");
plugin lws server status Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-15 14:01:29 +08:00
return buf - orig;
}
#endif
LWS_WITH_STATS
2017-05-07 10:02:03 +08:00
#if defined(LWS_WITH_STATS)
LWS_VISIBLE LWS_EXTERN uint64_t
lws_stats_get(struct lws_context *context, int index)
{
if (index >= LWSSTATS_SIZE)
return 0;
return context->lws_stats[index];
}
LWS_VISIBLE LWS_EXTERN void
lws_stats_log_dump(struct lws_context *context)
{
stats: ah and ssl restriction stats
2017-05-15 07:30:06 +08:00
struct lws_vhost *v = context->vhost_list;
LWS_WITH_PEER_LIMITS Introduces an optional, settable restriction on the number of simultaneous wsi connections and ah that a single peer IP can obtain. The default is disabled for build, and if enabled, unlimited. However when enabled at CMake, setting info.ip_limit_ah and / or info.ip_limit_wsi at context creation time will enforce the limits. Connections past the info.ip_limit_wsi limit are dropped, and connections needing an ah are forced to stay in the ah waiting list even when ah are available, while the peer is at the number of ah in info.ip_limit_ah.
2017-09-14 13:14:11 +08:00
int n, m;
(void)m;
stats: ah and ssl restriction stats
2017-05-15 07:30:06 +08:00
LWS_WITH_STATS
2017-05-07 10:02:03 +08:00
if (!context->updated)
return;
context->updated = 0;
lwsl_notice("\n");
lwsl_notice("LWS internal statistics dump ----->\n");
lwsl_notice("LWSSTATS_C_CONNECTIONS: %8llu\n", (unsigned long long)lws_stats_get(context, LWSSTATS_C_CONNECTIONS));
lwsl_notice("LWSSTATS_C_API_CLOSE: %8llu\n", (unsigned long long)lws_stats_get(context, LWSSTATS_C_API_CLOSE));
lwsl_notice("LWSSTATS_C_API_READ: %8llu\n", (unsigned long long)lws_stats_get(context, LWSSTATS_C_API_READ));
lwsl_notice("LWSSTATS_C_API_LWS_WRITE: %8llu\n", (unsigned long long)lws_stats_get(context, LWSSTATS_C_API_LWS_WRITE));
lwsl_notice("LWSSTATS_C_API_WRITE: %8llu\n", (unsigned long long)lws_stats_get(context, LWSSTATS_C_API_WRITE));
lwsl_notice("LWSSTATS_C_WRITE_PARTIALS: %8llu\n", (unsigned long long)lws_stats_get(context, LWSSTATS_C_WRITE_PARTIALS));
lwsl_notice("LWSSTATS_C_WRITEABLE_CB_REQ: %8llu\n", (unsigned long long)lws_stats_get(context, LWSSTATS_C_WRITEABLE_CB_REQ));
lwsl_notice("LWSSTATS_C_WRITEABLE_CB_EFF_REQ: %8llu\n", (unsigned long long)lws_stats_get(context, LWSSTATS_C_WRITEABLE_CB_EFF_REQ));
lwsl_notice("LWSSTATS_C_WRITEABLE_CB: %8llu\n", (unsigned long long)lws_stats_get(context, LWSSTATS_C_WRITEABLE_CB));
wrapper: make SSL_accept nonblocking The mbedTLS / OpenSSL wrapper spins for the duration of the handshake, which is a slow 31ms on my x86_64 box (by comparison it's ~1ms on actual OpenSSL on the same box). This doesn't change the 31ms but it stops us spinning during the accept and has us retry on POLLIN instead like OpenSSL. Note this also fixes the endemic mismatches in returncode between mbedTLS and OpenSSL semantics...
2017-10-06 21:23:59 +08:00
lwsl_notice("LWSSTATS_C_SSL_CONNECTIONS_ACCEPT_SPIN: %8llu\n", (unsigned long long)lws_stats_get(context, LWSSTATS_C_SSL_CONNECTIONS_ACCEPT_SPIN));
LWS_WITH_STATS
2017-05-07 10:02:03 +08:00
lwsl_notice("LWSSTATS_C_SSL_CONNECTIONS_FAILED: %8llu\n", (unsigned long long)lws_stats_get(context, LWSSTATS_C_SSL_CONNECTIONS_FAILED));
lwsl_notice("LWSSTATS_C_SSL_CONNECTIONS_ACCEPTED: %8llu\n", (unsigned long long)lws_stats_get(context, LWSSTATS_C_SSL_CONNECTIONS_ACCEPTED));
esp32: multi ap slots
2017-05-13 10:26:59 +08:00
lwsl_notice("LWSSTATS_C_SSL_CONNS_HAD_RX: %8llu\n", (unsigned long long)lws_stats_get(context, LWSSTATS_C_SSL_CONNS_HAD_RX));
LWS_WITH_PEER_LIMITS Introduces an optional, settable restriction on the number of simultaneous wsi connections and ah that a single peer IP can obtain. The default is disabled for build, and if enabled, unlimited. However when enabled at CMake, setting info.ip_limit_ah and / or info.ip_limit_wsi at context creation time will enforce the limits. Connections past the info.ip_limit_wsi limit are dropped, and connections needing an ah are forced to stay in the ah waiting list even when ah are available, while the peer is at the number of ah in info.ip_limit_ah.
2017-09-14 13:14:11 +08:00
lwsl_notice("LWSSTATS_C_PEER_LIMIT_AH_DENIED: %8llu\n", (unsigned long long)lws_stats_get(context, LWSSTATS_C_PEER_LIMIT_AH_DENIED));
lwsl_notice("LWSSTATS_C_PEER_LIMIT_WSI_DENIED: %8llu\n", (unsigned long long)lws_stats_get(context, LWSSTATS_C_PEER_LIMIT_WSI_DENIED));
esp32: multi ap slots
2017-05-13 10:26:59 +08:00
LWS_WITH_STATS
2017-05-07 10:02:03 +08:00
lwsl_notice("LWSSTATS_C_TIMEOUTS: %8llu\n", (unsigned long long)lws_stats_get(context, LWSSTATS_C_TIMEOUTS));
lwsl_notice("LWSSTATS_C_SERVICE_ENTRY: %8llu\n", (unsigned long long)lws_stats_get(context, LWSSTATS_C_SERVICE_ENTRY));
lwsl_notice("LWSSTATS_B_READ: %8llu\n", (unsigned long long)lws_stats_get(context, LWSSTATS_B_READ));
lwsl_notice("LWSSTATS_B_WRITE: %8llu\n", (unsigned long long)lws_stats_get(context, LWSSTATS_B_WRITE));
lwsl_notice("LWSSTATS_B_PARTIALS_ACCEPTED_PARTS: %8llu\n", (unsigned long long)lws_stats_get(context, LWSSTATS_B_PARTIALS_ACCEPTED_PARTS));
lwsl_notice("LWSSTATS_MS_SSL_CONNECTIONS_ACCEPTED_DELAY: %8llums\n", (unsigned long long)lws_stats_get(context, LWSSTATS_MS_SSL_CONNECTIONS_ACCEPTED_DELAY) / 1000);
if (lws_stats_get(context, LWSSTATS_C_SSL_CONNECTIONS_ACCEPTED))
lwsl_notice(" Avg accept delay: %8llums\n",
(unsigned long long)(lws_stats_get(context, LWSSTATS_MS_SSL_CONNECTIONS_ACCEPTED_DELAY) /
lws_stats_get(context, LWSSTATS_C_SSL_CONNECTIONS_ACCEPTED)) / 1000);
esp32: multi ap slots
2017-05-13 10:26:59 +08:00
lwsl_notice("LWSSTATS_MS_SSL_RX_DELAY: %8llums\n", (unsigned long long)lws_stats_get(context, LWSSTATS_MS_SSL_RX_DELAY) / 1000);
if (lws_stats_get(context, LWSSTATS_C_SSL_CONNS_HAD_RX))
lwsl_notice(" Avg accept-rx delay: %8llums\n",
(unsigned long long)(lws_stats_get(context, LWSSTATS_MS_SSL_RX_DELAY) /
lws_stats_get(context, LWSSTATS_C_SSL_CONNS_HAD_RX)) / 1000);
LWS_WITH_STATS
2017-05-07 10:02:03 +08:00
lwsl_notice("LWSSTATS_MS_WRITABLE_DELAY: %8lluus\n",
(unsigned long long)lws_stats_get(context, LWSSTATS_MS_WRITABLE_DELAY));
lwsl_notice("LWSSTATS_MS_WORST_WRITABLE_DELAY: %8lluus\n",
(unsigned long long)lws_stats_get(context, LWSSTATS_MS_WORST_WRITABLE_DELAY));
if (lws_stats_get(context, LWSSTATS_C_WRITEABLE_CB))
lwsl_notice(" Avg writable delay: %8lluus\n",
(unsigned long long)(lws_stats_get(context, LWSSTATS_MS_WRITABLE_DELAY) /
lws_stats_get(context, LWSSTATS_C_WRITEABLE_CB)));
stats: ah and ssl restriction stats
2017-05-15 07:30:06 +08:00
lwsl_notice("Simultaneous SSL restriction: %8d/%d/%d\n", context->simultaneous_ssl,
context->simultaneous_ssl_restriction, context->ssl_gate_accepts);
lwsl_notice("Live wsi: %8d\n", context->count_wsi_allocated);
context->updated = 1;
while (v) {
WITH_STATS: remove accidental dependency on ssl
2017-06-28 09:52:22 +08:00
if (v->lserv_wsi) {
stats: ah and ssl restriction stats
2017-05-15 07:30:06 +08:00
struct lws_context_per_thread *pt = &context->pt[(int)v->lserv_wsi->tsi];
struct lws_pollfd *pfd;
pfd = &pt->fds[v->lserv_wsi->position_in_fds_table];
lwsl_notice(" Listen port %d actual POLLIN: %d\n",
v->listen_port, (int)pfd->events & LWS_POLLIN);
}
v = v->vhost_next;
}
for (n = 0; n < context->count_threads; n++) {
struct lws_context_per_thread *pt = &context->pt[n];
struct lws *wl;
int m = 0;
lwsl_notice("PT %d\n", n + 1);
lws_pt_lock(pt);
lwsl_notice(" AH in use / max: %d / %d\n",
pt->ah_count_in_use,
context->max_http_header_pool);
wl = pt->ah_wait_list;
while (wl) {
m++;
wl = wl->u.hdr.ah_wait_list;
}
lwsl_notice(" AH wait list count / actual: %d / %d\n",
pt->ah_wait_list_length, m);
lws_pt_unlock(pt);
}
LWS_WITH_PEER_LIMITS Introduces an optional, settable restriction on the number of simultaneous wsi connections and ah that a single peer IP can obtain. The default is disabled for build, and if enabled, unlimited. However when enabled at CMake, setting info.ip_limit_ah and / or info.ip_limit_wsi at context creation time will enforce the limits. Connections past the info.ip_limit_wsi limit are dropped, and connections needing an ah are forced to stay in the ah waiting list even when ah are available, while the peer is at the number of ah in info.ip_limit_ah.
2017-09-14 13:14:11 +08:00
#if defined(LWS_WITH_PEER_LIMITS)
m = 0;
WITH_STATS: fix signed build warnings
2017-11-05 09:16:42 +08:00
for (n = 0; n < (int)context->pl_hash_elements; n++) {
lws_start_foreach_llp(struct lws_peer **, peer,
context->pl_hash_table[n]) {
LWS_WITH_PEER_LIMITS Introduces an optional, settable restriction on the number of simultaneous wsi connections and ah that a single peer IP can obtain. The default is disabled for build, and if enabled, unlimited. However when enabled at CMake, setting info.ip_limit_ah and / or info.ip_limit_wsi at context creation time will enforce the limits. Connections past the info.ip_limit_wsi limit are dropped, and connections needing an ah are forced to stay in the ah waiting list even when ah are available, while the peer is at the number of ah in info.ip_limit_ah.
2017-09-14 13:14:11 +08:00
m++;
} lws_end_foreach_llp(peer, next);
}
lwsl_notice(" Peers: total active %d\n", m);
if (m > 10) {
m = 10;
lwsl_notice(" (showing 10 peers only)\n");
}
if (m) {
WITH_STATS: fix signed build warnings
2017-11-05 09:16:42 +08:00
for (n = 0; n < (int)context->pl_hash_elements; n++) {
LWS_WITH_PEER_LIMITS Introduces an optional, settable restriction on the number of simultaneous wsi connections and ah that a single peer IP can obtain. The default is disabled for build, and if enabled, unlimited. However when enabled at CMake, setting info.ip_limit_ah and / or info.ip_limit_wsi at context creation time will enforce the limits. Connections past the info.ip_limit_wsi limit are dropped, and connections needing an ah are forced to stay in the ah waiting list even when ah are available, while the peer is at the number of ah in info.ip_limit_ah.
2017-09-14 13:14:11 +08:00
char buf[72];
lws_start_foreach_llp(struct lws_peer **, peer, context->pl_hash_table[n]) {
struct lws_peer *df = *peer;
if (!lws_plat_inet_ntop(df->af, df->addr, buf,
sizeof(buf) - 1))
strcpy(buf, "unknown");
lwsl_notice(" peer %s: count wsi: %d, count ah: %d\n",
buf, df->count_wsi, df->count_ah);
if (!--m)
break;
} lws_end_foreach_llp(peer, next);
}
}
#endif
LWS_WITH_STATS
2017-05-07 10:02:03 +08:00
lwsl_notice("\n");
}
void
lws_stats_atomic_bump(struct lws_context * context,
struct lws_context_per_thread *pt, int index, uint64_t bump)
{
lws_pt_lock(pt);
context->lws_stats[index] += bump;
if (index != LWSSTATS_C_SERVICE_ENTRY)
context->updated = 1;
lws_pt_unlock(pt);
}
void
lws_stats_atomic_max(struct lws_context * context,
struct lws_context_per_thread *pt, int index, uint64_t val)
{
lws_pt_lock(pt);
if (val > context->lws_stats[index]) {
context->lws_stats[index] = val;
context->updated = 1;
}
lws_pt_unlock(pt);
}
#endif
lws_ring: introduce generic flexible ringbuffer abstraction
2017-09-20 10:37:59 +08:00
Reference in a new issue Copy permalink